Immunodominant responses to the influenza virus M1_58–66 epitope: Stealth or protection?

Keskin et al. elegantly used mass spectrometry to detect viral peptides on the surface of influenza A virus (IAV)-infected epithelial cells (1). The authors found that the IAV cytotoxic T lymphocyte (CTL) epitope M1_58–66 was abundantly present on virus-infected cells, which corresponds with the immunodominant nature of this epitope. The authors argue that this might represent another “stealth” mechanism of influenza viruses, because a T-cell line specific for this epitope displayed low functional avidity. Furthermore, Keskin et al. conclude that the conservation of this particular epitope is not the result of functional constraints but merely represents an immune-evasion strategy. We disagree on several points.

First, in the paper of Berkhoff et al. cited by Keskin et al. (reference 31 in ref. 1), none of the alanine replacements were tolerated in genetically modified IAVs without loss of viral fitness. Alanine replacements at each of the positions of the 9-mer affected the kinetics of virus replication, resulting in reduced virus titers with at least one 10log, showing that these amino acid substitutions caused a reduction of progeny virus of >90%. Of interest, the M1_58–66 sequence largely overlaps with a functionally important nuclear export signal of the M1 protein (2).

Second, the M1_58–66-specific T-cell line that was used by Keskin et al. (1) was obtained after stimulation with peptide. This may have selected for M1_58–66-specific T cells of low functional avidity. This is a well-known phenomenon and was published in PNAS almost two decades ago (3). In contrast, Boon et al., tested the functional avidity of polyclonal CTL populations specific for nine IAV epitopes, including M1_58–66, obtained from 16 partially HLA class I-matched individuals after stimulation with live virus (4). In the Boon et al. study it was shown that M1_58–66-specific CTLs had the highest functional avidity of all conserved epitopes tested. Furthermore, the immunodominance of the epitopes directly correlated with the functional avidity of the T cells directed against them, which also contrasts with the message of the Keskin et al. paper (1). In addition, Keskin et al. fail to demonstrate that CTLs directed to epitopes other than M1_58–66 eluted from infected cells have a higher functional avidity, or are more abundantly expressed on the IAV-infected cell surface in the absence of the M1_58–66 epitope.

Interestingly, the overall IAV-specific CTL response is smaller in HLA-A*02:01⁻ subjects, as was demonstrated in groups of blood donors with matched HLA class I-alleles (5). Because the size of the virus-specific memory CD8⁺ T-cell pool is smaller in HLA-A*02:01⁻ subjects, it is likely that these individuals are less-well protected from infection with IAV. It was recently demonstrated that a higher frequency of preexisting IAV-specific CTLs correlated with protection against disease severity after infection with the 2009 pandemic virus. It would be of considerable interest to also take into account the epitope-specificity or MHC class I restriction of CTLs in this type of prospective study to definitely demonstrate that M1_58–66-specific CTLs also contribute to protective immunity, rather than being an immunodominant, “nonfunctional” stealth mechanism.