WO2010040572A2 - Antibodies - Google Patents

Abstract

Provided are antibodies, or fragments thereof, useful in recognising multiple clades of H5N1 avian influenza virus and methods of immunotherapy and vaccination comprising said antibodies or fragments.

Description

ANTIBODIES

The present invention relates to antibodies capable of recognising multiple clades of H5N1 and methods and uses therefor.

INTRODUCTION

Influenza A (H5N1) highly pathogenic (HP) virus causes severe disease in humans and poses an unprecedented pandemic threat (13, 23). The antiviral agents such as amantadine and rimantadine or neuraminidase inhibitor oseltamivir constitute an important treatment option. However, pharmacological treatment is often associated with the selection of drug resistant virus strain (13, 16) and H5N1 oseltamivir resistant variants have been isolated during oseltamivir treatment (4, 12). The unprecedented spread of highly pathogenic avian influenza virus subtype H5N1 in Asia and Europe is threatening animals and public health systems. Effective diagnostic and therapeutic strategies are needed to control and combat the disease.

WO 2007/089753 (Xia et al) discloses mouse monoclonal antibodies (obtained by hybridoma technology) that specifically bind to the HA protein of avian influenza virus subtype H5, as well as monoclonal antibodies capable of blocking at least 50% of the haemagglutinin binding activity of these antibodies. The authors also tested the original mouse antibodies, the corresponding mouse scFvs (single chain fragment variable) and the chimaeric antibodies derived from them in ELISA assays, haemagglutinin inhibition assays and in micro-well neutralization tests.

WO 2007/031550 (Throsby et al) relates to the preparation of immunoglobulin libraries from specific cell population and, in particular, the identification and generation of immunoglobulins derived from libraries created using RNA from subset of antibody producing B cells, IgM memory B cells.

However, there is a need in the art for antibodies against H5N1 that have a broad strain or clade specificity, particularly against the H5 subtype (including, for instance, H5N1, H5N2 and H5N3). SUMMARY OF THE INVENTION

Surprisingly, we have now discovered novel antibodies and fragments thereof that are capable of recognising multiple clades or strains of H5N1. These are particularly useful in widespread vaccination programmes because of their multi-strain specificity, something that was previously lacking in the art. They are also useful in methods of detection or diagnosis of subtypes, particularly H5N1.

Accordingly, in a fist aspect, the present invention provides an antibody or antibody fragment capable of recognising the same epitope or epitopes as an antibody which comprises at least one Heavy Chain Complementarity Determining Region (CDR) CDR selected from the group consisting of: SEQ ID NOS. 12, 28, 44, 60 and 76.

In a further aspect, the present invention provides an antibody or antibody fragment capable of recognising the same epitope or epitopes as an antibody which comprises at least one Light Chain CDR selected from the group consisting of: SEQ ID NOS. 15, 31, 47, 63 and 79.

In a still further aspect, the present invention provides an antibody or antibody fragment capable of recognising the same epitope or epitopes as an antibody which comprises at least one Heavy Chain CDR selected from the group consisting of: SEQ ID NOS. 12, 28, 44, 60 and 76; and at least one Light Chain CDR selected from the group consisting of: SEQ ID NOS. 15, 31, 47, 63 and 79.

It is particularly preferred that the antibody is capable of recognising H5N1, and comprises at least one of the present Heavy Chain CDRs and/or at least one of the present Light Chain CDRs. Therefore, the invention also provides an antibody or antibody fragment, for instance any of those defined herein, which is capable of recognising H5N1, and comprises at least one Heavy Chain CDR selected from the group consisting of: SEQ ID NOS. 12, 28, 44, 60 and 76; and/or at least one Light Chain CDR selected from the group consisting of: SEQ ID NOS. 15, 31, 47, 63 and 79. The antibody preferably comprises two or more CDRs. These may all be Heavy Chain CDRs, all Light Chain CDRs or may be any mixture of the two. It is preferred, however, that the Antibody comprises at least one of each (i.e. at least Heavy Chain CDR and at least one Light Chain CDR). Preferred combinations could include two Heavy Chain CDRs and one Light Chain CDR; or two Light Chain CDRs and one Heavy Chain CDR, for instance. The total number of CDRs can be as few as one, but 2 or 3 are more preferred; 4, 5 or 6 are particularly preferred; whilst 7, 8 or 9 or more are also preferable. However, as the present CDRs are highly specific, i.e. they recognise or bind to the H5N1 antigen with high affinity, large numbers of CDRs are not though to be necessary.

The CDRs are preferably positioned within framework regions, such as are well known in the art. The framework regions are preferably humanised to minimise self-recognition by the human host's immune system. The positioning of any of the one or more CDRs is not important, provided that there is no stearic hindrance preventing recognition of the epitope.

SEQ ID NOS. 12, 28, 44, 60 and 76 correspond to the Heavy Chain CDR3s of the antibodies identified herein as AVC4, AVDl, AVE2, AV A6 and AVG4, respectively. SEQ ID NOS. 15, 31, 47, 63 and 79 correspond to the Light Chain CDR3s regions of the antibodies identified herein as AVC4, AVDl, AVE2, AV A6 and AVG4, respectively.

CDRs, alone or in any combination, from AVC4 and/or AVDl (SEQ ID NOS; 12, 28, 15 and 31) are particularly preferred. The CDR3 sequences (SEQ ID NO; 44 and 47) from AV.E2 are also preferred.

Preferably, the antibody or fragment comprises a Heavy Chain. Alternatively, the antibody or fragment comprises a Light Chain. It is particularly preferred that the antibody or fragment comprises both a Heavy and Light Chain. In an aspect, the antibody or antibody fragment is capable of inhibiting H5N1 infection, preferably across more than one clade or serotype. Therefore, the antibody or antibody fragment is preferably specific for H5N1 and inhibits (reduces or ameliorates) infection by more than one H5N1 serotype or clade.

The Heavy and/or Light Chains preferably comprise one or more additional Complementarity Determining Regions (CDR). These CDRs may be selected from other CDRs known in the art, which would allow a degree of cross-reactivity to be "designed" into the antibody or fragment.

However, it is preferred that these additional CDRs are those designated CDRIs selected from those from AVC4, AVDl, AVE2, AV A6 and AVG4, which CDRl protein sequences are the same. These additional CDRs may also be those designated CDR2s selected from those from AVC4, AVDl, AVE2, AV A6 and AVG4, which CDR2 protein sequences are the same, in the context of a Heavy or a Light Chain. The additional CDRs may be provided in the context of a Heavy or a Light Chain.

Thus, it will be appreciated that the CDR3s and, optionally the additional CDRs, may be from the Heavy or Light Chains found herein and so any combination of the present CDR3s with any of the CDRIs and/or CDR2s disclosed herein is envisaged.

The Heavy and/or Light Chains may preferably comprise the entire variable regions of AVC4, AVDl, AVE2, AV A6 or AVG4, as appropriate. Preferably, therefore, the antibody or fragment comprises a single chain variable region, preferably with at least three CDRs.

The antibody or fragment may preferably comprise first and second chains, each chain comprising a variable region, the chains being linked by a suitable linker. Preferably, the first chain is a Light Chain and the second chain is a Heavy Chain, although both chains may be Heavy Chains or both c may be Light Chains. Therefore, although two (or more) Heavy Chains or two (or more) Light Chains are envisaged, one of each is preferred. The linker is preferably in order of 10-50 nucleotides and preferably predominantly Ser and Glycine, although the linker can be designed by the skilled person to incorporate suitable features such as glycosylation sites, cleavage sites or disulphide bridges, and so forth. Ideally, the linker should be sufficiently long and lacking in secondary or tertiary structure so as not to constrain the ability of either chain to bind to its epitope or epitopes.

Examples of preferred linkers are provided below. This antibody or fragment comprises, therefore, a single chain variable fragment of two chains, preferably one Light and one Heavy Chain, joined by a linker, known as an scFv antibody. This may be encoded by a single cistron, allowing for increased ease of delivery, especially in vivo via a suitable vector, as discussed further below.

Multimers of the any of the above chain combinations are also envisaged, particularly the Heavy and Light Chain combinations. Preferably, these would comprise multiple scFv units linked via a multifunctional linker. However, is also envisaged that any antibody or fragment as described herein could be linked to one or more others via suitable bi- or multi functional linkers.

For example, genetic engineering of scFv-scFv tandems linked with a third polypeptide linker has been carried out in several laboratories (Gruber, M., Schodin, B. A., Wilson, E. R. & Kranz, D. M. "Efficient tumor cell lysis mediated by a bispecific single chain antibody expressed in Escherichia coli". J. Immunol. (1994) 152, 5368-5374; Kurucz, I., Titus, J. A., Jost, C. R., Jacobus, C. M. &Segal, D. M.. "Retargeting of CTL by an efficiently refolded bispecific single-chain Fv dimer produced in bacteria". J. Immunol. (1995) 154, 4576-4582; Sergey M. Kipriyanov, Gerhard Moldenhauer, Jochen Schuhmacher, Bjorn Cochlovius, Claus-Wilhelm Von der Lieth, E. Ronald Matys and Melvyn Little. "Bispecific Tandem Diabody for Tumor Therapy with Improved Antigen Binding and Pharmacokinetics". J. MoI. Biol. (1999) 293, 41-56). Single chain Fv (scFv) fragments have been also genetically fused with adhesive polypeptides (de Kruif, J. & Logtenberg, T. "Leucine zipper dimerized bivalent and bispecific scFv antibodies from a semi-synthetic antibody phage display library". J. Biol. Chem. (1996) 271, 7630-7634.) or protein domains to facilitate the formation of heterodimers (Muller, K. M., Arndt, K. M., Strittmatter, W. & Pluckthun, A.. The first constant domain (CHl and CL) of an antibody used as heterodimerization domain for bispecific miniantibodies. FEBS Letters (1998), 422, 259-264).

Single chain Fv fragments can be also fused to an immunoglobulin CH3 domain, resulting in a self-assembling bivalent or bispecific protein, generally called a "minibody" (Hu SZ, Shively I et al., "Minibody: a novel engineered anti- carcinoembryonic antigen antibody fragment (single chain Fv-CH3) which exhibits rapid high level targeting of xenografts". Cancer Res (1996), 56:3055-61.) Such minibodies are a preferred aspect of the invention.

The antibody or fragment of the invention may be a monoclonal antibody (mAb) or a polyclonal antibody.

Preferably, the antibody or fragment of the invention is the full length AVC4, AVDl, AVE2, AV A6 or AVG4 sequence disclosed herein, although it will be appreciated that the linkers provided herein, although useful, are only examples and can be replaced.

Without being bound by theory, it is believed that the present CDRs recognise the H5 haemagglutinin protein and, in particular, the haemagglutinin HAl subunit thereof.

The AVC4, AVDl, AVE2, AV A6 and AV G4 sequences referred to are, in the most part, amino acid sequences. We have also provided nucleotide sequences encoding these amino acid sequences. Although these nucleotide sequences are preferred, it will be appreciated that the redundancy of the genetic code allows for variations in the nucleotide sequence that still encodes the same amino acid sequence. Indeed, although it is generally not preferred to alter the CDR sequences provided herein, some degree of variation is envisaged, particularly conservative amino acid substitutions. Outside of the CDRs, i.e. in the variable region frameworks provided, variants comprising insertions, mutations or substitutions are envisaged.

Preferred degrees of sequence homology for amino acids sequences, most preferably outside of the CDRs, are sequences having at least 60%, more preferably at least 70%, more preferably at least 75%, more preferably at least 80%, more preferably at least 85%, more preferably at least 90%, more preferably at least 95% and more preferably at least 99% sequence homology with those provided for AVC4, AVDl, AVE2, AV A6 and AV G4 herein. At least 80 or 85% are particularly preferred.

Similarly, preferred degrees of sequence homology for nucleotide sequences, most preferably outside of the CDRs, are sequences having at least 50%, more preferably 60%, more preferably at least 70%, more preferably at least 75%, more preferably at least 80%, more preferably at least 85%, more preferably at least 90%, more preferably at least 95% and more preferably at least 99% sequence homology with those provided for AVC4, AVDl, AVE2, AV A6 and AVG4 herein. At least 70 or 80% are particularly preferred.

Where reference is made to a polynucleotide sequence, then complementary or partially complementary sequences are also envisaged. These are preferably capable of hybridising to the reference sequence under highly stringent conditions. Generally, in order to maximize the hybridization rate, relatively low-stringency hybridization conditions are selected: about 20 to 25° C. lower than the thermal melting point (T _m ). The T _m is the temperature at which 50% of specific target sequence hybridizes to a perfectly complementary probe in solution at a defined ionic strength and pH. Generally, in order to require at least about 85% nucleotide complementarity of hybridized sequences, highly stringent washing conditions are selected to be about 5 to 15° C. lower than the T _m . In order to require at least about 70% nucleotide complementarity of hybridized sequences, moderately-stringent washing conditions are selected to be about 15 to 30° C. lower than the T _m . Highly permissive (very low stringency) washing conditions may be as low as 50° C. below the T _m , allowing a high level of mis-matching between hybridized sequences. Those skilled in the art will recognize that other physical and chemical parameters in the hybridization and wash stages can also be altered to affect the outcome of a detectable hybridization signal from a specific level of homology between target and probe sequences. Preferred highly stringent conditions comprise incubation in 50% formamide, 5^χSSC, and 1% SDS at 42° C, or incubation in 5><SSC and 1% SDS at 65° C, with wash in 0.2^χSSC and 0.1% SDS at 65° C.

Suitable sequence alignment or comparison methods are known in the art and may include programs such as BLAST.

In the case of both amino acid and nucleotide sequence variants, the ability to recognise the H5 protein of the H5N1 subtype, and preferably the other subTypes such as H5N2 and H5N3, is required.

Influenza A viruses are negative sense, single-stranded, segmented RNA viruses. There are several subtypes, labelled according to an H number (for the type of haemagglutinin) and an N number (for the type of neuraminidase). There are 16 different H antigens (Hl to H 16) and nine different N antigens (Nl to N9). Therefore, where reference is made herein to H5N1, it will be appreciated that this extends to the other subtypes such as at least H5N2 and H5N3, but also to any of H5N4, H5N5, H5N6, H5N7, H5N8 and/or, H5N9. Thus, whilst any H5 subytpe, especially H5N4-9 are preferred, H5N2 and H5N3 are more preferable and H5N1 is particularly preferred.

The antibody or fragment may be humanised, although it is noted that if the full sequences (or variants thereof) provided for AVC4, AVDl, AVE2, AVA6 and AVG4 are used, then these are already human sequences, so no humanisation will necessarily be required, although further optimisation is envisaged. Chimaeric or fusion antibodies, such as the present antibody or fragment fused to a reporter protein (for instance a fluorescent protein such as GFP or a reporter system such as elements from LacZ, for example beta-galactosidase) are also provided. The antibody or fragment may be provided with a constant region. Suitable constant regions are those provided in human IgGA, IgGB, IgGC, IgGD, IgGE, IgGF and, most preferably, IgG. Indeed, the whole CDRs of the invention may be provided in the context of any of said IgGs, by using the variable region frameworks of said immunoglobulins and, optionally, their constant regions as well, although the constant regions may be used interchanged as discussed. The IgG may be of various subtypes, including IgG2, IgG3 or IgG4, although IgGl is preferred.

We have developed human monoclonal antibodies (mAbs) in single chain fragment format towards the H5N1 avian influenza virus, useful in gaining new insights in the role of specific target antigens in infection and in methods of immunotherapy against human cases of H5N1.

Using a biopanning approach based approach, a large array of scFvs against H5N1 were isolated from the human synthetic ETH-2 phage antibody library. H5N1 ELISA -positive scFvs possessing unique VH and VL sequences were analyzed for H5N1 recognition and H5N1 neutralization. This investigation demonstrated that scFv antibodies possess different biochemical properties and neutralization activity across H5N1 viral strains. In particular, the scFv clones AV.D1 and AV. C4 exert a significant inhibition of the H5N1 VietNam/ 1194/04 virus infection in a pseudotype-based neutralization assay. Interestingly, these two scFvs displayed also a cross-clade neutralizing activity versus whooping swan/Mongolia/244/05 and Indonesia5/05 strains. These studies show that human mAbs in scFv format with well defined H5N1 recognition patterns and neutralizing activity can be isolated by biopanning selection of the synthetic ETH-2 phage antibody library using inactivated H5N1 virus as a bait. These scFvs or genetically derived recombinant human IgG are useful reagents for prophylaxis or adjunctive treatment of human cases of H5N1 influenza. It will be appreciated that where reference to H5N1 is made herein, this applies to all clades and strains, especially H5N1 VietNam/1194/04, whooping swan/Mongolia/244/05 and Indonesia5/05 strains.

Several factors could be involved in the different biological activity exerted by the antibodies including the affinity and function of the bound epitope domain. Without being bound by theory, it is likely that the antibodies and fragments of the invention, especially scFvs AV. C4 and AV. Dl (both of which are involved in neutralization activity), recognise distinct epitopes with different molecular structures (linear and conformational respectively) as suggested by Western blot analysis.

The antibodies and fragments of the invention are also useful in methods of detection or diagnosis H5N1. Thus, the invention also provides a method of diagnosing whether a patient has an H5N1 infection comprising using the antibodies and fragments of the invention to assay a sample of body fluids from an individual, said the antibodies and fragments being capable of detecting at least one epitope from or derived from H5N1. Suitable methods include an ELISA, Western Blotting or other suitable methods or kits comprising the antibodies and fragments of the invention.

Thus, the invention also provides such a kit for the detection of H5N1 in a sample, the kit comprising the antibodies and fragments of the invention and a suitable means of detection, such as luminescence means.

Also provided is a chip comprising such a kit or merely the antibodies and fragments of the invention, for instance in the form of a multi-well plate.

AV.E2, G4 and A6 are particularly preferred for methods and kits for diagnosis or detection of H5N1.

BRIEF DESCRIPTION OF THE FIGURES

The present invention will now be further described with reference to: Figure 1. Screening of H5N1 -specific scFv clones. IPTG -inducted bacterial supernatants of individual colonies from the third round of the ETH-2 library selection were assayed for H5N1 specificity. H5N1 /NIBRG- 14 virus or irrelevant antigen (glucose oxidase) coated 96-well microtiter plates were tested by ELISA. OD values of the scFv clones against inactivated, purified H5N1 /NIBRG- 14 virus (panel A) or irrelevant antigen (panel B) are shown.

Figure 2. Gene encoding for scFv antibodies. In panel A, the amino acid sequences of VH and VL chains in a single letter code and their residue positions in the CDR3 regions are reported. In panel B, the schematic representation of the scFv antibody displayed on Ml 3 phage as pill fusion proteins is depicted. Single amino acid codes are used according to standard IUPAC nomenclature. ^(* ^Numbering is according to Tomlinson et al. (1995);

Figure 3. Western blot of scFv antibodies. The anti H5N1 scFvs antibodies were analyzed by Western blot for H5N1 protein recognition in non reducing (A) and reducing (B) conditions. Note that the scFv clones AV. C4 and AV. G4 bind with proteins with molecular mass of about 70 kDa (A) and 50 KDa (B) corresponding to the molecular weight of the H5 haemagglutinin protein and its major subunit HAl, respectively. Monospecific rabbit immunoserum raised against H5 polypeptide (lane 7) and scFv to glucose oxidase (lane 6) were used as positive and negative controls, respectively; Figure 4. Neutralizing activity of the scFvs against H5N1 pseudotypes. Pseudotypes expressing HA from A/Vietnam/ 1194/2004 (A), A/Mongolia/244/2005 (B), A/Indonesia/5/2005 (C) and A/Anhui 1/2005 (D), were incubated with serial dilutions of the indicated scFv antibodies. The percentage of inhibition of infection was defined as the reduction in RLU compared to virus control wells after subtraction of background RLU. Sheep immune-serum against H5N1 by NISBC (poly- Ab) and irrelevant scFv GO against glucose oxidase were used as a positive and negative control, respectively. The profile of scFv AV.A6 is representative of the non-neutralizing clones (AV.A6, AV.E2 and AV.G4);

Figure 5 shows the nucleotide sequence of antibody AV. C4;

Figure 6 shows the protein sequence of antibody AV. C4;

Figure 7 shows the nucleotide sequence of antibody AV.D1;

Figure 8 shows the protein sequence of antibody AV.D1;

Figure 9 shows the nucleotide sequence of antibody AV. E2;

Figure 10 shows the protein sequence of antibody AV.E2;

Figure 11 shows the nucleotide sequence of antibody AV. A6;

Figure 12 shows the protein sequence of antibody AV. A6;

Figure 13 shows the nucleotide sequence of antibody AV. G4; and Figure 14 shows the protein sequence of antibody AV. G4.

In each of Figures 5-14, the sequences comprise, respectively, the heavy chain (VH) and the light chain (VL) of the scFv antibody joined by a "linker." The linker is shown in italics. The CDRs (Complementarity-Determining Regions) of the VH chain (HCDRs) and the CDRs of the VL chain (LCDRs) are underlined. Both the nucleotide and amino acid sequences are shown in the single-letter code.

DETAILED DESCRIPTION

Suitable framework regions are shown in the Figures either side of the CDRs, but do not include the linker, which is separate. Such preferred framework regions are, therefore, easily derivable from the sequences shown in the Figures and Sequence Listing.

Particularly preferred embodiments are scFv antibodies comprising one Heavy Chain and one Light Chain, each chain comprising at least one and preferably 2 or 3 CDRs as defined herein, linked by a suitable linker. Especially preferred examples comprise one or more CDRs from each of the Heavy and Light chains shown in the antibodies described herein as AVC4, AVDl, AVE2, AV A6 and AVG4.

It is noteworthy, in respect of WO 2007/089753 that the present scFv antibodies anti- H5N1 virus are fully human antibody fragment and that they show neutralization activities without subsequent engineering such as the conversion in whole immunoglobulin (for example human-mouse chimaeric antibodies). The present antibodies therefore offer the following advantages: a) they are fully human, hence poorly or not in the least immunogenic; b) these human fragments can be produced on a large scale in a prokaryotic system, so reducing costs and time spent on safe and effective drug substance production; c) these scFvs meet important criteria for a potential pharmaceutical anti H5N1 compound as their neutralizing activities, due to their lack of an Fc region, avoid any possible adverse reaction related to constant immunoglobulin domain (for instance a cytokine storm). Whilst the authors in WO 2007/031550 (Throsby et al) also reported the isolation of monoclonal antibodies against H5 protein from these libraries, no evidence of any neutralizing activities were shown in in vitro neutralization studies of these antibodies in scFv format. Thus, with respect to WO 2007/031550, we have shown the neutralization activity of anti-H5Nl virus by our scFv antibodies in in vitro experiments.

A study by Simmons et al (PIoS Medicine, VoI 4, 2007, pp 178- 186) describes whole human neutralizing antibodies against influenza H5N1 generated via EBV immortalization of human lymphocytes. However, EBV immortalized cells are characterized by phenotypic instability, poor secretion, and pose important safety concerns for the presence of a potential infective agent and/or its degradation product in the drug substance (Center for Biologies Evaluation and Research, Food and Drug Administration, 1997; Committee for Medicinal Products for Human Use, European Medicines Agency, 2007). Again, the neutralization activities of our scFv antibodies lacking an Fc region avoid any possible adverse reaction related to constant immunoglobulin domain (such as a cytokine storm, as mentioned above).

A paper by de Kruif et al (J. MoI. Biol., VoI 21, 1995, pp 97-105) and a protocol by ETH Zurich (ETH-2 human antibody phage library protocol dated 13 December 2005, available for their website) describe the construction of naive or semi-synthetic antibody repertoires. These, together with a paper by Lim et al (Vir. J., VoI 5, 2008, pp 130-139), are believed to describe the technological background.

It will also be noted that the CDR3 for each Light Chain in the preferred embodiments shown in the Figures and described herein, comprise an NSS-XXX-XXX motif, and in all cases but one an NSS-XXP-XXX, where X is any amino acid. All the Light Chain CDR3s are proline rich, in that they contain 2 or more Proline residues, which is surprising in a CDR of only 9 amino acids. These motifs are particularly preferred.

Monoclonal antibodies could to be of assistance in early detection and rapid typing of epidemic viruses, providing information about the evolution and geographic spread of newly emerging strains [M Tkacova, E Vareckova, I C Baker, J M Love, and T Ziegler: Evaluation of monoclonal antibodies for subtyping of currently circulating human type A influenza viruses. J Clin Microbiol. 1997, 35(5): 1 196-1198; Eva Vareckova, Nancy Cox, and Alexander Klimov : Evaluation of the Subtype Specificity of Monoclonal Antibodies Raised against Hl and H3 Subtypes of Human Influenza A Virus Hemagglutinins. J Clin Microbiol. 2002, 40(6): 2220-2223; T Ziegler, H Hall, A Sanchez-Fauquier, W C Gamble, and N J Cox: Type- and subtype-specific detection of influenza viruses in clinical specimens by rapid culture assay. J Clin Microbiol. 1995, 33(2): 318-321].

The present scFv Antibodies represent useful tools for studies of viral replication and antiviral activity, as well as for the development of diagnostic tools for monitoring H5N1 infection in in vitro and in vivo. Moreover, these scFv may show different properties in vivo, if armed with an antibody constant region (Fc region). For example, Cameron P Simmons et al. (Simmons C. P., N. L. Bernasconi, A. L. Suguitan Jr., K. Mills, J. M. Ward, N. V. V. Chau, T. T. Hien, F. Sallusto, D. Q. Ha, J. Farrar, M. D. de Jong, A. Lanzavecchia and K. Subbarao. 2007. Prophylactic and Therapeutic Efficacy of Human Monoclonal Antibodies against H5N1 Influenza. PLoS Med 4: el 78) report one mAb that was effective in neutralizing a Clade II virus in vivo but not in vitro, suggesting the neutralizing activity of this Mab is dependent upon a factor found in vivo, such as complement. In addition, since the high speed of H5N1 mutations, it cannot in principle be ruled out that such antibodies, in the presence of mutated virus, may become neutralizing.

The scFv antibodies that are preferred consist of VL and VH domains, connected by a linker (small peptide). These immunoglobulin fragments are functional per se in inhibiting avian virus infection. However, the conversion of the scFv into a full IgG immunoglobulin (with an Fc region) may enhance the functional properties of such immunoglobulin fragments. This could be obtained by cloning the only CDRs regions in an antibody scaffold or linking each V domain (VH and VL from scFv) with the respective C domain (CH and CL). However, it will be appreciated that there are meaningful differences in the four subclasses of IgG. Human antibodies of the IgGl and IgG3 isotypes are preferred as these can potentially support the effector functions of antibody dependent cellular cytotoxicity (ADCC) and complement-dependent cytotoxicity (CDC) (ABBAS A.K., LICHTMAN A.H. BASIC IMMUNOLOGY. Functions and Disorders of the Immune System.(2008) Editor: ELSEVIER SAUNDERS.).

ADCC is triggered by an interaction between the Fc region of an antibody that has bound, through its antigen-binding region, to a target antigen and the Fcγ receptors (FcγRs), particularly FcγRI and FcγRIII, on immune effector cells such as neutrophils, macrophages and natural killer cells. The target is eliminated by phagocytosis or lysis, depending on the type of mediating effector cell. A prerequisite first step for CDC is recruitment of the complement component CIq by IgG bound to its target antigen. This triggers a proteolytic cascade to activate complement. This can lead to the formation of a membrane attack complex that kills the target cell by disrupting its cell membrane. Alternatively, tumour-cell-bound CIq can bind to complement receptors, such as CIqR, CRl (CD35) and CR3 (CDl lb/CD18), on effector cells, such as neutrophils, macrophages and natural killer cells. This can trigger cell-mediated tumour-cell lysis or phagocytosis, depending on the type of effector cell (ABBAS A.K., LICHTMAN A.H. BASIC IMMUNOLOGY. Functions and Disorders of the Immune System.(2008) Editor: ELSEVIER SAUNDERS; David Filpula. "Antibody engineering and modification technologies". Biomolecular Engineering. (2007) 24, (2):201-215; Presta LG. "Engineering of therapeutic antibodies to minimize immunogenicity and optimize function". Adv Drug Deliv Rev. 2006 Aug 7;58(5-6):640-56. Epub 2006 May 23. Review).

Thus, it is preferred that the Ab has a portion, preferably an Fc region, adapted to bind Fc Receptors and/or adapted to bind complement, preferably to initiate the formation of the membrane attack complex. IgGl is a preferred format/context for implementation of the ADCC and CDC effector functions, whereas IgG3 are potent CDC activators and IgG2 and IgG4 are poor CDC activators (Aalberse and Schuurman, 2002 R.C. Aalberse and J. Schuurman, IgG4 breaking the rules, Immunology 105 (2002), pp. 9-19.)- IgGl isotypes have been chosen as much for their relatively dependable physical behaviour as their functional properties. For example, IgG3 mAbs have been reported to form aggregates, and IgG4 mAbs exhibit intermolecular exchange between heavy chains creating a mixture of bispecific antibodies. Very likely, the IgGl remains the most appropriate isotype for its functional properties. The complete human IgGl sequence is well known. Thus, IgGl or vagrants thereof is the preferred context for the present antibody or fragment, although other IgG' s are envisaged.

The invention also provides polynucleotides encoding the present antibodies or fragments, said polynucleotides being preferably DNA or RNA. The RNA is preferably mRNA.

Also provided are vehicles comprising one or more of the antibodies or fragments of the present invention, for delivery thereof to a patient. The vehicle may be a pharmaceutical composition or a vector (for instance a viral vector) comprising the antibody or fragment. Preferably, the vector comprises polynucleotides encoding the antibody or fragment.

The patient is preferably a mammal, more preferably a primate and most preferably a human. The patient may be infected with, or suspected of being infected with avian flu, especially H5N1, or any other virus mentioned herein. Preferably, the patient may not be infected, but it is desired to immunize the patient. In fact, the present antibodies or fragments are particularly useful in passive vaccines. Thus, in a further aspect, the present invention provides a passive vaccine comprising said antibodies or fragments, preferably in a suitable vehicle, as described above.

The antibody or fragment is capable of recognising at least one epitope on, or derived from, H5N1. The fragment is any portion of an antibody comprising at least one CDR, as described herein, capable of biding to and recognising an H5N1 epitope. The same CDR may recognise more than one epitope on H5N1, although is this thought to be unlikely. However, where multiple CDRs are provided, then more than one separate epitope may be recognised. Recognition may lead to identification of the presence of an H5N1 epitope, for instance in an assay. Alternatively, recognition may lead to neutralisation, for instance by stimulation immune clearance via the complement pathway in the case of an antibody or fragment administered to a patient.

The epitope may have been derived from H5N1 in the sense that it was processed, either by an antigen presenting cell or by enzymatic cleavage.

It will be appreciated that the nature of the linker, if present, is not essential to the invention, provided that binding and recognition of the antigen is not impaired, unless this is desired. The length of the linker may affects the folding of the scFv polypeptide, and a joining peptide of 15-20 amino acids seems to be optimal and is, therefore preferred. Single-chain antibodies with very short linkers had a tendency to form multimers which led to a higher apparent affinity. The fragments with linkers longer than 11 residues remained monomeric (Alfthan K, Takkinen K, Sizmann D, Soderlund H, Teeri TT. Properties of a single-chain antibody containing different linker peptides. Protein Eng. 1995 Jul;8(7):725-31).

A preferred and common scFv linkers is (Gly₄-Ser)₃. The structural properties of this linker have been examined by NMR spectroscopy (Freund C, Ross A, Guth B, Plϋckthun A, Holak TA. Characterization of the linker peptide of the single-chain Fv fragment of an antibody by NMR spectroscopy. FEBS Lett. 1993 Apr 5;320(2):97-100) and in this particular case it was found that the linker had essentially no influence on the structure of the variable domains. It was also noted that the chemical shifts of the serine and glycine residues were similar to those obtained for these amino acids in water, indicative of the exposure of the linker to solvent. Recent work with phage display technology has enabled the optimization of the linker peptide sequence of an scFv by utilizing phage display technology (Turner DJ, Ritter MA, George AJ. Importance of the linker in expression of single-chain Fv antibody fragments: optimisation of peptide sequence using phage display technology. J Immunol Methods. 1997 Jun 23;205(l):43-54.). The choice of variable domain orientation ((VL-linker-VH construction or VH-linker-VL construction) may be useful in optimising binding activity and this may be characteristic of individual scFv binding sites and selected linkers.

Methods of gene or immunotherapy, or vaccination, comprising administering the antibody or fragment of the invention, are also provided.

The antibody or fragment may also be used in the manufacture of a medicament for the vaccination or treatment of H5N1. Also provided is an antibody or fragment for use in the vaccination or treatment of H5N1.

We applied a very effective and safe in vitro phage display approach (23) for the isolation of single chain fragment variable (scFv) human antibodies using inactivated H5N1 virus as bait. Phage-displaying scFv H5N1 reactive antibodies were obtained after three rounds of selection on H5N1 virus coated immunotubes and subsequent amplification in TGl E. coli cells.

Distinct high-reactive phage antibody clones were selected, amplified, developed in protein scFv format and purified. These human scFvs were able to recognize the H5N1 virus using laboratory techniques such as ELISA and Western blot. Furthermore, two of them were found able to neutralize H5N1 strains. Finally, the genes encoding for the selected antibody fragments have been isolated and sequenced, thus facilitating various molecular approaches, including the construction of whole recombinant human immunoglobulin (IgG) to improve the efficacy of immunotherapy of H5N1 disease.

The invention will now be described with reference to the accompanying non-limiting figures and Examples. All references cited herein are incorporated by reference, to the extent that they do not conflict with the present invention. EXAMPLES

MATERIALS AND METHODS

Antigens.

The H5N1 virus used in this study was an inactivated, purified A/Vietnam/ 1194/2004 (H5N1) virus strain provided by National Institute for Biological Standards and Control (NIBSC).

H5N2 (A/Mallard/Italy/80/93), H5N3 (A/Mallard/Italy/208/00) and HlNl (A/Mallard/Italy/185/96) avian influenza virus (AIV) strains were a generous gift from Dr. Delogu (University of Bologna).

ETH-2 antibody phage library

The synthetic human recombinant antibodies library ETH-2 was used in the screening for anti-H5 Hemagglutinin scFv antibodies. The ETH-2 scFv library (24) consists of a large array (more than 10⁹ antibody combination) of scFv polypeptides displayed on the surface of Ml 3 phage. It was built by random mutagenesis of the CDR3 of only three antibody germline gene segments (DP47 for the heavy chain, DPK22 and DPL 16 for the light chain). Diversity of the heavy chain was created by randomizing four to six positions replacing the pre-existing positions 95 to 98 of the CDR3. The diversity of the light chain was created by randomizing six positions (91 to 96) in the CDR3. Antibody residues are numbered according to Ref. 22.

Isolation of phage antibodies from ETH-2 library

Specific scFvs were recovered by absorbing an aliquot of the ETH-2 library, containing 10¹³ cfu phage, in 4 ml PBS to the immunotubes (Nunc Maxisorp; Denmark) that were coated overnight (ON) at room temperature (RT) with 5 μg/ml of inactivated, purified A/Vietnam/ 1194/2004 (H5N1) virus in PBS. After biopanning, phages were eluted with 1 ml of 100 mM triethylamine and the solution was immediately neutralized by adding 0.5 ml of 1 M Tris-HCl pH 7.4. Eluted phages were used to infect log phase TGl E. coli bacteria [supE hsdΔ5 thi Δ(lac-proAB) F'(traD36 proAB+ lac!qlacZΔM15)] and amplified for the next round of selection as described elsewhere (5, 24). Briefly, 50 ml of 2xYT medium (Sigma-Aldrich; MO, USA) with 100 μg/ml ampicillin (2^χYTA) and glucose 1% were inoculated with enough bacterial suspension to yield an OD ₆o_{O nm} ~ 0.05-0.1.

The culture was grown up to OD_{600 nm} 0.4-0.5 and infected with M 13 K07 helper phage in a ratio of around 20:1 phage/bacteria. The rescued phages were concentrated by precipitation with PEG 6000 and used for next round of panning (usually three to recover antigen-specific antibody phages from the ETH-2 library). Plating on agar of TGl cells infected with a pool of phage antibodies from third selection allowed individual clones harbouring phagemid to grow. Several bacterial clones were tested for their ability to secrete functional, anti-H5Nl scFvs. For soluble scFv preparation, the individual colonies were grown in 96 flat bottomed wells (Nunc) for 2 hours at 37⁰C in 180 μl 2^χYTA medium and glucose 0.1% in 96 well plates and inducted with 50 μl 2^χYTA medium and 6 mM isopropyl-β-D-thiogalactopyranoside (IPTG) (Sigma). The following day the plates were spinned down at 1800 g for 10 minutes and the supernatants containing soluble scFvs were recovered and tested for specificity.

ELISA

96 well ELISA-plates (Nunc Maxisorp) were coated ON at RT with 0,5 μg of antigen (inactivated, purified virus or irrilevant proteins) in PBS. T he following day a blocking solution, consisting of 2% non fat dry milk in PBS (MPBS), was added; plates were washed with PBS containing 0.05% Tween 20 (TPBS) and incubated for 2 hours at RT with 50 μl of supernatants containing soluble scFv antibody, anti- Flag M2 antibody (1.6 μg/ml Sigma-Aldrich; MO, USA) and anti-mouse HRP conjugated antibody (1.6 μg/ml Dako; Denmark). The reaction was developed using 3, 31- 5, 51- tetramethylbenzidin BM blue, POD-substrate soluble (Roche Diagnostics; IN, USA) and stopped by adding 50 μl of 1 M sulfidric acid. The reaction was detected with an ELISA reader (Biorad; CA, USA), and the results were expressed as A (absorbance) = A(450 nm)-A(620 nm). Wells equal to and above three time as much the background signal were considered positive. DNA characterization and sequences

Plasmid DNA from individual bacterial colonies was digested with specific endonucleases and CDR regions were sequenced with an automated DNA sequencer (M- Medical/Genenco, Pomezia, Italy) using Fdseql (5'-GAA TTT TCT GTA TGA GG-3' SEQ ID NO. 81) and pelBback (5'-AGC CGC TGG ATT GTT ATT AC-3' SEQ ID NO. 82) primers.

Western Blot

1 μg of inactivated, purified A/Vietnam/ 1194/2004 (H5N1) virus or 0,5 μg of H5N1 or irrelevant proteins were loaded onto 10% SDS-PAGE in reducing and non- reducing conditions, and then transferred to a nitrocellulose membrane using standard procedures. The membrane was blocked in 4% MPBS ON at RT. Blotted proteins were incubated for 2 hours with supernatant containing soluble scFvs, washed with 0,05% TPBS and incubated again with 5 μg/ml anti-Flag M2 mouse antibody (Sigma- Aldrich) in 2% MPBS. After an additional incubation for 1 hour at RT in presence of 5 μg/ml of a goat anti-mouse antibody HRP-conjugated (Dako), the reaction was developed and visualized with a chemiluminescence detection kit (Pierce; IL, USA).

Soluble ScFv purification

The positive clones were cultured for large-scale scFv production. TGl E. CoIi infected cells were cultured at 30°C in 2><YT containing 100 μg/mL ampicillin and 0.1% glucose till OD_{60O nm} ⁼ 0.5. After induction of antibody expression by adding 1 mM IPTG to culture, cells were incubated for another 3 h at 3O⁰C. Periplasmic extracts were prepared by resuspending the bacterial pellet in 1/20 the original volume of 30 mM Tris, pH 7.0, 20% sucrose and 1 mM EDTA. After incubation for 20 min on ice, the debris was removed by centrifugation and supernatants (periplasmic fraction) were filtered (0.2 μm, Millipore, Bedford, MA). His-tagged scFv fragments were purified by immobilized metal affinity chromatography using Ni2+- nitriloacetic acid agarose (Qiagen, Madison, WI). ScFv fragments were eluted with 250 mM imidazole in PBS, dialyzed against PBS, tested for specific antigen recognition and stored at -80⁰C in small aliquots. Hemagglutination-inibition (HI) test

ScFv fragments were tested for antibodies to the influenza A/Vietnam/I 194/2004 H5N1 virus strain by HI test. The HI test was performed using horse red blood cells (HRBCs) according to standard procedures with minor modifications (20). Basically, a 1% suspension of HRBCs in 0.5% BSA/PBS was used throughout the procedure and HI titres were read after 60 min.

Generation of pseudotypes bearing H5 hemagglutinin

Influenza lentiviral pseudotypes, expressing luciferase reporter gene and bearing HA on the surface, were generated as recently described (21). Briefly, 293T cells were co-transected with the following four plasmids: the gag-pol and rev constructs, the pi.18 plasmid expressing HA from Influenza A/VietNam/ 1194/04 (H5N1, clade 1), and the reporter plasmid expressing renilla luciferase gene (21, 14). In order to obtain pseudotypes bearing HA from H5N1 clade 2 strains, nucleotide sequences encoding Indonesia/5/05, whooping swan/Mongolia/244/05 and Anhui/1/05 HAs were synthesized by GeneArt. HA sequences were subsequently cloned into the suitable pi.18 vector. The multi-basic cleavage site of VietNam/1194/04 HA was substituted into the other H5 sequences in an effort to generate functionally active pseudotyped vectors and to confer similar entry (11). Twenty-four hours post infection, 1 U of exogenous neuraminidase (Sigma) was added to allow the release of pseudotypes. Culture supernatants were collected 48 and 72 hrs post-infection, filtered and stored at -80 ⁰C. Pseudotypes were than titrated on 293T cells to calculate the TCID50.

Neutralization assay

Neutralizing antibodies were measured as reductions in Luciferase reporter gene expression after a single round of pseudotypes infection in 293 T cells as recently described (21). Briefly, 200 pseudotypes giving 100 RLU (Relative Luminescence Unit) was incubated with various dilutions of sheep serum or ScFv at various dilutions for 1 h at 37⁰C in 96-well flat-bottom culture plates. Freshly trypsinized cells were added to each well. After a 48-h incubation, cells were lysated and transferred to 96-well white solid plates for measurements of luminescence using a Molecular Devices Luminometer. The 50% inhibitory dose (IDs₀) was defined as the serum dilution required to cause 50% reduction in RLU compared to virus control wells after subtraction of background RLU.

Hyper-immune sheep seruA were obtained by immunizing 2 animals subcutanely with A/VietNam/1194/04 virus (NIBRG- 14).

RESULTS AND DISCUSSION

ScFv antibodies to H5N1

To isolate specific scFv antibodies, an aliquot of the human synthetic ETH-2 library containing about 1x10¹² cfu phages was introduced for panning into Maxisorp immunotubes coated with inactivated, purified H5N1 virus as a bait. Non-specifically absorbed phage were removed by washing. Bound phage were eluted, amplified and used for next round of panning as described (5, 24). After three rounds of selection phage antibody populations specifically recognizing the H5N1 virus were isolated. Plating on agar TGl phage antibody-infected cells allowed growth of individual phagemid clones. Soluble scFvs derived from IPTG induced colonies, were screened by ELISA and several of them proved to be specific for H5N1 virus (Fig.l). Molecular studies showed that all scFv H5N1 -positive antibodies had a MW of about 27 kDa as expected from the biochemical constitution of these immunoglobulin fragments (24) and that the genes encoding for VH and VL displayed on Ml 3 phage were correctly expressed (data not shown).

VH and VL gene sequences indicated the scFvs fall into five different classes possessing a unique DNA sequence encoding for the CDR3 region (Fig.2). CDRl and CRD2 regions were 100% identical as expected (24). For biochemical and functional studies were chosen the most reactive ELISA clones AV.G4, AV. C4, AV.E2, AV. A6 and AV.D1, each one representative of the five classes of the scFv antibodies.

Specificity of the scFv antibodies for H5 subtype of HA

In order to assess the antibody specificity, scFv clones AV2.G4, AV. C4, AV. E2, AV.A6 and AV. Dl were tested on different AIV strains (Table 1). No one scFvs reacts towards HlNl virus. Cross-reactivities were instead observed with viruses strains of H5 subtype. The scFv clones AV. C4, AV.D1 and AV.E2 bound with a conserved epitope shared by three different virus strains belonging to the H5 subtype, i.e. the homologous strain A/Vietnam/ 1194/04 (H5N1), and two viruses isolated form wild birds, A/Nammalr/Italy/80/93 (H5N2) and A/Mallard/Italy/208/00 (H5N3). Conversely the scFv clones AV. G4 and AV. A6 displayed a more restricted recognition pattern since no reactivity with the H5N3 strain was observed. The biochemical nature of the recognized target antigen from the scFvs were also investigated by western blot. The results of this study showed that the scFv clones AV. C4 and AV. G4 detected protein bands of about 70 and 50 kDa, corresponding to the expected MW of the uncleaved HA protein (Fig.3, A) and major subunit (HAl) of the HA protein (Fig.3, B) in reducing and non reducing condition, respectively. Conversely, the other scFvs do not stain AIV molecules in western blot analysis. From this study we may hypothesize that the scFv antibodies recognized a different biochemically structured conformed target antigen. In particular, AV. C4 and AV. G4 epitopes could be linear while it cannot in principle be ruled out the conformational structure of the AV. A6, AV. E2 and AV.D1 epitopes.

Neutralization assay

To evaluate the functional activity of those monoclonal antibodies, all five scFvs were analysed using the HI test against the influenza A/Vietnam/I 194/2004 H5N1 virus strain, as described (20). Among all scFv antibodies, only the clone A V. Dl demonstrated significant HI activity, even though at very high concentration (data not shown). Since HI requires both antibody binding to, or nearby, the HA receptor binding site (RBS), and antibody ability to cause steric hindrance, the poor or absent HI activity observed in this study could be due to antibody binding outside the virus RBS as well as to a less effective barrier between virion and red blood cells posed by the reduced size of these IgG fragments (27 kDa) compared to intact immunoglobulins (18). It should be emphasized here that even infection-neutralizing antibodies may have poor or absent HI activity (8, 10, 15).

To further investigate whether the scFv binding may interfere with H5N1 infection, the monoclonal antibodies were analysed in a pseudotype-based neutralization assay. HA sequences of the majority of HP H5N1 viruses in avian species segregate into two major phylogenetic lineages, termed Clade 1 and Clade 2 (25). Clade 1 viruses circulated in Cambodia, Thailand and Viet Nam and caused human infections during 2004 and 2005 and in Thailand in 2006. Multiple sub-clades of clade 2 have been identified, three of which (subclades 2.1, 2.2 and 2.3) have been responsible for human cases and differ in geographical distribution (26). Pseudotypes expressing HA from the homologous VietNam/ 1194/04 (clade 1), and from Indonesia/5/05, whooping swan/Mongolia/244/05 and Anhui/1/05 H5N1 strains, representatives of subclades 2.1, 2.2 and 2.3 respectively, were incubated with serial dilution of purified scFv antibodies, and the reduction in luciferase activity in the target cells was measured in a single-round neutralization assay. Under these conditions, the scFvs AV. Dl and AV. C4 were able to neutralize approximately 50% of VietNam/1194/04 pseudotype infection at the highest concentration tested, with an ID50 of 2.25 μg.

Interestingly, the same scFv clones were able to fully neutralize the whooping swan/Mongolia/244/05 and Indonesia5/05 strains (ID50 of 2.3 and 1.6 μg, respectively), whereas no inhibition of infection was observed versus the clade 2.3 Anhui 1/05 strain. (Fig. 4). All the others scFv antibodies (AV.A6, AV.E2 and AV. G4) did not interfere with the infection and their neutralization activity was similar to that observed with the irrelevant scFv antibody to glucose oxidase (in Figure 4, the neutralization profile of scFv AV. A6 is representative of the negative clones). The sheep iperimmune serum were used as positive control.

Several factors could be involved in the different biological activity exerted by the scFv antibodies including the affinity and function of the bound epitope domain. Very likely the two scFvs AV. C4 and AV.D1, both of which are involved in neutralization activity, recognize distinct epitopes with different molecular structure (linear and conformational respectively) as suggested by Western blot analysis.

CONCLUSIONS

The prospect of passive immunotherapy for infectious disease has received considerable attention in recent years (2, 6, 17). However, the administration of immunoserum or monoclonal antibodies generated via conventional hybridoma technology, while feasible and effective, presents several limitations. For example, the treatment of human patients with rodent monoclonal antibodies is limited by the severe adverse effects due to its xenogenic origin. Hence, these monoclonal antibodies require expensive and very complex genetic manipulations to generate chimaeric and/or humanized mAbs with lower but still undesired immunogenicity.

Recently, human mAbs generated via EBV immortalization of human lymphocytes have shown potent inhibition of H5N1 in infected animals (19). However, these mAbs require biotechnical challenging for large scale antibody production and pose important safety issues for the presence of an infective agent in the cell bank system. The construction of libraries of recombinant antibody fragments that are displayed on the surface of filamentous phage have become an important technological tool in generating new monoclonal antibodies for research and clinical applications (1). Human antibodies obtained by this method do not induce harmful immune response in patients, in comparison to murine, chimaeric or humanized monoclonal antibodies. Moreover, large amount of scFv antibodies can be obtained in simple and very safe prokaryotic expression systems, such as fermentation systems.

The efficacy of this strategy has been recently confirmed with the construction of combinatorial antibody libraries from the lymphocytes of survivors of the Turkish H5N1 avian influenza outbreak (9). In particular, six different antibody fragments isolated from these immuno-libraries were chosen for conversion into whole IgGl antibodies, which were expressed and proved effective in in vitro system versus H5N1 infection. Conversely, the scFvs with potent neutralization activity here described originated from a synthetic human antibody library constructed using the principles of protein design (24). Using this strategy we have shown for the first time that human scFv antibodies endowed with broadly neutralizing properties against H5N1 can be easily obtained from an entirely artificial antibody repertoire. The scFv antibodies here described meet important criteria for a potential anti-H5Nl compound:

1) they are human, hence poorly or not at all immunogenic, and

2) they neutralize H5N1 infection.

Furthermore, the small molecular size should provide for efficient tissue penetration, yet give rapid plasma clearance (7) and are, therefore, useful new tools to fight H5N1 infection in humans.

The scFv clones AV. C4 and AV. G4 detected protein bands of about 70 and 50 kDa in non reducing and reducing conditions respectively. This data matches well with the MW of HA protein and its major subunit HAl (see Figure 3). While none of the five scFv antibodies recognizes HlNl (A/Mallard/Italy/185/1996; A/Solomon Islands/3/2006) and H3N2 (A/Wisconsin/67/2005) strains, the clones AV.C4, AV.D1 and AV.E2 bound with a conserved epitope shared by three different virus strains belonging to the H5 subtype, i.e. the reassortant (viruses containing two or more pieces of nucleic acid (segmented genome) from different parents - such viruses are produced in cells co-infected with different strains of a given virus) H5N1/ NIBRG-14 (A/Vietnam/I 194/2004), and two viruses isolated from wild birds, A/Mallard/Italy/80/1993 (H5N2) and A/Mallard/Italy/208/2000 (H5N3) (see Table 1).

We have, therefore shown the clones AV.G4 and AV.C4 recognize the H5 haemagglutinin (and precisely its major subunit HAl) of H5N1. The clones AV.D1 and AV.E2 are also expected to recognize the haemagglutinin protein of H5 subtype.

The scFv AV. A6 reacts in ELISA against the reassortant H5N1/ NIBRG-14 (A/Vietnam/I 194/2004) and it also cross-reacts with A/Mallard/Italy/80/1993 (H5N2). However, AV.A6 displayed a more restricted recognition pattern since no reactivity with the H5N3 [A/Mallard/Italy/208/2000] strain was observed, as with the scFv AV.G4. But, unlike scFv AV. G4, the scFv AV.A6 does not react in western blot (see Figure 3). Therefore, it is reasonable to assume that the scFv AV. A6 does recognize the Hemagglutin H5 protein, albeit with a more restricted epitope (displayed on H5 protein of H5N1/ NIBRG-14 (A/Vietnam/I 194/2004) and on A/Mallard/Italy/80/1993 (H5N3), but not displayed on H5 protein of H5N3 [A/Mallard/Italy/208/2000] (see Table. I)).

The CDRs from the clone AV.C4 (SEQ ID NOS. 12 and SEQ ID 15 corresponding to the Heavy and Light Chain CDR3s respectively) are the most preferred, followed by AV. Dl (SEQ ID NOS. 28 and SEQ ID 31).

Obviously, for each single Antibody, the Heavy Chain CDR3 is preferred over the Light Chain CDR3. Since 1986, when the first three-dimensional structure of an antigen — antibody complex was solved (Amit et al., 1986), it is known that most of the contacts with the antigen are made by the heavy chain and in particular by HCD R3s. Moreover, HCDR3s (Heavy Chain CDR3s) are the only CDRs which are not structurally constrained to canonical structures (Barre et al., 1994).

It will be appreciated that not all CDRs, in a single chain having more than one CDR, will be equally important for antibody-antigen interaction. In fact, VHCDR3 (Heavy Chain CDR3) is the most variable CDR and forms the centre of the antigen combining site (Amit et al., 1997). However in many cases, the VLCDR3 (Light Chain CDR3) and/or the other CDR regions (CDRl and CDR2) may also be helpful for accuracy and higher affinity of binding toward the target epitope.

Avian influenza viruses

HlNl HlNl' H3N2^a H5N2 H5N3 H5N1¹ scFvs (A/Mallard/Italy/ (A/Solomon (A/Wisconsin/ (A/Mallard/Italy/ (A/Mallard/It (A/Vietnam/ 185/96) Islands/3/06) 67/05) 80/93) aly/ 1194/04)

AV.G4 - - - + - ++++

AV.C4 - - - +++ -H-+ ++++

AV.A6 - - - ++ - ++++

AV.D1 - - - ++ +++ ++++

AV.E2 - - - +++ +++ ++++

Ul

Table 1. ELISA reactivity of the scFvs against H5N1 to different avian influenza viruses.

Microtiter plates were coated with the indicated antigens diluted in coating buffer and incubated ON at RT. After blocking, plates were incubated for 2 hours at RT with supernatants containing soluble scFv antibody, anti- Flag M2 antibody and anti-mouse HRP conjugated antibody, washed, and later detected using POD-substrate (Roche Diagnostics). The reactivity was read with an ELISA reader at absorbance of 450 run.

[+, above background and <2-fold; ++, between 2- and 3 -fold; +++, between 3- and 10-fold; ++++, > 10-fold above background; not measurable above background]. aMutagrip Pasteur vaccine (season 2007-2008). Reverse genetically modified reassortant of H5N1 also named H5N1 /NIBRG- 14.

REFERENCES

1. Azzazy, H. M. and W. E. Highsmith, Jr. 2002. Phage display technology: clinical applications and recent innovations. Clinical Biochemistry 35: 425—445.

2. Casadevall, A., E. Dadachova, and L. Pirofski. 2004. Passive antibody therapy for infectious diseases. Nat Rev Microbiol 2: 695-703.

3. de Jong, M.D., C. P. Simmons, T. T. Thanh, V. M. Hien, G.J. Smith, T. N. B. Chau, D. M. Hoang, N. V. V. Chau, T. H. Khanh, V. C. Dong, P. T. Qui, B. V. Cam, D. Q. Ha, Y. Guan, J. S. M. Peiris, N. T. Chinh, T. T. Hien and J. Farrar. 2006. Fatal outcome of human influenza A (H5N1) is associated with high viral load and hypercytokinemia. Nat Med 12: 1203-1207.

4. de Jong, M.D., T. T. Tran, H. K. Truong, M. H. Vo, G. J. Smith, N. V. Chau, B. V. Cam, P.T. Qui, D Q. Ha, Y. Guan, J.S. M. Peiris, T. T. Hien, and J. Farrar. 2005. Oseltamivir resistance during treatment of influenza A (H5N1) infection. N Engl J Med 353: 2667-2672.

5. Flego, M., P. Di Bonito, A. Ascione, S. Zamboni, A. Carattoli, F. Grasso, A. Cassone and M. Cianfriglia. 2005. Generation of human antibody fragments recognizing distinct epitopes of the nucleocapsid (N) SARS-CoV protein using a phage display approach. BMC Infectious Diseases 5: 73.

6. Hanson, B. J., A. C. Boon, A. P. Lim, A. Webb, E. E. Ooi, and R. J. Webby. 2006. Passive immunoprophylaxis and therapy with humanized monoclonal antibody specific for influenza A H5 hemagglutinin in mice. Respir Res 7: 126.

7. Holliger, P. And P. J. Hudson. 2005. Engineered antibody fragments and the rise of single domains. Nature Biotechnology 23: 1126-1136.

8. Imai, M., K. Sugimoto, K. Okazaki and H. Kida. 1998. Fusion of influenza virus with the endosomal membrane is inhibited by monoclonal antibodies to defined epitopes on the hemagglutinin. Virus Res 53: 129-39.

9. Kashyap, A. K., J. Steel, A. F. Oner, M. A. Dillon, R. E. Swale, K. M. Wall, K. J. Perry, A. Faynboym, M. Ilhan, M. Horowitz, L. Horowitz, P. Palese, R. R. Bhatt, and R. A. Lerner. 2008. Combinatorial antibody libraries from survivors of the Turkish H5N1 avian influenza outbreak reveal virus neutralization strategies. Proc Natl Acad Sci U S A 105: 5986- 5991.

10. Kida, H., S. Yoden, M. Kuwabara and R. Yanagawa. 1985. Interference with a conformational change in the haemagglutinin molecule of influenza virus by antibodies as a possible neutralization mechanism. Vaccine 3 (Issue 3): 219-222.

11. Kong, W.P., C. Hood, Z. Y. Yang, et al. 2006. Protective immunity to lethal challenge of the 1918 pandemic influenza virus by vaccination. Proc Nat Acad Sci USA. 103:15987-91.

12. Le, Q. M., M. Kiso, K. Someya, Y. T. Sakai, T. H. Nguyen, K. H. L. Nguyen, N. D. Pham, H. H. Ngyen, S. Yamada, Y. Muramoto, T. Horimoto, A. Takada, H. Goto, T. Suzuki, Y. Suzuki and Y. Kawaoka. 2005. Avian flu: Isolation of drug-resistant H5N1 virus. Nature 437: 1108.

13. Li, K.S., Y. Guan, J. Wang, G. J. Smith, K. M. Xu, L. Duan, A. P. Rahardjo, P. Puthavathana, C. Buranathai, T. D. Nguyen, A. T. S. Estoepangestie, A. Chaisingh, P. Auewarakul, H. T. Long, N. T. H. Hanh, R. J. Webby, L. L. M. Poon, H. Chen, K. F. Shortridge, K. Y. Yuen, R. G. Webster and J. S. M. Peiris. 2004. Genesis of a highly pathogenic and potentially pandemic H5N1 influenza virus in eastern Asia. Nature 430: 209- 213.

14. Naldini L. 1998. Lentiviruses as gene transfer agents for delivery to non-dividing cells. Current opinion in biotechnology 9:457-63. 15. Okuno, Y., Y. Isegawa , F. Sasao, and S. Ueda. 1993. A Common Neutralizing Epitope Conserved between the Hemagglutinins of Influenza A Virus Hi and H2 Strains. Journal of Virology 67: 2552-2558.

16. Puthavathana, P., P. Auewarakul, P. C. Charoenying, K. Sangsiriwut, P. Pooruk, K. Boonnak, R. Khanyok, P. Thawachsupa, R. Kijphati and P. Sawanpanyalert. 2005. Molecular characterization of the complete genome of human influenza H5N1 virus isolates from Thailand. J Gen Virol 86: 423-433.

17. Sawyer, L. A. 2000. Antibodies for the prevention and treatment of viral diseases . Antiviral Research 47: 57-77.

18. Schofield, D. J., J. R. Stephenson and N. J. Dimmock. 1997. Variations in the neutralizing and haemagglutination-inhibiting activities of five influenza A virus-specific IgGs and their antibody fragments. J Gen Virol 78: 2431-9.

19. Simmons C. P., N. L. Bernasconi, A. L. Suguitan Jr., K. Mills, J. M. Ward, N. V. V. Chau, T. T. Hien, F. Sallusto, D. Q. Ha, J. Farrar, M. D. de Jong, A. Lanzavecchia and K. Subbarao. 2007. Prophylactic and Therapeutic Efficacy of Human Monoclonal Antibodies against H5N1 Influenza. PLoS Med 4: el78.

20. Stephenson, L, J. M. Wood, K. G. Nicholson, and M. C. Zambon. 2003. Sialic acid receptor specificity on erythrocytes affects detection of antibody to avian influenza haemagglutinin. J Med Virol 70: 391-398

21. Temperton, N. J., K. Hoschler, D. Major, C. Nicolson, R. Manvell, V. M. Hien, D. Q. Ha, M. de Jong, M. Zambon, Y. Takeuchi, R. A. Weiss. 2007. A sensitive retroviral pseudotype assay for influenza H5N1 -neutralizing antibodies. Influenza and Other Respiratory Viruses 1: 105-112.

22. Tomlinson, I. M., J. P. Cox, E. Gherardi, A. M. Lesk and C. Chothia. 1995. The structural repertoire of the human V kappa domain. EMBO J. 14: 4628-4638.

23. Ungchusak, K., P. Auewarakul, S. F. Dowell, R. Kitphati, W. Auwanit, P. Puthavathana, M. Uiprasertkul, K. Boonnak, C. Pittayawonganon, N. J. Cox, S. R. Zaki, P. Thawatsupha, M. Chittaganpitch, R. Khontong, J. M. Simmerman, and S. Chunsutthiwat. 2005. Probable person-to-person transmission of avian infl uenza A (H5N1). N Engl J Med 352: 333-40.

24. Viti, F., F. Nilsson, S. Demartis, A. Huber and D. Neri. 2000. Design and use of phage display libraries for the selection of antibodies and enzymes. Methods Enzymol 326: 480-505.

25. World Health Organization Global Influenza Program Surveillance Network. 2005 Oct [date cited]. Evolution of H5N1 avian influenza viruses in Asia. Emerg Infect Dis [serial on the Internet]. Available from http://www.cdc.gov/ncidod/EID/voll lnolO/05-0644.htm

26. World Health Organization. No authors listed. 2007. Antigenic and genetic characteristics of H5N1 viruses and candidate H5N1 vaccine viruses developed for potential use as pre-pandemic vaccines. Available from http://www.who.int/csr/disease/avian influenza/guidelines/summarvH520070403.pdf Sequences

Reference is also made to the figures, where Figures 5-6, 7-8, 9-10, 11-12 and 13-14 show the nucleotide and amino acid sequences of AVC4, AVDl, AV E2 AV A6 and AVG4, respectively.

Claims

Claims

1. An antibody or antibody fragment, which is capable of recognising H5N1, and comprises at least one Heavy Chain CDR selected from the group consisting of: SEQ ID NOS. 12, 28, 44, 60 and 76.

2. An antibody or antibody fragment, which is capable of recognising H5N1, and comprises at least one Light Chain CDR selected from the group consisting of: SEQ ID NOS. 15, 31, 47, 63 and 79.

3. An antibody or antibody fragment, which is capable of recognising H5N1, and comprises; at least one Heavy Chain CDR selected from the group consisting of: SEQ ID NOS. 12, 28, 44, 60 and 76; and at least one Light Chain CDR selected from the group consisting of: SEQ ID NOS. 15, 31, 47, 63 and 79.

4. An antibody or antibody fragment according to any preceding claim, which is capable of inhibiting H5N1 infection across more than one H5N1 clade or serotype.

5. An antibody or antibody fragment according to any preceding claim, wherein the Heavy and/or Light Chains comprise at least one additional CDR.

6. An antibody or antibody fragment according to claim 5, wherein the at least one additional CDR are CDRIs selected from AVC4, AVDl, AVE2, AV A6 and AVG4; or are CDR2s selected from AVC4, AVDl, AVE2, AV A6 and AVG4.

7. An antibody or antibody fragment according to any preceding claim, comprising first and second chains, each chain comprising a variable region, the chains being linked by a suitable linker.

8. An antibody or antibody fragment according to claim 8, wherein the first chain is a Light Chain and the second chain is a Heavy Chain.

9. An antibody or antibody fragment according to claim 7 or 8, which is an scFv antibody of two chains joined by a linker.

10. An antibody or antibody fragment according to claim 9, which is a single chain Fv fragment fused to an immunoglobulin CH3 domain, thereby providing a bivalent or bispecific protein.

11. An antibody or antibody fragment according to any preceding claim, which is a multimer of more than one antibody or fragment linked via a multifunctional linker.

12. An antibody or antibody fragment according to any preceding claim, wherein the antibody or fragment is provided with a constant region selected from those consisting of the constant regions of human IgGA, IgGB, IgGC, IgGD, IgGE, IgGF and IgG.

13. An antibody or antibody fragment according to claim 12, wherein the one or more CDRs are provided in the context of an IgGl framework and constant region.

14. An antibody or antibody fragment according to any of claims 4-13, wherein the H5N1 clade or strain is H5N1 VietNam/1194/04, whooping swan/Mongolia/244/05 or Indonesia5/05.

15. A polynucleotide encoding the antibody or fragment according to any preceding claim.

16. A pharmaceutical composition or a viral vector comprising an antibody or fragment according to any of claims 1-14 or a polynucleotide according to claim 15.

17. A kit for the detection of H5N1 in a sample, the kit comprising an antibody or fragment according to any of claims 1-14 or a polynucleotide according to claim 15 and a suitable means of detection, such as luminescence means.

18. A chip comprising a kit according to claim 16 or an antibody or fragment according to any of claims 1-14 or a polynucleotide according to claim 15.

19 A method of diagnosing whether a patient has an H5N1 infection comprising using the antibody or fragment according to any of claims 1-14 or a polynucleotide according to claim 15 to assay a sample of body fluids from an individual, said the antibodies and fragments being capable of detecting at least one epitope from or derived from H5N1.

20. A method of gene or immunotherapy, or vaccination, comprising administering a polynucleotide according to claim 15.

21. The use of an antibody or fragment according to any of claims 1-14 or a polynucleotide according to claim 15 in the manufacture of a medicament for the vaccination or treatment of H5N1.

22. An antibody or fragment according to any of claims 1-14 or a polynucleotide according to claim 15 for use in the vaccination or treatment of H5N1.

23. A method of vaccinating or treating an H5N1 infection comprising administering to a patient an antibody or fragment according to any of claims 1-14 or a polynucleotide according to claim 15.