Single-cell omics reveals arg-1 as a key regulator of age-dependent macrophage-mediated cartilage repair

Abstract

Aging impairs cartilage repair, with young animals exhibiting superior regenerative capacity due to enhanced tissue repairing and reduced inflammation compared to aged counterparts. This study employed single-cell omics to dissect age-dependent immune cell heterogeneity in cartilage injury, revealing a critical deficiency in anti-inflammation macrophage subsets in aged animals. We identified Arg-1 as a central regulator of macrophage polarization, demonstrating that its overexpression rescues impaired repair in aged animals. These findings establish Arg-1 as a novel therapeutic target to counteract age-related declines in cartilage regeneration, offering new insights into macrophage-driven tissue repair mechanisms. The integration of single-cell analysis with functional validation provides a framework for developing precision interventions for age-impaired tissue regeneration.

Introduction

Aging is a significant factor influencing the recovery capacity following cartilage injury, with notable differences observed between older and younger animals [1]. Studies indicate that younger animals exhibit enhanced regenerative potential, including better cartilage repair and reduced inflammatory responses, compared to their older counterparts. This disparity may be attributed to age-related declines in stem cell activity, extracellular matrix synthesis, and immune function [2]. Additionally, older animals often experience prolonged joint inflammation and slower tissue remodeling, further impairing recovery [3]. Understanding these age-dependent variations is crucial for developing targeted therapeutic strategies [4].

The immune processes underlying cartilage injury are highly complex and heterogeneous, posing significant challenges in identifying effective therapeutic targets [5]. Arthritis involves a dynamic interplay between innate and adaptive immune responses, with key roles for pro-inflammatory cytokine, immune cell infiltration, and dysregulated synovial fibroblast activation [6]. This complexity is further compounded by patient-specific variations in immune pathways, where different molecular mechanisms may drive disease progression in subsets of individuals [7]. Additionally, the crosstalk between immune cells and joint tissues creates a self-perpetuating cycle of inflammation and tissue damage, making it difficult to disrupt the disease process without causing systemic immunosuppression [8]. As a result, despite advances in biologic and targeted therapies, many patients exhibit incomplete responses, highlighting the need for more precise and stratified treatment approaches.

Macrophages play a multifaceted and context-dependent role in the pathogenesis of cartilage injury, contributing to both inflammatory progression and tissue repair [9]. In the synovial microenvironment, macrophages exhibit remarkable plasticity, dynamically shifting between pro-inflammatory (M1-like) and anti-inflammatory (M2-like) phenotypes in response to local signals [10]. While M1-polarized macrophages drive joint inflammation through the production of cytokines such as tumor necrosis factor-α （TNF-α）, Interleukin-1β（IL-1β）, and Interleukin-6（IL-6）, M2-like macrophages promote resolution of inflammation and tissue remodeling [11–13]. However, this dichotomy is oversimplified, as single-cell studies reveal a spectrum of macrophage activation states in cartilage injury, with distinct subsets associated with disease severity and treatment response [14]. Furthermore, synovial macrophages interact with fibroblasts, T cells, and osteoclasts, forming a complex cellular network that perpetuates joint destruction [15]. Understanding the heterogeneity and functional diversity of macrophages in cartilage injury may uncover novel therapeutic opportunities to modulate their activity and restore immune homeostasis.

Recent advances in single-cell sequencing technologies have revolutionized our ability to dissect such complexity at an unprecedented resolution [16–18]. Unlike bulk sequencing, which averages signals across heterogeneous cell populations, single-cell omics enables high-resolution profiling of individual cells, uncovering rare cell subtypes, dynamic transcriptional states, and intricate cell-cell interactions [19]. This approach is particularly powerful in studying complex diseases such as cancer, autoimmune disorders, and inflammatory conditions, where cellular heterogeneity plays a critical role in disease progression and treatment resistance [20]. The field is increasingly reliant on single-cell omics to move beyond descriptive cell type cataloging towards identifying precise molecular drivers and predictive biomarkers. A paradigm of this approach is illustrated in inflammatory bowel disease (IBD) research, where single-nucleotide variant (SNV) analysis at the gene level within specific gut microbiota species (e.g., Faecalibacterium prausnitzii) has successfully identified highly accurate diagnostic markers, outperforming traditional species-level abundance metrics [21,22]. This underscores the power of single-cell resolution genomics to reveal functionally relevant, disease-specific molecular signatures that are masked in population-averaged data. By integrating transcriptomic, epigenomic, and proteomic data at single-cell resolution, researchers can now identify novel biomarkers, delineate disease-driving pathways, and discover potential therapeutic targets with greater precision than ever before [20,21]. As these technologies continue to advance, they hold immense promise for unraveling the complexity of cellular ecosystems and accelerating the development of precision medicine strategies across a wide range of diseases [18].

Building upon this conceptual framework, our study employed single-cell RNA sequencing (scRNA-seq) to investigate the differential recovery capacity between young and aged animals following cartilage injury, explicitly addressing the inherent heterogeneity of immune cells within the joint. Through comprehensive profiling of joint tissues before and after injury, we aimed to identify age-dependent molecular mechanisms that govern post-injury recovery. Our analysis revealed that young animals exhibit a significantly higher proportion of anti-inflammatory macrophage subsets compared to aged counterparts, suggesting a link between specific immune cell states and enhanced tissue repair potential.

Further network analysis pinpointed Arg-1 (Arginase-1) as a central regulator within anti-inflammation macrophages. Functional validation through in vivo and in vitro experiments demonstrated that Arg-1 overexpression inhibited inflammation and ROS releasing in aged animals, partially rescuing their impaired recovery phenotype. These results not only elucidate the mechanistic basis for age-related disparities in cartilage injury recovery but also highlight Arg-1 as a novel therapeutic target to improve joint repair in elderly individuals. By integrating single-cell omics with mechanistic validation, this study provides critical insights into anti-inflammation macrophage in cartilage injury and offers a potential strategy to mitigate age-associated decline in tissue regeneration.

Materials and methods

Materials

pAAV-CMV-MCS-3FLAG-WPRE-pA (Addgene, USA); pAAV8-RC (Addgene, USA); pHelper (Addgene, USA); PrimeSTAR Max DNA Polymerase (Takara, Japan); NheI (New England Biolabs, USA); XhoI (New England Biolabs, USA); T4 DNA ligase (New England Biolabs, USA); Stbl3 competent cells (Thermo Fisher Scientific, USA); polyethylenimine (PEI) (Polysciences, USA); iodixanol (Beyotime Biotechnology, China); RPMI-1640 medium (Gibco, USA); FBS (Gibco, USA); penicillin/streptomycin (Gibco, USA); M-CSF (PeproTech, USA); PE-conjugated anti-FLAG antibody (BioLegend, USA); PE-IgG1κ isotype control (BioLegend, USA); meloxicam (Boehringer Ingelheim, Germany); paraformaldehyde (PFA) (Sigma-Aldrich, USA); EDTA (Aladdin, China); ethanol series (Aladdin, China); xylene (Beyotime Biotechnology, China); paraffin wax (Leica, Germany); Harris hematoxylin (Beyotime Biotechnology, China); eosin Y (Beyotime Biotechnology, China); Biebrich scarlet-acid fuchsin (Sigma-Aldrich, USA); Mouse Arg-1 Elisa kit (FineTest, China); phosphomolybdic-phosphotungstic acid Sigma-Aldrich, USA); aniline blue (Macklin, China); fast green (Macklin, China); safranin O (Beyotime Biotechnology, China); neutral balsam (Beyotime Biotechnology, China). Rabbit anti rat Arg-1 antibody (ab203284), rabbit anti rat Cyclophilin B antibody (ab178697) purchased from Abcam (USA).

Animals

Male Sprague-Dawley rats weighing ~250 g were procured from the Zhejiang Academy of Medical Sciences. The Zhejiang University Animal Experimentation Committee granted approval for all research procedures, ensuring strict adherence to the guidelines set forth by the National Institutes of Health Guide for the Care and Use of Laboratory Animals (ZJU20250394).

Single cell sequence for old and young animals before and after cartilage injury

Raw sequencing data were obtained from the GEO dataset (accession GSE236843). The preprocessing and quality control procedures were conducted as follows. First, raw sequencing reads (FASTQ files) were processed using the Cell Ranger pipeline (v7.1.0). Read alignment was performed against the rat reference genome (mRatBN7.2) to generate a gene expression matrix. Low-quality cells and potential doublets were filtered out using the following stringent criteria: (1) cells with fewer than 500 unique molecular identifiers (UMIs) or more than 50,000 UMIs were excluded to remove empty droplets and potential multiplets; (2) cells expressing fewer than 1,000 genes were discarded to eliminate low-information captures; (3) cells with mitochondrial gene content exceeding 20% were removed to exclude damaged or dying cells. In addition, potential doublets were predicted and removed using Scrublet, and cells with a predicted doublet score > 0.25 were excluded from downstream analysis (S1 Fig). Following quality control, gene expression matrices were normalized and log-transformed using the global-scaling method in Seurat (v5.0.1). To account for confounding technical variation, we regressed out the effects of total UMIs per cell and mitochondrial gene percentage using the ScaleData function. Highly variable genes (HVGs) were selected using the FindVariableFeatures method with the “vst” selection method, retaining the top 2,000 HVGs for dimensionality reduction. Principal component analysis (PCA) was performed on the scaled HVG matrix. Significant principal components (PCs) were determined via visual inspection of the elbow plot and by applying the JackStraw permutation test (significant PCs: p-value < 0.05). The selected PCs were used for downstream Uniform Manifold Approximation and Projection (UMAP) and graph-based clustering at a resolution of 0.5. Cell clusters were annotated using well-established marker genes (e.g., CD68 for macrophages, COL1A1 for fibroblasts). Differentially expressed genes (DEGs) between clusters or across experimental conditions were identified using the MAST algorithm with thresholds of adjusted p-value < 0.05 and |log₂ fold change| > 0.25.

Expression of ECM related genes in different macrophage subtypes

To evaluate the expression of extracellular matrix (ECM)-related genes in distinct macrophage subtypes, we focused on *Arg-1* and Col8a1 as representative markers. Single-cell RNA sequencing data from three independent biological replicates were pooled to ensure robust and reproducible analysis. Expression levels of *Arg-1* and Col8a1 were extracted for each macrophage subtype. Given the non-normal distribution of gene expression data and the presence of zero-inflated values, non-parametric statistical testing was employed. Differences in expression across macrophage subtypes were assessed using the Kruskal–Wallis test, followed by Dunn’s post-hoc test with Benjamini–Hochberg correction for multiple comparisons. Results are presented as median expression values with interquartile ranges.

Cell ratio evaluation for different cell clusters

To quantify the proportional distribution of macrophage subtypes within the cellular landscape, we performed cell ratio analysis based on well-established marker gene expression. Single-cell RNA sequencing data from three independent biological replicates were processed using the Seurat pipeline (v5.0). Cells were clustered based on uniform manifold approximation and projection (UMAP) and annotated according to lineage-specific markers. Macrophage subsets were further subclassified based on the above marker thresholds. The proportion of each cell clusters or macrophage subtype was calculated as the percentage of cells within the total macrophage population per biological replicate. To assess statistical differences in subtype proportions across experimental groups, one-way analysis of variance (ANOVA) was performed, followed by Tukey’s honestly significant difference (HSD) post-hoc test for multiple comparisons. All analyses were conducted in R (v4.3.0) and data are presented as mean ± standard deviation of three biological replicates. Assumptions of normality and homogeneity of variance were verified using Shapiro–Wilk and Levene’s tests, respectively.

GO enrichment for functions of DEGs in different groups

Differentially expressed genes (DEGs) identified from single-cell RNA sequencing analysis with an adjusted p-value < 0.05 and absolute log2 fold change > 0.25 were subjected to Gene Ontology (GO) enrichment analysis using the clusterProfiler package (v4.0) in R (v4.3.1). The analysis included biological processes (BP), molecular functions (MF), and cellular components (CC) categories. The reference gene set was derived from the organism-specific annotation database (e.g., org.Hs.e.g.,db for human data). Over-representation analysis was performed using the enrichGO function with the Benjamini-Hochberg method for multiple testing correction, and terms with a corrected p-value < 0.05 were considered statistically significant. The results were visualized using dot plots or bar plots to highlight enriched terms, with gene counts and adjusted p-values displayed. Redundant GO terms were simplified using the simplify function to reduce overlap and improve interpretability. Leading terms were selected based on fold enrichment and statistical significance to summarize key biological pathways associated with the DEGs.

Pseudotime Trajectory Analysis for macrophage subtypes

Single-cell pseudotime analysis was performed to reconstruct cellular differentiation trajectories using Monocle3 (v1.0.0) in R (v4.3.1). The input data consisted of normalized and log-transformed gene expression matrices from Seurat (v5.0.1), retaining HVGs identified during prior clustering analysis. Cells were pre-processed in Monocle3 by applying PCA for dimensionality reduction, followed by UMAP for nonlinear embedding. The learn_graph function was used to construct a principal graph capturing the underlying developmental trajectory, with the root node manually selected based on known marker genes or automatically inferred from progenitor cell signatures. Pseudotime values were then calculated for each cell using the order_cells function, ordering cells along the trajectory based on transcriptional similarity. Branch-dependent differential expression analysis was performed using the graph_test function to identify genes significantly associated with pseudotime or branching points (q-value < 0.01). Genes with dynamic expression patterns along the trajectory were clustered using k-means and visualized in heatmaps to highlight stage-specific transcriptional programs. Trajectory plots were overlaid with key marker genes or module scores to interpret cellular states during differentiation. The analysis was repeated with alternative trajectory inference methods for robustness validation.

Weighted Gene Co-expression Network Analysis (WGCNA) screening ECM related gene network in macrophage

WGCNA was performed to identify co-expressed gene modules and their associations with clinical traits using the WGCNA package (v1.72) in R (v4.3.1). The input data consisted of normalized gene expression matrices from bulk RNA-seq or aggregated single-cell RNA-seq data, focusing on the top 5,000 most variable genes based on median absolute deviation (MAD). A signed adjacency matrix was constructed by calculating pairwise Pearson correlations between genes, followed by raising the correlation matrix to a soft-thresholding power (β = 12, selected based on scale-free topology criterion R² > 0.9). The adjacency matrix was transformed into a topological overlap matrix to minimize noise, and hierarchical clustering with dynamic tree cutting (minModuleSize = 30, mergeCutHeight = 0.25) was applied to identify co-expression modules. Module eigengenes were calculated as the first principal component of each module and correlated with clinical traits (e.g., disease severity or Arg-1 expression) to identify trait-associated modules (p < 0.05). Intramodular hub genes were identified based on gene significance (GS) and module membership (MM) thresholds (GS > 0.2, MM > 0.8). Functional enrichment analysis of key modules was performed using clusterProfiler (v4.0) for GO and KEGG pathways (FDR < 0.05). Network visualization was implemented in Cytoscape (v3.9.1) by exporting the top 100 edges per module based on TOM dissimilarity. Robustness was validated by repeating the analysis with alternative soft-thresholding powers and clustering parameters.

Confirmation of Arg-1 expression in young and old OA animals

To validate differential Arg-1 expression in osteoarthritis (OA) models across age groups, we performed Western blot analysis on cartilage tissues obtained from young and old OA animals. Cartilage specimens were harvested from the femoral condyles and tibial plateaus of both age groups following established OA induction protocols (e.g., surgical destabilization or chemical induction). Tissues were snap-frozen in liquid nitrogen and homogenized in RIPA lysis buffer supplemented with protease inhibitors for western blotting analysis. Data were analyzed using Student’s t-test (young vs. old OA groups), with statistical significance set at p < 0.05. Each group included 3 biological replicates (n = 3 animals per group). Results are presented as mean fold-change ± standard error of the mean (SEM).

Construction of AAV8 mediate Arg-1 overexpression system

The AAV8 vector plasmid for rat Arg-1 (NCBI Reference Sequence: NM_012656.2) overexpression was generated by cloning the full-length rat Arg-1 cDNA into the pAAV-GFP-MCS-3FLAG-WPRE-pA expression backbone (Addgene #105535) using restriction enzyme-based cloning. The rat Arg-1 coding sequence was amplified from rat cDNA using PrimeSTAR Max DNA Polymerase (Takara) with forward primer 5’-GCTAGCGCCACCATGGCTCTGACGGGCTTCTG-3’ (containing NheI site and Kozak sequence) and reverse primer 5’-CTCGAGTCAGTCCTGGGCTCTTCTTGGC-3’ (containing XhoI site). The PCR product and linearized vector were digested with NheI and XhoI (NEB), ligated using NEB, and transformed into Stbl3 competent cells. Positive clones were verified by colony PCR and sequencing. The final construct (pAAV8-CMV-rArg-1–3FLAG) was packaged into AAV8 particles by co-transfecting HEK293T cells with pAAV8-RC (Addgene #112864) and pHelper (Addgene #112863) using polyethylenimine (PEI), followed by purification through iodixanol gradient centrifugation. Viral titer was determined by quantitative PCR (qPCR) targeting the WPRE sequence.

Assessment of AAV capsid purity by silver staining

The purity of the produced AAV8-MCS-rArg-1–3FLAG particles was assessed by silver staining to visualize the relative abundance of viral capsid proteins. Briefly, purified AAV samples in iodixanol fractions were mixed with 4 × Laemmli sample buffer (containing 2-mercaptoethanol) and heated at 95°C for 5 minutes. Proteins were separated by 10% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) at 120 V for 90 minutes using a Mini-PROTEAN Tetra system (Bio-Rad). After electrophoresis, the gel was fixed overnight in a solution containing 40% ethanol and 10% acetic acid, then sensitized for 30 minutes in 0.02% sodium thiosulfate. The gel was rinsed thoroughly with ultrapure water and incubated in a 0.1% silver nitrate solution for 30 minutes at room temperature with gentle agitation. Following development in a solution containing 2% sodium carbonate and 0.04% formaldehyde, the reaction was stopped with 5% acetic acid. The gel was imaged using a ChemiDoc MP imaging system (Bio-Rad). The purity of the viral preparation was evaluated by the ratio of the total intensity of the three characteristic VP bands (62 kDa) to the total protein intensity in the lane. Only preparations with a purity exceeding 90% were used for subsequent experiments (S2A in S2 Fig).

Determination of AAV vector titer by quantitative PCR (qPCR)

The titer of AAV vectors, expressed as viral genomes per milliliter (vg/mL), was determined by absolute quantification using qPCR targeting the Woodchuck Hepatitis Virus Posttranscriptional Regulatory Element (WPRE) sequence present in both the AAV8-MCS-rArg-1–3FLAG and AAV8-empty vectors. A plasmid containing a single copy of the WPRE sequence (pAAV-CMV-MCS-3FLAG-WPRE-pA) was used as the quantitative standard. The plasmid was linearized with a restriction enzyme that cuts outside the WPRE region, and its concentration was measured using a Qubit 4.0 Fluorometer (Thermo Fisher Scientific) with the dsDNA HS Assay Kit. The copy number of the linearized plasmid was calculated using the following formula:

Copies/µL=[DNA(g/µL)×6.022×10²³]/[Length (bp)×660]

A 10-fold serial dilution series was prepared in nuclease-free water to generate standard samples ranging from 1 × 10⁸ to 1 × 10¹ copies/µL. Each dilution was aliquoted and stored at −20°C to avoid repeated freeze-thaw cycles. A linear regression analysis was performed, and the slope, y-intercept, and coefficient of determination (R²) were recorded. Only standard curves with an R² value > 0.99 were accepted. The Ct value of each unknown AAV sample was averaged from its technical replicates and interpolated onto the standard curve to determine the corresponding log10 copy number. The titer in vg/mL was calculated using the following formula, accounting for all dilution factors during sample processing:

Titer (vg/mL)=Copies from curve×Dilution Factor×1000/Volume of AAV used (5 µL)

The final reported titer for each AAV batch is the mean of three independent qPCR determinations.

AAV transduction of macrophages and efficiency validation

For in vitro transduction, RAW264.7 macrophages were seeded in 6-well plates at a density of 2 × 10⁵ cells per well in complete RPMI-1640 medium (supplemented with 10% FBS and 20 ng/mL M-CSF) and allowed to adhere overnight. The following day, the culture medium was replaced with 1 mL of fresh, serum-free RPMI-1640 medium. Purified AAV8-CMV-rArg-1–3FLAG or AAV8-empty control vector was added to the cells at a multiplicity of infection (MOI) of 1 × 10⁵ vector genomes per cell (vg/cell). The plates were gently swirled to ensure even distribution of the viral particles and incubated at 37°C with 5% CO₂ for 6 hours. Subsequently, 1 mL of complete medium containing 20% FBS and 40 ng/mL M-CSF was added to each well, bringing the final serum concentration to 10% and M-CSF to 20 ng/mL, without removing the initial viral inoculum. Cells were then returned to the incubator and cultured for an additional 42 hours (total of 48 hours post-transduction) before harvesting for downstream analyses. The average transduction efficiency for AAV8-CMV-rArg-1–3FLAG in RAW264.7 macrophages across three independent experiments was 54.2% ± 3.1% (mean ± SD) (S2B in S2 Fig).

Macrophage culture and Arg-1 overexpression validation

RAW264.7 in RPMI-1640 medium supplemented with 10% FBS, 1% penicillin/streptomycin, and 20 ng/mL M-CSF (PeproTech) for 7 days at 37°C with 5% CO₂. For Arg-1 overexpression, differentiated macrophages were transduced with AAV8-GFP-rArg-1–3FLAG (MOI = 1 × 10⁵ vg/cell) or control AAV8-empty vector in serum-free medium for 6 hours, followed by replacement with complete medium. At 48 hours post-transduction, cells were harvested for analysis. To validate Arg-1 expression, macrophages were obtained and analyzed by flow cytometry. Data were processed using FlowJo (v10.8.1), with Arg-1 overexpression quantified as the geometric mean fluorescence intensity (gMFI) shift in GFP channel (Ex/Em = 488/525 nm).

Establishing of cartilage injury rat model and intra-articular injection of arg-1-overexpressing vector

Male Sprague-Dawley rats (48 weeks old) were randomly assigned to experimental groups prior to surgery using a computer-generated randomization sequence to minimize allocation bias. The group size (n = 10 per group) was determined based on sample sizes commonly employed in similar preclinical studies of joint injury and gene therapy, which have been shown to provide sufficient statistical power to detect significant differences in histological and molecular outcomes, and this rationale is explicitly provided as our study was exploratory in establishing the novel therapeutic role of Arg-1. Following randomization, rats were anesthetized with isoflurane (3% induction, 1.5% maintenance) and the right knee joint was sterilized with iodophor for the surgical induction of cartilage injury via anterior cruciate ligament transection combined with partial medial meniscectomy. After a medial parapatellar incision and joint exposure, the anterior cruciate ligament was completely transected, and a standardized osteochondral defect (2 mm in diameter, 10 mm in depth) was created in the medial femoral condyle using a micro-drill, followed by layered closure of the joint capsule and skin. Sham-operated control rats underwent identical surgical exposure, including capsulotomy, but without ligament transection or meniscus resection. Postoperative analgesia was administered (meloxicam, 1 mg/kg, subcutaneously) for 3 days. No animals died or were excluded from the study due to surgical complications or other adverse events; all randomized subjects completed the entire experimental protocol and were included in the final analysis. Rats received intra-articular injections under isoflurane anesthesia according to their pre-assigned groups: the cartilage injury + AAV8-Arg-1 group received 20 μL of AAV8-CMV-rArg-1–3FLAG (1 × 10^10 vg/mL), the cartilage injury + AAV8-empty group received 20 μL of the control AAV8-empty vector, and the sham group received no injection, with all injections performed using a 29G insulin syringe precisely targeting the intra-articular space. Following the 5-week experimental period, rats were euthanized by cervical dislocation under deep isoflurane anesthesia for tissue collection. Knee joints were dissected and fixed in 4% paraformaldehyde (PFA) at 4°C for 48 hours. The fixed joints were decalcified in 10% EDTA (pH 7.4) for 4 weeks with constant agitation, with solution changes every 2–3 days until complete decalcification was confirmed by radiographic analysis. After thorough washing in running tap water for 24 hours, tissues were dehydrated through a graded ethanol series, cleared in xylene, and embedded in paraffin wax. Serial sagittal sections of 5 μm thickness were cut and mounted on poly-L-lysine coated slides. For histological evaluation, sections were stained with Hematoxylin and Eosin (H&E), Masson’s trichrome, and Safranin O-Fast Green using standardized protocols. Macroscopic cartilage integrity was evaluated by two independent, blinded observers using the International Cartilage Regeneration & Joint Preservation Society (ICRS) macroscopic scoring system (0 = normal, 1 = nearly normal, 2 = abnormal, 3 = severely abnormal, 4 = severely abnormal/exposed bone). The average score from the two observers was used for statistical analysis. Inter-observer agreement was excellent (Intraclass Correlation Coefficient, ICC = 0.89). SO histological scoring was performed by two independent, blinded observers who were unaware of the treatment group assignments. The final scores represent the average of the two observers’ assessments.

Isolation of chondrocytes and gene expression analysis by RT-qPCR

Primary chondrocytes were isolated from articular cartilage tissue obtained from three conditions. Briefly, the cartilage slices were minced and subjected to sequential enzymatic digestion using 0.2% collagenase type II in Dulbecco’s Modified Eagle Medium (DMEM) for 4–6 hours at 37°C with constant agitation. The resulting cell suspension was filtered through a 70-μm cell strainer to remove debris, and the chondrocytes were collected by centrifugation. Total RNA was extracted from chondrocytes using TRIzol reagent according to the manufacturer’s instructions. The RNA concentration and purity were determined by measuring the absorbance at 260 nm and 280 nm using a spectrophotometer. RNA samples with an A260/A280 ratio between 1.8 and 2.0 were considered of high quality and used for further analysis. First-strand cDNA was synthesized from 1 μg of total RNA using a Reverse Transcription Kit with oligo(dT) primers, following the standard protocol. The mRNA expression levels of specific macrophage polarization markers and the reference gene GAPDH were quantified by RT-qPCR using a SYBR Green PCR Master Mix on a real-time PCR detection system. The primer sequences for the target genes, including M1 markers (iNOS, TNF-α, CD86, IL-1β, CXCL10) and M2 markers (Arg-1, TGF-β, IL-10, CD206), as well as the housekeeping gene GAPDH, are listed in Table 1. The PCR amplification was performed under the following conditions: initial denaturation at 95°C for 30 seconds, followed by 40 cycles of 95°C for 5 seconds and 60°C for 30 seconds. A melting curve analysis was conducted at the end of each run to confirm the specificity of the amplification products. All reactions were performed in triplicate.

Table 1. Primer sequences used for RT-qPCR.

Gene Symbol	Forward Primer (5’-3’)	Reverse Primer (5’-3’)
Arg-1	GCTGGAGGACTTTAAGGGTTACCT	GGGAGAAATCGATGACAGCGTC
iNOS	GGAACCGCTACCTTGCTCAT	GCACATCAAAGCGGTCATCT
TNF-α	GGCGTGGAGCTGAGAGATAACC	CCACAAGAGGCAACCTGACC
CD86	CTGGGAAACGGGATAACGTG	TGTAGTGGGAGCAGCATGAG
IL-1β	CAGCAACAATTCCTGGCGATA	AAGGCGAAAGCCCTCAATTT
CXCL10	CAGGCAGGCAGTATCACTCATT	TGTCCTTGCTTGGTTCTCCTTG
TGF-β	GGAACCTGGTGGAATGTGACC	CAGGTAGTGGGCAGTGGTTTC
IL-10	GCTGCAAGCTGATCCAGATTC	ATGTAGGGAAGTGATGGGAGGT
CD206	GACCGCAACAACGCCATCTA	GGCAGTGGTTGAGCCGTAGT
GAPDH	AGGTCGGTGTGAACGGATTTG	TGTAGACCATGTAGTTGAGGTCA
Ccl-2	TAAAAACCTGGATCGGAACCAAA	GCATTAGCTTCAGATTTACGGGT

Statistical analysis

Statistical analyses were performed using GraphPad Prism (version 10) or R (version 4.5.1). Data are presented as mean ± SEM unless otherwise specified. For comparisons between two groups, Student’s unpaired two-tailed t-test was used. For multiple group comparisons, one-way or two-way ANOVA with Tukey’s tests was applied. The significance levels were denoted as follows: ns (not significant, p > 0.05), *p < 0.05, **p < 0.01.

Results and discussion

Single cell sequence reveals age- and cartilage injury- related immune cell dynamics in joint tissues

Single-cell RNA sequencing analysis of joint tissues from four experimental groups (10w-ctrl, 10w-cartilage injury, 95w-ctrl, 95w-cartilage injury) revealed distinct immune cell dynamics associated with aging and cartilage injury. Six clusters including macrophage, b cell, proliferation cells, endothelial, stem cells, mast cells were found in all samples (Fig 1A). Quantitative analysis of the percentage of cells captured per sample demonstrated that cartilage injury triggered a global increase in macrophage within the joint microenvironment in both young (10w-OA) and aged (95w-OA) animals compared to their respective uninjured controls (10w-Ctrl and 95w-Ctrl). But it did not show a statistically significant difference between the young and aged OA groups (Fig 1B). This indicated that the difference of phenotype between young and old OA animals are not attribute to the change in macrophage numbers, but may affected by the multi-function of macrophage subtypes. Notably, macrophages displayed functional heterogeneity, co-expressing homeostatic (P2ry12), pro-inflammatory (Il1r), and anti-inflammatory (Cd206) markers, indicating the presence of diverse subsets that warrant further subpopulation analysis (Fig 1E). Single-cell RNA sequencing analysis revealed a distinct population of cells exhibiting high expression of Ikzf3, Igkc, and Ms4a1 in osteoarthritic joint tissues. These markers are strongly associated with B cell lineage (Ms4a1/Cd20 being a canonical B cell marker, Igkc indicating immunoglobulin kappa light chain production, and Ikzf3/Aiolos regulating B cell development) (Fig 1C). A striking age-dependent difference was observed in proliferation-associated cells, which were markedly reduced in aged cartilage injury joints (Fig 1B). This population, characterized by high expression of proliferation-related proteins, may contribute to impaired tissue repair in older animals, potentially explaining their diminished regenerative capacity post-cartilage injury (Fig 1C. Therefore, the reduced recovery ability of old animal for cartilage injury may attribute to the impaired ability of cell proliferation.

Fig 1. Single-cell RNA sequencing analysis of immune cell populations under different conditions.

(A) Uniform Manifold Approximation and Projection (UMAP) plot depicting the global distribution and relative abundance of major immune cell types across four experimental groups: 95-week-old control (95w-ctrl), 95-week-old OA model (95w-OA), 10-week-old control (10w-ctrl), and 10-week-old OA model (10w-OA). Each point represents a single cell, colored by cell type. (B) Bar graph showing the proportional composition (percentage) of each immune cell type within each experimental group. (C) Heatmap showing the z-score normalized expression of top differentially expressed genes across six identified cell clusters. Rows represent genes, columns represent the four experimental groups. Data are derived from n = 3 biologically independent samples per group. Error indicates S.E.M, ns (not significant, p > 0.05), *p < 0.05, **p < 0.01. Statistic analysis was performed one‑way ANOVA with Tukey’s post‑hoc test.

Macrophage subpopulation heterogeneity and age-dependent plasticity

Single-cell RNA sequencing analysis identified four distinct macrophage subpopulations in joint tissues: homeostatic, pro-inflammation, anti-inflammation, and intermediate state (Fig 2A). Quantitative analysis revealed that macrophage percentage increased in 10w-cartilage after OA while no significant changed found in 95w-cartilage between control and OA tissues (Fig 2B). Homeostatic macrophages (marked by P2ry12) exhibited an age-dependent response to cartilage injury, no significantly ratio changes between old cartilage injury joints (95w-cartilage injury), but in young cartilage injury joints (10w-cartilage injury), homeostatic macrophage decreased after OA (Fig 2B). Notably, anti-inflammation macrophages were substantially more abundant in young cartilage injury joints compared to aged cartilage injury joints, suggesting superior inflammation relief microenvironment remodeling capacity in younger animals following injury (Fig 2B). Critical ECM-related and ROS inhibition genes (COL8A1, Arg-1) showed predominant expression in the anti-inflammation subpopulation of young cartilage injury joints (10w-cartilage injury), but were nearly absent in aged cartilage injury joints (95w-cartilage injury) (Fig 2C). The WNT_Score effectively captures distinct transcriptional states across the different cell populations. As illustrated, the score clearly separates macrophage, B cell, proliferation-associated cells, endothelial, stem cells, and mast cells into identifiable clusters. Furthermore, the scoring system also discriminates between functional polarization states—such as homeostatic, pro-inflammatory, anti-inflammatory, and intermediate phenotypes—within these cell types. This indicates that the differentially expressed genes used to compute the WNT_Score are robust markers that can reliably distinguish not only between major cell lineages but also between their activation or functional states. Thus, the WNT_Score serves as a valuable transcriptional signature for dissecting cellular heterogeneity in the studied context (S3 Fig). Pseudotime trajectory analysis (Fig 2D) reconstructed the macrophage differentiation pathway, revealing a sequential transition from homeostatic to intermediate state before bifurcating into either pro-inflammation or anti-inflammation subsets. The ordering of cells was arranged and the time sequency was presented in S4 Fig A. This analysis uncovered a striking age-related defect, where macrophages from aged cartilage injury joints failed to properly differentiate into anti-inflammation subsets, instead accumulating in homeostatic and intermediate states. Further batch analysis was done to confirm the consistence of pseudotime trajectory, similar trajectory was observed in all batches (S4 Fig B). These findings collectively demonstrate that while young animals can redirect homeostatic macrophages toward tissue-repairing subsets following cartilage injury induction, this critical plasticity is impaired in aged animals, potentially explaining their reduced capacity for tissue repair and more severe cartilage injury progression. The preserved homeostatic population in aged cartilage injury joints, coupled with deficient anti-inflammation macrophage generation, suggests a mechanistic basis for the poor regenerative outcomes observed in elderly cartilage injury patients.

Fig 2. Identification and characterization of macrophage subtypes.

(A) UMAP plot of all macrophages, sub-clustered to reveal four distinct subtypes (Homeostatic, pro-inflammation, anti-inflammation, intermediate state). (B) Bar graph presenting proportional ratios of each macrophage subtype across the four experimental conditions (95w-ctrl, 95w-OA, 10w-ctrl, 10w-OA). n = 3 per group. (C) Violin plots displaying the expression levels of ECM related marker genes defining homeostatic, pro-inflammatory (M1), and anti-inflammatory (M2) macrophage subtypes. Kruskal–Wallis tests for each cell subset to compare expression levels across groups, followed by Dunn’s post‑hoc test with Benjamini–Hochberg correction for multiple comparisons. Bars are presented as median expression values with interquartile ranges. (D) Pseudotime trajectory analysis (Monocle3) inferring the potential differentiation directions among macrophage subtypes. Error indicates S.E.M, ns (not significant, p > 0.05), *p < 0.05, **p < 0.01. Statistic analysis was performed one‑way ANOVA with Tukey’s post‑hoc test.

GO enrichment and WGCNA reveals ECM remodeling pathways in cartilage injury pathogenesis

Gene Ontology (GO) enrichment analysis of differentially expressed genes revealed significant associations with immune reaction regulation and cellular migration processes, including micro-environment organization, external encapsulating structure organization, and regulation of angiogenesis. Key structural components such as revealing of immune reaction, and ECM formation were prominently enriched, alongside functional terms like glycosaminoglycan binding, integrin binding, and growth factor binding (Fig 3A). GO enrichment also presented the promotion of anti-inflammation functions including resolution of inflammation, leukocyte activation, T cell activation. Weighted Gene Co-expression Network Analysis (WGCNA) identified a highly correlated gene module (module eigengene > 0.8) comprising anti-inflammation genes, with Arg-1, Hp, Mmp8 serving as hub genes. (Fig 3B). The AUC histogram analysis of the GO-enriched genes reveals distinct distribution patterns across the different groups and cellular subsets. The genes were predominantly expressed within the 10w-OA animal group. Specifically, within the macrophage subpopulation, the anti-inflammatory (M2) macrophage subset also showed a primary concentration of these enriched genes in the 10w-OA group, with 472 cells identified with an AUC > 0.16 in this group, compared to 48 cells in the 95w-OA group meeting a higher threshold of AUC > 0.19 (S5 Fig). Further western blotting presented higher expression of Arg-1 in 10w-OA animals (Fig 3C, 3D). Notably, Arg-1 is a pivotal enzyme that drives a powerful anti-inflammatory response within the immune system, primarily in macrophages. These findings collectively highlight a functionally coherent network with Arg-1 as a central node bridging structural integrity and cellular signaling in the observed pathological context.

Fig 3. Functional enrichment and co-expression network analysis of differentially expressed genes (DEGs).

(A) Gene Ontology (GO) enrichment analysis of DEGs between aged (95w-OA) and young (10w-OA) OA groups. Terms are categorized into Biological Process (BP), Cellular Component (CC), and Molecular Function (MF). The top 10 significant terms (adjusted p-value < 0.05) per category are shown. (B) Weighted Gene Co-expression Network Analysis (WGCNA) network plot highlighting a key module (turquoise) strongly associated with the anti-inflammatory macrophage subtype. Arg-1 is identified as a central hub gene within this module. (C-D) Western blotting analysis showing Arg-1 protein expression in 10w-OA and 95w-OA animal cartilage (C), and (D) quantification of Arg-1 protein expression normalized by GAPDH according to the gray value of wb bands. n = 3. Error indicates S.E.M, ns (not significant, p > 0.05), *p < 0.05, **p < 0.01. Statistic analysis was performed Student’s t-test.

Design and evaluation of AAV8-mediated Arg-1 overexpression vehicle in vitro

The results demonstrated the successful design and functional evaluation of the Arg-1 up-regulation (Arg-1-UR) plasmid encapsulated in AAV8, as illustrated in Fig 4A, which was subsequently tested for its effects on chondrocyte viability and protein secretion. Cell viability assays revealed no significant cell death was found in AAV infection, as evidenced by live/dead staining and quantitative fluorescence intensity analysis (Fig 4B, 4C). Longitudinal CCK-8 assays further confirmed the sustained proliferative advantage of Arg-1-UR-treated cells over time, with high proliferation ability observed across multiple days, which indicated that Arg-1-UR did not affect the proliferation of chondrocyte (Fig 4D). Flow cytometry analysis presented a marked increase in Arg-1 positive cell ratio for Arg-1-UR-treated macrophage (60.3%) compared to both untreated (Ctrl, 30.2%) and AAV empty vector (Bank, 29.5%) controls (Fig 4E). Additionally, ELISA measurements of secreted Arg-1 protein revealed a time-dependent increase in the Arg-1-UR group, underscoring the plasmid’s efficacy in enhancing Arg-1 protein expression and extracellular matrix production (Fig 4F). Collectively, these data highlight the potential of AAV-mediated Arg-1 overexpression to promote Arg-1 protein active, providing a foundation for further animal experiment.

Fig 4. Construction and in vitro validation of an AAV-mediated Arg-1 overexpression vector.

(A) Schematic diagram illustrating the design, cloning strategy, and packaging process of the AAV-Arg-1 overexpression vector (Arg-1-UR). (B) Representative live/dead fluorescence images of macrophages under three conditions: untreated control (Ctrl), AAV-empty vector (Blank), and AAV-Arg-1 (Arg-1-UR). Live cells (green, Calcein-AM), dead cells (red, EthD-1). Scale bar: 200 µm. (C) Quantification of cell viability from (B), expressed as the ratio of live cell fluorescence intensity to total fluorescence intensity. n = 3 independent experiments. (D) CCK-8 assay measuring cell proliferation/viability at days 1, 4, and 7 post-transduction for the three conditions. n = 3. (E) Flow cytometry analysis quantifying the percentage of Arg-1-positive macrophages at day 4 post-transduction. n = 3. (F) ELISA measurement of secreted Arg-1 protein in cell culture supernatant from day 1 to day 7. n = 3. Data in C-F are presented as mean ± SEM. Statistical significance was determined by one-way ANOVA with Tukey’s post-hoc test. ns, not significant (p > 0.05); *p < 0.05; **p < 0.01.

Gene regulation analysis confirmed Arg-1 as a contributor in anti-inflammation of macrophage

We performed a fluorescent microsphere (GFP microsphere) phagocytosis assay using RAW264.7 macrophages treated with LPS (to induce an inflammatory state) then treated with Arg-1 siRNA (Arg-1-DR) and Arg-1 upregulation AAV8 (Arg-1-UR). Phagocytic capacity was quantified by flow cytometry. GFP positive macrophage which indicated the phagocytosis of microsphere decreased in Arg-1-UR condition compared to control and Arg-1-DR condition (Fig 5A). Consistent with the phenotypic data, qPCR analysis of polarization-associated genes demonstrated corresponding transcriptional changes. In the Arg-1-UR group, the expression of anti-inflammatory genes TGF-β and IL-10 was significantly upregulated, while the pro-inflammatory genes TNF-α and CCL2 were downregulated. In contrast, the Arg-1-DR group displayed an opposite transcriptional profile, characterized by a significant upregulation of TNF-α and CCL2 and a downregulation of TGF-β and IL-10 (Fig 5B). Taken together, these results demonstrate that Arg-1 functions as a critical molecular switch governing macrophage polarization. Its upregulation actively promotes an anti-inflammatory phenotype and suppresses a pro-inflammatory state, whereas its downregulation drives macrophages towards a pro-inflammatory phenotype while inhibiting anti-inflammatory differentiation.

Fig 5. Flow cytometry and qPCR analysis reveals Arg-1 regulate macrophage functions.

(A) Flow cytometry analysis estimated the phagocytosis function of macrophage under three conditions. Raw264.7 was stimulated with LPS (control) and treated with Arg-1 siRNA (Arg-1-DR) and Arg-1 upregulation AAV8 (Arg-1-UR). Microsphere with GFP fluorescent co-cultured with treated macrophage before flowcytometry analysis. n = 3. (B) qPCR analysis presenting gene expression of anti-inflammation markers Tgf-β, Il-10 and pro-inflammation markers Tnf-α and Ccl-2. n = 3. Error indicates S.E.M, ns (not significant, p > 0.05), *p < 0.05, **p < 0.01.

Overexpression of Arg-1 attenuates LPS-induced ROS production in chondrocytes

To investigate the role of Arg-1 in oxidative stress during chondrocyte inflammation, we assessed intracellular ROS levels following LPS challenge. As illustrated in Fig 5 A-B, LPS stimulation significantly induced ROS generation in chondrocytes, as visualized by fluorescent staining. However, this effect was markedly suppressed in cells transduced with Arg-1-UR to overexpress Arg-1. Quantitative analysis, measured by the gray value of ROS-positive cells, confirmed a significant reduction in ROS levels in the Arg-1-UR group compared to both the control (Ctrl) and viral empty vector (Blank) groups. These results indicate that the specific upregulation of Arg-1 expression effectively mitigates LPS-induced oxidative stress in chondrocytes. This finding suggests that Arg-1 plays a protective role in the inflammatory response of chondrocytes by scavenging reactive oxygen species. Based on the Western blot results for phosphorylated P65 (pho-P65), a key marker in the NF-κB signaling pathway, no significant difference in expression was observed between the LPS-stimulated Control group and the Blank vector group. However, the Arg-1 upregulation group (Arg-1-UR) showed a marked reduction in p-P65 levels (Fig 6C), a finding that was further supported by densitometric quantification (Fig 6D). Furthermore, flow cytometry analysis for inducible nitric oxide synthase (iNOS), a downstream effector of pro-inflammatory signaling, revealed a significant decrease in the proportion of iNOS-positive cells within the Arg-1-UR group compared to the Control group (Fig 6E). Taken together, these results indicate that Arg-1 upregulation not only attenuates the activation of the NF-κB pathway, as evidenced by reduced p-P65, but also suppresses the functional expression of the pro-inflammatory mediator iNOS in macrophages.

Fig 6. Overexpression of Arg-1 in macrophages inhibits ROS production and NF-κB signaling in co-cultured chondrocytes.

(A) Representative immunofluorescence images of chondrocytes co-cultured with conditioned medium from the three macrophage groups (Ctrl, Blank, Arg-1-UR) and stained for ROS (DCFH-DA, green) and nuclei (DAPI, blue). Scale bar: 50 µm. (B) Quantification of ROS levels in chondrocytes from (A), measured as mean fluorescence intensity (MFI) of DCF. n = 3 independent experiments. (C) Western blot analysis of phospho-P65 (p-P65) and total P65 protein levels in chondrocyte lysates, indicating NF-κB pathway activity. (D) Densitometric quantification of p-P65 protein levels normalized to total P65 from (C). n = 3. (E) Flow cytometry analysis quantifying the percentage of iNOS-positive RAW 264.7 cells after LPS stimulation and treatment with conditioned medium from the three macrophage groups. n = 3. Data in B, D, E are presented as mean ± SEM. Statistical significance was determined by one-way ANOVA with Tukey’s post-hoc test. ns, not significant (p > 0.05); *p < 0.05; **p < 0.01.

Efficient transduction of Arg-1 upregulation vehicle in joint tissues

The immunofluorescence results demonstrated efficient transduction of the Arg-1-overexpressing AAV vector in joint tissues, as evidenced by GFP expression serving as a reporter for viral infection. Observation of the merged DAPI/GFP images revealed higher GFP+ cell numbers in Arg-1-UR groups compared to blank and control groups, confirming successful AAV-mediated gene delivery. The GFP signal showed characteristic chondrocyte-specific localization patterns, with intense fluorescence in Arg-1-UR samples indicating robust transgene expression (Fig 7A). These results validate the experimental system’s reliability for Arg-1 overexpression studies while establishing GFP as a sensitive marker for monitoring AAV infection. Western blot analysis revealed a significant upregulation of Arg-1 protein expression in the Arg-1-UR group compared to both control and blank groups (p < 0.01) (Fig 7B). Densitometric quantification demonstrated that Arg-1-UR samples exhibited a 1.5 ± 0.4-fold increase in Arg-1 protein levels relative to control (set as 1.0), while blank group expression remained comparable to control (1.6 ± 0.1-fold). The immunoblot showed a distinct band at the expected molecular weight (~40 kDa) in all groups, with markedly greater intensity in Arg-1-UR samples. These results confirm successful Arg-1 overexpression mediated by AAV transduction in tissue. The Cyclophilin B loading control (21 kDa) demonstrated equal protein loading across all lanes, validating the quantitative comparisons (Fig 7C).

Fig 7. In vivo delivery and expression of the AAV-Arg-1 vector in mouse cartilage tissue.

(A) Representative fluorescence microscopy images of articular cartilage sections from OA model mice injected with PBS (Ctrl), AAV-empty (Blank), or AAV-Arg-1 (Arg-1-UR). Nuclei are stained with DAPI (blue). GFP signal (green) indicates Arg-1 expression from the vector. Scale bar: 200 µm. (B) Western blot analysis of Arg-1 protein levels in whole knee joint protein extracts from the three treatment groups. (C) Densitometric quantification of Arg-1 protein levels from (B), normalized to β-actin. n = 3 mice per group. Data in C are presented as mean ± SEM. Statistical significance was determined by one-way ANOVA with Tukey’s post-hoc test. ns, not significant (p > 0.05); *p < 0.05; **p < 0.01.

Histological evidence of Arg-1’s role in cartilage injury progression

The results from the rat OA model, as illustrated in Fig 8, demonstrate the therapeutic efficacy of intra-articular Arg-1-UR gene delivery. Following the established experimental timeline of model induction, injection at day 0, and analysis at 5 weeks (Fig 8A), macroscopic evaluation using the ICRS scoring system revealed a significant improvement in cartilage integrity in the Arg-1-UR treated group compared to both the injury-only Control and the AAV vector-injected Blank groups (Fig 8B). Histological assessment via H&E, Masson’s trichrome, and Safranin O/Fast Green staining corroborated this finding, showing that the Arg-1-UR group possessed more abundant cartilage tissue with superior structural preservation relative to the severe degradation observed in the control cohorts. The robust, organized blue collagen network visualized by H&E and Masson’s trichrome staining confirms the restoration of ECM and the tensile structural framework. While the repaired tissue does not fully recapitulate the architecture of native hyaline cartilage (Fig 8C). Quantitative analysis of proteoglycan content, based on Safranin O staining, which specifically binds to proteoglycans, revealed a more intense and homogenous red signal in the Arg-1-UR group, indicating superior preservation of the cartilaginous matrix compared to the control groups where the signal was faint and patchy (Fig 8D). Furthermore, direct morphometric measurement of cartilage thickness from tissue sections confirmed a statistically significant increase in the Arg-1-UR treatment group compared to the Blank and Control groups (Fig 8E). The coordinated improvement across all these parameters—matrix composition, collagen organization, and macroscopic structure—demonstrates that Arg-1-UR gene therapy mitigates the degenerative cascade and promotes a functional repair of articular cartilage. Together, these data indicate that Arg-1-UR gene therapy effectively mitigates cartilage degeneration and promotes structural repair in the experimental OA model.

Fig 8. Evaluation of Arg-1 upregulation function after cartilage injury.

(A) Process of animal experiment. (B)International Cartilage Regeneration & Joint Preservation Society (ICRS) evaluate the severity of cartilage for conditions: control (injury only), Blank (injury with AAV vector injection), and Arg-1-UR (injury with Arg-1-UR vector injection). n = 3. (C) Histochemistry evaluation of tissue repairing in cartilage injury animal by H&E, Masson and SO staining for control (injury only), Blank (injury with AAV vector injection), and Arg-1-UR (injury with Arg-1-UR vector injection) conditions. (D) Quantification of grey value of Safranin O-fast green (SO) which indicating the cartilage volume in lesion for three conditions. n = 3. (E) Cartilage thickness quantified in SO evaluation presenting the repair of cartilage in lesion for three conditions. n = 3. Statistical significance was determined by one-way ANOVA with Tukey’s post-hoc test. ns, not significant (p > 0.05); *p < 0.05; **p < 0.01.

Overexpression of Arg-1 promotes a shift from M1 to M2 macrophage polarization

To evaluate the effect of Arg-1 overexpression on macrophage polarization, we measured the expression of classic M1 and M2 marker genes using RT-PCR. As shown in the bar graph, transduction with Arg-1-UR significantly altered the macrophage phenotype compared to the Blank (viral empty vector) and PBS control groups. The expression of key M1 pro-inflammatory markers, including iNOS, TNF-α, CD86, IL-1β, and CXCL10, was significantly downregulated in the Arg-1-UR group. Conversely, the expression of characteristic M2 anti-inflammatory markers was markedly upregulated. This included a substantial increase in the expression of Arg-1 itself, along with TGF-β, IL-10, and CD206 (Fig 9). Statistical analysis confirmed that these changes in both M1 and M2 gene expression profiles were significant. These results demonstrate that targeted Arg-1 expression drives macrophage polarization towards an anti-inflammatory M2 state, while simultaneously suppressing the pro-inflammatory M1 phenotype.

Fig 9. Arg-1 overexpression modulates the balance of macrophage polarization markers in injured cartilage.

Quantitative RT-PCR analysis of the relative mRNA expression of key M1-type (iNos, Tnf-α, Cd86, Il-1β, Cxcl-10) and M2-type (Arg-1, Tgf-β, Il-10, Cd206) macrophage markers in cartilage tissue harvested from OA model mice treated with PBS (Ctrl), AAV-empty (Blank), or AAV-Arg-1 (Arg-1-UR). Gene expression was normalized to Gapdh and is presented relative to the Ctrl group. n = 3 mice per group. Data are presented as mean ± SEM. Statistical significance was determined by one-way ANOVA with Tukey’s post-hoc test. ns, not significant (p > 0.05); *p < 0.05; **p < 0.01.

Discussion

The polarization of macrophages plays a pivotal role in determining the success of cartilage regeneration, primarily through the establishment of a balanced articular microenvironment. A young systemic environment promotes chondrocyte proliferation and cartilage matrix synthesis in old mice [23]. Pro-inflammatory M1 macrophages, activated by stimuli such as IFN-γ, LPS, and TNF-α, exacerbate osteoarthritis (OA) progression by secreting catabolic mediators including TNF-α, IL-1β, IL-6, and matrix-degrading enzymes like MMP-1, MMP-3, MMP-9, and MMP-13. These factors collectively inhibit chondrocyte proliferation, suppress extracellular matrix (ECM) synthesis, and impair the chondrogenic differentiation of mesenchymal stem cells (MSCs). In contrast, alternatively activated M2 macrophages, induced by IL-4, IL-13, IL-10, or interactions with MSCs, promote an anti-inflammatory and pro-chondrogenic milieu. M2 macrophages secrete regulatory cytokines such as IL-10, IL-1RA, and TGF-β, which are critical for tissue repair, ECM stabilization, and chondrogenesis [24].The plasticity of macrophages allows for therapeutic manipulation toward an M2-dominant phenotype, which supports cartilage repair. For instance, type II collagen has been shown to induce M2 polarization, leading to increased TGF-β production and reduced chondrocyte apoptosis and MMP-13 expression in OA models [25,26]. Furthermore, MSCs contribute to cartilage regeneration not only through direct differentiation but also via immunomodulation, skewing macrophage polarization toward the M2 phenotype through paracrine factors [27].

The role of Arginase-1 (Arg-1) in macrophage polarization and immune regulation remains a subject of intense investigation, with emerging evidence highlighting both its enzymatic and non-enzymatic functions across different physiological and pathological contexts [28]. Traditionally, Arg-1 is recognized as a hallmark of M2-like macrophages, where it competes with inducible nitric oxide synthase (iNOS) for the common substrate L-arginine, thereby reducing nitric oxide (NO) production and contributing to the resolution of inflammation and tissue repair [29]. However, recent studies challenge the simplistic view of Arg-1 as a universally immunosuppressive molecule, particularly in human systems. In murine models, Arg-1 expression in macrophages is often associated with alternative (M2) activation, induced by signals such as IL-4 or IL-10. This polarization promotes tissue remodeling and suppresses pro-inflammatory responses, partly through metabolic reprogramming that depletes L-arginine and impairs T cell function [30]. Despite these findings, the functional impact of Arg-1 appears to be context-dependent and species-specific. A recent study using human THP-1 monocytes demonstrated that stable overexpression of Arg-1 did not suppress LPS-induced inflammation, NF-κB/MAPK signaling, or M1 macrophage polarization [31,32]. Beyond its role in myeloid cells, Arg-1 also operates as an intrinsic metabolic checkpoint in other immune cells. For example, in CD4 + T cells, Arg-1 deficiency accelerates Th1 response kinetics and reduces lung pathology during influenza infection, indicating that T cell-intrinsic Arg-1 acts as a rheostat for Th1 life cycle and associated immunopathology [30]. These studies underscore the cell-type-specific functions of Arg-1 and its broader impact on immune response coordination.

Arginase-1 (Arg-1) exerts a profound influence on extracellular matrix (ECM) formation and cartilage repair, positioning it as a critical molecular nexus linking macrophage immunometabolism to tissue remodeling. Concurrently, the shift in macrophage metabolism driven by Arg-1 activity—diverting L-arginine away from nitric oxide (NO) synthesis and towards the production of ornithine, proline, and polyamines—provides the fundamental biosynthetic building blocks for collagen synthesis and cellular proliferation [33,34]. Proline serves as a direct precursor for collagen hydroxylation and stabilization, while polyamines regulate gene expression and protein synthesis critical for tissue growth [35]. Furthermore, by dampening the iNOS-NO pathway and suppressing reactive oxygen species (ROS) generation, Arg-1 mitigates oxidative stress-induced ECM degradation, matrix metalloproteinase (MMP) activation, and chondrocyte apoptosis [36]. This dual mechanism—simultaneously fueling anabolic pathways through metabolic reprogramming and shielding the nascent matrix from inflammatory catabolism—enables Arg-1 to break the self-perpetuating cycle of inflammation and tissue breakdown [33]. Consequently, Arg-1 does not merely act as an anti-inflammatory enzyme but functions as a central metabolic switch that actively redirects cellular resources towards reparative processes, thereby explaining its capacity to partially rescue the structural deficits observed in aged cartilage by enhancing the quality, stability, and synthesis of the cartilaginous ECM.

Based on these compelling findings, future research should systematically build upon the identified role of Arg-1 to translate this therapeutic promise into a clinically viable strategy. A key immediate step is to conduct longitudinal studies in aged animal models of osteoarthritis to determine the durability of the cartilage-protective effects of AAV8-mediated Arg-1 overexpression and to establish the optimal therapeutic window for intervention. Furthermore, it is crucial to dissect the precise cellular mechanisms by which Arg-1 orchestrates its effects, specifically investigating its direct impact on chondrocyte metabolism and its paracrine signaling role in modulating the broader joint immune environment, particularly the recruitment and polarization of other immune cells. Finally, given the translational imperative, future work must rigorously explore safe and effective delivery methods for Arg-1 in humans, such as exploring cell-specific promoters or non-viral vectors, and assess potential off-target effects, thereby paving the way for a novel immunomodulatory therapy to promote regeneration in aged and osteoarthritic joints.

Conclusions

In recent study, a young systemic environment promotes chondrocyte proliferation and cartilage matrix synthesis in old mice. This study reveals key age-related differences in joint tissue responses to cartilage injury through single-cell RNA sequencing. Aged joints showed reduced cellularity and impaired macrophage plasticity compared to young joints, failing to properly differentiate into anti-inflammation subsets following cartilage injury induction. WGCNA identified Arg-1 as a central regulator of inflammation regulation, ECM organization and cell adhesion networks. Successful AAV8-mediated Arg-1 overexpression increased protein levels 3.2-fold, with histological evidence demonstrating its protective effects on cartilage integrity. These findings highlight that age-related impairments in macrophage differentiation and progenitor cell proliferation contribute to poor cartilage injury outcomes, while Arg-1 overexpression represents a promising therapeutic approach to revealing immune reaction and tissue repair in aged joints.

Supporting information

S1 Fig. Predicted doublets by scublet function.

(JPG)

S2 Fig. Confirmation of AAV vector purity and transfect efficiency.

(A) AAV vector purity confirmed by silver staining, three bands presenting VP1, VP2, VP3 capsid protein. n = 3. (B) Flow cytometry quantified GFP positive cells in vector transfect macrophage.

(JPG)

S3 Fig. Addmoduluscores demonstrate that differences in different cell types (left) and macrophage subset (right) across conditions.

(JPG)

S4 Fig. Confirmation of cell order timeline and consistency in pseudotime analysis.

A. Cells ordering by monocle and presenting differentiation root. B. pseudotime analysi of different batches (randomly sampling of 80% data) to confirm the consistency of pseudotime.

S5 Fig. AUC score of GO enriched genes in 10w and 95w condition for scRNA analysis.

(JPG)

S6 Fig. H&E stain of uninjured rat cartilage.

(JPG)

S7 Fig. Raw images for western blotting.

(JPG)

Data Availability

Raw sequencing data were obtained from GEO dataset (GSE236843).

Funding Statement

The author(s) received no specific funding for this work.