All animal procedures were conducted in strict accordance with the Guidelines for the Care and Use of Laboratory Animals of Northeast Agricultural University (NEAU-[2011]-9), and the experimental protocols were approved by the University Animal Care Committee. Trp intervention in a mouse model: Male C57BL/6J mice were obtained from HFK Co. Ltd. (SPF, Beijing, China) at 5 weeks of age. Mice were housed under specific pathogen-free (SPF) conditions in plastic rodent cages at 22 to 24 °C with a 12-h light/dark cycle and had ad libitum access to food and water throughout the experimental period. The diets consisted of a normal CD (10% fat; 3,660 kcal/g; Xietong Biotechnology Company, Nanjing, China) and a high-fat diet (HFD) (60% fat; 5,128 kcal/g; Xietong Biotechnology Company, Nanjing, China). Mice were acclimated to the normal CD for 1 week before the experiments began. Each group had 8 replicates. Body weight and food consumption were monitored weekly. At the end of the experiment, all mice were euthanized for tissue collection.

In the treatment study, mice were weighed and randomly assigned to 2 diets. To induce obesity, mice were fed the HFD diet, while the CD diet served as the control. Obesity was considered established when the body weight of the HFD-fed mice was 20% higher than that of the CD-fed mice. After 8 weeks of HFD feeding to induce obesity, Trp was added to both diets at twice the original molar concentration and continued for another 8 weeks. Mice fed CD and HFD without Trp served as controls (Fig. 1A). The actual concentration of Trp in each group is as follows: CD group, 0.20%; CD + Trp group, 0.40%; HFD group, 0.27%; and HFD + Trp group, 0.54%. To ensure isonitrogenous diets and minimize the impact of other amino acids, additional nitrogen in the CD and HFD groups was compensated by supplementing with an amino acid mixture excluding Trp, added in proportion.

It is important to note that during the final week of the intervention, fresh fecal samples (gut microbiota with or without its metabolites) from the HFD + Trp group were aseptically collected for subsequent fecal microbiota transplantation (FMT) or sterile fecal filtrate (SFF) preparations. FMT and SFF were prepared as described in [21]. After feeding HFD for 8 weeks, mice underwent gavage with a mixed antibiotic solution (1 g/l neomycin sulfate, 1 g/l metronidazole, 1 g/l ampicillin, and 0.5 g/l vancomycin, 10 ml/kg) for 1 week. Subsequently, mice were orally gavaged with FMT (10 ml/kg/day) or SFF (10 ml/kg/day), while the control group received sterile saline orally (10 ml/kg/day). Gavage continued for 5 weeks with HFD maintained throughout (Fig. 3E).

In the FXR regulation study, mice were divided into 4 groups after 8 weeks on an HFD: (a) HFD vehicle (HFD), (b) HFD supplemented with 100 mg/kg/day Gly-MCA by gavage (HFD + Gly-MCA), (c) HFD vehicle supplemented with additional Trp (HFD + Trp), and (d) HFD supplemented with additional Trp and 100 mg/kg/day fexaramine by gavage (HFD + Trp + Fex). The intervention lasted 6 weeks (Fig. 4A). The actual concentration of Trp in each group is as follows: HFD group and HFD + Gly-MCA, 0.27%; HFD + Trp group and HFD + Trp + Fex, 0.54%.

Trp intervention in a finishing pig model: We purchased 12 healthy male finishing pigs (Large White) from a breeding pig farm (Jubao, Suihua City, Heilongjiang Province, China). A total of 12 finishing pigs were randomly allocated to 2 dietary treatments: a control group fed the basal diet, and a Trp group fed the basal diet supplemented with 0.78% Trp. The actual concentration of Trp in each group is as follows: control group, 0.14%; Trp group, 0.92%. To ensure isonitrogenous diets and minimize the effects of other amino acids, additional nitrogen in the control group was compensated by supplementing with an amino acid mixture excluding Trp, added proportionally. There were 6 replicates per group, with 1 pig per replicate. Pigs were individually housed in separate pens within the same room. The duration of the experiment was 30 days (Fig. 5A). Trp was purchased from Hebei Huayang Biological Technology Co. Ltd. (≥99.0% purity, Hengshui, China).

Polymerase chain reaction (PCR) amplification was performed using the following primers: 341F (CCTAYGGGRBGCASCAG) and 806R (GGACTACNNGGGTATCTAAT). All PCRs were carried out with Phusion High-Fidelity PCR Master Mix (New England Biolabs). The same volume of 1× loading buffer (containing SYB green) was mixed with PCR products, and electrophoresis was operated on 2% agarose gel for detection. Samples with bright main strip between 400 and 450 base pairs (bp) were chosen for further experiments. PCR products were mixed in equidensity ratios. Then, mixture PCR products were purified with Qiagen Gel Extraction Kit (Qiagen, Germany). Sequencing libraries were generated using TruSeq DNA PCR-Free Sample Preparation Kit (Illumina, USA). The library quality was assessed on the Qubit 2.0 Fluorometer (Thermo Fisher Scientific) and Agilent Bioanalyzer 2100 system. At last, the library was sequenced on an Illumina HiSeq 2500 platform and 250-bp paired-end reads were generated [22].

Sequences analysis was performed by Uparse software (Uparse v7.0.1001). Sequences with ≥97% similarity were assigned to the same operational taxonomic units (OTUs). Representative sequence for each OTU was screened for further annotation. For each representative sequence, the GreenGene Database was used based on RDP classifier (version 2.2) algorithm to annotate taxonomic information. Multiple sequence alignment was conducted using the MUSCLE software. OTU abundance information was normalized using a standard of sequence number corresponding to the sample with the least sequences. Subsequent analysis of alpha diversity and beta diversity was performed based on the output normalized data. Alpha diversity was applied to analyze the complexity of species diversity for a sample. Beta diversity analysis was used to evaluate differences of samples in species complexity. Beta diversity on both weighted and unweighted UniFrac was calculated by QIIME software (version 1.7.0). Principal coordinate analysis (PCoA) was performed to get principal coordinates and visualize from complex, multidimensional data. PCoA analysis was displayed in R software (version 2.15.3). LEfSe (LDA effect size) analyses were performed using the Kruskal–Wallis rank sum test for all characterized species, and significantly different species were obtained by detecting the difference in species abundance between different groups. The Wilcoxon rank sum test was then used to test whether all subspecies of the significantly different species obtained in the previous step converged to the same taxonomic level. Finally, linear discriminant analyses (LDAs) were used to obtain the final divergent species [23]. In addition, Spearman correlation between the BA profile and significantly different microbes and radar plots of the relative abundance of BSH-enriched microbes were produced using the ChiPlot website.

All data are presented as mean ± standard error of the mean (SEM). Statistical analysis was conducted using a 2-tailed unpaired Student's t test in SPSS Statistics (Chicago, USA). A significance level of P < 0.05 was adopted. The statistical significance is denoted as follows: *P < 0.05; **P < 0.01; ^#P < 0.05, ^##P < 0.01.

Other methods are in the Supplementary Materials.

To investigate the regulatory effects of Trp on MS, mice were fed an HFD for 8 weeks to induce MS, followed by 8 weeks of Trp intervention. The effect of Trp on a normal CD-fed mice was also examined (Fig. 1A). Compared to the CD group, the HFD group exhibited lower food intake, higher energy intake, and increased weight gain in mice (P < 0.05). Following Trp intervention, Trp treatment in the HFD group reduced weight gain (P < 0.05), while Trp treatment in the CD group had no effect on weight gain (Fig. 1B and Fig. S1A and B). The liver index and white adipose tissue (WAT) index also increased in the HFD group compared to the CD group (P < 0.05). Hematoxylin and eosin (H&E) and Oil Red O staining revealed larger lipid droplets and increased ballooning degeneration in the hepatocytes of HFD-treated mice. However, under HFD conditions, Trp intervention led to a reduction in WAT and remission of hepatic steatosis (P < 0.05) (Fig. 1C to E).

In order to investigate the impact of Trp supplementation on glucose homeostasis, we conducted glucose tolerance tests (GTTs) and insulin tolerance tests (ITTs). The results demonstrated that an HFD resulted in an increase in the area under the curve (AUC) of GTT and ITT assay, as well as an elevation in serum glucose levels. However, under HFD conditions, Trp intervention led to a reduction in AUC of the GTT and ITT assay and serum glucose levels (P < 0.05), indicating that Trp significantly improved glucose tolerance (Fig. 1F to H). The HFD also significantly increased serum and liver levels of triglycerides (TGs) and total cholesterol (TC), as well as serum aspartate aminotransferase (AST) and alanine aminotransferase (ALT). Trp alleviated HFD-induced hyperlipidemia, as evidenced by reduced TG and TC levels in serum and liver (P < 0.05). Trp significantly reduced serum AST and ALT levels under MS conditions, indicating that liver damage was reduced by Trp (P < 0.05) (Fig. 1J and K).

However, under CD conditions, Trp supplementation had no significant effect on liver weight, WAT weight, blood glucose levels, serum and liver lipid levels, and serum AST and ALT levels (Fig. 1C to K). Under MS conditions, Trp increased serum total bile acid (TBA) levels (P < 0.05). Interestingly, Trp increased liver TBA levels in mice with or without MS (P < 0.05) (Fig. 1I and L). These data suggest that Trp has no significant influence on lipid metabolism under normal metabolic conditions but can improve lipid metabolism to alleviate MS under MS conditions.

Our preliminary experiments indicated that dietary supplementation with aromatic amino acids resulted in a reduction of TG and an alleviation of hepatic steatosis, which was accompanied by an increase in BA synthesis in mice [21]. In the present study, we investigated the role of Trp, a member of the aromatic amino acids, in the BA cycle under MS conditions. The results showed that both the BAs synthesized in the liver and those excreted in the feces were altered by Trp supplementation. Trp markedly elevated the levels of numerous BA components in liver tissue, particularly non-12α-hydroxylated BAs (non-12-OH BAs), including chenodeoxycholic acid (CDCA), taurochenodeoxycholic acid (TCDCA), tauromuricholic acid (TMCA), tauroursodeoxycholic acid (TUDCA), hyocholic acid (HCA), hyodeoxycholic acid (HDCA), and glycohyodeoxycholic acid (GHDCA). In addition, the levels of deoxycholic acid (DCA) and taurodeoxycholic acid (TDCA) in 12-OH BAs were also enhanced. In fecal tissue, Trp also significantly elevated the levels of many non-12-OH BAs, including CDCA, TCDCA, β-muricholic acid (β-MCA), TMCA, HDCA, and taurohyodeoxycholic acid (THDCA). Overall, Trp increased not only TBA but also total non-12-OH and conjugated BAs in the liver and feces. It also increased the ratio of non-12-OH to 12-OH BAs in the liver and feces (P < 0.05). Notably, in terms of the percentage of BAs of each species, the pie charts show an increase in the proportion of MCA species and a decrease in the proportion of cholic acid (CA) species in the liver and feces of Trp-treated mice (Fig. 2A and B).

We next determined the mRNA expression levels of genes associated with BA metabolism. Consistent with elevated BA levels, the mRNA expression levels of BA synthesis genes Cyp7a1 and oxysterol 7α-hydroxylase (Cyp7b1) were significantly increased by Trp supplementation. The mRNA expression level of the hepatic BA receptor Fxr was unchanged, but the mRNA expression level of ileum Fxr was markedly suppressed. In addition, the mRNA expression levels of BA transport genes, hepatic bile salt export pump (Bsep) and ileum apical sodium-dependent bile acid transporter (Abst), were also significantly increased (P < 0.05) (Fig. 2C). Therefore, to determine whether BA metabolism plays a key role in MS, we correlated liver BA profiles, fecal BA profiles, and BA metabolism-related genes with MS indicators using Mantel tests. Overall, liver BA profiles were significantly correlated with body weight (BW), GTT AUC, serum TG, serum TC, AST, and serum TBA; fecal BA profiles were significantly correlated with BW, liver TC, and liver TBA; and BA metabolism-related genes were significantly correlated with BW, liver weight (LW)/BW, and liver TG (P < 0.05) (Fig. 2D). Therefore, the BA pathway was associated with BW, lipid levels, glucose metabolism, and the degree of hepatic steatosis in mice under MS conditions. Further protein level measurements showed that the levels of proteins involved in BA synthesis (CYP7A1, CYP27A1, and CYP7B1) were markedly elevated by Trp supplementation, whereas sterol 12α-hydroxylase (CYP8B1) exhibited no notable change. No significant alteration in protein levels was observed for FXR and its target gene SHP in the liver. Although ileum FXR protein expression was unchanged, protein expression of FGF15, a downstream molecule of FXR, was decreased, indicating that intestinal FXR–FGF15 signaling was inhibited (P < 0.05) (Fig. 2E and F). The protein expression of liver FGF15 was also decreased (P < 0.05) (Fig. S2A), thereby enhancing de novo BA synthesis.

Collectively, these results suggest that Trp ameliorates HFD-induced MS and is associated with enhanced BA metabolism in mice.

Gut microbiota is involved in BA metabolism, interacts with BAs, and influences BA composition. Therefore, we explored the effect of Trp on the gut microbiota composition in mice with MS through 16S rDNA gene sequencing. Principal coordinates analysis (PCoA) showed a separation in the gut microbiota structure between the HFD group and the Trp group. Alpha diversity analysis, including Shannon index and Simpson index, did not show significant differences, but phylogenetic diversity (PD) whole tree was significantly higher in the Trp group than in the HFD group. This indicates that the gut microbiota structure in mice with MS was affected by Trp (Fig. 3A). The predominant phyla in the fecal microbiota were Bacteroidetes, Firmicutes, Proteobacteria, and Actinobacteria (Fig. S3). To identify specific bacterial taxa or phylotypes at the species level and to ascertain the taxonomic differences of specific bacteria between HFD-fed and Trp-supplemented mice, a linear discriminative analysis (LEfSe) was conducted. At the phylum level, the cladogram revealed that many of the different bacteria in the Actinobacteria were enriched in the HFD-fed mice. At the genus level, the cladogram demonstrated that the HFD-fed mice exhibited 9 discriminative features, while the Trp-supplemented mice exhibited 4 (Fig. 3B). Spearman correlation analysis was further performed using these genus-level differential bacteria and fecal BAs to understand gut microbiota–BA interactions. Correlation analysis revealed that many non-12-OH BAs, including TCDCA, β-MCA, TMCA, and THDCA, have a strong negative correlation with Streptococcus, Lactobacillus, Parvibacter, Bifidobacterium, Acinetobacter, Gordonibacter, Enterorhabdus, and Coriobacteriaceae UCG-002, while they have a strong positive correlation with Eubacterium eligens group, Prevotella 9, Prevotella 1, and Catenibacterium (P < 0.05) (Fig. 3C). BSH enzymes deconjugate glycine- or taurine-conjugated BAs to unconjugated BAs. Since the abundance of 3 BSH-producing bacteria, Streptococcus, Lactobacillus, and Bifidobacterium, was significantly enriched in the HFD + Trp group, we quantified the abundance of the 5 BSH-producing bacteria (Streptococcus, Lactobacillus, Bifidobacterium, Enterococcus, and Lactococcus). Radar plots showed that the HFD + Trp group formed a smaller area and that Trp reduced their abundance, which is consistent with the elevation of conjugated BAs in the liver and feces (Fig. 3D). We also examined the colonic expression of genes downstream of aryl hydrocarbon receptor (AhR) (including Cypa1, Cyp1a2, and Cypb1) in response to Trp. Trp treatment significantly up-regulated Cypa1, Cyp1a2, and Cypb1 (Fig. S4), indicating AhR activation in colon tissues.

To investigate whether Trp treatment depends on the microbiota, mice were fed an HFD for 8 weeks and then gavaged with mixed antibiotic solutions for 1 week. Then, mice were orally gavaged with FMT or SFF from donor mice (donor mice were from the HFD + Trp group in the first mouse experiment), while the control group (HFD group) was orally gavaged with sterile saline. Gavage lasted for 5 weeks, during which time the HFD was maintained (Fig. 3E). As a result, the mice that received microbiota from HFD and Trp-treated mice or SFF from HFD and Trp-treated mice showed lower weight gain and serum glucose, TG, and TC concentrations compared to the HFD group. H&E and Oil Red O staining revealed smaller lipid droplets and less ballooning degeneration in the hepatocytes of FMT- or SFF-treated mice. Consistent with the pathological sections, liver TG and TC levels were also reduced after FMT and SFF (P < 0.05) (Fig. 3F to I). FMT and SFF alleviated MS similarly to their donors, suggesting that the ameliorative effects of Trp on MS depend on the gut microbiota. We further investigated the effects of microbiota on BA metabolism and found that liver TBA levels were greatly increased in FMT- and SFF-treated mice, although serum TBA levels remained unchanged. Consistent with the elevated liver TBA levels, the mRNA levels of BA synthesis genes Cyp7a1 and Cyp7b1 were considerably elevated by FMT and SFF. The mRNA level of liver Fxr and the protein levels of ileum FXR were unchanged, but the protein level of ileum FGF15 was markedly suppressed by FMT and SFF, indicating that FMT and SFF inhibit the intestinal FXR–FGF15 signaling pathway, enhancing de novo BA synthesis (P < 0.05) (Fig. 3J to L). Notably, It is worth noting that the impacts of improved MS caused by FMT and SFF depends on the microbes co-regulated by HFD and Trp, not just the Trp-regulated microbes.

The combination of these effects suggests that, under conditions of an HFD, gut microbiota play an essential and necessary role in inhibiting FXR signaling and improving BA metabolism induced by Trp. It is reasonable to assume that the gut microbiota–BA crosstalk exerts its ameliorative effects through the gut FXR–FGF15 axis.

To evaluate whether Trp alleviates MS through the intestinal Fxr pathway, mice were given different treatments starting from 8 weeks of HFD feeding: daily gavage of Gly-MCA (an intestine-specific FXR inhibitor), Trp intervention, and daily gavage of Fex (an intestine-specific FXR activator) in addition to the Trp intervention. The control group (HFD group) was left untreated, and the experiment lasted for 6 weeks, during which time all the mice were kept on the HFD (Fig. 4A). In the study of Fig. 1A, it can be observed that in the fourth week of the Trp intervention, there was a significant reduction in BW in the HFD + Trp group compared to the HFD group (Fig. 1B). This also means that Trp had a significant therapeutic effect on obesity after 4 weeks of intervention. Therefore, in the FXR regulation study, we changed the intervention period of Trp from 8 weeks to 6 weeks, which also ensured its therapeutic effect. In addition, the 14-week experimental period was consistent with the FMT or SFF experiment.

Western blot (WB) results confirmed that Gly-MCA treatment caused a strong inhibition of intestinal FGF15, and Trp treatment inhibited intestinal FGF15 similarly to Gly-MCA treatment, whereas Fex reversed the Trp-induced reduction in FGF15 protein levels (P < 0.05) (Fig. 4B). Consistent with changes in the FXR–FGF15 axis, Gly-MCA treatment and Trp treatment significantly increased serum and liver TBA levels, whereas Fex treatment reversed the Trp-induced increase in BA levels (P < 0.05) (Fig. 4C). In terms of gene expression levels of enzymes related to BA synthesis, Gly-MCA treatment promoted the mRNA expression levels of Cyp7a1, sterol 27-hydroxylas (Cyp27a1), and Cyp7b1. Trp treatment markedly promoted the mRNA expression levels of Cyp7a1 and Cyp7b1, whereas Fex reversed the promotion by Trp. Notably, Gly-MCA stimulated hepatic Fxr gene expression, which may be due to a compensatory effect of excessive inhibition of intestinal FXR–FGF15 axis (P < 0.05) (Fig. 4D). In terms of amelioration of MS, both Gly-MCA and Trp reduced body weight, serum glucose, TG, and TC levels in the liver and serum. H&E and Oil Red O staining showed that both Gly-MCA and Trp inhibited hepatic lipid accumulation. However, Fex reversed the ameliorative effect of Trp on MS, suggesting that Trp alleviates MS by improving BA metabolism through inhibition of intestinal FXR signaling (P < 0.05) (Fig. 4E to H).

In summary, Trp inhibits intestinal FXR signaling mediated by the gut microbiota–BA crosstalk, which in turn promotes de novo BA synthesis, thereby ameliorating MS.

Pigs are less susceptible to metabolic diseases due to their high levels of the unique beneficial BAs, HCA species. We next investigated the effect of Trp on BA metabolism in a pig model. Therefore, we chose the finishing pig model to further validate the role of Trp in improving lipid metabolism. We included 0.78% Trp in the basal diet of finishing pigs for 30 days and observed changes in growth performance as well as lipid metabolism (Fig. 5A). There were no significant differences in the initial and final weights between the 2 groups of finishing pigs. There were also no noticeable distinctions in average daily feed intake, average daily gain, and feed conversion ratio. However, Trp significantly reduced back fat thickness (P < 0.05) (Table). H&E and Oil Red O staining showed that Trp inhibited lipid accumulation in back fat, whereas this effect was not evident in the liver (Fig. 5B). Trp also reduced serum glucose and liver TC concentrations in finishing pigs (P < 0.05) (Fig. 5C and D).

Next, BA metabolism was measured. Trp had no significant effect on serum TBA but caused a significant increase in hepatic TBA. Trp markedly elevated the levels of numerous BA components in liver tissue, including taurocholic acid (TCA), DCA, glycodeoxycholic acid (GDCA), CDCA, glycochenodeoxycholic acid (GCDCA), lithocholic acid (LCA), ursodeoxycholic acid (UDCA), HCA, HDCA, THDCA, and GHDCA. Overall, 12-OH BAs, non-12-OH BAs, unconjugated BAs, conjugated BAs, and the ratio of non-12-OH to 12-OH BAs were significantly higher in Trp-supplemented finishing pigs (P < 0.05). Moreover, the proportion of HCA species has increased (Fig. 5E and F). Consistent with increased hepatic BA levels was a significant up-regulation of BA synthesis genes in the liver, including Cyp7a1, Cyp27a1, and Cyp7b1. The protein expression levels of intestinal FXR were unchanged, whereas FGF19 was significantly suppressed by Trp (P < 0.05) (Fig. 5G and H). The protein expression of liver FGF19 was also decreased (P < 0.05) (Fig. S2B), suggesting that the enhancement of hepatic BA synthesis is mediated by the intestinal FXR–FGF19 axis. Furthermore, Trp increased fecal TBA levels (Fig. S5), which is consistent with increased fecal TBA levels in the mouse model. This shows that Trp increases both hepatic BA synthesis and BA excretion in mice and finishing pigs. The levels of short-chain fatty acids (SCFAs) in the colonic contents were also measured. Radar plots showed that Trp remarkably raised levels of acetic acid and butyric acid, demonstrating that Trp improved the intestinal environment (P < 0.05) (Fig. 5I).

To investigate the influence of Trp on the hepatic lipid profile, we employed lipidomics to analyze the lipid profile of the liver. The total lipid concentration did not differ significantly (Fig. S6A). Orthogonal partial least squares discriminant analysis (OPLS-DA) revealed distinct lipid profiles between the 2 groups of pigs (Fig. 6A). A comparison of the levels of various lipid classes showed significant increases in wax esters (WEs) and zymosterol (ZyE) due to Trp supplementation (P < 0.05) (Fig. S6B and C). The volcano plot depicts changes in lipid molecules in the liver. Ten lipid molecules were significantly up-regulated [fold change (FC) > 1.5, P < 0.05], while 34 lipid molecules were significantly down-regulated (FC < 0.67, P < 0.05) by Trp. The different colors represent the classification of the lipid molecules, where the number of significantly different lipid molecules is higher in phosphatidylethanolamine (PE), phosphatidylserine (PS), phosphatidylcholine (PC), phosphatidylglycerol (PG), and ceramides (Cer) (Fig. 6B). PE, PS, PC, and PG are all glycerophospholipids (GPs). Twelve lipid molecules in PE, 7 in PS, 3 in PC, and 5 in Cer were clearly down-regulated. Additionally, 1 lipid molecule in PE and 3 in PG were clearly up-regulated (Fig. 6C). In order to characterize the intra-correlation of the same lipid and the correlations among the lipids of different classes, we analyzed the correlations of altered lipids in the liver. The results show that PE and PC have strong internal correlations not only within themselves but also with other different types of lipid molecules (Fig. S6D). Spearman correlation analysis was further conducted using these lipid molecules and hepatic BA species to explore the relationship between BAs and lipids in the liver. Correlation analysis showed that MCA species, CA species, HCA species, and DCA species are strongly negatively correlated with several PE, PS, PC, and Cer molecules but positively correlated with PG molecules (P < 0.05) (Fig. 6D). This suggests that BAs from these species are closely related to the hepatic lipid profile, particularly HCA species.

In conclusion, these data demonstrate that Trp inhibits intestinal FXR signaling and enhances BA synthesis, which in turn has a beneficial effect on lipid metabolism in finishing pigs.