Review Reports - A Nominally Safe Dose of Fumonisins Induces Mild Neuroinflammation in Chickens by Targeting Sphingolipids and Oxylipins but Not Cytokines

Round 1

Reviewer 1 Report

please see the pdf file uploaded

Comments for author File: Comments.pdf

Author Response

Reviewer 1

Referee evaluation of the manuscript “A nominally safe dose of fumonisins induces neuroinflammation in chickens by targeting sphingolipids and oxylipins but not cytokines” (Antioxidants, MDPI, Guerre et al.)

The study is a very detailed analysis of whole brain homogenates of chickens fed 14,6 mg/kg FB mycotoxins for 14 and 21 days, before termination. The omics type analysis was based on LC-MS/MS separation and detection of numerous components of the sphingolipidome, oxylipidome, cytokines, and PLA2c.

In general, the manuscript is very accurately prepared and the study design allow to address achievement of the goals of the work. I made in-text suggestions directly into the printout, these I will not repeat here but ask the Authors to address them in case of agreement. Please accept all my handwritten comments as suggestions and please feel free to implement them, my intention was to improve the work.

I did not indicate major things to implement during revision. Minor mistakes and suggestions are directly indicated with handwriting.

Thank you for taking the time to read this manuscript and for your very positive and constructive feedback. We have taken all your comments into account; they appear in red in the revised version of the manuscript. The changes requested by the other reviewer appear highlighted in yellow.

Some things I suggest for the Authors for consideration:

- The experimental design allows the exposure time based regressions, i.e. exposure time response. I kindly recommend to think over its implementation once the data support it. (At line 500 it is stated that prolonged exposure supports BBB permeability, so exposure dependence for it may be interesting.)

Thank you for your comment. It is difficult to model the passage of FB1 into the brain with only two measurement times. Tissue diffusion modeling generally follows first-order, sometimes second-order, kinetics, which are modeled by exponential equations, thus requiring at least three measurement times. In this study, determined by linear regression, the increase in FB1 concentrations in the brain occurs at a slope similar to that measured in the liver. Rather than a true change in BBB permeability, it is therefore possible that it is an accumulation of FB1 over time at a low rate that allows its detection in the brain. We have modified the manuscript accordingly; it seems premature at this stage to extrapolate these results further.

- L489-492: Was maybe the feed intake determined? If yes, please indicate the presence or absence of anorexic effect.

Thank you for your comment. We have clarified in the revised version of the manuscript that this study did not reveal any effects on the animals’ behaviour or feed intake.

- L498-499: please add here if the exposure length and the dietary dose were identical with the current work or if not, please describe them.

Thank you for your comment. The study was conducted in mice by intraperitoneal injection of large dose of FB1. We have clarified these points in the revised version of the manuscript.

- L505: do you also suppose slow accumulation without potent elimination?

Thank you for your comment. What we wanted to highlight with that statement is that FB1 was capable of accumulating in the brain over time, which could explain the latency period observed between exposure and the onset of clinical signs in the case of ELEM. This accumulation over time in the brain is also consistent with the accumulation we observed in the livers of chickens and, to a lesser extent, in the livers of pigs. This point has been clarified in the revised version of the manuscript. The existence of accumulation over time does not mean that the toxin is not eliminated, just that the rate of entry into the tissue is higher than the rate of exit.

- L532: instead of differential I would suggest: substrate specific.

Thank you for your comment; the text has been revised as suggested.

- L560-563: The biological meaning of the 1st sentence here is clear, but I was unable to understand the biological background of the second sentence here, based on this short description. Since there is a significant, complex relationship between SLs and the activity of cytosolic phospholipase A2, I suggest to describe this as not a new statement and to highlight/discuss those SL species (from the cohort analyzed) which may have cPLA2 modification activity. Indeed this part refers forward to the section 3., on the correlative relationships - is it maybe possible to allocate this sentence to the correlative part?

Thank you for your feedback. Yes, we used that sentence to introduce the following section, but it's true that it's clearer to put it at the end of section 4.3. In accordance with your suggestion, we have also listed the SLs that could be the cause of this effect.

- Section 4.3. Authors cite a report [48] on prioritized affinity of LOX and P450 enzymes towards AA and DHA (as compared to LA and alphaLA). Accepting this I would like to refer as well to marked differences in substrate availability, as DHA largely exceeds the abundance of LA and alphaLA, as well as sometimes AA does also, in the avian brain.

Thank you for your feedback. We have clarified this important point in the revised version of the manuscript, which reinforces our observations.

- L604 and earlier: the shortchain (e.g. 18:1/2:0) SLs are generally associated with membrane permeability. Accepting the hypothesis presented, can you as well think on altered membrane permeability developing from the increases amount of N-acetylsphingosine and dihidroCer?

Thank you for this comment. Indeed, an increase in membrane permeability secondary to the increased concentrations of 18:1/2:0 could also explain the higher passage of FB1 into the brains of chickens. This point has been clarified in the revised version of the manuscript.

- L647: I do not recommend to use the term lipidome here, as glycerophospholipids and neutral lipids were not analyzed, so I would shrink the focus only on those lipids studied.

Thank you for your comment; we have replaced “lipidome” with “sphingolipidome and oxylipidome” in the revised version of the manuscript.

Summarized: The manuscript is generally very well conceived. Its results are novel and correct. I suggest the minor/moderate modification of the ms. taking all my comments into consideration.

Thank you again for your insightful comments and for reviewing this work.

Reviewer 2 Report

This study used chickens as an animal model to investigate the mechanism of neuroinflammation induced by a nominally safe dose of fumonisins (FB1+FB2) as recommended by the EU. The authors simultaneously measured sphingolipids, oxylipins, cytokines, and related enzyme activities in the brain. The results showed that this dose of FB could cross the blood-brain barrier, significantly elevate levels of sphingosine, ceramide, and other sphingolipids, and activate PLA2c and COX2. Pro-inflammatory oxylipins were slightly increased at 14 days, while anti-inflammatory oxylipins were markedly enriched at 21 days; however, cytokines showed no significant changes. This study demonstrates that low-dose FB triggers neuroinflammation by regulating the sphingolipid-oxylipin axis independently of the cytokine pathway, suggesting potential neurotoxicity with long-term low-dose exposure. These findings provide new evidence for the safety assessment and neurotoxic mechanism of fumonisins. Publication is recommended after the following revisions:

Only two time points (14 days and 21 days) were set, which cannot reveal the critical inflection points of dynamic changes in sphingolipids and oxylipins. Is the time-gradient design reasonable?
The increased blood-brain barrier permeability was only inferred indirectly from FB1 content in the brain, without verification by gold-standard methods such as Evans blue leakage, tight junction protein (ZO-1, occludin, claudin-5) expression, or immunofluorescence staining. Can the conclusion of BBB impairment be reliably confirmed?
Only COX2 and PLA2c activities were measured in the oxylipin pathway, while the expression and activities of key enzymes in the LOX and CYP450 pathways were not determined.
Only four cytokines were detected; key pro-/anti-inflammatory cytokines including IFN-γ, IL-4, and IL-18, as well as inflammation-related chemokines in the brain, were not analyzed, so the potential role of cytokines cannot be fully excluded.
Neuroinflammation was only inferred from lipid and cytokine indicators, without detection of the expression and morphological changes of microglial marker Iba1 and astrocytic marker GFAP, leading to a lack of direct cellular evidence for neuroinflammation.
Obvious species differences exist between chickens and humans in blood-brain barrier structure, lipid metabolism, and inflammatory pathways. No verification was performed in human-derived neural cell models, resulting in insufficient scientific evidence for direct extrapolation to human health risks.
7. Professional academic editing is suggested to improve the language, optimize long sentence structures, and enhance the fluency and readability of the academic presentation.

Only two time points (14 days and 21 days) were set, which cannot reveal the critical inflection points of dynamic changes in sphingolipids and oxylipins. Is the time-gradient design reasonable?
The increased blood-brain barrier permeability was only inferred indirectly from FB1 content in the brain, without verification by gold-standard methods such as Evans blue leakage, tight junction protein (ZO-1, occludin, claudin-5) expression, or immunofluorescence staining. Can the conclusion of BBB impairment be reliably confirmed?
Only COX2 and PLA2c activities were measured in the oxylipin pathway, while the expression and activities of key enzymes in the LOX and CYP450 pathways were not determined.
Only four cytokines were detected; key pro-/anti-inflammatory cytokines including IFN-γ, IL-4, and IL-18, as well as inflammation-related chemokines in the brain, were not analyzed, so the potential role of cytokines cannot be fully excluded.
Neuroinflammation was only inferred from lipid and cytokine indicators, without detection of the expression and morphological changes of microglial marker Iba1 and astrocytic marker GFAP, leading to a lack of direct cellular evidence for neuroinflammation.
Obvious species differences exist between chickens and humans in blood-brain barrier structure, lipid metabolism, and inflammatory pathways. No verification was performed in human-derived neural cell models, resulting in insufficient scientific evidence for direct extrapolation to human health risks.
7. Professional academic editing is suggested to improve the language, optimize long sentence structures, and enhance the fluency and readability of the academic presentation.

Author Response

Reviewer 2

Thank you for taking the time to read this manuscript and for your very positive and constructive feedback. We have taken all your comments into account; they appear highlighted in yellow in the revised version of the manuscript. The changes requested by the other reviewer appear in red.

Only two time points (14 days and 21 days) were set, which cannot reveal the critical inflection points of dynamic changes in sphingolipids and oxylipins. Is the time-gradient design reasonable?

Thank you for your comment. It is true that a longer measurement period would have allowed for a better assessment of the evolution of the measured effects over time. However, this study complements our previous work conducted at 4 and 9 days of exposure in chickens to a similar exposure dose. The results obtained are therefore discussed in relation to this earlier work (ref 10 and 15 of the revised version of the manuscript). Strong consistency is observed between these two studies, both with regard to the detection of FB1 in the brain and with regard to the effects on SL and OL. Since the demonstration of FB effects on the sphingolipidome not attributable to its well-known inhibitory effect on ceramide synthesis is novel, it was important to confirm them in two independent studies. The same applies to its effects on OL. The importance of the exposure duration in the observed effects is discussed in the manuscript (L495-510). More generally, the results obtained in this study highlight the cumulative toxicity of FB1. Complementary studies on longer exposure are necessary to determine the long-term consequences of these effects. This point was clarified in the manuscript's conclusion (L666).

The increased blood-brain barrier permeability was only inferred indirectly from FB1 content in the brain, without verification by gold-standard methods such as Evans blue leakage, tight junction protein (ZO-1, occludin, claudin-5) expression, or immunofluorescence staining. Can the conclusion of BBB impairment be reliably confirmed?

Thank you for your comment. We completely agree that a further, specific study of BBB permeability would be necessary. As mentioned in the manuscript discussion, our previous work at 4 and 9 days of exposure did not detect FB1 in the brain, although high levels were found in the liver and the toxin was also detected in muscle. This study is therefore, to our knowledge, the first to reveal the presence of FB1 in the brain. These results are important because many studies report BBB permeability to FB1 without experimental evidence of the presence of FB1 in the brain. However, as you suggest, we did not actually demonstrate an increase in BBB permeability, but simply the passage of FB1 as the duration of exposure increased, which agrees with results observed in the liver. So, it is possible that the detection of FB1 in the brain is the result of a slow progressive accumulation, rather than a true alteration in BBB permeability. This point was clarified in the revised version of manuscript L505-510. Various modifications to the text were made L83, L503 and L660 to bring the document into consistency.

Only COX2 and PLA2c activities were measured in the oxylipin pathway, while the expression and activities of key enzymes in the LOX and CYP450 pathways were not determined.

Thank you for your comment. Yes, as already indicated in manuscript L376, we were unable to measure these activities due to a lack of available biological material. A significant amount of brain tissue was used to determine FB1 levels. For the reasons mentioned previously, this point seemed a priority to us.

Only four cytokines were detected; key pro-/anti-inflammatory cytokines including IFN-γ, IL-4, and IL-18, as well as inflammation-related chemokines in the brain, were not analyzed, so the potential role of cytokines cannot be fully excluded.

Thank you for your comment. We measured the cytokines most frequently measured during inflammatory processes, which are those for which fumonisins have previously shown effects. It is true that not all cytokines and chemokines were measured, and effects on these analytes could be considered. This point was clarified in the discussion of manuscript L637-640.

Neuroinflammation was only inferred from lipid and cytokine indicators, without detection of the expression and morphological changes of microglial marker Iba1 and astrocytic marker GFAP, leading to a lack of direct cellular evidence for neuroinflammation.

Thank you for your comment. We completely agree with this observation; specific measurements of the expression and activation of astrocyte and microglia markers would be necessary to characterize the inflammatory phenomenon and its intensity. This point was clarified in manuscript L643-646. We also emphasized the moderate nature of the observed inflammation, which led to a change in the manuscript's title.

Obvious species differences exist between chickens and humans in blood-brain barrier structure, lipid metabolism, and inflammatory pathways. No verification was performed in human-derived neural cell models, resulting in insufficient scientific evidence for direct extrapolation to human health risks.

Thank you for your comment. To our knowledge, the metabolism of sphingolipids and oxylipins is well conserved between animal species, as evidenced by the analytes used as standards and the numerous results we obtained with FB1 in the liver, kidneys, and plasma of chickens, turkeys, ducks, and pigs. We are not aware of any studies reporting differences in BBB permeability between poultry and mammal, and poultry are used in mechanistic studies to evaluate neurotoxicity in OECD guideline No 418 (https://doi.org/10.1787/9789264070905-en)

However, we obviously agree that the results obtained in this study cannot be directly extrapolated to humans. This point was clarified in the conclusion of the revised version of manuscript (L669-670).

Professional academic editing is suggested to improve the language, optimize long sentence structures, and enhance the fluency and readability of the academic presentation.

Thank you for your comment. The manuscript was edited by professional academic editing before submission, the proof of this work was submitted with the revised version of the manuscript. The other reviewer has considered the language as « The English is fine and does not require any improvement. ». Nevertheless, different spelling error have been corrected, they appear in red in the revised version of the manuscript.

Round 2

Reviewer 2 Report

After revision and improvement, the manuscript is well-organized with reliable conclusions. It is recommended to accept this manuscript.