Write a comment

This review reflects comments and contributions by Bhargy Sharma, Ryan Cubero & Anna Oliveras. Review synthesized by Ryan Cubero.

In this preprint, De Virgiliis and colleagues use a murine model of sciatic injury to provide functional evidence that peripheral nerve regeneration is affected by circadian rhythm. The authors show that regenerative capacity of mice DRG neurons is time-of-day dependent and that the disruption of the intrinsic neuronal circadian rhythm via knockdown or deletion of Bmal1, a non-redundant core clock gene, decreases regeneration capability after injury. They found that injuries performed at ZT20 induce a transcriptional program that enhances regeneration and targets long-lasting re-innervation. Also, lithium, a chono-active compound, is  pinpointed as a new therapeutic avenue for nerve repair. Although the role of the circadian clock has already been shown for other wound healing processes, these novel findings describe for the first time the regenerative potential in the PNS is controlled by neuronal intrinsic circadian clock and is time-of-day dependent. 

Positive aspects of the study:

• The preprint starts with a meta-analysis of previously published transcriptomes, which allowed them to establish “circadian rhythms” as a potential common pathway. Studies like this that make use of the vast published transcriptomes should be highly encouraged and commended.

• Overall, the experiments are well designed and well explained with images. The supplementary material provides extensive control experiments that help to draw and strengthen the conclusions.

• The conclusions suggest a careful re-evaluation of previous data interpretation taking in account the time dependency of data acquisition. This might help to solve previous controversial data. As well, the conclusions highlight  important considerations for future experimental designs. We believe that the findings provided by the authors are remarkably relevant for the scientific community, specially those in the field of axonal regeneration research.

• The implications of the study for future therapeutic strategies:

  • Focusing current neurorehabilitation therapies to time-of-day effects

  • Using chrono-active compounds for axon regeneration therapies

  • Potential gene therapy targeting circadian clock molecular machinery

Major aspects to be addressed:

• Although authors found transcriptional differences in ZT20 vs ZT8 after injury, we wonder if there is already a background regeneration machinery that is highly active at ZT20 in unchallenged DRGs that allows for increased regenerative capacity.

• Although authors clearly show that the regeneration ability depends on Baml1, they miss proving that Baml1 is directly binding regulatory gene sequences and actually orchestrating the transcriptional response. If not Baml1, who is the main orchestrator?

• While the authors denote lithium as a chrono-active drug, the way it affects circadian rhythm-dependent enhancement of regenerative ability is unclear. We suggest a thorough characterization of the effect of lithium treatment across all ZT during 24h.

• The authors should discuss the results of the recent publication https://www.nature.com/articles/s41467-023-40816-7 were they found that Baml1 controls axon regeneration via Tet3 epigenetics

• The authors show that the role of inflammation triggered by non-neuronal cells in nerve regeneration is not clock-dependent. Nevertheless, neutrophin, BDNF and NFG levels were measured in naive DRG neurons. The authors should specify if “naive” means uninjured neurons. In this case, we recommend comparing these data with post-injury levels.

Minor comments:

• In the introduction, we would suggest to include more background information (more details on the different mechanisms of PNS regeneration or evolutionary perspective on why regeneration might be controlled by circadian rhythms) and limit the description of the results to a very short summary.

• We consider that Figure 1 can gain some clarity by reworking the following aspects:

  • The heatmap shows biological processes instead of genes. Caption of the figure legend should be corrected accordingly.

  • Mention abbreviation of “DE genes” in figure legend

  • In the comparative GO analysis, we wonder how much overlap there is in the differentially expressed circadian rhythm genes across the regenerative models?

  • Mention abbreviation of SGC10 in the main text. A reference for this marker would be helpful as well.

  • We suggest having the fluorescence intensity distribution plotted beneath the representative image so readers can get a better sense of the regeneration index.

  • Legend for 1F is incorrect. Should be modified for G and H accordingly.

• We would like to encourage the authors to give some reason why they chose a sciatic nerve crush over sciatic nerve axotomy

• Supplementary figure 1 shows representative images ZT0, 4, 8, 12, 16 and 20 instead of only ZT8 and 20. Caption should be revised. In the same figure legend, authors explain that “Fluorescence intensity was measured in one series of tissue sections for each nerve”. We suggest providing more details on the average size of each tissue section, and how many sections in one series.

• Edit y-axis label of supplementary figure 2B to 'CTB+ DRG neurons' 

• In supplementary figures 3 and 4, describing abbreviations for LY6G and CD68 will help the reader. As well, we suggest to mark on the images the site of the injury and double-check consistency between the number of replicates shown in the plots and the figure legend.

• How many hours post-injury refers to the data shown in supplementary figure 5?

• The text of supplementary figure 7 legend should be checked for clarity,  “mRNA levels of Bmal1 mRNA levels” is confusing and also the references to other figures.

• Also, why in supplementary figure 7B, 5 replicates are chosen instead of 6 that were included in figure 1D?

• Figure 2B and supplementary Figure 8A should be consistently labeled using the same labels (ZT8_1, ZT8_2, and so on) in the heatmap and the PCA analysis of the clustered differentially expressed genes.

• Regarding the data presented in Supplementary Figure 8B, we wonder if there could be any reason why the ZT20 transcriptome is more likely to resemble IF than EE, if both non-injury models promote regeneration? Also, when comparing their data with previously published dataset, we find important to comment at which ZT the samples were collected, if available

• Clarify in Figure 2C how to interpret dot-plot data regarding size and color code of the dots

• In Figure 2D, we wonder whether there could be RAGs that are already upregulated in ZT20 even without the SNC? On that figure the y-label states -log12FC, while the figure legend states -log2 Fold Change.

• In Table 2, we recommend to add indications for green colored fold changes and specify whether  are these log2(FC) or log10(FC)

• In Figure 3B the x-labels don’t match between the histogram and the legend. The so

Comments on reporting:

• As the number of biologically independent groups varies from 3 to 6 for different experiments for similar samples throughout this study, an explanation for these differences in discussion would be good.

Competing interests

The author declares that they have no competing interests.