Influenza virus antibody compositions

Abstract

The present invention relates to unique CDR3 amino acid sequences, and to antibodies comprising said sequences. Also provided are therapeutic applications of said antibodies for use in treating or preventing influenza infection, diagnostic applications of said antibodies for use in detecting influenza virus, potency testing of influenza virus vaccines, and screening applications of said antibodies for use in designing a universal influenza vaccine.

Claims

1. A composition, comprising a single domain camelid antibody, wherein said single domain antibody comprises a CDR3 amino acid sequence selected from SEQ ID NOS: 1-38.

2. The composition of claim 1, wherein said single domain camelid antibody demonstrates antigenic cross-reactivity to HI and H5 influenza virus subtypes.

3. The composition of claim 1, wherein said single domain camelid antibody demonstrates antigenic cross-reactivity to H2 and H9 influenza virus subtypes.

4. The composition of claim 1, wherein said single domain camelid antibody demonstrates antigenic cross-reactivity to one or all of H2, H5, H6, H9, HI 1, H13, H16, H8, and H12 influenza virus subtypes.

5. The composition of claim 1, wherein said-single domain camelid antibody further comprises a CDR2 amino acid sequence selected from SEQ ID NOS: 39-76.

6. The composition of claim 1, wherein said single domain camelid antibody further comprises a CDR1 amino acid sequence selected from SEQ ID NOs: 77-114.

7. The composition of claim 1, wherein said single domain camelid antibody comprises an amino acid sequence selected from SEQ ID NOS: 115-133.

8. A composition comprising two or three or four or more covalently linked single domain camelid antibodies, wherein said two or three or four or more single domain camelid antibodies are the same or different, and each comprises a CDR3 amino acid sequence selected from SEQ ID NOs: 1-38.

9. The composition of claim 1, wherein said single domain camelid antibody comprises a humanized framework region.

10. A composition comprising a single domain camelid antibody, wherein said single domain antibody comprises one or more covalently linked CDR3 V.sub.HH amino acid sequence, each selected from SEQ ID NOs: 1-38.

11. The composition of claim 1, wherein said single domain camelid antibody lacks one or more human Vhh amino acid sequence, and in place thereof consists or comprises one or more covalently linked CDR3 VHh amino acid sequence, each selected from SEQ ID NOs: 1-38.

12. The composition of claim 1, wherein said single domain camelid antibody demonstrates neutralising activity against a HI (e.g. H1N1) influenza virus, for example against strain A/Califomia/06/09 and/or A/California/7/2009 and/or A/Brisbane/59/2007.

13. The composition of claim 1, wherein said single domain camelid antibody demonstrates neutralising activity against a H5 (e.g. H5N1 or H5N3) influenza virus, for example against strain A/Vietnam/1203/2004 and/or A/Vietnam/1194/04 and/or A/Vietnam/119/2004 and/or HongKong/213/2003 and/or A/turkey/Turkey/1/2005 and/or A/Indonesia/05/2005 and/or A/Duck/Singapore-Q/119-3/97.

14. The composition of claim 1, wherein said single domain camelid antibody demonstrates neutralising activity against a H2 influenza virus (e.g. H2N2 or H2N3), for example against strain A/Japan/305/1957 and/or A/Mallard/Eng/727/06.

15. The composition of claim 1, wherein said single domain camelid antibody demonstrates neutralising activity against a H9 influenza virus (e.g. H9N2), for example against strain A/HongKong/1073/99.

16. The composition of claim 1, wherein said single domain camelid antibody demonstrates neutralising activity against one, or more, or all Group I haemagglutinin influenza viruses selected from the group consisting of HI, H2, H5, H6, HI 1, H13, H16, H9, H8 and/or H12 influenza viruses.

Description

EXAMPLES

Example 1: Immunization Procedure and Analysis of Immune Response to Influenza

(1) Juvenile male alpacas were maintained at the Royal Veterinary College, Hertfordshire, UK. All animal experiments were performed in accordance with Home Office licence 80/2117. A blood sample prior to immunisation was obtained from the jugular vein and this was followed by 4 intramuscularly injections at 0 day, 21 day, 43 day and 71 day intervals divided between 4 legs. The primary immunisation consisted of 50 μg of recombinant H1 (A/California/07/09 (H1N1pdm) (Protein Sciences™) in 400 μl of sterile PBS and emulsified with 800 μl of complete Freunds adjuvant (Sigma) just prior to immunisation. Similarly three separate booster injections of 50 μg of recombinant H1 (A/California/07/09 (H1N1pdm) (Protein Sciences™) in incomplete Freunds adjuvant (Sigma) were administered. Approximately 3 days after each injection a 10 ml blood sample was collected from which serum was prepared after allowing the blood to clot overnight at 4° C. The serological response to A/California/07/09 at each stage of the immunisation schedule was evaluated by ELISA, haemagglutination inhibition assay (HI assay) on a panel of inactivated viral strains and microneutralisation assays (MN assays) using the pandemic H1N1 viral strain A/California/07/09 (Table 1). A clear response to the immunogen (H1N1 pdm) was seen which increased with each subsequent immunisation.

(2) TABLE-US-00002 TABLE 1 Assessment of the serological immune response in immunised alpacas. Immunisation strategy HI titre.sup.2 MN titre.sup.3 A/California/ B/Brisbane/ A/Uruguay/ A/HongKong/ A/Vietnam/ A/California/ Immuni- 07/09 60/08 716/07 213/03 1194/04 07/09 Antigen zation Bleed Day (H1N1) pdm Control (H3N2) (H5N1) (H5N1) (H1N1) pdm Primary Pre-immune 0 <8 <8 <8 <8 <8 <10 Rec HA.sup.1 (H1) Boost 1 First bleed 21 64 <8 <8 .sup. ND.sup.4 ND ND Rec HA (H1) Boost 2 Second bleed 43 2048 <8 <8 ND ND ND Rec HA (H1) Boost 3 Third bleed 71 8192 <8 <8 <8 <8 >1280 .sup.1Rec HA (H1), recombinant haemagglutinin derived from A/California/07/09 pandemic H1N1 viral strain. .sup.2HI titre, haemagglutination inhibition assay and titres are indicated as the reciprocal of the minimum serum dilution at which inhibition of agglutination of turkey erythrocytes is seen. .sup.3MN titre, microneutralisation titre is given as the reciprocal of the minimum dilution at which lysis inhibition is seen of MDCK cells by influenza strain A/California/07/09 (H1N1 pdm). .sup.4ND is not determined.

Example 2: Phage Displayed Single Domain VHH Antibody Library Construction

(3) For antibody library construction approximately 10 ml samples of blood were collected from an immunized alpaca (Example 1) into heparin tubes. Peripheral blood lymphocytes were purified using a ficol hypaque centrifugation procedure (Sigma). RNA was extracted using RiboPure™ RNA extraction kit (Novagen) according to manufacturer's instructions and stored at −70° C. prior to cDNA synthesis. First strand cDNA synthesis was performed according to manufacturer's instructions (Invitrogen). In short, up to 200 ng of total RNA was used per oligo dT primed cDNA synthesis reaction which was allowed to proceed for 50 minutes at 50° C. after which time the reaction was terminated by incubation at 85° C. for 5 minutes followed by placing the reaction on ice. The reactions were collected by brief centrifugation and 1 μl of RNase H was added to each tube and incubated for 20 minutes at 37° C. The resulting cDNA was either stored at −20° C. or used immediately in primary PCR reaction to amplify antibody gene sequences.

(4) Primary PCR was performed using two oligonucleotide primers designed to universally prime mammalian immunoglobulin genes in the CH2 domain and also at the 5′ end of alpaca VHH genes (Harmsen et al., 2000) within the signal sequence. High Fidelity taq polymerase (Roche) was used for all DNA amplification.

(5) TABLE-US-00003 Alp-CH2_Rev 5′ CGC CAT CAA GGT ACC AGT TGA AlpL-Fw_VHH 5′ GGT GGT CCT GGC TGC

(6) This resulted in the amplification of both a 600 bp and 900 bp antibody gene products. The 600 bp product corresponds to the heavy chain only antibody population and does not contain a CH1 domain. The 900 bp gene product corresponds to the conventional antibody heavy chain population and is not required. The 600 bp heavy chain only gene product was gel purified (QIAGEN) and then subjected to a secondary PCR step to amplify just the VHH antibody gene population (˜450 bp) and to append appropriate restrictions sites for cloning into the phage display vector pNIBS-1 (Example 3). Primers for secondary PCR were designed to amplify all VHH genes by annealing to the antibody FR1 and FR4 regions. The Sfi1 and Not1 restriction sites were chosen for cloning as they do not cut frequently within antibody genes.

(7) TABLE-US-00004 Alp_FR1_Sfi1 5′ CTG CAG GGA TCC GTT AGC AAG GCC CAG CCG GCC ATG GCA CAG KTG CAG CTC GTG GAG TCN Alp_FR4back_Not1 5′ GCT AGT GCA TGG AGC TCA TGC GGC C

(8) Approximately 5 μg of VHH antibody DNA was digested with Sfi1 restriction enzyme (New England Biolabs) in a 200 μl reaction overnight at 50° C. After further purification (QIAGEN) the VHH genes were digested with Not1 in a 100 μl reaction at 37° C. for 6 hours. The digested VHH genes were then ligated into the phage display vector pNIBS-1 (Example 3) which was similarly digested with Sfi1 and Not 1 restriction enzymes. Ligations were set up in a ratio of 4:1 insert to vector in a 20 μl reaction with each reaction containing approximately 100 ng of vector. After the addition of T4 DNA ligase (New England Biolabs) the reactions were incubated overnight at 16° C. The individual ligations were pooled and purified (QIAGEN) according to manufacturer's instructions. The purified ligation mix was then transformed into TG1 electrocompetent cells (Agilent) using electroporation (BIO-RAD). Cells were recovered from electroporation cuvettes using 1 ml SOC medium (Sigma) pre-warmed to 37° C. and then incubated at 37° C. for 1 hour with shaking to allow expression of antibiotic resistance. Libraries were spread on 22 cm dishes containing 2 TY agar ampicillin (100 μg/ml) and 20% v/v glucose. The library size was determined by preparing a dilution series counting viable colonies and was 5×10.sup.7 independent clones. After overnight incubation of the 22 cm dishes at 37° C. the libraries were harvested by flooding the plates 2 TY agar carbenicillin (100 μg/ml), 20% v/v glucose and scraping off the colonies. The resuspended colonies representing the antibody library were then stored at −70° C. in 20% v/v glycerol.

Example 3: Construction of Phage Display Vector pN1BS-1

(9) A polylinker sequence of 1782 bp was designed comprising the key features below: (i) A pelB signal sequence which drives the secretion of expressed antibody into the periplasm and whose c terminal amino acid and DNA sequence is compatible with encoding a Sfi1 restriction site. (ii) A hexa-histidine tag which is fused to the C-terminal end of a cloned VHH antibody and facilitates both purification and detection of expressed antibodies. (iii) A c-Myc epitope tag attached to the C-terminal end of the cloned antibody after the hexa-histidine tag and facilitates detection of expressed antibody (iv) Gene III sequence which encodes the M13 bacteriophage coat protein 3 and is in frame with antibody cassette to facilitate the production of antibody-gene III fusion products which can be displayed on the surface of phage. (v) An amber DNA stop codon between the c-Myc epitope tag and the gene III phage coat protein. This codon is read as a stop codon in Escherichia coli but in strains carrying a mutant suppressor tRNA gene it is frequently read as an encoding amino acid and read through to produce a antibody-gene 3 fusion product. This allows display of the antibody-gene 3 fusion on phage. Suppression of the amber codon is generally of low efficiency and this has the advantage that soluble antibody not fused to gene 3 can be produced from the same E. coli clone cell without sub-cloning to remove the gene 3 sequence. (vi) Sfi1 and Not1 restriction sites for cloning antibody genes were included as they cut infrequently in antibody genes and can be placed within DNA sequence whilst maintaining a functional expression cassette. Other restriction site like Mfe1 and BstEII were also incorporated as these are highly conserved in antibody germline genes and can be useful for cloning selected antibodies into other formats

(10) The polylinker sequence was synthesised from oligonucleotides (GENEART™) as a Hind111/EcoR1 fragment and cloned into the standard phagemid vector pUC119 to generate pNIBS-1 (FIG. 1A and FIG. 1B).

Example 4: Phage Antibody Library Selection Strategy

(11) The phage displayed antibody library (Example 2) was inoculated into 50 mls of 2×TY media supplemented with carbenicillin at 100 μg/ml (w/v) and 2% v/v glucose (2×TY AG). An inoculation of greater than 10× library size was used to ensure library diversity is maintained during the phage rescue procedure [Library size 5×10.sup.7 in size therefore 5×10.sup.8 cells were used as a starter inoculation from a glycerol stock]. The culture was grown at 37° C. to an OD600 of 0.5 and 5 mls of culture (corresponding to approximately 7.5×10.sup.8 bacteria) were transferred to a 50 ml sterile tube containing M13KO7 helper phage (New England Biolabs). The multiplicity of infection (MOI) was approximately 20:1 phage to bacteria so approximately 1.5×10.sup.10 phage were used to infect the bacterial culture. After the bacterial culture was added to the appropriate amount of M13KO7 helper phage it was incubated in a water bath at 37° C. for 30 minutes. The infected cells were then centrifuged for 10 minutes at 4000×g, the supernatant removed and the bacterial pellet resuspended in 25 mls of pre-warmed 2×TY supplemented with 25 μg/ml kanamycin and 100 μg/ml ampicillin (2×TY AK) and incubated overnight at 30° C. with shaking. The next day cells were pelleted by centrifugation for 15 minutes at 4,000×g and the supernatant containing the phage displayed antibody library collected. To purify the phage 1/5 volume of 20% polyethylene glycol 6000, 2.5M NaCl (PEG) and incubated on ice for 1 hour. The phage were then pelleted in a microfuge and resuspended in 1 ml of sterile PBS. The remaining bacteria were then removed by a brief centrifugation and the phage supernatant transferred to a new tube. A second purification was performed by adding 200 μl of PEG and incubating on ice for 20 mins. The phage were collected by centrifugation for 5 minutes at 4° C. The pelleted phage were then resuspended in 1 ml sterile PBS and any further contaminating bacteria removed by a further brief centrifugation. The purified phage were then used for selection.

(12) Phage antibody library selections were performed in immunotubes (Nunc) coated with 1 ml of 10 μg/ml recombinant haemagglutinin in PBS overnight at 4° C. [recombinant H1 [A/California/07/2009 (pandemic H1N1), recombinant H5 (A/Vietnam/1203/2004) protein Sciences™] (FIG. 2). The selection strategy was designed to recover both H1 specific antibodies and also H1/H5 cross reactive antibodies by alternating selection between immunotubes coated with H1 and H5 (FIG. 2). A parallel selection with an empty immunotube was also used to monitor antigen specific enrichment of the phage antibody library. The next day the immunotubes tube and control tube were blocked by filling completely with 2% w/v milk powder in sterile PBS (M-PBS) followed by a 2 hour incubation at room temperature. 500 μl of purified concentrated phage (10.sup.12-10.sup.13 phage) were added to 0.5 mls 4% MPBS and the phage blocked for 1 hour at room temperature. The immunotube was then washed with PBS Tween 20 (0.1% v/v) and 2×PBS and the phage mix transferred to the immunotubes covered with parafilm and incubated for 30 minutes on rotator and then 1.5 hours standing at room temperature. The immunotube was then washed 20 times with PBS Tween (0.1% v/v) and then 20 times with PBS [10 washes were used for the first round of selection to maximise the percentage recovery of antigen specific phage antibodies]. Specific phage antibodies were eluted by adding 1 ml of 100 mM triethylamine to the tube and covering with fresh parafilm followed by incubation for 10 min on rotator at room temperature. The eluted phage were then neutralised by transferring to a fresh microtube containing 0.5 ml 1M Tris HCl pH7.5. The eluted phage were then infected into E. coli to both prepare phage for a second round of selection and also to titrate the amount of phage recovered between sequential rounds of selection.

(13) A culture of E. coli ER2378 cells was grown to an OD 600 of approximately 0.5. To amplify the selected phage for a next round of selection 1 ml of eluted phage were mixed with 5 mls of ER2738 culture and 4 mls of 2×TY media. This was then incubated in a water bath at 37° C. for 30 minutes. This was then spread onto 22 cm bioassay dishes 2×TY AG plus tetracycline (20 μg/ml) and grown overnight at 30° C. The bioassay dishes were then scraped and the recovered cells used for a second round of selection as above. In order to monitor the progress of selections between sequential round of selection the titre of input phage and output phage were determined. A serial dilution of phage were made in 500 μl 2TY to which 500 μl of log phage ER2738 culture was added and incubated in a water bath for 30 mins at 37° C. The infected cells were spread on 2×TY AG tetracycline (20 μg/ml) agar plates and the colonies counted after overnight incubation at 30° C. The phage titre before and after selection for three rounds was calculated as pfu/ml and showed an increase in the recovery of phage indicative of antigen selective enrichment of specific phage antibodies (Table 2).

(14) TABLE-US-00005 TABLE 2 Selection and screening of phage displayed single domain antibody library Input.sup.2 Output.sup.2 Titre Titre Number Number Round Antigen.sup.1 (pfu/ml) (pfu/ml) binders H5.sup.3 binders H1.sup.3 R1a H1   2 × 10.sup.13 5.6 × 10.sup.8   5/24 10/24 R2a H1   4 × 10.sup.13   4 × 10.sup.10 — 12/24 R2b H5   4 × 10.sup.13   2 × 10.sup.10 16/24 — R3a H1 2.3 × 10.sup.13   8 × 10.sup.10 — 12/24 R3b H5 2.3 × 10.sup.13 4.8 × 10.sup.10 19/24 — .sup.1recombinant antigen is either H1 derived from A/California/07/09 (H1N1 pdm) or H5 derived from A/Vietnam/1203/04 (H5N1) .sup.2phage titres before and after selection are given as plaque forming units per ml (pfu/ml) .sup.3number of antibodies binding to either recombinant H1(A/California/07/09) or recombinant H5 (A/Vietnam/1203/04) out of total number screened

Example 5: Screening of Selected Library for Antibody Clones Binding to Recombinant H1 (A/California/07/09) and Recombinant H5 (A/Vietnam/1203/2004)

(15) Primary screening of selected antibodies was done using unpurified soluble VHH antibodies induced in a 96 well format. Individual colonies from each round of selection were inoculated into 100 μl of 2×TY AG (2%) in 96 well flat bottom plate (Costar) using sterile toothpicks. The plates were incubated overnight at 30° C. rotating at 255 rpm. The next day 35 ml of 60% v/v glycerol was then added to each well and the master plate stored at −70° C. From this master plate a new 96 well round bottom plate (Costar) containing 120 μl of 2×TY ampicillin 0.1% w/v glucose was inoculated and grown for 6 hours at 37° C. until OD 600 is approximately 0.9. After which 30 μl 2×TY ampicillin 100 μg/ml plus 5 mM IPTG (1 mM final concentration) was added and continue shaking at 37° C. overnight. The plates were then centrifuged 600×g for 10 minutes and the supernatant containing soluble VHH antibodies was used to test antigen specific reactivity in ELISA.

(16) A 96 well plate (Nunc) was coated with recombinant haemagglutinin at 1 μg/ml overnight in PBS at 4° C. The plates were washed three times with PBS and then blocked with 120 μl/well 2% w/v milk powder PBS (M-PBS) for 1 hour at room temperature. The plates were then washed twice each with PBS 0.1% (v/v) Tween and PBS. To each well 50 μl of 4% (w/v) milk powder PBS was added to each well and then 50 μl of culture supernatant containing soluble VHH. The ELISA plate was then incubated for 1.5 hours on a shaking platform at room temperature. After this incubation the plates were washed three times each with PBS 0.1% (v/v) Tween 20 and three times with PBS. To detect antibody binding 100 μl of anti myc 9E10-HRP (1/1000 dilution) (Roche) in 2% MPBS was added to each well and incubated for 1 hour at room temperature on a shaking platform. Plates were then washed three times each with PBS 0.1% (v/v) Tween 20 and three times with PBS and antibody binding developed using standard TMB staining. After stopping the reaction with 50 μl of 1M sulphuric acid the plates were read at 450 nM. An example of a primary ELISA screen using unpurified soluble VHH antibodies in a 96 well ELISA format is shown in FIG. 3A and FIG. 3B. Positive clones were scored as having at least twice background signal and were further characterised to analyse sequence diversity in particular the number of times a particular clone was recovered in the clones screened (Table 3). Using sequence analysis and evaluation by ELISA, nineteen antibodies with a unique VHH CDR3 sequence with specificity for recombinant H1 and/or recombinant H5 were isolated and progressed for more detailed secondary screening (Example 6).

(17) TABLE-US-00006 TABLE 3 CDR sequences of unique antibodies specific to pandemic H1N1 SEQ     Clone.sup.1 ID.sup.2 Number .sup.3 V.sub.HH CDR1 .sup.4 V.sub.HH CDR2 .sup.4 V.sub.HH CDR3 .sup.4 R1a G5  1  4 IVTMG VIGNYGNTNYADSVKR STTTPPYEY R2b D8  2  2 WYDVG DIASTRGTTNYADSVKG RRDWRDY R2b E8  3  4 RYRMG GITYDDSTNYAGSVKG NPPGNLY R2b D9  4  1 LYTMG AITSGESTNYADSVKG DPLSTGWGQYSY R1a F4  5  2 IVTMG VIGNYGNTNYADSVKG STTTPPHEF R1a A5  6 34 TYPMS AVTTDGSTSYADYAKG RDGFFNRYDY R1a C5  7  1 FYTMG VIGNGGNTNYAESVKG SGPGGLNV R1a E5  8  3 NYAIG FITSTSAVTKYADSVKG TRWVPTMKADEYNY R1a F5  9  1 GYAIA CRASDGNTYYAESLKG ASWVASLWSPSEYDY R1a B6 10  1 RYRMG SIAYDGSTSYADPVKG DPPGILY R1a E6 11  1 AYAIA CISPSDSFTEYGDSVKG DPVCTAGWYRPSRFDL R1a G6 12  1 IVTMG VIGNNDNTVYGDSVQG STTTPPYEY R1a H6 13  1 LYRVG AIDWGDGPTTYADSVKG GNTGSSDRSSSYVH R2a E8 14  1 MYMID SIDGRGTPMYADSVKG KSPLVDNEY R2a G8 15  1 NNAIG CMNSRDGTTNYADSVKG KGFAPFLIGCPWGKAEYDY R2a B9 16  1 INAMG AITAGGNTYYADSAKA TKAFGIATITADYEL R2a F9 17  1 IVTMG VIGNYGNTNYADSVKG SSTVAPHEY R2a G9 18  1 FYTMG VIGNGGNTNYADSVKG SGPGGVEV R2a H9 19  1 FYTMG VIGNGGTNYADSVKG SGPGGVIL .sup.1Clone identification number .sup.2Sequence identification .sup.3 Number of times a particular CDR3 sequence appeared in screening of 62 independently isolated clones .sup.4 CDR defined according to Kabat numbering scheme

Example 6: Analysis of Purified Single Domain VHH Antibodies on Different Influenza Sub-Types

(18) For more detailed screening of antibody specificity, purified soluble VHH antibodies were purified from the E. coli periplasm. The nineteen different antibodies identified in Example 4 were each inoculated into into 2 ml of 2YT medium supplemented with 100 μg/ml (w/v) ampicillin, 2% (w/v) glucose, Tet (5 μg/ml) in a 10 ml tube and grown overnight at 37° C. with shaking. From this starter a 50 μl inoculation was taken into 50 mls of 2×TY medium supplemented with 100 μg/ml ampicillin, 0.1% (v/v) glucose and grown to an OD of ˜0.6 at 37° C. Soluble antibody expression was induced with the addition of 1 mM IPTG followed by further incubation overnight at 30° C. The 50 ml culture was centrifuged for 20 minutes and the bacterial pellet re-suspended in 2.5 mls of periplasmic buffer (200 mM Tris HCl, pH 7.5, 20% (v/v) sucrose, 1 mM EDTA) followed by incubation on a rotator for 1 hour at 4° C. Cell debris was removed by centrifugation twice and the supernatant corresponding to the soluble VHH antibodies recovered from the bacterial periplasm were dialysed (Pierce) into PBS with two changes of buffer. After removal of EDTA by dialysis the antibody was purified by metal chelate chromatography (Clontech) according to manufacturer's instructions. Approximately 1 ml of purification resin was used per antibody. After incubation of the resin with the periplasmic preparations it was transferred to a gravity column (Biorad) and washed according to manufacturers instructions. The purified antibody was eluted into two bed volumes of elution buffer containing 100 mM imadazole. Antibody was further dialysed against PBS with two changes of buffer and antibody concentration determined using OD 280.

(19) Inactivated viral preparations were reconstituted in 1 ml PBS and a 96 well ELISA plate was incubated overnight at 4° C. with 50 μl of antigen resuspended in 10 mls of 0.5M bicarbonate buffer pH9.6. Plates were washed and processed as in Example 4 except a serial dilution of purified antibody was used of 30, 10, 3, 1, 0.3, 0.1, 0.03, 0.01, 0.003, 0.001, 0.0003, 0.0001 μg/ml and assays were performed in duplicate. Dose response curves were plotted and EC50 values calculated using GraphPad Prism (Table 4). From this analysis cross-reactive antibodies were seen as R2b-E8, R2b-D9, R1a-A5, R1a-C5, R1a-B6, R1a-H6, and R2a-G8.

(20) TABLE-US-00007 TABLE 4 Analysis of selected antibodies on different influenza viral strain sub-types A/Cali- A/Bris- A/Bris- A/ A/Viet- A/ A/Indo- A/Duck/ A/ A/ fornia/ bane/5 bane B/Bris- TurkTurk/ nam/1 HongKong/ nesia/ Sing- HongKong/ mallard/E 7/2009 9/2007 10/2007) bane/ 1/2005 19/2004 213/2003 05/2005 Q/97 1073/99 ng/727/06 (H1N1) (H1N1-like) (H3N2) 60/2008 (H5N1) (H5N1) (H5N1) (H5N1) (H5N3) (H9N2) (H2N3) EC50.sup.1 EC50 EC50 EC50 EC50 EC50 EC50 EC50 EC50 EC50 EC50 Clone (μg/ml) (μg/ml) (μg/ml) (μg/ml) (μg/ml) (μg/ml) (μg/ml) (μg/ml) (μg/ml) (μg/ml) (μg/ml) R1a-G5 1.73 0.29 — — — — — — — — — R2b-D8 — 14.8 — — — — — — — — — R2b-E8 3.64 18.4 — — 0.89 1.83 1.2 0.80 4.98 5.63 0.002 R2b-D9 0.90 0.80 — — 0.76 1.81 2.8 0.58 5.64 — 0.06  R1a-F4 0.79 — — — — — — — — — — R1a-A5 0.31 0.38 — — 0.13 0.25  0.27 0.15 0.27 — 0.006 R1a-C5 1.71 3.30 — — 0.15 2.89 2.9 1.17 7.57 4.48 0.06  R1a-E5 3.23 — — — — — — — — — — R1a-F5 0.55 0.70 — — — — — — — — — R1a-B6 0.39 0.58 — — 0.44 0.82  0.37 0.42 1.32 2.3  0.102 R1a-E6 2.46 — — — — — — — — — — R1a-G6 0.30 — — — — — — — — — — R1a-H6 3.85 — — — — — — — — 5.7  — R2a-E8 0.51 1.42 — — — — — — — — R2a-G8 4.11 1.86 — — 1.93 3.36 6.2 5.7  9.67 — 2.5  R2a-B9 1.00 9.84 — — — — — — — — — R2a-F9 0.19 — — — — — — — — — — R2a-G9 0.30 — — — — — — — — — — R2a-H9 0.97 — — — — — — — — — — control — — — — — — — — — — — .sup.1EC50 is the concentration of antibody in μg/m1 which gives 50% of the maximal binding effect of each purified VHH single domain antibody on a representative group of viral sub-types using ELISA. — No binding observed

Example 7: Affinity of Single Domain Antibodies on Recombinant Haemagglutinin from Different Influenza Viral Sub-Types

(21) Affinity of purified VHH antibodies (Example 5) was determined using the single cycle kinetics procedure on a BIAcore T100 machine (Karlsson et al., 2006). The procedure involves sequentially injecting an antibody concentration series over a haemagglutinin surface without regeneration between injections. Purified recombinant haemagglutinin of varying subtype (Protein Sciences, E enzymes) was immobilised onto a CM5 chip to approximately 3000 resonance units (RU). A series of antibody dilutions of 100 nM, 50 nM, 25 nM, 10 nM 5 nM were sequentially run over the different antigen surfaces without regeneration. A reference surface was subtracted prior to evaluation of the sensograms. Antibody affinity was determined using single cycle kinetics procedure of BIA evaluation software (GE healthcare) using a 1:1 fitting model (Table 5). Analysis on recombinant haemagglutinin and viral preparations (Example 6) showed antibodies R2b-E8, R2b-D9, R1a-A5, R1a-C5, R1a-B6, R1a-H6, and R2a-G8 to be cross-reactive antibodies.

(22) TABLE-US-00008 TABLE 5 Binding analysis on different recombinant haemagglutinin using surface plasmon resonance H1.sup.a H5.sup.b H2.sup.c H9.sup.d H3.sup.e H7.sup.f (H1N1 (H5N1) (H2N2) (H9N2) (H3N2) (H7N7) SEQ pdm) Kd Kd Kd Kd Kd Clone ID Kd (nM) (nM) (nM) (nM) (nM) (nM) R1a-G5  1 24 —.sup.g — — — — R2b-D8  2 —.sup.g — — — — — R2b-E8  3 8.2 3.6 — 87 — — R2b-D9  4 0.4 1.6 — 59 — — R1a-F4  5 7.3 — — — — — R1a-A5  6 3.3 0.7 186  — — — R1a-C5  7 1.9 2.7 67 22 — — R1a-E5  8 0.6 — — — — — R1a-F5  9 9.2 — — — — — R1a-B6 10 0.7 1.7 43   7.6 — — R1a-E6 11 22.8 — — — — — R1a-G6 12 4.4 — — — — — R1a-H6 13 — 241 — — — — R2a-E8 14 0.8 — 34 — — — R2a-G8 15 3.1 9 39 — — — R2a-B9 16 0.2 — — — — — R2a-F9 17 60 — — — — — R2a-G9 18 62 — — — — — R2a-H9 19 2.8 — 40 — — — -control — — — — — — — .sup.aSingle cycle affinity determination on full length recombinant H1 residues 18-530 derived from A/California/06/09. Affinity is given in nM to 1 decimal place and is the average of at least two separate assays. .sup.brecombinant H5 A/Vietnam/1203/04 (H5N1) .sup.crecombinant H2 A/Japan/305/1957 (H2N2) .sup.drecombinant H9 A/HongKong/1073/99 (H9N2) .sup.erecombinant H3 A/Brisbane/10/07 (H3N2) .sup.frecombinant H7 A/Netherlands/219/03 (H7N7). .sup.g— No Binding.

Example 8: Epitope Mapping

(23) Evaluation of whether the antibodies bound to the globular head of haemagglutinin was evaluated in BIAcore. Recombinant haemagglutinin HA1 domain (18-344) (A/California/06/09) H1N1 and whole haemagglutinin (18-530) (E Enzymes) was immobilised on a CM5 BIAcore chip (GE Healthcare). Affinity was determined using a single cycle kinetics procedure (Example 7) and the cross-reactive antibodies were shown not bind to the globular head domain HA1 and by inference likely bind to the HA2 stalk region. The HA2 stalk region is more conserved across viral sub-types and as such is consistent with the cross-reactivity of these antibodies.

(24) Further evidence localising cross-reactive antibodies to the more conserved haemagglutinin stalk region was obtained using an acid treatment which triggers an irreversible conformational change in the HA molecule. A viral preparation of H1N1 (A/California/06/09) was resuspended in 1 ml of PBS and 200 μl was diluted to 20 mls in PBS and the pH taken down to 2.83 with 1.5 mls of 1M HCl. After incubation for 2.5 hours at room temperature the antigen was neutralised with 20 mls of 0.5M bicarbonate buffer to give a final pH of ˜9.0. This acid treated antigen was then used to coat ELISA plates overnight at 4° C. The next day ELISA plates were blocked and antigen binding was determined as in Example 6. A parallel assay with antigen that had not been treated with acid was also tested. All antibodies were shown to bind H1N1 which had not been treated with acid whereas several antibodies were shown to lose binding after including the cross-reactive antibodies R2b-E8, R2b-D9, R1a-A5, R1a-C5, R1a-B6, R1a-H6. Antibodies shown to bind to the globular head generally retained binding after acid treatment. This provided further evidence that the cross-reactive antibodies were binding to the more conserved stalk region of haemagglutinin.

(25) Further evidence of cross reactive antibodies binding to the conserved stalk region was provided by assaying for the inhibition of heamaggluttination of turkey erythrocytes. Heamagglutinin (HA) is the major viral coat protein which mediates binding to cell surface sialic acid via the globular head domain (HA1). Heamagglutination inhibition assays were performed as in Harmon et al., (1988) and were used to determine which antibodies could block attachment of influenza virus to turkey erythrocytes, and to indicate localisation of the antibody epitope to the globular head domain (HA1). Briefly serial dilutions of serum from immunised alpacas or purified VHH antibodies starting from 250 μg/ml were prepared. Serial dilutions were incubated with 8 HA units of virus per well and turkey erythrocytes were added to a concentration of 0.5% and the plate was incubated at room temperature for 30 minutes. All antibodies which showed binding to the globular head (HA1) domain were also positive for inhibition of haeamgglutination, except R2a-G8. Of all the H1/H5 cross neutralising antibodies identified only R1a-C5 showed any inhibition of heamagglutiniation (Table 6).

(26) TABLE-US-00009 TABLE 6 Mapping of antibody epitopes to globular head or stalk region of haemagglutinin H1 HA1 HI titre (18-530).sup.1 (18-344).sup.2 (μg/ml) (pdm H1N1) (pdm H1N1) H1N1 Acid No acid Clone Kd (nM) Kd (nM) (X-181).sup.3 treatment .sup.4 treatment .sup.4 R1a-G5 24 26 >1.56 + + R2b-D8 — .sup.g — >25 + + R2b-E8 8.2 — >25 − + R2b-D9 0.4 — >25 − + R1a-F4 7.3 4.7 >2.08 + + R1a-A5 3.3 — >25 − + R1a-C5 1.9 734 >10.40 − + R1a-E5 0.6 — >25 − + R1a-F5 9.2 6.1 >25 ND + R1a-B6 0.7 — >25 − + R1a-E6 22.8 36 >25 ND + R1a-G6 4.4 4.6 >1.04 + + R1a-H6 — — >25 − + R2a-E8 0.8 — >25 − + R2a-G8 3.1 7.7 >25 .sup. ND.sup.4 + R2a-B9 0.2 — >25 − + R2a-F9 60 79 >1.04 + + R2a-G9 62 69 >1.30 + + R2a-H9 2.8 4.7 >1.30 + + -control — — >25 − + .sup.1Single cycle affinity determination on full length recombinant H1 residues 18-530 from A/California/06/09. Affinity given to 1 decimal place in nM .sup.2Single cycle affinity determination on globular head HA1 domain residues 18-344 from A/California/06/09. Affinity given to 1 decimal place in nM .sup.3Haemagglutination inhibition (HI) assay was given as the minimum dilution of antibody at which inhibition of agglutination of turkey erythrocytes in μg/ml of purified antibody. Values are given as an average of three independent assays. Laboratory adapted H1N1 strain (X-181) was used .sup.4 ELISA on viral preparation of A/California/07/09 (H1N1 pdm) treated with acid and corresponding control with no acid treatment .sup.4ND not determined

Example 9—Humanisation of a VHH Antibody (Exemplified by Reference to Camelid Sequence VHH R2a-F9)

(27) The framework regions (FR) of R2a-F9 were aligned with human germline sequence database using ClustalW and using sequences taken from V base (http://www.mrc-cpe.cam.ac.uk/vbase-ok)—see Table 7. The closest human germline sequence was DP53. The FR4 of R2a-F9 was aligned with the JH sequences and JH1 was chosen. The FR and CDR regions were defined according to Kabat et al., (1991). The IMGT numbering system (http://cines.fr) defines boundaries differently—for example cysteine 92 defines the start of CDR3 according to this nomenclature.

(28) Framework residues in R2a-F9 different from germline VH-DP53 sequence are bold and underlined. The sequence identity in the framework regions is 79%. There are a total of 17 residues in the frameworks which are different between alpaca and human germline variable gene DP53.

(29) The humanisation strategy involves replacing residues in FR that are not essential for antigen binding and stability. This involves choosing known typical camelid residues to be retained in the final sequence eg. ‘Hallmark residues’ (Vincke et al., 2008, Journal of Biological Chemistry, 284(5), 3273-3284, which is hereby incorporated in its entirety by reference thereto). In R2a-F9 five hallmark residues have been identified 4 of which are in the FR2. A series of variants are made and the importance of each lama residue to antigen binding and stability is assessed.

(30) All 12 constructs are tested in BIAcore for binding to H1, expression yield and thermal stability—see Table 8. Mutations to human germline which do not affect antigen binding are combined in a second round of constructs and the testing is repeated until a final humanised construct with the minimum amount of alpaca antibody sequence in the framework regions is obtained.

(31) Combine all mutations that do not affect function and stability in final constructs to create an antibody gene with maximum human sequence content in the framework regions. A total of 13 alpaca residues are converted to human germline are shown in Table 9.

(32) TABLE-US-00010 TABLE 7 1                            30 R2a-F9-h QVQLVESGGGLVQPGGSLRLSCSASGSISS DP53 EVQLVESGGGLVQPGGSLPLSCAASGFTFS CDR1- 3637         49 ------CDR2------- IVTMG  WYRQAPGKERELVA VIGNYGN-TNYADSVKG SYWMH  WVRQAPGKGLVWVS RINSDGSSTSYADSVKG 66  RFTVSRDNAKNTVYLQMNSLNVEDTAMYY  RFTISRNAKNTLYLQMNSLRAEDTAVYY  94 -CDR3---- R2a-F9-h CKF SSTVAPHEY WGQGTQVTVSS DP53(96) CAR --------- WGQGTLVTVSS

(33) TABLE-US-00011 TABLE 8 First round Mutant constructs R2a-F9_Q1E QVQLVESGGGLVQPGGSLRLSCSASGSISS IVTMG WYRQAPGKERELVA   VIGNYGNTNYADSVKG RFTVSRDNAKNTVYLQMNSLNVEDTAMYVCKF  SSTVAPHEY WGQGTQVTVSS R2a-F9_S23A EVQLVESGGGLVQPGGSLRLSCAASGSISS IVTMG WYRQAPGKERELVA   VIGNYGNTNYADSVKG RFTVSRDNAKNTVYLQMNSLNVEDTAMYVCKF  SSTVAPHEY WGQGTQVTVSS R2a-F9_S27F EVQLVESGGGLVQPGGSLRLSCSASGFISS IVTMG WYRQAPGKERELVA  VIGNYGNTNYADSVKG RFTVSRDNAKNTVYLQMNSLNVEDTAMYVCKF  SSTVAPHEY WGQGTQVTVSS R2a-F9_I28T EVQLVESGGGLVQPGGSLRLSCSASGSTSS IVTMG WYRQAPGKERELVA   VIGNYGNTNYADSVKG RFTVSRDNAKNTVYLQMNSLNVEDTAMYVCKF  SSTVAPHEY WGQGTQVTVSS R2a-F9_S29F EVQLVESGGGLVQPGGSLRLSCSASGSIFS IVTMG WYRQAPGKERELVA   VIGNYGNTNYADSVKG RFTVSRDNAKNTVYLQMNSLNVEDTAMYVCKF  SSTVAPHEY WGQGTQVTVSS R2a-F9_E46V EVQLVESGGGLVQPGGSLRLSCSASGSISS IVTMG WYRQAPGKERVLVA   VIGNYGNTNYADSVKG RFTVSRDNAKNTVYLQMNSLNVEDTAMYVCKF  SSTVAPHEY WGQGTQVTVSS R2a-F9_A49S EVQLVESGGGLVQPGGSLRLSCSASGSISS IVTMG WYRQAPGKERELVS   VIGNYGNTNYADSVKG RFTVSRDNAKNTVYLQMNSLNVEDTAMYVCKF  SSTVAPHEY WGQGTQVTVSS R2a-F9_V69I EVQLVESGGGLVQPGGSLRLSCSASGSISS IVTMG WYRQAPGKERELVA   VIGNYGNTNYADSVKG RFTISRDNAKNTVYLQMNSLNVEDTAMYVCKF  SSTVAPHEY WGQGTQVTVSS R2a-F9_N93R EVQLVESGGGLVQPGGSLRLSCSASGSISS IVTMG WYRQAPGKERELVA   VIGNYGNTNYADSVKG RFTVSRDNAKNTLYLQMNSLNVEDTAMYVCKF  SSTVAPHEY WGQGTQVTVSS R2a-F9_V94A EVQLVESGGGLVQPGGSLRLSCSASGSISS IVTMG WYRQAPGKERELVA   VIGNYGNTNYADSVKG RFTVSRDNAKNTVYLQMNSLRVEDTAMYVCKF  SSTVAPHEY WGQGTQVTVSS R2a-F9_M99V EVQLVESGGGLVQPGGSLRLSCSASGSISS IVTMG WYRQAPGKERELVA   VIGNYGNTNYADSVKG RFTVSRDNAKNTVYLQMNSLNAEDTAMYVCKF  SSTVAPHEY WGQGTQVTVSS R2a-F9_M99V EVQLVESGGGLVQPGGSLRLSCSASGSISS IVTMG WYRQAPGKERELVA   VIGNYGNTNYADSVKG RFTVSRDNAKNTVYLQMNSLNVEDTAVYVCKF  SSTVAPHEY WGQGTQVTVSS

(34) TABLE-US-00012 TABLE 9 R2a-F9 Hallmark VH-DP-53 residue Location diversity germline Comment Q1 FR1 — E Mutate Q1E S23 FR1 — A Mutate S23A S27 FR1 — F Mutate S27F I28 FR1 — T Mutate I28T S29 FR1 — F Mutate S29F Y37 FR2 Yes (FYHILV) V Retain E44 FR2 Yes (GEADQRSL) G Retain R45 FR2 Yes (LRCILPQV) L Retain E46 FR2 — V Mutate E46V L47 FR2 Yes (WLFAGIMRS) W Retain A49 FR2 — S Mutate A49S V69 FR3 — I Mutate V69I V78 FR3 — L Mutate V78L N93 FR3 — R Mutate N93R V94 FR3 — A Mutate V94A M99 FR3 — V Mutate M99V Q108 FR4 Yes (QLR) L Retain

Example 10: Introduction of a CDR3 VHH Sequence into a Human Antibody

(35) As the VHH CDR3 is the key determinant of binding specificity, this loop may be readily incorporated into an entirely human VH gene whilst preserving functionality. The resulting antibody is entirely human with a human CDR1 and CDR2 except for the VHH CDR3 loop which is derived from alpaca. Such an antibody has a lower risk of eliciting an adverse immune response when administered as a therapeutic monoclonal antibody due to the increased content of human sequence as compared to the antibody derived in Example 9.

(36) By way of example, the VHH CDR3 loop is grafted into an appropriate human V gene by designing an oligonucleotide corresponding to the VHH CDR3 sequence bracketed by human antibody germline sequence corresponding to FR3 and FR4 of either an individual human VH or a collection of human VH sequences. An example is shown below which includes a 3′ oligonucleotide corresponding to the VHH-CDR3 of R2a-F9 (SEQ ID 36) appended with a degenerate sequence designed to prime to all human VH3 gene family members. The VH3 gene was selected simply because camelid VHH antibodies show greatest homology to the human VH3 family. The corresponding 5′ primer is a degenerate oligonucleotide designed to anneal to the FR1 of all human VH3 gene members. These two oligonucleotides are used in a PCR reaction with human VH genes derived from cDNA prepared for example from peripheral blood lymphocytes of naive donors using established methods. The design shown below appends Sfi1 and BstEII restriction sites to facilitate cloning into the phage display vector pNIBS-1 (Example 3).

(37) TABLE-US-00013 GGCCCAGCCGGCC ATG GCA SAR GTG CAG CTG KTG GAG    Sfi1             FR1(Human VH3)                                          K   F  CTG TRB CGR MAY ATA RTG ACA TTC AAA             FR3(Human VH3)  S   S   T   V   A   P   H   E   Y   W   G   Q   G AGT TCA TGG CAA CGC GGC GTG CTC ATG ACC CCG GTC CCT     VHH CDR3  T   Q   V   T TGG GTC CAG TGG   BstEll

(38) After construction of a phage display library of human VH genes grafted with the VHH-CDR3 of R2a-F9, the library is then selected on recombinant H1 as in Example 4 and humanised antibodies screened as in Example 5.

Example 11: Assessment of Efficacy In Vivo

(39) The efficacy of cross-reactive antibodies for passive immunotherapy is assessed in a mouse challenge model. The seven H1/H5 cross neutralising antibodies (R2b-E8, R2b-D9, R1a-A5, R1a-C5, R1a-B6, R1a-H6, R2a-G8) are tested for their ability to protect BALB/c mice from challenge with H1N1 (A/California/07/09) and H5N1 (A/Vietnam/1203/2004) virus in a prophylactic setting. VHH antibodies are expressed, purified (Example 6) and administered intranasally as monovalent agents at 5 mg/kg, 24 hours before challenge with a lethal dose of virus. A non-relevant VHH antibody and a PBS only arm of the study are included as controls. Survival, body weight and viral titres in lung and other tissues are monitored to assess the antibodies' ability to protect against lethal challenge with H1N1 and H5N1 virus. Said antibodies demonstrate protection against such lethal challenge. Antibodies are also evaluated in a therapeutic setting where antibody is administered 24 hours post-infection with H1N1 and H5N1. Again survival, body weight and post challenge viral titres in lung and other tissues are monitored. Said antibodies demonstrate protection against such lethal challenge.

Example 12: Neutralisation Activity of Recombinant VHH Antibodies on H1N1 and H5N1

(40) The influenza microneutralisation assay was performed essentially as in Harmon et al., (1988). In brief, either serum from a immunised alpaca or purified VHH antibodies (at a starting concentration of 5 μg/ml) were serially diluted, two-fold, in a 50 μl volume of assay diluent (DMEM with the addition of; 2 mM glutamine, Pen/Strep 1/100, Amphotericin 1/100, non-essential amino acids and FCS 1/200) in a 96 well flat bottom plate (Costar) in duplicate columns 1-10. The reverse genetics reassortants NIBRG-14 (A/Vietnam/1194/04; H5N1) and the reassorted attenuated virus X-181 (A/California/07/09;H1N1) were used in viral neutralisation assays. X-181 virus and NIBRG-14 viruses were used at 10.sup.2TCID.sub.50 virus dose per 50 μl of assay diluents. In all assays a ‘virus only’ (VC) control and a ‘cells only’ (CC) control was included. Plates were incubated at room temperature for 1 hour and 100 μl of 1.5×10.sup.5/ml MDCK cells (Madin-Darby canine kidney cells) was added to all wells. Plates were incubated for 18 hours at 30° C. in 5% CO.sub.2. Cells were fixed with 100 μl of 0.6% w/v methanol/hydrogen peroxide per well and left for 20 minutes at room temperature. Plates were then washed 3 times in PBS Tween-20 (0.05% w/v). Growth of influenza virus and release of nucleoprotein was with a primary anti-influenza A nucleoprotein (Serotec) mouse monoclonal antibody. 100 μl of this antibody was added to each well at a 1:3000 dilution in 6 salt PBS/Tween-20 (0.1%) 5% w/v Marvel milk. 100 μl of this was added to all wells and plates and incubated for 1 hour at 37° C. Rabbit anti-mouse IgG polyclonal HRP conjugate (DAKO) at a dilution of 1:5000 in 6 salt PBS/Tween-20 (0.1% v/v), 5% (w/v) Marvel milk powder, 1% (v/v) bovine serum albumin was used and incubated for 1 hour at 37° C. Plates were washed 4 times and developed using TMB staining. The reactions were stopped by the addition of 100 μl of 0.5M HCl to each well and read at 450 nm & 620 nm. The 620 nm values were subtracted from the 450 nm values to correct for plate absorption. The antibody concentration at which 50% viral neutralisation was seen was determined using the equation;

(41) $\frac{\left(\mathrm{mean}OD\mathrm{of}VC\right)-\left(\mathrm{mean}OD\mathrm{of}CC\right)}{2}+\left(\mathrm{mean}OD\mathrm{of}CC\right)=x$

(42) Where x is the OD at which 50% of MDCK cells were infected.

(43) All antibody clones except R2b-D8, R1a-F4, R1a-E5, R1a-F5, and R1a-E6 were able to neutralise X-181 which is a laboratory adapted strain of H1N1 (Table 10). Of the H1N1 neutralising antibodies, R2b-E8, R2b-D9, R1a-A5, R1a-C5, R1a-B6, R1a-H6 and R2a-G8 were also able to neutralise NIBRG-14 (H5N1) and were deemed to be cross neutralising antibodies (Table 10). This was consistent with the cross reactivity data shown in examples 5, 6 and 7. For antibodies R2b-E8, R2b-D9, R1a-A5, R1a-B6 and R1a-H6 this was consistent with absence of haemagglutination inhibition activity and specificity for acid sensitive epitopes in the conserved stalk region of haemagglutinin (Table 6).

(44) TABLE-US-00014 TABLE 10 Antibody neutralisation of H1N1 and H5N1 viral infection of MDCK cells Neutralisation X-181 Neutralisation NIBRG-14 A/California/07/09 A/Vietnam/1194/04 Clone (H1N1 pdm) EC50 nM (H5N1) EC50 nM R1a-G5 30 nM — R2b-D8 — — R2b-E8 54 nM 29 nM R2b-D9 55 nM 7.8 nM R1a-F4 — — R1a-A5 23 nM 8.0 nM R1a-C5 23 nM 55 nM R1a-E5 — — R1a-F5 — — R1a-B6 31 nM 20 nM R1a-E6 — — R1a-G6 12 nM — R1a-H6 53 nM 12 nM R2a-E8 26 nM — R2a-G8 538 nM  >1 μM R2a-B9 29 nM — R2a-F9 68 nM — R2a-G9 62 nM — R2a-H9 16 nM — control — —

Example 13: Evaluation as a ‘Universal Antibody’ for Vaccine Potency

(45) Cross-reactive antibodies of the invention are used in potency assays for attenuated influenza vaccines. The cross-reactive antibodies of this application cross-react with (1) many/all haemagglutinins of one subtype/type of influenza virus, or with (2) haemagglutinins of more than one subtype of influenza A virus with all of these subtypes belonging to one of the two recognised phylogenetic groups of influenza HA, or with (3) HAs from both phylogenetic groups, group 1 and group 2.

(46) An antibody having cross-reactivity as in (1) is useful for the potency testing of multivalent, preferably trivalent and quadrivalent, influenza vaccines. An antibody having cross-reactivity as in (2) is useful for the potency testing of multivalent, preferably trivalent and quadrivalent, influenza vaccines; for the potency testing of pandemic or pre-pandemic influenza vaccines containing a single subtype of influenza A virus; and for the potency testing of vaccines containing more than one subtype/type, with each virus component belonging to a different phylogenetic group. An antibody having cross-reactivity as in (3) is useful for the potency testing of monovalent influenza vaccines, preferable pandemic and pre-pandemic influenza vaccines.

(47) The use of such cross-reactive antibodies in potency testing eliminates the need to frequently change the antibody reagent as is the case in currently used methods, for which specific sheep antisera have to be generated every time a strain change is announced and implemented; production of a new antiserum for every antigenically drifted virus that is used in influenza vaccines has been identified as a potential bottleneck in the pathway from virus strain change to vaccination with an updated influenza vaccine. To solve this problem, cross-reactive antibodies of this invention are produced in advance and are therefore available at all times; following a strain change in influenza vaccine production, (a) cross-reactive antibody/antibodies can be used immediately, with no lead time to availability of strain-specific antibodies.

(48) By way of example, a cross-reactive antibody is used in the single radial immunodiffusion (SRID or SRD) assay. Following standard protocols for SRD, the cross-reactive antibody is incorporated into a gel matrix and then used in the assay. A precipitin ring is formed when influenza haemagglutinin is allowed to diffuse into the gel and can be measured following standard protocols for visualisation of the SRD ring/zone.

(49) In another example, cross-reactive antibodies are used in ELISA assays to quantitatively measure the potency of vaccines and pre-vaccine samples, such as monovalent bulks. ELISA assays of various formats may be used. For instance, a sandwich ELISA using an antibody coated on a solid surface to capture haemagglutinin and an antibody to detect captured HA may be used. The capture and the detection antibody may be the same antibody, or may be different antibodies. Either of the two antibodies, or both, may be cross-reactive antibodies of this invention. In another format, a competition ELISA is used. HA is coated on a solid surface and reacted with cross-reactive antibody/antibodies of this invention. A reference antigen, or the sample to be measured, is used as a competitive reagent. Reduction of binding of the cross-reactive antibody to the coated HA can be used to quantify HA in solution, using a standard curve generated from the reference antigen reagent. Other formats of ELISAs may be employed to a similar effect; in all cases, a cross-reactive antibody, or a multitude of cross-reactive antibodies, are used as specific reagents to capture and/or detect HA.

(50) In another example, Surface Plasmon Resonance (SPR) technology is used to establish a potency assay for attenuated influenza vaccines. Cross-reactive antibodies are used to detect and quantify HA protein/antigen. Various formats of SPR-based assays may be used. In a preferred embodiment, a competitive assay, with reference antigen and cross-reactive antibody competing for binding to HA coated on a chip, as appropriate for the specific device used, is employed.

(51) In another example, cross-reactive antibody/antibodies are used in a method using proteins (in particular, HA) immobilised on a membrane. The membrane may be a nitrocellulose membrane, a PVDF membrane, or another suitable membrane. Immobilisation of protein and HA on the membrane may occur through various processes, such as Western Blotting using various devices, slot blotting, spotting onto membrane, and others.

(52) In another example, cross-reactive antibodies are used to enrich HA from samples, such as vaccines or monovalent bulks. Enrichment may be through use of cross-reactive antibodies in immune-precipitation, affinity chromatography, magnetic separation (with antibody being coupled by some means to magnetic substances, such as paramagnetic beads), or any other immunological method capable of binding and enriching HA out of a sample which will be, for the most part, a liquid containing HA in solution. HA adsorbed to substances, such as adjuvants, may also be subjected to such enrichment, either in untreated samples or following treatment to de-adsorb the HA. Enriched HA may then be quantified using any suitable method. Suitable methods include, but are not limited to, mass-spectrometry, in particular isotope dilution mass spectrometry as used in ‘immunocapture isotope dilution mass spectrometry’ (Pierce et al., Analytical Chemistry, 2011 dx.doi.org/10.1021/ac2006526) and HPLC.

Example 14: Design of an Influenza Vaccine which Elicits a Cross-Protective Immune Response Against Future Group 1 Influenza Strains

(53) Cross-reactive antibodies are used to define epitopes on haemagglutinin, which are structurally conserved across different viral subtypes. Such epitopes are then used to design a vaccine, which when administered to patients elicits a protective immune response against future viral subtypes so eliminating the need for repeated seasonal vaccination against a constantly evolving influenza virus.

(54) Mapping the epitope of a cross-reactive antibody of this invention is performed by co-crystallisation of the cross-reactive antibody in complex with haemagglutinin and subsequent X-ray crystallography. Alternatively, the cross-reactive antibody is constructed as a gene fragment phage display library derived from the HA gene of an influenza virus such as a pandemic H1N1 (e.g. A/California/07/2009). The H1-HA gene is synthesised, digested with DNAase into fragments and cloned into the phage display vector pNIBS-1. This gene fragment phage-displayed library is then selected with purified cross-reactive single domain antibodies described in this invention. Sequence analysis of the selected gene fragments is then used to delineate the cross-neutralising epitope(s) and map them against available HA sequences in the public databases to determine the extent of conservation across all 16 HA subtypes. For more ‘fine-tuned’ epitope mapping, alanine scanning mutagenesis of the HA gene fragment displayed on phage is used followed by screening for loss of specific antibody binding whilst preserving display on phage. The minimal haemagglutinin gene fragment which can generate a correctly folded product corresponding to the cross-reactive antibody epitope is then expressed and purified in a heterologous gene expression system or synthesised as a peptide using established chemical procedures. This recombinant HA epitope is then used to immunise BALB/c mice and to protect from lethal challenge with different viral strains including H5N1 and H1N1 viral sub-types.

(55) Alternatively, the cross-reactive antibodies of the invention can be used to isolate anti-idiotypic antibodies which represent a structural mimic of the relevant cross-reactive epitope on haemagglutinin. Such anti-idiotypic antibodies when used as a vaccine elicit an immune response which cross-reacts with relevant epitopes on viral haemagglutinin. An advantage of using the single domain VHH antibodies described in this invention is that they are structurally much simpler than a conventional antibody, which requires the stable association of both a VH and VL domain to form the paratope. By way of example, a cross-reactive single domain antibody (e.g. R1a-B6) is used to immunise an alpaca as in Example 1 and a phage display antibody library is constructed as in Example 2. The library is then selected on the initial cross-reactive antibody used for immunisations (Example 4). Anti-idiotypic antibodies are described as those which bind to the cross-reactive antibody (R1a-B6) used for immunisation and are negative for other antibodies carrying different CDR1, CDR2 and CDR3 sequences but having similar framework regions.