Developing CMV as a vaccine vector

There is an increasing need to develop anti-cancer vaccines that elicit strong and durable immune responses and are able to break immune tolerance to cancer antigens. Cytomegaloviruses (CMV) induce one of the strongest immune responses of any viral vector – as much as 20% of all CD8+ T cells can be specific for a single epitope, thus making them an exciting candidate for vaccine development. The majority of work, however, has been based on replication-competent viruses. In humans, the use of replication-competent human CMV (HCMV) vectors would not be permitted due to safety concerns, however replication-deficient vectors may not induce immune responses of the same magnitude as wildtype HCMV. In an attempt to make CMV vectors safer, I utilised a tetracycline repressor (tetR)-based system and generated a total of 55 mouse and human CMV vectors in which viral replication was dependent on inducible expression of multiple different viral genes. Having analysed the ability of the vectors to replicate in vitro and in vivo, I generated three vectors which replicated well in the presence of doxycycline, but in its absence replicated poorly in vitro, and did not appear to replicate in vivo. Importantly, in a direct comparison, the level of control was more stringent than viruses previously characterized in the literature. Adenoviruses are one of most widely used vectors in cancer immunotherapy, induce strong T-cell responses, and could be used in a prime-boost regiment with CMV vectors. However, we found that in a prophylactic regimen, a recombinant human adenovirus (Ad) encoding human 5T4 tumour-associated antigen alone did not protect against 5T4-expressing tumour challenge. We tested the immunogenicity of 5T4 in two mouse strains and established three 5T4-expressing cancer models in which vectors can be tested. In future work, we hope to use our inducible CMV/Ad vaccine regimen in these models