Synthetic Antigen-Presenting Vesicles for Selective Immunomodulation

The rapid advancement of generative artificial intelligence has enabled unprecedented progress for the field of computational protein design. A forthcoming challenge for generative protein design algorithms is the immunocompatibility of these de novo designed molecules with organism physiology, namely humans. A separate, but related, aspirational goal for synthetic biology is to perform cellular reprogramming in vivo so that cell-based therapies and biologics are generated endogenously by patients rather than being externally manufactured or expanded before delivery, as is the case with biologics, T cell therapies, and stem cell therapies; again, a major hurdle for the in vivo production of these therapies and in vivo cellular reprogramming is immunogenicity. To address these challenges, we first demonstrate a cell-like, cell-free approach for in vivo cellular reprogramming with the induced release of pMHCI and pMHCII-loaded synthetic antigen-presenting vesicles that are secreted from non-immune cells by DNA and mRNA transfection to facilitate the selective expansion or silencing of immune responses. Next, we show initial results for the use of human tonsil organoids as a quantitative assay for adenoviral vector immunogenicity, enabling future directed evolution approaches for immunogenicity reduction as well as generation of an immunogenicity dataset to tailor modern computational protein design algorithms for human immunocompatibility. Together, these projects represent complementary methods to control protein immunogenicity, either through rationally engineered or directedly evolved modifications identified by physiologically-relevant in vitro models, or with an administered mRNA therapeutic that selectively modifies the immune response to a protein that cannot be computationally redesigned.