Antibody-Mediated Suppression Regulates the Humoral Immune Response to Influenza Vaccination in Humans

Abstract

Background

Preexisting immunity, including memory B cells and preexisting antibodies, can modulate antibody responses to influenza in vivo to antigenically related antigens. We investigated whether preexisting hemagglutination inhibition (HAI) antibodies targeting the K163 epitope on the hemagglutinin (K163 antibodies) could affect antibody responses following vaccination with A/California/07/2009-like A(H1N1)pdm09 influenza viruses in humans.

Methods

Pre- and postvaccination sera collected from 300 adults (birth years, 1961–1998) in 6 seasons (2010–2016) were analyzed by HAI assays with 2 reverse genetics viruses and A(H1N1) viruses circulated from 1977 to 2018. Antibody adsorption assays were used to verify the preexisting K163 antibody–mediated suppression effect.

Results

Preexisting K163 antibody titers ≥80 affected HAI antibody responses following influenza vaccination containing A/California/07/2009-like antigens. At high K163 antibody concentrations (HAI antibody titers ≥160), all HAI antibody responses were suppressed. However, at moderate K163 antibody concentrations (HAI antibody titer, 80), only K163 epitope–specific antibody responses were suppressed, and novel HAI antibody responses targeting the non-K163 epitopes were induced by vaccination. Novel antibodies targeting non-K163 epitopes cross-reacted with newly emerging A(H1N1)pdm09 strains with a K163Q mutation rather than historic 1977–2007 A(H1N1) viruses.

Conclusions

K163 antibody–mediated suppression shapes antibody responses to A(H1N1)pdm09 vaccination. Understanding how preexisting antibodies suppress and redirect vaccine-induced antibody responses is of great importance to improve vaccine effectiveness.

(See the Editorial Commentary by Ji and Guthmiller on pages 299–302.)

During the 2009 pandemic, novel A(H1N1)pdm09 influenza A viruses (IAVs) infected children and adults who lacked preexisting cross-reactive neutralizing antibodies [1]. An estimated 65% of 2009 A(H1N1)pdm09 deaths worldwide occurred among people aged 18 to 64 years [2]. Since 2009, A(H1N1)pdm09 has continued to circulate as seasonal IAVs. From 2010–2011 through 2016–2017, the H1 component of influenza vaccine remained as A/California/07/2009 (CA/09)–like virus, one of the earliest isolates of the 2009 pandemic [3].

The 2013–2014 influenza season was the first A(H1N1)pdm09-predominant season in the United States since the 2009 pandemic, and persons aged 18 to 64 years accounted for 57% of reported hospitalizations [4]. The A(H1N1)pdm09 viruses isolated since 2013 possessed a mutation (K163Q, H1 numbering) in the Sa site of hemagglutinin (HA) [5]. Low vaccine effectiveness against the drifted A(H1N1)pdm09 field strains with the K163Q mutation (K163Q variants) was observed among middle-aged adults immunized with CA/09-like vaccines [3]. Additionally, some adults who were likely primed with A/USSR/90/1977 (USSR/77)–like A(H1N1) viruses possessed the dominant K163 antibodies [3, 6] and therefore were more susceptible to infection with K163Q variants [7]. The K163 antibody–dominant human sera failed to neutralize a reverse genetics (RG) virus with a single K163Q mutation (RG-K163Q) and the drifted K163Q variants [5, 6]. Moreover, K163-specific human monoclonal antibodies (K163-hmAbs) failed to protect mice against lethal challenge with a RG-K163Q virus [5]. Subsequently, the H1 vaccine component was updated to an A/Michigan/45/2015 (MI/15)–like virus in 2017 [3].

The reemergence of USSR/77-like viruses in 1977, after being absent since 1957, has been well documented. These viruses were antigenically and genetically similar to viruses that had circulated in 1950s, and most individuals were first infected with USSR/77-like viruses between 1977 and 1983 (age, ≤25 years in 1977; born after 1952) [8–11]. Only 4 USSR/77-like viruses were isolated in the United States during 1984 to 1986 [11]. A(H1N1) viruses that emerged in 1986 had major differences in their Sa and Sb antigenic sites from USSR/77-like viruses, including a double mutation (K125N and N127T) with the addition of a glycosylation motif in the Sa antigenic site, which likely shielded the K163 epitope [5, 12].

The specificity of K163 antibodies has been well characterized with K163-hmAbs and K163 antibody–dominant adult sera [5, 6, 13–15]. The K163 antibodies efficiently neutralized the 2009–2012 CA/09-like viruses and historical 1977–1983 USSR/77-like viruses but failed to neutralize RG-K163Q, 2013–2015 K163Q variants, and historical 1986–2007 A(H1N1) viruses possessing a glycosylation motif at HA-N125.

Antigenic imprinting by the first encounter with USSR/77-like viruses affected subsequent antibody responses to CA/09-like pandemic viruses in some middle-aged adults [3, 5, 6], a process described as the “original antigenic sin” (OAS) by Thomas Francis [16]. The first encounter with CA/09-like viruses in 2009 preferentially activated K163-specific memory B cells (MBCs) and resulted in dominant K163 antibody responses in some adults primed with USSR/77 [1, 17, 18]. Immune pressure from preexisting narrowly focused K163 antibodies in adults infected with CA/09-like viruses may lead to the emergence of K163Q variants by antibody-mediated selection and manifest severe disease [18, 19]. It has been reported that some middle-aged patients with lung failure possessed high levels of focused K163 antibodies at the time of the intensive care unit admission during the 2009 pandemic [18]. K163Q variants first appeared in 2010 to 2011 and later became epidemic during and after the 2013–2014 influenza season [15]. Additionally, rates of A(H1N1)pdm09 deaths increased among middle-aged patients infected with K163Q variants in the winter of 2013 to 2014 [20, 21]. Thus, A(H1N1)pdm09 viruses may be able to use this aspect of the “OAS antibody” to select K163Q variants that evade the host's immune system, which can be harmful at individual and population levels.

Our previous studies showed that ∼26% of middle-aged adults, who were likely primed with USSR/77-like viruses, mounted dominant (≥75%) K163 antibodies following immunization with CA/09-like vaccines during 2010 to 2016 [6]. Some adults also possessed substantial levels of focused K163 antibodies in their prevaccination sera. Antibody-mediated suppression of responses to specific antigens in vivo has been known for >100 years [22–25], and some studies have shown that epitope masking by preexisting antibodies may play an important role in limiting the boosting of anti–HA stalk antibody responses [26, 27]; yet, it is still largely unknown how preexisting immunity can influence the neutralizing antibody responses following influenza vaccination.

In this study, we generated experimental data to explore whether preexisting K163 antibodies affect the boosting of hemagglutination inhibition (HAI) antibody responses following immunization with vaccine containing CA/09-like antigens. We found antibody-mediated suppression to be key in recapitulating HAI antibody responses. Epitope masking could break the OAS in some middle-aged adults possessing focused preexisting K163 antibodies.

METHODS

Serum Samples

Three hundred adults were immunized with egg-derived inactivated trivalent influenza vaccines containing CA/09-like antigen during 2010 to 2016. Paired sera were collected at prevaccination and 21 to 28 days postvaccination from 6 seasons. The use of these anonymized sera was approved by the US Centers for Disease Control and Prevention human subject research determination review.

Influenza Viruses

IAVs were propagated in embryonated eggs or Madin-Darby canine kidney cells. Some viruses were concentrated and purified on a linear sucrose gradient as previously described [28]. We used X-179A vaccine strain, K163Q variants (egg- and cell-propagated 6B, 6B.1, and 6B.2 field strains), and 10 historical A(H1N1) viruses representing all major antigenic clusters circulated during 1977 to 2008. Two RG viruses were generated with the HA and neuraminidase genes from CA/09-like virus and 6 internal genes from A/Puerto Rico/8/1934, including RG-K163 possessing wild type CA/09 HA and RG-K163Q possessing CA/09 HA with a single K163Q mutation. All 19 viruses are described in Table 1; their HA genes were sequenced and published previously [6].

HAI Assay

Sera were treated with receptor-destroying enzyme to remove nonspecific inhibitors and adsorbed with packed turkey red blood cells to remove nonspecific agglutinins prior to testing with 4 HA units of virus and 0.5% turkey red blood cells [29].

Antibody Adsorption Assay

Serum was mixed with ∼10⁵ HA units of purified virus or a phosphate-buffered saline control. After incubation for 2 hours at 4 °C, the virus-serum mixture was centrifuged for 45 minutes at 100 000g to remove virus-antibody complexes and most of the unbound viruses. Residual viruses were removed by the addition of 100 µL of packed turkey red blood cells [6].

Statistical Analyses

Paired t tests were used to compare geometric mean titers (GMTs). P < .05 was considered statistically significant.

RESULTS

Preexisting K163 Antibody–Mediated Suppression Is Dose Dependent

IgG antibodies targeting the K163 epitope were induced in some adults primed with USSR/77 following infection or vaccination with CA/09-like viruses during 2009 to 2016 [13-15, 18]. IgG can suppress antibody responses against the antigens that they recognize in vivo [22–25]. IgG-mediated suppression is dose dependent: high doses of IgG suppress more effectively than low doses [24, 25]. However, whether preexisting anti–HA head IgG can suppress influenza vaccine–induced HAI antibody responses is unknown.

To explore whether preexisting K163 antibodies could decrease the magnitudes of boosting CA/09-like vaccine-induced HAI antibody responses, we measured antibody titers in pre- and postvaccination sera collected from 300 adults with 3 viruses: X-179A (vaccine strain), RG-K163, and RG-K163Q. The participants were immunized with trivalent influenza vaccines containing CA/09-like virus in the fall of 2010 to 2016. These adults were born between 1961 and 1998 (age range, 18–49 years) in the era during which A(H2N2), A(H3N2), and/or A(H1N1) IAVs were prevalent (Figure 1A and 1B) [6]. Dominant K163 antibodies in sera prevaccination and/or 21 to 28 days postvaccination are defined by a ≥4-fold reduction in HAI antibody titers against RG-K163Q vs RG-K163.

We identified 11 adults out of 300 for more detailed analysis. As shown in Table 2 and Figure 1B, 3 of 300 adults (1%, Nos. 1–3) had narrowly focused K163 antibodies in pre- and postvaccination sera; 4 adults (1.3%, Nos. 4–7) possessed narrowly focused K163 antibodies in prevaccination sera but not postvaccination sera. Of 300 adults, 26 (8.7%) displayed vaccine-induced focused K163 antibody responses in their postvaccination sera (Supplementary Table 1) [6], and 4 of these 26 adults (Nos. 8–11) possessed extremely focused postvaccination K163 antibodies (HAI antibody titers, 160–640) and very low or even undetectable prevaccination HAI antibodies (titers ≤10).

To determine whether preexisting K163 antibody–mediated suppression is dose dependent, we categorized the 11 adults into 3 groups based on their prevaccination HAI antibody titers against RG-K163 at 160 to 320 (group 1 [G1], Nos. 1–3), 80 (group 2 [G2], Nos. 4–7), or ≤10 (group 3 [G3], Nos. 8–11). We defined the magnitudes of boost as fold increases of HAI antibody titers from pre- to postvaccination. Different levels of fold increases of HAI antibody titers against RG-K163 and X-179A were observed in donors from G1 (1- to 2-fold), G2 (2- to 8-fold), and G3 (32- to 128-fold; Table 2). A clear negative correlation between prevaccination K163 antibody titers and fold increases of HAI antibody response against RG-K163 and X-179A was observed among the 3 groups (Figure 1C). We further compared fold increases of HAI antibody titers against RG-K163 and X-179A in G1 and G2 with all 26 adults (26/300) possessing the focused vaccine-induced K163 antibody responses but with prevaccination K163 antibody titers <80. Significantly lower fold increases of HAI antibody responses against RG-K163 and X-179A were observed in G1 (GMT >30-fold, P < .01) and G2 (GMT >6-fold, P < .01) as compared with the 26 controls (Supplementary Table 1).

These data indicated that preexisting K163 antibody titers ≥80 can effectively suppress vaccine-induced HAI antibody responses. Moreover, the K163 antibody–mediated suppression is dose dependent: high doses suppress more effectively than low doses.

Preexisting Antibody-Mediated Epitope- and Nonepitope-Specific Suppression

IgG-mediated suppression is antigen specific. Whether IgG suppresses antibody responses only to the epitope to which it binds (epitope-specific suppression) or to other epitopes on the same antigen (nonepitope-specific suppression) is an important question [22–25]. HA is a homotrimeric integral membrane glycoprotein and possesses 3 K163 epitopes near the receptor binding site on the CA/09 HA head [5, 15]. Infection or vaccination with CA/09-like viruses can induce K163 antibodies and HAI antibodies targeting ≥1 other epitopes in some adults [6, 13–15, 18]. Therefore, it could be an ideal system to study preexisting antibody-mediated epitope- and nonepitope-specific suppression effects for CA/09-like vaccine-induced antibody responses.

To study whether preexisting K163 antibodies could mediate epitope- or nonepitope-specific suppression, we tracked different CA/09-like vaccine-induced HAI antibody populations in the 11 adults (Table 3). In HAI assays, we used X-179A (containing an Q223R egg-adapted mutation), RG-K163, RG-K163Q, and 10 historical 1977–2007 A(H1N1) viruses (which caused almost all A[H1N1] epidemics/outbreaks between 1977 and 2008; Table 1). All vaccine-induced antibodies, including those targeting the egg-adapted R223 epitope (R223 antibody), were detected by X-179A (K163 + Q223R). All vaccine-induced HAI antibodies, except the R223 antibodies, were detected by RG-K163 (K163 + Q223), while the vaccine-induced HAI antibodies targeting epitopes (non-K163 + non-R223) were detected by RG-K163Q (K163Q + Q223). Because the K163 antibodies were produced from the MBCs that derived from the first exposure with USSR/77-like viruses [5, 6, 13, 18], we tracked the K163 antibodies with 3 USSR/77-like viruses: USSR/77, ENG/80, and CH/83. Other cross-reactive HAI antibodies were tracked with 1986–2007 A(H1N1) viruses.

As shown in Table 3, vaccination in the G3 adults (prevaccination K163 antibody titer ≤10, Nos. 8–11) resulted in a 32- to 128-fold increase in antibodies to X-179A and RG-K163, a 4- to 32-fold increase to 1977–1983 USSR/77-like viruses, and a ≤2-fold increase to RG-K163Q and 1986–2007 A(H1N1) viruses, suggesting that CA/09-like vaccine induced extremely focused K163 antibody responses. However, vaccination in the G2 adults (prevaccination K163 antibody titer, 80; Nos. 4–7) resulted in a 2- to 8-fold HAI antibody increase to X-179A and RG-K163, a 1- to 4-fold increase to 3 USSR/77-like viruses, and a ≤2-fold increase to 1986–2007 A(H1N1) viruses; surprisingly, vaccination induced a 32- to 64-fold antibody increase to RG-K163Q. As shown in Figure 2, fold increase of HAI antibodies against 3 USSR/77-like viruses in the G2 adults (GMT, 2-fold) were significantly lower than those in the G3 adults (GMT, 9-fold; P < .01). In addition, G2 adults displayed significantly higher HAI antibody increases against RG-K163Q (GMT, 54-fold) as compared with RG-K163 (GMT, 5-fold; P < .01). These data indicate that preexisting K163 antibodies in the G2 adults mainly suppressed K163 antibody responses but did not suppress HAI antibody responses targeting other non-K163 epitopes. Interestingly, these new non-K163 antibodies were not cross-reactive with all historical 1977–2007 A(H1N1) viruses. In contrast, vaccination did not result in any significant antibody increase (≤2-fold) against all 13 testing viruses in the G1 adults (Nos. 1–3) containing high levels of preexisting K163 antibodies (titers, 160–320), suggesting that such high levels of preexisting K163 antibodies could cause complete suppression of all HAI antibody responses targeting K163 and non-K163 epitopes.

Taken together, our data indicate that high levels of preexisting K163 antibodies (titers ≥160) could result in nonepitope-specific suppression, whereas the moderate levels of preexisting K163 antibodies (titer, 80) could cause epitope-specific suppression.

HAI Antibodies Targeting Non-K163 Epitopes Cross-react With Drifted A(H1N1)pdm09 Field Strains Possessing K163Q Mutation

To further study the specificity of CA/09 vaccine-induced novel non-K163 antibodies in the G2 adults (Nos. 4–7) who possessed narrowly focused K163 antibodies in prevaccination sera but not postvaccination sera, we performed HAI assays with 2013–2015 K163Q variants, including A/Bolivia/559/2013 (BO/13, 6B), MI/15 (6B.1), and A/Iowa/53/2015 (IW/15, 6B.2) viruses, which caused A(H1N1)pdm09 epidemics/outbreaks between 2013 and 2018 (Tables 1 and 4). The K163 antibodies in the adults from G1 (pre- and postvaccination sera), G2 (prevaccination sera), and G3 (postvaccination sera) displayed no cross-reactivity with K163Q variants, as expected. However, non-K163 antibodies in G2 donors (postvaccination sera) cross-reacted well with all K163Q variants.

A(H1N1)pdm09 Vaccine Preferentially Induced Strain-Specific HAI Antibody Responses in Adults Possessing Moderate Levels of Preexisting K163 Antibodies

Influenza vaccine can engage preexisting MBCs to produce cross-reactive antibodies targeting shared epitopes and naive B cells that induce de novo B-cell response to produce strain-specific antibodies targeting new epitopes [30–33].

We further explored the HAI antibody composition in the G2 adults (Nos. 4–7) with antibody adsorption. As shown in Figure 3 and Supplementary Table 2, the K163 antibodies in prevaccination sera can be completely removed by RG-K163 and X-179A (K163) but not RG-K163Q and MI/15 (K163Q). Interestingly, preexisting K163 antibodies in participants Nos. 4 and 5 could be completely removed by USSR/77, CH/83, TW/86, and NC/99; however, preexisting K163 antibodies in Nos. 6 and 7 could be removed only by USSR/77, CH/83, and TW/86, not NC/99. As expected, non-K163 antibodies in postvaccination sera could not be removed by all historical A(H1N1) viruses, but they could be removed by the A(H1N1)pdm09 viruses with or without K163Q mutation, including 6B.1 and 6B.2 field strains.

Taken together, our data indicate that CA/09-like vaccine induced A(H1N1)pdm09 strain-specific HAI antibody responses (likely naive B-cell origin) targeting new epitopes, which overcame the OAS and resulted in diversification of the humoral immune response in G2 adults. Importantly, these non-K163 antibodies bound very well to the K163Q variants that prevailed during and after 2013, suggesting a positive effect of K163 antibody–mediated suppression.

DISCUSSION

Understanding the mechanisms of how preexisting immunity, including MBCs and preexisting antibodies, modulate antibody responses to subsequent infections and vaccinations is particularly important for constantly evolving influenza viruses [34–38]. Here, we explored how preexisting K163 antibodies suppressed and redirected the generation of HAI antibody responses following immunization with CA/09-like vaccines in adults primed with USSR/77-like viruses. Our data indicate that different amounts of preexisting K163 antibodies could cause dose-dependent suppression of the antibody responses following vaccination containing CA/09 (Figure 1), and this suppression can be epitope or nonepitope specific depending on the levels of preexisting K163-specific antibodies (Figure 2, Tables 2 and 3).

The moderate level of preexisting K163 antibodies in G2 adults suppressed only K163 antibody responses, not antibody responses targeting the non-K163 epitopes (Tables 2 and 3, Figure 2, Supplementary Table 1). Such epitope-specific suppression is compatible with the epitope-masking hypothesis, with no involvement of the IgG Fc portion [39–42]. According to this model, preexisting K163 antibodies masked the K163 epitope, which prevented K163-specific MBCs from binding to it and resulted in a suppressed K163-specific antibody response. However, preexisting K163 antibodies did not mask other distinct HAI epitopes so that the antibody responses to these other epitopes could be efficiently induced. Such epitope-specific antibody-mediated suppression via epitope masking has been evidenced for multiple antigens, including HIV envelope glycoproteins, SARS-CoV-2 mRNA vaccine, and malaria vaccine [24, 43–45].

By contrast, in G1 adults, the high levels of preexisting K163 antibody suppressed all HAI antibody responses (Tables 2 and 3, Figure 2, Supplementary Table 1). Such nonepitope-specific suppression is not just due to epitope masking [24, 25]. The precise mechanisms behind such antibody-mediated nonepitope-specific suppression are still largely unknown, although several hypotheses have been proposed. This mode of suppression has been shown to be Fc dependent, implicating increased clearance of IgG-antigen complexes by FcrR⁺ phagocytic cells, complement-mediated lysis of entire antigens before they can stimulate an immune response, or negative regulation of B cells through the inhibitory FcrRIIB [22–25]. The high levels of preexisting K163 antibody in the G1 adults may form immune complexes with multiple viruses, which may cause rapid antigen clearance and prevent the activation of MBCs and naive B cells.

Recalled MBC-derived K163-hmAbs displayed variable binding affinities against CA/09-like viruses and priming USSR/77-like viruses [13–15]. The 4 G2 adults had equal levels of preexisting K163 antibodies (Table 3). However, the preexisting K163 antibodies in participants Nos. 6 and 7 had higher HAI activities against the priming USSR/77-like viruses and narrower binding breadth against the historical A(H1N1) viruses than Nos. 4 and 5 (Figure 3A, Supplementary Table 2). The binding affinities of hmAbs against CA/09 virus could be inversely correlated with the HAI breadth [17], suggesting that preexisting K163 antibodies in Nos. 6 and 7 might have higher binding affinities than those in Nos. 4 and 5. While the preexisting K163 antibodies in Nos. 6 and 7 exhibited more effective suppressive ability than Nos. 4 and 5 (Figure 3B), such different suppressive capacity may be associated with different binding affinities of the preexisting K163 antibodies. High-affinity IgG suppresses more efficiently than low-affinity IgG [22–25].

Novel non-K163 antibodies in G2 adults did not cross-react with historical 1977–2007 A(H1N1) viruses but did cross-react with K163Q variants, suggesting that these non-K163 antibodies were A(H1N1)pdm09 strain specific (Tables 3 and 4). Analysis of CA/09-like virus-activated B cells [17, 46] and the virus-induced serum HAI antibody responses (our unpublished data) demonstrated that the first viral exposure in 2009 mainly induced memory-driven cross-reactive HAI antibody responses in most adults but also detectable strain-specific HAI antibody responses in some adults with no preexisting anti-CA/09 HAI antibodies. Reexposure to the same vaccine mainly generated strain-specific B-cell responses in the adults possessing the preexisting antibodies [46]. However, the quality and quantity of antibodies required to limit B-cell activation against particular epitopes are still largely unknown [37]. Here, we found that moderate levels of preexisting K163 antibody can limit the boost of K163 epitope–specific antibody responses. In addition, high levels of preexisting K163 antibody can completely limit HAI⁺ MBCs and naive B-cell activation. Furthermore, not only can antibody-mediated suppression via epitope masking limit the K163 epitope–specific antibody response, it can also drive the diversification of the humoral responses by activating CA/09-specifc MBCs and/or naive B cells to produce strain-specific antibodies. These non-K163 antibodies may have reduced CA/09-like vaccine failure during 2013 to 2017, when K163Q variants circulated in humans. A similar effect has been described in SARS-CoV-2 mRNA vaccination and malaria vaccination [44, 45].

Positive and negative impacts of preexisting K163 antibodies to CA/09-like vaccine-induced antibody responses have been observed (Tables 3 and 4). On the one hand, the vaccine induced new non-K163 antibodies (G2 donors) that cross-reacted well with the newly emerging K163Q variants that prevailed since 2013 (Table 1). On the other hand, high levels of preexisting K163 antibodies (G1 donors) totally suppressed HAI antibody responses following vaccination. Such dual impacts were observed for other vaccines [33, 44, 45, 47]. Several approaches could be used to overcome the antibody-mediated suppression in future vaccine development and administration, such as modified immunization schedules, improved vaccine formulations (including the use of adjuvants), or increased vaccine dosage [26, 27, 45, 47].

Whether antibody-mediated suppression plays a role in modulating immune responses to influenza has recently been discussed [34–38] and explored in mathematical models and computer simulations [26, 27, 48]. Nevertheless, this strategy must be tested in experimental settings. To the best of our knowledge, this is the first experimental study to investigate how preexisting HAI antibodies modulate neutralizing antibody responses to subsequent influenza vaccination. However, there are certain limitations in this study. Although 300 adults were recruited during the 6 seasons, we were able to detect just a small number (n = 7) possessing focused preexisting K163 antibodies (Figure 1). Clearly, the impact of antibody-mediated suppression on immune responses following influenza infection and vaccination in different populations needs to be explored in larger studies. Imprinting by the first virus infection can also affect the subsequent antibody responses to A(H3N2) [49, 50]. Whether adjuvated, high-dose, or other enhanced influenza vaccines could influence antibody-mediated suppression and whether vaccine-induced antibody-mediated suppression could affect vaccine effectiveness should be further studied. Nevertheless, our data emphasize that antibody responses to influenza vaccination in humans are complex, adding the effects of epitope- and nonepitope-specific suppression to the effects of imprinting caused by the first exposure to historical IAVs (OAS).

In summary, our data suggest that K163 antibody–mediated suppression shapes the humoral immune responses to CA/09-like vaccines in some middle-aged adults. At high K163 antibody levels, antibody-mediated suppression prevents all HAI antibody responses. At moderate K163 antibody levels, antibody-mediated suppression limits the generation of the epitope-specific K163 antibodies but diversifies the humoral responses as new non-K163 antibody responses expand. Determining how the preexisting antibodies suppress and redirect vaccine-induced antibody responses is of great interest and has practical importance for the control of influenza as well as other diseases.

Supplementary Data

Supplementary materials are available at The Journal of Infectious Diseases online (http://jid.oxfordjournals.org/). Supplementary materials consist of data provided by the author that are published to benefit the reader. The posted materials are not copyedited. The contents of all supplementary data are the sole responsibility of the authors. Questions or messages regarding errors should be addressed to the author.

Notes

Acknowledgments. We thank the Influenza Genomics Team in the Virology Surveillance and Diagnostic Branch, Influenza Division, Centers for Disease Control and Prevention, for sequencing assistance.

Disclaimer. The findings and conclusions in this report are those of the authors and do not necessarily represent the views of the Centers for Disease Control and Prevention and the funding agency.

Financial support. This work was supported by the Centers for Disease Control and Prevention.

References

Author notes