AVOIDING NARCOLEPSY RISK IN INFLUENZA VACCINES

Abstract

The invention provides influenza vaccines and methods which improve the safety of influenza vaccines further, in particular in relation to the risk of causing narcolepsy in adjuvanted vaccines.

Claims

1. An influenza vaccine composition comprising influenza virus A nucleoprotein wherein a fragment of said nucleoprotein equivalent to amino acids 106 to 126 of SEQ ID NO: 2 binds to an MHC class II receptor comprising HLA DQB1*0602 with a lower affinity than a peptide having the amino acid sequence shown in SEQ ID NO: 1, with the proviso that if all influenza A nucleoprotein in the composition comprises the amino acid sequence shown in SEQ ID NO: 12, then the vaccine composition is not based on strain A/California/7/2009 (H1N1)-derived strain NYMC X-181.

2. An influenza vaccine composition comprising influenza A virus nucleoprotein wherein none of said nucleoprotein comprises a fragment equivalent to amino acids 106 to 126 of SEQ ID NO: 2 which binds to an MHC class II receptor comprising HLA DQB1*0602 with an equal or higher affinity than a peptide having the amino acid sequence shown in SEQ ID NO:1, with the proviso that if all influenza A nucleoprotein in the composition comprises the sequence shown in SEQ ID NO: 12, then the vaccine composition is not based on strain A/California/7/2009 (H1N1)-derived strain NYMC X-181.

3. An influenza vaccine composition according to claim 1 or claim 2 wherein not all of the nucleoprotein in the composition comprises the amino sequence shown in SEQ ID NO: 12.

4. An influenza vaccine composition according to claim 3 wherein not all of the nucleoprotein in the composition comprises the amino sequence shown in SEQ ID NO: 3.

5. An influenza vaccine composition comprising influenza virus A nucleoprotein wherein said nucleoprotein does not have an isoleucine residue at a position corresponding to amino acid 116 of the nucleoprotein amino acid sequence shown in SEQ ID NO: 2, with the proviso that if all influenza A nucleoprotein in the composition comprises the amino acid sequence shown in SEQ ID NO: 12, then the vaccine composition is not based on strain A/California/7/2009 (H1N1)-derived strain NYMC X-181.

6. An influenza vaccine composition comprising influenza virus A nucleoprotein wherein said nucleoprotein does not have an isoleucine at a position corresponding to amino acid 116 of the nucleoprotein amino acid sequence shown in SEQ ID NO: 2, with the proviso that if all of the nucleoprotein comprises a methionine at a position corresponding to amino acid 116 of the nucleoprotein amino acid sequence shown in SEQ ID NO: 2 then said nucleoprotein does not have the sequence shown as SEQ ID NO: 12.

7. An influenza vaccine composition comprising influenza virus A nucleoprotein wherein said nucleoprotein does not have an isoleucine or a methionine residue at a position corresponding to amino acid 116 of the nucleoprotein amino acid sequence shown in SEQ ID NO: 2.

8. An influenza vaccine composition comprising influenza virus A nucleoprotein wherein said nucleoprotein does not comprise the amino acid sequence shown in SEQ ID NO: 2, SEQ ID NO: 12, or SEQ ID NO: 13.

9. An influenza vaccine composition comprising influenza virus A nucleoprotein wherein the nucleoprotein has been modified to reduce or abolish its binding to an MHC class II receptor comprising HLA DQB1*0602 as compared with the unmodified nucleoprotein.

10. An influenza vaccine composition comprising influenza virus nucleoprotein wherein (i) the composition is a split virion vaccine and the amount of nucleoprotein present is less than 3 μg nucleoprotein per 10 μg of hemagglutinin or (ii) the composition is a subunit vaccine and the amount of nucleoprotein present is less than 0.5 μg nucleoprotein per 10 μg of hemagglutinin.

11. A vaccine composition according to claim 1, further comprising an adjuvant.

12. A vaccine composition according to claim 11 wherein the adjuvant is an oil-in-water emulsion.

13. A vaccine composition according to claim 12 wherein the adjuvant further comprises tocopherol.

14. A vaccine composition according to claim 1, further comprising a detergent.

15. A vaccine composition according to claim 1, which is a vaccine against one or more pandemic influenza strains.

16. A vaccine composition according to claim 11, which is a monovalent vaccine composition.

17. A vaccine composition according to claim 1, which is a split virion vaccine.

18. An adjuvanted split influenza vaccine, wherein the vaccine comprises antigens from at least 4 different influenza viruses and nucleoprotein from at least one influenza A virus having the amino acid sequence shown in SEQ ID NO: 2, characterized in that the adjuvant is an oil-in-water emulsion adjuvant which does not contain an additional immunostimulating agent, and whereby the composition contains a detergent.

19. A vaccine or a vaccine composition according to claim 1 for use in the pediatric population (0-36 months) and/or the adolescent population (4-19 years) and/or in subjects with a genetic predisposition to develop an autoimmune disease in connection with flu vaccination.

20. A method of testing an influenza A virus for suitability for vaccine production, comprising a step of determining whether the influenza virus's nucleoprotein, or a fragment thereof, can bind to HLA DQB1*0602 with lower affinity under the same conditions compared to nucleoprotein from H1N1 strain X-179A; wherein the influenza virus is suitable for vaccine production if its nucleoprotein can bind to HLA DQB1*0602 with lower affinity under the same conditions compared to nucleoprotein from strain X-179A.

Description

DESCRIPTION OF THE DRAWINGS

[0282] FIG. 1 Smith-Waterman alignments between the nucleocapsid X-179A/X-181 sequence fragments from Table 3 and orexin receptor 2 (Ox1R) and orexin receptor 1 (Ox2R). No other alignments between influenza fragments and orexin family sequences were less than 0.4. The alignments were generated with the program search from the FASTA package with default parameters. The regions of Ox2R and Ox1R in the alignments shown in are annotated as extracellular in Uniprot.

EXAMPLES

Narcolepsy and the H1N1 Pandemic

[0283] Molecular mimicry is an evolutionary adaptation whereby viruses and bacteria attempt to fool the body into granting them free access to the host tissue. Such mimicry works by showing the immune system stretches of amino acids that look like self. In responding to the microbe, the immune system becomes primed to attack the corresponding self-component (e.g. adenovirus type 2 and myelin basic protein in multiple sclerosis).

Autoimmune Diseases and Natural Infections

[0284] What is commonly observed by clinicians managing patients with autoimmune diseases is that natural infection can trigger and augment the severity of autoimmune disease activity. Similarly, natural infections are thought to play a major role in inducing disease in a genetically susceptible host. In the context of the H1N1 pandemic, 153 subjects were infected with H1N1 in Beijing, China, by May 2009 which increased to an estimated 1.18 million infected subjects by November 2009. By November, 1.36 million doses of unadjuvanted H1N1 vaccine had been administered to a population of 17 million (i.e. 0.8% of the population were vaccinated). Six months following the peak of H1N1 infection, there was a report of a 3 to 4-fold increase in new narcolepsy cases (n=142) in a Beijing cohort in which only 5.6% patients reported being vaccinated. This suggested to the inventors that the narcolepsy seen in Pandemrix™-treated patients might be connected to the H1N1 strain itself, perhaps due to molecular mimicry.

Pathological Mimicry Related to the Pandemrix™ Vaccine Antigen

[0285] As mentioned above, the Pandemrix™ vaccine was associated with narcolepsy while no increase in narcolepsy could be seen following vaccination with Focetria™. The source of the vaccine antigens used for Pandemrix™ is the high yielding A H1N1 reassortant X-179A while that used for Focetria™ is the higher-yielding A H1N1 reassortant X-181.

[0286] For pandemic H1N1 vaccine preparation, X-179A (used for Pandemrix™) was generated by the cross of high yielding strain X-157, with internal proteins traceable to A/Puerto Rico/8/1934 (PR8), with A/California/07/2009 (H1N1 subtype) which contributed the surface antigens HA and NA, and the internal protein PB1. Due to concerns of inadequate yield, the 32-fold yielding X-179A was re-reassorted by crossing again to X-157 on Jul. 14, 2009 leading to the generation of the 64-fold yielding X-181 (used in Focetria™ and other seasonal vaccines). This functional difference is related not only to the gene segment combinations from the reassortment, but also pre-existing variants and virus mutations selected during adaptation to the egg host (in ovo inoculation). A similar result occurred with X-53 and X-53a reassortants prepared for the swine flu vaccine in 1976 which were not antigenically distinguishable but had a single amino acid change in the HA gene of X-53a that increased the yield by 16-fold. Thus, X-179A used for Pandemrix™ is likely to have other qualitative differences compared to X-181 used for Focetria™ as will be explained below.

[0287] The purification processes for Pandemrix™ and Focetria™ have distinct differences. Split-virion vaccines (like Pandemrix™) and subunit vaccines (like Focetria™) are defined in the World Health Organization document on production technologies for influenza vaccines as follows: [0288] The majority of influenza vaccines are ‘split’ vaccines, which are produced by detergent-treating purified influenza virus. The splitting process breaks the virus allowing the relevant antigens to be partially purified (p.8) . . . . Compared to the whole-virus preparation, split vaccines are better characterized, contain less ovalbumin and are claimed to be less reactogenic (p.13). Subunit or surface surface-antigen vaccines are produced as for split virus, but more rigorous purfication is carried out so that the vaccine consists almost exclusively of highly purified HA and NA with minimal contaminating N, matrix protein, nucleoprotein and lipid.

[0289] The amount of nucleoprotein and matrix protein content in an intact influenza virus is two-fold and six-fold higher, respectively, to that of HA making these two internal proteins most likely to carry over depending on the type of purification process used to enrich for HA and NA. Indeed, a previous investigation characterized the split-virion vaccines Fluzone and MFV-Ject and the purified antigen/subunit influenza vaccines Fluvirin and Influvac that were available in the United Kingdom in 1992 [79]. Based on electron microscopy, viral nucleoprotein is readily evident in split-virion vaccines and by SDS-PAGE gel is detectable in all vaccines to varying levels (split-virion>purified antigen/subunit).

[0290] Since mutations of orexin receptors and loss of orexin cells have been implicated in narcolepsy, we have hypothesized that a vaccine antigen in Pandemrix™ could have a unique characteristic leading to molecular mimicry with either orexin (hypocretin) or its receptors. Therefore, sequence analysis was performed on the influenza proteins from X-179A, X-181 and orexin-related sequences (orexin, orexin A, orexin B, orexin receptor 1, orexin receptor 2 and HLA-DQB1). As the Pandemrix™ strain X-179A is implicated in the narcolepsy cases but Focetria™ is not, the influenza proteins were first compared for differences between the vaccine viral strains X-179A (Pandremix-associated) and X-181 (Focetria™-associated) that could account for the association with narcolepsy. Only proteins expected to be present in the vaccine preparations were included (HA, NA, NP and M1) as suggested by the electron microscopy and SDS-PAGE studies previously described.

[0291] To determine potential amino acid changes, influenza sequences were retrieved from the NCBI's Influenza Virus Resource [80] querying by the appropriate strain names (X-179A, X-181, A/California/07/2009(H1N1)) with duplicate sequences removed. Each influenza protein expected to be present in the vaccine preparations (HA, NP, M1, M2, NA) was compared across strains to identify sequences where X-179A differed from X-181 by at least one residue. Three amino acid differences were found with one in HA and two in NP (Table 1):

TABLE-US-00001 Genbank SEQ ID Protein Accession Strain Sequence NO: HA ACR47014 X-179A 136 KTSSWPNHDSNKGVTAACPHA 6 AFM72842 x-181 136 KTSSWPNHDSDKGVTAACPHA 7 NP ADE29096 X-179A 106 RELILYDKEEIRRIWRQANNG 1 AFM72846 X-181 106 RELILYDKEEMRRIWRQANNG 3 ADE2096 X-179A 130 WRQANNGDDAAAGLTHMMIWH 4 AFM72846 X-181 130 WRQANNGDDATAGLTHMMIWH 5

[0292] Sub-sequences including ten residues before and after the differences were then compared to the orexin related sequences using a Smith-Waterman alignment. Only four alignments had an e-value below 0.4 (the next best e-value is >0.4 giving a ratio of the best e-value alignment to the next best of 10:1 which is a good separation of the indicated alignment from the others). The alignments were between a nucleocapsid protein fragment from X-179A and X-181 (containing a single residue difference) and the orexin receptors 1 and 2 (FIG. 1). As the X-179A version of the nucleocapsid fragment overlaps with numerous epitopes reported in the Immune Epitope Database (www.iedb.org), this overlap with orexin receptors 1 and 2 (which are implicated in narcolepsy) suggests the potential for mimicry and is a possible explanation for the observations of narcolepsy in certain European countries following administration of the Pandemrix™ vaccine. Interestingly, the X-179A version of the nucleocapsid fragment is also identical to that from the H1N1 infection (consistent with the reported association of infection with narcolepsy [81]) while the X-181 version (Focetria™-linked) of the same nucleocapsid fragment contains a single amino acid substitution that could explain the lack of narcolepsy association with Focetria™.

[0293] Patients with narcolepsy associated with Pandemrix™ are exclusively HLA DQB1*0602. Thus in Sweden all 28 post-vaccination cases of narcolepsy were HLA DQB1*0602. In Finland, all 34 HLA typed narcolepsy patients in 2010 who were vaccinated with Pandemrix™ were HLA DQB1*0602. The binding motif for HLA DQB1*0602 is known. The core binding motifs for HLA DQB1*0602 are at positions 1, 3, 4, 6, and 9. There is a register for the orexin receptor 1 and 2 peptides that fits this motif well for the peptide LILYDKEEIRRIWRQANNG (SEQ ID NO: 10) with aliphatic amino acids at position 1 and 3, and a hydrophobic amino acid at position 4. Though position 6 does not fit the motif, position 9 is aliphatic and provides a good fit [82]. In narcolepsy the P4 binding pocket with the largest volume is critical for susceptibility, and some known examples of strong binders to 0602 have tyrosine at P4 [8].

[0294] The Pandemrix™ vaccine may further have included certain components of the H1N1 infectious agent responsible for immune responses to molecular mimics of self-antigens (hypocretin or one of its receptors) leading to narcolepsy. However, one must still exercise appropriate vigilance before drawing definitive conclusions because of the discordance in narcolepsy signals associated with the AS03-adjuvanted Pandemrix™ vaccine and the AS03-adjuvanted H1N1 pandemic vaccine Arepanrix (administered in Canada with no report of increased narcolepsy associated with vaccination). This might suggest that simply the presence of a pathogenic antigen may alone not be enough to explain the association or, alternatively, there may be differences in the presence or presentation of the split vaccine antigens in the AS03-adjuvanted vaccines due to differences in the splitting/purification process of the manufacturing sites (Dresden for Pandemrix™ and Quebec for Canada). An estimated 30.8 million doses of the GSK AS03-adjuvanted H1N1 vaccine manufactured in Dresden were used in more than 47 countries starting in October 2009 with high coverage in some countries including Finland. Sweden, Norway, and Ireland. The GSK AS03-adjvuanted H1N1 vaccine manufactured in Quebec was used with high coverage in Canada (Arepanrix) where an estimated 12 million doses were administered and also administered in several other countries. Epidemiological studies that are on-going in Canada will report on any association between narcolepsy and the AS03-adjuvanted H1N1 vaccine (Arepanrix) in due course.

[0295] The Dresden antigen contains Polysorbate 80 (Tween 80) and Triton X-100, while the Quebec antigen does not contain these excipients. It has been speculated that these detergents might have an effect on the development of Narcolepsy (see Assessment report Immunological differences between pandemic vaccines EMA/687578/2012). Therefore it might be of particular importance to avoid the presence of NP protein in vaccines produced from antigens which contain Tween 80 and/or Triton X-100 like in the Dresden antigen. This applies in particular to vaccines adjuvanted with oil-in-water emulsions like MF59 or AS03, as narcolepsy has not appeared in unadjuvanted seasonal vaccines derived from the Dresden antigen (see below).

[0296] The following table 2 shows the composition of 2 different inactivated split virion H1N1 antigen components prepared in Dresden (WO2011/051235 p 41).

TABLE-US-00002 Quantity per Quantity per 0.25 ml - 0.25 ml - DFLSA013A DFLS014A (initial (adapted Ingredient process) process) Unit Purified antigen fractions of 3.75 3.75 μg HA inactivated split virion A/California/7/2009 (H1N1)v NYMC X-179A Polyoxyethylene sorbitan ≧28.75 ≧28.75 μg monooleate (TWEEN-80 ™, or polysorbate 80) t-octylphenoxypolyoxyethanol 3.75 22.5 μg (TRITON X-100 ™) Sodium chloride 1.92 1.92 mg Disodium phosphate 0.26 0.26 mg Potassium dihydrogen phosphate 0.094 0.094 mg Potassium chloride 0.050 0.050 mg Magnesium chloride 0.012 0.012 mg Thiomersal 5 5 μg Water for injections q.s. ad. 0.25 0.25 ml

Absence of Narcolepsy in the GSK Non-Adjuvanted H1N1 Seasonal Influenza Vaccine

[0297] While both the pandemic and seasonal influenza vaccines manufactured by GSK contain the same H1N1 antigen, there is no narcolepsy signal reported with the non-adjuvanted H1N1 antigen in the seasonal vaccine that could immediately lead one to speculate that excessive immunostimulation by the adjuvant is causal for narcolepsy. However, in light of the previous discussion on cryptic antigens being revealed in the split-virus vaccines, one could as well explain the discordant narcolepsy association to the reservoir of pathological antigens being preferentially or more efficiently presented to the immune system by the AS03 adjuvant—keeping in mind that improved antigen presentation is a desirable and expected effect with any adjuvants. Adjuvants may act in several ways including the following: 1) delivering antigens to the immune system, 2) enhancing the uptake of the antigen by antigen-presenting cells (APCs), or 3) altering the structural conformation of the antigen within the vaccine, thereby allowing for progressive release, delayed clearance and better exposure to the immune system. The mode of action of oil-in-water emulsions is being better understood and currently is considered to involve modulation of innate inflammatory responses, APC recruitment and activation, enhancement of antigen persistence at the injection site, modulation of presentation of antigen to immune-competent cells, and elicitation of different patterns of cytokines.

Immunological Explanation

[0298] Pertinent findings from a conservancy analyses published in 2007 on antibody and T cell epitopes of influenza A virus using the Immune Epitope Database and Analysis Resources (IEDB) are the paucity of antibody epitopes in comparison to T-cell epitopes with the highest number of T-cell epitopes being derived from hemagglutinin protein and nucleoprotein [83]. Furthermore, T-cell epitopes are more conserved than antibody epitopes with 50% being conserved at 80% identity levels in human H1N1 strains suggesting significant levels of interstrain cross-reactivity for T-cell epitopes in influenza. Nucleoprotein of influenza A is efficiently presented by class I and class II major histocompatibility complexes and is capable of expanding both CD8+ and CD4+-specific effector T lymphocytes secreting gamma-interferon and tumor necrosis factor [84]. It is a major target antigen for cross-reactive anti-influenza A cytotoxic lymphocytes (CTL), and recombinant vaccinia virus containing the PR8 nucleoprotein gene can both stimulate and prime for a vigorous secondary cross-reactive CTL response [85]. It is precisely for this reason why split-virion influenza vaccines that contain significant quantities of non-surface proteins would be expected to increase cell-mediated immune responses compared to purer subunit vaccines. However, this is a double-edged sword because the same immunological mechanism generating a vigorous cross-reactive CTL response to a defined epitope of viral (or vaccine-modified) nucleoprotein can be problematic if it mimics host tissue (e.g., orexin receptors) leading to T-cell-mediated autoimmunity that degenerates in a genetically susceptible host (e.g., DBQ1*0602) into autoimmune disease (e.g., narcolepsy). The alignment of the Pandemrix™ nucleoprotein fragment with the orexin receptors is intriguing because distinct narcolepsy syndromes have also been generated in orexin 2 receptor knockout mice and orexin knockout mice (possibly through defective orexin to orexin 1 receptor signaling) [86]. Interestingly, if trace amounts of immunogenic nucleoprotein were present in Focetria™, there is the presence of one amino-acid substitution (methionine) in the nucleoprotein fragment aligning with orexin receptors that distinguishes it from that of Pandemrix™ nucleoprotein and H1N1 infection (both of which have isoleucine). This amino acid substitution in nucleoprotein contained in Focetria™ (inherited from the X-181 vaccine strain) may be functionally similar to that for influenza viruses in which a number of amino acid substitutions in the nucleoprotein enable escape from CTL-mediated immune surveillance in contrast to matrix protein that is highly conserved [87].

HLA Haplotype Binding Assays

[0299] Binding of peptides to various HLA-DRB1* haplotypes was analyzed by using the cell-free REVEAL™ class II binding technology (see ref. 88). This technology measures the ability of synthetic test peptides to stabilize MHC-peptide complexes. Detection is based on the presence or absence of the native conformation of the MHC-peptide complex, which is detected by a specific monoclonal antibody. Each peptide is given a score relative to a positive control peptide, which is a known T-cell epitope. The score is reported quantitatively as a percentage of the signal generated by the test peptide compared with the positive control peptide. Scores are assessed at time zero and again after 24 hours. The analysis also gives a stability index which represents the stability of the binding of each peptide with the MHC II complex being tested.

[0300] A total of 18 peptides were tested, plus a positive control:

TABLE-US-00003 Peptide SEQ ID # Sequence NO: Details  1 RELILYDKEEIRRIWRQANNG  1 X179-A NP fragment  2 RELILYDKEEMRRIWRQANNG  3 X181 NP fragment  3 LILYDKEEIRRIWRQ 18 X179-A NP fragment  4 LILYDKEEMRRIQRQ 19 X181 NP fragment  5 VGKMIGGIGRFYIQM 20 Common NP sequence  6 SGAAGAAVKGVGTMV 21 Common NP sequence  7 EKATNPIVPSFDMSN 22 Common NP sequence  8 IDPFKLLQNSQVVSL 23 Common NP sequence  9 LILYDKEERRRRWRQ 24 Mutant of #3 10 MNLPSTKVSWAAVTL 25 Orexin DQB1*0602 fragment 11 MNLPSIKVSWAAVTL 26 Mutant of #10, Thr.fwdarw.Ile 12 MNLPSMKVTSWAAVT 27 Mutant of #10, Thr.fwdarw.Met 13 LTVAAWSVKTSPLNM 28 Reverse of #10 14 GAGNHAAGILTLGKR 29 Orexin fragment HCRT56-68 15 ASGNHAAGILTMGRR 30 Orexin fragment HCRT87-99 16 AMERNAGSGIIISDT 31 Hemagglutinin fragment 17 ALNRGSGSGIITSDA 32 Hemagglutinin fragment 18 ALSRGFGSGIITSNA 33 Hemagglutinin fragment

[0301] Two HLA haplotypes were tested: DQA1*0102:DQB1*0602; and DQA1*0101:DQB1*0501. As discussed above, the DQB1*0602 haplotype has a known link to narcolepsy, but the DQB1*0501 haplotype seems to protect against development of narcolepsy based on HLA typing studies of patients.

[0302] Binding results were as follows:

TABLE-US-00004 Pep- DQB1*0501 DQB1*0602 tide REVEAL REVEAL Stability REVEAL REVEAL Stability # 0 hrs 24 hours index 0 hrs 24 hours index 1 27.8 7.4 3.5 24.4 16.9 11.1 2 1.1 1.0 1.1 1.5 0.5 0.2 3 4.7 1.2 0.6 1.2 0.4 0.2 4 0.1 0.1 0.1 0.5 0.4 0.6 5 43.6 13.9 6.4 77.3 52.3 33.0 6 0.0 0.0 0.0 0.1 0.1 0.1 7 0.0 0.0 0.0 0.0 0.0 0.0 8 0.0 0.0 0.0 0.1 0.1 0.1 9 19.0 6.4 2.9 22.3 17.3 14.7 10 0.2 0.1 0.1 1.7 0.3 0.2 11 18.0 2.1 1.4 47.5 20.8 9.6 12 2.5 0.8 0.4 12.3 3.7 1.7 13 0.3 0.3 0.4 0.9 0.4 0.2 14 0.0 0.0 0.0 0.1 0.1 0.1 15 0.1 0.1 0.1 0.2 0.2 0.3 16 0.1 0.1 0.1 0.0 0.0 0.0 17 0.0 0.0 0.0 0.0 0.0 0.0 18 0.0 0.0 0.0 0.0 0.0 0.0 Ctrl 100.0 14.5 8.7 100.0 67.4 42.5

[0303] There were four strong binders for the DQB1*0602 HLA haplotype, namely peptides #1, #5, #9 and #11 (>15% of the positive control signal). In contrast to the X179-A peptide (#1) which was strongly bound at time 0 and still 24 hours later, the corresponding fragment from X181 (#2) was a poor binder at both time points.

[0304] In general, the binding stability of all peptides is weaker with the DQB1*0501 haplotype. Peptide #1 still binds well but is clearly less stable after 24 hours. Peptide #3 (a shorter version of peptide #1) binds better to this haplotype than to DQB1*0602, whereas #4 (a shorter version of #2) remains a poor binder with the 0501 haplotype. Thus the, compared to DQB1*0602, which is clearly associated with narcolepsy patients, the NP peptides bound poorly to the haplotype which seems to protect against development of narcolepsy (DQB1*0501).

[0305] A 15-mer orexin fragment (#10) did not bind to either HLA haplotype, but a modified version of this peptide with a Thr.fwdarw.Ile mutation (#11) was a strong binder. The result for peptide #10 is not surprising in view of reference 8's report that the fragment needed a 15 aa linker in order to “facilitate complex formation” with the DQB1*06:02 HLA. The ability of the Thr.fwdarw.Ile mutation to convert the peptide into a strong binder supports the importance of the isoleucine in NP of strain X179-A in conferring affinity for the DQB1*06:02 HLA haplotype.

[0306] The result with peptide #12 demonstrates that changing the threonine in peptide #10 to methionine does not allow or improve binding to HLA DQB1*0602 (the allele associated with narcolepsy) and thus confirms the result seen with peptides #1 and #2 regarding DQB1*0602

[0307] Overall, these results show that peptides #1 and #2 exhibit markedly different binding to HLA-DQA1*01:02 DQB1*06:02. Peptide #1 exhibits good binding to this allele, and its high stability score suggests that the binding is strong. In contrast, peptide #2 exhibits low initial binding, and only a very small amount of complex remains after 24 hours, suggesting a short binding half-life, and an unstable complex. Thus peptide #1 (amino acids 106-126 of the X-179A nucleoprotein and containing an isoleucine residue at amino acid position 116) bound more strongly and with better stability than peptide #2 (amino acids 106-126 of the X-181 nucleoprotein and containing a methionine residue at amino acid position 116). Accordingly, these data support our hypothesis that the X-179A nucleoprotein, but not the X-181 nucleoprotein, may be involved in an autoimmune response specific to individuals with the HLA-DQB1*06:02 haplotype.

Modeling of Peptide Interactions with Orexin

[0308] The orexin (hypocretin) fragment 1-13 (SEQ ID NO: 16) is known to interact strongly with HLA DQB1*0602 [8]. Analysis shows that Leu-3, Thr-6 and Val-8 are key residues for the interaction. These three residues give a good 3D structural alignment with a 12-mer fragment of the X-179A nucleoprotein fragment (SEQ ID NO: 17) with the influenza NP peptide in the reverse orientation, with Ile-6 aligning with orexin Thr-6 and Ile-9 aligning with Leu-3. The NP Ile-6 is the residue which differs between SEQ ID NOs: 1 (X-179A) and 3 (X-181). Computer modelling shows that SEQ ID NO: 17 shows a very good fit in the binding groove of the DQ0602 crystal structure. The Ile-6 residue in X-179A fits well in the HLA protein's binding cavity for orexin's Thr-6, but the X-181A methionine residue in the corresponding 12-mer fragment of SEQ ID NO: 3 gives a severe spatial clash with the HLA protein.

[0309] For this modeling the PDB file 1uvq was used with PyMol (Schrodinger Inc). An analysis of the intermolecular interactions between the hypocretin peptide (SEQ ID NO: 16) and the HLA protein was performed in detail, both visually and through the use of the SiteMap software (Schrodinger Inc). Three residues from the X-ray structure of the peptide were identified as important to drive binding to the HLA protein, most likely through hydrophobic interactions (Leu-3, Thr-6 and Val-8).

[0310] An alignment of a fragment of the nucleoprotein (SEQ ID NO: 17) was carried out and it showed many clashes and unlikely binding poses, until it was docked in the reverse orientation. A putative binding mode was evaluated to the binding site of HLA-DQB1*0602 in terms of polarity and van der Waals clashes, and a good match was seen when Ile-6 was aligning with orexin Thr-6 and NP Ile-9 aligns with Leu-3. To avoid incorporating noise or bias into the alignment assessment, the following was done: (i) the orexin peptide was mutated into the NP peptide residue-by-residue using the mutation module of PyMol; (ii) at each position, the best conformers for each residue of the NP peptide were assessed to fit or not into the binding pocket by its proximity to the van der Waals surface of the HLA protein. The surface was generated with PyMol using the default normal external surface: (iii) the models generated for the NP peptide should fit as close as possible the location of the orexin template. A satisfactory model was generated by this method. One polarity mismatch originating in the alignment (NP Glu-4 vs. orexin Val-8) was further evaluated, but the Glu residue is spatially tolerated in the binding groove of the HLA protein (no clashes with the surface of the protein are observed).

[0311] To further analyse this binding hypothesis, the Ile/Met mutation (SEQ ID NOs: 1 and 3) was evaluated in the binding pocket following the same principles detailed above. When the Ile residue is mutated to Met, no conformer of Met could be found that did not severely clash with the surface of the HLA protein. From this analysis, it is concluded that the Met mutation cannot be tolerated in this pocket within the proposed structural alignment.

[0312] In parallel, both sequences (i.e. the Ile and Met variants) were modelled in DQB1*0302 (PDB 1jk8) and the pocket which would accommodate this residue (and which fits a Tyr residue from insulin in the published sequence) is flexible and so should not discriminate between Ile and Met.

[0313] Similarly, both NP sequences and the orexin fragment were modelled in a HLA-DQ2 structure (PDB 1s9v). The orexin peptide fits into the DQ2 groove with a loose fit and no severe clashes. In contrast, both NP peptides have a very severe clash protruding through the DQ2 surface, and this clash could not be cured in any rotamer.

[0314] Thus, these modelling studies are consistent with the MHC peptide binding study: of the three HLA molecules modelled, only DQB1*0602 can discriminate between the NP peptides which differ by the Ile/Met variation.

[0315] Accordingly, both the MHC peptide binding study and the modelling study point to differential binding of certain X-179A versus X-181 nucleoprotein-derived peptides to DQB1*0602 but not other HLA sub-types.

[0316] In addition, this sort of in silico modelling can be used to identify amino acid substitutions within the NP which should avoid strong binding to the DQB1*0602 haplotype.

Mass Spectrometry Analysis of Influenza Vaccines

[0317] Mass spectrometry was used to identify and quantify NP within five inactivated split influenza vaccines which include HA from the X-179A strain.

[0318] 100 μl samples of vaccines were acidified by addition of 10 μL 50% formic acid water. Vaccines were vortexed and sonicated in a sonicating water bath for 10 minutes. Protein precipitation was performed with the addition of 500 μL −80° C. acetone and stored at −80° C. for 2 hrs. Samples were centrifuged at 4° C., 10,000 rpm for 15 minutes. The supernatant was discarded and the protein pellet was dried for 10 minutes in a speed vac. Vaccines were reconstituted in 20 μL 8M urea, 50 mM ammonium bicarbonate and 20 μL 0.5% protease, 50 mM ammonium bicarbonate. Reduction was performed using DTT to a final concentration of 5 mM, reduced at 55° C. for 30 minutes. The samples were brought to room temperature and alkylated using propionamide at a final concentration of 10 mM, room temperature for 30 minutes. 60 μL of 50 mM ammonium bicarbonate was added to each sample followed by 300 ng of trypsin/LysC mix. The samples were digested overnight at 37° C., followed by acidification by the addition of 10 μL 10% formic acid in water. The peptides were purified on C18 microspin columns and dried to less than 3 μL in a speed vac.

[0319] Each sample was then reconstituted for HPLC-MSMS in 15 μL 0.2% formic acid, 2% acetonitrile 97.8% water and injected onto a self-packed fused silica 25 cm C18 reversed phase column. The flow rate was 300 nL/minute with a linear gradient from 8% mobile phase B to 50% mobile phase B over 90 minutes. The mass spectrometer was an LTQ Orbitrap Velos, set in data dependent acquisition (DDA) mode to fragment the 15 most intense multiply charged precursor ions, where these ions were placed on an exclusion list for 60 seconds. Results were searched on a sequence database using Byonic™ with tolerance settings of 10 ppm on the precursor ion and 0.25 Da on the fragment ions. A 1% FDR using a reverse decoy database approach was employed. The database contained all vaccine related protein sequences (310,490) from NCBI. For each vaccine the proteins were sorted by spectral count and a relative abundance calculation was made by summing the intensities of the top proteins identified by spectral counts in each vaccine. The reported intensity for each of the top 10 proteins was divided by the summed intensity for these 10, multiplied by 100 and presented as a percentage top 10 abundance.

[0320] Spectral counts and percentage abundance values for NP and HA from the X-179 strain were:

TABLE-US-00005 % Vaccine NP count % abundance HA count abundance Fluzone ™ 3-valent 457 13.9 219 8.4 Fluzone ™ 4-valent 444 11.2 252 5.7 Fluarix ™ 3-valent 517 18.4 336 6.1 Fluarix ™ 4-valent 417 7.6 196 3.9 Afluria ™ 3-valent 881 20.0 118 4.0

[0321] In further analysis, peptides identified by MS were precisely matched by sequence to the strains known to be present in the vaccines, rather than by homology. Using this method the results were as follows:

TABLE-US-00006 Vaccine NP count HA count NP:HA Fluzone ™ 3-valent 328 258 1.27 Fluzone ™ 4-valent 320 335 0.96 Fluarix ™ 3-valent 146 282 0.52 Fluarix ™ 4-valent 142 187 0.76 Afluria ™ 3-valent 212 604 0.35

[0322] It will be understood that the invention has been described by way of example only and modifications may be made whilst remaining within the scope and spirit of the invention.

TABLE-US-00007 SEQUENCES SEQ ID NO: 1 X-179A NP fragment 106-126) RELILYDKEEIRRIWRQANNG SEQ ID NO: 2 (X-179A NP full length; 49 8aa)   1 MASQGTKRSY EQMETDGERQ NATEIRASVG KMIGGIGRFY IQMCTELKLS DYEGRLIQNS  61 LTIERMVLSA FDERRNKYLE EHPSAGKDPK KTGGPIYRRV NGKWMRELIL YDKEEIRRIW 121 RQANNGDDAA AGLTHMMIWH SNLNDATYQR TRALVRTGMD PRMCSLMQGS TLPRRSGAAG 181 AAVKGVGTMV MELVRMIKRG INDRNFWRGE NGRKTRIAYE RMCNILKGKF QTAAQKAMMD 241 QVRESRNPGN AEFEDLTFLA RSALILRGSV AHKSCLPACV YGPAVASGYD FEREGYSLVG 301 IDPFRLLQNS QVYSLIRPNE NPAHKSQLVW MACHSAAFED LRVLSFIKGT KVLPRGKLST 361 RGVQIASNEN METMESSTLE LRSRYWAIRT RSGGNTNQQR ASAGQISIQP TFSVQRNLPF 421 DRTTIMAAFN GNTEGRTSDM RTEIIRMMES ARPEDVSFQG RGVFELSDEK AASPIVPSFD 481 MSNEGSYFFG DNAEEYDN SEQ ID NO: 3 (X-181 NP fragment 106-126) RELILYDKEEMRRIWRQANNG SEQ ID NO: 4 (X179A NP fragment 130-150) WRQANNGDDAAAGLTHMMIWH SEQ ID NO: 5 (X-181 NP fragment 130-150) WRQANNGDDATAGLTHMMIWH SEQ ID NO: 6 (X-179A HA fragment 136-157) KTSSWPNHDSNKGVTAACPHA SEQ ID NO: 7 (X-181 HA fragment 136-158) KTSSWPNHDSDKGVTAACPHA SEQ ID NO: 8 (X-179A NP fragment 108-116) LILYDKEEI SEQ ID NO: 9 LXLYXXXIXXXXXX SEQ ID NO: 10 (X-179A NP fragment 108-126) LILYDKEEIRRIWRQANNG SEQ ID NO: 11 LILYDKEEX SEQ ID NO: 12 (X-181 NP full length; 498 aa)   1 MASQGTKRSY EQMETDGERQ NATEIRASVG KMIGGIGRFY IQMCTELKLS DYEGRLIQNS  61 LTIERMVLSA FDERRNKYLE EHPSAGKDPK KTGGPIYRRV NGKWMRELIL YDKEEMRRIW 121 RQANNGDDAT AGLTHMMIWH SNLNDATYQR TRALVRTGMD PRMCSLMQGS TLPRRSGAAG 181 AAVKGVGTMV MELVRMIKRG INDRNFWRGE NGRKTRIAYE RMCNILKGKF QTAAQKAMMD 241 QVRESRNPGN AEFEDLTFLA RSALILRGSV AHKSCLPACV YGPAVASGYD FEREGYSLVG 301 IDPFRLLQNS QVYSLIRPNE NPAHKSQLVW MACHSAAFED LRVLSFIKGT KVLPRGKLST 361 RGVQIASNEN METMESSTLE LRSRYWAIRT RSGGNTNQQR ASAGQISIQP TFSVQRNLPF 421 DRTTIMAAFN GNTEGRTSDM RTEIIRMMES ARPEDVSFQG RGVFELSDEK AASPIVPSFD 481 MSNEGSYFFG DNAEEYDN SEQ ID NO: 13 (PR/8/34 NP full length; 498 aa)   1 MASQGTKRSY EQMETDGERQ NATEIRASVG KMIGGIGRFY IQMCTELKLS DYEGRLIQNS  61 LTIERMVLSA FDERRNKYLE EHPSAGKDPK KTGGPIYRRV NGKWMRELIL YDKEEIRRIW 121 RQANNGDDAT AGLTHMMIWH SNLNDATYQR TRALVRTGMD PRMCSLMQGS TLPRRSGAAG 181 AAVKGVGTMV MELVRMIKRG INDRNFWRGE NGRKTRIAYE RMCNILKGKF QTAAQKAMMD 241 QVRESRNPGN AEFEDLTFLA RSALILRGSV AHKSCLPACV YGPAVASGYD FEREGYSLVG 301 IDPFRLLQNS QVYSLIRPNE NPAHKSQLVW MACHSAAFED LRVLSFIKGT KVLPRGKLST 361 RGVQIASNEN METMESSTLE LRSRYWAIRT RSGGNTNQQR ASAGQISIQP TFSVQRNLPF 421 DRTTIMAAFN GNTEGRTSDM RTEIIRMMES ARPEDVSFQG RGVFELSDEK AASPIVPSFD 481 MSNEGSYFFG DNAEEYDN SEQ ID NO: 14 (Ox2R fragment)  1 TKLEDSPPCR NWSSASELNE TQEPFLNPTD YDDEEFLRYL WREYLHPKEY EWVLIAGYII 61 VFVVALIGNV LVCVAVWKNH HMRTVTNYFI VNLSLADVLV TITCLPATLV VDITETWFFG SEQ ID NO: 15 (Ox1R fragment)  1 MEPSATPGAQ MGVPPGSREP SPVPPDYEDE FLRYLWRDYL YPKQYEWVLI AAYVAVFVVA 61 LVGNTLVCLA VWRNHHMRTV TNYFIVNLSL ADVLVTAICL PASLLVDITE SWLFGHALCK SEQ ID NO: 16 (Orexin fragment) MNLPSTKVSWAAV SEQ ID NO: 17 (fragment of SEQ ID NO: 1) YDKEEIRRIWRQ SEQ ID NO: 18 (X-179A NP fragment) LILYDKEEIRRIWRQ SEQ ID NO: 19 (X-181 NP fragment) LILYDKEEMRRIWRQ SEQ ID NO: 20 (X-179A NP fragment) VGKMIGGIGRFYIQM SEQ ID NO: 21 SGAAGAAVKGVGTMV SEQ ID NO: 22 EKATNPIVPSFDMSN SEQ ID NO: 23 IDPFKLLQNSQVVSL SEQ ID NO: 24 LILYDNEERRRRWRQ SEQ ID NO: 25 MNLPSTKVSWAAVTL SEQ ID NO: 26 MNLPSTKVSWAAVTL SEQ ID NO: 27 MNLPSMKVSWAAVTL SEQ ID NO: 28 LTVAAWSVKTSPLNM SEQ ID NO: 29 GAGNHAAGILTLGKR SEQ ID NO: 30 ASGNHAAGILTMGRR SEQ ID NO: 31 AMERNAGSGIIISDT SEQ ID NO: 32 ALNRGSGSGIITSDA SEQ ID NO: 33 ALSRGFGSGIITSNA

REFERENCES

[0323] [1] Zaracostas J.; BMJ. 2011 Feb. 9; 342:d909 [0324] [2] Nohynek et al., PLoS One 2012; 7(3):e33536. [0325] [3] Partinen et al. PLoS One 2012; 7(3):e33723. [0326] [4] Dauvilliers et al.; Sleep 2010; 33(11):1428-1430. [0327] [5] Szakacs et al.; Neurology 2013; 80(14):1315-1321. [0328] [6] http:/www.who.int/influenza/vaccines/virus/recommendations/en/ [0329] [7] http://www.fda.gov/AdvisoryCommittees/CommitteesMeetingMaterials/BloodVaccinesandOther Biologics/VaccinesandRelatedBiologicalProductsAdvisoryCommittee/default.htm [0330] [8] Siebold et al.; Proc Natl Acad Sci USA 2004; 101(7):1999-2004. [0331] [9] Chaloupka et al 1996. Eur J Clin Microbiol Infect Dis. 1996 February; 15(2):121-7 [0332] [10] Williams et al, 2012, Vaccine 30: 2475-2482. [0333] [11] Getie-Kebtie et al., 2013, Influenza and Other Respiratory Viruses 7(4), 521-530. [0334] [12] Kalandadze et al., 1996, J. Biol Chem. 271(33): 20156-20162. [0335] [13] Justesen et al., 2009, Immunome Research 5: 2 [0336] [14] Kistner et al. (1998) Vaccine 16:960-8. [0337] [15] Kistner et al. (1999) Dev Biol Stand 98:101-110. [0338] [16] Bruhl et al. (2000) Vaccine 19:1149-58. [0339] [17] WO2006/108846. [0340] [18] Pau et al. (2001) Vaccine 19:2716-21. [0341] [19] http://www.atcc.org/ [0342] [20] http://locus.umdnj.edu/ [0343] [21] WO97/37000. [0344] [22] Brands et al. (1999) Dev Biol Stand 98:93-100. [0345] [23] Halperin et al. (2002) Vaccine 20:1240-7. [0346] [24] EP-A-1260581 (WO01/64846). [0347] [25] WO2006/071563. [0348] [26] WO2005/113758. [0349] [27] Grachev et al. (1998) Biologicals; 26(3):175-93. [0350] [28] WO97/37001 [0351] [29] WO02/28422. [0352] [30] WO02/067983. [0353] [31] WO02/074336. [0354] [32] WO01/21151. [0355] [33] WO02/097072. [0356] [34] WO2005/113756. [0357] [35] Huckriede et al. (2003) Methods Enzymol 373:74-91. [0358] [36] Vaccines. (eds. Plotkins & Orenstein). 4th edition, 2004, ISBN: 0-7216-9688-0 [0359] [37] Treanor et al. (1996) J Infect Dis 173:1467-70. [0360] [38] Keitel et al. (1996) Clin Diagn Lab Immunol 3:507-10. [0361] [39] Herlocher et al. (2004) J Infect Dis 190(9):1627-30. [0362] [40] Le et al. (2005) Nature 437(7062):1108. [0363] [41] WO2008/068631. [0364] [42] Rota et al. (1992)J Gen Virol 73:2737-42. [0365] [43] GenBank sequence GI:325176. [0366] [44] Gennaro (2000) Remington: The Science and Practice of Pharmacy. 20th ed, ISBN: 0683306472. [0367] [45] Banzhoff (2000) Immunology Letters 71:91-96. [0368] [46] Nony et al. (2001) Vaccine 27:3645-51. [0369] [47] EP-B-0870508. [0370] [48] U.S. Pat. No. 5,948,410. [0371] [49] WO2007/052163. [0372] [50] WO2007/052061 [0373] [51] WO90/14837. [0374] [52] Podda & Del Giudice (2003) Expert Rev Vaccines 2:197-203. [0375] [53] Podda (2001) Vaccine 19: 2673-2680. [0376] [54] Vaccine Design: The Subunit and Adjuvant Approach (eds. Powell & Newman) Plenum Press 1995 (ISBN 0-306-44867-X). [0377] [55] Vaccine Adjuvants: Preparation Methods and Research Protocols (Volume 42 of Methods in Molecular Medicine series). ISBN: 1-59259-083-7. Ed. O'Hagan. [0378] [56] Gargon et al. (2012) Expert Rev Vaccines 11:349-66. [0379] [57] WO2008/043774. [0380] [58] Allison & Byars (1992) Res Immunol 143:519-25. [0381] [59] Hariharan et al. (1995) Cancer Res 55:3486-9. [0382] [60] US-2007/014805. [0383] [61] US-2007/0191314. [0384] [62] Suli et al. (2004) Vaccine 22(25-26):3464-9. [0385] [63] WO95/11700. [0386] [64] U.S. Pat. No. 6,080,725. [0387] [65] WO2005/097181. [0388] [66] WO2006/113373. [0389] [67] Potter & Oxford (1979) Br Med Bull 35: 69-75. [0390] [68] Greenbaum et al. (2004) Vaccine 22:2566-77. [0391] [69] Zurbriggen et al. (2003) Expert Rev Vaccines 2:295-304. [0392] [70] Piascik (2003) J Am Pharm Assoc (Washington D.C.). 43:728-30. [0393] [71] Mann et al. (2004) Vaccine 22:2425-9. [0394] [72] Halperin et al. (1979) Am J Public Health 69:1247-50. [0395] [73] Herbert et al. (1979) J Infect Dis 140:234-8. [0396] [74] Chen et al. (2003) Vaccine 21:2830-6. [0397] [75] Needleman & Wunsch (1970) J. Mol. Biol. 48, 443-453. [0398] [76] Rice et al. (2000) Trends Genet 16:276-277. [0399] [77] Current Protocols in Molecular Biology (F. M. Ausubel et al., eds., 1987) Supplement 30. [0400] [78] Smith & Waterman (1981) Adv. Appl. Math. 2: 482-489. [0401] [79] Renfrey S, Watts A. Vaccine 1994; 12(8):747-752. [0402] [80] Carlson et al.; Arthritis Rheum 1999; 42(12):2705-2709. [0403] [81] Han et al.; Ann Neurol 2011; 70(3):410-417. [0404] [82] Ettinger et al.; J Immunol 2006; 176(3): 1988-1998. [0405] [83] Bui et al.; Proc Natl Acad Sci USA 2007; 104(1):246-251 [0406] [84] Doucet et al.; J Gen Virol 2011; 92(Pt 5):1162-1171. [0407] [85] Yewdell et al.; Proc Natl Acad Sci USA 1985; 82(6):1785-1789. [0408] [86] Willie et al.; Neuron 2003; 38(5):715-730. [0409] [87] Berkhoff et al.; J Virol 2005; 79(17):11239-11246. [0410] [88] Steinitz et al. Blood 2012; 119(17):4073-4082.