KR100494530B1 - Antibodies to Fibronectin ED-B Domains, Methods of Manufacture and Their Uses - Google Patents

Abstract

In the present invention, specific binding elements are provided which are specific for and directly bind to the tumor fetal domain of fibronectin (FN). In addition, the present invention provides a material and a production method for producing the urea.

Description

Antibodies to fibronectin ED-X domains, methods for their preparation, and uses

The present invention relates to an element that specifically binds to ED-B, a fetal isoform of fibronectin, which ED-B is expressed in neovascularization of the forming tumor, which is immunocytochemical and in vivo. Proven by tumor targeting studies. The present invention also relates to a substance related to the production of the specific binding member and a method of producing the same.

The primary goal of current tumor therapies is to kill as many tumor constituent cells as possible. The limitations of success experienced with chemotherapy and radiotherapy are related to the treatment of relatively specific lack of specificity and the tendency of toxic side effects on normal tissues. One method for improving tumor selectivity of therapy is to provide therapeutic agents using binding proteins that typically include a binding domain for antibodies that are specific for marker antigens expressed on the surface of tumor cells but not specific for normal cells. It is a method of transporting to the tumor site. This type of targeting therapy is roughly referred to as 'magic bullets', which are specific for the so-called tumor-associated antigens expressed on the cell surface and are obtained from rodent monoclonal antibodies (mABs). Mainly illustrated. Such mABs may be chemically conjugated to a cytotoxic component (eg, toxin or drug) or produced as a recombinant fusion protein, where the mAB and the gene encoding the toxin are linked together and expressed side by side.

The 'magic gun' approach has significant and limited effects in treating human cancers. The most pronounced effect is when targeting tumors of lymphatic origin, where malignant cells have the most freedom to access therapeutic doses through circulation. However, treatment of solid tumors remains a serious clinical problem because only a small fraction of the total cell mass, mainly the cells at the outermost end of the tumor, is exposed to circulating immunoconjugates; These peripheral targets form a so-called 'binding site barrier' to the interior of the tumor (Juweid et al., 1992, Cancer Res. 52, 5144-5153). Inside solid tumors, tissue structures are generally too dense due to fibroepilosis and dense tumor cells, allowing molecules of the size of antibodies to pass through. Moreover, solid tumors are known to have an elevated interstitial pressure due to the lack of a lymphatic release system, which impedes the influx of foreign molecules. A recent report on the factors affecting the uptake of therapeutic agents into the tumor site can be found in Jane's paper (Jain, R (1994), Sci. Am . 271, 58-65).

Although there are clearly limitations in treating solid tumors by targeting tumor-associated antigens, solid tumors have a common property of providing another antigen target for antibody therapy. Once the tumor has grown to a certain size, it generally relies on its own blood supply for the proper supply of oxygen and nutrients needed for growth. If this blood supply is interrupted or closed, there is a chance that thousands of growing tumor cells will actually starve. As the tumor grows, it is converted into an angiogenic phenotype, producing various antigenic factors that act on endothelial cells of adjacent capillaries and induce their proliferation and migration. The structure of these new blood vessels is not organized so that the distal end is invisible and there are holes in the blood vessels, increasing leakage, which is in sharp contrast with the regular capillaries of normal tissues. Increased expression of some cell surface antigens induces angiogenesis, many of which are common in the blood vessels of normal tissues. The identification of antigens unique to tumor neovascularization has been a limiting factor in developing methods for treating solid tumors through vascular targeting. Antigens, which are the subject of the present invention, solve this problem directly.

During tumor progression, extracellular epilepsy of surrounding tissue is remodeled through two main processes: (1) proteolysis of extracellular epilepsy components in normal tissue (2) by tumor cells and tumor-induced cytokines New Synthesis of Extracellular Epilepsy Components by Activated Interstitial Cells. These two processes produce 'tumor extracellular epilepsy' in steady state, which provides a more suitable environment for tumor progression, which is qualitatively different from normal tissue. Among the matrix components are large isoforms of tenasin and fibronectin (FN); Describing these proteins homogeneously is to recognize their strong structural heterogeneity that occurs at the transcriptional, post-transcriptional and post-translational stages (see section below). The subject of the present invention is one of the isoforms of fibronectin called a B + isoform (B-FN).

Fibronectin (FN) is versatile as a high molecular weight glycoprotein component in extracellular epilepsy and body fluids. They are involved in many different biological processes, such as establishment and maintenance of normal cell morphology, cell migration, homeostasis and thrombosis, wound healing and transformation of carcinogens (Alitalo et al., 1982; Yamada, 1983; Hynes, 1985; Ruoslahti et al. , 1988; Hynes, 1990; Owens et al., 1986). Structural diversity in FN is due to the selective splicing of the three regions of the early FN transcripts (ED-A, ED-B and IIICS) (Hynes, 1985; Zardi et al., 1987), thereby allowing more than 20 different ones. Other isotypes are produced, some of which are expressed differently in tumors and normal tissues. It is known that the splicing pattern of the FN-pre-mRNA is not only regulated by tissue-specific or developmental-specific methods but also in transformed and malignant cells (Castellani et al., 1986; Borsi et al. , 1987; Vartio et al., 1987, Zardi et al., 1987; Barone et al., 1989; Carnemolla et al., 1989; Oyama et al., 1989, 1990; Borsi et al., 1992b). In fact, FN isoforms comprising the ED-A, ED-B and IIICS sequences are expressed to a greater extent in transformed cells and malignant tumor cells than in normal cells. In particular, FN isoforms (B + isoforms) comprising the ED-B sequence have high expression in fetal and tumor tissues as well as during wound healing, but expression in normal tissues of adults is limited (Norton et al., 1987; Schwarzbauer et al., 1987; Schwarzbauer et al., 1987; Gutman and Kornblihtt, 1987; Carnemolla et al., 1989; ffrench-Constant et al., 1989; ffrench-Constant and Hynes, 1989; Laitinen et al., 1991). B + FN molecules are not detected in mature blood vessels but are elevated in normal tissues (eg, endometrial development), lesions (eg diabetic retinopathy), and angiogenic vessels at tumor growth sites (Castellani et al., 1994).

The ED-B sequence is a complete type III-homologous repeat encoded by a single exon and contains 91 amino acids. Unlike alternatively spliced IIICS isoforms, including cell-specific binding sites, the biological function of A + and B + isoforms remains a problem to be considered (Humphries et al., 1986).

The presence of the B + isoform itself constitutes a tumor-induced neoantigen, but in addition the expression of ED-B exposes the normally hidden antigen in type III repeat unit 7 (ED-B described above); Since such epitopes are not exposed in FN molecules lacking ED-B, expression of ED-B directly or indirectly leads to the expression of neoantigens. The concealed antigenic site is targeted by a monoclonal antibody (mAb) called BC-1 (Carnemolla et al., 1992). The specificity and biological activity of the mAb is described in EP 0 344 134 B1, from a hybridoma deposited under accession number 88042101 in the European Collection of Animal Cell Cultures (Salbury, Porton Down). You can get it. The mAb has been successfully used to localize angiogenic vessels of tumors without cross-reactivity to mature vascular endothelium, demonstrating the potential of FN isoforms in vascular targeting with antibodies.

However, some warning remains about the specificity of BC-1 mAb. The fact that BC-1 does not recognize the B + isoform indicates that in some tissues epitopes recognized by BC-1 will be exposed even without ED-B, and thus indirectly induce undesirable cross-reactions of BC-1. It caused a problem. Moreover, BC-1 has strict specificity for human B + isotypes, which means that it is impossible to study the biodistribution and tumor locality of BC-1 in animals. Although polyclonal antibodies against recombinant fusion proteins comprising B + have been produced (Peters et al., 1995), the antibodies respond only to FN treated with N-glycanase that exposes the epitope.

Another common problem with using mouse monoclonal antibodies is the human anti-mouse antibody (HAMA) response (Schroff et al., 1985; Dejager et al., 1988). The HAMA response has a range of effects, from neutralizing the administered antibody to reducing the therapeutic dose to allergic reactions, serum diseases and kidney damage.

Although polyclonal antisera that respond to recombinant ED-B have been identified (see above), it has been demonstrated that it is generally difficult to isolate mAbs with the same specificity as BC-1 after mouse immunity, for humans. This is because the ED-B protein of and mice shows virtually 100% sequence homology. Thus, in mice, human proteins may appear self-antigens and may not elicit an immune response to them. In fact, more than a decade of in-depth investigation in this field confirmed that only a single mAb has indirect reactivity to the B + FN isoform (BC-1) and does not recognize ED-B directly. Most importantly for sure, the specificity of BC-1 is not for parts of ED-B itself, but for concealed epitopes exposed as a result of ED-B, which epitope is unlikely to be present in mouse FNs. Large and therefore will not be recognized as a "self antigen" by the mouse's immune system.

FIG. 1 shows the amino acid sequences aligned for VH and VL of scFv CGS-1 and CGS-2. 1 (a) depicts the VH sequence; Figure 1 (b) shows the VL sequence. CDRs (1, 2 and 3) are shown. The human germline VH most homologous to the two types of scFv is the DP47 segment of the VH3 family; The most homologous VL segment for the two clones is DPL16, which is the light chain used to form the original scFv library (Nissim et al., 1994). Residues that distinguish the two clones are underlined.

2A of FIG. 2 illustrates a human FN subunit domain structure model. The IIICS, ED-A and ED-B variable regions generated due to the selective splicing of the FN pre-mRNA are shown. 2A also shows the internal homology as well as the major digestion products of termolysin including ED-B (Zardi et al., 1987). Figure 2B shows 4-18% SDS-PAGE for plasma and WI38VA FN, termolysin digests stained with coumasyl blue and immune points probed with BC-1, IST-6, CGS-1 and CGS-2 It is showing. Plasma FN digested with termolysin at undigested plasma FN (lane 1) and FN (lane 3) at 1 μg / mg and FN (lane 4) at 10 μg / mg are shown. Degradation with Termolysin at undigested plasma FN (lane 2) and 1 μg / mg FN (lane 5), 5 μg / mg FN (line 6) and 10 μg / mg FN (lane 7) WI38VA FN is shown. The numbers shown to the right represent the turmolysin degradation products shown in FIG. 2A. The figures on the left represent molecular weight criteria expressed in kilodaltons (kD).

3A of FIG. 3 shows the reactivity of the protein against FN type III repeat sequences and CGS-1, CGS-2, mAbs BC-1 and IST-6 contained in fusion and recombinant proteins expressed in E. coli. It is shown. FIG. 3B depicts a gel stained with Coomasyl blue and immune points probed with CGS-1, CGS-2, BC-1, IST-6 shown side by side. The numbers given in the lanes correspond to the numbers of the peptide structures at the top of the figure. The figures on the left represent molecular weight criteria in kD.

4 shows an infrared mouse imaging device; The mouse imaging device used in the targeting experiment consisted of a black non-fluorescent box equipped with a tungsten halogen lamp, an excitation and radiation filter and a computer-controlled 8-bit monochrome CCD-camera with specificity for the CY7 infrared fluorophore. .

FIG. 5 shows targeting fluorescently labeled antibody fragments to murine F9 teratocarcinoma using monomeric scFv (CGS-1) and dimeric scFv (CGS-1) 2 for B-FN. Dimer scFv (D1.3) 2 with binding specificity for lysozyme was used as a negative control.

FIG. 6 shows the targeting of fluorescently labeled antibody fragments to murine F9 teratomas using scFv (CGS-2) and low affinity scFv (28SI) with sufficient affinity for the same epitope of B-FN. . Targeting in large tumors (approximately 0.6 g) and small tumors (approximately 0.2 g) was detected (measured) over 48 hours using a black crust that partially blurred the image.

All documents mentioned herein are incorporated herein by reference.

The present invention has been implemented by a method of treatment which does not require prior immunization with fibronectin or ED-B as an alternative to the previous method; Antibodies specific for the ED-B isoform were obtained as single chain Fv (scFv) from a library of human antibody variable regions on the surface of filamentous bacteriophages (Nissim et al., 1994; WO 92/01047, WO 2/20791, WO 93/06213). , WO 93/11236, WO 93/19172.

The antibody phage library was used to find the following facts; When antigens are coated on a solid surface (“panning”), specific scFvs can be isolated by direct selection on the recombinant FN-fragment comprising the ED-B domain and on the recombinant ED-B itself. These same source antigens have also been successfully used to produce "second generation" scFv with improved properties over parental clones through "affinity maturation" processes. It was found that the isolated scFv reacted strongly and specifically with the B + isoform of human FN without being pretreated with N-glycanase.

In application to anti-tumor therapy, the advantage is that the antigen binding domains of the human antibodies provided herein do not undergo a HAMA response. Furthermore, as exemplified herein, the antibody antigen binding domain of the present invention is also useful for in vivo and in vitro immunohistochemical analysis of tumor tissues. These and other uses will be described in more detail later, and such uses will be apparent to those skilled in the art.

Terms

Specific binding elements

An element of a pair of molecules that specifically binds to each other. The specific binding pair elements may be derived from nature or part or all of them may be synthesized. One element of a molecular pair has one region that specifically binds to and complements a specific space and polar organization of the other element of the molecule pair on its surface or cavity. Therefore, the molecular pair element has a property of specifically binding to each other. Examples of specific binding pair classes are antigen-antibodies, biotin-avidin, hormone-hormone receptors, receptor-ligands, enzyme-substrates.

Antibodies

Immunoglobulins produced naturally or in part or in whole through synthetic methods. The term also includes all proteins or polypeptides having a binding domain, which domain is or is homologous to an antibody binding domain. They may be derived from natural sources or may be produced in part or in whole through synthetic methods. Examples of antibodies include immunoglobulin isotypes and their homologous subclasses; Fragments comprising Fab, scFv, Fv, dAb, Fd; There is a diabody.

Monoclonal and other antibodies can be chosen and recombinant DNA techniques can be used to produce other antibodies or chimeric antibodies that retain the specificity of the initial antibody. Such techniques may include the introduction of DNA encoding an immunoglobulin variable region or complementarity determining regions (CDRs) of an antibody into the constant or constant and framework regions of other immunoglobulins (EP-A). -184187, GB 2188638A or EP-A-239400. Hybridomas or other cells that produce antibodies may be subject to genetic mutations or other modifications, which may or may not alter the binding specificity of the antibodies produced.

Since an antibody can be modified in many ways, the term "antibody" should be interpreted to include all specific binding elements or substances comprising a binding domain with the required specificity. The term therefore includes antibody fragments, derivatives, functional equivalents and homologs to antibodies, including polypeptides comprising naturally occurring or semisynthetic or presynthesized immunoglobulin binding domains. Thus, chimeric molecules comprising immunoglobulin binding domains or their equivalents fused to other polypeptides are included. Cloning and expression of chimeric antibodies are described in EP-A-0120694 and EP-A-0125023.

Whole antibody fragments have been found to be able to perform antigen binding functions. Examples of binding fragments include (i) Fab fragments consisting of VL, VH, CL and CH1 domains; (Ii) an Fd fragment consisting of the VH and CH1 domains; (Iii) a Fv fragment consisting of the VL and VH domains of a single antibody; (Iii) a dAb fragment consisting of the VH domain (Ward et al., 1989); (Iii) an isolated CDR region; (Iii) a F (ab ′) ₂ fragment, which is a bivalentat fragment comprising two associated Fab fragments; (Iii) a single chain Fv molecule (scFv) (Bird et al., 1988; Huston et al., 1988), wherein the VH and VL domains are linked by a peptide linker, thereby combining the two domains to form an antigen binding site. ; (Iii) bispecific single chain Fv dimers (PCT / US92 / 09965) and (iii) “diabodies” (WO 94/13804; Holliger et al., 1993), which are multivalent or multispecific fragments constructed by gene fusion. .

Diabodies are polypeptide multimers, each polypeptide comprising a first domain comprising a binding region of an immunoglobulin light chain and a second domain comprising a binding region of an immunoglobulin heavy chain. The two domains are linked (eg, by a peptide linker) but do not associate with each other to form an antigen binding site; An antigen binding site is formed by the association of a first domain, one polypeptide in a multimer, and a second domain, another polypeptide in a multimer (WO 94/13804).

When bispecific antibodies are used, the bispecific antibodies may be prepared by a variety of methods, such as conventional bispecific antibodies, such as chemically prepared or from hybrid hybridomas (Holliger and Winter, 1993). It may also be one of the bispecific antibody fragments described above. It is preferred to use scFv dimers or diabodies rather than whole antibodies. Diabodies and scFv can be constructed without the Fc region using only variable domains, potentially reducing the effect of anti-genic response. Other forms of bispecific antibodies include the single chain "Janusins" (Traunecker et al., 1991).

Unlike bispecific whole antibodies, bispecific diabodies can be particularly useful because they are easy to prepare and can be expressed in E. coli . Diabodies (and many other polypeptides with antibody fragments) with appropriate binding specificities can be readily selected from phage using phage display (WO 94/13804). If one arm of the dabody has a specificity for eg antigen X and is constant, the library can be made to change the other arm and an antibody with appropriate specificity is selected.

Antigen binding domain

A portion of an antibody comprising a region that binds to and complements some or all of an antigen. If the antigen is large, the antibody can only bind to a specific site of the antigen, which is called an epitope. The antigen binding domain can be provided by one or more antibody variable domains. Preferably, the antibody binding domain comprises the light chain variable region (VL) of the antibody and the heavy chain variable region (VH) of the antibody.

Specific

A situation in which one element of a specific binding pair does not make any significant binding to a molecule other than its specific binding partner. The term also applies, for example, when the antigen binding domain is specific for a particular epitope held by multiple antigens, in which case the specific binding element having the antigen binding domain may bind to various antigens having the epitope. will be.

Functionally equivalent variant

A molecule (variant) that has structural differences with other molecules (parents) but retains some significant homology and at least retains some biological function of the parent molecule, such as its ability to bind to a specific antigen or epitope. . The variant may be in the form of a fragment, derivative or mutant. Variants, derivatives or mutants can be obtained by modifying the parent molecule by adding, deleting, replacing or inserting one or more amino acids or by linking other molecules. Such changes can be made at the nucleotide or protein level. For example, the encoded polypeptide can be a Fab fragment, which is linked to an Fc tail derived from another source. Alternatively, markers such as enzymes, fluorescein and the like may be linked.

Summary of the invention

In the present invention, specific binding elements having specificity for the ED-B tumor fetal domain of fibronectin (FN) are provided.

Specific binding elements according to the invention bind directly with the ED-B domain. In one embodiment, the specific binding element binds to one, some or all FNs including ED-B after FN treatment with protease termolysin. In another embodiment, the specific binding element binds to one, some or all FNs comprising a type III homologous repeat unit containing an ED-B domain. Known FNs have been identified in Carnemola et al. (1989; 1992). References to “all FNs including ED-B” may be taken as references to all FNs including ED-B identified in the paper.

The specific binding element preferably binds to human ED-B, preferably to one or more other species B + FN such as mice, rats and / or chickens. Preferably, the specific binding pair element can bind to non-human fibronectin ED-B such as human fibronectin ED-B and mouse, thereby testing and analyzing sbp elements in animal models. .

Specific binding pair elements according to the invention bind fibronectin ED-B without competing with the antibody BC-1 deposited as publicly available, as described elsewhere in the present invention. BC-1 has strict specificity for human B + isoforms. Specific binding pair elements according to the invention do not bind to the same epitope as BC-1.

The binding of B + FN with specific binding elements according to the invention can be inhibited by the ED-B domain.

In one embodiment of the invention, the binding domain, when measured as a purified monomer, has a dissociation constant (Kd) for ED-B FN of ⁶ × 10 ⁻⁸ M or less.

In one embodiment of the present invention, the binding domain may react with, or bind to, fibronectin ED-B even without pretreatment with fibronectin ED-B with N-glycanase.

Specific binding pair elements according to the present invention may be provided in separate or tablet form, that is, without other specific binding pair elements such as, for example, antibodies or antibody fragments, or other specificity capable of binding to fibronectin ED-B. Formulations or formulations lacking an appropriate binding pair element may be provided. Preferably, the specific binding element according to the invention is provided in substantially pure form. These may be "monoclonal" in the sense that they are from a single clone rather than limited to antibodies obtained using conventional hybridoma technology. As mentioned above, specific binding elements according to the invention can be obtained using bacteriophage display techniques and / or expression in recombinant cells such as bacterial cells or host cells. There is no prior document describing monoclonal specific binding pair elements that bind directly to fibronectin ED-B.

Preferably, the specific binding element comprises an antibody. Specific binding elements may comprise polypeptide sequences in the form of antibody fragments, such as single chain Fv (scFv). Other kinds of antibody fragments can also be used, such as Fab, Fab ', F (ab') 2, Fabc, Facb or diabody (Winter and Milstein, 1991; WO 94/13804). The specific binding element may be in the form of a whole antibody. The whole antibody may be in the form of one of the antibody isoforms such as, for example, IgG, IgA, IgD, IgE and IgM, or in the form of one of the isoforms such as IgG1 or IgG4.

The antibody may be derived from, for example, human, rat or rabbit. It will also be apparent to those skilled in the art that other things are derived. Preferably, the antibody is from human. The term "human" refers to an antibody from which some or all of it is derived from a human cDNA, protein or peptide library. The term includes humanoid peptides and proteins that are not derived from humans and that have been modified to impart human properties to the antibody molecules so that they can cross the human immune system barrier.

The specific binding element also differs from the antigen binding arm (ie specific domain) for fibronectin ED-B as described above, such as for example a bispecific antibody molecule (or a fragment such as F (ab ') ₂ ). It can be an engineered antibody type, or it can be in bivalent or polyvalent molecular form.

In addition to the antibody sequence, specific binding elements may include other amino acids, such as amino acids that form peptides or polypeptides, or amino acids that confer other functional properties beyond the antigen binding capacity to the molecule. For example, specific binding elements can include labels, enzymes or fragments thereof, and the like.

The binding domain may comprise part or all of the VH domain encoded by a germ line compartment or a rearrangement gene compartment. The binding domain may comprise part or all of the VL cappa domain or the VL lambda domain.

The binding domain may comprise a VH1, VH3 or VH4 point-line gene sequence or rearranged form thereof.

Specific binding elements according to the invention may comprise a heavy chain variable region (“VH” domain) derived from human dot line DP47, the sequence of which is shown in FIG. Has residues up to 98. The nomenclature of the 'DP' is described in Tomlinson et al. (1992). The amino acid sequence of CDR3 may be Ser Leu Pro Lys. The amino acid sequence of CDR3 may be Gly Val Gly Ala Phe Arg Pro Tyr Arg Lys His Glu. Thus, the binding domain of the specific binding element according to the invention may comprise a VH domain comprising the amino acid sequence shown in FIG. 1 (a) for CGS1 and CGS2.

The binding domain may comprise a light chain variable region (“VL” domain) derived from human dot line DPL16, the sequence of which is shown as codons 1 to 90 in FIG. 1 (b).

The VL domain may comprise a CDR3 sequence Asn Ser Ser Pro Val Val Leu Asn Gly Val Val. The VL domain may comprise the CDR3 sequence Asn Ser Ser Pro Phe Glu His Asn Leu Val Val.

Specific binding elements of the invention may include variants that are functionally equivalent to the sequence shown in FIG. 1, such as those in which one or more amino acids have been inserted, deleted, substituted or added while retaining the properties disclosed herein. For example, the CDR3 sequence may be different, under one or more changes to the framework region, or under the condition that the specific binding element binds to ED-B, or the framework region may be replaced with another framework region or modified form. .

One or more CDRs derived from the VL or VH domain of the antigen binding domain of an antibody disclosed herein may be used in a so-called "CDR-graft" wherein in the transplant one or more CDR sequences of the first antibody are not sequenced of that antibody. It is located within the backbone of the sequence, for example the backbone of other antibodies disclosed in EP-B-0239400. CDR sequences for CGS1 and CGS2 are shown in FIGS. 1 (a) and 1 (b).

The specific binding element according to the present invention may be one that competes with an antibody or scFv as described herein to bind fibronectin ED-B. The contention between binding elements can be easily tested in vitro, for example by attaching a specific reporter molecule to one binding element so as to be detectable even in the presence of other unlabeled binding element (s). Specific binding elements that bind to the same epitope or overlapping epitopes can be identified.

Specific binding elements according to the invention can be used in a method comprising the step of allowing or allowing a specific binding element to bind to its epitope. Binding occurs after administration of specific binding elements to mammals such as rodents such as humans or mice.

The present invention provides the use of specific binding elements used as tumor diagnostic reagents as above. The animal model experiments described below demonstrate that the binding elements according to the invention are useful for tumor localization in vivo.

Preferred specific binding elements according to the present invention include elements that bind to human tumors that exhibit an invasive and angiogenic phenotype, for example in cryostat sections, and elements that bind to fetal tissue, such as in cryostat sections. The binding can be demonstrated by immunocytochemical staining.

In a preferred embodiment of the present invention, the specific binding element does not bind to or binds to tenasin, an extracellular interstitial protein, to a significant extent.

In another preferred embodiment of the invention, the specific binding element binds to a significant extent even if it does not bind or binds to normal human skin, as demonstrated, for example, in cryostat sections and / or using immunocytochemical staining. I never do that.

In another embodiment for specific binding elements according to the invention, one selected from liver, spleen, kidney, stomach, small intestine, large intestine, ovary, uterus, bladder, pancreas, adrenal gland, skeletal muscle, heart, lung, thyroid and brain The above normal tissues (such as in sections at low temperature and / or as evidenced by immunocytochemical staining) do not bind or do not bind to any significant extent.

Specific binding elements for ED-B can also be used as in vivo targeting agents, which can also be used to specifically demonstrate the presence and location of tumors that express or relate fibronectin ED-B. The specific binding element can also be used as an imaging agent. In the present invention, a method of expressing fibronectin ED-B or determining the presence of a cell or tumor associated with the expression of fibronectin ED-B is provided, which method comprises contacting a cell with a specific binding element provided to the cell. Determining the binding of specific binding elements. The method can be carried out in an in vivo test or an in vitro test on test cell samples isolated from the body.

The reactivity of the antibody to the cell sample can be determined using an appropriate method. Attaching tags to individual reporter molecules is one possible method. The reporter molecule will directly or indirectly generate a signal that can be detected and preferably measured. The reporter molecule may be linked directly or indirectly, for example by covalent bonds such as peptide bonds or by non-covalent bonds. Linkage by peptide binding will be the result of recombinant expression of the gene fusion encoding the antibody and reporter molecule.

One preferred method is to covalently link each antibody with each fluorescent dye, phosphor or laser dye having spectroscopically separated light absorbing or emissive properties. Suitable fluorescent dyes are fluorescein, rhodamine, phycoerythrin and Texas red. Suitable dye source dyes include diaminobenzidine.

Other reporter molecules include, directly or indirectly, electronically detect or otherwise detect particulate matter, such as macromolecular colloidal particles or dyed latex beads, magnetic or paramagnetic materials, and detection signals that can be observed with the eye. Included are biological or chemical actives that can be recorded by the method. Such molecules can be, for example, enzymes that catalyze color development or discoloration reactions or change their electrical properties. They are susceptible to excited molecules and can exhibit characteristic spectroscopic absorbance or radioactivity as electrons move between energy levels. This may include chemicals used with biosensors. Biotin / avidin or biotin / streptavidin and alkaline phosphatase detection systems can be used.

The method of determining the binding is not a feature of the present invention and those skilled in the art can select an appropriate method according to their preferences and general knowledge.

The signal generated by each antibody-reporter conjugate can be used to derive quantifiable absolute data or comparative data for relevant antibody binding in cell samples (standard samples and test samples). In addition, general nuclei staining with dyes such as propidium iodine can be used to calculate the total cell population present in the sample, thereby revealing the quantitative ratio of each cell population to the total cells. ^When attaching a radionucleotide such as ¹²⁵ I, ¹¹¹ In or ^99m Tc to an antibody, if the antibody preferentially localizes to tumor tissue rather than normal tissue, a gamma camera is used to detect the presence of a radiolabel within the tumor tissue. And quantification. The quality of the tumor image obtained is directly related to the signal: noise ratio.

The antibody can also be used as a diagnostic agent for tracking newly vasculatured tumors and, for example, in modified form, to carry cytotoxic substances or to cause blood clotting in new blood vessels, thereby developing tumors. Indirect forms of tumor therapy are achieved by depleting oxygen and nutrients.

The present invention also provides the use of specific binding elements which are used as therapeutic agents when coupled, bound or engineered as fusion proteins to have effector function. Specific binding elements according to the invention are used to allow toxins, radioactive agents, T-cells, killer cells or other molecules to target tumors that express or are associated with the antigen of interest.

Accordingly, another object of the present invention is to use a method of treatment comprising administering a specific binding element provided in the present invention, a pharmaceutical composition comprising the specific binding element, and for example using a pharmaceutically acceptable excipient. To provide for the use of a specific binding element, which is used in the manufacture of a medicament for administration in a pharmaceutical composition or method of manufacturing a medicament comprising formulating a specific binding element.

According to the invention, provided compositions may be administered to a patient subject. Administration is preferably administered in a "therapeutically effective amount", wherein the effective amount is an amount sufficient to have a beneficial effect on the patient. The advantageous effect may mean that the at least one symptom should be at least as high as possible. The actual dosage, rate of administration and time of administration will depend on the nature and severity of the disease being treated. Treatment regimens, such as dose determination, etc., are determined at the responsibility of the general practitioner or other physician. Suitable antibody doses are known in the art (Ledermann et al. (1991); Bagshawe K.D. et al. (1991)).

The compositions may be administered alone or in combination with other therapeutic agents, which may be administered simultaneously or sequentially, depending upon the condition of the disease to be treated when combined.

Pharmaceutical compositions according to the invention and for use according to the invention may comprise, in addition to the active ingredient, pharmaceutically acceptable excipients, carriers, buffers, stabilizers or other substances well known in the art. These substances should be nontoxic and should not interfere with the efficacy of the active ingredient. The exact nature of the carrier or other material will depend on the route of administration such as oral administration or injection such as intravenous injection.

Pharmaceutical compositions for oral administration may be tablets, capsules, powders or solutions. Tablets may include solid carriers such as gelatin or adjuvants. Liquid pharmaceutical compositions generally include liquid carriers such as water, petroleum, animal and vegetable oils, mineral oils or synthetic oils. Physiological saline, dextrose or other saccharide solutions or glycols such as ethylene glycol, propylene glycol or polyethylene glycol may be included.

For intravenous injection or injection to a lesion site, the active ingredient will be in a parenterally acceptable liquid solution, in the form of a solution that is pyrogen free and has adequate pH, isotonicity and stability. Those skilled in the art will be able to prepare appropriate solutions using isotonic carriers such as, for example, sodium chloride injections, Ringer's injections, Ringer's injections with lactate added. Preservatives, stabilizers, buffers, antioxidants and / or other additives may be included as needed.

Specific binding elements according to the invention can be prepared by expression from nucleic acids encoding them. Nucleic acids encoding specific binding elements themselves form an aspect of the present invention, and methods for producing specific binding elements comprising the step of expression from the coding nucleic acid also form an aspect of the present invention. Expression can be conveniently accomplished by culturing the recombinant host cell comprising the nucleic acid under appropriate conditions.

The nucleic acid can encode both the amino acid sequence and the functional equivalent of the antigen-binding domain of the antibody described in the present invention. By adding, replacing, deleting or inserting one or more nucleotides, modifications can be made at the nucleotide level, which may or may not be reflected at the amino acid level depending on the degree of degradation of the genetic code.

Systems for cloning and expressing polypeptides in a variety of different host cells are well known. Suitable host cells include bacteria, mammalian cells, yeast and baculovirus systems. Mammalian cell lines available in the art for the expression of heterologous polypeptides include Chinese hamster ovary cells, HeLa cells, baby hamster kidney cells and other cells. A common and preferred bacterial host cell is E. coli .

Methods of expressing antibodies and antibody fragments in prokaryotic cells such as E. coli are well established in the art (Pluckthun, (1991)). Expression in eukaryotic cells in culture can also be used by those skilled in the art as a method for producing specific binding elements (Reff, (1993); Trill et al. (1995)).

Suitable vectors may be selected and prepared by including appropriate regulatory sequences, including promoter sequences, termination sequences, polyadenylation sequences, enhancer sequences, marker genes, and other suitable sequences. Vectors may be suitable such as plasmids, viruses such as phage or phagemids (Molecular Cloning: a Laboratory Manual; 2nd Edition, Sambrook et al., 1989, Cold Spring Harbor Laboratories). Numerous known techniques and protocols for nucleic acid engineering, such as preparation of nucleic acid constructs, mutagenesis, sequencing, introduction of DNA into cells and gene expression and protein analysis, are described in detail in the Short Protocols in Molecular. Biology, 2nd edition, Ausubel et al. Eds., John Wiley & Sons, 1992). Of these documents, Sambrook et al. And Ausubel et al. Are incorporated herein by reference.

Accordingly, another object of the present invention is to provide a host cell comprising the nucleic acid disclosed herein. Another object of the present invention is to provide a method comprising the step of introducing such nucleic acid into a host cell. As the introduction method, all available techniques may be used. For eukaryotic cells, suitable techniques include calcium phosphate transfection, DEAE-dextran, electroporation, liposome-mediated transfection, and transfection with retroviruses or other viruses such as baculoviruses for head virus or insect cells. Introduction may be included. In the case of bacterial cells, suitable techniques include transformation with calcium chloride, electroporation and transfection with bacteriophages.

After the introduction, expression from the nucleic acid can be induced, for example, by culturing the host cell under gene expression conditions.

In one embodiment of the invention, the nucleic acid of the invention is integrated into the genome (eg chromosome) of the host cell. This integration can be facilitated by enclosing sequences that facilitate recombination with the genome according to standard techniques.

After producing the specific binding element, the element may comprise, for example, a pharmaceutical or diagnostic agent, such as a kit comprising one or more reagents for determining the binding of the element to cells as described above in addition to the specific binding element. It will be used in the various methods disclosed in the present invention, such as used to make.

Other objects and embodiments of the invention will be apparent to those skilled in the art. In order to provide a thorough understanding of the present invention, it is intended to provide the following examples as examples, which are not intended to limit the invention.

Example list

Example 1 Isolation of Human scFv Specific to ED-B Domain of Human FN

Example 2-Affinity Maturation of Human scFv Specific to ED-B Domain of Human FN

Example 3 Specificity of Mature scFv Affinity to ED-B-containing Fibronectin

Example 4 Use of affinity matured anti-ED-B scFv used for immunocytochemical staining for human and mouse tumor sections.

Example 5 Use of Affinity Matured Anti-ED-B for In Vivo Targeting to Human Tumors

Example 6-Targeting to F9 teratocarcinoma graft in rats in nude mice

Example 1 Isolation of Human scFv Having Specificity to the ED-B Domain of Human FN

Recombinant antibodies were selected using a human scFv phage library (Nissim et al., 1994). Two different types of ED-B isoforms were used as antigen sources for selection, in which case the two types were recombinant human proteins.

Recombinant FN peptides comprising type III repeat units 2-11 (B−) and 2-11 (B +) were expressed in Escherichia coli .

Constructs were constructed using FN cDNAs obtained from pFH154 (Kornblihtt et al., 1985), λ F10 and λ F2 (Carnemolla et al., 1989) clones. The cDNA construct spanning bases 2229-4787 was inserted into the vector pQE-3 / 5 using a QIA expression kit from Qiagen (Chatsworth, CA). FN-III 2-11 (B-) and (B +) were purified by immunoaffinity chromatography using mAb 3E3 (Pierschbacher et al. 1981) conjugated with Sepharose 4B (Pharmacia). Type III homologous repeat unit 7B89 using cDNA from FN 2-11 (B +) and FN 2-11 (B-) as a template and PCR reaction (polymerase chain reaction) using UltMa DNA polymerase. DNA fragments for preparation of synthetic FN fragments, including 789, ED-B and FN-6, were produced. Using a QIA expression kit (Qiagen), primers were designed to clone the PCR product to pQE-12. They were then transformed into E. coli and expressed. All cDNA clones were sequenced using the Sequenase 2.0 DNA Sequencing Kit (USB).

Using the hexahistidine tag at the FN fragment carboxy terminus, the recombinant protein was purified by Ni-NTA chromatography (IMAC) according to the manufacturer's instructions (Qiagen). ED-B cDNA was cloned into the λgt11 bacteriophage vector to prepare an ED-B-βGal fusion protein to form the λED-B clone. The clone λchFN60 (including part of the ED-B sequence) was derived in the form of a fusion protein from pchFN60 cloning (Norton et al., 1987), which is a cloned FN of chickens.

For selection of human scFv phage libraries, each of the two different recombinant antigens (7B89 and ED-B) was panned three times. The antigen was coated in immunotubes (Nunc; mexif, Roskilde, Denmark) overnight in PBS (20 mM phosphate buffer, 0.15 M NaCl, pH 7.2) at a concentration of 50 μg / ml. The first of these antigens was recombinant FN fragment 7B89, which had adjacent type III FN homologous repeats on either side of the ED-B domain, which were coated on immunotubes overnight at 4 ° C. The second antigen used was recombinant ED-B (Zardi et al., 1987) with a hexahistidine tag at the carboxy terminus, since the protein does not contain lysine residues, the ED-B is sited on a reactive ELISA plate (Nunc; Covalink). The terminal amino group of the first amino acid can be used for immobilization with -specific covalent bonds. Coating was carried out overnight at room temperature.

After three pannings, eluted phages were infected with HB2151 E. coli cells and plated as described in the literature (Nissim et al., 1994). After each selection step, 95 ampicillin-resistant single colonies were screened and antigen-specific scFv identified using ELISA method. Clones showing the highest ELISA signal for the antigen used for panning were selected for use in other assays and affinity maturation tests. These clones were also demonstrated to specifically stain breast tumors and glioblastoma polymorphic sections by immunochemical staining, which will be described in more detail in Example 4.

Example 2-Affinity Maturation of Human scFv Specific for ED-B Domain of Human FN

Clones 35GE (selected using 7B89) and 28SI (selected using only the ED-B domain) were selected as candidate antibodies for affinity maturation. A simple affinity maturation method was investigated to change the L chain as one of the means of improving affinity, which randomized 6 middle residues (DSSGNH) of L chain CDR3 using degenerate oligonucleotides and PCR (FIG. 1). The method was based on a method capable of diversifying 20 ⁶ = 6.4 × 10 ⁷ sequences by randomizing. The region present with the H chain CDR3 is located at the center of the antigen binding site (Padlan, 1994). In addition, to avoid the possibility of electrostatic effects on the selection, the arginine residues preceding the six residues were directly mutated to serine.

Primers LMB3 (5 'CAG GAA ACA GCT ATG AC 3') and CDR3-6-VL-FOR (5 'CTT GGT CCC TCC GCC GAA TAC CAC MNN MNN MNN MNN MNN MNN AGA GGA GTT ACA GTA ATA GTC AGC CTC 3' Plasmids from bacterial single colonies expressing "parent" scFv fragments were amplified by PCR (94C [1 ']-55C [1']-72C [1'30 "], 25 cycles; buffers and conditions (See Marks et al., 1991.) The resulting product was gel-purified (removal of plasmid traces containing the original scFv gene), primers LBM3 and J1-Not-FOR (5'ATT GCT TTT CCT TTT TGC). GGC CGC GCC TAG GAC GGT CAG CTT GGT CCC TCC GCC 3 ') was used as the template for the second amplification step (94C [1']-55C [1 ']-72C [1'30 "], 25 cycles). . The crude PCR product is unfolded into a single band of correct molecular weight on an agarose gel, purified directly from the PCR mixture using Spin-Bind (FMC, Rockland, Maine, USA) and double digested with Nco1 / Not1. Next, gel-purified Nco1 / Not1-digested phagemid pHEN1 containing a dummy Ncol / Not1 insert to facilitate separation of the double-digested vector from the mono-digested vector (Hoogenboom et al., 1991). Connected. The vector was prepared using the Qiagen (Chatsworth, CA, USA) plasmid maxiprep kit. About 5 μg of plasmid and insert were used in the mixture for the ligation, and the mixture was extracted once with phenol, once with phenol / chloroform / isoamyl alcohol (25: 25: 1) and then glycogen (Boehringer Mannheim, Germany) was used as a carrier to precipitate in ethanol and then dried (speed-vac). The pellet was resuspended in 20 μl of water and electroporated into TG1 E. coli cells (Gibson, 1984) with electrosoluble. Typically, electrolytic cells with a concentration of 10 ⁹ transformants / μg are used when using a glycerol reservoir, or 10 ¹⁰ transformants / when using freshly prepared electrolytic cells. It was used at a concentration of μg. These values typically produced> 10 ⁷ clones by the method described herein.

The maturation library was then processed as in Nissim Library (Nissim et al., 1994) to produce phage particles, which were used for one selection in immunotubes using 7B89 (10 μg / ml) as antigen. Next, one kinetic choice was made (Hawkins et al., 1992). This optional step involves incubating the biotinylated 7B89 (10 nM) and phage suspension (about 10 ¹² tu) in 2% milk-PBS (2% MPBS) for 5 minutes, followed by non-biotanylated 7B89 (1 by adding M) and allowing to stand for 30 minutes to cause competition. 100 μl of streptavidin-coated dynabe (Dynal: M480) preblocked in 2% MPBS was added to the reaction mixture, followed by mixing for 2 minutes to capture the magnet, followed by (PBS + 0.1% Tween-20) It was washed 10 times alternately with PBS. Phage was eluted from the beads using 5 ml of 100 mM triethylamine. The solution was then neutralized with 0.25 ml 1M Tris at pH 7.4 and used to infect exponentially increasing HB2151 cells (Nissim et al., 1994). The scFv-containing supernatants were produced using 95 ampicillin-resistant single colonies (Nissim et al., 1994) and screened using ELISA, immunohistochemistry and BIA core methods to identify optimal binders. Next, they were subcloned between the Sfi1 / Not1 sites of the pDN268 expression vector (Neri et al., 1996), wherein a phosphorylated tag, a FLAG epitope and a hexahistidine tag were attached to the C-terminus of the scFv.

A single colony of the relevant antibody subcloned to pDN268 was grown at 37 ° C. in 2 × TY containing 100 mg / L ampicillin and 0.1% glucose. When the cell culture reached OD ⁶⁰⁰ = 0.8, IPTG was added to a final concentration of 1 mM and then continued to grow at 30 ° C. for 16-20 hours. After centrifugation (GS-3 Sorvall rotor, 7000 rpm, 30 minutes), the supernatant was filtered and concentrated, and loading buffer (50 mM phosphate, pH 7.4, 500 mM) using a tangential flow device (Minisette) NaCl, 20 mM imidazole). The solution thus obtained was loaded into 1 ml Ni-VTA resin (Qiagen), washed with 50 ml of loading buffer and eluted with elution buffer (50 mM phosphate, pH 7.4, 500 mM NaCl, 100 mM imidazole). Purified antibodies were analyzed by SDS-PAGE (Laemmli, 1970) and dialyzed against PBS at 4 ° C. Purified scFv formulations were further processed by gel-filtration using an FPLC device (Pharmacia) equipped with an S-75 column, due to the binding effect (Nissim et al., 1994; Crothers and Metzger, 1982). This is because scFv fragments are known to show good artificial binding to BIA cores. The concentration of the antibody in the FPLC-tablet monomer fraction was determined by spectrophotometry assuming that the absorbance of 1.4 units at 280 nm was observed for 1 mg / ml scFv solution.

The following (iv) to (v) antigens were used to determine the binding of various concentration ranges of monovalent scFv dissolved in PBS to the BIA core machine (Pharmacia Biosensor) in the range of 0.1 to 1 μM: ) Biotinylated recombinant FN fragment 7B89 1000 Resonance Units (RU) immobilized on a chip coated with streptavidin, which was specifically bound by 250 RU of scFv; (Ii) recombinant ED-B 200RU chemically immobilized to an N-terminal amino group, which was specifically bound by 600RU scFv; (Iii) ED-B-rich fibronectin WI38VA (see Example 3) 3500RU, which was specifically bound by 150RU scFv. In accordance with the manufacturer's instructions, a kinetic analysis of the data was performed. Based on the BIA core quantitative analysis for the antibody-containing supernatant, one affinity-mature version was selected for each scFv: clones CGS-1 and ED-B recombinant FN fragments derived from those selected using the 7B89 fragment. It is CGS-2 derived from what was selected using. The association rate constant (k _on ) and dissociation rate constant (k _off ) are shown in Table 1 along with the equilibrium dissociation constant (Kd) calculated for scFv and initial clone 28SI. Although CGS-1 and CGS-2 clones have equimolar dissociation constants in nanomolar units, clone CGS-2 has a Kd of 1 nM (improved at 110 nM) compared to all three proteins tested for the sensor chip. Showing that the most improved for the parent clone (Table 1). The reason for this improvement is mainly due to the slower kinetic dissociation constant (˜10 ⁻⁴ s ⁻¹ ), which was measured with monomeric antibody preparations (not shown).

The maturation strategy appears to be general and therefore maltose binding protein, cytochrome C, mouse extracellular domain endoglin (DN, L. Wyder, R. Klemenz), cytomegalovirus (AP, G. Neri, R. Botti, PN ), Affinity improving antibodies to nuclear tumor marker HMGI-C protein (AP, P. Soldani, V. Giancotti, PN) and ovarian tumor marker placental alkaline phosphatase (M. Deonarain and AA Epenetos) were obtained. Thus, this strategy appears to be at least as effective as other mutation strategies (Marks et al., 1992; Low et al., 1996), and has similar affinity to that derived from very large phage antibody libraries (Griffiths et al., 1994; Vaughan et al., 1996). An antibody having is obtained.

The affinity matured clones CGS-1 and CGS-2 were sequenced and aligned in the human pointline antibody V gene (V-BASE) database and then translated using MacVector software. The two clones had the most homology to the human dot line DP47 (VH3), and each clone had a different VH CDR3 sequence (Fig. 1). The two clones' VL genes were the DPL16 dot lines used to construct the synthetic scFv repertoire of humans described in Nissim et al. (Nissim, 1994). The VL CDR3 sequence differed by 4 out of 6 random sequences (FIG. 1B).

Kinetics and Dissociation Constants of Monomeric scFv Fragments CGS-1 and CGS-2 for ED-B Domain-containing Proteins antigen ED-B 7B89 FN WI38VA scFv CGS-1 SI28 CGS-2 CGS-1 SI28 CGS-2 CGS-1 SI28 CGS-2 k _off (s ^-1 ) * 7.0x10 ^-3 2.7 x 10 ^-2 1.5x10 ^-4 3.9 x 10 ^-3 3.0 x 10 ^-2 2.3 x 10 ^-4 5.0 x 10 ^-3 7.1 x 10 ^-2 6.5x10 ^-4 k _on (M ^-1 s ^-1 ) * 1.3 x 10 ⁵ 2.5 x 10 ⁵ 1.3 x 10 ⁵ 1.1 x 10 ⁵ 2.9 x 10 ⁵ 1.1 x 10 ⁵ 4.1 x 10 ⁵ 1.2 x 10 ⁶ 2.9 x 10 ⁵ Kd (M) * 5.4 x 10 ^-8 1.1x10 ^-7 1.1x10 ^-9 3.5 x 10 ^-8 1.0x10 ^-7 2.1 x 10 ^-9 1.2 x 10 ^-8 5.9 x 10 ^-8 2.4 x 10 ^-9

Description of [Table 1]

Experiments were conducted as described in the Materials and Methods section.

The k _off and k _on values are accurate to ± 30%, based on the concentration determination accuracy, and these results are obtained when different sensogram areas are used in the fitting process for slightly different results. . Kd = k _off / k _on .

Example 3 Specificity of Mature scFv Affinity to ED-B-containing Fibronectin

First, the ELISA method was used to assess the immunoreactivity of the two affinity mature scFv, CGS-1 and CGS-2, which were FB lacking mAb BC-1 (recognizing B-FN isoforms) and ED-B. Direct recognition was compared with mAb IST-6 (Carnemolla et al., 1989; 1992). The characteristics of these mAbs have already been reported (Carnemola et al., 1989; 1992). Subsequently, an extensive analysis of specificity was carried out using a panel of extensive FN fragments and recombinant fusion proteins induced by termolysin treatment.

Antigens used in ELISA and immunoblotting were prepared as follows. As previously reported (Zardi et al., 1987), FN was purified from conditioned medium of human plasma and WI38VA13 cell line. Purified FN was digested with termolysin (protease VII; sigma chemical) as reported in the Carnemolla et al. (Carnemolla et al., 1989). As reported previously (Borsi et al., 1991), the original FN 110kD (B−) and the original FN 120kD (B +) fragments (see FIG. 2) were purified from FN digests. As reported in Saginati's paper (Saginati et al., 1992), the large isoforms of tenasin-C were purified. Recombinant protein was expressed and purified as described in Example 1. SDS-PAGE and western blotting were performed as described in Carnemolla's paper (Carnemolla et al., 1989).

All of the antigens used in the ELISA method were diluted in PBS to a concentration of 50-100 μg / ml and coated in immuno-plate wells (Nunc, Roxkille, Denmark) overnight at 4 ° C. Unbound antigen was removed with PBS and the plates blocked for 2 hours at 37 ° C. with PBS containing 3% (w / v) bovine serum albumin (BSA). It was then washed four times with PBS (PBST) containing 0.05% Tween 20. Next, it was left for 1.5 hours at 37 ℃ to bind the antibody; ScFv was previously incubated with antiserum to the tag sequence: mAb M2 [Kodak, New Haven, Connecticut] for FLAG tags and 9E10 [ATCC, Rockville, Maryland] for myc tags. Control antibodies used in the tests were mAbs BC-1 and IST-6. After 4 washes with PBST, the plate was 37 ° C. with biotinylated anti-mouse IgG (Bio-SPA Division, Milan, Italy) of 1: 2000 diluted in PBST + 3% BSA. Incubated for 1 hour at. Wash repeatedly and dilute streptavidin-biotinylated alkaline phosphatase complex (Bio-SPA division, Milan, Italy) at 37 ° C. over 1 h (1: 800 in PBST containing ₂ mM MgCl ₂₎ ). The reaction was developed in 10% diethanolamine at pH 7.8 using a phosphatase substrate (cisma), which was a tablet type, and the optical density at 405 nm was recorded. The results are shown in Table 2 below.

CGS-1 CGS-2 BC-1 IST-6 Plasma FN 0.07 0.04 0.09 1.73 WI38VA FN 1.16 0.72 1.20 1.12 n110 kD (B-) 0.03 0.01 0.05 1.20 n120 kD (B +) 0.82 0.81 1.20 0.02 rec FN7B89 1.11 1.02 1.02 0.01 rec FN789 0.01 0.01 0.05 1.25 rec ED-B 1.12 1.32 0.15 0.04 rec FN-6 0.01 0.01 0.08 0.03 Tenasin 0.01 0.02 0.06 0.02

The immunoreactivity of scFv and monoclonal antibodies against fibronectin-derived antigens was measured using ELISA method. The measured value represents the OD value after subtracting the base signal with the value measured at 405 nm. The data are averages of four experimental values and show a standard deviation of up to 10%.

The different types of fibronectin used in the experiments are as follows: plasma FN = human plasma fibronectin; WI38-VA FN = fibronectin from the supernatant of SV40-transformed fibroblasts (Zardi et al., 1987); n110kD = FN domain 4 treated with termolysin without ED-B; n120kD = FN domain 4 treated with thermolysine, including ED-B; rec FN7B89 = adjacent Type III FN homologous repeat units are present on both sides of the ED-B domain; rec FN789 = Type III FN homology repeat unit with ED-B domain; rec ED-B = recombinant ED-B alone; rec FN6 = recombinant FN domain 6.

Both CGS-1 and CGS-2 recognize recombinant ED-B peptides as well as recognize native FN fragments and recombinant FN fragments containing ED-B sequences, while FN fragments lacking ED-B do not. Do not. In addition, CGS-1 and CGS-2 do not react with tenasin (which includes 15 type III homologous repeat units: Siri et al., 1991) and plasma FN to the extent that they are detected in the product of termolysin degradation. Does not contain the ED-B sequence (Zardi et al., 1987). In contrast, CGS-1 and CGS-2 react strongly with FN purified from SV40-transformed cell line WI38VA. About 70-90% of the FN molecules derived from the cell line include ED-B, which is a termolysin digestion experiment and S1 nuclease experiment conducted using purified RNA and total RNA prepared from the cell line. (Zardi et al., 1987; Borsi et al., 1992). The specificity of the scFv to the ED-B component of FN was demonstrated in more detail using soluble recombinant ED-B which inhibits the binding of CGS-1 and CGS-2 to FN on WI38VA cells (data not shown).

The data clearly showed that only CGS-1 and CGS-2 specifically reacted with the FN derivative containing the ED-B domain. Except in the case of recombinant ED-B, they all show the same reactivity as mAb BC-1, which is not recognized by BC-1. The intensity of the ELISA signal obtained as a value relative to the mAb control reflects that the two scFvs have high specificity for the ED-B-containing antigen.

The specificity of CGS-1 and CGS-2 was examined in more detail for immunoblots using FN and its termolysin digests derived from plasma and WI38VA cells. Upon degradation by termolysin, FNs obtained from WI38VA cells (most of which contain ED-B) produce 120kD fragments (including ED-B) and a few 110kD fragments lacking ED-B ( 2A; Zardi et al., 1987). Further digestion of the 120kD domain yields two fragments: one produces an 85kD fragment and a 35kD sequence comprising almost the entire ED-B sequence at the carboxy terminus (FIG. 2A; Zardi et al., 1987).

On the left side of FIG. 2B is shown the Coumasyl staining gel for the protein fraction analyzed by immunoblotting. Plasma FN (lane 1) and the termolysin digest of protein (lane 3 comprising 110 kD protein and lane 4 comprising digested 110 kD protein) were not recognized by CGS-1 and CGS-2. In contrast, ED-B-rich FNs obtained from WI38VA cells were recognized by both scFv fragments, both in the intact form (lane 2) and subsequently in the termolysin digest (lanes 5, 6 and 7). The smallest FN-derived fragment that can be specifically recognized by CGS-1 is the 120 kD protein (including the portion spanning Type III repeats 2-11), while CGS-2 is the ED-B N-terminus. In addition, it is possible to recognize 85kD fragments ranging from repeating units 2-7 (FIG. 2B; Zardi et al., 1987). These results indicate that the two types of scFv react with distinct epitopes in the ED-B sequence. The fact that CGS-2 binds to the 85 kD fragment indicates that an epitope for this clone is present at the amino terminus of ED-B. On the other hand, the fact that 120 kD fragments do not bind CGS-1 when degraded to 85 kD fragments indicates that CGS-1 recognizes epitopes located on the more carboxy terminus of the ED-B molecule.

With the use of recombinant FN fragments and fusion proteins with or without ED-B sequences, the fine specificity of CGS-1 and CGS-2 was examined in more detail by immunoblotting. The FN fusion protein was prepared according to the method described in Carnemolla et al. (1989). The results of this experiment are shown in FIG. 3; See Carnemolla et al., 1992, for a schematic comparison of the human FN domain structure. The binding profiles obtained were previously found for purified FN and proteolytic cleavage products by ELISA and immunoblotting methods: CGS-1 and CGS-2 were associated with ED-B-containing FN fragments (lanes 2 and 4). Strongly responded but confirmed essentially no reactivity to the FN sequences lacking ED-B (lanes 1 and 3). CGS-1 did not react with human ED-B fusion protein (lane 5) or chicken ED-B fusion protein (lane 6), but CGS-2 reacted strongly with both fragments (FIG. 3). These results may reflect some structural constraints of epitopes present in ED-B-containing FNs recognized by CGS-1; For example, the epitope is susceptible to protein denaturation or may not be properly provided when fractionated by SDS-PAGE and transferred to a solid support such as nitrocellulose.

Overall, the results indicate that CGS-1 and CGS-2 bind strongly and specifically to ED-B-containing FNs, and the binding sites are distinct from each other, but with ED-B structures recognized by mAb BC-1. Also proves to be a distinct area.

Example 4 Use of affinity matured anti-ED-B scFv used in immunocytochemical staining for human and mouse tumor sections

Both CGS-1 and CGS-2 have been used to immunologically localize ED-B-containing FN molecules present in various normal and neoplastic (tumor) tissues of humans. Since B-FN isoforms are known to be expressed in macrophages and fibroblasts during skin wound healing (Carnemolla et al., 199; Brown et al., 1993), skin was selected as normal tissue. Staining specificity was previously analyzed using anti-fibronectin mAbs for two selected tumors: glioma cell polymorphisms showed that endothelial cells present in the blood vessels of this tumor were in substantial proliferation and included expression of B-FN isoforms. The angiogenesis process has been studied in greater detail because it is in an increased state (Castellani et al. 1994). In addition, studies using various normal, hyperproliferative and neonatal breast tissues in humans have provided more detailed evidence of the correlation between angiogenesis and B-FN expression (Kaczmarek et al., 1994).

For the experiments described herein, immunohistochemical staining for CGS-1 and CGS-2 was performed and the results were stained for mAb BC-1 (recognizing B-FN isoforms) and all known FN isoform variants ( IST-4) compared to staining results for other mAbs known to respond to all or only to FN isoforms lacking ED-B (IST-6). The properties of all these control antibodies have been previously reported (Carnemolla et al., 1989; 1992).

Normal tissue and tumor tissue were obtained from samples taken during surgery. It is already established that tissue preparation and immobilization are important for accurate and sensitive detection of FN-containing molecules (Castellani et al., 1994). For immunohistochemistry, 5 μm thick cryostat sections were air dried and fixed in cold acetone for 10 minutes. Immunostaining was performed using streptavidin-biotin alkaline phosphatase complex staining kit (Bio-SPA Division, Milan, Italy), naphthol-AS-MX-phosphate and Fast Red TR (Sigma). Gill's hematoxylin was used as reverse staining and then mounted on slides in glycerol (Dako, Carpenteria, Calif.) As previously reported by Castellani (Castellani et al., 1994). To further analyze the specificity in experiments in which a positive response to tissue staining was obtained, the specificity for ED-B was examined by preincubating the antibody with the recombinant ED-B domain and then detecting as described above.

The experimental results showed that both CGS-1 and CGS-2 responded to the same histological structure as mAb BC-1. The staining pattern obtained from the skin with CGS-1, CGS-2 and BC-1 reflects the absence of ED-B in the FN expressed in the dermis. In staining invasive tubular carcinoma sections, CGS-1, CGS-2 and BC-1 showed a limited staining distribution, limited to the border between neoplastic cells and epilepsy. This is consistent with the fact that although the entire FN is evenly distributed in tumor epilepsy, the expression of B-FN is localized only in specific regions, which have previously been successfully localized with mAb BC-1 in invasive vascular carcinoma. (95% of cases) (Kaczmarek et al., 1994).

Previous findings in the staining of BC-1 in glioblastoma polymorphic tumors have been confirmed. Castellani et al. (1994) observed typical staining patterns of renal glomeruli-like vascular structures, and in our experiments it was observed that CGS-1 and CGS-2 yielded this quantitatively the same result.

However, there is an important difference between CGS-1 and CGS-2 and mAb BC-1: The two types of human scFv bind to both chicken B-FN and mouse B-FN while BC-1 is strictly It has been demonstrated that specificity only with human B-FN. CGS-2 responds to chicken embryos (data not shown), but both CGS-1 and CGS-2 respond to mouse tumors.

Vascular structures in rat F9 teratocarcinoma sections also showed CGS-1 staining. In contrast, all normal test tissues of the mouse (liver, spleen, kidney, stomach, small intestine, large intestine, ovary, uterus, bladder, pancreas, adrenal gland, skeletal muscle, heart, lung, thyroid and brain) are CGS-1 and CGS-2 There was a negative response to staining (data not shown). Structures stained in the F9 teratocarcinoma section were found to be specific for ED-B by using recombinant ED-B domains that completely inhibit staining (data not shown).

Example 5 Use of Affinity Matured Anti-ED-B for In Vivo Targeting to Human Tumors

Using the human melanoma cell line SKMEL-28 to subcutaneously inject 1 × 10 ⁷ cells / mouse into one flank of 6-10 week old nude nude mice (Balb-c or MF-1; Harlan, UK) to develop transplanted tumors I was. When the diameter of the tumor reached about 1 cm in the tail of the mouse with the tumor, 100 μl intravenously injected with a solution of scFv ₁ -Cy7 ₁ dissolved at ₁ mg / ㎖ in PBS.

100 μl of 1 M sodium bicarbonate and 200 μl of CY7-bis-OSu (Amersham; Cat. Nr. PA17000; 2 mg / ml in DMSO) was added to 1 ml of an antibody solution (1 mg / ml) dissolved in PBS. By addition, the recombinant antibody was labeled with CY7. After 30 minutes at room temperature, 100 μl of 1M Tris at pH = 7.4 was added to the mixture, and then labeled antibody using a disposable PD10 column (Pharmacia Biotech, Piscataway, NJ, USA) equilibrated with PBS. Was separated from unreacted dye. Eluted green antibody fractions were concentrated to approximately 1 mg / ml using Centricon-10 tubes (Amicon, Beverly, Massachusetts, USA). The label ratio achieved was generally close to one molecule of CY7: one molecule of antibody. This was measured spectroscopically using a 1 cm cuvette, assuming that 1 mg / mL antibody solution gave 1.4 units of absorbance at 280 nm, and the molar extinction coefficient of CY7 at 747 nm was 200,000 (M). ^-1 cm ^-1 ) and CY7 absorbance at 280 nm was neglected, using band-shift (Neri et al., 1996b), affinity-chromatography performed on an antigen column, or BIA core analysis of antibody samples after labeling. Immune responsiveness was confirmed. The mice were inhaled anesthesia with an oxygen / fluoromethane mixture and then mice were imaged at regular time intervals with a home-installed mouse-imaging device. Two to eight animals were tested for each sample to ensure reproducibility of the results. The process was carried out under the UK project license "Tumor Targeting" issued to D. Neri (UK PPL 80/1056).

Infrared mouse-imaging devices were fabricated by modifying the photodetection system of Folli et al. (1994), in which infrared fluorophore CY7 can be used. Infrared illumination was chosen for better tissue permeability. Fluorescence (> 760 nm) of CY7 was invisible to the human eye, so it was necessary to use a computer-controlled CCD camera. The mouse-imaging device is equipped with a 100W tungsten halogen lamp and is fitted with a 50mm diameter excitation filter (Chroma Corporation, Bratboro, Vermont, USA; 673-748nm) fixed specifically for CY7. It consists of a blocked black box. As a result, the illumination light evenly illuminates the entire 5x10 cm area in which the mouse to be imaged is placed. Fluorescent, 8-bit with C-mount lens and 50nm emission filter (Chroma Corporation, Brattleboro, Vermont, USA; 765-855nm) and interfaced with the ImageBOX system (Kinetic Imaging, Liverpool, UK) Monochromatic light was detected using a Pulnix CCD-camera. The system consists of a computer equipped with a frame-grabber and software for continuous image capture and integration. Three consecutive images obtained at 50 ms intervals were typically used for the averaging process; This number remained constant for a series of photographs of one animal, allowing direct comparison of tumor targeting at different points. Next, the image in TIFF format was converted to a PICT file using a program graphics converter, and then the image was obtained using a Macintosh 7100/66 computer and a MacDraw Pro program.

A schematic diagram of the design of this device is shown in FIG. 4.

These experiments show that scFv is localized in tumors when visualized to the level of visual observation.

The details at the microscopic level regarding the targeting of neovascularization of developing tumors using two types of anti-EDB scFv are described in detail below.

Nude mice and / or SCID mice with transplanted SKMEL-28 human melanoma or mouse F9 teratoma in one flank were injected with a non-FLAG tag labeled scFv fragment or biotinylated scFv fragment.

After injection, mice were sacrificed at different time points to obtain tumor sections and non-tumor sections, and then subjected to conventional immunohistochemistry protocols using anti-FLAG M2 antibodies (Kodak, 181) or streptavidin-based detection reagents. Stained accordingly. Optimal targeting was generally obtained 12 hours after injection. Both CGS-1 and CGS-2 have been shown to bind to neovascularization in both transplanted tumors in humans and teratocarcinomas in mice.

Example 6-Targeting to F9 teratocarcinoma graft in rats in nude mice

Solid tumors were developed by subcutaneous injection of 4 × 10 ⁶ rat F9 teratocarcinoma cells in one flank of nude mice. The tumor grew so rapidly in mice that, after about a week of injection, it became 1 cm in diameter and significantly formed blood vessels. To image the targeting of the antibodies, a modification to the photodetection method of Folli et al. (1994) was used to make kinetic assessments of tumor targeting and antibody disappearance in the same animal imaged at various time points. This has been described in detail above (see FIG. 4).

For tumor targeting and ease of detection of antibodies, scFv (CGS-1), scFv (CGS-2) and anti-lysozyme scFv (D1.3) (McCafferty et al.) By subcloning the antibody at the Sfil / Not1 site of the expression vector pGIN50. , 1990) has attached a homodimeric tag. The vector is a derivative of pDN268 (Neri et al., 1996b), in which the His6 sequence of the tag is substituted with the sequence: GGC LTD TLQ AFT DQL EDE KSA LQT EIA HLL KEK EKL EFI LAA H, and the cysteine residue and antibody fragment Includes amphiphilic Fos protein helix for covalent homodimerization (Abate et al., 1990). Full covalent dimerization was not achieved: approximately 30-50% of antibody fragments consisted of covalently linked dimers.

On a column obtained by coupling a ovarian zygozyme (D1.3) or 7B89 (anti-ED-B antibody; Carnemolla et al., 1996) and CNBr-activated sepharose (Pharmacia Biotech, Piscataway, NJ, USA), Affinity-chromatography was performed to purify antibody fragments. The supernatant was loaded onto an affinity support, which was washed with PBS, PBS + 0.5M NaCl and eluted with 100mM Et3N. The antibody was then dialyzed against PBS.

After labeling the antibody as described above, 100 μl of a ₁ mg / ml scFv ₁ -Cy7 ₁ solution dissolved in PBS was injected into the tail vein of the tumor-bearing mouse when the diameter of the tumor became approximately 1 cm. Injection.

As shown in FIG. 5, scFv (CGS-1) was localized at the tumor site for 3 days, and of course, antibody loss at the tumor site also occurred rapidly during this period. However, there was some degree of staining of the thighs. Tumor targeting ability of CGS-1 was dramatically improved by introducing amphiphilic helices containing cysteine residues at the C-terminus to promote antibody dimerization (Pack et al., 1993). Indeed, the localization of the dimer scFv (CGS-2) did not seem to decrease significantly during 24 to 72 hours. In contrast, the negative control (dimeric antibody scFv (D1.3) 2, anti-lysozyme antibody) showed rapid disappearance and no detectable localization in the tumor or thigh.

scFv (28SI) showed poor tumor targeting at 6 hours (not shown) but was not detected at all after 24 hours (FIG. 6). Affinity maturation has significantly improved targeting; Thus scFv (CGS-2) efficiently targeted large and small F9 tumors, either monomers (Figure 6) or dimers (not shown). Two days later, the percentage of antibody injected per gram of tumor was 2% for scFv (CGS-2) monomer and 3-4% for scFv (CGS-2) dimer. The amount delivered to the tumor by scFv (CGS-2) was also higher than that of scFv (CGS-1) (Figures 5 and 6), which is related to their respective affinity (Table 1). However, scFv (28SI) and scFv (CGS-2) both tend to be cleaved by proteases and show high liver uptake (FIG. 6), whereas scFv (CGS-1) antibodies are significantly more stable and lower Liver absorption is shown (FIG. 5).

references

Claims (5)

An isolated specific binding element, which is an antibody or antibody fragment comprising an antibody antigen-binding domain, wherein the antibody antigen-binding domain is isolated from a synthetic library and is specific for the ED-B morphogenic domain of fibronectin (FN). Direct binding, comprising the heavy chain (VH) variable region of the sequence derived from human dot line DP47 (including codon 1 Glu-codon 98 Arg in FIG. 1) and the CDR3 sequence Ser Leu Pro Lys Specific binding elements. An isolated specific binding element, which is an antibody or antibody fragment comprising an antibody antigen-binding domain, wherein the antibody antigen-binding domain is isolated from a synthetic library and is specific for the ED-B morphogenic domain of fibronectin (FN). As a direct binding, the heavy chain (VH) variable region of the sequence derived from the human point line DP47 (including codon 1 Glu-codon 98 Arg in FIG. 1) and the CDR3 sequence Gly Val Gly Ala Phe Arg Pro Tyr Arg Lys His A specific binding element comprising Glu. An isolated specific binding element, which is an antibody or antibody fragment comprising an antibody antigen-binding domain, wherein the antibody antigen-binding domain is isolated from a synthetic library and is specific for the ED-B morphogenic domain of fibronectin (FN). Direct binding includes the light chain (VL) variable region of the sequence derived from human dot line DPL16 (including codon 1 Ser-codon 90 Ser in FIG. 1) and the remaining CDR3 sequence Pro Val Val Leu Asn Gly Val Val. Specific binding element, characterized in that. An isolated specific binding element, which is an antibody or antibody fragment comprising an antibody antigen-binding domain, wherein the antibody antigen-binding domain is isolated from a synthetic library and is specific for the ED-B morphogenic domain of fibronectin (FN). Direct binding includes the light chain (VL) variable region of the sequence derived from human dot line DPL16 (including codon 1 Ser-codon 90 Ser of FIG. 1) and the remaining CDR3 sequence Pro Phe Glu His Asn Leu Val Val. Specific binding element, characterized in that. An isolated specific binding element, which is an antibody or antibody fragment comprising an antibody antigen-binding domain, wherein the antibody antigen-binding domain is isolated from a synthetic library and is specific for the ED-B morphogenic domain of fibronectin (FN). A direct binding element comprising a heavy chain (VH) variable region of the sequence derived from human dot line DP47 (including codon 1 Glu-codon 98 Arg in FIG. 1) and a CDR3 sequence .