AI辅助抗体设计进入快车道 药物安全问题仍需进一步验证

Core Insights - AI technology has shown unprecedented potential in therapeutic antibody design since the advent of "AlphaFold 2" for protein structure prediction [1] - Multiple research teams have successfully developed various therapeutic antibodies using proprietary AI tools, although safety and efficacy still require further validation [1] Antibody Drug Market - Antibodies are key proteins in the immune system that recognize specific targets and trigger protective responses, with over 160 engineered antibodies approved for treating cancer, infectious diseases, and autoimmune diseases globally [2] - The global antibody drug market is projected to exceed $455 billion in annual revenue by 2028, driven by the emergence of thousands of new antibodies [2] - Traditional antibody development faces challenges such as long cycles and high costs, but recent AI advancements are transforming the paradigm of antibody research and development [2] Innovative Developments - Several research teams have successfully developed a range of functional antibody drugs using AI platforms [3] - Absci announced the design of a specific antibody targeting a conserved region of the HIV virus, which could lead to a broad-spectrum anti-HIV drug [3] - The BoltzGen AI model, developed by a team led by Gabriel Corson, focuses on de novo design of proteins and peptides, achieving atomic-level precision in structural modulation [3] Significant Progress by Other Teams - A team led by David Baker discovered a broad-spectrum antibody that can bind to proteins common to all influenza viruses, paving the way for universal flu drugs [4] - Nabla and Chai Discovery successfully designed full-length antibodies that can specifically recognize GPCRs, which are traditionally difficult to target [4] - Nabla generated thousands of GPCR-binding antibodies, with some showing comparable or superior affinity to existing drugs [4] Impact on Clinical Development - The current wave of AI-driven antibody design is expected to significantly impact the number of clinical candidates and the efficiency of drug development [5] - AI-designed antibodies may soon enter human trials, as demonstrated by Genative's large-scale clinical trial for an antibody drug targeting severe asthma [6] Safety Validation Challenges - Despite advancements, AI-generated antibodies still face challenges in performance across different targets and predicting binding strength [6] - There is a need for rigorous preclinical safety evaluations to determine if AI-designed antibodies will be recognized as foreign by the human immune system [6] - Future AI designs may create antibodies with special functions, such as penetrating the blood-brain barrier or targeting multiple sites simultaneously [6]