Design of therapeutic proteins with enhanced stability

Abstract

Therapeutic proteins such as antibodies constitute the most rapidly growing class of pharmaceuticals for use in diverse clinical settings including cancer, chronic inflammatory diseases, kidney transplantation, cardiovascular medicine, and infectious diseases. Unfortunately, they tend to aggregate when stored under the concentrated conditions required in their usage. Aggregation leads to a decrease in antibody activity and could elicit an immunological response. Using full antibody atomistic molecular dynamics simulations, we identify the antibody regions prone to aggregation by using a technology that we developed called spatial aggregation propensity (SAP). SAP identifies the location and size of these aggregation prone regions, and allows us to perform target mutations of those regions to engineer antibodies for stability. We apply this method to therapeutic antibodies and demonstrate the significantly enhanced stability of our mutants compared with the wild type. The technology described here could be used to incorporate developability in a rational way during the screening of antibodies in the discovery phase for several diseases.aggregation ͉ antibody ͉ genetic engineering ͉ molecular simulation T herapeutic antibodies are glyco-proteins tailor-made to seek out and attach themselves to specific antigens, such as biomarkers on the surface of cancer cells. Because of their potential in the cure of various diseases, antibodies currently constitute the most rapidly growing class of human therapeutics (1). Since 2001, their market has been growing at an average yearly growth rate of 35%, the highest rate among all categories of biotech drugs (2). One of the major problems encountered in antibody-based therapies is that these antibodies tend to aggregate under the high concentration formulations required for disease treatment (3). Aggregation leads to a decrease in antibody activity and could elicit an immunological response (4, 5). It also has implications for regulatory approval and for delivery methods. Stabilization of therapeutic antibodies, however, is generally performed during the development phase using trial and error methods. These are both costly and time consuming. Thus, there is a need for a screening tool that will assess aggregation behavior. Such a tool could be used, for example, to evaluate the developability of a bio-pharmaceutical coming during the discovery phase. Protein aggregation in vivo is also shown to be directly responsible for many diseases such as type II diabetes and Alzheimer's (6, 7). Thus, an understanding of the fundamental mechanisms and regions involved in protein aggregation is of importance both for stabilizing protein therapeutics and for devising strategies to prevent in vivo aggregation.Although there are many possible mechanisms for aggregation, hydrophobic interactions were shown to be the predominant interactions in extensive studies of protein folding and protein-protein binding (8-13). Our results below show that these interactions also play a key role in antibody...

Homology modeling and structure-based design improve hydrophobic interaction chromatography behavior of integrin binding antibodies

“…In earlier work on this topic, properties were averaged over molecular dynamics simulations. 8 In this study, LowModeMD with implicit solvation was used to generate variable sidechain and flexible loop ensembles of each homology model. LowModeMD focuses velocities on low frequency motions, and is a much faster method for conformational sampling than classical dynamics performed in explicit solvent.…”

“…In the absence of a crystal structure, homology models have been used to derive predictions for properties such as aggregation, viscosity and clearance, 8–11 and structural surface hydrophobicity descriptors have been shown to correlate to HIC RT. 12–14 More recently, multi-parameter protein quantitative structure-property relationship (QSPR) models have been developed to predict high-concentration viscosities of mAbs using surface hydrophobicity and charge descriptors from full antibody homology models and similar protein QSPR models have been reported for predicting chromatographic behavior, including HIC.…”

“…15–17 The earlier work of building single-parameter hydrophobic patch predictors was validated appropriately using experimental data for fewer than twenty sequences. 8,12 For the multi-parameter models that are being constructed and applied to experimental antibody property prediction, 10,15,16 large sets of sequences and experimental data are required for model building and validation. Ideally, such datasets would include negative data in order to robustly train predictive models and advance the field in the direction of data driven computational predictions, e.g., available protein thermostability benchmark datasets have allowed machine learning to be applied, resulting in accurate thermostability predictions.…”

Boosting antibody developability through rational sequence optimization

“…Here, in contrast to cross-beta-sheet dependent aggregates, antibodies associate out of their natively folded conformation either through hydrophobic interactions or patches of locally accumulated charged amino acids on their surface. [11][12][13] These interactions can also affect the conformational equilibrium between folded and partially unfolded states such that association out of the folded state in combination with ever-present transient unfolding increases the likelihood for the formation of cross-beta-sheet aggregates. In addition, transiently formed associated proteins that retain their native conformation can give rise to an increase in the apparent particle size resulting in an exponential increase of viscosity at high protein concentrations.…”

A protein chimera strategy supports production of a model “difficult‐to‐express” recombinant target

“…Reports surrounding production of monoclonal antibodies (mAbs) suggest that unique sequence features influence the effectiveness of recombinant protein production 30, 31, 32. For bacterial expression systems, computational tools have been implemented to predict of surface properties to allow rational design of recombinant targets with increased protein solubility of otherwise insoluble or poorly soluble target proteins 29, 33, 34, 35, 36. Previous work has established that charge and polarity influences protein solubility 37.…”