Abstract Combination immunotherapy of cancer vaccines with immune checkpoint inhibitors (ICIs) represents a promising therapeutic strategy for immunosuppressed and cold tumors. However, this strategy still faces challenges, including the limited therapeutic efficacy of cancer vaccines and immune-related adverse events associated with systematic delivery of ICIs. Herein, we demonstrate the antitumor immune response induced by outer membrane vesicle from Akkermansia muciniphila (Akk-OMV), which exhibites a favorable safety profile, highlighting the potential application as a natural and biocompatible self-adjuvanting vesicle. Utilizing tumor cell-derived exosome as an antigen source and Akk-OMV as a natural adjuvant, we construct a cancer vaccine formulation of extracellular vesicles hybrid lipid nanovesicles (Lipo@HEV) for enhanced prophylactic and therapeutic vaccination by promoting dendritic cell (DC) maturation in lymph node and activating cytotoxic T cell (CTL) response. The Lipo@HEV is further loaded with plasmid to enable gene therapy-mediated PD-L1 blockade upon peritumoral injection. Meanwhile, it penetrates into lymph node to initiate DC maturation and CTL activation, synergistically inhibiting the established tumor. The fabrication of extracellular vesicles hybrid plasmid-loaded lipid nanovesicles reveals a promising gene therapy-guided and vesicle-based hybrid system for therapeutic cancer vaccination and synergistic immunotherapy strategy. Keywords: Cancer vaccine, Extracellular vesicles, Liposomes, Hybrid lipid nanovesicles, Akkermansia muciniphila, Immune checkpoint blockade Graphical abstract Schematic illustration of fabricated Lipo-PD-L1@HEV for synergistic cancer immunothearpy. After thin-film hydration of cationic liposome, incubation of cationic liposome with PD-L1 trap plasmid, and purification of tumor cell-derived exosome and Akk-OMV, the Lipo-PD-L1@HEV was fabricated by fusing extracellular vesicles with plasmid-loaded cationic liposome. Upon peritumoral injection, Lipo-PD-L1@HEV can localize in tumor for PD-L1 blockade, and penetrate into lymph node to efficiently initiate DC maturation, antigen presentation and CTL activation, synergistically inhibiting the established tumor. Image 1 [37]Open in a new tab 1. Introduction Among the tremendous efforts focused on immunotherapy to facilitate the host immune system in efficiently recognizing and killing tumor cells, cancer vaccines have undergone a resurgence over the past decade [[38]1], evidenced by the tumor-specific immunogenicity and clinical efficacy towards melanoma, glioblastoma and other cancers [[39]2,[40]3]. The basic principles of successful therapeutic vaccination involve delivery of high-quality antigens to dendritic cells (DCs), optimal DC activation, induction of robust and sustained cytotoxic T lymphocyte (CTL) responses, infiltration of the tumor microenvironment (TME) and durability of responses [[41]4], which can be achieved by co-administration of tumor antigens with adjuvants [[42]5], and reversal of tumor-induced immune exhaustion by immune checkpoint blockade [[43]6]. However, conventional cancer vaccines typically show limited therapeutic efficacy in clinical trials due to central and peripheral tolerance responses and elicit individual varying degrees of immune responses [[44]7]. Moreover, systemic delivery of immune checkpoint inhibitors (ICIs) can cause a series of side effects, including immune-related adverse events (irAEs) and chronic immune toxicity [[45][8], [46][9], [47][10]]. Thus, the precise therapeutic cancer vaccines and rational combination immunotherapies that induce the safe and robust antigen-specific immune responses warrant further development. Extracellular vesicles (EVs), secreted by both eukaryotic and prokaryotic cells, are a heterogeneous group of lipid-bound nano-sized membrane vesicles that contain hundreds of lipids, proteins, carbohydrates and nucleic acids, and serve as key mediators of various (patho)physiological processes [[48]11]. Tumor-derived EVs, which carry large quantities of cell components originated from the parent tumor cells, can be exploited as a tumor antigen source in a non-replicative form [[49]12]. On the other hand, the gut microbiota has been demonstrated as both a biomarker and an adjuvant for enhancing clinical response to ICIs [[50]13]. Akkermansia muciniphila, a gram-negative bacterium as the paradigm for next-generation beneficial microorganisms, is a promising candidate to enhance the clinical response to checkpoint inhibitor immunotherapies [[51][14], [52][15], [53][16]]. Clinical trials are crucial to confirm the role of Akkermansia muciniphila in increasing the success of immunotherapies, whereas further studies are needed for a deeper understanding of the mode of action [[54]17]. Recent studies have shown that Akkermansia muciniphila and its secreted protein 9 (P9) stimulate the secretion of IL-6 by macrophages and enterocytes [[55]18], and the outer membrane protein Amuc-1100 blunts colitis associated tumorigenesis by modulation of CD8^+ T cells in mice [[56]19]. Utilizing spectroscopic analysis and chemical synthesis instead of multi-omics technique, a newly identified PE lipid from Akkermansia muciniphila has been demonstrated to promote the release of TNF-α and IL-6 in DCs [[57]20]. Given that bacterial extracellular vesicles (BEVs) are nanometer-scale packages of adjuvant components inherited from the parent bacteria [[58][21], [59][22], [60][23]], we hypothesize that the outer membrane vesicle from Akkermansia muciniphila (Akk-OMV) mediates the enhancement of the parent bacterium to checkpoint inhibitor immunotherapy, and can be further exploited as a natural nanoadjuvant for therapeutic cancer vaccination. Herein, we purified three kinds of BEVs by means of differential centrifugation with size exclusion chromatography method, and demonstrated the antitumor immune response of Akk-OMV that prevented tumorigenesis and synergized with PD-L1 antibody against tumor growth. Upon different regimes of subcutaneous and intraperitoneal injections, Akk-OMV exhibited a favorable safety profile, which was suitable for further adjuvant immunotherapy strategies. Taking advantage of vesicle fusion technique that potentiated the hybrid vesicles with properties inherited from different mono-vesicles [[61]24,[62]25], the cancer vaccine formulation of EVs hybrid lipid nanovesicles (Lipo@HEV) was fabricated for co-delivery of antigens from tumor cell-derived exosome and adjuvant components from Akk-OMV, enabling the activation of DC maturation and CTL response for enhanced cancer vaccination. The Lipo@HEV based on the co-delivery vector of cationic liposome was further loaded with a PD-L1 trap plasmid to achieve PD-L1 blockade in tumor for reversing CTL exhausion and avoiding irAEs. Simultaneously, the plasmid-loaded Lipo@HEV penetrated into the lymph nodes to promote DC maturation and activate CTL response, improving the antitumor activity of PD-L1 blockade against the established tumor. The present study demonstrates the synergistically immunotherapeutic potential of the plasmid-loaded extracellular vesicles hybrid lipid nanoplatform, providing an extensible, biomimetic and gene therapy-based approach for therapeutic cancer vaccination and combination immunotherapy. 2. Material and methods 2.1. Cell culture and animals The murine melanoma cell line B16–F10 and murine breast cancer cell line 4T1 were obtained from American Type Culture Collection (ATCC). DC2.4 cell line was obtained from Procell Life Science & Technology Co. Ltd (Wuhan, China). B16–F10 and CT26 were cultured in DMEM medium (Gibco) with 10 % FBS, and 4T1 and DC2.4 cells were cultured in 1640 medium (Gibco) with 10 % FBS. All the cell lines were incubated at 37 °C with 5 % CO[2], and antibiotics (1 % penicillin/streptomycin, v/v) were added to cell culture medium. Female C57BL/6 and BALB/c mice (6–8 week) were purchased from Vital River Laboratory Animal Technology Co. Ltd (Beijing, China) and maintained under specific pathogen-free (SPF) condition. All animal experiments were conducted under guidelines approved by the Animal Care and Use Committee of Tongji Hospital, Tongji Medical College, Huazhong University of Science and Technology (SYXK 2019-0106). 2.2. Isolation and characterization of EVs Both eukaryotic and prokaryotic EVs were isolated by differential centrifugation method according to the previous studies with some modifications [[63]26,[64]27]. For isolation of tumor cell-derived exosomes, B16–F10, 4T1 and CT26 tumor cells were cultured in DMEM or 1640 medium with 10 % FBS until 60–70 % density. The cells were washed with PBS, and maintained in Opti-MEM medium (Gibco) for another 48 h. The supernatant was collected and centrifuged at 800 g for 5 min and 2000 g for 10 min at 4 °C. The supernatant was filtered using 0.4 μm polycarbonate membrane (Millipore), and the pellet was recovered after ultracentrifugation of 200000 g for 70 min at 4 °C. The supernatant was aspirated, and the pellet was resuspended in PBS and subsequently ultracentrifuged at 200000 g for 90 min at 4 °C. For isolation of BEVs, E. Coli DH5α were cultured in LB medium at 37 °C with 250 rpm shaking overnight, and the cultured cells were pelleted at 12000 g for 10 min at 4 °C. The supernatant was filtered using 0.4 μm polycarbonate membrane (Millipore), and the pellet was recovered after ultracentrifugation of 200000 g for 70 min at 4 °C. The supernatant was aspirated, and the pellet was resuspended in PBS and subsequently ultracentrifuged at 200000 g for 90 min at 4 °C. The collected pellet was purified by a size exclusion filter (Umibio, UR90102). Akkermansia muciniphila (ATCC BAA-835) were cultured in the brain heart infusion broth medium (BD BBL, 211059), and Bifidobacterium pseudolongum (ATCC 25526) were cultured in the modified reinforced clostridial broth medium (Shandong Tuopu Biol-engineering Co. Ltd., MD039) for 7 days under anaerobic conditions. The medium was supplemented with resazurin (5 mg/L) and l-cysteine (0.5 g/L). The cultured bacteria were pelleted twice at 12000 g for 10 min at 4 °C, and the isolation procedures were the same as those of E. Coli DH5α-OMV. The morphology of exosome and BEVs were imaged by TEM (Hitachi, HT7700) after negative staining with 2 % phosphotungstic acid. The diameter and zeta potential were determined by nanoparticle tracking analysis (NTA, Particle Metrix, Germany). The protein profiles were analyzed by SDS-PAGE with equal loading amounts of 10 μg vesicle protein, measured by the BCA assay (Beyotime, P0010). 2.3. Prophylactic antitumor evaluation of BEVs and transcriptome sequencing analysis Akk-OMV, DH5α-OMV and Bifi-CMV were subcutaneously injected at the tail base of BALB/c mice (10 μg) at 3 day-intervals for 5 times, respectively, and mice were subcutaneously inoculated with 2 × 10^5 4T1 tumor cells 7 days after the final injection. The tumor volume was measured every 2 days and calculated according to the formula V (mm^3) = length × width^2 × 0.52. The maximal tumor burden permitted was 1500 mm^3. After the prophylactic antitumor evaluation study of different BEVs, the collected spleens and tumors were snap frozen in liquid nitrogen. The total RNA was extracted using Trizol-reagent (Invitrogen) according to the manufacturer's protocol. Then eukaryotic mRNA was enriched by Oligo(dT) beads, and the enriched mRNA was fragmented into short fragments using fragmentation buffer and reverse transcribed into cDNA with random primers. Second strand cDNA were synthesized by DNA polymerase I, RNase H, dNTP and buffer. The cDNA fragments were purified, end repaired, poly(A) added and ligated to Illumina sequencing adapters. The ligation products were size selected by agarose gel electrophoresis, PCR amplified and sequenced using Illumina HiSeq2500. The data were analyzed by the online platform of Majorbio Cloud Platform. 2.4. Antitumor immune response induced by Akk-OMV and combination therapy with PD-L1 antibody BALB/c mice were subcutaneously inoculated with 2 × 10^5 4T1 tumor cells on day 0, and Akk-OMV was subcutaneously injected at the tail base (20 μg) on day 10, 13, 16, 19 and 22. Then mice were euthanized on day 25, and tumors, lymph nodes and spleens were collected for flow cytometry analysis. For combination therapy of Akk-OMV with PD-L1 antibody, BALB/c mice were subcutaneously inoculated with 2 × 10^5 4T1 cells on day −8. Akk-OMV was subcutaneously injected at the tail base (20 μg) on day 0, 3, 6, 9 and 12, and PD-L1 antibody (Leinco Technologies, P363) was intraperitoneally injected (150 μg) on day 6, 9 and 12. The tumor volume was measured every 3 days and calculated according to the formula V (mm^3) = length × width^2 × 0.52. The mice were euthanized on day 21, and the tumors and spleens were collected for flow cytometry analysis. 2.5. In vivo biosafety evaluation of Akk-OMV For a single-dose intraperitoneal injection, BALB/c mice were injected with Akk-OMV or DH5α-OMV at a dose of 20 μg, and the mice were euthanized at 8 h post-injection. For 3 times intraperitoneal injections, BALB/c mice were injected with Akk-OMV or DH5α-OMV at a dose of 20 μg on day 0, 3, 6, and the mice were euthanized on day 14. The survival rate was recorded during the treatments. For 3 times subcutaneous injections, C57BL/6 mice were injected with Akk-OMV or DH5α-OMV at a dose of 20 μg on day 0, 3, 6, and the mice were euthanized on day 9. The serum was collected, and hepatic and renal function indexes were evaluated by Chemray240 biochemical autoanalyzer. The major organs including hearts, livers, spleens, lungs and kidneys were collected, and H&E staining was conducted to evaluate the organ histological changes. 2.6. Preparation of EVs hybrid lipid nanovesicles The cationic liposome was prepared via thin-film hydration method. DOTAP (Avanti Polar Lipids, 890890P) and cholesterol (Sigma, 700000P) were dissolved in chloroform at a molar ratio of 1:1, and the solvent was evaporated under nitrogen flow. The dry film was hydrated with ultra-pure water by magnetic stirring overnight to obtain a coarse suspension with stock concentration of 10 mM each. The obtained suspension was extruded 11 times through polycarbonate membranes of 800 nm, 400 nm and 200 nm using a mini-extruder (Avanti Polar Lipids). To prepare the EVs hybrid lipid nanovesicles, cationic liposome was mixed with EVs and then sonicated with a probe ultrasonicator (50 % power, 130 W, 20 kHz) for 2 min (2 s interval) in ice water bath. Then the solution was co-extruded at least 5 times through polycarbonate membranes of 800 nm, 400 nm and 200 nm. To prepare plasmid-loaded EVs hybrid lipid nanovesicles, cationic liposome was first incubated with plasmid at N/P ratio of 5 for 15 min, and then sonicated and co-extruded with EVs. The pcDNA3.1 plasmid encoding PD-L1 trap was constructed by VectorBuilder (Guangzhou, China) according to the previous studies [[65]28,[66]29]. For the validation of PD-L1 trap plasmid, the DNA sequencing, restriction enzyme analysis (ApaLI and FspI) and agarose gel electrophoresis were conducted. The pRP-CMV-Luciferase plasmid was obtained from VectorBuilder (Guangzhou, China). The plasmids were extracted and purified using a EndoFree Maxi Plasmid Kit (Tiangen, China), and the accuracy was validated by DNA sequencing. 2.7. Characterization of EVs hybrid lipid nanovesicles The morphology of exosome, OMV and hybrid nanovesicles were imaged by TEM (Hitachi, HT7700) after negative staining with 2 % phosphotungstic acid. The size and zeta potential were analyzed by NTA. For fusion analysis of EVs with cationic liposome, OMV and exosome were incubated with PKH26 (Sigma, MINI26) and PKH67 (Sigma, MINI67) at final dye concentration of 5 μM for 10 min, respectively. After ultra-centrifugation at 200000g for 70 min at 4 °C, the pellets of dye-labeled exosome and OMV were resuspended in PBS and co-extruded with cationic liposome at different weight ratios of 100:1:1, 50:1:1, 20:1:1. The membrane colocalization was imaged using a confocal laser scanning microscope (CLSM, Nikon Eclipse TI), and quantified by nano-flow cytometry (NanoFCM). The tumor cell membrane (TCM) was extracted using a membrane and cytosol protein extraction kit (Beyotime, P0033) according to the manufacturer's instructions, and Lipo@HMV was prepared by fusing liposome with OMV and TCM following the same preparation procedure of Lipo@HEV. The protein profile was analyzed by SDS-PAGE, and all the samples were normalized to equivalent protein concentration of 10 μg measured by the BCA assay (Beyotime, P0010). The gel was stained in Coomassie brilliant blue dye solution (Servicebio, G2059) for 2 h and washed with in ultra-pure water for 4 h, and the gel was imaged by a Bio-Rad ChemiDoc MP Imaging System. For western blot analysis of exosome markers, proteins were transferred to polyvinylidene fluoride (PVDF) membranes, which were blocked with 5 % nonfat milk TBST for 1h and incubated with primary antibodies of CD9 (Abcam, ab92726) and TSG101 (Abcam, ab125011) overnight at 4 °C, followed by washing and incubating with a goat anti-rabbit IgG-HRP secondary antibody (Abcam, ab205718). The blots were imaged by a Bio-Rad ChemiDoc MP Imaging System. 2.8. In vitro cellular uptake and dendritic cell maturation analysis For cytotoxicity assay, B16–F10 and DC2.4 cells were seeded in 96-well plates and incubated with Lipo@HEV at different concentrations of total lipids for 24 h, and the cytotoxicity was determined by CCK8 assay. After PKH26/PKH67 staining of OMV/exosome and hybrid lipid nanovesicles preparation, B16–F10 and DC2.4 cells were seeded on glass coverslips and treated with PKH26/PKH67-labeled Lipo@HEV for 6 h. The cells were fixed with 4 % paraformaldehyde for 20 min and stained with DAPI for 10 min, then imaged by CLSM. BMDCs were differentiated from bone marrow cells isolated from the femurs and tibias of C57BL/6 mice according to the previous method [[67]30]. The cells were cultured in 1640 medium supplemented with 10 % FBS, 20 ng/mL GM-CSF (PeproTech, 315-03) and 10 ng/mL IL-4 (PeproTech, 214-14) in petri dishes. The non-adherent and loosely adherent cells were collected and treated with PKH26/PKH67-labeled Lipo@HEV for 6 h. The cellular uptake rate of PKH26/PKH67-labeled hybrid lipid nanovesicles was quantitative analyzed by flow cytometry (Beckman Coulter CytoFLEX). For DC maturation assay, BMDCs were plated into 24-well non-treated dishes and stimulated with different EVs and formulations of hybrid lipid nanovesicles (3 μg/mL vesicle protein) for 24 h. Then BMDCs were collected and incubated with anti-CD16/32 (Biolegend, 156603), and stained with FITC-anti-CD11c (Biolegend, 117306), PE-anti-CD80 (Biolegend, 104707) and APC-anti-CD86 (Biolegend, 105011) according to the manufacturer's protocols, followed by flow cytometry analysis. The data were analyzed by CytExpert 2.4 software. 2.9. In vivo biodistribution and plasmid transfection imaging For the biodistribution study of Akk-OMV, the purified Akk-OMV was first labeled with near-infrared fluorescent dye DiR at the concentration of 50 μM. After 1 h incubation at room temperature, unbound dye was removed by ultra-centrifugation at 200000g for 70 min at 4 °C, and the pellet was resuspended in PBS. Then the DiR-labeled Akk-OMV was subcutaneously injected at tail base of BALB/c mice at the dose of 10 μg. The major organs and inguinal lymph nodes were collected at 24, 36 and 48 h post-injection, and imaged by Caliper Spectrum In-vivo Imaging System. For the biodistribution study of hybrid lipid nanovesicles, DiR-labeled B16–F10 exosome and Akk-OMV were co-extruded with liposome to generate DiR-labeled Lipo@B16-EXO, Lipo@Akk-OMV and Lipo@HEV, and different formulations were subcutaneously injected at the tail base of C57BL/6 mice at the dose of 20 μg. The major organs and inguinal lymph nodes were collected and imaged at 48 h post-injection. For the biodistribution study of Lipo-PD-L1@HEV upon peritumoral injection, DiR-labeled B16–F10 exosome and Akk-OMV were co-extruded with plasmid-loaded liposome to generate the DiR-labeled Lipo-PD-L1@HEV. Then the DiR-labeled Lipo-PD-L1@HEV was peritumorally injected to B16–F10 tumor-bearing mice at the dose of 20 μg. The tumors, lymph nodes and major organs were collected and imaged at 48 h post-injection. For in vivo and ex vivo bioluminescence imaging, Lipo-luciferase and Lipo-luciferase@HEV were peritumorally injected into B16–F10 tumor-bearing mice at the dose of 15 μg plasmid, respectively. The mice were anesthetized and imaged at 48 h post-injection after intraperitoneal injection of D-luciferin potassium salt (ST196, Beyotime). Then the tumors and major organs were collected, and the ex vivo imaging were conducted. 2.10. In vivo antitumor evaluation of EVs hybrid lipid nanovesicles For the prophylactic study of EVs hybrid lipid nanovesicles, C57BL/6 mice were randomly divided into five groups: (1) control, (2) Lipo@B16-EXO, (3) Lipo@Akk-OMV, (4) Lipo@B16-HEV, (5) Lipo@4T1-HEV. The different formulations were subcutaneously injected at the tail base of the mice (40 μg vesicle protein), and the injections were performed at 3 day-interval for 3 times. Seven days after the final injection, mice were subcutaneously inoculated with 2 × 10^5 B16–F10 cells. For the therapeutic study of EVs hybrid lipid nanovesicles, C57BL/6 or BALB/c mice were randomly divided into four groups: (1) control, (2) Lipo@EXO, (3) Lipo@Akk-OMV, (4) Lipo@HEV. The mice were subcutaneously inoculated with 2 × 10^5 B16–F10 or CT26 cells on day 0, and the different formulations were subcutaneously injected at the tail base (40 μg vesicle protein) on day 4, 7 and 10. For the therapeutic study of Lipo-PD-L1@HEV, C57BL/6 or BALB/c mice were randomly divided into three groups: (1) control, (2) Lipo-PD-L1, (3) Lipo-PD-L1@HEV. C57BL/6 mice and BALB/c mice were subcutaneously inoculated with 2 × 10^5 B16–F10 cells or 4T1 tumor cells on day 0, and different formulations were peritumorally injected on day 9 at 3 day-intervals for 3 times with equal dose of 20 μg plasmid. The tumor volume was measured every 2–3 days and calculated according to the formula V (mm^3) = length × width^2 × 0.52. The maximal tumor burden permitted was 1500 mm^3. 2.11. Single-cell RNA-sequencing analysis After the prophylactic antitumor evaluation of EVs hybrid lipid nanovesicles, the tumors from Lipo@B16-HEV or control group were pooled and stored in the tissue storage solution (Miltenyi Biotec, 130-100-008). The sample preparation and sequencing were conducted by Berry Genomics Co., Ltd. (Beijing, China). The data were processed by 10 × cellranger software to get the basic expression file, cells with fewer than 200 genes or more than 6000 genes detected were removed. Genes that were expressed in fewer than 2 cells were also removed. The principal components were calculated using scanpy.pca. Dimensionality reduction and embedding was performed using UMAP analysis by the scanpy.tl.umap function. Differentially expressed genes were calculated using the Wilcoxon sum rank test with a fold change cutoff of 2. All p-values were adjusted for multiple testing using the Bonferroni correction. Initial annotation was ascribed by comparing the defined cell type markers. Pathway enrichment analysis was performed using the ssGSEA function from gseapy package with default parameters. Gene sets were obtained from the MSigDB database ([68]https://www.gsea-msigdb.org/gsea/msigdb). 2.12. Flow cytometry analysis The collected tumors, spleens and lymph nodes were stored in the tissue storage solution, followed by mechanical disruption and filtering through 70 μm cell strainers to obtain the single-cell suspensions. The tumor cells were incubated with anti-CD16/32 (Biolegend, 156603), and stained with APC-anti-CD3 (Biolegend, 100236) and PE-anti-CD8 (Biolegend, 100707), or FITC-anti-CD4 (Biolegend, 100405), APC-anti-CD25 (Biolegend, 102011), and BV421-anti-FOXP3 (Biolegend, 126419). The splenocytes were incubated with anti-CD16/32 (Biolegend, 156603), and stained with APC-anti-CD3 (Biolegend, 100236), PE-anti-CD8 (Biolegend, 100707), PerCP/Cyanine5.5-anti-CD44 (Biolegend, 103031), and FITC-anti-CD62L (Biolegend, 104405). The lymph node cells were incubated with anti-CD16/32 (Biolegend, 156603), and stained with FITC-anti-CD11c (Biolegend, 117306), APC-anti-CD86 (Biolegend, 105011), and PE-anti-CD80 (Biolegend, 104707), or APC-anti-CD3 (Biolegend, 100236), and PE-anti-CD8 (Biolegend, 100707). The staining procedure was performed according to the manufacturer's protocol, followed by flow cytometry analysis (Beckman Coulter CytoFLEX). The data were analyzed by CytExpert 2.4 software. 2.13. Immunofluorescence analysis The collected spleens were fixed in 4 % paraformaldehyde, embedded in paraffin and sectioned. After deparaffinization, rehydration, antigen recovery and blocking with 3 % hydrogen peroxide and 5 % BSA, the spleen sections were dual immunofluorescence stained with primary antibodies of CD4 (Servicebio, [69]G15064) and CD8 (Servicebio, [70]G15068) overnight at 4 °C, followed by incubation of Cy3-conjugated (Servicebio, GB21303) and FITC-conjugated secondary antibodies (Servicebio, GB22303) at room temperature for 1 h, respectively. Mounting was performed using a fluormount containing DAPI (Servicebio, G1407) for further image acquisition (Olympus BX53). 2.14. Enzyme-linked immunosorbent assay of serum cytokines The blood samples were collected from mice and stored at 4 °C overnight to obtain the serum samples after centrifugation at 8000g for 15 min. The serum samples were stored at −80 °C until further analysis. The concentrations of TNF-α (Invitrogen, 88-7324-22), IL-6 (Invitrogen, 88-7064-22), and IFN-γ (MultiSciences, EK280/3-48) were analyzed by enzyme-linked immunosorbent assay (ELISA) according to the instructions of the ELISA kits. 2.15. Statistical analysis Data analysis was conducted by Graphpad Prism 8.0 software. One-way ANOVA post hoc Tukey multiple comparison test was used for comparisons of more two groups, and student's t-test was used for comparisons between two groups. Graphs included means and error bars with all results presented as mean ± standard deviation (SD) or standard error of mean (SEM) as indicated in the figure legends. For survival rate study, log-rank (Mantel-Cox) test was used. The statistically significant was considered as follows: *p < 0.05, **p < 0.01, ***p < 0.001. 3. Results and discussion 3.1. Antitumor immune response induced by Akk-OMV Given the anaerobic characteristic of Akkermansia muciniphila, the culturing procedure was conducted under the strict anaerobic conditions with mucin-based brain heart infusion broth (BHI) medium. However, the transmission electron micrograph (TEM) images showed the impurity of Akk-OMV after bacterial culturing in mucin-based BHI medium or differential centrifugation without further purification process ([71]Fig. S1). After optimization of culture condition and purification procedure, the differential centrifugation with size exclusion chromatography method was utilized to isolate the OMVs from gram-negative bacteria Akkermansia muciniphila and Escherichia coli DH5α (DH5α-OMV), and cytoplasmic membrane vesicle from gram-positive bacterium Bifidobacterium pseudolongum (Bifi-CMV). TEM images indicated that Akk-OMV, Bifi-CMV and DH5α-OMV displayed nano-sized lipid-bilayer vesicular structures ([72]Fig. 1a), with average diameters of 127.5 ± 4.9 nm, 141.9 ± 2.3 nm and 127.8 ± 2.3 nm, and average zeta potentials of −44.73 ± 1.71 mV, −25.19 ± 0.61 mV and −35.11 ± 1.69 mV, respectively, as measured by nanoparticle tracking analysis (NTA) ([73]Fig. 1b). The SDS-PAGE analysis was conducted to evaluate the protein profile of three kinds of BEVs, and the results indicated the obvious different protein components originated from the parent bacteria ([74]Fig. 1c). Fig. 1. [75]Fig. 1 [76]Open in a new tab Characterization of BEVs and antitumor vaccination efficacy of Akk-OMV. (a) TEM images of BEVs (scale bar: 100 nm). (b) Average particle sizes and zeta potentials of BEVs (A, Akk-OMV; B, Bifi-CMV; D, DH5α-OMV) measured by nanoparticle tracking analysis (n = 3). (c) SDS-PAGE analysis of protein components in BEVs (M, Marker; A, Akk-OMV; B, Bifi-CMV; D, DH5α-OMV). (d) Experimental design: BALB/c mice were subcutaneously injected at the tail base with PBS, Akk-OMV, Bifi-CMV or DH5α-OMV (10 μg) at 3 day-intervals for 5 times, and then inoculated with 2 × 10^5 4T1 tumor cells on day 0. Mice were euthanized on day 28 for RNA-sequencing analysis. (e) Tumor growth kinetics after vaccination of different BEVs (n = 6). (f) Transcriptome heatmap of DEGs in tumors and spleens of mice vaccinated with different BEVs and PBS (n = 5). GO pathway enrichment analysis in (g) tumors and (h) spleens of Akk-OMV versus the PBS control. (i) KEGG pathway enrichment analysis in spleens in Akk-OMV, Bifi-OMV and DH5α-OMV compared to the PBS control. Data are presented as mean ± SD. Statistical significance was calculated via one-way ANOVA with Tukey multiple comparisons test. **p < 0.01 versus corresponding control group. To investigate the prophylactic vaccination capacity of different BEVs, Akk-OMV, Bifi-CMV and DH5α-OMV were subcutaneously injected to BALB/c mice at 3 day-intervals for 5 times before 4T1 tumor inoculation ([77]Fig. 1d). Akk-OMV vaccination significantly alleviated the tumor growth, while the effects of Bifi-CMV and DH5α-OMV were not obvious, which indicated the durable tumor-preventing capacity of Akk-OMV ([78]Fig. 1e). We next conducted RNA-sequencing to profile the differentially expressed genes (DEGs) in the transcriptomes of tumors and spleens from vaccinated mice. The DEGs analysis showed that BEVs vaccination altered various gene expressions in spleens and tumors compared to PBS control, of which Akk-OMV caused the most pronounced changes ([79]Figs. S2a–d, and [80]Figs. S3a–d). Kyoto Encyclopedia of Genes and Genomes (KEGG) annotation analysis revealed that the DEGs in spleens and tumors caused by Akk-OMV vaccination were mainly categorized in the immune system ([81]Fig. S2e, and [82]Fig. S3e). According to the gene expression heatmap in tumors and spleens among different groups ([83]Fig. 1f) and corresponding Gene Ontology (GO) pathway enrichment analysis, the tumors with Akk-OMV vaccination revealed the important role of major pathways such as regulation of B cell activation, production of molecular mediator of immune response and lymphocyte mediated immunity ([84]Fig. 1g), and DEGs in Akk-OMV vaccinated spleens were enriched in the major pathways including T cell differentiation, regulation of T cell activation and regulation of cell-cell adhesion ([85]Fig. 1h). In addition, KEGG pathway enrichment analysis of splenic DEGs indicated the predominant role of immune-related pathways after Akk-OMV vaccination, including T cell receptor signaling pathway, Th17 cell differentiation, Th1 and Th2 cell differentiation, and PD-L1 expression and PD-1 checkpoint pathway in cancer ([86]Fig. 1i). Collectively, these transcriptome analysis results explained the immunomodulatory effect of Akk-OMV vaccination, which lastly exhibited the tumor-preventing capacity. To verify the antitumor immune response induced by Akk-OMV, the near-infrared fluorescent dye DiR-labeled Akk-OMV was subcutaneously injected at the tail base of BALB/c mice, and the fluorescence signals of major organs indicated that Akk-OMV was mainly distributed in liver as analyzed by the In-vivo Imaging System (IVIS). In addition, the signals of inguinal lymph nodes were observed at different time points of 24, 36 and 48 h, revealing the migration and retention of Akk-OMV into lymph node ([87]Figs. S4a–c). We next performed 5-round subcutaneous injection of Akk-OMV to BALB/c mice after 4T1 tumor inoculation ([88]Fig. S5a). Flow cytometry analysis proved that Akk-OMV treatment promoted the expression of maturation markers on DCs (CD80^+CD86^+ in CD11c^+) in lymph node ([89]Fig. S5b). Moreover, Akk-OMV treatment caused a significant increase of CTLs (CD8^+ in CD3^+) and effector memory T cells (T[EM,] CD44^+CD62L^− in CD3^+CD8^+) in splenocytes ([90]Figs. S5c–d), and intratumoral infiltration of CTLs was upregulated as well ([91]Fig. S5e). Given the robust association of Akkermansia muciniphila with favorable antitumor responses of ICIs [[92][14], [93][15], [94][16], [95][17]], we further investigated the mediating role of Akk-OMV in the enhancement of parent bacterium to checkpoint inhibitor immunotherapies. 5-round subcutaneous injection of Akk-OMV in 4T1 tumor-bearing mice on day 0, 3, 6, 9 and 12 was conducted for the purpose of boosting the CTLs infiltration in tumor, which was combined with 3-round intraperitoneal injection of PD-L1 antibody on day 6, 9 and 12 ([96]Fig. S6a). The results showed that Akk-OMV or PD-L1 antibody treatment alone did not significantly reduced the tumor volume comparing with the control group at the endpoint of tumor growth on day 21, while the tumor volume was significantly reduced after combination treatment of Akk-OMV and PD-L1 antibody on day 15, 18 and 21. Moreover, there was a significant difference of tumor volume between Akk-OMV and combination treatment on day 15 ([97]Fig. S6b). The flow cytometry analysis confirmed that both Akk-OMV and combination treatment significantly increased the intratumoral infiltration of CTLs comparing with the control group, while PD-L1 antibody treatment did not ([98]Fig. S6c). Taken together, these results demonstrated that Akk-OMV activated DC maturation and CTL response, and exhibited the synergistically therapeutic potential with PD-L1 antibody. 3.2. Akk-OMV exhibits a favorable safety profile Owing to the high density of pathogen-associated molecular patterns (PAMPs), bacteria OMVs have been developed as versatilely self-adjuvanting vehicles to improve the immunogenicity of tumor antigens and achieve the therapeutic cancer vaccination [[99]26,[100]31]. However, the lipopolysaccharide and other toxins in OMVs become the major cause of severe toxic effects including cytokine storm and sepsis, hampering the further clinical applications [[101]32]. Given the safety and efficacy of Akkermansia muciniphila in various mouse models and in human trials [[102]33,[103]34], we assumed the favorable safety profile of Akk-OMV. Since Escherichia coli-derived OMVs were previously found to cause sever toxicity in vivo [[104]35], DH5α-OMV was chosen as the positive control. After single-dose intraperitoneal injection of Akk-OMV and DH5α-OMV to BALB/c mice, the serum and major organs were collected for evaluating the liver and renal function and organ histological changes at 8 h post injection ([105]Fig. S7a). There were no obvious histological changes of major organs and statistical differences of liver and renal function indexes between the groups of Akk-OMV and the PBS control, while the serum concentrations of liver enzymes alanine transaminase (ALT), aspartate aminotransferase (AST) and renal function marker blood urea nitrogen (BUN) in DH5α-OMV vaccinated mice were significantly increased ([106]Figs. S7b–c). We next performed intraperitoneal injection of Akk-OMV and DH5α-OMV to BALB/c mice every 3 days for 3 times ([107]Fig. 2a). DH5α-OMV caused 50 % death rate of mice at the endpoint on day 14 ([108]Fig. 2b), and remaining mice displayed visible splenomegaly ([109]Fig. 2c). There were significant differences in serum concentrations of liver albumin (ALB) and total bile acid (TBA) and renal BUN between the groups of DH5α-OMV and control ([110]Fig. 2d), and hematoxylin & eosin (H&E) staining showed obvious lymphatic nodule disappear and lymphocyte diffuse hyperplasia ([111]Fig. 2e). In addition, Akk-OMV and DH5α-OMV were subcutaneously injected to C57BL/6 mice at 3 day-intervals for 3 times ([112]Fig. 2f). DH5α-OMV treatment resulted in the obvious skin damage at the injection site, while mice received Akk-OMV treatment remained similar with the PBS control group ([113]Fig. 2g). H&E staining also indicated the disappear of lymphatic nodules and diffuse hyperplasia of lymphocytes in DH5α-OMV treated mice ([114]Fig. 2h), and the serum concentrations of liver alkaline phosphatase (ALP), ALB, direct bilirubin (DBIL) and total bilirubin (TBIL) and renal BUN and creatinine (CREA) were significantly increased compared to the control group ([115]Fig. 2i). On the contrary, the major organ histomorphology and biochemical indexes of mice received 3-round intraperitoneal or subcutaneous injection of Akk-OMV remained similar with the control group. Collectively, these results demonstrated the favorable safety profile of Akk-OMV, revealing the further application potential for adjuvant immunotherapy strategies. Fig. 2. [116]Fig. 2 [117]Open in a new tab Akk-OMV exhibited a favorable safety profile over E. Coli DH5α-OMV. (a) Experimental design for toxicity evaluation upon 3-round intraperitoneal injection. BALB/c mice were injected with Akk-OMV and DH5α-OMV (20 μg) on day 0, 3 and 6, and were euthanized on day 14. (b) Survival rate of each group at the end point (n = 6). (c) Spleen images and weights at the endpoint (n = 3 and 6). (d) Serum concentrations of hepatic and renal function indexes (n = 3 and 6). (e) Representative H&E staining images of spleens (Scale bar: upper images 1 mm, lower images 250 μm). (f) Experimental design for toxicity evaluation upon subcutaneous injection. C57BL/6 mice were injected at the tail base with Akk-OMV and DH5α-OMV (20 μg) on day 0, 3, 6, and were euthanized on day 9. (g) Skin damage images of different groups at endpoint (n = 7). The skin damage at the injection site was marked by red circle. (h) Representative H&E staining images of spleens (Scale bar: left images 1 mm, right images 250 μm). (i) Serum concentrations of hepatic and renal function indexes (n = 7). Data are presented as mean ± SD. Statistical significance was calculated via one-way ANOVA with Tukey multiple comparisons test. *p < 0.05, **p < 0.01, ***p < 0.001 versus corresponding control group. (For interpretation of the references to color in this figure legend, the reader is referred