Background Immune checkpoint inhibitors have transformed cancer therapy, yet many patients fail to respond, underscoring the need for complementary strategies. Personalized neoantigen cancer vaccines (NCVs), which stimulate highly specific T-cell responses against tumor-specific mutations, are a promising approach now advancing in clinical development. Methods We developed a clinically scalable platform based on high-purity linear hairpin DNA (hpDNA), produced by an entirely cell-free enzymatic method and delivered intramuscularly by electroporation (EP). An M20-hpDNA vaccine, encoding 20 neoantigens from the MC38 tumor model, was benchmarked against a plasmid DNA (pDNA) counterpart and a pseudouridine-modified messenger RNA (mRNA) version delivered by EP. Immune responses were characterized by enzyme-linked immunospot, flow cytometry, bulk RNA sequencing, and single-cell RNA/T-cell receptor sequencing. Antitumor activity was assessed in prophylactic, therapeutic, and metastatic mouse models. Results hpDNA and pDNA vaccines elicited robust and comparable CD8 + and CD4 + neoantigen-specific T-cell responses, leading to dose-dependent protection in both prophylactic and therapeutic settings. In contrast, pseudouridine-modified mRNA vaccine delivered by EP failed to elicit detectable immunity despite measurable transgene expression. Transcriptomic profiling revealed that hpDNA, but not EP-delivered mRNA, triggered a pro-inflammatory signature, including Il6, Ccl4, Cxcl2, Cd80, and Spp1. Single-cell sequencing of Adpgk-specific CD8 + T cells showed expansion of memory subsets and unique hyperexpanded clonotypes. Functionally, M20-hpDNA prevented lung metastases in a mixed-modality protocol and synergized with α-cytotoxic T-lymphocyte associated protein 4 therapy. Efficacy was further confirmed with C20-hpDNA, encoding 20 CT26-specific neoantigens. Conclusions These findings demonstrate that hpDNA delivered by EP is a versatile, potent, and clinically scalable platform for personalized NCVs. Under the delivery conditions tested here, EP-delivered, non-formulated mRNA did not elicit detectable antigen-specific T-cell responses, whereas hpDNA induces productive inflammation and durable T-cell memory. These results highlight the importance of delivery context in shaping vaccine immunogenicity and support further development of hpDNA as a safe and competitive strategy for individualized cancer immunotherapy. Importantly, our data do not challenge the well-established immunogenicity and clinical activity of formulated mRNA vaccines (eg, mRNA-lipid nanoparticle) in infectious disease and cancer settings.
Tonini et al. (2026) studied this question.