Inulin and metformin ameliorate polycystic ovary syndrome via anti-inflammation and modulating gut microbiota in mice

Abstract

Polycystic ovary syndrome (PCOS) represents an endocrine disorder, which is closely related with gut microbiota. Inulin, a kind of probiotics, has been proven to alleviate gut microbiota dysbiosis. Metformin, a biguanide agent, shows beneficial effects on chronic metabolic diseases. Our objective was to assess the effects and associated mechanisms of inulin and metforin on attenuation of PCOS in mice. Mice were divided into 4 groups: control group (CON), model group (MOD), inulin group (INU), metformin group (MET). The last three groups were fed 6 mg of dehydroepiandrosterone (DHEA) per 100 g body weight and 60% high-fat diet to generate mice model. After 21 days of intervention, mice were euthanized and associated indications were investigated. Body weight (BW) and testosterone (T) levels were significantly decreased, but estradiol (E2) levels were increased in INU or MET group, respectively. Ovary HE staining demonstrated that inulin or metformin ameliorated PCOS morphology. Inflammatory indicators from plasma and ovary including TNF-α, IL-6, and IL-17A were decreased in INU or MET group. Moreover, IL-10 in ovary of INU or MET group was increased. Sequencing and analysis of gut microbiota showed that compared to MOD group, Bifidobacterium was increased, but Proteobacteria, Helicobacter and Parasutterella were decreased in INU group. Helicobacter was decreased in MET group. Correlation analysis showed that gut microbiota was correlated with inflammatory factors. Our results revealed that inulin and metformin alleviated PCOS via anti-inflammation and modulating gut microbiota, which may contribute to potential clinical therapy for the disease.

POLYCYSTIC OVARY SYNDROME (PCOS) represents one of the most common endocrine disorders in reproductive-aged women, of which the prevalence is reported 6–18% worldwide [1]. The complicated clinical manifestations of PCOS include oligo-/anovulation, hyperandrogenism, insulin resistance, polycystic ovaries [2, 3], as well as a significantly increased risk of diabetes, cardiovascular disease, metabolic syndrome, endometrial cancer and chronic inflammatory disease [4]. Currently, the exact pathogenesis of PCOS still remains unclear. There is still no effective clinical applications or drugs used in clinic. Thus, the novel treatment of PCOS is urgently needed.

Accumulating studies have demonstrated that gut microbiota plays an important role in the progression and pathogenesis of PCOS [4-7]. In a rat model of letrozole-induced PCOS, the composition of gut microbiota was significantly altered inclusive of Lactobacillus, Ruminococcus, Clostridium and Prevotella. Further intervention with both Lactobacillus and fecal microbiota transplantation (FMT) of healthy rats showed beneficial for the treatment of PCOS [5]. The severity of insulin resistance and menstrual disorders as well as high levels of androgens in patients with PCOS are closely linked to alteration of gut microbiota [6]. A high fat diet and hyper-androgenic state induce LPS production from abnormal gut microbiota, causing systemic and ovarian inflammation in PCOS [7, 8].

Chronic inflammation also plays a crucial role in the pathogenesis of PCOS [7, 9, 10]. Multiple inflammatory indicators involving tumor necrosis factor-alpha (TNF-α), interleukin-6 (IL-6), and C-reactive protein (CRP), have remarkably increased in PCOS [9]. Moreover, occurrence and development of chronic inflammation in PCOS are closely linked with altered gut microbiota [10]. Ecological imbalance of gut microbiota leads to increased intestinal permeability. Intestinal-derived LPS enters into the circulation through the “intestine leaking” wall to induce systemic low-grade inflammation, leading to increased testosterone production of ovaries in the formation of PCOS [7, 11, 12].

Probiotic inulin, a plant-stored polysaccharide consisting of β-(2-1)-linked fructose residues, has been proven to alleviate diverse metabolic diseases via modulating gut microbiota. Inulin can alleviate endotoximia and inflammation in obesity by elevation of the relative abundance of intestinal Asteraceae, Bicenple cephalosporins and Bifidobacteria [13-15]. Metformin, a biguanide agent, reduces body weight (BW) and endotoxemia of patients by regulating the gut microbiota [16, 17]. However, the treatments of inulin in the inflammation and gut microbiota of PCOS remain poorly understood.

In the present study, we evaluated the effects and associated mechanisms of inulin and metformin on PCOS in mice. Our results may contribute to the understanding of complicated relationship among PCOS, gut microbiota, inflammation, and potential novel strategies for clinical applications.

Methods

Animal and diet

The experiments were approved by the Ethics Committee of General Hospital of Ningxia Medical University (No.2016-017). 21-day-old female C57BL/6J mice were purchased from Vital River Laboratory Animals Technology Co., Ltd., Beijing, China. All mice were housed in specialized cages in a temperature-controlled (25°C) and 12 h light/dark cycle room. All animal feed was obtained from TROPHIC Animal Feed High-tech Co., Ltd., Nantong, China. Inulin was purchased from Fengning Pingan High-tech Industrial Co., Ltd., Chengde, China. Trans-Dehydroandrosterone (DHEA) was obtained from Macklin Biochemical Co., Ltd., Shanghai, China.

Experimental design

All mice were randomly divided into 4 groups (10 mice/group). (a) Control group (CON), mice were fed the control diet and treated with same amount of normal saline (NS) as negative control; (b) Model group (MOD), mice were fed 6 mg DHEA per 100 g BW and 60% high-fat diet for 20 days with subsequently switched to the control diet, treated with the same amount of NS as the model control of the intervention group; (c) Inulin group (INU), mice were fed the same diet as MOD group and treated with inulin (0.05 g per 100 g BW) for 21 days while switched to a normal diet; (d) Metformin group (MET), mice were fed the same diet as MOD group but treated with metformin (1.9 g per 1,000 g BW) for 21 days while switched to a normal diet. 60% High-fat diet energy was consists of 19.4% protein, 20.6% carbohydrate and 60% fat. The control diet were the pair-fed controls for 60% high-fat diet. After 20 days of modeling and subsequent 21 days of treatments, all mice were euthanized and associated specimens were collected for further analysis.

Vaginal smear

Vaginal smears of all mice were collected every day at AM 9:00 to microscopically evaluate the estrous cycle by Wright-Giemsa staining. The observation period was lasted for 41 consecutive days (from the first day of modeling to the last day of treatment).

Determination of hormone and inflammatory factors

A series of plasma hormones including testosterone (T), follicle-stimulating hormone (FSH), luteinizing hormone (LH), prolactin (PRL), estradiol (E2), and progesterone (PROG), as well as the levels of TNF-α, IL-6, IL-10, and IL-17A in ovaries, were determined using enzyme linked immunosorbent assay (ELISA). All of indicators from plasma hormone and ovarian inflammatory cytokines were measured according to the manufacturer’s instructions (Shanghai Baoman Biotechnology Co., Ltd, Shanghai, China). The interassay variability of each ELISA kit was <15%, and the intra-assay variability of each ELISA kit was <10%.

Insulin tolerance tests

The insulin tolerance tests (ITT) were performed immediately after 21 days of inulin and metformin intervention. After the mice were fasted for 6 hours, blood glucose of the mice was measured. After intraperitoneal injection of insulin (1 IU/kg body weight) immediately, blood glucose was measured at 30, 60 and 90 minutes, respectively. GraphPad Prism software was used to calculated the total area under curve of glucose response (AUC).

Plasma LPS analysis

Plasma LPS levels of each group were determined using limulus amebocyte lysate kit (Xiamen Bioendo Technologky Co., Ltd, Xiamen, China). Briefly, 50 μL of diluted plasma (1:4 dilution with endotoxin-free water) was dispensed to each well in a 96-well plate. Then, 50 μL/well of the limulus amebocyte lysate reagent was added respectively. The microplate was incubated at 37°C for 30 min. After incubation, 100 μL of chromogenic substrate was warmed to 37°C was added for each well. Additional incubation was extended for 6 min at 37°C. The reaction was stopped by adding 100 μL of a 25% solution of glacial acetic acid. Optical density at 545 nm was used to measure microplate with the reader (Thermo Scientific, USA).

Hematoxylin-eosin (HE) staining

After sacrifice of mice, the ovaries were isolated and immediately fixed with formalin. The tissue was then dehydrated and embedded in paraffin. Paraffin-embedded sections were performed for HE staining. To evaluate ovary damage, every 10th section was mounted on a glass slide and analyzed using the Olympus light microscope (Melville, NY) by two persons blinded to the origin of the sections. Numbers of cystic follicle and corpora luta were counted. Cystic follicles were the follicles without oocytes, which showed as a large fluid-filled cyst with a thin granulosa cell layer and thick theca cell layer as previously described [18].

Cytometric bead array (CBA) analysis

The levels of plasma IL-6, IL-10, IL-17A and TNF-α were measured by BD^TM CBA mouse inflammatory cytokine kit (BD Bioscience, USA. No.560485) according to the manufacturer’s instruction. The operations were performed thrice. Briefly, a series of cytokine standard dilutions were prepared. Beads covered with 4 specific capture antibodies were mixed, respectively. Then 50 μL of mixed captured beads, 50 μL of the plasma sample or standard dilutions, with 50 μL of phycoerythrin (PE) detection reagent was added consecutively to each assay tube. After the mixture, the tubes were incubated for 2 h at room temperature in the dark. Next, each tube was added with 1 mL of wash solution and then centrifugated at 200 × g for 5 min. Finally, the supernatant of each tube was carefully removed, 300 μL of wash solution was added. After the measurement of resuspended microspheres by FACSCelesta^TM BD flow cytometer, cytokFCAP Array software (BD Bioscience, USA) was used for analysis of cytokine levels.

Sequencing and analysis of gut microbiota

After six weeks of treatment, 5 mice from each group were randomly selected and placed in sterilized cages. Fresh feces of each mouse was respectively collected and immediately stored at –80°C for subsequent DNA extraction.

Extraction of bacterial DNA by Cetyltrimethylammonium Ammonium Bromide (CTAB): the appropriate amount of lysozyme and sample were added to 1,000 μL CTAB lysate. The mixture was placed in a 65°C water bath and mixed by inversion several times in order to facilitate complete lysis of the sample. Next, phenol (PH 8.0), chloroform, and isoamyl alcohol were added to the supernatant after centrifugation so that the ratio of the three was 25:24:1. They were mixed by inversion and centrifuged at 12,000 × g for 10 min. In the same way, chloroform and isoamyl alcohol (24:1) were added to the obtained supernatant followed by centrifugation. The collected supernatant was added with isopropano. The mixture was precipitated at –20°C after shaking up and down. Then the mixture was centrifuged again according to the previous centrifugation conditions. The obtained precipitate was washed twice with 1 mL 75% ethanol. Then the precipitate was blown dry on a clean bench or air-dried at room temperature. It was particularly important that the DNA sample cannot be too dry. DNA samples were dissolved in ddH₂O. If the sample was difficult to dissolve, it needed to be incubated at 55–60°C for 10 min. Finally, 1 μL of RNase A was added to the dissolved DNA sample, which was allowed to placed at 37°C for 15 min to obtain bacterial DNA. The extracted DNA was stored at –20°C until application.

The DNA Sequences involving V3 and V4 of 16S rDNA hypervariable regions were amplified by Phusion^® High-Fidelity PCR Master Mix with GC Buffer (New England Biolab, America) using the following primers (5' to 3'): 341F-CCTAYGGGRBGCASCAG, 806R-GGACTACNNGGGTATCTAAT. The PCR product was analyzed and separated on a 2% agarose gel, which was purified using GeneJE Gel Recovery Kit (Thermo Scientific, America). The library was constructed using the TruSeq^® DNA PCR-Free Sample Preparation Kit in order to Qubit quantitation and library detection. After passing the test, the library was sequenced using iIllumina HiSeq 2500 platform by Beijing Novogene technology co., LTD., China.

Statistical analysis

All experimental data were analyzed using Prism (GraphPad Software Inc., CA, USA). The results were represented as mean ± SEM. According to the data conforming to a Gaussian distribution and equality of the variance, the independent t-tests were adopted for analysis of significant difference between two groups. In contrast, Mann-Whitney U were used to identify the significant difference in gut microbiota between two groups. Spearman’s correlation analysis was used to identify between gut microbiota and inflammatory indicators. Result were considered significant at p < 0.05.

Results

Routine parameters of mice in each group

There was no significant difference in body weights among diverse groups before treatment. However, after 21 days of intervention, BW in INU or MET group were lower than that in MOD group (INU vs. MOD, p = 0.0181; MET vs. MOD, p = 0.0001), demonstrating that inulin can reduce BWs in PCOS, but not as effective as metformin. T levels in MOD group were significantly higher than that in CON group (p = 0.0006). T levels in INU and MET groups were significantly reduced, compared to MOD group (INU vs. MOD, p = 0.0319; MET vs. MOD, p = 0.0019). Compared to CON group, E2 levels in MOD group were significantly lower (p = 0.0143). INU or MET treatment promoted the increase of E2 levels compared with that in MOD group (INU vs. MOD, p = 0.0221; MET vs. MOD, p = 0.0097). PROG levels in MOD group were significantly decreased, compared with CON group (p = 0.0137). In addition, there were no significant difference among treated and control groups in FSH, LH, PRL, and PROG levels (Table 1).

Table 1 Routine parameters of mice with diverse treatments in PCOS

Measurements	CON	MOD	INU	MET	Unpaired t tests
					a	b	c
Body Weight, g	20.48 ± 0.34	22.72 ± 0.34	21.66 ± 0.22	20.77 ± 0.21	0.0002	0.0181	0.0001
T, ng/mL	520.7 ± 45.26	920.35 ± 56.38	741.56 ± 39.61	578.15 ± 50.46	0.0006	0.0319	0.0019
FSH, mIU/mL	53.81 ± 3.45	45.71 ± 2.81	51.94 ± 3.52	52.42 ± 3.26	0.1059	0.2040	0.1577
LH, mIU/mL	6.71 ± 0.70	8.47 ± 0.38	7.13 ± 0.51	6.80 ± 0.65	0.0599	0.0697	0.0576
PRL, ng/mL	37.34 ± 2.31	32.51 ± 0.46	33.89 ± 0.75	37.13 ± 2.22	0.0742	0.1580	0.0763
E2, μmol/L	51.71 ± 1.050	40.64 ± 3.39	50.42 ± 0.66	52.22 ± 0.51	0.0143	0.0221	0.0097
PROG, ng/mL	12.80 ± 0.70	10.04 ± 0.52	11.44 ± 0.68	11.63 ± 0.75	0.0137	0.1420	0.1207

a: CON vs. MOD, b: MOD vs. INU, c: MOD vs. MET

Vaginal cytology

Alterations in vaginal smears of different estrous cycles in mice were shown in Fig. 1. Vaginal cytology for the first 20 days showed that the estrous cycle of PCOS mice was disordered and stopped in the diestrus phase. Intriguingly, intervention of inulin or metformin alleviated the estrous cycle disorder of PCOS in mice.

Fig. 1

Changes in estrous cycle of mice of each group. a: Vaginal smears of proestrus stage. b: Vaginal smears of estrus stage. c: Vaginal smears of metestrus stage. d: Vaginal smears of diestrus stage. e: Representative estrous cycle of each group, 1: diestrus stage, 2: proestrus stage, 3: estrus stage, 4: metestrus stage. Original magnification, ×100.

Inulin and metformin attenuated ovarian histopathological injury and reduced plasma LPS levels

According to HE staining of ovarian tissue in each group, granulosa layers of ovarian tissue in MOD group showed reduced cystic degenerating follicles. Whereas inulin or metformin increased the formation of granulosa layers and corpora lutea, demonstrating that inulin or metformin contributed to improving the pathological changes of ovarian tissue (Fig. 2).

Fig. 2

Effects of iunlin or metformin on ovarian tissues in PCOS with Hematoxylin-eosin (H&E) staining. a: CON, b: MOD, c: INU, d: MET. The boxed areas in a, b, c, d (40×) was shown changes in cystic follicles, with higher magnification (200×) in e, f, g, h. Changes in the corpus luteum (i and j). TCL: theca cell layer, GCL: granular cell layer, L: luteum. Original magnification (40×).

Plasma LPS was significantly increased in MOD group, compared with that in CON group (p = 0.0004). Meanwhile, the abnormal elevated LPS in PCOS was decreased in INU (p = 0.0023) or MET (p = 0.0006) group, suggesting that inulin and metformin possessed ability to attenuate LPS-induced endotoximia in PCOS (Fig. 3).

Fig. 3

Effects of inulin or metformin on plasma lipopolysaccharide (LPS) levels. Data are displayed as mean ± SEM, * p < 0.05, ** p < 0.01, *** p < 0.001.

Inulin and metformin improved insulin resistance in PCOS

Insulin sensitivity was significantly lower in MOD group than that in CON group (p = 0.0011), as determined by area under the curve AUC glucose during insulin tolerance tests. Meanwhile, insulin sensitivity was improved in INU group and in MET group compared to MOD group (Fig. 4, INU vs. MOD, p = 0.0103; MET vs. MOD, p = 0.0033).

Fig. 4

Insulin tolerance test. a. Serum glucose levels, b. AUC of glucose. * p < 0.05, ** p < 0.01.

Inulin and metformin decreased the levels of inflammatory cytokines of plasma and ovary in PCOS

We found a significant increase of plasma and ovary TNF-α (plasma: p = 0.0002, ovary: p = 0.0087), IL-6 (plasma: p = 0.0088, ovary: p = 0.0068), IL-17A (plasma: p < 0.0001, ovary: p = 0.0150), as well as a decrease of plasma and ovary IL-10 (plasma: p = 0.0365, ovary: p = 0.0055) in MOD group than that in CON group. In contrast, the IL-6 levels in INU (plasma: p = 0.0362, ovary: p = 0.0235) or MET group (plasma: p = 0.0029, ovary: p = 0.0171) were significantly reduced, compared to MOD group. Similarly, TNF-α concentrations were also lower in treated groups (plasma: INU vs. MOD, p = 0.0331; MET vs. MOD, p = 0.0315; ovary: INU vs. MOD, p = 0.0331; MET vs. MOD, p = 0.0315). Moreover, after the treatment, ovary anti-inflammatory IL-10 levels were increased (INU vs. MOD, p = 0.0318; MET vs. MOD, p = 0.0187). Finally, IL-17A levels in plasma and ovary were significantly lower in INU (plasma: p = 0.0024, ovary: p = 0.0437) or MET group (plasma: p < 0.0001, ovary: p = 0.0152) than that in MOD group (Fig. 5).

Fig. 5

Detection of plasma or ovary inflammatory cytokine levels in diverse groups of mice. a–d: Plasma in mice of each group was collected respectively for detection of TNF-α (a), IL-6 (b), IL-10 (c), and IL-17A (d). e–h: Ovary in mice of each group was collected respectively for detection of TNF-α (e), IL-6 (f), IL-10 (g), and IL-17A (h). Data were expressed as mean ± SEM. * p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001, NS, not significant.

Inulin and metformin modulated gut microbiota in PCOS

Accumulating evidences have demonstrated that the gut microbiota plays a crucial role in the occurrence and development of PCOS [4-7]. In order to investigate whether amelioration of inulin in PCOS was closely related to the alterations in the composition of gut microbiota, we conducted a metagenomic analysis of mice feces. Rarefaction curves were adopted to evaluate the rationality of the sequencing data (Supplementary Fig. S1), which tended to be flat when the sequences number increases to 8,429, showing that the amount of sequencing data were reasonable.

The whole composition bacterial community was analyzed using Unweighted-Unifrac of Pcoa (Fig. 6) and Weighted-Unifrac of NMDS (Supplementary Fig. S2). PCoA results showed a significant difference of species in fecal samples among the CON, MET and INU groups indicating that iunlin altered gut microbiota of PCOS. In addition, NMDS analysis presented similar results.

Fig. 6

PcoA analysis showed the difference in terms of species in fecal samples. Beta diversity was on Unweighted-Unifrac. a: CON vs. MOD; b: MOD vs. INU; c: MOD vs. MET.

At phylum level (Fig. 7a), we analyzed top six of phyla and found that Firmicutes and Bacteroidetes were the most common phylum in diverse groups. The proportion of Bacteroidetes was increased in CON (p = 0.0317), INU (p = 0.5317) or MET (p = 0.0952) group, compared to MOD group. There’s no alteration of Firmicutes in CON (p = 0.1508), INU (p = 0.9444) or MET (p = 0.8016) group, compared to MOD group. In addition, we found that Proteobacteria was significantly reduced in CON (p = 0.0079), INU group (p = 0.0317) or MET (p = 0.0556) group, compared with MOD group (Fig. 7c).

Fig. 7

Relative abundance of microbial species at phylum and genus levels in mice feces. a: The phylum analysis; b: The genus analysis; c: The significant analysis in phylum levels. d–f: The significant analysis in genus levels. * p < 0.05, ** p < 0.01, NS, no significance.

At the genus level (Fig. 7b), we analyzed top forty bacteria and found that Bifidobacterium, the most prevalent genus in each group, was significantly increased in CON and INU group (Fig. 7d, CON vs. MOD, p = 0.0079; INU vs. MOD, p = 0.0317). Moreover, the proportions of Helicobacter (Fig. 7e) and Parasutterella (Fig. 7f) were notably lower in CON (Helicobacter, p = 0.0079; Parasutterella, p = 0.0079) and INU group (Helicobacter, p = 0.0476; Parasutterella, p = 0.0159) than that in control MOD group. Moreover, Helicobacter also showed decreased in MET group (Helicobacter, p = 0.0238). But Parasutterella in MET group possessed no significant difference (Parasutterella, p = 0.0952). In general, these results indicated that dietary inulin or metformin significantly changed the initial proportion of genus components, mainly including Bifidobacterium, Helicobacter and Parasutterella.

Correlations between inflammatory cytokines and bacterial relative abundance

At the genus level, we analyzed the correlations between significant inflammatory cytokines and gut microbiota. In plasma, Bifidobacterium showed a negative correlation with TNF-α (p = 0.0006, Fig. 8a), IL-6 (p = 0.0040, Fig. 8b) or IL-17A (p = 0.0120, Fig. 8c), respectively. Helicobacter abundance was positively correlated with plasma inflammatory cytokines (TNF-α, p = 0.0184; IL-6, p = 0.0080; IL-17A, p = 0.0156, Fig. 8d–f). Bifidobacterium showed a negative correlation with TNF-α (p = 0.0004, Fig. 9a), IL-6 (p = 0.0150, Fig. 9c) and IL-17A (p = 0.0053, Fig. 9h). Parasutterlla abundance was positively correlated with ovary pro-inflammatory cytokines including TNF-α, IL-6, IL-17A (Fig. 9b: TNF-α, p = 0.0278; Fig. 9d: IL-6, p = 0.0020; Fig. 9i: IL-17A, p = 0.0132), but negatively correlated with ovary anti-inflammatory cytokine IL-10 (p = 0.0041, Fig. 9f). Meanwhile, Helicobacter abundance was positively correlated with IL-6 (p = 0.0022, Fig. 9e) and negatively correlated with IL-10 (p = 0.0092, Fig. 9g).

Fig. 8

Correlation analysis between relative abundance of gut microbiota and plasma inflammatory cytokines. a: Bifidobacterium and TNF-α; b: Bifidobacterium and IL-6; c: Bifidobacterium and IL-17A; d: Helicobacter and TNF-α; e: Helicobacter and IL-6; f: Helicobacter and IL-17A.

Fig. 9

Correlation analysis between relative abundance of gut microbiota and ovary inflammatory cytokines. a: Bifidobacterium and TNF-α; b: Parasutterella and TNF-α; c: Bifidobacterium and IL-6; d: Parasutterella and IL-6; e: Helicobacter and IL-6; f: Parasutterella and IL-10; g: Helicobacter and IL-10; h: Bifidobacterium and IL-17A; i: Parasutterella and IL-17A.

Discussion

In the present study, we investigated the efficacy of dietary inulin treatment in PCOS. By in vivo 21-day treatment of PCOS in mice, our study indicated that inulin and metformin showed more effective in the decrease of ovarian damage, suggesting that the intervention displayed preventive and therapeutic potential. Our further study revealed that amelioration of PCOS phenotype including hormonal disturbance may closely associated with the reduction of inflammation and the alteration of gut microbiota.

In this paper, dietary inulin and metformin were demonstrated to possess capacity to alleviate PCOS throughout a series of routine indicators including pathological examination, reduced BWs, as well as rectification of abnormal hormones (T and E2), which were consistent with previous studies [19-24].

Abnormal level of LPS derived from gut microbiota dysbiosis dramatically affects systemic metabolism through changing the intestinal epithelial barrier, leading to bacterial endotoxin into the blood, thereby upregulating pro-inflammatory signals. LPS induces inflammatory immune response to ultimately lead to chronic ovarian inflammation [3, 7]. In our study, plasma LPS was significantly decreased in INU and MET groups, suggesting that inulin and metformin may reduce intestinal permeability and decrease translocation of LPS from the intestine to the liver and blood circulation, eventually contributing to the reduction of ovarian inflammation.

As critical pro-inflammatory indicators, abnormal high levels of plasma TNF-α and IL-6 have been demonstrated to promote the proliferation, differentiation and maturation of follicles in PCOS patients [25-27]. Alteration of TNF-α in FF levels is associated with poor-quality oocytes undergoing IVF which decreased rates of fertilization, embryonic development, and pregnancy outcome [25]. IL-6 can reduce intra-follicular estradiol concentration, fertility and fertilizing capacity [26]. Studies have shown that pro-inflammatory cytokine (such as TNF-α and IL-6) in PCOS are significantly increased [27]. TNF-α, IL-6 and IL-10 play important role in the development of inflammation and oxidative stress [27]. Serum IL-17A, TNF-α and IL-6 are closely related to the number of meiotic I (MI) oocytes in PCOS patients [28]. In this study, we found that the levels of TNF-α, IL-6, and IL-17A of plasma and ovary in INU and MET groups were significantly reduced, suggesting that inulin and metformin may alleviate ovarian inflammation via anti-inflammatory cytokines. IL-10, a crucial anti-inflammatory cytokine, showed significant difference in ovary between treated groups and MOD group, suggesting that inulin and metformin may reduced PCOS inflammation partially via regulating IL-10 expression [29, 30].

Dysbiosis of gut microbiota is closely related with the pathogenesis of PCOS [4, 5, 31]. In our study, the proportion of Bacteriodetes and Firmicutes still remained predominant in all groups at phylum level, which was consistent with previous studies [31, 32]. Proteobacteria, a pathogenic bacterium that produces LPS, was significantly decreased in INU group compared to MOD group, indicating that reduction of LPS treated by inulin may due to inhibition of Proteobacteria proliferation [33]. However, the exact role of Proteobacteria in PCOS needs to be further investigated.

Accumulating studies have demonstrated that inulin may increase proportions of Bifidobacterium and Lactobacillus [32, 34]. Probiotic Bifidobacterium possessed capable of suppression of pro-inflammatory cytokine synthesis [35-37]. We also found that inulin elevated this beneficial Bifidobacterium, revealing that inulin may ameliorate PCOS by improving the proportion of Bifidobacterium. In addition, recent studies have showed the possible existence of a correlation between infection with Helicobacter and metabolic syndrome [37]. Furthermore, we found that the decreased number of Helicobacter after treatment with inulin or metformin showed a significantly positive correlation with pro-inflammatory factors (TNF-α, IL-6, and IL-17A). But the physiological role of Helicobacter in metabolic diseases, especially PCOS, is poorly understood. Pathogenic Parasutterella showed a significant decrease after intervention with inulin in our study. Taken together, in the present study, several differential candidates of gut bacteria were identified after treatment with inulin or metforin, of which, the alterations may contribute to improving gut dysbiosis and restoring the microbiota balance. But complicated interactions among above different bacteria and their metabolites with mucosal immune cells (NK cells, regulatory T lymphocytes, γδT cells, marrow derived suppressive cells and dendritic cells) and metabolism still remain largely unknown and need to be further investigated.

Conclusions

The present study highlighted that inulin and metformin may ameliorate PCOS via anti-inflammation and modulating gut microbiota in mice, which may potentially serve as an intervention for the prevention and treatment of PCOS.

Abbreviations

CON, control; INU, inulin; MOD, model; MET, metformin; T, Testosterone; FSH, Follicle-Stimulating Hormone; LH, Luteinizing Hormone; PRL, Prolactin; E2, Estradiol; PORG, Progesterone; ITT, insulin tolerance test; HE, Hematoxylin-eosin; IL, Interleukin; LPS, Lipopolysaccharide; CBA, Cytometric Bead Array; BW, Body weight; TNF, Tumor necrosis factor.

Acknowledgements

Not applicable.

Funding

This work was supported by the following funds:

1. The Research and Development Plan of the 13th Five-year Plan of Ningxia Autonomous Region (the major S&T projects) (Grant No. 2016BZ02), China.

2. The First Class Discipline Construction Project in Colleges and Universities of Ningxia (Grant No. NXYLXK2017A05), China.

3. Ningxia High School First-class Disciplines (West China first-class Disciplines Basic Medical Sciences at Ningxia Medical University) (Grant No. NXYLXK2017B07), China.

4. The National Key Research and Development Program of China (No. 2016YFD0400605), China.

Availability of Data and Materials

The datasets used and analyzed in the current study are available from the corresponding author on reasonable request.

Authors’ Contributions

WH, XJ and DXY designed and wrote the paper. LXR, LP, LK, SLP, YXL, ZLL, WZ, DYP, ZL, LH, and ZXX performed research. All authors have read and approved the final manuscript.

Competing Interests

The authors declare that they have no competing interests. The funding body plays no role in the design of the study and collection, analysis, and interpretation of data and in writing the manuscript.

Consent for Publication

Not applicable.

References

• 1   March  WA,  Moore  VM,  Willson  KJ,  Phillips  DI,  Norman  RJ, et al. (2010) The prevalence of polycystic ovary syndrome in a community sample assessed under contrasting diagnostic criteria. Hum Reprod 25: 544–551.
• 2  Homburg R (2008) Polycystic ovary syndrome. Best Pract Res Clin Obstet Gynaecol 22: 261–274.
• 3   Cani  PD,  Possemier  S,  Van de Wiele  T,  Guiot  Y,  Everard  A, et al. (2009) Changes in gut microbiota control inflammation in obese mice through a mechanism involving GLP-2-driven improvement of gut permeability. Gut 58: 1091–1103.
• 4   Liu  R,  Zhang  C,  Shi  Y,  Zhang  F,  Li  L, et al. (2017) Dysbiosis of gut microbiota associated with clinical parameters in polycystic ovary syndrome. Front Microbiol 8: 324.
• 5   Guo  Y,  Qi  Y,  Yang  X,  Zhao  L,  Wen  S, et al. (2016) Association between polycystic ovary syndrome and gut microbiota. PLoS One 11: e0153196.
• 6   Zhang  D,  Zhang  L,  Yue  F,  Zheng  Y,  Russell  R (2015) Serum zonulin is elevated in women with polycystic ovary syndrome and correlates with insulin resistance and severity of anovulation. Eur J Endocrinol 172: 29–36.
• 7   Tremellen  K,  Pearce  K (2012) Dysbiosis of Gut Microbiota (DOGMA)—a novel theory for the development of polycystic ovarian syndrome. Med Hypotheses 79: 104–112.
• 8   Zhang  F,  Ma  T,  Cui  P,  Tamadon  A,  He  S, et al. (2019) Diversity of the gut microbiota in dihydrotestosterone-induced PCOS rats and the pharmacologic effects of Diane-35, probiotics, and berberine. Front Microbiol 10: 175.
• 9   Escobar-Morreale  HF,  Luque-Ramirez  M,  Gonzalez  F (2011) Circulating inflammatory markers in polycystic ovary syndrome: a systematic review and metaanalysis. Fertil Steri 95: 1048–1058 e1–2.
• 10   Shen  J,  Obin  MS,  Zhao  L (2013) The gut microbiota, obesity and insulin resistance. Mol Aspects Med 34: 39–58.
• 11   Lindheim  L,  Bashir  M,  Munzker  J,  Trummer  C,  Zachhuber  V, et al. (2017) Alterations in gut microbiome composition and barrier function are associated with reproductive and metabolic defects in women with Polycystic Ovary Syndrome (PCOS): a pilot study. PloS One 12: e0168390.
• 12   Zeng  B,  Lai  Z,  Sun  L,  Zhang  Z,  Yang  J, et al. (2019) Structural and functional profiles of the gut microbial community in polycystic ovary syndrome with insulin resistance (IR-PCOS): a pilot study. Res Microbiol 170: 43–52.
• 13   Vandeputte  D,  Falony  G,  Vieira-Silva  S,  Wang  J,  Sailer  M, et al. (2017) Prebiotic inulin-type fructans induce specific changes in the human gut microbiota. Gut 66: 1968–1974.
• 14   Nicolucci  AC,  Hume  MP,  Martinez  I,  Mayengbam  S,  Walter  J, et al. (2017) Prebiotics reduce body fat and alter intestinal microbiota in children who are overweight or with obesity. Gastroenterology 153: 711–722.
• 15   Everard  A,  Lazarevic  V,  Gaia  N,  Johansson  M,  Stahlman  M, et al. (2014) Microbiome of prebiotic-treated mice reveals novel targets involved in host response during obesity. ISME J 8: 2116–2130.
• 16   Ejtahed  HS,  Tito  RY,  Siadat  SD,  Hasani-Ranjbar  Z,  Hoseini-Tavassol  Z, et al. (2018) Metformin induces weight loss associated with gut microbiota alteration in non-diabetic obese women: a randomized double-blind clinical trial. Eur J Endocrinol 180: 165–176.
• 17   Wang  JH,  Bose  S,  Shin  NR,  Chin  YW,  Kim  H, et al. (2018) Pharmaceutical impact of Houttuynia cordata and metformin combination on high-fat-diet-induced metabolic disorders: link to intestinal microbiota and metabolic endotoxemia. Front Endocrinol (Lausanne) 9: 620.
• 18   Lai  H,  Jia  X,  Yu  Q,  Zhang  C,  Qian  J, et al. (2014) High-fat diet induces significant metabolic disorders in a mouse model of polycystic ovary syndrome. Biol Reprod 91: 127.
• 19   John  GK,  Wang  L,  Nanavati  J,  Twose  C,  Singh  R, et al. (2018) Dietary alteration of the gut microbiome and its impact on weight and fat mass: a systematic review and meta-analysis. Genes (Basel) 9: 167.
• 20   Out  M,  Miedema  I,  Jager-Wittenaar  H,  van der Schans  C,  Krijnen  W, et al. (2018) Metformin-associated prevention of weight gain in insulin-treated type 2 diabetic patients cannot be explained by decreased energy intake: a post hoc analysis of a randomized placebo-controlled 4.3-year trial. Diabetes Obes Metab 20: 219–223.
• 21   Gholizadeh Shamasbi  S,  Dehgan  P,  Mohammad-Alizadeh Charandabi  S,  Aliasgarzadeh  A,  Mirghafourvand  M (2019) The effect of resistant dextrin as a prebiotic on metabolic parameters and androgen level in women with polycystic ovarian syndrome: a randomized, triple-blind, controlled, clinical trial. Eur J Nutr 58: 629–640.
• 22   Jeong  SY,  Kang  S,  Hua  CS,  Ting  Z,  Park  S (2017) Synbiotic effects of beta-glucans from cauliflower mushroom and Lactobacillus fermentum on metabolic changes and gut microbiome in estrogen-deficient rats. Genes Nutr 12: 31.
• 23   Naderpoor  N,  Shorakae  S,  de Courten  B,  Misso  ML,  Moran  LJ, et al. (2015) Metformin and lifestyle modification in polycystic ovary syndrome: systematic review and meta-analysis. Hum Reprod Update 21: 560–574.
• 24   Karateke  A,  Dokuyucu  R,  Dogan  H,  Ozgur  T,  Tas  ZA, et al. (2019) Investigation of therapeutic effects of erdosteine on polycystic ovary syndrome in a rat model. Med Princ Pract 27: 515–522.
• 25   Qiao  J,  Feng  HL (2011) Extra- and intra-ovarian factors in polycystic ovary syndrome: impact on oocyte maturation and embryo developmental competence. Hum Reprod Update 17: 17–33.
• 26   Nunez-Calonge  R,  Cortes  S,  Gutierrez Gonzalez  LM,  Kireev  R,  Vara  E, et al. (2016) Oxidative stress in follicular fluid of young women with low response compared with fertile oocyte donors. Reprod Biomed Online 32: 446–456.
• 27   Artimani  T,  Karimi  J,  Mehdizadeh  M,  Yavangi  M,  Khanlarzadeh  E, et al. (2018) Evaluation of pro-oxidant-antioxidant balance (PAB) and its association with inflammatory cytokines in polycystic ovary syndrome (PCOS). Gynecol Endocrinol 34: 148–152.
• 28   Sabbaghi  M,  Aram  R,  Roustaei  H,  Fadavi Islam  M,  Daneshvar  M, et al. (2014) IL-17A concentration of seminal plasma and follicular fluid in infertile men and women with various clinical diagnoses. Immunol Invest 43: 617–626.
• 29   Li  K,  Zhang  L,  Xue  J,  Yang  X,  Dong  X, et al. (2019) Dietary inulin alleviates diverse stages of type 2 diabetes mellitus via anti-inflammation and modulating gut microbiota in db/db mice. Food Funct 10: 1915–1927.
• 30   Kelly  B,  Tannahill  GM,  Murphy  MP,  O’Neill  LA (2015) Metformin inhibits the production of reactive oxygen species from NADH: ubiquinone oxidoreductase to limit induction of interleukin-1β (IL-1β) and boosts interleukin-10 (IL-10) in lipopolysaccharide (LPS)-activated macrophages. J Bio Chem 290: 20348–20359.
• 31   Kelley  ST,  Skarra  DV,  Rivera  AJ,  Thackray  VG (2016) The gut microbiome is altered in a letrozole-induced mouse model of polycystic ovary syndrome. PLoS One 11: e0146509.
• 32   Zhang  Q,  Yu  H,  Xiao  X,  Hu  L,  Xin  F, et al. (2018) Inulin-type fructan improves diabetic phenotype and gut microbiota profiles in rats. PeerJ 6: e4446.
• 33   Macfarlane  S,  Cleary  S,  Bahrami  B,  Reynolds  N,  Macfarlane  GT (2013) Synbiotic consumption changes the metabolism and composition of the gut microbiota in older people and modifies inflammatory processes: a randomised, double-blind, placebo-controlled crossover study. Aliment Pharmacol Ther 38: 804–816.
• 34   Drabińska  N,  Jarocka-Cyrta  E,  Markiewicz  LH,  Krupa-Kozak  U (2018) The effect of oligofructose-enriched inulin on faecal bacterial counts and microbiota-associated characteristics in celiac disease children following a gluten-free diet: results of a randomized, placebo-controlled trial. Nutrients 10: 201.
• 35   Angurana  SK,  Bansal  A,  Singhi  S,  Aggarwal  R,  Jayashree  M, et al. (2018) Evaluation of effect of probiotics on cytokine levels in critically ill children with severe sepsis: a double-blind, placebo-controlled trial. Crit Care Med 46: 1656–1664.
• 36  Xie Y, Guo Q, Li S, Liu M, Zhang Q, et al. (2017) Anti-inflammatory properties of Bifidobacterium longum expressing human manganese superoxide dismutase using the TNBS-induced rats model of colitis. J Microbiol Biotechnol 27.
• 37   Klemenak  M,  Dolinsek  J,  Langerholc  T,  Di Gioia  D,  Micetic-Turk  D (2015) Administration of bifidobacterium breve decreases the production of TNF-alpha in children with celiac disease. Dig Dis Sci 60: 3386–3392.