Identification and characterization of POSTN+ synovial progenitor cells in meniscus regeneration Authors: Wanting Yan, Xiaojia Huang, Jingsong Wang, Wennan Xu, Haozhi Zhang, Jiahao Zhu, Shuyang Guo, Weiping Li, Wei Zhao, Zhengzheng Zhang Published: 2026 ABSTRACT AbstractBackgroundThe meniscus exhibits a limited intrinsic self-repair capacity, particularly in its avascular inner region. A major obstacle in developing regenerative therapies is the lack of understanding at the cellular or molecular level regarding its repair and regeneration. This study aimed to characterize the critical cellular population that drives the meniscal healing process.MethodsWe performed single-cell RNA sequencing (scRNA-seq) on human clinical meniscus samples and established a meniscal injury model in New Zealand white rabbits for parallel analysis. Computational analyses, including pseudotime trajectory inference, were conducted to delineate cellular states and transitions. To dynamically trace cell fate, lineage tracing was employed. Functional validation was achieved through the isolation, genetic manipulation, and transplantation of the candidate progenitor cells into injury models in vivo.ResultsscRNA-seq revealed that a population of synovium-derived fibroblasts, characterised by high periostin expression, represents a unique injury-responsive population. Upon meniscal injury, these cells proliferated and rapidly migrated to the lesion site. Both pseudotime analysis and lineage tracing confirmed that this population transitioned into a chondrogenic state during the later stages of meniscal healing, showing progenitor characteristics; hence, they were termed POSTN+ synovial progenitor cells (POSTN+ SPCs). Mechanistically, the transcription factor TCF7L1 was identified as a candidate regulator of their progenitor characteristics, and may involve Wnt/β-catenin signaling pathway. Transplantation of POSTN+ SPC-enriched cells enhanced meniscus healing in the animal model.ConclusionThis study identifies POSTN + SPCs as a candidate injury-responsive progenitor-like cell population that plays an important role in meniscus repair. These cells undergo dynamic activation, proliferation, and chondrogenic differentiation at the lesion site, providing a promising cellular target for functional meniscal regeneration.The translational potential of this articleOur findings suggest that the recruitment or transplantation of POSTN + SPCs may hold promise as novel therapies to enhance functional meniscal regeneration.Keywords: Lineage tracing, Meniscus regeneration, POSTN+ SPCs, Single-cell RNA sequencing, TCF7L1, Wnt signaling Introduction The meniscus is a crescent-shaped structure within the synovial cavity of the knee joint, situated inferior to the femoral condyle and attached to the tibial plateau. It plays a vital role in load transmission, force dissipation, and joint stabilization [1]. The highly organized tissue structure of the meniscus confers its unique mechanical properties [2]. The outer region of the meniscus is predominantly composed of fibroblasts and is rich in type I collagen and fibronectin, providing tensile strength. This region is relatively vascularized and exhibits a high capacity for repair. In contrast, the inner region of the meniscus features chondrocytes and their secreted products—type II collagen and proteoglycans (containing GAGs)—which collectively form a hydrophilic and elastic fibrocartilaginous structure [3]. This structure effectively withstands and dissipates compressive stress while providing lubrication and cushioning. Between these two lies a transitional zone, where the cellular and matrix composition is intermediate, facilitating a smooth gradation in mechanical properties. Meniscus injury is one of the most common knee injuries, particularly prevalent among individuals engaged in physical activities. Meniscal repair and partial meniscectomy remain the most common arthroscopic procedures, yet current surgical approaches fail to prevent the articular degenerative changes that follow meniscal injury, which often lead to osteoarthritis [[4], [5], [6]]. Although significant advances have been made in tissue engineering and biomaterials in recent years, and commercially available artificial meniscus grafts have been applied clinically, these methods cannot fully relieve post-meniscectomy symptoms or effectively prevent chondrocyte degeneration, nor can they reconstruct meniscal tissue or serve as long-term substitutes [[7], [8], [9]]. Furthermore, while many studies have explored the role of various stem/progenitor cell sources in promoting meniscus regeneration, the majority of this research has been conducted in vitro or in animal models [10]. The corresponding cell types in humans and their mechanisms of action remain unclear, and elucidating such knowledge could provide critical insights for improving current repair strategies. Identification and characterization of key cellular populations driving repair processes represents a critical step toward enhanced regeneration. Understanding their spontaneous repair process—particularly the origin of the progenitor cells involved, as well as the regulation and differentiation during the repair process—not only offers key insights into their regenerative potential but also provides important theoretical support for the study of regeneration in other fibrocartilaginous tissues. In this study, we investigated the cellular heterogeneity of the meniscus repair process, the origins of key cells, and their dynamic lineage changes during repair. Through single-cell RNA sequencing of clinical samples and a rabbit injury model, we discovered that POSTN+ SPCs constitute a dynamically regulated, injury-responsive population that expands and persists throughout the repair process. Furthermore, we explored the underlying mechanisms of injury-induced activation in POSTN+ SPCs and identified TCF7L1 as a candidate regulator of their progenitor function. During the repair process, TCF7L1 modulates the lineage transitions of POSTN+ SPCs via the Wnt signaling pathway, thereby promoting tissue repair and regeneration. Finally, we evaluated the therapeutic potential of POSTN+ SPCs in meniscus regeneration. By elucidating the critical role of POSTN+ SPCs in meniscus repair, we aim to provide insights into the cellular and molecular mechanisms of fibrocartilage repair and to support future cell-based therapeutic strategies for meniscus tissue regeneration. Materials and methods Experimental design This study aims to identify the origin, characteristics, and lineage dynamics of the key cells involved in the natural repair and regeneration of the meniscus, and to evaluate their therapeutic potential in meniscus regeneration. We examined (i) the activation and lineage contributions of POSTN+ cells in human, rabbit, and mouse injury models; (ii) their origin and progenitor properties; (iii) their ability to proliferate, migrate, and regenerate; and (iv) the regulatory role of TCF7L1/Wnt signaling. Functional efficacy was tested by transplanting POSTN+ SPCs into rabbit meniscus defects. Sample sizes were determined from pilot experiments and prior reports; details are provided in figure legends. A complete list of antibodies, reagents, instruments, and software used in this study is provided in Supplementary Table S1. Human tissue collection Meniscus biopsies and synovium were collected during arthroscopic meniscus repair as part of standard procedures. Samples were derived from radial tears in the posterior horn of the lateral meniscus, spanning the red-white transitional zone. Samples were transported in Hanks' Balanced Salt Solution (HBSS) on ice to a central laboratory for processing. Primary cell isolation Meniscus and synovial cells were dissected and minced into 1 mm3 pieces. The tissue was digested in 2 mg/mL collagenase type II (245 U/mL) in DMEM at 37 °C overnight. The dissociated cells were then cultured in growth medium containing DMEM, 10% fetal bovine serum (FBS), and 1% penicillin/streptomycin. The complete culture medium was changed every 3 days. Cells from passages 1 to 3 were used in vitro differentiation assays and transplantation into injured rabbits. Surgical models Rabbit Model: Skeletal-mature male New Zealand white rabbits (2.5 kg) were anesthetized via inhalation of isoflurane (2.5%) following induction with pentobarbital sodium (25 mg/kg). Preemptive analgesia was administered via subcutaneous injection of 3 mL 1% lidocaine. Under sterile conditions, a medial parapatellar arthrotomy was performed to access the knee joint. The patella was laterally subluxated, and the medial collateral ligament was incised to fully expose the medial meniscus. A 2 mm trephine was used to create a standard defect in the medial meniscus spanning the red-white zones in the anterior horn. After saline irrigation, the joint capsule and skin were closed in layers with sutures. Cell Transplantation: POSTN+ SPC-enriched cells (1 × 106 in 20 μl PBS) were injected into the right knee after meniscus injury. The left knee received the same volume of PBS as a control. Mouse Model: Mice were anesthetized with 1.25% Avertin (0.2 mL/10 g). Unilateral knee injury was surgically induced in the right limb (n = 5–7/group). Briefly, medial parapatellar arthrotomy was performed under a dissection microscope. A full-thickness punch defect was created in the anterior horn of the meniscus using a 25G needle. The patella was repositioned, and the incision was closed with sutures. Contralateral unoperated knees served as internal controls. Mice were euthanized at 5 days, 2 weeks, or 6 weeks post-injury via CO2 overdose followed by cervical dislocation. Knee joints were harvested for downstream analysis. No formal randomization procedure was performed. Primary cell culture The culture medium for tdTomato+ cells consisted of DMEM-F12 supplemented with 10% high-quality fetal bovine serum (FBS), 1% Non-Essential Amino Acids, 1% GlutaMAX Supplement, and 1% penicillin/streptomycin. The cells were maintained in a 37 °C incubator with 5% CO2. Freeze-thaw assay (tissue devitalization) To devitalize the synovium, a stainless-steel probe pre-cooled in liquid nitrogen was applied to the target tissue. The synovium underwent 20 freeze-thaw cycles. Each cycle involved 30-s contact with the probe, followed by thawing with room-temperature saline. This process induced tissue whitening due to heat conduction, ensuring complete decellularization of the targeted tissues [11]. Cell viability assay Cell viability was evaluated using a dual-staining approach with 4′,6-diamidino-2-phenylindole (DAPI, nuclear stain) and propidium iodide (PI, dead cell stain). Fresh tissue samples (negative controls) and freeze-thawed tissues were incubated with PI (1.5 mM, 15 min), followed by PBS washing and counterstaining with DAPI (10 min) in accordance with the manufacturer's protocol. Positive controls were generated by freezing tissue at −80 °C for 48 h. Stained samples were imaged using a confocal laser scanning microscope (Olympus FV3000), and viability was quantified as the percentage of PI-positive (non-viable) cells relative to DAPI-positive (total) cells. Tissue histologic analyses Tissues were harvested at 5 days, 2 weeks, 6 weeks post-surgery. Healing was evaluated using a semi-quantitative scoring system assessing surface integrity, tissue thickness, and host integration. After collection, tissues were fixed in 4% paraformaldehyde (PFA; 24 h, 4 °C) and processed for either paraffin or OCT embedding. For paraffin embedding, samples underwent graded ethanol dehydration (70–100%), xylene clearing, and paraffin infiltration, while OCT-embedded tissues were cryoprotected in 30% sucrose (PBS, 48 h, 4 °C), equilibrated in OCT compound (24 h, 4 °C), and snap-frozen at −80 °C. Paraffin-embedded sections (5 μm thickness) were stained with hematoxylin and eosin (H&E) according to manufacturer guidelines. For Safranin-O/Fast Green staining, sections were sequentially stained with 0.02% Fast Green (Sigma–Aldrich, 2 min) and 0.1% Safranin-O (Sigma–Aldrich, 8 min) at room temperature to evaluate proteoglycan content. Masson's Trichrome staining (Masson's Trichrome Staining Kit, Servicebio) and Picrosirius Red (Picro Sirius Red Stain Kit, PHYGENE) were performed to assess fiber growth and collagen deposition. Immunofluorescence staining and in situ hybridization Frozen sections (10 μm thickness) were fixed in 4% PFA for 15 min, permeabilized with 0.1% Triton X-100 in PBS (10 min, room temperature), and blocked with 5% bovine serum albumin (BSA; Sigma–Aldrich) in PBS (30 min, room temperature). Sections were incubated overnight at 4 °C with primary antibodies. After washing with PBS three times, slides were incubated with secondary antibodies in PBS for 1 h. Nuclei were counterstained with DAPI for 5 min. Immunofluorescent images were acquired using a confocal laser scanning microscope (Olympus FV3000) and analyzed with ImageJ software. Col2a1 RNA localization was performed using the PinpoRNATM Multiplex Fluorescent RNA In Situ Hybridization Kit (GD Pinpoease Biotech Co. Ltd.) with proprietary probe design (China Patent ZL202110581853.9) and signal amplification systems (China Patent ZL202110575231.5). Briefly, formalin-fixed sections were treated with pre-A solution to quench endogenous peroxidase activity, followed by controlled protease digestion for target RNA exposure. Sequence-specific probes were hybridized at 40 °C for 2 h, with subsequent signal amplification through three iterative reactions: (1) probe-target binding stabilization, (2) horseradish peroxidase (HRP)-conjugated amplifier deposition, and (3) tyramide signal amplification (TSA) using OpalTM 520 fluorophore (1:1000 dilution; Akoya Biosciences). Fluorescent signals were visualized at 520 nm emission wavelength, with nuclear counterstaining (DAPI) for spatial reference. The multiplex immunohistochemical staining (mIHC) was conducted following manufacturer guidelines (Chengdu Zen-bioscience Co., Ltd): Tissue sections were fixed and permeabilized using a membrane disruption solution. Antigen retrieval was performed by incubating slides in 0.4% pepsin solution at 37 °C for 20 min, followed by endogenous peroxidase blockade with 3% hydrogen peroxide. Non-specific binding sites were blocked with 5% bovine serum albumin (BSA), then sections were incubated with primary antibodies at 4 °C overnight. HRP-conjugated secondary antibodies were applied, followed by tyramide signal amplification using TSA-520/570/690 fluorescent dyes. Iterative multiplex detection was achieved through sequential antibody stripping with 37 °C elution buffer. Nuclei were counterstained with DAPI prior to mounting in antifade medium. Image acquisition was performed using laser-scanning confocal microscopy. Optical clearing of mouse joint samples Optically cleared samples were generated as previously described [12], briefly, mice were perfused with 5 mM EDTA/PBS and perfusion-fixed with 4% paraformaldehyde (PFA). Knee joints were dissected free of muscle tissue, post-fixed in 4% PFA, then decalcified by perfusion with EDTA decalcification solution (pH 7.2) for 4 days, followed by a 2-h wash with pure water (pH 7.0) to remove residual salts. Decolorization was achieved by perfusing 25% Quadrol solution for 24 h. Delipidation was performed sequentially using toluidine blue solutions at 30% v/v, 50% v/v, and 70% v/v concentrations, each infused for 24 h. Tissues were dehydrated in toluidine blue -PEG medium for 48 h with daily medium replacement, followed by final clearing in BB-PEG medium for 24 h until optical transparency was achieved. FACS analysis For isolation of synovial cells, refer to section 2.18 on single-cell dissociation. Cells were filtered through a 40 μm cell strainer. For surface marker staining, cells were stained for 30 min at 4 °C in the dark with the following fluorochrome-conjugated antibodies: BV421-conjugated anti-CD34 (BD Horizon, Cat# 562577, RRID: AB_2687922), FITC-conjugated anti-CD90 (BioLegend, Cat# 328107, RRID: AB_893438), and PerCP-Cy5.5-conjugated anti-CD146 (BD Pharmingen, Cat# 562134, RRID: AB_10896319). Following staining, cells were washed twice with cold PBS containing 1% FBS. Cell sorting and analysis were performed using a Beckman CytoFLEX SRT Cell Sorter (Beckman Coulter). Data acquisition and analysis were conducted using Beckman CytExpert Software. Cell counting Kit-8 (CCK-8) assay 3000 cells in 100 μl of medium were seeded in a 96-well plate. At the designated time points, the original medium was replaced with medium containing 10% CCK-8 reagent (Dojindo Molecular Technologies) and incubated at 37 °C for 2 h. Afterward, the absorbance at 450 nm was detected with a microplate reader. Colony formation assay To determine colony forming efficiency, 1000 cells were seeded in triplicate wells of six-well plates. After 12 days, culture medium was removed and colonies were fixed in 4% PFA for 10 min, then stained with 1% Crystal Violet Stain solution for 10 min. CellTracker labeling SPCs were labeled using CellTracker™ fluorescent probes (Invitrogen). Briefly, cells were harvested and plated at the desired density, and culture media was removed. Cells were incubated with CellTracker™ working solution (1:1000 dilution in serum-free medium) at 37 °C for 30 min. After staining, the labeling solution was removed, and cells were washed twice with PBS and returned to normal culture conditions prior to downstream applications. Tcf7l1 overexpression and knockdown experiments Envelope (pMD2.G) and packaging (psPAX2) plasmids were obtained from Addgene (MA, USA). The preparation and packaging of the lentiviral vector were completed by IGEbio (Guangzhou, China). According to the viral infection protocol provided by the company, the packaged pCDH-CMV-Mus-Tcf7l1-EF1α-copGFP-T2A-Puro virus solution was used to infect sorted Postn+ cells from adult mice at a multiplicity of infection (MOI) of 100, with an empty lentiviral vector as a control. Polybrene was added to 200 μl of culture medium at a final concentration of 5 μg/mL, and the cells were infected in suspension at room temperature for 15 min. Then, both the cells and the virus solution were transferred into a culture dish and further cultured. The positive rate of cells was observed at 48–96 h. For knockdown of Tcf7l1, the packaged pLKO.1-U6-scramble-EF1a-copGFP-T2A-puro virus solution was used to infect sorted Postn+ cells from adult mice at a multiplicity of infection (MOI) of 100, with an empty lentiviral vector as a control. Polybrene was added to 200 μl of culture medium at a final concentration of 5 μg/mL, and the cells were infected in suspension at room temperature for 15 min. Then, both the cells and the virus solution were transferred into a culture dish and further cultured. The positive rate of cells was observed at 48–96 h. Real-time qRT–PCR Total RNA was extracted using the EZ-press RNA Purification Kit (EZBioscience), and cDNA was synthesized with the Color Reverse Transcription Kit (EZBioscience) following the manufacturer's instructions. qRT–PCR was performed on a LightCycler 96 (Roche) using 2× SYBR Green Color qPCR Mix (EZBioscience). Gene expression was normalized to GAPDH and analyzed using the 2−ΔΔCt method, with technical duplicates for each biological replicate. Primer sequences are listed in the Key Resources Table. In vitro multipotential differentiation assays Cells were expanded in growth medium (basal medium supplemented with 10% FBS and 1% penicillin/streptomycin) until 80% confluency. Following trypsinization, cells were allocated to lineage-specific differentiation protocols using commercial induction media (Pricella, Wuhan, China) according to the manufacturer's instructions. For adipogenesis, cells were cultured in differentiation medium for two weeks, then fixed in 10% formaldehyde and stained with Oil Red O. For osteogenesis, cells were cultured in osteogenic differentiation medium for 21 days, fixed, and stained with 2% Alizarin Red S (pH = 4.2). For chondrogenesis, cells were resuspended in control medium and seeded as micromass in the center of a 24-well plate. After allowing cells to attach for 1 h at 37 °C, differentiation medium was added, containing appropriate growth factors and reagents as specified by the differentiation protocol. Medium was refreshed every other day, and after 14 days, micro micromass were either stained with alcian blue or used for RNA isolation. Single-cell dissociation Fresh tissues were immediately preserved in MACS Tissue Storage Solution (Miltenyi Biotec) and processed through sequential enzymatic digestion. Tissues were rinsed with RPMI 1640 medium, minced into <1 mm3 fragments, and subjected to two-step digestion: (1) primary digestion with 4 mg/mL pronase (Roche) and 50 U/mL DNase I (Worthington) at 37 °C for 30 min (120 RPM, gentle agitation every 10 min), terminated by RPMI 1640 addition and 40 μm filtration; (2) secondary digestion of residual fragments with RPMI 1640 containing 1 mg/mL Collagenase I/II (Worthington) and 50 U/mL DNase I. Combined cell suspensions were centrifuged (400 g, 5 min), followed by erythrocyte lysis using Miltenyi Biotec's buffer per protocol. Pelleted cells were resuspended in RPMI 1640/10% FBS, with viability (>85%) confirmed by AO/PI staining (Countstar Fluorescence Cell Analyzer). Final single-cell suspensions were enriched using MACS Dead Cell Removal Kit (Miltenyi Biotec) prior to sequencing. Library construction The scRNA-Seq libraries were generated using the 10X Genomics Chromium Controller Instrument and Chromium Single Cell 3’ V3.1 Reagent Kits (10X Genomics, Pleasanton, CA). Briefly, cells were concentrated to approximately 1000 cells/uL and loaded into each channel to generate single-cell Gel Bead-In-Emulsions (GEMs). After the RT step, GEMs were broken and barcoded-cDNA was purified and amplified. The amplified barcoded cDNA was fragmented, A-tailed, ligated with adaptors and index PCR amplified. The final libraries were quantified using the Qubit High Sensitivity DNA assay (Thermo Fisher Scientific) and the size distribution of the libraries were determined using a High Sensitivity DNA chip on a Bioanalyzer 2200 (Agilent). All libraries were sequenced by illumina sequencer (Illumina, San Diego, CA) on a 150 bp paired-end run. Processing of single-cell RNA-seq data Single-cell experiments and sequencing were performed by 10X Genomics Chromium platform. We captured approximately 7000 to 120,000 cells each sample. Raw reads were processed with fastp with default parameter filtering the adaptor sequence and removed the low-quality reads to achieve the clean data, and the clean reads were processed with the Cell Ranger pipeline (v7.1.0) for alignment (with reference genome OryCun2.0, Ensembl version 110), gene counting, generate feature-barcode matrices. Quality control After obtaining the digital feature-barcode matrices, the R toolkit Seurat (v4.4.0) was applied to perform quality control, dimension reduction, clustering and differential gene expression analysis. For quality control, only the genes with expression detected in a minimum of 10 cells were retained, and cells with less than 300 detected genes were excluded. Furthermore, cells with mitochondrial gene content >20% were classified as dead cells and excluded from downstream analysis. The R package ‘DoubletFinder’ (v2.0.3) was applied to each sample to predict doublets, which were subsequently filtered out. scRNA-seq data analysis Following doublets removal and normalization, high-quality cells were subjected to downstream analyses, including normalization, dimensional reduction, and cell type annotation. Gene expression matrices were normalized using Seurat's LogNormalize method (scale factor = 10,000), and 2000 highly variable genes were used for further analysis. Batch correction and integration of all samples were performed using ‘FindIntegrationAnchors’ and ‘IntegrateData’ functions in Seurat. Subsequently, principal component analysis (PCA) dimensionality reduction was performed on the scaled data by ‘RunPCA’. Uniform Manifold Approximation and Projection (UMAP) was performed to place cells in similar local neighborhood together in two-dimensional space. ‘FindNeighbors’ and ‘FindClusters’ function was used to merged substantively similar clusters. To identify markers enriched in each cluster, we used the ‘FindAllMarkers’ function with routine parameters (min.pct = 0.25, logfc.threshold = 0.5) and obtained a DEGs list. Cell clusters were annotated based on established marker genes for meniscus cell types from published literature [[13], [14], [15]]. Gene enrichment analysis Gene Ontology (GO) and Gene Set Enrichment Analysis (GSEA) analysis were performed by the ClusterProfiler package [16]. GO terms with Benjamini-Hochberg adjusted P value < 0.05 were significantly enriched. GO analysis was utilized to explore the potential functions and crucial pathways. The normalized gene expression matrix of human meniscus was used to computed enrichment scores of a certain gene set by single-sample gene set enrichment analysis (ssGSEA). p values of each gene set were computed by comparing our dataset to the permutated dataset. The heatmaps showed –log10(p values) of enriched gene sets using the ComplexHeatmap package. Cell lineage trajectory analysis To investigate the temporal progression and developmental trajectories of Postn+ cells, we performed pseudotime analysis using Monocle (v1.0.0), Monocle3 (v2.20.0) and cytoTRACE (v0.3.3). This analysis was conducted based on the expression patterns of highly variable genes that were identified by Seurat. Pseudotime represents a continuous trajectory that captures the inferred developmental progression of individual cells within a population. RNA velocity was performed to predicts the future state of POSTN+ cells on a timescale. For the compute of RNA velocity, a loom file of spliced and unspliced mRNAs from aligned bam file using velocyto package (v0.17.17) in Python (V3.8) was created. Then RNA velocity was computed based on both steady state and dynamic models using scVelo (v0.3.2) package. The scVelo algorithm used raw read counts from sequencing. Gene regulatory networks For single cell regulatory network analysis, the pySCENIC package (v0.12.0) was used with default settings to identify master regulator and transcriptional regulatory networks. Firstly, the GENIE3 module was used to identify target genes that were significantly co-expressed with a certain transcription factor. Secondary, the RcisTarget module was used to predict the target genes by screening the enriched cis-regulatory motifs of candidate transcription factors. Lastly, the AUCell algorithm was used to score the activity of each regulon. Definition of cell signatures To evaluate the potential functions of target cell populations in menisci from different repair timepoints, we calculated the scores for Chondrogenic (GO number: 0051216) and fibrogenic (GO number: 0043062) functional signature gene sets of the cell populations using the “AddModuleScore” function in Seurat. The specific approach involved subtracting the aggregated expression of control gene sets from the average expression levels of the functional signature gene sets. All analyzed genes were binned based on their average expression, and control features were randomly selected from each bin. The relevant gene sets are detailed in the Supplementary Material (Supplementary Table S1). This method quantifies the functional states of cell populations through gene expression signatures, providing a quantitative basis for subsequent biological interpretation. Quantification and statistical analysis Statistical analysis was performed with GraphPad Prism software (v9.3). Continuous data are expressed as mean ± standard deviation (SD) from ≥3 biologically independent replicates (individual data points shown in figures), with sample sizes empirically determined through preliminary experiments and annotated in figure legends. Group comparisons utilized: two-tailed unpaired Student's t-test for two groups; one-way ANOVA with Tukey's post hoc test (parametric) for multi-group comparisons. No explicit data exclusion criteria or randomization protocols were applied, as all experimental arms were processed in parallel under identical conditions. Lineage tracing For in vivo lineage tracing, Postn-Cre; Rosa26-LSL-tdTomato double-transgenic mice were used. Tamoxifen (MCE) was dissolved in corn oil at a concentration of 20 mg/mL. To induce Cre-mediated recombination, adult mice received intraperitoneal injections of tamoxifen at a dose of 75 mg/kg body weight once every other day for five consecutive days (3 doses over 5 days), followed by a 3-day wash-out period before surgery. To trace the fate of Postn+ cells during meniscal injury, mice were subjected to meniscal injury surgery immediately after the tamoxifen induction. At the end of the experiment, menisci and synovial tissues were harvested for cell sorting, histological imaging, and analysis. Result Single-cell transcriptomics reveals the activation of POSTN+ fibroblasts during meniscal injury repair To understand the cellular basis of meniscus repair, we collected tissue from a patient who underwent surgery two weeks after acute meniscus injury. Arthroscopy revealed that, although the meniscus surface was less smooth and mild irregularity persisted in the inner region, the overall shape of the meniscus was retained, indicating spontaneous healing (Fig. 1A, Fig. S1A). We performed single-cell RNA sequencing on repair tissue biopsy and integrated the results with published normal meniscus data [dataset PRJCA008120] [13]. Unsupervised clustering identified nine major cell populations, including chondrocytes, fibroblasts, myo-fibroblasts, mural cells, endothelial cells, macrophages, NKT cells, mast cells and B cells (Fig. 1B, Fig. S1B). While normal meniscus was predominantly composed of chondrocytes (51.4%), fibroblasts became the dominant cell type (15.5%) in repair tissue, increasing from 8.2% in normal meniscus (Fig. S1C and D). Activation of POSTN + FB induced by meniscal injury (A) Schematic of the single-cell RNA sequencing of human meniscal injury clinical samples. The tibial plateau was covered by the repaired medial meniscus located in the posterior horn (White arrow). Human meniscus was collected via arthroscopy through the transseptal portal. (B) UMAP plot of all cells integrated from the repair meniscus (2 weeks post-injury) and normal meniscus. (C) Subclustering of fibroblasts and chondrocytes identifies significant alterations in POSTN+ cells. (D) Bar plot showing cluster proportions in normal versus repair samples. (E) Heatmap of enriched biological processes in POSTN+ FB and dividing D-POSTN+ FB derived from native versus repair meniscus samples. (F) Immunofluorescence staining showing the localization of POSTN, CD34, and CD90 in human injured meniscus (related to Fig. S1G). (green arrow: CD34+ cells; red arrow: POSTN+ cells; white arrow: CD90+ cells; pink arrow: POSTN+ CD90+ cells; yellow arrow: CD34+ POSTN+ cells). Focused analysis of extracellular matrix (ECM)-producing cells revealed five chondrocyte subclusters: (i) SOX9+ CH1, (ii) MT+ CH2, (iii) ANGPTL7+ CH3, (iv) CHAD+ CH4, (v) CCL2+ CH5, as well as four fibroblast subclusters: (i) Lining cell (LC, PRG4+), (ii) CFD+ Fibroblast (CFD-FB, CFD+), (iii) POSTN+ Fibroblast (POSTN+ FB, POSTN+), (iv) Dividing POSTN+ FB (D-POSTN+ FB) (Fig. 1C, Fig. S1E and F). Further analysis of cellular composition revealed a dramatic shift between normal and repair meniscus. Notably, POSTN+ FB increased from 3.1% in normal meniscus to 44.2% in the repair meniscus, and a significant fraction of POSTN+ cells exhibited a dividing state, suggesting its proliferative response to injury (Fig. 1D). Gene enrichment analysis further indicated that POSTN+ cells were associated with matrix remodeling processes such as collagen fibril crosslinking and proteoglycan synthesis, highlighting their potential role in driving meniscus repair (Fig. 1E). Histologically, the injured meniscus showed a loss of the native, well-aligned collagen bundles and replacement by hypercellular, loosely organized repair tissue, and immunofluorescence staining demonstrated the presence of POSTN+ cells within this lesion site, some of which showed co-localization with CD90 and CD34 (Fig. 1F, Fig. S1G). Together, these findings indicate that this injury-responsive POSTN+ population exhibits fibroblast-like characteristics and may play a key role in matrix synthesis and tissue reconstruction. Injury model reveals synovial origin and persistent engagement of POSTN + fibroblasts To further characterize meniscus repair over time and to trace the origin of POSTN+ FB, we established a full-thickness medial meniscus injury model, and select different time points during spontaneous repair as observation points (Fig. 2A). Consistent with clinical findings, the rabbit meniscus is primarily composed of chondrocytes and fibroblasts, and also includes endothelial cells, muscle cells, macrophages, and monocytes (Fig. 2B and C; Fig. S2A–C). Quantitative analysis revealed an inversion of the chondrocyte-to-fibroblast ratio during repair: in the intact meniscus, chondrocytes accounted for 88.6% of the cell population and fibroblasts for only 0.78%. By 6 weeks post-injury, chondrocytes declined to 31.6%, while fibroblasts expanded to 61.4% (Fig. S2B). Under steady-state conditions, POSTN+ FB are predominantly located within the fibroblast population and are present in extremely low numbers; however, in injured menisci, POSTN+ FB undergo significant expansion (Fig. 2D). Immunofluorescence staining showed that these cells rapidly accumulated at the lesion site after meniscal injury (Fig. 2E). Histological analysis also revealed pronounced synovial hyperplasia and a fibrous tissue response adjacent to the defect, accompanied by an increase in POSTN+ FB (Fig. 2D, Fig. S2D). POSTN + FB originating from the synovium rapidly proliferate during meniscus repair (A) Schematic and macroscopic image of the rabbit meniscus injury model. (B) UMAP plot showing major cell types captured from rabbit synovium and meniscus. (C) Dot plots depicting signature gene expression of different cell populations. Color intensity represents average gene expression, dot size indicates the proportion of cells expressing the corresponding gene. (D) Time-series plot shows the dynamic changes of POSTN+ fibroblasts during meniscal injury repair in rabbits. (E) Immunofluorescence staining of POSTN in the lesion site at 5 dpi and 2 wpi. Green: CNMD or COL2A1, chondrocyte markers. (F) Schematic of rabbit meniscus injury model with or without synovium devitalization (−SYNO). Tissues were harvested at 5 days, 2 weeks and 6 weeks post-injury for analysis. (G) Macroscopic images (a–e) and corresponding H&E-stained sections of rabbit meniscus at different repair stages (uninjured, instant, 5 dpi, 2 wpi, 6 wpi; -SYNO 5 dpi, -SYNO 2 wpi, -SYNO 6 wpi). Scale bars: 2 mm (macroscopic images); 200 μm (sections). (H) Macroscopic meniscal repair scores at 6 weeks post-injury with or without synovectomy. To clarify whether synovial fibroblast-derived cells play a key role in meniscus repair, we applied a freeze–thaw protocol to selectively devitalize the neighboring synovium, a method previously reported to disrupt synovial cell viability in joint repair models (Fig. 2F) [11]. Compared to controls, devitalization of the synovium markedly impaired the repair response: fewer cells migrated into the defect area, cell proliferation was diminished (Fig. 2G, Fig. S2H). The histological healing scores were also lower in the devitalized group (Fig. S2I). Immunofluorescence staining showed that after removal of the synovial tissue, POSTN expression in the meniscus at 2 weeks post-injury was confined to the lateral edge and did not accumulate at the lesion site (Fig. S2J). These results suggest that the POSTN+ cells that participate in meniscus repair may originate from the synovium. Pseudotime and lineage tracing reveal progenitor features and repair potential of POSTN + FB To gain deeper insights into the changes of POSTN+ FB in meniscus repair, we performed re-clustering analysis on synovial fibroblasts. Re-clustering of fibroblasts at higher resolution identified seven distinct subpopulations (Fig. 3A). A proliferative subpopulation D-POSTN+ FB, characterized by high expression of cell cycle genes (MKI67, BIRC5) [17] and high G2M score, increased dramatically following injury (Fig. S3A and B). Immunofluorescence staining also confirmed their primary location in synovium (Fig. S3C). Gene module correlation analysis revealed strong similarity and close developmental relationships among the three POSTN+ subpopulations (Fig. S3D). Dynamic trajectory and chondrogenic potential of POSTN + FB in meniscus repair (A) UMAP plot of fibroblast subclusters from rabbit single-cell data (Fig. 2B). Three POSTN+ fibroblast subsets are highlighted: POSTN+ FB, D-POSTN+ FB, and POSTN+ COL2A1+ FB. (B) Violin plots of cartilage extracellular matrix (ECM)-related genes, demonstrating a distinct chondrogenic secretion profile of POSTN+ COL2A1+ FB. (C) Bar plots showing temporal distribution of POSTN+ fibroblast subsets across different time points and tissue origins. (D) UMAP projection of POSTN+ fibroblast subpopulations across uninjured and repair stages, colored by phenotypic states (left) and potency score predicted by CytoTRACE2 (right). (E) Experimental schematic of lineage tracing using Postn-creER; Rosa26-LSL-tdTomato mice. Tamoxifen was administered every other day for 5 days (3 doses over 5 days), followed by a 3-day wash-out period before surgery. Tissues were harvested at 5 dpi, 2 weeks post-injury (2 wpi), and 6 wpi. (F) Representative fluorescence images of knee sections showing tdTomato+ cells (Td+, red; marking POSTN lineage) at indicated time points. Nuclei were counterstained with DAPI (blue). Dashed lines indicate the injury site. Scale bars: 200 μm. (G) Quantification of tdTomato+ cell percentages at each time point. Data are shown as mean ± s.d.; n = 4 (Intact, 5 dpi); n = 5 (2 wpi, 6 wpi). One-way ANOVA with Tukey's multiple comparisons test was used to determine statistical significance. (H) Col2a1-RNA fluorescence in situ hybridization images of the injured meniscus on Postn-creER; Rosa26-LSL-tdTomato mice at 2 wpi. Scale bars, 200 μm. M: meniscus. Red: tdTomato; Green: Col2a1; Blue: DAPI. (I) Quantification of Col2a1+ cell percentages at each time point. Data are shown as mean ± s.d.; n = 4 (2 wpi, 6 wpi). One-way ANOVA with Tukey's multiple comparisons test was used to determine statistical significance. Surprisingly, in addition to the D-POSTN+ FB, a POSTN+ cell subpopulation (POSTN+ COL2A1+ FB) expressing the chondrocyte markers COL2A1, COL9A1, ACAN emerged during the late stage of meniscus repair (Fig. 3B; Fig. S3A), suggesting the generation of new chondrocytes in the later stages of repair. The predominant POSTN+ FB accounted for 17.2% of fibroblasts under uninjured conditions, and the proliferative subpopulation D-POSTN+ FB increased from only 2% in uninjured meniscus to 23.7% at 5 days post-injury, then declining to 13.3% at 2 weeks and 7.2% by 6 weeks, while the POSTN+ COL2A1+ FB subpopulation formed a distinct chondrogenic center at 2 and 6 weeks post injury (Fig. 3A and C). To further assess the progenitor characteristics of POSTN+ FB, we employed pseudotime analysis. CytoTRACE2 analysis predicted a surge in the cellular potency of POSTN+ cells at 5 days post-injury, which is consistent with the expansion of proliferative D-POSTN+ FB, followed by progressive differentiation at later stages (Fig. 3D). RNA velocity analysis also suggested that D-POSTN+ FB originate from POSTN+ FB, and further indicated directional transitions between POSTN+ FB and the POSTN+ COL2A1+ FB as repair progressed (Fig. S3D). Consistently, pseudotime trajectory reconstruction showed a continuous progression without major branching, with D-POSTN+ FB occupying early stages and POSTN+ COL2A1+ FB emerging later, while POSTN+ FB spanned the full trajectory (Fig. S3E). To investigate the origin and fate of POSTN + FB in vivo, we performed lineage tracing using Postn-creER; Rosa26-LSL-tdTomato mice. Mice were administered tamoxifen for five days, followed by a seven-day washout, and then underwent full-thickness meniscus injury (Fig. 3E). Under uninjured conditions, less than 5% of cells were tdTomato+, and these cells were sparsely distributed in the synovial sub-lining region adjacent to the meniscus (Fig. 3F, Fig. S3F). Three-dimensional clearing and whole-mount imaging revealed that these tdTomato+ cells were localized at the meniscus–synovium junction (Supplementary Movie 1). Injury triggered a marked synovial thickening and a rapid increase in tdTomato+ lineage cells, which expanded to approximately 20% of synovial cells by 5 days post-injury, with some cells recruited to the injury site. At 2 weeks post-injury, the synovium showed massive hyperplasia, with tdTomato+ lineage cells stacking in multiple layers and occupying much of the articular space (Fig. 3F); these cells were also observed at the meniscus defect margin, proliferative activity decreased at this point, as shown by reduced Birc5 staining (Fig. S3G). Immunofluorescent co-staining for BIRC5 indicated that about half of the tdTomato+ cells were actively proliferating at this stage (Fig. S3G and H). By 6 weeks, the wound area was mostly closed and synovial thickness had partially regressed, but the number of tdTomato+ cells remained higher than in the uninjured state (Fig. 3F). To examine their differentiation potential, we performed RNA fluorescence in situ hybridization (RNA-Scope) and observed that a subset (∼10%) of Postn-lineage cells expressed the chondrogenic gene Col2a1 at 2 and 6 weeks (Fig. 3H and I). Furthermore, we found that tdTomato+ cells and Col2a1 immunofluorescence staining colocalized in the injured region of the meniscus at 2 weeks post-injury (Fig. S3I). Collectively, these findings demonstrate that POSTN+ FB, originally located in the synovium adjacent to meniscus, rapidly expand and proliferate after injury, persist throughout the repair process, and possess in vivo chondrogenic differentiation potential, supporting their role as injury-responsive progenitors in meniscus regeneration. TCF7L1/WNT signaling drives the regenerative capacity of POSTN+ SPCs To explore transcriptional regulatory mechanisms underlying the phenotypic changes in POSTN+ SPC, we applied single-cell regulatory network inference (SCENIC) to analyze regulon motif enrichment in the POSTN+ populations and projected their expression dynamics along pseudotime (Fig. 4A, Fig. S4A). Regulons related to cell proliferation (CTCF, SOX5, HOXA5) exhibited early activation following injury, regulons involved in chondrogenic differentiation (SOX9, PPARG, SREBF2, RUNX2) were more prominent at later stages (Fig. 4A). These analyses identified potential regulators of cell state transitions, and the transitions involving proliferation followed by differentiation are consistent with the aforementioned results. Specificity score analysis ranked TCF7L1 as a top candidate regulon in all three POSTN+ subclusters at 2 wpi (Fig. 4B, Fig. S4B), with its activity peaking at this stage (Fig. 4C). We performed network analysis and identified ECM genes (POSTN, COL1A1, COL3A1, LOXL2), chondrogenesis-related genes (SOX9, COL2A1, COL11A1, COL9A1), and proliferation-associated genes (H2AZ2, PCNA, SMC2) as predicted targets of the WNT/TCF signaling pathway regulated by TCF7L1(Fig. 4D). What's more, GO analysis revealed that these regulators are enriched in signaling pathways related to cell fate commitment, chondrocyte differentiation, and Wnt/β-catenin signaling (Fig. 4E). TCF7L1/WNT signaling regulates the phenotype of POSTN+ SPCs (A) Heatmap showing transcription factor expression dynamics across pseudotime in POSTN+ clusters. (B) Specificity scores of regulons in POSTN+ Dividing FB and COL2A1+ FB subsets at 2 weeks post-injury. Top five regulons are highlighted in red. (C) Boxplots of TCF7L1 regulon activity (AUC scores) across indicated timepoints and POSTN+ fibroblast subsets. (D) Regulatory network of TCF7L1 and its predicted target genes identified via SCENIC analysis. (E) Gene Ontology (GO) enrichment analysis of regulons (TFs) in POSTN+ SPCs. Dot size reflects gene count; color indicates adjusted p-value. (F) Immunofluorescence staining of rabbit knee joint sections at 2 weeks post-injury, co-stained for TCF7L1. Scale bars, 200 μm. Red: tdTomato; green: TCF7L1; blue: DAPI. (G) Bar graphs showing TCF7L1 mRNA levels in rabbit POSTN+ SPCs at 1, 2, 4 weeks post-injury. Data are presented as mean ± s.d. (H) UMAP plot showing the integrated single-cell transcriptome of injured rabbit menisci from adult and juvenile (4-week-old) rabbits, highlighting the POSTN+ subpopulation. (I) Gene Ontology (GO) enrichment analysis of DEGs upregulated in POSTN+ SPCs from 4-weeks-old injured rabbit meniscus. Dot size reflects gene count; color indicates adjusted p-value. (J) Violin plot shows TCF7L1 expression in adult normal synovium (S) and meniscus (M), adult injured meniscus, and injured meniscus at 4 weeks of age (4wM, 5 dpi). (K) Immunofluorescence staining of knee joint sections from Postn-creER; Rosa26-LSL-tdTomato mice, co-stained for TCF7L1. Scale bars, 50 μm. Red: tdTomato; green: TCF7L1; blue: DAPI. TCF7L1, a key transcriptional mediator of Wnt/β-catenin signaling, is known to be highly expressed in multipotent embryonic progenitors, becomes downregulated during maturation [[18], [19], [20]]. Immunofluorescence staining showed that POSTN and TCF7L1, a key molecule of the WNT signaling pathway, colocalized in the injured region of rabbit meniscus at 2 weeks post-injury (Fig. 4F). RT-qPCR analysis of SPCs isolated from rabbits at different developmental stages revealed that TCF7L1 expression peaked at one week postnatally, then declined with maturation (Fig. 4G). We then investigate the injury response in younger individuals. Meniscal injury was induced in 4-week-old rabbits and single-cell RNA sequencing of meniscal tissue was performed at 5 days post-injury (Fig. 4H, Fig. S4C). In uninjured adult rabbit, TCF7L1 expression was undetectable in POSTN+ SPCs, but following meniscal injury, TCF7L1 expression was reactivated in POSTN+ SPCs (Fig. 4I). What's more, in young Rabbit after injury, the differentially expressed genes of POSTN+ SPCs are significantly enriched in biological processes such as the Wnt signaling pathway and extracellular matrix adhesion (Fig. 4J). This aligns with the signaling pathways indicated by the regulons (transcription factors, TFs) of adult POSTN+ SPCs. However, a notable distinction is that the differentially expressed genes (DEGs) of POSTN+ SPCs are most significantly enriched in signaling pathways related to extracellular matrix remodeling (Fig. S4E). Immunofluorescence staining also showed that TCF7L1 is widely expressed in the knee joints of young mice, and exhibits a co-expression pattern with POSTN (Fig. 4K), but it is rarely observed in adults. These findings suggest that the Wnt pathway transcription factor TCF7L1 is associated with the repair and regeneration capabilities of POSTN + SPCs. Transplantation of POSTN + SPC-enriched cells promote meniscus regeneration To assess the functional potential of POSTN+ SPCs, we first established a flow cytometry sorting strategy to enrich POSTN+ SPCs from synovium using a combination of surface markers informed by single-cell RNA sequencing (Fig. 1F, Fig. S1E). A panel containing CD31, CD90, and CD34 enabled separation of the cell population into three distinct compartments by flow cytometry (Fig. 5A). We performed primary culture of sorted CD90+ cells. After 48 h, immunofluorescence staining for CD90 and POSTN showed that the purity of CD90+ cells was about 90%, and the proportion of cells expressing POSTN was about 30% (Fig. 5B). Quantitative RT-PCR confirmed that these CD31− CD34− CD90+ SPCs (POSTN+ SPC-enriched cells, as the SPC group) expressed significantly higher levels of POSTN compared to other fibroblasts (O-FB group) (Fig. 5C). Notably, they also exhibited strong chondrogenic capacity, with enhanced Alcian blue staining and significant upregulation of SOX9, ACAN, and COL2A1 expression after induction (Fig. 5D and E). In vitro, these cells demonstrated progenitor characteristic, including multilineage differentiation capacity, as evidenced by positive Alizarin Red and Oil Red O staining following osteogenic and adipogenic induction, respectively (Fig. 5F). Transplantation of POSTN + SPC-enriched cells promotes meniscal matrix remodeling and cartilage regeneration (A) Flow cytometry gating strategy for isolating CD31− CD34− CD90+ SPCs from rabbit synovial tissue. (B) Immunofluorescence showing the expression of CD90 and POSTN, as well as cell purity analysis, in flow-sorted CD90+ cells. Red: POSTN; green: CD90; blue: DAPI. (C) RT-qPCR showing significantly higher POSTN expression in SPCs compared to O-FBs (n = 4). Data are presented as mean ± s.d.; two-sided unpaired t-test. (D) Alcian blue staining showing chondrogenic differentiation potential of SPCs. (E) RT-qPCR analysis of SOX9, ACAN, and COL2A1 expression after 7 days of induction. Data are mean ± s.d.; two-sided unpaired t-test. (F) Multilineage differentiation of SPCs in vitro. Above: osteogenesis assessed by Alizarin Red staining; Below: adipogenesis assessed by Oil Red O staining. Scale bars, 100 μm. (G) Experimental schematic of 2-weeks-old POSTN+ SPC-enriched cells transplantation into injured meniscus. (H) Macroscopic view and HE staining of meniscus tissue 5 days post-transplantation of POSTN+ SPC-enriched cells. Scale bars: 2 mm (macro), 200 μm (section). (I) Representative Picrosirius Red-stained sections of injured meniscus at 6 wpi after POSTN+ SPC-enriched cells transplantation. Boxed areas are shown at higher magnification. Scale bars: 100 μm. (J) Macroscopic and histological assessment (HE, Masson's trichrome, Safranin O staining) of meniscus repair at 6 weeks post-injury with or without 2-weeks-old POSTN+ SPC-enriched cells transplantation. Boxed regions are magnified. Scale bars: 2 mm (macro), 100 μm (sections). (K) Quantification of meniscal histological healing evaluation at 6 wpi. Data are presented as mean ± s.d. (n = 5). Two-sided unpaired t-test was used for statistical analysis. To evaluate the therapeutic potential of POSTN+ SPC-enriched cells, we isolated cells from 2-week-old rabbits, expanded them in vitro, and transplanted them into meniscal defects by intra-articular injection (Fig. 5G). At 5 dpi, POSTN+ SPC-enriched cells showed increased cellularity and a greater accumulation of fibrous matrix at the defect site compared to controls (Fig. 5H). In addition, Sirius Red staining highlighted a trend towards better collagen fiber ingrowth at the periphery of the repaired tissue in the 2-week POSTN+ SPC-enriched cells transplantation group (Fig. 5I), suggesting improved matrix integration. By 6 wpi, gross examination suggested improved continuity between the regenerated and native meniscus in animals treated with POSTN+ SPC-enriched cells, whereas the control group exhibited a more evident gap at the injury site (Fig. 5J). Histological assessment showed that the SPCs group developed a repair tissue with increased volume, denser matrix organization. These features were reflected in higher histological healing scores in the SPC-transplanted group (Fig. 5J and K). Both groups demonstrated positive safranin-O and Masson trichrome staining, indicating the presence of glycosaminoglycan-rich and cartilaginous extracellular matrix. However, we observed a vessel-like structure in the center of the control group, while the transplanted group displayed more uniform fibrocartilage-like matrix (Fig. 5J). Taken together, these results indicate that transplantation of POSTN+ SPC-enriched cells promotes cellular proliferation, matrix deposition, and structural preservation following meniscal injury in vivo. The same results were validated in mouse tdTomato+ cells. We directly sorted tamoxifen-induced tdTomato+ cells from the meniscus and synovium of adult mice (Fig. S5A). Immunofluorescence staining analysis showed that approximately more than 80% of the tdTomato+ cells expressed Postn (Fig. S5B). Overexpression of Tcf7l1 in primary cultured adult tdTomato + cells (Fig. S5C and D) led to a significant upregulation of Tcf7l1 at both the RNA and protein levels (Fig. S5E and F); more importantly, overexpression of Tcf7l1 upregulated the expression of the cartilage-related genes Col2a1 and Acan, as well as the proliferation-related gene Mki67. Conversely, knockdown of Tcf7l1 in sorted juvenile tdTomato + cells (Fig. S5G and H) that originally highly expressed Tcf7l1 reduced the expression of Col2a1 and Mki67(Fig. S5I), suggesting that Tcf7l1 may promote proliferation and chondrogenic differentiation. Discussion This study is the first to identify POSTN+ synovial progenitor cells as an injury-responsive cell population that plays a critical role in meniscus repair and regeneration. Through single-cell transcriptomic analysis of clinical samples and an animal model, we reveal that synovium-derived POSTN+ fibroblasts are activated upon injury, enter a proliferative and chondrogenic state, and rapidly accumulate at the repair site, supporting the regeneration process. Pseudotime analysis and lineage tracing jointly suggest that these cells exhibit progenitor characteristics, with the potential to respond to injury, proliferate, and differentiate. More importantly, we identify TCF7L1 as a potential transcription factor associated with the regenerative capacity of these candidate progenitor cells by activating the Wnt signaling pathway. In vivo transplantation experiments further demonstrate that POSTN + SPC-enriched cells significantly enhance meniscus healing. These findings identify a novel cellular source for meniscus repair and provide experimental evidence for developing cell-based or target-specific regenerative therapies (Fig. 6). Schematic illustration of the regenerative mechanism of POSTN + SPCs. The results of this study align with previous observations regarding the important role of the synovium in joint repair [11,21], yet advance the understanding to specific cellular subsets and molecular mechanisms. While prior research often broadly focused on “synovial cells” or “mesenchymal stem cells” [[22], [23], [24]], our work precisely identifies the POSTN+ fibroblast subset as the true functional effector cells. It is important to highlight that previous studies have identified potential stem/progenitor cells in meniscus and synovium, including Gli1+ and Gdf5+ cells [25,26], but our scRNAseq data from both meniscus and synovium did not reveal a clear enrichment of these markers within specific cell clusters. This discrepancy could be attributed to several factors: (1) technical limitations, as transcriptomic data only reflect mRNA levels, while some markers require protein-based validation; (2) progenitor cell markers may exhibit differential expression during development, homeostasis, and injury response; (3) species specificity, as the molecular characteristics of progenitor cells may differ between rodents and humans. In addition, studies have relied on ontogenetically identified progenitor lineage markers, such as Gdf5, to study progenitor cells [14,24,25]. These markers often show significant changes in their expression patterns in adult tissues, and in many cases, become undetectable in adult tissues [27]. The plasticity of cell markers also becomes evident in different disease states. For example, CD146 has been proposed as a progenitor marker in osteoarthritic meniscus by Sun et al. [15], but we observed that CD146 in healthy meniscus and synovium was primarily enriched in endothelial cell, consistent with observations from other researchers [13,28,29]. This difference may reflect the distinct cellular phenotypic changes and molecular responses between healthy tissue and degenerative conditions such as osteoarthritis. This also suggests that comparing cellular responses across different biological processes may offer new perspectives for disease research. Collins et al. has generated an insightful taxonomy of fibroblasts in the synovial joint [27]. By reviewing Collins et al.’s single-cell data, we found that a new POSTN+ cell population appears in synovium following cartilage injury, which suggests that POSTN+ cells may play a shared role in tissue repair across joint tissues. Additionally, POSTN encodes periostin, a matricellular protein that serves as a scaffold for extracellular matrix assembly [30,31]. Extensive research has demonstrated periostin's involvement in various physiological and pathological processes, including cardiac tissue remodeling and fibrosis [[32], [33], [34]], tendon and ligament development [[35], [36], [37]], bone formation [38,39], and skin fibrosis [40]. Notably, Ackerman et al. found that Periostin provides a critical niche for tendon healing, as POSTN+ cell depletion significantly impairs tendon repair [36]. While their study primarily focused on Periostin as a matrix protein, we demonstrate that POSTN+ cells and their progeny directly participate in meniscus repair and even regeneration. In our subsequent research, we will employ a conditional apoptosis induction model to investigate the impact of POSTN+ SPCs on the structure and function of meniscus tissue from a developmental perspective, thereby paving the way for novel endogenous cell-based approaches to achieve complete meniscus regeneration. The elucidation of TCF7L1's role as an effector of the Wnt/β-catenin signaling pathway reveals a critical molecular switch that maintains progenitor cell activity and guides their differentiation toward cartilage during repair. Previous studies have shown that the Wnt signaling pathway governs the occurrence, development, self-renewal, proliferation, and regeneration of stem cells [41]. Wu et al. found that neuropeptide Y (NPY) may significantly increase the number of cells in the S phase and G2/M phase by promoting the expression of related proteins such as β-catenin in the Wnt pathway, thereby enhancing the proliferation of rat bone marrow mesenchymal stem cells [42], our PPI regulatory network analysis also indicated that TCF7L1 can directly regulate highly expressed mitosis-related genes in D-POSTN+ FB cells (Fig. 4D). Existing studies have shown that TCF7L1 typically closely interacts with other members of the TCF transcription factor family, such as TCF7L2 and TCF4, to co-regulate the Wnt/β-catenin signaling pathway [18]. Studies have also demonstrated that TCF family transcription factors can either directly regulate the expression of the key chondrogenic transcription factor SOX9 [43] or cooperate with SOX9 to regulate the characteristics of cancer stem cells [44], while our PPI regulatory network analysis indicated that TCF7L1 can regulate the expression of chondrogenesis-related genes by modulating SOX9 through its co-regulatory factor TCF4 (Fig. 4D). It is noteworthy that our data indicate that the expression level of TCF7L1 in POSTN+ SPCs gradually decreases after adolescence with advancing age (Fig. 4G), whereas its expression in POSTN+ SPCs following meniscal injury is significantly higher in young mice than in adult mice, accompanied by marked enrichment of the WNT signaling pathway (Fig. 4I and J). This may explain why meniscus repair outcomes in adolescents are often better than in adults [45], and why endogenous repair in adults is often incomplete—possibly due to an insufficient number of such cells in the injury microenvironment or inadequate intensity and duration of the TCF7L1-mediated activation signal (Fig. 4A). Our transplantation experiment provides a preliminary proof of concept for this therapeutic approach, demonstrating a potential strategy for enhancing adult meniscus regeneration. However, because these cells were harvested from juvenile donors, the accelerated repair response may partly reflect the baseline regenerative capacity of young tissue rather than an exclusive TCF7L1-driven effect. Isolating the exact functional contribution of TCF7L1 signaling during adult in vivo chondrogenesis remains a priority for future evaluation. There are still some other limitations in this study. First, the single-cell RNA-seq data were derived from a single patient (N = 1). Although we have observed the accumulation of POSTN+ cells in the injured meniscus areas of humans, mice, and rabbits (Figs. 1F, 2E and 3F and 3H; Fig. S1G), patient-specific variations and batch effects cannot be entirely discounted, future studies will expand these cohorts to robustly delineate true biological insights from technical artifacts. Second, while the functional dynamics of POSTN+ SPCs were validated in animal models, their clinical translation requires further investigation. Third, complete cryoablation of the synovium is anatomically restricted by the joint cavity and risks independently disrupting the local intra-articular nutrient supply. However, despite these inherent technical boundaries, our data consistently demonstrate that devitalizing the bulk synovial tissue severely suppresses POSTN+ cell accumulation and hinders downstream morphological healing. This supports a synovium-derived regenerative response and highlights the importance for further investigating the embryonic mechanisms and the downstream targets of TCF7L1, and the specific role of Wnt signalling during chondrogenic differentiation. In summary, this study establishes POSTN+ synovial progenitor cells and their key regulator TCF7L1 as core components of endogenous meniscus repair. This discovery extends beyond traditional tissue engineering and simple cell transplantation strategies, pointing toward a new therapeutic paradigm of “enhancing endogenous repair”. Future research directions include: developing biologics to target and mobilize or expand a patient's own POSTN+ SPCs; exploring small-molecule drugs to activate the TCF7L1/Wnt signaling pathway and enhance repair capacity; and combining POSTN+ SPCs with biomaterials to construct smarter composite grafts. These strategies hold promise for ultimately achieving functional meniscus regeneration and preventing the onset of osteoarthritis. Data and materials availability Single-cell RNA-seq data of human have been deposited in the Genome Sequence Archive (Genomics, Proteomics & Bioinformatics 2021) in National Genomics Data Center (Nucleic Acids Res 2022), China National Center for Bioinformation/Beijing Institute of Genomics, Chinese Academy of Sciences (GSA-Human: blinded) that are publicly accessible at https://ngdc.cncb.ac.cn/gsa-human. The raw data of mouse single-cell sequencing are deposited in the NCBI SRA database under accession number PRJNA (blinded). The data will be fully open access in April 2027. This study analyzes existing, publicly available data. All accession numbers for the datasets are listed in Supplementary Table S1. No original code is reported in this paper. Any additional information required to reanalyze the data is available from the lead contact upon request. All data are available in the main text or the supplementary materials. Author contributions Data curation: HXJ (Xiaojia Huang), WTY (Wanting Yan) Formal analysis: WTY (Wanting Yan), XJH (Xiaojia Huang) Funding acquisition: ZZZ (Zhengzheng Zhang), WPL (Weiping Li) Investigation: WTY (Wanting Yan), HXJ (Xiaojia Huang), JHZ (Jiahao Zhu), HZZ (Haozhi Zhang), SYG (Shuyang Guo) Methodology: WTY (Wanting Yan), HXJ (Xiaojia Huang) Project administration: ZZZ (Zhengzheng Zhang) Resources: ZZZ (Zhengzheng Zhang), WPL (Weiping Li) Software: XJH (Xiaojia Huang), WTY (Wanting Yan) Supervision: ZZZ (Zhengzheng Zhang), WPL (Weiping Li) Validation: ZZZ (Zhengzheng Zhang), WNX (Wennan Xu), JSW (Jingsong Wang) Visualization: HXJ (Xiaojia Huang), WTY (Wanting Yan) Roles/Writing – original draft: WTY (Wanting Yan), HXJ (Xiaojia Huang) Writing – review & editing: ZZZ (Zhengzheng Zhang), WTY (Wanting Yan), XJH (Xiaojia Huang), WZ (Wei Zhao), WNX (Wennan Xu), JSW (Jingsong Wang) Ethics Declaration and Ethical Approval Human tissue studies were conducted with written informed consent obtained from all participants or their legal representatives in accordance with the Declaration of Helsinki and relevant regulatory frameworks, with protocol approval granted by the Ethics Committee of Sun Yat-sen Memorial Hospital (Approval No. SYSEC2-2024-BA-364). Human biological materials were not sourced from executed prisoners or prisoners of conscience. Animal experimental procedures were performed in compliance with the ARRIVE (Animal Research: Reporting of In Vivo Experiments) guidelines and were approved by the Institutional Animal Care and Use Committee of Sun Yat-sen University (Approval No. 2023003448). Animals were randomly assigned to experimental groups where applicable, though blinding was not performed. Declaration of generative AI in scientific writing During the preparation of this work the authors used DeepSeek in order to check the grammar and sentence structure of this manuscript. After using this tool, the authors reviewed and edited the content as needed and take full responsibility for the content of the published article. Funding This work was supported by National Natural Science Foundation of China (82472410), National Natural Science Foundation of China (82022046), Natural Science Foundation of Guangdong Province China (2025A1515011351). Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Acknowledgment statement We thank Prof. Xiaochun Bai and Prof. Linjia Jiang for critical reading of the manuscript. We also acknowledge the core facility at Sun Yat-sen Memorial Hospital for technical support. We thank the editors and anonymous reviewers for their insightful comments that improved the manuscript. Footnotes Supplementary data to this article can be found online at https://doi.org/10.1016/j.jot.2026.101191. Contributor Information Wanting Yan, Email: wanting.yan@postgrad.manchester.ac.uk. Xiaojia Huang, Email: huangxiaojia@syxjnyy3.wecom.work. Jingsong Wang, Email: wjsworkmail@163.com. Zhengzheng Zhang, Email: zhangzhzh9@mail.sysu.edu.cn. Supplementary data The following are the Supplementary data to this article: Multimedia component 1 Multimedia component 2 Multimedia component 3 References Associated Data Supplementary Materials Multimedia component 1 Multimedia component 2 Multimedia component 3