Versatility of bacterial outer membrane vesicles in regulating intestinal homeostasis

Abstract

Outer membrane vesicles (OMVs) play vital roles in bacterial communication both intraspecifically and interspecifically. However, extracellular mechanisms of gut microbiota–derived OMVs in the intestine remain poorly understood. Here, we report that OMVs released from Akkermansia muciniphila are able to (i) restore disturbed balance of the gut microbiota by selectively promoting the proliferation of beneficial bacteria through membrane fusion, (ii) elicit mucosal immunoglobulin A response by translocating into Peyer’s patches and subsequently activating B cells and dendritic cells, and (iii) maintain the integrity of the intestinal barrier by entering intestinal epithelial cells to stimulate the expressions of tight junctions and mucus. We demonstrate that transplantation of gut microbiota–associated OMVs to the intestine can alleviate colitis and enhance anti–programmed cell death protein 1 therapy against colorectal cancer by regulating intestinal homeostasis. This work discloses the importance of gut microbiota–derived OMVs in intestinal ecology, providing an alternative target for disease intervention and treatment.

Bacterial vesicles can regulate the gut microbiota, modulate mucosal immunity, and maintain intestinal barrier integrity.

INTRODUCTION

The gut microbiota has emerged as a pivotal factor in maintaining gut and systemic homeostasis (1). In the gut ecosystem, bidirectional communication between microbes and the host is rarely mediated by direct cell-cell contacts due to the existence of epithelial barriers (2). Numerous studies have demonstrated that bacterial metabolites can mediate microbial interactions with the host (3–5). For instance, butyrate produced by Clostridia has been reported to stimulate colonic regulatory T cells, thereby suppressing host inflammatory responses (4). Gut microbiome–modified bile acid has been proven to be able to regulate the accumulation of natural killer T cells in the liver tumor (5). Undoubtedly, intricate cross-talk between the gut microbiota and its host is multidirectional and reciprocal, implying that the best-characterized derivative secreted from bacteria may be involved in their interplay. Gram-negative bacteria normally release outer membrane vesicles (OMVs) during growth, which are bilayered membrane nanostructures that range from 20 to 400 nm in size and contain multiple parental components, such as nucleic acids, proteins, enzymes, and lipopolysaccharide (LPS) (6). Provided with these characteristics, OMVs potentially play a critical role in microbe-microbe and host-microbe interactions. Recent works in mice have highlighted the roles of OMVs produced by pathogenic and engineered commensal bacteria in eliciting immunostimulatory responses after bypassing host barriers (7, 8). Despite the extension of biological functions beyond transporting active substance and defensing against outer membrane–acting stressors, such as phage predation and the invasion of antibacterial agents, it remains poorly understood whether bacterial OMVs could mediate other host responses than unilateral defense from the immune system (6, 9). It is widely acknowledged that long-term evolution has fostered advanced mutualism of the gut microbiota with its host (10). However, substantial evidence that OMVs can be extracellular mechanisms for effectively regulating gut homeostasis has not been reported yet. Here, we describe the underlying roles of commensal bacteria–derived OMVs in assisting the reversal of gut disorder by comprehensively modulating the gut microbiota, mucosal adaptive immune responses, and the physicochemical barrier (Fig. 1). As a proof-of-concept therapeutic target, we show that transplantation of OMVs instead of living bacteria to mouse gut can rescue colitis and enhance anti–programmed cell death protein 1 (PD-1) immunotherapy against colorectal cancer (CRC) by maintaining intestinal homeostasis.

Fig. 1. Schematic illustration of the versatility of bacterial OMVs in regulating intestinal homeostasis.

OMVs derived from commensal bacteria can restore imbalanced gut microbiome by selectively promoting the proliferation of beneficial bacteria via membrane fusion, trigger mucosal IgA response by translocating into PPs and activating B cells and DCs, and maintain the integrity of the gut barrier by entering intestinal epithelial cells to up-regulate the expressions of tight junctions and mucus. T_reg, regulatory T; T_H, T helper; MHC, major histocompatibility complex; TCR, T cell receptor.

RESULTS

Regulation of the gut microbiota

We chose Akkermansia muciniphila (Akk) strain, which colonizes intestinal mucous layer, as a model commensal bacterium to investigate the multifaceted roles of OMVs in gut ecology. OMVs derived from Akk were collected after bacterial reproduction, followed by purification through a set of centrifugation procedures. Nanoparticle tracking analysis (NTA) indicated that average size of the obtained OMVs was 108.60 ± 48.90 nm (fig. S1). To assess the impact of OMVs on the gut microbiota, dextran sulfate sodium (DSS)–induced mice, a model of gut disorder, were daily administered with Akk OMVs containing 20 μg of total proteins for five consecutive days by oral gavage and colonic contents were extracted for microbial composition analysis via 16S ribosomal RNA gene sequencing. Note that there were no significant alterations in the size and number of OMVs after incubation with simulated gastric fluid (pH 1.2) at 37°C for 4 hours, a period mimicking gastric residence time, indicating a satisfied stability to pass gastric cavity after oral ingestion (fig. S2). As shown in Fig. 2A (top), there was no sharp rise in rarefaction curves, implying that the amount of the sample sequencing was adequate for data analysis and operational taxonomic unit (OTU) number of OMV-treated mice was significantly larger than that of phosphate-buffered saline (PBS)–treated group. Shannon index of OMV-treated mice also increased greatly and was comparable to that of healthy control group (Fig. 2A, bottom). Overall, OMV-treated group exhibited marked improvements in both richness and diversity of the gut microbiota compared to PBS-treated mice. In contrast to PBS, Akk OMVs increased the relative abundances of several probiotic or commensal bacterial genera, including Bacteroides, Lactobacillus, and Alistipes, along with Lachnospiraceae NK4A136 group and bacterium f Lachnospiraceae (Fig. 2B) (11–14). In particular, beneficial members of the genus Bacteroides in OMV-treated group were significantly up-regulated (Fig. 2, C to E, and fig. S3A). The relative abundances of Bacteroides acidifaciens and Bacteroides thetaiotaomicron were separately increased from 0.47 ± 0.16% to 1.99 ± 0.44% and 0.02 ± 0.02% to 0.64 ± 0.11% (Fig. 2, D and E). B. acidifaciens has been reported to improve mucosal immune functions by switching immunoglobulin M (IgM) to IgA and promoting intestinal IgA production, while B. thetaiotaomicron can maintain intestinal barrier functions and ameliorate colonic inflammation by stimulating the productions of mucus and IgA and reversing epithelial damages (15–17). OMVs also elevated the relative abundances of Bacteroides stercorirosoris and Bacteroides caecimuris, which are associated with the alleviation of ulcerative colitis and colitis-associated tumorigenesis, from 0.14 ± 0.14% to 2.21 ± 0.54% and 0.10 ± 0.01% to 0.45 ± 0.13%, respectively (fig. S3A) (18, 19). However, different from above Bacteroides species, the relative abundance of Bacteroides vulgatus, with capacity to induce protease activity–dependent gut barrier dysfunction and ulcerative colitis, was declined (Fig. 2C) (20). In addition, OMVs reduced the relative abundances of bacteria belonging to the phylum Proteobacteria, which is the largest phylum containing numerous pathogenic bacteria and regarded as a microbial signature of dysbiosis in the gut microbiota (Fig. 2B and fig. S3B) (21). These results suggested that Akk OMVs could robustly reverse gut microbial imbalance by up-regulating beneficial bacteria, especially Bacteroides species, while reducing opportunistic pathogens.

Fig. 2. Regulation of the gut microbiota.

(A to E) 16S ribosomal RNA gene sequencing analysis of the gut microbiota. DSS mice were daily gavaged with 100 μl of Akk OMV suspension containing 20 μg of total proteins for 5 days and euthanized for sampling (n = 5). Healthy mice and DSS mice treated with PBS were separately used as controls. (A) Rarefaction curves (top) and Shannon curves (bottom). Species distribution of (B) bacterial genus and (C) Bacteroides species. Relative abundances of (D) B. acidifaciens and (E) B. thetaiotaomicron. (F and G) Interactions between Akk OMVs and Bacteroides species by directly incubating 1 ml of culture medium containing 50 μl of logarithmic phase bacterial solution and 100 μl of Akk OMV suspension at 37°C. Bacteria incubated with PBS were used as a control. (F) 3D CLSM images and flow cytometry histograms of Bacteroides after incubation for the indicated time points. Scale bars, 25 μm. (G) Growth curves measured by recording the OD value at 600 nm (n = 3). Data are means ± SEM. Significance was assessed using one-way or two-way ANOVA test, giving P values: *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001.

These findings encouraged us to investigate potential underlying mechanism by asking whether Akk OMVs could selectively interact with specific bacterial species among the gut microbiota. We covalently labeled the outer membrane of OMVs with Cy5.5 (fig. S4) and stained the internal nucleic acids with Hoechst. Representative strains of the genus Bacteroides, including B. acidifaciens, B. thetaiotaomicron, B. vulgatus, and B. fragilis, were chosen for verification. Bacteroides in logarithmic phase were incubated with labeled OMVs, and cells were collected at the indicated time points for analysis. Confocal laser scanning microscopy (CLSM) images displayed that Akk OMVs were fused with all these species of Bacteroides, despite that the uptake level varied from species (Fig. 2F and figs. S5 to S8). After coincubation for 1.5 hours, B. acidifaciens and B. fragilis were strongly stained with Cy5.5, while B. thetaiotaomicron and B. vulgatus were observed with relatively lowered fluorescence signals. Fluorescence intensity of Cy5.5 in these proliferating Bacteroides increased with coincubation time, suggesting continuous uptake of Akk OMVs. Three-dimensional (3D) CLSM images showed that after coincubation, these Bacteroides appeared Cy5.5-labeled cell membranes with intracellular Hoechst fluorescence signals, indicating favorable fusion of Akk OMVs with Bacteroides. To examine the specificity of OMV fusion, we chose a typical probiotic strain of Escherichia coli Nissle 1917 and a common pathogenic bacterium of Salmonella typhimurium SL1344 to study the uptake of Akk OMVs. Unexpectedly, there was no fluorescence signal observed for S. typhimurium even after 25-hour coincubation, while weak signals were observed in a few E. coli (figs. S9 and S10), supporting that Akk OMVs were able to preferentially fuse with specific genus. In good agreement with the results of CLSM imaging, flow cytometric analysis further quantitatively confirmed that, in contrast to S. typhimurium, these commensal bacteria fused with significantly much more Akk OMVs (Fig. 2F, right, and fig. S11). We further explored the effects of OMV fusion on bacterial proliferation by measuring corresponding growth curves by recording the optical density (OD) value at 600 nm. As shown in Fig. 2G, Akk OMVs markedly promoted the growths of B. acidifaciens, B. thetaiotaomicron, and B. fragilis, but had no benefit on B. vulgatus, implying their selective stimulation on the proliferation of specific strains in the gut microbiota that could be explained by effective fusion with Akk OMVs. Note that despite the presence of similar OMV fusion level, no beneficial effect occurred to the growth of B. vulgatus, pointing out the ineffectiveness of Akk OMVs to motivate the proliferation of potentially opportunistic Bacteroides species in DSS-induced gut disorder. As expected, negligible growth was found to E. coli and S. typhimurium after coincubation with Akk OMVs due to the absent or inefficient OMV fusion (fig. S12). In addition, OMVs released from probiotic E. coli and B. thetaiotaomicron showed limited effects on the proliferation of B. acidifaciens and B. vulgatus, further suggesting the specificity of Akk OMVs toward Bacteroides (fig. S13). Collectively, these data demonstrated that one of the mechanisms of Akk OMVs in restoring a balanced gut microbial structure was direct up-regulation of the proliferation of beneficial species by selective fusion.

Modulation of mucosal immune responses

During in vivo studies, Cy5.5-labeled Akk OMVs entered Peyer’s patches (PPs) 2 hours after direct delivery into the intestinal lumen and migrated to mesenteric lymph nodes (MLNs) with post-administration time extending to 4 hours (Fig. 3A and fig. S14). This observation prompted us to investigate whether OMVs could elicit mucosal immunomodulatory responses to regulate gut disorder. Given that the gut microbiota can affect IgA production and primary IgA production occurs mainly in PPs, we speculated that the entry of Akk OMVs might induce IgA responses, which is essential for the intestinal immune barrier to prevent pathogen invasion (22–24). For the analysis of mucosal immune responses, PPs and MLNs were collected from the intestinal tract of mice induced with gut disorder after oral gavage of Akk OMVs for five consecutive days. We first analyzed the role of Akk OMVs on the activation of dendritic cells (DCs) as the interaction of B cells with subepithelial DCs in PPs is required for IgA production (22). As expected, the expression of CD80 on DCs was significantly increased in the PPs of OMV-dosed mice (Fig. 3C). In vitro coculture with bone marrow–derived DCs (BMDCs) further confirmed that Akk OMVs were able to induce higher production of CD80⁺ DCs following internalization (Fig. 3, B and D). Subsequently, we analyzed the activation of B cells (B220⁺CD138⁻) and the production of plasma cells (B220⁻CD138⁺) in the sampled PPs. OMV-treated group significantly augmented the productions of CD69⁺ B cells and IgA⁺ plasma cells along with total B cells, thereby increasing intestinal IgA concentration from 18.71 ± 1.52 to 29.47 ± 2.73 mg/ml (Fig. 3, E to G). In addition to the direct influence of Akk OMVs on mucosal IgA responses, the improved gut microbial structure might stimulate IgA production by switching IgM to IgA by up-regulating the expressions of activation-induced cytidine deaminase (i.e., B. acidifaciens) and polymeric immunoglobulin receptor, which can transport IgA across the epithelium (i.e., B. thetaiotaomicron) (15, 16). On the other hand, IgA in the intestinal lumen can neutralize and exclude pathogenic species via antibody coating (23, 24). The counts of IgA⁺ fecal bacteria in OMV-treated group were significantly increased due to the enhanced production of IgA (Fig. 3H). This could be another reason that Akk OMVs reduced the relative abundance of harmful pathogens in the gut microbiota. In addition, gut commensal species can use IgA for mucosal colonization, implying the robustness of host-microbial symbiosis to recover a healthy microbial composition (25). Consequently, Akk OMVs played an essential and bidirectional role in intricate interactions between mucosal IgA responses and the gut microbiota, which could regulate the dysregulation of intestinal homeostasis.

Fig. 3. Modulation of mucosal immune responses.

(A) IVIS images of mouse PPs (top) and MLNs (bottom) sampled at 2 and 4 hours after enteral injection of 25 μl of Cy5.5-labeled Akk OMVs (0.1 mg/ml) or PBS in PP-located intestinal segments that were tied at both ends. (B) CLSM images of BMDCs after incubation with Cy5.5-labeled Akk OMVs (0.1 mg/ml) or PBS at 37°C for 24 hours. Scale bars, 10 μm. (C) Percentage of CD80⁺ DCs in the PPs of mice induced with gut disorder after daily oral gavage of 100 μl of Akk OMV suspension containing 20 μg of total proteins or PBS for 5 days (n ≥ 5). (D) Flow cytometric analysis (left) and CD80⁺ percentage (right) of BMDCs after incubation with Cy5.5-labeled Akk OMVs (0.1 mg/ml) or PBS at 37°C for 24 hours (n = 3). (E to I) Mucosal immune responses in the intestine of mice induced with gut disorder after daily oral gavage of 100 μl of Akk OMV suspension containing 20 μg of total proteins or PBS for 5 days (n ≥ 5). (E) Representative flow cytometry scatterplots of B cells (B220⁺CD138⁻) and plasma cells (B220⁻CD138⁺) in PPs (left) along with percentage of B cells (right). (F) Quantitative analysis of CD69⁺ B cells and IgA⁺ plasma cells in PPs. (G) IgA concentration in the intestine. (H) Flow cytometry histograms and quantification of IgA⁺ fecal bacteria. The naked bacterium was used as a control. (I) Quantitative analysis of IFN-γ⁺CD4⁺ T cells, IL-4⁺CD4⁺ T cells, IL-17⁺CD4⁺ T cells, and ratios of IL-4⁺/IFN-γ⁺CD4⁺ T cells and FOXP3⁺CD25⁺/IL-17⁺CD4⁺ T cells in MLNs. MFI, mean fluorescence intensity. Data are means ± SEM. Significance was assessed using t test, giving P values: *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001.

On the basis of these findings, we reasoned that Akk OMVs might simultaneously regulate the cellular immune responses in MLNs considering the migration through lymphatic channels. Previous studies have proven that mucosal immune cells constitutively probe gut environment, e.g., DCs capture antigens to coordinate T cell responses (26). The increased level of CD80⁺ DCs in PPs inspired us to investigate T cell reactions in MLNs. Flow cytometric analysis revealed increased ratios of interleukin-4 (IL-4)⁺/interferon-γ (IFN-γ)⁺CD4⁺ T cells and FOXP3⁺CD25⁺/IL-17⁺CD4⁺ T cells in MLNs collected from mice treated with Akk OMVs (Fig. 3I). The polarization toward anti-inflammatory phenotypes exhibited by T cells along with relatively low levels of immune responses in MLNs indicated that Akk OMVs could assist the host to maintain mucosal immune homeostasis (Fig. 3I and fig. S15). The variation of the adaptive immune responses in vivo could be ascribed to the evolved symbiotic relationship between the gut microbiota and its host in the presence of complex intestinal environment, although in vitro study of coculture BMDCs with splenic CD4⁺ T cells displayed an immune activation effect elicited directly by Akk OMVs (fig. S16) (27, 28). Mechanistically, intestinal pathogens that could cause cytokine storms were greatly reduced via OMV-assisted restoration of gut biological barrier and induction of mucosal IgA responses, which ultimately maintained host immune homeostasis.

Maintenance of the gut physicochemical barrier

We next studied the impact of Akk OMVs on shaping the physicochemical barrier. Mucus covering intestinal epithelial surface is the first protective line for fortifying gut barrier integrity, which is majorly modulated by the gut microbiota (29, 30). Goblet cells are predominant epithelial cells secreting mucus in the gastrointestinal (GI) tract, and their distribution reaches a maximum density in the colon (29, 30). The major form of O-glycosylated entity of Muc2, the best-characterized secreted mucin, was mainly produced by goblet cells distributed in the proximal colon of mice (29). Accordingly, we harvested proximal colon tissue from treated mice and found that Alcian blue–stained samples depicted marked increases of goblet cells that were fully filled with acidic mucus in the proximal colon of OMV-treated mice. Compared to individually scattered goblet cells with pathological cavities in PBS-administered gut disorder mice, the goblet cells in the proximal colon of OMV-treated mice were appreciably improved in both quantity and functionality, leading to the formation of a thicker mucus layer in the intestinal lumen (Fig. 4A and fig. S17). Moreover, the percentage of goblet cells was largely increased from 5.11 ± 2.34 to 19.89 ± 3.57 in the intestinal epithelium (Fig. 4B). The considerable efficacy of Akk OMVs to repair the mucus barrier could be attributed to the internalization of OMVs followed by the stimulation of goblet cells to produce mucus (fig. S18) and OMV-mediated increment in the abundance of beneficial bacterial species (i.e., B. thetaiotaomicron) responsible for mucus production (16, 17, 31).

Fig. 4. Maintenance of the gut physicochemical barrier.

(A to D) In vivo effects on repairing the gut physicochemical barrier of DSS mice after daily oral delivery of 100 μl of Akk OMV suspension containing 20 μg of total proteins or PBS for 5 days. (A) Alcian blue staining images of the proximal colon. Black and yellow arrows represent goblet cells with pathological cavities and the thickened mucus layer, respectively. Scale bars, 600 μm. (B) Percentage of goblet cells in the intestinal epithelium. (C) 3D CLSM images of the intestinal epithelial cells. Red fluorescence indicates Cy5.5-labeled Akk OMVs. Scale bars, 50 μm. (D) Immunofluorescence images of tight junctions ZO-1 and occludin expressed on the epithelial cells in the proximal colon. Scale bars, 20 μm. (E and F) In vitro effects on Caco-2 cells after treatment with Akk OMVs (0.1 mg/ml) along with LPS (5 μg/ml) at 37°C for 24 hours. Cy5.5-labeled OMVs were used for CLSM. (E) Immunofluorescence and (F) CLSM images. Scale bars, 25 μm. (G) Cytokine levels of IL-10, IL-13, and ACE2 in colon tissue together with IFN-γ and CRP in serum measured by ELISA. Samples were collected from DSS mice after treatment with Akk OMVs or PBS. Data are means ± SEM (n ≥ 5). Significance was assessed using t test, giving P values: *P < 0.05.

Different from bacteria with a large body size, Akk OMVs could cross the mucus layer and be internalized by intestinal epithelial cells (Fig. 4C and fig. S19), proposing the direct interaction of OMVs with the physicochemical barrier. The barrier function and gut permeability are closely associated with the tight junction complex, serving as a gatekeeper for the paracellular pathway (32). Occludin, a major component of tight junctions, decreases proportionally to the diminishment of epithelial barrier function, while ZO-1, an essential tight junction–associated protein, can promote epithelial proliferation to benefit mucosal repairment and homeostasis (33, 34). The expressions of occludin and ZO-1 in the epithelial cells in proximal colon were substantially increased, accompanied by the restoration of intestinal epithelial impairment as more overlapping and integral fluorescence signals were observed in the outlines of intestinal epithelium with normal morphology in OMV-treated group (Fig. 4D). In vitro immunofluorescence study of LPS-treated epithelial cell line Caco-2 further showed immediate enhancement of occludin and ZO-1 expressions after cocultivation with Akk OMVs (Fig. 4E). One of the mechanisms that Akk OMVs up-regulated intestinal tight junctions after cell internalization could be the regulation of the reassembly and stability of tight junctions by activating the adenosine 5′-monophosphate–activated protein kinase (AMPK) pathway (Fig. 4F) (35). In addition, OMV-mediated reversal of gut dysbiosis and maintenance of immune homeostasis jointly contributed to substantial up-regulation of anti-inflammatory cytokines including IL-10 and IL-13 as well as down-regulation of proinflammatory cytokines including angiotensin-converting enzyme 2 (ACE2), IFN-γ, and C-reactive protein (CRP), playing critical roles in fortifying the integrity of the gut barrier by regulating the tight junctions (Fig. 4G).

Therapeutic values of transplanting gut microbiota–associated OMVs

Having confirmed the versatility of Akk OMVs in regulating the gut microbiota, mucosal immune responses, and physicochemical barrier, we turned our attention to investigate whether OMV-mediated reversal of dysregulation of intestinal homeostasis could be applied for disease intervention and treatment. Inflammatory bowel disease (IBD), including Crohn’s disease and ulcerative colitis, is a group of autoimmune diseases without a completely clear etiology, still a major obstacle for conventional treatment (36, 37). To examine the therapeutic potential, mice induced with acute colitis by DSS were treated with Akk OMVs through oral administration. As previous studies have demonstrated that Akk can blunt colitis and associated tumorigenesis, here, both Akk- and PBS-treated colitis mice were used as controls (38, 39). After daily oral gavage for five consecutive days, the gained weight of OMV-treated mice was comparable to that of Akk-treated mice, which was markedly more than PBS control (Fig. 5A). Meanwhile, oral delivery of Akk OMVs corrected both the reduction in colon length and the escalation of colonic weight-to-length ratio in colitis mice (Fig. 5B and fig. S20). It has demonstrated that neutrophil infiltration serves as one of IBD features and the concentration of myeloperoxidase (MPO), an intracellular protein of neutrophils, is positively correlated with IBD activity (40). As displayed in Fig. 5C, MPO staining of the sectioned colonic tissues claimed that treatment with OMVs or Akk greatly reduced MPO-positive cell infiltration in corresponding colon tissues compared to PBS. Hematoxylin and eosin (H&E) staining highlighted that leukocyte infiltration–caused pathological symptoms, such as mucosal edema and crypt swelling and destruction in the proximal colon of DSS-induced mice, were appreciably reduced after OMV treatment (Fig. 5D). It has been reported that proinflammatory cytokines, e.g., IFN-γ, augment excessive infiltration of inflammatory cells and mucosal epithelium damage in mouse models of IBD, while anti-inflammatory cytokines, e.g., IL-10, exert a key anti-inflammatory role (41, 42). Consequently, the role of Akk OMVs on up-regulation of beneficial microbiota, avoidance of the bloom of gut opportunistic pathogens, and promotion of polarization toward anti-inflammatory phenotypes of T cells resulted in alleviated recruitment of inflammatory cell infiltration in colon tissues along with inflammatory pathological alterations. Furthermore, OMV-enabled enhancement of tight junctions could promote epithelial proliferation for mucosal repairment (34, 43). These results evidenced that oral administration of Akk OMVs was capable to ameliorate DSS-induced colitis in both symptomatic and pathological manifestations, potentially benefited from intestinal homeostasis constituted by OMV-mediated reduction of pathogen abundance, maintenance of immune homeostasis, and enhancement of tight junctions.

Fig. 5. Therapeutic values in ameliorating DSS-induced colitis.

Colitis mice induced by DSS were daily ingested with 100 μl of Akk OMV suspension containing 20 μg of total proteins for 5 days and euthanized for efficacy assessment. DSS mice treated with 10⁸ CFUs of Akk and PBS were used as controls, respectively. (A) Body weight fluctuation during treatment. (B) Average colon length after treatment (left) and digital photos of colon tissues resected from the cecum to the rectum (right). Scale bars, 1 cm. (C) Representative MPO staining images of the proximal colon (left) and quantitative analysis of MPO-positive cells (right). Black arrows represent MPO-positive cells. Scale bars, 100 μm. (D) Typical H&E staining images of the proximal colon. Green, yellow, and blue arrows represent inflammatory cell infiltration, mucosal edema, and crypt swelling and destruction, respectively. Scale bars, 625 μm. Data are means ± SEM (n = 5). Significance was assessed using one-way or two-way ANOVA test, giving P values: *P < 0.05 and ****P < 0.0001.

As one of the most widely used immune checkpoint inhibitors, anti– PD-1/its ligand PD-L1 monoclonal antibodies (mAbs) competitively bind to inhibitory receptors or their ligands on immune cells or tumor cells to unleash antitumor immune responses (44). However, most CRC patients are observed with primary resistance to PD-1 blockade (45). Recent clinical studies have revealed the relevance between the composition of the gut microbiota and tumor responses to PD-1–based immunotherapy (46). Several species, such as Akk and Enterococcus hirae 13144, have been reported to play vital roles in improving the efficacy of anti–PD-1 therapy (47). In particular, the commensal taxa (including B. acidifaciens, Alistipes spp., Clostridium spp., Ruminococcus spp., and Lachnospiraceae spp.) enriched by oral delivered Akk OMVs are abundant in PD-1–based immunotherapy responders or cancer patients, with progression-free survival longer than 3 months (46–49). We hence explored the possibility that beneficial regulation of the gut microbiome by Akk OMVs could enhance the efficacy of immunotherapy targeting PD-1 in a murine model of CRC. Mice were daily fed with Akk OMVs by oral gavage for two consecutive days and then received three doses of intraperitoneal injection of PD-1 mAbs (aPD-1) that were administered every 3 days. Notably, with the help of OMVs, aPD-1 achieved the most potent suppression of tumor growth among all the treated groups (Fig. 6A and fig. S21A). Immunological changes analyzed by flow cytometry and immunofluorescence imaging exhibited significant up-regulations of PD-L1 in both tumor cells (CD45⁻) and immune cells (CD45⁺) from mice cotreated with aPD-1 and OMVs, which were comparable to that of Akk-treated mice and in accordance with previous conclusions obtained from avatar mice that received fecal microbiota transplantation (FMT) from aPD-1 responders (Fig. 6, B and C, and fig. S21B) (46). These results pointed out that Akk OMVs reversed the initially faint efficacy of PD-1 blockade in CRC-bearing mice. Moreover, OMV combination with aPD-1 expanded the infiltration of CD8⁺ cytotoxic T lymphocytes, which improved the response of tumor to PD-1 blockade (Fig. 6D) (50). H&E staining images also clarified the expansion of necrotic regions in tumor tissue after treatment with OMVs and aPD-1, illustrating the enhanced efficacy against tumor growth (Fig. 6E and fig. S21C). Namely, Akk OMVs overcame primary resistance to PD-1 blockade in tumor immunotherapy by maintaining a beneficial gut microbial structure.

Fig. 6. Therapeutic values in blunting CRC.

Mice bearing CRC were daily dosed with 100 μl of Akk OMV suspension containing 20 μg of total proteins for 2 days before three doses of aPD-1 by intraperitoneal injection at a 3-day interval. Mice bearing CRC treated with 10⁸ CFUs of Akk and PBS under the same experimental condition were used as controls, respectively. (A) Individual tumor growth curves. (B) Quantitative analysis of PD-L1⁺ cells, PD-L1⁺CD45⁻ cells, and PD-L1⁺CD45⁺ cells in tumor. (C) Immunofluorescence images of PD-L1 and CD45 infiltrations in tumor beds. Scale bars, 50 μm. (D) Representative immunofluorescence images of CD3 and CD8 infiltrations in tumor beds. Scale bars, 50 μm. (E) Typical H&E staining images of tumor tissue. Scale bars, 300 mm. Data are means ± SEM (n ≥ 5). Significance was assessed using one-way or two-way ANOVA test, giving P values: *P < 0.05 and **P < 0.01.

DISCUSSION

In this study, taking Akk OMVs as an example, we investigated the underlying mechanisms of multiple roles of gut microbiota–derived OMVs in reversing gut disorders by regulating microbe-microbe and host-microbe interactions. We found that Akk OMVs significantly improved the richness and diversity of the gut microbiota and modulated its dysbiosis by increasing the abundance of beneficial species and suppressing the expansion of opportunistic pathogens. We also observed that Akk OMVs activated mucosal IgA responses to maintain immune homeostasis and eliminate hyperinflammatory responses, which in turn expelled pathogenic bacteria. In addition, we disclosed that these vesicles restored the intestinal physicochemical barrier by stimulating global cells to secrete mucus and activating epithelial cells to upgrade the expression of tight junctions. In light of the versatility in manipulating the multifaceted interplay between the gut microbiota, mucosal immune responses, and the physicochemical barrier, Akk OMVs were successfully used to intervene colitis and PD-1–based immunotherapy of CRC in mice by maintaining gut homeostasis.

Decades of studies on the responsibility of natural OMVs in the gut microbiota mainly focus on substance transfer between two bacterial species and passively protective function of pathogen-derived OMVs against extracellular stressors (6, 9). However, the role of commensal OMVs in intestinal microbial ecology is much more than that, especially in selective and contextual modulation of microbiota compositions. This study uncovered that Akk OMVs could improve the abundance of beneficial commensal taxa belonging to the phylum Firmicutes (i.e., Lachnospiraceae spp., Lactobacillus spp., and Ruminococcaceae spp.) and the phylum Bacteroidete (i.e., Bacteroides spp. and Alistipes spp.) but reduced potentially pathogenic taxa in the phylum Proteobacteria (i.e., Klebsiella spp.) (11–13, 21, 46). OMV-mediated regulation of the gut microbiota might be via selective and direct cross-talk with commensal species, as evidenced by OMV-induced promotion in the growth of specific beneficial bacteria following membrane fusion. Note that membrane fusion is one of the mechanisms of bacterial uptake, which depends on the zeta potential and hydrodynamic diameter of OMVs (51). In addition, our findings showed that the capability of OMVs on the stimulation of bacterial proliferation was not in line with bacterial uptake efficiency, which might be associated with inconsistent division rates of bacteria or intrinsic characteristics of OMVs. The underlying mechanisms of specific fusion selection between bacteria and OMVs need further investigation.

Mucosal IgA production, which is essential for intestinal barrier maintenance and pathogen expulsion, has been proven to be modulated by the gut microbiota (23, 24, 52). PPs that act as a niche to support the interactions between DCs and B cells are critical for IgA production (22). In our study, after entry into PPs, Akk OMVs augmented mucosal IgA production and elicited locally related immune responses including DC activation and B cell differentiation into plasma cells. Moreover, the improved bacterial compositions (i.e., increments in the abundances of B. acidifaciens and B. thetaiotaomicron) could promote IgA production and secretion by switching IgM to IgA and transporting IgA across the epithelium (15, 16). It has been reported that enteric pathogens can stimulate the intestinal immune system to produce pathogen-specific, high-affinity IgA, while homeostatic IgA elicited by commensal bacteria has common polyreactivity and low affinity (23). Once gut disorder occurs, IgA coating preferentially identifies proinflammatory pathogens as the mucosal immune system remains tolerant to commensal species for maintaining gut homeostasis and gut commensal species can use IgA for mucosal colonization (25, 27, 28, 53). This pointed out the key role of OMVs in promoting robust host-microbial symbiosis, which can be effectively maintained until new disorders caused by other disturbing factors (53). In addition, we speculated that advancement of IgA production induced by OMV-mediated homeostasis could compensate decrement caused by the reduction of pathogens based on both the obtained results and the fact that IgA arises prominently during homeostasis (24). Briefly, our results indicated that OMVs could regulate mucosal immune homeostasis by modulating mucosal IgA responses and restoring the disturbed balance of the gut microbiota.

The physicochemical barrier, primarily consisting of mucus, integrated mucosal epithelial cells, and intercellular tight junctions, provides crucial protection against the entry of harmful substances into the circulatory system (54). However, the epithelial-based barriers, particularly in the colon, prevent intimate microbiota-host contacts to restrain mutual interaction (2, 55). Our data suggested that Akk OMVs could pass through mucus layer and enter epithelial cells to play a dual role in enhancing mucus secretion and expression of tight junctions and reducing intestinal epithelial impairment, showing a reciprocal interaction. The vesicles also repaired goblet cells damaged by gut disorder in quantity and functionality. The formation of a thicker mucus layer might be attributed to the reversal of gut dysbiosis and direct stimulation of goblet cells by OMV internalization. Specifically, the increased B. thetaiotaomicron can promote goblet cell differentiation and thus stimulate mucus secretion (16, 17). In turn, mucus produced by proximal colon goblet cells is able to encapsulate enteral bacteria–contained feces to avoid potential pathogenic infection (29). As for the physical barrier, Akk OMVs may activate AMPK pathway to modulate the re-assembly and stability of tight junctions (35). Akk OMVs up-regulated anti-inflammatory cytokines, including IL-10 and IL-13, but down-regulated pro-inflammatory cytokines, such as ACE2, IFN-γ, and CRP. It was worth noting that IL-10 produced by mucosal immune cells is the governing regulatory cytokine of immunity against infection, but the upgradation of pro-inflammatory cytokines increases gut permeability and simultaneously escalates epithelial apoptosis to accelerate barrier disruptions (41, 42, 56). Moreover, B. thetaiotaomicron decreased the expression of ACE2 in intestinal tissues (57). Together, the mechanisms of Akk OMVs in recovering physicochemical barrier integrity include the regulation of immune homeostasis by reducing hyperinflammatory responses caused by pathogens, the reversal of gut dysbiosis by improving the abundance of beneficial bacteria, the promotion of functional epithelial cell proliferation, and the up-regulation of tight junctions.

Clearly, the triple roles of OMVs in restoring intestinal homeostasis were reciprocal and complementary, as evidenced by collaborative regulation of gut disorders and the efficacies of orally delivered Akk OMVs to ameliorate intestinal inflammation and enhance the sensibilization of CRC to anti–PD-1 immunotherapy. Uncontrolled microbial immune reaction is regarded as one of the pivotal features of IBD (37). We found that Akk OMVs assisted the host to manipulate the mucosal immune system of the GI tract to promote polarization toward anti-inflammatory phenotypes by decreasing pathogens and strengthening epithelial-based barrier function, resulting in relief of DSS-induced colitis. On the other side, upon combination with Akk OMVs, aPD-1 significantly delayed tumor growth in CRC-bearing mice by up-regulating PD-L1 in both tumor and immune cells. Mechanisms by which Akk OMVs restored the efficacy of PD-1/PD-L1 blockade in CRC potentially refer to their critical roles in up-regulating beneficial microbial species responsible for the efficacy of PD-1–based immunotherapy and maintaining homeostasis of intestinal ecology. Compared to living bacteria that often suffer from various environmental threats including gastric acid, bile salts, and mucus in the GI tract, OMVs were relatively insensitive to these stimuli and could communicate with the host by direct cell-cell interactions, thereby exhibiting potent therapeutic efficacies. Considering the safety issues of clinical application of FMT, the success of Akk OMVs on these treatments spotlighted the potential of using gut microbiota–derived OMVs as an alternative to FMT (58).

This work discloses the versatility of bacterial OMVs in regulating intestinal homeostasis via complementary and reciprocal mechanisms of modulating the gut microbiota, mucosal immunity, and the physicochemical barrier. Gut microbiota–derived OMVs claim their comprehensive yet critical position on engaging in both microbe-microbe and host-microbe interactions in the intestine. OMV-mediated gut homeostasis demonstrates the capacity to relieve intestinal inflammation associated with colitis and improve the sensibilization of CRC to PD-1–based immunotherapy in mice. It is anticipated that these findings will stimulate more innovative thinking focused on the significance of the existence of a large quantity of bacterial OMVs behind the gut microbiota and the promise of using these vesicles as a therapeutic target for disease intervention and treatment.

MATERIALS AND METHODS

Bacteria and cell strains

Akk (DSM 22959) and B. thetaiotaomicron (VPI-5482) were cultured in brain heart infusion (BHI) broth. B. vulgatus [American Type Culture Collection (ATCC) 8482] was cultured in BHI broth supplemented with cysteine-HCl (1 mg/ml), hemin (0.015 mg/ml), and yeast extract (BHI⁺) (5 mg/ml). B. acidifaciens (JCM 10556) was cultured in Gifu anaerobic medium (GAM). B. fragilis (ATCC 25285) was cultured in modified GAM. E. coli (Nissle 1917) and S. typhimurium (SL1344) were cultured in LB broth. Akk, B. thetaiotaomicron, B. vulgatus, B. acidifaciens, and B. fragilis were cultured under anaerobic conditions at 37°C, while E. coli and S. typhimurium were under normoxic environments. Caco-2 and HT-29-MTX-E12 cells were cultured with Dulbecco’s modified Eagle’s medium (DMEM) containing 10% fetal bovine serum (FBS), streptomycin (100 mg/ml), penicillin (100 U/ml), and 1% nonessential amino acid. CT-26 cells were cultured with DMEM containing 10% FBS, streptomycin (100 mg/ml), and penicillin (100 U/ml).

Purification of OMVs

Akk was cultured in BHI broth for 4 to 7 days at 37°C in an anaerobic chamber. After centrifugation at 15,000g for 30 min, the supernatant was filtered through 0.45-μm filters at 4°C to remove bacterial pellets. The OMV supernatant was washed once with sterile PBS and concentrated to 1 ml by centrifugation at 170,000g for 60 min at 4°C. Then, the concentrated solution was filtered through 0.45-μm filters and stored at −80°C for further use. The average size of Akk OMV was measured by NTA.

Fluorescent labeling of OMVs

The outer membrane of OMVs was labeled with Cy5.5–N-hydroxysuccinimide (NHS) by covalent bonding. Briefly, OMVs (0.1 mg/ml) were suspended in PBS with Cy5.5-NHS (1 μg/ml) and subsequently incubated in a thermo shaker at 600 rpm at 37°C for 2 hours. Nuclei of OMVs were stained with Hoechst for 10 min at room temperature. After washing with PBS for two times, the labeled OMVs were visualized and analyzed by CLSM and flow cytometry, respectively.

Resistance of OMVs to SGF

Equal amounts of Akk OMVs suspended in 100 μl of PBS were incubated in 1 ml of PBS and simulated gastric fluid (SGF) (pH 1.2, per 1000 ml of water containing 0.2 g of sodium chloride, 0.32 g of pepsin, and 700 μl of hydrochloric acid) at 37°C with gentle shaking. At the indicated time points, the average size and concentration of OMVs in the medium were measured by NTA for assessing resistance of OMVs to SGF.

Mouse model of gut disorder

All animals used here were 6- to 8-week-old female Balb/c mice, which were purchased from Charles River Laboratories (Beijing, China) and Jiesijie (Shanghai, China). All animal experiments were performed under the guidelines evaluated and approved by the ethics committee of Institutional Animal Care and Use Committee of Shanghai Jiao Tong University (A2020033). To induce gut disorder, mice were daily fed with distilled water containing 5% DSS (60316ES60, YEASEN) for 10 days. Then, pretreated mice were fed with 100 μl of OMV suspension containing 20 μg of proteins by oral gavage for five consecutive days. For OMV gavage, suspension of 20 μg of proteins of OMVs was measured using bicinchoninic acid assay (P0010, Beyotime Biotechnology). PBS was used as a control.

Microbiota 16S ribosomal RNA gene sequencing

First, total bacteria DNA was extracted from the colonic samples of gut disorder mice after daily oral delivery of 100 μl of Akk OMV suspension containing 20 μg of proteins or PBS for 5 days. The 16S full-length primer was designed according to the conservative region 27F-1492R (forward primer 27F: AGRGTTTGATYNTGGCTCAG and reverse primer 1492R: TASGGHTACCTTGTTASGACTT) for Solexa polymerase chain reaction amplification (10-μl system). The high-quality circular consensus sequencing (CCS) sequence was obtained by performing quality inspection on the formed sequencing library and then was processed to generate the optimized CCS, which was clustered at the level of 97% similarity (USEARCH, version 10.0). Moreover, the species classification was obtained on the basis of the sequence composition of the operational taxonomic unit. The diversity of gut microbiota along with species annotation and taxonomy was analyzed by Silva database and RDP Classifier. The richness and diversity of species in a single sample was examined by alpha diversity analysis, while beta diversity analysis was performed to compare the differences in the community composition and structure of different samples.

Bacterial uptake of OMVs

A volume of 50 μl of logarithmic phase bacterial solution was incubated in 1 ml of corresponding culture medium at 37°C with 100 μl of PBS containing OMVs (0.1 mg/ml). For simulating the gut disorder in vitro, a high initial amount of B. vulgatus (the OD value at 600 nm was around 0.1) was chosen to coculture with OMVs. At the indicated time points, 100 μl of the mixture was collected by centrifugation for 5 min at 7000g. Subsequently, the collections were washed with PBS three times to remove free labeled OMVs. Bacterial uptake of OMVs was visualized by CLSM and analyzed by flow cytometry. In addition, growth curves of bacteria were measured by microplate reader. At the indicated time points, 200 μl was taken from the medium to record the OD value at 600 nm. PBS was used as a control.

Entry of OMVs into intestinal lymph nodes

Entry of OMVs into PPs and MLNs was evaluated according to a previously reported method (59), which has been widely used to observe the entry of nanoparticles into epithelial cells and intestinal lymph nodes. A solution containing Cy5.5-labeled Akk OMVs (0.1 mg/ml) was injected into PP-located intestinal segment tied at both ends. At indicated time points, PPs and MLNs were extracted from the treated mice. The IVIS system was used to image the entry of OMVs into PPs and MLNs.

In vitro immune responses

BMDCs were extracted from bone marrow of mice and cultured in RPMI 1640 medium supplemented with granulocyte-macrophage colony-stimulating factor (20 ng/ml; 51048-MNAH, Sino Biological), 10% FBS, streptomycin (100 mg/ml), and penicillin (100 U/ml) for 3 days in vitro. Then, half of the supernatant was replaced with fresh culture medium, and cells were incubated for additional 3 days. To observe the uptake of Akk OMVs by BMDCs, 100 μl of PBS containing OMVs (0.1 mg/ml) was incubated with BMDCs for 24 hours at 37°C. The cells were washed for three times, followed by staining with Hoechst for 10 min at room temperature before observation by CLSM and flow cytometric analysis. The cells were stained with anti-CD11b (M1/70), anti-CD11c (N418), anti-CD80 (16-10A1), and anti-CD86 (24F) to assess the activation of BMDCs. For analysis of T cell responses, 100 μl of PBS supplied with Akk OMVs (0.1 mg/ml) was added into BMDC medium. After 2-day coincubation, CD4⁺ T cells extracted from mouse spleen were added and cultured for another 4 days. The cells were first stained with extracellular antibodies, including anti-CD3 (17A2), anti-CD4 (GK1.5), anti-CD8 (53-6.7), and anti-CD25 (3C7) antibodies, and then washed, fixed, and permeabilized before intracellular staining. Intracellular antibodies included anti-FOXP3 (FJK-16s) antibody and cytokines [anti–IL-17 (TC11-18H10.1), anti–IFN-γ (XMG1.2), and anti–IL-4 (11B11) antibodies]. The Foxp3/Transcription Factor Fixation/Permeabilization Kit (00-5521-00, eBioscience) was used to fix and permeabilize the cells. All antibodies were purchased from BioLegend and eBioscience. Data were acquired on a FACSVerse flow cytometer (BD Biosciences, USA) and analyzed with FlowJo software (Tree Star Inc.).

Mucosal immune responses

PPs and MLNs were extracted from gut disorder mice 5 days after oral administration of Akk OMV suspension containing 20 μg of proteins by gavage. The tissues were mashed through 70-μm cell strainers to generate single-cell suspensions in PBS buffers. After incubation with purified anti-CD16/32 (93) at 4°C for 15 min, the obtained cells were stained with antibodies. To measure activation and maturation of DCs and B cells, the cells were stained with anti-CD11b (M1/70), anti-CD11c (N418), anti-CD80 (16-10A1), anti-CD86 (24F), anti-B220 (RA36B2), anti-CD138 (281-2), anti-CD69 (H1.2F3), and anti-IgA (mA-6E1). To analyze T cell responses, the cells were first stained with extracellular antibodies, including anti-CD3 (CD3-12), anti-CD4 (GK1.5), anti-CD8 (53-6.7), and anti-CD25 (3C7) antibodies and then washed, fixed, and permeabilized before intracellular staining. Intracellular antibodies included anti-FOXP3, anti–IL-17 (TC11-18H10.1), anti–IFN-γ (XMG1.2), and anti–IL-4 (11B11) antibodies. The Foxp3/Transcription Factor Fixation/Permeabilization Kit (00-5521-00, eBioscience) was used to fix and permeabilize the cells. To measure the production of IgA, intestinal fluid and fecal samples were collected from treated mice 5 days after oral administration of OMVs. The intestinal fluid samples were diluted and measured via IgA enzyme-linked immunosorbent assay (ELISA) kits (EK274, MultiSciences) and read on a Synergy H1. IgA-bound bacteria from fecal samples were measured as described (32). Briefly, 100 mg of fresh stool sample was dissolved in 1 ml of sterilized PBS and then homogenized for 1 min. After multiple gradient centrifugations, large particles in the samples were removed and fecal bacteria were collected. For flow cytometric analysis, bacteria samples were stained with anti-IgA antibody (mA-6E1). All antibodies were purchased from BioLegend and eBioscience. Data were acquired on a FACSVerse flow cytometer (BD Biosciences, USA) and analyzed with FlowJo software (Tree Star Inc.).

Physicochemical barrier function

The proximal colons were extracted from gut disorder mice 5 days after oral administration of Akk OMV suspension containing 20 μg of proteins by gavage. To analyze mucus barrier, the samples were processed for Alcian blue staining (C0155S, Beyotime Biotechnology) in accordance with the manufacturer’s instructions. Fiji (ImageJ) was used to quantitatively analyze the area of stained cells and the percentage of goblet cells in the intestinal epithelium. To analyze physical barrier, the samples were processed for immunofluorescence staining in accordance with the manufacturer’s instructions, including occludin (sc-133256, Santa Cruz Biotechnology) and ZO-1 (AF5145, Affinity Biosciences). Levels of cytokines including IL-10 (EK210/4, MultiSciences), IL-13 (EK213/2, MultiSciences), and ACE2 (EK1188, Boster Biological Technology) in colon tissue as well as IFN-γ (EK280/3, MultiSciences) and CRP (EK294/2, MultiSciences) in serum were measured by commercially available ELISA kits. To observe uptake of Akk OMVs by intestinal epithelial cells, a solution of Cy5.5-labeled OMVs (0.1 mg/ml) was injected into the intestinal segment tied at both ends and the treated mice were euthanized for collecting intestinal tissues 2 hours later. Then, the tissues were processed for CLSM observation.

Uptake and activation of Caco-2 and HT-29-MTX-E12 cells

A volume of 100 μl of PBS containing Cy5.5-labeled Akk OMVs (0.1 mg/ml) was incubated with Caco-2 and HT-29-MTX-E12 cells for 24 hours at 37°C, respectively. The cells were washed three times with PBS and stained with Hoechst before visualizing uptake of OMVs under CLSM, respectively. Similarly, a volume of 100 μl of PBS containing Akk OMVs (0.1 mg/ml) was cultured with Caco-2 cells along with LPS (5 μg/ml) for 24 hours at 37°C. Then, the cells were processed for immunofluorescence staining, including occludin (sc-133256, Santa Cruz Biotechnology) and ZO-1 (AF5145, Affinity Biosciences), and then stained with Hoechst before CLSM observation. A volume of 100 μl of PBS containing Akk OMVs (0.1 mg/ml) was cultured with HT-29-MTX-E12 cells for 24 hours at 37°C. Then, the cells were processed for Alcian blue staining (C0155S, Beyotime Biotechnology).

Efficacy of OMVs in treating DSS-induced colitis

Colitis was induced in Balb/c mice with 5% DSS (60316ES60, YEASEN) dissolved in drinking water for 7 days as described (60). The development of inflammation was assessed by body weight changes and appearance of bloody stools. The colitis mice were daily fed with 100 μl of PBS containing Akk OMVs (0.2 mg/ml) by oral gavage for 5 days. DSS mice treated with 10⁸ colony-forming units (CFUs) of Akk (100 μl of bacterial suspension with an OD₆₀₀ value of 1, concentrated three times) and PBS were used as controls. Then, colon tissues were extracted from the mice for efficacy assessment. The colon length was recorded from cecum to rectum. Proximal colons were processed for MPO staining (Abcam, ab208670) and H&E staining (C0105M, Beyotime Biotechnology) in accordance with the manufacturer’s instructions.

Efficacy of OMVs in treating CRC mice

Balb/c mice were implanted with 1.5 × 10⁶ CT-26 cells subcutaneously. Upon reaching a tumor size of 100 to 400 mm³, the mice were treated with gavage of 100 μl of PBS containing Akk OMVs (0.2 mg/ml) along with intraperitoneal injection of 100 μg of anti–aPD-1. The mice received treatment of anti–aPD-1 three times every 3 days and fed with OMVs for two consecutive days before intraperitoneal injection. Tumor size was routinely monitored every day, and the mice were euthanized for sampling 10 days after treatment. The samples were processed for immunofluorescence staining using PD-L1 (64988, Cell Signaling Technology), CD45 (20103-1-AP, Proteintech), CD3 (CD3-12, Abcam), and CD8 (EPR20305, Abcam) antibodies in accordance with the manufacturer’s instructions and stained with Hoechst before they were observed by CLSM. The samples were also processed for H&E staining (C0105M, Beyotime Biotechnology). For flow cytometric quantitative analysis, tumor samples were cut into small pieces by surgical instruments and then digested by DMEM supplemented with collagenase type IV (100 U/ml) and deoxyribonuclease I (100 μg/ml) at 37°C for 30 min with shaking, followed by mashing through 70-μm cell strainers. The obtained single-cell suspension was washed and resuspended with PBS containing 0.5% bovine serum albumin. To measure expression of PD-L1 in tumor beds, the cells were stained with anti-CD45 (QA17A26, BioLegend) and anti–PD-L1 (10F.9G2, BioLegend). Data were acquired on a FACSVerse flow cytometer (BD Biosciences, USA) and analyzed with FlowJo software (Tree Star Inc.).

Statistical analysis

All statistical analyses were performed with GraphPad Prism 9 software. All the experimental data were analyzed using Student’s unpaired t test, one-way analysis of variance (ANOVA) test, or two-way ANOVA test and expressed as means ± SEM. Dunnett’s t test was used to correct multiple comparisons using statistical hypothesis testing. Differences were regarded to be statistically significant when *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001. Q test (Dixon criterion) was used to determine and eliminate abnormal data.

Supplementary Materials

This PDF file includes:

Figs. S1 to S21

View/request a protocol for this paper from Bio-protocol.