Dendritic Cell Derived‐Extracellular Vesicles Engineered to Express Interleukin‐12 and Anti‐CTLA‐4 on Their Surface for Combinational Cancer Immunotherapy

ABSTRACT

Dendritic cell (DC)‐derived extracellular vesicles (DEVs) are promising candidates for cancer vaccines, but their therapeutic effects still need further optimization. In this study, we utilized neoantigens, lipopolysaccharide and IFN‐γ to induce the maturation of DCs, and then isolated DEVs derived from these mature DCs. We showed that the immune checkpoint inhibitor (anti‐CTLA‐4 antibody, aCTLA‐4) can improve the immunostimulatory function of DEVs by directly activating T cells through immune checkpoint signal blockade. The cytokine interleukin‐12 (IL‐12), as one of the third signals for T cell activation, can also enhance the capability of DEVs to activate T cells directly. Based on these findings, we designed the engineered DEVs conjugated with IL‐12 and aCTLA‐4 (DEV@IL‐12‐aCTLA‐4) to improve the therapeutic potential of DEVs by providing sufficient immune regulatory signals. Moreover, the carrier property of DEVs also contributes to the delivery of IL‐12 and aCTLA‐4 to lymph nodes. This indicates that the conjugation of DEVs with IL‐12 and aCTLA‐4 constitutes a complementary approach, where IL‐12 and aCTLA‐4 help to enhance the T cell activation effect of DEVs, and DEVs facilitate the delivery of IL‐12 and aCTLA‐4. Our results showed that DEV@IL‐12‐aCTLA‐4 can enhance the Th1 immune response and reverse exhausted CD8⁺ T cells in the tumour microenvironment, effectively inducing robust T cell immune responses and inhibiting tumour growth in tumour‐bearing mice. Overall, this study expands the theoretical foundation of DEVs and provides a universal strategy for optimizing cancer combination immunotherapy by reprogramming DEVs.

1. Introduction

Mobilizing the immune system to fight against tumour cells is a promising strategy for cancer therapy (van Weverwijk and de Visser 2023; Kraehenbuehl et al. 2022; Riley et al. 2019). As one of the most important effector cells in the immune system, T cells play a key role in anti‐tumour immune responses (Schenkel and Pauken 2023; Oliveira and Wu 2023; Wang et al. 2020). Multiple immune signals regulate the specific activation of anti‐tumour T cells. After processing and presenting tumour antigen by antigen‐presenting cells (APCs), the antigen‐major histocompatibility complex (MHC) as the first signal and the costimulatory molecule as the second signal are provided for T cell activation (Chen and Flies 2013; Smith‐Garvin et al. 2009). At the same time, the cytokines secreted by APCs as the third signal are also necessary for activating T cells. In addition, the existence of immune checkpoints can balance the activation of T cells as a negative regulatory mechanism (Wei et al. 2018). The comprehensive effects of these complex signal regulatory networks determine the outcome of T cell activation. Rationally adjusting these regulatory signals is an effective strategy to develop the most optimal anti‐tumour T cell immunity.

Cancer vaccines aim to elicit a robust T‐cell response against tumour neoantigens (Hu et al. 2018). Dendritic cell (DC)‐based cancer vaccines exhibit potential in this field. However, their clinical outcomes remain unsatisfactory, with objective response rates seldom exceeding 15% (Harari et al. 2020). Extracellular vesicles (EVs) have been developed as new biological platforms for the preparation of cancer vaccines in recent years (Xu et al. 2018; Chen et al. 2022; Wang et al. 2023). EVs originating from diverse cells play distinct roles in various tumour therapy areas due to their heterogeneity, which is inherited from the parent cells, as well as created by different EV biogenesis pathways (Ye et al. 2023; Zhang et al. 2021; Zhang et al. 2023). Mature DC‐derived EVs (DEVs) retain the functional signalling molecules on the surface of parent DCs, such as the antigen‐MHC complex and costimulatory molecules, which are capable of inducing specific anti‐tumour immune responses mainly through the DC‐dependent and DC‐independent pathways (Nikfarjam et al. 2020; Matsumoto et al. 2020; Lindenbergh and Stoorvogel 2018). Upon engulfment of DEVs by immature DCs, the loaded immunostimulatory molecules and antigens induce DC maturation. Subsequently, these mature DCs cross‐present the antigen via MHC molecules, thereby initiating antigen‐specific T‐cell responses. This pathway, in this case, is DC‐dependent and represents an indirect route to T‐cell activation. Conversely, in the DC‐independent mechanism, DEVs possess the capability to directly activate T cells, bypassing the requirement for DCs. This direct activation is mediated by the binding of antigen‐MHC complexes and costimulatory molecules present on the DEVs to their cognate receptors on T cells (Pitt et al. 2014; Vincent‐Schneider et al. 2002).

After the phenomenon and mechanism of T‐cell activation were elucidated, DEVs are increasingly regarded as a suitable alternative to overcome the negative impact of the tumour immune microenvironment, which can impede the stimulation of T‐cell responses observed in DC‐based immunotherapy (Zhang et al. 2023; Lindenbergh and Stoorvogel 2018; Pitt et al. 2014). Moreover, DEVs as cell‐free formations are easier to store and transport. Despite these advantages, the therapeutic efficacy of DEV‐based cancer vaccines is still limited, and only a small portion of cancer patients can successfully generate effective anti‐tumour immune responses through DEV vaccination (Viaud et al. 2010; Escudier et al. 2005). How to further develop DEVs to design more effective cancer vaccines remains a great challenge.

Cytokines, as a traditional method, have been explored for cancer treatment throughout the past century (Propper and Balkwill 2022; Tait Wojno et al. 2019). However, their clinical utilization remains hindered due to the severe systemic toxicity associated with therapeutic dosages (Propper and Balkwill 2022; Briukhovetska et al. 2021). Achieving precise delivery of cytokines is a key development direction in this field. Interleukin‐12 (IL‐12) is mainly secreted by DCs to promote the Th1 anti‐tumour immune response, which is one of the typical representatives of the application of cytokines in cancer therapy (Tait Wojno et al. 2019; Garris et al. 2022; Vom Berg et al. 2013; Li et al. 2017). DEVs, functioning as mediators of intercellular communication among immune cells, are expected to deliver IL‐12 to specific immune cells. Meanwhile, given that DEVs, as cell‐free formations, only possess the first and second signals required to activate T cells and cannot secrete cytokines like parent DCs do to furnish the third signal, the inclusion of cytokines such as IL‐12 is anticipated to compensate for the missing third signals necessary for direct T cell activation, thereby enhancing the therapeutic efficacy of DEVs. Therefore, the combination of DEVs and IL‐12 could be a complementary strategy. Immune checkpoint inhibitor (ICI), another emerging treatment method, have revolutionized the pattern of cancer therapy in recent years (Garris et al. 2022). ICI can effectively reverse exhausted anti‐tumour T cells in the tumour microenvironment (TME). Increasing the potential response rate of ICI is the main task of this treatment method (Ramos‐Casals et al. 2020; Sharma et al. 2021). Recent studies showed that part of the T cells in tumour‐draining lymph nodes (LNs) are bona fide responders to ICI (Huang et al. 2022; Dammeijer et al. 2020). Therefore, targeted delivery of ICI to LNs may improve the response rate of ICI. DEVs, as biological platforms, represent a promising strategy to achieve this goal (Nikfarjam et al. 2020; Li et al. 2023). Meanwhile, the inhibition of ICI on immune checkpoints contributes to optimizing the direct activation of DEVs on T cells. Therefore, the combination of DEVs and ICI may also be a mutually reinforcing strategy.

In this work, we first showed that cytokine IL‐12 and ICI (anti‐CTLA‐4 antibody, aCTLA‐4) can facilitate the direct activation of DEVs on T cells in vitro. Based on this result, we further constructed a sufficient signal supply platform for T cell activation by conjugating IL‐12 and aCTLA‐4 onto the surface of DEVs (DEV@IL‐12‐aCTLA‐4). DEV@IL‐12‐aCTLA‐4 can provide three major immune stimulation signals and block immune checkpoints for the activation of anti‐tumour T cells (Scheme 1). The nanoscale and LNs homing feature derived from DEVs endowed DEV@IL‐12‐aCTLA‐4 with the function of transporting IL‐12 and aCTLA‐4 to the lymphatic system, which was beneficial to accurately utilize IL‐12 and aCTLA‐4. In tumour‐bearing mice, DEV@IL‐12‐aCTLA‐4 stimulated the Th1 immune response and reversed exhausted CD8⁺ T cells in the TME, significantly inhibiting tumour growth in established tumour models. DEV@IL‐12‐aCTLA‐4 is expected to become a universal combination immunotherapy platform combining the triple methods of cancer vaccines, cytokines and immune checkpoint blockade.

SCHEME 1.

DEVs engineered with IL‐12 and aCTLA‐4 provide sufficient immunostimulation signals for combinational cancer immunotherapy. aCTLA‐4, anti‐CTLA‐4 antibody; DC, dendritic cell; DEV, dendritic cell‐derived extracellular vesicle; LN, lymph node; SC, subcutaneous injection; UV, ultraviolet.

2. Materials and Methods

2.1. Materials, Cell Lines and Mice

Chicken OVA(SIINFEKL), Trp2 (SVYDEFVWL) and Adpgk (ASMTNMELM) were purchased from QYAOBIO (Shanghai, China). Interferon‐γ (IFN‐γ) and recombinant mouse IL‐12 p70 were obtained from AntGene (21‐8311‐U500, 21‐8121‐U100, Wuhan, China). Lipopolysaccharide (LPS) was purchased from Solarbio (L8880, Beijing, China). 1,2‐Dipalmitoyl‐sn‐glycero‐3‐phosphoethanolamine‐PEG‐succinimidyl ester (DSPE‐PEG‐NHS, 3 kDa) was obtained from Aladdin (D163643‐100 mg, Shanghai, China). Immune checkpoint inhibitor (ICI, the InVivoMAb anti‐mouse cytotoxic T‐lymphocyte antigen 4 (CTLA‐4) antibody) was purchased from BioXcell (773921A1, New Hampshire, USA). Foetal bovine serum (FBS), RPMI‐1640 medium, DMEM medium and penicillin/streptomycin were purchased from Gibco (MA, USA). Cell membrane fluorescence probes DiO and DiI were obtained from Beyotime (C1038, C1036, Shanghai, China). All fluorescence‐conjugated antibodies used in the flow‐cytometric analysis were obtained from BioLegend (CA, USA) or BD Pharmingen (USA). The FITC antibody labelling kit and APC antibody labelling kit were obtained from Abcam (UK).

DC2.4 was a kind gift from Professor Guanxin Shen of the Department of Immunology, School of Basic Medicine, Tongji Medical College, Huazhong University of Science and Technology. Unless otherwise specified, DC2.4 is referred to as DC in the article. B16‐OVA cells were generated by stably transfecting B16/F10 with the pcDNA3‐OVA plasmid (Genecreale, Wuhan, China) following the previous report (Phung et al. 2020). The production of LLC‐OVA cells is similar to that of B16‐OVA cells. Cells were grown in DMEM or RPMI‐1640 supplemented with 10% FBS and 100 IU/mL penicillin/streptomycin and incubated at 37^°C in 5% CO₂‐95% air.

All female C57BL/6 mice used in this study were approved and monitored by the Institutional Animal Care and Use Committee, Huazhong University of Science and Technology (IACUC Number: 3439).

2.2. Preparation of Mature DC‐Derived Extracellular Vesicles

DC2.4 was cultured in a culture dish under 37°C and 5% CO2‐95% air. The culture medium used was RPMI‐1640 supplemented with 10% foetal bovine serum, 100 U/mL penicillin and 100 µg/mL streptomycin. Passage was carried out when the cell culture reached a density of 75%–90%. To induce the maturation of DC2.4, OVA (100 µg/mL), IFN‐γ (10 ng/mL) and LPS (1 µg/mL) were added to the culture medium to stimulate DC2.4 for 24 h. Then the culture medium was replaced with the fresh culture medium. When the cells were cultured to a certain number, the culture medium was replaced with the free‐exosome culture medium and exposed to ultraviolet (UVB, 300 Jm⁻²) at room temperature for 30 min. After 24 h of incubation, supernatants were collected, centrifuged for 10 min at 600 × g to get rid of cells, and then centrifuged for 30 min at 2000 × g to remove debris. The supernatant was further centrifuged for 60 min at 16,000 × g to pellet EVs (Guo et al. 2019). The pellet was resuspended in PBS and stored at −80°C until use. The DEVs used were as fresh as possible. If storage was required, freeze‐thaw cycles were avoided.

2.3. Preparation of DEV@IL‐12‐aCTLA‐4

The method and principle of protein conjugation to DEV membrane were mainly referred from previous reports (Zheng et al. 2022; Phung et al. 2020; Kooijmans et al. 2016). First, the DEV (1.75 × 10⁸ particles) was resuspended in 200 µL PBS. Then 1.5 µL DSPE‐PEG‐NHS (3 kDa, 3.5 mg/mL) was mixed and fully shaken. After incubating at 4°C for 24 h, the excess DSPE‐PEG‐NHS was removed by high‐speed centrifugation (16000 × g, 60 min) twice to obtain pure DEV‐DSEP‐PEG‐NHS. Then the DEV‐DSEP‐PEG‐NHS was resuspended in 200 µL PBS. IL‐12 (35 ng) and aCTLA‐4 (350 ng) were added to DEV‐DSEP‐PEG‐NHS in PBS, fully mixed, shaken and incubated at 4°C for 24 h. Then the unconjugated IL‐12 and aCTLA‐4 were removed by high‐speed centrifugation (16000 × g, 60 min) twice. The obtained DEV@IL‐12‐aCTLA‐4 (1.8 × 10¹¹ particles/mL) was resuspended and stored in PBS at −80°C.

2.4. Characterization of DEV@IL‐12‐aCTLA‐4

The morphological examination of DEVs and DEV@IL‐12‐aCTLA‐4 was detected by transmission electron microscopy (TEM). Briefly, DEV and DEV@IL‐12‐aCTLA‐4 samples were added to 200‐mesh grids (Zhongjingkeyi Technology, Beijing, China, AZH200) drop by drop, and stained with 2% phosphotungstic acid for 3 min. Then the excess liquid was removed, and the grids were observed under the TEM (Hitachi, HT7700). Nanoparticle tracking analysis (NTA) was used to analyse the size distribution of DEV@IL‐12‐aCTLA‐4. The laser wavelength was adjusted to 520 nm, and the filter wavelength was set to scatter to gain videos about size distribution. Zeta potential of samples was measured on a ZetaView (Particle Metrix, DEU). DEVs and DEV@IL‐12‐aCTLA‐4 were diluted to 8 × 10⁷ particles/mL in PBS and then tested on the instrument ZetaView. The electrolyte was pure H₂O, the temperature was 23.15°C and the pH was 7.0. For the instrument parameters, the laser wavelength was set to 520 nm, and the filter wavelength was scatter. For the detection of the main functional membrane on the DEV@IL‐12‐aCTLA‐4, the western blot was performed. DEV@IL‐12‐aCTLA‐4 was centrifuged at 16,000 × g at 4°C for 60 min and resuspended in 100 µL lysis buffer (20 mM Tris‐HCl, pH 7.4, 2 mM EDTA, 25 mM NaF, and 1% Triton X‐100) containing protease inhibitors (Sigma‐Aldrich) for 30 min at 4°C. The protein concentration was quantified using the BCA Protein Assay Kit (Beyotime). The samples were then probed with antibodies against Alix, CD80, CD86, MHC‐I, CD63 and OVA (all purchased from Proteintech). The major functional proteins were visualized with HRP‐conjugated secondary antibodies (Proteintech) and ECL plus (Proteintech).

To demonstrate the successful conjugation of IL‐12 and aCTLA‐4, we used the APC‐labelled IL‐12 and FITC‐labelled aCTLA‐4 to prepare DEV@IL‐12‐aCTLA‐4. The obtained DEV@IL‐12‐aCTLA‐4 membrane was labelled with DiI. Then the DEV@IL‐12‐aCTLA‐4 was imaged by confocal laser scanning microscopy (Nikon Ax). Dimensions: 1024 × 1024, 12 bits. Numerical Aperture: 0.8. Refractive index: 1. Calibration: 0.86 µm/px. Objective: PLAN APO λD 20× or 40× OFN25 DIC N2. Modality: Widefield Fluorescence, laser scan confocal. Detector mode: Muti‐Channel. Scanner: Galvano Unidirectional. Scan Mode: Band. Temperature of acquisition: room temperature.

For the isothermal titration calorimetry experiment, DEVs were first coupled with DSPE‐PEG‐NHS to form DEV‐DSPE‐PEG‐NHS, which was then added to the sample cell. The mixed solution of IL‐12 and aCTLA‐4 was slowly and rhythmically injected into the sample cell through the injection needle. Monitored the change in heat of the reaction using an isothermal titration calorimeter (MicroCal PEAQ‐ITC). Instrument parameters: the temperature was 25°C, reference power was 10 μcal/s, feedback was high, stir speed was 750 rpm, and initial delay was 60 s.

For protein desorption studies, ultracentrifugation was a standard method to deplete the protein corona around nanoparticles (Wolf et al. 2022). A centrifugation speed greater than 100,000 × g is required to effectively separate the protein corona from the EV surface (Wolf et al. 2022; Dietz et al. 2023; Liam‐Or et al. 2024). The fraction of DEVs that requires speeds greater than 16,000 × g for separation was already discarded during the DEV isolation phase. Thus, even if a higher speed (110,000 × g) was used in subsequent experiments, it will not centrifuge down the fraction of DEVs that require speeds greater than 16,000 × g for separation and will not cause any changes in the composition of DEVs. Specifically, DEV was conjugated with APC‐labelled IL‐12 and FITC‐labelled aCTLA‐4 using DSPE‐PEG‐NHS. The obtained DEV@APC‐IL‐12‐FITC‐aCTLA‐4 was pelleted via ultracentrifugation at 4°C, 110,000 × g, for 90 min to remove adsorbed proteins. Then, the obtained precipitate was stained with DiI and observed under confocal laser scanning microscopy (Nikon Ax). The imaging conditions were the same as previously described.

2.5. Quantification of IL‐12 and aCTLA‐4 in DEV@IL‐12‐aCTLA‐4 Surface

For the detection of IL‐12 in DEV@IL‐12‐aCTLA‐4, we used direct and indirect methods by ELISA. In direct quantification, the proteins in DEV@IL‐12‐aCTLA‐4 were obtained by lysis buffer (20 mM Tris‐HCl, pH 7.4, 2 mM EDTA, 25 mM NaF, and 1% Triton X‐100), and the absorbance of the samples at 450 nm was measured using an IL‐12 ELISA Kit (Abcam). In indirect quantification, the supernatant was retained in the step of conjugating IL‐12. Then the supernatant was added to the IL‐12 ELISA Kit to detect unconjugated IL‐12. The amount of conjugated IL‐12 can be obtained by subtracting the amount of unconjugated IL‐12 from the amount of added IL‐12.

For the detection of aCTLA‐4 in DEV@IL‐12‐aCTLA‐4, we used the Antibody FITC Labeling Kit (Abcam) to mark aCTLA‐4. We set the concentration gradient of aCTLA‐4 and measured their OD₄₉₀ to make the standard curve of ‘aCTLA‐4 quantity‐absorbance.’ Then the OD₄₉₀ of DEV@IL‐12‐aCTLA‐4 was measured to determine the amounts of aCTLA‐4 in DEV@IL‐12‐aCTLA‐4.

2.6. In Vitro Immunostimulatory Capability of DEV@IL‐12‐aCTLA‐4 to T Cells

To assess the binding of DEV@IL‐12‐aCTLA‐4 to T cells, spleen T cells were labelled with Hoechst‐nucleus dye and DiD‐membrane dye and then seeded in a cell culture plate. DiO‐labelled EV samples were added to the T cells. For the labelling process, the dye concentration used in our experiment is 2.5 µM. A dye enhancer and dye buffer were used in the labelling process to improve the solubility of the dye and reduce micelle formation. Then cleaned the cells or EVs by high‐speed centrifugation. Both staining and centrifugation were performed using the EP tube. After centrifugation, cells or EVs settled at the bottom, while some micelles settled on the side walls of the EP tube. Then remove the supernatant and gently scrape off the micelles on the side wall. Added dye buffer to resuspend the cells or EVs and performed centrifugation again, repeated 3 times. After co‐incubation of EVs and cells for 1 h incubation at 37°C, cells were washed with PBS and fixed by incubating with 4% paraformaldehyde. After washing with PBS, cells were imaged by confocal laser scanning microscopy (Nikon Ax). Dimensions: 256 × 256, 12 bits. Numerical Aperture: 0.8. Refractive index: 1. Calibration: 1.23 µm/px. Objective: PLAN APO λD 20× or 40× OFN25 DIC N2. Modality: Widefield Fluorescence, laser scan confocal. Detector mode: Muti‐Channel. Scanner: Galvano Unidirectional. Scan Mode: Band. Temperature of acquisition: room temperature. For quantitative analysis, the cells were collected and analysed by flow cytometry.

For the detection of T cell activation by DEV@IL‐12‐aCTLA‐4, spleen T cells were sorted and stimulated to proliferate to a sufficient number by anti‐mouse CD3 and anti‐mouse CD28 (2 µg/mL). Then T cells were co‐cultured with different DEV samples for 24 h in the RPMI‐1640 medium with 10% FBS, 100 IU/mL penicillin/streptomycin, and 50 IU hIL‐2. Next, we collected the cells and used the flow cytometer (BD A1 Flow Cytometer) to analyse the proportion of CD4⁺ T cells and CD8⁺ T cells and the expression of CD69 and IFN‐γ in CD4⁺ T cells and CD8⁺ T cells. The total IFN‐γ secreted in the culture supernatant was detected by a Mouse IFN‐γ ELISA kit (Abcam).

To analyse the proliferation of T cells, the MTS assay was performed. DEV or DEV@IL‐12 or DEV@aCTLA‐4 or DEV@IL‐12‐aCTLA‐4 were added to the splenocytes in a 96‐well plate. After incubation for 48 h, the MTS reagent was added to each well and incubated for 1 h at 37°C. The absorbance was detected using a microplate reader at 490 nm.

2.7. Lymph Node Homing of DEV@IL‐12‐aCTLA‐4 and Targeted Delivery of IL‐12 and aCTLA‐4 In Vivo

For lymph node (LN) homing studies, DEV and DEV@IL‐12‐aCTLA‐4 were labelled with DIR dye and then administrated to C57BL/6 mice by SC injection at the tail base (Liu et al. 2022; Zhang et al. 2018). The injection volume was only 50 µL. The mice were immobilized in a supine position using a fixing device, and the tail was cleaned with 75% ethanol. A mark was made approximately 25 mm from the base of the tail, and the injection was performed at this distance. The needle was positioned parallel to the tail and inserted subcutaneously 7 to 10 mm towards the body of the mice. Withdrew the needle before administering the injection to observe if there was blood. If there was no blood, it indicated that the syringe had not entered the tail vein. Firm pressure was applied at the site of needle entry to prevent any back‐leakage of the liquid during the injection. The liquid was slowly injected subcutaneously into the base of the mouse tail. After the injection, the needle was gently removed, and the injection site was pressed using a cotton swab to stop bleeding. At different time points after injection, bioluminescence imaging was conducted (Bruker MS FX Pro Imaging System, Bruker Corp) to capture the fluorescent signal of EV at the inguinal LN and other major organs. The biological distribution experiment process and intravenous injection were similar to the subcutaneous injection. The amount of DEV formation used for the two injection routes is completely equal. Imaging was performed at different time points after injecting DiR‐labelled DEV or DEV@IL‐12‐aCTLA‐4 into the tail vein of mice.

To assess the distribution of IL‐12 and aCTLA‐4 to LN, IL‐12 or aCTLA‐4 were marked by using the FITC antibody labelling kit before DEV@IL‐12‐aCTLA‐4 production. For IL‐12, FITC‐IL‐12 or DEV@FITC‐IL‐12‐aCTLA‐4 was administered to C57BL/6 mice by SC injection at the tail base. For aCTLA‐4, FITC‐aCTLA‐4 or DEV@IL‐12‐FITC‐aCTLA‐4 was administered to C57BL/6 mice by SC injection at the tail base. At different time points after injection, the mice were euthanized, and the inguinal LN and major organs were collected for imaging (Bruker MS FX Pro Imaging System, Bruker Corp).

2.8. In Vivo Immunostimulatory Capacity of DEV@IL‐12‐aCTLA‐4

DEV@IL‐12‐aCTLA‐4 or other controls (30 µg/50 µL/shot) were subcutaneously injected into non‐treated C57BL/6 mice at the tail base at the indicated time points. Three days after the last immunization, mice were sacrificed, and then the inguinal lymph nodes and spleen were collected to make a single‐cell suspension. Flow antibodies like PE‐IFN‐γ, BV786‐CD3, FITC‐CD4, APC‐CD8, and PE/cy7‐CD69 were used to make the T cells. Using the flow cytometer (BD A1 Flow Cytometer) to analyse the results.

To assess the specific anti‐tumour immunity induced by DEV@IL‐12‐aCTLA‐4, the ELISPOT assay was conducted after immunization. Splenocytes selected from the spleen of immunized C57BL/6 mice were transferred to ELISPOT plates (2×10⁵ per well) and re‐stimulated with OVA (2 µg/mL). After 24 h of incubation, the cells were washed away, and the plates were incubated with biotinylated anti‐IFN‐γ antibody for 2 h at room temperature. Then, the ELISPOT plates were washed with deionized water and incubated for 1.5 h at room temperature with streptavidin–alkaline phosphatase followed by washing with deionized water. Finally, the ELISPOT plates were incubated with the BCIP/NBT substrate solution at room temperature until spots emerged. The colour development was stopped by repeated washings with deionized water. After drying, the spots were counted with an ELISPOT reader using ELISPOT software.

For the CTL assay, splenocytes separated from the spleen of immunized C57BL/6 mice were seeded in 96‐well plates as effector cells (E). Then B16‐OVA cells were separated as target cells (T) and incubated with effector cells (E) at the appropriate ratios of 12.5:1, 25:1, 50:1, and 100:1 for 24 h. The LDH levels in cell culture supernatants were tested by the CTL assay kit following the manufacturer's instructions (Beyotime Biotechnology).

2.9. In Vivo Anti‐Tumour Effect of DEV@IL‐12‐aCTLA‐4

All the animals used in the anti‐tumour study were female 6–8‐week‐old C57BL/6 mice. For the melanoma model, 1 × 10⁵ B16‐OVA cells were subcutaneously injected into the right flank of mice. For the LLC‐OVA model, 5 × 10⁵ LLC‐OVA cells were subcutaneously injected into the right flank of mice. Then the mice were divided randomly into 5 groups and immunized with different samples (30 µg per mouse) by subcutaneous injection at the tail base at the indicated time points. Tumour growth and body weight were monitored once every 2 days. Tumour volume was calculated using the equation: tumour volume = length × width² × 0.5. Three days after the third treatment, mice were humanely sacrificed, and the tumour, blood sample, and main organs were collected for further evaluation.

To assess intratumoural immunophenotypes, the tumour tissues were excised, cut into small pieces of 2–4 mm, and incubated in a digestion medium containing hyaluronidase (3 mg/mL, Biofroxx, 1141MG100) and collagenase IV (2 mg/mL, Biofroxx, 2091MG100) for 1 h at 37°C. The obtained cell suspension was passed through a 70 µm strainer and washed with PBS. The single cells were then co‐cultured with the medium containing PMA (0.1 µg/mL, Solarbio, P6741‐1 mg), ionomycin (1 µg/mL, Aladdin, 1139530–1 mg), and protein transport inhibitor (1 µL/mL, BD Pharmingen) for 5 h. After that, cells were blocked by anti‐mouse CD16/32 (BioLegend, 101319) and stained with Fixable Viability Stain 510 (BD Pharmingen, 564406) as a marker to identify live or dead cells. For surface antibody staining, CD45 (BioLegend, 103116), BV786‐CD3 (BD Pharmingen, 564010), BV421‐CD4 (BD Pharmingen, 562891), APC‐CD8 (Elabscience, E‐AB‐F1104E), and BV650‐PD‐1 (BD Pharmingen, 744546) were used. Afterwards, cells were washed and resuspended in the Transcription Factor Buffer (BD Biosciences, 2137553) for 30 min. Then cells were washed with Intracellular Staining Perm Wash Buffer (BD Biosciences, 2231080), stained with PE‐IFN‐γ (BD Pharmingen, 554412) and FITC‐Granzyme B (Invitrogen, 2527205), and followed by flow cytometric analysis.

2.10. Detection of Serum Cytokines and Biochemical Indicators

Peripheral blood of immunized mice was obtained by retro‐orbital puncture and centrifuged at 1200 rpm at 10 min. Then the supernatant serum was collected and stored at −80°C. Using the mouse IFN‐γ and IL‐10 ELISA kit (Abcam) to analyse the serum IFN‐γ and IL‐10 levels. Detected red blood cell (RBC), haemoglobin (HGB), aspartate aminotransferase (AST), alanine aminotransferase (ALT), total bilirubin (TBil), creatinine (Cr), and creatine kinase (CK) according to the manufacturer's instructions (Nanjing Jiancheng Bioengineering Institute).

2.11. Statistical Analysis

All experiments were repeated independently at least three times. All repeats in our study are biological repetitions. Except for the data of some mice that died before the completion of animal experiments, all repeated data were included in statistical analysis. Specifically, data were analysed using GraphPad Prism 9.0 and reported as mean ± SD. Data were checked by Gaussian distribution and homogeneity tests before analysis. For pairwise comparisons, we employed t‐tests, while one‐way analysis of variance (ANOVA) was used for comparisons involving more than two groups. Survival analyses were conducted using the log‐rank (Mantel–Cox) test. p < 0.05 was considered statistically significant.

3. Results

3.1. Enhancing the Capability of DEVs to Direct T Cell Activation In Vitro by IL‐12 and aCTLA‐4

To obtain the mature DC‐derived EVs (DEVs), DC2.4 was co‐cultured with neoantigen OVA, LPS and IFN‐γ. OVA was the model antigen of tumour neoantigens in this study. Flow cytometry showed the maturation of DC2.4 (Figures S1A–B). Then the mature DCs were exposed to UV irradiation to increase EV production, and the culture medium was collected to obtain DEVs by gradient centrifugation (Guo et al. 2019; Ruan et al. 2022). The results of DEV quantification through nanoparticle tracking analysis (NTA) and bicinchoninic acid (BCA) assay showed that 3.2 × 10¹¹ particles (8.4 µg proteins/10¹⁰ particles)of DEVs could be obtained from 10⁸ DCs (Figure S1C). The high‐yield DEVs obtained through UV irradiation had no physical property or T cell activation capability difference from natural DEVs (Figures S2A–D). DEVs were characterized based on morphology, size and protein content (Figures 1A–D). DEVs had a cup‐shaped morphology (Figure 1A), and the mean diameter was approximately 173 nm (Figures 1B–C). The protein profile of DEVs was similar to that of parental DCs (Figure 1D). The obtained DEVs expressed common marker proteins of EVs such as CD81, EPCAM, TSG101, CD63 and Alix, but not calnexin, indicating the purity of the obtained particles. Moreover, DEVs contained OVA and MHC as the first signal, and co‐stimulatory molecules (CD80 and CD86) as the second signal for T cell activation (Figure 1D and Figure S3).

FIGURE 1.

Enhance the capability of DEVs in direct T cell activation by IL‐12 and aCTLA‐4. (A) The images of DEVs were observed by transmission electron microscopy. Scale bars, 200 nm. (B) Size distribution of DEVs measured by nanoparticle tracking analysis (NTA). (C) Size distribution of DEVs measured by nanoflow cytometry. (D) The protein profile of DEV (right) and parental DC (left) was detected by western blot. (E)–(F) Analysis of CD80 and CD86 expression on BMDCs treated with PBS, free IL‐12, free aCTLA‐4, DEV, DEV mixed with IL‐12 (DEV + IL‐12), DEV mixed with aCTLA‐4 (DEV + aCTLA‐4), DEV mixed with IL‐12 and aCTLA‐4 (DEV + IL‐12 + aCTLA‐4) by flow cytometry. Mean ± SD, n = 5, one‐way ANOVA analysis. (G) Detection of TNF‐α levels secreted by BMDCs after 24 h of treatment with different formations using enzyme‐linked immunosorbent assay (ELISA). Mean ± SD, n = 5, one‐way ANOVA analysis. (H)–(I) Analysis of CD69 expression on CD4⁺ T and CD8⁺ T cells after 24 h of treatment with PBS, free IL‐12, free aCTLA‐4, DEV, DEV mixed with IL‐12 (DEV + IL‐12), DEV mixed with aCTLA‐4 (DEV + aCTLA‐4), DEV mixed with IL‐12 and aCTLA‐4 (DEV + IL‐12 + aCTLA‐4) by flow cytometry. Mean ± SD, n = 5, one‐way ANOVA analysis. (J) Proliferation of T cells measured using the MTS assay after 24 h of treatment with various formulations. Living cells reduce MTS to formazan, and the absorbance of formazan can be measured at approximately 490 nm (Abs. 490 nm). The absorbance value at 490 nm is directly proportional to the number of living cells, which reflects cell proliferation. Mean ± SD, n = 7, one‐way ANOVA analysis. (K) Detection of IFN‐γ levels secreted by T cells in the supernatant of the culture medium after 24 h of treatment with different formations using ELISA. Mean ± SD, n = 5, one‐way ANOVA analysis.

To explore the impact of IL‐12 and aCTLA‐4 on the DC‐dependent pathway, DEVs mixed with IL‐12 or aCTLA‐4 were used to stimulate immature BMDCs. The expression of CD80 and CD86 was detected by flow cytometry to reflect the maturation of BMDCs. The results showed that DEVs could promote BMDC maturation, but the existence of IL‐12 and aCTLA‐4 did not affect this process (Figures 1E–F). We also measured the secreted TNF‐α of BMDCs after stimulation. The results were consistent with the expression of CD80 and CD86; that is, IL‐12 and aCTLA‐4 did not affect the TNF‐α secretion induced by DEVs (Figure 1G), which means the existence of IL‐12 and aCTLA‐4 had no intervention on the DC‐dependent pathway. For the DC‐independent pathway, DEVs mixed with IL‐12 or aCTLA‐4 were used to directly stimulate spleen T cells isolated from C57BL/6 mice. We examined the activation of T cells by detecting the expression of CD69, a marker of T‐cell activation. Flow cytometry results showed that IL‐12 and aCTLA‐4 enhanced the capability of DEVs to activate T cells (Figures 1H–I). Cell proliferation experiments also indicated that IL‐12 and aCTLA‐4 promoted the function of DEVs in T‐cell proliferation, and DEVs mixed with IL‐12 and aCTLA‐4 induced the strongest T‐cell proliferation (Figure 1J). Then, we measured the cytokine IFN‐γ in conditioned media of DEV‐treated T cells. Similar to the observations on CD69 expression, the group of DEVs mixed with IL‐12 and aCTLA‐4 showed higher IFN‐γ secretion compared to DEVs alone (Figure 1K). All these results proved that the existence of IL‐12 and aCTLA‐4 enhanced the capability of DEVs in directly activating T cells but without exerting any influence on the DC‐dependent pathway.

3.2. Synthesis and Characterization of DEV@IL‐12‐aCTLA‐4

Based on the above findings and the fact that DEVs only have the first and second signals of T‐cell activation and lack other immunoregulatory signals, IL‐12 and aCTLA‐4 were conjugated onto the surface of DEVs to enhance the ability to directly activate T‐cell immune responses by providing more sufficient immune stimulation signals. The conjugation of IL‐12 and aCTLA‐4 was achieved through the coupling action of DSPE‐PEG‐NHS. DSPE‐PEG‐NHS containing alkyl chains was capable of anchoring to the lipid raft in the DEV membrane. At the same time, the N‐Hydroxysuccinimide (NHS) reactive groups of DSPE‐PEG‐NHS can be conjugated with proteins like IL‐12 and aCTLA‐4 through the amine‐coupling reaction (Zheng et al. 2022; Fan et al. 2022). The conjugation of DSPE‐PEG‐NHS did not affect the T‐cell activation ability of DEVs (Figure 2A).

FIGURE 2.

Synthesis and characterization of DEV@IL‐12‐aCTLA‐4. (A) Detection of T cell activation effects of natural DEV and DEV‐D (DEV conjugated with DSPE‐PEG‐NHS) by flow cytometry. Mean ± SD, n = 3, one‐way ANOVA analysis. (B) The images of DEV@IL‐12‐aCTLA‐4 were observed by transmission electron microscopy. (C) Comparing the size distribution of DEV and DEV@IL‐12‐aCTLA‐4 by NTA. (D) The protein profile of DEV (left) and DEV@IL‐12‐aCTLA‐4 (right) was detected by western blot. (E) The images of fluorescence‐labelled IL‐12 (APC) and aCTLA‐4 (FITC) on DEV@IL‐12‐aCTLA‐4 (DiI). Extracellular vesicles were labelled by Dil, IL‐12 was labelled by APC, and aCTLA‐4 was labelled by FITC. Scale bars, 100 nm. Scale bars, 5 µm. (F) Detection of IL‐12 (APC) and aCTLA‐4 (FITC) on DEV@IL‐12‐aCTLA‐4 by nanoflow cytometry. (G) Quantitative analysis of IL‐12 on DEV@IL‐12‐aCTLA‐4 via ELISA. Mean ± SD, n = 3. (H) Quantitative analysis of aCTLA‐4 on DEV@IL‐12‐aCTLA‐4 by measuring the standard curve of antibody quantity‐absorbance. Mean ± SD, n = 3.

The obtained DEV@IL‐12‐aCTLA‐4 was characterized based on morphology, size and protein content (Figures 2B–D). The shape of DEV@IL‐12‐aCTLA‐4 had not significantly changed compared to DEVs and still presents a cup‐shaped morphology (Figure 2B). NTA revealed that the average size of DEV@IL‐12‐aCTLA‐4 was 174 nm, which was similar to that that of DEVs without significant changes (Figure 2C). The cell surface markers of DEV@IL‐12‐aCTLA‐4 were assessed by western blot. The protein profile of DEV@IL‐12‐aCTLA‐4 was also similar to that of DEVs (Figure 2D). These results suggested that the conjugation of IL‐12 and aCTLA‐4 did not significantly affect the characteristics of DEVs, and DEV@IL‐12‐aCTLA‐4 retained the basic features of DEVs.

Next, we further confirmed the existence of IL‐12 and aCTLA‐4 on the membrane of DEV@IL‐12‐aCTLA‐4. Before preparing DEV@IL‐12‐aCTLA‐4, the DEV membrane was labelled with DiI, IL‐12 was marked with APC, and aCTLA‐4 was labelled with FITC. The image of DEV@IL‐12‐aCTLA‐4 obtained through confocal laser scanning microscopy showed that the fluorescence of APC‐IL‐12 and FITC‐aCTLA‐4 was localized on the membrane surface of DEV@IL‐12‐aCTLA‐4 (Figure 2E). Nanoflow cytometry also indicated DEV@IL‐12‐aCTLA‐4 was APC positive and FITC positive, while DEV was not (Figure 2F). To further validate the presence of IL‐12 and aCTLA‐4 on the membrane of DEV@IL‐12‐aCTLA‐4, we extracted the membrane proteins of DEV@IL‐12‐aCTLA‐4 and performed western blot and Coomassie blue staining. The results showed that the membrane proteins of DEV@IL‐12‐aCTLA‐4 contained IL‐12 and aCTLA‐4 (Figure S5A).

We sought to understand whether IL‐12 and aCTLA‐4 were conjugated to the DEVs through stable chemical reactions or adhesion. Firstly, we compared the fluorescence expression of IL‐12 and aCTLA‐4 in DEVs with and without the addition of DSPE‐PEG‐NHS by confocal laser scanning microscopy. The findings indicated that the introduction of DSPE‐PEG‐NHS significantly improved the anchoring efficiency of IL‐12 and aCTLA‐4 (Figure S5B). The isothermal titration calorimetry experiment revealed that the binding process after adding DSPE‐PEG‐NHS as a coupling agent underwent a heat‐related chemical reaction (Figure S5C), indicating that IL‐12 and aCTLA‐4 were conjugated in the DEV membrane through chemical reactions to form stable chemical bonds. Moreover, protein desorption cannot disrupt the binding of IL‐12 and aCTLA‐4 on DEV@IL‐12‐aCTLA‐4 (Figure S5D). All these results confirmed the successful anchoring of IL‐12 and aCTLA‐4 on the DEV@IL‐12‐aCTLA‐4 membrane through the coupling action of DSPE‐PEG‐NHS rather than adhesion. To quantify the conjugation amount of IL‐12, the enzyme‐linked immunosorbent assay (ELISA) was applied. Results showed that the average amount of IL‐12 was 24.5 ng in 1.75 × 10⁸ particles of DEV@IL‐12‐aCTLA‐4 (Figure 2G). Detecting the relationship between the amount of conjugation and the addition of IL‐12 found that the conjugation efficiency of IL‐12 was approximately 70% (Figure S5E). The presence of aCTLA‐4 on the DEV@IL‐12‐aCTLA‐4 was assessed by conjugating the FITC‐labelling aCTLA‐4. Detecting the fluorescence of DEV@IL‐12‐FITC‐aCTLA‐4 proved that the average amount of aCTLA‐4 was 205.7 ng in 1.75 × 10⁸ particles of DEV@IL‐12‐aCTLA‐4 (Figure 2H). The conjugation efficiency of aCTLA‐4 was approximately 60% (Figure S5F). The successful conjugation of IL‐12 and aCTLA‐4 constituted the foundation of DEV@IL‐12‐aCTLA‐4 with stronger T cell direct activation capability. As for the stability test, we examined the changes of DEV@IL‐12‐aCTLA‐4 (1.8 × 10¹¹ particles/mL) when stored in PBS at −80°C over a period of 7 days. The results showed there was no obvious change in morphology, size, zeta potential and particle numbers (Figure S6), indicating that DEV@IL‐12‐aCTLA‐4 performed with good stability.

3.3. Directly Activating T Cells by DEV@IL‐12‐aCTLA‐4 In Vitro

First, we assessed the direct binding of DEV@IL‐12‐aCTLA‐4 to T cells by labelling T cells and DEV@IL‐12‐aCTLA‐4 with different dyes. The images of confocal laser scanning microscopy showed that DEV@IL‐12‐aCTLA‐4 could directly bind to T cells after co‐incubation for 1 h (Figure 3A). The results of flow cytometry suggested that the binding was dose‐dependent on DEV@IL‐12‐aCTLA‐4 (Figure S7B). We next assessed the activation and proliferation of T cells after DEV@IL‐12‐aCTLA‐4 stimulation in vitro. The results of flow cytometry showed that the DEV@IL‐12‐aCTLA‐4 group had the highest CD69 expression in both CD4⁺ T cells and CD8⁺ T cells compared with other groups (Figures 3B–C). Cytokines like IFN‐γ appeared along with the activation of T cells, so we also detected the expression of IFN‐γ in T cells. Consistent with the result of CD69 expression, the DEV@IL‐12‐aCTLA‐4 group exhibited the highest proportion of IFN‐γ⁺CD4⁺ T cells and IFN‐γ⁺CD8⁺ T cells when compared to the other groups (Figures 3D–E). The results of ELISA also showed that the T cells in the DEV@IL‐12‐aCTLA‐4 group had the highest secretion of IFN‐γ (Figure 3F). The main function of IL‐12 is to induce the production of a Th1 immune response and maintain a robust increase in IFN‐γ. The highest proportion of IFN‐γ⁺ CD4⁺ T cells and the highest secretion of IFN‐γ indicated that DEV@IL‐12‐aCTLA‐4 inherits the property of IL‐12 and performs a strong capability to induce a Th1 immune response. At the same time, we also detected the corresponding Th2 response, and the results showed no differences among all groups (Figure S7C). Next, we checked the proliferation of T cells after co‐incubation with DEV@IL‐12‐aCTLA‐4. We observed that DEV@IL‐12‐aCTLA‐4 had a significantly stronger capacity to promote T cell proliferation compared to DEV and other formulations (Figure 3G). All these results demonstrated that DEV@IL‐12‐aCTLA‐4 can directly bind to T cells and promote T cell activation. Importantly, the stronger functions of DEV@IL‐12‐aCTLA‐4 compared to DEV are attributed to the conjugations of IL‐12 and aCTLA‐4. To demonstrate that the effect was DEV specific, we conjugated IL‐12 and aCTLA‐4 to EVs derived from HEK‐293T cells or mouse bone marrow mesenchymal stem cells (BMSCs). The results of flow cytometry showed that these non‐DC cell‐derived EVs conjugated with IL‐12 and aCTLA‐4 cannot directly activate T cells like DEVs (Figure S8).

FIGURE 3.

DEV@IL‐12‐aCTLA‐4 directly activated T cells in vitro. (A) The images of direct binding between DEV@IL‐12‐aCTLA‐4 and T cells. The nucleus of T cells was labelled with Hoechst, the cell membrane of T cells was labelled with DiD, and DEV@IL‐12‐aCTLA‐4 was labelled with DIO. Scale bars, 10 µm. (B)–(C) Analysis of CD69, a T‐cell activation marker, on CD4⁺ and CD8⁺ T cells following incubation with the PBS, DEV, DEV@IL‐12, DEV@aCTLA‐4, or DEV@IL‐12‐aCTLA‐4 by flow cytometry. Mean ± SD, n = 5, one‐way ANOVA analysis. (D)–(E) Detection of the proportion of IFN‐γ⁺ cells in CD4⁺ T cells and CD8⁺ T cells following incubation with the PBS, DEV, DEV@IL‐12, DEV@aCTLA‐4, and DEV@IL‐12‐aCTLA‐4 by flow cytometry. (F) Detection of IFN‐γ levels secreted by T cells after 24 h of treatment with different formations using ELISA. Mean ± SD, n = 5, one‐way ANOVA analysis. (G) Proliferation of T cells was measured using the MTS assay. Mean ± SD, n = 7, one‐way ANOVA analysis.

3.4. Lymph Node Accumulation and Delivery of IL‐12 and aCTLA‐4 by DEV@IL‐12‐aCTLA‐4 In Vivo

DEVs exhibit the capability to specifically target parent cells within LNs, and their nanoscale size also is beneficial for trafficking to LNs after subcutaneous injection. In theory, DEV@IL‐12‐aCTLA‐4 should possess a comparable LN trafficking ability to that of DEVs. To verify this, we examined the biodistribution of DEV@IL‐12‐aCTLA‐4 in vivo. DiR‐labelled DEVs or DEV@IL‐12‐aCTLA‐4 were subcutaneously injected into C57BL/6 mice at the base of the tail, and their penetration into the inguinal LNs was detected by imaging. The results showed efficient accumulation of DEV@IL‐12‐aCTLA‐4 in inguinal LNs, which was similar to DEVs (Figure 4A). At different time points after injection, the mice were euthanized, and their inguinal LNs and main organs were collected and imaged. In inguinal LNs, DEV@IL‐12‐aCTLA‐4 exhibited the same accumulation with DEVs in 12, 24, and 48 h (Figure 4B). Typically, in both the DEV@IL‐12‐aCTLA‐4 group and the DEV group, the strongest fluorescence was observed in the inguinal LNs, whereas no significant fluorescence distribution was observed in other major organs except the liver (Figure 4C). These observations demonstrated that DEV@IL‐12‐aCTLA‐4 retained the characteristics of DEVs and performed LNs targeting capability after subcutaneous injection. We also examined the biological distribution of DEV@IL‐12‐aCTLA‐4 following intravenous injection with equal DEV@IL‐12‐aCTLA‐4. The results indicated that, in contrast to subcutaneous injection, DEV@IL‐12‐aCTLA‐4 predominantly accumulated in the liver, spleen and lungs after intravenous injection. However, its presence in lymph nodes was notably lower compared to these organs (Figure S9). Compared to these two injection methods, subcutaneous injection was more beneficial for DEV@IL‐12‐aCTLA‐4 to achieve LN‐targeted delivery, so subsequent experiments used subcutaneous injection. To further verify that DEV@IL‐12‐aCTLA‐4 effectively targeted T cells within LNs and induced direct anti‐tumour immune responses following subcutaneous injection, we assessed the presence of DiO⁺CD4⁺ and DiO⁺CD8⁺ T cells in inguinal LNs by flow cytometry after administering various DiO‐labelled DEV formulations. As expected, we observed a certain proportion of DiO⁺CD4⁺ and DiO⁺CD8⁺ T cells in the DEV@IL‐12‐aCTLA‐4 group, indicating specific targeting of T cells within LNs by DEV@IL‐12‐aCTLA‐4 (Figures 4D–E).

FIGURE 4.

Lymph node accumulation and delivery of IL‐12 and aCTLA‐4 in vivo. (A) Inguinal LN (ILN) imaging of C57BL/6 mice after subcutaneous injection of PBS, DiR‐labelled DEV, or DiR‐labelled DEV@IL‐12‐aCTLA‐4 at the base of the tail. (B) Fluorescence signals in the draining inguinal LNs of mice treated with PBS, DiR‐labelled DEV, or DiR‐labelled DEV@IL‐12‐aCTLA‐4 were quantified with IVIS. Mean ± SD, n = 3, one‐way ANOVA analysis. (C) Analysis of the fluorescence signals in other major organs. Mean ± SD, n = 3, one‐way ANOVA analysis. (D‐E) Detection of DiO⁺ cells in CD4⁺ and CD8⁺ T cells in LNs by flow cytometry after DiO‐labelled DEV or DiR‐labelled DEV@IL‐12‐aCTLA‐4 administration. Mean ± SD, n = 3, one‐way ANOVA analysis. (F) Distribution of FITC‐IL‐12 in 24 h after subcutaneous injection of PBS, FITC‐IL‐12, or DEV@ FITC‐IL‐12‐aCTLA‐4 at the base of the tail. Mean ± SD, n = 3, one‐way ANOVA analysis. (G) Distribution of FITC‐aCTLA‐4 in 24 h after subcutaneous injection of PBS, FITC‐aCTLA‐4, or DEV@IL‐12‐FITC‐aCTLA‐4 at the base of the tail. Mean ± SD, n = 3, one‐way ANOVA analysis. ^* p < 0.05, ^** p < 0.01, ^*** p < 0.001, ^**** p<0.0001, one‐way ANOVA analysis.

To explore whether DEV@IL‐12‐aCTLA‐4 can deliver IL‐12 and aCTLA‐4 to LNs, we used fluorescence FITC to label IL‐12 or aCTLA‐4 to prepare DEV@IL‐12‐aCTLA‐4. Then the DEV@FITC‐IL‐12‐aCTLA‐4 or DEV@IL‐12‐FITC‐aCTLA‐4 was injected subcutaneously into mice at the base of the tail. The inguinal LNs and other major organs were collected to analyse the distribution of IL‐12 and aCTLA‐4. Results showed that DEV@IL‐12‐aCTLA‐4 facilitated the delivery of IL‐12 and aCTLA‐4 to LNs, exhibiting stronger delivery than free IL‐12 or aCTLA‐4 (Figures 4F–G). In addition, there was no distribution of IL‐12 or aCTLA‐4 in other major organs in the DEV@IL‐12‐aCTLA‐4 group except the liver, whereas in mice injected with free IL‐12 or aCTLA‐4, fluorescence was observed in the heart, spleen, or lung. These results indicated that DEV@IL‐12‐aCTLA‐4 can achieve the target delivery of IL‐12 and aCTLA‐4 to LNs, which is beneficial to more accurately utilize IL‐12 and aCTLA‐4.

3.5. Inducing Specific Anti‐Tumour Immunity by DEV@IL‐12‐aCTLA‐4 In Vivo

Based on our previous results of activating T cells in vitro and targeting immune cells in LNs in vivo, we next investigated the in vivo stimulation of CD4⁺ and CD8⁺ T cells after administration of the DEV@IL‐12‐aCTLA‐4 into mice bearing B16‐OVA tumours. We first assessed the production of local anti‐tumour immunity in LNs. By assessing the CD69 expression of T cells in LNs, we observed that DEV@IL‐12‐aCTLA‐4 induced significant CD4⁺ and CD8⁺ T cell activation in LNs. Compared with other groups, the DEV@IL‐12‐aCTLA‐4 group showed the highest expression of CD69 in T cells (Figures S10A–C). Consistent with that, the proportion of IFN‐γ⁺CD4⁺ T cells and IFN‐γ⁺CD8⁺ T cells in the DEV@IL‐12‐aCTLA‐4 group showed the highest levels in local LNs (Figures S10D–F). In contrast to its impact on the promotion of Th1 immunity (IFN‐γ⁺CD4⁺ T cells), DEV@IL‐12‐aCTLA‐4 did not affect the Th2 immune response (IL‐4⁺CD4⁺ T cells) within LNs (Figure S10D). Then we assessed the systemic anti‐tumour immunity response in vivo. Consistent with local lymphatic immunity, anti‐tumour immune activation, including Th1 immunity, was also detected in the spleen. The DEV@IL‐12‐aCTLA‐4 group showed the highest expression of CD69 and IFN‐γ in both CD4⁺ and CD8⁺ T cells (Figures 5A–E), while showing no effect on the Th2 immune response indicated by IL‐4⁺CD4⁺ T cells (Figure S11A). These results suggested that DEV@IL‐12‐aCTLA‐4 was more capable of activating T cells and inducing Th1 immune responses than any other formulations. By comparing the differences between DEVs and DEV@IL‐12‐aCTLA‐4, we found that IL‐12 and aCTLA‐4 conjugation enhanced the anti‐tumour immune induction effect of DEVs in vivo. Next, we evaluated the responsiveness of anti‐tumour immunity induced by DEV@IL‐12‐aCTLA‐4 to tumour‐specific antigens through ex vivo re‐stimulation with OVA. Splenocytes isolated from treated mice were re‐stimulated with OVA. The ELISPOT results showed that the DEV@IL‐12‐aCTLA‐4 group had the strongest immune response (Figure 5F). Correspondingly, the highest proportion of OVA‐tetramer⁺CD8⁺ T cells in the spleen of immunized mice was also found in the DEV@IL‐12‐aCTLA‐4 group by detecting neoantigens specific to CD8⁺ T cells (Figure 5G). Next, we assessed the specific killing ability of activated T cells on tumour cells. The activated T cells were co‐incubated with tumour cells and then lactic acid level was detected to reflect the apoptosis of tumour cells. Results showed that tumour cells expressing OVA were specifically killed, especially in the DEV@IL‐12‐aCTLA‐4 group (Figure 5H), while tumour cells lacking OVA expression were not killed (Figure S11B). These data implied that DEV@IL‐12‐aCTLA‐4 can promote T cell activation and elicit a robust and effective antigen‐specific immune response in vivo.

FIGURE 5.

DEV@IL‐12‐aCTLA‐4 induced specific anti‐tumour immunity in vivo. (A‐C) Analysis of CD69 on CD4⁺ and CD8⁺ T cells from the spleen of C57BL/6 mice following immunization with the PBS, DEV, DEV@IL‐12, DEV@aCTLA‐4, or DEV@IL‐12‐aCTLA‐4. Mean ± SD, n = 5, one‐way ANOVA analysis. (D)–(E) Detection of the proportion of IFN‐γ⁺ cells in CD4⁺ T cells and CD8⁺ T cells by flow cytometry from the spleen of immunized mice. Mean ± SD, n = 5, one‐way ANOVA analysis. (F) Responsiveness of specific anti‐tumour immunity in the treated mice assessed by ELISPOT. Mean ± SD, n = 3, one‐way ANOVA analysis. (G) Detection of OVA‐specific CD8⁺ T cells in the spleen measured by flow cytometry analysis of tetramer⁺ CD8⁺ T cells. Mean ± SD, n = 5, one‐way ANOVA analysis. (H) LDH release assay to analyse the specific lysis with different effector‐to‐target ratios. Splenocytes separated from the spleen of immunized C57BL/6 mice as effector cells and B16‐OVA as target cells. Mean ± SD, n = 3, one‐way ANOVA analysis.

3.6. Therapeutic Anti‐Tumour Efficacy and Biosafety of DEV@IL‐12‐aCTLA‐4

Insufficient treatment effect is a bottleneck of cancer vaccines. To investigate whether DEV@IL‐12‐aCTLA‐4 can induce an effective anti‐tumour effect in pre‐established tumours, a B16‐OVA tumour‐bearing mice model was constructed before vaccine immunization. Then PBS, DEV@IL‐12‐aCTLA‐4, or other formations were administered three times at intervals of 5 days to tumour‐bearing mice when the tumour volume reached approximately 80 mm³ (Figure 6A). Continuous monitoring of tumour volume showed that compared with free DEV, DEV@IL‐12 and DEV@aCTLA‐4 exhibited a stronger tumour growth inhibitory effect in the B16‐OVA tumour model. Meanwhile, DEV@IL‐12‐aCTLA‐4, which is modified with both IL‐12 and aCTLA‐4, can further enhance the therapeutic effect (Figure 6B and Figure S12A). Moreover, B16‐OVA tumour‐bearing mice in DEV@IL‐12‐aCTLA‐4 had the longest survival period (Figure 6C). To explore DEV@IL‐12‐aCTLA‐4's efficacy against poorly immunogenic tumours, the LLC‐OVA tumour‐bearing mice model was constructed before vaccine immunization, following a treatment protocol similar to that used for the previous B16‐OVA model (Figure 6D). Continuous monitoring of tumour volume showed that DEV@IL‐12‐aCTLA‐4 also significantly delayed LLC‐OVA tumour growth. The modification of IL‐12 and aCTLA‐4 also enhanced the therapeutic effect of DEV in the LLC‐OVA tumour model (Figure 6E and Figure S12B). LLC‐OVA tumour‐bearing mice in the DEV@IL‐12‐aCTLA‐4 group had the longest survival period (Figure 6F). To explore the broader applicability of DEV@IL‐12‐aCTLA‐4 beyond the OVA neoantigen, we sought to assess its effectiveness by introducing other neoantigens. To achieve this, we employed the mouse melanoma neoantigen Trp2 or the colon cancer neoantigen Adpgk to induce DC maturation. Subsequently, we isolated DEVs and synthesized DEV@IL‐12‐aCTLA‐4 for the treatment of B16 tumour‐bearing mice or MC38 tumour‐bearing mice corresponding to the respective neoantigens. The results showed that the DEV@IL‐12‐aCTLA‐4 could also effectively inhibit the growth of corresponding B16 tumours or MC38 tumours (Figures S12C–D). indicating the universality of DEV@IL‐12‐aCTLA‐4 for various neoantigens.

FIGURE 6.

Therapeutic efficacy and safety on tumour‐bearing mice. (A) Administration routes for the B16‐OVA model. (B) The subcutaneous tumour volume of B16‐OVA after PBS, DEV, DEV@IL‐12, DEV@aCTLA‐4, or DEV@IL‐12‐aCTLA‐4 treatment. Mean ± SD, n = 5, one‐way ANOVA analysis. (C) Overall survival of the B16‐OVA model after treatment of PBS, DEV, DEV@IL‐12, DEV@aCTLA‐4, or DEV@IL‐12‐aCTLA‐4. Mean ± SD, n = 5. Log‐rank (Mantel‐Cox) test. (D) Administration routes for the LLC‐OVA model. (E) The subcutaneous tumour volume of LLC‐OVA after treatment of PBS, DEV, DEV@IL‐12, DEV@aCTLA‐4, or DEV@IL‐12‐aCTLA‐4. Mean ± SD, n = 5, one‐way ANOVA analysis. (F) Overall survival of the LLC‐OVA model after treatment of PBS, DEV, DEV@IL‐12, DEV@aCTLA‐4, or DEV@IL‐12‐aCTLA‐4. Mean ± SD, n = 5. Log‐rank (Mantel‐Cox) test. (G) Changes in body weight of immunized mice. Mean ± SD, n = 5. (H) Detection of IFN‐γ in mice serum 3 days after the last treatment. Mean ± SD, n = 5, one‐way ANOVA analysis. (I) Analysis of RBC and haemoglobin in mice 3 days after the last treatment. Mean ± SD, n = 5, one‐way ANOVA analysis. (J) The haematoxylin and eosin (H&E) staining of major organs separated from the treated C57BL/6 mice. Scale bars, 100 µm.

Considering the potential systemic toxicity and immune‐related adverse events (irAEs) of IL‐12 and aCTLA‐4, we systematically assessed the safety of DEV@IL‐12‐aCTLA‐4 in vivo. Continuous monitoring of mouse body weight showed no significant decrease during the treatment process (Figure 6G), which preliminarily reflects the safety of DEV@IL‐12‐aCTLA‐4. Regarding the toxicity and irAEs of IL‐12, systemic IFN‐γ is the main contributor (Ryffel 1997; Mansurov et al. 2020). Therefore, after the last treatment, we tested the serum IFN‐γ in circulation in tumour‐bearing mice. Results showed that the systemic IFN‐γ level of the DEV@IL‐12‐aCTLA‐4 group had no significant changes compared to the PBS and DEV groups (Figure 6H). For aCTLA‐4, inhibition of the hematopoietic system is one of the main adverse reactions (Zhu et al. 2023). Thus, we assessed the blood safety of treated mice. The levels of red blood cells (RBC) and haemoglobin (HGB) in the DEV@IL‐12‐aCTLA‐4 group were in the normal ranges (Figure 6I), suggesting no risk of anaemia associated with DEV@IL‐12‐aCTLA‐4 treatment. Inflammation and injury of major organs are also common adverse reactions of IL‐12 and aCTLA‐4. Blood biochemical indicators (TBil, AST, ALT, Cr and CK) and inflammatory indicators CRP and cytokine IL‐10 did not significantly rise in the DEV@IL‐12‐aCTLA‐4 group (Figure S13). The haematoxylin and eosin (H&E) staining of major organs also showed that there was no inflammation or damage observed in the heart, liver, lung, or kidney when using DEV@IL‐12‐aCTLA‐4 immunization (Figure 6J). To assess the late‐onset toxicities or weight changes in the animals, we used normal C57BL/6 mice for similar treatment in Figure 6A and conducted long‐term safety monitoring. After a long period of treatment (41 days), safety testing was conducted, including mouse body weight (Figure S14A), blood routine (Figure S14B), organ HE staining (Figure S14C), blood biochemical indicators (TBil, AST, ALT, Cr; Figures S14D–G), and inflammatory indicators (CRP; Figure S14H). All results were within the normal range, collectively reflecting the long‐term safety of the DEV@IL‐12‐aCTLA‐4 treatment.

3.7. Inducing Th1 Anti‐Tumour Immunity and Reversing Exhausted CD8⁺ T Cells in TME by DEV@IL‐12‐aCTLA‐4

Successful immunotherapy for cancer depends on functional immune cell infiltration to the TME and improvement of the immune suppression microenvironment. To further confirm the tumour‐suppressive effects were attributed to the anti‐tumour immunity induced by DEV@IL‐12‐aCTLA‐4, we tested the infiltration of immune cells in TME. As the IL‐12 was known to induce a Th1 immune response, we first assessed the Th1 immunity in TME. The DEV@IL‐12‐aCTLA‐4 group exhibited the highest proportion of Th1 cells (IFN‐γ⁺CD4⁺ T cells), while both the DEV@IL‐12 group and DEV@IL‐12‐aCTLA‐4 group demonstrated greater infiltration of IFN‐γ⁺CD4⁺ T cells compared to the DEV group (Figure 7A). These results suggest that the conjugation of IL‐12 facilitated the activation of Th1 immunity at tumour sites by the DEVs. The upregulation of IFN‐γ⁺ CD4⁺ T cells may correspond to the downregulation of regulatory T cells (Treg). As expected, the DEV@IL‐12 group and DEV@IL‐12‐aCTLA‐4 group had a lower proportion of Treg (Foxp3⁺ CD4⁺ T cells) than the DEV group and PBS group (Figure 7B). Considering that ICI can reverse exhausted CD8⁺ T cells, we further examined the functional status of CD8⁺ T cells in TME. In contrast to the DEV group, the DEV@aCTLA‐4 group and DEV@IL‐12‐aCTLA‐4 group showed lower PD‐1⁺ CD8⁺ T cell and TIM‐3⁺ CD8⁺ T cell infiltration (Figures 7C–D), suggesting that the conjugation of aCTLA‐4 endowed DEVs with the capability to reverse exhausted CD8⁺ T cells in TME. Correspondingly, the DEV@IL‐12‐aCTLA‐4 group exhibited a higher frequency of functional CD8⁺ T cells, such as IFN‐γ⁺CD8⁺ T cells (Figure 7E) and granzyme B⁺ CD8⁺ T cells (Figure 7F), at the tumour site compared to both the PBS group and DEV group. In conclusion, DEV@IL‐12‐aCTLA‐4 with the functions of IL‐12 and aCTLA‐4 can induce a Th1 anti‐tumour immune response and reverse exhausted CD8⁺ T cells in TME, effectively improving the reactivity and killing capability of tumour‐infiltrating T cells.

FIGURE 7.

DEV@IL‐12‐aCTLA‐4 induced Th1 anti‐tumour immunity and reversed exhausted CD8⁺ T cells in TME. (A)–(B) Analysis of IFN‐γ⁺ CD4⁺ T cells and foxp3⁺ CD4⁺ T cells in tumour site by flow cytometry after treatment of PBS, DEV, DEV@IL‐12, DEV@aCTLA‐4, or DEV@IL‐12‐aCTLA‐4. Mean ± SD, n = 5, one‐way ANOVA analysis. (C)–(F) Detection of PD‐1⁺ CD8⁺ T cells, TIM‐3⁺ CD8⁺ T cells, IFN‐γ⁺ CD8⁺ T cells, and granzyme B⁺ CD8⁺ T cells in tumour microenvironment by flow cytometry after treatment of PBS, DEV, DEV@IL‐12, DEV@aCTLA‐4, or DEV@IL‐12‐aCTLA‐4. Mean ± SD, n = 5, one‐way ANOVA analysis.

4. Discussion

As a novel platform for cancer vaccines, DEVs with antigen‐MHC complexes and costimulatory molecules have been shown to activate anti‐tumour T cells through multiple pathways, including DC‐dependent and DC‐independent pathways. Although the research on DEVs as cancer vaccines has entered the stage of clinical trial, limited clinical effects hinder the further development of DEVs. Further research on the basic theory of DEVs is expected to break through the existing bottlenecks. Previous studies have explored the immune stimulation ability of DEVs in the absence of other T cell regulatory signals, emphasizing the DC‐dependent pathway as crucial for DEVs in eliciting adaptive immunity (Nikfarjam et al. 2020; Matsumoto et al. 2020). However, these investigations may overlook the potential impact of concurrent regulatory signals, such as IL‐12 cytokines, which are typically secreted by parent DCs and may coexist with DEVs under physiological conditions. The presence of other immune regulatory signals could subtly augment the DC‐independent pathway, potentially enhancing DEVs' direct activation of T cells beyond previous understanding. Therefore, combining appropriate immune regulatory signals with DEVs holds promise for further boosting the DC‐independent pathway and ultimately enhancing T cell activation. Based on this finding, we further attempted to conjugate IL‐12 and aCTLA‐4 onto the surface of DEVs and constructed an optimized vaccine platform (DEV@IL‐12‐aCTLA‐4). By providing more sufficient immune regulatory signals, DEV@IL‐12‐aCTLA‐4 can directly promote T cell activation and produce stronger anti‐tumour effects than DEVs. Our research may expand the basic theory of DEVs and provide a universal strategy to improve the therapeutic effect of DEVs.

The DEV@IL‐12‐aCTLA‐4 platform is not only a unilateral improvement of DEVs but also beneficial for the accurate utilization of IL‐12 and aCTLA‐4. IL‐12 and aCTLA‐4 exhibit superior efficacy in tumour immunotherapy. However, their clinical applications are hindered by potential systemic toxicity and irAEs (Ramos‐Casals et al. 2020). Addressing the challenge of precise delivery to mitigate irAEs is crucial for the clinical translatability of these agents. The LN homing capability of DEVs enables DEV@IL‐12‐aCTLA‐4 to specifically deliver IL‐12 and aCTLA‐4 to the lymphatic system. Following subcutaneous injection, DEV@IL‐12‐aCTLA‐4 is primarily located in inguinal LNs and the liver, while sparing other major organs, thereby minimizing the risk of potential systemic toxicity and irAEs. Our results showed that DEV@IL‐12‐aCTLA‐4 immunization did not cause weight changes, abnormal blood biochemical indicators, or histological inflammation in major organs. These data indicated that DEV@IL‐12‐aCTLA‐4 with good safety had the potential to facilitate further clinical translation of IL‐12 and aCTLA‐4 in humans.

Successful immunotherapy depends on the activation of specific anti‐tumour T cells and the improvement of the immunosuppressive microenvironment (Tie et al. 2022). After vaccination, DEV@IL‐12‐aCTLA‐4 accumulated in LNs, offering sufficient T cell stimulation signals and activating robust antigen‐specific immune responses. Importantly, the Th1 response was induced by the IL‐12 in DEV@IL‐12‐aCTLA‐4, which contributed to the production of IFN‐γ‐mediated anti‐tumour immunity. Following the establishment of anti‐tumour immunity, activated immune cells still need to overcome the immunosuppressive microenvironment to produce positive effects. Immunosuppressive factors in the TME assist tumour cells in immune escape and compromise the vaccine efficacy, with the immune checkpoint playing a pivotal role among these factors. Therefore, the combination of cancer vaccines with immune checkpoint blockade is considered a powerful therapy strategy (Ott et al. 2017). Cancer vaccines promote anti‐tumour immunity priming, while ICIs boost and maintain the effects of vaccines, leading to persistent immune responses (Ott et al. 2020; Zhao et al. 2019). DEV@IL‐12‐aCTLA‐4 implements this strategy by the conjugation of ICIs on the DEV surface. DEV@IL‐12‐aCTLA‐4 treatment reversed exhausted CD8⁺ T cells and induced more functional CD8⁺ T cells in TME, which indicated that DEV@IL‐12‐aCTLA‐4 had greatly improved the tumour immunosuppressive microenvironment, making the induced anti‐tumour immune cells successfully infiltrate into TME and eradicate cancer cells. However, future studies with larger sample sizes are essential to validate our findings and increase confidence in their applicability to broader contexts.

The difficulty in mass production and complex production processes are the major obstacles to EV application (Grangier et al. 2021; Cheng and Hill 2022). Our research also inevitably faces this issue. Moreover, conjugating IL‐12 and aCTLA‐4 to DEVs will increase the complexity and difficulty of EV‐based vaccine preparation to some extent. Future breakthroughs in the development of more simpler and more efficient methods, such as genetic engineering and charge coupling for EV conjugation, may help solve this problem (Tan et al. 2023). Another limitation of our study is the relatively low centrifugation speed used for DEV isolation, which predominantly yields microvesicles with a relatively low proportion of small EVs. Consequently, our DEVs may not fully represent the entire EV population. Given the importance of small EVs in biological activities, future studies should compare DEVs obtained at higher centrifugation speeds and use other EV isolation methods, such as size exclusion chromatography, to enhance the generalizability of our conclusions (Wang et al. 2022). In addition, we used DSPE‐PEG‐NHS to conjugate IL‐12 and aCTLA‐4 onto the DEVs. DSPE‐PEG polymers are biocompatible and amphiphilic polymers, which are widely used to prepare lipid nanoparticles, and have been applied to the delivery optimization of some clinical drugs (Milano et al. 2022; Wang et al. 2012). Although DSPE‐PEG polymers are feasible, there has been no systematic comparison with other conjugation methods, such as genetic engineering parent cells and charge coupling, to determine the most optimal conjugation method. Drawing on the experience of E M Jaffee et al. in constructing GAVX vaccines capable of secreting GM‐CSF (Jaffee et al. 2001; Heumann et al. 2023), a more efficient approach may involve directly constructing engineered DCs capable of secreting DEVs with additional T cell activation signals. Furthermore, given the complexity of T‐cell activation signals, further investigation is needed to determine whether alternative signal combinations yield superior effects compared to IL‐12 and aCTLA‐4.

In summary, based on the theoretical discovery that augmenting immune regulatory signals can enhance the ability of DEVs to activate T cells directly, we have constructed DEV@IL‐12‐aCTLA‐4 as an immunotherapy platform that integrates tumour vaccine, cytokine, and immune checkpoint blockade. DEV@IL‐12‐aCTLA‐4 ensures satisfactory biosafety and avoids the irAEs of IL‐12 and aCTLA‐4, while effectively eliciting a robust anti‐tumour immune response and inhibiting tumour growth. Our study expands the theoretical foundation of DEVs and paints an intriguing picture of DEVs combined with IL‐12 and aCTLA‐4 to optimize cancer combination immunotherapy by inducing Th1 immunity and reversing exhausted CD8⁺ T cells. Meanwhile, this strategy provides a fruitful approach to promote the clinical translatability of IL‐12 and aCTLA‐4. Applying this strategy to clinical practice in the future will shed light on the translation of DEVs, IL‐12, and aCTLA‐4 from the bench to the bedside for cancer combination immunotherapy.

Author Contributions

Yang Jin, Mengfei Guo, Jiangbin Chen, Qi Tan, and Zimo Yang conceived the concept. Yang Jin, Mengfei Guo, Jiangbin Chen, Qi Tan, and Zimo Yang designed experiments and wrote the manuscript. Jiangbin Chen, Qi Tan, and Zimo Yang performed most of the experiments and analysed data. Wenjuan Chen, E. Zhou, Minglei Li, and Jingjing Deng were involved in the tumour intervention experiments of the in vivo study. Yali Wu, Jiatong Liu, and Juanjuan Xu were involved in the discussion of the experiment design and reviewed the manuscript. Yang Jin, Mengfei Guo, and Juanjuan Xu supervised and/or provided funding to this project and reviewed and edited the manuscript. All authors have read and approved the manuscript.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Supporting Information

Chen, J. , Tan, Q. , Yang, Z. , Chen, W. , Zhou, E. , Li, M. , Deng, J. , Wu, Y. , Liu, J. , Xu, J. , Guo, M. , & Jin, Y. (2025). Dendritic Cell Derived‐Extracellular Vesicles Engineered to Express Interleukin‐12 and Anti‐CTLA‐4 on Their Surface for Combinational Cancer Immunotherapy. Journal of Extracellular Vesicles, 14, e70068. 10.1002/jev2.70068

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request. All data needed to evaluate the conclusions in the paper are present in the paper and/or the Supplementary Materials.