The evolution of vaccine strategies for colorectal cancer: from conventional approaches to the mRNA era

Abstract

Colorectal cancer (CRC) remains one of the leading causes of cancer-related mortality worldwide, with current treatment modalities frequently limited by tumor heterogeneity and immune evasion. Vaccine-based immunotherapy, particularly that utilizing messenger RNA (mRNA) technology, represents a promising innovative strategy due to its rapid development cycle, generally manageable safety profile in the short term, and potential for personalization. This review examines the limitations of traditional cancer vaccine platforms and contrasts them with the advantages offered by mRNA-based vaccines. It expounds on the various mechanisms through which mRNA vaccines target CRC, such as encoding tumor-associated antigens or tumor-specific neoantigens, delivering immune-stimulating cytokines, and modulating the immunosuppressive tumor microenvironment. An overview of recent clinical trials in the field of mRNA immunotherapy for colorectal cancer is provided. Preliminary evidence from early clinical studies indicates that these approaches exhibit a manageable safety profile and show signs of antitumor activity, showcasing significant potential for clinical application. Furthermore, we also discuss the challenges and prospects that mRNA vaccines are currently facing.

1. Introduction

Colorectal cancer (CRC) is the third most prevalent type of cancer worldwide (1). Despite significant advancements in endoscopic techniques for screening and early detection, which have contributed to a reduction in incidence in some regions, CRC remains the second leading cause of cancer-related mortality globally. Furthermore, in many developing countries, the incidence, prevalence, and mortality rates associated with CRC continue to be alarmingly high and are projected to increase further over the next decade, particularly in regions undergoing rapid economic transition (2).

The primary treatment for CRC involves surgical resection, which may be accompanied by adjuvant therapies such as chemotherapy or radiotherapy (3). Although immunotherapy combined with chemotherapy has been shown to significantly extend progression-free survival and overall survival in a subset of patients with CRC (e.g., those with deficient mismatch repair/microsatellite instability-high tumors), challenges including tumor heterogeneity and drug resistance continue to pose substantial obstacles to effective treatment for the majority of patients (4). Consequently, addressing these therapeutic challenges in individuals with advanced CRC may require alternative therapeutic strategies or new immunotherapeutic approaches (5, 6).

Vaccines traditionally function by stimulating the immune system to recognize and combat specific pathogens. Since the 18th century, vaccination has culminated in the complete eradication of smallpox and has significantly advanced the control of various infectious diseases (7). This success has inspired efforts to apply similar immunological principles to oncology, positioning vaccines as a promising frontier in cancer research. While early efforts focused on preventing virus-associated cancers, the broader goal of treating existing malignancies has driven the development of therapeutic cancer vaccines—designed to activate the immune system against established tumors. In 2010, FDA approval of Sipuleucel-T (Provenge)—the first therapeutic cancer vaccine—spurred research into neoantigen and vector-based platforms to overcome immune suppression and broaden efficacy (8). Subsequently, various platforms for cancer vaccines have been explored, each characterized by distinct mechanisms and challenges. These include cell-based vaccines, which utilize whole tumor cells or dendritic cells to present a broad array of antigens; microorganism-based vaccines, employing attenuated bacteria or viral vectors to deliver tumor antigens and provoke robust innate immune responses; exosome-based vaccines, which leverage exosomes for delivery to facilitate antigen presentation and immune activation; protein and peptide based vaccines, valued for their straightforward design, safety profile, and ease of production; and DNA-based vaccines, offering high stability and the capacity to induce sustained antigen expression (9). While each platform offers unique advantages, they are often constrained by limitations in immunogenicity, manufacturing complexity, or scalability, underscoring the imperative for more versatile and potent alternatives.

During the COVID-19 pandemic, messenger RNA (mRNA) vaccines rapidly gained prominence due to their unique advantages, including a favorable short-term safety profile, short development cycles, and flexible production capabilities (10). This attention has accelerated their application in cancer research. As a frontier technology derived from molecular biology and immunology (11), mRNA was first successfully demonstrated in an animal model using in vitro transcription (IVT) in 1990 (12). Over the past few decades, continuous advancements in nucleotide modification techniques and delivery strategies have significantly enhanced the efficacy, safety, and scalability of mRNA vaccines (13, 14). Collectively, these advancements underscore the growing promise of mRNA vaccines for applications in both tumor prevention and treatment.

The capacity of mRNA vaccines to engage both innate and adaptive immunity positions them as a uniquely adaptable platform. This inherent versatility, combined with their rapid and flexible manufacturing, makes them particularly attractive for personalised cancer immunotherapy. For instance, mRNA constructs encoding tumor-specific antigens (TSAs) offer a theoretical advantage: each patient translates the mRNA into proteins, which are then processed and presented in a manner that is unique and personalized. Antigen-presenting cells (APCs) can take up, process, and present these TSAs in the context of various human leukocyte antigen (HLA) class I and II molecules. If the resulting epitopes bind with high affinity to the patient’s HLA molecules, they can be recognized by T cells, leading to a broader and more efficient T-cell response in vivo (15–17).

In light of the profound heterogeneity and immune evasion that characterize CRC, mRNA vaccines offer a promising solution to current therapeutic challenges. Their rapid production and inherent customizability enable the development of personalized treatment platforms, facilitating the design of multi-epitope vaccines capable of targeting tumor clonal diversity (15, 18). This strategy may help reduce the risk of immune escape driven by heterogeneous antigen expression. Moreover, mRNA vaccines can activate both innate and adaptive immune pathways, eliciting robust CD8⁺ and CD4⁺ T cell responses specific to target antigens. This immune activation can promote the conversion of immunologically “cold” tumors into “hot” tumor microenvironments characterized by T cell infiltration, potentially overcoming mechanisms of immune evasion (19). Preclinical studies and early clinical trials have demonstrated promising results, underscoring their potential to activate tumor-specific immune responses and improve patient prognosis (5, 20, 21). Nonetheless, there remain challenges in augmenting immunological effects, refining delivery mechanisms and guaranteeing stability.

This review summarizes the advantages and disadvantages of different tumor vaccines, with particular emphasis on the potential biological functions underlying mRNA vaccines in CRC. We also discuss recent advancements in their clinical applications along with the challenges encountered and future directions for development in this field.

2. CRC vaccine strategy: from conventional to mRNA platforms

2.1. Conventional cancer vaccines: exploration and limitations

A wide variety of conventional vaccines for colorectal cancer exist, which can be categorized into several major types based on the final form that enters the human body: cell-based, microorganism-based, exosome-based, protein/peptide-based, and DNA-based vaccines. This section provides an overview of these conventional vaccine types and focuses on the limitations associated with each method in terms of formulation or delivery.

2.1.1. Cell-based vaccines

Cell-based vaccines leverage whole cells as either the source of antigens or as the delivery vehicle. Currently developed cell-based vaccines for CRC can be broadly categorised into two types based on their formulation and delivery strategy. The first type, tumour cell vaccines, utilize whole tumour cells (often autologous or allogeneic) as the antigenic formulation (22). The second type, dendritic cell (DC) vaccines, utilize DCs as a delivery system. By loading peptides or mRNA onto dendritic cells and reinfusing them into the patient’s body, they can exert their effects (23). Of these two types, vaccines that utilize tumor cells—especially autologous tumor cells—are well-suited for personalized treatment approaches, as they can be directly matched to the patient’s specific antigenic profile (22). In various preclinical cancer models and clinical trials evaluating similar vaccines in patients with CRC, it has been observed that the anti-tumor response correlates with the immune response elicited by the vaccine (24–27). GVAX^® and Vigil™ serve as representative examples of this category. In a Phase I clinical trial (NCT00656123), GVAX^® combined with cyclophosphamide was associated with extended survival in patients following radical resection of liver metastases from colorectal cancer; however, its efficacy may be limited in patients presenting with a significant burden of metastatic disease (27). Additionally, results from two other GVAX^® vaccine trials indicated limited clinical efficacy, as seen in NCT02981524, NCT01966289. Vigil™, while exhibiting promising outcomes in other malignancies (NCT01309230), was terminated in CRC (28). The clinical trial registry (NCT01505166) cites a “Business Decision to pursue other indications” as the official reason¹. This decision likely reflects the inherent and widely reported challenges within the field of autologous tumor cell vaccines, particularly their intricate manufacturing processes and the logistical hurdles associated with obtaining sufficient, viable tumor tissue from colorectal cancer patients (29).

DC vaccines leverage the capabilities of dendritic cells, which are recognized as the most potent professional APCs (30). Typically, DCs are utilized to load antigens and subsequently infuse them back into patients (31–34). In a Phase I clinical trial (NCT00558051), nine patients with invasive and recurrent malignancies received intranodal injections of DC vaccines loaded with killed autologous tumor cells, keyhole limpet hemocyanin, and pan DR helper T cell epitope. The overall cohort survival period was > 28 months (± 25 months). Among them, one patient achieved long-term disease-free survival for over 90 months post-treatment (31). In another Phase I/II clinical trial (NCT00228189), ten patients with CRC liver metastases received intradermal and intravenous injections of dendritic cells pulsed with carcinoembryonic antigen (CEA) peptide before undergoing resection of liver metastases. A significant number of CEA-specific T cells were detected in delayed-type hypersensitivity (DTH) biopsies from seven patients post-treatment, which produced substantial amounts of interferon-gamma (IFN-γ) upon stimulation with target cells loaded with CEA. Interestingly, within this study, a comparative analysis revealed that DCs transfected with CEA mRNA were not superior to CEA peptide-pulsed DCs in inducing tumor-specific immune responses (35). The underlying mechanisms for this unexpected observation remain elusive; however, they may stem from differential efficiencies in MHC class II antigen processing or the induction of distinct T-cell subsets by each vaccination strategy, as suggested by studies in other tumor types (36). These findings underscore the potential application value of DC vaccines in CRC; however, the development process for these vaccines is lengthy and necessitates autologous cell preparations, which may not align with the economic requirements for precision treatment (37).

It is noteworthy that a bioengineered allogeneic immune cell vaccine, AlloStim, has also been utilized in the treatment of colorectal cancer (NCT02380443, NCT01065441, NCT00861107)². This vaccine employs allogeneic activated CD4⁺ Th1-like cells and is intentionally mismatched with the recipient to induce reactivation of the immune response through a graft-versus-host-disease-like reaction (38). Specifically, a case report associated with a Phase IIb trial documented a rare objective response to ICIs in a single patient with pMMR/MSS mCRC following AlloStim^® treatment (39). While this individual case suggests a potential for therapeutic sensitization, preliminary results from the NCT02380443 trial, as reported on ClinicalTrials.gov, indicate that a majority of participants experienced varying degrees of adverse reactions³. These observations highlight the necessity for larger, controlled clinical trials to rigorously establish both the safety profile and the therapeutic efficacy of this combination strategy.

2.1.2. Microorganism-based vaccines

Microorganism CRC vaccines include bacterial, viral, and yeast-based vaccines. These vaccines possess inherent immunogenicity, and their genetic material can be modified to incorporate cancer antigens (40). The development of viral vector vaccines has progressed relatively rapidly (41, 42). A Phase I clinical trial involving patients with CRC demonstrated that a recombinant avian pox virus is safe but elicits only limited T-cell responses in cancer patients (42). This observation may be partly attributed to the induction of neutralizing antiviral antibodies—a known challenge for viral vector platforms. Yeast-based vaccines have been shown to be safe in multiple Phase I clinical trials (43–45). An ongoing Phase I clinical trial aims to integrate the yeast vaccine with personalized treatment approaches (NCT03552718). Concurrently, recent preclinical research indicates that antigen-anchored yeast vaccines can significantly activate intestinal DCs, thereby amplifying the immune response (46). Bacteria-based vaccines, such as Salmonella typhimurium, have been utilized in other malignancies due to their ability to achieve tumor-specific colonization (47, 48). However, microorganism-based vaccines require intricate manufacturing processes, which significantly extend production timelines. Furthermore, the potential presence of pre-existing antibodies against these microorganism vectors poses a substantial barrier to their implementation in personalized therapeutic strategies (49).

2.1.3. Exosome-based vaccines

Exosome-based vaccines refer to tumor vaccines formulated with exosomes. Exosomes are tiny membrane vesicles secreted by various cell types, containing multiple bioactive molecules, and can transfer information between cells. Their immunomodulatory function depends on their cell origin (50–52). For example, exosomes derived from dendritic cells can enhance antigen presentation and T cell activation by carrying MHC I/II complexes, thereby strengthening the immune response. In contrast, tumor-derived exosomes have been implicated in immune suppression and tumor progression, reflecting their complex “double-edged sword” roles in cancer biology (53, 54). A Phase I clinical trial involving 40 patients with metastatic CRC evaluated the therapeutic potential of exosomes from tumor ascites (Aex) in combination with granulocyte-macrophage colony-stimulating factor (GM-CSF). The results demonstrated that the Aex combined with GM-CSF group elicited a stronger anti-tumor cytotoxic T-cell response, suggesting its promise as a safe and effective therapeutic vaccine (55). However, the therapeutic index of exosome-based vaccines is highly sensitive to manufacturing purity and standardization. One of the primary challenges is that vaccines derived from insufficiently purified tumor-associated exosomes may inadvertently carry immunosuppressive cargoes, thereby counteracting the intended anti-tumor response (56–58). Additionally, large-scale production and storage of exosomes in clinical applications continue to pose significant challenges (59). Although exosomes are generally regarded as biocompatible (60), their immunogenicity requires thorough evaluation, which substantially impacts their clinical utilization.

2.1.4. Protein and peptide based vaccines

Peptide and protein vaccines are designed to trigger immune responses against specific tumor antigens. These vaccines mainly target immunogenic epitopes from tumor-associated antigens (TAAs) or TSA. Antigens are presented on the surface of APCs through MHC Class I or II molecules, activating T cells and inducing sustained antigen-specific immune memory (61). Generally, peptide cancer vaccines exhibit good tolerability among patients (62–65). A recent study demonstrated that PolyPEPI1018 combined with maintenance therapy is safe and well-tolerated in patients with microsatellite-stable (MSS) metastatic CRC (n=11). The vaccine induced CD8⁺ T-cell responses in 90% of patients, with 80% targeting at least three antigens, and achieved an objective response rate of 27.3%. Notably, two patients became eligible for curative surgery after vaccination, and those receiving multiple doses had significantly longer progression-free survival (12.5 vs. 4.6 months; P = 0.017), indicating promising clinical efficacy signals (63). In another investigation, the survivin-2B peptide vaccine was confirmed to be safe in HLA-A24-positive patients with advanced or recurrent colorectal cancer, with no severe adverse events reported. However, despite observed immunological responses—such as increased peptide-specific cytotoxic T lymphocytes (CTLs) frequency in one patient—the clinical benefits were modest: only transient tumor marker decreases in 6 out of 15 patients and a single minor response, while the majority showed disease progression (64). These findings underscore key limitations of peptide-based vaccines for personalized cancer immunotherapy. First, their strict HLA restriction limits applicability to specific patient subgroups (e.g., HLA-A24⁺ patients)—an inherent limitation of any epitope-based vaccine, but particularly pronounced in peptide vaccines due to their limited epitope capacity (typically 1–3 epitopes per peptide). Second, low intrinsic immunogenicity of short peptides often results in weak or transient T-cell activation, as reflected in the modest clinical responses observed. Unlike live or inactivated pathogens, peptides lack pathogen-associated molecular patterns (PAMPs) and fail to effectively stimulate innate immunity, necessitating the use of adjuvants, carrier proteins, or multimeric presentation systems (e.g., VLPs or nanoparticles) to enhance immunogenicity. Additionally, conformational limitations pose additional challenges for B-cell epitope design: linear peptides often fail to mimic native antigen structures, leading to antibodies that may not recognize the target protein in its natural conformation (66).

2.1.5. DNA-based vaccines

DNA based vaccines utilize DNA as a template for encoding antigens, facilitating their transfection into cells. While DNA vaccines are generally less costly than RNA vaccines and exhibit greater stability, they have not yet been widely implemented in clinical practice (67). This is primarily due to the potential risk of genomic integration, which may lead to insertional mutagenesis, as well as the challenges posed by the prolonged expression of encoded antigens (68). Specifically, chronic exposure to the same antigen can induce immune tolerance or T-cell exhaustion, where the immune system ceases to recognize the vaccine-derived antigen as a foreign threat, thereby diminishing the therapeutic effect (69). Consequently, safety remains a significant concern regarding the application of DNA vaccines in CRC treatment.

In conclusion, while conventional vaccine approaches have demonstrated some progress in cancer therapy, they face common hurdles. These include limited immunogenicity, complex and time-consuming manufacturing processes for personalised products, and the constraints of immune compatibility (e.g., HLA restriction). Therefore, the development of a novel vaccine capable of addressing these interrelated challenges is critically imperative.

2.2. Structural and advantages of mRNA vaccine platforms

2.2.1. Types and structural features of mRNA vaccines

Both mRNA vaccines and traditional vaccines are designed to activate the immune system (70). Compared with traditional vaccine technologies, mRNA platforms offer distinct advantages in terms of production speed, mechanisms of action, safety profiles, and personalization capabilities (70, 71). Research has highlighted that these characteristics endow mRNA technology with significant therapeutic potential and strategic value in addressing the complexity and heterogeneity of cancer (71). Consequently, this technology has increasingly become a pivotal direction in the field of cancer vaccine research in recent years. Currently, there are three primary types of mRNA cancer vaccines: traditional non-replicating mRNA, self-amplifying RNA RNA (saRNA) and Circular RNAs (circRNAs) (Figure 1). The fundamental structure of traditional non-replicating mRNA consists of an open reading frame (ORF) region that encodes the target peptides sequence, flanked by five prime (5’) and three prime (3’) untranslated regions (UTR). This structure is further stabilized by a 7-methylguanylate (m7G) 5’ cap and a poly(A) tail at the 3’ end. The addition of the 5’ cap and 3’ poly(A) tail can occur during IVT or through enzymatic addition following initial IVT. In contrast, saRNA contains two ORFs; one encodes the targeted antigen sequence while the other encodes a viral replication mechanism that facilitates long-term RNA amplification within cells (72). Beissert et al. developed an enhanced version of saRNA termed trans-amplified RNA (taRNA) (73). This approach is based on a dual-vector system comprising two distinct templates, which separately generate antigen-encoding alphaviral RNA and replicase-encoding RNA separately. Compared to conventional saRNA, the binary design of taRNA allows for both simpler manufacturing and greater flexibility in functionalization. CircRNAs are widely expressed RNA transcripts found in different species, which enhances molecular stability by conferring resistance to exonuclease-mediated degradation (74–76).

Figure 1.

Schematic diagram of mRNA types. (A) conventional mRNA. (B)self-amplifying mRNA, which contains four genes encoding nonstructural proteins (NSP1–NSP4) that together constitute the replicase complex. Expression of the target gene is enabled by a regulatory subgenomic promoter. (C) trans-amplifying mRNA, consisting of two mRNA species: one retains the genes encoding the replicase, and the other expresses the gene of interest. (D) circRNA, composed of an IRES, interest gene, and a linking part—the linking part primarily refers to the circularization system/signal and other auxiliary elements (such as translation-enhancing elements). UTR, untranslated region; NSP, non-structural protein; IRES, internal ribosome entry site. Created with BioRender.com.

CircRNAs can be genetically modified to facilitate protein synthesis by incorporating internal ribosome entry sites (IRES) or by including modified nucleotides, such as N⁶-methyladenosine, which can enhance translation efficiency (75). This structural resilience makes circRNAs particularly attractive for sustained antigen expression in vivo.

Upon cytosolic delivery, these mRNA constructs are translated by host ribosomes into antigenic proteins, which then undergo post-translational processing to yield functional, correctly folded immunogens. These structural characteristics enable mRNA vaccines to achieve efficient antigen expression and immune activation in CRC—an advantage of particular relevance in CRC, where inter- and intra-tumoral heterogeneity poses a major therapeutic challenge (11, 77).

2.2.2. Advantages of mRNA vaccines

mRNA vaccines exhibit several significant advantages over conventional methods and may prove more effective against various cancers. First, the development of RNA-based vaccines is relatively rapid and cost-effective. IVT has transformed mRNA manufacturing through an advanced industrial system, significantly reducing production costs. The IVT process eliminates cellular and their associated regulatory obstacles, resulting in a more expedited production timeline compared to other vaccine modalities (78). Second, mRNA is non-infectious and non-integrative, which substantially reduces the risk of adverse immune responses while preventing the potential integration of foreign genes into the host genome (79). Furthermore, mRNA is transiently expressed as it is degraded by normal cellular mechanisms, enhancing its safety profile. Moreover, due to rapid advancements in various nucleotide modification techniques and improvements in in vivo delivery methods, contemporary mRNAs demonstrate enhanced stability and customizability while facilitating swift uptake and expression (80–82).

2.2.3. Practical limitations and cross-platform comparison

Despite the transformative advantages of the mRNA platform, its practical limitations in the context of traditional vaccine strategies also warrant attention. These primarily include the following aspects:

(1) Delivery system: Precisely targeting intended cells (e.g., dendritic cells in situ) remains a primary challenge for ensuring vaccine efficacy (2). Response durability: While ensuring sustained protein expression, excessive reactogenicity must be tightly controlled to balance efficacy and safety (3). Manufacturing and logistics: Although production is rapid, the dependence on ultra-cold storage and process stability during scale-up remain bottlenecks for global deployment, especially in resource-limited regions.

Through the following table, a clear cross-platform comparison of key attributes of different vaccine platforms for CRC treatment can be provided. Table 1 presents a summary of the advantages and disadvantages of different vaccine platforms for colorectal cancer.

Table 1.

Comparison of colorectal cancer vaccines.

Vaccine platform	Subtype/personalization degree	Advantages	Disadvantages	References
Cell-Based Vaccines	tumour cell vaccines/Personalization	• Target a broad range of antigens • Induces both T- and B-cell responses	• Expensive, long production cycle, difficulty in tissue acquisition • Variable immunogenicity of antigens	(27, 29, 83)
	DC vaccines/Customizable	• Stimulates both CD4 + and CD8 + T cell responses • More durable immune responses	• Longer manufacturing time • Expensive and involves complicated cell culturing	(31, 37)
	allogeneic immune cell vaccine (AlloStim)/Non-personalized	• Off-the-shelf • Facilitates large-scale production and quality control	• Lack of specificity may prevent the induction of a strong adaptive immune response. • Risk of graft-versus-host disease or immune rejection response	(38, 39)
Exosome-based Vaccines	Personalization	• High safety, no cell vitality • Easy to store and transport	• Large-scale production and storage issues and standardization issues • Double-edged sword	(53, 54)
Protein and peptide based Vaccines	Customizable	• Straightforward GMP synthesis • Cheap, does not require cold-chain transportation	• Low immunogenicity; requires immune adjuvants to boost the immune response • The complexity of long peptide synthesis and purification leads to high HLA restriction.	(61, 63, 64)
Microorganism-based Vaccines	Bacterial vector vaccines/Customizable	• High immunogenicity and self-antigenicity • Can target tumor via intracellular infection	• Safety concerns due to live attenuated bacteria(such as systemic inflammation) • Complex production and manufacturing process	(47, 48)
	Viral vector vaccines/Customizable	• High immunogenicity and self-antigenicity • Can induce strong and durable immunity	• Pre-stored anti-vector antibodies may affect the therapeutic effect. • Safety issues (such as insertion mutations, excessive inflammatory responses)	(41, 42)
	Yeast vaccines/Customizable	• Stable and Cost-Effective • Can induce strong and durable immunity	• The production process and standardization still need to be improved. • Pre-stored antibodies may affect the therapeutic effect	(43–45)
DNA Vaccines	Customizable	• Lower production costs and degrade less readily vs mRNA neoantigen vaccines • Easy to encode multiple antigens in a single construct	• Low transfection efficiency, may require electroporation for efficient cellular uptake • Genomic integration risk	(69, 84)
mRNA Vaccines	Customizable	• Faster manufacturing time • Rapid induction of durable and functional CD8 + T cell responses • Easy to encode multiple antigens in a single construct	• Less stable, requires cold-chain logistics • Requires liposome or LNP for optimized delivery	(70, 85)

GMP, Good Manufacturing Practice; DC, Dendritic Cell; LNP, Lipid Nanoparticle; HLA, Human Leukocyte Antigen.

This comparison highlights that while mRNA technology offers unparalleled speed and flexibility for personalized therapy, challenges such as optimizing delivery systems for precise targeting(e.g., dendritic cells in situ) and ensuring durable protein expression without excessive reactogenicity remain critical hurdles for its widespread clinical application in CRC.

3. Mechanisms of mRNA vaccines in CRC

mRNA vaccines can treat tumors through at least two complementary mechanisms (1): remodeling the immunosuppressive tumor microenvironment (TME) to restore anti-tumor immunity; and (2) triggering specific immune responses to inhibit tumor growth (86, 87).

3.1. Tumor microenvironment of CRC

The TME refers to the specific environment that tumor cells rely on for their survival (88). The living environment of tumor cells is illustrated in Figure 2. The TME comprises various components, including tumor cells, immune cells, Cancer-associated fibroblasts(CAFs), extracellular matrix, and cytokines. CRC cells facilitate tumor growth and promote immune evasion by regulating the cells and molecules that favor tumor development. For instance, TGF-β signaling is highly activated in the CMS4/MSS subtype, while a subset of CMS1/microsatellite instability-high (MSI-H) tumors also exhibits a TGF-β-dependent stromal signature (89–91). It is an important factor driving the functional heterogeneity of immune and stromal cells in CRC TME. TGF-β is one of the major stimuli that promoting to differentiation of Treg cells. These Treg cells predominantly inhibit CD8+ T cell activity in colorectal cancer (88, 92). In murine models of CMS4 CRC, tumor cells secrete TGF-β2 to activate tumor-associated neutrophils, which subsequently inhibit T cell activity (93). Furthermore, TGF-β induces a myofibroblast phenotype in CAFs, resulting in aberrant extracellular matrix (ECM) protein production that facilitates the exclusion of CD8+ T cells from the TME (94).

Figure 2.

Schematic of the TME in CRC. The TME is composed of tumor cells, diverse immune cells (includingmacrophages, dendritic cells, lymphocytes, and neutrophils), cancer-associated fibroblasts (CAFs), the extracellular matrix (ECM), and aberrant tumor vasculature. Gut lumen-derived microbial components are localized near the tumor, implying their potential roles in tumor progression and immune modulation. The interactions among these cellular and non-cellular components constitute the complex TME. Created with BioRender.com.

Hypoxia and metabolic alterations are two common phenomena that significantly impact the immune response in solid tumors, including CRC (95). As the tumor progresses, cells in the center region upregulate the expression of pro-angiogenic cytokines and hypoxia-inducible factors in response to hypoxic conditions, thereby accelerating tumor growth. This phenomenon directly impacts immune cell function within the TME; for instance, hypoxia promotes apoptosis of γδ T cells (96). This may be particularly relevant in the context of MSI-H CRCs with HLA class I defects, where γδ T cells have been demonstrated to serve as key effectors of immunotherapy by eliminating tumor cells that have lost MHC expression (97). Hypoxia also stimulates the secretion of VEGF and osteopontin by various cell types such as CAFs, contributing to intratumoral angiogenesis (98). Furthermore, structural and topological vascular abnormalities exacerbate intratumoral hypoxia. In both in vitro studies and mouse models of colorectal cancer, by-products of hypoxia, including lactate and metabolites derived from anaerobic glycolysis, exert broad tumor-promoting and immunosuppressive effects on CAFs, T cells, and bone marrow-derived cells (99–101). These factors impair tumor cell recognition by the immune system, thereby hindering the efficacy of tumor vaccines.

Given the pivotal role of the TME in CRC progression and immune evasion, targeting and reshaping this immunosuppressive milieu has become a critical therapeutic strategy. The inherent programmability of mRNA vaccines provides a versatile platform for this purpose, enabling precise TME modulation through multiple strategies—including direct immune modulation and the encoding of specific antigens, immunostimulatory factors. These approaches can collectively reverse immune suppression and activate potent anti-tumor immunity.

3.2. Activation of innate immunity by mRNA vaccines

mRNA vaccines have the potential to induce both innate and adaptive immunity, thereby exerting effective anti-tumor effects. They initiate innate immune responses by recognizing PAMPs via pattern recognition receptors (PRRs) (102). Upon entering the body, in vitro-synthesized mRNA and its delivery vectors are recognized as exogenous material by PRRs. This recognition activates the innate immune response. APCs produce pro-inflammatory cytokines and co-stimulatory molecules that attract and promote the infiltration of immune cells—such as T cells, natural killer (NK) cells, basophils, and macrophages—into the tumor microenvironment (103). By recruiting key immune cells, this process drives a sustained antitumor immune response and the generation of adaptive B and T cell immunity.

Furthermore, the immunogenicity of mRNA is primarily mediated by Toll-like receptor 7 (TLR7) and TLR8 (11). TLR7 is expressed on B cells, macrophages, and DCs, where it functions to detect single-stranded RNA (ssRNA). TLR7 signaling enhances the production of pro-inflammatory cytokines and antigen presentation, while also improving the survival of memory B cells (104). Additionally, the myeloid differentiation factor 88 (MYD88)/TLR7 pathway drives the type I interferon (IFN-I) response and promotes a pro-inflammatory state via cytokine secretion (105). This pathway’s stimulatory effects upregulate the adaptive immune response induced by mRNA vaccines, and it also mediates B cell activation (106, 107). Vaccines encoding the B cell epitope trigger a specific antimetastatic effect (18). In contrast, TLR8 is mainly highly expressed in myeloid immune cells, including monocytes, myeloid dendritic cells and neutrophils (108). Similar to TLR7, TLR8 can also recognize ssRNA and its degradation products (such as fragments rich in GU sequences), thereby sensing the entry of exogenous mRNA vaccines. After activation, TLR8 also relies on MYD88 for signal transduction, initiating downstream NF-κB and MAPK pathways, and inducing pro-inflammatory cytokines (such as TNF-α, IL-12) (109, 110). It is noteworthy that the activation of TLR8 not only directly promotes the maturation of antigen-presenting cells but also, through synergy with TLR3/4, shapes an immune microenvironment biased towards Th1 type (111, 112), thereby enhancing the response of cytotoxic T cells. Additionally, the TLR8 signal may also play a certain role in regulating the inhibitory function of regulatory T cells (Treg cells), thereby indirectly influencing the intensity and persistence of the immune response induced by the vaccine (112).

In summary, these mechanisms demonstrate that mRNA vaccines can modulate acquired immunity and the tumor microenvironment through the regulation of innate immune pathways.

3.3. mRNA vaccines encoding tumor-associated antigens

TAAs are antigenic molecules found on both tumor cells and normal cells, including embryonic proteins, glycoprotein antigens, squamous cell antigens, among others (113). TAAs are commonly utilized in the clinical diagnosis of tumors; however, they are not exclusively characteristic of tumor cells. Normal cells can also synthesize these antigens in trace amounts, and their expression levels significantly increase during the proliferation of tumor cells. Most mammals demonstrate a significant degree of immune tolerance to TAAs, presenting a major challenge for the development of cancer vaccines that leverage these antigens due to central immune tolerance mechanisms (114). The trend in developing clinical targeted mRNA cancer vaccines has shifted towards employing multiple combinations of shared TAAs (115). When the mRNA encoding neoantigen enters the body, it will produce antigen proteins in the target cells. These proteins are processed into peptide epitopes that bind to MHC-I molecules. The resulting peptide–MHC-I complexes are presented to CD8⁺ T cells, activating them to differentiate into CTLs that directly kill antigen-expressing tumor cells (116). Additionally, secreted or released antigens can be taken up and presented by antigen-presenting cells via the MHC class II pathway. CD4⁺ T cells recognize these peptide–MHC-II complexes and differentiate into helper T cells, which support and modulate the ongoing immune response (117).

CEA is the most common TAA, expressed in nearly all colorectal cancers. Therefore, it represents a highly attractive option in clinical immunotherapy protocols. Multiple clinical studies have demonstrated that DCs loaded with CEA peptides can induce antigen-specific T cell responses in patients with colorectal cancer (23, 118, 119). Consequently, the use of CEA mRNA vaccines may theoretically elicit a broader and more robust T cell repertoire due to the expression of multiple epitopes. In other cancers, CEA has already shown clinical efficacy. For instance, several clinical trials have demonstrated that mRNA vaccines encoding TAAs can induce T cell responses in tumors such as melanoma (NCT04526899, NCT01278940, NCT01995708).

However, the results of TAA-based mRNA vaccines in CRC have been mixed. In a Phase I/II clinical trial (NCT00228189) involving CRC (35), peripheral blood mononuclear cells subjected to mRNA electroporation were used as stimulator cells and co-cultured with T cells. In the DTH skin test, CEA peptide-specific T cell reactivity was observed in 8 patients; conversely, no CEA peptide specificity was detected in 5 patients who received mRNA inoculation. This discrepancy may be attributed to a combination of factors including the patient’s baseline immune status, antigenic characteristics, vaccine delivery efficiency, and the sensitivity of the detection method employed. Further studies are warranted to assess the application of mRNA encoding TAAs, such as CEA mRNA, in CRC.

3.4. mRNA vaccines encoding tumor-specific antigens

TSAs are uniquely expressed on tumor cells and absent from normal tissues. Targeting TSAs represents a pivotal strategy in cancer immunotherapy, where the selection of appropriate antigens is critical for vaccine efficacy. Unlike TAAs, which are susceptible to central and peripheral immune tolerance, TSAs—particularly neoantigens arising from somatic mutations—exhibit high immunogenicity and can effectively overcome these tolerance barriers (120). Its mechanism of action within the body is the same as that described in Section 3.3, whereby antigen presentation leads to the activation of both CD8⁺ and CD4⁺ T cells, eliciting anti-tumor responses. Through these mechanisms, antigens encoded by mRNA vaccines targeting multiple TSAs can be presented through both the MHC I and MHC II pathways, thereby inducing a broad polyclonal immune response and effectively reducing immune escape caused by antigen loss (although it is important to acknowledge that these fundamental biological limitations cannot be completely eliminated) (121, 122). Preclinical evidence supports this approach. For instance, Zhang et al. developed a neoantigen−mRNA/DC vaccine using candidate neoantigens from mouse colon cancer (MC38) and evaluated its immune and antitumor effects (20). The results demonstrated that the neoantigen−mRNA/DC vaccine induced strong T cell immune responses and exhibited significant antitumor effects, effectively preventing tumor growth. A Phase I trial reported that three out of four patients treated with the neoantigen-targeting mRNA-4650 vaccine developed vaccine-induced CD8⁺ and CD4⁺ T-cell responses specific to the encoded neoantigens (5). Despite the limitations of a small sample size and tumor heterogeneity, the high response rate observed in three out of four cases preliminarily validates the technical feasibility of the TSA mRNA vaccine platform. These findings also underscore the necessity of utilizing biomarkers to identify patient subgroups in larger subsequent trials, thereby advancing the development of precision immunotherapy.

3.5. mRNA vaccines encoding immunostimulatory factors

Beyond delivering antigens, mRNA technology offers the flexibility to encode immunostimulatory factors that potentiate vaccine-induced antitumor immunity. While mRNA-encoded immunostimulatory factors are also being explored as standalone immunotherapies for direct intratumoral injection, this section focuses on their role in mRNA vaccine formulations, where they are co-delivered with TAAs or TSAs to enhance and sustain immune responses. By encoding immunomodulatory proteins (e.g., cytokines, co-stimulatory ligands, receptors, or enzymes), these vaccine components directly activate immune function and strengthen the body’s defense against tumors (115). A key advantage of mRNA technology in vaccine design is the ability to co-deliver multiple immunostimulatory sequences alongside antigen-encoding sequences, thereby activating synergistic pathways that amplify antigen-specific immunity. The most commonly utilized cytokines include Interleukin-2 (IL-2), IL-12, and OX40 Ligand (123). Although initially developed as intratumoral immunotherapies, studies of IL-12 mRNA have provided mechanistic insights relevant to vaccine development. For instance, in a Phase I trial of an IL-12 mRNA (NCT05392699), dose-dependent increases in C-X-C Motif Chemokine Ligand 9 (CXCL9), CXCL10, and CXCL11 levels were observed in peripheral blood samples, indicating activation of the IFN-γ downstream signaling pathway. By day 7 post-treatment, increased CD8⁺ T cell infiltration and elevated Programmed Death-Ligand 1 (PD-L1) expression were observed within the patients’ TME (124). These findings suggest that incorporating immunostimulatory factors such as IL-12 into mRNA vaccines could help reverse local immunosuppression and enhance the presentation of vaccine-encoded antigens.

Notably, mRNA-encoded immunostimulatory factors can also reprogram the TME by polarizing tumor-associated macrophages (TAMs) from a pro-tumorigenic M2 state toward an anti-tumor M1 phenotype (125). M1-like TAMs promote anti-tumor immunity by activating T cells, producing inflammatory cytokines, and enhancing phagocytosis (126). Within vaccine formulations, these immune-stimulating factors function as potent molecular adjuvants, augmenting both the innate immune response and the antigen-specific adaptive response triggered by TAAs or TSAs. The presence of these immune-stimulating factors can further amplify the immune response triggered by TAAs/TSAs, facilitating more effective recognition and targeting of tumor cells by the immune system. For example, in a preclinical study, combining tumor lysate with an IL-23A mRNA vaccine significantly promoted systemic immune activation and demonstrated promising efficacy in murine models of CT26 colon carcinoma abdominal and lung metastasis (127).

4. Current clinical translational status and combination therapy strategies

Clinical translational research on mRNA vaccine therapies for solid tumors is advancing expeditiously, with explorations in the field of CRC also exhibiting a trend toward diversification.mRNA-based therapeutic approaches are increasingly emerging as a promising strategy to combat tumor immune evasion. However, the scientific rigor of clinical trial design and safety management, alongside tumor heterogeneity and immunosuppressive TME in CRC, remain significant current challenges. mRNA-based therapeutic approaches are increasingly emerging as a promising strategy to combat tumor immune evasion. However, the scientific rigor of clinical trial design and safety management, alongside tumor heterogeneity and immunosuppressive TME in CRC, remain significant current challenges (77, 128, 129). The diversity of ongoing clinical trials is summarized in the table below Table 2.

Table 2.

The ongoing clinical trial involving mRNA vaccines for colorectal cancer.

NCT number	Platform	Antigen type	Combinatorial agent(s)	Phase	Enrollment (Estimated)	Primary endpoint	Status
NCT06497010	Traditional mRNA	Personalized neoantigen	PD-1 inhibitors	Early Phase 1	40	RP2D	Recruiting
NCT06577532	Traditional mRNA	Mutant KRAS neoantigens	Toripalimab	Early Phase 1	56	DLT,safety,ORR	Recruiting
NCT07182435	Traditional mRNA	Personalized neoantigen	/	Early Phase 1	36	DLT, safety, Immunogenicity of a personalized cancer vaccine	Not yet recruitingecruiting
NCT05359354	Traditional mRNA	Personalized neoantigen	PD-1 inhibitors	Not Applicable	36	MTD, DLT, Safety	Unknown status
NCT05940181	Traditional mRNA	Personalized neoantigen	Sintilimab	Not Applicable	9	MTD, DLT, Safety	Unknown status
NCT05949775	Traditional mRNA	Personalized neoantigen	Stintilimab	Not Applicable	20	PFS	Not yet recruiting
NCT06195384	Traditional mRNA	Personalized neoantigen	/	Phase 1	30	DLT	Recruiting
NCT05942378	Traditional mRNA	Personalized neoantigen	Adebrelimab	Phase 1	30	RP2D, Reaction of antigen-specific T cells	Unknown status
NCT07067385	Traditional mRNA	Personalized neoantigen	Stinlimab	Phase 1	40	Safety, Immunogenicity of a personalized cancer vaccine	Recruiting
NCT07245901	circmRNA	FAM53B-219aa	Toripalimab	Phase 1,2	60	Safety, RP2D, MTD	Not yet recruiting
NCT05141721	self-amplifying mRNA	Personalized neoantigen	heterologous chimpanzee adenovirus vaccine, Atezolizumab, Ipilimumab, Fluoropyrimidine plus leucovorin, Bevacizumab	Phase 2,3	700	ctDNA, PFS	Active, not recruiting

RP2D, recommended phase 2 dose; DLT, dose-limiting toxicity; ORR, objective response rate; MTD, maximum tolerated dose; ctDNA, circulating tumor DNA; PFS, progression-free survival.

4.1. Patient stratification: biomarkers for predicting CRC vaccine responsiveness

Personalized vaccines targeting tumor neoantigens are one of the core directions (such as NCT06195384, NCT05949775, etc.). This trend reflects the concept of precision medicine, but it also faces technical challenges such as standardization of neoantigen screening, HLA restriction analysis, and clonal assessment (130). In terms of clinical trial design, such studies usually require the use of composite endpoints: in the early stage of exploration, the main focus is on safety-related indicators (such as maximum tolerated dose, dose-limiting toxicity) (131). Furthermore, traditional response criteria—morpho- logical (such as RECIST), often fail to accurately and timely capture the therapeutic efficacy of immunotherapy, including personalized vaccines (132). Therefore, the application of alternative endpoints such as circulating tumor DNA (ctDNA) and immunogenicity of a personalized cancer vaccine (through ELISpot or TCR sequencing) as composite endpoints to precisely evaluate effectiveness has also become inevitable.

Secondly, combination therapy has been widely adopted, with its theoretical basis lying in the simultaneous activation of the immune system and the removal of its inhibitory mechanisms. Preliminary clinical studies have shown its potential (133, 134). In a study involving patients with advanced metastatic solid tumors (NCT03639714), a personalized vaccine regimen combining a chimpanzee adenovirus and a saRNA neoantigen vaccine was found to be safe and well-tolerated. Furthermore, improved overall survival (OS) was observed in several patients with MSS CRC (133). mRNA-5671/V941, administered either as a monotherapy or in combination with pembrolizumab, exhibited no dose-limiting toxicities (DLTs) in patients with colorectal cancer (NCT03948763), with signs of immune activation observed in a subset of participants. However, the majority of patients experienced varying degrees of adverse events, highlighting the complexity of its safety profile⁴. Consequently, how to distinguish the contributions of mRNA and other drugs such as immune checkpoint inhibitors (ICI) and how to avoid the superimposition of toxicity remain core issues (135, 136). To clarify the source of efficacy, a three-arm trial design (ICI monotherapy, vaccine monotherapy, and combination therapy) may be needed, or by tracking the clonal expansion of vaccine-specific T cells, to attribute the efficacy signal (137). The safety management of combination therapy is of vital importance. Based on the current clinical trial results, the background risks of mRNA vaccines themselves (such as fever and injection site reactions) are usually controllable (138). However, when combined with ICI, the theoretical risk of immune-related adverse events (irAEs) increases (135). Therefore, clear management strategies need to be formulated: closely monitor high-risk patients, establish irAE intervention plans, and explore preventive measures. Optimizing the timing of vaccination is key to enhancing therapeutic efficacy. For patients with postoperative minimal residual disease, the immune system has not yet been suppressed by advanced tumor burden, making it an ideal window for vaccination (139). Additionally, Grippin and colleagues demonstrated that SARS-CoV-2 mRNA vaccines can temporarily reset the tumor-immune interface, transforming “cold” tumors into those responsive to Programmed Cell Death Protein 1 (PD-1)/PD-L1 blockade (140). This could be a crucial step towards precision oncology, with treatment designs based on “opportunity windows” centered around the combined use of mRNA vaccines and immune checkpoint inhibitors (19).

Third, novel platforms continue to emerge, with circular RNA-based vaccines (e.g., circFAM53B, NCT07245901) representing an innovative approach to enhance antigen expression stability and durability. This technological breakthrough is expected to achieve more sustained antigen expression and stronger immune memory (141–143). However, the potential for conditional immune activation raises concerns about their safety and controllability (144). Early trials need to closely monitor excessive inflammatory responses such as cytokine storms and assess long-term risks.

In summary, the current trend in mRNA vaccine clinical trials is not only reflected in technological iterations but also in the in-depth consideration of design scientificity and safety. Future research should integrate biomarker stratification, precise endpoint selection, reasonable timing arrangements, and proactive toxicity management to promote the clinical translation of this therapy in colorectal cancer (145).

4.2. Biomarkers of vaccine reactivity in colorectal cancer and patient stratification

The clinical efficacy of mRNA vaccines in triggering anti-tumor immunity varies among different types of colorectal cancer patients. This heterogeneity highlights the importance of predictive biomarkers in patient selection and in guiding the development of rational combined treatment regimens (88).

4.2.1. MSI-H/dMMR versus MSS/pMMR

MSI-H/mismatch repair-deficient (dMMR) and MSS/proficient Mismatch Repair (pMMR) are the most fundamental molecular subtypes in colorectal cancer, which have a profound impact on clinical immunotherapy (146).

MSI-H/dMMR CRC, encompassing approximately 15% of all cases (with a higher prevalence in early-stage and right-sided tumors), is characterized by a defective DNA mismatch repair system (146). MSI-H/dMMR type colorectal cancer, due to its high antigen load and immune infiltration characteristics, is considered a promising candidate population for mRNA vaccine strategies (19).

In addition, the TME of MSS CRC typically exhibits immunosuppressive characteristics. Nevertheless, these hurdles do not preclude the therapeutic potential of vaccine-based strategies. Emerging clinical data suggest that mRNA-based platforms, particularly when targeting individualized neoantigens, hold promise for this patient population (133). By remodeling the TME, such integrated strategies can synergize with vaccines to elicit durable and robust immune responses.

Therefore, MSI status plays a crucial screening role in patient stratification. However, it must be acknowledged that there is heterogeneity within the MSI groups. For instance, MSH2/MSH6-deficient tumors typically exhibit a higher average Tumor mutational burden (TMB) than MLH1/PMS2-deficient tumors; MSS tumors with POLE/POLD1 mutations have an ultra-high mutation phenotype (usually over 100 mutations per megabase) (147, 148). This heterogeneity within the MSI groups further highlights the importance of other biomarkers, such as absolute TMB values, clonality of neoantigens, or specific genetic drivers of MMR deficiency, in optimizing patient selection and the efficacy of personalized vaccine approaches.

4.2.2. Tumor mutation burden, neoantigen quality, and HLA constraints

TMB provides a continuous and quantitative measure of genomic instability, which can further refine the aforementioned stratification. TMB, typically defined as the total number of somatic nonsynonymous mutations per megabase, serves as a proxy for the potential neoantigen pool (149). However, the relationship between TMB and vaccine responsiveness is not linear; a high mutational burden does not guarantee the presence of immunogenic epitopes, nor does it cover the complex interactions between antigen presentation and immune evasion. Therefore, the entire process from antigen generation to presentation needs to be considered. Evidence suggests that a high mutational burden does not consistently predict the presence of immunogenic epitopes, nor does it account for the complex interplay between antigen presentation and immune evasion mechanisms (150). Therefore, a multidimensional assessment of the entire process—spanning from antigen generation to presentation—is essential to better understand and predict vaccine efficacy.

The clonality and immunogenicity of neoantigens are also crucial for effectively activating immune responses (151). CRC exhibits extensive branched evolutionary features, with scarce clonal neoantigens and mainly subclonal neoantigens (152). Clonal neoantigens can drive complete tumor clearance; while subclonal neoantigens are prone to causing immune escape (153). In addition, the prediction of the immunogenicity of neoantigens and the actual expression of the mutated genes is also crucial for ensuring that neoantigens can be truly presented. The neoantigens expressed by mRNA vaccines are presented by HLA. However, approximately 21% to 28% of colorectal cancer patients have HLA-Loss of heterozygosity (HLA-LOH), which leads to the interruption of the antigen presentation pathway (153). This is also associated with a large number of subclonal neoantigens (154). Therefore, it is necessary to verify the quality of MHC expression, and ideally, new epitopes should be selected based on the corresponding MHC expression.

4.2.3. Immunosuppressive features of the tumor microenvironment

CRC TME usually has immunosuppressive effects, and its complex composition determines the efficacy of vaccine-induced T cell responses.From the perspective of the transcriptome, CMS classification further refines the immune phenotype of the TME. For instance, in CMS1 tumors, there is a dense infiltration of T cells, but the high expression of checkpoints such as PD-1 and TIM-3 reflects the exhausted state of T cells (155), suggesting the necessity of combining ICI when applying mRNA vaccines. In the CMS4 subtype, activated CAFs deposit a large amount of extracellular matrix, and patients with CAF enrichment may benefit from TGF-β inhibitors or matrix remodeling drugs (156). Mechanistically, this can collaborate with mRNA vaccines. In addition, CMS2/3 tumors are typically classified as immune desert subtypes with minimal T cell infiltration. For such “cold” tumors, inducing immunogenic cell death via chemotherapy could be a crucial step to transform them into “hot” tumors, thereby enhancing the efficacy of subsequent vaccination (157). It is also important to note that most CRC tumors exhibit intratumoral heterogeneity, which highlights the significance of multi-point dynamic detection and more precise TME biomarkers (such as T-cell exhaustion scores and CAF enrichment levels) in the clinical application of mRNA vaccines and combination therapy (158, 159).

5. Future directions

mRNA vaccines have great potential in the field of immunotherapy for colorectal cancer. However, further research and clinical trials are essential to optimize their efficacy, safety, and long-term benefits. Potential future directions and advancements include:

5.1. Optimizing mRNA vaccine design

mRNA can induce cytokine production by activating innate immune responses. However, excessive or sustained cytokine release may lead to severe side effects, including autoimmune reactions, and potentially interfere with specific immune responses against vaccine antigens (160). Excessive innate sensing may impair translation and shorten expression duration, whereas excessive suppression of intrinsic immunostimulatory signals may weaken dendritic cell activation and T-cell priming (161). Therefore, next-generation mRNA design should focus on fine-tuning, rather than eliminating, innate immune stimulation. Optimization at the RNA level should extend beyond nucleoside substitution to include coordinated engineering of the 5′ cap, untranslated regions, poly(A) tail, codon usage, and RNA secondary structure, with the goal of controlling not only expression intensity but also expression kinetics and immunological outcome. Additionally, the IVT preparation process often generates double-stranded RNA (dsRNA) by-products (162). These dsRNA impurities can activate intracellular immune sensing pathways, such as upregulating protein kinase R and oligoadenylate synthase, thereby triggering IFN-I-mediated immune responses that lead to rapid degradation of mRNA and compromise vaccine efficacy (163). While the removal of dsRNA via High-Performance Liquid Chromatography (HPLC) or RNase III enzymatic digestion is technically feasible in industrial settings, these downstream purification methods face significant challenges in terms of scalability, high operational costs (163). Therefore, future research should prioritize upstream process optimization—such as refining IVT parameters and engineering more precise RNA polymerases—to mitigate dsRNA formation at the source, thereby enhancing both the purity and potency of mRNA vaccines. Finally, delivery systems should be optimized not only to protect mRNA from degradation, but also to improve tissue distribution, antigen-presenting cell targeting, repeated-dose tolerability, and safety (164). Thus, future advances will likely depend on integrated strategies that combine RNA engineering, immune modulation, and precision delivery to generate more potent and clinically translatable mRNA cancer vaccines.

5.2. Innovations in mRNA vaccine storage methods

The development of new technologies is crucial for stabilizing vaccines while addressing some limitations associated with traditional freeze-dried storage (78). Spray drying and freeze-drying technologies provide a practical technical path for the room-temperature stable storage of mRNA preparations (165). For example, mRNA-1273 can be stored in a -20 °C freezer for 6 months, and can be kept in a refrigerated environment at 2 °C to 8 °C for one month (166). The focus of future research will be on optimizing the formulation to further extend the preservation of mRNA. In particular, the development of non-freezing liquid formulations will have a profound impact on the widespread availability of mRNA vaccine preparations in resource-poor regions. In addition, the establishment of a long-term stability evaluation system in the dried state, as well as the development of universal freeze-drying processes for different mRNA sequences (such as self-amplifying mRNA), will also be important directions for subsequent research.

5.3. The choice of injection route

The injection route is crucial for the translation efficiency of target proteins and the distribution of mRNA-based cancer vaccine. Intravenous injection remains the most common route for mRNA-based cancer vaccines in current clinical trials and effectively targets multiple lymphoid organs. However, this method may also lead to off-target effects such as systemic inflammatory responses (167). Meanwhile, subcutaneous administration is widely used in early-stage cancer RNA vaccines, such as CV9103. This approach utilizes the high density of dendritic cells and abundant lymphatic vessels in the skin to induce an effective immune response, but it may induce local side effects (168). Preclinical and clinical studies have explored various injection routes, including intravenous, subcutaneous, intradermal, intratumoral, and intranodal injection. Nevertheless, comprehensive experiments are still needed to fully exploit the great potential of mRNA-based cancer vaccines.

5.4. Combination therapy

The combined design of mRNA vaccines and immunotherapy has shown initial success. Through the combination of molecular pathways and immunodynamics, synergy is achieved, such as mRNA vaccines inducing the “interferon window” through type I interferons to synergize with the combination of immune checkpoint inhibitors, achieving the maximum efficacy (140). Future research should focus on clinical validation and combined administration under multimodal temporal schemes.

5.5. Biomarker development

As the clinical research on CRC mRNA vaccines progresses, future studies should focus on the dynamic biomarkers of the immune response driven by the vaccine. Since this therapy exerts its anti-tumor effect by activating the immune system, its clinical efficacy may not be reflected by conventional imaging evaluations (such as possible pseudo-progression (where the tumor temporarily enlarges) or good therapeutic effects observed in cases with a relatively mild disease burden but without obvious imaging remission) (169). For instance, ctDNA is a non-invasive and sensitive dynamic monitoring marker; a decline in its levels during treatment has shown potential correlation with prolonged overall survival, especially helpful for precisely stratifying molecularly remitted patients with stable disease as shown by traditional imaging (133). Biomarkers facilitate patient stratification, treatment monitoring, and assessment of treatment response, thereby enabling personalized treatment strategies.

6. Conclusions

With the implementation of numerous clinical trials, mRNA vaccines have shown promising early-stage results, demonstrating their safety, feasibility, and potential efficacy in the treatment of CRC. As a modular and customizable platform, these vaccines can elicit tumor-specific responses through various pathways while also modulating the tumor microenvironment. This positions them as a highly promising therapeutic approach. In conclusion, continuous research and progress in this field are expected to transform CRC immunotherapy, providing patients with more effective and personalized treatment options.

Glossary

CRC

Colorectal cancer

mRNA

messenger RNA

IVT

in vitro transcription

APCs

antigen-presenting cells

HLA

human leukocyte antigen

DC

dendritic cell

DTH

delayed-type hypersensitivity

CEA

carcinoembryonic antigen

IFN-γ

interferon-gamma

Aex

exosomes from tumor ascites

GM-CSF

granulocyte-macrophage colony-stimulating factor

MHC

major histocompatibility complex

TME

tumor microenvironment

VEGF

vascular endothelial growth factor

CAFs

tumor-associated fibroblasts

TGF

transforming growth factor

NK cell

natural killer cell

TAAs

tumor-associated antigens

TSAs

tumor-specific antigens

TAMs

tumor-associated macrophages

CTLs

cytotoxic T lymphocytes

saRNA

Self-amplifying RNA

circRNA

Circular RNA

MSS

microsatellite-stable

taRNA

trans-amplified RNA

MSI-H

Microsatellite Instability-High

PAMP

pathogen-associated molecular pattern

PRR

pattern recognition receptor

TLR7

Toll-like receptor 7

ssRNA

single-stranded RNA

MYD88

myeloid differentiation factor 88

Treg

regulatory T cells

ICI

immune checkpoint inhibitors

irAE

immune-related adverse events

dMMR

mismatch repair-deficient

pMMR

proficient Mismatch Repair

TMB

Tumor mutational burden

PD-1

Programmed Cell Death Protein 1

PD-L1

Programmed Death-Ligand 1

CXCL9

C-X-C Motif Chemokine Ligand 9

IL-2

Interleukin-2

LNP

lipid nanoparticles

ctDNA

circulating tumor DNA

Funding Statement

The author(s) declared that financial support was received for this work and/or its publication. This work was supported by the National Natural Science Foundation of China (grant number: 82172009), Natural Science Foundation of Shanxi Province (grant number: 202103021224357), Shanxi Provincial Key Laboratory of Integrated Traditional Chinese and Western Medicine for Digestivel Diseases (grant number: zyyyjs2024016).

Author contributions

TW: Visualization, Conceptualization, Investigation, Writing – review & editing, Writing – original draft. YM: Methodology, Conceptualization, Investigation, Visualization, Data curation, Writing – original draft, Software. SL: Data curation, Visualization, Conceptualization, Writing – original draft, Software, Investigation. HW: Writing – original draft, Visualization, Software, Investigation. MC: Conceptualization, Writing – review & editing, Funding acquisition, Project administration, Validation, Supervision. ZH: Writing – review & editing, Project administration, Supervision, Funding acquisition, Validation.

Conflict of interest

The author(s) declared that this work was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Generative AI statement

The author(s) declared that generative AI was used in the creation of this manuscript. During the preparation of this manuscript, the authors used DeepSeek-V3.2 for the purposes of Language polishing.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.