Anti-diabetic effect of di-caffeoylquinic acid is associated with the modulation of gut microbiota and bile acid metabolism

doi:10.21203/rs.3.rs-3875238/v1

Download PDF

Research Article

Anti-diabetic effect of di-caffeoylquinic acid is associated with the modulation of gut microbiota and bile acid metabolism

https://doi.org/10.21203/rs.3.rs-3875238/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Background

The human gut microbiome plays a critical role in both health and disease. A classic example of host-gut microbial co-metabolism involves bile acids, which biosynthesis in liver are excreted into the intestine where they are deconjugated and transformed by the gut microbiota, this process, in turn, activates signaling pathways, influencing host glycolipid and energy metabolism. Ilex tea exhibits properties that alleviate disruptions in lipid metabolism and inflammation by modulating the gut microbiota, yet the underlying mechanism remains unelucidated. DiCQAs is one of the most active and abundant polyphenolic pigments in Ilex tea. Here, we investigated diCQAs regulate diabetes through the BA-related pathway, using HFD + STZ-induced diabetic mice model and long-term mice group to exclude direct stimulatory effects, and studied gut microbiota structure and functions in mice.

Results

Here, we show that diCQAs alleviating symptoms of diabetic mice by alters gut microbiota carrying the BSH gene which associated with obesity and diabetes mellitus. DiCQAs protecting the intestinal barrier while increased enterohepatic circulation conjugated BAs, inhibited the FXR-FGF15 signaling axis in the ileum decreased hepatic FGFR4 protein expression, increased bile acid synthesis in liver, increased BA efflux to reduces hepatic BA stasis, decreased hepatic and plasma cholesterol levels. Moreover, diCQAs induce an upregulation of glucolipid metabolism-related proteins in the liver and muscle (AKT/GSK3β, AMPK), ultimately alleviating hyperglycemia. Additionally, they reduce inflammation by down-regulating the MAPK signaling pathway in the diabetic group.

Conclusions

Our findings provide insights into the mechanisms underlying the anti-diabetic effects of ilex tea. They suggest that reducing gut microbiota (specifically Acetatifactor sp011959105 and Acetatifactor muris) carrying the BSH gene could potentially serve as an anti-diabetic therapy by decreasing FXR-FGF15 signaling.

Anti-diabetic

Di-caffeoylquinic acid

Gut microbiota

Bile acids

Bile Salt Hydrolase

Metagenome-Assembled Genomes

Intestinal barrier

Glucolipid

Inflammation

The human intestinal tract harbors trillions of microorganisms known as the gut microbiota, which interact extensively with the host through metabolic exchange and co-metabolism of substrates [1]. Several mechanisms and metabolites through which gut microbiota may influence the development of metabolic diseases have been reported. For instance, Local immune system interacts with gut microbiota to control the immune responses, LPS from gram-negative bacteria outer leaflet of the outer membrane induce a chronic subclinical inflammatory process in T2DM through activation of TLR4 [2] and Bacteroides fragilis glycosphingolipids reduces colonic invariant natural killer T cell numbers to against autoimmune and allergic disorders [3]; In comparison to HFD mice, GF mice exhibited resistance to obesity. However, this resistance was diminished after fecal microbiota transplantation, because of the gut microbiota selectively suppressed Fiaf [4]; Bioactive metabolites produced by the gut microbiota, such as SCFAs improve insulin sensitivity and increase GLP-1 intestinal secretion, alleviating diabetes mellitus [5] and BCAA which elevated lead to mTOR activation, resulting in the phosphorylation of IRS-1 further inhibition of PI3K/AKT pathway, leading to impaired glucose transport [6, 7]. Addtionaly, Bile acids conversion is another important mechanisms of the gut microbiota with major impact on host metabolism, in particular, inhibited of intestinal FXR protect against metabolic disease [8, 9].

Over the last 20 years, an expanding body of research has demonstrated shifts in bile acids (BAs) metabolism among individuals with diabetes mellitus. It has been observed that manipulating the BAs pool can lead to enhanced glycemic control in these patients [9]. BAs are increasingly acknowledged for their role as metabolic regulators. Through the activation of distinct signaling pathways, they play a pivotal role in governing BAs, lipid, and glucose metabolism [10, 11]. The synthesis and metabolism of BAs is complex and highly regulated. In the liver, cholesterol undergoes synthesis into BAs through the classical pathway mediated by the enzyme CYP7A1. The synthesized BAs are secreted into bile by hepatocytes through the BSEP and stored in the gallbladder, then released into the intestine, where they undergo modifications by gut microbiota before being reabsorbed into the liver via the portal vein. Conjugated BAs are subjected to the bile salt hydrolase (BSH, EC 3.5.1.24)) enzyme, leading to their deconjugation. Research findings have demonstrated that germ-free mice exhibit a protective effect against obesity and insulin resistance when subjected to a high-fat diet, accompanied by significantly elevated levels of conjugated BAs compared to normal mice [4, 12]. Moreover, Li et al. established that inhibiting FXR signaling in the intestine alleviates high-fat diet-induced disruptions in glucose and lipid metabolism by modulating Lactobacillus relative abundance to reduce BSH activity, and by increasing T.β.MCA content to inhibit FXR signaling. Additionally, the study highlighted that intestinal inhibition of FXR signaling mitigated high-fat diet-induced metabolic disorders through the intestine-specific of knockout Fxr^ΔIE mice [13]. BAs play a pivotal role as signaling molecules, regulating the expression of the FXR protein in the ileum and influencing the production of FGF15, which is subsequently reabsorbed into the liver then binds to liver FGFR4 receptors, thereby a enterohepatic FXR-FGF15 signaling axis that regulates liver FGFR4 protein expression, leading to modulation of cholesterol levels and BA synthesis in the liver [14].

DiCQAs primarily consist of 3,4-diCQA, 3,5-diCQA, and 4,5-diCQA are extracted from Ilex tea which is a popular beverage in China. Our previous studies have demonstrated that diCQAs modulate the gut microbiota and alleviated lipid metabolism disruptions as well as inflammation induced by a HFD [15]. Given the pivotal role of the BSH gene carried by the gut microbiota in the hydrolyzation of conjugated BAs, and T.β.MCA as an inhibitor of the FXR-FGF15 signaling pathway in ileum. We hypothesize that the mechanism through which diCQAs regulate hyperglycemia is likely intricately connected to the gut microbiota. Through long-term gavage with diCQAs, we observed that diCQAs failed enhance glycolipid metabolism and also BA metabolism directly. However, In a diabetic mouse model induced by HFD + STZ, we observed that diCQAs regulated gut microbiota BSH ability by reducing the relative abundance of BSH-carrying bacteria, Lachnospiraceae, Acetatifactor, Acetatifactor sp011959105 and Acetatifactor muris, resulting in increased T.β.MCA content in plasma. Followed by inhibited the enterohepatic FXR-FGF15 signaling cascade, leading to heightened BA synthesis, reduced liver cholesterol accumulation, diminished activation of the inflammatory MAPK/NF-κB pathway, and concurrent activation of AKT/GSK3β and AMPK. These combined effects enhanced body glucolipid metabolism, alleviated hyperglycemia and inflammation induced by HFD + SZT.

Chemicals and reagents. Standard chow diet (1010001, Jiangsu Xietong Pharmaceutical Bio-engineering Co., Ltd.), control diet (D12450J, Research Diets Inc., New Brunswick, NJ, USA) contained 10% kcal% fat, while the high fat diet contained With 60 kcal% fat (D12492, HFD, Research Diets Inc., New Brunswick, NJ, USA). DiCQAs were prepared according to our previously reported [16]. The primary antibodies against FXR (#72105), p-AKT (Ser473, #4060), AKT (pan, #4685), p-PI3K (Tyr458, #4228), PI3K (p85, #4257), p-GSK3β (Ser9,#5558), GSK3β (#12456) ,p-AMPK (Thr172, #2535), AMPKα (#5831) and IRS-1 (#2382) were purchased from Cell Signal Technology (Beverly, MA). The primary antibodies against FGF15 (ab229630), FGFR4 (ab178396), TGR5 (ab72608), MUC2 (ab272692), Occludin (ab216327) iNOS (ab178945), SREBP2 (ab30682), TNF-α (ab183218), IL-1β (ab254360) and Caludin (ab211737) were purchased from abcam (Cambridge, MA, USA). The primary antibodies against FABP6 (ER62027), SREBP1 (ER1917-19), NK-κB (p65, ET1603-12) p-ERK(Erk1 (T202) + Erk2 (T185), ET1603-22), EKR1/2 (ET1601-29),p-JNK1/2/3 (T183 + T183 + T221, ET1609-42), JNK1/2/3 (ET1601-28), p-P38 (Thr180 + Tyr182) and P38 (ET1702-65) were purchased from HUBIO (Wuhan, China). The primary antibodies against a-Tubulin (AF2827), β-actin (AA128), GLP-1R (AF6996), PPARγ (AF7797) and PEPCK (AF7689) purchased from Beyotime Institute of Biotechnology (Shanghai, China).

Animal study. (1) The animal experiments strictly adhered to the Laboratory Animal Standards of Welfare and Ethics (DB32/T-2911-2016) and received approval from the Institutional Animal Care and Use Committee at Nanjing Agricultural University (SYXK-2017-0027). Specific-pathogen-free (SPF) male mice, aged six weeks, were sourced from Hangzhou Ziyuan Laboratory Animal Technology Co., Ltd. (SCXK(Zhejiang)2019-0004, Hangzhou, China). All mice were housed within a controlled environment conforming to SPF standards. This environment encompassed a 12-hour light-dark cycle, maintained at a temperature range of 20–22°C, alongside a relative humidity level of 45 ± 5%. After a one-week acclimatization period, mice were subjected to varying dietary conditions. The first group served as the control and fed at control diet (D12450J) while the remaining three groups were fed HFD (D12492). Following a period of six weeks on their respective diets, excluding the control group, the mice underwent intraperitoneal administration of STZ at a dose of 50 mg/kg, freshly prepared and dissolved in pre-frozen citric acid buffer at a pH of 4.5, repeated three times on alternate days in a week. The control group received an equivalent volume of citrate buffer as a control. On the third and seventh day’s post-STZ injection, fasting blood glucose (FBG) levels were quantified using a Roche Accu-Check Performa blood glucose meter (Roche Diagnostics (Shanghai) Co., Ltd., Shanghai, China). Mice exhibiting FBG levels of 16.7 mmol/L (based on two times measurement) were identified as diabetic and subsequently selected for the ensuing experiment. Diabetic mice were subjected to random assignment into three distinct cohorts: the Diabetes Mellitus (DM) group, the Low-Dose diCQAs (DL) group administered at a dosage of 200 mg/kg, and the High-Dose diCQAs (DH) group administered at a dosage of 400 mg/kg. Each group consisted of eight mice, housed in pairs within separate cages. The administration was conducted once daily via gavage over a duration of six weeks. Regular assessments were made regarding body weight, food intake, and water consumption on a weekly basis. Fasting blood glucose (FBG) levels were monitored biweekly. A seven-day period prior to the conclusion of the study was dedicated to the execution of an Oral Glucose Tolerance Test (OGTT) [17]. At the conclusion of the experiment, following a 12-hour fasting period, the mice were humanely euthanized using CO₂ anesthesia. Subsequently, plasma, liver, colon, muscle and ileum tissue and fecal samples were collected kept in liquid nitrogen and preserved at -80°C until analysis.

(2) The long-term group adhered to feeding with standard chow diet. Each group consisted of 8 mice, housed in pairs within separate cages. After a one-week acclimatization period, diCQA was administered via gavage to the LND group at a dosage of 200 mg/kg, consistent with the dosage given to the DL group. The control group (LNC) received an equivalent dosage of distilled water. OGTT and blood glucose assessments were conducted in accordance with previously described methods, spanning a duration of 12 weeks, at the end of this experiments, mice were fasted overnight before being euthanized by CO₂. Subsequently, plasma, liver, colon, muscle and ileum tissue and fecal samples were collected kept in liquid nitrogen and preserved at -80°C until analysis.

AGP cohort and MAGs study. The American Gut Project (www.americangut.org) undertook sample processing, sequencing, and core amplicon data analysis. All amplicon sequence data and associated metadata have been publicly disseminated through the AGP data portal (qiita.ucsd.edu; study ID 10317) [18]. Briefly, data filtration were executed in Qiime2 [19] pipline, wherein sequences with an aggregate read count less than 5000 and an inter-sample occurrence rate below 10%, in accordance with established empirical findings bloom sequences were also removed [20] by exclude the HASH value produced during the qiime2 process. Following this, original GTDK 202 database which trained by RESCRIPt (v 2023.2.0) was applied for taxonomy assignment [21]. Within AGP metadata, individuals manifesting BMI values within the range of 18 to 25 are categorized as the "normal" group (n = 6016), while those falling within the BMI range of greater than 25 to 30 are designated as the "overweight" group (n = 2516). Individuals with a BMI surpassing 30 are classified as the "obese" group (n = 884). Moreover, a subset of 186 individuals diagnosed with diabetes mellitus are delineated as the "T2DM group."

For Metagenome-Assembled Genomes (MAGs) analysis, sequences were obtained from the NCBI or ENA database. Trimmomatic (v0.33) was employed to remove contaminated sequences, including those originating from the sequenced organism itself or adapters. Subsequently, MEGAHIT (v1.1.1) was utilized for sample-by-sample assembly, with a minimum contig length parameter set at 500 and 20 processing threads. Sequences shorter than 1.5 kilobases were excluded from further consideration.To maximize the recovery of Bins, DAS Tool (v1.1.6) was employed for binning each sample, utilizing the parameters "-l concoct,maxbin,metabat –threads 20 –score_threshold 0.8". The refinement of MAGs involved the application of dRep (v3.4.2) to mitigate redundancy within each Bin [22], resulting in the acquisition of high-quality MAGs representing individual metagenomic species to the greatest extent possible, employing an Average Nucleotide Identity (ANI) cutoff of 0.95. The specific dRep running parameters were set as follows: "-pa 0.9 - sa 0.95 –cm larger -con 10". For near-complete MAGs, the -nc parameter in dReap to 0.60. For medium-quality MAGs possessing a Quality Score (QS) greater than 50, this parameter was adjusted to 0.30 [23].A total of 1081 bins underwent redundancy removal using dRep, resulting in the retention of 615 non-redundant bins and CheckM (v1.1.2)[24] was used to determine the quality of each bin. The supplementary table S5 contains crucial data to gene quality statistics, encompassing metrics such as contig N50, contig count, average contig length, open reading frame (ORF) count, tRNA and rRNA gene counts, as well as contig read depth, following the methodology described by Xie et al [25]. Relative bin abundances in each sample were calculated using CoverM, with parameters set as: "-m relative_abundance –min-read-aligned-percent 0.75 –min-read-percent-identity 0.95 –min-covered-fraction 0". This approach, which employs stringent identity criteria and truncates alignments, serves to mitigate result inflation due to spurious mappings. For phylogenetic analysis, we used PhyloPhlAn (v3.0.3) [26] to construct a phylogenetic tree from 615 high-quality Bins. The universal markers for prokaryotes utilized in this tree were sourced from Nicola et al.'s work, subsequently, protein sequences within the MAGs were aligned using MUSCLE (v3.8.31), and then inferred using FastTree (v2.1.9). ITOL was employed to visualize the phylogenetic trees. The taxonomic annotation of all MAGs was performed utilizing GTDB-Tk (v.2.1.0), relying on the database (v202).To function annotation, For functional annotation, all protein sequences predicted by Prodigal(v2.6.3) subjected to functional annotation using the default settings of KofamScan (v.1.1.0) and EggNOG-mapper (v2.1.10). KO numbers relevant to bile acid (BA) metabolism are: Bile salt hydrolase (BSH, ko.K01442), the hydroxysteroid dehydrogenases (HSDHs) 7α-HSDH(ko.K00076), 7β-HSDH(ko.K23231), 3α-HSDH(ko.K22604), 3β-HSDH(ko.K22607), 3alpha-hydroxy BA-CoA-ester 3-dehydrogenase (baiA,ko.K15869), BA-coenzyme A ligase (baiB, ko.K15868), 3-oxocholoyl-CoA 4-desaturase (baiCD, ko.K15870), BA 7alpha-dehydratase (baiE, ko.K15872), BA CoA-transferase (baiF, ko.K15871), 3-dehydro-bile acid Delta4,6-reductase (baiN, ko.K07007), 7beta-hydroxy-3-oxochol-24-oyl-CoA 4-desaturase (baiH, ko.K15873), and BA 7beta-dehydratase (baiI, ko.K15874). The carbohydrate-active enzyme (CAZyme, http://www.cazy.org/) profile for each MAG was forecasted utilizing the CAZyme database [38], employing a Hidden Markov Model (HMM) approach facilitated by HMMER, diamond and dbCAN_sub, we selectively considered results obtained from all three tools. We applied CD-HIT (v4.8.1) to generate a Sequence Similarity Network (SSN) for a total of 514 BSH genes within MAGs [27]. The BSH homologs were processed using CD-HIT with parameters set at a 40% identity threshold[28]. BWA-MEM2 (v2.2.1) [29] is employed for determining gene abundance in Metagenome-Assembled Genomes (MAGs). Briefly, establish index for each bin to the comparison with all sequencing data. Subsequently, samtools (v1.18)[30] and ‘jgi_summarize_bam_contig_depths’ get mapped reads. Finally, R (v4.3.0) was utilized to perform the transcripts per kilobase million (TPM) calculation process.

Plasma samples analyses. Fasting plasma glucose (FPG), total cholesterol (TC), triglyceride (TG), high-density lipoprotein cholesterol (HDL-C), low-density lipoprotein cholesterol (LDL-C), blood urea nitrogen (BUN), alanine aminotransferase (ALT), and aspartate aminotransferase (AST) levels in plasma were assessed using commercially available kits obtained from Nanjing Jiancheng Bioengineering Institute (Nanjing, China).Cytokine levels including tumor necrosis factor-alpha (TNF-a), interleukin-1b (IL-1 b), IL-6, and IL-10 in plasma were quantified using enzyme-linked immunosorbent assay (ELISA) kits from NeoBioscience Technology Company (Shenzhen, China). Insulin (INS) levels in plasma were determined using an ELISA kit from Mercodia AB (Uppsala, Sweden), while glucagon-like peptide 1 (GLP-1) levels in plasma were measured with an ELISA kit from Cusabio (Wuhan, China).

Bile acids and cholesterol analyses. A variety of bile acid (BA) species were quantified based on sample type, encompassing cholesterol, cholic acid (CA), α-muricholic acid (αMCA), β-muricholic acid (βMCA), ω-muricholic acid (ωMCA), chenodeoxycholic acid (CDCA), deoxycholic acid (DCA), hyodeoxycholic acid (HDCA), ursodeoxycholic acid (UDCA), lithocholic acid (LCA), taurocholic acid (TCA), tauro-α-muricholic acid (T.α.MCA), tauro-β-muricholic acid (T.β.MCA), taurodeoxycholic acid (TDCA), taurochenodeoxycholic acid (TCDCA), TUDCA, glycochenodeoxycholic acid (GCDCA), glycodeoxycholic acid (GDCA), and glycocholic acid (GCA). Standards for these BAs, including T.α.MCA, T.β.MCA, TDCA, CA-d4, GCDCA-d4, and DCA-d4, were procured from SHANGHAI ZZBIO CO., LTD. (Shanghai, China), while the remaining BA standards were obtained from Beijing Solarbio Science & Technology Co., Ltd. (Beijing, China). BAs in plasma, liver and feces were measured according to previous report(AB/Sciex QTrap 5500 LC-MS/MS system equipped with a Kinetex Biphenyl column 2.1 × 100 mm, 2.6 um, Phenomenex) [31]. Negative ion mode was used for all analyses except cholesterol, which was analyzed using positive ions in combination with APCI source [32], and data processing was performed using SCIEX OS-Q (Include quality-control and data preprocessing, including batch-effect adjustment and missing value imputation).

RT-qPCR and westen blot analyses. The total RNA from colon tissue was extracted using FastPure Total RNA Isolation Kit (Vazyme, Nanjing, China) in accordance with the manufacturer's instructions. Subsequently, qRT-PCR was performed, following our established protocol. All results were calculated with the 2^−ΔΔCt method with the normalization by GAPDH. For western blot analysis method: Sample Preparation: Total protein from the liver, colon, and ileum was extracted using radioimmunoprecipitation assay buffer supplemented with 1 mmol/L phenylmethylsulfonyl fluoride; Protein Quantification: Protein concentration was determined using a BCA kit (Vazyme, Nanjing, China); Gel Electrophoresis: Protein extracts were separated on 10% or 12% sodium dodecyl sulfate-polyacrylamide gels at 80-110V; Transfer and Blocking: Proteins were transferred to Polyvinylidene difluoride (0.22um) filter membranes (Millipore, USA) at 0.25A; Membranes were blocked in Tris-buffered saline and 0.1% Tween 20 buffer supplemented with 5% skim milk for 2 hours. Then membranes were incubated with primary antibodies at 4°C overnight. Afterward, membranes were washed three times with TBST solution and then secondary antibody incubation 2h with HRP-linked secondary antibody anti-rabbit/mice IgG, all membranes were scanned using ECL plus solution (Affinity Biosciences, Changzhou, China), and target band density was quantified using Image J software.

16S rRNA gene sequencing of gut microbiota and Acetatifactor.spp processing. The fecal genomic DNA was extracted using the FastDNA SPIN Kit (MP Biomedicals, Santa Ana, CA, USA). Subsequently, the 16S rRNA gene's V3-V4 hypervariable regions were targeted for amplification with the primers 341F (5’-CTACGGGNGGCWGCAG-3’) and 805R (5’-GACTACHVGGGTATCTAATCC-3’), followed by sequencing on an Illumina NovaSeq 6000 sequencer. All analyses were conducted using the QIIME2 (v2023.5) pipeline [19]. After quality control, 230 bp of the paired ends were retained. Dada2 was employed to generate representative sequences and record HASH values for subsequent sequence extraction. Diversity was calculated using q2-diversity, retaining ‘Unweighted UniFrac’ for further analyses. Taxonomic classification (sklearn v 0.24.0) was conducted using the same database as AGP taxonomy.

A total of 152 sequences from Acetatifactor.spp were download from NCBI, in addition to 7 sequences from MAGs for further analysis. The phylogenic tree and functional annotation results were acquired using the same methodology as described previously for MAGs. The 16S rRNA sequences were isolated using Barrnap (v0.9), followed by ‘makeblastdb’ in BlastN (v2.1.1), then comparison against the Acetatifactor.spp reprsent 16S rRNA sequences[33].

The Co-Abundance Group (CAG). Algorithm according to Liu et al. [34]: The OTUs that demonstrated a minimum presence of 20% across all samples, coupled with a collective abundance exceeding 0.01%, were selected. The SparCC algorithm [35], iterated over 100 bootstrap procedures, was applied to ascertain correlations among these selected OTUs. Subsequently, only OTUs exhibiting correlation scores surpassing 0.4 and a significance threshold of P < 0.05 were delineated as Co-Abundance Groups (CAGs, n = 123). The correlation values underwent transformation into a correlation distance (1-correlation value), and OTUs were subjected to clustering employing the Ward clustering algorithm implemented in the R package WGCNA (v1.72-1). The resultant CAG network was visually represented using Cytoscape (v 3.2.1).

Statistical analysis. Multiple-regression analysis by forward selection and single regression analysis was used to identify the gut microbiota associated with BMI and T2DM in AGP corhort (linear models in MaAsLin2 (v1.7.3))[36]. The output from the QIIME2 pipeline in Biom table format was imported into R (v4.3.0) (https://www.R-project.org/) for analysis. Alpha-diversity indices were calculated using the estimate_richness function from the R package ‘phyloseq’. For beta-diversity, the Bray–Curtis distance of genus-level data was computed using the vegdist function from the R package ‘vegan’. Principal Coordinate Analysis (PCoA) was conducted with the dudi.pco function from the R package ‘ade4’. PERMANOVA was employed to assess the statistical significance between groups, utilizing 9999 permutations. For the determination of the variation attributed to each of the collected host factors in diabetic groups, we conducted an Adonis test within R. Unless otherwise specified, statistical significance to lipid and glucose metabolism-related parameters and pro-inflammatory mediators in plasma, BAs and gut microbiota were conducted by using Mann-Whitney U and Kruskal-Wallis, respectively (rstatix v0.7.2), employing a Benjamini-Hochberg (BH) adjustment for multiple comparisons, with a significance threshold of p < 0.05. Functional gene differences analysis in MAGs were executed through DESeq2 (v1.40.2), with criteria set at log₂FoldChange > 1 and an adjusted p-value (padj) < 0.05. For the assessment of correlations between BAs and CAGs, the psych package (v2.3.6) was employed, utilizing the Spearman correlation method, and applying Benjamini-Hochberg adjustment for multiple testing (padj < 0.05). All results were visualized using ggplot2 (v3.4.2) and Adobe Illustrator 2020.

The AGP (American Gut Project) cohort study uncover gut microbiota associated with diabetes and obese

We downloaded a total of 13,422 data entries from the AGP database and subsequently filtered the data (reads ≥ 5000), resulting in 9,647 data entries. These included 6,016 entries in the normal weight group (body mass index 18<[BMI] ≤ 25), 2,561 in the overweight group (25 < BMI ≤ 30), 884 in the obese group (BMI > 30), and 186 in the diabetic group (Table S1). Consistent with other studies [37, 38], the fecal microbiota in all groups was dominated by three families (Fig. 1A): Bacteroidaceae (32.7% ± 1.8), Lachnospiraceae (18.8% ± 1.2) and Ruminococcaceae (12.9% ± 0.8). Our investigation into the association of BMI data in the AGP dataset to explore alterations in the gut microbiota identified 26 family-level associations with BMI (Fig. 1B, Table S2). Among these, 14 exhibited relative abundances that were positively correlated with BMI, whereas 12 exhibited relative abundances that were negatively correlated with BMI. Similarly, when analyzing patients with diabetes, we found that 9 family-level taxa exhibited relative abundances positively correlated with T2DM, whereas 19 exhibited relative abundances negatively correlated with diabetes (Fig. 1B, Table S3) and these correlations were independent of age and sex (Fig. S1). In particular, we observed increased relative abundance at Family level of Anaerotignaceae, Bacteroidaceae, Desulfovibrionaceae, Erysipelotrichaceae, Ruminococcaceae, and Streptococcaceae increased in obesity and diabetes, and Lachnospiraceae exhibited an extremely significant increase in the overweight and obese groups (Fig. 1C, Fig. S2, p < 0.01) but a slight increase in the diabetic group (Fig. S2, p = 0.118). Further analysis at the genus level (Fig. S3) revealed that Acetatifactor which belongs to Lachnospiraceae exhibited a highly significant increase in relative abundance in both diabetic and obese groups (p < 0.01). Alistipes exhibited a significant decrease in relative abundance in the obese group (p < 0.01), with a minor decrease in the diabetic group (p = 0.43). Lastly, Akkermansia displayed a highly significant decrease in relative abundance in the obese group (p < 0.005) and a slight decrease in the diabetic group (p = 0.318). Although several studies have implicated the involvement of Acetatifactor in the development of obesity and T2DM, little definitive information is available [17, 39]. Therefore, we assessed the effect of Acetatifactor on obesity and T2DM via evaluating the odds ratio and revealed that the Acetatifactor abundance was inversely correlated to the ORs of obesity and T2DM (Fig. 1D, Table S2). Due to Acetatifactor high discriminant value for obesity and T2DM in the AGP cohort analysis, we subsequently focused this study on the role of Lachnospiraceae and Acetatifactor in the control of diabetes.

Metagenome-assembled genomes (MAGs) analysis revealed the gut microbiota associated with BMI and T2DM in the AGP cohort are involved in BA transformation pathways

We utilized MAGs analysis against gut microbiota associated with BMI and T2DM in the AGP cohort. Utilizing extensive data totaling 1.1 Tb, we reconstructed 615 high-quality MAGs, each of these MAGs demonstrated a completeness exceeding 70%, with a contamination rate of less than 10%, as evaluated by CheckM (Fig. S4A-D, Table S5). All 615 MAGs were primarily represented at the family levels, Lachnospiraceae (169), Acutalibacteraceae (75), Ruminococcaceae (70), Bacteroidaceae (43), Oscillospiraceae (20) and Rikenellaceae (14) and revealed that total 404 of MAGs contained the enzyme bile salt hydrolase (BSH, EC: 3.5.1.24), capable of hydrolyzing conjugated BAs (Fig. 2A). BSH genes were found to be widely distributed among gut microbiota, including Lachnospiraceae (73.9%), Acutalibacteraceae (77.3%), Ruminococcaceae (77.1%), Bacteroidaceae (69.8%), Rikenellaceae (92.9%), and Oscillospiraceae (70.0%) (Table S6). Furthermore, the Enterobacteriaceae exhibited richness in K00076 (15, 78.9%) encoding 7-alpha-hydroxysteroid dehydrogenase (7α-HSDH), while MAGs alse carried the baiN (490, 79.8%), responsible for the sequential reduction of C6-C7 and C4-C5 after dehydroxylation which is essential for BA conversion, with Lachnospiraceae being particularly abundant in baiN (169, 34.5%). This collectively suggests that MAGs possess the capability to transform primary BAs into secondary BAs. To investigate the prevalence of regulatory genes in BA metabolism, we conducted taxonomic and functional annotation of high-quality genes obtained from representative species (mouse, n = 1,689; dairy cows, n = 718, human = 615) and BSH accounted for 65.9%, 55.7% and 51.4% of human, mouse and dairy cows MAGs, respectively (Fig. 2B). The results demonstrated widespread presence of BA metabolism-related genes capable of hydrolyzing conjugated BAs in humans, mice, and dairy cows. Specifically, BSH and 7α-HSDH exhibited the highest distribution. The MAGs of humans, mice, and dairy cows exhibit the presence of BSH genes predominantly within certain families (Fig. 2C): Lachnospiraceae (32.5%), Muribaculaceae (12.2%), Acutalibacteraceae (9.6%), Oscillospiraceae (6.9%), Ruminococcaceae (5.8%), Bacteroidaceae (5.3%), and Rikenellaceae (4.2%).

The classification and variation patterns of BSHs in gut microbiota reveal that BSH-carrying MAGs exhibit differences in carbon source utilization

To further analyze the homology between BSH genes, we clustered all BSH gene sequences using CD-HIT, establishing a protein sequence similarity network (SSN, Fig. 3A)[27, 40]. SSH builds were constructed with a sequence identity threshold of over 40% between protein sequences. The threshold setting allowed classification of 75.9% BSH homologs into 7 clusters (Fig. 3B, Table S6): Cluster 1 contained four BSH genes, all from species Bacteroides uniformis; Cluster2 included 34 BSH genes from MAGs, predominantly from the Bacteroidaceae family (41.2%), with major contributions from genera Bacteroides (29.4%), Tannerellaceae (17.6%), specifically sepecies Parabacteroides distasonis, and Rikenellaceae (17.6%), primarily represented by genera Alistipes (8.8%); Cluster3 contained 56 BSH genes, primarily from Bacteroidaceae (46.4%), with major contributions from genera Bacteroides (41.1%), Tannerellaceae (23.2%) with major contributions from primarily species Parabacteroides distasonis (19.6%), and Rikenellaceae (14.3%) with major contributions from genera Alistipes (10.7%); Cluster 4 included 283 BSH genes, primarily from the Lachnospiraceae family (46.2%), predominantly genera Agathobacter (19.9%) and Acetatifactor (3.4%), Ruminococcaceae (18.4%), primarily genera Ruminococcus_D (6.7%), and Acutaliococcaceae (13.2%), primarily genera Ruminococcus_E (6.7%); Cluster5 contained four BSH genes, primarily from the family of Lachnospiraceae (50.0%); Cluster6 included 25 BSH genes, primarily from the family of Lachnospiraceae (68.0%), and primarily represented by genera Bifidobacteriaceae (40%); Cluster7 comprised four BSH genes, all from species Fimenecus sp000432435. Among these, 74 signal peptides (Fig. 3C, Table S7) were predicted, consisting of 66 Sec signal peptides (Sec/SPI) and 8 lipoprotein signal peptides (Sec/SPII). These signal peptides are responsible for controlling protein secretion and translocation via short N-terminal amino acids in cluster2/3 MAGs [41]. Additionally, the majority of Sec/SPII peptides originated from Bacteroidaceae within Cluster2. Volcano plot (Fig. 3D) indicated significant differences in functional genes related to energy metabolism and carbohydrate transformation and utilization among all clusters. It's worth noting that sialidase (K01186, sialidase-1 [EC: 3.2.1.18]) hydrolyzes MUC2, a highly O-glycosylated polysaccharide in the intestinal whose terminal position is occupied by monosaccharide sialic acid and can be cleaved by sialidase[42]. The differential analysis of the gene encoding the starch-bound outer membrane protein SusE/F (K21572) revealed significant increases in lysosomal acid hydrolytic glycoside genes, namely K01186, K01206, and K12373 which can release the different components of mucin glycans[43]. This increase was accompanied by upregulation of genes such as K00688 (glycogen phosphorylase [EC: 2.4.1.1]), K00975 (glucose-1-phosphate adenylyltransferase [EC: 2.7.7.27]), and K01223 (6-phospho-beta-glucosidase [EC: 3.2.1.86]), suggesting variations in glucose metabolism. Furthermore, the annotation results from CAZy database demonstrate distinctions in carbon source utilization, indicating significant increases in genes related to MUC2 hydrolysis, specifically GH33, GH2, GH29, GH35, GH101, and GH109 (Table S8)[44, 45]. This implies that certain BSH-carrying MAGs possess the capability to directly utilize MUC2 as an energy source.

The BSH-MAGs associated with the AGP cohort exhibit functional traits that demonstrate their capability to directly utilize MUC2 as an energy source

Further analysis of functional gene differences in Lachnospiraceae carrying (BSH-MAGs) and those lacking BSH MAGs (non-BSH-MAGs), several noteworthy distinctions were identified (Fig. S5A). These encompassed genes included K09691 [EC: 7.5.2.14], K10188, K10190, K10542 [EC: 7.5.2.11], K10823, K15771, K01191 [EC: 3.2.1.24], K01192 [EC: 3.2.1.25], K01439, and K01176 [EC: 3.2.1.1]. These genes belong to lactose/L-arabinose transport systems (ABC transporters) permeability proteins. Notably, K10188 and K10190 were significantly reduced in non-BSH-MAGs compared to BSH-MAGs carrying BSH-MAGs (log₂FC > 1, padj < 0.05). This suggests that Lachnospiraceae in non-BSH-MAGs exhibit lower external energy utilization ability, especially for lactose/L-arabinose, than BSH-MAGs. Similar trends were observed for glycoside hydrolases (GHs) (log2 FC > 1, padj < 0.05, Fig. S5A), indicating differences in carbon source utilization. Importantly, GH33, a key gene for hydrolyzing MUC2, was not expressed in non-BSH-MAGs, mainly in Acetatifactor intestinalis and Acetatifactor sp003447295. Changes in GHs such including GH27 (a-galactosidase (EC 3.2.1.22) and GH112 (lacto-N-biose phosphorylase or galacto-N-biose phosphorylase (EC 2.4.1.211) clearly demonstrated that Acetatifactor. Spp, which partially carries GH33, can not only hydrolyzed conjugated BAs, but also use MUC2 as a carbon source and disrupt the intestinal barrier (Fig. S5C). Further analysis was performed for Acutalibacteraceae, Ruminococcaceae, Bacteroidaceae, Rikenellaceae, and Oscillospiraceae (log₂ FC > 1, padj < 0.05). Faecalibacterium and RuminococcusH without the BSH gene were examined for lysosome-related genes K01186, K01206, and K12373. However, MUC2 utilizes the key enzyme GH33 primarily in Bacteroides (45.7%), with Alistipes (8.6%) also annotated (Fig. S5B). Furthermore, Bacteroides and Alistipes were found to contain GH20, GH35, GH2, GH29, GH101, and GH109, necessary for hydrolyzing MUC2 (Fig. S5D). Combined with differential gene abundance results (Fig. S5), Acetatifactor, Bacteroides and Alistipes are confirmed to utilize intestinal mucin MUC2 as a carbon source.

Comparisons of the sequences and structures of BSHs

To assess the differences in BSH hydrolytic activity for conjugated BAs, we referred to three reported BSH sequences to examine the homology of BSH sequences with active sites of BSH hydrolyzing conjugated BAs in clusters and their hydrolytic ability. BSH sequences such as 5HKE_1 (49.3%, Lactobacillus salivarius), 2HF0_1 (41.14%, Bifidobacterium longum), and 4WL3_1 (59.8%, Enterococcus Feacalis) [28] exhibited homology to the representative sequence Agathobacter rectalis in cluster4 (Table S6). Cluster2 (Bacteroides eggerthii, 377aa), cluster3 (Morganella morganii, 372aa), cluster4 (Agathobacter rectalis, 335aa), and cluster6 (Collinsella sp900545445, 321aa) were selected as representative sequences (Fig. 4A). Alignment of the representative sequence BSH amino acids revealed the highest homology of 47.5% (comparison of BSHs from Collinsella sp900545445 and Agathobacter rectalis) and the lowest homology of 33.8% (comparison of BSHs from Morganella morganii to others). The secondary structure of these BSHs (Hydrogen bondingpattern between amino hydrogen and carboxyl oxygen molecules in the peptide backbone) suggests that despite differences in secondary topology (Fig. 4B), they share core active sites with known BSH genes [28] (Fig. 4C), including Cys2/2/2/2, Asp11, ARG18/18/18/18, Asp21/21/21, Trp22/22/27, Asn78/78/78/78, Asn168/168/168/168, Thr193/193/193, Asn194/194/194, and Arg248/248/248/248, capable of hydrolyzing bound BAs. To compare the hydrolytic capacity of representative BSH, we selected the representative sequence BSH using homology modeling by AlphaFold2 for molecular docking (Fig. 4C, Fig S6). The homology modeling structure in cluster4, which exhibited the strongest hydrolytic capacity for conjugated BAs, was characterized by lower binding energy in molecular docking, indicating amore stable complex, and the extraction of residues in contact with the substrate from the complex (Fig. S6A-B). These findings highlight differences in sequence, protein structure, and hydrolytic conjugated BA ability of the MAGs within different clusters. Additionally, considering the increased relative abundance of Lachnospiraceae in AGP corhort data and the stronger hydrolytic binding BA ability of BSH (Cluster4) in Lachnospiraceae compared to Bacteroidaceae and Rikenellaceae in Cluster2/3, it suggests that regulating the relative abundance of Lachnospiraceae could affect the composition of the BA pool. It is essential to note that Acetatifactor spp. containing BSH and GH33 have the potential to utilize MUC2 as a carbon source, potentially disruption the intestinal barrier.

Effects of diCQAs alleviate on physiological indicators in HFD/STZ-induced diabetic mice but not in long-term group

Results from the long-term gavage group (total 14 weeks, Fig S7A) indicated that there were no significant differences in fat content, blood glucose levels, and inflammatory factors between the LNC and LND groups (200 mg/kg). This suggests that the long-term regulation of diCQAs did not directly promote glucose and lipid metabolism in mice and had no side effects (Fig. 5A). In the high-fat diet (HFD) + STZ induced diabetic mice group (total 14 weeks, Fig S7A), diCQAs were administered via gavage at two different doses (DL, 200 mg/kg; DH, 400 mg/kg) in an attempt to uncover potential dosage effects. Interestingly, only the DL group, but not the DH group, exhibited a significant reduction in plasma fasting blood glucose levels compared to the DM group, which displayed symptoms typical of hyperglycemia in diabetes mellitus (Fig. 5B, p < 0.05). The liver index, and epididymal fat in mice were also significantly lower than those in the DM group (p < 0.05), with DL having a better effect than DH (Fig. S7). Additionally, DL group, but not DH group, resulted in significantly lower in HbA1C levels in the plasma compared to DM group, confirming the continuous blood glucose-regulating function (Fig. 5B, p < 0.05). The OGTT showed that the DL group demonstrated improved glucose tolerance compared to the DM group (Fig. 5B, p < 0.05), while the DH group exhibited no significant changes in OGTT. However, it was noted that insulin levels in the plasma of the DM, DL, and DH groups were significantly lower than it in NC group after STZ treatment, indicating that STZ had damaged islet B cells [46], resulting in hyperglycemia. Moreover, both does of diCQAs did not significantly restore B cell function. Nevertheless, GLP-1 levels, associated with glycemic regulation, significantly increased in the DL group compared to the DM and NC groups (Fig. 5B, p < 0.05), suggests that the potential mechanism by which the DL group exerts hypoglycemic effects is associated with gut microbiota. Compared with NC group, the DM group exhibited significantly increased levels of TC, TG, HDL-C, and LDL-C in plasma, all of which were effectively alleviated by both doses of diCQAs (Fig. 5B, p < 0.05). Additionally, chronic inflammation, another indicator of metabolic syndromes such as obesity and diabetes [47], was observed in DM group. Both doses of diCQAs (Fig. 5B, p < 0.05) were effective in reducing the expression of inflammatory factors TNF-α, IL-1β, and IL-6, while increasing the levels of IL-10, which may be attributed to the anti-inflammatory properties of diCQAs [48]. These findings suggest that DL group has the potential to reduce inflammation and alleviate hyperglycemia induced by HFD + STZ more effectively than the DH group.

DiCQAs promoted conjugated BAs levels in the enterohepatic circulation in diabetic group but slightily in long-term group

Analysis of total BAs in the enterohepatic system for both the long-term and diabetic groups revealed distinct patterns (Fig. S8A-C). Specifically, the concentration of conjugated BAs was found to be higher in the liver compared to plasma and feces. This suggests that conjugated BAs are synthesized within the liver, exocytosed into the intestine, and subsequently hydrolyzed by the gut microbiota, ultimately leading to the formation of secondary BAs [49]. Overall, in the LND and LNC groups, no significant trends in BA levels were observed in plasma and feces. However, a noteworthy observation was the significantly lower BA content in the liver of the LND group (Fig. S8C). This suggests that long-term regulation by diCQAs may promote BA exocytosis. Furthermore, due to the influence of the HFD + STZ modeling altered the gut microbiota, the DM group exhibited significant alterations in total BA, primary/secondary BA, and conjugated/unconjugated BA content in both plasma and feces relative to the NC group (p < 0.05). These changes disrupted the BA pools, which were subsequently ameliorated by diCQAs modulation, resulting in significant reductions in plasma levels of total BA, primary/secondary BA, conjugated/unconjugated BA, etc. (Fig. S8B, p < 0.05).Additionally, it was observed that diCQAs modulation mitigated the slight accumulation of total BAs in the liver (Fig. S8C, p = 0.164/0.215). Consistent with the results of total BA changes, PERMANOVAR analyses showed that there was no significant difference between LNC and LND in feces and plasma, but the content and proportion of BA in DM were significantly different from those in feces and plasma in the NC and DL groups (Fig. S8E-F, p < 0.05). Taken together, these results demonstrate that diCQAs play a role in restoring the BA pool by regulation of gut microbiota perturbed by HFD + SZT modeling.

Analysis of individual BAs revealed that taurine-conjugated CA and bMCA (TCA and T.b.MCA, respectively) were the most common conjugated BAs in both of long-term and diabetic groups (Fig. 6). Taurine-conjugated BAs were much higher than glycocholate-conjugated BAs showing differences in BA composition between mice and human [50]. Fecal BAs results in long-term and diabetic groups showed that conjugated BA was hydrolyzed and converted by gut microbiota to secondary BA, while there was no significant change in fecal DCA and lithocholic acid (LCA) content compared to the LNC and LND groups, but a significant elevation was observed in the DL group compared to the DM group (Fig. 6A, p < 0.05). The plasma BAs content closely resembled that of ileum BAs content, particularly in terms of conjugated BA like T.β.MCA, which exhibited a significant increase in the DL group compared to the DM group (Fig. 6B, p < 0.05). Notably, no significant change was observed in the long-term group. T.β.MCA, functioning as a natural antagonist of FXR, demonstrated its ability to inhibit the FXR-FGF15 signaling axis in the ileum, consequently leading to an upregulation of liver BA synthesis [14]. DCA and CDCA, known activators of TGR5 in the liver [51], experienced a significant increase in the DL group compared to the DM group (Fig. 6C, p < 0.05). The activation of TGR5 was associated with an enhancement in glucose metabolism and energy expenditure [52]. Furthermore, both plasma and liver cholesterol levels (Fig. S9) showed no significant change in the long-term group, whereas in the DL group, a notable decrease relative to the DM group was observed. This alteration may be linked to the heightened BAs synthesis in DL group, which increased T.β.MCA content in the ileum that activated the FXR-FGF15 signaling axis in the ileum.

Aggregating the measurements across the entire enterohepatic system, we demonstrated a reduction in the total content of conjugated BAs following hydrolysis by gut microbiota. In the DL group, there was a conditioned increase in the content of T.β.MCA compared to the DM group. Concurrently, cholesterol contents in both liver and plasma were concurrently reduced. These results indicate that diCQAs regulation in the DL group, at a dosage of 200mg/kg, plays a crucial role in alleviating hyperglycemic symptoms in mice by elevating T.β.MCA content through the modulation of gut microbiota.

Taxonomic analysis showed DL group more efficient than DH group in alleviating disorded gut microbiota in diabetic group.

We assessed changes in gut microbiota within the long-term and diabetic groups through sequencing of the fecal 16S rRNA V3–V4 region. No significant differences were observed in richness and evenness (Alpha diversity) in gut microbiota in both long-term and diabetic groups (Fig. S10A-B). PCoA of weighted pairwise Bray–Curtis distances indicated distinct clustering, with the LND group separate from the LNC group (Fig. S11A), while the microbiota composition of the NC group exhibited significant disparities from that of the others in diabetic groups, attributed to the perturbation caused by HFD + STZ modeling (Fig. S12A). This implies that the DL group exhibited a more effectiveness in regulating disrupted gut microbiota compared to the DH group. The top 15 most abundant families showed that the relative abundance of Rikenellaceae was significantly increased in LND and DL relative to the LNC and DM groups (Fig S11C, Fig S12C, P < 0.05), respectively, while, Lachnospiraceae exhibited significantly lower relative abundance in the DL group compared to the DM group (Fig S12C, P < 0.05). The Linear discriminant analysis effect size (LEfSe) result indicated a decrease in the relative abundance of genera Dubosiella and Desulfovibrio, while genera Alistipes, Akkermansia, and Ileibacterium showed an increase in the LND group (Fig S13); In the diabetic group, LefSe results indicated that the relative abundance of genera Acetatifactor, Anaerotignum, Erysipelatoclostridium, and Angelakisella increased in DM group compared to the other groups (Fig S13) and reports indicated that were significantly increased in HFD-induced obese/diabetic mice [53, 54]. In summary, DL group show more efficient than DH group in alleviating disorded gut microbiota in diabetic group.

DL reduced the relative abundance of BSH-carrying bacteria which disordered in gut microbiota in diabetic mice but not in long-term group

We explored the correlation between changes in gut microbiota and host glycemic indicators and BA profiles using Adonis [34]. Notably, significant indicators were correlated with variations in gut microbes based on unweighted UniFrac distances between samples (P < 0.05, Fig. 7A), which consistency with findings in most studies [12, 55]. Given the pivotal role of bacteria in gut ecosystems and BA pool modification [56], a co-abundance network was constructed based on the SparCC correlation coefficient. This network clustered operational taxonomic units (OUTs) that were present in at least 60% of samples or had a relative abundance greater than 0.01% into eight CAGs (OTU = 122, Fig. 6B). Within these CAGs (n = 8), relativate abundance of CAG-3 is significantly positively correlated with FBG and ALT in plasma. Conversely, it exhibited a negative correlation with BAs content, such as T.β.MCA in plasma (Fig. 6C). Moreover, CAG-6 (OTU = 12) showed a significant decrease in the DM group (p < 0.05), whereas CAG-3 (OTU = 39) exhibited a significant increase; CAG-3 was predominantly composed of Lachnospiraceae in family level (OTU = 16, Fig. 6D and Fig. S12C), and the relative abundance of the Acetatifactor increased significantly in the DM group at the genus level (Fig. 7E, Fig S14A) and these results were also consistent with AGP cohort anslysis (Fig. 1D). This increase was substantially reversed in DL group, with the most notable changes observed in the species Acetatifactor sp011959105 and Acetatifactor muris (Fig. 7F, Fig S14B). These findings suggest that DL may improve the BA pool and alleviate hyperglycemia and related metabolic syndromes by regulating gut microbiota in CAG-3. In addition, significant changes were observed in six genera and three species levels in the long-term group (Fig. S15A), as well as in five genera and six species levels in the NC and DL groups compared to the DM group (Fig. S15B), but only genera Alistipes, along with the species Alistipes sp003979135, exhibited consistent regulatory outcomes as observed in the LND from the long-term group. This suggests that the diCQAs (200mg/kg) could lead to an increased relative abundance of genera Alistipes, particularly Alistipes sp003979135 and decreased Acetatifactor sp011959105 and Acetatifactor muris in CAG-3.

Acetatifactor carries BSH and GH33 genes, enabling it to hydrolyze conjugated BAs and utilize MUC2 as a carbon source (Fig. S5). To further elucidate the functionality of Acetatifactor sp011959105 and Acetatifactor muris, we downloaded a total of 151 Acetatifactor spp. sequences from NCBI (Table S11) along with an additional seven Acetatifactor MAGs. The representative 16S rRNA sequences representing Acetatifactor muris, Acetatifactor sp011959105, and Acetatifactor sp003612485 were annotated and extracted, all of which were identified through BlastN alignment [33]. Acetatifactor muris was found to be most closely related to reference genome GCF_024623325.1, whereas Acetatifactor sp011959105 and Acetatifactor sp003612485 were most closely related to reference genomes GCF_011959105.1_ASM1195910 and GCF_949515085.1 (Fig S16). The genomic sequences of GCF_024623325.1 and GCF_011959105.1_ASM1195910 of Acetatifactor muris and Acetatifactor sp011959105 were found to contain genes such as GH33, GH2, GH20, GH29, GH35, GH42, and GH109 that hydrolyze MUC2 as a carbon source. These sequences also exhibited active sites similar to BSH, including Cys2/2/2, Asp11/11/11, AGR16/16/16, Asp21/21/21, Asn79/79/79, Asn182/182/182, and Asn194/194/194 (Fig. S17A-B). Molecular docking results (Fig. S17C) revealed their capability to hydrolyze conjugated BAs. In summary, both Acetatifactor muris and Acetatifactor sp011959105 have the capacity to hydrolyze conjugated BAs and utilize MUC2 as a carbon source lead to intestinal barrier disrption. In summary, CAG-3, which correlates with host glycemic indicators, show a significant reduction in the relative abundance of Acetatifactor muris and Acetatifactor sp011959105 in the DL group. This effectively reinstated the disrupted BA pool, leading to a restoration of the increased content of conjugated BAs, notably T.β.MCA.

DL regulate inhibit ileal FXR-FGF15 signaling axis and enhances the intestinal barrier

In both the long-term and diabetic groups, the gut microbiota exhibited a notable presence of genes, including the BSH gene capable of hydrolyzing conjugated BAs, as well as the GH33 gene responsible for MUC2 hydrolysis, which leads to the disruption of the intestinal barrier. Within the long-term group, Acetatifactor carrying both BSH and GH33 genes remained relatively stable, with relative abundance no significant changes observed. Conversely, in the diabetic group, the relative abundance of Acetatifactor, particularly in the DM group, was significantly higher than in the NC and DL groups. This led to a significant decaresed in T.β.MAC content in the plasma of the DM group (Fig. 5B). Therefore, we assessed markers proteins related to intestinal barrier (MUC2, Occludin and Claudin) and BA signaling in the intestine (FXR-FGF15). Notably, in the long-term group, the levels of tight junction proteins, including occludin, claudin1, and MUC2, were significantly higher in the LND group than in the LNC group. This suggests that diCQA (200mg/kg) gavage promoted the increase of intestinal mucins and tight junction proteins (Fig. 8A-B). The disrupted gut microbiota in the HFD resulted in compromised gut permeability in DM group, which was subsequently restored in DL group (p < 0.05, Fig. 8B). As mentioned above, the HFD-induced disturbance of the gut microbiota led to significant changes in the BA pool. After regulation with DiCQAs, the trend of BA changes was reversed, with DL concentration proving more effective than DH concentration (Fig. 6). It is noteworthy that BA content in plasma reflected BA content in the ileum. In the long-term group, no significant trend in conjugated BA content was observed in the LND group compared to the LNC group in ileum. Consequently, there was no significant difference in the expression of the BA receptor protein FXR in the LDN group (Fig. 8C), and no notable change in the expression of the FGF15-FGFR4 signaling axis protein in ileum. This indicates that long-term exposure to diCQAs did not directly increase the expression of FXR-FGF15 signaling axis in the ileum. Conversely, in the diabetic group, gut microbiota disturbance resulted in a significant decrease in BA content, particularly conjugated BA content in the plasma of the DM group compared to the NC and DL groups. As a result, the content of T.b.MCA, an FXR antagonist, was significantly increased in the NC and DL groups than in the DM group. Consequently, the protein content of the FXR-FGF15 signaling axis in the ileum of the DM group was significantly higher than that in the NC and DL groups (Fig. 8D). The inactivation of the ileum FXR-FGF15 signaling axis in DL group appears to be associated with the regulation of gut microbiota, leading to an increase in T.β.MCA content, rather than exerted a direct stimulatory effect. Furthermore, these findings have prompted an investigation into the enterohepatic FXR-FGF15-FGFR4 signaling axis and its effects on BA synthesis and glucolipid metabolism in the liver.

The mechanisms through which DL group improve liver BA synthesis and regulate inflammation and energy metabolism in the diabetic group involve the enterohepatic FXR-FGF15 signaling axis

Ileal epithelial cells secrete FGF15 into circulating, where it binds to the liver receptor FGFR4, thereby inhibiting BA synthase CYP7A1 [57]. In the long-term group, although no significant changes were observed in BA synthesis-related proteins, a notable decrease in total BA content was evident in the liver of the LND group compared to the LNC group. This reduction was primarily attributed to a significant increase in the expression of the BA efflux protein BSEP in the LND group (The bile salt export pump, Fig. 8A), suggesting that diCQA at 200mg/kg regulation could enhance BA efflux. As for the mechanisms of glucolipid metabolism and inflammation, no significant differences were detected in proteins related to glucose and lipid metabolism between the LND group and the LNC group (Fig. 10A, Fig. 11A). Similar observations were made in muscle tissue, with no significant changes in GLP-1R, GPR43, p-AKT/p-GSK3b and p-AMPK protein levels, except for p-PI3K (p < 0.05, Fig. S19A). In the ileum of DM group, the activation of the FXR-FGF15 axis led to a significant increase hormones FGF15, which secreted into the portal vein, circulates to its receptor, FGFR4, in the liver and acts to reduce liver BA synthesis via suppression of BA synthetic enzymes CYP7A1 [58]. Liver FGFR4 protein expression was found to decrease in parallel with the decreased ileal FGF15 protein expression and serum T.β.MCA levels in DLgroup (Fig. 9B). Additionally, the protein expression levels for liver BA synthetic proteins, primarily CYP7A1, were upregulated in the DL group compared to DM group (Fig. 9B). The observed increase in BA conjugation in plasma was correlated with an upregulation in the BA conjugation enzyme in DL group, BAAT, facilitating the conjugation of primary BAs with glycine or taurine (Fig. S21B). Taken together, these results indicate that the decreased in T.β.MCA levels in the ileum activation the ileal FXR-FGF15 axis heightened secretion of hormone FGF15 leads to a increased protein expression of FGFR4 in the liver of DM group than DL group, decreased BA synthetic enzyme proteins particularly CYP7A1 than DL group.

The biosynthesis of BAs is a central metabolic pathway for cholesterol, crucial in regulating cholesterol levels within the liver and serum. This process is pivotal for maintaining overall cholesterol homeostasis. About 5% of the total BAs, which are excreted through feces. This loss of BAs is compensated by de novo BA synthesis in the liver. The DL group exhibited a slightly higher excretion of total BAs in feces compared to the DM group. Moreover, total BA levels in the plasma were notably higher in the DL group as opposed to the DM group. Interestingly, the total BA content in the liver was slightly lower in the DL group, potentially attributed to the heightened expression of BSEP in DL group (Fig. 9B, Fig. S8). Cholesterol circulating in the bloodstream re-enters the liver via reverse cholesterol transporters, contributing to the BA synthesis process. Notably, in mice subjected to DL group, the expression of HMGR protein, a key enzyme in de novo cholesterol biosynthesis, was downregulated compared to DM group (Fig. 9B). Additionally, reverse cholesterol transporters ABCA1 and ABCG1 exhibited high expression levels in the liver (Fig. S20B). These findings suggest that diCQAs consumption reduce both liver cholesterol biosynthesis and reuptake, augmenting de novo BA synthesis, and enhancing BA efflux. The combined impact of these effect leads to a reduction in liver and serum cholesterol levels in DL group.

In the diabetic group, STZ-induced destruction of mouse islet B cells resulted in significantly lower insulin levels in the plasma of the DM, DL, and DH groups compared to the NC group. This led to a noticeable increase in IRS-1 protein levels in the liver of the NC group in comparison to the DM group. Notably, DiCQA regulation did not induce significant changes in IRS-1 levels (Fig. 10B). Similarly, GLP-1R and GPR43 protein levels remained relatively stable between the DM and DL groups (Fig. 10B). Activation of the AKT through TGR5 promoted the phosphorylation of the liver glycogen synthesis protein GSK3β, resulting in increased liver glycogen synthesis and reduced blood glucose levels [59]. In the DL group, TGR5, p-AKT, and p-GSK3β were significantly upregulated compared to the DM group (Fig. 10B). Additionally, p-AMPKα, SREBP2, and PCG-1α in the liver exhibited significant increases in the DL group compared to the DM group, enhancing glucolipid metabolism levels. In muscle tissue, p-PI3K/p-AKT/p-GSK3β, p-AMPKα protein levels, and PPARγ and GS, were significantly higher in the DL group than in the DM group (Fig. S19B). This improvement in glucolipid metabolism contributed to the decrease in blood glucose levels observed in the DL group. Meanwhile, in the long-term group, it was observed that diCQA did not directly activate p-PI3K/p-AKT/p-GSK3β, p-AMPKα to improve glucolipid metabolism, whether in the liver or muscle (Fig. 10A, Fig. S19A). FXR-FGF15 in the ileum causes signaling cascades activated by the liver FGFR4 receptor in addition to inhibiting liver BA synthesis; therefore, it can also induce the activation of the MAPK/NF-kB inflammatory signaling pathway [60]. In the long-term group, there were no significant differences observed in iNOS. Only p-ERK showed a significant increase in the MAPK signaling pathway in the liver. However, p-JNK, p-p38, and inflammatory cytokines i-NOS, TNF-α, and P65 remained unchanged (Fig. 10A), suggesting that long-term gavage of diCQAs had no significant effect on inflammation in mice, so that, the direct effects of diCQAs on reducing inflammation were ruled out. In the liver of the diabetic group, although iNOS did not show significant changes, the significant increase in FGFR4 in the DM group activated the MAPK inflammatory signaling pathway, resulting in significant increases in p-ERK, p-JNK, p-P38, and inflammatory cytokines TNF-α and IL-1β compared to the DL group (Fig. 11B). These findings suggest that the DL group reversed the level of the BA-related signaling protein FGFR4 result in reduced MAPK signaling pathway alleviated liver inflammation.

Our previous studies have established the beneficial effects of diCQAs in regulating liver lipid synthesis genes, thereby alleviating HFD-induced fatty liver, demonstrated its potential in reducing inflammatory factors in mouse plasma, mitigating DSS-induced colitis by restore intestinal barrier function [15, 61]. In this study, we utilized mouse models in both the long-term group and the HFD + STZ-induced diabeic group to further investigate the anti-diabetic effects of diCQAs in ameliorating metabolic disorders and gut microbiota dysbiosis. We found that diCQAs were effective in relieving diabeic symptoms while enhancing intestinal barrier function by regulating gut microbiota. Notably, the long-term group results excluded a direct effect of diCQAs on glucolipid metabolism and inflammation, thus supporting the notion that polyphenolic natural products primarily exert their effects indirectly to modulating metabolic syndrome through the regulation gut microbiota [62, 63].

Conjugated BAs secreted from the liver into the intestine are hydrolyzed through a "gateway" reaction facilitated by gut microbiota carrying the BSH gene, a critical step necessary for the subsequent 7α-dehydroxylation of secondary BAs [64]. Previous research has shown that reducing intestinal microbiota BSH activity, increasing conjugated BA content in the ileum, inhibiting liver BA synthesis, and improving glucose and lipid metabolism are interrelated processes [13, 65, 66]. Lowering the content of T.β.MAC can activate farnesoid X receptor (FXR) signaling in the ileum [67]. BA tolerance, crucial for the survival of the intestinal microbial community, is widespread among humans, mice, and cattle, as revealed by the analysis of MAGs, which contain the K01442 gene responsible for BA hydrolysis (Fig. 2). These microbial populations adapt to the complex intestinal microbial environment through mechanisms such as bile tolerance, host mucoglycan degradation, polysaccharide utilization, lysine biosynthesis, and other metabolic abilities [68].

In addition to well-known BSH carriers such as Lactobacilli and Bifidobacteria [69, 70], this study observed significant increases in bacterial families associated with obesity and T2DM, including Lachnospiraceae (73.4%), Acutalibacteracea (77.3%), Ruminococcaceae (77.1%), and Bacteroidaceae (69.8%), all of which carry BSH hydrolysis genes (Fig. 2A). Notably, the genus Acetatifactor within the Lachnospiraceae, which carries both BSH and GH33 genes [19, 71], showed a significant association with the odds ratio for obesity and T2DM (Fig. 1D). Experimental results confirmed that the relative abundance of CAG-3, primarily composed of Lachnospiraceae, was significantly correlated with host glycemic indicators and plasma conjugated-BAs level.The DL group significantly reduced the relative abundance of CAG-3, including Lactobacillaceae and the associated genera Acetatifactor sp011959105 and Acetatifactor muris (Fig. 7). These genera have been identified as proficient in utilizing intestinal mucin MUC2 as a carbon source while concurrently hydrolyzing conjugated BAs, consequently disrupting the intestinal barrier. Plasma conjugated BA concentrations, particularly T.β.MAC, an FXR antagonist, were significantly higher in the DL and NC groups than in the DM group, whereas no similar changes were observed in the LNC group (Fig. 6). Cholesterol contents in both the liver and plasma were significantly lower in the DL and NC groups than DM group. These findings imply that DL efficiently reduced BSH hydrolase activity in the intestine, leading to the increased of conjugated BA hydrolysis in the ileum. Additionally, the down-regulation of FGFR4 expression in the liver, achieved through the inhibition of the enterohepatic FXR-FGF15 signaling axis, played a pivotal role in regulating cholesterol levels and BAs synthesis in both the liver and plasma.

BA acts as a signaling factor that binds to the FXR receptor in the ileum and the TGR5 receptor in the liver, ultimately affecting BA synthesis and influencing glucolipid metabolism [65, 66, 72]. Hideaki et al [67] revealed inhibition of FXR gene expression in the ileum, achieved by intragastric ampicillin, led to a significant increase in T.β.MCA content in the ileum. The re-supplementation of TCA did not restore FXR gene expression, highlighting the pivotal role of gut microbiota in regulating FXR receptor expression. This indicates the direct role of intestinal microbiota in modulating the enterohepatic FXR-FGF15 signaling axis by facilitating the hydrolysis of conjugated BA in the intestine through the action of the BSH hydrolysis genes it carries. In this study, we observed that BA concentrations in plasma, especially conjugated BA concentrations which closely paralleled those in the distal ileum, and they were significantly increased in the DL group lead to FXR in the ileum was significantly upregulated in the DM group than DL group, finnaly increased FGF15 hormone expression in ileum. Similar results were not observed in the long-term group (Fig. 8), ruling out any direct stimulatory effects of diCQAs. The observations suggest that the high T.β.MCA content in the DL group inhibited FXR-FGF15 signaling in the ileum, rather than indicating a direct effect of diCQAs. Research by Chiara et al. demonstrated that FXR-^/− and FGF15^−/− mice models showed downregulation of the FXR-FGF15 signaling axis within the enterohepatic circuit, leading to increased de novo BA synthesis in the liver [73]. The DM group displayed a significant increase in FGFR4 expression in the liver, which subsequently led to the inhibition of CYP7A1 protein expression through the ileal FXR-FGF15 signaling pathway. This inhibition, in turn, resulted in the suppression of de novo BA synthesis. In parallel, the DM group exhibited significantly elevated levels of HMGCR, a rate-limiting protein involved in liver cholesterol synthesis, as well as its regulatory protein, SREBP2. This led to an increased promotion of cholesterol synthesis compared to the NC and DL groups. Notably, the DL group effectively reversed the observed elevated cholesterol levels and slight BA stasis in the liver of the DM group by enhancing BSEP expression. The activation of FGFR4 in the liver further triggered the downstream MAPK/NF-KB inflammatory signaling pathway [60]. FGFR4 protein expression was markedly activated by FGF15 in the DM group, surpassing its expression levels in the DL group. This activation triggered heightened levels of key phosphoproteins, including p-p38, p-ERK, and p-JNK, integral components of the inflammatory signaling pathway. Consequently, led to significantly elevated levels of MAPK/NF-κB pathway and inflammatory cytokines like TNF-α and IL-1β in the DM group compared to the DL group, despite no significant changes in iNOS when compared to the DL group (Fig. 9). Therefore, DL effectively mitigated the MAPK/NF-kB inflammatory signaling pathway by inhibiting the enterohepatic FXR-FGF15 signaling cascade.

In glucolipid metabolism, the BA receptor TGR5 plays a crucial role in regulating these processes when activated by CDCA and DCA [74]. In the DL group, although GLP-1, SCFAs in feces, and GLP-1R, GPR43, and IRS-1 protein expresstion in the liver remained unchanged, CDCA significantly increased TGR5 protein expression levels. This activation stimulated key pathways involved in glucolipid metabolism, such as PI3K/AKT/GSK3β and AMPK consequently enhancing liver glucolipid metabolism abilities. Additionally, the same protein levels related to glycogen synthesis and energy metabolism in muscle exhibited significant increases in the DL group compared to the DM group, enhanced muscle glycogen synthesis capacity. Thus, DL effectively attenuated the enterohepatic axis FXR-FGF15 signaling cascade, inhibited liver FGFR4 expression, and combinded activation of liver TGR5, ameliorating hyperglycemia and addressing disruptions in glucose and lipid metabolism in diabetic mice.

Despite our study of AGP cohort data, there is a continued need for more comprehensive cohort data containing information, such as BAs or FBG .et al. Limited by the analyses of the MAGs, it is possible that some of the BSH-carrying gut microbiota were not assembled or discovered. Therefore, large data is still needed for an integrated analysis to identify BSH-carrying bacteria. Additionally, the evaluation of diCQAs' potential side effects and the determination of an optimal dosage for humans will necessitate clinical trials before it can be established as a routine therapy. Nevertheless, our preliminary studies do show promise in demonstrating the efficacy of diCQAs in modulating gut microbiota and bile acid metabolism, leading to potential anti-diabetic effects.

In this study, we excluded the direct regulatory effects of diCQAs through long-term group mice models and mice administered with 200mg/kg diCQAs (DL group) demonstrated a more pronounced effect in regulating glucolipid metabolism in HFD + STZ induced diabetic mouse compared to 400mg/kg diCQAs (DH group). In summary, we uncovered that DL regulates mechanisms associated with glucolipid metabolism and inflammnation in diabetic mice via its effects on regulate gut microbiota involves the following: (1) The DL group enhances the intestinal barrier by reducing the relative abundance of gut microbiota such as Lachnospiraceae, Acetatifactor, Acetatifactor sp011959105, and Acetatifactor muris, while also increasing the relative abundance of Alistipes, including Alistipes sp003979135, a beneficial bacterium from the Rikenellaceae; (2) The DL group increases the content of conjugated BAs and inhibits the ileum FXR-FGF15 signal axis, ultimately resulting in the regulation of hepatic BA synthesis through the inhibition of liver CYP7A1 expression and the reduction of cholesterol accumulation, mediated by HMGCR and its regulatory protein SREBP2; (3) DL suppresses the FGFR4-MAPK/NF-kB signaling pathway, leading to the inhibition of hepatic inflammation. (4) In the DL group activated hepatic TGR5 signaling by CDCA, initiating downstream pathways such as AKT/GSK3β and AMPK. Consequently, this activation led to an increase in hepatic glycogen synthesis and energy expenditure. Similar mechanisms were observed in the muscle tissues of the DL group. In summary, diCQAs (200mg/kg) ameliorates glucolipid metabolism disorders and mitigates inflammation in diabetic mice, and achieve these benefits through regulate gut microbiota, increase conjugated BA levels, and inactivate the enterohepatic FXR-FGF15 signaling axis. These findings may provide new insights for diCQAs in ilex tea, understanding the role of gut microbiota in diabetes mellitus, and modulating gut microbiota as a therapeutic target.

AKT: v-akt murine thymoma viral oncogene homolog 1

AMPKa: 5′-AMP-activated protein kinase-alpha1

ANI: Average Nucleotide Identity

BA: Bile acid

BMI: Body mass index

BSEP: Bile salt export pump

BSH: Bile salt hydrolase

CDCA: Chenodeoxycholic acid

CYP7A1: Cholesterol 7α-hydroxylase

DCA: Deoxycholic acid

DiCQAs: Dicaffeoylquinic acid

ERK: Extracellular-signal-regulated kinase

FBG: Fasting blood glucose

Fiaf: Fasting-induced adipose factor

FXR: Farnesoid X Receptor

FGF15: Fibroblast growth factor 15

GF: Germ-free

GLP-1R: Glucagon-like peptide 1 receptor

GSK3b: glycogen synthase kinase 3-beta gene

HMGR: 3-hydroxy-3-methylglutaryl-CoA-reductase

MAG: Metagenome-Assembled Genomes

MAPK: Mitogen-activated protein kinases

MUC2: mucin-2

IL-1b: Interleukin-1 beta

IL-6: Interleukin-6

iNOS: Inducible nitric oxide synthase

IRS-1: Insulin receptor substrate-1

LPS: Lipopolysaccharides

JNK: c-Jun N-terminal kinases

NF-kB: Nuclear Factor-kappaB

OGTT: Oral glucose tolerance test

PCoA: Principal coordinate analysis

PI3K: phosphoinositide 3-kinase

PPARg: Peroxisome proliferator-activated receptor

SCFA: Short chain fatty acid

SREBP1: Sterol-regulatory binding protein 1

SREBP2: Sterol-regulatory binding protein 2

TGR5: Takeda G protein receptor 5

TLR4: Toll-like receptor 4

TNF-a: Tumor necrosis factor

T.β.MCA: Tauro-b-muricholic acid

SZT: Streptozotocin

Funding

This work was supported by Grants-in-Aid for scientific research from the National Natural Science Foundation of China (31871761), Guizhou Science and Technology Department (QKHJC-ZK [2021]-YB179) and a project funded by the Priority Academic Program Development of Jiangsu Higher Education Institutions.

Acknowledgements

We thanks the Bioinformatics Center of Nanjing Agricultural University, the Nanjing Key Research for support the hardware of supercomputing platforms.

Authors' contributions

Yujie Huang: Performed the experiments, the overall analysis and designed the research; Weiqi Xu: investigation; Guijie Chen: investigation; Yi Sun: investigation, project administration; Xiaoxiong Zeng: Conceptualization, supervision, project administration, writing-reviewing and editing.

Availability of data and materials

All sequencing data for the long-term and diabetes groups were deployed at National Center for Biotechnology Information (NCBI) under PRJNA1029050; Representative MAGs contains functional gene annotation results for the human gut (n=615) have been deployed at Figshare (https://figshare.com/account/articles/24328567), raw sequence data downloaded from NCBI under PRJNA373901; Raw sequence data for the AGP cohort (n=13,422) downloaded from NCBI under PRJEB11419; Representative MAGs for dairy cows were downloaded from Figshare (https://doi.org/10.6084/m9.figshare.14456637 and 14176574); Representative MAGs for mice were downloaded from github (https://github.com/tillrobin/iMGMC).

Ethics declarations

Ethics approval and consent to participate

The animal experiments strictly adhered to the Laboratory Animal Standards of Welfare and Ethics (DB32/T-2911-2016) and received approval from the Institutional Animal Care and Use Committee at Nanjing Agricultural University (SYXK-2017-0027)

Competing interests

The authors declare that they have no competing interests.

Nicholson JK, Holmes E, Wilson ID. Gut microorganisms, mammalian metabolism and personalized health care. Nature Reviews Microbiology. 2005;3(5):431-8.
Wada J, Makino H. Innate immunity in diabetes and diabetic nephropathy. Nature Reviews Nephrology. 2016;12(1):13-26.
An D, Oh SF, Olszak T, Neves JF, Avci FY, Erturk-Hasdemir D, et al. Sphingolipids from a symbiotic microbe regulate homeostasis of host intestinal natural killer T cells. Cell. 2014;156(1):123-33.
Bäckhed F, Manchester JK, Semenkovich CF, Gordon JI. Mechanisms underlying the resistance to diet-induced obesity in germ-free mice. Proceedings of the National Academy of Sciences. 2007;104(3):979-84.
Canfora EE, Jocken JW, Blaak EE. Short-chain fatty acids in control of body weight and insulin sensitivity. Nat Rev Endocrinol. 2015;11(10):577-91; doi: 10.1038/nrendo.2015.128.
Vanweert F, Schrauwen P, Phielix E. Role of branched-chain amino acid metabolism in the pathogenesis of obesity and type 2 diabetes-related metabolic disturbances BCAA metabolism in type 2 diabetes. Nutrition & Diabetes. 2022;12(1):35.
Lynch CJ, Adams SH. Branched-chain amino acids in metabolic signalling and insulin resistance. Nature Reviews Endocrinology. 2014;10(12):723-36.
Sun L, Cai J, Gonzalez FJ. The role of farnesoid X receptor in metabolic diseases, and gastrointestinal and liver cancer. Nature Reviews Gastroenterology & Hepatology. 2021;18(5):335-47.
Wahlström A, Sayin SI, Marschall H-U, Bäckhed F. Intestinal crosstalk between bile acids and microbiota and its impact on host metabolism. Cell metabolism. 2016;24(1):41-50.
Thomas C, Pellicciari R, Pruzanski M, Auwerx J, Schoonjans K. Targeting bile-acid signalling for metabolic diseases. Nature reviews Drug discovery. 2008;7(8):678-93.
Tremaroli V, Bäckhed F. Functional interactions between the gut microbiota and host metabolism. Nature. 2012;489(7415):242-9.
Just S, Mondot S, Ecker J, Wegner K, Rath E, Gau L, et al. The gut microbiota drives the impact of bile acids and fat source in diet on mouse metabolism. Microbiome. 2018;6:1-18.
Li F, Jiang C, Krausz KW, Li Y, Albert I, Hao H, et al. Microbiome remodelling leads to inhibition of intestinal farnesoid X receptor signalling and decreased obesity. Nature communications. 2013;4(1):2384.
Huang F, Zheng X, Ma X, Jiang R, Zhou W, Zhou S, et al. Theabrownin from Pu-erh tea attenuates hypercholesterolemia via modulation of gut microbiota and bile acid metabolism. Nature Communications. 2019;10(1):4971.
Xie M, Chen G, Wan P, Dai Z, Zeng X, Sun Y. Effects of dicaffeoylquinic acids from Ilex kudingcha on lipid metabolism and intestinal microbiota in high-fat-diet-fed mice. Journal of agricultural and food chemistry. 2018;67(1):171-83.
Xie M, Chen G, Wan P, Dai Z, Zeng X, Sun Y. Effects of Dicaffeoylquinic Acids from Ilex kudingcha on Lipid Metabolism and Intestinal Microbiota in High-Fat-Diet-Fed Mice. J Agric Food Chem. 2019;67(1):171-83.
Xu W, Huang Y, Zhou W, Peng Y, Kan X, Dong W, et al. Theasinensin A attenuated diabetic development by restoring glucose homeostasis, improving hepatic steatosis and modulating gut microbiota in high-fat-diet/streptozotocin-induced diabetic mice. Food Science and Human Wellness. 2023;12(6):2073-86.
McDonald D, Hyde E, Debelius JW, Morton JT, Gonzalez A, Ackermann G, et al. American gut: an open platform for citizen science microbiome research. Msystems. 2018;3(3):10.1128/msystems. 00031-18.
Bolyen E, Rideout JR, Dillon MR, Bokulich NA, Abnet CC, Al-Ghalith GA, et al. Reproducible, interactive, scalable and extensible microbiome data science using QIIME 2. Nature biotechnology. 2019;37(8):852-7.
Amir A, McDonald D, Navas-Molina JA, Debelius J, Morton JT, Hyde E, et al. Correcting for microbial blooms in fecal samples during room-temperature shipping. MSystems. 2017;2(2):10.1128/msystems. 00199-16.
Robeson MS, 2nd, O'Rourke DR, Kaehler BD, Ziemski M, Dillon MR, Foster JT, et al. RESCRIPt: Reproducible sequence taxonomy reference database management. PLoS Comput Biol. 2021;17(11):e1009581; doi: 10.1371/journal.pcbi.1009581.
Olm MR, Brown CT, Brooks B, Banfield JF. dRep: a tool for fast and accurate genome de-replication that enables tracking of microbial genotypes and improved genome recovery from metagenomes. bioRxiv. 2017:108142.
Almeida A, Mitchell AL, Boland M, Forster SC, Gloor GB, Tarkowska A, et al. A new genomic blueprint of the human gut microbiota. Nature. 2019;568(7753):499-504.
Parks DH, Imelfort M, Skennerton CT, Hugenholtz P, Tyson GW. CheckM: assessing the quality of microbial genomes recovered from isolates, single cells, and metagenomes. Genome research. 2015;25(7):1043-55.
Xie F, Jin W, Si H, Yuan Y, Tao Y, Liu J, et al. An integrated gene catalog and over 10,000 metagenome-assembled genomes from the gastrointestinal microbiome of ruminants. Microbiome. 2021;9(1):137.
Asnicar F, Thomas AM, Beghini F, Mengoni C, Manara S, Manghi P, et al. Precise phylogenetic analysis of microbial isolates and genomes from metagenomes using PhyloPhlAn 3.0. Nature communications. 2020;11(1):2500.
Foley MH, Walker ME, Stewart AK, O’Flaherty S, Gentry EC, Patel S, et al. Bile salt hydrolases shape the bile acid landscape and restrict Clostridioides difficile growth in the murine gut. Nature Microbiology. 2023;8(4):611-28.
Jia B, Park D, Hahn Y, Jeon CO. Metagenomic analysis of the human microbiome reveals the association between the abundance of gut bile salt hydrolases and host health. Gut Microbes. 2020;11(5):1300-13.
Vasimuddin M, Misra S, Li H, Aluru S: Efficient architecture-aware acceleration of BWA-MEM for multicore systems. In: 2019 IEEE international parallel and distributed processing symposium (IPDPS). IEEE; 2019: 314-24.
Danecek P, Bonfield JK, Liddle J, Marshall J, Ohan V, Pollard MO, et al. Twelve years of SAMtools and BCFtools. Gigascience. 2021;10(2):giab008.
Castellanos-Jankiewicz A, Guzmán-Quevedo O, Fenelon VS, Zizzari P, Quarta C, Bellocchio L, et al. Hypothalamic bile acid-TGR5 signaling protects from obesity. Cell Metabolism. 2021;33(7):1483-92. e10.
Oswald S, Scheuch E, Cascorbi I, Siegmund W. A LC–MS/MS method to quantify the novel cholesterol lowering drug ezetimibe in human serum, urine and feces in healthy subjects genotyped for SLCO1B1. Journal of Chromatography B. 2006;830(1):143-50.
Hosomi K, Saito M, Park J, Murakami H, Shibata N, Ando M, et al. Oral administration of Blautia wexlerae ameliorates obesity and type 2 diabetes via metabolic remodeling of the gut microbiota. Nature communications. 2022;13(1):4477.
Liu H, Chen X, Hu X, Niu H, Tian R, Wang H, et al. Alterations in the gut microbiome and metabolism with coronary artery disease severity. Microbiome. 2019;7(1):1-14.
Friedman J, Alm EJ. Inferring correlation networks from genomic survey data. PLoS computational biology. 2012;8(9):e1002687.
Mallick H, Rahnavard A, McIver LJ, Ma S, Zhang Y, Nguyen LH, et al. Multivariable association discovery in population-scale meta-omics studies. PLoS computational biology. 2021;17(11):e1009442.
Biagi E, Franceschi C, Rampelli S, Severgnini M, Ostan R, Turroni S, et al. Gut microbiota and extreme longevity. Current Biology. 2016;26(11):1480-5.
Newton RJ, McLellan SL, Dila DK, Vineis JH, Morrison HG, Eren AM, et al. Sewage reflects the microbiomes of human populations. mBio. 2015;6(2):e02574; doi: 10.1128/mBio.02574-14.
Li S, Li J, Mao G, Yan L, Hu Y, Ye X, et al. Effect of the sulfation pattern of sea cucumber-derived fucoidan oligosaccharides on modulating metabolic syndromes and gut microbiota dysbiosis caused by HFD in mice. Journal of functional foods. 2019;55:193-210.
Atkinson HJ, Morris JH, Ferrin TE, Babbitt PC. Using sequence similarity networks for visualization of relationships across diverse protein superfamilies. PloS one. 2009;4(2):e4345.
Teufel F, Almagro Armenteros JJ, Johansen AR, Gislason MH, Pihl SI, Tsirigos KD, et al. SignalP 6.0 predicts all five types of signal peptides using protein language models. Nat Biotechnol. 2022;40(7):1023-5.
Marcobal A, Southwick AM, Earle KA, Sonnenburg JL. A refined palate: bacterial consumption of host glycans in the gut. Glycobiology. 2013;23(9):1038-46; doi: 10.1093/glycob/cwt040.
Blakeley-Ruiz JA, McClintock CS, Shrestha HK, Poudel S, Yang ZK, Giannone RJ, et al. Morphine and high-fat diet differentially alter the gut microbiota composition and metabolic function in lean versus obese mice. ISME Communications. 2022;2(1); doi: 10.1038/s43705-022-00131-6.
Fang J, Wang H, Zhou Y, Zhang H, Zhou H, Zhang X. Slimy partners: The mucus barrier and gut microbiome in ulcerative colitis. Experimental & Molecular Medicine. 2021;53(5):772-87.
Bell A, Juge N. Mucosal glycan degradation of the host by the gut microbiota. Glycobiology. 2021;31(6):691-6.
Szkudelski T. The mechanism of alloxan and streptozotocin action in B cells of the rat pancreas. Physiological research. 2001;50(6):537-46.
Donath MY, Shoelson SE. Type 2 diabetes as an inflammatory disease. Nature reviews immunology. 2011;11(2):98-107.
Jin Y, Liu R, Xie J, Xiong H, He JC, Chen N. Interleukin-10 deficiency aggravates kidney inflammation and fibrosis in the unilateral ureteral obstruction mouse model. Laboratory investigation. 2013;93(7):801-11.
Trauner M, Boyer JL. Bile salt transporters: molecular characterization, function, and regulation. Physiological reviews. 2003;83(2):633-71.
Winston JA, Theriot CM. Diversification of host bile acids by members of the gut microbiota. Gut Microbes. 2020;11(2):158-71; doi: 10.1080/19490976.2019.1674124.
Fiorucci S, Mencarelli A, Palladino G, Cipriani S. Bile-acid-activated receptors: targeting TGR5 and farnesoid-X-receptor in lipid and glucose disorders. Trends Pharmacol Sci. 2009;30(11):570-80.
van Nierop FS, Scheltema MJ, Eggink HM, Pols TW, Sonne DP, Knop FK, et al. Clinical relevance of the bile acid receptor TGR5 in metabolism. The lancet Diabetes & endocrinology. 2017;5(3):224-33.
Hata S, Okamura T, Kobayashi A, Bamba R, Miyoshi T, Nakajima H, et al. Gut Microbiota Changes by an SGLT2 Inhibitor, Luseogliflozin, Alters Metabolites Compared with Those in a Low Carbohydrate Diet in db/db Mice. Nutrients. 2022;14(17); doi: 10.3390/nu14173531.
Zhang N, Wang Q, Lin F, Zheng B, Huang Y, Yang Y, et al. Neoagarotetraose alleviates high fat diet induced obesity via white adipocytes browning and regulation of gut microbiota. Carbohydrate Polymers. 2022;296:119903.
Fall T, Salihovic S, Brandmaier S, Nowak C, Ganna A, Gustafsson S, et al. Non-targeted metabolomics combined with genetic analyses identifies bile acid synthesis and phospholipid metabolism as being associated with incident type 2 diabetes. Diabetologia. 2016;59(10):2114-24.
Guzior DV, Quinn RA. Microbial transformations of human bile acids. Microbiome. 2021;9(1):1-13.
Uriarte I, Fernandez-Barrena MG, Monte MJ, Latasa MU, Chang HC, Carotti S, et al. Identification of fibroblast growth factor 15 as a novel mediator of liver regeneration and its application in the prevention of post-resection liver failure in mice. Gut. 2013;62(6):899-910.
Uriarte I, Fernandez-Barrena MG, Monte MJ, Latasa MU, Chang HC, Carotti S, et al. Identification of fibroblast growth factor 15 as a novel mediator of liver regeneration and its application in the prevention of post-resection liver failure in mice. Gut. 2013;62(6):899-910; doi: 10.1136/gutjnl-2012-302945.
Kida T, Tsubosaka Y, Hori M, Ozaki H, Murata T. Bile acid receptor TGR5 agonism induces NO production and reduces monocyte adhesion in vascular endothelial cells. Arteriosclerosis, thrombosis, and vascular biology. 2013;33(7):1663-9.
Chen H, Li J, Pi C, Guo D, Zhang D, Zhou X, et al. FGF19 induces the cell cycle arrest at G2-phase in chondrocytes. Cell Death Discovery. 2023;9(1):250.
Wan P, Peng Y, Chen G, Xie M, Dai Z, Huang K, et al. Modulation of gut microbiota by Ilex kudingcha improves dextran sulfate sodium-induced colitis. Food Research International. 2019;126:108595.
Xie M, Chen G, Wan P, Dai Z, Hu B, Chen L, et al. Modulating effects of dicaffeoylquinic acids from Ilex kudingcha on intestinal microecology in vitro. Journal of agricultural and food chemistry. 2017;65(47):10185-96.
Roopchand DE, Carmody RN, Kuhn P, Moskal K, Rojas-Silva P, Turnbaugh PJ, et al. Dietary polyphenols promote growth of the gut bacterium Akkermansia muciniphila and attenuate high-fat diet–induced metabolic syndrome. Diabetes. 2015;64(8):2847-58.
Ridlon JM, Harris SC, Bhowmik S, Kang D-J, Hylemon PB. Consequences of bile salt biotransformations by intestinal bacteria. Gut microbes. 2016;7(1):22-39.
Sayin SI, Wahlström A, Felin J, Jäntti S, Marschall H-U, Bamberg K, et al. Gut microbiota regulates bile acid metabolism by reducing the levels of tauro-beta-muricholic acid, a naturally occurring FXR antagonist. Cell metabolism. 2013;17(2):225-35.
Zheng X, Chen T, Jiang R, Zhao A, Wu Q, Kuang J, et al. Hyocholic acid species improve glucose homeostasis through a distinct TGR5 and FXR signaling mechanism. Cell metabolism. 2021;33(4):791-803. e7.
Kuribayashi H, Miyata M, Yamakawa H, Yoshinari K, Yamazoe Y. Enterobacteria-mediated deconjugation of taurocholic acid enhances ileal farnesoid X receptor signaling. European journal of pharmacology. 2012;697(1-3):132-8.
Lin L, Lai Z, Yang H, Zhang J, Qi W, Xie F, et al. Genome-centric investigation of bile acid metabolizing microbiota of dairy cows and associated diet-induced functional implications. The ISME journal. 2023;17(1):172-84.
Song Z, Cai Y, Lao X, Wang X, Lin X, Cui Y, et al. Taxonomic profiling and populational patterns of bacterial bile salt hydrolase (BSH) genes based on worldwide human gut microbiome. Microbiome. 2019;7:1-16.
Kumar RS, Brannigan JA, Prabhune AA, Pundle AV, Dodson GG, Dodson EJ, et al. Structural and functional analysis of a conjugated bile salt hydrolase from Bifidobacterium longum reveals an evolutionary relationship with penicillin V acylase. Journal of Biological Chemistry. 2006;281(43):32516-25.
Glover JS, Ticer TD, Engevik MA. Characterizing the mucin-degrading capacity of the human gut microbiota. Scientific reports. 2022;12(1):8456.
Wang XX, Edelstein MH, Gafter U, Qiu L, Luo Y, Dobrinskikh E, et al. G protein-coupled bile acid receptor TGR5 activation inhibits kidney disease in obesity and diabetes. Journal of the American Society of Nephrology: JASN. 2016;27(5):1362.
Degirolamo C, Rainaldi S, Bovenga F, Murzilli S, Moschetta A. Microbiota modification with probiotics induces hepatic bile acid synthesis via downregulation of the Fxr-Fgf15 axis in mice. Cell reports. 2014;7(1):12-8.
Jia W, Xie G, Jia W. Bile acid–microbiota crosstalk in gastrointestinal inflammation and carcinogenesis. Nature reviews Gastroenterology & hepatology. 2018;15(2):111-28.

No competing interests reported.

Download PDF

Version 1

posted

You are reading this latest preprint version

Anti-diabetic effect of di-caffeoylquinic acid is associated with the modulation of gut microbiota and bile acid metabolism

Status:

Version 1

Abstract

Background

Results

Conclusions

Figures

Background

Methods

Results

The AGP (American Gut Project) cohort study uncover gut microbiota associated with diabetes and obese

Comparisons of the sequences and structures of BSHs

DL regulate inhibit ileal FXR-FGF15 signaling axis and enhances the intestinal barrier

Discussion

Conclusion

Abbreviations

Declarations

References

Additional Declarations

Supplementary Files

Status:

Version 1