The microbiota-derived bile acid taurodeoxycholic acid improves hepatic cholesterol levels in mice with cancer cachexia

Abstract

Alterations in bile acid profile and pathways contribute to hepatic inflammation in cancer cachexia, a syndrome worsening the prognosis of cancer patients. As the gut microbiota impinges on host metabolism through bile acids, the current study aimed to explore the functional contribution of gut microbial dysbiosis to bile acid dysmetabolism and associated disorders in cancer cachexia. Using three mouse models of cancer cachexia (the C26, MC38 and HCT116 models), we evidenced a reduction in the hepatic levels of several secondary bile acids, mainly taurodeoxycholic (TDCA). This reduction in hepatic TDCA occurred before the appearance of cachexia. Longitudinal analysis of the gut microbiota pinpointed an ASV, identified as Xylanibacter rodentium, as a bacterium potentially involved in the reduced production of TDCA. Coherently, stable isotope-based experiments highlighted a robust decrease in the microbial 7α-dehydroxylation (7α-DH) activity with no changes in the bile salt hydrolase (BSH) activity in cachectic mice. This approach also highlighted a reduced microbial 7α-hydroxysteroid dehydrogenase (7α-HSDH) and 12α-hydroxysteroid dehydrogenase (12α-HSDH) activities in these mice. The contribution of the lower production of TDCA to cancer cachexia was explored in vitro and in vivo. In vitro, TDCA prevented myotube atrophy, whereas in vivo hepatic whole transcriptome analysis revealed that TDCA administration to cachectic mice improved the unfolded protein response and cholesterol homeostasis pathways. Coherently, TDCA administration reversed hepatic cholesterol accumulation in these mice. Altogether, this work highlights the contribution of the gut microbiota to bile acid dysmetabolism and the therapeutic interest of the secondary bile acid TDCA for hepatic cholesterol homeostasis in the context of cancer cachexia. Such discovery may prove instrumental in the understanding of other metabolic diseases characterized by microbial dysbiosis. More broadly, our work demonstrates the interest and relevance of microbial activity measurements using stable isotopes, an approach currently underused in the microbiome field.

Keywords: 12-alpha-hydroxysteroid dehydrogenase; 7-alpha-dehydroxylation activity; 7-alpha-hydroxysteroid dehydrogenase; Cancer cachexia; Xylanibacter; bile acids; cholesterol; colorectal cancer; gut microbiota; liver; muscle atrophy.

Graphical abstract

Figure 1.

Hepatic bile acid profiling revealed alterations in bile acid composition in C26, MC38 and HCT116 cachectic mice as compared to their respective CT mice. n = 4–8 mice per group, data are presented as mean ± SEM, *p < 0.5, **p < .01 and ***p < .001 vs CT groups. NQ, not quantified. Panel a has been previously published in Thibaut et al, J Cachexia Sarcopenia Muscle 2021, these data points are therefore indicated with empty symbols.

Figure 2.

Hepatic gene expression levels in C26, MC38 and HCT116 cachectic mice as compared to their respective CT mice. Genes involved in bile acid synthesis (a) and hepatobiliary transport system (b). Cytochrome P450 family 7 sub‐family a member 1 (Cyp7a1), cytochrome P450 family 8 sub‐family B member 1 (Cyp8b1), cytochrome P450 family 27 sub‐family a member 1 (Cyp27a1) and cytochrome P450 family 7 sub‐family B member 1 (Cyp7b1), Na+/taurocholate transport protein (Ntcp), bile salt export pump (Bsep), multidrug resistance‐associated protein 2 (Mrp2), organic solute transporter subunit beta (Ostb). n = 4–8 mice per group, data are presented as mean ± SEM, *p < 0.05, **p < 0.01 and ***p < 0.001 vs CT groups.

Figure 3.

Gut microbiota composition (at the ASV level) according to the different stages in the progression of cachexia in cachectic C26 mice. (a) Experimental design. (b) Body weight, food intake evolution and hepatic TDCA levels at 8, 9 and 10 days after C26 cell injection (C26) or sham injection (CT). Data are presented as mean ± SEM, *p < 0.05, **p < 0.01, ***p < 0.001 vs CT groups. (c) Principal component analysis at the ASV level (PERMANOVA, R² = 36%, p = 0.001). (d) Venn diagram generated from ASV significantly changed in the C26 groups on days 8, 9 and 10 (q-value <0.05), compared with their respective CT groups. (e) Relative abundance of the 6 ASVs that were significantly changed in the C26 groups on days 8, 9 and 10. Data are presented as mean ± SEM, * q < 0.05, **q < 0.01 vs CT groups. (f) Spearman correlations between the relative abundance of ASVs that were significantly changed in the C26 groups on days 8, 9 and 10, and their hepatic levels of TDCA (C26 groups). n = 7–8 mice per group. Panel B has been previously published in Thibaut et al, J Cachexia Sarcopenia Muscle 2021, these data points are therefore indicated with empty symbols.

Figure 4.

Gut microbiota composition (at the family level) according to the different stages in the progression of cachexia in cachectic C26 mice. (a) Principal component analysis at the family level (PERMANOVA, R² = 65%, p = 0.001). (b) Barplots representing the relative abundance of families in sham-injected mice (CT groups) and mice receiving an injection of C26 cancer cells (C26 groups) on days 8, 9 and 10. (c) Relative abundance of families that were significantly changed in the C26 groups on days 8, 9 and 10. n = 7–8 mice per group, data are presented as mean ± SEM, *q < 0.05 and **q < 0.01 vs CT groups.

Figure 5.

Reduced microbial 7α-dehydroxylation activity in cachectic C26 mice. (a) Deconjugation of d₄-taurocholic acid (d₄-TCA) by the fecal bile salt hydrolases into d₄-cholic acid (d₄-CA) (left), followed by the caecal 7α-dehydroxylation activity into deoxycholic acid (d₃-DCA) (right) in sham-injected mice (CT) or mice injected with C26 cells (C26). (b) Detection of deuterated epimerization intermediates in caecal content of CT and C26 mice after incubation of d₄-CA in anaerobic conditions at 37°C during 15 min. n = 6–12 mice per group, data are presented as mean ± SEM, **p < 0.01 and ***p < 0.001 vs CT group. HSDH, hydroxysteroid dehydrogenase. ECA, epicholic acid. ND, not detected.

Figure 6.

TDCA protected C₂C₁₂ myotubes against atrophy caused by dexamethasone. (a) Comparison of myotube diameter (µm) between the control myotubes (CT) and myotubes incubated with 50 µm of taurodeoxycholic acid (TDCA). (b) Comparison of myotube diameter between the CT myotubes and myotubes incubated with 1 µm of INT-777 (specific TGR5 agonist). (c) Comparison of myotube diameter between the 1 µm dexamethasone (DEXA) and vehicle-treated myotubes, myotubes incubated with 1 µM of DEXA and 50 µm of TDCA, and myotubes incubated with 1 µM of DEXA and 100 µm of TDCA. (d) Phase contrast microscopy pictures of myotubes treated with the vehicle (CT), treated with 1 µm DEXA, incubated with 1 µM of DEXA and 50 µm of TDCA, or incubated with 1 µM of DEXA and 100 µm of TDCA. Scale bar = 200 µm. Data are presented as mean ± SEM, representative of 3 independent experiments performed in triplicates (N = 3, n = 3). **p < 0.01, ***p < 0.001 vs CT or vs DEXA.

Figure 7.

TDCA reduced liver weight in cachectic C26 mice. (a) Body weight and food intake evolution in sham-injected mice (CT), mice injected with C26 cells (C26), and mice injected with C26 cells and treated with 10 mg/kg/day of TDCA (C26-TDCA), expressed in % of initial body weight (% iBW) or % of initial food intake (% iFI). Two-way anova with Tukey’s posttests. (b) Tumor, gastrocnemius (GAS), tibialis anterior (TIB) and liver weights in CT, C26 and C26-TDCA mice. One-way Anova with Dunnett’s posttests, C26 group as the reference group. n = 7–8 mice per group, data are presented as mean ± SEM, *p < 0.05, **p < 0.01 and ***p < 0.001 vs C26 group.

Figure 8.

TDCA improved cholesterol homeostasis in cachectic C26 mice. (a) Significantly modified gene pathways between C26 and CT mice and between C26-TDCA and C26 mice using gene set enrichment analysis. (b) Heatmap of the 50 differentially regulated genes in C26-TDCA mice versus C26 mice (abs(L2FC) > 1, padj < 0.05). Among the 48 induced genes, 1 gene was induced (dark red cluster), 18 were not affected (light red cluster) and 29 were repressed (green cluster) in C26 mice compared to CT mice. 1 gene downregulated by TDCA was not affected in C26 mice (light blue cluster) while 1 gene downregulated by TDCA was also downregulated in C26 mice (dark blue cluster). (c) Hepatic cholesterol levels in CT, C26 and C26-TDCA mice. One-way anova with Dunnett’s posttests, C26 group as the reference group. n = 6–8 mice per group, data are presented as mean ± SEM, *p < 0.05, **p < 0.01 and ***p < 0.001 vs C26 group.