chemically synthesised influenza peptides
World Health Organization
Influenza causes yearly epidemics of mild disease and, occasionally, pandemics with millions of fatalities. Currently, no vaccine is effective against all influenza strains. Analysis of influenza sequences from animal and human isolates using CLUSTALW and a novel proprietary epitope prediction algorithm identified six conserved T cell-reactive regions in several proteins. Immunisation of transgenic mice with a preparation of these six regions as chemically synthesised peptides (FLU-v) induced a specific HLA-A*0201-mediated CD8+ T cell response. This T cell population also reacted against human cells infected with three non-related influenza strains, confirming that the identified regions contain epitopes naturally presented by infected human cells and conserved amongst non-related viruses. Moreover, FLU-v immunisation significantly increased survival of transgenic mice against lethal challenge with influenza. Overall, FLU-v represents a promising influenza vaccine candidate, obviating the need for yearly vaccinations and allowing the stockpiling and initiation of a worldwide vaccination program in advance of a pandemic outbreak.
Influenza viruses cause yearly epidemics associated with a significant economic and social burden arising from treatment, work-hours lost 1 and, in some groups (e.g. elderly populations), increased morbidity and mortality 2. Influenza also causes pandemics characterised by high infection and mortality rates. The last two influenza pandemics (in 1957 and 1968) had a combined death toll of more than 6 million individuals, whereas the Spanish Flu pandemic of 1918 killed more than 40 million people worldwide. After recent outbreaks of highly pathogenic influenza viruses in the Far East, the World Health Organization (WHO) believes we are now closer to an influenza pandemic than at any time since 1968 (WHO 2005. Available at www.who.int/csr/disease/avian_influenza/phase/en/print.html).
For over 50 years, healthcare programs have relied on vaccines that provide antibody-mediated prophylactic immunity to Influenza. These antibodies target small sections of the hemagglutinin (HA) and neuraminidase (NA) antigens found on the virus capsid. However, the high variability of these antigens, which constitute the basis for influenza strain classification, means that current vaccines targeting specific influenza strains fail to provide universal protection against all other strains. As new variants of the virus emerge year to year, the vaccine preparations must be updated and re-administered annually. To compound this problem, vaccine mass production cannot start until the vaccine strains for the year have been identified and released by WHO, and the production process itself is heavily reliant on the availability of fertilised hen eggs 3. As a result, there is a delay of 6 months or more between strain identification and enough doses of the vaccine being available.
A new approach to influenza vaccination is necessary to allow effective vaccination in advance of a pandemic outbreak. Vaccines that induce an immune response via T lymphocytes rather than via B lymphocytes and antibodies are potential candidates. CD8+ CTL play a central role in the resolution of influenza infections 4–6 and, in the absence of antibodies, provide significant protection against re-infection 7.
T lymphocytes target influenza epitopes that are bound to HLA molecules and presented on the surface of the body's own APC. Vaccination with synthetic peptides containing T cell epitopes can induce a strong T cell response that provides protection against challenge with influenza 8–11. However, since many of these peptide vaccines contain a single T cell epitope and this is specific for a single HLA molecule, they are not perceived as viable vaccines against highly variable viruses in a heterogeneous HLA human population.
In this report, we describe a new candidate synthetic influenza vaccine containing multiple conserved T cell epitopes (polyepitope vaccine) naturally presented by human cells infected with non-related influenza strains and which elicits a protective T cell response against the virus.
Identification of highly-conserved T cell-reactive influenza protein fragments
In order to identify conserved T cell reactive regions in influenza A and B, all (i.e. animal and human) influenza A and B protein sequences stored at the National Center for Biotechnology Information Taxonomy database were aligned using the CLUSTALW program. Influenza A and B sequences for each protein were analysed together unless the resulting consensus was divergent in more than 40% of its sequence, in which case, A and B sequences were analysed independently. Sequence regions in which every aa was present in at least 70% of the influenza isolate population were considered to be conserved.
The identified consensus sequences were analysed for the presence of reactive T cell regions (i.e. containing epitopes) for multiple mouse MHC (including H-2Kb) and human HLA (including HLA-A*0201) alleles using a proprietary algorithm (PepTcell Ltd).
The analysis revealed that only M1, M2, NP and PB1 contained conserved T cell reactive regions (Table 1), although only in the case of M2 and PB1 were these sequences shared by A and B isolates.
|Name||Consensus sequence||Amino acid locationa)||Polymer length (amino acids)|
|M1Ab)||DLEALMEWLKTRPILSPLTKGILGFVFTLTVP||36 to 67||32|
|M1Bc)||LLYCLMVMYLNPGNYSMQVKLGTLCALCEKQASHS||124 to 158||35|
|NPAb)||DLIFLARSALILRGSVAHKSC||255 to 275||21|
|NPBc)||PGIADIEDLTLLARSMVVVRP||306 to 326||21|
|PB1d)||LLIDGTASLSPGMMMGMFNMLSTVLGVSILNLGQ||395 to 428||34|
|M2d)||IIGILHLILWILDRLFFKCIYRLF||32 to 55||24|
Induction of an IFN-γ-mediated immune response
Having identified conserved T cell epitopes in several influenza proteins, we sought to assess whether a preparation of synthetic peptides encompassing those epitopes (FLU-v) could induce an HLA-specific Th1-like immune response (i.e. an IFN-γ-mediated response). For this purpose, production of IFN-γ was measured in splenocyte cultures from transgenic mice immunised with either the FLU-v or a non-relevant peptide (NRP-v) preparation cocultured with mytomicin C (MMC)-treated HLA-A*0201-bearing (T1) or non-bearing (JURKAT) human cell lines intracellularly loaded with each of the polyepitope peptides.
Comparison of the cytokine response in FLU-v and NRP-v immunised animals, revealed that splenocytes from those receiving FLU-v produced a significantly (p <0.05) increased level of IFN-γ when cocultured with HLA-A*0201-bearing human T1 cells internally loaded with the M1A (Fig. 1A), M1B (Fig. 1B), NPA (Fig. 1C), NPB (Fig. 1D), PB1 (Fig. 1E) or M2 (Fig. 1F) peptides. No significant differences in IFN-γ production were observed when splenocytes were cocultured with non-HLA-A*0201-bearing human JURKAT cells loaded with the same peptides (Fig. 1). As the transgenic mice used in these experiments do not bear any other human HLA and there is no evidence that their CD8+ T cells would specifically recognise any FLU-v-derived epitopes in the context of other human HLA that they have never encountered 12, these results clearly show that the observed IFN-γ response is specifically caused by T cells recognising the HLA-A*0201 epitopes present in the M1A, M1B, NPA, NPB, PB1 and M2 peptides.
The level of IFN-γ production was also significantly increased in FLU-v- vs. NRP-v-vaccinated animals when soluble peptide antigen was added to the splenocyte cultures (Fig. 1). However, for M1A (Fig. 1A) and M2 (Fig. 1F), the level of this response was lower than that observed when the antigen was presented via the HLA-A*0201-bearing T1 cells. A possible explanation for this is that M1A and M2 contain fewer murine epitopes. As these HLA-A*0201 transgenics also express the complete mouse set of H-2b class I and class II MHC molecules and soluble antigen is processed by APC into both MHC class I and II pathways 13, only a small pool of H-2b specific CD8+ and CD4+ T cells may be stimulated whilst, at the same time, less antigen is being presented to HLA-A*0201 specific CD8+ T cells. This in turn could lead to lower levels of IFN-γ being produced.
No IL-4 response was detected against any of the polyepitope peptides (data not shown). As IL-4 production is antagonistic to IFN-γ production and hence to the creation of antigen-specific CD8+ T cell responses, the lack of IL-4 production again indicates that FLU-v immunisation induces a specific Th1-like immune response to each of the polyepitope peptides.
Reactivity to three non-related influenza strains
The peptides of the FLU-v preparation encompass conserved T cell reactive regions within the Influenza virus population. Therefore, CD8+ T cells from FLU-v-vaccinated mice should be capable of recognising epitopes naturally processed and presented by human cells infected with serologically unrelated strains of influenza. As expected, splenocytes from FLU-v-immunised mice produced significantly higher levels of IFN-γ than those from NRP-v-vaccinated mice when cocultured with HLA-A*0201 bearing (T1) human cells infected with either A/Puerto_Rico/8/34 (H1N1), A/NYMC/X-147 (H3N2) or B/Johannesburg/5/99 (Fig. 2). No differences were observed when the splenocytes of FLU-v- and NRP-v-vaccinated mice were cocultured with HLA-A*0201 non-bearing (JURKAT) cells infected in the same way. No IL-4 response was detected in any of the vaccinated mice (data not shown).
These results indicate that FLU-v vaccination induces a T-cell response that is effective against current unrelated strains, and hence most likely, against future strains that may arise from antigenic drift and/or genetic re-assortment between animal and human strains.
Murine MHC-mediated immune response to non-related influenza strains
Prior to testing the protective potential of the FLU-v preparation in mice, we wished to assess the contribution of the murine MHC-specific T cells to the response against Influenza following vaccination. For this purpose, splenocyte cultures from FLU-v- and NRP-v-immunised mice were cocultured with H-2b bearing (EL-4) and non-bearing (P815) mouse cells infected with the same influenza strains described above. Splenocytes from FLU-v-vaccinated animals produced a significantly (p <0.05) higher level of IFN-γ compared to those of NRP-v-vaccinated animals when cocultured with MMC-treated, influenza-infected syngeneic mouse cells (EL4), but not when cocultured with allogeneic cells (P815) treated in the same way (Fig. 3). No IL-4 response was detected in any of the vaccinated mice (data not shown).
These results clearly show that the conserved regions targeted by the FLU-v peptides contain not only human, but also murine T-cell epitopes.
Increased survival of transgenic mice to lethal influenza challenge
To assess whether the FLU-v-induced immune response and, in particular, its CD8+ T cell component, could improve survival, a challenge study was performed in normal and CD8-depleted transgenic mice using the Influenza A/Puerto_Rico/8/34 strain.
Most animals immunised with either FLU-v or NRP-v, but subject to CD8 depletion, succumbed to influenza infection by day 6 after intranasal challenge (Fig. 4A). In contrast, in the absence of CD8 depletion, there was a significant (p <0.05) improvement in survival rate in animals immunised with FLU-v versus NRP-v (Fig. 4B).
These results clearly indicate that a small dose of this vaccine preparation induces a CD8+-mediated response, which can significantly increase survival against a lethal dose of influenza.
The current primary tool against influenza is a vaccine preparation that induces antibody-mediated immunity. This vaccine, however, does not target the T cell arm of the immune response despite evidence that CTL play a significant protective role during infection 4–7 and the perceived advantage of such approach 14. The current licensed influenza vaccine confers protection by inducing a prophylactic antibody response against the virus. There is evidence the vaccine also induces a specific T cell response 15; however, several studies have shown that this response is significantly impaired in the elderly 16 and does not increase significantly the number of influenza-specific IFN-γ-producing T cells in the general adult population 17. Moreover, most T cell epitopes in influenza are found in its internal proteins, and the nature of the current vaccine preparation (i.e. a split whole virus containing mainly HA and NA) results in the loss of most of these internal proteins from the final product. As a result, the antibody and T cell response induced by the vaccine are primarily directed against HA and NA, which are highly variable and, as it is widely acknowledged, do not provide heterotypic protection.
Based on evidence that CTL play a significant protective role during influenza infection 4–7, we sought to identify conserved influenza protein sequences targeted by T cells that could be synthesised chemically. These synthetic polyepitope peptides could then constitute the basis for a protective multi-strain vaccine that overcomes the inadequacies of the current vaccine.
For the identification of conserved protein regions across most, if not all, influenza strains, we had to account for the virus’ high level of antigenic heterogeneity, which is attributed to two distinct evolutionary processes: antigenic drift and genetic re-assortment 18. The first involves continuous small random mutations in the sequence of viral proteins. The second involves the exchange of genetic material between two strains during infection, thus giving rise to a novel strain. Both these events have clear practical consequences. Antigenic drift affects all influenza proteins 19, but primarily HA and NA, and is responsible for the yearly need to update and administer the influenza vaccine composition. Genetic re-assortment, due to the susceptibility of some species (e.g. waterfowl and swine) to infection by both human and animal influenza virus, is responsible for both the incorporation of novel sequences in the pool of human infectious strains and for the origin of pandemic strains 19–21.
By definition, antigenic drift is a random process and it is not possible to forecast the exact HA or NA sequences of a future influenza strain, pandemic or not. However, over time, point mutations tend to accumulate in regions of the genome where they may or may not be biologically advantageous, but where they are certainly not disadvantageous, as that would result in the extinction of that virus line. Following this principle, it has been possible to define heterogeneic areas not only in variable viruses such as HIV 22, but also in highly stable viruses such as human papillomavirus 23. Multiple sequence alignment of available influenza protein sequences is an appropriate method for the identification of such variable regions.
In contrast to previous studies where only human influenza isolates have been analysed (for examples, see 24, 25), we chose to incorporate sequences from isolates of both human and animal origin, as it is widely expected that the next pandemic influenza strain will evolve from an animal, most likely avian, strain. Further justification for this approach is provided by published data showing that (i) animal influenza sequences have been incorporated into circulating human influenza viruses by genetic re-assortment 26, 27 and (ii) that swine act as a long-term reservoir and genetic ‘melting pot’ for avian, human and swine viruses 18.
Our analysis revealed several conserved sequences across influenza proteins, and identification of T cell epitopes for multiple mouse MHC and human HLA alleles within those sequences was carried out using a proprietary algorithm. Over the years, multiple algorithms have been developed to identify individual T cell epitopes within a protein (for review, see 28). Early algorithms that relied on the identification of biochemical ‘hot-spots’ along a protein sequence were shown to be wholly unreliable 29. More recent algorithms rely on calculating the binding strength of a sequence to a MHC/HLA molecule. However, experimental evidence shows that neither all high-affinity peptides are immunogenic 30 nor all immunogenic peptides are of high affinity 31. Hence, there is a poor correlation between predicted and experimental results 32.
The proprietary algorithm here used identifies and categorises T cell epitopes within a given protein based on the analysis of the structural affinity of a peptide for a given MHC/HLA allele and the reactivity of this complex to T cells. Many of the epitopes identified by the algorithm, but not included in the FLU-v preparation, had already been identified experimentally (a list is available at www.flu.lanl.gov/review/epitopes.html). However, their efficacy as multi-strain candidate vaccine products is limited as they fall within regions of increased sequence variability as defined by our multiple sequence analysis. Six conserved regions of less than 40 aa were identified in M1, M2, NP and PB1 that contain a high density (>5) of human T cell epitopes.
Most of the T cell epitopes found in the FLU-v preparation were not ranked as immunodominant by the algorithm. Typically, the bulk of the CD8+ T cell response is directed to a limited number of immunodominant epitopes 33, and it is probably due to the low frequency of CD8+ effectors to most FLU-v epitopes in infected humans and rodents that they have not been identified experimentally before. However, immunodominance is not a pre-requisite for vaccine efficacy, as several subdominant epitopes have been shown to serve as good vaccine targets for the control of respiratory viral infections 34, 35. Our results showing (i) the specific reactivity of each conserved sequence in the context of human and murine MHC class I molecules, (ii) the ability of T cells induced by FLU-v vaccination to specifically recognise epitopes naturally processed and presented by influenza-infected human and murine cells and (iii) the ability of a low-dose preparation of these peptide sequences to increase survival in a lethal challenge study, clearly support the assertion that a mixture of subdominant epitopes can act together, perhaps synergistically, to induce T cell-based cross-strain protection against influenza.
Our experimental results cannot, and are not intended to, rule out the possibility that antibody and/or CD4+ T cell responses against all or some of the peptides in the FLU-v preparation may play a role in conferring protection against challenge by influenza. However, our results certainly corroborate the fact that CD8+ T cells play a significant role in conferring protection against influenza infection.
It is generally accepted that protective immunisation of the entire human population via peptide vaccination is difficult to achieve. However, this argument is based on two specific issues 36, 37, neither of which apply to the FLU-v peptide preparation. First, peptide vaccines tend to target individual epitopes, which, by definition, are specific for individual HLA alleles. As a result, any immunity conferred by immunisation with that peptide will only arise in individuals bearing the appropriate HLA allele for that peptide. Although for technical reasons we have only analysed T cell reactivity experimentally for a single human HLA allele, the described polyepitope sequences in M1, M2, NP and PB1 were selected on the basis of containing a high density of T cell epitopes for several human HLA alleles. Moreover, despite having used incomplete Freund's adjuvant in this study, we have obtained similar results (unpublished observations; G. A. Stoloff and W. Caparros-Wanderley) using adjuvant combinations that are more appropriate to human use. As a result, and pending further experimental work, it would be expected that the vaccine preparation would provide protective immunity in the majority of individuals. Secondly, peptide vaccines have tended to target epitopes with no consideration for the variability or frequency of that particular epitope in the pathogen population. As a result, any immunity conferred by these epitopes could be easily overcome by highly variable viruses (e.g. influenza, HIV) or lead to protection against only a limited number of strains of the pathogen. However, it is unlikely that protein fragments that are not subject to antigenic drift in either animal or human isolates of influenza will start to become variable as a result of immunological pressure following mass vaccination of humans. We have not applied our algorithm to the identification of T cell epitopes in any species other than mice and humans. Therefore, we make no claim regarding the T cell reactivity of these fragments on other species. Indeed, it is possible that the identified fragments may not contain a single T cell epitope for any other animal species and, if that is the case, there will be no immunological pressure on the sequence to mutate independently of whether they are used in a mass vaccination programs.
It is for these reasons that we believe a synthetic peptide preparation (FLU-v) containing the identified sequences constitutes a viable and promising candidate Influenza vaccine preparation, which would (i) preclude the need for repeated yearly vaccinations, (ii) allow the stockpiling of vaccine prior to the identification of any pandemic strain and, more importantly, (iii) allow the initiation of a worldwide vaccination program well in advance of the pandemic outbreak.
Materials and methods
Identification of highly conserved T cell reactive regions in influenza proteins
In order to identify conserved T cell-reactive regions in influenza A and B, all influenza A and B protein sequences stored at the National Center for Biotechnology Information Taxonomy database (www.ncbi.nlm.nih.gov/Taxonomy/taxonomyhome.html) were downloaded and sorted according to their genomic origin, i.e. gene segments coding for hemagglutinin (HA), neuraminidase (NA), matrix proteins (M1, M2), nucleoprotein (NP), critical components of the viral polymerase (PB1, PB2, PA) and non-structural proteins (NS1, NS2). Conserved regions within these protein sequences were identified by means of multiple sequence alignment using CLUSTALW 38.
The identified consensus sequences were analysed for the presence of reactive T cell regions (i.e. containing epitopes) for multiple mouse MHC (including H-2Kb) and human HLA (including HLA-A*0201) alleles using a proprietary algorithm (PepTcell Ltd). The algorithm identifies and categorises T cell epitopes within a protein based on analysis of the structural affinity of a peptide for a given MHC/HLA allele and the reactivity of this complex to T cells.
The final selection of conserved polyepitope T-cell reactive fragments in the influenza isolate population was based on three criteria: (i) length (no more than 40 aa), (ii) presence of at least five human T cell epitopes, and (3) a probability of <10–10 for any isolate not to contain at least one of the identified T cell epitopes.
Transgenic homozygous C57BL/6-TgN(HLA-A2.1)1Enge mice were obtained from the Jackson Laboratory. These mice carry the human HLA-A*0201 transgene and express significant quantities of the human class I MHC antigen HLA-A2.1 on cells from the spleen, bone marrow and thymus. These transgenic mice have been used to identify HLA-A2.1 epitopes in HCV that are recognised by human A2.1-restricted CTL 39. Transgene carrier status and expression in these mice were monitored and confirmed every 6 months by quantitative PCR and reverse-transcription PCR, respectively.
Peptides and antibodies
All polyepitope peptides (Table 1) and an NRP control were synthesised by Fmoc chemistry and resuspended in 10% DMSO in PBS. Purified rat anti-mouse CD8 IgG2a (clone YTS169.4) was obtained from AbD Serotec (UK).
Cell lines and viruses
The T1 and JURKAT cell lines are human lymphoblastoid lines derived from HLA-A*0201-bearing and non-bearing individuals, respectively. T1 cells were maintained in IMDM (Invitrogen). JURKAT cells were maintained in RPMI-1640 medium (Sigma) containing 10 mM HEPES and 1 mM sodium pyruvate. Both media were supplemented with 50 IU/50 μg/mL of penicillin/streptomycin (Sigma) and, as complete medium, 10% FCS (Sigma). All cell lines were obtained from the ATTC collection.
The EL4 (H-2b) and P815 (H-2d) cell lines are chemically induced tumour lines derived from C57BL/6 and DBA/2 mice, respectively. Both cell lines were maintained in DMEM (Sigma) supplemented with 50 IU/50 μg/mL of penicillin/streptomycin and, as complete medium, 10% FCS.
Primary splenocyte cultures were maintained in IMDM medium supplemented with 0.02 mM β-mercaptoethanol (Sigma), 50 IU/50 μg/mL of penicillin/streptomycin and 10% FCS.
Three influenza strains were used for infection: influenza A strains Puerto_Rico/8/34 and NYMC/X-147, and influenza B strain Johannesburg/5/99. All three strains were obtained from the influenza WHO repository based at the National Institute for Biological Standards and Control (NIBSC; UK). The two A strains are serologically distinct (H1N1 and H3N2) and, together with the B strain, constitute the equivalent of the standard trivalent vaccine preparation prescribed for human use.
Preparation of target cells for cytokine analysis
Cell cultures in exponential phase were harvested by centrifugation (200 × g, 5 min) and transfected with the polyepitope peptide antigens (5 μg per 106 cells) using Lipofectin (Invitrogen) according to the manufacturer's instructions. After 8–10-h incubation in complete medium, cells were treated with MMC (Sigma). Alternatively, cell cultures were harvested by centrifugation and infected with live influenza virus (multiplicity of infection of 5–10) for 1 h, washed twice in serum-free medium, and incubated in complete medium for 18 h before MMC treatment.
For MMC treatment, cells were harvested by centrifugation (200 × g, 5 min) and resuspended in serum-free medium containing 50 μg/mL MMC. After 45 min incubation at 37°C, cells were washed four times in serum-free medium and resuspended in complete IMDM medium.
Immunisations and challenge studies
On day 1, 7–10-week-old C57BL/6-Tg(HLA-A2.1)1Enge/J mice were each immunised subcutaneously with 200 µL of antigen preparation. In the test group (FLU-v), each dose of the antigen preparation contained 60 nmol of an equimolar mixture of all six polyepitope peptides (10 nmol each) prepared in incomplete Freund's adjuvant (Sigma) according to the manufacturer's instructions. In the control group (NRP-v), each dose of the antigen preparation contained an equivalent dose of NRP, also prepared in incomplete Freund's adjuvant. On day 15 post-immunisation, all animals received a booster immunisation using the same doses and route of delivery.
For immunogenicity studies, animals (n = 8 for both groups) were culled on days 22 or 23, and their spleens collected. For challenge studies, on day 16, animals in the test and control groups (n = 14 for both) were split into two equal subgroups (n = 7 each; i.e. Control-1, Control-2, Test-1 and Test-2). On day 19, all animals in the Control-1 and Test-1 groups received a 200-μL intraperitoneal injection of rat anti-mouse-CD8 sera (100 μg), whereas all animals in the Control-2 and Test-2 groups received an equivalent injection of unrelated rat sera. On day 20, all groups were challenged intranasally under anaesthesia with a lethal dose (approximately 1.5 × 107 pfu) of influenza A/Puerto_Rico/8/34. On day 22, animals were again intraperitoneally injected with either rat anti-mouse-CD8 or unrelated sera as described above. From day 20, all animals were monitored daily for morbidity and mortality. All animals still alive on day 27 were culled and the study was terminated. All experimental work was carried out in accordance with the Scientific Procedures Act (1986) and under project licenses PPL 60/3418 and PPL 60/3573.
Mouse spleens were gently pressed through cell strainers and red blood cells were removed with red cell lysis buffer (nine parts 0.16 M NH4Cl and one part of 0.17 M Tris, pH 7.2). Splenocyte suspensions from each experimental group were plated in 24-well plates at a density of 4 × 106 cells/well containing either polyepitope peptides (5 μg/mL; antigen-specific positive control) or MMC-treated human or mouse cell lines (splenocyte to cell, S:C ratio 10:1) that had been either transfected with polyepitope peptides or infected with live influenza.
After 4 days of incubation, the supernatant was collected and analysed for IFN-γ and IL-4 production using a mouse cytokine ELISA kit (PharMingen) to assess the induction of an immune response. The threshold of detection for the assays was 39.06 pg/mL for IFN-γ and 9.77 pg/mL for IL-4. A positive response was defined as an increment of at least 30% over the background response with a statistical significance of p <0.05.
Statistically significant differences between FLU-v and NRP-vaccinated animals in the IFN-γ response to different antigens and in the survival against lethal challenge with influenza were established by non-parametric Mann–Whitney analysis and by Fisher's exact test analysis, respectively. Differences were considered statistically significant if the p value was below 0.05.
We are grateful to Ms. D.E. Adams and Mr. D. Burt for excellent animal husbandry and support and to Professor H. van der Heyde and Dr. I. Gramaglia for their critical review of the manuscript.