Universal influenza virus vaccines: what can we learn from the human immune response following exposure to H7 subtype viruses?

Abstract

Several universal influenza virus vaccine candidates based on eliciting antibodies against the hemagglutinin stalk domain are in development. Typically, these vaccines induce responses that target group 1 or group 2 hemagglutinins with little to no cross-group reactivity and protection. Similarly, the majority of human anti-stalk monoclonal antibodies that have been isolated are directed against group 1 or group 2 hemagglutinins with very few that bind to hemagglutinins of both groups. Here we review what is known about the human humoral immune response to vaccination and infection with H7 subtype influenza viruses on a polyclonal and monoclonal level. It seems that unlike vaccination with H5 hemagglutinin, which induces antibody responses mostly restricted to the group 1 stalk domain, H7 exposure induces both group 2 and cross-group antibody responses. A better understanding of this phenomenon and the underlying mechanisms might help to develop future universal influenza virus vaccine candidates.

Keywords

universal influenza virus vaccine / hemagglutinin stalk / H7N9

Introduction

Influenza viruses cause annual epidemics and, in irregular intervals, pandemics in the human population. Viral infection in humans leads to respiratory disease that can be associated with severe morbidity and mortality with up to half a million deaths caused by seasonal influenza viruses every year globally [1]. Pandemics are usually associated with higher numbers of deaths. An example is the 1918/19 H1N1 pandemic which caused an estimated 40 million deaths worldwide [2]. Influenza viruses are a significant challenge for vaccine design due to their ability to mutate and escape antibody-based immunity through introduction of antigenic changes in their surface glycoproteins hemagglutinin (HA) and neuraminidase (NA). The HA, specifically the immunodominant, membrane distal globular head domain, is the major target of influenza virus vaccines (Fig. 1A) [2]. The globular head domain harbors the receptor binding site of the virus and antibodies that block this site prevent the virus from attaching to host cells and therefore neutralize it. Unfortunately, this domain has a very high plasticity [3,4] and the virus escapes the antibody response by introducing point mutations that lead to changes in the major antigenic sites on the head domain, a phenomenon called antigenic drift. Therefore, seasonal influenza virus vaccines need to be updated frequently [5]. The situation is even worse for emerging pandemic viruses for which matched vaccines need to be manufactured, a process that usually takes up to six months during which the population is vulnerable to infection [2]. These problems have led to major initiatives to develop a universal influenza virus vaccine that would protect from all antigenically drifted seasonal influenza viruses, zoonotic influenza virus infections, and pandemic influenza viruses [6,7]. Several conserved targets for such a vaccine have been proposed including the membrane proximal stalk domain (Fig. 1A) of the HA which evolves much slower than the head domain [8]. In fact, antibodies against this domain have been isolated from mice and humans and can neutralize a broad panel of influenza virus subtypes and strains. The majority of these broadly neutralizing antibodies bind to either the stalk of influenza A group 1 HAs (H1, H2, H5, H6, H8, H9, H11, H12, H16, H17, H18) [9–11] or to the stalk of influenza A group 2 HAs (H3, H4, H7, H10, H14, H15) [12,13] with some notable exceptions that bind to both groups [14–16] or even all influenza A and B HAs [17]. It has been noted that exposure of humans with pre-existing immunity to pre-pandemic seasonal H1N1 to novel influenza virus strains/subtypes from group 1 that feature drastically different head domains, but conserved epitopes in the stalk domain leads to preferential induction of these broadly neutralizing stalk-reactive antibodies. This has been shown for infection and vaccination with 2009 pandemic H1N1 (which has a very different head domain compared to pre-pandemic seasonal H1N1) and for vaccination with H5N1 vaccines [18–25]. These observations have led to vaccine strategies based on chimeric HAs which have “exotic” head domains, but stalk domains from e.g. H1 or H3 HAs. Sequential vaccination with these constructs in animal models also leads to the preferential induction of broadly neutralizing antibodies [26]. While several studies have elucidated the immune response after sequential exposure to divergent group 1 HAs in humans, studies about sequential exposure to group 2 HAs are sparse. Here we will review what is known about the immune response after H7 (group 2) exposure in humans that have been primed by infection and/or vaccination with seasonal H3N2 (group 2) viruses. Lessons learned from these studies might inform further development of group 2 targeted universal influenza virus vaccine candidates.

Pre-existing immunity to H7 HA on a polyclonal and monoclonal level

The human population is constantly exposed to seasonal H3N2 viruses through both vaccination and infection. H3N2 has been circulating in humans ever since it caused a pandemic in 1968 (“Hong Kong Flu”) and can cause severe morbidity and also mortality. Therefore, the human population is primed for group 2 HAs. However, sera from human individuals have little or no baseline neutralizing activity to H7 subtype viruses [28–30]. This is not very surprising since most of the immune response induced by H3N2 exposure is directed toward the globular head domain [31] and H3 and H7 globular head domains only share approximately 30% amino acid identity [32]. In addition, H3 and H7 HAs are relatively far apart phylogenetically within group 2 HAs, with H3 being a member of the H3 clade (with H4 and H14) and H7 being a member of the H7 clade (with H10 and H15) (Fig. 1B). However, H3N2 infections (and potentially to a lower degree, vaccination) induces low levels of group 2 stalk-reactive antibodies in humans [28,33]. Of note, these levels are substantially lower than group 1 stalk reactive antibodies, likely because circulation of very different group 1 HA expressing viruses (H1N1, H2N2, pandemic H1N1) has boosted stalk titers against group 1 HAs to higher levels [34,35]. In fact, the reactivity to H7 is lower than most other group 2 or group 1 HA independently of the age group (Fig. 2). Nevertheless, some cross-reactivity to H7 HA has been detected in serum of human individuals [28]. This also translates to the monoclonal level. In a recent study Henry Dunand and colleagues found that of 83 H3 reactive antibodies isolated from plasmablasts after vaccination only six reacted to H7 [36]. Three of these six antibodies showed neutralizing activity and protected mice from a lethal challenge with H7N9 virus. Interestingly, two of the mAbs also reacted with group 1 HAs while one had a pan-group 2 profile. It is interesting to note that two of the mAbs were isolated after seasonal influenza virus vaccination (containing H1N1, H3N2 and influenza B components) while the third, cross-group mAb was isolated after vaccination with monovalent pH1N1 vaccine. The cross-group mAbs both featured variable heavy (VH) segments from the VH1-18 heavy chain germline [37]. Similarly, low numbers of H7 reactive memory-B cells were found in the same study. While human serum has low reactivity to H7, this reactivity is boosted to some degree after H3N2 infection [28]. However, even then antibody levels remain low. In summary, low levels of antibodies and B cells with specificities to H7 HA exist in humans, and the majority of this reactivity is likely induced by H3N2 exposure and directed to the stalk domain of HA. Of note, recent work by Gostic and colleagues hypothesized that group 2 imprinting by H3N2 infection during childhood could explain protection of certain segments of the population from severe infections and death caused by H7N9 viruses [38]. The authors speculate that this could be driven by antibodies to the group 2 stalk. This is an intriguing thought and while reactivity to H7 HA is very low in all age groups (including those likely first exposed in life to H3N2; Fig. 2) this phenomenon might be driven by memory B cells for which antibody products are not readily detected in serum.

Breadth of the polyclonal immune response after H7 vaccination

Before H7N9 emerged in 2013 [39,40], very few vaccine trials with H7 HA had been conducted. The general consensus from these trials was that H7 HAs are of very low immunogenicity, even lower than H5 HAs [29,41,42]. Similar observations have been made with H7N9 vaccines. However, the use of adjuvants [30], virus-like particles [43], or heterologous prime-boost regimens [44,45] improved the immunogenicity but results still lagged behind what is typically observed for seasonal influenza virus vaccines. Importantly, the typical readout used in these trials is the hemagglutination inhibition (HI) assay. In addition, microneutralization (MN) assays are now used to assess the induction of H7 specific immunity, but results from these assays — unlike results from the HI assay— are not accepted by regulatory agencies as “surrogates” or “correlates” of protection. The HI assay only detects antibodies against the globular head domain, while the MN assay detects mostly antibodies against the head domain but might detect stalk-reactive antibodies as well, when they are present at high levels. Both assays usually miss the induction of binding but non-neutralizing antibodies as well as levels of neutralizing antibodies below the limit of detection. Importantly, these types of antibodies might still provide protection, e.g. through effector functions, as demonstrated by several studies [46–49]. Enzyme-linked immunosorbent assays (ELISAs) were used to assess the immune response in only very few studies. These assays detect all antibodies that bind a certain HA, including antibodies that are not detected in HI and MN assays. A study based on an H7N1 inactivated vaccine prime-boost regimen found strong cross-reactivity induced by ELISA, while HI or MN activity was negligible [50]. The antibody response also extended to H15 and an induction of group 2 stalk-reactive antibodies was detected (albeit at low levels). An induction of antibodies to seasonal H1 HA was not detected. Another study examined the antibody response in humans vaccinated twice with an H7N7 live attenuated vaccines (LAIV) followed by an H7N7 inactivated vaccine (IIV) [51]. An antibody response against both the H7 head domain as well as the stalk domain was detected by several methods including analysis of memory B cell frequency and by mapping via a phage library. The memory B cell frequency against pandemic H1 HA was also measured but no significant increase was detected. From these limited data it can be concluded that H7 vaccines induce detectible levels of group 2 stalk-reactive antibodies.

Breadth of the polyclonal immune response after H7N9 infection

So far, two studies have analyzed the breadth of the antibody response after H7N9 infection in humans. In an early paper in 2014 Guo and colleagues analyzed sera from samples from 21 patients and found an increase of H7 binding antibodies, HI and MN titers post H7N9 infection [52]. Interestingly, they also detected an increase in H3 and pandemic H1 antibody binding by ELISA. The increase was higher and longer lasting for H3, which would be expected since it can be hypothesized that H7 exposure in H3 experienced individuals would boost stalk-reactive antibodies. The induction of antibodies to pandemic H1 HA however is highly unusual. A second, more detailed study that examined sero-reactivity of 18 individuals infected with H7N9 was recently published [53]. Liu and colleagues tested binding to all HA subtypes and found that H7N9 infection boosted binding antibody titers (measured by ELISA) against group 1 and group 2 HAs (but not influenza B virus HA). As expected, overall stronger induction (and higher absolute titers) was observed for group 2 HAs but reactivity against several group 1 HAs, specifically pandemic H1, was strongly induced as well. Of note, this was also reflected in neutralization titers which increased against many subtypes including pandemic H1N1.

These are important findings since they suggest that infection with H7N9 induces a much broader immune response than infection with H3N2 or even vaccination with H7 vaccines, which induce some group 2 specific stalk antibodies on a polyclonal basis. The response was comparable to the primary response to infection with pandemic H1N1 in humans previously exposed to seasonal H1N1, which can also trigger cross-group HA responses [28]. A direct comparison with the immune response after H5N1 vaccination cannot be made since only very limited data from H5N1 survivors exists [54].

It is important to keep in mind that the analyzed cohorts typically consisted of elderly individuals who were likely exposed to influenza viruses many times. In addition, many of the H7N9 infected individuals frequented wet markets, which increases the chance that they were exposed to avian influenza viruses of different subtypes in the past and had therefore a skewed immune response already. Finally, most of these individuals suffered from co-morbidities which might have influenced their immune responses in unknown ways.

Analysis of the monoclonal immune response after H7N9 vaccination

Three studies have so far analyzed the monoclonal antibody response to H7N9 vaccination. An initial study analyzed mAbs derived from memory B cells of 12 donors who had received an inactivated split H7N9 vaccine (adjuvanted with MF59 or AS03) twice. Twenty hybridomas were obtained and 11 mAbs were further characterized [55]. Eight mAbs showed binding that was restricted to H7 HAs, with 5 of them having classical HI activity. Three mAbs showed broader binding activity with two binding broadly to group 2 HAs and one binding to both group 2 and group 1 HAs. One of the group 2 cross-reactive mAbs showed strong neutralization against H3N2 while the other two seemed to be non-neutralizing antibodies, with the caveat that neutralizing activity was only tested up to a concentration of 2 μg/mL. This concentration might be not in the range of many anti-stalk antibodies which are typically less potent neutralizers as compared to HI-active antibodies [56,57]. Another study analyzed mAbs isolated from plasmablasts after a vaccination regimen that included two H7N9 LAIV inoculations followed by an inactivated H7N9 vaccine [46]. Of 12 analyzed mAbs, eight showed specificity (albeit mostly broad) for H7 and had HI and/or MN activity. One of these mAbs, an HI active antibody, cross-reacted to H15. Three mAbs had no neutralizing activity but showed broad binding to group 2 HAs, and in two cases to group 1 HAs. While these mAbs did not neutralize, they were still capable of protecting mice from a lethal challenge with H7N9 virus. The last antibody showed no HI activity but was neutralizing, and was identified as bona fide anti-stalk mAb. It bound strongly to group 2 HAs as well as pandemic H1 HA (and with lower affinity to some other group 1 HAs). Of note, this mAb featured a VH from the VH3-53 germline (see below). A third, very recent study [58] reported mAbs derived from memory B cells from a vaccine trial that tested H7 DNA prime-H7N9 inactivated vaccine boost vaccination regimens (with H7N9 inactivated vaccine prime-boost regimens or DNA plus inactivated vaccine twice as controls) [45]. Cells were obtained from all three vaccine groups and the authors stated that no difference was found between the groups. Memory B cells were sorted for reactivity to H7 and other HAs. The authors reported that the overall amount of cross-reactive cells was lower for H7 as compared to an H5 trial with a similar setup. However, while H5N1 vaccination induced approximately 50% group 1 stalk-reactive antibodies and very little cross-group reactivity, the percentage of cross-group reactive antibodies was close to 25% in the H7 vaccinees. The response after H5N1 vaccination was skewed toward the VH1-69 germline, which is heavily over-represented in group 1 stalk-reactive antibodies. After H7N9 vaccination the response was skewed toward a VH1-18 response. In this study, several mAbs were cloned out and analyzed for signatures. mAbs with a VH1-18 germline, or a VH6-1 germline typically showed cross-group reactivity. However, while VH6-1 mAbs neutralized both H1N1 and H3N2, VH1-18 mAbs either neutralized H1N1 or H3N2 (more often). Another signature found were VH3-53 mAbs, which showed pan group 2 activity and neutralized H3N2, but also had an affinity toward pandemic H1 HA (with no binding to seasonal H1 and very little binding to other group 1 HAs). Interestingly these VH1-18 and VH3-53 mAbs with similar signatures had also been isolated in earlier studies from plasmablast responses to seasonal trivalent, pandemic H1N1 and H7N9 vaccines [36,46].

These data are highly interesting and suggest that the human immune system is capable of reacting to H7 exposure with both group 2 specific as well as cross-group anti-stalk responses. Of note, some bona fide group 2 stalk antibodies (VH3-53) also seem to have an unusual, selective affinity for pandemic H1— a signature that was also detected in the polyclonal response after H7N9 infection.

Conclusions

Based on the limited data available it becomes clear that the human immune response to H7 HA is — like the response to H5 HA — strongly influenced by H1 and H3 primed pre-existing memory-B cells that have specificities for conserved epitopes which are shared between human and avian influenza viruses. While H5N1 vaccination seems to produce strong responses of stalk antibodies that cross-react between group 1 HAs (with a skewed, VH1-69 signature), exposure to H7 HA seems to trigger a slightly different response. H7 vaccination, as analyzed so far, seems to induce a polyclonal serological response that mostly targets H7 and other group 2 HAs with a substantial proportion of group 2 stalk antibodies present. H7N9 infection triggers a polyclonal serum response that is much broader than the one induced by vaccination and includes induction of antibodies against group 2 and group 1 HAs with a notable reactivity to H1. On a monoclonal level, broadly reactive mAbs with VH1-18 signatures can be isolated from plasmablasts even after vaccination with seasonal trivalent inactivated or monovalent pH1N1 vaccine. Plasmablast responses after H7N9 vaccination also reveal that neutralizing (VH3-53) and non-neutralizing broadly protective antibodies are induced. Similarly, analysis of the memory B cell compartment after H7N9 vaccination revealed broadly binding and neutralizing mAbs, in many cases spanning group 1 and group 2 (VH6-1, VH1-18) or with pan-group 2 plus pH1 reactivity (VH3-53). After infection and on a cellular, monoclonal level these responses seem to be much broader than the VH1-69 dominated group 1 stalk response after H5N1 vaccination. Currently, it is unclear why there is a difference in the response to H5 as compared to the response to H7. The stalk structure of group 1 and group 2 HAs is slightly different and the positions of conserved glycans differ as well. It might be that germlines like VH1-18, VH3-53 and VH6-1 have structure/sequence constellations that facilitate interactions with the H7 (group 2) stalk while binding is not well supported by the VH1-69 germline.

It is important to realize that both plasmablast and memory B responses — on which the current data are mostly based — are not necessarily reflective of long-term serum antibody responses that are driven by long lived plasma cells in the bone marrow. Just because a memory B cell or plasmablast response produces antibodies with broad specificities does not necessarily mean that the same cells (or clonally related cells) will migrate to the bone marrow to become long lived plasma cells that provide constant levels of serum antibody. Additional serological analysis of H7N9 vaccine trials (including heterologous prime-boost regimens and adjuvanted vaccination) are needed to reach a better understanding of the long-term serum response. Nevertheless, these data — although based on only a handful of studies — suggest that a group 2 stalk antigen might be a viable option to induce broad protection against group 2 HA expressing viruses including H3N2, H7N9 and H10N8. These limited data also suggest that a very strong stalk-based immunity induced by a group 2 construct might provide some protection against group 1 HA expressing viruses, specifically pH1N1.

References

[1] WHO. WHO Fact sheet N°211. 2014

[2] Krammer F, Palese P. Advances in the development of influenza virus vaccines. Nat Rev Drug Discov 2015; 14(3): 167–182

[3] Heaton NS, Sachs D, Chen CJ, Hai R, Palese P. Genome-wide mutagenesis of influenza virus reveals unique plasticity of the hemagglutinin and NS1 proteins. Proc Natl Acad Sci USA 2013; 110(50): 20248–20253

[4] Doud MB, Bloom JD. Accurate measurement of the effects of all amino-acid mutations on influenza hemagglutinin. Viruses 2016; 8(6): E155

[5] Gerdil C. The annual production cycle for influenza vaccine. Vaccine 2003; 21(16): 1776–1779

[6] Berlanda Scorza F, Tsvetnitsky V, Donnelly JJ. Universal influenza vaccines: shifting to better vaccines. Vaccine 2016; 34(26): 2926–2933

[7] Krammer F, García-Sastre A, Palese P. Is it possible to develop a “Universal” influenza virus vaccine? Toward a universal influenza virus vaccine: potential target antigens and critical aspects for vaccine development. Cold Spring Harb Perspect Biol 2017; a028845

[8] Nachbagauer R, Krammer F. Universal influenza virus vaccines and therapeutic antibodies. Clin Microbiol Infect 2017; 23(4): 222–228

[9] Ekiert DC, Bhabha G, Elsliger MA, Friesen RH, Jongeneelen M, Throsby M, Goudsmit J, Wilson IA. Antibody recognition of a highly conserved influenza virus epitope. Science 2009; 324(5924): 246–251

[10] Sui J, Hwang WC, Perez S, Wei G, Aird D, Chen LM, Santelli E, Stec B, Cadwell G, Ali M, Wan H, Murakami A, Yammanuru A, Han T, Cox NJ, Bankston LA, Donis RO, Liddington RC, Marasco WA. Structural and functional bases for broad-spectrum neutralization of avian and human influenza A viruses. Nat Struct Mol Biol 2009; 16(3): 265–273

[11] Throsby M, van den Brink E, Jongeneelen M, Poon LL, Alard P, Cornelissen L, Bakker A, Cox F, van Deventer E, Guan Y, Cinatl J, ter Meulen J, Lasters I, Carsetti R, Peiris M, de Kruif J, Goudsmit J. Heterosubtypic neutralizing monoclonal antibodies cross-protective against H5N1 and H1N1 recovered from human IgM+ memory B cells. PLoS One 2008; 3(12): e3942

[12] Friesen RH, Lee PS, Stoop EJ, Hoffman RM, Ekiert DC, Bhabha G, Yu W, Juraszek J, Koudstaal W, Jongeneelen M, Korse HJ, Ophorst C, Brinkman-van der Linden EC, Throsby M, Kwakkenbos MJ, Bakker AQ, Beaumont T, Spits H, Kwaks T, Vogels R, Ward AB, Goudsmit J, Wilson IA. A common solution to group 2 influenza virus neutralization. Proc Natl Acad Sci USA 2014; 111(1): 445–450

[13] Ekiert DC, Friesen RH, Bhabha G, Kwaks T, Jongeneelen M, Yu W, Ophorst C, Cox F, Korse HJ, Brandenburg B, Vogels R, Brakenhoff JP, Kompier R, Koldijk MH, Cornelissen LA, Poon LL, Peiris M, Koudstaal W, Wilson IA, Goudsmit J. A highly conserved neutralizing epitope on group 2 influenza A viruses. Science 2011; 333(6044): 843–850

[14] Corti D, Voss J, Gamblin SJ, Codoni G, Macagno A, Jarrossay D, Vachieri SG, Pinna D, Minola A, Vanzetta F, Silacci C, Fernandez-Rodriguez BM, Agatic G, Bianchi S, Giacchetto-Sasselli I, Calder L, Sallusto F, Collins P, Haire LF, Temperton N, Langedijk JP, Skehel JJ, Lanzavecchia A. A neutralizing antibody selected from plasma cells that binds to group 1 and group 2 influenza A hemagglutinins. Science 2011; 333(6044): 850–856

[15] Wu Y, Cho M, Shore D, Song M, Choi J, Jiang T, Deng YQ, Bourgeois M, Almli L, Yang H, Chen LM, Shi Y, Qi J, Li A, Yi KS, Chang M, Bae JS, Lee H, Shin J, Stevens J, Hong S, Qin CF, Gao GF, Chang SJ, Donis RO. A potent broad-spectrum protective human monoclonal antibody crosslinking two haemagglutinin monomers of influenza A virus. Nat Commun 2015; 6: 7708

[16] Gupta P, Kamath AV, Park S, Chiu H, Lutman J, Maia M, Tan MW, Xu M, Swem L, Deng R. Preclinical pharmacokinetics of MHAA4549A, a human monoclonal antibody to influenza A virus, and the prediction of its efficacious clinical dose for the treatment of patients hospitalized with influenza A. MAbs 2016; 8(5): 991–997

[17] Dreyfus C, Laursen NS, Kwaks T, Zuijdgeest D, Khayat R, Ekiert DC, Lee JH, Metlagel Z, Bujny MV, Jongeneelen M, van der Vlugt R, Lamrani M, Korse HJ, Geelen E, Sahin Ö, Sieuwerts M, Brakenhoff JP, Vogels R, Li OT, Poon LL, Peiris M, Koudstaal W, Ward AB, Wilson IA, Goudsmit J, Friesen RH. Highly conserved protective epitopes on influenza B viruses. Science 2012; 337(6100): 1343–1348

[18] Nachbagauer R, Wohlbold TJ, Hirsh A, Hai R, Sjursen H, Palese P, Cox RJ, Krammer F. Induction of broadly reactive anti-hemagglutinin stalk antibodies by an H5N1 vaccine in humans. J Virol 2014; 88(22): 13260–13268

[19] Ellebedy AH, Krammer F, Li GM, Miller MS, Chiu C, Wrammert J, Chang CY, Davis CW, McCausland M, Elbein R, Edupuganti S, Spearman P, Andrews SF, Wilson PC, García-Sastre A, Mulligan MJ, Mehta AK, Palese P, Ahmed R. Induction of broadly cross-reactive antibody responses to the influenza HA stem region following H5N1 vaccination in humans. Proc Natl Acad Sci USA 2014; 111(36): 13133–13138

[20] Wrammert J, Koutsonanos D, Li GM, Edupuganti S, Sui J, Morrissey M, McCausland M, Skountzou I, Hornig M, Lipkin WI, Mehta A, Razavi B, Del Rio C, Zheng NY, Lee JH, Huang M, Ali Z, Kaur K, Andrews S, Amara RR, Wang Y, Das SR, O’Donnell CD, Yewdell JW, Subbarao K, Marasco WA, Mulligan MJ, Compans R, Ahmed R, Wilson PC. Broadly cross-reactive antibodies dominate the human B cell response against 2009 pandemic H1N1 influenza virus infection. J Exp Med 2011; 208(1): 181–193

[21] Li GM, Chiu C, Wrammert J, McCausland M, Andrews SF, Zheng NY, Lee JH, Huang M, Qu X, Edupuganti S, Mulligan M, Das SR, Yewdell JW, Mehta AK, Wilson PC, Ahmed R. Pandemic H1N1 influenza vaccine induces a recall response in humans that favors broadly cross-reactive memory B cells. Proc Natl Acad Sci USA 2012; 109(23): 9047–9052

[22] Thomson CA, Wang Y, Jackson LM, Olson M, Wang W, Liavonchanka A, Keleta L, Silva V, Diederich S, Jones RB, Gubbay J, Pasick J, Petric M, Jean F, Allen VG, Brown EG, Rini JM, Schrader JW. Pandemic H1N1 influenza infection and vaccination in humans induces cross-protective antibodies that target the hemagglutinin stem. Front Immunol 2012; 3: 87

[23] Pica N, Hai R, Krammer F, Wang TT, Maamary J, Eggink D, Tan GS, Krause JC, Moran T, Stein CR, Banach D, Wrammert J, Belshe RB, García-Sastre A, Palese P. Hemagglutinin stalk antibodies elicited by the 2009 pandemic influenza virus as a mechanism for the extinction of seasonal H1N1 viruses. Proc Natl Acad Sci USA 2012; 109(7): 2573–2578

[24] Qiu C, Huang Y, Wang Q, Tian D, Zhang W, Hu Y, Yuan Z, Zhang X, Xu J. Boosting heterosubtypic neutralization antibodies in recipients of 2009 pandemic H1N1 influenza vaccine. Clin Infect Dis 2012; 54(1): 17–24

[25] Andrews SF, Huang Y, Kaur K, Popova LI, Ho IY, Pauli NT, Henry Dunand CJ, Taylor WM, Lim S, Huang M, Qu X, Lee JH, Salgado-Ferrer M, Krammer F, Palese P, Wrammert J, Ahmed R, Wilson PC. Immune history profoundly affects broadly protective B cell responses to influenza. Sci Transl Med 2015; 7(316): 316ra192

[26] Krammer F, Palese P. Influenza virus hemagglutinin stalk-based antibodies and vaccines. Curr Opin Virol 2013; 3(5): 521–530

[27] Russell RJ, Kerry PS, Stevens DJ, Steinhauer DA, Martin SR, Gamblin SJ, Skehel JJ. Structure of influenza hemagglutinin in complex with an inhibitor of membrane fusion. Proc Natl Acad Sci USA 2008; 105(46): 17736–17741

[28] Nachbagauer R, Choi A, Hirsh A, Margine I, Iida S, Barrera A, Ferres M, Albrecht RA, García-Sastre A, Bouvier NM, Ito K, Medina RA, Palese P, Krammer F. Defining the antibody cross-reactome directed against the influenza virus surface glycoproteins. Nat Immunol 2017; 18(4): 464–473

[29] Cox RJ, Madhun AS, Hauge S, Sjursen H, Major D, Kuhne M, Höschler K, Saville M, Vogel FR, Barclay W, Donatelli I, Zambon M, Wood J, Haaheim LR. A phase I clinical trial of a PER.C6 cell grown influenza H7 virus vaccine. Vaccine 2009; 27(13): 1889–1897

[30] Mulligan MJ, Bernstein DI, Winokur P, Rupp R, Anderson E, Rouphael N, Dickey M, Stapleton JT, Edupuganti S, Spearman P, Ince D, Noah DL, Hill H, Bellamy AR; DMID 13-0032 H7N9 Vaccine Study Group. Serological responses to an avian influenza A/H7N9 vaccine mixed at the point-of-use with MF59 adjuvant: a randomized clinical trial. JAMA 2014; 312(14): 1409–1419

[31] Moody MA, Zhang R, Walter EB, Woods CW, Ginsburg GS, McClain MT, Denny TN, Chen X, Munshaw S, Marshall DJ, Whitesides JF, Drinker MS, Amos JD, Gurley TC, Eudailey JA, Foulger A, DeRosa KR, Parks R, Meyerhoff RR, Yu JS, Kozink DM, Barefoot BE, Ramsburg EA, Khurana S, Golding H, Vandergrift NA, Alam SM, Tomaras GD, Kepler TB, Kelsoe G, Liao HX, Haynes BF. H3N2 influenza infection elicits more cross-reactive and less clonally expanded anti-hemagglutinin antibodies than influenza vaccination. PLoS One 2011; 6(10): e25797

[32] Hai R, Krammer F, Tan GS, Pica N, Eggink D, Maamary J, Margine I, Albrecht RA, Palese P. Influenza viruses expressing chimeric hemagglutinins: globular head and stalk domains derived from different subtypes. J Virol 2012; 86(10): 5774–5781

[33] Margine I, Hai R, Albrecht RA, Obermoser G, Harrod AC, Banchereau J, Palucka K, García-Sastre A, Palese P, Treanor JJ, Krammer F. H3N2 influenza virus infection induces broadly reactive hemagglutinin stalk antibodies in humans and mice. J Virol 2013; 87(8): 4728–4737

[34] Miller MS, Gardner TJ, Krammer F, Aguado LC, Tortorella D, Basler CF, Palese P. Neutralizing antibodies against previously encountered influenza virus strains increase over time: a longitudinal analysis. Sci Transl Med 2013; 5(198): 198ra107

[35] Nachbagauer R, Choi A, Izikson R, Cox MM, Palese P, Krammer F. Age dependence and isotype specificity of influenza virus hemagglutinin stalk-reactive antibodies in humans. MBio 2016; 7(1): e01996–e15

[36] Henry Dunand CJ, Leon PE, Kaur K, Tan GS, Zheng NY, Andrews S, Huang M, Qu X, Huang Y, Salgado-Ferrer M, Ho IY, Taylor W, Hai R, Wrammert J, Ahmed R, García-Sastre A, Palese P, Krammer F, Wilson PC. Preexisting human antibodies neutralize recently emerged H7N9 influenza strains. J Clin Invest 2015; 125(3): 1255–1268

[37] Watson CT, Breden F. The immunoglobulin heavy chain locus: genetic variation, missing data, and implications for human disease. Genes Immun 2012; 13(5): 363–373

[38] Gostic KM, Ambrose M, Worobey M, Lloyd-Smith JO. Potent protection against H5N1 and H7N9 influenza via childhood hemagglutinin imprinting. Science 2016; 354(6313): 722–726

[39] Liu D, Shi W, Shi Y, Wang D, Xiao H, Li W, Bi Y, Wu Y, Li X, Yan J, Liu W, Zhao G, Yang W, Wang Y, Ma J, Shu Y, Lei F, Gao GF. Origin and diversity of novel avian influenza A H7N9 viruses causing human infection: phylogenetic, structural, and coalescent analyses. Lancet 2013; 381(9881): 1926–1932

[40] Gao R, Cao B, Hu Y, Feng Z, Wang D, Hu W, Chen J, Jie Z, Qiu H, Xu K, Xu X, Lu H, Zhu W, Gao Z, Xiang N, Shen Y, He Z, Gu Y, Zhang Z, Yang Y, Zhao X, Zhou L, Li X, Zou S, Zhang Y, Li X, Yang L, Guo J, Dong J, Li Q, Dong L, Zhu Y, Bai T, Wang S, Hao P, Yang W, Zhang Y, Han J, Yu H, Li D, Gao GF, Wu G, Wang Y, Yuan Z, Shu Y. Human infection with a novel avian-origin influenza A (H7N9) virus. N Engl J Med 2013; 368(20): 1888–1897

[41] Talaat KR, Karron RA, Callahan KA, Luke CJ, DiLorenzo SC, Chen GL, Lamirande EW, Jin H, Coelingh KL, Murphy BR, Kemble G, Subbarao K. A live attenuated H7N3 influenza virus vaccine is well tolerated and immunogenic in a Phase I trial in healthy adults. Vaccine 2009; 27(28): 3744–3753

[42] Couch RB, Patel SM, Wade-Bowers CL, Niño D. A randomized clinical trial of an inactivated avian influenza A (H7N7) vaccine. PLoS One 2012; 7(12): e49704

[43] Fries LF, Smith GE, Glenn GM. A recombinant viruslike particle influenza A (H7N9) vaccine. N Engl J Med 2013; 369(26): 2564–2566

[44] Babu TM, Levine M, Fitzgerald T, Luke C, Sangster MY, Jin H, Topham D, Katz J, Treanor J, Subbarao K. Live attenuated H7N7 influenza vaccine primes for a vigorous antibody response to inactivated H7N7 influenza vaccine. Vaccine 2014; 32(50): 6798–6804

[45] DeZure AD, Coates EE, Hu Z, Yamshchikov GV, Zephir KL, Enama ME, Plummer SH, Gordon IJ, Kaltovich F, Andrews S, McDermott A, Crank MC, Koup RA, Schwartz RM, Bailer RT, Sun X, Mascola JR, Tumpey TM, Graham BS, Ledgerwood JE. An avian influenza H7 DNA priming vaccine is safe and immunogenic in a randomized phase I clinical trial. Vaccines (Basel) 2017; 2(1): 15

[46] Henry Dunand CJ, Leon PE, Huang M, Choi A, Chromikova V, Ho IY, Tan GS, Cruz J, Hirsh A, Zheng NY, Mullarkey CE, Ennis FA, Terajima M, Treanor JJ, Topham DJ, Subbarao K, Palese P, Krammer F, Wilson PC. Both neutralizing and non-neutralizing human H7N9 influenza vaccine-induced monoclonal antibodies confer protection. Cell Host Microbe 2016; 19(6): 800–813

[47] Tan GS, Leon PE, Albrecht RA, Margine I, Hirsh A, Bahl J, Krammer F. Broadly-reactive neutralizing and non-neutralizing antibodies directed against the H7 influenza virus hemagglutinin reveal divergent mechanisms of protection. PLoS Pathog 2016; 12(4): e1005578

[48] Jegaskanda S, Job ER, Kramski M, Laurie K, Isitman G, de Rose R, Winnall WR, Stratov I, Brooks AG, Reading PC, Kent SJ. Cross-reactive influenza-specific antibody-dependent cellular cytotoxicity antibodies in the absence of neutralizing antibodies. J Immunol 2013; 190(4): 1837–1848

[49] Jegaskanda S, Weinfurter JT, Friedrich TC, Kent SJ. Antibody-dependent cellular cytotoxicity is associated with control of pandemic H1N1 influenza virus infection of macaques. J Virol 2013; 87(10): 5512–5522

[50] Krammer F, Jul-Larsen A, Margine I, Hirsh A, Sjursen H, Zambon M, Cox RJ. An H7N1 influenza virus vaccine induces broadly reactive antibody responses against H7N9 in humans. Clin Vaccine Immunol 2014; 21(8): 1153–1163

[51] Halliley JL, Khurana S, Krammer F, Fitzgerald T, Coyle EM, Chung KY, Baker SF, Yang H, Martínez-Sobrido L, Treanor JJ, Subbarao K, Golding H, Topham DJ, Sangster MY. High-affinity H7 head and stalk domain-specific antibody responses to an inactivated influenza H7N7 vaccine after priming with live attenuated influenza vaccine. J Infect Dis 2015; 212(8): 1270–1278

[52] Guo L, Zhang X, Ren L, Yu X, Chen L, Zhou H, Gao X, Teng Z, Li J, Hu J, Wu C, Xiao X, Zhu Y, Wang Q, Pang X, Jin Q, Wu F, Wang J. Human antibody responses to avian influenza A(H7N9) virus, 2013. Emerg Infect Dis 2014; 20(2): 192–200

[53] Liu L, Nachbagauer R, Zhu L, Huang Y, Xie X, Jin S, Zhang A, Wan Y, Hirsh A, Tian D, Shi X, Dong Z, Yuan S, Hu Y, Krammer F, Zhang X, Xu J. Induction of broadly cross-reactive stalk-specific antibody responses to influenza group 1 and group 2 hemagglutinins by natural H7N9 virus infection in humans. J Infect Dis 2017; 215(4): 518–528

[54] Kashyap AK, Steel J, Oner AF, Dillon MA, Swale RE, Wall KM, Perry KJ, Faynboym A, Ilhan M, Horowitz M, Horowitz L, Palese P, Bhatt RR, Lerner RA. Combinatorial antibody libraries from survivors of the Turkish H5N1 avian influenza outbreak reveal virus neutralization strategies. Proc Natl Acad Sci USA 2008; 105(16): 5986–5991

[55] Thornburg NJ, Zhang H, Bangaru S, Sapparapu G, Kose N, Lampley RM, Bombardi RG, Yu Y, Graham S, Branchizio A, Yoder SM, Rock MT, Creech CB, Edwards KM, Lee D, Li S, Wilson IA, García-Sastre A, Albrecht RA, Crowe JE Jr. H7N9 influenza virus neutralizing antibodies that possess few somatic mutations. J Clin Invest 2016; 126(4): 1482–1494

[56] He W, Mullarkey CE, Duty JA, Moran TM, Palese P, Miller MS. Broadly neutralizing anti-influenza virus antibodies: enhancement of neutralizing potency in polyclonal mixtures and IgA backbones. J Virol 2015; 89(7): 3610–3618

[57] DiLillo DJ, Tan GS, Palese P, Ravetch JV. Broadly neutralizing hemagglutinin stalk-specific antibodies require FcgR interactions for protection against influenza virus in vivo. Nat Med 2014; 20(2): 143–151

[58] Andrews SF, Joyce MG, Chambers MJ, Gillespie RA, Kanekiyo M, Leung K, Yang ES, Tsybovsky Y, Wheatley AK, Crank MC, Boyington JC, Prabhakaran MS, Narpala SR, Chen X, Bailer RT, Chen G, Coates E, Kwong PD, Koup RA, Mascola JR, Graham BS, Ledgerwood JE, McDermott AB. Preferential induction of cross-group influenza A hemagglutinin stem-specific memory B cells after H7N9 immunization in humans. Sci Immunol 2017; 2(13): eaan2676

RIGHTS & PERMISSIONS

Higher Education Press and Springer-Verlag GmbH Germany