Flavin-Containing Monooxygenase 3 (FMO3) Is Critical for Dioxin-Induced Reorganization of the Gut Microbiome and Host Insulin Sensitivity

Round 1

Reviewer 1 Report

In this work the authors used either WT Fmo3+/+ or knockout Fmo3-/-  mice to study how the endocrine disruptor TCDD affects the expression and the activity of the enzyme. Furthermore, the manuscript is also aimed at dissecting the possible connections between the effects of TCDD toxicity on gut microbiome and the liver role of FMO3. The experiments presented are convincing for all the different aims such as study the effect of TCDD on:

1. mRNA and protein expression induction
2. glucose (in)tolerance
3. microbiome re-organization
4. AhR target expression.

Nevertheless, the manuscript requires a big effort to restructure both the introduction and the discussion to make it more comprehensible for a broader audience. In the current format, it is not straightforward to understand what is the significance of the work especially when the concepts of diet, microbiome and the specific role in metabolism of FMO3 in converting TMA to TMAO are brought together. A link is missing.

Comments:The authors have focused mainly on the effect of TCDD on cell biology, but more attention should be given to the biochemistry of the enzyme that is the actual subject of the study: FMO3. For a comprehensive review about FMO3 biochemistry, the connection to TMAO and the multiple roles of the metabolite please see: Catalysts 9 (12), 1028. Furthermore, while the experimental model of this works is centered on the differences emerging between FMO3 and knockout FMO3 it would be also interesting to put into this context the single nucleotide polymorphic variants of FMO3 within the introduction and what these mutations could mean in terms of structural-functional relationship for the enzyme, please see: International Journal of Biological Macromolecules 162, 1484-1493; Basic & Clinical Pharmacology & Toxicology 123 (6), 687-691; Scientific reports 7 (1), 1-11; Gene 593 (1), 91-99; International journal of molecular sciences 14 (2), 2707-2716

lines 303-309 the authors say: “Collectively, this work demonstrates an underappreciated connection between environmental pollutants, the gut microbiome, and host glucose homeostasis. Given the fact that the gut microbial TMA-FMO3-TMAO pathway is initiated by ingestion of trimethylamine containing nutrients (i.e., choline, carnitine, -butyrobetaine, etc.), this work has important implications in our understanding of how dietary practices may converge with environmental exposures to impact cardiometabolic disease.”

What are the implications? What practices are going to impact cardiometabolic disease? There seems to be no effort by authors to join the dots. Is it possible to draw a figure to make it more clear? Please be more specific.

lines 311-315 the authors say “AhR ligands such as TCDD have long been known to have endocrine disrupting properties that can contribute to diverse human diseases including obesity and diabetes [1-4], atherosclerosis [5, 6], diverse cancers [7-10], and kidney disease [11-13]. It is interesting to note that the TMA-FMO3-TMAO pathway has also been linked to these same diseases in humans [19-24]. This work bolsters the emerging concept that FMO3 may serve as a critical integrator of xenobiotic metabolism and cardiometabolic disease.”

Also in this case the authors are not straightforward. Beside the fact that level of interest might depend on the audience reading the paper – the point is: how do the authors believe FMO3 may serve as a critical integrator of xenobiotic metabolism and cardiometabolic disease?

Lines 326-346 should be summarized to 10-12 lines that are related to “study limitations and future perspectives”.

A separate “Conclusion” section could be very helpful.

• Minor:

Abstract, lines 31-35. The sentence is not clear, please re-phrase.

Author Response

Response to Reviewer 1 Comments

Point 1: The authors have focused mainly on the effect of TCDD on cell biology, but more attention should be given to the biochemistry of the enzyme that is the actual subject of the study: FMO3. For a comprehensive review about FMO3 biochemistry, the connection to TMAO and the multiple roles of the metabolite please see: Catalysts 9 (12), 1028. Furthermore, while the experimental model of this works is centered on the differences emerging between FMO3 and knockout FMO3 it would be also interesting to put into this context the single nucleotide polymorphic variants of FMO3 within the introduction and what these mutations could mean in terms of structural-functional relationship for the enzyme, please see: International Journal of Biological Macromolecules 162, 1484-1493; Basic & Clinical Pharmacology & Toxicology 123 (6), 687-691; Scientific reports 7 (1), 1-11; Gene 593 (1), 91-99; International journal of molecular sciences 14 (2), 2707-2716

Response 1: We appreciate the reviewers request to be more comprehensive in our introduction to/discussion of FMO3. We have made changes to the introduction and discussion sections to address this request using the suggested publications. In reference to adding discussion around FMO3 mutations and TMAO production capacity, we hesitate to overstate links. This is due to the fact that most of the mutations have not been assayed for TMA-to-TMAO conversion, but instead have been studied for xenobiotic metabolism using other known substrates.

Point 2: lines 303-309 the authors say: “Collectively, this work demonstrates an underappreciated connection between environmental pollutants, the gut microbiome, and host glucose homeostasis. Given the fact that the gut microbial TMA-FMO3-TMAO pathway is initiated by ingestion of trimethylamine containing nutrients (i.e., choline, carnitine, -butyrobetaine, etc.), this work has important implications in our understanding of how dietary practices may converge with environmental exposures to impact cardiometabolic disease.”

What are the implications? What practices are going to impact cardiometabolic disease? There seems to be no effort by authors to join the dots. Is it possible to draw a figure to make it more clear? Please be more specific.

Response 2: The reviewer’s desire for elaboration on these implications is well taken. We have elaborated to state “Specifically, clinical prevention and treatment strategies for cardiometabolic disease may someday take into account patient history of environmental exposure, diet, microbial TMA production capacity, and FMO3 genotype. Indeed, there is ongoing studies to phar-macologically target FMO3 to reduce its ability to convert TMA to TMAO [62-66], as well as studies using nonlethal, microbe-targeting inhibitors that limit production of TMA from dietary substrates [27, 38, 67-70].”.

Point 3: lines 311-315 the authors say “AhR ligands such as TCDD have long been known to have endocrine disrupting properties that can contribute to diverse human diseases including obesity and diabetes [1-4], atherosclerosis [5, 6], diverse cancers [7-10], and kidney disease [11-13]. It is interesting to note that the TMA-FMO3-TMAO pathway has also been linked to these same diseases in humans [19-24]. This work bolsters the emerging concept that FMO3 may serve as a critical integrator of xenobiotic metabolism and cardiometabolic disease.”

Also in this case the authors are not straightforward. Beside the fact that level of interest might depend on the audience reading the paper – the point is: how do the authors believe FMO3 may serve as a critical integrator of xenobiotic metabolism and cardiometabolic disease?

Response 3: To be more straightforward, we have elaborated further such that the link between xenobiotic metabolism and cardiometabolic disease is clear by rephrasing the quote to say “This work bolsters the emerging concept that FMO3 may serve as a critical integrator of xenobiotic metabolism and cardiometabolic disease through the microbe-dependent production of TMAO.”.

Point 4: Lines 326-346 should be summarized to 10-12 lines that are related to “study limitations and future perspectives”. A separate “Conclusion” section could be very helpful.

Response 4: We have removed the last few sentences of this paragraph and moved them to the newly created “Conclusion” section.

Point 5: Abstract, lines 31-35. The sentence is not clear, please re-phrase.

Response 5: We have rephrased this sentence to read: “Our results show that Fmo3 is a critical component of the transcriptional response to TCDD, impacting the gut microbiome, host liver transcriptome, and systemic glucose tolerance”

Reviewer 2 Report

Massey et al.,’ manuscript titled ‘Flavin-Containing Monooxygenase 3 (FMO3) is Critical for Dioxin-Induced Reorganization of the Gut Microbiome and Host   Insulin Sensitivity’ investigated the effect of TCDD on the gut microbiome and systemic. TCDD exposure enhanced the hepatic expression of flavin-containing monooxygenase 3 (Fmo3) expression, which is also responsible for producing the gut microbiome-associated metabolite trimethylamine N-oxide (TMAO). Furthermore, TCDD-induced alterations in the gut microbiome, host liver transcriptome, and glucose tolerance in Fmo3+/+ and Fmo3-/- mice. Generally, this study is well designed, and the MS is well-written with well-thought-out experiments. However, some areas of the manuscript are unclear that would benefit from clarification.

1. It would be helpful for the readers to present a summary figure to show these results.
2. Why is there no Fmo3 expression in the vehicle in wt (Fmo3+/+) samples in RT and wb (Figure 1B right panel and 1C)? That is so strange to me, and this issue is critical to your work.
3. For the RNA seq, the -log10 value, 2.0 means 0.01. why the author did not use p<0.05 (1.3) as a cut-off. Also, please deposit the RNA raw data to NCBI or other public databases. The DEGs list is also required for the MS. It would be good to put the DEGs in the supplemental table.

Author Response

Response to Reviewer 2 Comments

Point 1: It would be helpful for the readers to present a summary figure to show these results.

Response 1: We have now included a graphical abstract to summarize our findings. The current version of the graphical abstract is in proof form because our institutional art department requires copyright approval upon acceptance of all manuscript artwork. If accepted, we will include the final copyright approved version of this graphical abstract.

Point 2: Why is there no Fmo3 expression in the vehicle in wt (Fmo3+/+) samples in RT and wb (Figure 1B right panel and 1C)? That is so strange to me, and this issue is critical to your work.

Response 2: The extremely low levels of Fmo3 expression in WT male mice has been described in many previous publications. One of the best examples of this was shown by Bennett and colleagues (Bennett, B., et al. 2013. Trimethylamine-N-oxide, a metabolite associated with atherosclerosis, exhibits complex genetic and dietary regulation. Cell Metab. PMID: 23312283), which demonstrates that male mice have barely detectable protein levels of Fmo3, whereas females have abundant levels. This low Fmo3 expression seen in male mice is the result in part of complex circadian control (Schugar, R. et al. 2022. Gut microbe-targeted choline trimethylamine lyase inhibition improves obesity via rewiring of host circadian rhythms. Elife PMID: 35072627) as well as pure sexual dimorphism seen specifically in C57BL/6 mouse strains (Li, Q., et al. 2013. Synchronous evolution of an odor biosynthesis pathway and behavioral response. Curr. Biol. PMID: 23177478). In our previous work (Schugar, R., et al. 2022. Elife) we show that Fmo3 is expressed only in the dark cycle in male mice and is expressed overall at a higher level in the livers of female mice. We have added clarification to the results section 2.1. The key point of this manuscript is that TCDD-induced expression of Fmo3 is important in shaping the gut microbiome and host insulin sensitivity, so the basal level of expression in control mice is less important to this story.

Point 3: For the RNA seq, the -log10 value, 2.0 means 0.01. why the author did not use p<0.05 (1.3) as a cut-off. Also, please deposit the RNA raw data to NCBI or other public databases. The DEGs list is also required for the MS. It would be good to put the DEGs in the supplemental table.

Response 3: We have made changes to figures 3 and 4 such that the –log10 cutoff is 1.3 or p<0.05. Additionally, we have added a pathway analysis with all of the DEGs (new Figure 3B). The raw data has been submitted to the NCBI GEO portal (accession GSE191138). We now include the DEG excel file as a supplement.

Reviewer 3 Report

The authors investigated whether TCDD has any ability to affect the gut microbiome, host liver transcripts and systemic glucose tolerance depending on the appropriate Fmo3 transcriptional regulation driven by the aryl hydrocarbon receptor. The authors argue that the role of Fmo3 in integrating diet-pollutant-microbe-host interactions has been underestimated so far.
After a clear introduction, the research results were presented in some detail.
However, there was no reliable statistical analysis. Among others, in chapters 2.1, 2.2, 2.3 and the corresponding figures, the p value is missing. Figure 5 is illegible, especially parts D, E and the graphs in F.
In the Materials and methods section, the statistical tools used are only presented in subchapter 4.7. The descriptions of the methods are very laconic, so it would be difficult to repeat the experiments based on them.
The discussion is far too short, does not reflect the richness of the results, and there are no comparisons to similar studies. Literature is also somewhat out of date, with only about 20% relating to the last 4 years.
The manuscript requires redrafting and supplementing the missing content mentioned above.

Author Response

Response to Reviewer 3 Comments

Point 1: After a clear introduction, the research results were presented in some detail.

However, there was no reliable statistical analysis. Among others, in chapters 2.1, 2.2, 2.3 and the corresponding figures, the p value is missing.

Response 1: We appreciate the reviewers concern for reliable statistical analysis. All quantitative data in sections 2.1-2.3 and figures 1-4 were analyzed by one of three statistical tests. All pairwise analyses were performed on RNAseq data using the Cufflinks software package, which uses linear statistical modeling to estimate an abundance of each transcript that explains the observed reads with maximum likelihood (https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3334321/). Two-way ANOVA with a Tukey’s post-hoc test was used for 2x2 analyses, such as that for qPCR data, where genotype and treatment (corn oil or TCDD) were the two factors. When analyzing the glucose tolerance curve (Figure 2A), a Three-way ANOVA with a Tukey’s post-hoc test was used, where genotype, treatment, and time (relative to bolus glucose injection) were the three variables.We have added the relevent p-values to the text of these sections where appropriate. We additionally, added the results of the statistical testing to the text of secton 2.5.

Point 2: Figure 5 is illegible, especially parts D, E and the graphs in F.

Response 2: We agree with the reviwer that these data are difficult to interpret as prevously presented. To better display the data, we have separated figure 5 into three separate figures where old Figure 5A-C remains Figure 5, old Figure 5D,E is now Figure 6, and old Figure 5F is now Figure 7.

Point 3: In the Materials and methods section, the statistical tools used are only presented in subchapter 4.7. The descriptions of the methods are very laconic, so it would be difficult to repeat the experiments based on them.

Response 3: With regards to the statistical tools, we have added text to indicate that GraphPad 9 software was used for data visualization and analysis. We have also added more details to sections 4.2-4.6

Point 4: The discussion is far too short, does not reflect the richness of the results, and there are no comparisons to similar studies. Literature is also somewhat out of date, with only about 20% relating to the last 4 years.

Response 4: We have made numerous changes to the discussion section to address this critical reviewer comment. We have added further discussion of results and more recent literature to lengthen the discussion section. Unfortunately, to our knowledge there are no comparable studies investigating glucose tolerance, liver injury, or the microbiome in any knockout mouse line. The majority of studies using knockout animals treated with TCDD are with AhR-null mice, which display many abnormalities independent of TCDD. Additionally, we found one study where Cyp1a1/1a2 knockout mice were generated (https://dmd.aspetjournals.org/content/47/8/907.long); however, no enpoints were comparable to our study. Although beyond the scope of the current manuscript, it would be intersting to determine whether any of the effects observed in Fmo3-/- mice are due to the observed upregulation of Cyp1a1.

To address the reviewer’s comment about the age of the cited literature, we performed a literature search (PubMed, search term “tcdd”) and found several recent publications that have relevance to the findings and discussion in our paper that are now included in the revised manuscript.