Deciphering HLA-I motifs across HLA peptidomes improves neo-antigen predictions and identifies allostery regulating HLA specificity.

Michal Bassani-Sternberg,Philippe Guillaume,David Gfeller,George Coukos,Lana E Kandalaft,Marthe Solleder,Chloé Chong,Philippe O Gannon,Huisong Pak,Tomer Hertz

doi:10.1371/journal.pcbi.1005725

Abstract

The precise identification of Human Leukocyte Antigen class I (HLA-I) binding motifs plays a central role in our ability to understand and predict (neo-)antigen presentation in infectious diseases and cancer. Here, by exploiting co-occurrence of HLA-I alleles across ten newly generated as well as forty public HLA peptidomics datasets comprising more than 115,000 unique peptides, we show that we can rapidly and accurately identify many HLA-I binding motifs and map them to their corresponding alleles without any a priori knowledge of HLA-I binding specificity. Our approach recapitulates and refines known motifs for 43 of the most frequent alleles, uncovers new motifs for 9 alleles that up to now had less than five known ligands and provides a scalable framework to incorporate additional HLA peptidomics studies in the future. The refined motifs improve neo-antigen and cancer testis antigen predictions, indicating that unbiased HLA peptidomics data are ideal for in silico predictions of neo-antigens from tumor exome sequencing data. The new motifs further reveal distant modulation of the binding specificity at P2 for some HLA-I alleles by residues in the HLA-I binding site but outside of the B-pocket and we unravel the underlying mechanisms by protein structure analysis, mutagenesis and in vitro binding assays.

Highlights

Human Leukocyte Antigen class I (HLA-I) molecules play a central role in defence mechanisms against pathogens and immune recognition of cancer cells
The vast majority (>95%) of HLA-I alleles still lack documented ligands and despite very valuable algorithmic developments to generalize prediction methods to any allele [12], it remains more challenging to make accurate predictions for alleles without known ligands. These alleles are only found in a small fraction of the Caucasian population, they are more frequently encountered in other ethnic groups and gaining more understanding of their binding specificity would be desirable for expanding the scope of therapeutic strategies relying on HLA-I ligand predictions
We measured the HLA peptidome eluted from six B cell lines, two in vitro expanded tumor-infiltrating lymphocytes (TILs) samples and two leukapheresis samples selected based on their high diversity of HLA-I alleles

Summary

Introduction

HLA-I molecules play a central role in defence mechanisms against pathogens and immune recognition of cancer cells. Information about binding motifs (mathematically defined here as Position Weight Matrices and graphically represented as sequence logos) of HLA-I molecules has been mainly obtained from biochemical assays where chemically synthesized peptides are tested in vitro for binding. The vast majority (>95%) of HLA-I alleles still lack documented ligands and despite very valuable algorithmic developments to generalize prediction methods to any allele [12], it remains more challenging to make accurate predictions for alleles without known ligands These alleles are only found in a small fraction of the Caucasian population, they are more frequently encountered in other ethnic groups and gaining more understanding of their binding specificity would be desirable for expanding the scope of therapeutic strategies relying on HLA-I ligand predictions. Such potentially biased datasets can be sub-optimal for training HLA-I ligand predictors

Methods

Results

Conclusion