Next-generation anti-inflammatory protein designed by AI, 50% more effective than existing therapeutics

□ A research team led by Professor Jiwon Um of the Department of Brain Sciences at DGIST (President Kunwoo Lee) and CEO Iksoo Chang of iProtein Therapeutics (a former Professor of the Department of Brain Sciences at DGIST) developed a ‘next-generation anti-inflammatory protein’ using AI and supercomputing. They demonstrated its effectiveness through animal testing. The new protein showed up to 53% better anti-inflammatory effects compared to current treatments, and it is expected to significantly improve the treatment of various inflammatory diseases, including rheumatism, gout, and autoimmune diseases.

□ ‘Anakinra,’ a leading anti-inflammatory protein treatment currently in clinical use, blocks the IL-1 signal, which causes inflammation. However, it has limitations such as relatively low efficacy, a short duration of action, and an increased risk of infection and injection site inflammation at high doses. For these reasons, there has been a persistent demand for the development of stronger and safer alternative therapeutics.

□ To address these issues, the research team adopted a protein structure-guided design approach. Researchers from DGIST and iProtein Therapeutics utilized supercomputing-based molecular dynamics simulations and thermodynamic analyses to identify the key amino acid E127, which influences the binding strength of the IL-1Ra protein. Subsequently, they designed and produced six new protein variants centered around this region.

□ As a result, all variants showed anti-inflammatory effects 25-53% higher than the current treatment, with the ‘E127Q’ variant demonstrating the highest effectiveness. Notably, in brain neuron electrophysiological experiments, E127Q restored NMDA receptor signals, overactivated by inflammation, to normal levels, thereby showcasing its ability to effectively control ‘neuroinflammation,’ which had been difficult to address with existing anti-inflammatory proteins. In an animal model of chronic neuroinflammation, E127Q exhibited significantly better anti-inflammatory effects and neuronal function restoration compared to the original protein.

□ Professor Jiwon Um said, "This research is significant because a protein designed using AI and supercomputing has demonstrated higher efficacy than existing drugs in actual cell and animal experiments. In particular, the E127Q variant has great potential to be developed as a treatment for systemic inflammatory diseases and brain neuroinflammation, establishing it as a next-generation anti-inflammatory protein platform."

□ The results of this study were published on November 30 in Theranostics (IF=13.3, SCIE Q1, top 5% in the field), a globally recognized academic journal in biomedical and bioengineering fields. This research was funded by the Ministry of Science and ICT and the National Research Foundation of Korea through the Mid-Career Researcher Support Program, the Global Leader Research Program, the Basic Research Laboratory Support Program, and the Sejong Fellowship.