Our study examined the pattern of cellular responses in influenza vaccinated individuals using three different T cell assays, ELISPOT to quantitate the numbers of specific IFNγ producing cells and 3H thymidine and CFSE assays to quantitate the number of proliferating T cells, in relation to B cell responses examined using microneutralization assays. We studied the responses to two influenza A subtype viruses-H1N1 A/New Caledonia/20/99 and H3N2 A/Wisconsin/67/05 which contain the HA and NA proteins in the 2006–2007 inactivated trivalent influenza vaccine. We found that the kinetics of human influenza immune responses was variable and that preexisting immunity to this virus affected the ability of vaccination to induce significant increases in responses. A most interesting finding in this study was the discordance between cellular and humoral responses in influenza vaccinated individuals.
Given that the majority of participants worked in health care settings and received frequent influenza vaccinations, proliferation in prevaccination PBMC to live influenza viruses in thymidine assays was not unexpected. A more detailed analysis using CFSE flow cytometric assays detected small percentages (~5%) of proliferating CD4 CD45RA− and CD8 CD45RA− T cells consistent with previous exposure through natural infection and/or vaccination (). When examining peak responses elicited over the post vaccination period, significant increases were seen only in the CD8 CD45 RA+ subset. Traditionally associated with a naive phenotype, CD45RA+ cells have been shown in CMV and EBV systems to include cells capable of effector functions [28
]. Inactivated vaccines are not thought to primarily induce CD8 T cell responses given that the antigen processing pathway required by CD8 T cells is bypassed. The expansion of the CD8 CD45RA+ population may represent a reexpression of CD45RA+ on previously CD45RA− cells in an assay expanded for 6 days rather a true expansion of naive CD45RA+ cells. However, the inactivated influenza vaccine may have activated memory CD4 T cells and through their cytokine producing functions and ability to co- stimulate CD8 T cells, these CD4 cells were able to stimulate the proliferation of CD8 CD45RA+ T cells. Our data expand on the results from a recent, open, randomized, comparative clinical trial comparing an adjuvanted and non-adjuvanted subunit influenza vaccine, in which CFSE assays detected a significant increase in the frequency of the proliferating CD3 CD4 T cell population ( CD3/CD8 T cells were not analyzed) [32
]. Similarly, in a study of eight individuals who received the influenza vaccine, Long et al. found no increases in the frequencies of NK or CD8 T cells as a proportion of lymphocyte subset but did find, using intracellular cytokine staining, that there were increases in IFNγ producing CD8 T cells detected post vaccination after stimulation with virus [33
] Our results would seem to similarly suggest that the overall population of CD8 T cells is stable post vaccination but that there is a subset of cells (CD45RA+), that is altered after vaccination. Ideally, studies to verify these increases in CD8 T cell frequency using more quantitative assays such as tetramer staining with CD8 T cell epitopes or through the use of intracellular cytokine assays in our larger cohort would be helpful but availability of PBMC samples was limited.
Similar to the study of He et al [34
], we found statistically significant inverse correlations between the numbers of prevaccination IFNγ cells in ELISPOT assays to the individual H1 and H3 viruses and the magnitude of responses post-vaccination. In high responders post vaccination, IFNγ producing T cells may be responding to the HA and the NA proteins contained in the inactivated vaccine. Recent reports have found a considerable amount of nucleoprotein and M1 protein in the inactivated influenza vaccine; these may be additional targets for CD8 T cell responses [35
]. The presence of high baseline neutralizing antibodies may contribute to the lack of responses in individuals with high baseline numbers of IFNγ producing cells. However, we found no significant correlations between levels of prevaccination neutralizing antibodies and fold increases in the number of IFNγ producing cells. This suggests that the lack of an increase in the number of influenza – specific IFNγ-producing cells in the PBMC of these individuals with the highest prevaccination frequencies may be the presence of CD8 and CD4 influenza specific T cells which might rapidly eliminate antigen presenting cells containing vaccine antigens, limiting subsequent activation of T cells upon reexposure.
Vaccines that enhance T cell immunity in addition to robust antibody responses are likely to be the most protective. Understanding the relationships between influenza specific T cells and antibody responses elicited by vaccination will enable a more rational design of future vaccines to influenza virus. A previous study by McElhaney et al. in older adults found positive correlations between antibody levels and Th1/Th2 cytokine ratios in older adults who received a single dose of vaccine; a negative correlation was seen between these parameters in those who received a booster dose [38
]. A relationship between CD4 T cell and antibody responses would have been expected as this subset of T cells is involved in helping antibody producing B cells as well as helping CD8 T cells to proliferate. In our study, we found no correlations between neutralizing antibody titers and the number of IFNγ producing cells in terms of either prevaccination levels or fold increases. This discordance may be related to the high prevaccination immune responses present in these individuals. Though influenza virus causes an acute self limited infection, exposure to this virus occurs repeatedly, either through vaccination or through natural infection. Because of this, it was not unexpected that increases in T cell or B cell immune responses detected through ELISPOT assays or microneutralization assays were modest, findings which could account for some lack of correlation seen in our analysis. The time range of the blood samples taken may account for these results with the second sample ( 13–39 days) which falls within the likely peak of the CD8 T cell response having a more significant effect on the results than the last timepoint sampled. Studies have shown that age impacts T cell responses to influenza vaccination, with elderly individuals (>65 yrs old) showing a decline in type 1 T cell responses which may be a factor in these results; however, the average age of our subjects was 44 years old with no individuals over the age of 60 [39
]. The role of innate immunity components such as cytokines, dendritic cells, and toll like receptors in initiating influenza specific responses may also contribute to the discordance between these two arms of the immune system.
Current vaccine approaches depend on the induction of antibodies to the viral surface proteins hemagglutinin and neuraminidase that neutralize the infectivity of the virus and interfere with the release of newly replicated virus from the host cell. Given the fact that these surface proteins undergo frequent mutations, vaccines are reformulated each year to reflect the predominant circulating viruses of the previous influenza season. Current influenza vaccines are unable to elicit cross subtype reactive antibodies to new subtypes of virus such as the avian H5N1 virus which is of concern. Trivalent inactivated influenza vaccines should theoretically induce T cell responses at least to the HA and NA proteins present which contribute to overall immunity to influenza virus. However, our data and others suggest that given the high level of preexisting immunologic memory to influenza virus present in the population, current influenza vaccines are only able to modestly increase either T or B cell immunity to the virus [34
]. Induction of cellular immune responses to the more conserved internal proteins as opposed to the more variable surface proteins of the influenza virus contained in current formulations of the inactivated vaccine would be a more attractive candidate for long lasting immunity and potentially cross reactive responses to influenza virus. More detailed investigations on the T cell responses to these internal conserved proteins of the influenza virus and their functional qualities will be helpful in this regard.