Analysis and prediction of antibody stability
Lead Research Organisation:
University College London
Department Name: Structural Molecular Biology
Abstract
Strategic Research Priority: World Class Bioscience
In recent years, antibody based therapeutics have become essential in the treatment of a range of diseases, from cancer to neurological conditions (Li J. and Zhu Z, 2010). In fact, in 2015 seven of the top 10 grossing drugs were biologics, of which five were monoclonal antibodies (mAbs) (PharmaCompass.com, 2017). Additionally, the most successful drug, in terms of sales, Humira, is a monoclonal antibody used in the treatment of rheumatoid arthritis (Azevedo V. F. et al., 2015). The increasing number of regulatory approvals and market share of this type of therapeutics highlights its importance in modern medicine. The wide success of the application of antibodies in the treatment of disease is due to their ability to bind to a range of targets with high affinity, and with few side effects (Carter P. J., 2006). In fact, problems such as immunogenicity, where a patient develops an immune response to the therapeutic antibody, are being overcome with advances in the humanisation process (Safdari Y. et al., 2013). However, the issue with biologics, including antibodies, is their reduced effectiveness after long term storage. This is due their instability and aggregation propensity, which is not fully understood. Early detection of unsuitable mAbs during the antibody discovery process, and the identification of key residues involved in their stability, could enable the development of antibody therapeutics with improved biophysical properties.
The aim of this project is to develop a computational model that can predict antibody stability. To accomplish this over 900 Fabs (antigen-binding fragments) are being synthesised in partnership with UCB Celltech. These Fabs are derived by pairing non-cognate heavy and light chains from an in-house library of antibody variable regions isolated from human IgG. This medium-sized artificial library is currently being characterised in terms of the biophysical properties of each Fab. This data will then be used to find patterns, using machine learning algorithms, to improve the current understanding of antibody aggregation.
In recent years, antibody based therapeutics have become essential in the treatment of a range of diseases, from cancer to neurological conditions (Li J. and Zhu Z, 2010). In fact, in 2015 seven of the top 10 grossing drugs were biologics, of which five were monoclonal antibodies (mAbs) (PharmaCompass.com, 2017). Additionally, the most successful drug, in terms of sales, Humira, is a monoclonal antibody used in the treatment of rheumatoid arthritis (Azevedo V. F. et al., 2015). The increasing number of regulatory approvals and market share of this type of therapeutics highlights its importance in modern medicine. The wide success of the application of antibodies in the treatment of disease is due to their ability to bind to a range of targets with high affinity, and with few side effects (Carter P. J., 2006). In fact, problems such as immunogenicity, where a patient develops an immune response to the therapeutic antibody, are being overcome with advances in the humanisation process (Safdari Y. et al., 2013). However, the issue with biologics, including antibodies, is their reduced effectiveness after long term storage. This is due their instability and aggregation propensity, which is not fully understood. Early detection of unsuitable mAbs during the antibody discovery process, and the identification of key residues involved in their stability, could enable the development of antibody therapeutics with improved biophysical properties.
The aim of this project is to develop a computational model that can predict antibody stability. To accomplish this over 900 Fabs (antigen-binding fragments) are being synthesised in partnership with UCB Celltech. These Fabs are derived by pairing non-cognate heavy and light chains from an in-house library of antibody variable regions isolated from human IgG. This medium-sized artificial library is currently being characterised in terms of the biophysical properties of each Fab. This data will then be used to find patterns, using machine learning algorithms, to improve the current understanding of antibody aggregation.
People |
ORCID iD |
| Andrew Martin (Primary Supervisor) | |
| Gil Ferreira Hoben (Student) |
Studentship Projects
| Project Reference | Relationship | Related To | Start | End | Student Name |
|---|---|---|---|---|---|
| BB/M009513/1 | 01/10/2015 | 31/03/2024 | |||
| 1705637 | Studentship | BB/M009513/1 | 26/09/2016 | 25/09/2020 | Gil Ferreira Hoben |