Microbiota-derived self-adjuvanting exopolysaccharide-based codelivery system for potent cancer immunotherapy Authors: Xuemei Bao, Jiamin Shi, Haochi Zhang, Chunhe Li, Yanchen Liang, Na Pan, Yubing Fu, Xing Su, Wan Wang, Ting Wang, Lipeng Zhang, Xiao Wang Published: 2026 ABSTRACT AbstractCancer vaccines hold great promise by eliciting antigen-specific T cell responses against tumors; however, their efficacy remains limited by poor coordination between antigen delivery and innate immune activation. Here, microbiota-derived metabolites (exopolysaccharides, EPSs) are presented as a self-adjuvanting codelivery system that simultaneously delivers the Toll-like receptor 7/8 (TLR7/8) agonist R848 and ovalbumin (OVA). This bi-adjuvant nanovaccine (EPS@R848/OVA) rapidly accumulates in draining lymph nodes (dLNs) as early as 6 h after administration and significantly enhances the uptake and activation of CD103+CD11b− conventional dendritic cells (cDCs), thereby facilitating efficient antigen cross-presentation and eliciting robust antigen-specific cytotoxic CD8+ T lymphocyte (CTL) responses. Importantly, the nanovaccine markedly inhibits primary and metastatic melanoma growth and remodels the tumor microenvironment (TME) by increasing CD8+ T cell infiltration. Furthermore, combining the nanovaccine with the anti-PD-1 antibody yields potent synergistic antitumor efficacy and prolonged survival in melanoma models. Mechanistically, it markedly promotes DC maturation and type I interferon production through synergistic activation of the Dectin-2-Syk-CARD9 and TLR7-MyD88-IRF7 pathways. Collectively, this work demonstrates that the microbiota-derived EPS-based codelivery system offers a simple and effective strategy for developing potent cancer vaccines.Keywords: Cancer immunotherapy, Exopolysaccharides, Self-assembly, Codelivery, Conventional dendritic cells, CD8+ T cell responses Introduction Cancer immunotherapy has emerged as a transformative treatment paradigm that harnesses the host immune system to recognize and eliminate malignant cells, achieving substantial progress in oncology [1,2]. Several cancer immunotherapy approaches, including immune checkpoint blockade, cancer vaccines, chimeric antigen receptor T (CAR-T) cell therapy, and cytokine therapy, have been clinically tested [3,4]. Among these, cancer vaccines aim to elicit tumor-specific T cell immunity and constitute a promising treatment strategy [5,6]. However, the clinical efficacy of cancer vaccines remains suboptimal, as they generate limited CD8+ T cell responses [7,8]. This insufficient cellular immunity arises from inefficient delivery of tumor antigens and suboptimal cross-presentation of antigens by DCs to CD8+ T cells, as well as from the limited capacity of the FDA-approved adjuvants to adequately activate DCs and provide the necessary costimulatory signals for effective CD8+ T cell priming [9,10]. Increasing evidence suggests that the codelivery of adjuvants and tumor antigens within a single nanocarrier to prepare nanoparticle (NP)-based vaccines (nanovaccines) represents a promising strategy for the development of potent cancer vaccines [[11], [12], [13]]. However, efficiently co-encapsulating adjuvants and antigens with distinct physicochemical properties within a single nanocarrier, and precisely delivering them to lymphoid tissues and DCs to promote robust antigen-specific CD8+ T cell responses, remains a formidable obstacle. Extensive research has focused on liposomes, poly(lactic-co-glycolic acid) (PLGA) and poly(ethylene oxide)-block-poly(d,l-lactic acid) systems [[14], [15], [16], [17]]. Nevertheless, the clinical translation of these delivery platforms suffers from the limitation of suboptimal safety, the need for additional targeting modifications, and, importantly, insufficient intrinsic adjuvant activity, which limits their ability to coordinate antigen delivery with innate immune activation [[18], [19], [20], [21]]. Consequently, the development of efficient self-adjuvanting delivery platforms provides a simple and effective strategy to address these challenges. Wang et al. have developed a cholesterol azetidine derivative-based nanoparticle that specifically induces innate immune activation via the TLR9 pathway and promotes tumor-specific CD8 T cell responses [22]. Gong et al. also developed a palmitic acid-modified nanoemulsion with intrinsic self-adjuvanting activity, achieving cascade targeting delivery and potent activation of cellular immunity [11]. However, such platforms remain scarce, and their fabrication is often cumbersome and compositionally complex, limiting the rapid application and clinical feasibility of cancer vaccines. Emerging evidence indicates that microbiota-derived metabolites can enhance antitumor immunity and promote immune checkpoint blockade therapy (ICBT) [23,24]. Exopolysaccharides (EPSs) are high-molecular-weight sugar-based metabolites produced by microorganisms [25]. Microbiota-derived EPSs have emerged as promising vaccine adjuvants owing to structural diversity, biodegradability, biocompatibility, and natural affinity for pattern recognition receptors (PRRs) [[26], [27], [28], [29]]. Recent studies have demonstrated that orally administered Bifidobacteria-derived EPSs induce DC activation through the TLR2-MyD88 pathway and enhance antigen-specific Th1 immune responses when co-administered with antigens, thereby acting as adjuvants in cancer immunotherapy [30]. Interestingly, our previous work has shown that Lactobacillus plantarum-derived amphiphilic EPSs serve as a self-adjuvanting delivery platform that induces T cell-mediated antitumor immunity [31]. These findings indicate that microbiota-derived EPSs may enable the integration of antigen delivery with innate immune activation. To date, however, although several encouraging findings have emerged from studies using microbiota-derived EPSs in cancer immunotherapy, systematic investigations of their immunological mechanisms and combination therapy strategies remain lacking. Here, we present microbiota-derived amphiphilic EPSs as a class of self-adjuvanting codelivery systems, with EPS233 derived from Lacticaseibacillus paracasei emerging as the leading candidate through systematic screening. Moreover, we screened various TLR agonists and reported our finding that EPS synergized with the TLR7/8 agonist R848 to induce potent antitumor immunity. Toward systemic cancer immunotherapy, we developed a bi-adjuvant nanovaccine (EPS@R848/OVA) for the codelivery of R848 and OVA, which enabled efficient loading of both the antigen and the small-molecule agonist. We found that EPS@R848/OVA activates innate immunity in DCs through the Dectin-2 and TLR7/8-mediated signaling pathways, which synergistically induce type I interferon and proinflammatory cytokine production, thereby facilitating efficient antigen cross-presentation. Moreover, the nanovaccine efficiently improved antigen delivery to dLNs and CD103+CD11b− cDCs, leading to augmented antigen-specific CTL responses. This nanovaccine system achieved potent antitumor efficacy in both orthotopic and metastatic melanoma models and significantly enhanced responsiveness to anti-PD-1 antibody, resulting in strong synergistic antitumor effects. Overall, we developed a microbiota-derived EPS-based codelivery system to boost antitumor immunity for potent cancer immunotherapy (Fig. 1). Schematic illustration of a microbiota-derived EPS-based codelivery system for potent cancer immunotherapy. We prepare a microbiota-derived EPS-assisted bi-adjuvant nanovaccine, EPS@R848/OVA, in which EPS simultaneously delivers OVA and R848 to dLNs with subsequent localization to CD103+CD11b− cDCs. The nanovaccine specifically activates the Dectin-2 and TLR7 pathways to induce IFN-I production, thereby enhancing antigen cross-presentation and eliciting potent antigen-specific CD8+ and Th1 T cell responses. The nanovaccine markedly inhibits tumor growth and prolongs survival in both orthotopic and metastatic melanoma models, while exhibiting favorable systemic biosafety. Moreover, a combination of the nanovaccine with aPD-1 yields potent synergistic antitumor efficacy and prolongs survival in melanoma models. Materials and methods Materials The exopolysaccharide EPS233 was isolated from Lacticaseibacillus paracasei WXD233, which was obtained from our laboratory strain library. Pam3CSK4, poly(I:C), MPLA, resiquimod (R848), and CpG ODN (ODN 1826) were purchased from MedChemExpress (MCE). Model antigen OVA and FITC-labeled OVA were purchased from Sigma-Aldrich and Bersee. LinKine™ AbFluor™ 647 Labeling Kit was purchased from Abbkine Scientific Co., Ltd. OVA257-264 peptide (SIINFEKL) and GP33 (KAVYNFATM) peptide were produced by Genscript Biotech Co., Ltd. and purified by HPLC to a purity exceeding 95%. Anti-mouse PD-1 (CD279, clone RMP1-14) was obtained from Leinco Technologies. Cells Bone marrow cells were aseptically separated from the femurs and tibias and cultured in complete DC medium. The medium comprised 15% FBS (ExCell Biotech, FSD500), 1.5% penicillin-streptomycin (Servicebio, G4003), 55 μM β-mercaptoethanol (Gibco, 2466010), 19 ng mL−1 GM-CSF (Novoprotein, CK02), and 10 ng mL−1 IL-4 (Novoprotein, CK74). Cultures were incubated at 37°C in 5% CO2 environment. Fresh medium was added on day 3, and BMDCs were ready for experimental use on day 7 [32]. DC2.4 cells were maintained at 37°C in 5% CO2 in RPMI medium (Servicebio, G4535) containing 15% FBS and 1.5% penicillin-streptomycin. OT-I and OT-II T cells were separated from spleens of female OT-I and OT-II transgenic mice, and cultured in RPMI-1640 complete medium with 55 μM β-mercaptoethanol and 4 mM HEPES (Servicebio, G4210) at 37°C with 5% CO2. B16F10-OVA and B16F10-GP33 cells were maintained in DMEM (Servicebio, G4515) containing 15% FBS, 1% penicillin and streptomycin and 1% NEAA (Servicebio, G4219) at 37°C. EL4 cells were maintained in DMEM containing 15% FBS, 1.5% penicillin and streptomycin, and 1.5% NEAA at 37°C in 5% CO2. Animals Healthy female wild-type (WT) C57BL/6 mice (6-8 weeks old) were purchased from Beijing Speifu Biotechnology Co., Ltd. (Beijing, China) and maintained under specific pathogen-free (SPF) conditions. Healthy female OT-I and OT-II transgenic mice, as well as Rag−/− and CD8−/− mice, were maintained in the laboratory under SPF conditions. For the experiments presented in this study, a total of approximately 780 female WT C57BL/6 mice were used. In addition, 25 female OT-I transgenic mice, 18 female OT-II transgenic mice, 30 Rag−/− mice, and 30 CD8−/− mice were included. Mice were randomly assigned to experimental groups. All animal-related procedures were conducted in accordance with protocols approved by the Ethics Committee of Inner Mongolia University (IMU-Mouse-2024-149). Isolation of exopolysaccharides from microbiota Twenty strains of lactobacilli were selected from our laboratory strain library for screening. Bacteria were inoculated into MRS broth and incubated at 37°C for 48 h. The culture was centrifuged and the supernatant was collected. Proteins were removed by adding trichloroacetic acid, followed by precipitation with ethanol at a final concentration of 80% (v/v) overnight. The precipitate was collected by centrifugation, dialyzed against water using a dialysis membrane (8000–12,000 Da) for 72 h, and lyophilized to obtain the crude polysaccharide. Purification of exopolysaccharides Exopolysaccharides were dissolved in DI water and purified by DEAE Sepharose Fast Flow chromatography (GE, USA; 2.6 × 40 cm column) and CL-6B size exclusion chromatography (GE, USA; 1.6 × 100 cm column), followed by dialysis against distilled water and lyophilization, yielding homogeneous exopolysaccharides fractions. Physicochemical characterization of EPS233 The FTIR was recorded using a spectrometer (Bruker, VERTEX70). The morphological features of the purified exopolysaccharide powder were examined by SEM (S-4800, Hitachi, Tokyo, Japan). The EPS233 nanoparticles were stained with 2 wt% phosphotungstic acid negative stain solution (Solarbio, G1870) and visualized with a cryo-TEM (Talos F200C G2, Thermo Fisher). The average molecular weight of EPS233 was assessed by HPGPC. Monosaccharide composition was analyzed by Ion chromatography (IC). The linkage types of EPS233 were analyzed by Shimadzu QP 2010 gas chromatography-mass spectrometry (GC-MS). The PMAAs of EPS233 were identified by comparing their mass spectra with the standard spectra available from the Complex Carbohydrate Research Center (CCRC, University of Georgia, Athens, GA, USA) spectral database. The nuclear magnetic resonance (NMR) spectra were recorded using a Bruker Advance 500 MHz spectrometer (Bruker, Rheinstetten, Germany). Preparation and characterization of EPS@R848/OVA EPS@R848/OVA was prepared by a LINZ-8A microinjection pump. EPS (60 mg) was solubilized in 20 mL of DI water and stirred at 220 rpm. R848 (15 mg, from a 100 mg mL−1 stock solution in DMSO) was diluted in 20 mL of DI water, and OVA (6 mg) was solubilized separately in 20 mL of DI water. The R848 and OVA solutions were added dropwise into the EPS solution at a flow rate of 2 mL min−1 under continuous stirring (200 rpm). The mixture was then dialyzed against deionized water at 4°C for 72 h using a 50 kDa MWCO membrane (Solarbio, YA1077) to remove free cargo and residual solvent, and the nanoparticles were collected by lyophilization. EPS concentration was determined by the phenol-sulfuric acid method and adjusted to 10 mg mL−1 before further use. For OVA encapsulation analysis, the nanoparticle suspension was centrifuged using ultrafiltration tubes with a 50 kDa MWCO, and the solution at the bottom was obtained to measure free OVA. The amount of free OVA was measured by an Enhanced BCA Protein Assay Kit (Beijing Solarbio Biotechnology Co., Ltd.). The drug loading (DL) and encapsulation efficiency (EE) were calculated using the following equations: DL (%) = [(Initial OVA - Free OVA)/total weight of nanoparticles] × 100% EE (%) = [(Initial OVA - Free OVA)/Initial OVA] × 100% The R848 concentration in EPS@R848/OVA was measured using a Shimadzu LC-20AT HPLC system. Chromatographic separation was performed on a C18 column with acetonitrile and water containing 10% glacial acetic acid. The gradient was set as follows: 2:98 (v/v) from 0 to 2 min, 100:0 (v/v) from 2 to 12 min, and 2:98 (v/v) from 12 to 25 min. The measurement was carried out at 320 nm. The release rates of R848 and OVA from EPS@R848/OVA nanoparticles were determined using HPLC and the BCA assay, respectively. EPS@R848/OVA samples were placed in dialysis bags with 2 kDa or 50 kDa MWCO membranes and immersed in PBS at varying pH levels (pH 5.0 to simulate lysosomal conditions and pH 7.2 to simulate physiological conditions) at 37°C, with agitation at 300 rpm. The solution (0.2 mL) was sampled at predefined time points (0, 6, 24, 48, 72, 96, and 120 h) and immediately replenished with equal-volume PBS. For visualization, the nanoparticles were negatively stained with 2 wt% phosphotungstic acid and observed under TEM. Nanoparticle size and zeta potential were measured using a particle size analyzer (Malvern, 90PLUS). For stability characterization, EPS@R848/OVA was suspended in PBS containing 10% FBS at 5 mg mL−1 and maintained at approximately 25°C or 4°C. Particle size was measured weekly employing dynamic light scattering. Formulations were sterilized using a 0.22 μm filter before conducting in vitro and in vivo experiments. Molecular dynamics simulation MD simulations were conducted with the AMBER software suite. The EPS structure was built with GLYCAM-Web Carbohydrate Builder and exported in AMBER-compatible formats. Molecular docking between EPS and R848 was performed using AutoDockTools to evaluate binding affinity. The crystal structure of OVA (PDB ID: 1UHG) was retrieved. For the multi-component EPS@R848-OVA system, initial configurations were generated using PACKMOL. The complex was placed in a periodic truncated octahedral box filled with TIP3P water (10 Å padding) and balanced with appropriate counterions. OVA was described by the ff14SB force field, EPS by GLYCAM06, and R848 by GAFF2 parameters with AM1-BCC partial charges. A 100-ns production run was performed using a 2-fs integration step. Hydrogen-containing bonds were restrained via the SHAKE algorithm. Electrostatic interactions were treated with the particle mesh Ewald approach, applying a 10 Å cutoff for nonbonded interactions. Trajectory analysis with CPPTRAJ assessed structural stability (RMSD), conformational compactness (Rg), solvent exposure (SASA), intermolecular hydrogen bonds, and interaction energy decomposition (van der Waals and electrostatic) between EPS, R848 and OVA. Assessment of systemic toxicity induced by R848 Mice were subcutaneously injected into the inner thigh with either R848+OVA or EPS@R848/OVA (7.68 mg kg−1). Blood was collected from the tail vein after 6 h. Cytokines in serum were measured using ELISA and cytokine bead array kits. ELISA kits for mouse IL-12p70 (PM1270), IL-1β (DY401), IL-6 (DY406), and TNF-α (PMTA00B) were sourced from Peprotech (R&D). We evaluated systemic toxicity by tracking body weight changes in mice at multiple intervals post-administration. Cell viability assessment Cell cytotoxicity of EPS@R848/OVA was evaluated using a CCK-8 assay. BMDCs and L929 cells were treated with varying concentrations of EPS@R848/OVA (5, 10, 20, 50, 70, or 100 μg mL−1) for 24 h, and cell viability was determined according to the manufacturer's instructions. DC uptake and lysosomal escape To evaluate the internalization efficiency and lysosomal escape of EPS@R848/OVA, DC2.4 cells were exposed to FITC-labeled OVA (1.5 μM FITC) and EPS@R848/OVA (50 μg mL−1) in complete medium for 2, 4, 6, 8, and 12 h. Cells were then collected and washed three times. Lysosomes were labeled with LysoTracker Red (Solarbio, L8010) 25 min before the end of incubation. Then cells washed three times and incubated with DAPI solution (Solarbio, C0060, 0.5 μg mL−1) for 10 min. Final mounting was performed by sealing coverslips on microscope slides using antifade mounting medium. Fluorescence was imaged using a Nikon A1R confocal laser scanning microscope (CLSM, Japan). PCC between lysosomes and FITC-conjugated formulations was quantified utilizing ImageJ software. Characterization of BMDC maturation BMDCs were obtained in accordance with the method mentioned earlier. For cytokine secretion and co-stimulatory expression analysis, BMDCs were enumerated and cultured in complete DC medium at 37°C in a 5% CO2 atmosphere. Fresh medium was introduced on the third day of culture, and the BMDCs were ready for experimental use on day 7. BMDCs were treated with OVA, EPS@OVA, R848+OVA, or EPS@R848/OVA (50 μg mL−1) for 24 h. Subsequently, the supernatant was analyzed using ELISA kits (R&D) to measure IL-6, TNF-α, IFN-β (MIFNB0), and IL-12p70. BMDCs were treated with an Fc-blocker prior to staining with the following markers: CD11c, CD40, CD80, CD86, MHC-II, MHC-I, and SIINFEKL-H2-Kb. Data were acquired on a NovoCyte flow cytometer (ACEA Biosciences, USA) and analyzed using FlowJo software version 10 (Treestar). Antibodies used for flow cytometry are listed in Table S1. RNA-sequencing (RNA-seq) analysis BMDCs were incubated with PBS or EPS@R848/OVA (50 μg mL−1) for 12 h, and then harvested using TRIzol. The samples were then subjected to RNA-seq analysis using the DNBSEQ-T7 platform. Western blot analysis BMDCs were plated at 4 × 106 cells per six-well plate and incubated overnight and then treated with OVA, EPS@OVA, R848+OVA, or EPS@R848/OVA (50 μg mL−1) for 8 h. Then BMDCs were treated with cell lysis buffer (Beyotime Biotechnology, P0013) supplemented with PMSF (ST506) and protease and phosphatase inhibitor cocktail (P1046) to collect the supernatant. Samples were combined with 6×SDS-PAGE buffer and heated at 100°C for 8 min for denaturation. Equal protein loading was ensured by prior quantification, and the samples were applied to an 8-12% SDS-PAGE gel. Electrophoresis was conducted at 120 V for 60 min. The proteins were then transferred to a PVDF membrane at 400 mA on ice. The membrane was blocked with rapid blocking buffer (Yeasen) for 60 min at 37°C. The membrane was incubated with primary antibodies, including β-actin (TransGen Biotech, HC201), Dectin-2 (Yamay Bio, P013356), CARD9 (Proteintech, 85749-1-RR), SYK (Proteintech, 14858-1-AP), P-SYK (Proteintech, 83331-1-RR), TLR7 (Yamay Bio, P100315), MyD88 (Yamay Bio, P012343), IRF7 (GenuIN Biotech, 1266), and P-IRF7 (CST, 24129). Following triple rinses with TBST, the membrane was exposed to secondary antibodies for 2 h at 25°C. Chemiluminescence detection and analysis were then performed. Whole-mouse and lymph node fluorescence imaging Fluorescence imaging was conducted on both the entire mouse and its lymph nodes. Mice were injected subcutaneously into the inner thigh with 50 μL PBS containing either AF647-labeled OVA or AF647-labeled EPS@R848/OVA (7.68 mg kg−1). AF647 signals were measured in vivo using a Berthold NightOWL in vivo Imaging System (BERTHOLD, LB983) from day 0 to day 5. In vivo fluorescence at the injection site was expressed relative to initial values recorded at 30 min (day 0). Inguinal dLNs and other major organs were collected at different time points, and overall fluorescence signals were analyzed using a Lago spectral imaging system with a 640 nm excitation wavelength and a 690 nm emission wavelength. Cellular internalization study Mice were injected subcutaneously into the inner thigh with 50 μL PBS containing either AF647-labeled OVA or AF647-labeled EPS@R848/OVA (7.68 mg kg−1). Mice received two injections at a 14-day interval. Inguinal dLNs were harvested on day 21, dissociated using a grinding rod, and the cells were passed through 30 μm strainers to achieve single-cell suspension. Cells were treated with an Fc-blocker before staining with markers including CD103, CD11c, MHC-II, CD11b, Ly6C, B220, Ly6G, and CD86. Data were acquired on a NovoCyte flow cytometer and analyzed using FlowJo software. In vitro assay for OVA-specific T-cell proliferation Cocultures employed BMDCs as APCs alongside CFSE-labeled OT-I or OT-II T cells, sorted via negative selection from spleens and dLNs of transgenic mice. The selection utilized Dynabeads™ FlowComp™ mouse CD8+ or CD4+ T cell kits (Invitrogen, 11462D; 11461D). T cells were enriched and labeled with 0.5 μM CFSE (BioLegend, 423801) in PBS with 0.2% BSA for 15 min at 37°C. The reaction was halted using FBS, followed by three washes of the cells with complete RPMI medium. In the experiment, 1 × 104 BMDCs were incubated in RPMI, exposed to OVA, EPS@OVA, R848+OVA, or EPS@R848/OVA (50 μg mL−1), and maintained at 37°C. After 3 h, BMDCs were washed and then cocultured with OT-I or OT-II cells at a 1:9 APC to T cell ratio in 250 μL of DMEM. T cell proliferation was evaluated after 3 days using flow cytometry (FCM), with CFSE dilution serving as a proliferation marker. Evaluation of the killing efficacy of CTL cells to EL4 cells in vitro Total splenocytes harvested from OT-I mice were treated with OVA, EPS@OVA, R848+OVA, or EPS@R848/OVA (50 μg mL−1) for 24 h. The complete RPMI medium was supplemented with 20 ng mL−1 IL-2 (BioLegend, 503805). 1 × 106 CD8+ T cells per group were collected to assess their antigen-specific cytotoxicity against EL4 cells after 3 days. EL4 target cells were counted and split into two tubes, one tube was stained with 0.05 μL of CFSE (CFSE-Low), and the other was stained with 0.5 μL of CFSE and incubated with 6 μg of OVA257-264 peptide (CFSE-High) for 45 min at 37°C. Following incubation, cells were resuspended to a concentration of 2 × 106 cells mL−1 and combined in equal proportions (CFSE-High to CFSE-Low). Cytotoxic activity of primed CD8+ T cells was evaluated against tumor targets across different E:T ratios using flow cytometry after 6 h of incubation. Phenotypic analysis of T cells and B cells in dLNs Mice were injected subcutaneously into the inner thigh with 50 μL PBS containing either free OVA, EPS@OVA, R848+OVA, or EPS@R848/OVA (7.68 mg kg−1). Mice were primed and boosted 14 days later. Single-cell suspensions were prepared from inguinal dLNs on day 21. Then cells were reconstituted with FACS buffer (PBS with 2% FBS) and pretreated with Fc-blocker before staining with various markers. These markers included CD8, CD4, CD44, B220, GL7, CD95, CD138, IgD, CXCR5, and PD-1. Cells were mixed with FACS buffer following staining. For intracellular cytokine detection, cell suspensions were distributed into culture plates and stimulated for 5 h in RPMI medium supplemented with GolgiPlug (BD, 554715), PMA (Sigma, P8139), and ionomycin (Yeasen, 50401ES03). The cells were rinsed and subsequently incubated with fixation/permeabilization buffer for 25 min on ice. After washing with BD Perm/Wash buffer (BD Bioscience), cells were labeled with antibodies for 90 min on ice including IFN-γ, TNF-α, and IL-17A. Following staining, the cells underwent two washes with BD Perm/Wash buffer. Cells were mixed with FACS buffer following staining. Data were acquired on a NovoCyte flow cytometer and analyzed using FlowJo software. Assessment of antigen-specific T cell immunity in dLNs Adoptive transfer of CD8+ or CD4+ T cells were sorted from spleens and dLNs of OT-I or OT-II mice. After washing, CD8+ T cells were counted and intravenously transferred into naive CD8−/− recipient mice (1.0 × 106 cells per mouse), CD4+ T cells were intravenously transferred into naive Rag−/− recipient mice (1.0 × 106 cells per mouse). Following a one-day rest post-adoptive transfer, mice were injected subcutaneously with the EPS@R848/OVA (7.68 mg kg−1) in the inner thigh. Single-cell suspensions were prepared from excised iLNs. For flow cytometry assay of OVA-specific T cell responses, cells were resuspended in FACS buffer and using Fc-blocker before staining with CD8, CD4, and CD44. For intracellular cytokine detection, cell suspensions were distributed into culture plates and stimulated for 5 h in RPMI medium supplemented with GolgiPlug, PMA, and ionomycin. The cells were subsequently incubated with fixation/permeabilization buffer. After washing with BD Perm/Wash buffer, the cell preparations were dyed with IFN-γ and TNF-α markers. Data were acquired on a NovoCyte flow cytometer and analyzed using FlowJo software. Assessment of antigen-specific antibody levels Mice were injected subcutaneously into the inner thigh with 50 μL PBS containing either OVA, EPS@OVA, R848+OVA, or EPS@R848/OVA (7.68 mg kg−1) at week 0, followed by a booster at week 2. Blood samples were obtained from the tail vein and assessed via ELISA-based antibody profiling. After coating with OVA (4 μg mL−1 in PBS), plates were sequentially blocked (1.5% casein) and probed with serially diluted serum samples. Following an additional wash, plates were treated with HRP-conjugated anti-mouse IgG (Proteintech, SA00001-1) or IgG2b (Proteintech, SA00012-3) as the secondary antibody, then processed using TMB substrate (GE Healthcare). Absorbance at 450 nm was determined using a microplate spectrophotometer, and data analysis was conducted with SkanIt software. The antibody titer was calculated as the inverse of the maximum dilution factor yielding an OD450 value above the threshold. In vivo prophylactic efficacy Mice were injected intraperitoneally with either OVA, EPS@OVA, R848+OVA, or EPS@R848/OVA (7.68 mg kg−1). Mice received three administration doses at weekly intervals. The mice were euthanized 21 days post-inoculation, and both the iLNs and spleens were collected. Then the single-cell suspensions were blocked with an Fc-blocker before staining with markers, including CD8, CD4, CD44 and CD62L (MEL-14, BD eBioscience). For intracellular cytokine detection, the cell suspensions were stained with IFN-γ and TNF-α. Data were acquired on a NovoCyte flow cytometer and analyzed using FlowJo software. Mice were inoculated subcutaneously with 1 × 106 B16F10-OVA cells per mouse. Tumor progression and animal health were assessed by tracking growth rates, incidence, body weight changes, and survival. Volume was calculated as V = L × W × W × 0.5, where V represents tumor volume and L and W denote tumor length and width. Twenty-two days post-tumor injection, mice were euthanized for tumor weighing and photography. Mice were euthanized and final data recorded when the tumor size surpassed 1500 mm3 or the maximum dimension surpassed 1.5 cm. In vivo therapeutic efficacy In the therapeutic subcutaneous B16F10-OVA tumor model, mice were inoculated subcutaneously with 1 × 106 B16F10-OVA cells each mouse. Treatment was initiated when tumors reached approximately 30-50 mm3. Mice were immunized intraperitoneally with 200 μL PBS containing either OVA, R848+OVA, EPS@OVA, or EPS@R848/OVA (7.68 mg kg−1) on days 5, 11, and 17. In the therapeutic subcutaneous B16F10-GP33 tumor model, mice were inoculated in the same manner, and immunized with 200 μL PBS containing either GP33, R848+GP33, EPS@GP33, or EPS@R848/GP33 (7.68 mg kg−1) on days 5, 10, and 15. aPD-1 was administered intraperitoneally on days 6, 11, and 16 (200 μg per dose). Tumor progression and animal health were assessed by tracking growth rates, incidence, body weight changes, and survival. Twenty-two days post-tumor injection, mice were euthanized for tumor weighing and photography. Mice were euthanized and final data recorded when the tumor size surpassed 1500 mm3 or the maximum dimension surpassed 1.5 cm. In the therapeutic B16F10-GP33 lung metastasis model, mice were inoculated intravenously with 1 × 106 tumor cells. Mice were intraperitoneally immunized with 200 μL PBS containing either GP33, R848+GP33, EPS@GP33, or EPS@R848/GP33 (7.68 mg kg−1) on days 3, 6, and 11 post-inoculation. aPD-1 was administered intraperitoneally on days 4, 7, and 12 (200 μg per dose). Mice were euthanized and lung metastatic nodules were quantified. Lungs were then preserved in paraformaldehyde (PFA), paraffin-embedded, and sliced into 5 μm sections, and subjected to H&E staining for microscopic evaluation. Antitumor immunity dynamics in tumor-bearing mice Following the completion of tumor treatment experiment, tumors and tdLNs were separated, gently dissociated and enzymatically digested in 2 mg mL−1 collagenase type IV for 50-60 min at 37°C to generate single-cell preparations. Cells were mixed in FACS buffer. To quantify infiltrating immune cells within TME, tumor cell suspensions was washed and treated with an Fc-blocker before staining with various markers: Fixable Viability Dye, CD8, CD4, CD44, CD11b, Gr-1 (Ly6G), CD25, Foxp3 and PD-1. For intracellular cytokine detection, the cell suspensions were dyed with IFN-γ, TNF-α and granzyme B. TdLNs and spleens were treated with an Fc-blocker and subsequently stained with CD11c, CD86, CD80, CD8, CD4, CD44, CD62L and PD-1. Data were acquired on a NovoCyte flow cytometer and analyzed using FlowJo software. Toxicity assessment of drugs In vivo and In vitro Female mice bearing subcutaneous B16F10-GP33 tumors were injected intraperitoneally with different formulations. On day 7 after the final injection, blood was obtained for serum biochemical analysis, and the heart, liver, spleen, lung, and kidney were harvested for histology and H&E staining. For long-term safety assessment, healthy female mice were administered the same immunization schedule. Four weeks after the final injection, blood was collected for serum biochemical analysis, and major organs (heart, liver, spleen, lung, and kidney) were harvested for H&E staining. For hemolysis evaluation, fresh mouse red blood cells (RBCs) were collected and washed with PBS. The RBC suspension was incubated with different formulations at 37°C for 2 h. PBS and 1% Triton X-100 served as negative and positive controls, respectively. After centrifugation, the absorbance of the supernatant was measured at 540 nm, and the hemolysis rate was calculated. Data analysis All analyses were carried out in GraphPad Prism 8. Data are displayed as mean ± SEM. Differences between two groups were evaluated by unpaired t-test. For multiple-group comparisons, one-way ANOVA with Dunnett's multiple comparisons test was applied. Significance was defined as NS, not significant; *, p < 0.05; **, p < 0.01; ***, p < 0.001; ****, p < 0.0001. Results Isolation and activity screening of microbiota-derived exopolysaccharides To identify the most promising candidate with integrated delivery and self-adjuvant properties, we extracted 20 EPSs from lactic acid bacteria. For EPS isolation, proteins were precipitated and removed from bacterial supernatants using trichloroacetic acid, followed by an ethanol precipitation to obtain crude EPSs. The extracts were further purified by dialysis and freeze-drying (Fig. S1). Next, we evaluated the in vitro adjuvant activity of these EPSs. At the cellular level, activation of BMDCs was used as the primary readout. EPS131, EPS233, EPS250, and EPS288 significantly increased the proportions of CD40, CD80, CD86, and MHC-I on BMDCs compared with other candidates (Fig. S2A–D). Notably, these EPSs also robustly induced secretion of IFN-I and the pro-inflammatory cytokine IL-12p70 in the BMDC culture supernatants (Fig. S2E and F). At the animal level, mice were immunized intraperitoneally with 20 EPSs mixed with OVA. Serum obtained 14 and 28 days post-immunization showed that EPS233, EPS250, and EPS288 elicited the highest levels of OVA-specific IgG (Fig. S3A and B). To further assess their physicochemical properties, we examined the critical micelle concentration (CMC), a key parameter reflecting the ability of polymers to form nanoparticles [33]. These EPSs exhibited amphiphilic characteristics and formed nanoparticles (Fig. S4A–D). Moreover, evaluation of OVA encapsulation across the four EPSs revealed that EPS233 achieved the highest encapsulation efficiency (79%) and drug loading (29%) (Fig. S4E), indicating its superior antigen delivery capability. In summary, EPS233 was identified as the optimal candidate with integrated delivery and self-adjuvant properties following systematic screening. Molecular structure and characteristics of EPS233 The EPS233 was further purified and the elution curves were shown in Fig. S5A and B. The purified EPS exhibited high uniformity, as verified by UV spectroscopic analysis (Fig. S5D). The FTIR spectra of EPS233 indicated the presence of abundant functional groups, including hydroxyl, carbonyl, carboxyl, and epoxy groups (Fig. S5C). High-performance gel permeation chromatography estimated a mean molecular mass of approximately 77.6 kDa (Fig. S5E). The GC-MS analysis of the monosaccharide composition revealed that EPS233 comprises GalN, GlcN, Ara, Gal, Glc, Man, and GlcA in molar ratios of 8.1:7.1:1.0:4.5:3.0:22.4:20.6, respectively (Fig. S5F). The SEM of EPS233 showed a fibrous morphology (Fig. S5G), and the TEM images showed that EPS233 spontaneously formed spherical nanostructures with a diameter of about 27 nm when dispersed in water, with a bright and uniform morphology (Fig. S5H). The glycosidic linkage patterns of sugar residues were determined through methylation analysis followed by GC-MS. The glycosidic linkage patterns were listed in Table S2, with representative PMAAs spectra provided in Fig. S6. Twelve sugar residues were identified, including 2,3,5-Me3-Araf, 2,3-Me2-Araf, 2,3-Me2-Arap, 2,3,4,6-Me4-Manp, 2,3,4,6-Me4-Galp, 3,4,6-Me3-Manp, 2,3,6-Me3-Galp, 2,3,6-Me3-Glcp, 2,4,6-Me3-Manp, 2,3,4-Me3-Manp, 3,4-Me2-Manp, 2,4-Me2-Galp, with the molar ratios of 7:4.6:2.8:17.5:2.1:14.2:6.6:18.2:3.7:3.5:14.1:5.7. To further elucidate the detailed structure of EPS233, comprehensive NMR spectroscopic analyses were conducted, including one-dimensional (1H and 13C) and two-dimensional (1H-1H COSY, HSQC, HMBC and NOESY) experiments (Fig. S7A–F). All proton and carbon signals were fully assigned and compiled in Table S3. Taken together with the above analysis, the main chain connection of EPS233 is →3)-β-D-GlcpNAc-(1 → 4)-β-D-GlcpA-(1]6 → 3)-Manp-(1 → 2)-Man-(1 → 2)-Man-(1 → 2,6-Man-1→ (Fig. S8). EPS potentiates TLR agonist activities TLR agonists are potent cancer immune adjuvants [34]. However, their antitumor efficacy in cancer immunotherapy has been limited [35]. Our previous results have demonstrated the unique adjuvant capacity of EPS. Therefore, we investigated whether EPS could enhance TLR agonist activity to promote immune cell activation and bolster antitumor effects. Specifically, we examined IFN-I production and pro-inflammatory responses in murine BMDCs treated with EPS in combination with various TLR agonists. When EPS was added to Pam3CSK4, poly(I:C), MPLA, R848, and CpG ODN, it resulted in elevated secretion of IFN-β, IL-6, and IL-12p70 from BMDCs compared to TLR agonists alone (Fig. S9A–C). Notably, we did not observe a significant enhancement of IFN-β production when EPS was supplemented with TLR2 or TLR4 agonists (Fig. S9A). To validate these in vitro findings, we investigated whether EPS could enhance the in vivo antitumor efficacy of TLR agonists. Tumor-bearing mice were treated intraperitoneally with TLR agonists with or without EPS after subcutaneous inoculation of B16F10-OVA melanoma cells (Fig. S10A). In agreement with the in vitro findings, coadministration of EPS enhanced the antitumor efficacy of poly(I:C), R848, and CpG (Fig. S10B–E). Notably, the combination of EPS and R848 showed robust antitumor efficacy. Given the robust enhancement observed with R848 in the presence of EPS, we then focused our subsequent studies on the combination of R848 and EPS. Preparation and characterization of EPS@R848/OVA We exploited the self-assembling properties of exopolysaccharides to develop a nanoparticle system for codelivery of R848 and antigens for systemic cancer immunotherapy (Fig. 2A). EPS@R848/OVA was prepared in aqueous solution, and its CMC value was 1.7701 mg mL−1, which meant EPS may encapsulate the hydrophobic small molecule agonists and OVA to form new nanoparticles (Fig. 2B). After formulation optimization, the highest encapsulation efficiencies (wt/wt) were achieved at an EPS:R848:OVA ratio of 10:2.5:1, reaching 84.1% for R848 and 84.5% for OVA (Fig. S11A–F). The TEM images revealed that EPS@R848/OVA exhibited a uniform distribution and monodisperse spherical geometry in water (Fig. 2C). Preparation and characterization of EPS@R848/OVA. (A) Diagram illustrating the preparation of EPS@R848/OVA. (B) CMC value of EPS@R848/OVA. (C) TEM images of EPS@R848/OVA. Scale bars, 50 nm. (D) Schematic of molecular docking simulations between EPS and R848. (E) The assembly process of EPS@R848 with OVA within 100 ns in MD simulations. (F) DLS size measurements of EPS@R848/OVA. (G and H) The release profiles of (G) R848 and (H) OVA from EPS@R848/OVA were evaluated in PBS containing 10% FBS at pH 5.0 and 7.2. Molecular docking simulations revealed a favorable interaction between EPS and R848, with a predicted binding energy of −2.54 kcal mol−1, indicating spontaneous association. Structural analysis showed that the curled and folded EPS chain forms a local groove-like pocket that accommodates R848 with good structural complementarity. At the binding interface, hydrogen bonding was identified as the major stabilizing force. Specifically, polar atoms on R848, including nitrogen and oxygen atoms, formed multiple intermolecular hydrogen bonds with hydroxyl groups on EPS. OGB-6 and XGB-2 were identified as the key glycosyl residues involved in ligand recognition. The hydrogen-bond distances of 3.0, 3.1, 3.2, and 3.3 Å fall within the typical range for stable hydrogen bonds, supporting the stability of the interaction. Together, these results indicate that EPS interacts with R848 through favorable binding affinity, structural complementarity, and a hydrogen-bonding network (Fig. 2D). Molecular dynamics simulations were conducted to elucidate the nanoscale self-assembly of EPS, R848, and OVA in water. As shown in Fig. 2E, the components were initially randomly dispersed at 0 ns, began to form small aggregates by 20 ns, and gradually assembled into a stable spherical nanocluster by 100 ns, indicating spontaneous co-assembly in aqueous solution. To further define the molecular basis of this process, we analyzed the RMSD of backbone atoms, Rg, SASA, intermolecular hydrogen bonds, and interaction energies over time (Fig. S12A–E). These analyses showed that the assembly process was driven by the coordinated contribution of hydrogen bonding, electrostatic interactions, and van der Waals forces, providing mechanistic insight into the intermolecular interactions underlying the formation and stabilization of the nanostructure. DLS analysis indicated that EPS@R848/OVA had a mean size of 37.31 ± 1.51 nm with a narrow size distribution (polydispersity index = 0.245 ± 0.009) (Fig. 2F). Additionally, EPS exhibited a zeta potential of −39.24 ± 5.02 mV and EPS@R848/OVA had a zeta potential of −23.51 ± 0.31 mV (Fig. S13A). We next evaluated the release of R848 and OVA from EPS@R848/OVA at pH 7.2 and 5.0. R848 showed sustained release at pH 7.2 without an initial burst, whereas release was accelerated at pH 5.0 (Fig. 2G), consistent with enhanced cargo liberation under acidic conditions. OVA displayed a similar profile, with <25% cumulative release over 24 h at pH 7.2 and > 50% at pH 5.0 (Fig. 2H). These results indicate a pH-responsive release behavior favorable for intracellular delivery and immune activation [36]. The nanoparticles maintained stability in PBS for 3 weeks at 25°C and 4°C (Fig. S13B). Considering that systemic release of R848 may induce a cytokine storm associated with acute toxicity and morbidity [37], we compared the induction of systemic pro-inflammatory cytokines in mice following early subcutaneous immunization with either R848+OVA or EPS@R848/OVA. R848 induced elevated systemic pro-inflammatory cytokines at 6 h, leading to temporary body weight reduction, a response not observed with EPS@R848/OVA (Fig. S14A and B). This may result from the prolonged release and local retention of R848 by EPS@R848/OVA. Finally, the in vitro biocompatibility of EPS@R848/OVA was assessed in BMDCs and L929 cells. CCK-8 analysis showed detectable cytotoxicity only at 100 μg mL−1 (Fig. S15A and B), indicating favorable biocompatibility at lower doses. EPS@R848/OVA induced robust innate immune activation in vitro As an effective delivery system, efficient delivery of antigens to DCs is crucial. DCs must efficiently capture exogenous antigens, present them on cell-surface MHC molecules, and facilitate interactions with T cells through MHC-antigen complexes, thereby activating T-cell immune responses [38]. Beyond enhancing antigen endocytosis, promoting lysosomal escape is even more critical for improving antigen presentation efficiency [39]. To further elucidate intracellular trafficking, DC2.4 cells were incubated with free FITC-OVA or EPS@R848/OVA-FITC for 2, 4, 6, 8, and 12 h, followed by lysosomal staining and CLSM observation (Fig. 3A). Pearson's correlation coefficient (PCC) was calculated to quantify the colocalization between FITC and LysoTracker signals, with higher PCC values indicating stronger correlation. EPS@R848/OVA-FITC showed strong lysosomal colocalization at 4 h, indicating efficient internalization. The PCC value then decreased markedly over time, reaching a low level at 8-12 h, consistent with progressive lysosomal escape. In contrast, free FITC-OVA exhibited only moderate PCC values throughout the time course (Fig. 3B). This may be attributed to its limited cellular uptake and subsequent retention within lysosomes, leading to degradation rather than effective cytosolic escape. Following the release of antigens from lysosomes in DCs, they are processed and presented on the surface of DCs. Therefore, we further analyzed the expression levels of MHC molecules by flow cytometry to evaluate antigen presentation in BMDCs. The surface expression of MHC-I, SIINFEKL-H-2Kb, and MHC-II on BMDCs was markedly elevated in the group co-incubated with EPS@R848/OVA compared to the control, OVA, R848+OVA, and EPS@OVA groups (Fig. 3C–E). The maturation status of BMDCs was assessed by flow cytometry. The EPS@R848/OVA group demonstrated significantly enhanced immunogenicity compared with other groups, showing a stronger capacity to upregulate costimulatory molecules CD40, CD80, and CD86 on BMDCs (Fig. 3F–H). In parallel, the secretion of T-cell-activating cytokines including IFN-β, IL-6, TNF-α, and IL-12p70 by BMDCs was analyzed. In line with BMDC maturation, BMDCs treated with EPS@R848/OVA produced markedly elevated levels of these cytokines (Fig. 3I). These findings demonstrate that the codelivery system significantly enhances the cellular internalization and lysosomal escape of OVA, thereby promoting DC activation and antigen cross-presentation (Fig. 3J). EPS@R848/OVA enhanced antigen uptake and lysosomal escape to facilitate BMDC maturation and amplify antigen cross-presentation in vitro. (A and B) Representative CLSM images (A) and the corresponding PCC analysis (B) of DC2.4 cells exposed to FITC-labeled OVA and EPS@R848/FITC-OVA for various times at 37°C. Lysosomes and nuclei were stained with LysoTracker Red and DAPI, respectively. Scale bar, 10 μm. (C-E) FCM was performed to determine the percentages of (C) MHC-I+, (D) SIINFEKL-H-2Kb+ and (E) MHC-II+ BMDCs after different treatments (n = 3). (F-H) FCM analysis was used to determine the percentages of (F) CD40+, (G) CD80+, and (H) CD86+ BMDCs after various treatments (n = 3). (I) Concentrations of IFN-β, IL-6, TNF-α, and IL-12p70 in the supernatant of BMDCs were measured by ELISA after various treatments (n = 3). (J) The schematic illustrates that EPS@R848/OVA enhances DC internalization, lysosomal escape, BMDC maturation, and antigen presentation. (K) Cluster heatmap illustrating gene expression profiles in BMDCs exposed to control or EPS@R848/OVA (n = 3). (L) Western blot analysis of the activation of the Dectin-2 and TLR7 pathways in BMDCs following incubation with R848+OVA, EPS@OVA, or EPS@R848/OVA (n = 3). NS, not significant; *, p < 0.05; **, p < 0.01; ***, p < 0.001; ****, p < 0.0001. Mechanistic research To investigate the molecular mechanism through which the codelivery system stimulates DCs, we performed transcriptomic profiling of BMDCs incubated with EPS@R848/OVA or PBS. Transcriptomic analysis revealed markedly increased expression of Irf8 and Batf3 in EPS@R848/OVA-treated BMDCs. BATF3 is known to sustain IRF8 autoactivation and drive the terminal differentiation of CD103+CD11b− cDCs, which are uniquely capable of cross-presenting antigens on MHC-I molecules, a process essential for inducing protective CTL responses against tumors [40]. In addition, multiple immune activation-associated genes, including interleukins, tumor necrosis factor (TNF), interferons, and chemokines, were significantly upregulated in EPS@R848/OVA-treated BMDCs (Fig. 3K). KEGG pathway analysis further revealed that EPS@R848/OVA activated BMDCs through coordinated engagement of Toll-like receptor and C-type lectin receptor pathways (Fig. S16A). GO analysis additionally showed that EPS@R848/OVA enhanced the levels of MHC-I-related proteins involved in antigen cross-presentation and promoted IFN-I production and T cell-mediated immunity (Fig. S16B). Transcriptomic analysis revealed that Clec4n and Syk were markedly upregulated in EPS@R848/OVA-treated BMDCs, whereas Clec7a expression remained unchanged, suggesting that EPS is preferentially recognized through Dectin-2 [41]. To validate this possibility, we examined the activation of the Dectin-2-Syk signaling axis by western blot analysis. CARD9 functions as a central downstream adaptor in Dectin signaling, and its expression was also assessed [42]. As shown in Fig. 3L, both EPS@OVA and EPS@R848/OVA markedly increased the protein levels of Dectin-2 and CARD9, as well as Syk phosphorylation, compared with the control and R848+OVA groups. Given that R848 is known to activate innate immune cells through the TLR7-MyD88-IRF7 pathway to induce IFN-I responses [43], we next evaluated this signaling pathway by western blot analysis. Both R848+OVA and EPS@R848/OVA upregulated the expression of TLR7 and MyD88 and enhanced IRF7 phosphorylation. Overall, these results reveal that EPS@R848/OVA activates DCs through the coordinated engagement of the Dectin-2-Syk-CARD9 and TLR7-MyD88-IRF7 signaling pathways, thereby promoting IFN-I production and enhancing antigen cross-presentation. Lymph node targeting and cellular internalization Nanoparticles with specific properties can efficiently drain into dLNs, where they activate resident innate and adaptive immune cells. Furthermore, surface functionalization of nanoparticles with pattern recognition receptor (PRR) ligands has been demonstrated as a successful strategy to enhance lymph node targeting and accumulation [44,45]. To assess the dLN migration and retention of EPS@R848/OVA in vivo, mice were treated with Alexa Fluor (AF) 647-labeled OVA and EPS@R848/OVA, followed by whole-body fluorescence imaging. Compared with free AF647-OVA, which showed rapid systemic distribution, EPS@R848/OVA-AF647 remained locally accumulated and exhibited prolonged in vivo retention for more than 120 h (Fig. 4A). Additionally, the EPS@R848/OVA-AF647-treated group exhibited a significantly elevated fluorescent signal of AF647 in the inguinal LNs (iLNs) as early as 6 h following administration (Fig. 4B). To further confirm these findings, ex vivo fluorescence imaging of major organs was performed at 6 h and 24 h after subcutaneous injection. Consistent with the in vivo imaging results, EPS@R848/OVA-AF647 showed preferential accumulation in the inguinal lymph nodes at both time points. Semiquantitative ROI analysis further demonstrated that the fluorescence intensity in dLNs was higher in the EPS@R848/OVA group than in the free OVA group (Fig. S17). For comparison, we prepared a simple mixture of AF647-conjugated OVA, EPS, and R848 (denoted as EPS + R848+OVA). Whereas free OVA and EPS + R848+OVA failed to induce appreciable dLN expansion, EPS@R848/OVA significantly induced dLN expansion (Fig. 4D). Correspondingly, the weight of dLNs in the EPS@R848/OVA group increased by 1.27- and 1.34-fold at 6 h and 24 h, respectively, compared with EPS + R848+OVA, and by 1.36- and 1.73-fold compared with OVA (Fig. 4E). These findings suggest that EPS@R848/OVA preferentially targets and is retained in the dLNs much longer than free OVA or EPS + R848+OVA. To evaluate whether EPS@R848/OVA retention in dLNs promotes antigen uptake, various APC populations, including DCs, macrophages, and B cells, were examined by flow cytometry. Compared with OVA alone and EPS + R848+OVA, cellular uptake of EPS@R848/OVA was observably increased as early as 6 h following administration and sustained for at least 24 h (Fig. 4F–H, and I). The improved cellular uptake was associated with an increase in cell number and sustained activation of total DCs in dLNs (Fig. 4G). Interestingly, EPS@R848/OVA accelerated the uptake of OVA by CD103+CD11b− cDCs (Fig. 4J), a subset that is critical for antitumor immunity because of its potent capacity to prime CD8+ T cells [46]. Strikingly, EPS@R848/OVA induced sustained OVA+ CD103+CD11b− cDC activation, as evidenced by the activation marker CD86 (Fig. 4K). Notably, EPS + R848+OVA despite containing identical components, did not achieve comparable uptake levels, highlighting the essential role of the efficient codelivery. These results demonstrate that the codelivery system rapidly and effectively accumulates in dLNs, driving innate immune activation. Rapid accumulation of EPS@R848/OVA improved the uptake of OVA and activation of APCs in dLNs. Mice received subcutaneous administrations of OVA-AF647 (excitation/emission = 651/672 nm) or EPS@R848/OVA-AF647 into the inner thigh at 0 h. (A) Representative radiance maps (normalized total intensity) at the injection site, with longitudinal data (n = 3). (B) Representative radiance maps of harvested iLNs, with mean fluorescence intensity plotted at 6 h and 24 h (n = 3). (C) Schematic of the iLNs sampling strategy (n = 4-6). (D and E) Representative images and weights of iLNs were measured at 6 and 24 h post-injection (n = 4-6). (F, H, and I) Uptake of AF647-OVA by (F) total DCs (CD11c+OVA+), (H) B cells (B220+OVA+) and (I) macrophages (CD11b+Ly6G−Ly6C− OVA+) in iLNs at 6 and 24 h post-immunization (n = 4-6). (J) Uptake of AF647-OVA by CD103+CD11b− cDCs (CD11c+MHCII+CD103+CD11b− OVA+) in iLNs at 6 and 24 h post-immunization (n = 5). (K) The number of CD86+OVA+CD103+CD11b− cDCs (CD11c+MHCII+CD103+CD11b−OVA+CD86+) in iLNs at 6 and 24 h (n = 4). NS, not significant; *, p < 0.05; **, p < 0.01; ***, p < 0.001; ****, p < 0.0001. EPS@R848/OVA induced superior OVA-specific CD8+ T-cell immune responses We next assessed the ability of the codelivery system to stimulate the proliferation of antigen-specific T cells in vitro. BMDCs treated with OVA combined with various formulations were co-cultured with CFSE-labeled OVA-specific CD8+ T cells or CD4+ T cells sorted from OT-I (CD8+) and OT-II (CD4+) mice (Fig. 5A). Both OT-I and OT-II T cells exhibited significantly higher proliferation when cultured with BMDCs treated with EPS@R848/OVA compared to those cultured with BMDCs treated with other groups (Fig. 5B and Fig. S18A). To further validate the enhancement of CTL cytotoxicity by the codelivery system, we conducted an in vitro killing assay. The treated OT-I cells were co-cultured with EL4 target cells at different ratios. As demonstrated in Fig. 5C and D, the EPS@R848/OVA-treated group exhibited increased cytolysis of EL4 cells compared with the other groups. This indicates that EPS@R848/OVA treatment improves CTL cytotoxicity against EL4 target cells [47]. EPS@R848/OVA enhanced robust OVA-specific CTL responses in vitro and OVA-specific CD8+ T-cell responses in dLNs. (A) Schematic of BMDC and OT-I cell co-culture assay and in vitro CTL killing assay. (B) FCM plots illustrating CFSE dilution in OT-I T cells post-cultivation with BMDCs exposed to PBS, OVA, EPS@OVA, R848+OVA, or EPS@R848/OVA. Values indicate the percentage of dividing OT-I, identified by reduced CFSE fluorescence (n = 3). (C) Percentage of EL4 target cell lysis by CTLs after co-incubation at different ratios (n = 3). (D) Representative FCM plots of EL4 target cell lysis at a 1:1 ratio after co-incubation with CTLs treated with different formulations (n = 3). (E) Mice received subcutaneous administrations of PBS, OVA, EPS@OVA, R848+OVA, or EPS@R848/OVA into the inner thigh on days 0 and 14. (F and G) FCM analysis 7 days after the second vaccination illustrating representative plots and quantification of (F) CD8+CD44+ and (G) CD8+CD44+IFN-γ+TNF-α+ T cells in the iLNs (n = 5). (H) Purified CD8+ T cells were separated from OT-I mice and adoptively transferred into CD8−/− mice via tail vein injection 24 h before immunization. Mice received subcutaneous immunization into the inner thigh with the designated nanovaccines. (I and J) Representative FCM plots with quantitative analysis of (I) CD8+CD44+ and (J) CD8+CD44+IFN-γ+TNF-α+ T cells in the iLNs 5 d after immunization (n = 5-6). NS, not significant; *, p < 0.05; **, p < 0.01; ***, p < 0.001; ****, p < 0.0001. T cell immune activation triggered by the codelivery system in the dLNs was examined. Mice were treated with two doses (Fig. 5E). Flow cytometric analysis of the dLNs indicated that EPS@R848/OVA elicits more potent CD8+ T-cell responses compared to R848+OVA or EPS@OVA in the dLNs. Specifically, EPS@R848/OVA treatment expanded the populations of effector CD8+ T cells (Fig. 5F) and multifunctional CD8+ T cells in the dLNs, characterized by simultaneous secretion of IFN-γ and TNF-α (Fig. 5G), which is associated with enhanced tumor-killing efficacy. A similar trend was observed in CD4+ T-cell effector functions, where EPS@R848/OVA induced higher IFN-γ+ Th1 and IL-17A+ Th17 populations within CD4+ T cells than other formulations (Fig. S18B). Together, these findings demonstrate that EPS@R848/OVA elicits robust CD8+ and CD4+ T-cell responses in dLNs. To accurately characterize the OVA-specific CD8+ T cell response in the dLNs following EPS@R848/OVA immunization, an adoptive transfer model was established using OT-I cells (Fig. 5H). EPS@R848/OVA markedly expanded the magnitude of effector CD8+ T cells and multifunctional CD8+ T cells (Fig. 5I and J) relative to the other vaccine formulations. OVA-specific CD4+ T cell activation and differentiation were further assessed using an adoptive transfer model in Rag−/− recipient mice (Fig. S18C). Consistent with the findings for CD8+ T cells, dLN analysis revealed that EPS@R848/OVA induced a more potent OVA-specific CD4+ T cell response (Fig. S18D) and significantly enhanced the effector differentiation of IFN-γ-producing CD4+ T cells (Fig. S18E) compared to other formulations. These findings demonstrate that EPS@R848/OVA effectively enhances the magnitude and quality of OVA-specific CD8+ and CD4+ T-cell responses in dLNs. The robust T cell responses induced by EPS@R848/OVA prompted us to further assess its effects on humoral immunity, which is primarily dependent on germinal center (GC) responses [48]. Mice were treated with two doses and dLNs were analyzed by flow cytometry 7 days later (Fig. S19A). EPS@R848/OVA significantly enhanced GC B cell responses in dLNs compared with other formulations (Fig. S19B). Consistent with this finding, EPS@R848/OVA also markedly increased the expansion of T follicular helper (Tfh) cells, providing critical bolster to GC B cells (Fig. S19C). In addition to promoting GC responses, EPS@R848/OVA enhanced the extrafollicular B cell response, as evidenced by a significant increase in early plasmablasts (Fig. S19D). Serum immunoglobulin responses were then evaluated over time after prime and booster immunization (Fig. S19A). EPS@R848/OVA induced significantly stronger OVA-specific IgG responses than the other formulations as early as 2 weeks after priming and sustained this response after boosting. A similar trend was observed for OVA-specific IgG2b, consistent with the Th1-biased CD4+ T cell response observed above (Fig. S19E). Together, these results indicate that EPS@R848/OVA potently augments both GC-dependent and antibody-mediated humoral immune responses. The nanovaccine effectively prevented orthotopic melanoma development and induced long-term immune memory responses In view of the robust antigen-specific immune response induced by the bi-adjuvant nanovaccine, we proceeded to evaluate its immunoprotective effects. Mice were treated with three doses, then challenged with B16F10-OVA melanoma cells (Fig. 6A). The trajectories of tumor progression for each group are presented in Fig. 6B and C. Vaccination with EPS@R848/OVA significantly delayed tumor formation, evidenced by the slowest tumor growth rate and smallest tumor volume. On day 22 post-inoculation, mice were euthanized and tumors were collected. EPS@R848/OVA demonstrated notable anticancer efficacy, achieving approximately 52% reduction in tumor weight compared with R848+OVA and EPS@OVA (Fig. 6D and F). Importantly, EPS@R848/OVA demonstrated no significant toxicity, as reflected by the tumor-to-body weight ratio (Fig. 6E). EPS@R848/OVA induced long-term immune memory responses and effectively prevented B16F10-OVA orthotopic tumor development. (A) Schematic of the prophylactic melanoma treatment regimen in mice. (B) Graph illustrating tumor volume progression in B16F10-OVA tumor-bearing mice subjected to various treatments over 22 days (n = 6). (C) Individual growth curves of B16F10-OVA tumors in mice over 22 days (n = 6). (D) Excised tumor weights in B16F10-OVA tumor-bearing mice on day 22 (n = 5). (E) Excised tumor weights were expressed as a percentage of body weight in B16F10-OVA tumor-bearing mice on day 22 (n = 5). (F) Representative images of tumors from each group on day 22 (n = 5). (G, I) Numbers of CD8+ TCM and CD8+ TEM cells, along with representative FCM plots, in the dLNs (G) and spleens (I) collected from mice 21 days after the initial nanovaccine administration (n = 4-5). (H, J) Numbers of CD4+ TCM and CD4+ TEM cells in the dLNs (H) and spleens (J), collected from mice 21 days after the initial nanovaccine administration (n = 4-5). NS, not significant; *, p < 0.05; **, p < 0.01; ***, p < 0.001; ****, p < 0.0001. To clarify how the bi-adjuvant nanovaccine confers strong prophylactic protection against B16F10-OVA melanoma, we analyzed the immune response following vaccination. Memory T cell populations in dLNs and spleens were quantified 7 days after the final immunization. EPS@R848/OVA immunization markedly expanded memory T cell populations (both CD8+ and CD4+ subsets) in the spleens and iLNs compartments. Enhanced populations included central memory T cells (TCM) and effector memory T cells (TEM) across both T cell lineages (Fig. 6G–J). Furthermore, mice immunized with EPS@R848/OVA exhibited dramatically higher populations of CD8+CD44+ and CD4+CD44+ T cells, along with cytokine-producing T cells in the spleens compared with other groups (Fig. S20A–D). Accordingly, EPS@R848/OVA enhanced and sustained protective cellular immunity by increasing CD8+ TCM and TEM cells in the dLNs and spleens, facilitating rapid generation of effector CD8+ T cells and initiating CTL responses against tumor development. EPS@R848/OVA also increased CD4+ TCM and TEM cells, promoting increased tumor infiltration of CD4+ T cells, which augments the CTL response [49]. Together, these results indicate that the bi-adjuvant nanovaccine exhibits strong prophylactic efficacy and induces durable antitumor immune memory in vivo. The nanovaccine elicited potent antitumor efficacy and remodeled the TME We further investigated the bi-adjuvant nanovaccine's potential to induce tumor regression as a therapeutic cancer vaccine. The antitumor efficacy of EPS@R848/OVA was evaluated using a B16F10-OVA subcutaneous melanoma model (Fig. 7A). The EPS@R848/OVA-treated group exhibited pronounced tumor growth inhibition compared with groups treated with R848+OVA, EPS@OVA, OVA, and control (Fig. 7B). At the endpoint, the mean tumor volume in the EPS@R848/OVA group was reduced by 52.86% and 49.47% compared with the EPS@OVA and R848+OVA groups, respectively. Detailed tumor growth curves for each group are presented in Fig. 7C. Treatment with OVA alone showed minimal therapeutic effect, while the single-adjuvant vaccines led to only partial tumor growth inhibition. In contrast, EPS@R848/OVA produced the most pronounced suppression of tumor progression. This trend is supported by representative tumor images and the quantification of tumor weights and relative tumor weights at the endpoint (Fig. 7D–F). Quantitative analysis further showed that EPS@R848/OVA reduced tumor weight by approximately 48% and 58% compared with the R848+OVA and EPS@OVA groups, respectively (Fig. 7E). The nanovaccine effectively suppressed orthotopic and metastatic melanoma growth and remodeled the TME. (A) Schematic of the therapeutic melanoma treatment regimen in mice. (B) Graph illustrating tumor volume progression in B16F10-OVA tumor-bearing mice after various treatments over 22 days (n = 7). (C) Individual growth curves of B16F10-OVA tumors in mice over 22 days (n = 7). (D) Representative images of tumors from each group (n = 5). (E) Tumor weights of excised tumors from each group post-treatment on day 22 (n = 6-7). (F) Ratio of excised tumor weight to body weight in B16F10-OVA tumor-bearing mice on day 22 (n = 6-7). (G and H) Numbers of (G) CD8+CD44+ T cells and (H) CD8+CD44+IFN-γ+granzyme B+ T cells in tumors as determined by FCM on day 22 (n = 5). (I and J) Numbers of (I) MDSCs and (J) Tregs in tumors as determined by FCM on day 22 (n = 4-5). (K) Schematic of the therapeutic lung metastasis melanoma treatment regimen in mice. (L) Quantification of lung metastatic nodules in B16F10-GP33 tumor-bearing mice on day 18 (n = 6). (M) Ratio of excised lung weight to body weight in B16F10-GP33 tumor-bearing mice on day 18 (n = 6). (N) Histological examination of the excised lungs was performed using H&E staining. (O and P) FCM analysis quantified the number of (O) CD8+CD44+ T cells and (P) CD8+CD44+granzyme B+IFN-γ+ T cells in the lung of tumor-bearing mice on day 18 (n = 4-5). (Q and R) FCM analysis quantified the number of (Q) MDSCs and (R) Tregs in the lung of tumor-bearing mice on day 18 (n = 5). NS, not significant; *, p < 0.05; **, p < 0.01; ***, p < 0.001; ****, p < 0.0001. To further investigate the antitumor immunity of the bi-adjuvant nanovaccine, tumors and tumor-draining lymph nodes (tdLNs) were excised on day 22 (Fig. 7A). TdLNs are critical sites for priming adaptive immune responses and sustaining long-term antitumor immunity [50]. EPS@R848/OVA therapy effectively stimulated DCs in tdLNs, resulting in a marked increase in mature DCs, thereby promoting T cell activation. Conversely, other groups demonstrated minimal effect on DC maturation relative to the control (Fig. S21A). Flow cytometric analysis was performed to identify CTL and helper T cell responses. EPS@R848/OVA markedly expanded the populations of activated CD8+ T cells (Fig. 7G). Elevated levels of IFN-γ, TNF-α, and granzyme B were observed in CD8+ T cells from the EPS@R848/OVA group (Fig. 7H and Fig. S21D and E). The EPS@R848/OVA therapy enhanced both the abundance and effector function of CD8+ T cells. CD4+ T cells exert antitumor effects by directly killing tumor cells via cytokine secretion, such as IFN-γ and TNF-α, or by indirectly activating other immune cells, including CTLs, to enhance their cytotoxicity against tumor cells [51]. Notably, compared with other groups, the magnitude of activated CD4+ T cells and CD4+CD44+IFN-γ+TNF-α+ T cells was markedly augmented in tumors treated with EPS@R848/OVA (Fig. S21B and C). Furthermore, substantial infiltration of regulatory T cells (Tregs) into tumors promotes tumor progression and poor prognosis by facilitating immune tolerance and immune evasion [52]. Importantly, EPS@R848/OVA can inhibit Tregs, thereby remodeling the immunosuppressive microenvironment (Fig. 7I). In addition, myeloid-derived suppressor cells (MDSCs) are immunosuppressive regulatory cells that inhibit antitumor immunity through multiple mechanisms, including suppression of CD4+ and CD8+ T cell function and promotion of Treg expansion [53,54]. Administration of EPS@R848/OVA resulted in a decrease in the infiltration of MDSCs in the TME compared with other treatments (Fig. 7J). These findings demonstrate that the bi-adjuvant nanovaccine remodels the immunosuppressive TME and enhances antitumor immunity. Although nanovaccines have significantly advanced cancer immunotherapy, metastasis from primary tumors to distant organs remains a major challenge. We next evaluated the efficacy of the peptide GP33-based nanovaccines in a melanoma lung metastasis model and assessed their potential as a universal vaccine platform [55]. To generate the metastatic model, mice were intravenously injected with B16F10-GP33 cells, and then mice were treated with three doses (Fig. 7K), and lungs were collected on day 18 for analysis. As shown in Fig. 7L and Fig. S22A, both untreated and GP33-treated mice exhibited widespread metastatic nodules in the lungs. In contrast, vaccination with EPS@GP33, R848+GP33, or EPS@R848/GP33 significantly inhibited lung metastatic tumor burden, with EPS@R848/GP33 showing the most potent therapeutic effect. The number of metastatic nodules in the EPS@R848/GP33 group decreased by 17.2 and 13.8 nodules compared with the EPS@GP33 and R848+GP33 groups, respectively. Consistently, the lung-to-body weight ratio was lower in the EPS@R848/GP33 group than in the EPS@GP33 and R848+GP33 groups (Fig. 7M). H&E staining of lungs further validated the potent antitumor efficacy of EPS@R848/GP33 (Fig. 7N). Consistently, the tumor immune microenvironment analysis indicated that EPS@R848/GP33 induced the highest levels of activated cytotoxic T cells (Fig. 7O), helper T cells (Fig. S22B), and multifunctional T cells, including CD8+CD44+IFN-γ+TNF-α+ T cells (Fig. 7P and Fig. S22C), CD8+CD44+granzyme B+IFN-γ+ T cells (Fig. S22D), and CD4+CD44+IFN-γ+TNF-α+ T cells (Fig. S22E) compared with other groups, while reducing the intratumoral infiltration of Tregs and MDSCs (Fig. 7Q and R). Meanwhile, EPS@R848/GP33 therapy effectively activated DCs in dLNs (Fig. S22F). Collectively, these findings demonstrate that the EPS-based bi-adjuvant nanovaccine serves as a versatile therapeutic vaccine platform for cancer immunotherapy with both protein and peptide antigens, effectively enhancing the efficacy of peptide-based vaccines that often exhibit weak immunogenicity. The nanovaccine synergized with ICB therapy to elicit potent antitumor efficacy Given that upregulated PD-1 on the activated T cell surface contributes to tumor escape from immunosurveillance through the PD-1/PD-L1 pathway, we investigated the antitumor activity of combination therapy with the peptide-based nanovaccine EPS@R848/GP33 and anti-PD-1 antibody (aPD-1) in B16F10-GP33 orthotopic tumor-bearing mice. The schematic is shown in Fig. 8A. Mice treated with aPD-1 alone showed continued primary tumor growth, indicating limited therapeutic efficacy of monotherapy. In contrast, EPS@R848/GP33 treatment markedly suppressed primary tumor progression, as evidenced by the tumor growth curves and final tumor weights (Fig. 8B–E). Strikingly, the combination of EPS@R848/GP33 and aPD-1 (EPS@R848/GP33 + aPD-1) exhibited near-complete tumor elimination, achieving 98.1% inhibition in tumor weight (Fig. 8D). This pronounced effect was further reflected in the markedly suppressed tumor volume progression (Fig. 8B and C) and in the representative images of excised tumors (Fig. 8E). In addition, the survival time of EPS@R848/GP33 + aPD-1 group (41 days) was much longer than that of the aPD-1 group (32 days) (Fig. 8F). Analysis of the TME indicated that EPS@R848/GP33 + aPD-1 markedly enhanced the infiltration of activated CD8+ T cells and multifunctional CD8+CD44+IFN-γ+TNF-α+ T cells (Fig. 8G and H), as well as activated CD4+ T cells and CD4+CD44+IFN-γ+TNF-α+ T cells (Fig. S23A), compared to either monotherapy. Furthermore, the frequencies of immunosuppressive MDSCs (Fig. S23B) and Tregs (Fig. S23C) were markedly reduced in TME following combination treatment. Additionally, combination therapy significantly increased the infiltration of activated CD8+ T cells and activated CD4+ T cells in the tdLNs compared to either monotherapy alone (Fig. S23D). Systemic immune memory was also enhanced, as evidenced by increased numbers of CD8+ TEM and TCM cells in the spleen (Fig. S23E) and tdLNs (Fig. S23F) of tumor-bearing mice treated with EPS@R848/GP33 + aPD-1, indicating the induction of durable antitumor immunity. We next examined the expression of the exhaustion marker PD-1, a key indicator of intratumoral T cell exhaustion following engagement with PD-L1 [56]. Importantly, combination treatment effectively reduced PD-1 expression on both CD8+ and CD4+ T cells within tumors (Fig. 8I and Fig. S23G) and tdLNs (Fig. 8J and Fig. S23H), suggesting attenuation of T cell exhaustion and more effective blockade of the PD-1/PD-L1 pathway. The nanovaccine effectively enhanced the antitumor efficacy and antitumor immunity of anti-PD-1 therapy in melanoma models. (A) Schematic of the therapeutic orthotopic B16F10-GP33 melanoma treatment regimen in mice. (B) Graph illustrating tumor volume progression in B16F10-GP33 tumor-bearing mice after various treatments over 20 days (n = 7). (C) Individual growth curves of B16F10-GP33 tumors in mice over 20 days (n = 7). (D) Tumor weights of excised tumors in mice on day 20 (n = 5-6). (E) Representative images of tumors from each group post-treatment (n = 5). (F) Overall survival rates of tumor-bearing mice post-treatments (n = 7). (G and H) FCM analysis quantifying the numbers of (G) CD8+ CD44+ T cells and (H) CD8+CD44+IFN-γ+TNF-α+ T cells in the tumors on day 20 (n = 5). (I and J) Representative FCM plots and proportions of exhausted CD8+ T cells (CD8+CD44+PD-1+) in tumors (I) and in the tdLNs (J) on day 20 (n = 5). (K) Schematic of the therapeutic lung metastasis melanoma treatment regimen in mice. (L) Quantification of lung metastatic nodules in B16F10-GP33 tumor-bearing mice on day 18 (n = 6). (M) Ratio of excised lung weight to body weight in B16F10-GP33 tumor-bearing mice on day 18 (n = 6). (N) Overall survival rates of tumor-bearing mice after treatment (n = 7). (O) Representative images of lungs from each group post-treatment (n = 6). (P) FCM analysis quantifying the numbers of CD8+ CD44+ T cells and CD8+CD44+IFN-γ+TNF-α+ T cells in the lung of tumor-bearing mice on day 18 (n = 5). (Q and R) Representative FCM plots and proportions of exhausted CD8+ T cells (CD8+CD44+PD-1+) in lungs (Q) and spleens (R) of tumor-bearing mice on day 18 (n = 5). NS, not significant; *, p < 0.05; **, p < 0.01; ***, p < 0.001; ****, p < 0.0001. To further evaluate the therapeutic potential of the combination strategy, we assessed the antitumor efficacy of EPS@R848/GP33 + aPD-1 in a B16F10-GP33 lung metastatic model. As shown in Fig. 8L, mice treated with aPD-1 alone exhibited widespread metastatic nodules in the lungs, indicating limited therapeutic efficacy. In contrast, EPS@R848/GP33 + aPD-1 exhibited the most potent therapeutic effect, achieving 88.1% inhibition in tumor weight (Fig. 8L and O). The ratio of lung weight to body weight in aPD-1-treated group was markedly higher than that of EPS@R848/GP33 + aPD-1 group, reflecting extensive metastatic nodule formation in the former groups (Fig. 8M). In addition, the survival time of the EPS@R848/GP33 + aPD-1 group (33 days) was much longer than that of the aPD-1 group (25 days), whereas mice in the control group died quickly (Fig. 8N). TME analysis indicated that EPS@R848/GP33 + aPD-1 treatment significantly elevated the numbers of activated CD8+ T cells and multifunctional CD8+CD44+IFN-γ+TNF-α+ T cells (Fig. 8P), as well as activated CD4+ T cells and CD4+CD44+IFN-γ+TNF-α+ T cells (Fig. S24A), while decreasing the frequency of MDSCs and Tregs (Fig. S24B and C) in tumor tissues compared with either monotherapy alone. Importantly, combination treatment induced a long-term systemic immune memory, as evidenced by significantly elevated numbers of CD8+ TEM and TCM cells in the spleen (Fig. S24D) and iLNs (Fig. S24E). As expected, PD-1 expression on CD8+ T cells was markedly increased in the aPD-1 group, whereas T cells from the EPS@R848/GP33 + aPD-1 group exhibited reduced PD-1 expression in tumors (Fig. 8Q), spleens (Fig. 8R) and iLNs (Fig. S24F), indicating more efficient blocking of the PD-1/PD-L1 pathway [57]. Collectively, these results suggest that the EPS@R848/GP33 + aPD-1 enhances T-cell activation and infiltration within tumors and tdLNs, while reducing the abundance of immunosuppressive Tregs and MDSCs. Furthermore, the induction of robust systemic immune memory supports the potential for long-term protection against tumor recurrence. These effects provide sufficient action targets for improving the response to PD-1 blockade, thereby enhancing the antitumor efficacy of anti-PD-1 therapy and providing a promising strategy to enhance therapeutic efficacy in primary and metastatic melanoma, as well as potentially in other tumor types. In vitro hemocompatibility was assessed by a hemolysis assay. The hemolysis rate of EPS@R848/OVA remained below 5% and was comparable to the PBS control, indicating negligible hemolytic activity (Fig. S25). The in vivo safety of the codelivery system was further evaluated following repeated nanovaccine administration. H&E staining of major organs showed no obvious histopathological abnormalities in mice treated with EPS@R848/GP33 or EPS@R848/GP33 + aPD-1 compared with naïve mice (Fig. S26A). Consistently, serum biochemical analysis revealed no abnormal changes in liver-related markers (ALB, ALP, ALT, and T-BIL), creatinine (CREA), or total protein (TP) levels (Fig. S26B). Long-term safety was further examined in healthy mice. At 4 weeks after the third immunization, major organs retained normal tissue architecture without evident lesions (Fig. S27A), and serum biochemical parameters remained within the normal range (Fig. S27B). These results support the favorable safety profile of the EPS-based nanovaccine after repeated administration. Discussion To stimulate and amplify tumor-specific T cell responses, recent studies in nanovaccines have emphasized the simultaneous delivery of tumor antigens and adjuvants to DCs [14,58]. In this process, antigens transport to secondary lymphoid organs, cytosolic delivery and antigens cross-presentation in coordination with innate stimulation are essential for triggering potent antitumor immunity [8,59]. Thus, to develop efficient cancer nanovaccines, we screened and structurally characterized a microbiota-derived amphiphilic EPS as a versatile self-adjuvanting codelivery system. This EPS acts as a “molecular bridge” to self-assemble with OVA and R848 into a bi-adjuvant nanovaccine (EPS@R848/OVA), driven by coordinated hydrogen bonding, electrostatic interactions, and van der Waals forces. This assembly strategy ensures high antigen and hydrophobic small molecule agonist encapsulation efficiency, while also improving drug solubility and the pharmacokinetic properties of its payloads, thereby further eliciting potent antigen-specific CD8+ T cell responses and demonstrating pronounced antitumor effect as a potent nanovaccine, synergizing with PD-1 blockade. Microbiota-derived EPSs offer several advantages over existing self-adjuvanting nanocarriers, including straightforward synthesis, low cost, and eliminating the need for additional targeting modifications. Significantly, instead of being a modular composite of two discrete components for delivery and adjuvant, our EPSs platform functions as a single, multifunctional platform. We observed that EPS@R848/OVA nanovaccine elicits robust innate immune activation. Specifically, the nanovaccine markedly enhances the cytosolic codelivery of antigen and adjuvant, thereby driving DC activation and antigen cross-presentation to prime CD8+ T cell responses [60]. We further explored the molecular mechanism underlying EPS@R848/OVA-mediated immune activation. These effects depend on the sensing of EPS by DCs through Dectin-2, but not Dectin-1 [41]. Notably, structural characterization of EPS by mass spectrometry and nuclear magnetic resonance revealed a high abundance of mannose residues, consistent with its recognition by Dectin-2 [29], suggesting a close link between its structural features and its immunostimulatory potency, and providing valuable insight for future vaccine design. Moreover, this Dectin-2 pathway further acts synergistically with R848-triggered TLR7/8 signaling, primarily through the TLR7-MyD88-IRF7 axis, to drive the production of proinflammatory cytokines [43,61]. IFN-I plays a pivotal role in initiating antitumor immunity by strengthening DC-T cell crosstalk, promoting tumor-antigen cross-presentation, and enhancing the cytotoxic activity of T cells [62,63]. Notably, cooperative activation of the Dectin-2 and TLR7 pathways synergistically enhances IFN-I production. Lymph nodes act as critical sites for initiating antigen-specific immune responses; however, most nanovaccines still face a major bottleneck in achieving efficient antigen delivery to this critical site and in inducing sufficient DC maturation to support robust antigen cross-presentation [64,65]. EPS@R848/OVA rapidly accumulates in dLNs and is efficiently taken up by APCs within 6 h, thereby maximizing the coordinated delivery and utilization of OVA and R848. This behavior is consistent with the favorable physicochemical features of the nanovaccine, as medium-sized nanoparticles (5-100 nm) with a negative surface charge possess superior lymphatic drainage properties, leading to enhanced dLN accumulation and prolonged retention [65,66]. In addition to passive lymphatic transport, active cell-mediated migration has been shown to further increase nanovaccine accumulation in lymph nodes [44]. Mannan-modified nanovaccines may exploit Dectin-2-mediated interactions with APCs to promote their migration to dLNs and enhance intranodal delivery. More importantly, our nanovaccine substantially enhances the uptake and activation of the indispensable CD103+CD11b− cDC subset after dLN homing, which is intrinsically proficient at cross-presentation [46,60]. EPS@R848/OVA further promotes antigen cross-presentation and induces antigen-specific CTL and Th1 responses, key determinants of effective antitumor immunity [44]. Moreover, in tumor-bearing mice, we observed robust CD8+ T cell and memory immune responses in tdLNs, which are increasingly recognized as critical targets for cancer immunotherapy [50]. Collectively, these findings support an integrated delivery paradigm that coordinates lymph node targeting, efficient antigen uptake, DC activation, potent antigen cross-presentation, and antigen-specific T cell immunity within a minimalist nanovaccine platform. Most tumors exhibit inherently low immunogenicity and are frequently accompanied by substantial infiltration of immunosuppressive cells, like Tregs and MDSCs [52,54]. Moreover, these tumors are deficient in proinflammatory immune populations and functional effector T cells, which further limits the effectiveness of immunotherapy [67]. In this study, our nanovaccines induce durable systemic immune memory and remodel the TME by increasing the intratumoral infiltration of effector CD8+ T cells while decreasing the infiltration of Tregs and MDSCs. This reduction in Tregs and MDSCs likely limits the secretion of immunosuppressive cytokines, such as IL-10 and TGF-β, and reduces barriers to T cell infiltration, thereby suppressing both primary and metastatic tumor growth. Although ICBT has emerged as a powerful alternative to conventional cancer therapies, its clinical benefit remains limited to a subset of patients, and the effective treatment of cold tumors remains a major challenge [[68], [69], [70], [71]]. Notably, our nanovaccines promote an immunogenic tumor phenotype and sensitize otherwise anti-PD-1-nonresponsive tumors to checkpoint blockade, resulting in substantial tumor regression. These findings suggest that our codelivery system strategy may provide a promising immunotherapeutic approach for cold tumors that are refractory to ICB monotherapy [72]. Moreover, these therapeutic effects occur without significant changes in body weight or evident systemic toxicity, indicating a favorable safety profile of the nanovaccines. While these findings are encouraging, the present study is primarily confined to melanoma models expressing surrogate antigens. Further studies in wild-type and poorly immunogenic tumor models using tumor-associated antigens are needed to evaluate the broader applicability of this codelivery system. As cancer vaccine development increasingly moves toward personalized neoantigens derived from patient-specific mutations, incorporating such neoantigens into the EPS-based codelivery platform will be an important step toward further evaluating its translational potential. Conclusion Our study establishes microbiota-derived amphiphilic EPSs as a self-adjuvanting codelivery system that enables efficient and coordinated antigen delivery and innate immune activation. Formulated into a potent bi-adjuvant nanovaccine, this platform induces potent antigen-specific CD8+ T cell responses and achieves potent antitumor efficacy in both orthotopic and metastatic melanoma models, synergizing with PD-1 blockade. This work provides a strategy to address a central limitation in current cancer vaccine design for potent cancer immunotherapy. This microbiota-derived amphiphilic EPS-based codelivery system provides a simple, scalable, and versatile framework for next-generation cancer vaccines, with broad applicability to tumor-associated antigens and emerging neoantigens. CRediT authorship contribution statement Xuemei Bao: Writing – review & editing, Writing – original draft, Validation, Software, Methodology, Formal analysis, Data curation, Conceptualization. Jiamin Shi: Methodology. Haochi Zhang: Methodology. Chunhe Li: Methodology. Yanchen Liang: Software. Na Pan: Methodology. Yubing Fu: Software, Methodology. Xing Su: Methodology. Wan Wang: Methodology. Ting Wang: Methodology. Lipeng Zhang: Conceptualization. Xiao Wang: Writing – review & editing, Supervision, Software, Project administration, Funding acquisition, Formal analysis, Conceptualization. Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Acknowledgements Funding: Supported by the JieBangGuaShuai Program (2023-JieBangGuaShuai), the National Center of Technology Innovation for Dairy (No. 2023-JSGG-2), the Inner Mongolia Autonomous Region University Innovation Team Development Program (NMGIRT2504), and the 2024 Public Hospital Joint Research Fund Program (Matching Grant, 2024GLLH0347). Footnotes Supplementary data to this article can be found online at https://doi.org/10.1016/j.mtbio.2026.103677. Contributor Information Lipeng Zhang, Email: zlp_boy2008@163.com. Xiao Wang, Email: wangxiao@imu.edu.cn. Supplementary data The following is the Supplementary data to this article: Multimedia component 1 Data availability Data will be made available on request. References Associated Data Supplementary Materials Multimedia component 1 Data Availability Statement Data will be made available on request.