AI-designed proteins outperform viral and lipid nanoparticles in RNA delivery.

Researchers at Helmholtz Munich and the Technical University of Munich have developed STV-C8, a novel RNA transporter built from AI-designed proteins that significantly outperforms existing delivery mechanisms in preclinical trials. Published in Nature in 2026, the study introduces a bottom-up engineered platform that leverages generative artificial intelligence to create synthetic protein scaffolds with non-natural geometries optimized for intracellular RNA delivery. Traditional RNA therapeutics face delivery barriers due to the molecule's vulnerability outside cells. Current solutions rely on virus-derived carriers or lipid nanoparticles, both of which present limitations in efficiency, scalability, and targeting. The German research team bypassed these constraints by assembling functional protein domains onto an AI-generated structural core. During development, the team evaluated over one hundred variants and found that configurations diverging from natural viral capsid architectures yielded superior transfection rates. STV-C8 emerged as the optimal candidate, demonstrating a modular architecture capable of carrying diverse RNA payloads and enabling cell-type-specific delivery. Preclinical validation confirmed the platform's efficacy and safety. In cell culture assays, STV-C8 achieved substantially higher transfection efficiency than benchmark lipid nanoparticles and viral vectors, requiring significantly lower RNA doses to produce equivalent protein expression levels. In vivo testing in mice following intravenous administration showed targeted RNA expression primarily in pulmonary tissue, with no detectable immunological reactions or systemic toxicity. Further functional validation in a porcine model demonstrated successful CRISPR/Cas9 delivery to skeletal muscle, resulting in precise excision of a disease-relevant segment of the dystrophin gene, a genetic defect associated with Duchenne muscular dystrophy. The experimental system represents a scalable alternative to biologics reliant on biological templates. Researchers emphasize that the AI-driven design process systematically expands the protein engineering space, allowing rapid iteration and customization for therapeutic applications. While promising, STV-C8 requires additional optimization before clinical translation. Ongoing investigations will focus on refining tissue-specific targeting mechanisms and mapping systemic biodistribution profiles to ensure consistent performance across patient populations. The research team, led by Dr. Christoph Gruber and co-first authors Dr. Maren Kirstin Schuhmacher and Dr. Florian Giesert, has secured plans to transfer the technology to a dedicated spin-off company. Project leader Prof. Wolfgang Wurst characterized STV-C8 as a foundational platform ready for expansion across multiple therapeutic pipelines. The development marks a significant advancement in synthetic biology and RNA medicine, positioning AI-optimized protein vectors as a viable next-generation solution for targeted genetic therapies.