SRF-deficient astrocytes provide neuroprotection in mouse models of excitotoxicity and neurodegeneration
This article has been Reviewed by the following groups
Listed in
- Evaluated articles (Review Commons)
Abstract
Reactive astrogliosis is a common pathological hallmark of CNS injury, infection, and neurodegeneration, where reactive astrocytes can be protective or detrimental to normal brain functions. Currently, the mechanisms regulating neuroprotective astrocytes and the extent of neuroprotection are poorly understood. Here, we report that conditional deletion of serum response factor (SRF) in adult astrocytes causes reactive-like hypertrophic astrocytes throughout the mouse brain. These Srf GFAP-ER CKO astrocytes do not affect neuron survival, synapse numbers, synaptic plasticity or learning and memory. However, the brains of Srf knockout mice exhibited neuroprotection against kainic-acid induced excitotoxic cell death. Relevant to human neurodegenerative diseases, Srf GFAP-ER CKO astrocytes abrogate nigral dopaminergic neuron death and reduce β-amyloid plaques in mouse models of Parkinson’s and Alzheimer’s disease, respectively. Taken together, these findings establish SRF as a key molecular switch for the generation of reactive astrocytes with neuroprotective functions that attenuate neuronal injury in the setting of neurodegenerative diseases.
Article activity feed
-
-
-
-
-
-
Note: This rebuttal was posted by the corresponding author to Review Commons. Content has not been altered except for formatting.
Learn more at Review Commons
Reply to the reviewers
1. General Statements
We thank the editors for sending our manuscript for peer review and the reviewers for careful reading and their critical comments to improve the manuscript. Below, we describe the experiments that have been carried out in response to the reviewers and incorporated in the preliminary revision. We also describe our plan for the revisions that will address the remaining comments of the reviewers. Most of the comments are addressable with additional experiments (some of which are already ongoing) and these experiments will surely strengthen the study reported in this manuscript without affecting the fundamental findings. We would require …
Note: This rebuttal was posted by the corresponding author to Review Commons. Content has not been altered except for formatting.
Learn more at Review Commons
Reply to the reviewers
1. General Statements
We thank the editors for sending our manuscript for peer review and the reviewers for careful reading and their critical comments to improve the manuscript. Below, we describe the experiments that have been carried out in response to the reviewers and incorporated in the preliminary revision. We also describe our plan for the revisions that will address the remaining comments of the reviewers. Most of the comments are addressable with additional experiments (some of which are already ongoing) and these experiments will surely strengthen the study reported in this manuscript without affecting the fundamental findings. We would require up to 4-6 weeks to complete these experiments.
2. Description of the planned revisions
Insert here a point-by-point reply that explains what revisions, additional experimentations and analyses are planned to address the points raised by the referees.
Reviewer #1 (Evidence, reproducibility and clarity (Required)):
Summary: The authors used a conditional transgenic mouse model to demonstrate that deletion of serum response factor (SRF) from adult astrocytes provides neuroprotection in various insult/ diseases contexts without promoting any obvious phenotypic deficiencies. The work builds on the group’s previous study where SRF was embryonically deleted from astrocytes and their precursor cells. Given the role of SRF in promoting glial cell differentiation, the adult conditional KO used in the current study was designed to circumvent the limitations of the previous approach. The authors used a variety of complementary approaches (including immunohistochemistry, electrophysiology, transcriptomics, and behavior) to demonstrate the therapeutic potential of their approach. However, I have questions regarding the validity of the behavioral analyses as well as some of the imaging results that dampen my overall enthusiasm.
Major Comment #1
The synaptogenic factors probed in Figure 3C (e.g. glypicans, thrombospondins, etc.) are not likely to play major roles in the adult brain in a non-injury context, so I do not know that these analyses provide any significant insight into potential functional changes in the mutant mice. Along the same lines, the analysis of synapse count (Figure 3D-E) seems inconsequential given that SRF was knocked out well after the period of developmental synaptogenesis. It would have been much more interesting to have performed these analyses following insult (such as the kainate injury model used by the authors) or in one of the disease models presented later in the manuscript. As it stands, I don't think they add very much to the study.
Response: We are grateful to the reviewer for the careful reading of the manuscript. Astrocytes are known to regulate the formation, maintenance, and elimination of synapses. It has been previously shown that LPS-induced reactive astrocytes exhibit reduced expression of several synaptogenic factors, were unable to promote synapse formation and showed reduced phagocytic activity (PMID: 28099414). We wanted to determine whether the SRF-deficient reactive-like astrocytes were likely compromised in their ability to produce pro-synaptogenic factors and/or adversely affect synapse maintenance. We agree with the reviewer that analysis of synapses in the adult brain may not address the role of these mutant astrocytes in synaptogenesis. But our results indicate that the mutant astrocytes are likely not affecting synapse maintenance or exhibit altered phagocytotic activity that would result in increased or decreased synapse numbers. We will make this clearer in the revised manuscript.
Minor Comment #2:
The authors should note that the use of GluA1 as a postsynaptic marker will not identify silent synapses (i.e. structurally "normal" but functionally inert).
Response: We agree with the reviewer that GluA1 will not identify silent synapses. To study silent vs functional synapses, we will stain for Piccolo (presynaptic) and NMDA receptor NR1 subunit (post-synaptic) to label all synapses and compare this with Piccolo/GluA1 co-localized synapses to identify the functional synapses.
Reviewer #2 (Significance (Required):
The manuscript addresses the important area of the cellular mechanisms that underlie neuroprotection. The ms adds to our understanding of genetic control of neuroprotection and should be of significant interest to others in the field. The experimental approach systematic and the data presented are generally of high quality and believable. While the ms presents quite a bit of overall cellular data that underlies various areas of neuronal and brain function that may be affected by loss of SRF, it is still somewhat descriptive. It is unclear what aspect of astrocyte reactivity is determinative, how mechanistically in normal cells SRF suppresses reactivity, and how SRF -negative reactive astrocytes confer such broad neuroprotection. While the latter is well beyond the scope of this study, the authors do propose SRF may be involved in regulating oxidative stress and amyloid plaque clearance as a potential pathway to account for SRF's role, however a more systematic discussion based on the gene expression data and known pathways would be welcome. Overall, this is a high quality ms that should be of interest to the field that identifies a SRF as a novel player in neuroprotection.
Response: We thank the reviewer for the careful reading of the manuscript and for the positive comments. We will include a more detailed discussion on the genes and pathways based on our gene expression data that may provide insights into how SRF may regulate astrocyte reactivity and neuroprotection.
Additional considerations:
Quantification of the extent of SRF loss in astrocytes in conditional tamoxifen knockout would strengthen the quality of the data.
Response: We will provide this data in the revised manuscript.
While the authos did use a Sholl analysis to show hypertophic changes in SRF negative astrocytes, given that SRF is an important regulator of actin and other cytoskeletal related proteins in other cell types, and that cytoskeletal components can play an important role in cell signaling, it is somewhat surprising that the gene array analysis did not include actin and other cytoskeletal proteins, nor did the authors consider a more careful analysis of intracellular cytoskeletal changes and the potential mechanistic implications of this for observed reactivity and neuroprotection.
Response: We agree with the reviewer that SRF is a well-established regulator of actin cytoskeleton. However, we did not any significant changes in gene expression for actin or actin-regulatory proteins. We would have expected a decrease in astrocyte morphology similar to the neurite/axon defects exhibited by SRF-deficient neurons. It is unclear whether the hypertrophic morphology is due to transcriptional regulation of actin/actin-binding proteins or due to astrocyte reactivity. This would be a very interesting question and we will investigate these aspects in future studies.
Reviewer #3 (Evidence, reproducibility and clarity (Required):
Summary: The study by Thumu et al., suggests that astrocytic specific deletion of SRF in mice results in morphological changes in these cells that does not affect neuronal survival, synapse number, plasticity or cognition. However, in in vivo mouse models of excitotoxic damage and neurodegenerative disease, deletion of SRF reduced neurotoxicity. The authors provide sufficient evidence to suggest that astrocytic SRF contributes to neurotoxicity in various models however some claims are made that are currently not supported by evidence.
Major comments:
- The authors claim that SRF KO astrocytes undergo hypertrophy. However, the quantification of the number of intersections gives information about morphology rather than hypertrophy. Quantification of cell size (area of S100B staining) should be provided.
Response: We will provide the data suggested by the reviewer.
- For the RNAseq of isolated astrocytes did the authors confirm that other cell types (e.g microglia) did not contaminate their samples?
Response: We will provide the information requested by the reviewer.
Reviewer #3:
Minor comments:
- The authors say that in Figure 1B many astrocytes did not show any SRF expression. However, overall averages of SRF intensity are plotted in Figure 1C. It would support their claim to instead to calculate the percentage of SRF expressing cells above a certain threshold in each condition, rather than plotting the mean intensity. As a control for their method of quantifying SRF intensity in Figure 1B, demonstrating no change in SRF in neurons would provide confidence for the specificity of the knockout.
Response: We will provide the quantification of the extent of SRF loss in astrocytes (percent astrocytes that are deleted for SRF) as suggested by Reviewer 2. We will also provide SRF intensity from neurons as suggested by the reviewer.
- The authors use the term "reactivation" throughout the manuscript. This could be misconstrued as re-activation and so I would suggest using the terms "reactivity" or "reactive transformation". Furthermore, only one region is quantified in Figure 1C while in later figures multiple regions are quantified. The authors should justify this decision or update the figures with data from missing regions.
Response: We will make this change in using the term “reactivity” as suggested by the reviewer.
- In Figure S2 the authors should provide a positive control for their staining.
Response: We will provide the positive control data for this experiment.
- Can the authors explain the large amount of variability in number of synapses in 15 mpi in Figure 3E?
Response: We will perform more immunostainings and update the data presented in this figure.
- Images in Figure 2C are poorly visible and should be improved in terms of either quality or magnification.
Response: We will provide better quality image for Figure 2C.
- The authors should provide a list of differentially expressed genes from RNAseq of SRF KO mice. No information is currently given in the text about the number of differentially expressed genes in the conditional knockout.
Response: We will include this information in the revised manuscript.
- In figure 5A data would be better illustrated as a volcano plot (similar to Fig. S7C).
Response: We will provide this in the revised manuscript.
3. Description of the revisions that have already been incorporated in the transferred manuscript
Please insert a point-by-point reply describing the revisions that were already carried out and included in the transferred manuscript. If no revisions have been carried out yet, please leave this section empty.
Reviewer #1: Major Comment #2
There is considerable variability in the behavioral results, particularly the fear conditioning and Barnes maze tasks (Figures 4F-G). Given the extremely low sample size for mouse behavior (n=5 in on group, n=7 in the other), it is highly likely that the behavioral tests were done with a single cohort of animals (which would be far from ideal) and that these experiments are significantly underpowered. Furthermore, it does not appear that the fear conditioning task was properly optimized. For example, in the control mice in context A, there were two animals that were at or very close to 0 percent freezing; these were likely outliers, or even an indication that the foot shock conditioning protocol was not working as it should. The highest percent freezing of either group was ~70%, which would have been an ideal starting place as an average for the control group. In addition, sex of the animals was not reported for these experiments. If the authors combined sexes as they did in other analyses in this paper, it is possible that they missed reaching the appropriate reaction threshold for the foot shock for some of the animals, as sex differences have previously been demonstrated in mice (DOI: 10.1037/bne0000248). Given the age at which the animals are assessed with these tasks, these specific revisions would require greater than 6 months to complete. However, as currently presented, there simply are not enough data points to make any conclusions regarding behavior.
Response: We have performed the behavioural experiments with an additional cohort of animals for both control and mutant groups and reanalysed the data. We now have n=11 for control and n=9 for mutant group. Only males were used for the behaviour experiments, and we do not see any significant difference in behaviour between the two groups. These results are included in revised Figure 4E-G in the Preliminary Revision of the manuscript. However, we are waiting to perform the remote recall memory for the fear conditioning experiment and will include this date in the revised manuscript.
Minor Comment #1:
The representative GFAP images (Figure 1 E/G) do not appear to have been taken at the same magnification. This was particularly apparent in the comparison between the control and CKO hippocampus at 12mpi. It is difficult to say with certainty, due to the lack of fiducial markers in many of the images. Inclusion of a nuclear stain (DAPI) would be highly beneficial to allow the reader to make a more informed comparison.
Response: These images were taken at the same magnification. We have included the DAPI staining for these images in Suppl. Figure 2 in the Preliminary Revision of the manuscript.
**Referees cross-commenting**
After reading the comments of the other reviewer, I think we're in agreement that the cellular and molecular data, while descriptive, is of mostly excellent quality. Moreover, the significance of the study is high, and the potential readership broad. However, I stand by my initial assessment of the behavioral data and find the manuscript quite lacking in this regard. Proper revisions would take at least half a year or more, so the authors may be disinclined to go this route. That being said, if the behavioral data were to be excised, I would be happy to sign off on the rest of the manuscript provided that the other major criticisms are addressed.
Response: We thank the reviewer for the appreciation of our work. We have increased the number of animals in the behavioural experiments and do not see any significant difference between the two groups. These results are included in revised Figure 4E-G in the Preliminary Revision of the manuscript.
In response cross-comment of Rev 2:
Agreed that if properly conducted and presented, the behavioral data would indeed provide a nice functional correlate to the cellular work. In its current state, I'm afraid that it is instead a hindrance to the study and I would recommend that they just remove it if they choose not to address my concerns with the quality (particularly the extreme variability and the complete lack of freezing by several of the animals, especially in the controls).
Response: We hope that the revised behaviour data would provide a strong functional correlate to the other findings in the study.
Additional cross-comments:
I agree with the added criticisms raised by Reviewer #3, and I think that the manuscript would be greatly improved by revisions that address those and the original criticisms from myself and Reviewer #2. I still think that the behavioral data should be omitted, provided that the authors are not capable or willing to appropriately address those concerns within a reasonable time frame.
Response: We will address all the concerns raised by the reviewers with the required experiments to further strengthen the findings in this study.
Reviewer #3
Major Comment
- In Figure S1 the authors provide evidence showing lack of B-gal in cell types other than astrocytes (neurons/OPCs). However, microglia are missing, which could be important as later they show that microglia undergo changes in the SRF knockout model. This staining should be provided.
Response: We have performed double immunostaining for b-gal and IbaI and do not see any overlap between IbaI and b-gal, suggesting that there is no Cre expression in microglia. We have included this data in revised Figure S1F in the Preliminary Revision of the manuscript.
- The authors claim in the text that microglia have thicker processes and an amoeboid shape however no evidence of this is provided in Figure S5.
Response: We have provided data to show larger microglia area and morphology in revised Figure S5 in the Preliminary Revision of the manuscript.
- In the text "Enrichment analysis of Gene Ontology terms for Biological Process (GO BP) revealed that Srf deficient astrocytes showed enrichment of pathways related to cellular response to beta amyloid and beta-amyloid clearance." This is not shown in fig 5. It would be more accurate to say that there is a downregulation of genes involved in B amyloid metabolic process.
Response: We apologize for the omission in showing that this data was presented in Suppl. Fig. S8E. We have now indicated this in the main text.
Minor Comments:
- Figure 1E is missing body weight data noted in the figure legend.
Response: We apologize for this oversight. This data was actually included in Suppl. Figure S3E and not in Figure 1. We have made the appropriate correction to Figure legend 1.
- In Figure 2B figure labels are missing.
Response: We thank the reviewer for pointing out this omission. We have added the missing labels.
- Details of houskeeping gene normalisation are missing from qPCR data.
Response: We apologize for not providing this information. We have included this in the revised Methods section.
4. Description of analyses that authors prefer not to carry out
Please include a point-by-point response explaining why some of the requested data or additional analyses might not be necessary or cannot be provided within the scope of a revision. This can be due to time or resource limitations or in case of disagreement about the necessity of such additional data given the scope of the study. Please leave empty if not applicable.
Reviewer #3, Major Comment 1:
- The title of the manuscript is "SRF-deficient astrocytes provide neuroprotection in mouse models of excitotoxicity and neurodegeneration". It would be more accurate to say that SRF is involved in neurotoxicity in these models. To make a comment on the role of SRF in neuroprotection, experiments should be performed in spinal cord injury or ischaemia, where deficiency of SRF would be hypothesised to worsen recovery.
Response: We disagree with the reviewer with this assessment. There is no evidence to suggest that SRF is involved in neurotoxicity. What our data suggests is that SRF deficiency results in a reactive astrocyte state that is neuroprotective in these models. We hypothesize that in injury/infection/disease conditions that would result in generation of neuroprotective astrocytes, SRF expression or function may be negatively regulated. It would be interesting to see whether the SRF-deficient astrocytes alleviate or exacerbate pathology and recovery following spinal cord injury and ischaemia.
-
Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.
Learn more at Review Commons
Referee #3
Evidence, reproducibility and clarity
Summary: The study by Thumu et al., suggests that astrocytic specific deletion of SRF in mice results in morphological changes in these cells that does not affect neuronal survival, synapse number, plasticity or cognition. However, in in vivo mouse models of excitotoxic damage and neurodegenerative disease, deletion of SRF reduced neurotoxicity. The authors provide sufficient evidence to suggest that astrocytic SRF contributes to neurotoxicity in various models however some claims are made that are currently not supported by evidence.
Major comments: The title of the manuscript is "SRF-deficient astrocytes provide neuroprotection in …
Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.
Learn more at Review Commons
Referee #3
Evidence, reproducibility and clarity
Summary: The study by Thumu et al., suggests that astrocytic specific deletion of SRF in mice results in morphological changes in these cells that does not affect neuronal survival, synapse number, plasticity or cognition. However, in in vivo mouse models of excitotoxic damage and neurodegenerative disease, deletion of SRF reduced neurotoxicity. The authors provide sufficient evidence to suggest that astrocytic SRF contributes to neurotoxicity in various models however some claims are made that are currently not supported by evidence.
Major comments: The title of the manuscript is "SRF-deficient astrocytes provide neuroprotection in mouse models of excitotoxicity and neurodegeneration". It would be more accurate to say that SRF is involved in neurotoxicity in these models. To make a comment on the role of SRF in neuroprotection, experiments should be performed in spinal cord injury or ischaemia, where deficiency of SRF would be hypothesised to worsen recovery.
The authors claim that SRF KO astrocytes undergo hypertrophy. However, the quantification of the number of intersections gives information about morphology rather than hypertrophy. Quantification of cell size (area of S100B staining) should be provided.
In Figure S1 the authors provide evidence showing lack of B-gal in cell types other than astrocytes (neurons/OPCs). However, microglia are missing, which could be important as later they show that microglia undergo changes in the SRF knockout model. This staining should be provided.
Can the authors explain the large amount of variability in number of synapses in 15 mpi in Figure 3E?
The authors claim in the text that microglia have thicker processes and an amoeboid shape however no evidence of this is provided in Figure S5.
For the RNAseq of isolated astrocytes did the authors confirm that other cell types (e.g microglia) did not contaminate their samples?
In the text "Enrichment analysis of Gene Ontology terms for Biological Process (GO BP) revealed that Srf deficient astrocytes showed enrichment of pathways related to cellular response to betaamyloid and beta-amyloid clearance." This is not shown in fig 5. It would be more accurate to say that there is a downregulation of genes involved in B-amyloid metabolic process.
OPTIONAL: Figure 6 would be greatly strengthened by functional in vivo experiments showing reversal of motor/ cognitive phenotypes.
OPTIONAL: The study would be improved by isolating astrocytes from the models used in figure 6 and performing RNAseq to provide information about how SRF knockout affects astrocyte reactivity in these models.
Minor comments: The authors say that in Figure 1B many astrocytes did not show any SRF expression. However, overall averages of SRF intensity are plotted in Figure 1C. It would support their claim to instead to calculate the percentage of SRF expressing cells above a certain threshold in each condition, rather than plotting the mean intensity. As a control for their method of quantifying SRF intensity in Figure 1B, demonstrating no change in SRF in neurons would provide confidence for the specificity of the knockout.
The authors use the term "reactivation" throughout the manuscript. This could be misconstrued as re-activation and so I would suggest using the terms "reactivity" or "reactive transformation".
Furthermore, only one region is quantified in Figure 1C while in later figures multiple regions are quantified. The authors should justify this decision or update the figures with data from missing regions.
In Figure S2 the authors should provide a positive control for their staining.
Figure 1E is missing body weight data noted in the figure legend.
Images in Figure 2C are poorly visible and should be improved in terms of either quality or magnification.
In Figure 2B figure labels are missing.
Details of houskeeping gene normalisation are missing from qPCR data.
The authors should provide a list of differentially expressed genes from RNAseq of SRF KO mice. No information is currently given in the text about the number of differentially expressed genes in the conditional knockout. In figure 5A data would be better illustrated as a volcano plot (similar to Fig. S7C).
Significance
The strength of the manuscript is that the authors demonstrate in more than one model that astrocyte specific knockout of SRF rescues neuronal death, implicating SRF in astrocyte mediated neurotoxicity. The limitations of the study are that the mechanism by which SRF deletion reduces excitotoxicity is not addressed and there is no supporting data beyond neuronal survival in the excitotoxicity/OHDA models or plaque density in the APP/PS1 model.
This study adds SRF to an expanding understanding of the neurotoxic capacity of astrocytes in certain reactive states. It will be of broad interest to the astrocyte reactivity field.
My field of expertise is in astrocyte and microglia interactions in neurodegenerative diseases.
-
Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.
Learn more at Review Commons
Referee #2
Evidence, reproducibility and clarity
The manuscript, "SRF-deficient astrocytes provide neuroprotection in mouse models of excitotoxicity and neurodegeneration" by Thumu et al., describes the observation astrocyte specfic SRF deficient mice exhibit neuroprotection against a broad range of brain pathologies. The current ms follows up on previous work done by the corresponding author Ramanan and colleagues in which they showed that astrocyte-specific deletion of SRF early during mouse development resulted in persistent reactive-like astrocytes throughout the postnatal mouse brain. In the current ms the authors present data that adult astrocyte specific conditional …
Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.
Learn more at Review Commons
Referee #2
Evidence, reproducibility and clarity
The manuscript, "SRF-deficient astrocytes provide neuroprotection in mouse models of excitotoxicity and neurodegeneration" by Thumu et al., describes the observation astrocyte specfic SRF deficient mice exhibit neuroprotection against a broad range of brain pathologies. The current ms follows up on previous work done by the corresponding author Ramanan and colleagues in which they showed that astrocyte-specific deletion of SRF early during mouse development resulted in persistent reactive-like astrocytes throughout the postnatal mouse brain. In the current ms the authors present data that adult astrocyte specific conditional deletion of serum response factor results in reactive-like hypertrophic astrocytes that localize throughout the mouse brain. They further show that SRF deficient astrocytes do not affect neuron survival, synapse numbers, synaptic plasticity or learning and memory. Strikingly, they further show that brains of Srf knockout mice exhibit protection against neurodegenerative disease related pathologies including induced excitotoxic cell death and that SRF-deficient astrocytes abrogate dopaminergic neuron death and reduce beta-amyloid plaques in mouse models of Parkinson's and Alzheimer's disease. Based on their results, the authors proposes that SRF is a key molecular switch for the generation of reactive astrocytes with neuroprotective functions can attenuate neuronal injury in the setting of neurodegenerative diseases.
Referees cross-commenting
Reviewer #1 raises an important concern regarding the quality of the behavioral studies. I would also agree that the ms is still strong and the findings are significant without them, although they do extend the functional dimensions of the overall study.
Significance
The manuscript addresses the important area of the cellular mechanisms that underlie neuroprotection. The ms adds to our understanding of genetic control of neuroprotection and should be of significant interest to others in the field. The experimental approach systematic and the data presented are generally of high quality and believable. While the ms presents quite a bit of overall cellular data that underlies various areas of neuronal and brain function that may be affected by loss of SRF, it is still somewhat descriptive. It is unclear what aspect of astrocyte reactivity is determinative, how mechanistically in normal cells SRF suppresses reactivity, and how SRF -negative reactive astrocytes confer such broad neuroprotection. While the latter is well beyond the scope of this study, the authors do propose SRF may be involved in regulating oxidative stress and amyloid plaque clearance as a potential pathway to account for SRF's role, however a more systematic discussion based on the gene expression data and known pathways would be welcome. Overall, this is a high quality ms that should be of interest to the field that identifies a SRF as a novel player in neuroprotection.
Additional considerations:
- Quantification of the extent of SRF loss in astrocytes in conditional tamoxifen knockout would strengthen the quality of the data.
- While the authos did use a Sholl analysis to show hypertophic changes in SRF negative astrocytes, given that SRF is an important regulator of actin and other cytoskeletal related proteins in other cell types, and that cytoskeletal components can play an important role in cell signaling, it is somewhat surprising that the gene array analysis did not include actin and other cytoskeletal proteins, nor did the authors consider a more careful analysis of intracellular cytoskeletal changes and the potential mechanistic implications of this for observed reactivity and neuroprotection.
-
Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.
Learn more at Review Commons
Referee #1
Evidence, reproducibility and clarity
Summary: The authors used a conditional transgenic mouse model to demonstrate that deletion of serum response factor (SRF) from adult astrocytes provides neuroprotection in various insult/diseases contexts without promoting any obvious phenotypic deficiencies. The work builds on the group's previous study where SRF was embryonically deleted from astrocytes and their precursor cells. Given the role of SRF in promoting glial cell differentiation, the adult conditional KO used in the current study was designed to circumvent the limitations of the previous approach. The authors used a variety of complementary approaches (including …
Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.
Learn more at Review Commons
Referee #1
Evidence, reproducibility and clarity
Summary: The authors used a conditional transgenic mouse model to demonstrate that deletion of serum response factor (SRF) from adult astrocytes provides neuroprotection in various insult/diseases contexts without promoting any obvious phenotypic deficiencies. The work builds on the group's previous study where SRF was embryonically deleted from astrocytes and their precursor cells. Given the role of SRF in promoting glial cell differentiation, the adult conditional KO used in the current study was designed to circumvent the limitations of the previous approach. The authors used a variety of complementary approaches (including immunohistochemistry, electrophysiology, transcriptomics, and behavior) to demonstrate the therapeutic potential of their approach. However, I have questions regarding the validity of the behavioral analyses as well as some of the imaging results that dampen my overall enthusiasm.
Major Comments:
- The synaptogenic factors probed in Figure 3C (e.g. glypicans, thrombospondins, etc.) are not likely to play major roles in the adult brain in a non-injury context, so I do not know that these analyses provide any significant insight into potential functional changes in the mutant mice. Along the same lines, the analysis of synapse count (Figure 3D-E) seems inconsequential given that SRF was knocked out well after the period of developmental synaptogenesis. It would have been much more interesting to have performed these analyses following insult (such as the kainate injury model used by the authors) or in one of the disease models presented later in the manuscript. As it stands, I don't think they add very much to the study.
- There is considerable variability in the behavioral results, particularly the fear conditioning and Barnes maze tasks (Figures 4F-G). Given the extremely low sample size for mouse behavior (n=5 in on group, n=7 in the other), it is highly likely that the behavioral tests were done with a single cohort of animals (which would be far from ideal) and that these experiments are significantly underpowered. Furthermore, it does not appear that the fear conditioning task was properly optimized. For example, in the control mice in context A, there were two animals that were at or very close to 0 percent freezing; these were likely outliers, or even an indication that the foot shock conditioning protocol was not working as it should. The highest percent freezing of either group was ~70%, which would have been an ideal starting place as an average for the control group. In addition, sex of the animals was not reported for these experiments. If the authors combined sexes as they did in other analyses in this paper, it is possible that they missed reaching the appropriate reaction threshold for the foot shock for some of the animals, as sex differences have previously been demonstrated in mice (DOI: 10.1037/bne0000248). Given the age at which the animals are assessed with these tasks, these specific revisions would require greater than 6 months to complete. However, as currently presented, there simply are not enough data points to make any conclusions regarding behavior.
Minor Comments:
- The representative GFAP images (Figure 1 E/G) do not appear to have been taken at the same magnification. This was particularly apparent in the comparison between the control and CKO hippocampus at 12mpi. It is difficult to say with certainty, due to the lack of fiducial markers in many of the images. Inclusion of a nuclear stain (DAPI) would be highly beneficial to allow the reader to make a more informed comparison.
- The authors should note that the use of GluA1 as a postsynaptic marker will not identify silent synapses (i.e. structurally "normal" but functionally inert).
Referees cross-commenting
After reading the comments of the other reviewer, I think we're in agreement that the cellular and molecular data, while descriptive, is of mostly excellent quality. Moreover, the significance of the study is high, and the potential readership broad. However, I stand by my initial assessment of the behavioral data and find the manuscript quite lacking in this regard. Proper revisions would take at least half a year or more, so the authors may be disinclined to go this route. That being said, if the behavioral data were to be excised, I would be happy to sign off on the rest of the manuscript provided that the other major criticisms are addressed.
In response cross-comment of Rev 2:
Agreed that if properly conducted and presented, the behavioral data would indeed provide a nice functional correlate to the cellular work. In its current state, I'm afraid that it is instead a hindrance to the study and I would recommend that they just remove it if they choose not to address my concerns with the quality (particularly the extreme variability and the complete lack of freezing by several of the animals, especially in the controls).
Additional cross-comments:
I agree with the added criticisms raised by Reviewer #3, and I think that the manuscript would be greatly improved by revisions that address those and the original criticisms from myself and Reviewer #2. I still think that the behavioral data should be omitted, provided that the authors are not capable or willing to appropriately address those concerns within a reasonable time frame.
Significance
General assessment: Overall strengths of this study are the implications of SRF as a broad spectrum anti-neurodegeneration agent and the variety of techniques used. Limitations of this study include a lack of meaningful synaptic comparisons and underpowered behavioral assays.
Advance: Provided the above limitations are addressed, this study would provide a meaningful advance in our understanding of controlled reactive astrogliosis as a potential therapeutic strategy for neuroprotection.
Audience: This study would be of interest to a wide audience, particularly neuro- and gliobiologists as well as clinicians who deal with brain disorders and injury.
Expertise: imaging; behavior; synaptic development
-
