Vitamin A-coupled MSCs- derived extracellular vesicles relieve acute liver injury by targeting hepatic stellate cells to deliver REDD1

Abstract

Mesenchymal stem cells (MSCs)-derived extracellular vesicles (EVs) offer great potential for treating liver injury. However, owing to their intrinsic surface characteristics, bare EVs are non-specifically distributed in the liver tissue after systemic administration, leading to limited therapeutic efficacy. Acute liver injury, often induced by acetaminophen overdose, can progress to life-threatening fibrosis, with hepatic stellate cells (HSCs) recognized as central drivers of this pathological process. While regulated in development and DNA damage response 1 (REDD1) has demonstrated antifibrotic effects by inhibiting HSCs activation, its clinical application has been hindered by challenges in targeted delivery. Utilizing the natural affinity of vitamin A (VA) for retinol binding protein receptors on HSCs, this study addresses the critical need for targeted therapies in acute liver injury by developing a novel delivery system based on VA-conjugated EVs (V-EVs) to transport the therapeutic gene REDD1 specifically to activated HSCs (aHSCs). After loading with REDD1 (V-EV^REDD1), the system showed enhanced cellular uptake in activated HSCs in vitro and effective hepatic accumulation in vivo. Treatment with V-EV^REDD1 significantly suppressed HSCs activation, reduced inflammation and hepatocyte apoptosis, improved liver function, and alleviated liver fibrosis in experimental models. This work highlights a promising strategy that combines targeted vesicle delivery with gene therapy, offering a potential avenue for improving the precision and efficacy of treatments for liver injury and fibrosis.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12951-026-04686-7.

Introduction

Acute liver injury is a rapidly progressive, life-threatening condition which can induced by multiple infectious, immune, metabolic and neoplastic diseases. Of these, acetaminophen (APAP) overdose is a leading cause for acute liver injury in developed countries [1]. Clinically, acute liver injury treatment remains limited and mostly supportive. When acute liver injury progresses to acute liver failure (ALF), orthotopic liver transplantation (OLT) is only accessible to a small proportion of patients due to the shortage of donor organs [2]. Therefore, there is an urgent need for emerging strategies for the treatment of acute liver injury.

After acute liver injury, the liver initiates a fibrogenesis program as part of its repair mechanism to protect the liver structure. However, excessive repair may result in the abnormal accumulation of intrahepatic connective tissue, leading to the formation of pathological fibrous scar deposits that compromise liver function and even progress to the development of liver fibrosis or cirrhosis in severe cases. Hepatic stellate cells (HSCs) are recognized as the key drivers of liver fibrosis in response to liver injury and inflammation, undergoing a transition from a quiescent state into an α- Smooth Muscle Actin (SMA)-positive myofibroblast-like phenotype upon liver injury [3, 4]. Therefore, blocking or reversing HSCs activation and reducing the production of extracellular matrix (ECM) become the pivotal target for treating liver fibrosis. Activated HSCs (aHSCs) abundantly express the receptor for retinol-binding protein (RBP) receptor on their cell surface. Accumulated evidences suggest that vitamin A (VA) binds to RBP in the circulation to form a complex, which is selectively internalized by aHSC through endocytosis mediated by RBP receptors [5–7], which provided the exploration of VA-decorated delivery systems with specific HSCs-targeting abilities [8, 9].

Regulated in development and DNA damage response 1 (REDD1) is the stress-responsive protein induced by various stresses such as hypoxia, alcohol, excessive reactive oxygen species (ROS) production, and inflammation [10]. Previously, we demonstrated that REDD1 overexpression alleviated HSCs activation, thus suppressing liver fibrosis [11]. However, gene therapy still faces substantial challenges in nucleic acid delivery, particularly in terms of targeted delivery and transfection efficiency [12]. An excellent delivery carrier achieving accurate entry and effective intracellular release into HSCs is necessary.

Mesenchymal stem cells (MSC) are extensively derived from umbilical cords (UC), placental, bone marrow, adipose tissue, playing therapeutic effects in tissue repair and regeneration [13, 14]. Our previous studies demonstrated the roles of MSC in attenuating liver injury and liver fibrosis [11, 15]. However, MSC in vivo survival rates are less than 6%, thus exhibiting minimal therapeutic effects [16]. Although allogeneic transplantation is also available, its clinical application has been limited by immunogenicity in vivo. MSC-derived extracellular vesicles (MSC-EVs) as mediators of cell-to-cell communication have been reported to carry functional contents and exhibit similar therapeutic potentials as the parent MSC [6]. Given their high biosafety and advantages including low rates of tumorigenicity and pulmonary embolism, the current research on MSC-EVs replacing MSC is becoming increasingly hot [17–19]. MSC-EVs have been widely used for the delivery of drugs, or small molecules, which exhibit low-toxicity, better biocompatibility and higher transfection performance [20]. However, non-specific distribution of MSC-EVs in the target tissue results in limited therapeutic efficacy.

In this work, we engineered MSC-EVs derived from human umbilical cord by hydrophobic insertion of VA conjugate and demonstrated that VA- loaded EV (V-EV) targeted HSCs in vitro and in vivo. Then, we transfected REDD1 gene into V-EV (V-EV^REDD1) to observe the its efficacy. The in vivo and in vitro experiments showed that V-EV^REDD1 not only effectively inhibited HSCs activation, but also mitigated liver inflammatory, improved liver function, alleviated liver fibrosis, and reduced hepatocyte apoptosis. While the anti-fibrotic function of REDD1 in HSCs has been established by previous studies, and VA-mediated targeting of nanocarriers or EVs to aHSCs has been reported, the combination of these two strategies has not been exported. This study combined the gene therapy and the excellent drug-loading characteristics of MSC-EVs for the treatment of liver injury, providing a practical, synergistic strategy to accelerate the clinical translation of MSC-EVs targeted delivery therapy in liver diseases (Scheme 1).

Scheme 1.

Schematic diagram of the protective effect and mechanism of V-EV^REDD1 on APAP/CCl4-induced liver injury

Materials and methods

MSCs culture and isolation of EV^REDD1

Human umbilical cord MSCs (HUMSC) were obtained and identified through full-term caesarian section deliveries with the informed consent of the parents as previously described by our team [21]. For subsequent experiments, 1 × 10⁶ passage 6 HUMSC were seeded and cultured in a T-75 flask with DMEM/F12 (Gibco, USA) containing 10% fetal bovine serum (FBS) and 1% penicillin-streptomycin at 37℃ in a humidified atmosphere with 5% CO₂.

To obtain EV expressing high levels of REDD1, cells were transfected with adenovirus-mediated REDD1 or the negative control vector (Ad-REDD1/NC) at a titer of 1 × 10⁹ pfu, as produced by Jikai Technology Co., Ltd. The construct sequences for REDD1 were as follows: REDD1(91934-1)-p1, GTGGATCCGAGCTCGGTACCCGCCACCATGCCTAGCCTCTGGGATCG; and REDD1(91934-1)-p2, ATATTTTATTACCGGTTTAATTAATCAACACTCTTCAATGAGCAGCTGCTCGGAG. To determine the optimal multiplicity of infection (MOI) for Ad-REDD1 transduction, MOI gradients of 10, 30, 100, and 300 were tested. At 24, 48 and 72 h post-transduction, the proportion of positive cells and fluorescence intensity were qualitatively assessed using fluorescence microscopy. Based on these observations, an MOI of 100 and a transduction period of 48 h were selected as the optimal conditions for subsequent experiments.

After trypsin digestion, the HUMSC pellet was washed twice with phosphate-buffered saline (PBS) and resuspend in sterile water at a density of 1 × 10⁷ cells per ml. To prepare EVs, the suspension was then physically extruded 20 times through a 400 nm polycarbonate porous membrane (Whatman, Cytiva, USA) using a mini-extruder (Avanti Polar Lipids) at room temperature. The extruded product was ultracentrifuged at 120,000 × g for 70 min at 4℃ using a Type 70 Ti rotor (Beckman Coulter, USA) to pellet the vesicles. BCA protein assay kit (Beyotime, China) was used to calculate the concentration of EV. Finally, purified HUMSC-derived EV was suspended in PBS and stored at −80℃ until further use.

Synthesis of DSPE-PEG-VA

DSPE-PEG-VA was synthesized using simple EDC/NHS chemistry according to the literature [22]. Briefly, retinoic acid (RA), EDC·HCl, and NHS were dissolved in anhydrous DMSO at a molar ratio of 0.5:1.1:1, and then stirred at 40 °C for 48 h. DSPE-PEG-NH₂ was then added at a RA-to-DSPE-PEG-NH2 molar ratio of 1.1:1, and the reaction continued for another 36 h. The reaction mixture was transferred to the dialysis bag (Molecular weight: 1 KDa) to remove small molecule impurities, and freeze-dried for 3 days to afforded the desired conjugate, pale yellow-colored DSPE-PEG-VA.

Preparation and characterization of V-EV^REDD1

The phospholipid insertion method was used to modify HUMSC-EVs to prepare V-EV^REDD1. In brief, 1 µg/µL bare EV^REDD1 was incubated with equal mass of DSPE-PEG-VA at 4 ℃ for 30 min. The amphiphilic DSPE-PEG-VA spontaneously inserts into the EV membrane through hydrophobic interactions between the DSPE tail and the lipid bilayer, yielding V-EV^REDD1.

Western blot analysis for markers: proteins were extracted with RIPA lysis buffer and quantified by BCA kit (Beyotime). Equal amounts of proteins (40 µg) were separated by SDS-PAGE and transferred to PVDF membranes (Merckmillipore, DE). After blocking with 5% skim milk powder at room temperature for 1 h, the membranes were incubated with primary antibodies against CD9 (1:800, ab92726, Abcam), CD63 (1:800, ab217345, Abcam), CD81 (1:800, ab109201, Abcam), and Calnexin (1:3000, ab22595, Abcam). Then the goat anti-rabbit secondary antibodies coupled with horseradish peroxidase was incubated and developed with chemiluminescence HRP substrate (BIO-RAD). Band quantification was performed using ImageJ software (National Institutes of Health, USA). For each target protein, the integrated density of the target band was normalized to that of the loading control band from the same sample. All quantified data were then expressed as fold change relative to the control group.

Nanoparticle tracking analyzer (NTA): The size distribution of bare EV and V-EV were measured using a nanoparticle tracking analyzer (NTA, Malvern, England).

Dynamic light scattering (DLS): The zeta potential of bare EV and V-EV were measured by dynamic light scattering measurements (DLS, Malvern, England).

Transmission electron microscope (TEM) detection: EV and V-EV were fixed with 2% glutaraldehyde for 2 h, followed by 1% osmic acid for 1–2 h as our previous study [15, 23]. After washing with PBS, the samples were dehydrated through a graded ethanol series (30%–100%, 15 min per step) and then washed with 100% acetone twice (20 min for each time). Infiltration was carried out using acetone-epoxy resin mixtures in a stepwise gradient for 8–12 h, followed by embedding in pure epoxy resin and polymerization at 60 °C for 48 h. Ultrathin Sects. (50–70 nm) were cut, stained with 2% uranyl acetate for 15 min and then with lead citrate, and examined under a TEM (JEOL, Japan).

NMR and FTIR: Nuclear magnetic resonance (NMR, AVANCE III, Bruker), Fourier transform infrared (FTIR) spectroscopy (Nicolet 6700, Thermo, USA) were used to confirm the successful synthesis of EV and V-EV by Shanghai Yanshuo Technology Co., Ltd (China).

In vitro cellular uptake

The mouse hepatic stellate cell line JS-1 and the human hepatocyte cell line 7702 were cultured separately in DMEM (Gibco, USA) supplemented with 10% FBS (Gibco, USA) and 1% penicillin-streptomycin (Gibco, USA). All cells were incubated at 37℃ in a humidified 5% CO₂ atmosphere.

Identification of VA targeted EVs: JS-1 cells were seeded in a 24-well plate and treated with TGF-β (5 ng/mL) for 24 h through Lipofectamine 8000 (Thermo, USA), PKH26-labeled EV and CY2-labeled VA were identified by immunofluorescence staining.

For HSCs activation, JS-1 cells were treated with TGF-β (5 ng/mL) for 24 h, while 7702 cells were treated by APAP at a concentration of 10 mM for 12 h. After performing three washes with PBS, the medium was replaced with 500 µL medium containing PKH26-labelled EV^REDD1 or V-EV^REDD1 (30 µg/mL) was added to each well and incubated at 37℃ for 24 h. DNA endocytosis was determined using laser confocal (SP8, Leica).

Animal experiment

Animal experiments for this study were approved by the Animal Experimentation Ethics Committee of Shanxi Medical University and were performed according to the committee’s guidelines. All C57BL/6 male mice (8 weeks) were received from Shanxi Medical University Animal Center, and were randomly divided into 5 groups (n = 6 mice per group) using a random number table to ensure allocation concealment: Control, model and 3 treatment groups, including EV with high expression of Ad-NC (EV^NC), EV with high expression of REDD1 (EV^REDD1), V-EV with high expression of REDD1 (V-EV^REDD1). All experimental procedures were conducted using double-blind method.

For model groups, CCl4 was dissolved in olive oil at a volume ratio of 1:3, yielding a final CCl4 concentration of 25% (v/v). The CCl4 mixture was administered intraperitoneally at a volume of 10 ml/kg, corresponding to a CCl4 does of 2.5 ml/kg, once daily for 4 consecutive days. For APAP treatment, mice were subjected to overnight fasting with free access to water, followed by an intraperitoneal injection of APAP dissolved in saline at a dose of 300 mg/kg for 1 time; For control group, equal amount of olive oil or saline was done; For the treatment group, the mice were injected with EV^NC, EV^REDD1, or V-EV^REDD1 at a weight ratio of 100 µg/20 g, all of which were normalized to total protein content. The injection was performed using an insulin syringe to ensure accuracy and minimize tissue trauma. The therapy was carried out 2 h after the last modeling administration. The serum and liver tissue were collected 24 h after treatment for further biochemical and histological analysis. Humane endpoints were predefined, and affected animals were promptly euthanized by cervical dislocation under isoflurane anesthesia. No unexpected deaths occurred prior to these endpoints.

In vivo tracking of V-EV^REDD1

EV or V-EV were labeled with the lipophilic near-infrared dye DIR. Briefly, 100 µg of EVs were incubated with 5 µM DIR in PBS at 37 °C for 30 min with gentle shaking. Unbound free dye was removed by ultracentrifugation at 120,000 × g for 70 min at 4 °C, followed by two additional washes with PBS. Fluorescence microscope was used to confirm the labeling efficiency.

After CCl4 injection, 100 µL (1 µg/µL) of DIR-labeled bare EV or V-EV were injected into mice with liver injury through tail vein. Time-dependent biodistribution was observed at 2 h, 12 h, 24 h, 72 h, 120 h post-injection using an in vivo imaging system (IVIS Spectrum, Revvity, USA) with 28 filters covering 430–850 nm. Heart, liver, spleen, lung and kidney both EV and V-EV groups were obtained to observe organ accumulation at 72 h. All fluorescence images were obtained at wavelengths of 748 nm (excitation) and 780 nm (emission). Exposure time was adjusted between 50 and 300 s based on signal intensity. Autofluorescence was corrected by subtracting pre-injection baseline images from post-injection images using Living Image software. Fluorescence signals were quantified by drawing a fixed circular ROI over various organ regions, and average fluorescence intensity (p/s/cm²/sr) was automatically calculated.

For immunofluorescence co-localization of PKH26 labeled-EV and α-SMA, the liver sections were treated with antibodies of anti-α-SMA (1:1000, ab124964, Abcam), and then incubated with fluorescent secondary antibodies. The nuclei were stained with DAPI. Images were harvested using confocal microscopy.

Assessment of the effects of V-EV^REDD1 on activated HSCs

HSCs were treated with TGF-β (5 ng/mL) for 24 h for their activation and incubated with medium containing EV^REDD1 or V-EV^REDD1 (30 µg/mL) for 24 h, as described above. For the detection of cell survival, CCK8 assay kit (Solarbio, China) was used, and the optical density level at 450 nm was detected by enzyme labeling instrument (Bio-Rad, USA). The cell viability was analyzed by normalizing to the control group. REDD1 protein expression were detected by western blot with primary antibody against REDD1 (1:1000, ab106356, Abcam).

To clarify HSCs activation, RNA was isolated with TRIzol (Life Technology, USA), and reversely transcribed to cDNA through specific reverse transcription kit (Tiangen, Beijing, China). cDNA together SYBR Green (Tiangen, Beijing, China) and primers of α-SMA and collagen I were used for qRT-PCR. GAPDH served as internal reference. The relative RNA abundance was calculated according to 2 − ΔΔCt method. All sequence information of primers was listed in Table 1.

Table 1.

The sequences for all primers

To detect activated HSCs lipid droplets, BODIPY reagent kit (C2053, Beyotime, China) was performed. Meanwhile, Oil Red O stain Kit (C0157, Beyotime, China) was also been used for lipid staining within HSCs. Briefly, for BODIPY staining, cells were fixed with 4% paraformaldehyde for 15 min, incubated with BODIPY 493/503 (1:1000) for 30 min at 37 °C, and counterstained with DAPI. For Oil Red O staining, fixed cells were rinsed with 60% isopropanol, and stained with filtered Oil Red O solution (0.3%) for 15 min, and then counterstained with hematoxylin for 1 min. Images were captured by confocal microscopy.

In addition, immunofluorescence staining was performed to quantality HSCs autophagy and activation. Briefly, the samples were treated with the primary antibodies of anti-α-SMA (1:1000, ab124964, Abcam) and anti-LC3 (1:1000, T55992, Abmart), and then incubated with fluorescent secondary antibodies. The nuclei were stained with DAPI, and the cells were observed using a confocal microscope. For immunofluorescence quantification, at least five random fields per sample were captured under identical acquisition settings. All image analyses were conducted by an investigator blinded to the experimental groups.

Assessment of the effects of V-EV^REDD1 on hepatocytes

After a 12 h exposure to APAP (10 mM), 7702 cells culture medium was replaced with medium containing EV^REDD1 or V-EV^REDD1 (30 µg/mL) for 24 h. Then, alanine aminotransferase (ALT), aspartate aminotransferase (AST), and lactate dehydrogenase (LDH) were assessed using specific kits (C009-2-1, C010-2-1, A020-1–2, Jiancheng, Nanjing, China) following the instructions of the manufacturer. Reactive oxygen species (ROS) assay kit (S0033, Beyotime, China) was used to detect hepatocytes ROS. Image J was used to quantitative analysis fluorescence. Western blot of Bax (1:1500, T40051, Abmart), Bcl-2 (1:800, 381702, Zhengneng), cleaved caspase 3 (1:1000, 341034, Zhengneng) and caspase 3 (1:1000, 19677, Proteintech) were performed to detect hepatocytes apoptosis.

Evaluation of liver function and inflammation in vivo

Liver weight ratio was performed throughout the study. Before collecting liver tissue and blood at the end of the experiment, RFLSI ZW laser speckle blood flow imaging system (RWD, China) was used to observe the distribution of liver blood perfusion volume. Serum levels of ALT, AST, hydroxyproline (HYP), glutathione (GSH), propylene glycol (MDA) were determined using specific kits (C009, C010, A030, A006, A003, Jiancheng, Nanjing, China). RT-qPCR analysis was conducted to evaluate the expression levels of genes associated with interleukin (IL)−6, IL-10, TNF α, TGF β. Primers are outlined in Table 1. Immunofluorescence staining of F4/80 (1:150, 263101, Zhengneng), CD86 (1:150, R380350, Zhengneng), CD206 (1:200, 251716, Zhengneng) were performed to evaluate immune regulation. All biochemical and molecular assays were repeated three times, with calibration curves (R² > 0.99) using kit standards.

Assessment of liver regeneration in vivo

For immunohistochemistry, liver tissues were surgically removed and fixed 4% neutral formaldehyde before being embedded in paraffin and sectioned. the sections were incubated with primary antibodies, including cleaved caspase 3 (1:200, 341034, Zhengneng), caspase 3 (1:300, 19677, Proteintech), caspase 9 (1:300, 10380, Proteintech), which were followed by incubation with secondary antibodies and a DNA chromogenic reaction. ImageJ software was employed to quantify the positive area expression.

For TUNEL staining, TUNEL assay kit (40307ES60, Yeasen, Shanghai, China) was performed, and DAPI was used to stain the cell slides fixed with 4% paraformaldehyde fixation solution.

Histology quantification was performed using ImageJ. For each liver section, five randomly selected non-overlapping fields were analyzed, and the percentage of positive cells was quantified per field. The mean value of the five fields was calculated per sample. Three non-consecutive sections per animal were examined. All assessments were conducted by an investigator blinded to group allocation.

Antifibrosis Efficacy in vivo

Formalin-fixed, paraffin-embedded liver tissue specimens were sliced into 5-µm-thick sections. The histological architecture and fibrotic areas were evaluated using H&E, Sirius red and Masson staining. Moreover, immunostaining of tissue sections was performed with α-SMA (1:500, ab124964, Abcam), collagen I (1:400, bs-0578R, Bioss, Beijing, China) and mTOR (1:100, T56571, Abmart) antibodies to determine levels of collagen deposition. RT-qPCR analysis was performed to evaluate the mRNA levels of α-SMA and collagen I. Immunofluorescence co staining of REDD1 (1:500, ab106356, Abcam) and α-SMA to clarify the role of REDD1 in liver fibrosis.

Statistical analysis

All statistical analyses were conducted using GraphPad Prism 8 (GraphPad Software LLC, USA). Data were expressed as the mean ± standard deviation. Statistically significant differences among groups were analyzed using one-way or two-way ANOVA, or Student’s t-test, as appropriate. The number of biological replicates or animals per group is provided in the corresponding figure legends. Statistical significance was assigned as *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001.

Results

Preparation and characterization of V-EV

MSC-derived EVs demonstrate therapeutic potential in tissue regeneration through their unique paracrine signaling and proven efficacy in attenuating chronic hepatic injury [13, 24]. The human umbilical cord mesenchymal stem cells (HUMSC) used in this study were isolated and identified in our previous study [25]. First, the optimal MOI for Ad-REDD1 transfection into HUMSC was determined (Supplementary Fig. 1 A), and MOI of 100 with a 48-h transduction period was selected as the optimal condition for subsequent experiments. Moreover, live/dead staining confirmed that Ad-REDD1 had no effect on HUMSC viability (Supplementary Fig. 1B).

To achieve targeted delivery to HSCs, we engineered VA-conjugated small EVs (V-EVs) by incorporating DSPE-PEG-VA into the membranes of bare EVs derived from HUMSC (Fig. 1A). The amphiphilic DSPE-PEG-VA spontaneously inserts into the EV membrane through hydrophobic interactions between the DSPE tail and the lipid bilayer (Fig. 1B). The marker proteins of V-EV, including CD9, CD63, CD81, were highly presented while Calnexin was negative expression (Fig. 1C), indicating that VA modification did not alter the fundamental characteristics of EVs. Co-localization of fluorescent signals from CY2-labeled VA and PKH26-labeled EVs on the same particles confirmed successful conjugation of VA onto the EVs surface (Fig. 1D).

Fig. 1.

Preparation and characterization of V-EV. (A) Schematic illustration of the phospholipid insertion-based surface modification to prepare V-EV. (B) Conjugation stoichimetry modified by VA. (C) Western blot analysis of characteristic protein markers in EV and V-EV, including CD9, CD63, CD81 and Calnexin. (D) Fluorescence intensity of CY2 -labeled EV. Scale bars: 100 μm. (E) The TEM images of EV and V-EV. (F) The NTA of EV and V-EV. (G) The zeta potentials of EV and V-EV. (H) The FTIR spectra of EV and V-EV. (I) The NMR spectra of -NH₂ and -COOH. Data expressed as mean ± SD

V-EVs were further characterized by TEM, NTA, and zeta-potential measurement to evaluate isolation efficiency and physical properties. TEM images revealed that V-EVs exhibited intact, oval morphologies bounded by a distinct lipid bilayer (Fig. 1E). NTA showed that the average particle size of both unmodified EVs and V-EVs was approximately 160 nm (Fig. 1F; Table 2). In addition, the zeta potentials of EVs and V-EVs showed no significant difference (Fig. 1G), indicating preserved colloidal stability. Successful conjugation was further verified by FTIR spectroscopy and NMR, which displayed characteristic peaks corresponding to VA linkage (Fig. 1H, I). Notably, when EVs and V-EVs were incubated in FBS, their particle sizes gradually increased over time (Supplementary Fig. 2); nevertheless, no significant degradation or aggregation was observed, supporting their stability and suitability for subsequent in vivo applications. Collectively, these results confirmed the successful fabrication of V-EVs with retained structural integrity, providing a stable platform for subsequent functional studies.

Table 2.

NTA comparison of EV and V-EV

To evaluate homing capacity in vivo, DiR-labeled EV^REDD1, or V-EV^REDD1 were administered to CCl4-induced fibrotic mice. Longitudinal in vivo imaging revealed significantly higher hepatic fluorescence intensity in the V-EV^REDD1 group compared to EV^REDD1 controls at all time points (Fig. 2E, F). Ex vivo imaging revealed that the fluorescence was primarily localized in the liver and lung, with minimal presence in the heart, spleen and kidney (Fig. 2G). Further analysis indicated that the fluorescence intensity in the liver of the V-EV^REDD1 group was significantly greater compared to EV^REDD1 group (Fig. 2H). While some degree of pulmonary retention was observed for V-EV^REDD1 group, hepatic accumulation was significantly greater than pulmonary accumulation within this group, confirming improved liver-targeting efficiency. To support the claim of in vivo HSC targeting, we performed immunofluorescence co-localization of PKH26-labeled EV with α-SMA in liver sections from control, APAP-treated, and CCl4-treated mice. As shown in Fig. 2I, in control mice, both EV and α-SMA signals were minimal. In contrast, in APAP- and CCl4-induced liver injury models, strong co-localization of EV and α-SMA was observed. Notably, the V-EV group exhibited significantly greater fluorescence overlap of EV with α-SMA compared to the unmodified EV group, confirming that VA conjugation confers HSC-targeting capability in vivo. These data establish that VA functionalization confers efficient and specific targeting of aHSCs in both cellular and animal models.

Fig. 2.

In vitro and in vivo HSCs targeting capability of V-EV^REDD1. (A) Fluorescence image of HUMSC transfected with REDD1. Scale bars: 100 μm. (B) The mRNA levels of REDD1 in HUMSC (n = 3). (C) Cellular uptake of PKH26-labeled EV^REDD1 and V-EV^REDD1 by quiescent/activated JS-1 cells and 7702 cells. (D) Fluorescence intensity of PKH26-labeled EV^REDD1 and V-EV^REDD1 in cells (n = 3). (E) In vivo biodistribution images of EV^REDD1 and V-EV ^REDD1 in APAP mice. (F) The time-dependent fluorescence intensity of EV^REDD1 and V-EV^REDD1 in liver tissues (n = 3). (G) Fluorescence images of EV^REDD1 and V-EV^REDD1 across five organs including heart, liver, spleen, lung, kidney in APAP mice. (H) The fluorescence intensity of EV^REDD1 and V-EV^REDD1 in five organs (n = 3). (I) Representative images of immunofluorescence staining of PKH26 labeled-EV and α-SMA in control, APAP/CCl4-treated mice. Scale bars: 50 μm. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. Data expressed as mean ± SD

Inhibition of V-EV^REDD1 on activated HSCs

Previous studies have demonstrated that autophagy play pivotal roles in the processes of HSCs activation and liver fibrosis [11, 23, 26]. To determine the optimal time point for HSCs activation experiments, JS-1 cells were treated by TGF-β with 6 h, 12 h, 24 h, 48 h, and fluorescence signals peaked at 24 h (Fig. 3A). However, CCK8 assays indicated that the cell viability of JS-1 cells significantly downregulated at 24 h (Fig. 3B). Based on the optimal balance between activation efficacy and cell survival, the 24 h time point was selected for subsequent experiments, because it provided strong HSCs activation signals with acceptable cell viability and suitable morphology for downstream experiments.

Fig. 3.

Inhibition of V-EV^REDD1 on activated HSCs. (A) Fluorescent images of TGF-β plasmid transfection into JS-1. Scale bars: 100 μm. (B) CCK8 assay for cells viability of HSCs with different concentrations of TGF-β (n = 3). (C) The protein expression of REDD1 in HSCs from different treatment groups (n = 3). (D) Representative images of immunofluorescence staining of α-SMA and LC3 in HSCs from different treatment groups. Scale bars: 50 μm. (E) Representative images of immunofluorescence staining of BODIPY in HSCs from different treatment groups. Scale bars: 50 μm. (F) Oil Red O staining in HSCs from different treatment groups. Scale bars: 50 μm. (G) The BODIPY representative images are quantified by ImageJ software (n = 3). (H) The mRNA levels of α-SMA and collagen I in HSCs from different treatment groups (n = 3). *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. Data expressed as mean ± SD

Given the enhanced targeting, we performed western blot to quantify REDD1 protein levels in HSCs treated with EV^REDD1 versus V-EV^REDD1. The results showed that V-EV^REDD1 leads to significantly higher REDD1 expression in HSCs (Fig. 3C), indicating that the treatment of V-EV^REDD1 is not only due to improved HSCs targeting but also to more efficient intracellular delivery or expression of REDD1. We next investigated the functional impact of V-EV^REDD1 on HSCs autophagy and activation. Following activation with TGF-β, JS-1 cells were incubated with EV^NC, EV^REDD1 or V-EV^REDD1 for 24 h. Immunofluorescence assays showed that V-EV^REDD1 significantly suppressed the HSCs activation and autophagy stimulated by TGF-β, as evidenced by reduced α-SMA expression and decreased LC3-positive autophagic vesicles (Fig. 3D).

Autophagy is involved in regulating the dynamic balance of lipid droplets in HSCs. In aHSCs, stimulated autophagy decreased lipid droplet content while impaired autophagy function led to lipid accumulation [27]. As a marker of HSCs autophagy activity, BODIPY was markedly restored by V-EV^REDD1 compared to EV^REDD1 group (Fig. 3E, G). Simultaneously, Oil Red O staining further verified that V-EV^REDD1 promoted the lipid droplet accumulation in HSCs, approaching normal levels (Fig. 3F). Furthermore, RT-qPCR analysis revealed the most significantly decrease of key fibrogenic genes including α-SMA and collagen I levels following administration of V-EV^REDD1 (Fig. 3H). Taken together, these results strongly suggest that V-EV^REDD1 hold great potential for facilitating the reversion of aHSCs back to a quiescent state.

Protection of V-EV^REDD1 on APAP-induced hepatocytes injury in vitro

Hepatocytes are of utmost significance in the process of liver repair [28]. We next explored the protective effect of V-EV^REDD1 on hepatocytes. CCK8 assays was performed to evaluate 7702 hepatocyte cells activity. The results showed that the cell viability of 7702 cells remained above 90% when exposed to APAP at concentrations ranging from 2.5 mM to 10 mM for 6 h and 12 h; however, viability decreased significantly at 24 h and 48 h (Fig. 4A). The intracellular ROS levels also revealed that 10 mM APAP exhibited significantly increased ROS accumulation compared to 2.5 mM and 5 mM (Supplementary Fig. 3). Subsequently, 10 mM APAP-injured 7702 cells for 12 h were used for subsequent experiments. Serum biochemistry revealed significant decreases in liver injury markers (ALT, AST and LDH) in the EV^REDD1 and V-EV^REDD1 group relative to the APAP model group; while there was no further reduction in the V-EV^REDD1 group compared to EV^REDD1 group (Fig. 4B). Furthermore, intracellular ROS accumulation was also attenuated by EV^REDD1 and V-EV^REDD1, and V-EV^REDD1 has not a more recovery (Fig. 4C, E), indicating that V-EV^REDD1 targeted HSCs rather than hepatocytes.

Fig. 4.

Protection of V-EV^REDD1 on APAP-induced hepatocytes injury in vitro. (A) Cell viability of 7702 with the concentrations of APAP (n = 3). (B) ALT, AST, and LDH levels of the supernatant of different treated 7702 cells (n = 3). (C) Intracellular ROS levels of different treated 7702 cells by fluorescent microscope. (D) Mitochondrial membrane potential of different treated 7702 cells by fluorescent microscope. (E) The ROS representative images are quantified by ImageJ software (n = 3). (F) The fluorescence ratio of aggregate/monomer are quantified by ImageJ software (n = 3). (G, H) The protein expression of bax, bcl-2, caspase-3 and cleaved caspase-3 in hepatocytes from different treatment groups (n = 3). (I) The mitochondrial oxidative stress levels in hepatocytes from different treatment groups. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. Data expressed as mean ± SD

Mitochondrial damage could trigger oxidative stress, leading to an increase in ROS production [29]. Mitochondrial membrane potential detection is mainly used to evaluate the apoptosis of mitochondria. As shown as in Fig. 4D, F, EV^REDD1 and V-EV^REDD1 both increased the membrane potential reduced by APAP, indicating improved cellular energetics. Moreover, western blot analysis revealed that EV^REDD1 and V-EV^REDD1 led to a remarkable reduction in the expression of pro-apoptotic genes (Bax, Cleaved-caspase 3) and an increase in anti-apoptosis genes Bcl-2 (Fig. 4G, H), suggesting alleviated hepatocytes apoptosis.

Abnormal mitochondrial morphology is a critical indicator of liver cell dysfunction [30]. TEM analysis showed that the ultrastructure of hepatocyte mitochondria in APAP group was severely damaged. Compared with the control group, the mitochondria exhibited crista swelling, a disordered structure, and a reduced crista density (Fig. 4I). After treatment with EV^REDD1 and V-EV^REDD1, there was a remarkable improvement in the integrity of these mitochondrial structures. Collectively, these results indicated that, like EVs, V-EV^REDD1 also had cell protective and anti-apoptotic effects on liver cells, contributing to overall liver recovery.

Amelioration of drug-induced liver injury by V-EV^REDD1 in vivo

To assess whether V-EV^REDD1 was sufficient to enhance the therapeutic effect in vivo, APAP/CCl4-induced liver injury mice were administered systemic injections according to the strategy diagram (Fig. 5A). Liver/Body weight ratio demonstrated significantly decrease in V-EV^REDD1-treated mice versus EV^REDD1, both surpassing EV^NC group (Fig. 5B). Serum biochemical analysis showed that both EV^REDD1 and V-EV^REDD1 improved liver function, reflected by decreased serum ALT, AST and HYP levels, elevated GSH, and reduced MDA, while V-EV^REDD1 exhibited a more significant mitigation (Fig. 5C, E). Meanwhile, laser speckle imaging indicated that compared with EV^REDD1 treatment, there were more higher liver tissue perfusion and reduced stiffness in the liver tissue of V-EV^REDD1-treated group, suggesting amelioration of liver tissue ischemia (Fig. 5D, F). These data suggest that the transplantation of V-EV^REDD1 has a positive impact on the recovery of liver damage. V-EV^REDD1 exhibited superior therapeutic outcomes compared to EV^REDD1 group due to enhanced targeting.

Fig. 5.

Amelioration of drug-induced liver injury by V-EV^REDD1 in vivo. (A) The strategy diagram of V-EV^REDD1 for liver injury. (B) Liver/body ratio in APAP mice with different treatment (n = 3 mice). (C, E) Serum ALT, AST, GSH and MDA levels in APAP/CCl4 mice with different treatment (n = 3 mice). (D, F) Representative images of speckle imaging in APAP/CCl4 mice with different treatment. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. Data expressed as mean ± SD

V-EV^REDD1 reduces liver inflammation in vivo

Liver inflammation is the core driving factor leading to liver damage. If left uncontrolled, the persistent inflammatory response and liver cell damage can form a vicious cycle, jointly driving liver disease towards fibrosis, cirrhosis, and even liver cancer [31]. Here we evaluate whether the anti-inflammatory effect of V-EV^REDD1 better. It can be directly seen that the liver color of the mice with liver injury is slightly white with a grainy sensation, especially in the CCl4 model group. After V-EV^REDD1 treatment, the liver became rosy and smooth, similar to the liver in normal group than that of EV^REDD1 treatment (Fig. 6A, Supplementary Fig. 4 A). Histopathological assessment via H&E staining revealed severe inflammatory infiltrates around the hepatic sinusoids, disrupted hepatic lobular architecture and expanded interstitial spaces in APAP/CCl4-treated mice compared to healthy controls (Fig. 6A, Supplementary Fig. 4 A). Both EV^REDD1 and V-EV^REDD1 treatments substantially reduced inflammatory infiltration and architectural damage, with V-EV^REDD1 exhibiting superior restorative effects, which was quantitatively evaluated (Fig. 6B, Supplementary Fig. 4B). Correspondingly, RT-qPCR analysis demonstrated V-EV^REDD1 presented a more notably affect in the downregulation of pro-inflammatory mediators including IL-6, TNF-α, TGF-β and upregulation of the anti-inflammatory cytokine IL-10 compared with EV^REDD1 (Fig. 6C, Supplementary Fig. 4 C). These data highlight the ability of V-EV^REDD1 to modulate the hepatic immune microenvironment, thereby attenuating inflammation-associated tissue injury.

Fig. 6.

V-EV^REDD1 reduced APAP-induced liver inflammation. (A) General observation of liver and H&E staining representative images of liver tissue in APAP mice with different treatment, and (B) the corresponding histological score (n = 3 mice). (C) The relative mRNA expression of IL-6, IL-10, TGF-β, and TNF-α in APAP mice liver tissues with different treatment (n = 3 mice) (D) Representative images of immunofluorescence staining of F4/80, CD86 and CD206 in APAP mice liver tissues with different treatment. Scale bars: 100 μm. *p < 0.05, **p < 0.001, ***p < 0.001, ****p < 0.0001. Data expressed as mean ± SD

To validate the observed cytokine changes are associated with functional immune modulation in vivo, immunofluorescence staining was performed to assess the influence of V-EV^REDD1 on macrophage phenotypes. The results revealed that relative to the EV^REDD1 group, the expression of the M1 marker CD86 in liver macrophages was markedly reduced in the V-EV^REDD1-treated groups. Conversely, the M2 marker CD206 showed a remarkable increase (Fig. 6D, Supplementary Fig. 4D). These data strongly suggest that V-EV^REDD1 can effectively attenuate inflammatory responses and inhibit the polarization of M1 macrophages.

V-EV^REDD1 suppresses hepatocyte apoptosis in vivo

To further delineate the mechanisms underlying hepatic repair, we assessed apoptosis. Immunohistochemistry analysis revealed that APAP overdose triggered an anti-apoptosis phenotype, with elevated apoptosis genes including cleaved caspase-3, caspase-3 and caspase-9 in liver tissues (Fig. 7A-D). Treatment with EV^REDD1 and V-EV^REDD1 significantly reversed reduced cleaved caspase-3, caspase-3 and caspase-9 expression, with V-EV^REDD1 showed the most significant effect. Notably, all the phenomenon was also observed in CCl4 liver injury mice (Fig. 7E-H). Moreover, TUNEL staining showed that the APAP control group exhibited clear signs of hepatocyte apoptosis, whereas EV^REDD1 and V-EV^REDD1-treated groups showed minimal or no apoptotic activity, especially V-EV^REDD1 exhibited more significant (Fig. 7I, K). The liver injury model treatment induced by CCl4 showed the same results (Fig. 7J, L). Collectively, these findings suggested that V-EV^REDD1 not only inhibits apoptosis but also promotes hepatocyte regeneration in vivo.

Fig. 7.

V-EV^REDD1 suppressed hepatocyte apoptosis in vivo. (A, E) Representative immunohistochemical staining images of anti-cleaved caspase-3, anti-caspase3 and anti-caspase9 in APAP/CCl4 mice liver tissues with different treatment, and (B, C, D, F, G, H) the corresponding quantitative analysis of positive areas stained with cleaved caspase-3, caspase3 and caspase9 (n = 3 mice). (I, J) Representative TUNEL staining images in APAP/CCl4 mice liver tissues with different treatment, and (K, L) the corresponding quantitative analysis stained with apoptotic cells (n = 3 mice). *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. Data expressed as mean ± SD

Anti fibrotic effect of V-EV^REDD1 on liver fibrosis

Finally, we evaluated the anti-fibrotic potential of V-EV^REDD1 in models of APAP- and CCl4-induced fibrosis. Accordingly, Sirius Red and Masson’s trichrome staining revealed severe collagen deposition in APAP/CCl4-treated mice compared to controls, which was strongly decreased after by EV^REDD1 and V-EV^REDD1, with V-EV^REDD1 exhibiting superior restorative effects (Fig. 8A, Supplementary Fig. 5 A). Quantitative evaluation of fibrosis through Sirius red staining and Masson’s trichrome demonstrated a significant reduction in collagen deposition in the V-EV^REDD1 group versus EV^REDD1 and the APAP/CCl4-induced fibrosis group (Fig. 8C; Supplementary Fig. 5 C).

Fig. 8.

Anti fibrotic effect of V-EV^REDD1 on APAP-induced liver fibrosis. (A) Representative images of Sirius red, Masson in APAP mice liver tissues with different treatment, and (C) the corresponding histological score (n = 3 mice). (B) Representative images of immunohistochemical staining of anti-TGFβ, anti-α-SMA, anti-collagen I in APAP mice liver tissues with different treatment, and (D) the corresponding histological score (n = 3 mice). (E) The relative mRNA expression of α-SMA and collagen I in APAP mice liver tissues with different treatment (n = 3 mice). (F) Representative images of immunofluorescence co-staining of REDD1 and α-SMA in APAP mice liver tissues with different treatment. Scale bars: 100 μm. (G) H&E staining representative images of heart, liver, spleen, lung and kidney in control and EV^NC mice. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. Data expressed as mean ± SD

Immunohistochemical analysis further confirmed the strongest antifibrotic efficacy of V-EV^REDD1, showing a most significantly reduction in TGF-β (a factor promoting liver fibrosis), α-SMA (HSCs activation marker) and collagen I compared to the EV^REDD1 group (Fig. 8B, D; Supplementary Fig. 5B, D). Transcriptional analysis confirmed markedly downregulation of fibrotic genes (α-SMA and collagen I) in V-EV^REDD1 group (Fig. 8E; Supplementary Fig. 5E). Additionally, the immunofluorescence co-staining of liver sections from APAP/CCl4-induced fibrotic mice, revealed spatial proximity between REDD1 and α-SMA, indicating the increase in REDD1 led to a decrease of α-SMA in aHSCs (Fig. 8F; Supplementary Fig. 5 F). Safety assessments demonstrated no pathological abnormalities in major organs (heart, liver, lung, spleen, kidney) via H&E staining (Fig. 8G). These results collectively establish that V-EV^REDD1 delivers REDD1 specifically to aHSCs and robustly attenuates the progression of liver fibrosis.

Given the established role of REDD1 as a negative regulator of mTORC1, we further investigated the mTORC1 autophagy axis regulated by REDD1. Immunohistochemical staining for mTORC1 in liver tissues revealed that compared to EV^REDD1 group, V-EV^REDD1 significantly suppressed mTORC1 expression (Supplementary Fig. 6A-D), indicating the protective effect of V-EV^REDD1 on hepatocytes. To determine the HSCs specificity of V-EV^REDD1, we performed double immunofluorescence co localization of α SMA (a marker of activated HSCs) and LC3 (an autophagosome marker). Our results showed that V-EV^REDD1 reduced both α-SMA and LC3 signals in HSCs (Supplementary Fig. 6E, F), suggesting that V-EV^REDD1 overexpression attenuates HSCs activation by targeting suppression of HSCs autophagy. These data demonstrate that the modulatory effect of V-EV^REDD1 on autophagy is HSCs selective and does not broadly affect hepatic parenchymal cells, which likely contributes to the observed anti-fibrotic effects.

Discussion

Acute liver injury is often triggered by drug with a high risk of progression to life-threatening fibrosis and organ failure [32]. Currently, N-acetylcysteine is the only clinically approved antidote, but its efficacy is limited by a narrow therapeutic window and severe side effects [33]. Therefore, there is a pressing need to develop novel therapeutic strategies that can offer more effective protection against liver injury. In this study, we developed a precision-delivery platform by VA-conjugated EVs derived from HUMSCs to transport the therapeutic gene REDD1 specifically to aHSCs. The principal novelty of our work is not the identification of a new molecular mechanism or an unprecedented targeting ligand, but rather the combination of these two established elements into a unified VA-conjugated EV system for REDD1 gene delivery, which achieving HSCs-specific delivery therapy. Our results demonstrate that this targeted system, V-EV^REDD1, significantly alleviates acute liver injury by suppressing HSCs activation, attenuating hepatocyte apoptosis, modulating inflammation, and improving liver function in preclinical models.

The central innovation of our work lies in the strategic fusion of two promising yet individually limited therapeutic concepts: the inherent homing capability of VA toward HSCs and the excellent biocompatibility and cargo-delivery capacity of MSC-EVs. While the therapeutic potential of MSC-EVs in liver repair is well-documented [34, 35], their non-specific biodistribution significantly hampers efficacy and dose requirements [36]. Concurrently, strategies exploiting the VA-RBP receptor axis has emerged an effective strategy for HSCs-targeted delivery, as demonstrated by VA-coupled polymers or liposomes [37, 38]. Here, we constructed V-EVs through hydrophobic insertion of DSPE-PEG-VA, which endowed the vesicles with HSCs-targeting ability while preserving their native physicochemical characteristics. Both in vitro and in vivo imaging confirmed that VA modification markedly enhanced the accumulation of EVs within injured livers and specifically in aHSCs, thereby overcoming the off-target distribution observed with unmodified EVs. This finding aligns with and extends the work [22], confirming VA as a robust ligand for in vivo HSCs targeting within an EV-based system.

REDD1 gene is induced by various stresses and has been studied in cell proliferation and survival. Research showed that REDD1 is a novel molecule that reduced susceptibility to oxidant-induced liver injury and liver fibrosis [10, 39]. The choice of REDD1 as the therapeutic cargo is underpinned by our prior research identifying its potent role in suppressing HSCs activation and autophagy, key drivers of fibrogenesis [11]. However, the clinical translation of gene therapy like REDD1 delivery is notoriously constrained by the “delivery problem” – achieving sufficient nucleic acid concentration at the precise disease site without systemic toxicity [40]. Our V-EVs platform effectively addresses this limitation. The targeted delivery of REDD1 via V-EV^REDD1 not only replicated but significantly amplified the antifibrotic effects observed with bare EV^REDD1. V-EV^REDD1 exerted a dual cytoprotective effect in the fibrotic liver microenvironment. In hepatocytes, it robustly alleviated oxidative stress, restored mitochondrial membrane potential, and reduced apoptosis, thereby attenuating the release of paracrine pro-fibrotic signals. Concurrently, in HSCs, it directly suppressed autophagy and activation. By simultaneously mitigating the activator signals from injured hepatocytes and blocking the autophagic fuel required for HSCs transformation, V-EV^REDD1 synergistically disrupted the hepatocyte-HSCs crosstalk that drives fibrosis progression.

In murine models of APAP- and CCl4-induced liver injury, systemic administration of V-EV^REDD1 elicited comprehensive therapeutic benefits. This multi-faceted amelioration including improving liver function, reducing oxidative stress, modulating the hepatic inflammatory milieu, decreasing collagen deposition, and enhancing tissue perfusion, collectively underscored a shift from injury and fibrosis towards repair and regeneration. Importantly, the spatial co-localization of delivered REDD1 with α-SMA+ cells in fibrotic livers provides direct visual evidence of successful targeted gene delivery, indicating that the superior effects of V-EV^REDD1 over EV^REDD1 in all these parameters not only attributed to enhanced HSCs-specific biodistribution, but the broad hepatoprotection conferred by the intrinsic bioactivity of MSC-EV carrier.

The liver is a complex organ comprising a variety of cell types. Approximately 60%–80% of cells in the mammalian liver are hepatocytes, which are the primary cell groups responsible for various physiological functions [41]. Notably, in vitro experiments using APAP-injured hepatocytes revealed that the protective effects of V-EV^REDD1 were comparable to those of unmodified EV^REDD1. This observation demonstrates that the hepatocyte protection does not drive from REDD1-specific activity or the VA modification. Rather, it originates from the intrinsic cytoprotective properties of MSC-EV vector itself, which is well-documented to exert cytoprotective, anti- apoptotic, and anti-inflammatory effects on injured hepatocytes through the transfer of miRNAs, proteins, and lipids [42, 43]. The anti-fibrotic effect of V-EV^REDD1 evidenced by superior suppression of HSCs autophagy, activation and collagen deposition compared to EV^REDD1 is attributed to the targeted delivery of REDD1 to aHSCs, not to direct hepatocyte protection. Thus, V-EV^REDD1 retains the baseline, broad hepatoprotective efficacy of native MSC-EVs while adding a potent, targeted anti-fibrotic action through a greater proportion REDD1 delivery to aHSCs, thereby addressing multiple pathological axes of liver injury with a single agent.

Unlike synthetic nanoparticles that may face biocompatibility or clearance issues, MSC-EVs are natural biological entities with low immunogenicity and inherent tissue-repair signals. The modification strategy preserves these advantages while conferring precise targeting. This study thus moves beyond the current hotspot of using MSC-EVs as mere replacement therapies or non-targeted delivery vehicles, advancing them into the realm of precision medicine for liver diseases.

Conclusion

In summary, we developed a safe and effective delivery system for treating liver injury. By extracting HUMSC-derived EVs and engineering a VA-modified system loaded with REDD1 (V-EV^REDD1), we demonstrated their potential in protecting APAP/CCl4-induced liver injury. V-EV^REDD1 displayed HSCs autophagy and activation inhibition, anti-hepatocyte apoptosis, antioxidative, and hepatoprotective properties in vitro, with the ability to target the liver via RBP receptor-mediated uptake by aHSCs. In vivo, intravenous administration of V-EV^REDD1 effectively alleviated oxidative stress, reduced inflammation, suppressed cell apoptosis, attenuated liver fibrosis. These findings underscore V-EV^REDD1 as a promising platform, offering a novel strategy to prevent and treat liver injury and address drug-induced hepatotoxicity.

Author contributions

T-J H designed and performed the research, analyzed the data, and wrote the manuscript. Z-X L, X-R S, X-Y F, L-Y W performed the research and analyzed the data, J-J R and X-N L supervised the research and revised the manuscript. All authors read and approved the final manuscript.

Funding

This study was supported by Central guidance for local scientific and technological development funding projects (YDZJSX2025D068) and China postdoctoral science foundation (2025M782136).

Data availability

No datasets were generated or analysed during the current study.

Declarations

Ethics approval and consent to participate

The animal study was reviewed and approved by the Institutional Animal Care and Use Committee of Shanxi Medical University, and the license Key was SCXK2021-0006.

Competing interests

The authors declare no competing interests.