Induction of Broadly Neutralizing H1N1 Influenza Antibodies by Vaccination

See allHide authors and affiliations

Science  27 Aug 2010:
Vol. 329, Issue 5995, pp. 1060-1064
DOI: 10.1126/science.1192517


The rapid dissemination of the 2009 pandemic influenza virus underscores the need for universal influenza vaccines that elicit protective immunity to diverse viral strains. Here, we show that vaccination with plasmid DNA encoding H1N1 influenza hemagglutinin (HA) and boosting with seasonal vaccine or replication-defective adenovirus 5 vector encoding HA stimulated the production of broadly neutralizing influenza antibodies. This prime/boost combination increased the neutralization of diverse H1N1 strains dating from 1934 to 2007 as compared to either component alone and conferred protection against divergent H1N1 viruses in mice and ferrets. These antibodies were directed to the conserved stem region of HA and were also elicited in nonhuman primates. Cross-neutralization of H1N1 subtypes elicited by this approach provides a basis for the development of a universal influenza vaccine for humans.

Seasonal influenza outbreaks are driven by the evolution of diverse viral strains that evade human immunity. Immune protection is mediated predominantly by neutralizing antibodies directed to the hemagglutinin (HA) of these viruses, and the coevolution of HA and neuraminidase (NA) generates variant strains that become resistant to neutralization. Yearly influenza vaccine programs have relied on surveillance of circulating viruses and the identification of strains likely to emerge and cause disease (1). An alternative approach to influenza prevention is the generation of universal influenza vaccines. This strategy is based on the premise that invariant regions of the viral proteins can be identified as targets of the immune response. Several broadly neutralizing antibodies directed against the viral HA have been identified (26), and the structural basis of antibody recognition and neutralization has been recently elucidated (3, 4). Although this knowledge has identified at least one functionally conserved and constrained target of neutralizing antibodies, it has not been possible to elicit such broadly neutralizing antibodies by vaccination. Gene-based vaccination offers the potential to improve the priming of immune responses that can subsequently enhance immunity induced by the appropriate heterologous boost (7). In this study, we examined whether gene-based priming could potentiate the neutralizing antibody response elicited by the seasonal influenza vaccine or by replication-defective adenovirus 5 (rAd5) encoding HA by evaluating the potency, breadth, and efficacy of cross-protection in relevant animal models.

To elicit neutralizing antibody responses with greater breadth and potency, plasmid expression vectors encoding H1N1 or H3N2 HAs were prepared based on the 2006–2007 vaccine strains A/New Caledonia/20/99 (NC) (1999 NC) and A/Wisconsin/67/05 (2005 WI) (8), respectively. Mice were immunized with an empty plasmid (control) or a HA-encoding plasmid, followed by a boost with a trivalent 2006–2007 seasonal vaccine that expressed matching H1 or H3 HA. Gene-based vaccination with 1999 NC HA followed by 2006–2007 seasonal vaccine boosting stimulated a greater than 50-fold increase in neutralizing antibody titer than that produced by one dose of seasonal vaccine alone or DNA alone (Fig. 1A). To evaluate the breadth of neutralization, antisera were analyzed for their ability to neutralize heterologous H1N1 strains. Remarkably, the DNA/seasonal vaccine antiserum neutralized previous H1N1 strains dating back to 1934 (Fig. 1B). This antiserum also inhibited the activity of A/Brisbane/59/2007 (Fig. 1B). Priming with HA from a different subtype, H3N2 (2005 WI), failed to stimulate an increase in neutralization titer to 1999 NC after a 2006–2007 seasonal vaccine boost (Fig. 1C), though it did increase H3N2 neutralization titers against both autologous and heterologous H3N2 viruses (fig. S1). DNA priming with matched H1N1 HA was therefore required to boost the seasonal vaccine neutralizing antibody response to homologous and heterologous H1N1 strains. Because it stimulates strong boosting capacity for antibodies and also allows for evaluation of a single matched HA strain boost, we also evaluated the ability of a rAd5 HA vector to stimulate this response. Immune sera from mice immunized with H1 HA DNA/vaccine or DNA/rAd5 HA also neutralized other group 1 influenza strains such as H2N2 and H5N1 viruses (fig. S2), indicating that this prime/boost immunization strategy broadens neutralization beyond the H1N1 subtype.

Fig. 1

Increased titer and breadth of neutralizing antibodies to H1N1 strains elicited by DNA/seasonal flu vaccine immunization. (A) Pseuotyped neutralization assay to measure the neutralizing antibody response in mice immunized with homologous H1N1 1999 NC HA DNA vaccine, seasonal flu vaccine, or a DNA prime and seasonal flu vaccine boost regimen. (B) The neutralizing antibody response in mice against 1934 PR8, A/Singapore/6/1986 (1986 Sing), and 2007 Bris HA-pseudotyped lentivirus reporters after a DNA prime/vaccine boost regimen. (C) The neutralizing antibody response against a 1999 NC HA pseudotyped lentivirus reporter in response to an H3N2 HA DNA (A/Wisconsin/67/2005) prime/seasonal vaccine boost.

We next evaluated whether these antibodies would confer protection against lethal challenge in mice. Protective immunity was tested using the most distant H1N1 strain, derived from the 1934 virus (1934 PR8). Animals were immunized with DNA alone, seasonal vaccine alone, or the prime/boost combination. A 1934 PR8 DNA prime followed by boost with rAd5 encoding 1934 PR8 HA served as a positive control. Animals immunized with the 1999 NC DNA/seasonal vaccine showed significantly increased survival rates (Fig. 2A, left; P < 0.0001) and less body weight loss (Fig. 2A, right) as compared to DNA alone–, seasonal vaccine alone–, or sham-immunized controls. Although the survival rates for the matched DNA/rAd5 1934 PR8 trended higher than the 1999 NC DNA/seasonal vaccine group, the difference was not statistically significant (Fig. 2A, P = 0.3714). For the prime/boost vaccine, survival did not correlate with the hemagglutination inhibition (HAI) or microneutralization titer to 1934 PR8 or 1999 NC viruses. Rather, the increase in neutralization of the 1934 PR8 strain by DNA/seasonal vaccine sera was detected with a pseudotyped lentiviral reporter assay (Fig. 2A), which provides a more sensitive measure of HA-specific viral neutralization (911).

Fig. 2

Immune protection conferred against lethal challenge by 1934 PR8 influenza virus in mice and against infection by 1934 PR8 or 2007 Bris in ferrets. (A) Protection of prime/boost immune mice after heterologous virus challenge (upper panel). Mice were immunized with control vector (n = 5 mice), PR8 HA DNA/PR8 rAd5 (n = 5), H1 HA DNA (n = 20), seasonal vaccine (n = 5), or H1 HA DNA/seasonal vaccine (n = 20). Three weeks after the final immunization, the animals were challenged with 50 median lethal doses of 1934 PR8 virus, and survival (left) and weight loss (right) were recorded and evaluated. P = 0.3713 between the PR8 DNA/PR8 rAd and DNA/vaccine groups; P < 0.0001 for DNA/vaccine as compared to DNA-only or vaccine-only groups by Kaplan-Meyer analysis. (Lower panel) The antibody responses to homologous (1999 NC) or heterologous (1934 PR8) HAs elicited by HA DNA alone, seasonal vaccine alone, or HA DNA prime/seasonal vaccine boost immunization were measured by HAI (left), microneutralization (middle), and pseudotyping (right) assays. Median inhibitory concentration (IC50) titers for the pseudotyped lentiviral vector reporter assay are shown. Titers of 100 to 1000 are shown in yellow and of ≥2560 in red. (B) Protection of ferrets from 2007 Bris viral challenge. TCID50, median tissue culture infective dose. Two groups of four ferrets were immunized with 1999 NC HA DNA/seasonal flu vaccine or control vector and challenged with heterologous 2007 Bris virus [106.5 egg infective dosage (EID50)]. Virus titers in the nasal swabs from day 1 and day 5 after challenge were determined by means of end-point titration in Madin-Darby canine kidney cells. P = 0.0104 between day 5 control and day 5 prime/boost. (C) Protection of ferrets from 1934 PR8 viral challenge. Two groups of six ferrets were immunized with 1999 NC HA DNA/rAd5 vaccine or control vector and challenged with heterologous 1934 PR8 virus (106.5 EID50). Virus titers in the nasal swabs from day 1 and day 5 after challenge were determined in eggs from an initial dilution of 1:10 in phosphate-buffered saline and expressed as EID50/ml. The limit of virus detection was 101.5 EID50/ml. P = 0.0004 between day 5 control and day 5 prime/boost.

Infectious challenge in ferrets is widely considered a better model to predict vaccine efficacy in humans (12). To evaluate immune protection in ferrets, animals were first immunized with the prime/boost combination that conferred cross-neutralization in mice. Under this vaccine regimen and as seen in mice, we also observed cross-reactive neutralizing antibodies to H1N1 viruses (Table 1, A and B). These animals were next tested by challenge with a seasonal 2007 virus, the Brisbane (Bris) strain. The DNA prime/vaccine boost immunization conferred protection against the 2007 virus, as indicated by the significantly reduced viral titers in the nasopharynx (Fig. 2B). We also evaluated the ability of DNA prime/rAd5 HA immunization to elicit broadly neutralizing antibodies and protection. Ferrets immunized with this gene-based combination generated higher titers of cross-neutralizing antibodies (Table 1B) and were protected against challenge with a more divergent strain, 1934 PR8, showing a >2 log reduction in nasopharyngeal viral loads (Fig. 2C).

Table 1

Neutralization activity of mouse, ferret, and monkey antisera against H1N1 pseudotyped viruses. (A) Neutralization activity of murine antisera from DNA-, seasonal vaccine– (one or two doses), or DNA/seasonal vaccine–immunized mice against H1N1 pseudotyped viruses (1986 Sing; A/Beijing/262/1995, 1995 Bei; 1999 NC; A/Solomon Islands/3/2006, 2006 SI; 2007 Bris). (B) Neutralization activity of antisera from DNA/seasonal vaccine– or DNA/rAd5–immunized ferrets against the indicated H1N1 pseudotyped viruses. (C) Neutralization activity of antisera from DNA-, seasonal vaccine–, or DNA/seasonal vaccine–immunized monkeys against the indicated H1N1 pseudotyped viruses. IC50 titers are shown for all panels. Titers of <100 (low) are shown in green, of 100 to 1000 (medium) in yellow, and >1000 (high) in red.

View this table:

We analyzed the target of neutralization breadth further by testing DNA, vaccine (one or two doses as currently recommended for human vaccines), or DNA prime/seasonal vaccine boost sera against a variety of strains. In mice, the highest neutralization titers were generated against the homologous 1999 NC strain or an earlier strain, A/Beijing/262/1995 (1995 Bei), by all vaccine regimens; however, minimal cross-neutralization of other strains was observed with DNA or seasonal vaccine immune sera as compared to DNA/seasonal vaccine (Table 1A). Two doses of seasonal vaccine increased the neutralization titer against the homologous 1999 NC HA but had a minimal effect on heterologous neutralization. In nonhuman primates, a similar increase in titer and breadth of neutralizing antibodies to H1N1 viruses was elicited by this prime/boost immunization (Table 1C).

To document that neutralizing antibodies were directed to the highly conserved HA stem, we included wild-type 1999 NC HA trimer (WT) or a matched stem mutant protein (∆Stem) as competitors in the neutralization assay. The stem mutant trimer showed minimal reactivity with the previously defined C179 monoclonal antibody (mAb) directed to this region, in contrast to WT 1999 NC HA trimer (Fig. 3A). The specificity of this stem mutant was also confirmed by size-exclusion column chromatography, showing that it forms a trimer of the appropriate size (fig. S3A) recognized by a mAb to the 1999 NC HA head, and yet it fails to react with three mAbs specific for the conserved region of the HA stem (24): C179, CR6261, and F10 (fig. S3B). When included as competitors in the neutralization assay, WT 1999 NC HA trimer, but not the stem mutant, blocked neutralization against 1934 PR8 virus by the stem-directed C179 mAb (Fig. 3A), further documenting the specificity of this stem mutation. When mouse sera from DNA/vaccine- or DNA/rAd5 HA–immunized animals were analyzed in this way, both WT and stem mutant HAs inhibited neutralization against homologous 1999 NC virus, but the stem mutant failed to block neutralization against heterologous 1934 PR8 virus (Fig. 3B). In ferrets, which showed protection against the 1934 PR8 and 2007 Bris virus, as expected, both WT and stem mutant HA trimers blocked neutralization against homologous 1999 NC virus; however, only the WT protein inhibited the neutralization against heterologous 2007 Bris virus or 1934 PR8 virus (Fig. 3C). Additional competition assays were performed with the stem-directed CR6261 mAb to further document the specificity of these antisera. Antisera from DNA/rAd5 HA–immunized ferrets were preabsorbed with cells expressing the stem mutant of the 1999 NC HA to remove non–stem-directed HA antibodies. By enzyme-linked immunosorbent assay (ELISA), we examined the ability of the CR6261 antibody to block the binding of ferret sera to homologous or heterologous H1N1 HAs as compared to a control immunoglobulin G (IgG). CR6261, in contrast to control antibody, inhibited the binding of ferret sera to these HA trimers, further confirming the presence of stem-specific antibodies in the ferret sera (Fig. 3D). We also demonstrated the presence of antistem antibodies in monkeys immunized with a DNA/seasonal vaccine prime/boost combination. Monkey sera were absorbed with cells (10) expressing WT or stem mutant 1999 NC HA. Absorption with the WT HA removed reactivity to homologous and heterologous HAs (Fig. 3E). In contrast, antisera absorbed with cells expressing the stem mutant retained reactivity with heterologous HA trimers while retaining a lower level of homologous binding as expected (Fig. 3E). Together these results demonstrate the specificity of the antistem antibodies elicited by the prime/boost immunization in mice, ferrets, and nonhuman primates.

Fig. 3

Stem-directed antisera elicited by HA DNA/seasonal flu vaccine immunization. (A) WT or ΔStem HA protein was immunoprecipitated with C179 mAb or a nonreactive isotype control (IgG2a) and was detected with an antibody to a histidine tag (left panel). mAb C179 was preabsorbed with HIV Env (HIV), WT 1999 NC trimer (WT), or stem mutant 1999 NC trimer (∆Stem), and the neutralization activities of the preabsorbed antibody were measured with 1934 PR8 HA-pseudotyped lentiviral vectors (right panel). The percent reduction in neutralization was determined at 5 μg of C179 per milliliter. (B) Immune sera from mice immunized with H1 HA DNA/seasonal flu vaccine or H1 HA DNA/rAd5 were preabsorbed as described in (A), and the neutralization activities of the preabsorbed antisera were measured with 1999 NC or 1934 PR8 HA-pseudotyped lentiviral vectors (at a 1:200 serum dilution). (C) Immune sera from ferrets immunized with H1 HA DNA/seasonal flu vaccine or H1 HA DNA/rAd5 were preabsorbed as above, and the neutralization activities of the preabsorbed antibody or antisera were measured with 1999 NC, 2007 Bris, or 1934 PR8 HA-pseudotyped lentiviral vectors (at a 1:200 serum dilution). (D) Antisera from DNA/rAd5 HA immunized ferrets were preabsorbed with cells expressing the stem mutant of the 1999 NC HA to remove non–stem-directed HA antibodies. ELISA plates coated with 1999 NC or 1986 Sing HA trimers were preincubated with a control IgG or CR6261 before the addition of the preabsorbed sera. Detection of the presence of ferret antibodies was performed with an anti-ferret secondary antibody. (E) Antisera from monkeys immunized with H1 HA DNA/seasonal vaccine were preabsorbed with 293F cells expressing either WT or ∆Stem of 1999 NC HA, and the binding of preabsorbed sera to 1999 NC, 1986 Sing, or 2007 Bris HA trimers was examined by ELISA.

Protection by antibodies directed to the conserved stem of the HA in ferrets is probably relevant to influenza immunity in humans. The presence of these antibodies was highly correlated with efficacy and suggested that neutralization function contributes to protection. The generation of these antibodies was dependent on gene-based priming, which can increase the number and diversity of CD4 clones (7) that stimulate B cells to secrete antibodies of greater magnitude and diversity. In fact, we observed that the DNA prime/vaccine boost elicited higher HA-specific T cell responses as compared to vaccine alone (fig. S4). Multiple doses of vaccine with inclusion of a B cell adjuvant, or other immunization approaches, could possibly help achieve this effect. Recent publications have shown that priming with vaccine elicits cross-reactive CD4+ T cells (13), and the MF59 adjuvant expands the antibody repertoires against H5N1 influenza virus (14).

Vaccine-elicited antisera almost exclusively target the variable head region of HA rather than the conserved stem. Although broadly neutralizing antibodies to HA have been derived from mice (2), humans (5, 6), or recombinant antibody libraries (3, 4), it has not been possible to specifically elicit them through vaccination, a difficulty shared by other viruses, such as HIV-1 [reviewed in (15)]. In addition to H1N1, this prime/boost combination also elicited an increase in the titer and breadth of antibodies to H3N2 HAs (fig. S1), and it could potentially be applied to influenza B. Stem-focused HA immunogens could also be developed using rational structure-based protein design to increase breadth still more (16). We have recently assessed the efficacy of a DNA/rAd5 prime/boost immunization for enhancing antibody responses in humans with HIV-1 Env immunogens (17). As observed in other nonhuman primate and rodent studies (18, 19), this vaccine platform elicited an enhancement of antibody responses in humans similar to those to the DNA/rAd HA vaccine described here. Together, these data support the applicability of this vaccine strategy to humans. In such studies, it will be important to define immune correlates of protection, which will probably differ from those for seasonal vaccines. Pre-existing influenza immunity in humans could possibly affect vaccine efficacy. In this case, the vaccine could still be deployed in influenza-naïve children or infants. Evaluation of this first-generation universal H1N1 vaccine candidate in clinical studies will determine its ability to protect against natural infection and improve the public health benefit of influenza vaccination.

Supporting Online Material

Materials and Methods

Figs. S1 to S4


References and Notes

  1. Materials and methods are available as supporting material on Science Online.
  2. We thank A. Ault, J.-P. Todd, A. Zajac, and C. Chiedi for help with the animal studies; K. Dai, W. Shi, and S. Y. Ko for technical support; M. Lewis (BIOQUAL), B. Sanders (BIOQUAL), and members of the Nabel lab for helpful discussions; A. Tislerics and B. Hartman for manuscript preparation; and Y. Okuno for providing the C179 mAb. NIH has filed a patent application (U.S. patent E-341-2008l; international patent WO 2010/036958 A2) on this work (authors: C-J.W., Z-y.Y, and G.J.N.), related to gene- and protein-based approaches to influenza vaccination. This research was supported by the Intramural Research Program of the Vaccine Research Center, NIAID, NIH. The findings and conclusions in this report are those of the authors and do not necessarily reflect the views of the funding agency.

Stay Connected to Science

Navigate This Article