A combinatorial therapeutic platform based on graphene–collagen cryogel and apoptotic vesicles for modulating hypertrophic scars

Abstract

Graphene-based collagen hydrogels have demonstrated anti-fibrotic potential in various diseases; however, their therapeutic efficacy in hypertrophic scars (HS) remains largely unexplored. Adipose stem cells (ASCs) and their extracellular vesicles have been shown to regulate HS progression, yet the role of ASCs derived apoptotic vesicles (ASCs-ApoVs) has not been fully investigated. In this study, we developed graphene-incorporated type I collagen cryogel (G-GEL(C)) and collected ASCs-ApoVs. We evaluated their individual and combined capacity to regulate HS and explored the potential of G-GEL(C) as a delivery system for ASCs-ApoVs. G-GEL(C) significantly inhibited fibrosis in HS derived fibroblasts (HS-fibroblasts), as evidenced by the downregulation of COL1A1 (p < 0.001 for both protein and mRNA), α-SMA (p < 0.001 for protein; p < 0.0001 for mRNA), and Vimentin (p < 0.05 for protein; p < 0.0001 for mRNA). Additionally, G-GEL(C) suppressed cell proliferation (p < 0.0001) and lateral migration (p < 0.001). Treatment with ASCs-ApoVs also reduced COL1A1 (p < 0.01 for protein; p < 0.0001 for mRNA) and α-SMA (p < 0.05 for both protein and mRNA), while Vimentin transcription was also downregulated (p < 0.001). G-GEL(C), characterized by high porosity and selective adsorption capacity for ASCs-ApoVs, enabled efficient loading and delivery of these vesicles. In vivo, G-GEL(C) loaded with ASCs-ApoVs decreased the scar elevation index (SEI) (p < 0.05), reduced α-SMA expression (p < 0.0001), locally increased 8-OHdG levels and raised the proportion of M2 macrophages, indicating effective regulation of HS. ASCs-ApoVs exhibit anti-fibrotic effects in HS. G-GEL(C) functions both as a direct modulator of HS-fibroblast phenotypes and as an efficient carrier for apoptotic vesicles. Collectively, they form a potent combinatorial system that significantly attenuates hypertrophic scar.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1038/s41598-026-57477-3.

Introduction

Graphene, an allotrope of carbon, has emerged as a promising therapeutic agent for dermatology due to its exceptional biocompatibility¹, optoelectronic properties^2,3, and photothermal effects^4,5. Recent studies have shown that graphene and its derivatives exhibit prominent anti-fibrotic and immunomodulatory effects. For instance, graphene quantum dots (G-QDs, nanoparticles < 100 nm) can disrupt the cyclin-CDK-E2F axis, arrest the cell cycle, induce quiescence in myofibroblasts, and suppress their proliferation, migration, α-SMA expression, and collagen synthesis⁶. The observed modulation of fibroblast activity and anti-fibrotic pathways led us to postulate that this material holds promise for treating hypertrophic scars (HS).

However, the simple delivery of graphene is often limited by poor dispersion homogeneity and low systemic stability. Type I collagen, a natural polymer and a major component of human extracellular matrix, has been widely used in skin wound repair, corneal injury regeneration⁷ and so on. When employed as a delivery matrix for graphene, type I collagen hydrogel facilitates improved stability of the graphene delivery system⁸ and reduces direct graphene-cell contact, which may help mitigate potential cytotoxicity^9,10. Additionally, the graphene-type I collagen composite hydrogel has been reported to exhibit anti-fibrotic⁹, macrophage-modulating¹¹, cell alignment-inducing¹², and excellent biocompatibility properties. All characteristics collectively suggest that the type I collagen-graphene hydrogel may offer a more suitable platform for scar modulation compared to graphene alone.

In parallel, local transplantation of bone marrow mesenchymal stem cells (BM-MSCs) and adipose stem cells (ASCs) has demonstrated efficacy in scar treatment^13,14. However, several studies indicate that the majority of transplanted ASCs undergo apoptosis within days after implantation¹⁵, suggesting that their long-term therapeutic benefits may be mediated primarily through paracrine vesicles or necrotic debris rather than by the presence of living cells. ASCs derived apoptotic vesicles (ASCs-ApoVs) have been shown to promote cutaneous wound healing^16–18, pro-survival of ischemic flaps¹⁹, and hair follicle regeneration²⁰. Based on evidence above, we hypothesized that ASCs-ApoVs, as a specific type of extracellular vesicle (EV) released during ASCs apoptosis, could regulate the HS.

Despite their therapeutic potential, the efficient delivery and controlled release of EVs remain key challenges. Conventional delivery strategies often suffer from rapid vesicle clearance²¹, poor local retention, and uncontrolled release profiles, which limit their translational potential. Cryogel—a class of hydrogels formed under subzero temperatures—exhibit super-macroporous structures with pore diameters usually ranging from 10 to 200 μm²². Specifically, the type I collagen-graphene cryogel offers three synergistic advantages in ApoVs delivery. First, the graphene component selectively adsorbs ApoVs, firmly capturing and densely distributing them on the pore walls, thereby preventing rapid washout under fluid flow. Second, these µm pores allow cells to freely migrate into the scaffold²³ and directly contact the concentrated vesicles on the pore walls, significantly improving local utilization. Third, each pore could act as an independent vesicle reservoir; owing to local variations in cell arrival time and graphene distribution, the timing of vesicle release or uptake differs among pores, yielding a more sustained release profile than direct vesicle injection. Collectively, these features position the type I collagen-graphene cryogel as a promising delivery platform for therapeutic vesicles²⁴.

Compared to existing methods for HS regulation, the injection of cryogel loaded with ApoVs represents a novel therapeutic approach with fewer of the disadvantages commonly encountered in daily clinical practice. For instance, although intralesional steroid injection remains the first-line therapy for HS²⁵, it offers only temporary, non-specific suppression and is associated with considerable side effects, such as skin atrophy, pigmentation abnormalities, and telangiectasia²⁶. In contrast, the ASCs-ApoVs loaded type I collagen-graphene cryogel system could provide multi-targeted, sustained, and physiologically relevant modulation of scar-associated fibroblasts, allows host cell migration into the scaffold for direct vesicle interaction, and reduces the need for repeated dosing. These features position it as a promising next-generation therapy for HS, particularly in cases that are resistant to or intolerant of steroids.

Therefore, in this study, we fabricated a graphene-type I collagen cryogel (G-GEL(C)) and collected ASCs derived ApoVs (ASCs-ApoVs). We first investigated their respective anti-fibrotic effects in the context of HS, and further explored the feasibility of a combined delivery approach using ASCs-ApoVs loaded injectable G-GEL(C).

Materials and methods

Preparation of injectable ApoVs loaded G-GEL(C) gel

1. Graphene-type I collagen (G-GEL) gel: Purified rat tail collagen type-I (Corning^® 354236, 3–4 mg/ml in 0.02 N acetic acid; 1 ml) was neutralized to pH 7.4 (for crosslink and better cell contact), using 10× PBS (0.2 ml) and 0.1 N NaOH (approximately 0.1 ml) on ice. The solution was then incubated at 37 °C for 30 min for pre-crosslinking. 2.5/2.0/1.5/1.0 mg Graphene (Amethyst^® 966627; thickness: 0.55–3.74 nm, < 10 layers; specification in Supplementary Table 1) per 2 ml final system volume was dispersed in Milli-Q water by sonication in a water bath for 2 h at room temperature (RT). The dispersed graphene was subsequently added to the neutralized collagen solution and thoroughly mixed. The solution could be used to form the graphene-type Ⅰcollagen (G-GEL) hydrogel. Type Ⅰ collagen hydrogel (COL-GEL) without graphene was also fabricated at the same parameters as control.

2. Graphene-type I collagen cryogel (G-GEL(C)): 50mM EDC and 25mM NHS solution was first dissolved in adequate 0.1 M MES buffer (PH = 5.5) and then added to the reaction mixture mentioned above. After again neutralization (for crosslink and better cell contact), the whole system was filled into a 2 ml syringe and immediately kept at −20 °C for 15 h. The resultant cryogel was thawed in ultrapure Milli-Q water (1 ml/mg at least) for 1 h, freeze and finally lyophilized overnight. Type Ⅰ collagen cryogel (COL-GEL(C)) without graphene was also fabricated at the same parameters as control.

3. Injectable G-GEL(C) loaded with apoptotic vesicles (ApoVs): The lyophilized G-GEL(C) was mechanically crushed by manual grinding. The resulting powder was then mixed with PBS (1 ml/mg of dry gel) and subjected to vigorous vortexing/oscillation to break the fragments into smaller pieces. The obtained dispersion was drawn into a 1 mL syringe fitted with a 26-gauge needle (0.45 mm outer diameter, 0.26 mm inner diameter) to produce an injectable G-GEL(C) suspension.

To avoid disrupting the structure of ApoVs by vigorous shaking or vortexing, the injectability of the lyophilized material was first confirmed as described above. Subsequently, the injectable G-GEL(C) was lyophilized again and then immersed in a PBS solution containing ApoVs (1.0 µg/mL) at a volume of at least 1 mL per mg of dry gel. The immersion was carried out at 4 °C (to minimize premature release and preserve vesicle integrity) for at least 24 h with gentle shaking to finally obtain the injectable G-GEL(C) loaded with ApoVs.

Cell isolation and culture

Adipose-derived stem cells (ASCs). ASCs were isolated from liposuction tissues obtained from three donors (aged 18–48 years, with a BMI 18.5–30) at Shanghai Ninth People’s Hospital, following prior informed consent. The adipose tissue was first digested with 1 mg/mL Type I collagenase (Gibco™, 17100017)²⁷ in basic ASC culture medium (OriCell, HUXMD-90011; 37 °C, 2 h, shaking). The digested tissue was then filtered through a 70-µm strainer, centrifuged (1500 rpm, 5 min), and cultured in ASC culture medium, supplemented with 10% FBS (Gibco™, 10099141 C) and 1% penicillin/streptomycin (Gibco™, 15140122). ASCs from passages 2–4 were used for experiments.

Hypertrophic scar derived fibroblasts (HS-fibroblasts). Hypertrophic scar tissues were obtained from patients undergoing scar revision surgery at Shanghai ninth peoples’ hospital from 2023 to 2024, with written informed consent acquired (Number of donors: n = 5; Donor age range: 25–45 years old; Scar location: abdominal skin (1), neck skin (1), back skin (1), breast skin (1) and arm skin (1); Scar duration: 6–12 months post-injury). Epidermis was removed using 2 mg/mL DispaseII (Yeasen, 40104ES80; 4 °C, 12 h, shaking), followed by dermis dissociation. Dermis was then cut into small pieces and digested with 1 mg/mL Type I collagenase (37 °C, 2 h, shaking). The suspension was filtered (70 μm), centrifuged (1500 rpm, 5 min), and cultured in high-glucose DMEM (Gibco™, 11965092) supplemented with 10% FBS and 1% penicillin/streptomycin. HS-fibroblasts at passages 3–5 were used.

All cells were routinely passaged at 80–90% confluence using 0.25% trypsin-EDTA (Gibco™, 25200072; 37 °C, 3–5 min).

Apoptotic vesicles (ApoVs) collection and identification

1. ASCs-derived apoptotic vesicles (ASCs-ApoVs) collection: Cell apoptosis was induced in serum free ASC culture medium with 5µM staurosporine (STS; MCE, HY-15141; dissolved in DMSO and diluted in serum-free DMEM medium) for 12 h²⁸. With confirmation of apoptotic morphology under light microscopy, cells were trypsinized (37 °C, 2 min) and centrifuged at 300 g for 10 min to remove intact cells and debris. The supernatant was further centrifuged at 3,000 g for 20 min to collect ApoVs¹⁷, which were then washed twice with PBS to remove residual solvable STS and other soluble components. The protein concentration of cleaning ApoVs was quantified by BCA assay (Beyotime, P0012), adjusted to 1 mg/mL in PBS, and stored at −80 °C for long-term preservation. ASCs-ApoVs were added to the HS-fibroblasts’ culture medium at the designated proportion.

2. Shape and size of ASCs-ApoVs: Fresh ApoVs was characterized by transmission electron microscopy (TEM) and nanoparticle tracking analysis (NTA) to confirm their size and morphology.

3. Adequacy of apoptosis confirmation: Cytoskeleton staining (FITC-Phalloidin; Solarbio, CA1620; RT, 30 min) and flow cytometry of ASCs as parental cells of ApoVs were performed to demonstrate the adequacy of apoptosis. As for flow cytometry, ASCs were harvested through trypsinization (37℃, 2–3 min) after 12 h STS treatment, washed twice with ice-cold PBS, and subsequently resuspended in 1× Annexin V binding buffer (Sigma-Aldrich, APOAF-20TST) at a standardized density of 1 × 10^{^6} cells/mL for subsequent apoptosis analysis. Cell suspensions were then stained with 5 µL Annexin V-FITC (488 nm excitation/530 nm emission) and 5 µL propidium iodide (PI, 488 nm excitation/575 nm emission) for 15 min at room temperature in the dark.

4. Absorption of ASCs-ApoVs by HS-fibroblasts: a.DiO labeling of ASCs-ApoVs: ASCs were incubated with 5–10 µM DiO (MCE, HY-D0969; λex/λem = 484/501 nm) in culture medium for 20 min at 37 °C, washed, and then subjected to apoptosis induction (5 µM STS, 12 h) to generate DiO-labeled ASCs-ApoVs. b. CM-Dil labeling of HS-fibroblasts for subcellular localization: HS-fibroblasts’ membranous organelle was stained with CM-Dil (Invitrogen™, C7000; λex/λem = 553/570 nm) by incubating with 5 µL of labeling solution per 1 mL growth medium for 20 min at 37 °C, followed by washing. c. Cellular uptake assay: DiO-labeled ASCs-ApoVs were co-cultured with CM-Dil-pre-labeled HS-fibroblasts for 2 h at 37 °C. Nuclei were then counterstained with Hoechst 33,342 (MCE, HY-15559; λex/λem = 346/460 nm) for 15 min at room temperature. Confocal microscopy was used to visualize cytoplasmic internalization of ASCs-ApoVs, with bright-field imaging to delineate cell morphology.

5. Vehicle control of soluble STS To further rule out the potential interference from soluble STS in the ASCs-ApoVs solution, we collected the vehicle control from the apoptosis induction medium (serum-free HD-DMEM containing 5µM STS; 37 ℃, 12 h) using the same protocol as ASCs-ApoVs isolation.

Real-time quantitative reverse transcription PCR (RT-qPCR)

Total RNA was isolated from cells using Trizol™ reagent (Invitrogen™,15596018CN). RNA concentration and purity were determined by Nanodrop spectrophotometer. cDNA was synthesized from 1 µg RNA using reverse transcriptase (TaKaRa, RR036A). Reactions were performed in triplicate using Probe-based master mix (TaKaRa, RR820A) on a real-time PCR system. Detailed primer information is provided in Table 1: Primer information. GAPDH was used as the internal reference for data normalization to ensure reliable quantification. Relative gene expression was calculated using the 2^−ΔΔCt method.

Table 1.

Primer information.

Gene	Species	Forward	Reverse
COL1A1	Human	5’ AGG GCC AAG ACG AAG ACA TC 3’	5’ GTC GGT GGG TGA CTC TGA GC 3’
COL3A1	Human	5’ TGA AGG GCA GGG AAC AAC T 3’	5’ GGA TGA AGC AGA GCG AGA AG 3’
ACTA2	Human	5’ CGT GGC TAT TCC TTC GTT ACT A 3’	5’ ATC AGG CAA CTC GTA ACT CTT C 3’
VIM	Human	5’ CTG TAA GTT GGT AGC ACT GAG 3’	5’ TTA GGG GAA ACC GTT AGA C 3’
GAPDH	Human	5’ GGG AAG GTG AAG GTC GGA GT 3’	5’ GGG GTC ATT GAT GGC AAC A 3’

Western blot

Cells were lysed in RIPA buffer (Epizyme, PC101) with protease inhibitors (Thermo Scientific™, 78430; 4 °C, 30 min). After denaturation (100 °C, 10 min) with loading buffer (Epizyme, LT101), samples (20 µg/lane) were loaded onto the SDS-PAGE gel (FuturePAGE™, ET15420LGel; 4–20% 15 wells) and transferred to PVDF membranes (Amersham™, GE10600023) using a wet transfer system. Membranes were blocked with 5% non-fat milk (Epizyme, PS112), probed with primary antibodies (4 °C, overnight) and HRP-conjugated secondary antibodies (RT, 1 h), then detected with ECL (Epizyme, Omni-ECL™ SQ201). Protein bands were quantified using ImageJ (Fiji) software (National Institutes of Health, USA) with GAPDH as loading control. Detailed antibody information was provided in Tables 2 and 3.

Table 2.

Primary antibody information.

Primary antibodies	Catalog number
COL1A1 antibody	Proteintech 67288-1-Ig
COL3A1 antibody	ABclonal A0817
α-SMA antibody	Affinity AF1032
Vimentin antibody	CST #P08670
8-OHdG antibody	MCE HY-P81140
iNOS antibody	Proteintech18985-1-AP
CD163 antibody	Proteintech 68218-1-Ig
Caspase-3 antibody	Proteintech 19677-1-AP
GAPDH antibody	Proteintech 60004-1-Ig
β-actin antibody	Proteintech 66009-1-Ig

Table 3.

Secondary antibody information.

Secondary antibody	Catalog number
HRP-conjugated Goat Anti-Rabbit IgG	Proteintech SA00001-2
HRP-conjugated Goat Anti-Mouse	Proteintech SA00001-1
CoraLite488-conjugated Goat Anti-Rabbit IgG	Proteintech SA00013-2
CoraLite594-conjugated Goat Anti-Mouse IgG(H + L)	Proteintech SA00013-3
CoraLite594-conjugated Goat Anti-Rabbit IgG	Proteintech SA00013-4
Peroxidase AffiniPure® Goat Anti-Rabbit IgG (H + L)	Jackson 111-035-045

Immunofluorescence (IF)

Cells on coverslips were fixed with 4% PFA (Beyotime, P0099-100 ml; 30 min, RT), permeabilized with 0.1% Triton X-100 (Beyotime, ST1723-100 ml; 10 min, RT), and blocked with 5% BSA (Solarbio, SW3015; 30 min, RT). Primary antibody incubation was performed overnight at 4 °C, followed by fluorescent secondary antibodies (1 h, RT). Nuclei were stained with DAPI (MCE, HY-D1738), and samples were mounted for imaging using fluorescence microscopy. Detailed antibody information was provided in Tables 2 and 3.

For each image, Regions of Interest (ROIs) were randomly selected. The mean fluorescence intensity for each ROI was determined using ImageJ software. Data were collected from at least 30 cells per experimental group over the course of three biological replicates.

Cellular cytoskeleton staining

The HS-fibroblasts were separately cultured on glass coverslips for 48 h: regular coverslips as control, COL-GEL(C) coated coverslips, G-GEL(C) coated coverslips. Cells were first washed with PBS, followed by fixation with 4% PFA (30 min, RT). After fixation, the cells were permeabilized with 0.1% Triton X-100 for 10 min and stained with 1:100 FITC-Phalloidin (Solarbio, CA1620; RT, 30 min). After PBS wash, nuclei were counterstained with DAPI. Samples were mounted for imaging using confocal microscopy.

Scratch wound healing assay

Direct contact method. HS-fibroblasts were seeded into 6-well plates (Corning, 3516) coated with either COL-GEL(C), G-GEL(C) or left uncoated (regular control) and cultured until reaching 100% confluence under standard conditions (37 °C, 5% CO₂). A uniform scratch was then generated along the short axis (perpendicular to the reference lines) using a sterile 200 µL pipette tip. After removing detached cells by gentle PBS washing (2–3 times), 2 mL of serum-free HD-DMEM (containing 1% penicillin-streptomycin) was added to each well. Baseline images (0 h) were captured at the predefined reference points, and cell migration was monitored at 24 h, 48 h and 72 h.

Hydrogel extract method. To prepare hydrogel and cryogel extracts, 500 µL/well of the pre-gel solution was added to 6-well culture plates and allowed to form hydrogel or cryogel using the method described before. The resulting plate-coated hydrogel or cryogel were immersed in 2 mL/well of ultrapure Milli-Q water for 1 h, followed by three washes with 1 mL/well of PBS. After sterilization under UV light for 60 min, the hydrogel or cryogel were immersed in 1 mL/well of serum-free high-glucose DMEM at 37 °C for 48 h to obtain the extracts for scratch wound healing assay; and in 1 mL/well of complete high-glucose DMEM at 37 °C for 48 h to obtain the extracts for Western blot.

HS-fibroblasts were first seeded in uncoated 6-well plates (Corning, 3516) and grown to 100% confluence under the same culture conditions. The scratched cells were then treated with the corresponding cryogel extracts (serum-free HD-DMEM, COL-GEL(C) extract, G-GEL(C) extract). Baseline images (0 h) were captured at the predefined reference points, and cell migration was monitored at 24 h and 48 h.

The relative healing rate was quantified as: 1- (remaining scratch area at a given time point/scratch area at 0 h) × 100%.

Cell proliferation ability test (CCK-8 assay)

HS-fibroblasts (1 × 10³/well) were seeded into 96-well plates (Corning, 3516) coated with either COL-GEL(C), G-GEL(C), or left uncoated (regular control). After a 48-hour attachment period, the proliferation measurement was taken at the 2nd day, 4th day and 6th day. 100 µL culture medium with 10 µL of CCK-8 reagent was added per well. Following a 2-hour incubation at 37 °C, the absorbance at 450 nm was recorded.

Each group’s raw OD values were normalized to the blank control (culture medium without cells) to account for inter-assay variation, using the formula: (OD _{control or treatment}-OD_blank)/OD_blank. To avoid potential interference from graphene in the CCK-8 assay, a G-GEL(C) only coated group (without cells) was included in addition to the regular blank control, allowing proper normalization of the G-GEL(C)+cell group.

Furthermore, to eliminate any optical interference caused by graphene coating at the bottom of the plate, 100 µL of DMEM culture medium containing the CCK-8 reagent from each well was transferred to a fresh, empty 96-well plate for absorbance measurement.

Scanning electron microscope (SEM)

The surface of the freeze-dried COL-GEL(C), G-GEL(C) and injectable G-GEL(C) were coated with gold nanoparticles and evaluated by scanning electron microscopy (JSM-IT300, JEOL, Tokyo, Japan) at a high voltage of 20 kV. The pore quantity and relative area were analyzed by ImageJ software.

Water absorption capacity

Three specimens of freeze-dried G-GEL(C) and COL-GEL(C) were equally weighed as W₀ and placed in PBS for 2 min, 4 min, 6 min, 8 min,10 min, 20 min and 30 min at room temperature. After soaking, specimens were removed from PBS and weighed to obtain W_t. The weight absorption percentage was calculated by the following equation. Water absorption percentage (%) = [(Wₜ - W₀)/W₀] × 100%.

Interconnected porosity

To test the interconnected porosity, freeze-dried G-GEL(C) and COL-GEL(C) in similar dry weight (W_d) were first weighted and hydrated for 30 min, 12 h, 36 h, 60 h. Hydrated scaffolds were weighed as wet weight or total weight (W_t) on a scale. And a Kimwipe was lightly applied to the scaffold surface for 30s to wick away loosely held water, and the mass was again recorded as retained weight (W_r).

The interconnected volume was calculated as the mass of water wicked away divided by the total hydrated mass: Interconnected porosity percentage (%)= [(W_t-W_r)/W_t] × 100%.

Quantification of ASCs-ApoVs loading and release

To minimize premature release and preserve vesicle integrity, ASCs-ApoVs were loaded onto the cryogels at 4 °C for 24 h; release testing was then performed at 37 °C to simulate physiological conditions.

1. ASCs-ApoVs Loading Lyophilized G-GEL(C) and COL-GEL(C) samples were individually incubated in a known concentration of ASCs-ApoVs solution (C_initial, determined by BCA) at a volume of at least 1 mL per mg of dry gel at 4 °C for 24 h under gentle shaking. After incubation, the hydrogels were retrieved, and the protein concentration of the residual solution (C_after) was assessed by BCA again to determine the ApoVs adsorption efficiency.

2. Surface-associated ASCs-ApoVs releasing test To quantify the fraction of ASCs-ApoVs loosely associated with the cryogel surface, each sample after ASCs-ApoVs loading was gently rinsed three times by pipetting up and down with 37 °C PBS (100 µL per mg of dry gel). The protein concentration in the collected rinse solution was measured using a BCA assay against a PBS blank.

3. Short-term ASCs-ApoVs release test after surface wash Following removal of surface-associated ASCs-ApoVs, the same cryogel samples were incubated in fresh PBS at 37 °C (100 µL per mg of dry gel) for 2 h to allow short-term release. The release medium was then collected, and the cumulative protein release was quantified via a BCA assay against a PBS blank. (It should be noted that the short-term release data were obtained after removal of surface-associated ASCs-ApoVs; therefore, they do not represent the total initial release from intact cryogels.)

Rabbit ear hypertrophic scar model establishment

Healthy New Zealand white rabbits (n = 3, 2.5–3.0 kg) were acclimatized for one week. After depilating the ventral ear surfaces, general anesthesia was induced via intraperitoneal pentobarbital sodium (3%, 1 mL/kg). Under aseptic conditions, a full-thickness skin and perichondrium specimen (1.0 cm×1.0 cm) was excised from the mid-ventral region of each ear, preserving the underlying cartilage. Wounds were left open to heal.

Starting on postoperative 30 days, the scars in each HS group (Control (PBS), 1.0 µg/ml ASCs-ApoVs, injectable G-GEL(C), and 1.0 µg/ml ASCs-ApoVs loaded injectable G-GEL(C)) received intralesional injections of the corresponding agent twice weekly for a period of three weeks. All injections were administered in equal volumes. After three weeks, tissue samples were harvested for hematoxylin and eosin (H&E), Masson, immunohistochemical (IHC) and immunofluorescence (IF) analysis. Scar elevation index (SEI) = scar/normal skin thickness.

After the experiment, rabbits were euthanized under isoflurane (RWD, R510-22) anesthesia. Animals were placed in an induction chamber with 3%−5% isoflurane and 1 L/min oxygen. Once the righting reflex was lost, anesthesia was maintained with 2%−3% isoflurane via a face mask. After confirming deep anesthesia by the absence of a pedal withdrawal reflex, potassium chloride injection (China Otsuka Pharmaceutical Co., Ltd, 1–2 mEq/kg) was rapidly injected into the marginal ear vein. Death was confirmed by cessation of heartbeat and respiration, along with fixed dilated pupils.

For each rabbit (n = 3), three tissues were collected per normal skin and treatment group to account for intra-tissue variability. Prior to statistical analysis, the three replicates were averaged per rabbit per treatment group, yielding a final sample size of n = 3 per group (one value per rabbit). This approach provides a conservative estimate of treatment effects while avoiding pseudoreplication.

Statistics analysis

Continuous data are presented as the mean ± standard deviation. For comparisons between two groups, an independent samples t-test was employed. For comparisons across three or more groups, a one-way analysis of variance (ANOVA) was applied, followed by Student-Newman-Keuls test (SNK test) test for post-hoc pairwise comparisons. Statistical significance: *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

To enable cross-batch comparisons of independent experiments (including RT-qPCR and select Western blot analyses), quantitative data were normalized to their corresponding internal controls. This normalization was performed by dividing individual measurements in the treatment groups by their respective control groups, thereby defining the control as a baseline of 1. In the graphical representations, the control group is depicted as a baseline of 1 without error bars, while the normalized results in treatment groups are expressed with new mean ± standard deviation.

Results

Graphene modulates the phenotype of hypertrophic scars derived fibroblasts (HS-fibroblasts)

To investigate the anti-fibrosis effects of graphene in hypertrophic scars (HS), we conducted the phenotypic assessments of HS-fibroblasts cultured with graphene-incorporated type I collagen cryogel (G-GEL(C)) or type I collagen cryogel (COL-GEL(C)) coated plates.

Graphene was incorporated into the 2 mL system at varying concentrations (2.5, 2.0, 1.5, and 1.0 mg) to determine the optimal dosage. The Western blot results (Fig. 1, A-B) demonstrated that graphene supplementation significantly inhibited the expression of COL1A1 in HS-fibroblasts (P < 0.0001 for G-GEL(C)−1 to −3; P < 0.01 for G-GEL(C)−4) and COL3A1 (P < 0.0001 for G-GEL(C)−1 to −2; P < 0.001 for G-GEL(C)−3). A clear dose-dependent effect was observed particularly for G-GEL(C)−1, which most markedly suppressed α-SMA (P < 0.001) and Vimentin (P < 0.001) expression in HS-fibroblasts. Based on these findings, a graphene concentration of 2.5 mg per 2 mL was selected for subsequent experiments. The graphene-type I collagen cryogel prepared at this ratio is hereafter uniformly designated as G-GEL(C).

Fig. 1.

The graphene-incorporated type I collagen cryogel (G-GEL(C)) modulates phenotype of hypertrophic scars derived fibroblasts (HS-fibroblasts). (A) Western blot analysis of myofibroblast-related markers (COL1A1, COL3A1, α-SMA, and Vimentin) in HS-fibroblasts cultured for 48 h on normal plates (as control), type I collagen cryogel (COL-GEL(C)) coated plates or on graphene-incorporated type I collagen cryogel (G-GEL(C)) coated plates at different graphene concentrations (2.5, 2.0, 1.5, and 1.0 mg/2 mL). Original blots are presented in Supplementary Fig. 3. (B) Quantification of protein expression from section A by relative gray vale to GAPDH. (Mean ± SD, n = 8 independent experiments, data analyzed by one-way ANOVA with SNK test for posterior comparisons.) (C) Western blot analysis of myofibroblast-related markers (COL1A1, COL3A1, α-SMA, and Vimentin) in HS-fibroblasts cultured on normal plates (as control), COL-GEL(C) coated plates or G-GEL(C) (with 2.5 mg/2 ml graphene) coated plates for 48 h. Original blots are presented in Supplementary Fig. 3. (D) Quantification of protein expression from section C by relative gray vale to GAPDH. (Mean ± SD, n = 3 independent experiments, data analyzed by one-way ANOVA with SNK test for posterior comparisons.). (E) Quantitative analysis of RT-qPCR of related genes with 48 h normal plates (as control), COL-GEL(C) coated plates or G-GEL(C) coated plates culture, normalized to GAPDH and the control group. (Mean ± SD, n = 3 independent experiments, data analyzed by one-way ANOVA with SNK test for posterior comparisons.) Statistical significance: *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

In the repeated Western blot experiment using G-GEL(C) containing 2.5 mg graphene per 2 mL (Fig. 1, C-D), the protein expression levels of COL1A1 (p < 0.001), COL3A1 (p < 0.05), α-SMA (p < 0.001), and Vimentin (p < 0.05) were significantly decreased. Among these, the reductions in COL1A1 (p < 0.001), α-SMA (p < 0.01), and Vimentin (p < 0.05) were more pronounced compared with the COL-GEL (C) treatment. Consistent with the Western blot results, RT-qPCR analysis (Fig. 1, E) demonstrated that G-GEL(C) treatment significantly downregulated the transcriptional levels of the genes encoding COL1A1 (p < 0.001), COL3A1 (p < 0.0001), α-SMA (p < 0.0001), and Vimentin (p < 0.0001) in HS-fibroblasts.

In summary, graphene-incorporated type I collagen cryogel (G-GEL(C)) show robust anti-fibrotic effects on HS-fibroblasts.

Graphene modulates the proliferation, lateral migration and the arrangement of the cytoskeleton of HS-fibroblasts

The effect of COL-GEL(C) and G-GEL(C) on the proliferation of HS-fibroblasts was assessed using the CCK-8 assay (Fig. 2, A-B). On day 2 post-seeding, the COL-GEL(C) group showed no significant difference from the control group (p > 0.05), whereas the G-GEL(C) group exhibited significantly lower OD450 nm readings (p < 0.0001). By day 4, the OD value of the COL-GEL(C) group became significantly higher than that of the control group (p < 0.05), while the G-GEL(C) group remained lower than others (p < 0.0001). On day 6, the G-GEL(C) group continued to display significantly lower OD readings compared with the control (p < 0.0001) and the COL-GEL(C) group (p < 0.0001). However, no significant differences in cell density or morphology were observed under light microscopy after 2 days of seeding (Supplementary Fig. 1, A), indicating that G-GEL(C) could exhibit relatively low cytotoxicity despite suppressing cell proliferation.

Fig. 2.

The G-GEL(C) modulates proliferation, lateral migration and cytoskeletal characteristics of HS-fibroblasts. (A) Bar charts of the CCK-8 assay results of HS-fibroblasts cultured on normal plates (as control), COL-GEL(C) coated plates or G-GEL(C) coated plates at different time points. (Mean ± SD, n = 3 independent experiments, data analyzed by one-way ANOVA with SNK test for posterior comparisons.) (B) Proliferation curve of HS-fibroblasts cultured on normal plates (as control), COL-GEL(C) coated plates or G-GEL(C) coated plates at 2-, 4-, and 6- day following seeding by CCK-8 assay. (Lines were slightly offset on the x-axis for clarity; offsets do not represent actual time differences.) (C) Images of scratch wound healing assay of HS-fibroblasts cultured on normal plates (as control), COL-GEL(C) coated plates or G-GEL(C) coated plates. (D) The line chart of quantified wound closure rates of HS-fibroblasts at 0, 24, 48 and 72 h according to the section C. (Lines were slightly offset on the x-axis for clarity; offsets do not represent actual time differences.) (Mean ± SD, n = 3 independent experiments, data analyzed by one-way ANOVA with SNK test for posterior comparisons.) (E) Cytoskeletal staining of HS-fibroblasts cultured on normal coverslips (as control), COL-GEL(C) coated coverslips or G-GEL(C) coated coverslips for 48 h (scale bars = 75 μm). Statistical significance: *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

A scratch-wound assay was performed to further evaluate the effect of graphene direct-contact on the lateral migration ability of HS-fibroblasts (Fig. 2, C-D). Results showed that at 24 h post-scratch, no significant difference in cell lateral migration was observed among the groups (p > 0.05). After 48 h, cells on both COL-GEL(C) and G-GEL(C) coated surfaces exhibited significantly reduced lateral migration and lower wound closure rates compared to the control (p < 0.05), though no statistical difference was detected between the two treatment groups. By 72 h, G-GEL(C) displayed the lowest wound closure rate, followed by COL-GEL(C), while the control group showed the highest closure rate. These findings indicate that G-GEL(C) exerts a time-dependent inhibitory effect on the lateral migration of HS-derived fibroblasts.

However, material extracts failed to recapitulate the phenotypic modulation of HS-fibroblasts (Supplementary Fig. 1, B-C) and did not reproduce the 48 h inhibition of HS-fibroblast lateral migration (Supplementary Fig. 1, D-E). These results imply that the regulatory effects of graphene on HS-fibroblasts likely depend on direct physical contact.

We further hypothesized that graphene may act through cytoskeleton-mediated mechanisms. Therefore, cytoskeletal staining on HS-fibroblasts grown on G-GEL (C) and COL-GEL (C) coated coverslips was conducted (Fig. 2, E). HS-fibroblasts cultured on G-GEL (C) exhibited a significantly larger cell spreading area around the graphene, suggesting enhanced adhesion to the hydrogel matrix. This observation aligns with their inhibited migratory phenotype (Fig. 2, C-D).

In summary, G-GEL(C) significantly suppresses the proliferative capacity and lateral migration of HS-fibroblasts. Furthermore, the regulatory action of graphene is highly dependent on the direct contact.

Adipose stem cells (ASCs) derived apoptotic vesicles (ApoVs) modulates the phenotype of HS-fibroblasts

Apoptosis of adipose stem cells (ASCs) was induced by treatment with 5 µM staurosporine (STS) for 12 h to harvest apoptotic vesicles (ApoVs)²⁸. Cell histological changes observed by FITC-phalloidin staining (Fig. 3, A) and the proportion of PI(−)/Annexin V(+) cells detected by flow cytometry (Fig. 3, B) confirmed the efficacy of apoptosis induction. Transmission electron microscopy (TEM) revealed the vesicular structure and nanometer-scale diameter of the harvested vesicles (Fig. 3, C).

Fig. 3.

Identification of adipose stem cell derived apoptotic vesicles (ASCs-ApoVs). (A) FITC-phalloidin staining of normal ASCs and ASCs undergoing apoptosis induced by staurosporine (STS; 5 µM, 12 h) (scale bar = 75 μm). (B) Annexin V-FITC/PI dual staining assay of ASCs for apoptosis detection (5 µM STS, 12 h) by flow cytometry. (C) Transmission electron microscopy (TEM) images of ASCs-ApoVs (scale bars = 500 nm and 200 nm). (D) Nanoparticle tracking analysis (NTA) of ASCs-ApoVs. (E) Western blot analysis of caspase-3 expression in ASCs-ApoVs and the cell bodies of apoptotic ASCs, with β-actin used as a reference. Original blot images are presented in Supplementary Fig. 3. (F) Subcellular localization of internalized ASCs-ApoVs in HS-fibroblasts. (a Nuclear staining (Hoechst, blue) of ASCs-ApoVs treated HS-fibroblasts. (b Membrane staining (CM-Dil, red) of cellular organelles of HS-fibroblasts. (c ASCs-ApoVs were labeled with DiO dyestuff (green). (d Merged image showing dual localization patterns: some ASCs-ApoVs co-localized with HS-fibroblasts’ membranous organelles (yellow in merged channels); others dispersed in cytoplasm (green only) (Scale bar = 50 μm).

Our previous measurements showed that the zeta potential of ApoVs was approximately − 10 mV, indicating good colloidal stability, minimal vesicle disruption, and few protein impurities¹⁷. Nanoparticle tracking analysis (NTA) (Fig. 3, D) revealed a population of vesicles primarily ranging from 100 to 300 nm in diameter, along with a fraction of 800–900 nm particles, which is consistent with observations from TEM images (Fig. 3, C). Collectively, these findings suggest that ApoVs exhibit size heterogeneity.

Western blot analysis of Caspase-3 (Fig. 3, E) in both apoptotic ASCs and ASC-derived STS-induced vesicles further confirmed the apoptotic identity of these vesicles²⁹, hereafter referred to as ASCs’ apoptotic vesicles (ASCs-ApoVs). DiO-labeled ASCs-ApoVs were internalized by CM-Dil labeled HS-fibroblasts after 2 h of co-culture at 37 °C (Fig. 3, F), confirming that ASCs-ApoVs remain intact as membranous vesicles capable of regulating HS-fibroblasts.

ASCs-ApoVs, serving as biological vesicles, demonstrated excellent phenotypic modulation effects on HS-fibroblasts. Western blot analysis (Fig. 4, A-B) revealed that treatment with ASCs-ApoVs (1.0 µg/ml, 24 h) significantly decreased the protein expression levels of COL1A1 (p < 0.01) and α-SMA (p < 0.05) in HS-fibroblasts, while increasing COL3A1/COL1A1 ratio (p < 0.05). RT-qPCR results (Fig. 4, C) further indicated that after ASCs-ApoVs treatment (1.0 µg/ml, 24 h), the transcriptional levels of genes encoding COL1A1 (p < 0.0001), COL3A1 (p < 0.0001), α-SMA (p < 0.05) and Vimentin (p < 0.001) were significantly downregulated in HS-fibroblasts. And the COL3A1/COL1A1 transcript ratio was markedly elevated (p < 0.05). Immunofluorescence staining (Fig. 4, D) further confirmed α-SMA expression inhibition after ASCs-ApoVs treatment (p < 0.05). Additional western blot analysis rules out potential effects from solvable STS that may be residual in the ASCs-ApoVs preparation (Supplementary Fig. 1, F-G). In summary, ASCs-ApoVs significantly reduced the fibrotic level of HS-fibroblasts.

Fig. 4.

Adipose stem cells derived apoptotic vesicles (ASCs-ApoVs) modulate the fibrotic phenotype of HS-fibroblasts. (A) Western blot analysis of myofibroblast-related markers (COL1A1, COL3A1, α-SMA, and Vimentin) in HS-fibroblasts with PBS (same volume as control) or ASCs-ApoVs administration (1.0 µg/ml, 24 h). Original blots are presented in Supplementary Fig. 3. (B) Quantification of protein expression from section A by relative gray vale to GAPDH. (Mean ± SD, n = 3 independent experiments, data analyzed by independent sample t test.) (C) Quantitative analysis of RT-qPCR of related genes with ASCs-ApoVs administration (1.0 µg/ml, 24 h) or PBS (same volume as control), normalized to GAPDH and the control group. (Mean ± SD, n = 3 independent experiments, data analyzed by independent sample t test.) (D) Immunofluorescence staining (scale bars = 400 μm) and the scatter plot of α-SMA expression in HS-fibroblasts after ASCs-ApoVs treatment (1.0 µg/ml, 24 h) or PBS (same volume as control). (Mean ± SD, n = 3 independent experiments, data analyzed by Independent-samples t-test.) Statistical significance: *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

To preliminarily assess the potential advantages of the combined delivery of G-GEL(C) and ASCs-ApoVs, the protein and transcript levels of scar-related markers in HS fibroblasts were normalized and collectively analyzed (Supplementary Fig. 2, A-D). Across multiple key indicators, the two agents exhibited complementary and synergistic effects. Specifically, G-GEL(C) downregulated Vimentin protein expression (p < 0.05), an effect not significantly observed with ASCs-ApoVs alone; while ASCs-ApoVs exerted stronger inhibition on COL1A1 protein expression compared to G-GEL(C) alone (p < 0.0001). In summary, both G-GEL(C) and ASCs-ApoVs exhibit regulatory effects on HS-fibroblasts, though with distinct characteristics. Utilizing G-GEL(C) to deliver ASCs-ApoVs could have a coordinated regulatory potential on HS-fibroblasts and hypertrophic scar tissue.

G-GEL(C) as an effective vesicle carrier with porosity and ASCs-ApoVs adsorption

Microscopic observation (Fig. 5, B) revealed that both plate-coated and injectable G-GEL(C) formed a porous, cross-linked collagen structure, consistent with scanning electron microscopy (SEM) analysis (Fig. 5, C-b/c), indicating the relative structural stability of the G-GEL(C) cryogel during form modulation. Furthermore, mechanical fragmentation introduces more macropores (e.g., fracture-induced pores of approximately 40–100 μm) into the injectable G-GEL(C) compared to its intact counterpart, while still retaining small pores of similar diameters (around 20 μm). This increased porosity provides a larger exposure and contact area for adsorbing ASCs-ApoVs, thereby holding promise for enhancing the ASCs-ApoVs loading efficiency.

Fig. 5.

G-GEL(C) is an effective vesicle carrier with porosity and ASCs-ApoVs adsorption. (A) Photographic comparison of injectable G-GEL(C), and massive COL-GEL(C) as well as G-GEL(C) after PBS absorption. (B) Macroscopic photograph of the 6-well cell culture plates coated G-GEL(C) and injectable G-GEL(C) in PBS (scale bars = 300 μm). (C) Representative scanning electron microscopy (SEM) images of COL-GEL(C), G-GEL(C) and injectable G-GEL(C) (scale bars = 200 μm). (D) Quantitative comparison of the relative pore area in COL-GEL(C) and G-GEL(C) via SEM image analysis (Mean ± SD, n = 6 independent experiments, data analyzed by Independent-samples t-test.). (E) (a Swelling kinetics of COL-GEL(C) and G-GEL(C) over time. (b Quantitative comparison of the equilibrium swelling ratio in PBS. (Mean ± SD, n = 3 independent experiments, data analyzed by Independent-samples t-test.) (F) Calculated porosity of COL-GEL(C) and G-GEL(C) according to the statistics in G section. (Mean ± SD, n = 3 independent experiments, data analyzed by Independent-samples t-test.) (G) Dry weight, wet weight and retained weight of COL-GEL(C) and G-GEL(C) before and after 30 min of PBS uptake. (Mean ± SD, n = 3 independent experiments, data analyzed by Independent-samples t-test.) (H) (a Adsorption of ASCs-ApoVs (24 h, 4℃) by COL-GEL(C) and G-GEL(C), measured by BCA assay. (Mean ± SD, n = 3 independent experiments, data analyzed by one-way ANOVA with Student-Newman-Keuls post-hoc test.) (b Surface-associated ASCs-ApoVs releasing (3-time wash, 37℃) from COL-GEL(C) and G-GEL(C), measured by BCA assay. (Mean ± SD, n = 3 independent experiments, data analyzed by one-way ANOVA with Student-Newman-Keuls post-hoc test.) (c Short-time releasing of ASCs-ApoVs (2 h, 37℃) from COL-GEL(C) and G-GEL(C), measured by BCA assay. (Mean ± SD, n = 3 independent experiments, data analyzed by one-way ANOVA with Student-Newman-Keuls post-hoc test.) Statistical significance: *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

Besides, SEM structural characterization further showed that G-GEL(C) possessed a more porous architecture with larger pores (Fig. 5, D) compared to COL-GEL(C). This increased porosity could facilitate fibroblast infiltration upon dermis injection, thereby enhancing direct cell-material contact and potentially improving therapeutic efficacy.

Both lyophilized COL-GEL(C) and G-GEL(C) exhibited high water absorption capacity (Fig. 5, E). However, the incorporation of graphene conferred more interconnected pores (Fig. 5, F), higher water absorption quality (Fig. 5, G) and accelerated swelling equilibrium during the early absorption phase (Supplementary Fig. 2, E-F). These findings indicate that G-GEL(C) possesses superior ASCs-ApoVs liquid carrying efficiency and capacity.

Notably, the adsorption and retention of ASCs-ApoVs by G-GEL(C) were not solely attributable to physical swelling. Following incubation with lyophilized G-GEL(C), the total protein concentration in the ASCs-ApoVs solution decreased significantly (P < 0.0001), indicating selective vesicle adsorption from the PBS medium (Fig. 5, H-a). In contrast, COL-GEL(C) did not exhibit this function. We hypothesize that this difference may be due to interactions between the surface charge of graphene and the membrane properties of ASCs-ApoVs³⁰. Furthermore, the amount of surface-associated ApoVs released from COL-GEL(C) was much higher than that from G-GEL(C) (P < 0.0001) (Fig. 5, H-b), indicating that ASCs-ApoVs loaded in COL-GEL(C) are mainly localized on the surface, leading to a burst release. Whereas those in G-GEL(C) enable a milder release at the initial stage. Both COL-GEL(C) and G-GEL(C) allowed short-term ASCs-ApoVs release (Fig. 5, H-c). As cell infiltration into the cryogel takes time, the released ApoVs provide early therapeutic effects. Therefore, both short-term release and long-term cell-mediated mechanisms are realized.

In summary, owing to its favorable physicochemical properties and confirmed selective adsorption of ASCs-ApoVs, G-GEL(C) represents a promising injectable delivery platform for ASCs-ApoVs.

In vivo therapeutic efficacy in a rabbit ear hypertrophic scar model

To further evaluate the in vivo scar-modulating capabilities of ASCs-ApoVs, injectable G-GEL(C), and ASCs-ApoVs loaded injectable G-GEL(C), we utilized a rabbit ear hypertrophic scar model.

The induced scars exhibited typical hypertrophic characteristics (Fig. 6, A-PBS): raised and stiff tissue protruding from the surrounding normal skin. H&E-stained sections revealed hypercellularity and locally disorganized, whorl-like fibrous structures. Masson’s trichrome staining indicated an abundance of red-stained fibers. α-SMA immunohistochemistry showed intense staining within numerous fibroblasts, which displayed irregular morphology consistent with activated myofibroblasts.

Fig. 6.

Histological staining of hypertrophic scars (HS) in different groups. (A) Histological results of H&E, Masson’s Trichrome, and immunohistochemical staining of α-SMA in control group (normal skin tissues) and hypertrophic scar group with PBS, ASCs-ApoVs (1.0 µg/ml), G-GEL(C) and ASCs-ApoVs (1.0 µg/ml) loading G-GEL(C) injection separately. (B) Analysis of (a scar elevation index (SEI), (b area fraction of red fibers by Masson’s Trichrome and (c α-SMA expression level per relative units in different groups. (Mean ± SD, n = 3 independent experiments, statistical significance was determined by one-way ANOVA with Student-Newman-Keuls post-hoc test.) Statistical significance: *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

Local injection of ASCs-ApoVs (1.0 µg/ml) significantly reduced scar thickness, with markedly decreased cellularity and a substantial reduction in whorl-like fibrous arrangements (Fig. 6, A- ASCs-ApoVs). Tissue staining showed a significant decrease in scar elevation index (SEI) (Fig. 6, B-a; P < 0.05), red fiber content (Fig. 6, B-b; P < 0.0001), α-SMA positive fibroblasts (Fig. 6, B-c; P < 0.0001). All results indicate that local ASCs-ApoVs administration effectively modulates hypertrophic scar tissue remodeling in vivo.

G-GEL(C) injection resulted in the retention of black and multi-layered graphene within the scar tissue (Fig. 6, A- G-GEL(C)). G-GEL(C) injection also led to a significant reduction in scar elevation index (SEI) compared to the untreated control (Fig. 6, B-a; P < 0.05). While disorganized and coarse whorled fibers persisted in areas distant from the graphene, collagen fibers adjacent to the G-GEL(C) implant exhibited a notably aligned and unidirectional pattern. Correspondingly, tissue near the material showed a significant decrease in α-SMA expression (Fig. 6, B-c; P < 0.0001).

The combination of ASCs-ApoVs (1.0 µg/ml) loaded injectable G-GEL(C) achieved the lowest α-SMA expression among all groups (Fig. 6, B-c; P < 0.0001). Scar elevation index (SEI) was also significantly reduced (Fig. 6, B-a; P < 0.05). Masson’s staining confirmed the aligned fiber pattern near the G-GEL(C) material.

All animal vital organ sections, blood biochemistry (Supplementary Table 2), and routine blood tests (Supplementary Table 3) were normal, indicating that local injection of G-GEL(C) or with ASCs-ApoVs for scar treatment had minimal individual toxicity.

Histological examination by H&E staining revealed that graphene retained in the dermis was typically surrounded by marked infiltration of mononuclear and multinuclear cells (Fig. 6, A), suggesting potential immunomodulatory effects beyond direct fibroblast regulation. As graphene sheets are readily phagocytosed by macrophages³¹—a process reported to be associated with intracellular reactive oxygen species (ROS) generation³² and macrophage polarization³³—further staining for macrophage polarization (iNOS/CD163) and ROS-related DNA/RNA damage (8-OHdG) was performed in the G-GEL(C) and ASCs-ApoVs-loaded G-GEL(C) groups (Fig. 7).

Fig. 7.

Tissue staining of macrophage polarization and ROS-related DNA/RNA damage in G-GEL(C) and ASCs-ApoVs loaded G-GEL(C) groups. HE, Masson and immunofluorescence staining was performed to evaluate macrophage polarization (iNOS for M1, CD163 for M2) and ROS-related DNA/RNA damage (8-OHdG) in the G-GEL(C) group and the ASCs-ApoVs loaded G-GEL(C) group (Scale bars = 100/200/500 µm).

In the G-GEL(C) group alone, increased 8-OHdG (green) staining was observed inside the scar tissue; however, this staining was markedly reduced in the immediate vicinity of the graphene material, suggesting a local inhibitory effect of graphene on oxidative stress. Double staining for iNOS (green) and CD163 (red) revealed the presence of both M1 and M2 macrophages within the scar. Notably, only a few M1 macrophages accumulated around the material, while M2 macrophages were predominantly located at a distance from the material.

The ASCs-ApoVs loaded G-GEL(C) group exhibited significant differences from the G-GEL(C) alone. Although increased 8-OHdG staining was similarly observed inside the scar, this signal extended consistently into the interior of the material. Furthermore, unlike the G-GEL(C) alone group, which predominantly showed M1 macrophages around, the ApoVs loaded G-GEL(C) group exhibited significant M2 macrophage recruitment both inside and around the material. As low level ROS are beneficial for initiating macrophages with M2³⁴, we hypothesize that local aggregation and sustained release of ApoVs from G-GEL(C) create a ROS regulatory niche that skews macrophage polarization toward the M2 phenotype. In this context, the ASCs-ApoVs loaded G-GEL(C) acts as a spatial scaffold that concentrates M2 macrophages in the vicinity of the material, thereby amplifying the anti-fibrotic immune response and more effectively regulating hypertrophic scar development than G-GEL(C) alone.

Discussion

The relationship between graphene and fibrosis has recently gained increasing attention. Several studies have demonstrated anti-fibrotic effects of graphene-based materials in different organs. For instance, graphene quantum dots were reported to alleviate renal fibrosis by reducing oxidative stress and restoring mitochondrial membrane potential³⁵. Incorporation of graphene oxide into collagenous biomaterials was shown to attenuate the neural scar-forming phenotype transition of reactive astrocytes in vitro³⁶. Moreover, graphene oxide nanomaterials decreased the expression of TGFβRⅡ, Smad2, and Smad3 in hepatic stellate cells, indicating their potential to attenuate liver fibrosis³⁷. In this study, we demonstrated that the graphene-type I collagen cryogel (G-GEL(C)) regulates the phenotype (COL1A1, α-SMA and Vimentin), proliferation, and lateral migration of hypertrophic scar derived fibroblasts (HS-fibroblasts) through direct contact, broadening its application in skin fibrosis.

Beyond the direct effects on HS-fibroblasts, graphene-based hydrogels may also modulate scar formation via macrophage polarization and anti-oxidative damage. Previous studies have shown that graphene-based materials promote M2 macrophage polarization in rheumatoid arthritis^38,39 and spinal cord injury⁴⁰, reducing tissue damage and inflammation. However, in the present study, we observed a spatial distribution with M1 macrophages near the material and M2 macrophages farther away—a pattern warranting further investigation. Additionally, G-GEL(C) locally reduced ROS-related markers (8-OHdG), indicating a significant anti-oxidative damage function. Given that low levels of ROS are beneficial for initiating M2 macrophages³⁴, whereas high levels of ROS drive M1 polarization⁴¹, both ROS regulation and macrophage polarization warrant further study.

Comparison between COL-GEL(C) and G-GEL(C) confirmed that graphene is essential for achieving the fibrosis regulatory functions described above. Beyond its biological activity, graphene endows the type I collagen cryogel with unique physicochemical properties, including a highly porous structure, rapid and higher water absorption, active adsorption of apoptotic vesicles (ApoVs), and gentler release kinetics. This selective loading capacity makes G-GEL(C) a promising scaffold for composite-based scar modulation via injectable delivery. Nevertheless, the incorporation of type I collagen into the graphene system is equally important, as it overcomes several limitations of graphene delivery alone. First, the high stiffness of graphene relative to human dermis may activate fibroblasts, upregulate α-SMA⁴², and trigger YAP-mediated mechanosensitive signaling⁴³, thereby promoting myofibroblast differentiation and excessive extracellular matrix deposition⁴⁴. Therefore, a higher graphene content does not necessarily lead to better outcomes. Second, graphene dispersions suffer from poor long-term stability due to its limited solubility⁴⁵ in both aqueous and organic media, even after chemical modification⁴⁶. The type I collagen cryogel system improves the homogeneity and uniformity of graphene dispersion.

Apoptotic vesicles (ApoVs) are a specific type of extracellular vesicle derived from apoptotic cells. Compared with other conventional approaches for HS regulation, ApoVs exhibits remarkable advantages. Mechanistically, ApoVs enable solid and comprehensive therapeutic effects due to their diverse cargo. Structurally speaking, the nanoscale diameter of ApoVs facilitates enhanced biodistribution and has the ability to traverse biological barriers⁴⁷—features usually unattainable with intact cellular therapies⁴⁸. Besides, the phospholipid bilayer of ApoVs could provide cargo protection, significantly improving molecular bioavailability^49,50. Clinically speaking, ApoVs exhibits higher yields⁵¹ compared with other cell-free therapies.

However, highly efficient and sustainable delivery methods for ApoVs and other extracellular vesicles are still under investigation. In this study, G-GEL(C) exhibited selective adsorption of ApoVs, as evidenced by a significant decrease in the protein concentration of the ApoVs solution after material incubation. This vesicle adsorption mechanism is more stable and efficient for clinical translation than the simpler method of loading vesicles via passive swelling commonly used in other hydrogels. Moreover, compared with G-GEL(C) alone, the ApoVs-loaded G-GEL(C) group displayed distinct spatial patterns of ROS-related damage index and macrophage distribution within the material, indicating that ApoVs were stably retained by G-GEL(C). The ability to achieve long-term spatially controlled regulation—which can be conveniently executed by manual injection—further underscores the great potential of this combined system for clinical translation.

This study has several limitations that warrant further investigation. First, the preparation of G-GEL(C) involved basic physical grinding and syringe filtration to achieve injectability. Future work could develop more standardized processing methods and explore alternative formulations such as microspheres or microneedles. Second, although we have proposed hypotheses for graphene mediated regulation of HS-fibroblasts, macrophages, and local ROS homeostasis, these mechanisms require further validation through standardized experimental approaches. Third, only one type of graphene reagent was evaluated in this study. The effects of graphene with different sizes, layer numbers, or compositional ratios on HS remain to be compared. Finally, the inconsistency between mRNA and protein levels observed in this study may arise from post-transcriptional regulation, differential protein stability, post-translational modifications, or temporal mismatches between transcription and translation. As ASCs-ApoVs carry a variety of regulatory cargos (e.g., miRNAs, cytokines) that can independently affect mRNA turnover and translation efficiency, the discordance likely reflects genuine multi-layer regulation. Elucidating the precise mechanism will require dedicated kinetic and pharmacological studies, which are beyond the scope of the current work but will be pursued in future investigations.

Conclusion

Graphene directly inhibits HS progression by downregulating fibrotic markers, suppressing proliferation, and impeding lateral migration of HS-fibroblasts. At the tissue level, it also regulates ROS homeostasis and M1/M2 macrophage polarization. The highly porous G-GEL(C) with selective protein adsorption efficiently incorporates anti-fibrosis ASCs-ApoVs. Together, this combined system represents a translatable therapeutic strategy for in vivo modulation of HS and skin fibrosis.

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

HS

Hypertrophic scar

ASCs

Adipose stem cells

ApoVs

Apoptotic vesicles

ASCs-ApoVs

ASCs derived apoptotic vesicles

G-GEL(C)

Graphene-incorporated type I collagen cryogel

SEI

Scar elevation index

BM-MSCs

Bone marrow mesenchymal stem cells

EV

Extracellular vesicle

G-GEL

Graphene-type I collagen hydrogel

RT

Room temperature

COL-GEL

Type Ⅰ collagen hydrogel

COL-GEL(C)

Type Ⅰ collagen cryogel

HS-fibroblasts

Hypertrophic scar derived fibroblasts

STS

Staurosporine

TEM

Transmission electron microscopy

NTA

Nanoparticle tracking analysis

RT-qPCR

Real-Time quantitative reverse transcription PCR

ROIs

Regions of interest

SEM

Scanning electron microscopy

HE

Hematoxylin and eosin

IHC

Immunohistochemical

IF

Immunofluorescence

ANOVA

One-way analysis of variance

SNK test

Student-Newman-Keuls test

Author contributions

Authors contribution: Mengyuan Jiang: Methodology, Investigation, Formal analysis, Writing - original draft. Xiyuan Mao: Conceptualization, Supervision, Project administration, Writing - review & editing. Lu Zhang: Supervision, Methodology, Writing - review & editing.

Funding

This research was funded by the National Natural Science Foundation of China (Grant No. 82272295 and Grant No. 82002053).

Data availability

All data generated or analyzed during this study are included in this article and its supplementary information files.

Declarations

Competing interests

The authors declare no competing interests.

Ethical approval

This study involving human sample collection was conducted in accordance with the Declaration of Helsinki and was reviewed and approved by the Institutional Ethics Review Committee of Shanghai Jiao Tong University School of Medicine (SH9H-2025-TK282-1). Written informed consent was obtained from all human participants or their legal guardians prior to their inclusion in the study. All animal experiments were performed in compliance with the ARRIVE guidelines and the National Institutes of Health Guide for the Care and Use of Laboratory Animals, and were approved by the Institutional Ethics Review Committee of Shanghai Jiao Tong University School of Medicine (SH9H-2024-A961-SB).