Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2022 Sep 1.
Published in final edited form as: Aliment Pharmacol Ther. 2021 Jul 4;54(6):792–804. doi: 10.1111/apt.16496

Decreased Secondary Faecal Bile Acids in Children with Ulcerative Colitis and Clostridium difficile Infection

Sarah Rotondo-Trivette 1, Beibei Wang 2,3, Christopher Gayer 1,4, Riddhi Parsana 4, Yihui Luan 3, Fengzhu Sun 2, Sonia Michail 1,4
PMCID: PMC8384671  NIHMSID: NIHMS1713680  PMID: 34218431

Abstract

Background:

Patients with ulcerative colitis (UC) have an increased risk of Clostridium difficile infection (CDI). There is a well-documented relationship between bile acids and CDI.

Aims:

To evaluate faecal bile acid profiles and the gut microbial changes associated with CDI in children with UC.

Methods:

This study was conducted at Children’s Hospital Los Angeles. Faecal bile acids and gut microbial genes related to bile acid metabolism were measured in 29 healthy children, 23 children with mild to moderate UC without prior CDI (UC group), 16 children with mild to moderate UC with prior CDI (UC+CDI group), and 10 children without UC with prior CDI (CDI group).

Results:

Secondary faecal bile acids, especially lithocholic acid (3.296 versus 10.793, p <= 0.001) and ursodeoxycholic acid (7.414 versus 10.617, p <= 0.0001), were significantly lower in children with UC+CDI when compared to UC alone. Secondary faecal bile acids can predict disease status between these groups with 84.6% accuracy. Additionally, gut microbial genes coding for bile salt hydrolase, 7α-hydroxysteroid dehydrogenase, and 7α/β-dehydroxylation were all diminished in children with UC+CDI compared to children with UC alone.

Conclusions:

Bile acids can distinguish between children with UC based on their prior CDI status. Bile acid profile changes can be explained by gut microbial genes encoding for bile salt hydrolase, 7α-hydroxysteroid dehydrogenase and 7α/β-dehydroxylation. Bile acid profiles may be helpful as biomarkers to identify UC children who have had CDI and may serve as future therapeutic targets.

Keywords: Pediatrics, ulcerative colitis, Clostridium difficile infection, primary bile acids, secondary bile acids

Graphical Abstract

graphic file with name nihms-1713680-f0001.jpg

Introduction

Ulcerative Colitis (UC), a subtype of inflammatory bowel disease (IBD), is characterized by chronic inflammation of the colon1. It is an important pediatric disease, as 20–25% cases present during childhood. The incidence of UC is constantly rising with alarming increases in pediatrics2–4.

Patients with UC are at an increased risk of developing Clostridium difficile infection (CDI), have higher rates of colectomy and death from CDI, and experience higher rates of recurrence5. UC patients with CDI also tend to be younger and have less prior antibiotic exposure5. Most cases of CDI in these patients represent outpatient acquired infections5. While it is known that CDI is transmitted via ingestion of spores, it is unknown why UC patients are at an increased risk of developing CDI. Interestingly, the relationship between bile acids and C. difficile spore germination has been well documented in vitro6–8. Primary bile acids promote germination of C. difficile spores, whereas secondary bile acids generally inhibit germination and growth9. Primary bile acids, cholic acid (CA) and chenodeoxycholic acid (CDCA), are synthesized by the human liver10. Bile acids are further conjugated with taurine and glycine. Conjugated primary bile acids are then secreted into the gallbladder and released into the duodenum with meals. Most bile acids are then re-absorbed in the terminal ileum in the process of enterohepatic circulation. Conjugated bile acids that travel to the large intestine are deconjugated by bile salt hydrolases (BSH)10. After deconjugation, bile acids may be transformed by the gut microbiota into secondary bile acids via 7α/β-dehydroxylation or 7α/β-hydroxysteroid dehydrogenation (HSDH).11 These transformations can be visualized in Figure 1.

Figure 1.

Figure 1

Open in a new tab

Intestinal bile acid transformations as described by Ridlon et al.10 Circled enzymes/genes are decreased in children with UC+CDI compared to UC alone, see results section for details.

Secondary bile acids include ursodeoxycholic acid (UDCA), lithocholic acid (LCA), and deoxycholic acid (DCA). A portion of these secondary bile acids are reabsorbed in the colon and will be returned to the liver where they will be conjugated again before being re-released into the small intestine. The rest will be lost in feces.

It has been previously demonstrated that while adult patients with UC have similar total faecal bile acids, they have a lower proportion of faecal secondary bile acids compared to healthy subjects.1 Additionally, total faecal secondary bile acids are lower in adult patients with active CDI compared to healthy controls11. However, faecal bile acids have not been previously described in children with both UC and a history of CDI.

In the present study, faecal bile acids were measured in healthy children, children with a history of CDI, children with mild to moderate UC, and children with mild to moderate UC and a history of CDI. We present novel findings of the differences in faecal bile acids and gut microbial genes related to bile acid metabolism between children with UC and children with UC+CDI. This could help to elucidate the mechanism behind the increased risk of CDI in UC patients and explore future therapeutic options targeting bile acid metabolism in this population.

Methods

This study was conducted at Children’s Hospital Los Angeles (CHLA) under IRB # CCI-11–00148. The subjects were recruited by the investigator and study team during their regular clinic visit, while inpatient, or through the department list from January 2014 to June 2019.

After providing informed consent, patients were asked to provide a stool sample and complete a questionnaire. Stool collected was immediately placed on ice and transferred to CHLA within 24 hours. At CHLA, stool was stored at −80 degrees Celsius. Data collected in the questionnaire included: gender, month & year of birth, race, weight, height, medications, dietary intake, and if the child is healthy or has any disease. In this study, we examined the stool of healthy patients (defined as absence of a medical diagnosis or symptoms such as abdominal pain/diarrhea), patients with a history of UC, patients with a history of CDI (defined as positive C. difficile PCR test, most patients also had a positive toxin assay confirmation), and patients with a history of both UC and CDI.

Bile Acid Extraction

Bile acid extraction was performed at the University of California, Davis using the procedure previously described by La Merrill et al12. Briefly, 15mg stool samples were kept frozen in 2-mL Eppendorf tubes until used. After recording exact weights, 10 μL of antioxidant solution, 0.2 mg/mL solution BHT/EDTA in 1:1 methanol:water, was added. Next, 10 μL of bile acid surrogate was added. Lastly, 500 μL of cold methanol and stainless steel grinding balls were added to the tubes. The samples were then homogenized using Geno/Grinder (2×30 sec) and centrifuged. Supernatant was transferred to a 1.5 mL Eppendorf tube containing 10 μL 20% glycerol solution in methanol. Meanwhile, an addition 500 μL of cold methanol was added to the centrifugation pellet; homogenization and centrifugation was repeated in the same manner as above. The second resulting supernatant was combined with the first in the 1.5 mL Eppendorf tube. Vials were transferred to Speed-vac and evaporated to dryness before being reconstituted in 100 μL of PHAU/CUDA 100nM in methanol/ACN 50:50, vortexed for 10 seconds, and sonicated for 5 minutes. The rack of samples sat on wet ice for 15 minutes before being centrifuged for 3 minutes at highest speed. Supernatant was transferred to glass insert in amber HPLC vials and stored at −20 degrees C until Liquid Chromatography/Mass Spectroscopy (LCMS). Bile acids were finally isolated and quantified using LCMS on Thermo Vanquish UPLC/AB Sciex Qtrap with targeted multiple reaction monitoring method.

Microbial DNA Extraction

Faecal DNA was extracted using the QIAamp Power faecal DNA kit (Qiagen, Germantown MD) by following the manufacturer’s instruction. Vortex-Genie 2 with a horizontal tube holder adaptor was used for mechanical lysis of faecal cells. The DNA quantity and quality were confirmed using Nanodrop.

16s rRNA Amplicon Sequencing

The 16s bacterial DNA V4 region from stool DNA and negative controls were amplified by PCR using uniquely barcoded primers (515FB: 5′-GTG YCA GCM GCC GCG GTA A-3′; 806RB: 5′-GGA CTA CNV GGG TWT CTA AT-3′)13 which were modified from the original 515F-806R primer pairs14. Each PCR sample was run in triplicate. Library qualities were assessed on Agilent High Sensitivity DNA Bioanalyzer chips. All samples were sequenced using custom sequencing primers: Read 1 (5′-TAT GGT AAT TGT GTG YCA GCM GCC GCG GTA A-3′), Read 2 (5′-AGT CAG CCA GCC GGA CTA CNV GGG TWT CTA AT-3′), and Index (5′-AAT GAT ACG GCG ACC ACC GAG ATC TAC ACG CT-3′). Paired-end sequencing (2 × 150bp) was performed using Illumina MiSeq Reagent Kit v2 flowcell on an Illumina MiSeq System.

Whole Genome Metagenomic Sequencing

Metagenomic libraries were constructed using the Nextera XT DNA Library Preparation Kit (Illumina) and Illumina Nextera XT Index v2 Kit A and B following the manufacturer’s protocols. Library qualities were assessed on Agilent High Sensitivity DNA Bioanalyzer chips. Libraries were pooled and sequenced on an Illumina NextSeq500 High Output v2 flowcell on an Illumina NextSeq 500 System, producing 2×150bp paired-end reads. All samples and control reads were pre-processed, and quality filtered using trim_galore (https://www.bioinformatics.babraham.ac.uk/projects/trim_galore/). Host-derived reads were removed using KneadData (https://github.com/biobakery/kneaddata). The number of reads in each sample ranged from 26 million to 2 million. The mean reads of the samples was 11 million with standard deviation of 5.12. The histogram of the numbers of reads in the samples is given in supplementary Figure S2.

Functional Metagenomics Analysis

The functional assignment of metagenomic data, particularly the presence and abundance of BSH (EC 3.5.1.24), 7α-HSDH, 7β-HSDH, and bai genes, was explored using CLC Genomic19 workbench version 20.0.4 (CLC, Bio-Qiagen, Aarhus, Denmark). An average of 11 million reads per sample were trimmed and assembled using the “De Novo Assemble Metagenomic” tool. The resulting average number of contigs was 74,000 (of length >200bp). The binning of the contigs was carried out using Bin Pangenomes by sequence tool in CLC Genomic workbench. To identify gene and coding DNA sequences (CDS) on contigs, Annotate CDS with Best DIAMOND HIT tool was used. DIMOND index was created in CLC using customized BSH, 7α-HSDH, 7β-HSDH, and bai gene databases downloaded from UniProtKB. The input reads were mapped back to the annotated contigs using the map reads to reference tool and DIMOND hit functional profiles were created. The functional abundance profiles for all samples were merged into one profile using the Merge Abundance table tool.

Statistical Analysis

The ranges of bile acid concentrations as well as BSH, 7α-HSDH, 7β-HSDH, and bai gene abundances were very large, therefore we used logarithm to transform the data. We added 10e-5 to zero to avoid infinite values when doing the log transformation. Outliers were removed from analysis. Q-Q plots of log-transformed concentrations were used to see whether the data followed a normal distribution. Shapiro-Wilk tests were used to further test whether the data was normally distributed. If data was normally distributed, we used the ANOVA to evaluate statistically significant differences and student’s t tests for pairwise comparison. If data was not normally distributed, Kruskal-Wallis tests and Wilcoxon rank sum tests were used instead. Measurements with p-value less than 0.05 were declared as significant.

We also predicted disease status of UC versus UC+CDI using the concentrations of each individual deconjugated bile acid. A patient was predicted to have UC if their log-transformed value was greater (or smaller) than threshold t and predicted as UC+CDI if their value was less (or greater) than threshold t. The threshold t was chosen by giving the highest accuracy. The accuracy was defined as the total number of correct predictions over the total number of individuals.

Results

Demographics

Key patient information and demographics can be found in Table 1. We included 29 healthy children, 23 children with mild to moderate UC, 10 children with a history of CDI, and 16 children with mild to moderate UC and a history of CDI. Again, heathy was defined as an absence of a medical diagnosis or symptoms such as abdominal pain/diarrhea. The average age of the 73 children included was approximately 15 years and they were 55% female. There was no statistically significant difference in the age of the children in each group, as seen in Figure 2. Healthy children had an average of 0.86 bowel movements per day, while children with a history of CDI alone had an average of 1.90 (p = 0.0045). Children with UC had an average of 4.39 bowel movements per day, while children with UC and a history of CDI had an average of 3.85 (p = 0.5729).

Table 1:

Demographics

Healthy CDI UC UC+CDI
Number of patients 29 10 23 16
Age (average in years) 14.1 11.2 15.4 15.2
Gender (% males) 48% 60% 52% 56%
Average PUCAI score (p=0.7322) NA NA 37 34
Average duration of UC (years) (p=0.3432) NA NA 2.62 1.98
Average number of daily bowel movements (p=9.5e-12) 0.86 1.90 4.39 3.85
Active CDI NA None NA None
Average number of past CDI (p=0.8215) NA 3.5 NA 3.25
Current therapy for CDI None None None None
Number of patients with CDI within the last 6 months (p=0.6364) NA 6 (60%) NA 8 (50%)
Use of antibiotics within 24 months of faecal sample collection (excluding CDI therapy) None 10 (100%) 3 (13.04%) 3 (18.75%)
Number of patients taking proton pump inhibitors (p=0.0678) None 3 (30%) 1 (4.35%) 1 (6.25%)
Number of patients taking biologic therapy (p=0.6597) None None 16 (69.57%) 10 (62.5%)
Number of patients taking immunomodulators (p=0.1597) None None 5 (21.74%) 1 (6.25%)
Number of patients taking 5-aminosalicylates (p=0.3455) None None 4 (17.39%) 5 (31.25%)
Number of patients currently taking corticosteroids (p=0.4559) None None 5 (21.74%) 2 (12.5%)
Open in a new tab

Figure 2.

Figure 2

Open in a new tab

Age distribution for CDI, Healthy, UC, and UC+CDI groups using Kolmogorov-Smimov tests. Pairwise p values were all greater than 0.05, indicating subjects in different groups had similar age distribution.

Of the children with UC, the average pediatric ulcerative colitis index (PUCAI) scores were approximately 36 and the average duration of disease was 2.3 years. UC patients were taking medications such as biologics, immunomodulators, 5-aminosalicylates and corticosteroids. Differences between the rates at which these medications were used between UC alone and UC+CDI can also be observed in Table 1, though there were no significant differences between these groups.

Of the children with a history of CDI, they had experienced an average of 3.35 recurrences of CDI. None of the children had active CDI during this study or were currently taking CDI therapy. The last episode of CDI was within the last 12 months but had been resolved for at least 1 month prior to stool collection. However, 60% of CDI patients and 50% of UC+CDI patients had CDI within the last 6 months. It is worth noting that 16 patients had taken antibiotics, other than CDI therapy, in the 24 months prior to stool collection. However, only 2 patients had non-CDI antibiotics within 6 months prior to stool collection. The most common non-CDI antibiotic was overwhelming amoxicillin, followed by a cephalosporin. It is also worth noting that no subjects in this study have a history of liver disease or have taken UDCA.

Bile Acids

First, we compared the differences in total primary, total secondary and the ratio of secondary to primary among the four groups as seen in Figure 3A–C. The totals included conjugated and deconjugated forms of each bile acid. Consistent with prior studies1, UC subjects had a significantly reduced ratio of secondary to primary bile acids than healthy subjects (0.713 versus 4.192, p <= 0.0001). CDI and UC+CDI had an even further reduced ratio (−0.532 versus 4.192 and −1.357 versus 4.192, p <= 0.0001). Interestingly, CDI, healthy and UC subjects had no significant differences in total secondary bile acids, but UC+CDI subjects had significantly fewer total secondary bile acids than all other groups (10.618 versus 13.144, 13.839, and 13.615, p <= 0.001). Healthy subjects had the least total primary bile acids (9.584, p <= 0.001), while CDI had the highest (14.169, p <= 0.01). UC and UC+CDI had similar levels of total primary bile acids (12.575 versus 12.776, p > 0.05).

Figure 3.

Figure 3

Open in a new tab

Boxplots of the log transformed concentrations (picograms/milligram feces) of total primary bile acids (A), total secondary bile acids (B), the ratio of total secondary to total primary (C), total conjugated bile acids (D), total deconjugated bile acids (E), and total overall bile acids (F) for CDI, Healthy, UC, and UC+CDI groups. Pairwise comparisons were performed using students’ t test (Wilcoxon rank sum test for ratio comparisons). P values obtained by ANOVA (Kruskal-Wallis test for ratio comparison) were included, with ns (not significant) for p>0.05, * for p <= 0.05, ** for p <= 0.01, *** for p <= 0.001, and **** for p <= 0.0001.

We then compared levels of conjugated, deconjugated and total bile acids between the four groups: Healthy, UC, CDI, UC+CDI. As shown in Figure 3D–F, we found that UC+CDI had the lowest levels of total faecal bile acids (12.866, p <= 0.05), CDI had the highest (14.613, p <= 0.05), and there was no significant difference between healthy and UC (13.851 versus 14.027, p > 0.05). Additionally, we found that UC+CDI had the lowest levels of total deconjugated bile acids (10.529, p <= 0.001) while all other groups were not significantly different. CDI had the highest levels of total conjugated bile acids (14.142, p <= 0.05), while healthy had the lowest (9.699, p <= 0.01). Total conjugated bile acids were not significantly different between UC and UC+CDI (11.375 versus 11.860, p > 0.05).

Relative concentrations of all forms (deconjugated, glycine conjugated, and taurine conjugated) of each bile acid (CA, CDCA, DCA, LCA, and UDCA) in each of the 4 groups can be found in Figure S1.

Differences between UC and UC+CDI bile acid profiles:

As previously stated, we found that UC+CDI patients had significantly fewer secondary bile acids and deconjugated bile acids than children with UC alone. In order to understand what was driving this difference, we compared the total (Figure 4A1–A5) and deconjugated forms (Figure 4B1–B5) of each individual bile acid. We found that LCA and UDCA were the primary drivers of this difference.

Figure 4.

Figure 4

Open in a new tab

Boxplots of the log transformed concentrations (picograms/milligram feces) of total (conjugated and deconjugated) CA (A1), CDCA (A2), DCA (A3), LCA (A4), UDCA (A5), and deconjugated CA (B1), CDCA (B2), DCA (B3), LCA (B4), UDCA (B5) for CDI, Healthy, UC, and UC+CDI groups. P values obtained by Kruskal-Wallis tests were included, with ns (not significant) for p>0.05, * for p <= 0.05, ** for p <= 0.01, *** for p <= 0.001, and **** for p <= 0.0001.

To assess whether the decreased LCA and UDCA were simply a result of increased disease activity, we also created a scatter plot of PUCAI scores in relation to LCA and UDCA concentrations for both patients with UC and UC+CDI (Figure 5). In our analysis, R2 values were close to 0, and F tests were not significant, indicating PUCAI and LCA or UDCA did not have a linear correlation. Therefore, an increase in PUCAI score would not explain a decrease in LCA or UDCA.

Figure 5.

Figure 5

Open in a new tab

Scatter plot for the log-transformed abundance changes of LCA (red) or UDCA (green) along with PUCAI for UC and UC+CDI patients. Linear fitted curve, the Goodness of Fit (R2), as well as the p-value from F test are shown. PUCAI and the log LCA or log UDCA didn’t have linear correlations.

In Figure 6, we assessed whether time since last CDI would affect LCA or UDCA concentration in patients with UC+CDI. We found that there was no significant difference in LCA or UDCA concentrations between UC+CDI patients who had CDI within the last 6 months versus UC+CDI patients who did not have CDI within the last 6 months. There were no significant differences between the two groups.

Figure 6.

Figure 6

Open in a new tab

Log transformed LCA and UDCA concentrations in patients with UC+CDI who had CDI <6 months ago vs >6 months ago. P values obtained by Wilcoxon rank sum tests are included.

Lastly, in Figure 7, we created a scatter plot to assess whether there was a relationship between the number of CDI recurrences and LCA or UDCA concentrations in UC+CDI patients. Again, R2 values were close to 0 and F tests were not significant, indicating the number of CDI recurrences and LCA or UDCA did not have a linear correlation. Therefore, an increase in the number of CDI recurrences would not explain a decrease in LCA or UDCA.

Figure 7.

Figure 7

Open in a new tab

Scatter plot for the log-transformed abundance changes of LCA (red) or UDCA (green) along with the number of CDI recurrences for UC+CDI patients. Linear fitted curve, the Goodness of Fit (R2), as well as the p-value from F test are shown. The number of CDI recurrences and the LCA or UDCA did not have linear correlations.

Predicting past CDI in UC patients

After completing these analyses, we then attempted to predicted UC versus UC+CDI using the concentrations of each individual deconjugated bile acid. Table 2 indicates that the prediction accuracy of using LCA is up to 0.846. UDCA had similar results, with prediction accuracy of 0.821. This means LCA or UDCA concentration in stool could be used as a criterion to determine whether a UC patient had CDI in the past or not. Interestingly, combining LCA and UDCA did not increase the prediction accuracy compared to each bile acid alone.

Table 2:

Prediction results for UC versus UC+CDI using the concentration of single deconjugated bile acids. The threshold t is chosen by giving the highest accuracy. A person is predicted to be UC if its log-transformed value is greater than t and predicted as UC+CDI if its value is less than t.

Input Features Threshold Accuracy
CA 8.346 0.718
CDCA 1.546 0.59
DCA 6.496 0.744
LCA 4.505 0.846
UDCA 7.868 0.821
LCA+UDCA 0.897 0.597
Open in a new tab

Differences between UC and UC+CDI gut microbial genes

In order to understand why UC+CDI patients had fewer deconjugated bile acids and decreased LCA and UDCA concentrations, we compared bacterial genes related to bile acid metabolism between UC and UC+CDI. Again, please refer to Figure 1 to visualize the transformations involved in bile acid metabolism.

BSH is required to deconjugate bile acids from glycine or taurine prior to being further metabolized to secondary bile acids. We compared the level of BSH in Figure 8A. UC+CDI patients were found to have significantly less BSH than all other groups (p <= 0.01).

Figure 8.

Figure 8

Open in a new tab

Boxplot of the log transformed abundance of bile salt hydrolase (BSH) (A), 7α-hydroxysteroid dehydrogenase (HSDH) (B), and 7β-hydroxysteroid dehydrogenase (HSDH) (C) for CDI, Healthy, UC, and UC+CDI groups. P values obtained by Kruskal-Wallis tests were included, with ns (not significant) for p>0.05, * for p <= 0.05, ** for p <= 0.01, *** for p <= 0.001, and **** for p <= 0.0001.

We also compared 7α-HSDH (Figure 8B) and 7β-HSDH (Figure 8C), as these are responsible for the conversion of CDCA to UDCA.20 We found 7α-HSDH was reduced in UC+CDI (p <= 0.05), however there was no significant difference between UC and UC+CDI in 7β-HSDH. As this is a two-step process, reduction in just one of these genes could account for the decreased UDCA in UC+CDI.

In order to convert CA to DCA and CDCA to LCA, primary bile acids must undergo microbially mediated 7α-dehydroxylation. Some bacteria are also capable of 7β-dehydroxylation, which converts UDCA to LCA. Both processes involve multiple genes and have been previously described by Ridlon et al10. BaiG, baiB, and baiA are common to both processes. BaiCD and baiE are stereospecific for 7α-dehydroxylation, while baiH and baiI are stereospecific for 7β-dehydroxylation. As seen in Figure 9, we found that baiCD, baiE, baiG, and baiI had no significant difference between UC and UC+CDI, while baiA, baiB, and baiH were significant decreased in UC+CDI compared to UC (p <= 0.05). As baiA and baiB are common to both pathways, this demonstrates that both 7α-dehydroxylation and 7β-dehydroxylation are decreased in UC+CDI, however the decrease in baiH demonstrates that 7β-dehydroxylation is even further reduced.

Figure 9.

Figure 9

Open in a new tab

Boxplot of the log transformed abundance of baiA (A), baiB (B), baiCD (C), baiE (D), baiG (E), baiH (F), and baiI (G) for CDI, Healthy, UC and UC+CDI groups. P values obtained by Kruskal-Wallis tests were included, with ns (not significant) for p>0.05, * for p <= 0.05, ** for p <= 0.01, and *** for p <= 0.001.

Bacterial Genera

Using 16S sequencing, we were able to compare the levels of certain bacterial genera involved in bile acid metabolism. As seen in Figure 10, certain BSH producing genera, including Bacteroides, Blautia, Dorea, Enterococcus, Finegoldia, Roseburia, and Ruminococcus, are all significantly reduced in UC+CDI compared to UC (p <= 0.05), Interestingly, Blautia, Dorea, Roseburia and Ruminococcus also produce the genes involved in the 7α-dehydroxylation. There were no other genera that we measured that showed significant differences between UC and UC+CDI.

Figure 10.

Figure 10

Figure 10

Open in a new tab

Boxplot comparing the log transformed abundances of bile acid metabolizing bacteria for CDI, Healthy, UC and UC+CDI groups at genus level. P values obtained by Kruskal-Wallis tests were included, with ns (not significant) for p>0.05, * for p <= 0.05, ** for p <= 0.01, and *** for p <= 0.001.

Discussion

In this study, we describe novel findings of faecal bile acids in pediatric patients with UC who have a history of CDI (UC+CDI). We found that their bile acid profile is distinct from those who have UC without prior CDI, those who are healthy and those who have CDI without UC. We also delineate the gut microbiome genomic differences in bile acid metabolite pathways that explain the variations seen in the bile acids among these groups. We demonstrate that secondary faecal bile acids, especially LCA and UDCA are significantly reduced in UC+CDI compared to UC alone (3.296 versus 10.793, p <= 0.001) and (7.414 versus 10.617, p <= 0.0001)

As described in the introduction, the relationship between bile acids and C. difficile spore germination has been well documented in vitro6–8. Primary bile acids promote germination of C. difficile spores, whereas secondary bile acids generally inhibit germination and growth9. Our finding that secondary bile acids, especially LCA and UDCA, are significantly reduced may provide insight into why UC patients are at an increased risk of CDI even without traditional risk factors for developing the infection, such as advanced age and antibiotic exposure, however this may also simply be a consequence of past infection.5

We also found that deconjugated faecal LCA and UDCA levels could predict which UC patients had CDI in the past with up to 84.6% accuracy. This means that LCA and UDCA may be useful biomarkers for determining which UC patients have had CDI in the past. In future studies, this may provide insight into which UC patients have a gut microbiome conducive to CDI and therefore have an increased risk of developing CDI. Identifying UC patients most at risk for developing CDI may be helpful to clinicians so that strategies to minimize future infection can be implemented21,22. CDI is a major cause of morbidity and mortality in UC patients and has been shown to be associated with an increased risk of colectomy, post-operative complications, and death23–27. Therefore, preventing CDI in UC patients is an important area of study and finding ways to identify children at risk of developing CDI would be a valuable first step.

The decrease in LCA and UDCA is most likely due to gut microbial differences between UC alone and UC+CDI, as these secondary bile acids are formed by microbial transformations that can be visualized in Figure 1. LCA is formed via 7α-dehydroxylation of CDCA, but can also be formed by 7β-dehydroxylation of UDCA20. UDCA is formed from CDCA by a two-step process, involving 7α-HSDH followed by 7β-HSDH20. However, these transformations can only take place once the bile acids are deconjugated by BSH. In Figure 8, we report that UC+CDI had significantly less BSH than UC alone (p <= 0.01). We also found that 7α-HDSH, the first step in conversion of CDCA to UDCA, is significantly reduced in UC+CDI (p <= 0.05), while there is no difference in the second step, 7β-HDSH. Additionally, in Figure 9, we demonstrate that the genes baiA, baiB, and baiH, which are involved in 7α-dehydroxylation and 7β-dehydroxylation, were significantly reduced in UC+CDI compared to UC (p <= 0.05). The differences in the abundance of these enzyme genes provide a potential mechanism behind the reduced LCA and UDCA we observed. We additionally demonstrated that UC disease activity, time since last CDI infection, and number of CDI recurrences was not associated with changes in the LCA or UDCA concentrations.

Currently, faecal microbiota transplantation (FMT) is recommended after the third recurrence of CDI in pediatric patients28. However, recent evidence has reported the occurrence of multi-drug resistant organism infection following FMT, which lead to death in at least one case29. More targeted therapies for preventing and treating CDI would be useful in avoiding the risks associated with FMT, many of which may not yet be fully understood. Examples of more targeted therapies include a study that found that supplementing patients with Clostridium scindens alone, which is capable of 7α-dehydroxylation, replenished secondary bile acids and restored resistance to CDI in patients that had undergone hematopoietic stem cell transplantation30. Other studies support the benefit of UDCA in mitigating CDI.31–35. Data obtained from this study and from similar future studies can serve as groundwork to develop gut microbial and bile acid targeted therapies.

To summarize, in this study we compared faecal bile acids in healthy children, children with a history of CDI, children with mild to moderate UC, and children with mild to moderate UC and a history of CDI. We describe novel findings of faecal bile acids in patients with UC+CDI and found that secondary bile acids, especially LCA and UDCA, were significantly reduced in this group compared to those with UC alone. Therefore, bile acid profiles can distinguish between children with UC alone from those with UC who have had CDI in the past. Those with UC+CDI may therefore have a microbiome permissive to the development of CDI and faecal bile acid may be helpful biomarkers to evaluate their risk. The bile acid profiles among these two groups can be interpreted based on the changes we see in BSH, 7α-HDSH, and bai genes as well as changes in bacterial genera involved in producing these genes. Since bile acid levels in children approach adult levels by 4 years of age36 and our study population had an average age of approximately 15 years old, these findings can likely be generalized to adults as well.

Future studies would be important to validate our findings in a larger cohort. It would also be interesting to assess whether faecal bile acids could be useful as biomarkers to distinguish between UC flare and active CDI. CDI is diagnosed when patients develop new onset diarrhea and have a positive stool test.37 However, in patients with UC, new onset diarrhea may represent a disease flare and a positive stool test may represent asymptomatic colonization. Therefore, it would be useful to have a biomarker that could differentiate between CDI causing colitis in a patient with UC and worsening colitis due to UC flare. Previous studies evaluated the use of procalcitonin and other systemic inflammatory response markers for this purpose without success.38,39 While our findings suggest that LCA and UDCA are useful for determining whether a UC patient has had CDI in the past, future studies of UC patients with active CDI would be necessary to assess the role of bile acids in distinguishing UC flare from active CDI infection.

Supplementary Material

fS2

Figure S2 Histogram of the number of reads in the samples

fS1

Figure S1 Boxplots of the log transformed concentrations (picograms/milligram feces) of deconjugated, glycine conjugated, and taurine conjugated of each individual bile acid (CA, CDCA, DCA, LCA, UDCA) for CDI, Healthy, UC, and UC+CDI groups. P values obtained by ANOVA or Kruskal-Wallis tests are included, with ns (not significant) for p>0.05, * for p <= 0.05, ** for p <= 0.01, *** for p <= 0.001, and **** for p <= 0.0001.

Acknowledgements:

Sonia Michail will act as guarantor of the article. Sarah Rotondo-Trivette analyzed the data and wrote the paper. Beibei Wang performed statistical analysis. Christopher Gayer edited the paper. Riddhi Parsana analyzed microbial DNA. Yihui Luan and Fengzhu Sun oversaw statistical analysis. Sonia Michail designed the study, collected data, and assisted with writing of the paper. All authors approved the final version of this manuscript.

Grant Support:

Sonia Michail, National Institutes of Health R01HD081197. Beibei Wang, China Scholarship Council.

Footnotes

COI Statements:

Sarah Rotondo-Trivette has no conflicts of interest to disclose.

Beibei Wang has no conflicts of interest to disclose.

Christopher Gayer has no conflicts of interest to disclose.

Riddhi Parsana has no conflicts of interest to disclose.

Yihui Luan has no conflicts of interest to disclose.

Fengzhu Sun has no conflicts of interest to disclose.

Sonia Michail has no conflicts of interest to disclose.

References

  • 1.Duboc H et al. Connecting dysbiosis, bile-acid dysmetabolism and gut inflammation in inflammatory bowel diseases. Gut 62, 531–539, doi: 10.1136/gutjnl-2012-302578 (2013). [DOI] [PubMed] [Google Scholar]
  • 2.Rocchi A et al. Inflammatory bowel disease: a Canadian burden of illness review. Can J Gastroenterol 26, 811–817, doi: 10.1155/2012/984575 (2012). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3.Foster A & Jacobson K Changing incidence of inflammatory bowel disease: environmental influences and lessons learnt from the South asian population. Front Pediatr 1, 34, doi: 10.3389/fped.2013.00034 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Schildkraut V et al. Sixty-year study of incidence of childhood ulcerative colitis finds eleven-fold increase beginning in 1990s. Inflamm Bowel Dis 19, 1–6, doi: 10.1002/ibd.22997 (2013). [DOI] [PubMed] [Google Scholar]
  • 5.Nitzan O, Elias M, Chazan B, Raz R & Saliba W Clostridium difficile and inflammatory bowel disease: role in pathogenesis and implications in treatment. World J Gastroenterol 19, 7577–7585, doi: 10.3748/wjg.v19.i43.7577 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Theriot CM & Fletcher JR Human faecal metabolomic profiling could inform Clostridioides difficile infection diagnosis and treatment. J Clin Invest 129, 3539–3541, doi: 10.1172/JCI130008 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Tam J et al. Intestinal bile acids directly modulate the structure and function of C. difficile TcdB toxin. Proc Natl Acad Sci U S A 117, 6792–6800, doi: 10.1073/pnas.1916965117 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Theriot CM & Petri WA Jr. Role of Microbiota-Derived Bile Acids in Enteric Infections. Cell 181, 1452–1454, doi: 10.1016/j.cell.2020.05.033 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Brown JR et al. Changes in microbiota composition, bile and fatty acid metabolism, in successful faecal microbiota transplantation for Clostridioides difficile infection. BMC Gastroenterol 18, 131, doi: 10.1186/s12876-018-0860-5 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Ridlon JM, Kang DJ & Hylemon PB Bile salt biotransformations by human intestinal bacteria. J Lipid Res 47, 241–259, doi: 10.1194/jlr.R500013-JLR200 (2006). [DOI] [PubMed] [Google Scholar]
  • 11.Allegretti JR et al. Recurrent Clostridium difficile infection associates with distinct bile acid and microbiome profiles. Aliment Pharmacol Ther 43, 1142–1153, doi: 10.1111/apt.13616 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.La Merrill M et al. Perinatal exposure of mice to the pesticide DDT impairs energy expenditure and metabolism in adult female offspring. PLoS One 9, e103337, doi: 10.1371/journal.pone.0103337 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Walters W et al. Improved Bacterial 16S rRNA Gene (V4 and V4–5) and Fungal Internal Transcribed Spacer Marker Gene Primers for Microbial Community Surveys. mSystems 1, doi: 10.1128/mSystems.00009-15 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Caporaso JG et al. Global patterns of 16S rRNA diversity at a depth of millions of sequences per sample. Proc Natl Acad Sci U S A 108Suppl 1, 4516–4522, doi: 10.1073/pnas.1000080107 (2011). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Bolyen E et al. Reproducible, interactive, scalable and extensible microbiome data science using QIIME 2. Nature biotechnology 37, 852–857 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Callahan BJ et al. DADA2: high-resolution sample inference from Illumina amplicon data. Nature methods 13, 581–583 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Rognes T, Flouri T, Nichols B, Quince C & Mahé F VSEARCH: a versatile open source tool for metagenomics. PeerJ 4, e2584 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.DeSantis TZ et al. Greengenes, a chimera-checked 16S rRNA gene database and workbench compatible with ARB. Applied and environmental microbiology 72, 5069–5072 (2006). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Wei PL et al. Characterization of Faecal Microbiota with Clinical Specimen Using Long-Read and Short-Read Sequencing Platform. Int J Mol Sci 21, doi: 10.3390/ijms21197110 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Ridlon JM & Bajaj JS The human gut sterolbiome: bile acid-microbiome endocrine aspects and therapeutics. Acta Pharm Sin B 5, 99–105, doi: 10.1016/j.apsb.2015.01.006 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Khanna S Management of Clostridioides difficile infection in patients with inflammatory bowel disease. Intest Res, doi: 10.5217/ir.2020.00045 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Mahida YR New concepts in C. difficile management. Br Med Bull 131, 109–118, doi: 10.1093/bmb/ldz029 (2019). [DOI] [PubMed] [Google Scholar]
  • 23.Khoudari G et al. Rates of Intestinal Resection and Colectomy in Inflammatory Bowel Disease Patients After Initiation of Biologics: A Cohort Study. Clin Gastroenterol Hepatol, doi: 10.1016/j.cgh.2020.10.008 (2020). [DOI] [PubMed] [Google Scholar]
  • 24.Waljee AK et al. Mapping the relationships between inflammatory bowel disease and comorbid diagnoses to identify disease associations. Eur J Gastroenterol Hepatol 32, 1341–1347, doi: 10.1097/MEG.0000000000001869 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Hellmann J et al. Microbial Shifts and Shorter Time to Bowel Resection Surgery Associated with C. difficile in Pediatric Crohn’s Disease. Inflamm Bowel Dis 26, 1212–1221, doi: 10.1093/ibd/izz263 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.Balram B et al. Risk Factors Associated with Clostridium difficile Infection in Inflammatory Bowel Disease: A Systematic Review and Meta-Analysis. J Crohns Colitis 13, 27–38, doi: 10.1093/ecco-jcc/jjy143 (2019). [DOI] [PubMed] [Google Scholar]
  • 27.Negron ME et al. Ulcerative Colitis Patients With Clostridium difficile are at Increased Risk of Death, Colectomy, and Postoperative Complications: A Population-Based Inception Cohort Study. Am J Gastroenterol 111, 691–704, doi: 10.1038/ajg.2016.106 (2016). [DOI] [PubMed] [Google Scholar]
  • 28.Davidovics ZH et al. Faecal Microbiota Transplantation for Recurrent Clostridium difficile Infection and Other Conditions in Children: A Joint Position Paper From the North American Society for Pediatric Gastroenterology, Hepatology, and Nutrition and the European Society for Pediatric Gastroenterology, Hepatology, and Nutrition. J Pediatr Gastroenterol Nutr 68, 130–143, doi: 10.1097/MPG.0000000000002205 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Michail SK, Stacy; Nicholson, Maribeth; Kellermyer, Richard. Addendum for: Faecal Microbiota Transplantation for Recurrent Clostridium difficile Infection and Other Conditions in Children: A Joint Position Paper From the North American Society for Pediatric Gastroenterology, Hepatology, and Nutrition and the European Society for Pediatric Gastroenterology, Hepatology, and Nutrition. J Pediatr Gastroenterol Nutr 70, 404 (2020). [DOI] [PubMed] [Google Scholar]
  • 30.Buffie CG et al. Precision microbiome reconstitution restores bile acid mediated resistance to Clostridium difficile. Nature 517, 205–208, doi: 10.1038/nature13828 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31.Weingarden AR et al. Ursodeoxycholic Acid Inhibits Clostridium difficile Spore Germination and Vegetative Growth, and Prevents the Recurrence of Ileal Pouchitis Associated With the Infection. J Clin Gastroenterol 50, 624–630, doi: 10.1097/MCG.0000000000000427 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.Winston JA et al. Ursodeoxycholic Acid (UDCA) Mitigates the Host Inflammatory Response during Clostridioides difficile Infection by Altering Gut Bile Acids. Infect Immun 88, doi: 10.1128/IAI.00045-20 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33.Palmieri LJ et al. Inhibitory Effect of Ursodeoxycholic Acid on Clostridium difficile Germination Is Insufficient to Prevent Colitis: A Study in Hamsters and Humans. Front Microbiol 9, 2849, doi: 10.3389/fmicb.2018.02849 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34.Webb BJ, Brunner A, Lewis J, Ford CD & Lopansri BK Repurposing an Old Drug for a New Epidemic: Ursodeoxycholic Acid to Prevent Recurrent Clostridioides difficile Infection. Clin Infect Dis 68, 498–500, doi: 10.1093/cid/ciy568 (2019). [DOI] [PubMed] [Google Scholar]
  • 35.Camilleri M & Gores GJ Therapeutic targeting of bile acids. Am J Physiol Gastrointest Liver Physiol 309, G209–215, doi: 10.1152/ajpgi.00121.2015 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36.Cheng S, Zhu L & Faden HS Interactions of bile acids and the gut microbiota: learning from the differences in Clostridium difficile infection between children and adults. Physiol Genomics 51, 218–223, doi: 10.1152/physiolgenomics.00034.2019 (2019). [DOI] [PubMed] [Google Scholar]
  • 37.McDonald LC et al. Clinical Practice Guidelines for Clostridium difficile Infection in Adults and Children: 2017 Update by the Infectious Diseases Society of America (IDSA) and Society for Healthcare Epidemiology of America (SHEA). Clin Infect Dis 66, e1–e48, doi: 10.1093/cid/cix1085 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 38.Limsrivilai J et al. Systemic Inflammatory Responses in Ulcerative Colitis Patients and Clostridium difficile Infection. Dig Dis Sci 63, 1801–1810, doi: 10.1007/s10620-018-5044-1 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39.Reinink AR et al. Differentiating Clostridium difficile Colitis from Clostridium difficile Colonization in Ulcerative Colitis: A Role for Procalcitonin. Digestion 96, 207–212, doi: 10.1159/000481133 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

fS2

Figure S2 Histogram of the number of reads in the samples

NIHMS1713680-supplement-fS2.png (38.2KB, png)
fS1

Figure S1 Boxplots of the log transformed concentrations (picograms/milligram feces) of deconjugated, glycine conjugated, and taurine conjugated of each individual bile acid (CA, CDCA, DCA, LCA, UDCA) for CDI, Healthy, UC, and UC+CDI groups. P values obtained by ANOVA or Kruskal-Wallis tests are included, with ns (not significant) for p>0.05, * for p <= 0.05, ** for p <= 0.01, *** for p <= 0.001, and **** for p <= 0.0001.

NIHMS1713680-supplement-fS1.tif (708.6KB, tif)

RESOURCES