Development of monoclonal antibody formulations to decrease aggregation during long-term storage using molecular simulations

Therapeutic proteins such as monoclonal antibodies (mAbs) have been approved for the treatment of different kinds of diseases in humans. They are often administrated into patients through subcutaneous injection which requires high concentration formulations in order to reach the dosage requirements in a limited volume. At such high concentration, the proteins are likely to form aggregates in a short period of time from days to weeks. However, for a therapeutic protein to be commercially available, a shelf life of greater than 12 months is required. Therefore, the stability of mAbs during manufacturing and storage is vital. And reducing the aggregation propensity is the key to success.

Addition of excipients (such as ions, free amino acids, sugar and small organic molecules) can be used to stabilize mAbs during manufacturing and storage. By altering the types and concentration of excipient molecules the stability can be varied. Studies that focused on applications of formulation stabilisation have been carried out recently, but the molecular mechanisms are undiscovered. From pioneers' experimental data, we know that certain amino acids are used to reduce attractive protein-protein interactions (PPIs) and therefore reducing the viscosity and improving stability. But further knowledge of how excipients affect the viscosity of highly concentrated mAbs as a function of PPIs is vital for processing protein-based therapeutics.

Homology models of clinically approved human mAbs will be built and then this project will focus on designing and optimizing the formulation for those aggregation-prone therapeutic monoclonal antibodies using atomistic cosolvent molecular dynamics (MD) simulations. Cosolvent MD simulations are simple and effective techniques employed for prediction and characterisation of protein allosteric binding sites, but their applications to formulation development remains underexplored. The simulations can give detailed pictures of nonbonded interactions as well as bonded interactions that stabilise proteins at the atomistic level.

In the project, a Cosolvent Analysis Toolkit (CAT, recently developed in Bronowska's lab) will be used to analyse trajectories of mAbs with different combinations and concentrations of excipients in different temperature and pressure ranges. CAT can accurately rank the dynamic interactions between mAbs and excipients using a novel hybrid empirical force filed scoring function, therefore the aggregation-prone "hot spots" within mAbs can be identified and aggregation will be mitigated by adjusting the excipient molecules. Properties such as solvent-exposed surface area (SASA), root-mean-square deviation (RMSD), radial distribution factor (RDF) and per residue conformational entropy will be calculated from the trajectories and relative aggregation propensity can be derived per region. The results will then be compared with experimental data such as size-exclusion chromatography (SEC), small-angle x-ray scattering (SAXS) and dynamic light scattering (DLS) data.
Inclusion of the international industrial partner (Iksuda Therapeutics) will provide a set of experimental data on antibody-drug conjugates (ADC) and will engage in joint iterative design of improved formulations for therapeutic mAbs and ADC.

During the project, an in-depth understanding of the factors and mechanisms causing the protein aggregation on an atomistic level will be developed and excipients that can disrupt the aggregations by non-covalent interactions will be spotted. Although the project is not concentrating on drug discovery directly, it will lead a way to new potential therapeutics based on engineered proteins with improved formulations which is attractive to the pharmaceutical industry.

The project is aligned with EPSRC Biophysics and Soft Matter Physics, and Chemical Biology and Biological Chemistry areas.