Multicellular senescence impairs skeletal muscle recovery following disuse in aging

Abstract

Aged skeletal muscle has a diminished capacity to recover after disuse. Although muscle regrowth requires coordinated interactions between immune and progenitor cells, the mechanisms of impaired remodeling in aged skeletal muscle remain poorly understood yet possibly involve the accumulation of senescent cells. We used a flow cytometry approach coupled with scRNAseq to determine the muscle senescent cell identity and transcriptional landscape during skeletal muscle recovery following disuse atrophy. Young and aged mice underwent 14 days of hindlimb unloading followed by reloading (7 or 14 days). At recovery, old mice showed smaller myofibers and abnormal muscle macrophage dynamics corresponding to greater collagen content. These outcomes coincided with elevated markers of muscle senescence (p21 and γH2AX) and increased SPiDER-β-Gal⁺ cells, which inversely correlated with muscle mass. Single-cell resolution of SPiDER⁺ cells unmasked several senescent interstitial muscle vascular and stromal populations. Senescent interstitial cell populations were enriched in aged muscle and displayed a senescence-associated secretory phenotype (SASP) across multiple stromal, vascular, and immune cell types. Senolytic treatment reduced overall senescent cell burden, attenuated macrophage accumulation, and restored muscle mass and function in aged mice following disuse. These findings identify a multicellular senescence environment within the muscle interstitial niche as a hallmark of impaired muscle recovery following disuse.

Senescent interstitial cells accumulate in aged muscle and are related to impaired cellular remodeling.

INTRODUCTION

Aging is accompanied by a progressive decline in the capacity of skeletal muscle to recover after a period of disuse (1–3). This impaired recovery accelerates the age-associated decline in muscle mass, strength, and function, thereby placing older adults at risk for frailty, loss of independence, and increased mortality (4, 5). Unfortunately, little progress has been made toward establishing effective strategies to improve muscle regrowth in older adults, partly because the mechanisms underlying this process are poorly understood.

Recovery of skeletal muscle after disuse relies on a coordinated interplay between myogenic and nonmyogenic interstitial muscle cells, as well as immune cells (1, 6, 7). Macrophages are key regulators of muscle recovery that clear debris, activate satellite cells, and promote tissue remodeling during their transition from a pro-inflammatory to an anti-inflammatory phenotype (8–10). In aged muscle, macrophages exhibit a skewed anti-inflammatory profile at rest. In contrast, during regrowth from disuse atrophy, aged muscle is characterized by a dysregulated pro-inflammatory state, which is tied to poor muscle remodeling (11–13). These data suggest that macrophage function and polarization are disrupted in aged muscle and during regrowth.

Cellular senescence is a process by which cells undergo permanent cell cycle arrest in response to various stressors, functioning as a physiological mechanism to prevent the proliferation of damaged cells (14). These senescent cells enter a state of permanent growth arrest and acquire a senescence-associated secretory phenotype (SASP), which is composed of cytokines, chemokines, growth factors, and extracellular matrix remodeling proteins (15). Transient senescent cells can support tissue repair by releasing factors that attract immune cells and promote regeneration (16, 17). However, with aging, immunosenescence impairs immune surveillance and reduces the ability to clear senescent cells, resulting in their accumulation (18, 19). This persistent accumulation contributes to the promotion of chronic inflammation, fibrosis, and tissue dysfunction. Prior work has shown that muscle interstitial cells—including macrophages, muscle stem cells/satellite cells, and fibro-adipogenic progenitors (FAPs)—undergo senescence with age and that the pharmacological reduction of senescent cells improves muscle function (20, 21). Furthermore, experimental studies have demonstrated that an increased senescence burden is associated with muscle atrophy, fibrosis, and reduced physical function (15). Therefore, key interstitial muscle cells, which are essential for cellular remodeling during regrowth, may enter a senescent state, contributing to poor muscle recovery outcomes in advanced age.

Despite evidence linking senescent cells to age-related tissue dysfunction, it remains unclear which senescent populations accumulate in skeletal muscle following disuse and whether their presence corresponds with impaired recovery. To address this gap, we examined functional, morphological, and muscle cellular outcomes during recovery from disuse in young and old mice. We coupled these outcomes with a flow cytometry methods and single-cell RNA sequencing to catalog senescent cells in skeletal muscle. We hypothesized that aged mice during regrowth following disuse atrophy would be characterized by heightened senescent macrophages, which would be related to poor muscle recovery.

RESULTS

Aged mice have impaired recovery of muscle size and function following disuse

Young and old male C57BL/6 mice underwent 14 days of hindlimb unloading (HU), followed by reloading (RL) for 7 or 14 days, with measurements obtained at ambulatory baseline (AMB), HU, RL7, and RL14 (Fig. 1A). Both age groups exhibited a loss of grip strength after unloading. However, only young mice regained baseline grip strength upon reloading, whereas old mice displayed a sustained deficit in grip strength (Fig. 1B). Consistent with these functional changes, gastrocnemius muscle mass normalized to body weight was reduced following HU in both groups, indicating hindlimb muscle atrophy. The recovery of muscle mass occurred only in young mice, while old mice failed to return to baseline levels during reloading (Fig. 1C).

Fig. 1. Disuse causes loss of skeletal muscle mass and function, with impaired recovery during reloading in old mice.

(A) Experimental design schematic. (B) Whole-body grip strength of young and old mice at AMB, HU, and during reloading (RL4, RL7, and RL14). (C) Gastrocnemius muscle mass normalized to body weight of young and old mice at AMB, HU, RL7, and RL14. All data are presented as means ± SD. Statistical analyses were performed using two-way analysis of variance (ANOVA), followed by Tukey’s multiple comparisons test. ***P ≤ 0.001 (*) indicate group differences at the same time point. #P < 0.05, ##P < 0.01, and ###P ≤ 0.001 (#) indicate within group differences from their baseline time point (n = 5 to 10 per group) for AMB, HU, RL4, RL7, and RL14. Created in BioRender, P. Bourrant (2026; https://BioRender.com/wkn45qd).

We then examined gastrocnemius fiber cross-sectional area (CSA) using histochemistry (Representative image, Fig. 2A). Mean fiber CSA decreased in old mice after HU, with significant reductions at both RL7 and RL14 compared to the AMB. In contrast, young mice maintained fiber size across all time points (Fig. 2B). The analysis of fiber size distribution revealed a significant age-by-time interaction, indicating age-dependent differences between groups. At baseline, young mice had a greater proportion of larger fibers, while old mice displayed a shift toward smaller fibers (Fig. 2C). These differences became more pronounced during reloading after disuse, with young mice showing an increase in mid-to-large fibers, whereas old mice displayed an increase in small fibers at both RL7 and RL14 (Fig. 2, D and E). Overall, these data suggest that advanced age impairs the recovery of myofibrillar size and function after disuse.

Fig. 2. Old mice experience impaired recovery of myofiber size following disuse and reloading.

(A) Representative histochemical image of fiber CSA, wheat germ agglutinin (WGA; magenta), and scale bar of 20 μm. (B) Gastrocnemius mean fiber area. Gastrocnemius fiber size distribution at AMB (C), RL7 (D), and RL14 (E). All data are presented as means ± SD. Statistical analyses were performed using two-way ANOVA, followed by Tukey’s multiple comparisons test. *P < 0.05, **P < 0.01, and ***P ≤ 0.001 (*) indicate group differences at the same time point. #P < 0.05, ##P < 0.01 (#) indicate within group differences from their baseline time point. Significant main effects from the two-way ANOVA are shown in the panels above (n = 6 per group) for AMB, RL7, and RL14.

Old mice accumulate macrophages and exhibit higher collagen content in the gastrocnemius following disuse

We next evaluated cell type–specific remodeling processes in skeletal muscle using immunofluorescence to investigate cellular dynamics during recovery in old mice. The proportion of CD68⁺/DAPI⁺ (4′,6-diamidino-2-phenylindole–positive) macrophages increased in old mice during reloading, with a significant increase at RL14 compared to AMB (Fig. 3, A and B). Similarly, M2-like macrophages identified by CD68⁺/CD163⁺/DAPI⁺ colocalization were more common in old mice than in young mice across reloading time points (Fig. 3, C and D). Flow cytometry analysis supported an age-related increase in monocytes, although total macrophage abundance by fluorescence-activated cell sorting (FACS) did not fully mirror the immunofluorescence findings (fig. S1, A to C). The frequency of centrally nucleated fibers (CNFs) was elevated in old mice, indicative of a persistent degeneration-regeneration state and consistent with prior reports (Fig. 3, E and F) (22, 23). Satellite cells, identified by colocalization of Pax7⁺/DAPI⁺ staining, were significantly more abundant in young mice at all time points, a pattern supported by FACS (Fig. 3, G and H, and fig. S1D). In contrast, the quantification of FAPs by FACS showed no difference across age or time (fig. S1E). Furthermore, muscle collagen content, assessed by Sirius Red staining, was significantly higher in old mice compared to young mice across all time points, with the greatest accumulation observed at RL14 (Fig. 3, I and J). The collagen content was correlated with both total CD68⁺/DAPI⁺ macrophages and M2-like CD68⁺/CD163⁺/DAPI⁺ macrophages at RL14, indicating a link between immune cell abundance and increased fibrotic burden in aged muscle (Fig. 3, K and L).

Fig. 3. Old mice display impaired skeletal muscle remodeling following disuse and reloading.

(A) Representative immunofluorescent staining for macrophage CD68⁺ (red), DAPI⁺ (blue), and WGA (green). (B) Gastrocnemius macrophage content (colocalization of CD68⁺/DAPI⁺) normalized to tissue area. (C) Representative immunofluorescent staining for M2-like macrophage CD68⁺ (red), CD163⁺ (magenta), DAPI⁺ (blue), and WGA (green). (D) Gastrocnemius M2-like macrophage content (colocalization of CD68⁺/CD163⁺/DAPI⁺) normalized to area. (E) Representative image of CNF for DAPI⁺ (blue) and WGA (green). (F) Gastrocnemius CNF content normalized to fiber number. (G) Representative image of satellite cells PAX7⁺ (green), DAPI⁺ (blue), and WGA (red). (H) Gastrocnemius satellite cell content (colocalization of DAPI⁺/PAX7⁺) normalized to fiber number. (I) Representative image of Sirius Red staining (collagen I and III). (J) Percent Sirius Red staining in the gastrocnemius. Correlation between CD68⁺DAPI⁺ cells/mm² (K) and CD68⁺CD163⁺DAPI⁺ cells/mm² (L) compared to Sirius Red (% area) in the gastrocnemius at RL14. All data are presented as means ± SD. Statistical analyses were performed using two-way ANOVA, followed by Tukey’s multiple comparisons test. *P < 0.05, **P < 0.01, and ***P ≤ 0.001 (*) indicate group differences at the same time point. ##P < 0.01, and ###P ≤ 0.001 (#) indicate within group differenced from their baseline time point. Significance in (K) and (L) was assessed using Pearson’s correlation (n = 5 to 8 per group) for AMB, RL7, and RL14. Cell-specific identification is indicated by a white arrow. Scale bars, 20 μm.

Old mice exhibit elevated markers of senescence (p21 and γH2AX) in the skeletal muscle following reloading

We next assessed markers of cellular senescence in the skeletal muscle during recovery. The number of p21⁺/DAPI⁺ nuclei was higher in old mice compared to young mice at all time points, with a notable increase observed after 14 days of reloading (RL14) (Fig. 4, A and B). The frequency of p21⁺ nuclei was inversely related to gastrocnemius muscle mass at RL14, indicating that increased senescence burden corresponded to impaired recovery (Fig. 4C). Similarly, γH2AX⁺/DAPI⁺ myonuclei, a marker of DNA damage and cellular senescence, were significantly increased in old mice across all time points, with a marked accumulation observed at RL14 (Fig. 4, D and E). γH2AX⁺ myonuclei also showed a negative correlation with gastrocnemius mass at RL14 (Fig. 4F). Overall, these results demonstrate that aged age in mice is associated with an increased burden of senescent cells in the skeletal muscle during recovery, which correlates with impaired muscle recovery.

Fig. 4. Old mice exhibit heightened accumulation of senescent cells in skeletal muscle after RL14.

(A) Representative image of p21⁺ (red), DAPI⁺ (blue), and WGA (green). (B) Gastrocnemius p21 content (colocalization of DAPI⁺/p21⁺) normalized to fiber number. (C) Correlation between p21⁺/fiber and gastrocnemius mass at RL14. (D) Representative image of γH2AX⁺ (red), DAPI⁺ (blue), and WGA (green). (E) Gastrocnemius γH2AX⁺ content (colocalization of DAPI⁺/γH2AX⁺ normalized to myonuclei. Correlation between % γH2AX⁺ myonuclei and gastrocnemius mass at RL14 (F). All data are presented as means ± SD. Statistical analyses were performed using two-way ANOVA followed by Tukey’s multiple comparisons test. *P < 0.05, and ***P ≤ 0.001 (*) indicate group differences at the same time point. #P < 0.05, ##P < 0.01, and ###P ≤ 0.001 (#) indicate within group differences from their baseline time point. Significance in (C) and (F) was assessed using Pearson’s correlation (n = 4 to 6 per group) for AMB, RL7, and RL14. Cell-specific identification is indicated by a white arrow. Scale bars, 20 μm.

SPiDER-β-Gal⁺ senescent cells accumulate in the gastrocnemius muscle of old compared to young mice after reloading

A representative histogram illustrates SPiDER β-Gal staining and cell distribution at RL14 in fluorescence minus one (FMO), young, and old mouse gastrocnemius samples (Fig. 5A), with the complete gating strategy described in the Supplementary Materials. The gating strategy was validated using FMO and true-positive controls to ensure accurate identification of SPiDER-β-Gal⁺ populations (fig. S2A). In combination with SPiDER-β-Gal labeling, we also stained for specific cell populations (macrophages, FAPs, and satellite cells) known to undergo cellular senescence in aging skeletal muscle (21). The quantification of % SPiDER-β-Gal⁺ cells confirmed a significantly higher proportion of senescent cells in old mice across all time points compared to young mice, with abnormal and most pronounced accumulation at RL14 in the old mouse group (Fig. 5B). Consistent with p21⁺ and γH2AX⁺ cells, the percentage of SPiDER-β-Gal⁺ cells was inversely correlated with gastrocnemius mass at RL14 (Fig. 5C). The cell type–specific analysis of SPiDER-β-Gal⁺ cells further revealed that old mice had a greater proportion of senescent macrophages, satellite cells, and FAPs compared to young mice at RL14, reflecting the presence of multiple senescent interstitial populations in aged mouse muscle identified through this profiling approach (Fig. 5, D to F). Notably, senescent macrophages represented the largest of these fractions of SPiDER-β-Gal⁺ cells among these labeled cell types, suggesting that immune cells constitute a substantial part of the overall senescent cell burden in aged skeletal muscle.

Fig. 5. Old mice show greater senescent cell accumulation across different cell types after RL14.

(A) Representative histogram of gating strategy for FMO SPiDER-β-Gal and SPiDER-β-Gal–stained gastrocnemius samples at RL14 in young and old mice. (B) Quantification of percent SPiDER-β-Gal⁺ cells. (C) Correlation between SPiDER-β-Gal⁺cells (%) and gastrocnemius mass at RL14. Quantification of % SPiDER-β-Gal⁺ cells within total, macrophages population (D), satellite cells (E), and FAPs (F). The violet bar represents the percentage of SPiDER-β-Gal⁺ cells within each population. All data are presented as means ± SD. Statistical analyses were performed using two-way ANOVA, followed by Tukey’s multiple comparisons test. *P < 0.05, **P < 0.01, and ***P ≤ 0.001 (*) indicate group differences at the same time point. #P < 0.05, ##P < 0.01, and ###P ≤ 0.001, (#) indicate within group differences from their baseline time point. Significance in (C) was assessed using Pearson’s correlation (n = 5 to 6 per group) for AMB, RL7, and RL14.

Senescent macrophages accumulate and adopt an M2-like transcriptional program in aged muscle after disuse

To further investigate the heterogeneity of senescent cell transcriptional signatures in aged mouse muscle, we performed single-cell RNA sequencing on the gastrocnemius muscle from old mice at RL14 (RL14) after disuse (fig. S3). Uniform Manifold Approximation and Projection (UMAP) clustering identified major skeletal muscle resident cell types, including macrophages, FAPs, endothelial subtypes, and satellite cells (Fig. 6A). Mapping the senescence status of SPiDER-β-gal⁺ and SPiDER-β-gal⁻ cells onto the UMAP revealed that senescent cells were found across multiple clusters, including macrophages (Fig. 6B). This unbiased approach unmasked other vascular and stromal interstitial cell populations (e.g., vascular smooth muscle, neuroglial, and endothelial vein) that exhibited comparable SPiDER-β-Gal⁺ proportions. We focused our analyses on macrophage, satellite cell, and FAP populations previously reported to be susceptible to senescence (21), and our flow cytometry data similarly revealed SPiDER-β-Gal⁺ cells across these populations. Macrophages exhibited one of the highest senescent burdens, followed by satellite cells and FAPs (Fig. 6C). This aligns with the number of differentially expressed genes in the SPiDER-β-gal⁺ cell fraction compared to SPiDER-β-gal⁻ cells. However, SPiDER-β-gal⁺ satellite cells had considerably fewer differentially regulated genes (table S1).

Fig. 6. Single-cell transcriptomic of SPiDER-β-Gal^+/− unravels a senescent macrophage-specific accumulation after disuse in skeletal muscle.

Gastrocnemius muscles were harvested from old mice after RL14 following disuse. SPiDER-β-galactosidase staining was used for FACS to isolate SPiDER-β-Gal⁺ and SPiDER-β-Gal⁻ populations for downstream single-cell RNA sequencing analysis (n = 20 samples, pooled). (A) UMAP plot showing the distribution of single-cell clusters annotated by cell type. (B) UMAP plot illustrating the localization of SPiDER-β-Gal⁺ (blue) versus SPiDER-β-Gal⁻ (gray) in gastrocnemius muscle of old mice after RL14 following disuse. (C) Proportion of senescent versus nonsenescent cells within each cell type. (D) Bar graph showing up-regulated Hallmark pathways from gene set enrichment analysis (GSEA) in senescent versus nonsenescent macrophages. (E) Volcano plot depicting differentially expressed genes in macrophages between senescent (SPiDER-β-gal⁺) and nonsenescent (SPiDER-β-gal⁻) populations. (F to H) Violin plots showing single-cell transcript abundance (log₁p-normalized expression) of canonical M2-like macrophage–associated genes (Cd163, Mrc1, and Retnla) across the two experimental groups (F), expression of SASP-related factors (Ccl2, Cxcl1, Igfbp4, and Timp2) (G), and expression of senescent-associated genes (Cdkn1a, Glb1, and Trp53) (H). NicheNet-derived circos plot illustrating predicted ligand-target interactions between sender cell populations (bottom) and macrophages as receiver cells (top). Link width and transparency correspond to the interaction weight of each ligand-target pair. Colors on the lower half denote the sender cell type: Magenta, ligands expressed in multiple cell types (common); blue, endothelial shear-stress cells; light green, FAPs; dark green, satellite cells; and red, macrophages. (I) Venn diagram showing overlapping and unique genes detected in FAPs, macrophages, and satellite cells from single-cell RNA sequencing data. (J) Overlapping regions represent genes shared between both cell populations, while nonoverlapping regions indicate cell type–specific transcripts.

Given prior evidence of macrophage dysfunction during muscle recovery and the pronounced senescent population signature, we therefore evaluated their transcriptional profile. Senescent macrophages exhibited enhanced inflammatory signaling, characterized by activation of tumor necrosis factor–α (TNF-α)/nuclear factor κB (NF-κB), complement, and inflammatory responses, which is consistent with a chronic SASP that can worsen tissue inflammation. Beyond inflammatory signaling, senescent macrophages were characterized by pathways that sustain the senescent state, including the mammalian target of rapamycin complex 1 (mTORC1) and phosphatidylinositol 3-kinase (PI3K)/AKT/mTOR pathways (nutrient sensing and SASP amplification), the unfolded protein response, and the p53 pathway (cell cycle arrest) (Fig. 6D). To complement the pathway analysis, we then investigated gene-level differences between senescent (SPiDER-β-gal⁺) and nonsenescent (SPiDER-β-gal⁻) macrophages. Given the comparable abundance of other senescent cell types made up of vascular and stromal populations (Fig. 6C and fig. S4, C to N), senescent macrophages exhibited heightened up-regulation of transcripts characteristic of an M2-like macrophage senescence-associated phenotype, including Mrc1/CD206, Cd163, and Retnla/Fizz (Fig. 6E). Additional senescent macrophages expressed genes linked to matrix remodeling and fibrotic SASP activity, such as Dab2 and Timp2, which reinforce transforming growth factor–β (TGF-β) signaling and extracellular matrix deposition. Cd63, Itm2b, and Gnaq indicated increased lysosomal and vesicle trafficking consistent with SASP secretion. Elevated Pf4 and Igf1 further highlighted a profibrotic program characteristic of senescent macrophages (Fig. 6E).

To expand the pathway and differential expression analyses for macrophages (Fig. 6D and E), we visualized gene-level distributions using violin plots. Senescent (SPiDER-β-Gal⁺) macrophages exhibited higher expression of M2-associated markers (Cd163, Mrc1, and Retnla) compared to SPiDER β-Gal⁻ cells (Fig. 6F). SASP-related transcripts (Ccl2, Cxcl1, Igfbp4, and Timp2) were also elevated in SPiDER-β-Gal⁺ macrophages, indicating heterogeneous secretory output (Fig. 6G). Last, senescence markers (Cdkn1a, Glb1, and Trp53) were enriched in the SPiDER-β-Gal⁺ subset (Fig. 6H). Together, SPiDER-β-Gal⁺ cell proportions and SASP-associated transcriptional markers provide evidence that multiple interstitial vascular and stromal cell populations including macrophages exhibited a robust senescent cell phenotype.

We then used NicheNet to identify ligand-receptor interactions and their ligand-target regulatory potential, characterizing cell-to-cell communication with macrophages specified as the receiver. FAPs emerged as the primary sender, contributing to the maintenance of senescence and fibrotic macrophage phenotypes (e.g., Csf1, Junb, Hmox1, Tgfbi, and Ccr2). An autocrine signal within macrophages was also evident, with Tnf providing feedback to stress/senescence pathways (e.g., Junb, Cxcl2, Sqstm1, and Sdc4) (Fig. 6I). When FAPs or satellite cells were designated as receivers, the predicted ligand-receptor activity was markedly lower, indicating macrophages as the main communication hub within the senescent muscle niche (fig. S4, A and B). Consistent with cell-to-cell interactions, a Venn diagram comparing genes detected in FAPs, macrophages, and satellite cells revealed 168 shared genes between FAPs and macrophages, supporting their cellular cross-talk, whereas minimal interactions occurred in satellite cells (Fig. 6J and fig. S4B).

Acute senolytic treatment during reloading enhances skeletal muscle recovery and reduces senescent interstitial cell burden in aged muscle

Given the accumulation of senescent interstitial cells in aged skeletal muscle and the regulatory role these cells play in muscle regeneration and regrowth (21, 24), we next assessed whether an acute pharmacological clearance of senescent cells during reloading could improve recovery. Old mice received intermittent dasatinib plus quercetin (D + Q) or a vehicle control following HU (Fig. 7A). Grip strength improved in D + Q treated mice during reloading, with significant differences observed at RL7 and RL14 compared to controls (Fig. 7B), indicating the improved recovery of muscle function. Consistent with these functional changes, gastrocnemius muscle mass normalized to body weight was higher in D + Q treated mice at RL14 (Fig. 7C). The mean fiber CSA also increased following D + Q treatment (Fig. 7E). The fiber size distribution analysis showed a decrease in the proportion of small fibers and an increase in mid-to-large fibers in D + Q–treated mice relative to controls (Fig. 7F). In contrast, CNF frequency and satellite cell abundance were not different between the senolytic compared to the vehicle control groups at RL14 (Fig. 7, H and J). Markers associated with tissue remodeling and senescence were also reduced following senolytic treatment. Sirius Red staining supported reduced collagen deposition in D + Q–treated muscle (Fig. 7L), along with decreased senescence-associated β-galactosidase activity (Fig. 7R). Meanwhile, p21⁺ nuclei normalized to fiber count showed no significant difference between groups at RL14 (Fig. 7T). Immunofluorescence analyses supported reduced muscle CD68⁺ macrophage content and a marked decrease in CD68⁺CD163⁺ macrophages in D + Q–treated mice (Fig. 7, N and P). Transcriptional analysis confirmed these results, noted by decreased expression of senescence and SASP-related genes, including p53, p21, p16, Ifn, Igfbp4, and Il-6 in senolytic-treated mice compared to control (Fig. 7U). Overall, these findings suggest that an acute senolytic treatment lessened the senescent cell burden during reloading, thereby improving skeletal muscle recovery in aged mice.

Fig. 7. Senolytic treatment during reloading improves skeletal muscle recovery and reduces senescence burden in aged muscle.

(A) Experimental design schematic. (B) Whole-body grip strength of control (CON) and senolytic cocktail D + Q old mice at AMB, HU, RL4, RL7, and RL14 (n =9 per group). (C) Gastrocnemius muscle mass normalized to body weight at RL14. (D and E) Representative histological images and quantification of fiber CSA stained with WGA and (F) fiber size distribution. (G and H) Representative image of CNFs (DAPI and WGA) and quantification of CNF content normalized to fiber number. (I and J) Representative staining and quantification of satellite cells (PAX7⁺/DAPI⁺) normalized to fiber number. (K) Representative image of Sirius Red staining for collagen and (L) quantification of percent Sirius Red staining. Representative immunofluorescent staining and quantification of macrophages (CD68⁺/DAPI⁺) (M and N) and M2-like macrophages (CD68⁺/CD163⁺/DAPI⁺) (O and P) normalized to tissue area. (Q and R) Representative image and quantification of senescence-associated β-galactosidase (SA-β-Gal)–positive cells. (S and T) Representative image and quantification of p21⁺ cells (p21⁺/DAPI⁺) normalized to fiber number. For (C), (E), (F), (H), (J), (L), (N), (P), (R), (T), and (U), n =9 per group for CON and D + Q at RL14. RT-qPCR analysis of Trp⁵³ (p53), Cdkn1a (p21), Cdkn2a (p16INK4a), Ifn-γ, Cxcl10, Igfbp4, Il6, Il1β, Timp2, Ccl2, and Col3a1 in the gastrocnemius (U; n =7 per group). Data are presented as means ± SD. Statistical analyses were performed using two-way ANOVA with Tukey’s multiple comparisons test for grip strength (B) and two-tailed unpaired Welch’s t test for RL14 comparisons [(C), (E), (H), (J), (L), (N), (P), (R), (T), (U)]. *P < 0.05, **P < 0.01, and ***P ≤ 0.001. (*) indicate group differences at the same time point; # indicates differences from baseline within group. Cell-specific identification is indicated by arrows. Scale bars, 20 μm. Created with BioRender, P. Bourrant (2026; https://BioRender.com/689x5mj).

DISCUSSION

In the present study, we investigated the cellular mechanisms of how advanced age influences skeletal muscle regrowth following disuse atrophy. We found that aged mice during recovery exhibited muscle deficits, along with increased markers of senescence (p21 and γH2AX) and an abnormal accumulation of senescent cells, as measured by flow cytometry (SPiDER-β-Gal⁺). Notably, several interstitial muscle cell populations, including macrophages, exhibited comparable senescent burdens, highlighting a heterogeneous multicellular senescence landscape in aged muscle following reloading from disuse. Moreover, using single-cell analysis, we demonstrate that the transcriptional profile of SPiDER⁺ macrophages exhibited large skewing of SASP and senescent markers (compared to other senescent cell populations) and resembled an anti-inflammatory phenotype and may be reinforced by FAPs. Last, we demonstrate that the attenuation of senescence cells using a senolytic treatment lessened the senescent muscle phenotype, decreased the accumulation of macrophages, and improved muscle and functional recovery in aged mice, suggesting that the regulatory role senescence cells have on overall muscle recovery in aged mice. Overall, these findings reveal a heterogenous population of senescence cells in aging muscle during recovery and provide mechanistic insight into how cellular senescence contributes to impaired regrowth in aged skeletal muscle after disuse.

We found that old mouse skeletal muscle during recovery from disuse atrophy was marked by macrophage accumulation, increased collagen content, and a greater burden of senescence markers (p21 and γH2AX), indicating dysregulated tissue remodeling. These observations are consistent with prior studies demonstrating a disrupted macrophage inflammatory phenotype and function during regrowth in aged rodents (10, 11, 25, 26). Moreover, these results align with the accumulation of senescent cells during muscle regeneration or hypertrophy in aged mice, where senescent cells contribute to fibrosis and impaired muscle remodeling (21, 27, 28). Our work demonstrates that clearance of senescent cells improves muscle quality and functional outcomes, supporting a mechanistic role for senescent cells to contribute to age-related fibrosis and impaired muscle recovery dynamics. We further extend our observation of macrophage dysregulation and cellular senescence using a flow cytometry approach with a SPiDER-β-Gal probe. As a result, we validated that aged muscle is markedly burdened with senescent cells after reloading from disuse atrophy, with the highest accumulation at RL14. Single-cell SPiDER-β-Gal analyses indicate that cellular senescence during muscle recovery is not limited to a single lineage but instead represented a heterogeneous, multicellular process. Senescent cells were detected across immune, stromal, vascular, and myogenic populations, including FAPs, endothelial cells, satellite cells, and macrophages. Although the relative abundance of senescent cells was comparable across several of these populations, macrophages were prioritized due to the scope of this work, their known immunoregulatory role in tissue remodeling, and prior work demonstrating dysfunctional state in aged skeletal muscle (10, 11, 29). Dungan et al. (24) provided evidence that some senescent cells in regenerating aged muscle were indeed CD11b⁺ immune cells. Similarly, satellite cells in our study accounted for a relatively large proportion of the senescent cell population during recovery in old mice, consistent with (21) during muscle regeneration using a similar SPiDER-β-gal⁺ label. However, the transcriptional profile of senescent satellite cells was minimal, and interaction plots suggested little cellular cross-talk in this cell population when compared to macrophages and FAPs. Furthermore, the lack of changes in the satellite cell pool in both young and old mice during recovery questions their role, especially given that there was minimal muscle damage and regeneration in this model outside of normal aging. Moreover, the work from Jackson and colleagues (30) demonstrates that the young adult skeletal muscle recovers normally after HU despite satellite cell depletion. Together, our data suggest that the senescence phenotype in our model arises predominantly from nonmyogenic populations, particularly immune and stromal cells. Furthermore, the alignment of elevated senescence markers with abnormal macrophage accumulation and relationship to elevated collagen content further supports the idea that senescent macrophages may contribute to maladaptive remodeling in aged muscle during recovery from disuse atrophy.

We also demonstrate that senescent macrophages in aged muscle adopt an M2-like transcriptional program, characterized by an enrichment in SASP, matrix remodeling, and profibrotic signaling. Pathway analysis highlighted the activation of TNF-α/NF-κB, mTORC1, and PI3K/AKT/mTOR cascade pathways, which are known to sustain the senescent program and promote chronic inflammation. Consistent with recent mechanistic work, the depletion of CD206⁺ M2-like macrophages increased FAP activation and follistatin secretion, promoting muscle regeneration by reducing TGF-β signaling (31). These findings highlight the essential role of cellular interactions between macrophages and FAPs in shaping the muscle remodeling environment. Previous work by Zhang et al. (32) demonstrated in vitro that senescent FAPs and their SASP factors from aged muscle can recruit and influence macrophage polarization toward an M2-like phenotype. Likewise, Thomas and colleagues (33) demonstrate the critical intercellular interplay among FAPs, macrophages, and satellite cells during muscle overload, yet this interplay is dysregulated in aging. Our data extend these findings by suggesting that advanced age disrupts the macrophage-FAP axis through the emergence of senescent, M2-like macrophages. The ligand-receptor analysis further showed that FAPs and autocrine signaling from neighboring macrophages were the dominant cell type converging on macrophage receptors to reinforce SASP signaling.

Our study had some limitations. First, analyses were limited to the gastrocnemius muscle and might not fully reflect responses across different muscles or fiber types. Second, while our data support a mechanistic connection between senescent cell accumulation and muscle fibrosis during muscle recovery, we did not directly assess how cellular senescence affected other age-related mechanisms linked to muscle regrowth such as neuromuscular integrity, protein turnover, mitochondrial function, or anabolic signaling pathways (34–36). Last, we are also unclear of the regulatory contribution of specific senescent cell populations on muscle recovery with advanced age, including but not limited to the specific role of senescent macrophages. Gain- and loss-of-function experiments targeting specific myogenic and nonmyogenic cellular populations in the skeletal muscle during recovery are warranted.

In summary, our research adds a previously unexplored dimension to the aged skeletal muscle literature by identifying multiple cellular senescent nonmyogenic cell populations including anti-inflammatory macrophages as a key mechanism that regulates cellular remodeling after disuse atrophy in aged skeletal muscle (Fig. 8).

Fig. 8. Overview of the senescent interstitial cells and regulatory role on skeletal muscle recovery in aged mice.

Schematic illustrating that, after a period of HU and reloading, aged skeletal muscle shows impaired mass, strength, and myofiber size, accompanied by increased fibrosis and elevated senescence markers (p21 and γH2AX). Senescent interstitial cells accumulate across multiple stromal, vascular, and immune populations and exhibit a SASP, including factors such as Ccl2, Cxcl1, Timp2, and Igfbp4. Among these, macrophages display a senescent M2-like phenotype (CD206 and CD163) with elevated senescence-associated markers. Together, the accumulation of senescent interstitial cells is associated with impaired muscle recovery with advanced age. Created in BioRender, P. Bourrant (2026; https://BioRender.com/z1vnnyk).

MATERIALS AND METHODS

Experimental design

Young and old mice were evenly distributed by weight across the following experimental groups: AMB, 14 days of HU, RL7, or RL14 following HU. AMB mice remained freely ambulatory in standard cages (two to four per cage) without undergoing HU or reloading. HU was performed using a modified Morey-Holton HU design, with two mice housed per cage during the unloading period (11, 29, 37). Body weight, food, and water intake were monitored daily. For the reloading groups, the HU apparatus was removed, allowing mice to move freely in standard cages. At the end of the experiment, mice were fasted for 3 hours before euthanasia under isoflurane, followed by cervical dislocation. Hindlimb muscles were dissected, weighed, and processed for flow cytometry, snap-frozen in liquid nitrogen, or embedded in optimal cutting temperature (OCT) compound and isopentane for storage at −80°C.

Animals

All animal experiments were conducted in compliance with the Institutional Animal Care and Use Committee guidelines and received official approval (protocol 1476 to M.J.D.). Mice were housed at the University of Utah Animal Resources Center, an Association for Assessment and Accreditation of Laboratory Animal Care (AAALAC) accredited and approved by the National Institutes of Health. All animal procedures were approved by the University of Utah Institutional Animal Care and Use Committee. Experiments followed ARRIVE 2.0 guidelines. Male C57BL/6 young (~4 months) and old (~26 months) mice were obtained from the National Institutes of Health/National Institute on Aging. The mice were maintained on a 12-hour light/12-hour dark cycle with ad libitum access to a standard chow diet (Envigo, 2920X) and water. In mice, the progression of muscle aging tends to be more severe in males, providing a model that better reflects the human (male and female) condition (38). Although females were not examined in this study, it is plausible that overlapping mechanisms contribute to muscle aging across sexes and deserve future exploration (39). Animals were randomly assigned to experimental groups with baseline body weight matched across groups. Animals were monitored daily for health status and body weight. Mice showing signs of illness, rapid weight loss, impaired mobility unrelated to the protocol, or tumor burden were excluded.

Senolytic treatment

Following 14 days of HU, old (~26-month) male C57BL/6 mice were assigned to either a control (CON) or senolytic treatment group (D + Q). Mice received oral gavage of a senolytic cocktail containing dasatinib (5 mg/kg; Sigma-Aldrich, SML2589) and quercetin (50 mg/kg; Q4951, Sigma-Aldrich) dissolved in 10% polyethylene glycol molecular weight 400 (PEG-400; Sigma-Aldrich, 202398), while control mice received vehicle (10% PEG-400) alone, as previously described (36). Gavage was performed at the start of reloading (day 0) and on days 4, 7, and 10. Mice were euthanized on reload day 14 for tissue collection and analysis. This acute senolytic cocktail treatment was chosen on the basis of prior published senolytic studies demonstrating effective clearance of senescent cells with intermittent administration (26, 27).

Functional in vivo assays

A grip strength meter (Columbus Instruments, Columbus, OH, USA) was used to measure whole-body grip strength at AMB, HU, RL4, RL7, and RL14. Mice underwent a brief acclimation period on the grid, followed by three consecutive grip attempts. To ensure consistency, all four limbs were in contact and gripped to the grid before each trial, and mice were gently pulled by the base of their tail until the grip was released. Peak force was recorded following each grip test, with the average peak force used as the final measure of whole-body grip strength.

Immunofluorescence and histology

Frozen OCT-embedded gastrocnemius muscles were sectioned transversely at 10 μm using a cryostat (CM1860, Leica Biosystems, Wetzlar, DE). Muscle sections were processed for analysis of myofiber CSA, CNFs, satellite cell content (Pax7⁺), senescence and DNA damage markers (SA-β-Gal⁺, p21⁺, and γH2AX⁺), collagen content (Sirius red), and macrophages (CD68⁺ and CD163⁺) as previously described (25, 27, 29, 37, 40). Image acquisition and quantitative analyses were performed blinded to the experimental groups. Detailed methodologies are provided in Supplementary Methods (41, 42).

RNA isolation and RT-qPCR

RNA was isolated from gastrocnemius muscle (~30 mg) using QIAzol lysis reagent (QIAGEN, 79306), with chloroform phase separation and isopropanol precipitation according to the manufacturer’s instructions. The RNA was resuspended in nuclease-free water, and the concentration was measured using an Epoch spectrophotometer (Take3, BioTek, Winooski, VT). One microgram of RNA was reverse-transcribed using the iScript cDNA synthesis kit (Bio-Rad, 17088-91) on a Bio-Rad T100 Thermal Cycler (20-μl reaction; 25°C for 5 min, 46°C for 20 min, 95°C for 1 min, and 4°C hold).

The reverse transcription quantitative polymerase chain reaction (RT-qPCR) was performed using cDNA diluted 1:8 in nuclease-free water on a CFX Connect Real-Time PCR Detection System (Bio-Rad) with SsoAdvanced Universal SYBR Green Supermix (Bio-Rad, #17252-70). Gene expression was normalized to ribosomal protein L32, and relative fold changes were calculated using the ΔΔCt (2^−ΔΔCt) method. Target genes were selected to represent senescence and SASP markers as described previously (21). Primer sequences were purchased from the University of Utah Health Sciences Center Core. (Primer sequences are provided in table S2.)

FACS analysis

Gastrocnemius muscles were dissected, minced, and digested in Ham’s F12 medium containing Liberase and deoxyribonuclease I at 37°C for 40 min. Following filtration, centrifugation, and red blood cell lysis, the single-cell suspension was resuspended in FACS buffer for staining. Immune and progenitor cell populations were identified using CD31, CD45, CD11b, Sca-1, F4/80, and α7-integrin, with DAPI used to exclude dead cells.

Gating strategy defined macrophages (CD31⁻CD45⁺F4/80⁺CD11b⁺), monocytes (CD31⁻CD45⁺F4/80⁻CD11b⁺), satellite cells (CD31⁻CD45⁻Sca-1⁻α7-integrin⁺), and FAPs (CD31⁻CD45⁻Sca-1⁺α7-integrin⁻).

For senescent cell detection, muscles were processed as described above and incubated with the Senescence Detection Kit (Dojindo, SG04-01) following the manufacturer’s instructions. A detailed description of methodology and reagents can be found in Supplementary Methods.

Senescent cells (SPiDER-β-Gal⁺) and nonsenescent cells (SPiDER-β-Gal^-) single-cell RNA sequencing

Single-cell RNA sequencing was performed on senescent and nonsenescent cells isolated by FACS using the Cytek Aurora Spectral Cell Sorter. Gastrocnemius muscles from both hindlimbs (n = 10 mice per group) were collected at 14 days of recovery using a separate cohort of aged mice. Muscles were pooled, dissected, and digested as described above. Cell pellets were incubated with the Senescence Detection Kit (Dojindo, SG04-01). To improve sorting, 7-aminoactinomycin D (7-AAD) (1:200; BioLegend, 420404) and Hoechst 33342 (1:40,000; Tocris, 5117) were added to exclude debris and to enrich for nucleated cells, as recommended by 10x Genomics technical support. Cells were sorted into SPiDER-β-Gal⁺ (senescent) and SPiDER-β-Gal⁻ (nonsenescent) fractions at the University of Utah Flow Cytometry Core, and libraries were generated at the University of Utah Genomics Core. Single-cell RNA sequencing was carried out using the Chromium GEM-X Single Cell 3′ v4 Gene Expression Kit (10x Genomics). Approximately 2,000–4,000 cells were captured per dataset, with a viability of ~90%, and mean reads per cell ranged from 30,000 to 40,000. Detailed data processing can be found in Supplementary Methods (43–49)

Statistical analyses

Data are presented as means ± SD, with individual data points shown. To assess the main effects of age (young versus old) and disuse/recovery time (AMB, HU, RL7, and RL14), a two-way analysis of variance (ANOVA) was performed. When a significant interaction or main effect was observed, Tukey’s multiple comparisons test was applied to identify differences between groups at each time point. Statistical significance between age groups at the same time point is indicated by *P < 0.05, **P < 0.01, and ***P ≤ 0.001. In contrast, differences from AMB within the same age group are indicated by #P < 0.05, ##P < 0.01, and ###P ≤ 0.001. Correlations were evaluated using Pearson’s correlation. Comparisons between two independent groups (CON versus D + Q) were performed using a two-tailed unpaired Welch’s t test. The statistical significance was set at P < 0.05. Sample sizes were determined using a power analysis.

All in vivo outcomes represent biological replicates, where n represents individual mice. Histological quantifications were performed on full sections or multiple fields per muscle and averaged to generate a single value per animal. RT-qPCR reactions were run in technical duplicates, and the mean value was used for statistical analysis. All analyses and figure preparation were performed using GraphPad Prism (v10.6, La Jolla, CA, USA).

Supplementary Materials

This PDF file includes:

Supplementary Methods

Figs. S1 to S4

Tables S1 and S2