Plant-Derived Exosome-Like Nanoparticles: Biological Functions, Therapeutic Applications, and Translational Challenges

Abstract

Plant-derived herbal medicines remain important sources of bioactive compounds, but their clinical application is often hindered by poor bioavailability and variable pharmacokinetic profiles. Recent studies have identified plant-derived exosome-like nanoparticles (PELNs) as naturally occurring bioactive nanoparticle preparations containing proteins, lipids, small RNAs, and secondary metabolites. This review examines advances in PELN isolation, biological properties, and their therapeutic applications, with particular emphasis on immune regulation, gut-microbiota modulation, and bone homeostasis. We discuss how isolation and purification approaches affect particle characteristics and the interpretation of biological outcomes. Preclinical studies suggest that PELNs from selected plant sources may modulate osteoblast and osteoclast activities, potentially contributing to bone-remodeling balance. Certain PELN preparations have shown promising characteristics for scalable production, gastrointestinal stability, and oral administration in experimental settings, but these potential advantages are source- and method-dependent and have not been established through systematic comparisons with mammalian exosomes. Interpretation of reported bioactivity is further complicated by inconsistent particle characterization, uncertain vesicular identity and purity, and limited evidence distinguishing vesicle-associated effects from those of co-isolated compounds. Further standardization of isolation, characterization, and dosing is essential for evaluating the clinical potential of PELNs. Pharmacokinetics, biodistribution, and long-term safety also remain insufficiently investigated. Overall, PELNs offer a promising area of plant-based nanomedicine research, but their clinical utility remains to be demonstrated.

Graphical Abstract

Introduction

Plants and herbal medicines have long provided important sources of bioactive compounds for drug discovery and therapeutic development.¹ Their multicomponent nature has attracted interest in the management of complex diseases, including inflammatory, metabolic, and neurodegenerative disorders.^2–5 However, challenges such as limited bioavailability, variable pharmacokinetics, and inconsistent therapeutic outcomes can restrict their clinical applications.⁶ These limitations have encouraged interest in naturally occurring nanoscale systems that may contribute to the transport and biological activity of plant-derived compounds.⁷

In parallel with the long-standing herbal medicine field, recent advances in cellular biology have highlighted the critical role of EVs, particularly exosomes, as mediators of intercellular communication.⁸ Exosomes are nanoscale vesicles secreted by most cell types, which carry proteins, lipids, and nucleic acids that influence various physiological and pathological processes.⁹ They are involved in immune modulation, tissue regeneration, tumor progression, and metabolic regulation, making them promising candidates for therapeutic development.^10–13 Interestingly, emerging studies have shown that plants also produce EVs that share certain morphological characteristics with mammalian exosomes, although their biogenesis pathways remain incompletely understood.¹⁴ These plant-derived exosome-like nanoparticles (PELNs) represent a naturally occurring, biocompatible class of nanovesicles that encapsulate lipids, proteins, RNAs, and metabolites from their parent plants.¹⁴ Although the term PELNs is widely used in the literature, the classification and biological origin of these particles remain under investigation. The designation “exosome-like” reflects their structural and functional similarities to mammalian exosomes but does not necessarily imply identical biogenesis pathways. According to the Minimal Information for Studies of Extracellular Vesicles 2023 (MISEV2023) guidelines, the term “exosome” should be reserved for vesicles whose endosomal origin has been established.¹⁵ Furthermore, PELN preparations obtained through plant tissue homogenization may contain extracellular vesicles, intracellular membrane fragments, and other co-isolated nanoparticles, making their precise identity difficult to determine.¹⁶ Therefore, in this review, the term PELNs is retained to maintain consistency with the existing literature while acknowledging the current limitations in distinguishing genuine extracellular vesicles from other plant-derived nanoparticles.

Beyond their potential as delivery vehicles, certain PELN preparations have demonstrated intrinsic biological activities, including the ability to modulate immune responses, regulate the gut microbiota, suppress tumor growth, and alleviate inflammatory or metabolic disorders.^17–20 This growing body of evidence suggests that PELNs are not merely passive delivery systems, but possess intrinsic therapeutic activities derived from the natural compositions of their source plants. Despite these promising findings, several important challenges remain in understanding the biological functions and therapeutic potential of PELNs. Variations in isolation and purification protocols can influence particle purity, reproducibility, and molecular composition, making comparisons across studies difficult. Moreover, the extent to which the reported biological effects are attributable to vesicle-associated cargo rather than co-isolated bioactive compounds remains unclear. The lack of standardized dosing strategies and sufficient mechanistic validation further complicates the interpretation of experimental findings. Addressing these limitations is essential for establishing the therapeutic relevance and reproducibility of PELNs. From a translational perspective, clinical translation will require scalable manufacturing processes, standardized quality control, and systematic evaluation of pharmacokinetics, biodistribution, and long-term safety. Furthermore, the development of appropriate regulatory frameworks remains an important consideration for translating heterogeneous plant-derived nanoparticle preparations into clinically applicable therapeutics.^21,22

Although previous reviews have summarized the composition, isolation, and therapeutic applications of PELNs, the relationship between methodological heterogeneity and the interpretation of their biological activities warrants further critical evaluation.^21,23 Accordingly, this review focuses on the direct therapeutic potential of PELNs, rather than their use as drug carriers. We first summarize the evolution of plant-based therapeutics and introduce the biological characteristics and mechanisms of PELNs, while critically examining how isolation strategies and particle heterogeneity affect the reliability and comparability of reported findings. We then highlight the emerging role of PELNs in tissue regeneration and bone metabolism, discussing recent advances that suggest their potential in treating osteoporosis as a representative example of a multifactorial disorder involving interconnected cellular and systemic regulatory mechanisms. By integrating evidence from diverse botanical sources with current knowledge of PELN isolation, biological mechanisms, and translational limitations, we aim to provide a critical perspective on their potential development as natural product-based nanomedicines.

Evolution of Plant-Based Therapeutics

Herbal medicines and the concept of “food-as-medicine” have served as foundational elements of health care across human civilization.^24,25 Throughout history, diverse cultures have independently developed sophisticated systems of plant-based therapies, ranging from Ayurveda in India and Kampo in Japan to traditional healing practices in Africa, the Americas, and Europe. TCM is also a prominent example of a medical system that emphasizes the use of medicinal plants.^26–29

Although these systems differ in theory, practice, and cultural background, they share a common understanding that botanical resources can be used not only for nutrition, but also for disease prevention and treatment. Modern pharmacology has repeatedly validated the clinical value of these botanical traditions. Many essential medicines used today were originally identified through the study of plants used in traditional or empirical medicine.³⁰ Notable examples include morphine isolated from Papaver somniferum,^31,32 artemisinin from Artemisia annua,^33,34 and potent anticancer agents such as paclitaxel from the Pacific yew tree³⁵ and camptothecin from Camptotheca acuminata.³⁶ These breakthroughs demonstrate that traditional knowledge, when combined with modern scientific investigation, remains an important driver of drug discovery and therapeutic innovations.

In contrast to the reductionist strategy of isolating single active molecules, many traditional and dietary plant-based interventions rely on the combined effects of multiple bioactive components.³⁷ Fruits, vegetables, grains, tea leaves, herbs, and other edible or medicinal plants contain complex mixtures of polyphenols, flavonoids, terpenoids, alkaloids, saponins, lipids, proteins, and small (s)RNAs. These components may act through multiple targets and pathways, which is particularly relevant for complex disorders. This multicomponent and multitarget nature has made plant-based therapeutics an important area of interest in food science, functional foods, nutraceuticals, and natural products research.

Despite their broad biological potential, clinical and functional applications of plant-derived bioactive compounds still face several important limitations. Many phytochemicals exhibit poor water solubility, chemical instability, rapid degradation in the gastrointestinal (GI) tract, and limited oral bioavailability.³⁸ In addition, isolating single molecules might not fully preserve the coordinated activity of whole plant materials or complex botanical preparations. These limitations encouraged researchers to reconsider how plants exert their biological effects and to explore whether naturally existing structures within plants may contribute to the stability, delivery, and activity of their bioactive components. This perspective provides an important foundation for understanding the emerging interest in plant-derived nanoscale vesicles in modern biomedical and food science research.

Exosomes

Among the various EV subtypes, exosomes represent the most extensively studied population. Exosomes are typically defined as small vesicles ranging from 30 to 150 nm in diameter, that originate from the endosomal pathway, where early endosomes mature into late endosomes and subsequently form multivesicular bodies (MVBs).³⁹ Within these MVBs, the inward budding of the limiting membrane produces intraluminal vesicles (ILVs), which are eventually released into the extracellular space as exosomes when MVBs fuse with plasma membranes.⁴⁰ Structurally, exosomes are lipid bilayer-enclosed vesicles enriched in cholesterol, sphingomyelin, ceramides, and phosphatidylserine, which confer stability and protect their cargo from enzymatic degradation.^41,42 Their molecular contents include diverse bioactive molecules, such as various proteins and RNAs.^43,44 In addition, they also include lipids and metabolites.⁴² These complex molecular cargos reflect the physiological state of their parent cells and underpin their function as vehicles of cell-cell signaling,^45,46 gene regulation,⁴⁷ and immune modulation.^9,48

Plant-Derived Exosome-Like Nanoparticles (PELNs)

PELNs are nanoscale particles obtained from various plant species, although not all preparations can be definitively classified as naturally secreted extracellular vesicles.⁴⁹ Plant EVs can be released through several cellular trafficking pathways and can be isolated from plant tissues, cell cultures, or apoplastic fluids.⁵⁰ However, their exact biogenesis in plants remains less clearly defined.⁵¹ While mammalian exosomes are generated through the endosomal MVB pathway, current evidence suggests that plant EVs may originate from multiple secretory pathways, including MVB-mediated secretion and exocyst-positive organelle (EXPO)-associated trafficking.^52,53 Other mechanisms, such as vacuole-associated and autophagy-related secretion, have also been proposed, although their contributions remain incompletely understood.^54–56 Importantly, the biogenesis of specific plant EV populations does not necessarily establish the origin of nanoparticles recovered from homogenized plant tissues. The sizes of PELNs generally range from 30 to 150 nm.⁵⁷ However, larger particle populations have also been reported, depending on the plant source and isolation method. Their morphology is often described as spherical or cup-shaped when viewed under electron microscopy.^58,59 Although the structural features of PELNs are broadly similar across species, their molecular compositions reflect the biochemical characteristics and physiological roles of the parent plant.

Proteins contained within PELNs are among their most abundant and functionally diverse constituents. Proteomic analyses identified proteins involved in membrane trafficking, stress responses, enzymatic metabolism, and signal transduction.⁶⁰ For instance, tetraspanins (TETs), a family of transmembrane proteins that contribute to organizing membrane microdomains, mediate vesicle transport, and facilitate molecular interactions.⁶¹ Another notable component is the plant-specific syntaxin, PEN1, which is involved in cell wall formation and modification, and plays an essential role in defense against fungal invasion.⁶² The interaction between PEN1 and Rab GTPases, members of the Ras-related small GTPase family, promotes membrane fusion events and directs PELN trafficking within the cellular environment.⁶³ Additionally, PELNs carry enzymes related to redox homeostasis, such as peroxidases, superoxide dismutases, and catalases, which are essential for maintaining cellular defense against oxidative stress, reflecting their potential involvement in plant immune mechanisms.⁶⁴ These conserved protein families suggest that PELNs play active roles in intracellular protection, nutrient transport, and intercellular signaling within plant systems.

PELNs also contain abundant sRNAs, including microRNA (miRNA)-like molecules and small interfering (si)RNAs. These RNAs are selectively packaged during vesicle formation and are thought to participate in regulating gene expressions within plants, particularly under environmental stress or pathogen challenge.⁶⁵ Among them, miRNAs have attracted considerable interest for their capacity to modulate gene expressions at the post-transcriptional level and to facilitate interspecies communication, including between plants and mammalian cells.^66,67 In addition to miRNAs, specific siRNAs, such as IGN-siR1, TAS2-siR453, and TAS1c-siR483, were reported to mediate cross-kingdom RNA interference by silencing fungal virulence genes.⁶⁸ Collectively, these findings highlight the potential of PELN-derived sRNAs as functional mediators in both plant defense and interspecies communication.

Lipids are the principal structural components of PELNs, forming the bilayer membrane that protects the vesicular contents and determines vesicle stability. Lipid profiles of PELNs differ from those of mammalian exosomes, as they contain high proportions of phosphatidic acid (PA), phosphatidylcholine (PC), digalactosyldiacylglycerol (DGDG), and monogalactosyldiacylglycerol (MGDG).^18,69 These lipids are common constituents of plant cellular membranes and are crucial for maintaining vesicle curvature, rigidity, and fusion capacity.⁷⁰ This unique lipid composition is also believed to play an important role in vesicle formation and cargo sorting within the plant endomembrane system.⁷¹ Additionally, they are not merely structural. PA, for example, was implicated in modulating immune responses⁷² and enhancing membrane fusion and cellular uptake,⁷³ while galactolipids contribute to antioxidant activity.⁷⁴ Collectively, these lipid species not only define the biophysical properties of PELNs but also actively contribute to their biological functionality and intercellular communication potential.

In addition to macromolecules, PELNs carry a wide spectrum of secondary metabolites that mirror the phytochemical composition of their source plants. Metabolomic and chromatographic analyses have identified numerous plant-specific metabolites encapsulated within these vesicles. For instance, ginger-derived PELNs contain gingerols, shogaols, and paradols, compounds known for their antioxidative and anti-stress functions within the plant rhizome.⁷⁵ Grape-derived PELNs are rich in polyphenols and resveratrol derivatives, which contribute to defense against UV irradiation and pathogen invasion.⁷⁶ Lemon and grapefruit PELNs contain flavonoids such as hesperidin, naringenin, and eriocitrin, molecules involved in regulating reactive oxygen species in citrus tissues.^77,78 PELNs isolated from ginseng were shown to include ginsenosides, triterpene saponins that play roles in plant growth and immune signaling,⁷⁹ while Pulsatilla chinensis-derived vesicles may carry oleanane-type saponins, compounds associated with the plant’s antimicrobial defense.⁸⁰ These secondary metabolites serve multiple functions in plants, including antioxidant protection, pathogen resistance, and intercellular communication during stress responses. Encapsulation within vesicles may help stabilize these molecules, facilitating their storage and targeted transport between plant cells or tissues.

Overall, PELNs can be viewed as complex, naturally assembled nanostructures that carry molecular components reflecting the characteristics of their parent plant. Understanding this compositional diversity is important for interpreting their biological activities and evaluating their therapeutic potential.

Advantages of PELNs Compared to Mammalian Exosomes

Mammalian cell-derived exosomes have attracted significant attention in recent years as potential tools for diagnostics and therapy. These vesicles can mediate intercellular communication by transferring proteins, RNAs, and lipids⁸¹ and have been engineered to deliver chemotherapeutic agents, nucleic acids, and small molecules to target cells.⁸² Clinical studies have explored the use of exosomes derived from immune cells, mesenchymal stem cells, and tumor cells for applications such as immune regulation, regenerative therapy, and drug delivery.⁸³ However, despite their therapeutic promise, the use of mammalian exosomes remains limited by several challenges, including low production yields, complex isolation processes, potential immunogenicity, and the risk of transmitting animal- or cell-derived pathogens.^84,85 Furthermore, large-scale manufacturing of mammalian exosomes is expensive and difficult to standardize, restricting their clinical translation.⁸⁶

In contrast, PELNs represent a naturally abundant, biocompatible alternative. PELNs can be obtained from edible and medicinal plants using potentially cost-effective procedures, although large-scale production and quality control remain challenging.⁸⁷ These nanoparticle preparations can be obtained from various plant parts, such as roots, leaves, fruits, and seeds, but their structural characteristics and batch-to-batch reproducibility depend on the source material and isolation procedures.⁸⁸ Importantly, PELNs have been employed in two principal ways: as delivery vehicles for therapeutic compounds and as direct therapeutic agents.²⁰ While a few studies have engineered PELNs to encapsulate exogenous drugs or RNA molecules, an increasing number of investigations have highlighted their intrinsic biological activities, derived from the natural metabolites, lipids, and sRNAs they carry.²¹ This dual potential supports the investigation of PELNs as both naturally bioactive preparations and potential delivery vehicles.

From a physicochemical and practical perspective, PELNs may offer several potential advantages, although direct comparisons with mammalian exosomes remain limited. Their plant-derived lipid compositions may contribute to stability under certain gastrointestinal conditions, as reported for selected preparations.⁸⁹ The edible origin of many plant sources may offer certain practical advantages and reduce concerns associated with animal-derived pathogens, although the immunogenicity and safety of PELNs, particularly following repeated administration, require further assessment.⁹⁰ Collectively, these characteristics highlight the potential of PELNs as plant-derived nanotherapeutic candidates, while emphasizing the need for standardized production, comparative characterization, and safety evaluation.

Uptake and Bioavailability of PELNs in Mammalian Systems

Interactions between PELNs and mammalian cells have increasingly been the focus of research, as they provide critical insights into how orally or systemically administered vesicles exert biological activity. Unlike synthetic NPs, which often require chemical modification to enhance their stability or cellular uptake, PELNs possess naturally evolved lipid and surface compositions that facilitate their absorption and internalization.¹⁸ Experimental evidence from multiple plant sources, including ginger, grape, grapefruit, broccoli, and ginseng, suggests that PELN-associated materials may reach mammalian tissues following administration, although the biodistribution of intact vesicles remains incompletely understood.⁹¹ One of the notable properties of PELNs is their structural resilience in the GI tract. Their membranes, which are enriched in PA, galactolipids, and other plant-specific lipids, exhibit strong resistance to acidic conditions and digestive enzymes.⁹² In vitro simulations and in vivo tracking experiments showed that a significant proportion of orally administered PELNs remain structurally intact after exposure to gastric fluid and pancreatic enzymes.^93,94 This stability allows PELNs to reach the small intestine without undergoing rapid degradation, thereby preserving the biological activities of their cargos. The unique lipid composition of PELNs is believed to play a central role in this process, although the precise molecular mechanisms that confer this protection are still under investigation.⁹⁵

After reaching the intestinal lumen, PELNs interact with epithelial cells and resident immune cells and are taken up through several endocytic pathways. Fluorescence-labeling studies demonstrated that PELNs can enter mammalian cells through clathrin-mediated endocytosis, caveolin-dependent endocytosis, macropinocytosis, or direct membrane fusion, depending on the plant source and lipid profile of the vesicle.^96–99 Once internalized, PELNs are trafficked through endosomal compartments, where changes in pH and membrane composition can facilitate the release of their proteins, lipids, secondary metabolites, and sRNAs into the cytoplasm. This intracellular delivery process mirrors certain aspects of mammalian exosome uptake, although direct comparative evidence of greater uptake efficiency remains limited.¹⁰⁰ Once internalized by mammalian cells, PELNs undergo endosomal maturation and partial lysosomal processing, which facilitate the release of their molecular cargos.^101–104 Proteins and metabolites may diffuse into the cytoplasm, while sRNAs can associate with mammalian RNA-binding complexes.^7,105 In some cases, bioactive metabolites, such as flavonoids, terpenoids, or saponins, may also be liberated during vesicle degradation, allowing them to interact directly with intracellular signaling pathways.^70,106 These demonstrate that PELNs act not merely as passive particles but as biologically active nanostructures capable of delivering functional molecules into mammalian cells.

Together, these findings suggest that selected PELN preparations may exhibit stability under certain physiological conditions and undergo cellular internalization, although the extent of intact-vesicle uptake and systemic biodistribution requires further investigation.

Isolation Strategies for PELNs

Isolating PELNs represents a crucial step for evaluating their structural, molecular, and functional properties. Because plant tissues contain abundant fibers, polysaccharides, pigments, and secondary metabolites, the methods used for PELN isolation must balance purity, yield, scalability, and preservation of vesicle integrity.²³ Various strategies have been adapted from mammalian EV research,¹⁰⁷ although plant-derived vesicles often require additional clarification and purification steps due to the complexity of plant matrices. A schematic representation of the isolation strategies as well as the pros and cons of each strategy are presented in Figure 1.

Figure 1.

Schematic illustration of the isolation workflows and the advantages and limitations of different strategies. In general, plant materials are homogenized and subjected to centrifugation to obtain a clarified supernatant containing nanoscale vesicles. (A) Differential ultracentrifugation isolates PELNs through sequential centrifugation steps at increasing speeds, allowing separation based on particle size and sedimentation rate. (B) Density-gradient ultracentrifugation separates particles according to their buoyant density using gradient media such as sucrose solutions. (C) Polyethylene glycol (PEG) precipitation enriches PELNs through polymer-mediated precipitation based on PEG–vesicle interactions. (D) Ultrafiltration and tangential-flow filtration separate PELNs from other components using membranes with defined molecular weight or size cutoffs. (E) Size-exclusion chromatography (SEC) separates particles via porous matrices (eg, Sepharose), followed by fractionated elution according to particle size. The centrifugation conditions illustrated are representative examples and may vary depending on the plant source and isolation protocol. Created in BioRender. Chen, (Y) (2026) https://BioRender.com/7pjo72y.

Differential Ultracentrifugation

Differential ultracentrifugation (Figure 1A) is the most widely used method for isolating PELNs. It relies on sequential centrifugation steps that progressively remove coarse debris, cell fragments, and larger vesicles, followed by high-speed ultracentrifugation that pelletizes nanosized vesicles. Typical workflows involve initial low-speed centrifugation to eliminate plant fibers, intermediate speeds to remove larger particles, and final ultracentrifugation at approximately 100,000 to 150,000 ×g to collect PELNs.¹⁰⁸

This method generally produces preparations with moderate purities and acceptable yields, although the final products often contain small contaminants if plant extracts are highly viscous or rich in polysaccharides.¹⁰⁹ Differential ultracentrifugation is appropriate for laboratory-scale experiments but is less suitable for industrial scaling due to its long processing time, dependence on large ultracentrifuge rotors, and lower throughput. In addition, exposure to extreme g-forces may affect vesicle morphology, occasionally leading to deformation or aggregation, which can influence vesicle integrity.¹¹⁰

As summarized in Table 1, differential ultracentrifugation has been widely applied to PELNs isolated from diverse plant sources. The reported average particle diameters vary considerably, ranging from 123 nm for Camellia sinensis leaves to 397.5 nm for red onion-derived nanoparticles. Intermediate particle sizes were also reported for apple-, plum-, yam-, and bitter melon-derived PELNs. These findings further demonstrate that the final particle size may be influenced by both the isolation conditions and the intrinsic properties of the source plant.

Table 1.

Comparison of Isolation Methodologies and Average Particle Diameters of Isolated Plant-Derived Exosome-Like Nanoparticles (PELNs)

Plant Source (Scientific Name)	Average Particle Diameter	Characterization Methods	Reference
Differential ultracentrifugation
Fenugreek seeds (Trigonella foenum-graecum)	176 to 361 nm	SEM; DLS; ZP	[111]
Pueraria lobata	119 to 163 nm	TEM; NTA	[112]
Houttuynia cordata Thunb	143.8 nm	TEM; NTA; WB (CD9/CD63/TSG101)	[113]
Phellinus linteus	100 to 260 nm	TEM; DLS	[114]
Houttuynia cordata	168 nm	TEM; NTA; ZP	[115]
Tangerine peel (Citrus reticulata)	172 nm	TEM; NTA; LC–MS/MS	[116]
Fuji apple (Malus pumila)	236.8 nm	NA	[117]
Red onion (Allium cepa)	397.5 nm	TEM; DLS; ZP	[118]
Bitter melon (Momordica charantia L).	155 nm	EM; NTA	[119]
Yam (Dioscorea japonica)	168 and 328 nm	TEM; cryo-TEM; NTA	[120]
Leaves of Camellia sinensis	123 nm	TEM; NTA	[121]
Robinia pseudoacacia L. flowers	176.3 nm	TEM; particle-size analysis (instrument unspecified)	[122]
Plum (Prunus salicina)	211 nm	TEM; cryo-EM; NTA	[123]
Sucrose density-gradient ultracentrifugation
Ginseng (Panax ginseng)	100 to 130 nm	TEM; particle-size analyzer (model unspecified)	[124]
	71.42 nm	TEM; DLS; NTA; ZP	[125]
Andrographis paniculata	180 nm	TEM; DLS	[126]
Goji berries (Lycium barbarum L).	171 nm	TEM; NTA	[127]
Grapefruit (Citrus paradisi)	400 nm	TEM; NTA	[128]
Portulaca oleracea L.	160 nm	Cryo-TEM; NTA; DLS; ZP	[129]
Garlic (Allium sativum)	160 nm	NA	[130]
Differential centrifugation followed by sucrose density gradient ultracentrifugation
Catharanthus roseus	75 nm	TEM; DLS; nanoFCM; ZP	[131]
Panax notoginseng	151 nm	TEM; DLS; ZP	[132]
Platycodon grandiflorum	141 nm	TEM; DLS; NTA; ZP; SDS–PAGE	[133]
Mulberry bark (Morus alba)	151 nm	EM; NTA	[134]
Goji berries (Lycium barbarum L).	128 nm	TEM; NTA	[135]
Houttuynia cordata	100 nm	TEM; NTA; DLS; ZP	[136]
Citrus reticulata (Unshiu)	175 nm	TEM; NTA; ZP	[137]
Peucedanum japonicum	140 nm	TEM; NTA	[138]
Leaves of Camellia sinensis	167 nm	TEM; AFM; DLS; ZP	[139]
	150 nm	TEM; NTA	[140]
	166.5 nm	TEM; NTA	[141]
Lavender (Lavandula angustifolia)	160 nm	TEM; DLS; NTA; ZP	[142]
Tea flowers (Camellia sinensis)	131 nm	TEM; DLS; ZP	[143]
Polyethylene glycol precipitation
Ginseng (Panax ginseng)	150 nm	TEM; NTA; WB (CD63/CD81)	[144]
Curcuma amada	324.7 nm	SEM; DLS; ZP	[145]
Polyethylene glycol-based precipitation method combined with differential centrifugation
Trifolium pratense	167.3 nm	DLS; ZP	[146]
Brucea javanica	105 nm	TEM; NTA; DLS; ZP	[147]
Lycium ruthenicum Murray	114 nm	TEM; NTA; ZP; LC–MS/MS	[148]
Papaya (Carica papaya L).	169 nm	TEM; DLS; ZP	[149]
Black nightshade berry (Solanum nigrum L).	107 nm	TEM; NTA; DLS; ZP	[150]
Kiwi fruit (Actinidia deliciosa)	30 to 150 nm	TEM; DLS; LC–MS/MS	[151]
Wheat (Triticum aestivum)	40 to 100 nm	SEM; flow cytometry (CD9/CD63/HSP70)	[152]
Tangential-flow filtration
Artemisiae argyi	68 to 290 nm	TEM; DLS; ZP	[153]
Size-exclusion chromatography
Brassica oleracea	137.8 nm	TEM; NTA	[154]
Membrane affinity spin columns (commercial)
Hemp seed (Cannabis sativa)	95 nm	NTA; WB (Annexin V)	[155]

Notes: NA indicates that characterization methods could not be verified from the information accessible to the authors and does not necessarily indicate that these methods were not reported in the original study. Reported particle diameters should be interpreted cautiously because isolation and characterization protocols vary across studies.

Abbreviations: NA, not available; DLS, dynamic light scattering; NTA, nanoparticle tracking analysis; SEM, scanning electron microscopy; TEM, transmission electron microscopy.

Density-Gradient Ultracentrifugation

Density-gradient ultracentrifugation (Figure 1B) separates vesicles on the basis of their buoyant density by layering plant extracts over sucrose or iodixanol gradients. During ultracentrifugation, vesicles migrate to density fractions ranging from approximately 1.12 to 1.19 g/mL, allowing removal of contaminants such as proteins, pigments, and membrane fragments.¹⁵⁶

This technique produces highly purified vesicle preparations and is often considered the gold standard for distinguishing true nanoscale vesicles from co-isolated particles. However, the long centrifugation time and the need for precise gradient preparation reduce the scalability of this approach. Although it yields highly pure vesicles and preserves morphology due to controlled sedimentation, overall yields are generally low, and the method is not optimal for large-volume processing.¹⁵⁷

Density-gradient ultracentrifugation has also been applied to PELNs derived from various fruits, vegetables, and medicinal plants, as summarized in Table 1. Most reported particle diameters ranged from approximately 70 to 180 nm, including those derived from ginseng, goji berries, garlic, and Portulaca oleracea. However, grapefruit-derived nanoparticles were reported to reach approximately 400 nm. This broad variation indicates that density-based purification improves sample purity but does not necessarily produce a uniform particle-size distribution across different plant sources.

Polyethylene Glycol (PEG) Precipitation

PEG precipitation (Figure 1C) concentrates vesicles by reducing their solubility in aqueous solutions.¹⁵⁸ Plant extracts are mixed with PEG and incubated at cold temperatures, after which precipitated vesicles are collected by low-speed centrifugation. The principle is based on polymer-driven exclusion forces that aggregate nanosized particles.¹⁵⁹

This strategy is simple, inexpensive, and highly adaptable to large-scale workflows, making it valuable when high yields are required. However, PEG precipitation typically results in vesicles of lower purity because soluble proteins, polysaccharides, and secondary metabolites efficiently co-precipitate. The presence of these additional molecules may obscure particle characterization or alter vesicle surface properties. Although the method preserves the vesicle integrity better than high-speed ultracentrifugation, the purity and reproducibility may widely vary across plant species.^160,161

As shown in Table 1, PEG-based isolation methods have produced PELNs with variable particle diameters across different plant sources. Ginseng-derived nanoparticles exhibited an average diameter of approximately 150 nm, whereas Curcuma amada-derived nanoparticles reached 324.7 nm. When PEG precipitation was combined with differential centrifugation, particle sizes ranging from 30–150 nm for kiwi fruit to approximately 169 nm for papaya were reported. This variation may reflect differences in the plant matrix as well as the co-precipitation of non-vesicular components during PEG-based isolation.

Ultrafiltration and Tangential-Flow Filtration

Ultrafiltration (Figure 1D) isolates PELNs by passing plant extracts through membranes with defined molecular weight cutoffs, typically between 100 and 300 kDa. Tangential-flow filtration is a modified form of this method in which the sample flows parallel to the membrane surface, reducing clogging and allowing continuous processing.¹⁶²

These filtration-based techniques are advantageous for their rapid processing times and scalability, especially when large plant volumes must be concentrated. They maintain vesicle morphologies well, since they do not require extreme centrifugation forces.¹⁶³ However, membrane fouling can occur with viscous plant extracts, which may reduce yields. Furthermore, the efficiency of separation heavily depends on the selected pore size, influencing both the purity and vesicle size distribution.¹⁶⁴

Compared with centrifugation- and precipitation-based approaches, relatively limited information is currently available for filtration-based isolation methods in Table 1. Artemisiae argyi-derived nanoparticles isolated using tangential-flow filtration exhibited a broad particle-size range of 68–290 nm. This broad distribution suggests that filtration-based methods may recover heterogeneous nanoparticle populations. The final particle characteristics may therefore depend on both the selected membrane conditions and the complexity of the plant matrix.

Size-Exclusion Chromatography (SEC)

SEC (Figure 1E) separates particles based on their hydrodynamic radius using porous matrices such as Sepharose. When plant extracts are applied to the column, larger vesicles elute earlier, while smaller molecules such as proteins, pigments, and free metabolites remain within the porous matrix and elute later. This produces a clear physical separation between nanosized vesicles and soluble contaminants.¹⁶⁵

SEC generates vesicles of high purity and excellent structural integrity. It is particularly effective when used after a concentration step such as ultrafiltration or initial centrifugation. Although the method is reproducible and gentle, its limited loading capacity reduces its scalability. Yields are typically lower than with PEG precipitation but higher in purity than differential ultracentrifugation alone.¹⁶⁶

The currently available data for SEC-based isolation remain limited in Table 1. Brassica oleracea-derived nanoparticles isolated using this method had an average particle diameter of 137.8 nm. This result is consistent with the ability of SEC to separate nanosized vesicles from smaller soluble contaminants while preserving vesicle integrity. However, additional data from other plant sources are needed to determine whether similar particle characteristics can be consistently obtained using this approach.

Methodological Comparisons and Biological Heterogeneity of PELNs

To systematically evaluate current preparation strategies, representative studies were organized in Table 1 according to their primary isolation methods. Overall, the reported particle-size distributions substantially overlap among differential ultracentrifugation, density-gradient ultracentrifugation, PEG-based precipitation, filtration, and SEC, indicating that no single isolation approach consistently produces PELNs within a narrowly defined size range.

The data also reveal considerable variation among different plant sources processed using similar isolation methods. This finding suggests that the reported particle size of PELNs may be strongly influenced by intrinsic biological factors, including the tissue type, viscosity, fiber and polysaccharide contents, lipid composition, and molecular cargo of the source plant, although the relative contributions of these factors and isolation conditions cannot be determined from cross-study comparisons.¹⁶⁷ Nevertheless, methodological variables, such as centrifugation force and duration, membrane cutoff, purification sequence, and particle-characterization technique, may also contribute to the observed differences among studies.

Recent Studies on Therapeutic Applications of PELNs

In recent years, numerous studies have explored the therapeutic potential of PELNs across a wide spectrum of disease models. These investigations demonstrated that PELNs obtained from different botanical sources possess distinct biological activities and can modulate mammalian physiology through diverse molecular mechanisms. The following subsections discuss these findings based on their botanical origins, while Table 2 summarizes representative studies according to their primary biological regulatory axes and experimental models, allowing a clearer distinction between in vitro and in vivo evidence.

Table 2.

Therapeutic Potential and Cross-Kingdom Molecular Mechanisms of Plant-Derived Exosome-Like Nanoparticles (PELNs)

Plant Source (Scientific Name)	Therapeutic Applications	Research Models	Administration Route	Biological Functions and Mechanisms	Reference
Immune and Inflammatory Axis
Ginseng (Panax ginseng)	Rheumatoid arthritis	Collagen-induced arthritis (CIA) mouse model	Oral	1. Lowering serum levels of IL-6 and TNF-α 2. pgi-miR6135j suppresses KRAS expression and downstream MAPK signaling in bone macrophages	[124]
Catharanthus roseus	Chemotherapy-induced immunosuppression	RAW264.7 cells and cyclophosphamide-induced immunosuppressed BALB/c mice	Intraperitoneal	Restoring immune function by activating the TNF-α/NF-κB/PU.1 pathway, enhancing macrophage and lymphocyte activity and reversing leukopenia	[131]
Panax notoginseng	Cerebral ischemia-reperfusion injury	Transient middle cerebral artery occlusion (tMCAO) rat model and primary microglia cultures	Intravenous	Inducing M1-to-M2 microglial polarization through lipid-mediated PI3K/Akt activation, suppressing inflammation and promoting neuroprotection	[132]
Platycodon grandiflorum	Acute lung injury (ALI)	RAW264.7 cells and lipopolysaccharide-induced ALI in C57BL/6J mice	Intravenous	Inducing M2 polarization of RAW264.7 cells via modulation of glycolytic and lipid metabolic pathways, linking metabolic reprogramming to inflammation control	[133]
Mulberry bark (Morus alba)	Dextran sulfate sodium (DSS)-induced colitis	MC38 and Caco-2 cells and DSS-treated mice	Oral	1. HSPA8 activates AhR signaling in intestinal epithelial cells, inducing COPS8 and antimicrobial peptide expression 2. Epithelial COPS8 deletion abolishes protection, establishing the AhR–COPS8 axis as essential for anti-inflammatory activity	[134]
Houttuynia cordata	Virus infection	HEK293T cells and influenza A virus (H1N1), SARS-CoV-2, and RSV	N/A	1. miR858a/miR858b targeting the NP gene of influenza A virus (H1N1) and miR166a-3p targeting ORF1ab of SARS-CoV-2 2. miR168a-3p, miR168b-3p, miR8175, and novel_mir2 suppress human MAPK3 and AKT1 expressions	[115]
Houttuynia cordata	DSS-induced colitis	NLRP3-deficient C57BL/6 mice	Oral	1. Ameliorating colitis via inhibition of NLRP3 inflammasome activation and cytokine production 2. Modulating gut microbiota to promote mucosal protection	[136]
Artemisiae argyi	Ulcerative colitis (UC)	HT-29 cells and dextran sulfate sodium (DSS)–induced colitis in BALB/c mice	Oral	1. Ameliorating colitis by restoring barrier integrity and reducing inflammation, oxidative stress, and epithelial apoptosis 2. Restoring gut microbiota homeostasis through integrated miRNA, lipidomic, and metabolomic regulation	[153]
Papaya (Carica papaya L).	Inflammation-related conditions	RAW 264.7 cells and zebrafish larvae	Immersion exposure	1. Attenuating LPS-induced inflammation by reducing nitric oxide, IL-1β, and IL-6 expressions while increasing IL-10 levels 2. Reducing macrophage and neutrophil migration to the wound site	[149]
Black nightshade berry (Solanum nigrum L).	Macrophage-driven inflammation	RAW264.7 cells	N/A	Reducing IL-6 expression at both mRNA and protein levels	[150]
Peucedanum japonicum	Colitis	RAW264.7, RAW-Blue, THP-1–derived macrophage-like cells, and HL-60 cells and DSS-induced colitis model in C57BL/6J mice	Oral	1. Suppressing LPS-induced inflammation by reducing IL-6 and IL-1β expressions and inhibiting TNFα-NF-κB, IL-6-JAK-STAT3, and IFNγ signaling pathways 2. Protecting colonic epithelium and suppressing IL-6-mediated inflammation via PjELN-derived small RNAs	[138]
Portulaca oleracea L.	Colitis	Dextran sulfate sodium–induced colitis in both C57BL/6 and IL-10−/− mice	Oral	Enriching L. reuteri-derived indoles to activate AhR and reprogram CD4⁺ T cells into anti-inflammatory CD4⁺CD8⁺ cells	[129]
Lavender (Lavandula angustifolia)	UVB-induced skin photoaging	HaCaT cells and ICR mice	Topical	1. Reducing the secretion of IL-1β, IL-6, and TNF-α 2. cpa-miR166e and zma-miR166h-3p regulate inflammation, oxidative stress responses, and collagen metabolism	[142]
Hemp seed (Cannabis sativa)	Androgen-induced follicle growth inhibition	Dihydrotestosterone-induced feather growth inhibition model in chicken embryos	In ovo	1. Restoring SHH and Wnt/β-catenin signaling and suppressing AR/BMP4 expression under DHT conditions 2. Modulating immune-related markers associated with follicular inflammation	[155]
Kiwi fruit (Actinidia deliciosa)	Antifungal activity	Fusarium oxysporum, Fusarium solani, and Botrytis cinerea	N/A	Reducing spore germination and lowering spore density	[151]
Gut Microbiota and Barrier Function Axis
Andrographis paniculata	Ulcerative colitis	NCM460, HCT116, and RAW 264.7 cells and DSS-treated C57BL/6 and IL-10−/− mice	Oral	1. Upregulating claudin-1, ZO-1, occludin, and mucin2 2. Increasing the abundance of the probiotic Lactobacillus murinus 3. Inducing M2 polarization of colonic macrophages through PI3K-AKT and JAK-STAT activation with increased IL-4R expression	[126]
Citrus reticulata (Unshiu)	Metabolic and microbiota-related disorders	HFD-induced gut dysbiosis C57BL/6J mice	Oral	1. Decreasing pathogenic bacteria while enriching beneficial commensals 2. Increasing acetic acid production and transferring lipid/miRNA cargos (crt-miR166a/168/171) to modulate microbial gene regulation	[137]
Garlic (Allium sativum)	Inflammatory bowel disease (IBD)	Caco-2 cells and dextran sulfate sodium–induced colitis models in C57BL/6J mice	Oral	1. han-miR3630-5p suppresses TLR4/MyD88/NF-κB signaling 2. Modulating gut microbiota by increasing Lachnospiraceae and reducing Helicobacter abundance	[130]
Leaves of Camellia sinensis	Intestinal barrier damage	Caco-2 and T84 cells and rats	Oral	Preserving intestinal epithelial and immune barriers through tight junction protection and IL-22 signaling.	[140]
	Irritable bowel syndrome (IBS)	IBS induced by psychological stress in a rat model	Oral	Restoring body weight, preserving tight junction integrity, reducing visceral hypersensitivity, and suppressing CRH and CRHR1 expressions	[141]
Oxidative Stress and Cellular Homeostasis Axis
Houttuynia cordata Thunb	Ischemic brain injury	OGD/R-treated HT22 cells and rat middle cerebral artery occlusion model	Intravenous	miR159a targets ACSL4 to inhibit ferroptosis and reduce mitochondrial damage	[113]
Phellinus linteus	UV-induced skin aging	HaCaT cells, human skin fibroblasts, UV-irradiated mouse skin, and human volunteer skin	Dorsal skin	1. Reducing UV-induced oxidative stress and senescence by lowering ROS/MDA/SA-β-Gal/MMP1 and enhancing SOD and collagen gene expressions 2. miR-CM1 targets Mical2, enhancing COL1A2 expression and reducing UV-induced skin aging	[114]
Lycium ruthenicum Murray	Amyloid-β (Aβ)-induced neuronal injury	PC12 cells	N/A	Reducing Aβ-induced apoptosis through miRNA-dependent regulation of MAPK and PI3K/AKT pathways	[148]
Bitter melon (Momordica charantia L).	Doxorubicin-induced cardiotoxicity	Doxorubicin-treated C57BL/6J mice and Sprague–Dawley rats and H9c2 cells and neonatal rat ventricular myocytes	Intravenous	Activating the p62-Keap1-Nrf2 axis via p62 stabilization to induce antioxidant signaling and cardioprotection	[119]
Leaves of Camellia sinensis	Skin hyperpigmentation	B16-F10 cells and UVB-induced pigmentation mouse model	Topical	1. Enhancing radical scavenging, tyrosinase inhibition, and suppression of melanin synthesis 2. miR-828b-MYB4 axis inhibits melanogenesis and inflammation through PI3K/AKT modulation and enhanced melanoautophagy	[121]
Robinia pseudoacacia L. flowers	Hypoxia-induced injury	NGEC and HIEC cells and C57BL/6J mice	Oral	Downregulating HIF-1α/HIF-2α and lipid peroxidation enzymes (NOX4, ALOX5), reducing ROS accumulation, and inhibiting hypoxia-induced ferroptosis	[122]
Tissue Regeneration and Bone Homeostasis Axis
Ginseng (Panax ginseng)	Osteoporosis	Bone marrow–derived macrophages (BMMs) and lipopolysaccharide-induced bone loss in mice	Subcutaneous injection	Suppressing RANKL-IκBα/JNK/ERK signaling and NFATc1/c-Fos, with Rb1/Rg1-enriched GDNs showing superior inhibition of osteoclastogenesis	[125]
Curcuma amada	Diabetic wound healing	HaCaT cells and streptozotocin-induced diabetic mouse model	Topical	Upregulating FSTL1 in keratinocytes, enhancing wound repair, and restoring FSTL1 levels in diabetic wounds	[145]
Goji berries (Lycium barbarum L).	Pro-osteogenic and fracture-healing	MC3T3-E1 cells and femoral fracture model in C57BL/6J mice	Local injection	Promoting osteogenesis and fracture repair through PI3K/Akt/mTOR activation with p70S6K and 4EBP1 phosphorylation	[127]
Pueraria lobata	Osteoporosis	Primary human bone mesenchymal stem cells and ovariectomized rat models	Intraperitoneal	Suppressing TMAO-bone resorption axis via microbiota modulation and autophagy activation	[112]
Grapefruit (Citrus paradisi)	Tendinopathy	Collagenase-induced tendinopathy mouse model and primary tendon cells	Hyaluronic acid-based microneedle injection	1. Enhancing proliferation, migration, collagen I production, and M2 macrophage polarization 2. Improving functional recovery and ECM remodeling while attenuating fibrosis	[128]
Fuji apple (Malus pumila)	Osteoporosis	MC3T3-E1 cells and C57BL/6 mice	Oral	Upregulating osteoblast markers (Runx2, ALP, OPN, and BMP2/4) via BMP2/Smad1 activation with downstream ERK/JNK MAPK signaling	[117]
Brassica oleracea	Wound healing and inflammation	HaCaT cells	N/A	Enhancing cell migration and regeneration-related gene expression while reducing NO production and suppressing IL-1β and TLR2 expressions	[154]
Red onion (Allium cepa)	Cutaneous wound healing	Full-thickness excisional skin wound model in C57BL/6 mice	Local intradermal	Promoting pro-repair M2 polarization with NF-κB suppression and increased antioxidant activity	[118]
Yam (Dioscorea japonica)	Osteoporosis	Ovariectomy-induced osteoporotic mice	Oral	Promoting osteoblast proliferation, differentiation, and mineralization via BMP-2/p38-Runx2 activation, increasing ALP, OPN, and type I collagen, independent of conventional saponins	[120]
Wheat (Triticum aestivum)	Skin wound healing	HDFs, HaCaT, and HUVECs cells	N/A	Enhancing collagen I expression and angiogenesis while suppressing apoptosis without affecting cell cycle	[152]
Plum (Prunus salicina)	Osteoporosis	MC3T3-E1 cells and primary mouse osteoclast cultures	N/A	1. Activating BMP-2–Runx2 axis to promote osteogenic differentiation 2. Suppressing osteoclast differentiation by downregulating NFATc1/c-Fos signaling and reducing TRAP⁺ cells	[123]
Metabolic Regulation and Antitumor Axis
Ginseng (Panax ginseng)	Non-small cell lung cancer (NSCLC)	A549 and H1299 cells and C57BL/6 syngeneic LLC tumor model	Intravenous	1. Suppressing PD-L1 2. Enhancing T cell antitumor activity via increased IL-2 and IFN-γ secretion	[144]
Fenugreek seeds (Trigonella foenum-graecum)	Iron deficiency anemia	Caco-2 and RAW264.7 cells and diet-induced iron deficiency anemia rat model	Oral	Facilitating stable gastrointestinal delivery of phytoferritin nanocages, restoring hemoglobin levels and hematological indices.	[111]
Trifolium pratense	Diabetic	Streptozotocin-induced diabetic rat model	Intraperitoneal	1. Decreasing fasting glucose and nitric oxide while elevating insulin, C-peptide, and antioxidant capacity 2. Upregulating β-cell functional and regenerative genes, including insulin, PDX1, NGN3, and SIRT1	[146]
Brucea javanica	Triple-negative breast cancer	4T1 cells and subcutaneous 4T1 breast tumor BALB/c mice	Intratumoral	1. Inhibiting 4T1 cell proliferation, migration, angiogenesis, and VEGF secretion via miRNA delivery 2. Inhibiting the PI3K/Akt/mTOR axis while activating ROS- and caspase-dependent apoptosis	[147]
Goji berries (Lycium barbarum L).	Glucocorticoid-induced skeletal muscle atrophy	C2C12 cells and dexamethasone-induced muscle atrophy mouse model	Intramuscular	1. Enhancing muscle mass and strength via AMPK/SIRT1/PGC1α activation 2. Promoting myocyte proliferation and mitochondrial metabolism primarily via AMPK activation	[135]
Tangerine peel (Citrus reticulata)	Type 2 diabetes-associated hepatic steatosis	3T3-L1 and AML-12 cells and C57BL/KsJ-db/db mice	Oral	1. Restoring glucose and lipid metabolic homeostasis via regulation of fatty acid β-oxidation, glycolysis, and lipogenesis pathways 2. Ameliorating gut microbiota dysbiosis and normalizing short-chain fatty acid and bile acid profiles	[116]
Leaves of Camellia sinensis	Breast cancer	4T1 cells and orthotopic breast cancer models	Oral and intravenous	1. Inducing excessive intracellular ROS, triggering mitochondrial dysfunction, cell-cycle arrest, and apoptosis 2. Modulating gut microbiota to suppress tumor-associated and inflammatory taxa while enriching protective commensals	[139]
Tea flowers (Camellia sinensis)	Cancer	MCF-7 cells and MCF-7 xenograft tumor models and 4T1 lung metastasis models in mice	Oral and intravenous	1. Promoting ROS-mediated mitochondrial apoptosis to inhibit cancer cell growth and invasion 2. Modulating gut microbiota composition	[143]

Fruits

Fruits represent a diverse group of nutrient-rich plant organs that produce exosome-like NPs with strong immunomodulatory and metabolic regulatory properties. Recent evidence demonstrated that fruit-derived vesicles are potent regulators of inflammation and metabolic homeostasis across various biological models. For instance, NPs isolated from papaya¹⁴⁹ and black nightshade berry¹⁵⁰ exerted significant anti-inflammatory effects by reducing the recruitment of macrophages and neutrophils to injury sites or by suppressing expressions of proinflammatory cytokines such as interleukin (IL)-6. Beyond inflammation, fruit-derived vesicles, like those from kiwi fruit,¹⁵¹ provide a natural defense against pathogenic fungi by inhibiting spore germination through the action of specialized defense-related proteins. The regulatory capacity of citrus species is particularly noteworthy, as vesicles from Citrus reticulata¹³⁷ and tangerine peel¹¹⁶ can reshape the gut microbiota and normalize lipid metabolism to alleviate conditions like hepatic steatosis and type 2 diabetes. Furthermore, grapefruit-derived nanovesicles¹²⁸ have shown potential as localized regenerative agents for tendinopathy, where they facilitate extracellular matrix (ECM) remodeling and functional recovery without systemic toxicity when delivered via specialized microneedle systems.

Vegetables

Vegetables are essential dietary components that yield bioactive nanovesicles capable of reinforcing epithelial barriers and promoting systemic tissue repair. Vegetable-derived exosome-like NPs frequently target the epithelial barrier integrity and cellular tissue repair mechanisms to maintain organ health. Vesicles from B. oleracea¹⁵⁴ and red onion¹¹⁸ were shown to accelerate cutaneous wound healing by promoting keratinocyte migration and coordinating antioxidant responses within the local wound microenvironment. In the context of gastrointestinal health, nanovesicles from Peucedanum japonicum,¹³⁸ P. oleracea,¹²⁹ and garlic¹³⁰ provide robust protection against experimental colitis by suppressing mucosal inflammation and reinforcing the intestinal barrier through modulating the gut microbiota. Additionally, bitter melon-derived NPs¹¹⁹ offer specialized benefits in cardioprotection, where they stabilize mitochondrial integrity and activate antioxidant signaling to mitigate damage associated with chemotherapy-induced cardiotoxicity. Together, these findings illustrate how common vegetables serve as a source of bioactive particles that support both systemic and localized homeostasis.

Herbs and Medicinal Plants

Medicinal herbs and perennial plants provide a rich repertoire of specialized metabolites and signaling molecules encapsulated within NPs for treating complex systemic diseases. These medicinal plants represent the most extensively researched category of PELNs due to their complex phytochemical profiles and long history of clinical use in traditional medicine systems. In cancer immunotherapy, vesicles from ginseng,¹⁴⁴ Brucea javanica,¹⁴⁷ and Catharanthus roseus¹³¹ can attenuate immune evasion or reverse chemotherapy-induced immunosuppression by modulating T-cell activity and macrophage polarization. Beyond oncology, ginseng-derived exosome-like NPs¹²⁴ demonstrated significant potential in the management of autoimmune conditions such as rheumatoid arthritis. In such models, these vesicles act as orally active immunomodulatory agents that preferentially accumulate in inflamed joints to reduce arthritis severity and attenuate joint inflammation. Their therapeutic effects also include protection against bone erosion in collagen-induced arthritis models, suggesting that orally administered PELNs can effectively regulate systemic inflammatory responses beyond the GI tract.

The therapeutic reach of herbal PELNs also extends to the central nervous system, as NPs from Houttuynia cordata¹¹³ and Panax notoginseng¹³² can cross the blood-brain barrier to protect against ischemic injury and ferroptosis. For inflammatory and metabolic disorders, Andrographis paniculate,¹²⁶ mulberry bark,¹³⁴ and A. argyi¹⁵³ demonstrate strong colon-targeted effects that restore the gut microbial balance and suppress ulcerative colitis. Furthermore, specialized vesicles from goji berries¹²⁷ support skeletal health by promoting osteogenesis and fracture repair, while those of mango ginger,¹⁴⁵ and Phellinus linteus¹¹⁴ provide dermatological protection against diabetic wounds and skin photoaging. Nutritional support is also provided by fenugreek¹¹¹ and Trifolium pratense,¹⁴⁶ which enhance the bioavailability of essential minerals and improve glucose homeostasis.

Beyond these systemic effects, medicinal herbs offer specialized protection for pulmonary and neurological health through unique molecular cargos and site-specific delivery. Fresh Platycodon grandiflorum-derived exosome-like NPs¹³³ serve as a key bioactive component that alleviates acute lung injury by suppressing proinflammatory cytokines and enhancing anti-inflammatory signaling. Similarly, the antiviral potential of Houttuynia cordata¹¹⁵ vesicles is closely linked to their sRNA cargo which targets both viral genes and host signaling pathways to interfere with replication while simultaneously regulating host antiviral responses. These vesicles also demonstrate robust GI protection by inhibiting the NLR family pyrin domain containing 3 (NLRP3) inflammasome and remodeling the gut microbiota to restore homeostasis in experimental colitis models.¹³⁶ In the field of neurodegeneration, Lycium ruthenicum nanovesicles¹⁴⁸ protect neuronal cells from amyloid-induced toxicity by modulating the mitogen-activated protein kinase (MAPK) and phosphatidylinositol 3-kinase (PI3K) signaling axes through encapsulated miRNAs. Furthermore, goji berry-derived nanovesicles^135,153 provide localized bioactivity for mitigating muscle wasting by promoting mitochondrial metabolism and muscle preservation under catabolic conditions.

Cereals/Grains

Cereals and grains contain bioactive nanostructures that bridge fundamental nutrition with advanced tissue engineering and regenerative applications. Research into cereal-derived vesicles, particularly those from wheat,¹⁵² highlights their significant pro-regenerative activity in skin repair. These NPs are readily internalized by skin-related cell types to enhance cell viability and collagen expression while promoting angiogenesis without disrupting normal cell cycle progression. Such findings indicate that even common dietary grains contain bioactive nanostructures capable of coordinating the complex cellular processes required for cutaneous regeneration and ECM remodeling.

Leaves/Tea/Botanical Leaves

Leaves and tea-based extracts are prominent sources of nanovesicles that modulate the gut-brain signaling axis and provide robust antioxidant protection for epithelial tissues. Tea leaf-derived exosome-like NPs have emerged as versatile therapeutic agents with unique effects on both tumor suppression and stress-induced intestinal disorders. Oral administration of these vesicles can fortify the intestinal barrier against psychological stress¹⁴⁰ and alleviate symptoms of irritable bowel syndrome^140,141 by harmonizing neuroendocrine signaling pathways. In oncology, tea-derived nanotherapeutics retard breast tumor growth¹³⁹ by inducing mitochondrial apoptosis and reshaping the intestinal microbiota to create a tumor-suppressive environment. Beyond internal applications, these vesicles also function as effective topical agents for reducing skin hyperpigmentation by modulating melanogenesis-related signaling axes and enhancing the dermal collagen content.¹²¹

Other Plant-Derived Products

Beyond primary dietary sources, various botanical products such as flowers and seeds offer specialized nanovesicles with unique tissue-targeting and regenerative properties. Other botanical sources, including flowers and seeds, contribute further to the therapeutic diversity of PELNs in clinical and cosmetic applications. Lavender-derived¹⁴² and hemp seed-derived¹⁵⁵ NPs exhibit specialized regenerative effects in dermatological protection and follicular growth, offering natural alternatives for treating photoaging and androgen-induced hair loss. Additionally, orally administered vesicles from Robinia pseudoacacia flowers¹²² preferentially accumulate in the GI tract to protect the gastric and small intestinal mucosa against hypoxia-induced injury. These flower-derived nanostructures effectively maintain the intestinal barrier by providing direct cytoprotection and stabilizing the local microenvironment. Notably, oral delivery of tea flower vesicles¹⁴³ achieves significant anti-metastatic effects while avoiding the systemic toxicity often associated with intravenous administration, highlighting the importance of delivery routes in the development of plant-based nanomedicines.

Integration of Botanical Diversity and Functional Mechanisms

The evidence presented throughout this section highlights a fundamental convergence between the diverse botanical origins of these NPs and their shared therapeutic targets in mammalian systems. While the preceding subsections emphasized the unique pharmacological contributions of specific herbs and fruits and vegetables, the systematic organization in Table 2 reveals how these distinct vesicles often operate through common regulatory axes. However, the strength of the supporting evidence varies considerably across studies. For example, the biological activities of black nightshade berry- and wheat-derived nanoparticles have primarily been investigated in cell-based models, whereas studies involving mulberry bark- and bitter melon-derived nanoparticles incorporated genetic or pathway-disruption approaches to provide stronger mechanistic evidence.^{119,134,150,152} Nevertheless, variations in experimental models and particle characterization complicate direct comparisons of therapeutic outcomes. Differences in the use of appropriate experimental controls limit the generalizability of these findings. This synthesis suggests that plant-derived nanoparticles from various botanical sources may influence conserved pathways, including the immune-inflammatory axis, the gut-microbiota and bone-homeostasis axis, and pathways associated with oxidative stress and cellular homeostasis. By bridging the source-based narrative with a mechanism-centered framework, this review highlights the potential of PELNs as a multi-component, multi-target therapeutic approach, while recognizing that further mechanistic and preclinical validation is required. Such an integrated perspective may help identify shared biological mechanisms and guide future investigations into the therapeutic relevance of PELNs from diverse botanical sources.

Potential Role of PELNs in Osteoporosis

In the preceding sections, we comprehensively reviewed applications of PELNs across a broad range of biomedical fields (Table 2), including immune regulation, inflammation, metabolic disorders, cancer therapy, neurodegenerative diseases, aging, as well as tissue repair and regeneration. Accumulating evidence suggests that PELNs exert multifaceted biological effects by modulating cellular signaling, intercellular communication, and tissue homeostasis. Notably, recent studies have begun to reveal a previously underappreciated role of PELNs in regulating bone metabolism and osteoporosis.¹⁶⁸ Osteoporosis is a prevalent and progressive skeletal disorder, particularly affecting elderly populations,¹⁶⁹ in which the capacity for tissue repair and regeneration is markedly compromised.¹⁷⁰ Given the intrinsic regenerative limitations associated with aging, the emerging application of PELNs in osteoporosis not only expands their therapeutic scope but also highlights their potential as a novel, biocompatible strategy for restoring bone homeostasis.¹⁷¹

Bone homeostasis is maintained through a tightly regulated balance between bone formation and bone resorption, processes primarily mediated by osteoblasts and osteoclasts, respectively. Osteoblasts are responsible for ECM synthesis and mineralization, whereas osteoclasts degrade mineralized bone to facilitate remodeling and calcium homeostasis. Under physiological conditions, the coordinated activity of these two cell types ensures skeletal integrity and adaptability.¹⁷² However, osteoporosis arises when this balance is disrupted, resulting in net bone loss and a compromised bone microarchitecture. In elderly populations, this imbalance is driven by multiple age-associated factors, including a reduced osteogenic differentiation capacity of mesenchymal stem cells, increased osteoclast activity, chronic low-grade inflammation, hormonal changes, and impaired cellular communication within the bone microenvironment.¹⁷³ Additionally, aging is accompanied by declines in tissue repair and regenerative potential, further exacerbating bone fragility and delaying recovery following injury.¹⁷⁴ These multifactorial alterations collectively shift bone remodeling toward resorption-dominant states, ultimately leading to osteoporosis. Given the central role of intercellular communication and microenvironmental regulation in age-related bone loss, PELNs, which possess intrinsic bioactivity and a signaling-modulatory capacity, have garnered increasing attention as promising candidates for osteoporosis interventions (Figure 2).

Figure 2.

Roles of PELNs in the treatment of osteoporosis. Under physiological conditions, bone homeostasis is maintained through a dynamic balance between osteogenesis mediated by osteoblasts and osteoclastogenesis driven by osteoclasts. In osteoporosis, this balance is disrupted, characterized by reduced osteogenic activity and enhanced osteoclastogenesis, leading to net bone loss. Emerging evidence indicates that PELNs derived from selected plant sources have been reported to modulate key signaling pathways to promote osteoblast differentiation and activity and/or suppress osteoclast formation and function, potentially contributing to the regulation of bone homeostasis. The illustrated mechanisms are based on findings from specific plant sources and experimental models and should not be interpreted as universal effects of PELNs. Created in BioRender. Chen, (Y) (2026) https://BioRender.com/ohqjbe3.

Plum-derived exosome-like nanovesicles exert a dual regulatory effect by promoting osteogenic differentiation while suppressing osteoclast formation, as reflected by reduced tartrate-resistant acid phosphatase (TRAP)-positive multinucleated cells and downregulation of osteoclast related markers, including peroxisome proliferator-activated receptor γ (PPARγ), nuclear factor of activated T-cells, cytoplasmic 1 (NFATc1), and c-Fos.¹²³ Similarly, ginseng-derived nanovesicles primarily act on the bone resorption side of remodeling by inhibiting osteoclast differentiation and reducing inflammatory bone loss, leading to improved bone volume, increased bone mineral density, and decreased calvarial bone resorption in lipopolysaccharide-induced bone loss models.¹²⁵

In contrast, yam- and apple-derived nanovesicles appear to exert predominantly pro-osteogenic effects. Oral administration of yam-derived nanovesicles improved longitudinal bone growth, bone mineral density, and osteoblast-associated parameters in mice with ovariectomy-induced osteoporosis, with efficacy comparable to estradiol treatment.¹²⁰ Apple-derived nanovesicles also enhanced osteoblast proliferation and osteogenic differentiation, and a biodistribution analysis further suggested that these vesicles can be absorbed through the GI tract and reach bone tissues after oral administration.¹¹⁷ Beyond direct regulation of bone cells, Pueraria lobata-derived exosome-like nanovesicles improved bone formation in ovariectomized rats through mechanisms associated with autophagy activation and gut microbiota modulation. Their ability to reduce harmful bacterial populations and attenuate trimethylamine N-oxide (TMAO)-associated bone resorption further highlights the potential involvement of the gut bone axis in PELN-mediated osteoporosis protection.¹¹²

Although these studies collectively suggest a role for PELNs in bone regulation, their experimental approaches differ substantially. For example, ginseng-derived nanovesicles were evaluated in an inflammation-associated bone loss model, whereas yam- and Pueraria lobata-derived nanovesicles were investigated in ovariectomy-induced osteoporosis models.^112,120,125 In contrast, the evidence for plum- and apple-derived nanovesicles primarily supports their effects on osteogenic or osteoclast-related cellular processes.^117,123 These differences in disease models, administration routes, and experimental endpoints limit direct comparisons of therapeutic efficacy across plant sources. Nevertheless, the reported findings suggest that PELNs may influence bone remodeling through distinct mechanisms, including the promotion of osteoblast differentiation and mineralization, suppression of osteoclastogenesis, and modulation of gut microbiota-associated metabolic pathways. Although certain studies have reported gastrointestinal uptake, tissue distribution, and preliminary safety findings, these observations remain insufficient to establish consistent oral bioavailability or long-term safety across different PELN preparations. Further investigations using standardized experimental protocols and rigorous mechanistic validation are therefore necessary to clarify their therapeutic potential in osteoporosis.

Discussion

In this review, we summarized recent advances in plant-derived exosome-like nanovesicles (PELNs), with particular emphasis on studies published during the past 3 years. Collectively, these studies support the concept that PELNs are biologically active nanostructures rather than passive carriers, capable of mediating intercellular and even interspecies communication through their intrinsic cargos of lipids, proteins, RNAs, and plant-derived secondary metabolites. By integrating concepts from plant biology, EV research, and nanomedicine, PELNs have emerged as a promising therapeutic platform with potential applications in inflammation, metabolic disorders, neurodegeneration, cancer, tissue repair, and regenerative medicine (Table 2). Among these indications, osteoporosis appears to be a particularly promising target. PELNs derived from diverse plant sources were reported to enhance osteoblast activity, suppress excessive osteoclastogenesis, and thereby restore bone homeostasis. In addition, orally administered PELNs may influence the gut-bone axis through modulating the gut microbiota and its metabolites (Figure 2), suggesting potential advantages for treating age-related osteoporosis, a multifactorial disorder that is often insufficiently controlled by single-target therapies.

Despite these encouraging findings, several important limitations remain insufficiently addressed in the current literature. A major challenge is the lack of standardization in the isolation and characterization as well as cargo profiling of these vesicles which substantially limits reproducibility among studies. The comprehensive synthesis of data in Table 1 highlights considerable variation in reported particle sizes across different plant sources and isolation methods. Different isolation approaches frequently yield vesicle preparations with markedly different purities, particle size distributions, and molecular compositions. Consequently, biological effects reported for a given plant source may reflect methodological differences as much as intrinsic properties of the PELNs themselves.

As discussed in the isolation section, different preparation strategies involve trade-offs among yield, purity, vesicle integrity, and scalability (Figure 1). These methodological differences should be considered when interpreting biological findings and selecting preparation workflows for future therapeutic development. For future pharmaceutical development, a combined workflow incorporating initial concentration by ultrafiltration or ultracentrifugation followed by density-gradient purification or SEC might provide the best balance between yield, purity, and scalability.

Importantly, methodological variability has also contributed to substantial inconsistency in the reported sizes of PELNs. Even when the same plant source and similar isolation protocols are used, different groups frequently report vesicles spanning a broad size range. Interestingly, many of these distinct preparations still exhibit therapeutic activity. This raises an unresolved but highly important question. Which particle population actually mediates the observed biological effects? It remains unclear whether the active fraction corresponds to a specific vesicle subtype, a broader mixture of nanoscale particles, or even non-vesicular contaminants that co-purify during isolation. Some studies have attempted to address this uncertainty by comparing PELN preparations with individual bioactive compounds. For example, ginseng-derived nanovesicles exhibited stronger inhibitory effects on osteoclastogenesis than isolated ginsenosides Rb1 and Rg1, suggesting that additional vesicle-associated components may contribute to their biological activity.¹²⁵ Nevertheless, such comparisons do not establish whether the observed effects depend on intact vesicles or other components within the isolated nanoparticle fraction. Future studies should therefore incorporate appropriate controls, such as vesicle-depleted fractions, purified free-compound controls, and experiments assessing the effects of vesicle disruption, to distinguish vesicle-dependent mechanisms from the activities of co-isolated substances. From a translational perspective, distinguishing vesicle-dependent activity from the effects of co-isolated components is critical because future manufacturing strategies will require a clearly defined and reproducible target product. In addition, characterization of PELNs is similarly inconsistent across the literature. Most studies primarily rely on particle size and concentration measurements obtained by an NP tracking analysis or dynamic light scattering. While these methods are useful for estimating the size distribution, they are insufficient to confirm vesicular identity. Unlike mammalian exosomes, PELNs currently lack universally accepted surface markers (such as cluster of differentiation 9 (CD9), CD63, and CD81), and therefore are more appropriately described as “exosome-like” NPs. Many studies do not include complementary validation by electron microscopy, density profiling, lipid composition, membrane-associated proteins, or other biochemical analyses. As a result, it is often difficult to distinguish genuine membrane-bound vesicles from other plant-derived NPs of similar sizes. Therefore, greater standardization and mechanistic clarity will be essential before these particles can progress from an emerging experimental concept to clinically applicable medicines.

This review also highlights the favorable translational characteristics of PELNs. They can be isolated in large quantities from inexpensive and widely available plant materials, and many studies suggest that oral administration produces minimal acute toxicity. In contrast to many synthetic nanomaterials or mammalian cell-derived exosomes, PELNs appear to relatively efficiently cross the GI barrier and may therefore provide a convenient route for non-invasive drug delivery. In addition, increasing commercial usage in topical PELN-based formulations for skin and hair applications suggests that these NP are generally well tolerated when applied locally. However, the apparent safety of PELNs should not be interpreted as definitive evidence of long-term safety following systemic or oral administration. Because PELNs contain highly heterogeneous mixtures of RNAs, metabolites, proteins, and other plant-derived compounds, whether chronic exposure could influence normal physiology in unexpected ways remains uncertain. In particular, plant miRNAs and secondary metabolites may potentially interact with mammalian signaling pathways beyond the intended therapeutic target. At present, few studies have examined the long-term biodistribution, pharmacokinetics, metabolism, or clearance of PELNs after repeated administration. Accordingly, evaluating the therapeutic potential of PELNs requires an integrated assessment linking botanical source, isolation and purification methods, particle identity and purity, and molecular cargo to the observed biological mechanisms and outcomes. Differences in dose normalization, exposure, and biodistribution must also be considered when comparing findings across experimental models. Integrating these parameters with safety assessments and manufacturing feasibility may provide a more reliable basis for evaluating the reproducibility and translational potential of PELN preparations.

Conclusion

PELNs represent a naturally assembled and biologically active nanoplatform that connects edible and medicinal plants with modern nanomedicine. Their diverse cargoes may contribute to multi-component and multi-target regulation. Emerging preclinical evidence also suggests their potential in osteoporosis, although clinical efficacy has not been established. Their natural abundance, potential biocompatibility, and reported gastrointestinal stability in selected preparations suggest opportunities for developing PELNs as oral nanotherapeutics. However, these properties and the feasibility of scalable production require further validation across different plant sources and isolation methods. Major challenges to clinical translation include inconsistent isolation and characterization methods, uncertain particle identity and purity, the absence of universally accepted markers, and insufficient cargo characterization. These limitations also complicate the attribution of biological effects specifically to vesicular components. In addition, pharmacokinetic, biodistribution, and long-term safety data remain limited. Standardized mechanism-based potency assays, dose normalization strategies, and batch-release criteria are also needed to improve cross-study comparability and ensure reproducible product quality. Together, these efforts will be essential for advancing PELNs toward reproducible and clinically applicable nanomedicines.

Funding Statement

This work was supported by grants from the National Science and Technology Council (grant nos.: NSTC113-2320-B-038-052-MY3 and NSTC114-2320-B-038-035-MY3), Taiwan.

Author Contributions

All authors made a significant contribution to the work reported, whether that is in the conception, study design, execution, acquisition of data, analysis and interpretation, or in all these areas; took part in drafting, revising or critically reviewing the article; gave final approval of the version to be published; have agreed on the journal to which the article has been submitted; and agree to be accountable for all aspects of the work.

Disclosure

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