Thrombin-Induced Microglia Activation Modulated through Aryl Hydrocarbon Receptors

Round 1

Reviewer 1 Report

The study by Sheu and Pan et al. is a very interesting work investigating the modulatory role of aryl hydrocarbon receptors in thrombin-mediated microglia activation, both in vitro and in vivo. The methodology is rigorous and precise and the results are very clearly presented despite the complexity of the study. Furthermore, the Authors correctly cite the literature on the subject, clearly supporting the results obtained. I only have a few minor suggestions for the Authors: 

- In the Results section, explanatory sentences are sometimes given that would perhaps be more appropriate in the methods and/or introduction (e.g. lines 95-99, 116-119, 217-219, 233-237). This could make the results presented more fluent and linear. 

- Pragraph 2.4: The Authors rightly focus on the measurement of pro-inflammatory cytokines. However, were the anti-inflammatory ones also measured? If so, does the change in their levels also follow a consistent trend in the deleted AhR models? 

- Line 321: probable repetition (images)

- Considering the link between thrombin and endothelial cells, has eNOS activity been assessed? If so, are there differences in its activity in AhR deleted models? 

- I would recommend that the Authors include a small final table to summarise the results in a concise and qualitative way, perhaps dividing between in vitro and in vivo, in order to make the final message even clearer and more accessible 

- Paragraph 5: I suggest that the Authors include a short sentence in the conclusion in which possible therapeutic applications (also from a translational point of view) and future studies could be hypothesised in order to better evaluate the effects of the results obtained 

Author Response

Response to comments

Reviewer 1:

The study by Sheu and Pan et al. is a very interesting work investigating the modulatory role of aryl hydrocarbon receptors in thrombin-mediated microglia activation, both in vitro and in vivo. The methodology is rigorous and precise and the results are very clearly presented despite the complexity of the study. Furthermore, the Authors correctly cite the literature on the subject, clearly supporting the results obtained. I only have a few minor suggestions for the Authors: 

- In the Results section, explanatory sentences are sometimes given that would perhaps be more appropriate in the methods and/or introduction (e.g. lines 95-99, 116-119, 217-219, 233-237). This could make the results presented more fluent and linear. 

Response to comments:

Yes, it is. We edited these paragraph as you recommended and deleted these unnecessary explanatory sentences.

- Pragraph 2.4: The Authors rightly focus on the measurement of pro-inflammatory cytokines. However, were the anti-inflammatory ones also measured? If so, does the change in their levels also follow a consistent trend in the deleted AhR models? 

Response to comments:

Yes, it is. It is a good point that we should measure the inflammatory and anti-inflammatory cytokines to assess the effect of AhR on the regulation of the inflammatory response. Based on the hypothesis of deletion of AhR to accelerated the inflammatory response induced by thrombin, we only measure the inflammatory response without the work on the anti-inflammatory cytokines. 

- Line 321: probable repetition (images)

Response to comments:

Yes, it is. In Figure 9, we focus the AhR effect on the neuron and the imaging did not reveal the typical alteration, which is ambiguous to the reviewers. We also re-edited Figure 8 and 9 to make it much clearer.

- Considering the link between thrombin and endothelial cells, has eNOS activity been assessed? If so, are there differences in its activity in AhR deleted models? 

Response to comments:

Yes, it is. We measured the expression of eNOS in the wild or AhR deleted microglia cells treated with thrombin. But we did not measure the expression of e-NOS in the above said condition.

I would recommend that the Authors include a small final table to summarise the results in a concise and qualitative way, perhaps dividing between in vitro and in vivo, in order to make the final message even clearer and more accessible 

Response to comments:

Yes, it is. We add one paragraph in the results” To summarize the results in a concise and qualitative way, we made the graph consisting of hypothesis of AhR involved in thrombin induced microglia cell activation and the tables of in-vitro and in-vivo experiment to make this study clearer (Supplementary Figure 1).

- Paragraph 5: I suggest that the Authors include a short sentence in the conclusion in which possible therapeutic applications (also from a translational point of view) and future studies could be hypothesised in order to better evaluate the effects of the results obtained 

Response to comments: Yes, it is. We revised the conclusion section” In conclusion, thrombin caused the activation of microglia through increased vessel permeability, expression of inflammatory response, and phenotype of M1 microglia as well the MMP activity. Deleting AhR augmented the above detrimental effects. These findings indicate the modulation of AhR is essential for the regulation of thrombin-induced brain damages and the AhR agonist may harbor the potential therapeutic effect in thrombin induced neurodegenerative disorder.    

Reviewer 2 Report

[Comments]

The authors reported that thrombin induced microglia activation modulated through aryl hydrocarbon receptors using multiple methods. However, the reviewer has some major concerns.

1)      The authors performed Mile’s assay with Evans Blue dye to examine cerebrovascular permeability. The showed dye leakage in both ears of mice in Fig. 3A and quantitative permeability data in cerebral cortex endothelial cells in Fig. 3D. However, they did not show images showing increased vascular permeability in cerebral cortex. The authors should show images showing thrombin-induced increased vascular permeability in cerebral cortex. 

2)      In Figs. 6, 8, and 9, the authors showed images of iNOS, CD68, Arginase-1-positive microglia, Iba1-positive cells, 8-oxo-dG, Fluoro-jade b, and TUNEL-positive cells, respectively. Semi-quantitative evaluation using the images were performed. However, Western blotting analysis should be performed to evaluate thrombin-induced increased numbers of them.

3)      There are differences in morphology between resting and activated microglia. The authors should show morphological characteristics of resting and activated microglia.

I have no comment on English Language.

Author Response

Response to comments

Reviewer 2:

The authors reported that thrombin induced microglia activation modulated through aryl hydrocarbon receptors using multiple methods. However, the reviewer has some major concerns.

• The authors performed Mile’s assay with Evans Blue dye to examine cerebrovascular permeability. The showed dye leakage in both ears of mice in Fig. 3A and quantitative permeability data in cerebral cortex endothelial cells in Fig. 3D. However, they did not show images showing increased vascular permeability in cerebral cortex. The authors should show images showing thrombin-induced increased vascular permeability in cerebral cortex. 

Response to comments: Yes, it is. It is a good point that we revised the Fig 3 by adding the representative photography of Evan blue dye distribution in four groups and the quantitative analysis.   

• In Figs. 6, 8, and 9, the authors showed images of iNOS, CD68, Arginase-1-positive microglia, Iba1-positive cells, 8-oxo-dG, Fluoro-jade b, and TUNEL-positive cells, respectively. Semi-quantitative evaluation using the images were performed. However, Western blotting analysis should be performed to evaluate thrombin-induced increased numbers of them.

Response to comments: Yes, it is. We add the data of western blot analysis in Figure 6. We also revised Fig 8 and 9 to them more recognizable.

• There are differences in morphology between resting and activated microglia. The authors should show morphological characteristics of resting and activated microglia.

Response to comments: Yes, it is. We revised the Figure 5. In Figure 5C, we found that resting microglia cells either in wild or AhRKO mice showed little ramification. But in wild or AhRKO mice subjected to thrombin injection, the remarkable ramification in microglia were found. The number of microglia showed the significantly increased in AhRKO compared to wild type.        

Reviewer 3 Report

The paper by Pan and colleagues deals with effects of the interplay of the two pleiotropic factors involved in the pathology of neurodegeneration i.e. thrombin and AhR. These effects were studied on microglial cells, hippocampus slices and in vivo on mice.

General problems:

I do not find explanation how relevant to physiological conditions are the concentrations of thrombin used in the authors experiments. In the method section there is 4000 u/ml. In the text of Results differ from 5 to 300 u/ml. Which one is the true one? In the platelets clotting assays traces of thrombin are enough to start the process. Therefore, authors should discuss this problem.

The methods section is lacking the description of docking simulations with ZDOCK. The reader cannot find what PDB structures are used. Are they dissected from complex structures or not. On the other hand, docking alone is a weak evidence for the interaction between proteins. A physical interaction should be evidenced by, for example, pull- down assay. Last section of methods - statistical analysis has some leftovers from the IJMS template.

Summarising: I do not recommend major revision since, the paper needs really extensive rewriting and new experiments. It’s a pity, because it has a large experimental part and a lot of work has been done.

I found some difficulties reading the description of docking results.

Author Response

Response to comments

Reviewer 3:

The paper by Pan and colleagues deals with effects of the interplay of the two pleiotropic factors involved in the pathology of neurodegeneration i.e. thrombin and AhR. These effects were studied on microglial cells, hippocampus slices and in vivo on mice.

General problems:

I do not find explanation how relevant to physiological conditions are the concentrations of thrombin used in the authors experiments. In the method section there is 4000 u/ml. In the text of Results differ from 5 to 300 u/ml. Which one is the true one? In the platelets clotting assays traces of thrombin are enough to start the process. Therefore, authors should discuss this problem.

Response to comments: Yes, it is. Thank you for your reminding. It is our calculation error and we revised our data according to methodology.

The methods section is lacking the description of docking simulations with ZDOCK. The reader cannot find what PDB structures are used. Are they dissected from complex structures or not. On the other hand, docking alone is a weak evidence for the interaction between proteins. A physical interaction should be evidenced by, for example, pull- down assay. Last section of methods - statistical analysis has some leftovers from the IJMS template.

Response to comments: Thank you for your comments, which are our weakness in this study. In our upcoming study using the docking method, we should add the data in the physical interaction. 

Summarizing: I do not recommend major revision since, the paper needs really extensive rewriting and new experiments. It’s a pity, because it has a large experimental part and a lot of work has been done.

Response to comments: Thank you for your prestigious comments.

We tried to respond to comments with points to points to make the manuscript readable and meaningful.

Reviewer 4 Report

The manuscript by Meei-Ling Sheu and Liang-Yi Pan et al shows some interesting findings on the role of AHR in thrombin-associated inflammatory response. The in vitro results are compelling but the in vivo studies must be significantly improved to be able to come to the conclusions made in the manuscript. In general, there is very less information in the results section about the experimental plan or the methodology employed. What mice were compared, at what age, and which part of the brain was assessed? Readers shouldn’t have t juggle between figure legends and the results section to understand what’s going on. The rationale behind experiments is not explained, it looks a lot like the authors just wanted to put in as much information without trying to make a story of their findings.

If the authors want to say that all these effects upon AHRKO are mediated by microglia, the knockout should be microglia specific maybe using a cx3cr1 cre? If not, they can leave it at the effect of thrombin/AHR on the various aspects of inflammation.

1.       Authors should include representative images for microglial culture in Figures 2 and 4 or include them in the supplementary. With such drastic changes in the inflammatory response, it is important to see how the cells appear, and the density of the cells before and after treatment. Were there any morphological changes in the cells?

2.       Blot in Figure 5 was performed from only microglia or whole brain lysate? MMP9/MMP2 are produced in huge amounts from the border-associated macrophages. If the result is represented from the whole brain, it is most likely from the BAMS and not microglia. Also, the images are not very convincing in Figure 5 C. It would be better if they showed separate green and red channels followed by merge.

3.       The figure legend for Figure 6 says the microglial marker used is IBA1. That is not represented in the figure. Again, to confirm that iNOS expression or ARG1 is being shown in microglia colocalization of these signals must be shown with IBA1, preferably single channel followed by merge. A suggestion would be to use flow cytometry to show these expressions in microglia. That could be a perfect quantitative result. Also, iNOS and ARG1 are not the cardinal M1/M2 markers, to be able to really distinguish M1/M2 a panel of multiple markers is preferable. So the language has to be mellowed to state that as assessed by these two markers we observed so and so...

4.       Figure 7 requires representative images. Just the quantification is not enough to show such drastic effects represented.

5.       Authors should rephrase, “The inflammatory response in microglia induced by thrombin was found to be highly 317 correlated with survival of nearby neurons.” They have shown an increase in inflammatory signature in microglia in vitro, it was not similarly shown in vivo in microglial cells by sequencing or flow, or any immunoassay. It is acceptable that AHR deletion is affecting microglia and inducing cell death. But this by no means shows that microglial inflammation is inducing this neuronal cell death. Thrombin by itself can result in the production of high molecular peptides that are capable of inducing cell death.

English is fine. But a lot of the words are hyphenated. Consider checking that. 

Author Response

Response to comments

Reviewer 4:

The manuscript by Meei-Ling Sheu and Liang-Yi Pan et al shows some interesting findings on the role of AHR in thrombin-associated inflammatory response. The in vitro results are compelling but the in vivo studies must be significantly improved to be able to come to the conclusions made in the manuscript. In general, there is very less information in the results section about the experimental plan or the methodology employed. What mice were compared, at what age, and which part of the brain was assessed? Readers shouldn’t have t juggle between figure legends and the results section to understand what’s going on. The rationale behind experiments is not explained, it looks a lot like the authors just wanted to put in as much information without trying to make a story of their findings.

If the authors want to say that all these effects upon AHRKO are mediated by microglia, the knockout should be microglia specific maybe using a cx3cr1 cre? If not, they can leave it at the effect of thrombin/AHR on the various aspects of inflammation.

1. Authors should include representative images for microglial culture in Figures 2 and 4 or include them in the supplementary. With such drastic changes in the inflammatory response, it is important to see how the cells appear, and the density of the cells before and after treatment. Were there any morphological changes in the cells?

Response to comments: It is a good point. But we did not take photograph in Figure 2 and 4 during the experiment. Only in the revised figure 5, we found that resting microglia cells either in wild or AhRKO mice showed little ramification. But in mild or AhRKO mice subjected to thrombin injection, the remarkable ramification in microglia were found. The number of microglia showed the significantly increased in AhRKO compared to wild type.        

2. Blot in Figure 5 was performed from only microglia or whole brain lysate? MMP9/MMP2 are produced in huge amounts from the border-associated macrophages. If the result is represented from the whole brain, it is most likely from the BAMS and not microglia. Also, the images are not very convincing in Figure 5 C. It would be better if they showed separate green and red channels followed by merge.

Response to comments: Yes, it is. The tissue for analysis was obtained from the region of the hippocampus. We also revised Figure 5C including immunohistochemistry staining in IBA1, MMP 9 and merged imaging of IBA1 and MMP 9.  

3. The figure legend for Figure 6 says the microglial marker used is IBA1. That is not represented in the figure. Again, to confirm that iNOS expression or ARG1 is being shown in microglia colocalization of these signals must be shown with IBA1, preferably single channel followed by merge. A suggestion would be to use flow cytometry to show these expressions in microglia. That could be a perfect quantitative result. Also, iNOS and ARG1 are not the cardinal M1/M2 markers, to be able to really distinguish M1/M2 a panel of multiple markers is preferable. So the language has to be mellowed to state that as assessed by these two markers we observed so and so...

Response to comments: Yes, it is. We add the data of western blot analysis in Figure 6 D and E. 

4. Figure 7 requires representative images. Just the quantification is not enough to show such drastic effects represented.

Response to comments: Yes, it is. We revised Figure 7 by adding the representative of photograph of measurement in lesion volumes. 

5. Authors should rephrase, “The inflammatory response in microglia induced by thrombin was found to be highly 317 correlated with survival of nearby neurons.” They have shown an increase in inflammatory signature in microglia in vitro, it was not similarly shown in vivo in microglial cells by sequencing or flow, or any immunoassay. It is acceptable that AHR deletion is affecting microglia and inducing cell death. But this by no means shows that microglial inflammation is inducing this neuronal cell death. Thrombin by itself can result in the production of high molecular peptides that are capable of inducing cell death.

Response to comments: Yes, it is. We should reduce the intonation in these descriptions “The inflammatory response accelerated in microglia by thrombin or combined effects of activated microglia or thrombin itself may be attributed to the neuronal death.”    

Round 2

Reviewer 2 Report

The authors appropriately responded to the reviewer’ comments and revised the manuscript. The revised manuscript has been improved

Author Response

Thank you for your comments and we learn a lot.

Reviewer 3 Report

The authors did not improved the description of docking results. Please follow the suggestions from 1st round of review.

The description of docking results is poor

Author Response

Response to comments: Thank you for your comments, which are our weakness in this study. In our upcoming study using the docking method, we should add the data in the physical interaction. We added the description of docking simulations with ZDOCK in methodology section recommend by the reviewer. First, in the protein structure preparation, the three-dimensional structure of AHR (UniProt ID: P35869) and thrombin (UniProt ID: P25116) in the Protein Data Bank (PDB) format were retrieved and prepared for docking. This procedure got involved in checking for any missing residues, adding hydrogen atoms, and addressing any structural irregularities or inconsistencies. Second, In ZDOCK input preparation, the retrieved structures of AHR and thrombin, including the specific ligand position at Ser-36 in AHR, were used as input for ZDOCK. The PDB files of the two proteins information were transferred to the software. Third, in receptor-ligand definition, the ligand position at Ser-36 in AHR was defined as the binding site for thrombin. Fourth, in grid generation, ZDOCK employed a grid-based approach to facilitate the initial search process. A three-dimensional grid was generated around the receptor protein of AHR, focusing on the ligand position at Ser-36. The grid size and resolution parameters were determined based on the characteristics and proximity of the ligand site to the binding partner. Fifth, in Fast Fourier Transform (FFT) Search, ZDOCK utilized a fast Fourier transform algorithm to efficiently explore the six-dimensionally translational and rotational space of AHR and thrombin, with specific emphasis on the ligand position at Ser-36 in AHR. Sixth, in Monte Carlo refinement, ZDOCK performed a refinement step using a Monte Carlo algorithm. This step aimed to improve the accuracy of the predicted docking models by adjusting the protein-protein interface and exploring potential conformational changes while maintaining the specific ligand position at Ser-36 in AHR. Seventh, in Scoring Function- ZRANK: ZDOCK employed a scoring function called ZRANK to evaluate and rank the predicted docking models. ZRANK incorporated various energy terms, such as electrostatic interactions, van der Waals forces, and desolvation energy, to assess the quality of each model. This model was ranked based on their ZRANK scores, with lower scores indicating a more favorable binding model. Finally, in Output Analysis, ZDOCK generated an output file containing a ranked list in the predicted docking models between AHR and thrombin. Each model was assigned a ZRANK score, allowing for the identification of the most favorable docking configurations that accommodated the ligand position at Ser-36 in AHR, which was used in this experiment.

Reviewer 4 Report

The authors have tried to address most comments. 

Still not convinced with:

1. Figure 5 that the MMP9 activity increase is from the microglia. It colocalizes-yes, there are differences in IBA1 microglia-yes but there are no visible differences in the MMP9-positive microglia in both groups. There should be a quantification of MMP9 in microglia to conclude that, " The increased 223 MMP 9 activity was found in microglia subjected to thrombin induction in wild type but 224 further escalated in AhR deletion."

2. I think as mentioned in the first review report Figures 2 and 4 need representative images. It is not appropriate to support in vitro microglial culture results with findings from the in vivo study.

Once again, in Figure 6. they say they have shown Iba1, But I don't see it in the figure. 

They could improve their writing. A. The language keeps changing from active to passive randomly, with many typos and  B. There is still a lack of coherence between results.

They could improve their writing. A. The language keeps changing from active to passive randomly, with many typos and  B. There is still a lack of coherence between results.

Author Response

1. Figure 5 that the MMP9 activity increase is from the microglia. It colocalizes-yes, there are differences in IBA1 microglia-yes but there are no visible differences in the MMP9-positive microglia in both groups. There should be a quantification of MMP9 in microglia to conclude that, " The increased 223 MMP 9 activity was found in microglia subjected to thrombin induction in wild type but 224 further escalated in AhR deletion."

Response to comments: Thank you for your comments. I revised Figure 5 by adding the calculation of MMP 9 positive cells presented as number/surface area(mm²) shown in supplementary data as supplementary Figure 1.   

2. I think as mentioned in the first review report Figures 2 and 4 need representative images. It is not appropriate to support in vitro microglial culture results with findings from the in vivo study. Once again, in Figure 6. they say they have shown Iba1, But I don't see it in the figure.

Response to comments: Thank you for your comments. We revised Figure 2 by adding the light photography of primary culture of microglia in wild and AhRKO mice subjected to the simulation by PBS and thrombin and shown in Figure 2A.