Dual‐Function Plant‐Derived Nanovesicles From Regenerated Cannabis sativa Roots for Immunotherapy and Vaccine Delivery

ABSTRACT

Cannabis sativa is a medicinal plant that produces a diverse array of pharmacologically active metabolites, making it a valuable resource for pharmaceutical applications. In this study, an adventitious root (AR) culture system was established from C. sativa using two representative plant growth regulators—naphthaleneacetic acid (NAA; hereafter referred to as N‐ARs) and indole‐3‐butyric acid (IBA; hereafter referred to as I‐ARs) —from which plant‐derived nanovesicles (PDNVs) were subsequently isolated (hereafter N‐PDNVs and I‐PDNVs, respectively). The resulting N‐PDNVs and I‐PDNVs exhibited average diameters of 128 ± 2 and 124 ± 4 nm, respectively, with zeta potentials of −12.9 and −15.7 mV. Both PDNV types maintained structural integrity and colloidal stability under diverse external stress conditions, underscoring their physicochemical robustness. Metabolite profiling of PDNVs revealed 25 distinct metabolites. Functionally, I‐PDNVs markedly enhanced dendritic cell maturation through Toll‐like receptor 2 (TLR2)‐ and TLR4‐dependent pathways, promoted T cell proliferation and activation (notably IFN‐γ‐ and IL‐17A‐producing subsets), and increased natural killer (NK) cell activity compared with N‐PDNVs. In immunosuppressed and tumour‐bearing mouse models, I‐PDNVs further augmented NK cell, Th1 and cytotoxic T lymphocyte (CTL) responses, thereby confirming their superior potential as immunotherapeutic agents. Moreover, in immunized mouse models, OVA_257‐264‐encapsulated I‐PDNVs demonstrated a clear advantage as a vaccine delivery platform by eliciting a potent OVA_257‐264‐specific CTL response. When applied as a prophylactic cancer vaccine, they not only delayed tumour growth but also reshaped the antitumour immune landscape, characterized by enhanced CTL responses, reduced regulatory T cell frequencies and diminished exhausted CD8⁺ T cell populations. Collectively, these findings highlight the potential of I‐PDNVs as dual‐function PDNVs, serving both as immunotherapeutic agents and as vaccine delivery platforms for applications requiring reinforced Th1, CTL and NK cell responses.

Schematic diagram illustrating the immunotherapeutic and vaccine delivery potential of I‐PDNVs.

1. Introduction

Nanovesicles, commonly referred to as extracellular vesicles (EVs) are secreted by various organisms or are artificially isolated under various conditions, with their sizes ranging from 30 nm to 10 µm (Kalluri and LeBleu 2020). Owing to their functional properties, such as immunity promotion, inflammation control and antioxidant activities, nanovesicles have attracted attention as promising candidates for drug delivery systems (DDS) and for treating various immune‐related diseases, including cancers, infectious diseases and autoimmune disorders (Zhang et al. 2020; Andaloussi S et al. 2013; Harrell et al. 2019; Kar et al. 2023). Their therapeutic potential is largely attributed to the immunomodulatory cargos they carry, including lipids, small molecules, proteins, nucleic acids and metabolites, which are either encapsulated within or presented on the surface of EVs (Chettimada et al. 2018; Deng and Miller 2019; Zhang et al. 2023; Segura et al. 2005). Furthermore, their structural resemblance to synthetic liposomes allows them to effectively load and protect various drugs from degradation by external stimuli, enhancing their viability as DDS platforms (Samanta et al. 2018).

EVs derived from human mesenchymal stem cells (MSCs) are a notable example of EVs as therapeutic agents (Lotfy et al. 2023). MSC‐derived EVs contain microRNAs (miRNAs), which act as a driving force to induce inflammatory polarization of macrophages towards the anti‐inflammatory M2 phenotype. Such EVs ultimately regulate excessive inflammatory responses, being promising candidates for the treatment of various inflammatory diseases (Hu et al. 2022; Roszkowski 2024). Another example of EVs as drug delivery systems is those derived from dendritic cells (DC‐EVs). DC‐EVs promote the activity of immune cells—particularly that of Th1 and natural killer (NK) cells—by presenting surface ligands recognized by toll‐like receptor (TLR) 1, 2 and 4, thereby contributing to the orchestration of inflammatory responses (Sobo‐Vujanovic et al. 2014). Such properties make DC‐EVs promising DDS candidates for cancer vaccines, leading to the development of new platform technologies that enhance anticancer immunity, with clinical research currently underway (Xia et al. 2022).

According to current studies, the effective use of EVs isolated from various organisms for disease treatment or DDS requires the elucidation of their unique bioactive functions and the validation of their efficacy in controlling target diseases (Yang et al. 2021; Guo et al. 2022; Lin et al. 2024; Nouri et al. 2024). Additionally, the advancement of EV‐based therapeutics and DDS systems requires overcoming key challenges such as sustainable production, enhanced extraction efficiency and functional optimization for targeted disease applications. For example, cancer‐derived EVs contain antigenic subunits of various cancer cells and can be used as vaccines; however, their efficacy is limited by immune evasion mechanisms, particularly immune tolerance (Ning et al. 2018; Poggio et al. 2019; Szczepanski et al. 2011; Gao et al. 2020). To address these limitations, new technologies have been developed that apply various stresses to cancer cells, including heat shock and irradiation, resulting in higher yields of more immunogenic EVs suitable for use as vaccines (Tan et al. 2010; Kim et al. 2020). In addition to these techniques, several methods have been reported to enhance the yield of cell‐derived EVs, including the use of chemical inducers targeting the cell cortex, particularly actin, as well as modulating cultivation conditions, such as hypoxia, temperature shifts and pH changes (Glebovskii and Pashkevich 1989; Chen et al. 2011; Parolini et al. 2009). Ultimately, to successfully develop EVs for disease treatment or DDS, four critical criteria must be met: (1) establishment of a sustainable, scalable and consistent EV production platform; (2) implementation of strategies to enhance EV yield and augment their intrinsic therapeutic potential; (3) comprehensive evaluation of the stability, bioactivity and mechanisms of action of EVs; and (4) optimization of their therapeutic efficacy tailored to disease‐specific treatment strategies.

In this study, we investigated the adventitious root (AR) culture system as a strategy to overcome key limitations in developing EVs for therapeutics and DDS applications. ARs, regenerated from plant tissues through the action of plant growth regulators—particularly auxins that drive root initiation and elongation—are recognized as a valuable platform for the large‐scale production of bioactive compounds in medicinal plants (Gonin et al. 2019; Murthy et al. 2008; Murthy et al. 2024). Among the auxins employed for AR induction, the naturally occurring indole‐3‐butyric acid (IBA) and the synthetic naphthaleneacetic acid (NAA) are most commonly used, with the applied auxin type influencing AR morphology, growth kinetics and the accumulation of bioactive compounds (Kim et al. 2003; Jeong et al. 2009). Notably, this culture system can be maintained under sterile and controlled conditions, minimizing the impact of environmental fluctuations, pest invasions and harmful soil pathogens, while ensuring a relatively uniform and consistent biomass quality. Consequently, this culture system is anticipated to provide a robust platform for the sustainable production of high‐quality EVs.

Building on these strengths, our primary objective was to compare and characterize the metabolic and immunobiological profiles of plant‐derived EVs isolated from ARs of C. sativa L. cultured with either NAA or IBA. In this study, we specifically focused on plant‐derived nanovesicles (PDNVs), defined as encompassing all vesicular fractions—including those potentially representing vesicle‐like artificial structures—obtained from plant tissues through controlled disruptive processes followed by filtration (Pinedo et al. 2021). C. sativa, a medicinal plant widely recognized for its diverse applications in pharmaceuticals, cosmetics and health supplements (Leinen et al. 2023; Martins et al. 2022; Fordjour et al. 2024), served as the source for NAA‐induced ARs (N‐ARs) and IBA‐induced ARs (I‐ARs), from which N‐PDNVs and I‐PDNVs were isolated, respectively. We systematically assessed the immunomodulatory effects of these PDNVs on dendritic cells (DCs)—key regulators of immune activation and disease pathogenesis—and further delineated their underlying molecular mechanisms. In addition, we investigated their therapeutic utility in immunotherapy and their potential as DDS platforms. Collectively, these studies aim to establish PDNVs from C. sativa ARs as a versatile biotherapeutic modality integrating robust immunomodulatory activity with effective vaccine delivery capability.

2. Materials and Methods

2.1. AR Production and Growth Conditions

Seeds from C. sativa Cheongsam (Korean hemp) were sterilized with a 1% hydrogen peroxide solution (H₂O₂) by shaking in darkness at 25°C for 2 days, with the 1% H₂O₂ solution being refreshed daily. After removing the seed coat from the germinated seed, the embryos were rinsed twice with sterile water and dried using 3M paper. The dried embryos were then placed on 1/2 Murashige and Skoog (MS) medium (Duchefa Biochemie, Haarlem, Netherlands) containing 15 g/L sucrose and 4 g/L gelite at pH 5.8. Seedlings were grown under 16 h light and 8 h dark at 25°C for 7 days. To regenerate ARs, root explants dissected from 7‐day‐old seedlings were incubated on MS‐NAA agar (MS including vitamins, 15 g/L sucrose, 0.5 g/L 2‐morpholinoethanesulfonic acid (MES), 0.1 mg/L NAA (Duchefa Biochemie), 4 g/L gelite, pH 5.8) or MS‐IBA agar (MS including vitamins, 15 g/L sucrose, 0.5 g/L MES, 0.1 mg/L IBA (Duchefa Biochemie) and 4 g/L gelite, pH 5.8) at 22°C in the dark. The ARs regenerated on MS‐NAA or MS‐IBA agar (named N‐ARs and I‐ARs, respectively) were maintained by subculturing at 2‐week intervals.

2.2. PDNVs Isolation From N‐ARs and I‐ARs

N‐ARs were cultured in an MS‐NAA liquid medium, while I‐ARs were cultured in an MS‐IBA liquid medium at 22°C in the dark with shaking for 2 weeks. After culturing, ARs were collected and removed from the conditioned medium, followed by washing in sterile distilled water. Approximately 100 g of N‐AR and I‐AR were separately homogenized with 1 L of cold phosphate‐buffered saline (PBS) using a blender at maximum speed for 3 min. The homogenized N‐ARs or I‐ARs were centrifuged at 4000 × g for 20 min at 4°C to discard any fibres and large particles. The supernatants were centrifuged at 12,000 × g for 20 min at 4°C and filtered through a 0.45‐µm membrane filter (GVS Filter Technology, Baldwinsville, NY, USA). The filtered supernatants were concentrated using a tangential flow filtration (TFF) system (Pall Corporation, Port Washington, NY, USA) with a 500‐kDa membrane filter. To obtain PDNVs of high quantity and quality, the concentrated supernatants (30 mL) were layered on an Optiprep solution (3 mL, 60% iodixanol; Axis‐Shield, Oslo, Norway) loaded in ultracentrifuge tubes (Beckman Coulter, California, USA) and ultracentrifuged at 120,000 × g for 60 min at 4°C. The PDNVs located between the Optiprep solution and the supernatant were harvested by carefully removing the supernatant and Optiprep solution, leaving approximately 1 mL. Finally, the isolated PDNVs were resuspended in PBS and filtered through a 0.45‐µm syringe filter (Advantec, Tokyo, Japan), followed and stored at −80°C until further use. PDNVs isolated from N‐ARs and I‐ARs were designated as N‐PDNVs and I‐PDNVs, respectively. The protein contents of N‐PDNVs and I‐PDNVs were determined using a BCA protein assay kit (Thermo Fisher Scientific, Waltham, MA, USA).

2.3. Measurement of the Zeta Potential, Particle Size and Size Distribution by Nanoparticle Analyzer

The size distribution and zeta potential of the isolated PDNVs were measured using an IZON nanoparticle analyzer (IZON Science Ltd., Christchurch, New Zealand), utilizing tunable resistive pulse sensing (TRPS). To assess their enzymatic resistance, PDNVs resuspended in PBS were incubated with proteinase K (PK; 100 µg/mL), RNase (6 µg/mL) and DNase (3 µg/mL) at 37°C for 30 min. Following treatment, their physical characteristics—including particle size, zeta potential and size distribution (expressed as the d90/d10 ratio)—were measured to assess structural integrity. PDNVs resuspended in PBS without exposure to enzymatic agents were used as untreated controls. DNase, PK and RNase were procured from Thermo Fisher Scientific. Additionally, the stability of PDNVs was assessed under varying conditions, including temperature fluctuations, storage duration and the presence or absence of foetal bovine serum (FBS). PDNVs were incubated in either 10% EV‐depleted FBS (prepared via ultracentrifugation at 120,000 × g for 18 h at 4°C to remove the EVs included in FBS) or PBS. The incubation was performed at both 37°C and 4°C, and the physical size, zeta potential and size distribution were analyzed after 6, 24 and 48 h to monitor any changes in their stability.

2.4. Morphological Analysis of PDNVs

Cryogenic transmission electron microscopy (cryo‐TEM) was performed by applying 3 µL of each PDNV sample onto Quantifoil R 2/2 holey carbon grids (Quantifoil Micro Tools GmbH, Germany). Excess liquid was removed using Whatman grade 1 filter paper (Thermo Fisher Scientific), and the grids were immediately vitrified in liquid ethane using a Vitrobot Mark IV system (Thermo Fisher Scientific). The vitrified samples were stored in liquid nitrogen and imaged using a Glacios cryo‐TEM (Thermo Fisher Scientific) operating at 200 kV.

2.5. Western Blotting

N‐PDNVs and I‐PDNVs were further characterized using antibodies against Arabidopsis thaliana PDNVs (At‐PDNVs)‐specific positive and negative markers, as previously described (Lopez de Las Hazas et al. 2023). Positive markers included anti‐lumenal binding protein (anti‐BiP; Agrisera, Vannas, Sweden), anti‐heat shock protein 70 (anti‐HSP70; Agrisera) and anti‐PENETRATION 1 (anti‐PEN1; Cusabio, Wuhan, China). The negative marker was anti‐rubisco large subunit (anti‐RbcL; Agrisera). As positive controls for PDNV‐associated proteins (BiP, HSP70 and PEN1), PDNVs were isolated from 100 g of A. thaliana leaves (designated as At‐PDNVs) using the same procedures applied for the preparation of N‐PDNVs and I‐PDNVs from N‐ARs and I‐ARs. To provide a positive control for proteins absent from plant EVs, 1 g of A. thaliana leaves was directly lysed in RIPA buffer (Pierce, Rockford, IL, USA) to obtain total leaf protein extracts (designated as At‐leaf proteins), which served as the control for RbcL. Western blotting was performed under conditions previously described (Kim et al. 2020). For each assay, 50 µg of protein from PDNV preparations or At‐leaf protein extracts was used. Antibodies were used at the following dilutions: anti‐BiP (1:2000), anti‐HSP70 (1:3000), anti‐PEN1 (1:2000) and anti‐RbcL (1:5000).

2.6. Metabolite Analysis

The metabolites constituting PDNVs were identified and analyzed via gas chromatography‐mass spectrometry (GC‐MS) and ultra‐performance liquid chromatography coupled with quadrupole time‐of‐flight mass spectrometry (UPLC‐Q‐TOF‐MS). Briefly, 1 mg of PDNVs was lyophilized using a freezing/vacuum‐drying system (VD‐800F; Taitec, Saitama‐ken, Japan) and dissolved in 100 µL of 80% methanol for UPLC‐Q‐TOF‐MS analysis. For GC‐MS analysis, 20 µg of N‐PDNVs and I‐PDNVs were dissolved in 70 µL of hydroxymethyl amine and incubated at 37°C for 90 min, followed by the addition of 70 µL of N, O‐bis(trimethylsilyl)trifluoroacetamide (BSTFA) and heating at 90°C for 30 min. After derivatization, 1.5 mL of sodium chloride (NaCl) and 1 mL of n‐hexane were added, and the mixture was centrifuged at 2000 rpm for 2 min to collect the upper layer for analysis. Subsequently, UPLC‐Q‐TOF‐MS and GC‐MS analyses were conducted following the protocols established in our previous studies (Lee et al. 2024). All metabolites identified and annotated in this study were classified according to the metabolomics standards initiative (MSI) Levels 1–4 (Sumner et al. 2007). According to the MSI, Level 1 identification was achieved by comparing accurate mass, retention time and MS/MS spectra with those of authentic standards analyzed under identical experimental conditions. For LC‐MS analysis, metabolites without authentic standards were annotated as Level 2 based on accurate mass (mass error ≤ 5 ppm), retention time and MS/MS spectral similarity against public databases, including MassBank and MoNA. For GC‐MS analysis, Level 2 annotation was supported by comparison of experimental retention indices (RI) and electron ionization (EI) mass spectra with the NIST and Wiley libraries, considering similarity scores ≥ 80% as acceptable. Features assigned only to compound classes were reported as Level 3 and unresolved peaks were categorized as Level 4. The GC‐MS and UPLC‐Q‐TOF‐MS results were analyzed using multivariate statistical methods in SIMCA P+ v.14.0.1 (Umetrics, Umea, Sweden) and visualized through a partial least squares discriminant analysis (PLS‐DA). The quality of the PLS‐DA model was assessed based on three parameters (R₂X, R₂Y and Q²) and validated through permutation tests.

2.7. Cell Culture and Cell Viability

Bone marrow‐derived DCs (BMDCs) were obtained from 8‐week‐old female C57BL/6 mice, sourced from Orient Bio Inc. (Seoul, Republic of Korea). The cells, specifically red blood cell (RBC)‐depleted bone marrow cells, were extracted from the femurs and tibiae. For differentiation, bone marrow cells were cultured in complete Roswell Park Memorial Institute (cRPMI) medium (Gibco BRL, Grand Island, NY, USA) containing 10% FBS (Gibco BRL), 1% penicillin‐streptomycin (Gibco BRL), 0.5 ng/mL IL‐4 (JW Creagene, Gyeonggi, Republic of Korea) and 20 ng/mL GM‐CSF (JW Creagene). This culture was maintained for 7 days at 37°C in a 5% CO₂ atmosphere, with the medium being added or replaced every 3 days. To evaluate the impact of the PDNVs treatment on cell viability, 0.5 × 10⁶ BMDCs in 500 µL of cRPMI medium were seeded into each well of a 48‐well plate. The cells were then exposed to various concentrations (1, 2 and 5 µg/mL) of PDNVs or 50 nM staurosporine (STS) for 19 h at 37°C. After this treatment, 50 µL of EZ‐Cytox Cell Viability kit solution (DoGen, Seoul, Korea) was added to each well, followed by incubation at 37°C for 90 min. The absorbance was measured at 450 nm using a microplate reader (Molecular Devices Inc., San Jose, CA, USA).

2.8. Extracellular and Intracellular Cytokine Measurement

To evaluate the extracellular cytokine levels, BMDCs (a density of 0.25×10⁶ cells in 500 µL of cRPMI medium per 48‐well plate) were treated with PDNVs (1, 2.5 and 5 µg/mL), lipopolysaccharide (LPS; 100 ng/mL; Invivogen, San Diego, CA) or Pam3CSK4 (Pam3; 100 ng/mL; InvivoGen) for 19 h. LPS and Pam3 were used as positive controls for DC maturation. Following treatment, the culture supernatants were collected, and the levels of the extracellular cytokines, including the tumour necrosis factor‐α (TNF‐α), interleukin (IL)‐12p70 and IL‐10, were quantified using enzyme‐linked immunosorbent assay (ELISA) kits (Thermo Fisher Scientific) according to the manufacturer's instructions. For intracellular cytokine analysis, BMDCs were treated with 5 µg/mL PDNVs or 100 ng/mL LPS in the presence of a transport inhibitor cocktail (Thermo Fisher Scientific) for 12 h. Post‐incubation, the cells were harvested and stained with the CD11c‐PE‐Cy7 antibody (BD Bioscience, San Jose, CA, USA) for 30 min at 4°C in the dark. After surface staining, the cells were incubated with Cytofix/Perm solution (BD Bioscience) and incubated for an additional 30 min at 4°C in the dark. Subsequent intracellular cytokine staining (for 30 min at 4°C in the dark) was performed using TNF‐α‐APC (Thermo Fisher Scientific), IL‐12p70‐FITC (BD Bioscience) and IL‐10‐PE (BD Bioscience). Finally, the stained cells were analyzed using a Life Launch Attune Nxt Flow Cytometer (Thermo Fisher Scientific).

2.9. Quantitative Analysis of the Surface Molecules on BMDCs

To assess the impact of PDNVs on the expression of surface molecules in BMDCs, cells were cultured with 5 µg/mL PDNVs or 100 ng/mL LPS for 19 h at 37°C. Post‐treatment, the cells were collected and washed twice with cold PBS. The cells were then stained with CD11c‐PE‐Cy7, MHC‐I‐PE, MHC‐II‐PerCp‐Cy5.5 and CD80‐APC for 30 min at 4°C in the dark. All surface molecule antibodies and isotype controls (APC‐conjugated Hamster IgG2, PE‐conjugated Rat IgG2a and PerCp‐Cy5.5‐conjugated rat IgG2b) for surface molecules were sourced from BD Bioscience. The stained cells were subsequently analyzed via flow cytometry.

2.10. Determination of Endotoxin Contamination in PDNVs

To examine endotoxin contamination in the isolated PDNVs, PDNVs (5 µg/mL) or LPS (100 ng/mL) were co‐incubated with 5 µg/mL polymyxin B solution (PMB; Sigma‐Aldrich) for 1 h at 25°C. BMDCs (0.5 × 10⁶ cells per 500 µL) were incubated with PMB‐treated N‐PDNVs, I‐PDNVs and LPS for 19 h at 37°C. Following incubation, the cells were collected and stained with CD11c‐PE‐Cy7, MHC‐I‐PE, MHC‐II‐PerCp‐Cy5.5 and CD80‐APC, followed by flow cytometry analysis. The supernatants were collected to quantify the extracellular cytokine (TNF‐α, IL‐12p70) levels using ELISA kits.

2.11. Antigen Uptake and Antigen‐Presenting Ability

To evaluate the antigen uptake ability of BMDC, cells were incubated with 5 µg/mL PDNVs or 100 ng/mL LPS for 19 h. Following this, the cells were further incubated at either 4°C or 37°C for 30 min. Subsequently, 0.5 mg/mL of FITC‐conjugated dextran (40,000 Da, Sigma‐Aldrich) was added to each well, mixed thoroughly by pipetting and incubated for an additional 30 min under the same temperature conditions (4°C or 37°C). The cells were harvested, washed twice with PBS and stained with the CD11c‐PE‐Cy7 antibody for 30 min at 4°C in the dark. The proportion of the CD11c⁺dextran⁺ cells was quantified via flow cytometry. To evaluate the antigen‐presenting ability via the MHC‐I pathway, DC were treated with either the full‐length ovalbumin (OVA) protein alone or co‐treated with 5 µg/mL PDNVs and 500 µg/mL OVA protein, followed by incubation at 37°C for 19 h. The well‐characterized OVA_257–264 peptide (SIINFEKL) was used as a positive control. Regarding the MHC‐II pathway, cells were treated with either 25 µg/mL of the Eα_44–76 peptide (RLEEFAKFASFEAQGALANIAVDKANLDVMKKR) alone or co‐treated with PDNVs and the Eα_44–76 peptide under identical conditions. The Eα₅₂–₆₈ peptide (ASFEAQGALANIAVDKA) served as a positive control. Post‐incubation, the cells were harvested and washed twice with PBS. The cells were stained with CD11c‐PE‐Cy7 and 25‐D1.16‐PE (specific for SIINFEKL; Thermo Fisher Scientific), and Y‐Ae‐FITC (specific for ASFEAQGALANIAVDKA; Thermo Fisher Scientific) antibodies for 30 min at 4°C in the dark. Flow cytometry was subsequently performed to analyze the proportion of the CD11c⁺H‐2Kb⁺ and CD11c⁺Y‐Ae⁺ cells.

2.12. Evaluation of the Cellular Uptake Ability of PDNVs

To assess their cellular uptake, the isolated PDNVs were fluorescently labelled with 5 µM CellTrace CFSE (Thermo Fisher Scientific) by incubation at 37°C for 15 min in the dark. Labelled PDNVs were then collected by ultracentrifugation at 120,000 × g for 60 min at 4°C, resuspended in 1 mL of PBS and filtered through a 0.45 µm syringe filter. To compare the cellular uptake efficiency of CFSE‐labelled PDNVs, BMDCs were differentiated from 8‐week‐old wild‐type (WT), Toll‐like receptor 2 knockout (TLR2^−/−; Jackson Laboratory, Bar Harbor, ME, USA) and Toll‐like receptor 4 knockout (TLR4^−/−; Jackson Laboratory) mice, all on a C57BL/6 background. The cells were incubated with the CFSE‐labelled PDNVs (5 µg/mL) at 37°C for varying durations (30, 60, 90 and 120 min). Post‐incubation, the cells were collected and stained with the CD11c‐PE‐Cy7 antibody for 30 min at 4°C. Flow cytometry was used to analyze the CFSE‐labelled PDNVs⁺CD11c⁺ cell populations, with the fluorescence intensities compared against those of non‐treated cells.

2.13. Mixed Lymphocyte Reaction Assay

Splenocytes were isolated from the spleen of BALB/c mice and subjected to RBC depletion using an RBC lysis solution (Thermo Fisher Scientific). The RBC‐depleted splenocytes were then labelled with 1 µM CFSE for 10 min at 37°C in the dark. After labelling, the CFSE‐labelled splenocytes were washed and resuspended in an RPMI medium supplemented with 10% FBS and 1% penicillin‐streptomycin. The cells were then seeded at a density of 5 × 10⁵ cells per well in a 96‐well U‐bottom plate. BMDCs derived from C57BL/6 mice were treated with 100 ng/mL LPS, 5 µg/mL PDNVs for 2 h at 37°C. After incubation, DCs (1 × 10⁵ cells per well) were co‐cultured with the CFSE‐labelled splenocytes at a 5:1 ratio for 3 days. Post‐culture, the splenocytes were analyzed via flow cytometry. Additionally, the levels of the extracellular cytokines, including IFN‐γ, IL‐5 and IL‐17A, were quantified in the culture supernatants using ELISA kits (Thermo Fisher Scientific).

2.14. Analysis of the NK Cell Activity

NK cells were isolated from spleens derived from 8‐week‐old female C57BL/6 mice using an NK cell isolation kit (Miltenyi Biotec, Bergisch Gladbach, Germany) based on negative selection. BMDCs (2 × 10⁶) were pre‐treated with LPS (100 ng/mL) or PDNVs (5 µg/mL) for 2 h. Subsequently, the pre‐activated DCs were co‐cultured with the isolated NK cells at a ratio of 1:4 (2 × 10⁶ NK cells) for 24 h at 37°C. Following the co‐culture period, the cells were treated with a transport inhibitor cocktail for 4 h. Surface staining was performed with NK1.1‐APC‐Cy7 (BD Bioscience) and CD3e‐FITC (BD Bioscience) for 30 min at 4°C. After surface staining, the cells were fixed and permeabilised using a Cytofix/Perm solution. Intracellular staining was conducted with CD107a‐PE‐Cy7, IFN‐γ‐PE and Granzyme B (G/B)‐APC for 30 min at 4°C. All antibodies were sourced from BD Bioscience. The stained cells were analyzed via flow cytometry to assess the activation status of the NK cells, focusing on the expression levels of CD107a, IFN‐γ and G/B.

2.15. Mouse Immunosuppression Model

The Institutional Animal Care and Use Committee (IACUC) of the Korea Research Institute of Bioscience and Biotechnology (KRIBB) approved all animal procedures used in the immunosuppression model (Approval no.: KRIBB‐AEC‐21330). Eight‐week‐old female C57BL/6 mice were randomly assigned to three groups (n = 5 per group): a control group (PBS‐injected), a cyclophosphamide (CTX; Sigma‐Aldrich) plus PBS (CTX+PBS)‐injected group and a CTX plus I‐PDNVs (CTX+I‐PDNVs)‐injected group. The control group received intraperitoneal injections of PBS (100 µL) once daily for 5 consecutive days, and 3 days after the final PBS injection, mice were orally administered PBS (200 µL) once daily for an additional 5 days. In the CTX‐induced immunosuppressed group (CTX+PBS), CTX was administered intraperitoneally at 80 mg/kg (100 µL) once daily for 5 consecutive days to induce immunosuppression, and 3 days after the final CTX injection, mice received oral PBS (200 µL) once daily for 5 days. For the I‐PDNVs‐treated group (CTX+I‐PDNVs), CTX was administered in the same manner, and 3 days after the final CTX injection, mice received oral I‐PDNVs (100 µg/mouse in 200 µL) once daily for 5 days. Following the treatment period, mice were assessed for cytotoxicity, immune cell distribution and immune function. Body weights were recorded at four designated time points throughout the experimental period. For cytotoxicity analysis, splenocytes were isolated from each mouse, adjusted to a concentration of 2 × 10⁶ cells/mL and subjected to Annexin V and PI staining following the manufacturer's protocol (Thermo Fisher Scientific). Flow cytometry analysis was conducted to quantify the apoptotic (Annexin V⁺) and necrotic (PI⁺) cell populations. For T cell subset analysis, splenocytes (2 × 10⁶ cells) were stained with fluorochrome‐conjugated antibodies specific for CD3‐Alexa 700, CD4‐PerCp‐Cy5.5 and CD8‐APC‐Cy7 surface markers (BD Bioscience). Flow cytometry was utilized to assess the distribution of CD3⁺CD4⁺ and CD3⁺CD8⁺. To evaluate T cell activation, splenocytes were stimulated with a cell stimulation cocktail containing a transport inhibitor (Thermo Fisher Scientific) for 4 h at 37°C. Cells were then stained with surface markers (CD3, CD4 and CD8), followed by intracellular staining for IFN‐γ‐PE (Thermo Fisher Scientific), IL‐5‐APC (BD Bioscience) and IL‐17A‐PE‐Cy7 (Thermo Fisher Scientific) using a fixation/permeabilization kit. Flow cytometry analysis was performed to assess cytokine production in T cells. The NK cell distribution was evaluated by staining splenocytes (2 × 10⁶ cells) with NK1.1‐APC‐Cy7 and lineage‐specific antibodies (CD19‐FITC, CD14‐FITC and CD3e‐FITC; BD Biosciences). Flow cytometry was used to quantify the proportion of NK cells. For analysis of NK cell activation, splenocytes were stimulated with a cell stimulation cocktail containing transport inhibitors for 4 h. After stimulation, cells were first surface‐stained with NK1.1‐ APC‐Cy7 and lineage‐specific antibodies (CD19‐FITC, CD14‐FITC and CD3e‐FITC), followed by fixation and permeabilization, and subsequently stained intracellularly with CD107a‐PE‐Cy7, IFN‐γ‐PE and G/B‐APC. The frequency of NK cells expressing CD107a, IFN‐γ and G/B was quantified by flow cytometry.

2.16. Tumour Inoculation and Treatment Protocol

IACUC of KRIBB authorized the use of animals for the tumour‐bearing mouse model (Permission no.: KRIBB‐AEC‐24248). Eight‐week‐old female C57BL/6 mice (n = 5 per group) were randomly assigned to 3 groups: Control (PBS‐injected), Tumour+PBS and Tumour+I‐PDNVs. For tumour‐injected groups (Tumour+PBS and Tumour+I‐PDNVs), E.G7 cells (a mouse lymphoma cell line; 1 × 10⁵ cells in 100 µL PBS) were intradermally injected into the left lower backs of the mice. Control mice received an equivalent intradermal injection of PBS (100 µL). Six days after tumour inoculation, mice were orally administered either PBS (200 µL; Control and Tumour+PBS groups) or I‐PDNVs (100 µg/mouse in 200 µL PBS; Tumour+I‐PDNVs group) at 2‐day intervals for a total of 5 doses. The tumour dimensions were measured every 2 days thereafter, and the tumour area was calculated using the following formula: area = (length × width²)/2.

2.17. Immune Cell Analysis in Tumour‐bearing Mice

Spleen cells from tumour‐bearing mice were isolated 22 days post‐tumour inoculation for immune cell activation analysis. For splenic DC activation, spleen cells (2 × 10⁶ cells) were stained with NK1.1‐APC‐Cy7, lineage cocktail (CD19‐FITC and CD3e‐FITC) and DC surface markers (CD11c‐PE‐Cy7, CD80‐APC, MHC‐I‐PE and MHC‐II‐PerCp‐Cy5.5) followed by flow cytometry analysis. NK and T cell activation was assessed by stimulating the cells for 4 h with a cell stimulation cocktail containing transport inhibitors, followed by surface and intracellular staining. For NK cell analysis, the following antibodies were used: NK1.1‐APC‐Cy7; lineage‐specific antibodies (CD19‐FITC, CD14‐FITC and CD3e‐FITC); CD107a‐PE‐Cy7; IFN‐γ‐PE; and G/B‐APC. For T cell subtype analysis, the following antibodies were used: CD3‐Alexa Fluor 700, CD4‐PerCP‐Cy5.5, CD8‐APC‐Cy7, IFN‐γ‐PE, IL‐5‐APC and IL‐17A‐PE‐Cy7. For multifunctional T cell analysis, the following antibodies were used: CD3‐Alexa Fluor 700, CD4‐PerCP‐Cy5.5, CD8‐APC‐Cy7, IFN‐γ‐PE, IL‐2‐PE‐Cy7 and TNF‐α‐APC.

2.18. Encapsulation of OVA_257‐264 Within I‐PDNVs

To encapsulate OVA_257‐264 within I‐PDNVs, 200 µg of I‐PDNVs were diluted in 100 µL of PBS and mixed with 50 µg of OVA_257‐264. The mixture was sonicated using a Vibra‐Cell CV33 ultrasonic processor (Sonics & Materials Inc., CA, USA) at 20% amplitude for 30 s, followed by a 30 s incubation on ice for stabilization. This procedure was repeated 6 times in total. And then, the sample was incubated at 37°C for 1 h and washed 5 times with PBS using ultrafiltration tubes (100 kDa MWCO; MilliporeSigma, Burlington, MA, USA), at 6000 × g for 2 min per cycle at 4°C to remove free OVA_257‐264. I‐PDNVs were then recovered, and their protein‐based mass was determined using a BCA assay, showing recovery comparable to the initial input. The OVA_257‐264 content in I‐PDNVs was quantified by dissolving PDNVs in 1% Triton X‐100, and the sample was analyzed via liquid chromatography‐tandem mass spectrometry (LC‐MS/MS) using an Acquity UPLC BEH C18 column (2.1 × 50 mm, 1.7 µm; Waters). The mobile phases consisted of water with 0.1% formic acid (A) and acetonitrile with 0.1% formic acid (B). A gradient elution method with a flow rate of 0.3 mL/min and a 10 µL injection volume was employed. Several multiple reaction monitoring conditions were established to determine the quantitative and qualitative ions, and the MS conditions were optimized before sample analysis. The standard curve for OVA_257‐264 ranged from 1000 to 0.1 ng/mL. The content of the OVA_257‐264 peptide in I‐PDNVs was approximately 10 µg, corresponding to an encapsulation efficiency of 20% based on the initial input of 50 µg OVA_257‐264. For all in vitro and in vivo experiments, the I‐PDNVs used had an OVA_257‐264 loading capacity of approximately 10 µg OVA_257‐264 per 200 µg I‐PDNVs. The encapsulation efficiency of OVA_257‐264 and drug loading efficiency were evaluated according to the following equations: Drug loading capacity = (mass of the encapsulated OVA_257‐264 in I‐PDNVs)/(protein‐based mass of I‐PDNVs after purification), Encapsulation efficiency = (mass of the encapsulated OVA_257‐264 in I‐PDNVs)/(initial OVA_257‐264 input) × 100.

2.19. Analysis of the Functional and Physicochemical Stability of OVA_257‐264‐Encapsulated I‐PDNVs

For physicochemical analysis, particle size, size distribution and zeta potential of intact I‐PDNVs and OVA_257–264‐encapsulated I‐PDNVs (OVA_257–264‐I‐PDNVs) were measured using an IZON nanoparticle analyzer. In addition, vesicular morphology and structural integrity were examined via cryo‐TEM. To assess functional stability, BMDCs were treated with either 5 µg/mL of intact I‐PDNVs or OVA_257–264‐encapsulated I‐PDNVs for 19 h at 37°C. Following treatment, extracellular levels of TNF‐α and IL‐12p70 were quantified by ELISA, and the surface expression of CD80, MHC‐I and MHC‐II was analyzed using flow cytometry.

2.20. Immunization and Anticancer Vaccine Evaluation of OVA_257‐264‐Encapsulated I‐PDNVs

An anticancer vaccine model was established using 8‐week‐old female C57BL/6 mice immunized intramuscularly at 1‐week intervals (3 times in total) with one of the following: PBS (100 µL), OVA_257–264 peptide (10 µg/mouse in 100 µL), I‐PDNVs (200 µg/mouse in 100 µL), or OVA_257–264 (10 µg)‐encapsulated I‐PDNVs (200 µg/mouse in 100 µL). To evaluate OVA_257–264‐specific CD8⁺ T‐cell responses, spleens (n = 5 per group) were harvested 1 week after the final immunization. Splenocytes (2 × 10⁶ cells) were restimulated ex vivo with 20 µg/mL OVA_257–264 peptide for 12 h. After stimulation, cells were stained with CD3‐Alexa 700 and CD8‐APC‐Cy7 antibodies for 30 min at 4°C, fixed and permeabilized and then stained intracellularly with IFN‐γ‐PE antibody. The frequency of CD3⁺CD8⁺IFN‐γ⁺ T cells was quantified by flow cytometry. To evaluate antitumour efficacy, mice (n = 5 per group) were challenged subcutaneously with 1 × 10⁵ OVA‐expressing E.G7 lymphoma cells in the left flank 1 week after the final immunization. Tumour growth was monitored on Days 3, 5, 10, 12 and 14 post‐challenge. Two weeks after tumour inoculation, spleens were again harvested to evaluate multifunctional CD8⁺ T cell responses. Splenocytes were stimulated with a cell stimulation cocktail containing a transport inhibitor for 4 h, followed by surface staining with anti‐CD3‐Alexa 700 and anti‐CD8‐APC‐Cy7 antibodies and intracellular staining with IFN‐γ‐PE, IL‐2‐PE‐Cy7 and TNF‐α‐APC antibodies. The frequency of multifunctional CD8⁺ T cells producing IFN‐γ, IL‐2 and TNF‐α was analyzed by flow cytometry. For the analysis of regulatory T cells (Treg), splenocytes were first surface‐stained with CD3‐Alexa 700 and CD4‐PerCp‐Cy5.5 antibodies, followed by fixation and permeabilization and subsequently stained intracellularly with Foxp3‐PE. The frequency of Treg cells was quantified by flow cytometry. For the analysis of exhausted CD8⁺ T cells, splenocytes were stained with the following surface antibodies: CD3‐Alexa 700, CD8‐APC‐Cy7, TIM‐3‐APC and PD‐1‐PE‐Cy7 (all from Thermo Fisher Scientific), and subsequently analyzed by flow cytometry.

2.21. Statistical Analysis

Statistical analysis of the processed GC‐MS and UPLC‐Q‐TOF‐MS metabolite intensities was performed using one‐way analysis of variance (ANOVA) with Duncan's test (p < 0.05) in SPSS 24.0 (SPSS Inc., Chicago, USA). The metabolite fold changes were visualized as a heatmap using GraphPad Prism version 10.0 (GraphPad, San Diego, CA, USA), with the heatmaps representing the mean log₂ fold‐change values of N‐PDNVs relative to I‐PDNVs. All other statistical analyses were performed in GraphPad Prism version 10.0. Data are presented as mean ± standard deviation (SD). For in vitro immunological assays, significance was determined using one‐way ANOVA with Tukey's post hoc test or an unpaired t‐test. For in vitro immunological analysis, the levels of significance for comparison between samples were determined by one‐way ANOVA with Tukey's post hoc test or unpaired t‐test. For in vivo experiments, differences between the two groups were analyzed using the Mann–Whitney rank test, while multiple group comparisons were evaluated using Kruskal–Wallis with Tukey's post hoc test. For statistical analysis, GraphPad Prism version 10.0 was used. Statistical significance was defined as *p < 0.05, **p < 0.01 and ***p < 0.001.

3. Results

3.1. Comprehensive Characterization and Stability Analysis of N‐PDNVs and I‐PDNVs

To isolate AR‐derived PDNVs, N‐ARs and I‐ARs (Figure 1A) were homogenized and subjected to a TFF system followed by cushioned ultracentrifugation, resulting in the successful isolation of N‐PDNVs and I‐PDNVs of high purity (Figure 1B). According to TRPS analysis using an IZON nanoparticle analyzer, the average particle concentrations of N‐PDNVs and I‐PDNVs were 1.25 × 10¹⁰ particles/mL and 1.35 × 10¹⁰ particles/mL, respectively (Figure 1C). The corresponding protein concentrations were approximately 0.56 mg/mL and 0.54 mg/mL, indicating comparable production yields between the two preparations (Figure 1D). Cryo‐TEM analysis revealed that both N‐PDNVs and I‐PDNVs exhibited spherical nanoscale vesicular structures ranging approximately from 100 to 150 nm in diameter (Figure 1E). Western blot analysis demonstrated that N‐PDNVs and I‐PDNVs shared PDNV‐associated markers (BiP and HSP70) with At‐PDNVs, while PEN1 was detected exclusively in At‐PDNVs. The chloroplast protein RbcL—abundant in leaf and stem tissues but absent or only minimally present in PDNVs—was undetectable in N‐PDNVs and I‐PDNVs, with At‐leaf proteins serving as a positive control (Figure 1F). To evaluate their structural and physicochemical stability under enzymatic and environmental stress conditions, both types of PDNVs were further characterized using an IZON nanoparticle analyzer and cryo‐TEM (Figure 2). Intact N‐PDNVs and I‐PDNVs exhibited an average diameter of 128 ± 2 and 124 ± 4 nm, respectively, and zeta potential values of −12.9 and −15.7 mV, respectively. Upon treatment with PK (100 µg/mL), DNase (3 µg/mL) and RNase (6 µg/mL) for 30 min, both N‐PDNVs and I‐PDNVs retained their physicochemical stability, showing no significant changes in particle size, d90/d10 ratio, or zeta potential, and remained within the stable range of −30 to +30 mV (Figure 2A). In addition, cryo‐TEM analysis confirmed the preservation of vesicular morphology and structural integrity following enzymatic treatment (Figure 2B). Stability was also confirmed under various suspension conditions (10% FBS or PBS, at 4°C or 37°C) up to 48 h (Figure 2C–F). After incubation in 10% FBS for 6, 24 and 48 h, no statistically significant changes were observed in particle size, d90/d10 ratio, or zeta potential (Figure 2C), and cryo‐TEM analysis demonstrated sustained structural integrity (Figure 2D). Similar stability was observed in PBS, with consistent size parameters and zeta potential values (Figure 2E), further supported by cryo‐TEM images confirming preserved morphology (Figure 2F). Collectively, these results indicate that both N‐PDNVs and I‐PDNVs exhibit favourable colloidal and enzymatic stability under physiologically relevant aqueous conditions, including prolonged suspension in serum‐containing and buffer environments.

FIGURE 1.

Comprehensive characterization of N‐PDNVs and I‐PDNVs. (A) Representative images of C. sativa adventitious roots (ARs) induced by naphthaleneacetic acid (N‐ARs) or indole‐3‐butyric acid (I‐ARs). (B) N‐PDNVs and I‐PDNVs were concentrated using a TFF system and isolated via cushioned ultracentrifugation with 60% iodixanol. (C) Particle concentrations of N‐PDNVs and I‐PDNVs measured by tunable resistive pulse sensing (TRPS) using an IZON nanoparticle analyzer. (D) Protein concentrations of N‐PDNVs and I‐PDNVs determined by BCA assay. (E) Representative cryo‐TEM images showing spherical nanoscale vesicular morphology of N‐PDNVs and I‐PDNVs. (F) Western blot analysis of PDNV‐associated markers (BiP, HSP70, PEN1) in N‐PDNVs and I‐PDNVs, compared with Arabidopsis thaliana‐derived PDNVs (At‐PDNVs). The chloroplast protein RbcL—abundant in leaf and stem tissues—was used as a negative control for PDNV preparations, with A. thaliana total leaf protein (At‐leaf proteins) serving as a positive control. The results are presented as the mean ± SD (n = 3) and are representative of one of three independent experiments. Statistical analysis was performed using an unpaired t‐test. The term “ns” indicates not significant.

FIGURE 2.

Structural and physicochemical stability of N‐PDNVs and I‐PDNVs under enzymatic and environmental stress. (A) Particle size, size distribution (d90/d10 ratio), and zeta potential of N‐PDNVs and I‐PDNVs, measured by TRPS (IZON nanoparticle analyzer), before (non) and after treatment with proteinase K (PK; 100 µg/mL), DNase (3 µg/mL) or RNase (6 µg/mL) for 30 min. (B) Representative cryo‐TEM images confirming preservation of vesicular morphology following enzymatic treatment. (C) Particle size, d90/d10 ratio and zeta potential of N‐PDNVs and I‐PDNVs after incubation in 10% foetal bovine serum (FBS) at 4°C or 37°C for 0, 6, 24 and 48 h. (D) Cryo‐TEM images showing intact vesicular structure of N‐PDNVs and I‐PDNVs after FBS incubation at 4°C or 37°C for 48 h. (E) Particle size, d90/d10 ratio and zeta potential of N‐PDNVs and I‐PDNVs after suspension in phosphate‐buffered saline (PBS) at 4°C or 37°C for up to 48 h. (F) Cryo‐TEM images confirming preserved morphology of N‐PDNVs and I‐PDNVs following PBS incubation at 4°C or 37°C for 48 h. The results are presented as the mean ± SD (n = 3) and are representative of one of three independent experiments. Statistical analysis was performed using one‐way ANOVA with Tukey's post hoc test. **p < 0.01 and ***p < 0.001. The term “ns” indicates not significant.

3.2. Metabolite Identification in N‐PDNVs and I‐PDNVs

To further elucidate the biochemical characteristics of N‐PDNVs and I‐PDNVs, comprehensive metabolite profiling was performed via UPLC‐Q‐TOF‐MS (Figure S1A and Figure 3A). The metabolites were compared using multivariate statistical analysis, and the discrimination of N‐PDNVs and I‐PDNVs was visualized using a PLS‐DA score plot (Figure 3A; left panel). The PLS‐DA model was validated based on the goodness‐of‐fit (R₂X = 0.646 and R₂Y = 0.991), predictability (Q₂ = 0.967), p values (1.35E‐03), and cross‐validation using permutation tests (Figure 3A; central panel), indicating a clear separation between N‐PDNVs and I‐PDNVs with statistical reliability. The differences between N‐PDNVs and I‐PDNVs were attributed to a total of 12 metabolites identified, exhibiting variable importance in the projection (VIP > 1) and p values (< 0.05) derived from all normalized chromatogram intensities (Table 1). These metabolites were categorized into 5 nucleobase/nucleoside‐related metabolites (guanine, adenine, hypoxanthine, guanosine and methyladenosine isomer), 4 amino acids and peptides (alanyl‐isoleucine, L‐tryptophan, isoleucyl‐threonine and valyl‐valine), 2 alkaloids (Cannabisativine, Terestigmine), and 1 lipid (phytosphingosine). Additionally, the differential metabolites between N‐PDNVs and I‐PDNVs were identified by visualizing heatmaps of the average log₂‐fold change values (Figure 3A; right panel). Furthermore, PLS‐DA modelling based on the GC‐MS metabolite profiles (Figure S1B and Figure 3B) showed a clear discrimination between N‐PDNVs and I‐PDNVs, as indicated by the goodness‐of‐fit (R₂X = 0.836 and R₂Y = 0.999), predictability (Q₂ = 0.992), p values (1.81E‐05) and cross‐validation from the permutation test (Figure 3B; left and central panel). Thirteen major metabolites with VIP > 1 and p values < 0.05 were identified, including 8 sugars (D‐(‐)‐tagatofuranose, D‐(‐)‐fructofuranose, D‐psicopyranose, D‐(‐)‐fructose, D‐fructose, β‐D‐glucopyranose, D‐glucose [22.39 min] and D‐glucose [22.67 min]), 3 acidic compounds (malic acid, D‐gluconic acid and galactaric acid), 1 lipid (palmitic acid), and 1 sugar alcohol (glycerol) (Table 2). These different metabolites were displayed using heatmaps of the average log₂‐fold change values (Figure 3B; right panel).

FIGURE 3.

N‐PDNV and I‐PDNV metabolite composition. Partial least squares discriminant analysis (PLS‐DA) score plots of the N‐PDNV and I‐PDNV metabolites (n = 5) analyzed using quadrupole time‐of‐flight mass spectrometry (UPLC‐Q‐TOF‐MS; A) and gas chromatography‐mass spectrometry (GC‐MS; B). The PLS‐DA model was qualified based on the R²X, R²Y, Q² and p values, and validated via cross‐validation with a permutation test. The heat maps represent the mean of the log₂ fold‐change values of I‐PDNVs relative to N‐PDNVs.

TABLE 1.

Major N‐PDNVs and I‐PDNVs metabolites identified via UPLC‐Q‐TOF‐MS analysis.

No.	Tentative identification	RT (min)	Observed m/z [M+H]+	Theoretical m/z [M+H]+	Predicted molecular formula	Mass error (ppm)	MSI level	Evidence	Fragment ions (m/z)	VIP	p value
1	Guanine	0.76	152.0573	152.0567	C5H5N5O	4.10	2	MassBank	136, 110, 84	1.4775	8.290E‐11
2	Adenine	0.77	136.0607	136.0618	C5H5N5	−7.84	2	MassBank	119, 92	1.4777	7.972E‐08
3	Hypoxanthine	1.05	137.0448	137.0458	C5H4N4O	−7.20	2	MoNA	119, 110, 94	1.4783	2.802E‐07
4	Guanosine	2.23	284.0987	284.0989	C10H13N5O5	−0.83	2	MassBank	152, 135, 110	1.4043	2.255E‐05
5	Methyladenosine isomera	2.70	282.1195	282.1197	C11H15N5O4	−0.63	2	MassBank	136, 119, 94	1.4790	4.951E‐09
6	Alanyl‐isoleucine	3.10	203.1382	203.1390	C9H18N2O3	−4.03	3	CFM‐ID (predicted)	140, 84	1.3295	3.54E‐07
7	L‐tryptophan	3.48	205.0964	205.0971	C11H12N2O2	−3.64	2	MassBank	188, 146, 143, 130, 118, 115	1.4776	7.131E‐13
8	Isoleucyl‐threonine	3.89	233.1493	233.1496	C10H20N2O4	−1.19	3	CFM‐ID (predicted)	215, 169, 88	1.4762	5.504E‐07
9	Valyl‐valine	4.04	217.1547	217.1547	C10H20N2O3	0.15	3	CFM‐ID (predicted)	200, 169	1.4763	6.679E‐07
10	Cannabisativine	4.69	382.3058	382.3064	C21H39N3O3	−1.61	3	CFM‐ID (predicted)	364, 250, 198	1.3249	1.49E‐05
11	Terestigmine	9.96	376.2589	376.2595	C21H33N3O3	−1.51	4	Not available	218, 209, 125	1.4259	4.823E‐04
12	Phytosphingosine	10.31	318.2999	318.3003	C18H39NO3	−1.15	2	MassBank	300, 282	1.4674	1.436E‐05

Note: p values were analyzed by Duncan's test.

Abbreviation: MSI, metabolomics standards initiative; No, number; RT, retention time; UPLC‐Q‐TOF‐MS, ultra‐high‐performance liquid chromatography quadrupole time‐of‐flight mass spectrometry; VIP, variable importance in the projection.

^a Methyladenosine isomer mainly refers to N⁶‐methyladenosine (m⁶A) or N¹‐methyladenosine (m¹A), although other minor methylation sites (e.g., m³A, m⁷A, Am) cannot be excluded.

TABLE 2.

Major N‐PDNVs and I‐PDNVs metabolites via GC‐MS analysis.

No.	RT (min)	Compounds	Observed RI		Similarity index	Molecular formula	VIP	p value
1	11.27	Glycerol (3TMS derivative)	1270	87		C12H32O3Si3	1.0802	1.6866E‐04
2	15.53	Malic acid (3TMS derivative)	1483	78		C13H30O5Si3	0.9590	1.3852E‐03
3	20.96	D‐(‐)‐Tagatofuranose (5TMS derivative)	1696	90		C21H52O6Si5	1.1036	8.0685E‐12
4	21.09	D‐(‐)‐Fructofuranose (5TMS derivative)	1804	91		C21H52O6Si5	1.1028	5.8132E‐11
5	21.15	D‐Psicopyranose (5TMS derivative)	1808	90		C21H52O6Si5	1.0810	1.0002E‐06
6	22.01	D‐(‐)‐Fructose (5TMS derivative, methyloxime, anti)	1863	92		C22H55NO6Si5	1.0968	1.6527E‐05
7	22.16	D‐Fructose (5TMS derivative, methyloxime, syn)	1873	88		C22H55NO6Si5	1.0970	1.1626E‐05
8	22.34	β‐D‐Glucopyranose (5TMS derivative)	1884	93		C21H52O6Si5	1.1005	1.1988E‐09
9	22.39	D‐Glucose (5TMS derivative, methyloxime, 1Z)	1888	90		C22H55NO6Si5	1.1037	3.7559E‐12
10	22.67	D‐Glucose (5TMS derivative, methyloxime, 1E)	1906	78		C22H55NO6Si5	1.1042	1.5482E‐12
11	23.82	D‐Gluconic acid (6TMS derivative)	1990	69		C24H60O7Si6	1.0793	1.2571E‐06
12	24.37	Galactaric acid (6TMS derivative)	2034	76		C24H58O8Si6	0.9162	3.1384E‐03
13	24.48	Palmitic acid (TMS derivative)	2043	92		C19H40O2Si	0.7970	1.8571E‐02

Note: p values were analyzed by Duncan's test. RI values were calculated using an n‐alkane series. According to the Metabolomics Standards Initiative (MSI), four compounds (D‐(‐)‐fructose, D‐fructose, and D‐glucose) were identified at Level 1 (authentic standards) and the remainder were annotated at Level 2 (putative identification without authentic standards), except for D‐gluconic acid and galactaric acid, which were classified as Level 3.

Abbreviations: GC‐MS, gas chromatography‐mass spectrometry; No, number; RI, retention index; RT, retention time; TMS, timethylsiyl; VIP, variable importance in the projection.

3.3. Maturation and Activation of DCs by N‐PDNVs and I‐PDNVs

To assess the immunostimulatory potential of N‐PDNVs and I‐PDNVs, their effects on cell viability, as well as the maturation and activation of DCs, were evaluated. For cell viability assessment, DCs were treated with varying concentrations (1, 2 and 5 µg/mL) of PDNVs for 19 h. To serve as a positive control for cytotoxicity, cells were treated with STS, a known inducer of cell death. Neither N‐PDNVs nor I‐PDNVs affected cell viability, even at the highest concentration tested (Figure 4A). Next, extracellular cytokine production was evaluated in DCs treated with PDNVs at concentrations of 1, 2.5 and 5 µg/mL for 19 h. In all in vitro experiments, LPS was used as a positive control for DC maturation and activation, and was applied at a concentration of 100 ng/mL. As shown in Figure 4B, both N‐PDNVs and I‐PDNVs significantly enhanced the secretion of pro‐inflammatory cytokines TNF‐α and IL‐12p70. Notably, at 5 µg/mL, I‐PDNVs induced a statistically greater secretion of these cytokines than N‐PDNVs. In contrast, IL‐10 levels remained unaffected by treatment with N‐PDNVs or I‐PDNVs. The ability of PDNVs to induce pro‐inflammatory cytokines was further validated by intracellular staining. Following a 12 h treatment with 5 µg/mL PDNVs, intracellular expression of IL‐12p70 and TNF‐α was analyzed in DCs, confirming elevated cytokine production (Figure 4C). Moreover, DCs treated with 5 µg/mL of N‐PDNVs or I‐PDNVs for 19 h exhibited increased expression of surface maturation markers, including CD80, MHC‐I and MHC‐II, compared with non‐treated controls. I‐PDNVs induced a significantly higher expression of these markers than N‐PDNVs (Figure 4D). Since mature DCs typically exhibit reduced antigen uptake and enhanced antigen presentation capacity, we further explored these functional properties. DCs treated with LPS or PDNVs (5 µg/mL) for 19 h were incubated with FITC‐conjugated dextran at 37°C (active uptake) or 4°C (negative control) for 30 min. Both LPS‐ and PDNVs‐treated DCs demonstrated a significantly reduced uptake of dextran at 37°C compared with non‐treated controls. Notably, DCs treated with I‐PDNVs exhibited lower dextran uptake than those treated with N‐PDNVs (Figure 4E). Antigen‐presenting capacity was further assessed by evaluating MHC‐I and MHC‐II pathways using OVA protein and Eα_44–76 peptide, respectively. Non‐treated DCs and those treated with LPS or PDNVs (5 µg/mL) were incubated with either OVA protein or Eα_44–76 peptide for 19 h. Both the LPS‐ and PDNVs‐treated DCs exhibited an enhanced formation of the OVA_257–264/MHC‐I and Eα_52‐68/MHC‐II complexes compared with non‐treated DCs, indicating increased antigen presentation (Figure 4F). However, no statistically significant difference was observed between N‐PDNVs‐ and I‐PDNVs‐treated groups. Finally, to rule out the possibility that the immunostimulatory effects were due to endotoxin contamination, PDNVs (5 µg/mL) and LPS were pre‐incubated with 5 µg/mL PMB (a cyclic polypeptide antibiotic known to bind and neutralize LPS) for 1 h at 25°C before treatment (Figure S2). Subsequently, DCs were treated with either PMB‐treated or untreated LPS and PDNVs for 19 h, and the expression levels of pro‐inflammatory cytokines (TNF‐α and IL‐12p70; Figure S2A) and surface molecules (CD80, MHC‐I, MHC‐II; Figure S2B) were analyzed. As expected, the expression of pro‐inflammatory cytokines and surface molecules induced by LPS was abrogated when LPS was pretreated with PMB. In contrast, PDNVs retained their ability to induce pro‐inflammatory cytokines and surface markers even after PMB exposure, supporting the endotoxin‐free nature of PDNVs. The flow‐cytometric gating strategies used to evaluate DC maturation, along with the complete antibody fluorescence panels, are provided in Figure S3. Taken together, these results demonstrate that both N‐PDNVs and I‐PDNVs effectively induce DC maturation. I‐PDNVs, in particular, exhibit superior immunostimulatory activity, as evidenced by enhanced production of Th1‐polarizing cytokines and greater upregulation of maturation markers compared with N‐PDNVs.

FIGURE 4.

Immunostimulatory capacity of N‐PDNVs and I‐PDNVs in dendritic cells (DCs) maturation. (A) Cell viability of BMDCs treated with various concentrations of N‐PDNVs (1, 2 and 5 µg/mL), I‐PDNVs (1, 2 and 5 µg/mL), and 50 nM staurosporine (STS) for 19 h, measured using an EX‐cytox solution. (B) Extracellular cytokine levels (TNF‐α, IL‐12p70 and IL‐10) in the culture supernatant of BMDCs treated with N‐PDNVs (1, 2.5 and 5 µg/mL), I‐PDNVs (1, 2.5 and 5 µg/mL), or LPS (100 ng/mL) for 19 h, determined by ELISA. (C) Intracellular cytokine levels (TNF‐α, IL‐12p70 and IL‐10) in BMDC treated with N‐PDNVs (5 µg/mL), I‐PDNVs (5 µg/mL) or LPS (100 ng/mL) plus a transport inhibitor cocktail for 12 h, were assessed using flow cytometry following intracellular cytokine staining. (D) Surface molecule expression of CD80, MHC‐I and MHC‐II on CD11c⁺BMDC treated with N‐PDNVs (5 µg/mL), I‐PDNVs (5 µg/mL) or LPS (100 ng/mL) for 19 h, analyzed by flow cytometry. (E) Antigen uptake ability of BMDCs treated with N‐PDNVs (5 µg/mL), I‐PDNVs (5 µg/mL) or LPS (100 ng/mL) for 19 h, assessed via incubation with FITC‐conjugated dextran at 37°C and 4°C for 30 min, followed by flow cytometric analysis of CD11c⁺dextran⁺ cells. (F) Antigen‐presenting capacity of MHC‐I and MHC‐II on CD11c⁺BMDCs, assessed after treatment with N‐PDNVs (5 µg/mL), I‐PDNVs (5 µg/mL) or LPS (100 ng/mL) for 19 h in the presence of the ovalbumin (OVA) protein or Eα_44–76 peptide. The OVA_257–264/MHC‐I and Eα_52–68/MHC‐II complexes on the CD11c⁺ cells were detected using fluorescence‐conjugated 25‐D1.16 and Y‐Ae antibodies, respectively. Data are shown as mean ± SD (n = 3) and represent one of three independent experiments. Statistical analysis was performed using one‐way ANOVA with Tukey's post hoc test or unpaired t‐test. *p < 0.05, **p < 0.01 and ***p < 0.001. Non, non‐treated control.

3.4. DC Maturation by N‐PDNVs and I‐PDNVs via TLR2 and TLR4 Recognition

To elucidate the signalling pathways underlying DC maturation induced by N‐PDNVs and I‐PDNVs, we focused on the role of TLRs, which serve as essential pattern recognition receptors that initiate innate immune responses by detecting pathogen‐associated molecular patterns (Kawai and Akira 2011; Duan et al. 2022). Specifically, TLR2 and TLR4 were selected for investigation due to their well‐established roles in recognizing diverse microbial components and promoting immune activation (Mukherjee et al. 2016; Yang et al. 2015). To determine whether the maturation and activation of DCs by PDNVs occur via TLR2 or TLR4 signalling, we examined the functional phenotypes of BMDCs derived from WT, TLR2^−/− and TLR4^−/− mice (referred to as WT‐DCs, TLR2^−/−‐DCs and TLR4^−/−‐DCs, respectively). BMDCs were treated with N‐PDNVs or I‐PDNVs (5 µg/mL), Pam3 (100 ng/mL), or LPS (100 ng/mL) for 19 h, followed by assessment of extracellular cytokine production (Figure 5A) and surface molecule expression (Figure 5B). Pam3 and LPS, well‐established agonists of TLR2 and TLR4 respectively, were used as reference controls for DC maturation. In this context, Pam3 served as a positive control in WT‐DCs and TLR4^−/−‐DCs, but as a negative control in TLR2^−/−‐DCs; conversely, LPS acted as a positive control in WT‐DCs and TLR2^−/−‐DCs, but as a negative control in TLR4^−/−‐DCs. Interestingly, analysis of TNF‐α production revealed that I‐PDNVs treatment led to reduced TNF‐α secretion in both TLR2^−/−‐DCs and TLR4^−/−‐DCs compared with WT‐DCs. In contrast, N‐PDNVs treatment induced TNF‐α production in TLR2^−/−‐DCs to levels comparable with WT‐DCs, while a reduction was only observed in TLR4^−/−‐DCs. Regarding IL‐12p70, both N‐PDNVs and I‐PDNVs effectively stimulated IL‐12p70 production in WT‐DCs and TLR2^−/−‐DCs, whereas IL‐12p70 levels were markedly diminished in TLR4^−/−‐DCs (Figure 5A). We next assessed the expression of surface maturation markers CD80, MHC‐I, and MHC‐II using the flow‐cytometric gating strategy outlined in Figure S3B. For CD80, both PDNVs induced comparable expression in WT‐DCs and TLR2^−/−‐DCs, but failed to upregulate CD80 in TLR4^−/−‐DCs. Moreover, the expression levels of MHC‐I and MHC‐II induced by PDNV treatment were significantly lower in TLR2^−/−‐DCs and TLR4^−/−‐DCs than in WT‐DCs (Figure 5B). Finally, to determine whether TLR2 or TLR4 are involved in the uptake of PDNVs, 5 µg/mL CFSE‐labelled N‐PDNVs (CFSE‐N‐PDNVs) and I‐PDNVs (CFSE‐I‐PDNVs) were incubated with WT‐DCs, TLR2^−/−‐DCs and TLR4^−/−‐DCs for 30, 60, 90 or 120 min. Flow cytometric analysis showed no significant differences in cellular uptake across all DC types for either PDNVs (Figure 5C). Collectively, these findings suggest that TLR2 and TLR4 are not required for the internalization of PDNVs but are critical mediators of the DC maturation responses induced by N‐PDNVs and I‐PDNVs.

FIGURE 5.

N‐PDNVs‐ and I‐PDNVs‐induced DC maturation through TLR2‐ and TLR4‐dependent pathway. (A) Extracellular cytokine levels (TNF‐α, IL‐12p70) in BMDCs derived from wild type (WT), TLR2 (TLR2^−/−)‐ and TLR4 (TLR4^−/−)‐deficient mice following treatment with Pam3 (100 ng/mL), LPS (100 ng/mL), N‐PDNVs (5 µg/mL) or I‐PDNVs (5 µg/mL) for 19 h using ELISA. (B) Surface molecule expression of CD80, MHC‐I and MHC‐II on BMDCs from WT, TLR2^−/− and TLR4^−/− mice after stimulation as described in (A), measured by flow cytometry and presented as mean fluorescence intensity (MFI) (C) Cellular uptake of CFSE‐labelled N‐PDNVs (CFSE‐N‐PDNVs, 5 µg/mL) and CFSE‐labelled I‐PDNVs (CFSE‐ I‐PDNVs, 5 µg/mL) in BMDCs derived from WT, TLR2^−/− and TLR4^−/− mice at 30, 60, 90 and 120 min, analyzed by flow cytometry. Statistical analysis was performed using one‐way ANOVA with Tukey's post hoc test or unpaired t‐test. Data are presented as mean ± SD (n = 3) and are representative of three independent experiments. *p < 0.05, **p < 0.01, ***p < 0.001. Non, non‐treated control.

3.5. T and NK Cell Activation by N‐PDNVs‐ and I‐PDNVs‐Treated DCs

To evaluate whether the N‐PDNVs‐ and I‐PDNVs‐treated DCs have the capacity to activate T cells and NK cells, their immunostimulatory potential was assessed using co‐culture systems. DCs were first stimulated with PDNVs (5 µg/mL) or LPS for 2 h and then co‐cultured with CFSE‐labelled T cells or NK cells for 3 days and 1 day, respectively. As shown in Figure 6A, DCs treated with N‐PDNVs, I‐PDNVs, or LPS significantly promoted T cell proliferation—as indicated by increased CFSE dilution—and induced the production of Th1‐type cytokine (IFN‐γ) and Th17‐type cytokine (IL‐17A), compared with untreated DCs (non‐DC). In contrast, neither N‐PDNVs‐treated DCs (N‐PDNVs‐DCs)‐ nor I‐PDNVs‐treated DCs (I‐PDNVs‐DCs) stimulated the secretion of the Th2‐type cytokine IL‐5, whereas LPS‐treated DCs (LPS‐DCs) significantly increased IL‐5 levels. Interestingly, the I‐PDNVs‐DCs induced significantly higher levels of IFN‐γ and IL‐17A in T cells than those treated with N‐PDNVs. We next assessed the capacity of PDNVs‐treated DCs to activate NK cells by measuring intracellular IFN‐γ levels, the degranulation marker CD107a, and the cytotoxic granule marker G/B. As shown in Figure 6B, NK cells co‐cultured with the I‐PDNVs‐DCs exhibited a marked increase in the proportion of triple‐positive (CD107a⁺IFN‐γ⁺G/B⁺) cells, compared to NK cells co‐cultured with non‐DCs. In contrast, N‐PDNVs‐DCs did not elicit a comparable increase in this triple‐positive population. However, both the N‐PDNVs‐DCs and I‐PDNVs‐DCs enhanced the proportion of the CD107a⁺G/B⁺ double‐positive NK cells. These findings suggest that N‐PDNVs and I‐PDNVs can effectively stimulate DCs to drive robust T cell responses and enhance NK cell activation. Importantly, I‐PDNVs exhibited a stronger immunostimulatory capacity than N‐PDNVs, underscoring their superior potential as immunomodulatory agents.

FIGURE 6.

Activation of T cells and NK cells by N‐PDNVs‐ and I‐PDNVs‐treated DCs. (A) T cell proliferation and activation induced by non‐treated DCs (non‐DC), LPS‐treated DCs (LPS‐DC, 100 ng/mL), N‐PDNV‐treated DCs (N‐PDNVs‐DC, 5 µg/mL) and I‐PDNV‐treated DCs (I‐PDNVs‐DC, 5 µg/mL). CFSE‐labelled T cells were co‐cultured with the indicated DCs for 3 days. T cell proliferation was analyzed by flow cytometry based on CFSE dilution, and T cell activation was assessed by quantifying extracellular cytokine levels (IFN‐γ, IL‐5, IL‐17A) in the culture supernatants using cytokine‐specific ELISA kits. Flow‐cytometric analysis of CFSE‐labelled T cells was performed by first identifying lymphocyte populations based on forward scatter (FSC) and side scatter (SSC) parameters, followed by subsequent analysis of CFSE fluorescence within this gated population. Data are presented as mean ± SD (n = 3) and are representative of three independent experiments. (B) Analysis of NK cell activation induced by N‐PDNVs‐DC and I‐PDNVs‐DC, conducted by co‐culturing NK cells with non‐DC, LPS‐DC, N‐PDNVs‐DC and I‐PDNVs‐DC for 24 h. Following co‐culture, intracellular cytokine staining for IFN‐γ, CD107a and Granzyme B (G/B) was performed, and the frequencies of IFN‐γ⁺, CD107a⁺ and G/B⁺ cells within the CD3⁻NK1.1⁺ NK cell population were evaluated by flow cytometry. CD3⁻NK1.1⁺ NK cells were identified from the lymphocyte gate (FSC/SSC). Data are presented as mean ± SD (n = 5) and are representative of three independent experiments. Statistical analysis was performed using one‐way ANOVA with Tukey's post hoc test or unpaired t‐test. *p < 0.05, **p < 0.01 and ***p < 0.001.

3.6. Immune‐Enhancing Effect of I‐PDNVs on CTX‐Induced Immunosuppressed Mice

Based on the superior immunostimulatory effects of I‐PDNVs compared with N‐PDNVs observed in the in vitro experiments, we next evaluated their potential to restore immune function in a cyclophosphamide (CTX)‐induced immunosuppressed mouse model. Briefly, the control group received PBS intraperitoneally (100 µL) for 5 days, followed 3 days later by oral PBS for 5 days. The CTX+PBS group received CTX (80 mg/kg, intraperitoneal) for 5 days, then oral PBS for 5 days. The CTX+I‐PDNVs group received CTX as above, followed 3 days later by oral I‐PDNVs (100 µg/mouse) for 5 days (Figure 7A). Therapeutic efficacy was evaluated by monitoring changes in body weight (Figure 7B), cytotoxicity (Figure 7C) and immune cell responses (Figure 7D–G). Body weight was recorded on Days 6, 9, 12 and 14 after the start of the experiment (Figure 7B). As expected, CTX‐injected mice exhibited a significant reduction in body weight compared with the control group, whereas continuous oral administration of I‐PDNVs significantly mitigated CTX‐induced weight loss. To evaluate cell viability, splenocytes from each group were subjected to Annexin V and PI staining (Figure 7C). The CTX+PBS group displayed a marked increase in apoptotic and necrotic cell populations (Annexin V⁻PI⁺, Annexin V⁺PI⁺ and Annexin V⁺PI⁻) compared with the control group, while these populations were notably reduced in the CTX+I‐PDNVs group. T cell profiling further revealed that CTX injection decreased both the frequencies (Figure 7D) of CD4⁺ and CD8⁺ T cells and the functional activity (Figure 7E) of multiple subsets, including Th1 (IFN‐γ⁺CD3⁺CD4⁺), Th2 (IL‐5⁺CD3⁺CD4⁺), Th17 (IL‐17A⁺CD3⁺CD4⁺) and cytotoxic T lymphocyte (CTL; IFN‐γ⁺CD3⁺CD8⁺), relative to the control group. In contrast, the CTX+I‐PDNVs group exhibited recovery of CD4⁺ and CD8⁺ T cell frequencies (Figure 7D), along with significant restoration of Th1, Th17 and CTL functional activity (Figure 7E), compared to the CTX+PBS group. The flow‐cytometric gating strategies used to define T‐cell frequencies and T‐cell functional subsets are presented in Figure S4A,B, respectively. Finally, to determine whether I‐PDNVs could reverse CTX‐induced NK cell suppression, we examined both NK cell frequency (Figure 7F) and activation status (Figure 7G), including the expression of IFN‐γ, G/B, and CD107a. The flow‐cytometric gating strategies used to evaluate NK‐cell frequency and activation are provided in Figure S5B,C, respectively. The CTX+I‐PDNVs group showed restored NK cell frequency (Lineage⁻NK1.1⁺ cells; Figure 7F) and activation (Figure 7G) relative to the CTX+PBS group, with a particularly notable increase in the proportion of CD107a⁺IFN‐γ⁺G/B⁺ triple‐positive NK cells. Collectively, these results demonstrate that I‐PDNVs can counteract CTX‐induced systemic immunosuppression by restoring Th1, Th17 and CTL responses and effectively enhancing NK cell activation.

FIGURE 7.

Immune response induced by I‐PDNVs administration in immunosuppressed mice. (A) Schematic overview of the experimental design showing the oral administration of I‐PDNVs in the cyclophosphamide (CTX)‐induced immunosuppressed model. At Day 14 after the start of the experiment, spleens from each group of mice were collected for analyses of cytotoxicity, immune cell distribution and activation. (B) Body weight changes of mice in the PBS‐injected (Control), CTX+PBS‐injected (CTX+PBS), and CTX+I‐PDNVs‐injected (CTX+I‐PDNVs) groups, monitored on Days 6, 9, 12, and 14 after the start of the experiment. (C) Splenocytes were stained with Annexin V and propidium iodide (PI), and the proportions of necrotic cells (Annexin V⁻PI⁺), late apoptotic cells (Annexin V⁺PI⁺), and early apoptotic cells (Annexin V⁺PI⁻) were determined by flow cytometry. (D) Splenocytes were stained with CD3, CD4, CD8 antibodies, and the proportions of CD3⁺CD4⁺ and CD3⁺CD8⁺ T cells were analyzed by flow cytometry. (E) Splenocytes were stimulated with a cell stimulation cocktail containing a transport inhibitor for 4 h, followed by surface staining with CD3, CD4 and CD8 antibodies and intracellular staining with IFN‐γ, IL‐5 and IL‐17A antibodies. T cell subtypes were subsequently analyzed by flow cytometry: Th1 (IFN‐γ⁺CD3⁺CD4⁺), Th2 (IL‐5⁺CD3⁺CD4⁺) Th17 (IL‐17A⁺CD3⁺CD4⁺) and CTL (IFN‐γ⁺CD3⁺CD8⁺). (F) Splenocytes were stained with NK1.1 and lineage cocktail antibodies and NK (Lineage⁻NK1.1⁺) cells were quantified by flow cytometry. (G) Splenocytes were stimulated with a cell stimulation cocktail containing a transport inhibitor for 4 h, followed by surface staining with NK1.1 and lineage cocktail antibodies and intracellular staining with IFN‐γ, granzyme B (G/B) and CD107a antibodies. The expression of these cytokines and effector molecules was analyzed in Lineage⁻NK1.1⁺ NK cells. The data represent two independent experiments and are expressed as the mean ± SD (n = 5 mice per group). Statistical analysis was performed using Mann‐Whitney rank test for body weight comparison and Kruskal–Wallis with Tukey's post hoc test for multiple group comparisons. *p < 0.05, **p < 0.01 and ***p < 0.001.

3.7. Suppression of Tumour Progression by I‐PDNVs

Based on our findings that I‐PDNVs robustly activate innate immune cells, including DCs and NK cells, and induce Th1 and CTL responses, we next evaluated their therapeutic potential in a tumour model that requires potent cell‐mediated immunity. Mice were subcutaneously inoculated with E.G7 cells, a well‐established murine lymphoma cell line. Six days after tumour implantation, oral administration of I‐PDNVs (100 µg/mouse) was initiated and continued for five consecutive doses to assess their effects on tumour progression and anticancer immune activation (Figure 8A). To begin with, tumour growth was monitored following I‐PDNVs administration. The tumour+I‐PDNVs group exhibited attenuated or delayed tumour progression compared with the tumour+PBS group throughout the observation period (Figure 8B). We then assessed antitumour immune responses. On Day 22, the activation status of innate immune cells—particularly DCs (Figure 8C) and NK cells (Figure 8D)—as well as T cells (Figure 9), was examined. DC maturation was determined by flow cytometric analysis of CD80, MHC‐I and MHC‐II expression on splenic DCs (Lineage⁻NK1.1⁻CD11c⁺ cells) (Figure 8C; top panel). The flow‐cytometric gating strategies used to evaluate splenic DC activation are provided in Figure S5A. The tumour+PBS group showed reduced expression of these maturation markers compared with the control group, whereas the tumour+I‐PDNVs group displayed markedly higher expression levels than the tumour+PBS group (Figure 8C; bottom panel). Next, NK cell activation was assessed by stimulating splenocytes with a cell stimulation cocktail for 4 h, followed by intracellular staining of NK cell (Lineage⁻NK1.1⁺ cells)‐associated functional markers, including IFN‐γ, CD107a, and G/B (Figure 8D). Similar to the trend observed in DC maturation, the tumour‐only group exhibited a pronounced reduction in the proportion of IFN‐γ⁺CD107a⁺G/B⁺ triple‐positive and CD107a⁺IFN‐γ⁺ double‐positive NK cells compared with the control group. These NK cell subsets were restored in frequency in the tumour+I‐PDNVs group. Flow‐cytometric analysis of NK‐cell activation was performed according to the gating strategy presented in Figure S5C. We further analyzed T cell responses following I‐PDNVs administration (Figure 9). Splenocytes were stimulated for 4 h with a cell stimulation cocktail, and activated T cells (CD3⁺CD4⁺ and CD3⁺CD8⁺ subsets) were evaluated for intracellular IFN‐γ, IL‐5, IL‐17A and Foxp3 expression (Figure 9A). The flow‐cytometric gating strategies used to define T‐cell functional subsets are presented in Figure S4B. The tumour‐only group exhibited a reduced frequency of Th1 cells (CD4⁺IFN‐γ⁺) and an increased proportion of Th2 (CD4⁺IL‐5⁺), Th17 (CD4⁺IL‐17A⁺) and Treg (CD4⁺Foxp3⁺) cells compared with the control group. Importantly, I‐PDNVs treatment (tumour+I‐PDNVs group) reversed the tumour‐induced suppression of Th1 cells, and decreased the proportions of Th2 and Treg cells, and concurrently increased the proportion of CTLs (CD8⁺IFN‐γ⁺). Given these results, we next examined whether I‐PDNVs could also promote the generation of polyfunctional effector T cells (Figure 9B,C)—characterized by concurrent production of multiple cytokines and linked to superior antitumour immunity (De Groot et al. 2019). In multifunctional Th1 cell analysis (Figure 9B), the tumour+PBS group displayed reduced frequencies of triple cytokine–positive (IFN‐γ⁺TNF‐α⁺IL‐2⁺) and double cytokine–positive (particularly IFN‐γ⁺TNF‐α⁺) CD4⁺ T cells compared with the control group. In contrast, the tumour+I‐PDNV group exhibited higher frequencies of triple cytokine–positive CD4⁺ T cells and increased proportions of IFN‐γ⁺IL‐2⁺ and TNF‐α⁺IL‐2⁺ double‐positive subsets, whereas the proportion of IFN‐γ⁺TNF‐α⁺ cells, reduced in the tumour+PBS group, remained unchanged. Similarly, multifunctional CTL analysis (Figure 9C) revealed that the tumour+PBS group had lower frequencies of triple cytokine–positive (IFN‐γ⁺TNF‐α⁺IL‐2⁺) and double cytokine–positive (particularly IFN‐γ⁺TNF‐α⁺ and TNF‐α⁺IL‐2⁺) CD8⁺ T cells compared with the control group. In contrast, the tumour+I‐PDNVs group displayed elevated frequencies of triple cytokine–positive CD8⁺ T cells and all double cytokine–positive subsets (IFN‐γ⁺IL‐2⁺, IFN‐γ⁺TNF‐α⁺, and TNF‐α⁺IL‐2⁺) relative to the tumour+PBS group. The flow‐cytometric gating strategies used to define multifunctional CD4 and CD8 T‐cell subsets are presented in Figure S4C. Collectively, these findings indicate that I‐PDNVs not only restore Th1 and CTL responses during tumour progression but also promote the development of multifunctional T cell subsets with potent effector capabilities. This coordinated enhancement of DC maturation, NK cell activation, and polyfunctional T cell responses likely underlies the observed suppression of tumour growth following I‐PDNVs administration.

FIGURE 8.

Antitumour efficacy and innate immune cell activation induced by I‐PDNVs administration in tumour‐bearing mice. (A) Schematic representation of the experimental design showing the oral administration of I‐PDNVs in the E.G7 tumour‐bearing mouse model. (B) Tumour volume in the tumour+PBS‐injected (tumour+PBS) and tumour+I‐PDNVs‐injected (tumour+I‐PDNVs) groups was monitored at 2‐day intervals for 22 days following tumour inoculation. (C and D) At Day 22, spleens from each group were collected for analyses of innate cell activation. (C) Splenocytes were stained with CD80, MHC‐I, MHC‐II, NK1.1, CD11c and lineage cocktail antibodies. The expression of CD80, MHC‐I and MHC‐II in lineage⁻CD11c⁺ splenic DCs was analyzed by flow cytometry and presented as mean fluorescence intensity (MFI). (D) Splenocytes were stimulated with a cell stimulation cocktail containing a transport inhibitor for 4 h, followed by surface staining with NK1.1 and lineage cocktail antibodies and intracellular staining with IFN‐γ, granzyme B (G/B), and CD107a antibodies. The expression of these cytokines and effector molecules was analyzed in Lineage⁻NK1.1⁺ NK cells. The data represent two independent experiments and are expressed as the mean ± SD (n = 5 mice/group). Statistical analysis was performed using Mann–Whitney rank test for tumour volume comparison and Kruskal–Wallis with Tukey's post hoc test for multiple group comparisons. *p < 0.05, **p < 0.01 and ***p < 0.001. Control: PBS‐injected group.

FIGURE 9.

Anti‐tumour T cell immunity induced by I‐PDNVs administration in tumour‐bearing mice. The experiments shown here were performed using splenocytes collected from the mice described in Figure 8. (A and C) Splenocytes were stimulated with a cell stimulation cocktail containing a transport inhibitor for 4 h. (A) Cells were surface‐stained with CD3, CD4 and CD8 antibodies, followed by intracellular staining with IFN‐γ, IL‐5, IL‐17A and Foxp3 antibodies. T cell subtypes were analyzed by flow cytometry: Th1 (IFN‐γ⁺CD3⁺CD4⁺), Th2 (IL‐5⁺CD3⁺CD4⁺), Th17 (IL‐17A⁺CD3⁺CD4⁺), Treg (Foxp3⁺CD3⁺CD4⁺) and CTL (IFN‐γ⁺CD3⁺CD8⁺). (B and C) Cells were surface‐stained with CD3, CD4 and CD8 antibodies, followed by intracellular staining with IFN‐γ, TNF‐α and IL‐2 antibodies. Multifunctional CD4⁺ (B) and CD8⁺ (C) T cells co‐expressing IFN‐γ, TNF‐α and IL‐2 were evaluated by flow cytometry. The data represent two independent experiments and are expressed as the mean ± SD (n = 5 mice/group). Statistical analysis was performed using Kruskal–Wallis with Tukey's post hoc test. *p < 0.05, **p < 0.01, and ***p < 0.001. Control: PBS‐injected group. Tumour+PBS: tumour+PBS‐injected group, Tumour+I‐PDNVs: tumour+I‐PDNVs‐injected group.

3.8. Induction of OVA‐Specific T Cell Responses by Vaccination With OVA Peptide‐Encapsulated I‐PDNVs

Given the potent cellular immune responses elicited by I‐PDNVs, we further investigated their potential as a vaccine delivery system to induce antigen‐specific immunity and corresponding protective effects. Before evaluating in vivo immunogenicity, we first compared the functional and physicochemical stability of OVA_257–264 (known to elicit OVA‐specific CD8⁺ T cell responses)‐encapsulated I‐PDNVs (OVA_257–264+I‐PDNVs) with that of intact I‐PDNVs. Physicochemical stability was assessed by measuring particle size, size distribution, zeta potential and structural integrity using a nanoparticle analyzer (Figure 10A) and cryo‐TEM (Figure 10B). Functional stability was evaluated by treating DCs with intact I‐PDNVs (5 µg/mL) or OVA_257–264‐encapsulated I‐PDNVs (5 µg/mL) for 19 h, followed by assessment of DC maturation (Figure 10C,D). The flow‐cytometric gating strategies used to define DC maturation based on surface molecule expression are presented in Figure S3B. The results showed that OVA_257–264+I‐PDNVs exhibited particle size, size distribution, zeta potential and structural integrity comparable to those of intact I‐PDNVs (Figure 10A,B). Functionally, treatment of DCs with OVA_257–264‐encapsulated I‐PDNVs effectively induced TNF‐α and IL‐12p70 production (Figure 10C), as well as the upregulation of CD80, MHC‐I and MHC‐II expression (Figure 10D), to a level equivalent to that observed with intact I‐PDNVs. These findings indicate that encapsulation of OVA_257–264 within I‐PDNVs does not alter their intrinsic immunostimulatory activity or structural properties. Based on these results, we next evaluated the efficacy of OVA_257–264+I‐PDNVs as a vaccine delivery platform in vivo. Mice were immunized three times at 3‐week intervals with either I‐PDNVs alone, OVA_257–264 peptide alone, or OVA_257–264+I‐PDNVs (Figure 10E). To assess OVA‐specific CD8⁺ T cell responses, splenocytes were collected 1 week after the third immunization, restimulated ex vivo with 20 µg/mL OVA_257–264 peptide, and analyzed by flow cytometry for OVA‐specific IFN‐γ⁺CD8⁺ T cells (Figure 10F). Strikingly, a marked increase in the frequency of OVA‐specific IFN‐γ⁺CD8⁺ T cells was observed only in the OVA_257–264+I‐PDNVs–immunized group, whereas neither I‐PDNVs alone nor OVA_257–264 peptide alone increased these cells relative to the PBS control. Collectively, these findings demonstrate that I‐PDNVs can serve as an effective vaccine delivery system, whereby encapsulation of OVA_257–264 enables the induction of potent OVA‐specific CD8⁺ T cell responses.

FIGURE 10.

Physicochemical, morphological and functional stability of intact and OVA_257–264‐encapsulated I‐PDNVs, and induction of OVA‐specific CD8⁺ T cell responses. (A) Physicochemical stability of intact I‐PDNVs and OVA_257–264‐encapsulated I‐PDNVs (OVA_257–264+I‐PDNVs) was evaluated by TRPS to determine particle size, size distribution (d90/d10 ratio), and zeta potential. Data are presented as mean ± SD (n = 3) and are representative of three independent experiments. Statistical analysis was performed using unpaired t‐test. (B) Vesicular morphology was assessed by cryo‐TEM. Images are representative of two independent experiments. (C and D) Functional stability analysis: BMDCs were treated for 19 h with intact I‐PDNVs (5 µg/mL) or OVA_257–264+I‐PDNVs (5 µg/mL). (C) TNF‐α and IL‐12p70 production in culture supernatants was quantified by ELISA. (D) Cells were surface‐stained with CD11c, CD80, MHC‐I and MHC‐II antibodies, and expression of CD80, MHC‐I and MHC‐II in CD11c⁺ DCs was analyzed by flow cytometry. Data are presented as mean ± SD (n = 4) and are representative of three independent experiments. Statistical analysis was performed using one‐way ANOVA with Tukey's post hoc test. (E) Schematic of the immunization protocol: C57BL/6 mice were immunized intramuscularly 3 times at 1‐week intervals with PBS (Control), I‐PDNVs alone, OVA_257–264 peptide alone, or OVA_257–264+I‐PDNVs. (F) Flow cytometric analysis of OVA‐specific CD8⁺ T cells: splenocytes were collected 1 week after the final immunization, restimulated ex vivo with OVA_257–264 peptide (20 µg/mL) for 12 h, and stained with CD3 and CD8 antibodies, followed by intracellular staining with IFN‐γ antibody. The frequency of CD8⁺IFN‐γ⁺ T cells was analyzed by flow cytometry. The data represent two independent experiments and are expressed as the mean ± SD (n = 5 mice/group). Statistical analysis was performed using Kruskal–Wallis with Tukey's post hoc test. ***p < 0.001. The term “ns” indicates not significant.

3.9. Antitumour Efficacy Induced by OVA Peptide‐Encapsulated I‐PDNVs Immunization

Our previous results demonstrated that encapsulation of the tumour antigen OVA_257–264 within I‐PDNVs could elicit potent tumour antigen–specific immune responses, highlighting their potential to confer strong protective immunity against OVA‐expressing tumours. Based on this rationale, we next assessed the antitumour efficacy in OVA_257–264+I‐PDNVs immunization in vivo. Specifically, mice were immunized three times at 1‐week intervals with PBS alone, I‐PDNVs alone, OVA_257–264 alone, or OVA_257–264+I‐PDNVs. One week after the final immunization, OVA‐expressing E.G7 tumour cells were subcutaneously inoculated into all groups except the PBS‐injected controls, and tumour progression was subsequently monitored (Figure 11A). Tumour growth analysis revealed that only the OVA_257–264+I‐PDNVs‐immunized group exhibited a pronounced delay in tumour progression compared with the tumour‐only group, whereas no protective effect was observed in the I‐PDNV‐alone or OVA_257–264‐alone groups (Figure 11B). To evaluate vaccine‐induced immune modulation, we also analyzed the induction of multifunctional CD8⁺ T cells (Figure 11C), as well as the frequencies of Treg cells (Figure 11D) and exhausted CD8⁺ T cells (Figure 11E). For multifunctional CD8⁺ T cell analysis, splenocytes were isolated 2 weeks after tumour inoculation, stimulated for 4 h with a cell stimulation cocktail, and analyzed by flow cytometry for CD8⁺ T cells expressing IFN‐γ, TNF‐α and IL‐2 (Figure 11C). The flow‐cytometric gating strategies used to define multifunctional CD8 T‐cell subsets are presented in Figure S4C. Compared with the control group, the tumour‐only group displayed reduced proportions of IFN‐γ⁺TNF‐α⁺IL‐2⁺, IFN‐γ⁺TNF‐α⁺ and TNF‐α⁺IL‐2⁺ CD8⁺ T cells. Notably, these multifunctional subsets were significantly increased in the OVA_257–264+I‐PDNVs‐immunized group relative to the tumour‐only group, while no enhancement was detected in the other immunization groups. We next examined the frequency of Treg cells (Foxp3⁺CD4⁺CD3⁺), which typically expand during tumour progression (Figure 11D). As expected, the tumour‐only group showed higher Treg frequencies compared with control mice. Interestingly, suppression in tumour‐associated Treg cell expansion was observed exclusively in the OVA_257–264+I‐PDNVs‐immunized group. Finally, to assess T cell exhaustion, we analyzed the expression of exhaustion‐associated surface markers (TIM‐3 and PD‐1) on CD8⁺ T cells by flow cytometry (Figure 11E). The flow‐cytometric gating strategies used to define exhausted CD8 T cells are presented in Figure S4D. The tumour‐only group exhibited elevated expression of these markers (TIM‐3⁺PD‐1⁺ cells) compared with control mice, whereas a marked reduction was observed only in the OVA_257–264+I‐PDNVs‐immunized group. Collectively, these findings highlight the strong potential of I‐PDNVs as an effective vaccine delivery system. In particular, this platform is characterized by its ability to enhance CD8⁺ T cell responses, reduce Treg cell frequencies and alleviate CD8⁺ T cell exhaustion, thereby promoting robust antitumour immunity.

FIGURE 11.

Antitumour effects induced by OVA_257‐264‐encapsulated I‐PDNVs immunization. (A) Experimental schematic illustrating the antitumour study design. Briefly, mice were divided into the following groups: Control, tumour‐only, I‐PDNVs, OVA_257–264 and OVA_257–264‐encapsulated I‐PDNVs (OVA_257–264+I‐PDNVs). Mice were immunized intramuscularly 3 times at 1‐week intervals with PBS (for Control and tumour‐only groups), I‐PDNVs, OVA_257–264, or OVA_257–264+I‐PDNVs. One week after the final immunization, OVA‐expressing E.G7 tumour cells were inoculated into all groups except the Control group. Antitumour immunity was evaluated 2 weeks after tumour inoculation by analyzing splenocytes. (B) Tumour size was monitored on Days 3, 5, 10, 12 and 14 following tumour inoculations. (C) Splenocytes were stimulated ex vivo with a cell stimulation cocktail containing a transport inhibitor for 4 h. Cells were surface‐stained with CD3 and CD8 antibodies, followed by intracellular staining with IFN‐γ, TNF‐α and IL‐2 antibodies. Multifunctional CD8⁺ T cells co‐expressing IFN‐γ, TNF‐α and IL‐2 were analyzed by flow cytometry. (D) Splenocytes (non‐stimulated condition) were surface‐stained with CD3 and CD4 antibodies, followed by intracellular staining with Foxp3 antibody. Foxp3⁺CD4⁺ regulatory T cells were analyzed by flow cytometry. (E) Splenocytes (non‐stimulated condition) were surface‐stained with CD3, CD8, PD‐1 and Tim‐3 antibodies. The proportion of PD‐1⁺Tim‐3⁺ cells among CD3⁺CD8⁺ T cells was assessed by flow cytometry. The data represent two independent experiments and are expressed as the mean ± SD (n = 5 mice/group). Statistical analysis was performed using Mann–Whitney rank test for tumour growth comparison and Kruskal–Wallis with Tukey's post hoc test for multiple group comparisons. *p < 0.05, **p < 0.01 and ***p < 0.001.

4. Discussion

In this study, we demonstrated that PDNVs derived from C. sativa ARs exhibit distinct immunomodulatory properties depending on the plant hormone used for root induction. I‐PDNVs derived from IBA exposure demonstrated significantly superior immunostimulatory effects compared with N‐PDNVs formed through NAA exposure. While both PDNVs exhibit similar physical characteristics, with average diameters of 118 ± 4 nm (N‐PDNVs) and 123 ± 4 nm (I‐PDNVs) and stable zeta potentials under external stress, their biological effects differed substantially. Compared with N‐PDNVs, I‐PDNVs more effectively promoted DC maturation via the TLR2‐ and TLR4‐dependent pathways, enhanced T‐cell proliferation (particularly IFN‐γ‐ and IL‐17A‐producing T cells), and increased NK cell activity. These enhanced immunostimulatory properties of I‐PDNVs afforded significant therapeutic effects, demonstrating their potential as an effective immunotherapeutic agent in both immunosuppressed (weight recovery, alleviation of cytotoxicity and restoration of Th1, Th17, CTL and NK cell activity) and tumour‐bearing (robust anti‐tumour immunity) mouse models. Furthermore, I‐PDNVs also exhibited superior performance as DDS for anticancer vaccines, as evidenced by the significantly strong OVA_257–264‐specific CD8⁺ T cell response and reduced tumour growth rate in mice immunized with OVA_257–264‐encapsulated I‐PDNVs. These findings highlight the potential of I‐PDNVs as both standalone immunotherapeutic agents and effective DDS with inherent immunostimulatory properties.

Notably, PDNVs derived from C. sativa ARs showed distinct differences in their metabolic composition and immunostimulatory efficacy depending on the hormones used for AR induction. The observed differences between N‐PDNVs and I‐PDNVs are beyond mere technical variation, reflecting the ways hormone‐induced metabolic changes fundamentally shape the biological activity of PDNVs. This is supported by the characteristics of EVs, which can exhibit metabolic variation depending on the culture condition of the parental cells, potentially influencing their bioactivity (Woith et al. 2021; Kocholata et al. 2024; Kim et al. 2022). Specifically, the IBA and NAA hormones used in this study are structurally distinct synthetic auxins. IBA belongs to the indole class and is converted into the natural auxin indole‐3‐acetic acid (IAA) within the cell, whereas NAA, a naphthalene‐based compound, is not converted into IAA. Interestingly, it has been reported that these two auxins not only regulate the intracellular metabolic pathways differently but also influence the metabolite composition of ARs (Strader et al. 2010; Damodaran and Strader 2019; Khadr et al. 2020). Recent studies corroborate this perspective by demonstrating that ARs induced by IBA in Polygonum multiflorum and Allochrusa gypsophiloides (Regel) Schisch exhibited higher total phenolic and flavonoid contents compared with those formed by NAA, leading to enhanced antioxidant and antimicrobial activities (Mursaliyeva et al. 2023; Ho et al. 2021). Consequently, the differences in the immunostimulatory efficacy between N‐PDNVs and I‐PDNVs are most likely attributable to auxin‐specific metabolic reprogramming. showing that I‐PDNVs are enriched in metabolites previously reported to drive DC maturation and immune activation. Specifically, guanine and hypoxanthine promote DC maturation by inducing reactive oxygen species–dependent upregulation of surface molecules and pro‐inflammatory cytokine secretion (Batal et al. 2014; Hahm et al. 2022;Pazmandi et al. 2019; Xu et al. 2016). Similarly, L‐alanine enhances phagocytosis via the TLR4 pathway and sustains T‐cell activation (Jiang et al. 2023; Ron‐Harel et al. 2019), while palmitic acid modulates immune cell activation and inflammatory signalling cascades (Tzeng et al. 2019). Threonine contributes to both adaptive and innate immunity by promoting T‐cell proliferation, enhancing activation and supporting B‐cell and macrophage differentiation (Matthews and Cantrell 2006; Tang et al. 2021). Valine has likewise been implicated in the maturation and functional activation of monocyte‐derived DCs, thereby reinforcing immune responses (Chen et al. 2017; Kakazu et al. 2007). Collectively, these findings indicate that I‐PDNVs harbour a metabolite repertoire capable of potentiating DC maturation and downstream immune activation, providing a mechanistic explanation for their superior immunostimulatory efficacy. This hormone‐specific metabolic enrichment underscores the promise of PDNV engineering as a strategy for developing tailored plant‐based nanotherapeutics for immunotherapy.

It is worth mentioning that the therapeutic effects of I‐PDNVs are accompanied by their ability to activate innate immune cells and T cells in both immunosuppressed and tumour‐bearing mouse models. The induction of DC maturation, NK cell activation and Th1‐ and CTL‐biased T cell responses following I‐PDNVs treatment suggests their potential as effective immunotherapeutic agents for inhibiting disease progression or promoting recovery in various pathological conditions. In particular, these mechanisms play a critical role in the development of treatments and vaccines for intracellular infectious diseases (e.g., tuberculosis) and chronic viral infections (e.g., influenza). Similar to the tumour microenvironment, these infectious diseases can suppress DC maturation or induce a tolerogenic state at the infection site, impairing the interactions between the innate and adaptive immune cells required for pathogen clearance (Kim and Shin 2022; Fernandez‐Sesma et al. 2006; Su et al. 2019; Kim et al. 2017). This immune dysfunction can ultimately lead to disease exacerbation due to insufficient activation of the NK cells, Th1 cells and CTLs (Su et al. 2019; Georgieva et al. 2018; Horst et al. 2011; Guo et al. 2011). Consequently, although this study did not directly confirm the therapeutic effects of I‐PDNVs against infectious diseases, our findings suggest that I‐PDNVs exhibit significant potential to elicit protective immunity, thereby preventing the onset of various infections. This highlights their applicability not only in cancer treatment but also as a promising immunotherapeutic strategy for infectious diseases. Furthermore, the ability of I‐PDNVs to induce immunogenic DCs with a Th1 bias in vitro, as well as their capacity to enhance Th1, CTL, NK cell and multifunctional T cell responses in vivo, further supports this hypothesis.

Finally, our study provides crucial evidence supporting the dual functionality of I‐PDNVs as both DDS and immunotherapeutic agents, serving as a valuable foundation for the development and advancement of more effective DDSs. Most DDSs are designed to serve as carriers that enhance the stability and efficacy of drugs or vaccines, including DNA, RNA and proteins (Vargason et al. 2021; Adepu and Ramakrishna 2021; Ezike et al. 2023). A well‐known example is synthetic liposomes, which share structural similarities with EVs and are commonly used as vaccine delivery systems (Sercombe et al. 2015; Giddam et al. 2012). However, liposome‐based DDSs often face limitations in effectively inducing vaccine‐specific immune responses, particularly cellular immunity, which is essential for vaccine efficacy. To overcome these limitations, many studies focus on developing vaccines by formulating adjuvants—particularly TLR agonists—with liposomes and incorporating vaccines to enhance cellular immunity (Nijen Twilhaar et al. 2022; Shen et al. 2020; Luna et al. 2024; Boks et al. 2015). Ultimately, in vaccine development incorporating DDS, the selection of an appropriate adjuvant has become indispensable. Notably, our findings demonstrate that I‐PDNVs can induce cellular immunity via TLR2 and TLR4 signalling activation, underscoring their potential as an adjuvant. Furthermore, the ability of I‐PDNVs to function as a DDS for cancer vaccines was demonstrated by their capacity to induce vaccine‐specific T cell responses, suppress tumour growth, and reprogram the antitumour immune landscape. In particular, the reduction of Tregs—which are characterized by their immunosuppressive activity on various T cell subsets—and the decrease of exhausted CD8⁺ T cells—defined by progressive loss of effector functions such as cytokine production and cytotoxicity—together with the enhancement of CTL responses, represent a remarkable feature of I‐PDNVs as a vaccine delivery system (Wang et al. 2015; So et al. 2023). These findings further substantiate the dual role of I‐PDNVs as both an adjuvant and a vaccine delivery platform.

Despite these promising findings, our study has several limitations. First, although clear metabolic differences between N‐PDNVs and I‐PDNVs were identified, further investigation is needed to delineate the direct immunological activities of individual metabolites within these PDNVs, particularly to pinpoint the key compounds driving immune enhancement. Elucidating these bioactive components would provide essential mechanistic evidence for the immunostimulatory capacity of I‐PDNVs. Second, beyond metabolomic profiling, integrated lipidomic and genomic analyses are warranted. Such multi‐omics approaches would yield deeper insights into the structural composition, molecular signatures and functional pathways of N‐PDNVs and I‐PDNVs derived from N‐ARs and I‐ARs, respectively. Addressing these gaps will not only refine the mechanistic understanding of PDNV‐mediated immune modulation but also support the rational design of PDNV‐based therapeutics optimized for specific immunotherapy and drug delivery applications.

5. Conclusion

Our study is the first to report the differences in the metabolic properties and immunoenhancing efficacy between I‐PDNVs and N‐PDNVs isolated from C. sativa AR induced by two distinct plant hormones, IBA and NAA. I‐PDNVs strongly promoted a Th1‐biased immune response by inducing both innate and adaptive immune activation, demonstrating their superior potential as an immunotherapeutic agent for immunosuppression and cancer treatment. Furthermore, the multifunctionality of I‐PDNVs is highlighted by validating their potential as an integrated adjuvant and DDS in a cancer vaccine model. Our findings suggest that I‐PDNVs are promising immunotherapeutic candidates not only for cancer treatment, but also for intracellular infectious diseases and chronic viral infections, emphasizing their role as a multifunctional nanomaterial capable of integrating adjuvant and vaccine delivery functions for next‐generation vaccine development.

Author Contributions

Su Hyun Park: writing – original draft, conceptualization, methodology, software, data curation, investigation, formal analysis, validation, resources. Han‐Gyu Choi: methodology, software, data curation, investigation, validation, visualization, formal analysis. Zhun Li: data curation, formal analysis, visualization. Yun Hye Kim: data curation, formal analysis, visualization. Hyeon Jin Lee: data curation, formal analysis, visualization. Ki‐Won Shin: data curation, formal analysis. Hwa‐Jung Kim: methodology, investigation, validation, visualization. Hyung‐Jun Kwon: methodology, investigation. Gimoon Seo: methodology, investigation, validation. Jae Cheol Jeong: funding acquisition, methodology, resources. Young Bae Ryu: funding acquisition, data curation, resources, methodology, supervision. Woo Sik Kim: writing – review and editing, resources, conceptualization, methodology, software, data curation, supervision, formal analysis, validation, investigation, visualization, funding acquisition, writing – original draft, project administration.

Funding

This work was supported by the Korea Research Institute of Bioscience and Biotechnology Research Initiative Programs (Grant/Award Nos. KGM1052511 and KGM5382521), and the Industrial Technology Innovation Infrastructure Program funded by the Ministry of Trade, Industry and Energy and the Korea Institute for Advancement of Technology (Grant No. RS‐2025‐11582968).

Conflicts of Interest

The authors declare that there are no financial or institutional conflicts of interest related to the affiliation of one co‐author with the Korea Vaccine Global Industrialization Foundation Project Group K‐Bio CMO Centre (Andong, Republic of Korea). This organization had no role in the study design, data collection, analysis, interpretation, or manuscript preparation.

Supporting information

Figures: jev270206‐supp‐0001‐FigureS1‐S5.docx

Park, S. H. , Choi H.‐G., Li Z., et al. 2025. “Dual‐Function Plant‐Derived Nanovesicles From Regenerated Cannabis sativa Roots for Immunotherapy and Vaccine Delivery.” Journal of Extracellular Vesicles 14, no. 12: e70206. 10.1002/jev2.70206

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.