CD90 identifies distinct fractions of muscle stem cells with different modalities of activation and quiescence maintenance

  1. Eyemen Kheir
  2. Michela Libergoli
  3. Samuele Metti
  4. Jonas Brorson Jensen
  5. Jakob Wang
  6. Niels Jessen
  7. Francesca Florio
  8. Francesca Murganti
  9. Serena Di Savino
  10. Marzia Belicchi
  11. Chiara Villa
  12. Kristian Vissing
  13. Mayank Verma
  14. Lorenzo Giordani
  15. Paolo Bonaldo
  16. Yvan Torrente
  17. Jean Farup
  18. Stefano Biressi

Reviewed by

Read the full article See related articles

Discuss this preprint

Start a discussion What are Sciety discussions?

Listed in

Log in to save this article

Abstract

Stem cell transition from quiescence to activation is crucial to guarantee productive tissue regeneration. Here we show that CD90 diversifies quiescent muscle stem cells (MuSCs) in murine and human muscle into two subpopulations differing in the kinetics of activation, CD90 +ve MuSCs exhibiting a faster exit quiescence and predominating the initial phases of regeneration compared to CD90 -ve MuSCs. In the absence of injury, the CD90 +ve fraction is primed toward activation through an active CD90-AMPK axis but is maintained in quiescence through signals from the extracellular matrix. Our studies show that Collagen VI, which is preferentially expressed by CD90 +ve MuSCs, binds to the Calcitonin receptor and plays a role in this context. Moreover, while the number of CD90 +ve and CD90 -ve subpopulations is similar in healthy muscles, the CD90 -ve fraction predominates in the muscles of murine models of Duchenne and Ullrich congenital muscular dystrophies. These findings provide novel insights into the mechanistic determinants of MuSCs functional heterogeneity and have implications for understanding the stimulation of repair in dystrophic muscle.

Article activity feed

  1. Review Commons

    Note: This response was posted by the corresponding author to Review Commons. The content has not been altered except for formatting.

    Learn more at Review Commons

    Reply to the reviewers

    Reviewer #1 Major comments:

    1. Data demonstrated the statistical differences in MuSC behaviors between CD90+ve and CD90-ve cells. However, the difference is small. For example, it is unclear whether the minimal difference in CALCR expression level between CD90+ve and CD90-ve cells gives rise to any biological difference.

    We thank the reviewer for this thoughtful comment and agree that it is important to distinguish statistical significance from biological relevance. However, describing the differences as “small” is somewhat subjective, as it is not clear whether this refers to per-cell expression differences or the magnitude of downstream functional consequences. In the case of CALCR, even modest shifts in receptor abundance can be biologically meaningful in threshold-dependent ligand–receptor systems, and in our study, the CALCR difference is supported by concordant orthogonal readouts (flow cytometry/immunofluorescence and transcript levels, see revised Fig. 6 F-I) together with functional evidence showing differential sensitivity of CD90+ve versus CD90−ve MuSCs to Col6-mediated restraint (see Fig. 6 D-E). In addition, to directly address the Reviewer’s question, we now include in this revised version of the manuscript further experimental data showing a differential response between CD90+ve and CD90−ve MuSCs to pharmacological inhibition of the CALCR signaling pathway (see revised Fig. S9G, H, I). Taken together, these complementary molecular and functional findings indicate that, although the quantitative differences may not appear large in absolute terms, they are sufficient to generate consistent, measurable, and biologically meaningful outcomes.

    1. Negative controls of FACS analyses are required because different sizes of cells might exert different background intensities. (Figure 2I, 2L, and 6F).

    We have now added negative controls to Fig. 2L and 6F, demonstrating the specificity and size independence of the measured signal. Panel 2I, which highlighted size differences between CD90+ve and CD90−ve MuSCs in the previous version, is now relocated to Fig. S3J.

    1. If CD90+ve MuSCs express Col6 higher than CD90-ve MuSCs, they should also highly express the primary target of Notch target genes, Hes1, Hey1, and HeyL. The authors should examine the expression levels of these genes.

    We have now expanded Suppl Table 1 to include all genes upregulated in CD90+ve MuSCs with p-values

    1. As described above, the quantifications of many results, including MyoD, were based on the fluorescent intensity. I know the difficulty of preparing enough cells for experiments, but the authors need to present data supporting these results.

    The evaluation of fluorescent intensity as a reliable and sensitive readout of protein content has been used in multiple publications in the myogenesis field by independent authors and by us (de Morree et al. 2019; Florio et al. 2023; Vetter and Lawlor 2026; Zanotti et al. 2022). The availability of sufficient amounts of materials is an important limitation for proteomic studies, and we thank the Reviewer for acknowledging this. To demonstrate the reliability of our pixel quantification-based assay, we have confirmed selected datasets using an alternative quantification approach based on visual discrimination of MyoD+ve and MyoD-ve cells (see Fig. S3E and associated Fig. S3D in the revised version). The results we obtained confirm those from pixel-quantification. Moreover, to further corroborate the trustworthiness of our analytical approach based on pixel quantification, we performed in parallel western blot analysis and the evaluation of fluorescent intensity on proliferating and differentiating myoblasts using antibodies recognizing MyoD, one of the markers we used throughout the manuscript. The data derived from this parallel evaluation (see Additional Fig. 1 below) clearly demonstrate robust parallelism between the two quantification methods.

    1. Figure 7G-H; More quantitative analyses should be included. In addition, the sample number was different between Fig7E and H. There is no significant difference in the CD90 expression in Fig7G. The authors need to confirm the reproducibility.

    We thank the reviewer for the opportunity to clarify these points. We apologize for not having clearly explained the design of this experiment. Figure 7E refers to a different time-point than Fig. 7F-I (i.e., 1.5 vs. 4.5 days after injury; see Fig. 7D for a visualization of the experimental design). The analysis at 1.5 days post injury was performed in 3 independent biological replicates. The difference in sample numbers between Fig. 7F and Fig. 7H is due to the fact that the absolute cell count was not performed in one biological replicate of CD90-depleted muscle and one biological replicate of control muscle. As a result, those samples were included in the qualitative and morphometric analyses but could not be incorporated into the cell number quantification. Importantly, this discrepancy does not affect the overall outcome or statistical interpretation of the results. We have clarified this in the revised figure legends. Regarding Fig. 7G, we would like to specify that CD90 staining is not shown in this panel. We really apologize for the confusion. The images display laminin and embryonic MyHC (eMyHC) staining to highlight the size and regeneration status of newly formed myofibers. Therefore, CD90 expression in this panel is not relevant to the analysis presented. We have revised the legend to explicitly clarify this point and avoid potential confusion. Overall, we confirm the reproducibility of the findings and have improved the clarity of presentation in the revised manuscript.

    Minor comments:

    1. Figure S4. The authors need to show evidence that these cells are proliferating. Without the evidence, CD90 expression my just be retained in non-dividing cells. If it is difficult, the results should be removed.

    We thank the reviewer for highlighting this possibility, and we have now removed panels D and E from Figure S4.

    1. Heterogeneity in cell cycle progression in MuSCs is well documented as fast and slow dividing cells. This reviewer recommends discussing the relevance of CD90 expression to these reports. PMID: 22349695 PMID: 8608871

    In the revised version of the manuscript, we have included the indicated reports in a paragraph of the discussion centered on a possible “division of labor” between CD90+ve and CD90-ve subsets of MuSCs during muscle regeneration. See also the response to the first major point raised by Reviewer #2.

    Reviewer #2

    Major Comments:

    1. It is perplexing that the CD90+ fraction is implicated in activation, proliferation, and differentiation (Mgn+ data) while simultaneously contributing to the CD90-ve population (Fig. 3E). However, the reverse does not seem to occur, with CD90-ve cells not replenishing the CD90+ fraction. If the CD90+ subpopulation indeed accounts for the majority of myogenesis, this provokes the question: what is the functional role of the CD90− fraction? Notably, CD90-ve MuSCs appear to divide effectively during regeneration (Fig. 2E-G), further emphasizing the need to clarify their contribution to the overall regenerative process. The presence of a substantial number of CD90-ve MuSCs across conditions suggests they cannot simply be dismissed as irrelevant and understanding their role will help clearly establish the +/- subpopulations as functionally different.

    We thank the reviewer for raising this important point. We would like to clarify that we do not suggest at any stage that the CD90−ve MuSC population is insignificant or dispensable for regeneration. On the contrary, our data consistently show that CD90−ve MuSCs are numerically substantial across homeostatic and regenerative conditions and retain clear proliferative capacity during muscle repair (Fig. 2E-G). We fully agree that their persistence strongly argues for a functional role. Our study was designed primarily to identify and characterize functional differences between CD90+ve and CD90−ve MuSCs in terms of activation dynamics and quiescence control, rather than to comprehensively define the lineage hierarchy or long-term fate of each subpopulation. In this context, the observation that CD90+ve MuSCs can give rise to CD90−ve cells in vitro (Fig. 3E) suggests a degree of plasticity and may indicate that CD90 expression marks an activation-prone state rather than a rigid lineage boundary. The lack of reciprocal conversion under the tested conditions does not imply that CD90−ve cells are functionally inert, but rather that the two fractions may occupy distinct positions along a continuum of activation states. Importantly, our in vivo data demonstrate that CD90−ve MuSCs do enter the cell cycle during regeneration, albeit with slower kinetics compared to CD90+ve cells. This supports a model in which CD90+ve cells are primed for rapid early activation, while CD90−ve cells may represent a more dormant or reserve-like fraction that contributes to regeneration with delayed kinetics or plays a stabilizing role during regeneration. Such division of labor would be consistent with emerging concepts of functional heterogeneity within stem cell compartments. Similar “division-of-labor” models have already been proposed in other stem cell systems, including muscle, where subsets differ in proliferation kinetics, differentiation, or self-renewal behavior, as well as in other tissues. We have included a dedicated paragraph in the discussion to highlight this aspect in light of classical and recent literature. A detailed dissection of the long-term lineage contribution and self-renewal capacity of the CD90−ve fraction would be highly informative; however, addressing this question would require dedicated clonal tracing and transplantation experiments beyond the scope of the present study. We have now clarified this point in the revised Discussion, explicitly stating that our goal is to highlight differential activation modalities between the two subpopulations rather than to assign exclusive regenerative responsibility to one fraction. Taken together, our findings support the view that CD90+ve and CD90−ve MuSCs represent functionally distinguishable, yet complementary, subpopulations within the muscle stem cell pool, rather than hierarchically “major” versus “minor” or “relevant” versus “irrelevant” fractions.

    1. The depletion of CD90+ cells (Fig. 7D-I) is the correct experimental approach to assess the function of these cells in vivo. However, the method employed, using IP injections of a CD90 antibody, can lack specificity. Even with optimal specificity, CD90 is expressed on numerous cell types across the body. This raises the possibility that observed effects may result from targeting other CD90+ cells in skeletal muscle or other tissues, both locally and systemically. To mitigate these confounding factors, the authors should attempt strategies to reduce off-target effects. While the technical challenges are acknowledged by this reviewer and may be prohibitory, addressing these limitations would substantially enhance the impact of this work. Additionally, the embryonic myosin heavy chain (eMHC) images (Fig. 7G, H) should be more representative of the quantification data to ensure consistency.

    We thank the reviewer for this constructive comment and agree that antibody-mediated depletion strategies may raise concerns regarding specificity. As correctly pointed out, CD90 is expressed in additional cellular compartments beyond MuSCs, a limitation that we have explicitly acknowledged in the revised manuscript. Importantly, the anti-CD90 antibody used in this study is highly specific, as validated by flow cytometry and immunofluorescence analyses (see Fig. S1 and Fig. 1H). Moreover, the same clone (30-H12) has been previously employed by other groups for in vivo depletion approaches with comparable experimental aims, supporting its reliability for targeting CD90+ve cells (Powell et al. 2012; Zhou et al. 2022). While we cannot completely exclude effects on other CD90-expressing cells, our depletion strategy was performed in the context of acute muscle damage, with local intramuscular administration at the time of injury following systemic priming, which may partially limit potential broader systemic confounding effects. The timing of the phenotype - restricted to early regenerative stages - argues in favor of a local MuSCs-related contribution. We agree that genetic or lineage-restricted strategies would provide a more selective approach; however, such models are currently unavailable for selectively targeting CD90+ve MuSCs without affecting other CD90-expressing populations. Finally, as requested, we have replaced the representative eMHC images in Fig. 7G and 7H to better reflect the quantification data and ensure improved consistency between images and measured outcomes.

    1. Similar concerns about off-target effects noted in point #2, apply to the use of the Col6 KO mouse model, which appears to be a full body KO, meaning Col6 is absent not only in MuSCs but also in other cell types that typically express Col6. This deficiency would have been present throughout development, complicating the interpretation of the observed effects. The authors do acknowledge Col6 expression by non-MuSC cell types, but the in vivo impact remains challenging to interpret, particularly due to the potential developmental and systemic effects of removing Col6. Also, the observation that the CD90-ve subpopulation still expresses Calcr raises further questions about Col6 acting only on the CD90+ fraction and expression by MuSCs being consequential in vivo. The trend observed in Fig. 6M for CD90-ve cells suggests that this mechanism might not be exclusive to CD90+ cells, warranting further investigation or explanation since an outlier in the Col6KO CD90-ve group may have influenced interpretation.

    We thank the reviewer for this thoughtful comment and agree that the use of a constitutive Col6a1−/− model introduces interpretative limitations. As correctly noted, Col6 is expressed by multiple cell types, and its absence throughout development may potentially result in niche-level and systemic effects that complicate attribution of the phenotype exclusively to MuSC-derived Col6. We have clarified this point in the revised Discussion and tempered our conclusions accordingly. Importantly, we do not propose that the Collagen VI - CALCR axis operates exclusively in the CD90+ve fraction. CALCR is also expressed, although at lower levels, in CD90−ve MuSCs, and the in vivo data in Col6a1−/− mice indicate that both subpopulations are affected. Indeed, parameters such as MyoD modulation shift in the same direction in both fractions (see Fig. 6L), supporting the idea that this signaling axis is functionally active across the MuSC pool. However, several observations indicate that components of the Col6–CALCR axis are more pronounced in CD90+ve MuSCs and that certain responses are more robust in this fraction. These include higher Col6 and CALCR expression levels (Fig. 6A-C and 6F-I), a more pronounced increase in cell size upon Col6 ablation (Fig. 6K), and a clearer modulation of activation-associated readouts (levels of pAMPK and MyoD, and EdU incorporation) upon inhibition of the Col6-CALCR axis in vitro (see the newly introduced experiments with PKA inhibitor in Fig. S9 G-H of the revised version). Thus, rather than invoking strict differential sensitivity, our data support a quantitative model in which the pathway operates in both subpopulations but with greater amplitude or prominence in CD90+ve MuSCs. Regarding the trend observed in Fig. 6M for CD90−ve cells, we acknowledge that variability might have played a role and have revised the text to avoid overstatement. As mentioned above, we expanded characterization of the Col6-CALCR axis in the two subpopulations by performing additional ex vivo experiments to investigate the activation of the signaling pathway downstream of CALCR and the impact of its pharmacological inhibition on the two subpopulations of MuSCs. Overall, we conclude that the Collagen VI–CALCR axis is not exclusive to CD90+ve MuSCs, but that its components and functional consequences appear particularly evident in this fraction.

    1. The siCD90 experiment in Fig. 5 demonstrates effective KD at both the transcript and protein levels, but the observed impact on the proliferation of CD90+ cells (Fig. 5G), while statistically significant, appears to be less than expected. This result is also confusing given the substantial reduction in pAMPK levels observed in Fig. 5L, leading to the expectation of a more pronounced effect on proliferation if the proposed CD90-pAMPK mechanism is a driving pathway. Additionally, Fig. 5N suggests that pAMPK supports proliferation in both CD90+ and CD90− subpopulations. While the AICAR treatment in CD90− cells does not achieve significance, the data exhibit a bimodal distribution among replicates, with an apparent outlier in the control group potentially skewing the analysis. This variability necessitates further clarification for the relationship between CD90, pAMPK, and MuSC proliferation.

    We thank the reviewer for this careful evaluation of the siRNA and AICAR experiments. First, to improve clarity and better reflect the relationship between CD90 knockdown efficiency and biological outcome, we have now re-expressed the proliferation data in Fig. 5 F, G, L as percentage reduction relative to the CD90+ve scrambled control condition. Presenting the data in this normalized manner highlights a strong correlation between the extent of CD90 ablation (Fig. 5E) and the magnitude of the physiological effect. While the absolute change in proliferation may appear moderate, it scales consistently with the degree of CD90 protein reduction, supporting a dose-effect relationship rather than an all-or-none response. Our data support the interpretation that the CD90-AMPK axis contributes to, but does not solely determine, proliferative behavior. MuSC proliferation is likely governed by multiple converging pathways, and CD90-dependent modulation of AMPK represents an important component of this regulatory network. To further strengthen the mechanistic link, we have complemented the primary MuSCs data with gain-of-function experiments in C2C12 reserve cells (see Fig. S6 in this revised version). Overexpression of CD90 in this model enhances activation-associated features, including increased Myod upregulation, pAMPK levels, and augmented entry into the cell cycle upon stimulation, compared to control untransduced cells. These findings provide independent evidence that CD90 expression is sufficient to potentiate AMPK signaling and bias cells toward a more activation-prone state, supporting the causal nature of the CD90–AMPK axis beyond observations in primary MuSCs. Concerning the AICAR experiments (Fig. 5N), we acknowledge that in vivo pharmacological activation can be influenced by biological variability. It is indeed possible that AICAR exerts effects on both CD90+ve and CD90−ve populations, as AMPK is present in both fractions (Additional Fig. 2).

    1. The CD90 related findings in human samples appear less robust compared to those in mice. While the sorting successfully identifies sizable CD90+ and CD90-ve populations (Fig. 4A), the sequencing data show only small regions of high CD90 expression, as highlighted in red by the authors (Fig. 4C, D). Have the authors considered replicating the sequencing experiments within their own laboratory? While it is acknowledged that sourcing human tissue may be a limitation, it may strengthen the translational impact if possible.

    We thank the reviewer for this thoughtful comment. We agree that the CD90-related signal in the human scRNA-seq dataset appears less striking than in the murine Cy-TOF data; however, we believe that the most parsimonious explanation lies in the well-documented technical limitations of single-cell RNA sequencing, particularly its limited sensitivity for low-abundance transcripts. It is widely recognized that in scRNAseq experiments, the number of genes detected depends on the sequencing depth, and therefore scRNA-seq suffers from “dropout” effects and reduced detection efficiency, especially for transcripts expressed at moderate-to-low levels, such as those that do not encode abundant structural proteins (Kharchenko et al. 2014; Svensson et al. 2017). CD90 falls within this category, as it is not an abundant structural protein and may therefore be underrepresented at the mRNA level despite robust protein detection with FACS. Indeed, discrepancies between protein-level heterogeneity and scRNA-seq signal intensity are commonly reported, particularly for surface markers (Linderman et al. 2022; Stuart and Satija 2019). Importantly, in our study, the presence of substantial CD90+ve and CD90−ve human MuSC populations is robustly demonstrated by flow cytometry-based sorting (Fig. 4A-B), which directly measures protein abundance and shows a clear bimodal distribution. The scRNA-seq data were used as supportive, orthogonal evidence and are consistent with enrichment of CD90-expressing clusters, even if the signal is spatially restricted. Furthermore, functional assays (ex vivo EdU incorporation and activation parameters following injury) independently validate that CD90 marks a functionally distinct fraction in human muscle. While we agree that performing scRNA-seq in-house would be valuable, access to freshly isolated human MuSCs in sufficient numbers for high-depth single-cell sequencing remains technically and ethically challenging. Given that our conclusions rely primarily on protein-level stratification and functional validation, we believe that the translational relevance of the findings is adequately supported, and we hope the Reviewer will agree.

    Minor Comments:

    1. Fig. 1D - the MuSC population has an uncharacteristically low representation amongst cells of uninjured muscle. Can the authors comment on this in text?

    We thank the reviewer for raising this point. We have re-examined our calculations on a per-sample basis, and the proportion of MuSCs among total mononuclear cells isolated from uninjured muscle ranges between approximately 0.8% and 3%. This frequency is within the lower end of the range typically reported for MuSCs isolated by FACS from adult uninjured murine muscle, which is commonly described to fall around ~1-4% depending on digestion protocol, gating strategy, and muscle type (Liu et al. 2015; Machado et al. 2017; Montarras et al. 2005). Importantly, we intentionally applied a conservative gating strategy to minimize contamination from non-myogenic populations and to ensure that CD90 detection was strictly restricted to bona fide MuSCs. While this approach may reduce the apparent overall frequency of MuSCs, it increases confidence in the purity of the analyzed population and in the interpretation of CD90-based subfractionation. To improve clarity and transparency, we have now included an additional figure (Additional Fig. 3) below detailing the full gating scheme, along with absolute numbers and percentages of MuSCs across samples. We have also added explanatory text in the legend of Figure 1D of the revised manuscript to explicitly address this point.

    1. Fig. 2 - the g-alert phenotype corresponding with CD90 expression is interesting. Can the authors add a molecular marker to confirm this phenotype?

    We have now added to our size, rosa locus activity, and mitochondrial content analysis the quantification of phosphorylated S6, a marker of cells in the alert state according to preexisting literature (Rodgers et al. 2014) (see Fig. 2J of the revised version). The more pronounced presence of phospho-S6 in CD90+ve MuSCs under “alerting” conditions supports our conclusion that CD90+ve MuSCs present a more rapid tendency to enter the G0alert state compared to their CD90-ve counterpart.

    1. The authors mention "significantly higher fraction of CD90+ve MuSCs incorporated EdU in vivo at 1.5, 2.75, and 6 days after injury..." in 2G, but it only seems 1.5 and 2.75 are different. The text should be corrected.

    We apologize for the mistake and thank the Reviewer for noticing this typo, which we have corrected in the current version of the manuscript.

    1. 3B, C - 7dpi seems late for the analyses of Myogenin at a single-cell level since most differentiating MuSCs are fused at this point. Can the authors comment in the text why 7 dpi was chosen?

    In adult murine skeletal muscle, most of the events associated to the regenerative process following acute injury typically unfolds over approximately 14 days, with early activation and proliferation occurring within the first few days, followed by differentiation and fusion, and progressive maturation of newly formed fibers until ~30 days post-injury. Thus, 7 days post-injury represents an intermediate stage of the regenerative timeline, during which differentiation is still ongoing and newly formed fibers are not yet fully matured. At this time point, Myogenin-positive cells can still be detected, reflecting ongoing differentiation of MuSC progeny before complete fusion and maturation. Told that, the major goal of this experiment was to begin exploring potential mechanisms restoring the initial 1:1 ratio between CD90+ve and CD90-ve MuSCs after injury resolution. Our observations reported in Fig. 3A suggest that ~6 days post-injury, the reequilibration process has started. The choice of 7 days post-injury for the myogenin analysis is a consequence of this observation. Moreover, around this time-point, CD90+ve and CD90-ve MuSCs showed similar EdU incorporation rates (see Fig. 2G). We have clarified this rationale in the revised manuscript to better contextualize the timing of the analysis.

    1. Note - 7L is very interesting, but emphasizes the major Q#1 above. Mdx mice are known to lose MuSC capacity due to continuous rounds of proliferation. So, the CD90+ loss is supportive, but there are still many CD90-ve cells present. In other figures, the authors demonstrated the negative fractions still harbor decent proliferative potential, so why is there no rescue?

    We thank the reviewer for highlighting this important aspect of our findings and for recognizing the relevance of the observation in Fig. 7L. We agree that in mdx mice, where MuSCs undergo repeated cycles of activation and proliferation, the preferential loss of the CD90+ve subset is consistent with the idea that this subpopulation may be particularly vulnerable to chronic regenerative stress. At the same time, we acknowledge that a substantial number of CD90−ve MuSCs remain present. As the Reviewer correctly notes, our data indicate that CD90−ve cells retain proliferative potential under acute regenerative conditions. However, our study was designed primarily to define differences in activation dynamics and quiescence control between CD90+ve and CD90−ve MuSCs, rather than to establish their relative capacity to sustain long-term regeneration in pathological contexts. Understanding why the remaining CD90−ve population does not compensate for the loss of CD90+ve cells in mdx muscle would require dedicated lineage-tracing, transplantation, or long-term functional assays, which go beyond the scope of the present work. One possible interpretation, consistent with our data, is that CD90−ve MuSCs exhibit slower activation kinetics and may not efficiently support the rapid or repeated regenerative demands characteristic of dystrophic muscle. Thus, they may be insufficient to fully rescue regeneration under chronic pathological stress. Future studies specifically addressing the regenerative potential of purified CD90−ve MuSCs in transplantation or chronic injury models will be required to resolve this question in detail. We have clarified this point in the revised Discussion and tempered our interpretation accordingly.

    1. The authors should clearly state the number of cells and number of replicates for their single cell distribution graphs in all legends.

    We have added this information to all relevant legends, where it was not already present.

    1. In their stats section of methods, ns= p{greater than or equal to}0.15, please clarify.

    We thank the reviewer for pointing this out and apologize for the lack of clarity. The threshold “ns ≥ 0.15” was introduced as an arbitrary and conservative criterion to avoid labeling as “non-significant” comparisons with p values only marginally above the conventional 0.05 threshold. Our intention was to distinguish likely truly non-significant results (p ≥ 0.15) from those showing a statistical trend (0.05

    Reviewer #3

    Major comments

    1. While the manuscript provides valuable insights into the functional heterogeneity of MuSCs, there are some critical aspects that remain unclear. Specifically, the mechanism by which the quiescence of CD90+ MuSCs, maintained through the COL6-CALCR pathway, confers an advantage for their rapid activation is not sufficiently addressed. Understanding why this pathway enables such responsiveness would significantly strengthen the authors' conclusions. Additionally, the manuscript does not elucidate how the quiescence of CD90-negative MuSCs is maintained, leaving a gap in the characterization of MuSC heterogeneity. Without this information, the functional significance of this heterogeneity, particularly in the context of muscle regeneration, remains incomplete.It would be interesting to explore why CD90-negative cells appear less responsive to injury or why CD90+ cells are more readily activated. Addressing these questions would provide a more comprehensive understanding of the biological implications of MuSC heterogeneity and enhance the impact of the study.

    We thank the reviewer for this thoughtful and conceptually important comment. We agree that clarifying the mechanistic basis underlying the differential activation propensity of CD90+ve and CD90−ve MuSCs strengthens the interpretation of functional heterogeneity. To address this point, we have expanded the mechanistic component of the study in two directions. First, to better understand why CD90+ve MuSCs display a more pronounced activation profile, we performed gain-of-function experiments in C2C12 reserve cells by overexpressing CD90. These experiments, reported in Figure S6 of this revised version, demonstrate that CD90 overexpression enhances activation-associated features, supporting a causal link between CD90 expression and activation propensity. This complements the loss-of-function data in primary MuSCs and reinforces the concept of a CD90-AMPK axis contributing to a primed metabolic state. Second, regarding quiescence, we have further dissected the COL6-CALCR pathway by analyzing downstream signaling components and comparing its mechanistic features with the previously described COL5-CALCR axis (Baghdadi et al. 2018). Our new data show that COL6 engagement leads to modulation of downstream effectors (e.g., YAP localization and PKA-dependent signaling), consistent with a CALCR-mediated quiescence program that shares similarities with COL5-driven regulation (see Fig.S9 of the revised version). We have expanded the Discussion to more clearly articulate this mechanistic convergence. Importantly, we do not propose that COL6-mediated quiescence directly “confers” activation capacity in a deterministic sense. Rather, our model suggests that CD90+ve MuSCs exist in a poised state: they exhibit an intrinsically primed activation program (via CD90-AMPK), while concurrently maintaining quiescence through a COL6-CALCR-dependent restraint. This dual regulatory architecture may allow rapid transition upon injury without premature exhaustion, thereby providing a kinetic advantage. In this context, the destruction of muscle extracellular architecture associated with injury would release the “break” imposed by Collagen 6 on the activation of CD90+ve cells. Regarding CD90−ve MuSCs, we acknowledge that the precise mechanisms maintaining their quiescence remain incompletely defined in this study. However, our data suggest that they are less reliant on the COL6–CALCR axis. A full dissection of these pathways would require dedicated transcriptional and signaling analyses beyond the scope of the present work. We have clarified this aspect in the revised Discussion. Finally, we have further elaborated in the Discussion that CD90+ve and CD90−ve MuSCs may represent functionally complementary subpopulations: CD90+ve cells being primed for rapid early activation, and CD90−ve cells potentially contributing under different temporal or regenerative contexts. We believe that these additions provide a more comprehensive framework for understanding the biological implications of MuSC heterogeneity while maintaining appropriate caution regarding unresolved mechanistic aspects.

    1. I am particularly concerned about Figure 7. As the authors mentioned, CD90 is not specific to MuSCs. Therefore, the conclusion that CD90+ MuSCs are important for muscle regeneration based on the current experiment is not fully convincing. I suggest incorporating additional approaches to confirm this point. For example, transplantation of CD90+ or CD90- MuSCs into injured muscles would provide stronger support for their findings.

    We thank the reviewer for raising this important concern, which was also emphasized by Reviewer #2. We fully agree that the lack of absolute specificity of CD90 for MuSCs represents a limitation when interpreting in vivo depletion experiments. A number of observations somehow mitigate the concern (see answer to point #2 of Reviewer 2). At present, however, there are no available genetic tools that would allow selective targeting of the CD90+ve MuSC subpopulation. Addressing this question definitively would likely require the generation of compound mouse models combining at least two independent genetic modifications (e.g., a MuSC-specific driver together with a CD90-dependent conditional ablation system), which are currently not available. We have explicitly clarified this limitation in the revised manuscript. Regarding the suggestion of transplantation experiments, we agree that this approach is often used to assess regenerative potential. However, in the specific context of our study, transplantation may not directly resolve the key mechanistic question we are addressing. Indeed, isolation and transplantation procedures inevitably activate MuSCs due to enzymatic digestion and removal from their niche, thereby erasing differences related to activation kinetics and quiescence maintenance. Since the central focus of our work is the differential propensity for activation and the regulation of quiescence between CD90+ve and CD90−ve subpopulations, transplantation of already activated cells may obscure precisely the phenotypic differences we aim to characterize. Importantly, our conclusions do not rely solely on the depletion experiment in Fig. 7. The functional relevance of CD90+ve MuSCs is supported by multiple complementary lines of evidence, including differences in activation kinetics, AMPK signaling, response to Collagen VI, and their preferential depletion in dystrophic muscle. The in vivo antibody-mediated depletion, therefore, serves as supportive, rather than exclusive, functional validation. We have revised the Discussion to explicitly acknowledge these technical constraints (and, therefore, temper our conclusions), while emphasizing that the available evidence supports a functional contribution of the CD90+ve fraction to early regenerative dynamics.

    1. A previous study has demonstrated that the COL5-CALCR pathway is essential for maintaining MuSC quiescence. In this manuscript, the authors propose the COL6-CALCR pathway; however, the current study lacks specific experiments to clarify the differences and similarities between these pathways. Additionally, the discussion section does not adequately address these points, leaving the interpretation incomplete. A more thorough discussion comparing these pathways would significantly improve the manuscript.

    We thank the reviewer for highlighting this important conceptual point. We agree that a clearer comparison between the previously described Col5-CALCR pathway and the Col6-CALCR axis proposed in our study strengthens the interpretation of our findings. To directly address this issue, we have now included two additional sets of experiments in the revised manuscript. First, we performed a comparative analysis of Col5, Col6, and Col4 as substrates, evaluating their ability to activate downstream CALCR signaling, using the reduction of nuclear YAP accumulation as a functional readout (Zhang et al. 2019). Although only the effect induced by Col6 was statistically different from those induced by Col 4, these experiments suggest that both collagen 5 and 6 can activate this pathway (see Fig. S9E-F of the revised manuscript). Second, we assessed activation and proliferation parameters in freshly isolated CD90+ve MuSCs plated on Col5, Col6, or Col4 substrates. This allowed us to directly compare the functional consequences of these different ECM components on activation kinetics and proliferative behavior within the same experimental framework. The results indicate overlapping effects of Col6 and Col5, distinct from those induced by Col4. These observations support the idea that Col6 contributes to quiescence regulation in a manner that is at least partially convergent with the previously described Col5 pathway (see Fig S9 A-B of the revised manuscript). Importantly, the commune effect induced by Col6 and Col5 appears to be specific, as Col1 behaves similarly to Col4 under similar testing conditions (see Fig S9 C-D of the revised manuscript). In addition to incorporating these new data, we have substantially expanded the Discussion to more thoroughly compare the COL5-CALCR and COL6-CALCR axes, emphasizing both shared mechanisms (CALCR engagement and quiescence modulation) and potential differences in expression patterns, subpopulation bias, and magnitude of response. We believe that these additions significantly clarify the relationship between the two pathways and strengthen the overall mechanistic framework of the manuscript.

    Minor comments

    1. It would be interesting to see the spatial localization of CD90+ and CD90- MuSCs in skeletal muscle tissue.

    We would like to point out that the spatial localization of CD90+ve and CD90−ve MuSCs within skeletal muscle tissue is already shown in Fig. 1H and Fig.6B, where immunofluorescence analysis of muscle cryosections demonstrates the presence of both subpopulations in their native niche. In these images, CD90 staining is visualized in combination with MuSC markers, allowing identification of CD90+ve and CD90−ve MuSCs in situ. We have added to this response to Reviewers additional examples, in which Col6 staining is also highlighted (Additional Fig. 4). To improve clarity, we have revised the legend for Fig. 1H to more explicitly highlight this aspect and guide the reader to the relevant panel.

    1. I suggest conducting a more thorough investigation to characterize the quiescent CD90+ and CD90- MuSCs, particularly focusing on aspects such as protein translation machinery.

    We thank the reviewer for this insightful suggestion. We agree that a deeper characterization of quiescent CD90+ve and CD90−ve MuSCs, including analysis of protein translation machinery and related metabolic features, would provide valuable additional insight into the molecular basis of their functional differences. However, such an in-depth investigation would require dedicated molecular investigations, such as proteomic and/or ribosome profiling approaches, and goes beyond the scope of the present study, which is focused primarily on differential activation dynamics and quiescence regulation between the two subpopulations. We believe this represents an important and promising direction for future work.

    1. The authors should include a discussion of previously identified markers of MuSC heterogeneity, such as CD34, to provide better context for their findings.

    We thank the reviewer for this helpful suggestion. We agree that placing our findings in the context of previously described markers of MuSC heterogeneity is important. In the manuscript, we have explicitly evaluated the relationship between CD90+ve and CD90−ve MuSCs and previously reported heterogeneity markers, including CD34. These analyses are presented in Fig. S2, where we show that CD90-based stratification does not simply recapitulate previously defined subsets. For clarity, we have also added a new summary table (Additional Table 1 below) highlighting the limited overlap between CD90-defined fractions and other reported markers of MuSC heterogeneity. Furthermore, we have expanded the Discussion to note that CD90 does not align with markers such as CD34 and others described in the literature, emphasizing that CD90 identifies a functionally distinct layer of heterogeneity, primarily related to activation kinetics and quiescence regulation, rather than directly overlapping with previously characterized subpopulations.

    1. In Figure 6, the authors used COL4 as a negative control; however, this is insufficient to conclusively demonstrate the importance of COL6 in maintaining CD90+ MuSC quiescence. Including additional substrates beyond collagen, such as fibronectin or laminin, along with COL5, would strengthen the conclusions drawn from these experiments.

    We thank the reviewer for this valuable suggestion. We agree that expanding the range of substrates strengthens the interpretation of the role of COL6 in regulating MuSC quiescence. In the revised manuscript, we have now included additional comparative experiments using Collagen V and Collagen I as alternative substrates. Collagen V, consistent with previous reports implicating the COL5–CALCR axis in MuSC quiescence, produced effects qualitatively similar to those observed with Collagen VI, supporting a partially convergent mechanism at the level of CALCR signaling (see also the response to major comment #3 above) (Fig. S9A-B of the revised version). In contrast, Collagen I was less effective at promoting quiescence-associated features in CD90+ve MuSCs, yielding results similar to Col4 in terms of EdU incorporation and expression of pAMPK (see Fig. S9C-D of the revised version). These findings reinforce the idea that Collagen VI (and Collagen V) are not merely generic ECM components, but exert specific regulatory effects on MuSC activation dynamics. We have incorporated these new data into Figure S9, where we have also created a graphical scheme to summarize and better contextualize similarities and differences among ECM substrates in shaping MuSC behavior (Fig. S9J).

    ADDITIONAL REFERENCES Almada AE, Horwitz N, Price FD, Gonzalez AE, Ko M, Bolukbasi OV, Messemer KA, Chen S, Sinha M, Rubin LL, et al. 2021. FOS licenses early events in stem cell activation driving skeletal muscle regeneration. Cell Reports 34: 108656. Baghdadi MB, Castel D, Machado L, Fukada S-I, Birk DE, Relaix F, Tajbakhsh S, Mourikis P. 2018. Reciprocal signalling by Notch-Collagen V-CALCR retains muscle stem cells in their niche. Nature 557: 714–718. Beauchamp JR, Heslop L, Yu DSW, Tajbakhsh S, Kelly RG, Wernig A, Buckingham ME, Partridge TA, Zammit PS. 2000. Expression of CD34 and Myf5 defines the majority of quiescent adult skeletal muscle satellite cells. Journal of Cell Biology 151: 1221–1233. Cerletti M, Jurga S, Witczak CA, Hirshman MF, Shadrach JL, Goodyear LJ, Wagers AJ. 2008. Highly Efficient, Functional Engraftment of Skeletal Muscle Stem Cells in Dystrophic Muscles. Cell 134: 37–47. Chakkalakal JV, Christensen J, Xiang W, Tierney MT, Boscolo FS, Sacco A, Brack AS. 2014. Early forming label-retaining muscle stem cells require p27kip1 for maintenance of the primitive state. Development (Cambridge, England) 141: 1649–59. Chakkalakal JV, Jones KM, Basson MA, Brack AS. 2012. The aged niche disrupts muscle stem cell quiescence. Nature 490: 355–360. de Morree A, Klein JDD, Gan Q, Farup J, Urtasun A, Kanugovi A, Bilen B, van Velthoven CTJ, Quarta M, Rando TA. 2019. Alternative polyadenylation of Pax3 controls muscle stem cell fate and muscle function. Science 366: 734–738. Der Vartanian A, Quétin M, Michineau S, Auradé F, Hayashi S, Dubois C, Rocancourt D, Drayton-Libotte B, Szegedi A, Buckingham M, et al. 2019. PAX3 Confers Functional Heterogeneity in Skeletal Muscle Stem Cell Responses to Environmental Stress. Cell Stem Cell 24: 958-973.e9. Florio F, Vencato S, Papa FT, Libergoli M, Kheir E, Ghzaiel I, Rando TA, Torrente Y, Biressi S. 2023. Combinatorial activation of the WNT ‐dependent fibrogenic program by distinct complement subunits in dystrophic muscle. EMBO Molecular Medicine 15: 1–20. García-Prat L, Perdiguero E, Alonso-Martín S, Dell’Orso S, Ravichandran S, Brooks SR, Juan AH, Campanario S, Jiang K, Hong X, et al. 2020. FoxO maintains a genuine muscle stem-cell quiescent state until geriatric age. Gayraud-Morel B, Chrétien F, Jory A, Sambasivan R, Negroni E, Flamant P, Soubigou G, Coppée J-Y, Di Santo J, Cumano A, et al. 2012. Myf5 haploinsufficiency reveals distinct cell fate potentials for adult skeletal muscle stem cells. Journal of cell science 125: 1738–1749. Guardiola O, Iavarone F, Nicoletti C, Ventre M, Rodríguez C, Pisapia L, Andolfi G, Saccone V, Patriarca EJ, Puri PL, et al. 2023. CRIPTO-based micro-heterogeneity of mouse muscle satellite cells enables adaptive response to regenerative microenvironment. Developmental Cell 58: 2896-2913.e6. Kharchenko PV, Silberstein L, Scadden DT. 2014. Bayesian approach to single-cell differential expression analysis. Nat Methods 11: 740–742. Kuang S, Kuroda K, Le Grand F, Rudnicki MA. 2007. Asymmetric Self-Renewal and Commitment of Satellite Stem Cells in Muscle. Cell 129: 999–1010. Linderman GC, Zhao J, Roulis M, Bielecki P, Flavell RA, Nadler B, Kluger Y. 2022. Zero-preserving imputation of single-cell RNA-seq data. Nature Communications 13: 1–11. Liu L, Cheung TH, Charville GW, Rando TA. 2015. Isolation of skeletal muscle stem cells by fluorescence-activated cell sorting. Nature Protocols 10: 1612–1624. Machado L, Esteves de Lima J, Fabre O, Proux C, Legendre R, Szegedi A, Varet H, Ingerslev LR, Barrès R, Relaix F, et al. 2017. In Situ Fixation Redefines Quiescence and Early Activation of Skeletal Muscle Stem Cells. Cell Reports 21: 1982–1993. Montarras D, Morgan J, Collins C, Relaix F, Zaffran S, Cumano A, Partridge T, Buckingham M. 2005. Direct isolation of satellite cells for skeletal muscle regeneration. Science (New York, NY) 309: 2064–2067. Powell N, Walker AW, Stolarczyk E, Canavan JB, Gökmen MR, Marks E, Jackson I, Hashim A, Curtis MA, Jenner RG, et al. 2012. The Transcription Factor T-bet Regulates Intestinal Inflammation Mediated by Interleukin-7 Receptor+ Innate Lymphoid Cells. Immunity 37: 674–684. Relaix F, Montarras D, Zaffran S, Gayraud-Morel B, Rocancourt D, Tajbakhsh S, Mansouri A, Cumano A, Buckingham M. 2006. Pax3 and Pax7 have distinct and overlapping functions in adult muscle progenitor cells. Journal of Cell Biology 172: 91–102. Rocheteau P, Gayraud-Morel B, Siegl-Cachedenier I, Blasco MA, Tajbakhsh S. 2012. A subpopulation of adult skeletal muscle stem cells retains all template DNA strands after cell division. Cell 148: 112–125. Rodgers JT, King KY, Brett JO, Cromie MJ, Charville GW, Maguire KK, Brunson C, Mastey N, Liu L, Tsai CR, et al. 2014. MTORC1 controls the adaptive transition of quiescent stem cells from G 0 to GAlert. Nature 510: 393–396. Scaramozza A, Park D, Kollu S, Beerman I, Sun X, Rossi DJ, Lin CP, Scadden DT, Crist C, Brack AS. 2019. Lineage Tracing Reveals a Subset of Reserve Muscle Stem Cells Capable of Clonal Expansion under Stress. Cell Stem Cell 24: 944-957.e5. Stuart T, Satija R. 2019. Integrative single-cell analysis. Nat Rev Genet 20: 257–272. Svensson V, Natarajan KN, Ly L-H, Miragaia RJ, Labalette C, Macaulay IC, Cvejic A, Teichmann SA. 2017. Power analysis of single-cell RNA-sequencing experiments. Nat Methods 14: 381–387. Vetter TA, Lawlor MW. 2026. Automated Quantification of Dystrophin Expression by Immunofluorescence in Humans and Animal Models. Methods Mol Biol 2975: 67–87. Zanotti S, Magri F, Poggetti F, Ripolone M, Velardo D, Fortunato F, Ciscato P, Moggio M, Corti S, Comi GP, et al. 2022. Immunofluorescence signal intensity measurements as a semi-quantitative tool to assess sarcoglycan complex expression in muscle biopsy. Eur J Histochem 66: 3418. Zhang L, Noguchi Y-T, Nakayama H, Kaji T, Tsujikawa K, Ikemoto-Uezumi M, Uezumi A, Okada Y, Doi T, Watanabe S, et al. 2019. The CalcR-PKA-Yap1 Axis Is Critical for Maintaining Quiescence in Muscle Stem Cells. Cell Rep 29: 2154-2163.e5. Zhou L, Zhou W, Joseph AM, Chu C, Putzel GG, Fang B, Teng F, Lyu M, Yano H, Andreasson KI, et al. 2022. Group 3 innate lymphoid cells produce the growth factor HB-EGF to protect the intestine from TNF-mediated inflammation. Nature immunology 23: 251–261.

  2. Review Commons

    Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

    Learn more at Review Commons

    Referee #3

    Evidence, reproducibility and clarity

    Summary: In this manuscript, the authors explore the heterogeneity of adult muscle stem cells (MuSCs) in murine and human skeletal muscles. They identify diverse expression levels of CD90 on MuSCs and demonstrate that CD90+ MuSCs are primed for myogenic commitment during muscle regeneration. The authors show that CD90+ MuSCs rapidly enter the cell cycle upon MYOD activation, mediated by rapid AMPK phosphorylation. Furthermore, they investigate the characteristics of both CD90+ and CD90- MuSCs in the quiescent state, revealing that the quiescence of CD90+ MuSCs is maintained through the COL6-CALCR pathway-an original finding of this study. Lastly, using an antibody-mediated depletion model targeting CD90+ MuSCs, they confirm the critical role of this population in muscle regeneration. Antibody-induced depletion or imbalance of CD90+ MuSCs results in impaired or delayed muscle regeneration.

    Major comments

    1. While the manuscript provides valuable insights into the functional heterogeneity of MuSCs, there are some critical aspects that remain unclear. Specifically, the mechanism by which the quiescence of CD90+ MuSCs, maintained through the COL6-CALCR pathway, confers an advantage for their rapid activation is not sufficiently addressed. Understanding why this pathway enables such responsiveness would significantly strengthen the authors' conclusions. Additionally, the manuscript does not elucidate how the quiescence of CD90-negative MuSCs is maintained, leaving a gap in the characterization of MuSC heterogeneity. Without this information, the functional significance of this heterogeneity, particularly in the context of muscle regeneration, remains incomplete.It would be interesting to explore why CD90-negative cells appear less responsive to injury or why CD90+ cells are more readily activated. Addressing these questions would provide a more comprehensive understanding of the biological implications of MuSC heterogeneity and enhance the impact of the study.
    2. I am particularly concerned about Figure 7. As the authors mentioned, CD90 is not specific to MuSCs. Therefore, the conclusion that CD90+ MuSCs are important for muscle regeneration based on the current experiment is not fully convincing. I suggest incorporating additional approaches to confirm this point. For example, transplantation of CD90+ or CD90- MuSCs into injured muscles would provide stronger support for their findings.
    3. A previous study has demonstrated that the COL5-CALCR pathway is essential for maintaining MuSC quiescence. In this manuscript, the authors propose the COL6-CALCR pathway; however, the current study lacks specific experiments to clarify the differences and similarities between these pathways. Additionally, the discussion section does not adequately address these points, leaving the interpretation incomplete. A more thorough discussion comparing these pathways would significantly improve the manuscript.

    Minor comments

    1. It would be interesting to see the spatial localization of CD90+ and CD90- MuSCs in skeletal muscle tissue.
    2. I suggest conducting a more thorough investigation to characterize the quiescent CD90+ and CD90- MuSCs, particularly focusing on aspects such as protein translation machinery.
    3. The authors should include a discussion of previously identified markers of MuSC heterogeneity, such as CD34, to provide better context for their findings.
    4. In Figure 6, the authors used COL4 as a negative control; however, this is insufficient to conclusively demonstrate the importance of COL6 in maintaining CD90+ MuSC quiescence. Including additional substrates beyond collagen, such as fibronectin or laminin, along with COL5, would strengthen the conclusions drawn from these experiments.

    Significance

    Overall, this study provides a unique perspective on the functional heterogeneity of MuSCs by using CD90 as a marker to delineate distinct MuSC subpopulations. This approach sheds light on the specific roles of CD90+ MuSCs in muscle regeneration and offers new insights into the regulatory mechanisms governing MuSC function.

  3. Review Commons

    Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

    Learn more at Review Commons

    Referee #2

    Evidence, reproducibility and clarity

    Summary:

    Kheir et al. explore the heterogeneity within the MuSC compartment, identifying a CD90+ subpopulation with enhanced activation and proliferative capacity compared to its CD90-ve counterpart. While functional heterogeneity in MuSCs is a well-recognized and intriguing area of study, prior research has often focused on differences inferred from transcriptional profiles. This current study advances the field by intriguingly linking CD90 expression to distinct functional outcomes, thereby providing more compelling evidence for the existence of this subpopulation. The authors further investigate potential mechanisms, suggesting a connection between CD90 expression and intrinsic pAMPK activity as a driver of proliferation, as well as a role for Col6-Calcr binding in maintaining quiescence. While the majority of experiments are rigorously conducted, additional studies are suggested to determine whether the CD90+ fraction represents a subpopulation with substantial functional relevance. Detailed critiques and recommendations for further studies are outlined below.

    Major Comments:

    1. It is perplexing that the CD90+ fraction is implicated in activation, proliferation, and differentiation (Mgn+ data) while simultaneously contributing to the CD90-ve population (Fig. 3E). However, the reverse does not seem to occur, with CD90-ve cells not replenishing the CD90+ fraction. If the CD90+ subpopulation indeed accounts for the majority of myogenesis, this provokes the question: what is the functional role of the CD90− fraction? Notably, CD90-ve MuSCs appear to divide effectively during regeneration (Fig. 2E-G), further emphasizing the need to clarify their contribution to the overall regenerative process. The presence of a substantial number of CD90-ve MuSCs across conditions suggests they cannot simply be dismissed as irrelevant and understanding their role will help clearly establish the +/- subpopulations as functionally different.
    2. The depletion of CD90+ cells (Fig. 7D-I) is the correct experimental approach to assess the function of these cells in vivo. However, the method employed, using IP injections of a CD90 antibody, can lack specificity. Even with optimal specificity, CD90 is expressed on numerous cell types across the body. This raises the possibility that observed effects may result from targeting other CD90+ cells in skeletal muscle or other tissues, both locally and systemically. To mitigate these confounding factors, the authors should attempt strategies to reduce off-target effects. While the technical challenges are acknowledged by this reviewer and may be prohibitory, addressing these limitations would substantially enhance the impact of this work. Additionally, the embryonic myosin heavy chain (eMHC) images (Fig. 7G, H) should be more representative of the quantification data to ensure consistency.
    3. Similar concerns about off-target effects noted in point #2, apply to the use of the Col6 KO mouse model, which appears to be a full body KO, meaning Col6 is absent not only in MuSCs but also in other cell types that typically express Col6. This deficiency would have been present throughout development, complicating the interpretation of the observed effects. The authors do acknowledge Col6 expression by non-MuSC cell types, but the in vivo impact remains challenging to interpret, particularly due to the potential developmental and systemic effects of removing Col6. Also, the observation that the CD90-ve subpopulation still expresses Calcr raises further questions about Col6 acting only on the CD90+ fraction and expression by MuSCs being consequential in vivo. The trend observed in Fig. 6M for CD90-ve cells suggests that this mechanism might not be exclusive to CD90+ cells, warranting further investigation or explanation since an outlier in the Col6KO CD90-ve group may have influenced interpretation.
    4. The siCD90 experiment in Fig. 5 demonstrates effective KD at both the transcript and protein levels, but the observed impact on the proliferation of CD90+ cells (Fig. 5G), while statistically significant, appears to be less than expected. This result is also confusing given the substantial reduction in pAMPK levels observed in Fig. 5L, leading to the expectation of a more pronounced effect on proliferation if the proposed CD90-pAMPK mechanism is a driving pathway. Additionally, Fig. 5N suggests that pAMPK supports proliferation in both CD90+ and CD90− subpopulations. While the AICAR treatment in CD90− cells does not achieve significance, the data exhibit a bimodal distribution among replicates, with an apparent outlier in the control group potentially skewing the analysis. This variability necessitates further clarification for the relationship between CD90, pAMPK, and MuSC proliferation.
    5. The CD90 related findings in human samples appear less robust compared to those in mice. While the sorting successfully identifies sizable CD90+ and CD90-ve populations (Fig. 4A), the sequencing data show only small regions of high CD90 expression, as highlighted in red by the authors (Fig. 4C, D). Have the authors considered replicating the sequencing experiments within their own laboratory? While it is acknowledged that sourcing human tissue may be a limitation, it may strengthen the translational impact if possible.

    Minor Comments:

    1. Fig. 1D - the MuSC population has an uncharacteristically low representation amongst cells of uninjured muscle. Can the authors comment on this in text?
    2. Fig. 2 - the g-alert phenotype corresponding with CD90 expression is interesting. Can the authors add a molecular marker to confirm this phenotype?
    3. The authors mention "significantly higher fraction of CD90+ve MuSCs incorporated EdU in vivo at 1.5, 2.75, and 6 days after injury..." in 2G, but it only seems 1.5 and 2.75 are different. The text should be corrected.
    4. 3B, C - 7dpi seems late for the analyses of Myogenin at a single-cell level since most differentiating MuSCs are fused at this point. Can the authors comment in the text why 7 dpi was chosen?
    5. Note - 7L is very interesting, but emphasizes the major Q#1 above. Mdx mice are known to lose MuSC capacity due to continuous rounds of proliferation. So, the CD90+ loss is supportive, but there are still many CD90-ve cells present. In other figures, the authors demonstrated the negative fractions still harbor decent proliferative potential, so why is there no rescue?
    6. The authors should clearly state the number of cells and number of replicates for their single cell distribution graphs in all legends.
    7. In their stats section of methods, ns= p{greater than or equal to}0.15, please clarify.

    Significance

    General Assessment: The study is well conducted and addresses MuSC functional heterogeneity. There seems to be substantial evidence that CD90 fractionates the MuSC population and is related to proliferative capacity. Functional assessment in vivo needs some clarification with additional experiments, but the study seems promising. Also, interpretation of graphs should be updated as well since some distribution of replicates may be impacting statistical significance that can alter interpretation/outcomes.

    Advance: Again, MuSC heterogeneity has been an area of intense investigation for many years. The advancement would be mechanistic/functional.

    Audience: Specialized in skeletal muscle. There is potential for the CD90 fractionation to extend to other cell and tissue types, but this extent is unknown until this work is expanded.

  4. Review Commons

    Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

    Learn more at Review Commons

    Referee #1

    Evidence, reproducibility and clarity

    Summary: Kheir and colleagues found that CD90 expression levels could be utilized to divide MuSCs into two populations. The authors demonstrated that CD90+ve MuSC became activated state faster than CD90-ve cells. Mechanistically, AMPK is more activated in CD90+ve cells in response to niche loss. To suppress the high responsiveness to activation signalings, CD90+ ve cells highly express ColVI and CALCR than CD90-ve cells. The authors carefully examined the heterogeneous expression of CD90 in MuSCs. Overall, however, the differences in the characteristics of CD90+ve and CD90-ve cells are small. In addition, most of the data were based on fluorescent intensity. This reviewer does not feel that this study will have a significant impact on our understanding of MuSC biology.

    Major comments:

    1. Data demonstrated the statistical differences in MuSC behaviors between CD90+ve and CD90-ve cells. However, the difference is small. For example, it is unclear whether the minimal difference in CALCR expression level between CD90+ve and CD90-ve cells gives rise to any biological difference.
    2. Negative controls of FACS analyses are required because different sizes of cells might exert different background intensities. (Figure 2I, 2L, and 6F)
    3. If CD90+ve MuSCs express Col6 higher than CD90-ve MuSCs, they should also highly express the primary target of Notch target genes, Hes1, Hey1, and HeyL. The authors should examine the expression levels of these genes.
    4. As described above, the quantifications of many results, including MyoD, were based on the fluorescent intensity. I know the difficulty of preparing enough cells for experiments, but the authors need to present data supporting these results.
    5. Figure 7G-H; More quantitative analyses should be included. In addition, the sample number was different between Fig7E and H. There is no significant difference in the CD90 expression in Fig7G. The authors need to confirm the reproductivity.

    Minor comments:

    1. Figure S4. The authors need to show evidence that these cells are proliferating. Without the evidence, CD90 expression my just be retained in non-dividing cells. If it is difficult, the results should be removed.
    2. Heterogeneity in cell cycle progression in MuSCs is well documented as fast and slow dividing cells. This reviewer recommends discussing the relevance of CD90 expression to these reports. PMID: 22349695 PMID: 8608871

    Significance

    The heterogeneity of muscle stem cells is of great interest to muscle stem cell biologists, including this reviewer. The orchestrated expression and regulation of activation and quiescence pathways is conceptually new. Several molecules are heterogeneously expressed in muscle stem cells, but the expression pattern of CD90 does not correlate with them. However, as noted above, the difference between CD90-positive and CD90-negative cells is relatively small.

  5. Version published to 10.1101/2024.10.25.619970 on bioRxiv