Computational re-design of protein structures to improve solubility

Introduction: The rapid development of protein therapeutics is providing life-saving therapies for a wide range of human diseases. However, degradation reactions limit the quality and performance of these protein-based drugs. Among them, protein aggregation is the most common and one of the most challenging to prevent. Aggregation impacts biopharmaceutical development at every stage, from discovery to production and storage. In addition, regulators are highly concerned about the impact of protein aggregates on drug product safety. Area covered: Herein, the authors review existing protein aggregation prediction approaches, with a special focus on four recently developed algorithms aimed to predict and improve solubility using three-dimensional protein coordinates: SAP, CamSol, Solubis and Aggrescan3D. Furthermore, they illustrate their potential to assist the design of solubility-improved proteins with a number of examples. Expert opinion: Aggregation of protein-based drugs is, traditionally, addressed via wet lab experiments, using trial and error approaches that are expensive, difficult to perform and time-consuming. The structure-based in silico methods we describe here can predict accurately aggregation propensities, allowing researchers to work with pre-selected, well-behaved, protein candidates. These methods should contribute to the reduction of the time to the marketplace along with industrial costs and improve the safety of future therapeutic proteins.