Influenza Virus H1N1pdm09 Infections in the Young and Old: Evidence of Greater Antibody Diversity and Affinity for the Hemagglutinin Globular Head Domain (HA1 Domain) in the Elderly than in Young Adults and Children

Abstract

The H1N1 2009 influenza virus (H1N1pdm09) pandemic had several unexpected features, including low morbidity and mortality in older populations. We performed in-depth evaluation of antibody responses generated following H1N1pdm09 infection of naïve ferrets and of 130 humans ranging from the very young (0 to 9 years old) to the very old (70 to 89 years old). In addition to hemagglutination inhibition (HI) titers, we used H1N1pdm09 whole-genome-fragment phage display libraries (GFPDL) to evaluate the antibody repertoires against internal genes, hemagglutinin (HA), and neuraminidase (NA) and also measured antibody affinity for antigenic domains within HA. GFPDL analyses of H1N1pdm09-infected ferrets demonstrated gradual development of antibody repertoires with a focus on M1 and HA1 by day 21 postinfection. In humans, H1N1pdm09 infection in the elderly (>70 years old) induced antibodies with broader epitope recognition in both the internal genes and the HA1 receptor binding domain (RBD) than for the younger age groups (0 to 69 years). Importantly, post-H1N1 infection serum antibodies from the elderly demonstrated substantially higher avidity for recombinant HA1 (rHA1) (but not HA2) than those from younger subjects (50% versus <22% 7 M urea resistance, respectively) and lower antibody dissociation rates using surface plasmon resonance. This is the first study in humans that provides evidence for a qualitatively superior antibody response in the elderly following H1N1pdm09 infection, indicative of recall of long-term memory B cells or long-lived plasma cells. These findings may help explain the age-related morbidity and mortality pattern observed during the H1N1pdm09 pandemic.

INTRODUCTION

The 2009 pandemic of swine origin influenza virus H1N1 (H1N1pdm09) exhibited an unusual pattern of age-related morbidity and mortality, as it disproportionately affected children and young adults (4). Compared with seasonal influenza outbreaks, in which >90% of deaths and over half of hospitalizations occur among those ≥65 years of age, only 13% of deaths and 10% of hospitalizations are estimated to have occurred in that age group (4, 8, 14, 22, 41). It was postulated that the lower attack rate and frequency of severe disease in the elderly reflected earlier exposure to 1918 H1N1-like viruses prior to 1940 and in 1957 and to the swine origin H1N1 (A/NJ/76) virus in 1976 or was simply due to repeated vaccinations against seasonal strains (13, 38, 39, 40). However, data supporting each of these possibilities were not fully conclusive (23, 32, 33, 36).

Influenza subtypes are classified based on the antigenic variation within influenza hemagglutinin (HA) as measured by a hemagglutination inhibition (HI) assay. The HI assay is dependent on the antibodies that inhibit the interaction between the sialic acid receptor on the red blood cells (RBC) and the receptor binding domain (RBD) within the HA1 domain of influenza virus hemagglutinin. Therefore, the antigenic differences within influenza viruses are primarily due to mutations within the HA1 domain, while the protein sequence within the HA2 stalk domain is highly conserved among multiple influenza virus subtypes. Human polyclonal responses against one subtype can show significant cross-reactivity to hemagglutinins of other subtypes due to this high sequence conservation in the HA2 domain, as previously shown. But this binding cross-reactivity does not translate into cross-protection, since most of the antibodies against the HA2 stalk do not block virus infectivity. Recently, rare antibodies with broad neutralizing cross-reactivity that target the HA2 stem were reported, but they are not easily elicited by traditional vaccination (5, 15, 37). In our previous studies, we demonstrated that most of the polyclonal-neutralizing-antibody responses following influenza virus infections or inactivated-subunit vaccination, as measured in HI or microneutralization (MN) assays, targeted the HA1 domain (16, 18, 19). Furthermore, HI titers did not reflect the entire spectrum of infection- or vaccination-induced antibody repertoires and their affinities, which are likely to contribute to influenza virus clearance in vivo. Therefore, it is important to use multiple analytical assays to evaluate the humoral immune response against different domains within the influenza virus hemagglutinin that evolves independently for HA2 and HA1 antigenic regions (18).

Our laboratory has developed molecular and analytical tools to probe the complete antibody repertoires against influenza virus and to measure the antibody kinetics of polyclonal serum binding to different antigenic domains within influenza virus hemagglutinin. To that end, we have applied whole-genome-fragment phage display libraries (FLU-GFPDL), surface plasmon resonance (SPR) technologies, and resistance of polyclonal serum IgG to 7 M urea treatment to better understand the antibody immune response following pandemic influenza virus infection or vaccination (16, 17, 18, 19). In toto, these assays capture the majority of in vivo circulating influenza virus-specific antibodies derived from both long-lived plasma cells and newly activated naïve and memory B cells, all of which contribute to the control of virus replication in vivo and determine clinical outcome.

In the current study, these technologies were used to elucidate the magnitude, epitope diversity, and affinity of polyclonal serum antibodies from naïve ferrets and from multiage human cohorts that were infected with H1N1pdm09 during the second wave of the influenza pandemic in 2009 (mid-November and early December). The samples were collected anonymously from extra laboratory specimens at the University of Pittsburgh Medical Center's (UPMC) Presbyterian Hospital and the Children's Hospital of Pittsburgh (30, 41).

Our findings provide evidence that elderly adults had antibody responses to H1N1pdm09 infection that were qualitatively superior to those elicited in younger adults and children. Specifically, elderly infected individuals (≥70 years old) had more diverse circulating antibodies against both the internal genes and the HA1 RBD. Importantly, the affinity of antibody binding to the HA1 domain of H1N1pdm09 was significantly higher for polyclonal sera of older adults and the elderly (>60 years) than for all the younger age groups.

MATERIALS AND METHODS

Infection of ferrets and blood collection.

The ferrets used in the study tested seronegative for circulating seasonal influenza A (H1N1 and H3N2) and influenza B viruses by HI. Animal experiments with influenza virus A/California/07/2009 were performed in the AALAC-accredited animal biosafety level 3 (ABSL-3) enhanced facility. Female Fitch ferrets (n = 6 in each group) were infected and monitored as previously described (20, 31). Briefly, the ferrets were anesthetized with isoflurane and infected intranasally with 1 × 10⁶ 50% egg infectious doses (EID₅₀) of A/California/07/2009. Blood was collected from the anesthetized ferrets via the anterior vena cava every other day. Serum was separated, aliquoted, and stored at −80 ± 5°C. All procedures were in accordance with the National Research Council (NRC) Guidelines for the Care and Use of Laboratory Animals, the Animal Welfare Act, and the Centers for Disease Control (CDC)/National Institutes of Health (NIH) Bio-Safety Guidelines in Microbiological and Biomedical Laboratories and approved by the Institutional Animal Care and Use Committee (IACUC).

Description of the samples obtained from infected human subjects.

The samples analyzed were excess serum samples collected anonymously at the University of Pittsburgh Medical Center's Presbyterian Hospital and the Children's Hospital of Pittsburgh in mid-November and early December 2009 (41). Pediatric samples were obtained from blood samples collected in outpatient clinics during the week of 16 November 2009. Adult (older than 20 years) samples were obtained from the clinical laboratories of the UPMC hospitals during the week of 23 November 2009. University of Pittsburgh Institutional Review Board (IRB approval [(exempt) PRO09110164]) was obtained. Blood samples were collected using the honest-broker system at the University of Pittsburgh Laboratories, organized by decade of birth without other identifying information, and given to investigators. No information is available in regard to other comorbidities for these individuals. Sera were obtained from hospital laboratories and tested to find those that had antibody to pandemic H1N1 during the second wave of pH1N1 infections in 2009 (30). None of the donors were vaccinated with pH1N1, since no pH1N1 vaccine was available in the city at the time of collection. However, it is anticipated that the vaccination history (i.e., seasonal trivalent inactivated vaccines [TIVs]) probably varied significantly among the different age groups. Each serum sample was classified by the decade of birth of the donor and tested in an HI assay against pandemic H1N1 (A/California/7/2009) (Table 1).

Table 1.

Postinfection HI titers against H1N1pdm09 in human subjects stratified by age^a

Group	Age (yr)	Total no. of samples	HI GMT (range)b
1	0–9	15	208 (80–640)
2	10–19	37	226 (80–640)
3	20–29	15	347 (80–640)
4	30–39	8	320 (80–640)
5	40–49	16	365 (80–640)
6	50–59	19	383 (80–640)
7	60–69	5	320 (80–640)
8	70–89	11	160 (80–320)

^aSamples were collected from H1N1pdm09-infected human subjects from November to December 2009 and confirmed by HI assay. No information is available in regard to other comorbidities for these individuals.

^bGMT, geometric mean titer.

Construction of H1N1 gene fragment phage display libraries and panning of H1N1pdm09 GFPDL with H1N1pdm09-infected polyclonal ferret and human sera.

The H1N1pdm09 phage display library affinity selection in the current study was performed as previously described (19). For GFPDL panning using ferret postinfection sera, equal volumes of sera collected on days 3, 5, 7, 14, and 21 postinfection (p.i.) from 6 ferrets were pooled. For GFPDL panning of postinfection human sera, equal volumes of sera from 5 individuals in group (gp) 1 (0 to 9 years old), 4 individuals in group 2 (10 to 19 years old), and 6 individuals in group 8 (70 to 89 years old) were pooled (Table 2).

Table 2.

Human serum samples for GFPDL analysis

Serum samples	Age	HI titer
Group 1 (0–9 yr old)	10 mo	80
	11 mo	80
	3 yr	640
	4 yr	160
	8 yr	640
Group 2 (10–19 yr old)	10 yr	640
	14 yr	80
	18 yr	160
	19 yr	160
Group 8 (70–89 yr old)	82 yr	320
	82 yr	160
	84 yr	160
	85 yr	80
	85 yr	160
	88 yr	320

Avidity measurements for 7 M urea-resistant antibodies.

IgG avidity was determined directly for each individual postinfection serum by a modified enzyme-linked immunosorbent assay (ELISA) method as previously described (19). The percentages of 7 M urea-resistant antibodies were calculated from the dose-response curves (50% effective concentration [EC₅₀]) of 7 M urea-treated compared with untreated samples.

Affinity measurements of polyclonal serum antibodies by surface plasmon resonance.

Steady-state equilibrium binding of post-H1N1pdm09 infection individual human sera was monitored at 25°C using a ProteOn surface plasmon resonance biosensor (Bio-Rad). Antibody dissociation constants, which describe the stability of the complex, i.e., the fraction of complexes that decay per second, were determined directly from the serum/plasma sample interaction with recombinant HA1 (rHA1) (1 to 330) and rHA2 (331 to 514) proteins using SPR in the dissociation phase (600 s) and calculated using the Bio-Rad ProteOn manager software for the heterogeneous sample model as described previously (19). The dissociation constants were determined from two independent SPR runs.

Statistical analyses.

Differences between groups were examined for statistical significance using Student's t test. An unadjusted P value of less than 0.05 was considered to be significant.

RESULTS

Epitope spreading following infection of naïve ferrets with wild-type H1N1pdm09 virus.

Since most humans in North America have been exposed to circulating influenza viruses from a very early age, it was important to initially identify a model of a truly naïve animal in order to follow the development of antibodies against different gene products after H1N1pdm09 infection. This was further complicated by the high degree of conservation in the internal genes between H1N1pdm09 and seasonal influenza virus. To that end, we first evaluated influenza-naïve ferrets after H1Npdm09 challenge. Sequential blood samples were collected from H1N1pdm09-infected ferrets on days 3, 5, 7, 14, and 21. For GFPDL analyses, the postinfection samples were pooled for each day (6 animals per group). Panning of GFPDL-expressing inserts from all the internal genes (FLU-6) and the HA/neuraminidase (NA) genes was conducted separately and is shown in Fig. 1. To determine the diversity of the epitopes, the eluted phage clones were sequenced after panning. Figure 1B and D show the graphical distribution of the representative clones obtained after affinity selection. Only clones with a frequency of ≥2 are shown. The total number of bound phage in the FLU-6 GFPDL increased from 164 on day 3 p.i. to 7,890 on day 21 p.i. (Fig. 1A). The numbers of clones reactive to different antigen fragments within influenza virus proteins were extrapolated from the proportions obtained after sequencing 192 bound phage clones for the FLU-6 library or 384 clones for the HA/NA GFPDL per day (Fig. 1A and C). During the first week after infection, the majority of inserts mapped to the polymerase genes, but on day 14 and day 21, the majority of bound clones contained epitopes that mapped predominantly to M1 and to a lesser extent to other genes (Fig. 1A and B). This shift in antibody focus and development of antibody response mirrored the expression of early versus late structural genes during viral replication. These findings suggested that the immune response in naïve ferrets to influenza virus proteins following virus challenge is delayed in relation to the kinetics of viral replication. It most likely depends on a threshold of viral-protein expression and on recruitment of key immune cells (antigen-presenting cells [APC], Th cells, and B cells) to the site of replication in order to initiate the adaptive immune responses. All ferrets cleared the infection between days 7 and 10 (20, 31). Binding of epitopes in HA/NA also evolved in a stepwise manner, with a big increase (>3-fold) taking place between day 5 and day 7 postinfection. The virus HI titers were ≤40 on days 3 and 5, while high HI titers were measured on days 7, 14, and 21, ranging between 640 and 2,560 for individual ferrets (Fig. 1C). Most of the epitopes bound by sera from days 3 and 5 postinfection mapped to HA2, including inserts spanning the C terminus of HA1 and the N terminus of HA2, but not to the RBD. In contrast, from day 7 onward, the antibody focus shifted to the HA1 globular domain, with a significant fraction of epitopes mapping to the RBD (red and yellow inserts, respectively). The HA1/HA2 ratio jumped from 0.04 on day 5 to 5.95 on day 7 (Fig. 1C and D). NA epitopes were also recognized by pooled sera from days 5 to 21 p.i., and the greatest diversity of inserts was seen on day 21 (Fig. 1C and D).

Fig 1.

Elucidation of antibody repertoires elicited following H1N1pdm09 infection in ferrets. (A) Distribution of phage clones after affinity selection using GFPDL expressing all internal proteins (FLU-6) of influenza virus A/California/07/2009 on sera obtained from H1N1pdm09-infected ferrets. (B) Schematic alignment of the peptides recognized by post-H1N1pdm09 infection ferret sera as identified with GFPDL (FLU-6). To determine the diversity of the epitopes, the eluted phage clones were sequenced after panning, and the peptide sequences displayed on the selected phage clones were aligned with the influenza virus proteins in the construction of FLU-6 GFPDL. The graphical distribution of the representative clones (192 clones per group) with a frequency of ≥2 obtained after affinity selection is shown. The colors of the bars representing phage-displayed peptides are labeled on the arrows above the graph. The horizontal positions and lengths of the bars indicate the peptide sequence displayed on the selected phage clone aligned with the corresponding influenza virus protein. The thickness of each bar represents the frequencies of repetitively isolated phage inserts. The numbers on the x axis depict the amino acid residues for the corresponding proteins used for alignment. (C) Distribution of phage clones after affinity selection using GFPDL expressing HA and NA proteins of influenza virus A/California/07/2009 on sera obtained from H1N1pdm09-infected ferrets. (D) Schematic alignment of the peptides recognized by post-H1N1pdm09 infection ferret sera as identified with GFPDL (HA and NA). The colors of the bars representing inserts are labeled on the arrows above the graph. The peptide sequences displayed on the selected phage clones were aligned with the influenza virus proteins in the construction of FLU-HA plus NA GFPDL. The graphical distribution of the representative clones (384 clones per group) with a frequency of ≥2 obtained after affinity selection is shown. The horizontal alignment and the thickness of the individual inserts are presented as in panel B.

GFPDL analyses of serum samples from H1N1pdm09-infected individuals during November-December 2009.

The various cohorts of H1N1pdm09-infected subjects used in this study were previously described (41). The HI titers postinfection were not statistically different among the various age groups (Table 1), in agreement with previous studies (36, 41). Furthermore, when available, prepandemic HI titers against A/California/07/2009 in samples collected in March 2009 were <10 for the different age groups analyzed in this study (30). Since we were interested in comparing the epitope diversity recognized by elderly subjects with that recognized by the most naïve populations, we initially pooled serum samples from three age groups, 0 to 9 years, 10 to 19 years, and 70 to 89 years, for GFPDL analyses. All the pooled sera had similar HI titers, ranging between 80 and 640 (Table 2). The exact dates of infections were not available; however, since all individuals were infected during the second wave of the pandemic, it was estimated that all serum samples were collected between 3 and 6 weeks postinfection. To determine the diversity of the epitopes, the eluted phage clones after panning were sequenced. Figure 2B and D graphically depict the distribution of the representative clones (192 clones per group) obtained after affinity selection using postinfection human sera. Only clones with a frequency of ≥2 are shown. The numbers of phage clones reactive to different antigenic fragments within influenza virus proteins were extrapolated from the proportions obtained from sequencing 192 bound clones (Fig. 2A and C). Based on the ferret results (Fig. 1A and B), it was expected that the majority of epitopes in the internal genes (GFPDL FLU-6) would be in the M1 protein. As can be seen in Fig. 2A and B, the pooled sera from the very young (0 to 9 years old; gp 1) were indeed focused on epitopes in the M1 gene. Slightly more epitope diversity was observed with pooled sera from the 10 to 19 year olds (gp 2), with a clear predominance of M1 epitopes (90% of total inserts) (Fig. 2A and B). Interestingly, in the elderly individuals (gp 8), a more diverse epitope repertoire (typical of a recall response) was found, especially in the polymerase complex and the structural (NP) gene (Fig. 2A and B).

Fig 2.

Elucidation of antibody repertoires elicited in children, teens, and the elderly following infection with H1N1pdm09. (A) Distribution of phage clones after affinity selection using A/California/07/2009 GFPDL FLU-6 on pooled sera obtained from gp 1 (0 to 9 years), gp 2 (10 to 19 years), and gp 8 (70 to 89 years) following infection with H1N1pdm09 (Table 2). (B) Schematic alignment of the peptides recognized by post-H1N1pdm09 infection human sera as identified with H1N1-Flu6 GFPDL. To determine the diversity of the epitopes, the eluted phage clones were sequenced after panning, and peptide sequences displayed on the selected phage clones were aligned with the influenza virus proteins in the construction of FLU-6 GFPDL. The graphical distribution of the representative clones with a frequency of ≥2 obtained after affinity selection are shown. The color code is identical to that in Fig. 1B. The horizontal positions and the lengths of the bars indicate the peptide sequence displayed on the selected phage clone aligned with the corresponding influenza virus protein. The numbers on the x axis depict the amino acid residues for the corresponding proteins used for alignment. (C) Distribution of phage clones after affinity selection using GFPDL expressing HA and NA proteins of influenza virus on the same pooled sera as in panel A. (D) Schematic alignment of the peptides recognized by post-H1N1pdm09 infection human sera as identified with GFPDL (HA and NA). The peptide sequences displayed on the selected phage clones were aligned with the influenza virus proteins in the construction of FLU-HA plus NA GFPDL (see Table S1 in the supplemental material). The horizontal alignment and the thickness of the individual inserts are presented as in Fig. 1D, and the color code is identical to that in Fig. 1D. (E) Numbers of unique sequences in the HA1 RBD recognized by pooled sera from the three age groups following GFPDL analysis (see Table S1).

In the HA/NA GFPDL analyses, all groups had similar numbers of phages bound, ranging between 3,200 and 4,700. In the very young (gp 1), the HA1/HA2 ratio was 0.4, while in children (gp 2) and the elderly (gp 8), the HA1/HA2 ratios were 1.3 and 0.9, respectively, suggesting a more balanced response. Importantly, pooled sera from the three age groups recognized an increasing number of unique sequences in the HA1 domain, and specifically within the RBD, which is the primary target of neutralizing antibodies (Fig. 2E; see Table S1 in the supplemental material).

The increased diversity of antibody repertoires with age suggested that the antibody response in the elderly reflects activation of long-term memory B cells and/or long-lived plasma cells compared with mainly naïve B cells in the younger age groups, which resembled the antibody epitope repertoire in the ferrets 14 to 21 days postinfection (Fig. 1).

Evidence for high-affinity anti-HA1 antibodies in older adults infected with H1N1pdm09 as measured by 7 M urea resistance and SPR-based real-time kinetics.

To further investigate the effect of age on the quality of antibody responses, we extended our analysis to ELISA and SPR to measure the antibody affinity of the polyclonal human serum antibodies following H1N1pdm09 infection. In ELISA, as in other equilibrium-based assays (e.g., HI and MN), it is not feasible to discriminate between the contributions of antibody affinity and the antibody concentration to the binding titer. It is possible, however, to approximate antibody avidity by measuring the effect of denaturants on antibody binding, since low-avidity antibodies are more rapidly eluted (25, 28, 29).

To that end, individual postinfection serum samples from all age groups (including those analyzed by GFPDL) (Table 1) were examined by ELISA with recombinant H1N1 HA1 (1 to 330) or HA2 (331 to 514) (19, 20). To evaluate the relative avidity of bound serum IgG, the antibody-antigen complexes were briefly exposed to 7 M urea prior to addition of the secondary labeled anti-human IgG antibody, as previously described (19). The percentages of 7 M urea-resistant IgG were calculated from the dose-response curves (EC₅₀) of 7 M urea-treated compared with untreated samples. As can be seen in Fig. 3A, the fractions of HA1-bound 7 M urea-resistant antibodies were relatively low for younger age groups (gp 1 to 6), ranging between 17% and 22%. In contrast, in the older age groups (gp 7 and 8), significantly higher percentages of 7 M urea-resistant IgG were bound to rHA1 (46% and 49%, respectively). On the other hand, in the case of rHA2, no clear age difference was observed, with the percentages of 7 M urea-resistant antibodies ranging between 56% and 76% for all groups (i.e., significant proportions of high-affinity antibodies) (Fig. 3B).

Fig 3.

IgG affinities of individual sera for HA1 and HA2 domains in hemagglutinin following infection by H1N1pdm09 in humans. (A and B) Postinfection serum IgG binding to bacterially expressed H1N1 rHA1 (A) and rHA2 (B) was analyzed by comparing the EC₅₀s of 3-fold serial serum dilutions in the absence or presence of 7 M urea for various age group sera collected and stratified by decade from 1920 to 2000. The percentages of 7 M urea-resistant binding of human sera from individuals >60 years old to rHA1 were statistically higher than for all younger groups (P < 0.02). The data shown are means ± standard deviations (SD) for each serum from three independent experiments. (C and D) Antibody affinity as measured by antibody dissociation rates for rHA1 and rHA2 in different age groups. SPR analysis of post-H1N1pdm09 infection human sera from various age groups was performed with functional H1N1 rHA1 (1 to 330) and rHA2 (331 to 514) (20). Serum antibody off rate constants were determined as described previously (19) and in Materials and Methods. The correlation statistics of the anti-HA1 antibody off rate constants of the postinfection human sera between different vaccine groups were highly significant for the 0- to 20-year-old versus 50- to 89-year old age groups (P < 0.05 [t test]) but not significant between any of the age groups for anti-HA2 binding antibodies.

To further corroborate the 7 M urea ELISA data regarding polyclonal antibody affinity for HA1 in the various age groups, the postinfection plasma samples were subjected to SPR-based real-time kinetics measurements of polyclonal antibody off rates against H1N1 rHA1 (1 to 330) and HA2 (331 to 514) for each individual serum, as recently described (19). As can be seen in Fig. 3C, the average anti-HA1 polyclonal serum antibody dissociation off rates for groups 6 to 8 (50 to 59, 60 to 69, and 80 to 89 years old) were significantly lower than for the younger age groups, which reflects higher-affinity antibody binding to the HA1 domain for the older age groups. In the case of HA2, the polyclonal antibody off rates were low (indicating high-affinity binding), but no significant differences among the age groups were identified (Fig. 3D).

Altogether, our study demonstrates for the first time in humans that the influenza virus-specific antibody epitope repertoire and affinity of binding to the HA1 globular domain following H1N1pdm09 infections in the elderly were superior to those in younger adults and children, most likely reflecting contributions of recall memory B-cell responses.

DISCUSSION

Our findings provide new insight into the antibody responses generated following H1N1pdm09 infection in the older populations that was not previously appreciated using the traditional HI and MN assays. Following H1N1pdm09 infection, the antibody repertoires against the polymerases and structural genes, as well as the HA1 RBD, which is targeted by most neutralizing antibodies that inhibit influenza virus entry, were broader in elderly adults than in all younger age groups. The use of 7 M urea ELISA provided evidence of higher-avidity IgG binding to HA1 in the older age groups (60 to 89 years) than in the younger age groups (0 to 59 years). In the case of HA2, the polyclonal antibody avidity to this conserved domain was high, with no significant differences among the age groups. Adaptation of SPR technology to follow the real-time kinetics of polyclonal antibody binding to properly folded HA domains and the calculation of antibody-antigen dissociation rates (i.e., direct measurement of antibody affinity) for the heterogeneous sample model (19) has multiple advantages (including conformational epitopes and quaternary structures) over ELISA-based assays in which only a single time point is measured. The SPR analyses demonstrated significantly higher-affinity antibody binding to HA1 in the older age groups (including the 50- to 89-year-old group) than in young adults, children, and toddlers (0 to 49 years).

The affinity of serum antibodies is likely to play a key role in vivo, especially very early after infection, due to unfavorable antibody/viral load ratios. During the H1N1 pandemic, low-affinity antibodies in some infected individuals were associated with more severe disease (25) and lung pathology.

In a previous publication, Wrammert et al. conducted an evaluation of plasmablast-derived monoclonal antibodies (MAbs) isolated from H1N1pdm2009-infected individuals. They reported the finding of antibodies that had undergone somatic mutations, and some of the isolated MAbs demonstrated broad cross-reactivity. Of those, several mapped to the HA2 stem domain and others to the HA1 globular domain (37). Importantly, the above-mentioned study was conducted with MAbs derived from only 4 individuals ranging in age between 21 and 30 years. While such analysis provided important information regarding the variable regions of immunoglobulin heavy and light chains (V_H-V_L) usage and number of mutations from the presumed germ line genes, they cannot be easily extrapolated to the complete antibody repertoires in the sera/plasma of infected individuals. In order to achieve the power to compare multiple age groups and draw meaningful conclusions about antibody diversity and avidity, a very large number of plasmablast-derived MAbs must be established to represent the diversity observed in human sera/plasma. There is also the possibility of bias during B-cell selection and cloning. Therefore, the assays conducted in our study with polyclonal sera from infected individuals provide important unbiased information on the complete repertoire and avidity of the polyclonal antibodies that were elicited following H1N1pdm09 infection of old versus young individuals in vivo. The infection outcome is more likely to correlate with the rate and quality of the polyclonal antibodies and their ability to restrict virus replication to the upper respiratory tract (URT), followed by viral clearance in vivo.

Our results support the hypothesis that older individuals have long-term memory B cells, and possibly long-lived plasma cells with B-cell receptors that cross-react with the H1N1pdm09 (A/California/07/2009) HA1 and were rapidly recruited and activated following H1N1pdm09 infection (6, 7, 34, 35).

Antibody affinity maturation following antigenic exposure has been linked to somatic hypermutation in B cells, which is dependent on induction of the activation-induced cytidine deaminase (AID) enzyme (26, 27). However, age-related immune senescence includes a loss of AID expression in B cells due to impairment of the transcription factor E47 (11). Furthermore, the drop in AID expression in the elderly (≥60 years old) was shown to correlate with lower responses to annual vaccination with seasonal influenza virus vaccines (3, 10). Our findings of high-affinity antibodies against H1N1pdm09 HA1 in the elderly strongly suggest that these antibodies are the product of long-term memory B cells or long-lived plasma cells that have undergone somatic hypermutation during early exposure to H1N1 viruses that circulated in the United States until 1957-1958 (1, 2, 9, 21, 33, 39). It is also possible that exposure to or vaccination with the 1976 swine flu generated long-term memory B cells with cross-reactivity with H1N1pdm09 (24, 38). Indeed, in a previous study with the same serum, we found that 48% and 59% of the samples from individuals aged 70 to 79 years and 80 to 89 years, respectively, had HI titers against the 1918 pandemic H1N1 strain (30). Furthermore, 45 to 58% of samples from the 50- to 79-year-old group had HI titers against A/Denver/1/1957. In a separate study, we demonstrated that elicitation of anti-1918 immunity in young mice prevented morbidity and led to lower levels of lung infection by 2009 pandemic H1N1 in aged mice (12).

In agreement with the current findings in infected individuals, we recently evaluated the antibody responses of older and younger adults following inactivated H1N1pdm09 vaccination. We found that H1N1pdm09 vaccination induced 10-fold-higher antibody levels in elderly than in younger adults, including in subjects with no prevaccination HI titers. Importantly, post-H1N1 vaccination sera from the elderly subjects demonstrated substantially higher avidity than sera from younger subjects (>60% versus <30% 7 M urea resistance, respectively) and lower antibody dissociation rates using SPR (18). Thus, similar observations were made in postvaccination and postinfection scenarios.

Our findings of greater diversity of epitope specificity and higher affinity of antibodies to HA1 in older individuals post-H1N1pdm09 exposure provide an additional explanation for the unusually low rate of severe respiratory disease during the 2009 H1N1 pandemic in this age group that is usually most susceptible to morbidity and mortality due to seasonal influenza (22).