Monoclonal antibodies (mAbs) are a promising class of therapeutics, since ca. 25% of all biotech drugs in development are mAbs. Even though their therapeutic value is now well established, human and murine derived mAbs do have deficiencies, such as short in vivo lifespan and low stability. However, the most difficult obstacle to overcome, towards the exploitation of mAbs for disease treatment, is the prevention of the formation of protein aggregates.
ANTISOMA is a pipeline for the reduction of the aggregation tendency of mAbs through the decrease in their intrinsic aggregation propensity, based on an automated amino acid substitution approach. The computational models of mAbs produced by this method could prove important for the experimental improvement of their aggregation propensity, and thus the wider usage of this promising therapeutic approach.Method Description
DSSP detects the amino acid residues that are accessible to the solvent. Residues that belong to CDRs or those that are involved in epitope recognition are also identified. Then, substitutions are applied on critical sites that are located on "aggregation-prone" regions (APRs) based on the prediction by the algorithm AMYLPRED2. The default substitution matrix utilized in the online implementation of ANTISOMA is based on the experimental aggregation propensities of the standard amino acids, described by Sanchez de Groot et al.. Researchers can cautiously proceed to change the substitution matrix if they wish. As an extra step to ensure that the applied mutations will not lead to reduction of protein stability, the FoldX algorithm is integrated in ANTISOMA. FoldX is used to determine whether an applied point mutation increases or decreases the stability of the molecule, taking into consideration solution conditions. All destabilizing mutations are rejected from any further analysis. With the completion of this step all critical mutations that will be applied to a mAb have been detected by our method. Finally, the Modeller software is utilized to construct the three-dimensional model structure of the antibody after mutations have been applied.
ANTISOMA needs either the PDB ID or a .pdb file as input.
The primary output contains:
- Complementarity Determing Regions (CDRs) of light and heavy chain
- A table of rationally designed substitutions, customized for each input file
- A sequence alignment highlighting residues before and after substitutions
- "Aggregation-prone" regions (APRs) predicted by our consensus method AMYLPRED2 (Tsolis, et al. (2013), PLOS ONE 8(1):e54175) before and after substitutions
Advanced output files are also available for download:
- A text file (.txt) containing all primary output results
- Fasta sequence files (.txt) of the mAbs before and after substitutions
- A new model created by Modeller (Webb & Sali (2014), Current Protocols in Bioinformatics, John Wiley & Sons, 5.6:1-32).