(567c) De Novo Design of Engineered Cytokine Mimetics As Targeted Immunotherapeutics.

Recent efforts to re-engineer native cytokines to improve their selectivity and stability have significantly advanced progress in cytokine therapeutic development. However, there are inherent challenges to developing therapeutics from natural proteins. First, most natural cytokines are only marginally stable, hence mutations aimed at increasing efficacy can decrease expression or cause aggregation, making manufacturing and storage difficult. More substantial changes, such as the deletion or fusion of functional or targeting domains, can also dramatically alter pharmacokinetic properties and tissue penetration. In addition, immune responses targeting a variant of a natural protein may cross-react with the endogenous molecule with potentially catastrophic consequences. Finally, the high potency and multifarious functions of cytokines hinder their efficacy and can result in harmful off-target effects and systemic toxicities.

Here, we combine a best-in-class computational protein design software with directed evolution technologies to generate de novo cytokine mimetics that circumvent limitations for naturally-derived cytokine drugs. Rather than modifying existing molecules, we adopt a bold new approach to create exceptionally stable proteins with customized receptor interaction properties as robust and efficacious targeted therapeutics. We applied this strategy to develop a mimetic of the interleukin-2 (IL-2) cytokine, which plays a pivotal role in T and natural killer (NK) cell function. We created a hyper-stable de novo protein that is biased toward activation of IL-2 signaling pathways on immune effector cells, and showed that this molecule inhibits tumor growth and prolongs survival in mouse models of colorectal cancer and melanoma, while also mitigating the toxic side effects typically associated with IL-2 therapy. We built upon this exciting advance to design de novo cytokines that mimic IL-4, which promotes pro-regenerative immune programs following tissue damage. We demonstrated the extreme thermal stability and biased receptor activity of our engineered IL-4 mimetic, and showed that it elicited pro-regenerative immune responses in both cellular and animal models. Collectively, our work pioneers a novel cytokine engineering platform that integrates computational and experimental approaches to establish a general paradigm for advancing the clinical translation of cytokine therapeutics.