ABSTRACT
Cancer cell-killing by CD8+ T cells demands effective tumor antigen presentation by human leukocyte antigen class I (HLA-I) molecules. Screening and designing highly immunogenic neoantigens require quantitative computations to reliably predict HLA-peptide binding affinities. Here, with all-atom molecular dynamics (MD) simulations and free energy perturbation (FEP) methods, we design a collection of antigenic peptide candidates through in silico mutagenesis studies on immunogenic neoantigens, yielding enhanced binding affinities to HLA-B*44:02. In-depth structural dissection shows that introducing positively charged residues such as arginine to position 6 or lysine to position 7 of the candidates triggers conformational shifts in both peptides and the antigen-binding groove of the HLA, following the "induced-fit" mechanism. Enhancement in binding affinities compared to the wild-type was found in three out of five mutated candidates. The HLA pocket, capable of accommodating positively charged residues in positions from 5 to 7, is designated as the "dynamic pocket". Taken together, we showcase an effective structure-based binding affinity optimization framework for antigenic peptides of HLA-B*44:02 and underscore the importance of dynamic nature of the antigen-binding groove in concert with the anchoring motifs. This work provides structural insights for rational design of favorable HLA-peptide bindings and future developments in neoantigen-based therapeutics.
Subject(s)
Molecular Dynamics Simulation , Peptides , Protein Binding , Humans , Peptides/chemistry , Peptides/immunology , HLA-B44 Antigen/chemistry , HLA-B44 Antigen/immunology , HLA-B44 Antigen/genetics , Computer Simulation , Binding Sites , Protein ConformationABSTRACT
Cross reactivities are known for human leukocyte antigen inside HLA-B7 related Cross-Reactive Group (B7CREG). Some CE-IVD flow-cytometry kits use double monoclonal antibodies (mAb) to distinguish HLA-B27 and HLA-B7 but practice reveals more complexes results. This study explores the performances of this test. Analysis of 466 consecutive cases using HLA-B27 IOTest™ kit on a Navios™ cytometer from Beckman-Coulter, partially compared to their genotypes. Expected haplotypes HLA-B27-/HLA-B7- (undoubtedly HLA-B27 negative) and HLA-B27+/HLA-B7- (undoubtedly HLA-B27+) were clearly identified according to the manufacturer's instructions. On the opposite, patients strongly labeled with anti-HLA-B7 showed three different phenotypes regarding anti-HLA-B27 labeling: (1) most of the cases were poorly labeled in accordance with cross reactivity inside B7CREG (HLA-B27-/HLA-B7+ haplotype); (2) rare cases had strong B7 and B27 labeling corresponding to HLA-B27+/HLA-B7+ haplotype; (3) even less cases had strong labeling by anti-HLA-B7 but non for anti-HLA-B27, all expressing HLA-B44 and no B7CREG molecules. Surprisingly, more cases were not labeled with anti-HLA-B7 antibody but partially labeled with anti-HLA-B27 suggesting another cross reactivity out of B7CREG. mAb HLA typing suggests new, cross reactivities of anti-HLA-B27 antibody to more molecules out of B7CREG and of anti-HLA-B7 antibody but not anti-HLA-B27 to HLA-B44 molecule also out of B7CREG. HLA-B27 could surely be excluded in most samples labeled with HLA-B27, below a "grey zone" on intermediate intensity. More comparison is needed in future studies.