Advances in delivery technologies-powered cancer vaccines

Abstract

Over the past decade, cancer vaccines have shown great promise in immunotherapy for solid tumors. However, the efficacy of cancer vaccines is unsatisfactory due to the complexity, heterogeneity, and immune evasion of cancer, as well as the instability of the core components of the vaccines (antigens and adjuvants), the weak immunogenicity and low presentation efficiency of antigens, and the inability to effectively activate immune cells. Nanotechnology is considered to be a transformative approach to address these challenges by improving vaccine delivery. As carriers and/or adjuvants, nanoparticles take advantage of their superior physicochemical properties to enhance the stability of antigens and adjuvants, achieve controlled release in time and space, enable flexible and synergistic combination therapies, and create highly targeted delivery systems, thereby optimizing the efficacy and durability of antitumor immunity and minimizing side effects. In this review, the key components of the nanovaccine strategy are highlighted, which covers antigen forms, adjuvant types, nanovaccine platforms, and more. We provide the latest advances in the development of cutting-edge biomaterials and carrier systems for controlled vaccine delivery. Finally, integrating the current progress strategy, we offer critical perspectives on future applications of nanomaterials in cancer vaccines for clinical translation.

Graphical abstract

The efficacy of cancer vaccine relies on a well-orchestrated design: starting with the screening, selection, and modular assembly of antigens and adjuvants, followed by optimization of the delivery formulation, ultimately leading to a potent immune response. This integrated approach harnesses advances in delivery technology to enhance vaccine effectiveness.

Highlights

• •Rationally designed cancer vaccines integrate antigen and adjuvant into a single system to potentiate the immune response.
• •Nanocarrier systems for cancer vaccines enhance drug stability and targeted delivery to propel the cancer-immunity cycle.
• •Beyond conventional injections, vaccine delivery now incorporates hydrogels, microneedles, and mucosal systems.

1. Introduction

According to the latest statistics from the International Agency for Research on Cancer, there were nearly 20 million new cancer cases and 9.7 million new cancer deaths worldwide in 2022 [1]. The number of new cancer cases is projected to reach 35 million by 2050, a 77% surge compared to 2022, and a 90% spike in cancer deaths [2]. Despite significant progress in multiple areas of cancer treatment, cancer is a complex and heterogeneous disease that continues to pose daunting and complex challenges to human public health [[3], [4], [5]]. Most cancer patients undergo traditional modalities such as surgery, chemotherapy, radiotherapy, or targeted therapy, while all face some serious problems. Cancer immunotherapy has the potential to be an emerging strategy to improve cancer treatment outcomes using the patient's immune system to fight cancer [6,7]. Although many cancer immunotherapies, such as immune checkpoint antibodies and chimeric antigen receptor (CAR) T cell therapy, have shown great success in treating blood cancers, the treatment of solid tumors remains challenging [8,9]. Cancer vaccines are ideal for eradicating tumors through presenting tumor-specific antigens (TSAs) to the immune system, as a result, not only triggering an immediate and targeted immune response against primary cancer cells, but also building a long-term immune memory to prevent tumor recurrence and metastasis [[10], [11], [12]]. In the 70s of the 20th century, studies found that hepatitis B virus (HBV) is the main cause of liver cancer, and cervical cancer is strongly related to human papillomavirus (HPV), highlighting the potential of vaccines to prevent virus-induced cancer [13]. Subsequently, several generations of the United States (U.S.) Food and Drug Administration (FDA)-approved HBV vaccines and HPV vaccines were developed based on recombinant deoxyribonucleic acid (DNA) technology and genetically engineered viral proteins that could elicit an immune response without causing disease, thereby preventing viral infections and reducing the risk of liver cancer or cervical cancer by stimulating the immune system to produce antibodies against HBV or HPV [14]. In addition to prophylactic and viral cancer vaccines, there is a great deal of interest in developing vaccines that can treat non-viral cancers by TSAs with the deepening understanding of the role of the immune system in cancer. For example, the U.S. FDA has approved two therapeutic cancer vaccines: Sipuleucel-T, a prostate acid phosphatase-targeting dendritic cell (DC) vaccine for the treatment of metastatic castration-resistant prostate cancer, and Adstiladrin, an adenovirus vector vaccine for the treatment of high-risk Bacillus Calmette-Guérin (BCG)-unresponsive, non-muscle-invasive bladder cancer [15]. With the advent of the era of precision cancer treatment, personalized vaccines are expected to bring significant clinical benefits to different patients with cancer [16]. Talimogene laherparepvec (T-VEC), an oncolytic virus-based vaccine that stimulates a systemic immune response to cancer cells by exposing tumor antigens to the immune system after infecting and killing cancer cells, has been approved by the U.S. FDA for the treatment of certain metastatic melanoma that cannot be completely removed by surgery [17]. PCNAT-01, the first personalized neoantigen peptide vaccine that has received orphan drug designation from the U.S. FDA, has officially entered clinical trials for pancreatic cancer patients treated with radical resection and postoperative chemotherapy. Notably, cancer vaccines can be synergistically combined with other immunotherapy approaches to potentiate therapeutic efficacy. Moderna and Merck announced that they have received breakthrough therapy designation from the U.S. FDA for the adjuvant treatment of high-risk melanoma with tumor neoantigen messenger ribonucleic acid (mRNA) vaccine (mRNA-4157) in combination with programmed cell death protein 1 (PD-1) monoclonal antibody [18]. Last but not least, cancer vaccines are simple to prepare and relatively inexpensive, and have a wide range of application prospects and important scientific value in current cancer immunotherapy.

Despite the promising future, the effectiveness of cancer vaccines faces several limitations related to antigen recognition, immune tolerance, potential off-target risk, tumor heterogeneity, and immunosuppressive tumor microenvironment (TME) [19]. These barriers pose significant challenges to both the progress of cancer-related vaccines and their wider clinical application. In this context, combining cancer vaccines with nanomedicine to boost the treatment efficacy while minimizing the side effects may soon lead to a revolution in immuno-oncology. Nanotechnology has driven rapid advances in the field of drug delivery. By designing drug delivery systems on demand, it is possible to alter the biodistribution of various drugs, actively or passively target specific tissues and organs, thereby reducing off-target effects. Besides, delivery systems can be built with stimulus-responsive materials to achieve cargo release under specific conditions [20]. These characteristics make delivery technologies particularly well-suited to overcome the challenges faced by cancer vaccines. Nanoparticle-based vaccines, with their unique structure and physicochemical properties, have exhibited great promise in many preclinical studies [21].

In this review, we first summarize the classification and function of antigens and adjuvants, the core components of cancer vaccines. And the benefits of nanoparticle-based cancer vaccines are discussed. Then, the latest progress of various types of nano-biomaterials and delivery systems in cancer vaccines research is introduced (Fig. 1). Finally, we consider the challenges and opportunities in the emerging field of nanovaccines, with the aim of inspiring future cancer vaccine development by harnessing the potential of nanotechnology to provide more targeted, effective and safer treatments.

Fig. 1.

(A) The source of antigens in cancer vaccines. (B) The classification of adjuvants in cancer vaccines. (C) Different nano-biomaterials were used as cancer vaccine carriers. (D) The vehicles of cancer vaccination. (E) The efficacy of nanovaccines within the cancer-immunity cycle.

2. Antigens

Tumor antigens are a class of proteins that are specifically expressed on cancer cells or relatively overexpressed in cancer cells [22,23]. Antigens are the most essential and decisive parameter in developing potent cancer vaccines. Cancer vaccines clear tumors by directly activating the immune system. What's more, cancer vaccines can elicit immune memory to prevent recurrence and metastasis.

2.1. The classification of tumor antigens

Tumor antigens can be classified into tumor-associated antigens (TAAs) and TSAs based on their relationship to the tumor. TAAs represent a group of antigens that are widely present in cancer cells, including “self-antigens” such as tissue differentiation antigens (TDAs), overexpressed antigens, cancer-testis antigens (CTAs), and “non-self” antigens like viral or bacterial antigens [24,25]. TAAs can be found both on tumor surfaces and in certain normal tissue cells that share the same origin. While TAAs represent attractive candidates for universal cancer vaccine development due to their broad expression profiles, their limited tumor specificity poses significant challenges in avoiding toxicity against normal tissues.

Due to tumor heterogeneity, identical vaccines may exhibit differing therapeutic efficacy and prognostic outcomes among individuals with the same tumor type. TSAs are a class of neoantigenic antigens produced by viral infection, tumor genome variation or somatic cell mutation during tumorigenesis. Among TSAs, antigens derived from cellular gene mutations are known as neoantigens. They exhibit extremely high immunogenicity and represent the most ideal and safe targets among TSAs. In fact, neoantigens have become the central focus of current research and development in personalized tumor vaccines. Unlike TAAs, the TSAs do not exist in normal cells. Thus, as targets for cancer immunotherapy, TSAs can attack cancer cells selectively without affecting healthy cells.

The identification of TSAs is to screen out abnormal sequences by comparing the complete genetic sequence of the patient's tumor tissue with that of healthy cells. With the continuous maturation of artificial intelligence (AI) algorithms, numerous neoantigen prediction algorithms have been developed. With well-defined human leucocyte antigen (HLA) typing, directly assessing HLA binding affinity serves as a relatively mature and computationally efficient prediction method. However, it only reflects physical binding capacity and cannot predict immunogenicity [26]. In contrast, an alternative screening approach predicts the likelihood of peptide-triggered T cell responses by integrating features such as peptide-major histocompatibility complex (MHC) binding strength, T cell receptor recognition probability, and tumor expression levels through deep learning or ensemble modeling [27,28]. This method aligns more closely to the biological immune process and enables the prioritization of highly immunogenic candidate peptides. Furthermore, clonality analysis and multi-dimensional integrated ranking can enhance precision, but they also significantly increase computational and procedural complexity. In the last decade, several research groups successively used mass spectrometry (MS)-based shotgun proteomics analysis and whole-exome sequencing (WES)/ribonucleic acid sequencing (RNA-seq) techniques to identify TSAs and effectively treated patients with metastatic cholangiocarcinoma, advanced melanoma, or advanced non-small cell lung cancer (NSCLC) [[29], [30], [31], [32]]. While traditional methods can effectively predict mutations, they lack the ability to assess the actual presentation status of tumor antigens. NeoDisc integrates multi-omics data, including mass spectrometry immunopeptidomics, to directly detect peptides bound to HLA on the surface of cancer cells, demonstrating significant advantages in prioritizing immunogenic neoantigens [33]. Besides, researchers have already developed several new tools for screening candidate neoantigens, including personalized Variant Antigens by Cancer sequencing (pVACseq), Mutant Peptide eXpression Identifier (MuPeXI), and Neoantigen peptide screening and evaluation engine (Neopepsee) [[34], [35], [36]]. In addition to binary interactions (such as peptide-HLA or peptide-T cell receptor), Feng et al. focused on the ternary interaction among T cell receptor, antigen, and HLA. By incorporating the three core modules, they proposed a novel neoantigen screening pipeline called PISTE [37].

The TSAs can be categorized into shared TSAs (public TSAs) and personalized TSAs. Shared TSAs are common in different tumors and these mutated peptides can be used to prepare a broad-spectrum therapeutic TSA vaccine that can be shared by patients with the same mutated gene. Personalized TSAs are unique mutant peptides predicted based on the mutant profile of the patient's tumor genome. According to these patient-specific peptides, personalized TSA vaccines tailored to each patient are developed to trigger a powerful and long-lasting specific immune response to eliminate tumors and prevent recurrence effectively. Gritstone is developing two TSA-based cancer vaccines: a patient-specific cancer vaccine requiring a manufacturing period for each patient and an off-the-shelf cancer vaccine targeting shared neoantigens (NCT03794128). Notably, most of these studies are conducted in combination with chemotherapy, radiotherapy, or immune checkpoint inhibitor therapy. This design likely aims to overcome potential limitations of monotherapy, which may be particularly ineffective against tumors with low immunogenicity. Although early clinical trials have produced encouraging results, the large number of predicted TSAs often rarely triggered substantial antitumor responses [38]. This may be attributed to low throughput in experimental validation, heterogeneity of immune responses, insufficient correlation between theoretical predictions and actual clinical efficacy, and a lack of standardization in computational workflows. Additionally, the limitations of gene sequencing and bioinformatics technology lead to inaccurate identification of the type and number of TSAs during the screening process in each tumor sample, which may be one of the reasons why therapeutic vaccines based on TSAs have not yet been approved for standard cancer treatment.

2.2. Predefined peptide-based vaccines

The antigens of peptide-based cancer vaccines are polypeptide molecules prepared by chemical or biosynthetic techniques based on the amino acid sequence of a known or predicted epitope in a tumor antigen gene. These vaccines are designed to trigger a more focused immune response against critical epitopes by presenting the antigens to T cells [39]. Currently, multiple tumor peptide vaccines have entered clinical trial stages, covering various types of tumors, including: Melanoma Peptide Vaccine (NCT03849105) for melanoma, IMA950 (NCT01920191) for glioblastoma, Tedopi (NCT04884282) for NSCLC, NeuVax (NCT02297698) for breast cancer, among others. Peptide-based vaccines in clinical trials have evolved from early designs using short, broad-spectrum TAAs toward advanced strategies incorporating artificial synthetic growth peptides and personalized TSAs. Subunit cancer vaccines utilizing TSA peptides are superior to broad-spectrum TAAs, as the former can induce a precisely targeted immune response with higher safety and efficacy.

Although peptide vaccines have made substantial advances in clinical development, additional research is warranted to address existing limitations. During the delivery of predefined peptide-based vaccines, the peptides are susceptible to degradation by serum proteases, leading to poor in vivo stability. To address this issue, the design of suitable nanodelivery carriers provides an effective strategy. For instance, encapsulating peptides inside nanoparticles offers a protective physical barrier. Optimizing nanoparticle size facilitates lymphatic drainage and enhances passive targeting to lymph nodes. Surface polyethylene glycol (PEG)ylation is commonly used to extend circulation time, while stimulus-responsive release systems can be designed to release peptides only under specific conditions at the target site. Moreover, given the diverse properties of antigenic peptides, some cannot be efficiently taken up and presented by DCs. Chen et al. developed a nanovaccine platform based on Pox that enables the chemical conjugation of neoantigen peptides with diverse physicochemical properties. This self-assembling nanovaccine system, with a particle size of approximately 50 nm, significantly enhances antigen accumulation and lymph node infiltration, thereby improving antigen presentation [40]. Furthermore, since peptide vaccines are typically derived from self-antigens, central tolerance leads to the deletion of high-affinity T cell clones, ultimately compromising the potency of vaccine-elicited immune responses. The high heterogeneity and mutational burden of tumors further reduce the objective response rates of predefined peptide-based vaccines in clinical settings. Moreover, soluble antigens are degraded within lysosomes, leading to limited cross-presentation. This pathway mainly activates helper T cells (Ths) via the MHC II pathway, with limited activation of cytotoxic T cells (CTLs). To address this issue, promoting antigen escape from lysosomes into the cytoplasm is essential for enabling exogenous antigens to activate the MHC I antigen presentation pathway.

In the preparation process of tumor antigen peptide vaccines, several expression platforms have been used, including Escherichia coli (E. coli), plant, yeast, insect cells, and mammalian cells. Among these platforms, prokaryotic systems exhibit shorter production timelines but carry an elevated risk of endotoxin contamination. In addition, the absence of glycosylation in prokaryotic systems often reduces neutralizing antibody titers, necessitating the use of potent adjuvants to overcome suboptimal immunogenicity. In contrast, mammalian cell-expressed proteins retain native conformational epitopes that closely resemble natural tumor antigens, while presenting minimal risks associated with impurities. Nevertheless, this approach entails higher production costs and more complex bioprocessing workflows.

2.3. Nucleic acid-based vaccines

Nucleic acid vaccines, including mRNA and DNA cancer vaccines, are composed by genetic information of tumor antigens and will be the best choice for treating metastatic and mutation-prone tumors because of relatively short preparation cycle and long effective time. mRNA encoding tumor antigens can be delivered into host cells, enabling endogenous antigen expression and eliciting a potent, antigen-specific immune response against malignant cells. First, they are easy to be rapidly developed and produced through a cell-free process. Second, mRNA does not need to enter the nucleus, which reduces the risk of induced insertion mutations. Third, mRNA vaccines overcome the HLA restriction inherent to peptide vaccines while achieving more physiologically relevant antigen expression through host cell translation.

LK101 Injection (IND: CXSL2200612), China's inaugural self-developed autologous mRNA-DC vaccine engineered to deliver patient-specific tumor neoantigen mRNA, had received phase I clinical trial approval for advanced solid tumors. The vaccine utilized an mRNA-DC platform wherein DCs were transfected with mRNA encoding multiple neoantigens derived from the patient's tumor mutational signature. This approval represented the first clinical-stage evaluation of a personalized neoantigen mRNA-DC vaccine in China. Although the clinical trials have shown good application prospects, traditional mRNA vaccines face certain obstacles due to the instability of mRNA and the low translation efficiency caused by the difficulty in internalization of cells and the inability to replicate. To address the limitations of transient protein expression and high dosage requirements in non-replicating mRNA vaccines, self-amplifying RNA (saRNA) vaccines (such as Arcturus' ARCT-154) utilize encoded RNA replicase or insert viral replicase genes to enable prolonged mRNA expression. saRNA, as a second-generation mRNA antigen platform, enables sustained antigen expression at reduced doses and elicits more potent anti-tumor immune responses. However, the increased molecular size of saRNA presents ongoing challenges in manufacturing processes, delivery systems, and immunogenicity regulation. Furthermore, unmodified mRNA exhibits rapid clearance in vivo. To solve this problem, delivery carrier optimization can protect mRNA from enzymatic degradation while enhancing cellular uptake efficiency. Lipid nanoparticles with stability and scalable production capabilities are currently the most advanced delivery vectors for mRNA vaccines. The rational design of their lipid components can further enhance the efficacy of the mRNA vaccines. In addition to ensuring the stability of mRNA molecules, modified nucleotides similar to natural mRNA can also reduce the immunogenicity of mRNA and improve expression efficiency. Commonly used modified nucleosides are pseudouridine, 5-methylcytosine, and 2-thiouracil, which are integrated into the in vitro transcription of mRNA vaccines for improving mRNA stability and increasing translation efficiency while reducing the activation of innate immune responses.

DNA vaccines are usually bacterial plasmids encoding tumor antigens. Exogenous antigen genes are inserted into closed-loop DNA plasmids containing eukaryotic expression systems, and the recombinant plasmids are then introduced into host cells via physical methods such as gene gun or electroporation [41]. DNA-based vaccines combine multiple advantageous features: the capacity to encode diverse antigens with high specificity, rapid and scalable production capabilities, and superior stability that facilitates convenient storage and distribution. Although no commercially DNA cancer vaccines are available worldwide, there are several DNA cancer vaccines in clinical trials. As a DNA vaccine targeting HPV-16 and HPV-18, VGX-3100 has achieved landmark status as the first nucleic acid-based therapeutic to show statistically significant efficacy in phase III clinical trials for HPV-induced precancerous lesions. The latest phase I/II trial data of GNOS-PV02, a DNA plasmid-based personalized therapeutic cancer vaccine (PTCV), combined with the PD-1 inhibitor pembrolizumab, demonstrated a favorable safety profile and significantly enhanced survival outcomes in patients with advanced hepatocellular carcinoma (HCC) [42]. Although DNA vaccines have superior advantages in production process and antitumor capabilities, applications are limited due to the potential risks of their genome integration.

A multi-epitope strategy can be employed in the design of nucleic acid vaccines. This strategy employs the tandem encoding of multiple antigenic epitopes within a single nucleic acid sequence, thereby enabling the concurrent induction of broad-spectrum immune responses against diverse targets. This not only expands the immunological coverage of the vaccine but also enhances overall protective efficacy.

2.4. Cancer cell membrane-based vaccines

There are massive antigen epitopes and immunogenic proteins on cancer cell membranes, which are cell derivatives without genetic information and can be used as antigen sources for preparing cancer vaccines [43]. Compared with pre-encoded antigens, obtaining antigens directly from cancer cells is more convenient and specific, making it easier to generate personalized cancer vaccines [44]. Professor Liangfang Zhang reported a biomimetic nanodisc derived from cancer cell membranes (Fig. 2A) [45]. To form the nanodiscs, MC38 murine colorectal cancer (CRC) cell-derived plasma membranes were combined with monophosphoryl lipid A (MPLA) via ultrasonic blending, then incubated with styrene-maleic acid (SMA) copolymer under vortex mixing overnight. The small nanodiscs were easily taken up by antigen presenting cells (APCs) and effectively drained to lymph nodes, stimulating the immune system and promoting tumor-specific immunity.

Fig. 2.

Representative antigens in cancer vaccines. (A) Cancer cell membranes isolated from whole cancer cells and subsequently incubated with MPLA and SMA to form MPLA-loaded cancer cell membrane nanodiscs (CCND/MPLA) [45]. (B) The schematic of SCCNVs preparation for personalized cancer vaccination, along with their proposed mechanism of action [54]. (C) Cells engineered to produce EVs displaying an Fc-binding domain (Fc-EVs), enabling the conjugation of various monoclonal antibodies (mAbs) for precise delivery the loaded cargo [55]. (D) The preparation process and scanning electron microscopy (SEM) images of alginate cryogel sponge vaccines. The irradiated cancer cells were encapsulated with immune adjuvants cytosine-phosphate-guanine oligodeoxynucleotide (CpG ODN) and granulocyte-macrophage colony-stimulating factor (GM-CSF) [66,67]. (E) The schematic illustration of liposome-mediated immunochemotherapy [68].

In addition to excellent tumor antigen carrying and homologous targeting capabilities, cancer cell membranes retain the corresponding functions and surface physicochemical properties of the original cell membrane structure, which can reduce the uptake by the reticuloendothelial system and exhibit higher biocompatibility, biodegradability and safety, lower preparation cost and wider sources than synthetic antigens. Furthermore, due to the existence of specific homologous binding adhesion molecules, products derived from cancer cell membranes can specifically homologous bind to the same membrane protein [46]. The utilization of cancer cell membranes not only represents an effective way to introduce membrane-bound antigens but also a new approach to drug delivery as well as a novel personalized strategy for nanovaccines.

2.5. Cancer cell extracellular vesicle (EV)-based vaccines

EVs are closed vesicles secreted by cancer cells and consist of cancer cell membranes and cancer cell contents, containing various types of substances, such as multiple RNAs, lipids, and proteins. Numerous studies have shown that cancer cell EVs carry a large number of tumor antigens, which represent promising candidates for cancer vaccine development [[47], [48], [49]]. Like cancer cell membranes, cancer cell EVs have the advantage of low immunogenicity and can evade immune surveillance and recognition to successfully target tumor tissues. Simultaneously, cancer cell EVs possess inherent nanoscale architecture and exceptional biocompatibility, making them naturally suitable for direct application as drug delivery vehicles [50].

In addition to direct application of cancer cell EVs, more functional EVs can be obtained by pre-treating cancer cells [51]. The researchers described a strategy for constructing a vaccine using radiation-induced cancer cell-derived particles with a personalized and broad-spectrum antigen library that can induce comprehensive antitumor effects [52]. Li et al. developed a novel cancer vaccine platform based on EVs derived from pyroptotic cancer cells. These pyroptotic vesicles exhibit a unique immunogenic profile, enriched with TSAs and damage-associated molecular patterns (DAMPs), which collectively potentiate a robust antitumor immune response [53]. Hong et al. developed senescent cancer cell-derived nanovesicles (SCCNVs) as a PTCV that delivered patient-specific tumor antigens and improved vaccine immunogenicity without exogenous adjuvants (Fig. 2B) [54]. SCCNVs were produced by continuous extrusion of doxorubicin (DOX)-induced senescent cancer cells. These senescence-induced cancer cells exhibit a senescence-associated secretory phenotype (SASP), characterized by the expression of interferon (IFN)-γ and tumor necrosis factor (TNF)-α. These cytokines function as endogenous adjuvants to enhance the immunogenicity of vaccines, offering superior safety profiles compared to conventional exogenous adjuvants. Intradermal injection of SCCNVs can co-deliver tumor antigens and endogenous adjuvants to DCs, effectively activating immune response. In addition, EVs can be engineered to display certain proteins. Wiklander proposed a high modular technology by using molecular engineering tools to develop EVs that bind the fragment crystallizable (Fc) portion of antibodies, such as programmed cell death ligand 1 (PD-L1) or tumor-specific antigen peptides, improving the ability to specifically target cancer cells or further increasing the antigen abundance in EVs (Fig. 2C) [55]. Ito et al. transfected B16 cells with plasmid DNA containing the gene encoding a highly immunogenic bacterial antigen, the 6 kDa early secretory antigenic target (ESAT-6), to obtain EVs (ESAT-EV) expressing the immunostimulatory antigen ESAT-6, which could stimulate DCs and induce more production of CD86 antibody [56]. However, cancer cell EVs are involved in almost all aspects of tumor formation and invasion (including angiogenesis, proliferation, growth, and metastasis) and promote tumor immune escape by impelling DCs, natural killer (NK) cells, and T cells function disruption, which have somewhat hindered their direct use as safe cell-free cancer vaccines. Researchers from South Korea inhibited Yes‐associated protein (YAP) and autophagy to attenuate malignancy-related features of cancer cell EVs while improving immunogenicity [57]. Although genetic engineering is currently a common approach for preparing and modifying EVs, this method is also limited in practical applications. For example, the construction process is rather cumbersome, and it is necessary to consider whether the overexpression of the target protein will affect the normal viability of the donor cells.

2.6. Whole-cell tumor antigen-based vaccines

Apart from cancers caused by viruses and bacteria, such as cervical cancer, whole cancer cells/tissues are the best antigen source for developing remarkably effective cancer vaccines to induce specific immune responses and overcome the high heterogeneity of tumors. Compared with other tumor antigens, whole cancer cells as antigens have the merits of simpler preparation process, shorter preparation time and better safety [58].

The preparation of cancer cell lysates typically involves physical or chemical fragmentation methods, with repeated rapid freeze-thaw cycles being the most extensively adopted approach, as it effectively eliminates tumorigenic potential while retaining immunogenic features. Another established method for preparing cancer cell lysates involves inducing apoptosis or necrosis via ultraviolet (UV) irradiation before additional treatment. In addition, UV radiation can affect tumor neoantigens by inducing DNA mutations in cancer cells, which may enhance the host's antitumor immune response. OncoVAX, an autologous cancer cell lysate generated by UV irradiation and freeze-thaw cycles, was shown to significantly improve recurrence-free interval, overall survival, and recurrence-free survival in stage II colon cancer patients [[59], [60], [61]].

However, whole cancer cell lysates include water-soluble and water-insoluble components, making it challenging for a single APC to efficiently internalize all constituent molecules. Palacios et al. developed a breast cancer vaccine by combining the soluble fraction of autologous tumor tissue homogenate with formaldehyde and BCG suspension [62]. Preclinical and preliminary clinical studies demonstrated comparable therapeutic efficacy to conventional treatments, while exhibiting enhanced safety and tolerability. The water-insoluble component of cancer cell lysates may elicit stronger tumor-specific immune responses, whereas its poor solubility presents significant challenges for in vivo delivery. To address this limitation, Ma et al. developed nano-micron carriers by solubilizing insoluble components with urea, enabling concurrent delivery of both water-soluble and water-insoluble tumor antigens [63].

Beyond developing methods to utilize all cellular components post-lysis, researchers are also exploring alternative approaches that preserve intact cellular structures for tumor antigen presentation [64]. Inspired by single-celled diatoms, Guo et al. improved the biosilicification process and preserved TSAs from patients through low-temperature silicification [65]. Notably, the whole-cell antigens obtained by this approach can achieve stable, long-term storage at room temperature. Furthermore, rapid freezing in liquid nitrogen can be employed to inactivate cancer cells through a process analogous to viral inactivation. Mooney et al. designed a cavernous frozen gel loaded with cancer cells to deliver intact antigens derived from the frozen cancer cells to the tumor site (Fig. 2D) [66,67]. While inactivated cancer cells offer broad-spectrum antigenic coverage, this therapeutic approach faces challenges including potential incomplete inactivation and suboptimal immunogenicity.

Senescent cells represent promising candidates for whole-cell vaccines due to their limited proliferative capacity, prolonged in vivo persistence, and enhanced immunogenicity. The high immunogenicity of senescent cancer cells comes from their ability to activate IFN signaling, facilitate efficient antigen presentation, and modify immune peptide conformations. Senescence in cancer cells can be obtained through gene editing, which enhances immune recognition and strengthens immune surveillance against cancer cells.

2.7. In situ tumor antigen-based vaccines

Conventional cancer cell-based antigen vaccines require surgical tumor resection and ex vivo processing to generate therapeutic antigens, thereby excluding patients who are ineligible for surgery. This limitation has spurred interest in in situ vaccine strategies that induce cancer cell death while simultaneously releasing native antigens within the TME. Such approaches effectively transform solid tumors into endogenous antigen reservoirs, eliciting systemic tumor-specific immune responses. Importantly, in situ antigen generation offers superior safety compared to exogenous antigen administration, with significantly reduced risks of off-target toxicity.

Immunogenic cell death (ICD) is a common approach for providing personalized tumor antigens for in situ vaccines. This regulated cell death process is characterized by endoplasmic reticulum stress and reactive oxygen species (ROS) burst, which disrupt membrane integrity and promote the release of TAAs and neoantigens. Moreover, ICD is accompanied by the secretion of immunostimulatory DAMPs, which can be recognized by pattern recognition receptors (PRRs) on APCs, promoting their recruitment, activation, and maturation to ultimately prime potent T cell-mediated antitumor immunity. Numerous preclinical studies have shown that various therapeutic modalities, including anthracyclines, epigenetic modulators, molecularly targeted agents, radiotherapy, phototherapy, hyperthermia, electromagnetic therapy, and oncolytic viruses, can effectively induce ICD, thereby synergizing with immunotherapy. For example, Gu et al. developed liposomes encapsulating oxaliplatin (an ICD inducer) in combination with the stimulator of IFN genes (STING) agonist ADU-S100 to reprogram the immunosuppressive TME in CRC (Fig. 2E) [68]. Importantly, clinical evidence from hundreds of ongoing trials indicates that ICD-inducing chemotherapy can convert immunologically “cold” tumors into “hot” lesions, significantly improving their responsiveness to immune checkpoint blockade (ICB) therapy.

CryoVax represents a cutting-edge immunotherapy technology that has advanced to phase IIb clinical trials in U. S., targeting metastatic CRC patients who have developed chemotherapy resistance. This innovative approach utilizes tumor antigens generated through “cryoablation”-a minimally invasive procedure performed under computed tomography (CT) or ultrasound (US) guidance that destroys tumors via extreme cold temperatures, subsequently releasing their intracellular contents. The therapeutic protocol involves intravenous administration of AlloStim, a bioengineered allogeneic immune cell product, which localizes to the cryoablation lesion microenvironment containing tumor debris and serves as an adjuvant. This combination creates an in situ, patient-specific vaccine that can reprogram the local immune response and induce systemic antitumor immunity capable of targeting metastatic lesions throughout the body.

However, in situ-generated tumor antigens face several critical limitations [69]. On the one hand, necrotic cancer cells may paradoxically support tumor progression by either providing nutrients to surviving malignant cells or activating pro-tumorigenic mechanisms [70,71]. On the other hand, the release of tumor antigens remains an uncontrolled process, and the water-soluble antigens generated in situ are prone to rapid degradation while demonstrating poor permeability across lipid-rich cell membranes, ultimately limiting their effective uptake by APCs. To address these limitations, researchers have developed innovative strategies to enhance antigen utilization, including the employment of nanoparticles with specialized antigen-capture capabilities.

3. Adjuvants

Immune adjuvants can non-specifically enhance the immune response of antigens or modulate the type of immune response when injected into the body before or mixed with antigens [72]. Immune adjuvants act through the following mechanisms, including: (1) activating innate immunity via PRRs, (2) forming antigen depots to prolong exposure, (3) enhancing antigen uptake by immune cells, (4) stimulating the maturation of APCs, (5) promoting the release of cytokines and chemokines, (6) triggering inflammasome activation. Currently, seven adjuvants are approved by the U.S. FDA: aluminum salts, MF59, AS01, AS03, AS04, CpG 1018, and Matrix-M adjuvant authorized for COVID-19 vaccine emergency use [73].

3.1. Toll-like receptors (TLRs) agonists

As critical sensors of danger signals, TLRs orchestrate multiple immunological functions, including: recruiting immune cells to the vaccine administration site through chemokine production, and promoting antibody affinity maturation. Critically, TLR engagement not only enhances antigen capture, processing, and presentation via MHC molecules, but also induces the upregulation of APC maturation markers such as the co-stimulatory molecules CD80 and CD86.

Among TLRs expressed on the plasma membrane, targeting TLR 4 and TLR 5 is a common strategy for immune activation. Common TLR 4 agonists such as MPLA and glucopyranose lipid A (GLA) are structurally related to lipopolysaccharide (LPS) with significantly reduced pyrogenicity while maintaining potent immunostimulatory activity, thus increasing their feasibility for clinical use. Aminoalkyl glucosaminide 4-phosphate (AGP, marketed as RC-529) represents a synthetic structural analog of MPLA that demonstrates superior immunogenicity to traditional aluminum salts, inducing significantly higher antibody titers in clinical studies. Pan et al. developed a vaccine platform co-encapsulating ovalbumin (OVA)-encoding mRNA and MPLA within LNPs (sLNPs-OVA/MPLA) (Fig. 3A) [74]. MPLA can increase the expression of co-stimulatory molecules and the secretion of pro-inflammatory cytokines through TLR 4 pathway activation, stimulating APC maturation and then priming immune effector cells. In contrast to the predominantly cellular immune responses triggered by TLR 4 agonists, TLR 5 agonists can rapidly initiate innate immune responses [75]. Notably, flagellin-based adjuvants effectively stimulate mucosal immunoglobulin A (IgA) secretion, rendering them suitable for mucosal vaccine design. However, flagellin often triggers undesirable hyperreactivity and dose-dependent inflammatory responses by inducing excessive antibody production. Some studies have addressed this problem by removing B-cell epitope regions from flagellin while preserving TLR 5 adjuvant activity [76].

Fig. 3.

Representative immune adjuvants in cancer vaccines. (A) Schematic representation of spleen-selective co-delivery of mRNA and TLR 4 agonist-loaded LNPs for potent cancer immunotherapy through synergistic immunostimulation [74]. (B) The schematic illustration of self-assemble nanoparticles mediated activation of cGAS-STING pathway via DNA double damage [96]. (C) Schematic illustration of immune mobilization based on OMV-induced TRIM to enhance cancer vaccination [113].

Another major category of adjuvants requires intracellular activation of the corresponding TLRs. TLR 3 is an endosomal transmembrane receptor that primarily recognizes viral double-stranded RNA (dsRNA) and endogenous dsRNA released from damaged cells [77]. Several TLR 3 agonists have been developed, such as polyinosinic:polycytidylic acid (poly I:C), Riboxxol® G100, and artificial RNA-based nucleic acid immunostimulant X (ARNAX). Despite its immunostimulatory potency, poly I:C's clinical use is limited by dose-dependent inflammatory responses and cytokine release syndrome (CRS) [78]. The optimized Riboxxol® G100 and ARNAX improve their safety profiles by enhancing solubility and targeting capability, respectively [79,80]. Notably, the modified Poly-ICLC formulation (poly I:C stabilized with poly-L-lysine and carboxymethylcellulose) has shown both safety and efficacy in immune response activation [[81], [82], [83]]. More than 15 novel heterocyclic compounds, such as imidazoquinolines, imidazolines, pterins, pyrimidines, pyrropyrimidines, and benzimidazoles, have been synthesized as TLR 7/8 agonists, with resiquimod (R848) and imiquimod (IMQ) representing the most clinically advanced candidate [[84], [85], [86], [87]]. To avoid systemic inflammatory responses caused by small-molecule agonists, cholesterol conjugation is an effective strategy to achieve lymph node targeting and reduce off-target effects, which is a critical advancement for overcoming clinical translation barriers [88]. Kim et al. developed a cationic liposome ProLNG-001 for delivering ProLNG-S, a cholesterol-conjugated form of R848, to enhance immune efficacy and reduce systemic toxicity [89]. Following antigen uptake, TLR 9 activation promotes DCs trafficking to lymph nodes, thereby amplifying adaptive immune responses. CpG ODN can bind to TLR 9-expressing cells and initiate innate immunity followed by bridging to adaptive immune responses through perfectly mimicking the immunostimulatory activity of bacterial and viral DNA. Wang et al. developed a smart TME-responsive nanorobot to effectively deliver CpG to TLR 9-positive tumors, which activates TLR 9 signaling pathways, induces autophagy, and reprograms the immunosuppressive TME, collectively resulting in potent tumor growth inhibition and prevention of recurrence [90].

3.2. C-type lectin receptors (CLRs) agonists

CLRs represent a large superfamily comprising over 1000 proteins, including Dectin-1, Dectin-2, macrophage-inducible C-type lectin (Mincle), DC-specific intercellular adhesion molecule-3-grabbing non-integrin (DC-SIGN), CD206, and CD205. These receptors are ubiquitously expressed on the surface of diverse immune cells and play crucial roles in antigen capture and presentation. The mannose receptor (MR/CD206), a prominent member of the CLR family, serves as a key endocytic receptor predominantly expressed on macrophages and DCs. In 2006, White et al. employed mannose-functionalized liposomes to significantly enhance OVA uptake in monocyte-derived DCs [91]. The mannose targeting also contributes to increased cellular internalization and transfection of mRNA vaccines in immune cells.

3.3. Nucleotide-binding oligomerization domain (NOD)-like receptors (NLRs) agonists

NLRs agonists represent a class of immune adjuvants capable of triggering inflammasome activation and stimulating the immune system. Flagellin represents a unique dual-function PRR agonist capable of simultaneously activating both extracellular TLR 5 and intracellular NLR family CARD domain-containing protein (NLRC) 4 inflammasome pathways. For example, Vibrio vulnificus flagellin B (FlaB) has shown remarkable adjuvant potential when combined with various pathogens or tumor antigens [92]. Engineered Salmonella expressing FlaB can effectively remodel the TME upon localized delivery [75]. Based on these findings, the authors developed an innovative therapeutic cancer vaccine formulation combining Vibrio vulnificus FlaB, tumor antigen (HPV16E7ΔNLS), and DC-permeating peptide (DCpep6). This tripartite system successfully elicited durable tumor-specific cellular immunity, with mechanistic studies identifying the NLRC4 inflammasome pathway as the primary driver of vaccine potency [93]. It is worth noting that most NLR inflammasome sensors are located in the cytoplasm, so effective activation of NLR pathways necessitates substantially enhanced cellular uptake and cytoplasmic delivery than surface-expressed PRRs.

3.4. Cyclic guanosine monophosphate-adenosine monophosphate synthase (cGAS)-STING agonists

As an innate cytoplasmic DNA sensing pathway, cGAS-STING signaling pathway has garnered significant attention due to its pivotal role in activating both innate and adaptive immune responses to fight against tumor. Cyclic dinucleotides (CDNs) are the natural ligands of the STING pathway. However, the clinical translation of CDNs has been constrained by rapid enzymatic hydrolysis, poor cellular uptake, and suboptimal activation across all STING variants. IMSA101 is a CDN analogue that functions by activating the STING signaling pathway. Currently, it is in the phase I/II clinical trial (NCT04020185) to evaluate the safety and preliminary efficacy of IMSA101 as a monotherapy or in combination with PD-1 inhibitors in advanced solid tumors. Notably, Uslu et al. demonstrated that combining IMSA101 with CAR-T cell therapy enhances antitumor activity, as STING activation by IMSA101 promotes interleukin (IL)-18 secretion, thereby potentiating CAR-T cell function [94].

Compared to CDNs, small-molecule STING agonists offer advantages in chemical synthesis simplicity and pharmacokinetic properties. While the most advanced candidate, 5, 6-dimethylxanthenone-4-acetic acid (DMXAA), failed in phase III clinical trials, several other small-molecule agonists have progressed to clinical evaluation [95]. Beyond direct STING-targeting small molecules, DNA replication-disrupting therapies, such as DNA damage induced by chemo/radiotherapy, can also activate the STING pathway. However, conventional chemotherapeutic agents exhibit limited intrinsic STING-activating capacity, necessitating novel drug delivery strategies to enhance DNA damage efficiency and amplify STING-mediated immune responses. Cao et al. utilized ROS-responsive nanoparticles to deliver a hybrid platinum prodrug (CPT-PT(IV)), effectively activating the cGAS-STING pathway in a CRC mouse model (Fig. 3B) [96]. The nanoparticles facilitated tumor-specific accumulation of CPT-PT(IV), which subsequently released cisplatin and camptothecin (CPT), inducing synergistic DNA double-strand damage. This damage triggered robust cGAS-STING activation, leading to DC maturation and enhanced tumor infiltration of CD8⁺ T cells. Furthermore, strategies to amplify cytoplasmic DNA accumulation, such as inhibiting DNA repair via poly adenosine diphosphate-ribose polymerase (PARP) inhibitors or blocking cytoplasmic DNA degradation with deoxyribonuclease inhibitors, can potentiate cGAS-STING pathway activation.

Certain metal-based compounds exhibit the ability to directly activate the STING pathway. The immunomodulatory effects of manganese ions (Mn²⁺) were first documented three decades ago, with studies demonstrating that intraperitoneal administration of MnCl₂ significantly enhanced NK cell activity in mice, an effect postulated to be mediated by type I IFN signaling [97]. Mn²⁺ as a potent enhancer of innate immunity through cGAS-STING pathway activation, positioning it as a highly promising adjuvant candidate for next-generation immunotherapies.

Despite the promising potential of activating the cGAS-STING pathway for cancer immunotherapy, several significant challenges remain to be addressed. First, the STING pathway functions as a double-edged sword, which can also upregulate immunosuppressive molecules including PD-L1 and indoleamine 2, 3-dioxygenase (IDO) [98]. This limitation could potentially be overcome through combination therapies incorporating PD-L1 antibodies and IDO inhibitors. Second, the administration of high-dose STING agonists at frequent intervals may lead to undesirable T cell apoptosis, highlighting the need for careful optimization of both dosage and treatment schedules based on individual patient tolerance levels. Third, the presence of single nucleotide polymorphisms in the STING gene results in varied patient responses to STING agonists, necessitating careful consideration of patient selection criteria during clinical trials. Finally, the ubiquitous expression of STING across diverse cell types poses a major challenge, as systemic activation may trigger cytokine storms and cause damage to normal tissues. To address this, developing targeted delivery systems that can specifically activate STING within the TME while sparing normal tissues will be crucial. Importantly, many aspects of STING's dual mechanisms remain poorly understood, warranting further investigation to fully harness its immunostimulatory potential while minimizing adverse effects.

3.5. Inorganic compounds

Inorganic compounds are relatively safe and cost-effective adjuvants with a long history of use in both human and animal immunization. First described as an adjuvant by Glenny et al., in 1926, aluminum-based adjuvants have become the most widely used vaccine adjuvants, typically formulated as crystalline aluminum hydroxide, amorphous aluminum hydroxy phosphate sulfate, or potassium aluminum sulfate. These adjuvants offer several advantages, including excellent biocompatibility, high stability, well-established production processes, and low manufacturing costs. Functioning as antigen-delivery systems, aluminum salts form nanoparticle clusters in vaccines that provide stable adsorption surfaces for antigens through electrostatic interactions and ligand exchange. Following injection, this binding allows for gradual antigen release, prolonging immunostimulation while reducing both the required antigen dose and vaccination frequency. While aluminum adjuvants are effective at inducing Th2-type responses and enhancing humoral immunity, their ability to stimulate cellular immune responses remains a significant limitation. Furthermore, due to the poor biodegradability of aluminum adjuvants, their long-term biosafety warrants careful consideration. Other inorganic compound adjuvants such as calcium phosphate, zinc oxide and cobalt oxide exhibit comparable immunological functions and practical advantages [99].

3.6. Cytokines

Cytokines are soluble polypeptide signaling molecules secreted by immune cells and certain non-immune cells upon specific stimulation. As potent molecular adjuvants, they mediate intercellular communication and immune regulation. Well-studied cytokines include IFN, IL, TNF, GM-CSF, and chemokines [100,101]. Cytokines can enhance NK cell activity, direct T lymphocyte differentiation, broadly upregulate immune responses, and provide defense against bacterial, viral, and parasitic pathogens. Extensive research has demonstrated the adjuvant potential of cytokines in recombinant DNA and protein vaccines, where they effectively modulate host immunity [[102], [103], [104]].

3.7. Microbial adjuvants

Vaccines have traditionally been designed to induce adaptive immunity for protective effects. However, emerging evidence underscores the critical role of trained immunity (TRIM) in boosting innate immune responses, thereby enabling broad-spectrum defense against diverse pathogens. TRIM refers to the process where innate immune cells, after the first exposure to microbial products or inflammatory signals, undergo epigenetic and metabolic reprogramming to generate stronger and faster immune responses when subsequently stimulated again [105]. A wide range of biological agents, particularly bacteria-derived BCG and fungal-derived polysaccharides such as β-glucans, are recognized for their capacity to induce TRIM [106,107]. This immunomodulatory potential has been extensively studied in β-glucans sourced from organisms including Candida albicans, Trametes versicolor, Saccharomyces cerevisiae, and Peribacillus frigoritolerans [108]. Mechanistically, β-glucan-mediated TRIM initiates epigenetic reprogramming within monocytes and macrophages, resulting in sustained enhancement of their inflammatory responsiveness and antimicrobial effector functions. As a new type of adjuvant, the detailed mechanism for TRIM still needs further exploration.

Bacterial outer membrane vesicles (OMVs), naturally secreted by gram-negative bacteria, are nanoscale structures containing various bacterial bioactive components, including pathogen-associated molecular patterns (PAMPs) such as LPS [109]. These components endow OMVs with the ability to simultaneously activate both innate and adaptive immune responses, functioning as potent adjuvants. Studies have shown that DCs can effectively recognize and uptake OMVs, promoting antigen presentation and T cell activation through induced inflammatory responses [[110], [111], [112]]. Research conducted by Guangjun Nie and Xiao Zhao further demonstrated that pre-immunization with OMVs can induce TRIM, enabling long-term maintenance of innate immune cells in a highly active state. This process elicits innate immune memory, thereby enhancing the response and efficacy of subsequent tumor vaccinations (Fig. 3C) [113]. OMVs can be rapidly and mass-produced through fermentation, remain stable at room temperature, and can be stored long-term at 4 °C in lyophilized powder form. With high biocompatibility and significant industrial scalability, OMVs represent an ideal carrier for tumor vaccines. Their stable biomimetic nanostructure (20-300 nm) effectively protects and concentrates bioactive components such as mRNA, enabling direct delivery of sufficient payloads to distant target organs such as the spleen or tumor-draining lymph nodes, thereby activating immune cells. The bilayer lipid membrane structure allows for easy surface functionalization, enabling camouflage strategies to overcome challenges such as antibody-dependent clearance and high toxicity caused by highly immunogenic PAMPs during intravenous administration. Notably, the deposition of calcium and phosphate ions on the OMV surface to form a biomineralized shell not only enhances biocompatibility, hydrophilicity, and mechanical stability but also neutralizes the acidic pH of the TME upon degradation under acidic conditions, thereby enhancing both chemotherapeutic efficacy and immune responses.

3.8. Combination of adjuvants

Combined adjuvant systems constitute a forefront area in contemporary vaccine research, capitalizing on synergistic mechanisms to generate enhanced, balanced, and durable immune responses. The FDA-approved adjuvant AS04 (GSK) is a commercial product of MPLA combined with aluminum salts. Simultaneously activating different TLRs using multiple agonists will lead to a more intense secretion of cytokines and chemokines, thereby promoting the recruitment of immune cells. Li et al. selected various commercial adjuvants (TLR 4, TLR 7/8, and TLR 9 agonists) to test the effects of different adjuvant combinations on the maturation of DCs. They found that the combined application of three adjuvants was the most effective. Based on this, they successfully constructed vesicles expressing multiple TAAs and carrying three TLR agonists, which effectively stimulated the maturation of DCs and the activation of T cells, inducing systemic immune responses, and being effective in treating various tumor models and preventing tumor metastasis [114].

It is also possible to use adjuvants that jointly activate different immune pathways, such as concurrently engaging the TLR pathway while stimulating the cGAS-STING pathway. For instance, the research team led by Prabhani U. Atukorale designed a lipid-based super-adjuvant nanoparticle encapsulating both the STING agonist cyclic di-guanosine monophosphate (cdGMP) and the TLR 4 agonist MPLA [115]. This system achieved synergistic production of IFN I, induced polyfunctional tumor-specific T and B cell responses, and enhanced tumor-free survival after vaccination and challenge with aggressive cancer cells.

When using multiple adjuvants in combination, the physical and chemical compatibility, stability, production processes, and quality control standards among the various components need to be taken into consideration. Additionally, the combined action of multiple components makes it difficult to analyze the detailed molecular and cellular mechanisms. The computer screening method is used for the more rigorous screening of adjuvant combinations. Through iterative screening methods, Luo et al. screened 22 vaccine adjuvants and determined an effective LNP vaccine adjuvant combination C27a, which includes CL401 (TLR 2/7 agonist), Motolimod (TLR 8 agonist), 3D6A-PHAD (TLR 4 agonist), and QS-21 (inflammasome agonist). The C27a combination significantly enhanced the immune response of antigen-specific CD8⁺ T cells and increased the level of cytokine secretion. Moreover, the combination of C27a vaccine with anti-PD-1 antibody further enhanced the antitumor effect, demonstrating a good synergistic effect [116]. Furthermore, they also confirmed that the temporal and spatial co-localization of adjuvants and antigens can enhance the immune responses.

4. Benefits of nanovaccines

To develop effective vaccine delivery technology based on the aforementioned antigens and adjuvants, two fundamental challenges must be addressed: (1) ensuring efficient delivery of these components to the APCs, and (2) achieving effective activation of the APCs to trigger robust immune response. The rapid development of nanotechnology holds promise for addressing the limitations of free antigens and adjuvants by enhancing stability, reducing dosage requirements and boosting delivery efficiency, thereby achieving superior vaccine efficacy.

4.1. Amplifying/exerting the function of adjuvants

Recent advances in cancer immunotherapy have highlighted the importance of co-delivering antigens and adjuvants to APCs to potentiate tumor-specific T-cell responses. Optimal immune activation requires precise delivery of both components to the same DCs, particularly within intracellular compartments such as TLR-containing endosomes, which enhances DC maturation and antigen presentation [117]. Baljon et al. reported that co-encapsulating the adjuvants cGAMP and MPLA into nanocarriers was able to induce robust and tailored innate immune responses by increasing the immunogenicity of peptide antigens, thereby synergistically enhancing the expression of costimulatory markers, the secretion of pro-inflammatory cytokines, and the cross-presentation of peptide antigens in DCs [118]. The sequence and mode of antigen and adjuvant delivery also affect the immune response, so it is critical to design the fractionation/zonalization of the two components onto the nanoparticles and achieve their controlled delivery in a spatiotemporal manner. Wang et al. developed a nanovaccine system comprising a poly(lactic-co-glycolic acid) (PLGA) core loaded with the STING agonist cGAMP, enveloped by a fibrosarcoma cell membrane shell with high surface expression of CRT [119]. Exposing CRT to the surface of the nanovaccines helped to recruit DCs and stimulate uptake, thereby facilitating the simultaneous and efficient delivery of tumor antigens and STING agonists to DCs. This synergistic delivery strategy robustly activated the STING signaling pathway in DCs and elicited sustained immune stimulation.

Nanoparticles demonstrate unique immunostimulatory advantages by modulating immune activation to achieve appropriate immune responses. Notably, certain nanoparticle platforms possess inherent adjuvant-like characteristics, virus-like particles (VLPs) naturally exhibit adjuvant properties, while bacteria-derived nanoparticles can effectively activate DCs and enhance immunotherapeutic outcomes. This intrinsic immunogenicity represents a key advantage of nanoparticle-based systems in immunotherapy. In recent years, the development of antigen carriers that can act as immune adjuvants to further improve the efficiency and translational potential of cancer vaccines is a promising direction. To date, different polymer carriers have been studied as possible alternatives to conventional adjuvants. For example, polyethylenimine (PEI) has proven effective as a mucosal adjuvant for viral glycoprotein antigens, while its fluorinated derivatives with reduced cytotoxicity have shown potential as a self-adjuvant for cancer immunotherapy [120]. Xiaoyuan Chen's team have developed a tumor nanovaccine platform that integrated adjuvant functions into the delivery vehicle using branched-chain polyguanidine (PolyGu), where the guanidyl groups converted the cytotoxicity of PEI into innate immune activation [121]. The PolyGu-based nanovaccines potently stimulated DCs and promoted their maturation through the TLR 4 and NLRP 3 pathways, resulting in potent in vivo antitumor immunity.

4.2. Improving stability

The inactivation of antigens or adjuvants frequently arises from enzymatic degradation or certain chemical interactions (e.g., oxidation, hydrolysis, or transesterification) [122]. This instability not only limits vaccine shelf life but also hinders effective in vivo delivery through various administration routes. Overcoming the instability of free antigens and adjuvants is therefore critical for maintaining their structural integrity during storage while ensuring optimal bioavailability and therapeutic performance following administration. Encapsulation within the nanoparticles can protect antigens and adjuvants from enzymatic degradation, thereby improving vaccine stability, enabling sustained release, prolonging immune activation duration, and reducing systemic toxicity [15,123]. Nanovaccines have superior pharmacokinetics and bioavailability, which are essential for inducing robust and long-lasting antitumor immune responses. Tao et al. developed biocompatible aminated mesoporous silica nanoparticles (MSNs) to load CpG, which effectively prevented degradation by serum nucleases while simultaneously amplifying their immunostimulatory effects [124]. Similarly, LNPs have proven highly effective for mRNA vaccine delivery by shielding nucleic acids from both exogenous and endogenous nucleases. The clinical success of BioNTech/Pfizer and Moderna's COVID-19 mRNA vaccines exemplifies this technology's advantages, where LNPs not only protect mRNA from degradation but also prolong antigen exposure at injection sites compared to naked mRNA formulations [125].

4.3. Targeting immune organs or cells

After vaccination, the DCs-mediated T cell activation cascade involves four key stages: (1) lymphatic trafficking of vaccine components to secondary lymphoid organs, (2) antigen internalization and processing by DCs, (3) antigen presentation to naïve T lymphocytes, and (4) subsequent clonal expansion of effector T cells. The efficient production of antigen-specific effector T cells in organs such as lymph nodes and spleen plays a key role in cancer immune circulation, making the strategic targeting of cancer vaccines to these sites particularly promising [126]. Nanomaterials are considered as good candidates for immunotherapy due to their intrinsic ability to passively or actively target lymph nodes, spleen, and immune cells. With the targeting approach, nanoparticles can help achieve a more controlled immune response through selective cargo transport and limit unwanted immune overreactions or systemic toxicity caused by some immunotherapeutic agents.

Lymph nodes play a vital role in activating the adaptive immune system by providing an optimal environment for the proliferation, survival, and functional differentiation of diverse immune cell populations. Lymph node-targeting nanovaccines are a promising immunotherapeutic strategy, as their selective delivery to DCs within lymph nodes can potentiate robust antigen presentation and subsequent activation of tumor-specific T cell responses against both primary and metastatic malignancies. Particle size, surface charge, shape, and functionalization modification are the key physicochemical characteristics that determine the lymphatic trafficking efficiency of nanoparticles [127]. Nanovaccines are delivered to draining lymph nodes primarily through three size-dependent pathways. Nanovaccines smaller than 10 nm can diffuse into the bloodstream and are subsequently cleared. Those ranging between 10 and 100 nm benefit from the high permeability of lymphatic vessels, allowing them to enter lymphatics via convection and diffuse into lymph nodes. In contrast, nanovaccines larger than 100 nm tend to remain at the injection site, where they are internalized by migrating DCs and subsequently transported to lymph nodes through a DC-dependent homing mechanism. After entering the lymphatic system, tumor vaccines must also be effectively retained within lymph nodes to avoid premature drainage through efferent lymphatics. This retention is essential as it increases the opportunities for vaccine-DC interactions, which are critical for inducing a potent immune response. However, the retention efficiency largely depends on the internalization capacity of DCs. Positively charged nanoparticles are known to enhance DC internalization significantly due to electrostatic interactions with the negatively charged cell membrane. Additionally, the shape of nanovaccines also influences their retention in lymph nodes. Zhao et al. systematically investigated the impact of nanoparticle morphology on lymphatic uptake using antigen-loaded gold nanoparticles (∼40 nm diameter) with distinct shapes (stars, cages, and rods), and found that gold nanostars showed superior follicular accumulation and humoral immune response, while gold nanocages exhibited prolonged paracortical retention and stronger cellular immune activation [128]. Passive targeting in nanovaccine design can enhance the uptake efficiency of antigens by DCs through modifications in physical properties such as size, shape, and surface charge. Moreover, active targeting strategies that utilize receptor-ligand interactions can improve the specific internalization of vaccines by DCs. To date, active targeting of multiple surface receptors on DCs has been studied and well identified for precise vaccine delivery. Among these, CD40 has shown better antigen internalization efficiency compared to other surface molecules, enhancing vaccine efficacy and the generation of T cell responses. Rosalia et al. co-encapsulated OVA, TLR agonists, and anti-CD40 monoclonal antibody (aCD40-mAb) in PLGA nanoparticle (NP-CD40) to actively target DCs via the CD40 receptor, thereby promoting nanoparticle internalization [129]. Compared to non-targeted nanoparticle vaccines, NP-CD40 vaccines exhibited selective and improved antigen delivery to DCs in vivo, leading to enhanced DC maturation and robust activation of antigen-specific CD8⁺ T cell. NP-CD40 vaccines not only inhibited tumor progression but also significantly prolonged survival in preclinical models. Beyond CD40, subpopulation-specific targeting of C-type lectin receptors (e.g., CD205 and mannose receptors) can further refine nanovaccine delivery to distinct DC subsets, optimizing antigen uptake and immunogenicity. Expanding on DC-targeted delivery, Yin et al. developed a VLPs as mRNA vaccine vectors with an engineered Sindbis-virus glycoprotein that recognized surface markers on DCs [130], which elicited substantially stronger and more durable antigen-specific cellular immune responses compared to non-targeted VLPs or LNP formulations.

Notably, the spleen as the largest peripheral lymphoid organ is another critical target for nanovaccines delivery due to its ability to robustly trigger APCs activation and subsequent T cell responses [131]. For targeted delivery to the spleen, the accessory lipid 1, 2-dioleoyl-sn-glycero-3-phosphate (DOPA) can be incorporated into LNPs containing ionizable lipids or senescent erythrocyte membranes can be coated onto nanoparticles [132]. However, the intricate lipid composition required for such delivery systems may hinder their broad application in peptide vaccines, while ionizable lipids pose potential dose-dependent toxicity concerns [133]. He et al. showed that liposomes with a complement C3-containing protein crown and high membrane fluidity could efficiently accumulate in the spleen [134]. By further engineering these nanocarriers to co-deliver antigenic peptides, CpG, and a sonosensitizer (termed LNVS), the team demonstrated that intravenous administration of LNVS efficiently target the spleen compared to soluble vaccines (a mixture of free antigen and adjuvant), thereby in markedly amplifying antigen-specific CD8α⁺ T cell responses.

4.4. Enhancing lysosomal escape and cross-presentation of antigens

Following antigen uptake, APCs process and present antigens through two distinct pathways: MHC II molecules present exogenous antigens to activate CD4⁺ T cells, whereas MHC I molecules present endogenous antigens to activate CD8⁺ T cells. CTLs (CD8⁺ T cells) are crucial immune cells that kill cancer cells directly, while Ths (CD4⁺ T cells) play a role in coordinating the immune response. Therefore, vaccine design should focus on promoting efficient antigen loading onto MHC I molecules to achieve optimal CTL-mediated antitumor immunity. This process critically depends on the successful translocation of protein or peptide antigens from the endosome/lysosomal to the cytoplasm. Thus, cytoplasmic delivery of antigens is a pivotal step in cross-presentation, and nanoparticles that facilitate this process through either endosomal/lysosomal escape or membrane fusion mechanisms can significantly enhance MHC I presentation of antigens, leading to robust CTL activation.

The proton sponge effect stands as the predominant approach for enhancing antigen cross-presentation. When cationic polymer- or lipid-based delivery systems (such as PEI and 1, 2-dioleoyl-3-trimethylammonium-propane (DOTAP)) containing protonable amine groups are internalized by APCs and subsequently buffer the acidic environment of endosomes or lysosomes through proton absorption. This proton sponge effect disrupts ionic equilibrium within the vesicles, leading to an influx of chloride ions and water. The resulting osmotic pressure causes vesicular swelling and eventual rupture, thereby facilitating antigen release into the cytoplasm and presentation via MHC I molecules. Alternative strategies employ pH-responsive nanocarriers that exploit the acidic endosomal environment to destabilize endosomal/lysosomal membranes through membrane fusion or gas generation, thereby enabling endosomal escape of antigens [135]. The third method is to use photosensitizers that are excited to form singlet oxygen when exposed to a specific light source, resulting in lipid peroxidation, disruption of endosomal membranes and subsequent antigen release into the cytoplasm [136].

4.5. Regulating tumor immune microenvironment

The poor infiltration of APCs and CTLs as well as the presence of an abundance of immunosuppressive components (such as M2-polarized TAMs, MDSCs) in immunologically “cold” tumors inevitably hinder the host's natural immune response, making these tumors easy to evade immune surveillance and immune clearance, and prone to inducing progression and metastasis, thereby limiting the satisfactory efficacy of cancer vaccines. Accumulating evidence suggests that inducing ICD in the tumor region following specific treatments, such as chemotherapy, radiotherapy, phototherapy, or sonodynamic therapy, can trigger the release of DAMPs and pro-inflammatory molecules, thus facilitating immune cell infiltration and remodeling the TME to form immunologically “hot” tumors [137,138]. Nanoparticles enable precise delivery of ICD inducers (including chemotherapeutic agents [139], radiosensitizers [140], photosensitizers [141], and sonosensitizers [134]) to the tumor sites, while minimizing systemic toxicity, creating favorable immune-activating microenvironments, and improving therapeutic efficacy [142].

Furthermore, when integrated with nanovaccines, strategies designed to improve deep tumor penetration can significantly amplify ICD-mediated antitumor immune responses through synergistic mechanisms. Wang et al. constructed US-driven nanovaccines (G5-CHC-R) by combining the sonosensitizer Chenghai chlorin and the immunomodulator R848 onto an ultra-small dendritic nanoscaffold (Fig. 4) [143]. Once G5-CHC-R entered the tumor, R848 released in response to the tumor hypoxic microenvironment, thereby modifying the TME by switching the macrophage phenotype. Under a free-field-based whole-body US system, US-triggered G5-CHC-R exhibited remarkable tissue penetration capabilities and spatiotemporal precision, which could amplify local immune responses by driving tumor ICD and the “cold-warm-hot” tristate transition of TME. Besides, the sono-nanovaccines also elicited robust adaptive antitumor immunity for successfully inducing systemic tumor suppression.

Fig. 4.

Reprogramming the tumor immune microenvironment by nanovaccines. (A) Schematic diagram of the sono-nanovaccines G5-CHC-R structure and the therapeutic mechanism driven by free-field whole-body US irradiation. (B) Transmission electron microscopy (TEM) image of G5-CHC-R. (C) High-performance liquid chromatography (HPLC) analysis showed catalytically induced dissociation of R848 from G5-CHC-R. (D) Flow cytometry quantification of the effect of released R848 on macrophage phenotypic transformation. (E) In vivo fluorescence imaging tracking G5-CHC-R biodistribution at different time points post-injection, and ex vivo tissue imaging at 12 h. (F) Immunofluorescence staining of CRT and HMGB1 in tumor tissues resected after various treatments. Quantitative flow cytometry analysis of immune cell populations in pancreatic tumors: (G) M1 TAMs, (H) M2 TAMs, (I) mature DCs, and (J) MDSCs. (K) Immunofluorescence analysis of CD8⁺ T cells and IFN-γ production in tumor sections. (L) Representative photographs showing the tumor nodules in the intestine and the corresponding hematoxylin & eosin (H&E) staining of tissue sections [143].

4.6. Forming in situ nanovaccines

Tumor heterogeneity exists both inter-individually (among patients with identical tumor types) and inter-tumorally (across distinct tumor classifications), consequently impeding the identification of universally applicable TAAs. In addition, the development of cancer vaccines faces bottlenecks due to the biological complexity and the technical hurdles in manufacturing personalized vaccines ex vivo. Therefore, the urgency of developing simple yet effective therapeutic cancer vaccines is highlighted, and the identification of TAAs is the basis for eliciting specific and productive antitumor immunity [139]. Nevertheless, the process of tumor antigen identification remains complex and resource-intensive. To overcome the challenges, in situ cancer vaccination has emerged as a promising alternative strategy that directly utilizes antigens from cancer cells without requiring prior antigen identification or isolation, thereby enabling the induction of robust, personalized immune responses. Orthotopic vaccination typically promotes the release of antigens by inducing ICD [144,145], which allows for the development of vaccines in vivo, thereby reducing the extensive resources needed for conventional ex vivo vaccine production. Notably, this strategy enables comprehensive utilization of the tumor's entire antigen repertoire, circumventing the limitations of single TAA.

Normally, TAAs are released from apoptotic cancer cells and engulfed by APCs, thus inducing their activation and maturation. However, due to the body's intrinsic immune clearance, only a fraction of TAAs is successfully processed and presented by APCs to T cells, while the majority undergo lysosomal degradation, resulting in an ineffective activation of antitumor immune response and induction of abscopal effect. Recent advances in biomaterials and nanotechnology have revolutionized the field of in situ cancer vaccination. These well-designed nanocarriers can deliver therapeutics to tumor sites, induce tumor ICD and capture TAAs through multiple interactions, including both non-covalent binding mechanisms (e.g., hydrophobic interaction, electrostatic attraction, and metal coordination) and covalent conjugation strategies (e.g., maleimide (Mal)-mediated thiol-ene reactions and Schiff base reactions) (Fig. 5A–C) [6,146,147]. In situ nanovaccines demonstrate superior antigen retention capability and enable efficient delivery of tumor antigens in the form of nanoparticles to DCs, significantly enhancing immune activation. In particular, certain engineered nanoparticles not only capture tumor antigens but also promote endosomal/lysosomal escape, markedly improving antigen cross-presentation efficiency. Huang et al. developed cationic lipid-based system (FCD-NPs) using 2 mg DOTAP modification, which achieved effective antigen capture while maintaining excellent biocompatibility (Fig. 5D and E) [148]. Furthermore, multifunctional FCDP-NPs were engineered by coating the FCD-NPs with photothermal agent polydopamine (PDA). Upon intratumoral administration and near-infrared (NIR) laser irradiation, the positively charged FCD-NPs were exposed to capture the TAAs produced after photothermal therapy (PTT) and form in situ nanovaccines (FCD-NPs@TAAs) (Fig. 5F). In situ nanovaccines have demonstrated remarkable capabilities in stimulating adaptive immunity, transforming immunologically “cold” tumors into “hot” phenotypes while establishing long-term immune memory to prevent tumor metastasis. Yang et al. reported that IMQ-loaded polymeric micelles (IP-NPs) coated with PEI could efficiently capture TAAs generated during DOX-induced ICD. The formed in situ nanovaccines effectively promoted DCs maturation and T lymphocyte activation, inducing potent and durable immune memory [149]. Yu et al. developed Mal-modified Pluronic F127-chitosan nanoparticles coated with Astragalus polysaccharide (AMNPs) that captured cryoablation-released tumor antigens, targeted lymph nodes, enhanced lysosomal escape, and activated distant DCs, thereby generating robust systemic antitumor immunity (Fig. 5G) [150].

Fig. 5.

In vivo antigen-capturing nanovaccines. (A) Schematic illustration of AC-NPs binding tumor antigens released by radiotherapy and enhancing their presentation to DCs. (B) Quantitative analysis of protein binding capacity to AC-NPs. (C) Comparative profiling of proteins bound to AC-NPs with varying surface modifications [147]. (D) Zeta potential shift of FCD-NPs following TAA adsorption. (E) Quantification of TAA loading efficiency by FCD-NPs and their cytotoxicity toward DC2.4 cells. (F) Schematic diagram of FCDP-NPs assembling nanovaccines in vivo that reverse the immunosuppressive TME and potentiate PTT against tumor [148]. (G) Schematic diagram of the synthesis process of AMNPs and antigen grasping for in situ vaccination [150].

5. Nano-biomaterial-based delivery systems for cancer vaccines

Given the above advantages, nanotechnology is considered a promising alternative for improving the efficacy and safety of cancer vaccines. The nanoscale vectors have given rise to a new generation of vaccine formulations, including several clinically approved products and numerous candidates currently under investigation [151]. Nanocarrier systems for cancer vaccines can be categorized according to their distinct functional characteristics and compositional elements, each with its own set of strengths and limitations (Table 1). Recent significant advances and promising proof-of-concept applications of these platforms are further highlighted by selected preclinical studies (Table 2). The following sections detail these nanoparticle-based vaccine delivery systems.

Table 1.

Different types of nanocarriers and their main advantages and disadvantages.

Platform	Advantages	Disadvantages
Lipid-based nanoparticles	•Excellent delivery systems for nucleic acids •Effectively taken up by the cells and released into the cytoplasm •High biocompatibility •Adjustable composition, ratio, particle size, encapsulation rate, and surface modification •The mature preparation and production process	•Low-temperature storage (−80 °C) •Inflammatory responses and dose-dependent toxicity •Strong hepatotoxicity •Complex design for targeting extrhepatic tissues
Polymer-based nanoparticles	•A wide range of material options •Flexible design •Easy to chemical modification and functionalization •Controllable drug release •Good stability, facilitating preservation and transportation	•Poor biodegradability with long-term retention and toxicity •Lower efficiency of nucleic acid delivery and transfection •Difficulty in quality control
Nucleic acid nanoparticles	•Realize atomic-level controllable construction •Composed of natural nucleic acids, the degradation products are usually non-toxic •Predictable physical and chemical properties •Intelligent drug release or logical gate-controlled therapy •Multifunctional platform for integrating diagnosis and treatment	•Biological stability is poor: unmodified nucleic acids are easily degraded •Difficult to achieve large-scale production: low synthesis yield and difficulty in controlling uniformity •Difficult to achieve efficient cell uptake •Potential immunogenicity risk •Unestablished quality control standards
Inorganic nanoparticles	•Easy to synthesize: the size, shape and crystal structure can be precisely controlled •High load capacity •Stable, not easily affected by the biological environment •Unique characteristic: responsive to external stimuli (light, magnetic, US)	•Poor biocompatibility and biodegradability •Relatively complex chemical modification and functionalization •The problem of toxic material leakage
Carrier-free nanoparticles	•Extremely high drug loading and drug utilization efficiency •Fundamentally eliminating the issues related to biocompatibility, immunogenicity and long-term toxicity of the carrier •Relatively simple and environmentally friendly preparation process •Suitable for local administration	•Relying heavily on the molecular structure of drugs •Required examination of in vivo physical stability •Relatively complex chemical modification and functionalization
Biomimetic nanoparticles	•Outstanding biocompatibility and low immunogenicity •Long-term circulation and active homing ability •Automatically integrate multiple biological functions •Enhanced intracellular uptake and penetration ability of biological barriers •Applicable to personalized treatment	•Complex preparation and standardization challenges •Limited drug loading capacity and controllable release •Potential biosecurity risks •Difficult to store and transport •Inadequate production for clinical demand

Table 2.

Some of the significant preclinical studies of cancer nanovaccines. EE: encapsulation efficiency, TCPTP: T cell protein tyrosine phosphatase, MAFLD: metabolic dysfunction-associated fatty liver disease, HCC: hepatocellular carcinoma, CCR5: C-C motif chemokine receptor 5, CD: cluster of differentiation, 4-1BBL: 4-1BB ligand, PSMA: prostate-specific membrane antigen, PCa: prostate cancer, PSMA: prostate specific membrane antigen, GBM: glioblastoma, CPPs: cell-penetrating peptides, iNOS: inducible nitric oxide synthase, NHC: nanofiber-hydrogel composite, MITF: microphthalmia-associated transcription factor, PBA: phenylboronic acid, FSP1: ferroptosis suppressor protein 1, TAZ: transcriptional coactivator with PDZ-binding motif, LncRNA: long non-coding RNA, IRP: immune-related protein, MMP: matrix metalloproteinase, γ-PGA: poly(γ-glutamic acid), PL: polylysine, CLNs: central lymph nodes, ILNs: inguinal lymph nodes, PCD: protein condensate, aPD-L1: anti-PD-L1 antibodies, A2: angiopep-2, RIF: rifampicin, EGCG: epigallocatechin gallate, Pep: peptide, MLKL: mixed lineage kinase domain-like protein, TDR848: tetraphenylethylenyl diphenylamine R848, MB: molecular beacon, RNP: ribonucleoprotein, APE1: apurinic/apyrimidinic endonuclease 1, RNase: ribonuclease, AP: apurinic/apyrimidinic, PDAC: pancreatic ductal adenocarcinoma, OSCC: oral squamous cell carcinoma, POI: protein of interest, STAT3: signal transducer and activator of transcription 3, OG: origami, NZs: nanozigzags, GEM: gemcitabine, TDAs: tumor-derived antigens, R837: imiquimod, ZOL: zoledronic acid, ENR: epigenetic nano-regulator, TNBC: triple-negative breast cancer, ES: elesclomol, IRG1: immune-responsive gene 1.

Platform	Antigen/Adjuvant	Tumor model	Size (nm)	EE	Surface functionalization	Stimuli-responsiveness/Targeted	Immunomodulatory actions and antitumor effects	Ref
Lipid-based nanoparticles	TCPTP mRNA	MAFLD, HCC	∼100	-	Vitamin E	Liver	·Reprogrammed the hepatic microenvironment ·Promoted antigen presentation, the infiltration of CCR5 and CD8+ T cells, and specific antitumor immunity	[236]
	OVA mRNA	B16F10-OVA	185	∼85%	-	Spleen	·Induced APC maturation and activation, triggering antigen-specific immune response ·Promoted the secretion of endogenous cytokines, further enhancing T cell activation and cytotoxic activity ·Potent tumor growth inhibition	[224]
	IL-21, IL-7, and 4-1BBL mRNA	MC38, E0771, B16F10	79.97 ± 4.62	98.50 ± 0.45%	-	-	·Resensitization to anti-PD1 therapy in both treated and distal tumors ·Alteration of immune cell composition and activation in the TME with induction of tumor-specific CD8+ T cells and protective immunological memory	[237]
	PSMA mRNA	PCa	∼100	98.15 ± 0.56%	DSPE-PEG-VHH	PSMA+ cancer cells	·Enhanced and specific uptake, and mRNA transfection ·Increased LNP accumulation, but not functional mRNA delivery	[238]
	Engineered IL-12 mRNA	GBM	120∼180	83.2 ± 2.5%	SR-57227, CPPs (Tat)	Brain	·Upregulation of M1 phenotype markers, including CD80, CD86, and iNOS ·Augmented immune activation ·Suppressed tumor growth and improved overall survival	[239]
	CD19/CD20 CAR mRNA	Nalm6	<200	-	anti-CD5/CD8	Spleen (specific T cell subsets)	·Naïve B cell repopulation following B cell depletion, indicative of an immune reset	[240]
	Tumor antigens (C10/D6/F5) mRNA, NHC	MC38-OVA B16-OVA	208.9 ± 33.9, 133.0 ± 2.3, 148.6 ± 1.3	-	-	A local immunostimulatory niche	·Generated immunostimulatory microenvironment ·Robust antigen presentation and activation of multiple immune cell subsets ·Antigen-specific responses ·Robust Th1, Th2, and Th17 responses for bolstering antitumor efficacy	[241]
	siRNA targeting MITF	B16F10	80.3 ± 5.2	Greater than 90%	PBA	Cancer cells	·Superior tumor targeting capabilities and effectively silences genes ·Anti-melanoma and anti-metastasis efficacy	[242]
	FSP1 siRNA	Colon cancer	145.43 ± 2.55	90.59 ± 2.15%	-	Passively target tumor tissues	·Ferroptosis-driven oxidative damage to promote apoptosis, tumor antigen release, and DC maturation ·Passive tumor tissue targeting to enhance systemic antitumor immunity ·Suppression of primary and distant tumor progression with reduced side effects	[243]
	YAP and TAZ siRNAs	PCa	125.2 ± 3.2	89.30 ± 3.16%	CD44-specific targeting peptide A6	CD44 specific tumor	·Improved delivery to breast cancer cells, silencing and increased cell death ·Reduced proliferation in prostate cancer organoids and tumor growth in a patient-derived xenograft model	[244]
	Circular RNA	GBM	-	-	LncRNA (H19-IRP)	Clinical GBM samples	·Strong CTL response against GBM with associated growth inhibition	[245]
Polymer-based nanoparticles	PVP@Pt nanozyme, OVA257–264 peptides	Melanoma	∼50	-	DSPE-PEG	Acidic LN microenvironments	·Efficient LN targeting, robust CD8+ T cell activation, and significant tumor inhibition in both prophylactic and therapeutic melanoma models ·Excellent biocompatibility and minimal systemic toxicity	[246]
	OVA, immune adjuvant, STING agonist	B16-OVA, EG7-OVA	186 ± 2	85.1%	NH-Boc group	LN	·Simultaneous STING pathway activation and TLR 4-mediated DC maturation triggered by primary amine ·Durable tumor growth suppression accompanied by long-term protective immune memory ·Prolonged mice survival rate combination with aPD-L1	[247]
	Ruthenium complex NAMI-A	Subcutaneous breast cancer	132	23.26%	NAMI-A	-	·Reduction of MMP2 and MMP9 expression in cancer cells ·Decreased invasive and migratory capacity of cancer cells ·Eradication of cancer cells	[248]
	OVA-manganese oxide, γ-PGA, ε-PL	Melanoma	∼160	MnOx ∼4.9%, OVA ∼20.2%	-	-	·Enhanced cytosolic antigen release and DC maturation ·Reprogramming of the immunosuppressive TME ·Robust immune response against B16-OVA melanoma, preventing postsurgical tumor recurrence	[249]
	The SAP sequence, peptide-CpG ODN	GBM	40-48	CpG 95%	PEG-P(TMC-DTC)-spermine, SAP-PEG-P(TMC-DTC)	CLNs, GBM, spleen, ILNs	·Eliciting robust local and systemic immune responses combined with anti-CTLA-4 antibody therapy ·Significantly enhanced T cell and humoral immune responses leading to prolonged survival and 43% complete tumor regression	[250]
	OVA, IMDQ	B16-OVA, MC38	∼20	OVA 81%	-	Cathepsin B responsive, LN drainage	·Enhanced antigen presentation ·Induced 100% robust CD8+ T cell responses and superior tumor clearance	[40]
	OVA, PCD	Melanoma	∼150	∼80%	-	LN	·Induced robust antigen-specific CTL responses and humoral immunity, demonstrating potent antitumor efficacy ·Activating the cGAS-STING pathway and promoting DC maturation	[251]
	aPD-L1, paclitaxel	GBM	58.8	82.31 ± 4.30% for antibody, 87.63 ± 2.18% for paclitaxel	A2 peptide	Redox-responsive	·Enhanced ICB efficacy ·Facilitating efficient antibody delivery ·Reprogramming local immunity to empower immunotherapy against GBM	[252]
	RIF, EGCG	CRC (CT26-Luc)	107.2 ± 9.2	14.4% for RIF and 96.0% for EGCG	FadA-Pep, PEG shell	Fn biofilms targeting, pH-responsive release	·Eliciting systemic immune responses characterized by M1 macrophage polarization ·Suppressed recruitment of MDSCs ·Enhanced T-cell infiltration, potentiating antitumor efficacy	[253]
	DNA hairpins (Hi and HCpG)	Melanoma model	195	67.0% for Hi and 92.6% for HCpG	-	Acidic lysosomal environment	·Robust innate immune responses ·Promoted DC activation, enhanced CD8+ T cell proliferation, and increased the secretion of pro-inflammatory cytokines ·Improved cancer immunotherapy outcomes	[254]
	CpG and loxoribine	Melanoma model (B16)	∼186.7	Loxoribine loading capacity was 0.5%	3-MAPBA	pH responsive	·Conversion of M2 type macrophages to the M1 phenotype ·Promoted DC maturation and activated T cell ·Reprogramming the immunosuppressive TME ·Enhanced immune response, robust CD8+ T cell response and efficient tumor regression	[255]
	OVA mRNA, MLKL mRNA	Pancreatic tumor	180∼400	∼100%	POctS	Pancreatic cancer tissue	·Activation of innate immune responses at designated sites ·Simultaneous prevention and treatment of tumor	[256]
	TLR agonist TDR848	MB49 cancer	123.7 ± 35.7	85.42%	DSPE-PEG2000, fluoroethyl-phenylboronic ester	Inflammatory TME-responsive	·Induction of ICD and enhanced TAA delivery, leading to DC maturation, T cell activation, and immunogenic TME reprogramming ·Activation of adaptive immune responses against tumor	[257]
Nucleic acid nanoparticles	CpG ODN, NK cell-derived exosomes	Melanoma	46.1 ± 17.0	-	-	Tumor-validating environment	·Activated APCs to enhance the immunotherapy ·Cellular debris functioning as an immune antigen to amplify immunotherapeutic efficacy ·Achieved a remarkable tumor suppression rate of 91.2%	[258]
	MB and Cas9 RNP	Pancreatic cancer	342	Cas9 ∼100%	-	APE1 and RNase H	·Achieved effective tumor site accumulation, cellular uptake, and specific tumor detection ·Reduced PLK1 expression, induced cancer cell apoptosis, and inhibited the tumor growth	[259]
	Cas9 RNP, hemin	PDAC	384.3	Cas9 ∼100%	-	-	·Efficient gene editing and robust downregulation of Nrf2 ·Remarkable antitumor efficacy	[260]
	Tumor antigens, CpG	Melanoma and lymphoma models	35.0 × 22.5 × 27.0	-	PEGylated oligolysine	Draining LNs from the injection site	·Enhanced DC activation, antigen cross-presentation, CD8+ T cell activation, Th1-polarized CD4 activation and NK cell activation ·Improved cancer immunotherapy inducing long-term T cell memory and synergistic activity with anti-PD-L1	[261]
	OVA257-264SIINFEKL, CpG	B16-OVA	132.6 ± 71.7	-	PNA-derivatized LLP2A ligand	Immune and melanoma cells	·Boosting antigen presentation by DCs and eliciting strong CD8+ T cell and NK cell responses ·Significant tumor regression and prolonged survival	[262]
	CpG ODN	OSCC	23.78 ± 0.80	-	-	APCs	·Effective activation of APCs leading to an elevated proportion of M1-like macrophages and mature DCs ·Heightened production of inflammatory cytokines ·Facilitation of T cell phenotypic differentiation, leading to significantly increased CTL infiltration in lymph nodes and tumor tissues ·Executing a potent antitumor effect and inhibiting the progression of tumors	[263]
	PROTAC	MCF-7	-	-	Two STAT3 recognition modules	Cancer cell membranes	·Improving the tumor specificity, degradation efficiency, and overall antitumor effectiveness of PROTACs	[264]
	Antigen-coded mRNA, CpG ODNs	Melanoma	254.58 ± 25.90	-	Acidic-responsive DNA sequence, DCs targeting aptamer	Responding to the acidic lysosomal environment	·Promoted lysosomal escape, TLR activation ·Eliciting potent antigen-specific antitumor immunity ·Inducing strong effector and memory T cell responses ·Rejection of progressive tumors	[265]
	Peptide, RNA-OG	Melanoma	∼60	13∼200 peptide	-	DCs	·The same vaccine format with a low number of peptides enhanced CD8+ T cell responses without provoking T cell exhaustion ·Strong protective antitumor immunity	[266]
Inorganic nanoparticles	Silica NZs	Melanoma	245	-	-	BMDCs	·Generation of mature mBMDCs ·Priming of antigen-specific CTLs toward PD-1lowCD44high memory phenotypes ·Inhibition of tumor progression	[267]
	OVA, chitosan	Melanoma	238 ± 14	OVA 41.1%	Mannose and phenylboronate ester	APCs	·Activated the cellular immune response ·Inducing systemic immunity to eliminate cancer cells ·Inhibited the growth of B16-OVA melanoma tumors ·Extended the survival time	[268]
	Cell lysates, CpG ODN	GEM-resistant 4T1 cells	∼200	-	-	Draining LN	·Enhanced maturation of DCs ·Delayed progression of therapy-resistant tumors ·Increased T lymphocyte infiltration in conjunction with ICB therapy, amplifying antitumor immunity and efficacy against chemotherapy-resistant tumors	[269]
	Ultrasmall PEG-coated fluorescent silica nanoparticle	B16-GMCSF, B16F10	-	-	-	-	·Extension of survival through combined inhibition of IL-6 and PD-L1 ·Activation of the stimulator and reprogramming of the TME towards a pro-inflammatory phenotype ·Induction of significant cytotoxic and antitumor inflammatory responses ·Significant reductions in cell populations and receptors driving suppressive activities	[270]
	A manganese dioxide shell	4T1	∼59	-	-	High levels of intracellular GSH	·Modulated the unfavorable TME to restore CTLs infiltration and function and efficiently induced ICD ·The Mn2+-mediated stimulator of the interferon genes pathway ·Elicited long-lasting antitumor immunity to considerably inhibit primary, recurrent, and metastatic tumors	[271]
	Gold nanoparticles	B16-OVA, A549	2.3 ± 0.3	97.5% or 91.7%	Mitochondria membrane-binding and LC3-targeting peptides	Mitochondria	·Decreased mitochondria in cancer cells ·Enhancing the specific killing by CD8+ T cells and increasing the immunogenicity of cancer cells to overcome immune tolerance	[272]
	Whole TDAs, gold nanoxanthium (AuNX)	Melanoma and breast cancer	∼12.5	-	PDA	DCs	·Absorption of released TDAs and delivery to the DCs for cross-presentation ·Activating CD8+ T cells for efficient tumor-specific immunity ·Inhibited primary and distant tumors by producing robust antitumor immune responses	[273]
	Gold nanoparticles	MCF7, DLD-1	∼30	-	high-affinity α/β-peptide analogs	SM3 antibody	·Induction of immunogenic and cross-reactive antibodies against the natural antigen ·Elicited cellular immunity	[274]
	Eutectic gallium-indium nanoparticles	Hepa1-6 tumor	155.15 ± 1.05	-	A targeting CFR peptide	Target protein αvβ6 on the cancer cells membrane	·Resolution of the trade-off between deep tumor penetration and prolonged intratumoral retention ·Merging nanotechnology and immunology to overcome TME barriers ·Resolving the critical challenges of targeted delivery, sustained mitochondrial stress, and immune cell engagement	[275]
	Ag2S QDs with OVA and R837	4T1 breast tumor	13.5	R837 53.7%	-	-	·Activated DCs ·Improved long-term survival rates ·Reduced tumor size ·Elicited robust immune responses	[276]
	Ca(CO3)x(PO4)2(1−x)/3, IL-2	Melanoma	4-132	100%	-	-	·Reprogramming of the immunosuppressive TME ·Selective activation of cytotoxic and memory T lymphocytes ·Enhanced tumor-suppressive effects that inhibit local tumor recurrence ·Suppression of growth in distant, untreated tumors ·Maintenance of long-term T cell immunity against rechallenged metastatic tumors	[277]
Carrier-free nanoparticles	CIRTAs and R848	4T1	84.64 ± 1.44	R848 45.20% ± 7.14%	-	-	·Promoting DC maturation ·Significant tumor suppression ·Increased the infiltration of T cells into the tumor while decreasing the proportion of Tregs and modulating the TME ·Long-term immune memory	[178]
	DOX, adjuvant MPLA	B16 melanoma model	163.2 ± 1.9	DOX 11.96 ± 0.001%, MPLA 18.76 ± 0.002%	Tumor-homing peptide iRGD, MMP2 enzyme responsive peptide	TME-responsive properties	·Inducing ICD and simultaneously triggering tumor specific immune response ·Amplified the antitumor immunity, significantly inhibiting tumor growth and metastasis ·Demonstrating remarkable suppression of tumor growth, metastasis inhibition and recurrence prevention	[278]
	CyI and celastrol	4T1	76.8 ± 2.2	47.98 ± 1.8% and 10.01 ± 0.18%	-	Tumor	·Long-lasting antitumor immunity that inhibits metastatic spread and eliminates distant tumor ·Promising capacity to prime robust innate and adaptive immune responses	[279]
	OVA, InP-derived Fe-Ace	E.G7-OVA lymphoma cells	40∼80	OVA about 50%	-	DCs in LN targeting	·Uptaken by the same APC to effectively induce their activation ·Triggering tumor antigen-specific antitumor immune responses and inhibiting the tumor growth at distant site	[280]
	Aluminum (Al3+), and hydrophilic drug ZOL	4T1	274	ICG 55.85%, ZOL 5.93%	-	Tumor	·Enabling the release of TAAs ·Induction of TAM polarization ·Reversing the immunosuppressive TME ·Triggered antitumor immune response effectively suppresses tumor growth	[281]
	F-PEI, CpG	4T1	177.27 ± 1.23	CpG 98.4%	-	DCs in LN	·Enhanced cellular uptake by DCs and activated DCs ·Promoted cross-presentation of antigens ·Induction of macrophage repolarization	[282]
Biomimetic nanoparticles	Pten and STING mRNAs, Rab27a siRNA	YUMM1.7 cells, MC38	∼100	-	-	Tumor	·Knockdown of Rab27a ·Decreased levels of sEV with improved nanoparticle accumulation in tumor tissue	[283]
	Mannan and PEI	CT26	230 ± 10.3	∼93% mannan and ∼7% PEI by weight	Saccharomyces cerevisiae mannan	Dectin-2 and TLR 4 on DCs	·Induction of cytokine and chemokine release ·Recruitment and activation of antitumor CD8+ T cells, NK cells and M1-like macrophages within the TME	[284]
	Apoptotic cancer cell membranes, MnO2 NPs, cGAMP, Ce6	B16F10	164.5 ± 2.1	69.3% Ce6, 58.3% cGAMP	-	pH/ROS-responsiveness	·Efficient tumor prevention and direct tumor treatment through comprehensive immunomodulation	[285]
	OVA, DMXAA	EG7-OVA	200∼300	OVA 95.6%	-	LN	·Delivered the OVA protein to DCs and successfully activated the STING pathway	[286]
	ALDH1A1-overexpressing tumor-cell-derived nanovesicle, CSC-specific antigen and TAA, siYthdf1-based ENR	Breast and melanoma cancers	∼155	∼80%	An aptamer targeting the DC-specific intercellular adhesion molecule-3	Grabbing the nonintegrin (DC-SIGN) receptor	·Suppression of lysosomal protease activity ·Improved cross-presentation of CSAs/TAAs through minimized degradation and promoted cytosolic antigen export ·Potent activation of T cell immune responses ·Suppression of postoperative tumor recurrence and metastasis, leading to prolonged survival	[287]
	PTX-STeS-PTX (PSTeSP), R848	B16F10	142.2	PSTeSP 99.99%, R848 11.60%	PEGylated, cancer cell membranes	Tumor	·Enhanced homologous targeting specificity toward cancer cells ·Promoting cellular uptake and selectively inducing cytotoxicity in cancer cells ·Increased immune complexity at tumor sites and reshaping of the TME ·Suppression of cancer cell metastasis	[288]
	Plasmids encoding the ECD of HER2	4T1	77.83	-	aPD-L1 scFv (C5H9V2)	Tumor	·Blockade of the PD-1/PD-L1 signaling pathway ·Preventing the premature exhaustion of CAR-T cells ·Innate immune activation by OMVs leading to immunosuppressive TME remodeling and enhanced CAR-T cell infiltration and effector function ·Addressing the resistance associated with tumor antigen heterogeneity	[289]
	Atox1 siRNA and copper ionophores ES	TNBC, CRC	212.35 ± 23	siAtox1 70%, ES 12.3%	Akkermansia-derived OMVs	The elevated copper levels in tumors	·Enhanced copper accumulation ·Stimulated the tumor immune microenvironment ·Promoting T cell infiltration and specific tumor killing	[290]
	OVA, plasmid, OMVs	Pa02-OVA, B16-OVA	∼30	OVA 91.7 ± 1.6%	PD-L1 antibodies	Tumor	·Dual capacity for secondary immune signal modulation and PD-L1/PD-1 interaction blockade ·Enhanced binding of CD80 on DCs to CD28 on T cells ·Improved antigen presentation efficiency and T cell activation ·Enhanced antitumor effects ·Inhibition of cancer cell immune escape during the immune-mediated killing stage	[291]

5.1. Lipid-based nanoparticles

Lipids are integral components of cell membranes and biological frameworks, characterized by their phospholipid bilayer organization, which makes them highly suitable for the manufacture of advanced cancer nanovaccines [152,153]. Lipid-based nanocarriers exhibit amphiphilic properties, allowing hydrophilic antigens and adjuvants to be encapsulated within their aqueous core, while hydrophobic components can be embedded within the lipid bilayer. Incorporating lipids into nanovaccine design leverages their exceptional biocompatibility, significant biodegradability, ability to mimic natural cellular structures, and potential for tailored modifications to elicit precise immunogenic responses. By tuning the composition, key properties such as particle size and surface charge of lipid-based nanomaterials can be customized, directly influencing biodistribution, cellular uptake, and interactions with immune cells. Ionizable lipids remain neutrally charged under physiological pH but become protonated in acidic endosomal environments. This pH-responsive charge conversion facilitates endosomal escape of encapsulated cargo, a feature that has contributed to the clinical success of LNPs in mRNA vaccines.

Surface engineering of lipid-based nanovaccines can be achieved through conjugation of targeting ligands (antibodies, peptides, or aptamers) to phospholipids, enabling specific recognition of receptors overexpressed on immune cells or cancer cells. In addition, certain phospholipids, such as phosphatidylserine, can act as “eat-me” signals when incorporated into nanovaccines, promoting immune cell uptake and macrophage-mediated antigen processing. Yuan et al. successfully prepared liposome nanoparticles (Lipo-NPs) with varying elasticities and demonstrated that soft Lipo-NPs exhibited enhanced membrane affinity, while Lipo-NPs with moderate elasticity facilitated cargo delivery to macrophages through membrane fusion (Fig. 6A–C) [154]. Importantly, the soft Lipo-NPs showed superior in vivo transport efficiency to tumor-draining lymph nodes, which was attributed to their enhanced deformability and lower elastic modulus.

Fig. 6.

Nano-biomaterial-based delivery systems for cancer vaccines. (A) Flow cytometry analysis of RAW 264.7 macrophages and B16F10-EGFP cancer cells co-treated with DiI-labeled Lipo-NPs, showing enhanced macrophage uptake of soft Lipo-NPs compared to hard Lipo-NPs. (B) Diagram of macrophage membrane interactions with three distinct Lipo-NPs. (C) Fluorescence images of tumor-draining lymph node sections 24 h after intratumoral injection of each Lipo-NPs [154]. (D) Schematic diagram of the spatiotemporal release of CpG and antigens in PPE [159]. (E) Schematic diagram of RNA-OG-peptide nanovaccines assembly. (F) Atomic force microscopy (AFM) imaging of RNA-OG-peptide [168]. (G) Fabrication process and therapeutic mechanism of MSNs-ABC@PDA-OVA nanovaccines, demonstrating synergistic PTT and immunotherapy against B16-OVA melanoma tumors [135]. (H) Design and mechanism of CIRTAs@R848 for suppressing drug-resistant tumor growth [178]. (I) Schematic diagram of the ClyA-Catcher OMVs system for antigen display [185].

The rapid clinical deployment of LNPs, particularly for mRNA vaccines, has unveiled a nuanced safety and biodistribution profile that must be considered for their adaptation in cancer vaccines, which may require repeated administration. Two key considerations are dose-dependent inflammatory responses and pronounced hepatosplenic tropism. The underlying mechanisms are multifactorial, involving immune activation by both lipid components and the delivered nucleic acids. Perhaps more relevant for vaccine efficacy is the biodistribution profile: conventional LNPs exhibit significant hepatic accumulation after intravenous administration, mediated by apolipoprotein E adsorption. While this has enabled revolutionary liver-targeted therapies, for a cancer vaccine aiming to prime immune responses in lymphoid tissues, this represents off-target delivery. Although overt, clinically significant hepatotoxicity has been rare in the context of prophylactic vaccination, the sequestration of vaccine cargo in the liver could theoretically reduce the dose available for immune cell uptake in lymph nodes. This off-target accumulation warrants particular attention for cancer patients who may have compromised liver function due to underlying conditions or concurrent hepatotoxic therapies. Consequently, strategies to re-engineer LNPs for lymphoid system targeting are a major focus of cancer nanovaccine development.

5.2. Polymer-based nanoparticles

Polymer-based nanoparticles have become ideal for the delivery of immune adjuvants, proteins, and peptides due to their excellent biodegradability, minimal immunogenicity, superior biocompatibility, and configurable release kinetics. In our previous work, we constructed polymersomes and hybrid nanoparticles based on poly(ε-caprolactone) (PCL)-PEG-PCL triblock copolymers to exquisitely load model antigen OVA and TLR agonists in spatially separated compartment, which could effectively activate DCs through spatiotemporal co-delivery, stimulate strong CD8⁺ CTLs response and immune memory, and prolong survival in treated mice [[155], [156], [157], [158]].

PLGA nanoparticles are the most widely used polymeric carriers for single-dose nanovaccines, which can release their contents within days, weeks, or even months for extending the duration of immune stimulation and enhancing vaccine efficacy. Yan et al. used PLGA nanoparticle-stabilized pickering emulsions (PPE) to improve the release kinetics for CpG and antigens (Fig. 6D) [159]. CpG inserted at the oil-water interface exhibited a rapid release profile, while antigens encapsulated within PLGA nanoparticles displayed sustained release. The resulting distinct intracellular release kinetics (rapid initial burst followed by prolonged release) triggered immediate and potent activation of TLR 9 while maintained long-term antigen exposure. Conjugation of targeting ligands on the surface of PLGA nanoparticles facilitates specific recognition of immune cells or cancer cells, and the addition of PEG to PLGA nanoparticles further increases their circulation half-life [160].

Chitosan is a natural cationic polysaccharide and has gained significant attention as an effective antigen delivery vehicle. Chitosan-based nanoparticles offer multiple advantages for vaccine delivery, including: enhanced epithelial penetration of antigenic macromolecules, protection of antigens from premature degradation, prolonged antigen residence time, and improved antigen presentation [161]. The positively charged surface of chitosan nanoparticles facilitates strong electrostatic interactions with negatively charged cell membranes, while its intrinsic immunostimulatory properties can further enhance vaccine immunogenicity. These unique characteristics make chitosan nanoparticles a promising platform for developing effective mucosal and systemic vaccines.

Compared to LNPs, polymeric systems generally exhibit a more favorable preliminary toxicity profile with less reported hepatotoxicity, making them attractive for certain applications. However, their transfection efficiency for nucleic acid vaccines typically lags behind LNPs. The critical development path lies in hybridizing polymer durability with efficient cytosolic delivery mechanisms, perhaps through fusion with membrane-destabilizing peptides or lipids, positioning them as ideal carriers for subunit vaccines and sustained-release adjuvants.

5.3. Nucleic acid nanoparticles

Over the past few decades, nucleic acid nanotechnology has emerged as a promising platform for drug delivery systems, and diverse two- and three-dimensional DNA/RNA nanostructures with various geometries have been developed, such as DNA origami, single-stranded tile, DNA nanostructures, and single-stranded DNA/RNA origami [162]. Nucleic acid-based nanostructures offer great opportunities for the rational design and delivery of immunomodulators due to their homogeneous geometry, feasible modifications, programmability, surface addressability, and intrinsic biocompatibility [[163], [164], [165], [166]]. The unique molecular recognition properties of nucleic acids enable these nanostructures to achieve cell-specific targeting through sequence-specific interactions with immune cells or cancer cells, thereby enhancing the precision of therapeutic delivery. What's more, nucleic acid nanostructures can be designed to be located in intracellular compartments and release cargo based on biochemical markers such as pH, redox potential, and enzymes in the diseased microenvironment [167].

However, conventional DNA origami nanostructures are typically assembled by folding long scaffold DNA strands with hundreds of surplus short DNA strands, a process that is not only relatively costly but also requires multiple tedious purification steps that significantly reduce the final product yield [166]. In particular, it remains uncertain whether nucleic acids trigger an immune response in vivo, and the characteristics of such responses are still largely unexplored [165]. Yip et al. presented a nanovaccine platform based on self-assembled RNA origami (RNA-OG) nanostructures, which function as potent TLR 3 agonist [168]. In addition, due to its robust synthesis and versatile modification capabilities, RNA-OG could be easily linked to a wide variety of peptides to construct RNA-OG-peptide nanovaccines with uniform size, precise peptide loading, and high stability (Fig. 6E and F). The assembled RNA-OG-peptide nanovaccines induced DC maturation, reduced immunosuppression, and elicited tumor-specific CD8⁺ T cell responses, thereby improving the survival rate of tumor-bearing mice.

5.4. Inorganic nanoparticles

Inorganic biomaterials include a range of metals, metal oxides, and inorganic salts with unique physicochemical and biological properties that make them highly suitable for constructing advanced nanovaccine platforms. Metal-based nanomaterials can enhance cancer nanovaccine development through enabling efficient antigen/adjuvant delivery, targeted localization, and immune stimulation via ligand targeting, surface modification, and other strategies. MSNs have recently attracted attention due to their unique properties, such as large surface area, high porosity, good biocompatibility, easy surface functionalization, making them suitable for delivering diverse drugs, antibodies, genes, proteins, and peptides [169]. Biomolecules or supramolecules are conjugated onto the outer surface of MSNs as stimuli-responsive gatekeepers, enabling on-demand cargo release from the mesopores exclusively under specific stimuli [170].

Huang et al. designed MSNs-based cancer nanovaccines (MSNs-ABC@PDA-OVA) by loading ammonium bicarbonate (ABC) within the mesopores, where PDA served as both a pH-responsive gatekeeper to prevent premature ABC release and a functional platform for antigen conjugation (Fig. 6G). The abundant amino groups on the PDA surface enabled efficient covalent loading of OVA through sulfhydryl-reactive chemistry [135]. Following cellular uptake, the acidic environment triggered ABC decomposition, generating CO₂/NH₃ that physically disrupted lysosomal membranes. This promoted efficient OVA translocation to the cytoplasm, significantly enhancing MHC I antigen presentation and subsequent CTL activation. In vivo evaluation revealed the MSNs-ABC@PDA-OVA nanovaccines exhibited remarkable tumor growth inhibition, durable protection against tumor recurrence, and significant suppression of metastatic progression.

5.5. Carrier-free nanoparticles

Traditional drug delivery systems often suffer from inefficient drug loading and instability due to poor compatibility between drugs and carrier materials, leading to premature drug leakage during storage and in vivo circulation [171]. Moreover, excessive use of carrier materials may cause vehicle-related toxicity concerns [172]. Although some novel nanodrug delivery systems have progressed, their structures and preparation techniques may be too cumbersome for large-scale manufacturing and clinical translation [173]. To solve these challenges, carrier-free nanomedicines have emerged as a groundbreaking alternative. Carrier-free nanomedicines are made from the self-assembly or co-assembly of pure drug molecules through covalent and/or non-covalent bonds, including hydrophobic interactions, electrostatic interactions, and hydrogen bonds, without any carrier materials involved [174]. All components in carrier-free nanomedicines play a positive role, often exhibiting high drug loading capacity, enhanced biological stability, better therapeutic outcomes, and reduced side effects [175]. In particular, carrier-free nanomedicines offer advantages in quality control and industrialized mass production, and their simple preparation techniques pave the way for clinical translation [176].

Zeng et al. proposed a carrier-free nanovaccine strategy by self-assembly of fluorinated CpG (F-CpG) with melanoma neoantigen peptide (Obsl1) [177]. The nanovaccine without additional carrier materials exhibited an exceptionally high antigen loading capacity (83.19 wt%), increasing the uptake of Obsl1 by APCs. As a result, the F-CpG/Obsl1 nanovaccine activated the innate and adaptive immune responses for the long-term preventive and therapeutic efficacy against B16F10 melanoma. Pan et al. reported a therapeutic nanovaccines (CIRTAs@R848) composed solely of chemotherapy-induced resistant tumor antigens (CIRTAs) and TLR 7/8 agonist R848 (Fig. 6H) [178]. CIRTAs@R848 required no additional vectors, featured a simple production process, and demonstrated significant tumor suppression.

5.6. Biomimetic nanoparticles

Biomimetic delivery platforms exhibit excellent biocompatibility and enhanced stealth properties, enabling them to evade recognition by the reticuloendothelial system and prolong their circulation time. These features collectively improve both safety and efficacy profiles, demonstrating strong potential for clinical translation [179]. EVs derived from various immune cells (such as APCs, T cells, B cells, NK cells, and neutrophils) stand out as particularly effective natural nanocarriers. These EVs inherently contain a complex mixture of bioactive lipids, proteins and nucleic acids that facilitate intercellular communication and immune regulation, allowing them to remodel the TME and enhance antitumor immunity through both cellular and humoral immune responses. Significant progress has also been made in the development of biomimetic cell membrane-coated nanoparticles. By preserving native surface proteins and functions from immune cells, these platforms achieve extended circulation times while enabling controlled release of therapeutic payloads [180]. Their unique composition grants them an unparalleled ability to activate comprehensive immune responses, bridging both innate and adaptive immunity [181]. In recent years, cancer cell membrane-based vaccines have gained attention for their intrinsic homologous targeting capability and ability to present antigens to prime tumor-specific immune responses [182]. The formidable challenge for EVs is manufacturing scalability and standardization. Issues include low yield, heterogeneous vesicle populations, and difficulties in loading exogenous cargo with high efficiency. Unlike the well-defined chemistry of LNPs, EV-based products face significant hurdles in characterization, potency assays, and regulatory approval. The future likely lies in applying LNP-like engineering principles to EV manufacturing (e.g., bottom-up assembly of synthetic vesicles with selected natural membrane components) or decorating synthetic nanoparticles with precise, isolated bionic elements (e.g., specific exosomal surface proteins) to create “hybrid biomimetics” that balance nature's sophistication with engineering control. Semi-biological nanocarriers are composed of biomaterials and synthetic materials, which combine and offer engineering advantages, including low toxicity, high biocompatibility, scalable and reproducible manufacturing, while maintaining the immunological benefits of biological components [183].

Bacterial OMVs serve as potent immunomodulators by initiating pro-inflammatory cascades and regulating immune system functions. Their dual capacity as both antigen carriers and natural adjuvants makes OMVs particularly valuable for vaccine development and immunotherapeutic applications [184]. Cheng et al. designed a multifunctional OMV-based vaccine platform that first displayed tumor antigens on the surface of OMVs by fusing with ClyA protein, and then simplified the antigen presentation process by using a plug-and-play system consisting of tag/capture protein pairs (Fig. 6I) [185]. OMVs decorated with different protein traps could simultaneously display multiple different tumor antigens to elicit a robust and synergistic antitumor immune response. Bacterial membranes are also one of the common biomaterials used for antitumor immunotherapy, owing to the various immunostimulatory components on the bacterial surface that can induce immune activation or promote tumor-specific enrichment [186]. Bacterial membranes-derived nanoparticles possess inherent advantages for targeted drug delivery, including their nanoscale dimensions, structural stability, and excellent biocompatibility in vivo [187].

VLPs mimic the size, shape, and molecular organization of complete viruses, and are composed of structural viral proteins such as capsid or envelope proteins. VLPs are highly immunogenic, non-infectious, and non-replicating. VLPs can directly induce B cell activation and proliferation and participate in related antibody subclass transformation, recombination, and somatic hypermutation. Pathogen-mimicking nanoparticles have become the frontier of vaccine delivery technology, showing potent immune activation and favorable biocompatibility. Notably, while VLP-based designs are generally considered safer than traditional vaccine approaches, careful consideration must still be given to potential safety implications, even for attenuated or inactivated formulations.

6. Administration routes

While the selection and rational design of nanocarriers are critical for enhancing antigen presentation and immune activation, the choice of macroscopic administration routes is equally pivotal to the ultimate therapeutic success of these engineered nanoparticles. “Delivery technology” encompasses not only the nanoscale carriers but also the macroscale strategies that ensure these carriers effectively reach and stimulate the immune system. The administration routes determine the physiological environment in which the vaccine first interacts, the types of local immune cells involved, and the overall efficiency of systemic distribution. Besides the conventional injection administration, transdermal and mucosal immunization have both demonstrated good therapeutic efficacy [188,189]. Upon delivery to the lymph nodes, injected vaccines can elicit robust cellular and humoral immune responses, effectively activating CTLs and stimulating antibody production. In contrast, transdermal administration primarily interacts with APCs in the skin, such as Langerhans cells in the epidermis and DCs in the dermis, thereby promoting stronger cellular immunity and efficiently activating skin-resident memory T cells. Mucosal vaccination primarily targets mucosa-associated lymphoid tissue, the first line of defense against infection at entry portals. It mainly induces both local and systemic humoral immunity, characterized by the production of secretory mucosal immunoglobulin A (S-IgA) and serum immunoglobulin G (IgG) antibodies, while also eliciting cellular immune responses at mucosal sites [190]. However, systemic cellular immunity induced via mucosal vaccination is generally weaker than that achieved through injection. Although different administration routes tend to elicit distinct immune response patterns, the final outcome is fundamentally determined by the antigen, adjuvant, and delivery platform. By selecting or engineering an administration route that is physiologically compatible and functionally tailored to the specific tumor vaccine formulation, the appropriate immune-stimulating signals can be delivered to the precise anatomical and cellular niches at the optimal time. Such strategic targeting not only enhances the bioavailability and lymphoid trafficking of vaccine components, but also actively steers the immune response toward a desired functional phenotype, thereby enabling precise and sustained reprogramming of the antitumor immune landscape.

6.1. Injected vaccines

Traditionally, vaccines are mainly administered via local injection, including intramuscular injection, subcutaneous injection or intradermal injection. After such vaccination, soluble antigens often diffuse and degrade rapidly at the injection site, resulting in poor targeting to lymph nodes and inefficient uptake by immune cells. In the context of deep-seated tumors, hematological malignancies, or metastatic tumors, intravenous administration enables systemic delivery to target and eliminate disseminated cancer cells. The primary challenge in systemic administration is the low proportion of active components that successfully reach the target site, coupled with the risk of systemic toxicity. To address this, nanovaccines can be functionally modified to specifically recognize and bind to APCs in the spleen, lymph nodes, or tumor. The specific modification strategies and detailed methodologies are elaborated in chapter 4.3.

Injectable hydrogels, which can be administered via subcutaneous or intratumoral injection to form in situ drug reservoirs, represent an emerging delivery system that has demonstrated considerable advantages for cancer vaccine applications. First, hydrogels can co-load multiple antigens and immune adjuvants while effectively protecting labile biomolecules (such as mRNA and DNA) from degradation, thereby prolonging the retention time of antigens. Their porous structure enables controlled and sustained payload release to the local tissues, effectively extending immune response duration [191]. Second, hydrogels confine drugs to the injection site, reducing the side effects caused by systemic dispersion, such as chemotherapy toxicity and cytokine storms [192]. Hydrogels incorporated immune checkpoint inhibitors can effectively reverse the immunosuppressive state of tumors and increase the efficacy of vaccines [193]. Third, hydrogels can be designed as intelligent drug repositories to respond to environmental stimuli (e.g., pH, enzymes, heat) for on-demand release of immunotherapeutic agents within the TME [[194], [195], [196], [197]]. Importantly, natural and select synthetic hydrogels are biocompatible and undergo safe metabolic clearance. Cheng et al. proposed a thermo-responsive hydrogel (NvIH) co-encapsulating ICB antibodies and nanovaccines loaded with three immune-stimulating agonists for TLR 7/8/9 and STING pathways (Fig. 7A–D). [198]. Following in situ vaccination, the rapid sol-gel transition significantly prolonged tumor retention time, enabled sustained release of immunotherapeutic payloads, and reduced acute systemic inflammation. Remarkably, the NvIH system induced significant regression of both local and distant poorly immunogenic tumors, including challenging orthotopic glioblastoma models.

Fig. 7.

Representative vehicles for cancer vaccine applications. (A) Schematic diagram of the antitumor immune mechanism of the hydrogel NvIH loaded with triple immunostimulants and anti-PD-1 antibody. (B) Hydrodynamic diameter and TEM image characterization. (C) Gelation properties evaluated by vial inversion method. (D) Injectable property demonstration through syringe administration of NvIH [198]. (E) Fabrication process and structural characterization of cryoMNs [203]. (F) Schematic showing IL-12 mRNA loading into exosome or liposomes for pulmonary immunization in lung cancer models using nebulization administration. (G) Size distribution analysis and (H) TEM micrographs of IL-12 mRNA loaded exosome or liposomes. (I) Representative immunofluorescence images of tumor-bearing mouse lungs [208]. (J) Schematic illustration of orally administered Ce6/R837@Lp₁₂₇NP for CRC immunotherapy [210].

6.2. Transdermal vaccines

Transdermal vaccines specifically denote the administration of tumor vaccines through the skin, as opposed to via subcutaneous or intramuscular injection. Microneedles (MNs) are a novel transdermal delivery platform for cancer vaccines that can painlessly penetrate the outer physical barrier of the skin (stratum corneum) and directly deliver vaccines to the immune cell-rich dermis, while avoiding the drawbacks of traditional injections [199]. The MN system prevents vaccine diffusion loss and eliminates first-pass metabolism effects, while demonstrating excellent sustained release properties, ease of administration, high bioavailability, and improved patient compliance. Soluble MNs can be functionally optimized, such as by using cross-linked materials to prolong the release of antigens and adjuvants, or by using environmentally responsive materials or nanoparticles to precisely control the release of vaccines [200]. More sophisticated designs utilizing double or triple-layer MNs facilitate implementation of combination therapies through spatiotemporal and sequential drug administration strategies [201]. Yang et al. developed antitumor vaccine MNs (TCV-DMNs) to co-deliver GM-CSF and the autophagy promoter (Tat-beclin 1) [202]. Following transdermal vaccination, GM-CSF acts as an adjuvant to recruit DCs and enhance their antigen uptake capacity. Concurrently, Tat-beclin 1 promoted DC maturation through autophagy upregulation. The study demonstrated that vaccination with TCV-DMNs could effectively prevent and treat malignant tumors, achieving an impressive extension of relapse-free survival exceeding 40 days.

MNs enhance the stability of loaded bioactive components by shielding them from direct exposure to external moisture, oxygen, or enzymes. For instance, certain matrix materials (e.g., hyaluronic acid) form hydrophilic networks that stabilize protein structures and prevent denaturation, while hydrophobic materials (e.g., PLGA) protect small molecule drugs from hydrolysis. However, conventional MNs are unsuitable for fragile therapeutics like antibodies and living cells. Chang et al. employed a cryomicroneedles (cryoMNs) to co-deliver OVA-pulsed DCs and anti-PD-1 antibody for combination immunotherapy, demonstrating the great potential of cryoMNs as an integrated delivery platform of the therapeutic cells and biomacromolecules (Fig. 7E) [203,204].

6.3. Mucosal vaccines

Mucosal vaccines represent a unique class of immunizations administered directly at mucosal surfaces, capable of inducing robust local and systemic immune responses [205]. When antigens are presented to mucosal tissues, they trigger a comprehensive immune activation characterized by the production of antigen-specific antibodies in both secretory fluids and circulation, and the generation of diverse T cell populations in mucosal-associated lymphoid tissues and systemic compartments. This immunological responsiveness stems from the mucosal immune system's natural ability to constantly monitor and respond to environmental antigens, supported by its extensive network of immune cells including APCs, T lymphocytes, and other immune cells distributed throughout mucosal surfaces [206]. Mucosal vaccination offers several distinct advantages, including significantly reduced production and distribution costs compared to injectable vaccines, painless administration that improves patient acceptance, and particular suitability for mass vaccination during public health emergencies. However, despite these benefits, the field faces notable challenges, as evidenced by the limited number of approved mucosal vaccines (currently only nine, comprising eight oral formulations and one nasal vaccine). The relatively slow progress in mucosal vaccine development can be attributed to several key obstacles: the brief retention time of conventional vaccine formulations at mucosal sites, the induction of immune tolerance at mucosal surfaces, and various physical barriers that hinder effective antigen uptake and presentation. These challenges highlight the need for innovative formulation strategies to overcome the unique hurdles posed by mucosal immunization. With the innovation of responsive materials, carrier systems and preparations, activating mucosal immunity for cancer immunotherapy is an emerging research direction. This approach primarily targets malignancies associated with mucosal tissues, particularly NSCLC, head and neck squamous cell carcinoma, CRC, gastric cancer, bladder cancer, and cervical cancer, along with pulmonary metastases from breast cancer.

With the highest mortality rate among all cancers globally [1], lung cancer presents an urgent need for more effective treatment strategies. Inhalation delivery has emerged as the optimal administration route for lung cancer, enabling simultaneous induction of robust humoral immunity, potent cellular immunity, and comprehensive mucosal immunity, which is the most significant advantage compared with traditional injectable vaccines [207]. Professor Ke Cheng developed an inhalable exosome loaded with IL-12 mRNA that promoted IFN-γ-mediated immune activation, systemic immunity, and immune memory in lung cancer-bearing mouse models, successfully suppressing lung cancer and preventing recurrence (Fig. 7F–I) [208]. Due to their appropriate aerodynamic size, the exosomes can accurately reach the alveoli and deposit, thereby achieving the sustained drug release and prolonging therapeutic duration.

CRC remains the second leading cause of cancer-related deaths worldwide [209]. Although oral or rectal immunization has been proposed as the preferred strategy to induce mucosal immune responses in the intestine, poor mucosal penetration and the instability of the antigen and carrier system in the intestinal environment remain significant challenges to overcome. Zu et al. developed an oral vaccine (Ce6/R837@Lp₁₂₇NP) for CRC treatment featuring a protective coating of plant-derived lipids and pluronic F127 (Fig. 7J) [210]. This formulation maintains stability during gastrointestinal transit while enabling effective colonic mucosal infiltration and tumor-specific penetration. Moreover, the development of mucosal vaccines capable of inducing robust and systemic antitumor immunity at intestinal surfaces represents a crucial advancement in CRC immunotherapy. Li et al. successfully developed a new mRNA vaccine to stimulate mucosal immune responses in the gut by co-delivering all-trans-retinoic acid (ATRA) and mRNA using LNP [211]. ATRA-adjuvanted mRNA-LNP significantly improved tumor inhibition and prolonged animal survival compared to conventional mRNA-LNP without ATRA.

7. Next-generation nanovaccine development directions

To more effectively address the complex tumor microenvironment and overcome the limitations of current immunotherapies, next-generation nanovaccines are evolving along two core directions: intelligent responsiveness and combinatorial synergy. On the one hand, stimuli-responsive nanovaccines are designed to leverage endogenous or exogenous triggers to achieve spatiotemporally precise release and delivery of vaccine components. On the other hand, combinatorial therapeutic strategies integrate nanovaccines with other modalities, such as ICB, radiotherapy, or chemotherapy, to systematically enhance antitumor immune responses and remodel the immunosuppressive TME. The following sections elaborate on these two advancing frontiers in nanovaccine development.

7.1. Stimuli-responsive nanovaccines

Leveraging the unique features of the TME, such as acidic pH, elevated enzymatic activity, and high levels of reduced GSH, stimuli-responsive nanocarriers enable precise spatiotemporal release of vaccine components, thereby minimizing off-target toxicity. Among these, pH-sensitive nanoparticles take advantage of the acidic conditions commonly found in tumor tissues to enhance targeting specificity. Chen et al. developed a STING nanovaccine responsive to both acidic pH and NAD(P)H quinone oxidoreductase 1 for the treatment of HPV-induced cancers [212]. This nanovaccine consists of di-amidobenzimidazoles conjugated to a STING-activating polymer PSC7A through an azobenzene linker as an adjuvant, and co-assembled with the HPV16 E7 protein as an antigen. Following intravenous administration, nanovaccines measuring approximately 25-30 nm in diameter efficiently co-deliver the payload to the myeloid cell populations in the secondary lymphoid organs and tumors, boosting immune activation while reducing the systemic toxicity associated with conventional STING agonist delivery. On the other hand, redox-sensitive nanocarriers release their cargo in response to the highly reductive conditions of the TME, often achieved through the incorporation of cleavable bonds such as disulfide linkages. Xu et al. covalently attached the TLR 7/8 agonist R848 via a disulfide bond to fifth-generation polyamidoamine (G5-PAMAM) dendrimer, forming the carrier G5-R848, which was subsequently complexed with the model antigen OVA to yield the nanovaccine RINV [213]. Upon intracellular reduction, R848 is released, promoting cytosolic antigen delivery and cross-presentation, leading to significant prophylactic and therapeutic efficacy in a B16F10-OVA melanoma model. Similarly, Xing et al. engineered a dual-responsive nanovaccine (DRNV) through the self-assembly of antigen with an amphiphilic polymer, polyethylene glycol modified poly (lipoic acid) (PTA) [214]. The TLR 7/8 agonist IMDQ was conjugated to PTA via an enzyme-cleavable linker, preventing rapid systemic clearance and potential toxicity. In the presence of intracellular GSH, the PTA core rapidly degrades, resulting in DRNV disintegration and enzyme-triggered release of IMDQ, ultimately activating endosomal TLR 7/8 and eliciting robust antitumor immunity capable of inhibiting tumor growth and establishing long-term immune memory.

Furthermore, exogenous physical stimuli such as light and US have been integrated into responsive nanovaccine platforms. Light-responsive systems often combine photothermal or photodynamic therapy to induce ICD, release tumor antigens, and remodel the immunosuppressive TME. Jiang et al. constructed a nanoprobe composed of lanthanide nanoparticles, M2-like macrophage-targeting peptide coupled with OVA, and a photo-controlled cysteine protease inhibitor, which precisely modulates antigen cross-presentation in TAMs and enhances immunotherapy against glioblastoma [215]. US-responsive nanovaccines offer distinct advantages in deep-tissue penetration and non-invasiveness. Cheng et al. developed a sequentially targeted sonodynamic nanovaccine (Stars NV) capable of crossing the blood-brain barrier, homing to TAMs, and ablating tumors and stimulating systemic antitumor immunity upon US activation, significantly extending survival in a murine intracranial glioblastoma model [216]. To further improve responsiveness and therapeutic outcomes, multi-stimuli systems are also being explored. Yang's team designed a carbon dot-based nanovaccine (Cu-N-CDs@OVA) responsive to both light and US. Under laser irradiation, it mediates photothermal ablation of primary tumors and antigen release, while US promotes nanoparticle penetration and diffusion into deep lesions, synergistically activating DCs and inducing systemic antitumor immunity [217]. Stimuli-responsive nanovaccines, whether activated by endogenous (pH, enzymes, redox) or exogenous (light, US) triggers, offer a powerful strategy for the precise control of vaccine release and enhanced delivery, paving the way for next-generation cancer immunotherapy.

7.2. Combination with other therapies

The efficacy of nanovaccine monotherapy is often limited by tumor heterogeneity, poor T cell infiltration hindered by the tumor stroma, and insufficient T cell responses. Therefore, combining nanovaccines with other treatment modalities has become a crucial research direction, leveraging synergistic mechanisms to enhance therapeutic outcomes. In particular, the combination of nanovaccines with ICB has shown significant promise in multiple studies. Immune checkpoint inhibitors, including those targeting cytotoxic T-lymphocyte antigen 4 (CTLA-4), PD-1, and PD-L1, are widely used in clinical practice. Currently, immune checkpoint inhibitors have become the standard of care for advanced cancers such as melanoma. However, drug resistance and recurrence remain significant challenges. As reported at the 2023 American Association for Cancer Research (AACR) Annual Meeting, the Moderna and Merck mRNA cancer vaccine (mRNA-4157/V940) combined with pembrolizumab reduced the risk of melanoma recurrence and death by 44% in a phase IIb clinical trial. In July 2023, mRNA-4157/V940 entered phase III clinical trials, and the final results will be published in 2029, just one step away from the market. To further investigate the mechanisms of combination therapy, Carreira et al. developed a mannose-grafted PLGA nanovaccine loaded with the melanoma neoantigen for targeted delivery to APCs [218]. This study further integrated three distinct PD-L1 modulation strategies: a monoclonal antibody, a small-molecule inhibitor SM56, and a small interfering RNA (siRNA), enabling a systematic comparison of their characteristics. While monoclonal antibodies are effective, they face limitations such as high cost and limited tissue penetration. SM56 demonstrated better bioavailability and tumor penetration. In contrast, siRNA can specifically silence PD-L1 in DCs, helping to elucidate its cell-specific role in immunosuppression.

The combination of nanovaccines with ICD-inducing modalities such as radiotherapy and chemotherapy is also becoming an emerging trend. Utilizing ICD inducers to enhance vaccine immunogenicity helps activate systemic antitumor immune responses. Furthermore, radiotherapy upregulates MHC I expression on cancer cells, thereby enhancing antigen presentation and promoting CD8⁺ T cell-mediated recognition and killing. To optimize therapeutic efficacy and minimize toxicity, radiotherapy should be administered locally with appropriate dosing schedules, while chemotherapy can be integrated into targeted delivery systems to improve tumor specificity and reduce systemic immunosuppression. Under controlled ICD induction, even normal tissue cells may trigger antitumor immune activation, contributing to enhanced cancer immune surveillance. Meng et al. proposed the concept of an “inflamed-cell-as-vaccine” and designed tumor-targeting biomimetic nanoparticles (rVAR2-M-NPs) loaded with digoxin-OVA (Dig-OVA) complexes [219]. These nanoparticles induce mild ICD stress, promote cancer cell autoinflammation, downregulate hypoxia-inducible factor-1α (HIF-1α), enhance radiosensitivity, and avoid excessive ROS generation. In both in vitro and in vivo studies, cancer cells treated with rVAR2-M-NPs adopted an immunogenic “inflamed” phenotype. When combined with γ-ray irradiation, this strategy significantly suppressed the growth and metastasis of orthotopic 4T1 tumors, prolonged survival, and exhibited no notable adverse effects.

8. Challenges to the clinical translation of nanovaccines

Advancing tumor vaccines into clinical practice is fundamentally about translating scientific discoveries into global health solutions, ultimately to address the growing challenge of rising cancer incidence and mortality worldwide. Personalized neoantigen and DC vaccines have shown promising efficacy in early-stage clinical trials for several cancers, including melanoma and colon cancer. The success of the COVID-19 mRNA-LNP vaccine has provided a measure of confidence for the nanoparticle vaccine, demonstrating the potential for clinical translation and market viability. However, it represents but one milestone, and the broader path is still lined with considerable obstacles.

8.1. Clinical safety issues

Biological safety is the primary concern during the clinical translation. Most clinical trials of nanovaccines have reported common grade 1-2 adverse reactions, while a few have reported individual cases of grade 3 or even grade 4 adverse reactions [[220], [221], [222]]. Unlike conventional small-molecule drugs, nanocarrier materials have the potential to elicit immunotoxicity or inflammatory responses due to their unique properties. For instance, the immunogenicity and toxicity of the LNP system have not been resolved. Issues such as the disruption of endosomal and cell membranes by ionizable cationic lipids, the blood clearance and allergic reactions caused by repeated injections of PEGylated LNPs, as well as the immune rejection reactions resulting from increased dosage or repeated administration, all constitute major obstacles for clinical translation. A 2020 phase I clinical trial of MRX34 was closed early due to serious immune-mediated adverse events that resulted in four patient deaths [223]. The inherent liver accumulation of LNP carriers represents a critical translational challenge, particularly for extrahepatic oncology applications. Significant off-target uptake by hepatic cells not only diminishes therapeutic efficacy through reduced antigen presentation but also raises clinically relevant safety concerns regarding hepatotoxicity. Apart from LNPs, other delivery vehicles also have their own safety issues: (1) incomplete understanding of the metabolic pathways and long-term biodistribution of exogenous materials like polymers and inorganic nanomaterials, (2) the potential risk of genomic integration associated with gene editing or nucleic acid therapies, and (3) the inherent immunogenicity of biologically derived nanovaccines, which may provoke severe immune reactions such as cytokine storms.

The successful clinical translation of nanovaccines hinges on resolving critical safety issues. A key strategy to mitigate the toxicity of LNP carriers involves simplifying or modifying their composition, such as reducing the standard four-component formulation to three- or two-component systems. Demonstrating this approach, Yin et al. recently developed a dimethylamino-lipidoid-based OncoLRC formulation through systematic screening and optimization. Compared to conventional LNPs, OncoLRC not only achieved almost exclusive targeted mRNA delivery to the spleen but also maintained high efficiency at a markedly reduced lipid-to-mRNA weight ratio of 1.5:1, contrasting with the typical 10:1 ratio in standard formulations [224]. Further biodistribution control can be attained by modifying nanoparticle composition to reduce hepatic uptake or by surface-functionalization with targeting ligands. In parallel, the risk of off-target effects and dose-related adverse reactions can be reduced by optimizing the administration route and implementing precise dosage control. In addition, as a novel therapeutic platform, it is particularly important to continuously gather long-term follow-up data on potential unknown risks during clinical trials.

8.2. Clinical effectiveness issues

Statistical analysis of global clinical trials for tumor nanovaccines registered on ClinicalTrials.gov indicates notable advancements in the field, with nearly two hundred related clinical trials completed to date. A selection of recent and active studies is provided in Table 3. Nevertheless, a sharp decline in success rates occurred in phases II and III, despite a promising 94% success rate in phase I trials. This failure stemmed primarily from insufficient clinical efficacy rather than immunogenicity or toxicity, halting the progression of many candidates. The therapeutic efficacy of nanovaccines is primarily constrained by three factors: the immunosuppressive TME, inadequate antigen screening, and poor scalability.

Table 3.

Clinical trials of cancer nanovaccines in the past three years. NCT: National Clinical Trial, LA cSCC: locally resectable advanced cutaneous squamous cell carcinoma, KRAS: Kirsten Rat Sarcoma, NRAS: Neuroblastoma RAS viral oncogene homolog, ESCC: esophageal squamous cell carcinoma, EC: endometrial cancer, ODG: oligodendroglioma, IDH: isocitrate dehydrogenase, EOC: epithelial ovarian cancer, CEA: carcinoembryonic antigen, CNS: central nervous system.

Delivery system	NCT number	Active molecules	Diseases	Combination therapy	Study phase	Status	First posted
LNPs	NCT07245901	circFAM53B-219aa mRNA	Advanced solid tumors	Toripalimab	I/II	Not yet recruiting	2025-11-24
	NCT07100210	IL-22 binding protein mRNA	Refractory malignant solid tumors	-	I	Recruiting	2025-08-03
	NCT07053072	PD-1 mRNA	Primary Hepatocellular Carcinoma	-	I/II	Recruiting	2025-07-08
	NCT06689540	MTS105	Advanced Hepatocellular Carcinoma	-	I	Recruiting	2024-11-14
	NCT06623422	Up to 34 personalized neoantigens mRNA	NSCLC	Pembrolizumab	III	Recruiting	2024-10-02
	NCT06389591	pp65 mRNA	Recurrent GBM	-	I	Recruiting	2024-04-29
	NCT06305767	Up to 34 personalized neoantigens mRNA	Bladder Cancer	Pembrolizumab	I/II	Recruiting	2024-03-12
	NCT06295809	Individualized neoantigen mRNA	LA cSCC	Pembrolizumab	II/III	Active, not recruiting	2024-03-06
	NCT06077760	Up to 34 personalized neoantigens mRNA	NSCLC	Pembrolizumab	III	Recruiting	2023-10-11
	NCT05933577	Up to 34 personalized neoantigens mRNA	Melanoma	Pembrolizumab	III	Active, not recruiting	2023-07-06
	NCT05726864	Amph-Peptides 7P	KRAS, NRAS, mutated PDAC, solid tumors	Amph-CpG-7909	I/II	Active, not recruiting	2023-02-14
DC vaccine	NCT07288112	Tumor lysate and mRNA	Refractory melanoma	pIFN	I/II	Not yet recruiting	2025-12-17
	NCT06805305	DOC1021	GBM	Tumor resection, Temozolomide, pIFN	II	Recruiting	2025-02-03
	NCT06751849	Personalized tumor neoantigen	Advanced NSCLC	Radiotherapy, PD-1 inhibitor	II	Recruiting	2024-12-30
	NCT06749925	Hybrid DCs	GBM	Pembrolizumab	III	Not yet recruiting	2024-12-27
	NCT06739226	GM-CSF and B7-2 (CD86)	Melanoma	CAR-T cell therapy, CTL therapy	I/II	Recruiting	2024-12-18
	NCT06675201	Neoantigen	ESCC	Standard ICIs	II	Recruiting	2024-11-05
	NCT06253494	AdHER2DC	EC	Pembrolizumab, N-803	I/II	Recruiting	2024-02-12
	NCT06254326	Total tumor RNA-pulsed DCs	Recurrent and progressive ODG	IDH1/2 inhibitors	I	Recruiting	2024-02-12
	NCT05773859	Tumor-lysate loaded XP-DC	EOC	Standard-of-care chemotherapy, surgery	I/II	Recruiting	2023-03-17
	NCT05767684	CEA	Refractory tumor and solid tumor	Lenvatinib, Nivolumab	I	Recruiting	2023-03-14
	NCT05765084	Tumor antigen WT1	Malignant pleural mesothelioma	Chemotherapy	I/II	Recruiting	2023-03-13
Whole cancer cell	NCT07199413	SV-BR-1-GM	Solid tumor patients with CNS metastases	Cyclophosphamide, Pembrolizumab	I	Not yet recruiting	2025-09-30
	NCT06366490	Innocell	Recurrent EOC	-	I	Not yet recruiting	2024-04-16

The most common barrier to success in clinical trials is the limited efficacy of vaccines as monotherapy, a consequence of the immunosuppressive TME in which even activated T cells may be excluded from the tumor or become functionally exhausted by suppressive cells and signals. Regarding antigen selection, both TAAs and personalized neoantigens are associated with their own limitations. The former may compromise effectiveness due to tumor heterogeneity, while the latter faces challenges in prediction accuracy and manufacturing complexity. Furthermore, the increasing need for individualized manufacturing and the complexity of production processes have significantly hampered the clinical scalability. Various factors, including patient demographics, geographic location, environmental conditions, and clinical implementation protocols, can all influence vaccine responsiveness. For instance, many patients in the advanced stage or those who have undergone other treatments have exhausted their immune systems and are unable to respond to immunotherapy. Additionally, in some underdeveloped regions, there are limitations in cold chain transportation and the capabilities of medical staff, making it impossible to obtain the actual data on the effectiveness of the drugs.

Multiple strategies have been attempted to address the aforementioned issues. Combination therapy has been shown to enhance intratumoral immunogenicity, alleviate the immunosuppressive TME, thereby improving patient immunotherapy response rates. The specific mechanism has been confirmed in numerous preclinical studies. In clinical trials, tumor vaccines are frequently administered in combination with immune checkpoint inhibitors or integrated into standard first-line chemotherapy or radiotherapy, demonstrating superior efficacy. To overcome issues such as neoantigen inactivation, more accurate and efficient algorithms are needed to screen sequencing data and predict neoantigens with optimal immunogenicity and clinical potential. Furthermore, to better evaluate clinical benefit, vaccine validation should prioritize settings such as postoperative adjuvant therapy or early-stage neoadjuvant therapy, where lower tumor burden and milder immunosuppression are more conducive to eliciting an effective immune response.

8.3. Other issues

As multi-component complex systems, tumor vaccines still face many issues that must be overcome to advance their clinical translation, such as manufacturing, cost-effectiveness, and regulatory issues. The complex composition and structure of nanovaccines lead to an ambiguous relationship between their physicochemical properties and biological effects. Utilizing AI or deep learning computational models to correlate material properties with in vivo performance can accelerate the screening of optimal formulations. Furthermore, by integrating genomic analysis and other techniques, the in vivo treatment process can be visualized as effectively as possible. To address the challenge of scaling up from laboratory preparation, the researchers aim to develop a minimalistic formulation that retains full functionality, with the goal of transitioning rapidly from the laboratory research to industrial production. This approach is designed to ensure process consistency throughout the expansion from preclinical to commercial manufacturing. To ensure consistency between batches in mass industrial production, it is essential to standardize the manufacturing process and establish strict controls over the basic key quality attributes, including particle size, encapsulation efficiency, zeta potential, and stability. To address high costs and improve accessibility, it is imperative to develop modular and automated rapid-production platforms, which are aimed at shortening production cycles, enhancing cost control, and tackling the complexities of personalized formulations. The clinical translation potential of vaccine risks being constrained by the lack of regulatory frameworks. Therefore, the primary challenge lies in fostering international coordination to unify regulatory standards and assessment methods, thereby ensuring a more transparent and clear approval pathway.

Finally, welfare ethics constitutes a paramount consideration that must be incorporated. Only by adhering to the ethical foundation can innovation truly and sustainably serve human health. While most preclinical studies rely on rodent models, the transition to larger mammals such as pigs, dogs, or non-human primates involves significantly higher costs. Furthermore, their immune pathways and key biological targets still differ from those in humans. Advances in alternative models, such as micro-physiological systems, now enable researchers to obtain robust preclinical data through various organoid chips while supporting the FDA's stated goal of reducing reliance on animal experiments. This approach addresses critical ethical and welfare issues and simultaneously offers a model system that more accurately reflects human biology. As an innovative therapy, tumor vaccines demand stricter ethical oversight in clinical trials, primarily due to the inherent complexity of underlying concepts like nanotechnology and immune mechanisms. This complexity may hinder meaningful comprehension among lay participants, thereby challenging the practical exercise of their right to informed consent. Meanwhile, due to the lack of long-term safety data, the ethics of clinical research is a dynamic process that requires the full participation of multiple parties to safeguard the dignity, rights and well-being of patients.

9. Conclusions and outlook

This review provides a comprehensive overview of recent advances in engineered delivery systems for cancer vaccines, covering fundamental components such as antigens and adjuvants as well as innovative platforms including nanoparticle formulations, hydrogels, MN patches, and inhalation devices. Overall, the review highlights the benefits and transformative potential of nanocarrier-based vaccine delivery systems, which are poised to revolutionize the field by enabling safer, more precise, and efficacious immunization strategies through the careful nanocarriers design and optimal delivery system engineering as well as the selection and co-delivery of antigens and adjuvants. Furthermore, we explore the major challenges faced by these innovative vaccine delivery systems in terms of clinical translation. In the future, the synergistic integration of expertise spanning nanomaterials science, vaccinology, oncology, immunology, biomedicine, engineering, and pharmaceuticals, coupled with the application of cutting-edge disruptive technologies such as AI, bioinformatics, and machine learning algorithms may be needed to optimize nanovaccine design for enhanced clinical translation and therapeutic efficacy. Modern vaccine design has evolved beyond traditional empirical “trial-and-error” approaches. On the contrary, with the advancement of new computational tools and “big data” analytics, a wealth of data can be provided to aid researchers with unprecedented insights into immunological mechanisms and vaccine efficacy, accelerating the discovery of novel effective combinations of antigens and adjuvants [225]. Next-generation sequencing and other omics technologies enable comprehensive genomic profiling and immunophenotyping to reveal the unique immune and molecular signatures of the heterogeneous TME, which can drive personalized nanovaccine development [226]. Concurrently, advances in machine learning and computing power have significantly enhanced the prediction, prioritization, and validation of neoantigens using bioinformatic tools such as pVACseq, NetMHCpan, and NeoFlow [227,228], while high-resolution imaging and advanced image processing techniques are beginning to uncover hidden spatial and temporal patterns in immune cell and neoantigen interactions within the TME [229]. Computational models of vaccine carrier systems enable a systematic precise mapping between nanocarrier design parameters and their corresponding antigen/adjuvant release kinetics. It would be helpful to explore the immune responses elicited by various administration routes (e.g., intramuscular, subcutaneous, inhaled, needle-free). Collectively, these sequencing technologies and computational algorithms can be applied at all stages of nanovaccine design with the aim of antigen selection, adjuvant optimization, delivery platform engineering, and administration route determination, thereby changing the inherent paradigm of vaccine development from empirically to rationally designed and improving both therapeutic efficacy and clinical translation success rates. Although mice remain the most widely used animal model, critical differences exist between murine and human immune systems [20]. Even conserved immune receptors often exhibit substantial interspecies variations in their cell-type-specific expression patterns. Besides, current preclinical studies predominantly rely on simple subcutaneous xenograft models due to their cost-effectiveness and technical convenience. However, more physiologically relevant models, including spontaneous tumor systems and patient-derived xenografts models, should be increasingly incorporated in future research, as they better recapitulate the complex immunosuppressive TME observed in human cancers [230]. It is worth noting that organoids that can closely mimic the structure and function of human tissues in vivo may also be a good choice for the development of nanovaccines with precision and enhanced patient-specific applicability [231].

The intrinsic properties of nanocarriers with design flexibility make nanoparticle-based cancer vaccines have the potential for improved stability, high loading capacity, precise targeting ability, enhanced bioavailability, and long-term therapeutic effects [232]. These properties have facilitated the development of in situ vaccination strategies that combine direct tumor-killing methods with immune-stimulating drugs for localized delivery, paving the way for personalized cancer immunotherapy [69]. In addition, considering the complexity of TME during tumorigenesis and development, nanovaccines are further combined with radiotherapy, chemotherapy, phototherapy, immune checkpoint inhibitors, and other conventional therapies to amplify antitumor immune response and generate sufficiently durable immune memory. Despite the promising potential of cancer nanovaccines demonstrated in numerous preclinical studies and their emergence as a breakthrough technology in vaccine development, several daunting challenges must be addressed before successful clinical translation can be achieved. The foremost challenge lies in overcoming the safety concerns surrounding nanomaterial-based formulations, which currently represent a major barrier to clinical application. A primary issue is the nonspecific accumulation of nanomedicines in healthy organs, particularly the liver and spleen. Furthermore, the toxicity assessment protocols for nanovaccines remain underdeveloped compared to those for conventional pharmaceuticals, with standardized biosafety evaluation criteria for novel nanomaterial platforms yet to be fully established. Comprehensive safety profiling must extend beyond acute toxicity to include rigorous long-term chronic toxicity evaluations. Notably, careful consideration must be given to the possible formation of a protein corona on the surface of nanoparticles, which critically influences biocompatibility, functional performance, and in vivo biodistribution [233]. Consequently, the development of effective strategies to modulate protein corona formation are necessary in nanovaccine design. The manufacturing process of nanovaccines presents another significant hurdle that requires careful optimization. Streamlined synthesis protocols must be elaborately developed to ensure scalability, batch-to-batch consistency, product stability, and cost-effectiveness, while simultaneously minimizing the use of potentially harmful organic solvents. The integration of automated systems and robotic technologies into production workflows could substantially enhance manufacturing precision and efficiency, thereby facilitating the clinical translation of nanovaccines. Finally, the incomplete understanding of the complex immune networks involved in tumorigenesis and progression hinders the widespread application of nanovaccines in cancer immunotherapy. The innate and adaptive immunity are intricate, and current limitations in comprehending how modulation of specific immune components affects the overall network functionality continue to generate debate within the scientific community. This knowledge gap significantly impedes the rational design and optimization of nanovaccine-based immunotherapies [234,235].

Looking ahead, the future development of cancer vaccines is rapidly advancing toward greater intelligence, systematization, and ecological integration. First, computationally driven rational vaccine design platforms will deeply integrate multi-omics data, patient-specific immune profiles, and AI predictive models. This integration will enable full-process digital simulation including antigen selection, adjuvant matching, and delivery system optimization. It will significantly shorten development timelines and improve success rates. Second, the development of responsive smart vaccine systems has become a key focus. Such systems can dynamically release antigens and adjuvants in response to microenvironmental signals such as pH, enzyme activity, or immune cell activation. They may even be programmed to fine-tune local and systemic immune balance. Third, dynamic monitoring and adaptive adjustment of vaccine efficacy will become feasible. Leveraging tools such as liquid biopsy and wearable sensors to provide real-time feedback on immune status will enable follow-up dosing strategies to be personalized, forming a closed-loop therapeutic system. Finally, the establishment of a global collaborative and data-sharing ecosystem is essential. Integrating clinical trial data, real-world evidence, and AI analytics on open science platforms will accelerate the discovery of universal principles. This will drive the iteration of next-generation universal and personalized vaccines.

CRediT authorship contribution statement

Chenlu Huang: Writing – original draft, Data curation. Hanyong Wang: Writing – original draft, Data curation. Shamei Luo: Data curation. Chenxi Yu: Data curation. Linhua Zhang: Writing – review & editing, Supervision, Conceptualization. Dunwan Zhu: Writing – review & editing, Supervision, Funding acquisition, Conceptualization.

Ethics approval and consent to participate

As a review article, there is no need to include an ethical report with the submission.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.