The transcription factor Slug represses p16Ink4a and regulates murine muscle stem cell aging

Activation of the p16Ink4a-associated senescence pathway during aging breaks muscle homeostasis and causes degenerative muscle disease by irreversibly dampening satellite cell (SC) self-renewal capacity. Here, we report that the zinc-finger transcription factor Slug is highly expressed in quiescent SCs of mice and functions as a direct transcriptional repressor of p16Ink4a. Loss of Slug promotes derepression of p16Ink4a in SCs and accelerates the entry of SCs into a fully senescent state upon damage-induced stress. p16Ink4a depletion partially rescues defects in Slug-deficient SCs. Furthermore, reduced Slug expression is accompanied by p16Ink4a accumulation in aged SCs. Slug overexpression ameliorates aged muscle regeneration by enhancing SC self-renewal through active repression of p16Ink4a transcription. Our results identify a cell-autonomous mechanism underlying functional defects of SCs at advanced age. As p16Ink4a dysregulation is the chief cause for regenerative defects of human geriatric SCs, these findings highlight Slug as a potential therapeutic target for aging-associated degenerative muscle disease.

. This is not quantitative and they need to show reproducibility. It is also advisable to include negative control regions. Are these cells 'primary' or (somehow) immortalised myoblasts? To conclude that "Slug is a transcriptional repressor for p16Ink4a", they need to provide functional validation using reporter assay with the wt/mutant promoter, otherwise tone down. Fig. 4h: it would be more informative if the authors show the difference between wt resting and wt cultured. Using the C hIP assay, can they see the reduction of slug on the p16 promoter during culture? Fig. 4i,j: "these results indicated that Slug deficiency leads to an increase in p16Ink4a transcription in SC s, and replicative stress signaling triggered by SC activation and proliferation concurrently increases the stability of p16Ink4a mRNA". This is not supported by the data. Actually, the logic doesn't make sense. How can one hypothesise that mRNA is unstable when it is high. It would make more sense to speculate that p16 protein is unstable in resting SC s? Fig. 5c: cumulative PD is decreased between p2 and 3, suggesting a substantial number of cells might undergo cell death (or plating efficiency becomes low). The authors need to clarify whether slug ko induces both senescence and cell death thus residual (survival) cells are used for the assays in vitro.
Minor points: Reviewer #2: Remarks to the Author: In general this is a well written manuscript with significant evidence provided that the EMT transcription factor Slug/Snai2 can control muscle stem cell quiescence/senescence though the repression of p16Ink4a . The SC -specific deletion of a novel conditional Slug allele using Pax-7 C re mouse line is very convincing concerning the cell-autonomous role of Slug in regulating SC quiescence/senescence. In addition, the rescue of muscle regeneration of Slug deficient mice by breeding onto p16Ink4a deficient background provides very strong genetic evidence of the important of p16 in the Slug null phenotype.
Major concerns: 1) The authors show that there is a difference in Snai2 expression in SC in the muscle between young and old mice and that this correlates with changes in p16 expression. However, no mechanistic insight or discussion is provided concerning exactly how or why Snai2 levels fail off over time in the SC population. Is a similar correlation observed in human muscle samples in young vs old or in control vs diseased muscle biopsies? From a regenerative medicine perspective, how would one potentially modulate Snai1 expression in the adult? 2) Given the gene changes that occur in Snai2 null SC populations that are involved in cytokinesis, chromosome segregation and microtubule assembly how sure are the authors that asymmetric vs symmetric cell division is not affected and contributing to the regeneration defect given the importance of asymmetric self-renewal in the SC cells of the muscle (see Kuang et al., C ell 2007).
Minor issues: 1) In the discussion the authors mention the possibility of Snai1 partial compensation in Snai2 SC phenotype but do not mention recent paper by Sieiro et al., eLife 2016 whereby it was demonstrated that Snai1 levels can indirectly influence important muscle transcription factors such as Myf5. 2) Figure 4F is confusing and needs to be changed. It is Snai2 locus that has been FLAG tagged at C -terminus and not p16Ink4a. This should be removed as it is depicted in the Sup. Fig 3? In this Figure primers should be indicated that flank E box element. How conserved is this E-box element? Is this E-box element present in the human p16Ink4a promoter (see major concern 1 above).
3) How sure are the authors that the C -terminal tag on SNAI2 that was used for C hIP experiments does not interfere with the normal function of SNAI2? How does this compare to N-terminal tags? Would be good to show that protein half life and localization is not affected by this Tag.
Reviewer #3: Remarks to the Author: Review of Zhu et al., 2018 Nature C ommunications manuscript The manuscript entitled "Transcription factor Slug/Snail2 reinforces self-renewal in aged skeletal muscle stem cell" by Zhu et al., (2018) attempt to demonstrate that the zinc-finger transcription factor Slug is highly expressed in quiescent muscle stem cells, Satellite C ells (SC s), and function as a direct transcriptional repressor of the cell cycle inhibitor, p16ink4a. Then, loss of Slug promotes increases expression of p16ink4a in SC s and accelerates the entry of SC s into a senescent state upon damage-induced stress. Therefore, p16ink4a depletion partially rescues defects of Slug null SC s. The study also, through the body of the results section, infers that decreased level of Slug expression and consequent de-repression of p16ink4a in muscle SC s during aging, intrinsically impact their self-renewal potential after multiple round of regeneration. It should be noted that the authors of the manuscript under review have previously shown that Slug repress self-renewal of hematopoietic stem cells (HSC s) and is essential for controlling the transition of HSC s from relative quiescence under steady-state condition to rapid proliferation under stress conditions (Sun et al, 2010). In summary, the manuscript by Zhu et al, (2016) has identified of a novel upstream regulator of p16Ink4a. However numerous technical flaws minimize the authors conclusions General points: First, all the experiments are conducted with Slug germline knockout or Pax7-C re conditional knockout mice, raising the possibility that the phenotype is not confined only to muscle SC s, calling into question the proposed Slug intrinsic role in SC s. Also, the phenotype can derive from post-natal maturation mechanisms that influence adult homeostasis, proliferation, differentiation and self-renewal potential of SC s, during muscle regeneration. So, to this point, it is essential to use Pax7C reER tamoxifen inducible mice models to determine the specific functions of Slug solely in quiescent SC s and their downstream progeny during muscle repair. Second, to determine a possible SC s self-renewal potential defects, the author take advantage of transplantation assay and quantify self-renew SC s mainly by FAC S analysis of VC AM+/GFP+ SC s from WT or Null mice. As they state in the text, one-month post transplantation some of the donor SC s may be still proliferating, therefore the use of VC AM marker (VC AM marks SC s and their progenitors) cannot be used to determine the fraction of self-renewed SC s. There are two important criteria that SC s must fulfill to be considered 'self-renewed': 1) location: SC s must reenter the muscle fiber niche and reside under the basal lamina; 2) quiescent state: upon muscle injury SC s activate and acquire proliferation markers (Ki67+/MyoD+). Once they self-renew, they must go back to quiescence (Ki67-/MyoD-). Therefore, the authors need to perform more detailed experiments to determine if Slug impact SC s self-renewal.
Third, the authors need to perform a more careful analysis of cell fate during the repair processwhile they argue for a senescence mediated control of self-renewal, in vitro experiments suggest that a large number of progenitors undergo senescence-not the cells that return to quiescence. If a large number of progenitors are senescing in vitro-why does this not manifest as a phenotype in vivo? Is this due to the use of germline mouse in senescence experiments, or a decline in the number of myonuclei during regeneration that was not scored by the authors.
Fourth, based on their title a reader would expect more figures related to SC s self-renewal decline during aging. Only the final figure attempt to connect Slug, self -renewal and aging. During aging, the number and function of muscle SC s decline. Recently, it has been showed also that maintenance of quiescence in adult mouse life relies on active repression of senescence pathway (Sousa-Victor P et al., 2014), mainly though p16Ink4a. Therefore, does loss of Slug in young cells make them like aged/geriatric SC s? Detailed points: Figure 2A: Rather than having SC s compared to DN cells, which are almost unknown population, it will be more informative and relevant to compare population previously characterized such as FAPS, C D31 endothelial cells (where the role of Slug is well characterized) or blood C D45+ cells. Figure 2B: Please quantify slug expression in activated SC s, for example 2-3 days in proliferation conditions. If Slug is increased in proliferating cells, how would this change the authors conclusions? Figure 2F: Please quantify Slug+ Pax7+SC s; quantification of Pax7+ SC s need to be more accurate: please stain adult tissue section from C trl and Slug cKO mice for Pax7, Laminin and Ki67/MyoD. Figure 2: The authors showed defect in skeletal muscle repair after multiple round of injury despite the increase SC s fraction. C an you please address if SC s isolated from SlugcKO are precociously activated and consequentially fail to differentiate? Brdu experiments to show expansion due to activation of SC s and analysis of myogenic fate are required. Please quantify the number of myonuclei in regenerating muscle fibers. After single /double injury test the propensity of Pax7SC s to proliferate/differentiate/senesce. Figure 4B-C : The author should pull out genes that are involved in muscle SC s self-renewal and maintenance of stem cell pool, if they want to focus on self-renewal defect. Figure 4H: Please show absolute value of mRNA analysis for p16ink4a level during quiescence and activation in vitro and in vivo in Slug wt versus Slug null. Moreover, a time course analysis of Slug expression during activation in relation to p16 is recommended. Figure 4I: Please stain for laminin, it is important to know where Pax7+SC s/P16+ cells are located. Figure 5 A: Please stain and quantify terminally differentiated myotubes for Myosin Heavy C hain and reserve cells for Pax7. Is there a defect in reserve cell number? P16ink4a has been used as a marker of senescence, but not all senescent cells are p16ink4a+ (Rodier and C ampisi, 2011). Please confirm that increased in senescence is related to increased p16ink4a+ cells. Please stain reserve cells after 7d subgrowth in vitro with p16 antibody. Figure 5H: Please quantify Pax7+ senescent cells after 10d post injury. Their title of figure 5 is misleading, because they did not look at self-renewal, except for the in vitro reserve cells assay. They observed entrance into cellular senescence under proliferative stress. In vivo self-renewal senescence has not been shown, please do that. Figure 6A-B: After deletion of p16 (again at germline level) is senescence decreased? How many Pax7+ cells are also SAbgal+? Please quantify self-renewed cells on aged muscle section and SC s contribution to muscle fibers. SC s repopulation experiments: please specify if your donor SC s are from aged mice and include also young/adult mice as a control. Figure 7I Quantification of dystrophin+ fibers is needed Is Slug overexpression impacting cell proliferation and reducing senescence? Provide experimental The authors show that slug plays a role for muscle regeneration through regulating p16. They show loss of slug in SCs leads to p16 upregulation and induction of senescence in a p16 dependent manner. Consistently, the loss of slug in SCs results in decreased regenerative capacity during repeated injury. In addition, this is partially rescued by p16 ablation as well as enforced expression of slug. Data suggest a new link between slug and p16/senescence and its key role in muscle regeneration. The experiments are generally well designed.

Supp Fig. 1d shows a reduced myofiber size in whole body slug KO mice, but this doesn't seem to be the case in cKO? The authors might want to comment on this.
Answer: This is an interesting question. It is worth noting that although whole-body Slug KO mice exhibit reduced myofiber size and smaller hindlimb muscle mass than those in wild-type mice; but, when normalized to body weight, none of the relative weights of these muscles were affected by loss of Slug. This is because the whole-body Slug KO mice also exhibit smaller body size, which might be due to involvement of Slug in skeletal stem cell homeostasis and osteogenesis (Tang et al., 2016 Nat Cell Biol 18(9):917). Therefore, we think less muscle mass and smaller myofiber size are only phenotypes in scale with the skeletal bone and body size, rather than reflecting a major role in prenatal skeletal muscle development.

Fig. 3g/h. This is a nice experiment but not quantitative. The figure only shows a single result. They need to show reproducibility of the result.
Answer: Thank the Reviewer for the compliment. This is an adapted competitive repopulation assay that was applied in muscle stem cell study in this manuscript for the first time. We did repeat independently this experiment to have obtained similar result indicating the compromised self-renewing capacity of Slug null SCs ( Supplementary Fig. 6 in this revised manuscript).

Fig. 4c. please provide normalized enrichment scores (NES) and adj-p values.
Answer: Per request by the other reviewer, pathway enrichment analysis displayed in the previous Fig. 4c was revised and presented as Supplementary Fig. 7 (in this revised manuscript). NES and adjp values were added correspondingly in each panel.

Fig. 4g. This is not quantitative and they need to show reproducibility. It is also advisable to include negative control regions. Are these cells 'primary' or (somehow) immortalized myoblasts? To conclude that "Slug is a transcriptional repressor for p16Ink4a", they need to provide functional validation using reporter assay with the wt/mutant promoter, otherwise tone down.
Answer: Thanks for the suggestion. We designed another pair of primers targeting the 3' untranslated region with no E-box element as a negative control. The binding affinity of Slug on the putative E-box element was quantified relative to the non-specific binding by ChIP-qPCR assay ( Fig. 4f in this revised manuscript). For primary myoblasts, they might be somehow immortalized during selection process after CRISPR/Cas9-mediated tagging. As suggested by the reviewer, we performed p16 luciferase reporter assay in freshly isolated SCs to further support the conclusion from the ChIP-qPCR results (Supplementary Fig. 9 in this revised manuscript).

Fig. 4h: it would be more informative if the authors show the difference between wt resting and wt cultured. Using the ChIP assay, can they see the reduction of slug on the p16 promoter during culture?
Answer: It is a constructive suggestion. We were unable to show reduced binding of Slug on the p16 promoter during culture by ChIP assay because there is a lack of a validated ChIP graded Slug antibody for ChIP assay in freshly isolated SCs. That's why we performed Slug affinity tagging at its Cterminus in SC-derived myoblasts by CRISPR/Cas9 technique to facilitate assessing the occupancy of endogenous Slug at the promoter region of p16 by ChIP assay. However, we did perform a timedependent expression analysis of both Slug and p16 Ink4a in ex-vivo cultured myoblasts, showing that Slug was reduced but p16 Ink4a was increased along with the culture (Supplementary Fig. 10 in this revised manuscript). Together with the data from the ChIP-qCPR and p16 luciferase reporter assays, it should be evident that Slug actively represses p16 Ink4a in SCs.

Fig. 4i,j: "these results indicated that Slug deficiency leads to an increase in p16Ink4a transcription in SCs, and replicative stress signaling triggered by SC activation and proliferation concurrently increases the stability of p16Ink4a mRNA". This is not supported by the data. Actually, the logic doesn't make sense. How can one hypothesize that mRNA is unstable when it is high? It would make more sense to speculate that p16 protein is unstable in resting SCs?
Answer: We really appreciate the Reviewer's comments and suggestion. We changed the conclusion into "these results indicated that Slug deficiency leads to an increase in p16Ink4a transcription in SCs, and replicative stress signaling triggered by SC activation and proliferation concurrently increases the stability of p16 Ink4a protein." in this revised manuscript. Indeed, it was reported that p16 translation is suppressed by miR-24 (Lal et al., 2008 PLos One 3(3):e1864), which is highly expressed in quiescent SCs but significantly down-regulated in activated SCs (Sun et al., 2018 Mol Ther Nucleic Acids 11:528). We have added this post-translational regulation mechanism in the Discussion section (in this revised manuscript).

Fig. 5c: cumulative PD is decreased between p2 and 3, suggesting a substantial number of cells might undergo cell death (or plating efficiency becomes low). The authors need to clarify whether slug ko induces both senescence and cell death thus residual (survival) cells are used for the assays in vitro.
Answer: Thank the Reviewer for this instructive suggestion. Although decrease of cumulative PD might indicates senescence and cell death or a low plating efficiency, we think that it was more likely ascribed to cellular senescence, due to the following reasons: i) Gene ontology enrichment analysis of biological processes showed that Slug deletion derepressed a set of genes related to cellular senescence (Fig. 4c in the revised manuscript), but apoptosis-related GO terms were enriched neither in Slug null SCs ( Fig. 4c in the revised manuscript) nor in ex-cultured Slug-silenced primary myoblasts (Fig. 4a in the revised manuscript); ii) New data from SC transplantation experiment in the revised manuscript demonstrated a robust engraftment of Slug-deficient SCs after transplantation (Supplementary Fig.  4e-g in the revised manuscript). Little or no Slug null SCs-derived myofiber should be detected If Slug deletion caused cell death in SCs. (Fig. 6a,b)

". 'Statistical significance' is not shown.
Answer: We added 'Statistical significance' in Figure 6b (in this revised manuscript).

Fig. 4f seems to show tagging of p16, not slug?
Answer: We apologize for this mistake. Flag-tagging was actually introduced in Slug gene. The diagram illustrating epitope tagging of endogenous Slug in myoblasts by CRISPR/Cas9-mediated genome editing has been shown in Supplementary Fig. 8 (in this revised manuscript).

Fig. 7a: which dots represent slug?
Answer: It is indicated with an arrow in each panel (Fig. 7a).

Reviewer #2 (Remarks to the Author):
In general this is a well written manuscript with significant evidence provided that the EMT transcription factor Slug/Snai2 can control muscle stem cell quiescence/senescence though the repression of p16Ink4a. The SC-specific deletion of a novel conditional Slug allele using Pax-7 Cre mouse line is very convincing concerning the cell-autonomous role of Slug in regulating SC quiescence/senescence. In addition, the rescue of muscle regeneration of Slug deficient mice by breeding onto p16Ink4a deficient background provides very strong genetic evidence of the important of p16 in the Slug null phenotype.

1) The authors show that there is a difference in Snai2 expression in SC in the muscle between young and old mice and that this correlates with changes in p16 expression. However, no mechanistic insight or discussion is provided concerning exactly how or why Snai2 levels fail off over time in the SC population. Is a similar correlation observed in human muscle samples in young vs old or in control vs diseased muscle biopsies? From a regenerative medicine perspective, how would one potentially modulate Snai1 expression in the adult?
Answer: We appreciate the Reviewer for these constructive comments and suggestions. We did observe an aging-associated reduction of Slug/Snai2 expression in mouse SCs. Mechanistically, Slug expression is under control of a number of signaling pathways such as fibroblast growth factor (FGF), Wnt, transforming growth factor β (TGFβ), Notch, Stem cell factor (SCF), integrins and estrogens etc.  (12):2300-2311), which are known as inducers of Slug, were reported to decline with age in mouse SCs. Therefore, we speculate these impaired upstream signaling pathways might account for Slug insufficiency in aged SCs.
Unfortunately, we are unable to provide a direct experimental evidence showing a similar trend of aging-associated decrease of Slug in human SCs since our current study is a pilot investigation in mouse, and there lacks referable database comparing gene expression between young and old human muscle stem cells. Nevertheless, we believe our finding in aged mouse SCs may be also conserved in their human counterpart because 1) the potential Slug-binding site (E-box) consensus sequence is also detected in the promoter region of human p16 Ink4a (Supplementary Fig. 9c); 2) Knockdown or overexpressing Slug in primary human myoblasts causes up and down-regulation of p16 Ink4a expression, respectively (Supplementary Fig. 9d-g); 3) active Notch and Notch ligand Delta are declined in old human muscle compared to that of in young's (Carlson et al., 2009 EMBO Mol Med 1:381-391).
Our findings in current study offer a novel therapeutic target for aging-associated degenerative muscle disease. Indeed, there are some small molecules have been reported to induce or suppress Slug expression in some types of cells (Barrallo-Gimeno and Nieto, 2005 Development 132(14):3151). We have confirmed that forced expression of activated Notch1 is able to induce Slug expression in cultured myoblasts. Therefore, it would be interesting to conduct in vivo studies by intramuscular or intraperitoneal administrating small molecules that induce Slug activators or suppress Slug repressors in aged mice to test the improvement of aging-associated muscle stem cell defects

2) Given the gene changes that occur in Snai2 null SC populations that are involved in cytokinesis, chromosome segregation and microtubule assembly how sure are the authors that asymmetric vs symmetric cell division is not affected and contributing to the regeneration defect given the importance of asymmetric self-renewal in the SC cells of the muscle (see Kuang et al., Cell 2007).
Answer: These are excellent suggestions for the potential mechanistical study on self-renewal defect in Slug-deficient SCs. As revealed by Kuang et al. (Cell 2007), Notch signaling plays a critical role in the maintenance of SC self-renewal by asymmetric division. Unfortunately, neither gene ontology analysis nor differential gene expression heatmap results in our Slug null SC data set showed significant changes in components involved in Notch signaling pathways. However, whether Slug directly regulates asymmetric cell-fate determinant Numb segregates is worthy of future investigation. Regarding those significantly changed genes involved in cytokinesis, chromosome segregation and microtube assembly upon Slug deletion, we did observe a relatively faster activation of Slug-null SCs from G 0 to S stage during first cell division under stimulus. But after a couple rounds of cell proliferation, p16 was significantly increased in Slug-deficient SCs, and caused cellular senescence.

1) In the discussion the authors mention the possibility of Snai1 partial compensation in Snai2 SC phenotype but do not mention recent paper by Sieiro et al., eLife 2016 whereby it was demonstrated that Snai1 levels can indirectly influence important muscle transcription factors such as Myf5.
Answer: According to the Reviewer's suggestion, we included this paper in the discussion section in our revised manuscript.

2) Figure 4F is confusing and needs to be changed. It is Snai2 locus that has been FLAG tagged at C-terminus and not p16Ink4a. This should be removed as it is depicted in the Sup. Fig 3? In this Figure primers should be indicated that flank E box element. How conserved is this E-box element? Is this E-box element present in the human p16Ink4a promoter (see major concern 1 above).
Answer: We apologize for the mistake. A diagram for epitope tagging of endogenous Slug in myoblasts by CRISPR/Cas9-mediated genome editing has been shown in Supplementary Fig. 8 (in this revised manuscript). This E-box element is highly conserved, and also present in the human p16 Ink4a promoter region (Supplementary Fig. 9c).

3) How sure are the authors that the C-terminal tag on SNAI2 that was used for ChIP experiments does not interfere with the normal function of SNAI2? How does this compare to Nterminal tags? Would be good to show that protein half-life and localization is not affected by this Tag.
Answer: FLAG tag is one of the most popular tags and has been widely used to immunoprecipitation and ChIP assays. FLAG tag has never been reported to interfere normal function of target genes. It has been shown that the C-terminal FLAG tag on Snail2/Slug did not affect nuclear localization of Slug in cells and its function in promoting EMT and regulating its target genes (Gastroenterology 2014, 146:1386-1396).
Although we did not compare C-terminal tag with N-terminal tag in our manuscript, we believe that N-terminal tag should work with ChIP assay since it is unlikely for a N-terminal tag to affect the DNAbinding domain of Slug locate in the C-terminal.
We confirmed by western blot that protein half-life was not affected by this tag (Supplementary  Fig. 8d).

Review of Zhu et al., 2018 Nature Communications manuscript
The manuscript entitled "Transcription factor Slug/Snail2 reinforces self-renewal in aged skeletal muscle stem cell" by Zhu et al., (2018)

First, all the experiments are conducted with Slug germline knockout or Pax7-Cre conditional knockout mice, raising the possibility that the phenotype is not confined only to muscle SCs, calling into question the proposed Slug intrinsic role in SCs. Also, the phenotype can derive from post-natal maturation mechanisms that influence adult homeostasis, proliferation, differentiation and self-renewal potential of SCs, during muscle regeneration. So, to this point, it is essential to use Pax7CreER tamoxifen inducible mice models to determine the specific functions of Slug solely in quiescent SCs and their downstream progeny during muscle repair.
Answer: According to the Reviewer's suggestions, we crossed Slug fl/fl mice with Pax7 creER mice to obtain Slug fl/fl Pax7 creER and Slug fl/+ Pax7 creER (used as control) mice. Slug gene was efficiently deleted in adult MuSC by consecutive intraperitoneal injection of tamoxifen for 5 days (Supplementary Fig. 3a-c in this revised manuscript). Similar to what we found in SlugKO (germline knockout) and Slug cKO models (Pax7-Cre conditional knockout), tamoxifen-induced deletion of Slug in adult SCs also caused severe regenerative defect in secondarily but not primarily injured muscles (Supplementary Fig. 3d-f).
In addition, we repeated the flow cytometry and transplantation-based quantitative assay for SC self-renewal using primary SCs isolated from Slug fl/fl mice. Slug fl/fl SCs were infected with control retrovirus or retrovirus-expressing Cre recombinase, which deleted Slug efficiently through catalyzing homologous recombination of DNA fragment between loxP sites (Supplementary Fig. 5a,b). Controland Cre-virus-infected cells were then transplanted into each side of pre-injured TA muscles, respectively. Similar to the results using SCs from previous Slug cKO mice, the Cre retrovirus-infected SCs yielded about 5-fold less GFP positive fraction in the total SC population from the recipients compared to that of Control virus-infected SCs (Supplementary Fig. 5c,d). More importantly, we confirmed by immunohistochemistry analysis that significantly less Cre retrovirus-infected SCs returned to native SC niche than Control virus-infected SCs did (Supplementary Fig. 5e,f). Together, these data strongly affirmed our previous conclusions based on the Slug cKO mice in this study.

Second, to determine a possible SCs self-renewal potential defects, the author take advantage of transplantation assay and quantify self-renew SCs mainly by FACS analysis of VCAM+/GFP+ SCs from WT or Null mice. As they state in the text, one-month post transplantation some of the donor SCs may be still proliferating, therefore the use of VCAM marker (VCAM marks SCs and their progenitors) cannot be used to determine the fraction of self-renewed SCs. There are two important criteria that SCs must fulfill to be considered 'self-renewed': 1) location: SCs must reenter the muscle fiber niche and reside under the basal lamina; 2) quiescent state: upon muscle injury SCs activate and acquire proliferation markers (Ki67+/MyoD+). Once they self-renew, they must go back to quiescence (Ki67-/MyoD-). Therefore, the authors need to perform more detailed experiments to determine if Slug impact SCs self-renewal.
Answer: The flow cytometry and transplantation-based quantitative assays for SC self-renewal was performed according to the method described by Arpke and Kyba (2016 Skeletal Muscle Regeneration in the Mouse pp163-179). The similar assay has been also used to compare the cell-autonomous stem cell self-renew capability between young and aged mice (Cosgrove et al., 2014 Nat Med 20:255-264).
According to the paper by Hardy et al. (PLos One 2016 11(1):e0147198), 95% percent of undifferentiated exogenous SCs were quiescent; only a small portion of the donor SCs may be still proliferating even by one-month post transplantation and these cycling SCs could be activated during preparation of mononucleated cells (Velthoven et al., 2017Cell Reports 21:1994-2004Machado et al., 2017Cell Reports 21:1982-1993. Nevertheless, in response to the reviewer's question regarding if Slug impacts SC self-renewal, we have provided more evidence in the revised manuscript including 1) immunohistochemistry assay demonstrating significantly reduced Slug-null SCs that returned to the native stem cell niche after transplantation in comparison to wild-type control SCs (Supplementary Fig. 5e,f); 2) in vitro myofiberassociated SC culture assay showing that Slug-deficient SCs have an intrinsic defect in self-renewal ( Fig. 3i-k).

Third, the authors need to perform a more careful analysis of cell fate during the repair processwhile they argue for a senescence mediated control of self-renewal, in vitro experiments suggest that a large number of progenitors undergo senescence-not the cells that return to quiescence. If a large number of progenitors are senescing in vitro-why does this not manifest as a phenotype in vivo? Is this due to the use of germline mouse in senescence experiments, or a decline in the number of myonuclei during regeneration that was not scored by the authors.
Answer: It is notable that in vivo regenerating environment is more complicated than in vitro culture condition. SCs cultured in serum-rich growth medium mainly undergo proliferation; However, SCs in regenerating muscle go through not only proliferation but also differentiation and self-renewal.
Experimental data obtained from in vitro culture model reflected Slug-deficient SCs acquire features of senescence more easily than wildtype SCs under proliferative stress. We showed in the revised manuscript that differentiation of Slug null SCs is normal as indicated by comparative in vitro induced differentiation and in vivo engraftment after transplantation (Supplementary Fig. 4c-g). It is thus reasonable that first regeneration was normal in Slug -/mice. However, when looking into those activated but not differentiated SCs in regenerating muscle, p16 Ink4a was significantly derepressed after a few rounds of proliferation (Fig. 4g-i in this revised manuscript), which was similar to those ex-vivo cultured Slug -/myoblasts. These cells had impaired self-renewing capability, and were in senescence ( Fig. 5f-h) and irreversible quiescence (Fig. 5i,j) after first muscle regeneration. Indeed, derepression of p16 Ink4a has been reported to provoke defective self-renewal in hematopoietic (Smith et al., 2003 Mol Cell 12 (2) We also applied tamoxifen-induced adult SC-specific SlugKO mice to confirmed conclusions based on the Slug cKO mice (Supplementary Fig. 3) in this revised study. -Victor P et al., 2014), mainly though p16Ink4a. Therefore, does loss of Slug in young cells make them like aged/geriatric SCs?

Fourth, based on their title a reader would expect more figures related to SCs self-renewal decline during aging. Only the final figure attempt to connect Slug, self -renewal and aging. During aging, the number and function of muscle SCs decline. Recently, it has been showed also that maintenance of quiescence in adult mouse life relies on active repression of senescence pathway (Sousa
Answer: This is an excellent question. Many of previous studies from others had already demonstrated that aged SCs are characterized by self-renewal defect (Shefer et al., 2006 Dev Biol 294 (1) (7420):355-60). In addition to evidences including reduced Slug and increased p16 Ink4a expression in aged SCs (Fig. 7a-d), we further demonstrated that Slug null SCs resemble aged SCs in both similarly altered metabolic reprogramming and cell cycle regulator signatures ( Supplementary Fig. 12 in this revised manuscript), indicating great similarity in intrinsic propensities of Slug -/-SCs and aged SCs.

Figure 2A: Rather than having SCs compared to DN cells, which are almost unknown population, it will be more informative and relevant to compare population previously characterized such as FAPS, CD31 endothelial cells (where the role of Slug is well characterized) or blood CD45+ cells.
Answer: Following the Reviewer's suggestion, we compared Slug expression in endothelial cells (CD31 + ), pan-lymphocytes (CD45 + ), fibro-adipogenic progenitors (CD31 -CD45 -Scal1 + ) and satellite cells (CD1 -CD45 -Scal1 -Vcam I + ) in this revised manuscript (Fig. 2a).

Figure 2B: Please quantify slug expression in activated SCs, for example 2-3 days in proliferation conditions. If Slug is increased in proliferating cells, how would this change the authors conclusions?
Answer: According to the Reviewer's suggestion, we quantified Slug expression in freshly isolated quiescent SCs (QSC), activated SCs (ASC) upon culture in growth medium for 3 days, as well as the fully differentiated myotubes (MT) (Fig. 2b in this revised manuscript). Our data showed that compared to quiescent SCs, Slug expression was slightly decreased in activated SCs upon culture for 3 days and fell into a very low level in fully differentiated myotubes. Our result indicated that Slug is not increased in proliferating SCs. Fig. 2f, all Pax7 + SCs in Ctrl mice (Slug fl/+ Pax7 Cre/+ ) expressed concomitantly Slug, while all Pax7 + SCs from Slug cKO mice (Slug fl/fl Pax7 Cre/+ ) were negative for Slug staining. However, when assessing efficiency of tamoxifen-induced Slug knockout in adult SCs of Ctrl (Slug fl/+ Pax7 CreER ) and cKO (Slug fl/fl Pax7 CreER ) mice, quantification of Slug + Pax7 + SCs was provided (Supplementary Fig.  3c in this revised manuscirpt).

Answer: As shown in
According to the Reviewer's suggestion, we quantified Pax7 + SCs by staining resting adult skeletal muscle sections from Ctrl and Slug cKO mice for Pax7 and MyoD (Supplementary Fig. 4a,b in this revised manuscript). We apologize for limiting condition of the fluorescent microscope in our lab, which has only a maximum of three fluorescence channels. Therefore, Laminin was not co-stained in these sections. But based on all other IHC results staining Pax7 and Laminin in current study, all Pax7 + SCs were located beneath basal lamina, a classical SC anatomical location ( Fig. 3b and Supplementary  Fig. 5e in this revised manuscript). We hope that our answers could satisfy the Reviewer's comments.  Fig. 4e-g). We detected a 4-fold increase in the number of SA-β-Gal + cells in transverse sections of Slug cKO TA on day 10 post 1 st injury ( Fig. 5g in this revised manuscript). Pax7 and Ki67 co-immunostaining result demonstrated that these SA-β-Gal + cells were positive for Pax7 but negative for Ki67 staining (Fig. 5h in this revised manuscript), suggesting a status of senescence in SCs.
According to the Reviewer's suggestion, we tested the re-activation capacity of Pax7 + SCs in Ctrl and Slug cKO mice after first round of muscle regeneration. The results showed that most majority of SCs in Slug cKO mice failed to re-activate as indicated by a markedly lowered percentage of Ki67 + SCs on day 2.5 post 2 nd BaCl 2 injury comparing to Ctrl mice (Fig. 5i, j in this revised manuscript). Aswer: Thank the Reviewer for this excellent suggestion. We re-evaluated the gene enrichment analysis results, and demonstrated that there was a switched metabolic reprogramming with relatively higher energy-consuming status in Slug null SC as indicated by enriched glycolysis and oxidative phosphorylation pathways (Supplementary Fig. 7a,b in this revised manuscript).
These results indicate that Slug null might disturb balance between self-renewal and differentiation after activation. Indeed, it was previously demonstrated that the mitochondrial-associated metabolism pathway is more silent in Pax7 Hi SCs being of higher level of stemness and responsible for self-renewal (Rocheteau et al., 2012 Cell 148, 112-125). In addition, a more recent study using in vitro SC culture model showed that enhanced oxidative phosphorylation negatively affects the return to quiescence of activated SCs (Theret et al., 2017EMBO J 36, 1946-1962.

Figure 4H: Please show absolute value of mRNA analysis for p16ink4a level during quiescence and activation in vitro and in vivo in Slug wt versus Slug null. Moreover, a time course analysis of Slug expression during activation in relation to p16 is recommended.
Answer: We would like to interpret this question as that the reviewer wants to see the fold change of p16 Ink4a before and after SC activation both in vitro and in vivo. We still presented the data as relative expression by 2 -ΔΔCt method in this revised manuscript, but p16 Ink4a expression in all other groups were presented in a form of fold change comparing to that of in quiescent SCs of wildtype mice (Fig. 4g). Per suggested, we also performed a time course expression analysis of both Slug and p16 Ink4a in ex-vivo cultured myoblasts (Supplementary Fig. 10 in this revised manuscript).

Figure 4I: Please stain for laminin, it is important to know where Pax7+SCs/P16+ cells are located.
Answer: It has been shown that most majority of the Pax7 + SCs should have returned to native stem cell niche 30 days after injury (Comprehensive Physiology 2015, 5:1027-1059). Our new data also showed that activated SCs returned to the niche 30 days after transplantation (Supplementary Fig.  5e). Our previous Figure 4I (Fig. 4h in this revised manuscript) was mainly performed to ask whether p16 protein was elevated in Pax7+ SCs after first round of injury.

Figure 5 A: Please stain and quantify terminally differentiated myotubes for Myosin Heavy Chain and reserve cells for Pax7. Is there a defect in reserve cell number? P16ink4a has been used as a marker of senescence, but not all senescent cells are p16ink4a+ (Rodier and Campisi, 2011). Please confirm that increased in senescence is related to increased p16ink4a+ cells. Please stain reserve cells after 7d subgrowth in vitro with p16 antibody.
Answer: Experiments assessing differentiation of Slug-deficient SCs was specifically performed elsewhere in this revised manuscript (Supplementary Fig. 4c-g). We did not compare the reserve cell number because we only used the "reserve cell" model to prove that Slug-deficient SCs acquire features of senescence during in vitro proliferation. SCs used for this experiment were sorted from Pax7-zsGreen transgenic mice. By day 21 of induced differentiation, equal number of Slug +/+ -and Slug -/--zsGreen + (indicating Pax7 + ) mononucleated cells were sorted again for subculture and SA-β-gal staining experiment.
According to the Reviewer's suggestion, we demonstrated that the increased senescence in Slugnull reserve cells after culture was related to increased p16 Ink4a expression by showing that Slug -/-p16 -/cells exhibited no senescence under the same condition (Fig. 6f,g in this revised manuscript). Per suggested, we also stained reserve cells with p16 antibody after 7-day subculture in vitro. It turned out that there was significantly higher proportion of p16 Ink4a+ cells in Slug -/reserve cell-derived progeny ( Supplementary Fig. 11 in this revised manuscript).

Figure 5H: Please quantify Pax7+ senescent cells after 10d post injury. Their title of figure 5 is misleading, because they did not look at self-renewal, except for the in vitro reserve cells assay. They observed entrance into cellular senescence under proliferative stress. In vivo self-renewal senescence has not been shown, please do that.
Answer: Figure 5h was to demonstrate that the SA-β-Gal + cells detected in Fig. 5f were Pax7 + SCs. The quantification of these senescent SCs was shown in Fig. 5g.
We apologize for the misleading description in title of Figure 5. Quiescent SCs activate and proliferate in both in vitro culture and in vivo injury conditions. What we had expected to summarize in this title was that lack of Slug facilitates entry of SCs into cellular senescence under proliferative pressure. To this end, we revised the title of Fig. 5 (in this revised manuscript) as 'Slug-deficient SCs acquire features of senescence during in vitro and in vivo proliferation'. Regarding in vivo self-renewal senescence, we performed SA-β-Gal and Pax7 staining on cryosections of TA muscles from Ctrl and Slug cKO mice on day 10 post 1 st injury ( Fig. 5f-h), when most majority of SCs has already self-renewed, and growth of new muscle fibers is very advanced (Dumont et al., 2015Comp Physiol 5:1027-1059. These results indicated senescence in self-renewed SCs in vivo. We further provided in vivo experimental evidence showing that self-renewed SCs in Slug cKO mice failed to re-activate upon second injury (Fig. 5i,j in this revised manuscript), indicating status of senescence and irreversible quiescence.

Figure 6A-B: After deletion of p16 (again at germline level) is senescence decreased? How many Pax7+ cells are also SAbgal+? Please quantify self-renewed cells on aged muscle section and SCs contribution to muscle fibers. SCs repopulation experiments: please specify if your donor SCs are from aged mice and include also young/adult mice as a control.
Answer: Yes, deletion of p16 is sufficient to significantly decrease senescent Slug -/-SCs (less than 1%) under proliferative stress (Fig. 6f,g in this revised manuscript). All the experimental mice used in Fig. 6 were actually young adult mice. Because loss of Slug caused regenerative and self-renewing defects in SCs in young adult mice. We demonstrated that removal of p16 could partially rescue Slug-deficiency caused defects (Fig. 6 in this revised manuscript).

Figure 7I Quantification of dystrophin+ fibers is needed. Is Slug overexpression impacting cell proliferation and reducing senescence? Provide experimental
Answer: According to the Reviewer's suggestions, we have quantified dystrophin + fibers (Figure 7j in this revised manuscript). We also performed gain-of-function studies and showed that Slug overexpression does not impact cell proliferation, but robustly suppressed p16 Ink4a expression in longterm cultured SCs (Figure 7f in this revised manuscript). As we have added in this revised manuscript, deletion of p16 Ink4a rescued the cellular senescence phenotype in Slug-null SCs under proliferative stress, , forced expression of Slug should be able to actively suppress cellular senescence. This is also supported by experimental data showing that forced expression of Slug could restore self-renewing capacity of long-term cultured SCs (Fig. 7g,h in this revised manuscript).