Newly emerging and re-emerging infectious diseases pose a continuous threat to the health of our society. Studies of host immune responses against these microbes provide insights on both the pathogenic mechanisms of the organisms and new approaches in vaccine development. Identification of immunodominant T cell epitopes within the infectious organism is one of the key initial steps that are essential for understanding host cellular immune responses. Since CD4+ T cells play a central role in the regulation of the adaptive immune response against viruses, it is particularly important to identify HLA-class II restricted epitopes for studying cellular immunity to influenza A. In this study, we applied TGEM technology to identify H5HA specific antigenic epitopes prior to a potential H5N1 epidemic-pandemic shift.
An inherent caveat in the TGEM approach is that identified T cell responding epitopes by tetramer staining may not be naturally processed. We were pleased to see that at least 4 out of 6 epitopes we identified are naturally processed from whole protein by both in vitro and in vivo methods. It should be mentioned that, while our results indicate that these 4 H5HA peptides correspond to naturally processed epitopes, the precise length of the naturally processed protein fragment could differ from those of the synthetic peptides. Although TGEM is a synthetic peptide based assay system, the HLA class II restricted epitopes identified by TGEM were highly correlated to naturally processed epitopes. This may reflect that epitopes identified by TGEM technique have high affinity to ‘empty’ HLA class II molecules, because tetramer preparation relies on the affinity of peptide loading onto ‘empty’ HLA class II monomer. Peptides with higher affinity compete better to bind in HLA class II and therefore have a better chance to be presented on cell surface in vivo. These aspects of the tetramer methodology lead to a second caveat in the TGEM approach. Because the assay relies on high affinity peptide binding, the assay may fail to detect low affinity peptides that are legitimate epitopes. Our cumulative results over the past several years indicate that the TGEM method detects epitopes with binding affinities as low as 10–20 μM. Therefore, it is possible that there are additional low affinity epitopes (below 10–20 μM) or epitopes with low naïve frequencies within H5HA. These epitopes could play an important role in natural immunity.
Influenza A virus has 3 surface proteins (HA, NA and M2) and 8 internal proteins (NP, PA, PB1, PB2, M1, NS1, NS2 and PB1-F2) [15
]. The internal proteins and MP2 are relatively conserved between viral subtypes. Therefore, cellular immunity against epitopes derived from these proteins are often cross reactive because the epitopes are actually identical or only slightly different among different Influenza A viral strains [35
]. Since the surface glycoprotein HA is subject to antigenic shift and drift due to host immunological defense pressures, novel HA T cell epitopes are constantly emerging from most new influenza A strains. Identification of H5N1 unique, specific, immunodominant CD4 T cell epitopes is of particular interest for 1) tracking viral strain specific T cell immunity, 2) evaluation of strain specific vaccine efficacy, and 3) development of peptide-based strain specific vaccines. In addition, some studies have suggested that antibody-producing B cells may favor T-cell help towards T cells recognizing the same protein antigen toward which the antibody is produced to bind to[37
]. Therefore, CD4+ T cell immunity against H5HA may be best suited to provide help to H5HA specific B cells in antibody production.
In this study, we identified two immunodominant epitopes, H5HA57-76
(). The H5HA57-76
epitope is a novel strain specific epitope that has very little homology to other human tropic influenza strains (). T cells specific to the H5HA57-76
epitope were only found in the naïve (CD45RA+) T cell population (). However despite its uniqueness among influenza strains, H5HA57-76
epitope is a highly conserved among H5HA clades or subclades. Actually, it is completely identical in all candidate H5N1 vaccine reference viruses recommended by the World Health Organization [39
]. Therefore, it could be a very useful T cell epitope to evaluate specific H5N1 vaccine efficacy. In addition, it could also be an important epitope as a peptide vaccine candidate. First, our in vitro experiment demonstrated that H5HA57-76
epitope could be efficiently processed and presented by antigen-presenting cells (). Second, when HLA-DR0401 mice were immunized with H5N1 subvirion vaccine or H5HA recombinant protein, T cell reactivity towards H5HA57-76
epitope was most frequent and strongest (). T cell immunity against H5HA57-76
epitope was detected in 3 out of 3 H5N1 subvirion immunized mice. Third, epitope mapping data showed H5HA57-76
epitope was only identified in 1 out of 5 subjects () indicating that there is a lack of T cell immunity towards H5HA57-76
epitope in the general HLA-DR0401+ population. Collectively, we conclude that H5HA57-76
epitope could be a necessary and efficient epitope as a peptide based vaccine candidate.
In contrast to H5HA57-76
, sequence homology analysis revealed that the H5HA441-460
epitope is highly conserved in HA proteins among H1N1 and H2N2 viral strains, two human tropic influenza subtypes (data not shown)[20
]. Therefore, T cell responses to H5HA441-460
epitope could result from a cross () or a recall response from annual immunization of Influenza vaccine comprised of the H1N1 subtypes. Indeed, when CD4+ T were sorted into CD45RA+ (naïve phenotype) and CD45RA− (memory phenotype) fractions, H5HA441-460
response T cells were clearly detected in CD45RA− T cell pool (). Furthermore, T cell reactivity towards H5HA441-460
epitope showed a consistently potent response throughout the epitope mapping experiments in all the subjects used in this study (). Again, like H5HA57-76
was confirmed as a naturally processed immunodominant epitope in immunized mice in vivo. Collectively, these data strongly argued that H5HA441-460
is a cross-reactive epitope for H1N1, H2N2 and H5N1 strains. The strong H5HA441-460
epitope specific reactivity is likely a record of vaccination or exposure outcome of current flu-vaccines.
Besides H5HA57-76 and H5HA441-460, the remaining 3 naturally processed epitopes: H5HA17-36, H5HA121-140 and H5HA377-396, H5HA121-140 and H5HA377-396 showed only moderate immunogenicity in the immunized transgenic mice (). Therefore, although these epitopes could be naturally processed by antigen-presenting cells, they may not be immunodominant epitopes.
In summary, our study indicated that 1) novel HLA class II restricted epitopes can emerge from a novel influenza antigen, i.e., H5HA antigen, 2) novel epitopes could be effectively identified by our TGEM technology, 3) T cells specific for these novel epitopes exist in antigen unexposed individuals in the naïve cell population, 4)T cell responses to these novel epitopes may be suboptimal in the unexposed population, 5) knowledge of epitopes could be used to further determine the immunodominant epitopes using a convenient in vitro protein stimulation and tetramer staining process, 6)immunodominant epitopes defined from this in vitro assay could fully mirror in vivo study results. We successfully identified two immunodominant H5HA epitopes, one is unique for H5N1 substrains but conserved within H5N1 clades and subclades, the other is a cross-reactive epitope for multiple subtypes. The H5HA57-76 immunodominant epitopes should be a valuable marker for tracking H5N1 specific immunity possible in the development of peptide-based vaccines.