WO2017025698A1 - Bispecific, cleavable antibodies - Google Patents

Abstract

The present invention provides a specific binding molecule comprising an immunoglobulin light chain (VL) domain and immunoglobulin heavy chain (VH) domain of a first antibody to a tissue specific antigen in which each domain is conjugated separately via a peptide linker comprising a proteolytic cleavage site to an immunoglobulin light chain (VL) domain and immunoglobulin heavy chain (VH) domain respectively of a second antibody, compositions comprising such molecules and uses in medicine.

Description

BISPECIFIC, CLEAVABLE ANTIBODIES

The present invention relates to modified therapeutic antibodies which can be prepared in a prodrug form and targeted to a site of disease where the antibodies are subsequently released in order to exert the desired biological activity of the antibodies at the specific site of interest.

Biologic drugs, such as the anti TNF-a antibody adalimumab (Humira® / Exemptia™), have represented a breakthrough in the treatment of rheumatoid arthritis. Despite this, concerns remain over the lack of efficacy in a sizable proportion of patients and the potential for systemic side effects such as infection. The design of improved biologic pro-drugs specifically targeted to the site of inflammation has the potential to alleviate current concerns surrounding the safety of biologic anti-cytokine therapies and increase pharmacological potency.

Rheumatoid arthritis (RA) is a systemic inflammatory condition that primarily affects synovial joints, characterised by persistent synovitis and destruction of bone and cartilage. RA affects around 1 % of the adult population, with a higher prevalence (2%) in the population over 60 years of age and a 3-fold higher incidence in women (1). While the aetiology of the disease remains incompletely understood, it is known that pro-inflammatory cytokines play a role in disease pathogenesis, sustaining inflammation which leads to joint destruction (2). Key cytokines in the development of RA include Tumour Necrosis Factor (TNF)-a, Interleukin (ΙΙ_)-1 β and IL-6. These cytokines can stimulate the production of matrix metalloproteinase (MMP) enzymes, destroying the extracellular matrix leading to cartilage and bone damage (3). The collagenases, MMP-1 and MMP-13, play a significant role in RA as they are shown to be the rate limiting step in the process of collagen degradation (4).

In recent years biologic therapy has revolutionized the field of RA treatment. Nonetheless the disease continues to be linked to severe pain, depression and function impairment, with 20-40% of patients failing to respond to current therapy (5, 6). The cost of treating RA by biologic agents is far higher than 'conventional' disease modifying anti-rheumatic drugs (DMARDs) and continues to be linked to negative consequences of organ toxicity (7). Targeting TNF-a using monoclonal antibodies (mAb) such as adalimumab (Humira®, Abbvie) and infliximab (Remicade®, Janssen Biologies), as a monotherapy or in combination with other DMARDs, has become the gold standard for RA therapy (8). However, while TNF-a has a highly deleterious effect in inflammatory joint diseases, it plays a vital role in the body's defence against infection (9). In the immune response to mycobacteria tuberculosis TNF-a plays a critical protective role, leading to macrophage activation, cell recruitment, granuloma formation and maintenance of granuloma integrity (10-12). Thus, systemic TNF-a blockade increases risk of tuberculosis infection and re-activation in patients with latent disease compared to alternative DMARD therapy (13). On the other hand, though the exact mechanism of non-response to anti-TNF-a biologic therapy is not clear (14), it is plausible to hypothesize that lack of efficacy may be due to sub-optimal TNF-a blockade at the site of inflammation which cannot be improved by increased systemic administration due to potential general toxicity at doses higher than those recommended (15, 16).

The past two decades have marked a substantial revolution in the treatment paradigm for RA. The advent of biologic agents has provided a new avenue for successful treatment of RA. However, there remains a considerable number of patients that do not respond to the available therapies and a treatment-free remission is rarely achieved (5, 6). Despite the obvious success of the current treatments, little effort has been invested in improving the safety profiles of the available therapeutics.

The identification of tissue specific or overexpressed antigens in the inflamed synovium may provide a solution to these concerns and allow the development of therapeutics for tissue specific drug delivery (32). Several candidates are now being considered for arthritic synovium targeting, including the onco-foetal extra-domain A of fibronectin (EDA). Indeed, a single chain antibody fragment targeting EDA has recently entered clinical evaluation as scFv-IL10 fusion protein for the treatment of RA (NCT02076659). An alternative strategy could involve the use of bispecific antibodies (BsAb) to combine tissue targeting with therapeutic function. To date no BsAb has been clinically evaluated for RA, however, two antibodies catumaxomab and blinatumomab, have been recently approved and several constructs are currently in clinical trials for cancer treatment (33).

The BsAb format DVD-lg™, has shown potential as a versatile platform for dual antigen targeting (17). One of the crucial aspects for conventional BsAb is the capacity for real synergistic activity between the two binding moieties. Differences in binding affinities may result in targeting skewed towards one antigen and, as a consequence, a sub-optimal therapeutic activity. This is probably the main drawback when combining tissue targeting moieties with existing therapeutic domains, usually characterised by very high affinities. Interestingly, the size and composition of the linker between the outer and inner variable domain of DVD-lg™ antibodies, has been shown to significantly impact on the kinetic activity of the inner region (19). The generation of activatable bispecific antibodies is described in US 2013/0266568. However, the immunoglobulin heavy chain and immunoglobulin light chain (VH / VL) domains used to mask the therapeutic antibody are not entirely cleaved off from the therapeutic antibody after activation. The different CDR regions in the antibodies bind simultaneously to separate targets. The bispecific antibodies prepared demonstrate relatively weak reduced binding affinity (blocking) compared to after protease cleavage up to a maximum reported reduction of 350-fold.

US 201 1/0178279 describes an antibody construct in which a "masking" epitope is connected by a peptide linker to an antibody to render it inactive. The masking epitope contains an epitope specific to the antibody, for example specific to one more antigen recognition sites of the antibody. However, the construct in US 201 1/0178279 has no tissue specificity. A possible solution would be the development of new agents with dual specificity with one domain targeting molecules expressed at high level [e.g. cell adhesion molecules (CAM)] at the disease site (synovium in the case of RA) and a second domain targeting cytokines of interest (e.g. TNF-a). Bispecific agents such as dual-variable-domain immunoglobulins (DVD-lg™) (17), theoretically, could deliver higher local concentrations with lower systemic exposure. In this format, the variable domains of two distinct mAbs are linked creating a tetravalent, dual-targeting single agent. While it has been shown that viable DVD-lg molecules can be identified through optimization of antibody pair, antibody variable domain orientation and linkers, an ongoing limitation of the technology is the lower binding affinity observed by the "inner domain" compared to the "outer domain". Several bodies of work have investigated the possibility of increasing the viability of the inner domain using variable linkers and it has been suggested that each antibody needs to be optimized individually in terms of inner/outer domain arrangement and linker length construction to derive the best molecule (18, 19).

The inventors have surprisingly found that by reversing this concept, the intrinsic reduced activity of the inner domain is an advantage in designing an antibody based 'pro-drug' targeting molecule. For example, in one embodiment of the invention, a pro-drug targeting molecule has been devised with an outer domain capable of targeting the inflamed synovium, and an inner domain that binds TNF-a principally in the inflamed disease tissue, while decreasing binding in the systemic circulation to the cytokine target, as the inner anti-cytokine domain as fully masked prior to enzymatic cleavage.

According to a first aspect of the invention there is provided a specific binding molecule comprising an immunoglobulin light chain (VL) domain and an immunoglobulin heavy chain (VH) domain of a first antibody to a tissue specific antigen in which each domain is conjugated separately via a peptide linker comprising a proteolytic cleavage site to an immunoglobulin light chain (VL) domain and an immunoglobulin heavy chain (VH) domain respectively of a second antibody. In other words, the immunoglobulin light chain (VL) domain of the first antibody to a tissue specific antigen is conjugated separately via a peptide linker comprising a proteolytic cleavage site to the immunoglobulin light chain (VL) of the second antibody, and immunoglobulin heavy chain (VH) domain of the antibody to a tissue specific antigen is conjugated separately via a peptide linker comprising a proteolytic cleavage site to the immunoglobulin heavy chain (VH) of the second antibody. The immunoglobulin light chain (VL) and immunoglobulin heavy chain (VH) domains of the first antibody can be considered to represent an equivalent to a single chain variable (scFv) fragment. The immunoglobulin light chain (VL) and immunoglobulin heavy chain (VH) domains of the first antibody are sufficient to provide steric hindrance to inhibit the activity of the second antibody until the specific binding molecule reaches the site of disease and is cleaved to release the second antibody. The specific binding molecule of the present invention is therefore a true pro-drug in the sense that the activity of the second antibody is fully abolished when part of the specific binding molecule and then fully active once released at the site of disease. The specific binding molecules of the invention provide a reduction in specific affinity (blocking) capacity of at least up to 1000-fold, up to 1500-fold, up to 2000-fold, or up to 2500-fold inhibition compared for the second antibody compared to after protease cleavage. In some embodiments, the activity of the second antibody may be totally blocked before cleavage of the domains of the first antibody off from the second antibody.

The present invention represents a significant improvement over the prior art in terms of the blocking capacities for the antibodies when released at the site of disease. The specific binding molecules of the invention provide a reduction in specific affinity (blocking) capacity of at least up to 2500-fold inhibition compared to 350-fold as reported in US 2013/0266568 for the second antibody compared to after protease cleavage. The construct of the present invention has only one binding region available at a time in contrast to US 2013/0266568. In the present invention, the second antibody (i.e. the therapeutic antibody) is released following proteolytic cleavage of both the linker sequences and the removal of the domains of the first antibody (i.e. the tissue targeting antibody). The construct in US 201 1/0178279 would not provide the requisite level of tissue targeting capacity as demonstrated by specific binding molecules of the present invention.

Further the specific binding molecules of the invention carries a cleavable linker on both chains (heavy and light) to provide a complete removal of the external binding regions in order to release a fully functional antibody where needed. This action will ensure that (1) the molecule in the initial phase is only capable of binding the antigen on the target tissue, ensuring accumulation in the diseased area, and (2) that following cleavage the antibody will unmask and activate the therapeutic domain as a natural IgG antibody. The external binding domain, having finished its purpose will be cleaved, released from the antibody and removed by the organism. A key advantage of the specific binding molecules of the invention is that the approach can be applied to existing therapeutic antibodies (e.g. anti-cytokine antibodies in use in rheumatoid arthritis) and used to prepare a construct which can release a final product that is in shape and activity comparable to the original therapeutic product. The domains in the first antibody to a tissue specific antigen may be from any suitable antibody which recognises and binds specifically to a tissue specific antigen or which binds to an antigen undergoing overexpression or reactivation in a disease condition.

For example, the first antibody to a tissue specific antigen may be an antibody to a Cell-adhesion molecule (CAM) (for example ICAM1 , ICAM3, VCAM1 , EpCAM), Fibronectin, Extra domain A of Fibronectin, Tenascin C, an integrin (for example aVp3, aVpl), LFA-1 , Annexin A1 , Nucleotin, Tie- 1 , Tie-2, Aminopeptidase N, CD13, CD44 and spliced variants (e.g. CD44v4, CD44v6), CD90, CD55, Folate receptor, Collagen type II and modifications thereof occurring during inflammation, Citrullinated proteins (e.g. citrullinated Fibrinogen, citrullinated Vimentin), Vascular endothelial growth factor receptor 1 (VEGFR-1), Vascular endothelial growth factor receptor 2 (VEGFR-2), Antigen recognised by the A7 synovium targeting antibody, Antigen recognised by the synovial endothelial targeting peptide CKSTHDRLC (SyETP). Without wishing to be bound by theory, these antigens may be useful in embodiments of the invention relating to the treatment of an inflammatory disease condition such as rheumatoid arthritis, other acute / chronic inflammatory conditions (e.g. vasculitis, dermatitis, inflammatory bowel disease, inflammatory neurological conditions) and cancer.

In one embodiment, the first antibody to a tissue specific antigen may be an antibody to ICAM1 , ICAM3 or the antigen recognised by the A7 synovium targeting antibody. ICAM1 (also known as CD54) encodes a surface glycoprotein expressed on endothelial cells and cells of the immune system. ICAM1 binds to CD1 1 a/CD18 or CD1 1 b/CD18 integrin proteins and has a role in the immune response. ICAM3 (also known as CD50) binds to leukocyte adhesion LFA-1 protein and has a role in the immune response. The tissue specific antigen may also be the A7 epitope identified as specific to the microvasculature of arthritic synovium. In one embodiment, the domains in the first antibody to a tissue specific antigen may be from an antibody to ICAM1 .

Alternative sources of targeting antigens for the first antibody to a tissue specific antigen may be ICAM1 , VCAM1 , EpCAM, Extra domain B of Fibronectin, Melanoma-associated Chondroitin sulfate proteoglycan (MCSP), Melanoma-associated proteoglycan (MAPG), High molecular weight melanoma associated antigen (HMV-MAA), Prostate-specific membrane antigen (PSMA), Epidermal Growth factor Receptor (EGFR), Hepatocyte growth factor receptor (HGFR), Fibroblast activation protein (FAP), Carcinoembryonic Antigen (CEA), Cell-adhesion molecule (CAM), Human B-cell maturation target (BCMA), Placental growth factor (PLGF), Folate receptor, Insulin-like growth factor receptor (ILGFR), CD133, CD40, CD37, CD33, CD30, CD28, CD24, CD23, CD22, CD21 , CD20, CD19, CD13, CD10, HER3, HER2, Non-muscle myosin heavy chain type A (nmMHCA), Transferrin, Epithelial cell adhesion molecule (EpCAM), Annexin A1 , Nucleotin, Tenascin, Vascular endothelial growth factor receptor 1 (VEGFR-1), Vascular endothelial growth factor receptor 2, (VEGFR-2), Aminopeptidase N, Tie-1 , Tie-2, or c-Met. Without wishing to be bound by theory, these antigens may be useful in embodiments of the invention relating to the treatment of a disease condition such as cancer.

Other suitable domains may be from antibodies to PNAd, α4β7, MAdCAM-1 .

The second antibody can be a therapeutic antibody, i.e. an antibody which exerts a pharmacologically useful effect. The second antibody may find use in the treatment of diseases or conditions after it is released by cleavage of the domains of the first antibody from the specific binding molecule by a protease at a site of disease.

The second antibody may be a complete antibody, for example an IgG molecule, or it may be an antibody fragment as defined below, for example an scFv, or two single-domain antibodies.

In one embodiment, the second antibody may be an anti-cytokine antibody for example an anti- TNF antibody, an anti-interferon antibody or an anti-interleukin antibody. The anti-cytokine antibody may be an antibody to an TNF selected from TNF-a, lymphotoxin-a, lymphotoxin-β, CD27L, CD30L, FASL, 4-1 BBL, OX40L or TRAIL, or an interleukin selected from IL1 , IL-2, IL-3, IL- 4, IL-5, IL-6, IL-7, IL-8, IL-9, IL-10, IL-1 1 , IL-12, IL-13, IL-14, IL-15, IL-16, IL-17, IL-18, IL-19, IL-20, IL-21 , IL-22, IL-23, IL-24, IL-25, IL-26, IL-27, IL-28, IL-29, IL-30, IL-31 , IL-32 and IL-33 (either -a or -β), an interferon selected from IFN- , IFN-β and IFN-γ), , or IFN-γ inducing factor (IGIF), a Bone morphogenetic protein (BMP), or an antibody to a chemokine, such as an antibody to a macrophage inflammatory protein (for example, MIP-1 A or MIP-1 B), Monocyte Chemotactic Protein (MCP) (for example MCP1 , 2 or 3), RANTES, CCL19, CCL21 , IP-10, GROB, Eotaxin TARC, or CD20, CD44, CD80, CD86 or CTLA-4, or angiopoietin (for example, Ang-1 or Ang-2), or an antibody to a trophic factor (for example, Epidermal growth factor (EGF), Platelet derived growth factor (PDGF), Fibroblast growth factor (FGF), Nerve growth factor (NGF), Colony stimulating factor (CSF), Granulocyte/macrophage colony stimulating factor (GM-CSF), Hepatocyte growth factor, !nsulin-like growth factor, Placenta growth factor, Vascular endothelial growth factor (e.g. VEGF-A, VEGF-C, VEGF-D), o TGF-β). In one embodiment, the anti-cytokine antibody may be an anti-TNF-α antibody, for example adalimumab, infliximab or other anti-TNF-α antibodies such as TN3-19.12 from rat.

The peptide linker comprising a proteolytic cleavage site may comprise a specific amino acid sequence recognised by a protease active at the site of disease where the second antibody is desired to exert its biological effects, for example in an inflammatory disease or cancer where inflammation or tissue remodelling may typically occur. Suitable proteolytic cleavage sites may be those recognised by any one of the following proteases: Collagenases (MMP-1 , MMP8), Gelatinases (MMP-2, MMP-9), MMP-3, M M P-7, MMP-10, MMP-12, MMP-13, MMP-14, glutamate carboxypeptidase II, cathepsin B, cathepsin L, cathepsin S, cathepsin K, cathepsin F cathepsin H, cathepsin U, cathepsin O , neutrophil elastase, plasma kallikrein, KLK3, Disi nteg ri ns and Meta llo prote in ases (e .g . ADAM 9 , ADAM10, ADAM17, ADAM-28, ADAMTS-1 , ADAMTS-4, ADAMTS-5) AMSH, y-secretase component, uPA, FAP, APCE, decysin 1 , Calpain 2, (m/ll) large subunit, Caspase 1 , apoptosis-related cysteine peptidase (IL-I P con- vertase), Granzyme A (granzyme 1 , CTL-associated serine esterase 3), fu ri n , Kallikrein-related peptidase 1 1 , Legumain, N-acetylated alpha-linked acidic dipeptidase-like 1 and Rep-sin. In certain embodiments, the proteolytic cleavage site may be a matrix metalloproteinase (MMP) or aggrecanase cleavage site such as described in Figure 12.

It may be advantageous for the peptide linker to comprise the cleavage sequence without any additional flanking sequences being present in the peptide linker. In some embodiments, the linker may be of from 5 to 26 amino acids in length, 6 to 16 amino acids, 6 to 26 amino acids, e.g. of from 5, 6, 7, 8, 9, 10, 1 1 , 12, 13, 14, 15, 16, 17, 18, 19, 20, 21 , 22, 23, 24, 25, or 26 amino acids in length. A MMP cleavage site may comprise a number of amino acid residues recognisable by MMP. Preferably, the MMP cleavage site comprises the minimum number of amino acid residues required for recognition and cleavage by MMP. Moreover, the amino acids of the MMP site may be linked by one or more peptide bonds which are cleavable, proteolytically, by MMP. MMPs which may cleave the MMP site include, but are not limited to, MMP1 , MMP2, MMP3, MMP7, MMP8, MMP9, MMP10 and MMP13 (Yu and Stamenkovic, Genes and Dev. 14, 163-176 (2000); Nagase and Fields, Biopolymers, 40, 399-416 (1996); Massova et al., J. Mol. Model. 3, 17-30 (1997); reviewed in Vu and Werb; Genes and Dev. 14, 2123-2133 (2000)).

The MMP cleavage site may be any site recognised by any matrix metalloproteinase as shown in Figure 12 selected from any one of SEQ ID NO: 1 , or SEQ ID NO: 14 to SEQ ID NO: 121 , e.g. a cleavage site from any one of MMP1 , MMP2, MMP3, MMP7, MMP8, MMP9, MMP10 or MMP13 or a sequence of amino acids which has at least 50%, 60%, 70%, 80%, 90%, 95% or 99% identity with the various published consensus cleavage sequences for such matrix metalloproteinases, using the default parameters of the BLAST computer program provided by HGMP, thereto. In one embodiment the MMP proteolytic cleavage site may have the amino acid sequence PLGLWA (SEQ ID NO: 1).

Aggrecanases which may cleave the aggrecanase site include, but are not limited to ADAMTS-4 (aggrecanase-1), ADAMTS-5 (aggrecanase-2) and ADAMTS-1 1 (Tortorella, M.D., et al Osteoarthritis Cartilage, 2001 . 9(6): p. 539-552); Abbaszade, I., et al J Biol Chem, 1999. 274(33): p. 23443-23450).

The sequences of the protein cleavage sites of ADAMTS-4 (aggrecanase-1) are shown below. Suitable ADAMTS-4 sites include:

HNEFRQRETYMVF (SEQ ID NO: 2)

QEFRGVTAV (SEQ ID NO: 3)

TEGEARGS (SEQ ID NO: 4) ADAMTS-5 (aggrecanase-2) cleavage sites may be FREEEGLGS (SEQ ID NO: 5)

SELEGRGT (SEQ ID NO: 6)

GELEGRGT (SEQ ID NO: 7)

TAQEAGEG (SEQ ID NO: 8)

KEEEGLGS (SEQ ID NO: 9)

VSQELGQR (SEQ ID NO: 10)

RPAEARLE (SEQ ID NO: 1 1) The proteolytic cleavage site may be flanked by one or more additional amino acid flanking sequences on either or on both sides of the cleavage site. Suitable amino acid residues may be glycine (G) and/or serine (S). The additional sequences may be present singly or in multimeric repeating units composed of monomers of 5 amino acids, such as dimers, trimers and/or tetramers. One flanking sequence may be G₄S which may be independently present as G₄S or in multimers of 2, 3, or 4, for example (G₄S)₂, (G₄S)₃, or (G₄S)₄. Shorter flanking sequences may be advantageous in the practice of the invention, for example, G₄S or (G₄S)₂.

In one embodiment, where the proteolytic cleavage site is PLGLWA (SEQ ID NO: 1), the cleavage site may be surrounded by adjacent flanking sequences, for example GGGGSPLGLWAGGGGS (SEQ ID NO: 12). In another embodiment the proteolytic cleavage site and linker sequence may be

GGGGSGGGGSPLGLWAGGGGSGGGGS (SEQ ID NO: 13)

The specific binding molecule of the invention is therefore a protein sequence composed of an anti- cytokine antibody in which the complementarity determining regions (CDRs) have been masked by the immunoglobulin light chain (VL) and immunoglobulin heavy chain (VH) of an antibody to a tissue specific antigen.

The terms "antibody" and "immunoglobulin" are used herein interchangeably. The term antibody includes antigen binding fragments such as the Fab, Fab', F(ab')2 fragments (one light chain and half a heavy chain), scFv (two variable domains, one from a light chain VL and one from a heavy chain VH) and single-domain antibodies (one single variable domain, VH or VL). The antibody may be monoclonal or polyclonal. The monoclonal antibodies may be monospecific or bispecific. The antibody may be humanised in which the sequence of an antibody is modified to increase the similarity to antibody variants produced naturally in humans. The antibody may be a chimeric antibody in which one or more regions of an antibody are substituted for those from another species, for example the Fc region in a murine antibody may be substituted for the Fc region from a human antibody. The term "isolated" in this text means isolated from its natural environment. The term "protein" in this text means, in general terms, a plurality of amino acid residues joined together by peptide bonds. It is used interchangeably and means the same as peptide, oligopeptide, oligomer or polypeptide, and includes glycoproteins and derivatives thereof. The term "protein" is also intended to include fragments, analogues and derivatives of a protein wherein the fragment, analogue or derivative retains essentially the same biological activity or function as a reference protein. The protein may be synthetic or otherwise isolated from a natural source. The protein may be produced recombinantly or chemically synthesized. A fragment, analogue or derivative of the specific binding molecule as defined in this text, may be a protein sequence of at least 6, preferably 10 or 20, or up to 50 or 100 amino acids long.

The fragment, derivative or analogue of the protein may be (i) one in which one or more of the amino acid residues are substituted with a conserved or non-conserved amino acid residue (preferably, a conserved amino acid residue) and such substituted amino acid residue may or may not be one encoded by the genetic code, or (ii) one in which one or more of the amino acid residues includes a substituent group, or (iii) one in which the mature polypeptide is fused with another compound, such as a compound to increase the half-life of the polypeptide (for example, polyethylene glycol), or (iv) one in which the additional amino acids are fused to the specific binding molecule, such as a leader or secretory sequence which is employed for purification of the specific binding molecule. Such fragments, derivatives and analogues are deemed to be within the scope of those skilled in the art from the teachings herein.

Particularly suitable are variants, analogues, derivatives and fragments having the amino acid sequence of the protein in which several e.g. 5 to 10, or 1 to 5, or 1 to 3, 2, 1 or no amino acid residues are substituted, deleted or added in any combination. Especially suitable among these are silent substitutions, additions and deletions, which do not alter the properties and activities of the specific binding molecule of the present invention. Also especially suitable in this regard are conservative substitutions.

An example of a variant of the present invention is a specific binding molecule as defined above, apart from the substitution of one or more amino acids with one or more other amino acids. The skilled person is aware that various amino acids have similar properties. One or more such amino acids of a substance can often be substituted by one or more other such amino acids without eliminating a desired activity of that substance.

Thus the amino acids glycine, alanine, valine, leucine and isoleucine can often be substituted for one another (amino acids having aliphatic side chains). Of these possible substitutions it is preferred that glycine and alanine are used to substitute for one another (since they have relatively short side chains) and that valine, leucine and isoleucine are used to substitute for one another (since they have larger aliphatic side chains which are hydrophobic). Other amino acids which can often be substituted for one another include: phenylalanine, tyrosine and tryptophan (amino acids having aromatic side chains); lysine, arginine and histidine (amino acids having basic side chains); aspartate and glutamate (amino acids having acidic side chains); asparagine and glutamine (amino acids having amide side chains); and cysteine and methionine (amino acids having sulphur containing side chains). Substitutions of this nature are often referred to as "conservative" or "semi- conservative" amino acid substitutions.

Amino acid deletions or insertions may also be made relative to the amino acid sequence for the specific binding molecule referred to above. Thus, for example, amino acids which do not have a substantial effect on the activity of the specific binding molecule, or at least which do not eliminate such activity, may be deleted. Such deletions can be advantageous since the overall length and the molecular weight of a polypeptide can be reduced whilst still retaining activity. This can enable the amount of polypeptide required for a particular purpose to be reduced - for example, dosage levels can be reduced.

Amino acid insertions relative to the sequence of the specific binding molecule above can also be made. This may be done to alter the properties of a substance of the present invention (e.g. to assist in identification, purification or expression, as explained above in relation to fusion proteins). Amino acid changes relative to the sequence for the specific binding molecule of the invention can be made using any suitable technique e.g. by using site-directed mutagenesis. It should be appreciated that amino acid substitutions or insertions within the scope of the present invention can be made using naturally occurring or non-naturally occurring amino acids. Whether or not natural or synthetic amino acids are used, it is preferred that only L- amino acids are present. A specific binding molecule according to the invention may have additional N-terminal and/or C-terminal amino acid sequences. Such sequences can be provided for various reasons, for example, glycosylation.

The term "fusion protein" in this text means, in general terms, one or more proteins joined together by chemical means, including hydrogen bonds or salt bridges, or by peptide bonds through protein synthesis or both.

The term "identity" as known in the art is the relationship between two or more polypeptide sequences or two or more polynucleotide sequences, as determined by comparing the sequences. In the art, identity also means the degree of sequence relatedness (homology) between polypeptide or polynucleotide sequences, as the case may be, as determined by the match between strings of such sequences. While there exist a number of methods to measure identity between two polypeptide or two polynucleotide sequences, methods commonly employed to determine identity are codified in computer programs. Preferred computer programs to determine identity between two sequences include, but are not limited to, GCG program package (Devereux, et al., Nucleic acids Research, 12, 387 (1984), BLASTP, BLASTN, and FASTA (Atschul et al., J. Molec. Biol. 215, 403 (1990).

The invention further provides nucleic acid encoding the specific binding molecule of the first aspect of the invention as defined above. A second aspect of the invention provides a nucleic acid construct comprising a first nucleic acid sequence encoding a specific binding molecule as defined above.

The terms "nucleic acid construct" and "nucleic acid sequence" generally refer to any length of nucleic acid which may be DNA, cDNA or RNA such as mRNA obtained by cloning or produced by chemical synthesis. The DNA may be single or double stranded. Single stranded DNA may be the coding sense strand, or it may be the non-coding or anti-sense strand. For therapeutic use, the nucleic acid construct or sequence is preferably in a form capable of being expressed in the subject to be treated. The nucleic acid sequence may be synthetic or otherwise isolated from a natural source. The nucleic acid sequence may be produced recombinantly or chemically synthesized.

The nucleic acid sequence of the second aspect of the invention may be in the form of a vector, for example, an expression vector, and may include, among others, chromosomal, episomal and virus- derived vectors, for example, vectors derived from bacterial plasmids, from bacteriophage, from transposons, from yeast episomes, from insertion elements, from yeast chromosomal elements, from viruses such as baculo-viruses, papova-viruses, such as SV40, vaccinia viruses, adenoviruses, fowl pox viruses, pseudorabies viruses and retroviruses, and vectors derived from combinations thereof, such as those derived from plasmid and bacteriophage genetic elements, such as cosmids and phagemids. Generally, any vector suitable to maintain, propagate or express nucleic acid to express a polypeptide in a host, may be used for expression in this regard. The vector may comprise a plurality of the nucleic acid constructs defined above, for example 2 or more.

The nucleic acid sequence of the second aspect of the invention preferably includes a promoter or other regulatory sequence which controls expression of the nucleic acid. Promoters and other regulatory sequences which control expression of a nucleic acid have been identified and are known in the art. The person skilled in the art will note that it may not be necessary to utilise the whole promoter or other regulatory sequence. Only the minimum essential regulatory element may be required and, in fact, such elements can be used to construct chimeric sequences or other promoters. The essential requirement is, of course, to retain the tissue and/or temporal specificity. The promoter may be any suitable known promoter, for example, the human cytomegalovirus (CMV) promoter, the CMV immediate early promoter, the HSV thymidine kinase, the early and late SV40 promoters or the promoters of retroviral LTRs, such as those of the Rous Sarcoma virus (RSV) and metallothionine promoters such as the mouse metallothionine-l promoter. The promoter may comprise the minimum comprised for promoter activity (such as a TATA element without enhancer elements) for example, the minimum sequence of the CMV promoter. Preferably, the promoter is contiguous to the first and/or second nucleic acid sequence.

As stated herein, the nucleic acid sequence of the second aspect of the invention may be in the form of a vector. Vectors frequently include one or more expression markers which enable selection of cells transfected (or transformed) with them, and preferably, to enable a selection of cells containing vectors incorporating heterologous DNA. A suitable start and stop signal will generally be present. One embodiment of the invention relates to a cell comprising the nucleic acid construct of the second aspect of the invention. The cell may be termed a "host" cell, which is useful for the manipulation of the nucleic acid, including cloning. Alternatively, the cell may be a cell in which to obtain expression of the nucleic acid. Representative examples of appropriate host cells for expression of the nucleic acid construct of the invention include virus packaging cells which allow encapsulation of the nucleic acid into a viral vector; bacterial cells, such as Streptococci, Staphylococci, E.coli, Streptomyces and Bacillus subtilis; single cells, such as yeast cells, for example, Saccharomyces cerevisiae, and Aspergillus cells; insect cells such as Drosophila S2 and Spodoptera Sf9 cells, animal cells such as CHO, COS, C127, 3T3, HEK.293, and Bowes Melanoma cells and other suitable human cells; and plant cells e.g. Arabidopsis thaliana.

Introduction of an expression vector into the host cell can be affected by calcium phosphate transfection, DEAE-dextran mediated transfection, microinjection, cationic - lipid-mediated transfection, electroporation, transduction, scrape loading, ballistic introduction, infection or other methods. Such methods are described in many standard laboratory manuals, such as Sambrook et al, Molecular Cloning, a Laboratory Manual, Second Edition, Coldspring Harbor Laboratory Press, Coldspring Harbor, N.Y. (1989).

Mature proteins can be expressed in host cells, including mammalian cells such as CHO cells, yeast, bacteria, or other cells under the control of appropriate promoters. Cell-free translation systems can be employed to produce such proteins using RNAs derived from the nucleic acid construct of the second aspect of the present invention. Appropriate cloning and expression vectors for use with prokaryotic and eukaryotic hosts are described by Sambrook et al, Molecular Cloning, a Laboratory Manual, Second Edition, Coldspring Harbor Laboratory Press, Coldspring Harbor, N.Y. (1989).

Proteins can be recovered and purified from recombinant cell cultures by well-known methods including ammonium sulphate or ethanol precipitation, acid extraction, anion or cation exchange chromatography, phosphocellulose chromatography, hydrophobic interaction chromatography, affinity chromatography, hydroxylapatite chromatography, high performance liquid chromatography, lectin and/or heparin chromatography. For therapy, the nucleic acid construct e.g. in the form of a recombinant vector, may be purified by techniques known in the art, such as by means of column chromatography as described in Sambrook et al, Molecular Cloning, a Laboratory Manual, Second Edition, Coldspring Harbor Laboratory Press, Coldspring Harbor, N.Y. (1989). According to a third aspect of the invention, there is provided a composition comprising the specific binding molecule or nucleic acid construct of the first or second aspects of the invention. Therefore, the specific binding molecules or nucleic acid constructs of the present invention may be employed in combination with the pharmaceutically acceptable carrier or carriers. Such carriers may include, but are not limited to, saline, buffered saline, dextrose, liposomes, water, glycerol, ethanol and combinations thereof. The composition may be provided in the form of a pharmaceutical formulation.

The pharmaceutical compositions may be administered in any effective, convenient manner effective for treating a patient's disease including, for instance, administration by oral, topical, intravenous, intramuscular, intranasal, or intradermal routes among others. In therapy or as a prophylactic, the active agent may be administered to an individual as an injectable composition, for example as a sterile aqueous dispersion, preferably isotonic.

For administration to mammals, and particularly humans, it is expected that the daily dosage of the active agent will be from 0.01 mg/kg up to 10mg/kg body weight, typically around 1 mg/kg. The physician in any event will determine the actual dosage which will be most suitable for an individual which will be dependent on factors including the age, weight, sex and response of the individual. The above dosages are exemplary of the average case. There can, of course, be instances where higher or lower dosages are merited, and such are within the scope of this invention

References to uses of the specific binding molecules, nucleic acid constructs, vectors, or host cells of the present invention in the treatment of diseases, such as inflammatory diseases or cancer, includes embodiments relating to the use of the specific binding molecules, nucleic acid construct, vector, or host cell in the manufacture of a medicament for the treatment of said diseases.

According to a fourth aspect of the invention, there is provided a specific binding molecule in accordance with the first aspect of the invention for use in the treatment of inflammatory conditions or cancer. This aspect of the invention therefore extends to and includes a method for the treatment of inflammatory conditions or cancer comprising the administration to a subject of a composition comprising a specific binding molecule as defined above.

The present invention provides a composition as described above for use in the treatment of inflammatory conditions or cancer. Inflammatory conditions include, without limitation, atherosclerosis, acute and chronic lung inflammation (e.g., chronic bronchitis, asthma, lung infection including bacterial and viral infections such as SARS and influenza, cystic fibrosis, etc.), inflammation of virus-infected tissues (e.g., viral lung infections, viral myocarditis, viral meningitis, etc.), ulcerative colitis, endotoxic shock, arthritis (e.g., rheumatoid arthritis, juvenile arthritis, osteoarthritis, psoriatic arthritis, reactive arthritis, viral or post-viral arthritis, ankylosing spondylarthritis, etc.), psoriasis, Crohn's disease, inflammatory bowel disease, insulin dependent diabetes mellitus, injury independent type II diabetes, ischemia induced inflammation, otitis media (middle ear infection), gout, multiple sclerosis, cachexia, and Ataxia Telangiectasia. Arthritis defines a group of disease conditions (or arthropathies) where damage is caused to the joints of the body and includes osteoarthritis (also known as degenerative joint disease) which can occur following trauma to the joint, following an infection of the joint or as a result of aging. Other forms of arthritis include rheumatoid arthritis and psoriatic arthritis, which are autoimmune diseases, and septic arthritis is caused by infection in the joints. Inflammatory disease conditions also include vasculitis, dermatitis, inflammatory bowel disease, inflammatory neurological conditions). Cancer defines a group of diseases characterized by an abnormal proliferation of cells in the body, which can be defined as tumors, for example glioma. Types of gliomas include ependymomas, astrocytomas, oligodendrogliomas and mixed gliomas. A Grade 4 astrocytoma is also known as a glioblastoma.

As used herein, the term "treatment" includes any regime that can benefit a human or a non-human animal. The treatment of "non-human animals" extends to the treatment of domestic animals, including horses and companion animals (e.g. cats and dogs) and farm/agricultural animals including members of the ovine, caprine, porcine, bovine and equine families. The treatment may be in respect of any existing condition or disorder, or may be prophylactic (preventive treatment). The treatment may be of an inherited or an acquired disease. The treatment may be of an acute or chronic condition.

In a fifth aspect, the invention provides a nucleic acid sequence in accordance with the second aspect of the invention for use in the treatment of inflammatory conditions or cancer as defined above. This aspect therefore extends to and includes a method for the treatment of inflammatory conditions or cancer comprising the administration to a subject a nucleic acid construct of the second aspect of the invention. Where the nucleic acid construct is used in the therapeutic method of the invention, the construct may be used as part of an expression construct, e.g. in the form of an expression vector such as a plasmid or virus. In such a method, the construct may be administered intravenously, intradermal^, intramuscularly, orally or by other routes. The nucleic acid construct of the second aspect of the invention, and proteins derived therefrom, may be employed alone or in conjunction with other compounds, such as therapeutic compounds, e.g. anti-inflammatory drugs, cytotoxic agents, cytostatic agents or antibiotics. The nucleic acid constructs and proteins useful in the present invention are preferably provided in an isolated form, and preferably are purified to homogeneity. The nucleic acid construct of the second aspect of the invention may be used therapeutically in a method of the invention by way of gene therapy. Alternatively, protein encoded by the nucleic acid construct may be directly administered as described herein. Administration of the nucleic acid construct of the second aspect may be directed to the target site by physical methods. Examples of these include topical administration of the "naked" nucleic acid in the form of a vector in an appropriate vehicle, for example, in solution in a pharmaceutically acceptable excipient, such as phosphate buffered saline, or administration of a vector by physical method such as particle bombardment according to methods known in the art. Other physical methods for administering the nucleic acid construct or proteins of the third aspect of the invention directly to the recipient include ultrasound, electrical stimulation, electroporation and microseeding. Further methods of administration include oral administration or administration through inhalation. Particularly preferred is the microseeding mode of delivery which is a system for delivering genetic material into cells in situ in a patient as described in US 5697901 . The nucleic acid construct according to the second aspect of the invention may also be administered by means of delivery vectors. These include viral delivery vectors, such as adenovirus, retrovirus or lentivirus delivery vectors known in the art. Other non-viral delivery vectors include lipid delivery vectors, including liposome delivery vectors known in the art. Administration may also take place via transformed host cells. Such cells include cells harvested from the subject, into which the nucleic acid construct is transferred by gene transfer methods known in the art. Followed by the growth of the transformed cells in culture and grafting to the subject. As used herein the term "gene therapy" refers to the introduction of genes by recombinant genetic engineering of body cells (somatic gene therapy) for the benefit of the patient. Furthermore, gene therapy can be divided into ex vivo and in vivo techniques. Ex vivo gene therapy relates to the removal of body cells from a patient, treatment of the removed cells with a vector i.e., a recombinant vector, and subsequent return of the treated cells to the patient. In vivo gene therapy relates to the direct administration of the recombinant gene vector by, for example, intravenous or intravascular means. Preferably the method of gene therapy of the present invention is carried out ex vivo. Preferably in gene therapy, the expression vector of the present invention is administered such that it is expressed in the subject to be treated. Thus for human gene therapy, the promoter is preferably a human promoter from a human gene, or from a gene which is typically expressed in humans, such as the promoter from human CMV. For gene therapy, the present invention may provide a method for manipulating the somatic cells of human and non-human mammals, and also a method which may involve the manipulation of the germ line cells of a non-human mammal.

The present invention therefore provides a method for providing a human with a specific binding molecule comprising introducing mammalian cells into a human, the human cells having been treated in vitro to insert therein a nucleic acid construct according to the second aspect of the invention.

Each of the individual steps of the ex vivo somatic gene therapy method are also covered by the present invention. For example, the step of manipulating the cells removed from a patient with the nucleic acid construct of the second aspect of the invention in an appropriate vector. As used herein, the term "manipulated cells" covers cells transfected with a recombinant vector. Also contemplated is the use of the transfected cells in the manufacture of a medicament for the treatment of inflammatory conditions, such as arthritis or cancer, as defined herein above.

The present invention may also find application in veterinary medicine for treatment/prophylaxis of domestic animals including horses and companion animals (e.g. cats and dogs) and farm animals which may include mammals of the ovine, porcine, caprine, bovine and equine families. In a sixth aspect of the invention, there is provided is a kit of parts comprising a specific binding molecule of the first aspect of the invention, a nucleic acid construct of the second aspect of the invention, and an administration vehicle including, but not limited to, tablets for oral administration, inhalers for lung administration and injectable solutions for intravenous administration. In a seventh aspect of the invention, there is provided a process for preparing the specific binding molecule of the first aspect of the invention comprising production of the specific binding molecule recombinantly by expression of a suitable nucleic acid sequence encoding said specific binding molecule in a host cell, followed by purification of the expressed specific binding molecule. In an eighth aspect of the invention, there is provided a process for preparing a nucleic acid construct of the second aspect of the invention, comprising ligating together nucleic acid sequences encoding a specific binding molecule as defined above.

In one embodiment, there is provided a specific binding molecule comprising the immunoglobulin light chain (VL) and immunoglobulin heavy chain (VH) domains of an anti-ICAM1 in which each domain is conjugated separately via a peptide linker comprising a proteolytic cleavage site to the immunoglobulin light chain (VL) and immunoglobulin heavy chain (VH) respectively of an anti-TNFa antibody. The proteolytic cleavage site may be from a metalloproteinase such as MMP-1 , -3, -9, or -13. One suitable cleavage sequence is PLGLWA. The specific binding molecule may be of particular use in the treatment of an inflammatory disease, such as rheumatoid arthritis (RA).

The present invention therefore provides a biological pro-drug with enhanced specificity for sites of disease, for example inflammatory sites (synovium) and reduced specificity to off-target effects, for example with respect to binding / neutralising cytokines in the systemic circulation or sites of physiologic immune response, for example TNF-a. This construct has the potential to form a platform technology that is capable of enhancing the therapeutic index of drugs for the treatment of diseases, for example inflammatory diseases, including rheumatoid arthritis.

The present inventors have also shown that by modulating the linker length it is be possible, while maintaining the specificity of the outer domain to the target of interest, to reduce the affinity of binding of the inner domain to TNF-a in its circulating unbound form. This coupled with further engineering of the linker to contain an MMP cleavable sequence allows a fully functional antibody to be released and act above all locally at the site of inflammation. Such a molecule that is provided by the present disclosure (identified as activatable DVD (aDVD)) has benefits in reducing systemic toxicity and increasing the therapeutic dosage at disease sites improving the therapeutic index.

The present inventors have shown that by reducing the linker length it is possible to selectively impair antigen accessibility to the internal domain in a reversible manner, via the presence of an MMP cleavable site within the linker. By placing the therapeutic moiety on the inner region three important effects were obtained: 1) binding capacity skewed towards tissue targeting provided by the outer variable domain, 2) inhibition of systemic engagement of the inner therapeutic binding region prior to target tissue localisation and enzymatic cleavage and 3) selective activation of the therapeutic antibody on the site of local inflammation. This construct therefore provides a tissue- specific delivery of an antibody pro-drug.

The present inventors have developed an activatable DVD-like construct, aDVD, with an anti- ICAM1 outer domain, for inflamed synovium tissue targeting, linked to the anti-TNF-a antibody adalimumab. The linkers designed which contained an MMP cleavable sequence, were readily cleaved by the proteolytic MMP-1 enzyme both in recombinant form and in human RA synovial fluid (SF), providing insight for efficient in vivo antibody activation. Although MMP overexpression is generally associated with inflammation, angiogenesis and wound repair, different tissues/conditions are characterised by increased expression of specific MMP subgroups. In the context of inflammation, elevated levels of MMP-2, -7, -8 and -9 have been reported in experimental autoimmune encephalomyelitis (34-36), while MMP-3 and -9 have been associated with cutaneous inflammation (37). In inflammatory arthritis, overexpression of MMP-1 , -3, -9 and -13 has been correlated with disease progression and joint damage (26, 38, 39). MMPs levels in RA SF and serum where found to be significantly higher than in healthy controls, with SF levels several hundred-fold higher than serum (39, 40). As synoviocytes of the lining represent the predominant source of MMPs in the arthritic synovium (41), the synovial tissue can be expected to display a greater concentration of MMP-1 and -3 than the associated SF and peripheral blood. The slower antibody cleavage rate observed when using RA SF, and the lack of activation in RA sera, may result in an advantage in vivo, where only the antibody actively accumulated in the arthritic synovial tissue may be held long enough for an efficient MMP cleavage. Further, in the context of infections such as Tuberculosis (a major risk for anti-TNFa treatment), the secreted levels of key MMPs, e.g. MMP-1 are detected up to 100-fold lower compared to RA SF) (42, 43). This may result in an increased safety due to reduced risk of unwanted antibody activation in other tissues and/or in the presence of concomitant infections.

TNF-a binding characterisation in ELISA and inhibition of TNF-a biological activity in L-929 assays, proved that short linkers could substantially restrict antigen accessibility and binding capacity. The binding / neutralisation capacity, however, could be fully resolved upon MMP cleavage. Among the three MMP containing linkers, the short PLGLWA linker showed the highest binding inhibition, reasonably due to the vicinity of the outer domain and increased steric hindrance. Indeed, SPR binding kinetic measurements proved an impaired antigen access to the inner domain, resulting in a slower K_a. Once bound, however, K_d kinetics for TNF-a were unchanged, suggesting unaltered functionalities for adalimumab in aDVD format. This was further confirmed by the identical kinetics of the cleaved aDVD and the parent adalimumab antibody. This effect was more pronounced using variable regions with smaller binding interface (e.g. infliximab). With regard to RA synovium, peptides and scFv antibody fragments with selective synovium specificity have been identified previously (45-47). The incorporation of the scFv A7 in the aDVD format as demonstrated in the present invention for ICAM1 , will provide human arthritic synovium specificity to the molecule, with great potential for tissue-specific drug delivery in RA therapy. These constructs have the potential to increase safety, due to selective activation in situ and inhibited systemic cytokine engagement, and potency, due to enhanced therapeutically relevant concentrations achieved at the site of active disease. It is envisaged that the development of tissue-targeting pro-drugs such as the described aDVD could represent optimal flexible platforms for rational design of therapeutics with significant impact on the treatment landscape for RA.

Preferred features of the second and subsequent aspects of the invention are as for the first aspect mutatis mutandis.

The invention will now be further described by way of reference to the following Examples which are presented for the purposes of illustration only and are not to be construed as being limitations on the invention. Reference is also made to the following figures in which:

FIGURE 1

Figure 1 shows the structure and characterisation of aDVD antibodies. Schematic of the general structure for DVD constructs with anti-ICAM1 outer domain linked to the anti-TNF-a antibody adalimumab (a). VH, CH, VL and CL refer to variable heavy, constant heavy, variable light and constant light chain regions, respectively. Linker length and amino acidic composition is summarised in the table with the MMP cleavable sequence highlighted in bold and cutting position marked with a slash (a). Time course of aDVD antibodies cleavage with recombinant MMP resolved in SDS-PAGE. The gel shows a gradual conversion from DVD heavy chain (HC) to IgG HC, due to cleavage and removal of the outer anti-ICAM1 variable region (b). Similar processing was detected on the light chain (data not shown). Western blot analysis under reducing conditions shows time dependent increase of IgG HC content due to cleavage of the biotinylated aDVD antibody carrying the short MMP cleavable linker, following incubation with RA synovial fluid at 37°C, while no cleavage was detected at 72h for the antibody carrying the scrambled MMP linker (c). Further, No cleavage was detected upon incubation with sera form RA patients or in the presence of the MMP inhibitor GM6001 (c). Cutting kinetics of biotinylated aDVD with recombinant MMP-1 in Figure 7. FIGURE 2

Figure 2 shows anti-TNF-a activity of DVD antibodies. TNF-a binding capacity for DVD antibodies (Scrambled MMP and Long linker) and aDVD antibodies (MMP, G₄S-MMP-G₄S and (G₄S)₂-MMP- (G₄S)₂) was evaluated in ELISA (a). Reduced binding capacity was detected for uncut aDVD constructs while full potency could be restored upon cleavage with MMP enzyme, as compared with adalimumab IgG. Neutralisation of TNF-a induced cytotoxicity in L-929 cell line was impaired in uncut aDVD antibodies, with stronger impairment for ICAM-MMP-adalimumab antibody (b). Similar potencies to adalimumab IgG were obtained following MMP enzymatic digestion. Results are expressed as effective concentration 20 (EC20) for (a) and inhibitory concentration 20 (IC20) for (b) corresponding at the dose necessary to obtain 20% of activity.

FIGURE 3

Figure 3 shows analysis of aDVD antigen binding kinetics. SPR sensorgrams with binding kinetics for TNF-a of adalimumab IgG, aDVD and aDVD cleaved with recombinant MMP enzyme (a). The reduced binding capacity for TNF-a of aDVD can be reverted following digestion with MMP enzyme, restoring full binding potential compared to adalimumab IgG (for kinetic measurements see Table 1). TNF-a concentrations: 20 nM (red), 8 nM (yellow), 3.2 nM (green), 1 .28 nM (blue) and 0.512 nM (purple). Dynamic binding kinetics for TNF-a and ICAM1 are shown in panel (b). When the aDVD antibody had been saturated with TNF-a, with limited binding capacity, the second antigen was injected showing retention of ICAM1 specificity in the presence of TNF-a. On-chip digestion of the construct with recombinant MMP enzyme was sufficient to cleave the antibody, releasing the outer domain and the coupled ICAM1 antigen. Finally, injection of TNF-a highlights the restored binding potency of the internal anti-TNF-a domain.

FIGURE 4

Figure 4 shows human synovial tissue reactivity of aDVD antibodies. The reactivity of ICAM1 - MMP-adalimumab with RA, OA and non-arthritic human synovial tissues was examined using immunohistochemistry. Bound biotinylated aDVD antibodies were detected using streptavidin- horseradish peroxidase (HRP) complex and compared to staining pattern of anti-ICAM-1 IgG. The presence of blood vessels was depicted using anti-vWF antibody in combination with anti-CD31 antibody. Scale bar = 100 μηι. It can be seen that the strongest binding of aDVD ICAM1 -MMP- adalimumab and anti-ICAM1 is observed in the RA synovium, followed by OA synovium but no/neglible binding to normal synovium, while control antibodies (anti-vWF and anti-CD31) bind equally effectively to all synovial tissues. FIGURE 5

Figure 5 shows molecular interactions determine aDVD inhibitory properties. Schematic of infliximab and adalimumab interaction with TNF-a reveals a reduced contact area (red residues) for infliximab compared to adalimumab which may predict an increased binding inhibition in aDVD format (a) (adapted from Hu et al. 2013 (31)). TNF-a binding capacity for aDVD ICAM-MMP- Infliximab pre and post cleavage were compared to Infliximab in ELISA assay, a 2500 fold inhibition for the infliximab aDVD construct compared to Infliximab was observed, the binding was rescued upon cleavage of the construct with MMP (b). A greater blocking capacity for TNF binding was demonstrated with the infliximab aDVD construct compared to the adalimumab containing construct (ICAM-MMP-Adalimumab ELISA assay was adapted from Figure 2). TNF-a neutralization in L-929 functional assay showed complete loss of function for ICAM-MMP-lnfliximab compared to Infliximab, demonstrating greater inhibition compared to the same construct containing the adalimumab anti-TNF domain (ICAM-MMP-Adalimumab L-929 assay was adapted from figure 2) (c). Kinetics of binding to TNF-a of aDVD ICAM-MMP-lnfliximab and Infliximab were compared. Data demonstrates binding of 20 nM TNF to Infliximab and ICAM-MMP-lnfliximab coupled to sensor surface at the same density (d). Comparison of ICAM-MMP-Adalimumab and ICAM-MMP-lnfliximab kinetic measures can be found in Table 1 .

FIGURE 6

Figure 6 shows anti ICAM-1 activity of DVD antibody. ICAM1 binding capacity for ICAM-MMP- Adalimumab aDVD antibody was evaluated in ELISA. Unprocessed aDVD antibody showed similar EC50 compared to parent anti ICAM1 IgG antibody. Binding capacity to ICAM1 was lost upon MMP cleavage of the aDVD construct, due to removal of the outer variable region.

FIGURE 7

Figure 7 shows MMP cleavage activity on biotinylated aDVD. Time course cleavage of biotinylated aDVD antibodies with recombinant MMP resolved in SDS-PAGE. The gel shows a gradual conversion from DVD heavy chain (HC) to IgG HC, due to cleavage and removal of the outer anti- ICAMI variable region in the aDVD antibody carrying the short MMP linker. Cutting kinetics was comparable to the aDVD antibodies in Figure 1 b. No cleavage was detected for the scrambled MMP linker.

FIGURE 8

Figure 8 shows the anti-mouse (m)ICAMI -MMP-Adalimumab antibody. (a) Schematic representation of mlCAM1 -MMP-Adalimumab antibody. (b) Digestion of mlCAM1 -MMP- Adalimumab antibody with recombinant MMP1 enzyme, (c) ELISA assay on human TNFa with mlCAM1 -MMP-Adalimumab antibody, (d) L-929 TNF induced cytotoxicity assay with mlCAMI - MMP-Adalimumab antibody.

FIGURE 9

Figure 9 shows schematic representations of antibodies developed according to the invention. FIGURE 10

Figure 10 shows synovium staining with A7-MMP-Adalimumab antibody. It can be seen that the strongest binding of A7-ICAM1 -MMP-Adalimumab is observed in the RA synovium, followed by OA synovium but no/negligible binding to normal synovium, while control antibodies (anti-vWF/anti- CD31) bind equally effectively to all synovial tissues.

FIGURE 11

Figure 1 1 shows SPR kinetic measurements for the anti-mouse antibodies described in the invention.

FIGURE 12

Figure 12 shows list of sequences for alternative cleavage sites for MMP proteases. FIGURE 13

Figure 13 shows the amino acid and nucleotide sequences for activatable DVD antibody A7 anti- Human TNF (Adalimumab) with PLGLWA linker (SEQ ID NO: 1) (a) VH amino acid sequence, (b) VL amino acid sequence, (c) VH nucleotide sequence and (d) VL nucleotide sequence. FIGURE 14

Figure 14 shows the amino acid and nucleotide sequences for activatable DVD antibody A7 anti- Human TNF (Adalimumab) with GGGGSPLGLWAGGGGS linker (SEQ ID NO: 12) (a) VH amino acid sequence, (b) VL amino acid sequence, (c) VH nucleotide sequence and (d) VL nucleotide sequence.

FIGURE 15

Figure 15 shows the amino acid and nucleotide sequences for activatable DVD antibody A7 anti- Human TNF (Adalimumab) with GGGGSGGGGSPLGLWAGGGGSGGGGS linker (SEQ ID NO: 13) (a) VH amino acid sequence, (b) VL amino acid sequence, (c) VH nucleotide sequence and (d) VL nucleotide sequence.

FIGURE 16

Figure 16 shows the amino acid and nucleotide sequences for activatable DVD antibody A7 anti- Human TNF (Infliximab) with PLGLWA linker (SEQ ID NO: 1) (a) VH amino acid sequence, (b) VL amino acid sequence, (c) VH nucleotide sequence and (d) VL nucleotide sequence. FIGURE 17

Figure 17 shows the amino acid and nucleotide sequences for activatable DVD antibody A7 anti- mouse TNF (TN3.19.12) with PLGLWA linker (SEQ ID NO: 1) (a) VH amino acid sequence, (b) VL amino acid sequence, (c) VH nucleotide sequence and (d) VL nucleotide sequence.

FIGURE 18

Figure 18 shows the amino acid and nucleotide sequences for activatable DVD antibody anti- human ICAM1 (perlan) anti-Human TNF (Adalimumab) with PLGLWA linker (SEQ ID NO: 1) (a) VH amino acid sequence, (b) VL amino acid sequence, (c) VH nucleotide sequence and (d) VL nucleotide sequence.

FIGURE 19

Figure 19 shows the amino acid and nucleotide sequences for activatable DVD antibody anti- human ICAM1 (perlan) anti-Human TNF (Adalimumab) with GGGGSPLGLWAGGGGS linker (SEQ ID NO: 12) (a) VH amino acid sequence, (b) VL amino acid sequence, (c) VH nucleotide sequence and (d) VL nucleotide sequence.

FIGURE 20

Figure 20 shows the amino acid and nucleotide sequences for activatable DVD antibody anti- human ICAM1 (perlan) anti-Human TNF (Adalimumab) with GGGGSGGGGSPLGLWAGGGGSGGGGS linker (SEQ ID NO: 13) (a) VH amino acid sequence, (b) VL amino acid sequence, (c) VH nucleotide sequence and (d) VL nucleotide sequence. FIGURE 21

Figure 21 shows the amino acid and nucleotide sequences for activatable DVD antibody anti- human ICAM1 (perlan) anti-Human TNF (Infliximab) with PLGLWA linker (SEQ ID NO: 1) (a) VH amino acid sequence, (b) VL amino acid sequence, (c) VH nucleotide sequence and (d) VL nucleotide sequence.

FIGURE 22

Figure 22 shows the amino acid and nucleotide sequences for Activatable DVD antibody anti- mlCAMI (YN1/1 .7.4) anti-Human TNF (Adalimumab) with PLGLWA linker (SEQ ID NO: 1) (a) VH amino acid sequence, (b) VL amino acid sequence, (c) VH nucleotide sequence and (d) VL nucleotide sequence.

FIGURE 23

Figure 23 shows the amino acid and nucleotide sequences for Activatable DVD antibody anti- mlCAMI (YN1/1 .7.4) anti-mouse TNF (TN3.19.12) with PLGLWA linker (SEQ ID NO: 1) (a) VH amino acid sequence, (b) VL amino acid sequence, (c) VH nucleotide sequence and (d) VL nucleotide sequence.

FIGURE 24

Figure 24 shows results in a collagen induced arthritis (CIA) mouse model. DBA/1 mice challenged with Type II collagen at day 0 according to standard protocol and boost at day 21 . Mice were treated at day 14 with 50 μg of antibody (or appropriate control) 3 times a week until day 28. Arthritic score was assessed by measure of paw swelling and number of small joints involved. aDVD anti-mouse ICAM1 and anti-mouse TNF showed protection against development of rheumatoid arthritis compared to PBS and negative IgG control treatments.

FIGURE 25

Figure 25 shows results in human transgenic mouse model of arthritis (Tg197). 50 μg of Cy5.5 labelled Adalimumab and equivalent concentration of aDVD anti-mouse ICAM1 -MMP-Adalimumab injected IP in Tg197 mice. Tissue localisation analysed using MS imaging at 0, 1 , 2, 3, 4, 5, 6 and 7 days post-injection. The bispecific aDVD antibody shows increase joint localisation compared to standard Adalimumab anti-TNF antibody.

Examples 1 to 6:

In Examples 1 to 6 methods are described for the design and construction of activatable dual variable domain (aDVD) antibodies to target the intercellular adhesion molecule (ICAM)1 , up- regulated at sites of inflammation, and the anti-TNF-a antibodies (adalimumab and infliximab). These bispecific molecules include one arm targeting the outer domain of ICAM1 and the other the therapeutic domain of anti-TNF-α, both arms were linked to a Matrix Metalloproteinase cleavable linkers. Constructs were tested both for their ability bind and neutralize targets in vitro and ex-vivo.

Example 1 : Cloning and expression of DVD antibodies

Sequences of variable regions of antibody anti human ICAM1 and human TNF-a have been previously described (20, 21). Sequence data management was performed using serial doner 2.6. Variable sequences were generated by gene synthesis (Genscript, New Jersey, USA) and combined into various constructs using overlapping extension PCR (22). PCR products were cloned into the AbVec-hlgG1 and AbVec-hlgK vectors (23) using the restriction sites Agel/Sall and Agel/BsiWI respectively. Clones were sequence verified prior to protein expression.

Vectors encoding the Heavy and Light chains of the DVD antibody were transfected into HEK-293T cells in DMEM medium containing 10% foetal bovine serum, 100 U/ml penicillin, 100 μg/ml streptomycin and 0.5 mg/ml geneticin, 24h before transfection. Transfection was performed with JetPRIME reagent (Polyplus) according to the manufacturer's protocol. The antibodies were purified from the supernatant via affinity chromatography using protein A Sepharose CL-4B (GE Healthcare). DVD antibodies were biotinylated using EZ-Link Sulfo-NHS-SS-Biotinylation kit (Thermo-Fisher Scientific) according to manufacturer's protocol. Example 2: MMP enzymatic digestion

Antibodies were incubated at 100 μg/ml with 35 U of recombinant MMP-1 enzyme (Enzo Life Sciences) in 50 mM Tris, 0.15 M NaCI, 10 mM CaCI₂, 50 mM ZnCI₂, 0.02% Brij35, at 37°C. Antibodies used for kinetic analysis were digested for 1 hour at 37°C. Digestion with RA synovial fluid (SF) and RA serum was performed by incubating 500 ng of biotinylated antibody in 200 μΙ of fluid at 37°C for 24 to 72 hours and in the presence of 20 μΜ MMP inhibitor GM6001 .

Example 3: Protein characterisation

Protein purity and molecular weight were assessed by resolution in reducing SDS-PAGE using Mini-Protean 4-20% TGX gels (Biorad) followed by Sypro® Ruby protein gel stain according to manufacturer's instruction. Western blot analysis of RA SF and serum digested antibodies was performed via nitrocellulose transfer. Biotinylated antibody Heavy and Light chains were detected using streptavidin-horseradish peroxidase (HRP).

Example 4: Quantification of anti-TNF-a activity

Enzyme-linked immunosorbent assay (ELISA) for anti-TNF-a activity was performed in 96-well plates (Thermo-Fisher Scientific) coated with 100 ng/ml of TNF-a in PBS overnight at 4 °C. Plates were blocked with PBS 2% BSA for 2h at room-temperature before incubation with serial dilutions of DVD antibody. Bound antibodies were detected with anti-human IgG HRP conjugated antibody (Jackson Immunotools). Plates were then incubated with TMB substrate (GE Healthcare) and reactions stopped with 1 N H₂S0₄. Optical absorption measured at 450 nm. EC20 was calculated using dose response non-linear fit curve in GraphPad Prism v5.

Inhibition of TNF-a induced cytotoxicity was conducted on L-929 cell line. Briefly, 3x10⁴ cells were seeded in 96-well plates in 100 μΙ of DMEM medium supplemented with 10% FBS, 100 U/ml penicillin, 100 μg/ml streptomycin and 10 μΜ of the MMP inhibitor GM6001 for 18 hours at 37°C. The medium was then replaced with 100 μΙ of complete medium with 1 μg/ml actinomycin D, 0.45 ng/ml TNF-a (Sigma) or TNF-a and the antibody of interest (with 1 :2 serial dilution) for 24 hours at 37°C. 500 μg/ml of thiazolyl blue tetrazolium bromide in PBS (Sigma) was added to the wells and incubated for 3 hours at 37°C. Medium was then removed and cells resuspended in 100 μΙ 90% isopropanol 10% DMSO for 15 minutes. Optical absorption was measured at 595 nm. % viability was calculated [(OD595nm X 100)/OD595nm of sample without TNF-a]. IC20 was determined using dose response non-linear fit curve in GraphPad Prism v5.

Example 5: Surface Plasmon Resonance (SPR) Experiments were performed on Biacore T200 instrument using HBS-P+ as running and dilution buffer (GE Healthcare Bio-Sciences). BIAevaluation software Version 2.0 (GE Healthcare) was used for data processing. For binding kinetics, mouse anti-human IgG (GE Healthcare) was covalently coupled to CM5 Sensor Chip (GE Healthcare). Human antibody or DVD antibody was captured, and various concentrations of interaction partner protein injected over the flow cell at a flow rate of 30 μΙ/min. A double reference subtraction was performed using buffer alone. Kinetic rate constants were obtained by curve fitting according to a 1 :1 Langmuir binding model.

Example 6: Immunohistochemical analysis

Formalin-fixed paraffin embedded tissue sections were dewaxed and antigens retrieved with 10 minutes boiling in citrate buffer pH 6 (Dako). Slides were stained with 10 μg/ml of biotinylated DVD antibodies for 1 hour at RT and visualised with streptavidin-HRP complex using 3,3'- diaminobenzidine chromogen (Dako). Rabbit anti-ICAM1 IgG (Abeam) was detected with goat anti- rabbit IgG HRP conjugated (Jackson Immunotools). Mouse anti-vWF (Dako) and mouse anti-CD31 (R&D) were used to depict human vascular endothelial cells and were detected using goat anti- mouse HRP (SantaCruz Biotech). Sections were counterstained with hematoxylin, mounted with Depex mounting medium (Dako) and acquired with CellSens imaging system (Olympus).

Examples 7 to 10

In Examples 7 to 10 the results of experiments are described as follows. Intact aDVD constructs demonstrated significantly reduced binding and anti TNF-a activity in the pro-drug formulation compared to parent antibodies. Physiological concentrations of MMP enzyme, and human synovial fluid were capable of cleaving the external domain of the antibody revealing a fully active molecule. Activated antibodies retained the same binding and anti-TNF-a inhibitory capacities as parent molecules.

Example 7: Design and cleavage of an activatable dual variable domain antibody

To create a bispecific antibody (BsAb) format with therapeutic activity in RA and targeting capacity for the inflamed synovium, the gold standard for anti-TNF-α biologies adalimumab was coupled with an ICAM1 targeting antibody, using an adaptation of the well-established DVD-lg™ format (17). The construct described contains the anti-ICAM1 VL and VH domains linked to the light chain and heavy chain, respectively, of the anti-TNF-α adalimumab via a small peptide linker, schematised in Figure 1a. To create DVD BsAb with impaired binding capacity for the internal variable domain, a series of linkers with varied length and amino acidic composition were designed to test for the desired activity (Table in Figure 1 a). The Long linker was derived from natural linker found in human IgG antibodies and was previously described in the context of DVD-lg™ format (24). Reducing the linker length can substantially alter the kinetic properties of the internal binding domain (24). Without wishing to be bound by theory, it is hypothesized that short linkers could impair the accessibility for the ligand to the internal domain in such a way that could be reverted upon cleavage of the internal linker, thus forming an activatable DVD pro-drug (aDVD). The remaining four linkers contained an MMP cleavable site (PLGLWA) (25) alone or in the presence of G₄S flanking regions, and a scrambled MMP cleavable sequence (AGPLLW). To test for the ability of the MMP enzyme to access, cleave and activate the internal anti-TNF-a domain, the aDVD constructs where incubated with physiologically relevant concentrations of recombinant MMP enzyme. Reduced SDS-PAGE analysis of the digested aDVD constructs in Figure 1 b, showed a rapid processing of the aDVD carrying the MMP cleavable site, with the formation of molecular weight products coherent with an IgG format.

Incubation of the aDVD construct carrying the short MMP linker (PLGLWA) with SF of RA patients, also showed time dependent activation of the construct, confirming the processing capacity in physiological conditions (Figure 1 c). Activation using SF was less efficient than observed with recombinant protein. This may be due to saturation of MMP activity in ex-vivo assays, which one would not anticipate to occur in vivo during chronic inflammation where MMP expression in the synovial tissue is expected to be higher than in the surrounding SF (26). Additionally, the cleavage could be inhibited by the MMP inhibitor GM6001 while no cleavage could be detected for the aDVD carrying the scrambled MMP linker, further confirming MMP mediated activation of the constructs.

Example 8: aDVD shows impaired binding to TNF-a, which is rescued by MMP cleavage

To be effective as a targeting pro-drug it is important that the aDVD molecules retain binding to their target antigen via the outer binding domain, while the inner domain is shielded. Binding of aDVD molecules to ICAM1 (outer domain) and TNF-a (inner domain) was investigated using ELISA. The uncut aDVD molecules retained binding to ICAM1 to the same extent as the parent anti ICAM1 antibody (Figure 6). However, the molecules before MMP cleavage showed a 275-fold binding reduction to TNF-a compared to adalimumab IgG. Binding to TNF-a was fully rescued for all the constructs following MMP cleavage (Figure 2a). To assess the capability of the aDVD constructs to inhibit ligand binding to its receptor and prevent TNF-a-induced cytotoxicity the L-929 assay was employed. The ability of the uncleaved aDVD construct to block and inhibit TNF-a was severely impaired, consistent with binding data observed by ELISA. The uncleaved aDVD antibodies showed up to 132-fold increase in IC20 compared to adalimumab IgG while cleavage with MMP completely rescued the inhibitory capacity (Figure 2b). As expected, the short MMP cleavable linker (PLGLWA) was characterised by a greater TNF-a binding impairment and was further validated using SPR. A comparison of affinities of restricted and processed forms of the aDVD molecule can be found in Table 1. Binding of TNF-a to the uncleaved molecule was greatly impaired as demonstrated by a 365-fold reduction in K_D. Observation of the kinetics of binding indicated that the difference in affinity was predominantly driven by reduction of the association rate constant (k_a), 189-fold lower than the uncleaved molecule, whereas the dissociation rate constant (k_d) was largely unchanged (Figure 3a, Table 1 ). This result demonstrated that blocking of the external domain mainly acts by inhibiting association through steric hindrance, however once bound the antibody retains similar binding characteristics indicating that the internal domain has not been modified and remains fully functional. Importantly, cleaved aDVD molecules not only showed identical affinity but also component kinetics of binding to the parent adalimumab antibody (Table 1). In both cases binding kinetics were in good agreement with previously reported data (27).

In order to target proteins in a disease setting, the aDVD needs to maintain cleavage capacity in the presence of both targeting and effector antigens as is likely to be the scenario in the cytokine rich environment of the inflamed synovium. This is particularly pertinent as the aDVD is still capable of binding to TNF-a with a slow dissociation rate which could conceivably block the cleavage site by steric hindrance (Figure 3a). In order to observe whether the aDVD molecule could be cleaved and activated in this environment, the molecule was immobilised on an SPR sensor chip and saturating concentrations of TNF-a were injected, followed by ICAM1 , prior to MMP cleavage on the sensor surface (Figure 3b). TNF-a showed the same restricted level of binding as had been previously demonstrated with the uncut material. ICAM1 however was capable of binding to the molecule in the presence of TNF-a as demonstrated by the change in response units (5RU) observed which was of the same magnitude as ICAM1 injected on free antibody at the same concentration (data not shown). In the presence of both saturating concentrations of ICAM-1 and TNF-a, MMP enzyme was injected over a period of 30 minutes. Following on chip cleavage and a period of stabilization to remove unbound material from the chip surface, the chip was re- challenged with TNF-a. Post-cleavage the TNF-a binding capacity was rescued, as demonstrated by the enhanced 5RU, which was measured at the same level as the injected concentration on the unrestricted antibody (Figure 3a, 3b).

Example 9: aDVD antibody platform for tissue specific targeting

One of the key characteristics of the aDVD format is the ability to present the anti-TNF-a therapeutic function in a pro-drug format that can be activated following encounter with MMP enzymes in the site of arthritic inflammation. The presence of the outer variable domain targeting ICAM1 , an integrin overexpressed in inflammatory conditions such as RA (28, 29), would allow preferential accumulation of the antibody in the target tissue, facilitating the encounter with proteolytic enzymes. MMP cleavage causes the removal of the anti-ICAM1 external domain, resulting in loss of ICAM1 specificity (Figure 6). To test the ability of the aDVD to retain tissue targeting capacity when in full conformation, the reactivity with the microvasculature in human synovial samples from RA (n=3), OA (n=3) and non-arthritic (n=1) patients was examined via immunohistochemistry (Figure 4).

The ICAM-MMP-adalimumab aDVD was able to selectively target the human inflamed synovium in both RA and OA patients with similar efficacy when compared to an anti-ICAM1 IgG antibody. Importantly, no detectable reactivity was identified in the synovial sample from a non-arthritic patient. The specificity for arthritic synovium further strengthens the potential of the aDVD for targeted drug delivery in rheumatoid arthritis. Furthermore, the aDVD format may represent a flexible platform for targeted delivery of pro-drugs that can be easily adapted to other cytokines and to other disease conditions with a simple exchange of the outer targeting domain. Example 10: Improving structural design of aDVD molecules

Since the reduced binding of aDVD molecules can be mediated by the blocking of the internal domain, it was then investigated whether further inhibition of binding could be predicted through knowledge of the interaction between TNF-a and the inner domain antibody. The crystal structures of adalimumab and infliximab in complex with TNF-a have recently been solved (30, 31). Crystal structure data showed that adalimumab bound to trimeric TNF-a via a broader binding interface with a total buried surface area of 2540A² (31), while infliximab bound to the TNF-a trimer via a reduced binding interface of 1977A² (Figure 5a). Additionally, adalimumab engages the TNF-a trimer through interactions with two monomers of the trimer, while the binding of infliximab is mediated almost exclusively through the loop region of a single TNF-a monomer. It was therefore predicted that the smaller interaction surface area in the infliximab-TNF-a complex would translate to a binding interface that would be more readily blocked by the outer domain. To test this, an aDVD molecule was engineered with infliximab as the inner binding domain and tested for binding and functionality. Infliximab bound to TNF-a with an EC20 of 0.004 nM, while the aDVD-infliximab bound with an EC20 of 21 .6 nM, once processed by MMP cleavage the antibody demonstrated binding that was comparable to the original infliximab antibody (see Figure 5b). The fold difference between cleaved and uncleaved aDVD-infliximab antibody was 3000, 10-fold higher than the difference measured for the aDVD-adalimumab construct. The ability of the antibody to inhibit TNF- α binding to its receptor in L-929 functional studies was also greatly diminished, as no anti-TNF-a functionality could be detected for the uncleaved aDVD-infliximab over the range of concentrations tested while activity was fully rescued after linker cleavage (Figure 5c). SPR data demonstrated 2500-fold reduction in K_D that was predominantly driven by a reduction in k_a with a 1000-fold reduction in the association constant observed (see Figure 5d and Table 1 ). This data further demonstrated the flexibility through which different binding moieties can be introduced, and highlighted the fact that the molecular interactions between parent antibody and target antigen can be used to design aDVD molecules with a more potent blocking capacity.

Table 1

SPR kinetic measurements of aDVD antibodies

References

1. Helmick et al Arthritis Rheum. 2008;58(1):15-25.

2. Wasserman AM.. American family physician. 201 1 ;84(1 1):1245-52.

3. Manicourt et al Arthritis Rheum. 1995;38(8):1031 -9.

4. Johansson et al J Cell Sci. 2000;1 13 Pt 2:227-35.

5. Weyand et al Nat Rev Rheumatol. 2009;5(10):583-8.

6. Nurmohamed et al Drugs. 2005;65(5):661 -94.

7. Lawrence et al Arthritis Rheum. 2008;58(1):26-35.

8. Smolen et al Ann Rheum Dis. 2014;73(3):492-509.

9. Cairns et al The Ulster medical journal. 2002;71 (2):101 -5.

10. Roach et al Journal of immunology. 2002;168(9):4620-7.

1 1 . Mohan et al Infection and immunity. 2001 ;69(3):1847-55.

12. Bean et al Journal of immunology. 1999;162(6):3504-1 1 .

13. Tubach et al Joint, bone, spine : revue du rhumatisme. 2005;72(6):456-60.

14. Taylor et al Nature reviews Rheumatology. 2009;5(10):578-82.

15. Nestorov I. Seminars in arthritis and rheumatism. 2005;34(5 Suppl1):12-8.

16. Maini et al Arthritis Res. 2002;4 Suppl 2:S22-8.

17. Wu et al Nat Biotechnol. 2007;25(1 1 ):1290-7.

18. Gu et al Methods in enzymology. 2012;502:25-41 .

19. Digiammarino et al MAbs. 201 1 ;3(5):487-94.

20. Luo et al J Immunol Methods. 2003;275(1 -2):31 -40.

21 . Salfeld et al Arthritis Rheum. 1998;41 (suppl 9).

22. Higuchi et al Nucleic acids research. 1988;16(15):7351 -67.

23. Di Niro et al. Journal of immunology. 2010;185(9):5377-83.

24. DiGiammarino et al Methods Mol Biol. 2012;899:145-56.

25. Adams et al Nature Biotechnology. 2003; 21 (45cbb8a9-f268-504c-a369- 4896f3795e44):1314-34.

26. Yoshihara et al Ann Rheum Dis. 2000;59(6):455-61 .

27. Kaymakcalan et al Clinical immunology. 2009;131 (2):308-16.

28. Yang et al J Cell Physiol. 2010;224(2):516-26.

29. Smith et al Rheumatology (Oxford). 2001 ;40(9):965-77.

30. Liang et al Journal of biological chemistry. 2013;288(19):13799-807.

31 . Hu et al Journal of biological chemistry. 2013;288(38):27059-67.

32. Ferrari et al Nat Rev Rheumatol. 2015.

33. Chames et al MAbs. 2009;1 (6):539-47.

34. Kieseier et al Brain. 1998;121 ( Pt 1):159-66.

35. Graesser et al J Neuroimmunol. 2000;109(2):121 -31 .

36. Nygardas et al Clin Exp Immunol. 2002;128(2):245-54.

37. Wang et al Proc Natl Acad Sci U S A. 1999;96(12):6885-9.

38. Maeda et al Ann Rheum Dis. 1995;54(12):970-5. 39. Green et al Rheumatology (Oxford). 2003;42(1):83-8.

40. Tchetverikov et al Ann Rheum Dis. 2004;63(7):881 -3.

41 . Hembry et al Ann Rheum Dis. 1995;54(1):25-32.

42. Hoheisel et al Tuberculosis (Edinb). 2001 ;81 (3):203-9.

43. Elkington et al Eur Respir J. 201 1 ;38(2):456-64.

44. Pitzalis et al Current opinion in rheumatology. 2013;25(3):334-44.

45. Kamperidis et al Arthritis and rheumatism. 201 1 ;63(12):3758-67.

46. Lee et al Arthritis and rheumatism. 2002;46(8):2109-20.

47. Wythe et al Ann Rheum Dis. 2013;72(1):129-35.

Claims

1 . A specific binding molecule comprising an immunoglobulin light chain (VL) domain and an immunoglobulin heavy chain (VH) domain of a first antibody to a tissue specific antigen in which each domain is conjugated separately via a peptide linker comprising a proteolytic cleavage site to an immunoglobulin light chain (VL) domain and an immunoglobulin heavy chain (VH) domain respectively of a second antibody.

2. A specific binding molecule as claimed in claim 1 , in which the first antibody to a tissue specific antigen binds to an antigen undergoing overexpression or reactivation in a disease condition

3. A specific binding molecule as claimed in claim 1 or claim 2, in which the first antibody to a tissue specific antigen is an antibody to an antigen expressed in tissue in a subject with an inflammatory disease, for example Rheumatoid Arthritis.

4. A specific binding molecule as claimed in claim 3, in which the first antibody to a tissue specific antigen is an antibody selected from the group consisting of antibodies to a Cell-adhesion molecule (CAM) (for example ICAM1 , ICAM3, VCAM1 , or EpCAM), or the Antigen recognised by the A7 synovium targeting antibody, Fibronectin, Extra domain A of Fibronectin, Tenascin C, an integrin (for example aVp3, aVpl), LFA-1 , Annexin A1 , Nucleotin, Tie-1 , Tie-2, Aminopeptidase N, CD13, CD44 and spliced variants (e.g. CD44v4, CD44v6), CD90, CD55, Folate receptor, Collagen type II and modifications thereof occurring during inflammation, Citrullinated proteins (e.g. citrullinated Fibrinogen, citrullinated Vimentin), Vascular endothelial growth factor receptor 1 (VEGFR-1), or Vascular endothelial growth factor receptor 2 (VEGFR-2).

5. A specific binding molecule as claimed in claim 1 or claim 2, in which the first antibody to a tissue specific antigen is an antibody to an antigen expressed in tissue in a subject with cancer.

6. A specific binding molecule as claimed in claim 5, in which the first antibody to a tissue specific antigen is an antibody to ICAM1 , VCAM1 , EpCAM, Extra domain B of Fibronectin, Melanoma-associated Chondroitin sulfate proteoglycan (MCSP), Melanoma-associated proteoglycan (MAPG), High molecular weight melanoma associated antigen (HMV-MAA), Prostate- specific membrane antigen (PSMA), Epidermal Growth factor Receptor (EGFR), Hepatocyte growth factor receptor (HGFR), Fibroblast activation protein (FAP), Carcinoembryonic Antigen (CEA), Cell-adhesion molecule (CAM), Human B-cell maturation target (BCMA), Placental growth factor (PLGF), Folate receptor, Insulin-like growth factor receptor (ILGFR), CD133, CD40, CD37, CD33, CD30, CD28, CD24, CD23, CD22, CD21 , CD20, CD19, CD13, CD10, HER3, HER2, Non- muscle myosin heavy chain type A (nmMHCA), Transferrin, Epithelial cell adhesion molecule (EpCAM), Annexin A1 , Nucleotin, Tenascin, Vascular endothelial growth factor receptor 1 (VEGFR- 1), Vascular endothelial growth factor receptor 2, (VEGFR-2), Aminopeptidase N, Tie-1 , Tie-2, or c- Met.

7. A specific binding molecule as claimed in claim 1 or claim 2, in which the first antibody to a tissue specific antigen is an antibody selected from the group consisting of antibodies to PNAd, α4β7, MAdCAM-1 .

8. A specific binding molecule as claimed in claim 1 , in which the second antibody is an anti- cytokine antibody, an anti-chemokine antibody, or an anti-trophic factor antibody.

9. A specific binding molecule as claimed in claim 8, in which the anti-cytokine antibody is an anti-TNF antibody, an anti-interleukin antibody or an anti-interferon antibody.

10. A specific binding molecule as claimed in claim 9, in which the anti-cytokine antibody is an antibody to TNF-a, lymphotoxin-a, lymphotoxin-β, CD27L, CD30L, FASL, 4-1 BBL, OX40L, TRAIL

IL1 , IL-2, IL-3, IL-4, IL-5, IL-6, IL-7, IL-8, IL-9, IL-10, IL-1 1 , IL-12, IL-13, IL-14, IL-15, IL-16, IL-17, IL-18, IL-19, IL-20, IL-21 , IL-22, IL-23, IL-24, IL-25, IL-26, IL-27, IL-28, IL-29, IL-30, IL-31 , IL-32, IL- 33 (either -a or -β), IFN- , IFN-β and IFN-γ), IFN-γ inducing factor (IGIF), or to a Bone morphogenetic protein (BMP),

1 1 . A specific binding molecule as claimed in claim 8, in which the anti-chemokine antibody is an antibody to MIP-1A, MIP-1 B, MCP1 , MCP2, MCP3, RANTES, CCL19, CCL21 , IP-10, GROB, Eotaxin TARC, CD20, CD44, CD80, CD86, CTLA-4, Ang-1 or Ang-2),

12. A specific binding molecule as claimed in claim 8, in which the antibody to a trophic factor is an antibody to Epidermal growth factor (EGF), Platelet derived growth factor (PDGF), Fibroblast growth factor (FGF), Nerve growth factor (NGF), Colony stimulating factor (CSF), Granulocyte/macrophage colony stimulating factor (GM-CSF), Hepatocyte growth factor, insulin- like growth factor, Placenta growth factor, VEGF-A, VEGF-C, VEGF-D, or TGF-β.

13. A specific binding molecule as claimed in claim 10, in which the anti-TNF-a antibody is an antibody selected from the group consisting of adalimumab, infliximab or TN3-19.12.

14. A specific binding molecule as claimed in any one of claims 1 to 13, in which the peptide linker comprising a proteolytic cleavage site is a matrix metalloproteinase (MMP) cleavage site or an aggrecanase cleavage site.

15. A specific binding molecule as claimed in claim 14, in which the amino acid sequence of the MMP proteolytic cleavage site is PLGLWA (SEQ ID NO:1).

16. A nucleic acid sequence encoding the specific binding molecule as defined in any one of claims 1 to 15.

17. A vector comprising a nucleic acid sequence as defined in claim 16.

18. A host cell comprising a nucleic acid or vector as defined in claim 16 or claim 17.

19. A composition comprising the specific binding molecule as defined in any one of claims 1 to 15, the nucleic acid sequence as defined in claim 16, the vector as defined in claim 17, or the host cell as defined in claim 18.

20. A pharmaceutical composition comprising the specific binding molecule as defined in any one of claims 1 to 15, the nucleic acid sequence as defined in claim 16, the vector as defined in claim 17, or the host cell as defined in claim 18.

21 . A specific binding molecule as claimed in any one of claims 1 to 15, nucleic acid sequence as defined in claim 16, a vector as claimed in claim 17 or a host cell as claimed in claim 18 for use in the treatment of inflammatory conditions or cancer.

22. A method for the treatment of inflammatory conditions or cancer comprising the administration to a subject of a composition comprising a specific binding molecule as defined in any one of claims 1 to 15, nucleic acid sequence as defined in claim 16, a vector as claimed in claim 17 or a host cell as claimed in claim 18.

23. A kit of parts comprising a specific binding molecule as defined in any one of claims 1 to 15, nucleic acid sequence as defined in claim 16, a vector as claimed in claim 17 or a host cell as claimed in claim 18 and an administration vehicle including, but not limited to, tablets for oral administration, inhalers for lung administration and injectable solutions for intravenous administration.

24. A process for preparing the specific binding molecule as claimed in any one of claims 1 to 15 comprising production of the specific binding molecule recombinantly by expression of a suitable nucleic acid sequence encoding said specific binding molecule in a host cell, followed by purification of the expressed specific binding molecule.

25. A process for preparing a nucleic acid construct as defined in claim 16, comprising ligating together nucleic acid sequences encoding a specific binding molecule as defined in any one of claims 1 to 15.