From targeted delivery to inflammation suppression: engineering plant exosomes for heart failure therapy post-infarction

Abstract

Plant-derived exosomes are natural nanovesicles rich in bioactive compounds and show promise for tissue regeneration. However, their clinical use is limited by poor stability and low targeting accuracy. To overcome these issues, we developed an engineered approach to enhance the targeted delivery of plant exosomes, focusing on chronic inflammation—a key driver of heart failure—evaluating their therapeutic potential in myocardial infarction. By optimizing isolation methods and fusing exosomes with synthetic liposomes, we created collagen-targeted hybrid nanovesicles (GEP-NPs). In a mouse model of myocardial infarction, GEP-NPs efficiently accumulated in damaged heart tissue. Compared to non-engineered vesicles, GEP-NPs more effectively reduced fibrosis, suppressed ventricular remodeling, and modulated chronic post-infarction inflammation, thereby preventing progression from acute injury to chronic heart failure. Mechanistic studies showed that GEP-NPs may inhibit the overactive PI3K-AKT-mTOR pathway, which regulates inflammation, cell survival, and metabolism. The bioactive components from plant exosomes likely act effectively within the engineered vesicles to suppress this pathway, reducing persistent inflammation and promoting tissue repair. This work provides a scalable, reproducible method for engineering plant exosomes, improving both delivery precision and stability, and offering a promising strategy for treating chronic cardiac inflammation and preventing heart failure.

Graphical abstract

1. Introduction

Myocardial infarction (MI) is the most common manifestation of ischemic heart disease (IHD). According to data from the Global Burden of Disease (GBD) study, IHD continues to be the leading cause of death worldwide [1,2]. Despite significant advancements in current therapeutic strategies—including revascularization of occluded vessels and optimized pharmacotherapy for ventricular remodeling [3], patients with acute coronary syndrome (ACS)—including MI—demonstrate an increased risk of stroke during long-term follow-up [4]. Consequently, there is an urgent need to move beyond the traditional therapeutic approach centered exclusively on "rapid vascular recanalization" and to develop novel treatment strategies for MI that aim to achieve multiple objectives, such as maximal myocardial salvage, preservation of cardiac function, and improvement of long-term clinical outcomes.

In recent years, various novel bioengineering technologies have been applied in the treatment of cardiovascular disease (CVD). For instance, transplantation of healthy mitochondria into macrophages enhances their reparative capacity, and subsequent delivery to cardiomyocytes significantly reduces infarct size while improving cardiac function [5]. Therapeutic mRNA delivery via lipid nanoparticles (e.g., VEGF-C) promotes lymphangiogenesis and mitigates inflammatory responses, thereby improving MI prognosis [6]. Recently, plant-derived bioactive compounds have garnered significant attention in the field of cardiovascular disease treatment. Plant-derived bioactive compounds can modulate multiple nodes within disease networks through the synergistic actions of diverse active constituents. For example, flavonoids may concurrently exert antioxidant, anti-inflammatory, and endothelial protective effects, thereby attenuating the pathological progression of MI via multiple mechanistic pathways [7]; glycyrrhizic acid hydrogels and chloroplast constituents effectively reduce infarct area and facilitate cardiac functional recovery through synergistic antioxidant effects and energy supplementation mechanisms [8]. Furthermore, plant-derived components typically exhibit excellent biocompatibility and low toxicity. For instance, Chebulinic acid, as a potential natural alternative to statins, was initially investigated to explore novel therapeutic agents with reduced side effects [9]. Meanwhile, plant-derived materials demonstrate considerable feasibility in terms of large-scale cultivation and extraction. Among these, plant exosomes—are emerging as a highly promising novel approach for cardiovascular disease CVD treatment. These plant-derived exosomes possess distinct advantages, including a natural origin, excellent biosafety profile, multifaceted therapeutic potential, and high adaptability as drug delivery carriers [10]. In comparison with exosomes derived from animal cells, they are associated with reduced production costs and show enhanced potential for the development of standardized and scalable manufacturing processes, thereby more effectively addressing the substantial clinical demand for therapeutic materials [11]. For example, Bitter melon-derived exosome-like nanovesicles (MC-ELNs) can ameliorate doxorubicin-induced cardiomyocyte cytotoxicity and enhance cell survival by stabilizing p62 protein, promoting nuclear translocation of Nrf2, and augmenting cellular antioxidant responses [12]. Meanwhile, Lycium barbarum-derived exosomes effectively enhance cardiac function by suppressing cardiomyocyte apoptosis and mitigating myocardial fibrosis progression [13]. Building upon the remarkable therapeutic efficacy demonstrated by the aforementioned bioengineering interventions, plant-derived biomaterials—particularly those with proven biological activity and biocompatibility—have emerged as promising candidates for further development and clinical application.

However, despite significant advancements in research on plant-derived exosomes, their practical application continues to face challenges that hinder widespread adoption and utilization in cardiovascular disease treatment. For example, plant-derived exosomes inherently lack the capacity for specific targeting of cardiac and vascular lesion sites. Following intravenous administration, they are rapidly cleared by the mononuclear phagocyte system in organs such as the liver and spleen, leading to markedly reduced delivery of therapeutic agents to the MI site [11,14]. Plant-derived exosomes encounter multiple challenges within the systemic circulation, including enzymatic degradation, pH fluctuations, and immune system attacks. Although their structure can maintain a certain degree of stability under specific conditions (e.g., 4 °C), their biological activity may significantly decrease during long term storage or oral delivery processes [15]. Therefore, enhancing the targeting capability of plant-derived exosomes while preserving their biological activity is crucial for advancing their clinical applications.

Recent researches have demonstrated that engineering modifications of plant-derived exosomes can significantly enhance their targeting capability, therapeutic efficacy, and applicability across a wide range of pathological conditions. Accordingly, herein, we present a streamlined and efficient methodology for the extraction and engineering of red cabbage exosomes (Rabex). Red cabbage (Brassica oleracea var. capitata f. rubra) is a nutrient-rich cruciferous vegetable high in anthocyanins, glucosinolates, vitamin C, and polyphenolic compounds [16]. These bioactive components provide strong antioxidant and anti-inflammatory effects, supporting its potential in preventing cancer [17,18], protecting cardiovascular health [[19], [20], [21]], and promoting intestinal health [22]. Research has demonstrated that chronic inflammation plays a pivotal role as a key driver in the process of cardiac remodeling following MI. Persistently activated immune cells not only exacerbate myocardial tissue damage but also promote fibroblast activation and excessive collagen deposition, thereby inducing myocardial fibrosis and ultimately leading to the onset of heart failure [23,24]. Recent studies have found that Rabex have significant anti-inflammatory and cell-protective effects: they can inhibit the activation of immune cells, alleviate inflammatory responses, and reduce the apoptosis of human keratinocytes and fibroblasts [25]. In inflammatory bowel disease, Rabex has also been confirmed to inhibit macrophage-mediated inflammatory responses, while promoting epithelial tissue regeneration and delaying disease progression [26]. However, although Rabex has demonstrated potential therapeutic value in various inflammation-related pathological conditions, there are no reports on whether it can intervene in the chronic inflammation and subsequent fibrosis process after MI.

Here, we developed an easy-to-operate and scalable workflow to produce abundant engineered plant-derived exosomes. Given that aberrant collagen deposition in cardiac tissue following MI constitutes a core pathological feature driving fibrosis and cardiac functional deterioration [27], this study functionalized GEP-NPs by conjugating the collagen-targeting peptide GKWHCTTKFPHHYCLY [28] onto their surface. We aimed to explore whether GEP-NPs could selectively target and regulate inflammation and fibrosis after MI, providing a new intervention for heart failure. Multi-omics analysis showed that GEP-NPs well preserved the bioactivity of endogenous biomolecules in Rabex and displayed improved collagen specificity. In a mouse MI model, GEP-NPs exerted strong cardioprotection, likely by inhibiting the pathological overactivation of the PI3K-AKT-mTOR pathway (Fig. 1). This simple and rationally designed Rabex engineering strategy combines the low toxicity and high biocompatibility of natural carriers with the targeting, efficiency and multifunctionality of synthetic nanomaterials. It offers a promising therapeutic strategy for MI and a universal platform for studying various exosomes and plant materials, greatly expanding their applications in regenerative medicine.

Fig. 1.

Schematic diagram of the research.

2. Materials and methods

2.1. Isolation of exosomes from red cabbage

In this experiment, commercially available red cabbage (Brassica oleracea var. capitata f. rubra) sourced from a domestic cooperative farm was utilized as the raw material. The samples were thoroughly washed three times with distilled water to effectively remove surface contaminants, including dust, soil, and pesticide residues. Subsequently, a low-temperature cold-pressing technique was employed to obtain the crude extract. To further eliminate cellular debris and other large particulate impurities from the juice, a multi-step gradient centrifugation process was conducted: initial centrifugation at 1000×g for 10 min at 4 °C, followed by sequential centrifugation at 8000×g for 1 h, twice at 14000×g (each for 1 h), and finally at 20000×g for 1 h. All supernatants were collected, aliquoted, and stored at −80 °C in an ultra-low temperature freezer for subsequent experimental use. During the exosome extraction process, 200 mL of the pre-treated supernatant was subjected to ultracentrifugation using a Beckman Optima XE-100 ultracentrifuge at 120000×g for 2 h. After discarding the supernatant, the pellet containing nanoscale vesicles was resuspended in phosphate-buffered saline (PBS) and adjusted to a final volume of 4 mL, yielding a particle concentration of 6.6 × 10¹⁰ particles/mL. To further enhance vesicle purity, size-exclusion chromatography (SEC, Unique Autopre25D-M402, Suzhou Insight Intelligent Technology Co., Ltd.) was employed for fine separation, with PBS as the mobile phase. Elution fractions were collected at 200 μL per tube, and each fraction was analyzed for vesicle count and protein content. Finally, fractions 7 to 13 were pooled, and the exosome solution was concentrated tenfold using ultrafiltration for subsequent analysis.

2.2. Nanoparticle tracking analysis (NTA) of exosomes

To characterize the isolated exosome-like nanovesicles, this study employed nanoparticle tracking analysis (NTA) to determine their size distribution and particle concentration using the ZetaVIEW nanoparticle tracking analysis system (Particle Metrix, serial number: 22-756). Protein quantification was performed using the Enhanced BCA Protein Assay Kit (Beyotime, product number P0010). During the establishment of the protein standard curve, a 0.5 mg/mL protein standard was serially diluted to create concentration gradients of 0, 0.025, 0.05, 0.1, 0.2, 0.3, 0.4, and 0.5 mg/mL. Each well of a 96-well plate was loaded with 20 μL of the standard solution, followed by the addition of 200 μL of BCA working solution. After incubation at 37 °C for 30 min, the absorbance was measured at 562 nm using a BioTek Synergy H1 multifunctional microplate reader. The components of the exosome-like nanovesicles were analyzed under identical conditions, and their protein concentrations were calculated based on the standard curve. Finally, the purity of the exosome-like nanovesicles was assessed by the ratio of particle concentration to protein concentration.

2.3. Transmission electron microscopy (TEM) analysis of exosomes

During the sample preparation process for transmission electron microscopy, the nano-capsule colloidal suspension is initially loaded onto the surface of a specialized copper grid and allowed to deposit at room temperature for 10 min. Subsequently, excess liquid is removed using filter paper, and the sample is left to air-dry under ambient conditions. Following this, the sample undergoes negative staining with a phosphotungstic acid solution provided by SPI-Chem, with a staining duration of 3 min. Upon completion of staining, residual staining solution is again absorbed using filter paper, and the sample is left to air-dry at room temperature. Finally, the prepared sample is subjected to morphological observation using a JEM-1400 transmission electron microscope from JEOL Ltd. to analyze the structural characteristics of the nano-vesicles.

2.4. Liposomes were synthesized by the thin-film hydration extrusion method

The preparation of liposomes was conducted according to the following protocol: Precisely weigh 10 mg of soybean lecithin (CS-95, Shanghai Avit Pharmaceutical Technology Co., Ltd.), 1 mg of targeting peptide-lipid conjugate GKWHCTTKFPHHYCLY-PEG550-DSPE (Xi'an Ruixi Biological Technology Co., Ltd.), and 2.5 mg of cholesterol (Shanghai Macklin Biochemical Co., Ltd.), and dissolve them together in 4 mL of chloroform. The solution was subjected to rotary evaporation under reduced pressure at 45 °C and 100 rpm to remove the organic solvent until a uniform lipid film formed on the walls of the flask. As a control, non-targeted liposomes were prepared using the same procedure, consisting of 10 mg of soybean lecithin, 1 mg of PEG550-DSPE (Xi'an Ruixi Biological Technology Co., Ltd.), and 2.5 mg of cholesterol. After complete solvent evaporation, 5 mL of phosphate-buffered saline (PBS) was added to the flask containing the lipid film, followed by vortexing to ensure thorough hydration and dispersion of the lipids. To obtain uniformly sized small unilamellar vesicles, the crude dispersion was extruded 21 times through a 100 nm polycarbonate membrane. The resulting liposome suspension was purged with argon to prevent oxidation, with a particle concentration of 6.4 × 10¹⁰ particles/mL, and aliquoted for storage at 4 °C for subsequent studies.

2.5. Evaluation of fusion efficiency

This study employed Fluorescence Resonance Energy Transfer (FRET) technology to conduct quantitative analysis of membrane fusion efficiency. The experiment utilized a FRET pair composed of fluorescent probes NBD-DOPE and RhB-DOPE labeled in the lipid bilayer, monitoring fluorescence signal changes to reflect the membrane fusion process. Fluorescence detection was performed using the Edinburgh FLS 980 fluorescence spectrometer, with characteristic emission peaks of NBD-DOPE and RhB-DOPE collected at 522 nm and 578 nm respectively under an excitation wavelength of 462 nm. Membrane fusion induced dilution of the lipid bilayer, manifested as an increase in fluorescence intensity at 522 nm and a decrease at 578 nm. The fusion efficiency was quantified through FRET Dissolution Efficiency (EFD), calculated as follows: EFD = F522/(F522 + F578) where F522 and F578 represent fluorescence intensity values at 522 nm and 578 nm respectively. To establish quantitative standards, a series of liposomes containing varying molar ratios of NBD-DOPE to RhB-DOPE were prepared, with probe molar percentages of 0.650%, 0.500%, 0.250%, 0.185%, 0.125%, and 0.063%, corresponding to lipid dilution ratios of 0.65, 1, 2, 3, 4, and 5. The EFD values of each liposome sample were measured, and a standard curve was plotted with lipid dilution ratio as the abscissa and EFD as the ordinate. Based on this curve, the measured EFD values of the labeled liposome mixture were converted to corresponding lipid dilution ratios, thereby evaluating membrane fusion efficiency after freeze-thaw cycle treatment.

2.6. Basic characterization of GEP-NPs

To evaluate the stability of GEP-NPs, the samples were stored in a light-protected environment at 4 °C and systematically analyzed at time points of 0 h, 3 h, 8 h, 24 h, 2 days, 3 days, 7 days, and 14 days to monitor changes in particle size, polydispersity index (PDI), and drug release rate. Zeta potential was measured using a dynamic light scattering system (PSS, 3000DLS), while particle size and PDI were characterized using a nano laser particle size analyzer (Malvern, Nano-ZS90). The drug release rate was assessed via ultraviolet spectrophotometry, with absorbance data collected using a microplate reader (Tecan, A-5082).

2.7. Rabex/GEP-NPs labelling and live imaging

To investigate the distribution characteristics of Rabex and GEP-NPs in the hearts of MI model mice, this study employed the cell membrane dye DiR (Aladdin, Cat. No. D131031) for fluorescent labeling of these two nanoparticles. Twenty-four hours following MI modeling, Rabex or GEP-NPs were administered to the mice via tail vein injection. After an additional 24-h post-dosing period, the animals were euthanized and subjected to systematic dissection. Major organs including the heart, liver, spleen, lungs, and kidneys were harvested for ex vivo fluorescence imaging. Quantitative fluorescence detection and analysis of all tissue samples were performed using the PerkinElmer IVIS Spectrum in vivo imaging system.

2.8. Animal experiments

All animal experimental procedures in this study were approved by the Experimental Animal Ethics Committee of Enshi Central Hospital (Approval Number: W202503001) and strictly adhered to the guidelines outlined in the National Institutes of Health's "Guide for the Care and Use of Laboratory Animals." Male C57BL/6J mice (8 weeks old) were purchased from Jiangsu Gemma Pharmaceutical Technology Co., Ltd. and housed in a controlled environment with a constant temperature of 25 °C, humidity of 40%–60%, and a 12-h light-dark cycle, with free access to standard feed and water. After one week of acclimatization, the mice were anesthetized with sodium pentobarbital, and tracheal intubation was performed to connect them to a ventilator for maintaining ventilation. A thoracotomy was performed at the fourth intercostal space to expose the heart, and the proximal left anterior descending coronary artery was ligated using 7-0 suture. Successful MI modeling was confirmed by the appearance of significant pallor in the area distal to the ligation towards the apex. The thoracic cavity was then closed layer by layer, and the mice were placed on a warming pad for recovery. After confirming the absence of complications such as pneumothorax, the mice were allowed to regain consciousness.

2.9. Assessment of cardiac function in mice

On the 28th day after MI modeling, the cardiac function of mice was evaluated by ultrasound. During the experiment, the mice were anesthetized with isoflurane inhalation and placed in a prone position on a 37 °C constant temperature operating table. The short-axis M-mode dynamic images of the left ventricle were collected using a small animal ultrasound imaging system (Vevo 3100, equipped with a 40 MHz linear probe). Based on the image data, the left ventricular inner diameters during systole and diastole were measured respectively, and the left ventricular ejection fraction, fractional shortening, and end-diastolic and end-systolic diameters were calculated accordingly. Data from three consecutive complete cardiac cycles of each mouse were recorded for analysis. All ultrasound scans and subsequent image analyses were performed in a single-blind manner.

2.10. Histology assay

On the 28th day after MI, the mice were euthanized, and their hearts were taken and immersed in a 4% polyformaldehyde solution (pH 7.4) for fixation. After dehydration and embedding of the cardiac tissue, the tissue sections were cut into 5-μm thick slices. Then, the tissue sections were stained with hematoxylin-eosin (Cat: G1005-1/2; Servicebio; China) and Masson's staining agent (Cat: G1006; Servicebio; China). The sections were scanned in bright-field mode at a magnification of 20x using a high-resolution microscope (Leica; Japan). Heart sections were stained with Masson's trichrome. The percentage of collagen area was quantified using ImageJ (v1.53t) software. A fixed color threshold was applied to all images to select blue-stained collagen fibers, and the area fraction was calculated automatically.

2.11. Immunohistochemistry

For conducting immunofluorescence and histochemical staining on cardiac tissue, the sections are first fixed with 4% paraformaldehyde, and then blocked with goat serum for 1-2 h. After the blocking process is completed, the primary antibody is added to the slide and incubated at 4 °C overnight. After removing or recovering the primary antibody on the second day, the secondary antibody is incubated at room temperature for 2 h (fluorescent secondary antibody needs darkness for incubation). Finally, DAPI is used for staining and the slides are sealed. The required images are obtained using a confocal microscope (Olympus; Japan).

2.12. Tests of liver and kidney function

After the mice were euthanized, whole blood was collected through the abdominal aorta, and after centrifugation at 3000g for 20 min at 4 °C, serum was mounted. Subsequently, the contents of alanine aminotransferase (ALT), aspartate aminotransferase (AST), urea nitrogen (BUN) and creatinine (Cr) in the mouse serum were determined using commercial kits (Cat: C009-2-1, C010-2-1, C013-2-1, C011-2-1; Nanjing Jiancheng Bioengineering Institute; China).

2.13. Quantitative proteomics analysis

Proteins were extracted from cardiac tissue samples using SDT lysis buffer (4% SDS, 100 mM DTT, 100 mM Tris-HCl, pH 8.0). Following a 3-min boiling step, ultrasonic disruption was carried out to further lyse the tissue samples. Centrifugation was then performed at 16,000×g for 15 min to remove insoluble cellular debris. The resulting supernatant was collected and quantified using the BCA protein assay kit (BeyoTime, China). Subsequently, protein digestion was performed using the FASP method as described by Wisniewski, Zougman et al. [PMID: 19377485]. Briefly, the reducing cysteine residues were alkylated by adding detergent, DTT, and IAA in UA buffer. The protein suspension was then subjected to overnight digestion with trypsin (Promega) at an enzyme-to-substrate ratio of 50:1 under 37 °C incubation. After reaction termination, the peptide mixture was collected by centrifugation at 16,000×g for 15 min, followed by desalting using C18 StageTip columns for subsequent LC-MS analysis. The reconstituted peptides were quantified at OD280 wavelength using a Nanodrop One instrument (Thermo Scientific, USA). The LC-MS/MS analysis was performed on an Orbitrap Astral mass spectrometer coupled with a Vanquish Neo UHPLC system (Thermo Fisher Scientific). Peptides from each sample source were loaded onto the column (50 cm low-load μPAC™ Neo HPLC column, Thermo Scientific) at a flow rate of 2.2 μL/min. The mobile phases for reversed-phase high-performance liquid chromatography (RP-HPLC) consisted of 0.1% formic acid in water (phase A) and 0.1% formic acid in 80% acetonitrile (phase B). Elution was achieved using a linear gradient program with phase B at a flow rate of 1.25 μL/min, yielding a total elution time of 8 min. The specific gradient conditions were as follows: 0–0.1 min, 4%–6% B; 0.1–1.1 min, 6%–12% B; 1.1–4.3 min, 12%–25% B; 4.3–6.1 min, 25%–45% B; 6.1–6.5 min, 45%–99% B; 6.5–8 min, 99% B held constant. Eluted peptides were detected by the Orbitrap Astral mass spectrometer. Full MS scans were acquired in the range of 380–980 m/z at a resolution of 240,000, with an AGC target of 500% and an injection time of 5 ms. MS/MS scans were performed in data-independent acquisition (DIA) mode, controlled by Astral software, covering 150–2000 m/z with an isolation window of 2 m/z, an AGC target of 500%, and an injection time of 3 ms. The normalized collision energy was set at 25 eV with a cycle time of 0.6 s. Full MS and DIA scan data were recorded in profile mode and centroid mode, respectively. The DIA-MS data were analyzed using DIA-NN 1.8.1 software [2,3], with trypsin selected as the proteolytic enzyme. During database searching, the maximum allowed missed cleavage sites for precursor and fragment ions were set to 1, with a mass tolerance of 10 ppm. Carbamidomethylation of cysteine was specified as a fixed modification, while protein N-terminal acetylation and methionine oxidation were included as variable modifications, with a maximum of 1 variable modification allowed per peptide. The peptide length was restricted to 7-30 amino acids, and peptide charge states were set from 1+ to 4+. The fragment ion m/z scan range was defined as 150-2000. Database search results were filtered at both the peptide spectrum match (PSM) and protein levels using a false discovery rate (FDR) threshold of <1%, and subsequently exported for downstream analysis.

2.14. Metabolomics analysis

The samples were gradually thawed at 4 °C, and an appropriate volume of sample (0.5–1.0 mL; for excessive volumes, lyophilization was performed for concentration) was accurately transferred to a centrifuge tube. An equal volume of extraction solvent (methanol/acetonitrile, 1:1, v/v) was added, followed by vortex mixing for 60 s and low-temperature ultrasonic extraction for 30 min, with the extraction repeated twice. The mixture was then incubated at −20 °C for 1 h to precipitate proteins, followed by centrifugation at 4 °C and 12,000 rpm for 20 min. The supernatant was collected and subjected to lyophilization. The resulting residue was reconstituted in 200 μL of 30% acetonitrile solution, vortex-mixed, and centrifuged at 4 °C and 14,000 rpm for 15 min. The final supernatant was used for instrumental analysis. The data acquisition system primarily comprises an ultra-high-performance liquid chromatography instrument (Vanquish UPLC, Thermo Fisher Scientific, USA) and a high-resolution mass spectrometer (Q Exactive HFX, Thermo Fisher Scientific, USA). The raw data was initially preprocessed using Progenesis QI software (Waters Corporation, Milford, USA), which involved baseline filtering, peak identification, peak matching, retention time alignment, and peak alignment, resulting in a data matrix containing retention times, mass-to-charge ratios, and peak intensities. Ultimately, the chromatographic peaks containing secondary mass spectrometry data were identified by integrating commercial databases, the in-house plant metabolite standard database developed by Sanshu Biotech, and the corresponding secondary mass spectral fragmentation patterns.

2.15. Antibodies

Antibodies against to TSG101 (Cat: 28283-1-AP; Proteintech; 1:1000), CD63 (Cat: 25682-1-AP; Proteintech; 1:1000), HSP70 (Cat: 80735-1-RR; Proteintech; 1:1000), Vimentin (Cat: A19607; ABclonal; 1:500), CD68 (Cat: A24386PM; ABclonal; 1:200), cTNT (Cat: A4914; ABclonal; 1:200), Col1 (Cat: 67288-1-Ig; Proteintech; 1:500), IL-6 (Cat: ab233706; Abcam; 1:50), IL-1β (Cat: ab283818; Abcam; 1:50), p-mTOR (Cat: 5536T; Cell Signaling Technology; 1:1000), p-AKT (Cat: 4060T; Cell Signaling Technology; 1:1000), t-AKT (Cat: 10176-2-AP; Proteintech; 1:1000), t-mTOR (Cat: 66888-1-Ig; Proteintech; 1:1000), PTEN (Cat: 60300-3-Ig; Proteintech; 1:1000), GAPDH (Cat: sc-137179; SANTA; 1:1000) as well as Alexa Fluo 594 or 488 conjugated Goat anti Rabbit or mouse secondary antibodies (1:200) were purchased from Cell Signaling Technology. WGA staining kit was purchased from Sigma-Aldrich (Cat: L4895).

2.16. Data acquisition and statistical analysis

The experimental animals were randomly divided into four groups. Relevant data were collected and processed using Excel 2021, Image J 1.8.0, Living Image 4.4 (PerkinElmer), FV31S-SW Observer 2.4 version (Olympus Company), and Vevo 3100 LT-40MHz. Statistical analysis was performed using GraphPad Prism 10 (version 10.1.2). Data were presented as mean ± standard deviation. Comparisons among multiple groups with the same variance were conducted using one-way analysis of variance, and comparisons among multiple groups with unequal variances were conducted using Brown-Forsythe and Welch variance analysis. A P value < 0.05 was considered to be statistically significant.

3. Results

3.1. Fabrication and characterization of GEP-NPs

To realize efficient isolation of Rabex, a combined method of ultracentrifugation (UC) and size exclusion chromatography (SEC) — denoted as UC-SEC — was adopted in this work. As illustrated in Fig. 2A, juicing was conducted with a cold-press juicer, which compresses plant tissues to achieve solid-liquid separation. This process minimizes mechanical stress and heat generation, thereby preserving the integrity of Rabex. After separation and extraction, the juice was subjected to low-speed centrifugation to remove cell debris, followed by UC for enriching the Rabex dispersion. Subsequently, SEC was applied to purify the UC-enriched samples, as SEC separates Rabex based on size. Smaller molecular impurities, including proteins, nucleic acids, and small-molecular fragments, diffuse into the porous column matrix, which extends their elution time. In contrast, larger Rabex elute more rapidly via the shortest flow path. Following the successful isolation of high-purity nanovesicles, the nanovesicles were hybridized with GKWHCTTKFPHHYCLY peptide-conjugated liposomes through freeze-thaw cycles, resulting in the formation of hybrid vesicles referred to as GEP-NPs.

Fig. 2.

Fabrication and characterization of GEP-NPs. (A) Schematic diagram of the synthesis process of GEP-NPs. (B) Quantification of nanovesicle and protein concentrations across all fractions obtained from SEC. (C) Representative profile illustrating the size distribution of nanovesicles isolated through UC-SEC. Inset: TEM image of purified Rabex. (D) Comparison of nanovesicle purity obtained using ultracentrifugation alone versus UC-SEC. ****P < 0.0001 (n = 3). (E) Changes in the particle size of nanovesicles before and after fusion with liposomes, as measured after five freeze-thaw cycles. ***P < 0.001 (n = 3). (F) Representative profile showing the size distribution of nanovesicles isolated by UC-SEC. Inset: TEM image of isolated GEP-NPs. (G) Fluorescence spectra of labeled liposomes before and after fusion with nanovesicles, recorded using an excitation wavelength of 462 nm. (H) The Zeta potentials of liposomes, Rabex, and GEP-NPs (n = 3). (I) Variation in particle size of GEP-NPs when stored under light-protected conditions in PBS at 4 °C and pH 7.0. (J) Drug Release Kinetics of GEP-NPs in Phosphate Buffer at 4 °C and pH 7.0 under light-shielding conditions. (K) Western blot analysis of conventional exosomal protein markers. (L-M) Detection and quantitative analysis of exosomal protein markers following multiple freeze-thaw cycles (n = 3).

For evaluating the purity of Rabex, nanoparticle tracking analysis (NTA) was utilized to quantify the Rabex concentration in each SEC fraction, while the bicinchoninic acid (BCA) assay was employed to determine the protein concentration. Rabex purity was then calculated based on the particle-to-protein ratio. As shown in Fig. 2B, Rabex-rich fractions were primarily collected between fractions 7 and 13, whereas protein contaminants were mainly detected in fractions 19–37. This result confirms the effective separation of Rabex from protein impurities. Therefore, fractions 7 to 13 were pooled for subsequent experiments, as they exhibited high Rabex concentration and minimal protein contamination. The average diameter of the collected Rabex was 147 nm, which falls within the typical size range of Rabex (Fig. 2C).

Fig. 2D presents a comparative analysis of nanovesicle purity obtained via two isolation methods: ultracentrifugation alone (UC-only) and the UC-SEC method. The results indicate that nanovesicles isolated by UC-SEC had significantly higher purity, reaching 8.62 × 10⁸ particles per microgram of protein. By contrast, the purity of nanovesicles from the UC-only method was 2.77 × 10⁷ particles per microgram of protein — meaning the purity of the former was 31 times that of the latter. This significant improvement demonstrates that size exclusion chromatography offers superior efficiency in removing protein contaminants.

The high-purity nanovesicles obtained via UC-SEC possess several advantages, including reduced immunogenicity and toxicity, making them highly suitable for drug delivery applications. Given its excellent performance, the UC-SEC method can serve as a standardized approach for isolating plant-derived nanovesicles. It is not limited to red cabbage but can also be extended to various plant sources, thereby expanding its potential applications in drug delivery, nutraceuticals, and regenerative medicine.

Following the successful isolation of Rabex, GEP-NPs—hybrid nanovesicles formed via freeze-thaw-mediated fusion of nanovesicles with GKWHCTTKFPHHYCLY-modified liposomes—were prepared. The vesicle size increased from 148.2 nm (before fusion) to 225.6 nm (after fusion) (Fig. 2E), which confirms the successful incorporation of liposomes into the nanovesicles. Fluorescence resonance energy transfer (FRET) analysis was used to evaluate the fusion efficiency between nanovesicles and liposomes. FRET is a highly sensitive technique that detects molecular proximity at the nanoscale by measuring energy transfer between a donor fluorophore and an acceptor fluorophore. When these fluorophores are in close proximity, the excited donor fluorophore triggers non-radiative energy transfer to the acceptor fluorophore, leading to fluorescence emission at a longer wavelength. However, as the fluorophores dilute and their mutual distance increases, the efficiency of energy transfer decreases. This distance dependence of FRET makes it a powerful tool for assessing membrane fusion events. In this study, FRET-labeled lipids were integrated into the liposome membrane to quantitatively evaluate fusion efficiency, followed by measuring the fluorescence intensity before and after fusion with nanovesicles. Liposomes were prepared using NBD-DOPE (N-(7-nitro-2,1,3-benzoxadiazol-4-yl)dioleoyl-phosphatidylethanolamine) as the donor fluorophore carrier and RhB-DOPE (1,2-dioleoyl-sn-glycero-3-phosphoethanolamine-N-(lissamine rhodamine B sulfonyl)) as the acceptor fluorophore carrier. The labeled liposomes were then mixed with nanovesicles, and after 5 freeze-thaw cycles, an increase in fluorescence at 522 nm and a corresponding decrease at 578 nm were observed — a phenomenon indicating FRET disruption caused by lipid dilution (Fig. 2G). These results confirm that the fluorescent lipids in the liposome membrane mixed with the membrane components of the nanovesicles, verifying the success of hybridization. Lipid dilution rate, a key indicator of fusion efficiency, was determined using calibrated fluorescence spectroscopy, yielding a value of 2.8 ± 0.2 after 5 freeze-thaw cycles. Transmission electron microscopy (TEM) was used to observe morphological changes of Rabex (before fusion) and nanovesicles (after fusion). Although TEM analysis did not provide statistical insights into the average morphology, no significant structural differences were observed between Rabex before fusion with liposomes (inset of Fig. 2C) and after fusion (inset of Fig. 2F), which further supports the structural integrity of the hybrid vesicles. The zeta potentials of the hybrid vesicles (GEP-NPs), Rabex, and liposomes were measured. The zeta potential absolute values of these three substances consistently remained above 20 mV (Fig. 2H), indicating that GEP-NPs, similar to Rabex and liposomes, exhibit certain electrostatic repulsion effects among them, which is conducive to maintaining particle stability.

The stability of GEP-NPs was evaluated by monitoring their particle size (Size) and polydispersity index (PDI) at different time points using dynamic light scattering (DLS)., GEP-NPs showed excellent stability within 0–3 d, with particle size around 200 nm and PDI <0.4. At 7–14 d, particle size slightly increased but PDI remained <0.6, indicating reduced but still acceptable long-term colloidal stability (Fig. 2I). To evaluate the drug release stability of GEP-NPs, this study used the main natural anthocyanin polyphenols rich in red cabbage as the model active component to investigate its in vitro release behavior. The results indicated that GEP-NPs had excellent sustained-release controllability: the cumulative release rate was less than 25% in the first 3 days, and exceeded 50% by the 14th day, confirming that this nano-delivery system had superior carrier structural stability and long-term sustained-release performance (Fig. 2J).

The classical protein markers, including HSP70, TSG101, and CD63, in liposomes, Rabex, and GEP-NPs were detected via Western blot (WB) analysis (Fig. 2K). Given that Rabex contain bioactive macromolecules such as proteins and are highly sensitive to repeated freeze-thaw cycles, we subjected GEP-NPs to different numbers of freeze-thaw cycles and assessed the protein content across experimental groups after treatment via WB analysis. The results demonstrated that GEP-NPs were capable of retaining the activity of most internal biological macromolecules after three cycles of repeated freeze-thaw treatment. However, this capacity exhibited a significant decline with further increases in the number of freeze-thaw cycles. (Fig. 2L–M).

3.2. Proteomic and metabolomic analysis of GEP-NPs and Rabex

To identify the specific bioactive molecules present in Rabex and to evaluate potential changes in these molecules following its engineered transformation into GEP-NPs, this study utilized proteomics and comprehensive plant metabolomics for systematic analysis. Fig. 3A illustrated the protein density distribution profiles of GEP-NPs and Rabex obtained through rigorous quality control of proteomics data. “Quantifiable” proteins required detection in ≥2 technical/biological replicates (n = 3) with a minimum MS1 intensity of 1 × 10⁴. Fig. 3B showed the quantitative distribution histogram of proteins in GEP-NPs and Rabex. To thoroughly elucidate the functional characteristics of diverse proteins in GEP-NPs and Rabex, a comprehensive functional annotation was conducted on the identified proteins, encompassing multiple dimensions including Gene Ontology (GO), protein domains, KEGG pathways, COG/KOG functional classifications (Fig. 3C and D). Subsequently, the identified proteins of GEP-NPs and Rabex were classified according to their subcellular localization, and the results were presented in the form of a pie chart (Fig. 3E and F). Fig. 3G and H presented the KEGG pathway classification analysis results of the identified proteins in GEP-NPs and Rabex. The horizontal axis represents the number of identified proteins in each category, and the vertical axis shows the secondary functional categories under the first-level KEGG classification (including metabolism, genetic information processing, environmental information processing, cellular processes, organismal systems, human diseases, and drug development). Fig. 3I presented a heatmap of the top 30 proteins commonly identified in both GEP-NPs and Rabex, ranked by their expression levels. Furthermore, KEGG pathway enrichment analysis was conducted on the proteins identified in GEP-NPs and Rabex, with the results presented in a bar chart format (Fig. 3J and K). Finally, metabolomic analysis was conducted on the metabolites in GEP-NPs and Rabex, with the distribution of various metabolites presented in the form of pie charts (Fig. 3L and M). The proteomic and metabolomic analyses indicate that following the engineering modification of Rabex to form GEP-NPs, while certain proteins and active metabolites exhibited reduction, the majority of bioactive macromolecules were effectively preserved.

Fig. 3.

Proteomic and metabolomic analysis of GEP-NPs and Rabex. (A) The density distribution of GEP-NPs and Rabex proteins (The Density value on the Y-axis represents the relative frequency of the number of proteins/peptides detected within different Log10 Intensity (logarithm of intensity) intervals; n = 3). (B) Quantifiable proteins of GEP-NPs and Rabex. (C) Comprehensive functional annotation of GEP-NPs proteins. (D) Comprehensive functional annotation of Rabex proteins. (E) Subcellular localization classification of GEP-NPs proteins. (F) Subcellular localization classification of Rabex proteins. (G) KEGG pathway classification analysis of GEP-NPs Proteins. (H) KEGG pathway classification analysis of Rabex Proteins. (I) Heatmap analysis of the top 30 shared proteins between GEP-NPs and Rabex. (J-K) KEGG pathway enrichment analysis of GEP-NPs and Rabex Proteins; enrichment analysis was performed using Fisher's exact test. Only pathways with a p-value <0.05 and FDR (False Discovery Rate) < 0.05 (corrected by the Benjamini-Hochberg procedure) were considered significantly enriched and retained in the figure. The x-axis represents the -log10 transformed p-value, indicating the statistical significance of the enrichment. (L) Quantitative analysis of metabolites in GEP-NPs. (M) Quantitative analysis of metabolites in Rabex.

3.3. In vitro phagocytic capacity, functional efficacy, and in vivo distribution and toxicity of GEP-NPs

Cardiomyocytes, fibroblasts, and macrophages are crucial cellular components in the cardiac repair process following MI. These cells exert essential functions during this process through their respective biological activities. Therefore, in the in vitro phagocytosis experiment, we selected three representative cell lines to verify that GEP-NPs could be phagocytosed. Concurrently, immunohistochemical analysis of cardiac tissue sections confirmed the targeted binding capability of GEP-NPs to collagen I in myocardial fibrosis. The flow diagram of the experiment was shown in Fig. 4A. After co-culturing Dil-labeled GEP-NPs with different cell lines for 24 h, we found that the myocardial cell line (H9c2), fibroblast cell line (NIH/3T3) and macrophage cell line (RAW 264.3) could phagocytize GEP-NPs (Fig. 4B–D). Quantitative analysis of GEP-NPs uptake across three cell types revealed a marked cell-type–dependent hierarchy: RAW 264.7 macrophages exhibited near-complete uptake (∼100%), NIH/3T3 fibroblasts showed moderate uptake, and H9c2 cardiomyocytes displayed the lowest uptake efficiency (<50%) (Fig. S1). Subsequently, the antioxidant, anti-fibrotic, and anti-inflammatory activities of GEP-NPs were evaluated in three distinct models: the H₂O₂-induced oxidative stress injury model in H9c2 cells (Fig. 4E), the TGF-β-induced cellular fibrosis model (Fig. 4F), and the LPS-stimulated macrophage inflammation model (Fig. 4G). The results demonstrated that GEP-NPs exhibited significant antioxidant, anti-fibrotic, and anti-inflammatory effects, consistent with the anticipated outcomes. To verify the ability of GEP-NPs to target collagen I in mice with MI, fluorescence imaging was performed in important organs such as heart, liver, spleen, lung and kidney of mice in different groups: MI group (MI), MI + Rabex group (MI + Rabex) and MI + GEP-NPs group (MI + GEP-NPs) (Fig. 4H). The findings indicate that DIR-labeled GEP-NPs, administered via tail vein injection, while distributed across other major organs, effectively target localized cardiac tissue following MI. Subsequently, immunofluorescence staining of collagen I and DiR-labeled Rabex/GEP-NPs was performed on cardiac tissue sections, and the characteristic images were shown in Fig. 4I–J. In the cardiac tissue sections of the MI + GEP-NPs group, the colocalization signals (yellow) between GEP-NPs (red) and Col1 molecules (green) were significantly increased compared to the MI + Rabex group. These findings further substantiate that GEP-NPs can specifically target the collagen deposition regions following MI. Biocompatibility is a fundamental and critical requirement for biomaterials, as it not only directly affects life and health but also ensures that materials can function effectively while maintaining harmonious integration within biological systems. Therefore, we evaluated the biocompatibility of GEP-NPs in mice following MI. We measured the serum levels of alanine aminotransferase (ALT), aspartate aminotransferase (AST), blood urea nitrogen (BUN), and creatinine (Cr) in each group of mice, and no significant changes were observed (Fig. 4K–N). Fig. 4O presented representative HE staining images of liver and kidney from mice in different experimental groups. Our findings demonstrate that treatment with Rabex and GEP-NPs for MI did not cause significant histological changes or abnormal inflammatory cell infiltration in the vital organs in mice. The representative Masson staining in Fig. 4P also demonstrated the absence of abnormal fibrosis in the liver and kidney. The aforementioned findings indicate that GEP-NPs exhibit excellent phagocytic properties in vitro, along with potent antioxidant, antifibrotic, and anti-inflammatory capabilities. In vivo, these nanoparticles demonstrate superior collagen-targeting specificity while maintaining favorable biosafety and biocompatibility profiles.

Fig. 4.

In vitro phagocytic capacity and functional efficacy, and in vivo distribution and toxicity of GEP-NPs. (A) Schematic diagram of experimental design. (B) Representative images of the phagocytosis of GEP-NPs in vitro by cardiomyocytes (H9c2; Scale bar: 50 μm). (C) Representative images of the phagocytosis of GEP-NPs in vitro by fibroblasts (NIH/3T3; Scale bar: 50 μm). (D) Representative images of the phagocytosis of GEP-NPs by macrophages in vitro (Raw 264.7; Scale bar: 25 μm; exposure times: DAPI - blue 100 ms, CD68 - green: 200 ms, GEP-NPs - red: 150 ms). (E) Evaluation of the antioxidant capacity of GEP-NPs in H9c2 Cells (Scale bar: 50 μm; exposure time: 200 ms). (F) Evaluation of the anti-Fibrotic effects of GEP-NPs in NIH/3T3 Cells (Scale bar: 50 μm). (G) Evaluation of the anti-inflammatory effects of GEP-NPs in RAW 264.7 Cells (Scale bar: 25 μm). The images are representative of cells treated with 20 μg/mL GEP-NPs; GEP-NPs were labeled with Dil dye. (H) Fluorescence images of the fibrotic heart and important organs were obtained in mice 4 weeks after MI (tail vein injection of DiR-labeled Rabex or GEP-NPs once a week). (I)Characteristic co-localization images of Rabex and Col1 in MI tissue sections. (J) Characteristic co-localization images of GEP-NPs and Col1 in MI tissue sections. (K-N) Measure serum levels of ALT, AST, Cr, and BUN (indicators of hepatic and renal function) in each group of mice (n = 5). Significant differences were detected by two-tailed one-way ANOVA with Tukey's multiple comparisons test. All data were shown as means ± SD, P < 0.05. (O-P) Representative H&E and Masson staining images of liver and kidney tissue from each group of mice with MI (Scale bar: 50 μm).

3.4. GEP-NPs improved cardiac function in mice with MI

To determine whether GEP-NPs have therapeutic effects on MI, we injected GEP-NPs weekly via tail vein for four weeks after the successful establishment of MI model in mice. All mice were randomly divided into 4 groups: Sham group (In this group of mice, the chest was closed after only scratching the pericardium after thoracotomy, Sham), MI group (After opening the chest and ligating LAD, the mice were treated with chest closure after confirming the successful establishment of MI model, MI), MI + Rabex group (The mice in this group received intravenous administration of Rabex via the tail vein following MI, MI + Rabex) and MI + GEP-NPs group (The mice in this group received intravenous administration of conjugated collagen-targeted peptide-fused Rabex via the tail vein following MI, MI + GEP-NPs). Cardiac function was evaluated using small-animal echocardiography on day 28 following MI in each group of mice, with representative M-mode echocardiographic images presented in Fig. 5A. Compared with the MI group, the left ventricular ejection fraction (LVEF) and fractional shortening (LVFS) in the Rabex and GEP-NPs groups were significantly improved, with the GEP-NPs group showing a more pronounced improvement than the Rabex group (Fig. 5B and C). As important indicators for assessing ventricular remodeling, the left ventricular end-systolic diameter (LVDs) and end-diastolic diameter (LVDd) in both the Rabex and GEP-NPs groups were significantly reduced compared to mice in the MI group, with a more pronounced improvement observed in the GEP-NPs group (Fig. 5D and E). Histopathological evaluation of MI extent and fibrosis degree was performed using Hematoxylin-Eosin (HE) staining and Masson's trichrome (Masson) staining (Fig. 5F). Masson staining results demonstrated that compared with the MI, both Rabex and GEP-NPs treatments significantly attenuated myocardial fibrosis, with GEP-NPs exhibiting superior therapeutic efficacy over Rabex (Fig. 5G). Quantitative analysis further confirmed these observations, showing a consistent trend in the preservation of infarct zone wall thickness (Fig. 5H). MI can lead to cardiomyocyte necrosis and a reduction in cardiac contractility, thereby substantially increasing the mechanical load and ventricular wall stress experienced by the adjacent viable myocardium. This pathological mechanism initiates maladaptive myocardial hypertrophy, which ultimately results in cardiac dilation and structural changes, commonly referred to as ventricular remodeling. To investigate the effects of GEP-NPs on cardiomyocyte hypertrophy, cellular identification was performed using 4′,6-diamidino-2-phenylindole (DAPI) and wheat germ agglutinin (WGA) staining (Fig. 5I). Quantitative analysis of cardiomyocyte cross-sectional area demonstrated that both Rabex and GEP-NPs effectively inhibited cardiomyocyte hypertrophy, with GEP-NPs exhibiting a more pronounced inhibitory effect (Fig. 5J). Simultaneously, angiogenesis represents a critical factor influencing the prognosis of MI. Consequently, through immunofluorescence detection of the vascular endothelial cell marker CD31, we observed that both the MI + GEP-NPs group and the MI + Rabex group demonstrated superior pro-angiogenic effects in the peri-infarct region compared to the MI group, with GEP-NPs exhibiting particularly pronounced efficacy (Fig. 5K and Fig. S2). The infarcted myocardial region elicits a pronounced inflammatory response, marked by the infiltration of inflammatory cells such as macrophages, which secrete a variety of cytokines, including interleukin-1β (IL-1β) and interleukin-6 (IL-6). These cytokines play a role in promoting fibrotic remodeling of cardiomyocytes. Representative immunohistochemical images of IL-1β and IL-6, along with semi-quantitative analysis results, demonstrate that administration of GEP-NPs significantly downregulates the expression of pro-inflammatory factors in the infarcted area compared to both the MI group and Rabex group (Fig. 5L–. O and P). Matrix metalloproteinases (MMPs) serve as pivotal drivers of ventricular remodeling following myocardial injury, exhibiting a strong correlation with heart failure and adverse clinical outcomes. Consequently, we employed immunohistochemical analysis to evaluate the expression levels of MMP2 and MMP9, the most extensively studied and strongly associated biomarkers, in cardiac tissues across experimental groups. The findings demonstrated a significant downregulation of both MMP2 and MMP9 expression in MI mice treated with either GEP-NPs or Rabex, with the GEP-NPs group exhibiting the lowest percentage of positive cells in cardiac tissues (Fig. 5M–. Q and R). Finally, we measured the reactive oxygen species (ROS) levels in cardiac tissues across all groups. The result demonstrated a consistent trend with previous findings, indicating that GEP-NPs exhibit significant antioxidant capacity (Fig. 5N). Building upon the aforementioned findings, this study further demonstrates that GEP-NPs exert significant anti-fibrotic, anti-inflammatory, pro-angiogenic, and anti-hypertrophic effects, thereby ameliorating ventricular remodeling and improving cardiac function following MI.

Fig. 5.

Cardiac ultrasound imaging and histopathological correlation analysis in MI mice following 4-week treatment with Rabex or GEP-NPs. (A) Representative M-mode echocardiographic images of MI in each group of mice at 28 days post-treatment. (B-E) Cardiac function, including left ventricular ejection fraction (LVEF), fractional shortening (LVFS), left ventricular end-systolic diameter (LVDs) and end-diastolic diameter (LVDd), was measured by echocardiography 28 days after MI (n = 6). (F) Representative H&E and Masson staining images of myocardial tissue from each group of mice with MI (Scale bar: 500 μm). (G-H) Quantitative analysis of infarcted area wall thickness and collagen deposition fraction in MI mice across experimental groups (n = 5). Significant differences were detected by two-tailed one-way ANOVA with Tukey's multiple comparisons test. (I) Representative WGA staining images of myocardial tissue from each group of mice with MI (Scale bars: 20 μm and 10 μm). (J) Violin plots showing CAS analysis of cardiomyocytes from 5 independent heart specimens (n = 5) per group. For each specimen, 150 cardiomyocytes (infarct border zone, 3 sections, 5 high-power fields/section) were quantified, and the mean CAS of the 150 cells was calculated for each heart to generate the data for statistical analysis. Data are presented as median with interquartile range. Significant differences were detected by two-tailed Brown-Forsythe and Welch ANOVA with Dunnett's T3 multiple comparisons test. (K) Representative immunofluorescence staining images for CD31 in infarct border zone across experimental mice groups. (L) Representative immunohistochemical images of inflammatory factors in MI hearts across experimental mice groups. (M) Representative immunohistochemical images of MMP2/MMP9 in MI hearts across experimental mice groups. (The cellular source of MMP expression was not determined in the current study.) (N) Representative immunofluorescence images of ROS in MI hearts across experimental mice groups. (O-R) Quantitative analysis of IL-1β⁺/IL-6⁺/MMP2⁺/MMP9⁺ cell percentage in MI areas across experimental mice groups (n = 5). Significant differences were detected by two-tailed one-way ANOVA with Tukey's multiple comparisons test. All data were shown as means ± SD.

3.5. Quantitative proteomic analysis reveals potential mechanisms underlying the protective effects of GEP-NPs in MI

Quantitative proteomics stands as one of the cutting-edge core technologies in contemporary life sciences, enabling the dynamic, comprehensive, and precise elucidation of global protein expression alterations in organisms under specific physiological or pathological conditions (e.g., health, disease states, and pharmacological interventions). Through bioinformatics analysis, this approach facilitates the identification of key dysregulated signaling pathways associated with diseases, thereby uncovering their underlying molecular mechanisms. To investigate the underlying mechanisms of GEP-NPs in treating MI, we conducted qualitative and quantitative statistical analyses of cardiac tissue proteomic data from the Sham, MI, and GEP-NPs groups in murine models. Following initial quality control procedures, principal component analysis (PCA) results of the proteomic datasets from these three groups were obtained (Fig. 6A). The functional manifestation of distinct gene products relies on their specific subcellular localization, rendering the prediction of protein subcellular localization pivotal for elucidating protein structure, functional characteristics, and molecular interactions. To this end, we performed systematic subcellular localization annotation and statistical analysis on differentially expressed proteins across three murine cardiac groups by leveraging cellular component information from the Gene Ontology (GO) database (Fig. 6B). Subsequently, we conducted classification and statistical analysis of the differentially expressed transcription factors by leveraging the transcription factor family information provided in the AnimalTFDB database (Fig. 6C). Fig. 6D presented the clustered heatmap of differentially expressed proteins across three groups of cardiac samples, obtained through differential protein clustering analysis, with red representing higher relative expression levels and blue representing lower relative expression levels. The heatmap depicting differential transcription factors is presented in Fig. 6E. GO functional enrichment analysis was performed using all identified proteins as the background. The bubble chart in Fig. 6F displayed the top 10 most significantly enriched differential proteins across three categories: biological process (BP), molecular function (MF), and cellular component (CC). KEGG pathway analysis serves as the most direct and essential approach to systematically and comprehensively elucidate biological processes, phenotypic or pathological mechanisms, and pharmacological actions at the cellular level, thereby facilitating the understanding of protein functions and their interrelationships. Fig. 6G displayed the 30 most significantly enriched pathways resulting from the KEGG pathway enrichment analysis of differentially expressed proteins. To elucidate the specific mechanism of action of GEP-NPs, we focused on the comparative analysis of cardiac proteomic data from mice in the MI group and the GEP-NPs group. Bioinformatic pathway analysis using Wiki pathways revealed that the therapeutic effect of GEP-NPs in improving MI is significantly associated with the suppression of aberrant activation of the PI3K-AKT-mTOR signaling pathway (Fig. 6H). To verify the accuracy of the quantitative proteomics analysis results, we used Western blotting to detect the expression of key proteins in the PI3K-AKT-mTOR pathway. The characteristic images were shown in Fig. 6I. The quantitative Western blot results demonstrated that the PI3K-AKT-mTOR signaling pathway was indeed aberrantly activated after MI, and that this activation was markedly attenuated upon treatment with GEP-NPs (Fig. 6J–L). Fig. S3 illustrates enrichment analysis of protein–protein interaction networks derived from differentially expressed genes. To verify the causal mechanism of the PI3K-AKT-mTOR pathway and identify the key cell type involved, we examined the effects of GEP-NPs in three major cardiac cell types.GEP-NPs significantly inhibited the abnormal activation of the PI3K-AKT-mTOR pathway only in TGF-β-stimulated NIH/3T3 fibroblasts, but not in cardiomyocytes or macrophages (Fig. S4). Furthermore, pharmacological activation of the PI3K-AKT-mTOR pathway abolished the inhibitory effects of GEP-NPs on fibroblast activation (Fig. S5). These results establish a causal role of the PI3K-AKT-mTOR pathway in mediating the function of GEP-NPs and identify cardiac fibroblasts as the primary responsive cell type. The aforementioned results suggest that GEP-NPs ameliorate chronic cardiac inflammation, ventricular remodeling and enhance cardiac function following MI through the inhibition of aberrant PI3K-AKT-mTOR pathway activation.

Fig. 6.

Quantitative cardiac proteomics analysis in mice elucidated the underlying mechanisms of GEP-NPs in MI treatment. (A) Score plot of PCA for the three distinct sample groups. (B) Subcellular localization and statistical analysis of differentially expressed proteins across the three experimental groups. (C) Statistical analysis of domain protein counts across the three experimental groups. (D) Heatmap visualization of the top 50 differentially expressed proteins ranked by total expression intensity across the three experimental groups. (E) Heatmap of differential transcription factors across the three groups. Differentially expressed proteins (DEPs) or differential transcription factors were identified by one-way ANOVA for multi-group comparison, with a threshold of p < 0.05 (no FDR adjustment). (F) Functional enrichment bubble chart of differentially expressed proteins among the three groups based on GO analysis. (G) Bubble plot of KEGG pathway enrichment analysis for differentially expressed proteins among the three groups. (H) Wiki pathway pathway enrichment bubble chart of the differentially expressed proteins between the GEP-NP and MI groups (TOP 20). (I) Representative WB images of key proteins in the PI3K-AKT-mTOR signaling pathway. (J-L) Quantitative statistical analysis results of WB data for the PI3K-AKT-mTOR signaling pathway (n = 5). Significant differences were detected by two-tailed one-way ANOVA with Tukey's multiple comparisons test. All data were shown as means ± SD.

4. Discussion

In this study, we developed a simple yet effective engineering approach to successfully transform exosomes derived from red cabbage into GEP-NPs with type I collagen targeting capability. Subsequently, through the integration of multi-omics technologies, we systematically analyzed the key bioactive molecular composition in Rabex and its engineered product GEP-NPs. Comparative component analysis revealed that the core bioactive molecular profile of Rabex remained largely unchanged after engineering modification. In vitro cellular experiments demonstrated that GEP-NPs not only exhibited efficient uptake by multiple cell types but also possessed antioxidant, antifibrotic, and anti-inflammatory functions. Animal model studies further revealed that GEP-NPs specifically bound to collagen components in MI regions and significantly accumulated in mouse cardiac tissues, confirming their excellent collagen targeting capability. Moreover, while maintaining favorable biocompatibility and safety, GEP-NPs significantly reduced myocardial fibrosis and inflammatory responses post-MI and improved cardiac function.

In recent years, plant-derived biomaterials have garnered extensive attention due to their excellent biocompatibility, low toxicity, antioxidant and anti-inflammatory properties, as well as sustainability advantages, and have been widely applied in therapeutic research for various diseases [[29], [30], [31]]. Among them, red cabbage (Brassica oleracea var. capitata f. rubra), a common cruciferous vegetable, has demonstrated promising biomedical applications in recent years due to its rich content of bioactive compounds. Research indicates that its exosomes exert significant anti-inflammatory effects by suppressing immune cell activation, while mitigating apoptosis in human keratinocytes and fibroblasts [25]. In inflammatory bowel disease models, Rabex not only inhibits macrophage-mediated inflammatory responses but also promotes the regeneration of colonic epithelial cells, thereby effectively alleviating intestinal inflammation [26]. It is noteworthy that, although Rabex has shown promising therapeutic potential, there have been no systematic investigations to date into its efficacy or underlying mechanisms in the context of MI.

However, considering the histopathological features of the damaged myocardium and the rapid hemodynamic environment within the heart, non-engineered plant-derived nanovesicles exhibit limited therapeutic efficacy in cardiac applications due to their inability to achieve effective local concentrations at the target site [11,14,32,33]. To address this challenge, engineered plant-derived exosomes can be utilized to achieve precise active targeting, thereby significantly enhancing drug delivery efficiency. Phage display technology is an advanced molecular screening method that enables the identification of target molecules with tissue-specific binding properties and supports their sequence characterization [34,35]. The GKWHCTTKFPHHYCL peptide, identified through in vivo phage display screening, exhibits high specificity for type I collagen, a major structural component of fibrotic myocardial tissue [28,36].

In this study, we developed a simple and efficient surface modification strategy for plant-derived exosomes, enabling targeted delivery to damaged myocardial tissue based on pathological characteristics post-MI. Given that the necrotic myocardium and the damaged myocardium caused by inflammation are gradually replaced by scar tissue, and accompanied by abnormal fibrosis of adjacent healthy myocardium, in this study, artificial liposomes were conjugated with the collagen-targeting peptide GKWHCTTKFPHHYCL, and then fused with successfully isolated Rabex through the freeze-thaw method. After membrane filtration, we obtained uniformly sized fusion Rabex (GEP-NPs) capable of specifically targeting type I collagen. To evaluate the impact of this engineering process on key bioactive molecules (such as proteins, nucleic acids, and essential metabolites) in Rabex, we systematically analyzed the major protein composition and active metabolite profiles of GEP-NPs and Rabex using proteomics and plant-wide metabolomics technologies. The results demonstrated that although GEP-NPs lost a small amount of bioactive components during the engineering process, most functional proteins and metabolites were preserved. Leveraging their superior ability to target collagen deposition sites in the MI region, GEP-NPs effectively delivered bioactive components to cardiac fibroblasts, cardiomyocytes, and inflammatory cells in heart tissue, significantly improving cardiac contractile function and alleviating ventricular remodeling in MI model mice. According to the biodistribution results shown in Fig. 4H, GEP-NPs effectively accumulated in the injured myocardium, while relatively low fluorescence signals were observed in other major organs including the liver and kidney. Consistently, liver and kidney function tests and histological analysis revealed no obvious fibrosis or pathological damage in these organs, indicating favorable in vivo biosafety and negligible off-target toxicity of GEP-NPs at the tested time point. Since the targeting peptide binds to type I collagen, which is abundant in various fibrotic tissues, potential accumulation in other fibrotic organs (such as liver or kidney under comorbid fibrosis) may exist and warrants further investigation in future translational studies. Histopathological analysis revealed that GEP-NP treatment markedly reduced infarct size, increased ventricular wall thickness, decreased collagen deposition and abnormal fibrosis, alleviated cardiomyocyte hypertrophy, promoted angiogenesis, and significantly suppressed inflammatory responses. To further elucidate the potential mechanisms by which GEP-NPs improve MI, we employed quantitative proteomics combined with bioinformatics analysis. The findings indicate that GEP-NPs likely exert their therapeutic effects by suppressing the aberrant sustained activation of the PI3K-AKT-mTOR signaling pathway—whose hyperactivation post-MI can induce fibroblast activation and excessive collagen deposition [37]. Biocompatibility and biosafety are fundamental requirements for biomaterials, not only concerning the organism's life activities and health status but also serving as prerequisites for material functionality and effective integration with biological systems. Experimental results confirmed that GEP-NPs exhibit excellent biocompatibility and biosafety.

Several aspects of the present study can be further refined in future investigations. First, the specific active components within plant-derived exosomes remain to be identified, which will help reduce the “black box” nature and facilitate standardized production and clinical translation. Second, the detailed molecular mechanism underlying the regulation of the PI3K-AKT-mTOR pathway warrants further validation, including identification of key functional molecules. Third, additional control groups and more systematic pharmacokinetic studies will help strengthen the in vivo experimental design. Finally, future studies will benefit from identifying the cellular sources of MMP expression and distinguishing between capillaries and arterioles in vascular analysis.

5. Conclusion

In summary, this study has established a facile, efficient, and highly operational strategy for surface modification of plant-derived exosomes, which has been validated through multi-omics detection methods to maximally preserve their bioactive components. Concurrently, by designing and optimizing the targeting of biomaterials based on disease pathological characteristics, the study successfully achieved specific delivery of GEP-NPs to damaged myocardial tissue, significantly enhancing their retention efficiency in cardiac tissue and prolonging the therapeutic duration. In addition, GEP-NPs have demonstrated outstanding therapeutic effects in improving inflammation and ventricular remodeling after MI. Finally, quantitative proteomic analysis systematically revealed the underlying therapeutic mechanisms. This study posits that this innovative drug delivery system not only provides novel insights for the development of disease treatment strategies and drug delivery platforms but also holds transformative potential for the treatment of cardiovascular diseases and other systemic disorders.

CRediT authorship contribution statement

Xi Tan: Formal analysis, Investigation, Validation, Visualization, Writing – original draft, Writing – review & editing. Simin Lu: Formal analysis, Investigation, Validation, Visualization, Writing – original draft, Writing – review & editing. Wangbei Qin: Investigation, Writing – original draft. Xiaoqin Zhu: Investigation, Writing – original draft. Xinlu Wang: Investigation, Validation. Kaiwei Ren: Investigation, Validation. Biwei Cheng: Writing – original draft. Jiahui Li: Writing – original draft. Dongze He: Writing – original draft. Ziran Hu: Validation. Yuhua Lei: Conceptualization, Funding acquisition, Supervision. Lixing Wu: Conceptualization, Funding acquisition, Supervision. Jing Zhang: Funding acquisition, Methodology, Supervision, Writing – original draft. Jiaxin Hu: Conceptualization, Funding acquisition, Methodology, Project administration, Supervision, Writing – review & editing.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

The following is the Supplementary data to this article:

Data availability

Data will be made available on request.