Research ArticleVaccines

Vaccines with MF59 Adjuvant Expand the Antibody Repertoire to Target Protective Sites of Pandemic Avian H5N1 Influenza Virus

See allHide authors and affiliations

Science Translational Medicine  20 Jan 2010:
Vol. 2, Issue 15, pp. 15ra5
DOI: 10.1126/scitranslmed.3000624

Abstract

Vaccines against influenza viruses with pandemic potential, including H5N1, are under development. Because of a lack of preexisting immunity to these viruses, adjuvants (immune potentiators or enhancers) are needed to improve immune responses, to conserve scarce vaccine, and for cross-protection against strains that have drifted evolutionarily from the original. Aluminum-based adjuvants do not improve vaccine immunogenicity for influenza subunit vaccines, whereas oil-in-water adjuvants are effective, especially with H5N1-inactivated vaccines. We used whole-genome-fragment phage display libraries followed by surface plasmon resonance (SPR) technologies to elucidate the effect of different adjuvants on the antibody repertoire against H5N1 vaccine in humans. The oil-in-water adjuvant MF59 induced epitope spreading from HA2 to HA1 in hemagglutinin (HA) and neuraminidase relative to unadjuvanted or aluminum-adjuvanted vaccines. Moreover, we observed an increase by a factor of 20 in the frequency of HA1-to-HA2–specific phage clones in sera after MF59-adjuvanted vaccine administration and a factor of 2 to 3 increase in the avidity of antibodies binding to properly folded HA1(28–319), as measured by SPR. The adjuvant-dependent increase in binding to conformational HA1 epitopes correlated with broadening of cross-clade neutralization and predicted improved in vivo protection. Thus, MF59 adjuvant improves the immune response to a H5N1 vaccine by inducing qualitative and quantitative expansion of the antibody repertoires with protective potential.

Introduction

The recent rapid global spread of swine-origin H1N1 highlights the need for rapid development of effective vaccines against pandemic influenza viruses (1, 2). Much of our earlier knowledge about vaccines for pandemic influenza strains was derived from studies with the highly pathogenic H5N1 avian influenza A viruses. The H5N1 viruses still cause substantial lethality in bird populations, with numerous instances of human transmission resulting in severe human disease with >60% mortality. Adjuvants are groups of immune modulators that have been used with vaccines to boost the immune responses by various mechanisms. However, only a limited number of adjuvants have been approved by regulatory authorities for general use.

In the case of vaccination against avian H5N1 influenza, the immune responses to unadjuvanted or aluminum-adjuvanted H5N1 vaccines are significantly lower than those observed with vaccines that include oil-in-water adjuvants, which induced higher seroconversion rates (factor of 4 increase in neutralizing antibody titers) and cross-clade protection (against diverse H5N1 types) and allowed antigen dose sparing (37).

MF59 (the prototype oil-in-water adjuvant) has been shown, in multiple clinical trials with both seasonal and pandemic vaccines, to enhance antibody titers to homologous influenza strains and to also provide immunity to drifted (antigenically different) virus strains (810). Does this enhanced immunity to drifted strains result from increased antibody titers or from a qualitative impact on the antibody repertoire? To address this question, we evaluated sera from subjects immunized with unadjuvanted, aluminum-adjuvanted, or MF59-adjuvanted H5N1-inactivated vaccine with influenza-specific whole-genome-fragment phage display libraries (FLU-GFPDLs) expressing protein fragments from all the open reading frames of avian influenza H5N1 hemagglutinin (HA) and neuraminidase (NA) genes. In a previous study, similar methods allowed the identification of viral epitopes recognized by antibodies in sera of individuals who recovered from highly pathogenic H5N1 infection. It also allowed the mapping of broadly neutralizing human monoclonal antibodies (mAbs) derived from these H5N1 survivors. These mAbs were found to recognize nonlinear conformational epitopes presented by large HA fragments, encompassing the receptor binding domain (RBD) (11). Sera from these H5N1 convalescent patients also showed broad recognition of NA, and high reactivity was found with a C-terminal fragment, encompassing the region required for sialic acid binding and the catalytic activity of NA.

Results

MF59-adjuvanted H5N1 vaccine generates broader antibody profiles

To attain information on the epitope profile after vaccination, we analyzed samples from two independent clinical trials sponsored by the National Institute of Allergy and Infectious Diseases (NIAID) and Novartis Vaccines and Diagnostics (NVD). In the NIAID multicenter phase I trial, the immunogenicity of unadjuvanted, aluminum hydroxide–adjuvanted, and MF59-adjuvanted H5N1 (A/Vietnam/1203/2004, clade 1) subunit vaccines was compared after two doses. In that study, aluminum hydroxide–adjuvanted H5N1 vaccine (30 μg of HA per dose) did not enhance the low antibody response rate (29% seroconversion) observed with unadjuvanted vaccine at 45 μg of HA per dose (6). In contrast, antibody titers were higher by a factor of 3, and 63% of subjects achieved the predetermined endpoint (hemagglutination inhibition titer of >40) in the MF59-adjuvanted group (15 mg of HA per dose).

To evaluate the impact of the adjuvants on the quality of the antibody response to the H5N1 vaccine, we analyzed serum samples taken before and after the second vaccination from a subset of subjects enrolled in the NIAID study at Vanderbilt University with hemagglutination inhibition titers of 1:80 (two per group) via FLU-GFPDL (Table 1 and Fig. 1A). Surprisingly, sera taken before vaccination reacted with the H5-GFPDL, but the bound inserts primarily mapped to an HA2 region with high sequence homology (98%) to the seasonal H1N1 strains (Table 1, Fig. 1A, and table S1). Sera taken after the second dose of vaccination from the unadjuvanted vaccine group primarily reacted with phages expressing HA2 epitopes (60% of clones), with a significant boost in the number and diversity of HA2 inserts bound. The sera from unadjuvanted vaccine recipients also recognized phage-displayed HA1 inserts that were relatively short, mapping to the RBD and the N and C termini of HA1, with a low frequency (23% of clones) (Table 1, Fig. 1A, and table S1). The antibody profile of the aluminum-adjuvanted vaccine sera was similar to that of the unadjuvanted vaccine group, with a slight increase in the number of HA1 inserts bound in the aluminum-adjuvanted group (32% of clones) (Table 1, Fig. 1A, and table S1). This finding correlates with the poor hemagglutination inhibition seroconversion rates reported after unadjuvanted and aluminum-adjuvanted vaccination in this trial (6).

Table 1.

Distribution of phage clones after affinity selection on post-H5N1 vaccination sera (all sequenced inserts are shown in tables S1 and S2).

View this table:

Sera from individuals vaccinated with MF59-adjuvanted H5N1 vaccine in the NIAID trial generated a significantly different antibody epitope profile. The total number of bound phage clones increased by a factor of 3 relative to unadjuvanted or aluminum-adjuvanted vaccine sera (Table 1), with a significant expansion of antibodies directed against HA1 fragments (60% of inserts) and a broad recognition of NA (Table 1, Fig. 1A, and table S1). Only the sera from MF59-adjuvanted vaccine group recognized long HA1 sequences, encompassing the RBD, similar to those bound by broadly neutralizing human mAbs and with sera derived from H5N1 survivors (11). Thus, despite similar hemagglutination inhibition titers (1:80) in the post–second dose serum samples, the three vaccine formulations generated markedly different antibody repertoires, with MF59 inducing frequency ratios of HA1-to-HA2–bound phages that were higher by factors of 12 or 8 relative to unadjuvanted or aluminum-adjuvanted H5N1 vaccine, respectively (Table 1).

The potential of MF59 to broaden the diversity of the antibody response against avian H5N1 influenza was further evaluated with samples from a second clinical trial conducted by NVD and designed to compare the immunogenicity of unadjuvanted (15 μg of HA) and two doses (7.5 and 15 μg of HA) of MF59-adjuvanted H5N1 vaccine (12). By panning with the H5N1-GFPDL, serum samples from each of the groups with neutralization titers of ≥80 were pooled. In a preliminary experiment, the pooled sera were adsorbed with the H5N1-GFPDL followed by HA-specific enzyme-linked immunosorbent assay (ELISA). The H5N1-GFPDL (HA+NA) adsorbed 93% and 89% of total HA-specific antibodies from unadjuvanted and MF59-adjuvanted immune sera, respectively (fig. S2).

Affinity selection with H5N1-GFPDL on unadjuvanted vaccine sera exhibited a predominant HA2 response (90% of clones) with much lower frequency (10%) of bound phages expressing HA1 fragments (Table 1, Fig. 1B, and table S2), similar to the findings in the NIAID trial (fig. S2). In contrast, sera from individuals vaccinated with MF59-adjuvanted vaccine (either 7.5 or 15 μg of HA per dose) bound 4 to 5 times as many phages, with the bound phages expressing significantly more sequences from HA1 and NA (Table 1, Fig. 1B, and table S2). A factor of 17 to 20 increase in HA1-to-HA2 ratio of bound clones was observed with MF59-adjuvanted vaccine relative to unadjuvanted H5N1 vaccine (Table 1). Phages expressing large HA1 inserts, encompassing the RBD, were selected only by sera from subjects that had received the MF59-adjuvanted vaccines. The predominant de novo MF59-induced HA epitopes (solid rectangles in Fig. 1, A and B; tables S1 and S2, in colors) mapped to the exposed surface on the HA structure (Fig. 1C).

In recipients of MF59-adjuvanted vaccine in both trials, NA binding antibodies were also generated at high frequency (Fig. 1, A and B, and tables S1 and S2). We compared the patterns of recognition in the two clinical studies and concluded that only sera from subjects immunized with the MF59-adjuvanted vaccines contained antibodies that recognized segments encompassing the amino acid sequence 984 to 1004 (Fig. 1, A and B, and tables S1 and S2), a region that juxtaposes the sialic acid binding site of NA (Fig. 1D).

Fig. 1.

Elucidation of antibody repertoires elicited in humans after vaccination with unadjuvanted and adjuvanted subunit H5N1 vaccines. (A) Schematic alignment of the peptides recognized by post–second vaccination sera in the NIAID-sponsored trial, identified by panning with H5N1-GFPDL A/Vietnam/1203/2004. Amino acid designation is based on the HA+NA protein sequence (fig. S1). Bars with arrows indicate identified inserts in the 5′-3′ orientation in HA1 (red bars), HA2 (blue bars), and NA (black bars). Numbers next to bars indicate inserts that were isolated repetitively (only clones with frequency of >2 are shown; sequenced clones are shown in table S1). (B) Schematic alignment of the peptides recognized by post–third vaccination sera in the NVD trial was identified with H5N1-GFPDL A/Vietnam/1203/2004 [symbols as in (A)] (sequenced clones are shown in table S2). (C and D) Main antigenic clusters in the structures of HA (C) and NA (D) recognized strongly only by sera from MF59-adjuvanted H5N1-vaccinated individuals [corresponding colored sequences appear as solid bars in (A) and (B) and in tables S1 and S2]. (C) Antigenic clusters in HA with marked amino acid residues are shown as surface-exposed colored patches on one HA monomer within the HA trimer structure [Protein Data Bank (PDB) identifier 2IBX (16)]. (D) The immunodominant epitopes in the NA are shown on the tetrameric NA structure [PDB identifier 2HTY (17)], with the predicted site of bound sialic acid shown in red. Side view (bottom) and bird’s-eye view (top) are shown.

MF59-adjuvanted H5N1 vaccine induces higher reactivity to properly folded HA1 domain

Neutralization of influenza virus occurs by antibodies recognizing conformational epitopes primarily located on the globular head of the HA molecule (13, 14). Recently, we demonstrated that sera from H5N1-infected convalescing individuals and human mAbs derived from H5N1 survivors bound to cross-clade neutralizing conformational epitopes presented on large HA1 fragments identified by panning on the same H5N1-GFPDL (11). Therefore, it was important to establish whether the inactivated H5N1 vaccine administered with or without MF59 adjuvant elicited similar conformation-dependent antibodies that bind to properly folded large HA1 fragments. To that end, we expressed HA1(28–319) fragment, identified in the previous study (11), in a bacterial system and purified from inclusion bodies either after controlled redox folding conditions and slow dialysis at pH 7.2 or under denaturing conditions (pH 3.0) (fig. S3A). After purification, SDS–polyacrylamide gel electrophoresis gel analysis (fig. S3A) and circular dichroism (CD) melt spectroscopy (fig. S3B) confirmed that only the pH 7.2 purified HA1 fragment behaved as a properly folded protein with a melting point around 55°C, whereas the melting curve of the pH 3.0 purified protein was typical of an unfolded protein.

Proper folding of HA(28–319) was confirmed by surface plasmon resonance (SPR), in which all three H5N1-neutralizing human mAbs (11, 15) bound exclusively to the properly folded HA(28–319) fragment (pH 7.2) and not to the unfolded protein (pH 3.0) (Fig. 2, A and B).

Fig. 2.

Binding of H5N1-neutralizing human mAbs and post-H5N1 vaccination human sera to properly folded HA(28–319) protein. (A and B) Steady-state equilibrium analysis of human mAbs and vaccine sera to unfolded versus properly folded bacterially expressed H5 HA(28–319) fragment were measured with SPR. Human mAbs FLA5.10, FLD21.140, and FLA3.14 at 10 μg/ml were injected simultaneously onto recombinant HA(28–319) purified at pH 7.2 (A) or pH 3.0 (B), immobilized on a sensor chip through the free amine group, and onto a blank flow cell, free of peptide. Binding of the antibodies to the immobilized protein is shown as resonance units (RU). (C and D) Hundred-fold–diluted individual postvaccination sera from the three groups of the NVD vaccine trial, all with neutralization titers of 160, were passed over a sensor chip–immobilized HA(28–319) purified at pH 7.2 (C) or pH 3.0 (D). Serum from nonresponder (MN titer of <20) is included (green). (E) Cross-neutralization of H5N1 clades with post-H5N1 vaccine samples. Endpoint titers (mean of three replicates) using the same postvaccination sera as in (C) and (D) in a MN assay performed with rgH5N1xPR8 (2:6) reassorted viruses H5N1 A/Vietnam (clade 1.0), A/Indonesia (clade 2.1), and A/Anhui (clade 2.3.4).

To determine whether the MF59-adjuvanted vaccine sera also increased binding to the properly folded HA(28–319) segment that was recognized by the human mAbs with cross-clade protective activities, we applied SPR analysis to postvaccination sera from unadjuvanted and MF59-adjuvanted H5N1 vaccine recipients in the NVD clinical trial. Sera with equivalent microneutralization (MN) titers were analyzed (MN titers against A/Vietnam/1203/2004 of 1:160) (Fig. 2, C and D). A serum with MN titer of <20 from a nonresponder in the unadjuvanted vaccine group was included as control. Binding to the unfolded (pH 3.0) protein was similar with all four postvaccination sera (Fig. 2D). In contrast, binding to the properly folded HA1 fragment demonstrated a clear hierarchy; minimal binding of serum with MN titer of <20 (similar to its binding to the unfolded protein), low binding with the unadjuvanted vaccine serum, and higher binding with sera from the MF59-adjuvanted vaccine (HA at 7.5 or 15 μg per dose). This demonstrates a significant qualitative and quantitative increase in conformation-dependent antibody repertoire to properly folded HA1 fragment in sera from recipients of MF59-adjuvanted subunit H5N1 vaccine (Fig. 2C).

Properly folded bacterial H5N1-HA1 protein adsorbs neutralizing activity in post-H5N1 vaccination sera

The increase in binding to properly folded HA(28–319) correlated with a factor of 4 increase in the cross-neutralization of H5N1 clade 2.1 (A/Indonesia) and clade 2.3.4 (A/Anhui) observed in sera from the MF59-adjuvanted vaccine recipients relative to sera from unadjuvanted vaccine recipients (Fig. 2E). The increase of MN titers against A/Indonesia from 20 to 80 seen with the MF59-adjuvanted vaccine sera predicts in vivo cross-protection against the clade 2.1 H5N1 strain.

The functional importance of binding to properly folded bacterially expressed HA(28–319) protein was further confirmed in adsorption experiments. The properly folded HA(28–319), like the mammalian cell–expressed HA0, adsorbed most of the neutralizing activity of MF59-H5N1 immune sera (Fig. 3, A and B). In contrast, the unfolded HA(28–319) (pH 3.0) and a mixture of short HA1 peptides did not reduce the neutralizing activity (Fig. 3A). Some adsorption of neutralizing activity was also observed with a large properly folded N-terminal HA1(1–102) peptide recognized by most postvaccination sera, which overlaps with HA(28–319) (Fig. 3, A and B).

Fig. 3.

Properly folded HA(28–319) at pH 7.2 can adsorb most of the neutralizing activity in H5N1 + MF59 postvaccination sera. Serum from an individual immunized with MF59 + H5N1-HA (15 μg) with a MN of 640 against the vaccine strain was adsorbed on different mammalian and bacterially expressed H5 HA proteins or synthetic peptides and subjected to a MN assay performed with rgA/Vietnam/1203/2004 virus. (A) End-point neutralization titers of preadsorption or postadsorption sera. Similar results were obtained with sera from four individuals immunized with MF59 + H5N1-HA (15 μg) with MN titers ranging from 468 to 1184. (B) Ten-fold dilution of preadsorption and postadsorption sera [from (A)] was passed on sensor chip–immobilized HA(28–319) purified at pH 7.2, and binding kinetics was followed. Binding of the antibodies to the immobilized ligand is shown as resonance units (RU). SPR analysis confirmed that only adsorption with HA(28–319) (pH 7.2) and mammalian HA completely removed binding to the chip-bound HA. GST, glutathione S-transferase.

Kinetics of antibody development after H5N1 vaccination

To assess the kinetics of binding to properly folded HA1 proteins over the entire NVD study, we evaluated sequential samples of sera taken from individuals vaccinated with unadjuvanted vaccine (responder and nonresponder) and from MF59-adjuvanted vaccine recipients by SPR with HA(28–319) (pH 7.2) (Fig. 4, A to D). Sera from the nonresponder bound minimally to the HA fragment with no increase in binding after consecutive vaccinations (Fig. 4A). Vaccination with unadjuvanted vaccine resulted in a modest increase of binding to the properly folded HA(28–319) protein after the second and third doses (Fig. 4B). In contrast, sera from individuals who received MF59-adjuvanted vaccine exhibited a significant increase in antibody binding and avidity to the properly folded HA(28–319) protein after both the second and the third vaccine doses (Fig. 4, C and D). Similar observations were made in sera from five individuals from each vaccination group (Fig. 4E). The post–second vaccination SPR values tended to be higher in the 15-μg HA group, but the post–third vaccination SPR binding titers were similar for the MF59-adjuvanted vaccine at 7.5 and 15 μg of HA per dose (Fig. 4E). These data suggested that the ability of MF59-adjuvanted vaccines to elicit more potent and broader neutralizing antibodies correlated well with binding to properly folded HA(28–319) protein encompassing the RBD. Isotyping on HA(28–319) in SPR revealed that the increase in antibody binding after vaccination was immunoglobulin G (IgG)–specific.

Fig. 4.

Kinetics of antibody binding to HA epitopes after multiple immunizations with adjuvanted and unadjuvanted H5N1 subunit vaccine. Sequential SPR analysis of vaccine sera (before and after first, second, and third vaccination with unadjuvanted or MF59-adjuvanted H5N1 vaccine) with properly folded bacterially expressed H5 HA(28–319) fragment purified at pH 7.2. Ten-fold–diluted individual sera from three arms of the NVD vaccine trial at prevaccination (green) and at 28 days after each immunization were evaluated. Binding of the sera to the immobilized ligand is shown as RU values. (A) Sera from a nonresponder, with MN titer of <20 after three immunizations with unadjuvanted vaccine. (B) Sera for unadjuvanted vaccine recipient. (C) Sera from recipient of MF59 + HA (7.5 μg). (D) Sera from recipient of MF59 + HA (15 μg). Values in parentheses denote the endpoint MN titers (mean of three replicates) using individual sera in a MN assay performed with A/Vietnam/1203/2004-rgH5N1xPR8 reassorted virus. (E) Maximum RU values in SPR with H5 HA(28–319) fragment purified at pH 7.2 are shown for five individuals from each of the three vaccine arms in the NVD trial after first, second, and third immunizations. The average of maximum RU values for the five individuals after second immunization (in blue) and third immunization (in red) is depicted by horizontal lines.

Discussion

GFPDL and SPR analyses enabled a detailed assessment of the humoral responses to three different vaccine formulations: traditional unadjuvanted inactivated split influenza vaccine, aluminum-adjuvanted vaccine, and MF59-adjuvanted vaccine. The main findings of the study were as follows: (i) The oil-in-water adjuvant MF59 induced epitope spreading from HA2 to HA1 (which contains most of the neutralizing epitopes) and NA; (ii) a factor of 8 to 20 increase in frequency of HA1-to-HA2–specific phage clones was observed in MF59-adjuvanted vaccine sera relative to sera obtained after unadjuvanted or aluminum-adjuvanted vaccine; (iii) some of the newly induced antibodies in the MF59-adjuvanted vaccination sera targeted large conformational epitopes in HA1 that correlated with broad cross-clade neutralization and were not induced by unadjuvanted or aluminum-adjuvanted vaccines; and (iv) antibodies against potentially protective epitopes in the C terminus of NA, close to the sialic acid binding enzymatic sites, were also induced primarily after vaccination with MF59-adjuvanted vaccine and not with the other two presentations. These findings can explain the superior performance of oil-in-water–adjuvanted vaccines in inducing higher and broader neutralizing antibody responses in preclinical and clinical trials.

Despite the strong similarities in the recognition patterns of HA observed in recipients of the MF59-adjuvanted vaccines in both trials, we noted some differences in the antibody repertoires, among them a more frequent recognition, in the second NVD study, of N-terminal HA1 sequences and of HA2 inserts spanning the N-terminal sequence of HA2 next to the junction to the HA1 that includes the HA2 fusion peptide (Fig. 1, A and B, and tables S1 and S2). One possible explanation is that the two vaccine formulations were not totally identical: The two different subunit vaccines were produced in two different manufacturing sites and the formulation with MF59 was extemporaneous (at the bed side) in the first study (Fig. 1A), whereas vaccine and adjuvants were preformulated in prefilled syringes in the second study (Fig. 1B).

Our study revealed a previously unappreciated capacity of oil-in-water adjuvants, of which MF59 is the prototype, both to enhance the antibody responsiveness to influenza and to directly reshape the quality of the antibody response against the influenza antigens. It was previously reported that a rapid expansion of influenza-specific CD4+ T cells after the first dose of MF59-adjuvanted vaccine preceded the rise and maintenance of neutralizing antibody titers during the NVD H5N1 vaccine trial (12). The early rise in helper T cell activity could have promoted the selection and expansion of antigen-specific B cells against protective targets in HA1 and NA and their affinity maturation in the germinal centers. Our current findings, coupled with a previous study on long-term memory B cell induction after MF59-adjuvanted vaccine administration (8), predict that MF59-adjuvanted vaccines will produce broader and more effective antiviral activity in vivo, which should translate into higher vaccine efficacy.

Materials and Methods

Plasma samples and human mAbs

Serum or plasma samples from two H5N1 human vaccine trials were tested; the NIAID-sponsored A/Vietnam/1203/2004 vaccine trial was conducted at Vanderbilt University, St. Louis University, and Cincinnati Children’s Hospital Medical Center (6), wherein the antigen and adjuvant were mixed at bedside before immunization. Full approval for this research was obtained from the Vanderbilt Medical Center Institutional Review Board. V87P2 trial (A/Vietnam/1194/2004 with and without MF59 in 40 adults) was conducted by Novartis Vaccine and Diagnostics (NVD), where the antigen and adjuvant were premixed and were available as prefilled syringes (12). All samples were deidentified. The protocols were evaluated by Center for Biologics Evaluation and Research–National Institutes of Health Research Involving Human Subjects Committee (RIHSC) and conducted under RIHSC exemption 03-118B.

Human mAbs FLA5.10, FLD21.140, and FLA3.14 were derived from two H5N1-exposed survivors in Vietnam and have been described (11, 15).

Construction of H5N1 gene-fragment phage display libraries

Complementary DNA corresponding to all eight gene segments of the A/H5N1/Vietnam/1203/2004 were generated from the RNA isolated from egg-grown virus strain and were henceforth used for cloning. fSK-9-3 is a gIIIp display-based phage vector where the desired polypeptide can be expressed as gIIIp fusion protein. The phage display libraries used in the current study expressed inserts spanning the HA and NA genes (referred to as HA-NA) as previously described (11).

Panning of H5N1-GFPDL with polyclonal human vaccine sera

For each round of panning, an equal volume of sera was used for each group. Before panning of GFPDL with plate-bound polyclonal human antibodies, serum components, which might nonspecifically interact with phage proteins, were removed by incubation with ultraviolet-killed M13K07 phage-coated petri dishes. Subsequent GFPDL selection was carried out in antibody-coated wells or in solution (with protein A/G) as previously described (11). Inserts of bound phages were polymerase chain reaction–amplified and sequenced.

Adsorption of polyclonal human sera on H5N1-GFPDL and measurements of residual binding to H5N1-Vietnam recombinant HA in ELISA

Five hundred microliters of 10-fold–diluted pooled serum antibodies from H5N1±MF59 vaccinees (NVD trial) were adsorbed by incubation with H5N1 (HA+NA) GFPDL-coated petri dishes. To ascertain the residual antibody specificity, we performed an ELISA with wells coated with 200 ng/100 μl of recombinant H5 HA (A/Vietnam/1203/2004, Protein Sciences). After blocking with phosphate-buffered saline (PBS) with Tween 20 containing 2% milk, serial dilutions of human vaccine sera (with or without adsorption) in blocking solution were added to each well and incubated for 1 hour at room temperature, followed by addition of 2000-fold–diluted horseradish peroxidase–conjugated goat-specific antibody against human IgG-Fc, and developed by 100 μl of o-phenylenediamine dihydrochloride substrate solution. Absorbance was measured at wavelength of 490 to 492 nm.

Affinity measurements by SPR

Steady-state equilibrium binding of mAbs FLA5.10, FLD21.140, and FLA3.14 and post-H5N1 vaccine sera was monitored at 25°C with a ProteOn SPR biosensor (Bio-Rad). The HA(28–319)-His6 was coupled to a GLC sensor chip with amine coupling with 500 resonance units (RU) in the test flow cells. Samples of 60 μl of freshly prepared antibody at 10 μg/ml were injected at a flow rate of 30 μl/min (contact time, 120 s). Flow was directed over a mock surface to which no protein was bound, followed by the HA(28–319)-His6–coupled surface. Responses from the peptide surface were corrected for the response from the mock surface and for responses from a separate buffer-only injection. The mAb 2D7 (antibody to CCR5) was used as a negative control antibody in the experiments. Binding kinetics for the mAbs and human vaccine sera and the data analysis were performed with Bio-Rad ProteOn Manager software (version 2.0.1). HA(28–319)–bound antibodies were isotyped by sequential injection of antibodies specific to human IgA, IgG, and IgM (Invitrogen).

Neutralizing antibody adsorption with HA proteins

Fivefold–diluted postvaccination sera (500 μl) were added to 0.5 mg of purified HA-His6 proteins (or shorter HA1-derived peptides) or to control glutathione S-transferase–His6 protein and incubated for 1 hour at room temperature. Nickel–nitrilotriacetic acid magnetic beads (200 μl) (Qiagen) were added for 20 min at room temperature on end-to-end shaker to capture the His-tagged proteins and the antibodies bound to them, followed by magnetic separation. Supernatants containing the unbound antibodies were collected. The preadsorbed and postadsorbed sera were subjected to virus MN assay and SPR.

Neutralization assay

Virus-neutralizing activity was analyzed in a MN assay on the basis of the methods of the pandemic influenza reference laboratories of the Centers for Disease Control and Prevention (CDC). Low-pathogenicity H5N1 viruses, generated by reverse genetics, were obtained from CDC: A/Vietnam/1203/2004 (SJCRH, clade 1), A/Indo/5/2005 (PR8-IBCDC-RG2, clade 2.1), and A/Anhui/1/05 (IBCDC-RG5, clade 2.3.4). The experiments were conducted with three replicates for each serum sample and performed at least twice.

CD-monitored equilibrium unfolding experiment

For CD spectroscopy in solution, HA(28–319) proteins were dissolved in 20 mM PBS (pH 7.4) at 0.5 mg/ml. The change in ellipticity at 222 nm during unfolding was monitored with a J-715 CD system (JASCO). The unfolding reaction was initiated by subjecting the protein in PBS to 0.5°C/min increments. The experiments were carried out in triplicate.

Supplementary Material

www.sciencetranslationalmedicine.org/cgi/content/full/2/15/15ra5/DC1

Fig. S1. The H5N1 A/Vietnam/1203/2004 HA+NA protein sequence.

Fig. S2. Adsorption of HA-specific antibodies in post-H5N1 vaccine sera using H5N1-GFPDLs.

Fig. S3. (A) Schematic of expression and purification of H5N1 HA(28–319). (B) CD melt spectroscopy shows that HA(28–319) purified at pH 7.2 (but not at pH 3.0) is properly folded.

Table S1. Frequency of selected phage clones using H5N1-GFPDLs after panning with sera from H5N1-Vietnam subunit vaccination in NIAID study.

Table S2. Frequency of selected phage clones using H5N1-GFPDLs after panning with sera from H5N1-Vietnam subunit vaccination in NVD study.

Footnotes

  • Citation: S. Khurana, W. Chearwae, F. Castellino, J. Manischewitz, L. R. King, A. Honorkiewicz, M. T. Rock, K. M. Edwards, G. Del Giudice, R. Rappuoli, H. Golding, Vaccines with MF59 Adjuvant Expand the Antibody Repertoire to Target Protective Sites of Pandemic Avian H5N1 Influenza Virus. Sci. Transl. Med. 2, 15ra5 (2010).

References and Notes

  1. Acknowledgments: We thank S. Epstein and M. Eichelberger for comments on early version of the manuscript and valuable discussions.Funding: U.S. Department of Health and Human Services Biomedical Advanced Research and Development Authority and Vanderbilt–NIH General Clinical Research Center grant RR00095. K.M.E. and M.T.R. are members of the Vaccine Trials and Evaluation Unit, NIH–National Institute of Allergy and Infectious Diseases (contract NO1-AI-25462).Author contributions: International Committee of Medical Journal Editors criteria for authorship read and met: S.K., W.C., F.C., J.M., L.R.K., A.H., M.T.R., K.M.E., G.D.G., R.R., and H.G. agree with the manuscript’s results and conclusions; S.K., W.C., A.H., F.C., M.T.R., K.M.E., and H.G. designed the experiments and conducted the study; S.K., F.C., and H.G. analyzed the data; J.M. and L.R.K. conducted the neutralization assays; S.K., H.G., K.M.E., G.D.G., and R.R. contributed to the writing of the manuscript.Competing interests: The authors have no competing interests.
View Abstract

Stay Connected to Science Translational Medicine

Navigate This Article