Plant-Derived Nanovesicles for Ischemic Stroke Therapy via the Gut Microbiota-Gut-Brain Axis: A New Paradigm of Systemic Regulation

Introduction: From Localized Intervention to Systemic Regulation – The Evolution of the Therapeutic Paradigm for Ischemic Stroke

Ischemic stroke (IS), with its high incidence, disability, and mortality rates, imposes a heavy global disease burden.¹ This burden is expected to continue increasing with the intensification of population aging.² For decades, IS therapeutic research has primarily focused on early vascular recanalization (eg., thrombolysis, thrombectomy)³ and neuroprotection.⁴ However, neuroprotective strategies are often limited by the natural barrier of the blood-brain barrier (BBB), leading to low delivery efficiency for most candidate drugs.^5,6 Concurrently, the post-IS pathological process involves a complex interwoven network of mechanisms including excitotoxicity,⁷ oxidative stress,⁸ neuroinflammation,^9,10 and apoptosis,¹¹ rendering single-target interventions often minimally effective. Therefore, there is an urgent need for a new strategy capable of systemically and multi-targetedly regulating this disease network.

The burgeoning concept of the “gut microbiota-gut-brain axis” provides a theoretical breakthrough for this need,¹² particularly in the pathophysiology of neurological diseases like IS.¹³ Critically, this axis is bidirectional. IS not only damages the brain but also rapidly induces gut dysbiosis and increased intestinal permeability, leading to a “leaky gut”.^14–16 This, in turn, facilitates the translocation of bacteria and bacterial products into the systemic circulation, exacerbating peripheral and central neuroinflammation, thereby creating a vicious cycle that worsens stroke outcomes.^17,18 The gut, as the body’s largest immune and endocrine organ, engages in continuous bidirectional communication with the Central Nervous System(CNS) through neural, endocrine, immune, and metabolic pathways.^19–21 The gut microbiota and its metabolites can profoundly influence systemic immune status and the structural and functional repair of the brain’s neurovascular unit (NVU).^22,23 This suggests that intervening at the gut, a “peripheral hub,” may remotely and systemically modulate the brain’s pathological environment, opening a new front for IS treatment.

In recent years, Plant-Derived Nanovesicles(PDNVs), a class of natural nanocarriers derived from plants that are morphologically and functionally similar to mammalian exosomes, have emerged prominently. They are nanoscale lipid bilayer vesicles secreted by plant cells,^24–26 capable of crossing the BBB,²⁷ and naturally loaded with various bioactive components from their parent plants, including proteins, lipids, functional RNAs, and metabolites,^28,29 playing a key role in intercellular communication. Compared to synthetic nanocarriers, PDNVs possess inherent low immunogenicity, excellent biocompatibility,^30,31 and unique intrinsic bioactivity.^32–34 In contrast to mammalian exosomes, which face substantial challenges in oral delivery due to their susceptibility to degradation in the harsh gastrointestinal environment,^35,36 PDNVs exhibit remarkable stability. Their robust lipid bilayer structure enables them to resist enzymatic digestion and maintain structural integrity under extreme pH conditions, facilitating effective transit through the stomach and intestine.^36,37 This inherent stability, combined with their nanoscale size and biocompatibility, positions PDNVs as particularly well-suited for oral administration—a critical advantage for engaging the gut microbiota-gut-brain axis in a non-invasive and sustained manner.³⁸ More importantly, their multi-component synergistic nature reflects the essence of traditional Chinese medicine’s “holistic view” and “compound formula” multi-target synergistic treatment.^39,40

Therefore, breaking through the bottlenecks in IS treatment requires an innovative strategy capable of systemic regulation, multi-target intervention, and effective BBB traversal. PDNVs, with their unique attributes, offer an ideal vehicle for achieving this systemic therapy via the “gut microbiota-gut-brain axis.” This article aims to demonstrate that PDNVs are not merely drug “delivery tools” but are “functional messengers” capable of performing ordered, multi-level systemic regulation through the gut-brain axis. However, to argue for the rationale and urgency of this novel strategy, it is first necessary to fully understand the complexity of IS itself and the fundamental limitations of existing treatment paradigms.

IS: Disease Burden, Complex Mechanisms, and Therapeutic Bottlenecks

This section will systematically elaborate the urgency of developing new therapies and the potential entry points for PDNVs from three levels: epidemiology, pathophysiology, and current treatment dilemmas (Figure 1).

Figure 1 Global burden, complex mechanisms, and therapeutic bottlenecks of ischemic stroke (IS). (A) Geographical disparities in the global prevalence of IS, highlighting its pervasiveness as a major public health issue. Reprinted with permission from ref.2 Copyright (2025) Springer Nature. (B) Core network of the “ischemic cascade” triggered by the energy crisis after IS, including multiple positive feedback loops such as excitotoxicity, oxidative stress, organelle dysfunction, neuroinflammation, and cell death. The simplified workflow illustrates the progression from energy failure to blood‑brain barrier disruption. Reprinted with permission from ref.,41 Copyright (2022) American Heart Association. (Black Arrows: Describe conventional biological processes and signal transduction. Red Arrows: Highlight pathological states. Cyan Arrows: Marking the pathological process of calcium overload).; (DAMP: damage‑associated molecular pattern; IFN: interferon; IL: interleukin; MMP: matrix metalloproteinase; Th: helper T cell; TNF: tumor necrosis factor). (C) Inherent limitations of current mainstay treatments regarding time window, hemorrhage risk, BBB penetration, and multi-target regulation, emphasizing the urgency for developing a new paradigm.

Global Epidemiology and Socioeconomic Burden

According to the latest report in The Lancet Neurology, among non-communicable diseases, stroke remains the world’s second-leading cause of death and the third-leading cause of death and disability combined (measured in disability-adjusted life-years lost - DALYs). By 2021, the global number of stroke survivors reached 93.8 million, with 11.9 million new stroke cases.⁴² It is estimated that the total global cost of stroke exceeds $890 billion, accounting for 0.66% of global GDP.⁴³ IS, as the most common form of stroke, constitutes approximately 65.3% (62.4–67.7%).⁴² Currently approved vascular reperfusion therapies (eg., intravenous thrombolysis) not only have a narrow therapeutic time window but also carry risks of inducing cerebral hemorrhage and secondary injury to ischemic tissue, with significant variability in patients’ long-term functional outcomes.⁴⁴ Given the time window limitations and the current lack of definitively effective interventions for ischemia/reperfusion injury or neuroprotective strategies, the therapeutic bottleneck for IS is stark.

The Complex Network Mechanism of Progressive Injury

IS initiates a complex, multi-stage pathological cascade. The immediate energy crisis—characterized by ATP depletion and ionic homeostasis disruption—triggers glutamate excitotoxicity and pathological Ca²⁺ overload.^7,45–48 This Ca²⁺ overload acts as a central signaling hub, simultaneously inducing mitochondrial dysfunction, reactive oxygen species burst, and endoplasmic reticulum stress, forming a self-amplifying “iron triangle” of organelle injury that drives the energy crisis toward inflammation and cell death.^49–55

These molecular disturbances subsequently activate systematic neuroinflammation. Damage-associated molecular patterns released from necrotic neurons and damaged mitochondria are recognized by pattern recognition receptors on microglia and astrocytes, activating NF-κB and other pro-inflammatory pathways.^56–59 This drives glial polarization toward pro-inflammatory M1/A1 phenotypes and promotes the release of TNF-α, IL-1β, and other inflammatory mediators.^10,60–62 Simultaneously, peripheral immune cells infiltrate the ischemic brain,^63,64 forming a positive feedback loop with mitochondrial damage that perpetuates injury.^65,66

A core consequence of this cascade is disruption of the BBB and NVU. Pro-inflammatory cytokines downregulate endothelial tight junction proteins, while activated microglia and infiltrating leukocytes secrete matrix metalloproteinases (MMPs) that degrade the basement membrane.^67–74 The resulting BBB breakdown induces vasogenic edema and allows uncontrolled entry of plasma proteins and inflammatory cells into the brain parenchyma, further exacerbating cell death and creating a positive feedback loop that amplifies injury.^70,75

Advances and Core Limitations of Current Treatment Modalities

Acute IS treatment plans are primarily formulated based on time since onset, severity of neurological deficit, and imaging findings. For eligible patients, standard treatments include intravenous thrombolysis and mechanical thrombectomy.⁷⁶ The most widely used thrombolytic drug, alteplase, has an extremely short half-life (4–6 minutes) and must be administered intravenously within 4.5 hours of symptom onset.^77,78 Although alteplase reduces the risk of disability at 3 months by approximately 30%, the recanalization rate with thrombolysis alone is less than 50%.⁷⁹ Although studies have explored combination therapies to mitigate risks, no clear breakthrough has yet been achieved.^80,81 Mechanical thrombectomy is the gold standard for large-vessel occlusion IS, but its application is similarly constrained by factors such as patient age,^82,83 time window, stroke location,⁸⁴ and post-procedural complications.^85–87

Given the accessibility, time window limitations, and inherent bleeding risks of intravenous thrombolysis and thrombectomy, clinical IS management also includes medications such as anticoagulants, antiplatelets, and neuroprotective agents.⁸⁸ However, the BBB as a key barrier for drug delivery,⁸⁹ coupled with the inherent shortcomings of single-target strategies for complex network diseases,⁹⁰ are major reasons for the suboptimal efficacy of current drug therapies. In summary, the pathological mechanism of IS is a complex systemic network with multiple nodes and positive feedback loops. Any intervention targeting a single link is easily compensated for or bypassed by the network. Therefore, IS treatment urgently requires a new paradigm capable of overcoming BBB limitations and synergistically acting on multiple pathological aspects. Against this backdrop, PDNVs, which combine multi-route delivery, efficient BBB crossing, and systemic regulatory potential, are increasingly becoming a research focus, with their potential to revolutionize the therapeutic paradigm garnering widespread attention.^91–93

Beyond Delivery: PDNVs as Multifunctional Therapeutic Messengers

Exosomes are a subclass of extracellular vesicles(EVs) secreted by eukaryotic cells, featuring a phospholipid bilayer structure and a diameter in the nanometer range.⁹⁴ They appear disc-shaped or cup-shaped under transmission electron microscopy.⁹⁵ Acting as intercellular communication mediators, they can transfer bioactive substances to recipient cells.^96,97 In recent years, the application of exosomes in brain drug delivery has advanced.⁹⁸ In contrast, PDNVs, widely studied only after 2009,⁹⁹ have demonstrated higher efficiency in intracellular substance delivery and transfer due to their low immunogenicity, lack of cytotoxicity, high delivery efficiency, and good biocompatibility.¹⁰⁰ PDNVs show therapeutic potential for neurological diseases by regulating multiple pathways including calcium signaling, anti-oxidation, neuroinflammation, and apoptosis,¹⁰¹ and can also serve as novel carriers for delivering exogenous drugs.¹⁰⁰. This chapter will delve into the unique biological characteristics of PDNVs compared to traditional synthetic nanocarriers, elucidating their potential as multifunctional therapeutic messengers (Figure 2).

Figure 2 Plant-derived nanovesicles (PDNVs) as natural multifunctional nano-therapeutic messengers. PDNVs are nanoscale lipid bilayer vesicles extracted from various plant cells. They are naturally loaded with a variety of bioactive components including proteins, lipids, nucleic acids (eg., miRNAs), and plant active metabolites, constituting a multi-component synergistic system. Their unique composition confers superior stability. Upon oral administration, PDNVs resist gastrointestinal digestion and are efficiently taken up by intestinal epithelial cells, immune cells, and even gut microbiota via multiple endocytic pathways. More importantly, PDNVs exhibit good biocompatibility and low immunogenicity, and can effectively cross the blood‑brain barrier, enabling targeted delivery and regulation from peripheral sites (eg., the gut) to the central nervous system.

Natural Multi-Component Synergistic System

PDNVs are rich in biomolecules, including lipids, nucleic acids, proteins, and plant active ingredients,¹⁰² with some biological functions resembling those of animal-derived exosomes.¹⁰³ Oral administration is a common and ideal route for PDNVs, offering advantages such as non-invasiveness, safety, and high compliance, but its bioavailability is often limited and variable due to the complex gastrointestinal environment.¹⁰⁴ The excellent physicochemical properties of PDNVs (small particle size, negative charge, lipid bilayer membrane, and hydrophilic surface) enable them to effectively evade physiological barriers in the gastrointestinal tract such as extreme pH,^105,106 digestive enzymes,¹⁰⁷ and the mucus barrier,¹⁰⁶ safely target lesion sites, prolong residence time in the gut,^108,109 and achieve responsive delivery.¹¹⁰

Specifically, the lipid bilayer of PDNVs effectively protects their carried miRNAs from degradation by ribonucleases and harsh gastrointestinal conditions.²⁷ The inherent 2′-O-methylation modification of miRNAs also confers acid resistance and stability, allowing them to maintain long-term activity in the human intestinal environment.¹¹¹ The protein composition of PDNVs is mainly intracellular, including enzymes related to cell wall remodeling with potential anti-pathogen activity,¹¹² and metabolic enzymes conferring antioxidant capacity.¹¹³ Their peripheral and transmembrane proteins may be involved in vesicle formation and specific targeting.¹¹⁴ The RNA in PDNVs, especially miRNAs, is believed to play roles in intercellular and even cross-species gene regulation.¹¹⁵ Beyond their structural role, specific lipid species in PDNVs actively mediate interactions with host cells and gut microbes, determining tissue tropism and cellular uptake efficiency.^116–118 For instance, phosphatidic acid, a key lipid component in ginger-derived PDNVs, is preferentially recognized and internalized by Lactobacillaceae in a lipid-dependent manner, facilitating targeted delivery of vesicle-encapsulated miRNAs to specific bacterial taxa.¹¹⁷ Similarly, galactolipids such as digalactosyldiacylglycerol, abundant in oat PDNVs, play a crucial role in their interaction with microglial cells and contribute to their anti-neuroinflammatory effects after crossing the BBB.¹¹⁹ Grapefruit-derived PDNVs, which are rich in phosphatidylcholine, preferentially accumulate in the liver and spleen following systemic administration, demonstrating the role of lipid composition in dictating tissue tropism.¹²⁰ These examples illustrate that lipid composition is not merely a structural determinant but actively dictates PDNV targeting, uptake, and biological function.

Furthermore, differences in the lipid composition of PDNVs determine their different functional and distribution tendencies. For example, specific phospholipids may facilitate their accumulation in the intestine or promote transport to the liver and exert varying attraction to different gut microbes.^116–118 Homologous plant active molecules contained in PDNVs can also be delivered into the body to produce corresponding biological effects.^121,122 These biomolecules are not randomly combined but form a therapeutically functional unit with intrinsic connections.

Source-Determined Functional Programming

The functional properties of PDNVs exhibit significant “parental memory.” Compared to plant crude extracts or single compounds, PDNVs are rich in numerous endogenous active components, possess higher lipid solubility, can significantly promote drug absorption and membrane permeability, increase drug concentration in target organs, thereby enhancing efficacy.¹²³ PDNVs have been confirmed to possess various effects such as anti-inflammatory, anti-tumor, immunomodulatory, and antioxidant stress.^124,125 Simultaneously, PDNVs from different sources demonstrate specific therapeutic effects on different disease systems.^123,126 For example, ginseng, known for its anti-inflammatory, antioxidant, and anticancer properties, yields PDNVs reported to improve symptoms related to dementia, diabetes, respiratory infections, and cancer^127,128 (more examples summarized in references).^123,126,129 This aligns closely with the traditional efficacy classification of Chinese medicinal herbs, providing a rational basis for screening and applying specific PDNVs based on therapeutic goals.¹³⁰

Excellent Biocompatibility, Stability, and BBB-Crossing Ability

The excellent biocompatibility of PDNVs is primarily manifested as low immunogenicity and potential natural targeting. They are free of zoonotic or human pathogens,¹³¹ suitable for oral administration, non-toxic to healthy tissues, and possess outstanding biocompatibility.^109,132 Their potential natural targeting is often attributed to their unique phospholipid bilayer structure and specific biomolecules they carry, which may confer tissue tropism, ensuring therapeutic components accumulate at intended sites.^129,133 For instance, grape PDNVs rich in phosphatidylcholine preferentially accumulate in liver tissue;¹³⁴ ginger PDNVs rich in phosphatidic acid can be preferentially taken up by Lactobacillaceae in a lipid-dependent manner and induce IL-22 production via miRNA delivery, alleviating colitis.¹¹⁷ Ceramide lipids in ginseng PDNVs can promote macrophage M2 polarization via the TLR4/MyD88 pathway, inhibiting tumor growth.^135,136 Garlic PDNVs express lectins that specifically bind to the CD98 receptor on hepatocytes, promoting targeted uptake and anti-inflammatory responses.¹¹⁴ Coffee PDNVs exert therapeutic effects on chronic liver disease through their miRNA-mediated regulation of related genes.¹³⁷ Broccoli PDNVs rich in the secondary metabolite glucoraphanin can activate AMP-activated protein kinase (AMPK) and target dendritic cells (DCs), thereby alleviating colitis.¹²¹ Further elaboration on the natural targeting potential of PDNVs can be found in reference.¹³⁸

The particle size of PDNVs typically ranges from 10 to 1000 nanometers, with morphology similar to animal exosomes.¹⁰⁰ Compared to the latter, PDNVs demonstrate superior biostability, mainly due to their robust lipid membrane, enabling resistance to enzymatic degradation and maintenance of structural integrity in gastric and intestinal fluids, achieving long circulation.¹³⁹ Studies show that ginger PDNVs can maintain integrity in extreme pH environments.¹⁴⁰ Their nanoscale size also grants them inherent potential to penetrate the BBB, intestinal barrier, and skin barrier.¹⁴¹ This stability is crucial for overcoming the challenges of drug delivery via the gastrointestinal tract and BBB traversal.^142,143 Recent studies report that PDNVs from various sources, such as Panax notoginseng,¹⁴⁴ grapefruit,¹⁴⁵ and celery seed,¹⁴⁶ can effectively penetrate the BBB, increase their distribution in the brain, and exert neuroprotective effects. Table 1 (located at the end of the text) compiles the effects of different administration routes of representative PDNVs on their BBB-crossing efficiency and mechanisms of action.

Table 1 BBB-Crossing Efficiency and Main Mechanisms of Action of Representative PDNVs

To further leverage the “intervention + delivery” advantages of PDNVs, drug delivery systems based on PDNVs have become a research hotspot.¹⁵³ Compared to synthetic lipid carriers, PDNVs are not only non-cytotoxic but are also taken up by cells more rapidly.¹²⁰ Even at high doses, their cellular uptake efficiency remains high.¹⁵⁴ In summary, as a natural bionanosystem shaped by evolution, the core advantage of PDNVs lies in their ability to intelligently respond to complex internal microenvironments, achieving protective delivery, targeted enrichment, and conditional release of their cargo. This “friendly and intelligent” interactive ability with host biological systems constitutes the physical and biological foundation for their effectiveness as “functional messengers” capable of traversing multiple biological barriers and implementing multi-level, systemic regulation via the gut-brain axis.

However, despite these compelling advantages, several limitations must be acknowledged. The targeting efficiency of unmodified PDNVs to specific brain lesions remains modest, and their multi-component nature, while synergistic, also raises the possibility of off-target effects from bioactive molecules that are not therapeutically relevant. Furthermore, the challenges in large-scale, good manufacturing practice-compatible production and batch-to-batch consistency, as discussed later, currently hinder their clinical translation.^{129,138,155–157}

Targeting the “Gut Microbiota-Gut-Brain Axis” for IS Therapy: The Systemic Therapeutic Pathway of PDNVs

Within 24 hours after IS onset, gut dysbiosis and increased intestinal mucosal permeability (ie., “leaky gut”) can occur, subsequently inducing endotoxemia and bacterial translocation.^16,158–160 Concurrently, the dysregulated microbiota and compromised gut barrier jointly alter the signaling pattern of microbial metabolites to the brain. These metabolites are now regarded as potential biomarkers and important pathophysiological mediators for IS.^161,162 Based on this, this chapter constructs a three-stage continuous action model of PDNVs treating IS via the “gut microbiota-gut-brain axis,” systematically elaborating this new paradigm of multi-level, holistic regulation (Figure 3).

Figure 3 Three-stage model of PDNV‑mediated systemic therapy for ischemic stroke via the gut-brain axis. Stage 1 (Gut Remodeling): Orally administered PDNVs are taken up by intestinal epithelial cells, immune cells, and specific microbes. They collectively reshape a stable, anti inflammatory gut microenvironment “bridgehead” by repairing tight junctions, modulating immune cell phenotypes (eg., macrophage M2 polarization), and precisely modulating microbial structure (increasing beneficial bacteria, suppressing harmful ones). Stage 2 (Signal Amplification): Two parallel pathways operate. Direct pathway: a fraction of intact PDNVs are absorbed via the mesenteric lymphatic system and portal vein, enter the systemic circulation, cross the blood‑brain barrier, and act as “direct messengers” delivering native bioactive cargo (eg., miRNAs, proteins, lipids) to brain parenchyma. Indirect pathway: most PDNVs remain in the gut, where they promote the production of short-chain fatty acids, bile acid metabolites, indole derivatives, and DHA, which enter the bloodstream as a “humoral signal tide” and act as “indirect modulators” without requiring the vesicles themselves to reach the brain. Stage 3 (Central Repair): The above signals synergistically reprogram brain microglia/macrophages toward the M2 phenotype (anti‑inflammatory), shift astrocytes toward the neuroprotective A2 type, support oligodendrocyte precursor cell survival and remyelination, restore blood-brain barrier integrity, and enhance neuronal survival and synaptic plasticity. This ultimately achieves the structural and functional holistic reconstruction of the neurovascular unit.

To better illustrate the unique advantages of PDNVs, Table 2 provides a comparative analysis with synthetic liposomes and mammalian exosomes across key parameters including cost, stability, scalability, BBB crossing efficiency, and clinical translation status.

Table 2 Comparative Analysis of PDNVs with Synthetic Liposomes and Mammalian Exosomes

Stage 1: Remodeling the Gut Microenvironment – Establishing the “Bridgehead”

Studies show that when PDNVs are administered intravenously or intraperitoneally, they primarily accumulate in the liver and spleen; whereas upon oral administration, they preferentially distribute to the gastrointestinal tract.¹³⁵ After oral dosing, their fluorescent signals can be detected in multiple regions of the intestine, confirming effective delivery to the target site.¹⁶⁶ PDNVs are not passively decomposed in the gastrointestinal tract but actively engage in multi-level, complex interactions with the gut microenvironment, leveraging their nano-size, stability, and surface properties. This interaction begins with efficient internalization by different cells within the gut,¹⁶⁷ such as intestinal epithelial cells^139,168 and lamina propria immune cells,^139,169,170 via various endocytic pathways. These pathways include: clathrin-mediated endocytosis (a classic ligand-specific uptake pathway),¹⁷¹ caveolin-mediated endocytosis (potentially involved in trans-epithelial transport of PDNVs and can evade lysosomal degradation, protecting their bioactive cargo),¹⁷² macropinocytosis (nonspecifically and efficiently internalizing large amounts of PDNVs and surrounding extracellular fluid),¹⁷³ and phagocytosis (primarily the way professional phagocytes like gut macrophages take up PDNVs, directly related to their subsequent immunomodulatory function).^154,174 The specific pathway depends on cell type, PDNVs surface composition, and receptor-ligand interactions.¹⁷⁵ Upon internalization, PDNVs release their loaded active components, achieving cross-kingdom intercellular regulation.¹⁷⁶

Particularly important is that PDNVs can also be directly recognized and selectively internalized by gut microbes.¹¹⁸ Mechanistically, PDNVs reshape the gut microbiota through two distinct but potentially synergistic modes of action. First, PDNVs can function as “prebiotic-like” agents, serving as a metabolic substrate that selectively promotes the growth of beneficial bacterial taxa. For example, garlic PDNVs restore the abundance of beneficial Lachnospiraceae while suppressing pro-inflammatory Helicobacter genus;¹⁷⁷ lemon PDNVs enhance the intestinal survival of Lactobacillus rhamnosus and Streptococcus thermophilus ¹⁶⁹. Second, and more remarkably, PDNVs can act as “cross-kingdom regulators” by delivering functional small RNAs (eg., miRNAs) that are internalized by specific bacteria, where they can modulate bacterial gene expression. One hypothesized mechanism involves direct fusion of their lipid bilayer with the bacterial cell membrane, a process dependent on the specific lipid composition of the PDNVs.¹⁷⁵ For example, labeled garlic PDNVs were taken up by major phyla like Bacteroidetes and Firmicutes within 3 hours of oral administration.¹⁷⁸ Ginger PDNVs can be preferentially taken up by Lactobacillaceae in a lipid-dependent manner and deliver miRNAs targeting Lactobacillus rhamnosus genes.¹¹⁷ Furthermore, PDNVs and their components can serve as “ecological substrates,” metabolized and transformed by microbes into more active secondary products, representing a third layer of microbiota modulation.

Through the above composite network, PDNVs synergistically remodel the gut microenvironment from three core dimensions: First, precise modulation of gut microbiota: Different PDNVs exhibit regulatory capabilities toward specific microbes. For instance, garlic PDNVs restore the abundance of beneficial Lachnospiraceae and reduce pro-inflammatory Helicobacter genus abundance;¹⁷⁷ their miRNA can also enhance Bacteroides thetaiotaomicron abundance.¹⁷⁹ PDNVs from sources like ginseng,¹⁸⁰ fresh Rehmanniae Radix,¹⁸¹ lemon,¹⁸² pueraria lobata,¹⁸³ and tea¹⁸⁴ all demonstrate the ability to modulate microbial structure, promote beneficial bacteria, and suppress harmful ones. Second, repair and enhancement of the intestinal physical barrier: PDNVs directly promote barrier repair by acting on intestinal epithelial cells. Grape PDNVs can stimulate intestinal stem cell proliferation.¹³⁴ PDNVs from aloe,¹⁸⁵ allium tuberosum,¹⁸⁶ etc., can upregulate tight junction protein (ZO-1, Occludin) expression; tea PDNVs can promote antimicrobial peptide secretion;¹⁸⁷ mulberry root bark PDNVs protect epithelial cells via the aryl hydrocarbon receptor (AhR) pathway.¹⁸⁸ PDNVs from ginseng,¹⁸⁹ turmeric,¹⁹⁰ folium artemisiae argyi,¹⁹¹ honeysuckle,¹⁹² etc., also repair or enhance the intestinal barrier through different mechanisms. Furthermore, regulating intestinal immune homeostasis is another key aspect. PDNVs shift the gut immune state toward anti-inflammatory and reparative by influencing various immune cells within the gut-associated lymphoid tissue (GALT), including macrophages, DCs, and T cells. A core function is driving macrophages toward the M2 phenotype,¹⁹³ as seen with PDNVs from Centella asiatica,¹⁷⁵ ginger,¹⁹⁴ grapefruit,¹²² and Houttuynia cordata.¹⁶⁷ Simultaneously, PDNVs can modulate DC function; Petasites japonicus PDNVs promote their maturation,¹⁹⁵ whereas broccoli¹²¹ and Boehmeria japonica¹⁹⁶ PDNVs induce a tolerogenic phenotype, subsequently modulating T cell responses. Importantly, these locally modulated immune cells in the GALT do not remain confined to the gut. Following PDNV-mediated education, they can migrate via the mesenteric lymphatics and thoracic duct into the systemic circulation, eventually trafficking to the brain where they contribute to the resolution of neuroinflammation.^117,178 This process represents a critical link between peripheral immune modulation and central repair.

Beyond macrophages and DCs, PDNVs also directly modulate T cell populations within the GALT. For instance, Portulaca oleracea L PDNVs promote Lactobacillus reuteri growth to produce indole derivatives, activating the aryl hydrocarbon receptor (AhR) in CD4⁺ T cells and driving their reprogramming into regulatory CD4⁺CD8⁺ T cells, which exhibit potent immunosuppressive functions.¹⁹⁷ Honeysuckle PDNVs modulate T cell immune responses by restoring the Treg/Th17 balance, thereby alleviating both local and systemic inflammation.¹⁹⁸ These PDNV-educated T cells, once mobilized into the circulation, can suppress systemic inflammatory tone and limit the infiltration of pro-inflammatory immune cells into the brain. Furthermore, PDNVs can influence neutrophil dynamics, a key player in post-stroke neuroinflammation. Although direct evidence in IS models is emerging, studies in colitis models indicate that PDNVs can reduce neutrophil infiltration into inflamed tissues by downregulating chemokine expression and inhibiting neutrophil extracellular trap formation.^167,177 This suggests that PDNVs may similarly limit neutrophil-mediated damage in the ischemic brain by modulating the gut-neutrophil axis.

A particularly compelling example that bridges the two arms of the gut-brain axis comes from Momordica charantia (bitter melon)-derived PDNVs. Intravenous administration of these vesicles has been directly shown to protect against ischemic brain injury by preserving BBB integrity and inhibiting neuronal apoptosis via the AKT/GSK3β pathway.¹⁴⁷ Concurrently, oral administration of the same PDNV source effectively alleviates ulcerative colitis by inhibiting macrophage inflammation, scavenging reactive oxygen species, and protecting mitochondrial integrity in the gut.¹⁹⁹ This parallel evidence—demonstrating both direct neuroprotection and oral gut-remodeling efficacy for PDNVs from a single plant source—provides unique support for the feasibility of oral PDNV-based IS therapy via the gut-brain axis.

These three dimensions of remodeling constitute a highly synergistic, self-reinforcing network: barrier repair reduces harmful substance translocation, providing a stable environment for beneficial microbes; beneficial microbes and their metabolites support epithelial health and enhance immune tolerance; a balanced immune microenvironment in turn safeguards barrier repair and microbial homeostasis. Through this network, PDNVs reshape the post-IS dysregulated gut into a structurally intact, micro-ecologically balanced, and immunologically tolerant robust “bridgehead,” laying the foundation for the generation and dissemination of subsequent systemic therapeutic signals. Representative studies are summarized in Table 3 (located at the end of the text).

Table 3 Summary of Representative Studies on PDNVs Remodeling the Gut Microenvironment

Stage 2: Generation and Amplification of Systemic Signals – Initiating “Humoral Communication”

Following the successful establishment of the gut “bridgehead,” therapeutic signals need to be efficiently transmitted systemically and centrally. This process is achieved primarily through two parallel and synergistic pathways: direct delivery of intact vesicles and indirect mediation by gut-derived active molecules, together constituting a powerful systemic “humoral communication” network.

On the one hand, a fraction of structurally intact PDNVs are absorbed from the gut—likely a small but functionally significant portion, depending on the plant source and vesicle surface properties.^38,147 Consistent with the observation that orally administered PDNVs predominantly localize to the gastrointestinal tract,^135,139 the majority of these vesicles remain within the gut lumen and mucosa, where they exert their primary remodeling effects. Nevertheless, a detectable fraction enters systemic circulation via the mesenteric lymphatic system and portal vein, and crosses the BBB by virtue of their nanoscale size and biomembrane properties.^119,152 Studies have confirmed that intravenously injected PDNVs from sources like Panax notoginseng,¹⁴⁴ grapefruit,¹⁴⁵ and Momordica charantia¹⁴⁷ can efficiently reach the brain. More importantly, orally administered PDNVs from oats and garlic have also been detected in the brain, demonstrating that they can act as intact “therapeutic messengers” reaching brain lesions after intestinal absorption.^119,152 In this context, PDNVs act as “direct messengers”, delivering their native bioactive cargo (eg., miRNAs, proteins, lipids) directly to the brain parenchyma, where they can be internalized by neurons, microglia, and other NVU cells.¹⁴⁴

On the other hand, the majority of orally administered PDNVs remain localized within the gastrointestinal tract.^135,139 There, they engage in extensive interactions with the gut microbiota, intestinal epithelial cells, and GALT. These interactions lead to the remodeling of the gut ecosystem, including modulation of microbial composition, repair of the intestinal barrier, and induction of tolerogenic immune responses.^117,167,177 The beneficial effector molecules generated from this gut-level remodeling converge into a broad “humoral signal tide” absorbed into the bloodstream, exerting systemic regulation on distal organs without requiring the PDNVs themselves to reach the brain. Specifically, PDNVs can effectively regulate levels of key metabolites: (1) By promoting the abundance of specific SCFA-producing bacteria (eg., Lachnospiraceae, Bifidobacterium, and Lactobacillus),^117,177,186 PDNVs significantly increase intestinal production of SCFAs (eg., acetate, butyrate, and propionate).^167,186 These SCFAs are absorbed into the bloodstream, cross the BBB via monocarboxylate transporters, and exert dual protective effects: they enhance BBB integrity by upregulating tight junction proteins (eg., occludin, claudin-5, ZO-1),^200,201 and they promote microglial polarization from the pro-inflammatory M1 phenotype toward the anti-inflammatory, neuroprotective M2 phenotype, thereby reducing neuroinflammation and enhancing synaptic plasticity;^201,202 (2) Promoting bile acid absorption and modulating their metabolism;¹⁹⁸ (3) Increasing levels of immunomodulatory indole derivatives by promoting Lactobacillus reuteri growth;¹⁹⁷ (4) Through delivered aly-miR159a-3p inhibiting bacterial phospholipase C expression, leading to accumulation of the potent neuroprotective agent docosahexaenoic acid (DHA) in the body.¹¹⁸ In this scenario, PDNVs act as “indirect modulators”.^117,178

The balance between these two pathways—direct vesicle delivery versus indirect gut-derived signaling—likely depends on factors such as the PDNV source, surface lipid composition, particle size, and the integrity of the host gut barrier. Nonetheless, both pathways converge to create a systemic anti-inflammatory and pro-regenerative environment conducive to central nervous system repair.

Stage 3: Synergistic Activation of the CNS Repair Network – Achieving “Ultimate Repair”

When therapeutic signals arrive in the brain, their ultimate task is to reverse ischemic damage and initiate orderly repair. PDNVs and the systemic signals they mediate, through synergistic regulation of key cell populations within the NVU, initiate an endogenous, multi-level repair program aimed at systemically restoring the structural and functional integrity of the NVU.

Core Regulation: Phenotypic Reprogramming of Microglia/Macrophages

Activation of the repair network begins with phenotypic reprogramming of the brain’s core immune cells—microglia and infiltrating macrophages. Research shows components of orally administered PDNVs can be transported to the brain and preferentially taken up by microglia, lowering brain levels of pro-inflammatory factors like IFN-γ and TNF-α and inhibiting neuroinflammation.^178,203 In vitro experiments confirm PDNVs can be internalized by microglial cell lines, suppressing downstream inflammatory mediator release by downregulating receptors like TLR4.^204–206 In a traumatic brain injury model, intravenous injection of PDNVs inhibited excessive microglial activation and M1 polarization while reducing astrocyte activation, neuronal oxidative stress and apoptosis, and protecting dendritic structures.²⁰⁷ This process can also benefit from gut-derived signals (eg., SCFAs); circulating SCFAs can cross the BBB, synergistically promoting microglia/macrophage polarization from M1 to M2 phenotypes, playing a central role in combating neuroinflammation and enhancing synaptic plasticity.²⁰²

Synergistic Support: Reparative Responses of Astrocytes and Oligodendrocytes

The anti-inflammatory and neurotrophic microenvironment created by M2-type microglia/macrophages provides key conditions for repair in other supportive cells of the NVU. In this environment, activated astrocytes are more inclined to shift toward the neuroprotective A2 type, thereby enhancing metabolic and trophic support for neurons.²⁰¹ Metabolites like SCFAs may participate in regulating astrocyte activation via pathways such as SGK1/IL-6. Simultaneously, this favorable microenvironment also promotes the survival, proliferation, and differentiation of oligodendrocyte precursor cells, helping to reduce their loss post-ischemia, thereby supporting remyelination and remodeling in the ischemic penumbra, laying a structural foundation for restoration of neural conduction function.²⁰⁸

Integrated Outcome: Structural and Functional Reconstruction of the NVU

Building upon the synergistic action of the above cellular events, repair ultimately manifests as the holistic reconstruction of the NVU. On one hand, neuronal apoptosis in the penumbra is effectively inhibited, and synaptic plasticity is enhanced. On the other hand, PDNVs and the systemic signal molecules they elevate (represented by SCFAs) can act on cerebrovascular endothelial cells, promoting angiogenesis in the ischemic region, and restoring and consolidating BBB integrity by upregulating tight junction protein expression, among other means.^147,200 This establishes a virtuous cycle of reduced harmful substance leakage, improved local microcirculation, and support for neural regeneration, creating a sustainable homeostatic environment for the functional recovery of the entire NVU.

In summary, through direct and indirect signals transmitted via the gut microbiota-gut-brain axis, PDNVs initiate within the CNS a cascade repair network centered on immune reprogramming, supported by multi-cellular synergistic responses, and culminating in holistic NVU repair, marking the completion of their therapeutic strategy’s systemic leap from peripheral intervention to central functional reconstruction. The multi‑level actions of PDNVs along the gut‑brain axis are systematically summarized in Table 4.

Table 4 Overview of Plant-Derived Nanovesicles Mechanisms of Action Along the Gut-Brain Axis

Toward the Clinic: Translational Challenges and Prospective Strategies

Although the prospects of PDNVs for IS therapy via the gut-brain axis are exciting, the path to clinical translation remains fraught with challenges. These are not mere technical optimization issues but involve systemic hurdles spanning production paradigms, regulatory frameworks, and therapeutic concepts. This chapter aims to clarify the translational challenges of PDNVs from laboratory research to clinical application, analyze core bottlenecks, and propose forward-looking yet feasible coping strategies based on existing scientific research and regulatory developments.

Core Challenge One: The Standardization Leap from “Phytochemistry” to “Nanodrug”

The complexity of PDNVs is both their strength and the primary obstacle to industrialization.²⁰⁹ Their production currently heavily depends on plant source (variety, part, origin, growth conditions), extraction method, and post-processing techniques, leading to significant batch-to-batch variability in product characteristics such as particle size distribution, vesicle yield, biomolecular composition, and functional activity. Among these variables, the lack of a standardized, scalable, and good manufacturing practice-compatible isolation protocol represents a particularly critical roadblock. Currently, the most commonly employed methods—ultracentrifugation, size-exclusion chromatography, and polymer-based precipitation—each present distinct advantages and limitations that directly impact the final product.

Ultracentrifugation, the most widely used method in research settings, offers high purity but suffers from low yield, potential vesicle aggregation, and poor scalability.^163,209 Size-exclusion chromatography preserves vesicle integrity and maintains native bioactivity but often yields dilute samples that require concentration, increasing processing time and cost.¹⁴² Polymer-based precipitation is simple and scalable but frequently co-precipitates non-vesicular contaminants such as proteins and lipoproteins, compromising purity and potentially introducing confounding biological effects.^109,140 These methodological differences result in PDNV preparations with divergent physicochemical properties (size, zeta potential, yield) and biomolecular cargo profiles (lipid, protein, and RNA composition), which in turn affect functional outcomes—making direct comparison between studies challenging and hampering the establishment of reproducible efficacy profiles necessary for clinical development.²¹⁰ This heterogeneity not only affects experimental reproducibility but is also the “Achilles’ heel” for clinical studies and mass production.

Establish a “Quality by Design” (QbD) production framework: Drawing on experience from biologics and advanced therapy medicinal products (ATMPs), define the Critical Quality Attributes (CQAs) of PDNVs, such as: (1) Physical attributes (particle size, Zeta potential, concentration); (2) Chemical composition signature (characteristic lipid, protein, plant functional RNA, or metabolite marker profiles); (3) Functional activity (eg., ability to regulate specific cytokine secretion, selective promotion or inhibition of target microbes). Critical Process Parameters during production (eg., homogenization intensity, centrifugation force, purification column packing) must be tightly controlled to ensure consistency of CQAs.

Develop multi-omics-based characterization and quality control standards: Single physicochemical characterization is insufficient to guarantee functional consistency. Lipidomics, proteomics, small RNA sequencing, and metabolomics should be integrated to establish a “multi-omics fingerprint” for each therapeutic PDNVs, serving as the core basis for batch release. Simultaneously, develop in vitro potency assays related to in vivo efficacy (eg., testing the ability to regulate specific immune cell phenotypes in a simulated gut environment), linking chemical consistency to biological functional consistency.

Advance synthetic biology and cultivation technologies: To break free from dependence on natural plant raw materials, long-term exploration should focus on utilizing plant cell suspension cultures for controlled PDNVs production, or reconstructing “artificial plant vesicles” with specific therapeutic functions in model microorganisms via synthetic biology, enabling fully controllable design of composition and function.

Core Challenge Two: A New Regulatory Science Paradigm – How to Define and Evaluate a “Living” Nanosystem

PDNVs are neither single chemical entities nor traditional botanical drug crude extracts, but functional nanosystems with intrinsic bioactivity, complex and potentially interacting components. Existing drug regulatory frameworks (eg., FDA/EMA/NMPA) with their core review logic based on defined active ingredients are ill-suited. The fundamental question facing regulators is: when the active ingredient is a “whole” composed of hundreds of dynamically interacting molecules, how do we define its identity and assess its quality, safety, and efficacy?

Advocate for an “evidence-based batch” and “totality of evidence” assessment philosophy: Learn from regulatory experience with ATMPs (eg., stem cell products, CAR-T cells) and botanical drugs (eg., FDA’s Botanical Drug Guidance). The core of regulatory submission should not be exhaustive characterization of all components, but providing sufficient evidence demonstrating: (1) Stable manufacturing processes yielding “quality-similar” batches consistently (via multi-dimensional quality control); (2) Plausible mechanism of action, even if complex, elucidated through systems biology methods (network pharmacology, multi-omics analysis) for main pathways and networks; (3) Clear and reproducible clinical efficacy.

Construct a risk-based tiered regulatory pathway: Differentiate non-clinical and clinical study requirements based on PDNVs source (edible vs. medicinal plant), route of administration (oral vs. intravenous), degree of modification (natural vs. engineered), and therapeutic area (life-threatening disease vs. chronic management). For example, orally administered PDNVs from edible plants, based on their Generally Recognized as Safe (GRAS) background, might have simplified toxicology requirements, with focus on newly discovered pharmacological actions and long-term safety.

Strengthen early dialogue with global regulatory agencies: Researchers and industry should proactively engage in Pre-Investigational New Drug (Pre-IND) meetings with innovative drug review departments of agencies like the FDA, EMA, and NMPA to seek scientific advice on key issues such as product classification, Chemistry, Manufacturing, and Controls (CMC) requirements, and non-clinical study strategies, jointly promoting the formation of regulatory guidelines adapted to such innovative products.

Core Challenge Three: Personalized Medicine – from “One-Size-Fits-All” to “Tailor-Made”

The core of the “gut microbiota-gut-brain axis” mechanism lies in its high degree of individualization. Patients’ baseline gut microbiota composition, immune status, and metabolic phenotype vary greatly, inevitably leading to differential therapeutic responses to the same PDNVs preparation. Ignoring this heterogeneity may result in clinical trial failure or mediocre efficacy.

Discover predictive biomarkers for patient stratification: Systematically collect patients’ baseline gut metagenomic, metabolomic, and immunomic data in preclinical and early clinical studies, correlating them with treatment response (eg., improvement in neurological function scores, imaging-based repair, decrease in inflammatory markers). The goal is to identify biomarker combinations predictive of PDNVs efficacy (eg., specific microbial signatures, metabolite profiles) for precise patient enrollment in later-stage clinical trials.

Develop “enterotype-guided” PDNVs matching strategies: Based on gut microbiota composition, populations can be broadly categorized into different “enterotypes.” Future exploration could establish an “enterotype-PDNVs functional profile” correspondence. For instance, for patients with a Prevotella-dominant enterotype, match PDNVs that specifically promote their beneficial metabolic functions; for patients with extremely low microbial diversity, consider PDNVs with stronger microbiota-reshaping capabilities or combination with prebiotics.

Explore modular engineering and combination therapies: Leveraging the ease of engineering PDNVs, develop “chassis vesicles.” After clarifying core mechanisms, surface modification (targeting peptides) or cargo loading (specific siRNA, neuroprotectants) of natural PDNVs can enhance their intervention capability for specific patient subgroups or disease-critical nodes. Simultaneously, explore synergistic effects of PDNVs with existing therapies (eg., antiplatelets, statins) to form combination regimens that enhance efficacy and reduce side effects.

Future Directions: Toward Oral PDNV Formulations for IS Therapy

While the current evidence establishes a strong mechanistic foundation for PDNVs in IS therapy, direct validation of orally administered PDNVs in IS models remains a key priority. Encouragingly, recent studies have demonstrated that orally administered nanoparticles can effectively treat IS via gut-brain axis mechanisms, providing compelling proof-of-concept for this strategy. Yang et al showed that oral administration of Lactobacillus plantarum-derived EVs inhibited neuronal apoptosis in a mouse mouse Middle Cerebral Artery Occlusion(MCAO) model by delivering miR-101a-3p that targets the c-Fos/TGF-β axis.²¹¹ This study represents the first demonstration that orally administered, non-mammalian EVs can treat IS through gut-brain axis signaling. More recently, Weng et al reported that oral administration of Tianma Gouteng decoction-derived nanoparticles maintained blood-brain barrier integrity, reduced infarct area, and improved neurological function in a mouse MCAO model by regulating the S1PR1/ERK/MEK signaling axis.²¹² These findings collectively validate the feasibility of oral nanoparticle-based gut-brain axis targeting for IS therapy. Future research should prioritize the development of orally delivered PDNV formulations optimized for gastrointestinal stability, targeted gut microbiota modulation, and efficient systemic signaling to the brain.

Conclusion

PDNVs represent an innovative fusion between the “holistic view” of traditional herbal medicine and the “precision” of modern nanomedicine. Through the systems biology pathway of the gut microbiota-gut-brain axis, they operate in a multi-component synergistic, multi-target regulatory, multi-level progressive manner, dynamically reshaping pathological microenvironments and activating endogenous neural repair networks, thereby offering a novel therapeutic paradigm for complex systemic diseases like IS with intertwined mechanisms. A pivotal advantage underpinning their clinical potential is their origin from edible plants, which confers a “generally recognized as safe” status. Unlike synthetic nanoparticles, which often raise concerns regarding long-term toxicity and immunogenicity, PDNVs exhibit inherent low immunogenicity and excellent biocompatibility, positioning them as a particularly attractive candidate for clinical translation.^30,31,131 Looking ahead, research in this field should focus on the following key directions: (1) Deepening mechanistic understanding: Utilizing multi-omics, approaches, including transcriptomics, metabolomics, and metagenomics, along with single-cell technologies and spatiotemporal kinetic analyses, to finely map the dynamic network of interactions between PDNVs multi-components and the host/microbiota, elucidating the molecular and cellular basis of their systemic efficacy. (2) Advancing translational research: Concurrently developing QbD-based standardized production processes, innovative regulatory science evaluation systems, and biomarker-based personalized application strategies, building a complete translational pathway from basic discovery to clinical application. (3) Expanding the therapeutic horizon: Validating and exploring the therapeutic potential of PDNVs in other neurological diseases characterized by neuroinflammation, barrier damage, and gut-brain axis dysregulation (eg., Alzheimer’s disease, Parkinson’s disease, multiple sclerosis, depression).

Ultimately, through interdisciplinary collaboration and an integrated innovation system spanning basic mechanisms, standardized processes, preclinical evaluation, and regulatory science, there is potential to successfully advance this innovative strategy—merging traditional wisdom with modern technology—to the clinic, not only providing transformative treatment options for IS patients but also opening new avenues for systemic intervention against other complex diseases.