Human interleukin-4 (IL-4) is a critical therapeutic target for allergic diseases and cancer, yet current biologics face stability and specificity limitations. We report a novel strategy combining de novo computational design with directed evolution to engineer a D-protein inhibitor targeting IL-4. Unlike stochastic screening, our approach enables epitope-specific design against the mirror-image D-IL-4 structure. Crucially, we integrated WALTZ-guided aggregation prediction into the evolution cycle to simultaneously optimize both binding affinity and solution behavior. The resulting D-protein, D-18252-evo, binds native IL-4 with nanomolar affinity (∼87 nM) and effectively blocks receptor engagement. Functional assays confirm potent inhibition of IL-4-induced STAT6 phosphorylation and cell proliferation. Furthermore, D-18252-evo exhibits exceptional biophysical properties, including high thermal stability and resistance to proteolytic degradation. This work establishes a scalable framework for generating robust mirror-image therapeutics, positioning D-proteins as a promising next-generation platform for treating cytokine-driven disorders with enhanced stability and targeted efficacy.
Xu et al. (2026) studied this question.