What evidence gaps are holding back mRNA cancer vaccines?

Why can't we get enough mRNA into the right immune cells?

The single biggest technical gap is delivery. mRNA is fragile and must reach antigen-presenting cells (like dendritic cells) inside lymph nodes, but current lipid nanoparticles (LNPs) — the same technology used in COVID-19 vaccines — struggle with two biological barriers. First, after LNPs are taken up by cells, most of the mRNA gets trapped inside endosomes and lysosomes and is destroyed before it can be translated into protein. One study found that adding esomeprazole (a common heartburn drug) to LNPs raised the pH inside these compartments, helping mRNA escape and boosting antigen expression both in lab dishes and in mice [4]. Second, the dense extracellular matrix around tumors physically blocks LNPs from reaching their targets; the same esomeprazole strategy also loosened this matrix by dampening fibroblast activity [4]. Without such fixes, the majority of the injected mRNA dose is wasted.

Even when LNPs do reach cells, they can trigger inflammation that actually shuts down protein production. A 2024 study showed that LNPs generate reactive oxygen species during translation, which both reduces how much antigen is made and causes unwanted side effects [6]. The researchers engineered a biodegradable polymer that mops up these reactive oxygen species, improving translation efficiency and reducing inflammation in mice [6]. This suggests that current LNP formulations are not just inefficient — they may be actively counterproductive.

Another delivery gap is targeting. Most LNPs end up in the liver or spleen by default, not in lymph nodes where immune responses are orchestrated. A 2025 review notes that strategies to direct mRNA vaccines specifically to dendritic cells and lymph nodes are still in early development, and significant work is needed to understand the underlying trafficking mechanisms [7]. Without this targeting, the vaccine's potency is severely limited.

How do we know which tumor targets to aim for?

A second major gap is that we cannot reliably predict which mutated proteins (neoantigens) on a patient's tumor will actually provoke a strong immune response. In a 2024 clinical trial of personalized mRNA vaccines for non-small cell lung cancer and melanoma, only 20–30% of the predicted MHC-I and MHC-II epitopes triggered measurable CD8+ and CD4+ T-cell responses [1]. That means 70–80% of the vaccine's payload is essentially wasted — it encodes targets the immune system ignores. Another study using a lipopolyplex delivery system in mice and two human patients also found that neoantigen prediction accuracy is suboptimal, compromising in vivo therapeutic efficacy [2]. The problem is that current algorithms predict which peptides will bind to MHC molecules, but binding does not guarantee that T cells will recognize and attack those targets. Improving these prediction models is a critical evidence gap.

Beyond prediction, there is the issue of tumor heterogeneity. A single tumor can have dozens of different mutations, and different metastases can have different neoantigens. A 2023 review highlights that tumor heterogeneity makes it difficult to select a universal or even a broadly effective set of vaccine targets [5]. Most current vaccines target only a handful of neoantigens, which may allow the tumor to escape immune pressure by downregulating those specific targets. The field lacks robust data on how many neoantigens need to be included, and from which tumor regions, to prevent escape.

Even when the vaccine works, why doesn't the tumor shrink?

The third critical gap is that mRNA vaccines often generate a strong immune response in the blood and lymph nodes, but the tumor itself actively suppresses those immune cells once they arrive. A 2023 review notes that the immunosuppressive tumor microenvironment — full of regulatory T cells, myeloid-derived suppressor cells, and inhibitory cytokines — is a major obstacle that blunts the effectiveness of otherwise potent vaccines [5]. This is why virtually all successful clinical trials to date have combined mRNA vaccines with immune checkpoint inhibitors (like anti-PD-1 antibodies). In the melanoma trial mentioned earlier, the personalized vaccine was given alongside pembrolizumab (Keytruda), and even then, the response was not universal [1]. Another study in mice found that combining the vaccine with an immune checkpoint inhibitor boosted antitumor activity compared to the vaccine alone [2].

A 2025 study uncovered a specific metabolic mechanism that may explain part of this suppression. After injecting an mRNA-LNP vaccine into mice, the researchers found that macrophages in the draining lymph nodes produced high levels of itaconate, a metabolite that actually dampens dendritic cell antigen presentation and T-cell function [8]. When they knocked down the gene (Irg1) responsible for itaconate production using a small interfering RNA (siRNA) delivered in the same LNP, the vaccine's efficacy improved significantly, both alone and in combination with anti-PD-1 therapy [8]. This suggests that the vaccine itself may inadvertently trigger immunosuppressive pathways, and we are only beginning to understand these counter-regulatory loops.

Finally, there is a gap in understanding how to optimally combine mRNA vaccines with other treatments. While combination with immune checkpoint inhibitors is common, a 2024 review points out that the optimal timing, dosing, and sequence of these combinations have not been established in rigorous clinical trials [3]. Without this data, we are essentially guessing at the best way to use these vaccines in the clinic.

About These Sources

This answer is built on 8 peer-reviewed studies — published from 2023 to 2025, 7 from 2024 or later, 6 in Q1 journals, collectively cited 162 times — selected as the most relevant from 15 studies that passed quality screening, drawn from 50 papers retrieved from a database of over 500 million.

Sources used in this answer

Related research questions