Quercetin Ameliorates Skeletal Muscle Fibrosis After Injury by Interfering with Macrophage-Myofibroblast Transition

Introduction

The incidence of acute skeletal muscle injury is rising, paralleling the increase in sports participation. The NCAA Injury Surveillance Program Database estimates that ~63 per 1000 NCAA athletes were exposed to musculoskeletal-related injuries.¹ Acute skeletal muscle injuries are often capable of recovery; however, in some cases, the severity of the injury makes complete restoration challenging, leading to decreased muscle function and an increased risk of re-injury.² Therefore, understanding the mechanisms of skeletal muscle injury and identifying effective solutions for better restoration are crucial.³

Excessive inflammation following muscle injury can divert the repair process towards a pathological outcome: fibrosis. This condition is characterized by the aberrant accumulation of extracellular matrix (ECM), which disrupts tissue architecture and function.^4,5 In skeletal muscle, fibrosis alters biomechanical properties, reducing stress tolerance and strength.⁶

Myofibroblasts are the principal ECM-synthesizing cells in connective tissue remodeling due to their dual role in promoting wound closure and facilitating fibrotic deposition.⁷

The origin of myofibroblasts is still controversial. In skeletal muscle, the sources of fibrosis include myoblasts, FAPs and perivascular cells.^8–10 Recent studies have also identified macrophage-myofibroblast transition (MMT) as a source in skeletal muscle. Based on our previous research, MMT exists in skeletal muscle following acute injury. MMT mainly occurred in M2 macrophages at five days post-injury. The complement system plays a key role in MMT, with C3ar1 being highly expressed during this transition.¹¹ The complement system is pivotal for both innate and adaptive immunity. Proteins of the complement system are mainly produced by the liver and widely distributed in blood and tissues, among which the most abundant complement protein is C3. ¹² C3 molecules are continuously broken down into two fragments, namely C3a and C3b. The former then binds with C3aR to initiate a series of biological procedures. Apart from its role in infection and inflammatory tissue damage, some literatures have indicated the role of C3a/C3aR in tissue and organ regeneration.^13,14 However, whether this axis is therapeutically causal in MMT and skeletal muscle fibrosis is still unclear.

Given the identification of C3a‑driven MMT as a pivotal and druggable event in post‑injury fibrosis, we reasoned that a compound capable of disrupting this transition could offer a novel therapeutic strategy. Quercetin is a flavonoid found in vegetables and fruits, having anti-inflammation properties.¹⁵ The effects of this herb against renal fibrosis,¹⁶ liver fibrosis,¹⁷ lung fibrosis,¹⁸ myocardial fibrosis¹⁹ and muscle atrophy have been demonstrated by research. Quercetin has been shown to attenuate the differentiation capacity of PDGFRα/CD201 double-positive muscle stem cells, specifically impairing their transition toward adipogenic and fibrogenic phenotypes.¹⁵ While quercetin attenuates fibrosis in human skeletal muscle, the ingredient target interactions are not distinct.

Therefore, in the current study, we first conducted weighted gene correlation network analysis (WGCNA) to find MMT-related genes using scRNA-seq. Then, network pharmacology and experimental validation were performed to confirm the anti-fibrosis mechanism of quercetin in skeletal muscle injury.

Methods

Single‑Cell Sequencing Analysis

All primary datasets analyzed in this work were sourced from the GEO repository. (https://www.ncbi.nlm.nih.gov/geo/). GSE143437 and GSE159500 included 12 samples of hindlimb muscle after acute injury, ie., 7 days (D7, n=3), 5 days (D5, n=3), and 2 days post-injury (D2, n=3), as well as non-injured control (D0, n=3).

We processed all raw data in R (version 4.4.0), performing initial conversion to Seurat objects with the Seurat package (v5.1.0). To ensure data quality, cells with fewer than 300 detected genes or with >20% mitochondrial gene content were filtered out. The two datasets were integrated using the IntegrateData function in Seurat (with standard parameters) to correct for potential batch effects. Subsequent steps included normalization to adjust for technical variations, dimensionality reduction, and clustering/visualization using UMAP. Clustering analysis identified cell types and subpopulations, differential expression analysis detected significantly changed genes, and functional enrichment analysis explored relevant biological functions and pathways. Functional enrichment analyses were conducted using the clusterProfiler R package.

Pseudo-Bulk Differentially Expressed Gene Analysis

Differentially expressed genes (DEGs) identification for macrophage populations (D5 vs. D0) was performed with Seurat FindMarkers function, retaining all default algorithmic parameters (log2fc.threshold > 0.25, adjusted p-value < 0.05). Comparative studies established pseudobulk methods as optimal for differential expression analysis across conditions in scRNA-seq, where cellular counts are aggregated per biological sample within defined cell states.²⁰ In this study, we also implemented a pseudobulk strategy using AggregateExpression (Seurat v5.1.0) and identified the differentially expressed genes (DEGs) in skeletal muscle on D5 via DESeq2.

WGCNA

We implemented WGCNA using the R package hdWGCNA to construct gene co-expression networks in macrophages and identify modules strongly correlated with C3a expression. Standard metrics including module eigengenes, gene significance (GS), module membership (kME), and module significance were computed.

Network Pharmacology Prediction

To decipher the potential mechanism by which quercetin might interfere with C3a-driven MMT, we employed a network pharmacology approach. This systems biology strategy was utilized to bridge our bioinformatics findings with the known pharmacologic profile of quercetin. The aim was to generate a testable hypothesis by identifying key molecular targets at the intersection of the compound’s action and the disease process.

Quercetin-containing herbal preparations were systematically identified through keyword searches (“quercetin”) in the TCMSP (https://tcmsp-e.com) and cross-referenced with the Encyclopedia of Traditional Chinese Medicine (ETCM). Protein targets of quercetin were retrieved from TCMSP. All targets were consolidated, deduplicated, and standardized using UniProt (https://www.uniprot.org) for nomenclature harmonization.

DEGs identified from single‑cell RNA sequencing data between D5 and D0 cells, together with the genes from brown module of scRNA seq, were utilized as targets related to skeletal muscle fibrosis. Then a custom Venn diagram was generated by the Evenn web tool (ehbio.com) to demonstrate common key genes. PPI networks were reconstructed using STRING (v12.0) under stringent criteria (confidence≥0.7, experimental evidence only). Topological features were extracted via Cytoscape (v3.7.1) with node importance ranked by betweenness centrality.

The intersectional genes identified through multi-omics analysis represent core therapeutic targets through which quercetin exerts its anti-fibrotic effects in skeletal muscle.

Molecular Docking

The 3D structures of the target protein (Ctsd, Stat1, Il1b, Cxcl10) were retrieved from the protein database (PDB) (http://www.rcsb.org/). The chemical structure of quercetin was acquired from PubChem (CID: 5280343) as the primary source, with supplementary verification using the TCMSP Database. AutoDock 4.2.6 software was used to identify the binding site between Ctsd, Stat1, Il1b, Cxcl10 and quercetin. After pretreatment of ligands and receptors, molecular docking was accomplished by configuring operating methods and docking parameters. Visualization was realized through software PyMol2.5.

Cell Culture and Stimulation

Bone marrow-derived macrophages (BMDMs) were extracted based on our previous protocol.¹¹ Primary bone marrow-derived cells were maintained in DMEM/F12 complete medium (10% FBS + 1% penicillin/streptomycin) with 50 ng/mL M-CSF (Gibco, PMC2044) for 1 week to induce macrophage differentiation. Cell purity was confirmed by F4/80 immunostaining (>90% positive cells).

Following 72-hour stimulation with 10 ng/mL mouse C3a (Novoprotein, CM99), cells were assessed for fibrotic marker α-SMA via Western blot prior to collection. The influence of quercetin on C3a stimulated BMDMs was examined by treating the cells with a range of quercetin concentrations (0 µM, 10 µM, 20 µM, 40µM,60µM, 80µM, 100 µM and 120µM) for 24, 48, and 72 hours. The vehicle control group received 0.1% DMSO, which corresponded to the highest solvent concentration present in quercetin-treated groups.

Through systematic evaluation using CCK-8 viability assays, we identified 40 μM quercetin treatment for 48 hours as our low-dose intervention. The high-dose condition was established at 100 μM quercetin for 48 hours. These optimized treatment parameters were subsequently employed in all mechanistic investigations of quercetin’s anti-fibrotic effects.

CCK-8 Assay

Bone marrow-derived macrophages (BMDMs) were seeded in 96-well plates at 5×10³ cells/well and exposed to quercetin (0–120 μM) for 24, 48, or 72 h. Cell viability was assessed using the CCK-8 assay (MedChemExpress, HY-K0301). 10μL reagent was added per well and incubated for 120 min. Absorbance at 450 nm was quantified using a Varioskan LUX microplate reader, with viability normalized to untreated controls.

Animal Feeding and Model Establishment

This study was conducted under the ethical approval from Shanghai Pudong Hospital (IACUC-20231011-01) in full compliance with international standards (NIH Guide) for humane animal research. Thirty 10-week-old male C57BL/6J mice underwent standardized gastrocnemius contusion injury under [specify anesthesia, eg., 2% isoflurane] anesthesia following established methods.¹¹ After the injury, the animals were allowed to move freely. Quercetin was intramuscularly injected at the fifth day post-injury into twenty-one mice. Quercetin (50µL, 20 mg/kg) was dissolved in 5% DMSO and injected to the injury site. Control group received an equal volume of PBS. At 7, 14 and 28 days post-injury (labeled as groups D7, D14 and D28, respectively), six randomly selected mice from each group were euthanized to harvest the injured muscle. Additionally, three mice from the same batch served as un-injured controls (group D0) and were euthanized for bilateral gastrocnemius collection at the same time as the D28 group.

Histological Measurements

Intact, injured and quercetin treatment gastrocnemius samples were dissected, embedded in OCT, and frozen using precooled liquid nitrogen. The samples were cut into 8-μm sections and processed for routine HE staining and Masson staining following established protocols.²¹

Immunofluorescent Staining

Immunofluorescence staining of cells was carried out as previously described.²¹ The primary antibodies included anti-CD206 (18,704-1-AP, Proteintech), anti-α-SMA (ab7817, Abcam), anti-CD68 (ab125212, Abcam), anti-Col 1 (sc-293182, Santa Cruz) and anti-Dystrophin (12715-1-AP, Proteintech). Nuclei were stained with DAPI. Images were captured using a fluorescence microscope (ECHO Revolve, America) and quantified using ImageJ software.

For comparisons in vivo, the relative fluorescence intensity of CD206, α-SMA, and Col 1 was quantified with D0 as reference, while the co-expression of CD68 with α-SMA was quantified using D5 as the reference.

Fluorescence images were acquired using a fluorescence microscopy and analyzed with ImageJ software. Expression levels were quantified from six randomly selected high-power fields, normalized to DAPI-positive cell counts, and expressed relative to the NC group.

Western Blot Analysis

Protein was extracted using RIPA. Following 30-minute ice incubation with periodic mixing, samples were centrifuged (12,000 × g, 15 min, 4°C) to pellet cellular debris. The resulting supernatants were heat-denatured (95°C, 5 min) prior to immunoblotting. Primary antibodies against α-SMA (Abcam ab7817), Col 1 (Abcam ab270993), and GAPDH (Proteintech 60004-1-Ig) were employed Intensity was detected using ChemiDoc XRS+ system (Bio-Rad), with band quantification in ImageJ (v1.53) after background correction and GAPDH normalization (n=3).

Behavioral Evaluation

Gait performance was evaluated at 28-days post-injury using CatWalk XT 10.6 (Noldus, Netherlands). Three compliant runs were collected per animal, with primary endpoints including: (a) print length (stance phase contact pattern); (b) swing speed (inter-limb coordination).

Statistical Analysis

Quantitative data are reported as mean ± SD. Intergroup differences were assessed using Student’s t-test (two groups) or one-way ANOVA with Tukey’s correction (multiple groups) in GraphPad Prism 9.5.1, with p-values below 0.05 deemed statistically significant. For all enrichment analyses (GO, KEGG) and identification of hub genes, p-values were adjusted for multiple comparisons using the Benjamini-Hochberg false discovery rate (FDR) method.

Results

scRNAseq Reveals Unique Cell States After Contusion Injury

Through quality control, a total of 43,791 cells were screened out, and 8 cell types were identified based on cell-specific markers and significantly enriched genes (Figure S1A). The 8 populations included endothelial cells9 (Pecam1, Cdh5), fibroblasts/FAPs (Pdgfra, Col1a1), glial cells (Kcna1, Plp1), macrophages (Cd68, Lyz2), muscle stem cells (Pax7, Myf5), neutrophils (S100a8, S100a9), stromal cells (Pgam2, Myoz1), and tenocytes (Tnmd, Scx) (Figure S1B). From this UMAP plot split by time group, it could be seen that only a few macrophages were present in healthy skeletal muscle (D0), but a significant increase in macrophages could be observed one day after injury (D2), reaching its peak on the D5. Conversely, muscle stem cells (MuSCs) were a trace population in D0 or D2, and they expanded explosively at D5, indicating an extremely active status in the repair process. Post injury, similar to endothelial cells, fibroblast population experienced a dramatic decline in number, and then expanded gradually from D2 to D7 (Figure S1A). Given the pivotal role of macrophages in orchestrating inflammatory and reparative responses, we performed an in-depth sub-clustering analysis of this population.

To further explore macrophage heterogeneity, we performed sub-clustering analysis on the macrophage population, followed by dimensionality reduction, obtaining 9 subclusters (Figure 1A). Genes related to fibrosis were highly expressed in macrophages (Figure S2). The UMAP plot and relative proportions demonstrated the distribution of these populations in different samples. Cluster 1 had the largest quantity at D2 (Figure 1B). Cluster 1 macrophages, hereinafter referred to as pro-Inflammatory macrophages (M1), were marked by Ccl2, Ccl7, Ccl6, and S100a4. The upregulation of chemokines and S100a4 assists in recruiting immune cells and promoting inflammation, which in turn facilitated the activation and differentiation of satellite cells. At D5, cluster 0 had reached their peaks, mainly consisting of anti-inflammatory macrophages (M2). In cluster 0, C1qa, C1qc, C1qb, and Apoe were upregulated, which have been proven to be M2 markers. The temporal dominance of this M2 at D5, characterized by high expression of C1q and Apoe—factors involved in immunomodulation and ECM organization—suggests its potential role in facilitating tissue remodeling and possibly contributing to the fibrotic niche during the peak of MMT. By regulating the functions of macrophages, these genes may influence the structure and function of the extracellular matrix, facilitating the repair of damaged tissue.

Figure 1 scRNA seq analysis of macrophages and major interactions between key cells in injured skeletal muscle. (A) UMAP embedding of macrophages data colored by meta-clusters to simplify visualization and split by time point to visualization. (B) Relative proportion of subclusters at each time point in macrophages. (C) netVisual_circle to demonstrate number of interactions in any two cell types. (D) Dot plot of all ligand receptor results in the specified receptor ligand cell type. The red font indicates the cell interaction on D2, while the blue font represents the cell interaction on D5.

To gain a better understanding of the ligand-receptor interactions among macrophages, fibroblasts, and MuSCs, the data of D2, D5, and D7 were employed for cell–cell communication analysis. There was a vigorous cross-talk between macrophages and interstitial cells, particularly fibroblasts. Additionally, a comparison was conducted between D2 and D5, and the results showed that the reactions of Col1a1/Col1a2 - (Itga11 + Itgb1), Col1a1/Col1a2 - Sdc1, and Pdgfa - Pdgfra/Pdgfrb were upregulated (Figure 1C and D). This enhanced macrophage-fibroblast communication via PDGF and collagen-integrin pathways at D5 provides a mechanistic basis for how macrophages directly stimulated fibroblast activation, proliferation, and collagen production, thereby driving the fibrotic progression observed at this critical phase.

Cells were classified as undergoing MMT if they co‑expressed the macrophage marker Cd68 and the myofibroblast marker Acta2 (encoding α‑SMA). We categorized cells into four types based on a-SMA and Cd68, namely MMT (a-SMA+, Cd68+), IAF (a-SMA+, Cd68-), IAM (a-SMA-, Cd68+), and others (a-SMA-, Cd68-). Quantity calculation and UMAP visualization indicated that the quantity of MMT was the largest on D5 (Figure S1C).

Skeletal Muscle Highly Expresses Genes Related to Immunity on D5

To explore the gene expression and corresponding functional analysis of skeletal muscle cells on D5, we conducted differential gene expression analysis. Volcano plots show genes that were differentially expressed in macrophages (Figure 2A). GSEA and GO revealed significant activation of muscle system process and tissue remodeling pathways in D5 macrophages (Figure 2B–E). And the lysosome pathway was significantly upregulated in D5 macrophages (Figure 2C and F).

Figure 2 Identification of high-confidence differentially expressed genes in macrophages (Single-cell-level analysis). (A) Volcano plots showing genes that are differentially expressed in macrophages. Green dots indicate genes upregulated in D5 group; yellow dots indicate genes upregulated in D0 group. (B) Dot plot displays the top 8 enriched gene sets (GO) ranked by normalized enrichment score (NES). (C) Dot plot displays the top 8 enriched gene sets (KEGG) ranked by normalized enrichment score (NES). (D) The muscle system process, tissue remodeling (E) and lysosome signaling pathway (F) enrichment plots for the differentially expressed gene set in D5 vs. D0 macrophages. The red curve represents the enrichment score (ES), and the vertical bars indicate the positions of DEGs in the ranked gene list. FDR < 0.25.

The correlation analysis of the D0 and D5 groups was presented in the figure (Figure 3A). The immune-related cell groups and the non-immune-related cell groups each had an overall correlation. After filtering according to the conditions of p - value < 0.05 and | log2FoldChange | > 0.25. Among 1,638 DEGs identified by single-cell analysis and 1,301 DEGs from pseudobulk analysis, 1,022 genes overlapped and were classified as high-confidence DEGs (Figure 3B).

Figure 3 Pseudo-bulk differentially expressed gene analysis on D5. (A) The correlation analysis of the D0 and D5 groups. (B) Venn diagram shows the overlap between DEGs identified by single-cell and pseudobulk approaches. The intersecting genes (n=1022) were considered high-confidence DEGs. (C) The result of GO and KEGG (D) enrichment analysis of genes upregulated in D5 group.

The GO results based on all the high-confidence DEGs of the D5 groups, encompassing three parts: BP, MF and CC, were demonstrated in the figure (Figure 3C). Specifically, GO_BP involves the regulation of innate immune response, positive regulation of innated immune response and positive regulation of response to biotic stimulus. GO_CC emphasizes the vacuolar membrane, lytic vacuole membrane, and lysosomal membrane. GO_MF primarily focuses on structural constituent of ribosome. In addition, the figure presented the KEGG secondary classification, which listed the top 20 entries. Significantly, the lysosome signaling pathway had the highest number of genes and the p-value of lysosome signaling pathway was the smallest (Figure 3D).

The Co-Expressed Genes and Pathways with C3ar1

The macrophage transcriptome was analyzed through WGCNA. Initial hierarchical clustering of macrophage subpopulations employed Pearson correlation metrics with average linkage. Network construction required selection of an optimal soft-thresholding power, with β=2 chosen to achieve scale-free topology (Figure 4A). Gene clustering was performed using topological overlap matrix (TOM)-based dissimilarity (1-TOM), revealing distinct expression patterns (Figure 4B). Dynamic tree cutting identified 11 co-expression modules (Figure 4C), each characterized by unique eigengene expression profiles that potentially represent functional gene sets. WGCNA revealed a high interconnectivity among Modules 2, 4, 6, and 10. There were genes such as Cd68, C3ar1, C1qa, C1qb, C1qc (Figure 4E) in the modules. Therefore, we selected Module 2, 4, 6, 10 as the interesting modules. GO and KEGG analysis indicated modules were associated with lysosome (Figure 4F–G). The lysosome signaling pathway enriched the highest number of genes in cellular processes.

Figure 4 The results of high dimensional weighted gene correlation network analysis (hdWGCNA). (A) The power of β = 2 (scale-free R 2 = 0.95) was selected as the soft-thresholding parameter. (B) A dendrogram of all differentially expressed genes was clustered based on a dissimilarity measure. (C) The presentation of the kME calculation results of the genes in 11 modules. (D) The functionality of the correlations between each module based on their hME, ME, or hub gene scores. (E) A network diagram depicting the interaction relationships between genes within the module 2, 4, 6 10. (F) The result of GO and KEGG (G) enrichment analysis of genes in the 4 modules.

Identifying Hub Genes and Functional Annotation

Among 246 module genes and 1,022 DEGs, 89 overlapping genes were analyzed via PPI network and KEGG pathways (Figure 5A and B). These genes were enriched in cellular processes, particularly the lysosome pathway, which had the most genes linked to cell growth and death (Figure 5C and D). The PPI network identified the top 30 hub genes (Figure 5E), and their KEGG analysis also highlighted the lysosome pathway (Figure 5F).

Figure 5 Results of network pharmacology analysis. (A) Venn diagram depicting the target intersection of the DEGs and the module genes of macrophages on D5. (B) Protein-protein interactions network of 89 targets. (C) The result of the KEGG and GO (D) enrichment analysis for 89 targets. (E) PPI network indicates lists of top 30 core targets. (F) The result of KEGG for top 30 core targets.

Network Pharmacology Ctsd, Stat1, Il1b and Cxcl10 as Targets of Quercetin

These bioinformatics analyses collectively highlighted the lysosomal pathway, with genes such as Ctsd, as being tightly co-expressed with the MMT driver C3ar1. This suggested a potential functional role for this pathway in the C3a-driven transition process. To explore whether a therapeutic compound could intervene at this nexus, we next employed network pharmacology to intersect these MMT and fibrosis-associated genes with known targets of the natural flavonoid quercetin.

Network pharmacology identified Ctsd, Stat1, Il1b, and Cxcl10 as potential quercetin targets among 167 candidate genes (Figure 6A). Molecular docking confirmed strong binding between quercetin and these proteins, with binding energies of −7.8 kcal/mol (Ctsd), −7.1 kcal/mol (Stat1), −7.0 kcal/mol (Il1b), and −6.3 kcal/mol (Cxcl10) (Figure 6B–F, Figure S3A–D).

Figure 6 Molecular docking results. (A) Venn diagram depicting the target intersection of 89 targets and quercetin targets. (B) The 3D structure of quercetin. (C) Schematic diagram of docking and binding mode Ctsd and quercetin. (D) Schematic diagram of docking and binding mode Il1b and quercetin. (E) Schematic diagram of docking and binding mode Stat1 and quercetin. (F) Schematic diagram of docking and binding mode Cxcl10 and quercetin.

Quercetin Inhibits MMT in C3a-Stimulated BMDMs

To determine whether quercetin could suppress MMT development, we established an in vitro model using C3a-stimulated BMDMs. As expected, C3a stimulation significantly induced MMT, as evidenced by markedly increased α-SMA expression compared to untreated controls (Figure 7A–C).

Figure 7 Quercetin Inhibits MMT in C3a-Stimulated BMDMs. (A) CCK-8 assay determining BMDMs viability post-quercetin treatment at different doses (0, 20μM, 40μM, 60μM, 80μM, 100μM, 120μM). (B) The representative immunofluorescence images demonstrate the location and relative expression of CD68 and α-SMA, and quantified (C). Data was presented as mean ± SD. *p<0.05, **p<0.01, ****p<0.0001.

Treatment with quercetin produced a concentration-dependent suppression of α-SMA expression in C3a-stimulated BMDMs. The higher concentration of quercetin (100μM) exhibited more pronounced inhibition compared to its lower concentration counterpart (40μM), clearly demonstrating a dose-response relationship in quercetin’s ability to counteract C3a-induced MMT (Figure 7A–C).

Quercetin Inhibited Skeletal Muscle Fibrosis and Promote the Gait Recovery of Mice

The expression levels of fibrosis markers (α-SMA and Col1a1) were examined via Western blot, revealing that contusion injury can upregulate the expression of fibrosis-related markers, while quercetin could mitigate this pathology (Figure 8A–C).

Figure 8 The effect of quercetin in inhibiting skeletal muscle fibrosis on D7. (A) Representative Western blot images showing protein levels of α-SMA and collagen I (Col1) in gastrocnemius muscles from control (Con), injured (Inj), and quercetin-treated (Inj+Que) groups at day 7. (B and C) Quantitative analysis of α-SMA (B) and Col1 (C) protein expression normalized to GAPDH. Data are presented as mean ± SD (n=3). *p < 0.05. (D) Representative immunofluorescence images of muscle sections stained for CD206 (green), α-SMA (red), and DAPI (blue). Scale bar: 50 μm. (E and F) Quantification of fluorescence intensity for CD206 (E) and α-SMA (F). Data are presented as mean ± SD (n=3). **p < 0.01. (G) Representative immunofluorescence images stained for dystrophin (green). Scale bar: 50 μm. (H) Quantification of fluorescence intensity for dystrophin, ***p < 0.001.

Immunofluorescence staining results confirmed that excessive CD206 and α-SMA expression was detected in the injured tissue and the effect of the drug was manifested on D7 (Figure 8D–F). Besides, quercetin treatment significantly upregulated dystrophin expression (Figure 8G and H).

Additionally, quercetin decreased the expression of COL1 in comparison with the injured group on D14 (Figure 9A and B). HE and Masson staining revealed quercetin can reduce the accumulation of collagen and promote the healing of muscles on 28 (Figure 9C and D).

Figure 9 The effect of quercetin in inhibiting skeletal muscle fibrosis on D14. (A) Representative immunofluorescence images of gastrocnemius muscle sections at day 14 post-injury, stained for collagen I (Col1, red) and DAPI (blue, nuclei). Scale bar: 50 μm. (B) Quantification of Col1-positive area from images in (A), expressed as a percentage of the total field area. Data are presented as mean ± SD (n=3). *p < 0.05 (C) HE staining was utilized to demonstrate the newborn muscle fibers, while MASSON staining showed the collagen deposition. Scale bar: 100 μm. (D) Quantification of fibrotic area from Masson’s trichrome-stained sections, expressed as the percentage of blue-stained collagen area relative to the total muscle area. Data are presented as mean ± SD (n=3). *p < 0.05. (E) Representative paw print patterns from the CatWalk XT gait analysis system at day 28. The yellow box highlights the right hind (RH) limb track, and the Orange box highlights the left front (LF) limb track. (F) Quantitative gait analysis. Left panel: Print length (the distance of paw contact with the glass plate during stance phase). Right panel: Swing speed (the speed of the paw movement during swing phase). ***p < 0.01.

To assess the functional recovery, the catwalk analysis was conducted across groups on the 28th day after injury. The result of gait analysis showed that the rear limb function of the quercetin group was greatly improved compared to contusion group (Figure 9E and F).

Discussion

In this study, we confirmed the presence of MMT in skeletal muscle, with its peak abundance on D5. Additionally, our findings indicated that the lysosome pathway might play a critical role. Quercetin exerted a therapeutic effect by regulating Ctsd, Stat1, Il1b and Cxcl10 within the lysosomal pathway. Consequently, validations were performed both in vitro and in vivo, demonstrating that quercetin effectively reduced fibrosis following skeletal muscle injury and facilitated the recovery of the locomotor function in mice.

Fibrosis following skeletal muscle injury is a troubling issue. Severe fibrosis can lead to muscle atrophy and weakness. A large number of scholars have studied how to slow down or reverse fibrosis, but it remains controversial. Despite their undisputed role as the primary extracellular matrix-producing cells in fibrosis, the developmental lineages of myofibroblasts continue to elude complete characterization, with emerging single-cell RNA sequencing studies revealing previously unappreciated progenitor populations.²²

A growing body of research has demonstrated MMT as a key event in fibrogenesis.²³ The transformation of macrophages into myofibroblasts was first proposed by Nikolic-Paterson et al They identified MMT processes in human and experimental fibrotic nephropathy through co-expression of macrophage (CD68) and myofibroblast (α-SMA) antigens.²⁴ Yang et al systematically examined macrophage-myofibroblast transition (MMT) in a rat unilateral ureteral obstruction model of pulmonary fibrosis. Their findings demonstrated that MMT-derived cells constitute a significant proportion of activated myofibroblasts and contribute substantially to fibrotic progression in pulmonary tissues.²⁵ We have previously proven the existence of MMT after skeletal muscle injury via complement system activation.¹¹ In this study, we further certificated that inhibiting MMT can antagonized skeletal muscle fibrosis.

We conducted an analysis using publicly available scRNA-seq data and found that the number of MMT cells peaked on D5. Further analysis of macrophages alone revealed that, compared to D2, macrophages on D5 expressed significantly higher levels of *COL1A1* and *PDGFA*indicating a fibrotic phenotype. CellChat analysis also demonstrated increased communication between macrophages and fibroblasts on D5, primarily through the PDGF/PDGFR signaling pathway and COL1A1-related pathways. Stimulation of the PDGF/PDGFR signaling cascade promotes key cellular processes including proliferative expansion, directional migration, and enhanced synthesis of extracellular matrix components,²⁶ while COL1A1 is directly associated with collagen synthesis.²⁷ Collectively, these dual mechanisms synergistically drive fibrotic progression. Our experimental evidence conclusively establishes macrophage-myofibroblast transition (MMT) as a pivotal contributor to both fibrogenesis initiation and disease advancement.

In an effort to explore how C3a stimulates the MMT process of macrophages, we performed an intersection analysis between genes upregulated in day-5 macrophages and C3ar1-related module genes. This approach identified 89 candidate genes sharing biological functions with C3ar1 that were significantly upregulated in our model system. Notably, several of these genes are associated with lysosomal signaling pathways, including Ctsd, Galns, Hexb, Cd68, Ctsz, Glb1, Lipa, Psap, and Tpp1. We postulate that ROS accumulation, induced by skeletal muscle injury, functions as a primary mediator of lysosomal membrane destabilization.^28,29 This results in the leakage of cathepsins into the cytosol and subsequent impairment of normal lysosomal degradation function, potentially explaining the sustained upregulation of lysosomal-related genes observed in our model.²⁹ Protein-protein interaction (PPI) network analysis further revealed Ctsd and Ctsz as central hub genes in this process. Ctsd and Ctsz (cathepsin D/Z) are lysosomal aspartic proteases. Emerging evidence indicates that calpain and cathepsins play pivotal roles in mediating muscle fiber degradation in inflammatory myopathies.³⁰ And this finding aligns with current literature demonstrating that inhibition of lysosomal proteases, particularly Ctsd, can attenuate renal fibrosis progression.³¹ Furthermore, beyond skeletal muscle, Ctsd upregulation has been implicated in fibrogenesis in other organs, such as the liver, highlighting its broader role in myofibroblast activation.³² We observed concurrent upregulation of cathepsin Z (Ctsz) and Ctsd, consistent with established roles of cathepsins in NLRP3 inflammasome activation and subsequent IL-1β/IL-18 maturation and secretion.^33–36

Quercetin was a flavonoid found in vegetables and fruits, exhibiting anti-inflammation properties.¹⁵ Research has demonstrated the efficacy of quercetin against renal fibrosis,¹⁶ liver fibrosis,¹⁷ lung fibrosis,¹⁸ myocardial fibrosis¹⁹ and muscle atrophy. Recent studies demonstrate that quercetin inhibits the adipogenic and fibrogenic differentiation potential of muscle-resident PDGFRα+/CD201+ progenitor cells, effectively blocking their transition into adipocytes and fibroblasts.¹⁵ While quercetin attenuated fibrosis in human skeletal muscle, the interactions between its active ingredients and molecular targets were not delineated. Consequently, we employed network pharmacology, an established systems biology approach that systematically identifies bioactive components in herbal medicines, maps their target interactions, and deciphers the complex biological networks underlying drug-target-disease relationships.³⁷ By intersecting 167 potential targets of quercetin with 89 muscle fibrosis-related targets, we obtained 4 common genes (Ctsd, Stat1, Il1b and Cxcl10). The molecular docking of 4 protein and quercetin demonstrated excellent binding ability. Notably, Ctsd is most likely to bind with it.

Our cell experiments provided direct evidence that quercetin dose-dependently inhibits C3a-induced MMT in BMDMs. The observed suppression of α-SMA expression, particularly at higher quercetin concentrations, demonstrates its capacity to interrupt this critical fibrogenic process.

To further validate these findings, we performed comprehensive analyses of gastrocnemius muscle samples from control, contusion-injured, and quercetin-treated groups. Quercetin significantly mitigated contusion-induced fibrosis, lowering the level of both CD206 and α-SMA. These findings suggest that quercetin effectively inhibits MMT following muscle contusion injury.

Notably, quercetin treatment also markedly enhanced dystrophin expression. Dystrophin, a cytoskeletal protein localized at costameres, serves as a critical structural component that links the intracellular cytoskeleton to transmembrane proteins and the ECM.³⁸ Previous studies have established dystrophin as a robust biomarker for evaluating myofiber maturity at the single-fiber level.³⁹ Corroborating these molecular improvements, CatWalk gait analysis revealed significant functional recovery in quercetin-treated mice on D28. The coordinated improvement in both dystrophin expression and locomotor function strongly suggests that quercetin promotes structural and functional restoration of injured skeletal muscle.

Quercetin treatment significantly reduced collagen I-positive myofibroblasts by day 14 post-injury, demonstrating its potent anti-fibrotic effects. Histological analyses further confirmed the reduction in activated myofibroblasts, with H&E staining demonstrating improved muscle structure and Masson’s trichrome staining revealing significantly reduced collagen deposition compared to untreated injured controls.

Integrating our multi-omics and experimental data, we propose a coherent model for how quercetin ameliorates skeletal muscle fibrosis (Graphical Summary). Following contusion injury, C3a drives a subset of macrophages toward an MMT phenotype, a process associated with lysosomal pathway activation. Quercetin is predicted to interact with key nodes within this process, most notably Ctsd. This interaction is postulated to modulate lysosomal function, thereby inhibiting the C3a-induced MMT program. The downstream consequence of this inhibition is a significant reduction in the generation of MMT-derived myofibroblasts, leading to decreased fibrosis, improved muscle structure, and ultimately, enhanced functional recovery.

Limitations

This study has several limitations that should be considered. First, while our integrated bioinformatics approach identified compelling candidate targets and pathways through which quercetin may inhibit MMT, the direct relationship between these specific targets and the observed anti-fibrotic effects remains to be experimentally established. Second, although we demonstrate a reduction in MMT-derived myofibroblasts, the relative contribution of this source compared to other potential sources to the overall fibrotic burden in this model was not quantified. Third, the pharmacokinetics and optimal long-term dosing regimen of quercetin for muscle fibrosis require further investigation.

Conclusion

Our results identify quercetin as a promising inhibitor of skeletal muscle fibrosis, highlighting its therapeutic potential for fibrotic muscle disorders. We provide evidence that quercetin’s anti-fibrotic effect is associated with the inhibition of MMT. Through integrated RNA-seq and network pharmacology analyses, we further implicate the lysosomal signaling pathway, with Ctsd as a central candidate, in this process. These findings advance our understanding of muscle fibrosis and offer a rationale for exploring quercetin-based treatments.