Animals

Healthy female Sprague-Dawley (SD) rats (6-8 weeks old, weight 240 ± 10 g) were purchased from the Experimental Animal Center of Ningxia Medical University. All animal experiments were conducted with every effort to minimize animal suffering and were approved by the Ethics Committee of Ningxia Medical University (IACUC No. Iacuc-nylac-2023-094). The animals were kept in a temperature-controlled chamber for a week (temperature maintained at 22 ± 2°C and air humidity controlled at 40-70%) and subjected to a 12-hour cycle of light and darkness in a polycarbonate cage. The feed was purchased from Beijing Gaoli Feed Co., LTD. (crude protein 46.65%, moisture 20.73%, crude fat 0.09%, crude ash 0.13%, crude fiber 0.07%, trace elements calcium and phosphorus).

Isolation and culture of human placental mesenchymal stem cells (PMSCs)

We collected full-term placenta tissue from healthy mothers in strict compliance with routine elective cesarean section procedures at Ningxia Medical University General Hospital. Our meticulous process ensures the high quality and purity of the research materials. We guarantee that all research activities are based on the explicit informed consent of the mother and that the research protocol has been approved by the Human Research Ethics Committee of Ningxia Medical University (IACUC No. KYLL-2023-0269) to maintain the legitimacy and transparency of the research. The collected placental tissue was finely trimmed by ophthalmic scissors, avoiding blood vessels and accurately obtaining the chorionic tissue. Subsequently, the obtained placenta tissue was thoroughly rinsed more than three times with clean phosphate-buffered saline (PBS, 0.0067 M PO4, pH 7.0-7.2) to ensure clean and contaminant-free. The rinsed tissue is finely broken by mechanical action and then transferred to the centrifuge tube, DMEM culture-medium (GIBCO, Thermo-Fisher Scientific, USA), collagenase type A of 1 mg/mL(GIBCO, Thermo-Fisher Scientific, USA), and DNase enzyme of 0.05 mg/mL(GIBCO, Thermo-Fisher Scientific, USA). The tissues were digested for two hours at a constant 37°C water bath. Once this was done, we removed the waste liquid by centrifugation, resuspended the cells with PBS, and filtered them step by step using a two-stage cell sieve filtration system (100 μm vs. 40 μm cell sieve) to obtain a pure cell suspension. After centrifuging the collected cell suspension, we used a serum-free mesenchymal stem cell culture medium (Yinfeng Biological Group Co. Ltd, Jinan, Shandong Province, China) for re-suspension. The cells were inoculated in a 10 cm sterile petri dish and cultured under constant conditions of 37°C and 5% CO₂. To ensure cell quality and activity, we change the culture medium every two days. The first cultured cells were defined as primary cells [19]. When the cells had grown to a 70-80% fusion state, we performed cell passage using TrypLE™ Express (1x concentration, GIBCO, Thermo-Fisher Scientific, USA). This paper used the fifth-generation PMSCs as experimental materials in the subsequent cell therapy experiments.

Identification and differentiation of PMSCs

First, cells cultured to the third generation were digested and collected with TrypLE™ Express. Next, we adjusted the cell concentration to 1 × 10⁶ cells/mL using a PBS suspension. Subsequently, we added PE-labeled CD45, CD73, CD90, and CD105 antibodies and FITC-labeled HLA-DR, CD34, and CD14 antibodies (BD, USA) to test tubes containing 200 μL cell suspension, strictly following the manufacturer’s operating guidelines. To ensure the accuracy of the results, we also set up a negative control group. All test tubes were incubated at room temperature for 30 min away from light, and the cells were washed with PBS and then detected by flow cytometry.

We inoculated human placental mesenchymal stem cells cultured to the 5^th generation into a 6-well plate, ensuring that the cell density per well was 1 × 10⁵ cells/mL. When the degree of cell fusion approached 80%, we initiated the differentiation process. According to the product description, osteogenic, lipogenic, and chondrogenic differentiation fluids were used to induce the differentiation of these cells. At the same time, control cells continued to be added to the mesenchymal stem cells-specific culture medium. After 21 days of induced differentiation, the cells in the differentiated group were meticulously fixed and stained. First, the cells were fixed at room temperature for 30 min using 4% paraformaldehyde. We then stained the cells with three specific staining agents, alizarin red, oil red O, and Yasin blue, respectively, to distinguish adipoblast, osteoblast, and chondroblast. Finally, they were carefully observed under a microscope and photographed to verify the effect of induced differentiation on different cell types.

POI model construction and PMSC treatment experimental animal grouping

As shown in Fig 1, 40 female SD rats aged 6 to 8 weeks were selected for this study. After one week of adaptive feeding, they were equally divided into four groups with ten rats in each group. The use of a random number table method for group division ensured the fairness and objectivity of the study. The groups were as follows: the CON group, serving as the control group, received the same amount of PBS intraperitoneal injection as the other groups, providing a baseline for comparison. The MOD group, as the POI (premature ovarian failure) animal model group, received an intraperitoneal injection of 1.5mg/kg cisplatin solution daily for seven consecutive days. After POI modeling, the MED group was given 0.3mg/kg/d of estradiol and 3mg/kg/d of dydrogesterone for 14 consecutive days according to Femoston instructions.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 1. Flow chart of animal experiment.

https://doi.org/10.1371/journal.pone.0313763.g001

The PMSC group was used to study the treatment of cisplatin-induced POI models with PMSCs, a promising regenerative therapy due to their immunomodulatory and tissue repair properties. After establishing the POI model, the rats’ body weight and estrus cycle were diligently monitored throughout the experiment. The POI model was evaluated by measuring ovary weight, steroid hormone level, and morphological observation.

Estrous cycle monitoring and ovarian follicle count in rats

Vaginal smears of rats were collected daily at 9:00 am, and the stages of the estrous cycle were determined by Wright-Giemsa staining, which was performed by microscopically analyzing the dominant cell types in the smears. The estrus cycle consists of four stages [20], composed of different dominant cells: the preestrus was mainly composed of nucleated epithelial cells; In the estrus period, the keratinoid epithelial cells were dominant. The late stage of estrus contains white blood cells, nucleated epithelial cells, and keratinized epithelial cells. White blood cells were the main cell types of interest. After sampling, the ovarian tissue samples were immediately fixed with 4% paraformaldehyde (Servicebio, Wuhan Servicebio Technology Co., LTD., Wuhan, Hubei, China), then paraffin-embedded and cut into 4μm slices. To assess ovarian damage, ten sections (total n =  8) were selected from each sample placed on glass slides and treated with hematoxylin and eosin (H&E) staining (Solarbio, Beijing, China). Subsequently, two experts, unaware of the sample source, used an Olympus optical microscope (Melville, NY, USA) to analyze the follicles (primary, secondary, mature, and atretic) in the section classified and counted [21]. The percentage of follicles at each stage was calculated and compared between groups.

Rat sample collection and handling

Our experimental protocol, meticulously designed to ensure the utmost accuracy of the results, is of paramount importance to our work. The rats in each group were required to fast for 12 hours after the last dose, and then to be anesthetized by inhaling isoflurane. During anesthesia, we continuously monitored the rats’ respiratory status and corneal reflexes, demonstrating our commitment to their welfare and the reliability of the experimental results. Once the rat was successfully anesthetized, we placed it in a supine position and immobilized its limbs. The skin of the chest and abdomen was thoroughly disinfected and the heart was carefully exposed with surgical instruments for cardiac blood extraction. After blood collection, the blood in the collection vessel was left to rest for 30 min and centrifuged for 10 min at 4°C at 12000 × g. After centrifugation, we removed the supernatant and subpack to label it. To maintain the activity of the sample, we quickly froze it in liquid nitrogen, and then transferred it to the -80°C refrigerator for storage for subsequent experiments. After the blood collection procedure was completed, bilateral ovarian tissue was further collected. One ovary was immobilized in 4% paraformaldehyde for subsequent HE staining and immunohistochemical analysis. The other ovary was rinsed with normal saline (Shandong Qidu Pharmaceutical CO., LTD.) to remove excess tissue, such as fat and fallopian tubes, leaving only the ovarian tissue itself. The tissue was then placed in a centrifuge tube and immediately transferred to a -80°C ultra-low temperature refrigerator for further use.

Determination of steroid sex hormones

To accurately assess steroid sex hormone levels in rats, we followed the instructions of Proteintech (Wuhan, China) enzyme-linked immunosorbent assay (ELISA) kit and performed detailed tests on collected plasma samples. Specifically, we measured concentrations of follicle-stimulating hormone (FSH), luteinizing hormone (LH), estrogen (E2), testosterone (T), progesterone (PROG), and prolactin (PRL); the detection sensitivities of these indexes were 0.1 IU/L, 0.1 mIU/ml, 0.1 pmol/L, 1.0 pg/mL, 0.1 nmol/L and 0.1 ng/mL, respectively. To ensure the accuracy and reliability of the data, we conducted three independent measurements for each sample with utmost precision. By measuring each pore’s absorbance value (OD value) at 450 nm and combining it with the standard curve, we calculated the exact levels of FSH, LH, E2, T, PROG, and PRL in rat plasma.

Plasma and ovarian inflammatory markers and determination of LPS

The measurements of inflammatory cytokines in plasma and ovary, including IL-1β, IL-6, IL-10, IL-33, and TNF-α, were conducted with thoroughness and attention to detail using ELISA kits provided by Proteintech (Wuhan, China). The sensitivity of IL-1β, IL-10, and IL-33 was 0.1 pg/mL, and the sensitivity of IL-6 and TNF-α was 1.0 pg/mL, respectively. Each sample was independently measured three times to ensure the accuracy of the results. In addition, the plasma LPS levels in each group were measured using Limulus reagents provided by Xiamen Baiyuan Technology Co., LTD. (Xiamen, China). The procedure was meticulously followed: 50 μL of diluted plasma (diluted with endotoxin-free water in a ratio of 1:4) was added to each hole of the 96-well plate, and then 50 μL of Limulus reagent was added at the initial time point. After incubation at 37°C for 30 min, 100 μL of chromogenic substrate was added to each well and continued incubation at 37°C for 6 min. Finally, a termination solution was added to stop the reaction, and the optical density values were measured at 545 nm using a microplate reader, ensuring the validity and reliability of the results.

Intestinal microbiota analysis

Sample collection and pretreatment: At 24 hours before the start of the experiment, six rats were randomly selected from each group and transferred to separate sterile cages to collect their fresh feces. Immediately after collection, the fresh feces were frozen in liquid nitrogen and stored at − 80°C for subsequent DNA extraction. To extract total DNA, first weigh 0.5 g of feces into 2 mL centrifuge tubes, add 1 mL of SLX-Mlus Buffer and 500 mg of magnetic beads, and grind using a grinder at 250 seconds and 45 Hz parameter setting. The total DNA was then extracted according to the instructions on the E.Z.N.A.^® Soil DNA Kit. To verify the quality of the DNA, we used 1% agarose gel electrophoresis and measured the concentration and purity of the DNA with a Nanodrop 2000 c U.V.-vis Spectrophotometer. PCR amplification targets the V3 and V4 hypervariable regions of the 16S rRNA gene analysis using specific primers (338F-ACTCCTACGGGAGGCAGCAG and 806R-GGACTACHVGGGTWTCTAAT) [22]. The PCR reaction mixture including 4 μL 5 ×  Fast Pfu buffer, 2 μL 2.5 mM dNTPs, 0.8 μL each primer (5 μM), 0.4 μL Fast Pfu polymerase, 10 ng of template DNA, and ddH₂O to a final volume of 20 µL. PCR amplification cycling conditions were as follows: initial denaturation at 95°C for 3 min, followed by 27 cycles of denaturing at 95°C for 30 s, annealing at 55°C for 30 s and extension at 72°C for 45 s, and single extension at 72°C for 10 min, and end at 4°C. PCR product processing and sequencing: To confirm the quality and purity of PCR products, we used 2% agarose gel electrophoresis. Subsequently, an equal volume of 1 × loading buffer (including SYB green) was mixed with PCR products, and agarose gels containing target DNA were purified using the Qiagen Gel Extraction Kit. After purification, the PCR products were quantified using Quantus™ Fluorometer (Promega, Madison, Wisconsin, USA). Libraries were generated by the TruSeq DNA PCR-Free Sample Preparation Kit (Illumina, San Diego, USA), and whole genome sequencing was carried out using the Illumina platform NovaSeq sequencer to obtain 250 bp peer read data. Data analysis: The received data is first subjected to strict quality control by Fastp software and sequentially spliced by FLASH software. Then, using UPARSE software, we performed OTU cluster analysis of the non-redundant sequences with 97% similarity. In this process, we use UCHIME software to eliminate potential chimera sequences and use the RDP classification method combined with the Silva database (SSU132) for detailed species classification annotation of sequences. This thorough data analysis process ensures the validity of our findings. These data will be used for OTU relative abundance analysis, α and β diversity assessment, and comparison of typical and endemic OTUs between samples to study differences in community composition across samples at different taxonomic levels. Finally, we will use correlation analysis software to explore the potential relationship between strains and abnormal indicators. This research has the potential to inspire and motivate further studies in the field of microbiology and genetics.

Determination of SCFAs in rat stool

We used the gas-mass spectrometry system (Thermo Scientific™ TRACE™ 1310) to accurately measure SCFAs in rat stool samples. In the process of chromatographic analysis, we selected the Agilent HP-INNOWAX capillary column (30 m × 0.25 mm, film thickness 0.25 μm), and the sample was injected with 1 μL sample volume and 10:1 shitter ratio. The inlet and transmission line temperature is 250°C, and the ion source temperature is 300°C. In terms of temperature control, we use a programmed temperature rise method: the initial temperature of 90°C rises to 120°C at 10°C/min, then rises to 150°C at 5°C/min, and finally rises to 250°C at 25°C/min quickly and maintains it for 2 min. In the whole process, the carrier gas is helium with a flow rate of 1.0 mL/min. For mass spectrometry, we selected a selective ion monitoring (SIM) scanning mode, combined with an electron bombardment ionization (EI) source with an electron energy of 70 eV, to ensure the acquisition of high-quality mass spectrometry data. To ensure measurement accuracy, we configured standard curves involving precisely weighed pure standards such as acetic acid, propionic acid, butyric acid, isobutyric acid, valeric acid, isovaleric acid, and caproic acid. We formulated a series of mixed standard solutions with a concentration gradient using ether. These standard solutions are stored at 0°C and used to draw standard curves. The curve is based on the linear regression relationship between the ratio of the concentration of the standard substance and the corresponding peak area, and the linear regression equations for all substances show a high correlation (correlation coefficient > 0.99). In addition, we verified the stability and reliability of the measurement method by evaluating precision, recovery rate, repeatability, limit of quantitation, and QC quality control results. Metabolite extraction and calculation: In the metabolite extraction phase, we first moved a precisely weighed fecal sample into a 2 mL centrifuge tube and added 50 μL of 15% phosphoric acid, 100 μL of internal standard (isocaproic acid) solution with a concentration of 125 μg/mL, and 400 μL of ether. After homogenization for 1 min and centrifugation for 10 min (12000 × g, 4°C), the supernatant was taken for machine test. Finally, the SCFAs content was calculated according to the following formula: Content (μg/g) =  (C × 0.5)/sample size × 1000, where C unit is μg/mL and sample size unit is mg.

Statistical analysis

GraphPad Prism 7 (GraphPad Software, San Diego, California, USA) was utilized. The data were considered significant at p < 0.05 after using the one-way analysis of variance (ANOVA), followed by Tukey’s post hoc statistical test. The results were presented as mean±standard deviation (SD).

Culture and identification of PMSCs in vitro

After isolation and purification, PMSCs extracted from fresh placenta showed unique cell morphology and excellent differentiation potential. These cells showed the characteristics of adhesive growth, the shape of a long spindle, tightly intertwined in a swirl arrangement, and the cytoplasm filled. The nucleoli was visible, arranged, parallel, or rotated, morphologically similar to fibroblasts (Fig 2A). Under specific induction conditions, PMSCs exhibited the ability to differentiate into a variety of cell types. Among them, after 21 days of lipid-induced culture, oil red O staining showed the formation of significant bright red lipid droplets in the cells, indicating that PMSCs had successfully differentiated into adipocytes (Fig 2B). After 21 days of culture under osteogenic induction conditions, alizarin red staining revealed the deposition of a large amount of red mineralized matrix, indicating the potential of PMSCs to differentiate into osteoblasts (Fig 2C).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 2. Morphological observation and identification of PMSCs.

A: PMSCs culture; B: PMSCs lipogenesis induced oil red O staining; C: PMSCs osteogenesis induced alizarin red staining; D: PMSCs chondrogenesis induced Yasine blue staining; E: PMSCs Surface Marker detection (A-D:200x).

https://doi.org/10.1371/journal.pone.0313763.g002

Similarly, when PMSCs were cultured under chondroblast-induced conditions for 21 days, Asin blue staining showed that all cells were blue because the cells expressed type II collagen, which is specific to chondrocytes, confirming their ability to differentiate into chondroblasts (Fig 2D). Combined with these experimental results, we could confirm that PMSCs isolated from placental tissues were capable of differentiation into adipocytes, osteoblasts, and chondroblasts in three ways. Flow cytometry analysis further revealed that mesenchymal stem cell markers CD105, CD90, and CD73 were highly expressed on the surface of these PMSCs, while hematopoietic stem cell markers CD34, CD45, monocyte marker CD14, and HLA-DR were at very low or no expression on the cell surface (Fig 2E).

General state, body weight, and food intake of rats in each group

Beginning with the adaptive feeding phase, we measured the rats’ weight regularly every other day. The results showed that the rats in the CON group had good general health, entire mental state, bright hair, and average daily eating and excretion. The MOD rats, on the other hand, showed an unhealthy state, with slow movement and dark, yellowish hair. Before modeling, all groups had no significant difference in body weight (P > 0.05). However, statistical analysis of body weight after modeling showed that compared to the CON group, rats in the MOD group showed a significant decrease in body weight (P < 0.0001, Fig 3A). To assess the feeding of the rats more accurately, we adopted a calculation method of feeding rations (500 grams per cage) and measuring the remaining amount after 24 hours. The results showed that the food intake of the MOD group was notably lower than that in the CON group (P < 0.0001, Fig 3B). The body weight and food intake after PMSCs intervention showed no significant difference.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 3. Comparison of body weight and food intake of rats in each group.

A: Analysis of body weight difference among each group; B: Comparison of food intake difference among all groups (****p <  0.0001).

https://doi.org/10.1371/journal.pone.0313763.g003

Ovarian weight and ovarian index in diverse groups

Compared with the CON and PMSC groups, the ovarian weights of rats in the MOD group were significantly reduced. At the same time, the ovarian weight of rats in the MED group showed decreased compared with the CON group (P < 0.01 in the MOD group versus the CON group, P < 0.05 in the MOD group versus the PMSC group, Fig 4A). By accurately calculating the ovarian index of rats, we observed that the ovarian index in the MOD group was significantly lower than that in the CON or PMSC groups (P < 0.01 in the MOD group vs. the CON and PMSC groups, Fig 4B, 4C). These accurate and precise findings suggested that intervention with PMSCs significantly improved ovarian weight and index in POI (primary ovarian insufficiency) rats, showing its potential therapeutic effect.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 4. Comparison of ovarian weight and ovarian index of rats in each group.

A: Comparison of ovary weight in each group; B: Comparison of ovarian index in each group; C: unilateral ovarian tissue (*p <  0.05, **p < 0.01).

https://doi.org/10.1371/journal.pone.0313763.g004

PMSCs intervention rectified the estrous cycle of rats with POI

We compared the differences in the estrous cycle of each group of rats by a vaginal smear assay using Wright-Giemsa staining. The results showed that the rats in the CON group presented a standard estrus cycle, including pre-estrus, estrus, post-estrus, and inter-estrus, and the whole cycle lasted steadily for 4-5 days, as shown in Fig 5A–5E. However, after cisplatin-induced POI, the estrous cycle of MOD group rats showed significant interference, which was characterized by irregular cycles and mainly stayed in the late estrous period (Fig 5G). It is worth noting that after the intervention of PMSCs or estrogen drugs, we observed that the estrous cycle of these two groups of rats gradually restored to a complete and regular cycle state in the late intervention period (Fig 5F and 5H).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 5. Comparison of the estrus cycle of rats in each group.

A: vaginal smear staining during pre-estrus; B: Vaginal smear staining during estrus; C: vaginal smear staining during post-estrous; D: vaginal smear staining during inter-estrus; E: estrous cycle of CON group; F: estrous cycle in MED group; G: estrus cycle of MOD group; H: estrous cycle of PMSC group (A-D:200×).

https://doi.org/10.1371/journal.pone.0313763.g005

PMSCs intervention attenuated the pathological damage of ovarian tissue in rats with POI

The HE staining was used to observe and analyze the pathological changes of ovaries in each group in detail (Fig 6A–6D). In the CON group, the follicles within the ovarian tissue showed normal morphology at different stages of development. However, in the MOD group, the ovarian tissue structure showed significant disorder, as the number of primary, secondary, and mature follicles decreased significantly (P < 0.001, Fig 6C, 6E). In contrast, the atretic follicles increased (P < 0.05, Fig 6C, 6E). These changes indicated the impaired ovulation in the cisplatin-induced POI model. However, the study brings a ray of hope as it is gratifying to note that the ovarian tissue structure of the rats was significantly improved after treatment with PMSCs. Specifically, primary, secondary, and mature follicles increased significantly (P < 0.05). In contrast, the number of atretic follicles decreased significantly (P < 0.01), indicating that PMSCs could dramatically improve the abnormal status of ovarian follicles in POI rats. This potential of PMSCs offers a promising outlook for the treatment of POI.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 6. Comparison of HE staining results and the number of follicles at different levels in ovarian tissue of rats in each group.

A: The hematoxylin-eosin (H&E) stain of CON group ovarian tissue; B: The hematoxylin-eosin (H&E) stain of ovarian tissue in MED group; C: The hematoxylin-eosin (H&E) stain of MOD group ovarian tissue; D: The hematoxylin-eosin (H&E) stain of ovarian tissue in PMSC group; E: Comparison of the number of follicles at all levels in each group (A-D:40 × , * p < 0.05, **p < 0.01, ***p < 0.001).

https://doi.org/10.1371/journal.pone.0313763.g006

PMSCs intervention ameliorated the homeostasis of steroid hormones in rats with POI

Six steroid sex hormones including follicle-stimulating hormone (FSH), luteinizing hormone (LH), estrogen (E2), testosterone (T), progesterone (PROG), and prolactin (PRL), were accurately quantified. After statistical analysis, we found that FSH levels in the MOD group were significantly increased compared with those in the CON group (P < 0.01). In contrast, FSH levels were significantly decreased after PMSCs intervention (P < 0.05, Fig 7A). At the same time, the E2 level in the MOD group was considerably lower than that in the CON or MED groups (P < 0.05, Fig 7B). Although there was no statistically significant difference in E2 level between the MOD group and the PMSC group (P = 0.0534, Fig 7B), there was a trend difference between the two groups. In addition, E2 level in the MOD group was also lower than that in the MED group (P < 0.05, Fig 7B). Another notable finding was that T level was elevated in the MED group compared to the MOD group (P < 0.05, Fig 7C). However, no significant difference was found in the levels of PROG and PRL among the groups (P > 0.05, Fig 7D, 7F). In addition, LH level in the MOD group was significantly lower than that in the CON and MED groups (P < 0.01 and P < 0.05, Fig 7E). These results suggested that PMSCs intervention significantly improved hormone levels in POI rats by predominantly regulating FSH and E2.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 7. Comparison of steroid hormone levels in rats in each group.

A: Follicle-stimulating hormone FSH level; B: Estrogen E2 level; C: Testosterone T level; D: Progesterone PROG level; E: luteinizing hormone LH level; F: Prolactin PRL level (*p < 0.05, **p < 0.01).

https://doi.org/10.1371/journal.pone.0313763.g007

Changes in levels of inflammation-related factors in each group

As shown in Fig 8, plasma level of proinflammatory cytokine IL-1β in the MOD group was significantly higher than that in the CON and MED groups (MOD vs CON P < 0.01, MOD vs MED P < 0.001, Fig 8A). Similarly, the level of IL-1β in ovarian tissue of the MOD group also exerted an increase compared with that in the CON group (P < 0.01) and MED group (P < 0.05, Fig 8A). In addition, IL-6 level in ovarian tissue of the MOD group was significantly elevated compared to the CON and MED groups (P < 0.001, Fig 8B). The level of the anti-inflammatory factor IL-10 in ovarian tissue was reduced in the MOD group and MED group (P < 0.001, Fig 8C). In plasma, TNF-α level was down-regulated only in the MED group (P < 0.05), but in ovarian tissue, TNF-α level was higher in the MED group than that in the CON group (P < 0.001, Fig 8D). It was noteworthy that the ovarian TNF-α level in the MOD group was significantly higher than that in the CON group (P < 0.001, Fig 8D). The plasma IL-33 level in the MOD group was also higher than that in the MED group (P < 0.05, Fig 8E). Notably, after intervention with PMSCs, multiple proinflammatory cytokines in POI rats were significantly inhibited, including IL-1β in plasma (P < 0.001, Fig 8A), IL-6 in ovarian tissue (P < 0.05, Fig 8B), and TNF-α in ovarian tissue and plasma (P < 0.01, Fig 8D), and plasma IL-33 (P < 0.05, Fig 8E). Meanwhile, the anti-inflammatory factor IL-10 in ovarian tissue was significantly up-regulated after the intervention (P < 0.01, Fig 8C). These results suggested that PMSCs significantly improved POI in rats by inhibiting the inflammatory response.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 8. Comparison of levels of inflammatory cytokines and LPS in each group.

A: IL-1β levels in plasma and ovarian tissue of rats in each group; B: IL-6 levels in plasma and ovarian tissue of rats in each group; C: IL-10 levels in plasma and ovarian tissue of rats in each group; D: Levels of TNF-α in plasma and ovarian tissue of rats in each group; E: Plasma IL-33 level of rats in each group; F: LPS levels in each group (*p < 0.05, **p < 0.01, ***p < 0.001).

https://doi.org/10.1371/journal.pone.0313763.g008

PMSCs intervention improved the endotoxemia in rats with POI

To further investigate the effects of PMSCs on gut dysbiosis, particularly on gram-negative bacteria, permeability, and intestinal barrier integrity, the pathogenic bacteria-derived LPS that translocated from intestine to plasma was measured after PMSCs treatment. The results showed that the plasma LPS concentration in the MOD group was significantly higher than that in the CON group, indicating the presence of endotoxemia in POI rats (P < 0.05, Fig 8F). After PMSC intervention, LPS levels in peripheral blood showed a downward trend, suggesting that PMSCs may play a role by affecting intestinal flora.

PMSCs intervention modulated the difference of overall community structure of intestinal microbiota

Subsequently, to study the effects of PMSCs on the structure and abundance of intestinal flora in POI rats, we applied 16S rRNA sequencing to conduct a comprehensive analysis of fecal samples. First, we used sparse curves to evaluate the rationality of each group of rats’ intestinal flora sequencing data. This evaluation method collects the α-diversity index of the corresponding samples by randomly sampling the sequences. It drawed the α-diversity index curve that changes with the increased amount of extracted data. In this process, we pay special attention to Sobs (number of observed species) and Shannon index as evaluation indicators. The results showed that the data tended to be stable when the number of sequenced samples in each group reached 4000. This ensured that our sequencing data was sufficient to cover all microbial species in the samples with high reliability and rationality, as shown in Fig 9A.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 9. Analysis of the difference in overall bacterial community structure in each group.

A: Dilution curves of 16S rRNA high-throughput sequencing groups were used to evaluate the rationality of sequencing data; B: α diversity of Simpson and Shannon indices; C, D: β diversity of PCoA based on weighted UniFrac distance and NMDS based on Bray Curtis; E: LEfSe analysis, LDA >  3. N =  5/ group; The data were expressed as mean ± SEM and analyzed by unpaired Student’s t-test (*p < 0.05, **p < 0.01, ***p < 0.001).

https://doi.org/10.1371/journal.pone.0313763.g009

Further analysis showed no significant difference in α diversity of gut microbiome by the Shannon index between the MOD and PMSC groups. However, community diversity was significantly lower in the MED group than that in the CON and MOD groups (P < 0.05), and the difference was statistically significant (Fig 9B). To better understand the differences in gut microbial community structure among groups, we assessed β diversity using unweighted principal coordinate analysis (PcoA) and weighted distance matrix (non-metric multidimensional scale, NMDS). Both PcoA and NMDS analysis revealed significant differences in the intestinal microbial community structure of rats in different groups, and the clustering patterns among the groups were significantly separated (Fig 9C, 9D). In addition, to identify specific bacterial taxa that were closely related to each group, we performed LEfSe analysis. Analysis results showed that the MOD group was enriched in Rikenella, Micrococcaceae, Bacteroidales, and Desulfovibrionaceae, while the CON group enriched in Lachnospiraceae and Clostridia (Fig 9E). These findings provided important clues for us to understand further the effects of PMSCs on POI by modulating the intestinal microbiota. Specifically, PMSCs might exert their effects on POI by modulating the balance of these microbial communities, potentially reducing inflammation, improving intestinal barrier function, and enhancing motility. The enrichment of beneficial bacteria and suppression of harmful microbes could be key mechanisms through which PMSCs alleviate POI symptoms.

The difference in intestinal flora in rats of each group at the phylum level

To further compare the differences among the diverse groups of gut microbiota, we conducted a detailed analysis of the gut microbiota at the phylum level. Firmicutes and Bacteroidetes dominate the intestinal bacterial community at the phylum level, accounting for about 90% of the entire bacterial community structure. In addition, Actinobacteria and Desulfobacterota also accounted for a certain amount, about 8%, while the abundance of other bacteria remained relatively low (Fig 10A). Specific to the experimental groups, the intestinal microbial community of the MOD group was Bacteroidetes as the main dominant bacteria, while the CON group was significantly inclined to Firmicutes (Fig 10B). Meanwhile, compared to the CON group, in the MOD group, the proportions of Actinomyces (P < 0.01), unclassified_k_norank_d_Bacteria (P < 0.05), Proteobacteria (P < 0.05), Cyanobacteria (P < 0.05), Desulphurizing bacteria (P < 0.05) and Deferrobacteria (P < 0.05) were significantly increased. This showed that cisplatin significantly affected the composition and structure of the intestinal flora in rats. Further, compared with the MOD group, the flora distribution of Firmicutes, Bacteroides, and Proteobacteria in the PMSC group showed significant differences (P < 0.05, Fig 10C, 10D), indicating that the intestinal flora composition of the PMSC group was more similar to that in the CON group. However, there was no significant difference in the distribution of Firmicutes and Campilobacterota between the MED group and the PMSC group (P > 0.05, Fig 10E). These findings provided valuable insights into understanding the effects of cisplatin on gut microbiota and the potential regulatory role of PMSCs.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 10. The relative abundance of different microflora at intestinal portal level of rats in each group was different.

A: Relative abundance of microbial species at phyla level; B: Comparison of intestinal microbial species abundance in CON/MOD group; C: Comparison of intestinal microbial species abundance in MOD/PMSC groups; D: Comparison of intestinal microbial species abundance in MED/MOD group E: Comparison of intestinal microbial species abundance in MED/PMSC group (*p <  0.05, **p <  0.01, ***p <  0.001).

https://doi.org/10.1371/journal.pone.0313763.g010

The difference of intestinal flora in rats of each group at genus level

On the genus level, there were differences in microflora among all groups (Fig 11A).Compared to the CON group, the proportions of Unclassified_f_Lachnospiraceae (P < 0.05), norank_f_norank_o_Clostridia_UCG-041 (P < 0.05) and norank_f_norank_o_RF39 in the MOD group were decreased (P < 0.05). The relative abundance of 8 species with relatively small proportions, such as Bacteroides (P < 0.05), Bifidobacterium (P < 0.01), unclassified_k__norank_d__Bacteria (P < 0.05), norank_f_UCG_010 (P < 0.05) was significantly increased (Fig 11B).The results showed that POI rats had dysregulation of intestinal flora. After PMSC intervention, it was significant microbial community differencesat the genus level. Specifically, the abundance of Turicibacter and Desulfovibrio in the PMSC group is significantly increased. However, Alloprevotella, Parabacteroides, Rikenellaceae_RC9_gut_group and Rikenella were decreased (P < 0.05, Fig 11C). In addition, the abundances of Romboutsia (P < 0.05), Turicibacter (P < 0.05) and Dubosiella (P < 0.05) in MOD group were lower than those in MED group. The relative abundance of unclassified_k__norank_d__Bacteria (P < 0.05), UCG_005 (P < 0.05), and Oscillibacter (P < 0.05) were increased compared with those in MED group (Fig 11D). The relative abundance of Ruminococcus (P < 0.01), norank_f_UCG_010 (P < 0.05), UCG_005 (P < 0.05), Christensenellaceae_R-7_group (P < 0.05) were lower in the MED group than in the PMSC group (Fig 11E). In summary, PMSC intervention can significantly change intestine dysmicrobiota of POI in which is mainly reflected in the significant increase of Turicibacter and Desulfovibrio and the decrease of Alloprevotella, Parabacteroides, Rikenellaceae_RC9_gut_group and Rikenella.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 11. Relative abundance of different intestinal flora of rats in each group at the genus level.

A: Relative abundance of microbial species at genus level B: comparison of intestinal microbial species abundance in CON/MOD group C: Comparison of intestinal microbial species abundance in MOD/PMSC group D: Comparison of intestinal microbial species abundance in MED/MOD group E: Comparison of intestinal microbial species abundance in MED/PMSC groups (*p < 0.05, **p < 0.01, ***p < 0.001).

https://doi.org/10.1371/journal.pone.0313763.g011

PMSCs improved intestinal microbiota derived SCFAs in POI rats

Our research has uncovered significant findings about the role of PMSCs in improving SCFAs in the intestinal microbiota of POI rats. The linear regression equations and corresponding concentration ranges for each short-chain fatty acid standard are presented in Table 1. The critical metabolites of the intestinal microbiota, SCFAs mainly including acetic acid, propionic acid, butyric acid, isobutyric acid, isovaleric acid, valeric acid, and caproic acid, showed clear and easily identifiable peak shapes in the TIC chromatogram, which fully validates the accuracy of the analytical method and the reliability of the data (Fig 12A). In the RSD distribution map, the Y-axis represents the target metabolite’s RSD value; the data quality is good when the RSD value is less than 15% (Fig 12B). Global metabolite clustering heat maps reveal metabolic patterns of metabolites under different experimental conditions. Metabolites with similar metabolic patterns tend to have similar functions or participate in the same metabolic processes or cellular pathways. The relative content of metabolites in the heat map is represented by the depth of color, with darker red indicating a higher expression level and darker blue indicating a lower expression level. The columns represent the different samples, while the rows indicate the names of the metabolites. The cluster tree on the left results from cluster analysis based on the similarity between metabolite molecules to show the correlation between metabolites and their grouping (Fig 12C). After detailed comparison, we found that the levels of acetic acid (P < 0.05), butyric acid (P < 0.05), and caproic acid (P < 0.01) in the MOD group were significantly reduced compared with those in the CON group (Fig 12E, 12F, 12I). However, other SCFAs show no significant differences (Fig 12D, 12G, 12H, 12J). In the PMSC group, compared with the MOD group, the level of caproic acid was significantly increased (P < 0.05, Fig 12I). In addition, compared with the CON group, the contents of butyric acid (P <  0.05), caproic acid (P < 0.05), and propionic acid (P < 0.05) in the MED group were significantly reduced (Fig 12F, 12I, 12J). In addition, KEGG (Kyoto Encyclopedia of Genes and Genomes) provided a strong support for metabolic analysis and metabolic network research. In Fig 12K, the green nodes represented metabolic pathways, while the other nodes displayed metabolite molecules. The size of the node of a metabolic pathway reflected the number of metabolite molecules associated with it, and the larger the number, the larger the node. The color change of the metabolite molecular nodes indicated the magnitude of the log2(FC) value, and the metabolites that did not label the log2(FC) information indicated that they did not differ significantly between the diverse groups. This analysis further revealed the critical role of SCFAs such as butyric acid in body metabolism, reinforcing the reliability of our data.

Download:

• PNG
  larger image
• TIFF
  original image

Table 1. Linear regression equation and concentration range of each SCFAs standard.

https://doi.org/10.1371/journal.pone.0313763.t001

Download:

• PNG
  larger image
• TIFF
  original image

Fig 12. Determination of SCFA content in fecal stool and intestinal metabolites of rats in each group.

A: Chromatogram of rat stool sample; B: RSD distribution diagram (Y-axis indicates the RSD value distribution of the target metabolite. RSD < 15%, indicating good value); C: Cluster heat map results; D-J: Contents of isobutyric acid, acetic acid, butyric acid, isovaleric acid, valeric acid, caproic acid and propionic acid (μg/g by gas chromatography-mass spectrometry) (n =  5/ group). The data were expressed as mean ±  standard error (SEM) and were statistically analyzed using unpaired Student’s t-test: *  p <  0.05, ** p <  0.01.). K: KEGG metabolite molecular network diagram (Green dots represent metabolic pathways, other dots represent metabolite molecules (the size of the metabolic pathway site indicates the number of metabolite molecules associated with it; the larger the number, the dot). The molecular points of metabolites represent the magnitude of log2(FC) value through gradient change, and there is no metabolite log2(FC) information in many groups.

https://doi.org/10.1371/journal.pone.0313763.g012

Correlation analysis of intestinal flora with SCFAs, hormones, and inflammation in POI rats after PMSC treatment

We conducted a series of correlation analyses to analyze further the relationship between intestinal flora, hormones, inflammation, and SCFAs in POI (Fig 13). At the classification level, we observed a significant positive correlation between the abundance of Firmicutes and the content of caproic acid (P < 0.01, Fig 13A). In contrast, the abundance of Bacteroides was negatively correlated with the content of caproic acid (P < 0.05, Fig 13A). These findings suggest a direct influence of the gut microbiome on SCFA production. On the other hand, the abundance of Proteobacteria was positively correlated with inflammation-related indicators such as LPS (P < 0.05) and IL-1β (P < 0.05, Fig 13B), while negatively correlated with steroid sex hormone E2 (P < 0.05, Fig 13C), suggesting that gut microbes may indirectly influence the course of POI by influencing inflammation levels and interacting with sex hormone levels. Further analysis at the taxonomic level of the genus, we found that the abundance of Lachnospiraceae (p < 0.001) and Clostridia (P < 0.001) were significantly positively correlated with caproic acid levels (Fig 13D). In addition, the abundances of Blautia and Prevotellaceae were positively correlated with IL-33 (P < 0.05, Fig 13E) and FSH (P < 0.05, Fig 13F). These data further reinforce the vital link between gut microbes, inflammation, and POI-related hormones. We also found a significant correlation between butyric acid and various bacteria such as Desulfovibrio, Dubosiella, Faecalibaculum, Roseburia, Turicibacter, Romboutsia. These findings shed light on the complex network of interactions among gut flora, inflammation, steroid sex hormones, and metabolites SCFAs that collectively contribute to the disease progression of POI.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 13. Heat map of correlation between intestinal flora, SCFAs, hormones, and inflammation in POI rats after PMSC treatment.

A: Analysis of the correlation between bacteria and SCFAs at the gate level. B: Analysis of the correlation between bacteria and inflammatory factors at the gate level. C: Analysis of the correlation between bacteria and hormones at the gate level. D: Analysis of correlation between bacteria and SCFAs at the generic level. E: Analysis of correlation between bacteria and inflammatory factors at the generic level. F: Analysis of correlation between bacteria and hormones at the generic level. Spearman correlation analysis was used, and the intensity of correlation coefficients among samples was visually demonstrated by color depth. A total of 5 samples/groups were included in this study for analysis, and the results were marked by significance test (*p < 0.05, **p < 0.01, ***p < 0.001).

https://doi.org/10.1371/journal.pone.0313763.g013