Antibiotic-Induced Alterations of the Gut Microbiota Alter Secondary Bile Acid Production and Allow for Clostridium difficile Spore Germination and Outgrowth in the Large Intestine

Groups of C57BL/6 mice were treated with various antibiotics (cefoperazone, clindamycin, vancomycin, metronidazole, and kanamycin), and some groups were allowed to recover off of the antibiotic cefoperazone for up to 6 weeks in order to create different microbial and metabolic environments in the small intestine (ileum) and the large intestine (cecum) (Fig. 1). At the time of necropsy (depicted by red circles in Fig. 1), ileal and cecal contents were collected for ex vivoC. difficile spore germination and outgrowth assays. After a 6-h incubation, C. difficile VPI 10463 spores (approximately 10⁴ spores/ml) were able to germinate and initiate outgrowth in ileal content from non-antibiotic-treated mice and almost all antibiotic-treated mice (Fig. 2A). Spores added to phosphate-buffered saline (PBS [negative control]) did not germinate or outgrow due to the lack of germinant present. The only two metabolic environments from ileal content that were able to initiate germination and outgrowth of spores but failed to reach significance were from mice allowed 3 and 4 weeks of recovery after cefoperazone.

Unlike ileal content, C. difficile spores were only able to germinate and outgrow in cecal content from mice treated with specific antibiotics (Fig. 2B). We previously demonstrated non-antibiotic-treated mouse cecum does not support spore germination or outgrowth ex-vivo, whereas cefoperazone-treated mouse cecum does (12, 31). The only metabolic environments that allowed for significant spore germination and outgrowth in the mouse cecum ex vivo were directly after cefoperazone, 1 and 2 weeks of recovery after cefoperazone, 6 weeks of recovery after cefoperazone plus clindamycin, clindamycin alone, and vancomycin.

Since bile acids are required for C. difficile spore germination and are also able to inhibit outgrowth (22, 32), we defined the bile acids present in paired ileal and cecal samples from ex vivo germination and outgrowth assays. Prior to antibiotics, ileal content contains mostly primary bile acids, specifically taurine-conjugated and unconjugated primary bile acids, with a lesser concentration of unconjugated secondary bile acids (Fig. 3A). Primary bile acids taurocholate (TCA) and cholate (CA), both germinants of C. difficile spores, are always present at high concentrations in the ileum (average concentrations in vivo, TCA, 0.03%, and CA, 0.05%). This is in contrast to the cecum, where unconjugated secondary bile acids make up the majority of bile acids prior to antibiotics (Fig. 3B). Antibiotic treatments that supported C. difficile spore germination and outgrowth ex vivo in Fig. 2B (directly after cefoperazone, 1 and 2 weeks of recovery after cefoperazone, 6 weeks of recovery after cefoperazone plus clindamycin, clindamycin alone, and vancomycin) resulted in significantly more TCA and less secondary bile acids in the cecum (Fig. 3B and susceptible in panel C; see Fig. S1 and S2 in the supplemental material). Murine cecal contents that did not support C. difficile spore germination and outgrowth are highlighted in the black boxes in Fig. 3B and are associated with an increase in unconjugated secondary bile acids (resistant in Fig. 3C) (average concentrations in vivo, ωMCA, 0.004%, HDCA, 0.002%, UDCA, 0.004%, LCA, 0.001%, and DCA, 0.023%).

We previously demonstrated using untargeted metabolomics that directly after cefoperazone treatment there was a significant decrease in unconjugated secondary bile acids in the murine cecum, and 6 weeks later, the levels returned to baseline or prior to antibiotic treatment (12). Here we show using a more sensitive targeted bile acid assay that secondary bile acids start to return to baseline levels within 2 weeks of recovery of cefoperazone, and by 6 weeks of recovery, they return to baseline levels (see Fig. S3 in the supplemental material).

To define the relationship between relevant in vivo levels of secondary bile acids and colonization resistance to C. difficile in Fig. 2, we tested the ability of these bile acids to inhibit or enhance the first step of C. difficile colonization, spore germination. Previous studies have shown TCA is required for maximal spore germination, and chendeoxycholate (CDCA) is able to inhibit this interaction (26). Figure 4A shows maximal spore germination with TCA (0.1%) alone and significant inhibition of spore germination with the addition of CDCA (0.04%). Since most bile acids are soluble in ethanol (ωMCA, HDCA, UDCA, and LCA) a positive control of TCA made with ethanol was used for these comparisons. Secondary bile acids that significantly interfered with TCA-mediated spore germination in a concentration-dependent manner were ωMCA (0.001% and 0.004%) and LCA (0.001% and 0.01%), whereas HDCA (0.0001% and 0.001%) enhanced spore germination at lower concentrations.

Since the concentration of TCA (average concentration in vivo TCA, 0.0001%) is not high enough to support spore germination in the murine cecum during resistance to C. difficile, we turned our attention to the secondary bile acid DCA, which is highly abundant during resistance to C. difficile (average concentration in vivo DCA, 0.023%) and is a known germinant of C. difficile spores (12, 22). In Fig. 4B, spore germination with DCA (0.1%) alone does not reach the same maximal level of spore germination as TCA (0.1%) alone (Fig. 4A). The addition of CDCA (0.04%) significantly interfered with DCA-mediated spore germination as well as supplementation with ωMCA (0.001% and 0.004%). Secondary bile acids present in the cecum during resistance to C. difficile (ωMCA and LCA) were able to inhibit germination of C. difficile spores.

To further characterize the dynamics between the pathogen and secondary bile acids, we evaluated growth of C. difficile over a 24-h period in vitro in the presence of relevant in vivo concentrations of secondary bile acids. In media supplemented with secondary bile acids HDCA, UDCA, LCA, and DCA, C. difficile exhibited decreased growth rates compared to controls (C. difficile alone and C. difficile with ethanol added and no bile) without the addition of secondary bile acids. (Ethanol was added to a C. difficile culture to account for the ethanol in the bile acids.) C. difficile growth curves with secondary bile acids that significantly inhibited the growth rate from Fig. 5A are displayed in Fig. 5B (see Fig. S4 in the supplemental material). Secondary bile acids present during resistance to C. difficile were able to inhibit growth of C. difficile.

Since many secondary bile acids shape resistance to C. difficile, and members of the gut microbiota are responsible for the biotransformation of primary bile acids to secondary bile acids, we defined the microbiome of each ileal and cecal sample (see Fig. S5 and S6 in the supplemental material). To examine the relationship between members of the gut microbiota and bile acids, we calculated the Spearman’s rank correlation coefficient for all operational taxonomic units (OTU) within the microbiome and bile acids using data from the mouse ileum and cecum (Fig. 6). To visualize these correlations, we performed unsupervised clustering of OTU and bile acids from all groups, which revealed three distinct OTU clusters (O1 to O3 in Table S1 in the supplemental material) and two bile acid clusters (B1 and B2). The organization of the correlation revealed distinctive relationships between OTU and bile acids in the different groups. OTU in the first OTU cluster (O1) were positively correlated with bile acids in the first bile acid cluster (B1) and negatively correlated with most of the bile acids in cluster B2, which contains all of the secondary bile acids (black boxes in Fig. 6). The OTU in cluster O1 include many members from the Proteobacteria and Firmicutes phyla, more specifically from the Enterobacteriaceae and Lactobacillaceae families (see Table S1). Many of these bacteria were present in samples that supported C. difficile germination and outgrowth. This is in contrast to the relationship between clusters O2 and B2, which has a positive correlation. O2 is made up of members from the Firmicutes phylum: specifically, these Lachnospiraceae and Ruminococcaceae family members are positively correlated with all secondary bile acids (see Table S1). Many of these bacteria were present in samples that provided resistance against C. difficile germination and outgrowth. Cluster O3 is made up of many members from the Bacteroidetes phylum, from the Porphyromonadaceae family (see Table S1) and is positively correlated with all bile acids from clusters B1 and B2.

The University Committee on the Use and Care of Animals (UCUCA) at the University of Michigan approved this study. The University of Michigan laboratory animal care policies follow the Public Health Service Policy on Humane Care and Use of Laboratory Animals (59). Animals were assessed twice daily for physical condition and behavior, and those assessed as moribund were humanely euthanized by CO₂ asphyxiation. Trained animal technicians performed animal husbandry in an AAALAC-accredited facility.

Five- to 8-week-old C57BL/6 wild-type (WT) mice (male or female) were obtained from a breeding colony at the University of Michigan that was originally established using animals purchased from Jackson Laboratories (Bar Harbor, ME). Mice were housed with autoclaved food, bedding, and water. Cage changes were performed in a laminar flow hood. Mice had a 12-h cycle of light and darkness.

Groups of C57BL/6 WT mice (male and female) ranging from 5 to 8 weeks in age were treated with a variety of antibiotics (Fig. 1). Groups of mice (n = 7 to 10 per group) were given cefoperazone (0.5 mg/ml) in sterile drinking water for 10 days followed by a recovery period of 2 days and 1, 2, 3, 4, 5, and 6 weeks with regular drinking water. Three additional groups of mice were treated with a 5-day course of vancomycin (0.5 mg/ml), metronidazole (0.5 mg/ml), or kanamycin (0.5 mg/ml) in their drinking water followed by a 2-day recovery period with regular drinking water. Mice treated with clindamycin were given a subcutaneous injection of 1 mg/ml clindamycin solution so that each mouse received 10 mg/kg of body weight followed by a 1-day recovery. An additional group of mice was given an injection of clindamycin after 6 weeks of recovery from a 10-day treatment with cefoperazone (0.5 mg/ml) followed by a 1-day recovery after the clindamycin injection. Non-antibiotic-treated controls for this experiment included mice given sterile drinking water during the duration of these experiments. After each treatment, mice were euthanized, and at the time of necropsy, ileal and cecal contents were collected, flash-frozen in liquid nitrogen, and stored at −80°C until further analysis.

A two-step extraction of gut content with ethanol and 1:1 chloroform-methanol was done with isotope-labeled internal standards spiked in. Extraction solvent was added to gut samples for homogenization with a probe sonicator. Homogenized samples were kept on ice for 10 min and centrifuged at 45°C at 13000 rpm for 10 min. Supernatants containing metabolites were transferred to another tube, and pellets were extracted with 1:1 chloroform-methanol. Supernatants from the two steps were combined and dried under vacuum at 45°C by using a vacuum centrifuge. Dried samples were reconstituted in 50:50 methanol-water for LC-MS analysis. A series of calibration standards were prepared and analyzed along with samples to quantify metabolites.

An Agilent 1290 liquid chromatograph (Agilent, Santa Clara, CA) and waters Acquity BEH C₁₈ column with 1.7-µm particle size (2.1 by 100 mm; Waters, Milford, MA) was used for chromatographic separation. Mobile phase A (MPA) was 0.1% formic acid in LC-MS-grade water. Mobile phase B (MPB) was 0.1% formic acid in acetonitrile. The mobile phase gradient consisted of 0 to 0.5 min of 5% MPB, 3 min of 25% MPB, 17 min of 40% MPB, and 19 to 21 min of 95% MPB and was run on an Agilent 6490 triple-quadrupole mass spectrometer (Agilent, Santa Clara, CA). Data were processed by MassHunter workstation software, version B.06. Table 1 lists the bile acids detected by name and abbreviation as well as their MRM transitions (i.e., parent m/z→daughter m/z).

C. difficile VPI 10463 spores were prepared as described in previous studies (12, 60).

C. difficile spores were subjected to heat treatment (65°C for 20 min) prior to use for these studies. Mouse ileal and cecal contents (non-antibiotic treated and antibiotic treated) were weighed and passed into an anaerobic chamber, diluted at a 1:1 ratio with PBS, and spun down for 5 s in a microcentrifuge. Approximately 10³C. difficile spores (10⁴ spores/ml) were added to diluted ileal content, cecal content, and a PBS-only control at a final volume of 0.1 ml and then incubated at 37°C anaerobically for 6 h. After incubation, bacterial enumeration was done on taurocholate-cycloserine-cefoxitin-fructose agar (TCCFA) selective medium (12). The remaining ileal and cecal contents were heat treated at 65°C for 20 min to kill off any vegetative cells, and a subsequent bacterial enumeration was done. Samples plated prior to heating represent spores and vegetative cells (germination and outgrowth). Samples plated after heat treatment represent the spores remaining.

Spores were subjected to heat treatment (65°C for 20 min) prior to use for these studies. Sterile bile acid solutions dissolved in either water or ethanol were prepared and passed into an anaerobic chamber and added to brain heart infusion (BHI) plus 100 mg/liter l-cysteine broth with either 0.1% taurocholate (TCA) or 0.1% deoxycholate (DCA) as a positive control. C. difficile spores were added to broth and allowed to incubate for 30 min at 37°C anaerobically. Bacterial enumeration of the samples was performed on both BHI plates (vegetative cells) and BHI agar supplemented with 0.1% TCA (total CFU). Since 0.04% chendeoxycholate (CDCA) is a known inhibitor of TCA-mediated spore germination, this was used as a negative control in this assay (32). TCA at 0.1% and DCA at 0.1% made in water or ethanol (10%) were used as positive controls of spore germination. The percentage of germination of the positive controls TCA and DCA alone was calculated as [(CFU on BHI)/(CFU on BHI + TCA or DCA)] × 100. Inhibition of germination with TCA with the addition of secondary bile acids ωMCA, HDCA, UDCA, LCA, and DCA was calculated as [(CFU on BHI + TCA + secondary bile acids)/(CFU on BHI + TCA alone)] × 100. Inhibition of germination with DCA with the addition of secondary bile acids ωMCA, HDCA, UDCA, and LCA was calculated as [(CFU on BHI + DCA + secondary bile acids)/(CFU on BHI + DCA alone)] × 100. Germination under each condition was performed in triplicate. Controls were always used and compared to the proper bile acid backgrounds of water or ethanol.

C. difficile VPI 10463 was cultured overnight at 37°C in BHI plus 100 mg/liter·l-cysteine broth in an anaerobic growth chamber (Coy labs). After 14 h of growth, C. difficile was subcultured 1:10 into BHI plus 100 mg/liter·l-cysteine and allowed to grow for 3 h anaerobically at 37°C. The culture was then diluted in fresh BHI so that the starting optical density at 620 nm (OD₆₂₀) was 0.01. The cell suspension was added to a 96-well plate at a final volume of 0.2 ml. Sterile bile acid solutions were added to the wells in triplicate. Cell density (OD₆₂₀) was monitored every 30 min for 24 h, shaking the plate for 90 s before each reading, in a Tecan plate reader (reference no. 16039400) inside an anaerobic chamber. The specific growth rate was determined as μ = ln(X/X_o)/T, where X is the OD₆₂₀ value during the linear portion of growth and T is time in hours. Values given are representative of the mean μ values from two independent cultures done in triplicate. The growth curve data in Fig. S5 in the supplemental material start at an OD of 0.1 because BHI medium alone has a starting OD₆₂₀ of 0.1.

The bile acids used in all in vitro assays include taurocholic acid sodium salt hydrate (TCA) (Sigma-Aldrich, CAS no. 345909-26-4), sodium deoxycholate (DCA) (Sigma-Aldrich, CAS no. 302-9504), chenodeoxycholic acid (CDCA) (Alfa Aesar, CAS no. 474-25-9), lithocholic acid (LCA) (Sigma-Aldrich, CAS no. 434-13-9), ursodeoxycholic acid (UDCA) (Sigma-Aldrich, CAS no. 128-13-2), ω-muricholic acid (ωMCA) (Steraloids, catalog ID no. C1888-000), hyocholic acid (HCA) (Steraloids, CAS no. 547-75-1), and hyodeoxycholic acid (HDCA) (Steraloids, CAS no. 83-49-8). TCA and DCA were soluble in water, and the rest of the bile acids were soluble in ethanol.

Microbial DNA was extracted from mouse ileal and cecal tissue snips that also included luminal content using the PowerSoil-htp 96-well soil DNA isolation kit (Mo Bio Laboratories, Inc.). The V4 region of the 16S rRNA gene was amplified from each sample using a dual-indexing sequencing strategy (61). Each 20-µl PCR mixture contained 2 µl of 10× Accuprime PCR buffer II (Life Technologies), 0.15 µl of Accuprime high-fidelity Taq (catalog no. 12346094) high-fidelity DNA polymerase (Life Technologies), 2 µl of a 4.0 µM primer set, 1 µl DNA, and 11.85 µl sterile double-distilled water (ddH₂O) (free of DNA, RNase, and DNase contamination). The template DNA concentration was 1 to 10 ng/µl for a high bacterial DNA/host DNA ratio. PCR was performed under the following conditions: 2 min at 95°C, followed by 30 cycles of 95°C for 20 s, 55°C for 15 s, and 72°C for 5 min, followed by 72°C for 10 min. For low-biomass samples (ileal content), a “touchdown PCR” protocol was performed. Each 20-µl PCR mixture contained 2 µl of 10× Accuprime PCR buffer II (Life Technologies), 0.15 µl of Accuprime high-fidelity Taq (catalog no. 12346094) high-fidelity DNA polymerase (Life Technologies), 2 µl of 4.0 µM primer set, 1 µl DNA, and 11.85 µl sterile ddH₂O (free of DNA, RNase, and DNase contamination). The template DNA concentration was 1 to 10 ng/µl for a high bacterial DNA/host DNA ratio. PCR was performed under the following conditions: 2 min at 95°C, followed by 20 cycles of 95°C for 20 s, 60°C for 15 s, and 72°C for 5 min (with a 0.3°C increase of the 60°C annealing temperature each cycle), followed by 20 cycles of 95°C for 20 s, 55°C for 15 s, and 72°C for 5 min, followed by 72°C for 10 min. Libraries were normalized using a Life Technologies SequalPrep normalization plate kit (catalog no. A10510-01) following the manufacturer’s protocol. The concentration of the pooled samples was determined using the Kapa Biosystems library quantification kit for Illumina platforms (KapaBiosystems KK4854). The sizes of the amplicons in the library were determined using the Agilent Bioanalyzer high-sensitivity DNA analysis kit (catalog no. 5067-4626). The final library consisted of equal molar amounts from each of the plates, normalized to the pooled plate at the lowest concentration.

Sequencing was done on the Illumina MiSeq platform, using a MiSeq reagent kit V2 with 500 cycles (catalog no. MS-102-2003) according to the manufacturer’s instructions, with modifications (61). Libraries were prepared according to Illumina’s protocol for preparing libraries for sequencing on the MiSeq (part 15039740 Rev. D) for 2 or 4 nM libraries. The final load concentration was 4 pM (but it can be up to 8 pM) with a 10% PhiX spike to add diversity. Sequencing reagents were prepared according to Illumina’s protocol for 16S sequencing with the Illumina MiSeq personal sequencer (61). (Updated versions of this protocol can be found at http://www.mothur.org/wiki/MiSeq_SOP .) Custom read 1, read 2, and index primers were added to the reagent cartridge, and FASTQ files were generated for paired-end reads.

Analysis of the V4 region of the 16S rRNA gene was done using mothur (version 1.33.3) (61, 62). Briefly, the standard operating procedure (SOP) at http://www.mothur.org/wiki/MiSeq_SOP was followed to process the MiSeq data. The paired-end reads were assembled into contigs and then aligned to the SILVA 16S rRNA sequence database (63, 64) and were classified to the mothur-adapted RDP training set v9 (65) using the Wang method and an 80% bootstrap minimum to the family taxonomic level. All samples with <500 sequences were removed. Chimeric sequences were removed using UCHIME (66). Sequences were clustered into operational taxonomic units (OTU) using a 3% species-level definition. The OTU data were then filtered to include only those OTU that made up 1% or more of the total sequences. The percentage of relative abundance of bacterial phyla and family members in each sample was calculated. A cutoff of 0.03 (97%) was used to define operational taxonomic units (OTU) and used to calculate the inverse Simpson index as a measure for diversity. Standard packages in R were used to create average heat maps per mouse treatment group and changes in the inverse Simpson index per treatment group (see Fig. S5 and S6 in the supplemental material).

We computed the Spearman’s rank correlation coefficients (r) for all pairs among the 26 bile acids and 121 OTU (Fig. 6) using “method = ‘spearman,’ use = ‘pairwise.complete.obs’” in R package as seen in reference 12.

Statistical tests were performed using Prism version 6.00d for Mac Os X (GraphPad Software, La Jolla, CA, USA). Significance between groups of ex vivo spore germination and outgrowth studies was done by Mann-Whitney nonparametric t test (Fig. 2). Significance between controls and all other groups in in vitro germination and growth studies were calculated by Student’s parametric t test (Fig. 4 and 5). The Kruskal-Wallis one-way analysis of variance test followed by Dunn’s multiple comparison test was used to calculate the significance of bacterial diversity of non-antibiotic-treated controls compared to all other groups in Fig. S6 in the supplemental material. Statistical significance was set at a P value of <0.05 for all analyses.