Locomotor recovery following contusive spinal cord injury does not require oligodendrocyte remyelination

Remyelination occurs after spinal cord injury (SCI) but its functional relevance is unclear. We assessed the necessity of myelin regulatory factor (Myrf) in remyelination after contusive SCI by deleting the gene from platelet-derived growth factor receptor alpha positive (PDGFRα-positive) oligodendrocyte progenitor cells (OPCs) in mice prior to SCI. While OPC proliferation and density are not altered by Myrf inducible knockout after SCI, the accumulation of new oligodendrocytes is largely prevented. This greatly inhibits myelin regeneration, resulting in a 44% reduction in myelinated axons at the lesion epicenter. However, spontaneous locomotor recovery after SCI is not altered by remyelination failure. In controls with functional MYRF, locomotor recovery precedes the onset of most oligodendrocyte myelin regeneration. Collectively, these data demonstrate that MYRF expression in PDGFRα-positive cell derived oligodendrocytes is indispensable for myelin regeneration following contusive SCI but that oligodendrocyte remyelination is not required for spontaneous recovery of stepping.

Using a conditional Myrf KO in oligodendrocytes, this manuscript definitively shows that remyelination is not essential for hindlimb locomotor recovery after a mid-thoracic contusive SCI. The experiments are well conceived, data for the most part properly analyzed, and the manuscript itself is extremely well written. These data are important as myelinating cell grafts have been reported by many investigators to improve locomotor function after these injuries. In light of the current findings, the conclusions from those studies will have to be re-interpreted. These data emphasize that the important variables for the extent of recovery after thoracic contusive SCI are the extent of spared white matter and the reorganization of the lumbar locomotor circuitry. While this conclusion has been inferred from previous studies, current data definitively rule out remyelination as a reparative process that impacts on the extent of recovery. These data thus have important implications with respect to rationale design of cellular therapies for SCI. Individual concerns are detailed below.
Lines 61-2: The wrong verb is used. I believe the context would be better conferred in the authors used 'identify' not 'act as' and 'which differentiate into' rather than 'produce'.
Neither the methods nor the results indicate if the Pdgfr-CreERT2:Myrffl/fl mice were heterozygous or homozygous on the two transgenes. This must be provided.
There is inconsistent nomenclature with respect to how the mouse gene and proteins are denoted. The authors are referred to the following website for current guidelines. http://www.informatics.jax.org/mgihome/nomen/gene.shtml Line 102+: The authors do not use the most appropriate controls in these studies. Ideally, they should use Pdgfr-CreERT2:Myrffl/fl mice + vehicle, rather than Myrffl/fl mice + tamoxifen. While I appreciate the difficulties in breeding double transgenics, double transgenics often show phenotypes that are distinct from the parent floxed strain. At the very least, they must show that there is no difference between these two controls.
Line 408+: To unequivocally count the absolute number of objects in a section, unbiased stereology must be used (e.g. PMID: 28798525, PMID: 25743692, PMID: 25743692). Those methods are not used here. The authors should just why they were not and why the methods that did employ were indeed devoid of 2D counting artifacts.
There are numerous places throughout this manuscript where the authors use quantitative descriptors and show only single histological images. If the authors want to use the quantitative word 'majority', quantification with appropriate statistical analyses must be done. It is hard to believe that Fig 5d would have nearly complete classes of thin, none, and thick myelin respectively. These sections seem somewhat cherry picked. The authors should provide quantitative data across multiple animals.  This is a very interesting study that is well-conducted and with conclusions supported by convincing data. It is also a very important study that should, it is to be hoped, put to bed the notion promoted by some that traumatic spinal cord injury will benefit from pro-remyelination interventions. My main issue concerns are less the way the study has been done but rather the way that the rationale is presented. Personally, I think the authors can more critical of the evidence for persistent demyelination in SCI on which much of the claims that remyelinationenhancing therapies will be useful are based. For example, the paper by Totoui and Keirstead (J Comp Neurol) contains an image claiming persistent demyelination that is widely recognised as misleading. I think the bold statement that chronic oligodendrocyte loss and demyelination is a feature of SCI (lines 54-55 page 2) is not one that many would agree with -and the phrasing should perhaps be tempered. It also somewhat contradicts the claim made in line 70-72 on page 3. Surely the issue here is that 1) when demyelination occurs it is generally followed by rather efficient remyelination (the paper by Jeffery and Smith in Brain Pathol 2006 provides convincing evidence that this occurs in clinical disease that supports the strong body of experimental data and should perhaps be cited), and 2) it is such a minor component of the overall pathology of SCI (compared with frank axonal loss for example) that it is unlikely to be a major driver of functional recovery?
Minor points 1. Many in the field prefer the term oligodendrocyte progenitor cell rather than oligodendrocyte precursor cell, since the term precursor implies a lineage restriction that is not a feature of adult OPCs (the authors own recent work on the origin of Schwann cells in the CNS being part of the evidence that these cells are not lineage restricted).
2. Page 10, line 264 -'it is' rather than 'it's' 3. The role of astrocytes in constraining the expansion of Schwann cell remyelination in the absence of oligodendrocyte remyelination should probably be given more prominence (see, for example, Monteiro de Castro et al Am J Pathol 2015).
Reviewer #3 (Remarks to the Author): After traumatic injury to the spinal cord, some axons crossing the injury site are severed or die back post-injury while others survive but become demyelinated post-injury. The question has arisen, therefore, whether preventing this demyelination or encouraging rapid and more extensive remyelination might be a useful therapeutic goal. The field appears divided on this. The current article by Duncan et al. tries to resolve this issue by examining the role of oligodendrocyte (OL) and myelin production in the limited spontaneous locomotor recovery that follows experimental traumatic injury to mouse spinal cord.
The authors prevented production of new OLs and OL-derived myelin by conditional deletion of the transcription factor Myrf in OL precursors (OPCs) using Pdgfra-CreER(T2), an approach that was devised previously to examine the role of new OLs in motor learning. The present study is wellcontrolled and demonstrates that new OL production is knocked down >90% in the spared ventrolateral white matter, following a contusion injury to the dorsal funiculus. This resulted in a ~50% reduction in remyelination of spared axons at the injury site, almost all of the observed remyelination being attributable to newly-generated Schwann cells. These Schwann cells were generated within 2 weeks post-injury, mainly from Pdgfra-positive precursors (possibly OPCs) since they were labelled by the Rosa-mGFP reporter. This is something that has been observed previously by the Tetzlaff lab following spinal cord injury, and by others in the context of gliotoxin-induced demyelination. This Schwann-cell-mediated remyelination, rather than OL-mediated remyelination, might be what drives locomotor recovery in the first 2 weeks post-injury, because OL-mediated remyelination occurs subsequent to that. Production of Schwann cells and Schwann cell-derived myelin was not diminished by conditional KO of Myrf in the present experiments so this study does not rule out the possibility that Schwann cell remyelination is important in locomotor recovery. This could be examined in future by an analogous approach using e.g. Po-CreER(T2).
An interesting and useful side-shoot of the study is the observation that almost all OL-mediated remyelination is prevented by deletion of Myrf, providing strong evidence that OL remyelination is via newly-generated OLs (from OPCs) rather than by elaboration of new myelin sheaths by preexisting OLs. Perhaps more could be made of this -in the Discussion for example -with reference to previous attempts to address this very issue (e.g. Crawford et al., 2016 Am J Pathol 186, 511).
Overall, I thought that this is an excellent study, well-conducted, well thought-through and well written on the whole, although the text could be shortened significantly if required. The main conclusion, that the limited locomotor recovery observed in the first few weeks of a contusion injury does not rely on OL-mediated remyelination seems sound, and could be important by casting doubt on the utility of attempting to improve outcomes by targeting OL-remyelination. It also focusses attention on the possible role of Schwann cells in functional recovery.
Minor points: 1. In Figures 1A and 3A, the lox sites should be re-drawn in the same orientation. In the opposite orientation as shown, cre-recombination would cause flip-flopping of the intervening DNA, not deletion. Conventionally the direction is shown 5' to 3' (L to R).
2. In most figures the labelling of some panels is far too small to be visible at final size.

Response to Editor and Reviewers
We would like to thank the reviewers and editor for finding time in their busy schedules to read our article and provide insightful comments and critiques. These comments have been the basis for our revisions which have greatly strengthened this manuscript.

Reviewer comments are in bold.
Author responses are in plain text. Statements from the manuscript that address reviewers' comments are italicized, and new text in the manuscript is italicized and bolded.

Reviewers' comments:
Reviewer #1 (Remarks to the Author): Using a conditional Myrf KO in oligodendrocytes, this manuscript definitively shows that remyelination is not essential for hindlimb locomotor recovery after a mid-thoracic contusive SCI. The experiments are well conceived, data for the most part properly analyzed, and the manuscript itself is extremely well written. These data are important as myelinating cell grafts have been reported by many investigators to improve locomotor function after these injuries. In light of the current findings, the conclusions from those studies will have to be re-interpreted. These data emphasize that the important variables for the extent of recovery after thoracic contusive SCI are the extent of spared white matter and the reorganization of the lumbar locomotor circuitry. While this conclusion has been inferred from previous studies, current data definitively rule out remyelination as a reparative process that impacts on the extent of recovery. These data thus have important implications with respect to rationale design of cellular therapies for SCI. Individual concerns are detailed below.
We thank the reviewer for their encouraging comments regarding the impact of this study. We agree that previous cell transplantation studies, which were thought to confer functional benefits through remyelination, will need to be re-interpreted. We highlight the significance of these data to cellular transplantation studies and more broadly to whether remyelination is a validated clinical target at several points in the manuscript: Lines 61-2: The wrong verb is used. I believe the context would be better conferred in the authors used 'identify' not 'act as' and 'which differentiate into' rather than 'produce'.
We agree with the reviewer and have used the suggested verbs to clarify the sentence. We have changed it from: Line 60-61: Platelet-derived growth factor receptor α (PDGFRα), act as oligodendrocyte precursor cells (OPCs) to produce new oligodendrocytes after SCI 20,22,26 To Line 60-62: Platelet-derived growth factor receptor α (PDGFRα) expression in resident, nonvascular associated cells, identifies these cells as oligodendrocyte progenitor cells (OPCs) 28,29 , which differentiate into new oligodendrocytes after SCI 20,22,26 .
Neither the methods nor the results indicate if the Pdgfr-CreERT2:Myrffl/fl mice were heterozygous or homozygous on the two transgenes. This must be provided.
We agree that this is an important methodological detail that was overlooked in the initial manuscript. We have now added whether the gene was heterozygous or homozygous for each transgene used in both the results and methods section of the manuscript. There is inconsistent nomenclature with respect to how the mouse gene and proteins are denoted. The authors are referred to the following website for current guidelines. http://www.informatics.jax.org/mgihome/nomen/gene.shtml We thank the reviewer for this suggestion and have corrected all cases of nomenclature for both genes and proteins that did not adhere to guidelines in both the text and figures. All gene and allele names now have their first letter capitalized and the whole name is italicized, while protein symbols have been capitalized in alignment with these conventions.
Line 102+: The authors do not use the most appropriate controls in these studies. Ideally, they should use Pdgfr-CreERT2:Myrffl/fl mice + vehicle, rather than Myrffl/fl mice + tamoxifen. While I appreciate the difficulties in breeding double transgenics, double transgenics often show phenotypes that are distinct from the parent floxed strain. At the very least, they must show that there is no difference between these two controls.
To address the reviewer's concern, and to ensure PDGFRα transgene insertion alone did not alter recovery after SCI, we conducted an additional experiment in which Myrf fl/fl mice heterozygous with the PDGFRα-CreERT2 transgene were compared to littermate Myrf fl/fl mice without the PDGFRα-CreERT2 transgene in the absence of tamoxifen. These mice did not differ in their response to T9/10 thoracic spinal cord injury, indicating there is no inherent difference between these two controls (Supplementary Figure 1 -see below).  Fig. 1

)'
We recognize that the transgenic mice breeding strategy could have been articulated more clearly in the text. As a result we have added or changed the following lines.

yielded litters in which all mice had both copies of Myrf surrounded by LoxP sites (Myrf fl/fl ) with individual mice either without (control mice) or with the PDGFRα-CreERT2 transgene (Myrf ICKO). '
This description indicates that all control mice used in the experiments were littermates of Myrf ICKO mice. This breeding strategy ensured that the control mice were on the same genetic background as the knockout mice. It had the added benefit of being ideal for blinding during behavioural experiments, as cages would contain both control and knockout animals.
Additionally, using Myrf fl/fl PDGFRα-CreERT2-negative mice administered tamoxifen as controls for behavioural experiments avoided known confounds that tamoxifen has on recovery. Tamoxifen is a highly biologically active molecule. At lower doses, tamoxifen has been demonstrated by several groups to be neuroprotective after spinal cord injury (Williams et al., 1996;Tian et al., 2009;Guptarak et al., 2014;Colon et al., 2016;de la Torre Valdovinos et al., 2016;Osuna-Carrasco et al., 2016), traumatic brain injury (Franco Rodriguez et al., 2013) and can even enhance remyelination by directly acting on oligodendrocyte progenitor cells (Gonzalez et al., 2016). Higher doses, like those administered for inducible Cre-LoxP experiments, yield increased cellular stress in select neuronal populations (Denk et al., 2015). By administering tamoxifen to all mice used in behavioural analyses, we control for its known biological effect.
Line 408+: To unequivocally count the absolute number of objects in a section, unbiased stereology must be used (e.g. PMID: 28798525, PMID: 25743692, PMID: 25743692). Those methods are not used here. The authors should just why they were not and why the methods that did employ were indeed devoid of 2D counting artifacts.
We appreciate the reviewer highlighting the importance of using proper stereological methodology in counts of absolute objects in a section. We agree that without the use of stereological principles there is an increased probability of counting artifacts.
In this manuscript we did counts of oligodendrocyte density in thick (3D) sections and 2D counts of myelin from thin sections. All of our counts of objects (either cells, or myelin sheaths) used the key principles of stereology to reduce counting artifacts and bias. For example, all 3D counts are conducted using an optical disector to guard against overestimates or duplicate object counts by counting only unique elements. This ensures that there is not a change in estimate due to an alteration in the size, shape or orientation of the cell. The experimental details from the manuscript demonstrating that stereological principles were used are highlighted in further detail below. Additionally, we discuss the stereology of our counts of relative cell density in 3D space versus the absolute counts of myelin sheaths from 2D sections. We recognize now that elements could have been stated more explicitly and these points are included within the methods section.
A key aspect of stereology is that systematic uniform random sampling (SURS) is employed (Gundersen et al., 1988;Brown, 2017). In all of our 3D cell counts, the entirety of a given section was uniformly sampled, as indicated in the methods, to ensure that objects had an equal chance of being counted (bolded is added text).

Line 432-434: 'We performed systematic uniform random sampling within each section 69 by overlying a grid (individual grid size 103 µm x 108 µm) onto a low magnification preview image of a cross section of spinal cord. One counting square for every 3 x 3 grid area was imaged at 400x magnification with a randomized start location (40x objective NA 1.3).'
Sections were also systematically sampled in the rostral-caudal orientation every 200 µm from lesion epicenter. Our goal in these cell counts was to determine the relative density and capacity of Myrf ICKO mice to generate new oligodendrocytes, which primarily occurs in close proximity to the lesion (Tripathi and McTigue, 2007;Hesp et al., 2015). For this reason, sections underwent systematic uniform random sampled in the epicenter and 200 or 400 µm rostral and caudal for counts of oligodendrocytes.

Lines 431-432: 'For analysis of cell densities, we imaged the epicenter of injury and the next two sections 200 and 400 µm rostral and caudal for each animal for a total of five sections per mouse.'
Crucially, for all counts of oligodendrocyte cell density from thick sections, optical disectors were used to ensure that only unique objects were counted. The entire Z-stack through a section was taken on a confocal microscope before cells were counted within an optical disector as in (Gundersen et al., 1988). Nuclei that came into focus were only counted, and nuclei at the edge of the optical disector where excluded if they touched two of the adjacent sides and included if they touched either of the two opposing sides. Further details (bolded italics) have been added to the methods section.
Line 436-437: 'Z-stacks were imaged through the entire depth of the 20µm thick section with 1 µm spacing between optical sections and cells were counted in three dimensional space within a 100 x 100 µm optical disector.

Nuclei that came into focus and were within the optical disector or in contact with right and upper edge were counted to ensure only unique objects were quantified.'
Therefore, we combined SURS with stereological counting probes in thick 3D sections to reduce sampling artifacts in these counts. This is now clear within the text. For 2D counts of total myelin content, myelin was quantified at the lesion epicenter because both the level of demyelination and remyelination is highest. Therefore, at this location myelin dynamics are most likely to impact locomotor recovery. The gold standard for measuring either axon number (Larsen, 1998;Zarei et al., 2016) or myelin thickness (g-ratio) is using 2D crosssections of axons. 2D counts combined with systematic sampling have been used to determine the number of axons within the tibial and optic nerve (Williams et al., 1996;Larsen, 1998).
We sectioned every 20 µm through our resin blocks to guarantee the area with the most severe pathology was measured to ensure that the lesion epicenter was compared between mice. For these counts, systematic uniform random sampling was again employed as well as a counting frame was used with 2 inclusion and 2 exclusion lines to ensure only unique objects were counted. This approach should greatly reduce the chance that individual myelinated axons were oversampled. Larsen and colleagues estimated that in the tibial nerve employing SURS and ensuring that at least 150-200 axons were counted was sufficient to reduce empirical variance to just 5% (Larsen, 1998). We adapted this protocol for use in the spinal cord, but counted a much larger 1500-2500 axons per animal due to the variability in axon density regionally across the injured cord. The area we counted was equivalent to 1/7 of the area of the spinal cord. We have now quantified myelin thickness using electron microscopy. We find that on average myelin tends to be thinner in control mice relative to Myrf ICKO throughout the spinal cord at 6 WPI (Figure 5e, f). The paucity of examples of thin myelin sheaths in Myrf ICKO (g ratio > 0.85), which are normally rarely found in the uninjured mouse spinal cord (James et al., 2011;Ishii et al., 2014), are suggestive of little remyelination (Figure 5e). Combined with increased number of unmyelinated axons > 1µm in Myrf ICKO (8336 ± 1072 axons relative to 1298 ± 327 in controls), provides compelling evidence that remyelination is ablated and chronic demyelination is present (Figure 5g). Broadly, these data indicate that images in figure 5d are representative and Myrf ICKO rarely have evidence of thinly myelinated fibers and outright demyelination is widely observed at 6 WPI. We thank the reviewer for this suggestion and think   This typographic error has been corrected.

In Fig 7, the schematics of oligodendrocyte myelin and Schwann cell myelin do not appear to accurately reflect the quantitative data in Figs. 2 & 4. They should be better drawn to scale.
We have changed the schematic in Fig. 7b to reflect the actual quantified values in the manuscript for myelin. Likewise, the BMS scores follow the recovery curves of the group means over time. We felt this figure was crucial for highlighting the temporal discordance between remyelination and behavioural recovery. This is a very interesting study that is well-conducted and with conclusions supported by convincing data. It is also a very important study that should, it is to be hoped, put to bed the notion promoted by some that traumatic spinal cord injury will benefit from proremyelination interventions.
My main issue concerns are less the way the study has been done but rather the way that the rationale is presented. Personally, I think the authors can more critical of the evidence for persistent demyelination in SCI on which much of the claims that remyelinationenhancing therapies will be useful are based. For example, the paper by Totoui and Keirstead (J Comp Neurol) contains an image claiming persistent demyelination that is widely recognised as misleading. I think the bold statement that chronic oligodendrocyte loss and demyelination is a feature of SCI (lines 54-55 page 2) is not one that many would agree with -and the phrasing should perhaps be tempered.
We agree with the reviewer that there is little evidence of persistent demyelination after chronic SCI. We have stated as much in recent reviews (Plemel et al., 2014;Assinck et al., 2017a). We did not intend to imply that chronic demyelination was a feature of SCI, only that oligodendrocyte loss occurred for several weeks after SCI. However, we recognize this phrasing could be construed as such and so we have reworded the sentence to ensure our rationale is clear.
Lines 54-55 changed from: 'However, chronic oligodendrocyte death 6 and demyelination of spared axons are characteristic after SCI 7, 8, 9 ,10 and could diminish connectivity of spared circuits' To: 'However, oligodendrocyte death in the weeks after SCI 6 presumably results in the demyelination of spared axons 7,8,9,10 , which could diminish the connectivity of spared circuits.' It also somewhat contradicts the claim made in line 70-72 on page 3. Surely the issue here is that 1) when demyelination occurs it is generally followed by rather efficient remyelination (the paper by Jeffery and Smith in Brain Pathol 2006 provides convincing evidence that this occurs in clinical disease that supports the strong body of experimental data and should perhaps be cited), and 2) it is such a minor component of the overall pathology of SCI (compared with frank axonal loss for example) that it is unlikely to be a major driver of functional recovery?
We have amended lines 54-55 (see previous response) so that we no longer contradict ourselves in lines 70-72. Many studies demonstrate that remyelination is a highly efficient process after SCI, including this one and previous work from our laboratory (Powers et al., 2012;  We thank the reviewer for highlighting these problems with how the rationale is presented, and think it is now much clearer and more accurate.

Minor points
1. Many in the field prefer the term oligodendrocyte progenitor cell rather than oligodendrocyte precursor cell, since the term precursor implies a lineage restriction that is not a feature of adult OPCs (the authors own recent work on the origin of Schwann cells in the CNS being part of the evidence that these cells are not lineage restricted).
We have changed the term 'precursor' to 'progenitor' in all instances in the text.

Page 10, line 264 -'it is' rather than 'it's'
This change has been implemented. We agree that astrocytes likely have a prominent role in constraining Schwann cell myelination after SCI. We have elaborated on this point in the discussion, and added the citation mentioned. We also pointed out how the Myrf ICKO did not alter astrocyte coverage, which could have restricted Schwann cell compensation following oligodendrocyte remyelination failure with Myrf ICKO.

Reviewer #3 (Remarks to the Author):
After traumatic injury to the spinal cord, some axons crossing the injury site are severed or die back post-injury while others survive but become demyelinated post-injury. The question has arisen, therefore, whether preventing this demyelination or encouraging rapid and more extensive remyelination might be a useful therapeutic goal. The field appears divided on this. The current article by Duncan et al. tries to resolve this issue by examining the role of oligodendrocyte (OL) and myelin production in the limited spontaneous Professor and ICORD Director ICORD, Blusson Spinal Cord Centre, 818 W 10 th Ave. Vancouver, BC Canada V5Z 1M9.
Email: tetzlaff@icord.org locomotor recovery that follows experimental traumatic injury to mouse spinal cord.
The authors prevented production of new OLs and OL-derived myelin by conditional deletion of the transcription factor Myrf in OL precursors (OPCs) using Pdgfra-CreER(T2), an approach that was devised previously to examine the role of new OLs in motor learning. The present study is well-controlled and demonstrates that new OL production is knocked down >90% in the spared ventrolateral white matter, following a contusion injury to the dorsal funiculus. This resulted in a ~50% reduction in remyelination of spared axons at the injury site, almost all of the observed remyelination being attributable to newly-generated Schwann cells. These Schwann cells were generated within 2 weeks post-injury, mainly from Pdgfra-positive precursors (possibly OPCs) since they were labelled by the Rosa-mGFP reporter. This is something that has been observed previously by the Tetzlaff lab following spinal cord injury, and by others in the context of gliotoxin-induced demyelination. This Schwann-cell-mediated remyelination, rather than OL-mediated remyelination, might be what drives locomotor recovery in the first 2 weeks post-injury, because OL-mediated remyelination occurs subsequent to that.

Production of Schwann cells and Schwann cell-derived myelin was not diminished by conditional KO of Myrf in the present experiments so this study does not rule out the possibility that Schwann cell remyelination is important in locomotor recovery. This could be examined in future by an analogous approach using e.g. Po-CreER(T2).
We thank the reviewer for the kind words on the construct and importance of this study. We agree that Schwann cell myelination may be a driver of functional recovery and stated this explicitly in the text: Line 307-309: 'Importantly, we found Schwann cell myelination, in contrast to oligodendrocyte remyelination, occurs early enough after injury to potentially mediate recovery.' An interesting and useful side-shoot of the study is the observation that almost all OLmediated remyelination is prevented by deletion of Myrf, providing strong evidence that OL remyelination is via newly-generated OLs (from OPCs) rather than by elaboration of new myelin sheaths by pre-existing OLs. Perhaps more could be made of this -in the Discussion for example -with reference to previous attempts to address this very issue (e.g. Crawford et al., 2016 Am J Pathol 186, 511).
We agree with the reviewer that this is an important point and have emphasized at several points in the manuscript. We have also added the citation mentioned above to the discussion and the following lines (bolded italics). Overall, I thought that this is an excellent study, well-conducted, well thought-through and well written on the whole, although the text could be shortened significantly if required. The main conclusion, that the limited locomotor recovery observed in the first few weeks of a contusion injury does not rely on OL-mediated remyelination seems sound, and could be important by casting doubt on the utility of attempting to improve outcomes by targeting OL-remyelination. It also focusses attention on the possible role of Schwann cells in functional recovery.

Lines 260-263: 'Oligodendrocyte genesis by resident PDGFRα+ OPCs cannot be compensated for by other cell sources like ependymal cells or Schwann cells, even when resident OPC
Minor points: 1. In Figures 1A and 3A, the lox sites should be re-drawn in the same orientation. In the opposite orientation as shown, cre-recombination would cause flip-flopping of the intervening DNA, not deletion. Conventionally the direction is shown 5' to 3' (L to R).
We apologize for this error, and have corrected this in Figures 1A and 3A. 2. In most figures the labelling of some panels is far too small to be visible at final size.
We have made an effort in all of the figures to increase the font size where possible.