AI Maps Out a Multi-Target CMV Vaccine Blueprint: EVX-V1 Animal Studies Target Infection, Latency, and Reactivation

Once cytomegalovirus (CMV) enters the human body, it may remain latent for life and reactivate when immunity is weakened. This ability to span primary infection, latency, and recurrence is also one of the barriers that vaccine development has struggled to overcome for decades. Danish biotechnology company Evaxion has now released a new set of preclinical data for its candidate program EVX-V1, proposing the use of artificial intelligence to select multiple antigens and immune epitopes to contain the virus at different stages.

The results were presented at the 2026 International Herpesvirus Workshop in Montreal. According to the company’s announcement, novel T-cell epitopes identified by AI improved control of acute infection, viral latency, and reactivation in CMV-infected mice. The key here is not for the algorithm to directly predict whether a vaccine will work, but to use it to screen large volumes of viral sequences and immune characteristics for fragments more likely to be recognized by T cells, which are then validated in animal models.

EVX-V1 uses a multi-component design. It combines newly discovered protective B-cell antigens and T-cell epitopes with structurally optimized known antigens, including an AI-adjusted prefusion conformation of glycoprotein B (gB). gB is involved in viral entry into cells; stabilizing it in a specific conformation is intended to make it easier for antibodies to recognize surface structures with protective value.

Evaxion said earlier cell and animal studies showed that its novel B-cell antigens could reduce viral infection and cell-to-cell spread, while the optimized structural antigens produced stronger virus-neutralizing activity. The new data add evidence on the T-cell side, extending the vaccine concept from preventing viral entry to clearing infected cells and suppressing the re-emergence of latent virus. However, the magnitude of the relevant comparisons, sample sizes, statistical methods, and full experimental design were not disclosed in the news announcement.

CMV typically does not cause serious disease in most healthy adults, but congenital infection may lead to hearing, vision, or neurodevelopmental problems. Organ transplant recipients and other immunosuppressed populations may also develop severe disease. There is still no approved CMV vaccine, making a multi-target strategy biologically appealing: antibodies can intercept the virus outside cells, while T cells may be able to address cells that are already infected or harboring the virus.

However, EVX-V1 remains at the antigen-selection and preclinical-validation stage. Immune responses to CMV in mice may not replicate human infection, and the candidate formulation has not yet been shown to induce durable, broad, and safe protection in humans. Multi-antigen combinations also increase the complexity of manufacturing, dose configuration, and quality control. These conference data are better viewed as an early signal supporting the design approach rather than a definitive conclusion about vaccine efficacy. The next critical step will be to determine the final antigen combination and evaluate its safety and immune effects through comprehensive studies and human trials.