Optimization algorithms for functional deimmunization of therapeutic proteins.

Area of Science:

• Biotechnology
• Immunology
• Computational Biology

Background:

• Protein therapeutics can trigger anti-biotherapeutic immune responses, primarily through T-cell epitope recognition by antigen-presenting cells.
• Deimmunization strategies aim to reduce immunogenicity by minimizing or altering these epitopes.
• Crucially, mutations for deimmunization must not negatively impact the protein's therapeutic efficacy.

Purpose of the Study:

• To develop advanced computational methods for deimmunizing protein therapeutics.
• To simultaneously reduce immunogenicity and maintain therapeutic activity through strategic point mutations.
• To provide a flexible and globally optimal approach applicable to diverse protein classes.

Main Methods:

• A dynamic programming algorithm was employed to identify sets of conservative point mutations.
• The method integrates predictions of T-cell epitope occurrence with protein stability and activity analyses.
• The approach allows user-defined constraints on the number of mutations and targeted MHC alleles.

Main Results:

• The developed methods identify optimal and near-optimal mutation sets to minimize predicted T-cell epitopes.
• Case studies demonstrate that the generated variants exhibit improved immunogenicity and/or stability compared to previous methods.
• The output provides a scored list of mutations predicted to enhance deimmunization while preserving protein function.

Conclusions:

• Global optimization algorithms combined with immunogenicity and stability prediction offer a powerful tool for protein engineers.
• This approach facilitates exploration of sequence space to design more effective and safer protein therapeutics.
• The methodology provides valuable insights into the complex interplay between sequence modifications and biological function.