Polarized macrophage-derived microvesicles differentially modulate intestinal barrier dysfunction and endoplasmic reticulum stress in ulcerative colitis

Abstract

Macrophage polarization is critically involved in the initiation and progression of ulcerative colitis (UC); however, whether microvesicles (MVs) released from macrophages with distinct polarization states exert different effects on the intestinal epithelial barrier and endoplasmic reticulum (ER) stress remains unclear. Here, MVs were isolated from M0, M1, and M2 macrophages and evaluated in a dextran sulfate sodium (DSS)-induced mouse colitis model and an inflammatory cytokine-stimulated NCM460 epithelial cell model. M2 macrophage-derived MVs (M2-MVs) mitigated body-weight loss, disease activity, colon shortening, and histopathological injury and partially restored Claudin-1, Occludin, ZO-1, and MUC2 expression. In contrast, M1 macrophage-derived MVs (M1-MVs) tended to aggravate inflammation and barrier disruption. Mechanistically, M2-MVs reduced IL-1β, TNF-α, and IL-6, increased IL-10, and inhibited GRP78, PERK/eIF2α phosphorylation, CHOP, and cleaved caspase-3. Consistent effects were observed in vitro, where M2-MVs restored barrier proteins and reduced ER stress and apoptosis. Moreover, the ER-stress inhibitor 4-phenylbutyric acid attenuated the detrimental effects of M1-MVs, supporting ER stress as an important downstream mechanism. By systematically comparing MVs from M0, M1, and M2 macrophages in parallel, this study reveals a polarization-dependent, bidirectional pattern of macrophage–epithelial communication in experimental colitis and provides preclinical evidence supporting M2-MVs as candidate cell-free effectors for intestinal mucosal repair.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s13062-026-00923-z.

Introduction

Ulcerative colitis (UC) represents a major clinical entity within the diverse range of inflammatory bowel diseases (IBD) [1]. The onset and progression of this disorder involve complex interactions among numerous elements, including genetic susceptibilities, external triggers, dysregulated intestinal microflora, aberrant activation of mucosal defences, and compromised epithelial integrity, which collectively result in persistent intestinal inflammation and structural damage [2, 3]. Despite available therapies, many patients experience poor responses, recurrent flares, adverse effects, and require surgery [4–6]. Consequently, the development of innovative therapeutic approaches that can concurrently inhibit inflammation and repair intestinal epithelial barrier stability remains a crucial unmet need in the management of UC [7].

Intestinal macrophages occupy a vital junction between immune modulation and mucosal restoration, with their phenotypic shift influencing the progression of intestinal inflammation [8]. Within the inflammatory milieu of UC, macrophages frequently transition into a pro-inflammatory M1-like state, secreting various pro-inflammatory cytokines and intensifying mucosal damage. Conversely, M2-like macrophages are more intimately linked with inflammatory resolution, epithelial regeneration, and the preservation of mucosal stability [9, 10]. Consequently, altering the functional states of macrophages or exploiting their paracrine activities could potentially disrupt the pathological loop of inflammatory expansion and barrier dysfunction in UC [11, 12].

Extracellular vesicles (EVs) perform a vital function in mediating cellular interactions, as they transport bioactive signals. These particles can encapsulate origin-specific proteins, lipids, nucleic acids, and metabolites, thus modulating inflammatory reactions, tissue healing, and barrier function within target cells [13–15]. As a significant subset of EVs, microvesicles (MVs) are primarily produced via outward budding of the cell membrane. Relative to direct cellular delivery, macrophage-derived MVs might act as a cell-free delivery vehicle that maintains the functional traits of the parent cells while minimising constraints linked to cell therapy, including cellular viability, phenotypic instability, and safety risks [16, 17].

Dysregulated engagement of the endoplasmic reticulum (ER) pathway serves as a critical mechanistic bridge connecting pro-inflammatory signals with the impairment of epithelial permeability [18, 19]. Persistent ER stress can promote inflammatory amplification, epithelial cell dysfunction, and cell death through PERK/eIF2α/CHOP-related signaling, thereby further weakening the intestinal mucosal barrier and facilitating UC progression [20, 21]. Consistent with this mechanistic framework, clinical studies have revealed activation of the PERK/eIF2α pathway in the colonic mucosa of patients with UC, while experimental evidence indicates that excessive CHOP expression impairs epithelial regeneration and increases susceptibility to intestinal injury [22, 23].

Based on this background, we hypothesized that macrophage polarization may determine the regulatory effects of macrophage-derived MVs on UC-like intestinal inflammation. In particular, M2-MVs may preserve intestinal epithelial integrity and promote barrier homeostasis by suppressing PERK/eIF2α/CHOP-associated ER stress, while M1-MVs may be involved in a pro-inflammatory signal transmission and in the worsening of ER stress and mucosal damages. Although macrophage polarization and extracellular vesicles have each been implicated in UC, the polarization-dependent functional heterogeneity of macrophage-derived MVs has not been systematically compared in parallel. Therefore, we generated MVs from M0, M1, and M2 macrophages and evaluated their effects in a DSS-induced colitis model, an inflammatory cytokine-stimulated intestinal epithelial cell model, and a 4-phenylbutyric acid rescue experiment. This design was intended to define the bidirectional effects of vesicle polarity and to test ER stress as a downstream mechanistic link.

Materials and methods

Macrophage culture and polarization

RAW264.7 cells, a murine macrophage cell line, were purchased from Shanghai Jinyuan Biotechnology Co., Ltd., Shanghai, China. The cells were cultured in DMEM medium from Servicebio (Wuhan, China) supplemented with high-glucose content. The medium contained 10% fetal bovine serum (FBS) purchased from Gibco, New Zealand. Subsequently, the cells were cultured in a humidified incubator set to 37 °C, with an environment comprising 5% CO₂. Polarization induction was carried out according to the experimental grouping once the cells reached about 70%–80% confluence. Untreated cells served as the M0 group. Cells stimulated with 100 ng/mL LPS (Sigma-Aldrich, USA) combined with 10 ng/mL recombinant murine IFN-γ (PeproTech, USA) for 24–48 h were defined as the M1 group, whereas cells stimulated with 20 ng/mL recombinant murine IL-4 and 20 ng/mL recombinant murine IL-13 (PeproTech, USA) for 24–48 h were defined as the M2 group. After induction, cells were collected for subsequent experiments, and the expression levels of markers related to polarization were assessed through Western blotting and RT-qPCR methods. The verification of M1 polarization was established through heightened expression of CD86 and iNOS, whereas M2 polarization was validated by the increased expression of CD206 and ARG1.

Isolation and purification of microvesicles from different macrophage sources

To minimize the influence of FBS-derived extracellular vesicles on the isolation of macrophage-derived MVs, EV-depleted FBS was prepared before the experiment. In short, fetal bovine serum (FBS) underwent ultracentrifugation at a force of 120,000 × g for a duration of 12 h at a temperature of 4 °C. The upper approximately 70% of the serum was carefully collected without disturbing the pellet. The gathered serum underwent sterilization via a 0.22-micrometer filter. Subsequently, it was divided into aliquots and kept at a temperature of − 20 °C until it was ready for utilization. Following the treatment of M0, M1, and M2 macrophages, the existing culture medium was swapped out for a new medium that contained FBS depleted of EVs. Subsequently, the cells were cultivated for an extra 48 h. Once that time was completed the supernatants of the cell cultures were collected. After the second ultracentrifugation at 4 °C, MVs were pelleted by differential centrifugation. First, centrifugation of supernatant at 2,000×g for 20 min to remove suspended cells and large debris. Subsequently, homogenate was centrifuged at 16,500×g for 75 min on secondary centrifugation in order to recover MVs. Pellets obtained were finally resuspended in sterile PBS (Servicebio, China), and spun down at 20,000×g for 60 min to remove remaining soluble proteins and medium components. Finally, they were resuspended in sterile PBS, divided into aliquots, and preserved at − 80 °C until needed.

Establishment of an inflammatory NCM460 cell model

The human-derived colonic epithelial cell line NCM460 was obtained from Shanghai Jinyuan Biotechnology Co., Ltd. (China). These cells were cultured in high-glucose DMEM (Servicebio, China), enriched with 10% FBS (Gibco, New Zealand), and maintained at 37 °C in a humidified environment containing 5% CO₂. Log-phase cells in good condition were selected for experiments. When the confluence reached around 80%, a model for in vitro inflammation was created through the simultaneous stimulation with 30 ng/mL of recombinant human IFN-γ, TNF-α, and IL-1β (ABclonal, China) for a duration of 48 h. This cytokine mixture was referred to as the cytokine cocktail (CK) in subsequent experiments.

Assessment of macrophage-derived MVs internalization

To examine the uptake of macrophage-derived MVs by NCM460 cells, NCM460 cells in the log phase were placed in confocal culture dishes and grown to a suitable density. An equal amount of MVs, based on protein concentration, was resuspended in 200 µL PBS and labeled with DiI fluorescent dye (Yuanye, China) at a final concentration of 5–10 µM. The solutions were maintained at a temperature of 37 °C for 30 min in a dark setting. After the labeling procedure, MVs suspensions were combined with PBS and centrifuged at 20,000×g for a duration of 60 min to remove any unbound dye. Microvesicles labeled with DiI were then added to the NCM460 cell culture system at a concentration of 15 µg/mL and co-incubated for 12 h in darkness. After this incubation, the cells were washed three times with PBS and fixed in 4% paraformaldehyde for a duration of 15 min. The nuclei and cell membranes were stained using Hoechst 33,342 and DiO (Yuanye, China), respectively, at concentrations ranging from 5 to 10 µM. Following a 20-minute incubation at 37 °C in the absence of light, the cells were washed with PBS. Images were taken using either a fluorescence or confocal microscope after the mounting process. DiI-labeled MVs appeared as red fluorescence, Hoechst 33,342-labeled nuclei as blue fluorescence, and DiO-labeled cell membranes as green fluorescence.

Animal housing and ethical compliance

C57BL/6JNifdc male mice 6–8 weeks of age and weighing approximately 20 g were purchased from Beijing Vital River Laboratory Animal Technology Co., Ltd. (China). They were kept in specific pathogen-free (SPF) conditions with a room temperature of 24 ± 1 °C, 55 ± 5% relative humidity, and a 12/12 light/dark cycle. All procedures involving the animals were conducted in accordance with established ethical guidelines and standards for animal welfare.

DSS-induced acute colitis model and treatment protocol

After a 7-day acclimatization phase, mice were assigned randomly into five groups: Control, DSS, DSS+M0-MVs, DSS+M1-MVs, and DSS+M2-MVs, each consisting of three mice. Except for the Control group, all other mice had ad libitum access to drinking water containing 2.5% (w/v) dextran sulfate sodium (DSS; APExBIO, USA) for a duration of 7 days to establish an acute model of UC-like colitis. Mice assigned to the Control group received autoclaved water to drink throughout the experiment’s duration. MVs intervention took place on days 0, 2, 4, and 6 during the modeling phase. The groups of mice designated as DSS+M0-MVs, DSS+M1-MVs, and DSS+M2-MVs were administered intraperitoneal injections of 50 µg of the respective M0-MVs, M1-MVs, or M2-MVs, all diluted in 200 µL of sterile PBS. In contrast, the DSS mice received an identical volume of PBS at the corresponding time points. All mice were sacrificed on day 8 for sample collection. During the experiment, general condition, activity, stool consistency, fecal bleeding, and body weight were recorded daily to evaluate DSS-induced colitis progression.

Protein extraction and western blot analysis

Tissues from the colon of each group were promptly transferred to pre-chilled DPBS that was free of Ca²⁺ and Mg²⁺, and they were washed to eliminate any intestinal contents and blood. After colonic epithelial cells were isolated using the EDTA method, frozen in liquid nitrogen and kept at − 80 °C until further processed for protein analysis. The total proteins were isolated by RIPA lysis buffer(Solarbio, China) that supplemented with 1 mM PMSF (Solarbio, China), and the combine uses of phosphatase inhibitor(cocktail) (MedChemExpress, USA) to avoid the degradation of protein and dephosphorylation. Supernatants were then collected after centrifugation of samples following lysis. Protein concentrations were measured using a BCA protein assay kit (Thermo Fisher Scientific, USA). Protein samples were mixed at equal amounts with 5×SDS loading buffer (Biosharp, China), denatured by boiling at 100 °C for 10 min. The resulting protein samples were then separated using 10% SDS-PAGE gels (Servicebio, China) before being transferred to PVDF membranes (Millipore, USA). Following the transfer process, membranes were incubated at room temperature for 1 h with a BSA blocking buffer (Servicebio, China) to reduce nonspecific interactions. Subsequently, the membranes were exposed to the appropriate primary antibodies overnight at 4 °C. These primary antibodies included those targeting ER stress-related proteins such as p-PERK, PERK, GRP78, p-eIF2α, eIF2α, and CHOP (Affinity Biosciences, USA); proteins associated with the epithelial barrier like Claudin-1 (Affinity Biosciences, USA), Occludin (Boster Biological Technology, China), and ZO-1 (Proteintech, USA); as well as β-actin (ABclonal, China), which served as the internal control. The next day, membranes underwent three washes with TBS buffer (Servicebio, China) supplemented with Tween 20 (Merck, Germany), each lasting 5 min. The samples were subsequently incubated at room temperature for one hour with the suitable secondary antibodies. After the washes, protein bands were identified with an ECL chemiluminescence substrate (Biosharp, China), and images were acquired using a chemiluminescence imaging system. The protein expression levels were normalized against β-actin or the relevant total protein.

Total RNA isolation and quantitative real-time PCR analysis

RNA extraction from tissue samples was performed following the guidelines specified in the E.Z.N.A.^® Total RNA Kit I (Omega Bio-tek, USA). The concentration and purity of the RNA were evaluated, and those samples that met the quality criteria were utilized for reverse transcription. An equivalent amount of total RNA was transformed into cDNA utilizing a reverse transcription kit (Thermo Fisher Scientific, USA). Quantitative real-time PCR was performed using the generated cDNA as the template in conjunction with SYBR Green qPCR reagent (CWBIO, China). Primers specific to the genes were generated by Sangon Biotech Co., Ltd. (China), and the sequences of these primers are listed in Table 1. Each sample consisted of three technical replicates, and the experiments were performed independently a minimum of three times. The mRNA expression levels of the target gene were standardized against an internal reference gene, and the relative expression levels were calculated using the 2^−ΔΔCt methodology.

Table 1.

Primer sequences used for RT-qPCR

Gene	Forward sequence (5′→3′)	Reverse sequence (5′→3′)
Mouse CD86	GGTGGCCTTTTTGACACTCTC	TGAGGTAGAGGTAGGAGGATCTT
Mouse iNOS	GTTCTCAGCCCAACAATACAAGA	GTGGACGGGTCGATGTCAC
Mouse CD206	CTCTGTTCAGCTATTGGACGC	CGGAATTTCTGGGATTCAGCTTC
Mouse ARG1	CTCCAAGCCAAAGTCCTTAGAG	AGGAGCTGTCATTAGGGACATC
Mouse IL-1β	GCAACTGTTCCTGAACTCAACT	ATCTTTTGGGGTCCGTCAACT
Mouse IL-6	TAGTCCTTCCTACCCCAATTTCC	TTGGTCCTTAGCCACTCCTTC
Mouse TNF-α	ACTGAACTTCGGGGTGATCG	CCACTTGGTGGTTTGTGAGTG
Mouse IL-10	GCTCTTACTGACTGGCATGAG	CGCAGCTCTAGGAGCATGTG
Mouse β-actin	GGCTGTATTCCCCTCCATCG	CCAGTTGGTAACAATGCCATGT
Human IL-1β	TGAGCTCGCCAGTGAAATGA	TGAAGCCCTTGCTGTAGTGG
Human IL-6	CCGGGAACGAAAGAGAAGCTC	ACCGAAGGCGCTTGTGGAG
Human TNF-α	CCCTCACACTCAGATCATCTTCT	GCTACGACGTGGGCTACAG
Human IL-10	GACTTTAAGGGTTACCTGGGTTG	TCACATGCGCCTTGATGTCTG
Human GAPDH	AATGGGCAGCCGTTAGGAAA	GCCCAATACGACCAAATCAGAG

4-PBA treatment and annexin V/PI apoptosis assay

NCM460 cells that were stimulated with CK underwent treatment with 2 mM of the ER stress inhibitor 4-phenylbutyric acid (4-PBA; Sigma-Aldrich, USA), in conjunction with either M1-MVs or M2-MVs, according to the specified experimental groups. After a 48-hour incubation period, the cells were collected, stained using an Annexin V-FITC/PI apoptosis detection kit (Yeasen Biotechnology, China), and evaluated by flow cytometry.

Statistical analysis

Quantifications were performed with the program ImageJ to measure grayscale values and fluorescent intensities in the experimental images, while statistical analysis and processing of the data were performed with SPSS version 22.0 as well as GraphPad Prism version 9.5 for creating the graphs. Data are expressed as the mean ± SD. All in vitro studies had at least three biological replicates with each group being composed of a minimum of three animals. For comparisons involving more than two groups, one-way analysis of variance (one-way ANOVA) followed by Tukey’s multiple-comparisons test was used. The threshold of statistical significance was set at the P < 0.05 level (Scheme 1).

Scheme 1.

MVs derived from differently polarized macrophages exert distinct regulatory effects on intestinal epithelial injury during ulcerative colitis

Results

RAW264.7 macrophages exhibit distinct polarization phenotypes

Untreated RAW264.7 cells were, therefore, chosen as M0 macrophages to set up a working system, and will be considered so in the context of the macrophage and MVs with diverse activation states. M1-like polarization was induced in combination with LPS and IFN-γ, while M2-like polarization was achieved with a combination of IL-4 and IL-13. After polarization, culture supernatants were harvested and MVs were obtained through differential centrifugation (Fig. 1A). Bright-light microscopy indicated clear morphological differences between the types of induction. Cells in M0 culture were predominantly round or oval, and displayed a partial adherence characteristic. Group M1 cells had an increased spreading, while some cells even appeared flattened. On the other hand, M2 cells were more elongated and showed spindle-like changes, indicating that the skewing stimuli seemed to bring out different morphology in RAW264.7 cells (Fig. 1B). Western blotting analyses of polarization markers revealed that, as compared to M0 counterparts, M1 macrophages had higher levels of CD86 and iNOS proteins, and M2 macrophages had increased CD206 and ARG1 levels (Fig. 1C-G). In the same line of evidence, qRT-PCR confirmed these protein expression profiles, revealing a significant induction of CD86 and iNOS in M1, and also a marked enhancement of the mRNA levels of CD206 and ARG1 in M2-differentiated macrophages (Fig. 1H-K). These findings indicate that RAW264.7-derived M1-like and M2-like macrophage polarization models were successfully established, providing a cellular basis for the subsequent isolation and functional investigation of MVs derived from different macrophage phenotypes.

Fig. 1.

Identification of macrophage polarization. A: Schematic illustration of macrophage polarization and MVs isolation. B: Representative bright-field images of M0, M1, and M2 macrophages. Yellow arrows indicate cells with polarization-associated morphological changes. C–G: Protein expression levels of CD86, iNOS, CD206, and ARG1 were evaluated through Western blot analysis accompanied by densitometric evaluation. H–K: The mRNA levels of CD86, iNOS, CD206, and ARG1 were measured using RT-qPCR. Results are shown as mean ± SD, with n = 3. The figure indicates statistical significance

MVs derived from differently polarized macrophages are isolated, characterized, and taken up by NCM460 cells

After successful generation of M0, M1, and M2 macrophages, transmission electron microscopy showed that M0-MVs, M1-MVs, and M2-MVs all exhibited typical membranous vesicle-like structures with clear particle boundaries and predominantly round or oval morphologies (Fig. 2A). Western blot analysis showed that MVs from all three groups expressed the extracellular vesicle-associated marker proteins CD9, CD63, and TSG101, indicating that the isolated particles possessed EV-like molecular characteristics (Fig. 2B). Nanoparticle tracking analysis further demonstrated that M0-MVs, M1-MVs, and M2-MVs were mainly distributed within the nanoscale range, with relatively concentrated peak sizes, suggesting a stable particle size distribution of the isolated MVs (Fig. 2C–E). To verify the uptake capacity of colonic epithelial cells for macrophage-derived MVs, DiI-labeled M0-MVs, M1-MVs, or M2-MVs were co-incubated with NCM460 cells, with DiO used to label cell membranes and Hoechst used to label nuclei. Fluorescence imaging showed distinct red fluorescent puncta within NCM460 cells after incubation with MVs from different sources, mainly distributed in the cytoplasm and perinuclear regions (Fig. 2F). This work shows that MVs from macrophages are indeed able to be taken up by colonic epithelial cells and experimentally supports their functional role in subsequent behavior of these cells.

Fig. 2.

Isolation, characterization, and NCM460 cell uptake of macrophage-derived MVs. A: Images obtained via transmission electron microscopy that illustrate the morphological features of M0-MVs, M1-MVs, and M2-MVs. B: Analysis using Western blot to assess the MVs marker proteins CD9, CD63, and TSG101. C–E: Nanoparticle tracking analysis revealing the distribution of particle sizes and concentrations for M0-MVs, M1-MVs, and M2-MVs. F: Uptake of MVs from different macrophage phenotypes by NCM460 cells, as observed using DiI-labeled MVs, DiO-labeled cell membranes, and Hoechst-labeled nuclei. Scale bars are indicated in the figure

M2 macrophage-derived MVs alleviate whereas M1 macrophage-derived MVs exacerbate DSS-induced acute colitis in mice

In order to evaluate the potential effects of MVs secreted by differently polarized macrophages in lesions mimicking ulcerative colitis, an experimental acute colitis model was generated by administering DSS 2.5% in the drinking water. M0-MVs, M1-MVs, M2-MVs or PBS were administrated systemically to mice i.p. on day 0, 2, 4 and 6 and harvested on day 8 (Fig. 3A). The DSS mice showed typical colitis symptoms when compared to the Control group, with decreased body weight, increased disease activity index (DAI), colon shortening and more severe gross damage, thus corroborating model efficiency. After MVs intervention, disease phenotypes differed markedly among groups. M0-MVs showed only limited protective effects against weight loss, a rise in DAI levels, and a reduction in colon length. Conversely, M1-MVs worsened weight loss, heightened DAI ratings, and intensified damage to the colon. In comparison, M2-MVs significantly mitigated weight loss, diminished DAI ratings, ameliorated colon length, and substantially lowered gross injury ratings (Fig. 3B–F). Hematoxylin and eosin staining further confirmed these phenotypic changes. The DSS group displayed obvious destruction of mucosal architecture, crypt damage, inflammatory cell infiltration, and tissue edema. Pathological injury was further aggravated in the DSS+M1-MVs group, as indicated by expanded mucosal destruction and enhanced inflammatory infiltration. In contrast, the DSS+M2-MVs group showed relatively preserved colonic mucosal architecture, partial restoration of crypt arrangement and epithelial continuity, reduced inflammatory cell infiltration, and markedly decreased histological scores (Fig. 3G, H). These results indicate that MVs derived from macrophages with different polarization states exert opposite biological effects in DSS-induced colitis. M1-MVs exhibit pro-inflammatory and injury-promoting effects, whereas M2-MVs effectively alleviate DSS-induced colitis phenotypes and histological damage.

Fig. 3.

Effects of MVs derived from macrophages with different polarization states on DSS-induced colitis phenotypes in mice. A: Diagram illustrating the acute colitis model induced by DSS and the intervention protocol involving MVs. B: Changes in body weight. C: Scores for the disease activity index (DAI). D: Exemplary macroscopic images of colonic tissues. E: Quantitative assessment of colon length. F: Gross colonic injury scores. G: Histopathological scores of colonic tissues. H: Hematoxylin and eosin staining showing pathological changes in colonic tissues from each group. The outcomes are presented as the average value along with the standard deviation, derived from a sample size of three (n = 3). Statistical significance is indicated in the figure

M2-MVs restore the colonic epithelial barrier after DSS-induced injury

The alteration of the intestinal epithelial barrier represents a significant pathological characteristic linked to colitis induced by DSS. A study was carried out to assess the impact of extracellular vesicles (MVs) from various sources on the integrity of barriers by analyzing the expression levels of tight-junction proteins such as Claudin-1, Occludin, and ZO-1 in colonic tissues. Western blot analysis showed that DSS treatment significantly reduced tight junction protein expression compared with the Control group, suggesting impaired colonic epithelial tight junction structures. M0-MVs had no obvious restorative effect on these proteins, whereas M1-MVs further suppressed tight junction protein expression. In contrast, M2-MVs markedly upregulated the levels of Claudin-1, Occludin, and ZO-1 (Fig. 4A–D). These results suggest that M2-MVs improve DSS-induced epithelial barrier injury at the protein expression level. Immunofluorescence analysis further supported the Western blot findings. In the Control group, tight junction proteins showed a relatively continuous distribution in the colonic epithelial region, indicating intact epithelial junctional structures. In the DSS group, these fluorescence signals were markedly weakened and disorganized. The DSS+M1-MVs group showed a further reduction in barrier protein signals, whereas the DSS+M2-MVs group exhibited enhanced signals with a relatively continuous distribution in the epithelial region. In addition, MUC2, an important component of the mucus barrier, was decreased in the DSS group, while M2-MVs treatment enhanced MUC2 fluorescence intensity, suggesting that M2-MVs may contribute not only to the restoration of the tight junction barrier but also to the maintenance of mucus barrier integrity (Fig. 4E-I). Collectively, these findings indicate that M2-MVs effectively alleviate DSS-induced colonic epithelial barrier disruption, whereas M1-MVs may further exacerbate barrier injury.

Fig. 4.

Effects of MVs derived from macrophages with different polarization states on intestinal epithelial barrier injury in DSS-treated mice. A–D: The levels of Claudin-1, Occludin, and ZO-1 proteins in colonic tissues were evaluated using Western blotting followed by densitometric analysis. E–I: Immunofluorescence assessments were performed to investigate the expression and distribution of Claudin-1, Occludin, ZO-1, and MUC2, along with the measurement of the area that tested positive for staining. The results are shown as mean ± SD, based on n = 3. The figure indicates statistical significance

M2-MVs suppress ER stress and inflammatory responses in the colon of DSS-treated mice

To gain a deeper understanding of the mechanisms through which M2-MVs alleviate colitis and barrier damage, alterations in pathways related to ER stress were investigated. Transmission electron microscopy showed that colonic epithelial cells in the Control group had relatively intact ultrastructural features. In contrast, the DSS group showed obvious cellular ultrastructural damage, characterized by disordered organelle distribution, marked ER dilation, reduced mitochondrial number with blurred cristae, increased autolysosomes, impaired intercellular tight junctions, and sparse, fractured, or even detached microvilli. These injuries were further aggravated after M1-MVs treatment. By contrast, M2-MVs markedly alleviated DSS-induced ultrastructural abnormalities, suggesting a protective effect on ER homeostasis in colonic epithelial cells (Fig. 5A).Western blotting showed that DSS significantly upregulated GRP78, increased PERK and eIF2α phosphorylation, and elevated CHOP and Caspase-3 levels, indicating activation of signaling related to ER stress and cell damage linked to PERK/eIF2α/CHOP in colitis induced by DSS was examined. While M0-MVs showed minimal impact, M1-MVs notably elevated levels of GRP78, p-PERK/PERK, p-eIF2α/eIF2α, CHOP, and cleaved caspase-3. In contrast, M2-MVs significantly inhibited these elevations (Fig. 5B–G). These findings suggest that M2-MVs suppress DSS-induced ER stress responses in colonic tissues, whereas M1-MVs may further promote stress signaling activation. Subsequently, inflammatory cytokine expression was detected by qRT-PCR. The DSS group exhibited a notable increase in mRNA levels for IL-1β, TNF-α, and IL-6, whereas the expression of IL-10 experienced a reduction. M1-MVs further promoted pro-inflammatory cytokine expression and suppressed IL-10. Conversely, the expression levels of IL-1β, TNF-α, and IL-6 were reduced by M2-MVs, while the amounts of IL-10 were elevated at the same time (Fig. 5H–K). These findings indicate that the protective role of M2-MVs in DSS-induced colitis could be associated with the inhibition of pro-inflammatory responses and the endoplasmic reticulum stress pathway involving PERK/eIF2α/CHOP.

Fig. 5.

Effects of MVs derived from macrophages with different polarization states on ER stress and inflammatory responses in colonic tissues of DSS-treated mice. A: Transmission electron microscopy reveals ultrastructural alterations in the colonic epithelial cells across all groups. The red arrows highlight the endoplasmic reticulum, while the yellow arrows point to mitochondria, the blue arrows show autolysosomes, green arrows indicate tight junctions, and purple arrows mark microvilli. B–G: The assessment of GRP78, p-PERK/PERK, p-eIF2α/eIF2α, CHOP, and cleaved caspase-3 levels in the colon was conducted using Western blotting coupled with densitometric analysis. H–K: mRNA expression levels of colonic IL-1β, TNF-α, IL-6, and IL-10 were evaluated through RT-qPCR. Results are presented as mean ± SD with n = 3. The figure indicates statistical significance

M2-MVs restore barrier protein expression and suppress ER stress in inflammation-stimulated NCM460 cells

To evaluate the immediate regulatory effects of microvesicles (MVs) on intestinal epithelial cells, an inflammatory injury model in vitro was established by exposing NCM460 cells to CK, which was then followed by treatment with M0-MVs, M1-MVs, or M2-MVs. Western blot analysis indicated that stimulation of CK resulted in a notable reduction in the protein levels of Claudin-1, Occludin, and ZO-1 in comparison to the Control group, implying a direct disturbance of protein homeostasis associated with the barrier in colonic epithelial cells. M0-MVs showed a certain tendency to restore barrier protein expression, but the effect was limited. M1-derived MVs resulted in a reduction of Claudin-1, Occludin, and ZO-1 expression. In contrast, M2-derived MVs significantly elevated the amounts of these tight junction proteins (Fig. 6A–D). These results are consistent with the in vivo findings and indicate that M2-MVs can directly act on intestinal epithelial cells to ameliorate inflammation-induced barrier protein loss. Detection of ER stress-related proteins showed that CK stimulation significantly increased the expression of GRP78, p-PERK/PERK, p-eIF2α/eIF2α, CHOP, and cleaved caspase-3, suggesting that the inflammatory microenvironment induced marked ER stress in NCM460 cells. M1-MVs treatment further enhanced the expression of these stress-related proteins, whereas M2-MVs markedly inhibited CK-induced GRP78 upregulation, PERK/eIF2α phosphorylation, and increased CHOP and cleaved caspase-3 expression (Fig. 6E–J). The qRT-PCR results showed that IL-1β, TNF-α, and IL-6 mRNA levels were markedly increased after CK stimulation, whereas IL-10 expression was decreased. M1-MVs further amplified pro-inflammatory cytokine transcription, while M2-MVs significantly suppressed IL-1β, TNF-α, and IL-6 expression and upregulated IL-10 levels (Fig. 6K–N). These in vitro findings are highly consistent with the results obtained from DSS-treated mice, further demonstrating that M2-MVs directly alleviate barrier injury and ER stress responses in colonic epithelial cells under inflammatory stimulation. Taken together, the in vivo and in vitro results indicate that M2 macrophage-derived MVs may maintain intestinal epithelial barrier homeostasis by inhibiting PERK/eIF2α/CHOP-mediated ER stress and inflammatory cytokine expression, thereby exerting protective effects against DSS-induced colitis. In contrast, M1-MVs tend to aggravate inflammatory responses, barrier disruption, and ER stress activation.

Fig. 6.

Effects of MVs derived from macrophages with different polarization states on barrier proteins, ER stress, and inflammatory cytokine expression in inflammation-stimulated NCM460 cells. A–D: The protein expression levels of Claudin-1, Occludin, and ZO-1 were analyzed in NCM460 cells through Western blotting and densitometric quantification across several groups: Control, CK, CK + M0-MVs, CK + M1-MVs, and CK + M2-MVs. E–J: Western blotting and densitometric analysis were utilized to quantify the expression levels of GRP78, p-PERK/PERK, p-eIF2α/eIF2α, CHOP, and cleaved caspase-3. K–N: RT-qPCR was conducted to evaluate the mRNA expression levels of IL-1β, TNF-α, IL-6, and IL-10. In this context, CK represents stimulation with IFN-γ, TNF-α, and IL-1β. The outcomes are shown as mean ± SD, with a sample size of n = 3. The figure indicates statistical significance

4-PBA-mediated inhibition of ER stress confirms the involvement of ER stress in MVs-mediated regulation of intestinal epithelial cell injury

To elucidate the role of ER stress in the modulation of intestinal epithelial cell damage induced by MVs, the ER stress inhibitor 4-PBA was administered to the CK-triggered inflammatory injury model involving NCM460 cells, allowing for further observation of its influence on responses mediated by M1-MVs and M2-MVs. Flow cytometry analysis with Annexin V/PI double staining revealed a significant increase in the apoptosis rate of NCM460 cells following CK stimulation, while treatment with 4-PBA significantly mitigated CK-induced apoptosis. Compared with the CK group, M1-MVs further increased apoptosis, while combined treatment with 4-PBA markedly alleviated the M1-MVs-induced increase in apoptosis. In contrast, M2-MVs reduced apoptosis levels under CK stimulation, and the apoptosis rate remained at a relatively low level after combined treatment with M2-MVs and 4-PBA (Fig. 7A–H).Western blot analysis showed that CK stimulation activated PERK/eIF2α/CHOP-related ER stress and apoptotic signaling. M1-MVs further enhanced these changes, whereas combined treatment with 4-PBA markedly attenuated M1-MVs-induced upregulation of ER stress- and apoptosis-related proteins. M2-MVs showed an inhibitory trend similar to that of 4-PBA, and the related indicators generally remained at low levels after combined treatment (Fig. 7I–N). These results suggest that inhibition of ER stress can reduce CK- and M1-MVs-induced apoptosis in intestinal epithelial cells and is consistent with the protective changes induced by M2-MVs. Therefore, PERK/eIF2α/CHOP-related ER stress may represent an important mechanistic link through which MVs regulate intestinal epithelial cell injury and cell fate under inflammatory conditions.

Fig. 7.

Effects of 4-PBA on M1-MVs and M2-MVs mediated regulation of apoptosis and ER stress-related protein expression in inflammation-stimulated NCM460 cells. A–H: Annexin V-FITC/PI double-staining flow cytometry analysis of apoptosis in NCM460 cells from the Control, CK, CK + 4-PBA, CK + M1-MVs, CK + M1-MVs + 4-PBA, CK + M2-MVs, and CK + M2-MVs + 4-PBA groups, together with quantitative analysis of the total apoptosis rate. I–N: In every group, the protein expression levels of GRP78, p-PERK, PERK, p-eIF2α, eIF2α, CHOP, and cleaved caspase-3 proteins were evaluated via Western blotting and measured using densitometric techniques. Results are expressed as mean ± standard deviation, where n = 3. The figure illustrates significant statistical differences

Discussion

As a long-lasting and recurring inflammatory bowel condition, UC has demonstrated a steadily rising incidence worldwide [24]. Current therapies still face substantial challenges in achieving durable mucosal healing and preventing complications [7, 25]. In the present study, we systematically compared MVs derived from M0, M1, and M2 macrophages using parallel in vivo and in vitro models. The principal finding is a polarization-dependent bidirectional effect: M2-MVs alleviated DSS-induced colitis, restored the intestinal epithelial barrier, and suppressed PERK/eIF2α/CHOP-associated ER stress, whereas M1-MVs amplified inflammatory injury and stress signaling. This parallel positive–negative comparison extends the conventional focus on macrophage phenotypes themselves and identifies their shed MVs as functionally distinct paracrine mediators of macrophage–epithelial communication.

The preservation of the intestinal epithelial barrier’s integrity is crucial for sustaining intestinal homeostasis [26]. The downregulation of tight junction proteins such as Claudin-1, Occludin, and ZO-1, induced by DSS, along with the disruption of the MUC2-mediated mucus barrier, may result in the leakage of luminal antigens and the exacerbation of inflammatory responses [27–29]. In this study, M2-MVs significantly upregulated the expression of these barrier markers, whereas M1-MVs further aggravated their disruption. Previous studies have shown that macrophage-derived EVs can act as intercellular communication carriers. By delivering miRNAs, proteins, and other bioactive molecules, they regulate the functions of recipient cells and thereby participate in inflammatory responses, macrophage polarization, and tissue repair. For example, macrophage-derived microvesicles can mediate the transfer of miR-223 and influence monocyte/macrophage differentiation, while lactadherin-enriched macrophage-derived sEVs have also been shown to alleviate intestinal inflammation by reprogramming macrophage polarization [30–32]. In line with these results, the current research additionally illustrated, focusing on the MVs subgroup, that MVs originating from M1 and M2 macrophages have different impacts on the intestinal epithelial barrier. Notably, this polarization-dependent bidirectional effect indicates that macrophage-derived MVs may represent an important class of paracrine regulatory mediators in the inflammatory microenvironment of UC.

Aberrant activation of ER stress is a key link connecting inflammatory responses, epithelial cell dysfunction, and barrier disruption [33]. Sustained activation of the PERK/eIF2α/CHOP pathway can induce CHOP upregulation and cleaved caspase-3-mediated apoptosis, thereby promoting epithelial cell death and barrier dysfunction [34, 35]. Previous studies have observed alterations in ER stress-related markers, such as GRP78 and XBP1, in intestinal tissues from patients with UC, suggesting that ER stress is associated with susceptibility to intestinal inflammation and disease progression [19, 36]. In this study, both the in vivo DSS model and the in vitro inflammatory cell model demonstrated that M2-MVs substantially reduced the phosphorylation cascade of this pathway, while M1-MVs intensified this response. These findings indicate that both inflammatory models were accompanied by marked ER stress and cell injury signaling, and suggest that MVs may participate in barrier injury or protection by regulating ER stress status in epithelial cells.Inflammation and ER stress may mutually reinforce each other. On one side, cytokines that promote inflammation, including TNF-α, IL-1β, and IL-6, have the capability to enhance ER stress by triggering oxidative stress, disrupting calcium homeostasis, and increasing the burden of protein folding. On the other hand, excessive ER stress can further enhance inflammatory cytokine release and cell death, thereby aggravating barrier injury [37–39]. The uptake or delivery of bioactive cargo by M2-MVs could then lead to the downregulation of ER stress and apoptotic proteins observed in the present study. In addition, they reduced the levels of IL-1β, TNF- α, IL-6, and augmented the production of IL-10. But, M1-MVs showed an opposite effect. Taken together, it seems that the protective effect of M2-MVs is not a single-factor process but likely involves crosstalk among inflammation attenuation, ER stress alleviation, and restoration of barrier proteins. As a chemical chaperone, 4-PBA can reduce protein misfolding and ER burden [40]. In this research, the suppression of endoplasmic reticulum (ER) stress diminished apoptosis induced by CK and M1-derived microvesicles (MVs) and lowered the levels of proteins associated with ER stress. M2-MVs and 4-PBA showed similar regulatory trends in most indicators, further supporting the involvement of PERK/eIF2α/CHOP-related ER stress in MVs-mediated regulation of intestinal epithelial cell injury.

From the perspective of existing research, EVs have attracted increasing attention in UC. EVs from different sources can influence intestinal inflammation by regulating immune cell activation, epithelial cell repair, and inflammatory signal transduction [15, 41, 42]. The findings of the present study are generally consistent with these observations and further suggest that macrophage polarization status may influence the functional direction of macrophage-derived MVs. Compared with broadly discussing the paracrine effects of macrophages, distinguishing between M1-MVs and M2-MVs may help provide a more detailed understanding of the complexity of macrophage–epithelial cell communication under different inflammatory conditions. In addition, this study revealed a potential association between MVs and ER stress, suggesting that MVs, as intercellular information carriers, may participate in the regulation of stress homeostasis in intestinal epithelial cells. This finding supplements previous studies on ER stress regulation in UC, although the precise underlying mechanisms require further investigation.

This study has several limitations. The biodistribution, pharmacokinetics, dose–response characteristics, batch-to-batch consistency, and long-term safety of macrophage-derived MVs were not evaluated. Future studies should validate these findings in larger animal cohorts and primary cell or organoid models, identify the key functional cargo of MVs, and further standardize their preparation and characterization.

From a translational perspective, M2-MVs have potential as cell-free effector carriers. Compared with live-cell therapy, MVs, as medium-to-large EVs, may offer certain advantages in storage, delivery, and engineering modification, while potentially reducing uncertainties related to cell survival, phenotypic drift, and safety [43]. Several studies support this potential. For example, HucMSC-derived EVs have been shown to alleviate DSS-induced colitis by promoting M2 macrophage polarization [44], whereas engineered macrophage-derived EVs have been reported to regulate inflammation and epithelial regeneration in studies of atherosclerosis and diabetic wound healing [45, 46]. However, considerable challenges remain before clinical application, including clarification of the effective components of MVs, biodistribution, targeted delivery, and long-term safety [47, 48]. Taken together, these findings suggest that M2-MVs may have potential value in the modulation of UC-related mucosal injury by attenuating inflammatory responses, reducing PERK/eIF2α/CHOP-associated ER stress, and supporting intestinal epithelial barrier repair. Further studies are required to determine their translational applicability.

Conclusion

This research demonstrates that the polarization-dependent functional heterogeneity of macrophage-derived MVs is a key determinant of their regulatory effects on UC-like intestinal inflammation. In DSS-induced acute colitis, M2-MVs reduced disease severity, histopathological injury, and epithelial barrier disruption, whereas M1-MVs tended to aggravate inflammatory damage. M2-MVs also suppressed inflammatory responses and PERK/eIF2α/CHOP-associated ER stress, preserved tight-junction protein expression, and reduced epithelial apoptosis. Consistent in vitro findings, together with the 4-PBA rescue experiment, support the involvement of ER stress as an important downstream mechanism in MV-mediated epithelial injury. Collectively, these findings reveal a polarization-dependent regulatory pattern of macrophage-derived MVs in experimental colitis and identify M2-MVs as promising cell-free therapeutic effectors for intestinal mucosal repair. This work provides experimental evidence for MVs mediated macrophage–epithelial communication and offers a theoretical basis for developing EVs based therapeutic strategies targeting barrier dysfunction, inflammation, and ER stress in ulcerative colitis.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

C.L., Q.K., and Y.W. conceptualized the study, conducted the investigation, and wrote the original draft of the manuscript. S.W. contributed to conceptualization, investigation, and data curation. H.W. and Y.J. developed the methodology and performed the formal analysis. Y.X. and C.Q. supervised the study, acquired funding, provided resources, and administered the project. All authors reviewed and approved the final manuscript.

Funding

This work was financially supported by the Science and Technology Program of XPCC (2023ZD025), the Shihezi University Postdoctoral Research Project (336483), the Health Project under the Third Batch of the “2 + 5” Key Talent Program (TSYC202401B137), and the First Affiliated Hospital of Shihezi University Doctoral Fund (BS202104).

Data availability

All data supporting the findings of this study are available within the paper.

Declarations

Ethical approval

Animal experiments were carried out in accordance with the National Institutes of Health guidelines for the care and use of laboratory animals. All procedures were reviewed and authorized by the Animal Research Ethics Committee of Shihezi University (Approval No. A2024-748).

Competing interests

The authors declare no competing interests.