Compositions monovalent for cd40l binding and methods of use   

This application is a continuation application of Ser. No. 12/387,421, filed May 1, 2009, which is a divisional application of U.S. Ser. No. 11/102,512, filed Apr. 8, 2005, now U.S. Pat. No. 7,563,443, which claims the benefit of U.S. 60/610,819, filed Sep. 17, 2004, the entirety of each is incorporated herein by reference.

BACKGROUND OF THE INVENTION

CD40 is a 50 kD cell surface glycoprotein molecule expressed on the surface of mature and immature B cells, macrophages, follicular dendritic cells, thymic epithelium, normal basal epithelium, and some tumor-derived cell lines. The CD40 molecule is a member of the TNF receptor family, and has important signaling functions leading to a variety of downstream effects in various cell types. Early studies showed that cross-linking of CD40 on the B cell surface with an antibody resulted in B cell proliferation and activation. Antibody cross linking of CD40 in the presence of IL-4 induces proliferation and class switching in vitro, B cell aggregation via LFA-1 (Gordon et al., 1988, J. Immunol. 140: 1425), and serine/threonine and tyrosine phosphorylation of a number of intracellular substrates (Gordon et al., 1988, supra; Uckun et al., 1991, J. Biol. Chem. 266:17478). Anti-CD40 monoclonal antibodies also prime B cells to proliferate in response to agents such as PMA (Gordon et al., 1987, Eur. J. Immunol. 17: 1535) and anti-CD20 antibody (Clark & Ledbetter, 1986, Proc. Natl. Acad. Sci. U.S.A. 83: 4494).

The receptor homology of CD40 and the antibody cross-linking studies showing a central role for CD40 in B cell activation prompted the search for a natural ligand. A mutant of the Jurkat T cell line was found to constitutively activate human B cells to secrete immunoglobulin (Yellin et al., 1991, J. Immunol. 147: 3389-3395). A monoclonal antibody, termed 5c8, was raised which specifically reacted with the mutant line, but not with the parental Jurkat cell line. The 5c8 antibody immunoprecipitated a 30 kD (more accurately, 29.3 kD, 261 amino acids) cell surface polypeptide and was found to specifically inhibit the B cell helper function of the mutant cell line. (Lederman et al., 1992, J. Exp. Med., 175: 1091-1101; Lederman et al., 1992, J. Immunol. 149: 3817-3826; Lederman et al., 1993, Curr. Opin. Immunol. 5: 439-444). The 30 kD polypeptide ligand of the 5c8 antibody was termed T-BAM, for T-B-cell Activating Molecule. A second line of studies used molecular cloning techniques to identify polypeptides that specifically bind the CD40 molecule. cDNA clones for a specific ligand of CD40 were identified in a CD40 binding assay and alternately termed CD40 Ligand (CD40L), gp39, CD154, or TRAP (Graf et al., 1992, Eur. J. Immunol. 22: 3191-3194; Armitage et al., 1992, Nature 357: 80-82; and Aruffo et al., 1993, Cell 72: 291-300). Subsequently, the CD40L clone was found to have the same structure as T-BAM (Covey et al., 1994, Mol. Immunol. 31: 471-484). Human CD40L protein shows 82.8% and 77.4% identity at the nucleic acid and amino acid levels, respectively, to a similar protein isolated from murine EL4 thymoma cells. Both of these proteins are ligands for CD40 cell surface antigen expressed on resting B cells. CD40L has also been described as IMD3, a protein involved in hyper-IgM immunodeficiency syndrome.

The human gene encoding CD40L maps to chromosome Xq26.3-q27. The gene contains five exons. Deletions, point mutations and frameshift mutations clustering within a limited region of the CD40L extracellular domain have been found to be the basis of a rare X-linked immunodeficiency syndrome (Hyper-IgM immunodeficiency syndrome, HIGM1) characterized by recurrent bacterial infections, very low or absent IgG, IgA and IgE, and normal to increased IgM and IgD serum levels. Causally-related mutations have been found to consist of clustered deletions arising by splice-donor mutations with exon skipping, splice-acceptor mutations with utilization of a cryptic splice site, and deletion/insertion events with the creation of a new splice site.

CD40L is expressed on activated, but not resting CD4+ T cells, and was found to play a particularly important role in the humoral immune response, being linked to B cell proliferation, antibody and cytokine production, and cell viability. In vivo, deletion or mutation of CD40L leads to severe immunodeficiency, both in mice and in humans, characterized by hypogammaglobulinemia and T cell deficits in cell-mediated immunity (Chess, C., 2001, in Therapeutic Immunology, 2nd edition, Austen, K. F., Burakoff, S., Rosen, F. and Strom, T., eds., Blackwell Sciences, pp. 441-456). Human CD4+ T cells infected by HIV1, which causes severe dysfunction of cellular immunity, but paradoxically results in intense polyclonal activation of B cells, do not express CD40L. Gene and cell surface expression of the CD40L by activated T cells has been shown to be depressed in a subgroup of patients with common variable immunodeficiency (CVI). Thus, inefficient signaling via CD40 may be responsible, at least in part, for the failure of B cell differentiation in these patients.

The functional consequences of CD40L binding to CD40 include, for example, a) rescuing B cells from apoptosis induced by Fas or cross-linking of IgM, b) induction of the co-stimulator molecules CD80 (B7-1) and CD86 (B7-2) which interact with CD28 and CD152 (CTLA-4) on the surface of activated T cells; c) increased expression of other cell surface activation molecules including CD23, CD54, CD95 and lymphotoxin-a; and d) inducing immunoglobulin class switching (see Chess, supra, and references 25, 44, and 47-60 cited therein). CD40L binding to CD40 also augments the antigen-presenting functions of dendritic cells, inducing maintenance of high levels of MHC class II antigens and upregulation of accessory molecules including CD58 (LFA-3). CD40L induces cytokine production and tumoricidal activity in peripheral blood monocytes. CD40L also co-stimulates the proliferation of activated T cells, and the co-stimulation is accompanied by the production of IFN-γ, TNF-α and EL2. The expression of CD40L on murine T-helper cells and CD4+ T cells is inhibited by IFN-γ, and is inhibited on T-helper-type 2 cells by TGF-β.

CD40L upregulates the expression of CD54 by cultured Hodgkin and Reed-Stemberg cells. The increased CD54 surface expression is accompanied by increased shedding of surface-bound CD54.

CD40L has also been suggested to be important in the induction of tolerance—CD80 and CD86, which are upregulated by CD40L, interact with CD28 to provide essential co-stimulation of T cells, in concert with T cell receptor activation, that results in full activation of T cells. In the absence of CD80 and CD86-triggered activation of CD28, energy or tolerance occurs as a consequence of antigen triggering (Linsley & Ledbetter, 1993, Ann. Rev. Immunol. 11: 191-212; Jenkins et al., 1993, Curr Opin. Immunol. 5: 361-367; and Boussiotis et al., 1996, Immunol. Rev. 153: 5-26).

The CD40L/CD40 pathway has been implicated in the in vivo priming of CD8+ cytotoxic T lymphocytes (CTSs) by CD4+ T cells. As noted, CD40L expressed on the surface of activated CD4+ T cells interacts with CD40 expressed on dendritic cells, inducing the dendritic cells to express more MHC, and signaling through CD40 can replace the requirement for CD4+T-helper cells in priming CD8+CTL responses. Blockade of CD40L inhibits CTL priming, emphasizing the vital role of CD40L/CD40 interactions in CTL priming by helper T cells (Ridge et al., 1998, Nature 393: 474-478; Schoenberger et al., 1998, Nature 393: 480-483; Bennett et al., 1998, Nature 393: 478-480).

CD40L can also mediate functional interactions of CD4+ T cells with other cells that express CD40, such as fibroblasts, synovial cells and endothelial cells (Yellin et al., 1995, J. Leuko. Biol. 58: 209-216; Yellin et al., 1995, J. Exp. Med. 182: 1857-1864). CD40L induces the expression of CD54 (ICAM-1) and CD106 (VCAM-1) by fibroblasts, as well as increasing fibroblast IL-6, collagenase and collagen production and inducing fibroblast proliferation. Thus, CD40L/CD40 interactions may be involved in the induction of fibrosis associated with autoimmunity and immune responses.

CD40L interaction with CD40 induces endothelial cells to express CD62E (E-selectin), ICAM-1 and VCAM-1. The upregulation of these adhesion molecules may be involved in the binding of inflammatory cells to vascular endothelium and the subsequent migration of the inflammatory cells to sites of inflammation. CD40L blockade retards the migration of leukocytes through endothelial cell barriers. In animal models of autoimmunity, antibodies to CD40L interfere with the accumulation of inflammatory cells at the site of inflammation.

CD40/CD40L interactions have been implicated in diseases having an immune or autoimmune connection. Animal models of immune-related disease in which the CD40L/CD40 pathway has been demonstrated to play a role in the pathology include, for example, murine models of systemic lupus erythematosis (Lupus or SLE; see, e.g., Kalled et al., 1998, J. Immunol. 160: 2158-2165), arthritis (collagen-induced arthritis, see, e.g., Durie et al., 1993, Science 261: 1328-1330), multiple sclerosis (experimental autoimmune encephalomyelitis, EAE; see, e.g., Howard et al., 1999, J. Clin. Invest. 103: 281-290), autoimmune thyroiditis (experimental autoimmune thyroiditis, EAT; see, e.g., Caryanniotis et al., 1997, Immunology 90: 421-426), colitis (hapten-induced colitis; see, e.g., Stuber et al., 1996, J. Exp. Med. 183: 693-698), atherosclerosis and coronary artery disease (see, e.g., Mach et al., 1998, Nature 394: 200-203), and allograft rejection (see, e.g., Parker et al., 1995, Proc. Natl. Acad. Sci. U.S.A. 92: 9560-9564; Kirk et al., 1997, Proc. Natl. Acad. Sci. U.S.A. 94: 8789-8794; Larsen et al., 1996, Nature 381: 434-438 and Blazar et al., 1997, J. Immunol. 158: 29-39).

CD40L antibody trials for treatment of human immune-related diseases include studies in patients with Lupus (see, e.g., Huang et al., 2002, Arthritis Rheum. 46: 1554-1562). A phase I trial demonstrated that anti-CD40L humanized monoclonal antibody (IDEC-131) is safe and well tolerated by patients with Lupus (Davis et al., 2001, J. Rheumatol. 28: 95-101). A phase II study with the IDEC-131 antibody showed improvement in clinical symptoms, but efficacy of the drug over placebo controls was not demonstrated (Kalunian et al., 2002, Arthritis Rheum. 46: 3251-3258). In a phase II study with BG9588 anti-CD40L antibody, clinical efficacy was demonstrated, but the study was terminated due to the occurrence of thromboembolic events (Boumpas et al., 2003, Arthritis Rheum. 48: 719-727).

U.S. Pat. Nos. 5,474,771 (Lederman et al.) and 5,876,950 (Siadak et al.) disclose murine monoclonal antibodies specific for different epitopes of human gp39. WO95/06666 (Noelle & Foy) discloses murine anti-gp39 antibodies.

U.S. Pat. No. 6,328,964 (Noelle & Claassen) discloses methods for the treatment of multiple sclerosis using gp39-specific antibodies.

U.S. Pat. No. 5,747,037 (Noelle et al.), and EP0721469B1 (Ledbetter et al.) and its U.S. counterpart U.S. Pat. No. 5,869,049 disclose anti-human monoclonal (mouse) antibodies specific for gp39. U.S. Pat. No. 5,876,718 (Noelle et al.) discloses methods of inducing T cell non-responsiveness to transplanted tissues and of treating graft-versus-host disease with anti-gp39 monoclonal (mouse) antibodies. EP0742721131 (Noelle et al.) discloses methods of inhibiting a humoral immune response to a thymus-dependent antigen that use anti-gp39 monoclonal (mouse) antibodies. U.S. Pat. No. 6,375,950 describes methods for inducing T cell unresponsiveness to donor tissue or organs in a transplant recipient through use of anti-gp39 monoclonal (murine) antibodies.

EP1005372B1 (De Boer et al.) describes methods for the selective killing of autoreactive CD40L+ T cells using anti-CD40L monoclonal (mouse) antibody-toxin fusion proteins.

U.S. Pat. No. 6,340,459 (Yellin et al.) describes the use of murine anti gp39 monoclonal antibody 5c8 for the treatment or prevention of reperfusion injury.

EP0831906B1 (Claassen et al.) describes methods for the treatment of T cell-mediated tissue destruction in autoimmune diseases such as multiple sclerosis using anti-gp39 monoclonal (mouse) antibodies. Antibodies used in therapeutic approaches in the prior art have been divalent antibodies of murine origin.

A number of smaller antigen binding fragments of naturally occurring antibodies have been identified following protease digestion. These include, for example, the “Fab fragment” (VL-CL-CH1-VH), “Fab′ fragment” (a Fab with the heavy chain hinge region) and “F(ab′)2 fragment” (a dimer of Fab′ fragments joined by the heavy chain hinge region). Recombinant methods have been used to generate even smaller antigen-binding fragments, referred to as “single chain Fv” (variable fragment) or “scFv,” consisting of VL and VH joined by a synthetic peptide linker.

While the antigen binding unit of a naturally-occurring antibody (e.g., in humans and most other mammals) is generally known to be comprised of a pair of V regions (VE/VH), camelid species express a large proportion of fully functional, highly specific antibodies that are devoid of light chain sequences. The camelid heavy chain antibodies are found as homodimers of a single heavy chain, dimerized via their constant regions. The variable domains of these camelid heavy chain antibodies are referred to as VHH domains and retain the ability, when isolated as fragments of the VH chain, to bind antigen with high specificity ((Hamers-Casterman et al., 1993, Nature 363: 446-448; Gahroudi et al., 1997, FEBS Lett. 414: 521-526). Antigen binding single VH domains have also been identified from, for example, a library of murine VH genes amplified from genomic DNA from the spleens of immunized mice and expressed in E. coli (Ward et al., 1989, Nature 341: 544-546). Ward et al. named the isolated single VH domains “dAbs,” for “domain antibodies.” The term “dAb” will refer herein to an antibody single variable domain (VH or VL) polypeptide that specifically binds antigen. A “dAb” binds antigen independently of other V domains; however, as the term is used herein, a “dAb” can be present in a homo- or heteromultimer with other VH or VL domains where the other domains are not required for antigen binding by the dAb, i.e., where the dAb binds antigen independently of the additional VH or VL domains.

Antibody single variable domains, for example, VHH, are the smallest antigen-binding antibody unit known. For use in therapy, human antibodies are preferred, primarily because they are not as likely to provoke an immune response when administered to a patient. As noted above, isolated non-camelid VH domains tend to be relatively insoluble and are often poorly expressed. Comparisons of camelid VHH with the VH domains of human antibodies reveals several key differences in the framework regions of the camelid VHH domain corresponding to the VH/VL interface of the human VH domains. Mutation of these residues of human VH3 to more closely resemble the VHH sequence (specifically Gly 44→Glu, Leu 45→Arg and Trp 47→Gly) has been performed to produce “camelized” human VH domains that retain antigen binding activity (Davies & Riechmann, 1994, FEBS Lett. 339: 285-290) yet have improved expression and solubility. (Variable domain amino acid numbering used herein is consistent with the Kabat numbering convention (Kabat et al., 1991, Sequences of Immunological Interest, 5th ed. U.S. Dept. Health & Human Services, Washington, D.C.)) WO 03/035694 (Muyldermans) reports that the Trp 103→Arg mutation improves the solubility of non-camelid VH domains. Davies & Riechmann (1995, Biotechnology N.Y. 13: 475-479) also report production of a phage-displayed repertoire of camelized human VH domains and selection of clones that bind hapten with affinities in the range of 100-400 nM, but clones selected for binding to protein antigen had weaker affinities.

While many antibodies and their derivatives are useful for diagnosis and therapy, the ideal pharmacokinetics of antibodies are often not achieved for a particular application. In order to provide improvement in the pharmacokinetics of antibody molecules, the present invention provides single domain variable region polypeptides that are linked to polymers which provide increased stability and half-life. The attachment of polymer molecules (e.g., polyethylene glycol; PEG) to proteins is well established and has been shown to modulate the pharmacokinetic properties of the modified proteins. For example, PEG modification of proteins has been shown to alter the in vivo circulating half-life, antigenicity, solubility, and resistance to proteolysis of the protein (Abuchowski et al., J. Biol. Chem. 1977, 252:3578; Nucci et al., Adv. Drug Delivery Reviews 1991, 6:133; Francis et al., Pharmaceutical Biotechnology Vol. 3 (Borchardt, R. T. ed.); and Stability of Protein Pharmaceuticals: in vivo Pathways of Degradation and Strategies for Protein Stabilization 1991 pp 235-263, Plenum, N.Y.).

Both site-specific and random PEGylation of protein molecules is known in the art (See, for example, Zalipsky and Lee, Poly(ethylene glycol) Chemistry: Biotechnical and Biomedical Applications 1992, pp 347-370, Plenum, N.Y.; Goodson and Katre, 1990, Bio/Technology, 8:343; Hershfield et al., 1991, PNAS 88:7185). More specifically, random PEGylation of antibody molecules has been described at lysine residues and thiolated derivatives (Ling and Mattiasson, 1983, Immunol. Methods 59: 327; Wilkinson et al., 1987, Immunol. Letters, 15: 17; Kitamura et al., 1991, Cancer Res. 51:4310; Delgado et al., 1996 Br. J. Cancer, 73: 175; Pedley et al., 1994, Br. J. Cancer, 70:1126).

SUMMARY

The invention relates to antibody polypeptides that monovalently bind CD40L. Because of the clear importance of CD40L in the production of antibodies, the CD40/CD40L interaction and pathways present important targets for the development of therapeutic approaches for the treatment of diseases and disorders that involve inappropriate or excessive antibody responses, such as autoimmune diseases. Antibody polypeptides that are monovalent for binding of CD40L can inhibit CD40L activity, including binding and activation of CD40 on the B cell surface and downstream effects, while avoiding potential undesirable effects that can occur with antibodies capable of divalent or multivalent binding of CD40L. Monovalent anti-CD40L antibody polypeptides can also be applied to any of a number of uses for which standard divalent antibodies are also used, e.g., in vivo imaging and diagnosis.

In one aspect, the antibody polypeptide consists of or comprises a single immunoglobulin variable domain that specifically binds and antagonizes the activity of CD40L, preferably without substantially agonizing CD40 activity. In another aspect, because human antibodies will avoid the generation of an immune response to the antibodies when administered to human subjects for the treatment or prevention of disease, the antibody polypeptide is a human antibody polypeptide that monovalently binds CD40L, preferably without substantially agonizing CD40 activity.

In summary then, in one embodiment, the invention provides a human antibody polypeptide that is monovalent for binding to CD40L (gp39).

In one embodiment, the human antibody polypeptide dissociates from human CD40L with a Kd in the range of 50 nM to 20 pM, inclusive, as measured by surface plasmon resonance. For example, the IQ for human CD40L can be 25 nM to 20 pM, 10 nM to 20 pM, 5 nm to 20 pM, 1 nM to 20 pM, 0.5 nM to 20 pM, 0.1 nM to 20 pM, 0.1 nM to 50 nM, 75 pM to 20 pM or even 50 pM to 20 pM.

Unless otherwise stated, all ranges described herein are inclusive of the specific endpoints.

In another embodiment, the antibody polypeptide inhibits the binding of CD40L to CD40.

In another embodiment, the binding of the antibody polypeptide to CD40L does not substantially agonize CD40 activity.

In another embodiment, the human antibody polypeptide inhibits the binding of CD40 to CD40L, and does not substantially agonize signaling by CD40.

In another embodiment, the binding of the antibody polypeptide to CD40L does not substantially induce JNK phosphorylation in Jurkat T-cells.

In another embodiment, the binding of the antibody polypeptide to CD40L does not substantially induce IFN-γ secretion by Jurkat. T-cells co-stimulated with anti-CD3 antibody.

In another embodiment, the presence of the antibody polypeptide in a standard platelet aggregation assay does not result in aggregation of more than 25% over the aggregation observed in a negative control assay performed without the addition of antibody.

In another embodiment, the human antibody polypeptide comprises a single immunoglobulin variable domain that binds CD40L. In a preferred embodiment, the single immunoglobulin variable domain is a VH or a VL domain.

In another embodiment, the antibody polypeptide is selected from the group consisting of a dAb, a FAb, an scFv, an Fv, or a disulfide-bonded Fv.

In another embodiment, the human antibody polypeptide is PEG-linked. In one embodiment, the PEG is covalently linked to the human antibody polyepeptide. In one preferred embodiment, the PEG-linked human antibody polypeptide has a hydrodynamic size of at least 24 kD. In another preferred embodiment, the PEG is linked to the antibody polypeptide at a cysteine or lysine residue. In another preferred embodiment, the total PEG size is from 20 to 60 kD, inclusive. In another preferred embodiment, the PEG-linked human antibody polypeptide has a hydrodynamic size of at least 200 kD.

In one embodiment, the antibody polypeptide has an increased in vivo half-life relative to the same antibody polypeptide composition lacking polyethylene glycol.

In another embodiment, the tα-half life of the antibody polypeptide composition is increased by 10% or more. In another embodiment, the tα-half life of the antibody polypeptide composition is increased by 50% or more. In another embodiment, the tα-half life of the antibody polypeptide composition is increased by 2× or more. In another embodiment, the tα-half life of the antibody polypeptide composition is increased by 5× or more, e.g., 10×, 15×, 20×, 25×, 30×, 40×, or more. In another embodiment, the tα-half life of the antibody polypeptide composition is increased by 50× or more.

In another embodiment, the PEG-linked antibody polypeptide has a tα half-life of 0.25 to 6 hours, inclusive. In another embodiment, the tα-half-life is in the range of 30 minutes to 12 hours, inclusive. In another embodiment, the tα-half life of the antibody polypeptide composition is in the range of 1 to 6 hours.

In another embodiment, the tβ-half life of the antibody polypeptide composition is increased by 10% or more. In another embodiment, the tβ-half life of the antibody polypeptide composition is increased by 50% or more. In another embodiment, the tβ-half life of the antibody polypeptide composition is increased by 2× or more. In another embodiment, the tβ-half life of the antibody polypeptide composition is increased by 5× or more, e.g., 10×, 15×, 20×, 25×, 30×, 40×, or more. In another embodiment, the tβ-half life of the antibody polypeptide composition is increased by 50× or more.

In another embodiment, the antibody polypeptide composition has a to half-life of 1 to 170 hours, inclusive. In another embodiment, the tβ-half life is in the range of 12 to 48 hours, inclusive. In another embodiment, the tβ-half life is in the range of 12 to 26 hours, inclusive.

In addition, or alternatively to the above criteria, the present invention provides a dAb containing composition comprising a ligand according to the invention having an AUG value (area under the curve) in the range of 1 mg.min/ml or more. In one embodiment, the lower end of the range is 5, 10, 15, 20, 30, 100, 200 or 300 mg.min/ml. In addition, or alternatively, a ligand or composition according to the invention has an AUC in the range of up to 600 mg.min/ml. In one embodiment, the upper end of the range is 500, 400, 300, 200, 150, 100, 75 or 50 mg.min/ml. Advantageously a ligand according to the invention will have an AUC in the range selected from the group consisting of the following: 15 to 150 mg.min/ml, 15 to 100 mg.min/ml, 15 to 75 mg.min/ml, and 15 to 50 mg.min/ml.

In another embodiment, the antibody polypeptides described herein can be linked to human serum albumin (HSA), which also has the effect of increasing the in vivo half life of the molecule. The human serum albumin coding sequences can be obtained by PCR using primers derived from the cDNA sequence available at GenBank Accession No. NM000477. Such coding sequences can be fused to the coding sequence for a monovalent anti-CD40L antibody polypeptide as described herein, and the fusion can be expressed by one of skill in the art.

In another embodiment, the tα-half life of the HSA-linked human antibody polypeptide composition is increased by 10% or more.

In another embodiment, the tα-half life of the HSA-linked human antibody polypeptide composition is in the range of 0.25 hours to 6 hours.

In another embodiment, the tβ-half life of the HSA-linked human antibody polypeptide composition is increased by 10% or more.

In another embodiment, the 43-half life of the HSA-linked human antibody polypeptide composition is in the range of 12 to 48 hours.

In another embodiment, the human antibody polypeptide comprises an amino acid sequence selected from the group consisting of SEQ ID NOs 7-82.

In another embodiment, the human antibody polypeptide inhibits binding of CD40L to CD40 with an IC50 in the range of 20 μM to 1.5 pM, inclusive; IC50 for inhibition of CD40L binding to CD40 in any embodiment described herein is preferably measured as described herein in Example 6. The ICso can preferably be in the range of 20 pM to 1 μM, 20 pM to 900 nM, 20 pM to 800 nM, 20 pM to 700 nM, 20 pM to 600 nM, 20 pM to 500 nM, 20 pM to 400 nM, 20 pM to 300 nM, 20 pM to 200 nM, 20 pM to 100 nM, or 20 pM to 50 nM. Further acceptable or preferred ranges include, for example, 50 pM to 1 pM, 100 pM to 500 nM, 125 pM to 250 nM, 150 pM to 200 nM, 150 pM to 100 nM and 200 pM to 50 nM.

In another embodiment, the antibody polypeptide is fused to a second antibody polypeptide which binds a ligand other than CD40L. In a preferred embodiment, the antibody polypeptide which binds a ligand other than CD40L binds a ligand selected from the group consisting of HSA, TNFα, IL-1, IL-2, IL-4, IL-6, IL-8, IL-12, IFN-γ, CD2, CD4, CD8, CTLA4, LFA1, LFA3 and VLA4.

In another embodiment, the human antibody polypeptide is free of an Fc domain. The limits of an Fc domain are set out in Kabat et al. (1991, Sequences of Immunological Interest, 5th ed. U.S. Dept. Health & Human Services, Washington, D.C.; incorporated herein by reference). In the alternative, an Fc domain consists of the CH2-CH3 regions, optionally including a hinge region linked to the CH2. In a preferred embodiment, the human antibody polypeptide does not mediate platelet aggregation in a standard platelet aggregation assay.

The invention further encompasses a human antibody polypeptide which has an amino acid sequence at least 85% identical to a sequence selected from the group consisting of SEQ ID NOs 7-82, which antibody polypeptide specifically and monovalently binds CD40L.

The invention further encompasses an antigen-binding polypeptide, the polypeptide comprising a single immunoglobulin variable domain which specifically and monovalently binds CD40L. Recited differently, the invention further encompasses a polypeptide comprising a moiety which specifically binds CD40L, which moiety consists of a single immunoglobulin variable domain.

In one embodiment, the polypeptide consists of a human single immunoglobulin variable domain.

In another embodiment, the polypeptide has a Kd for human CD40L in the range of 50 nM to 20 pM, inclusive, as determined by surface plasmon resonance. For example, the Kd for human CD40L can be 25 nM to 20 pM, 10 nM to 20 pM, 5 nm to 20 pM, 1 nM to 20 pM, 0.5 nM to 20 pM, 0.1 nM to 20 pM, 75 pM to 20 pM or even 50 pM to 20 pM.

In another embodiment, the polypeptide inhibits the binding of CD40L to CD40.

In another embodiment, the polypeptide inhibits the binding of CD40 to CD40L and has an IC50 in the range of 20 pM to 1.5 μM, inclusive. For example, the IC50 can be in the range of 20 pM to 1 μM, 20 pM to 900 nM, 20 pM to 800 nM, 20 pM to 700 nM, 20 pM to 600 nM, 20 pM to 500 nM, 20 pM to 400 nM, 20 pM to 300 mM, 20 pM to 200 nM, 20 pM to 100 nM, or 20 pM to 50 nM. Further acceptable or preferred ranges include, for example, 50 pM to 1 μM, 100 pM to 500 nM, 125 pM to 250 nM, 150 pM to 200 nM, 150 pM to 100 nM and 200 pM to 50 nM.

In another embodiment, the binding of the polypeptide to CD40L does not substantially agonize CD40 activity.

In another embodiment, the binding of the polypeptide to CD40L does not substantially induce JNK phosphorylation in Jurkat T-cells.

In another embodiment, the binding of the polypeptide to CD40L does not substantially induce IFN-γ secretion by Jurkat T-cells co-stimulated with anti-CD3 antibody.

In another embodiment, the presence of the antibody polypeptide in a standard platelet aggregation assay does not result in aggregation more than 25% over the aggregation observed in a negative control assay lacking antibody polypeptide.

In another embodiment, the single immunoglobulin variable domain is a human single immunoglobulin variable domain.

In another embodiment, the single immunoglobulin variable domain is a VH or a VL domain.

In one embodiment, the polypeptide is PEG-linked. In one embodiment, the PEG is covalently linked. In one preferred embodiment, the PEG-linked antigen-binding polypeptide has a hydrodynamic size of at least 24 kD. In another preferred embodiment, the PEG is linked to the antigen-binding polypeptide at a cysteine or lysine residue. In another preferred embodiment, the total PEG size is from 20 to 60 kD, inclusive. In another preferred embodiment, the PEG-linked antigen-binding polypeptide has a hydrodynamic size of at least 200 kD.

In another embodiment, the PEG-linked polypeptide has an increased in vivo half-life relative to the same polypeptide composition lacking linked polyethylene glycol. In another embodiment, the tα-half life of the polypeptide composition is increased by 10% or more. In another embodiment, the tα-half life of the polypeptide composition is increased by 50% or more. In another embodiment, the tα-half life of the polypeptide composition is increased by\'2× or more. In another embodiment, the tα-half life of the polypeptide composition is increased by 5× or more, e.g., 10×, 15×, 20×, 25×, 30×, 40×, or more. In another embodiment, the tα-half life of the polypeptide composition is increased by 50× or more.

In another embodiment, the PEG-linked antibody polypeptide has a tα half-life of 0.25 to 6 hours, inclusive. In another embodiment, the tα half-life is in the range of 30 minutes to 12 hours, inclusive. In another embodiment, the tα-half life of the polypeptide composition is in the range of 1 to 6 hours.

In another embodiment, the tβ-half life of the polypeptide composition is increased by 10% or more. In another embodiment, the tβ-half life of the polypeptide composition is increased by 50% or more. In another embodiment, the tβ-half life of the polypeptide composition is increased by 2× or more. In another embodiment, the tβ-half life of the polypeptide composition is increased by 5× or more, e.g., 10×, 15×, 20×, 25×, 30×, 40×, or more. In another embodiment, the tβ-half life of the polypeptide composition is increased by 50× or more.

In another embodiment, the antibody polypeptide composition has a tβhalf-life of 1 to 170 hours, inclusive. In another embodiment, the t3-half life is in the range of 12 to 48 hours, inclusive. In another embodiment, the tβ-half life is in the range of 12 to 26 hours, inclusive.

In another embodiment, the composition has an AUC value (area under the curve) in the range of 1 mg.min/ml or more. In one embodiment, the lower end of the range is 5, 10, 15, 20, 30, 100, 200 or 300 mg.min/ml. In addition, or alternatively, a ligand or composition according to the invention has an AUC in the range of up to 600 mg.min/ml. In one embodiment, the upper end of the range is 500, 400, 300, 200, 150, 100, 75 or 50 mg.min/ml. Advantageously a ligand according to the invention will have an AUC in the range selected from the group consisting of the following: 15 to 150 mg.min/ml, 15 to 100 mg.min/ml, 15 to 75 mg.min/ml, and 15 to 50 mg.min/ml.

In another embodiment, the antibody polypeptide is linked to human serum albumin (HSA). In another embodiment, the antibody polypeptide has an increased in vivo half-life relative to the same polypeptide composition lacking linked HSA. In another embodiment, the antibody polypeptide has a tα-half life that is increased by 10% or more relative to a molecule lacking linked HSA. In another embodiment, the tα-half life of the polypeptide composition is in the range of 0.25 minutes to 6 hours. In another embodiment, the tβ-half life of the polypeptide composition is increased by 10% or more. In another embodiment, the tβ-half life is in the range of 12 to 48 hours.

In another embodiment, the antigen-binding polypeptide comprises an amino acid sequence selected from the group consisting of SEQ ID NOs 7-82.

In another embodiment, the antigen-binding polypeptide is free of an Fc domain.

In another aspect, the invention encompasses an immunoglobulin variable domain polypeptide which has an amino acid sequence at least 85% identical to a sequence selected from the group consisting of SEQ ID NOs 7-82, which polypeptide specifically and monovalently binds CD40L.

In one embodiment, the immunoglobulin variable domain polypeptide antagonizes the binding of CD40L to CD40.

In another embodiment, the immunoglobulin variable domain polypeptide inhibits the binding of CD40 to CD40L and has an IC50 in the range of 20 pM to 1.5 μM, inclusive. For example, the IC50 can be in the range of 20 pM to 1 μM, 20 pM to 900 nM, 20 pM to 800 nM, 20 pM to 700 nM, 20 pM to 600 nM, 20 pM to 500 nM, 20 pM to 400 nM, 20 pM to 300 nM, 20 pM to 200 nM, 20 pM to 100 nM, or 20 pM to 50 nM. Further acceptable or preferred ranges include, for example, 50 pM to 1 μM, 100 pM to 500 nM, 125 pM to 250 nM, 150 pM to 200 nM, 150 pM to 100 nM and 200 pM to 50 nM.

In another embodiment, the immunoglobulin variable domain polypeptide inhibits the interaction of CD40 with CD40L, but does not substantially agonize intracellular signaling by CD40. In a preferred embodiment, the binding of the polypeptide to CD40L does not substantially induce JNK phosphorylation in Jurkat T-cells. In another preferred embodiment, the binding of the polypeptide to CD40L does not substantially induce IFN-γ secretion by Jurkat T-cells co-stimulated with anti-CD3 antibody. In another preferred embodiment, the binding of the antibody polypeptide to CD40L does not substantially induce platelet aggregation in a platelet aggregation assay.

In another embodiment, the antigen-binding polypeptide further comprises a second antibody polypeptide which binds a ligand other than CD40L. In a preferred embodiment, the second antibody polypeptide binds a ligand selected from the group consisting of HSA, TNFα, IL-1; IL-2, IL-4, IL-6, IL-8, IL-12, IL-18, CD2, CD4, CD8, CTLA4, LFA1, LFA3 and VLA4.

The invention further encompasses a method of antagonizing the binding of CD40 to

CD40L in an individual, the method comprising administering a monovalent anti-CD40L antibody polypeptide as described herein to the individual, wherein the polypeptide antagonizes the binding of CD40 to CD40L in the individual.

The invention further encompasses a method of antagonizing an activity of CD40 or CD40L in an individual, the method comprising administering a a monovalent anti-CD40L antibody polypeptide as described herein to the individual, wherein the polypeptide antagonizes an activity of CD40 or CD40L or both.

The invention further encompasses a composition comprising an extended release formulation comprising a monovalent anti-CD40L antibody polypeptide, preferably, but not limited to, a polypeptide comprising a single immunoglobulin variable domain that binds CD40L. In one embodiment, the single immunoglobulin variable domain is a non-human mammalian single immunoglobulin variable domain. In another embodiment, the single immunoglobulin variable domain is a human single immunoglobulin variable domain.

The invention further encompasses a method of treating or preventing a disease or disorder mediated by CD40L in an individual in need of such treatment, the method comprising administering to the individual a therapeutically effective amount of a composition comprising a monovalent anti-CD40L antibody polypeptide, preferably a composition comprising a single human immunoglobulin variable domain that binds CD40L. In one embodiment, the disease or disorder is an autoimmune disease or disorder.

The invention further encompasses a method of treating or preventing a symptom of systemic lupus erythematosus (SLE) in an individual, the method comprising administering a monovalent anti-CD40L antibody polypeptide to said individual in an amount effective to treat or prevent a symptom of SLE.

The invention further encompasses an antibody polypeptide that is monovalent for binding to CD40L, wherein the antibody polypeptide comprises a universal framework.

In one embodiment, the universal framework comprises a VH framework selected from the group consisting of DP47, DP45 and DP38, and/or the VL framework is DPK9.

In another embodiment, the antibody polypeptide comprises a generic ligand binding site.

In another embodiment, the generic ligand binding site binds a generic ligand selected from the group consisting of protein A, protein L and protein G.

In another embodiment, the antibody polypeptide comprises a variable domain having one or more framework regions comprising an amino acid sequence that is the same as the amino acid sequence of a corresponding framework region encoded by a human germline antibody gene segment, or the amino acid sequences of one or more of the framework regions collectively comprises up to 5 amino acid differences relative to the amino acid sequence of the corresponding framework region encoded by a human germline antibody gene segment.

In another embodiment, the antibody polypeptide comprises a variable domain, wherein the amino acid sequences of FW1, FW2, FW3 and FW4 are the same as the amino acid sequences of corresponding framework regions encoded by a human germline antibody gene segment, or the antibody sequences of FW1, FW2, FW3 and FW4 collectively contain up to 10 amino acid differences relative to the amino acid sequences of corresponding framework regions encoded by the human germline antibody gene segment.

In another embodiment, the antibody polypeptide comprises an antibody variable domain comprising FW1, FW2 and FW3 regions, and the amino acid sequence of said FW1, FW2 and FW3 are the same as the amino acid sequences of corresponding framework regions encoded by human germline antibody gene segments. In another embodiment, the human germline antibody gene segment is selected from the group consisting of DP47, DP45, DP48 and DPK9.

As used herein, the term “human” when applied to an antibody polypeptide or to an immunoglobulin variable domain means that the polypeptide has a sequence derived from a human immunoglobulin. A sequence is “derived from” a human immunoglobulin coding sequence when the sequence is either: a) isolated from a human individual or from cells or a cell line from a human individual; b) isolated from a library of cloned human antibody gene sequences (or a library of human antibody V domain sequences); or c) when a cloned human antibody gene sequence (or a cloned human V region sequence (including, e.g., a germline V gene segment)) was used to generate one or more diversified sequences that were then selected for binding to a desired target antigen. The term “human” as applied herein to an antibody polypeptide or to an immunoglobulin variable domain does not encompass an immunoglobulin from another species, e.g., mouse, camel, etc., that has been “humanized” through grafting of human constant region sequences onto an antibody polypeptide (i.e., replacing non-human constant regions with human constant regions) or through grafting of human V region framework sequences onto an immunoglobulin variable domain from a non-human mammal (i.e., replacing non-human framework regions of a V domain with human framework regions).

At a minimum, a human variable domain has at least 85% amino acid similarity (including, for example, 87%, 90%, 93%, 95%, 97%, 99% or higher similarity) to a naturally-occurring human immunoglobulin variable domain sequence.

As used herein, the term “domain” refers to a folded protein structure which retains its tertiary structure independently of the rest of the protein. Generally, domains are responsible for discrete functional properties of proteins, and in many cases may be added, removed or transferred to other proteins without loss of function of the remainder of the protein and/or of the domain.

By “single immunoglobulin variable domain” is meant a folded polypeptide domain which comprises a sequence characteristic of immunoglobulin variable domains and which specifically binds an antigen (e.g., dissociation constant of 500 nM or less). A “single immunoglobulin variable domain” therefore includes complete antibody variable domains as well as modified variable domains, for example in which one or more loops have been replaced by sequences which are not characteristic of antibody variable domains, or antibody variable domains which have been truncated or comprise N- or C-terminal extensions, as well as folded fragments of variable domains which retain a dissociation constant of 500 nM or less (e.g., 450 nM or less, 400 nM or less, 350 nM or less, 300 nM or less, 250 nM or less, 200 nM or less, 150 nM or less, 100 nM or less) and the target antigen specificity of the full-length domain. Where necessary or in case of any doubt, the numbering convention and boundaries set forth by Kabat et al. (1991, supra) are applicable to immunoglobulin variable and constant domains referred to herein.

An antibody single variable domain polypeptide, as used herein refers to a mammalian single immunoglobulin variable domain polypeptide, preferably human, but also includes rodent (for example, as disclosed in WO00/29004, the contents of which are incorporated herein in their entirety) or camelid VHH dAbs. Camelid dAbs are antibody single variable domain polypeptides which are derived from species including camel, llama, alpaca, dromedary, and guanaco, and comprise heavy chain antibodies naturally devoid of light chain: VHH. VHH molecules are about 10× smaller than IgG molecules, and as single polypeptides, they are very stable, resisting extreme pH and temperature conditions. Moreover, camelid antibody single variable domain polypeptides are resistant to the action of proteases. Camelid antibodies are described in, for example, U.S. Pat. Nos. 5,759,808; 5,800,988; 5,840,526; 5,874,541; 6,005,079; and 6,015,695, the contents of each of which are incorporated herein in their entirety. Camelid VHH antibody single variable domain polypeptides useful according to the invention include a class of camelid antibody single variable domain polypeptides having human-like sequences, wherein the class is characterized in that the VHH domains carry an amino acid from the group consisting of glycine, alanine, valine, leucine, isoleucine, proline, phenylalanine, tyrosine, tryptophan, methionine, serine, threonine, asparagine, or glutamine at position 45, such as for example L45, and further comprise a tryptophan at position 103 according to the Kabat numbering. Humanized camelid VHH polypeptides are taught, for example in WO04/041862, the teachings of which are incorporated herein in their entirety. It will be understood by one of skill in the art that naturally occurring camelid antibody single variable domain polypeptides may be modified according to the teachings of WO04/041862 (e.g., amino acid substitutions at positions 45 and 103) to generate humanized camelid VHH polypeptides.

The phrase “single immunoglobulin variable domain polypeptide” encompasses not only an isolated single immunoglobulin variable domain polypeptide, but also larger polypeptides that comprise a monomer of a single immunoglobulin variable domain polypeptide sequence. A “domain antibody” or “dAb” is equivalent to a “single immunoglobulin variable domain polypeptide” as the term is used herein. With regard to a single immunoglobulin variable domain polypeptide, the binding to antigen, e.g., CD40L, is mediated by the single immunoglobulin V domain without a requirement for a complementary V domain.

According to the invention, the terms “antibody single variable domain polypeptide”, “antibody single variable domain”, “single antibody variable domain”, and “immunoglobulin single variable domain” are understood to be equivalent.

As used herein, the phrase “sequence characteristic of immunoglobulin variable domains” refers to an amino acid sequence that is homologous, over 20 or more, 25 or more, 30 or more, 35 or more, 40 or more, 45 or more, or even 50 or more contiguous amino acids, to a sequence comprised by an immunoglobulin variable domain sequence.

As used herein, the terms “homology” or “similarity” refer to the degree with which two nucleotide or amino acid sequences structurally resemble each other. As used herein, sequence “similarity” is a measure of the degree to which amino acid sequences share similar amino acid residues at corresponding positions in an alignment of the sequences. Amino acids are similar to each other where their side chains are similar. Specifically, “similarity” encompasses amino acids that are conservative substitutes for each other. A “conservative” substitution is any substitution that has a positive score in the blosum62 substitution matrix (Hentikoff and Hentikoff, 1992, Proc. Natl. Acad. Sci. USA 89: 10915-10919). By the statement “sequence A is n % similar to sequence B” is meant that n % of the positions of an optimal global alignment between sequences A and B consists of identical amino acids or conservative substitutions. Optimal global alignments can be performed using the following parameters in the Needleman-Wunsch alignment algorithm:

For polypeptides: Substitution matrix: blosum62. Gap scoring function: −A−B*LG, where A=11 (the gap penalty), B=1 (the gap length penalty) and LG is the length of the gap.

For nucleotide sequences: Substitution matrix: 10 for matches, 0 for mismatches. Gap scoring function: −A−B*LG where A=50 (the gap penalty), B=3 (the gap length penalty) and LG is the length of the gap.

Typical conservative substitutions are among Met, Val, Leu and Ile; among Ser and Thr; among the residues Asp, Glu and Asn; among the residues Gln, Lys and Arg; or aromatic residues Phe and Tyr.

As used herein, two sequences are “homologous” or “similar” to each other where they have at least 85% sequence similarity to each other, including, e.g., 90%, 95%, 97%, 99% or even 100% sequence similarity, when aligned using either the Needleman-Wunsch algorithm or the “BLAST 2 sequences” algorithm described by Tatusova & Madden, 1999, FEMS Microbial Lett. 174:247-250. Where amino acid sequences are aligned using the “BLAST 2 sequences algorithm,” the Blosum 62 matrix is the default matrix.

As used herein, the terms “inhibit,” “inhibits” and “inhibited” refer to a decrease in a given measurable activity by at least 10% relative to a reference. Where inhibition is desired, such inhibition is preferably at least 20%, 30%, 40%, 50%, 60%, 70%, 80%, 90% or more, up to and including 100%, i.e., complete inhibition or absence of the given activity. One way that inhibition of CD40L binding to CD40 is measured is as described in Example 6 herein.

As used herein, the terms “activate,” “activates” and “activated” refer to an increase in a given measurable activity by at least 5% relative to a reference, for example, at least 10%, 25%, 50%, 75% or even 100%.

As used herein, the term “antagonist” refers to an agent that inhibits at least one activity mediated by CD40L. An activity is “antagonized” if the activity is reduced by at least 10%, and preferably at least 20%, 30%, 40%, 50%, 60%, 70%, 80%, 90%, 95%, 97% or even 100% (i.e., no activity) in the presence, relative to the absence of an antagonist. An antagonist as the term is used herein preferably comprises a single immunoglobulin variable domain that binds monovalently to CD40L.

As used herein, the term “agonist” refers to an agent that activates at least one activity mediated by CD40L, either alone or when combined with another co-stimulus, relative to a reference. An activity is “agonized” if the activity is increased by at least 10%, e.g., 50%, in the presence, relative to the absence of an agonist.

As used herein, the term “epitope” refers to a unit of structure conventionally bound by an immunoglobulin VHNL pair. Epitopes define the minimum binding site for an antibody, and thus represent the target of specificity of an antibody. In the case of a single immunoglobulin variable domain, an epitope represents the unit of structure bound by a single variable domain in isolation. That is, the binding site is provided by one, single immunoglobulin variable domain.

As used herein, the term “extended release” or the equivalent terms “controlled release” or “slow release” refer to drug formulations that release active drug, such as a polypeptide drug, over a period of time following administration to an individual. Extended release of polypeptide drugs, which can occur over a range of desired times, e.g., minutes, hours, days, weeks or longer, depending upon the drug formulation, is in contrast to standard formulations in which substantially the entire dosage unit is available for immediate absorption or immediate distribution via the bloodstream. Preferred extended release formulations result in a level of circulating drug from a single administration that is sustained, for example, for 8 hours or more, 12 hours or more, 24 hours or more, 36 hours or more, 48 hours or more, 60 hours or more, 72 hours or more 84 hours or more, 96 hours or more, or even, for example, for 1 week or 2 weeks or more, for example, 1 month or more.

As used herein, a “CD40L activity” is an activity involving or resulting from the binding of CD40L to CD40, and includes, but is not limited to binding to CD40 (assayed, for example, according to the method described in Example 6), activation of Jun-N-terminal Kinase (JNK), the induction of T cells to produce and secrete cytokines including, for example, IL-10, IFN-γ and TNF-α., and the mediation of platelet aggregation. Assays for these activities are provided herein below.

As used herein, the term “does not substantially agonize” means that a given agent, e.g., an anti-CD40L antibody polypeptide, does not activate one or more of the CD40L activities including Jun-N-terminal kinase activation (phosphorylation) in Jurkat T cells and induction of IFN-γ production or secretion in anti-CD3-stimulated Jurkat T cells, as the term “activate” is defined herein.

As used herein, the term “antibody polypeptide” refers to a polypeptide which either is an antibody or is a part of an antibody, modified or unmodified, which retains the ability to specifically bind antigen. Thus, the term antibody polypeptide includes an antigen-binding heavy chain, light chain, heavy chain-light chain dimer, Fab fragment, F(ab′)2 fragment, dAb, or an Fv fragment, including a single chain Fv (scFv). The phrase “antibody polypeptide” is intended to encompass recombinant fusion polypeptides that comprise an antibody polypeptide sequence that retains the ability to specifically bind antigen in the context of the fusion.

As used herein, the term “monovalent” means that a given antibody polypeptide or single immunoglobulin variable domain polyepeptide can bind only a single molecule of its target. Naturally-occurring antibodies are generally divalent, in that they have two functional antigen-binding arms, each comprising a VH and a VL domain. Where steric hindrance is not an issue, a divalent antibody can bind two separate molecules of the same antigen. In contrast, a “monovalent” antibody has the capacity to bind only one such antigen molecule. As the term is used herein, a “monovalent” antibody can also comprise more than one antigen binding site, e.g., two antigen binding sites, but the binding sites must be for different antigens, such that the antibody can only bind one molecule of CD40L at a time. The antigen-binding domain of a monovalent antibody can comprise a VH and a VL domain, but preferably comprises only a single immunoglobulin variable domain, i.e., a VH or a VL domain, that has the capacity to bind CD40L without the need for a corresponding VL or VH domain, respectively. A monovalent antibody lacks the capacity to cross link molecules of a single antigen.

As used herein, the term “standard platelet aggregation assay” means the assay described in the section herein below, entitled “Platelet Aggregation Assay.”

As used herein, the terms “VH domain” and “VL domain” refer to immunoglobulin variable regions as defined by Kabat et al. (supra), which is incorporated herein by reference.

As used herein, “linked” refers to the attachment of a polymer moiety, such as PEG to an amino acid residue of an antibody polypeptide. Attachment of a PEG polymer to an amino acid residue of an antibody polypeptide, e.g., an anti-CD40L dAb, is referred to as “PEGylation” and may be achieved using several PEG attachment moieties including, but not limited to N-hydroxylsuccinimide (NHS) active ester, succinimidyl propionate (SPA), maleimide (MAL), vinyl sulfone (VS), or thiol. A PEG polymer, or other polymer, can be linked to an antibody polypeptide at either a predetermined position, or may be randomly linked to the an antibody polypeptide molecule. It is preferred, however, that the PEG polymer be linked to an antibody polypeptide at a predetermined position. A PEG polymer may be linked to any residue in the an antibody polypeptide, however, it is preferable that the polymer is linked to either a lysine or cyseine, which is either naturally occurring in the antibody polypeptide, or which has been engineered into the antibody polypeptide, for example, by mutagenesis of a naturally occurring residue in the antibody polypeptide to either a cysteine or lysine. PEG-linkage can also be mediated through a peptide linker attached to an antibody polypeptide. That is, the PEG moiety can be attached to a peptide linker fused to an antibody polypeptide, where the linker provides the site, e.g., a free cysteine or lysine, for PEG attachment. As used herein, “linked” can also refer to the association of two or more antibody polypeptides, e.g., dAb monomers, to form a dimer, trimer, tetramer, or other multimer. Antibody polypeptide monomers can be linked to form a multimer by several methods known in the art, including, but not limited to, expression of the antibody polypeptide monomers as a fusion protein, linkage of two or more monomers via a peptide linker between monomers, or by chemically joining monomers after translation, either to each other directly, or through a linker by disulfide bonds, or by linkage to a di-, tri- or multivalent linking moiety (e.g., a multi-arm PEG). While dAb multimers are specifically contemplated herein, e.g., in the context of dual- or multi-specific antibody polypeptide constructs, it is emphasized that for any given antibody polypeptide construct, the construct should only be able to bind one molecule of CD40L, i.e., the constructs can have only one CD40L-binding element, and cannot cross link CD40L.

As used herein, “polymer” refers to a macromolecule made up of repeating monomeric units, and can refer to a synthetic or naturally occurring polymer such as an optionally substituted straight or branched chain polyalkylene, polyalkenylene, or polyoxyalkylene polymer or a branched or unbranched polysaccharide. A “polymer” as used herein, specifically refers to an optionally substituted or branched chain poly(ethylene glycol), poly(propylene glycol), or polyvinyl alcohol) and derivatives thereof.

As used herein, “PEG” or “PEG polymer” refers to polyethylene glycol, and more specifically can refer to a derivitized form of PEG, including, but not limited to N-hydroxylsuccinimide (NHS) active esters of PEG such as succinimidyl propionate, benzotriazole active esters, PEG derivatized with maleimide, vinyl sulfones, or thiol groups. Particular PEG formulations can include PEG-O—CH2CH2CH2—CO2—NHS; PEG-O—CH2—NHS; PEG-O—CH2CH2—CO2—NHS; PEG-S—CH2CH2—CO—NHS; PEG-O2CNH—CH(R)—CO2—NHS; PEG-NHCO—CH2CH2—CO—NHS; and PEG-O—CH2—Co2—NHS; where R is (CH2)4)NHCO2(mPEG). PEG polymers useful in the invention may be linear molecules, or may be branched wherein multiple PEG moieties are present in a single polymer. Some particularly preferred PEG conformations that are useful in the invention include, but are not limited to the following:

As used herein, a “sulfhydryl-selective reagent” is a reagent which is useful for the attachment of a PEG polymer to a thiol-containing amino acid. Thiol groups on the amino acid residue cysteine are particularly useful for interaction with a sulfhydryl-selective reagent. Sulfhydryl-selective reagents which are useful for such attachment include, but are not limited to maleimide, vinyl sulfone, and thiol. The use of sulfhydryl-selective reagents for coupling to cysteine residues is known in the art and may be adapted as needed according to the present invention (See Eg., Zalipsky, 1995, Bioconjug. Chem. 6:150; Greenwald et al., 2000, Crit. Rev. Ther. Drug Carrier Syst. 17:101; Herman et al., 1994, Macromol. Chem. Phys. 195:203).

The attachment of PEG or another agent, e.g., HSA, to an antibody polypeptide or to a single immunoglobulin variable domain polypeptide as described herein will preferably not impair the ability of the polypeptide to specifically bind CD40L. That is, the PEG-linked antibody polypeptide or single immunoglobulin variable domain polypeptide will retain its binding activity relative to a non-PEG-linked counterpart. As used herein, “retains activity” refers to a level of activity of a PEG-linked antibody polypeptide which is at least 10% of the level of activity of a non-PEG-linked antibody polypeptide, preferably at least 20%, 30%, 40%, 50%, 60%, 70%, 80% and up to 90%, preferably up to 95%, 98%, and up to 100% of the activity of a non-PEG-linked antibody polypeptide comprising the same antigen-binding domain or domains. More specifically, the activity of a PEG-linked antibody polypeptide compared to a non-PEG linked antibody variable domain should be determined on an antibody polypeptide molar basis; that is equivalent numbers of moles of each of the PEG-linked and non-PEG-linked antibody polypeptides should be used in each trial. In determining whether a particular PEG-linked antibody polypeptide “retains activity”, it is preferred that the activity of a PEG-linked antibody polypeptide be compared with the activity of the same antibody polypeptide in the absence of PEG.

As used herein, the term “in vivo half-life” refers to the time taken for the serum concentration of a ligand (e.g., a single immunoglobulin variable domain) to reduce by 50%, in vivo, for example due to degradation of the ligand and/or clearance or sequestration of the ligand by natural mechanisms. The anti CD40L antibody polypeptides or single immunoglobulin variable domain polypeptides described herein can be stabilized in vivo and their half-life increased by binding to molecules, such as PEG, which resist degradation and/or clearance or sequestration. The half-life of an antibody polypeptide is increased if its functional activity persists, in vivo, for a longer period than a similar antibody polypeptide which is not linked to a PEG polymer. Typically, the half life of a PEGylated antibody polypeptide is increased by 10%, 20%, 30%, 40%, 50% or more relative to a non-PEGylated antibody polypeptide. Increases in the range of 2×, 3×, 4×, 5×, 10×, 20×, 30×, 40×, 50× or more of the half life are possible. Alternatively, or in addition, increases in the range of up to 30×, 40×, 50×, 60×, 70×, 80×, 90×, 100×, 150× of the half life are possible. According to the invention, a PEG-linked antibody single variable domain has a half-life of between 0.25 and 170 hours, preferably between 1 and 100 hours, more preferably between 30 and 100 hours, and still more preferably between 50 and 100 hours, and up to 170, 180, 190, and 200 hours or more.

As used herein, “resistant to degradation” or “resists degradation” with respect to a PEG or other polymer-linked antibody polypeptide monomer or multimer means that the PEG- or other polymer-linked antibody polypeptide monomer or multimer is degraded by no more than 10% when exposed to pepsin at pH 2.0 for 30 minutes and preferably not degraded at all.

As used herein, “hydrodynamic size” refers to the apparent size of a molecule (e.g., a protein molecule) based on the diffusion of the molecule through an aqueous solution. The diffusion, or motion of a protein through solution can be processed to derive an apparent size of the protein, where the size is given by the “Stokes radius” or “hydrodynamic radius” of the protein particle. The “hydrodynamic size” of a protein depends on both mass and shape (conformation), such that two proteins having the same molecular mass may have differing hydrodynamic sizes based on the overall conformation of the protein. Hydrodynamic size is measured, for example, by size exclusion chromatography. The hydrodynamic size of a PEG-linked antibody polypeptide, e.g., a single immunoglobulin variable domain (including antibody variable domain multimers as described herein), can be in the range of 24 kDa to 500 kDa; 30 to 500 kDa; 40 to 500 kDa; 50 to 500 kDa; 100 to 500 kDa; 150 to 500 kDa; 200 to 500 kDa; 250 to 500 kDa; 300 to 500 kDa; 350 to 500 kDa; 400 to 500 kDa and 450 to 500 kDa. Preferably the hydrodynamic size of a PEGylated antibody polypeptide of the invention is 30 to 40 kDa; 70 to 80 kDa or 200 to 300 kDa. Where a single immunoglobulin variable domain polypeptide is desired for use in imaging applications, the polypeptide should have a hydrodynamic size of between 50 and 100 kDa. Alternatively, where a single immunoglobulin variable domain polypeptide is desired for therapeutic applications, the polypeptide preparation should have a hydrodynamic size of greater than 200 kDa.

As used herein, the term “IC50” refers to the concentration of an inhibitor necessary to inhibit a given activity by 50%. IC50 is determined by assaying a given activity, e.g., binding of CD40L to CD40, in the presence of varying amounts of the inhibitor (e.g., monovalent anti-CD40L antibody polypeptide), and plotting the inhibitor concentration versus the activity being targeted. Binding of CD40L to CD40 is measured herein by the method described in Example 6. Alternatively, SPR can be used.

As used herein, the term “fused to an antibody polypeptide” means that a polypeptide is fused to a given antibody through use of recombinant DNA techniques. Thus, an antibody “fused to” another polypeptide, e.g., to another antibody of different binding specificity, does not exist in nature and is generated through recombinant means. The term “fused to an antibody polypeptide” also encompasses the linkage of a polypeptide to a given antibody polypeptide through, for example, disulfide or other chemical linkages, where the fused polypeptide is not naturally found fused to the antibody polypeptide. Recombinant and chemical methods of fusing a polypeptide to another polypeptide, e.g., to an antibody, are well known in the art.

As used herein, the term “Fc domain” refers to the constant region antibody sequences comprising CH2 and CH3 constant domains as delimited according to Kabat et al., supra. The Fc portion of the heavy chain polypeptide has the ability to self-associate, a function which facilitates the formation of divalent antibodies. The term “lacks an Fc domain” means that a given antibody polypeptide lacks at least the portion of an immunoglobulin Fc domain (as such domains are defined according to Kabat et al., supra) sufficient to mediate the dimerization of Fc-containing antibody polypeptides. Dimerization of Fc-containing antibody polypeptides is measured, for example, by chromatographic methods or by surface plasmon resonance. An antibody polypeptide lacking an Fc domain avoids Fc-platelet interactions and therefore avoids induction of platelet aggregation.

As used herein, the term “symptom of systemic lupus erythematosus” refers to any of the clinically relevant symptoms of SLE known to those of skill in the art. Non-limiting examples include the accumulation of IgG autoantibodies (e.g., against nuclear antigens such as chromatin, snRNPs (especially UI, Sm, Ro/SSA and La/SSB), phospholipids and cell surface molecules), hemolytic anemia, thrombocytopenia, leukopenia, glomerulonephritis, vasculitis, arthritis, and serositis). A reduction in such a symptom is a reduction by at least 10% in a clinically measurable parameter, or by at least one point on a clinically-accepted scale of disease severity.

As used herein, the phrase “specifically binds” refers to the binding of an antigen by an immunoglobulin variable domain with a dissociation constant (Kd) of 1 μM or lower as measured by surface plasmon resonance analysis using, for example, a BIAcore™ surface plasmon resonance system and BIAcore™ kinetic evaluation software (e.g., version 2.1). The affinity or Kd for a specific binding interaction is preferably about 500 nM or lower, more preferably about 300 nM or lower.

As used herein, a “generic ligand” is a ligand that binds a substantial proportion of functional members in a given repertoire, e.g., in a phage display library. Thus, the same generic ligand can bind many members of the repertoire regardless of their target ligand specificities. In general, the presence of a functional generic ligand binding site indicates that the repertoire member is expressed and folded correctly. Thus, binding of the generic ligand to its binding site provides a method for preselecting functional polypeptides from a repertoire of polypeptides. Generic ligands include, for example, Protein A, Protein G and Protein L.

As used herein, the term “universal framework” refers to a single antibody framework sequence corresponding to the regions of an antibody conserved in sequence as defined by Rabat (supra) or corresponding to the human germline immunoglobulin repertoire or structure as defined by Chothia and Lesk, (1987) J. Mol. Biol. 196:910-917. The invention provides for the use of a single framework, or a set of such frameworks, which has been found to permit the derivation of virtually any binding specificity though variation in the hypervariable regions alone.

FIG. 1 shows gel analysis of the quality of biotin-labeled CD40L used in the screening procedures described herein. (a) 1 μg of non-biotinylated-CD40L (Lane 1) and 0.3 μg of biotin-CD40L (Lane 2) were analysed on SDS-PAGE and detected by Simply Blue Safe-Stain. (b) 0.1 μg of biotin-CD40L (Lane 1) and 0.02 μg of biotin-CD40L (Lane 2) were detected by Western-blot probing with 1:5000 Streptavidin-HRP.

FIG. 2 shows a graphical representation of a dose response receptor binding assay (RBA) readout, analysing the inhibition of CD40L binding to CD40-Fc by dAbs DOM-10, -20, -27, -30, -31, -62, -77, titrated from 1 μM down to 10 pM. dAbs DOM-20, -30, and -31 are the most potent with IC50 values of approximately 8 nM.

FIG. 3 shows a graphical representation of a dose response receptor binding assay readout, analysing the inhibition of CD40L binding to CD40-Fc by dAbs DOM-4 and DOM-5, titrated from 1 μM down to 500 pM. The IC50 values for dAbs DOM-5 and DOM-4 are approximately 3 nM and 100 nM respectively.

FIG. 4 shows a graphical representation of a dose response receptor binding assay readout, analysing the inhibition of CD40L binding to CD40-Fc by dAb DOM-24, titrated from 100 nM down to 0.5 pM. The data were curve-fitted using GraphPad Prism software.

FIG. 5 shows the sequence of the VH framework based on germline sequence DP47-JH4b (SEQ ID NO: 1, amino acid sequence; SEQ ID NO: 2, nucleotide sequence—both sense and antisense strands are shown—SEQ ID NO: 2 is the top strand). HCDRs 1-3 are indicated by underlining.

FIG. 6 shows the sequence of the Vκ framework based on germline sequence DPK9 JK1 (SEQ ID NO: 3, amino acid sequence; SEQ ID NO: 4, nucleotide sequence—both sense and antisense strands are shown—SEQ ID NO: 4 is the top strand). LCDRs 1-3 are indicated by underlining.

FIG. 7 shows a schematic representation of the CD40L binding assay used herein, e.g., in Example 6.

FIG. 8 shows various GAS1 secretion signal peptide coding sequences.

FIG. 9 is a schematic showing the CD40L RBA. The soluble proteins obtained from R&D Systems are CD40/Fc=homodimer and CD40L=homotrimer).

FIG. 10 shows the results of a receptor binding assay which demonstrates the affinity of PEGylated DOME-24cys with either 30K PEG MAL or 40K PEG2-MAL.

FIG. 11 shows the results of an assay to assess the simultaneous binding of a dual specific dimer to HSA and CD40L (shaded bars). Binding to control BSA antigen is also shown (solid bars).

FIG. 12 shows the results of an assay to assess the simultaneous binding of a dual specific Fab to HSA and CD40L (shaded bars). Binding to control skimmed mild powder antigen is also shown (solid bars).

FIG. 13 shows the results of FACS analysis of the inhibitory effect of monomeric DOM-24 (grey dotted line). Control stimulated cells as shown as the solid black line and a control dAb is shown as the grey solid line.

FIG. 14 shows the results of FACS analysis of the inhibitory effect of the Vk dAb DOM-116 (dotted line). Control stimulated cells are shown as the solid black line and a control dAb is shown as the grey solid line.

DETAILED DESCRIPTION

The invention provides antibody polypeptides that are monovalent for binding to CD40L. Monovalency for CD40L binding removes the possibility for cross-linking that occurs with prior art antibodies, and which plays a role in undesirable side effects observed with anti-CD40L monoclonal antibodies. Further, while not wishing to be limited to any specific mechanism or theory, because antibody polypeptides monovalent for CD40L cannot cross link CD40L, the possibility is eliminated that cross-linked CD40L may in turn cross-link cell surface CD40 and result in agonism of CD40 signaling activity. Thus, in a preferred aspect, the anti-CD40L antibodies disclosed herein not only inhibit or antagonize the binding of CD40L to CD40, they do not substantially agonize CD40 activity.

In one aspect, the antibodies monovalent for CD40L binding are human antibody polypeptides. Human antibody polypeptides can be administered to human patients while largely avoiding the anti-antibody immune response often provoked by the administration of antibodies from other species, e.g., mouse. While murine antibodies can be “humanized” by grafting human constant domains onto the murine antigen-binding domains, human antibodies as disclosed herein are produced without the need for laborious and time-consuming genetic manipulation of a murine antibody sequence.

The heavy and light polypeptide chains of antibodies comprise variable (V) regions that directly participate in antigen interactions, and constant (C) regions that provide structural support and function in non-antigen-specific interactions with immune effectors. The antigen binding domain of a conventional antibody is comprised of two separate domains: a heavy chain variable domain (VH) and a light chain variable domain (VL: which can be either Vκ or Vλ. The antigen binding site itself is formed by six polypeptide loops: three from the VH domain (H1, H2 and H3) and three from the VL domain (L1, L2 and L3). In vivo, a diverse primary repertoire of V genes that encode the VH and VL domains is produced by the combinatorial rearrangement of gene segments. C regions include the light chain C regions (referred to as CL regions) and the heavy chain C regions (referred to as CH1, CH2 and CH3 regions). A naturally-occurring antibody generally comprises two antigen binding domains and is therefore divalent.

A number of smaller antigen binding fragments of naturally occurring antibodies have been identified following protease digestion. These include, for example, the “Fab fragment” (VL-CL-Cu1-VH), “Fab′ fragment” (a Fab with the heavy chain hinge region), and “F(ab′)2 fragment” (a dimer of Fab′ fragments joined by the heavy chain hinge region). Recombinant methods have been used to generate such fragments and to generate even smaller antigen-binding fragments, e.g., those referred to as “single chain Fv” (variable fragment) or “scFv,” consisting of VL and VH joined by a synthetic peptide linker (VL-linker-VH). Fab fragments, Fab′ fragments and scFv fragments are monovalent for antigen binding, as they each comprise only one antigen binding domain comprising one VHNL dirtier. Even smaller monovalent antibody fragments are the “domain antibodies,” or “dAbs,” which comprise only a single immunoglobulin variable domain, e.g., VH or VL, that alone specifically binds antigen, i.e., without the need for a complementary VL or VH domain, respectively.

The term “dAb” will refer herein to a single immunoglobulin variable domain (VH or VL) polypeptide that specifically binds antigen. A “dAb” binds antigen independently of other V domains; however, a “dAb” can be present in a homo- or heteromultimer with other VH or VL domains where the other domains are not required for antigen binding by the dAb, i.e., where the dAb binds antigen independently of the additional VH or VL domains. The preparation of single immunoglobulin variable domains is described and exemplified herein below.

Monovalent antibody polypeptides can be generated in several different ways. For example, the nucleic acid sequence encoding heavy and light chains of an antibody known to bind. CD40L can be manipulated to generate a number of different antibody polypeptides that are monovalent for CD40L binding. Thus, given the sequences encoding the heavy and light chain polypeptides that constitute an antibody and standard molecular cloning methodologies, one can generate monovalent antigen-binding polypeptide constructs such as Fab fragments, scFv, dAbs, or even bispecific antibodies (i.e., antibodies that comprise two different antigen-binding moieties and can therefore bind two separate antigens, preferably simultaneously) that are monovalent for CD40L.

Thus, one means of generating monovalent antibody polypeptides specific for CD40L is is to amplify and express the VH and VL regions of the heavy chain and light chain gene sequences isolated, for example, from a hybridoma (e.g., a mouse hybridoma) that expresses anti-CD40L monoclonal antibody. The boundaries of VH and VL domains are set out by Kabat et al. (1991, supra). The information regarding the boundaries of the VH and VL domains of heavy and light chain genes is used to design PCR primers that amplify the V domain from a heavy or light chain coding sequence encoding an antibody known to bind CD40L. The amplified V domains are inserted into a suitable expression vector, e.g., pHEN-1 (Hoogenboom et al., 1991, Nucleic Acids Res. 19: 4133-4137) and expressed, e.g., as a fusion of the VH and VL in an scFv or other suitable monovalent format. The resulting polypeptide is then screened for high affinity monovalent binding to CD40L. For all aspects of the present invention, screening for binding is performed as known in the art or as described herein below.

Alternatively, library screening methods can be used to identify monovalent CD40L-specific binding proteins. Phage display technology (see, e.g., Smith, 1985, Science 228: 1315; Scott & Smith, 1990, Science 249: 386; McCafferty et al., 1990, Nature 348: 552) provides an approach for the selection of antibody polypeptides which bind a desired target from among large, diverse repertoires of antibody polypeptides. These phage-antibody libraries can be grouped into two categories: natural libraries which use rearranged V genes harvested from human B cells (Marks et al., 1991, J. Mol. Biol., 222: 581; Vaughan et al., 1996, Nature Biotech., 14: 309) or synthetic libraries whereby germline V gene segments or other antibody polypeptide coding sequences are ‘rearranged’ in vitro (Hoogenboom & Winter, 1992, J. Mol. Biol., 227: 381; Nissim et al., 1994, EMBO J., 13: 692; Griffiths et al., 1994, EMBO J., 13: 3245; De Kruif et al., 1995, J. Mol. Biol., 248: 97) or where synthetic CDRs are incorporated into a single rearranged V gene (Barbas et al., 1992. Proc. Natl. Acad. Sci. USA, 89: 4457). Methods involving genetic display packages (e.g., phage display, polysome display) are well-suited for the selection of monovalent CD40L-specific antibody constructs because they generally express only monovalent fragments, rather than whole, divalent antibodies, on the display packages. Methods for the preparation of phage display libraries displaying various antibody fragments are described in the preceding references. Such methods are also described, for example, in U.S. Pat. No. 6,696,245, which is incorporated herein by reference. The methods described in the \'245 patent generally involve the randomization of selected regions of immunoglobulin gene coding regions, in particular VH and VL coding regions, while leaving other regions non-randomized (see below). The \'245 patent also describes the generation of scFv constructs comprising individually randomized VH and VL domains.

The VH gene is produced by the recombination of three gene segments, VH, D and JH. In humans, there are approximately 51 functional VH segments (Cook and Tomlinson (1995) Immunol Today 16: 237), 25 functional D segments (Corbett et al. (1997) J. Mol. Biol. 268: 69)- and 6 functional JH segments (Ravetch et al. (1981) Cell 27: 583), depending on the haplotype. The VH segment encodes the region of the polypeptide chain which forms the first and second antigen binding loops of the VH domain (H1 and H2), while the VH, D and JH segments combine to form the third antigen binding loop of the VH domain (H3).

The VL, gene is produced by the recombination of only two gene segments, VL and JL. In humans, there are approximately 40 functional Vκ segments (Schäble and Zachau (1993) Biol. Chem. Hoppe-Seyler 374: 1001), 31 functional Vλ, segments (Williams et al. (1996) J. Mol. Biol. 264: 220; Kawasaki et al. (1997) Genome Res. 7: 250), 5 functional Jκ segments (Hieter et al. (1982) J. Biol. Chem. 257: 1516) and 4 functional 5 segments (Vasicek and Leder (1990) J. Exp. Med. 172: 609), depending on the haplotype. The VL segment encodes the region of the polypeptide chain which forms the first and second antigen binding loops of the VL domain (L1 and L2), while the VL and JL, segments combine to form the third antigen binding loop of the VI, domain (L3). Antibodies selected from this primary repertoire are believed to be sufficiently diverse to bind almost all antigens with at least moderate affinity. High affinity antibodies are produced in vivo by “affinity maturation” of the rearranged genes, in which point mutations are generated and selected by the immune system on the basis of improved binding.

Analysis of the structures and sequences of antibodies has shown that five of the six antigen binding loops (H1, H2, L1, L2, L3) possess a limited number of main-chain conformations or canonical structures (Chothia and Lesk (1987) J. Mol. Biol. 196: 901; Chothia et al. (1989) Nature 342: 877). The main-chain conformations are determined by (i) the length of the antigen binding loop, and (ii) particular residues, or types of residue, at certain key positions in the antigen binding loop and the antibody framework. Analysis of the loop lengths and key residues has enabled the prediction of the main-chain conformations of H1, H2, L1, L2 and L3 encoded by the majority of human antibody sequences (Chothia et al. (1992) J. Mol. Biol. 227: 799; Tomlinson et al. (1995) EMBO J. 14: 4628; Williams et al. (1996) J. Mol. Biol. 264: 220). Although the H3 region is much more diverse in terms of sequence, length and structure (due to the use of D segments), it also forms a limited number of main-chain conformations for short loop lengths which depend on the length and the presence of particular residues, or types of residue, at key positions in the loop and the antibody framework (Martin et al. (1996) J. Mol. Biol. 263: 800; Shirai et al. (1996) FEBS Letters 399: 1.

While, in one approach, diversity can be added to synthetic repertoires at any site in the CDRs of the various antigen-binding loops, this approach results in a greater proportion of V domains that do not properly fold and therefore contribute to a lower proportion of molecules with the potential to bind antigen. An understanding of the residues contributing to the main chain conformation of the antigen-binding loops permits the identification of specific residues to diversify in a synthetic repertoire of VH or VL domains. That is, diversity is best introduced in residues that are not essential to maintaining the main chain conformation. As an example, for the diversification of loop L2, the conventional approach would be to diversify all the residues in the corresponding CDR (CDR2) as defined by Kabat et al. (1991, supra), some seven residues. However, for L2, it is known that positions 50 and 53 are diverse in naturally occurring antibodies and are observed to make contact with the antigen. The preferred approach would be to diversify only those two residues in this loop. This represents a significant improvement in terms of the functional diversity required to create a range of antigen binding specificities.

Immunoglobulin polypeptide libraries can advantageously be designed to be based on predetermined variable domain main chain conformation. Such libraries may be constructed as described in International Patent Application WO 99/20749, the contents of which are incorporated herein by reference. Thus, in one aspect, an antibody polypeptide comprises the amino acid sequence of a given human germline V region gene segment, e.g., VH germline gene segment DP-47, or Vκ germline gene segment DPK9. Such variable region polypeptides can be used for the production of scFvs or Fabs, e.g., an scFv or Fab comprising (i) an antibody heavy chain variable domain (VH), or antigen binding fragment thereof, which comprises the amino acid sequence of germline VH segment DP-47 and (ii) an antibody light chain variable domain (VL), or antigen binding fragment thereof, which comprises the amino acid sequence of germline Vκ segment DPK9. Diversification of sequences within the context of the selected heavy and light chain germline gene segments, e.g., DP-47, DPK 9, DP45, DP38, etc. can generate a repertoire of diverse immunoglobulin coding sequences. One approach to diversification is described below in the context of generating a library of diversified dAb or scFv sequences. These variable region polypeptides can also be expressed as dAbs and screened for high affinity binding to CD40L. The repertoire can be cloned into or generated in a vector suitable for phage display, e.g., a lambda or filamentous bacteriophage display vector and is then screened for binding to a given target antigen, e.g., CD40L.

A single immunoglobulin variable domain is a folded polypeptide domain which comprises sequences characteristic of immunoglobulin variable domains and which specifically binds an antigen (e.g., dissociation constant of 500 nM or less), and which binds antigen as a single variable domain; that is, there is one binding site provided by a single immunoglobulin variable domain without any complementary variable domain. A single immunoglobulin variable domain therefore includes complete antibody variable domains as well as modified variable domains, for example in which one or more loops have been replaced by sequences which are not characteristic of antibody variable domains or antibody variable domains which have been truncated or comprise N- or C-terminal extensions, as well as folded fragments of variable domains which retain a dissociation constant of 500 nM or less (e.g., 450 nM or less, 400 nM or less, 350 nM or less, 300 nM or less, 250 nM or less, 200 nM or less, 150 nM or less, 100 nM or less) and the target antigen specificity of the full-length domain. Preferably an antibody single variable domain useful in the invention is selected from the group of VH and including Vkapps and Vlambda. The single immunoglobulin variable domains of use herein are preferably “human” as that term is defined herein.

Single immunoglobulin variable domains are prepared in a number of ways. For each of these approaches, well-known methods of preparing (e.g., amplifying, mutating, etc.) and manipulating nucleic acid sequences are applicable.

One means of preparing single immunoglobulin variable domains is to amplify and express the VH or VL region of a heavy chain or light chain gene for a cloned antibody known to bind the desired antigen. That is, the VH or VL domain of a known anti-CD40L antibody coding region can be amplified and expressed as a single domain (or as a fusion of a single domain) and evaluated for binding to CD40L. The boundaries of VH and VL domains are set out by Kabat et al. (1991, supra). The information regarding the boundaries of the VH and VL domains of heavy and light chain genes is used to design PCR primers that amplify the V domain from a cloned heavy or light chain coding sequence encoding an antibody known to bind CD40L. The amplified V domain is inserted into a suitable expression vector, e.g., pHEN-1 (Hoogenboom et al., 1991, Nucleic Acids Res. 19: 4133-4137) and expressed, either alone or as a fusion with another polypeptide sequence.

In a preferred approach, a repertoire of VH or VL domains, preferably human VH or VL domains, is screened by, for example, phage display, panning against the desired antigen. Methods for the construction of bacteriophage display libraries and lambda phage expression libraries are well known in the art, and taught, for example, by: McCafferty et al., 1990, Nature 348: 552; Kang et al., 1991, Proc. Natl. Acad. Sci. U.S.A., 88: 4363; Clackson et al., 1991, Nature 352: 624; Lowman et al., 1991, Biochemistry 30: 10832; Burton et al., 1991, Proc. Natl. Acad. Sci. U.S.A. 88: 10134; Hoogenboom et al., 1991, Nucleic Acids Res. 19: 4133; Chang et al., 1991, J. Immunol. 147: 3610; Breitling et al., 1991, Gene 104: 147; Marks et al., 1991, J. Mol. Biol. 222: 581; Barbas et al., 1992, Proc. Natl. Acad. Sci. U.S.A. 89: 4457; Hawkins and Winter (1992) J. Immunol., 22: 867; Marks et al. (1992) J. Biol. Chem., 267: 16007; and Lerner et al. (1992) Science, 258: 1313. Fab phage display libraries are taught, for example, by U.S. Pat. No. 5,922,545. scFv phage libraries are taught, for example, by Huston et al., 1988, Proc. Natl. Acad. Sci. U.S.A. 85: 5879-5883; Chaudhary et al., 1990, Proc. Natl. Acad. Sci. U.S.A. 87: 1066-1070; McCafferty et al., 1990, supra; Clackson et al., 1991, supra; Marks et al., 1991, supra; Chiswell et al., 1992, Trends Biotech. 10: 80; and Marks et al., 1992, supra. Various embodiments of scFv libraries displayed on bacteriophage coat proteins have been described. Refinements of phage display approaches are also known, for example as described in WO96/06213 and WO92/01047 (Medical Research Council et al.) and WO97/08320 (Morphosys, supra).

The repertoire of VH or V1 domains can be a naturally-occurring repertoire of immunoglobulin sequences or a synthetic repertoire. A naturally-occurring repertoire is one prepared, for example, from immunoglobulin-expressing cells harvested from one or more individuals. Such repertoires can be “naïve,” i.e., prepared, for example, from human fetal or newborn immunoglobulin-expressing cells, or rearranged, i.e., prepared from, for example, adult human B cells. Natural repertoires are described, for example, by Marks et al., 1991, J. Mol. Biol. 222: 581 and Vaughan et al., 1996, Nature Biotech. 14: 309. If desired, clones identified from a natural repertoire, or any repertoire, for that matter, that bind the target antigen are then subjected to mutagenesis and further screening in order to produce and select variants with improved binding characteristics.

Synthetic repertoires of single immunoglobulin variable domains are prepared by artificially introducing diversity into a cloned V domain. Synthetic repertoires are described, for example, by Hoogenboom & Winter, 1992, J. Mol. Biol. 227: 381; Barbas et al., 1992, Proc. Natl. Acad. Sci. U.S.A. 89: 4457; Nissim et al, 1994, EMBO J. 13: 692; Griffiths et al., 1994, EMBO J. 13: 3245; DeKriuf et al., 1995, 3. Mol. Biol. 248:97; and WO 99/20749.

In one aspect, synthetic variable domain repertoires are prepared in VH or Vκ backgrounds, based on artificially diversified germline VH or Vκ sequences. For example, the VH domain repertoire can be based on cloned germline VH gene segments V3-23/DP47 (Tomlinson et al., 1992, J. Mol. Biol. 227: 7768) and JH4b. The VK domain repertoire can be based, for example, on germline Vκ gene segments O2/O12/DPK9 (Cox et al., 1994, Eur. 1. Immunol. 24: 827) and Jκ1. Diversity is introduced into these or other gene segments by, for example, PCR mutagenesis. Diversity can be randomly introduced, for example, by error prone PCR (Hawkins, et al., 1992, J. Mol. Biol. 226: 889) or chemical mutagenesis. As discussed above, however it is preferred that the introduction of diversity is targeted to particular residues. It is further preferred that the desired residues are targeted by introduction of the codon NNK using mutagenic primers (using the IUPAC nomenclature, where N=G, A, T or C, and K=G or T), which encodes all amino acids and the TAG stop codon. Other codons which achieve similar ends are also of use, including the NNN codon (which leads to the production of the additional stop codons TGA and TAA), DVT codon ((A/G/T) (A/G/C)T), DVC codon ((A/G/T)(A/G/C)C), and DVY codon ((A/G/T)(A/G/C)(C/T). The DVT codon encodes 22% serine and 11% tyrosine, asgpargine, glycine, alanine, aspartate, threonine and cysteine, which most closely mimics the distribution of amino acid residues for the antigen binding sites of natural human antibodies. Repertoires are made using PCR primers having the selected degenerate codon or codons at each site to be diversified. PCR mutagenesis is well known in the art.

In one aspect, diversity is introduced into the sequence of human germline VH gene segments V3-23/DP47 (Tomlinson et al., 1992, J. Mol. Biol. 227: 7768) and JH4b using the NNK codon at sites 1130, H31,1133, H35,1150, H52, 1152a, 1153, H55, H56, H58, 1195, 1197 and 1198, corresponding to diversity in CDRs 1, 2 and 3, with the numbering as used in U.S. Pat. No. 6,696,245.

In another aspect, diversity is also introduced into the sequence of human germline VH gene segments V3-23/DP47 and JH4b, for example, using the NNK codon at sites 1130, H31, H33, H35, H50, H52, H52a, 1-153, H55, H56, H58, H95, H97, H98, H99, H100, H100a and H100b, corresponding to diversity in CDRs 1, 2 and 3,with the numbering as used in U.S. Pat. No. 6,696,245.

In another aspect, diversity is introduced into the sequence of human germline Vκ gene segments O2/O12/DPK9 and Jκ1, for example, using the NNK codon at sites L30, L31, L32, L34, L50, L53, L91, L92, L93, L94 and L96, corresponding to diversity in CDRs 1, 2 and 3,with the numbering as used in U.S. Pat. No. 6,696,245.

Diversified repertoires are cloned into phage display vectors as known in the art and as described, for example, in WO 99/20749. In general, the nucleic acid molecules and vector constructs required for the performance of the present invention are available in the art and are constructed and manipulated as set forth in standard laboratory manuals, such as Sambrook et al. (1989). Molecular Cloning: A Laboratory Manual, Cold Spring Harbor, USA.

The manipulation of nucleic acids in the present invention is typically carried out in recombinant vectors. As used herein, “vector” refers to a discrete element that is used to introduce heterologous DNA into cells for the expression and/or replication thereof. Methods by which to select or construct and, subsequently, use such vectors are well known to one of skill in the art. Numerous vectors are publicly available, including bacterial plasmids, bacteriophage, artificial chromosomes and episomal vectors. Such vectors may be used for simple cloning and mutagenesis; alternatively, as is typical of vectors in which repertoire (or pre-repertoire) members of the invention are carried, a gene expression vector is employed. A vector of use according to the invention is selected to accommodate a polypeptide coding sequence of a desired size, typically from 0.25 kilobase (kb) to 40 kb in length. A suitable host cell is transformed with the vector after in vitro cloning manipulations. Each vector contains various functional components, which generally include a cloning (or “polylinker”) site, an origin of replication and at least one selectable marker gene. If a given vector is an expression vector, it additionally possesses one or more of the following: enhancer element, promoter, transcription termination and signal sequences, each positioned in the vicinity of the cloning site, such that they are operatively linked to the gene encoding a polypeptide repertoire member according to the invention.

Both cloning and expression vectors generally contain nucleic acid sequences that enable the vector to replicate in one or more selected host cells. Typically in cloning vectors, this sequence is one that enables the vector to replicate independently of the host chromosomal DNA and includes origins of replication or autonomously replicating sequences. Such sequences are well known for a variety of bacteria, yeast and viruses. The origin of replication from the plasmid pBR322 is suitable for most Gram-negative bacteria, the 2 micron plasmid origin is suitable for yeast, and various viral origins (e.g. SV 40, adenovirus) are useful for cloning vectors in mammalian cells. Generally, the origin of replication is not needed for mammalian expression vectors unless these are used in mammalian cells able to replicate high levels of DNA, such as COS cells.

Advantageously, a cloning or expression vector also contains a selection gene also referred to as selectable marker. This gene encodes a protein necessary for the survival or growth of transformed host cells grown in a selective culture medium. Host cells not transformed with the vector containing the selection gene will therefore not survive in the culture medium. Typical selection genes encode proteins that confer resistance to antibiotics and other toxins, e.g. ampicillin, neomycin, methotrexate or tetracycline, complement auxotrophic deficiencies, or supply critical nutrients not available in the growth media.

Because the replication of vectors according to the present invention is most conveniently performed in E. coli, an E. coli-selectable marker, for example, the β-lactamase gene that confers resistance to the antibiotic ampicillin, is of use. These can be obtained from E. coli plasmids, such as pBR322 or a pUC plasmid such as pUC18 or pUC19.

Expression vectors usually contain a promoter that is recognized by the host organism and is operably linked to the coding sequence of interest. Such a promoter may be inducible or constitutive. The term “operably linked” refers to a juxtaposition wherein the components described are in a relationship permitting them to function in their intended manner. A control sequence “operably linked” to a coding sequence is ligated in such a way that expression of the coding sequence is achieved under conditions compatible with the control sequences.

Promoters suitable for use with prokaryotic hosts include, for example, the β-lactamase and lactose promoter systems, alkaline phosphatase, the tryptophan (trp) promoter system and hybrid promoters such as the tac promoter. Promoters for use in bacterial systems will also generally contain a Shine-Dalgarno sequence operably linked to the coding sequence.

In libraries or repertoires as described herein, the preferred vectors are expression vectors that enable the expression of a nucleotide sequence corresponding to a polypeptide library member. Thus, selection is performed by separate propagation and expression of a single clone expressing the polypeptide library member or by use of any selection display system. As described above, a preferred selection display system uses bacteriophage display. Thus, phage or phagemid vectors can be used. Preferred vectors are phagemid vectors, which have an E. coli origin of replication (for double stranded replication) and also a phage origin of replication (for production of single-stranded DNA). The manipulation and expression of such vectors is well known in the art (Hoogenboom and Winter (1992) supra; Nissim et al. (1994) supra). Briefly, the vector contains a β-lactamase or other selectable marker gene to confer selectivity on the phagemid, and a lac promoter upstream of a expression cassette that consists (N to C terminal) of a pelB leader sequence (which directs the expressed polypeptide to the periplasmic space), a multiple cloning site (for cloning the nucleotide version of the library member), optionally, one or more peptide tags (for detection), optionally, one or more TAG stop codons and the phage protein pill. In one embodiment, the vector encodes, rather than the pelB leader sequence, a eukaryotic GAS1 leader sequence which serves to direct the secretion of the fusion polypeptide to the periplasmic space in E. coli or to the medium in eukaryotic cell systems. Using various suppressor and non-suppressor strains of E. coli and with the addition of glucose, iso-propyl thio-β-D-galactoside (IPTG) or a helper phage, such as VCS M13, the vector is able to replicate as a plasmid with no expression, produce large quantities of the polypeptide library member only, or produce phage, some of which contain at least one copy of the polypeptide-pIII fusion on their surface.

An example of a preferred vector is the pIII phagemid vector (Hoogenboom et al., 1991, Nucl. Acids Res. 19: 4133-4137; sequence is available, e.g., as SEQ ID NO: 7 in WO 03/031611), in which the production of pill fusion protein is under the control of the LacZ promoter, which is inhibited in the presence of glucose and induced with IPTG. When grown in suppressor strains of E. coli, e.g., TG1, the gene III fusion protein is produced and packaged into phage, while growth in non-suppressor strains, e.g., HB2151, permits the secretion of soluble fusion protein into the bacterial periplasm and into the culture medium. Because the expression of gene III prevents later infection with helper phage, the bacteria harboring the phagemid vectors are propagated in the presence of glucose before infection with VCSM13 helper phage for phage rescue.

Construction of vectors according to the invention employs conventional ligation techniques. Isolated vectors or DNA fragments are cleaved, tailored, and re-ligated in the form desired to generate the required vector. If desired, sequence analysis to confirm that the correct sequences are present in the constructed vector is performed using standard methods. Suitable methods for constructing expression vectors, preparing in vitro transcripts, introducing DNA into host cells, and performing analyses for assessing expression and function are known to those skilled in the art. The presence of a gene sequence in a sample is detected, or its amplification and/or expression quantified by conventional methods, such as Southern or Northern analysis, Western blotting, dot blotting of DNA, RNA or protein, in situ hybridization, immunocytochemistry or sequence analysis of nucleic acid or protein molecules. Those skilled in the art will readily envisage how these methods may be modified, if desired.

Screening Single Immunoglobulin Variable Domains for Antigen Binding:

Following expression of a repertoire of single immunoglobulin variable domains on the surface of phage, selection is performed by contacting the phage repertoire with immobilized target antigen, washing to remove unbound phage, and propagation of the bound phage, the whole process frequently referred to as “panning.” This process is applicable to the screening of single immunoglobulin variable domains as well as other antibody fragments that can be expressed on a display library, e.g., scFv, Fab, etc. Alternatively, phage are pre-selected for the expression of properly folded member variants by panning against an immobilized generic ligand (e.g., protein A or protein L) that is only bound by folded members. This has the advantage of reducing the proportion of non-functional members, thereby increasing the proportion of members likely to bind a target antigen. Pre-selection with generic ligands is taught in WO 99/20749. The screening of phage antibody libraries is generally described, for example, by Harrison et al., 1996, Meth. Enzymol. 267: 83-109.

Screening is commonly performed using purified antigen immobilized on a solid support, for example, plastic tubes or wells, or on a chromatography matrix, for example Sepharose™ (Pharmacia). Screening or selection can also be performed on complex antigens, such as the surface of cells (Marks et al., 1993, BioTechnology 11: 1145; de Kruif et al., 1995, Proc. Natl. Acad. Sci. U.S.A. 92: 3938). Another alternative involves selection by binding biotinylated antigen in solution, followed by capture on streptavidin-coated beads.

In a preferred aspect, panning is performed by immobilizing antigen (generic or specific) on tubes or wells in a plate, e.g., Nunc MAXISORP™ immunotube 8 well strips. Wells are coated with 150 μl of antigen (100 μg/ml in PBS) and incubated overnight. The wells are then washed 3 times with PBS and blocked with 400 μl PBS-2% skim milk (2% MPBS) at 37° C. for 2 hr. The wells are rinsed 3 times with PBS and phage are added in 2% MPBS. The mixture is incubated at room temperature for 90 minutes and the liquid, containing unbound phage, is removed. Wells are rinsed 10 times with PBS-0.1% tween 20, and then 10 times with PBS to remove detergent. Bound phage are eluted by adding 200 μl of freshly prepared 100 mM triethylamine, mixing well and incubating for 10 min at room temperature. Eluted phage are transferred to a tube containing 100 μl of 1M Tris-HCl, pH 7.4 and vortexed to neutralize the triethylamine. Exponentially-growing E. coli host cells (e.g., TG1) are infected with, for example, 150 ml of the eluted phage by incubating for 30 min at 37° C. Infected cells are spun down, resuspended in fresh medium and plated in top agarose. Phage plaques are eluted or picked into fresh cultures of host cells to propagate for analysis or for further rounds of selection. One or more rounds of plaque purification are performed if necessary to ensure pure populations of selected phage. Other screening approaches are described by Harrison et al., 1996, supra.

Following identification of phage expressing a single immunoglobulin variable domain that binds a desired target, if a phagemid vector such as pHEN1 has been used, the variable domain fusion proteins are easily produced in soluble form by infecting non-suppressor strains of bacteria, e.g., HB2151 that permit the secretion of soluble gene III fusion protein. If a GAS1 secretion signal peptide is encoded by the vector, the fusion polypeptide can be secreted by eukaryotic (e.g., yeast or mammalian) or prokaryotic (e.g., E. coli) cells. Alternatively, the V domain sequence can be sub-cloned into an appropriate expression vector to produce soluble protein according to methods known in the art.

Purification and Concentration of Single Immunoglobulin Variable Domains:

Single immunoglobulin variable domain polypeptides or other monovalent antibody polypeptides secreted into the periplasmic space or into the medium of bacteria are harvested and purified according to known methods (Harrison et al., 1996, supra). Skerra & Pluckthun (1988, Science 240: 1038) and Breitling et al. (1991, Gene 104: 147) describe the harvest of antibody polypeptides from the periplasm, and Better et al. (1988, Science 240: 1041) describes harvest from the culture supernatant. For some antibody polypeptides, purification can also be achieved by binding to generic ligands, such as protein A or Protein L. Alternatively, the variable domains can be expressed with a peptide tag, e.g., the Myc, HA or 6×-His tags, which facilitates purification by affinity chromatography.

If necessary, monovalent anti-CD40L antibody polypeptides are concentrated by any of several methods well known in the art, including, for example, ultrafiltration, diafiltration and tangential flow filtration. The process of ultrafiltration uses semi-permeable membranes and pressure to separate molecular species on the basis of size and shape. The pressure is provided by gas pressure or by centrifugation. Commercial ultrafiltration products are widely available, e.g., from Millipore (Bedford, Mass.; examples include the Centricon™ and Microcon™ concentrators) and Vivascience (Hannover, Germany; examples include the Vivaspin™ concentrators). By selection of a molecular weight cutoff smaller than the target polypeptide (usually ⅓ to ⅙ the molecular weight of the target polypeptide, although differences of as little as 10 kD can be used successfully), the polypeptide is retained when solvent and smaller solutes pass through the membrane. Thus, a molecular weight cutoff of about 5 kD is useful for concentration of anti-CD40L single immunoglobulin variable domain polypeptides described herein.

Diafiltration, which uses ultrafiltration membranes with a “washing” process, is used where it is desired to remove or exchange the salt or buffer in a polypeptide preparation. The polypeptide is concentrated by the passage of solvent and small solutes through the membrane, and remaining salts or buffer are removed by dilution of the retained polypeptide with a new buffer or salt solution or water, as desired, accompanied by continued ultrafiltration. In continuous diafiltration, new buffer is added at the same rate that filtrate passes through the membrane. A diafiltration volume is the volume of polypeptide solution prior to the start of diafiltration—using continuous diafiltration, greater than 99.5% of a fully permeable solute can be removed by washing through six diafiltration volumes with the new buffer. Alternatively, the process can be performed in a discontinuous manner, wherein the sample is repeatedly diluted and then filtered back to its original volume to remove or exchange salt or buffer and ultimately concentrate the polypeptide. Equipment for diafiltration and detailed methodologies for its use are available, for example, from Pall Life Sciences (Ann Arbor, Mich.) and Sartorius AG/Vivascience (Hannover, Germany).

Tangential flow filtration (TFF), also known as “cross-flow filtration,” also uses ultrafiltration membrane. Fluid containing the target polypeptide is pumped tangentially along the surface of the membrane. The pressure causes a portion of the fluid to pass through the membrane while the target polypeptide is retained above the filter. In contrast to standard ultrafiltration, however, the retained molecules do not accumulate on the surface of the membrane, but are carried along by the tangential flow. The solution that does not pass through the filter (containing the target polypeptide) can be repeatedly circulated across the membrane to achieve the desired degree of concentration. Equipment for TFF and detailed methodologies for its use are available, for example, from Millipore (e.g., the ProFlux M12™ Benchtop TFF system and the Pellicon™ systems), Pall Life Sciences (e.g., the Minim™ Tangential Flow Filtration system).

Protein concentration is measured in a number of ways that are well known in the art. These include, for example, amino acid analysis, absorbance at 280 nm, the “Bradford” and “Lowry” methods, and SDS-PAGE. The most accurate method is total hydrolysis followed by amino acid analysis by HPLC, concentration is then determined then comparison with the known sequence of the single immunoglobulin variable domain polypeptide. While this method is the most accurate, it is expensive and time-consuming. Protein determination by measurement of UV absorbance at 280 nm faster and much less expensive, yet relatively accurate and is preferred as a compromise over amino acid analysis. Absorbance at 280 nm was used to determine protein concentrations reported in the Examples described herein.

“Bradford” and “Lowry” protein assays (Bradford, 1976, Anal. Biochem. 72: 248-254; Lowry et al., 1951, J. Biol. Chem. 193: 265-275) compare sample protein concentration to a standard curve most often based on bovine serum albumin (BSA). These methods are less accurate, tending to undersetimate the concentration of single immunoglobulin variable domains. Their accuracy could be improved, however, by using a VH or Vκ single domain polypeptide as a standard.

An additional protein assay method is the bicinchoninic acid assay described in U.S. Pat. No. 4,839,295 (incorporated herein by reference) and marketed by Pierce Biotechnology (Rockford, Ill.) as the “BCA Protein Assay” (e.g., Pierce Catalog No. 23227).

The SDS-PAGE method uses gel electrophoresis and Coomassie Blue staining in comparison to known concentration standards, e.g., known amounts of a single immunoglobulin variable domain polypeptide. Quantitation can be done by eye or by densitometry.

Single human immunoglobulin variable domain antigen-binding polypeptides described herein retain solubility at high concentration (e.g., at least 4.8 mg (˜400 μM) in aqueous solution (e.g., PBS), and preferably at least 5 mg/ml (˜417 μM), 10 mg/ml (˜833 μM), 20 mg/ml (˜1.7 mM), 25 mg/ml (˜2.1 mM), 30 mg/ml (˜2.5 mM), 35 mg/ml (˜2.9 mM), 40 mg/ml (˜3.3 mM), 45 mg/ml (˜3.75 mM), 50 mg/ml (˜4.2 mM), 55 mg/ml (˜4.6 mM), 60 mg/ml (˜5.0 mM), 65 mg/ml (˜5.4 mM), 70 mg/ml (˜5.8 mM), 75 mg/ml (˜6.3 mM), 100 mg/ml (˜8.33 mM), 150 mg/ml (˜12.5 mM), 200 mg/ml (˜16.7 mM), 240 mg/ml (˜20 mM) or higher). One structural feature that promotes high solubility is the relatively small size of the single immunoglobulin variable domain polypeptides. A full length conventional four chain antibody, e.g., IgG is about 150 kD in size. In contrast, single immunoglobulin variable domains, which all have a general structure comprising 4 framework (FW) regions and 3 CDRs, have a size of approximately 12 kD, or less than 1/10 the size of a conventional antibody. Similarly, single immunoglobulin variable domains are approximately V2 the size of an scFv molecule (−26 kD), and approximately ⅕ the size of a Fab molecule (˜60 kD). It is preferred that the size of a single immunoglobulin variable domain-containing structure disclosed herein is 100 kD or less, including structures of, for example, about 90 kD or less, 80 kD or less, 70 kD or less, 60 kD or less, 50 kD or less, 40 kD or less, 30 kD or less, 20 kD or less, down to and including about 12 kD, or a single immunoglobulin variable domain in isolation.

The solubility of a polypeptide is primarily determined by the interactions of the amino acid side chains with the surrounding solvent. Hydrophobic side chains tend to be localized internally as a polypeptide folds, away from the solvent-interacting surfaces of the polypeptide. Conversely, hydrophilic residues tend to be localized at the solvent-interacting surfaces of a polypeptide. Generally, polypeptides having a primary sequence that permits the molecule to fold to expose more hydrophilic residues to the aqueous environment are more soluble than one that folds to expose fewer hydrophilic residues to the surface. Thus, the arrangement and number of hydrophobic and hydrophilic residues is an important determinant of solubility. Other parameters that determine polypeptide solubility include solvent pH, temperature, and ionic strength. In a common practice, the solubility of polypeptides can be maintained or enhanced by the addition of glycerol (e.g., ˜10% v/v) to the solution.

As discussed above, specific amino acid residues have been identified in conserved residues of human VH domains that vary in the VH domains of camelid species, which are generally more soluble than human VH domains. These include, for example, Gly 44 (Glu in camelids), Leu 45 (Arg in camelids) and Tip 47 (Gly in camelids). Amino acid residue 103 of VH is also implicated in solubility, with mutation from Trp to Arg tending to confer increased VH solubility.

In preferred aspects of the invention, single immunoglobulin variable domain polypeptides are based on the DP47 germline VH gene segment or the DPK9 germline Vκ gene segment. Thus, these germline gene segments are capable, particularly when diversified at selected structural locations described herein, of producing specific binding single immunoglobulin variable domain polypeptides that are highly soluble. In particular, the four framework regions, which are preferably not diversified, can contribute to the high solubility of the resulting proteins.

It is expected that a single human immunoglobulin variable domain that is highly homologous to one having a known high solubility will also tend to be highly soluble. Thus, as one means of prediction or recognition that a given single immunoglobulin variable domain would have the high solubility recited herein, one can compare the sequence of a single immunoglobulin variable domain polypeptide to one or more single immunoglobulin variable domain polypeptides having known solubility. Thus, when a single immunoglobulin variable domain polypeptide is identified that has high binding affinity but unknown solubility, comparison of its amino acid sequence with that of one or more (preferably more) human single immunoglobulin variable domain polypeptides known to have high solubility (e.g., a dAb sequence disclosed herein) can permit prediction of its solubility. While it is not an absolute predictor, where there is a high degree of similarity to a known highly soluble sequence, e.g., 90-95% or greater similarity, and particularly where there is a high degree of similarity with respect to hydrophilic amino acid residues, or residues likely to be exposed at the solvent interface, it is more likely that a newly identified binding polypeptide will have solubility similar to that of the known highly soluble sequence.

Molecular modeling software can also be used to predict the solubility of a polypeptide sequence relative to that of a polypeptide of known solubility. For example, the substitution or addition of a hydrophobic residue at the solvent-exposed surface, relative to a molecule of known solubility that has a less hydrophobic or even hydrophilic residue exposed in that position is expected to decrease the relative solubility of the polypeptide. Similarly, the substitution or addition of a more hydrophilic residue at such a location is expected to increase the relative solubility. That is, a change in the net number of hydrophilic or hydrophobic residues located at the surface of the molecule (or the overall hydrophobic or hydrophilic nature of the surface-exposed residues) relative to a single immunoglobulin variable domain polypeptide structure with known solubility can predict the relative solubility of a single immunoglobulin variable domain polypeptide.

Alternatively, or in conjunction with such prediction, one can determine limits of a single immunoglobulin variable domain polypeptide\'s solubility by simply concentrating the polypeptide.

Affinity Determination:

Isolated single immunoglobulin variable domain- and antibody polypeptide-containing polypeptides as described herein preferably have affinities (dissociation constant, Kd,=Koff/Kon) of at least 500 nM or less, and preferably at least 400 nM-50 pM, 300 nM-50 pM, 200 nM-50 pM, and more preferably at least 100 nM-50 pM, 75 nM 50 pM, 50 nM-50 pM, 25 nM-50 pM, 10 nM 50 pM, 5 nM 50 pM, 1 nM-50 pM, 950 pM 50 pM, 900 pM-50 pM, 850 pM-50 pM, 800 pM-50 pM, 750 pM-50 pM, 700 pM 50 pM, 650 pM 50 pM, 600 pM 50 pM, 550 pM 50 pM, 500 pM-50 pM, 450 pM 50 pM, 400 pM-50 pM, 350 pM-50 pM, 300 pM-50 pM, 250 pM-50 pM, 200 pM-50 pM, 150 pM 50 pM, 100 pM-50 pM, 90 pM 50 pM, 80 pM 50 pM, 70 pM-50 pM, 60 pM-50 pM, or even as low as 50 pM.

The antigen-binding affinity of an antibody polypeptide, e.g., a single immunoglobulin variable domain polypeptide or other monovalent antibody polypeptide, can be conveniently measured by SPR using the BIAcore system (Pharmacia Biosensor, Piscataway, N.J.). In this method, antigen is coupled to the BIAcore chip at known concentrations, and variable domain polypeptides are introduced. Specific binding between the variable domain polypeptide and the immobilized antigen results in increased protein concentration on the chip matrix and a change in the SPR signal. Changes in SPR signal are recorded as resonance units (RU) and displayed with respect to time along the Y axis of a sensorgram. Baseline signal is taken with solvent alone (e.g., PBS) passing over the chip. The net difference between baseline signal and signal after completion of antibody polypeptide injection represents the binding value of a given sample. To determine the off rate (Koff), on rate (Kon) and dissociation rate (Kd) constants, BIAcore kinetic evaluation software (e.g., version 2.1) is used.

Thus, SPR can be used to monitor antagonism of CD40L binding to CD40 by a monovalent anti-CD40L antibody preparation by measuring the displacement or inhibition of binding of CD40L to CD40 caused the monovalent antibody preparation. SPR can also be used to monitor the dimerization, or preferably, the lack of dimerization, occurring via Fc region in antibody preparations as described herein.

High affinity is dependent upon the complementarity between a surface of the antigen and the CDRs of the antibody or antibody fragment. Complementarity is determined by the type and strength of the molecular interactions possible between portions of the target and the CDR, for example, the potential ionic interactions, van der Waals attractions, hydrogen bonding or other interactions that can occur. CDR3 tends to contribute more to antigen binding interactions than CDRs 1 and 2, probably due to its generally larger size, which provides more opportunity for favorable surface interactions. (See, e.g., Padlan et al., 1994, Mol. Immunol. 31: 169-217; Chothia & Lesk, 1987, 7. Mol. Biol. 196: 904-917; and Chothia et al., 1985, J. Mol. Biol. 186: 651-663.) High affinity indicates antibody polypeptide/antigen pairings that have a high degree of complementarity, which is directly related to the structures of the variable domain and the target.

In one aspect, a monovalent anti-CD40L antibody polypeptide, e.g., a single immunoglobulin variable domain polypeptide, is linked to another antibody polypeptide to form a heterodimer in which each individual antibody polypeptide is capable of binding a different cognate antigen. Fusing antibody polypeptides, such as single immunoglobulin variable domains, as heterodimers, wherein each monomer binds a different target antigen, can produce a dual-specific ligand capable, for example, of bridging the respective target antigens. Such dual specific ligands may be used to target cytokines and other molecules which cooperate synergistically in therapeutic situations in the body of an organism. Thus, there is provided a method for synergising the activity of two or more cytokines, comprising administering a dual specific antibody heterodimer capable of binding to the two or more cytokines.

Non-limiting examples of second targets for anti-CD40L dual specific antibody polypeptides include the following: TNF-α; IL-1; IL-2; 1L-4; IL-6; IL-8; IL-12; IL-18; IFN-γ; CD2; CD4; CD8; CTLA4; LFA1; LFA3 and VLA4.

Homologous Sequences:

The invention encompasses anti-CD40L antibody polypeptides, e.g., CD40L-binding single immunoglobulin variable domain clones, and clones with substantial sequence similarity or homology to them that also bind target antigen with high affinity. As used herein, “substantial” sequence similarity or homology is at least 85% similarity or homology.

Calculations of “homology” or “sequence identity” between two sequences (the terms are equivalent and used interchangeably herein) are performed as follows. The sequences are aligned for optimal comparison purposes (e.g., gaps can be introduced in one or both of a first and a second amino acid or nucleic acid sequence for optimal alignment and non-homologous sequences can be disregarded for comparison purposes). In a preferred embodiment, the length of a reference sequence aligned for comparison purposes is at least 30%, preferably at least 40%, more preferably at least 50%, even more preferably at least 60%, and even more preferably at least 70%, 80%, 90%, 100% of the length of the reference sequence. The amino acid residues or nucleotides at corresponding amino acid positions or nucleotide positions are then compared. When a position in the first sequence is occupied by the same amino acid residue or nucleotide as the corresponding position in the second sequence, then the molecules are identical at that position (as used herein amino acid or nucleic acid “homology” is equivalent to amino acid or nucleic acid “identity”). The percent identity between the two sequences is a function of the number of identical positions shared by the sequences, taking into account the number of gaps, and the length of each gap, which need to be introduced for optimal alignment of the two sequences.