Bile Acid Analogs with Anti-Germination Activities for Prophylaxis of Clostridioides difficile Infection Alter Bile Acid Homeostasis in the Enterohepatic Cycle

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

This manuscript describes an analysis of bile acid (BA) species and hepatic transcriptomic changes in mice treated with CamSA and CA-Quin, two compounds targeting C. difficile infection, for seven days. The authors analyzed BA profiles in the liver, feces, and intestinal chyme. Both compounds exhibited similar effects on BA composition and significantly increased fecal BA excretion. Furthermore, liver RNA-seq analysis revealed that neither compound caused observable hepatotoxicity, suggesting they are safe. Overall, this study complements the authors’ previous findings and supports the conclusion that CamSA and CA-Quin may inhibit C. difficile germination by modulating BA homeostasis and improving gut microbiota composition.

Comments:

1. The authors should clarify the units used for BA quantification in Figures 2 and 3 (e.g., µM, nM). Details on how BA data were normalized should also be included in the Methods section.

2. Since ALT and AST are well-known markers of liver toxicity, it would strengthen the study to include measurements of plasma ALT and AST levels.

3. It appears that different numbers of mice were used for the BA analysis (DMSO: n=4, CamSA: n=3, CA-Quin: n=4) and RNA-seq analysis (DMSO: n=8, CamSA: n=8, CA-Quin: n=7). The authors should clarify the rationale for excluding some mice from the BA analysis.AI-assisted tool was used to check the grammar of this report.

Author Response

Comment 1: The authors should clarify the units used for BA quantification in Figures 2 and 3 (e.g., µM, nM). Details on how BA data were normalized should also be included in the Methods section

Response 1: Units added in figures 2 and 3, as suggested

BA data normalization explanation has been added to the experimental section

Comment 2: Since ALT and AST are well-known markers of liver toxicity, it would strengthen the study to include measurements of plasma ALT and AST levels.

Response 2: As this was an ancillary study, based on available biospecimen, to assess the effect of anti-germinants on bile acid metabolism, we unfortunately did not collect blood samples from the tested animals. However, previous published work has shown that mice and hamsters treated with multi-day high doses of CamSA do not develop any overt signs of acute or sub-chronic toxicity. Furthermore, CamSA-treated mice show no gross hepatic or intestinal anomalies. Similar lack of toxicity was observed with other synthetic bile salt anti-germinants (PMID: 23420906, PMID: 24023628, PMID: 30012758, PMID: 34780262). Similarly, CA-Quin treated animals in this study did not show overt signs of toxicity or organ alterations. We further clarified this in the discussion of our manuscript.

Comment 3: It appears that different numbers of mice were used for the BA analysis (DMSO: n=4, CamSA: n=3, CA-Quin: n=4) and RNA-seq analysis (DMSO: n=8, CamSA: n=8, CA-Quin: n=7). The authors should clarify the rationale for excluding some mice from the BA analysis.AI-assisted tool was used to check the grammar of this report.

Response 3: The number of animals used was the same for BA analysis and RNA-seq. The discrepancy is due to the fact that for RNAseq, two separate liver samples were analyzed from each animal (technical replicates). We now have removed the duplicated samples, and RNA-seq and qPCR have been reanalyzed properly to convey this.

Reviewer 2 Report

Comments and Suggestions for Authors

The manuscript, "Bile Acid Analogs with Anti-germination Activities for Prophy- 2
laxis of Clostridioides difficile Infection Alter Bile Acid Homeo- 3
stasis in the Enterohepatic Cycle", reported the results of these two analogs on the liver ranscriptome and bile acid profiling, following a 7 day treatment of the bile acid analogs, CamSA and CA-Quin for antigemination activities against C. difficile.

The study is highly significant because of the novelty in using bile acid modifiers in the treatment of GI infection. Using these novel treatments, it was shown no obvious toxicity even though the compound seems highly effective. Due to the nature of these compounds as bile acid modifiers, liver gene expression at the mRNA level, and bile acid profile in a few tissues were determined. It is interesting that there is little impact to the bile acid or FXR regulated genes in the intestine, with Cyp7a1 induced in the liver, along with changes in induction of other genes involved in drug metabolism.

Overall, the paper is well written, easy to follow and the results were convincing. There is only one minor suggestion is to check the RNAseq data to see if PXR pathway is altered, given teh change in Cyp3a11 and Cyp7a1.

 

Author Response

Comment 1: Overall, the paper is well written, easy to follow and the results were convincing. There is only one minor suggestion is to check the RNAseq data to see if PXR pathway is altered, given the change in Cyp3a11 and Cyp7a1.

Response 1: This is a great suggestion. To investigate this in a more objective way, we performed an enrichment analysis for the significant DEGs in each treatment group to examine whether there are critical upstream regulators of these DEGs that mediate the impact of drug treatment. Our analysis did reveal that Nr1i3 and PPARA, instead of PXR, are potential transcription factors mediating the transcriptomic changes in the CA-Quin-treated group, but not in the CamSA-treated group. We have included this analysis in our manuscript.

Reviewer 3 Report

Comments and Suggestions for Authors

The manuscript examines two bile-acid analogs, CamSA and CA-Quin, previously shown to inhibit C. difficile spore germination, and evaluates 7-day oral dosing effects on murine bile-acid (BA) profiles (liver, ileal chyme, feces) and hepatic transcriptome. Headline findings are increased fecal BA excretion—especially unconjugated species—reduced fecal TCA, increased fecal CDCA, and modest hepatic transcriptional changes; the authors infer limited hepatotoxicity and propose that BA shifts may further hinder C. difficile germination.

Major:

1. The authors attribute increased fecal unconjugated BAs to “elevated gut microbiota activity.” However, host changes could also drive this: you report downregulation of hepatic BAAT (conjugation), BSEP, and ileal OSTβ—all of which can reduce conjugation and reabsorption, increasing fecal unconjugates without requiring higher microbial BSH activity. Please temper the claim, and (ideally) add fecal BSH activity assays and/or 16S/shotgun metagenomics to quantify microbial capacity.
2. Methods say “Serum and liver BA were extracted and quantified,” yet results present liver, chyme, feces (and no serum). Please clarify whether serum was measured and where those data are reported. If not measured, amend Methods.
3. Results show only hepatic RNA-seq and ileal qPCR. Please either present ileal RNA-seq analyses (including FXR–FGF15 axis genes: Fxr/Nr1h4, Fgf15, Asbt/Slc10a2, Ostα/β, Fabp6/I-BABP) or state why ileal RNA was not analyzed transcriptomically. This is central to interpreting EHC changes and host feedback on Cyp7a1.
4. The Introduction/overview states liver histology was analyzed, but no histology results or images are shown. Please include H&E/Pathology scoring (e.g., inflammation, cholestasis, steatosis).
5. You infer “minimal hepatotoxicity” from PCA/DEG patterns. Please strengthen with serum ALT/AST/ALP/total bilirubin, liver weight-to-body-weight ratios, and histopathology. Given BSEP downregulation, cholestatic risk should be considered explicitly.
6. The translational claim would be much stronger if you tested ex vivo spore germination with collected chyme/fecal extracts from treated mice (standard in the field), or linked BA shifts to germination assay readouts in parallel. At minimum, quantify the TCA:CDCA germination index and show correlations with the germination outcome.
7. Please Justify 50 mg/kg QD × 7 days for both analogs. For CA-Quin, the statement “not metabolized by the murine microbiota (data not shown)” needs supporting data or removal.
8. You posit decreased EHC due to ileal OSTβ downregulation; consider measuring ASBT (Slc10a2) and FGF15–SHP (Nr0b2) to strengthen the FXR-FGF15 feedback link to Cyp7a1 upregulation reported.

Minor:

1. Group size is n=4 per treatment for most endpoints, but feces on day 7 reports CamSA n=3. And RNA-seq sizes differ (DMSO 8, CamSA 8, CA-Quin 7). Please explain these missing samples.
2. Methods state no p-value correction for t-tests and extensive one-way ANOVAs across many BA species/timepoints and qPCR targets. Please apply Benjamini–Hochberg FDR (or similar) for BA panels and qPCR, and report adjusted p-values alongside effect sizes and CIs.
3. DEGs were filtered at unadjusted p<0.05 for Venn/enrichment; volcano plots then highlight genes by “−log10(adjusted p) < 0.05,” which is mathematically inconsistent and likely a text error.
4. The text/figure legend refers to “Bsep (Abcc11)”, but BSEP is Abcb11, not Abcc11. Please correct gene symbols consistently across RNA-seq, figures, and qPCR sections.
5. Where you describe a “trend of separation” in PCA analysis, please add PERMANOVA/Adonis on distance matrices and report variance explained with confidence intervals.
6. “TCA taurocholic acid” is correct, but you also list “TCA taurodeoxycholic acid”, which should be TDCA—please correct. Also fix duplication.
7. “PCA principle component analysis” should be corrected into principal.
8. Where you infer increased SCFA production from hepatic enrichment of “monocarboxylic/organic acid” terms, please temper or support with fecal/serum SCFA quantification; pathway enrichment alone is indirect.
9. Ensure consistent use of Greek symbols (α, β, ω) and capitalization across text, figures, and Supplementary Tables.

Author Response

Major Comments:

Comment 1: The authors attribute increased fecal unconjugated BAs to “elevated gut microbiota activity.” However, host changes could also drive this: you report downregulation of hepatic BAAT (conjugation), BSEP, and ileal OSTβ—all of which can reduce conjugation and reabsorption, increasing fecal unconjugates without requiring higher microbial BSH activity. Please temper the claim, and (ideally) add fecal BSH activity assays and/or 16S/shotgun metagenomics to quantify microbial capacity.

Response 1: We agree with the reviewer that multiple mechanisms could lead to increased fecal unconjugated BA. The manuscript has been modified to discuss these possibilities.

Comment 2: Methods say “Serum and liver BA were extracted and quantified,” yet results present liver, chyme, feces (and no serum). Please clarify whether serum was measured and where those data are reported. If not measured, amend Methods.

Response 2: This was a typo. Serum was not tested. We have changed this in the manuscript.

Comment 3: Results show only hepatic RNA-seq and ileal qPCR. Please either present ileal RNA-seq analyses (including FXR–FGF15 axis genes: Fxr/Nr1h4, Fgf15, Asbt/Slc10a2, Ostα/β, Fabp6/I-BABP) or state why ileal RNA was not analyzed transcriptomically. This is central to interpreting EHC changes and host feedback on Cyp7a1.

Response 3: We tried to perform RNA-Seq on ileum samples. However, the extracted ileum RNA was not of sufficient quality for RNA-Seq. Thus, we were limited to ileal q-PCR.

Comment 4: The Introduction/overview states liver histology was analyzed, but no histology results or images are shown. Please include H&E/Pathology scoring (e.g., inflammation, cholestasis, steatosis).

Response 4: This was a typo. Given that this is an ancillary study based on available biopsy samples, we only have frozen liver tissue. Unfortunately, we failed to obtain high-quality histological staining based on these frozen liver tissue. Therefore, no histology was performed. We have now removed this sentence from the manuscript.

Comment 5: You infer “minimal hepatotoxicity” from PCA/DEG patterns. Please strengthen with serum ALT/AST/ALP/total bilirubin, liver weight-to-body-weight ratios, and histopathology. Given BSEP downregulation, cholestatic risk should be considered explicitly.

Response 5: As this was an ancillary study, based on available biospecimen, to assess the effect of anti-germinants on bile acid metabolism, we unfortunately did not collect blood samples from the tested animals. However, previous published work has shown that mice and hamsters treated with multi-day high doses of CamSA do not develop any overt signs of acute or sub-chronic toxicity. Furthermore, CamSA-treated mice show no gross hepatic or intestinal anomalies. Similar lack of toxicity was observed with other synthetic bile salt anti-germinants (PMID: 23420906, PMID: 24023628, PMID: 30012758, PMID: 34780262). Similarly, CA-Quin treated animals in this study did not show overt signs of toxicity or organ alterations. We further clarified this in the discussion of our manuscript.

Comment 6: The translational claim would be much stronger if you tested ex vivo spore germination with collected chyme/fecal extracts from treated mice (standard in the field), or linked BA shifts to germination assay readouts in parallel. At minimum, quantify the TCA:CDCA germination index and show correlations with the germination outcome.

Response 6: Our current model is that CDI prophylaxis is the result of anti-germination activity. As such, any effect from BA pool changes on ex vivo spore germination, would be masked by the presence of CamSA or CA-Quin. We do agree that it is interesting to show the TCA:CDCA ratios. This was added to the manuscript.

Comment 7: Please Justify 50 mg/kg QD × 7 days for both analogs. For CA-Quin, the statement “not metabolized by the murine microbiota (data not shown)” needs supporting data or removal.

Response 7: Since the stability of CA-Quin was performed by TLC, these results are only semi-quantitative. The statement was removed from the manuscript

Comment 8: You posit decreased EHC due to ileal OSTβ downregulation; consider measuring ASBT (Slc10a2) and FGF15–SHP (Nr0b2) to strengthen the FXR-FGF15 feedback link to Cyp7a1 upregulation reported.

Response 8: ASBT and FGF15 were analyzed by ileal qPCR and were found not to be altered in the presence of CamSA or CA-Quin.

 

Minor Comments:

Comment 1: Group size is n=4 per treatment for most endpoints, but feces on day 7 reports CamSA n=3. And RNA-seq sizes differ (DMSO 8, CamSA 8, CA-Quin 7). Please explain these missing samples.

Response 1: The number of animals used was the same for BA analysis and RNA-seq. The discrepancy is due to the fact that for RNAseq, two separate liver samples were analyzed from each animal (technical replicates). We now have removed the duplicated samples, and RNA-seq and qPCR have been reanalyzed properly to convey this.

Comment 2: Methods state no p-value correction for t-tests and extensive one-way ANOVAs across many BA species/timepoints and qPCR targets. Please apply Benjamini–Hochberg FDR (or similar) for BA panels and qPCR, and report adjusted p-values alongside effect sizes and CIs.

Response 2: We reported the Tukey test results. The Tukey test compares the difference between each pair of means with appropriate adjustment for the multiple testing. Therefore, no more adjustment for multiple testing is needed.

Comment 3: DEGs were filtered at unadjusted p<0.05 for Venn/enrichment; volcano plots then highlight genes by “−log10(adjusted p) < 0.05,” which is mathematically inconsistent and likely a text error.

Response 3: Corrected.

Comment 4: The text/figure legend refers to “Bsep (Abcc11)”, but BSEP is Abcb11, not Abcc11. Please correct gene symbols consistently across RNA-seq, figures, and qPCR sections.

Response 4: Corrected.

Comment 5: Where you describe a “trend of separation” in PCA analysis, please add PERMANOVA/Adonis on distance matrices and report variance explained with confidence intervals.

Response 5: Corrected as requested. Table S5 details results from R.

Comment 6: “TCA taurocholic acid” is correct, but you also list “TCA taurodeoxycholic acid”, which should be TDCA—please correct. Also fix duplication.

Response 6: Corrected.

Comment 7: “PCA principle component analysis” should be corrected into principal.

Response 7: Corrected.

Comment 8: Where you infer increased SCFA production from hepatic enrichment of “monocarboxylic/organic acid” terms, please temper or support with fecal/serum SCFA quantification; pathway enrichment alone is indirect.

Response 8: We removed this from the discussion. 

Comment 9: Ensure consistent use of Greek symbols (α, β, ω) and capitalization across text, figures, and Supplementary Tables.

Response 9: Corrected as requested.

Round 2

Reviewer 3 Report

Comments and Suggestions for Authors

The authors have addressed most of the concerns raised in the previous round of review, particularly regarding the clarification of methods and the addition of ileal qPCR data. However, the revisions have exposed a discrepancy between the molecular data and the authors' conclusions regarding "safety," which needs to be addressed in the discussion before publication.

1. From a physiological perspective, increased bile acid synthesis combined with inhibited apical secretion is a classic molecular signature of potential intrahepatic cholestasis. Without serum liver enzymes (ALT/AST) or histology to definitively rule out cholestasis, the conclusion should reflect that while overt toxicity wasn't observed in previous studies, the current molecular profile suggests a potential risk of bile acid accumulation or cholestatic stress that warrants monitoring.

2. The authors show Fgf15 is unchanged, yet Cyp7a1 is upregulated in the liver (CA-Quin group). Usually, Cyp7a1 is suppressed by FGF15. Please briefly discuss this discrepancy in the discussion section.

Author Response

Comment 1: From a physiological perspective, increased bile acid synthesis combined with inhibited apical secretion is a classic molecular signature of potential intrahepatic cholestasis. Without serum liver enzymes (ALT/AST) or histology to definitively rule out cholestasis, the conclusion should reflect that while overt toxicity wasn't observed in previous studies, the current molecular profile suggests a potential risk of bile acid accumulation or cholestatic stress that warrants monitoring.

Response 1: We agree with this point. We have discussed this further in the manuscript.

Comment 2: The authors show Fgf15 is unchanged, yet Cyp7a1 is upregulated in the liver (CA-Quin group). Usually, Cyp7a1 is suppressed by FGF15. Please briefly discuss this discrepancy in the discussion section.

Response 2: This is a great suggestion. We have discussed more about this issue in the manuscript.