High-throughput engineering of bispecific antibodies

Protabit LLC received $304,602 in Project Grant funding from the National Institute of Allergy and Infectious Diseases (NIAID) under the Allergy and Infectious Diseases Research program (CFDA 93.855). The award, effective August 7, 2025, with completion targeted for July 31, 2026, supports the development of high-throughput engineering technologies for bispecific antibodies—therapeutic molecules that combine two different antigen-binding domains. The project addresses a critical gap in bispecific antibody development, where current low-throughput, trial-and-error approaches fail to efficiently identify and characterize candidates with desirable properties while minimizing liabilities such as polyreactivity, poor expression, and high viscosity.

The primary products and technical capabilities to be delivered include: (1) a yeast-display platform enabling correct display and pairing of intact, full-length bispecific immunoglobulin G (IgG) antibodies on cell surfaces at significantly higher throughput than existing single-domain approaches, and (2) long-read sequencing methodologies capable of identifying complete bispecific IgG sequences in single reads. Collectively, these capabilities will enable the expression, assay, and characterization of millions of unique bispecific IgGs, providing the technical foundation necessary to apply artificial intelligence and machine learning approaches to engineering high-performing bispecific antibodies for therapeutic development.

Abstract

Bispecific antibodies are extremely valuable therapeutic biologic molecules and make up a growing number of biologic drugs on the market. However, they are more complicated and riskier to develop than monoclonal antibodies, as there are no high-throughput engineering strategies that reflect the final structural form of the antibody. As a result, liabilities emerge when two antigen-binding domains are put together, like polyreactivity, poor expression, and high viscosity. The standard corrective actions are low-throughput, trial and error approaches, often requiring the discovery of new parental antibodies. Thus, there is an unmet need for new technologies for bispecific IGG-like antibody development. We propose a novel therapeutic, CO-evolved bispecific antibodies (cebabs). These therapeutics comprise two different antigen binding (FAB) domains that are engineered with high-throughput technologies. Cebabs are distinguished from prior bispecific antibodies in that the FAB domains are affinity-matured simultaneously, allowing assays for downstream liabilities to be performed while the sequence-function landscape is being explored. Developing cebabs will require a protein engineering platform that accommodates orders of magnitude more throughput than those used for single fabs. In this proposal we will develop the technical capabilities to correctly display intact, properly paired, full-length bispecific iggs on the surface of yeast cells. The second major innovation in this work will be the development of long read sequencing capabilities that identifies a full-length bispecific IGG with a single read. Having both of these capabilities will allow US to express, assay and characterize millions of unique bispecific iggs at high - throughput, providing the necessary technology to gather the data that is essential for engineering high performing bispecific antibodies using AI and machine learning approaches. In phase II we will exploit this innovative technology to discover and develop an array of therapeutic bispecific antibodies for known targets where bispecific antibodies are expected to provide a new mode of action.