Advances in AI-designed base editors pave the way for next-generation gene therapy

Researchers have made significant strides in developing advanced gene-editing tools through the integration of artificial intelligence and structural biology, marking a pivotal moment in the field of genetic medicine. A collaborative effort led by Daesik Kim, a professor at Sungkyunkwan University, alongside Yong-Sub Kim from the University of Ulsan College of Medicine and Jae-Hyun Park from Sungkyunkwan University School of Medicine, has resulted in the creation of "OpenABE," a new class of adenine base editors designed to enhance therapeutic outcomes. The breakthrough comes after years of challenges faced by earlier AI-generated gene editors, which often exhibited limited efficacy and raised concerns regarding off-target effects. These issues hindered their application in clinical settings due to safety risks. The newly developed OpenABE variants, specifically "OpenABE 1.1" and "OpenABE 1.2," demonstrate a marked improvement, achieving up to 36 times higher gene-editing efficiency compared to previous models. This advancement was achieved through a meticulous process involving structure-guided protein engineering, which allowed the researchers to refine the molecular architecture of the base editors. Using AlphaFold-based structural predictions, the team mapped the three-dimensional configurations of the base editors, identifying crucial elements that contribute to their stability and function. By modifying these structures with targeted mutations and incorporating specialized appendages inspired by successful conventional editors, the researchers enhanced the performance of the new tools. The result is a set of gene editors that not only match the effectiveness of the widely recognized ABE8e but also exhibit superior precision in targeting specific genetic sequences. In addition to improving efficiency, the team addressed long-standing issues related to off-target edits, ensuring that modifications occur exclusively at the intended sites. This level of accuracy is essential for safe therapeutic applications, particularly when dealing with complex genetic conditions. Moreover, the research extends beyond nuclear DNA, demonstrating the capability to perform precise edits within mitochondrial DNA, a domain previously challenging to manipulate due to its distinct structural characteristics. To facilitate safe delivery of these editors into cells, the researchers utilized engineered virus-like particles (eVLPs), enabling controlled and selective modification of specific genes. This method ensures minimal interference with other parts of the genome, further enhancing the safety profile of the new tools. Kim emphasized that overcoming the shortcomings of early AI-designed editors through innovative structural approaches has opened new avenues for treating genetic disorders effectively. The implications of this research extend beyond individual therapies. The team has also unveiled "OpenCRISPR-1," another AI-driven gene editor that matches the performance of traditional Cas9 systems while drastically reducing off-target mutations. By integrating OpenCRISPR-1 with prime editing technologies and eVLP delivery systems, the researchers are laying the groundwork for a versatile platform capable of addressing a wide array of genetic conditions with unprecedented precision and safety.