Heterogeneous levels of delta-like 4 within a multinucleated niche cell maintains muscle stem cell diversity

  1. Susan Eliazer
  2. Xuefeng Sun
  3. Emilie Barruet
  4. Andrew S Brack

  • Curated by eLife

    eLife logo

    eLife assessment

    The premise behind this manuscript is important and timely for muscle biologists and for stem cell biologists. The identification of heterogenous distribution of factors across the myofiber is an important contribution for dissecting how muscle stem cell diversity in a tissue is achieved. However, the mechanism of action proposed by the authors will require additional experimental support.

Reviewed by

Read the full article See related articles

Listed in

Log in to save this article

Abstract

The quiescent muscle stem cell (QSC) pool is heterogeneous and generally characterized by the presence and levels of intrinsic myogenic transcription factors. Whether extrinsic factors maintain the diversity of states across the QSC pool remains unknown. The muscle fiber is a multinucleated syncytium that serves as a niche to QSCs, raising the possibility that the muscle fiber regulates the diversity of states across the QSC pool. Here, we show that the muscle fiber maintains a continuum of quiescent states, through a gradient of Notch ligand, Dll4, produced by the fiber and captured by QSCs. The abundance of Dll4 captured by the QSC correlates with the protein levels of the stem cell (SC) identity marker, Pax7. Niche-specific loss of Dll4 decreases QSC diversity and shifts the continuum to cell states that are biased toward more proliferative and committed fates. We reveal that fiber-derived Mindbomb1 (Mib1), an E3 ubiquitin ligase activates Dll4 and controls the heterogeneous levels of Dll4. In response to injury, with a Dll4-replenished niche, the normal continuum and diversity of the SC pool is restored, demonstrating bidirectionality within the SC continuum. Our data show that a post-translational mechanism controls heterogeneity of Notch ligands in a multinucleated niche cell to maintain a continuum of metastable states within the SC pool during tissue homeostasis.

Article activity feed

  1. Version published to 10.7554/elife.68180 on eLife
  2. eLife

    Author Response

    Reviewer #3 (Public Review):

    In the submitted manuscript, Eliazer et. al. conclude that Dll4 and Mib present on myofibers maintain a continuum of SC fates providing SCs capable of regenerating muscle and repopulatin the SC niche. The data provide new insights into the maintenance of SCs, demonstrating niche-derived factors are responsible for regulating SC behavior. Loss of either Dll4 or Mib from the myofiber reduces SC numbers and impairs muscle regeneration. Overall the data provide compelling evidence that niche-derived Dll4 and Mib regulate SC fate, however, whether the interaction maintains a continuum of SC fates as concluded by the authors is insufficiently supported by the data provided.

    We thank the reviewer for their comments.

    One significant issue with the manuscript is the "discovery" of an SC continuum related to the relative levels of Pax7 expression. A similar continuum was established nearly a decade ago by Zammit et al., 2004 and Olguin et al., 2004 and thus, is not new. The authors need to reference the work and discuss the prior published data with regard to the observations in the current manuscript. The data establishing a continuum of SCs and the relationship to Pax7 protein levels can largely be eliminated and referenced by the two former manuscripts. For example, these manuscripts establish that elevated Pax7 levels drive quiescence and low Pax7 levels correlate with differentiation. The data from these manuscripts establish that SCs with modest Pax7 protein levels can acquire quiescence accompanied by increases in Pax7 protein

    The omission of these two seminal papers was a massive oversight on our behalf. They have now been included. In the original manuscript we acknowledged that SCs exist on a continuum-a gradual transition from one state to another, based on scRNA-seq studies and the present data (Dll4, Pax7 and Ddx6 expression). The references for the sequencing data were included. But with all due respect to the reviewer, the Zammit and Olguin papers binned Pax7 into discrete classes once satellite cells had activated. This is not a demonstration of a continuum. Moreover, we do not make any statements about Pax7 levels in activated conditions. Therefore, the reviewer is drawing comparisons between two different contexts. The statements we have made as they pertain to a continuum under homeostatic conditions are accurate with publications to date.

    The data relating the level of Pax7 expression with Dll4a and Mib are intriguing but the authors do not establish a direct relationship, demonstrating that Dll4 or Mib regulate Pax7 levels. An alternative explanation is that Dll4 and Mib inhibit differentiation and thus promote SC quiescence indirectly. This is a critical distinction, as the authors could be correct and Dll4 via Mib regulate SC fate.

    We don’t make the claim that Dll4/Mib1 regulates Pax7 directly. We would side with the majority of publications showing that Notch signaling directly regulates Pax7. We have now added further experiments to examine whether Dll4 regulates Notch signaling. We crossed a transgenic mouse line harboring a Notch reporter with MF-Dll4 mice to analyze Notch signaling in SCs. The first experiment we performed with this reporter was to correlate the levels of Pax7 and Notch signaling on a cell-by-cell basis. In control mice, we found a linear positive relationship between levels of Pax7 and the Notch reporter. Next, we compared Notch reporter levels in control versus Dll4-null. We observed that Notch reporter levels decreased to below detectable levels in Dll4 null muscle. Therefore, Dll4 acts non-autonomously to regulate Notch signaling in SCs during homeostasis (refer to Reviewer 1 comment 1, and Essential revisions #3).

    The reviewer raises an important point: Does Notch regulate quiescence directly or a differentiation/commitment program when SCs are in a quiescent state. We never claimed that Dll4/Mib1 regulates quiescence. The only way to conclude anything about quiescence would be to examine expression of proliferative markers in vivo. Rather, throughout the manuscript we referred to Dll4 regulating the state of the quiescent SC pool, as measured by changes in Pax7 and Ddx6 expression. In the discussion section we had discussed that Notch signaling may regulate differentiation/commitment of cells in a quiescent state.

    It is unclear that the loss of Dll4 or Mib1 reduce diversity of SCs. If these repress differentiation then their loss would be expected to enhance differentiation and reduce SC numbers, which is what the data demonstrate.

    Diversity can be restated as the variability across a population. We demonstrate that the variance of Pax7 and Ddx6 expression decreases after Dll4 deletion. Important to note that we are analyzing the SCs that are not lost through differentiation. The fact that some of the SCs are lost through differentiation is not inconsistent with a shift in the continuum. We expect SCs to be lost through differentiation as they shift along the continuum towards a Dll4/Notch/Pax7 low state.

    We observe reduced number of Dll4/Pax7 high cells, which is consistent with a shift in continuum. The counterpoint would be that Dll4/Notch/Pax7 high cells commit to differentiation. There is no evidence for that conclusion in this work or any other work published to date. We discussed this issue in the results section.

    We have also performed an experiment where mice were treated with a lower dose of TMX to reduce rather than delete Dll4. We find that the total number of SCs does not change, while the relative number of Dll4/Pax7 high cells is reduced while mid and low are increased (Figure 4). This is consistent with a shift in a continuum of states.

    Finally, the injury data provided are for 4d post injury and thus, the data may represent a delay in regeneration as opposed to a failure to regenerate. At 30 d post injury regeneration is typically considered complete. How do wild type and Dll4 null as well as Mib null muscle compare at 30d post injury.

    We analyzed muscle regeneration of MF-Dll4fl/fl tissue, 40 days after injury. The mean CSA of muscle fibers are significantly smaller than the control fibers, suggesting a defect in tissue regeneration. This is now included in Figure 5-figure supplement 2. Due to time constraints, we have not performed the same experiment with Mib1 mutants.

  3. eLife

    eLife assessment

    The premise behind this manuscript is important and timely for muscle biologists and for stem cell biologists. The identification of heterogenous distribution of factors across the myofiber is an important contribution for dissecting how muscle stem cell diversity in a tissue is achieved. However, the mechanism of action proposed by the authors will require additional experimental support.

  4. eLife

    Reviewer #1 (Public Review):

    The manuscript by Eliazer et al. identifies the Notch ligand Dll4 as a myofiber-derived regulator of muscle stem cells (satellite cells, SCs). The amount of Dll4 surrounding individual SCs on single fiber preparations correlates with the level of Pax7 protein in those cells. Genetic removal of Dll4 from fibers results in: 1) a distribution of Pax7 levels in remaining SCs that skews towards the lower end; and 2) a phenotype similar to, but weaker than, that previously published for removal of the essential Notch pathway transcriptional regulator RBP-J from SCs (including propensity of SCs to spontaneously enter the differentiation program and a deficient regenerative response). Genetic removal of Mib1 from fibers led to loss of Dll4 clustering at SCs and a phenotype similar to loss of Dll4. The authors conclude that Dll4 maintains a continuum of diverse SC states during quiescence, perhaps contributing to which SCs are prone to self-renewal vs. differentiation.

    It is accepted that the myofiber is a key niche cell for SCs, but the number of known myofiber-derived niche factors is very small and mechanisms are not well characterized. Furthermore, it is established that Notch signaling in SCs is critical to maintenance of SC quiescence, yet the source and identify of the relevant Notch ligands is not clear. Therefore, the elegant genetic identification of Dll4 as a myofiber niche factor is of high significance. The conclusion about SC states may be somewhat premature, and I have questions about how some of the experiments were performed, but overall this is a very useful paper for the field.

  5. eLife

    Reviewer #2 (Public Review):

    The work by Eliazer et al investigates the role of Dll4 spatial heterogeneity on myofibers in maintaining MuSC diversity. The authors show on isolated myofibers that individual MuSC exhibit different intensities, by immunofluorescence analysis, of Pax7 and Ddx6, expressed in quiescent MuSC, and that there is a positive correlation between the intensities of the two quiescence markers. They further isolated MuSC high, medium and low Pax7 from the Pax7-nGFP transgenic mouse and validated in vitro that that the Pax7 high are slower in entering the cell cycle and expressing myogenin. To understand whether diversity of factor on myofibers could regulate this spatial diversity, the authors focused on Notch signaling. By comparing by microarray data Notch ligands during postnatal muscle growth, they show that Dll4 showed the most enrichment as cells transition to quiescence. By immunofluorescence on isolated myofibers, the authors show heterogeneity of Dll4 localization across the myofiber, with enriched clusters around MuSC. The authors monitored along individual myofibers the distribution of Dll4 and found no correlation with the distance from the NMJ. Upon myofiber specific deletion of Dll4, the authors show that MuSC exhibit downregulation of Pax7 and Ddx6, as well as a reduced number of MuSC in tissues and increased expression of MyoD and myogenin. Upon injury, mice in which Dll4 was deleted in myofibers exhibited reduced myofiber cross-sectional area, indicating a defect in the repair process. By using mice in which Mib1, an activator of Notch signaling, is deleted in myofibers, the authors show reduced Dll4 intensity and reduced diversity of Pax7 expression in MuSC as well as impaired regeneration. Understanding how the microenvironment regulate MuSC diversity is relevant to dissect their heterogeneity. The findings are interesting and novel and the manuscript is well written. However, while the authors report diversity of Dll4 and Mib1 in myofibers, the approach of genetic deletion complete ablates gene expression, and it does not necessarily modulate spatial distribution. Thus, additional experiments are required in order to fully support the authors' interpretation.

  6. eLife

    Reviewer #3 (Public Review):

    In the submitted manuscript, Eliazer et. al. conclude that Dll4 and Mib present on myofibers maintain a continuum of SC fates providing SCs capable of regenerating muscle and repopulatin the SC niche. The data provide new insights into the maintenance of SCs, demonstrating niche-derived factors are responsible for regulating SC behavior. Loss of either Dll4 or Mib from the myofiber reduces SC numbers and impairs muscle regeneration. Overall the data provide compelling evidence that niche-derived Dll4 and Mib regulate SC fate, however, whether the interaction maintains a continuum of SC fates as concluded by the authors is insufficiently supported by the data provided.

    One significant issue with the manuscript is the "discovery" of an SC continuum related to the relative levels of Pax7 expression. A similar continuum was established nearly a decade ago by Zammit et al., 2004 and Olguin et al., 2004 and thus, is not new. The authors need to reference the work and discuss the prior published data with regard to the observations in the current manuscript. The data establishing a continuum of SCs and the relationship to Pax7 protein levels can largely be eliminated and referenced by the two former manuscripts. For example, these manuscripts establish that elevated Pax7 levels drive quiescence and low Pax7 levels correlate with differentiation. The data from these manuscripts establish that SCs with modest Pax7 protein levels can acquire quiescence accompanied by increases in Pax7 protein

    The data relating the level of Pax7 expression with Dll4a and Mib are intriguing but the authors do not establish a direct relationship, demonstrating that Dll4 or Mib regulate Pax7 levels. An alternative explanation is that Dll4 and Mib inhibit differentiation and thus promote SC quiescence indirectly. This is a critical distinction, as the authors could be correct and Dll4 via Mib regulate SC fate.
    It is unclear that the loss of Dll4 or Mib reduce diversity of SCs. If these repress differentiation then their loss would be expected to enhance differentiation and reduce SC numbers, which is what the data demonstrate. No direct experiments demonstrate that Dll4 regulates the levels of Pax7 protein, the data provided show a correlation of higher Pax7 protein if Dll4 is present.

    Finally, the injury data provided are for 4d post injury and thus, the data may represent a delay in regeneration as opposed to a failure to regenerate. At 30 d post injury regeneration is typically considered complete. How do wild type and Dll null as well as Mib null muscle compare at 30d post injury.

    In summary, the data are intruiguing and suggest that Dll4 regulates satellite cell fate and maintains quiescence of satellite cells or inhibits their differentiation. Some additional data will resolve which of these outcomes is likely.

  7. Version published to 10.1101/2020.10.20.347484v1 on bioRxiv