Tea-derived extracellular vesicles-mediated PDRN delivery activates cAMP-HIF-1α to restore intestinal homeostasis in inflammatory bowel disease

Abstract

The clinical translation of polydeoxyribonucleotide (PDRN), a bioactive agent with anti-inflammatory and tissue-repair properties, is hindered by significant oral delivery barriers. This study utilizes inflammatory bowel disease (IBD) as a disease model, which is characterized by a vicious cycle of excessive oxidative stress, immunological homeostasis imbalance and intestinal flora dysbiosis in the colon. Leveraging the targeted delivery advantages of tea-derived extracellular vesicles (EV), we have developed them as nanocarriers to overcome the oral delivery challenges of PDRN (PDRN-EV). PDRN-EV exhibits excellent structural stability under simulated gastrointestinal conditions. In vitro studies demonstrated that PDRN-EV exerts therapeutic effects via dual synergistic mechanisms of anti-oxidative stress and immunomodulation. Furthermore, the anionic surface properties of PDRN-EV promote selective accumulation at inflammatory sites, while surface-exposed monogalactosyldiacylglycerol and digalactosyldiacylglycerol (MGDG/DGMG) galactolipids mediate the specific targeting phagocytosis of macrophages. In vivo experiments conducted in a dextran sulfate sodium (DSS)-induced colitis model demonstrated that orally-administered PDRN-EV significantly alleviates adverse characteristics such as pro-inflammatory responses and impaired intestinal barrier function. The underlying mechanism involves driving macrophage M2 polarization through activation of the cAMP/HIF-1α signaling pathway, promoting DNA replication, and restoring microbiota equilibrium. This work establishes a novel oral nanotherapeutic strategy for IBD that circumvents gastrointestinal degradation and off-target effects.

Graphical abstract

The development of an orally-administered nanocomposite based on polydeoxyribonucleotide loaded tea-derived extracellular vesicles (PDRN-EV) is described for inflammatory bowel disease (IBD) therapy, which ameliorates the pro-inflammatory responses and restores intestinal barrier functions through synchronous ROS-scavenging, immunomodulation, and modulation of gut microbiota. This study bridges the critical gap between nucleic acid pharmacology and clinical translation.

Highlights

• •Tea-derived extracellular vesicle (EV)-based oral PDRN delivery system overcoming gastrointestinal degradation barriers.
• •EV carriers demonstrate bioactivities synergizing with PDRN DSS-induced colitis.
• •EV surface galactolipids enable receptor-mediated macrophage targeting.
• •PDRN-EV activates cAMP/HIF-1α axis to drive M2 polarization.
• •A multifunctional integrated nanoplatform for nucleic acid drug delivery.

1. Introduction

Inflammatory bowel disease (IBD), including ulcerative colitis and crohn's disease, is characterized by intestinal barrier dysfunction and immune dysregulation [1]. It has been found that IBD is characterized by a vicious loop of excessive oxidative stress, immunological homeostasis imbalance, and intestinal microflora disturbance. Nanotechnology has shown great potential in advancing IBD treatment through approaches focused on its three key pathological characteristics [2]. Antioxidative strategies employ nanozymes [3], bioactive compounds [4], and natural antioxidants [5]. Anti-inflammatory approaches use nanoparticles loaded with nucleic acids [6], clinical or natural drugs [7]. Microbial modulation involves nanomaterial-assisted antibiotic delivery and fecal microbiota transplantation (FMT) [8]. Compared to conventional systems, nanocarriers improve drug efficacy and reduce side effects [9]. However, limitations remain—such as rapid clearance by the reticuloendothelial system and reduced accumulation at lesion sites [10,11], which currently restrict their broader clinical application in IBD.

Polydeoxyribonucleotide (PDRN), a high-molecular-weight (50–1500 kDa) deoxyribonucleotide complex extracted from salmon (Oncorhynchus spp.) sperm DNA, has emerged as a cutting-edge therapeutic agent in regenerative medicine due to its multifaceted pharmacological activities [12]. Modern pharmacological studies have elucidated the molecular mechanisms underlying anti-inflammatory, pro-angiogenic, and tissue-repair effects of PDRN, primarily mediated through adenosine A_2A receptor activation and DNA salvage pathway modulation [13]. In preclinical models, PDRN have shown significant translational promise. For example, intraperitoneal administration suppressed pro-inflammatory cytokine release (e.g., TNFα, IL-6) and enhanced colonic epithelial regeneration in dextran sulfate sodium (DSS)-induced colitis [14]. In addition, topical delivery upregulated the expression of vascular endothelial growth factor (VEGF), promoting vascularization and re-epithelialization in diabetic cutaneous wounds [15]. Notably, the absence of immunogenic proteins or peptides in its molecular composition of PDRN confers exceptional biocompatibility, positioning it as a clinically viable therapeutic candidate [12].

Despite its therapeutic potential, the clinical value of PDRN is limited by poor administration techniques. Current research predominantly focuses mostly on intraperitoneal or transdermal administration, while oral delivery, as the most convenient and patient-friendly route, remains underexplored [[16], [17], [18], [19]]. From a translational perspective, oral formulations have distinct advantages for treating gastrointestinal disorders such as IBD. They not only maximize drug concentration at intestinal lesions while minimizing systemic toxicity, but also improve patient adherence, which is beneficial for long-term management of chronic conditions [20,21]. However, oral delivery of PDRN, as a representative nucleic acid drug, encounters three critical barriers. The first issue is gastrointestinal instability, where naked PDRN undergoes rapid degradation in gastric acid (pH 1.0–3.0) and pancreatic nucleases [[22], [23], [24]]. The second issue is inefficient targeting. Thickening of the mucus layers and interruption of epithelial tight junctions at inflamed sites can reduce drug accumulation, while the lack of active targeting mechanisms further exacerbates the drug accumulation. The third obstacle is cellular internalization barriers, where the high molecular weight and strong negative charge of PDRN impede efficient cellular uptake via endocytosis, resulting in low intracellular delivery efficiency of free PDRN [25,26].

To overcome these challenges, nano-delivery systems have been proposed as a key strategy. Current approaches employ calcium carbonate or chitosan-based carriers to encapsulate PDRN via electrostatic interactions, achieving modest improvements in cellular uptake [15,18]. Nevertheless, critical limitations still remain, mainly in the lack of sufficient gastrointestinal stability of synthetic carriers and the lack of inflammation-specific targeting capabilities, leading to off-target accumulation and compromised therapeutic efficacy. Crucially, no study has systematically addressed the delivery requirements for orally-administered PDRN formulations, including gastrointestinal protection, mucus penetration, and cellular targeting, which has hindered its clinical translation.

Plant-derived extracellular vesicles, characterized by a phospholipid bilayer enriched with proteins, RNAs, and secondary metabolites, exhibit superior gastrointestinal stability due to their unique lipid composition—such as high phosphatidic acid (PA) content and absence of cholesterol [27]. Recent advances have established plant EV as a revolutionary platform for oral drug delivery in IBD therapy [28,29]. These vesicles can be further engineered or hybridized to enhance therapeutic efficacy [30,31]. Moreover, their bilayer structure facilitates efficient nucleic acid loading via co-incubation or sonication, achieving high encapsulation efficiency [32]. Of particular interest, tea-derived EV enriched with surface galactolipids enable targeted phagocytosis by macrophages through galactose-receptor binding [33,34]. They have also been successfully applied for the oral delivery of siRNA in the treatment of acute aortic dissection [35]. These findings provide critical insights for developing PDRN-loaded tea-derived EV delivery systems.

This study attempted to load PDRN into tea-derived EV (PDRN-EV) as an ideal orally-administered candidate drug for IBD treatment by utilizing the dual effects of PDRN on anti-inflammation and tissue repair (Scheme 1). Consequently, this study presented two innovations. First, orally-administered PDRN delivery breakthrough: we pioneered the oral delivery feasibility of PDRN by tea-derived EV carriers that overcome gastrointestinal degradation and off-target distribution barriers. Second, synergistic therapy paradigm: we proposed a “carrier-drug collaborative therapy” framework, demonstrating that galactolipid-enriched tea-derived EV bilayers not only enable inflammation-targeted PDRN delivery but also confer intrinsic bioactivities, resulting in therapeutic amplification in IBD. Our designed PDRN-EV nanoplatform can be potentially applied as a breakthrough strategy for the IBD treatment.

Scheme 1.

Schematic illustration of tea-derived extracellular vesicles-mediated PDRN delivery activates cAMP-HIF-1α to restore intestinal homeostasis in inflammatory bowel disease.

2. Materials and methods

2.1. Materials

Polydeoxyribonucleotide (PDRN) was purchased from Sigma-Aldrich (St. Louis, MO, USA). 3, 3′-Dioctadecyloxacarbocyanine perchlorate (DiO) was obtained from Yeasen Biotechnology (China). Cellular senescence assay kit, DPPH kit, ABTS kit and JC-1 kit were purchased from Beyotime (China). DSS (36,000–50,000 Da) was purchased from MP biomedicals. ELISA kits were sourced from NeoBioscience Technology Co. Ltd. (China).

2.2. Extraction of EV and PDRN loading

Fresh green tea leaves were rinsed with deionized water to remove coarse stems and diseased tissues, followed by mechanical disruption and filtration through sterile gauze to obtain the primary extract. EV were purified via differential centrifugation: sequential centrifugation at 1000×g for 10 min, 3000×g for 20 min, and 10,000×g for 40 min to remove dead cells, cell fragments, and large vesicles. The final supernatant was subjected to ultracentrifugation at 150,000×g (Beckman Optima XPN-100, SW32 Ti rotor) for 70 min to pellet EV, which were resuspended in sterile 0.9 % NaCl. EV were mixed with PDRN (Sigma-Aldrich) at a mass ratio of 1: 4, followed by sonication (Bioruptor Plus, Diagenode; low power, 30 s pulse/30 s interval, 6 cycles) and incubation at 37 °C with constant shaking (50 rpm, 2 h) to facilitate membrane fusion. Free PDRN was removed by ultracentrifugation (150,000×g, 70 min), and the resulting PDRN-EV were resuspended in 0.9 % NaCl. Protein concentration was determined using the BCA assay.

2.3. Characterization of PDRN-EV and stability in gastrointestinal fluids

Particle size distribution was analyzed via nanoparticle tracking analysis (NTA), and zeta potential was measured by dynamic light scattering (DLS). Morphology was observed using transmission electron microscopy (TEM). To evaluate oral delivery stability, PDRN-EV (1 mg/mL) were dispersed in PBS (pH 7.4), simulated gastric fluid (SGF: 0.3 % pepsin, pH 1.2), simulated intestinal fluid (SIF: 1 % pancreatin, pH 6.8), or simulated colonic fluid (SCF: pH 7.4) and incubated at 37 °C with shaking for 2 h. Post-incubation, particle size changes were assessed by DLS, and agarose gel electrophoresis was performed to analyze PDRN protection against enzymatic digestion. Lipidomics analysis was conducted using methyl tert-butyl ether (MTBE) extraction, separated via a Thermo Accucore C30 column, and detected using a Q Exactive HF-X mass spectrometer. LipidSearch software was employed for lipid identification and quantification.

2.4. In vitro activity Evaluation of PDRN-EV

Raw 264.7 macrophages were seeded in 24-well plates at 3 × 10⁴ cells/well, stimulated with LPS (1 μg/mL) for 24 h [[36], [37], [38], [39]], and then co-cultured with EV (98.8 μg/mL), PDRN (50 μg/mL), or PDRN-EV (PDRN = 50 μg/mL,EV = 98.8 μg/mL). Intracellular reactive oxygen species (ROS) were detected using the DCFH-DA probe (10 μM), with fluorescence intensity quantified by flow cytometry (CytoFLEX) and spatial distribution visualized via inverted fluorescence microscopy (OLYMPUS IX73). mRNA expression of inflammatory cytokines (Tnfα, Il6, Il1β, Nos2 for M1 polarization; Il10 for M2 polarization) was assessed by qPCR. M1 polarization inhibition and M2 polarization promotion were evaluated via CD86 and CD206 immunofluorescence staining, respectively. LPS-stimulated cells served as the positive control, while untreated Raw 264.7 macrophages were the negative control.

2.5. Targeting specificity and uptake mechanism of PDRN-EV

DiO and DiR was dissolved in DMSO to prepare a 10 mM stock solution. PDRN-EV or EV were diluted to 1 mg/mL in an appropriate buffer, followed by the addition of DiO and DiR dye solution to achieve a final concentration of 10 μM. The mixture was incubated at 37 °C for 5 min under light-protected conditions. Subsequently, the labeled nanoparticles were isolated by centrifugation at 170,000×g for 2 h at 37 °C. The supernatant was discarded, and the pellet was resuspended in a sufficient volume of PBS to yield DiO and DiR labeled PDRN-EV or EV.

To investigate the mechanisms of uptake, PDRN-EV (DiO-labeled EV, PI-labeled PDRN) were co-incubated with RAW264.7 macrophages that were either unstimulated or pre-treated with LPS (1 μg/mL). Competitive inhibition was assessed using a galactose co-treatment group, while the normal uptake group received no galactose. Internalization mediated by galactose receptors was analyzed via flow cytometry (CytoFLEX) and confocal laser scanning microscopy (CLSM, OLYMPUS FV3000). For in vivo targeting, 8-week-old male C57BL/6J mice with DSS-induced colitis received DiR-labeled or DiO-labled PDRN-EV by oral gavage. Intestinal distribution was tracked using an in vivo imaging system at specified time points. At 2, 4, 6, 8, 12h post-gavage, colon tissues were harvested, stained with F4/80 antibody and DAPI, and analyzed by CLSM to evaluate macrophage targeting efficiency.

2.6. In vivo animal studies

The animal experimental protocols were conducted according to the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health and were approved by the medical ethics committee of Northwestern Polytechnical University (No: 202501153). Eight-week-old male C57BL/6J mice were acclimated for 1 week and randomly divided into five or four groups (n = 5). Colitis was induced via adding libitum 2.5 % DSS administration. Each animal was administered 250 μL of EV (2.34 mg/mL), PDRN (1.28 mg/mL) [14], or PDRN-EV (PDRN = 1.28 mg/mL, EV = 2.34 mg/mL). Disease activity index (DAI) was calculated to judge the severity of symptoms based on the summation of body weight loss (0−4), stool consistency (0−4), and fecal bleeding (0−4). Mice were sacrificed on day 8, and colons and major organs were harvested for histopathological analysis.

2.6.1. HE staining and AB-PAS staining

The harvested colonic specimens were fixed with 4 % formalin solution for 48 h. After fixation, samples were immersed in turn with 70 %, 80 % and 90 % ethanol solutions for 30 min, respectively, and subsequently put into 95 % and 100 % ethanol solutions twice for 20 min each time. Then, the dehydrated samples were infiltrated with xylene, embedded in paraffin, and cut into sections (5 μm). Hematoxylin-Eosin (H & E) and alcian blue-periodic acid Schiff's (AB-PAS) staining were performed, and then sections were analyzed with a light microscope (OLYMPUS IX73).

2.6.2. ELISA

Accurately weigh the colon tissue and add the homogenization medium (RIPA lysis buffer containing 1 % protease inhibitor cocktail) at a tissue-to-medium ratio of 1:9. Subsequently, add grinding beads and set the homogenization program as follows: 60 Hz, operate for 60 s, pause for 10 s, repeat this cycle four times. The resulting homogenate was centrifuged at 12,000×g for 5 min, and the supernatant was carefully collected. This procedure was repeated once to ensure consistency. Finally, the levels of inflammatory factors (TNFα, IL-10, IL-6, and IL-1β) in the supernatant were quantitatively analyzed using the QuantiCyto® mouse ELISA kit (NeoBioscience).

2.6.3. qRT-PCR assay

After the cell or colon tissue washing, the total RNA in different groups was extracted using a total RNA extract kit (Accurate Biology), followed by reverse transcription of RNA into cDNA. The qRT-PCR was performed on a Bio-Rad CFX Manager system with a SYBR Premix EX Taq kit (Accurate Biology). The qRT-PCR reaction solution contained a total of 20 μL, including 2 μL of cDNA, 2 μL of primer (1 μM), 10 μL of SYBR green dye (2 × ), and 6 μL of Nuclease-Free H₂O. The procedure of two-step cycling amplification was: 95 °C for 30 s, followed by 39 cycles of 95 °C for 5 s and 60 °C for 30 s. The Ct values of the target gene were obtained based on the PCR curve. The relative expression levels were calculated with a value of 2^−ΔΔCt. The relative expression levels of different genes were normalized to the housekeeping gene GADPH. The primers used in this study were listed in Table S1.

2.6.4. Immunofluorescence staining

Immunofluorescence staining was performed to evaluate the expression of macrophage surface markers (CD86, CD206, and F4/80), and tight junction-associated proteins (ZO-1 and occludin-1). Briefly, the pretreatment process of tissue sections was similar to that mentioned earlier. Treated sections were incubated with the primary antibodies, including anti-CD86 (YM80223), antiCD206 (ab64693), anti-F4/80 (ab6640), anti-ZO-1 (ab221547), anti-occludin-1 (ab216327), PCNA (ab92552) and HIF-1α (sc-13515) at 4 °C overnight. Next, samples were incubated with the corresponding Alexa 488-labeled or Cy3-labeled fluorescent secondary antibody for 1 h, followed by incubation with DAPI for 5 min. Stained slides were visualized by using a confocal laser scanning microscope (CLSM, OLYMPUS FV3000).

2.6.5. Gut microbiota analysis

Fecal samples were collected aseptically, snap-frozen at −80 °C, and subjected to 16S rRNA sequencing. DNA was extracted, quality-checked via agarose gel electrophoresis and an Agilent 5400 automated capillary electrophoresis system, and amplified using primers targeting the V4 region (515F: 5′-GTGCCAGCMGCCGCGGTAA-3′; 806R: 5′-GGACTACHVGGGTWTCTAAT-3′). PCR products were pooled, purified, and sequenced on an illumina platform.

2.6.6. Transcriptomics analysis

Total RNA was extracted from colon tissues using TRIzol (Invitrogen), with RNA integrity verified via an Agilent 2200 Bioanalyzer (RIN >7.0). Clean reads were aligned to the mouse genome (mm10) using STAR (v2.7.9a). Gene expression levels (RPKM-normalized) were quantified using HTSeq (v0.13.5). Differentially expressed genes (DEGs) were identified via DESeq2 (fold change >2 or <0.5; P < 0.05, FDR <0.05). Gene Ontology (GO) terms were annotated using NCBI, UniProt, and GO databases, with significant associations determined by Fisher's exact test (P < 0.05).

2.7. Statistical analysis

Data were presented as mean ± SEM of a minimum of three samples (n ≥ 3). The number of repeated experiments or the number of experimental samples is indicated in the figure captions. All of the statistical analysis was carried out with Graphpad Prism 9 software. Specifically, the two-sample t-test and one-way analysis of variance (ANOVA) with a Tukey's honestly significant difference (HSD) multiple comparison post-hoc test were used to determine the statistical significance among different groups. Values of ∗p < 0.05, ∗∗p < 0.01 and ∗∗∗p < 0.001 were considered statistically significant.

3. Results and discussion

3.1. Preparation and characterization of PDRN-EV nanoplatform

EV was obtained from fresh tea leaves by mechanical homogenization and differential centrifugation. To prepare crude extracts, tea leaves were mechanically homogenized in phosphate buffer (PBS) (pH 7.4). To achieve monodisperse EV, sequential centrifugation was performed under optimized conditions: 1000×g for 10 min (dead cells removal), 3000×g for 20 min (cell fragments removal), and 10,000×g for 40 min (large vesicles removal), followed by ultracentrifugation at 150,000×g for 70 min. A low-intensity ultrasound-assisted membrane fusion strategy (40 W, 30 s pulse/30 s interval, 6 cycles) in conjunction with dynamic incubation (37 °C, 50 rpm, 2 h) was used to insert PDRN into EV at a 4: 1 mass ratio (PDRN: EV) for drug loading (Fig. 1a). As demonstrated by an encapsulation efficiency (EE) of 12.3 % and loading capacity (LC) of 35 % ascertained by ultrafiltration centrifugation, this method effectively loaded PDRN while preserving the structural integrity of the vesicle (Fig. 1b). Dynamic light scattering (DLS) revealed a minimal size variation between EV (140.69 ± 3.2 nm) and PDRN-EV (151.47 ± 4.1 nm) (Fig. 1c), comparable surface charges between EV (−10.44 ± 0.07 mV) and PDRN-EV (−11.02 ± 1.49 mV) (Fig. 1h). Nanoparticle tracking analysis (NTA) and Transmission electron microscope (TEM) images confirmed that the size distributions of both formulations were in the range of 140–150 nm (Fig. 1d & e), meeting the homogeneity requirements for oral nanocarrier.

Fig. 1.

(a) Schematic illustration of the extraction of tea-derived extracellular vesicles and preparation of PDRN-EV. (b) Encapsulation efficiency and loading capacity of PDRN in PDRN-EV. (c) Hydrodynamic size of EV and PDRN-EV determined by dynamic light scattering (DLS). (d & e) Size distribution (NTA) and TEM images of EV and PDRN-EV. Scale bar: 100 nm. (f) Stability test of PDRN and PDRN-EV after incubation in PBS, SGF, and SIF. (g) TEM images of PDRN-EV after incubation in PBS, SGF, SIF and SCF solution for 2 h, respectively. Scale bar: 100 nm. (h) Zeta potential of EV and PDRN-EV. (i) Changes of size and PDI of EV and PDRN-EV after incubation in PBS, SGF, SIF and SCF solution for 2 h, respectively. (J) Lipid composition analysis prior to (EV) and following ultrasonic drug loading (PDRN-EV).

EV-mediated preservation of PDRN was shown by simulations of digestion assays. EV-encapsulated PDRN retained its structural integrity in the simulated gastric fluid (SGF) and the simulated intestinal fluid (SIF) (clear electrophoretic bands), while free PDRN was completely degraded in the SGF (no distinguishable bands in gel electrophoresis) (Fig. 1f). These results demonstrated that tea-derived EV had an intrinsic capability to protect the encapsulated PDRN from the harsh environment of the upper gastrointestinal tract. Furthermore, both EV and PDRN-EV exhibited excellent colloidal stability under various physiological conditions. TEM images revealed that the bilayer structure of both EV and PDRN-EV remained intact without noticeable membrane disruption in environments include PBS, simulated gastric fluid (pH 1.2), simulated intestinal fluid (pH 6.8), and simulated colonic fluid (pH 7.4) (Fig. 1g). Throughout the digestion process, the variation in particle size was within 10 %, and the polydispersity index (PDI) remained below 0.23 (Fig. 1i). After oral gavage, EV and PDRN-EV sequentially pass through the stomach and small intestine before finally reaching the colon. In this process, the preservation of their intact nanostructure facilitates more efficient cellular uptake.

Subsequently, the major lipid components of tea-derived EV were quantitative evaluated (Fig. 1j). The result of lipidomics revealed that the major components included diacylglycerols (DGs, 12–13 %), phosphatidylcholines (PCs, 21.3–21.7 %), triacylglycerols (TGs, 22.6–23 %), and phosphatidylethanolamines (PEs, 7.9–8.2 %). As a consequence, the potential structure-function relationships were established. First, the structural robustness: high PCs and TGs content facilitated stable bilayer formation, enhancing mechanical strength and drug encapsulation. Second, the bioactivity modulation: PEs promoted intestinal epithelial proliferation and tight junction reinforcement [40,41]. Third, the targeting specificity: galactolipids (MGDG/DGDG) enabled the lectin-mediated targeting phagocytosis of macrophages via galactose receptor recognition [33,34,42]. Interestingly, no discernible changes in lipid composition were brought about by PDRN loading, indicating the stability of EV throughout the cargo integration.

3.2. In vitro antioxidant properties of PDRN-EV

Studies have shown that tea-derived EV are rich in various polyphenolic compounds with antioxidant activity, including gallic acid (GA), caffeine (CAF), epigallocatechin gallate (EGCG), epicatechin gallate (ECG), and epicatechin (EC), as well as flavonoids such as myricetin-3-O-galactoside, myricetin-3-O-rhamnoside, quercetin-3-O-glucoside, and quercetin [43,44]. To verify whether the PDRN-EV retains its antioxidant properties after ultrasound drug loading treatment, this study used ABTS·⁺ and DPPH· radical scavenging assays to evaluate its antioxidant capacity in vitro. In the ABTS·⁺ assay system, the blue-green ABTS·⁺ radical (λ max = 734 nm) generated by potassium persulfate (K₂S₂O₈) oxidation was used to assess the antioxidant effect of the samples [45]. The experimental results revealed that when the concentration gradient of PDRN-EV increased, the solution color gradually shifted from blue-green to colorless, resulting in a considerable drop in absorbance (Fig. 2a). Higher concentrations of PDRN-EV (1 mg/mL) could neutralize approximately 90 % of ABTS·⁺ radicals (Fig. 2c), displaying high antioxidant ability. To support this result, this study also performed DPPH· radical scavenging studies. In an anhydrous ethanol system, the typical absorption peak of DPPH· radicals (λ max = 517 nm) showed a concentration-dependent drop with the addition of PDRN-EV (Fig. 2b & d), which was consistent with the results of ABTS·⁺ radicals scavenging assay. Interestingly, we found that the antioxidant capacity of PDRN-EV does come from EV. Fig. S1a and S1b results demonstrated that the ABTS·⁺ and DPPH· scavenging capacity of PDRN-EV was comparable to that of EV, with no significant difference between them. In contrast, the antioxidant capacity of PDRN was significantly lower than both EV and PDRN-EV. This suggests that the antioxidant activity of PDRN-EV may be primarily attributed to the EV component.

Fig. 2.

(a & b) UV–vis spectra of ABTS·+ radicals and DPPH· radicals after incubation with different concentration of PDRN-EV. (c & d) Quantitative analysis of ABTS·+ scavenging ability and DPPH· scavenging ability (n = 3). (e) Representative fluorescent images of the intracellular ROS in macrophages after different treatments. Scale bar: 50 μm. (f) Flow cytometry analysis of the intracellular ROS in macrophages after different treatments. (g) Representative images of JC-1 staining showing the mitochondrial membrane potential of macrophages after different treatments. Scale bar: 50 μm. (h–j) The corresponding quantifications of JC-1 aggregates, JC-1 monomers, and JC-1 aggregates/JC-1 monomers, respectively. Representative images were taken from three independent samples (n = 3). (k) Flow cytometry analysis of JC-1 in macrophages after different treatments. Data are presented as mean ± SEM. ∗p < 0.05, ∗∗p < 0.01 and ∗∗∗p < 0.001 denote the statistical significance, calculated by the one-way ANOVA with Tukey's post-hoc test.

This study used lipopolysaccharide (LPS)-induced Raw 264.7 macrophages to further evaluate the antioxidant effect of PDRN-EV at the cellular level [46]. The LPS-stimulated macrophages were co-cultured with different formulations (PDRN = 50 μg/mL) for 24 h. According to the DCFH-DA staining, fluorescence images (Fig. 2e), semiquantitative ROS level (Fig. S1c) and flow cytometry analysis (Fig. 2f) showed that compared to the positive control, the levels of reactive oxygen species (ROS) in LPS-stimulated macrophages pretreated with EV, PDRN, and PDRN-EV were significantly reduced, with PDRN-EV exhibiting the strongest ROS scavenging activity. Cellular SOD activity, a critical marker of antioxidant capacity, was measured after different treatment, revealing that PDRN-EV possesses enhanced antioxidant activity (Fig. S1d). Considering that excessive ROS-induced oxidative stress can lead to mitochondrial dysfunction, this study further employed the JC-1 fluorescent probe to detect the changes in mitochondrial membrane potential (MMP). The results of fluorescence images, semi-quantitative statistics (Fig. 2g–j) and flow cytometry analysis (Fig. 2k) showed that LPS stimulation induced the conversion of JC-1 aggregates (indicative of high mitochondrial membrane potential, MMP) to monomers (reflecting low MMP), demonstrating LPS-induced oxidative stress and mitochondrial dysfunction in macrophages [47,48]. While both PDRN and EV partially reversed these pathological alterations, their efficacy remained limited. Notably, PDRN-EV treatment exhibited significantly enhanced restorative effects on MMP. Quantitative analysis revealed distinct MMP profiles across groups: Control: 53.0 % high-MMP cells (Q2), 44.6 % low-MMP cells (Q3); LPS: Severe MMP collapse (3.02 % Q2, 96.1 % Q3); EV: Partial restoration (21.1 % Q2, 78.2 % Q3); PDRN: Moderate restoration (15.9 % Q2, 83.0 % Q3); PDRN-EV: Near-normalization (45.4 % Q2, 52.3 % Q3). These results collectively demonstrate the superior antioxidant capacity of PDRN-EV, effectively counteracting LPS-induced mitochondrial damage and restoring membrane potential to near-physiological levels. The significantly higher proportion of high-MMP cells (Q2) in the PDRN-EV group (45.4 % vs ≤ 21.1 % in other treatment groups) underscores its enhanced bioactivity.

3.3. In vitro anti-inflammatory properties of PDRN-EV

Previous studies have demonstrated that LPS can induce polarization of M0 macrophages into the pro-inflammatory M1 phenotype [49]. To systematically evaluate the regulatory effect of PDRN-EV on macrophage polarization, this study used immunofluorescence double-labeling method to detect the expression levels of M1 marker (CD86) and M2 marker (CD206). According to fluorescence images (Fig. 3a & b), the number of CD86⁺ macrophages was significantly enhanced after LPS treatment, while CD206⁺ macrophages were almost absent. Interestingly, the phenomenon was alleviated by EV and PDRN treatment, indicating their potential immunomodulatory effects. Intuitively, PDRN-EV treatment predominantly increased the number of CD206⁺ macrophages, implying that the synergistic effect of EV and PDRN promoted the macrophage M2 polarization. Semi-quantitative analysis indicated that PDRN-EV reduced the proportion of M1-type macrophages by 61.44 % (Fig. 3c) and increased the polarization efficiency of M2-type macrophages by 1.7 times (Fig. 3d) compared to the LPS group.

Fig. 3.

(a & b) The immunofluorescence staining of CD86 (M1 marker), CD206 (M2 marker), and F4/80 (macrophage marker), respectively. Scale bar: 50 μm. (c & d) Semiquantitative analysis of CD86 and CD206 positive area/F4/80 positive area. Representative images were taken from three independent samples (n = 3). (e) The level of NO released of macrophages. (f–j) The mRNA expression levels of Nos2, Tnfα, Il6, Il1β, Il10 determined by qPCR assay (n = 3). Data are presented as mean ± SEM. ∗p < 0.05, ∗∗p < 0.01 and ∗∗∗p < 0.001 denote the statistical significance, calculated by the one-way ANOVA with Tukey's post-hoc test.

Furthermore, Griess method was employed to measure nitric oxide (NO) levels in cell supernatants, showing that LPS stimulation significantly promoted the production of pro-inflammatory mediator NO (Fig. 3e), while single EV or PDRN treatment effectively reversed this pathological process [50]. Meanwhile, NO level was predominantly reduced after PDRN-EV treatment. Given the strong correlation between macrophage polarization states and the progression of IBD, where M1 macrophages exacerbate intestinal mucosal damage by secreting pro-inflammatory factors such as NOS2, TNFα, IL-6, and IL-1β, while M2 macrophages promote tissue regeneration by releasing anti-inflammatory factors (IL-10), this study further detected the expression profile of inflammation-related cytokines in macrophages through qRT-PCR. The results showed that PDRN-EV administration dramatically decreased the mRNA expression of pro-inflammatory factors generated by LPS, including Nos2, Tnfα, Il6, and Ilβ mRNA levels (Fig. 3f–i). At the same time, the anti-inflammatory factor Il10 expression increased compared to the control group (Fig. 3j). The above data indicated that PDRN-EV could achieve anti-inflammatory effects by bidirectionally regulating macrophage polarization and the associated cytokine network.

3.4. Galectin-mediated PDRN-EV uptake in macrophages

Macrophage galactose-type lectin, a C-type galactose receptor, is highly expressed on the surface of activated macrophage subpopulations under inflammatory conditions [51,52]. Concurrently, PDRN-EV surface contains galactolipids (MGDG/DGDG), which enable lectin-mediated targeting phagocytosis by macrophages via galactose receptor recognition. Therefore, we investigated the uptake capacity of macrophages for PDRN-EV both in vivo and in vitro. For the in vitro experiments, the following groups were designed. As a control, macrophages without LPS pretreatment were co-cultured with PDRN-EV (DiO-labeled EV, PI-labeled PDRN) for 16 h. In addition, macrophages pretreated with LPS (1 μg/mL) for 24 h were co-cultured with PDRN, PDRN-EV, and PDRN-EV + galactose (Gal) for 16 h, respectively. Through quantitative analysis by flow cytometry and qualitative observation using confocal laser scanning microscopy (CLSM), it was found that LPS-activated macrophages had a significantly higher endocytosis efficiency of PDRN-EV compared to resting macrophages, indicating that pro-inflammatory M1 polarization was beneficial for its phagocytic activity. Furthermore, the fluorescence intensity of the PDRN-EV group was significantly higher than that of the single PDRN group, further confirming that the tea-derived EV carrier facilitated the phagocytosis of M1 macrophages, thereby improving the intracellular delivery of active components. After adding free Gal, the cellular uptake efficiency of PDRN-EV was lower than that of the group without Gal, clearly indicating that the galectin-dependent endocytic pathway is the core mechanism for the efficient uptake of PDRN-EV by M1 macrophages (Fig. 4a & S2a).

Fig. 4.

(a) Flow cytometry analysis and representative fluorescence images of cell internalization profiles of PI-labeled PDRN and DiO-labeled EV in macrophages after different treatments. Scale bar: 50 μm. (b) Representative fluorescence images (6 h) of colonic tissues from healthy mice (Control) or DSS-colitis mice (DSS) after oral gavage of DiO-marked PDRN-EV, which reveals F4/80 macrophages (red) internalizing PDRN-EV (green). White arrows highlight colocalized signals (DiO⁺/F4/80⁺). Scale bar: 100 μm. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

Encouraged by the above-mentioned results, we further evaluated the distribution characteristics of DiR-labeled PDRN-EV in normal mice and DSS-colitis mice through in vivo fluorescence imaging technology. The experiment was divided into three groups: Control + DiR-PDRN-EV group, DSS + DiR-PDRN-EV group, and DSS + free DiR group. The results showed that after the post-oral gavage for 6 h, specific fluorescence enrichment of DiR-PDRN-EV could be detected in the colon region of the DSS-colitis mice, and this signal maintained significant intensity at 12 h, which was higher than that of in the normal mice at the same time point. In stark contrast, the fluorescence signal in the DSS + free DiR group almost disappeared by 12 h (Fig. S2b). In vivo organ imaging further confirmed that after the post-oral gavage for 12 h, the fluorescence intensity of DiR-PDRN-EV in the colon tissue of the DSS-colitis mice was not only significantly higher than that of the normal mice but also markedly higher than that of the free DiR group (Fig. S2c). The phenomenon was attributed to the fact that the negatively charged surface of PDRN-EV enabled it to interact with the positively charged inflamed area. It is worth noting that no significant fluorescence accumulation was detected in major organs such as the heart, liver, spleen, lungs, and kidneys, fully confirming that the PDRN-EV nanoplatform possessed extremely low off-target effects (Fig. S2c).

In addition, PDRN-EV exhibited enhanced retention in inflamed lesions, benefiting from the increased endocytic activity of macrophages within the pathological microenvironment. As revealed by Fig. 4b, the intestinal tissues of DSS-colitis mice presented a significant accumulation of DiO-PDRN-EV in the areas surrounding F4/80⁺ macrophages, while the intestinal tissues of normal mice only detected a small amount of green fluorescence signals. Overall, these evidences suggested that PDRN-EV facilitated the accumulation in inflamed lesions via an active targeting transport mechanism mediated by macrophages. To further validate this result, we expanded the in vivo cellular uptake study to include additional early (2 h, 4 h) and late (8 h, 12 h) time points, as presented in Figure S2d-S2f. Notably, the extent of PDRN-EV uptake by macrophages was consistently higher in the DSS group compared to the control group over extended time periods. This comprehensive temporal analysis systematically tracked the biodistribution of PDRN-EV and provided unequivocal evidence that its internalization is mediated through macrophage-specific uptake mechanisms.

3.5. Therapeutic efficacy of PDRN-EV in DSS-induced colitis

To systematically evaluate the in vivo therapeutic efficacy of PDRN-EV against IBD, an acute colitis model was established by orally administering 2.5 % dextran sulfate sodium (DSS) to C57BL/6 mice for 7 consecutive days (Fig. 5a). Five experimental groups were designed: healthy mice (Control), DSS model (DSS + 0.9 % NaCl), EV carrier control (DSS + EV), PDRN monotherapy (DSS + PDRN), and PDRN-EV combination therapy (DSS + PDRN-EV). Each group was orally given 250 μL of the respective formulations every day. The body weight dynamics, disease activity index (DAI), and colonic pathological characteristics were continuously monitored throughout the trial. By the seventh day, the DSS model group had significantly reduced their body weight (18.60 ± 2.36 % from baseline), but the PDRN-EV treatment group had only reduced their body weight by 3.4 ± 1.12 %, outweighing both the EV carrier group (13.8 ± 1.93 %) and the PDRN monotherapy group (8.8 ± 1.88 %) (Fig. 5b). DAI scoring (integrating weight loss, hematochezia severity, and stool consistency) revealed that the PDRN-EV group achieved an 82.1 % reduction compared to the DSS model group (1.00 ± 0.45 vs 5.60 ± 0.40), and significantly lower reductions observed in the EV group and PDRN group (Fig. 5c). Anatomical analysis demonstrated that the colon length of DSS-colitis mice was shortened by 18.1 % compared to the healthy mice (6.66 ± 0.13 cm vs 8.13 ± 0.19 cm), and the PDRN-EV group maintained the colon length at 7.82 ± 0.23 cm, confirming its efficacy in suppressing inflammatory edema (Fig. 5d & S3).

Fig. 5.

(a) The timeline of the model fabrication of DSS-induced colitis and experimental procedures. DSS-colitis mice were orally administered with 0.9 % NaCl, EV, PDRN or PDRN-EV, respectively. (b) Daily bodyweight changes in each group for 8 days (n = 5). (c) DAI score in each group for 8 days (n = 5). (d) The representative colonic images of mice after different treatments (n = 5). (e) Representative images of HE staining of colonic tissues. Scale bar: 200 μm. (f) Representative images of AB/PAS staining of colonic tissues. Scale bar: 200 μm. (g–j) The immunofluorescence staining and semiquantitative analysis of tight junction indicators including ZO-1 and Occludin. Scale bar: 100 μm. Representative images were taken from three independent samples (n = 3). (k & l) The mRNA expression levels of Zo1 and Occludin determined by qPCR assay (n = 3). Data are presented as mean ± SEM. ∗p < 0.05, ∗∗p < 0.01 and ∗∗∗p < 0.001 denote the statistical significance, calculated by the one-way ANOVA with Tukey's post-hoc test.

Hematoxylin-eosin (HE) staining of colon tissues showed severe mucosal damage in the DSS model group, characterized by crypt atrophy, epithelial denudation, and neutrophil infiltration. The EV carrier group and the PDRN monotherapy group did not effectively alleviate these adverse effects. In contrast, the PDRN-EV treatment group preserved intact crypt architecture with markedly reduced inflammatory cell infiltration (Fig. 5e). In addition, Alcian Blue/Periodic Acid-Schiff (AB/PAS) staining demonstrated abundant goblet cells and mucin secretion in healthy controls, whereas the DSS model group showed a drastic depletion of goblet cells. Remarkably, PDRN-EV intervention boosted mucin secretion area and restored goblet cell density to nearly physiological levels (Fig. 5f), indicating its mucosal protective effects via mucus barrier reconstitution. According to immunofluorescence images and semi-quantitative analysis (Fig. 5g–j), the expression of tight junction proteins, including ZO-1 and Occludin, was significantly downregulated in the colonic tissues of the DSS model group, decreasing by 91.2 % and 79.0 % compared with the healthy control, respectively. Consistent with the results of HE staining, the EV carrier group and the PDRN monotherapy group did not restore intestinal barrier function. However, following PDRN-EV therapy, the expression levels of ZO-1 and occludin were significantly restored to nearly normal values, closely approximating those observed in healthy conditions. Similarly, subsequent qPCR analysis revealed the changes at gene level, with the mRNA expression patterns of Zo1 and Occludin congruent with their protein profiles (Fig. 5k & l). Taken together, these results fully demonstrated the importance of using tea-derived EV to deliver PDRN to colonic inflamed lesions, and PDRN-EV successfully restored the damaged intestinal mucosal barrier by rebuilding the paracellular tight junction structure.

3.6. Anti-inflammatory and antioxidant properties of PDRN-EV in DSS-induced colitis

Enzyme linked immunosorbent assay (ELISA) analysis demonstrated that PDRN-EV significantly suppressed the release of pro-inflammatory cytokine. TNFα, IL-1β, and IL-6 levels in colon tissues decreased by 37.5 % (1115.96 ± 82.80 pg/mL vs 1784.78 ± 168.78 pg/mL), 82.1 % (83.69 ± 17.96 pg/mL vs 466.69 ± 5.60 pg/mL), and 67.8 % (95.80 ± 32.39 pg/mL vs 297.63 ± 0.87 pg/mL) respectively compared to the DSS model group, while the expression of anti-inflammatory IL-10 increased by 37.3-fold (285.77 ± 53.68 pg/mL vs 7.47 ± 0.73 pg/mL) (Fig. 6a). In addition, qPCR results corroborated the ELISA results, showing downregulated Tnfα, Il1β, and Il6 mRNA levels alongside upregulated Il10 expression (Fig. 6b). In order to further investigate the antioxidative stress effect of PDRN-EV, we further detected the contents of several biological indicators (MDA and SOD) in colon tissues [53,54]. Malondialdehyde (MDA), a product of polyunsaturated fatty acid peroxidation, would be excessively generated when the free radicals increase. Superoxide dismutase (SOD) is an enzyme that catalyzes the conversion of superoxide radicals into oxygen and hydrogen peroxide, which plays an important role in reducing oxidative damage. Anti-oxidative stress related analysis revealed that PDRN-EV enhanced SOD activity to three times than that of the DSS model group (48.7 ± 4.2 U/mg vs 16.8 ± 2.1 U/mg) (Fig. 6c) and reduced MDA content by 70.56 % (0.63 ± 0.08 μmol/mg vs 2.14 ± 0.54 μmol/mg) (Fig. 6d).

Fig. 6.

(a) The levels of TNFα, IL-6, IL-1β and IL-10 in the colon tissue determined by ELISA assay (n = 3). (b) Relative mRNA expression of Tnfα, Il6, Il1β, Il10 in the colon tissue determined by qPCR assay (n = 3). (c–d) The levels of oxidation indicators including SOD and MDA (n = 3). (e–f) The immunofluorescence staining of CD86 (M1 marker), CD206 (M2 marker), and F4/80 (macrophage marker), respectively. Scale bar: 100 μm. (g–h) Semi-quantitative analysis of CD86 and CD206 positive area/F4/80 positive area. Representative images were taken from three independent samples. Data are presented as mean ± SEM. ∗p < 0.05, ∗∗p < 0.01 and ∗∗∗p < 0.001 denote the statistical significance, calculated by the one-way ANOVA with Tukey's post-hoc test.

Macrophages play an important role in the occurrence and development of IBD [55]. During the process of intestinal inflammation, the increased monocytes are recruited to the inflammatory site and mainly differentiate into pro-inflammatory M1 macrophages, generating pro-inflammatory cytokines and aggravating epithelial cell damage. In contrast, M2-type macrophages mainly produce anti-inflammatory and pro-generative molecules, which inhibit the expression of pro-inflammatory mediators and stimulate the renewal of epithelial stem cells, thus maintaining the integrity of the epithelial barrier [56]. Therefore, timely intervention to restore immune homeostasis and driving macrophage M2 type polarization is extremely important for alleviating the inflammatory response. Consequently, given that macrophages exerted an enhanced endocytosis of PDRN-EV in the pathological microenvironment, we further tested whether PDRN-EV may restore immunological homeostasis and produce anti-inflammatory effects by driving macrophage M2 polarization. Double immunofluorescence labeling was performed on M1 macrophages with F4/80/CD86 and M2 macrophages with F4/80/CD206, respectively. The damaged colonic epithelium of DSS-colitis mice infiltrated a larger number of CD86⁺ macrophages (M1), while less of CD206⁺ macrophages (M2) were presented in the inflamed lesions. Intuitively, after PDRN-EV treatment, the distribution of CD86⁺ macrophages showed an obvious decrease, whereas they presented a predominantly increase in the number of CD206⁺ macrophages (Fig. 6e & f). Semi-quantitative data revealed that it significantly reduced the proportion of CD86⁺ in F4/80⁺ macrophages, while the percentage of CD206⁺ in F4/80⁺ macrophages was higher than in the DSS-treated other control groups (Fig. 6g & h), indicating that PDRN-EV treatment facilitated macrophage polarization from M1 (pro-inflammatory) to M2 (anti-inflammatory) type. In short, orally-administered PDRN-EV simultaneously relieved oxidative stress and restored the immunological homeostasis of macrophages in DSS-colitis mice, thus reversing the adverse inflammatory response and mucosal damage into advantageous pro-regenerative microenvironment.

3.7. Therapeutic effect of PDRN-EV on colitis is superior to physical mixing

To better understand how ultrasound-mediated loading improves PDRN stability and therapeutic efficacy, a second cohort of animal experiments compared the therapeutic effects of physical mixing (PDRN + EV) and ultrasound-loaded (PDRN-EV) formulation (Fig. S4a). Four experimental groups were established: healthy mice (Control), DSS model (DSS + 0.9 % NaCl), physical mixture group (DSS + PDRN + EV), and ultrasound-loaded group (DSS + PDRN-EV). DSS-colitis mice showed a 16.12 ± 2.82 % drop in body weight, while the PDRN-EV treatment showed only an 8.06 ± 0.57 % decline, greatly surpassing the physical combination group (10.83 ± 1.48 %) (Fig. S4b). The PDRN-EV treatment displayed a 54.8 % reduction in DAI scoring compared to the DSS model (2.80 ± 0.37 vs 6.20 ± 0.58), and significantly lower than reductions observed in the physical mixture group (Fig. S4c). Anatomical examination revealed that the colon length of PDRN-EV treatment (7.52 ± 0.09 cm) increased by 16.8 % compared to the physical mixture group (6.44 ± 0.11 cm), nearing the normal level (7.82 ± 0.11 cm) (Fig. S4d and S4e). In addition, the PDRN-EV treatment showed a significant decrease in colonic IL-1β level compared to that of PDRN + EV group (78.01 ± 3.99 vs 109.45 ± 4.83 pg/mL, 28.7 % decrease) (Fig. S4f). Although there was no significant difference, the PDRN-EV treatment inhibited the secretion of TNFα and IL-6, while promoting the expression of IL-10. According to these results, although PDRN + EV treatment alleviated the DSS-induced colitis symptoms to some extent, its therapeutic effect was much lower than that of PDRN-EV group, which fully demonstrated the superiority of employing tea-derived EV to deliver PDRN.

HE staining of colon tissues revealed ongoing focal crypt atrophy in the physical combination group, but the PDRN-EV treatment showed crypt architectural restoration similar to healthy controls (Fig. S4g). AB/PAS staining showed that the density of goblet cells and mucin-secreting regions were higher in PDRN-EV treatment than that of in PDRN + EV group (Fig. S4h). Furthermore, immunofluorescence images and semi-quantitative data confirmed that PDRN + EV treatment facilitated the restoration of tight junction in DSS-colitiss mice, which was reflected in the increased protein expression of ZO-1 and occludin. In contrast, the expression levels of ZO-1 and occludin in the PDRN-EV group reached 70.4 % and 75.9 % of those in healthy control, respectively, outperforming the physical mixture group, which reached 52.3 % and 51.6 % of those in healthy control, respectively (Fig. S4i–S4l). These results suggested that compared with simple physical mixing, tea-derived EV could efficiently deliver PDRN to the inflamed lesions, thus alleviating inflammatory response and restoring intestinal barrier function in a synergistic manner.

3.8. The regulatory effects of PDRN-EV on the intestinal microbiota structure

Mounting evidence in recent years has elucidated the crucial roles of gut microbiota in concert with bile-derived organic compounds, polysaccharides, and other bioactive molecules in maintaining gastrointestinal homeostasis [57]. Notably, orally administered pharmaceuticals can substantially alter the structural composition of intestinal microbial communities through direct modulation of the physiological activities of specific bacterial strains and proliferative states [58,59]. Particularly, EV have been reported to influence the synthesis of short-chain fatty acids and other metabolites essential for intestinal health by modifying microbial metabolic profiles [33]. Building upon these findings, this study systematically evaluate the differential regulatory effects between PDRN + EV and PDRN-EV formulations on gut microbiota in DSS-colitis mice. Four experimental groups were established: healthy mice (Control), DSS model (DSS + 0.9 % NaCl), physical mixture group (DSS + PDRN + EV), and ultrasound-loaded group (DSS + PDRN-EV), aiming to 250 μL of different formulations were daily administered by oral gavage until the eighth day when fecal samples were collected for 16S rDNA sequencing.

A Venn diagram analysis found 157 core operational taxonomic units (OTUs) shared by groupings (Fig. 7a). The normal mice had the most microbial diversity (624 OTUs), whereas DSS administration drastically lowered OTU counts to 220. Notably, the PDRN + EV and PDRN-EV treatments recovered OTU numbers to 468 and 481, respectively. These indicate that both PDRN + EV and PDRN-EV can increase the microbial flora richness in DSS-induced colitis. Principal co-ordinates analysis (PcoA) revealed considerable microbiological structural differences between PDRN-EV treatment and DSS group (Fig. 7b). PDRN-EV treatment outperformed DSS control in terms of α-diversity analysis (Shannon index, Chao 1, observed features, and pielou-e) (Fig. 7c–f). Phylum-level study (Fig. 7g) indicated classic dysbiosis in the DSS group, with lower abundance of Firmicutes (Fig. 7h) and higher level of Proteobacteria (Fig. 7i) than that of normal group. In gut microbiota analysis, the Firmicutes/Bacteroidota ratio serves as a critical indicator for colitis models. Our results demonstrate that DSS induction markedly elevates this ratio, whereas therapeutic intervention, particularly with PDRN-EV, effectively restores it to normal levels, highlighting the superior efficacy of PDRN-EV treatment. (Fig. 7j). Given that Firmicutes contains butyrate-producing species with anti-inflammatory capabilities, and Proteobacteria overgrowth is associated with intestinal barrier dysfunction [60], these findings suggested that PDRN-EV exerted therapeutic benefits via core phylum modulation. Family-level analysis (Fig. 7k) revealed that DSS significantly induced depletion of anti-inflammatory bacterial families, including Muribaculaceae (79.2 % reduction) (Fig. 7l) and Lachnospiraceae (68.4 % reduction) (Fig. 7m). Muribaculaceae and Lachnospiraceae are known for their ability to reduce intestinal inflammation, regulate energy metabolism, and preserve mucosal integrity by producing short-chain fatty acids (SCFAs) [61,62]. Therefore, orally-administered PDRN-EV showed the ability to reverse colitis-associated intestinal microflora disorders. To summarize, both PDRN-EV and PDRN + EV exhibit a robust capacity to modulate the intestinal microbiota, thereby alleviating the symptoms of IBD through this regulatory mechanism. Moreover, there is no substantial difference in their ability to modulate the intestinal microbiota. It can be reasonably inferred that the primary function of intestinal microbiota modulation is mediated by EV. Simultaneously, this finding suggests that the active components of EV are preserved during the process of ultrasound-assisted drug loading.

Fig. 7.

(a) Venn diagram of common and unique bacterial species of mice in each group. (b) Principal coordinates analysis (PCoA). (c) Shannon, (d) Chao 1, (e) Observed Features and (f) pielou-e indices of operational taxonomic unit levels of intestinal flora from different mouse groups. (g) The relative abundance histogram of gut microbiota at phylum level. (h–i) Relative abundance of representative taxa: Firmicutes and Proteobacteria at phylum level. (j) The ratio of Firmicutes to Bacteroidata. (k) The relative abundance histogram of gut microbiota at family level. (l–m) Relative abundance of representative taxa: Muribaculaceae and Lachnospiraceae at family level. Data are presented as mean ± SEM. ∗p < 0.05, ∗∗p < 0.01 and ∗∗∗p < 0.001 denote the statistical significance, calculated by the one-way ANOVA with Tukey's post-hoc test.

3.9. Therapeutic mechanism of PDRN-EV on colitis revealed by transcriptomics

To further study the mechanism of PDRN-EV in the treatment of IBD, colonic tissues treated with PDRN-EV were collected for transcriptome analysis. Transcriptomic sequencing analysis of colonic tissues revealed that 2774 genes were significantly upregulated and 1244 genes were downregulated in DSS-colitis mice compared to the normal group. The PDRN-EV intervention precisely modulated 2527 downregulated genes and 790 upregulated (Fig. 8a). Venn analysis revealed a significant overlap between genes that were downregulated by PDRN-EV and those that were elevated by DSS, and heatmap clustering further confirmed this regulation pattern (Fig. 8b). These findings demonstrate that PDRN-EV treatment effectively ameliorated DSS-induced colitis manifestations, restoring pathological parameters to levels comparable to the healthy control group.

Fig. 8.

(a) Volcano plot visualization of transcriptome gene expression in DSS group vs Control group, PDRN-EV group vs DSS group, and PDRN-EV group vs Control group. P-value <0.05 and fold change (FC) > 2. (b) A Venn diagram with two overlapping circles was constructed to represent the shared up-regulated genes in DSS group and down-regulated genes in PDRN-EV group. A heatmap was generated to display the expression levels of all genes in overlapping circles. (c) GO enrichment bar-plot analysis of DEGs in overlapping circles. (d) KEGG enrichment scatter plot analysis of DEGs in overlapping circles. (e) GSEA analysis for inflammatory bowel disease, cAMP signaling pathway, HIF-1α signaling pathway and DNA replication in PDRN-EV group vs DSS Group.

Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analyses were performed to systematically characterize the biological features of differentially expressed genes (DEGs). As shown in (Fig. 8c), GO analysis revealed that DEGs were predominantly enriched in critical biological processes (BP), including cytokine-mediated signaling transduction, hypoxia response, inflammatory response, regulation of cell proliferation, and macrophage activation. Molecular function (MF) analysis indicated that it was significantly correlated with cytokine receptor activity, growth factor binding, G protein-coupled adenosine receptor activity, lipid binding, and DNA-binding transcription factor activity, RNA polymerase ll-specific. Cellular component (CC) enrichment highlighted cell surface, integral component of plasma membrane, cell junction, class IB phosphatidylinositol 3-kinase complex and interleukin-6 receptor complex. KEGG pathway analysis identified DEGs primarily involved in two major signaling networks: inflammation-related pathways, including cAMP signaling, PI3K-Akt signaling, and TNFα signaling, and oxidative stress response pathways, such as HIF-1 signaling and oxidative phosphorylation (Fig. 8d). Gene Set Enrichment Analysis (GSEA) further confirmed that PDRN-EV intervention markedly downregulated inflammatory response (NES = −1.868, FDR <0.001), cAMP signaling (NES = −1.69, FDR <0.05), and HIF-1α signaling (NES = −1.632, FDR <0.01), while upregulating DNA replication-related gene sets (NES = 2.195, FDR = 0) (Fig. 8e).

Mechanistically, PDRN-EV achieves anti-inflammatory therapy through coordinated modulation of the cAMP signaling pathway and DNA salvage mechanisms. PDRN-EV activates the cAMP pathway in macrophages, and cAMP activation suppresses hypoxia-inducible factor (HIF-1α) expression via the PKA signaling pathway. In pathological microenvironments, excessive oxidative stress significantly enhances HIF-1α transcriptional activity in macrophages. This molecular event not only exacerbates hypoxic stress [63], but also drives macrophage polarization toward the pro-inflammatory M1 phenotype. Notably, PDRN-EV specifically inhibits HIF-1α activation at colonic inflammatory sites, thereby blocking M1 polarization and downregulating pro-inflammatory gene expression. Simultaneously, as a nucleotide precursor, PDRN provides essential substrates for cellular proliferation at damaged sites through the DNA salvage mechanism when the de novo synthesis pathway is compromised, markedly accelerating mucosal repair processes [64].

3.10. Verification of the potential mechanism of action of PDRN-EV

Finally, we established a multi-group animal model to validate the central role of A_2A receptors (A_2AR) signaling in PDRN-EV-mediated IBD treatment. Previous studies confirmed the pharmacological properties of PDRN as a prototypical adenosine A_2AR agonist, and transcriptomic analyses revealed that PDRN-EV exerted the therapeutic effects by modulating the downstream cAMP/HIF-1α signaling pathway of A_2AR in macrophage polarization and inflammatory responses. The experimental groups included: normal control (Control), DSS colitis model (DSS + 0.9 % NaCl), DSS + A_2AR inhibitor DMPX (DSS + DMPX) [[65], [66], [67], [68]], PDRN-EV monotherapy (DSS + PDRN-EV), and PDRN-EV + DMPX co-treatment (DSS + PDRN-EV + DMPX) (Fig. 9a). To understand how the A_2AR-cAMP-HIF-1α axis coordinates the functional remodeling of macrophages and affects intestinal repair, we conducted multidimensional assessments of colon tissues, including colitis phenotypes (disease activity index, histological damage scores), macrophage polarization (M1/M2 marker expression), inflammatory cytokine profiles (TNFα, IL-10, IL-6, IL-1β), and the expression of key pathway protein (HIF-1α).

Fig. 9.

(a) The timeline of the model fabrication of DSS-induced colitis and experimental procedures. DSS-colitis mice were administered with saline, DMPX, PDRN-EV or PDRN-EV + DMPX, respectively. (b) Daily bodyweight changes in each group for 8 days (n = 5). (c) DAI score in each group for 8 days (n = 5). (d & e) The representative colonic images and length statistics of mice after different treatments (n = 5). (f) The levels of TNFα, IL-6, IL-1β and IL-10 in the colon tissue determined by ELISA assay (n = 3). (g & h) The immunofluorescence staining of CD86 (M1 marker), CD206 (M2 marker), F4/80 (macrophage marker). Scale bar: 100 μm. (i & j) The immunofluorescence staining of PCNA and HIF-1α, respectively. Scale bar: 100 μm. Data are presented as mean ± SEM. ∗p < 0.05, ∗∗p < 0.01 and ∗∗∗p < 0.001 denote the statistical significance, calculated by the one-way ANOVA with Tukey's post-hoc test.

Specifically, compared to the DSS model group, PDRN-EV treatment markedly reduced weight loss and DAI score, whereas PDRN-EV + DMPX co-treatment only partially reversed the phenomenon (Fig. 9b & c). Concurrently, colon length in the PDRN-EV group (6.62 ± 0.20 cm) was significantly greater than in the DSS model group (5.30 ± 0.20 cm), but this protective effect was attenuated by DMPX (PDRN-EV + DMPX group: 5.86 ± 0.14 cm) (Fig. 9d & S5a). In addition, PDRN-EV significantly suppressed the expression of pro-inflammatory cytokines TNFα, IL-6, and IL-1β (reduced by 63.3 %, 56.31 %, 27.19 %, respectively), while notably upregulating the level of anti-inflammatory cytokine IL-10 (2-fold increase). However, DMPX partially counteracted the anti-inflammatory effects of PDRN-EV (Fig. 9e). These results fully demonstrated that the inhibition of A_2AR significantly weakened the therapeutic effect of PDRN-EV.

Notably, after DMPX addition, the distribution of CD206⁺ macrophages showed an intuitive decrease compared to the PDRN-EV treatment, suggesting that the blocking of A_2AR effectively inhibited the PDRN-EV mediated macrophages M2 polarization (Fig. 9f and g, S5b, & S5c). As shown in (Fig. 9h & i) and (Fig. S5d and S5e), PDRN-EV significantly elevated the cell proportion of proliferating cell nuclear antigen (PCNA) and down-regulated the HIF-1α expression in colonic epithelial, suggesting its role of promoting hypoxia adaptation and tissue repair. However, DMPX addition suppressed the positive effects of HIF-1α activation and cell proliferation caused by PDRN-EV. This study demonstrates that PDRN-EV exerts therapeutic effects by modulating the A_2AR-cAMP-HIF-1α signaling axis, yet key mechanistic details remain unresolved. Notably, as A_2AR is a membrane surface receptor, in vivo and in vitro uptake studies confirm that PDRN-EV functions post-cellular internalization, a process potentially involving spatiotemporal regulatory mechanisms: During vesicle-cell membrane fusion, PDRN may be released into the extracellular microenvironment via exocytosis to directly activate membrane-bound A_2AR; alternatively, internalized PDRN could generate bioactive metabolites such as adenosine through lysosomal degradation, which may activate receptors via secondary secretion. Current experimental data cannot conclusively distinguish the relative contributions of these mechanisms, necessitating systematic validation through pharmacokinetic tracing (e.g., isotope-labeled PDRN) and conditional A_2AR knockout models.

3.11. Biological safety of the PDRN-EV

Good biosafety is a prerequisite for translating nanomedicine into clinical practice. First, the in vitro biocompatibility of PDRN, EV, PDRN-EV was tested separately. EV, PDRN, PDRN-EV all showed no cytotoxicity to the macrophages (Fig. S6a). In addition, seven days after oral administration of PDRN-EV, histopathological analysis showed no obvious inflammatory infiltration, necrosis, or fibrosis in the heart, liver, spleen, lungs, and kidneys as observed in HE staining (Fig. S6b). Whole blood analysis showed that the white blood cell count (6.4 ± 0.2 × 10⁹/L), red blood cell parameters (9.2 ± 0.1 × 10⁹/L), mean corpuscular volume (50.5 ± 0.1 fL), mean corpuscular hemoglobin concentration (14.2 ± 0.1 fL) and mean corpuscular hemoglobin (283.7 ± 4.7 g/L) were all within the normal physiological range (Fig. S6c). Serum biochemical indicators showed that alanine aminotransferase (ALT 29.7 ± 2.7 U/L), aspartate transaminase (AST 148.9 ± 2.3 U/L), uric acid (UA, 97.4 ± 12.8 mmol/L), creatinine (CREA-S, 18.7 ± 0.1 μmol/L), and urea (UREA, 8.7 ± 0.3 mmol/L) had no statistical difference compared to the control group (Fig. S6d), confirming that PDRN-EV has excellent biocompatibility.

4. Conclusions

The development of EV-encapsulated PDRN (PDRN-EV) represents a paradigm-shifting strategy for oral IBD therapy, integrating the intrinsic therapeutic properties of PDRN with the superior delivery capabilities of plant EV. Our findings demonstrate that the special lipid components of tea-derived EV confer an exceptional resistance to gastric acid and nucleases. Unfortunately, regarding the MGDG/DGDG lipids targeting macrophages in EV, we were unable to directly prove the existence of these lipids either inside or on the surface of the EVs. We will revisit this question if improved methods become available in future studies. Furthermore, the PDRN-EV system overcomes the critical limitations of conventional PDRN formulations through three synergistic mechanisms: gastrointestinal protection, inflammation-targeted delivery, and cellular internalization enhancement. Importantly, the drug-carrier combination amplifies the therapeutic outcomes, in which PDRN suppresses M1 polarization and promotes cell proliferation, while tea-derived EV scavenges ROS and modulates gut microbiota diversity, thereby collectively restoring epithelial barrier function and immune homeostasis in IBD models.

The PDRN-EV therapy highlights four transformative advantages over existing IBD therapies. First, disease-specific targeting: tea-derived EV endowed the PDRN with gastrointestinal stability and facilitated the accumulation and retention in inflamed lesions, reducing systemic exposure and off-target effects. Second, multimodal therapeutic synergy: the synergistic actions of anti-inflammatory, antioxidant and regulation of intestinal microflora comprehensively address the multifactorial pathogenesis of IBD. Again, enhancing patient compliance: oral administration eliminates injection-related discomfort, which is crucial for disease management. Finally, biocompatibility: both PDRN (non-immunogenic DNA fragments) and tea-derived EV exhibit excellent safety.

Beyond IBD, the PDRN-EV platform holds broad potential for treating other inflammation-driven disorders. Preliminary data suggest its effectiveness in wound healing. Future studies may focus on three key areas. Firstly, mechanistic depth: single-cell RNA sequencing was employed to elucidate how PDRN-EV modulates the A_2AR-cAMP-HIF-1α signaling axis. Secondly, translational expansion: engineering strategies were explored to enhance the circulation half-life for systemic applications. Thirdly, clinical validation: Good Laboratory Practice (GLP)-compliant toxicology studies and Phase I trials were conducted to establish dosage regimens.

In conclusion, this study not only establishes PDRN-EV as a breakthrough oral nanomedicine for IBD treatment, but also pioneers a versatile platform for nucleic acid delivery. By harmonizing natural therapeutic carriers with engineered functionalities, we bridge the critical gap between nucleic acid pharmacology and clinical translation, opening new frontiers in precision medicine.

CRediT authorship contribution statement

Tingting Cao: Writing – original draft, Methodology, Investigation, Data curation. Runrun Wan: Methodology, Investigation, Formal analysis, Data curation. Xueru Li: Investigation, Formal analysis, Data curation. Xin Hu: Investigation. Chengbiao Hu: Investigation. Yan Liang: Investigation. Meng Deng: Investigation. Xiangdong Wang: Supervision, Methodology, Investigation. Zhang Yuan: Writing – review & editing, Supervision, Project administration, Conceptualization. Chenghu Hu: Writing – review & editing, Supervision, Project administration, Funding acquisition, Conceptualization.

Declaration of competing interest

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

Appendix A. Supplementary data

The following is the Supplementary data to this article.

Data availability

Data will be made available on request.