Two complementary arms of mRNA cancer vaccination: non-personalized danger-response constructs and defined shared-neoantigen cassettes for cold tumors

Therapeutic cancer vaccination has moved from preclinical work into randomized trials, but the field remains split among strategies that are usually treated as alternatives. Personalized neoantigen vaccines give broad coverage at the cost of weeks of sequencing and patient-specific manufacturing. Shared tumor-associated antigen vaccines are available off the shelf but must overcome self-tolerance, and two large randomized trials of a cancer-germline antigen vaccine were negative. A third and more recent class encodes no tumor target at all and instead provokes a broad danger response. For immunologically cold tumors, which include several of the most lethal solid malignancies and respond poorly to checkpoint blockade, none of these has delivered a general solution. This review argues that two of these strategies are better understood as parts of one mRNA platform than as competitors. The first part is a non-personalized, non-tumor-specific construct that uses a foreign, highly immunogenic antigen to drive innate, type I interferon-dependent activation. In a single directly supportive preclinical study, lipid particles carrying tumor-agnostic antigen RNA converted poorly immunogenic tumors into an inflamed state, sensitized them to checkpoint inhibitors, and promoted epitope spreading toward endogenous tumor antigens in mice. Related RNA-particle work, including early translational observations in glioma, supports the general principle but does not test a fixed tumor-agnostic antigen product. Because all of the direct evidence is murine, the route, dose, carrier architecture, and schedule that produce a therapeutic interferon response in humans remain unknown and will require dedicated dose-finding and formulation trials. The second part is a defined cassette of shared and public neoantigens, recurrent driver mutations, cancer-germline antigens, and oncoviral antigens that recur across patients and can be built from public sequence records. This review synthesizes the mechanism of each part, sets both against the record of non-mRNA therapeutic cancer vaccines that have reached phase 3 or approval, presents a catalog of buildable shared-antigen constructs, and sets out the rationale for pairing a danger-response prime with a defined-antigen boost. It also treats adoptive cell therapy as a third partner, since tumor-infiltrating lymphocyte and engineered T-cell products now have regulatory approval in melanoma and synovial sarcoma and are in registrational testing in lung and endometrial cancer, and since an RNA vaccine has already been used clinically to amplify transferred cells. Throughout, established findings are separated from projections. A unified two-component mRNA platform is biologically coherent and is supported, component by component, by preclinical and early clinical data. Its clinical efficacy, its manufacturing economics, and its accessibility advantages remain to be shown prospectively.