(347d) Narrow Spectrum Antibiotic Treatment of Clostridium Difficile infection Improves Preservation and Restoration of Intestinal Metabolic Profile

Narrow spectrum antibiotic treatment of Clostridium difficileinfection improves preservation and restoration of intestinal metabolic profile

Karin Yanagi¹, Seoyoung Park¹, Xi Qian^2,3, Anne Kane⁴, Nicholas Alden¹, David Snydman⁴, Cheleste Thorpe⁴, Kyongbum Lee¹

1. Department of Chemical & Biological Engineering, Tufts University, Medford, MA
2. Department of Molecular Biology and Microbiology, Tufts University School of Medicine, Boston, MA
3. Clinical and Translational Science Institute, Tufts University, Boston, MA
4. Tufts Medical Center, Boston, MA, United States of America
5. Tufts University School of Medicine, Boston, MA, USA

Clostridium difficileis a spore-forming organism that can colonize the human intestine to cause disease, with severe, recurring cases leading to morbidity and mortality. Recurrence is Commensal organisms residing in the intestine have been shown to protect against C. difficileinfection (CDI) by producing metabolites that inhibit the germination and outgrowth of C. difficile. Metabolites that inhibit C. difficilespore germination include secondary bile acids, which are derived from liver-synthesized conjugated primary bile acids via deconjugation and biotransformation reactions catalyzed by bacterial enzymes. Treatment of CDI using a broad-spectrum antibiotic such as vancomycin (VAN) dramatically modifies the intestinal microbiota, which could deplete the intestine of commensal bacteria responsible for production of secondary bile acids and other metabolites that inhibit C. difficilespore germination. This dysbiosis can persist for several weeks after end-of-therapy (EOT), and has been associated with increased risk for recurrence.

In this study, we investigated the hypothesis that treating CDI with a more selective antibiotic reduces the collateral damage to the intestinal microbiota, improving preservation and restoration of a CDI-inhibitory metabolic profile. Stool samples from CDI patients treated with either a narrow- (ridinilazole, RDZ) or broad-spectrum antibiotic (VAN) were analyzed for microbiota composition using 16S rRNA sequencing, bile acids using targeted LC-MS assays, and global metabolite profiles using untargeted LC-MS experiments. The stool samples were collected both during treatment (day 1 and 10) and after EOT (day 25 and 40). A phase 2 clinical study showed that RDZ significantly reduced the rates of CDI recurrence compared to VAN. Patients treated with RDZ showed a significantly attenuated decrease in alpha diversity at EOT compared to patients treated with VAN. Taxonomically, RDZ treatment more narrowly altered the fecal microbiota, with modest reductions in Firmicutes, compared to broader and more sustained reductions in Firmicutes, Bacteroidetes, and Actinobacteria resulting from VAN treatment. Additionally, VAN treatment led to a dramatic increase in the relative abundance of Proteobacteria.

The bile acid analysis showed that RDZ-treated patients maintained significantly higher levels of secondary bile acids deoxycholate (DCA), ursodeoxycholate (UDCA), and hyodeoxycholate (HDCA) at EOT (day 10). Moreover, VAN treatment led to significant increases in the levels of conjugated primary bile acids taurocholate (TCA), taurochenodeoxycholate (TCDCA), and glycocholate (GCA) at EOT compared to baseline (day 1). Compositional analysis showed that VAN treatment led to a dramatic increase in the ratio of primary to secondary bile acids at EOT. By day 40 (30 days after EOT), this ratio improved to near healthy control levels in RDZ-treated patients, whereas the ratio remained elevated in VAN-treated patients.

We next investigated if the aberrant metabolite profile of VAN-treated patients (and to a lesser extent, RDZ-treated patients) is due to an overall reduction in bacterial abundance or due to more specific depletions of certain bacteria carrying out specialized metabolic functions. In samples from healthy controls, we did not find a significant correlation (Spearman’s rho) between secondary bile acids and Lachinospiraceae, a family of commensal gut bacteria that harbors the known operons encoding bile acid transformation enzymes required for secondary bile acid production. In contrast, samples from CDI patients showed significantly high positive correlation scores betweenLachinospiraceaeand secondary bile acids during and after antibiotic therapy. This suggests that the operon for secondary bile acid synthesis in Lachinospiraceaemay be more strongly induced in CDI patients, possibly due to the higher levels of primary bile acids. Interestingly, OTU counts of Enterobacteriaceaeare strongly correlated with secondary bile acids in the RDZ group, but not VAN, despite the large increase in the relative abundance of Enterobacteriaceaein the VAN group. This suggests that different antibiotics may influence intestinal bile acid metabolism by modulating the activities of bile acid enzymes, in addition to depleting the populations of bile acid metabolizing bacteria.

Untargeted analysis of fecal metabolites pointed to broad, functional differences in the intestinal microbiota of RDZ- and VAN-treated patients. At EOT, more than a quarter (27.5%) of almost 15,000 LC-MS features detected in both RDZ and VAN samples were differentially present in the two group (FDR corrected p-value < 0.05), and more than 83% of the differentially present features were elevated in the RDZ group, indicating a diminished capacity by the microbiota of VAN patients to generate diverse metabolic products. By day 40, broad differences persisted between the metabolic profiles of both RDZ- and VAN-treated patients compared to healthy controls; however, the profile of RDZ-treated patients were more similar to the healthy controls compared to VAN-treated patients as determined by principal component analysis (PCA). Annotation of the differentially present LC-MS features followed by pathway analysis found significant differences in purine, taurine, tyrosine, and bile acid metabolites in samples from RDZ- and VAN-treated patients. Consistent with the bile acid analysis, the VAN group showed a 5-fold decrease in free taurine, a major conjugation substrate of primary bile acids that is released by bacterial bile salt hydrolases (BSHs). Treatment with VAN also decreased most fermentation products of aromatic amino acids, including bacterial ligands of the aryl hydrocarbon receptor (AhR), a key regulator of intestinal immune cell activity. We also observed a depletion of amino sugars derived from mucin degradation. The major class of metabolites elevated in the VAN group are oligosaccharides, suggesting a diminished capacity for the intestinal microbiota to metabolize complex sugars. Interestingly, a recent study by Theriot et al. (2014) reported an increase in the levels of these sugars in the intestines of mice after antibiotic treatment during a CDI susceptible state.

Taken together, these findings suggest that RDZ treatment correlates with enhanced preservation and restoration of fecal bile acid composition as well as potential ligands of cellular receptors regulating intestinal immune function and substrates of bacterial metabolism. These metabolic profile differences between a narrow- and broad-spectrum antibiotic may contribute to their varying efficacy in preventing CDI recurrence. In ongoing work, we are testing the individual responses of bacterial strains harboring bile acid metabolizing enzymes to sublethal concentrations of narrow- and broad-spectrum antibiotics to investigate the hypothesis that antibiotics could directly regulate the metabolism of gut commensal bacteria.