Resveratrol Attenuates Trimethylamine- N -Oxide (TMAO)-Induced Atherosclerosis by Regulating TMAO Synthesis and Bile Acid Metabolism via Remodeling of the Gut Microbiota

In order to investigate the effect of RSV on TMAO synthesis in C57BL/6J mice, TMA and TMAO contents in serum were measured at 0, 1, 2, 4, 6 and 8 h after administration of a single dose of choline (400 mg/kg of body weight) or TMA (40 mg/kg) in mice by intragastric gavage. Via liquid chromatography-tandem mass spectrometry (LC/MS), we found that peak concentrations of TMA and TMAO occurred 4 h after choline administration and 1 h after TMA administration (Fig. 1A and B). Thus, in subsequent experiments we used serum samples taken at 4 h or 1 h after the administration of choline or TMA, respectively. Serum TMA and TMAO levels were much lower in RSV-treated mice given a single dose of choline than in those in the control group (Fig. 1C). For RSV-treated mice given TMA, TMAO levels were higher and TMA levels were lower than those in the control group (Fig. 1D), indicating that RSV may induce the metabolism of TMA to TMAO in the liver. Moreover, the addition of 1% choline over 30 days significantly increased the levels of TMA and TMAO in mice; however, this was reversed by RSV administration (Fig. 1E). Additionally, plasma RSV content was higher in mice fed RSV than in mice fed a standard diet (Fig. 1F). Our results suggested that RSV reduced TMAO synthesis levels in C57BL/6J mice.

TMAO generation is dependent on the gut microbiota, which can metabolize dietary choline to TMA. The TMA is then metabolized by enzymes of the FMO family in the liver, among which flavin monooxygenase 3 (FMO3) is the most active (23). Thus, we first examined the effects of RSV on FMO3 expression and activity in the liver. Treatment with RSV resulted in a significant increase in protein and mRNA levels of FMO3 (Fig. 2A and B). RSV also markedly increased FMO activity in the liver (Fig. 2C). These findings corresponded with our earlier results showing that RSV could increase TMAO levels in mice treated with a single dose of TMA (Fig. 1D). These results indicated that the RSV-induced decrease in TMAO levels was not due to its regulation of FMO3 in the liver.

To determine the role of the gut microbiota in the RSV-induced decrease of TMAO levels, we explored bacterial populations in the gut at the phylum and genus levels. Analysis at the phylum level revealed that the bacterial population of vehicle-treated mice was dominated by Bacteroidetes (35.1%) and Firmicutes (50.3%), with a low level of Proteobacteria (11.2%) (Fig. 2D). The proportion of sequences assigned to Bacteroidetes was significantly increased in metagenomes of RSV-fed animals at the expense of Firmicutes (Fig. 2D). Genus-level analysis showed that RSV induced an increase in the relative abundances of Bacteroides, Lactobacillus, Bifidobacterium, and Akkermansia in mice (Fig. 2E and F). RSV administration resulted in a decrease in the relative abundances of Prevotella, uncultured Ruminococcaceae (Ruminococcaceae_uncultured), Anaerotruncus, Alistipes, Helicobacter, and uncultured Peptococcaceae (Peptococcaceae_uncultured) (Fig. 2E and F). Using Pearson’s correlation tests, we found that, in general, taxa that were present at a relatively high proportion coincident with high plasma TMA concentrations also tended to be present at a relatively high proportion coincident with high plasma TMAO concentrations. Species of several bacterial taxa (Ruminococcaceae_uncultured, Prevotella, Bacteroides, Akkermansia, and Lactobacillus) remained significantly associated with plasma TMA and/or TMAO levels after false discovery rate (FDR) adjustment for multiple comparisons (Fig. 2G). Via a direct ex vivo assay, we found that the production of TMA from choline was lower in cecal content from RSV-treated mice and that the concentration of RSV in the cecal content was much higher than that in the plasma (Fig. 2H and I). Moreover, we found that the same mice fed with RSV could show markedly decreased TMA and TMAO levels after choline treatment; however, after washout with a chow diet for 1 month, they failed to reduce TMA and TMAO contents after choline administration (Fig. 2J). Additionally, real-time quantitative PCR (qPCR) analysis of suspected cecal microbes from mice treated with RSV for 2 months revealed that changes in the gut flora caused by RSV were similar to those observed in mice fed with RSV for 1 month (see Fig. S2A and B in the supplemental material). These results indicated that RSV remodeled the gut microbiome in mice, which might contribute to decreased gut microbial TMA production, thereby inhibiting TMAO synthesis.

The gut microbiota plays an important role in BA metabolism, by mediating deconjugation and 7-α-dehydroxylation of primary BAs (24). To assess the extent of BA biotransformation upon prebiotic colonization, we measured the effects of RSV on fecal BA excretion and composition. RSV significantly enhanced fecal BA loss, which was inhibited by Abs (Fig. 3A). It has been demonstrated that major secondary BAs are deoxycholic acid (DCA) and murideoxy cholic acid (MDCA), which are the 7-α-dehydroxylation products of cholic acid (CA) and beta-muricholic acid (β-MCA), respectively (25). Here, we found that the fecal CA/DCA ratio was unchanged by RSV administration, indicating that RSV had no effect on 7-α-dehydroxylation activity in the gut (Fig. 3B; see also Table S4 in the supplemental material). In contrast, there was a significant decrease in the fecal conjugated/unconjugated BA ratio, along with an increase in bile salt hydrolase (BSH) enzyme activity, in the feces of RSV-treated mice (Fig. 3C and D; see also Table S4). Moreover, we found that RSV treatment resulted in an increase in the proportions of Lactobacillus and Bifidobacterium that were present in the gut (Fig. 2E and F), likely contributing to its effects on BSH activity. Of note, administration of RSV resulted in a significant reduction in small-intestine (SI) BA content, including taurocholic acid (TCA), TβMCA, CA, chenodeoxycholic acid (CDCA), and DCA, in which the decrease in levels of TCA, CA, CDCA, and DCA might explain nearly 90% of the lower SI BA content (see Fig. S2C and Table S4). It has been demonstrated that the apical sodium bile acid transporter (ASBT) is responsible for ileal BA uptake, with a preference for conjugated over unconjugated BAs (26). We assessed mRNA and protein levels of ASBT and found that RSV had no significant influence on ASBT mRNA or protein levels (see Fig. S2D and E), suggesting that the RSV-mediated increase in fecal BA loss was independent of the presence of ASBT. Finally, we observed a reduction in the mRNA levels of the BA transporters organic solute transporter alpha (OSTα) and beta (OSTβ) in RSV-treated mice, which might contribute to the RSV-induced decrease in ileal BA content (see Fig. S2F). These results suggested that RSV-induced modulation of the gut microbiome contributed to increased BSH activity and subsequently promoted the generation of unconjugated BAs from conjugated BAs, thereby enhancing fecal BA loss.

It was reported previously that enhanced fecal BA loss is accompanied by enhanced hepatic BA neosynthesis (27, 28). We found that RSV significantly decreased the liver cholesterol level and increased the BA pool size and the gallbladder and SI luminal BA content; these effects were reversed by treatment with Abs (Fig. 4A to D). No differences were observed with respect to serum and hepatic BA content (Fig. 4E and F). In addition, the expression of CYP7A1 mRNA and protein, which is the rate-limiting enzyme in hepatic BA synthesis, was also investigated (29). As shown in Fig. 4G to I and in Table S4 in the supplemental material, RSV increased the bile TCA/TβMCA ratio and induced CYP7A1 mRNA and protein expression and the effects were inhibited by Abs treatment. These results suggested that RSV possibly increased hepatic BA synthesis by inducing CYP7A1 expression in a gut microbiota-dependent manner.

The feedback suppression of CYP7A1 gene transcription by FXR is one of the most important mechanisms in the maintenance of BA homeostasis (30). Previous studies have shown that FXR activation is a major mechanism in the suppression of BA synthesis and induces small heterodimer partner (SHP) and FGF15, which work together to suppress CYP7A1 (31, 32). Thus, we sought to determine the possible roles of the gut-liver FXR signaling in RSV-induced BA synthesis. RSV treatment resulted in a significant decrease in ileal FGF15 mRNA and protein levels; however, ileal FXR mRNA and protein levels were unchanged (Fig. 5A and B). These observations suggested that RSV downregulated FXR transcriptional activity, as demonstrated by the reduction in mRNA levels of its target genes. In the liver, RSV treatment appeared to have no significant effect on the mRNA level of the SHP gene and the mRNA and protein levels of FXR (see Fig. S3A and B in the supplemental material). These findings highlighted that RSV downregulated the ileal FXR-FGF15 axis, which might result from RSV-induced significant reduction in FXR agonists (TCA, DCA, CA, and CDCA) in SI tissues (see Table S4). Furthermore, we used Z-Guggulsterone (Z-Gug), a natural antagonist of FXR, to suppress FXR in mice as outlined previously (33). Treatment with Z-Gug markedly inhibited the expression of FGF15, thereby inducing CYP7A1 expression in mice (Fig. 5C and D). Additionally, similar results were seen in mice cotreated with RSV and Z-Gug (Fig. 5C and D). The presence of synthetic FXR agonist GW4064 was able to reverse RSV-induced alterations in the expression of FGF15 and CYP7A1 in mice (Fig. 5E and F). Taken together, these observations indicated that the gut-liver FXR-FGF15 axis played a key role in RSV-induced BA synthesis.

Plasma TMAO is strongly associated with AS (2). Therefore, we investigated whether changes in TMAO induced by RSV through microbiota remodeling were associated with AS regression in ApoE^−/− mice. The atherosclerotic lesion area and the cholesterol content in the whole aorta, as indexes of AS severity, were much higher in mice fed with choline than in chow-fed mice, and the effects were markedly reversed by RSV or Abs administration (Fig. 6). Moreover, when the gut microbiota was suppressed by Abs, RSV-induced inhibition of TMAO-caused AS was significantly abolished (Fig. 6). Additionally, RSV single treatment could also attenuate the pathogenesis of AS in ApoE^−/− mice (Fig. 6). These results suggested that RSV attenuated TMAO-induced AS in a gut microbiota-dependent manner.

As shown in Fig. 7A to D, RSV administration markedly decreased TMA and TMAO contents and increased the expression levels of liver FMO3 and FMO activity. These results were consistent with our findings in C57BL/6J mice, suggesting the important role of the gut microbiota in the RSV-induced decrease of TMAO levels.

The composition of the gut microbiota in cecal content was then determined. Via qPCR and sequencing 16S rRNA assays, we found that choline administration markedly increased the bacterial concentrations of Bacteroidetes phyla and decreased the concentrations of Firmicutes phyla compared to the control group (see Fig. S4A in the supplemental material). The Proteobacteria and Actinobacteria phyla showed no significant changes after choline treatment in ApoE^−/− mice (see Fig. S4A). Moreover, compared to chow-fed ApoE^−/− mice, choline significantly increased the number of Prevotella and decreased the concentrations of Bacteroides, Lactobacillus, and Bifidobacterium (see Fig. S4B). These results indicated that chronic choline supplementation might alter cecal microbial composition, thereby enhancing the synthesis of TMA and TMAO and ultimately increasing AS. As expected, treatment with RSV resulted in an increase in Bacteroidetes abundance (from 20.6 to 34.0%) at the expense of Firmicutes (from 60.1 to 50.1%) in choline-treated ApoE^−/− mice (Fig. 7E). RSV also increased the relative abundances of Bacteroides, Lactobacillus, Bifidobacterium, and Akkermansia and decreased the relative abundances of Prevotella, Ruminococcaceae_unclassified, and Biophila in choline-fed ApoE^−/− mice (Fig. 7F and G). After FDR adjustment for multiple comparisons, the abundance of genera Prevotella and Bacteroides was still significantly associated with TMAO and TMA levels in choline-treated ApoE^−/− mice (Fig. 7H). Additionally, via a direct ex vivo assay, we found that the production of TMA from choline was higher in cecal content from choline-treated mice and that the effect was reversed by RSV administration (Fig. 7I). Combined with the results seen in C57BL/6J mice, we postulated that RSV attenuated TMAO-induced AS by reducing TMAO synthesis via inhibition of the gut microbial TMA production.

TMAO appears to contribute to the development of AS, in part by affecting cholesterol metabolism through the regulation of the BA synthesis pathway (4). Therefore, the influence of RSV on BA synthesis was investigated in choline-treated ApoE^−/− mice. We found that the presence of choline significantly increased serum total cholesterol (TC) and liver cholesterol levels and decreased BA pool size and gallbladder and SI luminal BA content in ApoE^−/− mice; these effects were reversed by treatment with RSV (Fig. 8A to E). And there were no significant differences in liver BA content among the groups (Fig. 8F). Meanwhile, RSV also increased the bile TCA/TβMCA ratio (see Fig. S5A and Table S4 in the supplemental material) and reversed the choline-induced decrease of CYP7A1 mRNA and protein expression in ApoE^−/− mice (Fig. 8G and H). However, when the gut microbiota was suppressed by Abs, RSV-induced hepatic BA synthesis was significantly inhibited in choline-fed ApoE^−/− mice (Fig. 8). Additionally, RSV single treatment could also induce hepatic BA synthesis in ApoE^−/− mice (Fig. 8). These results suggested that RSV induced hepatic BA synthesis in a gut microbiota-dependent manner in choline-treated ApoE^−/− mice.

Finally, the effects of RSV on gut-liver FXR signaling were determined in choline-treated ApoE^−/− mice. The mRNA level of the SHP gene and the mRNA and protein levels of FXR in the liver were unchanged by RSV (Fig. 9A and B). The ileal FGF15 mRNA and protein levels were markedly reduced by RSV; however, ileal FXR mRNA and protein levels were unchanged in choline-fed ApoE^−/− mice (Fig. 9C and D). In addition, RSV had no effect on ASBT expression but still decreased the SI BA content, including TCA, CDCA, and TβMCA, in choline-fed ApoE^−/− mice (Fig. 9D and E; see also Table S4 in the supplemental material). RSV also increased fecal BA loss and fecal conjugated/unconjugated BA levels and increased BSH activity, with fecal CA/DCA ratios unchanged in choline-fed ApoE^−/− mice (Fig. 9F; see also Fig. S5B to D and Table S4). These results were similar to those found in C57BL/6J mice. Overall, we came to the conclusion that RSV attenuated TMAO-induced AS, partially by regulating BA synthesis through the enterohepatic FXR-FGF15 axis in a gut microbiota-dependent manner.

Female C57BL/6J mice and ApoE^−/− mice with a C57BL/6 genetic background were purchased from Jackson Laboratory (Bar Harbor, ME, USA). Animals were maintained at a controlled temperature (22 ± 2°C), with a 12-h light/dark period. For all the experiments, mice were housed individually in cages and had ad libitum access to water. Mice of the control group were fed a standard chow diet (NIH31 modified mouse/rat diet; Harlan Teklad). Animal experiments were carried out in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals published by the National Institutes of Health and were approved by the Animal Care and Use Committee of the Third Military Medical University (Chongqing, China; Approval SYXC-2014-00112). Pentobarbital sodium anesthesia (50 mg/kg of body weight) was administered prior to surgical procedures, with maximal effort exerted to minimize suffering. At the end of the experiments, mice were killed by cervical dislocation followed by decapitation.

Quantitative data are presented as means ± standard deviations (SD) of results from three independent experiments. Statistical analysis was conducted with either Student’s t test or one-way analysis of variance using SPSS 13.0 (SPSS Inc., Chicago, IL). A P value less than 0.05 was considered statistically significant, and the Tukey-Kramer post hoc test was applied if the P value was less than 0.05. Correlations between host parameters and bacterial populations were assessed by Pearson’s correlation tests using GraphPad Prism version 6.00 (GraphPad Software, La Jolla, CA). FDR in the multiple comparisons were estimated for each taxon based on the P values that resulted from correlation estimates in R statistical software.

Full descriptions of additional materials and methods are given inin the supplemental material. Text S1