Extremophyte-derived exosome-like nanovesicles remodel intestinal barrier dysfunction by multi-dimensional intervention: physical barrier repair with immune homeostasis and microbiome regulation

Abstract

Background

Intestinal barrier is the body’s largest immune structure and essential for nutrient absorption. Its dysfunction allows the translocation of pathogenic substances into circulation, thereby driving the pathogenesis of inflammatory bowel disease (IBD). Plant-derived exosome-like nanovesicles (ELNs), recognized for their biocompatibility and ability to traverse biological barriers, hold considerable potential for managing intestinal inflammation. Specially, plants cultivated under extreme environmental conditions typically adapt to be stress resistant with greater accumulation of associated biologics, which may in-turn confer unique bioactivities to their respective ELNs.

Results

This study investigated the protective effects and mechanisms of ELNs derived from the extremophyte Rosa roxburghii (R-ELNs) and Artemisia sphaerocephala Krasch (A-ELNs) against intestinal barrier dysfunction. In vitro and in vivo studies indicated that the ELNs, especially R-ELNs, provided enhanced protection against intestinal barrier dysfunction. Specifically, mucus secretion and tight junction protein expression were promoted, and macrophages were polarized toward M2 anti-inflammatory phenotypes. Furthermore, R-ELNs modulated the composition of the intestinal microbiota, thereby promoting a balanced microecological environment. Importantly, the protective effect of R-ELNs was suggested to be through an inhibitory effect on excessive activation of pro-inflammatory signaling proteins (AKT, p38). Notably, exosomes (Exos) derived from R-ELN-treated M2 macrophages had distinct miRNA profiles that can target inflammatory pathway genes (Tgfbr1, Map3k7, Met), enabling anti-inflammatory roles via intercellular communication.

Conclusions

These findings suggested that R-ELNs can restore intestinal barrier dysfunction via multiple synergistic mechanisms, positioning R-ELNs as a novel and promising preventive strategy for inflammatory bowel disease.

Graphical abstract

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s12951-026-04800-9.

Introduction

The intestinal barrier functions as the largest immune barrier in the body, playing pivotal roles in the digestion and absorption of nutrients while preventing the entry of pathogens, bacteria, and toxins into the bloodstream [1, 2]. Intestinal barrier dysfunction or “leaky gut” results in noxious molecule permeation of the mucosa and subsequent entry into circulation and driving inflammatory bowel disease (IBD) [3, 4]. The limited availability of efficacious treatments, significant adverse effects of current drugs, and the associated expenses seriously impact the quality of the IBD patient’s life and place a considerable financial burden on global healthcare systems [5, 6]. Consequently, there is an urgency in the development of innovative and clinically applicable strategies to repair the intestinal barrier.

Plant-derived exosome-like nanovesicles (ELNs) have emerged as promising tools for intercellular communication and therapeutic applications [7, 8]. They show similar structures to mammalian exosomes, and can deliver biologics such as lipids, proteins, and nucleic acids, as well as plant secondary metabolites to the intestine after oral administration to exert various bioactivities. For example, plant ELNs from tea leaves, Momordica charantia, Pueraria Lobata, and broccoli have been proven the effectiveness in alleviating intestinal inflammation [9–11]. Orally administered garlic ELNs stimulated gut Akkermansia muciniphila to release extracellular vesicles, which were subsequently taken up by microglia to alleviate brain inflammation induced by high-fat diet [12]. However, ELNs from different sources exhibit marked disparities in their functional potency. Consequently, generating ELNs with enhanced functionality remains a pivotal challenge within the current research field.

Plants grown in extreme environmental conditions (drought, salinity, high-altitude hypoxia) enable them to synthesize elevated concentrations of antioxidants and rare secondary metabolites, endowing the plants with greater environmental adaptability and functional activities. For example, Rosa roxburghii, a unique wild plant resource of the Yunnan-Guizhou and the Panxi plateau, is recognized for its rich profile of natural antioxidants, including vitamin C, superoxide dismutase (SOD), and polyphenols [13, 14]. Similarly, Artemisia sphaerocephala Krasch, a xerophytic perennial herb widely distributed in arid and semi-arid regions (deserts and sandy soils) of northwest China, can synthesize substantial quantities of antioxidant and anti-inflammatory bioactives, such as terpenoids, flavonoids, and polysaccharides to resist oxidative stress and water loss [15, 16]. Based on this, we speculate that ELNs secreted by these extremophyte species may serve as a source for concentrated nanoscale carriers of such protective metabolites. Thus, ELNs may represent as a class of naturally engineered “super nanovesicles”, functionally enhanced through evolutionary pressure and biomolecular optimization. We anticipate that these ELNs may exhibit exceptional efficacy in mitigating oxidative stress and inflammatory responses, potentially offering novel therapeutic opportunities for inflammatory and oxidative stress-related pathologies.

Accordingly, in this study, ELNs were isolated from wild Rosa roxburghii and Artemisia sphaerocephala Krasch. Following characterization, the regulatory effects of ELNs on the intestinal epithelial integrity and macrophage polarization were evaluated in vitro. Then, the protective effects of ELNs on gut barrier integrity in dextran sulfate sodium (DSS)-induced colonic inflammatory disease were explored in vivo. In light of the findings, we then analyzed the material basis of ELNs and explored the underlying mechanisms that promoted intestinal barrier repair following oral administration. Our findings elucidated the effects and mechanisms of ELNs derived from plants grown in extreme conditions on enhancing intestinal barrier integrity, thereby providing an innovative strategy for the prevention of inflammatory-related intestinal diseases.

Results

Extraction and characterization of ELNs

R-ELNs and A-ELNs were isolated from the fresh juice of Rosa roxburghii and Artemisia sphaerocephala Krasch via a multi-step differential ultracentrifugation at 4℃ (Fig. 1A). Transmission electron microscopy (TEM) images and dynamic light scattering (DLS) analysis showed that these ELNs exhibited a characteristic spherical morphology with diameters of approximately 150 nm (Fig. 1B and C) (Only one TEM image was displayed, as the two ELNs shared an analogous morphology). Additionally, nanoparticle tracking analysis (NTA) indicated that the majority of ELNs were ~ 150 nm in diameter, and the particle concentration was 1.23 × 10⁶ particles/mL for R-ELNs and 8.89 × 10⁵ particles/mL for A-ELNs (Fig. 1D). After separately digested in simulated gastrointestinal fluid for 2 h, and stored at 4 °C or −20 °C for 8 d, the ELNs exhibited no significant alterations in particle size or zeta potential (Fig. S1 and S2), demonstrating their well gastrointestinal tract and storage stability. Given the well-established antioxidant properties of Rosa roxburghii and Artemisia sphaerocephala Krasch [13–16], we sought to comprehensively verify the antioxidant activities of R-ELNs and A-ELNs in vitro. As shown in Fig. 1E, both types of ELNs, particularly the R-ELNs, displayed strong ABTS·⁺/DPPH scavenging and Ferric Reducing Antioxidant Power (FRAP) activity. Specifically, at a concentration of 2 mg/mL, the ABTS·⁺ quenching capacity and DPPH radical neutralization reached approximately 70%, and FRAP activity attained 0.3. Superoxide anions (O₂·⁻) are key metabolic free radicals with potent oxidative potential. Both ELNs, especially R-ELNs, exhibited exceptional O₂·⁻ scavenging activity, achieving ~ 90% at 2 mg/mL. Collectively, these findings confirmed that both ELNs exhibit significant anti-oxidative capacity, supporting their prospective application in the prevention or treatment of oxidative stress-related diseases.

Fig. 1.

Characterization of ELNs. (A) Scheme of the separation process of ELNs. (B) TEM images of ELNs. Scale bar: 200 nm. (C-D) Particle size distribution and particle concentration of ELNs. (E) Antioxidant capacity test of ELNs. (F) Schematic illustration of 3-week oral intervention with ELNs in C57BL/6J mice. (G) Weight changes of the mice. (H) Representative Hematoxylin and eosin (H&E) staining of the major organs. Scale bars: 50 μm. (I) Blood routine, (J) liver function, and (K) kidney function related biochemical indexes of the mice

For safety concerns, a comprehensive biosafety assessment of orally administered ELNs was conducted (Fig. 1F). The effective doses of the experiment were screened. Fig. S3 revealed that both 1.5 and 3.0 mg/kg doses of R-ELNs could attenuate DSS-induced colonic shortening and weight loss, with the 3.0 mg/kg dose demonstrating significantly greater efficacy. Consequently, the 3.0 mg/kg dosage was selected for the subsequent interventions. As illustrated in Fig. 1G, no statistically significant differences in body weight were observed in any of the experimental groups after 3 weeks of ELNs treatment. Histopathological analysis revealed no significant pathological lesions in the stomach, colon, as well as the heart, liver, spleen, lungs and kidneys (Fig. 1H). Hepatocytes exhibited uniform morphology with preserved cytoarchitecture, while myocardial fibers, splenic follicles, pulmonary alveoli, and renal tubules maintained regular organization without evidence of inflammation or degeneration. Furthermore, hematological and biochemical analyses confirmed no significant alterations in serum WBC, RBC, HGB and PLT levels between ELN-treated groups and control (Fig. 1I). Given that orally delivered xenobiotics tend to accumulate in hepatic and renal tissues, we further analyzed liver and kidney function biomarkers in serum (Fig. 1J and K). The results showed that no statistically significant variations were observed in these biochemical indicators among the groups. Together, these findings demonstrate that both ELNs induce no detectable inflammatory responses or pathological effects in mice, confirming their excellent oral biosafety.

In vitro internalization profiles and in vivo biodistribution of ELNs

The internalization of ELNs mediates the transport of encapsulated compounds to intracellular compartments, which is essential for eliciting cellular responses [17]. To investigate this process, we first assessed the mucosal penetration behavior of the two ELNs in porcine intestinal mucus through 3D images of confocal laser scanning microscopy (CLSM). Fluorescein Isothiocyanate (FITC) was utilized to label mucus (green fluorescence), while PKH26 was employed to label ELNs (red fluorescence). After co-incubation for 60 min, the majority of ELNs demonstrated the capacity to penetrate the mucus, with R-ELNs exhibiting a notably superior mucus penetration ability in comparison to the A-ELNs (Fig. 2A). To further confirm this finding, the micro-rheology technique was employed to assess the mean-square displacement (MSD) curve. An increase in the MSD value corresponds to an acceleration in the movement of the particles within the mucus. As shown in Fig. 2B, the MSD value of R-ELNs was significantly higher than that of the A-ELNs, indicating that the R-ELNs showed a faster movement rate within the mucus layer. Subsequently, the diffusion behavior of the ELNs in the intestinal mucus was further studied using a Transwell-based 3D mucus model, in which the apparent permeability coefficient P_app was used to evaluate the ability of the vesicles to penetrate the mucus layer. As predicted, the P_app of the R-ELNs remained higher than that of the A-ELNs following co-incubation for 120 min. However, both ELNs exhibited a P_app value greater than 1 × 10^{− 6} cm s^{− 1} (Fig. 2C), indicating their ability to effectively penetrate through the mucus layer. Following mucus layer penetration, the uptake behaviors of ELNs by human intestinal epithelial (Caco-2) cells were evaluated. The two ELNs exhibited no discernible cytotoxicity on Caco-2 cells within the concentration range of 0–50 µg/mL (Fig. S4). After incubation for 3 h, strong red fluorescence was observed in Caco-2 cells treated with the two different ELNs, validating their effective cellular internalization by the intestinal epithelial cells. Quantitative analysis further confirmed the time-dependent enhancement of ELNs uptake (Fig. 2D).

Fig. 2.

In vitro internalization and in vivo biodistribution of ELNs. (A) 3D visualization of ELN penetration through porcine intestinal mucus after 60 min co-incubation. Green: FITC-labeled mucus, Red: PKH26-stained ELNs. Scale bars: 20 μm. (B) MSD of ELNs in porcine intestinal mucus. (C) Apparent permeability coefficient (P_app) of ELNs across the porcine intestinal mucus barrier. (D) Cellular uptake of ELNs in Caco-2 cells analyzed by CLSM and quantitative fluorescence intensity. Green: cytoskeleton stained with FITC-phalloidin, Red: ELNs stained with PKH26. Scale bars: 20 μm. (E) In vivo biodistribution tracking of Cy5.5 labeled ELNs in mice, various organs and feces at indicated time points after oral administration. (F-H) Quantitative analysis of tissue distribution of Cy5.5-labeled ELNs. Groups with different letters represent significant differences (P < 0.05), whereas groups with the same letter are not significantly different. (The groups marked with A and B, a and b, x and y are significantly different from each other)

Furthermore, the biodistribution and degradation of the ELNs labelled with Cy5.5 in mice were evaluated by an animal fluorescence imaging system after oral administration. As shown in Fig. 2E, clear fluorescence signals were detected in the upper abdomen of the two ELNs treated mice at the first 2 h administration. Then, the fluorescence intensity underwent a significant decline and gradually disseminated to the lower abdomen after a period of 4 h. Following 10 h of administration, only extremely weak fluorescence signals could be detected in vivo. Quantitative analysis of fluorescence intensity in different organs revealed that orally administered ELNs exhibited strong and persistent accumulation specifically in the intestinal region. In contrast, their distribution in liver, spleen, kidney was minimal and cleared rapidly, demonstrating their potent intestinal targeting capability. Furthermore, the fluorescence signals of the feces of mice treated by the ELNs showed that no fluorescence signal of the feces was observed at the first 0.5 h, however, it gradually intensified over time, indicating the effective metabolic processing of the ELNs in vivo. Consequently, it was determined that the ELNs could remain within the intestinal tract for about 6 h, ensuring their capacity to exert biological activities.

Reparative effects of ELNs on LPS induced disruption of epithelial tight junctions in Caco-2 cells

Given that disruption of the intestinal barrier was tightly related with inflammation [2, 3], the anti-inflammatory effects of the ELNs on lipopolysaccharide (LPS)-induced epithelial cellular inflammation were firstly assessed. As shown in Fig. S5, Caco-2 cells exhibited a significant reduction in viability, accompanied by a notable augmentation in Reactive Oxygen Species (ROS) levels at an LPS concentration of 10 µg/mL. Consequently, 10 µg/mL of LPS was selected for subsequent induction experiments, a concentration that has been reliably used in prior studies [18, 19]. Further quantitative analysis of intracellular ROS levels demonstrated that co-incubated ELNs and LPS can significantly attenuate LPS-induced intracellular ROS accumulation (Fig. S6). Considering these findings, a Caco-2 cell monolayer model that mimics the intestinal barrier was established (Fig. S7) to systematically evaluate the preventive effect of ELNs on the disruption of intestinal barrier integrity (Fig. 3A). As demonstrated by Fig. 3B, the fluorescence intensities of the two ELNs on the basal side of the Transwell exhibited a time-dependent increase, confirming their ability to traverse the cell monolayer. Subsequently, we evaluated the regulatory effects of ELNs on LPS-induced intestinal barrier dysfunction in Caco-2 monolayers model. The data demonstrated that both ELNs, particularly R-ELNs, markedly restored LPS-impaired transepithelial electrical resistance (TEER) values (Fig. 3C), indicative of improved paracellular permeability and epithelial integrity. Notably, the ELNs treatment induced a biphasic TEER response. It showed no significant effects within 12 h, a notable decrease at 24 h, and recovery by 48 h. This pattern may be attributed to the fact that the ELNs require time for cellular uptake, internalization, and the subsequent release of their active cargo to exert biological activities. The observed reduction at 24 h may be caused by the initial complex crosstalk of the ELNs with the LPS-induced inflammatory milieu, which may initiated reparative processes that temporarily affect barrier stability. The subsequent rapid recovery in TEER at 48 h indicates that the ELNs-mediated reparative mechanisms were progressed to an efficient execution phase, ultimately restoring the barrier function. Concurrently, ELN treatment attenuated LPS-triggered inflammatory responses, as evidenced by reduced levels of pro-inflammatory cytokines (IL-6, IL-1β and TNF-α) and mediators (NO, iNOS and MPO), alongside mitigation of oxidative stress markers (MDA) and enhancement of antioxidant defenses (T-SOD and GSH) (Fig. 3D and E). Furthermore, the ELNs, especially R-ELNs, significantly upregulated the content of tight junction proteins (Fig. 3F), which are critical for maintaining barrier integrity [20]. Immunofluorescence staining further corroborated these results, revealing a pronounced protective effect of TJs in the integrity of intestinal barrier (Fig. 3G). Taken together, these findings collectively demonstrated the intervening potential of the ELNs, particularly R-ELNs, in improving intestinal barrier damage.

Fig. 3.

ELNs alleviate LPS-induced disruption of epithelial tight junctions in Caco-2 cell monolayers. (A) A schematic illustration of assessing LPS induced disruption of epithelial tight junctions. (B) Transmembrane efficiency of ELNs on Caco-2 cell monolayers. (C) Effect of ELNs on the TEER value of monolayer cells induced by LPS. Effects of ELNs on inflammatory factors (D), oxidative stress-related indicators (E) and the contents of tight junction proteins (F) in LPS-induced monolayer cells. (G) Immunohistochemistry assessment of tight junction proteins expression. Scale bars: 20 μm. Groups with different letters represent significant differences (P < 0.05), whereas groups with the same letter are not significantly different

Anti-inflammatory effects of ELNs in LPS-induced RAW264.7 macrophages

To further explore the anti-inflammatory effects of ELNs, the LPS-indued RAW264.7 macrophage cell line was utilized as a model to assess their anti-inflammatory potential [21]. The cytotoxicity assay showed that the ELNs did not significantly affect macrophage proliferation at concentrations ranging from 0 to 50 µg/mL (Fig. S8). The cellular uptake of ELNs by RAW264.7, as determined by CLSM, revealed that strong green fluorescence signals were detected in both native and LPS-induced macrophages, confirming efficient internalization of ELNs (Fig. 4A). The uptake exhibited time-dependent, with signal intensity progressively increasing over the incubation period (Fig. 4B). It is well established that macrophages can be polarized into the pro-inflammatory M1 phenotype by pathogens, accompanied by the release of pro-inflammatory cytokines and induction of ROS production. In contrast, M2 phenotype macrophages are characterized by their anti-inflammatory functions [22]. As demonstrated in Fig. 4C, the application of LPS markedly enhanced the production of pro-inflammatory cytokines (TNF-α and IL-1β), while concurrently suppressing the level of the anti-inflammatory cytokine IL-10. This imbalance was effectively reversed by the co-administration of both ELNs. Concurrently, oxidative stress markers, including T-SOD, GSH and MDA, demonstrated concomitant enhancement. Furthermore, the flow cytometry was used to assess intracellular ROS production (Fig. 4D). The results revealed that the inflammatory response induced by LPS was accompanied by an increase in ROS levels. However, these increases were significantly attenuated by both ELNs, with R-ELNs showing greater efficacy. Moreover, the expression of M1 type (iNOS) and M2 type (CD206) cells were assessed to investigate the polarization of RAW264.7 macrophages. Immunofluorescence staining revealed that LPS stimulation markedly increased the abundance of M1 type (iNOS) macrophages. Conversely, ELN intervention significantly enhanced CD206 expression, promoting a shift towards the M2-like phenotype (Fig. 4E-G). Notably, due to their potent ROS-scavenging capacity, ELNs, especially R-ELNs, greatly elevated the M2/M1 ratio compared with the LPS-treated group (Fig. 4H), suggesting an enhanced anti-inflammatory response. Collectively, these results indicate that R-ELNs can promote the polarization of macrophages toward the M2 anti-inflammatory phenotype by reducing inflammatory responses and oxidative stress. This shift could effectively alleviate tissue inflammation and ameliorate immune dysregulation, highlighting the great potential of R-ELNs in alleviating inflammation to restore intestinal barrier integrity.

Fig. 4.

Intracellular uptake and anti-inflammatory effects of ELNs in LPS-induced RAW264.7 macrophages. Cellular uptake of ELNs by RAW264.7 cells analyzed by (A) CLSM and (B) quantitative fluorescence intensity. Green: ELNs stained with Dio, Blue: cell nuclei stained with Hoechst 33,342. Scale bars: 20 μm. (C) Effect of ELNs on inflammatory factors and oxidative stress-related indicators in LPS-induced RAW264.7 cells after ELNs treated. (D) Intracellular ROS levels in LPS-induced RAW264.7 cells tested by flow cytometry. (E-G) Immunofluorescence assessment of M1 (iNOS) and M2 (CD206) expression. Scale bars: 20 μm. (H) Ratio of M1 to M2 in LPS-induced RAW264.7 after ELNs treated. Groups with different letters represent significant differences (P < 0.05), whereas groups with the same letter are not significantly different

In vivo protective effect of ELNs on DSS-induced colonic impairment

Inspired by the above results that the ELNs exhibit protective effects against inflammation and intestinal barrier dysfunction, we further assessed the preventive potential of R-ELNs and A-ELNs in ameliorating DSS-induced colonic injury in mice (Fig. 5A). As illustrated in Fig. 5B-C, oral administration of the two ELNs significantly attenuated DSS-induced body weight loss, disease activity index (DAI) and spleen index elevation, as well as colon shortening compared to the DSS-treated group. Notably, the DAI score of the mice exhibited a steady upward trend from day 14, while the body weight began to decrease slightly on day 17. This discrepancy can be attributed to the composite nature of the DAI, which incorporates fecal characteristics and hematochezia, indicative of early intestinal mucosal injury prior to measurable weight loss. DSS initially induces localized colonic inflammation and epithelial damage, resulting in an elevation in DAI. Conversely, a significant reduction in weight is observed subsequent to this stage, occurring only after the intestinal damage progresses and impairs nutrient absorption. Histopathological evaluation demonstrated that both R-ELNs and A-ELNs significantly ameliorated DSS-induced colonic injury, as evidenced by preserved crypt architecture, restored mucus layer integrity, and reduced inflammatory cell infiltration (Fig. 5D). Quantitative histopathological scoring further confirmed the results (Fig. 5E).

Fig. 5.

Evaluation of the protective effect of ELNs on DSS-induced colonic impairment. (A) Schematic illustration of ELNs alleviating colonic inflammation in C57BL/6J mice. Changes in (B) body weight, DAI score and spleen index, and (C) colon length. Histological assessment of the colonic sections using (D) H&E staining and (E) histological score. Scale bars: 50 μm. (F-H) Serum levels of proinflammatory cytokines, immune-related enzymes and oxidative stress-related indicators. Groups with different letters represent significant differences (P < 0.05), whereas groups with the same letter are not significantly different

Given that colonic injury is closely related to inflammation and oxidative stress, maintaining the balance of inflammation and redox is essential for the repair of colonic damage [9]. The findings from Fig. 5F indicated that two ELNs-based interventions significantly attenuated DSS-induced inflammatory responses, as evidenced by significant downregulation the levels of pro-inflammatory cytokines such as IL-1β, IL-6 and TNF-α. Additionally, accumulating evidence suggests that elevated content of intestinal immune-related enzymes can contribute to the amplification of colonic inflammation [23]. As shown in Fig. 5G, DSS treatment significantly elevated MPO, iNOS, and NO levels. Both ELNs interventions substantially reversed these effects, with R-ELNs exhibiting particularly pronounced efficacy. Furthermore, compared with the DSS-treated group, ELNs treatment markedly reduced serum MDA levels but increased SOD contents (Fig. 5H), indicating attenuated oxidative stress. Similarly, R-ELNs demonstrated a significantly stronger restorative effect. Overall, the in vivo findings demonstrated that the ELNs, particularly R-ELNs, exhibited significant efficacy in mitigating colonic inflammation. This was evidenced by their ability to effectively downregulate pro-inflammatory cytokines and immune-related enzymes while simultaneously enhancing endogenous antioxidant enzyme activity.

Restorative effect of ELNs on DSS-induced intestinal barrier dysfunction and associated liver inflammation

Given the promising effects of ELNs on DSS-induced colonic inflammation, we further examined whether oral administration of ELNs would ameliorate intestinal barrier dysfunction and downstream consequences. The mucus layer, which is chiefly consisted of mucin secreted by goblet cells located in the intestinal epithelium, constitutes an essential structural and functional element of the intestinal barrier, playing a critical role in maintaining its integrity and selective permeability [24]. Alcian blue staining of the colonic tissues (Fig. 6A) revealed that DSS-treated mice exhibited severe damage to colonic glands and crypts, accompanied by near-complete goblet cell depletion and a marked reduction in mucosal surface mucin. Following ELNs treatment, glandular structure was restored, goblet cell numbers significantly increased with an even distribution within the crypts, and mucus layer retention was enhanced. Notably, R-ELNs demonstrated a heightened efficacy in restoring these parameters. Transmission electron microscopy (TEM) can be employed to observe the structural alterations in TJs between colonic epithelial cells [25]. As shown in Fig. 6B, the intestinal epithelial cells in control group displayed well-organized microvilli with uniform morphology, intact intercellular junctional complexes, clearly visible desmosomes, and physiologically normal intercellular spacing. Conversely, DSS-treated exhibited a marked decrease in microvilli density and length. The intercellular connection complex structure was damaged, the continuity of tight junction proteins was interrupted, and the intercellular gap was significantly widened. After treatment with both types of ELNs, tight junctions between colonic epithelial cells were significantly restored compared with the DSS group. Notably, R-ELNs demonstrated a more pronounced effect, effectively reducing the colonic villus gap and promoting enhanced alignment. The content of colonic tissue TJs (Claudin-1, Occludin-1, ZO-1) further confirmed the result (Fig. 6C). Furthermore, we quantified the expression of these three TJs in mouse colonic tissues via immunofluorescence staining (Fig. 6D). Consistent with barrier disruption, DSS administration severely suppressed the expression of all three proteins. Strikingly, both ELNs interventions substantially counteracted this suppression, with R-ELNs achieving levels to near-baseline, indicating superior efficacy in restoring intestinal barrier integrity.

Fig. 6.

ELNs alleviated DSS-induced intestinal barrier dysfunction and liver inflammation. (A) Representative images of colon sections with Alcian blue staining. Scale bars: 50 μm. (B) Ultrastructural of the colon tissue. Scale bars: 2 μm. (C) Quantitative analysis of the content of tight junction proteins in colon tissue. (D) Immunohistochemical analysis of the expression of colonic tight junction proteins. Scale bars: 50 μm. (E) Liver histology analyzed by H&E and (F) liver index. Scale bars: 50 μm. Groups with different letters represent significant differences (P < 0.05), whereas groups with the same letter are not significantly different

Moreover, cumulative evidences have demonstrated that disruption of the intestinal barrier integrity permits the translocation of pathogenic microorganisms to the liver, thereby inducing hepatic inflammation [26]. Histopathological analysis (Fig. 6E) revealed that hepatocytes in control group were distributed uniformly and arranged in a regular pattern. Conversely, DSS-treated mice exhibited enlarged intercellular spaces, disorganized hepatic cords, and significant portal inflammatory cell infiltration. After the treatment with both types of ELNs, the infiltration of inflammatory cells into the liver decreased, and the arrangement of hepatic cords were observed to have a more regular morphology. Notably, R-ELNs treatment resulted in the most significant restoration of hepatic histology. These results were further confirmed by the liver index (Fig. 6F). Together, these findings indicate that oral administration of R-ELNs can effectively ameliorate intestinal barrier dysfunction and the associated hepatic inflammation through the synergistic promotion of goblet cell-derived mucin production and elevating the expression of TJs, thereby rebalancing the gut-liver axis.

Impact of ELNs on gut microbiota

Previous research has demonstrated that disruption of the intestinal barrier by inflammation drives dysbiosis of the gut microbiota [27, 28]. To further investigate the regulatory effect of extreme-environment plant ELNs on DSS-induced intestinal barrier injury, the gut microbiota profiling was analyzed by 16 S rDNA gene sequencing. Venn diagram analysis revealed significant disparities in the number of unique amplicon sequence variants (ASVs) among the four groups. The counts were 619 for the control group, 247 for the DSS group, 339 for the DSS + R-ELNs group, and 319 for the DSS + A-ELNs group (Fig. 7A). The analysis of α-diversity employing the Chao1 and Shannon indices indicated a decline in bacterial diversity within the DSS group. Conversely, treatment with ELNs resulted in a substantial augmentation in bacterial diversity (Fig. 7B and C). Principal coordinate analysis (PCoA) demonstrated that the composition of the gut microbiota at the genus level in the DSS group differed markedly from that in the Control group. Both the R-ELNs and A-ELNs groups exhibited an intermediate clustering between the Control and DSS groups (Fig. 7D), suggesting that ELNs treatment partially restored the disruption of microbial composition induced by DSS. Subsequently, we examined the phylum level composition of the intestinal microbiota after different treatments. As illustrated in Fig. 7E, Bacteroidetes and Firmicutes were the main bacteria. Notably, a decreased Firmicutes/Bacteroidetes (F/B) ratio is strongly associated with intestinal inflammation [29]. In the DSS group, this ratio exhibited a substantial decline to 0.54. However, R-ELNs intervention significantly increased the ratio to 0.76, thereby demonstrating their effective alleviation of intestinal inflammation. Heatmap analysis at the genus level (Fig. 7F) showed that DSS treatment significantly increased the abundance of harmful bacteria such as Bacteroides and Alistipes, which are positively associated with inflammation [30]. In contrast, the levels of beneficial bacteria, including Faecalibaculum, Bifidobacterium, Dubosiella, and Akkermansia, were markedly reduced. These beneficial genera contribute to gut health through complementary mechanisms. Faecalibaculum produces butyrate to nourish colonocytes and exert anti-inflammatory effects [31]. Bifidobacterium strengthens the mucosal barrier by enhancing tight junctions and promoting mucus secretion [32]. Dubosiella modulates immune balance by promoting Treg activity and suppressing Th17 responses and Akkermansia maintains homeostasis by enhancing mucin production and inhibiting TLR4 signaling [33]. Following the administration of R-ELNs and A-ELNs, particularly with R-ELNs, the elevated abundances of pathogenic bacteria were significantly reduced, while the depleted beneficial bacteria were restored. These shifts indicate that ELNs derived from Rosa roxburghii can partially counteract DSS-induced dysbiosis, thereby improving intestinal inflammation and barrier integrity through metabolic, barrier-protective, and immuno-regulatory pathways.

Fig. 7.

ELNs recovered the intestinal flora homeostasis in DSS-induced mice. (A) Venn diagram of species among the various groups. Microbial community α-diversity assessed by (B) Chao1 and (C) Shannon indices. (D) The principal coordinates analysis (PCoA) of microbiota after different treatments. (E) Bar graph of the microbiotal relative abundance at the phylum level and the ratio of Firmicutes/Bacteroidota. (F) Heatmap of the relative abundance at the microbial genus level. (G) LEfSe analysis of microbiota. (H) PICRUSt2-based functional prediction at KEGG Pathway Level 2. Groups with different letters represent significant differences (P < 0.05), whereas groups with the same letter are not significantly different

LEfSe analysis of the phylogenetic tree further revealed that R-ELNs and A-ELNs, especially R-ELNs, induced multi-level remodeling of the gut microbiota across phylogenetic lineages. Significant changes were observed at multiple taxonomic levels, including phyla (e.g., Actinomycetota, Bacteroidota), classes (e.g., Actinobacteria, Clostridia), orders (e.g., Bifidobacteriales, Lactobacillales) and families. At the level of the genus, the altered microbiota were principally divided into two categories. The promotion of genera such as Bifidobacterium, Faecalibaculum, and several Lachnospiraceae members is conducive to the fermentation of carbohydrates, the production of short-chain fatty acids, and the maintenance of the intestinal barrier. The occurrence of inflammation and mucosa damage is associated with genera including Escherichia, Shigella, Desulfovibrio, Bilophila, Mucispirillum, and Dubosiella, etc. These results indicate that DSS-induced dysbiosis primarily reflects a structural imbalance between beneficial bacteria (especially SCFA producers) and opportunistic pathogens (e.g., pro-inflammatory bacteria). Following R-ELNs intervention, these two functional groups shifted in opposite directions on the phylogenetic tree, suggesting that the treatment may restore intestinal microecological balance by synergistically promoting the growth of SCFA-producing taxa and suppressing the proliferation of pro-inflammatory bacteria (Fig. 7G). Furthermore, PICRUSt2-based functional prediction at KEGG Pathway Level 2 revealed that (Fig. 7H), compared with the control group, the DSS group exhibited marked alterations in multiple functional pathways associated with microbial basal metabolism and host-microbiota interactions, indicating DSS-induced functional dysbiosis of the gut microbiota. Intervention with either R-ELNs or A-ELNs partially restored these functional disturbances, with R-ELNs showing a more pronounced regulatory effect. The restoration was primarily characterized by an enhancement of activity in pathways related to carbohydrate metabolism, energy metabolism, glycan biosynthesis and metabolism, immune regulation, membrane transport, as well as replication and repair. The results demonstrate that DSS modelling significantly disrupts the functional profile of the gut microbiota at the secondary pathway level. In contrast, R-ELNs intervention more effectively mitigates this functional imbalance, highlighting its superior role in restoring microbiota functions related to metabolism, immunity, and cellular repair. Collectively, these results indicate that R-ELNs intervention can reshape the gut microbiota by promoting the proliferation of beneficial bacteria, including short-chain fatty acid producers and immuno-modulatory taxa, thereby alleviating intestinal inflammation and repairing the intestinal barrier damage.

Material basis for the role of R-ELNs

Given that plant-derived ELNs are known to contain abundant bioactive lipids, proteins, and RNA molecules that govern their biological functions [34], we conducted a comprehensive compositional analysis. Non-targeted lipidomic analysis identified a total of 26 lipids in R-ELNs, with triglycerides (TAG, 54.71%), phosphatidylcholine (PC, 31.51%), and phosphatidylethanolamine (PE, 8.42%) representing the predominant classes (Fig. 8A). Previous studies suggested that TAG can be internalized by macrophages and possesses immune-modulating properties [35]. PC is shown to exert anti-inflammatory effects and plays a significant role in ameliorating intestinal inflammatory diseases [36]. PE contributes to cellular structural stability and also demonstrates potent antioxidant functions [37]. Quantitative proteomic profiling identified 836 proteins and 1,908 polypeptides in R-ELNs. Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis revealed that these proteins participate in the regulation of multiple biological pathways, including oxidative phosphorylation, endocytosis, ascorbate and aldarate metabolism, and the MAPK signaling pathway, etc. (Fig. 8B). Additionally, the construction of RNA-seq libraries and subsequent sequencing of R-ELNs revealed that the top 20 microRNAs were implicated in the regulation of multiple inflammatory signaling pathways, predominantly the PI3K/AKT and MAPK pathways (Fig. 8C). Collectively, R-ELNs are enriched in diverse lipids, proteins, and miRNAs implicated in inflammation regulation. These constituents position R-ELNs as promising bioactives for ameliorating intestinal barrier damage. However, further research is required to identify the specific molecular classes responsible for their anti-inflammatory effects.

Fig. 8.

The material basis and primary anti-inflammatory constituents of R-ELNs. (A) Lipid profiles of R-ELNs. KEGG annotated charts of (B) proteins and (C) bubble charts of miRNAs in R-ELNs. (D) Schematic diagram for removing different components of R-ELNs after enzymatic digestion. (E) The level of inflammatory cytokines by LPS induced-RAW264.7 macrophage after incubation with R-ELNs of different treatments for 12 h. (F-G) Immunofluorescence assessment of AKT and p-AKT, p38 and p-p38 expression in RAW264.7 macrophage. (H) The level of inflammatory cytokines by LPS induced-RAW264.7 macrophage after the addition of activators and inhibitors of AKT and p38. (I-J) Immunofluorescence assessment of AKT and p-AKT, p38 and p-p38 expression in colon tissue. Scale bars: 100 μm. Groups with different letters represent significant differences (P < 0.05), whereas groups with the same letter are not significantly different

Previously, it has been demonstrated that RNase or protease can be used to remove RNA or protein in ELNs, respectively (Fig. 8D) [38]. As shown in Fig. 8E, the anti-inflammatory ability of R-ELNs markedly decreased after digestion of either RNA or protein. Among these treatments, protein digestion showed a stronger inhibitory effect on IL-1β and TNF-α. Simultaneous removal of protein and RNA significantly dampened the anti-inflammatory activity of R-ELNs. These findings collectively indicate that the anti-inflammatory properties of R-ELNs arise from the combined contribution of their lipid, protein, and RNA components, with RNA contributing playing the more critical role. Both protein and miRNA content in R-ELNs were suggested to regulate signaling pathways such as PI3K/AKT and MAPK (Fig. 8B and C). Such pathways play significant roles in intestinal barrier function [39, 40]. Therefore, we next sought to investigate the regulation effects of R-ELNs on AKT and p38 proteins. Immunofluorescence analysis revealed that R-ELN treatment significantly inhibited the phosphorylation levels of both AKT and p38 compared with the LPS-induced group (Fig. 8F and G), suggesting that R-ELNs may alleviate LPS-induced cell inflammation by blocking these signaling proteins. To further validate pathway dependency, the effects of ELNs on LPS-induced inflammatory factors in macrophages were examined using activators and inhibitors of AKT and p38. As demonstrated in Fig. 8H, the anti-inflammatory effect of ELNs was significantly attenuated upon the addition of either AKT or p38 activators. Conversely, the addition of inhibitors did not result in a further reduction in inflammatory factor levels, suggesting that ELNs already exert a significant degree of suppression on these pathways. These results confirm that ELNs can effectively alleviate inflammation by inhibiting both AKT and p38 signaling. Subsequent immunofluorescence staining of mouse colon tissues resulted in consistent protein-level changes in AKT and p38, further supporting the involvement of these pathways in vivo (Fig. 8I and J).

Anti-inflammatory effects of exosomes secreted by RAW264.7 macrophages

As key guardians of intestinal immunity, macrophages are indispensable for maintaining immune homeostasis, promoting the repair of damaged tissues, and ensuring the integrity of the epithelial barrier [41]. It has been demonstrated that Exos derived from M2-type macrophages have been demonstrated to promote epithelial cell proliferation, and the miR-590-3p they carry can enhance epithelial wound healing by inhibiting the expression of pro-inflammatory factors [42]. In this study, we demonstrated that R-ELNs intervention facilitated the transformation of LPS-induced macrophages from M1 type to M2 type (Fig. 4). To further explore the mechanism by which R-ELNs ameliorate intestinal barrier damage, the composition and functional profiles of miRNAs within Exos secreted by RAW264.7 macrophages subjected to different interventions were characterized. As shown in Fig. 9A, a total of 383 common miRNAs were identified in the Exos from both LPS-treated and R-ELNs-intervened macrophages. The LPS-treated group contained 32 unique miRNAs, whereas the number of unique miRNAs increased to 75 after R-ELNs intervention. NAT analysis showed that the diameters of these ELNs were approximately 120 nm, the particle concentration was 8.83 × 10⁵ particles/mL for Exos and 9.65 × 10⁵ particles/mL for R-Exos (Fig. 9B). Following prediction with the StarBase, miRWalk, and miRTarBase databases, putative target genes of exosomal miRNAs from differentially treated macrophages were functionally annotated via Gene Ontology (GO) and KEGG analyses. The GO functional annotation results (Fig. 9C) indicated that the predicted target genes were significantly enriched in inflammation-related molecular functions, including catalytic activity (GO:0003824), molecular transducer activity (GO:0060089), binding (GO:0005488), molecular function regulator activity (GO:0098772), molecular adaptor activity (GO:0060090), ATP-dependent activity (GO:0140657) and transcription regulator activity (GO:0140110). With regard to the biological processes, the target genes were involved in a variety of functions, including response to stimulus (GO:0050896), localization (GO:0051179), multicellular organismal process (GO:0032501), developmental process (GO:0032502), cellular process (GO:0009987), reproductive process (GO:0022414), homeostatic process (GO:0042592), biological regulation (GO:0065007), and immune system process (GO:0002376). These results indicate that the target genes of macrophage exosomal miRNAs play a significant role in the regulation of inflammatory processes. Furthermore, KEGG pathway analysis (Fig. 9D) revealed that these target genes were significantly enriched in multiple pathways associated with intestinal barrier function, including Bacterial invasion of epithelial cells, Pathogenic Escherichia coli infection, Adherens junction, Endocytosis, NF-κB signaling pathway, MAPK signaling pathway, Wnt signaling pathway, TNF signaling pathway and PI3K-Akt signaling pathway, etc. By screening the key target genes within inflammatory pathways, we found that the main target genes were Tgfbr1, Map3k7 and Met (Fig. 9E). The specific miRNAs in macrophage exosomes that target these genes were mmu-miR-122-5p, mmu-miR-425-5p, mmu-let-7b-5p, mmu-miR-301b-3p and mmu-miR-140-5p for Tgfbr1; mmu-miR-181a-5p for Map3k7; and mmu-miR-34b-5p for Met (Fig. 9F). According to the reported studies, the expression levels of mmu-miR-122-5p, mmu-miR-425-5p, mmu-let-7b-5p, and mmu-miR-181a-5p are negatively correlated with inflammation, whereas those of mmu-miR-301b-3p, mmu-miR-140-5p and mmu-miR-34b-5p exhibit a positive correlation with the inflammatory processes. It can be seen from Fig. 9F that R-ELNs intervention significantly modulated the expression of these inflammation-related miRNAs compared with the LPS group, redirecting the molecular profile toward an anti-inflammatory phenotype. Additionally, dual-luciferase reporter assays further demonstrated that mmu-let-7b-5p significantly reduced the luciferase activity of the wild-type TGFBR1 3′UTR reporter but not that of the mutant reporter, confirming TGFBR1 as a direct target of mmu-let-7b-5p in NIH3T3 cells (Fig. 9G). Similarly, the miRNAs mmu-miR-181a-5p and mmu-miR-34b-5p specifically suppressed the luciferase activity of wild-type reporters for Map3k7 and Met, respectively. However, mutations in their predicted binding sites abolished this suppression (Fig. 9H and I). This finding collectively verifies Map3k7 and Met as direct targets of mmmu-miR-181a-5p and mmmu-miR-34b-5p, respectively. Moreover, we evaluated the effects of the two Exo types on the levels of pro‑inflammatory cytokines (IL‑1β and TNF‑α) and the expression of TJs in LPS‑induced Caco-2 cells. Exos derived from macrophages pre‑treated with R‑ELNs significantly suppressed LPS‑induced production of IL‑1β and TNF‑α (Fig. 9J). Moreover, immunofluorescence staining showed that these R‑ELN‑primed Exos markedly enhanced the expression of ZO-1, Claudin-1 and Occludin, indicating their superior ability to restore intestinal barrier integrity (Fig. 9K). Consequently, the target genes of miRNAs derived from macrophage Exos are significantly enriched in pathways related to intestinal barrier function and inflammatory. Intervention with R-ELNs effectively reshaped the expression of these miRNAs, thereby modulating key inflammatory pathways and effectively suppressing inflammatory responses. These insights also provide potential molecular targets and new strategies for the treatment of inflammatory disease.

Fig. 9.

Anti-inflammatory effect of miRNAs in Exo secreted by RAW264.7 macrophage. (A) Venn diagram of shared and unique miRNAs species in each group. (B) Particle size distribution of Exos. (C) Molecular function and biological process of miRNAs in Exo secreted by RAW264.7 macrophage based on the GO classification. (D) Metabolic pathways of miRNAs in Exo secreted by RAW264.7 macrophage based on the KEGG analysis. (E) Main target genes of inflammation-related pathways involved in the miRNAs. (F) Expression of miRNAs in macrophages Exos that target the main genes involved in the inflammatory regulation-related pathways. (G-I) Predicted sites of TGFBR1 to bind mmu-let-7b-5p, Map3k7 to bind mmu-miR-181a-5p, Met to bind mmu-miR-34b-5p and the mutant sequences, as well as the Relative luciferase activity in NIH3T3 cells transfected with different types of target gens and miRNAs. (J) Levels of the pro-inflammatory factors IL-1β and TNF-α in Caco-2 cells. (K) Immunofluorescence analysis of the expression of TJs. Groups with different letters represent significant differences (P < 0.05), whereas groups with the same letter are not significantly different

Discussion

As research on plant-derived ELNs continues to expand, their biological components and functional properties have gained increasing attention in the field of nanobiomedicine [43]. Our study introduces a new category of ELNs sourced from plants growing in extreme environments, such as plateau or sand thorn forest biomes, and investigates their roles in repairing intestinal barrier damage. Plants grown in extreme environments can be considered as specialized “bioactive ingredient refineries” shaped by natural selection pressures [44, 45]. Their distinctive habitats endow them with significant anti-inflammatory and antioxidant biological activities. The plant ELNs possess similar bioactivities to those of the source plants [9–11]. In this study, we demonstrated that R-ELNs and A-ELNs, particularly R-ELNs, can efficiently promote the repair of intestinal barrier damage via a triplet of mechanisms: physical barrier enhancement, immune homeostasis, and microbial regulation. These findings not only advance the understanding of plant ELN biology, but also pave the way for developing effective and eco-friendly strategies for preventing intestinal barrier injury-related pathologies.

Previous work has demonstrated the potential that plant-derived ELNs hold in the treatment and intervention of various intestinal inflammatory diseases. For example, garlic-derived ELNs have been found to alleviate colonic inflammation by reshaping gut microbiota, marked by a significant increase in the relative abundance of beneficial Bacteroides. ELNs from Momordica charantia were shown to enhance intestinal barrier function by promoting mucus secretion, upregulating the expression of tight junction proteins such as Occludin and ZO-1, and improving the microbial composition. These effects are mainly attributed to the remarkable stability of ELNs, which can withstand the harsh acidic and enzymatic environment of the gastrointestinal tract following oral administration, ensuring their effective delivery to the colon [46, 47]. This study further confirmed that R-ELNs and A-ELNs could remain in the intestinal tract for up to 6 h after oral administration, providing sufficient time for their bioactivities to be exerted. Furthermore, ELNs derived from plants that have the capabilities to survive in extreme environments are likely to possess enhanced bioactivities, due to their high levels of antioxidant and anti-inflammatory components accumulated by the parent plants as an adaptation to environmental stressors. The dose of ELNs employed in the present study (3 mg/kg) was substantially lower than those reported for garlic ELNs (100 mg/kg) and Momordica charantia ELNs (30 mg/kg). The comparable outcomes observed at this reduced dose suggest that the ELNs employed in this study may possess higher specific activity.

As expected, both R-ELNs and A-ELNs effectively penetrated the mucus layer and were taken up by epithelial cells, a finding consistent with previous reports [48]. Additionally, these ELNs could effectively alleviate the LPS-induced epithelial cell inflammation. Concurrently, they could lower ROS levels in RAW264.7 macrophages and promote the polarization of macrophages from pro-inflammatory M1 phenotype toward the anti-inflammatory M2 phenotype. To further validate their bioactivities, a DSS-induced intestinal barrier injury model in mice was employed. The results indicated that these ELNs significantly ameliorated intestinal barrier damage by attenuating intestinal inflammation and oxidative stress, enhancing mucus secretion and tight junction protein expression, and promoting the growth of beneficial bacteria implicated in the regulation of intestinal inflammation. These findings indicated a bidirectional regulatory relationship exists between the intestinal barrier and the gut microbiota. The structural and functional integrity of the intestinal barrier is crucial for maintaining homeostatic balance of the gut microbiota, and any compromise in barrier function can directly lead to microbial dysbiosis. Conversely, the gut microbiota plays a fundamental role in the development and immunoregulation of the intestinal barrier. Disruption of the microbial community can promote the release of pro-inflammatory factors and activate intestinal immune responses, thereby heightening susceptibility to localized inflammation. Persistent inflammation subsequently disrupts tight junction proteins and degrades the mucus layer, establishing a self-perpetuating cycle of “inflammation-barrier dysfunction”.

Plant ELNs are known to carry diverse biologics, such as lipids, proteins, and RNAs, etc [9–11]. and our data revealed that the compositional profile of R-ELNs was indicative of anti-inflammatory outcomes. Thus, we further evaluated the anti-inflammatory effects of their protein and RNA hydrolysates. Notably, the anti-inflammatory efficacy was significantly reduced in both hydrolyzed forms compared with the intact ELNs, suggesting that the lipids, proteins, and RNAs within R-ELNs act synergistically to mediate the anti-inflammatory process, with RNA contributing the most important role. Further investigation into the expression of anti-inflammatory-related proteins revealed that the R-ELNs intervention can effectively inhibit the expression of AKT and p38MAPK proteins in macrophages, thereby contributing to the alleviation of inflammation. Subsequent research will focus on identifying which specific proteins or RNAs are pharmacologically active, thereby providing a foundation for the clinical application of R-ELNs.

Furthermore, the Exos secreted by living cells act as important intercellular messengers. They can deliver their cargo of lipids, proteins, miRNAs, and other bioactive molecules to the recipient cells, thereby directly regulating intercellular communication [49, 50]. Studies have demonstrated that Exos secreted by M2-type macrophages are enriched with a diverse of bioactive factors, which confer them potent anti-inflammatory functions [51]. Given that the R-ELNs intervention reduced ROS levels in RAW264.7 macrophages and facilitated their transition from M1 to M2 states (Fig. 4), we subsequently examined the RNA components of Exos secreted by these macrophages and evaluated their anti-inflammatory effects. The results demonstrated that Exos derived from R-ELNs treated macrophages exhibited enrichment of miRNA species, which can target inflammatory pathway genes, such as Tgfbr1, Map3k7 and Met, enabling them to play an anti-inflammatory role via intercellular communication, with potential implications in modulating inflammatory processes that arise in IBD. When applied to an LPS-induced Caco-2 cell model, these Exos notably attenuated the production of inflammatory factors and the enhanced the expression of TJs (Fig. 9), demonstrating their potent anti-inflammatory and intestinal barrier repair capacity.

Collectively, we propose that ELNs enhance the intestinal barrier through coordination of the physical barrier-immune-microbiota axis at multiple levels. Specifically, R-ELNs can reshape the gut microbiota, thereby promoting the proliferation of beneficial bacteria such as Faecalibaculum, Bifidobacterium, Dubosiella, and Akkermansia. These bacteria enhance the production of metabolites such as short-chain fatty acids, which directly strengthen the physical barrier by supporting intestinal epithelial tight junctions and mucus secretion. Moreover, microbial metabolites have been shown to facilitate the polarization of macrophages towards the reparative M2 phenotype. Activated M2 macrophages could suppress the release of pro-inflammatory factors (e.g., TNF-α, IL-1β) and promote epithelial repair, thereby alleviating immune-mediated barrier injury. The restored barrier integrity and resolved inflammation, in turn, further inhibit excessive M1-like macrophage activation and create a favorable niche for beneficial bacterial colonization (M2). Together, these interactions collectively establish a self-reinforcing cycle of barrier repair, immune alleviation and microbiota modulation. This multi-target regulatory network underlies the ability of R-ELNs to mitigate intestinal inflammation and restore barrier function. Subsequent studies may concentrate on the mechanistic intricacies of R-ELN-macrophage interactions, with a particular emphasis on the specific internalization of R-ELNs by macrophages. Concurrently, the therapeutic potential of Exos secreted by R-ELNs modulated macrophages should be assessed in models of intestinal barrier injury. By delineating the complete data chain from “R-ELNs - macrophages - intestinal barrier damage repair”, a new theory will be provided for precisely understanding the therapeutic action of R-ELNs on intestinal barrier impairment.

Conclusion

The results from in vitro and in vivo studies revealed that the R-ELNs derived from unique wild plant cultivated in Yunnan-Guizhou and the Panxi plateau, can protect intestinal barrier function via multiple synergistic mechanisms: (1) Effective intestinal targeting and barrier repair ability, whereby R-ELNs enhance the physical barrier function of the intestinal epithelium by promoting mucus secretion and upregulating tight junction protein expression; (2) pronounced anti-inflammatory and immunomodulatory effects from R-ELNs induction of macrophage polarization from the pro-inflammatory M1 phenotype towards the anti-inflammatory M2 phenotype, and markedly reducing levels of pro-inflammatory factors and ROS; and (3) modulation of gut microbial composition by R-ELNs, providing a positively regulated intestinal microecology and forming a more balanced microbiota and overall improvement to the gut microenvironment. Molecular analysis revealed that R-ELNs inhibited the excessive activation of pivotal pro-inflammatory signaling proteins AKT and p38, thereby blocking inflammatory signal transduction. Notably, R-ELNs treatment altered the miRNA cargo within Exos released by LPS-induced macrophages. These Exos are enriched in miRNAs targeting inflammatory pathway genes such as Tgfbr1, Map3k7, and Met, thereby exerting the anti-inflammatory and intestinal barrier repair effect through intercellular communication. Collectively, these findings demonstrate that R-ELNs possess favorable oral biosafety and are capable of effectively maintaining intestinal barrier integrity through multiple mechanisms, including physical barrier repair, immune response regulation and gut microbiota remodeling. This study thereby provides a promising innovative ingredient for developing preventive functional oral formulations for populations at high risk of colitis.

Experimental section

Isolation and characterization of ELNs

Rosa roxburghii (Longli, Guizhou) and Artemisia sphaerocephala Krasch seeds (Erdos City, Inner Mongolia) were thoroughly rinsed with distilled water and then homogenized in ice-cold PBS (pH 7.2). The resulting homogenate was allowed to stand at 4 °C for 12 h. Differential centrifugation was then performed at 4 °C: 1,200 g for 30 min, 3,000 g for 60 min, and 10,000 g for 60 min. After that, the supernatant was carefully collected and then ultracentrifuged at 100,000 g for 120 min at 4 °C to collect ELNs pellets. The ELNs obtained were quantified using a BCA protein quantification assay kit (Beyotime, Haimen, China) and then stored at −20 °C for further use.

The morphology of the ELNs was examined by TEM (Tecnai G2 20, Netherlands). The sample were prepared and observed according to a previously reported method [52]. The particle size and zeta potentials of the ELNs were determined by dynamic light scattering (DLS 2000, Malvern). The particle size and their concentration were measured by nanoparticle tracking analysis (NS300, Malvern). Triplicate measurements were performed for each sample, and the results were reported as the mean value.

Stability of ELNs

The gastrointestinal and storage stability of the ELNs were evaluated using DLS following a 2 h incubation in simulated gastric and intestinal fluids, and 12 days of storage at 4 °C and − 20 °C, respectively.

In vitro antioxidant ability of ELNs

ABTS·⁺ assay: The ABTS⁺ radical working solution was prepared by incubating a mixture of ABTS⁻ (7.4 mM) and K₂S₂O₈ (2.6 mM) at 25℃ for 12 h. Subsequently, the ELNs at concentrations of 0.5, 1, 2, 3, 4 and 5 mg/mL were individually mixed with the ABTS working solution, and incubated at 25℃ for 6 min. The absorbance was measured at 734 nm by a UV-Vis spectrophotometer. All procedures were performed in the dark.

DPPH assay: The DPPH dissolved in ethanol were initially prepared and then stored at 4 °C prior to use. Afterwards, the ELNs at specific concentrations were separately mixed with an equal volume of DPPH, and then incubated at 37 °C for 20 min in the dark. The absorbance was measured at 517 nm by a UV-Vis spectrophotometer.

FRAP assay: The FRAP working solution was prepared by mixing acetate buffer (300 mmol/L), tripyridyl-triazine (TPTZ) (10 mmol/L), and FeCl₃ (20 mmol/L) at a volume ratio of 10: 1: 1. Then, the ELNs at the specific concentrations were respectively mixed with the FRAP working solution, and incubated at 37 °C for 5 min in the dark. The absorbance was measured at 593 nm by a UV-Vis spectrophotometer.

O₂·⁻ assay: The ELNs at the specific concentrations were mixed with an equal volume of pyrogallol (0.2 M). After vortex-mixing, the reactions were carried out at 25 °C for 4 min and then quenched with HCl (8 M). The absorbance was conducted at 325 nm by a UV-Vis spectrophotometer.

Animals and ethics

Male C57 BL/6J mice (body weight 20–22 g) were purchased from the Xi’an Medical College (Xi’an, China). All animal experiments were performed and approved by the Animal Care and Use Committee of Xi’an Medical College (approval number: XYLS2024134).

Effective dose of ELNs for the prevention of intestinal barrier injury in vivo

The male C57BL/6J mice were randomly divided into four groups: Control, DSS, DSS + Low-dose ELNs and DSS+High-dose A-ELNs. The model of intestinal barrier injury was established by oral administration of 4% DSS for 7 consecutive days. Oral intervention with R-ELNs (1.5 mg/kg and 3 mg/kg) was implemented by starting 2 weeks prior to DSS treatment and continuing throughout the DSS exposure period. During the entire experimental period, the body weight was recorded daily. Following the experiment, the length of the colon in different treated groups were measured.

Long-term biosafety evaluation

To evaluate the long-term biosafety of ELNs, acclimated C57BL/6J mice were orally administered ELNs at a dose of 3 mg/kg for 21 consecutive days. After treatment completion, blood samples were collected for comprehensive hematological analysis and hepatic/renal function assessment. Subsequently, the mice were euthanized, and major organs (stomach, colon, heart, liver, spleen, lung and kidney) were excised for H&E staining to further evaluate the structural integrity and pathological alterations of the tissues.

Ex vivo mucus penetrating property

The ex vivo mucus penetrating property was assessed according to a previously reported method with some modifications [52]. Briefly, fresh porcine intestine segments were collected, thoroughly rinsed with 0.9% NaCl, and then meticulously dissected. Subsequently, mucus was scraped from the luminal surface using glass slides and then stored at 4 °C for subsequent use. To assess the mucus penetration of the ELNs, the freshly collected intestinal mucus of 1 mL was stained with 20 µL of FITC (10 mg/mL) and then transferred to a confocal dish. Subsequently, a 50 µL volume of the ELNs labeled with PKH26 was uniformly applied to the mucus surface. After 60 min of incubation, the ex vivo mucus penetrating property of the ELNs was recorded by CLSM (LSM800, Carl Zeiss AG) with excitation/emission wavelengths (λex/λem) set at 488 nm/520 nm for FITC and 551 nm/567 nm for PKH26, respectively.

Multi-particle tracking in mucus

For particle tracking studies, 20 mL of the freshly collected mucus was placed into a glass tube, followed by the addition of PKH26-labeled ELNs. Nanoparticle diffusion dynamics within the mucus matrix were tracked using diffusing wave spectroscopy (DWS) with a rheolaser MASTER instrument (Formulaction, France). The acquired data were then expressed as MSD.

Transport of ELNs across the mucus layer

The permeation capacity of ELNs through the intestinal mucus layer was further assessed using a Transwell system. Briefly, 5 mL of the freshly collected intestinal mucus was dispensed into the upper chamber, while 1.5 mL of PBS buffer was placed in the lower chamber. Subsequently, PKH26-labeled ELNs of 0.5 mL were introduced into the upper chamber. Incubation proceeded at 37 °C on an orbital shaker. Aliquots were collected from the lower chamber at 0.5, 1, 1.5, and 2 h for absorbance measurement by a microplate reader (Synergy H1, ThermoFisher Scientific). The quantification of permeability was conducted using the apparent permeability coefficient (P_app), calculated as follows:

where dQ/dt is the cumulative transport rate of ELNs across the membrane, A is the membrane surface area, and C₀ is the initial ELN concentration in the donor chamber.

Epithelial cellular uptake

The cytotoxicity of ELNs at varying concentrations was quantified using the MTT assay. Then, Caco-2 cells pre-cultured on coverslips in 12-well plates (37 °C, 24 h) were incubated with PKH26-labeled ELNs of 20 µg/mL at 37 °C for 3 h. After quenching with ice-cold PBS, cells were fixed with 4% paraformaldehyde (PFA) for 15 min, and then stained with FITC-phalloidin for 30 min in the dark. Finally, coverslips were mounted with antifade reagent for CLSM imaging, with λex/λem set at 488 nm/520 nm for FITC and 551 nm/567 nm for PKH26, respectively.

For uptake quantitative analysis, Caco-2 cells were incubated with PKH26-labeled ELNs at 37 °C for designated durations. Subsequently, the cells were trypsinized and centrifuged at 1,000 ×g for 5 min at 4 °C. The resulting pellets were lysed in ice-cold RIPA buffer, vortexed vigorously for 1 min, and incubated at 4 °C for 40 min to obtain cell lysate. The lysate obtained was then centrifuged at 13,000 ×g for 10 min at 4 °C. The fluorescence intensity of the obtaining supernatant was quantified by the microplate reader at λex/λem of 551 nm/567 nm.

In vivo biodistribution and degradation

Mice were orally administered Cy5.5-labeled ELNs (2 mg/mL) via gavage. At specific time points post-administration (0.5, 1, 2, 4, 6, 8, 10 h), mice were anesthetized with 2% isoflurane, and the in vivo distribution of ELNs was monitored using an IVIS fluorescence imaging system (PerkinElmer, Spokane, USA) at a wavelength of 680 nm for λex and 710 nm for λem. Subsequently, mice were euthanized to collect gastrointestinal tissues, liver, spleen, kidney and fecal pellets for ex vivo analysis of ELN accumulation patterns at each time point.

Construction of cellular inflammatory model

Caco-2 cells were seeded at a density of 1 × 10⁴ and cultured for 48 h. Then, the culture medium was discarded, and the cells were exposed to LPS at concentrations of 0, 0.1, 1, 10, 20, and 50 µg/mL. After incubation at 37 °C for 24 h, the cell survival rates of each group were measured using the MTT method. Furthermore, the level of ROS in cells was quantified to verify the successful construction of the inflammatory model. The diluted DCFH-DA probe was added to the cells and incubated for 30 min in the dark. The intracellular ROS levels were determined by the microplate reader at λex of 488 nm and λem of 525 nm.

Trans-cellular investigation

Caco-2 cell suspension at a density of 4 × 10⁵ cells was seeded in the upper chamber of 12-well Transwell plates, and 1.5 mL of culture medium was added to the lower chamber. After gentle swirling, plates were incubated at 37 °C for 21 days to grow into monolayers. The integrity of the cell monolayer was verified by FITC-labeled dextran (FD4) permeation and TEER. After that, PKH26-labeled ELNs of 0.5 mL were added onto the cell monolayer. Samples were taken from the base chamber at predetermined time points (0.5, 1, 2, 4, 6, 8, 10 h), and the permeation quantity of ELNs was measured using a microplate reader at wavelengths of 551 nm (λex) and 567 nm (λem).

To further investigate the effects of ELNs on cellular inflammation and epithelial barrier integrity, Caco-2 cells with monolayer integrity were co-treated with 10 µg/mL LPS and 20 µg/mL ELNs for 48 h. The relative TEER was measured. Subsequently, the cells were trypsinized, pelleted by centrifugation, and lysed at 4 °C for 40 min. The lysate was subjected to centrifugation at 13,000 × g for 20 min to collect the supernatant. Then, the levels of inflammatory mediators (IL-1β, IL-6, TNF-α, NO, iNOS, and MPO), oxidative stress markers (MDA, T-SOD and GSH) and tight junction proteins (ZO-1, Occludin and Claudin-1) in the supernatant were detected by commercial ELISA kits (ExCell, Shanghai, China). The expression levels of ZO-1, occludin, and claudin-1 were evaluated by immunohistochemistry.

Anti-inflammatory effect of ELNs in RAW264.7 cells

The anti-inflammatory efficacy of ELNs was further evaluated using RAW264.7 macrophages. Firstly, the cytotoxicity of ELNs on RAW264.7 macrophages at varying concentrations was quantified using the MTT assay. Then, the cellular uptake of ELNs by macrophages and LPS-intervened macrophages was determined. Briefly, macrophages were incubated with 20 µg/mL DiO-labeled ELNs in the absence or presence of 10 µg/mL LPS for 3 h. The cell nuclei were stained with Hoechst 33,342 for 30 min. The cellular uptake of ELNs was observed using CLSM at λex/λem of 484 nm/501 nm for DiO and 346 nm/460 nm for Hoechst 33,342. To quantify the uptake efficiency, the RAW264.7 cells were incubated with DiO-labeled ELNs at 37 °C for 1, 3, 6, and 9 h, respectively. Subsequently, the cell culture supernatant was collected, and the fluorescence intensity in the supernatant was measured using a microplate reader at λex/λem of 484 nm/501 nm. To assess cytokine production, the supernatant was collected and concentrations of TNF-α, IL-1β, IL-10, T-SOD, GSH, and MDA were quantified using commercial ELISA kits (ExCell, Shanghai, China). For intracellular ROS detection, cells treated with LPS and ELNs were stained with 5 µM DCFDA for 30 min. The ROS-scavenging capacity of ELNs was evaluated using a flow cytometer (CytoFLEX, Beckman Coulter, USA). For macrophage polarization analysis, cells under various treatments were immunostained for iNOS (M1 marker) and CD206 (M2 marker), followed by observation via CLSM.

In vivo efficacy preventive intestinal barrier injury

The male C57BL/6J mice were randomly divided into four groups: Control, DSS, DSS + R-ELNs and DSS + A-ELNs. The process of experimentation is consistent with the ELNs dose screening section. The dosage of the ELNs selected was 3 mg/kg. During the entire experimental period, the body weight was recorded daily, and the fecal consistency as well as the occurrence of bloody stools were assessed to calculate the DAI. Following the experiment, plasma samples were collected from mice for analysis using commercially available assay kits. Major organs were harvested for organ coefficient calculation. Colon tissue proteins were extracted with a commercial extraction kit to quantify the levels of TJs (ZO-1, Occludin. Claudin-1). Histopathological evaluation was performed via H&E staining, while Alcian blue staining was used to assess mucus secretion and goblet cell abundance. Ultrastructural morphology was examined using TEM, and the expression of key TJs was detected by immunohistochemical staining.

16 S rDNA gene sequencing analysis

The feces were collected on the final day of the trial and further analyzed using 16 S rDNA sequencing. Genomic DNA was extracted by the DNeasy PowerSoil Kit (QIAGEN, Inc., Netherlands), and its integrity was verified through 1.0% agarose gel electrophoresis. Samples meeting quality thresholds were selected for PCR amplification targeting the V3-V4 hypervariable region of the 16 S rDNA gene with specific primers. The resulting amplicons were purified and then sequenced using Illumina MiSeq platform. Taxonomic annotation of the sequencing data was performed using the Silva database to analyze compositional differences in the gut microbiota among groups. Further bioinformatic analysis was performed based on ASVs.

Lipidomic, proteomic, and small RNA analysis

Lipidomic and proteomic profiling of R-ELNs was outsourced to Baiqu Biomedical Technology (Shanghai, China) and mircoRNA sequencing analysis was outsourced to Science Compass (Shanghai, China). For lipidomic analysis, lipids were extracted from R-ELNs using a chloroform/methanol mixture (2:1, v/v). Separation was achieved using a Vanquish UHPLC system (Thermo Fisher Scientific) with a Phenomenex Kinetex C18 column (2.1 × 100 mm, 1.7 μm), followed by mass spectrometry on a Q Exactive Orbitrap instrument (Thermo Fisher Scientific). For proteomic analysis, proteins were enzymatically digested into peptides, separated on a nanoElute2 nano-UPLC system (Bruker Daltonics), and analyzed using a timsTOF Pro2 mass spectrometer. For small RNA sequencing, total RNA was extracted from R-ELNs with the miRNeasy Mini Kit (Qiagen, USA) and quantified using a Quantus Fluorometer (E6150, Promega). The construction of sequencing libraries was undertaken using the QIAseq RNA Library Kit. The expression levels of miRNA were quantified and annotated using miRbase. The top 20 most abundant miRNAs, ranked by Transcripts Per Million (TPM), were subsequently subjected to KEGG pathway enrichment analysis.

Determination the composition of exosomal RNA secreted by macrophages

RAW264.7 cells were seeded in 10 cm culture dishes at a density of 1 × 10⁴ cells/mL. After treatments with 10 µg/mL LPS for 12 h, 20 µg/ml of R-ELNs were added and co- incubated for additional 12 h. RAW264.7 macrophages that were only treated with LPS were used as the control. Then, the original culture medium was discarded, and the RAW264.7 cells were cultured in exosome-free serum medium for another 12 h to allow Exos secretion. Finally, the Exos secreted by macrophages was collected using the exosome isolation kit, and the miRNA sequencing library was constructed and sequenced via an Illumina NovaSeq™ XPlus platform. The target genes of miRNAs were predicted by the StarBase, miRWalk, and miRTarBase databases. Functional annotation of these target genes was then performed through GO and KEGG pathway analyses.

Dual-luciferase reporter assay

Dual-luciferase reporter vectors were constructed by inserting wild-type (WT) or mutant (Mut) sequences of TGFBR1, Map3k7, and Met, which were provided by ELK Biotechnology. These constructs were separately co-transfected into NIH3T3 cells with the corresponding miRNA mimics (mmu-let-7b-5p, mmu-miR-181a-5p, or mmu-miR-34b-5p) or a negative control. Luciferase activity was measured by the dual-luciferase reporter assay kit (Beyotime, Shanghai).

Statistical analysis

All experiments were performed at least three times, and the results were presented as mean ± standard deviation (SD). Statistical analyses were performed by one-way ANOVA or t-tests. A P value < 0.05 was considered statistically significant.

Abbreviations

IBD

Inflammatory bowel disease

ELNs

Exosome-like nanovesicles

SOD

Superoxide dismutase

DSS

Dextran sulfate sodium

R-ELNs

Rosa roxburghii exosome-like nanovesicles

A-ELNs

Artemisia sphaerocephala Krasch exosome-like nanovesicles

Exos

Exosomes

miRNAs

MicroRNAs

TEM

Transmission electron microscopy

DLS

Dynamic light scattering

FRAP

Ferric Reducing Antioxidant Power

O₂⁻

Superoxide anions

H&E

Hematoxylin and eosin

CLSM

Confocal laser scanning microscopy

FITC

Fluorescein Isothiocyanate

MSD

Mean-square displacement

LPS

Lipopolysaccharide

ROS

Reactive Oxygen Species

TEER

Transepithelial electrical resistance

DAI

Disease activity index

AVSs

Amplicon sequence variants

KEGG

Kyoto Encyclopedia of Genes and Genomes

GO

Gene Ontology

TPTZ

Tripyridyl-triazine

DWS

Diffusing wave spectroscopy

Author contributions

X.Y. and D.L. performed conceptualization; L.D., G.Y. and T.L. performed methodology; L.D., T. L. and Y. H. performed software; G.Y., Y. H. and Y.D performed data curation; D.L., L.D. and G.Y. performed investigation; G.Y., T.L. and Y.D performed validation; T.L., A.M., G.S. and G.L. performed formal analysis; X.Y. performed supervision; X.Y., D.L. and G.L. performed funding acquisition; G.Y. and Y. H. performed visualization; X.Y. performed project administration; G.S. performed resources; D.L. and L.D. performed the original draft; A.M. and X.Y. performed review & editing.

Funding

This work was supported by the Natural Science Foundation of China (32302276, 32472493), the Shaanxi Science and Technology association (2024JC-JCQN-23, 2023GXLH-078).

Data availability

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

Declarations

Ethics approval and consent to participate

All animal experiments were performed and approved by the Animal Care and Use Committee of Xi’an Medical College (approval number: XYLS2024134).

Consent for publication

All authors of this study agreed to publish.

Competing interests

The authors declare no competing interests.