- Top of page
- Materials and methods
Annual influenza epidemics represent a major cause of morbidity and mortality worldwide. Occasionally, as a result of antigenic shift, a new influenza virus subtype, which is not recognized by antibodies induced by previous infection or vaccination, appears in the human population. In the absence of specific immunity, such viruses can be transmitted rapidly to cause global pandemics.
Influenza vaccines have substantially reduced the burden of disease due to influenza infection, especially in vulnerable groups, such as the elderly and patients with chronic respiratory or cardiovascular disease.[3, 4] Although vaccination still represents the best way to prevent influenza, there is an urgent need for improvement. Current vaccination approaches aim to induce antibody responses against the variable viral surface antigens, mainly hemagglutinin (HA). Consequently, the overall success of seasonal vaccination depends mainly on the antigenic match between the vaccine and the circulating virus strain and may vary substantially from one season to the next.[5, 6] The antigenic composition of an emerging pandemic virus cannot be predicted at all, which makes it difficult to prepare sufficient vaccine stocks in due time. To restrict the impact of “between-season” strain variability and to attenuate the threat of a pandemic influenza outbreak, cross-protective influenza vaccines are desirable. Such vaccines should ideally target conserved viral antigens, such as the internal nucleoprotein (NP) or the matrix protein (M1).[7, 8]
Previously, we demonstrated that vaccination with whole inactivated virus (WIV), but not with subunit or split-virion vaccines, can protect mice from lethal heterosubtypic influenza challenge. This protection was due to the induction of a potent CD8+ T lymphocyte response against conserved virus proteins, such as NP.
In addition to the nature of the antigen and the presence of an adjuvant, the route of administration can strongly influence the immunogenicity of a vaccine.[10-12] For example, intranasally administered virus-like particles (VLPs) expressing influenza M2 protein induce superior antibody responses compared with the same vaccine administered subcutaneously.
Here, we investigated which route of WIV administration optimally induces heterosubtypic cross-protection against influenza. Specifically, we compared parenteral routes of administration (subcutaneous, SC; intramuscular, IM) with a mucosal vaccination route (intranasal, IN). After administration of H5N1 WIV, we determined the survival of mice after heterosubtypic challenge with H1N1 virus and measured the magnitude of induced flu-specific CD8+ T-cell responses. The main finding of the study is that full protection against lethal heterosubtypic challenge in mice was obtained only when WIV was delivered through one of the parenteral routes. The protection correlated with the presence of flu-specific CD8+ T cells. Only partial protection was observed in IN-vaccinated mice, which mounted very poor flu-specific CD8+ T-cell responses but developed cross-neutralizing IgA antibodies.
- Top of page
- Materials and methods
In a previous study, we demonstrated that immunization of mice with influenza WIV vaccine provides protection against heterosubtypic challenge, primarily through the induction of cross-reactive flu-specific CD8+ T-cell responses. Here, we show that parenteral administration, particularly IM injection, is superior to mucosal (IN) vaccine delivery for the induction of cross-reactive flu-specific CD8+ T-cell responses and cross-protection by WIV. However, mucosal immunization, as compared to IM or SC injection, induces a superior secretory IgA (SIgA) response, which may also contribute to heterosubtypic protection.
Our conclusion that parenteral vaccine administration is superior to mucosal delivery in terms of CD8+ T-cell priming is consistent with observations in several other systems. For example, Bessa et al. showed that virus-like particles containing a peptide derived from lymphocytic choriomeningitis virus induced a superior cellular immune response when the vaccine was administered through SC injection compared with mucosal administration. Furthermore, Decrausaz et al. observed that CD8+ T-cell responses were induced more effectively by parenteral rather than mucosal administration of a human papillomavirus vaccine. In addition, in early studies on the induction of flu-specific CD8+ T-cell activity by isolated NP, the antigen was administered IM or SC.
There are several potential explanations for the limited capacity of inactivated vaccines to induce CD8+ T-cell responses when administered mucosally. First, antigens delivered to mucosal surfaces are likely to be diluted in mucosal secretions and or quickly removed, limiting their availability for immune recognition.[18, 19] Importantly, low antigen availability is a limiting factor for cross-presentation and CD8+ T-cell induction. Also, antigen delivered through mucosal versus parenteral administration engages DCs with different cross-presenting capacities.[21, 22] Finally, mucosal immunization, as a consequence of the phenomenon of mucosal tolerance, may suppress the induction of systemic cellular responses.[23-25]
There appears to be a major difference between replicative and inactivated influenza vaccines in terms of their ability to induce CD8+ T-cell responses upon mucosal delivery to the respiratory tract. Several studies in mice, ferrets, and non-human primates have shown that mucosal administration, in particular pulmonary delivery, of infectious virus, resulting in a non-lethal infection of the (lower) respiratory tract, is very effective at inducing robust CD8+ T-cell responses and mediating cross-protection upon subsequent challenge with divergent virus variants or subtypes. Indeed, establishment of a pulmonary infection appears to be more effective at inducing flu-specific CD8+ T-cell responses than priming via the intraperitoneal, intravenous, or IN routes. In contrast, mucosal delivery of inactivated vaccines induces mainly SIgA antibodies, but not flu-specific CD8+ T-cell responses. Presumably, the barriers that impede the efficient development of T-cell responses upon mucosal delivery of inactivated vaccines do not affect replicative vaccines to the same extent. Productive infection results in the generation of comparatively high doses of viral antigen. In addition, active virus replication triggers DC activation, promotes the presentation of viral antigens in the context of both class I and class II MHC molecules, and may readily overcome mucosal tolerance. It is interesting in this respect that mucosal delivery of infectious influenza virus, as well as a recombinant adenovirus vector expressing influenza NP antigen, was more effective at inducing heterosubtypic cross-protection than IM injection.
Our finding that mucosal administration of influenza WIV vaccine is suboptimal for the induction of CD8+ T-cell immunity is at variance with findings from Alsharifi et al. They compared heterosubtypic protection induced by γ ray-inactivated WIV (γ-WIV) using different administration routes and found that IN administration was superior to SC delivery. In our hands, however, SC injection of WIV induced solid cross-protection that correlated closely with the magnitude of the NP-specific CD8+ T-cell response, while mucosal administration provided only partial protection that appeared to be mediated primarily by cross-reactive SIgA. It is difficult to explain this apparent discrepancy. It is unlikely that the dose of antigen was lower in our study, although a direct comparison of doses cannot be made. Alsharifi et al. measured antigen dose in pfu equivalents, whereas we used protein concentration. Nonetheless, a conservative estimate would suggest that the antigen dose was substantially higher in our study. It is possible that the protection observed in the study by Alsharifi et al. was mediated by cross-reactive antibodies, which were not investigated. Another important variable could be the use of different inactivation protocols for producing WIV; different inactivation protocols may yield vaccine formulations with varying capacities to activate cytosolic innate receptors and to induce cross-protective T-cell responses.[9, 35] Also, it is possible that the γ-WIV used by Alsharifi et al. was not entirely devoid of replication-competent virus, which, as discussed above, is very efficient in inducing CD8+ T-cell responses upon delivery to the respiratory tract. In this respect, it is interesting to note that infectivity of apparently completely inactivated γ-WIV may be reconstituted through genetic complementation. Indeed, upon multiple infection of a single cell, viral particles critically damaged at different parts of the genome may complement each other thereby reconstituting the capacity to produce infectious virus particles. A similar phenomenon has recently been described in a study by Brooke et al., showing that influenza virus often exists as a population of “abortive infectious forms” of virus that, through multiple infection, may reconstitute infectivity. A comparative study, involving head-to-head testing of similar doses of BPL-inactivated WIV or γ-WIV administered through different routes, would help to clarify the discrepancy between our findings and the findings by Alsharifi et al.
Our data are also at apparent variance with those of Bodewes et al. In the present study, IM injection of WIV induced robust cross-protective flu-specific CD8+ T-cell responses, whereas Bodewes et al. observed minimal flu-specific CD8+ T-cell induction and, as a result, no heterosubtypic cross-protection after IM vaccination of mice with formaldehyde-inactivated WIV. This apparent discrepancy may be explained by the use of divergent virus inactivation protocols and the use of different combinations of vaccine strain and challenge strain. As we showed previously,[9, 13] virus inactivation procedures using formaldehyde severely compromise the membrane fusion activity of WIV particles, which results in a significant decrease in the CD8+ T-cell-priming capacity of the vaccine. We prepared WIV using BPL as the inactivating agent, which preserves viral membrane fusion activity and, consequently, the CD8+ T-cell-priming ability of the vaccine to a considerable extent. Another variable that could explain the variance of our findings with findings of Bodewes et al. is the use of different combinations of vaccine strain and heterosubtypic challenge strain. Bodewes et al. used a reassortant H3N2 vaccine strain and a H5N1 A/IND challenge strain, while we used a reassortant H5N1 vaccine strain and a H1N1 A/PR8 challenge strain. The H5N1 reassortant contains internal virus proteins derived from A/PR8. While this optimizes internal viral antigen recognition by flu-specific CD8+ T cells, this model may not optimally reflect challenges that are faced in induction of cross-protection in humans.
Although mucosal administration of inactivated vaccine was suboptimal for the induction of flu-specific CD8+ T-cell activity in our experiments, we did observe partial protection from heterosubtypic challenge in mice immunized IN with WIV. The observed protection correlated with the presence of SIgA antibodies in mucosal secretions and cross-neutralizing serum antibodies which were found only after IN immunization, although it should be noted that the levels of antibodies in the vaginal washes might not fully reflect those of the respiratory organs. Sera of IM-immunized mice contained only IgG antibodies and did not show any in vitro neutralizing capacity (Figures 4A, B and 5). These results are in agreement with observations of others. Indeed, several studies have demonstrated the induction of full or partial protection against homosubtypic or heterosubtypic influenza infection by vaccination through a mucosal route and have also shown a close correlation between protection and the presence of mucosal SIgA antibodies.[31, 39-41] To establish whether a higher level of local protection (i.e., nasal cavity) is induced by IN immunization using WIV, a lower volume of challenge virus than used in the present study would be preferred. In this respect, aerosol inoculation of virus, for example, could mimicked natural influenza infection more closely.
In conclusion, the route of administration substantially influences the induction of cross-reactive flu-specific CD8+ T-cell responses and heterosubtypic cross-protection induced by influenza WIV in mice. Parenteral delivery of WIV, in particular IM vaccination, induces superior cross-reactive CD8+ T-cell responses and cross-protection compared with mucosal vaccine administration. On the other hand, antibody responses induced by mucosal (IN) vaccination with WIV, in particular SIgA, can contribute to heterosubtypic cross-protection in the absence of optimal flu-specific CD8+ T-cell immunity. Nonetheless, we conclude that parenteral vaccination is preferable for the induction of heterosubtypic cross-protection against influenza using WIV. Currently used vaccines either lack conserved target antigens for CD8+ T cells (e.g., subunit vaccines) or lack intrinsic adjuvant components such as viral RNA (e.g., subunit and split virus vaccines) that could help to boost (cellular) immunity through TLR7/8 activation. WIV, however, contains both conserved target antigens for CD8+ T cells and TLR-activating components and therefore holds promise as a candidate cross-protective influenza vaccine for use in humans. The findings from this study may further guide the development and implementation of such a cross-protective vaccine.