HK1195319A - Il-6 binding molecules - Google Patents

Description

IL-6 binding molecules

RELATED APPLICATIONS

Priority of U.S. provisional application serial No. 61/650,883 filed on day 5/23 of 2012, U.S. provisional application serial No. 61/720,102 filed on day 10/30 of 2012, and PCT/IB2012/056424 filed on day 14 of 2012, to which this application claims, are all incorporated herein by reference in their entirety.

Background

Interleukin-6 (IL-6) is the major proinflammatory cytokine. Which results in the proliferation and differentiation of immunocompetent and hematopoietic cells. Human IL-6 is a single glycoprotein consisting of 212 amino acids, with two N-linked glycosylation sites, and has a molecular weight of about 26 kDa. The structure of IL-6 comprises four alpha-helical domains with motifs of four cysteine residues, which are essential for the tertiary structure of IL-6. IL-6 signaling is mediated by IL-6 binding to the soluble or surface-bound alpha chain of the IL-6 receptor (IL-6 Ra) which enables the complex to interact with the cell surface transmembrane gp130 subunit, which mediates intracellular signaling.

IL-6 is involved in the pathogenesis of inflammatory diseases, including inflammatory autoimmune diseases such as Rheumatoid Arthritis (RA), ankylosing spondylitis, Systemic Lupus Erythematosus (SLE), Inflammatory Bowel Disease (IBD), and Castleman's disease. IL-6 is also implicated in the pathogenesis of cancers including prostate cancer, diffuse large cell lymphoma, multiple myeloma, and renal cell carcinoma. Also reported is the role of IL-6 in promoting cancer-associated anorexia (anorexia), oral mucositis and cachexia.

Although IL-6 binding molecules derived from immunization of non-human animals are known in the art, these molecules typically require extensive antibody engineering (e.g., CDR grafting and humanization) to reduce their immunogenicity. Furthermore, the resulting humanized variants are often plagued by undesirable binding affinities for the IL-6 target and require extensive antibody engineering and affinity maturation in an attempt to restore IL-6 binding affinity. The net result is that most IL-6 antibodies exhibit an undesirable binding affinity for the IL-6 target.

Thus, in view of the importance of IL-6 in disease development and the disadvantages of known IL-6 antibodies, it is apparent that there is a need in the art for improved (e.g., hardly engineered) IL-6 agents that can inhibit the biological activity of IL-6 and thus treat diseases associated with IL-6 activity.

Summary of The Invention

The present invention improves upon the prior art by providing binding molecules (e.g., antibodies or antigen-binding fragments thereof) that have an improved binding profile, i.e., specifically bind to IL-6 (e.g., human or non-human primate IL-6) with high binding affinity (e.g., picomolar binding affinity) and strongly inhibit its biological activity (e.g., bind to the IL-6 receptor). In certain exemplary embodiments, the IL-6 binding molecules of the invention are derived from the conventional repertoire of antibodies from camelid species (e.g., llamas) that have been subjected to active immunization with an IL-6 antigen. For example, camelid-derived IL-6 binding molecules of the invention may comprise paired VH/VL domains or other alternative frameworks, wherein one or more hypervariable loops (e.g. H1, H2, H3, L1, L2 and/or L3) of the VH or VL domain are derived from the camelid species. Also, in certain embodiments, at least one of the hypervariable loops adopts a canonical fold (or combination of canonical folds) that is identical or substantially identical to the fold of a human antibody. Such binding molecules exhibit high human homology (sequence and structure) and are therefore particularly useful for the treatment of IL-6 related diseases or disorders (e.g., inflammatory diseases and cancer) due to their low immunogenicity. Surprisingly, the IL-6 antibodies of the invention exhibit high binding affinity, manufacturability (manufacturability) and thermostability without the extensive and time-consuming antibody engineering and affinity maturation typically required for known IL-6 antibodies.

Thus, in one aspect, the invention provides a binding molecule that specifically binds to IL-6, said binding molecule comprising at least one antibody CDR, wherein said CDR comprises at least one amino acid residue that is buried in the F229 or F279 cavity of IL-6 when said binding molecule binds to IL-6. In certain embodiments, the binding molecule comprises a VH domain having the amino acid at position 98 according to Kabat, which is buried in the F229 hole of IL-6 when the antibody or fragment is bound to IL-6. In a specific embodiment, the amino acid at position 98 is tryptophan. In certain embodiments, the binding molecule comprises a VL domain having an amino acid at position 30 according to Kabat, which is buried in the F229 hole of IL-6 when the antibody or fragment is bound to IL-6. In a specific embodiment, the amino acid at position 30 is tyrosine, and in certain embodiments, the binding molecule comprises a VH domain having an amino acid at position 99 according to Kabat, which is buried in the F279 hole of IL-6 when the antibody or fragment is bound to IL-6. In a specific embodiment, the amino acid at position 99 is valine.

In certain embodiments, the binding molecule comprises a VH domain comprising hypervariable loops H1, H2 and H3, wherein the VH domain polypeptide is paired with a VL domain comprising hypervariable loops L1, L2 and L3, wherein at least one of the hypervariable loops H1-H3 and L1-L3 is obtained from a conventional antibody of an alpaca (Lama) species by immunization of the alpaca species with an IL-6 antigen. In a specific embodiment, at least one of the hypervariable loops H1, H2, L1, L2 and L3 exhibits a predicted or actual corresponding canonical fold structure which is identical or substantially identical to the canonical fold structure of the H1, H2, L1, L2 or L3 hypervariable loops found in human antibodies. In a particular embodiment, at least one of hypervariable loops H1 and H2 each exhibit a predicted or actual canonical fold structure which is identical or substantially identical to the corresponding human canonical fold structure. In a particular embodiment, at least one of the hypervariable loops L1, L2 and L3 each exhibit a predicted or actual canonical fold structure which is identical or substantially identical to the corresponding human canonical fold structure. In a specific embodiment, at least one of hypervariable loops H1 and H2 forms a combination of predicted or actual canonical fold structures which is identical or substantially identical to the corresponding combination of canonical fold structures known to occur in human germline VH domains. In a specific embodiment, at least one of hypervariable loops H1 and H2 forms a combination of canonical fold structures which corresponds to a combination of human canonical fold structures selected from the group consisting of: 1-1, 1-2, 1-3, 1-4, 1-6, 2-1, 3-1 and 3-5. In a specific embodiment, at least one of hypervariable loops L1 and L2 forms a combination of predicted or actual canonical fold structures which is identical or substantially identical to the corresponding combination of canonical fold structures known to occur in the human germline VL domain. In a specific embodiment, at least one of the hypervariable loops L1 and L2 forms a combination of canonical fold structures which corresponds to a combination of human canonical fold structures selected from the group consisting of: 11-7, 13-7(A, B, C), 14-7(A, B), 12-11, 14-11, 12-12, 2-1, 3-1, 4-1 and 6-1.

In certain embodiments, the binding molecule comprises a VH domain and a VL domain, wherein the VH domain and/or VL domain of the binding molecule has 90% or greater sequence identity within the framework regions FR1, FR2, FR3 and FR4 with one or more corresponding human VH or VL domains. In certain embodiments, the binding molecule comprises a VH domain and a VL domain and is a germlined variant of a parent binding molecule, wherein one or both of the VH and VL domains of the binding molecule comprises a total of 1 to 10 amino acid substitutions within the framework region as compared to the corresponding VH and VL domains of the parent non-human antibody. In a specific embodiment, the parent binding molecule is a conventional camelid antibody. In certain embodiments, the binding molecule is an antibody or antigen-binding fragment thereof.

In certain embodiments, the binding molecule comprises a VH domain comprising SEQ ID NO 500[ X₁PDVVTGFHYDX₂]The HCDR3 amino acid sequence set forth in (1), or a sequence variant thereof, wherein:

X₁is any amino acid, preferably D or Y;

X₂is any amino acid, preferably Y or N; and

wherein the sequence variant comprises one, two or three amino acid substitutions in the sequence. In a specific embodiment, the HCDR3 amino acid sequence is selected from the group consisting of: 497-499 is shown.

In certain embodiments, the VH domain further comprises SEQ ID NO 507[ VIX₁YX₂X₃DTYYSPSLX₄S]The HCDR2 amino acid sequence set forth in (1) or a sequence variant thereof, wherein:

x1 is any amino acid, preferably D, Y or N;

x2 is any amino acid, preferably D or E;

x3 is any amino acid, preferably a or G;

x4 is any amino acid, preferably E or K; and

wherein the sequence variant comprises one, two or three amino acid substitutions in the sequence. In a specific embodiment, the HCDR2 amino acid sequence is selected from the group consisting of: 501 and 506 in SEQ ID NO.

In certain embodiments, the VH domain is further comprised in SEQ ID NO 512[ X₁X₂YYX₃WX₄]The HCDR1 amino acid sequence set forth in (1), or a sequence variant thereof, wherein:

x1 is any amino acid, preferably T, S or P;

x2 is any amino acid, preferably R or S;

x3 is any amino acid, preferably a or V;

x4 is any amino acid, preferably S or T; and

wherein the sequence variant comprises one, two or three amino acid substitutions in the sequence. In a specific embodiment, the HCDR1 amino acid sequence is selected from the group consisting of seq id no:508 and 511 of SEQ ID NO.

In certain embodiments, the binding molecule comprises a VH domain comprising the HCDR3, HCDR2 and HCDR1 amino acid sequences set forth in SEQ ID NOs 497, 501 and 508, respectively.

In certain embodiments, the binding molecule further comprises a VL domain, wherein the VL domain comprises SEQ ID NO 524[ ASYX₁X₂X₃X₄X₅X₆X₇]The LCDR3 amino acid sequence set forth in (1), or a sequence variant thereof, wherein:

x1 is any amino acid, preferably R or K;

x2 is any amino acid, preferably N, H, R, S, D, T or Y;

x3 is any amino acid, preferably F, Y, T, S or R;

x4 is any amino acid, preferably N or I;

x5 is any amino acid, preferably N or D;

x6 is any amino acid, preferably V, N, G or a;

x7 is any amino acid, preferably V or I; and

wherein the sequence variant comprises one, two or three amino acid substitutions of the sequence. In a particular embodiment, the LCDR3 amino acid sequence is selected from the group consisting of seq id no:513-523.

In certain embodiments, the VL domain further comprises SEQ ID NO 535[ X₁VX₂X₃RX₄S]The LCDR2 amino acid sequence set forth in (1), or a sequence variant thereof, wherein:

x1 is any amino acid, preferably R, K, D, a or E;

x2 is any amino acid, preferably S, N or T;

x3 is any amino acid, preferably T, K or Y;

x4 is any amino acid, preferably a, T or V; and

wherein the sequence variant comprises one, two or three amino acid substitutions of the sequence. In a specific embodiment, the LCDR2 amino acid sequence is selected from the group consisting of seq id no:525 and 534, respectively.

In some embodimentsIn one embodiment, the VL domain further comprises SEQ ID NO 542[ AGX₁X₂X₃DX₄GX₅X₆X₇YVS]The LCDR1 amino acid sequence set forth in (1), or a sequence variant thereof,

wherein

X1 is any amino acid, preferably a or T;

x2 is any amino acid, preferably S or N;

x3 is any amino acid, preferably S, E or N;

x4 is any amino acid, preferably V or I;

x5 is any amino acid, preferably G, Y, T or F;

x6 is any amino acid, preferably G or Y;

x7 is any amino acid, preferably N, D or a; and

wherein the sequence variant comprises one, two or three amino acid substitutions of the sequence. In a specific embodiment, the LCDR1 amino acid sequence is selected from the group consisting of seq id no:538-541 of SEQ ID NO. In certain embodiments, the binding molecule comprises a VL domain comprising the LCDR3, LCDR2 and LCDR1 amino acid sequences set forth in SEQ ID NOs 513, 525 and 536, respectively. In certain embodiments, the binding molecule comprises: a VH domain having the amino acid sequences HCDR3, HCDR2 and HCDR1 set forth in SEQ ID NOs: 497, 501 and 508, respectively; and a VL domain having the LCDR3, LCDR2 and LCDR1 amino acid sequences set forth in SEQ ID NOs 513, 525 and 536, respectively. In certain embodiments, the binding molecule comprises a VH domain having at least 85% sequence identity to the amino acid sequence set forth in SEQ ID NO: 152. In certain embodiments, the binding molecule comprises a VH domain having an amino acid sequence selected from the group consisting of seq id nos: 127 and 232 and 569 and 571, respectively. In certain embodiments, the binding molecule comprises a VH domain having the amino acid sequence of SEQ ID NO: 152. In certain embodiments, the binding molecule comprises a VL domain having at least 85% sequence identity to the amino acid sequence set forth in SEQ ID No. 416. In certain embodiments, the binding molecule comprises a VL domain having an amino acid sequence selected from the group consisting of: 391 and 496 of SEQ ID NO. In certain embodiments, the binding molecule comprises a VL domain having the amino acid sequence of SEQ ID NO 416. In certain embodiments, the binding molecule comprises: a VH domain having the amino acid sequence set forth in SEQ ID NO: 152; and a VL domain having the amino acid sequence set forth in SEQ ID NO: 416.

In certain embodiments, the binding molecule comprises the H1 and H2 loops, which form a combination of canonical fold structures corresponding to combination 3-1 of the human canonical fold structures found in the human 1ACY antibody structure. In certain embodiments, the binding molecule comprises the L1 and L2 loops that form a combination of canonical fold structures corresponding to combination 6 λ -1 of the human canonical fold structures found in human 3MUG antibody structures. In certain embodiments, the binding molecule comprises the L1, L2, and L3 loops, which form a combination of canonical fold structures corresponding to the combination of human canonical fold structures found in human 3MUG antibody structures 6 λ -1-5.

In certain embodiments, the binding molecule comprises a VH domain comprising SEQ ID NO 544[ RAGX₁GX₂G]The HCDR3 amino acid sequence set forth in (1), or a sequence variant thereof, wherein:

X₁is any amino acid, preferably W;

X₂is any amino acid, preferably M, A, L, S or N; and

wherein the sequence variant comprises one, two or three amino acid substitutions of the sequence. In a specific embodiment, the HCDR3 amino acid sequence is selected from the group consisting of seq id no:543, 566, 567 and 568.

In certain embodiments, the VH domain further comprises SEQ ID NO 554[ X ]₁ISX₂X₃GX₄SX₅X₆YX₇DSVKG]The HCDR2 amino acid sequence set forth in (1), or a sequence variant thereof, wherein:

x1 is any amino acid, preferably a, P or R;

x2 is any amino acid, preferably a or S;

x3 is any amino acid, preferably S or G;

x4 is any amino acid, preferably G or V;

x5 is any amino acid, preferably a or T;

x6 is any amino acid, preferably Y, N or S;

x7 is any amino acid, preferably G, a or T; and

wherein the sequence variant comprises one, two or three amino acid substitutions of the sequence. In a specific embodiment, the HCDR2 amino acid sequence is selected from the group consisting of seq id no: 545-553.

In certain embodiments, the VH domain further comprises SEQ ID NO 562[ X ]₁X₂X₃X₄X₅]The HCDR1 amino acid sequence set forth in (1), or a sequence variant thereof, wherein:

x1 is any amino acid, preferably S or T;

x2 is any amino acid, preferably H or Y;

x3 is any amino acid, preferably a or R;

x4 is any amino acid, preferably M or L;

x5 is any amino acid, preferably S or Y; and

wherein the sequence variant comprises one, two or three amino acid substitutions of the sequence. In a specific embodiment, the HCDR1 amino acid sequence is selected from the group consisting of seq id no: SEQ ID NO 555-.

In certain embodiments, the VH domain comprises HCDR3 having an amino acid sequence selected from the group consisting of seq id nos: 543, 566, 567 and 568, and comprises the HCDR2 and HCDR1 amino acid sequences set forth in SEQ ID NOs 545 and 555, respectively. In certain embodiments, the binding molecule further comprises a VL domain, wherein the VL domain comprises the LCDR3 amino acid sequence set forth in SEQ ID NO:563, or a sequence variant thereof, wherein said sequence variant comprises one, two, or three amino acid substitutions of said sequence. In certain embodiments, the VL domain further comprises the LCDR2 amino acid sequence set forth in SEQ ID No. 564, or a sequence variant thereof, wherein the sequence variant comprises one, two, or three amino acid substitutions of the sequence. In certain embodiments, the VL domain further comprises the LCDR1 amino acid sequence set forth in SEQ ID No. 565, or a sequence variant thereof, wherein said sequence variant comprises one, two or three amino acid substitutions of said sequence. In certain embodiments, the VL domain comprises the LCDR3, LCDR2, and LCDR1 amino acid sequences set forth in SEQ ID NOS: 563, 564, and 565, respectively. In certain embodiments, the binding molecule comprises: a VH domain having the HCDR3, HCDR2 and HCDR1 amino acid sequences set forth in SEQ ID NOs 544, 545 and 555, respectively; and a VL domain having the LCDR3, LCDR2 and LCDR1 amino acid sequences set forth in SEQ ID NOS: 563, 564 and 565, respectively. In certain embodiments, the binding molecule comprises a VH domain having at least 85% sequence identity to the amino acid sequence set forth in SEQ ID No. 86. In certain embodiments, the binding molecule comprises a VH domain having an amino acid sequence selected from the group consisting of seq id nos: 39-126 and 569-571. In certain embodiments, the binding molecule comprises a VH domain having an amino acid sequence selected from SEQ ID NO:86, SEQ ID NO:569, SEQ ID NO:570 and SEQ ID NO: 571. In certain embodiments, the binding molecule comprises a VL domain having at least 85% sequence identity to the amino acid sequence set forth in SEQ ID NO: 350. In certain embodiments, the binding molecule comprises a VL domain having an amino acid sequence selected from the group consisting of: 303-390 SEQ ID NO. In certain embodiments, the binding molecule comprises a VL domain having the amino acid sequence of SEQ ID NO: 350. In certain embodiments, the binding molecule comprises: a VH domain having the amino acid sequence set forth in SEQ ID NO:86, SEQ ID NO:569, SEQ ID NO:570 or SEQ ID NO: 571; and a VL domain having the amino acid sequence set forth in SEQ ID NO: 350.

In certain embodiments, the binding molecule comprises loops H1 and H2 that form a combination of canonical fold structures corresponding to combinations 1-3 of the human canonical fold structures found in the structure of human 1DFB antibodies. In certain embodiments, the binding molecule comprises loops L1 and L2 that form a combination of canonical fold structures corresponding to combination 7 λ -1 of the human canonical fold structure found in the human 1MFA antibody structure. In certain embodiments, the binding molecule comprises loops L1, L2, and L3, which form a combination of canonical fold structures corresponding to combination 7 λ -1-4 of the human canonical fold structures found in human 3MUG antibody structures.

In certain embodiments, the binding molecule is a Fab fragment that is at less than 2x10^-5s^-1Off-rate (k measured by surface plasmon resonance)_off) Bind to human IL-6. In certain embodiments, the binding molecule binds to human IL-6 antigen with a subpicomolar binding affinity. In certain embodiments, the binding molecule binds to human IL-6 antigen with a single digit femtomolar binding affinity. In certain embodiments, the binding molecule comprises a hypervariable loop which is obtained from a conventional antibody of the alpaca without subsequent affinity maturation. In certain embodiments, the binding molecule inhibits IL-6-induced proliferation of B9 hybridoma cells with an IC50 of less than 0.1 pM.

In certain embodiments, the binding molecule exhibits a melting temperature (Tm) greater than 65 ℃. In certain embodiments, the binding molecule is a germlined variant of a parent camelid antibody, the germlined variant having a higher melting temperature than the parent camelid antibody.In certain embodiments, the binding molecule is expressed at a level of at least 20mg/ml following transient expression in HEK293 cells. In certain embodiments, the binding molecule is characterized as being less than about 10.0, such as less than about 6.0And (6) scoring. In certain embodiments, the binding molecule inhibits the binding of IL-6 to an IL-6 receptor. In certain embodiments, the binding molecule inhibits the binding of gp130 to the IL-6 receptor. In certain embodiments, the binding molecule specifically binds to human and cyno IL-6. In certain embodiments, the binding molecule comprises at least one CDR from a camelid antibody that specifically binds to IL-6.

In another aspect, the invention provides a pharmaceutical composition comprising a binding molecule according to any preceding claim, and one or more pharmaceutically acceptable carriers.

In another aspect, the invention provides a method of treating an IL-6 related disease or disorder comprising administering to a subject in need of such treatment an effective amount of a pharmaceutical composition of the invention.

In another aspect, the invention provides an isolated nucleic acid encoding a binding molecule disclosed herein.

In another aspect, the invention provides a recombinant expression vector comprising a nucleic acid molecule of the invention.

In another aspect, the invention provides a host cell comprising a recombinant expression vector of the invention.

Brief Description of Drawings

FIG. 1 depicts the results of a cell proliferation assay measuring the IL-6 neutralizing activity of antibodies of the invention in vitro.

FIG. 2 depicts the results of a tumor xenograft experiment in mice with epithelial ovarian cancer measuring the in vivo efficacy of the antibodies of the invention.

FIGS. 3A-B show camelid-derived hypervariable loops (L1-L3, H1 and H2) of the 61H7 antibody of the invention using a combination of canonical folds and canonical folds of predicted human antibodies.

Fig. 4A-B show camelid-derived hypervariable loops (L1-L3, H1 and H2) of the 68F2 antibody and its germline variant (129D3) of the invention using a combination of canonical folds and canonical folds of predicted human antibodies.

FIG. 5 depicts a space filling model for IL-6, which is overlaid (overlay) with the following, respectively: (A) f229 of IL-receptor; (B) w98 of F229 and 61H7VH of IL-6 receptor, (C) Y30 of F229 and 68F2VL of IL-6 receptor; and (D) W98 for F229, 61H7VH and V99 for 68F2VH for the IL-6 receptor, both according to Kabat numbering.

FIG. 6 depicts a space-filling model of two surface-bound cavities on IL-6 that are important for IL-6 receptor binding, which is superimposed with: residues F229 and F279 of the IL-6 receptor, and residues Y30 of 68F2VL and V99 of 68F2VH, all according to Kabat numbering (Y32 and V104 in this structure).

Figures 7A-B depict the thermostability of 68F2 and its germlined variant 129D3, measured in Biacore with immobilized glycosylated human IL-6 for: (A) other germlined variant IL-6 antibodies of the invention and (B) other reference antibodies. The upper part of each figure depicts the melting curve, while the lower part lists the Tm values for each antibody.

Figure 8 depicts the serum stability of antibodies clones 68F2, 129D3 (a germlined variant of 68F 2) and 103a1 (a variant of 61H 7). Reference antibody GL18 was also included.

Figure 9 depicts the low immunogenicity (Epibase) scores for IL-6 antibodies of the invention relative to a reference antibody (shown in bold), including the full human antibody adalimumab (Humira).

Fig. 10A-B depict an alignment of the VH and VL of (a)68F2 and (B)61H7, depicting high levels of sequence homology with the framework regions of their corresponding germlined variants 129D3 and 111a 7. The minimum number of framework changes introduced per molecule is also shown (13 in total).

Fig. 11A-B show an alignment of VH and VK of (a) CNTO328 and (B) VH _ rabbit (ALD518), depicting high levels of sequence homology to the framework regions of their corresponding germlined variants CNTO136 and VH _ human (ALD 518). The minimum number of framework changes introduced per molecule (36 and 46 in total) is also shown.

Figure 12 shows the pharmacokinetic profile of the 129D3IgG1 antibody and its variants in cynomolgus monkeys.

Figure 13 depicts the results of a serum amyloid a (saa) mouse model experiment measuring the in vivo efficacy of the antibodies of the invention.

Figure 14 depicts the results of a mouse psoriasis xenograft experiment measuring the in vivo efficacy of antibodies of the invention.

Figure 15 depicts tumor growth data observed in experiments measuring the in vivo efficacy of antibodies of the invention in a renal cell carcinoma mouse tumor xenograft model.

FIG. 16 depicts a Kaplan-Meier plot of survival data observed in experiments measuring the in vivo efficacy of antibodies of the invention in a renal cell carcinoma mouse tumor xenograft model.

Figure 17 depicts tumor growth data observed in experiments measuring the in vivo efficacy of antibodies of the invention in a renal cell carcinoma mouse tumor xenograft model, all agents administered at 3 mg/kg.

FIG. 18 depicts a Kaplan-Meier plot of survival data observed in experiments measuring the in vivo efficacy of antibodies of the invention in a renal cell carcinoma mouse tumor xenograft model, all agents administered at 3 mg/kg.

Detailed Description

I. Definition of

In order that the invention may be more readily understood, certain terms are first defined.

As used herein, the term "IL-6" refers to interleukin-6. IL-6 nucleotide and polypeptide sequences are well known in the art. An exemplary human IL-6 amino acid sequence is listed in GenBank accession GI:10834984, while an exemplary mouse IL-6 amino acid sequence is listed in GenBank accession GI: 13624311.

As used herein, the term "antibody" refers to an immunoglobulin molecule comprising four polypeptide chains, namely two heavy (H) chains and two light (L) chains, interconnected by disulfide bonds, and a multimer thereof (e.g., an IgM). Each heavy chain comprises a heavy chain variable region (abbreviated VH) and a heavy chain constant region. The heavy chain constant region comprises three domains: CH1, CH2 and CH 3. Each light chain comprises a light chain variable region (abbreviated VL) and a light chain constant region. The light chain constant region comprises a domain (CL 1). The VH and VL regions can be further subdivided into regions with hypervariability, called Complementarity Determining Regions (CDRs), interspersed with more conserved regions called Framework Regions (FRs).

As used herein, the term "antigen-binding fragment" of an antibody includes any naturally occurring, enzymatically obtainable, synthetic, or genetically engineered polypeptide or glycoprotein that specifically binds an antigen to form a complex. Antigen-binding fragments of antibodies may be derived, for example, from intact antibody molecules using any suitable standard technique, such as proteolytic digestion or recombinant genetic engineering techniques (involving manipulation and expression of DNA encoding antibody variable and optionally constant domains). Non-limiting examples of antigen-binding moieties include: (i) a Fab fragment; (ii) f (ab')₂A fragment; (iii) (ii) a fragment of Fd; (iv) (iv) an Fv fragment; (v) single chain fv (scFv) molecules; (vi) a dAb fragment; and (vii) a minimal recognition unit (e.g., an isolated Complementarity Determining Region (CDR)) consisting of amino acid residues that mimic the hypervariable region of an antibody. Other engineered molecules, such as bivalent antibodies (diabodies), trivalent antibodies (triabodies), tetravalent antibodies (tetrabodies) and minibodies (minibodies), are also encompassed within the expression "antigen binding portion".

As used herein, the term "variable region" or "variable domain" refers to the fact that: certain portions of the variable domains VH and VL differ widely in sequence between antibodies and are used for the binding and specificity of each particular antibody for its target antigen. However, the differences are not evenly distributed in the variable domain of the antibody. It aggregates into three segments called "hypervariable loops" in each of the VL and VH domains, which form part of the antigen-binding site. The first, second and third hypervariable loops of the V lamda light domain are referred to herein as L1(λ), L2(λ) and L3(λ) and can be defined as comprising residues 24-33(L1(λ) consisting of 9,10 or 11 amino acid residues), 49-53(L2(λ) consisting of 3 residues) and 90-96(L3(λ) consisting of 5 residues) in the VL domain (Morea et al, Methods20:267-279 (2000)). The first, second and third hypervariable loops of the V kappa light chain domain are referred to herein as L1(κ), L2(κ) and L3(κ) and may be defined as comprising residues 25-33(L1(κ), consisting of 6,7, 8, 11, 12 or 13 residues), 49-53(L2(κ), consisting of 3 residues) and 90-97(L3(κ), consisting of 6 residues) in the VL domain (Morea et al, Methods20:267-279 (2000)). The first, second and third hypervariable loops of the VH domain are referred to herein as H1, H2 and H3 and can be defined as comprising residues 25-33(H1, consisting of 7, 8 or 9 residues), 52-56(H2, consisting of 3 or 4 residues) and 91-105(H3, highly variable in length) in the VH domain (Morea et al, Methods20:267-279 (2000)).

Unless otherwise indicated, the terms L1, L2, and L3 refer to the first, second, and third hypervariable loops, respectively, of the VL domain and encompass the hypervariable loops obtained from the V kappa and V lambda isoforms. The terms H1, H2 and H3 refer to the first, second and third hypervariable loops, respectively, of a VH domain and encompass hypervariable loops obtained from any known heavy chain isotype, including gamma, epsilon, delta, alpha or mu.

Hypervariable loops L1, L2, L3, H1, H2 and H3 may each comprise a portion of a "complementarity determining region" or "CDR" as defined below. The terms "hypervariable loop" and "complementarity determining region" are not strictly synonymous, in that hypervariable loops (HV) are defined on the basis of structure, whereas Complementarity Determining Regions (CDRs) are defined on the basis of sequence variability (Kabat et al, Sequences of Proteins of Immunological Interest, 5th Ed. public Health Service, National Institutes of Health, Bethesda, MD., 1983), and the restrictions on HV and CDR may be different in some VH and VL domains.

The CDRs of the VL and VH domains can generally be defined as comprising the following amino acids: residues 24-34 of the light chain variable domain (CDRL1), 50-56(CDRL2) and 89-97(CDRL3) and residues 31-35 or 31-35b of the heavy chain variable domain (CDRH1), 50-65(CDRH2) and 95-102(CDRH3) (Kabat et al, Sequences of Proteins of Immunological Interest, 5th Ed. public Health Service, National Institutes of Health, Bethesda, Md. (1991)). Thus, HVs may be included in the respective CDRs, and references herein to "hypervariable loops" of the VH and VL domains are to be construed as also encompassing the corresponding CDRs and vice versa, unless otherwise indicated.

The more highly conserved portions of the variable domains are called Framework Regions (FR), as defined below. The variable domains of native heavy and light chains each comprise four FRs (FR 1, FR2, FR3 and FR4, respectively) which are predominantly in a β -stacked configuration, linked by three hypervariable loops. The hypervariable loops in each chain are held together tightly by the FRs and participate, together with hypervariable loops from other chains, in the formation of the antigen-binding site of the antibody. Structural analysis of antibodies revealed the relationship between the shape and sequence of the binding sites formed by the complementarity determining regions (Chothia et al, J.mol.biol.227: 799-. Despite its high degree of sequence diversity, five of the six loops adopt only a small repertoire of backbone conformations, referred to as the "canonical structure". These conformations are determined firstly by the length of the loop and secondly by the presence of critical residues at certain positions in the loop and in the framework regions, which determine conformation by packing, hydrogen bonding, or the ability to assume an unusual backbone conformation.

As used herein, the term "complementarity determining regions" or "CDRs" refers to non-contiguous antigen binding sites found within the variable regions of heavy and light chain polypeptides. These specific regions are described by Kabat et al, J.biol.chem.252, 6609-. Amino acid residues encompassing the CDRs as defined in each of the cited references above are listed for comparison. Preferably, the term "CDR" is a CDR defined by Kabat based on sequence comparison.

Table 1: CDR definition

¹Residue numbering follows Kabat et al, supra nomenclature

²Residue numbering follows the nomenclature of Chothia et al, supra

³Residue numbering follows MacCallum et al, supra nomenclature

As used herein, the term "framework region" or "FR region" includes amino acid residues that are part of the variable region, but not part of the CDRs (e.g., using the Kabat definition of CDRs). Thus, the variable region framework is about 100 to 120 amino acids in length, but includes only those amino acids outside the CDRs. For specific examples of heavy chain variable regions and for CDRs as defined by Kabat et al, framework region 1 corresponds to a domain encompassing the variable region of amino acids 1-30; framework region 2 corresponds to a domain encompassing the variable region of amino acids 36-49; framework region 3 corresponds to the domain encompassing amino acids 66-94, and framework region 4 corresponds to the domain of the variable region from amino acid 103 to the end of the variable region. The framework regions for the light chain are similarly separated by each light chain variable region CDR. Similarly, using the definition of CDRs by Chothia et al or McCallum et al, the framework region boundaries are separated by the corresponding CDR ends as described above. In a preferred embodiment, the CDRs are as defined by Kabat.

In naturally occurring antibodies, the six CDRs present on each monomeric antibody are short, non-contiguous amino acid sequences that are specifically positioned to form an antigen binding site when the antibody assumes its three-dimensional configuration in an aqueous environment. The remaining heavy and light chain variable domains show less intermolecular variability in amino acid sequence and are referred to as framework regions. The framework regions adopt predominantly a β -stacked conformation, and the CDRs form loops that connect, and in some cases form part of, the β -stacked structure. Thus, these framework regions act to form a scaffold (scaffold) that allows the six CDRs to be placed in the correct orientation by inter-chain, non-covalent interactions. The antigen binding site formed by the arranged CDRs defines a surface complementary to an epitope on the immunoreactive antigen. The complementary surface promotes non-covalent binding of the antibody to the immunoreactive epitope. The position of the CDRs can be readily identified by one of ordinary skill in the art.

As used herein, the term "F229 cavity" refers to the surface cavity of human IL-6 occupied by the phenylalanine 229 residue of the human IL-6 receptor in the IL-6/IL-6 receptor complex listed in Boulanger et al, 2003, Science27, 2101-2104 (incorporated herein by reference in its entirety).

As used herein, the term "F279 cavity" refers to the surface cavity of human IL-6 occupied by the phenylalanine 279 residue of the human IL-6 receptor in the IL-6/IL-6 receptor complex listed in Boulanger et al, 2003, Science27, 2101-2104 (incorporated herein by reference in its entirety).

As used herein, the term "camelid-derived" refers to antibody variable region amino acid sequences (e.g., framework or CDR sequences) that naturally occur in camelid (e.g., llama) antibody molecules. Antibodies of camelid origin may be obtained from any camelid species, including but not limited to llama (llama), dromedary (dromedary), alpaca (alpaca), vicuna (vicuna), guanaco (guanaco) or camel (camel). In certain embodiments, the camelid (e.g., llama) is actively immunized with IL-6 (e.g., human IL-6). In certain embodiments, the term "camelid-derived" is limited to antibody sequences derived from the conventional repertoire of camelids of antibodies, and specifically excludes antibody sequences derived from the repertoire of heavy chain-only antibodies (VHHs) of camelids.

As used herein, the term "conventional antibody" refers to an antibody of any isotype, including IgA, IgG, IgD, IgE or IgM. Natural or naturally occurring "conventional" camelid antibodies are typically heterotetrameric glycoproteins, composed of two identical light (L) chains and two identical heavy (H) chains. Each light chain is linked to a heavy chain by one covalent disulfide bond, and the number of disulfide linkages varies between heavy chains of different immunoglobulin isotypes. Each heavy and light chain also has regularly spaced interchain disulfide bridges. Each heavy chain has a variable domain (VH) at one end (N-terminus), followed by multiple constant domains. Each light chain has a variable domain (VL) at one end (N-terminus) and a constant domain (CL) at its other end; the constant domain of the light chain is aligned with the first constant domain of the heavy chain, and the light chain variable domain is aligned with the variable domain of the heavy chain. Specific amino acid residues are thought to form the interface between the light and heavy chain variable domains.

As used herein, the term "specifically binds" refers to an antibody or antigen-binding fragment thereof at least about 1x10^-6(e.g., 1x 10)^-6M，1x10^-7M，1x10^-8M，1x10^-9M，1x10^-10M，1x10^-11M，1x10^-12M，1x10^-13M，1x10^-14M，1x10^-15M or higher), preferably 1x10^-12M to 1x10^-15M or higher KD to IL-6, and/or greater than nonspecific antigen greater than at least two times the affinity binding to IL-6 ability. However, it is understood that an antibody or antigen-binding fragment thereof is capable of specifically binding to two or more antigens that are related in sequence. For example, the antibodies or antigen-binding fragments thereof disclosed herein are capable of specifically binding to human and non-human (e.g., mouse or non-human primate) IL-6.

As used herein, the term "antigen" refers to a binding site or epitope that is recognized by an antibody variable region.

As used herein, the terms "treatment" and "treating" refer to therapeutic or prophylactic measures described herein. Methods of "treating" and "treating" employ administering an antibody or antigen-binding fragment thereof of the invention to a subject, e.g., a subject having an IL-6 associated disease or disorder (e.g., inflammation and cancer) or susceptible to such a disease or disorder, to prevent, cure, delay, reduce the severity of, or alleviate one or more symptoms of the disease or disorder or relapsed disease or disorder, or to prolong the survival of the subject beyond that expected in the absence of such treatment.

As used herein, the term "IL-6 associated disease or disorder" includes disease states and/or symptoms associated with IL-6 activity. Exemplary IL-6-associated diseases or disorders include, but are not limited to, inflammatory diseases (e.g., inflammatory autoimmune diseases such as rheumatoid arthritis and systemic lupus erythematosus), cancer (e.g., prostate cancer, diffuse large cell lymphoma, multiple myeloma and renal cell carcinoma), and cancer-associated disorders (e.g., anorexia and cachexia).

As used herein, the term "effective amount" refers to an amount of an antibody or antigen-binding fragment thereof that, when administered to a subject, is sufficient to effect treatment, prognosis, or diagnosis of an IL-6 associated disease or disorder as described herein. The therapeutically effective amount may vary depending on the subject and disease condition being treated, the weight and age of the subject, the severity of the disease condition, the mode of administration, and the like, which can be readily determined by one of ordinary skill in the art. The dosage for administration may range, for example, from about 1ng to about 10,000mg, from about 1ug to about 5,000mg, from about 1mg to about 1,000mg, from about 10mg to about 100mg of an antibody or antigen-binding fragment thereof according to the invention. The dosage regimen may be adjusted to provide an optimal therapeutic response. An effective amount is also one in which any toxic or deleterious (i.e., side effects) effects of the binding polypeptide are minimized and/or less important than the beneficial effects.

As used herein, the term "subject" includes any human or non-human animal.

As used herein, the term "surface plasmon resonance" refers to a method that allows for detection by detecting within a biosensor matrixAnalysis of optical phenomena of real-time interactions by changes in protein concentration, e.g. using BIAcore^TMSystem (Biacore Life Sciences division of GE Healthcare, Piscataway, NJ).

As used herein, the term "K_D"refers to the equilibrium dissociation constant for a particular binding polypeptide/antigen interaction.

As used herein, the term "dissociation rate" refers to the dissociation rate (K) for a particular binding interaction_off)。

IL-6 binding molecules

In one aspect, the invention provides binding molecules (antibodies or antigen-binding fragments thereof) that specifically bind to IL-6 and inhibit its activity. Such binding molecules typically comprise at least one amino acid sequence of a CDR region as set forth in tables 13-18 herein.

Analysis of the crystal structure of human IL-6 complexed with the human IL-6 receptor revealed that two residues of the IL-6 receptor, F229 and F279, are critical for IL-6/IL-6 receptor interaction (see, e.g., Boulanger et al, 2003, Science27, 2101-2104, which is incorporated herein by reference in its entirety). In the IL-6/IL-6 receptor complex, F229 and F279 are buried in different cavities on the surface of IL-6. In certain embodiments, the binding molecules of the invention utilize these holes on IL-6 to achieve high affinity binding. In a specific embodiment, the binding molecule of the invention comprises an antibody CDR region, wherein said CDR region comprises amino acid residues that are buried in the F229 or F279 cavity on IL-6 when said binding molecule binds to IL-6.

Generally, the binding molecules of the invention inhibit IL-6 activity (e.g., by antagonizing IL-6 binding to the IL-6 receptor). In certain embodiments, the binding molecule also inhibits the binding of gp130 to the IL-6 receptor. However, in other embodiments, the binding molecule can bind to IL-6 without inhibiting gp130 binding to the IL-6 receptor.

Binding molecule pairs of the inventionIL-6 has high affinity, and usually in vivo and in vitro on inhibition of IL-6 activity is highly powerful. In certain embodiments, the binding molecules of the invention are present at about 1x10^-4s^-1(e.g., about 9x10^-5，8x10^-5，7x10^-5，6x10^-5，5x10^-5，4x10^-5，3x10^-5，2x10^-5And 1x10^-5) Dissociation rate (k measured by surface plasmon resonance)_off) Bind to human IL-6. In other embodiments, the binding molecules of the invention inhibit IL-6-induced proliferation of B9 hybridoma cells with an IC50 of less than 0.1 pM. In certain other embodiments, the binding molecules of the invention compete with a predetermined antibody that binds to IL-6, wherein such predetermined antibody comprises a VH sequence and a VL sequence selected from the group consisting of the VH and VL amino acid sequences set forth in tables 13-18. In certain other embodiments, the binding molecule of the invention competes against at least 30%, 40%, 50%, 60%, 70%, 80%, or 90% of the binding of the predetermined antibody to IL-6. In certain other embodiments, the binding molecules of the invention compete with the binding of 20a4, 24D10, 68F2, 61H7, 129D3 or 111a7 to IL-6, e.g., compete to exclude the binding of at least 50%, 60%, 70%, 80% or 90% of one of these antibodies to IL-6. In certain other embodiments, the binding molecules of the invention compete with the binding of 17F10, 24C9, 18C11, 29B11, 28a6 or 126A3 to IL-6, e.g., compete to exclude the binding of at least 50%, 60%, 70%, 80% or 90% of one of these antibodies to IL-6.

Generally, the binding molecules of the present invention also exhibit a high degree of thermal stability. In certain embodiments, the binding molecule exhibits a melting temperature (Tm) greater than 55 ℃ (e.g., at least 56, 57, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75 ℃ or higher). In certain exemplary embodiments, the Il-6 binding molecules of the invention are germlined variants that exhibit comparable or higher thermostability than their parent, camelid origin congeners. In certain exemplary embodiments, thermal stability is measured according to incubation in a suitable buffer (PBS) at a concentration of 100 μ g/ml for 1 hour. In other exemplary embodiments, the IL-6 binding molecule is thermostable to a full-length IgG format (e.g., comprising an IgG1 or IgG4Fc region).

The binding molecules of the invention are also characterized by high expression levels of functional antibodies, and low levels of non-functional contaminants, such as high or low molecular weight aggregates. For example, the IL-6 binding molecules of the invention can be characterized by a production level of at least 20mg/L (e.g., at least 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30mg/L or higher). In certain exemplary embodiments, the IL-6 binding molecule is a germlined variant that exhibits comparable or higher expression levels than its parent, camelid. In other exemplary embodiments, the expression level is determined using a full-length IgG format of the IL-6 binding molecules of the invention, e.g., by transient expression in HEK293 cells.

The binding molecules of the invention are also generally characterized by low predicted immunogenicity. For example, IL-6 binding molecules of the invention exhibit less than 15.0, less than about 12.0, or less than about 10.0Scores (e.g., total DRB1 scores). In certain exemplary embodiments, the binding molecule exhibits an immunogenicity score of about 9.0, about 8.0, about 7.0, or about 6.0. In still other embodiments, the immunogenicity score is less thanE.g., about 6.0, about 5.0, or about 4.0.

The binding molecules of the invention can bind to any IL-6, including but not limited to human and cynomolgus IL-6. Preferably, the binding molecule binds to human and cyno IL-6.

i) IL-6 antibodies or antigen-binding fragments thereof

In certain embodiments, the invention provides antibodies, or antigen-binding fragments thereof, that specifically bind to IL-6 (e.g., human IL-6) and antagonize IL-6's binding to the IL-6 receptor. The VH, VL and CDR sequences of exemplary Fab clones of the invention are listed in tables 13-18. The antibodies of the invention may comprise the framework and/or CDR amino acid sequences of any of these Fab clones.

The antibodies of the invention may comprise CDR region sequences having amino acid residues (e.g., aromatic amino acids such as tryptophan or tyrosine) that are buried in the F229 cavity on IL-6 when the antibody or fragment binds to IL-6. Exemplary antibodies comprise a VH domain with tryptophan at position 98 and/or a VL domain with tyrosine at position 30 according to Kabat. Such antibodies have a particularly high affinity for IL-6.

Additionally or alternatively, the antibodies of the invention may comprise CDR region sequences that have amino acid residues buried in the F279 cavity on IL-6 when the antibody or fragment binds to IL-6. An exemplary antibody comprises a VH domain having a valine at position 99 according to Kabat.

In certain embodiments, an anti-IL-6 antibody or fragment of the invention comprises a VH comprising 1, 2, or 3CDR amino acid sequences from the VH set forth in tables 13-16.

In certain embodiments, an anti-IL-6 antibody or fragment of the invention comprises a VL comprising 1, 2, or 3CDR amino acid sequences from the VLs listed in tables 13-16.

In certain embodiments, an anti-IL-6 antibody or fragment of the invention comprises: a VH comprising 1, 2 or 3CDR amino acid sequences from a VH listed in tables 13-18; and a VL comprising 1, 2, or 3CDR amino acid sequences from the VLs listed in tables 13-18. In a preferred embodiment, all six CDRs are from the same Fab clone.

In certain embodiments, an anti-IL-6 antibody or fragment of the invention comprises a VH set forth in tables 13-16.

In certain embodiments, an anti-IL-6 antibody or fragment of the invention comprises a VL set forth in tables 13-16.

In certain embodiments, an anti-IL-6 antibody or fragment of the invention comprises a VH and VL set forth in tables 13-16.

In certain embodiments, an anti-IL-6 antibody or fragment of the invention comprises a VH and a VL from a single Fab clone listed in tables 13-16.

In certain embodiments, the invention provides an antibody or antigen-binding fragment thereof that specifically binds to IL-6, said antibody or fragment comprising sequence variants of the CDR, VH and VL amino acid sequences listed in tables 13-18.

In certain embodiments, the sequence variant comprises a VH and/or VL amino acid sequence having about 80%, 85%, 86%, 87%, 88%, 89%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98% or 99% identity to a VH or VL region amino acid sequence set forth in tables 13-16.

In other embodiments, the sequence variant comprises a VH, VL, or CDR amino acid sequence selected from tables 13-18 altered by the introduction of one or more conservative amino acid substitutions. Conservative amino acid substitutions include substitutions of one type of amino acid with the same type of amino acid, where the type is defined by common physicochemical amino acid side chain properties and high substitution frequencies in homologous proteins found in nature, as determined by standard Dayhoff frequency exchange matrices or BLOSUM matrices. Six major classes of amino acid side chains have been assigned, including: type i (cys); type II (Ser, Thr, Pro, Ala, Gly); type III (Asn, Asp, Gln, Glu); type IV (His, Arg, Lys); type V (Ile, Leu, Val, Met); and type VI (Phe, Tyr, Trp). For example, substitution of another type III residue with Asp, such as Asn, Gln or Glu, is a conservative substitution. Thus, a predicted nonessential amino acid residue in an IL-6 antibody or antigen-binding fragment thereof is preferably replaced with another amino acid residue from the same type. Methods for identifying conservative substitutions of amino acids that do not eliminate antigen binding are well known in the art (see, e.g., Brummell et al, biochem.32:1180-1187 (1993); Kobayashi et al Protein Eng.12(10):879-884 (1999); and Burks et al Proc. Natl. Acad. Sci. USA94:412-417 (1997)).

In other embodiments, the sequence variant comprises a VH, VL, or CDR amino acid sequence selected from tables 13-18 altered to improve antibody production and/or manufacture, e.g., to exchange methionine for alanine, serine, or leucine. In certain other embodiments, the sequence variant comprises a VH, VL, or CDR amino acid sequence selected from tables 13-18 altered to improve antibody production, e.g., to exchange glutamine for glutamic acid, or asparagine for alanine or a related amino acid. In exemplary embodiments, one or more glutamines outside of the CDR regions of the VH amino acid sequences of table 16 have been changed to glutamate, e.g., one or more glutamines at positions 1, 3, 5 or 16 or any combination thereof in SEQ ID No.152 have been changed to glutamate to improve antibody production or stability. In a specific embodiment, the glutamine at position 1in SEQ ID No.152 has been changed to glutamic acid.

ii) IL-6 binding molecules with high human homology

In certain aspects, the IL-6 binding molecules of the invention are antibodies (or antigen-binding fragments) with high human homology. An antibody should be considered to have "high human homology" if its VH and VL domains together exhibit at least 90% amino acid sequence identity to the most closely matched human germline VH and VL sequences. Antibodies with high human homology may include antibodies comprising the VH and VL domains of natural non-human antibodies that exhibit sufficiently high% sequence identity with human germline sequences, including, for example, antibodies comprising the VH and VL domains of camelid conventional antibodies, as well as engineered, particularly humanized variants of such antibodies, and also "fully human" antibodies.

In one embodiment, the VH domain of the antibody having high human homology may exhibit 80% or greater amino acid sequence identity or sequence homology with one or more human VH domains in the framework regions FR1, FR2, FR3 and FR 4. In other embodiments, the amino acid sequence identity or sequence homology between the VH domain of the polypeptide of the invention and the closest matching human germline VH domain sequence may be 85% or more, 90% or more, 95% or more, 97% or more, or up to 99% or even 100%.

In one embodiment, the VH domain of the antibody having high human homology may contain fewer than 10 (e.g. 10, 9, 8, 7, 6,5, 4, 3, 2 or 1) amino acid sequence substitutions in the framework regions FR1, FR2, FR3 and FR4 compared to the closest matching human VH sequence.

In another embodiment, the VL domain of said antibody having high human homology may exhibit 80% or more sequence identity or sequence homology with one or more human VL domains in the framework regions FR1, FR2, FR3 and FR 4. In other embodiments, the amino acid sequence identity or sequence homology between the VL domain of the polypeptide of the invention and the closest matching human germline VH domain sequence may be 85% or more, 90% or more, 95% or more, 97% or more, or up to 99% or even 100%.

In one embodiment, the VL domain of the antibody having high human homology may contain less than 10 (e.g. 10, 9, 8, 7, 6,5, 4, 3, 2 or 1) amino acid sequence substitutions in the framework regions FR1, FR2, FR3 and FR4 compared to the closest matching human VL sequence.

Antibodies with high human homology may also comprise hypervariable loops or CDRs which have a canonical fold of human or human-like, as discussed in detail below. In one embodiment, at least one hypervariable loop or CDR in the VH domain or VL domain of an antibody with high human homology may be obtained or derived from the VH or VL domain of a non-human antibody, e.g. a conventional antibody from a species in the family camelidae, but which exhibits a predicted or actual canonical fold structure which is substantially the same as the canonical fold structure which occurs in a human antibody.

It should be noted that antibodies with high human homology do not necessarily possess the canonical fold structure of human or human-like. For example, primate antibodies have high sequence homology to human antibodies, but do not generally possess the canonical fold structure of human or human-like.

It is well established in the art that, although the primary amino acid sequences of the hypervariable loops present in the VH and VL domains encoded by the human germline are highly variable by definition, all hypervariable loops, except the CDR H3 of the VH domain, adopt only a few different structural conformations, termed canonical folds (Chothia et al, J.mol.biol.196:901-917 (1987); Tramotano et al Proteins6:382-94(1989)), depending on the length of the hypervariable loop and the presence of so-called canonical amino acid residues (Chothia et al, J.mol.biol.196:901-917 (1987)). The actual canonical structure of the hypervariable loops in the intact VH or VL domain can be determined by structural analysis (e.g., X-ray crystallography), but the canonical structure can also be predicted based on key amino acid residues that are characteristic of a particular structure (discussed further below). Basically, the specific pattern of residues that determines each canonical structure constitutes a "signature", which enables recognition of the canonical structure in the hypervariable loop of the VH or VL domain of unknown structure; the canonical structure may therefore be predicted based on the primary amino acid sequence only.

The predicted canonical fold structure for the hypervariable loops of any given VH or VL sequence in antibodies with high human homology can be analyzed using algorithms publicly available from www.bioinf.org.uk/abs/chothia. html, www.biochem.ucl.ac.uk/. about martin/antibodies. html and www.bioc.unizh.ch/antibodies/Sequences/Germlines/Vbase _ hvk. html. These tools allow the alignment of problem VH or VL sequences with human VH or VL domain sequences of known canonical structure and the prediction of the canonical structure of the hypervariable loops of the problem sequence.

In the case of VH domains, the H1 and H2 loops can be assessed as having a canonical fold structure "substantially identical" to a canonical fold structure known to occur in human antibodies if at least the first and preferably all of the following criteria are met:

1. the most closely matched human canonical structure type has the same length, determined by the number of residues.

2. At least 33% identity, preferably at least 50% identity, to the key amino acid residues described for the corresponding human H1 and H2 canonical structure types.

(Note: for purposes of the foregoing analysis, H1 and H2 rings were treated separately and each was compared against its most closely matching human canonical structure type)

The foregoing analysis relies on the prediction of canonical structures of the H1 and H2 loops of the antibody of interest. If the actual structures of the H1 and H2 loops in the antibody of interest are known, e.g., based on X-ray crystallography, the H1 and H2 loops in the antibody of interest can also be evaluated as having "substantially the same" canonical fold structures as canonical fold structures known to occur in human antibodies if the length of the loops is different (typically by ± 1 or ± 2 amino acids) from the length of the most closely matching human canonical structure type, but the actual structures of the H1 and H2 loops in the antibody of interest match the human canonical fold structures.

The key amino acid residues found in the human canonical structural types for the first and second hypervariable loops (H1 and H2) of the human VH domain are described in Chothia et al, J.mol.biol.227:799-817(1992), which is incorporated herein by reference in its entirety. Specifically, Chothia et al, 802 Table 3 (which is specifically incorporated herein by reference) lists the preferred amino acid residues at key positions for the H1 canonical structure found in the human germline, while 803 Table 4 (which is also specifically incorporated herein by reference) lists the preferred amino acid residues at key positions for the CDR H2 canonical structure found in the human germline.

In one embodiment, both H1 and H2 in the VH domain of an antibody with high human homology exhibit a predicted or actual canonical fold structure that is substantially identical to the canonical fold structure that occurs in human antibodies.

Antibodies with high human homology may comprise VH domains in which hypervariable loops H1 and H2 form a combination of canonical fold structures identical to the combination of canonical structures known to occur in at least one human germline VH domain. It has been observed that only certain combinations of canonical fold structures at H1 and H2 actually occur in VH domains encoded by the human germline. In one embodiment, H1 and H2 in the VH domain of an antibody with high human homology may be obtained from the VH domain of a non-human species, such as a species in the family camelidae, but which form the same combination of predicted or actual canonical fold structures as are known to occur in the VH domain of human germline or somatic mutations. In non-limiting embodiments, H1 and H2 in the VH domain of an antibody with high human homology may be obtained from VH domains of non-human species, such as species in the family camelidae, and which form the following canonical fold combinations: 1-1, 1-2, 1-3, 1-6, 1-4, 2-1, 3-1 and 3-5.

Antibodies with high human homology may contain VH domains that exhibit high sequence identity/sequence homology to human VH and contain hypervariable loops that exhibit structural homology to human VH.

The following may be advantageous: the canonical folds present at H1 and H2, and combinations thereof, in the VH domain of antibodies with high human homology are "correct" for the human VH germline sequence representing the closest match in overall primary amino acid sequence identity to the VH of the antibody with high human homology. For example, if the closest sequence match is to a human germline VH3 domain, then the combination of canonical folds formed by H1 and H2 may also occur naturally in the human VH3 domain to be advantageous. This may be particularly important where the antibody with high human homology is derived from a non-human species, for example an antibody containing VH and VL domains derived from a camelid conventional antibody, in particular an antibody containing humanized camelid VH and VL domains.

Thus, in one embodiment, the VH domain of an IL-6 antibody with high human homology may exhibit 80% or more, 85% or more, 90% or more, 95% or more, 97% or more, or up to 99% or even 100% sequence identity or sequence homology to the human VH domain in the framework regions FR1, FR2, FR3 and FR4, and further, H1 and H2 in the same antibody are obtained from a non-human VH domain (e.g. derived from a species in the family camelidae), but the combination of predicted or actual canonical fold structures formed is identical to a canonical fold combination known to occur naturally in the same human VH domain.

For example, in one exemplary embodiment, the H1 and H2 loops of an IL-6 antibody (e.g., 61H7) of the invention can comprise human canonical fold structure combinations 1-2 as found, for example, in human antibody structure 1 DFB. In another exemplary embodiment, the H1 and H2 loops of an IL-6 antibody of the invention (e.g., 68F2 or a germlined variant thereof 129D3) can comprise human canonical fold structure combination 3-1 as found, for example, in human antibody structure 1 ACY.

In other embodiments, both L1 and L2 in the VL domain of an antibody with high human homology are obtained from a VL domain of a non-human species (e.g., a camelid-derived VL domain) and each exhibits a predicted or actual canonical fold structure that is substantially identical to the canonical fold structure present in human antibodies.

Like the VH domain, the hypervariable loops of the VL domain of both the V lambda and V kappa types can adopt a limited number of conformations or canonical structures, determined in part by length, and also by the presence of key amino acid residues at certain canonical positions.

Within an antibody of interest having high human homology, loops L1, L2 and L3 obtained from the VL domain of a non-human species, such as camelidae species, can be assessed as having a canonical fold structure "substantially identical" to the canonical fold structure known to occur in human antibodies if at least the first and preferably all of the following criteria are met:

1. the most closely matched human structural types have the same length, determined by the number of amino acid residues.

2. At least 33% identical, preferably at least 50% identical, to the key amino acid residues described for the corresponding human L1 or L2 canonical structure types from the V lambda or V kappa repertoire.

(Note: for purposes of the foregoing analysis, the L1 and L2 loops were treated separately and each was compared against its most closely matching human canonical structure type).

The foregoing analysis relies on the prediction of canonical structures for the L1, L2, and L3 loops of the VL domain of the antibody of interest. If the actual structure of the L1, L2 and L3 loops is known, for example based on X-ray crystallography, the L1, L2 or L3 loops derived from the antibody of interest can also be evaluated as having a canonical fold structure "substantially identical" to the canonical fold structure known to occur in human antibodies if the length of the loops is different (typically by ± 1 or ± 2 amino acids) from the length of the most closely matching human canonical structure type, but the actual structure of the camelid loop matches human canonical fold.

The key amino acid residues found in the human canonical structural classes of CDRs for the human V lambda and V kappa domains are described in Morea et al Methods, 20:267-279(2000) and Martin et al, J.mol.biol., 263:800-815 (1996). The structural repertoire of human V kappa domains is also described in Tomlinson et al EMBO J.14:4628-4638(1995), while the structural repertoire of V lambda domains is described in Williams et al J.mol.biol., 264:220-232 (1996). The contents of all these documents are incorporated herein by reference.

L1 and L2 in the VL domain of antibodies with high human homology can form a combination of predicted or actual canonical fold structures that is identical to the combination of canonical fold structures known to occur in human germline VL domains. In non-limiting embodiments, L1 and L2 in the V lambda domain of an antibody having high human homology (e.g., an antibody containing a camelid-derived VL domain or a humanized variant thereof) can form one of the following canonical fold combinations: 11-7, 13-7(A, B, C), 14-7(A, B), 12-11, 14-11 and 12-12 (as defined in Williams et al J.mol.biol.264:220-32 (1996)) and as shown in http:// www.bioc.uzh.ch/antibody/Sequences/gelmins/VBase _ hVL. In non-limiting embodiments, L1 and L2 in the V kappa domain may form one of the following canonical fold combinations: 2-1, 3-1, 4-1 and 6-1 (as defined in Tomlinson et al EMBO J.14:4628-38(1995) and as shown in http:// www.bioc.uzh.ch/antisense/Sequences/Germlines/VBase _ hVK.html). For example, in one exemplary embodiment, the L1 and L2 loops of an IL-6 antibody (e.g., 61H7) of the invention may comprise the human canonical fold structure combination 7 λ -1 as found, for example, in human antibody structure 1 MFA. In another exemplary embodiment, the L1 and L2 loops of an IL-6 antibody of the invention (e.g., 68F2 or a germlined variant thereof 129D3) can comprise the human canonical fold structure combination 6 λ -1 as found, for example, in the human antibody structure 3 MUG.

In yet another embodiment, all three of L1, L2, and L3 in the VL domain of an antibody with high human homology may exhibit a substantially human structure. Preferably, the VL domain of an antibody with high human homology exhibits high sequence identity/sequence homology to a human VL, and also preferably the hypervariable loops in the VL domain exhibit structural homology to a human VL. For example, in one exemplary embodiment, loop L1-L3 of an IL-6 antibody (e.g., 61H7) of the invention can comprise the human canonical fold structure combination 7 λ -1-4 as found, for example, in human antibody structure 1 MFA. In another exemplary embodiment, the L1-L3 loop of an IL-6 antibody of the invention (e.g., 68F2 or a germlined variant thereof 129D3) can comprise the human canonical fold structure combination 6 λ -1-5 as found, for example, in the human antibody structure 3 MUG.

In one embodiment, the VL domain of an IL-6 antibody with high human homology may exhibit a sequence identity with the human VL domain of 80% or more, 85% or more, 90% or more, 95% or more, 97% or more, or up to 99% or even 100% in framework regions FR1, FR2, FR3 and FR4, and furthermore hypervariable loop L1 and hypervariable loop L2 may form a combination of predicted or actual canonical fold structures which is identical to a canonical fold combination known to occur naturally in the same human VL domain.

It is apparent that it is contemplated that a VH domain exhibiting high sequence identity/sequence homology to human VH and also having structural homology to the hypervariable loops of human VH, and a VL domain exhibiting high sequence identity/sequence homology to human VL and also having structural homology to the hypervariable loops of human VL, are combined to provide an antibody having high human homology comprising a VH/VL pair having maximum sequence and structural homology to a human-encoded VH/VL pair (e.g., a camelid-derived VH/VL pair).

iii) non-immunoglobulin binding molecules

In yet another aspect, the invention provides a non-immunoglobulin binding molecule that specifically binds to IL-6. As used herein, the term "non-immunoglobulin binding molecule" is a binding molecule whose binding site comprises a portion (e.g., a scaffold or framework) derived from a polypeptide other than an immunoglobulin, but which can be engineered (e.g., by addition of CDR region sequences) to confer a desired binding specificity to the binding molecule. The non-immunoglobulin binding molecules of the invention typically comprise one or more of the CDR regions listed in tables 13-18, which are removed from the non-immunoglobulin polypeptide.

In certain embodiments, the non-immunoglobulin binding molecule comprises a binding site portion derived from a member of the immunoglobulin superfamily that is not an immunoglobulin (e.g., a T cell receptor or cell adhesion protein (e.g., CTLA-4, N-CAM, telokinin)). Such binding molecules comprise a binding site portion that maintains the folded conformation of an immunoglobulin and is capable of specifically binding to IL-6 when modified to comprise one or more of the CDR regions listed in tables 13-18. In other embodiments, the non-immunoglobulin binding molecules of the invention comprise a binding site having a non-immunoglobulin fold-based protein topology (e.g., ankyrin repeat, tetranectin, and fibronectin), but nonetheless are capable of specifically binding to a target (e.g., IL-6) when modified to comprise one or more of the CDR regions listed in tables 13-18.

In one embodiment, the binding molecule of the invention comprises a tetranectin molecule. Tetranectin is a plasma protein with a trivalent structure. Each monomer of the tetranectin trimer comprises five unique amino acid loops that can be replaced or engineered to comprise antibody CDR sequences (e.g., CDR regions listed in tables 13-18). Methods of making tetranectin-binding polypeptides are described, for example, in US20110086770, which is incorporated herein by reference in its entirety.

In one embodiment, the binding molecule of the invention comprises a fibronectin molecule. Fibronectin binding molecules (e.g., molecules comprising fibronectin type I, II, or III domains) display CDR-like loops that can be replaced or engineered to include antibody CDR sequences (e.g., CDR regions listed in tables 13-18). Methods of making fibronectin binding polypeptides are described, for example, in WO01/64942 and U.S. patent nos. 6,673,901, 6,703,199, 7,078,490, and 7,119,171, each of which is incorporated herein by reference in its entirety.

In another embodiment, the binding molecule of the invention comprises a binding site from an affinity antibody (affibody). Affinity antibodies are derived from the immunoglobulin binding domain of staphylococcal protein (A) (SPA) (see, e.g., Nord et al, nat. Biotechnol., 15:772-777 (1997)). The affinity antibody binding sites employed in the present invention can be synthesized by mutagenesis of a SPA-related protein (e.g., protein Z) derived from a domain of SPA (e.g., domain B), and selection for mutant SPA-related polypeptides that have binding affinity for IL-6. Other methods for making affinity antibody binding sites are described in U.S. patent nos. 6,740,734 and 6,602,977 and WO00/63243, each of which is incorporated herein by reference.

In another embodiment, the binding molecule of the invention comprises a binding site from an anticalin. anticalins (also known as lipocalins) are members of a diverse family of β -barrel (beta-barrel) proteins that function to bind target molecules in their barrel/loop regions. The lipocalin binding site can be engineered to bind IL-6, i.e., the loop sequence of the strand connecting the barrel is randomized (see, e.g., Schlehuber et al, Drug Discov. today, 10:23-33 (2005); Beste et al, PNAS, 96: 1898-. The anticalin binding sites employed in the binding molecules of the invention may be obtained first from polypeptides of the lipocalin family which are mutated in four segments corresponding to the linear polypeptide sequence including the sequence positions 28 to 45, 58 to 69, 86 to 99 and 114 to 129 of the bile pigment (bilin) binding protein (BBP) of Pieris european (Pieris brassica). Other methods for preparing anticalin binding sites are described in WO99/16873 and WO05/019254, each of which is incorporated herein by reference.

In another embodiment, the binding molecule of the invention comprises a binding site from a cysteine-rich polypeptide. Cysteine-rich domains employed in the practice of the invention do not typically form alpha-helices, beta-sheets or beta-barrel structures. Typically, disulfide bonds promote the folding of this domain into a three-dimensional structure. Typically, the cysteine-rich domain has at least two disulfide bonds, more typically at least three disulfide bonds. An exemplary cysteine-rich polypeptide is an a-domain protein. The A domain (sometimes referred to as a "complement-type repeat") contains about 30-50 or 30-65 amino acids. In some embodiments, the domain comprises about 35-45 amino acids, and in some cases about 40 amino acids. Within the 30-50 amino acids, there are about 6 cysteine residues. Of the six cysteines, disulfide bonds are typically formed between the following cysteines: c1 and C3, C2 and C5, C4 and C6. The a domain constitutes a ligand binding module (motif). The cysteine residues of this domain are linked by disulfide bonds to form a compact, stable, functionally independent module. Clusters of these repeats constitute the ligand binding domain, and differential clustering (differential clustering) can confer specificity for ligand binding. Exemplary proteins containing the a domain include, for example, complement components (e.g., C6, C7, C8, C9, and factor I), serine proteases (e.g., enteropeptidases, matriptase, and corin), transmembrane proteins (e.g., ST7, LRP3, LRP5, and LRP6), and endocytic receptors (e.g., Sortilin-related receptors, LDL receptors, VLDLR, LRP1, LRP2, and ApoER 2). Methods for preparing A-domain proteins with desired binding specificities are disclosed in, for example, WO02/088171 and WO04/044011, each of which is incorporated herein by reference.

In other embodiments, the binding molecules of the invention comprise a binding site from a repeat protein. Repetitive proteins are proteins that contain successive copies of small (e.g., about 20 to about 40 amino acid residues) structural units or repeats that are stacked upon one another to form a continuous domain. Repeat proteins can be modified to accommodate a particular target binding site by adjusting the number of repeats in the protein. Exemplary repeat proteins include the designed ankyrin repeat protein (i.e., DARPin) (see, e.g., Binz et al, Nat. Biotechnol., 22: 575-. The tertiary structure of all ankyrin repeats identified so far is characterized by a β -hairpin followed by two antiparallel α -helices and finally by a loop connecting the repeat to the next. The ankyrin repeat constructed domain is formed by stacking the repeats as an extended and bent structure. The LRRP binding site forms part of the acquired immune system from sea lamprey and other jawbarvens and is similar to antibodies in that it is formed by recombining a set of leucine rich repeats during lymphocyte maturation. Methods for preparing DARpin or LRRP binding sites are described in WO02/20565 and WO06/083275, each of which is incorporated herein by reference.

Other non-immunoglobulin binding sites that may be employed in the binding molecules of the invention include binding sites derived from Src homology domains (e.g., SH2 or SH3 domains), PDZ domains, beta-lactamase, high affinity protease inhibitors, or small disulfide binding protein scaffolds such as scorpion toxin. Methods for preparing binding sites derived from these molecules have been disclosed in the art, see, e.g., Panni et al, J.biol.chem., 277: 21666-containing 21674(2002), Schneider et al, nat.Biotechnol., 17: 170-containing 175 (1999); legendre et al, Protein Sci, 11: 1506-; stoop et al, nat. Biotechnol., 21:1063-1068 (2003); and Vita et al, PNAS, 92: 6404-. Still other binding sites may be derived from binding domains selected from the group consisting of: EGF-like domain, Kringle domain, PAN domain, Gla domain, SRCR domain, Kunitz/bovine trypsin inhibitor domain, Kazal-type serine protease inhibitor domain, Trefoil (P-type) domain, von Willebrand factor type C domain, anaphylatoxin-like domain, CUB domain, thyroglobulin type I repeat, LDL receptor type a domain, Sushi domain, Link domain, Thrombospondin (Thrombospondin) type I domain, immunoglobulin-like domain, C-type lectin domain, MAM domain, von Willebrand factor type a domain, growth regulator B domain, WAP-type four disulfide core (four disulfide core) domain, F5/8 type C domain, blood bindin (Hemopexin) domain, laminin-type EGF-like domain, C2 domain and such other domains known to those of ordinary skill in the art, as well as derivatives and/or variants thereof.

Non-immunoglobulin binding molecules can be identified by selecting or isolating target binding variants from a library of binding molecules having artificially diversified binding sites. The variegated library may be generated by grouping into a library of CDR sequences (e.g., CDR sequences selected from those listed in tables 13-18) and/or entirely random methods (e.g., error-prone PCR, exon shuffling, or directed evolution) and/or by art-recognized design strategies. For example, the amino acid positions that are typically involved when a binding site interacts with its associated target molecule can be randomized by inserting degenerate codons, trinucleotides, random peptides, or entire loops at the corresponding positions within the nucleic acid encoding the binding site (see, e.g., U.S. publication No. 20040132028). The location of the amino acid position can be identified by investigating the crystal structure of the binding site that is complexed with the target molecule. Candidate positions for CDR sequence grouping (e.g., CDR sequences selected from those listed in tables 13-18) and/or randomization include binding cavities, helices, flat surfaces, and loops of the binding site. In certain embodiments, amino acids in the binding site that are likely candidates for diversification may be identified by their homology to the immunoglobulin fold. For example, residues within the CDR-like loops of fibronectin can be randomized to generate libraries of fibronectin binding molecules (see, e.g., Koide et al, J.Mol.biol., 284:1141-1151 (1998)). After the incorporation of CDR sequences (e.g., selected from those listed in tables 2-6) and/or randomization, the variegated library can then be subjected to a selection or screening step to obtain a binding molecule having the desired binding characteristics, e.g., specific binding to IL-6. Selection can be achieved by art-recognized methods such as phage display, yeast display, or nucleic acid display.

Germlining of camelid-derived VH and VL domains

Conventional antibodies to camelids provide an advantageous starting point for the preparation of antibodies useful as human therapeutics, due to the following factors (discussed in US12/497,239, which is incorporated herein by reference in its entirety):

1) high% sequence homology between camelid VH and VL domains and their human congeners;

2) high structural homology (i.e., a combination of human-like canonical fold structure and human-like canonical fold) between the CDRs of camelid VH and VL domains and their human congeners.

The camelid (e.g., llama) platform also provides significant advantages in the functional diversity of available IL-6 antibodies.

The usefulness of IL-6 antibodies comprising camelid VH and/or camelid VL domains for human therapy can still further be improved, e.g., rendered less immunogenic in a human host, by "germlining" the native camelid VH and VL domains. The overall goal of germlining is to produce a molecule in which the VH and VL domains exhibit minimal immunogenicity when introduced into a human subject, while retaining the specificity and affinity of the antigen binding site formed by the parent VH and VL domains.

Determining homology between a camelid VH (or VL) domain and a human VH (or VL) domain is a key step in this germlining process, whether for selecting the camelid amino acid residue to be altered (in a given VH or VL domain) or for selecting the appropriate alternative amino acid side chain.

A method of germlining camelids conventional antibodies has been developed based on alignment of a number of novel camelid VH (and VL) domain sequences (typically somatically mutated VH (or VL) domains known to bind target antigens) with human germline VH (or VL) sequences, human VH (and VL) consensus sequences, and germline sequence information available for alpacas (llama pacos).

The principles outlined in the following paragraphs may be applied to (i) selecting "camelid" amino acid sequences for substitution in a camelid-derived VH or VL domain or CDRs thereof, and (ii) selecting alternative "human" amino acid sequences for substitution when germlining any given camelid VH (or VL) domain. This method can be used to prepare germline variants of the VH and VL sequences listed in tables 13-16 herein.

Step 1: the human (germline) family and the members of this family are selected which show the highest homology/identity to the mature camelid sequence to be germlined. The general steps for identifying the closest matching human germline for any given camelid VH (or VL) domain are outlined below.

Step 2: specific human germline family members were selected for germline against. Preferably, this is the germline with the highest homology, or another germline family member from the same family.

And step 3: based on the amino acid utilization table of the camelid germline closest to the selected human germline, the positions deemed to be preferred for germlining are identified.

And 4, step 4: attempting to alter amino acids in the camelid germline that deviate from the closest human germline; germlining FR residues is preferred over CDR residues.

a. Preferably, the positions that deviate from the selected human germline used for germlining reference for which the amino acids found in the camelid sequence do not match the selected germline and are not found in other germline of the same subtype (both for V and for J-encoded FR amino acids).

b. It is also possible to locate (address) deviations from selected human germline family members during the germlining process, but for the location of other germline of the same family.

c. Other mismatches to the selected human germline can also be located (e.g., due to additional somatic mutations).

The following method may be used to determine the closest matching human germline for a given camelid VH (or VL) domain:

prior to analysis of percent sequence identity between camelidae and human germline VH and VL, canonical folds may first be determined, which allows the identification of families of human germline segments with combinations of the same canonical folds for H1 and H2 or L1 and L2 (and L3). Next, the human germline family member having the highest degree of sequence homology with the camelidae variable region of interest may be selected for scoring of sequence homology. Determination of the Chothia specification types of the hyper-variable rings L1, L2, L3, H1, and H2 may be performed using bioinformatics tools publicly available on web pages www.bioinf.org.uk/abs/Chothia. The output of this program shows the key residue requirements in the data file. In these data files, key residue positions are shown, as well as allowed amino acids at each position. The sequences of the variable regions of the antibodies are given as inputs and are first aligned to a common antibody sequence to assign (assign) the Kabat numbering scheme. Analysis of canonical folds used a set of key residue templates derived from automated methods developed by Martin and Thornton (Martin et al, J.mol.biol.263:800-815 (1996)). The boundaries of the individual framework regions can be assigned using the IMGT numbering scheme, which is adapted from Chothia's coding scheme (Lefranc et al, NAR27:209-212 (1999); IMGT. channels. fr).

Where specific human germline V segments are known, which use the same canonical fold combination for H1 and H2 or L1 and L2 (and L3), the best matching family members in terms of sequence homology can be determined. The percent sequence identity between camelid VH and VL domain framework amino acid sequences and corresponding sequences encoded by the human germline can be determined using bioinformatics tools, but manual alignment of the sequences can also be used. Human immunoglobulin Sequences can be identified from several protein databases, such as VBase (VBase. mrc-cpe. cam. ac. uk /) or Pluckthun/honeyger databases (http:// www.bioc.unizh.ch/antibody/Sequences/Germlines). To compare human sequences to V regions of Camelidae VH or VL domains, sequence alignment algorithms such as those available through websites such as www.expasy.ch/tools/# align may be used, although manual alignment may also be performed for a limited set of sequences. Human germline light and heavy chain sequences of families with the same canonical fold combination and the highest degree of homology to the framework regions 1, 2 and 3 of each chain can be selected and compared to camelidae variable regions of interest; FR4 was also examined for human germline JH and JK or JL regions.

Note that in the calculation of overall percent sequence homology, the residues of FR1, FR2 and FR3 were evaluated using the most closely matched sequences from human germline families with the same canonical fold combination. Only residues that differ from other members of the same family that most closely match or have the same canonical fold combination are scored (NB-excluding any primer-encoded differences). However, for germlining purposes, it is contemplated that residues in the same framework regions as members of other human germline families (which do not have the same canonical fold combinations) will be germlined, even though these residues have a "negative" score according to the stringent conditions described above. This assumption is based on the "mix and match" approach to germlining, in which FR1, FR2, FR3 and FR4 are compared with their most closely matching human germline sequences, respectively, so that the germlined molecules contain combinations of different FRs, as implemented by Qu and its colleagues (Qu et al, Clin. cancer Res.5: 3095-.

Modified binding molecules

In certain embodiments, a binding polypeptide of the invention may comprise one or more modifications. Modified forms of the binding polypeptides of the invention can be prepared using any technique known in the art.

i) Reducing immunogenic risk

In certain embodiments, the binding molecules of the invention (e.g., antibodies or antigen-binding fragments thereof) are modified using art-recognized techniques to further reduce their immunogenic risk. For example, the antibody or fragment thereof may be germlined according to the methods described above. Alternatively, the binding molecules of the invention may be chimerized, humanized, and/or deimmunized (deimmunize).

In one embodiment, an antibody or antigen-binding fragment thereof of the invention can be chimeric. Chimeric antibodies are antibodies in which different portions of the antibody are derived from different animal species, such as antibodies having variable regions derived from camelid (e.g., llama) monoclonal antibodies and human immunoglobulin constant regions. Methods of producing chimeric antibodies or fragments thereof are known in the art. See, e.g., Morrison, Science229:1202 (1985); oi et al, BioTechniques4:214 (1986); gillies et al, J.Immunol.Methods125:191-202 (1989); U.S. patent nos. 5,807,715; 4,816,567; and 4,816,397, all of which are incorporated herein by reference in their entirety. Techniques developed for the generation of "chimeric antibodies" (Morrison et al, Proc. Natl. Acad. Sci.81:851-855 (1984); Neuberger et al, Nature312:604-608 (1984); Takeda et al, Nature314:452-454(1985)) can be used to generate the molecules. For example, genetic sequences encoding the binding specificity of camelid anti-IL-6 antibody molecules can be fused to sequences from human antibody molecules having suitable biological activity. As used herein, a chimeric antibody is a molecule in which different portions are derived from different animal species, such as those having a variable region derived from a camelid (e.g., llama) monoclonal antibody and a human immunoglobulin constant region, e.g., a germlined or humanized antibody.

In another embodiment, the antibodies of the invention, or antigen binding portions thereof, are humanized. Humanized antibodies have a binding specificity comprising one or more Complementarity Determining Regions (CDRs) from a non-human antibody and a framework region from a human antibody molecule. Typically, framework residues in the human framework regions are substituted with corresponding residues from a CDR donor antibody to alter, preferably improve, antigen binding. These framework substitutions are identified by methods well known in the art, for example by modeling the interaction of the CDRs and framework residues to identify framework residues important for antigen binding, and performing sequence comparisons to identify unusual framework residues at specific positions (see, e.g., Queen et al, U.S. Pat. No. 5,585,089; Riechmann et al, Nature332:323(1988), which is incorporated herein by reference in its entirety). Antibodies can be humanized using a variety of techniques known in the art, including CDR grafting (EP239,400; PCT publication WO 91/09967; U.S. Pat. No. 5,225,539; 5,530,101; and 5,585,089), veneering (tunneling) or resurfacing (EP592,106; EP519,596; Padlan, Molecular Immunology28(4/5):489-498 (1991); Studnicka et al, Protein Engineering7(6):805-814 (1994); Roguska et al, PNAS91:969-973(1994)), and chain shuffling (U.S. Pat. No. 5,565,332).

In some embodiments, deimmunization can be used to further reduce the risk of immunogenicity of an IL-6 binding molecule (e.g., an antibody or antigen-binding portion thereof). As used herein, the term "deimmunization" includes altering a polypeptide (e.g., an antibody or antigen binding portion thereof) to modify a T cell epitope (see, e.g., WO9852976a1, WO0034317a 2). For example, VH and VL sequences from a starting IL-6 specific antibody or antigen binding portion thereof of the invention can be analyzed, and a human T cell epitope "map (map)" can be generated from each V region, showing the position of the epitope relative to the Complementarity Determining Regions (CDRs) and other critical residues within the sequence. Individual T cell epitopes from the T cell epitope map are analyzed to identify alternative amino acid substitutions with low risk of altering the final antibody activity. A series of alternative VH and VL sequences comprising combinations of amino acid substitutions are designed, then these sequences are grouped into a series of IL-6 specific antibodies or fragments thereof for use in the diagnostic and therapeutic methods disclosed herein, and then tested for function. Typically, 12 to 24 variant antibodies are generated and tested. The complete heavy and light chain genes containing the modified V and human C regions are then cloned into an expression vector and subsequent plasmids are introduced into cell lines for the production of whole antibodies. The antibodies are then compared in appropriate biochemical and biological assays and the optimal variant is identified.

ii) effector function and Fc modification

In certain embodiments, a binding molecule of the invention can comprise an antibody constant region (e.g., an IgG constant region, e.g., a human IgG1 or IgG4 constant region) that mediates one or more effector functions. For example, binding of the C1 component of complement to the constant region of an antibody activates the complement system. Activation of complement is important for opsonization and lysis of cellular pathogens. Activation of complement also stimulates inflammatory responses and may also be involved in autoimmune hypersensitivity. In addition, antibodies bind to receptors on a variety of cells via the Fc region, wherein the Fc receptor binding site on the Fc region of the antibody binds to an Fc receptor (FcR) on the cell. There are a number of Fc receptors, which are specific for different types of antibodies, including IgG (gamma receptor), IgE (epsilon receptor), IgA (alpha receptor) and IgM (mu receptor). Binding of antibodies to Fc receptors on cell surfaces triggers a variety of important and diverse biological responses, including phagocytosis (engulfment) and destruction of antibody-coated particles, clearance of immune complexes, lysis of antibody-coated target cells by killer cells (known as cell-dependent cell-mediated cytotoxicity or ADCC), release of inflammatory mediators, placental transfer, and control of immunoglobulin production. In a preferred embodiment, a binding molecule of the invention (e.g., an antibody or antigen-binding fragment thereof) binds to an Fc-gamma receptor. In other embodiments, a binding molecule of the invention may comprise a constant region that lacks one or more effector functions (e.g., ADCC activity) and/or is unable to bind to an Fc γ receptor.

Certain embodiments of the invention include anti-IL-6 antibodies in which at least one amino acid in one or more constant regions is deleted, or otherwise altered, to provide a desired biochemical characteristic such as reduced or enhanced effector function, the ability to non-covalently dimerize, increased ability to localize at a tumor lesion, reduced serum half-life, or increased serum half-life when compared to a complete, unaltered antibody having about the same immunogenicity. For example, certain antibodies or fragments thereof for use in the diagnostic and therapeutic methods described herein are domain deleted antibodies comprising polypeptide chains similar to immunoglobulin heavy chains, but lacking at least a portion of one or more heavy chain domains. For example, in certain antibodies, the entire domain of the constant region of the modified antibody may be deleted, e.g., all or part of the CH2 domain may be deleted.

In certain other embodiments, the binding molecule comprises constant regions derived from different antibody isotypes (e.g., constant regions from two or more of human IgG1, IgG2, IgG3, or IgG 4). In other embodiments, the binding molecule comprises a chimeric hinge (i.e., a hinge portion comprising hinge domains derived from different antibody isotypes, such as a hinge from the upper hinge domain of an IgG4 molecule and the middle hinge domain of IgG 1). In one embodiment, the binding molecule comprises an Fc region or portion thereof from a human IgG4 molecule and a Ser228Pro mutation (EU numbering) in the core hinge region of the molecule.

In certain embodiments, the Fc portion can be mutated to increase or decrease effector function using techniques known in the art. For example, deletion or inactivation of the constant region (by point mutation or other means) can reduce Fc receptor binding of the circulating modified antibody, thereby increasing tumor localization. In other cases, it may modulate (operate) complement binding for constant region modification consistent with the present invention, and thus reduce serum half-life and non-specific association of conjugated cytotoxins. Still other modifications of the constant region may be used to modify disulfide linkages or oligosaccharide modules, which allow for enhanced localization due to increased antigen specificity or flexibility. The resulting physiological profile, bioavailability and other biochemical effects such as tumor localization, biodistribution and serum half-life of the modifications can be readily measured and quantified using known immunological techniques without undue experimentation.

In certain embodiments, the Fc domain of the antibodies used in the invention is an Fc variant. As used herein, the term "Fc variant" refers to an Fc domain having at least one amino acid substitution relative to the wild-type Fc domain from which the Fc domain is derived. For example, wherein the Fc domain is derived from a human IgG1 antibody, the Fc variant of the human IgG1Fc domain comprises at least one amino acid substitution relative to the Fc domain.

Amino acid substitutions of the Fc variant can be located at any position within the Fc domain (i.e., EU regular amino acid positions). In one embodiment, the Fc variant comprises a substitution at an amino acid position located in the hinge domain or portion thereof. In another embodiment, the Fc variant comprises a substitution at the amino acid position located in the CH2 domain or portion thereof. In another embodiment, the Fc variant comprises a substitution at the amino acid position located in the CH3 domain or portion thereof. In another embodiment, the Fc variant comprises a substitution at the amino acid position located in the CH4 domain or portion thereof.

The binding molecules of the invention may employ any art-recognized Fc variant known to confer improved (e.g., reduced or enhanced) effector function and/or FcR binding. Such Fc variants may include, for example, any of the Fc variants disclosed in International PCT publications WO88/07089A1, WO96/14339A1, WO98/05787A1, WO98/23289A1, WO99/51642A1, WO99/58572A1, WO00/09560A2, WO00/32767A1, WO00/42072A2, WO02/44215A2, WO02/060919A2, WO03/074569A2, WO04/016750A2, WO 2/029207A 2, WO 2/2A 2, WO 2/2A 2, WO 2/2A 2, WO 2/2A 2, WO 2A 2/2A 2, WO 2A 2/2A 2, WO 2A2, WO 2/2A 2, WO 2/2A 2, WO 2A 36; 5,739,277; 5,834,250; 5,869,046; 6,096,871, respectively; 6,121,022; 6,194,551; 6,242,195, respectively; 6,277,375; 6,528,624, respectively; 6,538,124, respectively; 6,737,056; 6,821,505, respectively; 6,998,253, respectively; and 7,083,784 (each of which is incorporated herein by reference). In an exemplary embodiment, a binding polypeptide of the invention can comprise an Fc variant comprising an amino acid substitution at EU position 268 (e.g., H268D or H268E). In another exemplary embodiment, a binding polypeptide of the invention can comprise an amino acid substitution at EU position 239 (e.g., S239D or S239E) and/or EU position 332 (e.g., I332D or I332Q).

In certain embodiments, a binding polypeptide of the invention may comprise an Fc variant comprising an amino acid substitution that alters the antigen-independent effector function of the antibody, particularly the circulating half-life of the binding polypeptide. Such binding molecules exhibit increased or decreased binding to FcRn when compared to binding molecules lacking these substitutions, and thus have increased or decreased half-life in serum, respectively. Fc variants with improved affinity for FcRn are expected to have longer serum half-lives, and such molecules have useful uses in methods of treating mammals where the antibody half-life to be administered is longer (e.g. to treat chronic diseases or disorders). In contrast, Fc variants with reduced FcRn binding affinity are expected to have shorter half-lives, and such molecules may be used for administration to mammals, for example when reduced circulation time is advantageous (e.g. for in vivo diagnostic imaging, or when the starting antibody has toxic side effects when present in circulation for a longer time). Fc variants with reduced FcRn binding affinity are also less likely to cross the placenta and therefore may also be useful in the treatment of diseases or disorders in pregnant women. In addition, other applications where reduced FcRn binding affinity may be desired include applications where localization to the brain, kidney and/or liver is desired. In an exemplary embodiment, the altered binding molecules (e.g., antibodies or antigen-binding fragments thereof) of the invention exhibit reduced transport across the epithelium of the glomerulus from the vasculature. In another embodiment, the altered binding molecules (e.g., antibodies or antigen-binding fragments thereof) of the invention exhibit reduced transport across the Blood Brain Barrier (BBB) into the vascular space from the brain. In one embodiment, an antibody with altered FcRn binding comprises an Fc domain having one or more amino acid substitutions within the "FcRn binding loop" of the Fc domain. The FcRn binding loop comprises amino acid residues 280-299 (numbering according to EU). Exemplary amino acid substitutions that alter FcRn binding activity are disclosed in international PCT publication No. WO05/047327, which is incorporated herein by reference. In certain exemplary embodiments, a binding molecule (e.g., an antibody or antigen-binding fragment thereof) of the invention comprises an Fc domain having one or more of the following substitutions: V284E, H285E, N286D, K290E and S304D (EU numbering). In yet another exemplary embodiment, the binding molecules of the invention comprise a human Fc domain with the double mutation H433K/N434F (see, e.g., U.S. patent No. 8,163,881). In a specific embodiment, the binding molecule of the invention comprises one or more variable regions selected from table 16 and a human Fc domain with the double mutation H433K/N434F. In another specific embodiment, the binding molecule of the invention comprises one or more CDR sequences from table 17 and a human Fc domain with the double mutation H433K/N434F. In yet another specific embodiment, the binding molecule of the invention comprises a VH domain of SEQ ID No.152 and a human Fc domain with the double mutation H433K/N434F. In yet another specific embodiment, the binding molecule of the invention comprises a VH domain of SEQ ID No.152, in which glutamine at one or more positions (e.g. positions 1, 3, 5 or 16 or any combination thereof) has been changed to glutamate, and a human Fc domain with the double mutation H433K/N434F. In yet another specific embodiment, the binding molecule of the invention comprises a VH domain of SEQ ID No.152, in which the glutamine at position 1 has been changed to glutamic acid, and a human Fc domain with the double mutation H433K/N434F.

In other embodiments, the binding molecules for use in the diagnostic and therapeutic methods described herein have a constant region, such as an IgG1 or IgG4 heavy chain constant region, that is altered to reduce or eliminate glycosylation. For example, a binding molecule of the invention (e.g., an antibody or antigen-binding fragment thereof) can also comprise an Fc variant comprising an amino acid substitution that alters glycosylation of the Fc of the antibody. For example, the Fc variant may have reduced glycosylation (e.g., N-or O-linked glycosylation). In an exemplary embodiment, the Fc variant comprises a reduction in glycosylation of the N-linked glycan as is commonly seen at amino acid position 297(EU numbering). In another embodiment, the antibody has an amino acid substitution near or within a glycosylation motif, such as an N-linked glycosylation motif (e.g., an N-linked glycosylation motif comprising the amino acid sequence NXT or NXS). In a specific embodiment, the antibody comprises an Fc variant having an amino acid substitution at amino acid position 228 or 299(EU numbering). In more specific embodiments, the antibody comprises an IgG1 or IgG4 constant region comprising the S228P and T299A mutations (EU numbering).

Exemplary amino acid substitutions that confer reduced or altered glycosylation are disclosed in International PCT publication No. WO05/018572, which is incorporated herein by reference. In a preferred embodiment, the antibody or fragment thereof of the invention is modified to eliminate glycosylation. Such antibodies or fragments thereof may be referred to as "deglycosylated" antibodies or fragments thereof (e.g., "deglycosylated" antibodies). While not being bound by theory, it is believed that "deglycosylated" antibodies or fragments thereof can have an improved safety and stability profile in vivo. The deglycosylated antibody may be of any isotype or subclass thereof, for example IgG1, IgG2, IgG3 or IgG 4. In certain embodiments, the deglycosylated antibody or fragment thereof comprises a deglycosylated IgG4 antibody Fc region, which lacks Fc effector function, thereby eliminating the potential for Fc-mediated toxicity to normal vital organs that express IL-6. In still other embodiments, the antibodies or fragments thereof of the invention comprise altered glycans. For example, the antibody can have a reduced number of fucose residues on the N-glycan at Asn297 of the Fc region, i.e., be defucosylated. Defucose increases FcRgII binding to NK cells and strongly increases ADCC. Bivalent antibodies comprising anti-IL-6 scFv and anti-CD 3scFv have been shown to induce ADCC killing of IL-6 expressing cells. Accordingly, in one embodiment, defucosylated anti-IL-6 antibodies are used to target and kill IL-6 expressing cells. In another embodiment, the antibody may have an altered number of sialic acid residues on the N-glycan at Asn297 of the Fc region. "deglycosylated" antibodies or antibodies with altered glycans can be prepared using a number of art-recognized methods. For example, such antibodies can be produced using genetically engineered host cells (e.g., modified yeast, such as pichia (Picchia), or CHO cells) with modified glycosylation pathways (e.g., glycosyltransferase deletions).

iii) covalent attachment

The binding molecules of the invention may be modified, for example, by covalently attaching the molecule to the binding molecule such that covalent attachment does not prevent the binding polypeptide from specifically binding its cognate epitope. For example, but not limited to, the antibodies or fragments thereof of the invention may be modified by glycosylation, acetylation, pegylation, phosphorylation, amidation, derivatization with known protecting/blocking groups, proteolytic cleavage, attachment to cellular ligands or other proteins, and the like. Any of a variety of chemical modifications can be made by known techniques, including but not limited to specific chemical cleavage, acetylation, formylation, and the like. In addition, the derivative may contain one or more non-canonical amino acids.

The binding polypeptides (e.g., antibodies or fragments thereof) of the invention can be further recombinantly fused at the N-or C-terminus to a heterologous polypeptide, or chemically conjugated (including covalent and non-covalent conjugation) to a polypeptide or other composition. For example, anti-IL-6 antibodies can be recombinantly fused or conjugated to molecules useful as labels in detection assays, and effector molecules such as heterologous polypeptides, drugs, radionuclides, or toxins. See, e.g., PCT publication WO 92/08495; WO 91/14438; WO 89/12624; U.S. patent nos. 5,314,995; and EP396,387.

The binding molecules may be fused to heterologous polypeptides using methods known in the art to increase half-life in vivo or for use in immunoassays. For example, in one embodiment, PEG may be conjugated to a binding molecule of the invention to increase its in vivo half-life. Leong, S.R., et al, Cytokine16:106 (2001); in Drug Deliv.Rev.54:531 (2002); or Weir et al, biochem. Soc. transactions30:512 (2002).

Furthermore, the binding molecules of the invention may be fused to a marker sequence, such as a peptide, to aid in their purification or detection. In a preferred embodiment, the marker amino acid sequence is a hexa-histidine peptide, such as the markers provided in the pQE vector (QIAGEN, inc., 9259Eton Avenue, Chatsworth, calif., 91311), among others, many of which are commercially available. For example, hexahistidine allows for convenient purification of the fusion protein as described in Gentz et al, Proc.Natl.Acad.Sci.USA86:821-824 (1989). Other peptide tags that may be used for purification include, but are not limited to, the "HA" tag (which corresponds to an epitope derived from influenza hemagglutinin protein) (Wilson et al, Cell37:767(1984)) and the "Flag" tag.

The binding molecules of the invention may be used in unconjugated form, or may be conjugated to at least one of a variety of molecules, for example to improve the therapeutic properties of the molecule, to facilitate target detection, or to image or treat a patient. When purified, the binding molecules of the invention may be labeled or conjugated before or after purification. In particular, the binding molecules of the invention can be conjugated to therapeutic agents, prodrugs, peptides, proteins, enzymes, viruses, lipids, biological response modifiers (biological response modifiers), agents, or PEG.

The invention further encompasses binding molecules of the invention conjugated to diagnostic or therapeutic agents. The binding molecules can be used diagnostically, e.g., to monitor the development or progression of an immune cell disorder (e.g., CLL), as part of a clinical testing procedure, e.g., to determine the efficacy of a given therapeutic and/or prophylactic regimen. Detection may be facilitated by coupling the binding molecule to a detectable substance. Examples of detectable substances include various enzymes, prosthetic groups (prosthetic groups), fluorescent substances, luminescent substances, bioluminescent substances, radioactive substances, positron emitting metals for use in several positron emission tomography, and nonradioactive paramagnetic metal ions. For metal ions that can be conjugated to antibodies for use as diagnostics according to the present invention, see, e.g., U.S. Pat. No. 4,741,900. Examples of suitable enzymes include horseradish peroxidase, alkaline phosphatase, beta-galactosidase or acetylcholinesterase; examples of suitable prosthetic group complexes include streptavidin/biotin and avidin/biotin; examples of suitable fluorescent substances include umbelliferone, fluorescein isothiocyanate, rhodamine, dichlorotriazinylamine fluorescein, dansyl chloride or phycoerythrin; examples of luminescent substances include luminol (luminol); examples of bioluminescent materials include luciferase, luciferin and aequorin (phycerythrin); examples of suitable radioactive materials include 125I, 131I, 111In or 99 Tc.

The binding molecules for use in the diagnostic and therapeutic methods disclosed herein can be conjugated to a cytotoxin (e.g., a radioisotope, cytotoxic drug, or toxin), a therapeutic agent, a cytostatic agent, a biotoxin, a prodrug, a peptide, a protein, an enzyme, a virus, a lipid, a biological response modifier, an agent, an immunologically active ligand (e.g., a lymphokine or other antibody, wherein the resulting molecule binds to both neoplastic cells and effector cells, such as T cells), or PEG.

In another embodiment, anti-IL-6 antibodies for use in the diagnostic and therapeutic methods disclosed herein can be conjugated to a molecule that reduces tumor cell growth. In other embodiments, the disclosed compositions may comprise an antibody, or fragment thereof, conjugated to a drug or prodrug. Still other embodiments of the invention include the use of antibodies or fragments thereof conjugated to specific biotoxins or cytotoxic fragments thereof, such as ricin, gelonin, pseudomonas exotoxin or diphtheria toxin. The choice of which conjugated or unconjugated antibody to use will depend on the type and stage of cancer, the use of adjunctive therapy (e.g., chemotherapy or external radiation) and the patient's condition. It will be appreciated that such a selection may be readily made by those skilled in the art in light of the teachings herein.

It will be appreciated that in previous studies, isotope-labeled anti-tumor antibodies have been successfully used to destroy tumor cells in animal models and in some cases in humans. Exemplary radioisotopes include 90Y, 125I, 131I, 123I, 111In, 105Rh, 153Sm, 67Cu, 67Ga, 166Ho, 177Lu, 186Re and 188 Re. The radionuclide acts by generating ionizing radiation that causes multiple strand breaks in nuclear DNA, resulting in cell death. Isotopes used to produce therapeutic conjugates typically produce high energy alpha-or beta-particles, which have short path lengths. Such radionuclides kill cells in proximity thereto, e.g., neoplastic cells to which the conjugate attaches or enters. It has little or no effect on non-localized cells. Radionuclides are substantially non-immunogenic.

V. expression of binding molecules

After manipulation of the isolated genetic material to provide the binding molecules of the invention as described above, the gene is typically inserted into an expression vector for introduction into a host cell, which can be used to produce the desired amount of the claimed antibody or fragment thereof.

For the purposes of the present specification and claims, the term "vector" or "expression vector" is used herein to mean a vector according to the invention which is used as a vehicle for introducing and expressing a desired gene in a cell. As known to those skilled in the art, such vectors can be readily selected from the group consisting of: plasmids, phages, viruses and retroviruses. Generally, vectors compatible with the present invention will comprise a selection marker, suitable restriction sites to facilitate cloning of the desired gene, and the ability to enter and/or replicate in eukaryotic or prokaryotic cells.

For the purposes of the present invention, a variety of expression vector systems may be employed. For example, one type of vector utilizes DNA elements derived from animal viruses such as bovine papilloma virus, polyoma virus, adenovirus, vaccinia virus, baculovirus, retrovirus (RSV, MMTV or MOMLV) or SV40 virus. Others involve the use of polycistronic systems with internal ribosome binding sites. In addition, cells that integrate the DNA into their chromosomes can be selected by introducing one or more markers that allow selection of transfected host cells. The marker may confer prototrophy to an auxotrophic host, biocide (e.g., antibiotics) resistance, or resistance to heavy metals such as copper. The selectable marker gene may be directly linked to the DNA sequence to be expressed, or introduced into the same cell by co-transformation. Other elements may also be required for optimal synthesis of mRNA. These elements may include signal sequences, splicing signals, as well as transcriptional promoters, enhancers, and termination signals. In a particularly preferred embodiment, the cloned variable region genes are inserted into an expression vector along with synthetic heavy and light chain constant region genes (preferably human) as discussed above.

In other preferred embodiments, the binding molecules of the invention or fragments thereof may be expressed using a polycistronic construct. In such expression systems, multiple gene products of interest, such as the heavy and light chains of an antibody, can be produced from a single polycistronic construct. These systems advantageously use an Internal Ribosome Entry Site (IRES) to provide relatively high levels of the polypeptides of the invention in eukaryotic host cells. Compatible IRES sequences are disclosed in U.S. patent No. 6,193,980, which is incorporated herein by reference. One skilled in the art will recognize that such expression systems can be used to efficiently produce the full range of polypeptides disclosed in the present application.

More generally, once a vector or DNA sequence encoding the antibody or fragment thereof is prepared, the expression vector may be introduced into a suitable host cell. That is, the host cell may be transformed. Introduction of the plasmid into the host cell can be accomplished by a variety of techniques well known to those skilled in the art. This includes, but is not limited to, transfection (including electrophoresis and electroporation), protoplast fusion, calcium phosphate precipitation, cell fusion with enveloped DNA, microinjection, and infection with whole virus. See Ridgway, a.a.g. "Mammalian Expression Vectors" chapter24.2, pp.470-472Vectors, Rodriguez and Denhardt, Eds. (butterworth, Boston, mass.1988). More preferably, the introduction of the plasmid into the host is by electroporation. The transformed cells were grown under conditions suitable for the production of light and heavy chains and assayed for heavy and/or light chain protein synthesis. Exemplary assay techniques include enzyme-linked immunosorbent assay (ELISA), Radioimmunoassay (RIA), or fluorescence activated cell sorting analysis (FACS), immunohistochemistry, and the like.

As used herein, the term "transformation" shall be used in a broad sense to refer to the introduction of DNA into a recipient host cell that alters the genotype and subsequently causes a change in the recipient cell.

By analogy, a "host cell" is a cell transformed with a vector constructed using recombinant DNA techniques and encoding at least one foreign gene. In describing the process of isolating a polypeptide from a recombinant host, the terms "cell" and "cell culture" are used interchangeably to refer to the source of the antibody unless otherwise indicated. In other words, recovery of the polypeptide from the "cells" may mean from precipitated whole cells, or from cell cultures containing media and suspended cells.

In one embodiment, the host cell line used for antibody expression is of mammalian origin. One skilled in the art can determine the particular host cell line that is most suitable for the desired gene product to be expressed therein. Exemplary host cell lines include, but are not limited to, DG44 and DUXB11 (chinese hamster ovary line, DHFR-), HELA (human cervical carcinoma), CVI (monkey kidney line), COS (derivative of CVI with SV40T antigen), R1610 (chinese hamster fibroblast), BALBC/3T3 (mouse fibroblast), HAK (hamster kidney line), SP2/O (mouse myeloma), BFA-1c1BPT (bovine endothelial cells), RAJI (human lymphocytes), 293 (human kidney). In one embodiment, the cell line provides altered glycosylation of the antibody expressed thereby, such as defucosylation (e.g., per. c6.rtm. (Crucell) or FUT 8-knock-out CHO cell line (potelligent. rtm. cells) (Biowa, Princeton, n.j.)). In one embodiment, NS0 cells may be used. CHO cells are particularly preferred. Host cell lines are generally available from commercial services, American Tissue Culture Collection or published literature.

In vitro production allows for scale-up to provide large quantities of the desired polypeptide. Techniques for mammalian cell culture under tissue culture conditions are known in the art and include homogeneous suspension culture, e.g., in an airlift reactor or a continuous stirred reactor, or immobilized or captured (entrypped) cell culture, e.g., in hollow fibers, microcapsules, on agarose microbeads or ceramic cartridges (ceramic cartridges). If necessary and/or desired, the polypeptide solution may be purified by means of a custom-made chromatographic method, such as gel filtration, ion exchange chromatography, chromatography on DEAE-cellulose, and/or (immuno) affinity chromatography.

The gene encoding the binding molecule of the invention or a fragment thereof may also be expressed in a non-mammalian cell, such as a bacterial or fungal or plant cell. In this regard, it is understood that a variety of unicellular non-mammalian microorganisms such as bacteria may also be transformed; i.e. those capable of growing in culture or fermentation. Bacteria susceptible to transformation include members of the enterobacteriaceae family such as Escherichia coli (Escherichia coli) or Salmonella (Salmonella); bacillaceae (Bacillus) such as Bacillus subtilis; pneumococcus (Pneumococcus); streptococcus (Streptococcus); and strains of Haemophilus influenzae (Haemophilus influenzae). It is further understood that the polypeptide may become part of an inclusion body when expressed in bacteria. The polypeptides must be isolated, purified, and then assembled into functional molecules.

In addition to prokaryotes, eukaryotic microorganisms may also be used. Saccharomyces cerevisiae, a common baker's yeast, is the most commonly used among eukaryotic microorganisms, although a variety of other strains may generally be used. For expression in Saccharomyces (Saccharomyces), for example, the plasmid YRp7(Stinchcomb et al, Nature, 282:39 (1979); Kingsman et al, Gene, 7:141 (1979); Tschemper et al, Gene, 10:157(1980)) is commonly used. This plasmid already contains the TRP1 gene, which provides a selection marker for mutants of yeast lacking growth in tryptophan, such as ATCC No. 44076 or PEP4-1(Jones, Genetics, 85:12 (1977)). The presence of trpl deficiency, then, as a characteristic of the yeast host cell genome, provides an effective environment for detecting transformation by growth in the absence of tryptophan.

Pharmaceutical formulations and methods of administration of binding molecules

In another aspect, the invention provides a pharmaceutical composition comprising a binding molecule (e.g., an antibody or antigen-binding fragment thereof).

Methods for preparing the binding molecules of the invention and administering them to a subject are well known or can be readily determined by those skilled in the art. The administration route of the antibody or fragment thereof of the present invention may be oral, parenteral, inhalation or topical. The term parenteral as used herein includes intravenous, intraarterial, intraperitoneal, intramuscular, subcutaneous, rectal or vaginal administration. Parenteral administration in intravenous, intraarterial, subcutaneous and intramuscular forms is generally preferred. While administration of all of these forms is expressly contemplated as being within the scope of the present invention, exemplary forms for administration are solutions for injection, particularly for intravenous or intra-arterial injection or drip. Typically, a suitable pharmaceutical composition for injection may comprise a buffer (e.g., an acetic, phosphoric, or citric acid buffer), a surfactant (e.g., polysorbate), an optional stabilizer (e.g., human albumin), and the like. In other methods compatible with the teachings herein, the polypeptide may be delivered directly to the site of the detrimental (overture) cell population, thereby increasing exposure of the diseased tissue to the therapeutic agent. For example, the high thermal stability and solubility properties of the binding molecules of the invention make them ideal agents for topical administration, e.g., by subcutaneous (Sub-Q) injection. Furthermore, the extremely high affinity and potency of the antibodies of the invention allows for the use of lower effective doses, thereby simplifying subcutaneous injections. Accordingly, the binding molecules are particularly suitable for the treatment or prevention of inflammatory-related disorders (e.g., rheumatoid arthritis), cancer, or related conditions (e.g., anorexia or cachexia) for which localized delivery of the binding molecule may be desirable.

Preparations for parenteral administration include sterile aqueous or nonaqueous solutions, suspensions and emulsions. Examples of non-aqueous solvents are propylene glycol, polyethylene glycol, vegetable oils such as olive oil, and injectable organic esters such as ethyl oleate. Aqueous carriers include water, alcohol/water solutions, emulsions or suspensions, including saline and buffered media. In the subject invention, pharmaceutically acceptable carriers include, but are not limited to, 0.01-0.1M and preferably 0.05M phosphate buffer or 0.8% saline. Other common parenteral vehicles include sodium phosphate solution, Ringer dextrose, dextrose and sodium chloride, lactated Ringer, or fixed oils. Intravenous vehicles include liquid and nutritional supplements, electrolyte supplements such as those based on Ringer dextrose, and the like. Preservatives and other additives may also be present such as, for example, antimicrobials, antioxidants, chelating agents, and inert gases and the like. More specifically, pharmaceutical compositions suitable for injectable use include sterile aqueous solutions (when water-soluble) or dispersions and sterile powders for the extemporaneous preparation of sterile injectable solutions or dispersions. In such cases, the composition must be sterile and should be a liquid to the extent that easy syringability is achieved. It should be stable under the conditions of manufacture and storage and preferably preserved against the contaminating action of microorganisms such as bacteria and fungi. The carrier can be a solvent or dispersion medium containing, for example, water, ethanol, polyol (for example, glycerol, propylene glycol, and liquid polyethylene glycol, and the like), and suitable mixtures thereof. Suitable fluidity can be maintained, for example, by the use of a coating such as lecithin, by the maintenance of the required particle size in the case of dispersions and by the use of surfactants. Prevention of the action of microorganisms can be achieved by various antibacterial and antifungal agents, for example, parabens (parabens), chlorobutanol, phenol, ascorbic acid, thimerosal, and the like. In many cases, it is preferred to include isotonic agents, for example, sugars, polyalcohols such as mannitol, sorbitol, or sodium chloride in the composition. Prolonged absorption of the injectable compositions can be brought about by including in the composition an agent which delays absorption, for example, aluminum monostearate and gelatin.

In any event, sterile injectable solutions can be prepared by incorporating the active compound (e.g., the antibody itself or in combination with other active agents) in the required amount in a suitable solvent with one or a combination of ingredients enumerated herein, as required, followed by filtered sterilization. Generally, dispersions are prepared by incorporating the active compound into a sterile vehicle which contains a basic dispersion medium and the required other ingredients from those listed above. In the case of sterile powders for the preparation of sterile injectable solutions, the preferred methods of preparation are vacuum drying and freeze-drying which yields a powder of the active ingredient plus any additional desired ingredient from a previously sterile-filtered solution thereof. The preparations for injection are processed under sterile conditions according to methods known in the art, filled into containers such as ampoules, bags, bottles, syringes or vials, and sealed. Further, the preparations may be packaged and sold in the form of kits, such as those described in co-pending U.S. serial No.09/259,337 and U.S. serial No.09/259,338, each of which is incorporated herein by reference. Such articles preferably carry a label or package insert indicating that the relevant composition is useful for treating a subject suffering from or susceptible to an autoimmune or neoplastic disease.

The binding molecules of the invention can be formulated in a wide range of concentrations for pharmaceutical use. For example, the binding molecule can be formulated at a concentration of 5mg/ml to 50mg/ml (e.g., 5, 10, 20, 50 mg/ml). Alternatively, the binding molecules of the invention may be modified to higher concentration formulations for topical (e.g., subcutaneous) administration. For example, the binding molecule may be formulated at a concentration of 50mg/ml to 200mg/ml, such as about 50, about 75, about 100, about 150, about 175, or about 200 mg/ml.

For treatment under the conditions described above, the effective amount of the binding molecules of the invention will vary depending on a number of different factors, including the means of administration, the target site, the physiological state of the patient, whether the patient is a human or an animal, other drugs administered, and whether the treatment is prophylactic or therapeutic. Typically, the patient is a human, but non-human mammals including transgenic mammals can also be treated. Treatment doses can be titrated using conventional methods known to those skilled in the art to optimize safety and efficacy.

For passive immunization with an antibody of the invention, the dosage can range, for example, from about 0.0001 to 100mg/kg, and more typically 0.01 to 5mg/kg (e.g., 0.02mg/kg, 0.25mg/kg, 0.5mg/kg, 0.75mg/kg, 1mg/kg, 2mg/kg, etc.) of the host weight. For example, the dose may be 1mg/kg body weight or 10mg/kg body weight or in the range of 1 to 10mg/kg, preferably at least 1 mg/kg. Intermediate doses within the above ranges are also intended to be within the scope of the present invention.

Such doses may be administered to the subject daily, every other day, weekly, or according to any other schedule determined by empirical analysis. An exemplary treatment requires administration in multiple doses over a sustained period of, for example, at least six months. Other exemplary treatment regimens require dosing once every two weeks or once a month or once every 3 to 6 months. Exemplary dosage schedules include 1-10mg/kg or 15mg/kg daily, 30mg/kg every other day, or 60mg/kg weekly. In some methods, two or more monoclonal antibodies with different binding specificities are administered simultaneously, in which case the dose of each antibody administered may fall within the indicated ranges.

The binding molecules of the invention may be administered multiple times. The interval between single administrations may be, for example, daily, weekly, monthly or yearly. The interval may also be non-periodic, as indicated by measuring the blood level of the polypeptide or target molecule in the patient. In some methods, the dose is adjusted to achieve a certain plasma antibody or toxin concentration, e.g., 1-1000 μ g/ml or 25-300 μ g/ml. Alternatively, the antibody or fragment thereof may be administered as a sustained release formulation, in which case less frequent administration is required. The dose and frequency will vary depending on the half-life of the antibody in the patient. Typically, the germlined or humanized antibodies show the longest half-life, followed by chimeric and non-human antibodies. In one embodiment, the antibody or fragment thereof of the invention may be administered in unconjugated form. In another embodiment, the antibody of the invention may be administered multiple times in a conjugated form. In yet another embodiment, the antibody or fragment thereof of the invention may be administered in unconjugated form and then in conjugated form, or vice versa.

The dosage and frequency of administration may vary depending on whether the treatment is prophylactic or therapeutic. In prophylactic use, a composition comprising an antibody of the invention or a mixture thereof is administered to a patient who is not already in a disease state to enhance the patient's resistance. Such amount is defined as a "prophylactically effective amount". In this use, the exact amount will also depend on the health and general immune status of the patient, but will generally range from 0.1 to 25mg per dose, especially from 0.5 to 2.5mg per dose. Relatively low doses are administered at relatively infrequent intervals over a long period of time. Some patients continue to receive treatment for the remainder of their lives.

In therapeutic use, it is sometimes desirable to administer relatively high doses (e.g., about 1 to 400mg/kg antibody per dose, 5 to 25mg doses more commonly used for radioimmunoconjugates, and higher doses for cytotoxic-drug conjugate molecules) at relatively short intervals until the progression of the disease is reduced or terminated, and preferably until the patient exhibits partial or complete remission of the disease symptoms. After this, the patient may be subjected to a prophylactic regimen.

In one embodiment, a subject may be treated with a nucleic acid molecule encoding a polypeptide of the invention (e.g., in a vector). The dosage of nucleic acid encoding the polypeptide may range from about 10ng to 1g, 100ng to 100mg, 1 μ g to 10mg, or 30-300 μ g of DNA per patient. The dose for infectious viral vectors varies from 10 to 100 or more viral particles per dose.

The therapeutic agent may be administered by parenteral, topical, intravenous, oral, subcutaneous, intraarterial, intracranial, intraperitoneal, intranasal or intramuscular means for prophylactic and/or therapeutic treatment. For administration of the antibody of the invention, intramuscular injection or intravenous infusion is preferred. In some methods, the therapeutic antibody or fragment thereof is injected directly into the cranium. In some methods, the antibody or fragment thereof is administered as a sustained release composition or device, such as Medipad^TMThe device is applied.

The agents of the invention may optionally be administered in combination with other agents effective for treating a disorder or condition in need of treatment (e.g., prophylactic or therapeutic). Preferred additional agents are those recognized in the art and administered under standard conditions for the particular condition.

Effective single treatment dose (i.e., therapeutically effective amount)⁹⁰Y-labeled antibodies of the invention range from about 5 to about 75mCi, more preferably from about 10 to about 40 mCi. Effective single treatment non-myeloablative (ablative) dose¹³¹The I-labeled antibody ranges from about 5 to about 70mCi, more preferably from about 5 to about 40 mCi. Effective single treatment ablation dose (i.e. may require autologous bone marrow transplant)¹³¹The I-labeled antibody ranges from about 30 to about 600mCi, more preferably from about 50 to less than about 500 mCi.

Although used already¹³¹I and⁹⁰y has gained a great deal of clinical experience, but other radiolabels are also known in the art and have been used for similar purposes. Other radioisotopes are also used for imaging. For example, other radioisotopes compatible with the scope of the present invention include, but are not limited to¹²³I，¹²⁵I，³²P， ⁵⁷Co，⁶⁴Cu，⁶⁷Cu，⁷⁷Br，⁸¹Rb，⁸¹Kr，⁸⁷Sr，¹¹³In，¹²⁷Cs，¹²⁹Cs，¹³²I，¹⁹⁷Hg， ²⁰³Pb，²⁰⁶Bi，¹⁷⁷Lu，¹⁸⁶Re，²¹²Pb，²¹²Bi，⁴⁷Sc，¹⁰⁵Rh，¹⁰⁹Pd，¹⁵³Sm，¹⁸⁸Re， ¹⁹⁹Au，²²⁵Ac，²¹¹At，²¹³And (4) Bi. In this regard, the alpha, gamma and beta emitters are all compatible with the present invention. Furthermore, in view of the present disclosure, it is believed that one skilled in the art can readily determine which radionuclides are compatible with a selected therapeutic procedure without undue experimentation. For this purpose, other radionuclides that have been used in clinical diagnosis include¹²⁵I，¹²³I，⁹⁹Tc，⁴³K，⁵²Fe，⁶⁷Ga，⁶⁸Ga and¹¹¹in. Antibodies are also labeled with various radionuclides for potential use in targeted immunotherapy (Peiresz et al immunol. cell biol.65:111-125 (1987)). These radionuclides include¹⁸⁸Re and¹⁸⁶re and, to a lesser extent,¹⁹⁹au and⁶⁷and (3) Cu. U.S. patent No. 5,460,785 provides additional data regarding such radioisotopes and is incorporated herein by reference.

As discussed previously, the binding molecules of the invention can be administered in a pharmaceutically effective amount for the in vivo treatment of a mammalian condition. To this end, it is understood that the disclosed antibodies or fragments thereof are formulated to facilitate administration and promote stability of the active agent. Preferably, the pharmaceutical composition according to the present invention comprises a pharmaceutically acceptable, non-toxic, sterile carrier such as physiological saline, non-toxic buffer, preservative and the like. For the purposes of this application, a pharmaceutically effective amount of an antibody of the invention conjugated or unconjugated to a therapeutic agent shall mean an amount sufficient to achieve effective binding to the target and to achieve a benefit, such as alleviation of symptoms of a disease or disorder, or detection of a substance or cell. In the case of tumor cells, the polypeptide is preferably capable of interacting with a selected immunoreactive antigen on a neoplastic or immunoreactive cell and providing an increase in death of such cells. Of course, the pharmaceutical compositions of the present invention may be administered in single or multiple doses to provide a pharmaceutically effective amount of the polypeptide.

Consistent with the scope of the present disclosure, the binding molecules of the present invention may be administered to a human or other animal in an amount sufficient to produce a therapeutic or prophylactic effect in accordance with the treatment methods described previously. The polypeptides of the invention can be administered to such humans or other animals in conventional dosage forms prepared according to known techniques by combining the antibodies of the invention with conventional pharmaceutically acceptable carriers or diluents. It is recognized by those skilled in the art that the form and character of the pharmaceutically acceptable carrier or diluent is determined by the amount of active ingredient with which it is to be combined, the route of administration, and other well-known variables. It will further be appreciated by those skilled in the art that mixtures comprising one or more polypeptides according to the invention may prove particularly effective.

Methods of treating IL-6-associated diseases or disorders

The binding molecules of the invention are useful for antagonizing IL-6 activity. Accordingly, in another aspect, the invention provides a method of treating an IL-6-associated disease or disorder by administering to a subject in need thereof a pharmaceutical composition comprising one or more binding molecules of the invention.

IL-6-associated diseases or disorders suitable for treatment include, but are not limited to, inflammatory diseases and cancer.

One skilled in the art would be able to determine by routine experimentation what is an effective, non-toxic amount of an antibody (or other therapeutic agent) for the purpose of treating an IL-6 related disease or disorder. For example, the therapeutically active amount of a polypeptide can vary according to factors such as the disease stage (e.g., stage I versus stage IV), age, sex, medical complications (e.g., immunosuppressive conditions or diseases) and weight of the subject, and the ability of the antibody to elicit a desired response in the subject. Dosage regimens may be adjusted to provide the optimal therapeutic response. For example, several divided doses may be administered daily, or the dose may be reduced proportionally as indicated by the exigencies of the therapeutic situation. Generally, however, an effective dose is contemplated in the range of about 0.05 to 100 milligrams per kilogram of body weight per day, and more preferably about 0.5 to 10 milligrams per kilogram of body weight per day.

VIII example

Example 1: generation and selection of IL-6 specific antagonistic Fab

Llama (llama) was immunized with human IL-6 (generated in E.coli (catalog No. 130-093-934) purchased from MACS Miltenyi Biotec or generated in human embryonic kidney cells (catalog No. HZ-1044) purchased from Humanzyme). Immunization of llamas and harvesting of Peripheral Blood Lymphocytes (PBLs), as well as subsequent extraction of RNA and amplification of antibody fragments, were performed as described by De Haard and co-workers (De Haard H et al, JBC.274:18218-30, 1999). After the last immunization, blood was collected and total RNA was extracted from PBL prepared using a Ficoll-Paque gradient and the method described by Chomczynski P et al (anal. biochem.162: 156. Buchner 159, 1987). The extracted RNA was then used for random cDNA synthesis and PCR amplification of the heavy and light chain V regions (V.lambda.and V.kappa.) to construct a Fab-containing phagemid library as described by De Haard H et al (biol. chem.274, 1999).

Fab expressing phage were generated according to standard protocols and further immobilized human IL-6 was selected, biotinylated and captured by neutravidin, or coated directly onto maxisorp plates. Total or competitive elution of IL-6-binding phage with trypsin was performed according to standard phage display protocols.

IL-6 specific Fab's were then screened for cross-competition with IL-6 neutralizing antibody, B-E8 and IL-6 receptor using an ELISA based competition assay. The VH and VL amino acid sequences of exemplary antagonistic IL-6 specific fabs identified using this assay are listed in table 13 below.

The binding kinetics of IL-6 specific Fab, which is capable of cross-competing with the B-E8 antibody, was assessed using surface plasmon resonance (Biacore). Specifically, biotinylated (prokaryotic) IL-6 was captured on a streptavidin biacore sensor chip (CM5-SA) and different concentrations of purified Fab were injected over 3 minutes followed by a 5 minute wash with buffer. From the washing phase, the dissociation rate (kd) is determined, andfrom the injection phase, the on-rate (ka) was calculated using the concentration and the dissociation rate as parameters. The measured off-and on-rates and calculated affinities of the antagonistic fabs are shown in table 2. Fab24C9 and 24D10 have a molecular weight distribution at 10^-5s^-1A range of off rates, and affinities of 660 and 270pM, respectively. The binding kinetics of antagonistic purified fabs (table 3) and periplasmic fractions of Fab expressing bacteria (table 4) were also assessed by surface plasmon resonance (Biacore) using bacteria directly coated on a CM5 chip and eukaryotic produced human IL-6. The Fab tested had 6x10^-4To 2x10^-5s^-1The dissociation rate of (c).

Table 2: binding kinetics of selected purified antagonistic Fab clones

Table 3: binding kinetics of selected purified antagonistic Fab clones

Table 4: binding kinetics of Fab-containing periplasmic fraction

Example 2: VH/VL shuffling for improved affinity.

VL chain shuffling was used to improve the affinity of Fab17F10, 18C7, 18C9, 18C11, 20a4, 29B11, 16D2, and 28a 6. In this method, the heavy chains of these clones (e.g., fragments of VHCH 1) are reintroduced into the primary phagemid light chain library (see example 1). Affinity selection was performed to select chain-shuffled fabs with improved affinity for IL-6. The binding kinetics of chain-shuffled Fab was evaluated by surface plasmon resonance (Biacore) using human IL-6 produced by bacteria and eukaryotes, as well as cynomolgus monkey IL-6 (tables 5-7). The VH and VL amino acid sequences of exemplary IL-6 specific fabs selected by the VL chain shuffling method are listed in table 14 below.

VH chain shuffling was also used to improve the affinity of Fab24D 10. In this method, the light chain of 24D10 was reintroduced into the primary phagemid heavy chain library (see example 1) and selection was performed using the off-rate assay. In this type of selection, the phage is allowed to bind to the antigen on the substrate for 1.5 to 2 hours. In 2 nd round, in PBS-Tween washing 15 times after, in excess soluble IL-6 in the presence of another washing. The principle is that phage antibodies with poor off-rates and therefore faster dissociation are captured by excess soluble target and removed in the wash. This approach avoids re-binding of such phage to the coated target. The time for the second washing step increased with the number of runs performed and the temperature was also increased to 37 ℃ to select for more stable Fab variants. The binding kinetics of chain-shuffled Fab and baseline antibodies BE8 and GL18 were evaluated by surface plasmon resonance (Biacore) using human IL-6 produced by bacteria and eukaryotes (tables 8 and 9). The VH and VL amino acid sequences of exemplary IL-6-specific fabs selected by this VH chain shuffling process are listed in table 14 below.

Table 5: binding kinetics of purified 17F10, 18C11, 18C7 and 20A4 chain shuffled Fab

Table 6: binding kinetics of periplasmic fraction containing 29B11 chain shuffled Fab

Table 7: binding kinetics of periplasmic fraction containing 28A6 chain shuffled Fab

Table 8: binding kinetics of periplasmic fraction containing 24D10 chain shuffled Fab

Table 9: binding kinetics of purified 24D10 chain shuffled Fab

Example 3: crystal structure of Fab61H7 complexed with IL-6

i) Characterization of the IL-6/Fab61H7 Complex

Size exclusion chromatography was performed on an Alliance2695HPLC system (Waters) using a Silica Gel KW803 column (Shodex) eluting with 50mM Tris-HCl pH7.5, 150mM NaCl at a flow rate of 0.5 ml/min. The detection was performed using a triple-angle light scattering detector (Mini-DAWN)^TMTREOS, Wyatt technology, Santa Barbara, USA). Molecular weight determination was performed by the ASTRA V software (Wyatt technology).

ii) crystallization

Initial crystallization screening of the IL-6/Fab61H7 complex was performed using the commercial kits Structure Screen1 and 2, Proplex Screen and Stura Footprint Screen (Molecular Dimensions Ltd). Droplets protein (8.45mg/ml) to mother liquor set at a 1:1(v: v) ratio on a Greiner96 well plate using a Cartesian MicroSys SQ robot in a total volume of 200 nl. Diffraction-quality crystallization of the complex was obtained by vapor diffusion at 277K through static-drop (sitting-drop) after optimization in 27.14% PEG MME2K, 0.1M Na Hepes ph 7.14. The crystals belong to the C2 space group, with the following unit cell dimensions:，and are andand β =97 °. It contains an ILF/Fab complex in each asymmetric unit, havingVm value of (a), which corresponds to a solvent content of 48%.

iii) analysis of the Structure of the IL-6:61H7 Complex

The predicted canonical structures for the CDR loops of the 61H7mAb are 1 and 3 for H1 and H2, respectively. And for L1, L2, and L3 are 7,1, and 4, respectively (www.bioinf.org.uk/abs/chothia. html). Overlay of 61H7VH (derived from patient-derived antibodies against the HIV-1 Protein gp 41) with Protein Data Bank (PDB) accession No. 1dfb shows that H1 and H2 fold with the expected specifications (see fig. 3A). Coverage of 61H7VL with Protein Data Bank (PDB) accession number 1mfa indicates that all three light chain CDRs adopt the expected conformation (see FIG. 3B). Accordingly, structural analysis confirmed that the VH of 61H 7(VH 3 family member) belongs to the human 1-3 combination of canonical H1-H2 structures, while the VL of 61H7 (VL8 family member) belongs to the human 7 λ 1-4 combination of human canonical structures.

IL-6 has been previously crystallized and classified as a four-helix bundle connected by loops and additional minisils (Somers et al, 1997, EMBO Journal16, 989-. Superposition of apo IL-6(pdb1ALU) and IL-6 from the IL-6:61H7 complex showed good agreement between the two models: (). This confirms that the 61H7Fab recognizes and binds to the native conformation of IL-6.

The crystal structure of IL-6 complexed with its receptor and signaling receptor gp130 shows a hexameric complex (Boulanger et al, 2003, Science27, 2101-2104, which is incorporated herein by reference in its entirety). IL-6 and IL-6R through the site I formed non-signaling complexes. Site II is a complex epitope formed by a binary complex of IL-6 and IL-6R. The interaction of site III with gp130 forms a signaling complex.

Superposition of the IL-6: IL-6R structure (pdb1P9M) with the IL-6:61H7 structure shows good agreement between the two IL-6 structures: (). The two loops differ in conformation. The first loop covering residues Asn48-Asn61 is a longer loop that does not form a structure in the apo IL-6 and IL-6:61H7 complex. This ring is stabilized in the IL-6: IL-6R structure by the binding of IL-6R. The second ring that differs in conformation is the so-called BC ring.

The crystalline asymmetric unit contains a single 1:1 complex. The overall architecture of the IL-6:61H7 complex showed that both VH (60%) and VL (40%) were responsible for a larger interaction area ((R))) And make a contribution. From the crystal structure, the residues important for the interaction with IL-6 were identified. The hydrogen bonds and salt bridges formed between the 61H7Fab and the cytokines are listed in table 10. The interactions were limited to CDR1 and CDR3 of the light chain and CDR1, CDR2 and CDR3 of the heavy chain.

Table 10: 61H7 Hydrogen bonds and salt bridges in the IL-6 Complex

The superposition of the IL-6:61H7 complex and the IL-6: IL-6R complex shows that there is steric hindrance between the 61H7Fab and the IL-6R. The predominant cause of spatial collisions with IL-6R is VL. The epitopes of IL-6R and 61H7 will be very close to each other. To verify that there is overlap between the two epitopes, the spacer IL-6R isWithin residues and from 61H7FabThe inner residues were mapped and searched for overlap between the two epitopes. The overlap between the two epitopes is very small and is formed mainly by the VH paratope. The overlap is concentrated around the cavity occupied by both the HCDR3 loop of 61H7 and the IL-6R molecule. This binding site of IL-6R on IL-6 is referred to as site I (Boulanger et al, 2003, Science27, 2101-2104, which is incorporated herein by reference in its entirety). The cavity formation site I is occupied by the hydrophobic side chain of Phe229 of IL-6R. This amino acid is called a hotspot residue by Boulanger et al, since mutagenesis studies show its critical role in the interaction between receptor and cytokine. Mutation of this residue to valine or serine completely abolished IL-6R binding to IL-6 (Kalai et al, 1997, Blood, 1319-1333, all of which are incorporated herein by reference in their entirety). Trp98 in the center of the heavy chain CDR3 loop of Fab61H7 occupied the same hole, suggesting that it therefore struck a key epitope in IL-6 to block its interaction with IL-6R (shown in FIG. 5). Trp98 is likely to be a key residue for the ultrahigh affinity of Fab61H7 for IL-6.

Example 4: crystal structures of Fab68F2 and 129D3 complexed with IL-6

i) Generation, data collection, and structure determination of IL-6:68F2 crystals

8mg of 68F2mAb (4mg/ml) in Dulbecco phosphate buffered saline (d-PBS) pH7.2 was buffer exchanged on a Zeba TM Desalt Spin Column (Pierce Fab Preparation Kit Thermo Scientific) to digestion buffer containing 20mM cysteine-HCl. The samples were incubated with immobilized papain (Pierce Thermo Scientific) and digested at 37 ℃ for 6 hours. The Fc fragment was separated from the Fab fragment using CaptureSelect human Fc affinity matrix (BAC BV Unilever) equilibrated in d-PBS. The Fab fragment was recovered from the flow through and the Fc fragment was eluted using 0.1M glycine pH 2.0. The protein concentration was determined by UV spectrophotometry based on absorbance at 280 nm. 4.6mg (>50%) of purified Fab68F2 was recovered and concentrated to 1.53mg/ml on Amicon-Ultra (cut off 10 kDa).

2.5mg of rh IL-6(Immunotools) were incubated with 2.6mg of Fab68F2 in Dulbecco's phosphate buffered saline (d-PBS) pH7.2 for 1 hour at 4 degrees Celsius and then concentrated to 1ml on Amicon-Ultra (cut off 10 kDa). The IL-6:68F2 complex was then separated from excess free IL-6 by gel filtration chromatography on a Superdex75 column in d-PBS and finally concentrated to 8.1mg/ml on an Amicon-Ultra concentrator (cut-off of 10 kDa). The purification of the complexes was assessed on SDS-PAGE.

Size exclusion chromatography was performed on an Alliance2695HPLC system (Waters) using a Silica Gel KW803 column (Shodex) eluting with 50mM Tris-HCl pH7.5, 150mM NaCl at a flow rate of 0.5 ml/min. The detection was carried out using a triple angle light scattering detector (Mini-DAWN)^TMTREOS, Wyatt technology, Santa Barbara, USA). Molecular weight determination was performed by the ASTRA V software (Wyatt technology). Initial crystallization screening of the IL-6:68F2 complex was performed using the commercial kits Structure Screen1 and 2, Proplex Screen and Stura Footprint Screen (Molecular Dimensions Ltd). Ca is used for liquid dropsThe rtesian MicroSys SQ robot set a 1:1(v: v) ratio of protein (8.1mg/ml) to mother liquor on a Greiner96 well plate in a total volume of 200 nl. Diffraction quality of the complex was crystallized by placing in 25% PEG4K, 0.15M (NH)₄)₂SO₄Obtained by stationary-drop (sitting-drop) vapor diffusion at 277K after 9 months in 0.1M MES pH 5.5.

The crystals for data collection were transferred to a mother liquor containing 7.5% ethylene glycol and flash frozen in liquid nitrogen. Diffraction data were collected under standard low temperature conditions at an ESRF synchrotron (Grenoble, France) using an ADSC Quantum4 detector on a beam line (beamline) ID14-4, and processed using XDS and measured with XSCALE (scale). The crystal structure of IL-6 complexed with Fab68F2 was determined by molecular replacement of Fab129D3 and IL-6 structures (molecular replacement) using mollrep according to single-wavelength normal diffraction experiments (table 4). The refinement was performed with BUSTER. Crystals of IL-6:129D3 were similarly produced.

iii) analysis of the Structure of the IL-6:68F2 and IL-6:129D3 Complex

The crystal structure of the IL-6:68F2 complex hasThe resolution of (2). The model was refined to 26.7% R factor and 29.6% R_freeFactor, and has reasonable stereochemistry (meaning that more than 95% of the residues adopt the allowed conformation). The crystal structure of the IL-6:129D3 complex hasThe resolution of (2). The model was refined to 28.5% R factor and 31.3% R_freeFactor and has reasonable stereochemistry.

The crystallized 68F2 and 129D3Fab (F)) VH domain () And VL domain (r.m.s.d.0.4). These overlays showed no significant difference between the parent 68F2 and the germlined 129D3Fab structure.

The predicted canonical structures for the CDR loops of 68F2mAb and its germlined variant 129D3 were 3 and 1 for H1 and H2, respectively, and 6 λ,1 and 5 for L1, L2 and L3, respectively. Canonical folding of the heavy chain was predicted by the server at www.bioinf.org.uk/abs/chothia. The reference Fab (PG16) used for comparison of light chain canonical folds was found artificially by searching for the Antibody Structure Summary Page (www.bioinf.org.uk/abs/sacs/Antibody _ Structure _ Summary _ Page. html). The overlap of 68F2 and 129D3VL with the VL of PG16 (Protein Data Bank (PDB) accession number 3MUG) suggests that all three CDRs adopt the predicted conformation (see fig. 4A). The superposition of 68F2/129D3VH with the reference PDB accession number 1ACY shows that H1 and H2 adopt the predicted canonical fold (see FIG. 4B). Accordingly, structural analysis confirmed that the VL of 68F2/129D3 (VL2 family members) belongs to the human 6 λ -1-5 combination of structures of the specification L1-L3, while the VH of 68F2/129D3 (VH4 family members) belongs to the 3-1 combination of structures of the human specification H1-H2.

The crystalline asymmetric unit contains a single 1:1 complex. The overall architecture of the IL-6:68F2 complex showed that VH (50%) and VL (50%) were equally responsible for the larger interaction area ((50%))) A contribution is made. The interaction area is slightly larger than the interaction area of 61H7 and IL-6(). Similar to the 61H7 IL-6 complex, only the L2 loop is not directly involved in the interaction with IL-6.

The interface between the heavy and light chains corresponds to 68F2And for the 129D3 structure corresponds to. The interfacial area and the ratio of 61H7: () And 27B3() The areas measured for the antibodies were comparable. All interface areas were calculated using the EMBL Web services PISA.

From the crystal structure, the residues important for the interaction with IL-6 were identified. The hydrogen bonds and salt bridges formed between the 68F2Fab and the cytokines are listed in table 11. The interactions were limited to CDR1 and CDR3 of the light chain and CDR1, CDR2 and CDR3 of the heavy chain.

VL shuffling of the 20a4 clone resulted in the 68F2 clone having 10x better affinity than the parental 20a 4. The two antibodies differ primarily in their L1 and L2 CDRs. The light chain CDR2 loop did not contribute to IL-6 binding, therefore, the affinity improvement attributed to the L1 loop mutation. Two of the three residues that interact significantly with IL-6 are not conserved in the 20A4 clone: asn26, which forms a hydrogen bond with Ser76 of IL-6, Ser in 20A 4; and Thr31, which forms two hydrogen bonds with IL-6, is Gly in the 20A4 parent L1. The addition of three new hydrogen bonds when changing these two residues probably explains the partial affinity increase observed after VL shuffling. More importantly, the change in L1 may result in stabilizing Y30(Kabat numbering) and/or locating it correctly in the F229 cavity, allowing for additional high efficiency forces.

Table 11: 68F2 Hydrogen bonds and salt bridges in the IL-6 Complex

The overlap of the IL-6:68F2 complex and the IL-6: IL-6R complex showed steric hindrance between the 68F2Fab and the IL-6R, as observed for the 61H7: IL-6 complex. However, in contrast to the 61H7: IL-6 complex, the VH of mainly 68F2 caused a spatial collision with IL-6R in the 68F2: IL-6 complex. The 68F2Fab interacts with IL-6 exactly at the same site as IL-6R. Furthermore, 68F2 does not overlap with the gp130 binding site and therefore specifically competes only with IL-6R.

The overlap of the IL-6: IL-6R and IL-6:68F2 complexes indicates that the epitopes of IL-6R and 68F2 are very close to each other. Residues belonging to both epitopes were located on IL-6 and their overlap was determined. The overlap of the 68F2 epitope with IL-6R is almost complete. The binding site of IL-6R on IL-6 is referred to as site I (Boulanger et al 2003, Science27, 2101-2104). The cavity-forming site I is occupied by the hydrophobic side chain of Phe229 of IL-6R. This amino acid is called a hotspot residue by Boulanger et al, since mutagenesis studies show its critical role in the interaction between receptor and cytokine. Mutation of this residue to valine or serine completely abolished IL-6R binding to IL-6 (Kalai et al, 1997, Blood, 1319-1333). Examination of the hole depicted as position I in the IL-6:68F2 structure revealed that it was occupied by the CDR1 loop of the light chain of Fab68F 2. Specifically, Tyr32 (position 30 in Kabat numbering) in CDR1 of the light chain plays a crucial role in binding to this site (fig. 5C).

Another key residue in the interaction between IL-6 and IL-6R is Phe279 of IL-6R. The residue represents 20% of the total binding interface (C) ((C))) (in contrast to 28% for Phe 229: () Make it a second important interaction. Like Phe229, Phe279Bind to the cavities formed on the surface of IL-6. This cavity is also occupied by residue 68F2, more particularly by Val104 of the CDR3 loop of the heavy chain (position 99 in Kabat numbering) (fig. 6).

IL-6 has been previously crystallized and classified as a four-helix bundle connected by loops and additional minisils (Somers et al 1997 EMBO Journal16, 989-. The crystal structure of IL-6 in complex with its receptor and the signaling receptor gp130 has been resolved (Boulanger et al, 2003, Science27, 2101-2104). IL-6 and IL-6R through the site I formed non-signaling complexes. Site II is a complex epitope formed by a binary complex of IL-6 and IL-6R. The interaction of site III with gp130 forms a signaling complex.

Comparison of the structure of the IL-6:61H7 complex with the structure of IL-6:68F2 shows that although both mAbs compete with IL-6R for binding to site I of IL-6, both antibodies bind IL-6 at two different epitopes. 68F2 binds to the side of the barrel-shaped IL-6 and competes exclusively for IL-6R binding, while 61H7 interacts more at the base of the barrel-shaped IL-6, competing with IL-6R and gp 130. Interestingly, the unique overlapping epitope on IL-6 is the cavity filled by the hot spot residue Phe229 of IL-6R (FIG. 5). This indicates that the binding of this hole is critical for the high efficiency forces observed for 68F2 and 61H 7.

Example 5: structural functional analysis of W98 in HCDR3 of 61H7

A significant feature of the extremely highly potent antibodies disclosed herein is their ability to occupy the F229 binding cavity (referred to herein as the F229 cavity) of the IL-6R on IL-6. For 61H7 (and its germlined variants such as 111a7), the F229 cavity was occupied by the tryptophan 98 residue (W98) of HCDR 3. To evaluate the functional importance of W98 on IL-6 binding by 61H7, a 111a7 mutant was generated in which VH position 98 was mutated to all possible amino acids in the context of M100A or M100L. The binding kinetics of periplasmic fractions of bacteria containing mutant Fab were tested using surface plasmon resonance (Biocore) and the off-rate for each mutant was determined. The results of the mutation analysis are shown in Table 12. The data clearly show that tryptophan (W) at position 98 is the best possible amino acid to achieve the best off-rate. The mutation of W98 was always detrimental to the binding of 111A7 to IL-6.

Table 12: off-rates of 61H7Fab with VH mutations at positions 98 and 100

Example 6: germlining of VH and VL of Fab clones 61H7 and 68F2

The VH and VL sequences of clones 68F2 and 61H7 were aligned against human germline VH and VL sequences to identify the most closely related germline sequences. The germlining process was carried out as described in WO2010/001251 and by Baca et al (J.biol.chem. (1997)272: 10678-. Library/phage display methods were used in which deviating FR residues for both human and llama residues were incorporated.

The camel-derived IL-6 antibodies of the invention significantly resemble human-like in sequence and structure. As a result, only a minimal number of sequence changes (through germlining) incorporate the final germlined variants. For example, of the 87(VH) and 79(VL) amino acids in the VH and VL framework regions of the parental 68F2 antibody, only 6 (VH) and 7 (VL) amino acid changes were introduced (13 amino acid changes in total), resulting in a final germlined precursor (lead) (129D3) with 93.1% and 91.1% identity in its respective VH and VL framework (see alignment of fig. 10A). Similarly, of the 87(VH) and 79(VL) amino acids in the VH and VL framework regions of the parent 61H7 antibody, only 8 (VH) and 5 (VL) amino acid changes were introduced (13 amino acid changes in total), resulting in a final germlined precursor (111a7) with 90.8 and 93.7% identity in its respective VH and VL framework (see alignment of fig. 10B).

In contrast, the art-recognized IL-6 antibodies require extensive engineering (CDR grafting) and sequence alterations (back-mutations) to generate variants suitable for therapeutic use. For example, the reference mouse IL-6 antibody CNTO-328 requires 14 (VH) and 22 (VL) amino acid changes (36 in total) to generate a humanized variant CNTO136 that has only 84% and 72.5% homology in the VH and VL frameworks (for alignment, see fig. 11A). Another reference rabbit IL-6 antibody, ALD518, requires a total of 46 framework changes (26 in VH and 20 in VL) to generate the final humanized variants with only 70.5% and 74% sequence homology to the parent antibody. Thus, it is clear that IL-6 antibodies require only minimal engineering and result in molecules that are more human in sequence.

In addition, the small number of FR residues to be altered makes it possible to incorporate changes in CDR residues into the germlining process. Such CDR mutations can be used to remove amino acids that introduce production variability (glycosylation sites, oxidation, isomerization, etc.) or to germline CDR residues to amino acid changes found in different variants of antibodies.

Phage display (using stringent selection conditions) was used to select other functional fabs. Individual clones were screened for off-rate and the best hits were sequenced to determine human sequence identity. The VH and VL amino acid sequences of exemplary germlined IL-6 specific fabs are listed in tables 15 and 16 below.

CDR regions from all identified VH and VL domains (which were variant fabs of 129D3 and 111a7) were compared and CDR amino acid consensus sequences were determined. The CDR variants and derived CDR common sequences for 129D3 and 111a7 are listed in tables 17 and 18 below.

Example 7: in vitro potency assay

The in vitro potency of clones 129D3, 68F2, 61H7, 133a9, 133H2, 133E5 and 132E7 was determined using a cell-based neutralization bioassay using the B9 cell line, essentially as described by Helle et al 1988, eur.j. immunol 18; 1535 as described in 1540. B9 cells were derived from the murine B cell hybridoma cell line B13.29, which requires IL-6 for survival and proliferation, and responds to very low concentrations of human IL-6. The assay was performed using a concentration series of 10pg/ml (or 0.5pM) human IL-6 and purified 129D3, B-E8, CAT6001, CNTO136, 61H7, UCB124.g 1. B9 cells were seeded in IL-6 free medium at 5000 cells/200. mu.l in flat-bottom plates in the presence, presence or absence of IL-6, antibodies. Proliferation was measured by [3H ] thymine pulsing at 64-72H. The results (shown in figure 1 and table 23) demonstrate that all clones had high potency, with 129D3 having an IC50 of less than 0.1pM in this assay. Interestingly, as shown in FIG. 1, the IC50 of clone 129D3 was superior to the benchmark CNTO136, CAT6001, UCB124, and B-E8 antibodies.

The in vitro potency of clones 133E5, 133a9, 133H2, 111B1, 104C1, 129D3, 68F2, 61H7 was also determined using a cell-based neutralization bioassay using the 7TD1 cell line, essentially as Van Snick et al PNAS; 9679(1986), which is incorporated herein by reference in its entirety. 7TD1 cells were murine hybridoma cell lines formed from the fusion of the mouse myeloma cell line Sp2/0-Ag14 with spleen cells from C57BL/6 mice immunized with E.coli lipopolysaccharide prior to fusion. The growth of the 7TD1 cell line was dependent on IL-6, and removal of IL-6 resulted in apoptosis-induced cell death. The assays were performed using a concentrated series of 75pg/ml human IL-6 and purified 133E5, 133A9, 133H2, 111B1, 104C1, 129D3, 68F2, 61H7, B-E8, GL18LB, CNTO136 and hu 1U.

Briefly, 7TD1 cells (7.10E3) were incubated in RPMI1640 medium +10% FCS for 2-4h, followed by the addition of IL-6(75pg/ml final concentration) (200ul final volume) in microtiter plates. Cells were incubated at 37 ℃ for 3 days, then washed with PBS and 60ul of substrate solution (p-nitrophenyl-N-acetyl- β -D-glucosamine (Sigma N-9376); 7.5mM substrate in 0.05M sodium citrate pH 5; 0.25 vol.% triton X-100) was added for 4 h. The enzymatic reaction was stopped with 90ul of stop solution (100mM glycine +10mM EDTA pH10.4) and the OD measured at 405 nm. The results (shown in table 24) demonstrate that all clones have high potency in this assay with an IC50 of less than 1 pM.

Example 8: epithelial ovarian cancer mouse xenograft assay

The in vivo efficacy of clone 129D3 was determined using a mouse xenograft model. IGROV-1 epithelial cells (5X 10)⁶) Nude mice were injected subcutaneously. After 3 days, the mice were administered 129D3 or CNT0328 biweekly at a dose of 4 or 20 mg/kg. There were 5 mice per group and 10 mice in the control group. The percent survival of mice in each study group was determined weekly and the results plotted as a survival curve. The results (shown in FIG. 2) demonstrate that 129D3 exhibits in vivo efficacy superior to that of the baseline CNT0328IL-6 antibody.

Example 9: immunogenicity assays

The VH and VL regions of the IL-6 antibodies of the invention are evaluated for the presence of potential immunogenic sequences (e.g., putative HLA type II restricted epitopes, also known as TH-epitopes) and used with a variety of commercially available reference antibodiesThe analysis (profiling) method compares the immunogenicity scores.

Assays were performed at the allotype level for 18DRB1, 6DRB3/4/5, 13DQ and 5DP, i.e. a total of 42 HLA type II receptors. Strong and moderate binders of DRB1, DRB3/4/5, and strong binders of DQ and DP epitopes were identified. Epitope counting was performed for strong and medium affinity DRB1 binders, respectively. Peptides binding to the same group of multiple allotypes were counted once. An approximate score representing the worst case immunogenic risk is calculated as follows: score = Σ (epitope count x allotype frequency).

In other words, the number of epitopes affecting a particular HLA allotype is multiplied by the allele frequency of the affected allotype. For a given sequence, all DRB1 allotypes present in 2% or more of the white population for the product used in the study were summed.

The DRB1 scores for the IL-6 antibodies of the invention and representative reference antibodies are provided in figure 9. The total DRB1 score is a composite of the VH and VL scores for each antibody, whereas a low score indicates low immunogenicity of the antibody. Accordingly, FIG. 9 illustrates that the IL-6 antibodies of the invention are equally or less immunogenic than the benchmark IL-6 antibodies as well as other commercially available antibodies (e.g., Humira and Remicade).

Example 10: manufacturability of

The germlined version of VH and VL of 68F2 was re-cloned in the pUPE heavy and light chain expression vectors, respectively, for transient expression of full-length IgG1 antibody. Following transient expression in HEK293E cells, IgG1 antibody was purified with protein a and quantified by measuring OD 280. Table 19 below summarizes the levels of production of the germlined derivatives, as well as the levels of the 68F2 parent antibody. Potency (in pM) was also measured for each antibody in a proliferation assay based on 7TD 1.

Variants 126a3, 127F1, 129D3 and 129F1 (all selected from germlined libraries under stringent conditions) were found to have similar potency (i.e. 0.5 to 0.7pM) as wild-type 68F 2. Furthermore, all the germlined variants were well expressed, i.e., 24 to 28. mu.g/ml. The exception was the germlined variant 129F2 which gave a yield of 9. mu.g/ml.

Example 11: stability analysis

To examine the germlining and thermostability of the parental versions of 68F2 in the form of full-length human IgG1, antibody samples were incubated at a concentration of 100 μ g/ml (in PBS) at 4, 50, 55, 60, 65, 70 and 75 ℃ for 1 hour. After this, the sample was slowly cooled to 25 ℃ over a period of 15 minutes and maintained at this temperature for 2 hours, after which it was stored overnight at 4 ℃. After centrifugation to remove the pellet (of denatured antibody), the concentration of functional antibody remaining in solution was measured using Biacore (1/10 dilution in PBS and 1/10 in HBSEP). The slope of the dissociation curve obtained after injection on an IL-6 immobilized chip is a measure of the concentration of functional antibody.

As shown in fig. 7A, the melting curve and melting temperature of wild-type 68F2 and its germlined derivatives were apparently unaffected by germlining. Quite unexpectedly, the melting temperature even showed improvement. For example, 129D3 has a Tm around 70 ℃ that is 3 ℃ higher than the parent 68F2 antibody. The melting curves of the germline variant 129D3 and parent 68F2 were also compared together with the reference antibodies GL18 and CNTO136 (see fig. 7B). The advantageous thermostability (Tm of 70 ℃) of SIMPLE antibody 68F2 and in particular its germlined variant 129D3 is rather shockable when compared to the Tm of CNTO136 at 65 ℃ and the Tm of GL18 antibody at 61 ℃. Surprisingly, extensive antibody engineering (e.g., humanization) and in vitro affinity maturation applied to both reference antibodies strongly affected their stability, while minimal engineering of SIMPLE antibodies and germlining generated in vivo resulted in extremely good thermostability.

68F2 and its germline variant (129D3) (and derived from SIMPLE)^TMThe serum stability of the full-length human IgG1 version of the germlined variant 103a1 of antibody precursor (lead)61H 7) was compared to those of the reference antibody. After incubation in human serum at 37 ℃, functional concentrations of antibody were measured at weeks 1, 2, 4,8, 12, 16, 24, 32 and 56 and compared to pre-aliquoted standards. As depicted in fig. 8, the serum stability of the antibodies of the invention is advantageous compared to the reference antibody.

Example 12: CMC optimization

For CMC quality antibody production, several residues or motifs are not recommended. Including the presence of methionine in the CDR loops of an antibody. Methionine can be oxidized, resulting in chemically altered variants of antibodies with altered properties such as affinity, potency, and stability. Accordingly, the methionine present in CDR3 of 111A7 (and its germlined variant of 61H7) was mutated to alanine (111A7MA), leucine (111A7ML) or serine (111A7 MS). The resulting CMC-optimized sequences are provided in table 20 below. As shown in Table 21, the methionine residues of the mutation on the binding to IL-6 has negligible effect.

Example 13: pharmacokinetic (PK) study of clone 129D3 and its Fc mutant in cynomolgus monkeys

Pharmacokinetic analysis of antibody clone 129D3 was performed as a format of multiple IgG1 molecules. The following antibodies were analyzed: wild type IgG1129D3(129D3-WT), IgG1129D3(129D3-YTE) with mutations M252Y/S254T/T256E in the Fc region, and IgG1129D3(129D3-HN) with mutations H433K and N434F in the Fc region.

Cynomolgus monkeys (3 for each antibody tested) were injected intravenously with a single 5mg/kg dose of 129D3-WT, 129D3-YTE, or 129D 3-HN. Samples were taken at different time points and tested for plasma concentration of mAb by ELISA. Specifically, microtiter plates (Maxisorb Nunc) were coated with 1ug/ml of IL-6 (immunools) in PBS overnight at 4 ℃. The plates were washed 2 times with PBS-Tween and blocked with 300. mu.l PBS-1% casein for 2 hours. After washing 2 times with PBS-Tween, the samples were applied. All dilutions were performed with 1% pooled blank plasma (this is a pool from three non-immunized cynomolgus monkeys, see chapter 4.2). Samples were allowed to bind for 2 hours at RT. The plates were then washed 5 times with PBS-Tween and goat biotinylated anti-human IgG heavy and light chain monkey-adsorbed polyclonal antibodies were applied at 1000-fold dilutions (Bethyl, cat # A80-319B) and allowed to bind at RT for 1 hour. After washing the plates 5 times with PBS-Tween, HRP-conjugated streptavidin (Jackson Immunoresearch 016-030-. The plates were then washed 5 times with PBS-Tween and attenuated by the addition of TMB (Calbeiochem CL07) -s (HS) TMB1:1 mixture of reagents (Weakener) (SDT, # sTMB-W). The staining was allowed to proceed for 10 minutes and then 0.5M H₂SO₄After termination, the absorbance was measured at 450 nm. Samples were analyzed four times and 129D3-WT (from the same batch as injected into animals) was used for the standard curve.

The relevant PK parameters for the non-compartmental analysis (non-comparative analysis) are shown in table 22 below. The pharmacokinetic analysis for the different 129D3IgG1 antibodies is shown in figure 12 (results shown are the average results for this group of monkeys). The data clearly show that 129D3-YTE and 129D3-HN have a longer Mean Retention Time (MRT) compared to the parent 129D3-WT antibody. Moreover, 129D3-YTE and 129D3-HN have slower elimination rates and thus substantially extended half-lives compared to 129D 3-WT.

Interestingly, although both antibodies contained wild-type IgG1Fc region, the half-life of 129D3-WT was significantly longer compared to that of the MedImmune anti-IL-6 IgG1 antibody (GL18) described in US 201200344212. In particular, 129D3-WT had a half-life of about 15.6 days, compared to antibody GL18, which had a half-life of about 8.5 days. Thus, the extended half-life of the antibodies of the invention was shown to be due to the nature of their corresponding Fab regions.

Example 14: serum Amyloid A (SAA) mouse model

The in vivo efficacy of clones 68F2 and 61H7 was further investigated by measuring the ability of these antibodies to block serum amyloid a (saa) induction in response to injected IL-6. The general method for performing this assay is set forth in WO2006/119115A2, which is incorporated herein by reference in its entirety. Specifically, Balb/c mice were injected intravenously with 68F2, 61H7, the reference antibody GL18, or CNTO136, or saline solution (control). Four hours after antibody administration, mice were injected with 0.1ug of IL-6. After a further 16 hours, mice were bled and the concentration of serum amyloid a was determined by ELISA. Experimental groups, dosages and results are listed in table 26 herein. Dose responses are also graphically depicted in figure 13. The results show that antibody clones 68F2 and 61H7 have in vivo potency at least equal to the high potency reference controls GL18 and CNTO 136.

Example 15: humanized mouse psoriasis xenograft model

A mouse xenograft model was used to evaluate the prophylactic efficacy of clone 68F2 on induced psoriasis lesion formation. In particular, BNX mice were transplanted with 5mm diameter full thickness skin biopsy (biopsy) from non-involved (non-infiltrated) skin from psoriasis patients (1 slice per mouse). After 3 weeks, use 0.5x10⁶Activated PBMCs were injected with the grafts. Treatment of mice with clone 68F2, an anti-TNF antibody (Remicade), or betamethasone dipropionate (positive control) was initiated one day before injection of activated cells into the graft. Details of treatment groups and protocols are shown in table 25.

Treatment efficacy was determined by epithelial fold thickness (measured by light microscopy). Significance between groups was analyzed using an analysis of variance (ANOVA) followed by a post-hoc Least Square Difference (LSD) test to establish statistically significant differences between treatment groups. Values with p <0.05 indicate significant differences between groups.

The results of these experiments are presented in figure 14 herein. The epithelial fold thickness of the control group (group 2) was 156 μm ± 4 (mean ± s.e.m.). Treatment with betamethasone (group 1, n =3) significantly reduced the epithelial fold thickness to 83 μm ± 13(p <0.05) (fig. 12 and 13). The average epithelial fold thickness for the Remicade treatment group (group 3) was 125 μm + -12, compared to 125 μm + -12 for the 68F2 treatment group (group 4). These data show that clone 68F2 was as effective as the anti-TNF antibody Remicade in the humanized mouse psoriasis model.

Example 16: renal cell carcinoma mouse xenograft model

Investigation of the in vivo efficacy of clone 68F2 in a renal cell carcinoma mouse xenograft modelForce. The general method of performing this assay is set forth in WO2008/144763, which is incorporated herein by reference in its entirety. Briefly, RXF393 cells (2 × 10)⁶) Subcutaneous injections were made on both sides of nude mice. Allowing tumor growth to 50 to 300mm³And then antibody administration. 90% of the injected mice developed tumors. 40 mice were divided into 5 groups of 8 mice each. Each group received an intraperitoneal injection of PBS (control) or a specific dose of clone 68F2(1, 3, 10, or 30 mg/kg). Tumor size and survival were monitored twice weekly.

The survival data set forth in figure 15 herein shows that 68F2 is effective in delaying the death of mice relative to controls. In particular, median survival time was observed to be 15.5 days for the PBS group, 21 days for the 1mg/kg, 3mg/kg and 30mg/kg68F2 groups, and 27.5 days for the 10mg/kg68F2 group. Survival data presented in figure 18 herein shows that 129D3 and 111a7 at a dose of 3mg/kg effectively delayed the death of mice relative to controls. The tumor growth rate data set forth in figure 16 herein shows that clone 68F2 is effective in inhibiting tumor growth in a dose dependent manner, with saturation at the 10mg/kg dose. The tumor growth rate data set forth in fig. 17 herein shows that clones 129D3 and 111A7SDMA > a were effective at inhibiting tumor growth at a dose of 3 mg/kg.

Other tables:

table 13. Exemplary anti-IL-6 neutralizing Fab VH and VL amino acid sequences.

Table 14. Exemplary anti-IL-6 neutralizing Fab VH and VL amino acid sequences generated by VH or VL shuffling

Table 15. VH and VL amino acid sequences of an exemplary germlined variant of Fab clone 61H 7.

Table 16. VH and VL amino acid sequences of an exemplary germlined variant of Fab clone 68F2

Table 17. Sequence variants of the Fab129D3CDR amino acid sequence and CDR consensus sequences thereof

Table 18. Sequence variants of the Fab111A7CDR amino acid sequence and the CDR consensus sequence thereof

Table 19. Generation of level and efficacy (pM) of the germlining 68F2 variant

Table 20. CMC optimized sequence variants of Fab111A7

Table 21. IL-6 binding kinetics of CMC optimized sequence variants of Fab111a7

Table 22. Compartment-free PK assay for anti-IL-6 mAbs following single intravenous administration to cynomolgus monkeys

MRT = mean residence time; k = average exclusion rate constant; t1/2= exclusion half-life; cl = purge exclusion

TABLE 23 in vitro IL-6 neutralization assay using B9 cells

Table 24. In vitro IL-6 neutralization assay using 7TD1 cells

Table 25. Group and treatment regimens for use in psoriasis xenograft models

Group of	Treatment of	Group size	Route of administration	Processing frequency
					1	Betamethasone dipropionate	3	Surface of	2 times a day for three weeks
2	PBS	4	i.p.	200 μ l, 2 times a week for 3 weeks
					3	Remicade(10mg/kg)	7	i.p.	200 μ l, 2 times a week for 3 weeks
4	68F2(10mg/kg)	5	i.p.	200 μ l, 2 times a week for 3 weeks

Table 26. In vivo IL-6 neutralization in a SAA mouse model

Claims (25)

1. A binding molecule that specifically binds to IL-6, said binding molecule comprising at least one antibody CDR, wherein said CDR comprises at least one amino acid residue that is buried in either the F229 or F279 cavity on IL-6 when said binding molecule binds to IL-6.

2. The binding molecule of claim 1, comprising a VL domain having an amino acid at position 30 according to Kabat, which amino acid is buried in the F229 hole on IL-6 when the binding molecule binds to IL-6.

3. The binding molecule of claim 2, wherein the amino acid at position 30 is tyrosine.

4. The binding molecule of any preceding claim, comprising a VH domain having an amino acid at position 99 according to Kabat, which amino acid is buried in an F279 hole on IL-6 when the binding molecule binds to IL-6.

5. The binding molecule of claim 4, wherein the amino acid at position 99 is valine.

6. The binding molecule of any preceding claim which is an antibody or antigen-binding fragment thereof.

7. The binding molecule of any preceding claim, comprising a VH domain comprising the HCDR3 amino acid sequence selected from the group consisting of seq id nos: 497-Asonic 500, 543, 544, 566, 567 and 568.

8. The binding molecule of claim 7, wherein the VH further comprises an HCDR2 amino acid sequence selected from the group consisting of SEQ ID NO:501, 507 and 545, 554.

9. The binding molecule of claim 7 or 8, wherein the VH further comprises an HCDR1 amino acid sequence selected from the group consisting of SEQ ID NO:508 and 555 and 562.

10. The binding molecule of any one of claims 7-9, further comprising a VL domain, wherein the VL domain comprises an LCDR3 amino acid sequence selected from the group consisting of seq id nos: 513-524 and 563.

11. The binding molecule of claim 10, wherein the VL domain further comprises an LCDR2 amino acid sequence selected from the group consisting of seq id nos: 525-535 and 564.

12. The binding molecule of claim 10 or 11, wherein the VL domain further comprises an LCDR1 amino acid sequence selected from the group consisting of seq id nos: 536, 542 and 565.

13. The binding molecule of claim 6, comprising a VH domain having an amino acid sequence selected from the group consisting of SEQ ID NO: SEQ ID NO 1-232 and 569-571.

14. The binding molecule of claim 6, comprising a VL domain having an amino acid sequence selected from the group consisting of SEQ ID NO:233 and 496 are shown in SEQ ID NO.

15. The binding molecule of claim 6, comprising: a VH domain having an amino acid sequence selected from the group consisting of: 1-232 and 569-571; and a VL domain having an amino acid sequence selected from the group consisting of: 233 and 496 are shown in SEQ ID NO.

16. A binding molecule according to any preceding claim, comprising a VH domain comprising hypervariable loops H1, H2 and H3, wherein the VH domain polypeptide is paired with a VL domain comprising hypervariable loops L1, L2 and L3, wherein at least one of the hypervariable loops H1-H3 and L1-L3 is obtained from a conventional antibody of an alpaca species obtained by active immunization of the alpaca (Lama) species with an IL-6 antigen.

17. The binding molecule of claim 16, wherein:

a) at least one of the hypervariable loops H1, H2, L1, L2 and L3 exhibits a predicted or actual canonical fold structure which is identical or substantially identical to the corresponding canonical fold structure of the H1, H2, L1, L2 or L3 hypervariable loops present in human antibodies;

b) the hyper-variable loops H1 and H2 each exhibit a predicted or actual canonical fold structure that is identical or substantially identical to the corresponding human canonical fold structure;

c) the hypervariable loops L1, L2 and L3 each exhibit a predicted or actual canonical fold structure which is identical or substantially identical to the corresponding human canonical fold structure;

d) hypervariable loops H1 and H2 formed a combination of predicted or actual canonical fold structures which is identical or substantially identical to the corresponding combination of canonical fold structures known to occur in human germline VH domains;

e) the hypervariable loops H1 and H2 form a combination of canonical fold structures which corresponds to a combination of human canonical fold structures selected from the group consisting of: 1-1, 1-2, 1-3, 1-4, 1-6, 2-1, 3-1 and 3-5;

f) hypervariable loops L1 and L2 formed a combination of predicted or actual canonical fold structures which is identical or substantially identical to the corresponding combination of canonical fold structures known to occur in the human germline VL domain;

g) the hyper-variable loops L1 and L2 form a combination of canonical fold structures corresponding to a combination of human canonical fold structures selected from the group consisting of: 11-7, 13-7(A, B, C), 14-7(A, B), 12-11, 14-11, 12-12, 2-1, 3-1, 4-1 and 6-1;

h) hypervariable loops H1 and H2 formed a combination of canonical fold structures, which corresponds to the 3-1 combination of human canonical fold structures as found in the human 1ACY antibody structure;

i) the hypervariable loops L1 and L2 formed a combination of canonical fold structures which corresponds to the 6 λ -1 combination of human canonical fold structures as found in the human 3MUG antibody structure; and/or

j) The hypervariable loops L1, L2 and L3 formed a combination of canonical fold structures, which corresponds to the 6 λ -1-5 combination of human canonical fold structures as found in the human 3MUG antibody structure.

18. The binding molecule of any one of the preceding claims, which is

a) Inhibiting the binding of IL-6 to the IL-6 receptor;

b) inhibiting the binding of gp130 to the IL-6 receptor;

c) specifically bind to human and cynomolgus monkey (cynomolgus monkey) IL-6;

d) comprising at least one CDR from a camelid (camelid) antibody that specifically binds to IL-6;

e) characterized by having a molecular weight of less than about 10.0Grading;

f) expressed at least 20mg/ml by transient expression in HEK293 cells;

g) exhibits a melting temperature (Tm) greater than 65 ℃;

h) inhibits IL-6-induced proliferation of B9 hybridoma cells with an IC50 of less than 0.1 pM;

i) at less than 2x10^-5s^-1Dissociation rate (k measured by surface plasmon resonance)_off) Binds to human IL-6;

j) is a germlined variant of a parent camelid antibody, the germlined variant having a higher melting temperature than the parent camelid antibody;

k) comprising at least one CDR from a conventional antibody of the alpaca genus without subsequent affinity maturation;

l) has a serum half-life of at least 9 days, preferably at least 15 days, when administered intravenously in native IgG1Fc format to a cynomolgus monkey; or

m) is a germlined variant of a parent binding molecule, wherein the binding molecule comprises VH and VL domains, and wherein one or both of the VH and VL domains of the binding molecule comprises a total of 1 to 10 amino acid substitutions in the framework regions as compared to the corresponding VH and VL domains of the parent non-human antibody; and/or comprising VH and VL domains, wherein one or both of the VH and VL domains of the binding molecule has 90% or greater sequence identity in the framework regions FR1, FR2, FR3 and FR4 to one or more corresponding human VH or VL domains.

19. A binding molecule according to any preceding claim, comprising a VH domain having an amino acid sequence selected from the group consisting of seq id nos: 1-232 and 569-571, wherein at least one glutamine is changed to glutamic acid.

20. The binding molecule of any one of the preceding claims, comprising a human Fc domain having the double mutation H433K/N434F.

21. A pharmaceutical composition comprising a binding molecule according to any one of the preceding claims and one or more pharmaceutically acceptable carriers.

22. A method of treating an IL-6 associated disease or disorder comprising administering to a subject in need of such treatment an effective amount of the pharmaceutical composition of claim 21.

23. An isolated nucleic acid encoding the binding molecule of any one of claims 1-20.

24. A recombinant expression vector comprising the nucleic acid of claim 23.

25. A host cell comprising the recombinant expression vector of claim 24.