SUMMARY
The Immune Epitope Database and Analysis Resource (IEDB) was recently developed to capture epitope related data, and is publicly available at www.immuneepitope.org. An analysis resource linked to the database hosts various bioinformatics tools which can be used to identify novel epitopes as well analyze and visualize existing epitope data. Herein we report the results of a comprehensive analysis of antibody and T cell epitopes of influenza A virus. The first objective of the study was to compile and analyze all current information regarding antibody and T cell epitopes for influenza A virus. The second objective was to investigate possible cross-reactivity among avian (specifically H5N1) and human flu virus antibody and T cell epitopes.
As of May 22nd, 2006, the PubMed database contained approximately 2,000 references related to influenza antibody and T cell epitopes. After reviewing the abstracts and/or full-text of these references, 743 of them were found to contain relevant information and were curated into the database. As a result, more than 3,000 assay measurements related to approximately 190 antibody and 412 T cell distinct epitope molecular structures were captured. Despite the importance of antibody responses in protection from flu infection, T cell epitopes dominated the literature, compared to antibody epitopes. Approximately half of the reported antibody epitopes were discontinuous sequences (e.g., conformational) or defined as key residues identified by viral antibody escape mutant studies.
Influenza A epitopes were reported for 13 different subtypes and 58 different strains. The majority was from the human influenza H1N1 and H3N2 subtypes, and there were only two epitopes reported for the avian H5N1 subtype, highlighting the need for studies investigating the epitopic structure of avian flu strains. Antibody epitopes were identified from only five proteins, and mostly from the virus surface proteins HA, NA and M2. In contrast, T cell epitopes were identified from all 10 viral proteins. The HA and NP proteins contained the highest numbers of epitopes, with most CD4+ epitopes being derived from HA and most CD8+ epitopes from NP. While antibody epitopes were described mostly in mouse and rabbit, T cell epitopes were studied in mouse and also humans.
To predict possible cross-reactivity and protection from the avian H5N1 strains in previously vaccinated or infected hosts, we evaluated the conservancy of the reported epitopes utilizing the conservancy tool provided in the analysis resources of IEDB. The degree of conservation for each antibody and T cell epitope was calculated for 17 representative influenza strains of H1N1, H3N2 and H5N1 subtypes including A/Hong Kong/156/97 and A/VietNam/1194/2004. Overall, approximately 15% of the epitopes were conserved in various human flu strains, and 10% were also conserved in the avian flu strains considered. Only a small number of protective epitopes were characterized in challenge and virus neutralization assays, and they were mostly tested in mice. Protective T cell epitopes were highly conserved between human and avian influenza strains. For protective antibody epitopes, only those from M2 protein showed appreciable conservation, and may confer protection from avian H5N1 i nfection.
In summary, a comprehensive antibody and T cell epitope analysis of influenza A virus was undertaken using information collected from the IEDB. Although a great amount of knowledge is available, results from the current analysis identified five areas of knowledge gaps including: 1) Determination of protective antibody and T cell epitopes (only a few were reported in the literature), 2) Paucity of antibody epitopes in comparison to T cell epitopes, 3) Limited number of animal hosts from which the epitopes were studied, 4) Limited number of epitopes reported for avian influenza strains/subtypes, and 5) besides HA and NA proteins, there was relatively fewer epitopes reported for the other 8 proteins. As a step in preparing for possible pandemic influenza outbreaks, the knowledge gaps identified here would be a useful guide for future research directions in influenza A virus immune epitope identification studies.
For complete description of the study, see "Ab and T cell epitopes of influenza A virus, knowledge and opportunities." Following are the results of the study.
FIGURES
Figure 1. Process for selection and curation of relevant influenza A epitope literature references.
Figure 2 . PubMed query utilized to identify potentially relevant influenza epitope references The query consisted of four main parts. The first part contained epitope keywords constructed to capture references that are related to epitopes, such as “epitope”, “major histocompatibility complex”, “mimotope”, “antibody”, and “HLA”. The second part consisted of keywords specifically related to Influenza such as “influenza”, “influenza a virus”, and “influenza virus”. The third part was a filter consisting of keywords to limit the references by presence of an abstract and use of English language. Finally, “NOT” keywords were used to exclude abstracts related to reviews, editorials, meta-analyses, and comments
TABLES
Table 1. Total number of published influenza A epitopes by protein
Table 2. Total number of published influenza A epitopes by host species
Table 3. Proposed research agenda towards a more systematic and comprehensive collection of influenza immune epitopes
Table 4 . Distribution of epitope molecular structures by influenza strains
Table 5 . Collection of influenza strains for cross-reactivity analysis
Table 6 . Conservancy analysis of Ab linear epitope sequences
Table 7 . Conservancy analysis of T cell linear epitope sequences
Table 8 . Distribution of conserved linear epitopes at different identity levels
Table 9 . Conservancy analysis of Ab conformational epitope sequences
Table 10 . Conservancy analysis of protective Ab/T cell epitopes