1. Introduction
Famous traditional Chinese clinical prescriptions mainly consist of
Fructus gardeniae, the dried ripe fruits of
Gardenia jasminoides Ellis (GE) such as Zhi-Zi-Hou-Pu, Huang-Lian Jie-Du, and Zhi-Zi-Chi decoctions, which have been widely used in the treatment of insomnia, inflammation, hypertension, and gastrointestinal disorders
[1],
[2]. As the most abundant active iridoid glucoside ingredient in
Fructus gardeniae, geniposide possesses diverse physiological effects, including anti-inflammatory, analgesic, antidepressant, and antioxidant effects
[3],
[4],
[5],
[6],
[7]. However, prolonged usage or excessive doses of geniposide results in toxicity, of which hepatotoxicity is considered one of the most fundamental safety concerns for a few Chinese clinical prescriptions attributable to its glycone metabolite genipin
[8],
[9],
[10]. Consequently, attenuation of geniposide hepatotoxicity has become a focus of research. Genipin, a gut microbial reactive metabolite of geniposide, binds spontaneously to amino acids and exhibits higher hepatotoxicity than geniposide, indicating an important role of the gut microbiota in geniposide-induced liver injury
[11],
[12].
According to our previous study, isoflavones present in soy products such as fermented soybeans could potentially alter geniposide metabolism
in vivo and cause a time-dependent decrease in genipin exposure
[13],
[14],
[15]. However, the specific mechanisms and metabolic pathways underlying geniposide metabolism have not been systematically clarified, posing a great challenge for further studies on its toxicity attenuation and rational clinical applications. Interestingly, we also found that long-term addition of isoflavones to the diet caused changes in the gut microbiota composition of rats
[14]. Gut microbiota is known to produce and release numerous drug metabolic enzymes, including α-rhamnosidase, β-glucuronidase (β-Glu), β-glucosidase (β-GC), β-galactosidase (β-Gal), azoreductase, 7-α-hydroxylase, protease, and various carbohydrate enzymes. The components of the traditional herbal medicines and their bioactive compounds are metabolized following a series of metabolic reactions when exposed to the microbiota in the gastrointestinal tract
[16],
[17],
[18].
Inspired by this, the study was designed to make significant advancements in field of geniposide-induced liver injury by revealing the underlying mechanism by which isoflavones regulate geniposide metabolism, focusing on the gut microbiota (
Fig.1). First, the metabolic process of geniposide was elucidated
in vivo. The effects of isoflavones on metabolic enzyme activity and gut microbiota were comprehensively investigated using enzyme-linked immunosorbent assay (ELISA), 16S ribosomal RNA (rRNA) and metagenomic sequencing. Additionally, geniposide metabolism under varying gut microbiota conditions was studied to investigate the role of gut microbiota in attenuating geniposide hepatotoxicity with isoflavones. Based on the above findings, a potential microbial target was discovered to bring promising prospects for the prevention of geniposide-induced liver injury.
2. Materials and methods
2.1. Chemicals and supplies
Soybean (Lot #190321) was purchased from Shanghai Tonghanchuntang Traditional Chinese Medicine (China). Daidzin (Lot #P14M10F83057), daidzein (Lot #C06N6Y5504), genistin (Lot #P09M8F31018), genistein (Lot #H30A9Z69019), glycitin (Lot #G09M6F31015), glycitein (Lot #D16A8F32059), paeoniflorin (Lot #P16M06B25065), amoxapine (Lot #S28HS196364), and β-GC (Lot #A11HS191113) were purchased from Shanghai Yuanye Biotechnology (China) with the purity ≥ 98%. Geniposide (Lot #1203A024, purity ≥ 98%) was purchased from Sichuan Victory Biological Technology (China). Vancomycin, neomycin, penicillin, and metronidazole were purchased from BBI Life Science (China). Urethan was purchased from Meilunbio (China). Carboxymethylcellulose sodium (CMC-Na) and heparin were purchased from Sangon Biotech (China). Liquid de Man–Rogosa–Sharpe (MRS) medium was purchased from Qingdao Hi-Tech Industrial Park Hope Bio-Technology (China). Acetonitrile (mass spectrometry (MS) grade) was purchased from Merk (Germany). Acetic acid (high-performance liquid chromatography (HPLC) grade) was purchased from Tedia (USA). Lactobacilus sp. (BN CC195669) was acquired from the Beina Biologics-Henan Industrial Microbial Strain Engineering Technology Research Center (China). An E.Z.N.A Mag-Bind Soil DNA kit was purchased from OMEGA (USA). The 2× Hieff® Robust PCR Master Mix and Hieff NGS DNA Selection Beads was from Yisheng Biotechnology (China). The Qubit3.0 DNA kit was purchased from Life Technologies (USA). ELISA kits of β-Glu, β-GC, β-Gal, sulfotransferase (SULT), and glutathione S-transferase (GST) for rat along with ELISA kits of β-GC and SULT for human were purchased from Shanghai Enzyme-linked Biotechnology (China). All other reagents were purchased from Titan Technology (China).
2.2. Animals
Male Sprague–Dawley (SD) rats weighing (200 ± 20) g were purchased from Zhaoyan New Drug Research Center (China) with the license number SCXK (Su) 2018-0006. All animal experiments were conducted in accordance with institutional and ethical guidelines and approved by the Committee for Welfare and Ethics of Experimental Animals Welfare and Ethics of Naval Medical University (China).
2.3. Clinical study design and participants
Volunteers of both sexes were recruited for the study. At the screening visit, the following exclusion criteria were considered: current use of antibiotics; concomitant use of non-steroidal anti-inflammatory medications, including aspirin; current treatment for any chronic inflammatory condition or malignancy; previous colonic or small bowel resection; current smoking (minimum six months smoking cessation); long-term alcoholism and pregnancy. If an individual was eligible, the volunteer provided with written informed consent to confirm likely compliance. The clinical studies were approved by the Ethics Committee of Shanghai Tongji Hospital (approval code: K-2017-003-XZ-190130); and Shanghai Traditional Chinese Medicine Hospital (approval code: 2022SHL-KY-37-01). All the studies were conducted in accordance with the tenets of the Declaration of Helsinki.
2.4. Study on the metabolic profile of geniposide in vivo
Twelve rats were randomly divided into two groups: the control group was administered 0.5% CMC-Na solution by gavage, and the treatment group was administered geniposide solution by gavage for seven consecutive days. In this study, three dosage levels of geniposide (100, 200, and 300 mg∙kg
−1) were optimized according to famous traditional Chinese clinical prescriptions mainly consisting of
Fructus gardeniae such as Zhi-Zi-Chi decoction. Ultimately, 200 mg∙kg
−1 geniposide was used based on our previous study
[19]. Plasma, urine, feces, and bile samples were obtained and prepared as described in Appendix A. Volumes of 100 µL plasma, urine, feces, and bile samples were further prepared to achieve the supernatant for analysis with ultra high-performance liquid chromatography-quadrupole time-of-flight mass spectrometry (UHPLC-Q/TOF-MS) (Appendix A). The
in vivo metabolites of geniposide were identified based on mass-to-charge ratio, retention time, characteristic fragment ions, and reported literature
[20],
[21],
[22],
[23],
[24],
[25],
[26].
2.5. Geniposide metabolism in SD rats mediated by isoflavones
Eighteen rats were randomly divided into three groups. Except for group CON, the other groups were given drinking water containing 0.5 mg∙mL
−1 neomycin, 0.5 mg∙mL
−1 ampicillin, 0.5 mg∙mL
−1 metronidazole, and 0.25 mg∙mL
−1 vancomycin for two weeks to establish a pseudo-sterile rat model. In the third week, the CON and AC groups were gavaged with a 0.5% CMC-Na solution, and the ISO group was administered with an isoflavone extract (210 mg∙kg
−1) prepared according to our previous study (Appendix A)
[13],
[27]. Each group was administered geniposide solution at 200 mg∙kg
−1 continuously for 3 weeks according to our previous study
[19]. After geniposide administration, approximately 0.2 mL blood was taken from the posterior ophthalmic vein plexus of each group at 0, 0.25, 0.5, 1, 1.5, 2, 3, 4, 6, 8, 12, and 24 h, transferred into a 1.5 mL Eppendorf (EP) tube containing heparin to obtain the plasma and further prepared for analysis by high performance liquid chromatography-tandem mass spectrometry (HPLC-MS/MS) (Appendix A). The peak area ratio of each analyte to the internal standard (IS) was calculated to obtain the pharmacokinetic characteristics using DAS 2.2 software.
2.6. Molecular virtual docking analysis
Molecular virtual docking analysis was performed to predict geniposide conversion by intestinal enzymes. The crystal structure of the intestinal metabolic enzymes was obtained from the Protein Data Bank (PDB) database, and the structural information of the molecular ligands was downloaded from the PubChem database. The Chemdraw 3D software (v20.0) was used to convert the molecular ligand from sdf format to mol2 format. Molecular docking analysis of intestinal metabolic enzymes and geniposide or genipin was performed using SwissDock. During docking, the protein remained rigid, where the ligands were fully flexible. The parameters were set to default values and the docked conformation corresponding to the lowest binding energy was selected as the most probable binding conformation. The ViewDock plugin of UCSF Chimera (v1.16) was used to show binding patterns.
2.7. Effect of enzyme inhibition on geniposide metabolism
To validate the prediction of the molecular virtual docking analysis, 30 SD rats were randomly divided into five groups. The control group was administered a 0.5% CMC-Na solution by gavage, and the ISO group was administered a solution of the isoflavone extract (210 mg∙kg
−1) by gavage
[13]. After the administration of isoflavone extract, β-GC (100 mg∙kg
−1), amoxapine (AMXP) solution (2.7 mg∙kg
−1), and 2,6-dichloro-4-nitrophenol (DCNP) solution (9 mg∙kg
−1) were continuously administered to rats in group ISO + β-GC, ISO + AMXP, and ISO + DCNP for 21 d, respectively. After 21 d, each group was administered a single dose of geniposide solution (200 mg∙kg
−1)
[19]. Plasma was then obtained and analyzed according to the above steps described in Section 2.5.
2.8. Metabolic enzymes in SD rats mediated by isoflavones
Eighteen rats were randomly divided into three groups. Except for group CON, the other two groups were given drinking water containing 0.5 mg∙mL
−1 neomycin, 0.5 mg∙mL
−1 ampicillin, 0.5 mg∙mL
−1 metronidazole, and 0.25 mg∙mL
−1 vancomycin for 2 weeks to establish a pseudo-sterile rat model. In the third week, group CON, AC, and ISO were gavaged with 0.5% CMC-Na solution, 0.5% CMC-Na solution, and isoflavone extract (210 mg∙kg
−1)
[13] for 3 weeks, respectively. Finally, all the rats were intraperitoneally injected with urethane and subsequently perfused with normal saline to remove the cecum. The levels of β-Glu, β-GC, β-Gal, SULT, and GST in the cecum contents were determined respectively with the corresponding ELISA kits according to the manufacturer’s instructions.
2.9. Detection of total DNA from rat feces and PCR amplification
Thirty rats were randomly divided into five groups. Rats in group CON were gavaged with 0.5% CMC-Na solution for 7 d, while groups D7, D14, D21, and D28 were gavaged with isoflavone extract (210 mg∙kg
−1) for 7, 14, 21, and 28 d, respectively
[13],
[27]. Fresh fecal samples from each group were collected and prepared for extraction of total fecal DNA using an E.Z.N.A Mag-Bind Soil DNA Kit (Appendix A). The integrity and concentration of the extracted DNA were determined using an agarose gel method and a DNA detection kit (Qubit 3.0). The V3-V4 variable region of the 16S rRNA gene was amplified with primers 341F (5′-CCTACGGGNGGCWGCAG-3′) and 805R (5′-GACTACHVGGG-TATCTAATCC-3′) (Tables S6–S9 in Appendix A). After purification, the samples were sequenced using Illumina MiSeq to analyze the gut microbiota diversity.
2.10. Bioinformatics
For 16S rRNA sequencing, the original data files of the fecal samples were transformed into raw data using base recognition analysis. After barcode tag sequence identification, the effective data were obtained through quality control screening. Then the operational taxonomic unit (OTU) clustering, species annotation, α-diversity analysis, and β-diversity analysis were calculated and analyzed statistically with the Usearch software (v11.0.667) and ribosomal database project (RDP) classifier (v2.12). Statistical analysis of the community structure was carried out at each classification level according to taxonomic information. For metagenomic sequencing, Illumina paired-end 150 bp (PE 150) sequencing was performed after assessing DNA sample quality. Readfq was used to process the raw data to obtain clean data for subsequent analysis, and the data were assembled and analyzed using the MEGAHIT software (v1.0.4-beta). MetaGeneMark was used for gene prediction, CD-HIT software was used to delete redundant genes, DIAMOND software was used for sequence comparison, and the lowest common ancestor (LCA) algorithm of MEGAN software was used to determine the species annotation information of the sequences.
2.11. Metabolic enzymes and geniposide metabolism in rats regulated with different gut microbiota conditions
Thirty rats were randomly divided into five groups. Except for group CON, the other four groups were given drinking water containing 0.5 mg∙mL
−1 neomycin, 0.5 mg∙mL
−1 ampicillin, 0.5 mg∙mL
−1 metronidazole, and 0.25 mg∙mL
−1 vancomycin for 2 weeks to establish a pseudo-sterile rat model. In the third week, the CON, AC, ISO, fecal microbiota transplantation (FMT), and LAC groups were gavaged with 0.5% CMC-Na solution, 0.5% CMC-Na solution, isoflavone extract (210 mg∙kg
−1)
[13], 1 mL fecal bacterial suspension, and 1 mL
Lactobacillus resuspension respectively, for 3 weeks (Appendix A). Geniposide solution was then administered to each group at a dose of 200 mg∙kg
−1 according to our previous study. Plasma samples were obtained to study pharmacokinetic characteristics, as described in Section 2.5. Finally, the cecum of each rat was removed as Section 2.8 for metabolic enzymes analysis.
2.12. Clinical trials to verify the effect of isoflavones on human gut microbiota and intestinal enzymes
These studies were designed rigorously to verify the effect of isoflavones on human gut microbiota and intestinal enzymes, as volunteers were unable to consume geniposide, which is not yet an approved drug. In each study, volunteers with a soy-product-free diet and soy-product-rich diet were strictly recruited, and 20 eligible volunteers were divided into two groups, namely group CON and ISO with the screening criteria of not eating soy products and a long-term diet rich in soy products (minimum three months). Participants in the ISO group were asked to maintain a soy-product-rich diet, while those in the CON group were required to eat a soy-product-free diet for 2 weeks to collect fresh fecal samples (obtained with a fecal collection device) for systematic biological analysis. For 16S rRNA gene and metagenomic sequencing, microbial DNA extraction, PCR amplification and Illumina sequencing were performed as described in Section 2.10. The contents of β-GC and SULT in the fecal samples were determined with the corresponding ELISA kits.
2.13. Lactobacillus spp. intervened in geniposide-induced liver injury
Eighteen rats were randomly divided into three groups. The CON group was administered a 0.5% CMC-Na solution. Based on the established pseudo-sterile rat model, rats in the GPS group and LAC group were administered 0.5% CMC-Na solution and 1 mL
Lactobacillus resuspension respectively, for 3 weeks. The GPS and LAC groups were gavaged with geniposide solution at 200 mg∙kg
−1 for 5 consecutive days. Finally, plasma and liver samples were collected from all rats. Plasma samples were analyzed using an automatic biochemical analyser (HITACHI 7600, Japan) to determine alanine aminotransferase (ALT), aspartate aminotransferase (AST), and total bilirubin (TBIL). Liver tissues were fixed in formalin solution containing 10% neutral buffer and embedded in paraffin. Tissue was cut at 4 μm thickness and stained with hematoxylin and eosin (H&E). Pathological changes in the liver tissue were determined according to histopathological methods, and liver inflammation and hepatocyte death were assessed in a blinded manner using a modified histologic activity index (HAI)
[28].
2.14. Statistical analysis
Data were expressed as mean ± standard error (mean ± SD). Statistical differences were analyzed using one-way analysis of variance (ANOVA) or Student’s t-test. Data were considered statistically significant at P values less than 0.05. SPSS software (v16.0; SPSS, USA) and Excel 2016 (Microsoft, USA) were used for statistical analyses. Adobe Illustrator 2018 (Adobe Systems, USA) and GraphPad prism 6.0 were used for mapping.
3. Results
3.1. Metabolic profile of geniposide in vivo
Geniposide and its 24 metabolites were identified in rat plasma, urine, feces, and bile samples. The retention time, and secondary fragment ions of geniposide (M0) were consistent with those of the geniposide reference substance, indicating that M0 was geniposide (
Figs. 2(a)–(c)). In the positive ion mode, eight metabolites were identified (Table S10 in Appendix A), including four metabolites in feces, seven metabolites in urine, eight metabolites in bile, and four metabolites in plasma. In the negative ion mode, 19 metabolites were identified, including ten in feces, 15 in urine, 17 in bile, and 13 in plasma. Two different geniposide metabolic pathways were identified
in vivo (
Fig. 2(d)). One pathway involves geniposide metabolism in the form of a drug prototype, including the demethylation of geniposide to form geniposidic acid (M4), glucuronic acid reaction of M4 to form glucuronide conjugate of geniposidic acid (M8), demethylation of M8 to form demethylation of glucuronide conjugation of geniposidic acid (M9), and hydroxylation of geniposide to form 6α-hydroxygeniposide (M10). Another metabolic pathway involves the removal of one molecule of glucose from geniposide to form genipin (M11) followed by a series of metabolic reactions
in vivo, including the conjugation of taurine to genipin to form hydrolysis and taurine conjugation of geniposide (M12), conjugation of sulfate to genipin to form sulfate conjugate of genipin (M15), hydroxylation of M15 to form M16, and glucuronidation of genipin to form glucuronide conjugate of genipin (M17). As the peak area of each metabolite indicated that the drug prototype and its five main metabolites, including M11, M4, M8, M15, and M17 were the main forms of geniposide
in vivo, the above five metabolites were subsequently fixed to clarify geniposide metabolism.
3.2. Metabolism of geniposide mediated by isoflavones
According to pharmacokinetic characteristics of geniposide and its metabolites (Table S14 in Appendix A), the maximum plasma concentration (Cmax) and area under the curve from time 0 to infinity (AUC0–∞) of geniposide was significantly lower (P < 0.0001) while the Cmax and AUC0–∞ of M11, M4, M8, M15, and M17 (P < 0.001, P < 0.0001, P < 0.0001, P < 0.0001, and P < 0.001, respectively) were significantly higher in group ISO compared to group AC. These results indicated that geniposide metabolism was promoted by administration of isoflavones. Both geniposide and genipin showed distinct double peaks with a second peak at 2–4 h. This phenomenon may be attributed to enterohepatic recycling, suggesting that geniposide and genipin are likely reabsorbed into the intestinal tract after secretion into bile. Compared to the other two metabolites, geniposide, M11, M4, and M8 exhibited a relatively short time to maximum plasma concentration (Tmax), indicating that geniposide could be rapidly metabolized into genipin, geniposidic acid, and the glucuronide conjugate of geniposidic acid in vivo. The Tmax of M15 and M17 was 4–5 h, indicating that the sulfonation and glucuronidation of genipin mainly occurred from the ileum to the cecum. Besides, higher Cmax and AUC0–∞ of geniposide and lower Cmax and AUC0–∞ of genipin was noticed after pretreated with broad-spectrum antibiotic, which suggested that gut microbiota played an important role in geniposide metabolism.
3.3. Intestinal enzymes involved in geniposide transformation in vivo
Crystal structure of β-GC 3VKK and SULT 1CJM obtained from the PDB database was shown in
Fig. 3(a). The binding free energy of geniposide and β-GC, genipin and β-GC, geniposide and SULT, and genipin and SULT was −2.30, 8.30, 10.15, and −19.09 kcal∙mol
−1, respectively (
Figs. 3(b)–(e)). The results indicated that the molecules of geniposide and β-GC, along with genipin and SULT exhibited stronger docking capability. As shown in the docking results, many intermolecular bonds existed between geniposide and β-GC at the active site of β-GC, such as pink-marked VAL 168, VAL 171, LEU 176, ALA 246, PHE 249, and ALA 330 (
Fig. 3(b)). In addition, a few hydrophilic bonds (hydrogen bonds) were found at the active site of SULT, such as the pink-marked LYS 16, ALA 112, and GLN 116, which may be the main binding forces between genipin and SULT (
Fig. 3(e)).
AMXP and DCNP are the inhibitor of β-GC and SULT, respectively. In the experiment, AMXP, β-GC, and DCNP were administered orally with isoflavones for 21 d (
Fig. 3(f)). Compared to the control group, the
Cmax and AUC
0–∞ of geniposide increased with the combined pretreatment of isoflavones and AMXP while decreased with the pretreatment of isoflavones, isoflavones, and DCNP, or, isoflavones, and β-GC (
Figs. 3(g) and
(h)). Each exposure to genipin, the first metabolite of geniposide, was lowest in the ISO + AMXP group (
Figs. 3(g) and
(i)). The results suggested that β-GC levels were implicated in geniposide metabolism, also indicated that isoflavones regulated geniposide metabolism by altering β-GC levels. Additionally, the AUC
0–∞ and
Cmax of M15 and M17 were lower when pretreated with isoflavones and AMXP, or isoflavones and DCNP. With combined pretreatment with isoflavones and β-GC, the AUC
0–∞ and
Cmax of M15 and M17 were the highest (
Figs. 3(g) and
(j)). The results above draw the conclusion that geniposide was the substrate of β-GC, and genipin was the substrate of SULT 1CJM, which mainly abundant in gut and further confirm that by altering β-GC and SULT, geniposide metabolism and genipin transformation
in vivo could be promoted.
3.4. Effects of isoflavones on metabolic enzymes
The levels of β-Glu, β-GC, β-Gal, SULT, and GST in the cecum contents of normal rats were (53.49 ± 5.31), (33.37 ± 1.12), (38.79 ± 4.15), (1.74 ± 0.18), and (53.34 ± 10.25) ng∙mL−1, respectively. Compared with those of the control group, the contents of β-Glu ((47.21 ± 3.27) ng∙mL−1), β-GC ((29.69 ± 1.80) ng∙mL−1), β-Gal ((33.92 ± 2.16) ng∙mL−1), SULT ((1.49 ± 0.16) ng∙mL−1), and GST ((42.90 ± 2.77) ng∙mL−1) were significantly lower in the pseudo-sterile rats (P < 0.05). With isoflavones pretreatment by gavage for 3 weeks, the level of β-Glu, β-GC, β-Gal, SULT, and GST was (59.52 ± 3.55), (36.94 ± 3.30), (43.20 ± 2.06), (2.00 ± 0.16), and (58.25 ± 7.16) ng∙mL−1, respectively, significantly higher than the control group (P < 0.05). The results above suggested that the activities of β-Glu, β-GC, β-Gal, SULT, and GST in the cecum could be up-regulated with isoflavones. Interestingly, the association between the main pharmacokinetic characteristics such as Cmax and enzyme activity indicated that geniposide metabolism is closely related to enzyme levels. It was plain to see the transformation of geniposide and genipin showed an obviously positive correlation of enzyme activities. In addition, Cmax and AUC0–∞ of geniposidic acid, the glucuronide conjugate of geniposidic acid, the sulfate conjugate of genipin, and the glucuronide conjugate of genipin in the ISO group significantly increased (P < 0.05), while those in the AC group significantly decreased compared with normal rats, further confirming that administration of isoflavone could improve metabolic enzyme activity. The results also indicated that isoflavones regulated geniposide metabolism by altering the levels of β-GC and SULT.
3.5. Isoflavones regulated gut microbiota composition in rats
The animal experiment with different interventions was conducted according to
Fig. 4(a). Compared with the control group, group D7 showed no significant difference in gut microbiota diversity (
P > 0.05), where groups D14, D21, and D28 exhibited significant differences (
P < 0.05) as Shannon and Simpson indices indicated (
Figs. 4(b) and
(c)). The Chao and Ace indices in groups D7, D14, D21, and D28 were significantly higher than those in the control group (
P < 0.05) (
Figs. 4(d) and
(e)). These results suggest that the gut microbiota diversity in rats was altered by oral administration of isoflavones. In addition, the richness and diversity of the gut microbiota showed a positive trend with prolonged exposure to the isoflavones. The gut microbiota of groups D7, D14, D21, and D28 were significantly different from that of the control group at the OTU level (
Fig. 4(f)), indicating that isoflavones would cause changes in the gut microbiota structure. Additionally, with the extension of the isoflavone administration time, the composition of the gut microbiota continued to change until 21 d as group D21 and D28 were relatively close.
At the phylum level, the species with top 6 relative abundance were
Firmicutes,
Bacteroides,
Proteobacteria,
Verrucomicrobiota,
Actinobacteria, and
Candidatus_
Saccharibacteria (
Fig. 4(g)). Compared to the CON group, there were significant differences in the relative abundance of
Firmicutes,
Bacteroides,
Proteobacteria,
Verrucomirobia, and
Candidatus Saccharibacteria in groups D7, D14, D21, and D28. At the genus level, species with the top 30 relative abundance included
Lactobacillus,
Alistipes,
Akkermansia,
Desulfovibrio,
Ruminococcus,
Parabacteroides,
Bacteroides, and
Faecalibacterium (Table S15 in Appendix A). Notably,
Lactobacillus accounted for the largest relative abundance in the different groups (
Fig. 4(h)). And significant increase in the relative abundance of
Lactobacillus,
Bifidobacterium,
Alistipes and decreased relative abundance of
Desulfovibrio were observed with the pretreatment of isoflavones for 7, 14, 21, and 28 d (
Figs. 4(i)–(l)). Species exhibiting significant differences in abundance between different groups were identified under the conditions where the LDA score exceeds 2 (
Fig. 4(m)). The evolutionary branching diagram of different species, with circles radiating from the inside out representing the classification level from phylum to genus (or species) (
Fig. 4(n)), indicated that the relative abundance of probiotics increased significantly with prolonged administration time.
3.6. Metabolic enzymes and geniposide metabolism in rats regulated with different gut microbiota conditions
In the animal experiment (
Fig. 5(a)), it was found that the contents of β-Glu, β-GC, β-Gal, SULT, and GST decreased significantly in the pseudo-sterile group, but could restored to the extent far surpassing the level of the control group with the intervention of FMT. Notably, similar trends were observed in the levels of metabolic enzymes above after administration of
Lactobacillus spp., which confirmed that the gut microbiota played a vital role in the regulation of metabolic enzymes (
Figs. 5(b)–(g)).
To validate the effect of the gut microbiota on geniposide metabolism, the pharmacokinetic characteristic of geniposide increased in the pseudo-sterile group, whereas exposure to geniposide and genipin significantly decreased in the FMT group (
Figs. 5(h)–(j)). These results confirm that the disruption of gut microbiota homeostasis by antibiotics inhibits geniposide metabolism resulting in increased geniposide exposure and decreased exposure to its metabolites. After remodeling the gut microbiota with different interventions, the levels of gut metabolic enzymes can be reset, promoting the transformation of geniposide and genipin
in vivo. In addition, it was comprehensively shown that the exposure to both geniposide and genipin and the metabolic enzymes were closely related to gut microbiota conditions (
Figs. 5(k) and
(l)). These results further confirm that the gut microbiota play a critical role in geniposide metabolism.
3.7. Isoflavones rich diet was verified to regulate human gut microbiota and intestinal metabolic enzymes
To further confirm the effect of isoflavones rich diet on human gut microbiota and intestinal metabolic enzymes, fresh fecal samples from volunteers with the diets without soy products or soy product rich diets were collected for analysis of intestinal flora structure, and the activities of intestinal metabolic enzyme β-GC and SULT in fecal samples were characterized. According to the results, β-GC (
P < 0.05) and SULT (
P < 0.0001) (
Fig. 5(m)) was spotted in volunteers with soy product rich diets compared with group CON.
16S rRNA sequencing showed that 17 phyla, including
Firmicutes,
Bacteroides,
Proteobacteria,
Verrucomicrobiota, and
Actinobacteriota were present in fresh fecal samples from the volunteers (
Fig. 6(a)). And a total of 226 genera were identified, including
Bacteroides,
Faecalibacterium,
Parabacteroides,
Akkermansia,
Blautia,
Lachnoclostridium,
Ruminococcus, and
Lactobacillus (
Fig. 6(b)) at the level of genus. Compared to volunteers fed a soy-product-free diet, there was a significantly higher relative abundance of
Bacteroidota at the phylum level and
Bacteroides at the genus level in volunteers fed soy-product-rich diets (Tables S16 and S17 in Appendix A). A significant increased
Lactobacillus (
Fig. 6(c)) was observed in volunteers fed soy-product-rich diets compared to that in the CON group. Metagenomic sequencing identified 9 977 412 gene families and assembled the sequences of 11 023 species. After removing unannotated species, 9 126 microbial species representing 152 phyla and 2 638 genera were identified in the volunteers. The ISO group showed a significant difference in gut microbiota diversity (
P < 0.05) compared with the control group. Higher Shannon, Chao, and Ace indices and lower Simpson indices in the ISO group indicated that gut microbiota diversity in volunteers was altered by diets rich in soy products. Of the 9 126 identified microbial species, a gut microbiome signature comprised of 20 discriminatory species was identified (LDA > 3.0). Specifically, the identified signature included increases in the abundances of
Alistipes putredinis,
Lactobacillus rogosae,
Lactobacillus rhamnosus,
Lactobacillus sp.,
Lactobacillus ruminis CAG: 367,
Bacteroides thetaiotaomicron, and
Faecalibacterium prausnitzii, accompanied by a decrease in the abundances of
Prevotella copri,
Megamonas funiformis, and
Faecalibacterium longum. Notably, despite its low abundance,
Lactobacillus rhamnosus was one of the most critically beneficial species for discriminating between the control group with a soy-product-free diet and volunteers with a soy product rich diet in our dataset. A significant increase
Lactobacillus rhamnosus (
Fig. 6(d)) was observed in volunteers fed the soy-product-rich diets. Metabolism was a function of the marker genes (
Fig. 6(e)). Glycosidase hydrolases play an important role in the metabolism of all volunteers (
Fig. 6(f)). In addition, non-metric multidimensional scaling (NMDS) plots of carbohydrate enzyme based on CAZY levels 1 and 2 further revealed that the composition of carbohydrate enzyme changed with soy product rich diets (
Figs. 6(g) and
(h)). Especially, GH50 and GH116 (β-GC) would be the marker enzymes for volunteers with soy product rich diets (
Figs. 6(i) and
(j)). These results verify that isoflavones regulated human gut microbiota and intestinal metabolic enzymes as they do in rats.
3.8. Lactobacillus spp. intervened in geniposide-induced liver injury
In the animal experiment (
Fig. 7(a)), the body weight changes in rats treated with different interventions were recorded in detail (
Fig. 7(b)). Biochemical indices showed that the levels of ALT, AST, and TBIL significantly increased after treatment with multiple geniposide doses. However, no increase in biochemical indices was observed after pre-administration
Lactobacillus spp. for 21 consecutive days (
Fig. 7(c)). In addition, multiple doses of geniposide induced histological alterations in the rat liver, embodied by inflammatory cell infiltration and hepatocellular necrosis, whereas
Lactobacillus pretreatment resulted in notable histological improvements (
Fig. 7(d)). The significantly decreased modified HAI score in the LAC group indicates reduced liver damage (
Fig. 7(e)). These results suggested that
Lactobacillus spp. can prevent geniposide-induced liver injury.
4. Discussion
As the main active component of GE, geniposide has neuroprotective, anti-inflammatory, analgesic, and other activities
[4],
[29]. However, prolonged or high-dose use of geniposide results in hepatotoxicity
[8]. Notably, genipin, a reactive gut microbial metabolite of geniposide possesses much higher liver toxicity, indicating a significant role for the gut microbiota in geniposide-induced liver injury
[11]. We have previously revealed that isoflavones in soy products have the potential to modulate the gut microbiota and alter the pharmacokinetic profile of geniposide
[13],
[14],
[15]. Here, we focused on the role of the gut microbiota and uncovered how isoflavone-containing diets regulate geniposide metabolism to reduce liver toxicity.
4.1. Isoflavones regulated geniposide metabolism by altering intestinal enzymes
To investigate the mechanism of geniposide metabolism regulated by isoflavones, it is of great significance to clarify the metabolic process of geniposide. However, previous studies have mainly focused on the metabolites of geniposide in medicinal materials or its metabolites in plasma or urine after a single dose
[30],
[31]. In this study, the full-scale metabolic process of geniposide
in vivo was deduced, and geniposide and its five metabolites, the main form of geniposide
in vivo were selected for further study. Geniposide were primarily metabolized by intestinal enzymes β-Glu, β-GC, β-Gal, SULT, and GST, such as geniposide was metabolized into genipin by intestinal β-GC
[32],
[33],
[34]. The main toxic metabolite genipin, undergoes a further conjugation reaction through SULT to generate other non-toxic metabolites
[32],
[35]. Our study combining the main enzymes mentioned above and the pharmacokinetic characteristics of geniposide and its main metabolites with oral gavage of isoflavones confirmed that isoflavones could effectively alter the geniposide metabolism by regulating the activity of the main metabolic enzymes (
Figs. 5(k) and
(l)), and enzyme inhibition experiments further demonstrated this point.
Notably, β-GC, β-Glu, and other metabolic enzymes abundant in the intestine are mainly produced and released by the gut microbiota, such as
Escherichia coli would produce β-Glu during its growth process
[36], and the synthesis and the activities of extracellular β-GC would be promoted by
Lactobacillus [37],
[38]. Accordingly, changes in the structure of gut microbiota affect the levels of various metabolic enzymes. In the study, we showed that disruption of gut microbiota homeostasis via administration of antibiotics
[16],
[39] would result in a significant decrease in β-GC levels and increased geniposide exposure
in vivo. Similarly, the levels of β-Glu, SULT, and other intestinal enzymes decreased along with hindered biotransformation of genipin. These findings provide new insights into the regulation of these metabolic enzymes by altering the gut microbiota structure.
4.2. Enzymes involved in geniposide metabolism was regulated in a gut microbiota-dependent manner
The gut microbiota, a community of microorganisms living in the human gastrointestinal tract, participate in many important physiological activities by interacting with the host
[16],
[40],
[41],
[42],
[43]. Emerging studies have found that the gut microbiota can be modulated by drugs or dietary interventions, such as fermented soybeans rich in proteins and isoflavones that escape digestion in the upper digestive tract and become substrates for microbiota residing in the gut
[14],
[44],
[45],
[46],
[47]. We showed that oral gavage of isoflavones from fermented soybeans resulted in a unique community of gut microbiota, characterized by an increased relative abundance of probiotics such as
Lactobacillus and
Bifidobacterium, and a decreased relative abundance of harmful bacteria such as
Desulfovibrio (
Figs. 4(i)–(l)). Inspired by the fact that the structure of the gut microbiota in rats reached a stable state by the 21st day with continuous interventions of isoflavones, an intervention period of 21 days for the gut microbiota was adopted in the study.
In general, the gut microbiota mainly affects drug metabolism by secreting metabolites that directly, and indirectly affect related drug metabolism enzymes and transporters
[48],
[49],
[50]. To further clarify the correlation between isoflavones, intestinal levels, and gut microbiota, FMT was performed using feces from rats administered isoflavones for 21 consecutive days. In recent years, FMT, an effective means of reconstructing intestinal flora by implanting functional bacteria in the feces of healthy individuals into the gastrointestinal tract of the recipient, has been adopted to treat diseases such as
Clostridium difficile colitis and malignant tumors
[51],
[52],
[53]. With the intervention of FMT, the significant decreased contents of β-Glu, β-GC, β-Gal, SULT, and GST in the pseudo-sterile group were restored to the extent far surpassing the level of the control group (
Fig. 5(g)). With metagenomics sequencing, long-term consumption of soy isoflavones regulated human gut microbiota and intestinal metabolic enzymes characterized by significantly increased intestinal metabolic enzymes and
Lactobacillus, especially
Lactobacillus rhamnosus further comprehensively revealed that isoflavones alter intestinal drug metabolic enzymes by mediating gut microbiota.
4.3. Gut microbiota, a promising strategy for alleviating geniposide hepatotoxicity by interacting with isoflavones
Based on these important findings, isoflavones regulate geniposide metabolism by regulating metabolic enzymes through the gut microbiota, which may be helpful in alleviating geniposide hepatotoxicity in rats. Notably, as the dominant beneficial bacterium, the relative abundance of
Lactobacillus in the intestine is regulated by isoflavone treatment. Besides, consecutive gavage of
Lactobacillus would result in higher levels of intestinal metabolic enzymes such as β-GC and SULT with decreased exposure of geniposide and genipin
in vivo, which showed the same effect as consecutive gavage of isoflavones and significantly prevented from geniposide hepatotoxicity (
Figs. 7(d) and
(e)). These findings suggest that
Lactobacillus is a potential microbial target for the management of geniposide hepatotoxicity by interacting with isoflavones. Soy isoflavones also have been identified as a common dietary component. Generally, tofu, soymilk, soy cheese slices, and natto were known rich in isoflavones
[54],
[55]. In our study, volunteers tended to eat soy products such as instant soymilk, tofu, bean sprouts, tofu skin, and soy flour (full fat). Using the US Department of Agriculture (USDA) database for the Isoflavone Content of Selected Foods (Release 2.0), the daily consumption of isoflavones in the diet of each volunteer was calculated to be approximately 150–300 mg∙d
−1 [56]. Our clinical study revealed that soy isoflavones regulated human gut microbiota and intestinal metabolic enzymes in the same way as in rats, and a higher intake of soy isoflavones in the daily diet was accompanied by significantly higher levels of beneficial species and related intestinal metabolic enzymes. These findings highlight the importance of soy isoflavone-rich dietary strategies containing soy isoflavones in the range of 150–300 mg∙d
−1 to regulate geniposide metabolism and prevent drug-induced hepatotoxicity.
5. Conclusions
This study clarified the metabolic profile of geniposide in vivo, which was altered with altered with isoflavones by mediating specific enzyme activities, including β-GC and SULT with enzyme inhibitor experiments. FMT established the direct contribution of gut microbiota to the regulation of enzyme activities and geniposide metabolism, as key enzymes involved in geniposide metabolism were implicated in the alteration of gut microbiota composition with isoflavone intervention. Hence, this study comprehensively demonstrated that isoflavones in soy products regulate geniposide metabolism by mediating metabolic enzymes through the gut microbiota, especially Lactobacillus spp.. A soy-product-rich diet could effectively regulate the relative abundance of Lactobacillus and metabolic enzymes levels were further verified in our clinical trials. Furthermore, Lactobacillus spp. have been identified as potential microbial targets for reducing geniposide hepatotoxicity. These findings provide valuable insights into the rational use of drugs and diets to attenuate geniposide hepatotoxicity from the perspective of gut microbiota.
CRediT authorship contribution statement
Wen Yang: Methodology, Investigation, Formal analysis. Wen Zhang: Investigation. Xinhui Huang: Validation, Investigation, Formal analysis. Shuwen Geng: Validation, Data curation. Yujia Zhai: Validation, Formal analysis. Yuetong Jiang: Formal analysis. Tian Tian: Formal analysis. Yuye Gao: Formal analysis. Jing He: Methodology. Taohong Huang: Methodology. Yunxia Li: Investigation. Wenjing Zhang: Investigation. Jun Wen: Writing – review & editing. Jian-lin Wu: Writing – review & editing, Funding acquisition. Guangji Wang: Writing – review & editing. Tingting Zhou: Writing – review & editing, Supervision, Resources, Funding acquisition.
Declaration of competing interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgments
The research was supported by the National Natural Science Foundation of China (82461160264, 82474056, and 82104124), the grant from the Noncommunicable Chronic Diseases-National Science and Technology Major Project (2023ZD0502605), the Science and Technology Development Fund, Macau SAR (FDCT 0025/2021/A1), the Shanghai Municipal Health Commission (2022XD037), and the Shanghai Magnolia Talent Plan Pujiang Project (23PJD113). The authors would like to thank the National Demonstration Center for the Experimental Nautical Medicine Education of Naval Medical University for providing the necessary facilities and resources for this research.
Appendix A. Supplementary data
Supplementary data to this article can be found online at
https://doi.org/10.1016/j.eng.2024.10.023.
PDF
(4927KB)
