mRNA vaccines change personalized treatment for lung cancer

mRNA vaccines have recently emerged as an effective tool for preventing and treating various diseases. The use of vaccines in medicine has a long history, dating back to 1796 when Edward Jenner successfully vaccinated a boy against smallpox using the cowpox vaccine (21) (Figure 2). However, the discovery and application of mRNA are relatively recent, with only a few decades of history. mRNA was first identified in 1960; in the 1970s, researchers began synthesizing proteins using isolated mRNA; by 1985, mRNA was successfully synthesized in the laboratory.

A pivotal milestone in the development of mRNA vaccines occurred at the end of 1987 when Robert Malone discovered that human cells could absorb mRNA and produce the corresponding proteins following exposure to a mixture of mRNA strands and lipid droplets (22). This established a foundation for the subsequent research and development of mRNA vaccines. Research into this area has gradually increased over the years. In 1992, Jirikowski et al. successfully alleviated symptoms of diabetes insipidus in a murine model by administering mRNAs encoding oxytocin and vasopressin (23). The authors demonstrated the potential of mRNA in disease treatment. In the 21^st century, research on mRNA vaccines has made significant breakthroughs. In 2005, modified RNA was discovered to evade immune detection, providing novel insights for mRNA vaccine design. The translational potential of mRNA vaccine platforms was first demonstrated in clinical trials targeting rabies virus between 2010 and 2015, establishing foundational safety and immunogenicity profiles (24, 25). The unprecedented challenges posed by the COVID-19 pandemic catalyzed rapid technological advancements, with mRNA vaccines emerging as frontrunner candidates due to their modular design architecture, scalable production pipelines, and potent humoral/cellular immune induction (26). Global vaccine development efforts reached an unprecedented scale, with WHO tracking 337 candidate vaccines by 2022, including 47 mRNA-based formulations, of which 23 progressed to clinical evaluation (26). This accelerated trajectory validated mRNA platforms’ pandemic responsiveness and redefined vaccine development paradigms, positioning nucleic acid-based approaches as essential tools for addressing emerging infectious threats. BioNTech, Moderna, and CureVac, leveraging their profound technological expertise, constitute the core forces in the current pipeline landscape. These pipelines primarily follow two strategies: personalized neoantigen vaccines based on patient-specific mutations and “off-the-shelf” vaccines targeting shared tumor-associated antigens.

BioNTech has established the most extensive mRNA cancer vaccine pipeline, adopting a dual-track development strategy of both “personalized” and “off-the-shelf” approaches. Fixed-Antigen Vaccine BNT116 for Advanced NSCLC: BNT116 is a vaccine containing mRNAs encoding six common tumor-associated antigens in NSCLC (including CLDN6, PRAME, etc.). Its Phase I clinical data presented at the 2025 American Association for Cancer Research (AACR) Annual Meeting were encouraging. The study showed that BNT116 combined with the PD-1 inhibitor cemiplimab, used as first-line treatment for advanced NSCLC patients unsuitable for platinum-based chemotherapy, achieved an 80% disease control rate, a 45% objective response rate, and induced rapid ctDNA clearance. This indicates that mRNA vaccines can effectively break tumor immune tolerance and synergize with immune checkpoint inhibitors (27). Synergistic Layout of Personalized Neoantigen and Fixed-Antigen Vaccines: In addition to BNT116, its personalized neoantigen vaccine BNT122 (in collaboration with Roche) is advancing in late-stage clinical trials across multiple cancer types. Concurrently, the success of its fixed-antigen vaccine BNT111 in Phase II trials for melanoma further validates the breadth of its platform technology. The company also plans to initiate a global clinical trial in 2025 combining BNT327/PM8002 (a bispecific antibody) with an mRNA cancer vaccine, aiming to establish a next-generation backbone for lung cancer immunotherapy.

The collaboration between Moderna and Merck is a benchmark in the field of personalized mRNA cancer vaccines, with its core product, mRNA-4157 (V940), achieving breakthrough results in clinical research. Breakthrough Efficacy of mRNA-4157: This vaccine identifies patient-specific tumor mutations through sequencing and synthesizes mRNA encoding up to 34 neoantigens. In a Phase III clinical trial for post-surgical adjuvant treatment of high-risk melanoma, combining mRNA-4157 with pembrolizumab (Keytruda) alone significantly reduced the risk of recurrence or death compared to pembrolizumab monotherapy. Based on this success, the partners have initiated a Phase III study in non-small cell lung cancer (NSCLC) to validate this strategy in lung cancer (28). Pipeline Expansion and Technological Iteration: Moderna’s other candidate, mRNA-4359, is designed to target the immunomodulatory enzyme indoleamine 2,3-dioxygenase 1 (IDO1), tumor-associated antigens (TAAs), and the immune checkpoint molecule programmed death-ligand 1 (PD-L1), aiming to overcome the immunosuppressive tumor microenvironment and demonstrating potential immunostimulatory and anti-tumor activity. It has shown promise in early-stage clinical trials for patients resistant to immunotherapy. Furthermore, Moderna has established a highly automated, dedicated manufacturing facility for its personalized cancer vaccine mRNA-4157, laying the groundwork for addressing the production bottlenecks of personalized vaccines (29).

Following an adjustment in its infectious disease vaccine strategy, CureVac is shifting its focus to oncology and attracting significant attention due to its integration with BioNTech. Next-Generation Lung Cancer Vaccine CVHNLC: Its “off-the-shelf” candidate vaccine CVHNLC for squamous non-small cell lung cancer (sqNSCLC) received clinical trial approval from the European Medicines Agency (EMA) in 2025, marking its official entry into clinical development for lung cancer treatment. Merger Synergy with BioNTech: In 2025, BioNTech announced the acquisition of CureVac. This move aims to integrate their respective expertise in mRNA design, delivery technologies, and manufacturing to accelerate the development of mRNA cancer immunotherapies. This industry consolidation is expected to significantly enhance BioNTech’s R&D capabilities in key areas such as lung cancer. Beyond pipeline expansion, mRNA technology itself is continuously evolving. A study published in August 2025 proposed a “minimalist mRNA” design strategy. The research posited that for small antigenic molecules expressed by antigen-presenting cells, the untranslated regions (UTRs) and non-coding regions traditionally considered essential in conventional mRNAs might be dispensable. This simplified mRNA not only maintains effective immunogenicity but also accelerates candidate antigen screening and vaccine production processes, offering a new technological pathway for the rapid manufacturing of personalized cancer vaccines (30).

As key intermediate products from transcription to translation, mRNA carries the genetic code that guides the synthesis of the corresponding proteins. mRNA vaccines, a significant subset of nucleic acid vaccines, can be categorized into two primary types based on their characteristics: self-amplifying RNA (saRNA) and non-replicating mRNA. saRNA viral vaccines leverage the efficient amplification characteristics of single-stranded RNA within host cells to generate vaccines containing genes for viral replication and therapeutic purposes. Based on the methods used to acquire antigen expression, saRNA vaccines can be classified into three types: those derived from DNA plasmids, those delivered via virus-like particles, and those utilizing in vitro transcribed and amplified RNA. Immunological studies conducted in animal models have demonstrated that saRNA vaccines elicit robust cellular and humoral immune responses (31). Traditional non-replicating mRNA vaccines possess a relatively straightforward architecture, comprising a cap structure, a 5′-untranslated region (UTR), an open reading frame (ORF) encoding the vaccine antigen, a 3′-UTR, and a poly(A) tail. In addition to the ORF, which directly encodes the antigen, the other structural components are essential for maintaining mRNA stability and augmenting transcriptional efficiency. These elements also serve as modification sites that can prolong the in vivo half-life of mRNA and effectively mitigate unwanted immune responses (32). Compared with saRNA, conventional non-replicating mRNA vaccines are distinguished by their compact size and relatively simple structure, which includes a single ORF for encoding vaccine antigens. This design strategy ensures the vaccine elicits an immune response against specific antigens, avoiding unnecessary immune responses (33).

This fundamental mechanistic difference directly leads to distinct characteristics in antigen expression and immune responses between the two. Non-replicating mRNA enables rapid initiation of high-level antigen expression, but its duration is relatively short, typically lasting several days to a week. This “high-intensity, short-pulse” pattern effectively activates both cellular immunity (CD8⁺ T cells) and humoral immunity. In contrast, while saRNA may have a lower initial antigen expression level, its self-amplifying capability allows for sustained antigen expression lasting weeks or even months. This “low-dose, long-term” antigen exposure pattern is considered more conducive to inducing strong and long-lasting T cell immunological memory, which is crucial for eliminating tumor cells and preventing cancer recurrence (33). However, advantages coexist with challenges. The larger molecular size of saRNA—typically two to three times that of non-replicating mRNA—places higher demands on its production and delivery. Its complex sequence poses challenges for the design of plasmid DNA (pDNA) templates and the stability of in vitro transcription, while also requiring higher efficiency from delivery systems. In contrast, non-replicating mRNA benefits from its smaller molecular weight and simpler structure, resulting in a more mature and stable production process, as well as higher efficiency in encapsulation by delivery systems such as lipid nanoparticles (LNPs) (34).

Overall, these two platforms each have distinct emphases in their application strategies for cancer vaccines. Non-replicating mRNA technology is well-established and capable of rapidly eliciting potent immune attacks, making it highly suitable for developing therapeutic vaccines targeting personalized neoantigens with the aim of quickly eliminating existing tumors. In contrast, saRNA demonstrates unique potential in adjuvant vaccine scenarios that require the establishment of long-term immune surveillance to prevent tumor recurrence, thanks to its persistent antigen expression characteristics. Currently, personalized neoantigen vaccines represented by BioNTech and Moderna predominantly employ the non-replicating mRNA platform and have advanced to late-stage clinical trials. Meanwhile, saRNA vaccines developed by several companies are largely in early-stage clinical exploration for infectious diseases and oncology, with their clinical potential awaiting further validation.

Over the past few decades, the development of mRNA vaccines has encountered several significant challenges, including the intrinsic instability of mRNA molecules, their strong immunogenicity, and the lack of effective mRNA delivery systems (11). With continuous advancements in science and technology, these difficulties are gradually being resolved. First, mRNA vaccines offer unique advantages, as they do not use replicating vectors and hence lack issues such as antibiotic resistance, genomic integration risks, or strong immunogenic reactions (35). This characteristic has enabled mRNA vaccines to safely induce antibody responses in phase I clinical trials (21).

Besides tackling the issues of mRNA instability and immunogenicity, modified mRNA can remarkably minimize adverse reactions and enhance therapeutic effectiveness. Current investigations are delving into diverse mRNA delivery approaches. These encompass lipid-based, polymer-based, and peptide-based delivery platforms, along with cationic nanoemulsions. These delivery systems not only improve the stability of mRNA but also enhance its cellular uptake and penetration, making mRNA vaccines more effective in overcoming the initial challenges (36). Notwithstanding the substantial headway in advancing mRNA vaccines, this domain remains in an early and evolving phase. Consequently, it is essential to remain alert and prioritize safety-related factors when furthering the mRNA vaccine. Nevertheless, with ongoing advancements and technological improvements, the future of mRNA vaccines appears promising.

Currently, various traditional vaccines are being developed for lung cancer treatment. These vaccines can be broadly categorized into antigen-specific vaccines, which include peptide/protein, DNA, and vector-based vaccines, and whole-cell vaccines, such as allogeneic and autologous DC vaccines (Figure 3).

mRNA vaccines have the traits of not integrating into the genome and being non-infectious and are accompanied by a transient cellular expression state where repeated administration can be carried out. The encoded sequences in mRNA transcripts show great versatility, allowing for the expression of antigens and molecules that can regulate the immune system to trigger and adjust immune responses of both adaptive immunity and innate immunity. Full-length antigens containing numerous epitopes can be presented to the immune system via MHC-I and MHC-II molecules. Innate immunity serves as the primary defense against nonself antigens (23). Notably, mRNA vaccines primarily activate pro-inflammatory signaling pathways through two sets of pattern-recognition-receptor-mediated mechanisms, which initiate the body’s inborn immune response (56, 57) (Figure 4).

This first system is the Toll-like receptor (TLR) system, situated within the plasma membrane, endosomes as well as lysosomes present in epithelial cells and immune cells, such as DCs, monocytes, and macrophages (58). Double-stranded RNA (dsRNA) activates TLR3, which induces the generation of type I interferons (IFNs) through the TRIF (TIR domain-containing adapter-inducing IFN-β) pathway (59). In contrast, single-stranded RNA (ssRNA) activates TLR7 in addition to TLR8 and signals through the MYD88-dependent signaling pathway, resulting in the production of cytokines with pro-inflammatory properties regulated by either nuclear factor (NF)-κB or interferon regulatory factor (IRF) 3. The second system involves the retinoic acid-inducible gene I (RIG-I)-like receptors, being cytosolic RNA sensors playing a vital role in innate antiviral immunity (60, 61).RIG-I and melanoma differentiation-related protein 5 (MDA5) exhibit distinct activation patterns in response to 5′-triphosphorylated dsRNAs of varying lengths: short (18–19 bp) dsRNAs activate RIG-I, while longer (>1,000 bp) dsRNAs activate MDA5 (57, 61). However, excessive activation of these receptors can accelerate vaccine clearance and reduce their protective effects (57, 60). The effectiveness of the activation of the innate immune system instigated by mRNA-based immunotherapies significantly hinges upon the vaccine substances and also upon the manufacturing procedures (62). This encompasses the selection of carriers and the routes of administration, and these aspects represent pivotal areas of current research for mRNA vaccines (63, 64). Following vaccination, antigen-presenting cells (APCs) uptake the mRNA and transport it to the cytoplasm for antigen processing (65). The processed TAAs are then presented on MHC class I and II molecules to activate CD8+ and CD4+ T cells (66, 67). Additionally, CD4+ T cells can co-stimulate antigen-specific B cells, inducing humoral immune responses (68, 69). Conversely, B cells, which also serve in the capacity of APCs, can trigger the activation of CD4+ T cells after engulfing extracellular polypeptides and presenting them on MHC class II molecules, further enhancing the immune response (70). The BNT116 vaccine, currently undergoing clinical trials (NCT05142189↗), is specifically designed for non-small cell lung cancer (NSCLC). It utilizes lipid nanoparticles (LNPs) to deliver mRNA encoding tumor-associated antigens. After the vaccine is taken up by antigen-presenting cells (such as dendritic cells), the mRNA is translated into complete tumor antigen proteins in the cytoplasm. These antigens are processed and presented via both MHC class I and II molecules, simultaneously activating CD8⁺ cytotoxic T cells and CD4⁺ helper T cells, thereby inducing specific killing of lung cancer cells expressing these antigens. This ability to concurrently stimulate both cellular and humoral immunity represents an advantage of mRNA vaccines over certain traditional platforms (27). KRAS G12V Mutant-Targeting Vaccine: In preclinical studies, an mRNA vaccine specifically designed against the KRAS G12V mutation has demonstrated potential. In lung cancer models, this vaccine elicited robust T-cell responses, notably generating high levels of tumor necrosis factor-alpha (TNF-α) and interferon-gamma (IFN-γ)—cytokines critical for attacking and eliminating tumor cells. This highlights the precision of mRNA vaccines in targeting specific driver gene mutations such as those found in lung cancer (71). mPLA/mRNA Vaccine: A 2023 study reported an mRNA vaccine incorporating monophosphoryl lipid A (mPLA, a TLR4 agonist). In lung cancer models, including advanced models with bone metastasis, this vaccine not only effectively activated T cells but also demonstrated the ability to reprogram the tumor microenvironment: it reversed pro-tumorigenic M2-type macrophages into anti-tumor M1-type macrophages and promoted the infiltration and activation of dendritic cells and natural killer (NK) cells at the tumor site, thereby effectively inhibiting lung cancer growth and metastasis. This example illustrates that through rational design, mRNA vaccines can overcome the immunosuppressive nature of the lung cancer microenvironment (72). Targeted Therapy Precision mRNA functions as a blueprint for polypeptide chain biosynthesis, and the resultant polypeptide chains can undergo post-translational modifications to achieve effective functional conformational folding (23, 49). In addition, mRNA vaccines can produce multimeric polypeptide chains that cannot be correctly folded and assembled during normal biological responses. This enables both transmembrane and intracellular protein species to be easily transported to precise sub-cellular locations, ensuring the precision of targeted therapy. However, multiepitope mRNA vaccines also face significant challenges during the “folding” process from linear sequences to the correct three-dimensional structure, which is directly critical to the vaccine’s final efficacy. For example, if the polypeptides expressed by the vaccine fail to fold correctly, their conformation may significantly differ from that of the native antigen. This would render them ineffective for recognition by immune cells—akin to “a wrong key failing to unlock the door”—thereby preventing the activation of an immune response against the actual tumor and resulting in a loss of the vaccine’s immunogenicity (73). Furthermore, misfolded proteins are typically structurally unstable and more prone to rapid degradation by intracellular proteasomes. This means that even if the vaccine successfully expresses the antigen, it will be quickly “treated as waste” and cleared by the cell, failing to provide sufficiently sustained and potent immune stimulation (74).

Once fundamental production facilities and processes are established, adapting to different tumor antigens often requires only changing the mRNA sequence. This platform-based characteristic can, to a certain extent, save research, development, and production costs compared to other types of vaccines. Additionally, for traditional immunosuppressant therapies, the cost per treatment course can be as high as $11,733. For antibody-drug conjugates (ADCs), taking trastuzumab deruxtecan (DS-8201) as an example, the annual treatment cost in China reaches approximately 390,000 RMB. In contrast, it is projected that the annual treatment cost for universal mRNA cancer vaccines may be around $45,000.mRNA-based cancer immunotherapeutic agents employ mRNA to convey tumor-associated antigens or immunomodulatory molecular entities in conjunction with delivery vehicles and immunological adjuvants, aiming to evoke anti-neoplastic immune responses. The mRNA can be generated ex vivo by utilizing DNA templates, ribonucleoside triphosphates, and recombinant enzymatic substances (75, 76), during which a promoter of DNA-dependent RNA polymerase (such as T3, T7, or SP6) is integrated. The DNA template is subsequently linearized to act as a substrate for the synthesis of mRNA that is catalyzed by a DNA-dependent RNA polymerase. This is succeeded by template digestion by DNases. During the transcription process, a 5′ cap structure and a 3′ poly(A) tail are appended to boost translation efficiency in the in vivo environment. Subsequently, unbound nucleotides, enzymes, truncated RNA fragments, and remaining DNA residues are eliminated to purify the mRNA. This optimized procedure enables the expeditious production of mRNA, rendering it a desirable strategy for creating individualized cancer immunotherapies. Moreover, all the reagents and enzymes necessary to produce mRNA vaccines are commercially accessible. Notably, the manufacturing process of mRNA vaccines typically takes approximately 10 days, substantially shorter than that needed to produce other vaccine types (75). Therefore, the capacity to rapidly, cost-effectively, and readily scale up production significantly broadens the application prospects of mRNA-based vaccines for lung cancer.

mRNA-based immunization endeavors to trigger or augment potent anti-tumor immune reactions. Synthetic mRNAs encoding TAAs or TSAs can be delivered through ex vivo mRNA-engineered autologous DCs or formulated or unformulated mRNA injections. However, safety remains the primary concern for any new treatment method. Notably, the manufacturing of in vitro transcribed mRNA employs a cell-free methodology, effectively circumventing risks of protein or viral contamination commonly observed in alternative vaccine development systems, such as replicating viral carriers, attenuated pathogens, and protein-based formulations. Furthermore, the swift intracellular conversion of mRNA transcripts into functional polypeptides substantially mitigates potential microbial contamination risks. In contrast to DNA-based immunization strategies, mRNA-based formulations avoid genomic integration events, thereby minimizing the likelihood of insertional mutagenesis and associated oncogenic risks. The inherent susceptibility of mRNA molecules to endogenous ribonuclease-mediated degradation enables precise temporal regulation of protein synthesis through half-life modulation strategies (77, 78).

Initial human trials evaluating unformulated mRNA administration in melanoma patients, initiated in 2008, established foundational safety and tolerability profiles for this therapeutic modality. Notably, no grade III or IV adverse reactions, as defined by the WHO, were observed during the trial (78). Encapsulated mRNA vaccines, such as those complexed with protamine and administered via intradermal delivery in advanced-stage malignancies, predominantly trigger localized injection-site inflammation or systemic asthenia. Furthermore, multicenter phase III studies utilizing standardized toxicity grading systems have validated that mRNA-based formulations exhibit favorable safety profiles and align with precision oncology frameworks through mechanism-driven therapeutic targeting (79–81).

Nevertheless, the potential adverse reactions from mRNA vaccine injections cannot be overlooked. As mentioned, part of the advantage of mRNA vaccines stems from the fact that even without adjuvants, the lipid nanoparticles (LNPs) containing mRNA can activate innate immune responses, exerting a pro-inflammatory effect and primarily inducing cytokines such as IL-1β and IL-6. However, a strong innate immune response is a double-edged sword: it is necessary to initiate subsequent adaptive immunity, but if overactivated, it can accelerate vaccine clearance and lead to systemic inflammatory reactions, such as common adverse effects like fever, chills, and fatigue. In preclinical studies, it has even been observed that mRNA vaccines encoding secreted IL-12, while enhancing anti-tumor effects, could trigger systemic inflammatory cytokine storms and significant toxicity such as weight loss (82). The precise mechanisms underlying mRNA vaccine-associated myocarditis remain under investigation. Notably, studies have found that in immune checkpoint inhibitor (ICI)-related myocarditis (irMyocarditis), there is a significant upregulation of programmed death-ligand 1 (PDL1) expression on cardiomyocytes. This suggests that certain immune-related adverse events may be linked to the excessive or abnormal activation of the immune system. Therefore, publicly available data from more participants are still needed to establish the safety profile of mRNA vaccines (83).

Antigen selection is the first crucial step in vaccine development. Ideal candidate vaccines should have tumor cell specificity, be involved in tumor occurrence and progression, and can evoke an immune response without immune tolerance and stimulate antitumor immunity (84–86). Immunogenic targets currently used for cancer treatment include TAAs and TSAs (86). TAAs typically show higher expression levels in tumor cells than in normal cells. However, as self-antigens, TAAs may induce immune tolerance, which can reduce the efficacy of vaccines. In contrast, TSAs are only found on cancer cells, have a high ability to trigger immune responses, and have many different epitopes, making them suitable for personalized vaccines. Creating customized mRNA vaccines for lung cancer involves identifying unique mutations in the tumor by analyzing advanced sequencing information derived from both tumor samples and corresponding normal tissues (87). Computational methods for predicting neoantigens are employed to assess the expression levels and forecast the binding affinity of peptides originating from mutated genes to MHC alleles. A high transcript expression level is strongly linked to more robust T-cell responses and can offset the low MHC binding affinity in some mutations (88). However, among the neoantigens predicted by computational tools, many fail to elicit effective T-cell immune responses in vivo. These false-positive neoantigens represent a major obstacle in advancing vaccine development. An important source of false positives is homologous sequences. If a neoantigen’s sequence is too similar to certain normal human proteins, to maintain self-tolerance, the corresponding T cells may have already been eliminated in the thymus (i.e., central tolerance) or are rendered anergic in the periphery. Therefore, it is essential to incorporate alignment against the human proteome sequence in the algorithm design to filter out candidate neoantigens with high similarity to self-sequences (89). Computationally predicted candidate neoantigens must undergo in vitro and in vivo experimental validation to confirm their true immunogenicity. Typically, the predicted neoantigen peptides are synthesized and co-cultured in vitro with immune cells (such as peripheral blood mononuclear cells, PBMCs) isolated from the patient’s blood. If the neoantigen can be recognized by specific T cells, it will activate and induce their proliferation. This can be assessed by detecting T-cell activation markers (e.g., CD137), measuring cytokine secretion (e.g., IFN-γ), or directly observing T-cell proliferation.

In addition to MHC-I-binding TSAs, further MHC-II-binding TSAs are essential for a successful antitumor immune response. Prediction tools, such as NetMHCpan and MHCflurry, are utilized to evaluate the interaction strength between ligands and MHC molecules (90–93). However, the stability of the neoepitope-MHC complex is a more critical factor for predicting immunogenicity than its binding affinity (93). NetMHCstabpan is a tool designed to predict the stability of these complexes, proving valuable in identifying immunogenic mutations (94). Beyond the mere display on the cell surface, the engagement between peptide-MHC complexes and T-cell receptors is vital for triggering immune responses. This engagement is predicted based on the interaction between the amino acid side chains of T cell receptors and the MHC-bound peptides (94). It is noteworthy that the expression and function of MHC molecules are finely regulated. For instance, recent studies have revealed that MHC class II molecules in murine antigen-presenting cells undergo branched ubiquitination mediated by the E3 ubiquitin ligase MARCH1, dependent on K11- and K63-linked chains. This modification regulates the endocytosis and degradation of MHC II, thereby influencing its antigen-presentation function. This suggests that beyond binding affinity, the metabolic regulation of MHC molecules themselves is a key factor affecting the efficiency of neoantigen presentation (95). In addition, tumors are not composed of identical cells; there exist genetic differences within them, known as tumor heterogeneity. This results in vaccine-targeted neoantigens possibly being present only in a subset of tumor cells (subclones), thereby granting a growth advantage to other cell clones that do not express these neoantigens (immune escape). The CalicoST algorithm, published in Nature Methods in 2024, provides a powerful tool to address this issue. It can infer allele-specific copy number variations from spatial transcriptomics (SRT) data and reconstruct the evolutionary trajectory of tumor clones within the tissue spatial context. This technology can identify genomic events that are difficult to detect with conventional methods, such as copy-number-neutral loss of heterozygosity, which may lead to the loss of neoantigens. By mapping the “evolutionary landscape” of the tumor, it can guide us to prioritize neoantigens present in the trunk clonal of all tumor cells as vaccine targets, thereby more effectively attacking the root of the tumor (96). Recently, promising pipelines for identifying tumor antigens have been established by screening overexpressed and mutated genes and identifying prognostic and APC-related candidates. These pipelines utilize sophisticated methods like next-generation sequencing and mass spectrometry to identify peptides bound to HLA and forecast neoantigens with precision (97). Currently, there is a lack of widely accepted, unified standard protocols and quality control measures for neoantigen prediction pipelines. Different research groups may employ different combinations of algorithms and parameters, making it difficult to compare and validate prediction results across platforms and studies. Despite progress in machine learning methods, there is still significant room for improvement in prediction accuracy. A considerable number of false-positive and false-negative predictions persist, as many candidate peptides identified computationally ultimately fail to elicit T-cell responses in in vitro experiments, further highlighting the limitations of current algorithms. Computational tools are typically based on relatively simplified models. For instance, current algorithms still struggle to fully integrate and simulate the impact of complex biological regulatory layers—such as the ubiquitination modification of MHC class II molecules—or the dynamic evolution of tumor spatial heterogeneity on the final presentation of neoantigens. In the future, more advanced and reliable algorithms and methods are certain to emerge, enabling the identification of highly effective antigens for personalized mRNA lung cancer vaccines (98).

However, although these methods are theoretically feasible, no corresponding validation measures have been established. Therefore, selecting the optimal antigens for mRNA vaccine development remains a significant challenge. However, recent advancements in computational biology have accelerated the prediction of tumor antigens. In the future, more advanced and reliable algorithms and methods will likely emerge, enabling the identification of effective antigens for personalized mRNA lung cancer vaccines.

Prior to the clinical implementation of mRNA-based cancer immunotherapies, several crucial aspects need to be resolved, such as the delivery mechanism, stability characteristics, translation efficiency, and immunogenic potential (23, 99, 100). For instance, unformulated mRNA, due to its large molecular dimensions, susceptibility to degradation, and electrical charge properties, cannot penetrate cell membranes with high efficiency and reach the cell cytoplasm. However, in the case of immature dendritic cells, which possess a unique ability to take in mRNA through micropinocytosis (99, 101). To improve mRNA delivery to APCs, the selection and optimization of mRNA formulations—such as liposomes, polymer complexes, polynucleosomes, and lipocomplexes—and administration routes are crucial. Once mRNA is successfully delivered, its duration of existence within the body must be precisely modulated since multiple elements impact the pharmacodynamic and pharmacokinetic features of mRNA-based treatments. To boost the resilience of the mRNA, its configuration ought to be enhanced, such as fine-tuning the 5′ cap structure, the poly(A) tail length, the untranslated regions’ sequences, and the protein-encoding ORF (102–104).

State-of-the-art mRNA vaccine delivery platforms predominantly utilize engineered nanoscale carrier architectures, incorporating biologically derived viral vectors and synthetic non-viral nanoparticles. These vectors not only trigger cellular immunity but also evoke humoral immune reactions. Lipid nanoparticles (LNPs) represent the principal nanomaterials for mRNA vaccine conveyance and usually comprise four constituents: ionizable cationic lipids, lipid-linked polyethylene glycol, cholesterol, and naturally occurring phospholipids. The production of LNPs represents a crucial procedure that has an immediate impact on their dimensions and encapsulation efficacy. The crux of creating minute and homogeneous LNPs lies in the swift combination of an excess amount of water and the ethanol-lipid phases. Microfluidic mixing, which uses an interdigitated herringbone structure, is an alternative technique for manufacturing LNPs across various scales (104). Once the mRNA vaccine enters the target cell, the formation of complexes with the released cationic lipids is crucial to enable the delivery of nucleic acid. The negatively charged lipids within the cell membrane can counteract the electrical charge of the cationic lipid vectors, disturbing the electrostatic forces between the lipid vectors and the nucleic acid molecules. This, in turn, eases the liberation of LNPs and the subsequent transfer of mRNA into the cytoplasmic region (105, 106). The extracellular stability of mRNA can be augmented through encapsulation within LNPs, as they safeguard it from breakdown by the omnipresent ribonucleases (107). The application of nanocarriers has the potential to prolong the lifespan of biologics. This is achieved through measures such as forestalling untimely release and decomposition, as well as evading elimination. The organs responsible for this potential elimination are the kidneys and the liver (108). Consequently, delivery systems founded on nanomaterials can preserve the extracellular stability of mRNA-based vaccines.

Notably, LNP-based mRNA vaccines can address the limitations of traditional carriers, such as suboptimal cellular protein synthesis, inadequate antigen payload, and issues related to APC maturation. These mRNA vaccines are independent of host influence and can induce a durable immune response.

Besides delivery and stability aspects, immunogenicity is also a factor that demands consideration. Immunogenicity is determined not only by the delivery and stability of mRNA but is also closely related to its translation efficiency. The higher the translation efficiency, the greater the quantity of antigen synthesized by the cells, thereby enabling a more robust activation of the immune response (109, 110). Accumulating data indicates an inhibitory feedback loop among mRNA and the immune response it induces. For example, exogenous RNA can stimulate the generation of type I IFNs through primary immune activation (111). Over-production of type I interferons restricts translational processes and accelerates the breakdown of ribosomal RNA and cellular mRNA (102). Specifically, exogenous RNA activates proteins such as PKR, OAS, and IFIT, which suppress translation efficiency by phosphorylating eIF2α and degrading RNA, thereby reducing the protein expression of mRNA. To overcome this challenge, optimizing the structure of mRNA—such as lengthening the poly(A) tail, optimizing the 5’ cap structure, modifying the mRNA sequence, and performing post-transcriptional modifications (e.g., using demethylation modifications)—can effectively reduce innate immune responses and enhance translation efficiency (112). For example, N¹-methyl-pseudouridine (m1Ψ) modification significantly reduces type I IFN activation and increases mRNA translation efficiency by 2- to 4-fold. An optimized poly(A) tail length can increase translation efficiency by 1.5-fold by improving the interaction between poly(A)-binding protein and eIF4G, thereby promoting translation (113, 114). eIF4E, as a key factor in mRNA translation, initiates the translation process by binding to the 5’ cap structure of mRNA (115). Optimizing the 5’ cap structure of mRNA to bind more efficiently with eIF4E can significantly enhance translation efficiency and increase antigen expression (116). Studies show that an optimized Cap1 cap structure can improve translation efficiency by 3- to 5-fold while enhancing mRNA stability and translation efficiency [Furuichi 2015]. These optimizations will directly increase the expression level of the target antigen, thereby enhancing immunogenicity and further promoting the generation of specific immune responses (117). Adding poly(A) tails, sequence modification, and post-transcriptional refinement can mitigate innate immune activation without altering mRNA translation (118–120). For example, The TriMix strategy enhances dendritic cell maturation and promotes immune responses through T-cell activation by encoding CD70, CD40L, and TLR4, demonstrating notable efficacy in lung cancer immunotherapy (121, 122). Specifically, CD70 and CD40L enhance T-cell immune activity via co-stimulation, while TLR4 activation further strengthens the ability of dendritic cells (DCs) to present tumor antigens via MHC-I molecules by boosting the NF-κB pathway. This enables mRNA-encoded antigens to more effectively activate specific T-cell responses (123, 124). These optimization strategies significantly enhance the immunogenicity of mRNA vaccines and improve their immunotherapeutic outcomes by precisely regulating mRNA translation efficiency and immune cell activation. Thus, optimizing mRNA vaccines requires not only improving delivery and stability but also enhancing immunogenicity through optimized translation efficiency, thereby more effectively boosting their therapeutic efficacy.