Potential and development of cellular vesicle vaccines in cancer immunotherapy

Abstract

Cancer vaccines are promising as an effective means of stimulating the immune system to clear tumors as well as to establish immune surveillance. In this paper, we discuss the main platforms and current status of cancer vaccines and propose a new cancer vaccine platform, the cytosolic vesicle vaccine. This vaccine has a unique structure that can integrate antigen and adjuvant carriers to improve the delivery efficiency and immune activation ability, which brings new ideas for cancer vaccine design. Tumor exosomes carry antigens and MHC-peptide complexes, which can provide tumor antigens to antigen-processing cells and increase the chances of recognition of tumor antigens by immune cells. DEVs play a role in amplifying the immune response by acting as carriers for the dissemination of antigenic substances in dendritic cells. OMVs, with their natural adjuvant properties, are one of the advantages for the preparation of antitumor vaccines. This paper presents the advantages of these three bacteria/extracellular vesicles as cancer vaccines and discusses the potential applications of functionally modified extracellular vesicles as cancer vaccines after cellular engineering or genetic engineering, as well as current clinical trials of extracellular vesicle vaccines. In summary, extracellular vesicle vaccines are a promising direction for cancer vaccine research.

Introduction

The pivotal mechanism in controlling tumor development lies within the tumor-killing capacity of the immune system. The body identifies and presents tumor cell antigens via antigen-presenting cells, triggering the activation of T cells for targeted eradication, thereby potentially eliminating tumor cells [1–3]. However, in cancer patients, intricate interactions between the immune system and tumors disrupt immune function. Genetic mutations within tumors lead to the production of less immunogenic tumor subclones, which can evade immune recognition through the down-regulation of antigens or MHC, thereby failing to activate T cells for tumor-specific killing. Additionally, the tumor microenvironment harbors numerous immunosuppressive cells, including myeloid-derived suppressor cells (MDSCs), tumor-associated macrophages (TAMs), T-regulatory cells (Tregs), pre-tumor N2 neutrophils, and tumor-associated fibroblasts (CAFs) [4]. These cells can become activated, up-regulating the expression of immunosuppressive receptors (e.g., PD1 or CTLA-4) and secreting immunosuppressive cytokines (e.g., IL-6, IL-10, TGFβ, and VEGF), thereby impeding the tumor-killing effectiveness of T cells and leading to T cell inactivation or early apoptosis [5, 6]. The establishment of an immunosuppressive microenvironment, along with the emergence of tumor-hypoimmunogenic pressure clones, suppresses the immune system, pushing it into a state of tumor homeostasis or even tumor immune evasion. The theory of immune editing in cancer underscores the driving force behind tumor development [7, 8]. Consequently, the reactivation of the immune system and the stimulation of T-cell activity through targeted anticancer therapeutic interventions have emerged as novel strategies for tumor therapy. A fundamental principle of rational immunotherapy is to restore the cancer immune cycle, thereby enhancing and broadening the response of tumor-killing T cells (CTLs) against cancer cells [9].

Cancer immunotherapy has demonstrated significant promise in cancer treatment, with various approaches, including immune checkpoint blockade (ICB) therapy, cancer vaccines, and chimeric antigen receptor T-cell therapy, undergoing extensive investigation [10–12]. Among these, cancer vaccines stand out as a form of active specific immunotherapy capable of activating specific T cells by stimulating the body's immune system. Cancer vaccines offer the potential to target tumors with low immunogenicity, exhibiting low off-target lethality and high safety profiles. As such, they hold considerable promise in both cancer treatment and prevention efforts.

Cancer vaccines

Mechanism of action of cancer vaccines

The initiation of an adaptive immune response begins with the activation of the innate immune system, characterized by the recruitment of innate and adaptive immune cells via the release of inflammatory cytokines and chemokines [13]. Additionally, antigen-presenting cells (APCs), including macrophages, B cells, and dendritic cells (DCs), play a crucial role in this process. APCs capture soluble antigens through liquid-phase endocytosis or extract particulate antigens via receptor-mediated endocytosis. They then present these antigens to T cells, activating T cell-mediated killing by engaging co-stimulatory ligand receptors [3]. Among APCs, DCs primarily initiate antigen presentation and T cell activation. The rejection of tumors is predominantly attributed to cytotoxic CD8 + T cells, and the cellular immunity mediated by CD8 + cytotoxic T cells is crucial for eliminating malignant cells following cancer vaccine administration [14] (Fig. 1).

Fig. 1.

Cancer vaccine activates anti-tumor immunity. As a first step, antigen encounter by antigen presenting cells (APCs) such as dendritic cells (DCs) occurs at the injection site (or, in the case of DC vaccines, antigens may be exogenously loaded on APCs before injection). The antigen-loaded APCs traffic through the lymphatics to the draining lymph nodes, which are the primary site of T cell priming. In the lymph node, mature DCs present the tumor derived peptides on MHC class I molecules and MHC class II molecules to CD8 + and CD4 + T cells, respectively, of both naive and memory phenotypes. Immunosuppressive cells (such as MDSCs, Tregs, and M2 macrophages) and immunosuppressive cytokines can inhibit the activation of efector T cells and DC-mediated T cells directly or indirectly in TME. Cancer vaccines can promote antigen presentation and activate the body's anti-tumor immune response. There are two key components of cancer vaccines: tumor antigens, immune adjuvants. TAA, tumor-associated antigens; TSA, Tumor-specific antigens.GM-CSF, granulocyte–macrophage colony-stimulating factor; STING, stimulator of interferon genes protein; TLR, Toll-like receptor; DC, dendritic cell; MDSC, myeloid-derived suppressor cell; MHCI, MHC class I; NK cell, natural killer cell; PD-L1, programmed cell death 1 ligand 1; Treg cell, regulatory T cell. CTL, cytotoxic T lymphocyte; MDSC, myeloid-derived suppressor cell.Created in https://BioRender.com

Tumor antigens recognized by T lymphocytes form the cornerstone of cancer vaccine development, with antigen selection representing a pivotal aspect of vaccine design [15]. Initially, the focus lay on overexpressed antigens within tumors relative to normal tissues, known as tumor-associated antigens (TAAs) [16]. Examples include prostate-specific antigen (PSA), mammary erythropoietin-a, and differentiation antigens like Melan-A or gp100.

In 2021, OSE Immunotherapeutics unveiled the utilization of its cancer vaccine, Tedopi^®, for treating advanced non-small cell lung cancer patients unresponsive to immune checkpoint inhibitor (PD-1/PD-L1) therapy (NCT02654587). The vaccine targets 10 optimal neoantigenic epitopes from five TAAs typically expressed in lung cancer cells, stimulating T-lymphocytes to identify and combat cancer cells. Tedopi^® holds U.S. patent protection and is deemed a significant therapeutic for HLA-A2-positive non-small cell lung cancer. Nevertheless, TAAs must overcome immune tolerance to exhibit efficacy, and their expression in non-malignant tissues elevates the risk of vaccine-induced autoimmune toxicity. Clinical trials focusing on TAAs-based cancer vaccines have encountered limited success [17–19]. Tumor-specific antigens (TSAs) represent a distinct class of proteins expressed solely in tumor cells, arising from somatic mutations. Compared to TAAs, TSAs boast precise tumor specificity, heightened immunogenicity due to central tolerance absence, and limited “off-target” damage, rendering them ideal targets for cancer vaccines [20–22]. Various clinical trials have demonstrated promising efficacy with TSAs. Moreover, extracellular vesicles derived from heat shock-exposed dendritic cells and tumor cells have emerged as effective immunotherapeutic vaccines.

Tumor vaccines often exhibit low immunogenicity when relying solely on tumor antigens. Within tumor microenvironments, a plethora of immunosuppressive cells hinder T-cell function. Effective tumor vaccines typically necessitate immune adjuvants and nanocarriers to augment the immunogenicity of tumor antigens. Immune adjuvants typically act as stimulators of innate immunity, targeting APCs to furnish requisite co-stimulatory signals for successful antigen presentation [14, 23, 24]. Examples encompass Toll-like receptor (TLR) agonists, interferon gene-stimulating factor (STING) ligands, and lipopolysaccharide (LPS), such as poly IC [25]. Notably, the TLR3 ligand Poly I: C, renowned for inducing potent T1-type responses, when combined with vaccines, elicits robust CTL responses assessed in melanoma and glioma clinical trials. STING, a transmembrane protein activated by double-stranded DNA, triggers intracellular production of type I interferon, instigating innate immune signaling, APC maturation, and costimulatory signal upregulation. Efficient delivery vectors enhance APC uptake and processing efficiency for neoantigens. A diverse array of nanoparticles, including polymeric nanoparticles, liposomes, micelles, carbon nanotubes, mesoporous silica nanoparticles, gold nanoparticles, and viral nanoparticles, serve as delivery systems or adjuvants, either singly or in combination [26–28]. Liposomes, for instance, enhance target antigen immunogenicity for cancer vaccines and facilitate the delivery of RNA, DNA, and antigens [29]. Besides targeting APCs, liposomes shield delivered RNA from extracellular ribonucleases and ensure efficient uptake and expression by dendritic cells and macrophages across distinct lymphoid compartments. Thus, the judicious selection of tumor antigen, adjuvant, or delivery vehicle stands as paramount in cancer vaccine development. A robust cancer vaccine design must activate APCs presenting antigens to instigate efficacious effector T cell responses.

Current status of cancer vaccines

In 2010 the U.S. Food and Drug Administration (FDA) approved the first vaccine for the treatment of asymptomatic or minimally symptomatic hormone-refractory prostate cancer, Provenge (Sipuleucel-T), which was confirmed in a phase III clinical trial to reduce the risk of death and prolong the survival time of patients by an average of 4.1 months [30]. Multiple vaccines have since been developed and marketed and have shown potential applicability to significantly improve overall patient survival. Depending on the method of preparation, the current platform of cancer vaccines is divided into four main categories: cell-based vaccines, virus-based vaccines, peptide-based vaccines and nucleic acid-based vaccines.

Whole-cell vaccines

Vaccines that use whole cells as antigen carriers are known as cell-based cancer vaccines and can be divided into tumor cell-based vaccines and immune cell-based vaccines [31]. The effectiveness of whole tumor cell-based vaccines may be compromised by diluting the most immunogenic tumor antigen with all other autoantigens, essentially mimicking the endogenous presentation of tumor antigens to the immune system [32, 33]. Live tumor cells secrete soluble factors that inhibit immune cells and are less immunogenic [34]. For DC vaccines generated in vitro, dendritic cells are first differentiated and activated by cytokines and then loaded with antigen. This has been done in a variety of malignancies, including prostate cancer, renal cell carcinoma, melanoma, renal cell carcinoma, malignant glioma, and colon cancer [35]. DC vaccines and peptide vaccines have emerged as attractive anticancer immunotherapeutic options, but a number of barriers have hindered their successful application in clinical settings [36]. The low abundance of TAA-MHC complexes on the cell surface results in low yields, and it is cost prohibitive to produce enough dendritic cells for preparing cancer vaccines. Variations in the molecular composition of dendritic cells make vaccine quality control difficult to standardize production, and storage of DC vaccines is difficult to preserve. DC vaccines are sensitive to immunosuppressive molecules and signals present in the tumor microenvironment, and their function is inhibited in a tumor-mediated immunosuppressive environment [37]. Cell-based vaccines are an expensive therapeutic strategy, difficult to ensure standardized production, and lose efficacy in long-term storage.

Protein/peptide-based vaccines

Protein vaccines are typically composed of recombinant or purified proteins, and peptides are the most commonly used form of antigen in these vaccines. Unlike whole-cell vaccines, most peptide-based vaccines aim to target TAAs or TSAs via epitopes that stimulate CD8 + T cells or CD4 + T helper cells [38]. Compared to single-peptide formulations, multi-peptide vaccines generally demonstrate better efficacy, as they can be internalized and processed by professional APCs prior to MHC presentation, thereby providing optimal T-cell activation [39, 40]. Long peptides also typically contain MHC class II-restricted epitopes, which can stimulate both CD4 + and CD8 + T cells simultaneously.

However, TAAs must overcome immune tolerance to exert their effect, and since TAAs are also expressed in non-malignant tissues, this increases the risk of autoimmune toxicity induced by the vaccine. As a result, clinical trials of TAAs-based cancer vaccines have had limited success. In contrast, TSAs not only offer a more refined tumor specificity but also have higher immunogenicity due to the lack of central tolerance [41–43]. Additionally, TSAs induce limited "off-target" damage, making them ideal cancer vaccine targets.

The success of the neoantigen platform largely depends on tumor mutational burden (TMB). It is reasonable to assume that tumors with a high TMB may contain a higher number of neoantigens that can be targeted by vaccines and may respond better to immune checkpoint inhibitors (ICIs) [44, 45]. However, the identification and design of neoantigens present significant challenges. Although RNA sequencing provides useful information about the expression of mutated genes, helping to confirm mutations, not all mutations will lead to the creation of neoepitopes that can be recognized by the immune system. This is mainly due to HLA restrictions, as there are over 16,000 known alleles of HLA-A, HLA-B, HLA-C, and other genes. Therefore, when predicting potential immunogenic epitopes, individual HLA genotyping must be taken into account [38, 46–48]. Furthermore, the greater the heterogeneity of a neoantigen compared to the wild-type protein, the weaker its immunogenicity tends to be, as it is more easily eliminated through immune tolerance mechanisms. The affinity of the neoantigen for MHC presenting molecules is also a key factor in activating immune responses. Due to these challenges, peptide vaccines have limitations such as low immunogenicity, single epitopes, susceptibility to degradation, and vulnerability to immune escape by cancer cells.

In this context, exosome vaccines offer an effective solution. Exosomes provide a broad repertoire of tumor antigens, overcoming some of the limitations of traditional vaccine design. They carry a variety of antigens derived from tumor cells, which can induce a robust immune response and avoid the shortcomings of single-antigen vaccines. Exosomes not only provide a diverse range of antigens but also effectively activate dendritic cells, enhancing anti-tumor immune responses, and are therefore considered a powerful tool in cancer immunotherapy.

Virus-based vaccines

Viral vectors are another strategy for delivering antigens in vivo, the advantage of which is that the vaccine allows both innate and adaptive immunity to work together for an effective and long-lasting immune response. With the development of bioengineering techniques, methods such as virus-like particles are increasingly being used for therapeutic purposes. Enhancing the tumor microenvironment (TME) stands as a viable strategy for optimizing the effectiveness of viral vaccines. As our comprehension of the mechanisms underlying TME immunosuppression expands, various strategies for combining with viral vaccines have become feasible. [49, 50]. Combining virus-based vaccines with PD-1 inhibitors is the most common approach, showing long-term tumor-free survival in tumor models.M7824 is a novel bifunctional anti-PD-L1 and TGF-β that synergistically increases the effectiveness of virus-based cancer vaccines [51, 52].

Nucleic acid vaccines

Nucleic acid-based vaccines (DNA or RNA vaccines) represent a promising vaccine platform [53, 54]. These vaccines allow for the simultaneous delivery of multiple antigens that cover various TAAs or somatic tumor mutations. Additionally, nucleic acid vaccines can encode tumor antigens, enabling the simultaneous or cross-presentation of multiple class I and II specific HLA epitopes. As a result, they are less likely to be restricted by HLA types and have a higher potential to stimulate broader T cell responses [43, 48]. Due to the ubiquitous presence of RNases and structural differences between DNA and mRNA, DNA vaccines tend to have better stability and longer in vivo persistence than mRNA vaccines [55].

DNA vaccines offer several advantages, including the ability to encode a wide variety of antigens or larger antigens, high specificity, safety, low production costs, and ease of transport and storage [56]. However, they have relatively low immunogenicity, which often requires the use of adjuvants to enhance their efficacy, such as CpG motifs, polymers, nanoparticles, liposomes, and small molecule agonists [57, 58]. Moreover, DNA vaccines carry the potential risk of insertion mutations, whereas mRNA vaccines do not integrate into the host genome [59]. mRNA vaccines, on the other hand, are attractive and powerful platforms for cancer immunotherapy due to their efficiency, specificity, multifunctionality, rapid development potential, low manufacturing costs, and safety [60]. Recent advancements in mRNA vaccine design and delivery technologies have accelerated the development and clinical application of mRNA cancer vaccines.

mRNA vaccines combine the ability to encode virtually any protein with excellent safety, flexibility, and adjuvanticity. The adjuvant properties of single-stranded RNA, activated by TLR7 and TLR8, as well as the ability to encode multiple vaccine epitopes on the same RNA molecule, highlight their superiority as cancer vaccine formulations [61, 62]. mRNA activates innate immune responses through various RNA sensors, such as TLRs, RIG-I, and PKR. However, the application of mRNA in vaccine development has some limitations. First, naked mRNA can be rapidly degraded by extracellular RNases, preventing effective internalization by APCs [63]. Second, mRNA has inherent immunogenicity, which can activate downstream interferon-related pathways and induce innate immune responses. Although this intrinsic immunogenicity can serve as an adjuvant-like effect to enhance immune responses, it paradoxically accelerates mRNA degradation and reduces antigen expression [64].

To address these issues, the purity of mRNA products, modifications to mRNA sequences, delivery system design, and administration routes must be carefully optimized to appropriately activate innate immunity, initiate adaptive immune responses, and avoid excessive activation of toxic immune responses that suppress antigen protein expression [62]. Despite progress in overcoming these limitations over the past decades, the development of mRNA cancer vaccines still faces significant challenges.

Cancer vaccines represent promising immunotherapies aimed at stimulating the immune system to eliminate tumors and establish immune surveillance. A plethora of antigens, adjuvants, delivery strategies, and formulations have undergone testing, each presenting distinct advantages and limitations. While many cancer vaccines remain in preclinical and clinical research phases, the need persists for the development of more specific antigens and vaccine platforms. In this context, we propose a novel platform for cancer vaccines: the cellular vesicle vaccine. This innovative approach holds the potential to address multiple challenges inherent in current vaccine design. Its unique structure facilitates the integration of antigens, adjuvants, and vectors, thereby enhancing delivery efficiency and immune activation [65, 66]. Additionally, the cellular vesicle vaccine platform reduces production complexities, enabling easier quality control [67, 68]. This pioneering approach opens up new avenues for cancer vaccine design and warrants further exploration and development.

Cellular/bacterial vesicle vaccines

To enhance the therapeutic efficacy of antitumor vaccines, the utilization of nanocarriers for vaccine delivery has emerged as a prominent area of research. In comparison to the direct injection of peptide vaccines, employing nanocarriers offers several advantages. Firstly, it prevents the degradation of antigenic peptides, facilitating slow and sustained release. Prolonged antigen exposure, in turn, elicits a more robust immune response that effectively inhibits tumor growth [69]. Furthermore, nanocarriers enable the co-delivery of antigens and adjuvants, thereby enhancing T-cell responses [70, 71]. By co-loading antigens and adjuvants into nanocarriers, they can be efficiently delivered to lymphoid organs (such as lymph nodes or spleen) or antigen-presenting cells, leveraging the unique physicochemical properties of nanocarriers to improve delivery efficiency and enhance immunotherapy [72].Nanocarriers can be categorized into three main types based on their manufacturing methods: synthetic nanocarriers, semi-synthetic nanocarriers, and bio-derived nanocarriers [73–75]. Among these, bio-based nanocarriers are directly derived from living organisms, exemplified by extracellular vesicles (EVs), which are lipid bilayer-enclosed vesicles of cellular or bacterial origin. These bio-derived nanocarriers hold immense potential for tumor therapy owing to their low toxicity, excellent biodegradability, and biocompatibility properties.

Structure and properties of cellular and bacterial vesicle vaccines

The three main types of EVs secreted by cells are endosome-derived exosomes with diameters of 50–150 nm, microvesicles or “exosomes” with diameters of 200–1000 nm, and apoptotic vesicles with diameters of 100–5000 nm [72]. Bacteria can also secrete EVs, with both Gram-negative and Gram-positive bacteria producing bacterial membrane vesicles (BMVs) by different biogenesis mechanisms. EVs can facilitate antigen uptake by antigen precursor cells by mimicking the size, dimensions, and spatial structure of the pathogen [76]. They will preferentially target follicular dendritic cells (FDCs) and remain in FDCs [77, 78]. This is important for generating high affinity and durable antibody immune responses.

The composition of EV reflects that of the parental cells, and the nature and abundance of the substances it carries depend on the type and state of the cell, with properties similar to those of its cell of origin. Due to its bilayer structure and unique surface proteins, EV can serve as a highly efficient carrier of nucleic acids, proteins, and other antigens, and provide TAAs and other innate signals as a source of antigens presented by DCs. The special structure of EV can prolong the bioavailability of antigens (especially DNA and mRNA) by protecting them from being broken down by enzymes in vivo, thus maintaining their bioactivity [72]. especially DNA and mRNA) from being broken down by enzymes in the body to prolong their bioavailability and thus maintain their biological activity, and likewise provide an abundance of functional molecules (MHC-I, MHC-II, HSP70-90, and co-stimulatory molecules, etc.) capable of directly activating specific CTL responses in response to antigens as a viable option for cancer vaccination, and their lipid composition are more stable than cell-derived vaccines, more readily available on a large scale, have a longer half-life in the cycle of badness, and can induce more effective and longer-lasting immune responses compared to cell-derived vaccines [79].

In addition EVs mediate intercellular communication, their packaging contents may change with alterations in the cellular microenvironment or after therapeutic interventions, and their lipid membrane structure enables them to carry a wide range of molecules, which can be functionally modified either by modification of the parent cell or by direct functional modification of the isolated vesicles [73]. Genetic engineering modifications, such as CRISPR/Cas9 technology, allow flexible modification of exosomes, while lentiviral transfection techniques can induce exosomes to overexpress target proteins [80]. Physical surface modification of exosomes can be achieved by fusion with liposomes, insertion of lipophilic portions into membranes, and adsorption of molecules onto their surfaces. In addition to lipid insertion, the surface of exosomes can also be modified by adsorption [80]. In the chemical approach to exosome surface modification, vesicles are covalently attached by molecules to the amino groups of their surface proteins. Functional modification of vesicles can be achieved by utilizing the lipid membrane structure of exosomes for genetic engineering, as well as by enhancing the properties of biological membranes through chemical and physical modifications [81]. This has potentially important applications in improving the efficacy, persistence and targeting of vaccines. Its ease of cellular or genetic engineering, cargo loading, and subsequent functional modifications can be made to it to make the vaccine efficacy more powerful.

In conclusion, EVs possess distinctive characteristics that make them suitable as adjuvant carriers in cancer vaccine development. These characteristics include optimal size, biocompatibility, stability in the body's circulation, and the ability for target-specific delivery. EVs can be modified to co-deliver immune adjuvants, thereby synergistically activating APCs to enhance and regulate the immune response. This inherent capability positions cellular vesicles as a more advantageous platform for cancer vaccines compared to traditional antigen carriers. Presently, tumor cell exosomes, DC exosomes, and bacterial outer membrane vesicles represent the primary types of vesicles used in vaccine development. Their utilization holds promise for the advancement of cancer immunotherapy by harnessing the unique properties of EVs to optimize immune responses against tumors.

Sources of cellular/bacterial-based vesicle vaccines

Tumor cell exosomes

Tumor cells can secrete large numbers of exosomes to provide intercellular communication with surrounding and distant cells. These extracellular vesicles contain many types of mRNAs, functional surface proteins, enzymes, and lipids that allow them to exert local or systemic effects through direct interaction with cell surface receptors or transfer of their contents by recipient cells through plasma membrane fusion, endocytosis, phagocytosis, microcellular drinking, and lipid raft-mediated internalization [82]. Tumor-derived exosomes (TEXs) favor tumor progression and can carry immunosuppressive molecules such as Fas-L, TGF-β, TRAIL, PD-L1, and NKG2D ligands, which can directly inhibit the antitumor activity of effector CD8 + T cells and NK cells, and are involved in virtually all aspects of cancer progression, such as angiogenesis, proliferation, and metastasis. However, because of their ability to carry antigens and MHC-peptide complexes [83], they can provide tumor antigens to antigen-processing cells and enhance the recognition of tumor antigens by immune cells (Fig. 2). Exosomes isolated from tumor cells expressing heat shock proteins, including homologous HSP70 and HSP90, can also increase the immunostimulatory effect of TEXs [84]. In one trial, pulsed differentiated dendritic cells from autologous TEXs isolated from patient ascites could be used to induce tumor-specific CTLs, and in some cases even expand the restricted T-cell pool [85]. The immunoreactivity of exosomes in tumors is thus complex, and their concomitant pro- and anti-tumor effects in TME may be explained by their functional heterogeneity [86].

Fig. 2.

The role of tumor-derived exosomes in tumor immunity. Tumor-derived exosomes (TEVs) play a significant role in tumor progression and often contain immunosuppressive molecules like Fas-L, TGF-β, TRAIL, PD-L1, and NKG2D ligands. These molecules have the capability to directly inhibit the anti-tumor activity of effector CD8 + T-cells and NK-cells, contributing to various aspects of cancer progression including angiogenesis, proliferation, and metastasis. However, TEVs also possess the ability to carry antigens and MHC-peptide complexes. This enables them to provide tumor antigens to antigen-processing cells and enhance the recognition of tumor antigens by immune cells. Additionally, TEVs isolated from tumor cells expressing heat shock proteins, such as homologous HSP70 and HSP90, can further augment the immunostimulatory effects of TEVs.Created in https://BioRender.com

In vivo, TEXs inhibit the function of dendritic cells and immune cells and increase the activation of Tregs and myeloid-derived suppressor cells. However, through functional modifications, tumor exosomes can be transformed from an immunosuppressive state, to one that induces an anti-tumor immune response [87–89]. Under certain in vitro conditions, the antigen-presenting component of TEXs may predominate. And there are some subtypes of TEXs with the ability to present tumor antigens to CD8 + T cells and dendritic cells [90]. It has been shown that tumor-derived exosomes are effective in inducing antitumor responses both in vivo and in vitro, and that TS/A (breast cancer, H-2^d)-derived exosomes block autologous tumor progression in a CD4 + and CD8 + T-cell-dependent manner [91].Differences in the function of TEXs may be due to the heterogeneity of the TEXs themselves. It has been investigated that apparent ascites-derived exosomes (Aex) in combination with granulocyte–macrophage-colony stimulating factor (GM-CSF) have been used for immunotherapy of colorectal cancer (CRC). Aex + GM-CSF induces tumor-specific anti-tumor cytotoxic t-lymphocyte (CTL) responses [85]. Thus tumor-derived exosomes can act synergistically with adjuvants to activate anti-tumor immunity in vivo. The lipid-membrane structure and tiny size of exosomes can increase circulation time and promote efficient uptake and presentation of relevant antigens by APCs for efficient entry into the lymphatic drainage system, thereby initiating an anti-tumor response.

Compared with cellular vaccines, tumor cell membrane vesicles do not contain genetic material and have the advantages of better biosafety, easier large-scale manufacturing, long storage time, and good tumor-targeting ability due to the homology of the outer membrane. The main advantages of tumor vaccines derived from tumor cell vesicles over other vaccines are that they have their own library of tumor antigens and are easy to be genetically modified, and they have their own carrier function for antigen delivery. It is capable of fusing with specific target cells and transferring membrane proteins and internal components. Exosomes, in particular, are suitable for long-term ultra-low temperature preservation, which is very conducive to the production and storage of large quantities of tumor vaccines. Therefore, selecting appropriate tumor exosome subtypes, using genetic and chemical engineering of them with adjuvant synergism or extracellular vesicles to improve their targeting, and activation of immunity, and using tumor-derived exosomes as a source of shared tumor antigens could be a potential platform for cancer vaccines [92].

DC exosomes

DCs are the main cells for antigen presentation and play an important role in anti-tumor adaptive immune responses. DC-derived exosomes (DEVs) also express tumor antigens, MHC molecules, and co-stimulatory molecules on their surfaces, which can trigger the release of antigen-specific CD4 + and CD8 + T cells [93]. The surface of DEVs expresses antigen presentation molecules, including MHC-I and MHC-II molecules, and co-stimulatory molecules, including CD80 and CD86, which can directly induce corresponding T cell responses in vitro. -II molecules, as well as co-stimulatory molecules, including CD80 and CD86, which can directly induce the corresponding T cell responses in vitro [94]. However, this mechanism hardly occurs in vivo, probably because the T cell stimulatory activity of DEVs is less efficient than that of donor dendritic cells, and DEVs do not cause TCR cross-linking and co-stimulation to induce the release of naïve T cells. The aspect of antigen-specific T cell responses mediated by DEVs is mainly through indirect mechanisms.

DEVs can deliver functional peptide-MHC complexes carried to distal naïve dendritic cells. The target dendritic cells will be stimulated to mature and acquire the ability to activate antigen-specific T cells [95, 96]. Thus, exosomes act as carriers of antigenic substances disseminated in dendritic cells, exerting a mechanism that amplifies the amplification of the immune response (Fig. 3). In addition, DEVs can also induce T cell responses via tumor cells. DEVs is able to transfer MHC-peptide complexes to the surface of tumor cells, thus allowing tumor cells to be directly targeted by host T cells. In a recent in vitro study, it was observed that DEVs-treated human breast adenocarcinoma cells could re-stimulate previously activated T cells, resulting in an increase in T cell secretion of IFN-γ, suggesting that DEVs enhances tumor cell immunogenicity [97]. DEVs is not subjected to checkpoint signaling or immunosuppressive immune cells, such as Treg and MDSCs, and can also activate NK cells for innate tumor cell immunity [94, 98]. Expression of TRAIL, TNF, and Fas-L on the surface of DEVs activates NK cells and can directly induce apoptosis in cancer cells such as B16 melanoma, KLN205 lung squamous cell carcinoma, and MC38 colon adenocarcinoma cells [99].

Fig. 3.

DEVS activates T cells and NK cells in anti-tumor immunity. DEVs express on their surface antigen-presenting molecules, including MHC-I and MHC-II molecules, as well as co-stimulatory molecules, including CD80 and CD86, which can directly induce the corresponding T-cell responses in vitro. In vivo, T cells are activated mainly through an indirect mechanism, delivering the functional peptide-MHC complexes they carry to distal naïve dendritic cells. Thus, antigen-specific T cells are activated. The expression of TRAIL, TNF, and Fas-L on the surface of DEVs can activate NK cells.Created in https://BioRender.com

In contrast, the DEVs used in clinical trials were ordered from monocyte-derived dendritic cells (moDCs), and human-derived DEVs can be easily prepared using moDCs isolated from peripheral blood leukocytes treated with cytokines and other factors in vitro. The feasibility of using dimerization as a therapeutic anti-tumor vaccine has been validated in two phase I and one phase II clinical trials in patients with malignant melanoma and non-small cell lung cancer. However, externally differentiated moDCs have a lower migratory ability to activate T-cells, which leads to less than optimal clinical results. Few in vitro studies have used human dendritic cells as a source of DEVs, and most of the DEVs used in studies have used mouse bone marrow-derived dendritic cells (BMDCs) as a source for the production of DEVs [100]. For instance, loading α-fetoprotein in the HCC model [101, 102] or partner-rich lysates in GL261 glioma cell lines [103], antigen-loaded EVs can enhance overall survival and diminish tumor growth. Moreover, they can activate CD4 + and CD8 + T cell responses by increasing IFN γ and IL-2 production, promoting tumor infiltration, and facilitating the remodeling of the immune tumor microenvironment. This includes reducing Tregs numbers and decreasing IL-10 and TGF-β production. A study nonetheless used DEVs produced from human umbilical cord blood-derived dendritic cells and examined their antitumor capacity. In contrast to DEVs loaded with tumor peptides, total tumor RNA from a human gastric adenocarcinoma cell line (BGC823) was loaded into DEVs [104]. These RNA-loaded DEVs were superior to tumor cytotoxic T lymphocytes in inducing tumor-specific proliferation. This study also provides an alternative to moDCs, the source of human DEVs as mass-produced for use in anti-tumor vaccination methods.

The high levels of sphingomyelin and phosphatidylinositol in the lipid composition of exosomes make them more stable than cell membranes, and vaccine preparations can be cryopreserved at −80 °C for more than 6 months [105]. Compared with DCs, exosomes present 10- to 100-fold more TAA-MHC II complexes, and cell-free DEVs do not respond to immunosuppressive molecules, are less affected by the tumor microenvironment, and have a more potent stimulatory effect on T cells [106]. In contrast to chemokine-dependent migration, DEVs can be engineered for targeted delivery and can easily reach the appropriate locations in secondary lymphoid organs, the molecular composition of DEVs is more restricted and controllable than that of whole cells due to specific sorting and loading mechanisms [105], and the molecular composition of DEVs can be precisely defined, making quality control easier to achieve.

Thus DC vesicle vaccines have good immunogenicity and safety, and can be a promising platform for research on cancer vaccines by transferring to different host DC subpopulations, increasing the number of dendritic cells carrying tumor pMHC complexes, and acting as amplifiers of adaptive immune responses to induce anti-tumor immune responses in vivo.

OMVs

Bacterial membrane vesicles (BMVs) have emerged as new and promising platforms for the development of vaccines and immunotherapeutic strategies against infectious and non-infectious diseases. BMVs are rich in microbial-associated molecular patterns (MAMPs) and thus highly immunogenic, their nanoscale membrane vesicle structure allows them to possess a wide range of immune-modulatory properties, and they also have the ability to be genetically engineered, tolerant to exogenous big proteins and the ability to carry immunostimulants, making them an emerging and viable vaccine carrier [107]. Among them, both Gram-negative and Gram-positive bacteria produce bacterial membrane vesicles through different biogenesis mechanisms, and the presence of heterologous antigens in the MVs of Gram-positive bacteria has been relatively less investigated than in Gram-negative bacteria due to the latter's inability to release MVs through a thick cell wall thereby limiting the production of MVs [76, 108]. We focused on the prospects regarding Gram-negative bacterial membrane vesicles (OMVs) as bacterial outer membrane vesicles for the development of tumor vaccines.

OMVs have lipid-based spherical structures with diameters of 20–250 nm and are of appropriate sizes to be readily phagocytosed and efficiently processed by APCs thereby promoting adaptive immune responses. Mouse BMDCs were isolated and the ability of OMVs and non-vesicular OMV components (ultrasonication to disrupt the vesicular structure) to inject antigen into dendritic cells and stimulate dendritic cell maturation was compared [109]. The results showed that E7 carried by vesicles was taken up more by dendritic cells than by mixing with OMVs or strong adjuvants such as Freund's adjuvant, strongly suggesting that the vesicular structure of OMVs is important for activating the immune response. OMVs are enriched in bacterial MAMPs such as LPS, peptidoglycan, and flagellin with immunogenic components associated with their parental bacteria [65, 110]. The MAMPs of these OMVs enables them to bind to host pattern recognition receptors (PRRs), such as TLR4, initiating pro-inflammatory signaling cascades leading to cytokine and chemokine production [111, 112]. In addition, OMVs entering host cells are detected by other key components of intracellular host pattern recognition receptors, such as nucleotide-binding oligomerization structural domain protein 1 (NOD1) [113, 114]. OMVs can mediate inflammatory signaling through the NOD1 receptor, ultimately leading to the recruitment and activation of dendritic cells, which promotes the development of T-cell immunity. Therefore the natural adjuvant properties of OMVs are one of the advantages in the preparation of antitumor vaccines.

Anti-tumor vaccines based on OMVs mainly use genetic engineering techniques to make exogenous proteins expressed in vesicle lumen loading or on their membrane surface, and subsequently OMVs deliver loaded tumor antigens to APCs, thus activating anti-tumor immune responses. In contrast to eukaryotic cells, bacteria have rapid proliferative capacity, self-adjuvant properties, and easy-to-display gene editing techniques. They can be easily bioengineered to display desirable anti-tumor antigens [115]. (Fig. 4) Thus OMVs processing and translation is economical rather than cost-intensive. OMVs can be easily modified with exogenous antigens or epitopes by different synthetic biology methods, in particular by interfering with gene expression using protein antigen-encoding RNAs and siRNAs [116, 117]. Furthermore, OMVs formation involves a sorting mechanism in which proteins are screened and selectively localized to the periplasm or outer membrane. Using this mechanism, different tumor antigens can be delivered to the lumen or surface of the vesicle by fusing their coding sequences to those proteins that act as lead peptides [118]. Some tumor antigens can also be localized on the surface of the OMVs and thus be better recognized by the APCs, thereby inducing an immune response to kill tumor cells.

Fig. 4.

OMVs as tumor vaccines to activate anti-tumor immunity. OMVs are enriched in bacterial MAMPs such as LPS, peptidoglycan and flagellin with immunogenic components associated with their parental bacteria. The MAMPs of these OMVs enables them to bind to host pattern recognition receptors (PRRs) such as toll-like receptor 4 (TLR4), initiating pro-inflammatory signaling cascades that lead to the production of cytokines and chemokines. Antitumor vaccines based on OMVs mainly use genetic engineering techniques to enable exogenous proteins to be expressed in the vesicle lumen loading or on the surface of their membranes, followed by the delivery of loaded tumor antigens from OMVs to APCs to activate an antitumor immune response.Created in https://BioRender.com

In addition, OMVs has long-term stability. At milder conditions of 37◦ C and 50◦ C, OMVs did not show aggregation over an extended period of 4 weeks, and freeze-dried OMVs retained its antigenic and enzymatic activities even after long-term storage [116, 119]. Membrane composition and packaging of various proteins, phospholipids, and polysaccharides may provide superior resistance to environmental factors. It is suggested that their ease of preservation facilitates their production as tumor vaccines. In addition to the advantages of OMVs being stable, rigid and less susceptible to degradation in circulation, which reduces the degradation of loaded antigens, OMVs are unable to replicate compared to the bacteria themselves, and their ease of surface modulation allows them to establish a controlled immune response with a greater degree of safety [71]. The bioengineered OMVs-based vaccine platform enables flexible tumor antigen presentation and specific anti-tumor immunity in preclinical cancer models. OMVs have been successfully used to prepare vaccines against a variety of pathogens, including meningococcus. Although no cancer-related clinical trials have been reported yet, OMVs are emerging as extremely promising candidate cancer vaccine nanocarriers.

Design strategies for vesicle vaccines

The selection of vesicles as vaccine carriers has multiple advantages. However, a single vesicle is not sufficient to activate cancer immunity, especially as the cancer stage increases and the tumor microenvironment is formed, the immunosuppressive environment prevents the activation of T cells. In this case, strategies such as vesicle modification are needed to improve their immunogenicity or antigen presentation.

Co-delivery of antigenic adjuvants

Expression of these antigens on the surface of EVs may trigger stronger antigen-specific anti-tumor immune responses compared to soluble antigens, highlighting the great potential of EVs for vaccine development. Among them, tumor cell vesicles themselves carry antigens, enabling the co-delivery function of antigen and adjuvant. The released exosomes can all block the development of autologous tumors in a CD4 + and CD8 + T-cell-dependent manner, and tumor-derived exosomes are a shared source of tumor-rejecting antigens.

In the field of cancer immunotherapy, DCs play an important role in carrying and presenting tumor antigens as a crucial component of the immune system. DC exosomes are directly co-cultured with tumor lysates and directly loaded with tumor antigens. Antigens can also be carried indirectly by co-culturing DCs with tumor cells or lysate products and subsequent isolation of EVs [93]. For example, ovalbumin (OVA)-pulsed dendritic cell exosomes effectively induced the production of antigen-specific CD8 T cells [120], and AFP-enriched DEVs triggered an effective antigen-specific anti-tumor immune response by cell-free vaccines prepared by DC-derived EVs indirectly loaded with α-fetoprotein (AFP), remodeled the tumor microenvironment of HCC mice, and improved the overall survival rate [102]. Similarly, an EV vaccine with a direct device of HPV early antigen 7 peptide (E749-57) successfully induced an anti-tumor CTL response and activated effective protective and therapeutic immune responses in vivo [121]. In HER2/Neu + Tg1-1 mice with breast cancer HLA-A2 + HER2 + B16-F10 melanoma, this immunotherapeutic approach significantly activated potent tumor-specific CTLs and triggered protective immunity [122].

The OMVs can be realized in terms of antigen carrying either by surface modification or by embedding antigens internally. This approach has potential applications in the development of anti-tumor vaccines, mainly through genetic engineering techniques to achieve the expression of exogenous proteins inside OMVs or on their membrane surface. Internal loading of OMVs for MHC-I presentation and CTL activation was achieved by localizing HPV16 E7 protein between the inner and outer membranes of OMVs via Trx, which successfully mediated tumor regression [123]. Recently, the properties of the ClyA protein have been utilized to simplify the antigenic display process of OMVs, and "plug-and-play" technology has been developed [79]. By binding ClyA to the trap protein SpC/SnC, specific antigenic proteins bind to the labeling proteins SpT/SnT and are displayed on the surface of OMVs. In the B16-F10 melanoma model, the TRP2180- 188 (SVYDFFVWL) epitope of melanoma tyrosinase-related protein 2 (TRP2) was displayed on the surface of OMVs by plug-and-play technology. This increased the proportion of CD80 + and CD86 + dendritic cells, induced CD8 + CTL responses, inhibited tumor growth and lung metastasis, and triggered a robust anti-tumor immune response by tumor antigen-modified OMVs in a colorectal cancer model [79].Li et al. RNA-binding proteins on the surfaces of OMVs were modified by genetically engineering the RNA-binding protein L7Ae and lysosomal surface modification of the escape protein Listeria monocytogenes hemolysin O (OMV-LL) [124].OMV-LL-mRNA significantly inhibited melanoma progression and successfully led to complete tumor regression in 37.5% of colon cancer loaded mice.

Enhanced immunostimulatory capacity

Immunostimulation by vesicles themselves

OMVs are enriched with bacterial PAMPs, such as LPS, peptidoglycan and flagellin, which are immunogenic components related to their parental bacteria, and are able to directly interact with immune cells and epithelial cells through pattern recognition receptors, thus inducing the production of chemokines and cytokines and activating immune responses, and possessing the properties of natural adjuvants [125, 126].

Studies have shown that tumor-derived exosomes are effective in inducing anti-tumor responses both in vivo and in vitro, with apoptotic cell- and cumulus-derived exosomes being the focus of research due to the high immunogenicity they exhibit in cancer immunotherapy. Exosomes from apoptotic tumor cells exhibit stronger immunogenicity compared to normal tumor cells [127, 128]. And the immunogenicity of apoptotic vesicles in tumor cell vesicles is even higher than that of tumor cell exosomes. When mice were first immunized with apoptotic vesicles derived from the B16-F1 cell line and then faced with B16 melanoma implantation, they demonstrated a longer protective effect relative to mice receiving exosomes and microvesicles. This suggests that apoptotic vesicles may be uniquely advantageous in triggering a more durable and effective immune response [129]. Meanwhile, exosomes from peritonitis showed significant immune activation. In three-fifths of patients, it was found that cumulus-derived exosomes induced CTL initiation more effectively compared to tumor cell line-derived exosomes. Exosomes of cumulus origin showed more efficient CTL initiation compared to traditional exosomes of tumor cell line origin [91]. The immunogenicity of vesicles can be enhanced by a series of special treatments, including DEVs (HS-Co2-EV) from dendritic cells treated with heat stress and high CO2. In in vitro experiments, HS-CO2-EVs demonstrated an inhibitory effect on AGS gastric cancer cell proliferation (55% inhibition) and induced apoptosis (28% tumor apoptosis) [130]. Microvesicles were prepared by culturing C6 glioma cells with a single dose of 50 Gy radiation. It was found that irradiated C6 cell-derived microvesicles showed a significant reduction in tumor volume (more than 50%) in the treatment group [131]. These studies suggest that vesicle selection and treatment are critical for effective immunologic action, and these findings provide new perspectives for further research and development of vesicular cancer vaccines.

Adjuvant modification

Binding of activator ligands or co-stimulators to extracellular vesicles can significantly enhance their effectiveness during immunostimulation. For example, binding vesicles to TLR ligands, such as CpG-DNA and Poly: IC, or employing NKG2D ligands to activate NK cells. This strategy enhances the response of immune cells by binding to extracellular vesicles, thereby increasing the aggressiveness against cancer cells.

Immune adjuvants, which are usually stimulators of innate immunity, act on APCs and can provide the necessary co-stimulatory signals for successful antigen presentation. DC-derived EVs were activated by indirect loading of model antigen (OVA) and α-galactosylceramide (αGC), which is a ligand for i-NKT cells. Significantly delayed tumor growth was observed in a mouse B16-OVA melanoma model and increased the mean survival time of mice in the treatment group [132]. In the B16-OVA melanoma model, ev-loaded TLR3 ligand, LPS (TLR4 ligand) and TLR9 ligand adjuvant all significantly delayed tumor growth. Among them, EV-loaded Poly(I:C) was the most effective, activating CD8 + T cell infiltration within the tumor, inhibiting tumor growth and increasing survival [133]. In addition, EV produced by Poly(I:C) mature dendritic cells loaded with HPV early antigen 7 significantly inhibited the growth of TC-1 cervical tumors in mice and increased the survival rate in hormonal mice [121, 134]

The vesicles of tumor cells are themselves a good source of tumor antigens and have a powerful ability to activate autoantitumor immunity through adjuvant modification. TMV was prepared using subcutaneous tumors, and TMV was bound to the glycolipid-anchored immunostimulatory molecules GPI-B7-1 and GPI-IL-12 by protein transfer to generate a TMV vaccine. Tumor growth was inhibited in the HNSCC mouse model and the survival rate of SCCVII swollen mice was improved [135]. Engineered exosomes (Exos) exhibited potent antitumor activity by combining TLR3 agonists and inducers of cell death (ICDs), both in H mouse models and human breast cancer organoids, promoted in situ activation of cdc1, and ameliorated the subsequent tumor-responsive CD8 + T-cell responses [136]. IFN-γ fusion proteins were efficiently anchored on the surface of exosomes derived from prostate cancer cells and retained its biological activity, which significantly inhibited tumor growth and prolonged survival time in prostate cancer mice [137]. CpG DNA-modified exosomes were prepared by transfecting plasmid vectors containing endogenous tumor antigens and immunostimulatory CpG DNA into mouse melanoma B16BL6 cells using a genetic engineering approach. CpG DNA is a Toll-like receptor (TLR) 9 agonist, which is able to effectively activate DC cells and enhance their ability to present tumor antigens. It demonstrated potent in vivo anti-tumor effects in B16BL6 hormonal mice [138].

Targeted delivery

The targeting of vesicles can be enhanced by functional modifications to deliver antigen more efficiently. The exosomal DC cell lines selectively target cells expressing αv integrins by fusing the exosomal membrane protein lysosome-associated membrane protein 2 (LAMP2b) with the integrin-specific peptide iRGD through the DEVs system. This allows effective targeting of breast cancer cells in vitro and tumors in vivo [139]. DEVs can be engineered for targeted delivery and are capable of reaching appropriate locations in secondary lymphoid organs. Human DEVs directly loaded with the MART1 peptide (melanoma antigen; TAA) contain a complete assembly of functional peptide-MHC I complexes, which target dendritic cells and activate CD8 + T cells in vitro [95]. Modification of tumor antigens with DC surface receptor-specific ligands enables DC targeting, such as CD40 and sphingolipid Gb3, as well as DC cell targeting [140, 141]. Thus targeting DC cells and improving antigen delivery efficiency can efficiently activate anti-tumor immunity. By specifically modifying the glycocalyx of tumor cells, dendritic cell-specific intercellular adhesion molecule-3 grasping non-integrin (DC-SIGN), a natural ligand for high-mannose glycans, can be expressed on extracellular vesicles generated after induction of apoptosis. ApoEVs carrying DC-SIGN ligands can be efficiently internalized by DC cells, leading to enhanced initiation of tumor-specific CD8 + T cells [142, 143]. By inserting cholesterol-modified CpG and cholesterol-modified DC-SIGN aptamers as adjuvants into tumor cell membrane vesicles, a tumor cell-derived DC-targeted tumor vaccine (CMV-CpG/Apt) was successfully constructed to be able to specifically target dendritic cells, which resulted in improved delivery efficiency and triggered a stronger anti-tumor immune response [144].Huang et al. utilized lentivirus to internalize α lactalbumin (α-LA) mRNA transfection into breast cancer tumor cells [136] and enriched their secreted exosomes with a large amount of α-LA, thus enhancing the exosome targeting ability. The lipid outer layers of cell membranes or exosomes are also chemically and physically modified to enhance their targeting, immunostimulatory effects and improve in vivo delivery. Monophosphoryl lipid A is a potent lipid adjuvant that can embed in lipid membranes and induce dendritic cell maturation by targeting TLR4. [145, 146]. DC activation of CD40 is triggered by binding the antigen to a DEC205-specific antibody, leading to CD8 + T-cell activation and inhibition of B16-OVA tumor growth [141]. Choi and colleagues introduced a mannose-PEG coupling on the surface of EVs, which enhanced DC uptake via the mannose receptor. In vivo data demonstrated that mannose-targeted EVs exhibited increased accumulation in lymph nodes near the injection site compared to natural EVs.

Other strategies

Inhibition of the expression of immunosuppressive factors in exosomes can further amplify the anti-tumor immune response, as performed by Rossowska et al. They overexpressed exosomes secreted by genetically engineered MC38 and TGF-β1 shRNA cells expressing IL-12 and inhibited tumor growth by introducing IL-12 as well as by inhibiting TGF-β1, which promoted antigen-activated immunity by DC cell presentation [147]. In addition to loading loaded tumor antigens as a cancer vaccine, anti-tumor immunity can also be achieved by fighting stromal cells in the tumor microenvironment. OMVs loaded with full-length BFGF protein (154aa) as an antitumor vaccine successfully induced the body to produce high levels of anti-BFGF autoantibodies. These autoantibodies were effective in achieving anti-angiogenesis, disrupting the tumor immunosuppressive barrier, and promoting apoptosis [148].The study by Huang et al. employed TLR3 agonists and ICD inducers to specifically induce immunogenic death of breast cancer cells by designing the engineered exosome, HELA-Exos, which further amplified the antitumor immune response through the form of the original for vaccine. This study provides an innovative idea for cancer vaccines combining exosomes and immunostimulants [136].

Clinical trials

This bioengineered OMV based vaccine platform enables flexible tumor antigen display and specific anti-tumor immunity in preclinical cancer models. OMVs have been successfully used in the preparation of vaccines against Neisseria meningitidis, but there are currently no reported clinical trials related to cancer. Autologous monocyte-derived dendritic cell (MDC) cultures, loaded with HLA-elicited MAGE-A3, -A4, -A10, and MAGE-3DPO4 peptides (melanoma-associated antigens), have been investigated as a cell-free anticancer vaccine in Phase I clinical trials. Two main studies focused on patients with non-small cell lung cancer (NSCLC) [149] and malignant melanoma (MM) [150], demonstrating safety and feasibility in 2005.

In both trials, four doses of exosome vaccinations were administered at weekly intervals. Thirteen HLA A2 + patients with pre-treated Stage IIIb and IV NSCLC expressing MAGE-A3 or A4 were enrolled in this Phase I study. DCs were generated through leukapheresis, and DEX were produced and loaded with MAGE-A3, -A4, -A10, and MAGE-3DPO4 peptides. The results demonstrated that all three formulations of DEX were well-tolerated, with only grade 1–2 adverse events related to DEX use. Delayed-type hypersensitivity (DTH) reactivity against MAGE peptides was observed in 3 out of 9 patients. Immune responses included MAGE-specific T cell responses in 1 out of 3 patients and increased natural killer (NK) lytic activity in 2 out of 4 patients. This low-level T cell reactivity was attributed to the possible suppression of Tregs (CD4 + CD25 + T cells). In the second phase I trial, fifteen participants meeting the following criteria were enrolled: stage IIIb/IV, HLA-A1 + orHLA-B35 + , and HLA-DPO4 + leukocyte phenotype, with MAGE-3-overexpressed malignant melanoma (MM). No grade II toxicity was observed, and the maximum tolerated dose was not reached. One patient showed a partial response per RECIST criteria. Notably, a HLA-B35 + /A2 + patient vaccinated with A1/B35 defined CTL epitopes exhibited depigmentation around naevi, a MART1-specific HLA-A2 restricted T cell response in the tumor bed, along with a progressive loss of HLA-A2 and HLA-BC molecules on tumor cells during exosome therapy. Additionally, minor, stable, and mixed responses were observed in skin and lymph node sites. MAGE3-specific CD4 + and CD8 + T cell responses were not detected in peripheral blood. In the first Phase I trial using peptide-pulsed Dex, the observation of clinical regressions in the absence of T cell responses prompted the search for alternate effector mechanisms. Immature MCDC exosomes were shown to express NKG2D ligands on their membrane, facilitating binding to NKG2D on NK cells and resulting in pro-NK effects [151]. Following four weeks of vaccine administration, there was a significant enhancement in the number of circulatory NK cells. Post exosome immunotherapy, the expression of NKG2D and NK cytotoxicity were sustained in 50% of patients who initially had NK function deficits. Additionally, it was observed that exosome therapy could induce NK proliferation in vivo in an IL15Rα-dependent manner.

In the initial phase I studies, Dex proved ineffective in stimulating T-cell responses, prompting the exploration of innovative strategies to enhance Dex-based immunotherapy. Recognizing the superior efficacy of Dex derived from mature DCs compared to those from immature MoDCs in stimulating T cells, the decision was made to employ IFN-γ to stimulate human MoDCs in culture. A second generation of Dex (IFN-γ-Dex) was developed with the goal of enhancing NK and T cell immune responses. A total of 22 inoperable HLA-A2 + patients diagnosed with stage IIIB/IV disease received IFN-γ-Dex therapy [152]. Patients initially underwent three weeks of metronomic oral low-dose cyclophosphamide (CTX). This regimen was proven to decrease Tregs activity and induce IFN-γ and IL17-generating T cell clones based on several preclinical and clinical investigations [153, 154].The safety profile of IFN-γ-Dex therapy was favorable, with only one patient exhibiting grade III hepatotoxicity. Notably, one patient achieved long-term disease stability, experiencing complete regression of brain metastasis and becoming eligible for surgical tumor removal and local adjuvant radiotherapy. However, the trial’s primary endpoint, with the expectation of at least 50% of patients achieving a progression-free survival (PFS) of 4 months, was not met. Following nine injections, only seven patients (32%) exhibited stable disease and continued to receive vaccinations at 3-week intervals. Furthermore, the objective tumor response was limited. The median time to progression was 2.2 months, and the median overall survival (OS) was 15 months. However, despite the expectation of improved cooperation between T cells and NK cells, the second-generation Dex failed to induce cancer-specific T cell immune responses. Notably, BAG6 recovered in IFN-γ-Dex preparations demonstrated the ability to activate NK cells in an NKp30-dependent fashion.

The third phase I clinical trial involved the use of autologous ascites-derived exosomes (Aex) in combination with GM-CSF for colorectal cancer (CRC) Aex, isolated through sucrose/D2O density gradient ultracentrifugation, are 60–90-nm vesicles containing diverse immunomodulatory markers of exosomes and the tumor-associated carcinoembryonic antigen (CEA) [85]. Forty patients with advanced CRC (HLA-A0201 + CEA +) participated in the study and were randomly assigned to receive either Aex alone or Aex plus GM-CSF. Both groups received four subcutaneous immunizations at weekly intervals. Aex therapy demonstrated safety and good tolerance, with no autoimmune reactions and only grade 1–2 adverse events, including injection site reactions, pain, fever, nausea, and fatigue. The findings indicate that both therapies were safe and well-tolerated. Moreover, Aex combined with GM-CSF treatment showed superiority over Aex alone in activating CEA-specific T-cell responses, suggesting that GM-CSF could serve as a potential vaccine adjuvant. Despite the observed T-cell response, treatment response was only evident in one patient with stable disease and another patient showing a mild response after Aex plus GM-CSF treatment. Therefore, this study suggests that CRC immunotherapy with Aex in combination with GM-CSF is feasible, safe, and can be considered as an alternative in the treatment of advanced CRC. (Table 1).

Table 1.

EVs-based anti-tumour immunotherapy in clinical trials

EV origin	Phase	Loaded antigen	Cancer type	Immune cell effect	Clinical Outcome	Reference
Monocyte-derived DCs	I	MHC I Mage3A1/B35 and MHC II Mage3DP04	Melanoma (IIIB/IV Stage)	No MAGE3-specific CD4 +  and CD8 + T-cell responses; No DTH response; Enhanced NK cell function	Only grade 1 Toxicity,1/15 – partial response (according to RECIST criteria) In skin and lymph nodes: 1/15 – minor response 2/15 – stable response 1/15 – mixed response	[150]
Monocyte-derived DCs	I	MAGE-A3, -A4, -A10, and MAGE-3DPO4 peptides	NSCLC (stage IIIb and IV)	DTH reactivity against MAGE peptides in 3/9 MAGE-specific T cell responses in 1/3 NK lytic activity in 2/4 CMV responses Increase of Tregs in 2/3	Grade 1–2 toxicity,Stability in two progressive patients and continued disease stability over 12 months in 2/4 stable patients	[149]
Monocyte-derived DCs	I	MAGE3.A1 and MAGE3.DP04 peptides	Advanced melanoma	No MAGE3-specific CD4 +  and CD8 + T-cell responses 7/14 – restored the number and NKG2D-dependent function of NK cells		[151]
Exosomes from Ascites (Aex) with ± GM-CSF	I	Contain CEA	Colorectal cancer (stage III or IV)	DTH response as well as a CEA-specific CTL cell response	Grade 1–2 Toxicity,one patient with stable disease and another patient showing a mild response	[164]
IFNγ-matured moDC-derived DCs	II	IFN-γ, MHCI-peptides: MAGE-A1,MAGE-A3, NY-ESO-1, Melan-A/ MART-1, MHC II peptides:MAGE-A3-DP04, EBV	NSCLC (stage IIIb and IV)	No TAA-specific T-cell responses Increased NKp30- dependent NK cell function	Most grade 0–2 toxicity; one patient exhibiting grade III hepatotoxicity,one patient achieved long-term disease stability. 32% PFS at 4 months after chemotherapy	[152]

DCs dendritic cells, EVs extracellular vesicles, GM-CSF granulocyte–macrophage colony-stimulating factor, NK natural killer, Tregs T-regulatory cells, NSCLC non-small-cell lung carcinoma, CEA carcinoembryonic antigen, DTH delayed-type hypersensitivity, moDC monocytic-derived DC, EBV Epstein-Barr virus, CTL cytotoxic T lymphocytes, IFNΥ interferon-Υ

Summary of future perspectives

In this study, we proposed a new cancer vaccine platform, the cellular vesicle vaccine, aimed at overcoming several difficulties in current vaccine design. This platform has a unique structure that enables the integration of antigen and adjuvant carriers, thereby enhancing delivery efficiency and immune activation, while reducing production difficulties and facilitating quality control, providing new ideas for cancer vaccine design.

Extracellular vesicles, as lipid bilayer-enclosed vesicles of cellular or bacterial origin, play a multifunctional regulatory role in tumor immunity. They not only serve as mediators for mediating intercellular communication and are sensitive to changes in the tumor microenvironment, but also have homologous targeting and permeability across biological barriers, contributing to more efficient delivery of therapeutic substances to target cells [155]. Due to its specific structure and content, EV can be used as an efficient carrier for antigens such as nucleic acids and proteins, and its bilayer structure and surface proteins help to protect the antigens, prolong their bioavailability, and maintain their biological activity. In addition, EV has the ability to target FDCs, in which it is retained, providing favorable conditions for generating high affinity and durable antibody immune responses [77, 78]. Compared to cell-derived vaccines, vesicular vaccines have a more stable lipid composition and a longer half-life in circulation, increasing their stability and longevity in vivo. In addition, vesicular vaccines are easy to be cell-engineered or genetically engineered for subsequent functional modifications to make them more potent [20]. Therefore, vesicular vaccines, as a novel therapeutic strategy, have several advantages and offer new possibilities for progress in the field of cancer therapy. vaccine vectors such as OMVs, Dex, and TEVs have demonstrated efficacy in preclinical and clinical trials, providing new research directions for cancer therapy. In addition to therapeutic applications, researchers should increase their exploration of preventive cancer vaccines, such as specific cancers such as breast cancer for high-risk populations, or HPV vaccines [51, 52].

In patients with advanced NSCLC, vaccines may fail to generate appropriate targeted T-cell responses, which may involve insufficient MHC-I and MHC-II restriction antigens of the DEX used [152]. In addition, the presence of immunomodulatory molecules, such as PDL1, on the surface of IFN-γ-Dex may inhibit T-cell responses and, given the presence of an immunosuppressive milieu, especially in tumors with a highly immunosuppressive milieu, combining it with other immunotherapies may be useful to improve therapeutic efficacy. Thus combination with other immunotherapies may be a way and a future direction to improve the therapeutic efficacy of cancer vesicle vaccines [156].

However, like other immunotherapies, cancer vesicle vaccines face a series of challenges. Research on early-stage patients, as well as improvements in production quality and standardization, are key research directions. The tumor peptide library derived from exosomes secreted by long-term cultured tumor cell lines may not accurately reflect the natural peptide library processed by tumor cells growing in vivo. Additionally, the heterogeneity of tumor EVs is a major concern [157]. In animal experiments and clinical trials, naturally circulating dendritic cell (DC) vaccines have been shown to be safe and well-tolerated, with the ability to induce tumor-specific immune responses, leading to tumor eradication [92]. Moreover, their functionality is well preserved and sustained for extended periods, improving the reproducibility and efficacy in clinical applications. However, despite recent breakthroughs in cancer immunotherapy, the manipulation of dendritic cells or their EVs to develop more effective vaccine tools is still in the developmental stage. Subcellular and molecular-level optimization is still required to enhance their anti-tumor effects.

Various techniques have been developed for isolating and purifying outer membrane vesicles (OMVs), such as ultracentrifugation, ultrafiltration, precipitation, size exclusion chromatography, affinity separation, and density gradient centrifugation [71]. Although the process for isolating and purifying OMVs is relatively simple, these methods are limited to small-scale production with low yields. In industrial applications, there is a lack of standardization in the production conditions and preparation methods for OMVs, leading to significant variability in the composition and size of each batch [79]. The complexity of OMV production and the need for high uniformity also increase the difficulty of large-scale production. Therefore, compared to isolating exosomes, nanotechnology-based approaches may offer a more cost-effective and feasible strategy [158]. Establishing stable production lines and readily accessible technologies are crucial for clinical translation and large-scale application.

A tumor-derived nanoscale vesicle antigen complex, called nanoparticle vesicles (NV), has emerged as a new strategy for tumor cell-derived cancer vaccines [159]. NVs are directly purified from autologous tumor tissue via homogenization and sonication. These nanovesicles share similar characteristics with exosomes, such as surface marker proteins and morphology [160]. Artificial membrane vesicles can be obtained by extrusion and rupture of the tumor cell membrane. Similar to exosomes, these NVs carry various tumor antigens on their surface but lack excessive non-tumor-associated antigens, and their smaller particle size enables them to overcome biological barriers and remain in circulation for longer periods, making them an ideal choice for triggering tumor-specific immune responses [161]. Compared to exosomes, the extraction and preparation of membrane vesicles are relatively simple, and the vaccine preparation process has been established. Their yields are significantly higher than that of exosomes. Autologous tumor-derived NVs combined with adjuvants strongly induce immune responses in dendritic cells (DCs) in vitro and demonstrate anti-tumor activity against residual and metastatic tumors in vivo.

Genetic engineering techniques can also enable tumor-derived nanovesicles to carry specific molecules. For example, some studies have engineered tumor cells to express fibroblast activation protein (FAP) and then extruded them through a polycarbonate porous membrane to obtain large quantities of nanovesicle vaccines [162]. These vaccines contain both tumor-specific antigens and FAP antigens, allowing for the simultaneous targeting of tumor parenchymal and mesenchymal cells. This has resulted in significant anti-tumor effects in a breast cancer-bearing mouse model and induced ferroptosis in breast cancer cells. Another study developed a series of nanobodies (NBs) that included doxorubicin (DOX) and tyrosine kinase inhibitors (TKIs) loaded onto dendritic cell polymer-coated nanoparticles with hyaluronic acid (HA, for tumor targeting) to form HDDT NBs [163]. The study found that, after treating cancer cells with HDDT NBs, almost all cells (10–30 μm) were converted into uniformly sized microvesicles (1.6–3.2 um), achieving nearly 100% cell-to-vesicle conversion efficiency. The tumor tissue was thus transformed into an autologous vaccine, triggering systemic tumor-specific immune responses, including dendritic cell maturation, enrichment of M1-like tumor-associated macrophages (TAMs), and the accumulation of cytotoxic and helper T cells, as well as the establishment of immune memory.

Overall, the exosome vaccine platform shows promising potential in cancer therapy, but several challenges still need to be overcome. Through in-depth research, we hope to better understand its mechanisms and push it from the laboratory to clinical application, providing more effective treatment options for cancer patients.

Author contributions

WZ, XL, JG, and WT designed the topic, framework, and wrote the manuscript. SY, LT, XS, HW, and YD contributed to the figure creation and data organization. All authors read and approved the manuscript and agree to be accountable for all aspects of the research in ensuring that the accuracy or integrity of any part of the work are appropriately investigated and resolved. This study is not applicable to ethics approval and consent to participate.

Funding

This work was supported by the First Affiliated Hospital of Harbin Medical University Fund for Distinguished Young Medical Scholars (grant no. 2021J17); Beijing Medical Award Foundation (grant no. YXJL-2021–0302-0287); and Postgraduate Research & Practice Innovation Program of Harbin Medical University (grant no. YJSCX2023-63HYD).ORCID Id: Weiyang Tao (https://orcid.org/0000-0003-1107-5975).

Data availability

No datasets were generated or analysed during the current study.

Declarations

Competing interests

The authors declare no competing interests.