Glucagon-like peptide-1 receptor agonists decrease hyperinsulinemia and hyperandrogenemia in dehydroepiandrosterone-induced polycystic ovary syndrome mice and are associated with mitigating inflammation and inducing browning of white adipose tissue

Abstract

Polycystic ovary syndrome is a complicated hormonal and metabolic disorder. The exact pathogenesis of polycystic ovary syndrome is not clear thus far. Inflammation is involved in the progression of polycystic ovary syndrome. In addition, brown adipose tissue activity is impaired in polycystic ovary syndrome. Interestingly, glucagon-like peptide-1 receptor agonists have been reported to alleviate inflammation and promote browning of white adipose tissue. In this study, the effects of glucagon-like peptide-1 receptor agonists on polycystic ovary syndrome mice were explored. Mice were randomly assigned into four groups: control, dehydroepiandrosterone, dehydroepiandrosterone + liraglutide, and dehydroepiandrosterone + semaglutide. Relative indexes were measured after glucagon-like peptide-1 receptor agonist intervention. Glucose metabolism in polycystic ovary syndrome mice was ameliorated by glucagon-like peptide-1 receptor agonists, while the reproductive endocrine disorder of polycystic ovary syndrome mice was partially reversed. The messenger ribonucleic acid levels of steroidogenic enzymes and the expression of inflammatory mediators in serum and ovaries of polycystic ovary syndrome mice were improved. Furthermore, toll-like receptor 4 and phosphorylation of nuclear factor-kappa B protein levels were decreased by glucagon-like peptide-1 receptor agonists in ovary. Notably, after glucagon-like peptide-1 receptor agonist intervention, the expression of brown adipose tissue marker levels was considerably raised in the white adipose tissue of polycystic ovary syndrome mice. In conclusion, the hyperinsulinemia and hyperandrogenemia of polycystic ovary syndrome mice were alleviated by glucagon-like peptide-1 receptor agonist intervention, which was associated with mitigating inflammation and stimulating adipose tissue browning.

GLP-1RAs could improve the symptoms of polycystic ovary mice.

Graphical Abstract

Graphical Abstract.

Introduction

Polycystic ovary syndrome (PCOS) is a serious clinical condition that is related to metabolic and reproductive dysfunction and is characterized by abnormal menstrual cycles, anovulation, insulin resistance (IR), and hyperandrogenemia (HA) [1]. HA is the primary cause of anovulatory infertility and affects approximately 5–20% of women of childbearing age [2]. Notably, although genetic and environmental effects have been found to be the main drivers of PCOS, lifestyle improvements, including diet adjustment and exercise, are strongly recommended as the first and basic step in treatment [3].

HA and IR are the two main pathogenic mechanisms of PCOS. The ovary is one of the primary sites of androgen synthesis in women. Numerous steroid hormone synthesis-related enzymes, including steroidogenic acute regulatory protein (STAR), steroid 17 alpha-hydroxylase/17,20 lyase (CYP17A1), 3 beta-hydroxysteroid dehydrogenase (3β-HSD), and cytochrome P450 family 19 subfamily A member 1 (CYP19A1), are present [4]. For example, CYP17A1, a crucial enzyme in androgen production, is considerably increased in PCOS mice [5]. Furthermore, persistent low-grade inflammation is considered to be an important trigger in the development of PCOS [3, 6]. Importantly, there is a clear association between HA and inflammation in PCOS patients [7]. It is speculated that monocytes can be recruited into the ovary in PCOS, causing local inflammation and stimulating the ovarian steroidogenic enzyme (CYP17A1) to synthesize androgen [8]. Furthermore, one study revealed that TNF-alpha blocks insulin signaling by mediating serine phosphorylation of insulin receptor substrate 1 in insulin-sensitive tissues. As a result, the transport of glucose transporter-4 (GLUT-4) is reduced [6].

Patients with PCOS have been found to have abnormal adipose tissue function, which is linked to IR. Brown (BAT), white (WAT), and beige adipose tissue (BeAT) are three different forms of adipose tissue that are crucial for preserving the body’s overall energy balance [9]. BAT is a particular kind of adipose tissue that dissipates energy as heat and is composed of packed mitochondria and smaller lipid droplets [10]. BeAT is present in WAT and has the characteristics of WAT in a resting state but manifests the characteristics of BAT in response to cold and beta-adrenergic stimulation, a process known as the browning of WAT [11]. It has been found that BAT activity is reduced in both PCOS patients and dehydroepiandrosterone (DHEA)-induced PCOS rats [12, 13]. Emerging evidence has shown that activation of endogenous BAT by cold stimulation or compounds and BAT transplantation can ameliorate PCOS metabolic and reproductive disorders in various animal models [2]. Therefore, enhancing the quality or activity of BAT could be a new possible therapy for treating PCOS.

Glucagon-like peptide-1 (GLP-1) is an incretin released by intestinal L cells in the ileum and distal colon. GLP-1 receptor agonist (GLP-1RA) is a novel hypoglycemic agent that promotes insulin production in a glucose-dependent manner [14]. Liraglutide and semaglutide are common GLP-1 receptor agonists (GLP-1RAs) in clinical treatment [15]. Liraglutide has a half-life of 13 h in humans and is administered daily, while semaglutide has a longer half-life and is administered weekly. GLP-1RA has been found to reduce the levels of serum inflammatory cytokines in acute inflammatory models and downregulate several inflammatory signaling pathways [16]. It has also been demonstrated that peripheral administration of GLP-1RA can ameliorate obesity in high-fat-diet mice by promoting browning of WAT [17]. Furthermore, a few clinical studies have shown that GLP-1RA can improve menstrual irregularities in PCOS patients [18, 19]. The above evidence suggests that GLP-1RA can ameliorate reproductive and metabolic disturbances in PCOS by mitigating inflammation and stimulating WAT browning. However, little is known about the precise function and intricate mechanism of GLP-1RA in the treatment of PCOS.

In this research, we used DHEA to induce a PCOS model and demonstrated that GLP-1RAs ameliorated hyperinsulinemia and HA in PCOS mice, which was associated with mitigating inflammation and stimulating adipose tissue browning.

Materials and methods

Experimental animals

Female prepubertal C57BL/6 mice (3 weeks old, 10–12 g) were obtained from Shanghai SLAC Laboratory Animal Co. Ltd. (License No. SCXK (Shanghai) 2022-0004), maintained at 22–25°C under a 12 h light:12 h dark cycle, and fed freely. All animal experimental protocols were approved by the ethics committee of Animal Experiments of East China Normal University. At postnatal day 25, all mice were arbitrarily assigned to one of four groups: control, DHEA, DHEA + liraglutide, and DHEA + semaglutide. As mentioned above [20], PCOS mice (DHEA, DHEA + liraglutide, and DHEA + semaglutide) were induced with DHEA (dissolved in sesame oil, 6 mg/100 g body weight; Sigma-Aldrich), and control group mice were given the same dose of sesame oil for 3 weeks followed by four consecutive weeks of treatment. For treatment, mice in the control and DHEA groups were injected with normal saline, and mice in the DHEA + liraglutide group were intraperitoneally administered liraglutide (0.3 mg/kg body weight) [21] every day for 4 weeks, and mice in the DHEA + semaglutide group were intraperitoneally administered with semaglutide (0.1 mg/kg body weight) [22, 23] three times each week for 4 weeks. The glucose tolerance test (GTT) was assessed on the 40th day of the experiment, and the insulin tolerance test (ITT) was performed on the 44th day of the experiment. Vaginal smears were collected 8 days before the end of the intervention and continued to be collected until the end of the intervention. After the intervention, blood samples, ovarian tissues, and adipose tissues were obtained.

Assessment of the estrous cycle

Vaginal smears were performed at 9:00 every day from days 41 to 48 of the experiment. Under the microscope, the major cell types were identified in vaginal smears that had been stained with Wright’s Giemsa stain. The major cell types observed by microscopy were used to determine the current stage of the estrous cycle among the mice. As previously described [24], the majority of the cells in proestrus were rounded, nucleated epithelial cells. Cornified epithelial cells predominate during estrus. Metestrus was characterized by leukocytes and epithelial cells. Diestrus was characterized mostly by leukocytes and nucleated epithelial cells.

GTT and ITT

Prior to the GTT trial, mice were given a 12-h fast and had free access to water. Glucose was tested from blood collected from the tail vein and by utilizing an automatic reader and test strips (Bayer, Contour TS Blood Glucose Meter). After fasting glucose levels were measured, the mice received an intraperitoneal injection of D-glucose (2.0 g/kg body weight), and their glucose levels were checked 15, 30, 60, 90, and 120 min later. For ITT analysis, mice were given intraperitoneal injections of insulin (1.0 U/kg) after a 4-h fast, and glucose levels were checked at 0, 15, 30, 60, 90, and 120 min.

Serum analysis

The serum was extracted by centrifuging the mouse’s blood, and the serum was kept at −80°C. The levels of testosterone (ab108666; Abcam) and insulin (90080; Crystal Chem) were determined using enzyme-linked immunosorbent assay (ELISA) kits. The sensitivity of the ELISA kit for testosterone is 0.07 ng/ml and the sensitivity of the ELISA kit for insulin is 50 pg/ml. Mouse serum levels of follicle-stimulating hormone (FSH), luteinizing hormone (LH), sex hormone-binding globulin (SHBG), interleukin-6 (IL-6), interleukin-6(IL-10), and adiponectin were analyzed using ELISA kits (Jiangsu Jingmei Biological Technology Co. Ltd, China). The homeostatic model assessment of the IR (HOMA-IR) index was evaluated by the equation: [fasting glucose levels (mmol/l)] × [fasting serum insulin (IU/ml)]/22.5. The following equation was used to determine the free androgen index (FAI): [testosterone (nmol/l)]/[SHBG (nmol/l)] × 100.

Histological analysis

For hematoxylin and eosin (H&E) staining, the ovarian and adipose tissues of the mice were fixed with formalin and embedded in paraffin followed by sectioning at 5 μm. Hematoxylin and eosin were used to stain the sections.

For immunohistochemistry, adipose tissue slices were treated with bovine serum albumin for 10 min and then incubated overnight with UCP1 primary rabbit polyclonal antibody (1:200; GB112174; Servicebio) or rabbit anti-F4/80 antibody (1:1000; GB11027; Servicebio). Then, goat anti-rabbit IgG conjugated with horseradish peroxidase (HRP) was cultured for 1 h. Images were acquired using an Olympus microscope.

Real-time quantitative PCR

The TRIZOL reagent (Invitrogen) was used to isolate total ribonucleic acid (RNA). A complementary DNA (cDNA) reverse transcription kit (TransGen Biotech Co., Ltd.) was used to generate cDNA. SYBR Green Master Mix (Vazyme Biotech Co., Ltd.) was utilized for real-time quantitative PCR (Q5, Thermo Fisher), and the expression of the housekeeping gene glyceraldehyde-3-phosphate dehydrogenase (Gapdh) was used as a standard. The primer sequences are listed in Table 1.

Table 1.

Primer sequences for RT-PCR

Gene	Forward primer (5′ → 3′)	Reverse primer (5′ → 3′)
Gapdh	TGGCCTTCCGTGTTCCTAC	GAGTTGCTGTTGAAGTCGCA
Cyp11a1	GGATGCGTCGATACTCTTCTCA	GGACGATTCGGTCTTTCTTCCA
Cyp17a1	GTCGCCTTTGCGGATAGTAGT	TGAGTTGGCTTCCTGACATATCA
Cyp19a1	CATGGTCCCGGAAACTGTGA	CAGGCACTTCCAATCCCCAT
Hsd3b	GCCTGTATCCAAGCCAGTGT	TGCTCTTCCTCGTTGCCATT
Hsd17b	TTTGGGCCGCTAGAAGCAC	CAGTACACTTCGTGGAATGGC
Star   Setdb1   Slc44a1	ATGTTCCTCGCTACGTTCAAG TGCCTATCCAAACCGCCCAATG CGATGGTAGGTGGTGAGGAGGAG	CCCAGTGCTCTCCAGTTGAG GCTGCCATCCACCTCTTCAACTC GCAGTCAGTGAAGACACAGACAGAG
Il1b	GAAATGCCACCTTTTGACAGTG	TGGATGCTCTCATCAGGACAG
Il6	CTGCAAGAGACTTCCATCCAG	AGTGGTATAGACAGGTCTGTTGG
Il10	GCTGGACAACATACTGCTAACC	ATTTCCGATAAGGCTTGGCAA
Tnf	GATCGGTCCCCAAAGGGATG	CCACTTGGTGGTTTGTGAGTG
Ccl2	TTAAAAACCTGGATCGGAACCAA	GCATTAGCTTCAGATTTACGGGT
Dio2	CAGCTTCCTCCTAGATGCCTA	CTGATTCAGGATTGGAGACGTG
Ucp1	CAAAAACAGAAGGATTGCCGAAA	TCTTGGACTGAGTCGTAGAGG
Pgc1a	TATGGAGTGACATAGAGTGTGCT	CCACTTCAATCCACCCAGAAAG
Nos2	GTTCTCAGCCCAACAATACAAGA	GTGGACGGGTCGATGTCAC
Itgax	CTGGATAGCCTTTCTTCTGCTG	GCACACTGTGTCCGAACTCA
Arg1	CTCCAAGCCAAAGTCCTTAGAG	AGGAGCTGTCATTAGGGACATC
Mrc1	CTCTGTTCAGCTATTGGACGC	TGGCACTCCCAAACATAATTTGA

Western blotting

Total protein was extracted using radioimmunoprecipitation assay lysis buffer (Beyotime Institute of Biotechnology) added with protease inhibitors (Sigma) and phosphatase inhibitors (Beyotime Institute of Biotechnology). Protein concentrations were measured via a bicinchoninic acid protein assay kit (Beyotime Institute of Biotechnology). Protein from each sample was isolated by 10% sodium dodecyl sulfate polyacrylamide gel electrophoresis and then transferred to a polyvinylidene difluoride (PVDF) membrane (Millipore). The PVDF membrane was blocked with skim milk for an hour, followed by overnight incubation with the primary antibody and an hour-long incubation with the secondary antibody coupled to HRP. The primary antibodies utilized in this research included the following: anti-UCP1 (1:1000; Abcam), anti-inducible nitric oxide synthase (iNOS/nitric oxide synthase 2, inducible (Nos2); 1:1000; Abcam), anti-ARG1 (1:1000; CST), anti-Cleaved-Caspase 3 (1:1000; CST), anti-Caspase 3 (1:1000; CST), anti-toll-like receptor 4 (TLR4; 1:1000; CST), anti-phosphorylation of nuclear factor-kappa B (p-NF-κB; 1:1000; CST), anti-NF-κB (1:1000; Abcam), and anti-GAPDH (1:20 000; CST). Details of the antibodies are listed in Table 2. X-Ray film was used to detect signals by enhanced chemiluminescence. The protein expression of each band was analyzed by ImageJ software.

Table 2.

Details of the antibodies used in western blotting

Antibody	Company	Cat. no.	Species	Dilution
UCP1	Abcam	ab234430	Rabbit	1:1000
iNOS	Abcam	ab283655	Rabbit	1:1000
ARG1	CST	#93668	Rabbit	1:1000
Cleaved-Caspase 3	CST	#9664T	Rabbit	1:1000
Caspase 3	CST	#14220	Rabbit	1:1000
TLR4	CST	#14358	Rabbit	1:1000
p-NF-κB	CST	#3033T	Rabbit	1:1000
NF-κB	Abcam	ab32536	Rabbit	1:1000
GAPDH	CST	#5174	Rabbit	1:20 000
Anti-rabbit IgG (HRP)	CST	#7074	Goat	1:5000

Statistical analysis

Data were analyzed using GraphPad Prism 8.0. One-way ANOVA with Dunnett’s post hoc test or two-way ANOVA with Tukey’s post hoc test was used to evaluate the data. All statistics are presented as mean ± SEM. P value <0.05 was considered statistically significant.

Results

GLP-1RAs alleviate metabolic disorders in PCOS mice

The experimental procedure is presented in Figure 1A. Considering the appetite inhibitory of GLP-1 [25], we measured the body weight and food intake among these groups. As expected, both liraglutide and semaglutide reduced the increase in body weight caused by DHEA (liraglutide: 9% decrease, semaglutide: 14% decrease), which was consistent with another study [26] (Figure 1B). Then, we measured food intake and found that liraglutide and semaglutide reduced the increase in food intake induced by PCOS mice and that there was no difference in food intake between PCOS mice treated with GLP-1RAs and control mice (Figure 1C). These data suggested that liraglutide and semaglutide may improve PCOS mice metabolism.

Figure 1.

GLP-1RAs alleviate metabolic disorders in PCOS mice. (A–C) Three weeks after modeling, PCOS mice were treated with GLP-1 receptor agonists for 4 weeks. (A) Experimental flow chart, (B) body weight, (C) food intake. (D) Fasting glucose levels. (E) Fasting serum insulin levels. (F) HOMA-IR. (G) Glucose tolerance test (GTT). (H) Insulin tolerance test (ITT). n = 5–7 per group. Data are expressed as mean ± SEM. ^*P < 0.05, ^**P < 0.01.

To confirm the above hypothesis, we measured fasting glucose and insulin levels. Although there was no change in fasting glucose, both liraglutide and semaglutide decreased fasting insulin to normal levels (Figure 1D and E). Notably, both liraglutide and semaglutide significantly reduced HOMA-IR values (Figure 1F). Furthermore, liraglutide ameliorated the impaired glucose tolerance caused by DHEA in mice, and semaglutide showed a trend toward ameliorating glucose tolerance (Figure 1G). In addition, ITT revealed that the rate of decline in glucose slowed down and that the area under the curve increased in PCOS mice, suggesting that PCOS mice had reduced insulin sensitivity. Interestingly, the administration of GLP-1RAs ameliorated IR in DHEA-treated mice (Figure 1H).

GLP-1RAs improve ovarian morphology and serum sex hormone disorders in PCOS mice

Considering the vital role of estrous cycle in reproduction, vaginal smears were obtained over the course of eight consecutive days in each group. Figure 2A depicted the vaginal smears in different phases of the estrous cycle. As shown in Figure 2B, a normal regular estrous cycle was 4–5 days in length, while mice in the DHEA group were all in the diestrus. As we predicted, GLP-1RAs decreased the disorders of the estrous cycle. Histological analysis showed that the ovaries of the control group of mice had follicles at different developmental stages and no cyst-like follicles. However, the ovarian structure in PCOS mice was abnormal, characterized by fewer corpus luteum, more cystic follicles, and thinner granulosa cell layers in the follicles. Importantly, GLP-1RA administration increased the quantity of corpus luteum and decreased the quantity of cystic follicles, suggesting that GLP-1RAs reversed pathological development in ovarian tissue (Figure 2C–E).

Figure 2.

GLP-1RAs improve ovarian morphology in PCOS mice. (A) Vaginal smears. The normal estrous cycle sequence is proestrus, estrus, metestrus, and diestrus. (B) Estrous cycle. The estrous cycle curves were drawn based on vaginal smears collected on 8 consecutive days in each group of mice. P: prestrus, E: estrus, M: metestrus, D: diestrus. (C) Ovarian morphology by H&E staining. (D) Corpus luteum number. (E) Cystic follicle number. n = 6 per group. Data are expressed as mean ± SEM. ^*P < 0.05, ^**P < 0.01.

Given the important role of cell apoptosis in PCOS, we examined the expression levels of Caspase 3, which is the major promotor of apoptosis, and increased cleaved-Caspase 3 levels were strongly linked with apoptosis [27]. Molecular analysis indicated that the Cleaved-Caspase 3 protein level was notably increased in PCOS mice, but reduced after treatment with GLP-1RAs (Figure 3A and B). Sex hormone abnormalities are another important characteristic of PCOS. Thus, we measured the serum levels of testosterone, SHBG, LH, and FSH. Compared with the control group, the levels of serum SHBG in PCOS mice were lower, and the levels of serum testosterone and FAI were higher. GLP-1RAs markedly reduced serum testosterone levels and FAI values; however, they did not alter serum SHBG (Figure 3C–E). Although FSH levels did not differ between these groups (Figure 3F), the levels of LH were increased in PCOS mice. The LH/FSH ratio was also increased in PCOS mice. Notably, GLP-1RAs reversed high LH levels and high LH/FSH ratio (Figure 3G and H). Collectively, these results indicated that GLP-1RAs can significantly reverse the irregular estrous cycle and abnormal ovarian morphology in PCOS mice.

Figure 3.

GLP-1RAs improve serum sex hormone abnormalities in PCOS mice. (A) Specific antibodies against Cleaved-Caspase 3, Caspase 3, and GAPDH were used for Western blotting of ovarian proteins. (B) Protein quantification with Cleaved-Caspase 3 to Caspase 3 ratios. (C) Serum SHBG. (D) Serum testosterone. (E) FAI. (F) Follicle-stimulating hormone (FSH). (G) Luteinizing hormone (LH). (H) Serum LH/FSH ratio. n = 6 per group. Data are expressed as mean ± SEM. ^*P < 0.05, ^**P < 0.01.

GLP-1RAs regulate the expression of ovarian steroidogenic enzymes in PCOS mice

STAR mediates the transference of cholesterol from the mitochondrial outer membrane to the inner membrane, which is an important rate-limiting step in the synthesis of steroids. The messenger RNA (mRNA) levels of STAR were elevated in PCOS mice, and this was reversed by treatment with GLP-1RAs (Figure 4A). Cytochrome P450 11A1 (CYP11A1) converts cholesterol to pregnenolone. Here, its expression was significantly increased in DHEA-induced PCOS mice and reversed by GLP-1RA treatment (Figure 4B). CYP17A1 mediates the conversion of pregnenolone to synthesize 17 beta-hydroxypregnenolone via 17-hydroxylation, which is converted to DHEA by the action of 17,20-lyase. The levels of cytochrome P450, family 17, subfamily a, polypeptide 1 (Cyp17a1) were considerably induced in PCOS mice compared to those in the control group. GLP-1RAs reduced Cyp17a1 levels to near-normal levels (Figure 4C). Through the actions of 3β-HSD and 17 beta-hydroxysteroid dehydrogenase (17β-HSD), DHEA is transformed into androstenedione and ultimately into testosterone. In our study, hydroxy-delta-5-steroid dehydrogenase, 3 beta (Hsd3b) and hydroxysteroid 17-beta (Hsd17b) were increased in PCOS mice, and this was reversed to normal levels with the intervention of GLP-1RAs (Figure 4D and E). GLP-1RAs treatment also inhibited the increase in cytochrome P450, family 19, subfamily a, polypeptide 1 (Cyp19a1) expression induced by DHEA (Figure 4F). SET domain bifurcated histone lysine methyltransferase 1 (SETDB1) is a histone methyltransferase and solute carrier family 44 member 1 (SLC44A1) is a choline transporter. Both of them are common ovarian genes, and there is no significant difference in the expression of these two genes in each group of ovaries (Figure 4G and H).

Figure 4.

GLP-1RAs regulate the expression of ovarian steroidogenic enzymes in PCOS mice. Relative levels of mRNA: (A) steroidogenic acute regulatory protein (Star), (B) cytochrome P450, family 11, subfamily a, polypeptide 1 (Cyp11a1), (C) Cyp17a1, (D) Hsd3b, (E) Hsd17b, (F) Cyp19a1, (G) SET domain, bifurcated 1 (Setdb1), and (H) Slc44a1in the ovaries from control, DHEA, DHEA + liraglutide, and DHEA + semaglutide mice (n = 6/group). The data are presented as mean ± SEM. ^*P < 0.05, ^**P < 0.01.

GLP-1RAs attenuate inflammatory response and promote the polarization of M2 macrophages in PCOS mice

The levels of inflammatory cytokines in the ovarian tissue were measured to evaluate the effect of liraglutide and semaglutide in PCOS mice. The expression of interleukin 1 beta (Il1b), interleukin 6 (Il6), tumor necrosis factor (Tnf), and chemokine (C-C motif) ligand 2 (Ccl2) was obviously increased by DHEA, while the levels of interleukin 10 (Il10) were decreased. GLP-1RA treatment significantly reduced Il1b, Il6, Tnf, and Ccl2 levels while increasing the expression of Il10 (Figure 5A). Similar to gene expression, GLP-1RAs decreased serum IL-6 levels to normal (Figure 5B) and increased serum IL-10 levels (Figure 5C). TLR4 protein expression was also increased in PCOS mice, and GLP-1RAs reversed this change (Figure 5D and E). The levels of p-NF-κB/NF-κB in PCOS mice were significantly higher than those in the control group, whereas the levels were decreased after GLP-1RA intervention (Figure 5F). Macrophage infiltration of WAT is associated with chronic inflammatory response. As shown in Figure 6A, compared with the control group, there was a clear increase in the number of infiltrating macrophages in the WAT of PCOS mice, and GLP-1RA treatment reduced the number of infiltrating macrophages. The ovarian protein levels of ARG1 were much lower in the PCOS group than the control group, while the levels of iNOS protein expression were significantly higher. Notably, after treatment with GLP-1RAs, the expression levels of iNOS decreased, while the expression levels of ARG1 increased (Figure 6B and C). In comparison to the control group, PCOS mice had higher levels of Nos2 (encoding the iNOS protein) and integrin alpha X (Itgax) (encoding the CD11c surface protein) mRNA expression, but arginase1 (Arg1) and mannose receptor, C type 1 (Mrc1) (encoding the CD206 protein) expression levels were lower. GLP-1RA administration reduced Nos2 and Itgax expression while increasing Arg1 and Mrc1 expression (Figure 6D). These data suggested that GLP-1RAs attenuated the inflammatory response and promoted the polarization of M2 macrophages in PCOS mice.

Figure 5.

GLP-1RAs attenuate inflammatory response in PCOS mice. (A) Relative mRNA levels of inflammatory cytokines in ovarian tissue. (B) Serum IL-6 and (C) IL-10 levels were measured by ELISA. (D) Expression levels of TLR4 (E) and phosphorylation levels of NF-κB (F) in ovarian tissues. (n = 6/group). The data are presented as mean ± SEM. ^*P < 0.05, ^**P < 0.01.

Figure 6.

GLP-1RAs promote the polarization of M2 macrophages in PCOS mice. (A) Infiltration of macrophages in adipose tissue. (B) Expression of M1 (iNOS) and M2 (ARG1) marker proteins of macrophages in adipose tissue. (C) Protein quantification of iNOS and ARG1. (D) Relative mRNA levels of M1 (Nos2 and Itgax) and M2 (Arg1 and Mrc1) marker genes of macrophages in adipose tissue (n = 6/group). The data are presented as mean ± SEM. ^*P < 0.05, ^**P < 0.01. →: F4/80 positive macrophages.

GLP-1RAs promote the browning of WAT in PCOS mice

Given the browning of WAT by the GLP-1RAs, H&E staining was used to measure adipocyte size. As shown in Figure 7A, the adipocytes in PCOS mice were larger than those in the control group, and GLP-1RAs reduced the adipocyte size in PCOS mice. The expression of the BAT-specific genes uncoupling protein 1 (Ucp1), deiodinase, iodothyronine, type II (Dio2), and peroxisome proliferative activated receptor, gamma, coactivator 1 alpha (Pgc1a) in WAT was then assessed. Interestingly, in PCOS mice, the expression of all genes was dramatically reduced, but GLP-1RAs administration significantly raised the expression of these genes (Figure 7B). In line with the gene expression, the protein expression of UCP1 was considerably increased in the GLP-1RA intervention group compared with the DHEA group (Figure 7C–E). Research had shown that BAT transplantation improved the phenotype of PCOS mice by raising serum adiponectin levels [12]. In PCOS mice, adiponectin levels were reduced and intervention with GLP-1RAs increased adiponectin levels in PCOS mice (Figure S1). Taken together, the above results emphasized that the positive effects of GLP-1RAs were partially mediated by promoting the browning of WAT.

Figure 7.

GLP-1RAs promote the browning of WAT in PCOS mice. (A) H&E staining shows the size of adipocytes. (B) Relative mRNA levels of BAT marker genes in WAT. (C) Expression levels of UCP1 in adipose tissue. (D) Protein quantification with UCP1. (E) Immunohistochemical images showing UCP1 expression levels in adipose tissue (n = 6/group). The data are presented as mean ± SEM. ^*P < 0.05, ^**P < 0.01.

Discussion

PCOS is a serious condition related to reproductive, metabolic, and psychological defects, in which HA and IR play fundamental roles in its pathogenesis [28]. PCOS not only affects women of childbearing age but also leads to a series of complications, such as endometrial cancer and cardiovascular disease, in the later stage [29]. Recent clinical studies have shown that GLP-1RA treatment for PCOS patients may moderately reduce their androgen level; however, its mechanism remains unclear [30, 31]. To better understand the pathogenesis of PCOS, mice were injected with DHEA to induce PCOS-like symptoms, including anovulation, polycystic ovaries, and HA [32]. In this study, we found that treatment with GLP-1RAs can decrease hyperinsulinemia and HA in PCOS mice and that the protective benefits of GLP-1RAs in PCOS mice are associated with reduced inflammation and promotion of the browning of WAT.

Approximately 50% of women with PCOS are reported to be overweight or obese. Obesity, especially abdominal obesity, is strongly linked to infertility, HA, and hyperinsulinemia [33]. Consistent with a previous study, we found that administration of GLP-1RAs decreased PCOS mice’s weight, and this reduction may be due to the inhibition of appetite [34]. We also found that GLP-1RAs reduced the levels of fasting insulin and HOMA-IR and improved IR, which may be beneficial for weight loss [35]. IR-compensatory hyperinsulinemia enhances the stimulatory effect of LH on androgen production in theca cells [6]. In addition, some studies have revealed that insulin reduces the levels of SHBG released by the liver, resulting in an increase in free testosterone in serum and eventually leading to hyperandrogenism [36].

It is widely acknowledged that HA is closely linked to most of the reproductive and metabolic disorders connected to PCOS, even though the pathophysiology of the condition is still not fully clear [36]. Our study revealed that GLP-1RA treatment can reverse the increase in testosterone levels in PCOS mice. The LH/FSH ratio, an indicator in the clinical diagnosis of PCOS, was also reduced after liraglutide and semaglutide treatment. A previous meta-analysis has also provided evidence that GLP-1RA treatment can decrease serum testosterone and increase SHBG levels to relieve HA in women with PCOS [37]. HA reduces GLUT-4 expression and increases central obesity, which leads to IR [38]. Therefore, the development of PCOS can be regarded as a vicious cycle where HA contributed to IR, hyperinsulinemia, and enlarged adipocytes; these outcomes in turn facilitate the synthesis and secretion of androgen [39].

The synthesis and secretion of steroids are crucial for the follicle growth. Some research have found that the mRNA levels of steroid synthesis enzymes are significantly increased in PCOS mice, and our results agree with these studies [40, 41]. Global gene expression was not altered in our treatment groups as evidenced by the consistent expression levels of Setdb1 and solute carrier family 44, member 1 (Slc44a1), suggesting our treatment effects were specific for PCOS induction and subsequent amelioration of its negative impact. However, other studies have shown that the mRNA levels of steroid synthesis enzymes are decreased in PCOS mice due to the negative feedback of increased DHEA levels [4, 42]. Thus, changes in the mRNA levels of steroid synthesis enzymes in PCOS mice are controversial and require further study.

In addition, inflammation is associated with a variety of diseases. A large number of researchers have reported that chronic low-grade inflammation and metabolic disorders jointly promote the occurrence and development of PCOS [43–45]. Qi et al. [46] found that the levels of inflammatory cytokines in serum increased in mice treated with DHEA. The serum levels of TNF-alpha and IL-6 in women with PCOS were also considerably increased [7]. Several studies have reported the activation and high expression of upstream molecules in the inflammatory response signaling pathway in patients with PCOS and in animal models [47–49]. TLR4/NF-κB signaling mediates pathophysiological processes in a variety of inflammatory injuries [50, 51]. Liu et al. [52] found that the expression of TLR4 and phosphorylation of NF-κB were significantly increased in rats with PCOS. Interestingly, the phenotypes of PCOS were also significantly alleviated when TLR4/NF-κB signaling was suppressed by different agents [47, 52]. Consistent with these results, we discovered that PCOS mice had increased levels of pro-inflammatory cytokines in their ovarian tissues while they had lower levels of anti-inflammatory cytokines like IL-10. We also found that the TLR4/NF-κB signaling pathway was activated in the ovarian tissue of PCOS mice.

Inflammation is closely associated with obesity and IR [6, 53]. Considering that more than half of PCOS patients are overweight, even obese, and have higher serum inflammatory cytokine levels, it is reasonable to assume that adipose tissue inflammation associated with macrophage status is involved in PCOS. A previous study proved that M1 macrophages promote inflammation while M2 macrophages inhibit it [54, 55]. Specifically, iNOS and CD11c are classic markers of M1 macrophages, whereas M2 macrophages are identified by ARG1 and CD206 [56]. Most studies [3, 46, 57] involving the inflammatory response of PCOS suggested that there are a large number of M1 macrophages in the adipose tissue and peripheral blood of PCOS patients and animal models. Therefore, reversing macrophage polarization in adipose tissue may be a promising target for improving inflammation and IR in PCOS patients.

On the other hand, browning of WAT has a protective effect on multiple organ systems [58, 59]. In addition, BAT has a significant impact on enhancing energy metabolism and reducing IR [60, 61]. In contrast, WAT in the viscera increases the body’s risk of metabolic diseases [62]. The present research revealed the expression of BAT-related genes was dramatically decreased in the adipose tissue of PCOS mice, while the volume of white adipocytes was significantly increased. Notably, researchers have found that enhancing BAT function can ameliorate the phenotypes of PCOS [2]. A previous study had suggested that liraglutide may alleviate phenotypes of PCOS mice by increasing the expression of adiponectin [63]. This finding was supported by our study, which also demonstrated that liraglutide treatment can improve PCOS mice’s reproductive and metabolic disorders. Meanwhile, treatment with both liraglutide and semaglutide increased adiponectin levels in PCOS mice. Adiponectin has been reported to mediate the browning of WAT [64]. Moreover, recent studies have reported that GLP-1RA participated in the browning of adipose tissue and promote the thermogenesis of BAT [17, 65]. In line with previous findings, we found GLP-1RAs liraglutide and semaglutide promoted the browning of WAT in PCOS mice. Therefore, we hypothesized that GLP-1RAs may promote browning of WAT by increasing adiponectin level. However, adipose tissue inflammation inhibits WAT browning, which is regulated by the interaction between adipocytes and macrophages depending on integrin-α4 and vascular cell adhesion molecule 1 [66]. Macrophages and TNF-alpha also suppress adipocyte UCP1 expression by augmenting the extracellular signal-related kinase (ERK) pathway [67]. Guo et al. [68] found that GLP-1RAs not only attenuated macrophage infiltration but also decreased the secretion of inflammatory cytokines from macrophages. Consistent with previous findings, we found that both liraglutide and semaglutide reduced pro-inflammatory cytokine expression and induced anti-inflammatory cytokine expression. We also found that GLP-1RAs significantly decreased the markers of M1 macrophages and increased the markers of M2 macrophages in adipose tissue.

According to a pilot randomized clinical investigation [69], the increases in the pregnancy rates of infertile women with obesity and PCOS treated with the combination of metformin and liraglutide were higher than those of women treated with metformin alone. Another prospective clinical study [70] provided evidence that PCOS patients had lower levels of GLP-1 both during fasting and after meals than healthy women. Although clinical studies have confirmed that GLP-1 has a certain effect on improving PCOS phenotypes, the mechanism has not been thoroughly explored. Our research revealed the anti-inflammatory effects of GLP-1RAs and their role in stimulating the browning of WAT, thus revealing the possible mechanism of GLP-1RAs treatment in improving PCOS.

Our study had the following limitations. First, our study revealed that the expression of TLR4/NF-κB, an inflammatory signaling pathway, is enhanced in PCOS mice. We still need to further explore whether other signaling pathways mitigate inflammation after treatment with GLP-1RAs. Second, our study revealed that GLP-1RAs treatment can alleviate the symptoms of PCOS mice by relieving inflammatory response and stimulating the browning of WAT, but the cross-talk between inflammatory response and browning of WAT has not been thoroughly investigated. Finally, this study has the inherent limitations of animal experiments, and we will carry out clinical multicenter prospective research in the future.

Conclusion

The current study shows that GLP-1RAs relieve reproductive and metabolic disturbances in PCOS mice by attenuating local (ovarian) and systemic inflammatory responses and stimulating the browning of WAT. GLP-1RAs are considered a “panacea” because of their effects against inflammation, on insulin sensitivity, on ovarian morphology, against HA, and against other pathophysiological mechanisms of PCOS. GLP-1RA may be a promising treatment for PCOS in the future.

Authors’ contributions

YM, ZZ, and XD: conceptualization, funding acquisition, and project administration; YY and NL: data curation; XW and YM: formal analysis; GL: methodology; YL: resources; YZ: software; YM: supervision; YZ and XD: validation; CX and YY: visualization; YZ: writing—original draft; YL and GL: writing—review and editing.

Conflict of interest: The authors have declared that no conflict of interest exists.

Data availability

Data supporting the results of this study can be obtained from the corresponding authors.