Reprogramming of cis-regulatory networks during skeletal muscle atrophy in male mice

A comprehensive atlas of cis-regulatory elements and their dynamic activity is necessary to understand the transcriptional basis of cellular structure maintenance, metabolism, and responses to the environment. Here we show, using matched single-nucleus chromatin accessibility and RNA-sequencing from juvenile male C57BL6 mice, an atlas of accessible chromatin regions in both normal and denervated skeletal muscles. We identified cell-type-specific cis-regulatory networks, highlighting the dynamic regulatory circuits mediating transitions between myonuclear types. Through comparison of normal and perturbed muscle, we delineated the reprogramming of cis-regulatory networks in response to denervation, described the interplay of promoters/enhancers and target genes. We further unveil a hierarchical structure of transcription factors that delineate a regulatory network in atrophic muscle, identifying ELK4 as a key atrophy-related transcription factor that instigates muscle atrophy through TGF-β1 regulation. This study furnishes a rich genomic resource, essential for decoding the regulatory dynamics of skeletal muscle in both physiological and pathological states.

Authors used an established model of 1-/2-week-long denervation of the gastrocnemius muscle in the mouse.
Authors used separate single-nucleus RNA-Seq and ATAC-Seq, bulk RNA-Seq, ChIP-Seq for H3K27ac, and extensive bio-informatic analysis to identify cell-type specific regulatory networks acting in cis, and gene expression/signaling regulatory circuits underlying the interactions between myofibers and muscle-resident cells, controlling myofiber identity, and regulating the reprogramming of muscle metabolic features post-denervation.
Authors identified novel roles of specific transcription factors in regulating muscle neurogenic atrophy.
Major discoveries/results: 1) The generation of an atlas of regulatory networks by combining results of gene expression and chromatin accessibility of promoters and enhancers in eleven different types of muscle-resident cells after muscle denervation.
2) The identification of specific cis-regulatory elements (CREs) regulating differential accessible chromatin regions (ACRs) nearby the transcription start sites (TSSs) in eight muscle-resident cells.Authors show that more than 60% of ACRs are situated in proximal promoter regions or in distal enhancers.
3) The finding of more than 50,000 cis-regulatory links underlying co-accessibility between ACRs (potential enhancers).4) The reduction of ACRs distribution in the promoter regions of myonuclei upon denervation coupled by a parallel ACRs increment in their distal and intragenic regions.5) Myonuclei of the denervated muscle showing the highest level of differentially accessible regions (DARs) in enhancer regions linked to the expression of muscle atrophy-related genes, and the parallel reduction of distal DARs regulating genes for sarcomeric proteins.6) The identification of atrophy-responsive TFs upon muscle denervation affecting several muscleresident cell populations, and the main affected regulons in myonuclei of denervated Gas.7) The analysis of TFs landscape upon Gas denervation complemented with the H3K27ac genomic distribution to identify the epigenetic modulation of new candidate enhancers in the different muscle-resident cell populations.8) Generation of a myofiber-specific TFs hierarchy upon neurogenic atrophy induction.This may become a reference model useful to follow and predict the evolution of epigenetic and transcriptomic changes in different models of muscle atrophy.9) The identification of Elk4 as a major pro-atrophic/catabolic driver in the denervated muscle.
Overall, the study is well designed and conducted, using the appropriate methods of analysis.
I have minor comments and the request of some editing as follows: 1) Please better specify all over text and figure legends when the analysis was carried out on muscles TAs and Gas harvested after 7 and 14 days from denervation.2) Throughout the figures I would state in the legends the magnification for all the muscle section IF images.
3) Lines 125-133.Considering the evaluation of some pioneer transcription factors (PTFs) potentially involved in CREs activation would be interesting.In addition, a description of what is Tn5, also in Fig. 1g, would be helpful.4) Lines 152-154.Considering changing the wording of this sentence or just list nerve activity, exercise, or hormonal influences instead of "nerve activity or by exercise or hormonal influences".5) Line 186.Please describe in text and Fig. 2h legend which muscle is depicted, TA? Gas? Normal or denervated?Maybe add a co-staining for MyHC-I or MyHC-II for fiber typing.6) Line 209.Authors stated that denervated fibers show lower level of heterochromatin; based on which type of assay?Are these two images at the same magnification?Maybe an H&E would make it easier to see the atrophy.Also, the evidence of nucleoli by EM is not fully convincing.7) Line 214.In text Authors state ACRs, while in Fig. 3e legend differentially accessible regions (DARs).Please clarify.8) Line 219.Authors write denervated muscle, while the percentage shown in Fig. 3e is referring to myonuclei.Muscle might be perceived as a whole in this manner, including all the 11 cell populations, so please clarify in the text.In addition, always in Fig. 3e, because many other types of muscle-resident cells show little changes between normal vs. denervated, or changes in the opposite direction when compared to what is shown by myonuclei, it would be great to add a comment on that.9) Line 244.Skeletal structural components; please change to skeletal muscle structural components.10) Line 264.DARs with reduced expression.Is it better to say DARs with reduced expression of their linked genes?The same in Line 267.11) Line 272.Extended data fig.4; please specify Fig. 4b-g.12) Line 301.It is interesting the role of Ar with its strong downregulation upon denervation, as well as that of its target genes.Authors used male mice in this study; please add an additional comment on that in the discussion section.13) Extended Fig. 6g legend: oxidative capacity IN normal and… 14) Line 324.Author shows X-gal staining in mouse TA but does not comment on this in the text while looking at Smox and Gadd45a expression following denervation.Are we seeing increase senescence in these fibers with denervation with Gadd45a? 15) Lines 356-8.Authors wrote: "increasing motif activity at beginning of denervation followed by a decline as denervation progresses".Even if Authors are talking also about a pseudotime, this statement indicates that Authors conducted the analysis at different time points after denervation as stated in Methods, in Fig. 6c "7 days and 14 days", and in Line 402 "…as denervation progressed.Please reformulate these statements more clearly and consistently.16) Fig. 5g LogFC denvertion vs. nrormal; please edit.17  6p?Laminin?23) Lines 571-72.Please edit sentence "Mice underwent surgical of denervation was described previously".24) Line 600.About 10,000 cells?Probably 10,000 nuclei.25) Lines 619-20.Please better describe how data have been processed and analyzed.26) Line 729.Please add the processing for IFs for Gas, since something is missing somewhere, see Fig. 2h and Fig. 3a.27) Lines 749-752.Please better describe the staining.Prepare?And remove?Leave?28) Line 770.Please add AAV1-GFP as per Fig. 6k.29) Lines 773-777.When were the cells transfected with siRNA?Was it following 96 hrs of differentiation?How long were the siRNAs transfected before processing of myotubes?30) Line 792.How much tissue from the Gas was used for RNA extraction?
Reviewer #3 (Remarks to the Author): In this study, the authors profiled landscapes of the accessible chromatin regions in skeletal muscles of normal and denervated mice, identified cell-type-specific cis-regulatory networks, illustrated the reprogramming of cis-regulatory networks in response to denervation, and revealed the interplay of key promoters/enhancers and target genes.This is a meaningful study, which provides a theoretical basis for studying physiological and pathological metabolism in skeletal muscle, as well as facilitates further hypothesis-or data-driven research.The study is well designed and optimally organized.Three biological replicates make the data and of single-nucleus chromatin accessibility and RNA-sequencing fair and solid.The research on the regulatory function of ELK4 in denervated muscles further confirms the reliability and value of the foregoing analysis.However, there are some issues that should be addressed before considering publication in Nature communications.
Major comments: 1. Line 117: Figure 1e does not confirm the specific expression of these genes, but only that they are expressed in normal Gas muscle.2. Figure 1g: Authors should provide UMAP plots showing the distribution of all cell types.Otherwise, cell-type-specific activities of TFs motif cannot be distinguished.3. Line 156: Is the data used to analyze the myofiber type transition only from normal muscles?The authors should explain the significance of studying the myofiber type transition only in normal adult mice.4. Authors should add scale bars to all histological and fluorescent-stained photographs.5. Figure 2h: The immunofluorescence co-staining assays of these three differentially accessible genes with myofiber type markers (MYH 4,7, and1) should be performed to illustrate their dynamic expression changes in the process of myofiber type transition.6. Figure 3h: Does purple lines indicate co-accessibility link score in normal myonuclei or in denervated myonuclei?How does the authors compare the difference in distance regulation between normal myonuclei and denervated myonuclei?7. Lines 358-362: Does the beginning of the pseudotime trajectory represent the "beginning of denervation"?If the answer is "yes", it can be seen from Figure 5f that the motif activity of NR3C1 does not increase first and then decrease along with denervation, nor does the motif activity of MYOG increase persistently with denervation.If the answer is "no", what point in the pseudotime trajectory is the "beginning of denervation"?8. Line 376: How does the gene regulatory network with TF hierarchy exhibit a divergent pattern over time? 9.In Extended Data Fig. 6, SDH staining of denervated group appears to be selective.

REVIEWER COMMENTS
Reviewer #1 (Remarks to the Author): The study by Lin et al. utilizes matched single-nucleus chromatin accessibility and RNA-sequencing to generate landscapes of the accessible chromatin regions in the gastrocnemius muscle of juvenile (12-weekold) male C57BL6 mice under normal and denervated conditions.The manuscript provides a single-cell genomic resource for exploring the regulatory mechanisms underlying skeletal muscle atrophy.Additional studies will need to be performed to determine whether the landscapes developed for denervation-induced atrophy relate to other forms of muscle atrophy such as disuse, cachexia and sarcopenia.
Response: We greatly appreciate your insightful suggestions.We concur that different forms of muscle atrophy likely share certain commonalities while also exhibiting unique molecular mechanisms at the cellular As you pointed out, understanding the potential common and unique regulatory landscapes across different forms of muscle atrophy such as disuse, cachexia, sarcopenia, and atrophy associated with chronic conditions like cancer and chronic kidney disease, would indeed be valuable.We appreciate your insights regarding the need for additional studies on this topic.We fully agree and are committed to pursuing further research in this area.The methodologies and results from our current study will significantly guide these future investigations.
The identification of ELK1 as an atrophy-related transcription factor that promotes muscle atrophy through the regulation of TGF-b1 is novel and significant.The genomic data is strengthened by the in vivo experiments and the data presented in Figure 6.A limitation of the in vivo electroporation data are that they were performed only in the TA muscle.
Response: We greatly value your insight on the limitation of our study involving the in vivo electroporation data limited to the TA muscle, especially, it did not provide the insights of how ELK4 influence whole body metabolism and muscle hemostasis.We currently do not possess ELK4 knockout mice.However, recognizing the importance of this, we plan to utilize the LoxP-Cre technique to create ELK4 knockout in skeletal muscles in transgenic mice for more comprehensive studies.We intend to investigate the systemic impact of ELK4 on protein and energy metabolism.We are optimistic that these future investigations will further our understanding of the multifaceted roles of ELK4 in muscle atrophy and metabolism.

Specific comments:
Title: The title needs to specify that these data came from male mice.Response: We appreciate your suggestion to specify the gender of the mice used in our study.In line with your advice, we have revised the title to: "Reprogramming of cis-regulatory networks during skeletal muscle atrophy in male mice".
Abstract: The abstract needs to identify that these data were collected in juvenile (12 weeks) male C57BL6 mice.Also, the abstract states that you "examined the regulatory circuits that underpin the transition between oxidative and glycolytic myofibers".However, what was examined was the transition between myofibers expressing different myosin heavy chain isoforms.Muscle fibers differ in their metabolic and contractile properties.While it is true that in normal adult muscle there is a relationship between myosin heavy chain expression and oxidative (mitochondria) properties such that type I and IIa fibers are highly oxidative and type IIb fibers have the lowest oxidative properties; this relationship can change in response to exercise training or pathological conditions that induce muscle atrophy.It is incorrect to use glycolytic and oxidative to infer myosin heavy chain expression.
Response: We appreciate the reviewer's guidance.We have revised the phrase "examined the regulatory circuits that underpin the transition between oxidative and glycolytic myofibers" to "examined the regulatory circuits that underpin the transition between different myonuclear types".Additionally, we have included information on the gender, age, and strain of the mice used in our study in the abstract.Please refer to the updated abstract in our revised manuscript.
Line 80: "normal mouse skeletal muscle" -----It should be clearly stated that the atlas that was generated came from the gastrocnemius muscle of 12-week old male C57BL6 mice.
Response: We appreciate your suggestion for specificity.We have now revised the text on line 80 from "normal mouse skeletal muscles and those with neurogenic atrophy" to "normal gastrocnemius muscles and those undergoing neurogenic atrophy in 12-week-old male C57BL6 mice".Please see red text in page 3 line 77.
Line 152: This analysis is examining the transition between myosin heavy chain expression in myofibers, not the change in oxidative capacity.Type IIb fibers can increase their oxidative capacity in response to endurance training without changing MHC expression.Fibers can also decrease their oxidative capacity without changing their MHC expression.
Response: We thank you for this important clarification.We have updated the text on line 148 from "Evidence has shown switching of oxidative (type IIa/IIx) myofiber to glycolytic (type IIb) myofiber" to "Evidence has demonstrated the transition of myosin heavy chain expression between slow (type IIa/IIx) myofibers and fast (type IIb) myofibers."Please refer to the revised manuscript for this update.
Line 188-200: This analysis is not examining fiber type switching.This analysis is from fibers that are under a static condition.It is examining the relationships that have been established as a function of development and maturation to define the properties of fibers expressing different myosin heavy chain types.It is generally true that in normal adult muscle, type IIa fibers are more oxidative than IIb fibers.However, one should not predict metabolic properties from MHC properties.
Response: We agree with your observation that the metabolic characteristics of skeletal muscle cannot be directly equated to MHC properties.Previous studies have indeed shown that myofiber type can be influenced by factors such as nerve activity, exercise, and hormonal changes.However, the cis-regulatory networks governing the configuration of MHCs and metabolic properties under normal conditions are still not fully understood.To address this, we performed a trajectory analysis and observed an ordered transition from type IIa to type IIx and from type IIx to type IIb within the myonuclei of normal muscles.This led us to hypothesize a dynamic equilibrium existing between different myonuclear-types under normal conditions, and this equilibrium is necessary for maintaining myofiber-type configuration and muscle function.Similar principles apply to the regulation of energy metabolic enzymes.To avoid confusion, we've reframed our description in the manuscript, opting for the term "myonuclear transition" rather than "myofiber type switching".We have made the appropriate revisions to our manuscript to reflect these clarifications.
Thank you again for your astute observations and suggestions, which have helped us improve the clarity and accuracy of our work.
Line 346: Denervation does not stimulate a glycolytic to oxidative myofiber transition.The oxidative capacity of the muscle decreases following denervation.Again, do not conflate MHC isoform expression with oxidative capacity.
Response: We appreciate your important clarification regarding MHC isoform expression and oxidative capacity.In response, we have revised our manuscript, replacing "oxidative myofiber" with "type IIa myofiber" on Page 13, Line 345.Our observation of a shift in MHC isoform expression, specifically from Myh4 to Myh2 during denervation-induced muscle atrophy, aligns with previous studies.For instance, Matthieu Dos Santos et al. demonstrated the endogenous spatio-temporal expression of adult Myosin genes in normal skeletal muscle using a transgenic model with a super-enhancer at the locus containing Myosin genes, including Myh2 and Myh4 (A fast Myosin super enhancer dictates muscle fiber phenotype through competitive interactions with Myosin genes.Nat Commun.2022).Our claim about changes in Myh expression is corroborated by our bulk RNA-seq, snRNA-seq, and snATAC-seq data.We propose that these changes may contribute to the functional disorder of myofibers seen during muscle atrophy.We thank you for raising this point and enabling us to improve the clarity and precision of our manuscript.
Reviewer #2 (Remarks to the Author): Authors sought to determine transcriptomic, chromatin accessibility, and epigenetic changes at single-nucleus resolution to generate a comprehensive atlas of interaction between the muscle-resident cell populations during muscle neurogenic atrophy.
Authors used an established model of 1-/2-week-long denervation of the gastrocnemius muscle in the mouse.
Authors used separate single-nucleus RNA-Seq and ATAC-Seq, bulk RNA-Seq, ChIP-Seq for H3K27ac, and extensive bio-informatic analysis to identify cell-type specific regulatory networks acting in cis, and gene expression/signaling regulatory circuits underlying the interactions between myofibers and muscle-resident cells, controlling myofiber identity, and regulating the reprogramming of muscle metabolic features postdenervation.
Authors identified novel roles of specific transcription factors in regulating muscle neurogenic atrophy.
Major discoveries/results: 1) The generation of an atlas of regulatory networks by combining results of gene expression and chromatin accessibility of promoters and enhancers in eleven different types of muscle-resident cells after muscle denervation.
2) The identification of specific cis-regulatory elements (CREs) regulating differential accessible chromatin regions (ACRs) nearby the transcription start sites (TSSs) in eight muscle-resident cells.Authors show that more than 60% of ACRs are situated in proximal promoter regions or in distal enhancers.
3) The finding of more than 50,000 cis-regulatory links underlying co-accessibility between ACRs (potential enhancers).4) The reduction of ACRs distribution in the promoter regions of myonuclei upon denervation coupled by a parallel ACRs increment in their distal and intragenic regions.5) Myonuclei of the denervated muscle showing the highest level of differentially accessible regions (DARs) in enhancer regions linked to the expression of muscle atrophy-related genes, and the parallel reduction of distal DARs regulating genes for sarcomeric proteins.6) The identification of atrophy-responsive TFs upon muscle denervation affecting several muscle-resident cell populations, and the main affected regulons in myonuclei of denervated Gas.7) The analysis of TFs landscape upon Gas denervation complemented with the H3K27ac genomic distribution to identify the epigenetic modulation of new candidate enhancers in the different muscle-resident Response: We carried out the immunostaining in Gas muscles, as suggested.Detailed descriptions of the samples have been added to both the main text and the figure legend.We have also incorporated a co-staining for Myh2 (Type IIa) or Myh4 (Type IIb) for fiber typing.Please see revised Fig. 2h.6) Line 209.Authors stated that denervated fibers show lower level of heterochromatin; based on which type of assay?Are these two images at the same magnification?Maybe an H&E would make it easier to see the atrophy.Also, the evidence of nucleoli by EM is not fully convincing.
Response: Like previous electron microscopy studies, which revealed that heterochromatin preferentially localizes to the nuclear periphery and around the nucleolus (J Padeken, P Heun -Current opinion in cell biology, 2014), our assertion that denervated fibers exhibit lower levels of heterochromatin was based on electron microscopy.As suggested, we've performed H&E staining to illustrate muscle atrophy and replaced the EM image with one of higher resolution to offer more convincing evidence.The H&E staining clearly presents changes in the nuclei of male mice muscles during muscle atrophy (Fig. S4a).7) Line 214.In text Authors state ACRs, while in Fig. 3e legend differentially accessible regions (DARs).Please clarify.
Response: We apologize for the lack of clarity.In Fig. 3e, the accurate term should be differentially accessible regions (DARs) across all cell types.We have corrected and clarified this in the revised manuscript (Page 8, line 218).8) Line 219.Authors write denervated muscle, while the percentage shown in Fig. 3e is referring to myonuclei.Muscle might be perceived as a whole in this manner, including all the 11 cell populations, so please clarify in the text.In addition, always in Fig. 3e, because many other types of muscle-resident cells show little changes between normal vs. denervated, or changes in the opposite direction when compared to what is shown by myonuclei, it would be great to add a comment on that.
) Fig. 6I text ELH4; please change to ELK4.18) How can Authors comment the partial evidence in Fig. 6k that Elk4 knockdown does not increase C2C12 myotube size?The same question on Fig. 7p in vivo.I do agree that Authors did not find myostatin activation upon muscle denervation, but this evidence requires an additional comment.Since it seems that Elk4 does not bind autonomously to DNA, how Authors explain Elk4 action on chromatin and on its target genes?It might be mediated by serum response factor (SRF) TF? 19) Could ionomycin induce pro-atrophic pathways independent of Elk4 action?20) In Extended Fig. 7d it is not possible to appreciate myofibers size.Please reformulate that statement in Fig. 7d legend.21) It is possible to add the total Smad3 in the Western blotting of Fig. 6l and Fig. 6o? 22) What does represent the red color in Fig.
level.Prior research, including the studies by Nicole Almanzar et al. (Nature, 2020) and Deirdre D Scripture-Adams et al. (Commun Biol, 2022), has characterized single-cell RNA-seq gene expression profiles of muscle atrophy caused by aging and Duchenne Muscular Dystrophy (DMD).Our study indeed adds to this growing body of literature by providing a single-cell genomic resource specific to denervation-induced muscle atrophy.