CXCL12α/SDF‐1 from perisynaptic Schwann cells promotes regeneration of injured motor axon terminals

Abstract The neuromuscular junction has retained through evolution the capacity to regenerate after damage, but little is known on the inter‐cellular signals involved in its functional recovery from trauma, autoimmune attacks, or neurotoxins. We report here that CXCL12α, also abbreviated as stromal‐derived factor‐1 (SDF‐1), is produced specifically by perisynaptic Schwann cells following motor axon terminal degeneration induced by α‐latrotoxin. CXCL12α acts via binding to the neuronal CXCR4 receptor. A CXCL12α‐neutralizing antibody or a specific CXCR4 inhibitor strongly delays recovery from motor neuron degeneration in vivo. Recombinant CXCL12α in vivo accelerates neurotransmission rescue upon damage and very effectively stimulates the axon growth of spinal cord motor neurons in vitro. These findings indicate that the CXCL12α‐CXCR4 axis plays an important role in the regeneration of the neuromuscular junction after motor axon injury. The present results have important implications in the effort to find therapeutics and protocols to improve recovery of function after different forms of motor axon terminal damage.


Transaction Report:
(Note: With the exception of the correction of typographical or spelling errors that could be a source of ambiguity, letters and reports are not edited. The original formatting of letters and referee reports may not be reflected in this compilation.) Thank you for the submission of your manuscript to EMBO Molecular Medicine. I am sorry that it has taken so long to get back to you on your manuscript.

Editor: Céline Carret
While we are still missing one referee report, given that the two evaluations we have already are consistent, and a further delay cannot be justified, I have decided to proceed based on these evaluations.
You will see that both referees are rather positive about the study and their concerns are largely overlapping. Some shortcomings are noted and suggestions proposed that if satisfactorily addressed would improve the conclusiveness of the findings and provide additional clinical relevance to the data.
Given these evaluations, I would like to give you the opportunity to revise your manuscript, with the understanding that the referees' concerns must be fully addressed and that acceptance of the manuscript would entail a second round of review. Please note that it is EMBO Molecular Medicine policy to allow only a single round of revision and that, as acceptance or rejection of the manuscript will depend on another round of review, your responses should be as complete as possible.
Negro et al identified perisynaptic Schwann cell-derived CXCL12 and its neuronal receptor CXCR4 as being required for motor axon regeneration after a degenerating insult, and that these findings hold potential therapeutic value. This reviewer commends the authors for their good image quality, appropriate numbers of replicates, and blinded observers/experimenters to increase confidence in the results. However, some questions remain about the significance of the work and the following revisions are suggested to increase the impact and confidence in the findings.
Major revisions For figure 1D, qRT-PCR would be a more convincing means of determining changes in mRNA levels and offer an alternate means of validating the method that identified CXCL12 using independent biological samples. Also, the p value in the legend does not describe which time points are being compared. Figure 1E,F and EV1 aim to convince the reader that CXCL12 is expressed in PSCs by 3D reconstruction and colocalization of CXCL12 and lysotracker. While it's reasonably convincing that that CXCL12 protein is in acidic vesicles in PSCs, it is not clear what the significance of this is. The rationale for using lysotracker was not clearly stated. One interpretation is that PSCs are endocytosing CXCL12. The authors should perform in situ hybridization (FISH) to conclude that CXCL12 is in fact expressed in PSCs. Quantification of % of NMJs expressing CXCL12 and/or # of NMJs analyzed would be appropriate.
Assuming that CXCL12 expression is in fact increased in PSCs, is it limited to PSCs or is it expressed in non-terminal Schwann cells also? qPCR on nerve would be an appropriate control.
The EFP data in figure 2 shows CXCL12 is required for behavioral recovery and the authors correlate this with in vitro data showing CXCL12 as a growth factor for motor axons. The authors should also assess anatomical recovery of NMJs with their markers in figure 1A,B in the presence and absence of CXCL12 antibody injection to determine if fewer NMJs are reoccupied as a resulting of CXCL12 inhibition.
The authors show that CXCL12/CXCR4 is necessary for efficient recovery of EJPs, but do not show sufficiency which is the main determination of therapeutic potential. An experiment testing in vivo regeneration in response to recombinant CXCL12 by intramuscular or intrathecal injection would increase the significance of the results. Figure 3D,E aim to show that CXCR4 is expressed in neurons and not PSCs. In 3D, a counter stain with B3 tubulin to show axon tips is necessary. Also, why are there three images? In 3E, it is unclear whether CXCR4 is expressed in PSCs or axons as there is expression in domains of overlapping green (SCs-GFP) and axons in red (Smi81, although what SMI81 stains is not defined). Maybe in situ hybridization is a better option to determine cell specific expression. Figure 1B panels labeled for Snap25 should be labeled as Vamp1. Figure 1 sample size is not recorded in the figure legend as in other figures. For figure 1A,B, the qualitative nature of the data is very striking, but quantification of presynaptic marker recovery would be useful to determine if variability exists and to what extent across neuromuscular junctions within the muscle. Figure 1C may be kept but is somewhat unnecessary as an understanding of the experimental paradigm comes naturally from figure 1A,B. In 3E, SMI81 should be Snap25 to be consistent with Figure 1.

Minor points
Referee #2 (Comments on Novelty/Model System): I believe that the adopted model is fine. It would have been great to compare the poisoning done by treating the NMJ with alpha latrotoxin with another agent (e.g. taipoxin) acting via a completely different mechanism of action Referee #2 (Remarks): This manuscript by Negro and colleagues describes an interesting transcriptomic approach aimed to the identification of novel signals involved in the functional recovery of the neuromuscular junction (NMJ) upon damage. The chosen insult is the treatment of the NMJ with alpha-latrotoxin, a presynaptic poison that elicits a long lasting inhibition of this specialized synaptic terminal. The authors found that stromal derived factor 1 (SDF-1 or CXCL12 alpha) is released following NMJ degeneration by schwann cells and promote regeneration via the activation of its receptor CXCR4. Sequestration of SDF-1 or inhibition of its receptor halts the functional recovery of the NMJ, suggesting that this signal transduction pathway is pivotal in NMJ regeneration. The results are presented in a clear, succinct manner and in logical order, which makes the manuscript easy to read. In general, the claims are supported by the figures provided. However, specific aspects of this manuscript should revised prior to publication: 1. In contrast to the statement made in the main text and caption, both red channels in Figure 1A and B are labelled with SNAP25, rather than with SNAP25 and VAMP1. This should be amended. The specific treatment should be added to panel B. it would be useful to quantify the loss of the presynaptic markers using alpha-bungarotoxin as a mask. This analysis should address whether both plasma membrane and synaptic vesicle markers are depleted in the same extent. 2. The scheme in Figure 1C gives the wrong impression that samples are collected and/or imaged in a longitudinal rather than at fixed time points. This is misleading, especially for a scheme that should have the opposite goal. I suggest removing it. 3. No specific data emerging from the transcriptomic analysis have been provided (e.g. heat map) or if the raw data are going to be deposited in a suitable public database (e.g. Array Express at http://www.ebi.ac.uk/arrayexpress). 4. The staining of CXCL12 alpha is poorly visible and should be represented in white in the central panels to increase the contrast. The authors should comment on CXCL12 alpha distribution since its localization is unclear: some puncta seem to be present within the schwann cells, whereas some others map in areas not labeled in green (top in the alpha-latrotoxin 4 h sample). Likewise the statement that CXCL12 alpha puncta are colocalized with Lysotracker should be tuned down, since this colocalization is partial at best ( Figure EV1). Figure EV2 seems also redundant. 5. In Figure 2A, are control samples treated with alpha-latrotoxin and either with saline or an equivalent dose of an irrelevant antibody? Please clarify the methodology in the main text. 6. Figure 2D is not very informative, since the more informative quantification of these data is provided in panel E. Does treatment with CXCL12 alpha alter neuronal survival in the conditions described in this Figure? Furthermore, is the effect on axonal elongation detected if CXCL12 alpha is only added in the somatic chamber? 7. The resolution of Figure 3D is not optimal. 8. Does CXCR4 co-localize in motoneurons with classical growth cone markers? Why is SMI81 used in Figure 3E instead of SNAP25?
1st Revision -authors' response 28 March 2017 Referee #1 (Comments on Novelty/Model System): The model is appropriate and allowed to identify and then test the requirement of CXCL12/CXCR4 pathway for efficient recovery of EJPs. However, the model was not used to show sufficiency, i.e is administration of CXCL12 accelerating recovery. If another model is more appropriate for such experiments, the authors should investigate these options. This would be important to support the medical relevance to improve recovery of function after different forms of motor axon terminal damage.
Referee #1 (Remarks): Negro et al identified perisynaptic Schwann cell-derived CXCL12 and its neuronal receptor CXCR4 as being required for motor axon regeneration after a degenerating insult, and that these findings hold potential therapeutic value. This reviewer commends the authors for their good image quality, appropriate numbers of replicates, and blinded observers/experimenters to increase confidence in the results. However, some questions remain about the significance of the work and the following revisions are suggested to increase the impact and confidence in the findings.

The experiment showing that administration of rCXCL12α in mice accelerates NMJ recovery of function after nerve terminal degeneration is now reported in the revised paper. A positive effect obtained by simple injections of the small chemokine is indeed very remarkable, as the molecule is
expected to diffuse away rather rapidly. We are currently trying to devise means of achieving a slow release at the injected site.
Major revisions For figure 1D, qRT-PCR would be a more convincing means of determining changes in mRNA levels and offer an alternate means of validating the method that identified CXCL12 using independent biological samples.
We performed droplet digital PCR on cDNA from soleus NMJs treated with the toxin and collected by laser microdissection after 4, 16, 72 hours from toxin injection. We used soleus muscle in order to extend the validity of our transcriptome analysis performed on LAL to a different mice muscle that provides a better electrophysiological response. We compared the two muscles using immunofluorescence analysis of appropriate presynaptic markers and found that the time courses of their nerve terminal degeneration and regeneration induced by toxin injection are very similar. Digital PCR results are reported in panel D of Figure 1, and the kinetics of degeneration and regeneration in soleus muscle in Figure EV2 of the revised manuscript. Digital PCR data show the up-regulation of CXCL12α mRNA at 4 hours (fully degenerated NMJs in both muscles) with respect to controls, similarly to what obtained previously on LAL muscle by trascriptomics.
Also, the p value in the legend does not describe which time points are being compared. Figure 1 legend of the revised manuscript. Figure 1E,F and EV1 aim to convince the reader that CXCL12 is expressed in PSCs by 3D reconstruction and colocalization of CXCL12 and lysotracker. While it's reasonably convincing that that CXCL12 protein is in acidic vesicles in PSCs, it is not clear what the significance of this is. The rationale for using lysotracker was not clearly stated. One interpretation is that PSCs are endocytosing CXCL12.

The p value of panel C refers to control vs 4 hours samples. This is now specified in
Given the CXCL12α-positive spots observed in PSCs at degenerating NMJs, we wished to understand the nature of such vesicles. By Lysotracker staining we showed that the chemokine is accumulated in acidic structures, whose complete characterization is the goal of a future study. Accordingly, and following the Reviewer's comment, we deleted the former Figure EV1 from the revised manuscript.
The authors should perform in situ hybridization (FISH) to conclude that CXCL12 is in fact expressed in PSCs. Figure 1E of the revised manuscript, which confirm that CXCL12α mRNA is specifically expressed in PSCs during neurodegeneration.

Representative images of in situ hybridization at the NMJ are reported in
Quantification of % of NMJs expressing CXCL12 and/or # of NMJs analyzed would be appropriate.
The percentage of CXCL12α-positive NMJs (56 ± 5 %) is now reported in the Results section of the revised manuscript.
Assuming that CXCL12 expression is in fact increased in PSCs, is it limited to PSCs or is it expressed in non-terminal Schwann cells also? qPCR on nerve would be an appropriate control. The EFP data in figure 2 shows CXCL12 is required for behavioral recovery and the authors correlate this with in vitro data showing CXCL12 as a growth factor for motor axons. The authors should also assess anatomical recovery of NMJs with their markers in figure 1A,B in the presence and absence of CXCL12 antibody injection to determine if fewer NMJs are reoccupied as a resulting of CXCL12 inhibition.
We quantified the number of VAMP1-positive NMJs in soleus muscles 72 hours from toxin injection (regeneration under way), and compared it with muscles treated with both the anti-CXCL12α antibody and the toxin. Confocal images and relative quantification are reported in Figure EV3. We found that the delay in neurotransmission recovery measured by electrophysiology due to the action of CXCL12α neutralizing antibodies is paralleled by a reduced number of regenerating NMJs.
The authors show that CXCL12/CXCR4 is necessary for efficient recovery of EJPs, but do not show sufficiency which is the main determination of therapeutic potential. An experiment testing in vivo regeneration in response to recombinant CXCL12 by intramuscular or intrathecal injection would increase the significance of the results.
We performed EJP recordings on soleus muscles locally injected with the toxin with or without subsequent intramuscular administrations of rCXCL12α. A faster recovery of NMJ functionality was observed in the former condition. Results are reported in panel G of Figure 2 of the revised manuscript and in the Results section. This result is remarkable as CXCL12α is a small molecule that is expected to be washed out rapidly after injection. Therefore only those molecules that can bind their targets within a restricted time window, defined by the lymph flow rate in the muscle, will be able to exert their action at the NMJ. Moreover, it is believed that multiple factors are involved in recovery. We are currently trying to find a way of achieving a slow release of CXCL12α after injection. Figure 3D,E aim to show that CXCR4 is expressed in neurons and not PSCs. In 3D, a counter stain with B3 tubulin to show axon tips is necessary. Figure  3 of the revised manuscript. At toxin-treated LAL NMJs with GFP-SCs, CXCR4 becomes expressed in the motor axon stump, where it co-localizes with neurofilaments (please see the novel confocal images reported in Figure 3C of the revised manuscript).

The counterstain of CXCR4 with β 3 -tubulin in cultured SCMNs is now reported in panel A of
Also, why are there three images?
The three images simply aimed at showing different fields. The redundant images have been removed.
In 3E, it is unclear whether CXCR4 is expressed in PSCs or axons as there is expression in domains of overlapping green (SCs-GFP) and axons in red (Smi81, although what SMI81 stains is not defined). Maybe in situ hybridization is a better option to determine cell specific expression. Figure 3).

SMI81 refers to the monoclonal anti-SNAP 25 primary antibody used in the study. We apologize for the omission.
Minor points Figure 1B panels labeled for Snap25 should be labeled as Vamp1.
For figure 1A,B, the qualitative nature of the data is very striking, but quantification of presynaptic marker recovery would be useful to determine if variability exists and to what extent across neuromuscular junctions within the muscle.

Quantification of presynaptic markers recovery in LAL muscles exposed to α-LTx was reported in a recent publication of this lab (please refer to Duregotti et al., PNAS 2015).
Figure 1C may be kept but is somewhat unnecessary as an understanding of the experimental paradigm comes naturally from figure 1A,B.
Panel C has been removed from Figure 1, modified and reported as Figure EV1 in the revised manuscript.
In 3E, SMI81 should be Snap25 to be consistent with Figure 1.

Amended.
Referee #2 (Comments on Novelty/Model System): I believe that the adopted model is fine. It would have been great to compare the poisoning done by treating the NMJ with alpha latrotoxin with another agent (e.g. taipoxin) acting via a completely different mechanism of action.
We agree with the Referee. Indeed, the two types of presynaptic neurotoxins act via a completely different biochemical mechanism of action. However, we and others have shown before that, eventually, the common toxic event that leads to nerve terminal degeneration is the entry of an excessive amount of Ca 2+ . Therefore there is no reason to assume that the two types of neurotoxins will lead to a different regeneration program.

Referee #2 (Remarks):
This manuscript by Negro and colleagues describes an interesting transcriptomic approach aimed to the identification of novel signals involved in the functional recovery of the neuromuscular junction (NMJ) upon damage. The chosen insult is the treatment of the NMJ with alpha-latrotoxin, a presynaptic poison that elicits a long lasting inhibition of this specialized synaptic terminal. The authors found that stromal derived factor 1 (SDF-1 or CXCL12 alpha) is released following NMJ degeneration by schwann cells and promote regeneration via the activation of its receptor CXCR4. Sequestration of SDF-1 or inhibition of its receptor halts the functional recovery of the NMJ, suggesting that this signal transduction pathway is pivotal in NMJ regeneration. The results are presented in a clear, succinct manner and in logical order, which makes the manuscript easy to read. In general, the claims are supported by the figures provided. However, specific aspects of this manuscript should revised prior to publication: 1. In contrast to the statement made in the main text and caption, both red channels in Figure 1A and B are labelled with SNAP25, rather than with SNAP25 and VAMP1. This should be amended.
We apologize for this mislabeling. The correct label is reported in the revised paper.
The specific treatment should be added to panel B.

Done
It would be useful to quantify the loss of the presynaptic markers using alpha-bungarotoxin as a mask. This analysis should address whether both plasma membrane and synaptic vesicle markers are depleted in the same extent. 2. The scheme in Figure 1C gives the wrong impression that samples are collected and/or imaged in a longitudinal rather than at fixed time points. This is misleading, especially for a scheme that should have the opposite goal. I suggest removing it.
The scheme has been removed from Figure 1 and a modified version is now presented as Figure  EV1 of the revised manuscript.
3. No specific data emerging from the transcriptomic analysis have been provided (e.g. heat map) or if the raw data are going to be deposited in a suitable public database (e.g. Array Express at http://www.ebi.ac.uk/arrayexpress).
We are currently running an additional transcriptome analysis experimentally similar to the one that led to the identification of CXCL12α as a possible hit, but performed with Illumina Technology, rather than Ion Torrent. Once completed, we will perform a broad bioinformatic analysis to identify activated and depressed pathways, and we will deliver the entire set of transcriptomic data to a public data base in the context of a novel submitted paper.

In the present study we singled out CXCL12α hit (whose mRNA increases during neurodegeration to return to basal levels when regeneration is almost complete) as a potential candidate in motor axon terminal regeneration given its role in the immune response and also in motor axon development. The validation of this hit with a specific antibody revealed that the molecule is indeed involved in NMJ recovery of funtion. It is validation by different means (we also performed
CXCL12α mRNA FISH at the NMJ following Referee 1 suggestion, reported in Fig1E of the revised manuscript) that makes one trascriptomic hit relevant and biologically significant. The remaining experiments described in the present paper were built on this validation.
4. The staining of CXCL12 alpha is poorly visible and should be represented in white in the central panels to increase the contrast. The authors should comment on CXCL12 alpha distribution since its localization is unclear: some puncta seem to be present within the schwann cells, whereas some others map in areas not labeled in green (top in the alpha-latrotoxin 4 h sample).

CXCL12 staining is now reported in white and a new confocal image is presented.
After a careful inspection of dozen of pictures we can safely affirm that CXCL12 puncta are inside Schwann cells only. The few spots outside Schwann cells in the former picture may be attributed to some aspecificity of the secondary antibody, and also to the basal autofluorescence of muscle tissue.
In the revised manuscript we have reported different confocal images and the relative orthogonal projection that clearly show the localization of the chemokine in terminal Schwann cells. In addition, we have performed in situ hybridization of CXCL12α mRNA at the NMJ (as suggested by Referee 1), that further confirms the expression of the chemokine mRNA in PSCs ( Figure 1E).
Likewise the statement that CXCL12 alpha puncta are colocalized with Lysotracker should be tuned down, since this colocalization is partial at best ( Figure EV1). Figure EV2 seems also redundant.
Given the CXCL12α-positive spots, originally we wished to understand whether they were neutral classical secretory vesicles or acidic compartments destined to lysosomal exocytosis. This is the reason why we performed Lysotracker staining. Indeed, the chemokine appears to be accumulated in acidic structures, whose exact nature is currently under investigation. This point was arised also by Referee 1 and, accordingly, we have removed this image from the revised version. Indeed, establishing the nature of the secretory vesicle containing CXCL12α within PSCs is beyond the scope of the present study, and will be tackled in the future. Figure 2A, are control samples treated with alpha-latrotoxin and either with saline or an equivalent dose of an irrelevant antibody? Please clarify the methodology in the main text.

In
The bar labelled as "Ctr" corresponds to EJPs of soleus muscles from mice locally injected with saline in the hind limb, whereas the "anti CXCL12α" bar represents those from mice intraperitoneally injected with the sole anti-chemokine antibody. The methodology has been clarified accordingly in the revised manuscript (please see the Results section).
6. Figure 2D is not very informative, since the more informative quantification of these data is provided in panel E. Figure 2).

For clarity we have added a scheme of the microfluidic devices set-up and a new representative image (panel D of
Does treatment with CXCL12 alpha alter neuronal survival in the conditions described in this Figure? We have performed viability assay on SCMNs exposed to 500 ng/ml rCXCL12α for 5 DIV and found no differences in survival upon treatment.
Furthermore, is the effect on axonal elongation detected if CXCL12 alpha is only added in the somatic chamber?
When applied to the somatic chamber of microfluidic devices, rCXCL12α failed to promote axonal elongation. Results of the proposed experiment are reported in Figure EV4 of the revised manuscript.
7. The resolution of Figure 3D is not optimal.
A new confocal image is reported in Figure 3A of the revised manuscript.
8. Does CXCR4 co-localize in motoneurons with classical growth cone markers?
Co-localization of CXCR4 with the growth cone marker GAP43 has been performed and a representative image is shown in Figure 3B of the revised manuscript.
Why is SMI81 used in Figure 3E instead of SNAP25?
We apologize. SMI81 identifies the anti-SNAP25 antibody employed in the present study. Label has now been changed accordingly.
2nd Editorial Decision 19 April 2017 Thank you for the submission of your revised manuscript to EMBO Molecular Medicine. We have now received the enclosed reports from the referees that were asked to re-assess it. As you will see the reviewers are now globally supportive and I am pleased to inform you that we will be able to accept your manuscript pending final editorial amendments [not detailed]: Please submit your revised manuscript within two weeks. I look forward to seeing a revised form of your manuscript as soon as possible.
For animal studies, include a statement about randomization even if no randomization was used.
4.a. Were any steps taken to minimize the effects of subjective bias during group allocation or/and when assessing results (e.g. blinding of the investigator)? If yes please describe. Do the data meet the assumptions of the tests (e.g., normal distribution)? Describe any methods used to assess it.
Is there an estimate of variation within each group of data?
Is the variance similar between the groups that are being statistically compared?
Please ensure that the answers to the following questions are reported in the manuscript itself. We encourage you to include a specific subsection in the methods section for statistics, reagents, animal models and human subjects.

Captions
The data shown in figures should satisfy the following conditions: Source Data should be included to report the data underlying graphs. Please follow the guidelines set out in the author ship guidelines on Data Presentation. a statement of how many times the experiment shown was independently replicated in the laboratory.
Any descriptions too long for the figure legend should be included in the methods section and/or with the source data.

B--Statistics and general methods
the assay(s) and method(s) used to carry out the reported observations and measurements an explicit mention of the biological and chemical entity(ies) that are being measured. an explicit mention of the biological and chemical entity(ies) that are altered/varied/perturbed in a controlled manner. the exact sample size (n) for each experimental group/condition, given as a number, not a range; a description of the sample collection allowing the reader to understand whether the samples represent technical or biological replicates (including how many animals, litters, cultures, etc.). "Statistical analysis paragraph" "Statistical analysis paragraph" definitions of statistical methods and measures: 1. Data the data were obtained and processed according to the field's best practice and are presented to reflect the results of the experiments in an accurate and unbiased manner. figure panels include only data points, measurements or observations that can be compared to each other in a scientifically meaningful way. graphs include clearly labeled error bars for independent experiments and sample sizes. Unless justified, error bars should not be shown for technical replicates. if n< 5, the individual data points from each experiment should be plotted and any statistical test employed should be justified Please fill out these boxes ê (Do not worry if you cannot see all your text once you press return) a specification of the experimental system investigated (eg cell line, species name).

YOU MUST COMPLETE ALL CELLS WITH A PINK BACKGROUND ê
"Statistical analysis paragraph" "Statistical analysis paragraph" "Statistical analysis paragraph" "Statistical analysis paragraph" "Sample preparation for laser microdissection paragraph" "Statistical analysis paragraph" "Statistical analysis paragraph" "Statistical analysis paragraph" "Statistical analysis paragraph"