A new artificial intelligence-based pipeline has been developed to improve the design of intracellular antibodies, or intrabodies, which function within living cells. This method integrates protein structure prediction, sequence redesign, and live-cell screening to convert conventional antibody sequences into stable and functional intrabodies.
Antibodies are essential tools in biology and medicine due to their ability to selectively recognize specific targets. However, most traditional antibodies only operate outside of cells, limiting their usefulness for studying processes inside living cells. Developing intrabodies has been difficult because antibodies often misfold or lose activity within cellular environments.
The research team was led by Professor Hiroshi Kimura from the Institute of Integrated Research at the Institute of Science Tokyo (Science Tokyo), along with Mr. Daiki Maejima, a doctoral student at Science Tokyo, Associate Professor Timothy J. Stasevich from Colorado State University in the United States, and Professor Yasuyuki Ohkawa from Kyushu University in Japan. Their findings were published in Science Advances on January 2, 2026.
"We created a pipeline that combines artificial intelligence (AI)-guided protein structure prediction with sequence redesign and live-cell screening," said Kimura.
This approach maintains the antigen-binding regions of antibodies while modifying other framework regions to ensure proper folding and stability inside cells without losing target specificity.
The researchers tested 26 existing antibody sequences using their AI-driven strategy. Of these, 19 were successfully converted into functional intrabodies; notably, 18 had previously failed as intrabodies when using standard methods.
"Many antibodies that were not functional when expressed as intrabodies were ultimately found to be converted into functional ones in our study," Kimura commented. "This confirms that AI allowed us to redesign structures that were compatible within the cellular environment."
The study focused on intrabodies targeting modifications in histone proteins—key regulators of DNA packaging and gene activity—which are challenging to analyze with traditional techniques. The newly designed intrabodies detected changes in histone modifications inside living cells through fluorescence signals as modification levels varied.
Further testing showed that these redesigned molecules remained stable, soluble, specific to their targets, and performed consistently under different cellular conditions.
"By combining AI-based design with live-cell testing, we can now accelerate intrabody development with far greater confidence," Kimura explained.
The researchers suggest this approach could make developing new intrabodies faster and less expensive. As databases of antibody sequences grow, converting existing antibodies into effective intracellular probes may expand applications across diagnostics, imaging technologies, and therapeutic development.
