Comparing the Effects of Glyphosate and Mycotoxins in the Human Follicular Microenvironment: An Exploratory Exposome Study

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

This manuscript addresses a relevant topic by investigating glyphosate (GLY) in human follicular fluid and its association with oxidative stress and hormonal parameters in IVF patients. The detection of GLY in follicular fluid is potentially novel and of interest. However, significant methodological and interpretative limitations reduce the strength of the conclusions.

The primary concern is the exposure assessment. Glyphosate was measured using an ELISA-based method without confirmatory LC-MS/MS validation, and the assay is not fully validated for follicular fluid. The absence of AMPA measurement further limits exposure characterization, raising concerns about the accuracy of the reported concentrations.

The cross-sectional design with a small sample size and no control group limits inference. The study relies on univariate correlations without adjustment for key confounders such as age, BMI, dietary exposure, and IVF stimulation effects, all of which may influence hormonal and oxidative stress parameters. As a result, the reported associations may not reflect independent effects of glyphosate.

The statistical analysis is basic, with no correction for multiple testing and generally modest effect sizes. Despite this, several findings are overinterpreted. In particular, the associations with MDA, estradiol, and cortisol are discussed in mechanistic terms that are not supported by the data and, in some cases, contradict existing literature.

Additionally, pooling of follicular fluid samples may obscure biological variability, and there appear to be inconsistencies in unit reporting and figure scaling that should be carefully reviewed.

Overall, the manuscript has potential, but the conclusions should be substantially toned down, and the analysis strengthened to support the claims.

Comments on the Quality of English Language

The English is generally understandable but could be improved for clarity and precision, particularly in the discussion section.

Author Response

This manuscript addresses a relevant topic by investigating glyphosate (GLY) in human follicular fluid and its association with oxidative stress and hormonal parameters in IVF patients. The detection of GLY in follicular fluid is potentially novel and of interest. However, significant methodological and interpretative limitations reduce the strength of the conclusions.

 

First, we would like to thank you for your time and constructive comments. Below, we have provided a point-by-point response to the questions raised and detailed the modifications made to the manuscript.

 

The primary concern is the exposure assessment. Glyphosate was measured using an ELISA-based method without confirmatory LC-MS/MS validation, and the assay is not fully validated for follicular fluid. The absence of AMPA measurement further limits exposure characterization, raising concerns about the accuracy of the reported concentrations.

 

We fully agree with the Reviewer that using an ELISA-based method without LC-MS/MS validation has inherent limitations, particularly its inability to measure AMPA. We have significantly expanded the “Limitations” section to clarify that our study reports only parental exposure to GLY. Since AMPA is the primary biotransformation metabolite with its own endocrine-disrupting properties (Ojo et al., 2025) and distinct elimination patterns (Connolly et al., 2020; Franke et al., 2020), total exposure (GLY + AMPA) is likely underestimated.  In response to these valid concerns, we have revised the Abstract, Methods, and Limitations sections to explicitly state that our findings reflect parent glyphosate only, and we emphasize the need for future LC-MS/MS confirmatory studies.  Regarding the methodological choice of ELISA, we would like to respectfully provide our rationale: While we acknowledge that instrumental-chemical analysis (e.g., LC-MS/MS) is the gold standard, the available volume of human follicular fluid per patient is strictly limited. To overcome the volume limitation, we utilized a highly sensitive commercial ELISA kit validated for human serum. Given the substantial biological and biochemical similarities between serum and follicular fluid, we considered this validated assay a reliable and highly practical approach to this specific matrix constraint. To further support the biological validity of our ELISA measurements, we included a comparison with the latest European human biomonitoring (HBM) data in the Discussion. A recent study assessing Irish families using highly sensitive GC-MS/MS technology (Connolly et al., 2022) reported ultra-low urinary concentrations, consistent with our findings. Since the measurement units (ng/mL and µg/L) are mathematically equivalent, the concentrations we detected in the follicular microenvironment and serum closely align with the established environmental background exposure in the European adult population. This strong alignment provides confidence in the reliability of our method within this low-concentration range.  Nevertheless, we strictly frame our current findings as exploratory and are actively working on establishing LC-MS/MS protocols for parallel measurements in our future cohorts.

 

Connolly, A., Koslitz, S., Bury, D., Brüning, T., Conrad, A., Kolossa-Gehring, M., Coggins, M. A., & Koch, H. M. (2020). Sensitive and selective quantification of glyphosate and aminomethylphosphonic acid (AMPA) in urine of the general population by gas chromatography-tandem mass spectrometry. Journal of Chromatography B, 1158, 122348. https://doi.org/10.1016/j.jchromb.2020.122348

 

Franke, A. A., Li, X., & Lai, J. F. (2020). Analysis of glyphosate, aminomethylphosphonic acid, and glufosinate from human urine by HRAM LC-MS. Analytical and Bioanalytical Chemistry, 412(30), 8313–8324. https://doi.org/10.1007/s00216-020-02966-1

 

Ojo, A. B., Agbeye, O. D., Ogwa, T. O., Adedoyin, D., Rotimi, D. E., & Ojo, O. A. (2025). Implications of plastic-derived endocrine disruptors on human health. Toxicology Mechanisms and Methods, 35(8), 894–918. https://doi.org/10.1080/15376516.2025.2510525

 

 

 

 

The cross-sectional design with a small sample size and no control group limits inference. The study relies on univariate correlations without adjustment for key confounders such as age, BMI, dietary exposure, and IVF stimulation effects, all of which may influence hormonal and oxidative stress parameters. As a result, the reported associations may not reflect independent effects of glyphosate.

 

Unfortunately, recruiting such a control group is challenging. Ethically, it is nearly impossible to obtain permission for follicular fluid retrieval without a medical reason. 

Additionally, control patients may not meet the desired criteria, as there is no clinical diagnostic method to definitively confirm the absence of infertility issues. Even cases where treatment is due to the husband's fertilization difficulties are not entirely reliable. One possible approach is to examine follicular fluid from individuals applying for social egg retrieval and storage. In Hungary, this is only permitted for patients undergoing treatment for cancer, which can severely impair ovarian function. Moreover, this study was not primarily aimed at establishing a reference value, but rather exploring the relationship between mycotoxin levels, oxidative stress, and changes in antioxidant systems. 

The results presented in our study are only from a cross-sectional study, and the correlations may be influenced by age, BMI, and the effects of IVF stimulation. However, these parameters did not show a correlation or linear regression relationship with either malondialdehyde or hormone levels. This supplementary information has been corrected in the publication; the results have already been presented in our previous publication. (Szőke Z, Ruff E, Plank P, Molnár Z, Hruby L, Szentirmay A, Unicsovics M, Csókay B, Varga K, Buzder T, Sipos M, Sára-Popovics K, Holéci D, Posta K, Sára L. Mycotoxin-Induced Oxidative Stress and Its Impact on Human Folliculogenesis: Examining the Link to Reproductive Health. Toxins (Basel). 2025 Nov 28;17(12):574. doi: 10.3390/toxins17120574.)

 

Discussion: The "Limitations" subsection has been expanded to acknowledge the recommendations.

 

The statistical analysis is basic, with no correction for multiple testing and generally modest effect sizes. Despite this, several findings are overinterpreted. In particular, the associations with MDA, estradiol, and cortisol are discussed in mechanistic terms that are not supported by the data and, in some cases, contradict existing literature.

 

We agree with the reviewer. We were indeed looking only for simple correlation analyses and linear regression relationships. Since our number of cases is relatively low, our results are primarily intended to show thought-provoking trends. Of the multiple correlation relationships, we highlighted only those with a significant regression relationship. We have improved the manuscript according to the reviewer's suggestions, and the applied FDR corrections and their effects on significance are indicated throughout. The strong statistical correlation observed between GLY and MDA suggests a robust biological association rather than statistical noise. However, the significance of the association between E2 and GLY disappeared after the statistical correction. We strongly believe that the positive trend is remarkable and worthy of further investigation in a larger number of cases.

 

Additionally, pooling of follicular fluid samples may obscure biological variability, and there appear to be inconsistencies in unit reporting and figure scaling that should be carefully reviewed.

 

We apologize, but we do not clearly understand the Reviewer’s concern, likely due to language use. Using an existing technical/routine procedure, follicular fluid (ff) from a patient and 1 harvest (obtained from follicles at different stages, from both ovaries) were collected after the harvest and pooled to have only „1 ff sample” from 1 patient/harvest. Currently, sampling from a single follicle is technically feasible only in unique cases, when only one follicle develops in the right or left ovary. If multiple follicles prove suitable for oocyte retrieval, one puncture can be performed on each side to avoid unnecessary complications. The text has been revised to improve clarity. However, we generally agree with the Reviewer’s biological concern. Resolving that concern would require improving the procedure and/or conducting further experiments as well. That concern is also mentioned in the Limitations section.

 

Overall, the manuscript has potential, but the conclusions should be substantially toned down, and the analysis strengthened to support the claims.

 

Thank you! We have tried to modify the conclusions accordingly.

 

Comments on the Quality of English Language

The English is generally understandable but could be improved for clarity and precision, particularly in the discussion section.

 

An additional language proofreading has been applied on manuscript.

 

Thank you for your comments.

Reviewer 2 Report

Comments and Suggestions for Authors

This is a good paper. It addresses an important question, and the work is interesting and relevant for the field of environmental health and reproduction. The study is original, especially because it shows the presence of glyphosate in human follicular fluid, which is not well documented before . This is a strong point.

The design is clear and the methods are generally well described. The combination of exposure measurement with hormonal and oxidative stress markers is a good approach. The results are coherent and easy to follow. The association between glyphosate and MDA is convincing and biologically plausible. It is also good that the authors remain careful in the interpretation, especially for IVF outcomes where no clear effect is observed.

I also appreciate that the discussion is balanced. The authors compare their results with previous literature and try to explain differences, for example for estradiol. The section on non-monotonic dose response is interesting and important.

I have a few small comments that could improve the paper. In Figure 1, it would be better to show the individual data points, because the number of subjects is relatively low. Also, since the samples are paired (serum and follicular fluid from the same individuals), it would be more informative to add connecting lines between paired values. This would make the data easier to interpret.

It is also not very clear where Figure 6 and Table A1 come from. They seem partly disconnected from the main objective of the study. The inclusion of mycotoxins is not well explained, and it is difficult to understand how this relates to the main question about glyphosate and ovarian physiology. This should be clarified, or simplified if not essential.

The authors should also cite additional studies on endocrine-disrupting effects of glyphosate. For example, Mesnage, Robin et al. (2017) showed activation of estrogen receptor alpha by glyphosate-based herbicide constituents (Food and Chemical Toxicology, doi:10.1016/j.fct.2017.07.025). This would strengthen the discussion.

Finally, the authors could also go a bit further in the discussion and speculate on possible consequences of their findings. There are interesting epidemiological studies suggesting links between glyphosate exposure and reproductive outcomes. For example, Gerona et al. (2022) reported an association between early pregnancy glyphosate exposure and reduced fetal growth (Environ Health, doi:10.1186/s12940-022-00906-3). Discussing such work would help to place the present findings in a broader clinical context.

Overall, this is a solid and useful contribution. It brings new data and opens interesting questions for future research.

Author Response

This is a good paper. It addresses an important question, and the work is interesting and relevant for the field of environmental health and reproduction. The study is original, especially because it shows the presence of glyphosate in human follicular fluid, which is not well documented before . This is a strong point.

 

The design is clear and the methods are generally well described. The combination of exposure measurement with hormonal and oxidative stress markers is a good approach. The results are coherent and easy to follow. The association between glyphosate and MDA is convincing and biologically plausible. It is also good that the authors remain careful in the interpretation, especially for IVF outcomes where no clear effect is observed.

 

I also appreciate that the discussion is balanced. The authors compare their results with previous literature and try to explain differences, for example for estradiol. The section on non-monotonic dose response is interesting and important.

 

Thank you for your comments.

 

I have a few small comments that could improve the paper. In Figure 1, it would be better to show the individual data points, because the number of subjects is relatively low. Also, since the samples are paired (serum and follicular fluid from the same individuals), it would be more informative to add connecting lines between paired values. This would make the data easier to interpret.

 

It is also not very clear where Figure 6 and Table A1 come from. They seem partly disconnected from the main objective of the study. The inclusion of mycotoxins is not well explained, and it is difficult to understand how this relates to the main question about glyphosate and ovarian physiology. This should be clarified, or simplified if not essential.

 

We thank the Reviewer for this comment. We agree that the role of Figure 6 and Table A1, as well as the inclusion of mycotoxins, required clearer justification in relation to the study’s main objective. Our original rationale for including mycotoxin measurements was that they represent co-occurring environmental contaminants (as relevant natural elements of the exposome) with known endocrine-disrupting and oxidative stress–inducing properties. As originally described in the Methods section, these compounds were analyzed to better contextualize GLY exposure within a broader “real-life” exposome scenario, in which multiple toxicants may simultaneously influence the follicular microenvironment. In particular, Figure 6 was intended to provide a comparative perspective on the relative contribution of GLY versus selected mycotoxins to oxidative stress (as reflected by MDA levels in follicular fluid), while Table A1 presents the full correlation matrix to ensure transparency of all tested associations. Our results, evaluated using similar criteria, were recently published in a paper analyzing endometrial and serum samples from endometrial carcinoma cases (Unicsovics et al. 2026 Toxicology Reports. doi: 10.1016/j.toxrep.2026.102255).

However, we acknowledge that this aspect was not sufficiently integrated into the main narrative, which may give the impression that these elements are disconnected from the central research question.

To address this, we have revised the manuscript. We have more clearly explained the rationale for including mycotoxins in the Introduction and Methods sections, emphasizing the concept of combined environmental exposure. We have also integrated the interpretation of Figure 6 into the Results and Discussion, explicitly linking it to oxidative stress mechanisms relevant to glyphosate. We have clarified the purpose of Table A1 as supplementary correlation data supporting the main analyses.

We have streamlined the presentation to maintain focus on GLY; where appropriate, we have simplified or reduced emphasis on secondary analyses.

 

The authors should also cite additional studies on endocrine-disrupting effects of glyphosate. For example, Mesnage, Robin et al. (2017) showed activation of estrogen receptor alpha by glyphosate-based herbicide constituents (Food and Chemical Toxicology, doi:10.1016/j.fct.2017.07.025). This would strengthen the discussion.

 

We agree that including additional key studies on the endocrine-disrupting effects of GLY would further strengthen the Discussion. In response, we have now incorporated the recommended study by Mesnage et al. (2017), which demonstrated activation of estrogen receptor alpha by constituents of glyphosate-based herbicides, into the revised Discussion section. This reference complements our findings on the associations between exposure to GLY and estradiol levels and further supports the potential estrogenic and endocrine-disrupting activity of GLY and related formulations.

The added citation better positions our results within the existing mechanistic literature and reinforces the biological plausibility of our observations.

 

Finally, the authors could also go a bit further in the discussion and speculate on possible consequences of their findings. There are interesting epidemiological studies suggesting links between glyphosate exposure and reproductive outcomes. For example, Gerona et al. (2022) reported an association between early pregnancy glyphosate exposure and reduced fetal growth (Environ Health, doi:10.1186/s12940-022-00906-3). Discussing such work would help to place the present findings in a broader clinical context.

 

We agree that situating our findings within a broader clinical and epidemiological context strengthens the study's interpretation and relevance.

In response, we have expanded the Discussion to include additional epidemiological evidence linking glyphosate exposure to reproductive outcomes. Specifically, we now cite Gerona et al. (2022), who reported an association between early-pregnancy glyphosate exposure and reduced fetal growth. This study provides important human evidence that complements our findings at the ovarian microenvironment level.

We have also extended our discussion to more explicitly consider the potential downstream implications of our results. Although our study did not detect direct effects on IVF outcomes, the presence of glyphosate in follicular fluid, together with its association with oxidative stress (MDA) and hormonal alterations (e.g., estradiol and cortisol), suggests that even low-level exposure may influence oocyte quality and early developmental processes at a subclinical level. In this context, our findings may represent early mechanistic changes that precede clinically detectable reproductive outcomes, as suggested by epidemiological studies.

We have taken care to frame these considerations as hypothesis-generating, given the observational nature and sample size of our study.

 

Overall, this is a solid and useful contribution. It brings new data and opens interesting questions for future research.

 

Thank you for your comment!

 

Reviewer 3 Report

Comments and Suggestions for Authors

The authors demonstrated the presence of glyphosate (GLY) in human follicular fluid. Identifying that GLY crosses the "blood-follicle barrier" is a major contribution to reproductive toxicology and environmental health. The finding that GLY levels explain nearly 30% of the variability in follicular MDA (a marker of oxidative stress) is a strong statistical result for an environmental exposure study and warrants further investigation into oocyte quality.

 

Major comments

1. The "mycotoxin" disconnect
   Mycotoxins are suddenly introduced in the Methods section (Line 155), and Figure 6 represents a Pareto analysis comparing mycotoxins to GLY. However, the introduction provides absolutely no background or rationale for why mycotoxins are being measured alongside GLY. The abstract barely mentions them (Line 32) and reports no results regarding them. The authors must either integrate mycotoxins into the introduction (explaining that they are co-occurring dietary/environmental stressors that could synergize with GLY) OR remove the mycotoxin data and Figure 6 entirely, reserving it for a separate paper. Currently, Figure 6 feels entirely out of place.
2. Statistical analysis and multiple comparisons
   Table A1 shows that the authors ran dozens of correlations between GLY and various hormones, mycotoxins, and clinical parameters. Running this many statistical tests on a relatively small sample size (n=50) highly increases the risk of Type I errors (false positives). For example, the weak correlation between ffGLY and ffE2 (p=0.042) could easily be statistical noise.The authors should explicitly state whether they applied any corrections for multiple testing (e.g., Bonferroni or False Discovery Rate/Benjamini-Hochberg). If they did not, this must be explicitly stated as a limitation in the Discussion.
3. Confounding effect of IVF stimulation
   The authors note a positive correlation between GLY and estradiol (E2), which contradicts existing literature that suggests GLY suppresses steroidogenesis. The authors rightly hypothesize that this is due to the massive doses of recombinant gonadotropins given during IVF (Lines 381-383). Because IVF stimulation artificially overrules the natural HPG axis, the relationship between GLY and E2 in this specific cohort is highly confounded. This should be highlighted more prominently in the limitations section, as these hormone levels do not reflect a natural physiological state.
4. ELISA vs. analytical chemistry
   The authors acknowledge that using ELISA instead of LC-MS/MS is a limitation and that ELISA does not detect AMPA (the primary metabolite of GLY). Given that AMPA is often the primary driver of toxicity, the authors should explicitly state in the Abstract and Conclusion that these results represent parentglyphosate exposure only, and total exposure (GLY + AMPA) is likely higher.

 

Minor comments

Abstract: Line 25: There is a typo: "...in follicular fluid and serum, as6 their relationships..." This should likely be "...and their relationships...".

Missing keywords: Lines 50-51: The template text was left in the manuscript: " Keywords: keyword 1; keyword 2; keyword 3 (List three to ten pertinent keywords..." Please replace this with the actual keywords for the paper.

Editing: Line 436: There appears to be a fragmented sentence left over from editing: "...analyze this in more detail. reparative effect. Direct evidence..." Please delete or fix "reparative effect."

Figure consistency: Figures 2 and 3 show both the raw data scatterplot (A) and the square-root transformed scatterplot (B). However, Figures 4 and 5 only show one plot. For consistency and space-saving, it is recommended to only show the transformed data plots (where the linear regression is statistically valid) or ensure all figures follow the same A/B format.

Figure 5's caption references an "(A)," but there is no "(B)" in the figure.

Table A1 formatting: The formatting of Table A1 is very difficult to read. The rows for "Spearman r" and "r" are confusing, and splitting the tables across pages makes it hard to follow. Consider moving the entire mycotoxin dataset out of this table to simplify it, and format the p-values and confidence intervals to take up less vertical space.

Abbreviations: Line 517: Ensure all abbreviations used in the text are present here. For example, AMPA and GBH are listed, but MT (Melatonin) is missing from the list, despite its extensive discussion in the text.

Author Response

The authors demonstrated the presence of glyphosate (GLY) in human follicular fluid. Identifying that GLY crosses the "blood-follicle barrier" is a major contribution to reproductive toxicology and environmental health. The finding that GLY levels explain nearly 30% of the variability in follicular MDA (a marker of oxidative stress) is a strong statistical result for an environmental exposure study and warrants further investigation into oocyte quality.

 

We would like to thank you for your constructive comments. We have tried to provide a point-by-point response to the questions raised and detailed the modifications made to the manuscript.

 

 

Major comments

 

The "mycotoxin" disconnect

Mycotoxins are suddenly introduced in the Methods section (Line 155), and Figure 6 represents a Pareto analysis comparing mycotoxins to GLY. However, the introduction provides absolutely no background or rationale for why mycotoxins are being measured alongside GLY. The abstract barely mentions them (Line 32) and reports no results regarding them. The authors must either integrate mycotoxins into the introduction (explaining that they are co-occurring dietary/environmental stressors that could synergize with GLY) OR remove the mycotoxin data and Figure 6 entirely, reserving it for a separate paper. Currently, Figure 6 feels entirely out of place.

 

We thank the Reviewer for this clear and constructive comment. We fully agree that, in the original version, the inclusion of mycotoxins was insufficiently justified and not well integrated into the overall narrative, which may have made Figure 6 appear disconnected from the main objective.

Our intention in including mycotoxin measurements was to account for co-occurring environmental and dietary contaminants that may also act as endocrine disruptors and oxidative stressors within the same biological compartment. Because follicular fluid reflects a complex exposome, simultaneous exposure to multiple compounds (including mycotoxins, GLY, and possibly others) may better represent real-world conditions and could influence the follicle’s oxidative and hormonal milieu. Our findings, assessed using comparable criteria, were recently published in a paper analyzing endometrial and serum samples from endometrial carcinoma cases. (Unicsovics et al. 2026 Toxicology Reports. doi: 10.1016/j.toxrep.2026.102255).

This rationale was not adequately introduced or consistently applied in the Abstract and Discussion, leading to a lack of coherence. In response, we have revised the manuscript as follows:

We have explicitly introduced the concept of co-exposure in the Introduction, clarifying why mycotoxins were measured alongside GLY as relevant dietary/environmental stressors.

We have briefly summarized the mycotoxin-related approach in the Abstract to ensure consistency.

We have revised the Results and Discussion sections to better contextualize Figure 6, clearly stating that it serves as a comparative, exploratory analysis of oxidative stress contributors rather than a primary outcome. At the same time, to maintain focus, we have reduced the emphasis on mycotoxins and clarified that these analyses are secondary and hypothesis-generating.

We also carefully considered the Reviewer’s suggestion to remove these data entirely. However, we believe that with improved framing, their inclusion adds value by situating glyphosate exposure within a more realistic multi-exposure context without detracting from the manuscript's main message.

 

Statistical analysis and multiple comparisons

Table A1 shows that the authors ran dozens of correlations between GLY and various hormones, mycotoxins, and clinical parameters. Running this many statistical tests on a relatively small sample size (n=50) highly increases the risk of Type I errors (false positives). For example, the weak correlation between ffGLY and ffE2 (p=0.042) could easily be statistical noise.The authors should explicitly state whether they applied any corrections for multiple testing (e.g., Bonferroni or False Discovery Rate/Benjamini-Hochberg). If they did not, this must be explicitly stated as a limitation in the Discussion.

 

 

We thank the reviewer for highlighting the issue of multiple testing. We agree that the exploratory analyses involved numerous comparisons, which may increase the risk of type I error.

To address this, we have now applied false discovery rate (FDR) correction (Benjamini–Hochberg method) to all exploratory analyses where it was relevant. Importantly, our primary hypothesis-driven analyses (e.g., the association between follicular GLY and MDA, E2) were defined a priori and therefore were not subjected to multiple testing correction. Our key finding (GLY-MDA association) remained statistically significant and robust across multiple models after FDR correction.

We have clarified this distinction in the revised manuscript.

 

Confounding effect of IVF stimulation

The authors note a positive correlation between GLY and estradiol (E2), which contradicts existing literature that suggests GLY suppresses steroidogenesis. The authors rightly hypothesize that this is due to the massive doses of recombinant gonadotropins given during IVF (Lines 381-383). Because IVF stimulation artificially overrules the natural HPG axis, the relationship between GLY and E2 in this specific cohort is highly confounded. This should be highlighted more prominently in the limitations section, as these hormone levels do not reflect a natural physiological state.

 

We fully agree that controlled ovarian stimulation in IVF represents a highly non-physiological endocrine condition and may substantially confound the observed relationship between glyphosate exposure and estradiol levels.

As correctly pointed out, the administration of supraphysiological doses of recombinant gonadotropins overrides the natural HPG axis, resulting in hormone dynamics that do not reflect normal physiological regulation. This is particularly relevant when interpreting the positive association we observed between GLY and E2, which contrasts with much of the existing literature suggesting inhibitory effects of glyphosate on steroidogenesis.

In response to this comment, we have revised the manuscript to:

More prominently emphasize this issue in the Limitations section, explicitly stating that hormone levels measured in this cohort are strongly influenced by IVF stimulation and therefore may not reflect natural endocrine physiology.

Further temper our interpretation of the GLY–E2 association, clearly framing it as context-dependent and potentially confounded by treatment-related hormonal modulation.

Clarify that the observed associations should be interpreted with caution and cannot be directly extrapolated to non-stimulated populations.

 

ELISA vs. analytical chemistry

The authors acknowledge that using ELISA instead of LC-MS/MS is a limitation and that ELISA does not detect AMPA (the primary metabolite of GLY). Given that AMPA is often the primary driver of toxicity, the authors should explicitly state in the Abstract and Conclusion that these results represent parentglyphosate exposure only, and total exposure (GLY + AMPA) is likely higher.

 

We thank the Reviewer for emphasizing this important point. We agree that AMPA plays a significant role in toxicity and acknowledge that not including it is a limitation. We also concur that relying solely on an ELISA-based method without LC-MS/MS validation has its drawbacks, especially since it doesn't measure AMPA. We have expanded the “Limitations” section to clarify that our study only reports parental exposure to GLY. Because AMPA is a primary metabolite with its own endocrine-disrupting effects (Ojo et al., 2025) and different elimination patterns (Connolly et al., 2020; Franke et al., 2020), total exposure (GLY + AMPA) is probably underestimated. In response, we revised the Abstract, Methods, and Limitations sections to explicitly state that our findings pertain solely to parent glyphosate and to highlight the need for future LC-MS/MS confirmation. Concerning our choice of ELISA, we want to clarify our reasoning: While LC-MS/MS is the gold standard, the limited volume of human follicular fluid per patient guided us to use a highly sensitive commercial ELISA kit validated for human serum. Since serum and follicular fluid are biochemically similar, we believed this validated assay was a reliable and practical approach given the sample constraints. To support the biological relevance of our measurements, we compared our results with recent European human biomonitoring (HBM) data discussed in our study. A recent Irish study using sensitive GC-MS/MS technology (Connolly et al., 2022) found urinary concentrations that matched our low exposure levels. Because the units (ng/mL and µg/L) are equivalent, the concentrations we measured in follicular fluid and serum align well with environmental background levels in European adults. This consistency strengthens confidence in our method within this low-concentration range. We emphasize that our current findings are exploratory and that we are developing LC-MS/MS protocols for future measurements in parallel.

 

Connolly, A., Koslitz, S., Bury, D., Brüning, T., Conrad, A., Kolossa-Gehring, M., Coggins, M. A., & Koch, H. M. (2020). Sensitive and selective quantification of glyphosate and aminomethylphosphonic acid (AMPA) in urine of the general population by gas chromatography-tandem mass spectrometry. Journal of Chromatography B, 1158, 122348. https://doi.org/10.1016/j.jchromb.2020.122348

 

Franke, A. A., Li, X., & Lai, J. F. (2020). Analysis of glyphosate, aminomethylphosphonic acid, and glufosinate from human urine by HRAM LC-MS. Analytical and Bioanalytical Chemistry, 412(30), 8313–8324. https://doi.org/10.1007/s00216-020-02966-1

 

Ojo, A. B., Agbeye, O. D., Ogwa, T. O., Adedoyin, D., Rotimi, D. E., & Ojo, O. A. (2025). Implications of plastic-derived endocrine disruptors on human health. Toxicology Mechanisms and Methods, 35(8), 894–918. https://doi.org/10.1080/15376516.2025.2510525

 

 

Minor comments

 

Abstract: Line 25: There is a typo: "...in follicular fluid and serum, as6 their relationships..." This should likely be "...and their relationships...".

 

Thank you! It is now corrected.

 

Missing keywords: Lines 50-51: The template text was left in the manuscript: " Keywords: keyword 1; keyword 2; keyword 3 (List three to ten pertinent keywords..." Please replace this with the actual keywords for the paper.

 

We apologize. Keywords are now inserted.

 

Editing: Line 436: There appears to be a fragmented sentence left over from editing: "...analyze this in more detail. reparative effect. Direct evidence..." Please delete or fix "reparative effect."

 

We apologize. It has been deleted

 

Figure consistency: Figures 2 and 3 show both the raw data scatterplot (A) and the square-root transformed scatterplot (B). However, Figures 4 and 5 only show one plot. For consistency and space-saving, it is recommended to only show the transformed data plots (where the linear regression is statistically valid) or ensure all figures follow the same A/B format.

 

We have corrected all figures as requested by the reviewer. We have added A and B divisions everywhere.

 

Figure 5's caption references an "(A)," but there is no "(B)" in the figure.

 

Corrected.

 

Table A1 formatting: The formatting of Table A1 is very difficult to read. The rows for "Spearman r" and "r" are confusing, and splitting the tables across pages makes it hard to follow. Consider moving the entire mycotoxin dataset out of this table to simplify it, and format the p-values and confidence intervals to take up less vertical space.

 

Table A1 was corrected as requested. 

 

Abbreviations: Line 517: Ensure all abbreviations used in the text are present here. For example, AMPA and GBH are listed, but MT (Melatonin) is missing from the list, despite its extensive discussion in the text.

 

Thank you! Abbreviations are now reviewed and corrected.

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

The authors have addressed all the issues properly.

Author Response

On behalf of the authors, I would also like to thank you for your help!

Reviewer 2 Report

Comments and Suggestions for Authors

Thank you for taking into account my comments 

Author Response

On behalf of the authors, I would also like to thank you for you help!

Reviewer 3 Report

Comments and Suggestions for Authors

The authors have done a good job addressing standard methodological limitations, specifically by applying the False Discovery Rate (FDR) correction for multiple testing and acknowledging the limitations of using ELISA over LC-MS/MS for follicular fluid. However, more revisions are required.

Major comments

1. Post-FDR Results

• The application of the Benjamini–Hochberg FDR is good. However, the Discussion section heavily interprets the biological mechanisms of the GLY-estradiol (E₂) and GLY-cortisol relationships. Since the authors explicitly state in the Results that the associations between GLY and E2 (as well as GLY and cortisol) lost significance after FDR correction, the Discussion must reflect this more cautiously. Sentences like “The observed positive association between serum and ffGLY concentrations and E2 levels suggests that GLY acts as an endocrine disruptor at both systemic and follicular levels" should be softened. It should be explicitly framed as a "potential trend requiring validation in larger cohorts" rather than a confirmed finding.

2. The role of mycotoxins

• The inclusion of mycotoxins feels slightly disjointed from the rest of the narrative. The title focuses heavily on glyphosate, yet mycotoxins are abruptly introduced at the end of the introduction and featured prominently in Figure 6. If mycotoxins are meant to be a core part of this "exposome" study, they should be mentioned in the title (e.g., "Environmental Exposure to Glyphosate and Mycotoxins..."). Alternatively, if they are only acting as a negative control/comparator for oxidative stress potency (as Figure 6 suggests), this rationale needs to be clearly and explicitly stated in the Introduction and the Results so the reader understands why they were measured alongside GLY.

3. Figure 6

• Figure 6 is an interesting way to compare the effects of mycotoxins vs. GLY on MDA concentrations. However, the Methods section lacks a clear description of how the "relative (β) values of effect" were calculated for this specific Pareto chart. Please add a brief explanation in the Statistical Analysis section detailing the specific regression models used to generate the β values for Figure 6.

4. ELISA vs. LC-MS/MS

• The authors provide a robust and honest appraisal of the ELISA limitations in the limitations section (pages 13-14). Because follicular fluid is a complex, lipid-rich matrix, cross-reactivity is a genuine concern. To protect the manuscript from future critique, it is recommended to add the word "preliminary" or "exploratory" to the conclusion. For example: "In our study, we reported, probably for the first time, the preliminary evidence of GLY presence in human ff..."

 

Minor comments

1. Abbreviations:

• Make sure all abbreviations are consistent throughout the text and figures.
• Check the abbreviation for alpha-zearalenol. In the text and Figure 6, it is sometimes ffaZOL and, other times, α-ZOL. Standardize this throughout.

2. Figure:

• Figure 2: The caption states, "(A) and after a square root transformation (B)." However, the provided figure only shows one graph (which appears to be the transformed data, though the axes labels are just "ff Glyphosate" and "ff Malondialdehyde" without the square root symbol “√” on the actual axis numbers, though the axis title has it). Please ensure the visual panels match the caption.
• Figures 3 & 5: The text refers to a "shaded pink area,” but in some prints/screens this may just look like light red/grey. Ensure the color descriptions match the final rendered figures.
• Table A1: The formatting in the far-right columns is cut off (e.g., the 95% confidence intervals end in "-0.123" with no upper bound visible; "total r” and "sCorti"). Please adjust the table cell widths so all numerical values and variable names are fully visible.

• Page 14: “In this study, we do not have results for AMPA concentration in the ff, but we know this metabolite may also influence hormonal, antioxidant, and oxidative stress processes. “ Consider rephrasing slightly for academic tone: "Although AMPA concentrations in the ff were not quantified in this study, this major metabolite may also influence..."

Author Response

Comments and Suggestions for Authors

The authors have done a good job addressing standard methodological limitations, specifically by applying the False Discovery Rate (FDR) correction for multiple testing and acknowledging the limitations of using ELISA over LC-MS/MS for follicular fluid. However, more revisions are required.

 

Dear Reviewer, We are grateful for your thorough and constructive criticism. We are pleased that you found the methodological limitations and the application of FDR correction appropriate. We have fully incorporated your suggested changes into the manuscript, which have significantly improved the quality and scientific accuracy of the article.

Major comments

1. Post-FDR Results

• The application of the Benjamini–Hochberg FDR is good. However, the Discussion section heavily interprets the biological mechanisms of the GLY-estradiol (E₂) and GLY-cortisol relationships. Since the authors explicitly state in the Results that the associations between GLY and E2 (as well as GLY and cortisol) lost significance after FDR correction, the Discussion must reflect this more cautiously. Sentences like “The observed positive association between serum and ffGLY concentrations and E2 levels suggests that GLY acts as an endocrine disruptor at both systemic and follicular levels" should be softened. It should be explicitly framed as a "potential trend requiring validation in larger cohorts" rather than a confirmed finding.

We agree with the comment. The Discussion section has been extensively revised. Because the significance of the associations between GLY-estradiol and GLY-cortisol was lost after FDR correction, we have refined our earlier strong statements. We emphasize that these observations are "potential trends that require validation in larger cohorts" and refer to them as unconfirmed facts.

2. The role of mycotoxins

• The inclusion of mycotoxins feels slightly disjointed from the rest of the narrative. The title focuses heavily on glyphosate, yet mycotoxins are abruptly introduced at the end of the introduction and featured prominently in Figure 6. If mycotoxins are meant to be a core part of this "exposome" study, they should be mentioned in the title (e.g., "Environmental Exposure to Glyphosate and Mycotoxins..."). Alternatively, if they are only acting as a negative control/comparator for oxidative stress potency (as Figure 6 suggests), this rationale needs to be clearly and explicitly stated in the Introduction and the Results so the reader understands why they were measured alongside GLY.

 

Thank you for this insightful comment. We have revised the manuscript title to “Comparing the Effects of Glyphosate and Mycotoxins in the Human Follicular Microenvironment: An Exploratory Exposome Study” to make the exposome-level focus on mycotoxins clear from the outset. We have also clarified the rationale for including mycotoxins in the Introduction and Results sections, explicitly stating that they served as a negative control/baseline for assessing the oxidative stress (MDA) potential of glyphosate.

3. Figure 6

• Figure 6 is an interesting way to compare the effects of mycotoxins vs. GLY on MDA concentrations. However, the Methods section lacks a clear description of how the "relative (β) values of effect" were calculated for this specific Pareto chart. Please add a brief explanation in the Statistical Analysis section detailing the specific regression models used to generate the β values for Figure 6

We agree that the calculation of the beta values shown in the figure was not adequately explained. We have supplemented the caption and clarified the data’s origin and calculation.

 

4. ELISA vs. LC-MS/MS

• The authors provide a robust and honest appraisal of the ELISA limitations in the limitations section (pages 13-14). Because follicular fluid is a complex, lipid-rich matrix, cross-reactivity is a genuine concern. To protect the manuscript from future critique, it is recommended to add the word "preliminary" or "exploratory" to the conclusion. For example: "In our study, we reported, probably for the first time, the preliminary evidence of GLY presence in human ff..."

 

To avoid future criticism and for scientific accuracy, we have refined the Conclusions section by adding the words "preliminary" and "exploratory." 

 

Minor comments

1. Abbreviations:

• Make sure all abbreviations are consistent throughout the text and figures.
• Check the abbreviation for alpha-zearalenol. In the text and Figure 6, it is sometimes ffaZOL and, other times, α-ZOL. Standardize this throughout.

We have standardized the abbreviation of alpha-zearalenol throughout the text and in the figures: the αZOL format is used everywhere instead of the previous ffaZOL.

2. Figure:

• Figure 2: The caption states, "(A) and after a square root transformation (B)." However, the provided figure only shows one graph (which appears to be the transformed data, though the axes labels are just "ff Glyphosate" and "ff Malondialdehyde" without the square root symbol “√” on the actual axis numbers, though the axis title has it). Please ensure the visual panels match the caption.

Thank you. We revised Figure 2.

 

 

• Figures 3 & 5: The text refers to a "shaded pink area,” but in some prints/screens this may just look like light red/grey. Ensure the color descriptions match the final rendered figures.

Thank you for your comment. We have improved the illustrations.

 

 

• Table A1: The formatting in the far-right columns is cut off (e.g., the 95% confidence intervals end in "-0.123" with no upper bound visible; "total r” and "sCorti"). Please adjust the table cell widths so all numerical values and variable names are fully visible.

We have attached Table A1 in a separate file and corrected it as requested.

 

• Page 14: “In this study, we do not have results for AMPA concentration in the ff, but we know this metabolite may also influence hormonal, antioxidant, and oxidative stress processes. “ Consider rephrasing slightly for academic tone: "Although AMPA concentrations in the ff were not quantified in this study, this major metabolite may also influence..."

Thank you for your suggestion; it has been corrected.

Thank you for your many helps!