Plant-derived extracellular vesicles as emerging cardioprotective agents for cardiovascular diseases

Abstract

Cardiovascular diseases (CVDs) remain a leading cause of global morbidity and mortality. Their complex and multifactorial pathogenesis, involving endothelial dysfunction, chronic inflammation, oxidative stress, metabolic dysregulation, and pathological remodeling, limits the long-term effectiveness of current therapeutic strategies and underscores the need for novel treatment approaches. Plant-derived extracellular vesicles (PDEVs) have recently emerged as promising cardioprotective agents because of their favorable biocompatibility, relatively low immunogenicity, abundant endogenous bioactive cargoes, and engineering flexibility. Owing to these properties, PDEVs possess dual characteristics as natural nanocarriers and bioactive therapeutic agents. Preclinical evidence from various in vitro and in vivo cardiovascular disease models indicates that PDEVs exert antioxidative, anti-inflammatory, immunomodulatory, and tissue-reparative effects, thereby attenuating myocardial injury, reducing oxidative stress, and promoting cardiomyocyte survival. Beyond their intrinsic therapeutic activities, PDEVs can also serve as multifunctional drug delivery vehicles for small-molecule drugs, nucleic acids, proteins, and natural bioactive compounds, improving cargo stability, bioavailability, and therapeutic performance. Recent advances in surface functionalization, membrane fusion, and biomimetic design have further enhanced their targeting capacity and functional controllability. This review focuses on the application of PDEVs in cardiovascular disease therapy, systematically summarizing their preparation, characterization, quality evaluation, and relative advantages and limitations compared with conventional nanocarriers. It further highlights their therapeutic effects in different cardiovascular disease models, drug delivery applications, engineering strategies, and the current progress and key challenges in clinical translation. Continued advances in this field may promote the translation of PDEVs from experimental research to clinical application and broaden their value in cardiovascular nanomedicine.

Graphical abstract

Highlights

λ PDEVs exhibit excellent biocompatibility and minimal immunogenicity, making them ideal for cardiovascular applications.

λ Advanced techniques for PDEVs isolation and characterization enable stringent quality control and reproducibility.

λ PDEVs protect the heart through antioxidant, anti-inflammatory, anti-apoptotic, and reparative effects in various cardiovascular diseases.

λ PDEVs as therapeutic agents and engineered PDEVs for drug delivery promote clinical application and translational potential.

Introduction

Cardiovascular diseases (CVDs) remain the leading cause of death and disability worldwide. Their onset and progression are driven by interconnected pathological processes, including endothelial dysfunction, chronic inflammation, redox imbalance, metabolic dysregulation, and structural remodeling of the myocardium and vasculature [1–4]. CVDs primarily affect the heart and vascular system, presenting as impaired vascular function, inadequate myocardial perfusion and compromised cardiac pump performance, which ultimately lead to clinical sequelae such as myocardial ischemia-reperfusion injury, arrhythmias and chronic heart failure [5]. Their consequences can also extend beyond the heart and vessels, affecting organs such as the brain, kidneys and peripheral tissues and further increasing disability and mortality risks [6–8]. In addition, genetic susceptibility and common metabolic or lifestyle-related risk factors, such as hyperlipidaemia, hypertension, diabetes mellitus, smoking, obesity and sedentary behaviour, can promote endothelial injury, lipid deposition and inflammatory activation, thereby accelerating the chronic progression of CVDs [9, 10]. Mitochondrial dysfunction and immune dysregulation further reinforce these pathological cascades, promoting progressive structural and functional deterioration in the cardiovascular system [11]. Together, these convergent factors and pathways underpin the chronic, progressive, systemic and highly heterogeneous nature of CVDs at both clinical and pathological levels [5, 12].

Current management of CVDs includes risk-factor control, pharmacological therapy with agents such as statins, anticoagulants, β-blockers and renin-angiotensin-aldosterone system (RAAS) inhibitors, as well as revascularization and device-based interventions [4, 13–16]. However, long-term use of these agents is frequently accompanied by adverse effects, including myopathy and hepatotoxicity, increased bleeding risk, electrolyte disturbances and renal impairment, which can limit their sustained safe application [17–20]. Despite these advances, existing pharmacological and interventional approaches remain insufficient for precise and durable modulation of the multi-pathway processes that drive chronic cardiovascular progression. These limitations have stimulated interest in innovative therapeutic strategies capable of targeting disease-relevant tissues while reducing systemic adverse effects [21, 22].

In recent years, various nanodelivery platforms have been explored for CVD therapy; however, conventional synthetic nanocarriers still face limitations in biocompatibility, immune safety and suitability for long-term in vivo use [23, 24]. Against this background, extracellular vesicles (EVs) have attracted attention as endogenous nanoscale carriers with lipid-bilayer membranes, bioactive cargoes and surface molecules that support intercellular communication, cellular uptake and tissue interaction [25–29]. According to biological origin, EVs include animal-derived EVs, particularly mammalian EVs, and plant-derived extracellular vesicles (PDEVs); mammalian EVs are commonly classified by biogenesis as exosomes, microvesicles and apoptotic bodies [30, 31]. Unlike mammalian EVs, PDEVs still lack a unified biogenesis-based classification, so they are generally described by their plant origin, vesicular structure, nanoscale size and source-dependent molecular composition. Typically ranging from about 30 to 400 nm, PDEVs carry membrane-associated proteins, lipids, nucleic acids and secondary metabolites [32].

These structural and compositional features suggest that PDEVs should be considered not only as plant-derived nanoscale carriers but also as bioactive vesicular systems with source-dependent functions. Their membrane architecture and multifunctional cargo may support favourable biological compatibility and modulate redox balance, immune responses and inflammatory signaling, which are closely involved in cardiovascular injury and remodeling [33, 34]. In parallel, PDEVs can serve as delivery platforms for drugs, nucleic acids and other bioactive molecules, improving cargo stability, bioavailability and functional exposure in vitro and in vivo [35]. Engineering strategies further extend this dual function by improving delivery efficiency, targeting specificity and in vivo controllability. Surface functionalization, membrane fusion and biomimetic design can introduce targeting ligands, functional peptides or responsive modules to enhance lesion-specific delivery [36, 37]. With continued methodological refinement, plant-derived vesicles are progressing from basic research toward preclinical and early translational evaluation, although direct cardiovascular clinical evidence remains limited [38, 39]. Overall, available preclinical evidence supports the potential of PDEVs to limit myocardial injury, vascular inflammation, endothelial dysfunction and pathological remodeling, highlighting their dual identity as intrinsically active therapeutics and delivery vehicles for CVD therapy [40, 41].

In this review, we provide a comprehensive synthesis of the biological properties and cardiovascular relevance of PDEVs, with emphasis on their intrinsic therapeutic activities, drug-delivery roles and engineered PDEV platforms for CVD therapy. We also summarize the current status of clinical translation and the major barriers that must be addressed before cardiovascular application. First, we summarize the isolation and purification strategies, physicochemical characterization, stability, storage and compositional features of PDEVs, and further compare their advantages and limitations with conventional vesicular and nanocarrier systems. Next, in the context of the core pathological processes that drive CVDs, we delineate cellular uptake, disease-relevant targeting and the therapeutic pathways through which PDEVs may exert effects across distinct disease settings. We then discuss how engineering strategies such as membrane modification and biomimetic reconstruction can enhance vesicle stability, targeting performance and controllability of therapeutic efficacy. Finally, emerging clinical evidence for PDEV-based interventions is reviewed, with a critical discussion of key translational challenges and future priorities, including pharmacokinetics, biodistribution, cardiovascular targeting, scalable manufacturing, quality control, regulatory uncertainty, safety, immunogenicity and long-term biocompatibility.

An overview of PDEVs

Isolation and purification of PDEVs

PDEVs have shown significant potential in biological activity and therapeutic applications. The isolation and purification of PDEVs are crucial steps in their research and application, aiming to extract bioactive vesicles from various plant sources. PDEVs are derived from a wide range of plant parts, including roots, stems, leaves, flowers, bark, fruits, seeds, and dried aerial parts [42]. Therefore, specific pretreatment protocols must be implemented for different plant sources before isolation. The purpose of pretreatment is to release the vesicles, and common methods include tissue disruption and infiltration. Tissue disruption methods include mechanical disruption and enzymatic hydrolysis. Mechanical disruption (such as physical grinding or juicing) is simple and efficient, as demonstrated in the processing of Aloe vera [43], garlic [44], and ginseng [45]. However, high shear forces may damage cells, leading to the mixing of organelles and membrane structures with PDEVs, thus affecting their purity. Furthermore, the heat generated during the process may deactivate the bioactive compounds in PDEVs [46]. Some studies suggest that processing conditions can influence metabolite yields from the same plant material, with low-speed juicing showing higher metabolite recovery. However, this finding still requires further validation [47]. Another approach is enzymatic hydrolysis (commonly using a combination of cellulase and pectinase), which gently breaks down the cell wall while avoiding heat generation during processing. For example, Zhao et al. proposed a method for isolating PDEVs from Morinda officinalis by enzymatically degrading the cell wall using cellulase, pectinase, and other related enzymes [48]. However, this method is relatively time-consuming, with literature reports indicating that the treatment with enzyme complexes requires up to 12 h [49]. The infiltration method involves introducing a buffer solution into the plant tissue under vacuum pressure and then collecting the intercellular matrix by low-speed centrifugation and washing the solution. This method is milder than tissue disruption, reducing cell damage and resulting in higher purity. However, due to the dilution effect, the resulting PDEV concentration is relatively low [50, 51]. Thus, when higher concentrations of PDEVs are required, researchers often employ tissue disruption methods [52].

PDEVs isolation commonly relies on differential centrifugation (DC), ultrafiltration (UF), tangential flow filtration (TFF) and size-exclusion chromatography (SEC) to remove cellular debris and small-molecule contaminants, thereby improving vesicle purity and stability [53]. During purification, methods such as ultracentrifugation (UC), sucrose density gradient ultracentrifugation (SDG-UC) and immunoaffinity capture are often combined to further optimize the separation of exosome-like vesicles [54]. Particle size is typically characterized using nanoparticle tracking analysis (NTA), dynamic light scattering (DLS), transmission electron microscopy (TEM), cryo-electron microscopy and atomic force microscopy [55]. For example, in the study by Sriwastva and colleagues, exosome-like nanovesicles derived from mulberry bark were isolated using DC and subsequently purified using SDG-UC, yielding vesicles with anti-inflammatory activity [56]. In another study, Yang and colleagues isolated ginseng-derived exosome-like nanovesicles using DC and purified them by UC, obtaining vesicles with cardioprotective effects [57].

Physical characterization of PDEVs

Particle size distribution, surface charge, and nanoscale morphology constitute the three core dimensions of physical characterization for PDEVs. Table 1 summarizes these parameters across representative plant sources and provides useful information for assessing PDEV manufacturing feasibility, storage stability, and translational potential [58]. PDEVs are extracellular vesicle-like nanostructures released by plant cells and enclosed by a phospholipid bilayer. Figure 1 further shows that PDEVs commonly display spherical, ellipsoidal, or cup-shaped morphologies and usually range from approximately 30 to 400 nm in diameter [59]. For instance, You and colleagues measured cabbage-derived exosome-like nanovesicles by NTA and reported a mean diameter of approximately 100 nm. Their morphology was further confirmed by TEM, scanning electron microscopy, and cryo-electron microscopy imaging [60]. In a study by Emmanuela and colleagues, TEM-based characterization of exosome-like nanovesicles derived from Solanum nigrum identified predominantly spherical vesicles [61]. Consistently, Chen and colleagues used TEM to examine tea leaf-derived exosome-like nanovesicles and observed round or ellipsoidal structures, further supporting the typical morphological features of PDEVs [62]. The zeta potential of plant-derived EVs is commonly measured using electrophoretic light scattering (ELS) or related DLS-based platforms. As shown in Fig. 2, it usually ranges from near-neutral values to approximately −50 mV, indicating a net negative surface charge that may enhance colloidal stability in aqueous environments [63, 64]. Wang et al. reported a zeta potential of −15.2 mV for exosome-like nanovesicles derived from bitter melon (Momordica charantia), consistent with good dispersion stability [65]. Sánchez López et al. measured a zeta potential of −9.4 mV for pomegranate (Punica granatum)-derived exosomes, providing further support for the aqueous stability of plant-derived vesicles [66]. Beyond physicochemical properties, PDEVs comprise lipids, proteins, nucleic acids, and metabolites [52, 67]. These components can be profiled using western blotting, the bicinchoninic acid assay, LC-MS, HPLC, Raman spectroscopy, and other analytical techniques [68, 69]. Such compositional information is essential for understanding the biological activities of plant-derived EVs, particularly their roles in intercellular signaling, inflammation modulation, and therapeutic carrier functions [32, 70]. Accordingly, PDEVs are increasingly being investigated as an emerging therapeutic modality in translational and clinical research.

Table 1.

Summary of extraction, purification, size, morphology, zeta potential, biological functions, and disease applications of PDEVs

Plants	Extraction techniques	Purify	Size (nm)	Morphological characterization	Zeta potential (mV)	Biological function	Illnesses	Refs.
Mulberry bark	DC	SDG-UC	NTA: 196.4 ± 45.4	TEM: exosome-like bilayer	Not reported	Anti-inflammatory	Colitis	[56]
Tea (Camellia sinensis)	Cold press homogenization	TFF, SEC	NTA: 133.8 ± 9.2	TEM: spherical vesicle	DLS: −15.5	Anti-inflammatory, Antioxidant protection	MI/R	[57]
Cabbage	DC	UF, SEC	NTA: 100.0	TEM: spherical vesicle	ELS: −14.8	Anti-inflammatory	Inflammatory-related diseases	[60]
Black nightshade berries	NF, DC	PEG	NTA: 107.0	TEM: spherical vesicle	DLS: −0.6	Anti-inflammatory	Inflammatory-related diseases	[61]
Tea (Camellia sinensis)	Gradient centrifugation	UC	NTA: 166.9	TEM: round or oval vesicles	DLS: −10.4	Pro-apoptosis	Breast cancer	[62]
Bitter melon (Momordica charantia)	DC	UC	DLS: 106.0	TEM: cup-shaped vesicle	ELS: −11.2 ± 3.1	Anti-inflammatory, antioxidant protection	Colitis	[65]
Pomegranate (Punica granatum)	NF	SEC	DLS: 148.7 ± 9.2	TEM: spherical vesicle	ELS: −9.4	Antioxidant	Inflammatory-related diseases	[66]
Avocado	DC	UF	NTA: 120.0	TEM: round or oval vesicles	DLS: −12.8	Improve cardiac function	AS	[78]
Carrot	DC	SEC, UF	NTA: 100.0	TEM: spherical vesicle	DLS: −25.7	Antioxidant effect	MI/R	[79]
Danshen (Salvia miltiorrhiza)	DC, gradient centrifugation	UC	NTA: 10.1–110.0	TEM: membrane vesicle	DLS: −17.2	Angiogenesis promotion	MI/R	[41]
Broccoli	DC	UC	NTA: 150.0	TEM: spherical vesicle	DLS: −15.5	Pro-apoptotic effects	Pancreatic ductal adenocarcinoma	[80]
Ginseng (Panax ginseng)	DC	UC, PEG	NTA: 69.0	TEM: spherical vesicle	DLS: −22.4	Cardioprotective	Myocardial injury	[57]
Sanqi (Panax notoginseng)	DC, density gradient	UC	NTA: 151.3	TEM, Cryo-electron microscopy: round or oval vesicles	DLS: −21.8	Anti-inflammatory Neuroprotective	MI/R	[81]
Carrot root	CF	SEC	NTA: 143.9	TEM: spherical vesicle	DLS: −10.2	Antioxidant effects	Cardiac ischemia	[82]
Mandarin orange (Citrus reticulata Blanco)	PEG	Filtration, PEG	DLS: 150.0	TEM: spherical vesicle	Not reported	Antioxidant activity Anti-inflammatory effects	Atherosclerosis pathology	[83]
Avocado (Persea americana)	DC	Filtration, PEG	DLS: 30.0–170.0	TEM: cup-shaped bilayer vesicles	DLS: −20.0 to −30.0	Anti-inflammatory	Inflammatory-related diseases	[78]
Platycodon	UC	Density gradient	DLS: 30.0–150 0.0	TEM: spherical vesicle	Not reported	Systemic anti-tumor response	Triple-negative Breast Cancer	[84]
Gardenia (Gardenia jasminoides)	Crushing	Density gradient	NTA: 100.0–150.0	TEM: spherical vesicle	DLS: −5.2 ± 2.8	Improves mitochondrial function	Parkinson’s disease	[85]
Garlic chives (Allium tuberosum)	Homogenization, differential centrifugation	Sequential filtration, ultracentrifugation	NTA: 150.0	TEM: spherical vesicle	DLS: −14.0	Improves mitochondrial function	Dexamethasone-induced sarcopenia	[86]

Fig. 1.

Particle size distribution of extracellular vesicles from different plant sources. To ensure clarity, the median value was selected for plants with a range of particle sizes. Detailed data can be found in Table 1

Fig. 2.

Zeta potential distribution of extracellular vesicles from different plant sources. To ensure clarity, the median value was selected for plants with a range of zeta potentials. Detailed data can be found in Table 1

Stability and storage of PDEVs

The stability of PDEVs is a prerequisite for their characterization, functional evaluation, and therapeutic application. Increasing evidence suggests that pH, ultrasonic treatment, chemical reagents, and temperature can affect the concentration, physicochemical properties, biological functions, and overall stability of PDEVs. Feng et al. demonstrated that grapefruit-derived PDEVs exhibited better size uniformity in acidic solutions compared to neutral and alkaline solutions, where no significant changes were observed [71]. Moreover, grapefruit-derived PDEVs displayed a negative charge in neutral or alkaline environments, while they exhibited a weak positive charge in acidic conditions. Additionally, ginger-derived vesicle-like nanoparticles showed good stability in simulated gastric and intestinal environments [72]. The intact structure of PDEVs is crucial for their biological function. Boiling and ultrasonic treatments both disrupt the structure of PDEVs, leading to a loss of their biological activity. Another review reported that pretreatment with the chemical detergent Triton X-100 reduces the activity of PDEVs [73]. Temperature has a significant effect on the stability of PDEVs. Huang et al. found that Centella asiatica-derived nanovesicles stored at 20 °C exhibited enhanced stability compared to those stored at 4 °C, which experienced particle aggregation [74]. Yan et al. further confirmed the significant advantage of low-temperature storage, showing that exosome-like nanovesicles derived from Tripterygium wilfordii maintained stable size and diameter after one year of storage at −80 °C, with minimal changes [75]. Further studies are needed to systematically evaluate the stability of PDEVs from different plant sources and support consistent physicochemical properties across batches.

The storage of PDEVs is a critical stage for preserving their biological activity and integrity, particularly for the development of drug delivery systems. Low-temperature storage is the most common method for preserving PDEVs. Storage at temperatures of 4 °C, −20 °C, and −80 °C is typically used, with each temperature condition exerting different effects on the morphology and physicochemical properties of PDEVs. Repeated freezing and thawing can alter the biological characteristics, content, and surface markers of PDEVs. Short-term storage at 4 °C maintains the activity of PDEV proteins, but prolonged storage may lead to protein degradation and a loss of biological function. Storage at −20 °C is generally suitable for PDEV preservation over several weeks to several months and may help maintain physicochemical stability. Storage at −80 °C provides the most stable conditions for long-term storage, lasting several months or even years. In addition, PDEV storage solutions commonly include phosphate-buffered saline, and cryoprotectants such as glycerol, dimethyl sulfoxide, and trehalose are often used to enhance stability and minimize freezing-induced damage [76, 77].

Composition analysis of PDEVs

Component analysis is a central part of the quality control framework for PDEVs. Because plant sources vary substantially, their vesicles may differ in molecular composition, making systematic compositional profiling particularly important. Figure 3 illustrates the sources, structural features, characterization methods, and major molecular components of PDEVs.

Fig. 3.

Sources, characterization and composition of plant-derived extracellular vesicles. (a) Plant sources, pre-treatment and isolation workflow for obtaining PDEVs. (b) Physicochemical characterization of PDEVs, including representative TEM morphology images, reproduced with permission from [126, 127] and measurements of size distribution and zeta potential. (c) Endosome-related biogenesis and major molecular cargo of PDEVs, including lipids, proteins, nucleic acids and metabolites

Proteins

Of particular note are members of the annexin family, which are widely present in higher plants such as Arabidopsis thaliana, Oryza sativa, Triticum aestivum, and Zea mays. As Ca^²⁺-regulated membrane-bound proteins, they are involved in drought resistance and antioxidant stress responses, regulate reactive oxygen species (ROS) signaling, and can be encapsulated into EVs [87, 88]. Upon pathogen invasion, these vesicles are secreted extracellularly to deliver defense molecules to infection sites and enhance plant immunity [51, 89]. In a notable study, after stripping the outer membrane proteins of garlic-derived PDEVs using trypsin, the efficiency of their cellular internalization was significantly lower than that of untreated vesicles, suggesting that outer membrane proteins play a crucial promotional role in their internalization. In addition, broccoli-derived PDEVs are enriched in plasma membrane intrinsic protein type aquaporins (AQPs), which facilitate transmembrane water transport, as reflected by higher and more sustained osmotic water flux [90–92].

Heat shock proteins (HSPs), including HSP60, HSP70, HSP80, and HSP90, represent another group of key protein components widely present in PDEVs [93, 94]. Specifically, HSPs are particularly abundant in vesicles derived from the juice sac cells of sweet orange, lemon, grapefruit, and bitter orange. HSP70 and HSP80 are often used as reference marker proteins in EV research [54, 94, 95]. As highly expressed cytoplasmic proteins in vesicles, they are often used to verify the vesicular identity of samples and reflect their universal roles in vesicle biogenesis and homeostasis [96, 97]. AQPs, including members of the plasma membrane intrinsic protein subfamily, have been systematically identified in vesicles isolated from broccoli. They have also been confirmed to be enriched in nanovesicles derived from juice sac cells of sweet orange, lemon [92, 93, 98]. As transmembrane water channels, AQPs regulate membrane permeability and root water uptake. At the vesicular level, AQP content is positively correlated with osmotic water flux and membrane protein stability [99]. Arabidopsis thaliana-derived PDEVs are rich in stress-responsive proteins and attachment-related proteins, and carry RNA-binding proteins associated with the selective loading of small RNAs. By comparison, Poncirus trifoliata-derived EVs are enriched in Patellin-3, clathrin heavy chain, AQPs, glycolytic enzymes, and multiple classes of oxidoreductases [100]. Together, these proteins participate in vesicular trafficking, cytoplasmic and secretory pathway regulation, cargo sorting, and vesicle secretion, thereby contributing to the formation and functional properties of PDEVs [32].

Lipids

The lipid components of PDEVs have been confirmed to not only play a supportive role in maintaining their structural stability, but also serve as an indispensable part in regulating their biological functions [101]. PDEVs contain four major lipid classes: glycerolipids, phospholipids, sphingolipids, and phytosterols. Glycerolipids include triacylglycerol, diacylglycerol, monogalactosyldiacylglycerol, and digalactosyldiacylglycerol, whereas phospholipids include phosphatidic acid, phosphatidylglycerol, phosphatidylcholine, phosphatidylinositol, phosphatidylserine, phosphatidylethanolamine, and minor amounts of phosphatidylmethanol and phosphatidylethanol. Sphingolipids are mainly represented by glycosylinositol phosphorylceramides [101–104]. In some plant-derived EVs, particularly celery-derived EVs, diacylglycerol and phosphatidic acid are abundant and have been associated with efficient vesicle fusion with cell membranes. Ginger-derived EVs achieve targeted uptake of specific bacteria through their lipid fingerprints, which not only reshapes the gut microbiota and enhances mucosal immunity but also directly antagonizes pathogen virulence [105]. The role of lipids is not limited to regulating the uptake and internalization of PDEVs; they may also influence the inherent therapeutic activities of PDEVs themselves. For instance, grapefruit-derived EVs are rich in phosphatidylethanolamine and phosphatidylcholine, and these two lipid components are presumably the key factors underlying their antioxidant and anti-inflammatory effects [106, 107].

Nucleic acids

The nucleic acid components in PDEVs include deoxyribonucleic acid (DNA) and various types of RNA, such as miRNAs, small RNAs and tiny RNAs [108, 109]. These nucleic acid components have potential in cross-kingdom regulation, being capable of regulating gene expression in animals particularly, in processes related to immune responses and tumor responses [75, 110, 111]. Specifically, miRNAs in PDEVs can regulate cellular functions and pathological processes by targeting mammalian genes [112]. These functional RNA molecules have been identified in exosome-like nanoparticles from several edible plants, including ginger, grapefruit, carrot, turmeric, and garlic. Vesicle-associated RNAs can also be taken up by gut bacteria, thereby modulating gut microbiota composition and host immunity [89, 113, 114]. These findings highlight the regulatory potential of PDEV-derived miRNAs in cross-kingdom communication and RNA-based therapeutic strategies [115]. Studies have shown that Arabidopsis thaliana uses tetraspanin 8-positive exosome-like vesicles to deliver gene-silencing small RNAs to the fungal pathogen Botrytis cinerea across kingdoms. These small RNAs precisely target and silence the pathogen’s virulence genes, thereby forming a robust innate immune strategy [116, 117]. This mechanism clarifies that Arabidopsis thaliana enhances antifungal immunity through an EV-mediated cross-species RNA interference pathway [118]. However, miRNA identification in PDEVs remains technically challenging, mainly because complete precursor miRNA sequence data are unavailable for many plant species [119]. Based on existing research evidence, the function of PDEV-derived miRNAs in regulating gene expression during specific pathological processes has been initially confirmed, demonstrating high application value and promising development prospects.

Metabolites

PDEVs extracted from broccoli serve not only as natural nanocarriers but also exhibit certain anti-inflammatory activities themselves [120]. They may enhance the stability, delivery, and biological effects of bioactive compounds such as sulforaphane, supporting potential applications in nutrition, cosmetics, and therapeutics [121]. Meanwhile, limited evidence suggests that they may have cancer-protective effects [122]. Grapefruit-derived EVs contain naringenin, naringin, fructose, citric acid, glucose, sucrose, inositol, oxalic acid, and aucubin, as well as the amino acids leucine and isoleucine [123, 124]. Strawberry-derived EVs contain high levels of ascorbic acid [46]. Apple-derived EVs are enriched in flavonoids and furanocoumarins, compound classes that have been reported to exhibit antifungal and insecticidal activities [125]. Tea-derived EVs are rich in small-molecule metabolites, such as polyphenols and flavonoids, and may affect apoptosis, cell migration, and immune responses [62]. In general, PDEVs contain diverse metabolites, and their metabolic composition varies markedly among plant sources. This source-dependent metabolic diversity may contribute to the biological activity of PDEVs and support their development as therapeutic vesicles.

PDEVs and conventional vesicular carriers

Conventional and synthetic nanocarriers have provided important technical foundations for controlled drug delivery. According to carrier architecture and representative vesicular or vesicle-like formulations, these systems mainly include liposomes, polymeric nanoparticles, metal-based nanoparticles, animal-derived extracellular vesicles, synthetic exosome-like vesicles and PDEVs. Liposomes, including conventional liposomes, PEGylated liposomes and stimuli-responsive liposomes, are among the most established lipid-based vesicles and can accommodate both hydrophilic and lipophilic agents [128, 129]. Polymeric nanoparticles, represented by PLGA nanoparticles and chitosan nanoparticles, offer tunable architecture, surface chemistry and release behavior, making them suitable for sustained or responsive delivery [130–132]. Metal-based nanoparticles, such as gold nanoparticles, iron oxide nanoparticles, mesoporous silica nanoparticles and hybrid nanoparticles, provide structural stability and imaging-related functions, but their systemic use may be limited by poor biodegradability, metal ion release, oxidative injury and tissue accumulation [133, 134]. Animal-derived extracellular vesicles, including MSC-EVs, cardiac cell-derived EVs, endothelial cell-derived EVs and macrophage-derived EVs, have expanded vesicle-based delivery because they combine membrane-associated biological functions with cargo transfer capacity. Synthetic exosome-like vesicles, such as cell membrane-derived nanovesicles, artificial exosomes and biomimetic exosome-like vesicles, further improve controllability and functional design. However, these EV-related systems still face challenges in purification, scalability, heterogeneity and safety control [135, 136].

Compared with these systems, PDEVs, such as ginger EVs, Salvia miltiorrhiza EVs and onion EVs, should be considered bioactive vesicular carriers rather than inert delivery containers. Their plant-derived membrane structure and endogenous cargo may participate in cellular recognition, tissue interaction and biological regulation, which is particularly relevant to CVDs because cardiovascular pathology usually involves coordinated endothelial injury, immune-metabolic imbalance, oxidative stress, mitochondrial dysfunction and tissue remodeling [137]. At the same time, PDEVs occupy an intermediate position between natural vesicles and engineered nanocarriers. They preserve source-dependent biological complexity while remaining modifiable through ligand insertion, cargo loading, membrane fusion and responsive design [137–139]. To clarify their relative position among representative nanocarrier and EV-based systems, Table 2 summarizes their representative vesicles or carrier forms, architectures, loading mechanisms, major advantages, limitations and cardiovascular relevance.

Table 2.

Comparative advantages and limitations of PDEVs over conventional nanocarriers for CVDs

Platform	Representative vesicles or carrier forms	Loading mechanisms	Advantages	Disadvantages	Cardiovascular relevance	Refs.
Liposomes	Conventional liposomes; PEGylated liposomes; stimuli-responsive liposomes	Aqueous encapsulation; bilayer insertion; surface conjugation	Mature formulation; controllable size; dual cargo compatibility	Drug leakage; rapid clearance; limited intrinsic activity	Delivery for vascular inflammation, AS, MI/R injury and thrombosis	[128, 129]
Polymeric nanoparticles	PLGA nanoparticles; chitosan nanoparticles	Entrapment; adsorption; conjugation; controlled release	Sustained release; tunable degradation; flexible surface design	Polymer-related toxicity; burst release; degradation uncertainty	Controlled delivery for AS, restenosis, MI/R and vascular remodeling	[130–132]
Metal-based nanoparticles	Gold nanoparticles; iron oxide nanoparticles; mesoporous silica or hybrid nanoparticles	Surface adsorption; pore loading; coordination; electrostatic binding	High stability; imaging ability; stimulus responsiveness	Poor biodegradability; metal ion release; long-term toxicity risk	Imaging-guided therapy, ROS modulation and targeted vascular delivery	[133, 134]
Animal-derived EVs	MSC-EVs; cardiac cell-derived EVs; endothelial cell-derived EVs; macrophage-derived EVs	Donor-cell engineering; incubation; electroporation; sonication	Biological signaling; natural membrane proteins; efficient cargo transfer	Low yield; donor dependence; complex purification; cargo safety concerns	Cardiac repair, angiogenesis, immune regulation and post-MI/R remodeling	[135, 136]
Synthetic exosome-like vesicles	Cell membrane-derived nanovesicles; artificial exosomes; biomimetic exosome-like vesicles	Membrane reconstruction; extrusion; electroporation; surface engineering	Higher controllability; scalable design; easy functionalization	Incomplete mimicry of natural EV functions	Delivery of RNAs, proteins and small molecules for myocardial or vascular repair	[136]
PDEVs	Ginger EVs; grapefruit EVs; green tea EVs; Salvia miltiorrhiza EVs; onion EVs; Gouqi nanovesicles	Endogenous loading; passive incubation; membrane partitioning; ligand insertion; membrane fusion; electroporation; hydrogel encapsulation	Natural bioactive cargos; intrinsic anti-inflammatory and antioxidant activities; low risk of zoonotic pathogen transmission; modifiable targeting and engineering flexibility; biocompatibility; cross-kingdom bioactivity; protection of loaded cargos from degradation; suitability for combinational therapeutic and delivery functions	Source heterogeneity; unclear pharmacokinetics; limited quality standards; engineering complexity and regulatory uncertainty	Regulation of oxidative stress, inflammation, endothelial injury, immune imbalance, cardiomyocyte apoptosis, mitochondrial dysfunction and adverse remodeling; protection across AS, ischemic myocardial injury, cardiometabolic remodeling and therapy-related cardiotoxicity; targeted delivery to ischemic myocardium, injured vessels and vascular calcification lesions	[137–139]

The therapeutic role of PDEVs in CVDs

The therapeutic value of PDEVs in CVDs should be considered within both the shared pathological framework of cardiovascular injury and the distinct features of individual disease entities. Because different CVDs involve varied combinations of vascular, myocardial, immune and metabolic abnormalities, PDEVs may provide benefit by coordinately regulating multiple injury-related processes in a disease-specific manner.

The pathogenesis of CVDs

The pathogenesis of CVDs involves a complex interplay of genetic, environmental, lifestyle, and pathophysiological factors, with key mechanisms including endothelial dysfunction, inflammation, oxidative stress, metabolic dysregulation, hypertension, thrombosis, and unhealthy lifestyle habits [140–142]. Endothelial damage weakens the vascular anti-inflammatory and anticoagulant functions, promoting AS and thrombosis, and increasing the risk of cardiovascular events [143, 144]. Chronic low-grade inflammation activates immune cells and releases cytokines, exacerbating vascular injury and advancing the progression of AS, coronary artery disease, and other related conditions [145, 146]. Oxidative stress leads to the overproduction of free radicals, damaging the endothelium, inducing lipid peroxidation, and altering protein functions, thereby accelerating vascular aging [147, 148]. Metabolic dysregulation, particularly in diabetes and dyslipidemia, enhances inflammation and promotes lipid deposition, thereby accelerating AS [149, 150]. Hypertension causes mechanical damage to the vessel wall, stimulating arterial thickening and sclerosis, and increasing cardiac workload, ultimately leading to heart failure [151]. Thrombosis, a major cause of acute cardiovascular events, is typically triggered by hypercoagulability, endothelial injury, and platelet activation [152]. Genetic predisposition and poor lifestyle choices, particularly physical inactivity, poor diet, smoking, and excessive alcohol consumption, further increase the risk of CVDs by promoting obesity, enhancing inflammation, and altering lipid profiles [153–155]. These shared pathogenic processes provide the biological basis for exploring PDEVs as multi-component vesicular regulators in CVDs. Because PDEVs may act only after reaching disease-relevant cells and modulating intracellular signaling networks, their cellular uptake and mechanistic actions should be clarified before discussing their disease-specific therapeutic effects.

Cellular uptake and mechanistic actions of PDEVs

The biological effects of PDEVs in CVDs require effective contact with disease-relevant recipient cells rather than simple tissue exposure [4, 156]. Available evidence suggests that plant-derived vesicles can be internalized by major cardiovascular disease-relevant cells, including cardiomyocytes, endothelial cells, macrophages, vascular smooth muscle cells and fibrosis-associated cells, although their precise uptake routes within cardiovascular tissues remain incompletely characterized [79, 157]. Based on current EV research, this process may involve endocytic uptake, lipid raft-related internalization, macropinocytosis, phagocytosis-like uptake in macrophages and partial membrane fusion [158]. These entry patterns are likely influenced by vesicle physicochemical properties and the pathological state of recipient cells, suggesting that uptake efficiency is an important determinant of downstream functional responses [159].

Once internalized, PDEVs may influence cardiovascular pathology through coordinated regulation of stress-response, inflammatory and remodeling-related pathways [160]. Activation of Nrf2-associated antioxidant programs may enhance HO-1 and NQO1 expression and limit ROS-driven mitochondrial injury [57]. Suppression of NF-κB signaling may reduce endothelial activation, cytokine release, leukocyte adhesion and macrophage recruitment [57, 161]. Vesicle-associated miRNAs and small RNAs may further regulate mammalian targets involved in lipid handling, endothelial dysfunction, fibrosis, angiogenesis and maladaptive remodeling [81]. PDEVs may also affect NLRP3 inflammasome activation, apoptosis, ferroptosis, mitochondrial homeostasis and macrophage polarization, thereby linking molecular signaling changes to improved endothelial function, reduced fibrosis, smaller infarct burden and enhanced cardiac repair [161, 162].

The types and pathological mechanism of CVDs

Although cardiovascular diseases share core pathological processes, they diverge in initiating triggers, dominant cell types and progression patterns across indications. Figure 4 outlines a hierarchical, indication-specific view of pathogenesis, highlighting where shared mechanisms converge and where disease-defining nodes emerge. This organization clarifies the mechanistic context in which PDEVs may be deployed and helps rationalize differences in therapeutic responses across CVD entities. Given that PDEVs act through disease-specific mechanisms across different cardiovascular conditions, representative studies are summarized in Table 3, with emphasis on plant sources, vesicular contents, experimental models, administration regimens, molecular mechanisms, and cardiovascular protective outcomes.

Fig. 4.

Mechanistic landscape of CVDs and PDEVs therapeutic targets. Shared and disease-specific pathogenic pathways across major CVDs and their modulation by PDEVs

Table 3.

PDEVs as biotherapeutics to treat a variety of CVDs

CVDs type	Sources	Contents	Type	Experimental model(s)	Doses, route, and duration	Mechanisms/Effects	Cardiovascular protection	Refs.
Atherosclerosis (AS)
	Onions	Proteins, lipids, nucleic acids	In vitro	HUVECs	1 × 1010 particles/mL, co-incubated, for 24 h	IL-1β, IL-6, TNF-α, VCAM-1, ICAM-1, PPARγ and CD36 ↓; V-Onex ↑	Delaying the initiation atherosclerotic plaque formation	[157]
	Carthamus tinctorius L.	Proteins, small RNAs, a broad range of metabolites	In vitro	ox-LDL-treated HUVECs	40 µg/mL, co-incubation, for 12 h	CXCL12, VCAM-1 and ICAM-1 ↓; miR166a-3p and miR159a ↑	Attenuating arterial wall inflammation and plaque formation	[171]
			In vivo	Eight-week-old male ApoE−/− mice and age-matched wild-type C57BL/6 mice	40 mg/kg, i.g., for 12 weeks	IL-1β, IL-6 and CXCL12 ↓; heart accumulation ↑
	Mori fructus	miRNAs	In vivo	ApoE−/− mice	5 and 50 µg/day, i.p., for 12 weeks	Plaque burden and CD68+ macrophages ↓; Plaque stability ↑	Lipid-lowering and plaque-stabilizing	[173]
	Avocado	Proteins, lipids, nucleic acids	In vitro	Peritoneal macrophages from C57BL/6 mice	50 µg/mL, co-incubation, for 12 h	p-NF-kB p65, NLRP3, foam cell formation and ox-LDL uptake ↓	Delaying plaque initiation and progression	[78]
Myocardial ischemia-reperfusion (MI/R)
	Salvia miltiorrhiza	Not reported	In vivo	C57BL/6 mice	10 mg/kg, i.g., for 3 days	CD31+ microvessel density, EF%, FS%, HUVEC viability and migration ↑	Promoting neovascularization and reperfusion recovery	[41]
	Exocarpium Citri grandis	miRNAs, a broad range of metabolites	In vivo	BALB/c mice	10 mg/kg, i.v., for 14 days	Cleaved Caspase-9, apoptosis, Cxcl2 and CD8 + T cells ↓; Ly6C−Ly6G− and mitochondrial membrane potential ↑	Alleviating MI/R injury	[182]
	Panax notoginseng	Lipids, proteins, miRNAs	In vitro	Primary microglia	10 µg/mL, co-incubation, for 2 h	IL-6 mRNA and TNF-α ↓; IL-10 mRNA ↑	Alleviating I/R-associated cerebrovascular disorders	[81]
			In vivo	SD rats	3 mg/kg, i.v., for 3 consecutive days	Infarct volume, peri-infarct apoptosis, M1 microglia ↓, M2 microglia ↑
Chronic heart failure (CHF)
	Green tea	Endogenous antioxidant components	In vivo	C57BL/6 mice	0.625 mg/kg, i.g., for 21 days	HAAPIR, MEF2D and MMP ↓; α-SMA ↑	Preventing CHF secondary to severe aortic pathology	[190]
	Carrot	Lipids, nucleic acids	In vitro	H9c2 cardiomyoblasts	1.0 × 1011 particles/mL, co-incubation, for 24 h	ROS, caspase-3 and cardiomyocyte apoptosis rate ↓; Nrf2, HO-1 and NQO1 ↑	Reducing ongoing cardiomyocyte loss	[79]
	Momordica	Proteins, lipids, nucleic acids,	In vitro	H9c2 cardiomyocytes	0.5, 5, 10 and 25 µg/mL, co-incubation, for 48 h	Mitochondrial damage and apoptosis, PCNA, Cyclin D1, Cyclin B1 and cell cycle ↓; p-AKT/AKT and p-ERK/ERK ↑	Attenuating myocardial remodeling and counteracting myocardial injury	[193]
			In vivo	BALB/c nude mice	100 µg/kg, i.p., for 35 days	Cardiac fibrosis area and CK-MB ↓
Arrhythmia (ARR)
	Ginseng root	miRNAs, tRNAs, snRNAs, rRNAs, snoRNAs	In vitro	H9c2 rat cardiomyoblasts	20, 40 and 80 µg/mL, co-incubated, for 24 h	Bax and Cyt ↓; cell viability, mitochondrial membrane potential and Bcl-2 ↑	Limitting adverse remodeling	[197]
	Gouqi	Carbohydrates, lipids	In vitro	Dexamethasone-induced myotube atrophy in C2C12	5 × 10^8 particles/mL, co-incubation, for 24 h	Myotube diameter, mitochondrial membrane potential and ATP ↑	Reduce the risk of ARRs	[199]
			In vivo	C57BL/6J mice	A total of 1 × 108 particles per 50 µL of GqDNVs, i.m., for 14 days	Myosin heavy chain↓; SIRT1, SIRT, PGC-1α, myogenic factor 5 and myogenin ↑
	Ginger rhizomes	Lipids, proteins, RNA	In vitro	C57BL/6 mice BMDMs	3 × 1010 particles/mL, co-incubated, for 16 h	Caspase-1 p10, Lactate dehydrogenase, speck formation, IL-1β and IL-18 ↓	Attenuating chronic sterile inflammation in the myocardium	[200]
Acute myocardial infarction (AMI)
	Carrot	Proteins, lipids, carotene	In vitro	H9c2 cardiomyoblasts	1 × 1011 particles/mL, co-incubated, for 24 h	Apoptosis and caspase-3 activity ↓; Nrf2, HO-1 and NQO1 protein maintained ↑	Protection of cardiomyocytes	[79]
	Gouqi	Proteins, genetic materials, lipids, flavonoids	In vitro	HL-1 cardiomyocytes	5 µL/well; hypoxia, for 12 h	Bax mRNA ↓; cell viability, Ki67+ and PCNA+ ↑	Reduce the infarct size	[160]
			In vivo	C57BL/6 mice	1 × 10^8 particles, i.m., for 14 days	LVEdV, LVEsV, infarct size, p-p38 and p-p65 ↓; survival, LVEF, LVFS and wall thickness↑
Hypertensive heart disease (HHD)
	Semen Sinapis albae	Proteins, small RNAs, a broad range of metabolites	In vitro	HUVECs	1 × 108 particles/mL, pretreatment for 1 h, Ang II 10 µM for 24 h	ET-1, Ang II, IL-6, CXCL1 mRNA, VCAM-1, P53, P21 and CD38 ↓; NO and ATP ↑	Lower the risk of subsequent cardiovascular events	[227]
			In vivo	Spontaneously hypertensive rats	50 µg/kg, i.g., for 8 weeks	Smooth muscle cell proliferation and arterial stiffness ↓; eNOS and VEGF-A ↑
	Taraxacum officinale	Lipids, functional proteins	In vivo	SD rat	0.5 mg/kg, i.g., for 4 weeks	IL-1β, IL-6 and TNF-α ↓; butyrate, valerate, caproate and occludin protein expression ↑	Reducing the risk of HHD	[228]
Diabetic cardiomyopathy (DCM)
	Salvia miltiorrhiza	A diverse range of proteins, lipids and nucleic acids	In vitro	RAW264.7 macrophages	100 µg/mL, co-incubated, for 6 h	cTnT, CK-MB, NT-proBNPs ↓	Impeding the progression of DCM	[161]
			In vivo	C57BL/6 mice	10 mg/kg, i.v., for 14 days	Pyroptosis axis ↓; SOD, NEDD4 and glutathione ↑
Cardiotoxicity (CT)
	Beta vulgaris	Proteins, lipids, nucleic acids, vitamin C, carotenoids	In vitro	HL-1 cardiomyocytes	5 × 10^10 particles/mL, co-incubated, for 24 h	BELNV uptake by HL-1 cells and cell viability partially restored ↑	Reducing the risk of CT	[162]
			In vivo	C57BL/6 N mice	5 × 1011 particles/kg, i.p., for 4 weeks	Malondialdehyde ↓; glutathione, xCT and GPX4 ↑
Myocardial fibrosis (MF)
	Bitter melon	Not reported	In vitro	H9c2 cardiomyocytes	0, 0.5, 5, 10 and 25 µg/mL, co-incubated, for 48 h	DNA damage, cleaved-caspase-3 and cleaved-PARP ↓; cell viability and Ki-67 ↑	Suppressing radiation-induced myocardial fibrosis	[193]
			In vivo	BALB/c nude mice	100 µg/kg, i.p., for 10 days	cTnT, CK-MB and NT-proBNPs ↓

Atherosclerosis (AS)

AS is an inflammatory vascular disease characterized by lipid accumulation within the vascular wall, which leads to plaque formation and luminal stenosis of blood vessels [163, 164]. Elevated levels of fatty substances and cholesterol in plasma, driven by smoking, obesity, diabetes mellitus, and vascular injury, are recognized as major contributors to the development of AS [165]. AS progression involves endothelial dysfunction, vascular smooth muscle cell migration and proliferation, macrophage infiltration and foam-cell formation. In advanced lesions, vascular calcification and plaque destabilization further increase the risk of acute cardiovascular events [166, 167]. Current therapies mainly focus on lipid lowering and risk-factor control, but these approaches do not fully resolve endothelial activation, plaque inflammation or residual cardiovascular risk [168, 169]. These pathological features make AS a suitable indication for evaluating whether PDEVs can regulate endothelial activation, macrophage lipid handling and plaque stability [170].

Onion-derived extracellular vesicles (Onex) were isolated from onion juice and functionalized with a vascular cell adhesion molecule 1 (VCAM-1)-targeting peptide (VHPK) on the vesicle surface to enhance accumulation and uptake by injured endothelium. In lipopolysaccharide (LPS)-stimulated human umbilical vein endothelial cells (HUVECs), VHPK-engineered Onex (V-Onex) showed enhanced uptake, consistent with VCAM-1 upregulation under endothelial stress. Functionally, V-Onex reduced VCAM-1 and intercellular adhesion molecule 1 (ICAM-1) expression and decreased monocyte adhesion, indicating direct modulation of endothelial activation. In an oxidized low-density lipoprotein (ox-LDL)-induced foam-cell model, Onex and V-Onex reduced intracellular lipid accumulation and suppressed foam-cell-associated genes, including PPARγ and CD36. These findings suggest that Onex may interfere with both endothelial activation and macrophage lipid handling during early atherogenesis [157].

Carthamus tinctorius L.-derived nanovesicles (CDNVs) provide another example of PDEV-mediated vascular protection in AS. After oral administration, CDNVs entered the systemic circulation and showed tissue-related distribution, supporting their feasibility as orally delivered vesicular regulators. In AS models, CDNVs reduced circulating interleukin-1 beta (IL-1β), IL-6 and C-X-C motif chemokine ligand 12 (CXCL12) and decreased plaque burden in the whole aorta and aortic root. These effects occurred without marked changes in body weight or plasma lipid profiles, indicating that CDNVs may act mainly through vascular inflammatory signaling and chemokine regulation rather than direct lipid lowering. Their effects were more pronounced than those of hydroxysafflor yellow A in the reported model, highlighting the potential advantage of vesicle-associated multi-component regulation. In ox-LDL-injured HUVECs, CDNVs reduced ROS accumulation, apoptosis and monocyte adhesion. These changes were accompanied by lower VCAM-1 and ICAM-1 expression, linking CDNV treatment to suppression of endothelial activation. Mechanistically, the miR-166a-3p/CXCL12 axis was implicated in the endothelial-protective effect of CDNVs. This finding connects plant vesicle-derived small RNAs with chemokine-driven endothelial injury during atherogenesis. Therefore, CDNVs expand the AS-related actions of PDEVs from general vascular protection to RNA-associated modulation of endothelial activation, chemokine signaling and plaque development [171].

Mori fructus-derived extracellular vesicle-like nanoparticles (MFEVLPs) were isolated and purified from mulberry juice and subjected to basic characterization. In high-fat-diet-fed ApoE−/− mice, MFEVLPs showed anti-atherosclerotic activity after repeated administration [172]. MFEVLPs not only reduced plaque area in the whole aorta and aortic root but also improved plaque composition, as evidenced by decreased macrophage infiltration and an increased proportion of smooth muscle cells, consistent with enhanced plaque stability. These changes were accompanied by reduced total cholesterol, triglycerides, and LDL cholesterol, as well as attenuated hepatic lipid deposition. Mechanistically, MFEVLPs downregulated key lipogenic and cholesterogenic regulators, including sterol regulatory element-binding protein 1 (SREBP1), fatty acid synthase (FAS), acetyl-CoA carboxylase (ACC), and 3-hydroxy-3-methylglutaryl-CoA reductase. Small-RNA profiling and reporter assays further suggested that plant miRNAs carried by MFEVLPs may regulate targets such as 3-hydroxy-3-methylglutaryl-CoA reductase. This mechanism links vesicle-derived small RNAs to hepatic cholesterol biosynthesis, improved plasma lipid profiles and plaque stabilization [173]. Overall, by preferentially accumulating in the liver and dampening lipid and cholesterol biosynthetic pathways, MFEVLPs may mitigate and delay AS progression.

Myocardial ischemia-reperfusion (MI/R)

MI/R injury occurs when restoration of coronary blood flow triggers abrupt redox imbalance, inflammatory amplification and microvascular dysfunction after myocardial ischemia [174], which leads to oxidative stress damage to myocardial tissue. These events promote cardiomyocyte death, endothelial injury, microvascular obstruction and later fibrotic remodeling, thereby limiting functional recovery after reperfusion [175]. Current interventions mainly aim to mitigate myocardial damage after blood flow restoration, regulate excessive immune activation and promote tissue repair [176, 177]. However, drug therapy has limited efficacy and cannot reverse the progression of heart failure after myocardial infarction, while stem cell therapy also carries safety risks. This vesicle-mediated repair concept opens a path for evaluating whether PDEVs can support cardiomyocyte survival, endothelial recovery and neovascularization after MI/R injury [178, 179].

Salvia miltiorrhiza-derived exosome-like nanoparticles (DDNs) have been evaluated as pro-angiogenic vesicles for MI/R repair. Given the relatively low yield of mammalian exosomes, DDNs are considered more economical, readily accessible, and higher-yield, while exhibiting exosome-like size, surface charge, morphology, and stability [102]. The isolated vesicles displayed nanoscale size and negative surface charge, supporting their colloidal stability. Functionally, DDNs showed both endothelial and cardiac benefits in MI/R-related models. In vivo, a murine MI/R model was established by thoracotomy under isoflurane anesthesia, LAD ligation to induce ischemia, and reperfusion after 60 min. DDN-treated mice showed higher ejection fraction and fractional shortening, together with increased cluster of differentiation 31 (CD31)-positive microvessel density in the peri-infarct region. Histological assessment of major organs did not reveal obvious tissue injury, providing preliminary safety evidence [41, 180, 181]. Collectively, these findings link DDN-mediated neovascularization with improved reperfusion recovery and cardiac pump function. Thus, DDNs represent a PDEV example in which the main disease-relevant outcome is vascular repair rather than only reduction of cellular stress.

In the context of cold ischemia-reperfusion injury (IRI) after heart transplantation, Exocarpium Citri grandis-derived exosome-like nanovesicles (ENVs) were selected from multiple plant-derived nanovesicle candidates. In an H₂O₂-induced oxidative injury model using HL-1 cardiomyocytes, these ENVs significantly reduced ROS, decreased apoptosis, and restored mitochondrial membrane potential. In LPS-stimulated RAW264.7 macrophages, ENVs restrained NF-κB/Toll-like receptor-associated signaling. This was reflected by lower Ccl3, Tlr4, C-X-C motif chemokine ligand 2 (Cxcl2) and IL-6 expression. To improve macrophage-directed delivery during early IRI, ENVs were fused with calreticulin-externalized MSC membrane vesicles. Rapamycin was subsequently loaded via electroporation to generate rapamycin-loaded fusion nanovesicles (FNVs@RAPA). Leveraging the high-ROS microenvironment of the graft, the transplant was pre-labeled with a ROS-responsive azido sugar to introduce azide groups, followed by installation of dibenzocyclooctyne (DBCO) moieties on the FNVs@RAPA surface to enable bioorthogonal click chemistry, thereby promoting local enrichment within the transplanted heart. In a murine heart-transplant MI/RI model, FNVs@RAPA increased graft accumulation and preferential uptake by inflammatory macrophages. Treatment reduced Ly6C⁺ inflammatory macrophages, decreased IL-6/Cxcl2 expression and attenuated CD8⁺ T-cell responses. These changes were accompanied by less inflammatory infiltration and fibrosis and longer graft survival [182].

Beyond these representative studies, several PDEV preparations have shown activity in cardiomyocyte stress models relevant to reperfusion injury. For example, ginseng-derived PDEVs reduced ROS generation, stabilized mitochondrial membrane potential and limited apoptosis-related signaling in H9c2 cardiomyocyte injury models [57, 183]. Citrus-derived vesicles also reduced cytokine expression through NF-κB pathway regulation [183]. These findings are mechanistically relevant to MI/R injury, but most remain based on cellular stress models. Future studies should validate infarct size, microvascular obstruction, cardiac function and remodeling outcomes in standardized in vivo MI/R models.

Chronic heart failure (CHF)

CHF arises from primary myocardial or structural cardiac lesions that progressively impair ventricular pump function, which can manifest as systolic dysfunction, diastolic dysfunction, or a combination of both, leading to a reduction in cardiac output [184, 185]. Subsequently, the body chronically activates the sympathetic nervous system and the RAAS, and this excessive activation drives adverse remodeling and disease progression [186]. Current therapies mainly aim to reduce neurohumoral overload, improve hemodynamics and slow ventricular remodeling [187]. Direct evidence for PDEVs in established CHF models remains limited. Current studies mainly support their potential to regulate heart-failure-related processes, including vascular stiffness, adverse remodeling, cardiomyocyte injury and myocardial fibrosis [188].

Vascular calcification increases arterial stiffness and afterload, thereby contributing to ventricular remodeling and heart failure progression. Grape exosome-like nanoparticles (GENs) attenuated VC by limiting VSMC apoptosis and osteogenic transdifferentiation, with concurrent improvement in redox and inflammatory markers. These effects may indirectly reduce ventricular pressure burden and remodeling risk, particularly in heart failure phenotypes related to impaired vascular compliance [189]. These findings are significant because they connect GEN-mediated inhibition of vascular calcification with a potential reduction in ventricular afterload, a key contributor to CHF progression. By limiting VSMC osteogenic transition and arterial stiffening, GENs may help alleviate vascular remodeling that indirectly aggravates cardiac dysfunction. Nevertheless, this evidence remains indirect, and direct validation in established CHF models with ventricular functional endpoints is still required.

Green tea-derived vesicles loaded with an antisense oligonucleotide against heart-apoptosis-associated PIWI-interacting RNA (HAAPIR) reduced matrix metalloproteinase 9 (MMP9)-mediated vascular remodeling in an aortic dissection model. Oral treatment improved vascular architecture, increased survival and downregulated HAAPIR expression in cardiac tissue. These findings suggest that oral nucleic acid delivery via PDEVs and modulation of HAAPIR-the myocyte enhancer factor 2D (MEF2D)-MMP9 signaling axis may confer cardiovascular protection, providing indirect experimental support for potential applications in remodeling associated CVDs such as heart failure [190].

Carrot-derived nanovesicles protected H9c2 cardiomyoblasts against H2O2-induced injury by activating the Nrf2/HO-1/NQO1 antioxidant defense axis. This response was associated with reduced intracellular ROS accumulation, decreased caspase-3 activity and improved cell viability, indicating that these vesicles may strengthen endogenous cytoprotective programs under oxidative stress. Although this evidence remains limited to an in vitro injury model, it provides mechanistic support for exploring PDEVs in CHF-related cardiomyocyte damage [79]. In parallel, bitter melon-derived extracellular vesicles (BMDEVs) further support the role of PDEVs in myocardial injury and fibrosis control [191, 192]. BMDEVs reduced mitochondrial ROS accumulation, DNA damage and mitochondrial dysfunction under irradiation stress, thereby limiting cardiomyocyte injury. In vivo, BMDEVs attenuated radiation-induced myocardial fibrosis and cardiac remodeling [193]. These data lend mechanistic weight to a possible role for PDEVs in limiting fibrosis-related myocardial deterioration, although direct CHF models are still needed.

Arrhythmia (ARR)

Cardiac ARR arise from abnormal impulse formation or conduction, often in the setting of structural heart disease, conduction system defects or remodeling after cardiac injury [194]. Reversible triggers, such as electrolyte disturbance, endocrine abnormalities, pulmonary embolism and alcohol exposure, can further induce acute episodes [195]. Current management mainly relies on acute rhythm termination, rate control, rhythm control and catheter ablation. However, these approaches have limited ability to restore the damaged myocardial substrate that drives recurrent ARRs, including cardiomyocyte loss, scar formation, inflammatory remodeling and fibrosis [196].

Ginseng root-derived vesicles reduced doxorubicin-induced myocardial injury in experimental models [197]. Mechanistically, they limited mitochondrial apoptotic signaling, as shown by reduced cytochrome C release, increased Bcl-2 expression and decreased Bax expression [198]. By preserving cardiomyocyte viability, this effect could lessen electrical instability caused by cell loss after myocardial injury. Therefore, ginseng-derived vesicles may be more relevant to ARR through limiting myocardial damage and reducing substrate formation after cardiotoxic stress.

After delivery via fibrin gel, Gouqi-derived nanovesicles (GqDNVs) inhibited p38 mitogen-activated protein kinase (p38 MAPK) and NF-κB p65 signaling, reduced pro-apoptotic molecules and increased the anti-apoptotic/pro-apoptotic ratio, while simultaneously suppressing the expression of transforming growth factor beta 2 (TGF-β2) and fibrosis-related molecules. These changes attenuated cardiomyocyte apoptosis and myocardial fibrosis after infarction [210], thereby reducing pathological substrates associated with reentry and conduction heterogeneity [199]. In addition, the exosome-like nanovesicles (ELNs) from ginger rhizomes (G-ELNs) inhibited the assembly of the NLRP3 inflammasome and reduced the release of IL-1β and IL-18, indicating their potential to lower inflammatory burden [200]. Solanum lycopersicum-derived exosome-like nanovesicles delivered miR164a and miR164b-5p, inhibited Kelch-like ECH-associated protein 1 (KEAP1) and activated the Nrf2 axis, thereby reducing VSMC proliferation and migration after vascular injury. Although this evidence is more distant from ARR, it still connects with ischemia-related remodeling, a process that can increase rhythm vulnerability [48].

Acute myocardial infarction (AMI)

AMI is defined as acute myocardial necrosis caused by acute myocardial ischemia [201]. Most AMI cases result from thrombosis secondary to rupture or erosion of unstable atherosclerotic plaques. Clinically, AMI is commonly classified as ST-segment elevation myocardial infarction or non-ST-segment elevation myocardial infarction according to electrocardiographic and biomarker findings [202–204]. Rare etiologies of myocardial infarction include coronary artery spasm, coronary embolism, and thrombosis in non-atherosclerotic normal blood vessels [205]. The core therapeutic mechanism for AMI is to restore coronary blood flow as soon as possible, inhibit thrombotic or platelet processes, while addressing reversible coronary artery spasm and reducing myocardial oxygen consumption [206, 207]. Reperfusion and antithrombotic therapy have improved acute outcomes, but they do not fully prevent cardiomyocyte death, infarct expansion and later ventricular remodeling [208–210]. Targeted lipidomics linked PUFA-derived oxylipin signatures to recurrent major adverse cardiovascular events after ST-segment elevation myocardial infarction, while a pro-resolving oxylipin combination limited apoptosis, oxidative stress, fibrosis and cytokine activation in a mouse MI/R model [211].

After AMI, cardiomyocyte death and stress-related injury remain major contributors to infarct expansion and impaired functional recovery [212, 213]. Carrot-derived exosome-like nanovesicles (Carex) have been examined as natural vesicles for protecting cardiomyocytes under AMI-related injury conditions. Carex was isolated from carrot juice via centrifugation, ultrafiltration, and molecular sieve chromatography, and was identified as small vesicles with a particle size of approximately 140–150 nm and negative charge [214, 215]. These vesicles exhibited no significant toxicity to cardiac-derived H9c2 rat cardiomyoblast cell line cells and could be efficiently internalized by the cells. In an in vitro model relevant to myocardial infarction, H9c2 cells were treated with H₂O₂ to mimic oxidative stress. Carex reduced intracellular ROS accumulation, preserved cell viability and decreased caspase-3 activity in a dose-dependent manner [79, 216]. These findings indicate that Carex mainly provides preliminary evidence for cardiomyocyte protection rather than direct proof of AMI repair in vivo.

An experimental study investigated gouqi-derived nanovesicles (GqDNVs) for MI repair. GqDNVs were purified by sucrose density-gradient ultracentrifugation and rapidly gelled with fibrinogen to GqDNVs-gel, thereby enhancing local cardiac retention and sustained release. In a mouse MI model induced by permanent LAD ligation, local delivery of GqDNVs-loaded fibrin gel improved 14-day survival and echocardiographic function. It also reduced infarct size and decreased TUNEL-positivity, Bax expression and caspase-3/7 activity. In addition, fibrosis-related markers, including alpha-smooth muscle actin (α-SMA), collagen type I alpha 1 chain (COL1A1), COL3A1 and fibronectin 1, were downregulated. Angiogenesis and proliferation markers were increased in the infarct region. The treatment also lowered serum TGF-β2 and increased insulin-like growth factor 1 (IGF-1) and vascular endothelial growth factor A (VEGF-A), supporting enhanced tissue repair and neovascularization. Mechanistically, transcriptomic analysis and experimental validation linked these benefits to inhibition of p38 MAPK and modulation of NF-κB p65 signaling. Lipid metabolic pathway remodeling was also involved, indicating that GqDNVs may coordinate injury control, angiogenic repair and metabolic adaptation after MI [160]. Compared with cellular oxidative injury models, this study provides more direct support for the role of PDEVs in post-MI remodeling control.

Hypertensive heart disease (HHD)

The pathogenesis of HHD is primarily driven by sustained pressure overload combined with imbalances in humoral and local pro-remodeling molecules, which induce myocardial remodeling including increased cardiomyocyte apoptosis, interstitial and perivascular fibrosis, and structural and functional abnormalities of coronary microcirculation [217, 218]. These pathological changes consequently lead to impaired diastolic function, reduced perfusion reserve, and electrophysiological instability [219]. Current treatment focuses on sustained blood pressure control and prevention of remodeling, especially left ventricular hypertrophy, fibrosis, microvascular injury and cardiomyocyte loss [220, 221]. Mechanistically, therapies aim to rebalance remodeling signals, protect coronary microvessels and preserve myocardial compliance [222]. These effects are also associated with the repair of coronary arteriole structure and improvement of coronary flow reserve [223]. In addition, salt restriction may alleviate myocardial fibrosis and the deterioration of diastolic function not only through blood pressure reduction but also partially independent of blood pressure levels [224], and thus should be regarded as an important component of the comprehensive intervention for HHD.

In spontaneously hypertensive rats, Semen Sinapis albae-derived nanovesicles (SDNVs) were intragastrically administered for 8 consecutive weeks, with nifedipine and sinapine used as controls. This model allowed assessment of blood pressure, vascular remodeling and endothelial senescence [225, 226]. The results demonstrated that SDNVs reduced blood pressure and downregulated related vasoactive factors, while more prominently ameliorating vascular wall thickening and structural disorder, as well as attenuating the endothelial senescence phenotype. In Ang II-injured HUVECs, SDNVs were internalized and improved endothelial cell function. Subsequent detection of indicators including endothelin-1 secretion, NO/ATP levels, cell cycle, and senescence-associated beta-galactosidase staining revealed that SDNVs reversed Ang II-induced elevation of endothelin-1, reduction of NO/ATP levels, improved cell proliferation, and diminished senescent and inflammatory phenotypes [227]. Mechanistically, vesicle-delivered miR393a was linked to CD38 inhibition, which contributed to reduced endothelial senescence. This finding is significant because it connects a defined PDEV cargo with vascular remodeling control, an upstream process in HHD progression.

Using a rat model of intermittent hypoxia-induced hypertension, orally administered Taraxacum officinale (T. officinale)-derived ELNs reduced systolic blood pressure, primarily by restoring intestinal barrier integrity and suppressing systemic inflammation. T. officinale treatment alleviated cecal tissue injury, restored goblet cells and the tight-junction protein Occludin, and decreased inflammatory mediators including IL-1β, IL-6, and tumor necrosis factor alpha (TNF-α). Mechanistically, these effects were associated with recovery of short-chain fatty acid metabolism, with butyrate identified as a key mediator. Butyrate supplementation similarly lowered blood pressure and attenuated thoracic aortic medial thickening, supporting improved vascular remodeling [228, 229]. Collectively, these findings suggest that dandelion-derived vesicles may indirectly slow the progression of HHD by acting upstream to reduce blood pressure, dampen inflammation, and improve vascular remodeling.

Diabetic cardiomyopathy (DCM)

DCM is typically characterized initially by myocardial fibrosis and remodeling, predominantly manifesting as diastolic dysfunction, which subsequently progresses to systolic impairment and ultimately heart failure [230, 231]. Its progression is driven by metabolic disturbance, impaired insulin signaling, mitochondrial dysfunction, advanced glycation end-product accumulation and reduced NO bioavailability. Calcium-handling abnormalities, RAAS activation, autonomic neuropathy, endoplasmic reticulum stress and microvascular dysfunction further aggravate myocardial injury [230, 232, 233]. Mechanism-based therapeutic strategies for DCM first involve lifestyle interventions and long-term stable glycemic control to reduce the risk of hyperglycemia- or insulin resistance-driven myocardial remodeling and heart failure. Second, targeted interventions for vascular and neurohumoral abnormalities can improve microcirculation and myocardial load. Pathway-focused approaches targeting NF-κB, Nrf2 or c-Jun N-terminal kinase (JNK) signaling have also been explored to reduce myocardial injury and fibrotic remodeling [234]. Finally, improving Ca²⁺ homeostasis (including mitochondrial Ca²⁺ handling) and enhancing sarco/endoplasmic reticulum Ca²⁺-ATPase 2 (SERCA2)-mediated calcium cycling are recognized as therapeutic targets worthy of further optimization and translational research [235].

Salvia miltiorrhiza-derived exosome-like nanoparticles (SM-ELNs) were isolated by UF followed by sucrose density-gradient purification, yielding vesicles with a typical vesicular morphology and a size on the 100 nm scale [235]. In a mouse model induced by high-fat diet plus streptozotocin, SM-ELNs improved diabetic myocardial injury after systemic administration. SM-ELNs improved cardiac pump function, as reflected by increased left ventricular ejection fraction and fractional shortening, and alleviated myocardial structural injury and fibrosis. At the mechanistic level, SM-ELNs reduced macrophage-associated pyroptosis signals and NLRP3 inflammasome activity in diabetic myocardium [236, 237]. In vitro validation further focused on macrophage pyroptosis. Treatment with SM-ELNs (100 µg/mL) markedly reduced pyroptosis, assessed by a flow-cytometric assay combining a caspase-1 activity probe with propidium iodide staining [161, 238]. The key significance of this study is that it links SM-ELNs to the NEDD4/serum/glucocorticoid-regulated kinase 1 (SGK1) axis, NLRP3 inflammasome control and macrophage pyroptosis, providing a clearer mechanistic basis for PDEV-mediated protection in DCM.

PDEVs may confer cardioprotection against myocardial injury associated with diabetes complicated by myocardial infarction via antioxidant pathways [239]. Carrot-derived vesicles reduced ROS accumulation and apoptosis in H2O2-injured H9c2 cardiomyoblasts. This effect was accompanied by increased Nrf2, HO-1 and NQO1 expression, linking these vesicles to activation of endogenous cytoprotective programs [240]. Ginger-derived vesicles also promoted Nrf2 nuclear translocation and reduced ROS, adding supportive evidence for this stress-response pathway [241, 242]. However, these findings remain based on cellular injury models and should be distinguished from direct DCM evidence.

Other types

In cardiotoxicity (CT), Beta vulgaris-derived exosome-like nanovesicles (BELNVs) have shown protection against chronic doxorubicin-induced cardiac injury [243]. In vitro, BELNVs can be internalized by HL-1 cardiomyocytes, exhibit no significant intrinsic toxicity, and partially restore the doxorubicin-induced decline in cell viability. In a chronic doxorubicin-induced mouse model, BELNV treatment improved left ventricular ejection fraction and fractional shortening. It also enhanced exercise tolerance, reduced serum lactate dehydrogenase levels and alleviated myocardial histological damage [162]. The key significance of this study is that it links BELNV-mediated cardioprotection with ferroptosis regulation, expanding PDEV-related mechanisms beyond general stress reduction. This mechanism also provides a rationale for stimuli-responsive PDEV design, in which BELNV-based formulations could be engineered to respond to ROS accumulation, lipid peroxidation or ferroptosis-associated cues in chemotherapy-injured myocardium.

Bitter melon-derived extracellular vesicles (BMDEVs) have been investigated in radiation-induced cardiac injury and myocardial fibrosis. In vitro, BMDEVs were taken up by cardiomyocytes and improved cell survival under irradiation stress. They reduced DNA damage, mitochondrial ROS accumulation and mitochondrial dysfunction, thereby limiting radiation-related cardiomyocyte injury. In vivo, in a mouse model of thoracic irradiation, BMDEVs reduced myocardial injury and attenuated fibrosis. This finding highlights mitochondrial preservation and DNA damage control as important mechanisms for PDEV-mediated protection against therapy-related myocardial fibrosis [173].

PDEVs serving as therapeutic agents for CVDs treatment

As summarized in Fig. 5, PDEVs function as intrinsic therapeutic agents and versatile delivery platforms in chronic diseases. As natural secretory vesicles, they encapsulate diverse bioactive components, including proteins, lipids, nucleic acids and small metabolites, enabling modulation of complex pathological processes. They exhibit anti-inflammatory and antioxidant activities, regulate immune responses and promote tissue repair. Notably, growing evidence supports the direct therapeutic potential of PDEVs in cardiovascular diseases.

Fig. 5.

Intrinsic bioactivity and delivery functions of PDEVs. PDEVs act both as therapeutic entities and as nanocarriers, enabling targeted delivery of diverse cargos, with implications for cardiovascular diseases

Redox regulation

PDEVs can regulate redox imbalance through their endogenous bioactive cargoes, thereby reducing ROS-driven mitochondrial damage, lipid peroxidation and cardiomyocyte injury [244]. In a radiation-induced cardiac injury model, Momordica charantia-derived nanovesicles were internalized by H9c2 cardiomyocytes. Under irradiation stress, they improved cell survival, reduced DNA damage and preserved mitochondrial function by limiting mitochondrial ROS accumulation. These effects were accompanied by partial recovery of ROS-related protein phosphorylation, linking vesicle treatment to restoration of stress-response signaling. In mice exposed to thoracic irradiation, these nanovesicles reduced myocardial injury and fibrosis. This finding highlights mitochondrial preservation and DNA damage control as key outcomes of PDEV-mediated protection in therapy-related cardiac injury [245].

Similar redox-regulatory effects have also been observed under hypoxia, reperfusion and other cardiac stress conditions, where PDEVs may restrain ROS-amplified injury and downstream cardiomyocyte death. Carex and blueberry-derived exosome-like nanoparticles (B-ELNs) provide complementary cellular evidence from cardiomyoblast and vascular cell models. In H₂O₂-injured H9c2 cardiomyoblasts, Carex reduced ROS accumulation and caspase-3 activity while maintaining Nrf2/HO-1/NQO1-related cytoprotective signaling [246]. In a TNF-α-stimulated vascular cell model, B-ELNs preserved cell viability and reduced ROS production through modulation of stress-related gene expression [247]. Taken as a whole, PDEV-related redox regulation extends beyond direct ROS reduction and involves mitochondrial preservation, DNA damage control and endogenous cytoprotective pathways.

Immune remodeling

PDEVs can influence immune activation in transplantation- and reperfusion-related cardiovascular injury, particularly by reshaping macrophage behavior and local cytokine signaling [248]. Taking flavonoid-enriched vesicles derived from Exocarpium citri grandis as an example, in an animal model of ischemia-reperfusion injury associated with heart transplantation, a hybrid vesicle system of animal-plant dual origin was constructed by first preparing mesenchymal stem cell membrane-derived nanovesicles, then fusing them with ENVs and loading the immunosuppressive drug rapamycin. The system enriched rapamycin in the transplanted heart and reduced early graft injury. More importantly, it shifted Ly6C⁺Ly6G⁻ macrophages toward a reparative phenotype, thereby weakening immune-cell infiltration and cytokine amplification. Ultimately, it ameliorated myocardial reperfusion injury, prolonged the survival of cardiac grafts, and no significant toxicity was observed [249]. The value of this work lies in combining vesicle-associated biological activity with immunosuppressive cargo delivery, providing a clearer strategy for controlling graft-centered immune injury.

Grape-derived exosome-like nanoparticles (GELNs) further illustrate that PDEVs can regulate host immune repair responses rather than functioning only as carriers [250]. After oral administration, GELNs crossed the intestinal mucus barrier and were taken up by Lgr5 + intestinal stem cells. They activated Wnt/β-catenin signaling and increased stem-cell-associated markers, including axis inhibition protein 2 (AXIN2), Cyclin D1, MYC proto-oncogene (MYC) and epidermal growth factor receptor (EGFR). In a 3% Dextran sulfate sodium-induced colitis model, mice receiving continuous oral administration of GELNs exhibited markedly attenuated weight loss and mortality; the intestinal length and villus height were close to normal, whereas those in the control group were significantly shortened and atrophied. The increase in Lgr5 + Ki67+ epithelial cells links GELN activity to stem-cell-driven mucosal repair rather than simple cytokine suppression. Thus, they are regarded as promising candidates for low-immunogenic nanotherapeutics and key formulations for drug delivery platforms in the treatment of chronic inflammation-related diseases [251].

Vascular inflammation

As supporting evidence from plant-derived bioactive constituents, gingerenone A has been reported to restrain endothelial activation under LPS- or TNF-α-stimulated conditions. In TNF-α- or LPS-stimulated HUVECs, gingerenone A reduced monocyte-endothelial adhesion and lowered VCAM-1 and CCL2 expression. This result places the NF-κB/VCAM-1/CCL2 axis as a useful reference for evaluating how plant-derived vesicles may regulate endothelial activation, although gingerenone A itself should not be presented as direct PDEV evidence [252].

Stronger PDEV-specific evidence comes from vesicle studies that directly evaluate endothelial activation, chemokine signaling and monocyte adhesion in vascular disease models. In studies on AS, Carthamus tinctorius L.-derived nanovesicles (CDNVs) reduced plaque burden in ApoE−/− mice. In ox-LDL-injured HUVECs, CDNVs lowered CXCL12, VCAM-1 and ICAM-1 expression. These changes reduced monocyte adhesion and endothelial infiltration, linking CDNV activity to chemokine-adhesion control rather than simple lipid lowering [171]. Similarly, onion-derived vesicles provide another line of evidence. In endothelial and macrophage models, they reduced IL-1β, TNF-α and IL-6 expression, weakened adhesion-related responses and limited foam-cell formation. From this perspective, PDEVs should be described not simply as anti-inflammatory agents, but as regulators of cytokine release, chemokine gradients, endothelial adhesion and macrophage lipid handling. This mechanism-oriented interpretation better explains their relevance to AS progression and plaque stability [157].

Cell death regulation

Liu and colleagues reported that exosome-like nanovesicles derived from Rhodiola rosea markedly attenuate endothelial injury under ischaemic conditions, with the main mechanistic change centered on endothelial pyroptosis. In a hindlimb ischemia model, these vesicles suppressed the thioredoxin-interacting protein (TXNIP)/NLRP3 inflammasome axis, followed by lower caspase-1 activation and reduced GSDMD-dependent membrane pore formation. This reduced the pyroptotic burden in endothelial cells, dampened local inflammatory amplification, and promoted restoration of blood flow and structural repair in ischaemic tissue. The main value of this study lies in connecting PDEV activity with the TXNIP/NLRP3/caspase-1/GSDMD cascade, shifting the discussion from broad cytokine suppression to endothelial fate control in ischemic vascular injury [253].

In a separate review, Wang and colleagues systematically summarized the biological roles of PDEVs in tissue injury repair and regeneration, emphasizing that their intrinsic activities extend beyond promoting cell proliferation and migration to include regulation of pathological cell death programs. In injured microenvironments, plant vesicles may limit mitochondrial dysfunction and reduce sustained activation of apoptosis- and pyroptosis-related pathways. Such actions can reduce tissue damage and create conditions more permissive for regeneration and repair. Notably, the authors argued that, compared with solely suppressing downstream inflammatory outputs, upstream modulation of cell death signaling by plant vesicles may be more effective in breaking the self reinforcing cycle linking injury driven inflammation with cell death. This framework helps explain why PDEVs may protect vascular and myocardial tissues by regulating cell fate checkpoints, rather than only reducing soluble mediators [254].

PDEVs acting as a drug delivery system for therapeutic purposes

Because of their broad botanical sources, vesicular structure and modifiable membranes, PDEVs have attracted increasing interest as delivery platforms. Current studies have explored their use for loading small molecules, nucleic acids, proteins and natural bioactive compounds, as summarized in Table 4.

Table 4.

Summary of drug-loaded PDEVs: sources, cargos, particle size, loading efficiency, methods and applications

Source	Drug	Particle size (nm)	Loading efficacy	Method	Application	Refs.
Lemon	Doxorubicin	202.2 ± 6.2	18.8 ± 0.6%	Incubation	Ovarian cancer	[256]
Ginger	siRNA-CD98	190.0	61%	Sonication Extrusion	Inflammatory bowel disease	[271]
Cabbage	miR-184 Doxorubicin	100.0	Not reported	Incubation	Anti-tumour	[35]
Ginger	Adriacin	188.0	95.9% ± 0.3%	Sonication	Carcinoma of colon, Chemo-steroid	[257]
Green tea	piRNA-36,741	105.0–125.0	38–42%	Electroporation loading piRNA-36,741	TNBC	[190]
Ginger	siRNA	202.3	50% or more	Ultrasound Incubation	Hereditary, hemochromatosis	[260]
Tomato	HSP70-AF647 protein	140.0 ± 13.0	~ 1.1%	Sonication 35 kHz, 15 min Incubation 90 min	Glioma, Inflammation	[127]
Grapefruit	HSP70-AF647 protein	100.0–120.0	~ 1.1%	Sonication	Inflammatory bowel disease, hepatopathy	[127]
Broccoli	Astaxanthin	191.6 ± 2.2	72 ± 11%	Ultrasound	Carcinoma of colon	[266]
Banana	Curcumin	250.0	90%	Incubation Hydrogel bead encapsulation	Ulcerative colitis	[267]
Gouqi	GqDNVs, Endogenous bioactive components	100.0–150.0	Not reported	Encapsulation	Myocardial infarction	[160]
Cabbage	CELNs, tsRNA	100.0	Not reported	Encapsulation	Prevention and treatment of restenosis after vascular injury	[268]
Acerola	hsa-miR-340	340.0 ± 172.0	60%	Incubation for 30 min on bath ice	Target gene-suppressing effect in the small intestine	[265]
Aloe	Indocyanine green	220.0	Not reported	Not reported	Noninvasive transdermal administration and skin cancer therapy	[272]
Cucumber	DiI	167.0 ± 3.0	Not reported	Incubation Vortexing	Skin therapy	[48]
Orange	mRNA-based vaccines	167.0 ± 10.0	72 ± 11%	Cation-based interaction combined with controlled osmotic shock	COVID-19	[273]

Small-molecule agents

When PDEVs are used for delivering small-molecule drugs, their loading methods can be categorized into passive loading and active loading [255]. Taking the lemon-derived EV system as a typical example, a combined strategy of active loading and membrane hybridization was adopted: lemon-derived extracellular vesicles were ultrasonically fused with homologous tumor cell membrane fragments at 4 °C to construct hybrid vesicles. Fusion validation showed that the hybrid vesicles retained nanoscale morphology and membrane-associated proteins such as E-cadherin, CD44 and CD47. Subsequently, doxorubicin was incorporated by co-sonication, and unencapsulated drug was removed by ultrafiltration to obtain drug-loaded hybrid vesicles. Drug loading and pH-dependent release assays further supported the controllable delivery behavior of this formulation. Results demonstrated that lemon-derived hybrid vesicles exerted the most significant inhibitory effect on 4T1 tumor volume, accompanied by reduced Ki-67 expression and increased TUNEL-positive signals; meanwhile, hematoxylin and eosin staining of major organs showed no obvious histological damage, indicating that lemon-derived hybrid vesicles enhanced delivery efficiency while maintaining favorable safety profiles [256].

When cabbage-derived EVs were incubated with doxorubicin under the same conditions for 4 h, the particle size remained stable at 100 nm. This method features simple operation and preservation of the natural bioactivity of EVs, yet it suffers from low loading efficiency and poor repeatability for hydrophilic drugs [35]. By comparison, active loading improves cargo entry by transiently permeabilizing the vesicle membrane through sonication, electroporation, extrusion or lipid reassembly. For example, sonication-mediated loading of 5-fluorouracil into bitter melon-derived EVs enhanced its antitumor activity. In another formulation, doxorubicin-loaded ginger EVs prepared by lipid extraction, sonication and extrusion showed a particle size of approximately 188 nm and an encapsulation efficiency of 95.9 ± 0.26% [257]. Nevertheless, such methods are prone to causing membrane damage to EVs and loss of endogenous bioactive components, and lipid extraction in particular abrogates the intrinsic bioactivity of EVs. In practical formulation design, passive loading is preferable when preservation of native vesicle activity is important, whereas active loading is better suited for increasing encapsulation efficiency and tuning release behavior. Although most small-molecule loading studies have been conducted outside cardiovascular models, they provide useful design principles for CVD applications, including how to balance encapsulation efficiency, vesicle integrity, release control and preservation of endogenous activity.

Nucleic acid agents

Green tea-derived exosome-like nanovesicles have been used to deliver an antisense oligonucleotide targeting HAAPIR through electroporation [37, 190]. Specifically, pulsed electric currents transiently generated hydrophilic pores in the lipid bilayer membrane, enabling efficient cargo encapsulation [258]. Following purification by differential centrifugation combined with sucrose density-gradient ultracentrifugation, the resulting complexes exhibited resistance to nuclease-mediated degradation and retained integrity under acidic gastric conditions. In vitro, these complexes were readily internalized by human aortic smooth muscle cells, where they markedly attenuated nicotine-induced MMP9 overexpression and suppressed phenotypic switching. In vivo, after oral gavage in mice, the complexes traversed the gastrointestinal tract and preferentially accumulated at aortic lesions. In aortic dissection models induced by Ang II and β-aminopropionitrile, treatment reduced disease incidence, limited vascular dilation and improved survival. These effects were linked to regulation of the MEF2D/MMP9/α-SMA axis. Importantly, no overt toxicity was observed in major organs, and hematological parameters as well as liver enzyme levels remained within normal ranges [190]. Overall, this strategy offers notable advantages, including excellent biocompatibility, robust stability for oral nucleic-acid delivery, and potential synergistic antioxidant effects. Its limitations are also clear. Electroporation requires specialized equipment, and delivery efficiency may still be lower than that of some liposome-based transfection systems. The mechanisms governing intestinal transport, aortic lesion enrichment and cellular entry also need further clarification. For CVD applications, this example moves PDEV-based nucleic acid delivery from simple cargo protection toward oral administration and vascular lesion targeting.

To mitigate the potential toxicity arising from co-isolated impurities, vesicular lipids can be extracted and reconstituted into delivery systems by co-forming a thin lipid film with therapeutic cargos using the thin-film hydration method [259]. Ginger-derived vesicles loaded with CD98-targeting siRNA by sonication reduced macrophage cytotoxicity and apoptosis in vitro, supporting their feasibility for siRNA delivery [51]. Following oral administration, the siRNA-CD98/vesicle complexes were retained within the gastrointestinal tract and significantly downregulated CD98 expression throughout the gastrointestinal system. Nucleic-acid delivery with greater relevance to cardiovascular complications has been explored in iron-overload settings. In a hereditary hemochromatosis mouse model, ginger-derived vesicles were used to deliver divalent metal transporter 1 siRNA, leading to decreased intestinal epithelial divalent metal transporter 1 mRNA expression and reduced iron burden. Non-heme iron levels decreased in multiple organs, including the heart, together with lower ferritin and transferrin saturation [260]. This result links oral siRNA delivery by plant vesicles with systemic iron control and provides indirect support for reducing cardiac complications related to iron overload.

Protein agents

PDEVs can also be used to deliver protein therapeutics, and systematic experiments have demonstrated their capacity for stable loading while preserving protein activity [261]. Recombinant HSP70 has been loaded into tomato- and grapefruit-derived EVs through a two-step loading strategy. First, HSP70 was mixed with the EV suspension and passively incubated overnight at 4 °C to promote initial association with the lipid membrane [71]. Subsequent sonication transiently increased membrane permeability and promoted protein entry into the vesicle lumen. The samples were then further incubated to allow membrane recovery, and unencapsulated free protein was removed by centrifugal filtration [262–264]. In glioma cell assays, HSP70-loaded PDEV preparations showed no evident cytotoxicity and retained the stress-protective function of HSP70. In addition, treated cells displayed increased sensitivity to chemotherapeutic agents, supporting the conclusion that vesicle-associated HSP70 remained biologically active [127].

Garaeva and colleagues further evaluated grapefruit-derived extracellular vesicles as a protein delivery platform by loading recombinant HSP70 into the vesicles. In this system, vesicle encapsulation enabled HSP70 transfer to recipient cells without loss of its immunostimulatory activity. Compared with free HSP70, vesicle-based delivery achieved immune activation at lower doses and produced stronger in vivo antitumor effects. The importance of this work lies in showing that PDEV-mediated protein delivery can preserve cargo function and reduce the dose required for biological activity. For cardiovascular translation, the relevance of these studies remains mainly conceptual because most protein-loading evidence comes from non-cardiovascular models. Even so, they offer useful design principles for future delivery of proteins that modulate vascular repair, myocardial survival or remodeling after cardiac injury [265].

Natural bioactive compounds

Natural bioactive compounds are commonly incorporated into PDEVs through incubation, sonication or freeze-thaw processing, depending on their hydrophobicity, molecular size and membrane affinity [58]. Astaxanthin represents a typical lipophilic natural product that can be incorporated into plant EVs by sonication. After mixing the bioactive compound with the EV suspension, low frequency sonication was applied to transiently perturb membrane integrity, followed by recovery at room temperature to restore membrane structure. After loading, vesicle size remained at about 191.6 ± 2.23 nm and the zeta potential was minus 15.85 ± 0.92 mV, indicating good physicochemical stability [266]. Simple incubation has also been used for plant metabolites with suitable membrane affinity. In these formulations, vesicle diameters generally remained within roughly 95 to 224 nm, without clear signs of rupture or aggregation. Importantly, PDEV-mediated delivery can improve the solubility and intracellular uptake of natural bioactives while reducing apparent toxicity. In cell models, vesicle-loaded natural products generally show stronger intracellular availability and broader cytoprotective responses than free compounds, although direct cardiovascular validation remains limited [53].

Liu and colleagues developed a colon targeted delivery system for natural bioactives using PDEVs. Banana-derived vesicles were used to carry hydrophobic curcumin through co-incubation, allowing the compound to partition into membrane domains and improve aqueous dispersion. The drug-loaded vesicles were then embedded into hydrogel microbeads formed from alginate and banana pectin to protect the formulation in the upper gastrointestinal tract and enable controlled release in the colon. The value of this design lies in coupling vesicle-based loading with hydrogel shielding to achieve sequential gastrointestinal protection and local release [267]. Although the model is gastrointestinal rather than cardiovascular, the formulation logic is useful for oral CVD-oriented delivery, especially when natural compounds require protection before reaching systemic circulation or vascular targets.

Combination delivery

Recent work has combined plant vesicles with hydrogels, RNA cargos or lesion-retention strategies to address specific barriers in cardiovascular therapy. Zhou and colleagues embedded GqDNVs within a fibrin hydrogel to establish a local delivery platform that enhances retention in the infarct region after myocardial infarction and enables sustained release. In a mouse MI model, this formulation improved ventricular function, limited infarct expansion and reduced fibrotic remodeling. Mechanistically, these benefits were linked to inhibition of the p38 MAPK/NF-κB p65 axis [160]. Liu and colleagues reported another design with cardiovascular relevance, in which exosome-like nanovesicles from cabbage were used to carry tsRNA, a small noncoding RNA cargo, thereby combining vesicle transport with RNA regulation. By protecting the RNA cargo and enhancing injured-vessel accumulation, the formulation reduced neointimal hyperplasia in a post-injury restenosis model [268]. These examples shift PDEV delivery from single-carrier loading toward integrated therapeutic design, where vesicles are paired with matrix retention, sustained release or RNA regulation to match specific cardiovascular lesions. Such hydrogel-assisted vesicle retention also provides a practical basis for future in situ PDEV formulations, especially for myocardial infarction, restenosis and fibrotic remodeling where prolonged local exposure is required [269, 270].

Engineered PDEVs as nanocarriers for cardiovascular disease therapy

Native EVs are often constrained in cardiovascular applications by cargo heterogeneity and insufficient cardiac targeting, underscoring the need for engineering approaches to prolong their retention in injured myocardium and to amplify cardioprotective effects. Strategies such as surface functionalization, membrane modification and artificial biomimicry can enhance vesicle stability and targeting efficiency, improve biocompatibility, and reduce off-target effects, thereby playing a critical role in strengthening therapeutic efficacy. Representative engineered PDEV systems, including their engineering strategies, functional modifications, cargos, targets, experimental models and cardiovascular protection, are summarized in Table 5.

Table 5.

Representative studies of engineered PDEVs as nanocarriers for cardiovascular disease therapy

Source of PDEVs	Engineering strategy	Functional modification	Cargo	Target	Model	Mechanisms/Effects	Cardiovascular protection	Refs.
Onion (V-Onex)	Surface functionalization (hydrophobic insertion; chemical conjugation)	VCAM-1-targeting peptide (VHPK) conjugated via DSPE-PEG lipid anchor	Endogenous vesicular cargo	Inflamed vascular endothelium and atherosclerotic lesions	LPS-stimulated HUVECs; ox-LDL-induced foam-cell model	TNF-α, VCAM-1, ICAM-1, monocyte adhesion, lipid accumulation, PPARγ and CD36 ↓; endothelial uptake and V-Onex accumulation ↑	Reducing endothelial inflammation and foam cell formation	[157]
Grapefruit (pEVs)	Surface functionalization (hydrophobic insertion; click chemistry)	R11-3 aptamer conjugated via DSPE-PEG-maleimide lipid anchor	Endogenous bioactive components	Endothelial cells	Human cerebral microvascular endothelial cells	Cytotoxicity not evident and selective endothelial uptake ↑	Enables targeted delivery to inflamed endothelium, atherosclerotic plaques, and ischemic myocardium	[281]
Exocarpium Citri grandis (FNVs@RAPA)	Membrane fusion; ROS-responsive bioorthogonal click chemistry	Fusion with calreticulin-externalized MSC membrane vesicles; DBCO conjugation for ROS-responsive azide-click anchoring	Rapamycin	Inflammatory macrophages in transplanted heart	Murine heart-transplant IRI model	ROS, apoptosis, Ccl3, Tlr4, Ly6C+ and inflammatory macrophages ↓; cardiac graft accumulation, inflammatory macrophage uptake and graft survival ↑	Reduces myocardial damage and fibrosis, prolongs graft survival	[182]
Grapefruit (ESTP)	Biomimetic design; surface functionalization	Hydroxyapatite-binding targeting peptide conjugated via lipid anchor	Sodium thiosulfate	Calcified vascular lesions	Calcification-induced VSMCs; cholecalciferol-induced vascular calcification mouse model	Calcium deposition, apoptosis and inflammatory cytokines ↓; cellular uptake, calcified-aorta accumulation and survival ↑	Reduces arterial calcification and stiffness, preserves vascular compliance	[287]
Gouqi (GqDNVs)	Local hydrogel-based delivery and in situ retention	GqDNVs were embedded in fibrin hydrogel	Endogenous vesicular cargo	Infarcted myocardium	Mouse myocardial infarction model induced by permanent LAD ligation	Infarct size, cardiomyocyte apoptosis, p-p38 and p-p65 ↓; cell viability, survival, LVEF, LVFS and angiogenesis ↑	Reduces infarct size, inhibits fibrosis, promotes angiogenesis, improves cardiac function	[160]
Red cabbage (GEP-NPs)	Membrane fusion (freeze-thaw method)	Collagen type I-targeting peptide via liposome fusion	Endogenous cargo	Collagen-deposited fibrotic myocardium post-MI	Mouse MI model	Fibrosis, ventricular remodeling, chronic inflammation and PI3K-AKT-mTOR pathway activation ↓; damaged-heart accumulation and tissue repair ↑	Targets collagen-deposited fibrotic regions, inhibits ventricular remodeling, and prevents heart failure progression	[297]
Cabbage (CELNs)	Endogenous loading (natural encapsulation)	CELNs encapsulating small non-coding tsRNA	tRF-Trp-TCA	Injured blood vessels	Post-injury arterial restenosis model	Neointimal hyperplasia ↓; nucleic acid stability and injured-vessel accumulation ↑	Acts on vascular injury-induced restenosis, inhibiting neointimal formation and supporting vascular repair.	[268]
Coptis chinensis (CLENs)	Endogenous loading (natural encapsulation)	Utilizes intrinsic properties of CLENs for siRNA protection and delivery	CA1-siRNA	Aortic lesions	Endothelial cells; macrophages; in-stent restenosis animal model	CA1, PADI2, MCP-1, VCAM-1 neointimal formation and vascular inflammation ↓; CD31 expression, endothelial repair and vascular lesion accumulation ↑	Acts on ISR-related endothelial injury and vascular inflammation, promoting re-endothelialization and reducing restenosis after vascular injury.	[298]

Surface functionalization

Surface functionalization is among the most direct strategies for improving the cardiovascular performance of PDEVs, because native vesicles often lack defined lesion tropism and may therefore show limited accumulation in inflamed endothelium, atherosclerotic plaques or other pathological vascular sites [274–276]. By programming the vesicle interface with affinity ligands while preserving particle size, morphology and colloidal stability, membrane engineering can enhance selective adhesion and cellular internalization [277, 278]. In practice, this is commonly achieved through hydrophobic insertion of functional lipids into the vesicular bilayer, followed by covalent conjugation of peptides or aptamers to present targeting cues on the surface [279, 280].

Choi and Rhee developed onion-derived extracellular vesicles (Onex) as a vascular inflammation-targeted vesicle system. They first obtained a high-concentration, high-purity vesicle preparation by combining mechanical stirring and UF with SEC, and subsequently introduced vascular inflammation targeting via membrane engineering. The VCAM-1-targeting peptide VHPK was conjugated to a 1,2-distearoyl-sn-glycero-3-phosphoethanolamine-polyethylene glycol (DSPE-PEG) lipid anchor and inserted into the Onex membrane by hydrophobic insertion to generate V-Onex. After engineering, V-Onex maintained an average diameter of approximately 160 nm, remained stable under serum-containing conditions and showed no apparent cytotoxicity toward HUVECs or THP-1 cells [169]. These results indicate that peptide insertion did not markedly disrupt the nanoscale vesicular structure or short-term cellular compatibility of Onex [157].

Moon et al. established a scalable surface-functionalization strategy using grapefruit-derived extracellular vesicles (pEVs) as a modular carrier platform (Fig. 6). Figure 6 provides more than a schematic workflow; it summarizes the key evidence supporting the feasibility of this engineering strategy. Specifically, Fig. 6a shows the preparation route of surface-modified plant-derived vesicles, including vesicle isolation, hydrophobic lipid insertion and ligand conjugation, thereby illustrating how native grapefruit vesicles can be converted into targetable delivery carriers. Figure 6b presents the physicochemical characterization of the isolated pEVs, supporting that the vesicles retained nanoscale morphology and basic structural features after preparation. Figure 6c further shows cellular uptake and viability results, indicating that surface modification enhanced endothelial cell interaction without causing obvious cytotoxicity.

Fig. 6.

Surface-functionalizable grapefruit-derived extracellular vesicles. (a) Schematic illustration of the synthesis and treatment of surface-modified plant-derived extracellular vesicles for a targeting delivery carrier and a drug loading carrier. (b) The characterization of pEVs isolated from grapefruit. (c) Cellular uptake of pEVs and cell viability analysis. Reproduced with permission from [281]

Briefly, the functional lipid DSPE-PEG(2000)-maleimide was first incorporated into the vesicular lipid bilayer via hydrophobic insertion, and a 3′-thiolated R11-3 aptamer was subsequently conjugated onto the vesicle surface through a maleimide-thiol click reaction, thereby conferring endothelial cell-targeting capability. In vitro, aptamer-decorated pEVs showed no appreciable cytotoxicity across the tested dose range and enhanced uptake by human cerebral microvascular endothelial cells. The significance of this result is that ligand display improved endothelial recognition while preserving the basic cellular safety of plant-derived vesicles. Notably, this work was primarily oriented toward blood-brain barrier endothelial targeting and did not directly evaluate therapeutic efficacy in cardiovascular disease models. Nevertheless, because endothelial activation and endothelial injury are central features of many CVDs, this hydrophobic insertion and chemical conjugation framework provides a transferable strategy for future cardiovascular-targeted PDEVs delivery, especially for inflamed endothelium, atherosclerotic lesions and ischemic vascular beds [281].

Membrane fusion

Native PDEVs often show limited lesion-specific targeting and insufficient control over immune-cell interactions in complex cardiovascular microenvironments [104]. Membrane fusion technology, a widely used engineering modification strategy, addresses this limitation [282, 283]. Membrane fusion allows plant vesicles to integrate exogenous membrane components, thereby introducing targeting ligands, immune-regulatory signals or responsive anchoring modules that are absent from native PDEVs. Common physical and chemical fusion methods include extrusion, freeze-thaw cycling, PEG-mediated fusion and pH-assisted fusion. These approaches promote membrane mixing and recombination between different vesicular systems, enabling the incorporation of diverse lipids, membrane proteins and bioactive molecules into engineered PDEVs [284–287].

Lu et al. developed biomimetic hybrid nanovesicles for treating IRI in transplanted hearts. Plant-derived ENVs were first isolated from the exocarp of Shatian pomelo, and then fused with mammalian cell membrane vesicles engineered to express calreticulin to generate fusion nanovesicles. Rapamycin was subsequently loaded into the FNVs by electroporation, yielding FNVs@RAPA. To achieve precise enrichment at the graft site, a ROS-responsive metabolic glycoengineering strategy was used to expose azide groups specifically within the high-ROS microenvironment of the transplanted tissue. The vesicle surface was further functionalized with DBCO, enabling local anchoring through an azide-DBCO click reaction. In a murine heart-transplant IRI model, this system reduced pro-inflammatory Ly6C⁺Ly6G⁻ macrophages, increased reparative Ly6C⁻Ly6G⁻ macrophages, downregulated inflammatory signals such as IL-6 and Cxcl2, and decreased CD8⁺ T-cell infiltration, thereby alleviating inflammation and fibrosis and significantly prolonging graft survival [182].

The significance of membrane fusion engineering lies in reprogramming the vesicular interface of plant-derived nanovesicles from a relatively uniform membrane into a designable functional surface. This enables preservation of their intrinsic protective activities while introducing more defined tissue localization and immunomodulatory capabilities, thereby enhancing lesion enrichment and therapeutic efficiency. Beyond improved delivery, the key advantage of this strategy is its capacity to steer immune cell states toward a pro-resolving phenotype, reducing inflammatory injury and ultimately improving cardiovascular outcomes. Future progress will hinge on controllable and reproducible fusion processes to ensure robust in vivo performance and support translation.

Biomimetic designs

Natural product-based delivery research is increasingly moving toward multi-omics-guided and biomimetic design, providing a useful conceptual basis for engineered PDEVs [288]. A central challenge for PDEVs in cardiovascular therapy is that their native surfaces are not designed to recognize specific pathological cues within cardiovascular lesions, such as mineral deposition, exposed extracellular matrix or fibrotic remodeling [289–291]. Biomimetic design addresses this gap by matching engineered plant vesicles with disease-specific microenvironmental features, thereby improving selective adhesion, uptake and local retention [292–294]. In this context, membrane-level functionalisation with lesion-affinitive ligands provides a straightforward route to convert broadly biocompatible carriers into precision nanotherapeutics [295, 296].

Feng et al. developed a grapefruit-derived extracellular vesicle system for vascular calcification by exploiting hydroxyapatite deposition as a lesion-specific target. Sodium thiosulfate was first incubated with vesicles and then purified by centrifugation to remove unencapsulated drug, yielding sodium thiosulfate-loaded vesicles. A hydroxyapatite-binding peptide was then conjugated to a lipid anchor and inserted into the vesicle membrane to generate hydroxyapatite-targeting peptide-modified sodium thiosulfate-loaded vesicles (ESTP). In a calcification-induced vascular smooth muscle cell model, ESTP showed enhanced cellular uptake and more effectively reduced calcium deposition, oxidative injury and apoptosis. These results suggest that hydroxyapatite-targeted modification improved vesicle interaction with calcified vascular cells rather than merely increasing nonspecific uptake. In vivo, following administration in a cholecalciferol-induced murine vascular calcification model, fluorescence tracking indicated preferential accumulation of ESTP in calcified aortas, with reduced nonspecific retention in major organs and prolonged in vivo residence of the cargo. Therapeutically, ESTP decreased aortic calcification burden, reduced apoptosis, improved survival, and was associated with lower inflammatory cytokine levels and a more pro-repair immune phenotype. Although the primary benefit was directed toward alleviating vascular calcification rather than directly rescuing myocardial function, the improved survival together with reduced off-target exposure and tissue-injury risk in organs such as the heart suggests potential indirect cardioprotection and a lowered risk of CVDs at the whole-system level [287].

The key feature of biomimetic design is disease-microenvironment matching. Unlike general surface functionalization, this strategy uses pathological cues such as hydroxyapatite deposition in vascular calcification or collagen accumulation in post-MI fibrosis to guide vesicle localization. This logic is particularly relevant to CVDs, because many cardiovascular lesions are spatially restricted and structurally defined. Future biomimetic PDEVs should therefore be evaluated by whether microenvironment matching improves lesion retention and functional outcomes, rather than by cellular uptake alone.

Challenges and future opportunities of PDEVs for cardiovascular therapy

Despite the burgeoning preclinical evidence highlighting the therapeutic potential of PDEVs in cardiovascular diseases, their journey from bench to bedside is still in its infancy. Several critical hurdles spanning clinical evidence, biopharmaceutical characterization, manufacturing standardization, and long-term safety must be systematically addressed to enable successful clinical translation.

Current clinical evidence of PDEVs

Clinical evidence for PDEVs remains at an early exploratory stage. Although several human studies and registered clinical trials have evaluated plant vesicle preparations, few have directly targeted established CVDs, such as AMI, CHF, MI/R injury, or AS [64, 137]. Therefore, current clinical evidence mainly supports translational feasibility rather than definitive cardiovascular efficacy.

Among the currently available studies, lemon-derived vesicle preparations provide the most cardiovascular-relevant evidence. A pilot open-label study evaluated a natural supplement containing extracellular vesicles from lemon juice in healthy volunteers who received 1000 mg/day for three months. The study assessed safety and modifiable cardiometabolic risk parameters, and the reported changes in waist circumference and low-density lipoprotein cholesterol suggest that supplements containing plant vesicles may influence intermediate cardiovascular risk markers. However, this study was small, non-randomized, and not designed to assess cardiovascular outcomes [299]. In addition, NCT04698447 was designed as a randomized, double-blind, placebo-controlled trial to investigate a product containing lemon-derived nanovesicles in subjects with metabolic syndrome and healthy volunteers, with planned evaluation of cardiometabolic parameters over 1, 3, and 6 months.

Other clinical investigations involving plant vesicles have mainly focused on inflammatory bowel disease, polycystic ovary syndrome, oral mucositis, or curcumin delivery in colon cancer. These include ginger-derived vesicles for inflammatory bowel disease (NCT04879810), ginger or aloe-derived vesicles for polycystic ovary syndrome (NCT03493984), grape-derived exosomes for oral mucositis caused by chemoradiotherapy in head and neck cancer (NCT01668849), and plant exosomes for curcumin delivery in colon cancer (NCT01294072). Although these studies provide preliminary translational evidence for feasibility, oral administration, tolerability, and biomarker-based evaluation, their relevance to CVDs remains indirect, as summarized in Table 6. Future clinical studies in CVDs should therefore include disease-relevant outcomes, such as endothelial function, vascular inflammation, myocardial injury, cardiac remodeling, plaque stability, exercise capacity, and major adverse cardiovascular events [300, 301].

Table 6.

Clinical studies of PDEVs, including vesicle type, phase/status, associated indications, core functions, and NCT

Vesicle Type	Associated indication	Study details	Last update posted	Status	Core function	Study type	Phase	Study results
Lemon	Cardiometabolic and CVDs	(NCT04698447) The Role of a Natural Product, Containing Nanovesicles From Citrus Limon (L.) Juice, on Different CV Risk Factors	2021/01/06	Unknown	Dietary Supplement: Natural supplement containing nanovesicles delivered from Citrus Limon (L.) juice_MetS Dietary Supplement: Placebo	Interventional	Phase I/II	No
Ginger	Inflammatory bowel disease	(NCT04879810) Plant Exosomes +/- Curcumin to Abrogate Symptoms of Inflammatory Bowel Disease	2022/11/03	Completed	Procedure: Sigmoidoscopy and biopsy, blood work	Interventional	Phase I	No
Ginger/Aloe	Polycystic ovary syndrome	(NCT03493984) Plant Exosomes and Patients Diagnosed With Polycystic Ovary Syndrome (Polycystic ovary syndrome) 17	2021/03/16	Withdrawn	Other: Ginger exosomes Other: Aloe exosomes Other: Placebo	Interventional	Not applicable	No
Grape	Chemoradiotherapy-induced oral mucositis in head and neck cancer	(NCT01668849) Edible Plant Exosome Ability to Prevent Oral Mucositis Associated With Chemoradiation Treatment of Head and Neck Cancer	2022/08/09	Completed	Dietary Supplement: Grape extract Drug: Lortab, Fentanyl patch, mouthwash	Interventional	Phase I	No
Fruit	Colorectal cancer	(NCT01294072) Study Investigating the Ability of Plant Exosomes to Deliver Curcumin to Normal and Colon Cancer Tissue	2023/12/20	Recruiting	Dietary Supplement: curcumin Dietary Supplement: Curcumin conjugated with plant exosomes Other: No intervention	Interventional	Not applicable	No

Pharmacokinetics, biodistribution and cardiovascular targeting

The in vivo fate of PDEVs remains insufficiently defined. Orally administered Carthamus tinctorius L.-derived nanovesicles (CDNVs) can cross the gastrointestinal barrier, enter the bloodstream and distribute to the liver, lungs, heart and aorta; in the same ApoE−/− mouse model, CDNV treatment reduced atherosclerotic plaque burden [171]. This finding supports the feasibility of oral systemic exposure, but it does not provide absolute bioavailability, Cmax, Tmax, AUC, clearance rate or dose proportionality. For intravenously administered extracellular vesicles, rapid capture by clearance organs such as the liver and spleen remains a major limitation, and formulation-specific pharmacokinetic data for PDEVs are still scarce [64, 302]. Future studies should therefore quantify absorption efficiency, plasma exposure, organ clearance and dose-dependent biodistribution, rather than relying only on endpoint fluorescence imaging [284].

Cardiovascular targeting is another unresolved issue. No universal ideal parameter has been established for PDEV formulations, but several design attributes may guide cardiovascular delivery. Vesicles within approximately 50–200 nm may be more suitable for systemic transport, whereas very small particles are prone to renal clearance and larger particles are more likely to accumulate in the liver, spleen or mononuclear phagocyte system [302]. A narrow size distribution, preferably with a polydispersity index below 0.3, may improve batch consistency and biodistribution predictability. A moderately negative zeta potential, such as approximately −10 to −30 mV, may help maintain colloidal stability. These parameters should not be treated as fixed release standards at this stage, but as design references that need to be linked with lesion-to-organ distribution ratios, target-site retention and therapeutic exposure in cardiovascular tissues [284, 302].

Engineered PDEVs provide useful models for improving lesion selectivity. VCAM-1 peptide-modified onion vesicles enhanced interaction with activated endothelium in atherosclerosis-related models, while hydroxyapatite-targeted grapefruit vesicles preferentially accumulated in calcified aortas and reduced nonspecific organ retention [157, 287]. These examples suggest that future targeting studies should report not only cellular uptake, but also lesion-to-organ distribution ratios, retention duration, target-site cargo exposure and functional improvement in diseased cardiovascular tissues. A practical solution is to establish a tiered biodistribution framework: formulation attributes should first be linked to circulation and tissue exposure; targeting should then be tested in disease-relevant models such as AS plaques, ischemic myocardium, injured vessels or calcified lesions; finally, biodistribution data should be connected with therapeutic endpoints, including plaque burden, infarct size, vascular repair, fibrosis regression or cardiac function [8, 284].

Regulatory, manufacturing and quality control challenges

The regulatory pathway for PDEVs remains unclear because these vesicles occupy an intermediate position between botanical products, natural nanocarriers and engineered drug delivery systems [64, 138]. Native edible plant vesicles, drug-loaded PDEVs and ligand-modified PDEVs should not be evaluated under identical criteria, because their active components, exposure routes and safety risks differ. A practical solution is to establish source-to-product traceability, including plant origin, tissue part, cultivation condition, harvest time, pretreatment, extraction, purification and storage records [255]. This would help convert batch-to-batch variability in particle distribution, surface charge and bioactive cargo composition into measurable and controllable quality attributes.

Manufacturing remains a key obstacle to clinical translation. Current PDEV preparation still relies mainly on laboratory-scale procedures, including grinding, filtration, differential centrifugation, ultracentrifugation, ultrafiltration, size-exclusion chromatography and density-gradient purification [96]. These steps can affect vesicle recovery, purity and structural integrity. The heterogeneity summarized in Table 1 illustrates this problem: particle sizes differ markedly among sources, from approximately 69.0 nm for ginseng-derived vesicles to 196.4 ± 45.4 nm for mulberry bark-derived vesicles, while zeta potential ranges from near-neutral values such as −0.6 mV to more negative values such as −25.7 mV. Therefore, scale-up should be based on standardized operating procedures, in-process controls and reference batches, rather than simply increasing processing volume [284].

Quality control should integrate physicochemical characterization, impurity testing and function-related potency assays. Basic release criteria should include particle size, polydispersity index, zeta potential, morphology, vesicle purity, sterility, storage stability and residual impurities [96]. For loaded or engineered PDEVs, cargo loading, release behavior, ligand stability, targeting ability and vesicle integrity should also be assessed. As shown in Table 4, loading efficacy varies greatly among formulations, including 18.8 ± 0.6% for doxorubicin-loaded lemon vesicles, 90% for curcumin-loaded banana vesicles and approximately 1.1% for HSP70-loaded tomato or grapefruit vesicles. These differences indicate that a single universal assay is insufficient. Future studies should build product-specific potency tests according to intended function, such as endothelial protection, macrophage regulation, cardiomyocyte protection, redox injury reduction or lesion-targeted delivery [284, 303].

Safety, immunogenicity and long-term biocompatibility

Although PDEVs from edible plants are generally considered well tolerated, this should not be treated as a class-level safety conclusion [32, 256]. Current evidence is still mainly derived from short-term cell experiments, oral administration models and non-cardiovascular disease studies. For example, orally administered ginger-derived nanoparticles showed protective effects in a mouse model of alcohol-induced liver injury without obvious toxicity signals, supporting the short-term tolerability of some edible plant vesicles. However, these findings cannot be directly generalized to CVD therapy, where repeated dosing, higher systemic exposure or local retention in injured tissues may be required [72]. Therefore, safety evaluation should consider the vesicle source, dose, administration route and formulation type, rather than relying on general claims of biocompatibility.

Immunogenicity should be assessed according to administration route and engineering degree. Oral PDEVs mainly interact with the gastrointestinal barrier and mucosal immune system, whereas intravenously administered or locally retained PDEVs may directly contact complement proteins, circulating leukocytes and clearance organs [284]. In addition, loaded cargos, targeting ligands, hybrid membranes and chemical linkers may introduce new safety variables beyond those of native vesicles. Future studies should therefore include cytokine release, complement activation, anti-PDEV antibody formation, leukocyte activation, hemolysis, coagulation indices, liver and kidney function, and histopathology of major clearance organs [303].

Long-term biocompatibility is especially relevant to cardiovascular applications because many patients may require chronic or repeated treatment. Future studies should move from short-term toxicity observation to a tiered nonclinical safety framework, including single-dose toxicity, repeated-dose toxicity, dose escalation, recovery after treatment withdrawal, biodistribution, organ accumulation, cargo persistence and no-observed-adverse-effect levels [96, 303]. For engineered PDEVs, safety testing should further confirm that targeting modification or sustained retention improves therapeutic exposure without increasing off-target accumulation. This product-specific framework would provide a stronger basis for clinical translation of PDEVs in CVDs.

Conclusions and perspectives

Although current pharmacological strategies for cardiovascular disease have improved outcomes to some extent, their benefits often derive from modulation of a single target or a limited set of pathways, and long-term administration is frequently constrained by clinically meaningful adverse effects. As a result, these approaches remain insufficient to comprehensively address the highly complex, multifactorial pathobiology that drives cardiovascular disease progression. Collectively, these considerations position PDEVs as a naturally occurring nanoscale bioplatform that offers a distinct and emerging paradigm for cardiovascular prevention and therapy. PDEVs can be regarded as bioactive vesicular systems with both intrinsic therapeutic activity and delivery capacity. Their plant-derived membrane structure and endogenous cargos allow them to modulate key processes involved in cardiovascular injury and remodeling, including oxidative stress, immune imbalance, endothelial dysfunction, mitochondrial damage and fibrotic remodeling. Available evidence indicates that these effects are reflected in improved cardiomyocyte survival, reduced vascular inflammation, enhanced endothelial repair and attenuation of adverse myocardial or vascular remodeling across multiple preclinical CVD models. At the same time, their capacity for cargo loading and membrane engineering supports more targeted and controllable delivery of small molecules, nucleic acids, proteins and natural bioactive compounds. Engineered PDEVs further extend this potential by improving lesion-specific accumulation, local retention and functional controllability in diseased cardiovascular tissues. Current preclinical evidence therefore suggests that PDEVs may provide a useful bridge between natural vesicle-based therapy and engineered cardiovascular nanomedicine. However, most findings remain at the experimental stage and still need stronger disease-specific validation.

Notwithstanding the encouraging safety profile and pronounced cardioprotective effects of PDEVs reported in mechanistic studies and animal models, their clinical translation remains hindered by multiple challenges, particularly the paucity of robust clinical evidence. To date, published clinical explorations have largely focused on safety assessments and metabolic or inflammation-related surrogates, whereas systematic evaluation of cardiovascular specific endpoints remains scarce. Future studies should move from general feasibility assessment toward cardiovascular-oriented translational research. Priority should be given to standardized preparation, reproducible quality control, clear potency assays, pharmacokinetic and biodistribution studies, and systematic safety evaluation under different routes, doses and engineering designs. Clinical studies should further include meaningful cardiovascular endpoints, such as endothelial function, plaque stability, myocardial injury, cardiac remodeling, exercise capacity and major adverse cardiovascular events. With continued refinement in manufacturing, targeting design and mechanism-linked evaluation, PDEVs may gradually move from experimental vesicular systems toward clinically relevant therapeutic and delivery platforms for cardiovascular disease prevention and treatment.

Abbreviations

PDEVs

Plant-derived extracellular vesicles

CVDs

Cardiovascular diseases

EVs

Extracellular vesicles

RAAS

Renin-angiotensin-aldosterone system

DC

Differential centrifugation

UF

Ultrafiltration

TFF

Tangential flow filtration

SEC

Size-exclusion chromatography

UC

Ultracentrifugation

SDG-UC

Sucrose density gradient ultracentrifugation

NTA

Nanoparticle tracking analysis

DLS

Dynamic light scattering

TEM

Transmission electron microscopy

ELS

Electrophoretic light scattering

PEG

Polyethylene glycol

piRNA

PIWI-interacting RNA

tsRNA

Transfer RNA-derived small RNA

AS

Atherosclerosis

MI/R

Myocardial ischemia-reperfusion

IRI

Ischemia-reperfusion injury

CHF

Chronic heart failure

AMI

Acute myocardial infarction

HHD

Hypertensive heart disease

DCM

Diabetic cardiomyopathy

Nrf2

Nuclear factor erythroid 2-related factor 2

HO-1

Heme oxygenase 1

NQO1

NAD(P)H quinone dehydrogenase 1

NF-κB

Nuclear factor kappa B

NLRP3

NOD-like receptor family pyrin domain-containing 3

MAPK

Mitogen-activated protein kinase

VCAM-1

Vascular cell adhesion molecule 1

ICAM-1

Intercellular adhesion molecule 1

CXCL12

C-X-C motif chemokine ligand 12

Cxcl2

C-X-C motif chemokine ligand 2

IL-1β

Interleukin-1 beta

TNF-α

Tumor necrosis factor alpha

LPS

Lipopolysaccharide

ox-LDL

Oxidized low-density lipoprotein

HUVECs

Human umbilical vein endothelial cells

HSP70

Heat shock protein 70

HAAPIR

Heart-apoptosis-associated PIWI-interacting RNA

MMP9

Matrix metalloproteinase 9

MEF2D

Myocyte enhancer factor 2D

Onex

Onion-derived extracellular vesicles

VHPK

VCAM-1-targeting peptide

V-Onex

VHPK-engineered onion-derived extracellular vesicles

CDNVs

Carthamus tinctorius L.-derived nanovesicles

MFEVLPs

Mori fructus-derived extracellular vesicle-like nanoparticles

DDNs

Salvia miltiorrhiza-derived exosome-like nanoparticles

ENVs

Exosome-like nanovesicles

FNVs@RAPA

Rapamycin-loaded fusion nanovesicles

GqDNVs

Gouqi-derived nanovesicles

G-ELNs

Exosome-like nanovesicles from ginger rhizomes

SM-ELNs

Salvia miltiorrhiza-derived exosome-like nanoparticles

SDNVs

Semen Sinapis albae-derived nanovesicles

BMDEVs

Bitter melon-derived extracellular vesicles

BELNVs

Beta vulgaris-derived exosome-like nanovesicles

B-ELNs

Blueberry-derived exosome-like nanoparticles

GELNs

Grape-derived exosome-like nanoparticles

Carex

Carrot-derived exosome-like nanovesicles

pEVs

Grapefruit-derived plant extracellular vesicles

ESTP

Hydroxyapatite-targeting peptide-modified sodium thiosulfate-loaded vesicles

GEP-NPs

Red cabbage-derived engineered plant exosome-like nanoparticles

CELNs

Cabbage exosome-like nanovesicles

CLENs

Coptis chinensis-derived exosome-like nanovesicles

DSPE-PEG

1,2-distearoyl-sn-glycero-3-phosphoethanolamine-polyethylene glycol

DBCO

Dibenzocyclooctyne

CA1-siRNA

Carbonic anhydrase 1 small interfering RNA

PADI2

Protein arginine deiminase 2

ISR

In-stent restenosis

ACC

Acetyl-CoA carboxylase

AXIN2

Axis inhibition protein 2

EGFR

Epidermal growth factor receptor

ET-1

Endothelin-1

FAS

Fatty acid synthase

GPX4

Glutathione peroxidase 4

IGF-1

Insulin-like growth factor 1

JNK

c-Jun N-terminal kinase

KEAP1

Kelch-like ECH-associated protein 1

MCP-1

Monocyte chemoattractant protein 1

MYC

MYC proto-oncogene

P21

Cyclin-dependent kinase inhibitor p21

PARP

Poly(ADP-ribose) polymerase

PCNA

Proliferating cell nuclear antigen

PGC-1α

Peroxisome proliferator-activated receptor gamma coactivator 1 alpha

PI3K-AKT-mTOR

Phosphoinositide 3-kinase-protein kinase B-mechanistic target of rapamycin

SERCA2

Sarco/endoplasmic reticulum Ca2+-ATPase 2

SGK1

Serum/glucocorticoid-regulated kinase 1

SIRT1

Sirtuin 1

SOD

Superoxide dismutase

SREBP1

Sterol regulatory element-binding protein 1

TGF-β2

Transforming growth factor beta 2

TXNIP

Thioredoxin-interacting protein

VEGF-A

Vascular endothelial growth factor A

xCT

Cystine/glutamate antiporter

COL1A1

Collagen type I alpha 1 chain

α-SMA

Alpha-smooth muscle actin

CD31

Cluster of differentiation 31

Author contributions

**Qian Li** : Writing - original draft, Writing - review & editing, Visualization. **Zhen Wang** : Searching and Analyzing - literature results, Visualization. **Qianqian Huang** : Formal analysis, Visualization. **Nanbo Zheng** : Formal analysis, Funding acquisition. **Bingtao Zhai, Dongyan Guo** : Formal analysis, Methodology, Visualization. **Junbo Zou** : Reviewing, Conceptualization, Supervision, Funding acquisition. **Yajun Shi** : Reviewing, Methodology, Visualization. **Fei Luan** : Reviewing, Revising, Conceptualization, Supervision, Funding acquisition.

Funding

This study was financially supported by the Project of Shaanxi Administration of Traditional Chinese Medicine (No. SZY-KJCYC-2025-JC-043), the Basic Medicine Research Innovation Center for Cardiometabolic Diseases, Ministry of Education Open Projects Fund (No. xnykdxcxzx-2024-06), the Key Scientific Research Project Program of Shaanxi Provincial Education Department (No. 25JR059), the Science and Technology Innovative Talent Program of Shaanxi University of Chinese Medicine (No. 2024-CXTD-03), the Science and Technology Innovative Team Project of Shaanxi Administration of Traditional Chinese Medicine (No. 2025-CXTD-02), and the Shaanxi Province Postdoctoral Research Funding Project (No. 2025BSHSDZZ390).

Data availability

Clinical trial information was obtained from ClinicalTrials.gov under the corresponding NCT registration numbers and is available at the following URL: https://clinicaltrials.gov/.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.