Exosome-based vaccines in cancer immunotherapy: antitumor mechanisms, engineering strategies, and challenges.

Introduction
Vaccines are indispensable in modern medicine for disease prevention and treatment. Since the late 18th century, vaccine technology has evolved from live-attenuated and inactivated vaccines to subunit vaccines and, more recently, to novel platforms such as nucleic acid and viral vector vaccines (Kayser and Ramzan 2021; Plotkin 2014). However, traditional platforms face limitations in inducing durable immunity and overcoming immune evasion against complex diseases like cancer and highly mutable viral infections (Bouazzaoui et al. 2021; Cortese et al. 2025). These include safety concerns such as reversion to virulence in attenuated vaccines, insufficient immunogenicity of inactivated vaccines (Gebre et al. 2021; Guimarães et al. 2015; Vetter et al. 2018). instability and short half-life of mRNA vaccines (Vishweshwaraiah and Dokholyan 2022), as well as complex manufacturing processes and high costs (Zhang et al. 2023a). Therefore, new vaccine delivery strategies are urgently needed.
Exosomes, nanoscale extracellular vesicles secreted by cells, have recently emerged as a promising alternative platform for vaccine delivery. First identified in sheep reticulocytes in 1983 and named in 1987 (Johnstone et al. 1987; Pan and Johnstone 1983; Schubert 2020), exosomes (30–150 nm in diameter) are released via the fusion of multivesicular bodies with the plasma membrane (Han et al. 2022). They are produced by various cell types—including immune, mesenchymal stem, and tumor cells (Yoo et al. 2022; Zheng et al. 2018)—and are present in biofluids such as blood and saliva. By transporting proteins, lipids, and nucleic acids, exosomes mediate intercellular communication and modulate immune responses and physiological homeostasis (Isaac et al. 2021; Kalluri and LeBleu 2020; Liu et al. 2023; Shao et al. 2017; Thakur et al. 2021). Their molecular composition reflects the physiological and pathological state of parent cells (Wang et al. 2022a).
Owing to their biocompatibility, low immunogenicity, and tissue-penetrating ability, exosomes are attractive delivery vehicles (Cacciottolo et al. 2023; Kučuk et al. 2021; Sun et al. 2021). They can deliver diverse bioactive cargo to target cells and exhibit intrinsic immunomodulatory, anti-inflammatory, and pro-angiogenic activities, highlighting their potential as multifunctional vaccine platforms (Sadeghi et al. 2023). When engineered to carry antigens or immunostimulatory molecules, exosomes can mimic natural infection and elicit robust immune responses (Rezaie et al. 2022). Since the 1998 report that dendritic cell-derived exosomes inhibit tumor growth in mice (Zitvogel et al. 1998), exosome-based vaccine research has expanded from oncology to infectious diseases. For example, nebulized COVID-19 vaccines using lung cell exosomes induced strong mucosal and systemic immunity in animal models (Tan et al. 2024; Wang et al. 2022b).
Despite this promise, research remains fragmented. Existing reviews often focus on mechanisms and technologies without critically linking engineering strategies to the biological constraints they address. To bridge this gap, this review first outlines the inherent limitations of natural exosomes as vaccine platforms. It then evaluates how engineering approaches can overcome these bottlenecks and extends this analysis to application scenarios and translational challenges. This framework aims to guide future research from biological insight to rational design and, ultimately, clinical translation.

Mechanisms of Exosome-based vaccines in antitumor therapy
The antitumor immune mechanism of exosome-based vaccines is rooted in their dual capacity for efficient antigen delivery and precise dendritic cell (DC) activation. Exosomes display on their surface a repertoire of molecules that mediate cellular targeting, such as milk fat globule-EGF factor 8 (MFG-E8), CD11a, intercellular adhesion molecule-1 (ICAM-1, CD54), phosphatidylserine (PS), and the tetraspanins CD9 and CD81. These surface molecules interact with specific receptors on DCs, including integrin αvβ3 (CD51/CD61), CD11a, and CD54, thereby facilitating the complement-independent targeting and uptake of exosomes by DCs (Morelli et al. 2004). DCs are characterized by high surface expression of the mannose receptor (CD206). α-D-mannose-modified exosomes can therefore serve as specific ligands for this receptor, mediating targeted exosome adhesion to DCs and their subsequent endocytic internalization (Choi et al. 2019). Following uptake, exosomes are primarily processed via clathrin- or caveolin-dependent endocytic pathways and subsequently enter the endosomal-lysosomal compartment for antigen processing. Processed antigen peptides are then loaded onto major histocompatibility complex (MHC) molecules and transported to the DC surface for presentation. This sequence of events is fundamental for the subsequent activation of CD4⁺ helper T cells and CD8⁺ cytotoxic T lymphocytes (Zemanek et al. 2025).
Moreover, exosomes harboring MHC class I/II molecules, co-stimulatory molecules (e.g., CD40, CD80, CD86), and heat shock proteins (including HSP70 and HSP90) can directly promote the maturation of antigen-presenting cells (APCs) like DCs. This is accompanied by the upregulation of surface MHC and co-stimulatory molecules, ultimately enhancing antigen presentation efficiency (Karami Fath et al. 2022a). This process functions by targeting and downregulating the expression of SOCS1 (suppressor of cytokine signaling 1), a key negative regulator within DCs, thereby relieving the inhibition on the JAK-STAT signaling pathway. The consequent activation of the JAK-STAT pathway promotes the expression of maturation-associated transcription factors, including IRF4 and NF-κB, which collectively initiate the DC maturation program (Qiu et al. 2024). Mature DCs present processed antigens via MHC class II molecules to CD4⁺ helper T cells (particularly the Th1 subset), stimulating them to secrete crucial cytokines like interferon-γ (IFN-γ) and interleukin-2 (IL-2). This polarizes the immune response towards an effector phenotype, thereby augmenting antitumor immunity (Huda and Nurunnabi 2022).
To augment the innate immunostimulatory properties of exosomes, engineered modifications are applied to broaden and potentiate their functional repertoire. For instance, genetic engineering of tumor-derived exosomes enables the co-loading of tumor antigen-encoding mRNA and immunostimulatory nucleic acid adjuvants (e.g., the TLR ligand CpG DNA), thereby enhancing their targeting efficiency to DCs (Zhang et al. 2023b). This engineered approach, compared to that using natural exosomes, more effectively activates CD8⁺ T cells via DC-mediated cross-presentation, leading to the generation of antigen-specific cytotoxic T lymphocytes (CTLs). The resulting CTLs subsequently recognize tumor-associated antigen–MHC class I (TAA–MHC I) complexes on tumor cells and release cytotoxic molecules like perforin and granzymes, leading to direct tumor cell lysis (Qiu et al. 2024).
Exosome-based vaccines elicit antitumor immunity not only by activating adaptive T-cell responses but also through effective collaboration with innate immune killer cells. For example, exosomes derived from natural killer (NK) cells can deliver cytotoxic proteins like Fas ligand (FasL); these ligands bind to the Fas receptor on target cells, triggering apoptosis. Conversely, some tumor cell-derived exosome vaccines express ligands for NKG2D (such as MICA/B). These ligands engage the NKG2D receptor on NK cells, thereby directly activating their killing machinery in an MHC-unrestricted fashion (Luo et al. 2023a). Research has shown that embryonic stem cell-derived exosome vaccines engineered to carry granulocyte-macrophage colony-stimulating factor (GM-CSF) can markedly potentiate the tumor-specific activity of tumor-infiltrating lymphocytes and promote Th1 cytokine responses, primarily mediated by GM-CSF. Specifically, IFN-γ produced by Th1 cells further boosts NK cell cytotoxicity, thereby synergistically integrating adaptive and innate immunity into a coordinated antitumor network. This strategy has been demonstrated in murine models to effectively delay or prevent tumor progression (Xia et al. 2021).
Beyond their core function of initiating adaptive immunity in lymphoid organs, specially engineered “smart” exosome vaccines also possess the potential to directly intervene in the tumor immune microenvironment. For instance, by loading specific cytokines (e.g., IFN-γ), siRNA (e.g., targeting TGF-β), or utilizing M1 macrophage-derived exosomes as carriers, engineered exosomes hold promise to locally reprogram immunosuppressive macrophages or suppress the function of regulatory T cells at the tumor site. This approach synergizes with systemic immune activation to more thoroughly dismantle the tumor’s immune defenses (Meng et al. 2025).
Immune cells activated by exosome vaccination (e.g., T cells) persistently secrete IFN-γ. Functioning as a pivotal anti-angiogenic factor, IFN-γ inhibits the expression of vascular endothelial growth factor (VEGF)—a central driver of tumor angiogenesis by tumor cells and vascular endothelial cells. Consequently, the suppression of VEGF expression impedes tumor neovascularization, disrupts nutrient supply, and thereby significantly curbs tumor growth (Vlaeminck-Guillem 2022).
Beyond anti-angiogenic effects, dendritic cell-derived exosomes (DCexos) are enriched both on their surface and intracellularly with members of the TNF superfamily, such as tumor necrosis factor (TNF), Fas ligand (FasL), and TNF-related apoptosis-inducing ligand (TRAIL). Upon contact with tumor cells, the FasL and TRAIL presented by DCexos engage their cognate receptors—Fas and death receptors 4/5 (DR4/DR5), respectively—on the tumor cell surface. This interaction directly initiates the caspase-dependent apoptotic cascade within tumor cells, leading to programmed cell death independent of other immune cells (Dai et al. 2024).
In conclusion, exosome vaccines exert their antitumor effects through the synergistic action of the aforementioned multi-level mechanisms. It is crucial to emphasize that these functions are not universally shared by all exosomes but are profoundly dependent on the origin and state of their parent cells. To clearly illustrate this key concept, Fig. 1 compares the specific antitumor mechanisms predominantly mediated by exosomes derived from common cellular sources.

This figure comprehensively compares the specific anti-tumor mechanisms exhibited by exosomes derived from four key cell types when used as vaccine platforms or therapeutic carriers. As shown in A, dendritic cell-derived exosomes (Dex) possess a core advantage in their innate antigen-presenting capacity. Their surface enrichment of MHC-I/II molecules enables direct presentation of tumor antigens to CD8⁺ or CD4⁺ T cells, thereby efficiently priming cytotoxic T lymphocyte (CTL) responses and inducing long-term immune memory. Natural killer cell-derived exosomes (NK-Exo) (B) represent a rapid killing pathway independent of antigen priming. By delivering effector molecules such as perforin and granzyme, they directly induce tumor cell apoptosis or lysis. Macrophage-derived exosomes (M-Exo) (C) not only reprogram immunosuppressive M2 tumor-associated macrophages (TAMs) into anti-tumor M1 phenotypes via pathways like NF-κB, but also enhance cross-presentation by dendritic cells as antigen carriers. Engineered tumor-derived exosomes (TDEs) (D) leverage their inherent homotopic targeting to precisely deliver therapeutic payloads like chemotherapeutic drugs or siRNA while providing a complete tumor antigen spectrum, synergistically activating specific immune responses. (Created with Biogpd.com)
While the aforementioned mechanisms establish the biological basis for exosome-based vaccines, unmodified natural exosomes are far from an ideal vaccine platform. Their clinical translation faces three core limitations: insufficient targeting specificity, suboptimal immunogenicity, and poor controllability over antigen/adjuvant loading. These inherent drawbacks define the starting point for engineering. Consequently, the engineering strategies discussed in the subsequent section aim precisely at overcoming these bottlenecks, transforming the natural carrier into a potent and controllable therapeutic modality.

Engineering strategies and comparative analysis of Exosome-based vaccines
To guide the rational selection and translation of engineering strategies, this paper establishes a “goal-driven design framework” for exosome vaccines (Fig. 2), aiming to systematically map their complete pathway from engineering design to clinical application. The selection of engineering strategies should first serve to overcome specific biological limitations, then anchor to clear clinical application scenarios, and ultimately withstand the tests of challenges such as large-scale production and clinical heterogeneity. The following sections will use this framework to conduct in-depth analysis and comparison of each strategy, exploring their logical applicability in specific scenarios.

This figure systematically outlines the pathway from rational design to clinical translation for exosome-based vaccines. The starting point involves identifying the core biological limitations of natural exosomes (1). Targeted engineering strategies (2) are employed for functional reprogramming: physical drug loading enables rapid antigen delivery and enhanced T-cell activation; genetic engineering achieves sustained immune stimulation and precise DC targeting; membrane modification facilitates controlled antigen loading and specific recognition. Functionally enhanced engineered exosomes are then directed toward three core application scenarios (3): preventive vaccines, personalized neoantigen vaccines, and combination immunotherapies. Ultimately, all pathways converge on common clinical translation challenges (4) requiring systematic solutions, including scalable production and GMP compliance, function-oriented quality control standards, and clinical efficacy heterogeneity.

Overview of engineering strategies
Building upon these mechanisms, exosomes demonstrate considerable potential as a vaccine platform. However, their natural form faces limitations in targeting precision, immunogenicity, and controllable antigen loading. To develop stable, efficient, and programmable therapeutics, various engineering strategies have been employed to optimize these core functions. These include direct physical or chemical cargo loading, genetic/cellular engineering to modulate biosynthesis and composition, and membrane modification or biomimetic fusion to introduce novel properties.
Moving beyond descriptive listing, this analysis critically evaluates these approaches against the biological limitations outlined in Sect. 2. Table 1 details each method’s advantages while explicitly mapping its disadvantages and its capacity to address core bottlenecks in targeting, immunogenicity, and loading control. This perspective is synthesized in Fig. 3, which presents a rational selection framework. This framework guides strategy choice—for instance, physical loading for rapid encapsulation versus genetic engineering for stable functional display—based on the prioritized biological constraint, the nature of the therapeutic cargo, and the desired immunological outcome.

Physical Loading Strategies address poor loading controllability by applying external forces to transiently disrupt the exosomal membrane for cargo encapsulation. Sonication uses vibrations to permeabilize membranes, suitable for loading macromolecular antigens. For example, loading the CRISPR-Cas9 system into ovarian cancer cell-derived exosomes via sonication induced tumor cell apoptosis and enhanced cisplatin sensitivity (Liu et al. 2022). Electroporation creates transient pores via an electric field for efficient nucleic acid encapsulation. Optimized electroporation (150 V) achieved a 40% loading rate for SIRT6-targeting siRNA into prostate cancer cell-derived exosomes, leading to a 55% reduction in tumor volume in vivo (Lyu et al. 2024). Freeze-thaw cycling facilitates passive incorporation of water-soluble antigens like HIV-1 peptides, preserving exosome integrity while successfully eliciting antigen-specific CTL responses in mice (Habib et al. 2023). Physical processing can also enhance intrinsic immunogenicity; exosomes from gamma-irradiated melanoma cells show elevated HMGB1 levels, boosting dendritic cell and T-cell activation (Kim et al. 2020).
Genetic Engineering tackles targeting and immunogenicity limitations at the molecular source by overexpressing desired molecules or downregulating inhibitory ones. For instance, engineering triple-negative breast cancer cell-derived exosomes to overexpress α-lactalbumin improved targeting, and subsequent loading of immunogenic cell death inducers enhanced antitumor immunity (Huang et al. 2022a). Conversely, lentivirus-mediated PD-L1 silencing in leukemia cell-derived exosomes (LEX PD-L1si) enhanced DC maturation and CTL responses (Huang et al. 2022b; Yang et al. 2024). A multi-cargo system delivering IL4RPep-1, NF-κB p50 siRNA, and miR-511-3p enabled precise tumor-associated macrophage targeting and reprogramming (Lu et al. 2024).
Cellular Engineering modulates parental cells to shape exosome properties. Genetically modifying adipose-derived stem cells to secrete exosomes loaded with miR-7 suppressed cervical cancer growth by downregulating XIAP (Zhou et al. 2022). Transducing leukemia cells to express CD80/CD86 produced exosomes (LEX-CD8086) that promoted T-cell proliferation and directly stimulated antigen-specific CTLs (Li et al. 2022). Platforms like the DEXP&A2&N vaccine, constructed by coupling exosomes with tumor peptides and adjuvants, enhance dendritic cell activation and neoantigen cross-presentation (Zuo et al. 2022).
Advanced Hybrid Systems further augment functionality. Macrophage–tumor cell hybrid exosomes reduced tumor volume by 65% in murine models and induced protective memory T-cell responses (Wang et al. 2021). Engineered exosome-like nanovesicles expressing fibroblast activation protein promoted tumor ferroptosis via CTL-derived IFN-γ and depletion of cancer-associated fibroblasts, significantly enhancing antitumor efficacy (Hu et al. 2021).

This figure systematically elucidates three core engineering strategies for optimizing exosome vaccine functionality and establishes a rational selection framework based on key parameters. For studies aiming to rapidly deliver nucleic acid antigens to initiate an initial immune response, physical engineering strategies (A)—such as electroporation (a), ultrasound (b), and freeze-thaw cycles (c) are ideal choices. These strategies transiently modulate membrane permeability to efficiently encapsulate siRNA or mRNA, targeting dendritic cells (DCs) to rapidly induce antigen-specific CD8⁺ T cell cytotoxic responses. For achieving sustained, stable, long-term immune activation and memory, genetic modification strategies (B) are more suitable. By engineering parental cells to reprogram the molecular composition of exosomes at the source, these approaches target tumor-infiltrating lymphocytes (TILs) within the tumor microenvironment, enhancing the persistence and specificity of immune responses. For the complex challenge of synergistically stimulating multidimensional immunity and reversing immunosuppressive microenvironments, the cell fusion strategy (C) demonstrates unique value. By generating chimeric exosomes through hybridizing cells from different lineages, it endogenously integrates multiparental functions (e.g., delivering MHC-antigen complexes) and targets lymph node-resident T cells, thereby synergistically activating innate and adaptive immunity. This framework aims to elucidate that the selection of engineering strategies should begin with a clear definition of therapeutic objectives and be guided by systematic considerations of payload properties, target cell types, and anticipated immune effects.
These engineering strategies confer distinct platform advantages to exosome-based vaccines. Their inherent lipid bilayer ensures excellent biocompatibility and critical stability during storage and delivery. For example, dendritic cell-derived exosomes retain > 80% bioactivity after 7 days at 4 °C, surpassing the stability of conventional LNP-mRNA vaccines that typically require ultra-low temperature storage (Babaei et al. 2024; Dyball and Smales 2022; Mustajab et al. 2022). Functioning as native immune units, exosomes naturally display MHC–peptide complexes and co-stimulatory molecules on their surface, enabling direct antigen-specific T cell activation and potent cellular immunity without exogenous adjuvants (Thakur et al. 2022).
Compared to traditional platforms, exosome-based vaccines offer stable and direct antigen delivery, which may lead to stronger and more durable immune responses, yet their effectiveness is contingent on vaccine formulation and antigen properties (Araujo-Abad et al. 2025). For instance, γδ-T cell-derived exosome vaccines show direct tumoricidal activity and superior anticancer efficacy compared to dendritic cell-derived exosomes (Wang et al. 2023a). heir ability to co-deliver multiple antigens offers a novel strategy against rapidly mutating pathogens. Thus, multivalent exosome vaccines promise broad cross-protection across strains, overcoming the narrow spectrum of single-antigen vaccines (Montaner-Tarbes et al. 2021). Moreover, exosomes’ structural stability and facile surface modification make them more suitable for storage and standardized production than intact dendritic cell preparations (Schioppa et al. 2024).
In tumor immunotherapy, the nanoscale size and engineerable membrane of exosomes promote efficient penetration of dense tumor stroma, delivering antigens to deep lesions often inaccessible to conventional nanocarriers (Xu and Xiong 2024). Crucially, as acellular carriers, exosomes maintain functional integrity even within immunosuppressive tumor microenvironments rich in inhibitors like iNOS and TNF-α. They can thus persistently deliver activating signals to T cells, overcoming a key limitation of live-cell-based vaccines (Luo et al. 2023b).
No single strategy fits all applications. Selection depends on therapeutic goals: physical methods allow rapid antigen loading, whereas genetic engineering enables stable molecular expression. The future lies in “multi-modal engineering”—synergistically combining strategies to create multifunctional platforms capable of simultaneous tumor targeting, immune activation, and microenvironment modulation.

Comparative analysis and strategic selection
This section establishes a multidimensional framework to guide the selection and optimization of engineering strategies based on specific therapeutic goals, antigen characteristics, and translational stages. The analysis focuses on three key dimensions. First, a fundamental trade-off exists between loading efficiency and membrane integrity. Widely used exogenous physical methods, such as electroporation and sonication, achieve high loading efficiencies, particularly for nucleic acids (e.g., siRNA, miRNA, plasmids). For example, sonication outperforms co-incubation in paclitaxel loading, and electroporation-loaded siRNA targeting KRASG12D maintained potent activity in pancreatic cancer models (Ng et al. 2025). However, these methods can irreversibly disrupt exosomal membranes, leading to cargo leakage and impaired intrinsic functions (Choi et al. 2022). In contrast, genetically engineering parent cells enables endogenous cargo incorporation, better preserving vesicle structure and bioactivity. This approach, however, faces limitations in flexibility and efficiency due to constraints in cellular transfection and metabolic pathways (Lin et al. 2024). hus, optimized physical methods are suitable for rapidly loading membrane-insensitive, structurally stable molecules (e.g., certain siRNAs). Conversely, for complex functional proteins or when high vesicle integrity is critical, genetic engineering offers superior fidelity (Chen et al. 2021a).
Furthermore, different strategies prioritize distinct pathways to enhance targeting precision and modulate immunogenicity. Post-isolation membrane modification, such as click chemistry, offers notable modularity. This technique allows precise, flexible conjugation of diverse targeting ligands (e.g., RGE peptides, EGFR nanobodies) to harvested exosomes via efficient covalent reactions, typically without altering their core physicochemical properties, making it suitable for preliminary targeting screening (Salunkhe et al. 2020). However, chemical modifications can affect membrane fluidity, and batch-to-batch consistency presents scaling challenges. In contrast, genetically engineering parent cells enables stable expression of targeting ligands (e.g., fused to Lamp2b or CD63) or immunomodulatory molecules (e.g., MHC-II, cytokines, pathogen antigens). This approach produces functionally homogeneous exosome populations with integrated targeting and immunostimulatory capabilities, offering clear advantages for developing standardized therapeutics (Jafari et al. 2020). This implies that membrane modification techniques are more convenient when the primary goal is rapid validation of multiple targeting hypotheses; conversely, genetic engineering provides a more robust foundation for developing functionally integrated, process-stable therapeutic products.
Finally, strategy selection must balance technological complexity and scalability with regulatory feasibility. Physical or chemical methods are relatively straightforward to implement in the laboratory and allow rapid functional validation. However, they are highly sensitive to minor process variations (e.g., centrifugation speed, filter pore size, culture media batches), which can lead to significant inconsistencies in exosome yield, purity, size, and membrane integrity. These issues pose major challenges for scale-up, clinical-grade quality control, and regulatory approval (Ahn et al. 2022). n contrast, genetically engineered parental cells—developed into stable lines that secrete functionalized exosomes—combined with secondary modifications such as targeted conjugation, can address multiple challenges synergistically, including targeting specificity, cargo stability, and immunomodulation (Giovannelli et al. 2023). Although this approach requires more complex initial development, a stable cell bank provides a consistent, reproducible, and traceable production source. Thus, exploratory studies may favor simpler, flexible methods, but clinical translation should transition early to engineered, well-defined production processes that align with Good Manufacturing Practice (GMP) requirements for consistency and scalability (Chen et al. 2021b). In summary, future advances will likely involve precise secondary modifications of genetically engineered exosomes to systematically integrate targeting, loading, and immunomodulatory modules. This integrated strategy can synergistically enhance delivery efficiency, payload stability, and therapeutic activity, transforming exosome vaccines from promising carriers into standardized, designable therapeutics with definable efficacy and production profiles (Lai et al. 2022).

Applications of Exosome-based vaccines in cancer therapy
A critical test of the “biological constraints-engineering solutions” framework is its capacity to guide vaccine design against specific clinical challenges. This section analyzes how distinct therapeutic scenarios—personalized neoantigen targeting, immunoprophylaxis, and combination therapy—establish unique design priorities, thereby shaping the selection and integration of the engineering strategies previously outlined.

Personalized neoantigen vaccines
The core challenge for personalized neoantigen vaccines is the rapid delivery of patient-specific, weakly immunogenic neoantigens to effectively prime de novo T-cell responses against tumors with high mutational burden. This scenario prioritizes design principles such as flexible, high-efficiency antigen loading and potent dendritic cell (DC) activation/cross-presentation.
To meet the need for flexible loading, physical methods like electroporation are advantageous for encapsulating diverse neoantigen-encoding mRNAs. In the MC-38 model, engineered exosomes demonstrated superior tumor volume reduction and CD8⁺ T-cell infiltration compared to liposomes (Li et al. 2023). To achieve potent immune priming, further engineering can enhance the intrinsic immunogenicity of exosomes. For instance, exosomes derived from autologous tumors, which naturally home to lymph nodes, provide a dual-targeting platform that suppresses both primary and metastatic brain tumors without exogenous drug loading, underscoring their inherent optimization for central nervous system malignancies (Zou et al. 2025). Furthermore, precision can be achieved by targeting tumor-specific antigens such as CMV pp65 in glioblastoma; exosomes encapsulating such antigens leverage their biocompatibility to enable safe and effective neuro-oncological immunotherapy (Karami Fath et al. 2022b).

Preventive vaccines
Cancer preventive vaccines must induce broad and durable immune memory while ensuring an excellent safety profile for administration in healthy high-risk individuals. This scenario prioritizes design principles focused on cross-reactive antigen presentation, sustained immune education, and minimal carcinogenic risk.
Engineering strategies, therefore, concentrate on modifying the biological origin and cargo of exosomes to fulfill these principles. For example, engineered embryonic stem cell-derived exosomes delivering GM-CSF (ES-exo/GM-CSF) systemically remodel the immune microenvironment, inhibiting tumor growth and metastasis across multiple cancer models. Importantly, this platform demonstrates cross-reactivity to heterologous tumor antigens while avoiding the carcinogenic risks associated with intact cells, confirming its suitability for preventive design (Meng et al. 2023; Yaddanapudi et al. 2019). Likewise, dendritic cell-derived exosome vaccines conjugated with tumor-associated glycopeptides like MUC1 induce durable protective memory with good biocompatibility in preclinical studies (Zhu et al. 2022). These examples highlight how engineering prioritizes broad protection and safety, distinct from therapeutic vaccines focused on breaking established tolerance.
This engineered approach extends to creating sustained antigen exposure systems that mimic natural infection. For example, a DC + EXO vaccine using an antigen library from induced pluripotent stem cell (iPSC)-derived exosomes established long-term T cell memory and protective immunity against tumor rechallenge in melanoma models by providing continuous antigen signals (Wang et al. 2023b). This exemplifies a forward-looking design that optimizes immune education through engineered sources and delivery modalities.

Combination therapy
In advanced solid tumors, monotherapy efficacy is often limited by the immunosuppressive tumor microenvironment (TME). This context requires engineering exosome vaccines to act as synergistic partners with other therapies, focusing on TME remodeling and sequential immune modulation.
Accordingly, engineering aims to enhance specific combinatorial functions. First, exosomes can be designed to reverse local immunosuppression. For instance, Dex vaccines activated by TLR3 agonists like poly(I: C) recruit CD8⁺ T cells and NK cells to the tumor site, priming the TME for subsequent treatments (Damo et al. 2015). Beyond recruitment, exosomes can directly reprogram key immunosuppressive cells—such as M2-like tumor-associated macrophages and regulatory T cells—which drive immunotherapy resistance (Cheng et al. 2022). Engineered exosomes may deliver cytokines like IL-12 to polarize M2 to M1 macrophages, or carry TGF-β siRNA to neutralize immunosuppressive signals and inhibit Treg function (Santos and Almeida 2021). This local “immunosuppression reversal” combined with systemic immune activation forms a multipronged therapeutic paradigm.
Second, engineered exosomes can deliver synergistic payloads. When loaded with immunogenic cell death inducers, they not only present antigens but also convert tumor cell death into robust dendritic cell activation, thereby amplifying the antitumor immune cycle (Sheikhhossein et al. 2024). This evolution from a passive “carrier” to an active “immune regulatory hub” is pivotal for successful combination therapy.
The transition from a passive “carrier” to an active “immunomodulatory hub” is pivotal for successful combination therapies. Early clinical trials suggest antigen-loaded Dex vaccines are safe and exhibit a trend toward enhanced efficacy when combined with immune checkpoint inhibitors, providing preliminary clinical validation for this engineered synergy (Ma et al. 2024). However, deeper analysis reveals translational complexity. While Phase I trials confirmed the safety and immunogenicity of autologous Dex vaccines, both antigen-specific T-cell response rates (~ 6.7–33.3%) and objective response rates (e.g., 6.7%) remained limited, underscoring the inadequacy of monotherapy (Escudier et al. 2005; Morse et al. 2005). Subsequent Phase II studies showed that a polyantigen-loaded Dex vaccine, when combined with chemotherapy, enabled long-term disease stabilization in approximately 31.8% of patients (Ye et al. 2024). This indicates that even in combination regimens, efficacy is confined to a subset of patients, highlighting significant response heterogeneity and the need for precision in combinatorial strategies. The observed synergy with ICIs may arise because the vaccine acts as an “immune primer,” inducing specific T-cell clones in some patients. These clones then become critical cellular targets for subsequent ICI-mediated “de-suppression,” thereby amplifying the overall antitumor response.

Challenges facing Exosome-based vaccines
Despite their considerable promise in antitumor therapy, the translation of exosome-based vaccines from bench to bedside continues to confront several key challenges. First, their inherently small size and compositional complexity pose significant challenges for consistent quality control and standardized characterization. From a production standpoint, low yields and high preparation costs remain major barriers to scalability and industrial translation (Mukerjee et al. 2025; Tenchov et al. 2022). A fundamental issue is the substantial overlap in biophysical properties (e.g., buoyant density ranges of 1.10–1.21 g/mL) between exosomes and other extracellular vesicles, which renders complete purification exceptionally challenging (Lai et al. 2022). Addressing this will require the development of integrated purification strategies—for example, combining size-exclusion chromatography with affinity chromatography—or the optimization of scalable techniques such as tangential flow filtration. The goal is to achieve an optimal balance among purity, yield, and preserved bioactivity.
The intrinsic heterogeneity of exosomes represents a fundamental challenge to their development into standardized therapeutic agents. Substantial heterogeneity in size, molecular composition, and function across different exosome subpopulations leads directly to batch-to-batch variability and complicates the precise control of their cargo (Xu et al. 2024). Therefore, establishing a robust, standardized framework for characterization and quality control—encompassing physical properties (e.g., size, concentration), biochemical signatures (e.g., specific markers), and functional attributes (e.g., antigen-loading efficiency, immunostimulatory potency) is essential (Tan et al. 2024). To this end, advancing the integrated use of multi-parameter analytical technologies, such as high-resolution flow cytometry and single-particle analysis, will be key to identifying therapeutically active exosome subpopulations and establishing a foundation for process standardization.
Current research on tumor-derived exosomes (TEX) has notable limitations, including a narrow range of preclinical models and inadequate long-term observation, which constrain the development of antitumor vaccines. Most existing studies are predominantly conducted in a limited set of murine lung cancer models (e.g., 3LL, NSCLC) and assess short-term efficacy over relatively brief experimental timelines (typically 4–8 weeks). For instance, studies of TEX vaccines engineered with CD40L or overexpressing Rab27a have focused primarily on short-term endpoints such as tumor volume reduction and antigen-specific T cell responses. The lack of long-term follow-up data (e.g., > 6 months) impedes a comprehensive assessment of critical safety profiles, including risks of tumor recurrence and potential chronic organ toxicities. This gap significantly constrains the clinical translatability of these vaccine candidates (Wu et al. 2022).
In terms of efficacy, exosome-based vaccines frequently necessitate combination with other therapeutic modalities to achieve optimal outcomes. For example, in murine models, Dex vaccines enhance tumor-specific cytotoxic T lymphocyte (CTL) activation only when administered in combination with metronomic, low-dose oral cyclophosphamide. In human clinical trials, this combinatorial regimen still requires additional immunostimulatory agents to maximize efficacy, inevitably increasing treatment complexity and posing challenges for clinical implementation (Castillo-Peña and Molina-Pinelo 2023).
The clinical translation of exosome-based vaccines is significantly hindered by a paucity of robust safety and efficacy data. Their complex composition necessitates rigorous safety assessment; however, human pharmacokinetic and pharmacodynamic (PK/PD) profiles remain severely limited. Early-phase trials with small sample sizes are inherently underpowered to generate conclusive evidence (Shimon et al. 2022). A major contributing factor is that current evidence is predominantly derived from animal models, which cannot fully recapitulate the complex interplay of human tumor heterogeneity, interpatient variability, and comorbidities—all critical determinants of clinical safety and efficacy (Lv et al. 2022).
When incorporating engineered exosome vaccines into combination therapies such as with immune checkpoint inhibitors (ICIs), several translational barriers emerge beyond those seen in monotherapy. Beyond the well-documented risk of additive immunotoxicity and immune-related adverse events (irAEs) with ICI combinations (Tang et al. 2022), therapeutic heterogeneity presents a persistent challenge. For instance, exosome vaccines combined with chemotherapy achieve long-term disease stabilization in only about 31.8% of non-small cell lung cancer patients (Ye et al. 2024), mirroring the limited response rates of other immunotherapies—such as the sub‑30% efficacy of anti‑PD‑1 in hepatocellular carcinoma, attributable to variable tumor immune microenvironments like that mediated by high SRSF10 expression (Cai et al. 2024). Moreover, clinical development of such combinations is inherently complex, requiring systematic optimization of dosing, sequencing, and scheduling rather than simple drug stacking, with optimal regimens often tumor‑specific (Ott et al. 2017). Consequently, developing such combination regimens typically necessitates large-scale, multi-arm clinical trials to explore the vast parameter space, inevitably prolonging R&D cycles and escalating economic costs. For the emerging exosome vaccine platform, its combined development with ICIs will inevitably confront these formidable translational hurdles, urgently requiring innovative clinical trial designs to advance efficiently.

Conclusion
This review proposes a “biological constraint–engineering solution” framework that reframes exosome vaccine development as a targeted engineering challenge rather than merely an exploration of natural properties. Theoretically, it integrates fragmented research into a coherent design logic by evaluating each engineering strategy against its capacity to address specific biological shortcomings—such as poor targeting or weak immunogenicity. Practically, it guides prioritization: for solid tumor immunotherapy, genetic engineering may enable stable antigen presentation, while membrane modification can enhance targeting precision. This shifts the goal from producing generic exosome vaccines toward creating programmable, disease-specific therapeutics.
To realize this potential, future work must move beyond proof-of-concept studies and tackle the key industrialization and efficacy barriers highlighted here. First, scalable, standardized production processes must be established—potentially via synthetic biology to generate stable “engineered exosome-producing cells”—to overcome the “cell source and scale-up” bottleneck. Second, potency-based quality standards should replace purely physical characterization to ensure batch-to-batch consistency. Furthermore, as monotherapy efficacy in solid tumors may be limited by the immunosuppressive microenvironment, rational combination strategies with immune checkpoint inhibitors should be designed based on the mechanistic interplay between exosome-induced T-cell activation and immune checkpoint blockade, rather than empirical testing. Ultimately, the field must advance exosomes from heterogeneous biological mixtures toward a well-defined drug class in which origin, composition, function, and targeting can be precisely engineered and controlled.