Abstract

   Circadian rhythms orchestrate physiological processes such as
   metabolism, immune function, and tissue regeneration, aligning them
   with the optimal time of day (TOD). This study identifies an interplay
   between the circadian clock within muscle stem cells (SCs) and their
   capacity to modulate the immune microenvironment during muscle
   regeneration. We reveal that the SC clock triggers TOD–dependent
   inflammatory gene transcription after injury, particularly genes
   related to neutrophil activity and chemotaxis. These responses are
   driven by cytosolic regeneration of the signaling metabolite
   nicotinamide adenine dinucleotide (oxidized form) (NAD^+), as enhancing
   cytosolic NAD^+ regeneration in SCs is sufficient to induce
   inflammatory responses that influence muscle regeneration. Mononuclear
   single-cell sequencing of the regenerating muscle niche further
   implicates the cytokine CCL2 in mediating SC-neutrophil cross-talk in a
   TOD-dependent manner. Our findings highlight the intersection between
   SC metabolic shifts and immune responses within the muscle
   microenvironment, dictated by circadian rhythms, and underscore the
   potential for targeting circadian and metabolic pathways to enhance
   tissue regeneration.
     __________________________________________________________________

   The stem cell circadian clock orchestrates metabolic and immune
   signaling to time muscle regeneration following injury.

INTRODUCTION

   The circadian clock, the body’s intrinsic time-keeping mechanism,
   operates through a molecular clock in almost all cells within the body
   ([49]1). It comprises a central clock in the brain’s hypothalamic
   suprachiasmatic nucleus and peripheral clocks distributed throughout
   the rest of the body, influencing behaviors like sleep and
   physiological processes including core body temperature, hormone
   secretion, immune function, and metabolism ([50]2, [51]3). Circadian
   rhythms ensure that these activities align optimally with the time of
   day (TOD), maximizing fitness. On the molecular level, the clock
   functions as a transcription-translation feedback loop consisting of
   activators (CLOCK and BMAL1) that direct the expression of thousands of
   genes, including their own repressors (PER1-3 and CRY1-2). Accessory
   proteins can stabilize the oscillation, including the nuclear
   receptors, REVERBa/b and RORa/b ([52]1). Previous work from our group
   ([53]4) and others ([54]5) uncovered that the efficiency of muscle
   repair is subject to TOD variations, suggesting that circadian rhythms
   influence tissue regeneration. We also demonstrated that muscle stem
   cell–specific deletion of Bmal1 impairs hypoxic growth and
   differentiation of myoblasts in vitro and hinders muscle regeneration
   in vivo ([55]4). Despite these findings, the underlying mechanisms by
   which SC circadian clocks regulate muscle repair remain largely
   unknown.

   Muscle regeneration is a highly coordinated, dynamic process involving
   intricate interactions between muscle stem cells (SCs, a.k.a. satellite
   cells) and the immune system ([56]6). SCs are crucial for muscle
   repair, activated by injury to proliferate and differentiate, thereby
   replenishing the muscle fiber pool ([57]7). Neutrophils, the first
   responders to injury, play a critical role in initiating repair by
   clearing debris and secreting cytokines and chemokines ([58]8), thus
   fostering an environment conducive to healing and influencing the
   behavior of SCs and other progenitor cells ([59]9, [60]10).
   Nevertheless, the interactions between neutrophils and SCs, especially
   considering the metabolic changes triggered by muscle injury, are yet
   to be fully understood.

   This study identifies a role for the SC circadian clock in modulating
   these interactions through metabolic responses to injury. We
   demonstrate that the SC clock drives the redox cycling of the oxidized
   form of nicotinamide adenine dinucleotide (NAD^+) via anaerobic
   glycolysis in a TOD-dependent manner. Using a genetic mouse model
   expressing the cytosolic reduced form of NAD^+ (NADH) oxidase from
   Lactobacillus brevis (LbNOX) within SCs, we found that cytoplasmic
   NAD^+ regeneration is sufficient to boost myofiber repair. Notably, we
   uncovered a previously unrecognized role for NAD^+ regeneration in SCs
   in the activity of poly[adenosine diphosphate (ADP)–ribose] polymerase
   1 (PARP1) and nuclear factor κB (NF-κB), extending beyond previously
   identified NAD^+-dependent control of Sirtuins (SIRTs) and myogenic
   gene deacetylation. We show that hypoxic NAD^+ regeneration acts in SCs
   to drive neutrophil chemotaxis–related gene expression and enhance
   neutrophil infiltration and their interaction with SCs. In particular,
   the NAD^+-regulated chemokine Ccl2 exhibited varied expression levels
   across different times of day, influencing the quality and rate of
   muscle regeneration.

   Our findings not only provide insights into the circadian regulation of
   muscle regeneration but also highlight a complex metabolic-immune
   interface within the muscle microenvironment, potentially explaining
   the compromised muscle regeneration observed in conditions
   characterized by disrupted circadian rhythms, such as obesity,
   diabetes, and aging.

RESULTS

ASCs express genes modulating anaerobic glycolysis and neutrophil functions

   Under conditions of hypoxia, pyruvate, the end product of glycolysis,
   is converted into lactate by lactate dehydrogenase A (Ldha). This
   process involves the reduction of pyruvate with electrons from NADH,
   facilitating the regeneration of NAD^+ ([61]Fig. 1A). Using the
   myotoxin cardiotoxin to injure mouse muscles, we observed that the
   injury creates an acute hypoxic environment within skeletal muscle, a
   condition confirmed by the substantial stabilization of
   hypoxia-inducible factor (HIF) proteins in the initial days post-injury
   (fig. S1A). SCs within the muscle are also subjected to this hypoxic
   stress, as demonstrated by microarray analysis from a previous study
   ([62]9) that examined the temporal gene expression in SCs under both
   homeostatic conditions and after cardiotoxin-induced injury at various
   time points. Key HIF target genes, such as Hk2, Vegfα, and Epo, were
   markedly up-regulated in SCs by day 2 post-injury, followed by a
   decline over time (fig. S1B). Notably, Ldha expression, known for its
   higher pyruvate affinity and preference for converting pyruvate to
   lactate ([63]11), was increased during the hypoxia phase (fig. S1C).
   Conversely, its isoform Ldhb, which favors converting lactate back to
   pyruvate due to its higher affinity for lactate, was down-regulated in
   the days immediately following the injury (fig. S1C). This suggests
   that SCs generate NAD^+ through pyruvate fermentation in the early days
   post–myotoxin-induced muscle injury. We then performed RNA sequencing
   on quiescent SCs (QSCs) from homeostatic muscles and the activated SCs
   (ASCs) from injured muscles on day 1 post-injury, a time when
   cytoplasmic NAD^+ is predicted to be highly produced ([64]Fig. 1B). The
   transcription profiles of SCs in these two states showed segregation by
   principal components analysis, with the first principal component
   accounting for more than 90% of the variance ([65]Fig. 1C). In
   addition, thousands of genes were differentially expressed
   (|LogFC| > 0.5, adj. P < 0.05) (fig. S2A). As anticipated, genes
   related to hypoxia and glycolysis signaling pathways were significantly
   up-regulated in ASCs at 1 day post-injury (dpi) compared to QSCs
   ([66]Fig. 1D), confirming the hypoxic state of these ASCs. Genes
   associated with the inflammatory response, particularly those involved
   in neutrophil chemotaxis and migration, were also highly expressed in
   ASCs compared to QSCs ([67]Fig. 1E and fig. S2B). Given that
   neutrophils are the initial wave of immune cells to infiltrate the
   wounded site, peaking at day 1 post-injury ([68]12), we hypothesized
   that ASCs may play a role in modulating neutrophil activity at the
   early stages of muscle regeneration.

Fig. 1. Elevated expression of glycolysis and inflammation-related genes in
SCs following CTX injury.

   [69]Fig. 1.
   [70]Open in a new tab

   (A) Schematic representation of the anaerobic glycolysis pathway. (B)
   Schematic diagram of experiment using SCs extracted from the hindlimb
   muscles of Pax7-zsgreen mice, both uninjured and following
   cardiotoxin-induced injury at ZT16, on day 1 post-injury, and subjected
   to RNA sequencing for transcriptome analysis. (C) Principal components
   analysis depicting the transcriptional disparities between ASCs at 1
   dpi and QSCs. (D) Heatmap showing genes significantly up-regulated
   (adjusted P value <0.05) in ASCs at 1 dpi, highlighting associations
   with hypoxia and glycolysis, in comparison to QSCs. The color scale
   represents the log[2] fold change, indicating the magnitude of gene
   expression changes. (E) Pathway enrichment analysis, using MSigDB
   hallmarks and gene ontology biological processes, for genes
   significantly up-regulated (adjusted P value <0.05) in ASCs at 1 dpi
   relative to QSCs.

Table 1. Reagent table.

   List and source information for reagents and resources used in the
   study.
   Reagent or resource Source Identifier
   Antibodies
   Biotin anti-CD31 BioLegend Catalog no. 102404
   Biotin anti-CD45 BioLegend Catalog no. 103104
   Biotin anti-Sca1 BioLegend Catalog no. 108104
   PE anti-VCAM1 Miltenyi Biotec Catalog no. 130-104-712
   Anti-laminin Sigma-Aldrich Catalog no. L9393
   Ly-6G (Gr-1) antibody Thermo Fisher Scientific Catalog no. 14-5931-82
   Poly(ADP-ribose) polyclonal antibody Eno Life Science Catalog no.
   ALX-210-890A-0100
   Goat anti-rabbit IgG secondary antibody, Alexa Fluor 488 Thermo Fisher
   Scientific Catalog no. A11034
   Goat anti-rat IgG secondary antibody, Alexa Fluor 555 Thermo Fisher
   Scientific Catalog no. A-21434
   Goat anti-rabbit IgG secondary antibody, Alexa Fluor 594 Thermo Fisher
   Scientific Catalog no. A11012
   Anti-HIF1α Cayman Catalog no. 10006421
   PAR/pADPr R&D Systems Catalog no. 4335-MC-100
   Anti–NF-κB p65 Cell Signaling Technology Catalog no. 8242S
   Anti–phospho–NF-κB p65 (Ser^536) Cell Signaling Technology Catalog no.
   3033S
   β-Actin Cell Signaling Technology Catalog no. 4970
   InVivoMab anti-mouse Ly6G BioXCell Catalog no. BE0075-1
   InVivoMab anti-trinitrophenol Isotype control rat monoclonal BioXCell
   Catalog no. BE0089
   Chemicals, consumables
   Tamoxifen Sigma-Aldrich Catalog no. T5648
   Cardiotoxin Millipore Sigma Catalog no. 217503-1MG
   Insulin syringes BD Catalog no. BD-25150
   Tissue Plus O.C.T. compound Thermo Fisher Scientific Catalog no.
   23-730-571
   Formalin Thermo Fisher Scientific Catalog no. MR0458682
   Clarifier Sigma-Aldrich Catalog no. A3179
   Hematoxylin Thermo Fisher Scientific Catalog no. 72-11 L
   Eosin-Y Thermo Fisher Scientific Catalog no.71-11 L
   Permount mounting medium Electron Microscopy Science Catalog no.
   17986-01
   4% Paraformaldehyde Boster Bio Catalog no. AR1068
   VectaShield HardSet antifade mounting medium with DAPI Vector
   Laboratories Catalog no. H-1500-10
   Propidium iodide Ready Flow reagent Thermo Fisher Scientific Catalog
   no. [71]R37169
   Hams F-10 medium Corning Catalog no. 10-070-CV
   DMEM Corning Catalog no. 90-013-PB
   Collagenase D Roche Catalog no. 11088866001
   Dispase II Gibco Catalog no. 17105041
   40-μm cell strainer Corning Catalog no. 07-201-430
   70-μm cell strainer Corning Catalog no. 352350
   Pre-sort buffer BD Catalog no. BD-B563503
   Matrigel matrix Corning Catalog no. 356231
   Basic fibroblast growth factor ProSpec Catalog no. CYT-218
   Ad-CMV-iCre Vector Biolabs Catalog no. 1045
   Ad-CMV-Null Vector Biolabs Catalog no. 1300
   Fetal bovine serum Corning Catalog no. 35-010-CV
   AMPure XP beads Beckman Coulter Catalog no. A63880
   CelLytic mammalian tissue lysis reagent Sigma-Aldrich Catalog no. C3228
   Protease inhibitor cocktail Sigma-Aldrich Catalog no. P8340
   DC protein assay kit Bio-Rad Catalog no. 5000111
   0.45-μm nitrocellulose membrane Bio-Rad Catalog no. 1620115
   RBC lysis buffer (10X) BioLegend Catalog no. 420301
   BGP15 Tocris Catalog no. 6703
   EB 47 Tocris Catalog no. 4140
   Galloflavin Tocris Catalog no. 4795
   Mouse Ccl2 cDNA ORF clone GenScript Catalog no. OMu19558
   Phusion high-fidelity DNA polymerase and dNTP mix Thermo Fisher
   Scientific Catalog no. F530N
   MIGR1 Addgene Catalog no. 27490
   pCL-Eco Addgene Catalog no. 12371
   Lipofectamine 3000 transfection reagent Thermo Fisher Scientific
   Catalog no. L3000015
   Syringe filter, SFCA, 0.45 μm CELLTREAT Catalog no. 229766
   Polybrene Millipore Sigma Catalog no. TR-1003-G
   Critical commercial assays
   miRNeasy FFPE kit Qiagen Catalog no. 217504
   The SMART-seq v4 Ultra-Low Input RNA Kit Takara Bio Catalog no. 634890
   Agilent’s high-sensitivity DNA kit Agilent Catalog no. 5067-4626
   Nextera XT DNA library preparation kit Illumina Catalog no. FC-131-1024
   The Qubit dsDNA HS assay kit Thermo Fisher Scientific Catalog no.
   [72]Q32851
   NovaSeq 6000 S1 reagent kit v1.5 (100 cycles) Illumina Catalog no.
   20028319
   NovaSeq 6000 SP reagent kit v1.5 (100 cycles) Illumina Catalog no.
   20028401
   Applied Biosystems High-Capacity cDNA Reverse Transcription Kit Thermo
   Fisher Scientific Catalog no. 43-688-13
   TRIzol Reagent Thermo Fisher Scientific Catalog no. 15596018
   iTaq Universal SYBR Green Supermix Bio-Rad Catalog no. 1725121
   Mouse MCP-1 ProQuantum immunoassay kit Invitrogen Catalog no. A44837
   Software and algorithms
   Cell Ranger 10x Genomics
   STAR Dobin et al. ([73]96)
   Seurat Stuart et al. ([74]104)
   CellChat Jin et al. ([75]20)
   NicheNet Browaeys et al. ([76]34)
   Slingshot Street et al. ([77]46)
   tradeSeq Van den Berge et al. ([78]47)
   scDblFinder Germain et al. ([79]103)
   DESeq2 Love et al. ([80]98)
   Trimmomatic Bolger et al. ([81]95)
   RSEM Li et al. ([82]97)
   tximport Soneson et al. ([83]99)
   EnrichR Kuleshov et al. ([84]100)
   ShinyGO Ge et al. ([85]101)
   Fiji Schindelin et al. ([86]88)
   LabelsToROIs Waisman et al. ([87]89)
   [88]Open in a new tab

SCs express genes associated with neutrophil attachment, adhesion, and
chemoattraction

   Next, we performed single-cell RNA sequencing to explore the potential
   interactions between SCs and neutrophils during the early hypoxic
   stages of muscle regeneration using mononucleated cells collected
   following injury at two times of the day Zeitgeber 4 (ZT4 and ZT16)
   from cardiotoxin-injured tibialis anterior (TA) muscles at days 1 and 3
   post-injury, as well as intact uninjured muscles ([89]Fig. 2A).
   Following quality control procedures, the samples were integrated into
   a comprehensive transcriptomic atlas comprising 58,330 cells ([90]Fig.
   2B). We then performed unsupervised clustering and, through Uniform
   Manifold Approximation and Projection (UMAP), identified 21 cell types
   annotated on the basis of the expression of established marker genes
   ([91]Fig. 2B and fig. S3A) ([92]13–[93]15). Notably, within this atlas,
   two distinct clusters expressing classical neutrophil markers (S100a8
   and S100a9) were delineated, termed Neutrophils_1 and Neutrophils_2
   ([94]Fig. 2B and fig. S3A). Differential gene expression analysis
   revealed that Neutrophils_1 are more mature than the Neutrophils_2
   subtype, as evidenced by a significantly higher expression of the
   chemokine receptor Cxcr2 (fig. S3B) ([95]16). Correspondingly, gene
   ontology of biological process analysis showed that Neutrophils_1
   exhibited elevated expression of genes associated with inflammatory
   responses (fig. S3C), signifying their pronounced inflammatory nature
   and advanced maturation. In contrast, Neutrophils_2 showed increased
   expression of genes related to lysosomal lumen acidification (fig.
   S3C), a characteristic of granulocyte differentiation of neutrophilic
   leukocyte precursors, which is essential for degradation of
   phagocytosed microorganisms at the inflammation sites ([96]17).
   Intriguingly, enrichment analysis using the literature-based network in
   the Elsevier Pathway Collection highlighted surface expression markers
   regulated by myoblasts in Neutrophils_1 and by myeloblasts in
   Neutrophils_2, respectively (fig. S3C). This observation suggests that,
   like other well-established neutrophil modulators such as necrotic
   fibers ([97]18, [98]19), ASCs may contribute to the regulation of
   neutrophil activity, particularly influencing Neutrophils_1, following
   CTX injury.

Fig. 2. Interactions between SCs and neutrophils during early muscle
regeneration.

   [99]Fig. 2.
   [100]Open in a new tab

   (A) Schematic diagram of single-cell sequencing experiment using
   mononucleated cells from the TA muscles of both hindlimbs were
   collected at ZT4 and ZT16 on day 0, 1, and 3 after injury. Single cells
   were collected through fluorescence-activated cell sorting (FACS),
   excluding dead cells identified by propidium iodide (PI) staining.
   Samples from each time point and day were captured in droplet emulsions
   using a 10x Chromium Controller (10x Genomics), aiming for 10,000 cells
   per sample, and then sequenced. (B) UMAP visualization of 21 annotated
   cell types within pooled TA muscle cells from all six samples. EC,
   endothelial cells; FAPs, fibro-adipogenic progenitors. (C) UMAP of
   neutrophil and MuSC single-cell transcriptomes split by days
   post-injury. (D) Chord diagram showing the specific ligand-receptor or
   signaling pathway–mediated interactions among the three cell clusters.
   The edge colors correspond to the source (sender) of the signal, and
   edge weights are proportional to the interaction strength, with thicker
   lines indicating stronger signals. Circle sizes reflect the number of
   cells in each cell group. (E) Circle plot showing the aggregated
   cell-cell communication network sent from MuSCs and received by
   individual clusters of neutrophils at various days post-injury. The
   inner, thinner bar colors represent the targets receiving signals from
   the corresponding outer bars. The size of the inner bar is proportional
   to the signal strength received by the targets.

   On the basis of our finding that neutrophil-related genes are induced
   in MuSCs at day 1 post-injury ([101]Fig. 1E), we next focused on the
   regulatory interplay between MuSCs and neutrophils during early muscle
   regeneration. To this end, we separated neutrophils and MuSCs from the
   total cell populations and reexamined their dynamics over time using
   UMAP. A few Neutrophils_1 but no Neutrophils_2 were initially presented
   in homeostatic muscles (day 0) ([102]Fig. 2C). There was a marked
   increase in both neutrophil types on day 1 post-injury, which then
   sharply declined by day 3 post-injury ([103]Fig. 2C). Potential
   cross-communication between MuSCs and the two neutrophil types were
   delineated using CellChat, which can infer intercellular signaling
   network based on single-cell RNA sequencing data ([104]20). This was
   illustrated through vectorial circle plots representing interactions
   between any two cell groups ([105]Fig. 2D). The communicative links
   mediated by MuSCs were most prominent on day 1 post-injury, highlighted
   by the densest edge weights directed from MuSCs toward the neutrophil
   populations ([106]Fig. 2D). The nature of MuSC-initiated interactions
   evolved across the different days. Interactions involving laminin and
   collagen signaling, which facilitate cell attachment or adhesion, along
   with α chemokines featuring the CXC motif (CXCL) secreted by MuSCs,
   were prevalent at all observed time points. In contrast, β chemokines
   with the CC motif (CCL) were exclusively secreted by ASCs at 1 dpi,
   playing a crucial role in mediating MuSC communication with both
   neutrophil types ([107]Fig. 2E).

NAD^+ boosts chemokine gene expression and neutrophil recruitment in
regeneration

   CXCL and CCL chemokines are potent neutrophil chemoattractants
   ([108]21). Given that hypoxia is known to modulate chemokine and
   chemokine receptor expression ([109]22, [110]23), we investigated
   whether low oxygen tension or, more specifically, anaerobically
   regenerated NAD^+, could trigger chemokine gene expression in SCs.
   First, we cultured wild-type myoblasts under normoxia or hypoxia (1%
   O[2]) for short durations ranging from 1 to 6 hours and measured NAD^+
   levels by high-performance liquid chromatography. Our results showed a
   clear, duration-dependent increase in total NAD^+ levels in myoblasts
   exposed to hypoxia compared to those cultured under normoxia ([111]Fig.
   3A). To uncouple NAD^+ redox cycling from hypoxia, we expressed
   cytosolic NADH oxidase from L. brevis (LbNOX) in myoblasts (fig. S4, A
   and B), which allows for a compartment-specific elevation of the
   NAD^+/NADH ratio ([112]24). Primary myoblasts from cytoLbNOX^LSL/LSL
   mice were isolated and treated with Cre-expressing or control
   adenovirus ([113]Fig. 3B). CytoLbNOX-expressing myoblasts exhibited a
   significant increase in intracellular NAD^+ (fig. S4C), a decrease in
   NADH levels (fig. S4D), resulting in increased NAD^+/NADH ratio (fig.
   S4E). Seventy-two hours post infection, the cells were exposed to
   normoxic (21% O[2]) or hypoxic (1% O[2]) conditions for 24 hours. Among
   the candidate neutrophil chemotaxis–related genes identified as
   significantly up-regulated in 1-dpi ASCs compared to QSCs (fig. S2B),
   Ccl2 was notably the only chemokine that could be induced not only by
   hypoxia but also by cytoLbNOX, even under normoxia ([114]Fig. 3C and
   fig. S5). A two-way analysis of variance (ANOVA) revealed a significant
   interaction effect between oxygen concentration and cytoLbNOX-mediated
   NAD^+ regeneration (P = 0.008). Hypoxia treatment did not further
   increase gene expression in cytoLbNOX-expressing myoblasts ([115]Fig.
   3C), suggesting that hypoxia may induce Ccl2 expression via NAD^+
   regeneration. Consistent with this observation, the addition of
   galloflavin, a lactate dehydrogenase inhibitor that impedes NAD^+
   regeneration ([116]25), diminished Ccl2 expression under hypoxia
   conditions ([117]Fig. 3D) but failed to inhibit its up-regulation in
   cytoLbNOX-expressing myoblasts ([118]Fig. 3E), further supporting the
   pivotal role of NAD^+ regeneration, rather than lactate, in this
   context.

Fig. 3. NAD^+ regeneration-induced immunomodulatory gene expression via
PARP-mediated PARylation and NF-κB activation.

   [119]Fig. 3.
   [120]Open in a new tab

   (A) NAD^+ contents in wild-type myoblasts cultured in normoxia and
   hypoxia up to for 6 hours. n = 3 wells per condition. *P < 0.05 by
   one-way analysis of variance (ANOVA) test. (B) Schematic diagram of
   viral induction of LbNOX expression in primary myoblasts. (C) Ccl2
   expression in myoblasts under conditions described in (B). *P < 0.05,
   **P < 0.01 by two-Way ANOVA. ns, nonsignificant. (D) Ccl2 expression in
   myoblasts subjected to hypoxia (1% O[2]) for 24 hours, both in the
   absence and presence of galloflavin. Normoxic myoblasts (21% O[2])
   served as controls. ***P < 0.001 by one-way ANOVA, comparing each group
   to normoxic myoblasts. (E) Ccl2 expression in myoblasts treated for 24
   hours with or without galloflavin. *P < 0.05, ***P < 0.001 by two-way
   ANOVA analysis. (F) Schematic diagram of MuSC conditional LbNOX
   (LbNOX^musc) mouse experiment (n = 3 per group). (G)
   Immunohistochemistry and quantification of mouse TA muscle for Ly6G^+
   neutrophils and laminin. Scale bar, 100 μm. (H) Representative
   immunofluorescence image and quantification of neutrophils in mouse TA
   muscle under conditions described in (F). *P < 0.05 by unpaired
   Student’s t test. (I) Immunofluorescence of global PARylation in
   myoblasts. LbNOX expression is marked by GFP coexpression. Scale bar,
   200 μm. *P < 0.001 by unpaired Student’s t test. (J) Immunoblot
   analysis and quantification of global PARylation,
   serine-536–phosphorylated and total NF-κB p65 subunit, and β-actin in
   the absence (I) or presence (J) of BGP15. *P < 0.05 by unpaired
   Student’s t test. DMSO, dimethyl sulfoxide. (K) Immunomodulatory gene
   expression following a 6-hour treatment with or without BGP15.
   ***P < 0.001 by two-way ANOVA. (L) Ccl2 expression following a 6-hour
   treatment with or without EB47. *P < 0.05, ***P < 0.001 by one-way
   ANOVA.

   Considering the chemoattractant properties of CCL2, we examined whether
   cytoLbNOX-driven NAD^+ augmentation could enhance neutrophil
   recruitment within regenerating muscles in vivo. To this end, we
   crossed cytoLbNOX^LSL/LSL mice with a line expressing Cre recombinase
   under the control of a tamoxifen-inducible Pax7 promoter, resulting in
   excision of an upstream STOP codon to allow for muscle stem
   cell–specific cytoLbNOX expression (LbNOX^musc). One week following
   tamoxifen induction, we performed cardiotoxin-induced TA muscle injury
   and assessed neutrophil presence in both uninjured muscles and at 7 dpi
   via Ly6G immunohistochemistry ([121]Fig. 3F and fig. S6). Little or no
   Ly6G-positive signal was detected in the homeostatic uninjured muscles
   of both groups (fig. S6). At 7 dpi, neutrophil numbers in control
   muscles had largely diminished, whereas LbNOX^musc mice retained
   significantly more neutrophils ([122]Fig. 3G), aligning with our
   hypothesis that cytoplasmic NAD^+ mediates neutrophil activity in
   regenerating muscle.

   NAD^+ serves as a critical cosubstrate for two main families of
   posttranslational modification enzymes: the ADP-ribosyltransferase
   (ARTs) and the SIRTs ([123]26–[124]28). Given that PARP enzymes, a
   subset of ARTs, can activate the major immunoregulatory transcription
   factor NF-κB ([125]29), we asked whether cytoplasmic NAD^+ might
   regulate chemokines expression via PARP-mediated poly-ADP ribosylation
   (PARylation). CytoLbNOX-expressing myoblasts showed increased
   PARylation and NF-κB activation ([126]Fig. 3, H and I), which we also
   observed in hypoxia (fig. S7A). Notably, treatment with BGP15, a broad
   PARP inhibitor ([127]30), reduced intracellular protein PARylation and
   NF-κB activation ([128]Fig. 3J), leading to a significant decrease in
   Ccl2 expression in cytoLbNOX-expressing myoblasts ([129]Fig. 3K). This
   implies that PARP-mediated PARylation contributes to NAD^+-mediated
   Ccl2 expression via NF-κB. We found that Parp1 was significantly
   increased (log[2]FC = 1.68, adj. P = 1.14 × 10^−22) in 1-dpi ASCs
   compared to QSCs, while other Parp mRNAs were decreased (fig. S7B).
   This is consistent with previously published transcriptomic data from
   MuSCs following cardiotoxin injury, which found a peak of Parp1
   expression at 2 dpi (fig. S7C) ([130]9). The PARP1-specific inhibitor
   EB47 ([131]31), like BGP15, significantly suppressed Ccl2 expression in
   cytoLbNOX-overexpressing myoblasts but not in control cells ([132]Fig.
   3L). Together, these data indicate that PARP1 may mediate
   NAD^+-dependent neutrophil signaling and Ccl2 expression in ASCs.

Hypoxic regulation of immunomodulatory gene expression is under circadian
control

   Given the established role of the SC circadian clock in regulating
   hypoxic NAD^+ regeneration in myoblasts and muscle repair capacity
   following [injury] ([133]4), we hypothesized that NAD^+-induced
   immunomodulatory gene expression might also be under circadian control.
   In support of this, total RNA sequencing of wild-type and Bmal1^−/−
   primary myoblasts exposed to acute hypoxia (6 hours, 1% O[2]) (fig.
   S8A) revealed a marked down-regulation of chemokine signaling
   pathway–related genes (fig. S8, B and C). Previously, we demonstrated
   in cultured myoblasts that Bmal1-deficent myoblasts exhibit a reduction
   in hypoxic growth and myogenic capacity, which can be partially rescued
   by cytoLbNOX-expression ([134]4). Building upon these findings, we
   investigated the effect of cytoLbNOX expression in rescuing the muscle
   regeneration defects of SC-specific Bmal1 knockout (Bmal1^musc) mice in
   vivo (fig. S8D). While cytoLbNOX expression partially restored the
   regenerative defects of Bmal1^musc mice in terms of fiber size, the
   overall muscle integrity in cytoLbNOX-expressing mice was worsened
   compared to control mice (fig. S8, E and F). This may be attributed to
   the cytoLbNOX-induced neutrophil activation mechanism described in this
   study, as we observed a pronounced presence of mononucleated cells in
   the muscles of cytoLbNOX-expressing mice at day 7 post-injury (fig.
   S8E).

   Next, we performed cardiotoxin-induced injuries in TA muscles at ZT4
   and ZT16, corresponding to periods of low and high BMAL1 activity,
   respectively ([135]4, [136]32). ASCs were isolated at 24 hours (1 day)
   post-injury ([137]Fig. 4A) and subjected to total mRNA sequencing
   ([138]Fig. 4B). Aligning with our previous findings that the HIF
   signaling pathway is modulated by circadian clock ([139]4, [140]33),
   genes related to the hypoxia response were significantly more
   up-regulated in ASCs from muscles injured at ZT16 than at ZT4
   ([141]Fig. 4C). Notably, among the genes specifically induced in ASCs
   following injury at ZT16 versus ZT4 were those associated with
   inflammatory response, particularly neutrophil chemotaxis and
   chemokine-mediated signaling pathways ([142]Fig. 4C). These
   TOD-dependent differences were not observed in homeostatic QSCs at both
   ZT4 and ZT16, suggesting that they arise following injury (fig. S9).

Fig. 4. TOD-dependent immunomodulatory expression in ASCs at day 1
post-injury.

   [143]Fig. 4.
   [144]Open in a new tab

   (A) Schematic diagram of cardiotoxin-induced injury and SC isolation
   experiments in Pax7-zsgreen mice. (B) Volcano plot showing
   differentially expressed genes (DEGs, adjusted P < 0.05) in ASCs
   harvested from injured TA muscles at ZT4 compared to ZT16 at day 1
   post-injury. (C) Signaling pathway enrichment analysis for genes more
   highly expressed in ASCs at ZT16 compared to ZT4 at 1 dpi. (D) UMAP of
   neutrophil and MuSC single-cell transcriptomes split by the TOD on day
   1 post-injury. (E) Heatmap of ligand activities by NicheNet analysis
   showing ranked potential ligands expressed in the signal sender (MuSCs)
   that could predict the observed DEGs in neutrophil populations between
   ZT4 and ZT16 on day 1 post-injury. (F) Violin plot comparing expression
   of Ccl2 in ASCs at ZT4 and ZT16 on day 1 post-injury. (G) Schematic
   diagram of myoblast synchronization experiment. (H) Expression of Ccl2,
   Bmal1, and Per2 over the 24-hour cycle in dexamethasone-synchronized
   myoblasts under hypoxia. (I) Immunoblot analysis and relative
   quantification of total and serine-536–phosphorylated NF-κB p65 subunit
   and β-actin in synchronized myoblasts under hypoxia at CT4 and CT16,
   following a 6-hour treatment with or without EB47. (J) Ccl2 expression
   in synchronized myoblasts under hypoxia at CT4 and CT16, following a
   6-hour treatment with or without EB47. *P < 0.05 by two-way ANOVA
   analysis. ns, nonsignificant.

   Our mononucleated muscle cell samples for single-cell RNA sequencing
   described earlier were collected at both ZT4 and ZT16 from both
   homeostatic TA muscles and those injured at ZT4 and ZT16 on 1 and 3 dpi
   ([145]Fig. 2A). Reexamining the TOD-dependent dynamics within the UMAP
   encompassing neutrophils and MuSCs ([146]Fig. 4D), we used NicheNet
   ([147]34) to dissect TOD-dependent communication differences between
   MuSCs and neutrophils, designating MuSCs as the signal “sender” and
   neutrophils as the “receiver.” According to NicheNet’s guidelines, a
   Pearson correlation coefficient score around 0.10 or higher indicates
   statistically significant ligand-mediated cross-talk. Our results
   showed more substantial ligand-mediated interactions between MuSCs and
   the Neutrophils_1 population on day 1 post-injury, while interactions
   with Neutrophils_2 did not achieve a high Pearson score ([148]Fig. 4E).
   The differences in these interactions were not directly tied to cell
   proportion differences, as the Neutrophils_1 proportion remained
   similar between time points (35.2% at ZT16 versus 37.6% at ZT4) (fig.
   S10A). In contrast, although the Neutrophils_2 population was nearly
   twice as large at ZT4 compared to ZT16 (9.9% at ZT4 versus 4.6% at
   ZT16) (fig. S10A), the ligand-mediated interactions with MuSCs were
   weaker, suggesting that the observed differences in cell-cell
   communication were driven by factors beyond cell number differences.
   Ccl2 emerged as the sole CCL-type chemokine among the top ligands,
   exhibiting high activity in inducing target gene expression in
   Neutrophils_1 ([149]Fig. 4E and fig. S10, B and C). Furthermore, Ccl2
   gene expression was notably higher in ASCs at ZT16 compared to ZT4 on 1
   dpi ([150]Fig. 4F and fig. S2B).

   We next investigated whether the PARP1-mediated mechanism contributes
   to the TOD-dependent changes in Ccl2 expression levels. To address this
   question, we used an in vitro synchronization approach, a
   well-established model for studying the self-sustaining properties of
   circadian rhythms ([151]35–[152]38). Primary myoblasts were
   synchronized with dexamethasone (1 μM for 1 hour) to align their
   internal circadian clocks and then exposed to hypoxia to simulate the
   conditions that ASCs experience on the first dpi ([153]Fig. 4G). Cells
   were harvested every 4 hours to measure Ccl2 expression over the
   24-hour circadian cycle. Our quantitative polymerase chain reaction
   (qPCR) data revealed a clear circadian oscillation of Ccl2, mirroring
   the expression pattern of Bmal1, with Ccl2 expression being low at
   circadian time 4 (CT4) and high at CT16 ([154]Fig. 4H). We then
   assessed NF-κB activation in synchronized myoblasts at CT4 and CT16
   under hypoxia, in the presence and absence of the PARP1 inhibitor EB47,
   via Western blot analysis. NF-κB activity, indicated by an increase in
   phosphorylation of the NF-κB p65 subunit at serine-536, was higher in
   synchronized myoblasts at CT16, a time point that correlated with
   elevated Bmal1 expression levels ([155]Fig. 4H), compared to CT4 under
   hypoxia ([156]Fig. 4I). Following treatment with EB47 for 6 hours, we
   observed reduced NF-κB p65 phosphorylation at serine-536 in
   synchronized myoblasts at CT16 but not at CT4 ([157]Fig. 4I). While
   hypoxia-induced Ccl2 expression was significantly higher at CT16
   compared to CT4, treatment with EB47 abolished this difference
   ([158]Fig. 4J).

Ccl2 expression is associated with TOD-dependent differences in SC
proliferation

   A previous study demonstrated that conditioned medium from neutrophil
   cultures can promote myoblast proliferation while suppressing their
   differentiation ([159]39). This prompted us to investigate whether the
   TOD-dependent interactions between MuSCs and neutrophils might
   reciprocally influence SC fate decisions during muscle regeneration. To
   explore this, we analyzed the cellular heterogeneity within the MuSC
   populations from our single cell–transcriptomic atlas. Unsupervised
   clustering identified seven distinct subpopulations of myogenic cells
   ([160]Fig. 5A). Notably, cells in cluster 3 exhibited high expression
   of Pax7 and genes like Spry1, Cd34, Cdkn1b, and Pdk4, which are
   recognized markers of deep QSCs (dQSCs) ([161]40–[162]42). Further
   classification of the clusters was based on gene sets indicative of
   early commitment (e.g., Myf5, Egr2, Egr3, Myc, Jund, Dnajb1, Ccl8)
   ([163]41, [164]42), activation/proliferation (Myod1, Mki67, Cdk1,
   Ccnb1) ([165]43), differentiation (Myog, Myl4) ([166]7, [167]44), and
   self-renewal (Cdk6, Ccnd1, Tgfbr3, Smad4) ([168]45), leading to their
   designation as primed QSC 1 and 2, committed SCs, dividing SCs,
   differentiating SCs, and renewing SCs ([169]Fig. 5A and fig. S11A). A
   temporal analysis on UMAP highlighted that QSCs and some committed SCs
   were prevalent in homeostatic muscles, whereas dividing and
   differentiating SCs were mainly observed in regenerating muscles by 3
   dpi (fig. S11B).

Fig. 5. Ccl2 secretion by SCs promotes TOD-dependent proliferative
differences.

   [170]Fig. 5.
   [171]Open in a new tab

   (A) UMAP visualization of the subclustered MuSC populations, with
   subclusters defined as dQSC, primed quiescent SCs (pQSC). committed SCs
   (Com_SC), dividing SCs (Div_SC), differentiating SCs (Diff_SC), and
   renewing SCs (Ren_SC). (B) Pseudotime trajectory inference analysis of
   MuSC subclusters by Slingshot, with deep QSCs serving as the starting
   point of pseudotime. (C) Pseudotime ordered single-cell expression
   trajectories for the genes that were most differentially expressed
   between the time points of ZT4 and ZT16 in lineage 1 and 2. (D)
   Signaling pathway enrichment analyses on the DEGs from (C). (E)
   Pseudotime ordered single-cell expression trajectories for the cell
   cycle regulators Cdk1 and Mki67 at ZT4 and ZT16 within lineage 2. (F)
   Representative FACS plot for MuSC gating (zsGreen^+) at 3 dpi (top) and
   quantification of MuSC frequency within total mononucleated muscle
   cells (bottom). *P < 0.05 by unpaired Student’s t test (n = 5 mice per
   time point). (G) Schematic diagram of experiment using wild-type mice
   transplanted with myoblasts expressing GFP^+ retrovirus and either Ccl2
   (Ccl2-virus) or a control (R1) retroviruses (n = 4).
   Cardiotoxin-induced TA muscle injuries were performed at ZT16.
   Transplantation of 10^5 SCs occurred at 6 hours post-injury. At day 3
   post-transplantation, total MuSC frequency and the percentage of GFP^+
   SCs within the total MuSC population were assessed by flow cytometry.
   (H) Quantification of CCL2 by ELISA in the media after 24 hours of
   culture from equal number of myoblasts infected with either
   Ccl2-expressing or control retrovirus. (I) Quantification of total
   MuSCs and GFP^+ MuSCs in recipient TA muscles as described in (G).
   *P < 0.05 by unpaired Student’s t test.

   To better understand the hierarchical organization and validate the
   classification of these myogenic subclusters, we applied Slingshot
   ([172]46), a single-cell trajectory inference analysis, to the entire
   myogenic cell population within the atlas. This analysis delineated a
   sequential trajectory from deep quiescence through primed quiescence
   and commitment/activation, leading to division, which then bifurcated
   toward self-renewal (lineage 1) or differentiation (lineage 2)
   ([173]Fig. 5B). Along this pseudotime trajectory, Pax7 expression
   gradually decreased, while Myod1 expression increased in both lineages
   (fig. S11C). In the self-renewal lineage, Myog expression spiked before
   sharply decreasing, whereas in the differentiation lineage, Myog
   expression steadily rose (fig. S11C).

   To elucidate differential gene expression patterns between ZT4 and ZT16
   within each myogenic lineage, we used the “conditionTest” function from
   the tradeSeq package ([174]47). This analysis revealed 88 genes with
   significant differential expression (adjusted P value < 0.05, fold
   change > 1) across pseudotime ([175]Fig. 5C). Subsequent KEGG pathway
   enrichment analysis linked these genes predominantly to the cell cycle
   ([176]Fig. 5D). We further assessed the expression trajectories of two
   key cell division markers, cyclin-dependent kinase 1 (Cdk1) and Mki67,
   over pseudotime at both ZT4 and ZT16. Notably, while there was an
   initial surge in the expression of both regulators at lower pseudotime
   values within the differentiation lineage (lineage 2) of ZT4 myogenic
   cells, but not in ZT16 cells, this increase appeared to be driven by a
   single cell ([177]Fig. 5E). Overall, however, the expression levels of
   Cdk1 and Mki67 in ZT16 cells within this lineage rose more rapidly than
   in ZT4 cells, eventually reaching comparably high levels at later
   pseudotime stages ([178]Fig. 5E).

   The expression dynamics we observed suggest a potential regulatory
   effect of the TOD on key cell cycle regulators, which may in turn
   influence the proliferation trajectory of SCs after activation. To
   investigate this influence, we induced injury in TA muscles at ZT4 and
   ZT16 and then quantified the frequency of SCs in the total muscle cell
   population by day 3 post-injury, a peak time for dividing myonuclei
   ([179]48), using Pax7-zsgreen reporter mice ([180]49). Flow cytometry
   analysis revealed a significantly higher frequency of zsgreen^+ SCs in
   TA muscles injured at ZT16 compared to those injured at ZT4 at 3 dpi
   ([181]Fig. 5F). However, no difference in the frequency of zsgreen^+
   SCs was observed in homeostatic TA muscles between ZT4 (1.005% ± 0.26)
   and ZT16 (1.128% ± 0.32).

   To determine the extent to which SC-expressed CCL2 contributes SC
   proliferation, we transduced freshly isolated SCs with a
   Ccl2-expressing retrovirus and transplanted them into TA muscles of
   wild-type C57BL/6 mice 6 hours following cardiotoxin injury ([182]Fig.
   5G). The contralateral TA muscles were subjected to the same preinjury
   but received control virus (R1)–transduced SCs. The transplanted
   myogenic progenitors, labeled by green fluorescent protein (GFP)
   carried by the viral vector, allowed for the tracking of GFP^+-SCs
   within the total SCs population, characterized as
   VcamI^+Scal1^−CD31^−CD45^− ([183]50), at 3 dpi ([184]Fig. 5G). Before
   this experiment, we first verified the role of CCL2 as a secreted
   chemokine by comparing its levels in culture media from myoblasts
   infected with either Ccl2-expressing or control retroviruses using
   enzyme-linked immunosorbent assay (ELISA) ([185]Fig. 5H). Notably, we
   observed a significantly higher frequency of Ccl2-overexpressing SCs in
   the TA muscles of recipient mice compared to those transplanted with
   control virus-infected SCs ([186]Fig. 5I). This increase likely
   resulted from the CCL2 released into the niche by the exogenous SCs,
   which was also reflected in an increase in the total MuSC population
   compared to muscles receiving control virus-infected SCs ([187]Fig.
   5I).

Neutrophil depletion diminishes TOD effects on SC and muscle regeneration

   Our previous studies, along with others, showed a TOD-dependent
   variation in skeletal muscle repair following cardiotoxin-induced
   injury, with more effective regeneration occurring during active phase
   compared to the rest phase ([188]4, [189]5). The enduring impact of
   this TOD effect remains underexplored. To investigate whether the
   difference is transient or has long-lasting consequences, we induced TA
   muscle injuries with cardiotoxin at ZT4 or ZT16, repeating the process
   three times at 30-day intervals ([190]Fig. 6A). Histological
   examination conducted 30 days after the last injury revealed that TA
   muscles subjected to repeated injuries at ZT4, corresponding to a
   period of lower BMAL1 activity, exhibited a higher proportion of small
   regenerated myofibers and lower proportion of large myofibers compared
   to those injured at ZT16 ([191]Fig. 6, B and C). In addition, we
   assessed muscle strength by measuring torque under continuous
   electronic stimuli at multiple frequencies at the corresponding time of
   injuries ([192]Fig. 6D). As has been previously reported, max tetanic
   force production in the uninjured TA muscle exhibited variations
   depending on the TOD, with increased peak force during tetanic
   contractions (Max tetanic) observed in the early rest phase (ZT4)
   compared to the active phase (ZT16) ([193]Fig. 6E) ([194]51–[195]55).
   We also assessed recovery of muscles following repeated injuries at
   either ZT4 or ZT16 by comparing the Max tetanic of injured muscles to
   that of uninjured ones at identical times of day. While the interaction
   between TOD and injury status did not reach significance by two-way
   ANOVA, we observed a trend toward reduced recovery in mice injured at
   ZT4, as compared to ZT16 ([196]Fig. 6E). These findings point toward a
   potential lasting influence of TOD on muscle health and regeneration.

Fig. 6. Reduction of TOD-dependent variations in SC proliferation and muscle
regeneration through neutrophil depletion.

   [197]Fig. 6.
   [198]Open in a new tab

   (A) Schematic diagram of repeated cardiotoxin-induced injury at
   different times of day in wild-type mice (n = 3 per time point). Mice
   were euthanized at 30 days post-final injury, and TA muscles were
   assessed for muscle remodeling and function. (B) Representative
   hematoxylin and eosin (H&E) staining and (C) quantification of myofiber
   size distribution of the repeatedly injured TA muscles at ZT16 and ZT4
   described in (A). (D) Measurement of torque variations in response to
   different stimulation frequencies in the TA muscles that were
   repeatedly injured at ZT4 and ZT16. Torque at each frequency along the
   force frequency curve plotted as mean ± SEM. (E) Peak torque generated
   in the force frequency curve (mean; SEM) for each group: ZT16u (1.526;
   0.2237), ZT16i (1.200; 0.1101), ZT4u (2.470; 0.2365), ZT4i (1.614;
   0.1968). (F) Schematic diagram of Pax7-zsGreen mice following transient
   neutrophil depletion through three consecutive intraperitoneal
   injections of Ly6G antibody (n = 5 per group). Control mice received
   IgG antibody injections from the same host species as the Ly6G
   antibody. (G) Quantification of zsGreen^+ SCs on 3 dpi in the injured
   TA muscles from both IgG and Ly6G-treated mice (top), and in TA muscles
   from neutrophil-depleted mice injured at ZT4 and ZT16 (bottom).
   **P < 0.01 by unpaired Student’s t test. ns, nonsignificant. (H)
   Representative H&E staining of TA muscles from IgG and Ly6G treated
   mice that were injured at ZT4 and ZT16. (I) Quantification of myofiber
   size distribution in the injured TA muscles shown in (H). *P < 0.05,
   **P < 0.01, ***P < 0.001 by two-way ANOVA test.

   To further probe the role of neutrophils in mediating the TOD effect on
   muscle regeneration, we depleted neutrophils in the early stages
   post-injury. Depletion was achieved through three successive
   intraperitoneal injections of a Ly6G antibody. On the day of the final
   injection, we induced TA muscle injuries with cardiotoxin and examined
   neutrophil infiltration 18 hours later (fig. S12A). Flow cytometry
   analysis confirmed that the Ly6G antibody specifically reduced
   neutrophil infiltration compared to the control immunoglobulin G (IgG)
   antibody, with minimal impact on other immune cells like monocytes
   (fig. S12B). Leveraging this neutrophil depletion strategy, we
   evaluated its effect on SC proliferation at 3 dpi and on muscle
   regeneration at 7 dpi, following cardiotoxin-induced injuries at ZT4
   and ZT16 ([199]Fig. 6F). Flow cytometry–based cell quantification
   showed that neutrophil depletion reduced the frequency of SCs in TA
   muscles injured at ZT16 compared to those treated with control IgG
   antibody ([200]Fig. 6G, left). The TOD-dependent disparities in SC
   proliferation at 3 dpi are much less prominent in the absence of
   neutrophils ([201]Fig. 6G, right) and in muscle regeneration at 7 dpi
   ([202]Fig. 6, H and I), suggesting that neutrophils are key modulators
   of the TOD effect in muscle repair processes.

   Together, our study supports a model whereby hypoxic cytosolic NAD^+
   regeneration following muscle injury bolsters cell proliferation and
   concurrently modulates proinflammatory signaling, as induced by the
   activation of PARP1 and NF-κB pathways. This process is crucial for the
   chemotactic recruitment of neutrophils and their interaction with
   MuSCs, facilitating stem cell activation and proliferation. The
   circadian clock further influences this process by fine-tuning the
   hypoxia-induced anaerobic NAD^+ regeneration, imparting TOD-dependent
   differences in muscle repair efficacy ([203]Fig. 7).

Fig. 7. Proposed model depicting the process underlying TOD-dependent muscle
regeneration.

   [204]Fig. 7.
   [205]Open in a new tab

   Schematic depicting the process underlying TOD-dependent muscle
   regeneration. Following injury, skeletal muscle experiences passive
   hypoxia, triggering enhanced anaerobic glycolysis in SCs. The
   regeneration of cytoplasmic NAD^+ during this process acts as a
   substrate as activated PARP1, which then activates NF-κB through the
   phosphorylation of the p65 subunit at serine-536. This activation
   further promotes the expression of immunomodulator genes (e.g., Ccl2),
   aligning with a peak in neutrophil infiltration into the wounded site.
   This synchronization facilitates interactions between SCs and
   neutrophils, reciprocally enhancing SC proliferation and muscle repair
   efficiency. Given that the glycolysis pathway is regulated by the
   circadian clock, this signaling cascade contributes to variations in
   muscle repair efficacy depending on the TOD.

DISCUSSION

   Metabolic processes are essential for biological functions, with
   glycolysis being a key player in energy production by converting
   glucose into pyruvate, which fuels both aerobic and anaerobic pathways
   ([206]56). Notably, glycolysis is important for rapidly dividing cells,
   including tumor and immune cells, to meet their energy demands. Immune
   cells predominantly rely on glycolysis when activated, to bolster an
   inflammatory response, but switch back to oxidative phosphorylation
   after the inflammation subsides ([207]57). This metabolic shift is
   closely associated with inflammatory signaling, highlighted by
   increased glycolysis driving inflammatory markers in endothelial cells
   via HIF1α up-regulation ([208]58) and the dampening of NF-κB activation
   when glycolysis is inhibited ([209]59, [210]60). Similarly, our
   findings reveal elevated NF-κB signaling (fig. S7A) and inflammatory
   marker expression ([211]Fig. 3D and fig. S5) in myoblasts under hypoxic
   conditions, aligning with known glycolysis-inflammation dynamics.
   Mechanistically, glycolytic intermediates and enzymes have been
   implicated in modulating inflammation. For example, the accumulation of
   phosphoenolpyruvate in T cells encourages a proinflammatory phenotype
   ([212]61, [213]62), and lactate boosts proinflammatory activities
   through NF-κB pathways ([214]63, [215]64). In addition, certain
   glycolytic enzymes act as posttranscriptional regulators of
   inflammatory genes by binding to their mRNAs ([216]65, [217]66). These
   findings underscore the impact of glycolysis on immune cell activity.

   The reason why rapidly proliferating cells discard carbon by expelling
   pyruvate-converted lactate remains unclear, although recent insights
   suggest a preference for converting pyruvate to lactate via LDH when
   the need for NAD^+ regeneration surpasses adenosine triphosphate (ATP)
   demand ([218]67). While glycolysis facilitates NAD^+ regeneration
   critical for maintaining its flux and ATP production, our study
   suggests an added role of LDHA-mediated NAD^+ regeneration within SCs
   to regulate early muscle repair. Like cancer cells, glycolysis
   dominates in proliferating myoblasts even under highly oxygenated
   culture conditions ([219]68), underscoring its importance in supporting
   SC proliferation. Nonetheless, given that QSCs take more than 24 hours
   to commence and complete their first division ([220]69), while the rate
   of glycolysis in SCs is reported to increase as soon as they break
   quiescence, reaching threefold within 3 hours of activation ([221]70),
   ATP might not be their primary need immediately post-injury. Our
   findings indicate that beyond supporting glycolysis, NAD^+ regeneration
   is key for immune modulation, activating PARP and NF-κB pathways to
   induce chemokine production, thus orchestrating neutrophil recruitment
   and SC interaction. This mechanism is backed by the exclusive induction
   of Parp1 (fig. S7B) and enhanced expression of genes associated with
   DNA damage responses, a known trigger for PARP1 activation ([222]71)
   ([223]Fig. 1E) in 1dpi ASC versus QSCs. Given the rapid NAD^+-consuming
   capacity of PARP1 ([224]72) and its ability to prioritize NAD^+ over
   Sirtuin proteins ([225]73), our results suggest that alongside
   sustaining glycolytic flux, NAD^+ regeneration supports PARP1
   activation, which may be crucial for immune modulation in the initial
   muscle regeneration phase. However, this conclusion appears to
   challenge the established dogma of redox biology. If regenerated NAD^+
   is consumed for PARP1 activation, glycolytic flux would be expected to
   collapse because of insufficient replenishment of NAD(H). We
   hypothesize that anaerobic NAD^+ regeneration by LDHA plays a critical
   role in maintaining the NAD^+/NADH redox state, preventing glycolytic
   collapse under hypoxic conditions. Supporting this, our data
   demonstrate that galloflavin, a lactate dehydrogenase inhibitor that
   impedes NAD^+ regeneration, significantly reduced Ccl2 expression under
   hypoxia ([226]Fig. 3C). Notably, our RNA sequencing data revealed
   significant up-regulation of Nampt (log[2]FC = 0.84, adjusted P = 9.68
   × 10^−21) in myoblasts after 6 hours of hypoxia. This suggests that,
   while anaerobic glycolysis is NAD^+ neutral, increased NAD^+ synthesis
   via the salvage pathway may contribute to maintaining cellular NAD^+
   levels during hypoxia. Our data demonstrating temporal increases in
   intracellular NAD^+ levels in hypoxia-exposed myoblasts support this
   hypothesis ([227]Fig. 3A). This observation aligns with a previous
   report showing elevated NAD^+ levels in human umbilical cord
   endothelial cells under hypoxia, with increases at 10 and 60 min and a
   significant 57% increase after 120 min of incubation ([228]74).
   Similarly, significant increases in both NAD^+ levels and the
   NAD^+/NADH ratio were observed in endothelial progenitor cells under
   hypoxia ([229]75). We propose that regenerated NAD^+ contributes at
   least partially to PARP1 activation, which subsequently activates NF-κB
   and induces Ccl2 expression under hypoxic conditions. In addition,
   newly synthesized NAD^+ via the salvage pathway may further support
   PARP1 activity and help maintain glycolytic flux, reinforcing this
   mechanism. Whether lactate plays a role in this process remains an open
   question that warrants further investigation.

   Neutrophils play an indispensable role in the innate immune system,
   serving as the frontline defense against pathogens ([230]76). They
   navigate from the bone marrow to infection or inflammation sites by
   adhering to the leukocyte adhesion cascade ([231]77). This migration
   exhibits a circadian rhythm, potentially linked to the cyclic
   expression of genes related to migration, suggesting that NAD^+
   regeneration, triggering chemokine release, contributes to neutrophil
   recruitment across different muscle cell types. Furthermore, the
   circadian clock intricately modulates glycolytic activity by
   controlling the HIF1α-mediated response to hypoxia ([232]33, [233]78).
   In particular, our previous research highlighted that Bmal1 in SCs or
   myoblasts regulates the expression of critical glycolytic genes,
   including Glut1, Hk2, and Ldha, influencing glycolysis and the
   anaerobic regeneration of NAD^+ ([234]4). Thus, the circadian
   regulation of neutrophil infiltration is also influenced by peripheral
   clock residing in the surrounding cell populations.

   Muscle injury leads to the necrosis and degeneration of muscle fibers,
   accompanied by hematoma formation. Neutrophils rapidly migrate to the
   injury site within hours of damage ([235]79). Studies involving mice,
   where intraperitoneal injections of antisera targeting neutrophils were
   used alongside myotoxin injections, highlight the critical role of
   neutrophils in efficient muscle repair ([236]80). Neutrophils could
   contribute to muscle regeneration by clearing tissue debris ([237]81)
   and by recruiting monocytes and macrophages through the release of
   inflammatory cytokines and chemokines ([238]82). A recent in vitro
   study further underscored the influence of neutrophils on muscle
   regeneration. When myoblasts were treated with media conditioned by
   activated neutrophils, there was a notable promotion of myoblast
   proliferation and a suppression of differentiation compared to
   treatment with basal media ([239]39). Specifically, the study found
   that inhibiting the neutrophil-derived factor Ccl3 in the conditioned
   media nullified the proliferative effects, while blocking Mmp-9
   encouraged myoblast differentiation ([240]39). Intriguingly, Ccl3
   expression was significantly higher at ZT16 compared to ZT4 in the
   Neutrophil_1 population on day 1 post-injury (log[2]FC = 1.636927839,
   Adj. P value = 0) (fig. S10C), suggesting that the interaction between
   SCs and neutrophils leads to TOD-dependent variations in SC
   proliferation. Further studies are needed to understand whether
   TOD-dependent regulation of neutrophil activity can reciprocally affect
   SC proliferation and myogenic capacity.

   Within the subset of neutrophil chemotaxis–related genes that were
   up-regulated in ASCs at 1 dpi compared to QSCs, the expression of Ccl2
   showed TOD-dependent variations, being significantly higher in ASCs at
   ZT16 than at ZT4 ([241]Fig. 4F and fig. S2B). Ccl2 was the only
   chemokine gene among the top ligands that exhibited high activity in
   inducing target gene expression in Neutrophils_1 ([242]Fig. 4E and fig.
   S10C). CCL2 is recognized as a potent chemoattractant for neutrophils
   ([243]21), and previous study has indicated its capacity to attract
   substantial neutrophil numbers to the lung even without underlying
   inflammation ([244]83). Moreover, experiments in mice deficient in Ccl2
   have revealed impaired infiltration of monocyte/macrophages, leading to
   an altered regenerative process that favors adipogenic over myogenic
   differentiation ([245]84, [246]85). In addition, a decrease in Ccl2
   expression, resulted from CD8 knockout, has been associated with
   compromised muscle regeneration and increased fibrosis ([247]86).
   However, these studies did not explore the role of Ccl2 expressed by SC
   during muscle regeneration. Our findings, through transplantation of
   Ccl2-overexpressing SC-derived myoblasts, demonstrated an increase in
   the presence of these exogenous progenitors within the proliferating
   muscle stem cell pool of regenerating muscles at day 3 post-injury
   compared to control myoblasts ([248]Fig. 5I). This highlights the key
   role of neutrophil interactions in the activation and proliferation of
   SCs.

   The circadian regulation of muscle regeneration has been independently
   established in prior studies from our group and others
   ([249]4, [250]5). While the TOD-dependent differences in muscle repair
   with cardiotoxin are statistically significant and reproducible, they
   are not as pronounced as those observed in mice with systemic or
   SC-conditional core clock gene deletions when assessed histologically
   at day 7 post-injury ([251]4, [252]5, [253]87). Moreover, one study
   noted that the initial differences in regenerating fiber size observed
   between injuries inflicted at ZT14 and ZT2 seem to equalize by day 14
   post-injury ([254]5). Here, we demonstrate that the effects of TOD of
   muscle injury can be compounded with repeated injuries. Specifically,
   repeated injury during the rest period (i.e., ZT4 for a mouse) led to
   differences in the size of regenerated myofibers, even at 1 month
   following the last injury event. In addition, this coincided with a
   trend toward changes in muscle function, reinforcing the long-term
   negative impact of muscle injury during the rest phase, although
   further studies are required to establish the exact impact of TOD
   injury on long-term changes in muscle function. Overall, these findings
   raise important questions regarding the potential pathological aspects
   of impaired muscle regeneration in conditions such as obesity,
   diabetes, and aging, where circadian rhythms are continuously
   disrupted.

   The current study has certain limitations in its experimental design
   and data interpretation. While our signaling pathway enrichment
   analysis of QSCs at ZT4 and ZT16 (fig. S9) indicates that circadian
   regulation in the stem cell transcriptome differs significantly from
   the transcriptional changes observed in 1-dpi ASCs following injury at
   ZT4 and ZT16 ([255]Fig. 4C), we acknowledge that we cannot completely
   exclude the potential influence of circadian phase on gene expression
   in 1-dpi ASCs. Addressing this limitation would require performing RNA
   sequencing on ASCs collected at both ZT4 and ZT16 from muscles injured
   at these respective times. Last, while we observed an increase in total
   cellular NAD^+ levels in myoblasts following acute hypoxia exposure,
   the impact of hypoxia and the circadian clock on the free cytonuclear
   NAD^+ pool—crucial for regulating PARP activity—remains unclear.

   In summary, we have revealed a previously unrecognized mechanism that
   links NAD^+ regeneration during anaerobic glycolysis with neutrophil
   modulation. This interaction, in turn, affects the activation and
   proliferation of muscle stem cells, thereby influencing the
   effectiveness of muscle regeneration. Further research is needed to
   elucidate the role of clock-regulated NAD^+ metabolism in immune
   modulation, particularly in the context of acute and chronic disease
   states.

METHODS

Mice

   C57BL/6J mice (#000664) and B6.Cg-Pax7^tm1(cre/ERT2)Gaka/J mice
   (#017763) were acquired from The Jackson Laboratory. CytoLbNOX^LSL/LSL
   mice, characterized by a Lox-Stop-Lox (LSL) cassette upstream of
   CytoLbNOX and an internal ribosomal entry site–linked enhanced GFP,
   were provided by N. Chandel at Northwestern University. Pax7-zsGreen
   mice were supplied by M. Kyba at the University of Minnesota. To
   achieve SC-conditional expression of CytoLbNOX, we crossed
   CytoLbNOX^LSL/LSL mice with B6.Cg-Pax7^tm1(cre/ERT2)Gaka/J mice,
   resulting in CytoLbNOX^LSL/LSLPax7^tm1(cre/ERT2) and control
   CytoLbNOX^+/+Pax7^tm1(cre/ERT2) mice. CytoLbNOX expression was
   initiated through five consecutive daily intraperitoneal injections of
   tamoxifen (20 mg/ml in corn oil, 100 mg/kg body weight). The mice were
   housed under a 12:12 light:dark cycle. All procedures were carried out
   with male mice aged 8 to 12 weeks, in strict accordance with the
   guidelines of the Institutional Animal Care and Use Committee at
   Northwestern University (protocol #IS00026931).

Neutrophil depletion

   Neutrophil depletion in vivo was conducted through intraperitoneal
   injections of 300 μg the 1A8 monoclonal anti-mouse Ly6G antibody,
   administered daily over a period of three consecutive days. As a
   control, the corresponding rat IgG2a isotype control antibody was
   injected in a similar manner. The efficacy of neutrophils and monocyte
   depletion post-antibody treatment was verified by flow cytometry
   analysis.

Skeletal muscle injury

   Cardiotoxin-induced TA muscle injury was conducted according to
   methodologies previously described ([256]4). Mice underwent anesthesia
   via intraperitoneal administration of a ketamine (80 mg/kg) and
   xylazine (10 mg/kg) mixture. Subsequent to the removal of fur from the
   injection site, TA muscles received a 50-μl injection of a 10 μM
   cardiotoxin solution in saline, delivered using a 0.5-ml BD insulin
   syringe.

Histology

   To evaluate muscle regeneration, TA muscles were harvested at indicated
   time points after injury and enveloped in a thin layer of Tissue-Tek
   O.C.T. compound. The samples were snap-frozen in liquid
   nitrogen–chilled isopentane and preserved at −80°C for later
   processing. Continuous 10-μm-thick cross sections of the frozen TAs
   were prepared using a Leica CM1860 cryostat. For hematoxylin and eosin
   staining, sections were air-dried overnight at room temperature.
   Subsequent rehydration in phosphate-buffered saline (PBS) for 5 min was
   followed by fixation in 10% formalin. The nuclei were then stained with
   hematoxylin for 5 min, with excess stain removed by rinsing under tap
   water for 5 min and briefly dipping in clarifier. Eosin was applied for
   30 s to stain the cytoplasm and connective tissues. The sections were
   dehydrated through a graded series of ethanol, cleared in xylene, and
   mounted with Permount mounting medium. The prepared slides were
   examined using a Keyence BZ-X800 microscope. Cross-sectional area of
   myofibers was quantified using FIJI software ([257]88) with
   LabelsToROIs plugin ([258]89).

Immunohistochemistry

   Tissue sections were allowed to air-dry for 30 min at room temperature
   and were then fixed with 4% paraformaldehyde and permeabilized using
   0.25% Triton X-100 for 15 min. For blocking, sections were treated with
   PBS containing 5% goat serum, 2% bovine serum albumin (BSA), and 1%
   Tween-20 for 1 hour. This was followed by an overnight incubation at
   4°C with primary antibodies, including anti-laminin (1:100), anti–Ly-6G
   (1:100) in a diluent buffer of PBS with 0.5% goat serum, 2% BSA, and 1%
   Tween-20. After washing in PBS thrice, the sections were incubated with
   an Alexa Fluor–conjugated secondary antibodies at a 1:1000 dilution for
   1 hour at room temperature. Sections were then mounted using
   VectaShield HardSet antifade mounting medium with
   4′,6-diamidino-2-phenylindole (DAPI) for visualization under a Keyence
   BZ-X800 microscope.

Immunofluorescence

   Primary myoblasts were rinsed with ice-cold PBS and fixed with methanal
   for 15 min at −20°C. They were then permeabilized using 0.2% Triton
   X-100 and blocked to prevent nonspecific binding with 1% BSA for 1 hour
   at room temperature. This was followed by incubation with an
   anti-Poly(ADP-ribose) polyclonal antibody at a 1:500 dilution overnight
   at 4°C. The following day, the myoblasts were washed thrice with PBS
   and incubated with an Alexa Fluor 594–conjugated goat anti-rabbit IgG
   secondary antibody at a 1:1000 dilution in 1% BSA for 1 hour at room
   temperature. Nuclear staining was performed with DAPI (5 μg/ml in PBS)
   ffor 5 min, and the samples were visualized under an Evos FL
   cell-imaging microscope.

In vivo muscle contractility measurement

   Mice were anesthetized using a precision vaporizer (Somnoflo; Kent
   Scientific) with initial induction of 2.5% isoflurane. After induction,
   a mouse was transferred to a 37°C motor stage (Aurora Scientific), and
   the appropriate leg was shaved from ankle to hip.

   Dorsiflexion torque was measured as previously described ([259]90),
   with the foot fixed to the footplate attached to a 300C series dual
   mode lever. Briefly, two needle electrodes were placed subcutaneously
   on either side of the peroneal nerve to activate dorsiflexors. Then, a
   torque-frequency curve was generated from 10 to 150 Hz as previously
   described ([260]90). Torque was recorded at each frequency. ZT4 mice
   (injured and uninjured) were tested at 9:00 a.m., while ZT16 mice
   (injured/uninjured) were tested at 1 p.m.

Retrovirus production

   The cDNA open reading frame for mouse chemokine (C-C motif) ligand 2
   (Ccl2) was ordered from GenScript and subsequently amplified by PCR
   using Phusion High-Fidelity DNA Polymerase & dNTP Mix. The amplified
   sequence was cloned between the Xho I and EcoR I restriction sites of
   the retroviral mammalian expression vector MIGR1 ([261]91). To produce
   retrovirus carrying the Ccl2 gene, 293T cells were transfected with a
   DNA comprising 1.5 μg of pCL-Eco ([262]92) and 2.5 μg MIGR1-Ccl2, or
   MIGR1 vector using Lipofectamine 3000 Transfection Reagent according to
   the provided instructions. The retrovirus-enriched supernatant was
   collected 24 hours post-transfection and was then filtered through a
   syringe filter with a pore size of 0.45 μm to ensure purity.

Muscle cell isolation

   For the in vitro culture of myoblasts, muscle cells were extracted from
   the entire hindlimb muscles. Bulk RNA sequencing involved the use of
   Pax7-zsgreen mice, which were sequentially perfused with 50 ml of
   ice-cold PBS, followed by 50 ml of mild fixative (0.5%
   paraformaldehyde), and subsequently quenched with 50 ml of 2 M glycine,
   as previously reported ([263]43, [264]93). SCs were collected from the
   TA muscles on both sides. For single-cell RNA sequencing, mononucleated
   muscle cells were isolated from TA muscles on both hindlimbs. Muscle
   cell preparation was performed by following previously established
   procedures ([265]4). Specifically, excised muscle tissues were digested
   in Hams F-10 medium supplemented with collagenase D (5 mg/ml) and
   dispase II (5 mg/ml) for 30 min at 37°C with shaking at 500 rpm. The
   enzymatic digestion was terminated by adding a fivefold volume of
   ice-cold cell suspension buffer [Ham’s F-10 with 10% fetal bovine serum
   (FBS) and 3 mM EDTA]. The digested muscle tissue was then subjected to
   mechanical dissociation through vigorous pipetting (20 times with a
   10-ml pipette) to release MuSCs from the myofibers. The cell suspension
   was filtered through 70- and 40-μm nylon mesh strainers in sequence,
   and the filtrate was collected in a 15-ml conical tube and centrifuged
   at 500g for 5 min. The resulting cell pellets were resuspended in 1 ml
   of Pre-Sort buffer, with propidium iodide serving as an indicator of
   cell viability. For bulk RNA sequencing, zsgreen-positive mononuclear
   SCs were sorted using a BD FACSAria II cell sorter.

   Myoblasts preparation was performed with slight modification to a
   previously established preplating technique ([266]94). Initially, the
   crude suspension was seeded and expanded on a 60-mm dish coated with
   Matrigel, in myoblast growth medium consisting of Hams F-10 with 20%
   FBS, 1% penicillin-streptomycin (PS), and basic fibroblast growth
   factor (bFGF, 2.5 ng/ml) for 3 days. Following this, the cells
   underwent trypsinization and were subjected to a preplating process on
   a regular tissue culture–treated 60-mm dish for 1 hour, before being
   transferred to another Matrigel-coated dish. This procedure was
   repeated three times to reduce the presence of nonmyoblast cells. The
   resultant myoblast-enriched cultures were then maintained in a growth
   medium composed of Dulbecco’s modified Eagle’s medium (DMEM)/Hams F-10
   (1:1) with 20% FBS, 1% PS, and bFGF (2.5 ng/ml). The purity of
   myoblast, verified to be over 95% using immunofluorescence with a
   Pax7-antibody (Developmental Studies Hybridoma Bank), underscored the
   effectiveness of this method. Induction of Cyto-LbNOX was conducted no
   sooner than 48 hours after primary myoblasts were infected with either
   an adenovirus expressing Cre recombinase (Adv-Cre) or a control empty
   vector (Adv-Ctrl) from Vector Biolabs.

SC transplantation

   SCs from wild-type C57BL/6 mice were isolated on the basis of a
   combination of cell surface markers CD45^−CD31^−Sca1^−VcamI^+
   ([267]50). These cells were then transduced overnight with either a
   control or Ccl2-expressing retrovirus, in the presence of polybrene (8
   μg/ml) within wells coated with Matrigel. To prepare for
   transplantation, TA muscles in recipient wild-type C57BL/6 mice
   underwent injury preconditioning with a 50 solution of 10 μM
   cardiotoxin solution in saline, administered 6 hours before cell
   transplantation. Subsequently, 10,000 SCs, transduced with either
   control or Ccl2-expressing retrovirus as denoted by GFP expression,
   were sorted, and intramuscularly injected into the pre-injured TA
   muscles on each side. The cells were delivered in a small volume of 15
   μl using 0.3-ml insulin syringes to ensure precise administration.

Bulk RNA sequencing

   Total RNA was extracted from in vivo light-fixed Pax7-zsgreen^+ SCs
   using the miRNeasy FFPE Kit, adhering to the supplied instructions with
   a crucial modification that the digestion of RNA by proteinase K was
   performed at 56°C for 1 hour as previously reported ([268]43). For the
   initial double-stranded cDNA synthesis, amounts ranging from 10 pg to
   10 ng of total RNA were used using the SMART-seq v4 Ultra Low Input RNA
   Kit from Takara Bio, following the protocol included within the kit.
   After PCR amplification, the resulting cDNA was purified using the
   AMPure XP beads. This purified cDNA underwent assessment using the
   Agilent 2100 Bioanalyzer paired with Agilent’s High-Sensitivity DNA
   Kit. Three hundred picograms of the full-length cDNA, derived from the
   SMART-seq v4 Ultra Low-Input RNA Kit, was then prepared for sequencing
   using the Nextera XT DNA Library Preparation Kit. The PCR-amplified,
   size-selected cDNA fragments were both visualized and quantified using
   the Agilent 2100 Bioanalyzer in conjunction with Agilent’s
   High-Sensitivity DNA Kit, and the Qubit dsDNA HS Assay Kit,
   respectively. The pooled libraries were sequenced using NovaSeq 6000 SP
   Reagent Kit v1.5 (100 cycles) on a NovaSeq 6000 system, executing a
   paired-end run [51 base pairs (bp), repeated twice] to achieve a
   sequencing depth of approximately 50 million reads per sample.

RNA sequencing analysis

   Raw BCL files were processed into demultiplexed, paired-end FastQ files
   using bcl2fastq conversion software (v2.19.1). Subsequently, adaptor
   sequences were trimmed from these files using Trimmomatic (v0.33)
   ([269]95). The processed reads were then aligned to the mm10 Mus
   musculus reference genome using the STAR aligner (v2.5.2) ([270]96).
   Gene alignment counts were determined using RSEM (v1.3.3) ([271]97).
   Isoform-level quantifications obtained from RSEM were integrated into
   DESeq2 ([272]98) for differential expression analysis, facilitated by
   the tximport package ([273]99), with settings adjusted to “rsem” and
   both “txIn” and “txOut” flags set to “TRUE.” The DESeq2’s results
   function was used to identify differentially expressed genes, which
   were then subjected to multiple testing correction using the
   Benjamini-Hochberg method to calculate adjusted P values. Visualization
   of gene counts for DEGs was achieved using the pheatmap (v1.0.12) and
   EnhancedVolcano (version 1.2.0) packages within R (v3.6.1). Last, KEGG
   and Elsevier Pathway Collection pathways enrichment as well as gene
   ontology analyses were conducted using EnrichR ([274]100), ShinyGO
   (v0.66) ([275]101), and Fast Gene Set Enrichment Analysis (doi:
   [276]https://doi.org/10.1101/060012).

Single-cell RNA sequencing

   Muscle cells were isolated from both intact and cardiotoxin-injured TA
   muscles on both sides of the hindlimb at ZT4 and ZT16 on days 0
   (intact), 1, and 3 post-injuries. The initial step involved the
   elimination of red blood cells using RBC Lysis Buffer, followed by
   resuspension of the muscle cells in 1 ml of Pre-Sort buffer. Cell
   viability was assessed with propidium iodide. To ensure accurate cell
   sorting, doublet exclusion was strictly conducted by comparing the
   height to the area for both forward scatter and side scatter signals.
   Approximately 100,000 propidium iodide–negative single muscle cells
   were then sorted from each sample using a BD FACSAria II cell sorter.
   Subsequent sample processing was completed at the NUSeq Core Facility
   at Northwestern University. Briefly, the cell number and viability were
   further evaluated using the Nexcelom Cellometer Auto2000 using the
   acrodine orange/propidium iodide (AO/PI) fluorescent staining method.
   For single-cell RNA sequencing, 10,000 cells from each condition were
   loaded into a Chromium Controller (10X Genomics) using a Chromium Next
   GEM Chip G (10X Genomics). These were processed to form single-cell gel
   beads in the emulsion (GEM) following the manufacturer’s instructions.
   cDNA synthesis and library preparation were performed using the
   Chromium Next GEM Single Cell 3′ Reagent Kits v3.1 and Single Index Kit
   T Set A (10X Genomics) according to the provided manual. Quality
   control for constructed library was performed by Agilent Bioanalyzer
   High Sensitivity DNA kit (Agilent Technologies) and Qubit DNA HS assay
   kit for qualitative and quantitative analysis, respectively. The
   multiplexed libraries were pooled and sequenced using NovaSeq 6000 S1
   Reagent Kit v1.5 (100 cycles) on Illumina NovaSeq 6000 sequencer with
   2 × 50 paired-end kits using the following read length: 28 bp Read1 for
   cell barcode and unique molecular identifier, 8 bp I7 index for sample
   index, and 91 bp Read2 for transcript.

Single-cell RNA sequencing analysis

   For demultiplexing and read counting, Cell Ranger (v1.0.0) was used.
   The Seurat package (v4.0.0) ([277]102) in R facilitated data
   preprocessing and visualization. Quality control measures and cell
   selection criteria were stringently applied; cells exhibiting less than
   200 or more than 6000 feature counts (for cells from intact and 1 dpi
   injured TA muscles) or exceeding 7500 feature counts (for cells from 3
   dpi injured TA muscles) were excluded. Genes detected in fewer than
   three cells were discarded. To account for low-quality and dying cells
   typically showing high mitochondrial content, cells were selected on
   the basis of a threshold where only those with less than 5% of reads
   mapping to the mitochondrial were retained. In addition, MALAT1, a
   highly expressed large noncoding nuclear gene, was excluded from
   downstream analyses.

   To identify potential doublets, a two-tiered approach was adopted as
   per Germain et al. ([278]103): Initially, scDblFinder (v1.4.0)
   predicted doublets using default settings, followed by manual
   identification based on the expression of marker genes from different
   cell types. After refining the dataset to remove low-quality cells,
   normalization of single cells from all six samples was carried out
   using NormalizeData function, and samples were integrated using a
   canonical correlation analysis based anchor method. The dataset was
   further adjusted by regressing out the difference between G2-M and S
   phase scores, effectively preserving distinctions between noncycling
   and cycling cells while ignoring variations among proliferating cells.
   Clustering was performed using the first 50 principal components,
   derived from the top 2000 most variable features, in a K-nearest
   neighbor graph with a resolution at 0.5. Subsequently, UMAP provided
   nonlinear dimensional reduction and visualization of clusters.
   Differentially expressed features across clusters were identified using
   FindAllMarkers function. Cell identities within each cluster were
   ascribed on the basis of canonical marker genes for skeletal muscle
   cells, as described in previous studies ([279]14, [280]15), and markers
   from the CellMarker 2.0 database ([281]13). Differential gene
   expression analyses between clusters used the FindMarkers function,
   setting a threshold of 0.25 for log fold change and 0.05 for the
   adjusted P value. Cell lineage and pseudotime inference analysis within
   MuSC subclusters were conducted using the Slingshot package ([282]46),
   with downstream gene expression along trajectories analyzed via the
   tradeSeq package ([283]47). Intercellular interaction analysis and
   visualization among the two neutrophil and MuSC populations at various
   days post-injury were performed using the CellChat package ([284]20)
   with standard parameters. To examine TOD-dependent variations in
   communication between MuSC and neutrophils, the NicheNet package
   ([285]34) was used, positioning MuSC as the “signal sender” to explore
   the ligands affecting neutrophils.

Western blotting

   Myoblast whole-cell lysates were prepared using CelLytic mammalian
   tissue lysis reagent, with the addition of protease inhibitors, to
   ensure protein integrity. For the preparing total protein lysates from
   muscle tissue, each snap-frozen TA muscle specimen was placed in 500 μl
   of CelLytic mammalian tissue lysis reagent supplemented with protease
   inhibitors and homogenized using a TissueLyser II apparatus from Qiagen
   to achieve uniform cell disruption. Subsequent to homogenization,
   protein concentrations were determined using the DC Protein Assay Kit,
   enabling accurate loading for SDS–polyacrylamide gel electrophoresis.
   Proteins separated by electrophoresis were then transferred onto
   0.45-μm nitrocellulose membranes for immunoblotting. For the detection
   of specific proteins, the membranes were incubated overnight at 4°C
   with primary antibodies, including anti-HIF1α (1:1000 dilution),
   phospho–NF-κB p65 (Ser^536; 1:1000 dilution), NF-κB p65 (1:1000
   dilution), PAR/pADPr (1:1000 dilution), and β-actin (1:10,000
   dilution). Following primary antibody incubation, the membranes were
   washed and then incubated with appropriate horseradish
   peroxidase–conjugated secondary antibodies. The protein bands were
   visualized using an enhanced chemiluminescence detection system,
   allowing for the quantitative analysis of protein expression levels.

RNA extraction and real-time qPCR

   For the qPCR analysis, total RNA was isolated from myoblasts using
   TRIzol Reagent, adhering strictly to the guidelines provided by the
   manufacturer. The quality and quantity of the extracted total RNA were
   assessed using the NanoDrop 2000 Spectrophotometers. RNA samples
   displaying an OD260/280 ratio greater than 1.80 were deemed suitable
   and thus advanced to the first strand cDNA synthesis using Applied
   Biosystems High-Capacity cDNA Reverse Transcription Kit, following the
   instructions detailed in the kit’s protocol. The subsequent real-time
   qPCR was performed using iTaq Universal SYBR Green Supermix from
   Bio-Rad on a CFX Opus 384 Real-Time PCR System.

ELISA

   The concentration of CCL2 released by myoblasts into the culture medium
   was quantified using the mouse MCP-1 ProQuantum Immunoassay Kit
   following the instructions provided by the manufacturer.

Utilization of artificial intelligence tools

   Artificial intelligence (AI) tools were used within this manuscript as
   follows. ChatGPT-4o was used to check grammar and clarity of portions
   of the manuscript. The prompt given was: “Edit passage of manuscript
   for grammar and clarity.”

Acknowledgments