Abstract Recombinant proteins are indispensable tools in biomanufacturing, underpinning drug development and manufacturing processes ranging from biologics to cell therapies. Cellular pathways are driven by the modulation of varying signaling pathways comprising of numerous proteins and membrane receptors. However, reproducing certain proteins ex vivo poses a significant challenge, especially when the tertiary and quaternary structure are complex. This results in a protein that suffers from limited stability, short half-life, and inconsistent bioactivity which further restricts reproducibility, production costs, and poses a significant barrier to scalable manufacturing. In many cases, traditional protein engineering strategies, while valuable, are constrained by experimental throughput and limited predictive accuracy. To overcome this challenge, AcroAIxTM is an artificial intelligence-driven protein design platform that integrates structural modeling, machine learning, and sequence-function prediction to systematically generate high-performing protein variants. By optimizing key biophysical parameters, AcroAIxTM enables the rational enhancement of recombinant proteins for improved conformational stability, functional half-life, and binding efficiency. These advances are particularly impactful in cell culture applications, where growth factors such as interleukin-21 (IL-21) and fibroblast growth factor 2 (FGF-basic) are essential. Heat-stable forms of IL-21 and FGF-basic were developed to maintain their bioactivity and half-life in 37°C media for 3 days. The resulting cytokine remains active for a longer period and improves cell counts significantly under a reduced-feeding protocol compared to wild types. As such, engineering variants with enhanced stability and bioactivity can directly improves culture robustness, reduce factor replenishment requirements, and increase the efficiency of cell-based manufacturing workflows. By bridging computational design with experimental validation, the AcroAIxTM platform establishes a transformative framework for recombinant protein innovation, advancing both general biomanufacturing practices and cell culture-driven therapeutic production. Citation Format: Spencer Chiang, Jane Liu, Lisa Chou, An Ouyang, Lili Qin. Advancing protein engineering through AcroAIxTM: AI-driven strategies for enhanced structure and function optimization abstract. In: Proceedings of the American Association for Cancer Research Annual Meeting 2026; Part 1 (Regular Abstracts); 2026 Apr 17-22; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2026;86(7 Suppl):Abstract nr 4522.
Chiang et al. (Fri,) studied this question.