A new method for speeding the discovery and selection of targeted, biologically active antibodies, including agonists, is described by researchers at the Scripps Institute, who claim the technology also enables the creation and selection of asymmetric, bispecific antibodies. Current approaches to antibody discovery involve two steps; the first one to screen for binding antibodies expressed in, for example, yeast or phage, and the second assesses antibody hits for function in eukaryotic cells. In contrast, the approach developed by Hongkai Zhang, Ph.D., Ian A. Wilson, Ph.D., and Richard A. Lerner, Ph.D., effectively carries out both tasks at the same time.
The technique involves construction of a combinatorial antibody library in lentiviruses, which are then used to infect mammalian cells that express the antibodies and also secrete them, so both intracellular and extracellular targets can be accessed. Importantly, more than one virus can infect a single cell, which allows the study of combinatorial synergy of multiple proteins.
The invstigators tested the system through an experiment to identify erythropoietin agonist antibodies. Firstly, more traditional techniques were used to rifle through an antibody library to identify thousands of antibodies that bind to the EPO receptor. The antibody genes were then encoded in lentiviruses, and the viruses used to infect a mammalian cell population that expressed EPO receptors. Functional antibodies could then be selected simply by isolating the cells that proliferated most as a result of binding. The experiment yielded an antibody that showed about 60% of the biological activity of natural EPO.
Surprisingly, however, the researchers found that the system also provides opportunities to design highly potent, asymmetric bifunctional antibodies. They noticed that a number of the proliferating cells in the EPO assay had been infected by multiple lentiviruses, and contained sequences from more than one antibody. But when single antibodies were generated from these sequences and tested either individually or even in combination, they demonstrated no EPO agonist activity. Rather, the EPO mimic found in their test cells was actually a bispecific antibody that had self-assembled from two different heavy chains. Analysis of the molecule showed that it not only targeted two EPO receptor binding sites, but also demonstrated 100% of the biological activity of natural EPO.
“Our method differs from that of others who also attempted to generate antibody agonists, not only in the nature of the unique bispecific molecules that it yielded, but also because its power derives from selection rather than screening,” the authors state in their published paper in PNAS. Dr. Zhang et al say their system could just as easily be used to generate antagonist as well as agonist antibodies by changing the selection parameters, and potentially discover new therapeutic targets both for antibody and small molecule compounds. “The real power of this technique is its ability to help us discover the unknown,” concludes co-author Dr. Richard Lerner.
The investigators describe their system in a paper titled “Selection of antibodies that regulate phenotype from intracellular combinatorial antibody libraries.”