mRNA-based melanoma vaccines targeting gp100 and TRP2 macromolecules

Abstract

Melanoma, an aggressive form of skin cancer, is responsible for more than 80% of skin cancer mortality, despite representing only about 1% of all skin cancer cases. Against traditional treatments, a new generation of immunotherapies, including cancer vaccinations, has gained increasing attention for durable remission in advanced stages. Among them, Messenger RNA (mRNA) vaccines have arisen as a promising immunotherapy, with the ability to encode tumor-specific antigens to stimulate immune responses. In melanoma, shared melanocytic differentiation antigens such as glycoprotein 100 (gp100) and tyrosinase-related protein-2 (TRP-2) are targets due to their recurrent expression and immunogenic epitopes. In this review, we aimed to explain the mRNA vaccine design and delivery strategies with an antigen-centered focus on gp100 and TRP-2, emphasizing their design, immune interaction, preclinical and clinical utilization by classifying different types of studies. In vivo evidence demonstrated how LNP-based mRNA vaccines induce cytotoxic CD8⁺ T cell responses. mRNA vaccines have also demonstrated clinical potential in patients with melanoma; however, there are only a few clinical trials in phases I and II, where mRNA vaccines have been tested. Besides wide application of mRNA vaccines, there are Key challenges including delivery-related reactogenicity and safety considerations, especially in combination regimens. At the end of this review, we indicated computational Immunoinformatics in epitope prediction and design of immunotherapies targeting melanoma and Artificial Intelligence (AI) technologies. This review highlights the potential of mRNA vaccines in melanoma management, concentrating on ongoing improvements and future directions to address current challenges and improve therapeutic efficacy.

Introduction

Melanoma is an aggressive skin cancer that originates from melanocytes in the basal layer of the epidermis. [1, 2]. Melanoma can be exhibited as metastatic or non-metastatic [3]. Although melanoma accounts for only about 1% of all skin cancer cases, it causes over 80% of deaths associated with skin malignancy [4–6]. Ultraviolet radiation of sunlight, by accumulating genetic mutations in melanocytes, is the main environmental risk factor for their malignant transformation [7].

For decades, traditional treatments such as surgery, chemotherapy, and radiotherapy have been utilized as first-line treatments for melanoma [8]. Surgical resection, as the first line of treatment, is mainly used in patients up to the second stage of melanoma. For patients with higher stages of melanoma and metastasis, survival rates decrease significantly, and surgery alone has limited therapeutic potential [9–11]. Despite the radioresistant nature of melanoma, radiotherapy is used as a primary treatment in patients who are unable to undergo surgery. Patients at high risk of recurrence or those diagnosed with metastatic melanoma may still be eligible for radiotherapy [10, 11]. Chemotherapy, as one of the therapeutic options for melanoma management, lacks specificity for malignant cells and results in low drug accumulation in the tumor microenvironment [11–13]. As a result of the limited therapeutic advantages of traditional treatments and their significant side effects, novel targeted therapies have emerged to enhance melanoma management. Approaches such as immune checkpoint inhibitors (ICIs) have improved outcomes in many patients. However, despite their advances, a lot of patients fail to achieve durable remission or eventually experience relapse. For instance, a significant percentage of patients are unable to respond to ICIs due to innate and acquired resistance [14]. Given the increasing incidence of melanoma, demand for efficient treatment modalities still remains [15]. Novel immunotherapies, including next-generation cancer vaccines, have gained increasing attention [16]. Application of vaccines in oncology encompasses a broad range of platforms, including peptide- and protein-based vaccines, viral vectors, DNA plasmids, whole-cell or tumor lysate preparations, and dendritic cell (DC)–based approaches, each with distinct strengths and limitations regarding antigen breadth, immunogenicity, and manufacturing complexity. In this context, messenger RNA (mRNA) vaccines have gained considerable attention because they are proving capable of rapid deployment and potent immunogenicity [17]. The mRNA vaccines present an innovative preventive approach, providing genetic instructions directly to cells to facilitate accurate protein synthesis and elicit strong immune responses to fight cancer effectively [18, 19]. mRNA vaccines provide transient, non-integrating antigen expression and can be designed to encode either full-length proteins or selected epitopes, thereby supporting both major histocompatibility complex (MHC)I and II antigen presentation. With recent progress, particularly lipid-based delivery systems that are compatible with scalable production, mRNA platforms have become increasingly relevant to melanoma vaccine development. These approaches can be pursued either as “shared-antigen” vaccines targeting recurrent melanoma-associated antigens such as gp100 and TRP-2, or as individualized vaccines built around patient-specific neoantigens [18, 20]. In the present review, we aim to provide comprehensive insight into the mRNA vaccine platform and its design, examine how mRNA vaccines engage the immune system in melanoma, and summarize current therapeutic approaches. Finally, we discuss strategies to maximize the therapeutic potential of mRNA vaccines and highlight current limitations and future directions in melanoma management.

mRNA role in melanoma

Changes in how the mRNA is processed, structured, and regulated as a single-stranded molecule significantly impact gene expression patterns that help tumors form, grow, and bypass the immune system [21]. Alternative processing, such as splicing or polyadenylation, generates different mRNA isoforms with distinct untranslated regions (UTRs). In melanoma, dysregulated splicing changes the stability, export, and translation efficiency of mRNA [22]. Different RNA-binding proteins (RBPs) bind to mRNA untranslated regions or sequences to control the stability, location, and efficiency of translation of mRNA. In melanoma, the IGF2BP family members (IGF2BP1/2/3) are RBPs that cause cancer. Higher levels of IGF2BP1/3 in melanoma are linked to tumors that are more immunogenic and worse patient outcomes. Tumor cells that do not have IGF2BP1 respond better to interferon and are more likely to respond to anti–PD-1 therapy. Higher levels of IGF2BP1/3 in patient samples are linked to resistance to immunotherapy and lower survival rates. Because of this, RBPs like IGF2BPs keep and translate mRNAs that help melanoma cells grow and avoid the immune system [23]. Epitranscriptomic changes, such as N6-methyladenosine (m6A) methylation, are an important part of how mRNA is controlled in melanoma [24, 25]. In melanoma, enzymes like METTL3 (a methyltransferase) and FTO (a demethylase) are often out of balance. Studies have shown that overexpressing METTL3 makes pro-tumorigenic transcripts more stable and easier to translate, which helps the tumor evade immune detection and resist treatment [25, 26]. A recent study shows that m6A marks not only affect how fast melanoma cells grow, but they also control how the cells respond to immune checkpoint drugs [26, 27].

Non-coding RNAs, including circular RNAs (circRNAs) and long non-coding RNAs (lncRNAs), are essential for the post-transcriptional regulation of mRNA in addition to protein-based regulators [28]. Certain circRNAs in melanoma improve mRNA stability by acting as sponges for microRNAs that would normally eliminate carcinogenic mRNAs [29, 30]. Melanoma progression and metastasis are facilitated by the lncRNA DANCR's interaction with RNA-binding proteins, which increases the translation of transcripts associated with invasion and the epithelial-mesenchymal transition [31].

In melanoma, oncogenic signaling pathways all lead to the translational machinery. Changes that activate BRAF or NRAS make the MAPK pathway work too hard, while changes that make PTEN or mutations make the PI3K/Akt/mTOR pathway work better. Both pathways help cap-dependent mRNA translation. When MNK/eIF4E and PI3K/mTOR are both turned off at the same time, melanoma cells grow much less quickly. This shows that these pathways are responsible for making oncogenic proteins [32].

When cells are under stress, like when they do not get enough food or are taking medication, stress granules form in the cytoplasm to temporarily protect some mRNAs from being destroyed [33]. In melanoma, these granules keep only the transcripts that are important for cell survival and growth, which allows tumor cells to resist targeted therapy. Stress granules that are made or broken down in the wrong way have been linked to treatment resistance and worse clinical outcomes [34]. Using synthetic mRNA as a treatment is a new idea because mRNA is significant for the growth of melanoma and the immune system's ability to fight it. Recent advances have made it possible to use modified mRNA as a platform for vaccines against tumor-specific antigens [35].

mRNA-based vaccines

mRNA-based vaccines: mechanisms

Following intramuscular administration, mRNA vaccines transfect muscle cells, epidermal cells, and tissue-resident immune cells, including Antigen-presenting cells (APCs) such as DCs, macrophages, and secondary lymphoid organs [36]. Inside host cells, the mRNA is recognized by the innate immune system through endosomal receptors such as TLR3, TLR7, and cytosolic receptors, including 1RIG-I, MDA5, NOD2, and PKR. These receptors detect single- and double-stranded RNA, triggering inflammatory responses and the production of type I interferons. Although these responses enhance innate immunity, excessive activation may reduce antigen expression (37). Therefore, vaccine mRNAs are modified to minimize immune overstimulation and improve antigen expression (38–41). In transfected non-immune cells, antigen-encoding mRNA is translated into protein, which is subsequently degraded by proteasomes. The remaining antigenic epitopes are presented on the cell surface via MHCI. These MHC I – peptide complexes are recognized by cytotoxic CD8+ T lymphocytes, inducing cellular immune responses against the encoded antigen. In addition, transfected myocytes can activate bone marrow–derived DCs, which further stimulate CD8+ T cell responses [42]. Beyond MHC I presentation, transfected APCs can present antigens via the MHC II pathway, leading to activation of CD4+ T helper cells [43, 44]. Through the lymphatic system, residual vaccine components drain into lymph nodes, where they encounter monocytes and naïve T and B lymphocytes. As a result, transfection of lymph node–resident APCs leads to the activation of both cellular and humoral immune responses (Figure 2) [36, 45]. In response, the tumor microenvironment (TME) of melanoma may be resistant. As melanoma progresses, M2 macrophages are recruited more than M1 macrophages, and release cytokines, including tumor necrosis factor beta (TNF-β), cyclooxygenase-2 (COX-2), and interleukin-10 (IL-10), along with matrix metalloproteinases (MMPs). These cytokines, by destroying the extracellular matrix, provide more space for malignant cells to grow [46]. Meanwhile, myeloid-derived suppressor cells (MDSCs) activate various proinflammatory cytokines of TME, such as prostaglandin E2 (PGE2), granulocyte colony-stimulating factor (G-CSF), granulocyte-macrophage colony-stimulating factor (GM-CSF), and CCR5. By providing immunosuppressive molecules, such as nitric oxide synthase, reactive oxygen species, and arginase-1, MDSCs inhibit T-cell activation and elevate T-cell apoptosis and cell-cycle arrest [47–49]. Furthermore, regulatory T cells (Tregs) impair the function of effector T cells (Teff) and neutrophils, through releasing inhibitory cytokines such as IL-10, IL-35, and TGF-β. Tregs also reduce the efficacy of immune checkpoint inhibitors (ICI) by promoting immune tolerance, by lymphocyte-activation gene 3 (LAG-3), PD-1, and CTLA-4 [50–52]. Moreover, IL-10, VEGF, and TGF-β can alter the development of DCs and push them towards a tolerogenic state with weaker antigen presentation and higher production of immunosuppressive cytokines [53–55].

Fig. 2.

LNP-based transportation of mRNA vaccines for melanoma. Different types of LNPs, including liposomes, micelles, liposome-like nanoparticles, and nanostructured lipid carriers for the delivery of mRNA vaccines (A). LNPs disseminate in the blood circulation to passively target the neoplastic cells (B). LNPs containing mRNA are internalized into the neoplastic cells. Internalized LNPs are covered by plasma phospholipids and protonated. Afterward, the mRNA content escapes from the endosome carrier to disperse within the cellular cytoplasm. Following the escape, the mRNA is translated into a polypeptide chain and placed on the surface of the neoplastic cells. The placed polypeptide chain can be presented to T helper cells (C) DC-related: No Grade ≥ 3 AEs reported. Only mild (Grade 1/2) local skin reactions and flu-like symptoms IFN-α-2b-related: Grade 3 constitutional symptoms in 2/29 patients (7%)

mRNA-based vaccine: structure and vaccine design

Over the past decade, mRNA vaccines have been recognized as a futuristic treatment option (Figure 1). Compared to traditional cancer vaccines, mRNA vaccines offer a wider range of target options, including tumor-associated antigens (TAAs), tumor-specific antigens (TSAs), and tumor microenvironment (TME) antigens. Unlike peptide-based vaccines, these new targets are not restricted by the spatial limitation. mRNAs are enabled to encode multiple antigens or full-length proteins containing both MHCI and MHCII binding epitopes and boost humoral and cellular adaptive immune responses to strengthen anti-tumor immunity. mRNA vaccines do not mix with DNA, break down very easily, and do not have the potential to cause insertional mutagenesis [20, 56–61]. mRNA allows the genetic information from DNA to be translated into protein within the cytoplasmic ribosomes. The two primary types of mRNA used as vaccine candidates are non-replicating mRNA and self-amplifying mRNA. Both types consist of 5’ and 3’ untranslated regions (UTRs), an open reading frame (ORF) serving as the coding sequence, and a poly(A) tail. Self-amplifying mRNA carries an additional coding region within the ORF, enabling repeated intracellular RNA replication and thereby boosting antigen expression. These mRNA sequences are produced in vitro from a linearized DNA plasmid containing the target gene, resulting in mRNA vaccines that encode the selected antigen to elicit an immune response [62]. The 5’cap at the start of mRNA contains a modified guanine nucleotide called 7-methylguanosine (m⁷G). The first nucleotide (m7GpppN) is linked to m⁷G by three phosphate groups (ppp). m7GpppN protects RNA from exonuclease, modifies pre-mRNA splicing, starts mRNA translation, and exports the nuclear mRNA to the cytoplasm, and plays an important part in innate immune system recognition of host mRNA from foreign mRNA [62–64]. Post-transcriptional modifications, including 2′-O-methylation (Cap 1: m7GpppN1m; Cap 2: m7GpppN1mN2m), enhance mRNA translation while preventing activation of endosomal and cytosolic sensors, including RIG-I and MDA5, which recognize foreign RNA. [64, 65]. Choice of the first transcribed nucleotide (A, m6A, G, C, or U) can influence protein output, with A, Am, or m6Am promoting higher expression. A study by Sikorski et al. indicated that mRNA with A, Am, or m6Am as the first nucleotide exhibited elevated luciferase protein expression, while mRNA having G or Gm leads to reduced luciferase expression [66]. The 5’ and 3’ UTRs regulate mRNA stability, ribosome recruitment, translation efficiency, and post-transcriptional modifications. They play critical roles in ribosome recognition, mRNA stability, accurate and efficient translation, and further post-transcriptional modifications [67]. The poly(A) tail, a polyadenylated section at 3′, is an effective factor for the lifespan of the mRNA. The mammalian cells’ poly(A) tails, with about 250 nucleotides (nt) longer length, gradually get shorter over time. So, for the production of mRNA vaccines with a desired half-life, 100 nt is added to the poly(A) tail [68, 69]. All these different changes to the mRNA backbone and untranslated regions make RNases less effective, make the structure more stable, and make it much easier to translate.

Fig. 1.

Overview of mRNA-based vaccine benefits

mRNA-based vaccine: delivery format optimization

Following in vitro transcription (IVT) and administration, mRNA must reach the cytoplasm of target cells to enable antigen translation. However, owing to its large molecular size (10⁴–10⁶ Da), strong negative charge, and intrinsic susceptibility to extracellular and intracellular nucleases, mRNA alone is incapable of crossing cellular membranes and is at risk of rapidly degrading. Consequently, efficient and protective delivery systems are crucial for a successful mRNA vaccine.

Current mRNA delivery strategies can be broadly classified into ex vivo (also known as in vitro) and in vivo approaches. Ex vivo delivery typically involves the transfection of dendritic cells (DCs) using methods such as electroporation, gene gun–mediated delivery, or ex vivo transfection. Among them, electroporation is the most used due to its high transfection efficiency. Transfected DCs, loaded with either tumor-associated antigen (TAA)–encoding mRNA or total tumor RNA, are sent back to the patient to induce antigen-specific immune responses [70, 71]. This approach offers significant control over antigen presentation and predominantly induces a cell-mediated immunity. However, its clinical application is limited by labor-intensive character, high production costs, and challenges in large-scale GMP manufacturing [70–74].

In contrast, in vivo delivery strategies involve the direct administration of mRNA to transfect immune or non-immune cells with lipids or transfection agents, such as lipid nanoparticles (LNPs) [57, 62]. While direct injection of naked mRNA represents a rapid and cost-efficient approach, its main drawback is the short extracellular half-life of unprotected mRNA. To overcome this limitation, multiple physical and synthetic delivery systems have been developed, including gene guns, electroporation, virus-like particles, liposomes, lipoplexes, and cationic polymer-based complexes. These platforms protect in vitro–transcribed (IVT) mRNA from RNase-mediated degradation, facilitate cellular uptake, and enhance intracellular delivery efficiency [74–78] (Table 3).

Table 3.

Comparative features of mRNA vaccine platforms and antigen strategies in melanoma

A
Feature	DC-based mRNA vaccines	Direct LNP-based mRNA vaccines	References
Antigen delivery	Ex vivo loading of autologous DCs	In vivo uptake by APCs after injection	[21, 144]
Antigen presentation	Highly controlled, efficient MHC I/II presentation	Variable, formulation-dependent	[20, 41]
Manufacturing	Patient-specific, labor-intensive	Scalable, off-the-shelf	[26, 37]
Clinical scalability	Limited	High	[37]
Use in gp100/TRP-2 studies	Predominant in early melanoma trials	Emerging, limited clinical data	[211]

B
Feature	Shared antigens (gp100/TRP-2)	Personalized neoantigens	References
Antigen availability	Pre-defined, off-the-shelf	Patient-specific	[41, 211]
HLA restriction	Often HLA-A*02:01	Individual HLA profile	[211]
Tumor escape risk	Higher (antigen loss, immune editing)	Lower (but not absent)	[41, 211]
Clinical efficacy	Modest as monotherapy	Higher, esp. with ICI	[41]
Manufacturing timeline	Rapid	Longer, complex	[41]

Among these technologies, lipid nanoparticles (LNPs) have emerged as the most clinically advanced and widely utilized mRNA delivery platform. LNPs are nanoscale lipid formulations, typically ~100 nm in diameter, composed of four principal components: ionizable lipids, cholesterol, phospholipids, and lipid-linked polyethylene glycol (PEG). The ionizable lipids facilitate the endosomal escape of mRNA into the cytoplasm. These lipids also prolong extracellular mRNA stability and enhance tissue accumulation following administration [79–81]. However, certain lipid formulations may activate toll-like receptor (TLR) pathways, contributing to inflammatory responses and dose-dependent cytotoxicity [57, 82].Polyethylene glycosylated (PEGylated) lipids play an important role in improving LNP stability. Nevertheless, PEG density must be carefully optimized, as excessive PEGylation can stop cellular uptake and endosomal escape. Moreover, repeated administration of PEGylated LNPs may induce anti-PEG antibodies, leading to hypersensitivity reactions and accelerated blood clearance, which can compromise therapeutic consistency [83–86]. Cholesterol and phospholipids contribute to the structural integrity, membrane fusion capability, and phase transition behavior of LNPs. As endogenous membrane components, they are generally well tolerated and minimally immunogenic, supporting the favorable safety profile and scalability of LNP-based platforms [57, 87–89].

While LNPs have become the popular platform for mRNA vaccine delivery, these synthetic nanoparticles (NPs) face multiple challenges, such as inflammation and toxicity associated with ionizable lipids, and limited control over biodistribution [90]. To overcome these limitations, biomimetic NPs have been developed. Biomimetic nanoparticle platforms, through replicating key structural and functional characteristics of native cell membranes, enhance immune evasion, extend systemic circulation, and increase APC uptake. These biomimetic membrane coatings originate from natural cell sources or engineered analogues, including red blood cell (RBC) membranes, exosomes, cancer-associated fibroblast (CAF) membranes, virus-like particles (VLPs), and immune cell membranes, such as artificial antigen-presenting cells (aAPCs) [91, 92] APC engagement with biomimetic NPs depends on their coating composition. Their cDC1 uptake can boost MHC I presentation and CD8⁺ T-cell activation, while macrophage uptake can lead to antigen degradation or tolerance [93, 94]. Due to the absence of MHC molecules and the presence of CD47 in RBC membrane-coated nanoparticles, they can escape the reticuloendothelial system (RES) and prolong circulation time, and decrease unwanted innate immune activation [92, 95, 96] Exosome-coated nanoparticles contain a group of adhesion molecules, including integrins and tetraspanins that help them in better targeting and better uptake by APCs, especially DCs [92, 97–99]. Exosomes often preserve the targeting characteristics of the origin cells they came from. Therefore, tumor-derived exosomes (TEXs) tend to migrate back toward the tumor environment or draining lymph nodes. TEXs may also carry immunosuppressive molecules such as programmed cell death protein ligand 1 (PD-L1); hence, their combination with immune adjuvants can enhance vaccine performance. On the contrary, DC-derived exosomes (Dexs) already carry immunostimulatory MHC-peptide complexes and co-stimulatory molecules that enable them to activate both CD4⁺ and CD8⁺ T cells without the need for live DCs. The ability of specific targeting in exosomes enhances delivery of therapeutic cargo to the tumor microenvironment (TME) and secondary lymphoid organs, and highlights the potency of nanovaccine-mediated antitumor immune responses . Although exosome-coated nanovaccines improved delivery efficiency and targeting capability, they face multiple obstacles in clinical translation, such as source-cell heterogeneity, difficulties in large-scale purification, GMP standardization, and uncertainties in their immunological behavior [106, 107]. The development of artificial membrane-coated nanoparticles may address many of these limitations by optimizing lipid composition and ligand density, modulation of immunological activity, and scalable GMP-compatible manufacturing [108, 109].

Beyond formulation design, the route of administration critically influences mRNA vaccine biodistribution, immune activation, and therapeutic efficacy. Intramuscular and intradermal delivery are the most commonly employed routes, as they enable efficient uptake by APCs ,and promote sustained immune responses. In contrast, intravenous administration results in predominant hepatic uptake due to first-pass liver clearance, limiting its practicality for routine vaccination. Consequently, systemic delivery is typically reserved for specialized therapeutic applications, while localized delivery routes are favored for cancer immunotherapy [57, 84].

mRNA vaccine in melanoma treatment

Currently, mRNA-based vaccines have gained attention in the management of melanoma. These vaccines encode three groups of antigens, known as tumor-associated antigens, neoantigens, and inflammatory mediators. Tumor-associated antigens based on tumor composition, location, and stage can be divided into melanocyte differentiation antigens, cancer-germline antigens, and Membrane-associated proteins [16, 110, 111]. Melanocyte differentiation antigens are present in both melanocytes and melanoma cells, and include tyrosinase, tyrosinase-related proteins (TRP-1 and TRP-2), melanocyte antigen (MELAN-A/MART-1), glycoprotein 75 (gp75), and glycoprotein 100 (gp100) [111–114]. Cancer-germline antigens are mostly expressed on germ cells, and their presence in adults may be a sign of melanoma. These antigens include multiple families, such as the melanoma-associated antigen family (MAGE-family), the BM antigen family (BAGE-family), the G-antigen family (GAGE-family), the synovial sarcoma family of the X chromosome breakpoint (SSX-family), and NY-ESO-1 [115, 116]. Furthermore, neoantigens are only expressed on tumoral cells. These mutations exhibit high immunogenicity, and they vary based on tumor type and patients. [117–120]. Expression of TAA, such as gp100 and TRP-2, is heterogeneous. Studies have demonstrated that tyrosinase is present in all lesions, even in lesions with no gp100 expression (121, 122). In contrast to the heterogeneous expression of cutaneous melanoma, melanocyte differentiation antigens are expressed very abundantly in uveal melanoma [123]. TRP-2 is expressed in primary cutaneous and mucosal melanomas, but has a lower incidence in metastatic lesions. TRP-2 has no expression in desmoplastic melanomas, like other melanocyte differentiation antigens [124] (Table 3).

Preclinical evidence of mRNA vaccine in melanoma therapy

Multiple applications of mRNAvaccines have been assessed in preclinical models of melanoma cancer in different settings (Fig. 1). Oberly et al. demonstrated that LNP-based mRNA vaccines effectively induce cytotoxic CD8⁺ T cell responses by delivering tumor-associated antigens gp100 and TRP2 to APCs [44]. Wang et al. developed a lipid-coated calcium phosphate (LCP) nanoparticle to simultaneously deliver mRNA encoding TRP2 and siRNA targeting PD-L1 into DCs. Through this approach, they enhanced antigen presentation and achieved the benefits of immune checkpoint inhibitors. simultaneously. In addition, Mannose-functionalized, PEGylated LCPs enabled targeted delivery to lymph node-resident DCs. By implementing this strategy in a B16F10 melanoma model, they showed robust cytotoxic T cell and humoral responses, leading to significant inhibition of both tumor progression and metastatic spread [125].

Previously, He et al. investigated the function of ovalbumin (OVA), a protein known to boost neoantigen recognition by cytotoxic T lymphocytes. The delivery of this antigen into the mouse models with B16-F10 cells significantly reduced tumor growth by 70% [126]. In 2022, Chen et al. indicated that 113-O12B, as a novel NLP targeting lymph nodes, enabled efficient mRNA delivery without active targeting ligands. As a result, enhancing CD8⁺ T cell activation and therapeutic efficacy in B16F10 melanoma mice. Compared to synthetic ALC-0315 LNP, 113-O12B showed superior lymph node accumulation, leading to stronger antitumor immunity and long-term memory immunity. A combination of TRP-2 peptide-encoding mRNA vaccines with anti–PD-1 therapy showed complete tumor regression in 40% of mice receiving treatment. Meanwhile, administration of the ovalbumin-encoding mRNA produced both preventive and therapeutic antitumor effects [127]. The available preclinical studies evaluating mRNA vaccines for melanoma treatment are collected in Table 1.

Table 1.

Preclinical studies on mRNA-based vaccines for melanoma therapy

Study	Targeted antigen	Vaccine composition	Experiment subject	Results	References
Oberli et al. 2017	gp100 and TRP2	Lipid nanoparticles -formulated mRNA	C57BL/6 mice (B16-F10 melanoma)	LNP-formulated mRNA vaccines induced elevated CD8⁺ T-cell responses, inhibited tumor growth, and prolonged survival of the treated model	[44]
Wang et al. 2018	TRP2	lipid-coated calcium phosphate nanoparticle-based mRNA combined with PD-L1-targeting siRNA	C57BL/6 mice (B16-F10 melanoma)	A combination of mRNA vaccine and antigen-specific checkpoint inhibitor delivered through transfected dendritic cells in the lymph nodes enhanced CD8⁺ T-cell proliferation and response, which controlled melanoma tumor growth	[125]
Zhang et al. 2021	TLR4	C1 lipid nanoparticles -formulated mRNA	C57BL/6 J, Tlr4−/− and STING cKO mice (B16 murine melanoma)	Delivery of mRNA with C1 lipid nanoparticles increased expression of IL-12 via TLR4 signaling, enhanced dendritic cell activation and CD8⁺ T-cell responses, and reduced tumor growth	[204]
Li et al. 2021	M30	CpG2018B and the mRNA-based neoantigen vaccine + lipid nanoparticle (LNP)	C57BL/6 mice (B16-F10 melanoma)	A combination of CpG-B ODN (CpG2018B) with an mRNA-based neoantigen vaccine intratumorally enhanced antitumor effects by activating TLR9-mediated immune responses. Interestingly, each of these components alone can turn cold tumors into hot ones and reduce their growth	[205]
Chen et al. 2022	OVA, TRP-2	113-O12B lipid nanoparticle -formulated mRNA	C57BL/6 mice (B16-F10 melanoma)	LN-targeted mRNA delivery induced strong CD8⁺ responses, tumor inhibition, memory response, and synergy with anti-PD-1 therapy	[127]

Clinical evidence of mRNA vaccine in melanoma therapy

mRNA vaccines have also demonstrated clinical potential in patients with melanoma (Fig. 2). There are a few clinical trials in phases I and II, where mRNA vaccines have been evaluated in melanoma patients. Patel et al.'s phase I/II clinical study evaluated the DNA vaccines encoding gp100 and TRP-2. These melanocytic differentiation epitopes are immunogenic and restricted to specific HLA alleles, including HLA-A*0201. Targeting gp100 and TRP-2 induces CD8⁺ and CD4⁺ T-cell responses. In this trial, tumor-free patients responded more strongly than patients with tumors. Moreover, T cell response is increased after repeated dosing of the vaccine [120]. Kyte et al. developed a personalized melanoma vaccine using transfected autologous DCs with patient-derived tumor mRNA, which was shown to be safe and feasible in a phase I/II trial. In this study, among 22 patients with advanced melanoma, specific T cell responses were obtained in nearly half of the evaluable participants [128]. A clinical trial by Weide et al. showed that 21 metastatic melanoma patients administered an mRNA vaccine, coding for Melan-A, Tyrosinase, gp100, Mage-A1, Mage-A3, and Survivin with or without KLH and GM-CSF modulated immune regulation by diminishing regulatory T cells and myeloid-derived suppressor cells, elicited antigen-specific T cell responses, were well-tolerated with no adverse events above grade II, and resulted in one complete clinical response [129]. Benteyn et al. showed that melanoma therapy in patients with mRNA electroporated dendritic cells (TriMixDC-MEL), involving DCs electroporated with with one of four mRNAs encoding a TAA (tyrosinase, MAGE-A3, MAGE-C2, or gp100), effectively induces functional CD8⁺ T cell responses in patients with systemic metastases of melanoma [130]. A phase I trial of TRP2 mRNA-electroporated autologous DCs in resected melanoma patients, conducted by Chung et al., demonstrated the vaccine’s safety and immunogenicity. Mild delayed-type hypersensitivity reactions occurred in all patients, with no adverse event beyond grade 1 toxicity. Post-vaccination, both CD4⁺ and CD8⁺ T cells showed enhanced cytokine secretion and cytotoxic marker expression. Increased T-cell receptor (TCR) repertoire clonality changes were more noticeable in patients who remained relapse-free [131].BNT111’s phase 1 trial, has also demonstrated a boost in both novel and pre-existing T-cell immune responses against four common TAA, including NY-ESO-1,Tyrosinase, Melanoma-associated antigen A3 (MAGE-A3), and Trans-membrane phosphatase with tensin homology (TPTE) [92, 132]

Early antigen vaccines targeting shared tumor-associated antigens (TAAs), such as gp100 and TRP-2, showed immunogenicity, but their response did not exhibit a durable clinical benefit. Most of the studies in Table 2 are early-phase trials and mainly report limited objective responses when mRNA vaccines are used alone. Complete responses were uncommon, but partial responses and stable disease were observed more often. This suggests that vaccine-induced immune responses by themselves were often not enough for long-term tumor control. This limited efficacy may stem from various reasons, including central tolerance. Since mRNA vaccines encode gp100 and TRP-2, which are shared self-antigens, the induced cytotoxicity against melanoma may be weakened by central tolerance [133]. Furthermore, expression of gp100 and TRP-2 varies between patients and within tumors. This heterogeneity can lead to immune escape, since there is not sufficient antigen for the induced T-cell response to attach to [134]. In this case, using ICI, such as CTLA4, PD-1, and PD-L1, in combination therapy with the mRNA vaccines can maintain induced immunological responses and enable the vaccine-induced T cells to proliferate and promote anti-tumor effects [35]. Ugur Sahin et al. study indicated that melanoma combination therapy with mRNA vaccine FixVac and PD-1 inhibitors can present a synergistic effect. This Combination therapy can even restore the lost drug sensitivity to ICI treatment. After a two-year follow-up, most of the patients who experienced partial remission or remained stable exhibited longer disease control [135]. Furthermore, the combination of mRNA vaccines with pembrolizumab can significantly reduce the risk of disease relapse [136]. As a result, these findings supported a new approach in the treatment of melanoma. In terms of safety, mRNA vaccines were well tolerated. Most reported adverse events were mild to moderate. Immune-related adverse events were rare and were more frequently described in studies where mRNA vaccination was combined with immune checkpoint inhibitors rather than used alone. Overall, the findingsindicate that mRNA vaccines show an acceptable safety profile and induce immune responses, but their clinical impact in melanoma improves when applied as part of combination strategies [137, 138].

Table 2.

Clinical studies on mRNA-based vaccines for melanoma therapy

Trial ID	Phase	Formulation	Route	Combination	Status	Enrolled patients	Vaccine-encoded antigens	Findings	Grade ≥ 3 adverse events	References
NCT01278940	I/II	Electroporated autologous DCs with tumor mRNA	i.d. and i.n	None	Completed	22	Autologous tumor mRNA	Vaccine-specific T cell responses were detected in 9 of 19 evaluable patients. Delayed-type hypersensitivity responses were observed in 8 of 18 patients	None	Kyte et al. [128]
NCT00204607	I/II	Protamine-protected mRNA encoding Melan-A, Tyrosinase, gp100, MAGE-A1, MAGE-A3, Survivin	i.d	± KLH, GM-CSF	Completed	21	Melan-A, Tyrosinase, gp100, MAGE-A1, MAGE-A3, Survivin	The vaccine was safe and well-tolerated. Antigen-specific T cell responses were observed in 2 of 4 patients, and one patient achieved a complete response among seven	None	Weide et al. [129]
NCT00978913	I/II	TriMix mRNA (CD40L, caTLR4, CD70) + tumor antigen mRNAs (Tyrosinase, gp100, MAGE-A3, MAGE-C2)	i.d	None	Completed	35	Tyrosinase, gp100, MAGE-A3, MAGE-C2	One patient out of 17 achieved a partial response, and five patients experienced stable disease* Stable disease reported in 5 of 17 patients	None	Wilgenhof et al. [206]
NCT01066390	I	TriMix (CD40L, caTLR4, CD70) + one of four tumor-associated antigens (Tyrosinase, MAGE-A3, MAGE-C2, gp100)	i.d. and i.n	None	Completed	14	CD40L, TLR4, CD70 + TAA	T cell-specific responses were detected in 11 of 14 patients in blood samples and 12 of 14 patients in tissue samples. Two patients achieved complete response, one patient had partial response, and four patients had stable disease	None	Benteyn et al. [130]
NCT01676779	II	TriMix mRNA + tumor-associated antigens (MAGE-A1, MAGE-A3, MAGE-C2, gp100, Tyrosinase, MelanA/MART-1) fused to HLA II-targeting sequence	i.d	IFN-α-2b	Completed	30	Autologous mRNA	The median relapse-free survival was 22 months, and the 4-year overall survival rate was 70%. Mild DC-related local skin reactions and flu-like symptoms were observed. Two patients (7%) experienced grade 3 constitutional symptoms due to IFN-α-2b	DC-related: No Grade ≥ 3 AEs reported. Only mild (Grade 1/2) local skin reactions and flu-like symptoms IFN-α-2b-related: Grade 3 constitutional symptoms in 2/29 patients (7%)	Wilgenhof et al. [207]
NCT01302496	II	TriMix mRNA + tumor-associated antigen mRNAs (Tyrosinase, gp100, MAGE-A3, MAGE-C2)	i.v. and i.d	Ipilimumab	Completed	39	Tyrosinase, gp100, MAGE-A3, MAGE-C2	The 6-month disease control rate was 51%. Complete response was observed in 20.5% of patients, and partial response in 17.9%. T cell activation was observed in 12 of 15 evaluable patients, which correlated with clinical response	36% grade 3–4 immune-related adverse events	De Keersmaecker et al. [208]
NCT 02285413	II	Electroporated monocyte-derived DCs with gp100 and tyrosinase mRNA	i.v. and i.d	Cisplatin	Completed	41	tyrosinase, gp100	TriMixDC-MEL as adjuvant therapy improved the 1-year disease-free survival rate (71% vs. 35%) and delayed non-salvageable recurrence. Four predictive biomarkers were identified via mRNA profiling	None	Boudewijns et al. [209]
NCT05264974	I	Encapsulated in DOTAP liposomes	i.v	None	Suspended	18	LAMP1-pp65 mRNA	Preliminary antigen-specific T cell activation was observed. The trial was suspended, and no efficacy data were published	NA	Doonan et al. [210]

mRNA vaccines encoding tumor-associated antigens (TAAs) have demonstrated the ability to induce immune responses; however, their clinical efficacy has often been limited by tumor heterogeneity and immune escape. These challenges provide both a conceptual and practical foundation for the development of personalized neoantigen vaccines as the next generation of immunotherapies. Unlike gp100 and TRP-2, neoantigens are tumor-specific antigens that result from specific mutations. Due to their immunity to central tolerance, they can induce strong CD4⁺ and CD8⁺ T-cell responses [134, 139, 140]. Personalized neoantigen vaccines can address intratumoral heterogeneity by targeting multiple patient-specific epitopes and reducing the risk of immune escape [141]. The KEYNOTE-942 trial demonstrated that a personalized mRNA neoantigen vaccine (mRNA-4157/V940) combined with pembrolizumab significantly prolonged the recurrence-free survival compared with anti-PD-1 monotherapy [142]. These findings change the road of antigen vaccination toward individualized mRNA vaccine strategies optimized for durable clinical benefit.

Different mRNA vaccine formats and antigen strategies have been investigated in melanoma, but direct comparisons between them are limited. To address this issue, Table 3 summarizes the main differences between commonly used mRNA vaccine platforms and antigen approaches in melanoma research. The table compares ex vivo dendritic cell–based vaccines with direct LNP-based mRNA delivery, with emphasis on antigen presentation, manufacturing complexity, and clinical scalability. It also outlines key features of unmodified, nucleoside-modified, and self-amplifying mRNA constructs, including their innate immune activation and protein expression profiles. In addition, the table contrasts shared melanoma-associated antigens, such as gp100 and TRP-2, with personalized neoantigen strategies, focusing on HLA restriction, tumor immune escape, and translational considerations. Overall, this comparison is intended to support a more balanced evaluation of current and emerging mRNA vaccine strategies in melanoma [20, 37, 143, 144].

Side effects of mRNA vaccine in melanoma treatment

Despite mRNA-based vaccines’ promising outcomes in the treatment of melanoma, their favorable safety profile, and cost-effective manufacturing process, they have potential side effects limiting their broader application [16, 72, 145–148]. The prevalence of mRNA vaccine side effects ranges between 80 to 90%, with the majority classified as mild to moderate in severity [149, 150]. Most of these mild to moderate adverse reactions are pain-related and associated with inflammation [151]. However, adverse effects, including allergic reactions, renal failure, cardiac failure, and infarction, remain potential risks. mRNA vaccines may also undergo rapid degradation and trigger adverse immune reactions, including anaphylaxis and cytokine storm, as a severe form of immunologic response. These can lead to systemic inflammation and potentially life-threatening conditions [152].

mRNA is an unstable and easily degradable structure, which needs a delivery system to facilitate its entry into the intracellular environment. Lipid nanoparticles (LNPs) have shown efficient mRNA delivery. However, based on recent research, the inflammatory features of LNP components in mRNA vaccine formulations are responsible for their inflammatory side effects in clinics [152, 153]. Based on studies by Genetech and BioNTech, it has been indicated that LNPs are key players in inducing the inflammatory cytokine IL-1 [151]. Maja et al.'s study tried to investigate the side effects of the LNP delivery system for human erythropoietin (hEPO) mRNA in monkey and rat models. rats exhibited elevated liver enzymes, changes in coagulation parameters, and increased levels of IFN-γ–induced proteins, whereas monkeys showed splenic necrosis, lymphocyte depletion, and mild, reversible activation of the complement system [154]. Bol et al. investigated a dendritic cell (DC)-based vaccination strategy using autologous DCs matured with a cocktail (BCG, Typhim, Act-HIB, prostaglandin E2) of prophylactic vaccines. This approach, referred to as VAC-DC, involved electroporation of the DCs with mRNA encoding the antigens gp100 and tyrosinase, followed by pulsing with keyhole limpet hemocyanin (KLH) protein. The study demonstrated that VAC-DC causes local and systemic grade 2 and 3 toxicity defined by the Common Terminology Criteria for Adverse Events (CTCAE). The side effects were predominantly induced by the BCG component in the maturation cocktail, which were either self-limiting or resolved within a brief course of systemic steroid therapy [155]. Comparative analysis of two therapeutic cancer vaccine trials in patients with advanced melanoma, the Tumor Lysate (TL) vaccine by Weber et al. and the mRNA-4157 vaccine by Carpenter et al., revealed notable differences in safety profiles. In the TL vaccine trial, treatment-related grade 3 adverse events as classified by CTCAE occurred in less than 1% of patients which included symptomatic anemia and hospitalization. Whereas, the mRNA trial reported grade 3 adverse events in 12% of patients, with fatigue being the most frequently observed. A potential contributing factor to the higher incidence of adverse events in the mRNA trial is the difference in the use of immune checkpoint inhibitors (CPIs). While all participants in the mRNA vaccine trial were on CPI therapy, only 42% of the patients in the TL trial received CPIs. Given that CPIs are known to be independently associated with a higher risk of grade 3 adverse events in melanoma patients, this might be the cause of this discrepancy between the two trials [156–158]. Consequently, researchers and mRNA vaccine manufacturers are focusing on overcoming these difficulties through different strategies such as designing advanced delivery systems like non-invasive transdermal iontophoresis (IP) technology for enhancing cellular uptake and cytoplasmic translation efficiency, as well as eliminating double-stranded contamination, which is known to activate innate immune responses and contribute to mRNA degradation [20, 147, 159, 160]

Computational Immunoinformatics in epitope prediction and design of immunotherapies targeting melanoma

To improve the translational relevance of AI-assisted vaccine design in melanoma, computational workflows should be anchored to feasible antigen-selection questions. In particular, AI and immunoinformatics can support HLA-restricted epitope prioritization (with specific attention to HLA-A*02:01), alongside prediction of antigen processing and presentation, and evaluation of immune-evasion risks. This includes antigen downregulation and loss of HLA class I expression. This may help define the subgroups most likely to benefit from shared-antigen mRNA vaccination and inform rational combination strategies, especially with checkpoint inhibitors [161–163].

In the current clinical environment, melanoma is examined using the clinically and pathologically based Tumor Node Metastasis (TNM) framework. It classifies cases according to tumor dimensions, distant metastases, and lymphatic spread. This approach can result in a structured assessment. However, it doesn’t have the ability to show the complexity of the disease. Melanoma is marked by high variability both between tumors and within the same tumor. This can lead to noticeable differences in patient outcomes even in patients assigned the same stage. Ideally, a biomarker should capture biological characteristics of the tumor in every patient, but the TNM model is limited to anatomical features. [164].

Artificial intelligence (AI) technologies, such as both machine learning (ML) and deep learning (DL) are being used in medical imaging and histopathological analysis to analyze the tumor microenvironment (TME) in combination with immunohistochemistry [165]. These methods can identify biomarker distribution and expression in different tissue subtypes, which can help predict responses to immunotherapies or other treatments [166]. AI can also assist in selecting the suitable immunotherapy drug for the patient. An important advantage of AI is its capacity to process and learn from large datasets and identify complex patterns relevant to tasks such as diagnosing disease or mapping genetic mutations [167]. However, there are several challenges for broad integration of AI into clinical workflows. These include limited access to high-quality data, biases within datasets, insufficient data and code sharing practices, and the difficulty of clarifying complex AI models. [165]. Despite the challenges, AI is becoming a key factor in the development of personalized tumor vaccines. An important initial step in developing customized therapeutic cancer vaccines (TCVs) involves identifying tumor-specific neoantigens (TSNs) that are on the surface of malignant cells. The incorporation of AI into this process can mitigate identifying vaccine candidates for individual patients [168]. In recent years, significant progress has been made in the usage of AI tools to improve personalized cancer vaccines [169], such as deep learning models including convolutional neural networks, recurrent networks, and transformers. These approaches help to model interactions between peptides and MHC molecules and predict T cell responses. Generative algorithms have also been used to create improved genetic sequences, while integrated multitask systems combine multiple layers of biological information [170]. Early studies suggest that these AI approaches can perform better than previous methods and help in identifying neoantigens and producing vaccine components [171]. AI also enhances the precision of immune response selection by applying ML techniques to convert somatic mutations into viable neoantigen candidates [172, 173]. The combination of computational analysis, high-throughput genomics, and ML can more efficiently pinpoint therapeutic targets and biomarkers which are essential for the development of effective cancer vaccines [168]. In one of the previous works, graph-based dimensionality reduction was used to map patient immune profiles. This allows them to visualize how individuals clustered based on gene expression related to immunity. Pseudo-time analysis was then used to evaluate similarities in immune dynamics among patients. They also explored the correlations between 28 types of immune cells and principal components (PCA1 and PCA2) were assessed using Pearson analysis. The variations in immune cell populations were measured with the Wilcoxon test. Survival outcomes were examined across four immune subtypes. The results suggested that the identified tumor antigens could serve as valuable targets for mRNA vaccine development against skin cutaneous melanoma (SKCM).

In addition to identifying antigens, AI contributes to refining the delivery systems for mRNA vaccines. Deep Neural Networks (DNN) are particularly suitable for enhancing the immunogenicity of lipid nanoparticle (LNP)-based mRNA formulations to produce vaccines [174, 175]. DNNs consist of layered neural architectures that extract increasingly complex features from input data which enables them to uncover subtle dependencies in large and multidimensional datasets [169]. These models can also be trained to predict how variations in lipid composition, mRNA structure, and nanoparticle properties affect the activation of immune components such as dendritic cells, T cells, and B cells [175, 176]. DNNs offer valuable recommendations to fine-tune vaccine formulations By learning these relationships such as optimizing lipid ratios for better cell entry, tweaking nanoparticle size for improved lymphatic transport, or engineering mRNA sequences to boost antigen expression and immune recognition [177].

In addition to supervised learning, generative models have been employed to fine-tune codon usage for more efficient protein translation. One notable example is RiboCode. It is a deep generative framework that uses ribosome profiling data to engineer codon sequences that are optimized for ribosome engagement and elongation speed [178]. In another approach, Castillo-Hair et al. targeted the optimization of 5′ untranslated regions (UTRs) for mRNA-based delivery of gene-editing enzymes [179]. They used polysome profiling on large libraries of randomized 5′ across various cell types to assess their impact on translation. These data were then used to train a deep learning model to predict which UTR variants can improve ribosome loading. Similarly, Morrow et al. designed a machine learning framework for 3′ UTR design using high-throughput RNA stability assays [180]. Their dataset consists of thousands of synthetic 3′ UTRs tested in human cells. The model predicted mRNA half-life and identified new sequence designs that can significantly enhance transcript longevity.

Limitations and future perspectives

In the past decade, following the COVID-19 pandemic, mRNA vaccines have experienced remarkable success. Despite their advantage, mRNA vaccines have faced several challenges that limit their long-term application in infectious diseases and cancer immunotherapy. One of the main concerns is their potential for undesired immune response by multiple pathways, such as complement cascade stimulation, cytokine production, and immune cell engagement. These unwanted activations can start localized, systemic, or prolonged immunogenicity, which requires a comprehensive preclinical and clinical assessment. Another critical issue involves the biocompatibility of nanoparticle delivery systems. Certain inorganic nanoparticles may exhibit cytotoxic or genotoxic effects, emphasizing the need for comprehensive safety assessments to evaluate organ-specific toxicity and ensure minimal unwanted outcomes in vaccine recipients [181–183].

Recent advances in AI and ML have provided new opportunities to enhance vaccine design by predicting immune responses and optimizing nanoparticle characteristics. Furthermore, personalized mRNA vaccines, guided by each patient's individual genetic and immunological profiles, hold promise for more effective and tailored immunotherapies [184, 185]. Despite these bright sides, the ability to scale industrial production continues to be a significant problem. Small variations in the physicochemical characteristics of nanoparticles, such as size, composition, or surface chemistry, may critically alter their stability and immunization efficiency. Consequently, creating reliable large-scale production pipelines with strict processes for quality control is vital to guarantee consistency and reliability across batches [186]. Additionally, storage stability becomes a critical limiting factor. The preservation of nanoparticle-based mRNA vaccines relies on specific thermal conditions to ensure molecular stability and efficacy. Factors like nanoparticle aggregation, mRNA degradation, and diminished performance during storage or shipment present significant obstacles. Continuous advancements in formulation chemistry and packaging technologies are essential to improve vaccine stability and preservation, especially for use in limited resources and during international distribution programs [189].

Achieving scalable GMP production for personalized nanovaccines remains a major challenge. The manufacture of these individualized mRNA vaccines involves multiple complex steps, including patient-specific antigen sequencing, customized formulation, and full quality-assurance and quality-control (QA/QC) processes under time-sensitive conditions. Besides, many mRNA nanovaccine platforms demand firm cold-chain logistics for both distribution and long-term storage. To overcome these obstacles, strategies such as modular microfluidic manufacturing, AI- and large-language-model-assisted neoantigen design, and interchangeable nanocarrier platforms have been developed, enabling rapid and adaptable production of vaccines customized for each patient [92, 187, 188].

Utilization of patient-derived components, such as tumor lysates, membranes, and exosomes in biomimetic nanovaccines faces a complex regulatory condition, as they overlap biologics, cell therapies, and drug products. Since each dose is unique, GMP compliance is challenging, and batch variability, sterility assurance, and classification ambiguities should be addressed. Small-batch personalized manufacturing increases quality-control burden and cost, and any further combination therapies, such as nanovaccines with checkpoint inhibitors, require multi-agency review and safety attribution [92, 187, 190–194].

The safety profile of personalized nanovaccines remains a major obstacle to clinical translation. Major risks include cytokine release syndrome, off-target immune activation, accumulation in non-target organs, and residual contaminants from the manufacturing process. To mitigate these risks rooting from uncontrolled adjuvant exposure, excessive stimulation of innate immunity, or insufficient purification of biological components, biomimetic and advanced engineering strategies have been developed. Immune-evasive surface modifications, such as CD47 or PEG coatings, help reduce nonspecific immune recognition, while receptor-directed targeting localize immune activation to tumor tissue. Controlled-release formulations synchronize antigen and adjuvant delivery to prevent systemic cytokine storm, and restrict purification with microfluidic quality-control systems removes endotoxins and residual nucleic acids [92, 195–197].Finally, AI-guided personalization of vaccines introduces its own set of challenges. Machine learning and immunoinformatic tools are increasingly integrated into neoantigen selection and vaccine design, but the clinical validation of algorithmic predictions remains challenging [92].

Conclusion

The rising incidence of melanoma, combined with its aggressive behavior and low efficacy of traditional treatments, highlights the importance of finding new therapeutic strategies [4–6]. mRNA vaccines offer a novel method in melanoma immunotherapy with cost-effective production and potent immunogenicity, which can target various tumor-specific and tumor-associated antigens [17]. In this review, we explored the role of mRNA vaccines in melanoma treatment, explaining their molecular design, immune activation mechanisms, and clinical advancement, while discussing current challenges and future perspectives. Evidence from preclinical and clinical studies highlights the strong therapeutic potential of mRNA vaccines. As Oberli et al. and Wang et al. studies have indicated, LNP-based mRNA vaccines encoding gp100 and TRP2 antigens by inducing strong CD8⁺ T-cell responses can suppress tumor growth and improve long-term immune memory in melanoma models [44, 125]. Additionally, clinical trials conducted by Kyte et al. and Weide et al. have also shown that mRNA vaccines are safe and immunogenic, with some patients achieving complete clinical responses and no severe adverse events observed [128, 129]. Beyond studies of mRNA vaccines alone, the KEYNOTE-942 trial investigated the combination of mRNA vaccines with immune checkpoint inhibitors, enhancing melanoma treatment by exploiting potential synergistic effects [157]. These findings establish mRNA vaccines as a flexible platform capable of addressing the heterogeneity and immune evasion characteristics of melanoma. Advances such as 5’ cap methylation, poly(A) tail elongation, and designing LNPs have improved the stability, translational efficiency, and delivery of mRNA vaccines. Moreover, we have discussed how computational immunoinformatics, powered by AI and machine learning, is changing mRNA vaccine development by improving neoantigen prediction, optimizing nucleic acid sequences, and refining delivery strategies. These advancements enhance vaccine immunogenicity and personalization, addressing melanoma’s high inter- and intra-tumor variability, which conventional staging systems fail to capture. However, the integration of AI faces challenges, such as limited high-quality datasets and model interpretability, which highlights the need for more efficient data-sharing frameworks [118, 162, 168, 198, 199]. Despite their potential, challenges related to mRNA remain, including mRNA instability, potential inflammatory side effects from LNPs, and severe immune reactions like anaphylaxis or cytokine storms, while storage instability and cold-chain requirements complicate global distribution [200, 201]. Current studies focus on addressing these challenges by developing advanced delivery systems, such as non-invasive transdermal iontophoresis, and implementing strategies to reduce immunogenicity, including the elimination of double-stranded RNA contaminants [202]. Looking ahead, mRNA vaccines hold great potential for transforming melanoma therapy. Further advancements in mRNA engineering, nanoparticle biocompatibility, and AI-driven personalization are essential to overcoming current barriers. Comprehensive preclinical and clinical evaluations of mRNA vaccines, together with scalable production platforms, will ensure their safety and efficacy, allowing their broader use in melanoma and other cancers. Building on their success in combating infectious diseases, mRNA vaccines can become an essential tool of precision medicine, particularly in oncology, and present hope for improved survival and quality of life for patients with melanoma [198, 203].

Author contributions

M.H.K., A.R., K.Z., G.B.S., H.A., M.P., N.Y., K.S., and N.R. conceptualized the study and drafted the manuscript. M.H.K. designed the figure. K.S. N.Y., and N.R. supervised the study and critically appraised the manuscript. All authors confirmed the final submitted version of the manuscript to be published.

Funding

This study was supported by the USERN Foundation, Houston, TX (Grant No. 2025.06.21). This grant was awarded to the Network of Immunity in Infection, Malignancy and Autoimmunity (NIIMA), Universal Scientific Education and Research Network (USERN), Stockholm, Sweden.

Availability of data and materials

No datasets were generated or analysed during the current study.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.