A generalizable system for antigenic peptide targeting across HLA-I allotypes

Area of Science:

• Immunology and Cancer Therapeutics
• Protein Engineering and Structural Biology

Background:

• T cell receptors (TCRs) and TCR-mimicking antibodies target peptide antigens presented by Human Leucocyte Antigen (HLA-I) molecules.
• HLA polymorphism presents a major hurdle for developing broadly applicable immunotherapies, limiting coverage across diverse patient genetic backgrounds.
• Existing therapeutic modalities struggle to address the molecular diversity of HLA proteins and peptide antigens.

Purpose of the Study:

• To develop a generalized framework for engineering cross-HLA compatible binders (xTRACeRs) that maintain peptide specificity.
• To overcome the limitations imposed by HLA allotype diversity in the development of targeted immunotherapies.

Main Methods:

• Combined a peptide conformational prediction tool (PepPred) with a cross-HLA binding protein engineering system (TRACeR-I).
• Developed and validated xTRACeRs against clinically relevant peptide antigens across common alleles within five HLA-A/B/C supertypes.
• Utilized Cryo-electron microscopy (Cryo-EM) to determine the structural basis of xTRACeR-peptide-HLA complexes.

Main Results:

• Successfully developed xTRACeRs demonstrating compatibility across different HLA allotypes while maintaining high peptide specificity.
• Cryo-EM structures revealed mechanisms by which xTRACeRs navigate polymorphic HLA residues and interact with peptide antigens (e.g., PRAME, PHOX2B).
• Engineered xTRACeRs, when implemented as Chimeric Antigen Receptor (CAR) T cells, showed potent and specific tumor cell killing efficacy.

Conclusions:

• The developed framework provides a generalized approach to create binders effective across diverse HLA supertypes.
• Overcoming HLA supertype restriction significantly expands patient coverage for HLA-targeted immunotherapies.
• This strategy holds promise for advancing the development of more inclusive and effective cancer treatments.