Exploring the HLA-A Associations in EBV-Associated Classical Hodgkin Lymphoma : HLA-A Genotype Modifies the HLA-A*02:01 Restricted Cytotoxic T-Cell and Cytokine Response to EBV

Class I HLA genes are associated with risk of developing EBV-associated classical Hodgkin Lymphoma (EBV-cHL). HLA-A*01:01 is associated with increased risk and HLA-A*02:01 with decreased risk, resulting in an almost 10-fold difference in odds of developing EBV-cHL between HLA-A*02:01 and HLA-A*01:01 homozygotes with a gradation of risk in intermediate genotypes. Class I HLA genes are involved in EBV-specific cytotoxic T cell (CTL) responses, keeping infection under tight control in healthy hosts. HLA associations with EBV-cHL suggest that EBV-specific immune responses at time of initial infection, during chronic latent infection or at time of oncogenesis are critical in the pathogenesis of EBV-cHL. Many immunodominant EBV peptides are known to be presented by HLA-A*02:01. Since HLA-A*01:01 appears to increase risk of disease even in those individuals with HLA-A*02:01, we investigated whether the presence of an HLA-A*01:01 allele modifies the CTL response to HLA-A*02:01-restricted EBV epitopes.

Individuals with relevant HLA-A genotypes were selected from a pool of healthy adults who agreed to take part in this study; A*01:01/A*02:01 heterozygotes (n=15), A*02:01 homozygotes (n=12) and A*02:01/× heterozygotes (n=15), where x is neither A*01:01 or A*02:01. Assessment of A*02:01 restricted EBV response used IFN-γ ELISPOT and all 31 reported A*02:01 restricted EBV peptides. PBMCs were exposed to single EBV peptides and controls in duplicate. The highest negative control plus two standard deviations was taken as the cut-off for a “positive” response. Measures of response were: the number of peptides recognised; maximal response to any single peptide; and summative response to EBV peptides. Analyses examined all peptides, lytic peptides, latent peptides and peptides expressed in type II latency. Assessment of cytotoxicity by degranulation was performed by flow cytometric analysis of CD107 expression after exposure to peptide pools. Supernatants from these stimulations were used to measure TNF-α, IFN-γ, IL-2, IL-4, IL-5, IL-6, IL-10, IL-12 and IL-17. Results were compared by Kruskall-Wallis analysis across groups and Mann-Whitney analysis between pairs of groups, using SPSS.

Satisfactory ELISPOTs were obtained from 31 donors. No difference in response to A*02:01 restricted EBV peptides was observed between A*01:01/A*02:01 heterozygotes and A*02:01 homozygotes. A*02:01/x heterozygotes recognised more EBV peptides than either A*01:01/A*02:01 heterozygotes (p=0.023) or A*02:01/A*02:01 homozygotes (p=0.049), an effect which was greatest for lytic peptides (p= 0.012). The single maximal response to any EBV peptide was higher in A*02:01/x heterozygotes, reaching significance for lytic peptide responses (p=0.042). The summative response was also higher in A*02:01/x heterozygotes, and for lytic peptides these differences were significant (p=0.022).

A*02:01/x heterozygotes also demonstrated greater cytotoxicity assessed by CD107 than either A*01:01/A*02:01 heterozygotes or A*02:01 homozygotes. This effect was seen in response to latent and lytic peptides, with the differences greatest between A*02:01/x heterozygotes and A*02:01/A*01:01 heterozygotes.

A*02:01/A*01:01 heterozygotes produced significantly higher levels of IL-10 (mean 1427 vs 155 pg/ml p= 0.03), IL-17 (84 vs 70 pg/ml p=0.02) and IL-5 (35 vs 29 pg/ml p= 0.04) in response to lytic peptides compared to both other HLA-A groups. IFN-g secretion was higher in A*02:01/x heterozygotes (2552 vs 552 pg/ml, p=0.007).

The null hypothesis of this study, that all HLA-A*02:01 carriers should have the same response to HLA-A*02:01 restricted EBV peptides was disproved. A*02:01/x heterozygotes had CTL responses of greater breadth and magnitude than either A*01:01/A*02:01 heterozygotes or A*02:01 homozygotes, as measured by three different methods. A*01:01 did not inhibit CTL responses to A*02:01-restricted EBV peptides; however, in response to EBV lytic peptides, A*02:01/*01:01 heterozygotes demonstrated significant increases in cytokines associated with a Th2 response. These data suggest that overall HLA-A phenotype is important in determining CTL response through a single HLA allele. The cytokine profiles observed in the three groups following stimulation with EBV peptides may begin to explain the HLA-associated differences in risk of developing EBV-cHL.

No relevant conflicts of interest to declare.