Helpful Effects of a Specific Probiotic on Polycystic Ovary Syndrome Symptoms in Rats Treated with Letrozole

To investigate the effects of DSM 27449 on PCOS symptoms, we utilized a letrozole-induced PCOS-like rat model. Letrozole, a non-steroidal aromatase inhibitor, was used to induce PCOS-like characteristics, because it reliably mimics the reproductive and metabolic features of PCOS in humans, including increased androgen levels, polycystic ovaries, irregular estrous cycles, and metabolic irregularities [23,25]. In this study, the rats were divided into four groups: control, letrozole, Diane-35, and DSM 27449. Except for the control group, all rats were induced with PCOS-like symptoms using letrozole. Diane-35, a combination of cyproterone acetate and ethinylestradiol, was applied as a treatment control, because it is commonly used to manage PCOS symptoms in clinical settings. It helps reduce androgen levels and improve menstrual regularity [22]. Diane-35 and DSM 27449 were administered to respective groups to evaluate their effects on PCOS-like symptoms. A schematic representation of the experimental design is shown in Figure 1A.

In all experimental groups, body weights were similar at the beginning of the experiment. Three weeks of letrozole treatment significantly increased the body weight of rats in the letrozole and DSM 27449 groups compared to that of those in the control group, while rats in the Diane-35 group showed a smaller increase in body weight (Figure 1B). Overall, the PCOS-like rats were significantly heavier than the control rats at the end of the study (Figure 1C, Control vs. Letrozole, p < 0.0001; Control vs. Diane-35, p = 0.0115; Control vs. DSM 27449, p < 0.0001). The administration of DSM 27449 did not significantly reduce the weight gain caused by PCOS induction. To comprehensively assess the metabolic and reproductive health of the rats, glucose homeostasis and insulin sensitivity were assessed, as PCOS is highly correlated with metabolic diseases. The PCOS-like rats exhibited higher fasting glucose levels than those in the control group (Figure S1A, Control vs. Letrozole, p = 0.0338; Control vs. Diane-35, p = 0.0035); however, the level of fasting glucose was comparable between the DSM 27449 and control groups (p = 0.4977). The levels of fasting insulin and the homeostasis model assessment of the insulin resistance index were significantly increased in the PCOS-like rats compared to the control group (Figure S1B,C, p < 0.0001), indicating that insulin resistance was observed in this PCOS animal model. Nevertheless, the administration of DSM 27449 and Diane-35 did not rescue insulin resistance. The oral glucose tolerance test (OGTT) results show no delay in glucose clearance between the groups during the test (Figure S1D,E). Additionally, the lipid profile, including the levels of triglycerides, total cholesterol, high-density lipoprotein (HDL) cholesterol, and low-density lipoprotein (LDL) cholesterol were examined, as lipid abnormalities are common in PCOS. The levels of these lipids were similar in the control, letrozole, and DSM 27449 groups (Figure S1F–I), indicating that letrozole treatment and DSM 27449 administration have no significant effects on the serum lipid profile in PCOS-like rats. Remarkably, HDL-cholesterol levels were the lowest in the Diane-35 group (Control vs. Diane-35, p = 0.0304).

The reproductive health and ovarian function of the rats were evaluated by monitoring the estrous cycle and examining the ovarian morphology. The control rats exhibited regular estrous cycles of 4–5 days, whereas a majority of the PCOS-like rats were completely acyclic after the letrozole treatment (Figure 2A). Our results show that the PCOS-like rats spent a significantly higher percentage of time in the metestrus/diestrus phase than those in the control group, suggesting an irregular estrous cycle and reproductive dysfunction (Figure 2B, Control vs. Letrozole, p < 0.0001; Control vs. Diane-35, p = 0.0068; Control vs. DSM 27449, p = 0.0017). However, with the supplementation of DSM 27449, the disrupted estrous cycles improved, with 25% of the PCOS-like rats in the DSM 27449 group showing restored estrous cycles by the end of the experiment. A histological examination of the ovarian tissues showed multiple growing follicles at different developmental stages and a high number of corpora lutea in the control rats, whereas the PCOS-like rats exhibited an increased number of cystic follicles and fewer corpora lutea (Figure 2C). DSM 27449 significantly decreased the number of cystic follicles (Figure 2D, Letrozole vs. DSM 27449, p = 0.0135). A decreased number of corpora lutea was observed in the letrozole group, whereas DSM 27449 slightly recovered the number of corpora lutea compared to that in the letrozole group (Figure 2E). Overall, our data indicate that DSM 27449 has the potential to improve ovarian function and pathological changes, similar to the effects observed with Diane-35.

An elevated level of testosterone is one of the typical characteristics of PCOS. In this study, serum testosterone levels were significantly higher in the letrozole group than those in the control group (Figure 3A, p < 0.0001). The administration of DSM 27449 and Diane-35 significantly reduced the testosterone levels compared to those in the letrozole group, returning them to levels similar to the control rats (Letrozole vs. DSM 27449, p = 0.0404; Letrozole vs. Diane-35, p = 0.0064). The androgen receptor (AR) is responsible for mediating the effects of androgens, and AR-mediated effects play a critical role in the development of PCOS [26]. Therefore, we further examined the expression of AR in ovarian tissues. An immunohistochemistry (IHC) analysis of AR expression demonstrated its presence in the ovarian tissue (Figure 3B). The average AR-positive areas in the control, letrozole, Diane-35, and DSM 27449 groups were 25.4%, 48.5%, 34.8%, and 37.6%, respectively. An increased immunoreactive signal was observed in the PCOS-like rats, but the AR-positive areas in the DSM 27449 and Diane-35 groups were significantly lower than those in the letrozole group (Figure 3C, Letrozole vs. DSM 27449, p = 0.0050; Letrozole vs. Diane-35, p = 0.0004).

MiRNA dysregulation is often observed in patients with PCOS and PCOS animal models. To further clarify the role of miRNAs in PCOS, we evaluated the expression of a set of miRNAs that have been reported to be differentially expressed in PCOS, including miR-30a-5p, miR-93-5p, miR-223-3p, miR-320-3p, miR-324-3p, and miR-146a-5p, in the plasma and ovarian tissues of the rats [12,27,28,29,30]. We found that the expression of the plasma miR-30a-5p (p = 0.0335), miR-93-5p (p = 0.0469), and miR-223-3p (p = 0.0474) was significantly increased in the letrozole group compared to that in the control group (Table 1). The administration of DSM 27449 significantly reduced the expression of the plasma miR-30a-5p compared to that in the letrozole group (Letrozole vs. DSM 27449, p = 0.0066), and a decreasing trend in the expression of the plasmas miR-93-5p and miR-223-3p was observed. Similarly, the expression of miR-30a-5p was significantly increased in the ovarian tissues of the letrozole group (Control vs. Letrozole, p = 0.0227), whereas it was significantly reduced by DSM 27449 (Letrozole vs. DSM 27449, p = 0.0227). The expression levels of miR-320-3p, miR-324-3p, and miR-146a-5p in the plasma and ovarian tissues were comparable among all the groups. Our data demonstrate that DSM 27449 has the potential to restore the expression of miRNAs, as demonstrated by a significant increase in miR-30a-5p expression following letrozole induction, which was subsequently reduced by DSM 27449 administration in both the plasma and ovarian tissues.

Next, we investigated the correlation between miRNA expressions and PCOS-related parameters, including the serum testosterone, number of cystic follicles, and AR-positive area in the ovarian tissue. We observed that miR-30a-5p, miR-93-5p, and miR-223-3p were significantly associated with PCOS-related parameters, suggesting that these miRNAs play critical roles in PCOS pathology (Figure 4A). Coincidentally, a dysregulation of these miRNAs and excess androgen production have been found to be involved in the stimulation and promotion of apoptosis and autophagy in ovarian cells [31,32,33]. Thus, we examined apoptosis and autophagy in the ovarian tissues by performing IHC staining for the pro-apoptotic protein Bax, anti-apoptotic protein Bcl-2, and autophagy-related protein Beclin-1. As shown in Figure 4B,C, the expression of Bax was significantly increased in the letrozole group compared to that in the control group (p = 0.0083), which shows a 2.02-fold increase. The administration of DSM 27449 significantly reduced the expression of Bax in the ovarian tissues compared to that in the letrozole group (Figure 4B,C, Letrozole vs. DSM 27449, p = 0.0443). Interestingly, Diane-35 treatment did not reduce Bax expressions, suggesting a different mechanism for ameliorating symptoms in PCOS-like rats. The expression levels of Bcl-2 and Beclin-1 were not significantly different between the groups (Figure 4D,E).

Full-length 16S rRNA sequencing with a high quality and sufficient sequencing depth (Table S1) was performed to explore the effects of letrozole and DSM 27449 on the diversity and composition of the gut microbiome in PCOS-like rats. Alpha diversity was assessed using the Chao-1, Shannon, and Simpson indices to determine the species richness and evenness in the samples. No significant differences in alpha diversity were found among the control, letrozole, Diane-35, and DSM 27449 groups (Figure 5A–C). A non-metric multidimensional scaling (NMDS) plot based on weighted UniFrac distance matrices was generated to evaluate the differences in the beta diversity of the bacterial community. The results show no significant separation of the gut microbiome composition between the groups (Figure 5D, R = 0.047, p = 0.165). A linear discriminant analysis effect size was used to identify differentially expressed taxa in each group (Figure S2). Further analysis of the gut microbiome composition revealed 14 different taxonomic groups at the phylum level (Figure 5E). Firmicutes and Bacteroidetes were the dominant phyla, accounting for 97% of the reads. Although treatment with letrozole, Diane-35, or DSM 27449 had no significant effect on the alpha and beta diversity of the gut microbiome, we compared the relative abundance of the bacterial taxa at the genus level and identified six genera with differential expressions between the control and letrozole groups. Letrozole treatment significantly increased the relative abundances of the Prevotellaceae NK3B31 group (p = 0.0207), Ruminococcaceae UCG-013 (p = 0.0104), and Lactococcus (p = 0.0148) while decreasing the relative abundances of Acetitomaculum (p = 0.0193), Fusobacterium (p = 0.0204), and Bifidobacterium (p = 0.0056) compared to the control group (Figure 5F–K). Notably, the expression levels of these six genera were similar between the control and DSM 27449 groups, suggesting that DSM 27449 administration partially counteracted the alterations in the relative abundance induced by letrozole.

Spearman’s correlation analysis was used to investigate the correlation between differential microbial genera and the PCOS-related parameters. A majority of the microbial genera were significantly associated with the serum testosterone, number of cystic follicles, and AR-positive area in the ovarian tissues (Figure 6A). The Prevotellaceae NK3B31 group, Ruminococcaceae UCG-013, and Lactococcus were positively correlated with the PCOS-related parameters, whereas Acetitomaculum, Fusobacterium, and Bifidobacterium were negatively correlated with the PCOS-related parameters. These results indicate that the modulation of the microbial genera by DSM 27449 is positively associated with an improvement in PCOS symptoms.

We further assessed the association between microbial genera and plasma- and ovary-tissue-specific miRNAs across all cohorts (Figure 6B). The plasma miR-30a-5p demonstrated a statistically significant positive correlation with the Prevotellaceae NK3B31 group (r = 0.4065, p = 0.0318). Conversely, an inverse relationship was discerned between the plasma miR-30a-5p and Fusobacterium (r = −0.4777, p = 0.0101).

DSM 27449 was isolated from a healthy individual. Sub-culturing of DSM 27449 was performed according to previous studies [68,69]. Briefly, DSM 27449 was cultured in a de Man Rogosa Sharpe (MRS) broth (Becton-Dickinson, Sparks, MD, USA) at 37 °C for 18 h. After cultivation, the culture was harvested by centrifugation at 8000× g for 15 min. The supernatant was discarded, and the resulting pellet was resuspended in an MRS broth containing 20% glycerol. The suspension was then aliquoted into tubes and stored at −20 °C until use. For daily preparation, frozen aliquots of DSM 27449 were thawed and washed with a phosphate-buffered saline (PBS). The concentration was adjusted to 1 × 10⁹ colony-forming units (CFUs)/mL in PBS before oral administration.

Eight-week-old female Sprague-Dawley rats (BioLASCO Taiwan Co., Ltd., Taipei City, Taiwan) were purchased and housed under standard laboratory conditions with ad libitum access to food and water. This study complied with the ARRIVE guidelines, and all animal procedures were approved by the Institutional Animal Care and Use Committee of the National Yang Ming Chiao Tung University (protocol number: 1100802). The rats were divided into four groups (n = 8 per group): the control, letrozole, Diane-35, and DSM 27449 groups. The non-steroidal aromatase inhibitor letrozole (Cayman Chemical, Ann Arbor, MI, USA) was used to induce PCOS-like symptoms [25]. The PCOS-like rats received an oral administration of 1 mg/kg body weight of letrozole dissolved in 1% carboxymethyl cellulose (CMC) (Fujifilm Wako Pure Chemical Corporation, Osaka, Japan), while the control rats received 0.5 mL of 1% CMC for three consecutive weeks [25]. Diane-35 (Bayer Schering Pharma, Leverkusen, Germany) was prepared as one tablet suspended in 25 mL of 1% CMC and administered at a dosage of 4.5 mL/kg to rats in the treatment control group. DSM 27449 (1 × 10⁹ CFU/mL) was administered to the rats prior to PCOS induction and throughout the study period. The estrous cycle was monitored, and OGTT was performed. At the end of the experiment, all the rats were sacrificed for subsequent analyses.

Phases of the estrous cycle were determined by predominant cell types in vaginal smears [70]. Briefly, approximately 150–200 μL of a PBS was flushed into the rats’ vagina and back out two or three times to collect vaginal cells for microscopic examination [71]. The estrous cycle was classified into four phases (proestrus, estrus, metestrus, and diestrus), and the phases were identified based on the relative proportion of nucleated epithelial cells, anucleated keratinized epithelial cells, and leukocytes [72].

After the rats were sacrificed, their ovaries were harvested and immediately fixed in 10% formalin (JT Baker, Phillipsburg, NJ, USA). The ovaries were sent to the National Laboratory Animal Center (Taipei City, Taiwan) for the preparation of microscope slides with hematoxylin and eosin-stained histological sections for evaluation. The number of cystic follicles and corpora lutea were evaluated by two blinded observers.

To obtain the serum, blood samples were collected into serum-separator tubes (BD Vacutainer^®, Becton, Dickinson and Company, Franklin Lakes, NJ, USA); placed at room temperature for 0.5 h; and centrifuged at 1500× g for 15 min. To obtain the plasmas, blood samples were collected into tubes that contained dipotassium ethylenediaminetetraacetic acid (EDTA) (BD Vacutainer^®, Becton, Dickinson and Company) and centrifuged at 1500× g for 15 min. The serum or plasmas were collected, aliquoted, and stored at −80 °C for further analysis. Testosterone and HDL-cholesterol and LDL-cholesterol levels were measured using the Elecsys assay on Cobas immunoassay analyzers (Roche Diagnostics, Meylan, France). Glucose, triglyceride, and total cholesterol levels were measured using the Dri-Chem NX600 equipment (Fujifilm, Tokyo, Japan). Insulin levels were measured using an enzyme-linked immunosorbent assay kit (Cusabio, Houston, TX, USA).

IHC staining was conducted to evaluate the expression of AR, apoptosis-related proteins (Bax and Bcl-2), and autophagy-related protein (Beclin-1) in ovarian tissues. Ovarian sections were subjected to antigen retrieval in an EDTA solution for 90 min. The sections were then incubated with primary antibodies against AR (760-4605, Cell Marque Corporation, Rocklin, CA, USA); Bax (1:300 dilution, sc-7480, Santa Cruz Biotechnology Inc., Santa Cruz, CA, USA); Bcl-2 (790-4604, Ventana Medical Systems Inc., Tucson, AZ, USA); and Beclin-1 (1:50 dilution, sc-48341, Santa Cruz Biotechnology) for 2 h. The sections were washed, incubated with a specific secondary antibody for 1.5 h, and counterstained with hematoxylin. The slides were scanned, and positive staining was quantified using the ImageJ software (version 1.4.3.67, National Institutes of Health, Bethesda, MD, USA).

The ovarian tissues were harvested and stored at −80 °C until analysis. For RNA isolation from the ovaries, the total RNA was extracted from 20–30 μg of tissues using the RNeasy Mini Kit (Qiagen, Hilden, Germany), according to the manufacturer’s protocol. For RNA isolation from the plasmas, RNA fragments of <1000 nucleotides were isolated from 100 μL of the plasma samples using the LiqmiR microRNA Extraction Kit (Topgen Biotech, Kaohsiung City, Taiwan), according to the manufacturer’s instructions. The RNA isolated from the ovaries and plasmas was stored at −80 °C for further use. A complementary DNA synthesis was performed using the miRscript RT Kit (Topgen Biotech) based on reverse transcription with the specific miRNA stem-loop reverse-transcribed primers listed in Table S2. A real-time polymerase chain reaction (PCR) was performed on a StepOnePlus Real-Time PCR system (Applied Biosystems, Foster City, CA, USA) using the TopGE Probe qPCR Master Mix and TopQ miRNA probe qPCR assay for rno-miR-30a-5p, rno-miR-93-5p, rno-miR-223-3p, rno-miR-320-3p, rno-miR-324-3p, rno-miR-146a-5p, and U6 snRNA or cel-miR-39-3p (Topgen Biotech). Amplification data were normalized to U6 snRNA and cel-miR-39-3p expressions in the ovarian and plasma samples, respectively. The relative expression was quantified using the 2^-∆∆Ct relative quantification method [73].

Fresh fecal samples were collected into sterile tubes with the RNAlater Stabilization Solution (Thermo Fisher Scientific, Vienna, Austria) and stored at −80 °C for further analysis. Microbial DNA was extracted and purified using previously described techniques [69]. Briefly, a DNA-extraction buffer, glass beads, and phenols were added to the fecal mixture, homogenized, and centrifuged at 12,000× g for 5 min. The supernatants were collected for phenol-chloroform extraction and ethanol precipitation. The extracted DNA was dissolved in 60 μL of nuclease-free water and stored at −80 °C.

The full-length 16S rRNA gene (V1-V9) was amplified using the bacterial universal primer sets 27F: AGRGTTYGATYMTGGCTCAG and 1492R: RGYTACCTTGTTACGACTT, with PacBio barcodes in the PCR system. The PCR products were examined using 0.8% agarose gels and the Qubit 2.0 Fluorometer (Thermo Scientific, Waltham, MA, USA). Qualified PCR products were purified using AMPure PB beads (PacBio, Menlo Park, CA, USA) after mixing in equal amounts. The 16S rRNA amplicons were pooled and used to create an SMRTbell library. Single-molecule real-time (SMRT) sequencing was performed using a PacBio Sequel IIe sequencer with Chemistry version 2 (Pacific Biosciences, Menlo Park, CA, USA). After primary filtering on the Sequel IIe System, a secondary analysis was conducted using the SMRT analysis pipeline version 11.0.0 (Pacific Biosciences). A PacBio Sequel sequencer was used to sequence amplicon libraries (Genomics BioSci & Tech Co., New Taipei City, Taiwan). Sub-reads were transferred to circular consensus sequencing reads and demultiplexed. The filtered sequences were clustered into operational taxonomic units at 99% identity using the Mothur (version 1.44.0) software with the SILVA database (SILVA 132) and further analyzed using QIIME (version 1.9.0). An alpha-diversity analysis was performed using a species-richness estimator (Chao1) and species-evenness estimators (Shannon and Simpson). Beta diversity was analyzed using NMDS. The analysis of similarities, a non-parametric multivariate analysis, was applied to calculate the ranked dissimilarity matrix. The Mann–Whitney U test was employed to distinguish variations in the relative abundance of microbes, and the Bonferroni method was applied to correct for Type I errors.

The software GraphPad Prism 9 (GraphPad Software Inc., La Jolla, CA, USA) was used to process the experimental data, and all data are shown as a mean ± the standard error of the mean. Between-group differences were analyzed using a one-way or two-way analysis of variance followed by Tukey’s post-hoc test, Welch ANOVA followed by Dunnett’s T3 multiple comparisons test, or the Kruskal–Wallis test followed by Dunn’s multiple comparisons test, depending on the data distribution and assumption requirements. Spearman’s correlation coefficient was used for a correlation analysis of miRNA expressions and PCOS-related parameters, microbial genera and PCOS-related parameters, and microbial genera and miRNA expressions. The non-parametric Mann–Whitney U test was used to evaluate statistical significance in the relative abundance of bacteria between two groups, and the Bonferroni method was applied to correct for Type I errors.