The rigidity of a structural bridge on HLA-I binding groove explains its differential outcome in cancer immune response

The tremendous success of immune checkpoint blockade (ICB) therapy has raised the great demand for the development of predictive biomarkers. A recent cancer genomic study suggested that human leukocyte antigen (HLA)-B*44:02 and HLA-B*15:01 alleles may act as potential biomarkers for ICB therapies, however, the underlying molecular mechanisms remain largely elusive. Here, we investigated the molecular origins of differential responses to ICB therapies for four representative HLA alleles: HLA-B*44:02, HLA-B*15:01, HLA-B*07:02, and HLA-B*53:01, using extensive all-atom molecular dynamics simulations. We first demonstrated that the relatively more rigid peptide-binding groove of HLA-B*15:01, than those in the other three HLA alleles, may result in challenges in its recognition with T-cell receptors. Specifically, the "bridge" structure in HLA-B*15:01 is stabilized through both intramolecular electrostatic interactions between the HLA residues and intermolecular interactions between the HLA and the antigenic peptide. These observations were further confirmed by in silico mutagenesis studies, as well as simulations of several other HLA-B*15:01-peptide complexes. By contrast, the "bridge" structure is either completely absent in HLA-B*44:02 or easily perturbed in HLA-B*07:02 and HLA-B*53:01. Our findings provide detailed structural and mechanistic insights into how HLA genotype influences ICB responses and may have important implications for developing immune markers.