Temporal dynamics of ADSC therapy in skin fibrosis: unraveling the roles of ROS/NF-κB/TSG-6 signaling axis

Abstract

Background

Cutaneous fibrosis, particularly in localized scleroderma (LoS), poses a considerable therapeutic challenge owing to its progressive characteristics and the subsequent effects on quality of life. Although ADSCs exhibit therapeutic potential for fibrosis, their spatiotemporal mechanisms of action, particularly within fibrotic microenvironments, remain poorly characterized. This study sought to clarify the spatiotemporal dynamics and molecular mechanisms of ADSC-mediated fibrosis resolution in bleomycin (BLM)-induced murine LoS model.

Methods

Skin fibrosis was induced in C57BL/6J mice through daily subcutaneous injections of bleomycin (BLM) administered over a period of four weeks. GFP-labeled mouse or human ADSCs were injected into the fibrotic dorsum. ADSC distribution was tracked using fluorescence imaging and flow cytometry. Skin fibrosis was assessed histologically (H&E, Masson's trichrome, α-SMA, COL1) and molecularly (qRT-PCR for cytokines). Transcriptomic profiling (RNA-seq) of sorted GFP + ADSCs was performed on days 1, 7, and 14 post-injection. Key pathways (ROS, NF-κB, TSG-6) were validated in vitro using ADSCs and human LoS-derived fibroblasts (LoSFs) via pharmacological inhibition, gene knockdown (shTSG-6), co-culture, Western blotting, and dual-luciferase assays.

Results

ADSCs mitigated dermal thickening, collagen deposition, α-SMA expression, and inflammation (TNF-α, IL-6, IL-1β) over 21 days. Transcriptomics revealed a temporal hierarchy: early oxidative stress response (Day 1), followed by immunomodulation (Day 7, NF-κB, cytokine pathways), and later ECM remodeling (Day 14). Mechanistically, TGF-β induced ROS via NOX4, activating NF-κB, which directly bound the TSG-6 promoter to drive its expression. TSG-6 knockdown in ADSCs (ADSCshTSG-6) abolished their ability to suppress TGF-β/Smad signaling, collagen production, α-SMA expression, and inflammation in vitro and in vivo.

Conclusion

ADSCs resolve skin fibrosis through a biphasic mechanism involving initial adaptation and subsequent immunomodulation/ECM remodeling, centrally governed by a ROS-NF-κB-TSG-6 axis. TSG-6 is the critical downstream effector, disrupting the TGF-β/Smad pathway and inflammation. This study identifies TSG-6 as a key therapeutic mediator and a potential biomarker for optimizing ADSC-based therapies for fibrotic skin disorders.

Graphical Abstract

Supplementary Information

The online version contains supplementary material available at 10.1186/s13287-025-04746-z.

Introduction

Cutaneous fibrosis remains a formidable challenge in clinical dermatology due to its progressive nature, refractory course, and profound impact on patients' quality of life through skin thickening, contractures, and disfigurement. Scleroderma is a multifaceted autoimmune disorder that predominantly impacts connective tissues. Localized scleroderma (LoS) is characterized by widespread and progressive skin fibrosis, which results from abnormal collagen deposition, microvascular injury, and dysregulated immune responses [1]. The inflammatory processes in scleroderma contribute to the accumulation of extracellular matrix (ECM) components, leading to an increase in skin thickness, stiffness, reduced elasticity, functional impairment, and altered appearance [2].

Adipose-derived stem cells (ADSCs) have emerged as a promising therapeutic modality due to their multifaceted antifibrotic effects [3]. These multipotent adult stem cells exhibit the ability for self-renewal, long-term viability, and differentiation into various cell lineages. Their accessibility and abundance make ADSCs an attractive and practical option for cell-based therapy in scleroderma [4–6]. Research has demonstrated that ADSCs can mitigate ECM deposition, inhibit fibroblast activation, and enhance revascularization [7].

While the therapeutic effects of adipose-derived stem cells have been explored, a direct analysis of the behavior of injected ADSCs in vivo remains underexplored. This is likely due to the challenges posed by the diseased environment, which may influence ADSCs adaptation and survival [8]. Additionally, current studies of ADSC-treated Bleomycin (BLM)-induced skin fibrosis are predominantly limited to tissue-level analyses. To attain a more profound comprehension of the therapeutic mechanisms, transcriptomic approaches such as RNA sequencing can be employed to reveal cell interactions and the molecular basis of ADSC-mediated effects.

Here, we elucidate the temporal hierarchy and mechanistic underpinnings of ADSC-mediated fibrosis resolution using a BLM-induced murine LoS model. Through integrated transcriptomic profiling, functional validation, and spatial tracking of ADSCs, we identify TSG-6 as a critical effector molecule governed by a ROS-NF-κB axis (Fig. 1). By delineating this spatiotemporal mechanism, our work advances the therapeutic paradigm for fibrotic dermatoses and positions TSG-6 as a biomarker for optimizing stem cell-based interventions.

Fig. 1.

Schematic representation of the proposed ROS-NF-κB-TSG-6 signaling axis in ADSC-mediated fibrosis resolution

Methods

Animals

All animal experiments were conducted with the approval of the Institutional Animal Care and Use Committee of Peking Union Medical College Hospital. Six-week-old male C57BL/6J mice were obtained from the Laboratory Animal Research Center and housed in a specific pathogen-free (SPF) facility at the same institution. Mice were housed in groups of three per cage, maintained under a 12-h light/dark cycle, and provided with unrestricted access to sterilized food and water. The animals were randomly assigned to six experimental groups: the control group and several BLM-treated groups, with at least six mice in each group. To establish the localized scleroderma (LoS) animal model, mice in the BLM groups received daily dorsal subcutaneous injections of 0.1 mL of a 200 mg/mL BLM solution for a duration of four weeks. Mice in the control group were administered an equivalent volume of PBS via injection. The successful establishment of the LoS model was confirmed through visual skin assessment and histopathological examination.

Human adipose-derived stem cells and Green fluorescent protein (GFP)-labeled C57BL/6 mouse ADSCs (OriCell™, China) were cultured in Dulbecco's Modified Eagle Medium/F-12 (DMEM/F12, Gibco, USA), supplemented with 10% fetal bovine serum. The cultures were maintained under a humidified atmosphere containing 5% carbon dioxide at a temperature of 37 °C, as previously described. Mouse GFP-labeled ADSCs, human ADSCs, and transfected human ADSCs (ADSC^shTSG−6&ADSC^NC) were used between passages 3 and 6 and washed three times with PBS to eliminate any residual culture medium prior to injection. A 50 µL cell suspension containing 1 × 10⁸ cells/mL (equivalent to 5 × 10⁶ cells per mouse) was subcutaneously injected into the affected dorsal area of BLM-treated mice, except for those in the BLM control group. Follow-up analyses were conducted on days 1, 7, 14, and 21 post-injection based on the specific experimental design.

Histological evaluation of skin

Skin samples were fixed in 4% (v/v) paraformaldehyde in phosphate-buffered saline (PBS), subsequently embedded in paraffin, and sectioned into slices with a thickness of 5 μm, which were then mounted on glass slides. Histological evaluation of fibrosis was performed utilizing hematoxylin and eosin (H&E) staining as well as Masson's trichrome staining. The stained sections were scanned using a microscope equipped with a 20 × objective lens. Dermal thickness and collagen distribution were quantified using ImageJ 2.0 software, analyzing random, non-overlapping fields to ensure unbiased results.

To inhibit endogenous peroxidase activity, deparaffinized sections were treated with 3% hydrogen peroxide (H2O2) at 37 °C for a duration of 15 min. This was followed by blocking with 5% bovine serum albumin (BSA) in phosphate-buffered saline (PBS) for one hour to prevent nonspecific antibody binding. Immunohistochemical assays were performed using primary antibodies targeting IL-6 and TNF-α.

For immunofluorescence analysis, cells and sections were incubated with primary antibodies against COL-1 and α-SMA, followed by secondary antibodies. Images were acquired using an The Imager Z2 fluorescence microscope (Carl Zeiss) was utilized for the analyses. Quantitative assessments were conducted using ImageJ software to ensure accurate data interpretation.

Real-time quantitative polymerase chain reaction (qRT-PCR)

Total RNA was reverse-transcribed into complementary DNA (cDNA) utilizing the PrimeScript RT Reagent Kit (Takara, Japan). The synthesized cDNA was amplified with the Bio-Rad IQ5 Real-Time System using an Ultra SYBR Mixture (CWBIO, Beijing, China). The qRT-PCR protocol included an initial pre-denaturation step at 95 °C for 10 min, followed by 40 cycles of denaturation at 95 °C for 15 s and annealing at 60 °C for 1 min. A melting curve analysis was performed to confirm amplification specificity. Relative gene expression levels were calculated using the 2 − ΔΔCT method. For microRNA analysis, 800 ng of total RNA was reverse-transcribed into cDNA using the miRNA First Strand cDNA Synthesis Kit (Sangon Biotech, China). Primer sequences used in this study are listed in Table S1.

Preparation of single-cell suspensions

Single-cell suspensions of mouse skin tissue were prepared at 0, 1, 7, 14, and 21 days after GFP-labeled mouse ADSCs treatment. Mouse was sacrificed, and skin tissues were enzymatically digested using the Skin Dissociation Kit (Seekone, China), following the manufacturer’s instructions. The digested samples were subjected to filtration through 70 μm and 40 μm strainers to eliminate debris. Erythrocytes were lysed using a specialized red blood cell lysis buffer. Flow cytometry was then employed to sort GFP-positive resident skin cells from GFP-negative ADSCs for downstream analysis. GFP + cells were gated as follows: (1) FSC-A/SSC-A to exclude debris; (2) FSC-H/FSC-A to select singlets; (3) GFP fluorescence (488 nm excitation) with thresholds set against GFP − controls. To verify ADSC identity, sorted GFP + cells were stained with anti-CD29-FITC, anti-CD44-PE, anti-CD90-APC, anti-CD34-PE, and anti-CD45-PE (BioLegend).

RNA sequencing and analysis

Total RNA was extracted from tissues and cells utilizing TRIzol® Reagent in accordance with the manufacturer's instructions. The quality of the RNA was evaluated using the 5300 Bioanalyzer (Agilent Technologies, USA) and quantified with a NanoDrop ND-2000 spectrophotometer (Thermo Fisher Scientific, USA). Only high-quality RNA samples were selected for sequencing library preparation, meeting the following criteria: OD260/280 = 1.8–2.2, OD260/230 ≥ 2.0, RNA Quality Number (RQN) ≥ 6.5, a 28S:18S ratio of ≥ 1.0, and an amount greater than 1 μg.

RNA purification, reverse transcription, library construction, and sequencing were conducted by Shanghai Majorbio Bio-Pharm Biotechnology Co., Ltd. (Shanghai, China), adhering to Illumina’s protocol (Illumina, San Diego, CA, USA).

The RNA-seq transcriptome library was prepared using the Illumina® Stranded mRNA Prep Ligation protocol with 1 μg of total RNA. Messenger RNA was isolated via polyA selection with oligo(dT) beads and fragmented in a fragmentation buffer. Double-stranded cDNA synthesis utilized the SuperScript cDNA Synthesis Kit (Invitrogen, USA) with random hexamer primers. The cDNA underwent end-repair, phosphorylation, and adenylation according to Illumina’s library construction protocol. Libraries were size-selected for ~ 300 bp cDNA fragments on 2% Low Range Ultra Agarose and amplified by PCR using Phusion DNA Polymerase (NEB, USA) for 15 cycles. After quantification with Qubit 4.0, paired-end sequencing was conducted on the NovaSeq X Plus platform (2 × 150 bp read length).

Raw paired-end reads were quality-checked and trimmed using FastQC with default parameters. Clean reads were aligned to the reference genome with HISAT2. Transcript assembly was performed using StringTie in a reference-based manner. Differential expression analysis was conducted with DESeq2 or DEGseq, applying thresholds of |log2FC|≥ 1 and FDR ≤ 0.05 (DESeq2) or FDR ≤ 0.001 (DEGseq). Functional enrichment analysis, including Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG), was carried out using Goatools and KOBAS, respectively. Principal component analysis, clustering, heat map rendering, and time series analysis were performed on the Majorbio Cloud Platform.

ADSCs culture and treatments

ADSCs were cultured in DMEM/F12 medium supplemented with 10% FBS and 1% penicillin/streptomycin. For experiments, cells were serum-starved for 12 h before treatment. TGF-β treatment: Cells treated with 10 ng/mL recombinant human TGF-β (PeproTech, USA)for 6–24 h. TGF-β + antioxidant/ROS inhibitor: Cells pretreated with 5 mM N-acetylcysteine (NAC/ROSi, MCE, China) for 1 h before TGF-β exposure. NOX4 inhibition: Cells treated with 10 μM GKT137831 (NOX4 inhibitor/NOX4i, MCE, China) before TGF-β stimulation. NF-κB inhibition: Cells pretreated with 10 μM BAY11-7082 (NF-κB inhibitor/NF-κBi, MCE, China) before TGF-β.

ROS measurement

Intracellular levels of reactive oxygen species (ROS) were quantified using 10 μM CellROX Green reagent (Thermo Fisher Scientific, USA). Fluorescence intensity was assessed through flow cytometry (FACSCalibur, BD Biosciences) or fluorescence microscopy (Olympus IX73).

Plasmid engineering

Plasmid assembly was conducted following established molecular biology protocols (Sangon Biotech, China).The coding sequence of NF-KB was amplified utilizing primers that incorporated EcoRI and HindIII restriction sites at the 5' and 3' termini, respectively. Following PCR amplification, the product was directionally cloned into corresponding restriction sites of the pCMV-Tag2B plasmid. For TSG-6 promoter analysis, a 2 kb genomic fragment spanning nucleotides 151,357,592–151,359,592 (relative to transcriptional initiation site) was amplified from human genomic DNA using primers with NheI and HindIII terminal sites. This fragment was subsequently inserted into the NheI/HindIII-digested pGL3-Basic reporter vector. Site-specific mutations in NF-KB binding regions of the TSG-6 promoter were introduced using the QuikChange II mutagenesis system. All constructs underwent verification through Sanger sequencing.

Transfection protocol

Cells were seeded in 24-well plates at a density of 1 × 10⁵ cells per well and allowed to adhere for a period of 12 to 24 h. Transient co-transfections were conducted using Lipofectamine 2000 reagent (Invitrogen), incorporating 0.6 μg of expression vectors, 0.18 μg of promoter-reporter constructs, and 0.02 μg of the pRL-TK normalization plasmid. Following 6-h incubation, transfected cells were washed and maintained overnight in a complete medium containing 1% FBS. After 48-h serum starvation, cells were processed for subsequent analyses.

Dual-luciferase detection

Luciferase activity quantification was conducted using the Dual Luciferase Reporter System (Promega) in accordance with the manufacturer's guidelines. Cell lysates were prepared utilizing the provided lysis buffer and subsequently clarified through brief centrifugation. Luminescence measurements were obtained using a Modulus TM TD20/20 Luminometer (Turner Biosystems), with Renilla luciferase activity serving as an internal control for normalization purposes.

Isolation of LoS-derived fibroblasts

Skin tissues were obtained from 12 patients with localized scleroderma at Peking Union Medical College Hospital. Informed written consent was obtained from all participants, and the study received approval from the peking union medical college hospital's ethics committee (Approval No. ZS-2561). Dermal sections of LoS skin tissues were minced and cultured utilizing a tissue block explant method. Fibroblasts were cultured in DMEM supplemented with 10% fetal bovine serum (FBS), 100 U/mL penicillin, and streptomycin at 37 °C in a humidified atmosphere with 5% CO₂. Fibroblasts at passages 3–5 were used for subsequent experiments.

Virus transfection

shTSG-6 virus vector and its negative control were designed and synthesized (Sangon Biotech, China). According to the manufacturer's instructions, ADSCs were transfected with the shTSG-6 virus at a multiplicity of infection (MOI) of 40–50 per cell, utilizing 5 mg/L polybrene (Sangon Biotech, China). The culture medium was replaced 12 h post-transfection, and after an additional 48 h, puromycin was applied to select for positive cells (ADSC^shTSG−6&ADSC^NC).

Co-culture of human LoS-derived fibroblasts and ADSCs

For transwell co-culture, LoS-derived fibroblasts were seeded into six-well plates. The following day, 0.4-μm-pore Corning transwell inserts (Sigma-Aldrich, USA) containing 2 × 10⁵ ADSCs, ADSC^shTSG−6, and ADSC^NC were placed into the plates. Co-culture durations were 5, 24, or 48 h.

Western blotting

Proteins were extracted from fibroblasts and skin tissues using RIPA lysis buffer (Beyotime, China) with a protease inhibitor cocktail. Protein samples (50 μg) were separated by SDS-PAGE (10%) and transferred to PVDF membranes (Millipore, USA) at 100 V for 40–100 min. Membranes were blocked with 5% non-fat dry milk in TBST for 3 h at room temperature, followed by overnight incubation with primary antibodies at 4 °C (NOX4, p-IκBa, IκBa, NF-κb, TSG-6, TGF-β, p-Smad2/3, Smad2/3, Lamin A/C, GAPDH, Abcam). Secondary antibody incubation was performed at 37 °C for 1 h. Protein bands were visualized using an ECL kit and imaged on a FluorChem FC system (Tanon 5200, China). Protein expression levels were quantified and normalized using ImageJ software.

Statistical analysis

Data are presented as the mean ± standard error of the mean (SEM) derived from a minimum of three independent experiments. Statistical analyses were performed using SPSS version 17.0 software. Differences between groups were evaluated utilizing one-way analysis of variance (ANOVA) for multiple comparisons or the Student’s t-test for comparisons between two groups. A P-value of less than 0.05 was deemed statistically significant.

Results

ADSCs mitigate skin fibrosis by reducing ECM deposition and inflammation

There was no fluorescence signal in the control group and the model group before ADSCs injection. On the first day post-injection, ADSCs were primarily located in the subcutaneous and dermal layers. By day 7, ADSCs had begun to infiltrate the epidermis, and by day 14, their numbers had decreased, but they were distributed throughout the entire skin layer with a significant enrichment in hair follicles and perivascular areas. By day 21, the number of ADSCs had significantly diminished across all skin layers (Fig. 2A).

Fig. 2.

ADSCs mitigate BLM-induced skin fibrosis and inflammation in vivo. A Spatial distribution and temporal dynamics of GFP-labeled ADSCs within skin layers post-injection (days 1, 7, 14, 21) visualized by fluorescence imaging. Scale bar = 100 µm. B, C Representative H&E/Masson's trichrome staining of dorsal skin sections showing dermal thickness and collagen deposition and quantification of dermal thickness and collagen density (mean ± SEM, n = 6 mice/group; *p < 0.05 vs. PBS group; Scale bar = 100 µm). D qRT-PCR analysis of pro-inflammatory cytokine (TNF-α, IL-6, IL-1β) mRNA levels in skin tissues (mean ± SEM, *p < 0.05, vs. PBS group; one-way ANOVA)

Histopathological analysis of skin samples showed that, following BLM induction, the dermis thickened, and collagen fibers became increasingly dense and disorganized. α-SMA levels also increased, indicating the presence of more myofibroblasts. When ADSCs were injected subcutaneously, minimal changes were observed on the first day. However, by days 7, 14, and 21, there was evidence of reduced fibrosis, indicated by dermal thinning, a decrease in collagen density and α-SMA, and a more orderly collagen fiber arrangement (Fig. 2B, C). Similarly, quantitative real-time PCR showed that the proinflammatory cytokines tumor necrosis factor-alpha(TNF-α), interleukin-6 (IL-6), and interleukin-1 beta (IL-1β) were significantly upregulated following BLM injection and downregulated in the BLM + ADSCs group mouse skins at days 7, 14 and 21 days (Fig. 2D). These findings indicate that the administration of ADSCs resulted in a reduction of extracellular matrix (ECM) deposition in the skin, while also alleviating fibrosis and inflammation in the mouse bleomycin (BLM) model. However, an interesting phenomenon is that there was an explosive increase in inflammatory factors after the successful modeling, a sharp drop on the seventh day after intervention, and no significant change by the 14th day, but the fibrosis index decreased with a uniform and sustaining rate for 21 days.

Distinct transcriptomic profiling of ADSCs during in vivo fibrosis intervention

To investigate the effects of ADSCs within a skin fibrosis environment, we isolated GFP + cells from the skin tissue cell suspensions of the BLM modeling group using flow cytometry (Fig. 3A). Immediately post-sort, an aliquot was re-analyzed, confirming > 99% GFP + purity.Cells were analyzed on a BD LSRFortessa, with > 95% expressing CD29/CD44/CD90 and < 2% CD34/CD45 (Figure S1). We subsequently performed RNA sequencing to analyze the gene expression profiles of these GFP + cells at days 0, 1, 7, and 14. Flow cytometry analysis demonstrated a gradual decline in the number of GFP + cells over time, with significant reductions observed on days 1, 7, and 14. By day 21, the viability of ADSCs in vivo had declined to the extent that it was difficult to isolate sufficient cells for RNA-seq, so only data from days 1, 7, and 14 were included in the sequencing analysis. The remaining skin specimens were analyzed histologically.

Fig. 3.

Temporal transcriptomic profiling of ADSCs during in vivo fibrosis intervention. A Flow cytometry gating strategy for isolating GFP + ADSCs from skin cell suspensions at indicated time points. B Principal component analysis (PCA) of RNA-seq data from GFP + ADSCs isolated at days 0 (D0, pre-injection), 1, 7, and 14 post-injection into BLM-treated mice. C Unsupervised hierarchical clustering heatmap of significantly differentially expressed genes (DEGs) across time points. D Volcano plots depicting DEGs (|log2FC|≥ 1, FDR ≤ 0.05) for D1 vs D0, D7 vs D0, and D14 vs D0 comparisons. Red: Upregulated; Blue: Downregulated. E Significantly enriched GO & KEGG pathways in ADSCs at days 1, 7, and 14 post-injection (FDR ≤ 0.05)

Principal component analysis (PCA) revealed that the treatment groups formed distinct clusters, as shown in Fig. 3B. Despite this clustering, there was a notable overlap between the control group and the Day 1 group, as reflected in the unsupervised clustering heatmap (Fig. 3C). This cluster analysis highlighted significant transcriptomic differences across the four treatment groups.

We performed a comprehensive analysis of differentially expressed genes (DEGs). The volcano plot comparing D1 to D0 revealed a relatively small number of both upregulated and downregulated gene, similar to the results of PCA analysis. In contrast, the volcano plots on days 7 and 14 showed more differential genes, 2127 and 1720, respectively (Fig. 3D).

Gene set enrichment analysis was performed, and the findings were correlated with Gene Ontology (GO) terms. On day 1 post-intervention, several genes sets related to response to oxidative stress were upregulated. These included pathways involved in the negative regulation of axonogenesis, regulation of calcium ion transmembrane transport, and the cellular response to decreased oxygen levels. The KEGG pathway analysis conducted on day 1 demonstrated significant enrichment in several pathways, including the HIF-1 signaling pathway, oxidative phosphorylation, MAPK signaling, and central carbon metabolism in cancer (Fig. 3E).

By day 7, immunomodulatory pathways became more prominent. These included the regulation of interleukin-6, cellular response to interferon-beta, NF-κB signaling transduction, and regulation of endothelial cell proliferation. Noteworthy pathways enriched at this stage were cytokine-cytokine receptor interaction, pathways in cancer, IL-17 signaling pathway, and complement and coagulation cascades,

On day 14, the GO analysis revealed the enrichment of pathways associated with collagen fibril organization, blood coagulation fibrin clot formation, blood vessel morphogenesis, and branching morphogenesis. KEGG analysis on day 14 identified enrichment in pathways related to ECM-receptor interaction, complement and coagulation cascades, PPAR signaling, and pancreatic secretion.

Time-dependent pathway remodeling in ADSCs during fibrotic resolution

To delineate temporal patterns of ADSC-induced gene modulation during fibrotic remodeling, we conducted a comprehensive time-series analysis across Days 1, 7, and 14 (Fig. 4). Systematic identification of all potential expression trajectories revealed six predominant trends, categorized into upregulated and downregulated clusters. Subsequent functional characterization encompassed GO/KEGG enrichment analysis of trend-associated genes, with critical regulators visualized through heatmap analysis.

Fig. 4.

Expression trends of ADSC genes across days 1, 7, and 14 post-injection. Six predominant expression trajectories were identified by time-series analysis. The heatmap shows representative genes within each trend cluster. Significantly enriched GO terms/KEGG pathways associated with each major upregulated or downregulated trend are listed

Through temporal analysis, it was observed that stem cells activated the ROS pathway as early as the day after entering the organism. Immune regulation-related pathways became activated on day 7, accompanied by upregulated gene expression of multiple immune regulatory factors. From day 7 onward, sustained elevation was observed in both self-fibrosis markers and anti-fibrotic secretory factors. Among them, Tsg-6 and Cxcl14 have been increasing continuously since the first day. Continuous downregulation patterns correlated with immune response attenuation and environmental sensing mechanisms. Autoimmune pathways (Rheumatoid arthritis) and pathogen defense mechanisms demonstrated significant enrichment, with characteristic genes such as Ccl2, Tlr4, and Prkcb being identified. Pathways associated with developmental regulation and positive cellular modulation were also affected, indicating that the harsh fibrotic environment still impacted cell survival. Based on the above findings, we hypothesize that when ADSCs enter the fibrotic microenvironment, they may induce an oxidative stress response. Elevated ROS levels likely activate the NF-κB pathway, thereby stimulating the secretion of TSG-6 to exert anti-fibrotic effects. Therefore, cell and animal experiments were conducted to verify our conjecture.

Mechanistic validation of the ROS-NF-κB-TSG-6 regulatory axis

Our findings delineate a ROS-NF-κB-TSG-6 signaling axis activated by TGF-β through sequential molecular events. TGF-β initiates oxidative stress via coordinated NOX4 induction and endogenous antioxidant system suppression, resulting in sustained ROS accumulation that is pharmacologically reversible by NAC (Fig. 5A, B). This establishes NOX4 as the primary enzymatic source of TGF-β-induced oxidative damage, with ROS generation mechanistically linked to downstream signaling rather than being a bystander effect. The redox imbalance activates canonical NF-κB signaling through IκBa phosphorylation, triggering NF-κB nuclear translocation—a critical step evidenced by both TGF-β inducing nuclear NF-κB accumulation, while NAC pretreatment reduced nuclear localization to baseline levels (Fig. 5C). Western blot analysis confirmed TGF-β-induced upregulation of TSG-6 protein expression, which was abolished by pretreatment with ROS inhibitor (ROSi), NOX4 inhibitor (NOX4i) or NF-κB inhibitor (NF-κBi). This establishes TSG-6 as a downstream effector of ROS/NF-κB signaling (Fig. 5D). Through motif analysis, we identified a motif closely associated with NF-κB occupation (Fig. 5E). The schematic analysis revealed the presence of a canonical NF-κB binding site located within the promoter region of TSG-6 (Fig. 5F), proximal to the transcription initiation site, providing a structural basis for NF-κB-dependent transcriptional regulation. Dual-luciferase reporter assays showed TGF-β-induced activation of the TSG-6 promoter. Mutations in the NF-κB binding site significantly reduced promoter activity, functionally confirming NF-κB as the primary transcriptional regulator of TSG-6 (Fig. 5G).

Fig. 5.

Mechanistic validation of the ROS-NF-κB-TSG-6 regulatory axis (A) Intracellular ROS levels in ADSCs measured by flow cytometry after TGF-β stimulation with ROS inhibitor NAC. Western blot analysis confirming (B) TGF-β-induced ROS accumulation and its inhibition by NAC. (C) TGF-β-induced nuclear translocation of NF-κB in ADSCs and its suppression by NAC pretreatment. D TSG-6 protein expression in ADSCs treated with TGF-β ± inhibitors of ROS, NOX4, or NF-κB. GAPDH: loading control. Data are representative of ≥ 3 independent experiments (mean ± SEM; *p < 0.05, one-way ANOVA). E Sequence logo of the identified NF-κB binding motif. F Schematic diagram of the human TSG-6 promoter region showing the location of the predicted NF-κB binding site. G Dual-luciferase reporter assay. ADSCs co-transfected with wild-type or NF-κB binding site mutant TSG-6 promoter-luciferase constructs and treated with TGF-β. Firefly luciferase activity normalized to Renilla luciferase (mean ± SEM, n = 3 independent transfections, *p < 0.05)

These results collectively demonstrate that TGF-β exploits NOX4-derived ROS as second messengers to activate NF-κB, which directly binds the TSG-6 promoter to drive its expression. The production in promoter activity upon κB site mutation suggests this element is necessary for majority transcriptional activation, while residual activity may involve auxiliary redox-sensitive factors. The differential sensitivity of TSG-6 induction to pharmacological inhibitors reflects the signaling hierarchy, with NF-κB acting downstream of ROS. Importantly, the dual experimental approaches using both TGF-β establish the generalizability of the ROS-NF-κB-TSG-6 axis beyond specific TGF-β contexts.

TSG-6-mediated suppression of TGF-β/Smad signaling underlies ADSCs’ antifibrotic function

We also investigated the antifibrotic mechanism of TSG-6 secreted by ADSCs in a scleroderma-like fibrotic microenvironment by establishing an in vitro fibrosis system (Fig. 6A). To dissect the functional role of TSG-6, we employed a lentivirus-mediated gene silencing strategy, generating TSG-6-deficient ADSCs (ADSC^shTSG−6). Successful knockdown was confirmed at transcriptional levels, with significant reductions compared to negative control ADSCs (Fig. 6B). By co-culturing healthy human ADSCs with LoSFs isolated from localized scleroderma patients, we observed a progressive upregulation of TSG-6 gene expression in ADSCs under fibrotic stimulation (Fig. 6C). This dynamic response suggests that TSG-6 acts as a stress-responsive mediator, potentially enabling ADSCs to counteract pathological signals in fibrotic niches.

Fig. 6.

TSG-6 is essential for ADSC-mediated suppression of TGF-β/Smad signaling and fibrosis in vitro. A Schematic of the transwell co-culture system. LoS-derived fibroblasts (LoSFs) were co-cultured with ADSC, ADSC^NC or ADSC^shTSG−6. B qRT-PCR confirming TSG-6 knockdown efficiency in ADSCs. C Elisa's level of TSG-6 in ADSCs co-cultured with LoSFs. D Western blot analysis of TGF-β, p-Smad2/3, and Smad2/3 protein levels in LoSFs after co-culture with ADSCs. E Representative images of Col1 and α-SMA immunofluorescence staining in LSFs; the histogram demonstrates the relative immunofluorescence level (mean ± SEM, n = 3;*p < 0.05; **p < 0.01; ns: no differences)

Co-culture experiments with lesional scleroderma fibroblasts revealed that wild-type ADSCs and non-targeting shRNA controls substantially attenuated fibrotic progression, demonstrating reduced TGF-β signaling intensity and downstream Smad2/3 phosphorylation compared to PBS-treated disease baselines, while TSG-6-depleted ADSCs exhibited striking attenuation of this protective capacity (Fig. 6D). Meanwhile, elevated collagen I/α-SMA expression levels in the shTSG-6 group significantly exceeded those in ADSC and ADSC^NC interventions but remained moderately lower than PBS controls, collectively confirming TSG-6 as the indispensable molecular mediator enabling ADSCs to disrupt the fibrotic cascade through TGF-β/Smad modulation and extracellular matrix regulation (Fig. 6E).

Results of the animal model collectively validate the therapeutic efficacy of TSG-6-depleted ADSCs in ameliorating bleomycin-induced skin fibrosis in vivo(Fig. 7A). In the established fibrosis model, treatment with TSG-6-knockdown ADSCs (ADSC^shTSG−6) dramatically reversed therapeutical changes compared to PBS controls, wild-type ADSCs (ADSC), and non-targeting shRNA controls (ADSC^NC). Histopathological analysis (Fig. 7B, C) demonstrated that treatment with ADSC and ADSC^NC effectively normalized dermal architecture, significantly reducing bleomycin-induced dermal thickening and collagen deposition. However, the application of ADSCshTSG-6 diminished the anti-fibrotic effect.

Fig. 7.

Depletion of TSG-6 reduced the therapeutic efficacy of ADSCs in BLM-induced skin fibrosis in vivo. A Experimental design timeline. Mice received BLM injections for 4 weeks, followed by a single subcutaneous injection of PBS, ADSCs, ADSC^NC, or ADSC^shTSG−6. Tissues were harvested 21 days post-cell injection. B Representative H&E and Masson's trichrome staining of dorsal skin sections. Scale bar = 100 µm. C Quantification of dermal thickness and collagen density (mean ± SEM, n = 6 mice/group; *p < 0.05). D Representative images of Col1 and α-SMA immunofluorescence staining in skin (E) Quantification of COL-1 and α-SMA fluorescence intensity (mean ± SEM, n = 6; *p < 0.05)

Histopathological analysis (Fig. 7B, C) revealed that treatment normalized dermal architecture, reducing bleomycin-induced dermal thickening and collagen deposition, however, ADSC^shTSG−6 reduces the anti-fibrotic effect.

The bleomycin-induced skin fibrosis model also revealed that ADSC and ADSC^NC significantly attenuated α-SMA and collagen I expression, while ADSC^shTSG−6 abolished this therapeutic effect (Fig. 7D, E).

These results unequivocally establish TSG-6 as the key effector molecule driving fibrotic matrix remodeling downstream of the TGF-β/ROS/NF-κB axis and validate targeted TSG-6 in ADSCs as a superior therapeutic strategy.

Critically, the loss of TSG-6 not only exacerbated extracellular matrix (ECM) remodeling but also amplified pro-inflammatory signaling, as evidenced by upregulated IL-6 and TNF-α expression in immunohistochemical assays (Fig. 8A). This dual dysregulation—enhanced fibrogenesis coupled with unresolved inflammation—highlights TSG-6’s multifunctional role in disrupting the “fibrosis-inflammation” feedback loop.

Fig. 8.

TSG-6 deficiency in ADSCs restores inflammation and TGF-β/Smad signaling in fibrotic skin. A Representative immunohistochemical staining and quantification of IL-6 and TNF-α expression in skin sections B Western blot analysis and quantification of TGF-β protein levels and Smad2/3 phosphorylation ratio in skin tissues (mean ± SEM, n = 3;*p < 0.05; **p < 0.01; ns: no differences)

Mechanistically, Western blot analyses demonstrated that TSG-6 silencing restored TGF-β/Smad pathway activation, which was otherwise suppressed in ADSC and ADSC^NC groups. Specifically, TGF-β protein levels and the phosphorylation ratio of Smad2/3 were significantly reduced in ADSC-treated groups compared to controls and ADSC^shTSG−6 (Fig. 8B). These data align with the proposed model wherein ADSC-derived TSG-6 directly interferes with TGF-β/Smad signaling, thereby blocking downstream fibrogenic transcription. The restoration of TGF-β/Smad activity in TSG-6-deficient ADSCs underscores TSG-6 as a non-redundant inhibitor of this central profibrotic pathway.

Discussion

Fibrotic skin disorders, whether resulting from dysregulated wound healing—such as hypertrophic scars and keloids—or stemming from metabolic and immune-related conditions like scleroderma, exhibit several common pathological hallmarks. These include excessive fibroblast proliferation, abnormal overproduction of extracellular matrix, and subsequent loss of skin elasticity [9].

Scleroderma, a classic example of a fibrosing connective tissue disorder, manifests in two primary forms: systemic sclerosis and localized scleroderma. Systemic sclerosis can be life-threatening due to multi-organ involvement, including the lungs and kidneys, while localized scleroderma is confined to the skin and does not impact life expectancy. Despite their differences, both forms present with significant cutaneous manifestations that often result in pain, physical deformities, and psychological burden. However, only systemic sclerosis directly compromises internal organ function and life prognosis [10, 11]. The hallmark pathological features of scleroderma include. Microangiopathy: Damage to small blood vessels leading to vascular dysfunction. 2. Autoimmune dysregulation: Characterized by autoantibody production and T-cell activation. 3. Skin fibrosis: Excessive ECM deposition, primarily driven by the overproduction of type I and type III collagen fibers. In the inflammatory fibrotic response, the overproduction of collagen fibers results in the formation of a dense, waxy fibrotic matrix in the dermis, contributing to the progressive stiffness and thickening of the skin [1]. There is an urgent need for an integrated therapeutic approach that addresses epigenetic and genetic abnormalities, vascular issues, immune system defects, and the progression of fibrosis.

ADSCs and their secreted factors are thought to play a crucial role in the observed anti-fibrotic and immunomodulatory effects associated with the treatment of fibrotic skin in patients with scleroderma. In vitro models utilizing indirect co-culture of ADSCs or employing ADSC-conditioned media on cells derived from fibrotic tissue have demonstrated significant anti-inflammatory effects, a reduction in the production of extracellular matrix (ECM) components, and inhibition of fibroblast activation. [6–12].

The survival of ADSCs within the scleroderma skin microenvironment is a critical factor influencing their therapeutic efficacy. The fibrotic milieu in scleroderma is characterized by excessive ECM deposition, hypoxia, and chronic inflammation, which collectively create a hostile environment that may impair ADSC viability and function [8–13]. Compared to previous studies on ADSCs in healthy subcutaneous environments, our fluorescence imaging and flow cytometry results indicate a shortened survival time of ADSCs in the scleroderma microenvironment. By day 21, only a minimal number of surviving cells could be observed. Interestingly, our findings also suggest that ADSCs exhibit a degree of localization capability. Fluorescence imaging showed that ADSCs tend to aggregate around hair follicles and blood vessels. This observation aligns with previous studies, which have highlighted the potential effects of ADSCs on hair follicles and blood vessels in fibrotic environments [14, 15].

The RNA-seq analysis uncovered a choreographed transcriptional reprogramming of ADSCs across fibrotic resolution. Day 1 post-transplantation was dominated by oxidative stress pathways, reflecting ADSCs’ encounter with a hostile microenvironment rich in TGF-β. This early oxidative signature likely primes ADSCs for survival, as evidenced by upregulated NOX4 and antioxidant genes, mirroring ADSCs adaptation to ischemic niches [16]. Intriguingly, the transient ROS surge may serve a dual role: (1) activating cytoprotective autophagy via HIF-1α [17], and (2) triggering NF-κB-dependent cytokines secretion to initiate antifibrotic signaling.

By day 7, the transcriptome shifted toward immunomodulation, coinciding with peak Tsg-6 expression and maximal fibrosis reduction. The enrichment of NF-κB and IL-6 at this stage supports a model where ADSCs transition from stress adaptation to active immune modulation. However, by the seventh day, the immunomodulatory effects of ADSCs became increasingly evident, aligning with findings from previous studies. Traditionally regarded as possessing immune privilege due to their classification as mesenchymal stromal cells, ADSCs have been observed to provoke both cellular and humoral immune responses in vivo when injected into a foreign organism. This phenomenon may contribute to the rapid clearance of transplanted cell [18]. Nevertheless, ADSCs primarily operate through a "hit-and-run mechanism", exerting a limited impact on therapeutic efficacy, particularly in the short or medium term. Most ADSCs function via paracrine signaling, releasing cytokines, growth factors, and extracellular microvesicles into the local environment, rather than requiring direct cell-to-cell contact [19, 20]. The active substances secreted by ADSCs, including cytokines and extracellular vesicles, play a role in regulating the microenvironment surrounding fibroblasts and the ADSCs themselves [21–23].

Previous studies have demonstrated that following injection, ADSCs elicit adaptive cellular responses, which result in the secretion of IL-1, prostaglandin E2 (PGE2), IL-4, IL-10, and TGF-β. These secreted factors modulate and activate innate immune cells. Furthermore, ADSCs release various immunosuppressive factors, including nitric oxide, PGE2, hepatocyte growth factor, and indoleamine 2,3-dioxygenase. These factors downregulate TGF-β in skin fibrosis and recruit bone marrow-derived cells that are involved in tissue repair [24, 25]. Transcriptomic analysis revealed a significant upregulation of Nrf2, Il-10 Tsg-6, xcl14, Alpl, and Mmp-9 inflammation or fibrosis-associated secretory proteins (p < 0.05). Pathway verification through gene enrichment analysis aligned with established immunological mechanisms, particularly in immune cell regulation. Current evidence indicates mesenchymal stem cells ameliorate autoimmune-associated dermal fibrosis through multifaceted immunomodulation. As documented in sclerosis pathogenesis, MSCs maintain cutaneous homeostasis through three principal mechanisms: (i) suppression of CD4⁺ T lymphocyte and macrophage proliferation/chemotaxis; (ii) enhancement of regulatory T cell (Treg) differentiation and activation; (iii) negative regulation of B cell-mediated immune responses [5]. These mechanistic insights collectively advance our understanding of MSC-based immunotherapeutic strategies for inflammatory dermatoses [26].

Day 14 marked a shift toward ECM-receptor interaction and collagen remodeling, consistent with histological evidence of organized fibril deposition. By day 14, a temporal shift in therapeutic mechanisms was observed, characterized by attenuated immunomodulatory effects and predominant activation of anti-fibrotic pathways. In fibrotic pathologies, dermal fibroblasts undergo a phenotypic transition to myofibroblasts, marked by elevated α-SMA expression and enhanced contractility [27]. Experimental evidence highlights ADSCs as rich sources of anti-fibrotic mediators, including IL-10, adrenomedullin, and HGF. Notably, HGF exerts dual therapeutic effects by inhibiting myofibroblast differentiation and suppressing profibrotic activity through collagen synthesis downregulation and MMP induction [28, 29]. These mechanisms align partially with transcriptional profiling data from day 14 specimens. Current consensus identifies TGF-β1 suppression as a hallmark of ADSC-mediated anti-fibrotic efficacy, establishing it as a principal biomarker for therapeutic monitoring [30, 31]. ADSC-conditioned medium demonstrates consistent inhibition of collagen synthesis and downregulation of fibrogenic markers [32]. Paradoxically, in co-culture systems with human dermal fibroblasts, ADSCs enhance collagen I/III/VI deposition within the ECM [33], suggesting context-dependent modulation of fibroblast homeostasis. This dichotomy implies ADSCs preferentially target pathological fibroblasts to mitigate aberrant ECM accumulation. Transcriptomic analysis at Day 14 revealed significant enrichment of ECM-related pathways including 'collagen fibril organization' and 'blood vessel morphogenesis'. Importantly, these terms reflect reparative ECM remodeling rather than pathological fibrosis, as evidenced by: (1) concomitant downregulation of pro-fibrotic markers (COL1A1, α-SMA) in recipient tissues; (2) temporal coordination with preceding immunomodulatory responses (Day 7 NF-κB/cytokine pathways); and (3) histological confirmation of organized collagen architecture). This phased transition from inflammation resolution to structured matrix reorganization suggests ADSCs orchestrate quality-over-quantity ECM restoration.

Transcriptomic screening of engraftment-upregulated genes identified TNF-α-stimulated protein 6 (TSG-6) as a pivotal mediator underlying the anti-fibrotic efficacy of ADSCs, functioning through multifaceted immunoregulatory mechanisms. Experimental evidence indicates mesenchymal stromal cells mitigate fibrotic progression via TSG-6 secretion, a pleiotropic glycoprotein that orchestrates tissue microenvironmental homeostasis through coordinated pathway modulation [34–37]. This inflammation-associated mediator demonstrates constitutive expression in high metabolic-demand tissues, suggesting intrinsic protective roles against environmental stressors [38, 39]. Mechanistically, TSG-6 exerts dual regulatory control over canonical fibrotic pathways: (1) Direct suppression of TGF-β/Smad signaling through Smad2/3 phosphorylation inhibition, as evidenced by attenuated TGF-β1-mediated keloid fibroblast proliferation and BMP-4/Smad2-driven osteogenic differentiation [36–40, 40–42]; (2) Cross-talk modulation between inflammatory cascades, notably through IRE1α/TRAF2/NF-κB axis blockade in hyperplastic dermal fibroblasts and TLR2/MyD88/NF-κB pathway inhibition in spinal microglia [43, 44]. Our experimental data corroborate these mechanisms, thereby establishing TSG-6 as a master regulator of fibroinflammatory resolution through multi-level signaling network reprogramming.

Mechanistic studies delineated a novel signaling hierarchy wherein TGF-β-induced ROS generated via NOX4 activated NF-κB to transcribe TSG-6, which subsequently inhibited TGF-β/Smad signaling. This negative feedback loop established TSG-6 as both a product and regulator of TGF-β activity, forming a self-limiting circuit to prevent fibrotic overdrive. Key aspects of this axis included NOX4 serving as the primary ROS source, with TGF-β simultaneously inducing NOX4 expression and suppressing antioxidant enzymes such as SOD2, thereby creating a redox imbalance critical for NF-κB activation. Functionally, TSG-6 disrupted the fibrosis-inflammation loop by dual mechanisms inhibiting TGF-β/Smad signaling and pro-inflammatory cytokine production. Collectively, these findings positioned TSG-6 at the intersection of fibrosis and inflammation, resolving their interdependence through multi-level signaling reprogramming as supported by previous literature [41–45].

However, key questions remain unresolved, including whether TSG-6 preferentially remodels specific collagen isoforms such as COL1A1 versus COL3A1, how ADSC-derived ROS levels spatially fluctuate within fibrotic niches, and the clinical relevance of TSG-6 expression levels in predicting ADSC treatment response. Future studies should prioritize spatial transcriptomics to map ADSC-ECM interactions and validate TSG-6’s therapeutic efficacy in humanized fibrosis models to address these gaps and advance translational applications.

The present study provides a comprehensive mechanistic elucidation of how ADSCs mitigate skin fibrosis through TSG-6-dependent regulation of the TGF-β/Smad pathway, mediated by ROS-NF-κB signaling axis. By integrating in vivo tracking, transcriptomic profiling, pathway analysis, and molecular validation, we demonstrate that ADSCs orchestrate a dynamic, multi-stage antifibrotic response characterized by temporally resolved activation of oxidative stress signaling, immunomodulation, and extracellular matrix remodeling. The biphasic action of ADSCs (early immunomodulation to late ECM remodeling) suggests optimal therapeutic outcomes may require timed booster injections, a strategy warranting clinical trial validation.

Conclusion

This study elucidates the dynamic therapeutic mechanisms of ADSCs in mitigating bleomycin-induced skin fibrosis, revealing a biphasic mode of action characterized by immune modulation and fibrotic resolution. The ROS-NF-κB-TSG-6 axis emerges as a linchpin mechanism, linking TGF-β-induced oxidative stress to anti-fibrotic reprogramming. By decoding the temporal hierarchy of ADSC responses—from oxidative adaptation to matrix remodeling—this work provides a roadmap for optimizing stem cell therapies in fibrotic diseases. As the field advances toward precision antifibrotic strategies, TSG-6 stands out as both a biomarker and a therapeutic beacon, highlighting its dual potential for guiding next-generation interventions in scleroderma and related disorders. Further exploration of ADSC-derived paracrine factors and their temporal regulation will refine strategies to enhance clinical outcomes.

Abbreviations

ADSCs

Adipose-derived stem cells

BLM

Bleomycin

LoS

Localized scleroderma

ECM

Extracellular matrix

GFP

Green fluorescent protein

a-SMA

Alpha-smooth muscle actin

COL1

Collagen type I

TNF-α

Tumor necrosis factor-alpha

IL-6

Interleukin-6

IL-1β

Interleukin-1 beta

qRT-PCR

Quantitative real-time polymerase chain reaction

RNA-seq

RNA sequencing

IHC

Immunohistochemistry

MOI

Multiplicity of infection

shRNA

Short hairpin RNA

ROS

Reactive oxygen species

NF-kB

Nuclear factor kappa-light-chain-enhancer of activated B cells

TSG-6

TNF-alpha-stimulated gene 6

NOX4

NADPH oxidase 4

NAC

N-acetylcysteine

ROSi

ROS inhibitor

NOX4i

NOX4 inhibitor

NF-kBi

NF-kB inhibitor

TGF-β

Transforming growth factor-beta

PBS

Phosphate-buffered saline

DMEM/F12

Dulbecco's Modified Eagle Medium/F-12

FBS

Fetal bovine serum

BSA

Bovine serum albumin

GO

Gene Ontology

KEGG

Kyoto Encyclopedia of Genes and Genomes

PCA

Principal component analysis

DEGs

Differentially expressed genes

H&E

Hematoxylin and eosin

SEM

Standard error of the mean

ANOVA

Analysis of variance

HIF-1

Hypoxia-inducible factor 1

MAPK

Mitogen-activated protein kinase

PPAR

Peroxisome proliferator-activated receptor

TLR

Toll-like receptor

MyD88

Myeloid differentiation primary response 88

IRE1α

Inositol-requiring enzyme 1 alpha

TRAF2

TNF receptor-associated factor 2

PGE2

Prostaglandin E2

HGF

Hepatocyte growth factor

MMP

Matrix metalloproteinase

SOD2

Superoxide dismutase 2

SPF

Specific pathogen-free

Author contributions

Liquan Wang and Tianhao Li: designed and performed experiments, prepared figures/tables, prepared first draft of the manuscript; Xuda Ma, Ziming Li, Jieyu Xiang, Songlu Tseng: advised on data analysis and interpretation; Nanze Yu, Jiuzuo Huang and Xiao Long: conceived/designed the study, analyzed/interpreted the data, edited manuscript; All authors have reviewed and approved the final manuscript.

Funding

This study was supported by grants from National High Level Hospital Clinical Research Funding (2025-PUMCH-D-001, 2022-PUMCH-C-025, 2022-PUMCH-A-210), National Natural Science Foundation of China (82472565, 82302828), Beijing Natural Science Foundation of China (L254044, L244061, L244062),Peking Union Medical College Hospital Talent Cultivation Program (Category C) No.UBJ11557, Plastic Medicine Research Fund of Chinese Academy of Medical Sciences (2024-ZX-1-02, 2024-ZX-1-03).

Data availability

All relevant data from this study, including the proteomic data, are available from the corresponding author upon request. The proteomic data will be deposited in a searchable, public database.

Declarations

Ethics approval and consent to participate

The study was approved by the institutional review board of peking union medical college hospital (Date:2024-01, Approval No.ZS-2561) and Animal Ethics Committee (Date:2024-07, Approval No.MDKN-2024-100). For studies involving human participants/tissues, written informed consent was obtained from all donors, and all procedures were performed in accordance with the ethical standards. All animal experiments complied with ARRIVE guidelines and were conducted under the institutional guidelines for animal welfare.

Competing interests

The authors declare no competing interests.