EPITOPES

CROSS REFERENCE TO RELATED APPLICATIONS

[0001] This application claims the benefit of U.S. Provisional Patent Application No. 61/457,028, filed December 13, 2010; and U.S. Provisional Patent Application No.

61/626,792, filed October 3, 201 1.

BACKGROUND OF THE INVENTION

[0002] The present invention relates to a method of producing antigens that elicit an immune response to conserved epitopes and is therefore applicable to pathogens for which the primary immune response is directed at variable epitopes. Such a method is especially applicable to influenza vaccines. Accordingly, the invention also provides a universal vaccine against influenza.

[0003] A vaccine is designed to induce an immune response that recognizes a pathogen (or pathogen virulence factors) and thereby prevents or mitigates disease. The choice of antigens is, therefore, important. An immune response against surface exposed antigens is typically most effective against an infection. At the same time, because of this immune response, such surface exposed antigens are under constant evolutionary pressure to evolve and evade the immune system. Thus, a vaccine that elicits an immune response against a specific strain of pathogen may be extremely effective against that strain, but poorly effective against variant strains. To account for the evolution of virulent strains, the vaccine maker may therefore have to target multiple antigens, add new antigens as the pathogen evolves, or target conserved antigens.

[0004] A separate problem in vaccine design is that some epitopes elicit an undesirable immune response. For example, inducing non-neutralizing antibodies can enhance Fc- mediated infection of macrophages, which is the mechanism behind Dengue shock syndrome. Another problem is the induction of an immune response that cross reacts with host antigens. The most famous of these is Guillain— Barre syndrome which is associated with Campylobacter infection, but is also associated with influenza infection. Guillain- Barre syndrome was a reported side-effect of the 1976 swine flu vaccination program. Accordingly, the selection of epitopes for vaccines is far from routine. [0005] Influenza is well known for rapidly evolving different strains, requiring new vaccines every season. Influenza A causes seasonal epidemics affecting millions every year and resulting in the death of between 250,000 and 500,000 people every year, with up to millions in some pandemic years, according to WHO. These seasonal epidemics and pandemics arise because of the constant evolution of the virus both through mutations ("antigenic drift") and through genetic reassortment that occurs when two different influenza viruses infect the same cell ("antigenic shift"). Such reassortment is greatly enhanced by the ability of influenza A to infect a variety of host species, including birds, humans, and other animals, notably pigs. Thus, recombination between two or more viruses, with different primary hosts, may result in novel and highly pathogenic strains that are responsible for the great influenza pandemics.

[0006] Among Avian H5N1 influenzas, for example, there is concern that a human- adapted H5 influenza virus will evolve by mutational (genetic drift) and/or reassortment (genetic shift) mechanisms, to cause a catastrophic pandemic. It is believed that the virus that causes the pandemic will derive from H5 influenzas that are circulating in birds today, but differ from them in ways that are impossible to predict. Therefore, not only is there interest in producing vaccines against the circulating strains of H5, there is also interest in developing vaccines that would not be restricted by inherent strain-specificity.

[0007] Such "universal vaccines" target conserved and evolutionarily stable viral epitopes, rather than the continuously changing hemagglutinin (HA) and neuraminidase (N) epitopes targeted by seasonal flu vaccines (Gerhard, W et al. Prospects for Universal Influenza Virus Vaccine. Emerging Infectious Diseases, 2006. 12: p. 569., Subbarao, K, et al., Development of effective vaccines against pandemic influenza. Immunity, 2006. 24(1 ): p. 5-9.). Universal flu vaccines to date have focused on the highly conserved M2 and NP proteins (Kaiser, J., A One-Size-Fits-All Flu Vaccine. Science, 2006, 312:380). However, M2 and NP proteins are not abundant or easily accessible on the surface of infecting virions and the immune responses to M2 and NP do not directly prevent infection. Thus, an antibody response against M2 and NP is greatly inferior to that obtained by the standard seasonal influenza vaccine.

[0008] Haemagglutinin is abundant and surface exposed, and is a primary target of the immune response against the standard influenza vaccine. However, the HA molecule is highly variant, and the immune response to HA is overwhelmingly driven against the hypervariable regions of HA. Thus, in traditional influenza vaccination or natural infections, the protective humoral immune response is overwhelmingly directed at a limited number of continuously evolving, strain-specific, primary antigenic determinants on the surface of the influenza hemagglutinin, and there is minimal cross reaction with or protection against other serotypes of influenza. This creates a barrier to a "universal vaccine" as vaccination strategies are typically predicated on mimicking natural protective immunity.

[0009] There is now evidence of a weaker and more broadly protective type of

"heterotypic" immunity, which is not based on the response to primary antigenic determinants, but instead derives from responses to conserved viral antigens. It is now thought that heterotypic influenza protection does occur at low levels in human populations.

[0010] For example, a heterosubtypic response to seasonal influenza vaccine can be observed by isolating B-cells that produce antibodies that bind to conserved epitopes (Corti, D., et al., Heterosubtypic neutralizing antibodies are produced by individuals immunized with a seasonal influenza vaccine. J Clin Invest, 2010. 120(5): p. 1663- 1673). Natural infection can induce heterosubtypic antibodies that are cross protective, but only at very low titre (Sullivan, J.S., et al., Heterosubtypic antiavian H5N1 influenza antibodies in intravenous immunoglobulins from globally separate populations protect against H5N1 infection in cell culture. J Mol Genet Med, 2009. 3(2): p. 217 - 24; see also Sui, J., et al., Wide prevalence of heterosubtypic broadly neutralizing human antiinfluenza A antibodies. Clin Infect Dis, 201 1. 52(8): p. 1003-1009; Wrammert, J., et al., Broadly crossreactive antibodies dominate the human B cell response against 2009 pandemic H1N1 influenza virus infection. J Exp Med, 201 1. 208(1): p. 181-193). Epidemiological data collected before and during the 1957 flu pandemic suggested that heterosubtypic immunity to HA may be observed in adults but not in children (Epstein, S., Prior H1N1 influenza infection and susceptibility of Cleveland Family Study participants during the H2N2 pandemic of 1957: an experiment of nature. J Infect Dis., 2006. 193: p. 49-53.), and raises the possibility that elicitation of protective heterotypic responses may prove effective against avian influenza viruses.

[0011] More recent studies have advanced the concept of "seasoned" immunity. Through multiple infections with different strains, a "seasoned" response to conserved epitopes may be observed. (Lynch, G. W., et al., Seasoned adaptive antibody immunity for highly pathogenic pandemic influenza in humans. Immunol Cell Biol, 201 1 , pp 1 -10, Wrammert et al. J Exp Med, 201 1. 208(1 ): p. 1 81 -193). This response, while low, is sufficient to provide some degree of protection from heterotypic and heterosubtypic infection, and explains the greater heterosubtypic immunity observed in adults than in children. It is to be emphasized that the immunity offered by heterotypic and heterosubtypic immunity can be observed as a lower morbidity, mortality and viral shedding, but it is far inferior to the homotypic immunity usually obtained by standard vaccination or infection.

[0012] Heterotypic immunity has also been demonstrated by passive administration of a monoclonal antibody (CI 79) that recognizes a conserved conformational epitope on the hemagglutinin stem consisting of HAl 318-322 and HA2 47-58. CI 79 reduced the severity of illness and death rate in mice infected with HI , H2 or H5 influenzas (Okuno, Y., et al., A common neutralizing epitope conserved between the hemagglutinins of influenza A virus HI and H2 strains. J Virol., 1993. 67: p. 2552-8.; Okuno, Y., et al., Protection against the mouse-adapted A/FM/1/47 strain of influenza A virus in mice by a monoclonal antibody with cross-neutralizing activity among HI and H2 strains. J Virol., 1994. 68: p. 517-20.; Smirnov, Y., et al., Prevention and treatment of bronchopneumonia in mice caused by mouse-adapted variant of avian H5N2 influenza A virus using monoclonal antibody against conserved epitope in the HA stem region. Arch Virol., 2000. 145: p. 1733— 1.).

[0013] Recent attempts to create a universal vaccine have focused on eliciting an immune response against the stem/stalk domain. For example, Steel et a\.(Influenza virus vaccine based on the conserved hemagglutinin stalk domain, mBio 1 (1 ): 1 -9 (April 2010)) describes vaccination with a "headless" HA molecule to drive an immune response against the stalk domain of HA. Wei et al. (Induction of broadly neutralizing HINI influenza antibodies by vaccination, Science 329: 2060-2064 (27 August 2010, e-pub 15 July 2010)) describes how immunization with a DNA vector expressing H1N 1 HA and then boosting with H1N1 seasonal vaccine or replication defective adenovirus 5 vector encoding HA stimulated the production of broadly neutralizing antibodies that recognize HI from diverse HI isolates, with some cross-neutralization of H3 and H5. Further analysis indicated that the immune response was directed against stem antigens. Other research in this area has also been reported. (Bommakanti et al., Design of an HA2based Escherichia coli expressed influenza immunogen that protects mice from pathogenic challenge. Proc Natl Acad Sci U S A, 2010. 107(31): p. 13701-6; Wang et al. Vaccination with a synthetic peptide from the influenza virus hemagglutinin provides protection against distinct viral subtypes. Proc Natl Acad Sci U S A, 2010. 107(44): p. 18979-84.) [0014] However, such "stem" or "headless" vaccines miss other conserved epitopes, such as those that exist on the head (Khurana, Antigenic fingerprinting of H5N1 avian influenza using convalescent sera and monoclonal antibodies reveals potential vaccine and diagnostic targets. PLoS Med, 2009. 6(4): p. el 000049; rause et a\., A broadly neutralizing human monoclonal antibody that recognizes a conserved, novel epitope on the globular head of influenza H1N1 virus hemagglutinin. J Virol, 201 1. pmid_21849447; Whittle, et al., Broadly neutralizing human antibody that recognizes the receptorbinding pocket of influenza virus hemagglutinin. Proc Natl Acad Sci U S A, 201 1. 108(34): p. 1421621 ; Yoshida, et al., Crossprotective potential of a novel monoclonal antibody directed against antigenic site B of the hemagglutinin of influenza A viruses. PLoS Pathog, 2009. 5(3): p. e l 000350). Antibodies to the head domain block hemagglutination, and therefore should restrict access to the receptor binding site, and therefore preventing infection via interference with viron binding to host cell sialic acid receptors.

[0015] Another site outside the stem region is the cleavage site between the HA1 and HA2 domains of HA. This region is highly conserved between influenza A and B hemagglutinin precursors, and peptide conjugate vaccines with sequences from the highly conserved maturational HA1 HA2 elicited broadly protective immune responses against lethal challenge from other A and B influenzas (Bianchi, E., et al., Universal influenza B vaccine based on the maturational cleavage site of the hemagglutinin precursor. J Virol., 2005. 79: p. 7380-8., 14., Horvath, A., et al., A hemagglutinin-based multipeptide construct elicits enhanced protective immune response in mice against influenza A virus infection. Immunol Lett., 1998. 60: p. 127-36.).

[0016] Given that conserved epitopes that mediate broad neutralization are present on the HA head, as well as its stem, vaccine antigens comprised of entire trimeric hemagglutinins, rather than only the stem, or mimeitcs of selected broadly neutralizing epitopes, should offer the greatest opportunity of heterosubtypic protection.

[0017] The challenge of generating an immune response against the conserved epitopes on the head is that such conserved epitopes are structurally linked to the variable regions that are antigenically dominant. The immunodominant regions cannot be merely removed, however.

[0018] Epitopes on the surface of proteins are almost always discontinuous and conformation dependent (Barlow DJ, et al., Continuous and discontinuous protein antigenic determinants, Nature 1986; 322:747-748). Therefore, merely deleting the immunodominant region alters the structure of the head and thus the structure of the conserved epitope. By contrast, immunization against the stem region is less problematic because the entire stem may be used.

[0019] The challenge remains to generate through vaccination an immune response against conserved antigens, that is at sufficient titre to offer meaningful protection.

SUMMARY OF THE INVENTION

[0020] The invention provides a method of reducing the immune response to an epitope while retaining protein structure, comprising: (a) identifying amino acids high on the log odds relative propensity scale (LODrps, a measure of the likelihood of an amino acid being part of an epitope) and (b) replacing at least one high LODrps amino acid with a low LODrps amino acid. By reducing or ablating the immune response to a

primary/immunodominant epitope, the immune response is directed against secondary epitopes, including conserved epitopes that are weakly immunogenic.

[0021] In a related embodiment, the invention provides a method of making a vaccine that elicits an immune response against conserved epitopes on a protein antigen, comprising:

(a) identifying a primary immunodominant epitope in the antigen;

(b) replacing at least one high LODrps amino acid in the primary immunodominant epitope with a low LODrps amino acid thereby significantly eliminating the antigenicity of the primary immunodominant epitope, to create a modified antigen; wherein the modified antigen induces antibodies against conserved epitopes. In some embodiments, at least one high LODrps amino acid from each primary immunodominant epitope is replaced with a low LODrps amino acid.

[0022] The method of invention is suitable for the manufacture of a vaccine. In related embodiments, the vaccine is used for immunization against a disease by administration of an antigen as described herein. In some embodiments, the prime and boost antigens are different.

[0023] In other embodiments, the invention includes a modified protein antigen in which a primary immunodominant epitope in the native protein antigen is modified by replacement of at least one high LODrps amino acid with a low LODrps amino acid, thereby significantly eliminating the antigenicity of the primary immunodominant epitope. This embodiment may be usefully applied to influenza antigens, such as HA

[0024] Identification of the primary epitopes in influenza haemagglutinin (HA), and replacement of amino acids high on the log odds ratio propensity scale (LODrps) with corresponding low LODrps amino acids produces an HA molecule which induces antibody responses to conserved HA residues. Such modified HA molecules are suitable for a broadly neutralizing vaccine against influenza. Accordingly, the invention concerns an influenza haemagglutinin antigen in which all primary epitopes are modified to reduce antigenicity. In some embodiments, the haemagglutinin antigen is an H5 haemagglutinin.

[0025] Representative modifications include at Prol 25, Serl 29, Glu l 31 , Prol 40, Gln l 42, Lysl44, Serl45, Lys l 56, Lys l 57, Asnl 58, Thrl60, Argl 66, Aspl 87, and/or Lys l93 (H3 numbering). In some embodiments, the invention is an HA having one, two, three, four, five, six, seven, eight, nine, ten or all of these residues modified. Relatedly, the invention includes an influenza haemagglutinin antigen having a sequence at least 90%, 95, 98 and 99% identical with the HA portion found in any of SEQ ID NOs: 2-10.

[0026] Other suitable modifications may be grouped according to domains: dl (P140, Q142, K144); d2 (K156, K157, N158), d3 (E131), d4 (D187, K193) and d5 (P125, R166). In some embodiments, one to eleven of these amino acids are mutated. In another embodiment, mutagenesis of all of d l-d5 may require only a single mutation in each domain.

[0027] Amino acids are replaced with low LODrps amino acids, such as alanine or threonine.

[0028] In related embodiments, the invention includes a vaccine, comprising one or more modified haemagglutinins and a pharmaceutically acceptable carrier. In some embodiments the haemagglutinins are proteins. Such proteins may be administered directly, or attached to carrier such as a virus-like particle, incorporated into a replication-defective viral particle or inactivated virus. In other embodiments, the vaccine is in the form of a nucleic acid (DNA, RNA, etc) which is administered to a subject, whereupon the haemagglutinin is expressed. "DNA immunization" is well known in the art. [0029] The invention also includes a method of making an influenza vaccine, comprising

(a) identifying the primary epitopes on the HA molecule;

(b) replacing high LODrps amino acids in one or more primary epitope with low LODrps amino acids;

wherein the vaccine induces neutralizing antibodies that are cross-protective against distantly related HA molecules.

[0030] The invention also includes related methods of using the compositions of the invention. Accordingly, the invention includes a method of immunizing a subject, comprising administration of one or more doses of a vaccine made as described herein. In related embodiments, the subject is immunized with a composition comprising an antigen with an HA having modifications at Prol25, Ser 126, Serl29, Glul31 , Prol40, Glnl42, Lys l 44, Serl45, Lysl 56, Lysl 57, Asnl 58, Thrl 60, Argl 66, Aspl 87, and/or Lysl93; an antigen with an HA having modifications at Serl 26, Serl 29, Glul 31 , Prol40, Gln l42, Lys l44, Serl45, Lys l 56, Thrl 60, Argl 66, Aspl 87, and/or Lys l93, or an influenza haemagglutinin antigen having a sequence at least 90% identical with the HA portion found in any of SEQ ID NOs: 2-10. An immunization protocol may include immunization with more than one xHA together in a single dose, multiple doses, and multiple doses with different xHAs in each dose. An immunization protocol may include immunization with xHA and an adjuvant. Adjuvants, in the present context, include cytokines and other immunomodulatory molecules such as TLR (toll like receptor) agonists and their derivatives that stimulate the immune response.

[0031] In other embodiments, the xHA molecules may be bound to a structure that enhances the immune response, such as a virus-like particle, an immunostimulatory molecule (e.g. Tetanus toxin fragment), a dendrimer, and the like.

[0032] The HA molecules of the present invention are also suitable for the elicitation of antibodies that are broadly cross protective, including polyclonal and monoclonal antibodies. Such antibodies can provide "passive immunity" against infection and/or treatment of infected individuals. For example, such antibodies can be obtained, purified, concentrated, and stored. Accordingly, the invention includes methods of obtaining antibodies against the modified HA molecules, and antibodies obtained by immunization with such molecules. [0033] The methods and compositions described herein are suitable for generation of an immune response against influenza viruses, especially influenza A, and are therefore used in a vaccine against influenza A infection.

BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Figure 1 shows the 2.9 A 2FKO.pdb structure (Stevens, J., et al., Structure and receptor specificity of the hemagglutinin from an H5N1 influenza virus.. Science, 2006. 312: p. 404-10.) of the hemagglutinin from the A/Viet/1203/04 H5 influenza virus, with its monomeric subunits drawn in 3 different shades of gray.

[0035] The membrane-distal "head" of the trimeric HA is at the top of the illustrated molecule, and contains receptor binding and primary antigenic determinant structures. The membrane-proximal "stem" is at the bottom, and contains conserved fusion peptide and HA1/HA2 cleavage site sequences.

[0036] To create xHA antigens and shift the immune response from "head" primary antigenic determinants to conserved HA structures (including fusion-peptide CR6261 epitopes {white) on the HA stem (Ekiert, D. C, G. Bhabha, et al. (2009). "Antibody recognition of a highly conserved influenza virus epitope." Science 324(5924): 246-51.)), up to 1 1 high LODrps amino acids in primary antigenic determinant(s) {black) were replaced with amino acids having lower Discotope LODrps values.

[0037] Figure 2 illustrates the Drosophila Expression System (DES) (InVitrogen) construct for production of recombinant VI 203 hemagglutinin and xHA production.

[0038] Figure 3. Purified recombinant xHAs bearing mannosylated oligosaccharides for immune presentation a, Reduced and Non-reduced SDS-PAGE of purified xHAs (200 ng/lane) c, GNA lectin blot of duplicate of gel in a. GNA binds to terminal mannose d, Undigested (O) and PNGase F (P) digested xHAs. left, sypro red stained gel; right, GNA- stained blot.

[0039] Figure 4. Sequences of the parent hemagglutinin and xHAs.

[0040] The xHA.s were expressed in the Drosophila Expression System (InVitrogen) using the pmtbipv5hisa vector as shown in Fig. 2. Expressed polypeptides were comprised of: (1 ) the BIP signal sequence from pmtbipv5hisa, (2) followed by the dipeptide Arg-Ser from a Bgl2 linker sequence, (3) followed by HA0 encoding sequences incorporating (a) substitutions in primary antigenic determinant sites as indicated in Table 3, and (b) substitution of a "T" rather the "RRRKK"(R) sequence at the HA 1 /HA2 cleavage site of wildtype V I 203, (4) followed by a polypeptide encoding a thrombin cleavage site, foldon sequence, and hexa-histidine sequence tag, after the KREEIS sequence of HA2.

[0041] The expressed parental and xHA hemagglutinin sequences are shown with the BIP signal sequence in three letter code, primary antigenic substitutions primary antigenic determinant substitutions indicated as bold underlined residues, and C-terminal thrombin- foldon-hexaHis polypeptide in bold. Italics indicate Xba \-Kpn I fragment used for rapid construction of the xHA variants.

[0042] Figure 5: Immunoassays to verify proper folding of recombinant hemagglutinins and primary antigenic determinant knock out in xHAs.

[0043] Figure 6 A-D Polyclonal antisera to xHAs contain antibodies that compete for binding of a conserved fusion peptide-containing epitope on the HA stem. Panel A shows no competition by mouse non-immune serum (m NI, open circles), and competitive, concentration-dependent reduction of 1 F02 binding following exposure to the positive control CI 79 mAb (solid triangles). Each of the plots in the remaining B - D panels presents 1 F02 competition results from serum samples obtained at various stages during the immunization of a single animal with xHAs. The designation "pr" indicates antisera obtained after priming, "b l " and "b2" are respectively antisera obtained after boosts 1 and 2, etc. The animals in panel B were primed and boosted with xHA.par (10 ug/injection). The animals in panels C and D, were primed with xHA.4b, then repeatedly boosted with xHA.5b; those in C received 10 ug doses of the xHA antigens, while those in D received 20 ug doses.

[0044] Fig. 7 Hemagglutination inhibition (HAI) by antisera to xHAs.

[0045] Hemaggutination inhibition assays were based on the standard W.H.O. kit protocol for Hemagglutination testing. Antisera were treated overnight at 37°C with 3 volumes of receptor destroying enzyme (RDE, Denka), which was subsequently inactivated for 30m at 56 °C. Hemagglutination microplate wells were loaded with 25 ul of RDE-treated samples representing overall 18, 36 and 72-fold dilutions of the antisera, or with PBS. The four HA- pseudotyped lentviruses used for Fig.2 HA neutralization assays were employed as 'antigens' in the HA inhibition assays. The LV antigen stocks were adjusted to 8 HA units/ 50 ul, and 25 ul (4 HA units) added to the antiserum- and PBS-containing wells. After 30 minutes of incubation, 50 ul of glutaraldehyde fixed 0.5% turkey RBCs (Fitzgerald) were added to the antiserum+LV incubations, and plates were photographed 40 min later.

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0046] Because the primary immune response to influenza is directed against

continuously evolving primary antigenic determinant structures on the hemagglutinin, the inventor has constructed a series of HA antigens in which these primary epitopes have been modified to ablate immunogenicity but keep overall structure of the HA. The resulting antigen molecule(s) thereby stimulate the production of antibodies against secondary antigenic determinants, which are epitopes that do not map to the primary antigenic determinants. Antibodies against conserved elements are broadly neutralizing.

[0047] To reduce uncertainty, the following definitions are used throughout:

[0048] "About" is used as understood by the person of ordinary skill in the context of the variable to which "about " is applied. When in doubt, "about" indicates a variation of ± 10% of the stated value.

[0049] "Haemagglutinin" or "HA" refers to the Influenza haemagglutinin protein. In certain embodiments, this is the influenza A haemagglutinin, and recombinant variants thereof. In exemplary embodiments, the haemagglutinin is H5, or derived from H5. The term is used to describe a family of proteins, without regard to whether the protein actually possesses the property of haemagglutination. HAs are traditionally grouped by serotyping, into HI , H2, H3, H4, H5 and the like. The serotypes reflect antigenic and genetic variation. There is also variation within a given serotype, but this is less than between serotypes.

[0050] "xHA" refers to an HA for which one or more primary antigenic determinants are modified. "xHA.par" refers to the recombinant parental control hemagglutinin, which in the exemplary case is the HA from A/Vietnam/ 1203/2004 with the describe modification of the HA l/HA2 cleavage site.

[0051] "Antigenic determinants", or "epitopes," are structures recognized by antibodies and T-cell receptors of the immune system. Preferably, such epitopes are antibody epitopes. In the case of hemagglutinin, a small number of structures on the ectodomain (head) surface induce antibodies much more readily than the rest of the molecule and are referred to as being "immunodominant". [0052] "Primary antigenic determinants" are synonymous with "immunodominant epitopes" and are those to which the immune response is primarily directed. If expressed numerically, an immune response that is at least one order of magnitude greater to a given epitope would indicate that it is immunodominant. In the context of influenza HA, those epitopes in the highly variable regions of the HA head to which the immune system normally develops the strongest antibody response, and which are recognized following infection or immunization with conventional seasonal flu vaccines, are the "primary antigenic determinants." These determinants in influenza are constantly evolving.

[0053] "Secondary antigenic determinants" are not immunodominant. In the context of the present invention, secondary antigenic determinants are those epitopes that are recognized after ablation of the primary antigenic determinants. Secondary antigenic determinants may be located in the conserved regions of the HA, which occur both on the HA head and stem. "Ablation" of antigenic determinants traditionally occurred by deletion of the entire epitope but is meant here to indicate ablation of antigenicity, by substitution of amino acids. Following from the immunodominant terminology, Secondary antigenic determinants may also be referred to as "immunorecessive."

[0054] "Escape mutant" refers to a derivative influenza virus that does not bind to, and/or is not neutralized by, a particular antibody or antiserum. As used herein, "escape mutant" refers to a virus containing a mutation in an epitope targeted by a neutralizing antibody.

[0055] "LODrps" is the "log odds ratio propensity scale" which measures the natural logarithm of the odds-ratio of a given amino acid to be present on the antigen side of an antigen-antibody interaction, and is therefore a measure of the likelihood of being an epitope. A high LODrps value means that the amino acid is over-represented in the set of structurally well-defined epitopes, whereas under-represented amino acids have low log- odds ratio values. The term LODrps is described in greater detail below.

[0056] A "vaccine" describes a preparation designed to induce an immune response that is protective against disease. In the present context, a vaccine induces an immune response against influenza virus. A vaccine may be prophylactic or preventative, given prior to or shortly after exposure to an influenza virus; or therapeutic, given during infection to boost the immune response or drive the response in a specific direction. A vaccine does not have to induce a fully protective response that prevents all disease, as not all vaccines produce an immune response in all people, and the strength and nature of the immune response varies between people.

[0057] "Antibody" as used herein encompasses natural antibodies, chimeric and recombinant antibodies, and antibody fragments, such as Fab, scFv, and the like.

[0058] An "adjuvant" increases the immune response against an antigen with which it is presented. Adjuvants are known in the art and include aluminium hydroxide,

monophosphoryl lipid A, oils, cytokines, toll like receptor agonists, and the like.

[00591 Influenza hemagglutinin molecular structure HA is the major surface protein of the virus as well as the major viral target of neutralizing antibodies. The 2.9 A

2FKO.pdb structure (Stevens, J., et al., Structure and receptor specificity of the

hemagglutinin from an H5N1 influenza virus.. Science, 2006. 312 : p. 404- 10) of the hemagglutinin from the A Viet/1203/04 H5 influenza virus is shown in Fig. 1. It is a homotrimer with a large head (ectodomain) comprised of prominent beta sheet structures, and a stem (stalk) composed of long alpha helices. The base of the stem anchors the hemagglutinin in the viral membrane (or cellular membrane prior to budding), while its ectodomain head is exposed on the surface of the virus. HA plays an essential role in infection and the viral life cycle by (i) presenting binding sites for cellular receptors on its ectodomain surface and (ii) mediating fusion of viral and host-cell membranes to permit cellular entry of the genome-transcriptase complex following endocytosis.

[0060] Monomer subunits of the hemagglutinin trimer are synthesized as large HA0 precursor molecules, which are cleaved by host proteases at a site in the stem to generate HA1 and HA2 fragments of about 300 and 200 residues, respectively. The head is composed of HA1 sequences only, while the stem is a structurally complex structure containing entwined HA 1 and HA2 sequences.

[0061] Naturally contracted influenza infections and seasonal flu vaccines elicit antibodies which bind primary antigenic determinant epitopes on the HA head. The primary antigenic determinants are located adjacent to the receptor binding sites. Accordingly, antibodies to primary antigenic determinants are neutralizing and block infection by restricting host cell sialic acid receptor access to the HA receptor binding site. However, the protective function of these antibodies is short-lived due to rapid evolution of primary antigenic determinants under selective pressure. [0062] The primary antigenic determinants on HA have been identified by sequencing "escape mutants" selected with neutralizing antibodies, and correspond to hot spots of sequence variation during virus evolution (Stevens, J., et al., Structure and receptor specificity of the hemagglutinin from an H5N1 influenza virus.. Science, 2006. 312: p. 404- 10., Kaverin, N:, et al., Structure of antigenic sites on the haemagglutinin molecule of H5 avian influenza virus and phenotypic variation of escape mutants. J Gen Virol., 2002. 83 : p. 2497-505., Kaverin, N.V., et al., Epitope mapping of the hemagglutinin molecule of a highly pathogenic H5N1 influenza virus by using monoclonal antibodies. J Virol, 2007. 81(23): p. 1291 1 -7., Philpott, M., et al., Hemagglutinin mutations related to attenuation and altered cell tropism of avirulent avian influenza A virus. J Virol., 1990. 64: p. 2941-7., Wiley, D., I. Wilson, and J. Skehel, Structural identification of the antibody-binding sites of Hong Kong influenza haemagglutinin and their involvement in antigenic variation. Nature, 1981. 289: p. 373-8.). Figure 1 shows the locations (black) of H5 HA primary antigenic determinants identified using avian H5 escape mutant data and sequence information from human H5 drift isolates obtained during 1997 - 2004. The primary antigenic determinants surround and overlap the receptor binding domain (RBD) (Stevens, J., et al., Structure and receptor specificity of the hemagglutinin from an H5N1 influenza virus. Science, 2006. 312: p. 404- 10., Weis, W., et al., Structure of the influenza virus haemagglutinin complexed with its receptor, sialic acid. Nature, 1988. 333: p. 426-31.), which mediates binding of the virus to the cell during infection. A RBD and adjacent primary antigenic determinants are present on each monomer of the HA trimer.

[0063] In addition to the well known, narrowly-focused antibodies to the continuously evolving primary antigenic determinants on the HA head, there are also broadly-protective antibodies that recognize evolutionarily conserved and functionally critical structures located on HA stem and head surfaces.

[0064] For example, mAbs C179, CR6261 , F10 and 1 F02 (see, e.g., Okuno J Virol., 1993. 67: p. 2552-8; Ekiert et al. Science.324 (2009); Sui Virology 387: 473-481 (2009);

Wrammert et al. J Exp Med, 201 1. 208(1 ): 181 -93) recognize fusion peptide-containing HA stem epitopes that are conserved in Group 1 influenzas, which include the HI , H2, H5 subtypes. Similarly, mAb CR8020 recognizes a fusion peptide-containing HA stem epitope conserved in Group 2 influenzas, which includes the H3 and H7 subtypes (Ekiert et al. Science. 201 1. 333:843-50). Finally mAb FI6 recognizes fusion peptide-containing HA stem epitopes from both Group 1 and Group 2 viruses (Corti, D. et al., Science, 201 1.

333:850-6).

[0065] In addition to the broadly-protective antibodies recognizing conserved fusion- peptide containing structures located on HA stem, there are also broadly neutralizing antibodies that recognize conserved structures on the HA head .( hurana, Antigenic fingerprinting of H5N1 avian influenza using convalescent sera and monoclonal antibodies reveals potential vaccine and diagnostic targets. PLoS Med, 2009. 6(4): p. el 000049; rause et al, A broadly neutralizing human monoclonal antibody that recognizes a conserved, novel epitope on the globular head of influenza H1N1 virus hemagglutinin. J Virol, 201 1 . pmid_21849447; Whittle et al. Broadly neutralizing human antibody that recognizes the receptorbinding pocket of influenza virus hemagglutinin. Proc Natl Acad Sci U S A, 201 1. 108(34): p. 1421621 ; Yoshida et al., Crossprotective potential of a novel monoclonal antibody directed against antigenic site B of the hemagglutinin of influenza A viruses. PLoS Pathog, 2009. 5(3): p. el 000350).

[0066] The inventor hypothesized that reducing the immunogenicity of H5 HA primary antigenic determinants will increase immune responses against conserved HA epitopes that do not efficiently elicit immunological memory antibodies under routine infection and seasonal flu vaccine immunization conditions. These conserved epitopes are present in a broad spectrum of influenzas, and more likely to be retained in future H5N1 viruses than are the rapidly evolving primary antigenic determinants targeted by conventional vaccines, and therefore should represent superior targets for generating broad H5 and heterosubtypic cross-protection across all HAs.

[0067] B-cell epitope characteristics The H5 HA primary antigenic determinants to be neutralized in this work are defined based on extensive escape mutant and genetic drift data. However, in order to successfully knock them out and avoid their undesirable replacement with novel antigenic determinants, it is necessary to consider the properties of B-cell epitopes. Generally, B-cell epitopes locate to hydrophilic and dynamically flexible sites on a protein's surface (reviewed in Haste Andersen, P., M. Nielsen, and O. Lund, Prediction of residues in discontinuous B-cell epitopes using protein 3D structures. Protein Sci, 2006. 15(1 1): p. 2558-67., Parker, J.M., D. Guo, and R.S. Hodges, New hydrophilicity scale derived from high-performance liquid chromatography peptide retention data: correlation of predicted surface residues with antigenicity and X-ray-derived accessible sites.

Biochemistry, 1986. 25( 19): p. 5425-32., Ponomarenko, J.V. and P.E. Bourne, Antibody- protein interactions: benchmark datasets and prediction tools evaluation. BMC Struct Biol, 2007. 7: p. 64.; Jin L, Fendly BM, Wells JA, High resolution functional analysis of antibody-antigen interactions. J Mol Biol. 1992 Aug 5;226(3):851-65.). The recent exponential growth of antigen-antibody complex structures in the Protein Data Base enables detailed analysis of antigen-antibody contact sites and provides new information about the properties of surface substructures that form epitopes on protein antigens. Table 1 is an epitope log-odds ratio propensity scale (LODrps) derived by analyzing the distribution of amino acids present on the antigen side of antigen-antibody interfaces in 76 different x-ray structures of antigen-antibody complexes (Haste Andersen, P., M. Nielsen, and O. Lund, Prediction of residues in discontinuous B-cell epitopes using protein 3D structures. Protein Sci, 2006. 15(1 1 ): p. 2558-67.).

Table I

epitope log odds ratios

for 20 amino acids

N 1 :242 Y 0.03

R 1.18 W -0.064

P 1 .164 s -0.145

K 1.136 T -0.233

H 1.098 I -0.713

Q 1 .082 F -1 .147

D 0.691 V -1 .474

E 0.346 A -1 .522

M 0.273 L -1 .836

G 0.189 C -3.519

[0068] A high Discotope log-odds value means that the amino acid is over-represented in the set of structurally well-defined epitopes, whereas under-represented amino acids have low log-odds ratio values. The epitope log-odds ratio propensity scale is particularly useful when considered in conjunction with the extensive evolutionary drift and escape mutant mapping data available for influenza hemagglutinins. Of the ten H5 hemagglutinin residues identified as primary antigenic determinants on the basis of drift and escape mutant evidence, i.e., empirically, seven had positive log-odds ratios. [0069] The invention utilizes the epitope log-odds ratio propensity scale in the design of x-HA mutants with neutralized primary antigenic determinants. The strategy for reducing H5 HA primary epitope immunogenicity is to replace primary antigenic determinant residues for which there is strong drift and escape mutant evidence with amino acids that have lower log-odds ratios. This will not only destroy primary epitope(s), but should also reduce the likelihood of the new surface(s) serving as antigenic determinants. Without being bound by theory, it is believed that by disfavoring clonal selection of B cells for primary antigenic determinants, the immune response should be shifted to secondary epitopes, including conserved epitopes which do not normally elicit immune responses due to the immunodominance of HA primary antigenic determinants.

[0070] The invention is further understood by reference to the following examples, which are representative and not limiting.

EXAMPLES

[0071] EXAMPLE 1. HEMAGGLUTININ EXPRESSION AND PURIFICATION

[0072] Recombinant influenza hemagglutinins produced by two different baculovirus- based insect cell expression strategies have been demonstrated to be suitable for vaccine trials (Treanor, J.J., et al., Dose-related safety and immunogenicity of a trivalent baculovirus-expressed influenza-virus hemagglutinin vaccine in elderly adults. J Infect Dis, 2006. 193(9): p. 1223-8.) and for high-resolution receptor binding and structural studies (Stevens, J., et al., Structure and receptor specificity of the hemagglutinin from an H5N1 influenza virus. Science, 2006. 312: p. 404-10., Stevens, J., et al., Structure of the uncleaved human HI hemagglutinin from the extinct 1918 influenza virus. Science, 2004. 303: p. 1866-70.). Accordingly, baculovirus expression systems are appropriate for the expression of HA antigens.

[0073] The constructs of Stevens contain C-terminal foldon sequences to mediate correct folding of the HA, leading to crystal lographically verified trimeric HAs, an improvement over the original baculovirus-expressed Protein Sciences HAs of the Treanor reference. Accordingly, the DES expression approach used, below, is based on the Stevens constructs that contain the foldon domain.

[0074] A different insect cell expression system, the Drosophil Expression System (DES) (InVitrogen) was used for production of control hemagglutinins and xHAs described in the Examples. The DES parent construct for parental H5N 1 A/Vietnam/1203/2004 (V I 203) hemagglutinin production is illustrated in Figure 2. The sequence of the expressed recombinant hemagglutinin is SEQ DD NO: 1 of Figure 4. The DNA sequence encoding this protein was inserted into pMT BiP/V5-His vector, obtained from Invitrogen Corp. The VI 203 HAO precursor sequence was placed downstream of a metallothionein promoter for inducible expression, and a BIP signal peptide to mediate secretion. A thrombin cleavage site and 30-residue foldon and hexahistidine sequences were placed C-terminal to the hemagglutinin-encoding sequence to facilitate HA trimerization, purification and the subsequent removal of the foldon and his tag. The site of HAO cleavage into HA 1 and HA2 was modified (PQRERRRKKRGLFG to PQRETRGLFG) in order to maintain the prefusion conformation, reduce HA cleavage, and promote the production of HAO oligomers and trimers, which exhibit superior immunogenicity (Wei, C.J., et al., Comparative efficacy of neutralizing antibodies elicited by recombinant hemagglutinin proteins from avian H5N1 influenza virus. J Virol, 2008. 82(13): p. 6200-8.). The HA1 region in which the primary antigenic determinants are located is shown in a magnification, with the positions of the eleven primary antigenic determinant residues targeted by our study marked by black X marks. These primary antigenic determinant residues were selected on the basis of escape mutant data, genetic drift data, and Discotope high log odds values, as explained in Example 2, Table 2.

[0075] The HAO DNA sequence of the construct was optimized for expression in the DES Drosophila expression system by employing frequently utilized codons from a data base of highly expressed Drosophila proteins (Shields, D.C., et al., "Silent" sites in Drosophila genes are not neutral: evidence of selection among synonymous codons. Mol Biol Evol, 1988. 5(6): p. 704- 16.). The DNA sequence was further modified to include translationally silent, unique restriction sites at the locations shown in the map. Restriction sites in the magnified primary antigenic determinant region were utilized for rapid, efficient and economic substitution of cassettes with changes for x-HA variants.

[0076] Expression and purification of VI 203 hemagglutinin and related HAs DES- expressed control hemagglutinins and xHAs were purified using the strategy described by Stevens and Wilson (Stevens, J., et al., Structure and receptor specificity of the

hemagglutinin from an H5N1 influenza virus. Science, 2006. 312: p. 404-10., Stevens, J., et al., Structure of the uncleaved human HI hemagglutinin from the extinct 1918 influenza virus. Science, 2004. 303 : p. 1866-70.). This protocol produces HAs that are properly folded with polymerization into trimers resembling those obtained from bromelain-released virus hemagglutinin. Thus, such HAs mimic naturally occurring HAs.

[0077] The expression system used in the present examples features a thrombin cleavage site C-terminal to the HAO sequence, followed by a "foldon" sequence to promote efficient assembly of the trimer (Frank, S., et al., Stabilization of short collagen-like triple helices by protein engineering. J Mol Biol., 2001. 308: p. 1081-9.), and finally a hexa His-tag to facilitate protein purification. Expression plasmids were DNA sequence verified for the region encoding the signal sequence cleavage site through the His-tag and stop codon, then co-transfected with pCoHygro (InVitrogen) or pCoPuro (Iwaki, T., et al., Rapid selection of Drosophila S2 cells with the puromycin resistance gene. Biotechniques, 2003. 35(3): p. 482-4, 486) selection plasmids into Drosophila S2 cells. Stably transformed

hemagglutinin-expressing cell lines were selected by growth in the presence of hygromycin or puromycin for several weeks.

[0078] Stably transfected, hemagglutinin-expressing S2 cell lines were expanded up to 500 ml in 10% serum-containing Express Five media (InVitrogen). Then cultures were adapted to serum-depleted conditions by 1 : 1 dilution with serum-free media up to 2 liters. HA expression from the metallothionein promoter was induced by addition of CuSC>4 , and conditioned medium harvested 3-4 days later at trypan blue cell viabilities of 80-90%. Conditioned media supernatants were prepared, and frozen at -20C for cryoprecipitation of insect ferritin and storage. The xHAs were purified by Ni-NTA (Qiagen) chromatography with imidazole gradient elution. Peak fractions, identified by SDS-PAGE, were pooled, buffer exchanged and concentrated into 50 mM NaCl, 10 mM Tris pH 8. Protein concentration was measured by Coomassie Blue dye binding (Biorad). The xHAs used in this study were not thrombin digested, and retained the C-terminal foldon/his-tag sequence to promote recovery of trimeric and oligomeric hemagglutinins, which are efficient immunogens (Wei et al, J. Virol., 2008, 82:6200). Yields of -0.5 mg or more of hemagglutinin per liter of induced cells were obtained. Reducing and non-reducing SDS- PAGE was performed to assess purity and confirm HAO status, and the content of monomeric, trimeric and multimeric xHAs assessed by size exclusion chromatography on Sephacryl-S300 HR.

[0079] The above DES expression and purification protocols supported efficient production of 70 kD HAO parental VI 203 HA and xHAs, which are assembled into trimeric hemagglutinins. Figure 3 panel a and b show reduced and non-reduced SDS-PAGE gels, demonstrating the high level of purity of the purified xHA antigens, and the absence of inter-HAO disulfide bonding, which would be indicative of misfolding. Further evidence for the proper folding of the recombinant xHAs was provided by patterns of binding to (1) the 1F02 mAb to the fusion-peptide containing epitope on the HA stem (Wrammert et al, J Exp Med, 201 1. 208: 181) and (2) the VN04 panel of mapped mAb antibodies to V1203 HA primary antigenic determinant continuous and discontinuous epitopes (Kaverin et al., J Virol, 2007, 81 : 1291 1 ). (1 F02 and VN04 mAb binding data is presented below in Example 2, Figure 5.) The purified xHAs migrated on S300 Sephacryl gel exclusion

chromatography as -200 kD trimers and higher MW multimers, as previously described for baculovirus-expressed HAs (Stevens et al., Science, 2004, 303: 1866-70; Wei et al., J. Virol., 2008, 82:6200).

[0080] Production of xHAs antigens in Drosophila S2 cells is believed to results in their N-glycosylation with paucimannose and related glycans (Kim YK, et al., Production and N- glycan analysis of secreted human erythropoietin glycoprotein in stably transfected Drosophila S2 cells. Biotechnol Bioeng. 2005 Nov 20;92(4):452-61). Such glycosylation may be advantageous for efficient presentation of these molecules to the immune system via macrophage, monocyte and dendritic cell mannose receptors (Buzas EI et al, Carbohydrate recognition systems in autoimmunity. Autoimmunity. 2006 Dec;39(8):691-704; Gazi, U. and L. Martinez - Pomares, Influence of the mannose receptor in host immune responses. Immunobiology, 2009. 214(7): p. 554 - 61). Fig. 2c and 2d GNA lectin blotting and PNGase F digestion experiments confirmed the presence of mannose terminated Nglycans on the xHAs.

[0081] These results demonstrate successful Drosophila expression system synthesis and purification of recombinant hemagglutinins for induction of antisera and evaluation of the immune responses.

[0082] EXAMPLE 2. DESIGN AND PRODUCTION OF xHA MUTANTS

[0083] VI 203 primary antigenic determinants were ablated in x-HA variants by substitutions at epitope residues identified on the basis of escape mutant and genetic drift data. These positions were replaced with amino acids that are present at low frequencies on the antigen side of contact sites in a database of antigen - antibody crystal structures.

Knockout of VI 203 specific primary antigenic determinants in xHA variants was verified by screening against a panel of well-characterized monoclonal antibodies (Kaverin et al., J Virol, 2007, 81 : 1291 1) obtained from the Biodefense and Emerging Infections Research (BEIR) repository.

[00841 Mapping of the primary antigenic determinants The first step in the design of the xHAs was to carefully analyze H5 1 antigenic drift and escape mutant data and to develop operational definitions of the primary antigenic determinants that will be knocked out. Table 2 presents the H5 A Viet/1203/04 HA1 ectodomain sequence from residues 125 - 209 (H3 numbering), with primary antigenic determinant residues shaded. Black indicates primary antigenic determinant residues for which differences from the A/Viet/1203/04 sequence have been observed both in laboratory-generated escape mutants and in natural drift isolates (human viruses, 1997-2004). Gray indicates residues for which the assignment as a primary antigenic determinant was made solely on the basis of escape mutant evidence. Light gray indicates residues for which there is only evolutionary drift evidence.

Table 2

MAPPING OF H5 HEMAGGLUTININ PRIMARY ANTIGENIC DETERMINANTS

Table 2 (continued)

Footnotes for Table 2

^{100851 1} H5 A Viet 1203/04 HA residue ID; H3 numbering

^{100861 2} H = Primary antigenic determinant assignment based on drift in 1997 - 2005 H5N1 human isolates, Stevens, 2006, Fig. S5 (Stevens, J., O. Blixt, et al. (2006). "Structure and receptor specificity of the hemagglutinin from an H5N1 influenza virus." Science

312(5772): 404-10.)

|0087| 3 ^ _ p _r j _mar y antigenic determinant assignment based on drift in 2003 - 2006 H5N1 human and avian isolates, Kaverin, 2007 , Table 4 (Kaverin, N. V., I. A. Rudneva, et al. (2007). "Epitope mapping of the hemagglutinin molecule of a highly pathogenic H5N1 influenza virus by using monoclonal antibodies." J Virol 81(23): 1291 1-7.)

loose ^] ^ _ p _r i _mar y antigenic determinant assignment based on escape mutants of Ή5Ν2 A/Mallard/Pennsylvania/10218/84 from various mAbs, Kaverin, 2002, Table 3 (Kaverin, N., et al. (2002). "Structure of antigenic sites on the haemagglutinin molecule of H5 avian influenza virus and phenotypic variation of escape mutants." J Gen Virol. 83: 2497-505. )

^{100891 5} x = Primary antigenic determinant assignment based on escape mutants of H5 A/Vietnam/1203/04 from indicated monoclonal antibody, Kaverin, 2007, Tables 1 and 2

[0090] X = Primary antigenic determinant assignment based on H5N2

A/Mallard/Pennsylvania/10218/84 escape mutants (Kaverin, 2002) with the indicated mAb.

|oo9i | 6 j-j _ predicted using DiscoTope algorithm for the identification of discontinuous B- cell epitopes, Haste Andersen, 2006 www.cbs.dtu.dk/services/DiscoTope

[00 21 ^* Asterisk mark amino acids with epitope log odds ratio values of > 0.3 (Haste Andersen, P., M. Nielsen, et al. (2006). "Prediction of residues in discontinuous B-cell epitopes using protein 3D structures." Protein Sci 15(1 1 ): 2558-67.)

[0093] Shading shows primary antigenic determinant residues. Assignments were based on drift and escape mutant evidence (black), drift-only evidence (light gray), or escape mutant only data (gray)

[0094] The baculovirus expressed H5 A Viet/1203/04 HA of Stevens (2006) is N- glycosylated on Asn-169 but not Asn-158.

[0095] Ablation of primary antigenic determinants Discotope log odds ratio propensity scale (LODrps) values for the 20 amino acids have been assigned based their relative frequencies on the antigen side of the antigen-antibody interfaces in 76 different x- ray structures (Haste Andersen, P., M. Nielsen, and O. Lund, Prediction of residues in discontinuous B-cell epitopes using protein 3D structures. Protein Sci, 2006. 15(1 1 ): p. 2558-67). A high LODrps means that the amino acid is over-represented in known, structurally well-defined epitopes, whereas under-represented amino acids have low LODrps values.

[0096] In the above Table 2 presentation of H5 HA primary antigenic determinant sequences, residues with high LODrps values are marked with asterisks (*). Of the 34 shaded primary antigenic determinant residues in Table 2, 1 1 are surface-exposed and have log-odds ratio values of >0.3. The inventor hypothesizes that replacing these residues with low or negative log-odds ratio amino acids (see Table 3) should not only knock out primary antigenic determinants, but also reduce the likelihood of the modified surfaces also serving as new antigenic determinants.

TABLE 3

DESIGN PLAN FOR PRIMARY ANTIGENIC DETERMINANT KNOCK-OUT IN

CONTROL AND x- HEMAGGLUTININS

[00971 l £ _sca p _{e mu} tant substitutions are shown in bold face for A/Vietnam/ 1203/04 (H5N 1 ) escape mutants, and in italics for A/Mallard/Pennsylvania/10218/84 (H5N2) escape mutants

(00981

H3 numbering is used throughout the Table.

100991

Log odds ratio propensity score (LODrps) value is in parentheses.

⁴ Wildtype H5 A/Viet/1203/04 residues are shown in white cells, and substitution mutations are in gray cells.

loio ⁱ I 5 Lorjrpg index was calculated from sum of Discotope LODrps scores of HA primary antigenic determinant residues. Higher values indicate that primary antigenic determinants are richer in amino acids that occur commonly on the antigen side of antibody-epitope contact sites. Lower values indicate that primary antigenic determinants are composed of amino acids which are under represented on the antigen side of antibody-epitope contact sites.

[01021 x-HA Design Table 3 summarizes the design of H5 x-HA antigens. The parent molecule was HAO from A/Viet/1203/04 (clade 1 ), a highly pathogenic H5N 1 influenza isolated from a Vietnamese patient in 2004. A/Indonesia/5/05 (clade 2. 1.3) and

A/Egypt/2782/06 (clade 2.2) hemagglutinins serve as drift controls. [0103] For a given macromolecular antigen, and especially for influenza hemagglutinins, some substructures stimulate immune responses very effectively, while other substructures do not. The common properties of protein substructures that are effective B-cell antigens include surface location, hydrophilicity and dynamic flexibility (Haste Andersen, P., et al., Prediction of residues in discontinuous B-cell epitopes using protein 3D structures. Protein Sci, 2006. 15(1 1): p. 2558-67., Parker, J.M., D. Guo, and R.S. Hodges, New hydrophilicity scale derived from high-performance liquid chromatography peptide retention data:

correlation of predicted surface residues with antigenicity and X-ray-derived accessible sites. Biochemistry, 1986. 25(19): p. 5425-32., Ponomarenko, J.V. and P.E. Bourne,

Antibody-protein interactions: benchmark datasets and prediction tools evaluation. BMC Struct Biol, 2007. 7: p. 64). Certain amino acids occur more frequently in epitopes compared to others (Jin L et al, J Mol Biol. 1992, 226(3): 851), and this is reflected in their Discotope LODrps values as illustrated in Table 1 (Haste Andersen, P, Protein Sci, 2006. 15(1 1 ): p. 2558-67).

[0104] Accordingly, x-HAs were generated by altering the steric shapes of primary antigenic determinants towards reducing their immunogenicity. Replacement amino acids are low or negative LODrps residues that occur rarely in stable antigen-antibody interfaces. Primary antigenic determinant ablation is performed by replacing primary antigenic determinant residues identified by the drift and escape mutant data of Table 2 with amino acids that have lower LODrps values according to the scheme detailed in Table 3.

Reducing the tendency of antibodies to bind primary antigenic determinant sites by replacement of positive LODrps amino acids with low or negative LODrps amino acids should prevent clonal expansion of B-cells producing antibodies to the original primary antigenic determinants as well as the "knocked out" sites, and thereby facilitate expansion of B-cells making antibodies to conserved, normally "immunorecessive" sites on the hemagglutinin molecule that do not routinely elicit detectable immune responses.

[0105] During the design process, it is important to avoid changes that induce protein- folding problems in regions distal to the primary antigenic determinants, because these regions may contain structures that will become "secondary" epitopes. Therefore, modifications were limited to the confines of well-defined primary antigenic determinants. The X-HA.4 and xHA.5 variants, have all 5 primary antigenic determinants knocked out, but by different combinations of amino acid substitutions (right side of Table 3). In theory, ablation of all 5 epitopes should direct the immune response to conserved epitopes. In contrast, the x-HA. l , x-HA.2, and x-HA.3 hemagglutinins were designed to eliminate, as well as to retain, different subsets of primary antigenic determinants. These serve as experimental controls, but may also elicit immune responses to conserved epitopes.

[0106] For the x-HA.4 series, 6 high LODrps residues in the VI 203 primary antigenic determinants were replaced with escape mutant residues in x-HA.4a, or mainly threonine (LODrps = -0.233) in x-HA.4b, or mainly alanine (LODrps = -1.522) in x-HA.4c. Based on detailed analysis of the 2F 0 trimer structure, escape mutant and drift data, modification of just 6 residues should completely ablate the HA primary antigenic determinants. x-HA.5 series hemagglutinins, with 10 changes in primary antigenic determinant residues should further eliminate residual antigenicity in the primary epitopes.

[0107] Note that 2 residues, which are understood to contribute to the primary antigenic determinants based on escape mutant data, were not altered. These are 1155, which already has a low LODrps value (-0.713), and D187, which is at the edge of the receptor binding domain (RBD). This decision was based on a D187 substitution perturbing the nearby RBD, and thereby disrupting sialic acid binding. The other reason for keeping the conserved D187 residue is that, by preserving the RBS, it is possible to generate an immune response against the RBS. Such an antibody would be an ideal neutralizing antibody as has been recently demonstrated (Whittle, J.R., R. Zhang, S. Khurana, L.R. King, J.

Manischewitz, H. Golding, P.R. Dormitzer, B.F. Haynes, E.B. Walter, M.A. Moody, T.B. Kepler, H.X. Liao, and S.C. Harrison, Broadly neutralizing human antibody that recognizes the receptor binding pocket of influenza virus hemagglutinin. Proc Natl Acad Sci U S A, 201 1. 108(34): p. 14216-21 ).

[0108] In the bottom row of Table 3 a measurement called the LODrps index was introduced. It is calculated by summing the LODrps values for primary antigenic determinant residues of the HA in each column and serves as a relative indicator of the degree to which the primary antigenic determinants are neutralized in a given hemagglutinin molecule. The parental VI 203 HA control has a LODrps index value of 6.7, which is similar to the 4.6 and 6.3 values obtained for the other naturally occurring Indo05 and Egypt06 control hemagglutinins. Partial neutralization of VI 203 primary antigenic sites one mAb epitope at a time in x-HA. l , x-HA.2 and x-HA.3, modestly reduced the 6.7 index value to around 4. However, (almost) wholesale replacement of V I 203 primary antigenic determinant residues with escape mutant substitutions in x-HA.em reduced the index value to 0.6. For the x-HA.4 series in which 6 primary epitope residues were substituted with escape mutant residues, threonine, or alanine, index values were 0.4, -3.2 and -9.6, respectively. For the x-HA.5 series, the index values were -9.2 and - 15.6. The inventor hypothesizes that, by reducing the LODrps indices of x-HA mutants, modified surfaces of the HA becomes less antigenic, and thereby to switch the immune response to remaining higher LODrps residues of conserved HA surfaces that are not normally immunogenic.

[0109] Immunoassays verify proper folding of recombinant hemagglutinins and primary antigenic determinant knock out in xHAs

[0110] Recombinant xHA protein folding was assessed by testing for binding to well-characterized mAbs which recognize discontinuous epitopes on the HA head and stem. Human mAb 1 F02 recognizes the conserved fusion peptide-containing epitope on Group 1 HA stems and protects mice from in vivo virus challenge with antigenically distinct influenzas ( Wrammert et al. J Exp Med, 201 1. 208( 1 ): p. 181 - 193). The VN04 mAbs recognize 3 discontinous epitopes, and 1 linear epitope, within the primary antigenic determinant surfaces of the V 1203 HA ( averin et al., J Virol, 2007, 81 : 1291 1 ). Binding was assayed by immobilizing xHAs, or a control protein (human antithrombin IN), at 1 ug/ml on microplates, followed by blocking and serial exposure to 1 F02 and alkaline phosphatase - goat anti human Ig, or to the VN04 mAbs and alkaline phosphatase - goat anti mouse Ig. Fig. 5 shows that the 1 F02 mAb to the HA stem epitope bound all of the xHAs, whereas the VN04 mAbs to V I 203 primary antigenic determinants, bound to xHA.par (which has a wildtype HA head surface with intact primary antigenic determinants), but not xHA.4b and 5a (which have ablated immunodominant epitopes, see Table 3). Together, the binding data are supportive of proper folding of the xHA.par and the xHA.4b and 5a hemagglutinins, and also confirm successful knock-out of primary antigenic determinants in the latter.

[0111] EXAMPLE 3 PRODUCTION OF ANTISERA TO H5 CONTROL AND x-HA HEMAGGLUTININS, SURROGATE ASSAY FOR VIRAL NEUTRALIZATION

FUNCTION, AND IDENTIFICATION OF x-HA UNIVERSAL VACCINE

CANDIDATES

101121 Overview

[0113] The abilities of control and x-HA hemagglutinins to elicit broadly cross-protective humoral antibody responses were examined in mice. Antisera raised to individual control and x-HA hemagglutinins, as well as antisera raised by sequential challenge with different x-HAs, were screened in HA pseudotyped lentiviral vector reporter neutralization assays to identify molecules that stimulate responses to stable, conserved regions of H5

hemagglutinins.

[01141 Production of mouse antisera to recombinant hemagglutinins Non-immune serum is collected from mice. Groups of 4 or 5 Balb/C or fVB mice are vaccinated by injection with 10 or 20 ug of a control or x-HA hemagglutinin in Sigma Adjuvant System (S6322, formerly Ribi Adjuvant System). Three weeks after the primary injection and at >21 day intervals after each boost, animals received additional 10 ug boosts. Blood was drawn 1 -2 weeks after the booster immunizations. Mice are immunized with a single control or x-HA hemagglutinin repeatedly, or receive different x-HA antigens for the original immunization and subsequent boosts an alternative immunization protocol.

[01151 Characterization of x-HA cross-protective function using influenza hemagglutinin pseudotyped lentiviral vector reporter assays If an x-HA is to serve as an effective avian influenza pandemic vaccine and antigen for immunotherapeutic development, it must elicit immune responses to hemagglutinins from a wide range of H5 influenzas independently of sharing primary antigenic determinants With evolving strains, and those immune responses must lead to the neutralization of viral function in order to achieve cross protection. A lentiviral vector reporter assay is used to quantify the ability of antisera raised against various control and x-HA hemagglutinins to inhibit HA-mediated membrane fusion, which is a essential step in the infection process and viral life cycle.

[0116] The HA pseudo virus lentiviral vector reporter assay was developed, and the components were generously provided, by Dr. Gary Nabel and colleagues from the NIH Vaccine Research Center (Kong, W.P., et al., Protective immunity to lethal challenge of the 1918 pandemic influenza virus by vaccination. Proc Natl Acad Sci U S A, 2006. 103(43): p. 15987-91.). To generate the flu HA pseudotyped reporter viruses, 293T cells were cotransfected with 7 ug of pCMVAR8.2, 7 ug of pHR'CMV-Luc, and 125 ng of a CMV/R 8 B H5 construct, wherein the HA segment corresponds to H5 A/Vietnam 1203/04 (clade 1 ), A/Thai 1 (KAN- 1 )/04 (clade 1 ), A/Indonesia/5/05 (clade 2.1.3), A/Egypt/2782/06 (clade 2.2), A/Nigeria/641/06 (clade 2.2), or A/Iraq/207/06 (clade 2.2). The packaging cells are transfected overnight, then changed to fresh medium. At 48 h, virion-containing supematants were harvested, 0.45-um filtered, aliquoted, and used immediately, or frozen at -80°C. [0117] For neutralization assays, antiserum dilutions are mixed with lentiviruses that have been pseudotyped with the different H5 hemagglutinins, then added to 96-well plates containing 5,000 293 A cells per well. The medium is changed 14-16 h later, and at 72 or 96 h post infection, cells are lysed and the lysates assayed for luciferase activity (Promega Bright Glo assay). Percent neutralization of the LV reporters by tested antisera is calculated as { 1 - [(luminescence in wells with added antiserum) / (luminescence in wells with no added antiserum)]} . Lentivirus neutralization titers are obtained by analyzing neutralization as a function of antiserum dilution. For example, LVnt50 is the greatest antiserum dilution producing at least 50% inhibition of neutralization. The HA- pseudotyped LV reporter assay is used to screen antisera and identify xHAs inducing neutralizing antibodies to conserved HA features. The desired xHAs will elicit antisera with high titers for the neutralization of multiple HA pseudotypes. For the x-HA.4 and x- HA.5 series in which all primary antigenic determinants are knocked out and replaced in 4 out of 5 cases with very low LODrps amino acids, we predict that the immune response will be switched to substructures on the HA surface that are not normally immunogenic, and that if such substructures are conserved, the antisera will bind and neutralize reporter viruses pseudotyped for all the clades.

(01181 Targets of antibodies to x-HA s with universal vaccine function Epitope identification algorithms such as DiscoTope (Haste Andersen, P., M. Nielsen, and O. Lund, Prediction of residues in discontinuous B-cell epitopes using protein 3D structures. Protein Sci, 2006. 15(1 1): p. 2558-67) are used to predict the locations of the secondary antigenic determinants. On the right side of Table 2, residues of the A/Viet 1203/04 HA ectodomain that DiscoTope predicts to be epitopes are marked. The Discotope algorithm predicted 34 residues as epitopes. On the basis of genetic drift and escape mutant data, 12 of the 34 correspond to HA primary antigenic determinants. The remaining 18 residues, not mapped to known primary antigenic determinants, potentially include some conserved secondary antigenic determinants.

[0119] Conserved hemagglutinin structural features mediating viral functions whose interruption causes neutralization include: (1 ) sites in and bordering the receptor binding site, and (2) stem fusion peptide structural elements participating in the conformational change that mediates viral envelope fusion with the host cell membrane (Skehel, J.J. and D.C. Wiley, Receptor binding and membrane fusion in virus entry: the influenza hemagglutinin. Annu Rev Biochem, 2000. 69: p. 531-69.). [0120] The inventor hypothesizes that (1) structural neutralization of HA primary antigenic determinants shifts the immune response towards conserved epitopes, and that (2) sequential immunization with different x-HAs (Table 5, lines 13 and 14) will be superior for eliciting broadly protective immune responses.

[0121] EXAMPLE 4: CROSS NEUTRALIZATION RESULTS

[0122] Table 4 shows experiments investigating neutralization of clade 1 , clade 2.2 and clade 2.1.3 H5 HA-pseudotyped lentiviruses by antisera raised against recombinant xHA immunogens bearing intact wildtype (xHA.par), and partially (xHA.2) and completely (xHA.4b and xHA.5a) ablated primary antigenic determinants. Mouse non-immune sera did not neutralize any of the reporter viruses. Antisera generated with the VI 203 xHA.par parental control exhibited a focused pattern of neutralization, which was limited to the VI 203 clade 1 and Egypt clade 2.2 pseudotypes. Broader cross neutralization, extending to the H5 Indonesia clade 2.1 and HI PR8 reporters, was obtained with antisera to xHA.2, which has partial ablation of hemagglutinin primary antigenic determinants. Additionally, antisera raised by priming with xHA.4b and subsequent xHA.5a boosting also produced broad cross neutralization. For reference, the neutralization pattern obtained with purified mAb 1 F02 (Wrammert et al. J Exp Med, 201 1. 208(1): p. 181 -193) is shown in the Controls section at the bottom of the Table 4. The 1 F02 mAb binds a conserved, fusion peptide- containing epitope on the HA stem and provides in vivo protection against several antigenically distinct influenzas when administered therapeutically. The broad in vitro neutralization observed with the polyclonal antisera from xHA immunized animals, may result from contributions of antibodies like 1F02, which recognize conserved elements on the HA stem, and/or from recently described antibodies to conserved elements on the HA head, which interfere with receptor binding.

[0123] The broadened neutralization results were obtained for a portion of the mice immunized with xHAs, indicating that that ablation of primary epitopes can promote an immune response with potential to protect against different serovariants. Different adjuvant strategies may be applied to increase the reproducibility and magnitude of the response. TABLE 4

BROADENED NEUTRALIZATION OF INFLUENZA HEMAGGLUTININ PSEUDOTYPED LENTIVIRUS REPORTERS BY xHAANTISERA

LVNT50 ¹ LVNT50 LVNT50

mouse serum V1203 Indo Egypt PR8 ID from ³ H5 clade 1 H5 clade 2.1 H5 dade 2.2 H1 non-immune serum; 2 BALB/c mice

bD.LR n.a. o o o o bE.LR n.a. o o o o xHA par: 10ug prime and boosts; 2 BALB/c mice

bA.L b2 >912 >3648 o bA.R b3 >1706 >6823 o xHA.2: 20 ug prime and 10ug boosts; best 2 of 3 BALB/c mice

A4.0 b6 >5472 o >2736 >342

A4.L b6 >685 o >2742 >1371 xHA.2: 10 ug prime and boosts; best 2 of 5 BALB/c mice

bB.LR b4 >3655 o >3655 >114 bB.R b4 >914 >914 >3655 o

20 ug xHA.4b prime and 10 ug xHA.5a boosts; best 3 of 4 BALB/c mice

B2.0 b7 >914 >228 >3655 >457

B2.LR b4 o o >2742 >343

B2.R b5 o o >457 >457

10 ug xHA.4b prime and 10 ug xHA.5a boosts; best 2 of 5 BALB/c mice

bC.LR b4 o >171 >2742 o bC.R b5 >533 >133 >4265 >133

10 ug xHA.4b prime and 10 ug xHA.5a boosts; best 2 of 4 fVB mice

fC.L b3 >914 >914 >3655 o fC.LL b5 >685 >685 >5483 >685

CONTROLS

mAb 1F02, 32 ug/ml >228 o >1824 >114 rabbit a H5N1 rgAA iet 1203/04

(BEIR NR-4485) >22800 >11400

goat a H5 HA A tern/South Africa/61

(BEIR NR-3156) >91200 >18240 o goat a H1 HA A Puerto Rico/8/1934

(BEIR NR-3148) o >54825

¹ LVnt50, serum dilution producing >50% neutalization

² LVnt40, serum dilution producing >40% neutalization

³ b2, boost 2; b3, boost 3; n.a., not applicable [0124] EXAMPLE 5: xHAs ELICIT ANTIBODIES TO CONSERVED FUSION- PEPTEDE CONTAINING EPITOPES ON THE HA STEM

[0125] When administered therapeutically, monoclonal antibodies to conserved fusion peptide-containing epitopes on the hemagglutinin stem mediate broad influenza protection. Therefore, it is of interest to determine if xHAs elicit immune responses to the conserved stem epitopes and could be used to strengthen immunological memory for and induce production of broadly protective stem antibodies. Example 5 and Figure 6 show that mouse immunization with xHAs elicits antibodies to this critical HA stem element.

[0126] Monoclonal antibody 1 F02 binds to a conserved fusion peptide-containing epitope on Group 1 HA stems and protects mice from live virus challenge by antigenically distinct influenzas (Wrammert et al. J Exp Med, 201 1. 208(1): p. 181-193). mAb 1 F02 was used in a competition elisa to determine if antibodies to conserved fusion peptide-containing epitopes on the HA stem are present in polyclonal antisera from mice immunized with xHAs. The solid phase for the assay was BEIR NR-4143 rgA/Vietnam/1203/04 (H5N1) monovalent influenza subO virion vaccine, with its hemagglutinin element derived from V I 203. NR-4143-coated wells were exposed to dilutions of anti-xHA mouse antisera, or to non-immune serum, or control mouse mAb C I 79 (Okuno, Y., Y. Isegawa, F. Sasao, and S. Ueda, A common neutralizing epitope conserved between the hemagglutinins of influenza A virus HI and H2 strains. J Virol, 1993. 67(5): p.25520 8) as negative and positive competition controls, respectively; followed by incubation with 1.2 ug/ml human mAb 1 F02. 1F02 binding was measured with alkaline phosphatase goat anti-human Ig. Percent 1 F02 binding values, calculated as (A405 in the presence of antiserum) / (A405 in the absence of antiserum), were used to assess the presence and level of stem fusion-peptide epitope binding antibodies in the serum samples.

[0127] Results of 1 F02 competition experiments are presented in Fig. 6. Panel A shows no competition by mouse non-immune serum (m NI, open circles), and competitive, concentration-dependent reduction of 1 F02 binding following exposure to the positive control CI 79 mAb (solid triangles). Each of the plots in the remaining B - D panels presents 1 F02 competition results from serum samples obtained at various stages during the immunization of a single animal with xHAs. The designation "pr" indicates antisera obtained after priming, "bl " and "b2" are respectively antisera obtained after boosts 1 and 2, etc. The animals in panel B were primed and boosted with xHA.par (10 ug/injection). The animals in panels C and D, were primed with xHA.4b, then repeatedly boosted with xHA.5b; those in C received 10 ug doses of the xHA antigens, while those in D received 20 ug priming and 10 ug boosting doses.

[0128] Antisera obtained after priming did not contain measurable 1 F02 competitive activity. But in the selected animals shown, 1 F02-competing antibodies developed subsequently, appeared in sera collected after the first, second or third boost, and persisted once they had appeared. Surprisingly, anti-stem antibodies were present not only in xHA.4b-primed and xHA.5a-boosted mice (C and D), but also in animals challenged solely with wildtype xHA.par (B). Flu infections and seasonal vaccines are thought to elicit only limited production of anti-stem antibodies. However, the HA stem of the recombinant xHA.par immunogen, which is comprised of HA0 trimers and oligomers, may be more accessible than are HA stems contained in the virion envelope or vaccine preparations, which undergo formalin cross-linking during manufacture. Additionally, adjuvants expand and enhance antibody responses to influenza (Coler, R.N. et al. , A synthetic adjuvant to enhance and expand immune responses to influenza vaccines. PLoS One, 2010. 5(10): p. el 3677; Khurana, S., et al. , Vaccines with MF59 adjuvant expand the antibody repertoire to target protective sites of pandemic avian H5N1 influenza virus. Sci Transl Med, 2010. 2(15): p. 15ra5.). Accordingly, adjuvanted (Sigma Adjuvant System) xHA.par

administration may have been another factor in the production of anti-stem antibodies for this group.

[0129] Neutralization and 1 F02 competition results were not tightly correlated, consistent with a polyclonal immune response to xHA antigens. The neutralization and 1 F02 competition result disparities imply that antibodies to conserved structures above and beyond the fusion peptide-containing stem epitope also contribute to neutralization. These, for example, might include broadly neutralizing antibodies which bind to the HA head and inhibit RBS access.

[0130] EXAMPLE 6: xHAs ELICIT ANTIBODIES TO HA HEAD EPITOPES

[0131] Hemagglutination inhibition (HAI) assays were performed to determine if xHAs induced antibodies that interfered with red blood cell (RBC) sialic acid binding to the HA receptor binding sites of three different H5 and one H I hemagglutinin. Antigens were HA- pseudotyped lentiviruses of the specificity indicated by row labels. LVs were incubated for 30 min with dilutions of receptor destroying enzyme (RDE) treated antisera, or no serum, as indicated by column labels in panels B-D. Plates were photographed 40m after turkey RBC addition. RBC buttons in the bottom right corner wells of each panel are RBC-only, no- hemagglutination controls. Clear appearance of the four wells above the no- hemagglutination controls indicates HAI by the respective LV antigens. Comparisons of the three left wells of bottom rows with adjacent no-hemagglutination control show that RDE- treated antisera had no inherent hemagglutinating activity. Presence of RBC buttons of various sizes in wells where antiserum was incubated with LV antigen indicates inhibition of hemagglutination by antibodies contained in the antiserum, and demonstrates that xHA immunization can induce production of antibodies that reduce interactions between HA head receptor binding sites and RBC sialic acids. Panel A shows controls for mouse (M) and rabbit (R, BELR NR-4484) nonimmune (NI) sera, a polyclonal rabbit antiserum (BEIR NR-4487) to V 1203, and a polyclonal goat antiserum (NR-3148) to PR8. Antisera for the panel B experiment were from mice primed and boosted with xHA.par (10 ug/injection). Antisera for the panel C experiment were from mice primed with xHA.4b, then boosted 4 times with xHA.5b using 10 ug of the xHA antigens. Antisera for the panel D experiment were from mice primed with xHA.4b, then boosted 6 times with xHA.5b (20 ug primes and 10 ug boosts).

[0132] Sera from mice immunized with the parental control xHA.par hemagglutinin bearing intact primary antigenic determinants produced expected inhibition of

hemagglutination (panel A). Panels C and D show HAI results from animals primed with xHA.4b and boosted with xHA.5b. The primary antigenic determinants on xHA.4b and xHA.5a were ablated with different combinations of low LODrps amino acid substitutions (see Table 3 and Figure 5). Hemagglutination inhibition was also noted for some anti- xHA.4b-5a antisera, indicating production of antibodies to secondary epitopes that are adjacent to the receptor binding site, or which alter the conformation of the RBS when the antibodies are bound. Thus, xHA.4b and xHA.5a, which have different primary antigenic determinant surfaces than the HAs on the conventional hemagglutinating antigens, can elicit antibodies to HA head region sites that are distinct from the primary antigenic determinants, but nevertheless reduce RBS access. Several antibodies to conserved epitopes on the HA head mediate broad influenza protection ( hurana et al, PLoS Med, 2009. 6(4): p.

el 000049; rause et al., J Virol, 201 1. pmid_21849447; Whittle, et al., Proc Natl Acad Sci U S A, 201 1. 108(34): p. 1421621 ; Yoshida, et al., PLoS Pathog, 2009. 5(3): p. e l 000350). The HA inhibition observed with xHA.4b-5a induced antibodies observed suggests that the receptor-binding inhibition observed with those antibodies are important in contributing to the broad influenza neutralization in animals immunized with these xHAs.

[01331 EXAMPLE 7: xHA.5a PRIMARY ANTIGENIC DETERMINANT ABLATIONS BROADEN IMMUNE RESPONSE TO H5 INFLUENZAS

Table 5

LVnt50 ¹ titers from xHA.5a immunized mice

MOUSE ANTISERA HA-pseudotype of LV reporters immunogen serum ³ V1203 indo egypt PR8 xHA.par b2 >912 o >3648 o xHA.par b3 >1706 o >6823 o xHA.5a b2 o o >3648 o xHA.5a b2 >228 o >1824 o xHA.5a b2 >1824 >456 >3648 o xHA.5a b2 >912 >228 >3648 o none n.a. o o o o none n.a. o o o o none n.a. o o o o

CONTROLS

goat anti H5 H A A/tem/South Africa/61

(BEIR NR-3156) >91200 o > 182400 o rabbit anti H5N1 rgA/Viet 1203/04

(BEIR NR-4485) >22800 o >11400 o mAb 1 F02, 56 ug/ml >456 o >3648 >114

¹ LVnt50, serum dilution producing >50% neutalization

² each row is a different mouse

³ b2, boost 2; b3, boost 3; n.a., not applicable

[0134] Table 5 shows that mice challenged with xHA.5a, an H5 A/Viet 1203/2004 - derived hemagglutinin with knock out of all primary antigenic determinants (see Table 3 and Figure 5), produced polyclonal antisera with broader H5 neutralization than did mice challenged with the parental xHA.par control hemagglutinin, which retains the intact, wild type primary antigenic determinants of VI 203. Given that the VI 203 primary antigenic determinants were not present on the xHA.5a immunogen, it is likely that the broadly neutralizing antibodies it elicited in 2 mice recognize H5 structural elements that are well- conserved across the H5 subtype. The significance of Example 7 is that it demonstrates that a response can be elicted with only one xHA. This suggests that a broadly protective vaccine against circulating and future pandemic H5 avian influenzas could be developed from an xHA.5a platform.

[0135] The foregoing examples demonstrate that xHAs are properly folded and form the correct prefusion trimeric form, and therefore present as many of the conserved epitopes as possible, including epitopes that have not been identified. These xHAs elicit

POLYCLONAL responses containing antibodies to a variety of different conserved HA structures, each of which mediate distinct essential functions in the virus life cycle. For example, the combined presence of (a) antibodies to conserved epitopes on the head and (b) antibodies to conserved stem epitopes would block virus replication at 2 different points in the life cycle (infection and fusion), which are more effective than blocking either step alone.

[0136] The reported xHA approach is fundamentally different from widely proposed universal flu vaccine development strategies of isolating broadly neutralizing mAbs, precisely mapping their epitopes, and using this information for structure-based design of immunogens that will elicit antibodies focused around the mapped epitope. Instead, xHAs are 'generic' hemagglutinin molecules, which have knocked-out immunodominant epitopes, but retain normal overall HA tertiary and quartenary structure, including multiple conserved elements shared by distantly related influenzas. In contrast to peptide-based and truncated HA universal vaccine candidates, trimeric xHAs provided a single immunogen to elicit antibodies to multiple conserved epitopes, leading to polyclonal anti-xHA antisera containing antibodies to conserved, neutralization-mediating HA stem and head targets. This finding supports the feasibility of developing long—lasting, broadly neutralizing, subunit-type universal flu vaccines from an xHA platform. xHA vaccines are also suitable for both stimulating and amplifying "seasoned" pan-influenza immunity, wherein cross- protection is mediated by the combined effects of complementary, broadly neutralizing antibodies. This represents a promising development because it means that xHA-based universal flu vaccines should be able to work in adults through a mechanism of boosting conserved epitope memory cells, rather than priming.