A novel exopolysaccharide from Lactiplantibacillus plantarum H6 improves cholesterol metabolism via Muribaculum-mediated activation of the enterohepatic FXR–FGF15 axis

ABSTRACT

Hypercholesterolemia is a major risk factor for atherosclerotic cardiovascular disease; however, current therapeutic options such as statins are limited by issues including hepatotoxicity and patient intolerance. Probiotics and their metabolites show promise in modulating cholesterol metabolism through the gut‒liver axis, yet the specific commensal bacteria and molecular mechanisms underlying these effects remain poorly understood. In this study, we isolated and characterized EPS-D1, a novel exopolysaccharide (15.003 kDa) derived from Lactiplantibacillus plantarum H6, which is composed primarily of mannose (46.10%) and glucose (33.98%) and features a highly branched structure (branching degree of 29.5%). The administration of EPS-D1 significantly reduced the serum total cholesterol (TC), triglyceride (TG), and low-density lipoprotein cholesterol (LDL-C) by 40.31%, 37.55%, and 43.15%, respectively, in high-cholesterol diet (HCD) mice. Additionally, it improved hepatic steatosis and reduced markers of liver injury. Through 16S rRNA sequencing and fecal microbiota transplantation (FMT), we identified Muribaculum as the key commensal bacterium enriched by EPS-D1. Direct administration of Muribaculum (Muribaculum intestinale) replicated the cholesterol-lowering effects, decreasing ileal and fecal cholesterol levels by 74.79% and 53.16%, respectively. Mechanistically, both EPS-D1 and M. intestinale activated the enterohepatic FXR‒FGF15 axis, which resulted in the upregulation of hepatic cholesterol 7α-hydroxylase (CYP7A1) expression and the downregulation of ileal ASBT and NPC1L1, thereby promoting bile acid synthesis and inhibiting cholesterol absorption. Furthermore, M. intestinale increased intestinal short-chain fatty acids (SCFAs), particularly acetic acid and caproic acid, by 37.88% while also modulating the composition of the bile acid pool. These findings establish M. intestinale as a precise microbial target for cholesterol management and demonstrate that EPS-D1 from L. plantarum H6 enhances cholesterol metabolism through microbiota-mediated activation of the enterohepatic FXR‒FGF15 axis, providing a novel therapeutic strategy for managing hypercholesterolemia.

GRAPHICAL ABSTRACT

Introduction

High cholesterol level is the core pathological feature of metabolic syndrome, which is a major risk factor for atherosclerotic cardiovascular disease (ASCVD). Although statins are widely used in clinical practice, side effects such as hepatotoxicity, myalgia, and “statin intolerance” in some patients still need to be addressed.¹ Since cholesterol changes in the body are closely related to diet, dietary modification is considered to be one of the main strategies to prevent hypercholesterolemia. In recent years, the beneficial effects of probiotics and their metabolites have been widely noticed, while with the deeper study of the enterohepatic axis, probiotics and their metabolites affect the intestinal microbiota at the same time, influencing various metabolic disorders²; however, the current study on the reduction of in vivo cholesterol levels mediated by specific commensal bacteria in intestinal microorganisms is still in the exploratory stage.

Beneficial gut microbes can work with the body to fine-tune fat metabolism and cholesterol levels.³ In an experiment with healthy people, it was found that those with a lower baseline gut health index showed a stronger response to probiotic intervention, characterized by an increased abundance of the beneficial commensal bacterium.² Muribaculum intestinale (M. intestinale), as a core commensal genus in the gut, belongs to the family Muribaculaceae. It can metabolize foodborne and host polysaccharides, which has the ability to metabolize foodborne and host-derived polysaccharides.⁴ Its abundance is positively correlated with health. Fucosylated polysaccharide intervention effectively changed the harmful changes of intestinal microbiota in mice with colitis, and improved the relative abundance of Muribaculum taxa to maintain intestinal homeostasis.⁵ In addition, many studies have shown that it plays an important role in the regulation of cholesterol metabolism. The small molecule pectin can promote the growth of Muribaculum, improved cholesterol metabolism and reduced inflammation potential.⁶ These results suggested a potential association between Muribaculum and cholesterol reduction. However, no studies have yet systematically evaluated its functional role.

The farnesol X receptor (FXR)-fibroblast growth factor 15 (FGF-15) axis is a central pathway in the enterohepatic dialog. The activation of intestinal FXR induces the secretion of FGF15, which circulates through the portal vein to the liver to inhibit cholesterol 7α-hydroxylase (CYP7A1), thereby regulating bile acid synthesis and cholesterol metabolism. Lactobacillus johnsonii CCFM1376 could inhibit the expression of genes related to the FXR pathway in the ileum of mice and promote bile acid synthesis to improve hypercholesterolemia.⁷ Recent evidence has suggested that the intestinal microbiota can directly regulate FXR activity by modifying bile acid profiles,^8,9 but whether M. intestinale is involved in this process has not been reported.

Exopolysaccharides (EPS) are microbial extracellular glucose metabolites/polymers, including the conjugated EPS (cEPS) and the released EPS (rEPS), and various probiotics have been extensively studied for their ability to produce EPS. In previous studies, we have already investigated the cholesterol-lowering effect of active Lactiplantibacillus plantarum H6 (L. plantarum H6) and its inactivated strains,¹⁰ but it is unknown whether its specific cholesterol-lowering effect is related to secreted EPS. Fundamentally, EPS are carbon sources for gut microorganisms that are able to utilize EPS to convert them into beneficial substances, such as short-chain fatty acids (SCFAs) and bile acids (BAs).¹¹ In our previous research, it was found that both direct intragastric administration of L. plantarum H6 and fecal microbiota transplantation (FMT) could improve the phenomenon of reduced species richness and community evenness in mice induced by a high-cholesterol diet. Moreover, it was discovered that Muribaculacea, a key regulatory intestinal commensal strain, was the only strain that was enriched in all four treatments.¹⁰ However, critical microbial and host targets associated with the metabolic benefits of EPS from L. plantarum H6 remain elusive, especially its functional association with Muribaculacea.

In the present study, we investigated the effects of EPS from L. plantarum H6 on mice fed a high cholesterol diet and explored the mechanism of action. To investigate the regulatory effects of EPS on the intestinal microbiota of mice, gut microbiota sequencing and FMT were used to identify potentially beneficial intestinal commensal bacteria that regulate cholesterol metabolism by EPS. Finally, it was verified that the beneficial intestinal symbiotic bacteria were able to improve cholesterol metabolism through the EPS-regulated cholesterol metabolism pathway of action. This study demonstrated for the first time that EPS-D1 from L. plantarum H6 could improve cholesterol level by enriching the intestinal Muribaculum and regulating the FXR‒FGF15 axis in the gut and liver. This study provides a theoretical strategy for the microecological intervention of metabolic diseases with precision, safety and feasibility.

Materials and methods

Preparation of cEPS and rEPS of L. plantarum H6

L. plantarum H6 was cultured in MRS medium at 37 °C. Following 24 h of incubation, the bacterial culture was centrifuged at 4200 × g and 4 °C for 20 min to separate the supernatant and the cell pellet. For cEPS extraction, the cell pellet was washed twice and resuspended in 1 mol/L NaCl. The suspension was treated with an ultrasonic cell disruptor at 40 W and 4 °C for 3 min. The disrupted cell solution was centrifuged, and the supernatant was collected. The supernatant was concentrated using a rotary evaporator. The final concentrate was mixed with three times its volume of anhydrous ethanol to precipitate cEPS. For rEPS extraction, the 4% (w/v) trichloroacetic acid (TCA) was added to the supernatant, followed by overnight precipitation at 4 °C. The next day, the mixture was centrifuged again at 4200 × g and 4 °C for 15 min to remove proteins. The resulting supernatant was then concentrated by rotary evaporation. Subsequently, a three-fold volume of ice-cold anhydrous ethanol was added, and the solution was allowed to stand overnight at 4 °C. Flocculent precipitates formed and were collected. Dissolve the precipitated substances and add 4% trichloroacetic acid to repeat the impurity removal for three times. Perform dialysis with distilled water for 48 h, changing the dialysis solution every 6 h. Freeze-dry the dialyzed solution. Use the phenol-sulfuric acid method to detect the purity of the freeze-dried polysaccharides. Polysaccharides with a purity greater than 70% are used for subsequent experiments.

Isolation and purification of rEPS

For primary purification, the crude EPS extract was loaded onto a DEAE-cellulose column and eluted with sequentially increasing concentrations of NaCl. The fraction corresponding to the main peak of carbohydrate content (designated EPS-D1) was pooled. This fraction then underwent a secondary purification step on a Sephacryl S-400 HR column, eluted isocratically with distilled water. Throughout the purification process, the total sugar content in all the collected eluates was quantified using the sulfuric acid‒phenol method.

Characterization of EPS structure

Determination of molecular weight

The molecular weight and homogeneity of the fractions were determined by SEC-MALLS-RI. Analyses were performed using a DAWN HELEOS-II laser photometer and an Optilab T-rEX refractive index detector in tandem. The samples were eluted with a 0.1 M NaNO₃ aqueous solution through two serial columns at 45 °C and a flow rate of 0.6 mL/min. The system provided the weight–average and number–average molecular weights (M_w and M_n), the polydispersity index (M_w/M_n) and simultaneously determined the sample concentration and the specific refractive index increment (dn/dc).

Determination of monosaccharide composition

In order to analyze the monosaccharide composition, roughly 5 mg of EPS-D1 underwent hydrolysis using 2 M trifluoroacetic acid at 121 °C for a duration of 2 h in a sealed container. Following this, the hydrolysate was evaporated under a nitrogen stream, subjected to methanol washing, and subsequently redried. The resulting material was filtered through a 0.22 μm membrane prior to analysis. Monosaccharide analysis was performed using HPAEC-PAD on a Dionex ICS 5000+ system equipped with a CarboPac PA-20 column (3 × 150 mm). Data collection and processing were carried out using the Chromeleon 7.2 CDS software.

Ultraviolet (UV) and Fourier-transform infrared (FT-IR) analysis

1 mg/mL polysaccharide solution was used to analyze the samples using a UV spectrophotometer in the wavelength range from 190 to 600 nm.

Trace amount of EPS-D1 and dry KBr were mixed and ground thoroughly to make a tablet to be pressed, and then scanned at 4000–400 cm⁻¹ by FT-IR spectroscopy.

Methylation and GC–MS analysis

A small amount of EPS-D1 was dissolved in DMSO, followed by the addition of NaOH and methyl iodide to facilitate methylation. The resulting methylated product was then reacted in a TFA solution at 121 °C for 90 min. Subsequently, reduction was carried out using a NaBD4 solution at room temperature for 2.5 h, followed by acetylation with acetic anhydride at 100 °C for 2.5 h. The acetate was then dissolved in dichloromethane, and the lower layer was collected for GC‒MS analysis. The chromatographic analysis was performed with an injection volume of 1 μl, utilizing high-purity helium as the carrier gas. Mass spectrometry analysis was conducted using an electron impact ionization (EI) source. Analytes were detected in full scan (SCAN) mode, with the ion source temperature set to 200 °C, the ionization energy at 50 eV, and the mass scanning range (m/z) from 50 to 350.

Nuclear magnetic resonance (NMR) spectroscopy

EPS-D1 was dissolved in 0.5 mL D₂O to a final concentration of 40 mg/mL. 1D-NMR (¹H-NMR, ¹³C-NMR, DEPT-135) and 2D-NMR (COSY, NOESY, HMBC, HSQC and TOCSY) were recorded at 25 °C with a Bruker AVANCE NEO 500 M spectrometer system (Bruker, Rheinstetten, Germany) operating at 500 MHz.

Scanning electron microscopy (SEM) analysis

The surface of EPS-D1 powder was coated with gold, and the surface morphology was observed by scanning electron microscope, and observed at 1000×, 3000×, 5000×, and 10,000×.

In vitro determination of cholesterol-lowering activity

Cholesterol adsorption experiment, 1 mg/ml EPS-D1 and EPS-D2 were respectively added into a 50 ml centrifuge tube, and 10 ml cholesterol standard solution (10 mg/ml) was added, and no polysaccharide was used as a control. The adsorption was carried out by shaking at a constant temperature 37 °C. After 12 h, the supernatant was centrifuged at 10,000 × g for 10 min. The cholesterol content in the supernatant was analyzed by the kit, and the cholesterol adsorption rate was calculated by the formula. The formula used was as follows: cholesterol adsorption rate (%) = (cholesterol content of the blank groups−cholesterol content of the experimental group)/cholesterol content of the blank group × 100%.

Cholesterol micellar solubility experiment, 10 mmol/L sodium taurocholate, 2 mmol/L cholesterol, 5 mmol/L oleic acid, and 132 mmol/L NaCl were prepared with PBS, sonicated for 1 h, and incubated at 37 °C for 24 h to prepare cholesterol micellar. Different concentrations of EPS-D1 and EPS-D2 were mixed with 5 mL of cholesterol micelles and incubated at 37 °C for 2 h. After centrifugation, the cholesterol content in the supernatant was analyzed by the kit, and the cholesterol adsorption rate was calculated by the formula. The formula was as follows: cholesterol micelle solubility inhibition rate (%) = (sample cholesterol content−control cholesterol content)/sample cholesterol content × 100%.

Strain culture

The M. intestinale (GDMCC NO. 1.4667) used in this experiment was purchased from Guangdong Microbial Culture Collection Center. It was anoxically cultured at 37 °C for 48–72 h in blood plate medium (TSA + 5% defibrinized sheep blood). In the experiment to explore whether EPS-D1 could promote the growth of M. intestinal, the amount of EPS-D1 was added to the medium at 10 mg/mL. Ligilactobacillus salivarius QP1 (L. salivarius) was preserved in the Food Safety Laboratory of Jilin Agricultural University and cultured in MRS medium for 24 h for future use.

Animal experiments

All the animal experiments were approved by the Animal Experiment Ethics Committee of Jilin Agricultural University (animal experiment ethics numbers: 20240209002, 20241023001, and 20250310015, respectively), and the animal experiment process was strictly carried out in accordance with the requirements, strictly adhering to the international animal experiment ethics standards. Five-week-old C57BL/6 mice (Beijing Huafukang Biotechnology Co., Ltd. (SCXK (Beijing) 2019-0008). All the experimental mice were adaptively fed for one week in an SPF-level environment with a suitable temperature of 25 ± 1 °C and a light and dark cycle (12/12 h), and then subsequent experimental operations were carried out. The mice were randomly divided into different groups. During the experiment, the mice could drink water and eat freely.

Animal experiment 1: to explore the effects of two polysaccharide from L. plantarum H6 on reducing cholesterol levels in mice fed a high-cholesterol diet, the experimental mice were randomly divided into four groups, with six mice in each group, and were named as the normal diet group (ND), the model group (HCD), the cEPS experimental group (cEPS), and the rEPS experimental group (rEPS). During the experiment, the ND group was fed with maintenance feed, while the HCD group and the polysaccharide experimental groups were fed with high-cholesterol feed. At the same time, the polysaccharide experimental groups were intragastrically administered with polysaccharide solution (300 mg/kg BW), and the ND group and the HCD group were intragastrically administered with the same volume of normal saline. The experiment lasted for six weeks.

Animal experiment 2: in order to study the mechanism of L. plantarum H6 EPS-D1 on cholesterol reduction, the test mice were divided into four groups, with eight mice randomly assigned to each group, and named as the normal group (ND), model group (HCD), positive drug group (simvastatin, Sim), and EPS-D1 group (EPS). During the experiment, the ND group was fed with normal feed, while the HCD group, Sim group, and EPS group were fed with high-cholesterol feed. Meanwhile, the Sim group and EPS group were respectively administered simvastatin solution (3.8 mg/kg BW) and EPS-D1 solution (300 mg/kg BW) by gavage, and the ND group and HCD group were administered the same volume of normal saline by gavage for six weeks.

Animal experiment 3: to investigate the role of EPS in regulating the intestinal microbiota in reducing cholesterol levels in mice on a high-cholesterol diet, a FMT approach was adopted. Fresh feces were collected daily from mice in the HCD- or EPS group. Each 100 mg of fresh feces was resuspended in 1 mL of saline, and the supernatant was collected after low-temperature centrifugation. After one week of Abx gavage in recipient mice, the above supernatants (10 mL/kg BW) were gavaged daily to the recipient mice for five weeks.

Animal experiment 4: to verify the role of EPS-D1-derived intestinal microbiota, the most enriched dominant strains, Muribaculum and Ligilactobacillus from the EPS experimental group and the FMT experimental group were selected for validation. The experimental mice were treated with antibiotics for one week and divided into four groups, with six mice randomly assigned to each group, and were named as the normal group (ND), the model group (HCD), the M. intestinale group (Muri), and the L. salivarius group (Ligi). During the experiment, the ND group was fed with normal feed for mice, while the HCD, Muri and Ligi group were fed with high-cholesterol feed. At the same time, the Muri group and Ligi group were respectively administered 1 × 10¹⁰ CFU/mL M. intestinale or L. salivarius by gavage, and the ND group and HCD group were administered the same volume of normal saline by gavage for six weeks.

At the conclusion of the animal experiments, on the final day, mice feces were collected under aseptic conditions and preserved in sterile tubes at −80 °C for subsequent use. After anesthesia with sodium pentobarbital, mice blood was collected. Subsequently, the mice were sacrificed by spinal dislocation, and tissues such as the liver, ileum, and colon were harvested and stored at −80 °C for future experimental requirements.

Germ-free mice

The recipient mice in the FMT test were all germ-free mice. Antibiotics mixture (vancomycin: 0.5 mg/ml; gentamicin: 1 mg/ml; ampicillin: 0.5 mg/ml; streptomycin: 1 mg/ml). In order to verify the clear condition of the intestinal microorganisms, the feces of normal mice and mice treated with antibiotics for one week were mixed with normal saline at a concentration of 50 mg/ml, vortexed for 1 min, centrifuged at 500 × g for 3 min, and 800 µl was taken to measure the genomic DNA content according to the kit.

Biochemical assays of serum and tissue

The serum levels of total TC, TG, LDL-C, HDL-C, TBA, AST, and ALT were measured according to the instructions of the kit (Nanjing Jiancheng Bioengineering Institute, Nanjing, China). The liver weight of the collected mice was weighed and the ratio of the liver weight to the body weight of the mice was calculated as the mice liver index. Appropriate amount of liver, ileum or fecal samples were mixed with nine times the volume of absolute ethanol, homogenized mechanically under ice water bath, centrifuged at 2500 × g for 15 min, and the supernatant was collected. The levels of total TC, TG, LDL-C, and HDL-C were measured according to the instructions of the kit.

Histopathological analysis

The collected mice livers were fixed in 4% paraformaldehyde solution for 24 h, followed by dehydration, sectioning, and paraffin embedding for histomorphological analysis. The sections were stained with hematoxylin for 5 min, rinsed, and counterstained with eosin for 15 s. Following the staining process, the sections were mounted using neutral gum. Histological evaluation was conducted using a light microscope.

16S rRNA sequencing analysis

Genomic DNA was extracted from mice fecal samples, and the purity and concentration of the DNA were detected. Using this DNA as the template, the V3–V4 region of the 16S rRNA gene was amplified via polymerase chain reaction (PCR) with barcoded primers.¹² The PCR amplicons were separated by 2% (w/v) agarose gel electrophoresis, and the target bands were excised and purified using a gel extraction kit. The concentration of the purified products was determined and quantified. Subsequently, a sequencing library was constructed from the purified PCR products. Paired-end sequencing (PE300) was performed on an Illumina NextSeq 2000 sequencing system equipped with a P3 flow cell. The sequences obtained were subjected to quality control using the fastp¹³ software and then spliced with FLASH.¹⁴ OTU clustering and chimeric sequences removal were performed using the USEARCH¹⁵ software, followed by sequence normalization for all the samples. The OUTs were taxonomically annotated by comparing with the Silva 16S rRNA gene database using the RDP classifier.¹⁶ The PICRUSt2¹⁷ software was used for 16S functional prediction analysis.

Detection of targeted SCFAs

After centrifugation, 200 μL of supernatant was added with 100 μL of 15% phosphoric acid, 20 μL of internal standard solution (375 μg/mL), and 280 μL of ether solution for 1 min,¹⁸ and then centrifuged and the supernatant was taken for on-line testing. Chromatographic conditions¹⁹: Thermo Trace 1310 gas phase system, Agilent HP-INNOWAX capillary column (30 m × 0.25 mm ID × 0.25 μm); split-flow injection, injection volume of 1 μL, split ratio of 10:1; inlet temperature of 250 °C; the temperature of the ion source of 300 °C; and the temperature of the transmission line of 250 °C. The inlet temperature was 250 °C; the ion source temperature was 300 °C; and the transfer line temperature was 250 °C. The starting temperature of programmed heating is 90 °C, then 10 °C/min to 120 °C, then 5 °C/min to 150 °C, and finally, 25 °C/min to 250 °C for 2 min. The carrier gas is helium with a flow rate of 1.0 mL/min. Mass spectrometry on-board conditions²⁰: Thermo ISQ LT mass spectrometer, electron bombardment ionization (EI) source. SIM scanning mode, electron energy 70 eV. The short-chain fatty acid content of the samples was calculated according to the standard curve of short-chain fatty acids.

Detection of targeted BAs

The bile acid standard was diluted with 30% methanol to prepare ten gradient concentration points. Intestinal contents sample processing: Add 400 μL of methanol, shake, centrifuge and take the supernatant, filtered by 0.2 μm filter membrane and diluted 50 times for LC‒MS injection.²¹ Analytical conditions²²: the column was ACQUITY UPLC^® BEH C18 (2.1 × 100 mm, 1.7 μm) with an injection volume of 5 μL and a column temperature of 40 °C. Mobile phase: A (0.01% formic acid aqueous solution) and B (acetonitrile). Gradient elution: 0–4 min (25% B), 4–9 min (25% → 30% B), 9–14 min (30% → 36% B), 14–18 min (36% → 38% B), 18–24 min (38% → 50% B), 24–32 min (50% → 75% B), 32–33 min (75% → 90% B), and 33–35.5 min (90% → 25% B). min (90% → 25% B); flow rate 0.25 mL/min. The mass spectrometry was performed in ESI negative ion mode with the following parameters: ion source temperature 500 °C, voltage −4500 V, collision gas 6 psi, air curtain gas 30 psi, nebulizing gas and auxiliary gas 50 psi each; scanning mode was multiple reaction monitoring (MRM). The ratio of the peak area of the sample analyte to the peak area of the internal standard was used to calculate the corresponding sample concentration.

Real-time qPCR

An appropriate amount of liver or ileum tissue was weighed and added to RNA extraction solution (G3013, Servicebio) to extract total RNA and reverse transcribed to cDNA (G3337, Servicebio). The mRNA expression levels of HMGCR, CYP7A1, SREBP2, FGFR4, LDLR, SLC10A2, FGF15, NR1H4, NR1H3, NPC1L1, ABCG5, and ABCG8 were determined by RT–qPCR using SYBR Green qPCR Master Mix (G3326, Servicebio) in the CFX Connect system. The primer sequences are shown in Table S1. Using β-actin as an internal reference, the relative expression levels of target genes were calculated using the 2^−(ΔΔCt) method.

Western blot (WB) analysis

Appropriate amounts of liver and ileum tissue were weighed and lysed with RIPA lysates (G2002, Servicebio) containing protease inhibitors (G2008, Servicebio). The extracted proteins were electrophoresed using 8% or 12% acrylamide gels and transferred to PVDF membranes. The PVDF membranes were conjugated to specific antibodies and later to secondary antibodies conjugated to horseradish peroxidase. The following antibodies were used for western blotting: LDLR (A20808, ABclonal), FXR (A24015, ABclonal), CYP7A1 (A10615, ABclonal), FGF15 (PK15481S, Abmart), NPC1L1 (BD-PT8127, Biodragon), ASBT (IPB13168, Baijia), ABCG5 (IPB0314, Baijia), and ABCG8 (DF6673, Affinit). The target protein bands were visualized using an ultrasensitive chemiluminescence kit, and the intensity of the bands was quantified using ImageJ software.

Statistics and analysis

All the data results are expressed as the mean ± standard deviation and were statistically analyzed using GraphPad Prism 8.0. Student's t-test was used for comparisons between two groups. For comparisons of more than two groups, one-way ANOVA analysis was used, one-way ANOVA was used, and Dunnett's test was applied for multiple comparisons (experimental groups vs. HCD group). The 16S rRNA sequencing data were carried out using the Majorbio Cloud platform (https://cloud.majorbio.com) as follows: The mothur²³ software was used to calculate the α-diversity index. The Wilcoxon rank sum test and Student's t-test were used for intergroup difference analysis. Bray‒Curtis distance was applied to assess the similarity between microbial samples. Microbial differential analysis was performed using the Kruskal–Wallis rank sum test and the Wilcoxon rank sum test. LEfSe was employed to identify differentially abundant microbiota across various taxonomic levels (LDA > 2). The microbiota dysbiosis index was evaluated using the Wilcoxon rank sum test. Spearman correlation analysis and distance-based redundancy analysis (db-RDA) were conducted to explore the relationships between physicochemical indicators and microbial communities. Additionally, MaAsLin was used to investigate associations between physicochemical factors and the relative abundance of microbial taxa. When p < 0.05, the data difference between groups was significant.

Results

Extraction, structural elucidation, and hypocholesterolemic evaluation of polysaccharide EPS-D1 from L. plantarum H6

cEPS and rEPS were extracted from L. plantarum H6 by alcohol precipitation, as shown in Figure 1A, and the crude polysaccharides purity was purified to greater than 70%. After intragastric administration of cEPS and rEPS to high-cholesterol diet mice for six weeks (Figure 1B), it was found that rEPS could significantly reduce serum TC, LDL-C, and liver TC in mice, and significantly increase liver HDL-C levels, but cEPS did not show any improvement in host cholesterol (Figure 1C–H), indicating that rEPS made an important contribution to the cholesterol-lowering effect of L. plantarum H6.

Figure 1.

Isolation, purification, structural characterization, and cholesterol-lowering activity of the polysaccharide EPS-D1 from L. plantarum H6. (A) Extraction process of cEPS and rEPS; (B) experimental design for in vivo activity exploration of cEPS and rEPS; (C–H) serum and liver TC, LDL-C, and HDL-C contents in mice; (I) chromatogram of DEAE-cellulose exchange; (J, K) investigation of cholesterol-lowering activity in vitro; (L) gel chromatography purification of EPS-D1; (M) monosaccharide composition of EPS-D1; (N) UV spectra; (O) FT-IR spectra. Note: values are expressed as the mean ± SD (n = 6). *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001 vs. control HCD (same as below).

rEPS were separated by DEAE-cellulose column, as shown in Figure 1I, and two fractions, neutral EPS-D1 and acidic EPS-D2, were isolated. The in vitro cholesterol-lowering activity test of these two components showed that the adsorption rate of EPS-D1 to cholesterol was 36.734%, while the adsorption rate of EPS-D2 to cholesterol was only 7.457% (Figure 1J). Moreover, the solubility of EPS-D1 to cholesterol micelles increased with the increase of concentration (Figure 1K). These results indicated that EPS-D1 was an effective cholesterol-lowering component, so it was gel purified (Figure 1L) and characterized. The Mn and Mw of EPS-D1 were determined to be 15.003 kDa and 17.911 kDa, respectively, by SEC-MALLS-RI, with a narrow molecular weight distribution and good homogeneity (Figure S1A and B). By comparing the retention time of EPS-D1 with that of the mixed monosaccharide standards (Figures 1M and S1C), it was found that the monosaccharide composition of EPS-D1 included galactosamine (1.52%), arabinose (0.65%), glucosamine (3.09%), galactose (14.75%), glucose (33.98%), and mannose (46.10%). Among them, mannose and glucose were the main components, accounting for approximately 80.08% of the total sugar content. After UV scanning, no obvious absorption peaks were found at 260 and 280 nm, indicating that proteins and nucleic acids were not present in EPS-D1 (Figure 1N). In the FT-IR (Figure 1O), the absorption peaks at 3379.81 cm⁻¹ and 2932.46 cm⁻¹ were attributed to the C–H stretching vibration of hydroxyl group and alkyl group, respectively, which were the typical characteristics of sugars. The absorption at 1651.54 cm⁻¹ was attributed to the carboxyl group C=O asymmetric stretching vibration. The broad absorption peak at 1063.13 cm⁻¹ was caused by the stretching vibration between C–OH and C–O–C of the sugar ring, indicating that EPS-D1 had a pyranose ring configuration. The characteristic mannose absorption at 813.93 cm⁻¹ was consistent with the monosaccharide composition results. UV and FT-IR analysis showed that EPS-D1 was a high-purity polysaccharide with typical structural characteristics.

To elucidate the glycosidic bond type and structure of EPS-D1, it was subjected to methylation and NMR analysis. The obtained PMAA was analyzed by GC‒MS to generate total ion chromatograms (Figures S1D and S2). Glycosidic bonds were identified by comparing retention times and specific ion fragments with those in standard databases and are summarized in Table 1. 15 main modifications were found in EPS-D1. The main residue species were mannose (46.05%), glucose (40.76%), galactose (11.62%) and a small amount of arabinose (1.57%). With a branching degree of 29.5%, EPS-D1 is a highly branched polysaccharide. 2,6-Glc(p) (17.14%) and 6-Glc(p) (13.10%) were the most abundant branch points, and t-Man(p) (23.82%) and 2-Man(p) (14.38%) were the two most abundant bond types. It is strongly suggested that Man forms the major part of the branched chain and that galactose Gal and Glc are present in a variety of linked ways, possibly located at the branch point, the backbone, or inside the branched chain, adding to the structural complexity.

Table 1.

Glycosidic linkage type of EPS-D1 by methylation analysis.

Linkage type	PMAAs	Retention time (min)	Ion fragments (m/z)	Relative amount (%)
t-Man(p)	1,5-di-O-acetyl-2,3,4,6-tetra-O-methyl mannitol	13.894	87, 102, 118, 129, 145, 161, 162, 205	23.82
2-Man(p)	1,2,5-tri-O-acetyl-3,4,6-tri-O-methyl mannitol	16.672	88, 101, 129, 130, 161, 190, 205	14.38
3-Man(p)	1,3,5-tri-O-acetyl-2,4,6-tri-O-methyl mannitol	16.241	87, 101, 118, 129, 161, 202, 234	7.84
t-Glc(p)	1,5-di-O-acetyl-2,3,4,6-tetra-O-methyl glucitol	13.948	87, 102, 118, 129, 145, 161, 162, 205	2.47
4-Glc(p)	1,4,5-tri-O-acetyl-2,3,6-tri-O-methyl glucitol	17.288	87, 102, 113, 118, 129, 162, 233	5.06
6-Glc(p)	1,5,6-tri-O-acetyl-2,3,4-tri-O-methyl glucitol	17.555	87, 99, 102, 118, 129, 162, 189, 233	13.10
3,4-Glc(p)	1,3,4,5-tetra-O-acetyl-2,6-di-O-methyl glucitol	18.670	87, 118, 129, 143, 185, 203, 305	2.98
2,6-Glc(p)	1,2,5,6-tetra-O-acetyl-3,4-di-O-methyl glucitol	21.845	87, 88, 99, 100, 129, 130, 189, 190	17.14
t-Gal(p)	1,5-di-O-acetyl-2,3,4,6-tetra-O-methyl galactitol	14.468	87, 102, 118, 129, 145, 161, 162, 205	1.64
3-Gal(p)	1,3,5-tri-O-acetyl-2,4,6-tri-O-methyl alactitol	16.426	87, 101, 118, 129, 161, 202, 234	1.44
6-Gal(p)	1,5,6-tri-O-acetyl-2,3,4-tri-O-methyl galactitol	18.568	87, 99, 102, 118, 129, 162, 189, 233	2.18
4-Gal(p)	1,4,5-tri-O-acetyl-2,3,6-tri-O-methyl galactitol	17.124	87, 102, 113, 118, 129, 162, 233	2.77
3,6-Gal(p)	1,3,5,6-tetra-O-acetyl-2,4-di-O-methyl alactitol	21.086	87, 101, 118, 129, 189, 202, 234	2.25
2,6-Gal(p)	1,2,5,6-tetra-O-acetyl-3,4-di-O-methyl galactitol	21.647	87, 88, 99, 100, 129, 130, 189, 190	1.35
5-Ara(f)	1,4,5-tri-O-acetyl-2,3-di-O-methyl arabinitol	15.146	87, 102, 118, 129, 162, 189	1.57

The structure of EPS-D1 was further deduced by one-dimensional ¹H NMR, ¹³C NMR, DEPT-135, and two-dimensional COSY, HSQC, HMBC, NOESY and TOCSY spectra. In the ¹H NMR spectrum (Figure 2A), the signals were mainly concentrated in the range of δ 3.0–5.5 ppm. Multiple coupled signal peaks were identified in the anomeric signal region of δ 4.3–5.4 ppm, indicating that this sample contained multiple sugar residues. The chemical shifts of the anomeric hydrogens corresponding to these residues were δ 4.51, 4.52, 4.89, 4.96, 5.03, 5.09, 5.13, 5.27, etc. The non-anomeric hydrogen signals were mainly concentrated in the range of δ 3.1–4.2 ppm. Owing to severe overlap of some signals, the chemical shifts of H2–H6 of each sugar residue were assigned respectively based on the COSY and TOCSY spectra (Figure 2C and F). Based on the ¹³C NMR spectrum and HSQC spectrum (Figure 2B and D), the anomer signals of EPS-D1 were δ 5.13/102.08, 5.09/98.15, 5.27/100.55, 4.96/97.7, 5.03/102.17, 4.52/103.04, 4.89/99.35, and 4.51/102.5 ppm, which were denoted as sugar residues A, B, and C, respectively. C, D, E, F, G, and H. Combined with DEPT-135 and ¹³C NMR spectra, methylene (δ 61.08, 65.54, 60.12, 65.49, 60.97, 62.94, 60.56, and 68.86 ppm) signals were obtained. The speculation on residue A is as follows: the anomeric signal δ 5.13/102.08 ppm (H1/C1) indicates that residue A may be an α-configured mannose residue. In the COSY spectrum, the cross-peak δ 5.13/4.06 ppm identified H2 (4.06 ppm) of residue A, the cross-peak δ 4.06/3.83 ppm identified H3 (3.83 ppm) of residue A, the cross-peak δ 3.83/3.62 ppm identified H4 (3.62 ppm) of residue A, the cross-peak δ 3.62/3.74 ppm identified H5 (3.74 ppm) of residue A, and the cross-peak δ 3.74/3.74, 3.88 ppm identified H6 (3.74, 3.88 ppm) of residue A, thus completing the assignment of the chemical shifts of the hydrogen atoms on the sugar ring. Then, the chemical shifts of the carbon atoms on the sugar ring were assigned through HSQC signals. The chemical shift of C1 of residue A is δ 102.08 ppm, that of C2 is δ 70.02 ppm, that of C3 is δ 70.61 ppm, that of C4 is δ 66.83 ppm, that of C5 is δ 73.31 ppm, and that of C6 is δ 61.08 ppm. The chemical shift of C1 shifts to a lower field, indicating that the residue is substituted at the O-1 position of the sugar ring. Combined with the methylation analysis results and literature reports, it is inferred that the sugar residue A may be α-D-Manp-(1→²⁴ By analogy, sugar residue B is →2,6)-α-D-Glcp-(1→, sugar residue C is →2)-α-D-Manp-(1→²⁵ sugar residue D is →6)-α-D-Glcp-(1→²⁶ sugar residue E is →3)-α-D-Manp-(1→²⁵ sugar residue F is →4)-β-D-Glcp-(1→²⁷ sugar residue G is →4)-α-D-Galp-(1→²⁸ and sugar residue H is →3,6)-β-D-Galp-(1→²⁹). The assignments of the ¹H and ¹³C chemical shifts of all sugar residues are shown in Table 2.

Figure 2.

Structural characterization of EPS-D1. (A) ¹H NMR spectrum of EPS-D1; (B) ¹³C NMR spectrum of EPS-D1; (C) ¹H single bond ¹H COSY spectrum of EPS-D1; (D) HSQC spectrum of EPS-D1; (E) NOSEY spectrum of EPS-D1; (F) TOCSY spectrum of EPS-D1; (G, H) the deduced structure of EPS-D1; (I) SEM images.

Table 2.

Attribution of ¹H and ¹³C chemical shifts of each saccharide residue in EPS-D1.

Code	Glycosyl residues	Chemical shifts (ppm)
		H1/C1	H2/C2	H3/C3		H4/C4	H5/C5	H6/C6
A	α-D-Manp-(1→	5.13	4.06		3.83	3.62	3.74	3.74, 3.88
		102.08	70.02		70.61	66.83	73.31	61.08
B	→2,6)-α-D-Glcp-(1→	5.09	4.02		3.9	3.5	3.68	3.97, 3.75
		98.15	78.67		70.35	69.53	73.12	65.54
C	→2)-α-D-Manp-(1→	5.27	4.1		3.88	3.71	3.56	3.71
		100.55	78.48		70.31	66.91	71.43	60.12
D	→6)-α-D-Glcp-(1→	4.96	3.55		3.71	3.77	3.92	3.91, 3.73
		97.7	71.48		73.24	71.21	73.44	65.49
E	→3)-α-D-Manp-(1→	5.03	4.2		3.93	3.78	3.65	3.84
		102.17	69.56		77.91	66.47	73.36	60.97
F	→4)-β-D-Glcp-(1→	4.52	3.54		3.69	3.74	3.62	3.68
		103.04	71.23		71.19	n.d	74.92	62.94
G	→4)-α-D-Galp-(1→	4.89	3.97		3.72	3.56	3.67	3.82
		99.35	70.11		73.23	76.41	73.09	60.56
H	→3,6)-β-D-Galp-(1→	4.51	3.32		3.47	3.87	4.04	3.84
		102.5	73.16		76.01	70.57	70.34	68.86

The above results, combined with the analysis of HMBC and NOESY spectra (Figures 2E and S2F), revealed the structure and linkage of the polysaccharide. There was no signal in the HMBC spectrum. According to the NOESY spectrum, there were cross-peaks between H1 of sugar residue A and H6 of sugar residue B at δ 5.13/3.75 ppm and δ 5.13/3.97 ppm, between H1 of sugar residue A and H4 of sugar residue F at δ 5.13/3.74 ppm, between H1 of sugar residue A and H3 of sugar residue H at δ 5.13/3.47 ppm, and between H1 of sugar residue A and H6 of sugar residue H at δ 5.13/3.84 ppm. There were also cross-peaks between H1 of sugar residue B and H2 of sugar residue B at δ 5.09/4.02 ppm, between H1 of sugar residue B and H2 of sugar residue C at δ 5.09/4.1 ppm, between H1 of sugar residue B and H6 of sugar residue D at δ 5.09/3.73 ppm and δ 5.09/3.91 ppm, between H1 of sugar residue C and H2 of sugar residue B at δ 5.27/4.02 ppm, between H1 of sugar residue C and H2 of sugar residue C at δ 5.27/4.1 ppm, between H1 of sugar residue D and H6 of sugar residue D at δ 4.96/3.73 ppm, between H1 of sugar residue D and H3 of sugar residue E at δ 4.96/3.93 ppm, between H1 of sugar residue E and H2 of sugar residue B at δ 5.03/4.02 ppm, between H1 of sugar residue F and H6 of sugar residue B at δ 4.52/3.75 ppm, between H1 of sugar residue G and H6 of sugar residue B at δ 4.89/3.75 ppm and δ 4.89/3.97 ppm, and between H1 of sugar residue H and H4 of sugar residue G at δ 4.51/3.56 ppm. Based on the one-dimensional and two-dimensional NMR information and methylation results, it was inferred that EPS-D1 mainly consists of the main chain formed by the interconnection of →2,6)-α-D-Glcp-(1→, →2)-α-D-Manp-(1→, →6)-α-D-Glcp-(1→ and →3)-α-D-Manp-(1→, and the side chains are mainly composed of α-D-Manp-(1→ connected at the O-6 position of the sugar residue →2,6)-α-D-Glcp-(1→, etc and its possible structure is shown in the (Figure 2G–H). Additionally, through SEM observation of the surface morphology of EPS-D1 at different magnifications, it was found that EPS-D1 had a rough surface and an irregular-sized pore structure similar to that of a sponge (Figure 2I). In conclusion, we isolated and purified a novel small-molecule polysaccharide EPS-D1 from L. plantarum H6 through in vivo and in vitro cholesterol-lowering methods. It may have good cholesterol metabolism activity, and its specific role awaits further research.

EPS-D1 can improve the lipid level of mice fed with high cholesterol diet, regulate the composition of intestinal microbiota, and enrich beneficial intestinal symbiotic bacteria in mice

Next, we investigated the regulatory effects of L. plantarum H6 EPS-D1 on cholesterol metabolism in the host, the experimental design is shown in Figure 3A. There were no significant changes in body weight and weight gain between the experimental and HCD group during the study period (Figure 3B and C). Compared with the HCD group, the serum total TC, TG, and LDL-C of EPS-D1 mice were decreased by 40.31%, 37.55%, and 43.15%, respectively (Figure 3D–F), but there was no significant effect on the serum HDL-C (Figure 3G). Serum AST, ALT and TBA are important indicators of liver injury. Compared with that of the HCD group, the serum TBA content decreased by 55.33% (Figure 3H). AST and ALT decreased to 16.87 U/L and 31.84 U/L, respectively (Figure 3I and J). In addition, as an important metabolic organ of cholesterol, the liver index of the mice was also significantly decreased after the ingestion of EPS-D1 (Figure 3K). In HE sections of liver tissue, it was discovered that EPS-D1 intake could improve the liver steatosis (lipid droplet vacuoles), cell diffusion, and inflammatory cell infiltration brought on by the high-cholesterol diet. The inhibition of liver enlargement was also seen in visual images of the liver (Figure 3L). While EPS-D1 intake had no discernible effect on liver TG and LDL-C, it did dramatically raise HDL-C levels and decrease hepatic TC in EPS-D1 mice, as seen in Figure 3M‒P. These findings showed that in mice given a high-cholesterol diet, EPS-D1 might lower liver damage and raise serum and liver lipid levels.

Figure 3.

Improvement of lipid levels in mice fed a high-cholesterol diet by EPS-D1. (A) Experimental design of the mechanism underlying the reduction of cholesterol levels in EPS-D1; (B, C) body weight and weight gain of the mice; (D–J) serum TC, TG, LDL-C, HDL-C, TBA, AST, and ALT contents in the mice; (K) liver index; (L) visualization of the liver and HE staining; (M–P) liver TC, TG, LDL-C, and HDL-C contents in the mice. Note: values are expressed as the mean ± SD (n = 8).

The intestine (especially the ileum) and intestinal microbiota together form a precisely regulated system for cholesterol metabolism, playing a key regulatory role in the host's cholesterol metabolic balance. As shown in Figure 4A and B, after six weeks of EPS-D1 intake, the cholesterol levels in the ileum and feces of the mice were significantly decreased compared with the HCD group, indicating that the cholesterol metabolism in the host mice had reached a balanced state. Polysaccharides serve as an important carbon source for the intestinal microbiota. To explore the current effect of EPS-D1 on the intestinal microbiota of mice, 16S rRNA sequencing analysis was performed on the intestinal microbiota. Diversity PCA analysis showed that, at the OTU level, there was a significant difference in the overall community structure between the high-cholesterol diet group and the EPS-D1-treated mice (Figure 4C). In addition, the microbiota typing analysis showed that in the dominant microflora g__norank_f__Muribaculacea, the mice with EPS-D1 intake had a similar microflora typing as the mice with normal diet (Figure 4D). Moreover, the microbiota dysbiosis index (MDI) indicated that the MDI of the mice significantly decreased after the intake of EPS-D1 (Figure 4E). Further, the UpSet plot made using the OTU abundance table revealed that the number of OTU in the mice that consumed EPS-D1 increased by 5.25% and 45.17% compared to the ND and HCD group, respectively (Figure 4F).

Figure 4.

EPS-D1 improved gut microbiota disorder in mice fed a high-cholesterol diet. (A, B) Ileum and fecal TC contents in mice; (C) PCoA analysis; (D) typing analysis of bacterial microflora; (E) typing analysis of bacterial microflora; (F) upset analysis; (G, H) phylum level and genus level bar graphs of the microflora; (I) heatmap of community composition at the genus level; (J) multiple group difference test; (K) the difference between the two groups was tested; (L) LEfSe multi-stage discriminant analysis of species differences. Note: values are expressed as the mean ± SD (n = 6).

Specifically, Bacteroidota, Bacillota, Verrucomicrobiota, and Actinomycetota were the most dominant microorganisms at phylum level in all the groups (Figure 4G). At the genus level, norank_f__ Muribaculaceae, Lactobacillus, Akkermansia, Lachnospiraceae_NK4A136_group, Bacteroides, and Ligiactobacillus were the most relative content among all the components from the heatmap, it was found that the three experimental groups were significantly separated, and the EPS-D1 and ND groups were similarly clustered, indicating changes in the structure of the mice microbiota (Figure 4H and I). Further analysis of multi-group difference test showed that Bacteroides, norank_o__Clostridia_UCG-014 and Muribaculum were the most different microorganisms in the three groups (Figure 4J). In the two-group difference test, it was found that compared with the HCD group, the norank_o__Clostridia_UCG-014, Muribaculum, Fecalibaculum, Turicibacter, and Lactobacillus in the intestinal microflora of EPS-D1 mice administered by gavage were significantly increased, which were the significant difference microorganisms between the two groups (Figure 4K). LEfSe was used to analyze the differences in the gut microbiota between the different groups (Figure 4L). Compared with HCD, Lactobacillus, Turicibacter, norank_o__Clostridia_UCG-014, Fecalibaculum and Ligiactobacillus were the dominant bacterial strains in the intestinal microflora of EPS-D1 mice by gavage. Using the PICRUSt2 tool and combining with MetaCyc pathway analysis, it was found that most metabolic pathways in the EPS-D1 group were upregulated, indicating that the regulated intestinal microbiota could promote host metabolism. The pathways including glycolytic pathway (ANAGLYCOLTSIS-PWY, PWY-5484, and GLYCOLYSIS-PWY), pyruvate fermentation to acetate and lactate II (PWY-5100), and the pentose phosphate pathway (NONOXIPENT-PWY), etc. were all upregulated. These results suggest the functional potential of the intestinal microbiota is affected by EPS-D1, which in turn affects the lipid metabolism of the host (Figure S3L). These results suggested that EPS-D1 intake could ameliorate gut microbiota dysbiosis and improve the composition and abundance of gut microbiota in mice fed a high-cholesterol diet.

EPS-D1-derived gut microbes were able to improve lipid levels in mice fed a high-cholesterol diet

As a favorite substance of microorganisms, polysaccharides play an important role in regulating the intestinal microflora. In order to further explore the role of EPS-D1-regulated intestinal microflora in cholesterol metabolism, FMT was used to further explore (Figure 5A). After five weeks of FMT from antibiotic-treated mice, there was no significant difference in body weight between FMT-EPS and FMT-HCD mice (Figure 5B), but the serum and liver TC and LDL-C levels of FMT-EPS mice were significantly lower than those of FMT-HCD mice (Figure 5C,E,H,J). HDL-C levels were significantly increased (Figure 5F,K), but neither serum nor hepatic TG levels were significantly affected (Figure 5D,I). At the same time, FMT also significantly ameliorated liver damage and reduced the liver index in the mice (Figure 5L,G), which was consistent with the improvement results of EPS-D1 intake mice, indicating the important role of the intestinal microflora in the regulation of host cholesterol metabolism by EPS-D1.

Figure 5.

Fecal microbiota transplantation (FMT) using feces from mice fed with EPS-D1 improved lipid levels and gut microbiota disorders in mice fed a high-cholesterol diet. (A) Experimental design of FMT on EPS-D1; (B) changes in body weight in mice; (C–F) serum TC, TG, LDL-C, and HDL-C contents in mice; (G) liver index; (F–K) liver TC, TG, LDL-C, and HDL-C contents in mice; (L) visualization of the liver and HE staining; (M, N) TC content in the ileum and feces of mice; (O) MDI analysis; (P) upset analysis; (Q) typing analysis of bacterial microflora; (R) the difference between the two groups was tested; (S) LEfSe multi-stage discriminant analysis of species differences. Note: values are expressed as the mean ± SD (n = 6).

Similarly, cholesterol levels in the ileum and feces were significantly decreased in FMT-EPS mice (Figure 5M and N). Gut microbiota sequencing results showed that compared with FMT-HCD mice, the MDI of FMT-EPS mice was significantly reduced (Figure 5O), and the number ratio of ASV/OUT was increased by 39.10% (Figure 5P). In addition, FMT-HCD and FMT-EPS were obviously separated in the dominant bacteria Ligiactobacillus, g__norank_f__ Muribaculacea and Akkermansia (Figure 5Q). Finally, in the difference test between the two groups, it was found that norank_o__Clostridia_UCG-014, Odoribacter, Ligiactobacillus, Lachnospiraceae_NK4A136_group, and Muribaculum increased significantly in HCD-EPS. In LEfSe, Ligiactobacillus, Lachnospiraceae_NK4A136_group, norank_o__Clostridia_UCG-014, Acetatifactor, and g__unclassified_f__Muribaculacea was the dominant strain in HCD-EPS. In addition, functional analysis and prediction indicated that the FMT-EPS group not only increased pathways such as the glycolytic pathway (ANAGLYCOLTSIS-PWY, PWY-5484 and GLYCOLYSIS-PWY), pyruvate fermentation to acetate and lactate II (PWY-5100), and the pentose phosphate pathway (NONOXIPENT-PWY) but also increased the gluconeogenesis I (GLUCONEO-PWY) pathway (Figure S4Q). In conclusion, using the FMT assay, we confirmed the role of EPS-D1-derived gut microbes in improving host cholesterol levels.

EPS-D1-enriched intestinal microbiota-dominant Muribaculum spp. improved lipid levels in mice fed a high-cholesterol diet

In order to screen the key strains related to cholesterol metabolism, the intestinal dominant strains were found and screened by correlation analysis to clarify the mechanism of EPS-D1 in improving cholesterol metabolism by regulating the intestinal microflora. The Spearman algorithm was used to explore the correlation between the top 50 species in total abundance at the taxonomic level and the physicochemical parameters in the serum and liver of the mice. It was found that when EPS-D1 was administered by gavage, Lactobacillus, Ligiactobacillus, Limcsilactobacillus, Muribaculum, Odoribacter, Blautia and serum TC, ileum TC, liver LDL-C, liver index, fecal TC, serum AST, TBA and ALT were significantly negatively correlated, and Muribaculum had the strongest correlation (Figure 6A). As shown in Figure 6B, Lactobacillus and norank_f__ Muribaculacea were negatively correlated with serum TC, TG, ALT, AST, TBA, liver TC, ileum TC, and fecal TC. Lactobacillus and norank_f__ Muribaculacea were the key species affecting ND and EPS. In the FMT experiment, the correlation heatmap showed that, Ligiactobacillus, Lachnospiraceae_NK4A136_group, norank_o__Clostridia_UCG-014, Muribaculum, Odoribacter and Alistipes were negatively correlated with serum and liver LDL-C, liver TC, ileum TC, fecal TC, serum TC, and liver index (Figure 6C). In RDA/CCA analysis, Ligiactobacillus and norank_f__ Muribaculacea were negatively correlated with serum TC, LDL-C, liver TC, LDL-C, and fecal TC. Ligiactobacillus and norank_f__ Muribaculacea were the key species related to FMT-HCD group (Figure 6D). According to the above correlation analysis results, combined with the analysis of the dominant microflora of EPS-D1 by gavage and FMT, it was found that the dominant intestinal microbial strains enriched in EPS-D1 may be Muribaculum and Ligiactobacillus. Therefore, the correlation of Muribaculum and Ligiactobacillus with relevant physicochemical indices was further determined by linear model MaAsLin analysis. In the experiment of EPS-D1 gavage, Muribaculum showed a significant negative correlation with TC in the liver, feces and serum, and Ligiactobacillus showed a significant negative correlation with TC in liver and feces (Figure 6E). In the FMT experiment, Muribaculum was significantly negatively correlated with ileum and fecal TC, and Ligiactobacillus was significantly negatively correlated with serum, ileum, fecal TC and serum LDL-C (Figure 6F). These results indicated that Muribaculum and Ligiactobacillus were the dominant strains of EPS-D1-enriched gut microbes.

Figure 6.

Correlation between gut microbiota and serum and liver physicochemical parameters in mice. (A) Heatmap analysis of the correlation between the dominant intestinal strains and physical and chemical parameters in EPS-D1-treated mice, the relevance threshold of the significance marker/R/ ≥ 0.1; (B) RDA/CCA analysis among samples, physicochemical indices and microorganisms; (C) heatmap analysis of the correlation between dominant intestinal strains and physicochemical parameters in FMT mice; (D) RDA/CCA analysis among FMT samples, physicochemical indices and microorganisms; (E, F) MaAsLin analysis between physical and chemical indices and the relative abundance of microbial community species. Note: values are expressed as the mean ± SD (n = 6).

In order to explore whether the dominant strains Muribaculum and Ligiactobacillus also have the effect of improving cholesterol in mice fed a high-cholesterol diet, the two dominant strains were respectively gavaged to mice fed a high-cholesterol diet. As Ligiactobacillus has been previously reported to have the effect of reducing cholesterol.^30,31 Although few Muribaculum species have been isolated and cultured, M. intestinale remains the most characterized member of this genus. Here, we report for the first time that M. intestinale plays a direct role in modulating cholesterol metabolism. To explore whether M. intestinal is able to directly utilize ESP-D1, we added ESP-D1 to the medium of M. intestinal. The results showed that compared with the control group, the macroscopic colony formation of M. intestinal was increased in the plate supplemented with EPS-D1, indicating that ESP-D1 could promote the growth of M. intestinal (Figure S5A).In vivo, after six weeks of gavage, there was no significant change in food intake among the groups (Figure 7A and B). Compared with the HCD, the mice receiving M. intestinale showed a significant decrease in body weight after the third week. The same trend was observed in the control group treated with gavage of L. salivarius (Figure 7C). Compared with that of the HCD group, the intake of M. intestinale significantly reduced the levels of serum TC, TG, and LDL-C in mice (Figure 7D–F) but had no significant effect on serum HDL-C, though it did show an upward trend (Figure 7G). In addition, in the sections of liver tissue, it was found that Muri could improve hepatic steatosis (lipid droplet vacuoles), hepatocyte diffusion and inflammatory cell infiltration induced by high-cholesterol diet (Figure 7H), and improve liver enlargement and reduce liver index (Figure 7I). As shown in Figure 7J–M, after gavage of M. intestinale, the TC level in the liver of mice was significantly decreased and the HDL-C level was significantly increased, but there was no significant effect on the TG and LDL-C levels in the liver. Similarly, cholesterol levels were reduced by 74.79% and 53.16% in the ileum and feces of Muri-fed mice compared with HCD (Figure 7N‒O). In conclusion, EPS-D1 was enriched in the dominant genus Muribaculum by correlation analysis, and it was also verified for the first time to improve cholesterol in mice fed with high-cholesterol diet.

Figure 7.

M. intestinale improved lipid levels in mice fed a high-cholesterol diet. (A) Experimental design of M. intestinale to improve cholesterol levels; (B) changes in food intake and body weight in mice; (D–G) serum levels of TC, TG, LDL-C, and HDL-C in mice; (H) visualization of the liver and HE staining; (I) liver index; (J–M) liver levels of TC, TG, LDL-C, and HDL-C in mice; (N, O) TC content in the ileum and feces of mice. Note: values are expressed as the mean ± SD (n = 6).

EPS-D1 and M. intestinale can regulate cholesterol metabolism in mice via enterohepatic FXR‒FGF15 axis

RT-qPCR and WB were used to analyze the genes and proteins related to cholesterol metabolism to explore the mechanism of EPS-D1 improving high cholesterol diet mice. As shown in Figure 8A, HCD resulted in abnormal expression of hepatic cholesterol-related synthetase and related invertase as well as receptor genes, but CYP7A1, FGFR4 and LDLR gene expression was significantly upregulated after EPS-D1 intake, but HMGCR and SREBP2 gene expression was not significantly improved. In the ileum, the relative mRNA expression levels of genes involved in cholesterol absorption and bile acid efflux, including NPC1L1, NR1H4, FGF15, SLC10A2, and NR1H3, were significantly down-regulated after EPS-D1 administration compared with those in the HCD group (Figure 8B,E–H). In addition, the relative expression levels of ABCG5 and ABCG8 mRNA in the HCD group were significantly decreased, and their expression levels were up-regulated after ingestion of EPS-D1 (Figure 8C and D).

Figure 8.

EPS-D1 improved genes and proteins in signaling pathways related to cholesterol metabolism in high-cholesterol diet mice. (A) RT-qPCR analysis of liver genes involved in cholesterol metabolism; (B‒H) RT-qPCR analysis of ileum genes involved in cholesterol metabolism; (I‒K) western blot analysis of CYP7A1, LDLR, ABCG5, ABCG8, FXR, and FGF15 in the liver and a relative protein content expression heatmap; (L-M) western blot analysis of FXR, FGF15, NPC1L1, ASBT, ABCG5, and ABCG8 in the ileum and a relative protein content expression heatmap. Note: values are expressed as the mean ± SD (n = 5 or 3).

Further investigation was conducted on the protein expression in the liver and ileum. As shown in Figure 8I, the expression of CYP7A1 and LDLR in the liver of the HCD group was lower than that of the ND group (but not significantly), while the intake of EPS-D1 significantly upregulated the expression of these proteins. FXR, as a bile acid sensor, regulates bile acid metabolism and induces the secretion of FGF15. As shown in Figure 8K, the expression of FXR in the liver was significantly decreased by HCD group, which was significantly reversed by the intake of EPS-D1. However, the expression of FGF15 in the liver was significantly increased after the intake of a HCD and significantly decreased after the intake of EPS-D1. In addition, in the ileum, HCD resulted in a significant increase in the expression of FXR/FGF15, but its expression was significantly downregulated after EPS-D1 ingestion, regulating BAs metabolism through negative feedback and promoting cholesterol efflux. ABCG5/8 is a transport protein on the membrane of intestinal epithelial cells and liver cells, which can mediate the absorption of cholesterol. As found in Figure 8J, EPS-D1 intake can reverse the significant reduction of ABCG5 expression caused by a high-cholesterol diet but has no effect on ABCG8, and it is possible that these two proteins have a compensatory effect in liver cells. It is worth noting that the expression of ABCG5/8 in the ileum is completely opposite to that in the liver. The intake of EPS-D1 significantly reduced the expression of both (Figure 8M), which might be a feedback regulatory mechanism related to the decrease in TC content in the ileum. Additionally, NPC1L1 and ASBT were found to be significantly downregulated in ileum tissue after EPS-D1 administration, thereby reducing cholesterol absorption and decreasing the enterohepatic circulation, ultimately lowering the cholesterol level in mice (Figure 8N).

Next, we explored whether the dominant strain M. intestinale also regulates cholesterol metabolism through enterohepatic FXR-FGF15 signaling. As shown in Figure 9A–E, the Muri group could significantly increase the decreased expression of CYP7A1 and LDLR induced by HCD, and significantly up-regulate the protein expression of FXR and FGF15 in the liver. In addition, the protein expression levels of FXR, FGF15, NPC1L1 and ASBT in the ileum were also decreased after Muri intake (Figure 9F‒J). Based on these results, EPS-D1 and its dominant intestinal microbial strain M. intestinale could reduce enterohepatic circulation and increase the excretion of BAs by activating the enterohepatic FXR-FGF15 signaling pathway, upregulate hepatic CYP7A1 to promote cholesterol conversion and inhibit ileal NPC1L1 to reduce cholesterol absorption in BAs to regulate host cholesterol metabolism in mice fed a high-cholesterol diet.

Figure 9.

M. intestinale improved signaling pathways related to cholesterol metabolism in mice fed a high-cholesterol diet. (A) Western blot analysis of CYP7A1, LDLR, FXR, and FGF15 in the liver; (B–E) relative protein content expression heatmap in the liver; (F) western blot analysis of FXR, FGF15, NPC1L1, and ASBT in the ileum; (G–J) relative protein content expression heatmap in the liver. Note: values are expressed as the mean ± SD (n = 3).

M. intestinale can improve cholesterol level by regulating intestinal SCFAs and BAs in mice

SCFAs are the main metabolites of the intestinal microbiota and play a significant role in cholesterol metabolism. The effects of the dominant strain M. intestinale on intestinal SCFAs in mice fed a high-cholesterol diet are shown in Figure 10A–E. PCA analysis revealed that the Muri and ND groups were clustered together (Figure 10A), and the heatmap analysis also indicated a clear separation between the Muri and HCD groups (Figure 10B). Further analysis found that a high-cholesterol diet led to a significant decrease in the total SCFAs content in mice intestine. After Muri administration, the total SCFAs content increased by 37.88% (Figure 10C). Specifically, the contents of acetic acid and caproic acid significantly increased after Muri administration, while the contents of butyric acid, isobutyric acid, valeric acid, and isovaleric acid showed an upward trend but were not significant (Figure 10D). The species difference volcano plot and heatmap also indicated that acetic acid and caproic acid were the SCFAs significantly affected by Muri intake (Figure 10E and F).

Figure 10.

M. intestinale regulated mice intestinal SCFAs and BAs levels. (A, B) PCA analysis and heatmap analysis of SCFAs; (C) total SCFAs content; (D) acetic acid, propionic acid, butyric acid valerate caproic acid, isobutyric acid, and isovalerate acid content; (E, F) volcano map and heat map of differential metabolites; (G, H) PCA analysis and heatmap analysis of BAs; (I) Venn analysis; (J, K) content of total, primary and secondary BAs; (M) differential BAs heatmap; (N) difference BAs volcano plot between muri and HCD groups; (O) boxplot of differential BAs. Note: values are expressed as the mean ± SD (n = 6).

The conversion of cholesterol to BAs is an important pathway for reducing cholesterol. It has been confirmed by WB experiments that M. intestinale can negatively regulate BAs metabolism through the FXR-FGF15 axis and promote cholesterol excretion (Figure 9). Therefore, BAs-targeted metabolomics was further used to observe the effect of M. intestinale on the intestinal BA pool in mice fed a high-cholesterol diet. The PCA results showed a clear separation of the BA composition between the Muri group and the HCD group, indicating changes in the BA levels in the ileum of Muri mice. Additionally, heatmap analysis revealed that the BA composition of the Muri group was clustered together with the ND group (Figure 10G and H). Moreover, Venn analysis indicated that there were six differential BAs between the Muri vs HCD groups (Figure 10I). Specifically, compared with ND, the total BAs content in mice fed a high-cholesterol diet was significantly increased. After Muri intake, the total BAs content in the mice decreased significantly, especially the content of secondary BAs, but the total primary BAs did not change significantly (Figure 10J–L). The species difference heatmap indicated that there were 10 different BAs among the three groups, and the Muri group and ND group were clustered together and clearly separated from the HCD group (Figure 10M). The species difference volcano plot between the Muri and HCD groups showed that compared with HCD, the contents of TCA-3S, T-omega-MCA, 7-ketoLCA, and HCA significantly increased, while the content of TDCA significantly decreased (Figure 10N). Additionally, compared with HCD, after Muri intake, the contents of UDCA, a-MCA, and HCA in primary BAs in the ileum of mice significantly increased, while the contents of THCA and TDCA in secondary BAs significantly decreased (Figure 10O). These results suggested that the dominant strain M. intestinale could regulate the metabolic structure of SCFAs and BAs pools in the intestinal tract of mice, thereby maintaining the homeostasis of cholesterol level in mice.

Discussion

This study revealed for the first time a novel mechanism by which EPS-D1, a small exopolysaccharide from L. plantarum H6, improved cholesterol metabolism by remodeling the gut microbiota, promoting the growth of the Muribaculum and regulating the FXR‒FGF15 pathway along the enterohepatic axis. The monosaccharide composition of EPS-D1 included mannose, glucose, galactose, glucosamine, galactosamine and arabinose. In addition, the infrared spectrum analysis showed that the characteristic absorption peak of mannose is at 813.93 cm⁻¹, which is consistent with the results of the monosaccharide composition. Structural analysis revealed key features of EPS-D1, which was shown to be a small-molecule, highly branched, short-chain exopolysaccharide. In addition, low-molecular-weight polysaccharides can be efficiently and rapidly fermented by the intestinal microbiota to produce a large number of beneficial metabolites, such as SCFAs, which can exert their functions better.³² Studies have shown that the repeat units of probiotics derived exopolysaccharides are mostly glucose and galactose, and the mannose content is generally less.³³ However, the monosaccharide composition of EPS-D1 shows that the mannose component is the most abundant, followed by glucose. This difference in polysaccharide composition may be one of the reasons for the cholesterol-lowering activity of EPS-D1. In a new study, it was shown that the composition of exopolysaccharide extracted from Schleiferilactobacillus harbinensis Z171 was also high in mannose content and had better biological activity of lowering cholesterol,³⁴ which was better verified by this study. Methylation and NMR analysis showed that the main chain structure of EPS-D1 is →2,6)-α-D-Glcp-(1→, which makes its structure highly branched, provides a huge acting surface area and specific binding site, and may play an active role in reducing cholesterol through adsorption.³³ Moreover, EPS-D1 also has diversified terminal branches dominated by α-D-Manp-(1→, creating a complex polysaccharide framework with potential bioactive properties.³⁵ By SEM, EPS-D1 was found to have a rough surface and a sponge-like pore structure of irregular size, suggesting that this type of morphology may be beneficial for various biological activities, including its role in maintaining the integrity of the intestinal barrier and interacting with the gut microbiota.³⁶

In recent years, multiple studies have indicated that gut microbiota may serve as a potential therapeutic target for lowering cholesterol levels.³⁷ For instance, a human cohort study discovered that the gut bacterium Oscillibacter can metabolize cholesterol within the intestines, thereby contributing to reduced cholesterol levels and lowered cardiovascular disease risk.³⁸ In this study, analysis of the gut microbiota 16S rRNA in EPS-D1-gavage mice and their FMT mice revealed that EPS-D1 intake modulated the structural characteristics of gut microbiota. Differential analysis and correlation analysis identified dominant strains, including Muribaculum, Ligiactobacillus, Lactobacillus, and norank_o__Clostridia_UCG-014. Consistent with our previous findings, Muribaculum was the only genus that was significantly enriched in all four intervention groups (viable, inactivated, lysed, and FMT) of L. plantarum H6, with a 2.3- to 5.8-fold increase in relative abundance compared with the high-cholesterol model group. The increase was significantly higher than that of Lactobacillus and Bifidobacterium.¹⁰ This indicated that Muribaculum may serve as a common target for regulating host cholesterol metabolism across different intervention forms of L. plantarum H6.

The Muribaculaceae family primarily utilizes monosaccharides derived from mucins, which are capable of utilizing mucin O-glycosidic monosaccharides (such as sialic acid, N-acetylglucosamine, fucose, galactose, and N-acetylgalactosamine). This enables them to compete for nutrients with pathogens dependent on these mucosal sugar sources (such as Clostridium difficile), making them ecological protectors of the healthy gut.³⁹ In this study, the monosaccharide composition of EPS-D1 included galactosamine, glucosamine and galactose, which might be the reason for the enrichment of Muribaculum in mice after ingestion. However, whether the monosaccharide composition of EPS-D1 was directly related to the enrichment of Muribaculum needs further exploration. The present study further demonstrated that enriched Muribaculum monocultures recapitulated the cholesterol-lowering effects of EPS-D1 in animal experiments, providing the first direct demonstration of the regulatory effect of this strain on cholesterol metabolism in mice fed a HCD. Additionally, in this study, another dominant gut strain enriched by EPS-D1, Ligiactobacillus, was used as the control group for comparison with the Muri experimental group. Future research could investigate whether Muribaculum and Ligiactobacillus exhibit symbiotic effects and whether their combined use yields superior cholesterol-regulating effects compared to either strain alone.

In this study, we demonstrated for the first time the central targets of EPS-D1 and its enriched dominant intestinal microbiota species, Muribaculum spp. (M. intestinale), in improving cholesterol metabolism in mice fed a HCD, mainly by regulating the FXR-FGF15 signaling pathway of the enterohepatic axis. In the liver, HCD leads to a significant decrease in the expression of the cholesterol synthase SREBP2 gene and abnormal expression of HMGCR gene,⁴⁰ and EPS-D1 intake did not significantly improve it, which may be due to excessive intake of cholesterol, leading to the impairment of the cholesterol synthesis pathway.⁴¹ Additionally, HCD led to increased expression of the NR1H4 and FGF15 genes, whereas EPS-D1 intake significantly decreased NR1H4 expression. Consistent with our findings, prior studies indicated that tomato pectin negatively regulated intestinal and hepatic FXR signaling pathways through negative feedback mechanisms. Specifically, it promoted BAs synthesis by reducing ileal FXR and FGF15 levels while increasing gene expression associated with both the classical (CYP7A1) and alternative (CYP27A1) pathways.⁴²

FXR is expressed primarily in the liver and gut, acting as a bile acid sensor to regulate bile acid synthesis, metabolism, and transport.⁴³ When bile acid levels increase, FXR is activated, which inhibits CYP7A1 expression in the liver to reduce bile acid synthesis and induces FGF15 secretion in the intestine to inhibit hepatic CYP7A1 expression.⁴⁴ However, previous studies have shown that gut microbes can regulate the expression of FGF15 in the ileum and CYP7A1 in the liver through a FXR dependent mechanism,⁴⁵ and that increased excretion of BAs is accompanied by enhanced synthesis of BAs in the liver.⁴⁶ In the present study, the total FXR protein expression in the ileum of mice was decreased after M. intestinale ingestion. Although we did not directly localize FXR to the nuclear region, we found that the mRNA level of its coding gene NR1H4 was also decreased, suggesting that the reduction of FXR occurred at multiple levels of transcription and translation and its signaling pathway was also affected. In addition, we observed a significant reduction in the expression of FGF15, a core downstream target whose transcription is directly activated by FXR in the nucleus.⁴⁷ It is noteworthy that in this study, ABCG5/G8 mRNA levels were up-regulated and protein levels were down-regulated in the ileum, which may be due to the dominance of NPC1L1 absorption inhibition, while the decrease of ABCG5/8 in the liver and ileum may be compensated by other biliary excretion pathways (such as CYP7A1 up-regulation).⁴⁸

The important role of SCFAs in cholesterol metabolism has been well demonstrated by numerous experiments,^49,50 and we also found that M. intestinale is able to increase total SCFAs, acetate, and caproate. In the present experiments, mice with M. intestinale ingestion had a significant increase in CYP7A1 content in the liver, impaired FXR-FGF15 signaling in the ileum, and decreased NPC1L1 expression. Therefore, SCFAs may regulate cholesterol metabolism through the following mechanisms. One is to enhance the conversion of cholesterol to bile acids: CYP7A1 is the only rate-limiting enzyme for the conversion of cholesterol to BAs, and upregulation of CYP7A1 by SCFAs leads to increased conversion of cholesterol to BAs, thereby reducing cholesterol levels.^51,52 Second, inhibition of intestinal cholesterol absorption: SCFAs can down-regulate NPC1L1-mediated cholesterol absorption in the ileum.^52,53 In addition, SCFAs have been shown to be indirect antagonists/inhibitors of FXR signaling,⁵⁴ which was also found to be impaired in the present study, in turn reducing enterohepatic circulation and regulating cholesterol metabolism.

BAs is an important pathway for cholesterol excretion in the host, which can be impaired by a high-cholesterol diet, leading to abnormal cholesterol metabolism.⁵⁵ Probiotics can improve cholesterol metabolism by regulating cholesterol synthesis, production of secondary and secondary BAs, and excretion of BAs. For example, strain MCC2760, which is BA tolerant and has BSH activity, causes BA uncoupling in the gut and reduces their reabsorption through the ileum, leading to enhanced excretion of free BA through feces.⁵⁶ Conjugated BAs are converted to primary bile acids by BSH and are reabsorbed by ASBT proteins at the tip of enterocytes.⁵⁷ In this experiment, the expression of ASBT was decreased (Figure S5B), the enterohepatic circulation was reduced, the circulation of primary bile acids was reduced, and the primary bile acids in the ileum increased. This decreased source and increased route of primary bile acids in the ileum caused the content of primary bile acids in the ileum to remain almost unchanged. Secondary BAs are natural activators of FXR, and the reduction of its content directly reduces the activity of FXR signaling pathway in ileal epithelial cells, and then reduces the content of FGF15, the key messenger connecting the gut and liver, deregulates the inhibition of liver CYP7A1, and promotes the synthesis of BAs to accelerate cholesterol metabolism.^58,59 The aforementioned research revealed that Lactobacillus johnsonii CCFM1376 could alleviate hypercholesterolemia in mice by regulating bile acid composition and the FXR pathway.⁷ This study further verified the feasibility of regulating cholesterol metabolism by interfering with bile acid homeostasis. Specifically, the intake of M. intestinale achieved the regulation of cholesterol metabolic balance through negative feedback regulation of maintaining this homeostasis.

Limitations and prospects of this study

In this study, only the cholesterol of M. intestinale was evaluated, and the identified pathways were not verified. In addition, in this study, we observed that EPS-D1 could regulate cholesterol in the FXR-FGF15 pathway through Muribaculum, but whether EPS-D1 is directly involved in cholesterol metabolism or directly regulates the FXR-FGF15 axis should be further explored by molecular docking, cellular experiments, FXR knockout models, and other methods. In vitro experiments suggest that EPS-D1 has the ability to directly adsorb cholesterol, but its contribution in the complex in vivo environment still needs to be investigated. Concurrently, this study necessitates further clinical trials to investigate the effects of M. intestinale on human high-cholesterol metabolism. Additionally, further research is required to gain a comprehensive understanding of the mechanisms by which Muribaculum species in the gut contribute to host cholesterol homeostasis. Although the experiments demonstrated that one Muribaculum species in this study can uptake and metabolize cholesterol, we cannot fully validate the hypothesis that cholesterol is metabolized by other Muribaculum branch species. Culturing and characterizing other Muribaculum strains is essential for future research.

Conclusion

In this study, we demonstrated that EPS-D1, a novel exopolysaccharide from L. plantarum H6 effectively ameliorates hypercholesterolemia in high-cholesterol diet mice through selective enrichment of Muribaculum (Figure 11). This key commensal bacterium regulates the enterohepatic FXR‒FGF15 axis, leading to enhanced bile acid synthesis via CYP7A1 upregulation and reduced cholesterol absorption through ASBT and NPC1L1 downregulation. Additionally, Muribaculum increases specific SCFAs production and modulates the bile acid pool. These findings identify Muribaculum as a precise microbial target for cholesterol management and establish EPS-D1 as a promising therapeutic candidate for hypercholesterolemia, offering a novel microbiota-mediated approach to improve cholesterol metabolism through the gut‒liver axis.

Figure 11.

Mechanism of action of EPS-D1 and Muribaculum in improving cholesterol metabolism. Created in https://BioRender.com.

Supplementary Material

Supplementary material

Related Manuscript File.pdf

Supplementary material

Supplementary_Material_cleaned.docx

Funding Statement

This work was supported by the National Natural Science Foundation of China under Grant (32172189).

Disclosure of potential conflicts of interest

There are no conflicting interests to disclose, according to the authors.

Data availability statement

The Sequencing data for this study are available in the NCBI Sequence Read Archive database (Accession Numbers: PRJNA1285610 and PRJNA1285616).

Supplemental material

Supplemental data for this article can be accessed at https://doi.org/10.1080/19490976.2026.2623578.