KAIST Develops AI-Designed Nasal Spray That Blocks Respiratory Viruses

Korean researchers have developed a nasal antiviral platform that uses artificial intelligence technology to combat respiratory viruses.

KAIST announced Wednesday that a joint research team led by Professors Kim Ho-min and Chung Hyun-jung from the Department of Biological Sciences and Professor Oh Ji-eun from the Graduate School of Medical Science and Engineering has created a technology capable of broadly preventing various respiratory viruses. The team used AI to redesign interferon-lambda (IFN-λ) protein for enhanced stability and combined it with a delivery system that enables effective diffusion and prolonged retention in nasal mucosa.

Interferon-lambda is an innate immune protein that the human body naturally produces to fight viral infections and plays a crucial role in blocking respiratory viruses such as the common cold, influenza, and COVID-19. However, when formulated as a therapeutic and administered nasally, its effectiveness has been limited due to vulnerability to heat, degrading enzymes, mucus, and ciliary movement.

The research team used AI protein design technology to precisely address the structural weaknesses of interferon-lambda.

First, they significantly improved stability by converting loose loop structures in the protein, which caused instability, into helix structures that remain fixed like tight springs.

The team also applied "surface engineering" to make protein surfaces more water-soluble, preventing proteins from clumping together. Additionally, they introduced "glycoengineering" to add glycan structures to the protein surface, making the protein more robust and stable.

As a result, the newly designed interferon-lambda showed dramatically improved stability, capable of withstanding temperatures of 50°C for two weeks, and demonstrated rapid diffusion characteristics even in sticky nasal mucosa.

The researchers further protected the protein by encapsulating it in microscopic capsules called nanoliposomes and coating the surface with low-molecular-weight chitosan to significantly enhance mucoadhesion, allowing it to adhere to nasal mucosa for extended periods.

When applied to influenza-infected animal models, this delivery platform demonstrated powerful inhibitory effects, reducing nasal viral load by more than 85 percent.

This technology represents a mucosal immunity platform that can block viral infections at early stages through simple nasal spraying. It is expected to serve as a new treatment strategy capable of rapidly responding not only to seasonal influenza but also to unexpected new and variant viruses.

"By combining AI-based protein design with mucosal delivery technology, we have simultaneously overcome the stability and retention time limitations of existing interferon-lambda therapeutics," Professor Kim said. "This platform, which remains stable at high temperatures and adheres to mucosa for extended periods, is an innovative technology that can be utilized even in developing countries lacking strict cold chain infrastructure. It has great potential for expansion into various therapeutics and vaccine development, and represents meaningful results achieved through multidisciplinary convergence research spanning AI protein design, drug delivery optimization, and immune evaluation through infection models."

Dr. Yoon Jung-won from KAIST InnoCORE AI-Innovative Drug Research Center, Dr. Yang Seung-joo from the Department of Biological Sciences, and doctoral candidate Kwon Jae-hyuk from the Graduate School of Medical Science and Engineering participated as co-first authors in this research. The findings were published consecutively in the prestigious international journals Advanced Science (November 20) and Biomaterials Research (November 21).