Clinical studies show that both live, attenuated and inactivated, split influenza vaccines are safe and effective, although there are instances when greater protection against disease has been observed in young children vaccinated with live, rather than inactivated vaccines [17
]. This is likely due to differences in the immune mechanisms that contribute to protection: inactivated vaccines induce robust antibody responses [19
] that prevent infection or reduce spread of the virus, while live virus vaccines induce cellular responses [20
] that probably contribute to protection through secretion of anti-viral cytokines or direct killing of infected cells. Unfortunately the live, attenuated vaccine is administered intranasally, and is therefore contraindicated for individuals with asthma or other respiratory diseases, excluding many children from the possible benefit of cell-mediated immunity. In this report we show differences in the quality and quantity of immune responses to live and inactivated influenza virus administered IM, and demonstrate that as for live virus administered intranasally, IM vaccination with live virus induces robust antibody as well as CD8+ T cell responses, thereby providing protection against challenge with a heterosubtypic virus.
Comparison of cytokine and antibody profiles in sera of mice immunized IM with live and UV-treated virus preparations showed quantitative and qualitative differences, with greater amounts of acute inflammatory cytokines correlating with an increased antibody response following exposure to live virus preparation. Since we used a UV-inactivated virus preparation that retained some live virus, it is possible that even greater differences would have been observed in the complete absence of live virus.
The cytokines measured in this study are induced as a result of signaling through TLR7, RIG-I, and inflammasomes [8
]. IL-12 and IFN-γ are usually TLR7-driven, and since virus replication is not required for the interaction of this sensor with its ligand, single-stranded RNA, one would expect both cytokines to be induced following exposure to live and inactivated virus. However, significantly greater amounts of IL-12 and IFN- γ were present 1 day after IM vaccination with live virus than UV-treated virus, even though the latter treatment did not completely inactivate the virus. This may reflect differences in uptake of virions into cells or synergism with other responses that are replication-dependent, for example, an effect of IFN-α that is induced following immunization with the live virus preparation only. This latter idea is in accord with work that demonstrates the ligand for RIG-I is dsRNA [21
], and therefore this signal is replication-dependent. IFN-α may contribute to the enhanced immune response in direct as well as indirect means. For example, others have demonstrated synergism between IFN-α and IL-12 [22
], and IFN-α is known to increase immunogenicity by acting on B cells to induce early antibody responses [23
]. It is therefore not surprising that the magnitude of the antibody response following IM immunization with live virus was greater than UV-inactivated virus.
Both live and UV-inactivated virus preparations induced Th1-type responses after the first vaccination, resulting in expected IgG2a HA-specific antibodies. Unexpectedly, IgG1 (IL-4-dependent) and some IgG2b (IL-5 driven) HA-specific antibodies that are typically associated with Th2-type responses were amplified after a second dose of live but not inactivated vaccine. The live virus preparation did indeed result in increased concentrations of IL-4 and IL-5 in serum that could explain these responses.
Our results show that the most robust serum antibody response - measured as either total HA-specific IgG or HAI titers - is generated after IM immunization with live virus. Intranasal inoculation with live virus resulted in similar HAI titers after one dose, but unlike IM vaccination with live virus, HAI titers were not boosted after a second IN exposure to live virus. Priming at the mucosa and in the periphery certainly results in distinct responses that are often important for establishing protection at the appropriate site of infection [26
]; IN immunization provides the advantage of inducing local IgA and memory T cell responses that contribute to protection against infection of the upper respiratory tract, whereas IgG in the circulation provides effective protection against influenza infection in the lower respiratory tract of mice [27
]. While IM vaccination with live virus probably does not induce local antigen-specific IgA, robust serum HAI titers would contribute to vaccine-induced protection of the lung.
The proposal to administer a live vaccine parenterally is not a new idea - it has long been established that the live virus vaccine for measles, mumps and rubella (MMR) administered intradermally, is an effective pediatric vaccine. This vaccine has excellent immunogenicity when a single dose is delivered at 12 months of age, although a second dose is recommended, to provide a boost to the small percent of recipients (5%) that do not respond to one or more of the antigens after the first dose. Vaccine efficacy is robust and has eliminated endemic measles transmission in the United States [28
]. Increasing influenza vaccine immunogenicity by changing the form that is delivered intramuscularly may provide an important increase in effectiveness in the pediatric population.
Unfortunately, maternal antibodies inhibit responses to live virus MMR vaccine delivered parenterally [29
]. This is one reason the MMR vaccine is delivered at 1 year of age. Even though we demonstrate boosting of the response to a second dose of live virus delivered IM, careful studies need to be completed in order to establish whether the immune response to influenza can be primed in the presence of influenza-specific antibodies, whether of maternal origin or due to prior vaccination or infection. This will determine whether parenteral immunization with live virus is likely to be immunogenic in seropositive children and adults.
In summary, our results show that live virus is more immunogenic than inactivated virus when delivered intramuscularly. The increased antibody response corresponds with induction of greater amounts of inflammatory cytokines early after primary immunization. This form of antigen allows for activation of antigen-specific cytolytic CD8+ T cells that are the primary means of clearing influenza A viruses that do not have the same HA and NA subtype. Indeed, IM vaccination with live virus of H3N2 subtype induced large numbers of NP-specific CD8+ T cells and offered significantly more protection against heterosubtypic H1N1 virus challenge. Asthmatics or other individuals with respiratory ailments are excluded from receiving live, attenuated influenza vaccines, therefore, intramuscular delivery of this vaccine type may be a useful strategy to increase immunogenicity and efficacy in population groups that are most at risk of disease.