Accelerating Therapeutics with In Vitro Antibody Discovery

The development of in vitro antibody platforms is speeding up the pace of drug discovery. “In vitro antibody discovery platforms have transformed antibody generation by enabling the high-throughput screening of billions of antibody variants in a matter of weeks, rather than relying on lengthy animal immunization campaigns that can take months,” says John Cardone, Marketing Manager, Custom Antibody Services at Bio-Rad Laboratories. The phage display method, discussed here, is a technological jumping-off point for innovations like bispecific antibodies, antibody-drug conjugates (ADCs), and other complex therapeutic drug candidates. This article discusses how the advantages of in vitro antibody platforms are advancing therapeutics today.

The beauty of phage display

In phage display, antibody fragments (fused to bacteriophage coat proteins) are displayed on the surface of bacteriophages and are subsequently screened for their ability to bind antigens. A phage display library can display millions to trillions of unique antibody variants, with a broad diversity that increases the chance of finding a “hit.” Such libraries are screened iteratively, and each round of panning enriches for phages whose DNA encodes the much-sought antibody fragments that bind the target.

Phage display’s iterative process is a flexible and effective way to test many antibody sequences simultaneously. Once identified, antibody candidates are optimized before preclinical testing and clinical trials. “Using phage display, researchers can rapidly identify fully human, sequence-defined antibodies with the desired affinity, specificity, and functional properties directly from large recombinant libraries,” says Cardone. “This allows delivery of well-characterized antibody candidates in as little as 8 to 12 weeks while maintaining high success rates.” Bio-Rad’s Pioneer™ Antibody Discovery Platform and SpyLock™ technology support this by moving large libraries of antibodies through high-throughput screening.

In vitro antibody platforms make it possible to discover antibodies based on their binding properties—and link this directly to antibody sequence. “Once a desired clone is identified, its sequence can be recovered immediately, reformatted into the appropriate therapeutic backbone, and progressed into downstream characterization,” says Tracey Mullen, Chief Strategy Officer at Mosaic Biosciences. “The same platform can then be leveraged for affinity maturation, specificity engineering, species cross-reactivity, epitope targeting, and early developability optimization, making in vitro systems highly amenable to iterative engineering.”

Advanced therapeutics

Engineering antibodies to display specific characteristics is producing promising therapeutic modalities in bispecific antibodies and ADCs. A bispecific antibody can bind to two antigens, or two epitopes on one antigen. ADCs consist of a monoclonal antibody linked to a drug payload (e.g., an anti-cancer drug) that is released upon internalization by target cells. “Phage display platforms are ideally suited to these molecules because they enable precise selection against challenging targets, payload-associated epitopes, and specific domains within complex therapeutic constructs under highly controlled conditions,” says Cardone. In vitro platforms enable researchers to search for or engineer antibodies that show additional specific characteristics. “The antibody may need a specific epitope, affinity range, internalization profile, species cross-reactivity pattern, developability profile, or compatibility with a second binding arm or cytotoxic payload,” says Mullen.

Designing advanced therapeutics comes with increased complexity, which means additional optimization. “For an ADC, the antibody must distinguish tumor cells from healthy tissue, bind the appropriate epitope with an optimized affinity, internalize effectively, and maintain properties compatible with conjugation and manufacturing,” says Emily Leproust, CEO and Co-founder of Twist Bioscience. “For a bispecific, each arm must bind the intended target while the complete molecule also achieves the desired geometry, functional activity, stability, expression and developability.”

While this sounds like a tall order, technological advances are meeting the moment. For example, Twist offers optimized assays for the properties of ADC candidates, which allows researchers to make more informed decisions for candidate selection. For bispecifics, the B-Body^® platform can reformat a candidate antibody “into a human IgG-like scaffold that is designed to support efficient pairing, high expression, purity and stability,” says Leproust.

Tests for developability assess whether a candidate will be amenable to manufacturing, safety, storage, and other important stages of a drug’s clinical lifetime. “Increasingly, scientists have been screening for binding, specificity and developability much earlier and in parallel to reduce their downstream risks,” says Leproust. “This will be particularly important for ADCs and bispecifics, where the behavior of the complete molecule is judged on more than just binding affinity.”

Integrated workflows for therapeutics

The rise of next-generation biologics is fueled in part by the advent of integrated workflows —including elements such as automation, high-throughput screening, and/or AI and machine learning—that can speed up development timelines while reducing clinical risks. “Once hits are identified, integrated DNA-to-data workflows can produce tens to thousands of antibodies in parallel for binding affinity, biophysical, developability, and functional screening,” says Leproust. “This moves antibody discovery away from a series of slow, sequential steps and toward a scalable, data-rich process in which discovery, optimization, production and characterization work together.”

The use of machine learning with high-throughput, integrated workflows is likely to increase going forward. “In vitro platforms generate large, structured datasets linking antibody sequence to measured phenotype, which makes them particularly compatible with computational modeling,” says Mullen. “Machine learning can help identify sequence-activity relationships, propose improved variants, and reduce the number of experimental cycles required,” adds Mullen, though she notes that standardized datasets and validation practices still need to catch up.

Integrating other display methods (such as mammalian) into workflows could allow for different evaluative criteria in candidate selection. “Phage and yeast offer enormous diversity and throughput, while mammalian systems provide native IgG folding and more physiologically relevant post-translational processing,” says Mullen. “The future will likely be less about one platform replacing another and more about integrated workflows that use each technology where it adds the most value.” For ADCs, this may allow researchers to select earlier for characteristics such as internalization, trafficking, and tumor cell selectivity.

Recent advances in in vitro antibody platforms will continue to be refined in physiologically oriented directions. “Ultimately, the opportunity is not simply to discover more antibodies,” says Leproust. “It is to generate better antibodies, faster, with the data required to make stronger and more predictive decisions earlier.”