HSP25 and HSP25-P-Ser15 Prompt Innate Neuroprotection in Lobe X of the Cerebellum

Round 1

Reviewer 1 Report (New Reviewer)

Comments and Suggestions for Authors

The study by Hernandez-Perez et al investigates mechanisms and pathways of Purkinje cell degeneration and survival. The authors achieve this through the histological evaluation of heat shock protein HSP25 and other markers of cell death in Purkinje cells within the Purkinje cell degeneration (PCD) mouse model. The study provides useful and novel information regarding the cellular pathways involved in Purkinje cell survival and resistance to degenerative processes that may be utilised in future therapeutic strategies for cerebellar disease. This study is interesting and certainly worthy of publication, however I have several minor corrections that I feel should be addressed:

• It would be useful to include lower magnified images encompassing the entire sagittal cerebellar slice in Figures 1, 2 and 4, to demonstrate the specificity/localisation of HSP25 and HSP25-P-Ser15 expression to lobule X.
• An explanation should be provided within the figure legends to clarify what the top blue/red lines represent within the graphs (for example, in Figure 3).
• n numbers should be added to all figure legends.
• In addition to analysing Purkinje cell density (through cell counts/length) in lobule X over time (Figure 3a), it would be useful to also assess dendritic morphology/damage through analysing the thickness of the molecular layer and/or dendritic structure. This would help provide further evidence that the Purkinje cell is resistant to damage in lobule X until p25, rather than relying solely on the presence of the Purkinje cell soma.
• The text within the discussion feels proportionally overly long and would benefit from shortening.
• Product codes for the antibodies described within the study should be provided within the methods section.

Author Response

Response to REVIEWER #1

We sincerely thank the Reviewer for the thorough evaluation of our work and for the constructive feedback provided. We truly appreciate the recognition of the study’s strengths, as well as the proposed corrections. These observations have been very helpful in refining the manuscript, and we have addressed each point in detail below. Changes related to these suggestions appear in orange in the new version of the manuscript.

The study by Hernandez-Perez et al investigates mechanisms and pathways of Purkinje cell degeneration and survival. The authors achieve this through the histological evaluation of heat shock protein HSP25 and other markers of cell death in Purkinje cells within the Purkinje cell degeneration (PCD) mouse model. The study provides useful and novel information regarding the cellular pathways involved in Purkinje cell survival and resistance to degenerative processes that may be utilised in future therapeutic strategies for cerebellar disease. This study is interesting and certainly worthy of publication, however I have several minor corrections that I feel should be addressed:

1. It would be useful to include lower magnified images encompassing the entire sagittal cerebellar slice in Figures 1, 2 and 4, to demonstrate the specificity/localisation of HSP25 and HSP25-P-Ser15 expression to lobule X.

We appreciate this suggestion, and now, additional lower magnified images as well as a table gathering measurements of different lobes for different markers from P25 onwards (critical ages for our study) are included in the new version of the manuscript. Data correspond to HSP25-P-Ser15 (the active form of HSP25) and PKC-δ (the activator kinase) to ease the reader’s comprehension. In addition, data of these markers have also requested for other Reviewer. All of them constitute a new supplementary figure (Supp. Fig. 3).

2. An explanation should be provided within the figure legends to clarify what the top blue/red lines represent within the graphs (for example, in Figure 3).

Differences between genotypes (i.e. wild type or PCD) are signaled by asterisks at each time point. We have also compared the values of time points within each genotype. To avoid blurring graphs with many different symbols, especially when several time points may have statistically similar values, we decided to use the aforementioned horizontal lines. Note that those lines are drawn at different heights, thus corresponding to higher or lower values of an analyzed variable depending on the time point. Those time points with statistically similar values have a line at the same height. Finally, blue lines represent values of wild-type animals and red lines represent values of PCD mice. For example, in figure 1A, values of Purkinje cell density in wild-type mice (blue lines) at P15 are statistically higher than those of P20 onwards, which are statistically similar amongst them; besides, values of PCD mice (red lines) are statistically similar from P15 to P30, but at P35 are lower. A more detailed explanation is now included in the new version of the manuscript.

3. n numbers should be added to all figure legends.

Suggestion done.

4. In addition to analysing Purkinje cell density (through cell counts/length) in lobule X over time (Figure 3a), it would be useful to also assess dendritic morphology/damage through analysing the thickness of the molecular layer and/or dendritic structure. This would help provide further evidence that the Purkinje cell is resistant to damage in lobule X until p25, rather than relying solely on the presence of the Purkinje cell soma.

This is a very interesting suggestion that would provide a supplementary evidence of the HSP25 effect. However, these data would be somehow less relevant that the number of Purkinje cells itself. Indeed, both variables are closely related, as our group has previously demonstrated (Pérez-Martín et al., 2021, Neurotherapeutics). In this sense, although a neuroprotective molecule may temporary preserve the dendritic arborization of Purkinje cells, when this structure finally collapses, the neurons, in short, end to die. Moreover, the analyses proposed by the Reviewer could be somehow biased by the counting itself. That is to say, in the case of measuring the thickness of molecular layer, the comparisons should be made between lobe X (with many Purkinje cells) and other lobes (with few/scarce Purkinje cells), which would provide logical but over estimated results. Some similar may occur when comparing the dendrites of healthier Purkinje cells of lobe X with dendrites of ill/fated to die neurons of other lobes. All in all, it would be the comparison of characteristics of healthier neurons of longer survival with other cells that are damaged and prompt to die. Finally, the effect of HSP25 in other models of neuronal loss has classically analyzed in terms of cell death/survival. In sum, although the proposed analysis would be certainly interesting, in our humble opinion, the efforts ad time to address this question would not be worthy in a very limited response time, when stronger data (number of Purkinje cells) is already provided.

5. The text within the discussion feels proportionally overly long and would benefit from shortening.

We have shortened the Discussion section as much as possible. However, please, note that such shortening was delicate and it has not been done in a very extensive manner, as other previous and current Reviewers have asked several specific and deep explanations in this section.

6. Product codes for the antibodies described within the study should be provided within the methods section.

Following the advice of the Reviewer, these data are included in the new version of the manuscript.

Reviewer 2 Report (New Reviewer)

Comments and Suggestions for Authors

The authors present a study exploring mechanisms of resistance in cerebellar lobe X using the PCD mouse model of severe Purkinje cell degeneration. Through immunochemical analysis, they report increased expression of HSP25 and its phosphorylated form, suggesting a potential neuroprotective role. The work addresses a relevant topic and contributes to the understanding of differential vulnerability within the cerebellar cortex. However, I have several comments and points that require clarification.

 

Major Comments

1. In Figure 2C, the authors show HSP25-positive but Calbindin-negative cells located within the Purkinje cell layer. Could the authors clarify which cell population these might correspond to? Additionally, please explain the biological interpretation of this finding.
2. The meaning of continuous, dashed, and dotted lines used above each subpanel in Figures 3 and 5 is not clearly explained. I recommend clarifying the graphical code either in the figure legend or in the methodology section.
3. The manuscript states that, after applying a Kruskal–Wallis test, the Student–Newman–Keuls test was used as the post hoc comparison. However, this post hoc test is only appropriate for parametric data following ANOVA. Since Kruskal–Wallis is a nonparametric test based on ranks, it implies that data did not meet parametric assumptions. In the same line, nonparametric data are more appropriately represented using box plots rather than mean ± standard error, as neither Mann–Whitney U nor Kruskal–Wallis compare means or variance, but rather rank distributions. I suggest reassessing the statistical strategy, justifying the use of nonparametric analysis, explaining the rationale for using a parametric post hoc test, and considering replotting the results using box plots. The legend of Figure 3 also refers to differences “between means,” although the statistical test evaluates ranked data.
4. Is the expression level (cell density) of HSP25 expected in wild-type mice? Should it be considered as a basal or physiological level? Please clarify.
5. In relation to Figure 3B, the authors state: “However, at P25 and P30 the density of HSP25-expressing neurons was higher in PCD than in wild-type mice (pP25 = 0.08 and pP30 = 0.03; Fig. 3B)” (lines 183–184). However, p = 0.08 does not meet the conventional significance threshold (p < 0.05). Please clarify why this is interpreted as a significant difference, or consider modifying the statement.
6. The statement in lines 193–194 “In this second stage, at P35, the percentage of HSP25 expression was even greater than 100%” needs further explanation, since values above 100% require clarification in terms of quantification method or normalization.
7. In lines 262–264, the manuscript states that clear differences were observed among lobes I–IX and lobe X; however, only results for lobe X are shown. Please clarify whether analyses were performed in other lobes and, if so, consider including representative data.
8. Lines 269–270 indicate that TUNEL-positive elements were quantified in the Purkinje cell layer as an estimator of neuronal death. However, it is not clear which cellular elements were quantified, nor how this measurement specifically reflects Purkinje cell death. Additionally, Figure 4 provides low-resolution images with limited detail. Higher-magnification images showing individual cells for lobe X and for the remaining lobes quantified would strengthen the interpretation.
9. Regarding the effect of rottlerin, only illustrative images and qualitative descriptions are presented. Were quantitative measurements or cell counts performed? If not, please justify.
10. Section 5.1 would benefit from providing evidence supporting previous validation of the model and of the primers used.
11. According to line 565, several observations were made using only one animal. Is this sample size sufficient to support the conclusions presented? Please justify the sample size and statistical power.
12. I recommend including a section discussing study limitations.

 

Minor Comments

1. Reference 42 is cited for the first time in line 66, after reference 11. Please revise citation order to maintain numerical sequence.
2. Supplementary material cannot be accessed. Please verify and re-upload the files.
3. In Figure 3B, the title contains a typographical error (“HSP”).
4. The manuscript uses the prefix “p” combined with group or age labels (e.g., pPCD, pP25) to indicate statistical differences. Please explain this notation clearly to avoid confusion.
5. Please improve the quality and resolution of Figures 4 and 6.
6. I recommend updating the reference list, as only a few citations correspond to the last five years. Including more recent literature may improve context and relevance.

Author Response

Response to REVIEWER #2

We would like to thank Reviewer 2 for the careful evaluation of our work and for the constructive feedback provided. We truly appreciate the acknowledgement of our study, as well as the suggested corrections. The have been definitely very useful for refining the manuscript and, thus, increasing its scientific accuracy. We have addressed each point in detail below, and the related changes appear in blue in the new version of the manuscript.

The authors present a study exploring mechanisms of resistance in cerebellar lobe X using the PCD mouse model of severe Purkinje cell degeneration. Through immunochemical analysis, they report increased expression of HSP25 and its phosphorylated form, suggesting a potential neuroprotective role. The work addresses a relevant topic and contributes to the understanding of differential vulnerability within the cerebellar cortex. However, I have several comments and points that require clarification.

Major Comments

1. In Figure 2C, the authors show HSP25-positive but Calbindin-negative cells located within the Purkinje cell layer. Could the authors clarify which cell population these might correspond to? Additionally, please explain the biological interpretation of this finding.

Calbindin is a calcium-binding protein that is widely considered as a marker of Purkinje cells (Sillitoe et al., 2007, Annu Rev Cell Dev Biol; Muñoz-Castañeda et al., 2018, Sci Rep). However, there are scarce Purkinje neurons that constitute a population of calbindin-negative cells. This same phenomenon occurs with others classical marker for Purkinje cells, like parvalbumin. In addition, PCD neurodegeneration is accompanied by gene and protein silencing before neuronal death (Baltanás et al., 2011, J Biol Chem), which includes neuronal markers. Therefore, it is plausible to think that damaged Purkinje cells could stop expressing some genes/proteins, but maintaining the production of putative neuroprotective factors, as HSP25. This explanation is now included in the Discussion section of the new version of the manuscript.

2. The meaning of continuous, dashed, and dotted lines used above each subpanel in Figures 3 and 5 is not clearly explained. I recommend clarifying the graphical code either in the figure legend or in the methodology section.

The other Reviewer has addressed the same question. Differences between genotypes (i.e. wild type or PCD) are signaled by asterisks at each time point. We have also compared the values of time points within each genotype. To avoid blurring graphs with many different symbols, especially when several time points may have statistically similar values, we decided to use the aforementioned horizontal lines. Note that those lines are drawn at different heights, thus corresponding to higher or lower values of an analyzed variable depending on the time point. Those time points with statistically similar values have a line at the same height. Finally, blue lines represent values of wild-type animals and red lines represent values of PCD mice. For example, in figure 1A, values of Purkinje cell density in wild-type mice (blue lines) at P15 are statistically higher than those of P20 onwards, which are statistically similar amongst them; besides, values of PCD mice (red lines) are statistically similar from P15 to P30, but at P35 are lower. A more detailed explanation is now included in the new version of the manuscript.

3. The manuscript states that, after applying a Kruskal–Wallis test, the Student–Newman–Keuls test was used as the post hoc comparison. However, this post hoc test is only appropriate for parametric data following ANOVA. Since Kruskal–Wallis is a nonparametric test based on ranks, it implies that data did not meet parametric assumptions. In the same line, nonparametric data are more appropriately represented using box plots rather than mean ± standard error, as neither Mann–Whitney U nor Kruskal–Wallis compare means or variance, but rather rank distributions. I suggest reassessing the statistical strategy, justifying the use of nonparametric analysis, explaining the rationale for using a parametric post hoc test, and considering replotting the results using box plots. The legend of Figure 3 also refers to differences “between means,” although the statistical test evaluates ranked data.

The Reviewer addresses a very interesting question. Indeed, we had already thought a lot about this trouble, and after consulting with the statistician of our research unit, we decided to proceed as follows. First of all, we resolved to use non-parametric tests, since the sizes of experimental groups are not big and, thus, homoscedasticity and normality cannot be always ensured. In any case, the statistics program used (SPSS Statistics 26) allows the implementation of a post hoc analysis after Kruskal–Wallis test, which is analog to the Student–Newman–Keuls test. We have corrected this detail in the new version of the manuscript. Finally, although we agree that, strictly, the best form for representing non-parametric data are box plots rather than mean ± standard error, the use of this system would blur the data, especially in those graphs with both temporary and inter-genotype comparisons (the majority of this work). In this sense, the use of more intelligible data as mean ± standard error was chosen to ease the readers the comprehension of our work. We understand that this method for representing data is not the strictest, but we decided to sacrifice formal accuracy to gain comprehensibility and disseminative power. In any case, this procedure is not uncommon, as many scientific works use mean ± standard error for representing non-parametric data.

4. Is the expression level (cell density) of HSP25 expected in wild-type mice? Should it be considered as a basal or physiological level? Please clarify.

It should be. To our knowledge, previous works addressing the basal expression of HSP25 just describe qualitatively its presence or absence in wild-type or pathological conditions. They also relate the presence of this protein with neuroprotective effects, but the quantifications are circumscribed to the survival/number of Purkinje cells. Moreover, in many publications the analyses concerning HSP25 distribution are performed in coronal sections of cerebellum, which complicates a direct comparison of densities with our work. Moreover, to our knowledge, our work is the first one that offers temporal quantitative data of HSP25 distribution in wild-type animals. Note that, before the onset of degeneration (i.e. without additional known stimuli), the density of HSP25-positive Purkinje cells is similar in both control and PCD mice, which supports the HSP25 basal expression level.

5. In relation to Figure 3B, the authors state: “However, at P25 and P30 the density of HSP25-expressing neurons was higher in PCD than in wild-type mice (pP25 = 0.08 and pP30 = 0.03; Fig. 3B)” (lines 183–184). However, p = 0.08 does not meet the conventional significance threshold (p < 0.05). Please clarify why this is interpreted as a significant difference, or consider modifying the statement.

This is certainly a writing mistake. The p-value of comparisons at P25 is 0.008 instead of 0.08. That’s why in the Figure 3B two asterisks are shown at the point corresponding to P25. We sincerely thank to the Reviewer for noticing this error!

6. The statement in lines 193–194 “In this second stage, at P35, the percentage of HSP25 expression was even greater than 100%” needs further explanation, since values above 100% require clarification in terms of quantification method or normalization.

In a previous version of the manuscript we had included an explanation for this issue, but after a shortening of the discussion (suggested by other reviewers), we had to eliminate this point that, even very interesting, was considered secondary. We have included this explanation again in the corresponding part of Results of the new version of the manuscript. Briefly, at P35, we found that the percentage of HSP25 exceeds 100%: that is to say, there would be more cells expressing HSP25 than calbindin. This paradox is due to the fact that, at this point of neuronal degeneration, a percentage of Purkinje cells stops expressing calbindin, even though they are positive for HSP25 (see major comment 1; Baltanás et al., 2011, J Biol Chem).

7. In lines 262–264, the manuscript states that clear differences were observed among lobes I–IX and lobe X; however, only results for lobe X are shown. Please clarify whether analyses were performed in other lobes and, if so, consider including representative data.

Differences amongst lobes have been previously reported by our group (Hernández Pérez et al., 2023, Anatomia; Hernández-Pérez et al., 2025, Int J Mol Sci) and others, as it has been clearly described in the Introduction and Discussion sections. Moreover, Supplementary Figure 1 clearly shows such differences, although our study is focused in the lobe X. In any case, taking into account this commentary as well as other one from the Reviewer 1, we have included new data corresponding other lobes, which are gathered in new Supplementary Figure 3.

8. Lines 269–270 indicate that TUNEL-positive elements were quantified in the Purkinje cell layer as an estimator of neuronal death. However, it is not clear which cellular elements were quantified, nor how this measurement specifically reflects Purkinje cell death. Additionally, Figure 4 provides low-resolution images with limited detail. Higher-magnification images showing individual cells for lobe X and for the remaining lobes quantified would strengthen the interpretation.

TUNEL-positive elements are necessarily apoptotic nuclei, which has been specified in the new version of the manuscript. As it is explained in this section, we have counted these elements in the Purkinje cell layer, since the combination of an immunohistochemistry against calbindin and TUNEL was not suitable. The death of Purkinje cells is extremely fast and these neurons stop expressing many proteins (including neuronal markers) before dying. Then, the detection of double-labelled cells for calbindin and TUNEL is extremely complicated and does not reflect the real death rate of Purkinje cells. On the contrary, the quantification of apoptotic cell nuclei in the specific Purkinje cell layer gives a suitable estimation of this phenomenon (Pérez-Revuelta et al., 2025, Int J Mol Sci). Of course, we cannot discard a minority quantification of other apoptotic cells in this layer, but this number is negligible in comparison with the main Purkinje cell death rate. Moreover, the quantification in the control group (wild type) should compensate any bias. In any case, this limitation will be collected in the section purposed by this Reviewer (see below).

Concerning images of Figure 4, they have been taken with an epifluorescence microscope of high quality. However, the limitations pointed by the Reviewer could be due to the thickness of the tissue, at least, partially. Note that in order to do a quantification as complete as possible, we did not take focal planes, which may induce a loss of information. Then, we decided to include the real images that we used to show their pros and cons. Moreover, the double immunostaining against PKC-δ and HSP25-P-Ser15 was not as clean as other immunolabeling procedures, due to the sequential technique used (see Material and Methods). Even when this limitation, the technique did not provide false positives (see Material and Methods and Supplementary Figure 4) and allowed the quantification of neurons of interest. Furthermore, the acquisition of individual Purkinje cells is not easy, as they constitute a continuous layer, with a mixture of the ramifications of their dendritic trees. This is especially evident in lobe X, where even in PCD mice, this layering is protected. In this sense, only in the lobes I-IX of PCD mice in a clear degenerative moment, individual Purkinje cells could be distinguished and showed. In this situation, these neurons would be subjected to degeneration and their morphology would be altered, thus being not representative. In any case, the new Supplementary Figure 3 provides additional data concerning PKC-δ and HSP25-P-Ser15 data, not only in lobe X, but also in the other lobes of vermis. Then, this new figure may fulfill the requirements of the Reviewer.

9. Regarding the effect of rottlerin, only illustrative images and qualitative descriptions are presented. Were quantitative measurements or cell counts performed? If not, please justify.

This experiment was confirmative. Then, as it is shown in Material and Methods section (Table 1), just 2 animals per experimental group were used. In a previous version of the manuscript, the presence of HSP25/HSP25-P-Ser15 and the survival of more Purkinje cells in the lobe X was suggested as just correlative, but not consequential, even when it had been demonstrated in other neurodegenerative processes (Duffin et al., 2010, J Comp Neurol; Chung et al. 2016, PLoS gen). Then, we decided to perform this experiment, by inhibiting the action of the enzyme (PKC-δ) that turns HSP25 into its active form (HSP25-P-Ser15). As it was a confirmative experiment, the Bioethics Committee of the University of Salamanca did not allow us to use a big number of animals to avoid the sacrifice of an excessive quantity of experimental subjects. Fortunately, the qualitative results of this experiment were solid enough to corroborate a cause-consequence relationship between HSP25 and neuroprotection. In any case, the limitation inherent to qualitative results is also collected in the corresponding section.

• Section 5.1 would benefit from providing evidence supporting previous validation of the model and of the primers used.

The PCD model as well as the primers used for its genotyping have been perfectly validated by our group and others. Following the advice of the Reviewer, we have provided references for this purpose in the section 5.1 of the new version of the manuscript.

According to line 565, several observations were made using only one animal. Is this sample size sufficient to support the conclusions presented? Please justify the sample size and statistical power.

These observations were merely confirmative. Since Marzban et al (2011, Cerebellum) described the constitutive presence of HSP25 in the central-most area of the vermis in coronal sections, we would like to validate that our mice presented the same expression before analyzing other regions and conditions. Thus, we employed just one animal per experimental group, confirming such expression, as the majority of our work was developed in sagittal sections. The use of more mice for repeating non-strictly necessary cell quantifications or descriptions would not be allowed by the Bioethics Committee of the University of Salamanca.

• I recommend including a section discussing study limitations.

Following the advice of the Reviewer, we have included this new section at the end of Discussion.

Minor Comments

1. Reference 42 is cited for the first time in line 66, after reference 11. Please revise citation order to maintain numerical sequence.

The Reviewer is right. We have revised and corrected the order of references. Thanks for this observation.

2. Supplementary material cannot be accessed. Please verify and re-upload the files.

We are sorry about that inconvenience. We have uploaded all files following the guidelines of the journal and other reviewers have not reported any trouble. In any case, we will ensure their upload in the new resubmission. Moreover, we will try to include those files directly in the response to the Reviewer to ensure their assessment.

3. In Figure 3B, the title contains a typographical error (“HSP”).

Thanks to the Reviewer for showing this mistake. We have corrected it.

4. The manuscript uses the prefix “p” combined with group or age labels (e.g., pPCD, pP25) to indicate statistical differences. Please explain this notation clearly to avoid confusion.

The notation is explained in the new version of the manuscript.

5. Please improve the quality and resolution of Figures 4 and 6.

Concerning images of Figure 4, the limitations of its visualization have been previously explained (see Major comment 8).

Concerning Figure 6, we have kept the low-magnification images of sagittal sections, as they show the entire distribution of TUNEL-positive nuclei in the cerebellar vermis. In addition, the Reviewer 1 asked for these type of images. In any case, we have included new high-resolution confocal images (focal plane, maximum projection image and 3D reconstruction) in the new version of this figure, following the suggestion of this Reviewer.

6. I recommend updating the reference list, as only a few citations correspond to the last five years. Including more recent literature may improve context and relevance.

Following the advice of the Reviewer, new actualized references have been included in the new version of the manuscript.

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report (New Reviewer)

Comments and Suggestions for Authors

Thank you for the revised version of the manuscript and for the authors’ efforts to address the comments. After evaluation of the responses, I believe that some issues still require clarification before the manuscript can be considered fully supported.

First, the concern regarding the post hoc analysis following the Kruskal–Wallis test remains. This is a methodological issue rather than a matter of software implementation. For rank-based nonparametric tests, post hoc comparisons should be performed using appropriate nonparametric procedures with correction for multiple testing (e.g., Dunn or Conover tests). The authors should clearly specify the exact post hoc method applied in SPSS and indicate how multiple comparisons were controlled. Reanalysis using a standard nonparametric post hoc approach should be considered.

Second, the presentation of nonparametric data as mean ± SEM is not fully aligned with rank-based statistical inference and may be misleading. Alternative representations (e.g., median and interquartile range, individual data points, or distribution-based plots provided as supplementary material) would better reflect the statistical analysis performed. In addition, wording in the Results section and figure legends should avoid references to “differences between means” when nonparametric tests are used (fig. 3).

Finally, Supplementary Figure 3 contains visible word-processor spell-check or markup indicators. This figure should be replaced with a clean, final version.

Addressing these points would improve the clarity, methodological transparency, and consistency of the manuscript.

Author Response

We would like to thank again Reviewer 2 for a new evaluation of our work and for the constructive feedback provided. Changes derived from his/her suggestions appear in blue in the new version of the manuscript.

Thank you for the revised version of the manuscript and for the authors’ efforts to address the comments. After evaluation of the responses, I believe that some issues still require clarification before the manuscript can be considered fully supported.

We will try to satisfy all the Reviewer’s demands.

First, the concern regarding the post hoc analysis following the Kruskal–Wallis test remains. This is a methodological issue rather than a matter of software implementation. For rank-based nonparametric tests, post hoc comparisons should be performed using appropriate nonparametric procedures with correction for multiple testing (e.g., Dunn or Conover tests). The authors should clearly specify the exact post hoc method applied in SPSS and indicate how multiple comparisons were controlled. Reanalysis using a standard nonparametric post hoc approach should be considered.

The SPSS offers the possibility of using different post hoc comparisons after the nonparametric Kruskal-Wallis test in the last versions, included the current v.30, thus being perfectly validated. In our case, we have use an analog of the Student-Newman-Keuls comparison called Stepwise step-down comparisons (https://www.ibm.com/docs/en/spss-statistics/cd?topic=tests-choose-independent-samples-nonparametric). Such test allows the establishment of the most probable homogeneous subsets of data. Such subsets are based on asymptotic significations, at a level of significance set at 0.05. This explanation now appears in the corresponding part of Material and Methods section of the new version of the manuscript.

Second, the presentation of nonparametric data as mean ± SEM is not fully aligned with rank-based statistical inference and may be misleading. Alternative representations (e.g., median and interquartile range, individual data points, or distribution-based plots provided as supplementary material) would better reflect the statistical analysis performed. In addition, wording in the Results section and figure legends should avoid references to “differences between means” when nonparametric tests are used (fig. 3).

Following the advices of the Reviewer, we have included new box plot representations for all the analyzed data. Such information is placed next to the corresponding old graphs to ease the readers the interpretation and comprehension of data, and possible statistical differences between groups or ages, as well as their distribution.

Concerning wording, any reference to “differences between means” has been removed from the Results section and figure legends.

Finally, Supplementary Figure 3 contains visible word-processor spell-check or markup indicators. This figure should be replaced with a clean, final version.

The Reviewer is completely right. The rush to deliver the corrections on time led us to overlook errors such as this one. It has now been fixed.

Addressing these points would improve the clarity, methodological transparency, and consistency of the manuscript.

We hope these last amendments will fulfill the requirements of the Reviewer. If it not were the case, please, don’t hesitate to contact us and continue this fruitful round of discussion.

Author Response File: Author Response.pdf

Round 3

Reviewer 2 Report (New Reviewer)

Comments and Suggestions for Authors

The authors have adequately addressed the comments, and I consider the manuscript suitable for publication in its current form.

This manuscript is a resubmission of an earlier submission. The following is a list of the peer review reports and author responses from that submission.

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

This manuscript investigates the spatial patterns of neurodegeneration across cerebellar lobes
in the context of neurodegenerative conditions, with a focus on Purkinje cell (PC) survival in
the nodular lobe (lobe X). Using the Purkinje Cell Degeneration (PCD) mouse model, which
mimics childhood-onset cerebellar atrophy, the authors study the expression of heat shock
protein 25 (HSP25) and its phosphorylated form (HSP25-P-Ser15) through
immunohistochemistry (IHC). The findings suggest that in lobe X HSP25 may play a
neuroprotective role against PC apoptosis. Moreover, the phosphorylated form, HSP25-PSer15,
is weakly expressed but appears to be upregulated during neurodegeneration, possibly
mediated by PKC-δ.
1. Strengths of the Study
• The study addresses an important question regarding region-specific neuroprotection
in the cerebellum.
• The PCD mouse model is well-suited for studying neurodegeneration and the authors
convincingly demonstrate lobe-specific differences.
• Immunohistochemistry was well executed, with clearly presented spatiotemporal
expression patterns of HSP25, its phosphorylated derivative and PKC-δ with clear
presentation of technical limitations.
2. Limitations
• The sole experimental technique used is immunohistochemistry. While this provides
valuable spatial information, the conclusions would be strengthened by
complementary techniques (e.g., RT-qPCR or Western blot) to validate expression
levels of HSP25 and PKC-δ at the mRNA or protein level.
• The study lacks functional assays to establish causality. All observations are
descriptive and correlative. This limits the ability to firmly conclude that HSP25 or
HSP25-P-Ser15 plays a protective role in PC survival.
• The quantification of HSP25-P-Ser15 is based on a very small number of labeled cells,
which makes these data less convincing. Statistical robustness should be discussed.
Minor Comments
1. The introduction should explicitly state that Purkinje cell death in the PCD model
occurs via apoptosis, providing relevant references.
2. The strength of the conclusions should be moderated to reflect the correlative nature of
the study and the reliance on a single experimental approach.
Recommendations
While the data are compelling and the research question is interesting and timely, the authors
should address the study’s methodological limitations. Specifically, incorporating an additional
experimental technique such as in situ hybridization or HCR for HSP25 and PKC-δ would
substantially strengthen the manuscript. Alternatively, the authors should clearly acknowledge
these limitations and adjust their conclusions accordingly.

Comments for author File: Comments.pdf

Author Response

Response to REVIEWER #1 comments

We would like to sincerely thank the Reviewer for the thorough evaluation of our work and for the constructive feedback provided. We truly appreciate the recognition of the study’s strengths, as well as the insightful comments on methodological aspects and data interpretation. These observations have been very helpful in refining the manuscript, and we have addressed each point in detail below.

This manuscript investigates the spatial patterns of neurodegeneration across cerebellar lobes in the context of neurodegenerative conditions, with a focus on Purkinje cell (PC) survival in the nodular lobe (lobe X). Using the Purkinje Cell Degeneration (PCD) mouse model, which mimics childhood-onset cerebellar atrophy, the authors study the expression of heat shock protein 25 (HSP25) and its phosphorylated form (HSP25-P-Ser15) through immunohistochemistry (IHC). The findings suggest that in lobe X HSP25 may play a neuroprotective role against PC apoptosis. Moreover, the phosphorylated form, HSP25-PSer15, is weakly expressed but appears to be upregulated during neurodegeneration, possibly mediated by PKC-δ.

1. Strengths of the Study:

• The study addresses an important question regarding region-specific neuroprotection in the cerebellum.
• The PCD mouse model is well-suited for studying neurodegeneration and the authors convincingly demonstrate lobe-specific differences.
• Immunohistochemistry was well executed, with clearly presented spatiotemporal expression patterns of HSP25, its phosphorylated derivative and PKC-δ with clear presentation of technical limitations.

2. Limitations

• The sole experimental technique used is immunohistochemistry. While this provides valuable spatial information, the conclusions would be strengthened by complementary techniques (e.g., RT-qPCR or Western blot) to validate expression levels of HSP25 and PKC-δ at the mRNA or protein level.

We appreciate this suggestion and fully agree that complementary techniques such as RT-qPCR or Western blot could provide additional and valuable information. Unfortunately, due to current limitations in time, resources and the availability of experimental animals, it is not feasible for us to perform these experiments at this stage. Moreover, while genetic analyses are extremely informative, their results do not always correlate directly with protein expression, which is the primary focus of our study. Our objective was specifically to determine, at the single-cell level, the number and spatial distribution of Purkinje cells expressing HSP25 and PKC-δ in lobe X, something that cannot be directly obtained from mRNA analyses, which provide averaged values across heterogeneous cell populations. Indeed, in certain contexts, cells may express mRNA for a given protein without detectable protein expression, which would not align with the aims of this work. In this sense, protein-level data can serve to validate genetic results, but not necessarily the other way around.

• The study lacks functional assays to establish causality. All observations are descriptive and correlative. This limits the ability to firmly conclude that HSP25 or HSP25-P-Ser15 plays a protective role in PC survival.

We fully agree that functional assays would be the most direct way to establish a causal relationship between HSP25/HSP25-P-Ser15 expression and Purkinje cell survival. However, such experiments require specific genetic or pharmacological models that are not available in our facilities and could not be implemented within the current revision timeframe. As suggested, we have acknowledged this limitation in the revised manuscript (changes marked in green) and have incorporated into the Conclusions a statement explicitly noting the correlative nature of our study and the need for targeted manipulations. We believe our present results lay a solid foundation for future studies, and we plan to address causality in collaboration with groups possessing the appropriate models.

• The quantification of HSP25-P-Ser15 is based on a very small number of labeled cells, which makes these data less convincing. Statistical robustness should be discussed.

We thank the Reviewer for pointing out the limited number of HSP25-P-Ser15-positive Purkinje cells in our quantification. This low count reflects a genuine biological feature of the system rather than a technical shortcoming, as similarly low expression levels have been described in other models (e.g., Chung et al., 2016). While we acknowledge that small sample sizes may reduce statistical power, we have applied non-parametric test appropriate for this context, and the temporal patterns observed were consistent across animals. To address this point, we have included in the Discussion (changes marked in green) a statement explicitly acknowledging this limitation and clarifying its biological basis.

Minor Comments

1. The introduction should explicitly state that Purkinje cell death in the PCD model occurs via apoptosis, providing relevant references.
2. The strength of the conclusions should be moderated to reflect the correlative nature of the study and the reliance on a single experimental approach.

We are thankful for these suggestions. As recommended, we have modified the Introduction to explicitly state that Purkinje cell death in the PCD model occurs through apoptotic-like mechanisms, supported by published evidence (Baltanás et al., 2011; changes marked in green). In addition, following the advice, we have moderated the strength of the Conclusions to better reflect the correlative nature of our findings and the reliance on a single experimental approach. Specifically, we have included a statement acknowledging that the study is correlative and does not directly establish causality (changes marked in yellow, as it responds to comments from both reviewers).

Recommendations
While the data are compelling and the research question is interesting and timely, the authors should address the study’s methodological limitations. Specifically, incorporating an additional experimental technique such as in situ hybridization or HCR for HSP25 and PKC-δ would substantially strengthen the manuscript. Alternatively, the authors should clearly acknowledge these limitations and adjust their conclusions accordingly.

We appreciate these recommendations. As outlined above and due to limitations in time, we have acknowledged the methodological limitations of relying solely on immunohistochemistry and have moderated the conclusions accordingly. These changes, now incorporated into the revised Discussion and Conclusions, address the Reviewer’s concerns while maintaining the focus of the present study.

Reviewer 2 Report

Comments and Suggestions for Authors

This is a study that uses immunohistochemistry of the cerebellum in normal and PCD mice to examine what molecular factors contribute to the delayed degeneration of Lobe X in PCD mice. Based on their data, the authors conclude that the induction of expression of HSP25 and the p15-phosphorylated form of this protein contributes to the relative resistance of Lobe X to degeneration. Although addressing a significant question and providing information that is interesting, the data presented in the report does not adequately support the conclusions. Primarily, the correlation between HSP25/HSP25-P-15 expression and PC resistance in the cerebellum of PCD mice is weak. Other weaknesses include some inconsistencies, the lack of a cell death marker, and the poor quality of the figures, which appear somewhat fuzzy. Dendritic arborization is hardly discernible. A major concern is that the report is similar to a publication by Duffin et al. (2010), thereby lacking novelty. Although there is advancement of information, this is incremental.

Specific points

Evidence must be presented to show that degenerating PCs (based on TUNEL or other assays) are more likely not to express HSP25. As presented, the data does not adequately support the conclusion.

That authors state that “The expression of both PKC-δ and HSP25-P-Ser15 in Purkinje cells was much less abundant than that of calbindin and HSP25.”. This is confusing because all cells that express the phosphorylated form of HSP25 must express HSP25.

It is important to show evidence that HSP25-P-Ser15 is localized almost exclusively to lobe X. In Lobe X, is it primarily in PCs of the ventral portion of the lobe like total HSP25?

Supp. Fig- 1 shows that in the PCD mice, degeneration occurs unevenly in that most lobes display stripes of PCs that are resistant. If HSP25 / HSP25-Ser15 expression is a significant factor in protecting PCs, its expression should also be localized in the stripes of PCs that survive.  In other words, the expression patten of HSP25 does not reflect the pattern of resistance. Additionally, whereas Supp. Fig-1 shows expression of HSP25 uniformly across Lobe X, the authors state earlier that expression is restricted to the ventral portion of the lobe. Which one is it? In the introduction section also the authors state widespread HSP25 expression in Lobe X.

The authors claim that HSP25 is “exclusively located in the ventral-most area of lobe X”. However, much of Lobe X is resistant to degeneration undercutting the main conclusion of the report that HSP25 expression is a major contributor to the resistance of PCs to degeneration. 

The data does not adequately support the claim that expression of HSP25/HSP25-p-Ser15 is “induced”.

Author Response

Response to REVIEWER #2 comments

We would like to thank Reviewer 2 for the careful evaluation of our manuscript and for the constructive feedback provided. We truly appreciate the recognition of the interest of our study, as well as the insightful comments regarding methodological aspects, data interpretation, and novelty. While we acknowledge the contribution of Duffin et al. (2010), our work goes beyond by providing a detailed spatiotemporal analysis of HSP15 and its phosphorylated form in the PCD mouse model, including their relationship with PKC- δ and the region-specific resistance of lobe X. We believe these observations extend previous findings and provide additional insight into the mechanisms underlying differential vulnerability in cerebellar neurodegeneration. Concerning images, they have been taken with epifluorescence microscopes of high quality. However, the limitations pointed by the Reviewer can be due to the thickness of the tissue. Note that in order to do a quantification as complete as possible, we did not take focal planes, which may induce a loss of information. Concerning dendritic trees, it is possible that sections are not completely sagittal in all of their extension to take all these complete arborizations. In addition, some images correspond to PCD animals, whose Purkinje cells suffer impairments in its dendritic arborization. Then, in this case there are not any technical problem, but a biological feature. These comments have been very helpful in refining the manuscript and clarifying our conclusions. Below, we address each of the points raised in detail.

This is a study that uses immunohistochemistry of the cerebellum in normal and PCD mice to examine what molecular factors contribute to the delayed degeneration of Lobe X in PCD mice. Based on their data, the authors conclude that the induction of expression of HSP25 and the p15-phosphorylated form of this protein contributes to the relative resistance of Lobe X to degeneration. Although addressing a significant question and providing information that is interesting, the data presented in the report does not adequately support the conclusions. Primarily, the correlation between HSP25/HSP25-P-15 expression and PC resistance in the cerebellum of PCD mice is weak. Other weaknesses include some inconsistencies, the lack of a cell death marker, and the poor quality of the figures, which appear somewhat fuzzy. Dendritic arborization is hardly discernible. A major concern is that the report is similar to a publication by Duffin et al. (2010), thereby lacking novelty. Although there is advancement of information, this is incremental.

Specific points:

• Evidence must be presented to show that degenerating PCs (based on TUNEL or other assays) are more likely not to express HSP25. As presented, the data does not adequately support the conclusion.

We thank the Reviewer for this valuable suggestion. We fully agree that the most direct way to establish the relationship between HSP25 expression and Purkinje cell survival would be to combine HSP25 immunolabeling with apoptotic markers such as TUNES or caspase-3. We have in fact attempted TUNEL assays; however, the technical limitation is that this method requires capturing cells within a very narrow time window of apoptosis. This is particularly challenging in lobe X, where the number of degenerating Purkinje is lower and most of them express HSP25, making the detection of TUNEL-positive, HSP25-negative cells improbable. Instead, our conclusions rely on correlative evidence: Purkinje cells in lobe X, which express HSP25, are relatively resistant, whereas those in lobes I-IX, largely lacking HSP25, degenerate earlies. We acknowledge that this inference would be more directly supported by colocalization with apoptotic markers, and we have now clarified in the revised Conclusions (changes marked in blue) that future studies will be necessary to address this limitation.

• That authors state that “The expression of both PKC-δ and HSP25-P-Ser15 in Purkinje cells was much less abundant than that of calbindin and HSP25.”. This is confusing because all cells that express the phosphorylated form of HSP25 must express HSP25.

It is correct that cells expressing HSP25-P-Ser15 must also express HSP25. What we intended to convey was the relative abundance of immunolabeling: far fewer Purkinje cells were positive for HSP25-P-Ser15 compared to total HSP25. We have revised the sentence accordingly to avoid confusion (change marked in blue)

• It is important to show evidence that HSP25-P-Ser15 is localized almost exclusively to lobe X. In Lobe X, is it primarily in PCs of the ventral portion of the lobe like total HSP25?

We agree that clarifying the localization of HSP25-P-Ser15 within lobe X is important. In our material, the expression of this phosphorylated form was restricted to lobe X (with only rare positive cells in lobe VI of PCD mice), but its low abundance prevented us from identifying a clear ventral predominance comparable to total HSP25. We have modified the results section accordingly to better reflect this point (change marked in blue)

• Supp. Fig- 1 shows that in the PCD mice, degeneration occurs unevenly in that most lobes display stripes of PCs that are resistant. If HSP25 / HSP25-Ser15 expression is a significant factor in protecting PCs, its expression should also be localized in the stripes of PCs that survive. In other words, the expression patten of HSP25 does not reflect the pattern of resistance. Additionally, whereas Supp. Fig-1 shows expression of HSP25 uniformly across Lobe X, the authors state earlier that expression is restricted to the ventral portion of the lobe. Which one is it? In the introduction section also the authors state widespread HSP25 expression in Lobe X.

We thank the Reviewer for pointing out this potential source of confusion. We agree that the description of HSP25 expression in lobe X could be misinterpreted as inconsistent across the text and figures. As the Reviewer notes, the images in fact correspond to wild-type and PCD mice, as indicated in their respective labels, which explains the ventral-restricted versus widespread expression patterns. Nevertheless, to avoid any ambiguity for the reader, we have clarified this point at three levels: (1) in the Introduction, by rephrasing the description of HSP25 in lobe X; (2) in the Results, by explicitly stating that HSP25 is ventral in wild-type mice but generalized in PCD; and (3) in the legend of Supplementary Figure 1, by adding a note that calbindin-positive Purkinje cells observed in lobes I-IX at this age are only transiently preserved and not truly resistant populations. These changes, highlighted in blue, ensure consistency between the figures and the text (all changes marked in blue).

• The authors claim that HSP25 is “exclusively located in the ventral-most area of lobe X”. However, much of Lobe X is resistant to degeneration undercutting the main conclusion of the report that HSP25 expression is a major contributor to the resistance of PCs to degeneration.

We realize that the original wording could have created ambiguity. As clarified in our previous response, we have now specified that HSP25 expression is ventral-restricted just in wild-type mice, but generalized across the entire lobe X in PCD animals. We hope that with these clarifications, the apparent discrepancy is resolved.

• The data does not adequately support the claim that expression of HSP25/HSP25-p-Ser15 is “induced”.

We appreciate the observation regarding the use of the term “induced”. We agree that our data are correlative and do not directly demonstrate induction. To address this, we have revised the wording in the Discussion and in the Results (changes marked in blue). In addition, the Conclusions now explicitly state that “The present study is correlative in nature and does not directly establish causality” (highlighted in yellow, as both reviewers addressed this question), which we believe is fully in line with this comment.

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

The manuscript has been edited appropriately

Author Response

Many thanks for your comments, suggestions and your finally evaluation.