Researchers have developed a new protein nanoparticle platform called I53cs. Designed using artificial intelligence (AI), I53cs demonstrated promising performance in animal studies and offers a versatile approach for developing next-generation vaccines.

The study was led by Prof. WANG Xiangxi from the Institute of Biophysics of the Chinese Academy of Sciences, in collaboration with Jiangsu Recbio Technology Co., Ltd. The findings were published in Protein & Cell on July 10.

Protein nanoparticles have emerged as a promising vaccine platform because they display antigens in a highly ordered, multivalent manner, thereby enhancing immunogenicity and the durability of protective immune responses. However, many existing nanoparticle scaffolds are limited by scaffold-directed immune responses, suboptimal manufacturability, and restricted structural flexibility, highlighting the need for improved and more versatile antigen-display platforms.

Through the integration of AI-guided protein design, structural characterization, and animal immunization studies, the researchers established a comprehensive validation framework that provides a new technological strategy for developing next-generation vaccines.

The researchers first established a library of protein nanoparticle scaffolds. Using symmetry as the central design principle, they generated higher-order symmetric nanoparticles with diverse geometries through high-dimensional symmetry matching and hierarchical docking based on a single symmetric scaffold.

They subsequently evaluated the candidate scaffolds using structure prediction models, including AlphaFold and RosettaFold, together with sequence-design algorithms such as ProteinMPNN and SCUBA for sequence optimization and assembly screening. This workflow yielded a series of mature immunopotentiating scaffolds with tunable particle size, symmetry, and antigen-display valency.

Using this strategy, the researchers successfully constructed multiple nanoparticle architectures, including octahedral, icosahedral, and double-layer symmetric assemblies. The designed nanoparticles have diameters ranging from approximately 24 to 32 nm and can support 48 to 120 copies of displayed antigens. This provides a modular, customizable platform that can accommodate diverse antigens and immunization requirements.

Using this platform, the researchers generated monovalent, trivalent, and mosaic nanoparticle vaccines that display the hemagglutinin (HA) proteins of H1N1, H3N2, and influenza B/Victoria viruses.

Animal studies demonstrated that the I53cs platform rapidly elicits high levels of hemagglutination-inhibiting antibodies. Compared to soluble HA proteins and a licensed trivalent inactivated influenza vaccine, the nanoparticle vaccines generated stronger, more durable antibody responses and broad immunity against multiple influenza strains simultaneously.

These results suggest the potential of the I53cs platform as a versatile technology for developing next-generation influenza vaccines.

AI-designed protein nanoparticles enable multivalent display of influenza HA antigens, eliciting faster, stronger, more durable, and broader antibody responses. (Image by WANG Xiangxi's group)