Major Histocompatibility Complex Binding, Eluted Ligands, and Immunogenicity: Benchmark Testing and Predictions

Abstract

Antidrug antibody (ADA) responses impact drug safety, potency, and efficacy. It is generally assumed that ADA responses are associated with human leukocyte antigen (HLA) class II-restricted CD4+ T-cell reactivity. Although this review does not address ADA responses per se, the analysis presented here is relevant to the topic, because measuring or predicting CD4+ T-cell reactivity is a common strategy to address ADA and immunogenicity concerns. Because human CD4+ T-cell reactivity relies on the recognition of peptides bound to HLA class II, prediction, or measurement of the capacity of different peptides to bind or be natural ligands of HLA class II is used as a predictor of CD4+ T-cell reactivity and ADA development. Thus, three different interconnected variables are commonly utilized in predicting T-cell reactivity: major histocompatibility complex (MHC) binding, capacity to be generated as natural HLA ligands, and T-cell immunogenicity. To provide the scientific community with guidance in the relative merit of different approaches, it is necessary to clearly define what outcomes are being considered. Thus, the accuracy of HLA binding predictions varies as a function of what the outcome predicted is, whether it is binding itself, natural processing, or T-cell immunogenicity. Furthermore, it is necessary that the accuracy of prediction is based on rigorous benchmarking, grounded by fair, objective, transparent, and experimental criteria. In this review, we provide our perspective on how different variables and methodologies predict each of the various outcomes and point out knowledge gaps and areas to be addressed by further experimental work.

Comparison of HLA ligand elution data and binding predictions reveals varying prediction performance for the multiple motifs recognized by HLA-DQ2.5

“…These studies suggested that 1,000 nM is a generally applicable threshold for biologically relevant binding in the context of HLA class II. This threshold was also confirmed in a recent analysis which found that 83.3% of HLA class II restricted epitopes are predicted to bind their corresponding restricting allele with an affinity of 1,000 nM, or better (19). Of note, optimal thresholds for MHC class II are still unclear and are being discussed in the community.…”

“…Peptides with predicted binding affinities of >1,000nM were considered predicted nonbinders. This threshold was selected based on previous studies analyzing the binding affinity of HLA class II restricted T cell epitopes (16)(17)(18)(19).…”

Comparison of HLA ligand elution data and binding predictions reveals varying prediction performance for the multiple motifs recognized by HLA‐DQ2.5

“…Peptides with predicted percentile ranks of >20% were considered predicted non‐binders. This threshold was selected based on previous studies analysing the binding affinity of HLA class II‐restricted T‐cell epitopes 20‐23 …”

Comprehensive analysis of T cell immunodominance and immunoprevalence of SARS-CoV-2 epitopes in COVID-19 cases

“…Overall, the 49 most dominant epitopes showed significantly higher binding promiscuity (number of alleles bound at the 1,000 nM or better threshold) (Paul et al, 2019;Southwood et al, 1998) for the panel of common HLA class II than a control group of 49 non-epitopes derived from the same proteins (Average number of HLA predicted to be bind ± SD epitopes = 10.8 ± 6.5; non-epitopes = 5.7 ± 6; p=0.0001 by Mann-Whitney; Fig. S3C-D).…”