From Amphiphiles to mRNA platforms: emerging vaccination strategies for pancreatic cancer

Abstract

Pancreatic ductal adenocarcinoma (PDAC) remains among the deadliest cancers, with limited surgical eligibility, modest chemotherapy benefit, and resistance to immune checkpoint blockade. Two recent vaccine platforms have shown encouraging results. Wainberg et al. demonstrated that the amphiphile vaccine ELI-002 efficiently traffics to lymph nodes via albumin binding and induced KRAS-specific T-cell responses in most patients, correlating with survival. In parallel, Sethna et al. reported that an individualized uridine-modified mRNA vaccine elicited durable, polyfunctional CD8⁺ T cells with long-term persistence, especially when combined with PD-1 blockade. Amphiphiles provide rapid and efficient priming, whereas mRNA vaccines broaden and sustain clonotypic diversity. A hybrid prime–boost strategy may synergize these complementary mechanisms, while advances in multi-omics and AI-driven neoantigen prediction pave the way for personalized designs. Together, these developments suggest that PDAC, long regarded as immunologically “cold,” may become tractable to vaccination strategies. Importantly, these findings are based on early-phase clinical studies with limited patient numbers and should therefore be interpreted as preliminary clinical evidence requiring further studies.

Recently, two innovative vaccine platforms for pancreatic cancer have been developed. In Nature Medicine (2025), Wainberg et al.. showed that the ELI-002 amphiphile vaccine induced strong T-cell responses in patients with mutant KRAS (mKRAS) tumors [1]. In Nature (2025), Sethna et al.. demonstrated that lipid nanoparticles carrying uridine-modified mRNA encoding neoantigens primed long-lived CD8⁺ T cells [2].

Pancreatic ductal adenocarcinoma (PDAC) remains one of the most lethal malignancies, owing to its poor prognosis. Surgical resection is typically followed by adjuvant chemotherapy, including modified (m)FOLFIRINOX; however, fewer than 10–20% of patients are eligible for surgery at diagnosis. Immune checkpoint inhibitors have failed to yield meaningful benefits in PDAC and most colorectal cancer (CRC), in which poor antigen presentation and a highly immunosuppressive tumor microenvironment impose significant barriers to immune recognition and clearance. In addition, PDAC frequently exhibits stroma–associated T-cell exclusion and a myeloid-enriched immunosuppressive milieu, further limiting effective antitumor immunity. Notably, because KRAS mutations are observed in approximately 85% of pancreatic cancers, considerable research has focused on developing therapeutic approaches that specifically target KRAS. Nevertheless, targeting KRAS mutations has long been considered undruggable [3]. The development of KRAS G12C inhibitors has shown that direct targeting is possible; however, the clinical benefits are modest and short-lived owing to rapid resistance.

To address this issue, Wainberg et al. employed an amphiphile vaccine that couples KRAS-mutant peptides and a CpG adjuvant to a lipid tail capable of binding serum albumin. By hitchhiking on albumin, the constructs were efficiently trafficked into draining lymph nodes, optimizing their delivery to dendritic cells (Fig. 1a). In this trial, 84% (21/25) of patients developed KRAS-specific T-cell responses, encompassing both CD4⁺ helper and CD8⁺ cytotoxic subsets. Importantly, survival strongly correlated with response magnitude. Patients exhibiting a nine-fold higher expansion of KRAS-specific T cells experienced relapse-free survival not reached versus 3.0 months in low responders, and overall survival was not reached versus 16.0 months. Moreover, 67% of the patients demonstrated antigen spreading with secondary recognition of neoantigens beyond KRAS, suggesting broad immunological activation initiated by vaccination.

Fig. 1.

Model of amphiphile vaccine and mRNA neoantigen vaccine. a The amphiphile vaccine consists of albumin-binding lipids, a PEG linker and a G12D or G12R peptide. Following subcutaneous injection, the amphiphile vaccine binds to albumin and rapidly traffics to draining lymph nodes. b. The mRNA vaccine, encoding up to 20 neoantigens derived primarily from somatic passenger mutations, is encapsulated in lipoplex (mRNA-lipoplex). Upon intravenous injection, the mRNA-lipoplex internalized by antigen presenting cells (APCs), leading to CD8 + T cell activation and the induction of prolonged CD8 + T cell clones. c Approaches for future use of these vaccines. A proposed hybrid strategy combines amphiphile and mRNA-lipoplex cancer vaccines. Hybrid vaccination can be achieved through prime-boost strategy to promote both rapid priming and sustained T cell clonal persistence (top). Engineering mRNA-lipoplex vaccines fused with albumin-binding ligands enables efficient lymph node targeting, thereby enhancing antigen presentation and T cell priming (middle). AI-driven analysis of patient-specific genomic information can guide the design of personalized peptides or neoantigens, paving the way for the clinical implementation of individualized cancer vaccines (bottom). The figure was created using Biorender.com

Sethna et al. used an individualized uridine-based mRNA neoantigen vaccine (autogene cevumeran), mFOLFIRINOX, and a single vaccine boost dose. Once delivered intracellularly, the mRNA is translated, processed, and presented to both MHC-I and MHC-II, engaging in CD8⁺ and CD4⁺ immunity (Fig. 1b). When administered with PD-1 blockade, this vaccine elicited robust polyfunctional CD8⁺ responses. Notably, clone tracking revealed that vaccine-induced T cells were newly generated, not preexisting, and bore phenotypes similar to tissue-resident memory cells. Modeling predicted an average persistence of ~ 7.7 years, with ~ 20% of the clones expected to endure for decades. While the reported results were encouraging, it should be noted that both the amphiphile and neoantigen vaccine trials were early-phase studies conducted in limited cohorts (ELI-002 2P, n = 25; autogene cevumeran trial, n = 19).

In contrast to the vaccine platforms introduced above, the earlier generations of cancer vaccines are limited in many respects. Dendritic vaccines are often viewed as the most physiological approach; however, they require laborious ex vivo manipulation and show high donor variability, while rarely achieving efficient lymph node migration. Protein and peptide vaccines offer a high degree of compositional precision and are generally well tolerated; however, they are restricted by HLA type and tend to favor MHC-II presentation. DNA vaccines are attractive because they are cost-effective, stable, and straightforward to manufacture. However, in humans, they typically generate weak CD8⁺ T-cell responses [4]. Conversely, amphiphilic and mRNA vaccine platforms incorporate strategies to improve lymph node targeting and broaden antigenic coverage, thereby enhancing the persistence of T-cell responses. Both amphiphilic and mRNA vaccines elicit durable antitumor immunity in PDAC, which correlates with survival benefits, albeit through distinct pathways. Amphiphiles prioritize efficient antigen trafficking and priming in lymph nodes, as exemplified by the AMPLIFY-201 phase 1 trial in with 84% (21/25) of patients developed KRAS-specific T-cell responses and 67% (6/9) showed antigen spreading. Moreover, mRNA vaccines autogene cevumeran in resected PDAC correlated with delayed recurrence at a median follow-up of 3.2 years: responders (n = 8) exhibited median RFS not reached versus 13.4 months in non-responders (n = 8; HR = 0.14; P = 0.007). This vaccine enables in situ antigen synthesis with broad HLA coverage and long-term persistence. In the study, 86% of vaccine-expanded CD8⁺ T-cell clones remained detectable at ~ 3 years. Modeling based on early clinical follow-up projected an average persistence of ~ 7.7 years, with a subset of clones predicted to endure for decades. Furthermore, amphiphiles rely on direct peptide delivery, resulting in rapid but narrower immune activation, whereas mRNA vaccines leverage endogenous antigen expression, thereby enabling a broader epitope landscape.

However, each platform has its own limitations. Amphiphiles depend on predefined peptides that are inherently constrained by HLA presentation and have limited variant coverage. In addition, their reliance on synthetic peptides may restrict the induction of broader epitope responses and limit their adaptability to tumor heterogeneity. mRNA vaccines offer greater versatility, but face hurdles in terms of manufacturing complexity, cold-chain logistics, and innate immune sensing, which reduce translational efficiency. Nonetheless, both vaccine strategies demonstrated augmentation of CD4⁺ and CD8⁺ T-cell responses in clinical trials, suggesting that pancreatic cancer, traditionally regarded as an immunologically “cold tumor,” could potentially be rendered more immunologically active, thereby indicating a promising therapeutic potential.

Thus, hybrid amphiphile–mRNA regimens are promising. Amphiphiles excel at priming naïve T cells rapidly and efficiently in lymphoid tissues, whereas mRNA vaccines expand and sustain the clonotypic breadth. A prime boost or concurrent delivery can synergistically enhance antigen spread, reduce immune escape, and provide coverage across multiple KRAS and NRAS variants (Fig. 1c). Nevertheless, challenges remain, such as increased manufacturing complexity, the need for rigorous safety monitoring (e.g., excessive immune activation), and uncertainty regarding the optimal sequence, dosing, and patient stratification.

Notably, Sethna et al.. employed autogene cevumeran, which inherently embody the paradigm of personalized cancer vaccination. Unlike earlier decades constrained by technical barriers, modern multi-omics profiling and modularized manufacturing pipelines allow the design of patient-specific neoantigen vaccines (Fig. 1c). This personalization not only maximizes therapeutic efficacy but also enables rapid redesign at recurrence using archived patient data. Although several challenges remain, such as long manufacturing lead-times, high costs, limited tumor sample availability, and HLA diversity, multiple strategies are under active investigation to overcome these barriers, including accelerated workflows with modular CMC, AI-based epitope prediction, and hybrid designs that integrate both public and personalized neoantigens [5].

In summary, the two recent cancer vaccine platforms have shown promising clinical results for PDAC therapy. Amphiphilic and mRNA-based vaccines act through distinct mechanisms, but both platforms have shown an improvement over earlier vaccines by enhancing lymph node delivery, in situ antigen synthesis, antigenic breadth, and persistence. To overcome practical challenges, hybrid regimens combining rapid priming with sustained breadth and personalized design are promising for improving PDAC outcomes.

Author contributions

DG Lee and K Noh wrote the manuscript. K Noh supervised the manuscript. All authors read and approved the final manuscript.

Funding

This work was supported by NRF grants funded by Korea government (RS-2024-00338397) and KRIBB Research Initiative Program (KGM1062612 and KGM1322612).

Data availability

No datasets were generated or analysed during the current study.

Declarations

Competing interests

The authors declare no competing interests.