The treatment of acute tendon injury with small extracellular vesicles originating from TNFAIP6⁻ ADSCs subpopoulation both in vitro and in vivo

Abstract

Background

Small extracellular vesicles originating from adipose-derived mesenchymal stromal cells (ADSC-sEVs) have excellent therapeutic value in acute tendon injury. However, their mechanism and effects have not been fully elucidated. This study aimed to identify the key subsets and mechanisms of action of ADSC-sEVs involved in the repair of complete tendon tear caused by acute injury.

Methods

Based on our previous research demonstrating that ADSC-sEVs improve the quality of acute tendon injury repair, the present study utilized second-generation sequencing and bioinformatics to predict the key role of the TNFAIP6⁻ ADSC subgroup in acute tendon injury repair. We constructed different ADSC-sEVs through ADSC transfection and treated tendon stem cells for further exploration. EdU, cell scratch, and Transwell assays were used to evaluate cell proliferation and migration in vitro. Western blot and quantitative real-time polymerase chain reaction analyses were performed. Histopathological, immunohistochemical, and biomechanical testing were used for in vivo validation.

Results

TNFAIP6⁻ ADSC-sEVs significantly improved the therapeutic effect of ADSC-sEVs on acute tendon injury, which was related to the high expression of let-7c-5p. Application of different ADSC-sEVs in vitro and in vivo identified CRCT1/JAK2/STAT3 as a key downstream signaling pathway regulated by let-7c-5p.

Conclusions

Our findings enhance the current understanding of how TNFAIP6⁻ ADSC-sEVs exert healing properties in acute tendon injury through the let-7c-5p/CRCT1/JAK2/STAT3 signaling pathway. Furthermore, this study proposes a concept for constructing conditional ADSC-sEVs to enhance their inherent therapeutic effects.

Graphical Abstract

Supplementary Information

The online version contains supplementary material available at 10.1186/s13287-025-04789-2.

Introduction

Tendons are integral structures within the musculoskeletal system, playing a crucial role in maintaining and regulating the biomechanics of the body [1, 2]. The unique fiber structure of tendons contributes to their robust mechanical load performance [3]. However, certain sports activities or external forces can lead to acute tendon injuries, affecting approximately 30 million individuals annually and resulting in a substantial medical expenditure burden of up to $180 billion [4]. The management of complete tendon tear caused by acute injury has become a global health problem. Moreover, the sparse distribution of blood vessels and the low metabolic rate of cells in tendon tissues make restoring the original biological function of tendons difficult, rendering them susceptible to re-rupture even after healing [5]. Despite advancements in surgical techniques and rehabilitation methods for acute tendon injury, mitigating long-term complications and re-rupture remains challenging, leading to premature career terminations for many renowned athletes [6]. Consequently, there is a pressing need to develop high-quality repair methods after acute tendon injury.

Tendon stem cells (TSCs) are unique cells with a remarkable capacity for self-renewal and differentiation, playing a crucial role in tendon repair [7]. TSCs undergo extensive proliferation within a short period following tendon injury and are highly prone to tenogenic differentiation during the injury repair process [8]. Therefore, adjusting the activity of TSCs is anticipated to enhance tendon healing and minimize the need for surgery. Mesenchymal stromal cell (MSC) transplantation shows promise as a cutting-edge biological therapy, considering the ability of MSCs to regulate immune responses, improve the extracellular microenvironment, and activate endogenous stem cells [9–11]. For instance, Uysal et al. demonstrated that transplanting adipose-derived mesenchymal stromal cells (ADSCs) into injured tendons effectively promotes primary tendon repair [12]. However, limitations such as heterogeneity between cell populations, low cell retention, and ectopic osteogenesis persist in MSC transplantation [13, 14]. Consequently, small extracellular vesicles (sEVs), as the primary carriers of MSCs, have become a focal point in cell-free biotherapy research.

sEVs are membrane-bound extracellular vesicles facilitating cellular communication through processes such as endocytosis, membrane fusion, and receptor–ligand interactions [15]. MSC-sEVs exhibit therapeutic potential across several diseases [16–18]. However, MSCs consist of heterogeneous cell populations, and limited research has explored the functions and therapeutic effects of MSC-sEVs released by different MSC subsets. Owing to the challenges of restoring the normal strength and elasticity of tendons after acute injury, establishing efficient and high-quality methods to promote tendon healing is paramount. Considering the high specificity of tendon cell populations, complex repair process, and large demand for functional recovery, the advantages of cell-free regenerative medicine are increasingly prominent. Based on previous studies, we proposed for the first time to use sEVs derived from ADSC subsets to promote tendon healing. While further studying the therapeutic effect of ADSC-sEVs on acute tendon injury, we explored a novel method involving TNFAIP6⁻ ADSC-sEVs for tendon healing. Considering cellular heterogeneity, the present study constructed gelatin methacryloyl (GelMA)-loaded conditioned ADSC-sEVs to enhance the repair of acute tendon injury.

This study aimed to identify critical cell subsets and miRNAs in ADSC-sEVs involved in the healing of acute tendon injury. Through second-generation sequencing technology and bioinformatics analysis, we predicted and verified that TNFAIP6⁻ ADSC-sEVs regulate the biological characteristics of TSCs through the let-7c-5p-mediated CRCT1/JAK2/STAT3 signaling pathway. We believe the findings of this study may provide new ideas for cell-free regenerative medicine.

Materials and methods

Animals

A total of 108 male Sprague–Dawley (SD) rats (8–10 weeks old, 180–230 g) were sourced from the Animal Experimental Center of Harbin Medical University. The rats were fed an adequate diet and maintained on a regular 12-h light/dark cycle. This study followed the United States National Institutes of Health Guide for the Care and Use of Laboratory Animals and was approved by the ethics committee (No.ky2018-135).

Isolation and culture of ADSCs and TSCs

Following methods employed in previous studies, TSCs and ADSCs were isolated from the patellar tendon and inguinal adipose tissues of SD rats using collagenase type I (Sigma-Aldrich, St. Louis, MO, USA) digestion [19]. TSCs were cultured in Dulbecco’s Modified Eagle Medium (DMEM; Invitrogen, Carlsbad, CA, USA), and ADSCs were cultured in DMEM/F12 (Invitrogen). Each culture medium was supplemented with 10% fetal bovine serum (FBS) (Biological Industries, Kibbutz Beit-Haemek, Israel) and 1% penicillin-streptomycin (Beyotime, Shanghai, China). The cells were cultured in a cell incubator at 37 ℃ and 5% CO₂, and those from the Passage 3 to 5 were selected.

Isolation and identification of ADSC-sEVs

The isolation method for ADSC-sEVs mirrored that of a previous study [19]. Briefly, ADSCs were incubated in the small extracellular vesicle-free medium (Biological Industries) for 24 h, and ADSC-sEVs in the supernatant were isolated through differential centrifugation (300 g, 10 min; 3000 g, 10 min; 10,000 g, 30 min; 100,000 g, 2 h). ADSC-sEVs were identified using nanoparticle tracking analysis (NTA), transmission electron microscopy (TEM), and Western blotting. ADSC-sEVs were stained with the lipophilic membrane dye PKH26 (Sigma-Aldrich) for tracking experiments. GelMA loaded with 100 µg ADSC-sEVs was immersed in phosphate-buffered saline (PBS) for ADSC-sEVs release analysis. The supernatant was collected every 24 h, and the extent of release was evaluated using a BCA protein assay kit (Beyotime, China).

Cell transfection

Gene expression in ADSCs was regulated through transfection using Lipofectamine 3000 (Invitrogen). Let-7c-5p mimic, let-7c-5p inhibitor, and their corresponding negative controls were designed by GenePharma. For transfection, 7.5 µL of Lipofectamine 3000 was incubated with 75 µL of RNAs (let-7c-5p mimics, let-7c-5p inhibitors, and their negative controls) in 250 µL of MEM (Invitrogen) for 15 min and subsequently added to ADSCs for a 48-h incubation period. Similarly, Passage 3 TSCs were transfected using Lipofectamine 3000 (small interfering RNAs against CRCT1 and the negative control were purchased from GenePharma). The RNA sequences are presented in Additional file 1: Table S1. Quantitative real-time polymerase chain reaction (qRT-PCR) was performed to evaluate the expression level of target RNAs to verify the transfection efficiency of cells in each group.

Dual-luciferase reporter assay

The potential downstream target CRCT1 of let-7c-5p was predicted through TargetScan, DIANA-microT, and miRanda databases. We constructed wild- and mutant-type (WT/MUT-CRCT1) pmirGLO luciferase reporter plasmids for the CRCT1-3’ UTR. Using Lipofectamine 3000, we transfected these plasmids along with let-7c-5p mimics and their respective negative controls into TSCs. After 24 h of transfection, fluorescence was detected using a dual-luciferase reporter assay kit (Beyotime).

Treatment of TSCs with different ADSC-sEVs

TSCs were seeded in 6-well plates at a cell density of 5 × 10⁶/well to investigate the impact of various ADSC-sEVs subsets. Each group of TSCs was treated with 50 µg/mL of different ADSC-sEVs. The TSCs were randomly categorized into six groups as follows: (1) Control: TSC medium was replaced with an small extracellular vesicle-free medium. (2) ADSC-sEVs: normal ADSC-sEVs were added to the small extracellular vesicle-free medium. (3) NC-mimic ADSC-sEVs: negative control-mimic transfected ADSC-sEVs were added to the small extracellular vesicle-free medium. (4) let-7c-5p-mimic ADSC-sEVs: let-7c-5p-mimic transfected ADSC-sEVs were added to the small extracellular vesicle-free medium. (5) NC-inhibitor ADSC-sEVs: negative control-inhibitor transfected ADSC-sEVs were added to the small extracellular vesicle-free medium. (6) let-7c-5p-inhibitor ADSC-sEVs: let-7c-5p-inhibitor transfected ADSC-sEVs were added to the small extracellular vesicle-free medium. Subsequently, we explored the regulation of let-7c-5p on downstream targets and introduced inhibitor-let-7c-5p ADSC-sEVs and AG490 (a JAK2/STAT3 inhibitor, MedChemExpress, Monmouth Junction, NJ, USA) to TSCs transfected with si-CRCT1 or siNC. Additionally, we treated TSCs with TNFAIP6⁺ and TNFAIP6⁻ ADSC-sEVs to further explore the signaling pathways.

Cell proliferation assay

The proliferation capacity of TSCs was assessed by an EdU assay kit (UE, China). TSCs in each group were pretreated with 50 µmol/L EdU for 4 h. Following fixation with 4% paraformaldehyde, the Click-iT EdU working solution was prepared according to the manufacturer’s instructions. Subsequently, 5 µg/mL Hoechst 33,342 (UE) was used to label nuclei for 20 min, and the cell proliferation was quantified using a fluorescent fiber microscope (Leica, Wetzlar, Germany).

Cell migration assay

In the upper chamber of the Transwell plate, TSCs (1 × 10⁴) were seeded, and an small extracellular vesicle-free medium containing different ADSC-sEVs was added to the lower chamber based on the experimental groups. After incubation at 37℃ and 5% CO₂ for 24 h, the TSCs in the upper chamber were stained with 0.1% crystal violet, and the extent of TSCs migration was quantified under a light microscope (Leica).

Additionally, a cell scratch assay was performed to evaluate TSCs migration. TSCs were seeded in 6-well plates at a cell density of 2 × 10⁵ cells/well and cultured overnight at 37℃ and 5% CO₂. The assay involved scratching a straight line in the cultured cells and adding different ADSC-sEVs to an small extracellular vesicle-free medium. Changes in wound healing at 0 h and 24 h were recorded to evaluate TSC migration of TSCs.

Western blot analyses

Proteins were extracted from TSCs using RIPA buffer (Beyotime), and 20 µg of protein from each group was selected for analyses. Immunoblotting was conducted with the following primary antibodies: anti-CD9 (ab92726; Abcam, Cambridge, UK), anti-TSG101 (ab125011; Abcam), anti-HSP70 (ab2787; Abcam), anti-scleraxis (DF13293; Affinity Biologicals, Ancaster, ON, Canada), anti-tenomodulin (DF13715; Affinity), anti-collagen I (AF7001; Affinity), anti-phospho(p)-JAK2 (AF3024; Affinity), anti-JAK2 (AF6022; Affinity), anti-phospho(p)-STAT3 (AF3293; Affinity), anti-STAT3 (AF6294; Affinity), and anti-TNFAIP6 (PA5-75332; Thermo Fisher, Massachusetts, USA). Horseradish peroxidase-conjugated goat anti-rabbit immunoglobulin G (IgG) (BA1055; Boster, Wuhan, China) served as a secondary antibody. A chemiluminescence imaging system (ChemiScope 6200T, Clinx Science Instruments, Shanghai, China) was used to obtain images. Protein bands were quantified using ImageJ software.

qRT-PCR analyses

Total RNA in TSCs was extracted using Trizol reagent (Beyotime), and total RNAs from sEVs were extracted using a total RNA isolation kit (Thermo). RNAs were reverse transcribed into cDNA using Premix Ex Taq II. The qRT-PCR primers were obtained from GenePharma (Additional file 2: Table S2). RNA expression in the samples was assessed using real-time PCR with SYBR Green (Takara, Japan) on an ABI StepOnePlus system.

Experimental protocols and surgical procedures

A total of 108 SD rats were randomly divided into 5 groups: (1) Control (n = 36): Animals underwent partial patellar tendon resection surgery. (2) ADSC-sEVs (n = 24): Animals were treated with 200 µg ADSC-sEVs after partial patellar tendon resection surgery. (3) mimic-let-7c-5p ADSC-sEVs (n = 24): Animals were treated with 200 µg mimic-let-7c-5p ADSC-sEVs after partial patellar tendon resection surgery. (4) TNFAIP6⁺ ADSC-sEVs (n = 12): Animals were treated with 200 µg TNFAIP6⁺ ADSC-sEVs after partial patellar tendon resection surgery. (5) TNFAIP6⁻ ADSC-sEVs (n = 12): Animals were treated with 200 µg TNFAIP6⁻ ADSC-sEVs after partial patellar tendon resection surgery.

All animals were anesthetized with 0.3% sodium pentobarbital (30 mg/kg) before surgery. The rat patellar tendon injury model was constructed as previously described [19]. Briefly, an incision was made in the median patellar tendon to remove the middle one-third of the patellar tendon tissue. GelMA (EFL-GM-60, 10% w/v) mixed with different ADSC-sEVs was applied to the lesion and crosslinked into a gel state using ultraviolet light. Finally, the skin incision was closed using 4 − 0 sutures. Animals were euthanized by excessive CO₂ inhalation on days 7, 14, and 28, and patellar tendon tissue was extracted for subsequent studies.

Histopathological and immunohistochemical analyses

Patellar tendon tissues were fixed with 4% paraformaldehyde (Beyotime) and sectioned after paraffin embedding (0.4 μm). The tendon healing stage was evaluated using a light microscope after hematoxylin and eosin (HE) staining (Beyotime). For immunohistochemical analyses, tissue sections were stained with primary antibodies: anti-SCXA (DF13293; Affinity), anti-TNMD (DF13293; Affinity), anti-collagen I (AF7001; Affinity), and anti-CD146 (ab75769; Abcam), followed by incubation with a goat anti-rabbit IgG secondary antibody (ab6721; Abcam). For immunofluorescence analysis, tissue sections were incubated with primary antibodies: anti-CCR7 (ab32527; Abcam), anti-CD163 (ab182422; Abcam), anti-interleukin (IL)-6 (TA500067S; Origene), and anti-IL-10 ( ab33471; Abcam), followed by incubation with secondary antibodies (SA00013; ProteinTech, Chicago, IL, USA). Photographs were taken with a DM4 B microscope (Leica). Six tendon tissue samples were randomly selected from each group for analyses, and five fields per section were randomly selected for statistical analysis. Positive signals were quantified using ImageJ software.

Biomechanical testing

The bones at both ends (patella–patellar, tendon–tibia) of the six rat patellar tendon tissue samples obtained from each group were fixed on a Zwick I Z010 (Bavaria, Germany) for mechanical testing. The load-displacement curve before patellar tendon rupture was recorded at a rate of 5 mm/min. Subsequently, the failure load (N) and stiffness (N/mm) were obtained using testXpert software. Young’s modulus (N × 10³/mm²) was calculated after measuring the cross-sectional area (mm²) of the tendon with a vernier caliper.

Collection and analysis of single-cell RNA sequencing data

Rat inguinal adipose tissue single-cell RNA sequencing (scRNA-seq) data were obtained from the Gene Expression Omnibus (GSM3717978), Sequence Read Archive (SRR715485), and Array Express (E-MTAB-6677) databases. A unique molecular identifier (UMI) counting matrix was generated using the Seurat package in R. Cell data with a mitochondrial ratio exceeding 20% and single cells with over 6000 genes were filtered out. Finally, 20,764 single cells remained (2902 in GSM3717978, 9644 in SRR715485, and 8218 in E-MTAB-6677) and were applied in downstream analyses.

After quality control, the UMI count matrix was log normalized. Samples from three samples were processed and sequenced in batches to remove potential batch effects. In this process, the top 2000 variable genes were selected to create potential anchors using the FindIntgrationAnchors function of Seuart in R. Subsequently, the IntegrateData function was used to integrate data and create a new matrix with 2000 features, in which potential batch effects were mitigated.

To reduce the dimensionality of the scRNA-seq dataset, principal component analysis was performed on an integrated data matrix. Using the Elbowplot function of Seurat, the top 30 principal components were selected for downstream analysis. The main cell clusters were identified using the FindClusters function of Seurat, with the resolution set as default (res = 1.0), and subsequently visualized by two-dimensional tSNE or uniform manifold approximation and projection plots. Conventional markers described in a previous study were used to categorize each cell into a known biological cell type [20–22]. First, 20,764 cells were clustered into 10 major cell types. Subsequently, each major cell type was divided into subclusters to detect heterogeneity within each cell type. The Findallmaker function of Seurat was utilized to identify preferentially expressed genes in each cluster. Finally, heterogeneity among different cell clusters within the adipose tissue samples was detected based on the preferentially expressed genes in the cell clusters. scRNA-seq data from the ADSC clusters were extracted for further heterogeneity analysis and screened for different ADSC subgroups and specific marker genes.

Flow cytometry sorting

A total of 4 × 10⁶ Passage 3 ADSCs were collected and suspended in PBS containing 2% FBS, followed by incubation with anti-TNFAIP6 for 1 h and FITC-conjugated anti-rabbit IgG secondary antibody (S0008; Affinity) for 30 min. After filtering through a 300-mesh filter, TNFAIP6⁻/TNFAIP6⁺ ADSCs were sorted using a flow cytometer (BD FACSMelody, Franklin Lakes, USA) and added to DMEM/F12 containing 20% FBS. After the TNFAIP6⁻/TNFAIP6⁺ ADSCs were amplified to 80%, the culture medium was discarded and replaced with an small extracellular vesicle-free medium for 24 h. The cell supernatant was collected, and TNFAIP6⁻/TNFAIP6⁺ ADSC-sEVs were extracted using the above method. The adipogenic, osteogenic, and chondrogenic differentiation of various ADSC subpopulations was induced using assay kits from Cyagen, as described in our previous study [19].

Statistical analyses

All values are expressed as mean ± standard deviation. Quantitative data for each group were analyzed using one-way analysis of variance (ANOVA), followed by the Tukey–Kramer test. Analyses were performed using GraphPad Prism 9.5, and statistical significance was set at p < 0.05.

Statement

The work has been reported in line with the ARRIVE guidelines 2.0.

Results

Characteristics of ADSCs and internalization of ADSC-sEVs

ADSCs grew adherent and showed a spindle-shaped morphology (Additional file 3: Fig. S1A). ADSCs induced in vitro differentiated via adipogenesis, osteogenic, and chondrogenesis (Additional file 3: Fig. S1B). Flow cytometric analysis of ADSC surface markers (CD90- and CD105-positive and CD34-, CD45-, and CD11b-negative) is shown in Additional file 3: Fig. S1C. The PKH26 staining tracer experiment showed the successful uptake of ADSC-sEVs by TSCs (Additional file 3: Fig. S1D).

Let-7c-5p was highly expressed in ADSC-sEVs

Our previous investigations confirmed the efficacy of ADSC-sEVs in promoting the healing of tendon injuries [19]. To further explore the specific mechanism of action of ADSC-sEVs, we performed high-throughput sequencing of miRNA expression in ADSC-sEVs, revealing that let-7c-5p was significantly upregulated in ADSC-sEVs (Fig. 1A). A parallel study on the biological functions of miRNAs in TSC-sEVs demonstrated a pronounced enrichment of let-7c-5p (Fig. 1B) [23], suggesting that let-7c-5p expression in ADSC-sEVs has significant implications for TSCs.

Fig. 1.

The miRNA expression in different MSC-sEVs. A The high-throughput sequencing of miRNA expression in ADSC-sEVs, sorted by total read counts. B The high-throughput sequencing of miRNA expression in TSC-sEVs, sorted by total read counts. Data are presented as mean ± standard deviation

Characteristics of different ADSC-sEVs

ADSC-sEVs, mimic-let-7c-5p ADSC-sEVs, and inhibitor-let-7c-5p ADSC-sEVs exhibited quasi-circular structures under TEM (Fig. 2A). NTA showed that the diameters of these ADSC-sEVs were 110.3 nm, 112.7 nm, and 112.7 nm, respectively. (Fig. 2B). Western blot analysis confirmed the expression of CD9, TSG101, and HSP70 on the surface of all three ADSC-sEVs (Fig. 2C). In addition, we calculated the purity of these ADSC-sEVs were 4.04 × 10⁷ particles/µg, 4.11 × 10⁷ particles/µg, 4.17 × 10⁷ particles/µg, respectively.

Fig. 2.

Characterization of different ADSC-sEVs. A Morphology of ADSC-sEVs, let-7c-5p-mimic ADSC-sEVs, and let-7c-5p-inhibitor ADSC-sEVs under a transmission electron microscope. B The particle size distribution of ADSC-sEVs, let-7c-5p-mimic ADSC-sEVs, and let-7c-5p-inhibitor ADSC-sEVs. C The surface markers of ADSC-sEVs, let-7c-5p-mimic ADSC-sEVs, and let-7c-5p-inhibitor ADSC-sEVs were detected by western blot. Bars 100 nm

let-7c-5p expression in ADSC-sEVs regulates the biological characteristics of TSCs

ADSCs with different let-7c-5p expression levels (NC-mimic, let-7c-5p-mimic, NC-inhibitor, and let-7c-5p-inhibitor) were obtained through transfection (Fig. 3A), whereas ADSC-sEVs with different let-7c-5p expression levels were obtained through centrifugation (Fig. 3B). The expression of let-7c-5p in the let-7c-5p-mimic group was significantly increased, whereas that in the let-7c-5p-inhibitor group was not significantly suppressed. This result may be related to the fact that miRNA inhibitors act by competitive binding inhibition. NC-mimic and NC-inhibitor served as control groups. TSCs were treated with five different ADSC-sEVs. The EdU assay showed that let-7c-5p-mimic ADSC-sEVs significantly promoted TSC proliferation, whereas let-7c-5p-inhibitor ADSC-sEVs attenuated this effect (Fig. 3C, F). Transwell and scratch assays revealed that high let-7c-5p expression effectively promoted TSC migration (Fig. 3D, E, G, H). Additionally, alterations in let-7c-5p expression influenced the tenogenic differentiation of TSCs (Fig. 3I–L).

Fig. 3.

Effects of ADSC-sEVs with different let-7c-5p expression on the biological characteristics of TSCs. A The expression of let-7c-5p in different ADSCs. B The expression of let-7c-5p in different ADSC-sEVs. C, F Effect of different ADSC-sEVs on the proliferation of TSCs by EdU assays. D, E, G, H Effect of different ADSC-sEVs on the migration of TSCs by transwell and scratch assays. I–L Western blot analysis of protein levels of SCXA, TNMD, and COL Ⅰ induced by different ADSC-sEVs. Bars, 100 μm. Data are represented as mean ± SD. *vs ADSC-sEVs group; n = 3. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001

CRCT1 was identified as the target of let-7c-5p

CRCT1 was the only downstream target predicted by all three databases (TargetScan, DIANA-microT, and miRanda) (Fig. 4A). Therefore, we hypothesized that CRCT1 was the target gene of let-7c-5p. To verify this hypothesis, we constructed WT-CRCT1 and MUT-CRCT1 pmirGLO luciferase reporter plasmids and demonstrated targeting of let-7c-5p by CRCT1 using a dual-luciferase reporter assay (Fig. 4B). We found that let-7c-5p mimics significantly reduced the fluorescence expression of WT-CRCT1, while a lesser effect on MUT-CRCT1 expression was observed (Fig. 4C). In addition, we treated TSCs with let-7c-5p-inhibitor and NC-inhibitor ADSC-sEVs. qPCR analysis revealed that let-7c-5p expression was significantly correlated with CRCT1 expression (Fig. 4D). The CRCT1-related proteins, predicted using STRING, are shown in Fig. 4E. Kyoto Encyclopedia of Genes and Genomes (KEGG) analysis showed that the JAK/STAT signaling pathway was the most significantly enriched (Fig. 4F).

Fig. 4.

Let-7c-5p targets the expression of CRCT1. A Venn diagram of let-7c-5p potential target genes in the TargetScan, DIANA-microT and miRanda databases. B Schematic diagram of dual luciferase reporter plasmid construction. WT, wild-type; MUT, mutant. C Columnar statistics for relative luciferase activity. D Expression level of CRCT1 after treatment with let-7c-5p-mimic/inhibitor. E, F Bioinformatics analysis of CRCT1-related proteins and signaling pathways

let-7c-5p/CRCT1 affected the biological characteristics of TSCs through the JAK2/STAT3 signaling pathway

We introduced let-7c-5p-inhibitor ADSC-sEVs to TSCs transfected with si-CRCT1 or si-NC. Western blotting showed that reduced let-7c-5p content in ADSC-sEVs led to increased expression of p-JAK2 and p-STAT3 in TSCs, suggesting that let-7c-5p/CRCT1 may exert its biological role through the JAK2/STAT3 signaling pathway (Fig. 5A–C).

Fig. 5.

ADSC-sEVs promote the proliferation, migration, and tenogenic differentiation of TSCs via let-7c-5p/CRCT1/JAK2/STAT3 pathway. A–C Western blot analysis of protein levels of p-JAK2 and p-STAT3 expression in si-CRCT1 TSCs induced by let-7c-5p-inhibitor ADSC-sEVs. D–F AG490 inhibits the activation of p-JAK2 and p-STAT3 induced by let-7c-5p-inhibitor ADSC-sEVs. G, H EdU assay showed that AG490 enhanced the proliferation of let-7c-5p-inhibitor ADSC-sEVs on TSCs. I–L Transwell and scratch assays showed that AG490 enhanced the migration of let-7c-5p-inhibitor ADSC-sEVs on TSCs. M–P Western blot analysis of protein levels of SCXA, TNMD, and COL Ⅰ promoted by let-7c-5p-inhibitor ADSC-sEVs were enhanced by AG490. Bars, 100 μm. Data are represented as mean ± SD. *vs ADSC-sEVs group; ^#vs. let−7c−5p-inhibitor ADSC-sEVs; n = 3. *P < 0.05, **P < 0.01, ***P < 0.001, ^#P < 0.05, ^##P < 0.01, ^###P < 0.001, ^####P < 0.0001

Subsequently, we evaluated the effect of let-7c-5p/CRCT1/JAK2/STAT3 on TSCs using AG490. Western blotting showed that AG490 significantly inhibited the activation of the JAK2/STAT3 signaling pathway mediated by let-7c-5p-inhibitor ADSC-sEVs (Fig. 5D–F). Pretreatment with AG490 also mitigated the enhanced TSC proliferation and migration mediated by let-7c-5p-inhibitor ADSC-sEVs (Fig. 5G–L). Additionally, JAK2/STAT3 signaling pathway inhibition significantly reduced the effect of let-7c-5p-inhibitor ADSC-sEVs on the tenogenic differentiation of TSCs (Fig. 5M–P).

Release characteristics of ADSC-sEVs in the GelMA

Different ADSC-sEVs were obtained using differential centrifugation (Fig. 6A). The photoinitiator LAP was used to initiate GelMA polymerization, and the sieve-like biological scaffold structure was loaded with ADSC-sEVs (Fig. 6B, C). The in vitro ADSC-sEVs release analysis suggested that the GelMA-loaded ADSC-sEVs were gradually released in 7 days (Fig. 6D). Similarly, the in vivo PKH26-labeled sEV tracing experiments found that ADSC-sEVs acted locally for more than 7 days under GelMA loading (Fig. 6E, F). These results demonstrate that GelMA-loaded ADSC-sEVs can achieve a sustained release effect, enhancing the therapeutic efficacy of ADSC-sEVs in vivo.

Fig. 6.

Application and release characteristics of GelMA-loaded ADSC-sEVs in animal tendon injury model. A Steps for ultracentrifugal extraction of ADSC-sEVs. B The preparation method of GelMA. C The scanning electron microscopy image of GelMA-loaded ADSC-sEVs (ADSC-sEVs are shown by black arrow). Bars, 10 μm. D Profile of ADSC-sEVs released from the GelMA. E Treatment of GelMA-loaded ADSC-sEVs in animal tendon injury model. F In vivo imaging shew the retention time of PKH26-labeled ADSC-sEVs at the site of the tendon injury by loading GelMA

GelMA-loaded let-7c-5p-mimic ADSC-sEVs improved the quality of acute tendon injury healing in vivo

ADSC-sEVs and let-7c-5p-mimic ADSC-sEVs were administered to a rat patellar tendon injury model to evaluate their in vivo effects. HE staining revealed a more regular fibrous structure in the healing patellar tendon of the let-7c-5p-mimic ADSC-sEV group compared with that of the ADSC-sEV group at 2 and 4 weeks (Fig. 7A). The tendon stem cell marker CD146 displayed the highest expression in the let-7c-5p-mimic ADSC-sEV group, significantly exceeding that in the other two groups (Fig. 7A–C). These results suggest that let-7c-5p effectively promotes the proliferation and tenogenic differentiation of TSCs in vivo. The therapeutic effects of let-7c-5p on injured tendons were evaluated using biomechanical tests. The healing patellar tendon tissue in the let-7c-5p-mimic ADSC-sEV group showed significant superiority in terms of failure load, stiffness, and Young’s modulus compared with that in the other groups (Fig. 7D–G).

Fig. 7.

Let-7c-5p improved the healing of tendon injury. A The H&E staining and immunohistochemistry assay (SCX, TNMD, COL Ⅰ, and CD146) of tendon injury at week 2 (n = 6) and week 4 (n = 6). B, C Quantitative analysis of tenogenic related factors at week 2 (n = 6) and week 4 (n = 6). D–G Results of biomechanical tests (failure load, stiffness, and Young’s modulus) at 4 weeks (n = 6). Bars (H&E), 100 μm; bars (immunohistochemistry), 50 μm. Data are represented as mean ± SD. *vs Control group; ^#vs. ADSC-sEVs; n = 6. *P < 0.05, **P < 0.01, ***P < 0.001, ^#P < 0.05, ^##P < 0.01

GelMA-loaded let-7c-5p-mimic ADSC-sEVs reduced the tissue inflammatory response in vivo

The effect of let-7c-5p on the early inflammatory response to tendon injury were evaluated using immunofluorescence. The expression of inflammation-related factors in tendon tissues of the Control, ADSC-sEV, and let-7c-5p-mimic ADSC-sEV groups was assessed. The results indicated a significant decrease in the expression of IL-6 (a pro-inflammatory factor) in the let-7c-5p-mimic ADSC-sEV group, while IL-10 (an anti-inflammatory factor) expression was increased. Simultaneously, the expression of CCR7 (an M1 macrophage marker) significantly decreased in the let-7c-5p-mimic ADSC-sEV group, while the expression of CD163 (an M2 macrophage marker) significantly increased (Fig. 8A, B).

Fig. 8.

Let-7c-5p inhibits inflammatory expression in tendon injury. A The expression of CCR7⁺, CD163⁺, IL-6⁺, and IL-10⁺ cells were detected by immunofluorescence at week 1. B Positive ratio of inflammation-related factors (n = 6). Bars 50 μm. Data are represented as mean ± SD. *vs Control group; ^#vs. ADSC-sEVs; n = 6. *P < 0.05, **P < 0.01, ^#P < 0.05, ^##P < 0.01, ^###P < 0.001

TNFAIP6 as a key marker gene for the functional and stemness subgroups of ADSCs

The promoting effect of ADSCs on acute tendon injury repair were explored from the perspective of cellular heterogeneity. To improve the accuracy of the analysis results, we used a multicenter joint analysis of scRNA-seq data (GSM3717978, SRR715485, and E-MTAB-6677) from subcutaneous adipocyte populations in the groin of rats. After normalizing the obtained UMI matrix, the FindIntegrationAnchors function of the Seurat package was used to integrate three sets of scRNA-seq data (Fig. 9A), revealing that the three cell groups comprised very similar subgroups and high analytical quality (Fig. 9B-G). Based on relevant studies [20–22], we integrated and re-annotated 10 cell groups, including adipose-derived stem cells, pre-adipocytes, adipose-regulatory cells, adipocytes, endothelial cells, smooth muscle cells, T cells, B cells, macrophages, and dendritic cells (Fig. 9H–I). Subsequently, we extracted scRNA-seq data from ADSC cell populations for further analysis (Fig. 9J). Based on the screening and analysis of maker genes, we categorized ADSCs into a stemness subgroup (maintaining stemness related genes, such as TNFAIP6, CEBPB, and MT1) and a functional subgroup (multidirectional differentiation ability and cytokine expression related genes, such as CXCL13, SFRP2, and BMP7) [24–29]. A significant difference was found in the expression of TNFAIP6 between the two ADSC subgroups (Fig. 9K–N).

Fig. 9.

The scRNA-seq data analysis of ADSCs. A–G The multicenter joint analysis of scRNA-seq data from adipocyte populations. H, I tSNE 2D cell map displaying the clustering of scRNA-seq data (GSM3717978, SRR715485, and E-MTAB-6677) J tSNE 2D cell map colored by cell cluster identification. K Analysis of ADSCs subgroups (stemness and functional) based on scRNA-seq data. L–N Individual gene tSNE and violin plots showing the expression levels and distribution of representative marker genes

GelMA-loaded TNFAIP6- ADSC-sEVs improved the quality of acute tendon injury healing through let-7c-5p

ADSCs were successfully classified into TNFAIP6⁺ ADSCs and TNFAIP6⁻ ADSCs using flow cytometry sorting technology (Fig. 10A–C). The stemness of TNFAIP6⁺ ADSCs was significantly stronger than that of TNFAIP6⁻ ADSCs (Fig. 10D). Next, TNFAIP6⁻/TNFAIP6⁺ ADSCs-sEVs were extracted, and the expression levels of let-7c-5p in these sEVs were explored. As expected, let-7c-5p was significantly enriched in TNFAIP6⁻ ADSC-sEVs (Fig. 10E). The regulatory ability of TNFAIP6⁻ ADSC-sEVs on the tenogenic differentiation, proliferation, and migration of TSCs was significantly superior than that of TNFAIP6⁺ ADSC-sEVs (Fig. 10F–O). Finally, TNFAIP6⁻/TNFAIP6⁺ ADSCs-sEVs were transplanted into the rat patellar tendon injury model, revealing that TNFAIP6⁻ ADSC-sEVs strongly promoted acute tendon repair in vivo (Fig. 10P–T).

Fig. 10.

TNFAIP6⁻ ADSC-sEVs improved the healing of tendon injury through let-7c-5p. A–D Isolation and characterization (adipogenesis, osteogenic, and chondrogenesis) of TNFAIP6⁻ and TNFAIP6⁺ ADSCs. E The expression levels of let-7c-5p in TNFAIP6⁻ ADSC-sEVs and TNFAIP6⁺ ADSC-sEVs. F–O The effect of TNFAIP6⁻ ADSC-sEVs and TNFAIP6⁺ ADSC-sEVs on the proliferation, migration, and tenogenic differentiation of TSCs. P–T The therapeutic effects of TNFAIP6⁻ ADSC-sEVs and TNFAIP6⁺ ADSC-sEVs on tendon injuries in vivo. Bars (cells), 100 μm; bars (H&E), 100 μm; bars (immunohistochemistry), 50 μm. Data are represented as mean ± SD. *vs TNFAIP6⁺ ADSC-sEVs group; n (cell) = 3, n (immunohistochemistry) = 6. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001

Discussion

Tendon injury is a common condition affecting the motor system. It poses challenges due to the slow and inefficient natural tendon repair process, leading to unsatisfactory long-term rehabilitation outcomes [30]. Restoring the original physiological structure and biological characteristics of tendons poses an ongoing challenge for medical professionals. TSCs are a self-renewing cell population with multidirectional differentiation potential in tendon tissue [31]. Notably, Komatsu et al. observed that the transplantation of TSC sheets significantly accelerated tendon healing and regeneration [32].

Recent studies have shown that MSCs are beneficial in tissue healing, particularly through the secretion of sEVs. Cell-free biological therapy with MSC-sEVs has emerged as a promising approach in clinical applications [33, 34]. We previously found that TSC-sEVs could significantly enhance the biological activity of tendon cell populations and improve the quality of tendon healing following acute injury [23]. However, extracting TSCs from patients with pre-existing tendon injuries is challenging. Consequently, the application of TSC-sEVs and transforming this approach from basic research to clinical practice is hindered. Therefore, it is urgent to find other MSC-sEVs suitable for clinical application to optimize tendon healing. ADSCs are widely used in tissue engineering and regenerative medicine due to their abundant tissue sources, convenient cell separation, and strong cellular activity [35]. In a previous study, we introduced ADSC-sEVs for the first time in the context of tendon injuries and observed a notable improvement in the quality of acute tendon injury healing [19]. Therefore, the present study attempted to explore the specific mechanisms and key components of ADSC-sEVs in treating tendon injuries.

Considering the potential effect of estrogen on tendon healing, male rats were selected for in vivo exploration in the current study. ADSC-sEVs comprise a complex mixture of bioactive substances, such as proteins, mRNA, miRNA, and DNA. As key regulators in post-transcriptional gene expression, miRNAs play important roles in cell differentiation, biological development, and disease progression [36]. Liu et al. found that miRNA expression significantly impacts tendon growth and wound healing [37]. Although the specific mechanisms through which sEVs endogenously regulate the biological characteristics of TSCs are not fully understood, miRNAs have been identified as key players. Therefore, in the current study, we investigated miRNAs within ADSC-sEVs in the treatment of tendon injuries.

High-throughput sequencing determined that let-7c-5p was significantly enriched in ADSC-sEVs. This finding is consistent with that of a previous study showing that let-7c-5p was the most significantly enriched in TSC-sEVs [23]. This led us to hypothesize that ADSC-sEVs primarily promote acute tendon repair through let-7c-5p. The let-7 family, one of the earliest discovered miRNAs, is predominantly involved in cell differentiation and metabolic regulation [38–40]. Wang et al. demonstrated that let-7c-5p promotes the osteogenic differentiation of BMSCs via the tumor growth factor (TGF)-β signaling pathway [41]. Additionally, Wang et al. found that let-7c-5p and miRNA-21-5p in BMSC-sEVs promote the proliferation and migration of endothelial cells treated with rapamycin [42]. In the present study, we found that let-7c-5p effectively promoted the proliferation, migration, and tenogenic differentiation of TSCs.

Herein, we transfected ADSCs to change the expression level of let-7c-5p in ADSC-sEVs. Subsequently, we found that different ADSC-sEVs had similar NTA peak values and particle-to-protein ratios. Therefore, we considered that the ADSC-sEV subgroups had similar purity. We also noted that increasing let-7c-5p expression in ADSC-sEVs effectively improves the quality of tendon healing following acute injury. It is worth mentioning that let-7c-5p-inhibitor also had a positive regulatory effect on the biological characteristics of TSCs in vitro. On the one hand, we considered that although the expression of let-7c-5p was inhibited, it could still induce biological effects. On the other hand, other bioactive molecules in ADSC-sEVs can also effectively regulate the biological characteristics of TSCs, albeit their ability to regulate TSCs is lower than that of let-7c-5p. Overall, these results suggest a significant role of let-7c-5p in the ADSC-sEV-mediated treatment of tendon injuries.

To further understand the role of let-7c-5p in the treatment of tendon injuries, we employed TargetScan, DIANA-microT, and miRanda databases to predict possible targets. CRCT1 was selected as the focus, as it was the sole target intersecting all three databases. CRCT1 is encoded by an epidermal differentiation complex and plays a vital role in epidermal differentiation [43]. In the current study, we confirmed that let-7c-5p targets CRCT1 using a dual-luciferase reporter assay. PCR results confirmed that ADSC-sEVs with low expression of let-7c-5p could alleviate CRCT1 inhibition in TSCs. Therefore, we hypothesize that the biological efficacy of ADSC-sEVs depends on the let-7c-5p/CRCT1 interaction.

Subsequently, we predicted the downstream signaling pathways associated with let-7c-5p expression using bioinformatics. KEGG analysis showed that the JAK2/STAT3 pathway was potentially correlated with let-7c-5p/CRCT1. As a canonical signaling pathway present in various cells, the JAK2/STAT3 pathway is involved in numerous key biological processes in various diseases, such as tumors and inflammatory conditions [44–46]. In addition, Chen et al. showed that the JAK2/STAT3 signaling pathway plays a key role in regulating the aging process of TSCs [47]. Given the potential impact of the JAK2/STAT3 pathway on tendon healing, we explored its changes during ADSC-sEV treatment. Western blot analyses showed a significant correlation between p-JAK2, p-STAT3, and CRCT1. Subsequently, pretreatment of TSCs with AG490 (a JAK2/STAT3 inhibitor) and incubation with let-7c-5p-inhibitor ADSC-sEVs showed that AG490 attenuated the efficacy of ADSC-sEVs on the biological characteristics of TSCs. Therefore, the JAK2/STAT3 signaling pathway plays a significant role in the let-7c-5p/CRCT1-mediated healing qualities of ADSC-sEVs in acute tendon injury.

Tendon repair is an intricate process in which SCX serves as a pivotal molecule in tendon development, crucial for regulating TSC differentiation and proliferation [48]. TNMD is a key regulatory factor influencing the biological characteristics of TSCs, including proliferation, differentiation, aging, and tendon maturation [49]. Dex et al. found that TNMD and collagen I are colocalized in the extracellular matrix, and the expression of collagen I positively affects the mechanical strength and function of tendons [50]. Building on our previous research, we increased the content of let-7c-5p in ADSC-sEVs to promote TSC tenogenic differentiation in the current study. Furthermore, the expression of CD146 (a TSC marker) in tendons significantly and positively correlated with that of let-7c-5p in ADSC-sEVs. Similarly, let-7c-5p effectively promoted TSC proliferation and migration in vitro, a process related to the let-7c-5p/CRCT1/JAK2/STAT3 signaling mechanism.

In our previous study, ADSC-sEVs reduced early inflammatory responses by regulating macrophage polarization [19]. Interestingly, recent studies have shown that let-7c-5p inhibits inflammatory factor expression [51–53]. Accordingly, we explored the effect of let-7c-5p on early inflammation in tendon injury. Considering that acute tendon injury in the present study was induced by surgical trauma, rather than simple tendon rupture, and that the study focused on the local treatment effect after tendon injury, we evaluated the expression of inflammatory factors at the tendon healing site. As expected, immunofluorescence analysis showed that the increased expression of let-7c-5p effectively inhibited inflammatory factors in the early stages of tendon injury—a process potentially related to the polarization of M2 macrophages. Finally, we evaluated the quality of tendon healing through biomechanical testing. Tendon load, stiffness, and Young’s modulus in the mimic group showed the most significant improvement, indicating that let-7c-5p plays a crucial role in ADSC-sEV-mediated healing of acute tendon injury.

We found that the expression of let-7c-5p and its therapeutic effect on tendon injury varied between batches of ADSC-sEVs. Additionally, clinical trials have shown fluctuations in the efficacy of ADSC-sEVs [54]. Based on Wang et al.’s study on the heterogeneity of MSCs [55], we speculate that competition among different subgroups of ADSCs during culture leads to imbalanced proportions and changes in ADSC-sEVs function and efficacy. However, the current MSC biomarkers do not clearly define the heterogeneity of ADSC subgroups. Therefore, there is an urgent need to elucidate potential molecular markers for different ADSC subgroups and their heterogeneity. To enhance the accuracy of the analysis, we combined data from the Gene Expression Omnibus (GSM3717978), Sequence Read Archive (SRR715485), and Array Express databases (E-MTAB-6677) to integrate and extract ADSC scRNA-seq data from multiple centers for analysis. Adipose tissue is the body’s energy storage unit, and cell populations contain a high proportion of mitochondria; therefore, we screened cell data with a mitochondrial ratio < 20% to improve the analysis accuracy and filtered out single cells with >6000 genes to eliminate potential duplicate samples. The results of quality control showed that the cells in each group had high analytical quality. Subsequently, we used the Louvain algorithm offered by Seurat and related references to define the clustering resolution as 1.0, and 10 cell subpopulations were established for analysis. Finally, we categorized ADSCs into stemness and functional subgroups, with significant differential expression of TNFAIP6 observed between the two subgroups. TNFAIP6 is a glycoprotein with a molecular weight of 35–38 kDa, which plays an important role in maintaining the stemness and biological characteristics of MSCs [56]. Given the important role of the let-7 family in promoting stem cell differentiation [57], we hypothesize that the functional subgroup of ADSCs is the primary subgroup secreting let-7c-5p and that TNFAIP6 is the key marker for distinguishing ADSC subgroups. In subsequent experiments, we successfully isolated the TNFAIP6⁻ ADSCs. As expected, the differentiation ability of TNFAIP6⁻ ADSCs was significantly weaker than that of TNFAIP6⁺ ADSCs, while the expression of let-7c-5p in TNFAIP6⁻ ADSC-sEVs was significantly higher than that in TNFAIP6⁺ ADSC-sEVs. In vivo and in vitro experiments showed that TNFAIP6⁻ ADSC-sEVs significantly improved the quality of tendon healing following acute injury through let-7c-5p. To further explore the negative regulatory mechanism of TNFAIP6 on let-7c-5p, we used SCENIC analysis to explore transcription status differences among cell subsets and predicted the binding of let-7c-5p promoter using MACS software and the JASPAR database. We found that the transcription factor of IRF7 and STAT1/STAT2 may regulate the transcription of let-7c-5p in both the forward and reverse directions (Additional files 4: Figure S2). This finding provides a new direction for follow-up research.

In previous studies, sEVs have been found to exert their biological functions through local or intravenous injection. Currently, no practical carrier is available for in vivo treatment with sEVs. GelMA, a natural biomaterial derived from collagen widely used in biomedicine due to its excellent biocompatibility and low immunogenicity [58, 59], was selected as a carrier in the current study. We found that GeIMA targeted the therapeutic position of ADSC-sEVs and effectively regulated the biological characteristics of TSCs (proliferation, migration, and tenogenic differentiation); its sustained release effect enhanced the bioavailability of ADSC-sEVs. In addition, the therapeutic effect of ADSC-sEVs was significantly improved by constructing TNFAIP6⁻ ADSC-sEVs with high let-7c-5p expression. Therefore, constructing conditional ADSC-sEVs to optimize their therapeutic efficacy may be a potential strategy for sEV therapy in the future. Furthermore, we validated the benefits of TNFAIP6⁻ ADSC-sEVs in a rat tendon injury model. Thus, confirmation in clinical trials is required to realize the clinical transformation of basic research.

This study had some limitations. First, cell signaling pathways are complex and precise networks, and several signaling pathways (such as the TGF-β, MAPK, and Wnt families) are involved in the tendon healing process. This study only predicted and verified the key role of CRCT1/JAK2/STAT3 in the regulation of tendon healing by TNFAIP6⁻ ADSC-sEVs, and other related molecular biological mechanisms require further exploration. Second, we only evaluated the short-term quality of tendon healing, and long-term tendon treatment effects require further study. In addition, we used vernier calipers to measure the tendon cross-sectional area for biomechanical testing, which may have potential measurement deviation. We will use more accurate imaging evaluation methods for analysis in subsequent studies. Finally, we successfully constructed TNFAIP6- ADSC-sEVs and validated their superiority in treating tendon injuries. However, further exploration is needed to determine the specific reasons for the high expression of let-7c-5p in TNFAIP6- ADSC-sEVs and the efficacy of other cell subsets in adipose tissue.

Conclusions

In conclusion, this study explores the potential molecular mechanisms involved in acute tendon injury treatment using GelMA-loaded TNFAIP6⁻ ADSC-sEVs, focusing on the regulation of let-7c-5p expression. We combined scRNA-seq data to identify the key role of the TNFAIP6⁻ ADSCs subgroup in the treatment of acute tendon injury with ADSCs. Our results suggest that TNFAIP6⁻ ADSC-sEVs regulate the biological characteristics of TSCs through the let-7c-5p/CRCT1/JAK2/STAT3 signaling pathway, thereby promoting high-quality tendon healing. Additionally, this study provides a reference for constructing GelMA-loaded conditional ADSC-sEVs to optimize their therapeutic effects, which also provides new ideas for cell-free regenerative medicine.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

HCL, ADZ, and MYS contributed to cytology experiment, animal experiments, data acquisition, and manuscript writing; JYZ and TTZ revised this manuscript and analyzed the data; WJL, ZMZ, and ZNZ revised and editing this manuscript; YW and YBM guided the experiment, conceptualization; SYW and LMH provided experimental technical support and final approval of manuscript; QBC and ZZL took part in the experimental design, text revision, and final approval of manuscript.

Funding

This study was supported by the Natural Science Foundation of Zhejiang Provincial (LQN25H170001), National Natural Science Foundation of China (81871837, 81572117), the Natural Science Foundation of Anhui Provincial (2308085QH259), and the Specialized Research Fund for Doctoral Programs in Colleges and Universities of China (20132307110007).

Data availability

The data that support the findings of this study are available from the corresponding author upon reasonable request. The scRNA-seq datasets can be available in the Gene Expression Omnibus database with accession number GSM3717978 (DOI: 10.1126/science.aav2501), the Sequence Read Archive database with accession number SRR715485 (DOI: 10.1016/j.cmet.2018.05.025), and the Array Express database with accession number E-MTAB-6677 (DOI: 10.1038/s41586-018-0226-8).

Declarations

Ethics approval and consent to participate

This study received ethical approval from the Ethics Committee of Harbin Medical University (Mechanism of regulation of tendon stem cells differentiation through small extracellular vesicles) on Feb 23, 2018, with approval number Ky2018-135. All animal experiments were conducted in accordance with the ARRIVE guidelines and adhered to the applicable regulations, such as the U.K. Animals (Scientific Procedures) Act, 1986 and associated guidelines, EU Directive 2010/63/EU for animal experiments, or the National Research Council’s Guide for the Care and Use of Laboratory Animals.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.