(190a) A Hidden Light – Selection of Green Fluorescent Protein That Evades an Existing Antibody Response
AIChE Annual Meeting
2018
2018 AIChE Annual Meeting
Food, Pharmaceutical & Bioengineering Division
Poster Session: Engineering Fundamentals in Life Science
Monday, October 29, 2018 - 3:30pm to 5:00pm
The development and deployment of protein therapeutics have transformed modern medicine. Despite their exquisite functionality and specificity, the use of such biotherapeutics presents a unique set of challenges compared to conventional small molecule drugs. One of the major issues that can plague biotherapies is their propensity to interact with the host immune system, often resulting in production of deleterious anti-drug antibodies. Such anti-drug immune responses can result in changes in pharmacokinetics, toxic immune complex formation, decreased efficacy, and potentially severe allergic reactions. In light of these facts, the development of mitigation strategies that reduce the effects of immunogenicity are necessary for the continued use and development of protein based drugs.
In the case of an existing anti-drug antibody response, a biotherapeutic may be deimmunized by mutagenesis of antigenic sites on the protein surface. However, so called “B cell epitope deletion†requires structural knowledge of the protein’s antigenic sites, which is often the result of laborious epitope mapping studies or co-crystalization experiments. We have implemented a selection strategy whereby unknown epitopes, associated with a polyclonal anti-drug antibody response, are disrupted via flow cytometric screening of large mutagenic libraries. Error-prone libraries of a biotherapeutic agent are displayed on the yeast cell surface, and the library population is probed with polyclonal anti-drug serum from animal models or even human patients. Simultaneous staining for full length expression tags enables rapid selection of engineered variants bearing mutations that disrupt antibody binding while also passing the yeast protein quality control system. Here, we describe the redesign of green fluorescent protein (GFP) so as to evade a strong polyclonal antibody response generated in humanized mice. In the case of GFP, the protein’s molecular function is readily integrated into the high throughput screen. Presented with this opportunity, we have combined multicolor sorting logic and deep sequencing of distinct sorting trajectories to construct a map of GFP sequence space as it relates to evasion of antibody binding and maintenance of molecular function. Beyond GFP as a proof of principle, and with some modification, the methods described here could be utilized to generate biosimilars of existing therapeutic proteins that are known to elicit an anti-drug antibody response.
In the case of an existing anti-drug antibody response, a biotherapeutic may be deimmunized by mutagenesis of antigenic sites on the protein surface. However, so called “B cell epitope deletion†requires structural knowledge of the protein’s antigenic sites, which is often the result of laborious epitope mapping studies or co-crystalization experiments. We have implemented a selection strategy whereby unknown epitopes, associated with a polyclonal anti-drug antibody response, are disrupted via flow cytometric screening of large mutagenic libraries. Error-prone libraries of a biotherapeutic agent are displayed on the yeast cell surface, and the library population is probed with polyclonal anti-drug serum from animal models or even human patients. Simultaneous staining for full length expression tags enables rapid selection of engineered variants bearing mutations that disrupt antibody binding while also passing the yeast protein quality control system. Here, we describe the redesign of green fluorescent protein (GFP) so as to evade a strong polyclonal antibody response generated in humanized mice. In the case of GFP, the protein’s molecular function is readily integrated into the high throughput screen. Presented with this opportunity, we have combined multicolor sorting logic and deep sequencing of distinct sorting trajectories to construct a map of GFP sequence space as it relates to evasion of antibody binding and maintenance of molecular function. Beyond GFP as a proof of principle, and with some modification, the methods described here could be utilized to generate biosimilars of existing therapeutic proteins that are known to elicit an anti-drug antibody response.