A population response analysis approach to assign class II HLA-epitope restrictions

Abstract

Identification of the specific HLA locus and allele presenting an epitope for recognition by specific TCRs (HLA restriction) is necessary to fully characterize the immune response to Ags. Experimental determination of HLA restriction is complex and technically challenging. As an alternative, the restricting HLA locus and allele can be inferred by genetic association, using response data in an HLA-typed population. However, simple odds ratio (OR) calculations can be problematic when dealing with large numbers of subjects and Ags, and because the same epitope can be presented by multiple alleles (epitope promiscuity). In this study, we develop a tool, denominated Restrictor Analysis Tool for Epitopes, to extract inferred restriction from HLA class II-typed epitope responses. This automated method infers HLA class II restriction from large datasets of T cell responses in HLA class II-typed subjects by calculating ORs and relative frequencies from simple data tables. The program is validated by: 1) analyzing data of previously determined HLA restrictions; 2) experimentally determining in selected individuals new HLA restrictions using HLA-transfected cell lines; and 3) predicting HLA restriction of particular peptides and showing that corresponding HLA class II tetramers efficiently bind to epitope-specific T cells. We further design a specific iterative algorithm to account for promiscuous recognition by calculation of OR values for combinations of different HLA molecules while incorporating predicted HLA binding affinity. The Restrictor Analysis Tool for Epitopes program streamlines the prediction of HLA class II restriction across multiple T cell epitopes and HLA types.

Figure 1. Novel HLA-Epitope Restrictions Predicted by RATE

PBMCs were incubated with single HLA-transfected cells pulsed with (A) Rv1705c epitope FFGQNTAAIAATEAQ, (B) Rv1195 epitope SSYAATEVANAAAGQ, (C) Rv0288 epitope IMYNYPAMLGHAGDM, or (D) Rv3874 epitopes AAVVRFQEAANKQKQ and (E) AQAAVVRFQEAANKQ for 24 hours. IFN-γ release was measured by ELISPOT. The white bars show significant responses (p-values < 0.05), while the black bars represent non-significant responses. HLA alleles expressed by donor are presented in the table insert. HLA-transfected cell lines not available are labeled N/A. Predicted binding from IEDB is listed below each allele tested. RFs and Fisher's exact test p-values (when significant) are listed for predicted restrictions.

Figure 2. Novel HLA-Epitope Restrictions Validated by Tetramer Binding

A) Purified CD4⁺CD3⁺CD8⁻ cells stained with YYSNVTARTLLSSTNS tetramer-PE. B) CD3⁺CD4⁺CD8⁻ PBMCs stained with YYSNVTARTLLSSTNS tetramer-PE after 14 days of stimulation with peptide. C) CD3⁺CD4⁻CD8⁺ PBMCs stained with YYSNVTARTLLSSTNS tetramer-PE after 14 days of stimulation with peptide. D) Top: CD3⁺CD8⁻CD19⁻CD11b⁻CD56⁻ PBMCs stained with the indicated tetramer-PE combinations after anti-PE magnetic bead enrichment. Bottom: Flow-through from magnetic bead enrichment. Numbers shown are percentages of tetramer⁺ CD4⁺ or CD8⁺ cells.

Figure 3. Predicted Promiscuous Binding Epitope Validated by Tetramer Binding

A) Purified CD4⁺CD3⁺CD8⁻ cells stained with VKAQNITNKRAALIEA tetramer-PE. B) CD3⁺CD4⁺CD8⁻ PBMCs stained with VKAQNITNKRAALIEA tetramer-PE after 14 days of stimulation with peptide. C) CD3⁺CD4⁻CD8⁺ PBMCs stained with VKAQNITNKRAALIEA tetramer-PE after 14 days of stimulation with peptide. Numbers shown are percentages of tetramer⁺ CD4⁺ or CD8⁺ cells.