A genetic compensatory mechanism regulated by Jun and Mef2d modulates the expression of distinct class IIa Hdacs to ensure peripheral nerve myelination and repair
Curation statements for this article:-
Curated by eLife
Evaluation Summary:
This study is of interest to scientists working in the field of genetic control of glial cell differentiation, myelination and repair. The data are extensive, of high quality, support their main conclusions, and provide novel insights into regulation of genetic compensatory mechanisms. The presentation and interpretation of the data can be improved.
(This preprint has been reviewed by eLife. We include the public reviews from the reviewers here; the authors also receive private feedback with suggested changes to the manuscript. The reviewers remained anonymous to the authors.)
This article has been Reviewed by the following groups
Listed in
- Evaluated articles (eLife)
Abstract
The class IIa histone deacetylases (HDACs) have pivotal roles in the development of different tissues. Of this family, Schwann cells express Hdac4 , 5 , and 7 but not Hdac9 . Here, we show that a transcription factor regulated genetic compensatory mechanism within this family of proteins, blocks negative regulators of myelination ensuring peripheral nerve developmental myelination and remyelination after injury. Thus, when Hdac4 and 5 are knocked-out from Schwann cells in mice, a JUN-dependent mechanism induces the compensatory overexpression of Hdac7 permitting, although with a delay, the formation of the myelin sheath. When Hdac4 , 5 , and 7 are simultaneously removed, the myocyte-specific enhancer-factor d (MEF2D) binds to the promoter and induces the de novo expression of Hdac9 , and although several melanocytic lineage genes are misexpressed and Remak bundle structure is disrupted, myelination proceeds after a long delay. Thus, our data unveil a finely tuned compensatory mechanism within the class IIa Hdac family, coordinated by distinct transcription factors, that guarantees the ability of Schwann cells to myelinate during development and remyelinate after nerve injury.
Article activity feed
-
-
Author Response
Reviewer #1 (Public Review):
In the present manuscript, the authors investigate regulatory roles of class IIa histone deacetylases (HDACs) in Schwann cells on developmental myelination, as well as on myelin repair after acute nerve injury. The study directly builds on previous observations (Gomis-Coloma et al., 2018) where the authors have shown that the primary HDACs of Schwann cells, HDAC4 and HDAC5, have redundant functions and cause only a mild delay in myelination in a double knock out (dKO), suggesting compensatory mechanisms by other HDACs. In the present study the authors indeed show compensatory upregulation of HDAC7 in HDAC4/5 dKO. They furthermore show by ablating all three HDACs that, next to a induction of HDAC9 expression, myelination is further delayed and the architecture of Remak bundles even …
Author Response
Reviewer #1 (Public Review):
In the present manuscript, the authors investigate regulatory roles of class IIa histone deacetylases (HDACs) in Schwann cells on developmental myelination, as well as on myelin repair after acute nerve injury. The study directly builds on previous observations (Gomis-Coloma et al., 2018) where the authors have shown that the primary HDACs of Schwann cells, HDAC4 and HDAC5, have redundant functions and cause only a mild delay in myelination in a double knock out (dKO), suggesting compensatory mechanisms by other HDACs. In the present study the authors indeed show compensatory upregulation of HDAC7 in HDAC4/5 dKO. They furthermore show by ablating all three HDACs that, next to a induction of HDAC9 expression, myelination is further delayed and the architecture of Remak bundles even permanently altered. The authors provide high quality data employing a broad spectrum of methodology, including conditional mutagenesis in mice, electrophysiology, immunofluorescence, electron microscopy, RNAseq, ChIP, cell culture, qPCR and Western blotting to justify their hypothesis of a regulatory and compensatory role of HDACs in Schwann cells during development and regeneration. The physiological relevance of this compensatory network, however, is not intuitive. Better discussion and elaboration of central findings in triple KOs in comparison to single KOs (and vice versa) would strongly improve the manuscript.
In detail, the following points may improve the strength of the manuscript:
- With regard to the triple mutants (HDAC4,5 and 7) the authors present a data set from P2 to P21 and another at P60. Here, the manuscript would benefit from more comparable data sets for the respective timeline. E.g. the authors show an increased SC number at P21. What happens to these Schwann cells? Are they still present at P60? In line, the authors show that even in the triple mutants the expression of certain genes including cJun remains upregulated. How do the authors explain this upregulation? It would be helpful to know whether these genes remain upregulated in myelinating SC or whether persisting supernumerary SC are responsible for the expression of c-Jun and others at later timepoints (e.g. by IHC)?”
As is shown in Figure 5L, the total number of Schwann cells at P60 in the uninjured (UI) tKO nerve at P60 is slightly increased, although it doesn’t reach statistical significance in our analysis. Also, an increased area of p75Ngfr expression (a protein expressed by non-myelin forming Schwann cells, but downregulated in myelinating Schwann cells) in MPZ negative cells was consistently observed in the tKO nerves (Figure 3J). Together these data suggest that the existence of supernumerary non-myelin forming Schwann cells in the tKO nerve. To further explore this hypothesis, we have now performed IF co-localization studies of c-Jun expressing cells with MPZ (Figure 5-figure supplement 3B of the revised version of the manuscript) and found most of them are MPZ negative (arrowheads), supporting the idea of the existence of supernumerary Schwann cells at the Remak bundles. Nevertheless, we found some of them are MPZ positive as well (arrows). It is worth mentioning it has been previously shown that myelinating Schwann cells (in the c-Jun OE mice) can express moderate levels of c-Jun (doi: 10.1523/JNEUROSCI.0986-17.2017). Thus, our data suggest that although c-Jun is increased mainly by the existence of supernumerary Schwann cells at the Remak bundles, part of it comes from increased expression by myelinating Schwann cells.
“2) An important point is the description of the Remak- SC phenotype, which, in contrast to the only transient myelination phenotype, seems to persist in triple mutants. The authors suggest a defect of axonal segregation independent of a sorting defect and link this to a ectopic expression of genes of the melanocytic lineage. Given the importance of the Remak phenotype, a more detailed elaboration of this aspect also in dKO and cKO would be a strong benefit for the manuscript. In addition, the proposed ectopic expression of the melanocytic lineage genes would profit from a more extensive discussion and description with regard to their potential (transient) expression in wildtype Schwann cells and their functional relevance in relation to the observed Remak SC pathology. Moreover, the EM image in figure 2E suggests not only an increased number but also size of axons in the Remak bundles of triple mutants, in contrast to the respective quantification. As this point is crucial with regard to a potential sorting defect, the authors should carefully reevaluate the discrepancy between the presented image and data.”
We didn’t observe gross abnormalities in the structure of the Remak bundles in the single cKOs neither the dKO. However, and following the suggestion of the reviewer, we have now analyzed in more detail and quantified the Remak phenotype in these mice. As is shown in Figure 2-figure supplement 1 A of the revised manuscript, no major changes can be observed in these genotypes. Thus, the segregation defects of small size axons are found exclusively in the tKO.
Also, we have now discussed more extensively the expression of melanocytic markers and their putative role on the Remak phenotype in the results and discussion sections.
In both control and tKO, we occasionally observe axons larger than 1 µm in diameter, however, a detailed quantification showed us that axon size distribution is essentially identical in both genotypes (Figure 2E). We apologize for selecting a non-representative EM image of the Remak bundles of the tKO. We have now substituted this image in the revised manuscript by another one more representative of the quantification. We hope the reviewer will find it adequate.
“3) Regarding the expression changes of HDAC7 and HDAC9 in mutant mice: The authors only show HDAC7 expression at P60, while the proposed role of HDAC7 concerns early postnatal development. Could the authors comment on the expression of HDAC7 at earlier timepoints?
Furthermore, within the manuscript, the authors suggest a "de novo" expression of HDAC9 in triple mutants. However, the authors show a small, but significant upregulation of HDAC9 already in single cKO4 nerves (Fig S1A) as well as in single cKO7 mice (Fig. 9A), hence a more careful usage of the term "de novo" may be advisable.”
Following the suggestions of the reviewer, we have now included HDAC7 expression in the nerves of the dKO mice at P2 and P8. As is shown in Figure 1-figure supplement 1C of the revised manuscript, it is increased in both cases, although no so much as in the P60.
When we use the “de novo” term we refer that its expression is new when compared to the controls (cKO5) and wild types. We apologize for not being clear enough in the original manuscript. We have now revised the text and restricted and explained more carefully the meaning we pretend to communicate with the use of the term “de novo”.
“4) In general, the discussion of the single HDAC knockout mutants is sometimes too sparse. This applies especially to the description of the cKO4 mice, which show a number of, albeit subtle, important differences with regard e.g. to the number of unmyelinated axons at P2 and P8 as well as with regard to the number of Schwann cell nuclei. However, the authors conclude that the single KO does not show a prominent phenotype. Though, given the compensatory mechanisms between HDACs in SC and the fact that the double HDAC4 (in SC) and HDAC5 (global) knockout display a similar phenotype to single HDAC4 mutants, this point requires more discussion. This dKO dataset, however, is redundant to the previously published study by the authors (Gomis-Coloma et al., 2018)”.
We found no changes in the cKO7 and only a slight increase in the number of unmyelinated axons in the cKO5 at P8, that fades out when it is calculated as percentage (Figure1-figure supplements 4 and 5 in the revised manuscript).As pointed out by the reviewer, there are subtle although consistent differences in several myelination parameters in the cKO4 (Figure1-figure supplement 2). However, these differences are not as big as in the dKO (Figure1-figure supplement 5). Thus, whereas the cKO4 has a decrease in the 15 % of myelinated axons at P2 when compared with the wildtype, the dKO shows a more prominent decrease (24,5% compared to its control).
As we have used a large number of genotypes in the manuscript, we have tried to focus in those genotypes with more prominent phenotype, namely the dKO and the tKO. However, we agree with the reviewer that the existence of a subtle phenotype of the cKO4 deserves to be more clearly stated in the text. For this reason, we have rewritten the results and discussion sections in the revised version of the manuscript to highlight this phenotype more clearly.
On the other hand, we also agree with the reviewer that part of the dKO presented as supplementary data overlaps with our previous data (Gomis-Coloma et al., 2018). We repeated these experiments, in parallel with the other genotypes, and performed a slightly different quantification to be able to compare more accurately its phenotype with the phenotypes of the other genotypes used in the current study. We hope the reviewer will be able to see it more as a positive thing than a criticism.
“5) The authors then tested the mutants after injury. The presentation of data from these experiments, however, is a bit confusing as it is going back and forth between nerve crush and cut, different mutants (cKO4, KO5, dKO, tKO) and time points of analysis (10dpi, 20dpi, 21dpi, 30dpi). All mutants show a decreased remyelination after crush, the dKO and tKO further present increased c-Jun mRNA and protein at 10dpi and reduction of Krox20, Mbp, Mpz, Periaxin.”
We apologize we were not clear enough in the manuscript. We have tried to fix these problems and hope the reviewer will consider we have been clearer in the revised version of the manuscript.
As in the case of development, cKO7 did not show any significant change in remyelination after nerve crush injury (Figure 5-figure supplement 1). cKO5 showed minor and not consistent changes in remyelination after crush injury (Figure 4-figure supplement 2). Thus, for example, it shows a decrease in MBP mRNA but no changes in Prx nor MPZ (Figure 4-figure supplement 2M). Consequently, our conclusion is that they have no prominent phenotype for remyelination. However, and as the reviewer points out, cKO4 shows subtle but consistent delay in different remyelination parameters (Figure 4-figure supplement 1). We have now tried to highlight this fact in the revised manuscript.
“The sequencing results are said to be obtained after nerve injury, however, it is not clear whether this was a cut or crush.”
In this experiment it was a crush injury of the sciatic nerve. Along the whole paper we used the crush model of injury to study the regeneration and remyelination of the nerve, as it produces axotomy but maintains intact the perineurium allowing the growth of axons through the distal stump and the regeneration of the nerve (doi: 10.7554/eLife.62232). We have used a cut model of injury only for the studies on myelin clearance and the activation of the Schwann cell repair phenotype, as it avoids the entrance of the axons from the proximal stump. We have now modified the Results and Material and methods sections to explain this in more detail. We hope the reviewer will find it clearer in the revised version of the manuscript.
“Four days after nerve cut in tKO, the authors report increased expression of genes typical for repair Schwann cells, as well as a more rapid myelin debris clearance, although it is unclear how this was measured. Only by quantifying the number of still intact myelin profiles early after injury as in figure 5A? If the authors would like to stress the point of myelin clearance, additional information on degeneration profiles and autophagy (LC3bI-II, p62 Western blots) or data on macrophage abundance is needed and would gain meaningful insight.”
The aim of this experiment was to know if the remyelination delay is caused by a problem in the activation of the repair phenotype and/or in myelin clearance. Our results clearly show that this is not the case, as they seem to work properly in the tKO. Strikingly, we observed that both are even faster than in controls.
We apologize for not being clear enough in the text with the method used for quantifying myelin clearance. We have now included a more detailed description of the protocol used in the Material and methods section of the revised version of the manuscript. We quantified both the number of intact myelin profiles and the amount of myelin protein zero (MPZ) to monitor the elimination of myelin debris in the distal stumps. As suggested by the reviewer, we have now also quantified autophagy by measuring LC3bI-II by WB and found no changes in the tKO. Also, no changes in other autophagy markers were found by RTqPCR (Figure 5-figure supplement 2 G and H of the revised manuscript). After the suggestion of the reviewer we have also quantified the number of macrophages in the distal stumps at 4d after cut and found no changes between tKO and control (Figure 5-figure supplement 2I).Thus, the consistently increased myelin clearance found in the tKO mice is not caused by accelerated autophagy/myelinophagy neither increased numbers of macrophages. Although we don’t know the mechanism, we now entertaining the possibility that accelerated axonal degeneration induced by signaling molecules derived from tKO Schwan cells could underlie this phenomenon. Future studies are therefore necessary to address this point
“6) Mechanistically, the authors investigated the genes that respond to HDACs or to which HDACs bind. It is nicely shown that HDAC4 can bind the c-Jun promoter, thereby repressing its expression, but also to the TSS of Mcam, belonging to the melanocyte lineage. However, a potential role of this finding is not further clarified.
We have now discussed more extensively the putative role of class IIa HDACs in repressing the expression of Mcam and melanocytic lineage genes in SCP in the revised manuscript. We hope the reviewer will find it clearer.
“In addition, the generalized conclusion that "class IIa HDACs bind to and repress the expression of melanocyte lineage genes and negative regulators of myelination allowing myelination and remyelination proceed in a timely fashion" may be revised, considering that only HDAC4 has been tested”.
We apologize for this generalization. We have now removed this sentence and modified the text avoiding unnecessary generalizations in the revised manuscript.
“On the other side, it is nicely shown that c-Jun can bind to the HDAC7 promoter, inducing its expression. This is well analyzed both in vitro and in vivo using conditional c-Jun gain and loss of function in SC development. Here, although ectopic c-Jun overexpression in mice artificially increases HDAC7 expression in development, adding a more (patho-)physiological relevant experiment using c-Jun cKO in a nerve injury paradigm would be an asset.”
We agree with the reviewer that it will be very interesting to explore what happens with the induction of HDAC7 in a c-Jun cKO background. To this aim, we generated the dKO; c-Jun cKO genotype and measured HDAC7 gene expression in the sciatic nerves of these mice. As is shown in Figure 8G of the revised manuscript, the absence of c-Jun in Schwann cells totally prevents the compensatory overexpression of HDAC7 at P8 and P60.Interestingly we have also found an increase in the expression of HDAC9 in these mice, probably to compensate their incapacity to upregulate HDAC7. We believe that this, together with our previous data, strongly supports the view that c-Jun regulates the compensatory expression of HDAC7, and hope it will fully convince the reviewer.
“7) The final hypothesis from the authors is, that upon lack of the functionally redundant HDAC4/5 and the concomitant de-repression of c-Jun, HDAC7 is upregulated upon binding of c-Jun to compensate for the loss and ensure myelination, although delayed. If HDAC7 is also lost, Mef2d expression increases and induces "de novo" expression of HDAC9. The data presented in the manuscript indeed provide evidence of a role for HDAC4, HDAC5 and HDAC7 in developmental myelination and nerve repair with compensatory potential for each other. However, the physiological relevance of this compensatory functions is, although interesting, not quite clear and the manuscript may profit from a discussion of this point.”
We have discussed the potential physiological role of genetic compensation in the first paragraph of the discussion section (page 14). We apologize if we have not been clear enough. We have included now a more extensive discussion on the putative physiological relevance of gene compensation in myelination in the revised version of the manuscript. Fluctuations in gene expression (noise) is a well-known phenomenon that has been described from bacteria to mammalian cells, and may have dramatic effects on fitness if they persist long enough (DOI: 10.1126/science.1105891). Kafri et al 2006 (doi.org/10.1073/pnas.0604883103) suggested that gene redundancy has been evolutionarily selected because it can reduce the harmful effects of gene expression noise. On this basis, we speculate in the revised manuscript, that the genetic compensation could avoid fluctuations in the gene dose of class IIa HDAC in Schwann cells consequence of gene expression noise, allowing differentiation and the proper myelination of the peripheral nervous system. We thank the reviewer for helping us to see that a better clarification of this point was needed.
"Reviewer #2 (Public Review):
The classIIa Histone De-Acetylases (HDAC) play important roles in the transcriptional control of differentiation of a wide range of cell types. This class of HDACs is regulated by different signalling pathways and it involves the shuttling of the protein into the nucleus. Indeed, previous work from this lab has demonstrated that increased levels of cAMP shuttles HDAC4 into the nucleus of Schwann cells where it recruits NcoR1/HDAC3 to repress c-Jun expression and allows commencement of a myelin-related gene expression program. Thus, HDAC4 links cAMP signalling to repression of a 'repressor' to stimulate cell differentiation. However, genetic deletion of HDAC4 (or HDAC5 and HDAC4/HDAC5) does not have a significant effect on Schwann cell differentiation and myelination in vivo, suggesting that other compensatory mechanisms might exist.
Building upon their previous work, Velasco-Aviles and colleagues now demonstrate the existence of a genetic compensatory mechanism that relies on functional redundancies among the ClassIIa HDACs and the transcription factors c-Jun and Mef2d.
Using genetic ablation of multiple HDAC genes, extensive morphological analysis of developing and regenerating nerves combined with gene expression analysis, provide a description of the gene regulatory mechanisms that maintain adequate levels of ClassIIa HDACs required for peripheral nerve development and repair. Their data are of high quality and support their major finding.
One interesting finding is that in the tKO, in which myelination eventually appears to progress normally, Remak Schwann cells are deficient in segregating lower calibre axons into cytoplasmic cuffs (Figure 2E). The authors interpret this a segregation defect and not as a sorting defect (page 5). Now, it is difficult to see how these two cellular mechanisms can be distinguished or whether they are different mechanisms to begin with. Notably, the unsorted bundle of axons presented in Figure2E also contains larger calibre axons that should normally be myelinated. Therefore, a simpler interpretation is that tKO Schwann cells are moderately impaired in axonal segregation, which results in the failure to sort out the occasional larger calibre axons from bundles and ensheathment of the smaller calibre axons into mature Remak bundles.”
We apologize for selecting a non-representative image of the Remak bundles in the tKO. As it has been explained before in the response to the reviewer#1, when the Remak bundle phenotype was quantified we observed no changes in the number of axons bigger than 1 µm in diameter (see Figure 2E). We have now changed the image presented by another one more representative of the phenotype.
“There is no justification for proposing a 'segregation' mechanism different from the 'sorting' mechanism. As the sorting process critically depends on the elaboration of a basement membrane, it would be of interest to have a closer look at the basement membrane in EM and by IF in nerve sections and maybe WB. Is there any evidence for reduced laminin/collagen (or their receptors) expression in tKO nerves?”
We have now removed the proposition of a defect in segregation, in contraposition of a sorting defect, in the revised version of the manuscript. However, whatever the mechanism involved, it is worth mentioning that we couldn´t observe morphological differences in the basement membranes in the tKO by EM. Also, our RNAseq data shows that collagens, laminin and integrin receptors and other molecules associated with axon sorting and segregation (doi: 10.1177/1073858415572361) are not decreased in the tKO.
“It is argued throughout the manuscript that classIIa HDACs are involved in the repression of repressors of myelination. It is stated that in injured nerves a strong upregulation of such negative regulators of developmental myelination is observed (page 17). Regulators such as c-Jun, Runx2, Sox2 etcetera. To avoid confusion, it is important to clearly distinguish between developmental and repair functions (exemplified by c-Jun) and in Schwann cells cultured in the absence of axonal contact.”
We apologize for not being clear enough. We have now revised the text and changed it to distinguish more clearly between developmental and repair functions. We hope the reviewer will find it less confusing in the revised manuscript.
“Confusingly and erroneously, it is also stated that the POU domain transcription factor Oct6 blocks the transition from promyelinating Schwann cell into myelinating cells. The quoted paper does not support this idea at all. On the contrary, it demonstrates that Oct6 expression is required for the progression of promyelinating cells into fully myelinating cells.”
We agree with the reviewer in that this reference elegantly demonstrates that Oct6 is required for the progression of promyelinating cells into fully myelinating cells. However, it is worth noting that it has been also demonstrated that Oct6 needs to be properly downregulated in a timely fashion to allow myelination (Ryu et al 2007; DOI: 10.1523/JNEUROSCI.5497-06.2007). Thus, the upregulation of Oct-6 and its posterior downregulation are both necessary to permit peripheral nerves myelination. We apologize for forgetting to include this reference in the manuscript. We have now fixed it in the revised version of the manuscript. We hope the reviewer will find now coherent our arguments.
-
Evaluation Summary:
This study is of interest to scientists working in the field of genetic control of glial cell differentiation, myelination and repair. The data are extensive, of high quality, support their main conclusions, and provide novel insights into regulation of genetic compensatory mechanisms. The presentation and interpretation of the data can be improved.
(This preprint has been reviewed by eLife. We include the public reviews from the reviewers here; the authors also receive private feedback with suggested changes to the manuscript. The reviewers remained anonymous to the authors.)
-
Reviewer #1 (Public Review):
In the present manuscript, the authors investigate regulatory roles of class IIa histone deacetylases (HDACs) in Schwann cells on developmental myelination, as well as on myelin repair after acute nerve injury. The study directly builds on previous observations (Gomis-Coloma et al., 2018) where the authors have shown that the primary HDACs of Schwann cells, HDAC4 and HDAC5, have redundant functions and cause only a mild delay in myelination in a double knock out (dKO), suggesting compensatory mechanisms by other HDACs. In the present study the authors indeed show compensatory upregulation of HDAC7 in HDAC4/5 dKO. They furthermore show by ablating all three HDACs that, next to a induction of HDAC9 expression, myelination is further delayed and the architecture of Remak bundles even permanently altered. The …
Reviewer #1 (Public Review):
In the present manuscript, the authors investigate regulatory roles of class IIa histone deacetylases (HDACs) in Schwann cells on developmental myelination, as well as on myelin repair after acute nerve injury. The study directly builds on previous observations (Gomis-Coloma et al., 2018) where the authors have shown that the primary HDACs of Schwann cells, HDAC4 and HDAC5, have redundant functions and cause only a mild delay in myelination in a double knock out (dKO), suggesting compensatory mechanisms by other HDACs. In the present study the authors indeed show compensatory upregulation of HDAC7 in HDAC4/5 dKO. They furthermore show by ablating all three HDACs that, next to a induction of HDAC9 expression, myelination is further delayed and the architecture of Remak bundles even permanently altered. The authors provide high quality data employing a broad spectrum of methodology, including conditional mutagenesis in mice, electrophysiology, immunofluorescence, electron microscopy, RNAseq, ChIP, cell culture, qPCR and Western blotting to justify their hypothesis of a regulatory and compensatory role of HDACs in Schwann cells during development and regeneration. The physiological relevance of this compensatory network, however, is not intuitive. Better discussion and elaboration of central findings in triple KOs in comparison to single KOs (and vice versa) would strongly improve the manuscript.
In detail, the following points may improve the strength of the manuscript:
With regard to the triple mutants (HDAC4,5 and 7) the authors present a data set from P2 to P21 and another at P60. Here, the manuscript would benefit from more comparable data sets for the respective timeline. E.g. the authors show an increased SC number at P21. What happens to these Schwann cells? Are they still present at P60? In line, the authors show that even in the triple mutants the expression of certain genes including cJun remains upregulated. How do the authors explain this upregulation? It would be helpful to know whether these genes remain upregulated in myelinating SC or whether persisting supernumerary SC are responsible for the expressio of cJun and others at later timepoints (e.g. by IHC)?
An important point is the description of the Remak- SC phenotype, which, in contrast to the only transient myelination phenotype, seems to persist in triple mutants. The authors suggest a defect of axonal segregation independent of a sorting defect and link this to a ectopic expression of genes of the melanocytic lineage. Given the importance of the Remak phenotype, a more detailed elaboration of this aspect also in dKO and cKO would be a strong benefit for the manuscript. In addition, the proposed ectopic expression of the melanocytic lineage genes would profit from a more extensive discussion and description with regard to their potential (transient) expression in wildtype Schwann cells and their functional relevance in relation to the observed Remak SC pathology. Moreover, the EM image in figure 2E suggests not only an increased number but also size of axons in the Remak bundles of triple mutants, in contrast to the respective quantification. As this point is crucial with regard to a potential sorting defect, the authors should carefully reevaluate the discrepancy between the presented image and data.
Regarding the expression changes of HDAC7 and HDAC9 in mutant mice: The authors only show HDAC7 expression at P60, while the proposed role of HDAC7 concerns early postnatal development. Could the authors comment on the expression of HDAC7 at earlier timepoints?
Furthermore, within the manuscript, the authors suggest a "de novo" expression of HDAC9 in triple mutants. However, the authors show a small, but significant upregulation of HDAC9 already in single cKO4 nerves (Fig S1A) as well as in single cKO7 mice (Fig. 9A), hence a more careful usage of the term "de novo" may be advisable.In general, the discussion of the single HDAC knockout mutants is sometimes too sparse. This applies especially to the description of the cKO4 mice, which show a number of, albeit subtle, important differences with regard e.g. to the number of unmyelinated axons at P2 and P8 as well as with regard to the number of Schwann cell nuclei. However, the authors conclude that the single KO does not show a prominent phenotype. Though, given the compensatory mechanisms between HDACs in SC and the fact that the double HDAC4 (in SC) and HDAC5 (global) knockout display a similar phenotype to single HDAC4 mutants, this point requires more discussion. This dKO dataset, however, is redundant to the previously published study by the authors (Gomis-Coloma et al., 2018).
The authors then tested the mutants after injury. The presentation of data from these experiments, however, is a bit confusing as it is going back and forth between nerve crush and cut, different mutants (cKO4, KO5, dKO, tKO) and time points of analysis (10dpi, 20dpi, 21dpi, 30dpi). All mutants show a decreased remyelination after crush, the dKO and tKO further present increased c-Jun mRNA and protein at 10dpi and reduction of Krox20, Mbp, Mpz, Periaxin. The sequencing results are said to be obtained after nerve injury, however, it is not clear whether this was a cut or crush. Four days after nerve cut in tKO, the authors report increased expression of genes typical for repair Schwann cells, as well as a more rapid myelin debris clearance, although it is unclear how this was measured. Only by quantifying the number of still intact myelin profiles early after injury as in figure 5A? If the authors would like to stress the point of myelin clearance, additional information on degeneration profiles and autophagy (LC3bI-II, p62 Western blots) or data on macrophage abundance is needed and would gain meaningful insight.
Mechanistically, the authors investigated the genes that respond to HDACs or to which HDACs bind. It is nicely shown that HDAC4 can bind the c-Jun promoter, thereby repressing its expression, but also to the TSS of Mcam, belonging to the melanocyte lineage. However, a potential role of this finding is not further clarified. In addition, the generalized conclusion that "class IIa HDACs bind to and repress the expression of melanocyte lineage genes and negative regulators of myelination allowing myelination and remyelination proceed in a timely fashion" may be revised, considering that only HDAC4 has been tested. On the other side, it is nicely shown that c-Jun can bind to the HDAC7 promoter, inducing its expression. This is well analyzed both in vitro and in vivo using conditional c-Jun gain and loss of function in SC development. Here, although ectopic c-Jun overexpression in mice artificially increases HDAC7 expression in development, adding a more (patho-)physiological relevant experiment using c-Jun cKO in a nerve injury paradigm would be an asset.
The final hypothesis from the authors is, that upon lack of the functionally redundant HDAC4/5 and the concomitant de-repression of c-Jun, HDAC7 is upregulated upon binding of c-Jun to compensate for the loss and ensure myelination, although delayed. If HDAC7 is also lost, Mef2d expression increases and induces "de novo" expression of HDAC9. The data presented in the manuscript indeed provide evidence of a role for HDAC4, HDAC5 and HDAC7 in developmental myelination and nerve repair with compensatory potential for each other. However, the physiological relevance of this compensatory functions is, although interesting, not quite clear and the manuscript may profit from a discussion of this point.
-
Reviewer #2 (Public Review):
The classIIa Histone De-Acetylases (HDAC) play important roles in the transcriptional control of differentiation of a wide range of cell types. This class of HDACs is regulated by different signalling pathways and it involves the shuttling of the protein into the nucleus. Indeed, previous work from this lab has demonstrated that increased levels of cAMP shuttles HDAC4 into the nucleus of Schwann cells where it recruits NcoR1/HDAC3 to repress c-Jun expression and allows commencement of a myelin-related gene expression program. Thus, HDAC4 links cAMP signalling to repression of a 'repressor' to stimulate cell differentiation. However, genetic deletion of HDAC4 (or HDAC5 and HDAC4/HDAC5) does not have a significant effect on Schwann cell differentiation and myelination in vivo, suggesting that other …
Reviewer #2 (Public Review):
The classIIa Histone De-Acetylases (HDAC) play important roles in the transcriptional control of differentiation of a wide range of cell types. This class of HDACs is regulated by different signalling pathways and it involves the shuttling of the protein into the nucleus. Indeed, previous work from this lab has demonstrated that increased levels of cAMP shuttles HDAC4 into the nucleus of Schwann cells where it recruits NcoR1/HDAC3 to repress c-Jun expression and allows commencement of a myelin-related gene expression program. Thus, HDAC4 links cAMP signalling to repression of a 'repressor' to stimulate cell differentiation. However, genetic deletion of HDAC4 (or HDAC5 and HDAC4/HDAC5) does not have a significant effect on Schwann cell differentiation and myelination in vivo, suggesting that other compensatory mechanisms might exist.
Building upon their previous work, Velasco-Aviles and colleagues now demonstrate the existence of a genetic compensatory mechanism that relies on functional redundancies among the ClassIIa HDACs and the transcription factors c-Jun and Mef2d.
Using genetic ablation of multiple HDAC genes, extensive morphological analysis of developing and regenerating nerves combined with gene expression analysis, provide a description of the gene regulatory mechanisms that maintain adequate levels of ClassIIa HDACs required for peripheral nerve development and repair.
Their data are of high quality and support their major finding.One interesting finding is that in the tKO, in which myelination eventually appears to progress normally, Remak Schwann cells are deficient in segregating lower calibre axons into cytoplasmic cuffs (Figure 2E). The authors interpret this a segregation defect and not as a sorting defect (page 5). Now, it is difficult to see how these two cellular mechanisms can be distinguished or whether they are different mechanisms to begin with. Notably, the unsorted bundle of axons presented in Figure2E also contains larger calibre axons that should normally be myelinated. Therefore, a simpler interpretation is that tKO Schwann cells are moderately impaired in axonal segregation, which results in the failure to sort out the occasional larger calibre axons from bundles and ensheathment of the smaller calibre axons into mature Remak bundles. There is no justification for proposing a 'segregation' mechanism different from the 'sorting' mechanism. As the sorting process critically depends on the elaboration of a basement membrane, it would be of interest to have a closer look at the basement membrane in EM and by IF in nerve sections and maybe WB. Is there any evidence for reduced laminin/collagen (or their receptors) expression in tKO nerves?
It is argued throughout the manuscript that classIIa HDACs are involved in the repression of repressors of myelination. It is stated that in injured nerves a strong upregulation of such negative regulators of developmental myelination is observed (page 17). Regulators such as c-Jun, Runx2, Sox2 etcetera. To avoid confusion, it is important to clearly distinguish between developmental and repair functions (exemplified by c-Jun) and in Schwann cells cultured in the absence of axonal contact. Confusingly and erroneously, it is also stated that the POU domain transcription factor Oct6 blocks the transition from promyelinating Schwann cell into myelinating cells. The quoted paper does not support this idea at all. On the contrary, it demonstrates that Oct6 expression is required for the progression of promyelinating cells into fully myelinating cells.
-
