Antibodies are key molecules of our immune system and fight infections with exquisite specificity. As such, antibodies are natural therapeutics. In fact, so-called monoclonal antibodies are the fastest growing class of biological drugs and have revolutionized the treatment of human diseases. However, current platforms for design and selection of monoclonal antibodies are essentially stochastic, expensive and time-consuming. Therefore, there is an urgent need for more innovative and sophisticated strategies for designing the next-generation of tailored antibodies. Two major desirable features of therapeutic antibodies are: (1) long plasma half-life and (2) fine-tuned target specificity. Recent data have revealed surprisingly large differences in plasma half-life of such antibodies. The reasons for this, however, remain poorly understood, but are clearly related to the antibody variable region. In addition, our preliminary data suggest that antibodies binding the same target vary widely in their sequence composition. How specificity is encoded into the antibody variable region remains equally unclear. Thus, in order to digitize antibody discovery, there is a need to gain a comprehensive understanding of how antibody variable region sequences affect antigen binding and half-life, and hence therapeutic outcome. In this transdisciplinary project, we aim to develop a digital antibody discovery platform (DigiAb) that will allow machine-learning-driven computational selection and design of antibodies with both favorable target specificity and superior plasma half-life. This will be achieved by combining state-of-the-art biochemical methodology with computational modeling of single-cell antibody sequencing datasets in an academic-industry collaboration. Our project will reveal new principles in antibody biology and machine-learning for antibody discovery, which will pave the way for rule-driven fast and effortless digital design of tailored therapeutic antibodies.
Project leader: Victor Greiff
Institution: Institutt for klinisk medisin