Novel Bivalent Viral-Vectored Vaccines Induce Potent Humoral and Cellular Immune Responses Conferring Protection against Stringent Influenza A Virus Challenge

Seasonal influenza A virus (IAV) infections cause significant morbidity and mortality worldwide and remain a major public health concern. The novel avian-origin influenza A strain (H7N9), initially identified in 2013, is now circulating with almost annual frequency and almost one-third of all H7N9 cases occurred in the 2016/2017 influenza season. Most worringly, the case-fatality rate for this virus exceeds 40% (1–3). In addition, H7N9 influenza viruses have recently been assessed as having the highest potential pandemic risk of any novel IAV; this assessment is based on recent studies indicating that H7N9 viruses have increased genetic diversity and wider geographical distribution; additionally, in recent outbreaks, a significantly higher proportion of H7N9-infected patients have needed care in an intensive care unit (1–3). For the past 70 y, vaccination has been the mainstay healthcare strategy against influenza infection (4–6). However, traditional inactivated influenza vaccines (IIVs) confer strain-specific protection and do not typically induce the broad-spectrum immunity needed in the face of a newly emergent IAV (7–9). Therefore, the possible threat of a pandemic outbreak has catalyzed the development of broadly protective IAV vaccines.

Recent strategies to augment and broaden vaccine efficacy have shifted toward the development of “universal” vaccines capable of providing heterosubtypic protection against multiple, or possibly all, subtypes of IAV. Although humoral and cellular immunity can mediate heterosubtypic responses, inducing Abs against the more conserved stalk domain of hemagglutinin (HA) has been the recent focus of many vaccine programs (10, 11). However, multipronged approaches that target several Ags that induce humoral and cellular responses may limit the generation of viral escape mutants compared with vaccines that target a limited number of protective epitopes on the HA stalk. There are few vaccine technologies that will facilitate the delivery of multiple Ags to generate robust cellular and humoral immunity toward infectious disease Ags.

Viral-vectored vaccines have been developed for the induction of strong humoral and potent cellular immunity toward encoded Ags. An added benefit of this platform is that viral vectors can accommodate more than one Ag (12). Typically, for heterologous prime-boost vaccination strategies, one viral vector (e.g., Chimpanzee adenovirus [ChAd]) encoding the target Ag(s) is used for the priming vaccination, and a different platform, most often modified vaccinia Ankara (MVA), is used for the boost or repeat vaccination. In the current study, we describe novel ChAd- and MVA-vectored vaccines that are designed to simultaneously induce heterosubtypic and protective B and T cell responses against three influenza A Ags: HA, NP, and M1. Using a heterologous prime-boost strategy, we induce high levels of heterosubtypic and homologous immune responses targeting the major virion surface protein HA and the conserved internal viral Ags NP and M1. We demonstrate protection, after prime-boost vaccination, in a stringent challenge model of mouse-adapted avian IAV.

The construction of ChAdOx1 NP+M1 has been described previously (13). Details of the viral-vectored vaccines used in these studies are as described in Table I.

Procedures were performed according to the Animals (Scientific Procedures) Act 1986 (U.K.) and were approved by the University of Oxford Animal Care and Ethical Review Committee. Six- to eight-week-old female BALB/c (H-2^d) mice were obtained from Harlan Laboratories and were housed under specific pathogen–free conditions. All vaccines were formulated in endotoxin-free PBS and were administered intramuscularly (i.m.) in a total volume of 50 μl. BALB/c mice were immunized i.m. with 10 μg of HA7 protein, MVA (1 × 10⁶ PFU MVA-GFP or MVA-NP+M1 or MVA-NP+M1-H5 or MVA- NP+M1-H7), or ChAdOx1 (2.2 × 10⁶ infectious units of ChAdOx1 NP+M1 and/or 1 × 10⁸ infectious units of ChAdOx1-H7 or 1 × 10⁸ infectious units of ChAdOx1-GFP). For prime-boost regimens, mice were vaccinated with ChAdOx1 viral-vectored vaccines (at doses indicated), and all mice were boosted with 1 × 10⁶ PFU MVA 8 wk later.

Spleen ELISPOT was performed to measure Ag-specific IFN-γ, as described previously (14).The immunodominant H2-K^d–restricted (BALB/c) epitope NP_147–158 (TYQRTRALV) was used to measure responses following vaccination regimens with NP+M1 (15). H7HA responses were measured after stimulation with peptide pools. Two peptide pools were generated. The first pool contained peptides unique to A/Netherlands/219/2003, and the second pool contained peptides unique to A/Anhui/1/2013 and/or A/Shanghai/1/2013.

Briefly, peptides that were conserved between A/Netherlands/219/2003 and A/Anhui/1/2013 were pooled (NR-44011; Biodefense and Emerging Infections Research Resources Repository [BEI Resources], National Institute of Allergy and Infectious Diseases, National Institutes of Health). The resultant pool was representative of H7HA from A/Netherlands/219/2003 (H7N7). All peptides that differed between A/Netherlands/219/2003 (H7N7) and divergent strains (A/Anhui/1/2013 [H7N9] and A/Shanghai/1/2013) were pooled to generate a second peptide pool. This H7HA peptide pool is representative of regions of amino acid sequence diversity in the HA of A/Anhui/1/2013 (H7N9) and A/Shanghai/1/2013 (H7N9) and was generated from BEI Resources NR-44011 and NR-44012. H7 peptide pools were obtained through BEI Resources: Peptide Array, Influenza Virus A/Shanghai/1/2013 (H7N9) Hemagglutinin Protein Diverse Peptides, NR-44012. Medium alone was used as a negative control, and pools of overlapping peptides (H7HA) and NP_147–158 (TYQRTRALV) were typically added at 2 μg/ml.

ELISA was performed essentially as described (14). Nunc MaxiSorp 96-well plates were coated with 0.1 μg of recombinant protein [H7HA protein was produced in-house, as described (16)] and recombinant H5HA protein (A/Vietnam/1203/2004 [H5N1], recombinant from baculovirus, BEI Resources NR-10510) per well, and plates were washed until the fifth dilution of the reference standard (1:1600 dilution) reached an approximate OD₄₅₀ value of 1. This point was defined as one relative ELISA Unit, and relative ELISA units of test sera were calculated, essentially as described (14, 17).

A bone marrow IgG Ab-secreting cell (ASC) ELISPOT assay was performed as described (18, 19) using ∼1 × 10⁷ cells per milliliter in complete Iscove’s media that had been rested overnight. MultiScreen-IP Filter Plates were coated with 0.5 μg of recombinant HA, and negative control wells were coated with irrelevant protein (0.5 μg of OVA).

Starting with an initial 1:40 dilution, test sera were diluted 2-fold in complete DMEM and assayed as described (14). Results were normalized relative to cell-only wells and pseudotyped lentivirus-only wells and expressed as the percentage of inhibition of pseudotyped lentivirus entry (neutralization). IC₅₀ was calculated using GraphPad Prism 6 software.

All animal protocols were reviewed and approved by the Mount Sinai Institutional Animal Care and Use Committee. To assess the protective efficacy of the prime-boost vaccination regimen, 6–8-wk-old female BALB/c (H-2^d) mice (The Jackson Laboratory) were primed with ChAdOx1 NP+M1 (1.1 × 10⁷ infectious units, group 1), ChAdOx1-H7 HA (1 × 10⁸ infectious units, group 2), or ChAdOx1-GFP (1 × 10⁸ infectious units, group 4) (n = 10 mice per group). All viral vectors were administered i.m. in the musculus tibialis in a final volume of 50 μl. One group of animals received ChAdOx1 NP+M1 and ChAdOx1-H7 HA, and each virus was injected into a separate limb (n = 5 mice per group; group 3). Animals vaccinated with 10 μg of recombinant H7 from A/Anhui/1/13 (H7N9) supplemented with 5 μg of R848 (InvivoGen) served as a positive control (n = 10, group 5). Naive animals remained unvaccinated (n = 10, group 6). Eight weeks following the prime, groups 1, 2, and 3 were boosted i.m. with MVA-NP+M1 (1 × 10⁶ PFU). Group 4 received MVA-GFP as a boost, and group 5 was administered 10 μl of recombinant HA with R848. Blood ELISPOTs were performed at 2 wk after boost to ensure successful vaccine uptake. Three weeks following boost vaccination, all animals (n = 55) were anesthetized and challenged with five murine LD₅₀ of a 6:2 reassortment ratio of A/Shanghai/1/13 (H7N9) virus. Weight was monitored daily for 14 d; mice that lost ≥25% of their initial body weight were euthanized.

Statistical analyses were carried out using GraphPad Prism software version 6 (GraphPad, La Jolla, CA). Data were tested for normal distribution, and the appropriate statistical analysis was applied.

BALB/c mice were immunized i.m. with 1 × 10⁶ PFU MVA-H5, a vector expressing a single Ag in the group 1 HA, H5HA (A/Vietnam/1203/2004; H5N1), or MVA-NP+M1-H5, a bivalent vaccine expressing the same H5HA Ag, in addition to the T cell fusion Ag, NP+M1.

These new-generation bivalent constructs express NP+M1 using the early vaccinia promoter F11, whereas HA expression is driven by the P7.5 promoter. T cell immunogenicity was assessed 2 wk after vaccination by ex vivo IFN-γ ELISPOT against the immunodominant BALB/c epitope, NP_147–158 (TYQRTRALV) (Fig. 1). T cell responses to this epitope in mice vaccinated with the bivalent vaccine MVA-NP+M1-H5 were higher (median spot-forming units [SFU] = 206, p = 0.008) compared to the response in mice vaccinated with MVA-NP+M1 (P7.5) (median SFU = 60) (Fig. 1A). It has previously been shown that immunogenicity toward Ags expressed under the F11 MVA promoter is greater than the response toward P7.5-expressed Ags (20).

Two weeks after vaccination, total serum IgG responses were measured, by ELISA, against recombinant H5HA protein (A/Vietnam/1203/2004; BEI Resources) (Fig. 1B). No significant differences were observed between mice receiving MVA-H5 or MVA-NP+M1-H5 (Fig. 1B).

Because the immunogenicity of HA can vary greatly depending on subtype, humoral responses induced by vaccination with MVA-NP+M1-H7, a second bivalent construct expressing the group 2 HA, (A/Netherlands/219/2003; H7N7), in addition to NP+M1, was also investigated. BALB/c mice (n = 5–10) were immunized i.m. against H7HA and/or NP+M1.

As before, T cell responses following vaccination with the bivalent vaccine MVA-NP+M1-H7, in which the expression of NP+M1 is driven by the F11 promoter, were significantly higher compared with vaccination with MVA-NP+M1, in which expression is driven under the P7.5 promoter (**p < 0.01). These results indicate that the novel bivalent vaccine MVA-NP+M1-H7 elicits potent T cell responses against the NP+M1 fusion protein while also expressing a second Ag from the same viral vector (Fig. 2A).

Serum was collected at 2 and 8 wk after vaccination, and total IgG responses were measured against recombinant H7HA protein A/Netherlands/219/2003 (Fig. 2B). As a comparator, a group of mice was vaccinated with 10 μg of recombinant H7HA protein. Mice immunized with MVA-NP+M1-H7 had the highest median H7HA-specific IgG Abs at 2 wk after vaccination (Fig. 2B). Importantly, IgG Ab titers were maintained and remained high out to 8 wk after vaccination. Animals vaccinated with MVA-NP+M1-H7 also had the highest median responses at 8 wk after vaccination (Fig. 2B).

BALB/c mice were immunized with ChAdOx1 NP+M1 (2.2 × 10⁸ infectious units), ChAdOx1-H7 (1 × 10⁸ infectious units), or both, as described. Two weeks following vaccination, splenocytes were isolated, and T cell responses were measured by ex vivo IFN-γ ELISPOT, as before. Mice vaccinated with ChAdOx1 NP+M1 had higher responses compared with mice that received a mixture of ChAdOx1 NP+M1 and ChAdOx1-H7 (**p ≤ 0.01) (Fig. 3A). However, no significant difference in ELISPOT responses was observed between the responses in mice vaccinated with ChAdOx1 NP+M1 or a combination of ChAdOx1 NP+M1 and ChAdOx1-H7, administered into different limbs. This approach has previously been shown to augment immune responses and avoid competition between two vaccines administered concurrently (12).

No significant differences were detected between the H7HA-specific IgG Abs at 2 wk after viral vector vaccination (Fig. 3B). However, the response after ChAdOx1-vectored vaccines was significantly higher compared with vaccination with 10 μg of recombinant H7HA (**p ≤ 0.01) (Fig. 3B). These data demonstrate that vaccination with ChAdOx1-vectored vaccines expressing H7HA elicits superior humoral immunity compared with protein; moreover, these responses are maintained in multiantigen vaccination regimens.

Adenovirus–MVA prime-boost regimens are one of the leading strategies to induce potent immune responses against vaccine Ags (21, 22). BALB/c mice (n = 18 total, n = 6 per group) received a priming vaccination of ChAdOx1-NP+M1, ChAdOx1-H7, or both, administered separately into opposite limbs, as described. At 8 wk after prime, all groups were boosted with MVA-NP+M1-H7.

Splenic cells were isolated 2 wk after prime and after boost to assess T cell responses against NP+M1 and H7HA. Consistent with previous data, NP_147–158-specific T cell responses were slightly higher in mice primed with ChAdOx1 NP+M1 compared with ChAdOx1 NP+M1 and ChAdOx1-H7 administered into opposite limbs (Fig. 4A). However, following MVA vaccination, T cell responses against NP+M1 were ∼5-fold higher in all groups and there were no significant differences, after boost toward NP Ag, between mice that were primed with ChAdOx1 NP+M1 or coadministration of ChAdOx1 NP+M1 and ChAdOx1-H7 (Fig. 4A). T cell responses against H7HA were also measured by ex vivo IFN-γ ELISPOT against two H7HA peptide pools: one representative of H7HA from A/Netherlands/219/2003 (H7N7) (Fig. 4B) and another representative of amino acid sequence diversity between A/Netherlands/219/2003 (H7N7) and divergent strains [A/Anhui/1/2013 (H7N9) and A/Shanghai/1/2013 (H7N9)] (Fig. 4C). Sequence homology at the amino acid level for divergent strains [A/Shanghai/1/2013 HA and A/Anhui/1/2013 (H7N9)] and the vaccine insert, A/Netherlands/219/2003 HA, was 96%.

As expected, mice primed only with ChAdOx1 NP+M1 had no detectable H7HA-specific T cell responses (Fig. 4B, 4C, left panel). There were no significant differences between the H7HA-specific T cell responses at 2 wk after prime or after boost between mice vaccinated with ChAdOx1-H7 or ChAdOx1-H7 coadministered with ChAdOx1 NP+M1 (Fig. 4). Collectively, these results demonstrate that vaccination with ChAdOx1-H7 or coadministration of ChAdOx1 NP+M1 and ChAdOx1-H7, followed by immunization with MVA-NP+M1-H7, induces heterosubtypic T cell responses against H7HA.

Serum was collected at 2 and 8 wk after prime and after boost vaccinations, and the longevity of Ab responses was followed out to 26 wk following the initial immunization. Mice primed with ChAdOx1-H7 or ChAdOx1-H7 and ChAdOx1 NP+M1 had higher total IgG against H7HA at all time points compared with vaccination with either protein alone or with ChAdOx1 NP+M1 followed by MVA-NP+M1-H7 (Fig. 5A). As expected, we were unable to detect serum responses to HA in the ChAdOx1 NP+M1–only group until 2 wk after boost with MVA-NP+M1+H7. Peak boost responses for viral vector vaccinations were up to 50-fold higher than the response 2 wk following prime immunization (e.g., group 3: 1.34 × 10⁵ relative ELISA units [2 wk] versus 7.4 × 10⁶ relative ELISA units [16 wk]) and persisted for ≥26 wk after vaccination in all groups (Fig. 5A). These data suggest that strong humoral immune responses toward H7HA are generated and maintained over time by heterologous ChAd–MVA prime-boost regimens.

Long-lived humoral immunity is principally mediated by two B cell subsets: long-lived plasma cells (LLPCs) and memory B cells. LLPCs predominantly reside in the bone marrow (23, 24) and continuously secrete Ab. To further understand the basis of the humoral responses following heterologous prime-boost ChAd–MVA viral-vectored vaccination, LLPCs were enumerated 18 wk following the boosting vaccination.

Total IgG⁺ ASCs and H7HA-specific ASCs representative of LLPCs were measured using an IgG ASC ELISPOT assay. Elevated numbers of total IgG-secreting LLPCs were detected in the bone marrow of all immunized groups (Fig. 5B, 5C). However, elevated numbers of H7HA-specific LLPCs were detected in mice that had been primed with ChAdOx1-H7 or ChAdOx1-H7 and ChAdOx NP+M1 and boosted with MVA-NP+M1+H7 compared with naive BALB/c mice (Fig. 5B, 5C, 2 and 3). There was no significant difference between the number of H7HA-specific LLPCs detected in mice primed with ChAdOx1-H7 only (median SFU = 541) or ChAdOx1-H7 and ChAdOx1 NP+M1 (median SFU = 589). However, these numbers were higher compared with mice primed with ChAdOx1 NP+M1 (median SFU = 100) or protein alone (median SFU = 89) (Fig. 5B).

To assess the breadth of anti-HA Ab functionality, sera collected 8 wk after boosting with MVA-NP+M1-H7 were assayed against a number of pseudotyped lentiviruses. Two strains of H7HA pseudotypes were tested: A/chicken/Italy/1082/1999 (H7N1), a low pathogenic avian influenza strain closely related to the vaccine immunogen (98% homology at the amino acid level), and A/Shanghai/2/2013 (H7N9), the novel H7N9 first identified in humans in 2013 (96% homology at the amino acid level). A third group 2 HA lentivirus, expressing a different subtype, H3HA from A/Udorn/307/1972 (H3N2) (48% homology at the amino acid level), was also tested.

Pooled sera from mice primed with ChAdOx1-H7 or ChAdOx1-H7 and ChAdOx1 NP+M1 completely neutralized both H7 pseudotypes at all serum dilutions tested (Table II). In addition, IC₅₀ values of sera from mice primed with ChAdOx1 NP+M1 and boosted with MVA-NP+M1-H7 were higher than in mice vaccinated with protein alone. IC₅₀ values against the H3N2 pseudotype lentivirus were comparable among all groups vaccinated with viral vectors, but they were lower in the control group vaccinated with protein alone (Table II).

To assess heterosubtypic protective efficacy, mice were vaccinated, as described, and challenged with a lethal dose (5 × LD₅₀) of A/Shanghai/1/13. Amino acid sequence homology for A/Shanghai/1/13 HA7 (EPI439486) and A/Netherlands/219/2003 (AY340089.1) HA7 is 96%. Sequence homology, at the amino acid level, for the challenge strain NP+M1 and viral vector-encoded NP+M1 is 97 and 93%, respectively.

Negative controls (n = 10) were naive animals or animals that received a ChAdOx1 and MVA prime (both encoding an irrelevant Ag GFP) boost vaccination. Three weeks after the last immunization, animals were challenged with five murine LD₅₀ of SH1 (A/Shanghai/1/13) virus. Weight loss was monitored over a period of 14 d, and mice that lost >25% of their initial body weight were euthanized (Fig. 6).

Animals vaccinated with ChAdOx1-H7 alone or ChAdOx1-H7 and ChAdOx1 NP+M1 and boosted with MVA-NP+M1-H7 (groups 2 and 3) all survived lethal challenge. When these studies were repeated, groups 2 and 3 and group 1 (primed with ChAdOx1 NP+M1 boosted with MVA-NP+M1-H7) were equally protective as a positive control vaccination regimen previously shown to be protective (group 5, protein and a TLR agonist adjuvant) (25).

Importantly, in the first challenge experiment, groups 2 and 3 (animals vaccinated with ChAdOx1-H7 alone or ChAdOx1-H7 and ChAdOx1 NP+M1 and boosted with MVA-NP+M1-H7) retained their starting body weight throughout the monitoring period (Fig. 6A). Furthermore, in a second challenge experiment, group 1 also retained starting body weight (Fig. 6B). Comparison across the nadir of weight loss (days 4/5 through 8/9) between ChAdOx1 NP+M1–primed and MVA-NP+M1-H7–boosted animals (group 1) and those that received a protective regimen (group 5, positive control) demonstrates no significant difference in weight loss in the first challenge or in a second independent repeat challenge. It is evident that humoral and cellular immunity can offer improved efficacy in this stringent challenge model compared with protective vaccination regimens (protein and adjuvant) (Fig. 6).

In 2013, avian influenza A (H7N9) first caused an outbreak of severe respiratory illness in China. It re-emerged during winter 2013–2014, with ≥630 laboratory-confirmed infections documented by April 2015, and an associated mortality > 30% (1). However, most recently, in excess of 600 new cases have been reported during the fifth wave of H7N9 (start of 2017), which is now the biggest wave since human infection was first detected with a worryingly high case fatality rate (40%) (1–3). Consequently, there is an ongoing and pressing need for vaccines that can protect against avian-derived influenza viruses, especially given that H7N9 viruses now exhibit a seasonal pattern of circulation (2).

Clinical development of influenza vaccines is ongoing; however, split virus or subunit vaccines for avian influenza are known to be poorly immunogenic (26, 27) and often require multiple doses and/or formulation with potent adjuvants to achieve seroconversion (28). Although live attenuated vaccines can induce humoral and cellular immunogenicity, in adults these vaccines have previously been associated with lower seroconversion rates and higher rates of laboratory-confirmed influenza compared with trivalent influenza vaccine. These phenomena may be due to pre-existing immunogenicity at mucosal sites (7, 29, 30). Less than half of the vaccinees in a recent phase I clinical trial assessing safety and immunogenicity of a H7N9 live attenuated vaccine seroconverted (48%, 95% confidence interval 29.4–67.5) after one vaccination (31).

Vaccines that target conserved Ags, such as internal proteins of IAVs, may provide greater cross-protective responses toward diverse influenza strains, including newly emergent pandemic variants. We demonstrate that, although mice primed with ChAdOx1 NP+M1 and boosted with MVA-NP+M1-H7 had significantly less HA-specific Abs (Fig. 5), there was no difference in morbidity or mortality compared with a protein and adjuvant-only regimen (Fig. 6). These data highlight and confirm that T cells confer a degree of protection against the clinical symptoms of IAV infection. Most commercially available influenza vaccines primarily induce strain-specific Abs; however, it has been demonstrated that heterosubtypic T cells can confer broad-spectrum protection (32–34). A correlation between IAV-directed T cells and reduced viral shedding with less severe illness in humans has been demonstrated in a number of clinical studies (32, 33, 35). It has also been demonstrated that protective levels of NP-specific T cell responses are found in 43% of the adult population (36). Importantly, if a vaccine can boost the numbers of pre-existing influenza-specific T cells into this protective range, these vaccinees would be conferred a degree of protection toward newly pandemic influenza viruses. This level of boosting is achievable with clinical vaccination with MVA-NP+M1 in humans (37). In the event of a virulent pandemic outbreak, vaccination with MVA-NP+M1 could curb disease symptoms while the strain-specific HA protein or vaccine modality encoding the outbreak HA Ag could be manufactured. Follow-on vaccination with a strain-specific HA could then provide neutralizing Abs toward emergent viruses and curb disease transmission.

Advantageously, viral-vectored vaccines can facilitate delivery of multiple disease-specific Ags, which is thought to be key in curtailing viral escape mutants compared with vaccines that target a single Ag. However, delivery of multiple Ags can result in immune competition (12, 38), which can be largely circumvented by administration of viral-vectored vaccine-encoded Ags to separate sites, as described in this article. Although the exact mechanisms of antigenic interference following vaccination remain unknown, this phenomenon is thought to be influenced by spatial constraints on T cells (39, 40). The delivery of dual Ags by the bivalent MVA-vectored vaccine is less likely to induce immune interference, because distinct promoters drive Ag expression at different times following infection; the early F11 promoter is expressed before P7.5. In fact, very early expression of T cell–stimulating Ags by MVA has previously demonstrated higher T cell responses and reversal of immunodominance hierarchies (41). Promisingly, boosting with MVA-NP+M1-H7 significantly enhanced T cell responses against NP+M1 and H7HA, regardless of whether antigenic competition was observed following a priming vaccination.

In a stringent challenge model, inclusion of a viral-vector encoded HA at the prime and boost significantly outperformed all regimens in both challenges across the nadir of infection (day 4/5 through 8/9), and all animals in these groups retained their starting body weight throughout the heterologous challenge. Encouragingly, the two strains of H7-pseudotyped viruses used to assess responses were neutralized at all serum dilutions tested in mice primed with HA7 and NP+M1 Ags, and these humoral responses were maintained for up to 18 wk after vaccination. This is an important finding in light of the pandemic threat posed by currently circulating avian H7 viruses.

In summary, vaccination against HA, NP, and M1 at both prime and boost immunizations delivered by ChAd–MVA viral-vectored vaccines induces potent T and B cell responses. These novel bivalent MVA-vectored vaccines elicit potent T cell responses against NP+M1 while simultaneously inducing high levels of Abs that can recognize different HA subtypes. Furthermore, T cell responses against NP+M1 were significantly higher than responses induced by the first generation of clinically investigated MVA-vectored vaccines, a particularly encouraging result for future clinical work (37, 42). Our data show that these humoral and cellular responses, induced following a prime-boost vaccination, are heterologous and homologous in nature and can confer protection in a rigorous challenge model. Indeed, the simultaneous delivery of dual Ags (H7HA and NP+M1) outperformed a previously published efficacious vaccination regimen and, importantly, the dual delivery of Ags may alleviate the selective pressure currently thought to potentiate antigenic diversity in avian influenza vaccination (1, 43).

We thank V. Clark and H. Gray for animal husbandry and A. Worth and the Jenner Institute Vector Core facility for assistance. The following reagents were obtained through BEI Resources, National Institute of Allergy and Infectious Diseases, National Institutes of Health: peptide array, influenza virus A/Anhui/1/2013 (H7N9) HA protein, NR-44011; peptide array, influenza virus A/Shanghai/1/2013 (H7N9) HA protein diverse peptides, NR-44012; and H5 HA protein from influenza virus, A/Vietnam/1203/2004 (H5N1), recombinant from baculovirus, NR-10510.

S.C.G. is a named inventor on a patent application describing the ChAdOx1 vector (GB Patent application number 1108879.6) and is a named inventor on patents relating to methods of vaccination, including influenza vaccines. S.C.G. acts as a consultant for Vaccitech, a spin out company from the University of Oxford. The other authors have no financial conflicts of interest.