RU2774414C2 - Cytotoxicity-inducing therapeutic agent - Google Patents

Abstract

FIELD: biotechnology.

SUBSTANCE: proposed is a polypeptide complex exhibiting cell-mediated cytotoxicity, as well as a corresponding vector, cell, method, pharmaceutical composition.

EFFECT: method expands the range of means of treating various types of cancerous tumours.

38 cl, 26 dwg, 3 tbl, 11 ex

Description

The field of technology to which the invention belongs

The present invention relates to polypeptide complexes that allow for the treatment of cancer due to their ability to bring T cells closer to cancer target cells and the use of cytotoxic activity of T cells against cancer targets, to methods for obtaining polypeptide complexes and to therapeutic agents that contain the specified polypeptide complex as an active substance for the induction of cellular cytotoxicity. The present invention also relates to pharmaceutical compositions for the treatment or prevention of various types of cancer, which contain as an active ingredient the above therapeutic agent for inducing cellular cytotoxicity, and to therapeutic methods consisting in the use of these pharmaceutical compositions.

Prerequisites for the creation of the invention

To date, many therapeutic antibodies having very high antitumor activity have been developed as pharmaceuticals for cancer treatment (Non-Patent Document 1). These therapeutic antibodies are known to exert their antitumor effect on cancer cells by inhibiting signals critical for cancer cell growth, inducing cell death signals, antibody dependent cell cytotoxicity (ADCC) or complement dependent cytotoxicity (CDC) (Non-Patent Document 2). ADCC is cytotoxicity induced by effector cells such as NK cells and macrophages against antibody-associated target cancer cells when the Fc region of the antibody binds to the Fc receptor on the effector cells. In this case, the complement complex binds to the complement-fixing site in the antibody structure. CDC is a cytotoxicity that occurs when a component of the complement system in the complex forms a pore in the cell membrane of an antibody-bound cell, increasing the influx of water or ions into the cell. Although conventional therapeutic antibodies have a very high activity, so far the introduction of these antibodies has led only to unsatisfactory outcomes of therapeutic treatment. Thus, there is a need to develop therapeutic antibodies with a higher ability to kill cancer cells.

In addition to the above antibodies that acquire ADCC as their antitumor mechanism through the recruitment of NK cells or macrophages as effector cells, T cell recruitment-inducing antibodies (TR antibodies) which acquire cytotoxicity as their antitumor mechanism through T cell recruitment as effector cells have been known since the 1980s (Non-Patent Documents 3-5). A TR antibody is a bispecific antibody that contains an antibody to any of the subunits that complex the T cell receptor (TCR) on T cells, specifically an antibody that binds to the CD3 epsilon chain and an antibody that binds to the antigen on cancer target cells. The T cell approaches the cancer cell when the TR antibody simultaneously binds to both the CD3 epsilon chain and the cancer antigen, and this results in an antitumor effect on the cancer cell due to the cytotoxic activity of the T cell.

As a TR antibody, an antibody referred to as "trifunctional antibody" is also known (Non-Patent Documents 6 and 7). The trifunctional antibody is a complete IgG-type bispecific antibody in which one arm contains a Fab that binds to the cancer antigen and the other arm contains a Fab that binds to the CD3 epsilon chain. A therapeutic effect on malignant ascites has been demonstrated with the administration of catumaxomab, a trifunctional anti-EpCAM antibody, into the peritoneal cavities of patients with malignant ascites who have EpCAM-positive cancer cells. The use of catumaxomab is approved in the European Union for the above treatment.

In addition, a TR antibody designated as "bispecific T cell-attracting antibody (BiTE)" has now been found to have a strong antitumor effect (Non-Patent Documents 8 and 9). BiTE is a TR antibody with a molecular form in which an anti-cancer antigen scFv is linked to an anti-CD3 chain epsilon scFv via a short polypeptide linker. BiTEs are known to have an antitumor activity that is superior to that of various known TR antibodies (Non-Patent Documents 9 and 10). In particular, compared to other TR antibodies, BiTEs exhibit antitumor activity even when used at a significantly lower concentration and at a lower effector/cancer cell ratio (ET ratio). It has also been demonstrated that their action can be successfully exerted without the need for pre-activation of effector cells by IL-2, an anti-CD28 agonist antibody, or other. cells in vitro than rituxan, which is known to have a very good clinical effect. In addition, extremely high activity of blinatumomab has been found in Phase I and II clinical trials currently underway (Non-Patent Document 11).

Based on the fact that catumaxomab is approved as a clinically demonstrated therapeutic agent and that many BiTEs, including blinatumomab, have strong antitumor effects, it has been hypothesized that TR antibodies that mediate T- cells as effector cells may have a significantly higher potential as an antitumor agent compared to conventional antibodies, the mechanism of action of which is associated with ADCC.

However, it is known that a trifunctional antibody simultaneously binds to both a T cell and a cell such as a NK cell or a macrophage in a cancer antigen-independent manner, and as a result, the receptors that are expressed on the cells are cross-linked, and the expression of various cytokines is induced independently. from the cancer antigen. Systemic administration of the trifunctional antibody is expected to cause "cytokine storm" type side effects as a result of said induction of cytokine expression. In fact, in a phase I clinical trial, it was found that the maximum tolerated dose for systemic administration of catumaxomab to patients with non-small cell lung cancer is a very low dose of 5 μg/body, and that administration of a higher dose causes serious side effects (Non-Patent Document 12). With the introduction of catumaxomab at this low dose, its blood level can never reach an effective value. This means that the expected antitumor effect cannot be achieved with the administration of catumaxomab at this low dose.

At the same time, unlike catumaxomab, BiTE does not contain an Fcγ receptor binding site and, therefore, such an antibody does not cross-link with receptors that are expressed on T cells and cells such as NK cells and macrophages in a cancer antigen-dependent manner. Thus, it was demonstrated that BiTE does not cause cancer-antigen-independent cytokine induction, which is found with the introduction of catumaxomab. However, since BiTE is a modified small molecule antibody without an Fc region, the problem is that its blood half-life after administration to a patient is significantly shorter than that of IgG-type antibodies that are commonly used as therapeutic antibodies. In fact, according to published data, the blood half-life of BiTE when used in vivo is about a few hours (Non-Patent Documents 13 and 14). In clinical trials, blinatumomab was administered by continuous intravenous infusion using a minipump. Such an administration method is not only extremely inconvenient for patients, but also has the potential risk of medical complications due to device malfunction or the like. Thus, this method of administration cannot be considered desirable.

Documents characterizing the prior art

[Non-Patent Documents]

[Non-Patent Document 1]: Clin Cancer Res. 16(1), 2010, pp. 11-20.

[Non-Patent Document 2]: Drug Des Devel Ther 3, 2009, ss. 7-16.

[Non-Patent Document 3]: Nature 314 (6012), 1985, ss. 628-631.

[Non-Patent Document 4]: Int J Cancer 41 (4), 1988, ss. 609-615.

[Non-Patent Document 5]: Proc Natl Acad Sci USA 83 (5), 1986, ss. 1453-1457.

[Non-Patent Document 6]: Cancer Treat Rev. 36 (6), 2010, p. 458-467.

[Non-Patent Document 7]: Expert Opin Biol Ther 10 (8), 2010, ss. 1259-1269.

[Non-Patent Document 8]: Proc Natl Acad Sci USA. 92 (15), 1995, p. 7021-7025.

[Non-Patent Document 9]: Drug Discov Today, 10 (18), 2005, ss. 1237-1244.

[Non-Patent Document 10]: Trends Biotechnol 22 (5), 2004, ss. 238-244.

[Non-Patent Document 11]: Science 321 (5891), 2008, ss. 974-977.

[Non-Patent Document 12]: Cancer Immunol Immunother 56 (10), 2007, ss. 1637-1644.

[Non-Patent Document 13]: Cancer Immunol Immunother. 55 (5), 2006, pp. 503-514.

[Non-Patent Document 14]: Cancer Immunol Immunother. 58(1), 2009, pp. 95-109.

Brief summary of the invention

Objectives underlying the present invention

The present invention has been made in view of the above circumstances. The present invention was based on the task of creating polypeptide complexes suitable for cancer treatment, which have the ability to bring T cells closer to a cancer target cell and allow the use of cytotoxic activity of T cells against cancer target cells, methods for obtaining polypeptide complexes and therapeutic agents, which contain the specified polypeptide complex as an active substance for the induction of cellular cytotoxicity. Another object of the present invention is the development of pharmaceutical compositions for the treatment or prevention of various types of cancer, which contain as an active ingredient the above therapeutic agent for induction of cytotoxicity, and therapeutic methods consisting in the use of these pharmaceutical compositions.

Means for solving these problems

In the course of the invention, new polypeptide complexes have been discovered that retain the strong antitumor activity inherent in BiTE and have a long blood half-life, as well as very high safety, resulting in no induction of a cancer antigen-independent "cytokine storm" or the like. The present invention has also established that by replacing the antigen-binding domains of the polypeptide complexes, the polypeptide complexes can damage various target cells. Based on the above data, the present invention has demonstrated that the polypeptide complexes of the present invention damage cancer cells. During the creation of the present invention, it was also found that by regulating the association at the CH1/CL interface and introducing the protrusion-into-depression ( Knobs-into- Holes ( KiH )) modifications into polypeptide complexes achieve more efficient cellular cytotoxicity. In addition, the present invention has demonstrated that various cancers can be treated or prevented by using therapeutic agents for inducing cellular cytotoxicity that contain the polypeptide complex of the present invention as an active ingredient.

More specifically, the present invention provides:

[1] a polypeptide complex that contains:

(1) an antigen-binding domain;

(2) a domain containing an Fc region that has reduced Fcγ receptor binding activity; and

(3) a T-cell receptor complex binding domain;

[2] the polypeptide complex according to [1], wherein the T cell receptor complex binding domain is a T cell receptor binding domain;

[3] the polypeptide complex according to [1], wherein the T-cell receptor complex-binding domain is a CD3-binding domain;

[4] polypeptide complex according to one of paragraphs. [1]-[3], in which the antigen-binding domain is a bivalent antigen-binding domain;

[5] the polypeptide complex according to [4], in which the bivalent antigen-binding domain is a domain having the structure F(ab')2;

[6] the polypeptide complex according to [5], in which two polypeptides forming the heavy chain constant region of the domain, which has the structure F(ab')2, are individually linked to any of the two polypeptides forming the Fc domain;

[7] the polypeptide complex according to [6], wherein the CD3-binding domain is linked to either or both of the CH3s forming the Fc domain;

[8] the polypeptide complex according to [7], wherein the heavy chain Fv fragment forming the CD3 binding domain is linked to one of the CH3 forming the Fc domain, and the light chain Fv fragment forming the CD3 binding domain, linked to another CH3 forming an Fc domain;

[9] the polypeptide complex according to [8], in which the CH1 domain of the antibody is linked to the Fv fragment of the heavy chain, which forms the CD3 binding domain, and the CL domain of the antibody is linked to the Fv fragment of the light chain;

[10] the polypeptide complex according to [6], in which the CD3-binding domain is linked to any of two or both CLs forming F(ab')2;

[11] the polypeptide complex according to [6], in which the CD3-binding domain is linked to either or both VHs forming F(ab')2;

[12] the polypeptide complex according to [6], in which the CD3-binding domain is linked to either or both VLs forming F(ab')2;

[13] polypeptide complex according to one of p.p. [1]-[12], in which the CD3-binding domain is Fv;

[14] polypeptide complex according to one of p.p. [1]-[7] and [10]-[12], wherein the CD3 binding domain is a Fab;

[15] a polypeptide complex according to one of paragraphs. [1]-[7] and [10]-[12], wherein the CD3 binding domain is scFv;

[16] polypeptide complex according to one of p.p. [1]-[15], in which the CD3-binding domain is univalent;

[17] polypeptide complex according to one of p.p. [1]-[3], in which the antigen-binding domain is a monovalent scFv and a monovalent Fab;

[18] the polypeptide complex according to [17], in which the monovalent scFv is linked to one of the polypeptides that form the Fc domain through the scFv that forms the CD3-binding domain; The heavy chain Fv fragment of the monovalent Fab is linked to one of the Fc domain-forming polypeptides via a CH1 domain; and the Fv light chain fragment of Fab is linked to the CL domain;

[19] polypeptide complex according to one of p.p. [1]-[3], in which the antigen-binding domain is a divalent scFv;

[20] a polypeptide complex according to [19], in which one univalent scFv is linked to one of the polypeptides that form the Fc domain through the Fv fragment of the heavy chain that forms the CD3-binding domain, and the other univalent scFv is linked to another polypeptide, forming an Fc domain, through the Fv fragment of the light chain, forming a CD3-binding domain;

[21] a polypeptide complex according to [19], in which one univalent scFv is linked to one of the polypeptides that form the Fc domain through the scFv that forms the CD3 binding domain, and the other univalent scFv is linked to another polypeptide that forms the Fc domain ;

[22] a polypeptide complex according to one of paragraphs. [1]-[3], wherein the antigen-binding domain and the T-cell receptor complex-binding domain are each a monovalent Fab;

[23] the polypeptide complex according to [22], in which the Fv fragment of the heavy chain of a monovalent Fab forming an antigen-binding domain is linked to one of the polypeptides forming the Fc domain through the CH1 domain and the Fv fragment of the light chain of the Fab is linked to CL domain; and the heavy chain Fv fragment of a Fab forming a T cell receptor binding domain is linked to another Fc domain forming polypeptide via a CH1 domain, and the light chain Fv fragment of the Fab is linked to a CL domain;

[24] a polypeptide complex according to [22], in which the Fv fragment of a monovalent Fab that forms an antigen-binding domain is linked to one of the polypeptides that form an Fc domain through the CH1 domain and the Fv fragment of the Fab light chain is linked to CL- domain; and the light chain Fv fragment of a Fab forming a T cell receptor binding domain is linked to another Fc domain forming polypeptide via a CH1 domain, and the heavy chain Fv fragment of the Fab is linked to a CL domain;

[25] a polypeptide complex according to [22], in which the Fv fragment of the heavy chain of a monovalent Fab forming an antigen-binding domain is linked to one of the polypeptides forming the Fc domain through the CH1 domain and the Fv fragment of the light chain of the Fab is linked to CL domain; and the heavy chain Fv fragment of a Fab forming a T cell receptor binding domain is linked to another Fc domain forming polypeptide via a CL domain, and the light chain Fv fragment of the Fab is linked to a CH1 domain;

[26] a polypeptide complex according to [22], in which the Fv-fragment of the heavy chain of a monovalent Fab, which forms a T-cell receptor-binding domain, is linked to one of the polypeptides that form an Fc-domain through a CH1-domain and an Fv-fragment the light chain Fab is linked to the CL domain; and the Fv light chain fragment of the Fab forming the antigen binding domain is linked to another polypeptide forming the Fc domain via the CH1 domain, and the Fv heavy chain fragment of the Fab is linked to the CL domain;

[27] the polypeptide complex according to [22], in which the Fv fragment of the heavy chain of a monovalent Fab that forms a T-cell receptor-binding domain is linked to one of the polypeptides that form an Fc domain via a CH1 domain; and the Fv light chain fragment of Fab is linked to the CL domain; and a heavy chain Fv fragment of a Fab forming an antigen binding domain is linked to another polypeptide forming an Fc domain through a CL domain, and a light chain Fv fragment of a Fab is linked to a CH1 domain;

[28] polypeptide complex according to item [22] which contains:

(1) an antigen-binding domain in which a heavy chain Fv fragment of a monovalent Fab structure that binds an antigen is linked through a CH1 domain to one of the polypeptides constituting an Fc domain, and a light chain Fv fragment of the Fab structure is linked to a CL domain ; and

(2) a T-cell receptor complex-binding domain in which the heavy chain Fv fragment of the univalent Fab structure that binds to the T-cell receptor complex is linked via CH1 to another Fc domain-forming polypeptide and the light chain Fv fragment structure Fab is linked to the CL domain;

while the electrical charges of the CH1 and CL domains are controlled so that the Fv fragment of the heavy chain of the antigen-binding domain is connected to the Fv fragment of the light chain of the antigen-binding domain or the Fv fragment of the heavy chain of the T-cell receptor binding domain is connected to the Fv fragment light chain of the T-cell receptor binding domain;

[29] the polypeptide complex according to [28], in which the amino acid residue in the CH1 domain linked to the Fv fragment of the heavy chain of the T-cell receptor complex-binding domain has the same type of electric charge as the amino acid residue in the CL- a domain linked to an Fv fragment of an antigen-binding domain;

[30] the polypeptide complex according to [28], in which the amino acid residue in the CH1 domain linked to the Fv fragment of the heavy chain of the antigen-binding domain has the same type of electrical charge as the amino acid residue in the CL domain linked to the Fv- a fragment of the light chain of the domain that binds the T-cell receptor complex;

[31] a polypeptide complex according to [28], in which the amino acid residue in the CH1 domain linked to the Fv fragment of the heavy chain of the T-cell receptor complex-binding domain has the same type of electric charge as the amino acid in the CL domain linked with the Fv fragment of the light chain of the antigen binding domain, and the amino acid residue in the CH1 domain linked to the Fv fragment of the heavy chain of the antigen binding domain has the same type of electrical charge as the amino acid in the CL domain linked to the Fv fragment of the light chain of the domain, binding complex T-cell receptor;

[32] a polypeptide complex according to [29] or [31], in which the amino acid residue in the CH1 domain linked to the Fv fragment of the heavy chain of the T-cell receptor complex-binding domain has an electrical charge opposite to that of the amino acid a residue in the CL domain linked to the Fv light chain fragment of the T cell receptor binding domain;

[33] a polypeptide complex according to [30] or [31], in which the amino acid residue in the CH1 domain linked to the Fv fragment of the heavy chain of the antigen-binding domain has an electrical charge opposite to the charge of the amino acid residue in the CL domain, linked to the Fv-fragment of the light chain of the antigen-binding domain;

[34] a polypeptide complex according to one of paragraphs. [22]-[33], wherein the T cell receptor complex binding domain is a T cell receptor binding domain;

[35] the polypeptide complex according to [34], in which the T-cell receptor-binding domain is a CD3-binding domain;

[36] a polypeptide complex according to [32] or [33], wherein the amino acid residues in the CH1 and CL domains are one, two or more combinations of amino acid residues selected from the group consisting of combinations of:

(a) amino acid residue at position 147 (EU numbering) in the CH1 domain and amino acid residue at position 180 (EU numbering) in the CL domain;

(b) amino acid residue at position 147 (EU numbering) in the CH1 domain and amino acid residue at position 131 (EU numbering) in the CL domain;

(c) amino acid residue at position 147 (EU numbering) in the CH1 domain and amino acid residue at position 164 (EU numbering) in the CL domain;

(d) amino acid residue at position 147 (EU numbering) in the CH1 domain and amino acid residue at position 138 (EU numbering) in the CL domain;

(e) amino acid residue at position 147 (EU numbering) in the CH1 domain and amino acid residue at position 123 (EU numbering) in the CL domain; and

(e) amino acid residue at position 175 (EU numbering) in the CH1 domain and amino acid residue at position 160 (EU numbering) in the CL domain;

and in which the amino acid residue in the CH1 domain has an electrical charge opposite to that of the amino acid residue in the CL domain;

[37] the polypeptide complex according to [36], in which the amino acid residues are selected from the group, which additionally includes a combination of amino acid residues:

(g) amino acid residue at position 213 (EU numbering) in the CH1 domain and amino acid residue at position 123 (EU numbering) in the CL domain;

[38] a polypeptide complex according to [36] or [37], in which the amino acid residue having the opposite electrical charge is selected from an amino acid residue included in any of the following groups:

(X) glutamic acid (E) and aspartic acid (D); or

(Y) lysine (K), arginine (R) and histidine (H);

[39] a polypeptide complex according to one of paragraphs. [36]-[38], in which amino acid residues having an opposite electric charge are:

Lys at position 175 (EU numbering) in the CH1 domain and Glu at positions 131, 160 and 180 (EU numbering) in the CL domain;

[40] a polypeptide complex according to one of paragraphs. [36]-[38], in which amino acid residues having an opposite electric charge are:

Glu at positions 147 and 175 (EU numbering) in the CH1 domain and Lys at positions 131, 160 and 180 (EU numbering) in the CL domain;

[41] the polypeptide complex according to [40], wherein the amino acid residue at position 213 (EU numbering) in the CH1 domain is Glu and the amino acid residue at position 123 (EU numbering) in the CL domain is Lys;

[42] a polypeptide complex according to one of paragraphs. [1]-[41], in which the Fc domain is characterized by reduced activity of binding to the Fcγ receptor against FcγI, FcγIIA, FcγIIB, FcγIIIA and/or FcγIIIB;

[43] a polypeptide complex according to one of paragraphs. [1]-[42], in which the Fc domain is any of the Fc domains having SEQ ID NOs: 23, 24, 25 and 26, in which the amino acid(s) forming(s) the Fc domain is changed (s) as a result of a mutation;

[44] the polypeptide complex according to [43], in which the Fc domain contains any of the following amino acids:

amino acid sequence extending from position 118 to position 260 (EU numbering), which is the sequence of SEQ ID NO: 24; or

amino acid sequence extending from position 261 to position 447 (EU numbering), which is the sequence of SEQ ID NO: 26;

[45] the polypeptide complex according to [43], in which the amino acids forming the Fc domain contain a mutation in any of the following positions:

220, 226, 229, 231, 232, 233, 234, 235, 236, 237, 238, 239, 240, 264, 265, 266, 267, 269, 270, 295, 296, 297, 298, 299, 300, 325, 327, 328, 329, 330, 331 and 332 (EU numbering);

[46] the polypeptide complex according to [45], in which the Fc domain contains a mutation in the amino acids of SEQ ID NO: 23, forming the Fc domain;

[47] a polypeptide complex according to [46], in which the Fc domain is an Fc domain containing an amino acid substitution at position 233, 234, 235, 236, 237, 327, 330, or 331 (EU numbering) with an amino acid present at the corresponding position (EU numbering) in the corresponding IgG2 or IgG4;

[48] the polypeptide complex according to [46], in which the Fc domain contains an amino acid mutation at position 234, 235 or 297 (EU numbering);

[49] the polypeptide complex according to [48], in which the amino acid(s) at position 234, 235 and/or 297 is replaced(s) by alanine;

[50] a polypeptide complex according to one of paragraphs. [43]-[49], in which the sequences of the two polypeptides that form the Fc domain differ from each other;

[51] a polypeptide complex according to one of paragraphs. [1]-[50], in which the amino acid at position 349 is replaced by cysteine and the amino acid at position 366 (EU numbering) is replaced by tryptophan in the sequence of amino acid residues of one of the two polypeptides forming the Fc domain; and wherein the amino acid at position 356 is changed to cysteine, the amino acid at position 366 is changed to serine, the amino acid at position 368 is changed to alanine, and the amino acid at position 407 (EU numbering) is changed to valine in the amino acid residue sequence of another polypeptide;

[52] a polypeptide complex according to one of paragraphs. [1]-[50], in which the amino acid at position 356 (EU numbering) is replaced by lysine in the sequence of amino acid residues of one of the two polypeptides that form the Fc domain; the amino acid at position 439 (EU numbering) is changed to glutamic acid in another polypeptide; and the amino acid at position 435 (EU numbering) is changed to arginine in the amino acid sequence of either of the two polypeptides;

[53] a polypeptide complex according to [51] or [52], in which the GK sequence is removed by deletion from the carboxyl ends of the two polypeptides that form the Fc domain;

[54] a polypeptide complex according to one of paragraphs. [1]-[53], in which the antigen-binding domains bind to the same epitope;

[55] the polypeptide complex according to [54], where the specified epitope is present in a protein containing the amino acid sequence of SEQ ID NO: 2;

[56] the polypeptide complex according to [54], where the specified epitope is present in a protein containing the amino acid sequence of SEQ ID NO: 4;

[57] a polypeptide complex according to one of paragraphs. [1] to [53], in which each of the antigen-binding domains binds to a different (different) epitope;

[58] the polypeptide complex according to [57], where the specified other epitope is present in a protein containing the amino acid sequence of SEQ ID NO: 2;

[59] the polypeptide complex according to [57], where another epitope is present in a protein containing the amino acid sequence of SEQ ID NO: 4;

[60] a polynucleotide encoding a polypeptide complex according to one of paragraphs. [1]-[59];

[61] a vector containing a polynucleotide according to [60];

[62] a cell containing a vector according to [61];

[63] a method for obtaining a polypeptide complex, which consists in the fact that the cell is cultivated according to paragraph [62] and the polypeptide complex is isolated from the culture supernatant;

[64] a therapeutic agent for the induction of cellular cytotoxicity, which contains as an active substance a polypeptide complex according to one of paragraphs. [1]-[59];

[65] the therapeutic agent of [64], wherein the therapeutic agent for inducing cellular cytotoxicity is a therapeutic agent used in cancer;

[66] the therapeutic agent of [65], wherein the cancer is liver cancer or lung cancer;

[67] a method for the treatment or prevention of cancer, which consists in the fact that the polypeptide complex according to one of paragraphs. [1]-[59] are administered to a patient in need thereof; and

[68] the therapeutic or prophylactic method of [67], wherein the cancer is liver cancer or lung cancer.

The present invention also relates to kits for use in the method of the present invention, which contain a polypeptide complex of the present invention or a polypeptide complex obtained using the method of the present invention. The present invention also relates to the use of the polypeptide complex of the present invention or the polypeptide complex obtained by the method of the present invention for the preparation of a therapeutic agent for inducing cellular cytotoxicity. The present invention also relates to polypeptide complexes of the present invention, or polypeptide complexes obtained using the methods of the invention, for use in the method of the present invention.

Results of the invention

The present invention provides novel polypeptide complexes that retain the strong antitumor activity of BiTE and have a long blood half-life as well as very high safety, resulting in no induction of a cancer antigen-independent "cytokine storm" or the like. When the antigen-binding domain of the polypeptide complex of the present invention is changed, therapeutic agents that contain the polypeptide complex as an active ingredient to induce cellular cytotoxicity can target and damage various cells, including cancer cells. Thus, various types of cancer can be treated or prevented. This provides the desired treatment that is highly safe and comfortable and reduces the physical burden on patients.

Brief description of the drawings

The drawings show:

in fig. 1 is a graph comparing the cytotoxic activity of the following anti-GPC3 antibodies (an antigen belonging to the GPI-anchored receptor family): GPC3 ERY1 (GPC3 BiTE), GPC3 ERY2, and anti-GPC3 IgG-type antibodies. Solid square (■), solid triangle (▲), and open square (□) indicate the cytotoxic activity of GPC3 ERY1 (GPC3 BiTE), GPC3 ERY2, and anti-GPC3 IgG-type antibodies, respectively;

in fig. 2 is a graph showing a comparison of the cytotoxic activity of GPC3 BiTE and GPC3 ERY5. Solid square (■) and open circle (○) indicate cytotoxic activity of GPC3 BiTE and GPC3 ERY5, respectively;

in fig. 3 is a graph that compares the cytotoxic activity of GPC3 BiTE and GPC3 ERY6. Solid square (■) and solid triangle (▲) indicate the cytotoxic activity of GPC3 BiTE and GPC3 ERY6, respectively;

in fig. 4 is a graph that compares the cytotoxic activity of GPC3 BiTE and GPC3 ERY7. Solid square (■) and solid diamond (♦) indicate the cytotoxic activity of GPC3 BiTE and GPC3 ERY7, respectively;

in fig. 5 is a graph comparing the cytotoxic activity of GPC3 BiTE, GPC3 ERY8-2, GPC3 ERY9-1 and GPC3 ERY10-1. Solid square (■), solid triangle (▲), open circle (○), and open square (□) indicate the cytotoxic activity of GPC3 BiTE, GPC3 ERY8-2, GPC3 ERY9-1, and GPC3 ERY10-1, respectively;

in fig. 6 is a graph showing the in vivo antitumor activity of GPC3 ERY8-2 in a pre-mix PC-10 model. The open square (□) and the filled diamond (♦) indicate changes in tumor volume in the GPC3 ERY7 treated group and the control (PBS treated) group, respectively;

in fig. 7 is a graph showing the in vivo antitumor activity of GPC3 ERY10-1 in a pre-mix PC-10 model. Open square (□) and solid diamond (♦) indicate changes in tumor volume in the GPC3 ERY10-1 treated group and control (PBS treated) group, respectively;

in fig. 8 is a graph demonstrating the in vivo antitumor effect of GPC3 ERY10-1 in a PC-10 model of T-cell transfer. Open square (□) and solid diamond (♦) indicate changes in tumor volume in the GPC3 ERY10-1 treated group and control (PBS treated) group, respectively;

in fig. 9 is a graph showing the time course of plasma concentrations of GPC3 ERY9-1 and GPC3 ERY10-1 determined using GPC3-expressing Ba/F3 cells. Solid diamond (♦) and open square (□) denote the time dependence of plasma concentrations of GPC3 ERY9-1 and GPC3 ERY10-1, respectively;

in fig. 10 is a graph showing the time course of plasma concentrations of GPC3 ERY9-1 and GPC3 ERY10-1 determined using CD3-expressing Ba/F3 cells. Solid diamond (♦) and open square (□) denote the time dependence of plasma concentrations of GPC3 ERY9-1 and GPC3 ERY10-1, respectively;

in fig. 11 is a graph showing the ability of GPC3 BiTE, GPC3 ERY9-1, GPC3 ERY10-1, GPC3 ERY15-1 and catumaxomab to induce cytokines in a cancer antigen-independent manner;

in fig. 12 is a graph showing the in vitro cytotoxicity of GPC3 ERY18 L1, GPC3 ERY18L2, GPC3 ERY18L3, GPC3 ERY18L4 and GPC3 ERY18S1. Solid triangle (▲), solid circle (●), solid square (■), open square (□) and open diamond (◊) indicate the cytotoxic activity of GPC3 ERY18 L1, GPC3 ERY18 L2, GPC3 ERY18 L3, GPC3 ERY18 L4 and GPC3 ERY18 S1 respectively;

in fig. 13 is a graph comparing the in vitro cytotoxic activity of GPC3 ERY18 L3 and GPC3 ERY10-1. Solid square (■) and open square (□) indicate the cytotoxic activity of GPC3 ERY18 L3 and GPC3 ERY10-1, respectively;

in fig. 14 is a graph comparing the in vitro cytotoxic activity of GPC3 ERY19-3 and GPC3 BiTE. Open square (□) and solid square (■) indicate the cytotoxic activity of GPC3 ERY19-3 and GPC3 BiTE, respectively;

in fig. 15A is a chromatogram showing the results of a CM gel filtration analysis in which NTA1L/NTA1R/GC33-k0 is expressed. In FIG. 15B is a chromatogram showing the results of a CM gel filtration analysis in which NTA2L/NTA2R/GC33-k0 is expressed;

in fig. 16 - diagrams showing the domains that form the following polypeptide complexes described in the examples that are presented in this description: GPC3 BiTE, GPC3 ERY2, GPC3 ERY5, GPC3 ERY6, GPC3 ERY7, GPC3 ERY8-2, GPC3 ERY9-1, GPC3 ERY 10-1, GPC3 ERY15, GPC3 ERY18 and GPC3 ERY19-3. The cross-hatched domain denotes the variable region of the H chain of an anti-cancer antigen antibody (GPC3, EpCAM, EGFR); the domain with diagonal lines denotes the variable region of the L chain of an antibody to a cancer antigen (GPC3, EpCAM, EGFR); the dotted-line domain denotes the variable region of the H chain of an anti-CD3 antibody; the filled domain denotes the variable region of the L chain of the anti-CD3 antibody; open domain denotes the constant region of the antibody; a cross denotes a silent mutation in Fc and an asterisk denotes a mutation that enhances the heteromeric association of Fc;

in fig. 17 - diagrams of GPC3 BiTE (A); GPC3 ERY 10 (B); GPC3 ERY2 (B); GPC3 ERY5 (G); GPC3 ERY6 (D); GPC3 ERY7 (E); GPC3 ERY8-2 (W); GPC3 ERY9-1 (D); GPC3 ERY10-1 (I); GPC3 ERY15 (K); GPC3 ERY18 (L) and GPC3 ERY19-3 (M);

in fig. 18 amino acid residues forming the Fc domains of IgG1, IgG2, IgG3 and IgG4 and their EU numbering according to Cabot (in the context of this description, also referred to as "EU index");

in fig. 19 are diagrams showing the domains forming the polypeptide complexes described in the examples provided herein: GPC3 ERY17-2, GPC3 ERY17-3, EpCAM ERY17-2 and EpCAM ERY17-3. The cross-hatched domain denotes the variable region of the H chain of an anti-cancer antigen antibody (GPC3, EpCAM, EGFR); the domain with diagonal lines denotes the variable region of the L chain of an antibody to a cancer antigen (GPC3, EpCAM, EGFR); the dotted-line domain denotes the variable region of the H chain of an anti-CD3 antibody; the filled domain denotes the variable region of the L chain of the anti-CD3 antibody; open domain denotes the constant region of the antibody; a cross denotes a silent mutation in Fc and an asterisk denotes a mutation that enhances the heteromeric association of Fc;

in fig. 20 is a graph comparing the cytotoxic activity of GPC3 BiTE, GPC3 ERY17-2, GPC3 ERY17-3 and GPC3 ERY10-1. Solid square (■), solid triangle (▲), open circle (○), and open square (□) indicate the cytotoxic activity of GPC3 BiTE, GPC3 ERY17-2, GPC3 ERY17-3, and GPC3 ERY10-1, respectively;

in fig. 21 is a graph demonstrating the in vivo antitumor activity of GPC3 ERY17-2 in a PC-10 model of T cell transfer. Open square (□) and solid diamond (♦) indicate changes in tumor volume in the GPC3 ERY17-2 treated group and control (PBS treated) group, respectively;

in fig. 22 is a graph comparing the cytotoxic activity of GPC3 ERY17-2 and GPC3 ERY17-2-M20. The solid triangle (▲) and open circle (○) indicate the cytotoxic activity of GPC3 ERY17-2 and GPC3 ERY17-2-M20, respectively;

in fig. 23 is a graph comparing the cytotoxic activity of EpCAM ERY17-2 and EpCAM ERY17-3. The solid triangle (▲) and open square (□) indicate the cytotoxic activity of EpCAM ERY17-2 and EpCAM ERY17-3, respectively;

in fig. 24 - diagrams showing the domains that form the polypeptide complexes described in the examples that are presented in the present description: GM1, GM2 and GM0. "A" shows a polypeptide complex in which association at the CH1/CL interface is regulated and "knob-hole" (KiH) modifications are introduced. "B" shows a polypeptide complex in which association at the CH1/CL interface is not regulated or KiH modifications are not introduced. The cross-hatched domain denotes the variable region of the H chain of an anti-cancer antigen antibody (GPC3, EpCAM, EGFR); the domain with diagonal lines denotes the variable region of the L chain of an antibody to a cancer antigen (GPC3, EpCAM, EGFR); the dotted-line domain denotes the variable region of the H chain of an anti-CD3 antibody; the filled domain denotes the variable region of the L chain of the anti-CD3 antibody; open domain denotes the constant region of the antibody; a cross denotes a silent mutation in Fc and an asterisk denotes a mutation that enhances the heteromeric association of Fc;

and the donut-shaped symbol denotes a mutation governing interaction at the CH1/CL interface;

in fig. 25 is a graph comparing the cytotoxic activity of GM1, GM2 and GM0. The solid triangle (▲), open square (□), and open circle (○) indicate the cytotoxic activity of GM1, GM2, and GM3, respectively;

in fig. 26 is a graph showing the cytotoxic activity of EGFR ERY17-2. The solid triangle (▲) indicates the cytotoxic activity of EGFR ERY17-2.

Embodiments of the invention

The following definitions are given for the purpose of explaining the present invention.

Antibody

In the context of the present description, "antibody" refers to a naturally occurring immunoglobulin or an immunoglobulin obtained in whole or in part by synthesis. Antibodies can be isolated from naturally occurring sources, such as naturally occurring plasma and serum, or from culture supernatants of antibody-producing hybridomas. Alternatively, antibodies can be partially or completely synthesized using techniques such as genetic recombination. Preferred antibodies are, for example, antibodies belonging to any immunoglobulin isotype or subclass thereof. Known human immunoglobulins include antibodies of the following nine classes (isotypes): IgG1, IgG2, IgG3, IgG4, IgA1, IgA2, IgD, IgE, and IgM. Of these isotypes, the antibodies of the invention include IgG1, IgG2, IgG3 and IgG4.

Methods for obtaining an antibody with the desired binding activity are known to those skilled in the art. The following is an example that describes a method for producing an antibody (anti-GPC3 antibody) that binds to GPC3, which belongs to the family of GPI-anchored receptors (Int J Cancer. 103(4), 2003, pp. 455-465). According to the example described below, it is also possible to obtain antibodies that bind to an antigen other than GPC3.

Anti-GPC3 antibodies can be obtained as polyclonal or monoclonal antibodies using known methods. Anti-GPC3 antibodies are preferably obtained as monoclonal antibodies derived from mammalian organisms. Said mammalian-derived antibodies include antibodies produced by hybridomas or host cells transformed with an expression vector that carries the antibody gene using genetic engineering techniques.

Hybridomas producing monoclonal antibodies can be obtained using known methods, such as those described below. In particular, mammals are immunized using conventional immunization methods using the GPC3 protein as a sensitizing antigen. The resulting immune cells are fused with known parental cells using conventional fusion techniques. Anti-GPC3 antibody producing hybridomas can then be selected by screening for monoclonal antibody producing cells using conventional screening methods.

In particular, monoclonal antibodies are obtained according to the method described below. First, the GPC3 gene, the nucleotide sequence of which is presented in RefSeq under accession number NM_001164617.1 (SEQ ID NO: 1), can be expressed to obtain a GPC3 protein, the sequence of which is presented in RefSeq under accession number NP_001158089.1 (SEQ ID NO: 2), which can be used as a sensitizing antigen to produce an antibody. To do this, the gene sequence encoding GPC3 is inserted into a known expression vector and the appropriate host cells are transformed with this vector. The desired human GPC3 protein is purified from host cells or culture supernatants using known methods. For example, to obtain soluble GPC3 from culture supernatants, it is removed by deleting the amino acid at positions 564-580, which form the hydrophobic region corresponding to the GPI anchoring sequence used to anchor GPC3 to the cell membrane, from the GPC3 polypeptide sequence of SEQ ID NO: 2, and then the resulting protein is expressed in place of the GPC3 protein having SEQ ID NO: 2. Alternatively, a purified naturally occurring GPC3 protein can be used as the sensitizing antigen.

The purified GPC3 protein can be used as a sensitizing antigen for mammalian immunization. An incomplete GPC3 peptide can also be used as a sensitizing antigen. In this case, the partial peptide can be produced by chemical synthesis based on the human GPC3 amino acid sequence or by inserting a portion of the GPC3 gene into an expression vector. Alternatively, the partial peptide can be obtained by cleaving the GPC3 protein with a protease. Particular embodiments of the invention do not impose restrictions on the length and area of the incomplete GPC3 peptide. The preferred region can be arbitrarily selected from among the amino acid residues of the amino acid sequence located at positions 564-580 in the amino acid sequence of SEQ ID NO: 2. The number of amino acids forming a peptide that can be used as a sensitizing agent is preferably at least five or more, six or more or seven or more. More specifically, a peptide consisting of 8-50 residues, more preferably 10-30 residues, can be used as the sensitizing antigen.

Alternatively, a fusion protein obtained by fusing the desired partial polypeptide or GPC3 protein peptide with another polypeptide can be used as the sensitizing antigen. For example, antibody Fc fragments and peptide tags are preferably used to produce fusion proteins for use as sensitizing antigens. Vectors for the expression of these fusion proteins can be constructed by fusing in frame the genes encoding two or more of the desired polypeptide fragments and inserting the fusion gene into the expression vector described above. Methods for making fusion proteins are described in Molecular Cloning, 2nd ed. (Sambrook J. et al., Molecular Cloning, 2nd ed. 1989, 9.47-9.58, Cold Spring Harbor Lab. Press). Methods for producing GPC3 for use as a sensitizing antigen and immunization methods using GPC3 are specifically described in WO 2003/000883, WO 2004/022754 and WO 2006/006693.

There is no particular limitation regarding mammals to be immunized with the sensitizing antigen. However, it is preferable to select mammals based on their compatibility with the parental cells used for cell fusion. In general, rodents such as mice, rats, as well as hamsters, rabbits, and monkeys are preferably used.

The above animals are immunized with the sensitizing antigen using known methods. Conventional methods of immunization are, for example, intraperitoneal or subcutaneous injection of a sensitizing antigen into mammals. In particular, the sensitizing antigen can be appropriately diluted in PBS (phosphate buffered saline), physiological saline or the like. If necessary, a conventional adjuvant such as Freund's complete adjuvant is mixed with the antigen and the mixture is emulsified. The sensitizing antigen is then administered to the mammal several times at 4-21 day intervals. When immunized with a sensitizing antigen, appropriate carriers may be used. In particular, when a low molecular weight partial peptide is used as the sensitizing antigen, it is sometimes required for immunization to combine the peptide representing the sensitizing antigen with a carrier protein such as albumin or keyhole limpet hemocyanin.

Alternatively, hybridomas producing the desired antibody can be obtained using the DNA immunization described below. DNA immunization is an immunization method that provides immunostimulation through the expression of a sensitizing antigen in the body of an immunized animal by introducing a DNA vector designed in such a way that it provides the expression of a gene encoding an antigenic protein in the body of an animal. Compared to conventional immunization methods, in which a protein antigen is injected into the animals to be immunized, DNA immunization appears to have the following advantages:

- it is possible to carry out immunostimulation, while maintaining the structure of a membrane protein, such as GPC3; and

- there is no need to purify the antigen for immunization.

To obtain a monoclonal antibody of the present invention using DNA immunization, DNA expressing the GPC3 protein is first administered to the animal to be immunized. DNA encoding GPC3 can be synthesized using known methods such as PCR. The resulting DNA is inserted into an appropriate expression vector and then administered to the animal to be immunized. Expression vectors preferably used for this purpose include, for example, commercially available expression vectors such as pcDNA3.1. Vectors can be introduced into the body using conventional methods. For example, DNA immunization is carried out using a gene gun to introduce gold particles coated with an expression vector into the body cells of an animal to be immunized. Antibodies recognizing GPC3 can also be obtained by the methods described in WO 2003/104453.

After immunizing a mammal as described above, an elevated serum titer of GPC3-binding antibody is confirmed in the mammal. Thereafter, immune cells are obtained from the mammalian body and then used for cell fusion. In particular, splenocytes are preferably used as immune cells.

A mammalian myeloma cell is used as a cell to be fused with the above immunocyte. Myeloma cells preferably contain an acceptable selection marker for screening. The selection marker gives the cells characteristics that ensure their survival (or death) under specific culture conditions. As a selection marker, hypoxanthine-guanine phosphoribosyltransferase deficiency (abbreviated hereinafter as HGPRT deficiency) and thymidine kinase deficiency (abbreviated hereinafter as TK deficiency) are known. Cells deficient in HGPRT or TK are sensitive to hypoxanthine-aminopterin-thymidine (abbreviated hereinafter as HAT-sensitivity). Cells with GAT-sensitivity cannot synthesize DNA in the GAT selection medium and die as a result. However, when cells are fused with healthy cells, they can continue to synthesize DNA using the healthy cell's "recycling" pathway, and as a result, they can grow even in HAT selection media.

HGPRT-deficient and TK-deficient cells can be selected in medium containing 6-thioguanine, 8-azaguanine (abbreviated as 8AG hereinafter) or 5'-bromodeoxyuridine, respectively. Healthy cells are destroyed as they incorporate these pyrimidine analogs into their DNA. However, cells deficient in these enzymes can survive in the selection medium, since they cannot include these pyrimidine analogs. In addition, the selection marker includes resistance to G418, which is provided by the neomycin resistance gene, conferring resistance to 2-deoxystreptamine antibiotics (gentamicin analogues). Various types of myeloma cells are known that can be used for cell fusion.

For example, the following cells can preferably be used as myeloma cells:

P3(P3x63Ag8.653) (J. Immunol. 123 (4), 1979, pp. 1548-1550);

P3x63Ag8U.1 (Current Topics in Microbiology and Immunology 81, 1978, pp. 1-7);

NS-1 (C. Eur. J. Immunol. 6 (7), 1976, pp. 511-519);

MPC-11 (Cell 8 (3), 1976, pp. 405-415);

SP2/0 (Nature 276 (5685), 1978, pp. 269-270);

FO (J. Immunol. Methods 35 (1-2), 1980, pp. 1-21);

S194/5.XX0.BU.1 (J. Exp. Med. 148 (1), 1978, pp. 313-323);

R210 (Nature 277 (5692), 1979, pp. 131-133), etc.

Cellular fusion of immunocytes and myeloma cells is generally carried out using known methods, for example, the method described by Kohler and Milstein et al. (Methods Enzymol. 73, 1981, pp. 3-46).

More specifically, cell fusion can be performed, for example, in a conventional culture medium in the presence of a cell fusion enhancing agent. Cell fusion enhancing agents include, for example, polyethylene glycol (PEG) and Sendai virus (mouse haemagglutinating virus of Japan) (HVJ). If necessary, an auxiliary substance, such as dimethyl sulfoxide, is also added to improve the efficiency of the fusion.

The ratio of immunocytes to myeloma cells can be determined at the discretion of the investigator, preferably, for example, one myeloma cell for every 1-10 immunocytes. Culture media used for cell fusions include, for example, media suitable for growing myeloma cell lines such as RPMI1640 media and MEM media, as well as other conventional culture media used for this type of cell culture. Additionally, serum supplements such as fetal calf serum (FCS) can preferably be added to the culture medium.

For cell fusion, the above immune cells and myeloma cells, taken in predetermined amounts, are well mixed in the above culture medium. Then, a PEG solution (eg, average molecular weight of about 1000 to 6000) preheated to about 37° C. is added thereto at a concentration typically between 30% and 60% (w/v). The mixture is gently stirred until the desired confluent cells (hybridomas) are obtained. Then, the above culture medium is gradually added to the cells and centrifuged again to remove the supernatant. In this way, fusion agents that are undesirable for hybridoma growth can be removed.

The hybridomas thus obtained can be selected by culturing using a conventional selective medium such as HAT medium (culture medium containing hypoxanthine, aminopterin and thymidine). Cells other than the desired hybridomas (non-fused cells) can be killed by subsequent culturing in the above HAT medium for a sufficient period of time. As a rule, the period is from several days to several weeks. Hybridomas producing the desired antibody are then screened for and individually cloned using conventional serial dilution techniques.

The hybridomas thus obtained can be selected using a selection medium based on a selection marker carried by the myeloma cells used for cell fusion. For example, HGPRT- or TK-deficient cells can be selected by culturing using HAT medium (culture medium containing hypoxanthine, aminopterin and thymidine). In particular, when HAT-responsive myeloma cells are used for cell fusion, cells that successfully fuse with healthy cells can selectively proliferate in the HAT medium. Cells other than the desired hybridomas (non-fused cells) can be killed by culturing in the above HAT medium for a sufficient period of time. In particular, the desired hybridomas can be selected by culturing, typically for a period of time ranging from several days to several weeks. Hybridomas producing the desired antibody are then screened for and individually cloned using conventional serial dilution techniques.

Desired antibodies can preferably be selected and individually cloned using screening methods based on a known antigen/antibody reaction. For example, a GPC3-binding monoclonal antibody can bind to GPC3 expressed on a cell surface. You can screen these monoclonal antibodies using the method of cell separation based on fluorescence excitation (FACS). FACS is a system that evaluates antibody cell surface binding by analyzing cells in contact with a fluorescent antibody using a laser beam and by assessing the fluorescence emitted by individual cells.

For FACS screening for hybridomas that produce the monoclonal antibody of the present invention, GPC3 expressing cells are first obtained. Cells that are preferably used for screening are mammalian cells in which GPC3 is forced to be expressed. As a control, the ability of an antibody to bind to cell surface GPC3 can be selectively determined using non-transformed mammalian cells as host cells. In particular, anti-GPC3 monoclonal antibody-producing hybridomas can be isolated by selecting hybridomas that produce an antibody that binds to cells forcibly expressing GPC3 but not to host cells.

Alternatively, the ability of an antibody to bind to immobilized GPC3 expressing cells can be assessed based on the ELISA principle. For example, GPC3 expressing cells are immobilized on the wells of an ELISA plate. Hybridoma culture supernatants are brought into contact with the immobilized cells in the wells and antibodies are detected that bind to the immobilized cells. When the monoclonal antibodies are of murine origin, the cell-bound antibodies can be detected with an anti-mouse immunoglobulin antibody. Hybridomas producing the desired antibody that has antigen-binding activity are selected by the screening described above and can be cloned by serial dilution or a similar method.

Monoclonal antibody-producing hybridomas thus obtained can be subcultured into a conventional culture medium and stored in liquid nitrogen for an extended period of time.

The above hybridomas are cultured using a conventional method and the desired monoclonal antibodies can be obtained from culture supernatants. Alternatively, hybridomas are introduced and grown in suitable mammals and monoclonal antibodies are obtained from ascites. The first method is suitable for obtaining antibodies with a high degree of purity.

Preferably, antibodies encoded by antibody genes that are cloned from antibody-producing cells, such as the aforementioned hybridomas, can also be used. The cloned antibody gene is inserted into an appropriate vector and introduced into a host to express the antibody encoded by the gene. Methods for isolating antibody genes, inserting genes into vectors, and transforming host cells have been previously developed (see, for example, Vandamme et al., Eur. J. Biochem. 192(3), 1990, pp. 767-775). Methods for obtaining recombinant antibodies are also known and are described below.

For example, cDNA encoding the variable region (V region) of an anti-GPC3 antibody is obtained from hybridoma cells expressing an anti-GPC3 antibody. For this purpose, total RNA is first extracted from hybridomas. Methods that are used to extract mRNA from cells include, for example:

- method of ultracentrifugation in the presence of guanidine (Biochemistry 18(24), 1979, pp. 5294-5299) and

- AGPC method (Anal. Biochem. 162(1), 1987, pp. 156-159).

The extracted mRNAs can be purified using an mRNA purification kit (GE Healthcare Bioscience) or a similar kit. Alternatively, commercially available kits for extracting mRNA directly from cells, such as the QuickPrep mRNA purification kit (GE Healthcare Bioscience) can also be used. mRNA can be obtained from hybridomas using such kits. V-region-encoding cDNA antibodies can be synthesized from the resulting mRNAs using reverse transcriptase. cDNA can be synthesized using a Reverse Transcriptase First-strand cDNA Synthesis Kit (AMV) (Seikagaku Co.) or the like. In addition, the SMART RACE cDNA amplification kit (Clontech) and the PCR-based 5'-RACE method (Proc. Natl. Acad. Sci. USA 85(23 ), 1988, pp. 8998-9002; Nucleic Acids Res. 17(8), 1989, pp. 2919-2932). With this cDNA synthesis process, the appropriate restriction enzyme recognition sites described below can be introduced at both ends of the cDNA.

The cDNA fragment of interest is purified from the resulting PCR product and then inserted by ligation into the DNA vector. In this way, a recombinant vector is created and introduced into E. coli or a similar host. After colony selection, the desired recombinant vector can be obtained from colony-forming E. coli . Then, using a known method such as a nucleotide chain termination method, it is determined whether the recombinant vector has a cDNA nucleotide sequence of interest.

The 5'-RACE method, which uses primers to amplify a variable region gene, is generally used to isolate a gene encoding a variable region. First, a cDNA library used for 5'-RACE (5'-RACE cDNA library) is constructed using RNA extracted from hybridoma cells as a template. A commercially available kit, such as the SMART RACE cDNA amplification kit, can be used to synthesize a 5'-RACE cDNA library.

The antibody gene is amplified by PCR using the resulting 5'-RACE cDNA library as a template. Primers for amplifying a mouse antibody gene can be designed based on known antibody gene sequences. The nucleotide sequences of the primers vary depending on the immunoglobulin subclass. Thus, it is preferable to pre-subclass with an available kit, such as the Iso Strip mouse monoclonal antibody isotyping kit (Roche Diagnostics).

In particular, for example, to isolate genes encoding mouse IgG, primers are used that allow amplification of genes encoding γ1, γ2a, γ2b and γ3 heavy chains and κ and λ light chains. In general, a primer whose hybridization ("anneal") site with the constant region is located near the variable region is used as a 3'-terminal primer for amplifying the IgG variable region gene. In this case, the primer attached to the set of constructs of the 5' RACE cDNA library is used as a 5'-terminal primer.

The PCR products thus amplified are used to reconstruct immunoglobulins consisting of a combination of heavy and light chains. The desired antibody can be selected using the GPC3 binding activity of the reshaped immunoglobulin as an indicator. For example, when the goal is to isolate an antibody to GPC3, it is more preferred that the binding of the antibody to GPC3 is specific. You can screen for GPC3-binding antibodies, for example, using the following steps, in which:

(1) contacting a GPC3-expressing cell with an antibody containing a V region encoded by cDNA isolated from the hybridoma;

(2) determine the binding of the antibody to expressing GPC3 cell; and

(3) an antibody is selected that binds to a GPC3 expressing cell.

Methods for determining antibody binding to GPC3 expressing cells are known. In particular, antibody binding to GPC3 expressing cells can be determined using the techniques described above, such as FACS. Immobilized samples expressing GPC3 cells can be used to assess the binding activity of the antibody.

Preferred antibody screening methods that use binding activity as an indicator also include panning methods based on the use of phage vectors. Screening methods using phage vectors are advantageous when antibody genes are isolated from heavy chain and light chain subclass libraries from a population of cells expressing polyclonal antibodies. Genes encoding heavy chain and light chain variable regions can be fused with a suitable linker sequence to form a single chain Fv (scFv). Phages presenting scFv on their surface can be obtained by inserting the gene encoding scFv into a phage vector. The phages are brought into contact with the antigen of interest. DNA encoding an scFv that has binding activity of interest can then be isolated by harvesting antigen-bound phages. This process can be repeated as needed to enrich scFvs that have binding activity of interest.

After isolation of the cDNA encoding the V region of the anti-GPC3 antibody of interest, the cDNA is digested with restriction enzymes that recognize the restriction sites introduced at both ends of the cDNA. Preferred restriction enzymes recognize and cleave a nucleotide sequence that occurs at low frequency in the nucleotide sequence of an antibody gene. In addition, in order to insert the one-copy cleavage fragment in the correct orientation, a restriction site for an enzyme that forms a "sticky" end is preferably introduced into the vector of interest. The cDNA coding for the anti-GPC3 antibody V region is digested as described above and inserted into a suitable expression vector to create an antibody expression vector. When a gene encoding the constant region (C region) of an antibody and a gene encoding the above V region are fused in frame, a chimeric antibody is obtained. In the context of the present description, "chimeric antibody" means that the constant region and the variable region differ in their origin. Thus, in addition to murine/human heterochimeric antibodies to chimeric antibodies of the present invention, human/human allochimeric antibodies are also included. An expression vector for a chimeric antibody can be created by inserting the above V-region gene into an expression vector that already contains the constant region. In particular, for example, a sequence recognized by a restriction enzyme that cuts out the above V region gene should preferably be placed in the 5' region of an expression vector carrying DNA that encodes the constant region (C region) of the desired antibody. An expression vector for a chimeric antibody is created by in-frame fusion of two genes cleaved with the same restriction enzyme combination.

To obtain a monoclonal antibody to GPC3, the antibody genes are inserted into an expression vector in such a way that the expression of the genes occurs under the control of the expression regulating region. An expression control region for expressing antibodies includes, for example, enhancers and promoters. In addition, an appropriate signal sequence can be attached to the amino terminus so that the antibody is secreted from the cells to the outside. In the examples described below, a peptide having the amino acid sequence MGWSCIILFLVATATGVHS (SEQ ID NO: 72) is used as a signal sequence. Along with it, you can attach other acceptable signal sequences. The expressed polypeptide is cleaved at the carboxyl terminus of the above sequence and the resulting polypeptide is secreted from the cells to the outside as the mature polypeptide. Acceptable host cells are then transformed with an expression vector to produce recombinant cells expressing DNA that encodes an anti-GPC3 antibody.

DNAs that encode an antibody heavy chain (H chain) and an antibody light chain (L chain) are inserted separately into different expression vectors to express the antibody gene. An antibody molecule having H and L chains can be expressed by co-transfecting the same host cell for this purpose with vectors into which the H chain and L chain genes are inserted, respectively. Alternatively, host cells can be transformed with a single expression vector into which DNA encoding the H and L chains is inserted (see WO 94/11523).

Various combinations of host cells/expression vectors are known for producing antibodies by introducing isolated antibody genes into appropriate hosts. All of these expression systems can be used to isolate the antigen-binding domains and CD3-associated domains of the present invention. Suitable eukaryotic cells that are used as host cells are animal cells, plant cells and fungal cells. In particular, animal cells are, for example, the following cells:

(1) mammalian cells: CHO, COS, myeloma, baby hamster kidney (BHK), HeLa, Vero or the like;

(2) amphibian cells: clawed frog ( Xenopus ) oocytes or the like. and

(3) insect cells: sf9, sf21, Tn5 or the like.

In addition, as a plant cell, an antibody gene expression system is known in which cells derived from members of the genus Nicotiana such as Nicotiana tabacum are used. Cultured callus cells can be used to transform plant cells.

In addition, the following cells can be used as fungal cells:

yeast: the genus Saccharomyces , e.g. Saccharomyces cerevisiae , and the genus Pichia , e.g. Pichia pastoris, and

filamentous fungi: of the genus Aspergillus , e.g. Aspergillus niger .

In addition, antibody gene expression systems are known that use prokaryotic cells. For example, according to the present invention, bacterial cells can be used, i. E. coli cells, Bacillus subtilis cells, and the like. Expression vectors that carry the antibody genes of interest are introduced into these cells by transfection. The transfected cells are cultured in vitro and the desired antibody can be obtained from the culture of the transformed cells.

In addition to the above host cells, transgenic animals can also be used to produce a recombinant antibody. This means that the antibody can be obtained from an animal into which the gene encoding the antibody of interest has been introduced. For example, an antibody gene can be created as a fusion gene by inserting in frame into a gene encoding a protein that is specifically produced in milk. As a protein secreted into milk, for example, goat β-casein can be used. DNA fragments containing the fusion gene containing the antibody gene are injected into a goat embryo and then the embryo is introduced into a female goat. The desired antibodies can be obtained as a protein fused to milk protein from the milk of a transgenic goat born from a goat that is the recipient of the embryo (or its offspring). In addition, to increase the volume of milk containing the desired antibody that is produced by the transgenic goat, hormones can be administered to the transgenic goat if necessary (Ebert K. M. et al., Bio/Technology 12 (7), 1994, pp. 699-702).

When the polypeptide complex provided herein is administered to a human, an antigen-binding domain derived from an antibody produced by genetic recombination that has been artificially modified to reduce heterologous antigenicity to humans and other animals can also be used as the antigen-binding domain of the complex. These antibodies obtained by genetic recombination include, for example, humanized antibodies. These modified antibodies can be obtained using known methods.

The variable region of an antibody used to derive the antigen-binding domain of the polypeptide complex described herein typically consists of three hypervariable regions (CDRs) separated by four framework regions (FRs). The CDR is a region that largely determines the binding specificity of an antibody. The amino acid sequences of the CDRs are characterized by a high degree of variability. On the other hand, FR-forming amino acid sequences often have high identity even among antibodies with different binding specificities. Thus, in general, by CDR transplantation, the binding specificity of a particular antibody can be introduced into another antibody.

A humanized antibody is also referred to as a reshaped human antibody. In particular, humanized antibodies are known that are obtained by transplanting the CDR of an antibody of a non-human animal, such as a mouse antibody, into a human antibody, and the like. Conventional genetic engineering techniques for producing humanized antibodies are also known. In particular, for example, PCR with overlapping primers is a known method for transplanting mouse antibody CDRs into human FRs. When performing PCR with overlapping primers, the nucleotide sequence that encodes the CDR of the mouse antibody to be transplanted is added to the primers designed to synthesize the FR of the human antibody. Get primers for each of the four FR. It is generally accepted that when a murine CDR is transplanted into a human FR, it is advantageous to select a human FR having a high level of identity with the murine FR to maintain CDR function. Thus, it is generally preferred to use a human FR that contains an amino acid sequence having a high level of amino acid sequence identity with the amino acid sequence of the FR that is adjacent to the murine CDR to be transplanted.

The nucleotide sequences to be ligated are designed to be joined to each other in reading frame. Human FRs are individually synthesized using appropriate primers. As a result, products are obtained in which the DNA encoding the murine CDR is fused to the individual DNA encoding the FR. The nucleotide sequences encoding the mouse CDR of each product are designed to overlap with each other. Then, a complementary chain synthesis reaction is carried out to "anneal" the overlapping CDR regions of the products synthesized using the human antibody gene as a template. With this reaction, human FRs are inserted by ligation through mouse CDR sequences.

The full-length V region gene, into which three CDRs and four FRs are finally ligated, is amplified using primers hybridizing to the 5' or 3' end, which are added with sequences recognized by acceptable restriction enzymes. An expression vector for a humanized antibody can be obtained by inserting the DNA prepared as described above and the DNA that encodes the C region of a human antibody into an expression vector so as to be ligated in-frame. After the host is transfected with a recombinant vector to create recombinant cells, the recombinant cells are cultured and the DNA encoding the humanized antibody is expressed to produce the humanized antibody in cell culture (see European Patent Publication EP 239400 and International Patent Publication WO 1996/002576).

By qualitatively or quantitatively evaluating and measuring the antigen-binding activity of a humanized antibody prepared according to the method described above, human antibody FRs can be selected that allow the CDR to form a preferred antigen-binding site when ligated by the CDR. Amino acid residues in the FR can be changed as needed so that the CDRs of the reshaped human antibody form an acceptable antigen binding site. For example, amino acid sequence mutations can be introduced into FR using the PCR method used to transplant a mouse CDR into a human FR. More specifically, partial nucleotide sequence mutations can be introduced into primers that hybridize to FR. Nucleotide sequence mutations are introduced into FRs synthesized using the indicated primers. Mutant FR sequences having the desired characteristics can be selected by measuring and evaluating the antigen binding activity of the amino acid mutant antibody using the method mentioned above (Sato K. et al., Cancer Res. 53, 1993, pp. 851-856 ).

Alternatively, the desired human antibodies can be obtained by immunizing transgenic animals having the full range of human antibody genes (see WO 1993/012227; WO 1992/003918; WO 1994/002602; WO 1994/025585; WO 1996/034096; WO 1996 /033735), using DNA immunization.

In addition, techniques for producing human antibodies by panning using libraries of human antibodies are known. For example, the V region of a human antibody is expressed as a single chain antibody (scFv) on the surface of a phage using a phage display technique. You can select phages expressing scFv that binds to the antigen. The DNA sequence encoding the V region of a human antibody that binds to an antigen can be determined by gene analysis of selected phages. The DNA sequence of the scFv that binds to the antigen is determined. An expression vector is prepared by fusing the in-frame V region sequence to the C region sequence of the desired human antibody and then inserting it into a suitable expression vector. The expression vector is introduced into a cell suitable for the above expression. A human antibody can be produced by expressing a gene encoding a human antibody in cells. Such methods have already been described (see WO 1992/001047; WO 1992/020791; WO 1993/006213; WO 1993/011236; WO 1993/019172; WO 1995/001438; WO 1995/015388).

Antigen binding domain

In the context of the present description, "antigen-binding domain" refers to a region of an antibody containing a region that specifically binds to and is complementary to a complete antigen or region thereof. When the molecular weight of the antigen is large, the antibody can only bind to a specific site on the antigen. This particular region is referred to as an "epitope". An antigen binding domain may be formed by one or more variable domains of an antibody. Preferably, the antigen binding domains comprise both the antibody light chain variable region (VL) and the antibody heavy chain variable region (VH). Said preferred antigen binding domains include, for example, "single chain Fv (scFv)", "single chain antibody", "Fv", "single chain Fv2 (scFv2)", "Fab", and "F(ab')2".

The antigen binding domains of the polypeptide complexes of the present invention can bind to the same epitope. An epitope may be present in a protein that contains the amino acid sequence of SEQ ID NO: 2 or 4. Alternatively, the antigen-binding domains of the polypeptide complexes of the present invention may individually bind to different epitopes. An epitope may be present in a protein that contains the amino acid sequence of SEQ ID NO: 2 or 4.

Specificity

"Specific" means that a molecule that binds specifically to one or more binding partners does not have any significant ability to bind to molecules other than said partners. In addition, the term "specificity" is also used when the antigen-binding domain is specific for a particular epitope from among several epitopes included in the antigen. When the epitope to which the antigen-binding domain binds is included in several different antigens, then the polypeptide complex containing the antigen-binding domain can bind to different antigens that contain the epitope.

Antigen

The present description is not limited to any particular antigen, and any antigen except CD3 can be used. Said antigens include, for example, receptors, cancer antigens, MHC antigens, and differentiated antigens. The receptors include receptors belonging to the hematopoietic growth factor receptor family, the cytokine receptor family, the tyrosine kinase receptor family, the serine/threonine kinase receptor family, the TNF receptor family, the G protein-coupled receptor family, the GPI-anchored receptor family, the tyrosine phosphatase receptor family, the family of adhesion factors and the family of hormone receptors. The receptors belonging to these receptor families and their characteristics are described in various documents, for example in reviews such as New Comprehensive Biochemistry ed. Cooke B.A., King R.J.B., van der Molen H.J., vol. 18B "Hormones and their Actions Part II", 1988, pp. 1-46, Elsevier Science Publishers BV.; and SAIBO KOGAKU (Cell Technology) Supplementary Volume Handbook Series "Secchaku Inshi Handbook (Handbook for Adhesion factors)" ed. M. Miyasaka, 1994, Shujunsha, Tokyo, Japan; and Patthy, Cell, 61(1), 1990, ss. 13-14; Ullrich et al., Cell 61(2), 1990, ss. 203-212; Massagué, Cell 69(6), 1992, pp. 1067-1070; Miyajima et al., Annu. Rev. Immunol. 10, 1992, pp. 295-331; Taga et al., FASEB J. 6, 1992, pp. 3387-3396); Fantl et al., Annu. Rev. Biochem. 62, 1993, pp. 453-481; Smith et al., Cell 76(6), 1994, ss. 959-962, Flower D.R. at Biochim. Biophys. Acta: Flower Biochim. Biophys. Acta, 1422(3), 1999, pp. 207-234.

In particular, receptors belonging to the above receptor families preferably include, for example, human and mouse erythropoietin (EPO) receptors (Blood 76(1), 1990, pp. 31-35; Cell 57(2), 1989, pp. 277-285), human and mouse granulocyte colony stimulating factor (G-CSF) receptors (Proc. Natl. Acad. Sci. USA. 87(22), 1990, pp. 8702-8706; mG-CSFR, Cell 61(2) , 1990, pp. 341-350), human and murine thrombopoietin (TPO) receptors (Proc. Natl. Acad. Sci. USA 89(12), 1992, pp. 5640-5644; EMBO J. 12(7), 1993 , pp. 2645-53), human and murine insulin receptors (Nature 313(6005), 1985, pp. 756-761), human and murine Flt-3 ligand receptors (Proc. Natl. Acad. Sci. USA 91(2 ), 1994, pp. 459-463), human and mouse platelet growth factor (PDGF) receptors (Proc. Natl. Acad. Sci. USA. 85(10), 1988, pp. 3435-3439), human and murine interferon (IFN)-α and -β (Cell 60(2), 1990, pp. 225-234; and Cell 77(3), 1994, pp. . 391-400), human and mouse leptin receptors, human and mouse growth hormone (GH) receptors, human and mouse interleukin (IL)-10 receptors, human and mouse insulin-like growth factor (IGF)-I receptors, human and mouse leukosininhibitory receptors factor (LIF); and human and mouse ciliary neurotrophic factor (CNTF) receptors.

Cancer antigens are antigens that are expressed when cells become cancerous and are also referred to as "tumor-specific antigens". In addition, abnormal sugar chains that are expressed on the cell surface or protein molecules when cells become cancerous are also cancer antigens and are referred to as "cancer-associated carbohydrate antigens". These cancer antigens include, for example, GPC3 (Int J Cancer. 103(4), 2003, pp. 455-465), EpCAM (Proc. Natl. Acad. Sci. USA. 86(1), 1989, pp. 27- 31), EGFR, CA19-9, CA15-3 and sialyl SSEA-1 (SLX). GPC3 belongs to the above family of GPI-anchored receptors and is expressed in several types of cancer, including liver cancer. EpCAM is expressed in a variety of cancers, including lung cancer, and its polynucleotide and polypeptide sequences are listed in the RefSeq database under accession numbers NM_002354.2 (SEQ ID NO: 3) and NP_002345.2 (SEQ ID NO: 4), respectively.

Generally, MHC antigens are classified into class I and class II MHC antigens. MHC class I antigens include HLA-A, -B, -C, -E, -F, -G and -H, and MHC class II antigens include HLA-DR, -DQ and -DP.

Differentiated antigens include CD1, CD2, CD4, CD5, CD6, CD7, CD8, CD10, CD11a, CD11b, CD11c, CD13, CD14, CD15s, CD16, CD18, CD19, CD20, CD21, CD23, CD25, CD28, CD29, CD30 , CD32, CD33, CD34, CD35, CD38, CD40, CD41a, CD41b, CD42a, CD42b, CD43, CD44, CD45, CD45RO, CD48, CD49a, CD49b, CD49c, CD49d, CD49e, CD49f, CD51, CD54, CD55, CD56 , CD57, CD58, CD61, CD62E, CD62L, CD62P, CD64, CD69, CD71, CD73, CD95, CD102, CD106, CD122, CD126 and CDw130.

epitope

"Epitope" means an antigenic determinant in an antigen and refers to the antigenic site to which the antigen-binding domain of the polypeptide complex described herein binds. Thus, for example, an epitope can be characterized based on its structure. Alternatively, an epitope can be characterized based on the antigen-binding activity of the polypeptide complex that recognizes the epitope. When the antigen is a peptide or polypeptide, the epitope can be determined from the amino acid residues that form the epitope. Alternatively, when the epitope is a sugar chain, the epitope can be determined by its specific sugar chain structure.

A linear epitope is an epitope that contains an epitope in which the primary amino acid sequence is recognized. Said linear epitope typically contains at least three, and most often at least five, eg, about 8 to 10 or 6 to 20 amino acid residues in sequence.

In contrast to a linear epitope, a "conformational epitope" is an epitope in which the primary amino acid sequence constituting the epitope is not the sole determinant of the recognized epitope (e.g., it is not necessary that the primary amino acid sequence of the conformational epitope be recognized by an epitope-specific antibody). Conformational epitopes may contain more amino acid residues than linear epitopes. An antibody that recognizes a conformational epitope recognizes the three-dimensional structure of a peptide or protein. For example, when a protein molecule is folded into a three-dimensional structure, the amino acids and/or polypeptide backbones that form the conformational epitope align and the epitope becomes recognizable by the antibody. Methods for determining epitope conformations include, but are not limited to, for example, X-ray crystallography, 2D nuclear magnetic resonance, site-specific spin labeling, and electron paramagnetic resonance (see, for example, Epitope Mapping Protocols in Methods in Molecular Biology, ed. Morris , vol. 66, 1996).

Examples of methods for assessing epitope binding by a test polypeptide complex containing a GPC3 antigen binding domain are described below. The exemplary methods described below for assessing epitope binding by a test polypeptide complex containing an antigen-binding domain for an antigen other than GPC3 may also be used.

For example, to confirm that a test polypeptide complex containing a GPC3 antigen binding domain recognizes a linear epitope in a GPC3 molecule, the method described below can be used. For the above purpose, a linear peptide is synthesized containing the amino acid sequence forming the extracellular domain of GPC3. The peptide can be chemically synthesized or genetically engineered using a region in the GPC3 cDNA that encodes an amino acid sequence corresponding to the extracellular domain. The test polypeptide complex containing the GPC3 antigen binding domain is then evaluated for its binding activity to a linear peptide that contains an amino acid sequence forming an extracellular domain. For example, an immobilized linear peptide can be used as an antigen in an ELISA to evaluate activity for binding a polypeptide complex to a peptide. Alternatively, the linear peptide binding activity can be measured by the level of inhibition by the linear peptide of the binding of the polypeptide complex to GPC3 expressing cells. These assays can demonstrate the binding activity of a polypeptide complex to a linear peptide.

Whether a polypeptide complex containing a GPC3 antigen binding domain has the ability to recognize a conformational epitope can be determined as follows. For the above purpose, GPC3 expressing cells are obtained. A test polypeptide complex containing a GPC3 antigen-binding domain is said to recognize a conformational epitope if, upon contact, it binds strongly to GPC3-expressing cells, but binds only slightly to an immobilized linear peptide that contains the amino acid sequence that forms the GPC3 extracellular domain. In the context of the present description, "slightly binds" means that the binding activity is 80% or less, typically 50% or less, preferably 30% or less, and most preferably 15% or less, compared to the binding activity to human GPC3 expressing cells. .

Methods for evaluating the binding activity of a test polypeptide complex containing a GPC3 antigen-binding domain to GPC3 expressing cells include, for example, those described in: Antibodies: A Laboratory Manual (ed. Harlow, David Lane, Cold Spring Harbor Laboratory, 1988, pp. 359-420). In particular, the evaluation can be performed based on the principle of ELISA or fluorescence excitation-based cell separation (FACS) using GPC3-expressing cells as antigen.

In an ELISA format, the binding activity of a test polypeptide complex containing a GPC3 antigen binding domain to GPC3 expressing cells can be quantified by comparing the signal levels generated during the enzymatic reaction. Specifically, the polypeptide complex to be tested is added to an ELISA plate on which GPC3 expressing cells are immobilized. The cell-associated test polypeptide complex is then detected with an enzyme-labeled antibody that recognizes the test polypeptide complex. Alternatively, when FACS is used, serial dilutions of the test polypeptide complex are prepared and the titer of antibody binding to GPC3 expressing cells can be determined by comparing binding activity to GPC3 expressing cells.

Binding of the test polypeptide complex to an antigen that is expressed on the surface of cells suspended in buffer or a similar medium can be detected using a flow cytometer. Known flow cytometers are, for example, the following devices:

FACS ^CantoTM II,

FACSAria ^™ ,

FACSArray ^™ ,

^{FACSVantageTM} SE

^{FACSCaliburTM} (all trademarks of BD Biosciences),

EPICS ALTRA HyPerSort,

Cytomics FC 500,

EPICS XL-MCL ADC EPICS XL ADC,

Cell Lab Quanta/Cell Lab Quanta SC (all trademarks of Beckman Coulter).

Preferred methods for assessing the binding activity of a test polypeptide complex containing a GPC3 antigen binding domain to an antigen are, for example, the following method. GPC3 expressing cells are first reacted with the polypeptide complex to be tested and then stained with a FITC-labeled secondary antibody that recognizes the polypeptide complex. The test polypeptide complex is appropriately diluted with an acceptable buffer to obtain the complex at the desired concentration. For example, the complex can be used at a concentration of 10 μg/ml to 10 ng/ml. Then determine the intensity of fluorescence and the number of cells using FACSCalibur (firm BD). The fluorescence intensity, determined using analyzes based on the CELL QUEST program (company BD), namely, expressed as geometric mean values, reflects the level of antibody binding to cells. This means that the binding activity of a test polypeptide complex, which is characterized by the amount of bound test polypeptide complex, can be evaluated by determining the geometric mean.

Whether a test polypeptide complex containing a GPC3 antigen binding domain shares an epitope with another polypeptide complex can be judged by competition between the two complexes for the same epitope. Competition between polypeptide complexes can be determined by a cross-blockade assay or a similar assay. For example, a preferred cross-blockade assay is a competitive ELISA assay.

In particular, when performing a cross-blockade assay, the GPC3 protein immobilized in the wells of a microtiter plate is pre-incubated with or without the polypeptide complex considered as a candidate competitor, and then the test polypeptide complex is added. The amount of GPC3 protein-bound test polypeptide complex in the wells is inversely correlated with the binding activity of the candidate polypeptide complex that competes for binding to the same epitope. This means that the higher the affinity of the polypeptide complex considered as a candidate competitor for the same epitope, the lower the binding activity of the tested polypeptide complex with GPC3-sensitized wells.

The amount of the test polypeptide complex bound to the wells via the GPC3 protein is easily determined by prelabeling the polypeptide complex. For example, a biotin labeled polypeptide complex is evaluated using an avidin/peroxidase conjugate and an appropriate substrate. In particular, a cross-blockade assay that uses enzymes such as peroxidase as labels is referred to as "competition ELISA assay (competitive assay)". The polypeptide complex can also be labeled with other labeling substances that can be detected or evaluated. In particular, radioactive labels, fluorescent labels, and the like are known.

When the candidate polypeptide complex can block the binding of the test polypeptide complex that contains the GPC3 antigen binding domain by at least 20%, preferably at least 20-50% and preferably at least 50% compared with the binding activity established in the control experiment, which is carried out in the absence of the polypeptide complex considered as a candidate competitor, it is considered that the test polypeptide complex is characterized by a pronounced ability to bind to the same epitope that binds the polypeptide complex considered as a competitor candidate, or it competes for binding to the same epitope.

If the structure of the epitope associated with the test polypeptide complex, which contains the GPC3 antigen binding domain, has already been identified, then to decide whether the test and control polypeptide complexes have the same epitope, the binding activities of the two polypeptide complexes to the peptide can be compared. obtained by introducing amino acid mutations into the peptide that forms the epitope.

To evaluate the above binding activities, for example, the binding activities of the test and control polypeptide complexes are compared with the linear peptide in which the mutation has been introduced using the above ELISA format. In addition to the ELISA method, binding activity for a mutant peptide bound to a column can be determined by passing test and control polypeptide complexes through the column and then assessing the amount of polypeptide complex eluted into the elution solution. Techniques are known for adsorbing a mutant peptide onto a column, for example in the form of a GST-fused peptide.

Alternatively, when the identified epitope is a conformational epitope, the following method can be used to decide whether the test and control polypeptide complexes share a common epitope. First, cells expressing GPC3 and cells expressing GPC3 with a mutation introduced into the epitope are obtained. The test and control polypeptide complexes are added to a cell suspension obtained by suspending the cells in an acceptable buffer such as PBS. The cell suspensions are then washed appropriately with buffer and a FITC-labeled antibody that recognizes the test and control polypeptide complexes is added. Determine the intensity of fluorescence and the number of cells stained with labeled antibody using FACSCalibur (firm BD). The test and control polypeptide complexes are appropriately diluted with an acceptable buffer and used at the desired concentration. For example, they can be used at a concentration of 10 μg/ml to 10 ng/ml. The fluorescence intensity is determined by an assay whose results are evaluated using the CELL QUEST program (BD), namely, the geometric mean is determined, which reflects the amount of labeled antibody bound to the cells. This means that the binding activities of the test and control polypeptide complexes, which are characterized by the amount of labeled antibody bound, can be determined by evaluating the geometric mean.

When carrying out the method described above, for example, the following method can be used to decide whether the polypeptide complex is characterized by "minor binding to cells that express the mutant GPC3". First, test and control polypeptide complexes associated with cells that express the mutant GPC3 are stained with labeled antibody. Then determine the intensity of fluorescence in the cells. When a flow cytometer is used to evaluate fluorescence, the fluorescence intensity can be analyzed using the CELL QUEST program. Based on the geometric mean values in the presence of the polypeptide complex and without it, the value used for comparison (ΔGeo-Mean) can be calculated according to the formula below, characterizing the degree of increase in fluorescence intensity as a result of binding to the polypeptide complex.

ΔGeo-Mean = Geo-Mean (in the presence of the polypeptide complex)/Geo-Mean (in the absence of the polypeptide complex)

The comparative geometric mean value (ΔGeo-Mean value for the mutant GPC3 molecule) determined by the assay described above, which reflects the amount of test polypeptide complex bound to cells that express the mutant GPC3, is compared with the relative ΔGeo-Mean value, which reflects the amount of test polypeptide complex associated with expressing GPC3 cells. In this case, the concentrations of the test polypeptide complex used to determine the relative ΔGeo-Mean values for GPC3 expressing cells and cells expressing mutant GPC3 are most preferably adjusted to be the same or substantially the same. A polypeptide complex confirmed to recognize an epitope in GPC3 is used as a control polypeptide complex.

If the ΔGeo-Mean value of the test polypeptide complex used for comparison against cells expressing the mutant GPC3 is lower than the ΔGeo-Mean value used for comparison of the test polypeptide complex against cells expressing GPC3 by at least 80%, preferably by 50%, more preferably 30% and most preferably 15%, the test polypeptide complex is considered to "not significantly bind to cells that express the mutated GPC3". The formula for determining the Geo-Mean value (geometric mean) is described in the CELL QUEST software application manual (BD biosciences). When by comparison it is established that the relative values are practically equivalent, then the test and control polypeptide complexes can be considered to have the same epitope.

Variable Fragment (Fv)

As used herein, "variable fragment (Fv)" refers to the smallest unit of an antibody-derived antigen-binding domain, which consists of a pair comprising an antibody light chain variable region (VL) and an antibody heavy chain variable region (VH). In 1988, Skerra and Pluckthun discovered that homogeneous and active antibodies can be obtained from the periplasmic fraction of E. coli by inserting the antibody gene downstream of the bacterial signal sequence and inducing gene expression in E. coli (Science 240(4855), 1988, pp. 1038-1041). In Fv derived from the periplasmic fraction, VH associates with VL in such a way as to bind to antigen.

According to the present description, Fv preferably include, for example, a pair of Fv, which is a polypeptide complex or the like, containing:

(1) a bivalent antigen-binding domain, which is a bivalent scFv in which one monovalent scFv of a bivalent scFv is linked to one Fc domain-forming polypeptide via a heavy chain Fv fragment forming a CD3 binding domain, and the other monovalent scFv is linked to another an Fc domain forming polypeptide via a light chain Fv fragment forming a CD3 binding domain;

(2) a domain containing an Fc domain that does not have the ability to bind the Fcγ receptor and is derived from amino acids forming the Fc domain of IgG1, IgG2a, IgG3, or IgG4, and

(3) at least one univalent CD3 binding domain,

in which the Fv fragments of the light chain and the heavy chain as a result of association form such a CD3-binding domain that can bind to the CD3 antigen.

scFv, single chain antibody and sc(Fv)2

As used herein, "scFv", "single chain antibody", and "sc(Fv)2" all refer to a single polypeptide chain antibody fragment that contains the heavy and light chain derived variable regions but no constant region. Typically, a single chain antibody also contains a polypeptide linker between the VH and VL domains, allowing the formation of the desired structure, which is likely to provide antigen binding. The single chain antibody is described in detail in Pluckthun, The Pharmacology of Monoclonal Antibodies, vol. 113, ed. Rosenburg and Moore, Springer-Verlag, New York, 1994, pp. 269-315" (see also International Patent Publication WO 1988/001649; US Nos. 4946778 and 5260203). In a particular embodiment, the single chain antibody may be bispecific and/or humanized.

scFv is an antigen binding domain in which VH and VL forming Fv are linked to each other by a peptide linker (Proc. Natl. Acad. Sci. U.S.A. 85(16), 1988, pp. 5879-5883). VH and VL can be held in close proximity using a peptide linker.

sc(Fv)2 is a single chain antibody in which the four variable regions of two VLs and two VHs are linked by linkers such as peptide linkers to form one chain (J Immunol. Methods 231(1-2), 1999, pp. 177-189 ). Two VH and two VL can be derived from various monoclonal antibodies. Said sc(Fv)2 preferably include, for example, a bispecific sc(Fv)2 that recognizes two epitopes present on the same antigen as described in Journal of Immunology 152(11), 1994, ss. 5368-5374. sc(Fv)2 can be obtained by methods known to those skilled in the art. For example, sc(Fv)2 can be generated by linking the scFv with a linker such as a peptide linker.

According to the present description, the form of the antigen-binding domain forming sc(Fv)2 refers to an antibody in which two VH units and two VL units are organized in the following order: VH, VL, VH and VL ([VH]-linker-[VL]- linker-[VH]-linker-[VL]), starting at the N-terminus of the single chain polypeptide. The arrangement order of the two VH units and the two VL units is not limited to the above shape, and they can be arranged in any order. An example of the arrangement order in various forms is given below.

[VL]-linker-[VH]-linker-[VH]-linker-[VL]

[VH]-linker-[VL]-linker-[VL]-linker-[VH]

[VH]-linker-[VH]-linker-[VL]-linker-[VL]

[VL]-linker-[VL]-linker-[VH]-linker-[VH]

[VL]-linker-[VH]-linker-[VL]-linker-[VH].

The molecular form of sc(Fv)2 is also detailed in WO 2006/132352. Thus, according to these descriptions, specialists in this field can receive the required sc(Fv)2, designed to create the polypeptide complexes presented in the present description.

In addition, the polypeptide complexes of the present invention can be conjugated with a carrier polymer such as PEG or an organic compound such as an anticancer agent. Alternatively, an additional sugar chain sequence is preferably inserted into the polypeptide complexes in order for the sugar chain to provide the desired effect.

Linkers for linking variable regions of an antibody are arbitrary peptide linkers that can be introduced by genetic engineering, synthetic linkers, and linkers as described in, for example, Protein Engineering, 9(3), 1996, ss. 299-305. However, for the purposes of the present invention, peptide linkers are most preferred. The length of the peptide linkers is not specifically limited, and those skilled in the art can choose it depending on the task at hand. The preferred length is five or more amino acids (with the upper limit being (but not limited to) generally up to 30 amino acids or less, preferably 20 amino acids or less), and most preferably 15 amino acids. When sc(Fv)2 contains three peptide linkers, their length may be the same or different.

For example, said peptide linkers include:

Ser,

Gly Ser,

Gly Gly Ser,

Ser Gly Gly,

Gly Gly Gly Ser (SEQ ID NO: 5),

Ser Gly Gly Gly (SEQ ID NO: 6),

Gly Gly Gly Gly Ser (SEQ ID NO: 7),

Ser-Gly-Gly-Gly-Gly (SEQ ID NO: 8),

Gly Gly Gly Gly Gly Ser (SEQ ID NO: 9),

Ser-Gly-Gly-Gly-Gly-Gly (SEQ ID NO: 10),

Gly Gly Gly Gly Gly Gly Ser (SEQ ID NO: 11),

Ser-Gly-Gly-Gly-Gly-Gly-Gly (SEQ ID NO: 12),

(Gly·Gly·Gly·Gly·Ser (SEQ ID NO: 7))n,

(Ser Gly Gly Gly Gly (SEQ ID NO: 8))n,

where n is an integer of 1 or higher. Those skilled in the art may choose the length or sequence of the peptide linkers depending on the task at hand.

As a rule, synthetic linkers (chemical cross-linking agents) are used for cross-linking, they are, for example:

N-hydroxysuccinimide (NHS),

disuccinimidyl suberate (DSS)

bis(sulfosuccinimidyl)suberate (BS ³ ),

dithiobis(succinimidyl propionate) (DSP),

dithiobis (sulfosuccinimidyl propionate) (DTSSP),

ethylene glycol bis(succinimidyl succinate) (EGS),

ethylene glycol bis (sulfosuccinimidyl succinate) (sulfo-EGS),

disuccinimidyl tartrate (DST), disulfosuccinimidyl tartrate (sulfo-DST),

bis[2-(succinimidoxycarbonyloxy)ethyl]sulfone (BSOCOES) and

and bis[2-(sulfosuccinimidoxycarbonyloxy)ethyl]sulfone (sulfo-BSOCOES). These cross-linking agents are commercially available.

Typically, three linkers are required to link together the four variable regions of an antibody. The linkers used may be of the same type or of different types.

Fab, F(ab')2 and Fab'

"Fab" consists of one light chain and CH1 domain and variable region from one heavy chain. The heavy chain of a Fab molecule cannot form disulfide bridges with the heavy chain of another molecule.

"F(ab')2" or "Fab" is produced by treating an immunoglobulin (monoclonal antibody) with a protease such as pepsin and papain, and refers to an antibody fragment obtained by cleaving an immunoglobulin (monoclonal antibody) in the vicinity of disulfide bridges present between hinge regions in each of the two H-strands. For example, papain cleaves IgG in front of the disulfide bridges present between the hinge regions in each of the two H chains to form two homologous antibody fragments in which the L chain containing VL (L chain variable region) and CL (constant region L chain) are linked to an H chain fragment containing VH (H chain variable domain) and CHγ1 (γ1 region in the H chain constant region) via a disulfide bridge at their C-terminal regions. Each of these two homologous antibody fragments is referred to as Fab'.

"F(ab')2" consists of two light chains and two heavy chains, containing a CH1 domain and part of the CH2 domains in the constant region, as a result, disulfide bridges are formed between the two heavy chains. F(ab')2 forming the polypeptide complex presented in the present description, preferably can be obtained as follows. A complete monoclonal antibody or similar antibody containing the desired antigen-binding domain is partially cleaved with a protease such as pepsin; and Fc fragments are removed by adsorption to a protein A column. There is no limitation on a particular protease as long as it has the ability to selectively cleave a complete antibody to form F(ab')2 under appropriate reaction conditions for that enzyme, such as pH. Said proteases are, for example, pepsin and ficin.

Fc domain

The Fc domain that forms the polypeptide complex presented in the present description can preferably be obtained as follows. An antibody, such as a monoclonal antibody, is partially cleaved with a protease, such as pepsin. The resulting fragment is then adsorbed onto a protein A or protein G column and eluted with an appropriate elution buffer. There is no limitation on a particular protease as long as it has the ability to selectively cleave antibodies, such as monoclonal antibodies, under appropriate reaction conditions for that enzyme, such as pH. Said proteases are, for example, pepsin and ficin.

The polypeptide complexes provided herein contain an Fc domain with reduced ability to bind the Fcγ receptor, which include amino acids forming the Fc domain of IgG1, IgG2, IgG3, or IgG4.

The isotype of an antibody is determined based on the structure of the constant region. The constant regions of the IgG1, IgG2, IgG3 and IgG4 isotypes are designated as Cγ1, Cγ2, Cγ3 and Cγ4, respectively. The amino acid sequences of the Fc domain of the polypeptides that form human Cγ1, Cγ2, Cγ3 and Cγ4 are exemplified in SEQ ID NOs: 23, 24, 25 and 26, respectively. The relationship between the amino acid residues forming each amino acid sequence and Cabot's EU numbering (referred to herein as the EU index) is shown in FIG. eighteen.

The term "Fc domain" refers to the region outside of F(ab')2, which contains two light chains and two heavy chains, containing part of the constant region, which includes the CH1 domain and the region between the CH1 and CH2 domains, which allows the formation disulfide bridges between two heavy chains. The Fc domain forming the polypeptide complex presented in the present description can preferably be obtained as follows. An IgG1, IgG2, IgG3, or IgG4 monoclonal antibody or similar antibody is partially cleaved with a protease such as pepsin, followed by elution of the fraction adsorbed on the Protein A column. formation of F(ab')2 under appropriate reaction conditions for the enzyme, such as pH. Said proteases are, for example, pepsin and ficin.

Fcγ receptor

The term "Fcγ receptor" refers to a receptor having the ability to bind to the Fc domain of monoclonal antibodies in the form of IgG1, IgG2, IgG3 or IgG4, and it includes all members belonging to the family of proteins that are encoded primarily by the Fcγ receptor gene. In humans, the family includes FcγRI (CD64), including the isoforms FcγRIa, FcγRIb and FcγRIc; FcγRII (CD32), including isoforms FcγRIIa (including allotype H131 and R131), FcγRIIb (including FcγRIIb-1 and FcγRIIb-2) and FcγRIIc; and FcγRIII (CD16), including the FcγRIIIa isoform (including the V158 and F158 allotypes), and FcγRIIIb (including the FcγRIIIb-NA1 and FcγRIIIb-NA2 allotypes); as well as all unidentified human FcγRs, FcγR isoforms and their allotypes. However, the Fcγ receptor is not limited to these examples. FcγR includes, but is not limited to, FcγR derived from human, mouse, rat, rabbit, and monkey antibodies. FcγR can be derived from any organism. Mouse FcγR includes, but is not limited to, FcγRI (CD64), FcγRII (CD32), FcγRIII (CD16), and FcγRIII-2 (CD16-2), as well as all unidentified murine FcγRs, FcγR isoforms, and allotypes thereof. Said preferred Fcγ receptors include, for example, human FcγI (CD64), FcγIIA (CD32), FcγIIB (CD32), FcγIIIA (CD16) and/or FcγIIIB (CD16). The polynucleotide sequence and amino acid sequence of FcγI are shown in SEQ ID NOs: 13 (NM_000566.3) and 14 (NP_000557.1), respectively; the polynucleotide sequence and amino acid sequence of FcγIIA are shown in SEQ ID NOs: 15 (BC020823.1) and 16 (AAH20823.1), respectively; the polynucleotide sequence and amino acid sequence of FcγIIB are shown in SEQ ID NOs: 17 (BC146678.1) and 18 (AAI46679.1), respectively; the polynucleotide sequence and amino acid sequence of FcγIIIA are shown in SEQ ID NOs: 19 (BC033678.1) and 20 (AAH33678.1), respectively, and the polynucleotide sequence and amino acid sequence of FcγIIIB are shown in SEQ ID NOs: 21 (BC128562.1) and 22 (AAI28563. 1) respectively (the registration number of each sequence in the RefSeq database is indicated in brackets). Whether an Fcγ receptor has the ability to bind to the Fc domain of an IgG1, IgG2, IgG3, or IgG4 monoclonal antibody can be assessed using the ALPHA Screen® assay (Amplified L uminescent P roximity Homogeneous A ssay )), BIACORE method based on surface plasmon resonance (SPR), etc. (Proc. Natl. Acad. Sci. USA 103(11), 2006, pp. 4005-4010), in addition to the FACS and ELISA formats described above.

The term "Fc ligand" or "effector ligand" refers to a molecule, and preferably a polypeptide, that binds to the Fc domain of an antibody to form an Fc/Fc ligand complex. The molecule can be obtained from any organism. Binding of an Fc ligand to an Fc preferably induces one or more effector functions. These ligands include, but are not limited to, Fc receptors, FcγR, FcαR, FcεR, FcRn, C1q and C3, mannan-binding lectin, mannose receptor, Staphylococcus A protein, Staphylococcus G protein, and viral FcγRs. Fc ligands also include Fc receptor (FcRH) homologues (Davis et al., Immunological Reviews 190, 2002, pp. 123-136), which are a family of Fc receptors homologous to FcγR. Fc ligands also include unidentified molecules that bind to Fc.

Fcγ receptor binding activity

Disruption of Fc domain binding activity to any of the Fcγ receptors FcγI, FcγIIA, FcγIIB, FcγIIIA and/or FcγIIIB can be assessed using the FACS and ELISA formats described above, as well as ALPHA Screen analysis (homogeneous analysis of luminescence enhanced by proximity effect, BIACORE method based on surface plasmon resonance (SPR) (Proc. Natl. Acad. Sci. USA 103(11), 2006, pp. 4005-4010).

ALPHA Screen analysis is carried out on the basis of ALPHA technology, which is based on the principle described below, using two types of beads. The luminescent signal becomes detectable only when there is a biological interaction of the molecules associated with the donor beads with the molecules associated with the acceptor beads, and when both beads are in close proximity to each other. Excited by a laser beam, the photosensitizer in donor granules converts oxygen into excited singlet oxygen. When singlet oxygen diffuses from the donor granules and reaches the acceptor granules located in close proximity, a chemiluminescent reaction is induced in the acceptor granules. As a result, this reaction leads to the emission of light. If the molecules bound to the donor granules do not interact with the acceptor granules, then the singlet oxygen produced by the donor granules does not reach the acceptor granules and the chemiluminescent reaction does not occur.

For example, a biotin-labeled polypeptide complex is immobilized on donor beads, and a glutathione-S-transferase (GST) labeled Fcγ receptor is immobilized on acceptor beads. In the absence of a polypeptide complex containing a competing mutant Fc domain, the Fcγ receptor interacts with a polypeptide complex containing a wild-type Fc domain, resulting in a signal with a wavelength of 520 to 620 nm. The polypeptide complex containing the unlabeled mutant Fc domain competes with the polypeptide complex containing the wild-type Fc domain for interaction with the Fcγ receptor. Relative binding affinity can be assessed by quantifying the reduction in fluorescence due to competition. Methods for biotinylating polypeptide complexes such as antibodies with sulfo-NHS-biotin or the like are known. Acceptable methods for introducing a GST tag into the Fcγ receptor include methods that involve in-frame fusion of Fcγ and GST-encoding polypeptides, expression of the fusion gene using cells into which a vector carrying the gene has been introduced, and subsequent purification using a column containing glutathione. . The induced signal can preferably be analyzed, for example, by fitting to a single site competition model based on non-linear regression analysis using a program such as GRAPHPAD PRISM (GraphPad; San Diego).

One of the substances intended to study their interaction is immobilized as a ligand on a thin gold layer of the sensor chip. When light enters the rear surface of the sensor chip so that there is total reflection at the interface between the thin layer of gold and glass, then the intensity of the reflected light at a certain site is partially reduced (SPR signal). Another substance, intended to study its interaction, is injected as an analyte on the surface of the sensor chip. When an analyte binds to a ligand, the mass of the immobilized ligand molecule increases. This changes the refractive index of the solvent on the surface of the sensor chip. A change in refractive index causes a positive shift in the SPR signal (conversely, dissociation shifts the signal back to its original position). In the Biacore system, the level of the shift described above (i.e., mass change on the surface of the sensor chip) is plotted on the vertical axis, and thus a graph of mass change versus time (sensogram) is obtained as quantitative data. The kinetic parameters (association rate constant (ka) and dissociation rate constant (kd)) are determined from the sensorgram curves, and affinity (KD) is determined as the ratio of the two constants. BIACORE methods are preferably used for inhibition assays. Examples of such an inhibition assay are described in Proc. Natl. Acad. sci. USA 103(11), 2006, pp. 4005-4010.

In the context of the present description, "reduced activity of binding to the Fcγ receptor" means, for example, that when using the assay method described above, the competitive activity of the tested polypeptide complex is 50% or less, preferably 45% or less, 40% or less, 35% or less , 30% or less, 20% or less, or 15% or less, and most preferably 10% or less, 9% or less, 8% or less, 7% or less, 6% or less, 5% or less, 4% or less, 3% or less, 2% or less, or 1% less than the competitive activity of the control polypeptide complex.

Polypeptide complexes containing the Fc domain of a monoclonal antibody in the form of IgG1, IgG2, IgG3 or IgG4, respectively, can be used as control polypeptide complexes. The Fc domain structures are shown in SEQ ID NO: 23 (A added to the N-terminus of the sequence listed in RefSeq accession number AAC82527.1), SEQ ID NO: 24 (A added to the N-terminus of the sequence listed in RefSeq accession number AAB59393.1), SEQ ID NO: 25 (A added to the N-terminus of the sequence listed in RefSeq accession CAA27268.1) and SEQ ID NO: 26 (A added to the N-terminus of the sequence listed in RefSeq accession number AAB59394.1). In addition, when a polypeptide complex containing a mutant Fc domain of an antibody of a particular isotype is used as a test substance, the effect of the mutant mutation on Fcγ receptor binding activity is evaluated using the polypeptide complex containing the Fc domain of the same isotype as a control. As described above, respectively receive polypeptide complexes containing a mutant Fc domain with a truly reduced activity of binding to the Fcγ receptor.

Such known mutants include, for example, amino acid deletion mutants 231A-238S (EU numbering) (WO 2009/011941), as well as mutants C226S, C229S, P238S, (C220S) (J. Rheumatol 34, 2007, p. 11) ; C226S and C229S (Hum. Antibod. Hybridomas 1(1), 1990, pp. 47-54); C226S, C229S, E233P, L234V, and L235A (Blood 109, 2007, pp. 1185-1192).

Preferred polypeptide complexes are those that contain an Fc domain with an amino acid substitution at position 220, 226, 229, 231, 232, 233, 234, 235, 236, 237, 238, 239, 240, 264, 265, 266, 267, 269, 270, 295, 296, 297, 298, 299, 300, 325, 327, 328, 329, 330, 331, or 332 (EU numbering) in the amino acids that form the Fc domain of an antibody of a particular isotype. The invention is not limited to the isotype of the antibody from which the Fc domain is derived, and the corresponding Fc domain derived from a monoclonal antibody as IgG1, IgG2, IgG3 or IgG4 can be used. Preferably, an Fc domain derived from IgG1 antibodies should be used.

Preferred polypeptide complexes are, for example, complexes that contain an Fc domain containing one of the following substitutions, the position of which is determined according to the EU numbering (each number indicates the position of the amino acid residue according to the EU numbering; and the single letter amino acid symbol that precedes the number , denotes the amino acid residue before its substitution, and the one-letter amino acid symbol following the number denotes the amino acid residue after the substitution), amino acids that form the Fc domain of the antibody in the form of IgG1:

(a) L234F, L235E, P331S;

(b) C226S, C229S, P238S;

(c) C226S, C229S or

(d) C226S, C229S, E233P, L234V, L235A;

as well as complexes that contain an Fc domain with a deletion of amino acids at positions 231-238.

In addition, preferred polypeptide complexes are also complexes that contain an Fc domain containing one of the following substitutions, the position of which is determined according to the EU numbering, of the amino acids that form the Fc domain of an IgG2 antibody:

(e) H268Q, V309L, A330S and P331S;

(e) V234A;

(g) G237A;

(h) V234A and G237A;

(i) A235E and G237A or

(j) V234A, A235E and G237A. Each number indicates the position of the amino acid residue according to the EU numbering; and the one-letter amino acid symbol that precedes the number indicates the amino acid residue before it was changed, and the one-letter amino acid symbol that follows the number indicates the amino acid residue after the change.

In addition, preferred polypeptide complexes are also complexes that contain an Fc domain containing one of the following substitutions, the position of which is determined according to the EU numbering, of the amino acids that form the Fc domain of an IgG3 antibody:

(k) F241A;

(m) D265A or

(n) V264A. Each number indicates the position of the amino acid residue according to the EU numbering; and the one-letter amino acid symbol that precedes the number indicates the amino acid residue before it was changed, and the one-letter amino acid symbol that follows the number indicates the amino acid residue after the change.

In addition, preferred polypeptide complexes are also complexes that contain an Fc domain containing one of the following substitutions, the position of which is determined according to the EU numbering, of the amino acids that form the Fc domain of an IgG4 antibody:

(o) L235A, G237A and E318A;

(p) L235E or

(p) F234A and L235A. Each number indicates the position of the amino acid residue according to the EU numbering; and the one-letter amino acid symbol that precedes the number indicates the amino acid residue before it was changed, and the one-letter amino acid symbol that follows the number indicates the amino acid residue after the change.

Other preferred polypeptide complexes are complexes containing an Fc domain in which any amino acid at position 233, 234, 235, 236, 237, 327, 330 or 331 (EU numbering) in the amino acids forming the Fc domain of an IgG1 antibody , has been replaced by an amino acid which is in the corresponding position according to the EU numbering in the corresponding IgG2 or IgG4.

Preferred polypeptide complexes are also complexes containing an Fc domain in which one or more amino acids at positions 234, 235 and 297 (EU numbering) in the amino acids forming the Fc domain of an IgG1 antibody have been replaced by other amino acids . The invention is not limited to the type of amino acid after the substitution; however, most preferred are polypeptide complexes containing an Fc domain in which one or more amino acids at positions 234, 235 and 297 have been replaced by alanine.

Preferred polypeptide complexes are also complexes containing an Fc domain in which the amino acid at position 265 (EU numbering) in the amino acids forming the Fc domain of an IgG1 antibody is replaced by another amino acid. The invention is not limited to the type of amino acid after the substitution; however, most preferred are polypeptide complexes containing an Fc domain in which the amino acid at position 265 is replaced by alanine.

Fc domain derived from a bispecific antibody

According to the present description, the corresponding Fc domains derived from the bispecific antibody can also be used as an Fc domain with reduced Fcγ receptor binding activity. A bispecific antibody is an antibody that has two different specificities. An IgG-type bispecific antibody can be secreted from a hybridoma (quadroma) obtained by fusing two types of hybridomas producing an IgG antibody (Milstein C. et al., Nature 305, 1983, pp. 537-540).

Alternatively, an IgG-type bispecific antibody can also be secreted by introducing into cells a total of four genes, ie. L-chain and H-chain genes forming two types of IgGs of interest, and their co-expression. However, theoretically, there are ten combinations of H chains and L chains of IgG that can be obtained by such methods. It is difficult to purify IgG consisting of the required combination of H chains and L chains from ten IgG types. In addition, the theoretical amount of secreted IgG with the desired combination is also greatly reduced, so large-scale cultivation is required. This further increases the cost of production.

In such a case, an appropriate amino acid substitution can be introduced into the CH3 domain forming the H chain Fc domain in order to preferentially secrete IgG with a heterologous combination of H chains. In particular, this method can be carried out by replacing an amino acid that has a larger side chain (an overhang (meaning "bulge")) with an amino acid in the CH3 domain of one of the H chains, and replacing an amino acid that has a smaller side chain (cavity (meaning "void")), to an amino acid in the CH3 domain of another H chain, so that the ridge fits into the cavity. This promotes the formation of a heteromeric H chain and simultaneously inhibits the formation of a homomeric H chain (WO 1996/027011; Ridgway J. B. et al., Protein Engineering 9, 1996, pp. 617-621; Merchant A. M. et al., Nature Biotechnology 16, 1998 , pp. 677-681).

On the other hand, with respect to the L chain, the variable region of the L chain is less polymorphic than the variable region of the H chain, and therefore one would expect to obtain a common L chain that may have the ability to bind to both H chains. To achieve efficient expression of bispecific IgG, the introduction of such a common L chain and two H chain genes into cells that express IgG can be used (Nature Biotechnology 16, 1998, pp. 677-681). However, this idea is difficult to implement because the probability of obtaining two types of antibodies containing the same L chain by random selection is low. Thus, a method has been proposed for the selection of a common L-chain, which is characterized by a pronounced ability to bind to any different H-chains (WO 2004/065611).

In addition, there are also known techniques for producing a bispecific antibody by applying methods to control polypeptide association or association of polypeptide-derived heteromeric multimers to an association between two polypeptides that form an Fc domain. In particular, polypeptide association control methods can be used to produce a bispecific antibody (WO 2006/106905) in which the amino acid residues forming the interface between two polypeptides that form an Fc domain are altered to inhibit association between Fc domains having the same sequence. , and ensuring the formation of polypeptide complexes formed by two Fc domains having different sequences.

The above two polypeptides that form an Fc domain derived from a bispecific antibody can be respectively used as a domain including the Fc domain of the present invention. More specifically, these two preferred polypeptides that form an Fc domain include those in which the amino acids at positions 349 and 366 (EU numbering) in the amino acid sequence of one polypeptide are cysteine and tryptophan, respectively, and the amino acids at positions 356, 366, 368 and 407 (EU numbering) in the amino acid sequence of another polypeptide are cysteine, serine, alanine and valine, respectively.

In another embodiment, preferred domains comprising an Fc domain of the present invention comprise two polypeptides that form an Fc domain wherein the amino acid at position 409 (EU numbering) in the amino acid sequence of one polypeptide is aspartic acid , and the amino acid at position 399 (EU numbering) in the amino acid sequence of the other polypeptide is a lysine. In the above embodiment, the amino acid at position 409 may be glutamic acid instead of aspartic acid, and the amino acid at position 399 may be arginine instead of lysine. In addition, aspartic acid at position 360 or 392 can preferably also be used in combination with a lysine at position 399.

In another embodiment, preferred domains comprising an Fc domain of the present invention comprise two polypeptides that form an Fc domain wherein the amino acid at position 370 (EU numbering) in the amino acid sequence of one polypeptide is glutamic acid , and the amino acid at position 357 (EU numbering) in the amino acid sequence of the other polypeptide is lysine.

In a further embodiment, preferred domains comprising an Fc domain of the present invention comprise two polypeptides that form an Fc domain wherein the amino acid at position 439 (EU numbering) in the amino acid sequence of one polypeptide is glutamic acid , and the amino acid at position 356 (EU numbering) in the amino acid sequence of the other polypeptide is a lysine.

In addition, preferred domains, including the Fc domain proposed in the present invention, include options with their combination:

two polypeptides form an Fc domain in which the amino acids at positions 409 and 370 (EU numbering) in the amino acid sequence of one polypeptide are aspartic acid and glutamic acid, respectively, and the amino acids at positions 399 and 357 (EU numbering) in the amino acid sequence of the other the polypeptide is both lysine (in this embodiment, the glutamic acid at position 370 can be replaced by aspartic acid, and the glutamic acid at position 370 can be replaced by aspartic acid at position 392);

two polypeptides form an Fc domain in which the amino acids at positions 409 and 439 (EU numbering) in the amino acid sequence of one polypeptide are aspartic acid and glutamic acid, respectively, and the amino acids at positions 399 and 356 (EU numbering) in the amino acid sequence of the other the polypeptide is both lysine (in this embodiment, the glutamic acid at position 439 can be replaced by aspartic acid at position 360, 392, or 439);

two polypeptides form an Fc domain in which the amino acids at positions 370 and 439 (EU numbering) in the amino acid sequence of one polypeptide are both glutamic acid, and the amino acids at positions 357 and 356 (EU numbering) in the amino acid sequence of the other polypeptide are both a lysine; and

two polypeptides form an Fc domain in which the amino acids at positions 409,370 and 439 (EU numbering) in the amino acid sequence of one polypeptide are aspartic acid, glutamic acid and glutamic acid, respectively, and the amino acids at positions 399, 357 and 356 (EU numbering ) in the amino acid sequence of another polypeptide, all are lysine (in this embodiment, the amino acid at position 370 cannot be replaced with glutamic acid; alternatively, instead of replacing the amino acid at position 370 with glutamic acid, glutamic acid at position 439 can be replaced with aspartic acid or glutamic acid the acid at position 439 can be replaced by aspartic acid at position 392).

In a further embodiment of the invention, preferred domains comprising the Fc domain of the present invention comprise two polypeptides that form an Fc domain wherein the amino acid at position 356 (EU numbering) in the amino acid sequence of one polypeptide is lysine, and the amino acids at positions 435 and 439 (EU numbering) in the amino acid sequence of the other polypeptide are arginine and glutamic acid, respectively.

When the two above polypeptides that form an Fc domain derived from a bispecific antibody are used as the domain comprising the Fc domain of the present invention, the antigen binding domains and/or CD3 binding domains of the present invention can be placed in the desired combination.

Fc domain with reduced C-terminal heterogeneity

According to the present disclosure, Fc domains having improved C-terminal heterogeneity in addition to the above characteristics can be used as Fc domains with reduced Fcγ receptor binding activity. More specifically, the present invention provides Fc domains in which the glycine and lysine at positions 446 and 447 (EU numbering), respectively, in the amino acid sequences of two polypeptides constituting an Fc domain that is derived from IgG1, IgG2, IgG3, or IgG4 are deleted by deletion.

T-cell receptor complex binding domain

As used herein, a "T-cell receptor complex-binding domain" refers to a region of an antibody to a T-cell receptor complex that contains a region that specifically binds to and is complementary to all or part of the T-cell receptor complex. Said T cell receptor complex may be the T cell receptor itself or an adapter molecule which, in combination with the T cell receptor, forms the T cell receptor complex. The preferred adapter is CD3.

T cell receptor binding domain

As used herein, a "T cell receptor binding domain" refers to the region of an antibody to a T cell receptor that contains a region that specifically binds to and is complementary to all or part of the T cell receptor.

You can use the variable region or constant region of the T-cell receptor. However, the preferred epitopes to which the CD3 binding domain binds are epitopes located in the constant region. Constant region sequences include, for example, T-cell receptor α-chain (RefSeq, accession number CAA26636.1; SEQ ID NO: 67), T-cell receptor β-chain (RefSeq, accession number C25777; SEQ ID NO : 68), T cell receptor γ1 chain (RefSeq, accession number A26659; SEQ ID NO: 69), T cell receptor γ2 chain (RefSeq, accession number AAB63312.1; SEQ ID NO: 70) and δ - T-cell receptor chain (RefSeq accession number AAA61033.1; SEQ ID NO: 71).

CD3 binding domain

As used herein, "CD3 binding domain" refers to the region of an anti-CD3 antibody containing a region that specifically binds to and is complementary to all or part of CD3. The CD3 binding domain can be derived from one or more variable domains of an antibody. Preferably, the CD3 binding domain comprises both the light chain variable region (VL) of an anti-CD3 antibody and the heavy chain variable region (VH) of an anti-CD3 antibody. Said preferred CD3 binding domains include, for example, "single chain Fv (scFv)", "single chain antibody", "Fv", "single chain Fv2 (scFv2)", "Fab", and "F(ab')2".

The CD3 binding domain of the present invention can bind to any epitope, as long as the epitope is located within the γ chain, δ chain, or ε chain of the sequence forming human CD3. According to the present invention, preferred CD3 binding domains include domains that contain an anti-CD3 antibody light chain variable region (VL) and an anti-CD3 antibody heavy chain variable region (VH) that bind to an epitope that resides in the extracellular domain.

ε-chains of the human CD3 complex. These preferred CD3 binding domains include domains that comprise the anti-CD3 antibody light chain variable region (VL) and the OKT3 antibody heavy chain variable region (VH) (Proc. Natl. Acad. Sci. USA 77, 1980, pp. 4914-4917 ) or various known anti-CD3 antibodies. In addition, these acceptable CD3 binding domains include domains derived from an anti-CD3 antibody with the desired characteristics, which is obtained by immunizing the desired animal with a human CD3-forming γ-chain, δ-chain or ε-chain using the methods described above. Acceptable anti-CD3 antibodies from which the CD3 binding domain is derived are human antibodies and antibodies humanized by the appropriate method described above. The structures of the γ chain, δ chain and ε chain that form CD3 are shown as the polynucleotide sequences of SEQ ID NO: 27 (NM_000073.2), SEQ ID NO: 29 (NM_000732.4) and SEQ ID NO: 31 ( NM_000733.3), respectively, and in the form of polypeptide sequences SEQ ID NO: 28 (NP_000064.1), SEQ ID NO: 30 (NP_000723.1) and SEQ ID NO: 32 (NP_000724.1), respectively (RefSeq registration numbers are shown in brackets ).

Polypeptide complex

The structure of the polypeptide complex of the present invention is not limited if it contains

(1) an antigen-binding domain;

(2) a domain containing an Fc domain with reduced Fcγ receptor binding activity; and

(3) the T-cell receptor complex binding domain above.

According to the present invention, the preferred T cell receptor complex binding domain is the T cell receptor binding domain or CD3 binding domain. Each of the domains described above can be linked directly through a peptide bond. For example, (1) F(ab')2 is used as an antigen-binding domain, and (2) an Fc domain with reduced Fcγ receptor binding activity is used as a domain that contains an Fc domain with reduced Fcγ receptor binding activity. In this case, when the antigen-binding domain (1) is linked via a peptide bond to the domain that contains the Fc domain (2), the linked polypeptide forms an antibody structure. Said antibody can be obtained by purifying the culture media of the hybridomas described above, or by purifying the culture media of the desired host cells stably bearing a polynucleotide that encodes an antibody-forming polypeptide.

When the CD3 binding domain (3) is connected to an antibody structure, the CD3 binding domain can be connected via a peptide bond to the C-terminus of the constant region of the antibody structure. In another embodiment, the CD3 binding domain is linked via a peptide bond to the N-terminus of the heavy chain variable region or the light chain variable region of an antibody structure. In another embodiment, the CD3 binding domain can be linked via a peptide bond to the C-terminus of the light chain constant region of an antibody structure. The CD3 binding domain to be linked may have any desired structure; however, said acceptable CD3 binding domain includes preferably Fv and more preferably scFv. There are no restrictions on the valence of the CD3 binding domain that binds to the antibody structure. To connect a bivalent CD3 binding domain to an antibody structure, a monovalent CD3 binding domain can be linked via a peptide bond to the respective C-termini of the two Fc domains that form the constant region of the antibody structure. Alternatively, to link the bivalent CD3 binding domain to the antibody structure, a bivalent scFv (ie sc(Fv)2) can be linked via a peptide bond to the C-terminus of one of the two Fc domains. In this case, it is possible to successfully obtain a polypeptide complex in which a divalent scFv (i.e. sc(Fv)2) is linked to the C-terminus of only one of the two Fc domains that form the constant region of the antibody structure using the above Fc domain derived from the bispecific antibody. Alternatively, to link the monovalent CD3 binding domain to the antibody structure, the monovalent scFv can be linked via a peptide bond to the C-terminus of one of the two Fc domains. In this case, it is possible to successfully obtain a polypeptide complex of the present invention, in which a monovalent scFv is linked to the C-terminus of only one of the two Fc domains that form the constant region of the antibody structure, using the above Fc domain derived from a bispecific antibody.

In addition, when the CD3 binding domain (3) is fused via a peptide bond to the C-terminus of the constant region of the antibody structure, suitable polypeptide complexes include those in which the heavy chain Fv forming the CD3-binding domain is fused to the C-terminus of one constant region. a region (CH3 domain) forming an Fc domain and an Fv light chain fragment forming a CD3 binding domain are fused to the C-terminus of another constant region (CH3 domain) forming an Fc domain. In this case, a suitable linker such as Gly-Gly-Gly-Gly-Ser (SEQ ID NO: 7) is inserted to connect the Fv fragment of the heavy chain or light chain to the C-terminus of the constant region (CH3 domain). The number of repeats in the linker is not critical; however, they can be selected from 1-10, preferably 2-8, or more preferably 2-6. In particular, it is possible to insert an acceptable linker in which the number of repeats of [Gly-Gly-Gly-Gly-Ser] (SEQ ID NO: 7) is 1, 2, 3, 4, 5, 6, 7, 8, 9 or 10 .

Alternatively, when a polypeptide complex is prepared in which a heavy chain Fv fragment constituting a CD3 binding domain is fused to the C-terminus of one constant region (CH3 domain) constituting an Fc domain and a light chain Fv fragment of the CD3 binding domain is connected to the C-terminus of another constant region (CH3 domain) that forms the Fc domain, appropriate changes in amino acid residues can be made to increase the association between the Fv fragment of the heavy chain and the Fv fragment of the light chain, which allow the formation of disulfide bridges between Fv -fragment of the heavy chain and Fv-fragment of the light chain.

In another embodiment, when a polypeptide complex of the present invention is prepared wherein a heavy chain Fv fragment forming a CD3 binding domain is fused to the C-terminus of one constant region (CH3 domain) forming an Fc domain, and The Fv fragment of the light chain, which forms the CD3 binding domain, is connected to the C-terminus of another constant region (CH3 domain), which forms the Fc domain, then to increase the association between the Fv fragment of the heavy chain and the Fv fragment of the light chain CH1 - and CL The α-domains can be linked to both the Fv portion of the heavy chain and the Fv portion of the light chain.

According to another embodiment of the invention, in order to connect the bivalent CD3 binding domain to the antibody structure, the monovalent CD3 binding domain can be connected via a peptide bond to the respective C-termini of the two light chain constant regions or the respective N-termini of the light chain variable regions of the structure antibodies. Alternatively, in order to connect the bivalent CD3 binding domain to the antibody structure, the bivalent scFv (i.e. sc(Fv)2) can be linked via a peptide bond to the respective C-termini of the two light chain constant regions or the respective N-termini variable regions of the light chains of the antibody structure. In this case, using the Fc domain described above, derived from a bispecific antibody, it is possible to efficiently obtain polypeptide complexes in which a divalent scFv (i.e., sc(Fv)2) is linked to the C- or N-terminus of one of the two variable regions light chains of the antibody structure. Alternatively, in order to connect the monovalent CD3 binding domain to the antibody structure, the monovalent scFv can be linked via a peptide bond to the C- or N-terminus of one of the two light chain variable regions. In this case, using the above-described Fc domain derived from a bispecific antibody, one can efficiently obtain polypeptide complexes in which the monovalent scFv is linked to the N- or C-terminus of one light chain variable region from the two light chain variable regions of the antibody structure.

In another embodiment of the invention, in order to connect the bivalent CD3 binding domain to the antibody structure, the univalent CD3 binding domain can be connected via a peptide bond to the respective N-termini of the two heavy chain variable regions of the antibody structure. Alternatively, in order to connect a bivalent CD3 binding domain to an antibody structure, a bivalent scFv (ie sc(Fv)2) can be peptide bonded to the N-terminus of one of the two heavy chain variable regions. In this case, using the Fc domain described above, derived from a bispecific antibody, it is possible to efficiently obtain polypeptide complexes in which the divalent scFv (i.e., sc(Fv)2) is linked to the N-terminus of only one of the two heavy chain variable regions. antibody structures. Alternatively, in order to connect the monovalent CD3 binding domain to the antibody structure, the monovalent scFv can be linked via a peptide bond to the N-terminus of one of the two heavy chain variable regions. In this case, using the above-described Fc domain derived from a bispecific antibody, it is possible to efficiently obtain polypeptide complexes in which the monovalent scFv is linked to the N-terminus of one of the two heavy chain variable regions of the antibody structure.

In addition, the above-described polypeptide complex can be obtained by connecting each domain directly through a peptide bond or through peptide linkage through a peptide linker. In this case, as the linker to be used, for example, the linker described above and corresponding peptide-tagged linkers such as His-tag, HA-tag or FLAG-tag can be used. Furthermore, hydrogen bonding, disulfide bonding, covalent bonding, or ionic interaction, or a combination thereof, may preferably be used. For example, the affinity between CH1 and CL of an antibody can be used, or Fc domains derived from a bispecific antibody described above can be used to heteromerically associate Fc domains. In addition, intradomain disulfide bridges may preferably be used as described in the examples.

In another structure of the polypeptide complex of the present invention, for example, a monovalent Fv and a monovalent Fab are preferably used as (1) an antigen binding domain. In this case, the following structure is used. The heavy chain Fv fragment (VH) or the light chain Fv fragment (VL) of a monovalent Fv is linked via a peptide bond to the heavy chain CH1 domain. The heavy chain CH1 domain is linked via a peptide bond to one of (2) two Fc domains with reduced Fcγ receptor binding activity that form the polypeptide complex of the present invention. Another VL or VH fragment of the monovalent Fv is peptide-linked to the light chain CL domain, which is disulfide-linked to the heavy chain CH1 domain. Thus, VH and VL, respectively, connected to the ends of the heavy chain CH1 domain and the light chain CL domain form the binding domain of the antibody. sc(Fv)2, which forms both (1) an antibody binding domain and (3) a CD3 binding domain, can be linked via a peptide bond to the N-terminus of another Fc domain from the two domains described above. In this case, a polypeptide complex that has a structure in which the heavy chain CH1 domain is peptide bonded to one of the two Fc domains forming the polypeptide complex and sc(Fv)2 is peptide bonded to the other Fc domain can be obtain using the Fc domain described above, derived from a bispecific antibody. The polypeptide complex described above can be obtained by connecting each domain directly through a peptide bond or by peptide linkage through a peptide linker. In this case, the linker to be used is, for example, the linkers described above by way of example and the corresponding linkers with a peptide tag, eg a His tag, a HA tag or a FLAG tag.

In another preferred structure of the polypeptide complex of the present invention, as (1) antigen-binding domain is used, for example, also divalent scFv. In one of the variants of the structure, you can also receive a polypeptide complex, in which one of the divalent scFv is connected via a peptide bond through a VH, forming a (3) CD3-binding domain, with one of two (2) Fc domains with reduced Fcγ-binding activity. receptor, and the other divalent scFv is linked via a peptide bond via a VL forming a (3) CD3 binding domain to one of two (2) Fc domains with reduced Fcγ receptor binding activity. In this case, an Fc domain derived from a bispecific antibody as described above can be used. The polypeptide complex described above can be obtained by connecting each domain directly through a peptide bond or by peptide linkage through a peptide linker. In this case, the linker to be used is, for example, the linkers described above by way of example and the corresponding linkers with a peptide tag, eg a His tag, a HA tag or a FLAG tag.

In another version of the structure, which uses a divalent scFv as the (1) antigen-binding domain, a polypeptide complex can be obtained in which one of the divalent scFvs is connected via a peptide bond of the scFv forming the (3) CD3-binding domain to one of the two ( 2) Fc domains with reduced Fcγ receptor binding activity, and another divalent scFv linked via a peptide bond to another (2) Fc domain with reduced Fcγ receptor binding activity. In this case, using the above-described Fc domain derived from a bispecific antibody, a polypeptide complex can be obtained in which the scFv forming the antigen binding domain is connected via a peptide bond through the scFv forming the CD3 binding domain to one of the two Fc domains, forming a polypeptide complex, and scFv forming an antigen-binding domain is connected via a peptide bond to another Fc domain. The polypeptide complex described above can be obtained by connecting each domain directly through a peptide bond or by peptide linkage through a peptide linker. In this case, the linker that can be used is, for example, the linkers described above by way of example and the corresponding linkers with a peptide tag, eg a His tag, a HA tag or a FLAG tag.

In another preferred structure of the polypeptide complex of the present invention, for example, both the antigen-binding domain and the T-cell receptor complex-binding domain are each present in the structure of a monovalent Fab. To create one of the variants of the structure, it is possible to obtain a polypeptide complex in which the Fv fragment of the heavy chain of a monovalent Fab forming an antigen binding domain is connected via a CH1 domain to one of the polypeptides forming an Fc domain, and the Fv fragment of the light chain of the Fab is connected to CL domain, while the Fv heavy chain fragment of a Fab forming a T-cell receptor binding domain is linked via a CH1 domain to another polypeptide forming an Fc domain, and the Fv light chain fragment of a Fab is linked to a CL domain.

To create another variant of the structure, a polypeptide complex can be obtained in which the Fv fragment of the heavy chain of a monovalent Fab forming an antigen binding domain is connected via a CH1 domain to one polypeptide forming an Fc domain, and the Fv fragment of a Fab light chain is connected to a CL domain , and the Fv light chain fragment of a Fab forming a T cell receptor binding domain is linked via a CH1 domain to another polypeptide forming an Fc domain, and the Fv heavy chain fragment of a Fab is linked to a CL domain. Alternatively, a polypeptide complex can also be prepared in which the heavy chain Fv fragment of a monovalent Fab constituting a T cell receptor binding domain is linked via a CH1 domain to one polypeptide constituting an Fc domain, and the light chain Fv fragment of the Fab is conjugated with a CL domain, the light chain Fv fragment of a Fab forming an antigen binding domain is linked via a CH1 domain to another polypeptide forming an Fc domain, and the heavy chain Fv fragment of a Fab is linked to the CL domain.

To create another variant of the structure, a polypeptide complex can be obtained in which the Fv fragment of the heavy chain of a monovalent Fab forming an antigen binding domain is connected via a CH1 domain to one polypeptide forming an Fc domain, and the Fv fragment of a Fab light chain is connected to a CL domain , and the light chain Fv fragment of a Fab forming a T-cell receptor binding domain is linked via a CL domain to another polypeptide forming an Fc domain, and the light chain Fv fragment of a Fab is linked to a CH1 domain. Alternatively, a polypeptide complex can also be prepared in which the heavy chain Fv fragment of a monovalent Fab constituting a T cell receptor binding domain is linked via a CH1 domain to one polypeptide constituting an Fc domain, and the light chain Fv fragment of the Fab is conjugated with a CL domain, and the Fv heavy chain fragment of a Fab forming an antigen binding domain is linked via a CL domain to another polypeptide forming an Fc domain, and the Fv light chain fragment of a Fab is linked to a CH1 domain.

In another embodiment of the structure of the polypeptide complex of the present invention, in which both the antigen-binding domain and the binding domain of the T-cell receptor complex each have the structure of a monovalent Fab, preferred polypeptides of the present invention include polypeptides that have:

(1) an antigen-binding domain in which an Fv heavy chain fragment of a monovalent Fab structure that binds an antigen is linked via a CH1 domain of one of the above-described polypeptides constituting an Fc domain, and an Fv fragment of the Fab structure is linked to a CL domain; and

(2) a T cell receptor complex binding domain wherein the Fv heavy chain fragment of a monovalent Fab structure that binds to the T cell receptor complex is linked via a CH1 domain to another Fc domain forming polypeptide, and the Fv fragment the light chain of the Fab structure is linked to the CL domain;

and in which the electrical charges of the CH1 and CL domains are controlled such that there is an association of the Fv fragment of the heavy chain of the antigen-binding domain with the Fv fragment of the light chain of the antigen-binding domain or the association of the Fv fragment of the heavy chain of the T-cell receptor binding domain with Fv-fragment of the light chain of the domain that binds the T-cell receptor. In this embodiment, the structure (association controlled structure) of the polypeptide complex is not limited to a particular structure as long as the electrical charges of the CH1 and CL domains are controlled such that the Fv fragment of the heavy chain of the antigen-binding domain is associated with the Fv fragment of the light chain of the antigen-binding domain. domain or association of the Fv fragment of the heavy chain of the T-cell receptor binding domain with the Fv fragment of the light chain of the T-cell receptor binding domain.

To create a variant of the structure with controlled association, a polypeptide complex can be obtained in which the amino acid residues of the CH1 domain linked to the Fv fragment of the heavy chain of the T-cell receptor complex-binding domain have the same electrical charges as the amino acid residues of the CL domain, linked to the Fv light chain fragment of the antigen-binding domain.

To create a variant of the structure with controlled association, it is possible to obtain a polypeptide complex in which the amino acid residues of the CH1 domain linked to the Fv fragment of the heavy chain of the antigen-binding domain have the same electrical charges as the amino acid residues of the CL domain linked to the Fv fragment of the light a chain of the domain that binds the T-cell receptor complex.

To create a variant of the structure with controlled association, a polypeptide complex can be obtained in which the amino acid residues of the CH1 domain linked to the Fv fragment of the heavy chain of the T-cell receptor complex-binding domain have the same electrical charges as the amino acid residues of the CL domain, linked to the Fv fragment of the light chain of the antigen-binding domain, and the amino acid residues of the CH1 domain linked to the Fv fragment of the heavy chain of the antigen-binding domain have the same electrical charges as the amino acid residues of the CL domain linked to the Fv fragment of the light chain of the domain, binding complex of the T-cell receptor.

To create another variant of the structure with controlled association, a polypeptide complex can be obtained in which the amino acid residues of the CH1 domain linked to the Fv fragment of the heavy chain of the T-cell receptor complex-binding domain have the same electric charges as the amino acid residues of the CL domain , linked to the Fv fragment of the light chain of the antigen-binding domain, and the amino acid residues of the CH1 domain linked to the Fv fragment of the heavy chain of the domain binding the T-cell receptor complex, have electrical charges opposite to the amino acid residues of the CL domain linked to the Fv- a light chain fragment of the T-cell receptor-binding domain.

To create another variant of the controlled association structure, a polypeptide complex can be prepared in which the amino acid residues of the CH1 domain linked to the Fv fragment of the heavy chain of the T-cell receptor complex-binding domain have the same electric charges as the amino acid residues of CL- a domain linked to an Fv light chain fragment of an antigen-binding domain; the amino acid residues of the CH1 domain linked to the Fv fragment of the heavy chain of the antigen-binding domain have the same electrical charges as the residues of the CL domain linked to the Fv fragment of the light chain of the T-cell receptor complex-binding domain; the amino acid residues of the CH1 domain linked to the Fv heavy chain fragment of the T cell receptor complex binding domain have electrical charges opposite to those of the amino acid residues of the CL domain linked to the Fv light chain fragment of the T cell receptor binding domain.

To create another variant of the structure with controlled association, a polypeptide complex can be obtained in which the amino acid residues of the CH1 domain linked to the Fv fragment of the heavy chain of the antigen-binding domain have the same electric charges as the amino acid residues of the CL domain linked to Fv- light chain fragment of the T-cell receptor complex-binding domain and the amino acid residues of the CH1 domain linked to the Fv fragment of the heavy chain of the antigen-binding domain have electrical charges opposite to those of the amino acid residues of the CL domain linked to the Fv fragment of the light chain of the antigen-binding domain .

To create an alternative structure with controlled association, a polypeptide complex can be obtained in which the amino acid residues of the CH1 domain linked to the Fv fragment of the heavy chain of the T-cell receptor complex-binding domain have the same electric charges as the amino acid residues of the CL domain linked to the Fv fragment of the light chain of the antigen-binding complex; the amino acid residues of the CH1 domain linked to the heavy chain Fv fragment of the T-cell receptor complex-binding domain have the same electrical charges as the amino acid residues of the CL domain linked to the light chain Fv fragment of the T-cell complex-binding domain. receptor; and the amino acid residues of the CH1 domain linked to the Fv fragment of the heavy chain of the antigen-binding domain have electrical charges opposite to those of the amino acid residues of the CL domain linked to the Fv fragment of the light chain of the antigen-binding domain.

To create another variant of the controlled association structure, a polypeptide complex can be prepared in which the amino acid residues of the CH1 domain linked to the Fv fragment of the heavy chain of the T-cell receptor complex-binding domain have the same electric charges as the amino acid residues of CL- a domain linked to an Fv light chain fragment of an antigen-binding domain; the amino acid residues of the CH1 domain linked to the Fv fragment of the heavy chain of the antigen-binding domain have the same electrical charges as the amino acid residues of the CL domain linked to the Fv fragment of the light chain of the T-cell receptor complex-binding domain; the amino acid residues of the CH1 domain linked to the Fv heavy chain fragment of the T cell receptor complex binding domain have electrical charges opposite to those of the amino acid residues of the CL domain linked to the Fv light chain fragment of the T cell receptor binding domain; and the amino acid residues of the CH1 domain linked to the Fv fragment of the heavy chain of the antigen-binding domain have electrical charges opposite to those of the amino acid residues of the CL domain linked to the Fv fragment of the light chain of the antigen-binding domain.

Control of electric charges of CH1- and CL-domains

In order to produce a bispecific polypeptide complex that recognizes a T cell receptor binding domain epitope through the T cell receptor binding domain heavy and light chains and an antigen epitope through the antigen binding domain heavy and light chains, ten types of polypeptide complex molecules can theoretically be prepared, if, when creating a polypeptide complex, express each of the four chains.

However, the desired polypeptide complex molecule can preferably be obtained, for example, by controlling the domains in such a way as to inhibit the association between the heavy chain of the T cell receptor binding domain and the light chain of the antigen binding domain and/or the association between the heavy chain of the antigen binding domain and the light chain of the domain that binds the T-cell receptor.

Examples are amino acid residue changes that form the interface between heavy chain CH1 of the T-cell receptor binding domain and light chain CL of the antigen binding domain to positively charged amino acid residues, and amino acid residue changes that form the interface between heavy chain CH1 and CL light chain of the T-cell receptor binding domain into negatively charged amino acid residues. As a result of such changes, the undesirable association between the heavy chain CH1 of the T-cell receptor binding domain and the light chain CL of the antigen binding domain is inhibited, since the electrical charges of the amino acid residues that form the interface are positive in both cases, and the undesirable association between the heavy chain CH1 the antigen-binding domain and the light chain CL of the T-cell receptor-binding domain are also inhibited because the electrical charges of the amino acid residues that form the interface are negative in both cases. The desired polypeptide complex of the present invention can be efficiently generated by the required association between the heavy chain CH1 of the T cell receptor binding domain and the light chain CL of the T cell receptor binding domain, as well as by the required association between the heavy chain CH1 an antigen-binding domain; and a light chain CL of the antigen-binding domain. In addition, the required association between the heavy and light chains of the T-cell receptor binding domain is preferably enhanced because the amino acid residues that form the interface have opposite electrical charges to each other. The required association between the heavy and light chains of the antigen-binding domain is also preferably enhanced because the amino acid residues that form the interface have opposite electrical charges to each other. This allows you to effectively create a polypeptide complex, proposed in the present invention, with the desired association.

In addition, the association control of the present invention can also be used to inhibit association between CH1 domains (heavy chains of the T-cell receptor binding domain and antigen-binding domain) or between CL domains (light chains of the T-cell binding domain). receptor and antigen-binding domain).

Those skilled in the art can appropriately select the desired polypeptide complex in which association is controlled according to the present invention by the type of amino acid residues located in close proximity to the CH1/CL interface upon association.

In addition, using public databases or the like. those skilled in the art can appropriately select appropriate CH1 and CL antibody sequences in an organism such as a human, monkey, mouse, or rabbit. More specifically, CH1 and CL amino acid sequence information can be obtained using the methods described in the examples below.

In particular, as demonstrated in the examples described below, specific examples of combinations of amino acid residues located in close proximity (bordering or contacting) at the CH1/CL interface in association between CH1 and CL, which are appropriately linked to VH and VL forming a domain, T-cell receptor binding, or antigen-binding domain, include:

- lysine (K) at position 147 (EU numbering) in CH1 (for example, at position 147 in the amino acid sequence of SEQ ID NO: 1) and threonine (T) at position 180 (EU numbering) in the adjacent (contacting) CL;

- lysine (K) at position 147 (EU numbering) in CH1 and serine (S) at position 131 (EU numbering) in the adjacent (contacting) CL;

- lysine (K) at position 147 (EU numbering) in CH1 and threonine (T) at position 164 (EU numbering) in the adjacent (contacting) CL;

- lysine (K) at position 147 (EU numbering) in CH1 and asparagine (N) at position 138 (EU numbering) in CL, which border (contact) each other;

- lysine (K) at position 147 (EU numbering) in CH1 and glutamic acid (E) at position 123 (EU numbering) in the adjacent CL;

- glutamine (Q) at position 175 (EU numbering) in CH1 and glutamine (Q) at position 160 (EU numbering) in the adjacent (contacting) CL;

or

- lysine (K) at position 213 (EU numbering) in CH1 and glutamic acid (E) at position 123 (EU numbering) in the adjacent CL.

These statements are numbered according to the guidance of Cabot et al. (Kabat E.A. et al., Sequence of Proteins of Immunological Interest. NIH, 1991).

In the context of the present description, the numbers designated using the EU index correspond to the EU numbering (Sequences of proteins of immunological interest, NIH Publication No. 91-3242). According to the present invention, the terms "amino acid residue at position X (EU number)" and "amino acid at position X (EU number)" (where X is an arbitrary number) are used interchangeably with the terms "amino acid residue corresponding to position X (EU number) )”, “amino acid corresponding to position X (EU numbering)”.

As described in the examples below, the desired polypeptide complex can preferably be obtained by substitution of amino acid residues and using the methods proposed in the present invention.

The above amino acid residues are known to be highly conserved in humans and mice (J. Mol. Recognit. 16, 2003, pp. 113-120). Thus, by changing the amino acid residues corresponding to the above amino acid residues, it is also possible to control the association between CH1 and CL in the constant region of the polypeptide complex of the present invention, which differs from the polypeptide complexes described in the examples.

In particular, the present invention provides polypeptide complexes with controlled association between a heavy chain and a light chain, in which one, two or more pairs selected from the group consisting of pairs of amino acid residues specified in subparagraphs (a) to (e) below , have the same electric charges:

(a) amino acid residue at position 147 (EU numbering) in CH1 and amino acid residue at position 180 (EU numbering) in CL;

(b) amino acid residue at position 147 (EU numbering) in CH1 and amino acid residue at position 131 (EU numbering) in CL;

(c) amino acid residue at position 147 (EU number) in CH1 and amino acid residue at position 164 (EU number) in CL;

(d) amino acid residue at position 147 (EU number) in CH1 and amino acid residue at position 138 (EU number) in CL;

(e) the amino acid residue at position 147 (EU numbering) in CH1 and the amino acid residue at position 123 (EU numbering) in CL, and

(e) amino acid residue at position 175 (EU numbering) in CH1 and amino acid residue at position 160 (EU numbering) in CL.

In addition, another embodiment of the present invention are antibodies in which the amino acid residues in the pair of amino acid residues specified in subparagraph (g) below have the same electric charges:

(g) amino acid residue at position 213 (EU numbering) in CH1 and amino acid residue at position 123 (EU numbering) in CL.

The amino acid residues in each of the above pairs are localized in association in close proximity to each other, as described below in the examples. Using homology-based modeling methods or other methods using commercially available programs, those skilled in the art can appropriately identify in the desired CH1 or CL the amino acid positions corresponding to the amino acid residues specified in subparagraphs (a) to (g) above, and can appropriately change the amino acid residues at these positions.

Said "electrically charged amino acid residue" in the antibody described above is preferably selected from, for example, amino acid residues belonging to the following group (X) or (Y):

(X) glutamic acid (E) and aspartic acid (D); and

(Y) lysine (K), arginine (R), histidine (H).

With respect to the polypeptide complexes described above, "having the same electrical charge" means, for example, that two or more amino acid residues all belong to one of the groups (X) and (Y) described above. On the other hand, "has an opposite electrical charge" means, for example, that at least one of the two or more amino acid residues is an amino acid residue belonging to one of the groups (X) and (Y) described above, and the other amino acid residues are amino acid residues that belong to another group.

Preferred embodiments of the present invention are also methods for preparing the above-described polypeptide complexes and methods of the present invention for controlling association by replacing amino acid residues from the groups indicated in subparagraphs (a) to (g) above with amino acid residues that have the same electrical charge.

In the context of the present invention, the "substitutable" amino acid residues are not limited to the above constant region amino acid residues. Using homology-based modeling or other methods based on the use of commercially available programs, those skilled in the art can appropriately identify amino acid residues that form an interface in a mutated polypeptide or heteromeric multimer and appropriately replace amino acid residues at such positions. for the purpose of controlling the association.

In methods for inhibiting unwanted association between a heavy chain and a light chain by introducing repulsive charges at the interface between the heavy chain and light chain variable regions, amino acid residues contacting each other at the interface between the heavy chain variable region (VH) and the light chain variable region chains (VL) are, for example, glutamine (Q) at position 39 (for example, position 39 in the amino acid sequence of SEQ ID NO: 6 described in WO 2006/106905) in the FR2 heavy chain variable region and glutamine (Q) at position 38 (eg, position 44 in the amino acid sequence of SEQ ID NO: 8 described in WO 2006/106905) in the bordering (contacting) variable region of the FR2 light chain. Such preferred amino acid residues are also, for example, leucine (L) at position 45 (e.g., position 45 in the amino acid sequence of SEQ ID NO: 6 described in WO 2006/106905) in the FR2 heavy chain variable region and proline (P) at position 44 (eg, position 44 in the amino acid sequence of SEQ ID NO: 8 as set forth in WO 2006/106905) in the bordering (contact) variable region of the FR2 light chain. These positions are numbered according to the guidelines of Kabat co-authors (Kabat E.A. et al., Sequence of Proteins of Immunological Interest, NIH, 1991).

The above amino acid residues are known to be highly conserved in humans and mice (J. Mol. Recognit. 16, 2003, pp. 113-120). Thus, by changing the amino acid residues corresponding to the above amino acid residues, it is also possible to control the association between the VH and VL variable regions of antibodies other than the polypeptide complexes described in the examples.

More specifically, said antibodies that have heavy chain and light chain variable regions include those in which the following amino acid residues (1) and (2) or (3) and (4) have the same electrical charge:

(1) the amino acid residue corresponding to position 39 (EU numbering) in the variable region of the heavy chain;

(2) the amino acid residue corresponding to position 38 (EU numbering) in the variable region of the light chain;

(3) the amino acid residue corresponding to position 45 (EU numbering) in the variable region of the heavy chain;

(4) amino acid residue corresponding to position 44 (EU numbering) in the variable region of the light chain.

The above amino acid residues (1) and (2) or (3) and (4) are localized in association in close proximity to each other. Through the homology-based modeling method or other methods using commercially available programs, those skilled in the art can appropriately identify in the desired heavy chain or light chain variable region the amino acid positions corresponding to the amino acid residues indicated in (1) to (4) above. ), and can appropriately change the amino acid residues at these positions.

In the above antibody, the "electrically charged amino acid residue" is preferably selected from, for example, amino acid residues belonging to the following group (X) or (Y):

(X) glutamic acid (E) and aspartic acid (D); and

(Y) lysine (K), arginine (R), histidine (H).

In human and mouse antibodies, as a rule, the above amino acid residues (1)-(4) are:

(1) glutamine (Q);

(2) glutamine (Q);

(3) leucine (L) and

(4) proline (P), respectively.

In preferred embodiments of the present invention, the above amino acid residues are altered (eg, replaced with a charged amino acid). The types of amino acid residues (1) to (4) listed above are not limited to the above. These amino acids may be any other amino acids as defined above. For example, with respect to a human antibody, the amino acid corresponding to the amino acid at position 38 (EU numbering) in the variable region of the light chain may be histidine (H). Based on published documents (eg, J. Mol. Recognit. 16, 2003, pp. 113-120) or the like. those skilled in the art can determine the type of amino acid residue that corresponds to any position in the light chain, and therefore can change the amino acid residue accordingly (eg, change to a charged amino acid).

In techniques for inhibiting undesirable association between a heavy chain and a light chain by replacing amino acid residues that form a hydrophobic core at the interface between the heavy chain and light chain variable regions with electrically charged polar amino acids, preferred amino acid residues that have the ability to form a hydrophobic core at the interface between the heavy chain variable region (VH) and the light chain variable region (VL) are leucine (L) at position 45 in the heavy chain variable region and proline (P) at position 44 bordering the light chain variable region.

In general, the term "hydrophobic core" refers to the region in which the side chains of hydrophobic amino acids are combined within an associated polypeptide. Hydrophobic amino acids include, for example, alanine, isoleucine, leucine, methionine, phenylalanine, proline, tryptophan and valine. However, amino acid residues other than these hydrophobic amino acids (eg, tyrosine) may be involved in the formation of the hydrophobic core. In combination with the hydrophilic surface formed by the protruding side chains of hydrophilic amino acids, the hydrophobic core serves as a "driving force" to accelerate the association of water-soluble polypeptides. When hydrophobic amino acids of two different domains are present on the surface of the molecule and are accessible to water molecules, then entropy increases, resulting in an increase in free energy. As a result, the association of the two domains with each other must be such as to reduce the free energy, which is required for stabilization. Hydrophobic amino acids are immersed along the interface into the molecule to form a hydrophobic core.

It is believed that when hydrophobic amino acids that form a hydrophobic core upon association of polypeptides are changed to polar amino acids that carry an electrical charge, the formation of a hydrophobic core is inhibited. This leads to inhibition of the association of polypeptides.

Other known techniques can also be used to generate the polypeptide complexes of the present invention. For example, to enhance the association of the first VH (VH1) and the first VL (VL1) and/or the second VH (VH2) and the second VL (VL2), in addition to the "changes" proposed in the present invention, amino acids in one of the variable regions of the H chain are replaced with amino acids that have a larger side chain (“lump”; “bulge”), and amino acids in the variable region of the other H chain are replaced with amino acids with a smaller side chain (“trough”; “void”), resulting in "protrusion" is placed in the "depression". This enhances the association of VH1 and VL1 and/or VH2 and VL2, resulting in further inhibition of association between VH1 and VL2 polypeptides and/or between VH2 and VL1 (WO 1996/027011; Ridgway J.B. et al., Protein Engineering 9, 1996, ss 617-621; Merchant A. M. et al., Nature Biotechnology 16, 1998, pp. 677-681).

When creating the above-described polypeptide complex, each domain can be connected directly through a peptide bond, or the binding of peptides through a peptide linker. In this case, the linker used may be the linker described above by way of example and the corresponding peptide-tagged linkers, for example, His-tag, HA-tag, myc-tag or FLAG-tag. In addition, hydrogen bonding, disulfide bonding, covalent bonding, or ionic bonding, or a combination thereof, may preferably be used. For example, the affinity between CH1 and CL of an antibody can be used, or the Fc domains described above derived from a bispecific antibody can be used to heteromerically associate Fc domains. In addition, interdomain disulfide bridges can preferably be used as described in the examples.

The polypeptide complexes of the present invention include, for example, the variants shown in FIG. 17, 19 and 24.

The polypeptide complexes proposed in the present invention can be obtained by the same methods as the methods described above for obtaining recombinant antibodies.

In addition, the present invention relates to polynucleotides encoding the polypeptide complex proposed in the present invention. The polypeptide complex of the present invention can be inserted into any expression vectors. To obtain cells expressing the polypeptide complex, the appropriate host is transformed with an expression vector. The polypeptide complex encoded by the polynucleotide can be obtained by culturing cells expressing the polypeptide complex and collecting the expression product from the culture supernatant. In particular, the present invention relates to vectors carrying a polynucleotide that encodes a polypeptide complex of the present invention, cells that contain vectors, and methods for obtaining a polypeptide complex of the present invention, which consists in culturing the cells and collecting the polypeptide complex from culture supernatant. For the purposes described above, the same techniques as described above for recombinant antibodies can be used.

pharmaceutical composition

Another object of the present invention are pharmaceutical compositions that contain as an active substance a polypeptide complex, which includes:

(1) an antigen-binding domain;

(2) a domain containing an Fc domain with reduced Fcγ receptor binding activity, and

(3) CD3 binding domain.

The present invention also relates to therapeutic agents for the induction of cellular cytotoxicity, which contain as an active substance the complex described above (therapeutic agents for the induction of cellular cytotoxicity), agents that inhibit cell growth, and anticancer agents. The pharmaceutical compositions of the present invention can be used as an agent for the treatment or prevention of cancer. According to the present invention, the therapeutic agents for inducing cellular cytotoxicity, the cell growth inhibitory agents, and the anticancer agents of the present invention are preferably administered to individuals suffering from cancer or at risk of cancer recurrence.

According to the present invention, therapeutic agents for the induction of cellular cytotoxicity, agents that inhibit cell growth, and anticancer agents, which contain as an active substance a polypeptide complex, which includes:

(1) an antigen-binding domain;

(2) a domain containing an Fc domain with reduced Fcγ receptor binding activity, and

(3) CD3 binding domain,

may also be considered as a method for preventing or treating cancer, which comprises the step of administering the polypeptide complex to an individual, or the polypeptide complex may be used to prepare therapeutic agents for inducing cellular cytotoxicity, agents that inhibit cell growth, or anticancer agents .

In the context of the present description, the concept "contain as an active substance a polypeptide complex, which includes:

(1) an antigen-binding domain;

(2) a domain containing an Fc domain with reduced Fcγ receptor binding activity, and

(3) CD3 binding domain",

means that the composition contains a polypeptide complex as the main active ingredient; however, the relative content of the polypeptide complex is not specifically stated.

A pharmaceutical composition of the present invention, a therapeutic agent for inducing cellular cytotoxicity, a cell growth inhibitory agent, or an anticancer agent of the present invention can be prepared using various types of polypeptide complexes as needed. For example, the cytotoxic effect on cells expressing an antigen can be enhanced using a mixture of multiple polypeptide complexes of the present invention that bind to the same antigen. Alternatively, the therapeutic effect can be enhanced by preparing a polypeptide complex of the present invention containing an antigen-binding domain that binds to a cancer antigen in combination with other polypeptide complexes of the present invention that contain an antigen-binding domain for another antigen.

If necessary, the polypeptide complexes of the present invention can be encapsulated from microcapsules (microcapsules made from hydroxymethylcellulose, gelatin, poly(methyl methacrylate), etc.), and can be included in the components of colloidal systems intended for drug administration (liposomes, albumin microspheres, microemulsions, nanoparticles and nanocapsules) (for example, see Remington's Pharmaceutical Science 16th ed., ed. Oslo, 1980)). In addition, methods for preparing agents in the form of sustained release agents are known and can be applied to the polypeptide complexes of the present invention (J. Biomed. Mater. Res. 15, 1981, pp. 267-277; Chemtech. 12, 1982 , pp. 98-105; US No. 3773719; published European patents (EP) EP 58481 and EP 133988; Biopolymers 22, 1983, pp. 547-556).

Pharmaceutical compositions, cell growth inhibitors, or anticancer agents of the present invention may be administered to a patient either orally or parenterally. Parenteral administration is preferred. In particular, these administration methods include injection, nasal administration, transpulmonary administration, and transdermal administration. Injections include, for example, intravenous injections, intramuscular injections, intraperitoneal injections, and subcutaneous injections. For example, pharmaceutical compositions, therapeutic agents for inducing cellular cytotoxicity, agents that inhibit cell growth, or anticancer agents of the present invention can be applied topically or systemically by injection. In addition, acceptable methods of administration can be selected depending on the patient's age and symptoms. The dose to be administered can be selected, for example, from the following range: 0.0001 to 1000 mg per kg of body weight per administration. Alternatively, the dose may be selected, for example, from the following range: 0.001 to 100,000 mg/patient. However, the dose of the pharmaceutical composition according to the present invention is not limited to these doses.

Pharmaceutical compositions of the present invention can be prepared using conventional methods (for example, see Remington's Pharmaceutical Science, latest ed., Mark Publishing Company, Easton, U.S.A.) and may also contain pharmaceutically acceptable carriers and additives. Examples of these include, but are not limited to, surfactants, excipients, coloring agents, flavoring agents, preservatives, stabilizers, buffers, suspending agents, isotonicizing agents, binders, disintegrants, lubricants, flow agents and flavoring agents, and may be used. other common media. Specific examples of carriers are light anhydrous silicic acid, lactose, crystalline cellulose, mannitol, starch, calcium carmellose, sodium carmellose, hydroxypropyl cellulose, hydroxypropyl methylcellulose, polyvinyl acetal diethylaminoacetate, polyvinyl pyrrolidone, gelatin, medium chain triglyceride, polyoxyethylene hydrogenated castor oil 60, sucrose, carboxymethyl cellulose, corn starch, inorganic salt, and the like.

The present invention also provides methods for damaging cells expressing a cancer antigen or inhibiting cell growth by contacting cells expressing a cancer antigen with a polypeptide complex of the present invention that binds to a cancer antigen. Monoclonal antibodies that bind to a cancer antigen are described above as the cancer antigen-binding polypeptide complex of the present invention, which is included in the therapeutic agents for inducing cellular cytotoxicity, cell growth inhibitory agents, or anti-cancer agents of the present invention. The invention is not limited to cells to which the cancer antigen-binding polypeptide complex of the present invention binds if they express the cancer antigen. In particular, in the present invention, preferred cancer antigen-expressing cells are ovarian cancer cells, prostate cancer cells, breast cancer cells, uterine cancer cells, liver cancer cells, lung cancer cells, pancreatic cancer cells, gastric cancer cells, urinary cancer cells. bladder and colon cancer cells. If the cancer antigen is GPC3, then the invention is not limited to specific cells as long as they express GPC3. However, preferred cancer cells are hepatocarcinoma cells, lung cancer cells and ovarian cancer cells.

According to the present invention, "contacting" can be carried out, for example, by adding a cancer antigen-binding polypeptide complex of the invention to culture media of cancer antigen-expressing cells that are cultured in vitro . In this case, the added polypeptide complex can be used in a suitable form for this purpose, such as a solution or a solid obtained by lyophilization, or the like. When the polypeptide complex of the present invention is added in the form of an aqueous solution, the solution may be a pure aqueous solution containing only the polypeptide complex, or a solution containing, for example, the above surfactant, excipient, color, flavor, preservative , a stabilizer, a buffer, a suspending agent, an isotonicizing agent, a binder, a leavening agent, a lubricant, a flow agent, and a flavoring agent. The concentration of the added complex is not critical; however, the final concentration in the culture medium is preferably 1 pg/ml to 1 g/ml, more preferably 1 ng/ml to 1 mg/ml, and even more preferably 1 μg/ml to 1 mg/ml.

According to another embodiment of the present invention, "contacting" can also be carried out in vivo by treating non-human animals transplanted with cancer antigen-expressing cells or animals endogenously expressing the cancer antigen. The method of administration may be oral or parenteral. Parenteral administration is preferred. In particular, these administration methods include injection, nasal administration, lung administration, and transdermal administration. Injections include, for example, intravenous injections, intramuscular injections, intraperitoneal injections, and subcutaneous injections. For example, pharmaceutical compositions, therapeutic agents for inducing cellular cytotoxicity, agents that inhibit cell growth, or anticancer agents of the present invention can be applied topically or systemically by injection. In addition, acceptable methods of administration can be selected depending on the age of the animal and symptoms. When the polypeptide complex of the present invention is added in the form of an aqueous solution, the solution may be a pure aqueous solution containing only the polypeptide complex, or a solution containing, for example, the above surfactant, excipient, color, flavor, preservative , a stabilizer, a buffer, a suspending agent, an isotonicizing agent, a binder, a leavening agent, a lubricant, a flow agent, and a flavoring agent. The dose to be administered can be selected, for example, from the following range: 0.0001 to 1000 mg per kg of body weight per administration. Alternatively, the dose may be selected, for example, from the following range: 0.001 to 100,000 mg/patient. However, the dose of the polypeptide complex of the present invention is not limited to these examples.

The methods described below are preferably used as a method for assessing or determining cellular cytotoxicity due to contact of the polypeptide complex of the present invention with antigen-expressing cells to which the antigen-binding domain of the polypeptide complex of the present invention binds. Methods for evaluating or determining cytotoxic activity in vitro include methods for determining the activity of cytotoxic T cells or the like. Whether a polypeptide complex of the present invention has the ability to induce T cell-mediated cellular cytotoxicity can be determined using known methods (see, for example, Current protocols in Immunology, Chapter 7. Immunologic studies in humans, ed. John E, Coligan et al., John Wiley & Sons, Inc., 1993). When performing a cytotoxicity assay, a polypeptide complex whose antigen-binding domain binds to an antigen other than the antigen recognized by the antigen-binding domain of the polypeptide complex of the present invention and which is not expressed on cells is used as a control polypeptide complex. The control polypeptide complex is analyzed in the same way. The activity is then evaluated by determining whether the polypeptide complex of the present invention has more cytotoxic activity than the control polypeptide complex.

Cytotoxic activity in vivo is assessed or determined, for example, using the following procedure. Cells expressing the antigen to which the antigen-binding domain of the polypeptide complex of the present invention binds are transplanted intradermally or subcutaneously into a non-human animal. Then, starting on or after the day of transplantation, the test polypeptide complex is injected into the vein or peritoneal cavity every day or at intervals of several days. Over time, the size of the tumor is determined. The difference in tumor size change can be considered as cytotoxic activity. As in the in vitro assay, a control polypeptide complex is administered. The polypeptide complex of the present invention can be considered to have cytotoxic activity when the size of the tumors is smaller in the group treated with the polypeptide complex of the present invention than in the group treated with the control polypeptide complex.

The MTT method and the measurement of isotopically labeled thymidine uptake into cells are preferably used to evaluate or determine the effect of contact with the polypeptide complex of the present invention on growth inhibition of cells expressing the antigen to which the antigen binding domain of the polypeptide complex binds. While methods similar to those described above for assessing or determining cytotoxic activity in vivo, preferably can be used to assess or determine the ability to inhibit cell growth in vivo .

The present invention also provides kits that can be used in the method of the present invention, which contain a polypeptide complex of the present invention or a polypeptide complex obtained by the method of the present invention. The kits may be packaged with an additional pharmaceutically acceptable carrier or vehicle, or with instructions including instructions for use of the kits, etc.

In addition, the present invention relates to the polypeptide complexes proposed in the present invention, or polypeptide complexes obtained by the method proposed in the present invention, for use in the method proposed in the present invention.

All prior art documents cited herein are incorporated herein by reference.

Examples

Below, the present invention is described with reference to specific examples, which do not limit the scope of the invention.

Example 1. Construction and study of GPC3 ERY2

(1) Basic principles

There is a well-known method for prolonging the blood half-life of an administered protein in vivo , which is based on FcRn-mediated recycling using a protein of interest conjugated to the Fc domain of an antibody. However, conjugation of natural-type Fc to BiTE can lead to the induction of different cytokines, since the same molecule can bind to a T cell through the anti-CD3 scFv fragment of its BiTE residue and simultaneously bind to FcgR (Fcγ receptor) on the cell membrane. , for example, NK cells or macrophages, through its Fc domain, and the resulting cross-linking can activate these cells in a cancer antigen-independent manner. Thus, a molecule designated as ERY2 was obtained in which the BiTE antibody is linked via a polypeptide linker to an Fc domain having reduced Fcγ receptor binding activity (silent Fc), and ERY2 activity was evaluated by comparison with BiTE activity. An anti-CD3 chain epsilon scFv was coupled via a short peptide linker to an anti-glypican 3 (GPC3) scFv, which is a GPI-anchored protein known to be highly expressed on liver cancer cells, to generate BiTE to GPC3 (GPC3 BiTE ) (Fig. 17A). It was then coupled with silent Fc to give ERY2 to GPC3 (GPC3 ERY2) (FIG. 17B). In addition, antibodies to GPC3 of the conventional IgG type were generated for comparison. Anti-GPC3 IgG-type antibody was obtained as an antibody with a reduced content of fucose in the sugar chain fragment, i.e. a low fucose antibody known to have increased ADCC activity.

(2) Construction of GPC3 BiTE

PCR amplification using an anti-GPC3 antibody expression vector as template generated cDNAs each encoding either an H chain variable region (anti-GPC3 VH) or an L chain variable region (anti-GPC3 VL). PCR was performed using primers containing the appropriate additional sequences and the above cDNA as templates for constructing a cDNA fragment encoding anti-GPC3 scFv having an amino acid sequence in which anti-GPC3 VH and anti-GPC3 VL are linked via a linker with three Gly-Gly-Gly-Gly-Ser repeats (SEQ ID NO: 7).

In addition, a series of oligonucleotides was prepared, each having a nucleotide sequence encoding a partial sequence of the H chain variable region (M12 VH) or an L chain variable region (M12 VL) of an anti-CD3 antibody (M12), and having complementary sequences at the ends. The oligonucleotides were designed so that they could be linked to each other through regions of complementary sequences using a polymerase reaction to synthesize a polynucleotide corresponding to the H chain variable region (M12 VH) and the L chain variable region (M12 VL). The oligonucleotides were mixed and then linked to each other by PCR to obtain two cDNAs encoding the amino acid sequences of the respective variable regions. PCR was performed using primers containing the appropriate additional sequences and the above cDNA as templates for constructing a cDNA fragment encoding scFv M12 having an amino acid sequence in which M12 VL and M12 VH are linked through a linker having three Gly-Gly- Gly-Gly-Ser (SEQ ID NO: 7).

Then, using PCR using primers containing the appropriate additional sequences and the above cDNA fragments, each of which encoded either anti-GPC3 scFv or M12 scFv, as templates, a cDNA fragment was constructed that encoded an amino acid sequence, in which anti-GPC3 scFv and M12 scFv were connected through a linker consisting of Gly-Gly-Gly-Gly-Ser (SEQ ID NO: 7) and with a His-tag at the C-terminus (eight histidine residues) (sequence SEQ ID NO : 33 without 19 amino acids at the amino terminus).

PCR was carried out using primers containing the appropriate additional sequences, and as a template, a cDNA fragment encoding the amino acid sequence of SEQ ID NO: 33 lacking 19 amino-terminal amino acids, to obtain a cDNA fragment in which the Eco RI cleavable sequence, the Kozak sequence, and a nucleotide sequence encoding a secretory signal sequence was attached to the 5'end of the above cDNA fragment, and a Not I cleavable sequence was attached to the 3'end. The resulting cDNA fragment was digested with Eco RI and Not I and inserted into a mammalian cell expression vector to obtain an expression vector for GPC3 BiTE (SEQ ID NO: 33; the mature sequence lacks the 19 amino-terminal amino acids that serve as a signal sequence) .

The vector was introduced into CHO DG44 cells by electroporation. After serial dilution, the cells were cultured in the presence of 1 mg/ml geneticin to isolate drug-resistant cell lines. The culture supernatant of the resulting cell lines was analyzed by Western blotting using an anti-His tag antibody to select cell lines expressing GPC3 BiTE.

The culture supernatant resulting from large-scale cultivation of this cell line was applied to a column filled with SP Sepharose FF column (GE Healthcare). After washing the column, the fraction containing GPC3 BiTE was eluted using a NaCl concentration gradient. The fraction was applied to a HisTrap HP column (GE Healthcare). After washing the column, the fraction containing GPC3 BiTE was eluted using an imidazole concentration gradient. The fraction was concentrated by ultrafiltration and then the concentrate was applied to a column packed with Superdex 200 (GE Healthcare). Collected only monomer containing GPC3 BiTE fraction to obtain purified GPC3 BiTE.

(3) Construction of GPC3 ERY2

PCR was carried out using primers containing the same corresponding additional sequences as in the method described above, and using a method well known to those skilled in the art, such a site-directed mutagenesis method using the QuikChange Site-Directed Mutagenesis Kit (Stratagene), to obtaining expression vectors into which a polynucleotide is inserted encoding GPC3 ERY2_Hk (SEQ ID NO: 34; mature sequence that does not contain 19 amino-terminal amino acids serving as a signal sequence) or GPC3 ERY2_Hh (SEQ ID NO: 35; mature sequence that does not contain 19 amino-terminal amino acids serving as a signal sequence).

These expression vectors were co-introduced into FreeStyle293-F cell line (Invitrogen) to transiently express GPC3 ERY2. The resulting culture supernatant was applied to a column containing Anti FLAG M2 (Sigma). After washing, the column was eluted using FLAG peptide at a concentration of 0.1 mg/ml (Sigma). The fraction containing GPC3 ERY2 was applied to a HisTrap HP column (GE Healthcare). After washing, the column was eluted using an imidazole concentration gradient. The fraction containing GPC3 ERY2 was concentrated by ultrafiltration and then the concentrate was applied to a column packed with Superdex 200 (GE Healthcare). Collected only monomer containing GPC3 ERY2 fraction to obtain purified GPC3 ERY2.

(4) Construction of anti-GPC3 antibody with low fucose level

An expression vector for an anti-GPC3 antibody (designated humanized GC33 antibody in WO 2006/006693) was introduced into CHO DXB11 GDP-fucose gene-deleted cells (Cancer Sci. 101(10), 2010, pp. 2227-2233) by electroporation. After serial dilution, the cells were cultured in the presence of 0.5 mg/ml of geneticin to select for drug-resistant lines, and cell lines expressing the low-fucose GPC3 antibody were obtained. From the culture supernatant obtained by culturing these cells, an antibody-containing fraction was obtained by conventional purification by affinity chromatography using a Hitrap® column packed with protein A (Pharmacia). The antibody-containing fraction was then purified by gel filtration using Superdex 20026/60 (Pharmacia). A monomeric fraction was collected from the eluate to obtain an anti-GPC3 antibody with a low level of fucose.

(5) Cytotoxicity assay using human peripheral blood mononuclear cells

(5-1) Preparation of Human Peripheral Blood Mononuclear Cell (PBMC) Suspension

Received samples of 50 ml of peripheral blood of healthy volunteers (adults) using syringes pre-filled with 100 μl (1000 units/ml) of heparin solution (Novo-Heparin 5000 units per injection; Novo Nordisk). Peripheral blood was diluted twice with PBS(-), divided into four equal aliquots, and added to Leucosep lymphocyte separation tubes (cat. no. 227290; Greiner bio-one system), which were previously injected with 15 ml of Ficoll-Pak PLUS and centrifuged. After centrifuging the separation tubes (2150 rpm, 10 min, room temperature), the mononuclear cell fraction layer was collected. Cells in the mononuclear cell fraction were washed once with Dulbecco's Modified Eagle's Medium (SIGMA) containing 10% FBS (hereinafter referred to as 10% FBS/D-MEM), and then the cell density was adjusted to 4 × 10 ⁶ cells/ml with 10% FBS/D-MEM. The cell suspension thus obtained was used as the human PBMC suspension in subsequent experiments.

(5-2) Analysis of cytotoxic activity

Cytotoxic activity was evaluated based on the degree of inhibition of cell growth using a real-time cell analyzer xCELLigence (Roche Diagnostics). The target cells used were the SK-pca13a cell line, derived from the SK-HEP-1 cell line, which had upregulated the expression of human GPC3. SK-pca13a cells were detached from the plates and plated in an E-Plate 96 (Roche Diagnostics) at 1×10 ⁴ cells/well (100 μl/well). Viable cell analysis was then started using the xCELLigence real-time cell analyzer. The next day, the plate was removed from the xCELLigence real-time cell analyzer and 50 μl of each obtained antibody was added to the plate at various concentrations (0.004, 0.04, 0.4, and 4 nM). After reacting for 15 minutes at room temperature, 50 µl of human PBMC suspension (2×10 ⁵ cells/well) prepared according to the method described in section (5-1) was added. Cells were placed back into the xCELLigence Real Time Cell Analyzer to start viable cell analysis. The reaction was carried out in an atmosphere of 5% gaseous nitrogen dioxide at 37°C. The degree of cell growth inhibition (%) was determined using the formula below using the cell index value determined 72 hours after the addition of human PBMC. The cell index value used for the calculation was standardized in such a way that the cell index value determined immediately before the addition of the antibody was taken as 1.

Degree of cell growth inhibition (%) = (A-B) × 100/(A-1)

A is the average cell index in the well without antibody (target cell and human PBMCs only), and B is the average cell index in each well. The measurement was carried out in triplicate.

The cytotoxic activity of GPC3 BiTE, GPC3 ERY2 and anti-GPC3 IgG-type antibodies were evaluated using human blood derived PBMCs (peripheral blood mononuclear cells) as effector cells. GPC3 BiTE showed very strong activity (FIG. 1). This activity significantly exceeded the activity of the anti-GPC3 antibody with low fucose levels. Thus, GPC3 BiTE can serve as an effective therapeutic agent, exceeding in its activity an IgG-type antibody. On the other hand, the activity of GPC3 ERY2 was not as strong as that of GPC3 BiTE, although it also exceeded the activity of the anti-GPC3 IgG-type antibody. This suggests that adding only Fc to BiTE does not lead to the creation of the desired molecule.

Example 2: Construction and evaluation of GPC3 ERY5, GPC3 ERY6 and GPC3 ERY7

Further, in order to improve specific activity, the cancer agent binding domain (GPC3) was divalently converted to increase cancer cell binding activity. Another anti-GPC3 scFv was added to GPC3 ERY2, creating GPC3 ERY5 (FIG. 17D). In addition, a GPC3 Fab-type binding domain was added instead of scFv, creating GPC3 ERY7 (FIG. 17E). In addition, GPC3 ERY6 was also constructed (FIG. 17E) in which scFv to the CD3 GPC3 ERY5 epsilon chain was split into two arms.

Specifically, to obtain a series of expression vectors into which a polynucleotide encoding GPC3 ERY5_Hh, GPC3 ERY6_Hk, GPC3 ERY6_Hh, GPC3 ERY7_Hh, or GPC3 ERY7_L was inserted, a method known to those skilled in the art, such as PCR using primers containing the same corresponding additional sequences, was used. , as in the method described above.

For transient expression of each engineered molecule, the following combinations of expression vectors were introduced into FreeStyle293-F cells:

A. Engineered molecule: GPC3 ERY5

Polypeptides encoded by polynucleotides inserted into expression vectors: GPC3 ERY5_Hh (SEQ ID NO: 36; mature sequence that does not contain 19 amino-terminal amino acids serving as a signal sequence) and GPC3 ERY2_Hk.

B. Engineered molecule: GPC3 ERY6

Polypeptides encoded by polynucleotides inserted into expression vectors: GPC3 ERY6_Hk (SEQ ID NO: 37; mature sequence that does not contain 19 amino-terminal amino acids serving as a signal sequence) and GPC3 ERY6_Hh (SEQ ID NO: 38; mature sequence that does not contain 19 amino-terminal amino acids serving as a signal sequence).

B. Engineered molecule: GPC3 ERY7

Polypeptides encoded by polynucleotides inserted into expression vectors: GPC3 ERY7_Hh (SEQ ID NO: 39; mature sequence that does not contain 19 amino-terminal amino acids serving as a signal sequence), GPC3 ERY7_L (SEQ ID NO: 40; mature sequence that does not contain 19 amino-terminal amino acids serving as a signal sequence) and GPC3 ERY2_Hk.

The resulting culture supernatant was loaded onto an Anti FLAG M2 column (Sigma). After washing, the column was eluted with 0.1 mg/ml FLAG peptide (Sigma). The fraction containing the generated molecule was loaded onto a HisTrap HP column (GE Healthcare). After washing, the column was eluted with an imidazole concentration gradient. The fraction containing the generated molecule was concentrated by ultrafiltration. The fraction was then applied to a Superdex 200 column (GE Healthcare). Only the monomer fraction was collected from the eluate, yielding each purified molecule created.

These polypeptide complexes were compared with GPC3 BiTE in terms of cytotoxic activity. As a result, it was demonstrated that the cytotoxic activity of these polypeptide complexes was not as high as that of GPC3 BiTE (FIGS. 2-4). This result suggests that adding Fc to a BiTE structure or structure mimicking the design and having a configuration that allows bivalent binding to the cancer antigen does not result in the desired molecule.

Example 3: Construction and evaluation of GPC3 ERY8-2, GPC3 ERY9-1 and GPC3 ERY10-1

(1) Construction of GPC3 ERY8-2, GPC3 ERY9-1 and GPC3 ERY10-1

Next, molecules were created that did not have the BiTE structure, but had the required activity. IgG to cancer antigen (GPC3) was used as a basis and a molecule was created in which scFv to the CD3 epsilon chain was added to this scaffold. The IgG Fc used as a scaffold, as in the cases described above, was a silent Fc with reduced FcgR (Fcγ receptor) binding activity. GPC3 ERY8-2 (FIG. 17G), GPC3 ERY10-1 (FIG. 17I) and GPC3 ERY9-1 (FIG. 17H) were constructed in which scFv to the CD3 epsilon chain was attached to the N-terminus of the H chain, C- end of the H-chain and C-terminus of the L-chain of the anti-GPC3 IgG-type antibody, respectively.

Specifically, a method known to those skilled in the art, such as PCR using primers containing the same appropriate additional sequences as in the method described above, was used to obtain a series of expression vectors into which a polynucleotide encoding GPC3 ERY8-2_Hk (SEQ ID NO: 41; mature sequence that does not contain 19 amino-terminal amino acids serving as a signal sequence), GPC3 ERY8-2_Hh (SEQ ID NO: 42; mature sequence that does not contain 19 amino-terminal amino acids serving as a signal sequence) , GPC3 ERY9-1_H (SEQ ID NO: 43; mature sequence that does not contain 19 amino-terminal amino acids serving as a signal sequence), GPC3 ERY9-1_L-His (SEQ ID NO: 44; mature sequence that does not contain 19 amino terminus serving as a signal sequence), GPC3 ERY9-1_L-FLAG (SEQ ID NO: 45; mature sequence which does not contain 19 amino-terminal amino acids serving as a signal sequence) or GPC3 ERY10-1_Hh (SEQ ID NO: 46; a mature sequence that does not contain the 19 amino-terminal amino acids serving as a signal sequence).

For short-term expression of each generated molecule, the following combinations of expression vectors were introduced into FreeStyle293-F cells.

D. Engineered molecule: GPC3 ERY8-2

Polypeptides encoded by polynucleotides inserted into expression vectors: GPC3 ERY8-2_Hk (SEQ ID NO: 41; mature sequence that does not contain 19 amino-terminal amino acids serving as a signal sequence), GPC3 ERY8-2_Hh (SEQ ID NO: 42; mature sequence that does not contain 19 amino-terminal amino acids serving as a signal sequence) and GPC3 ERY7_L.

E. Engineered molecule: GPC3 ERY9-1

Polypeptides encoded by polynucleotides inserted into expression vectors: GPC3 ERY9-1_H (SEQ ID NO: 43; mature sequence that does not contain 19 amino-terminal amino acids serving as a signal sequence), GPC3 ERY9-1_L-His (SEQ ID NO: 44; mature sequence that lacks 19 amino-terminal amino acids serving as a signal sequence) and GPC3 ERY9-1_L-FLAG (SEQ ID NO: 45; mature sequence that lacks 19 amino-terminal amino acids serving as a signal sequence) .

E. Engineered molecule: GPC3 ERY10-1

Polypeptides encoded by polynucleotides inserted into expression vectors: GPC3 ERY10-1_Hh (SEQ ID NO: 46; mature sequence that does not contain 19 amino-terminal amino acids serving as a signal sequence) and GPC3 ERY8-2_Hk, GPC3 ERY7_L.

The resulting culture supernatant was loaded onto an Anti FLAG M2 column (Sigma). After washing, the column was eluted with 0.1 mg/ml FLAG peptide (Sigma). The fraction containing the generated molecule was loaded onto a HisTrap HP column (GE Healthcare). After washing, the column was eluted with an imidazole concentration gradient. The fraction containing the generated molecule was concentrated by ultrafiltration. The fraction was then applied to a Superdex 200 column (GE Healthcare). Only the monomer fraction was collected from the eluate, yielding each purified molecule created.

These molecules were evaluated for in vitro cytotoxic activity. The result indicates that all molecules had cytotoxic activity comparable to or greater than that of GPC3 BiTE (FIG. 5). In particular, it was found that GPC3 ERY9-1 and GPC3 ERY10-1 had a markedly higher cytotoxic activity than GPC3 BiTE. The present invention demonstrates for the first time that molecules created by adding scFv to the anti-cancer antigen CD3 epsilon chain had cytotoxic activity comparable to or greater than that of BiTE. In particular, the invention unexpectedly found that molecules such as GPC3 ERY9-1 and GPC3 ERY10-1 had a markedly higher cytotoxic activity compared to BiTE, despite the large distance between their cancer antigen binding domain and epsilon binding domain. -CD3 chains.

(2) Performance evaluation in vivo GPC3 ERY8-2 and GPC3 ERY10-1:

The in vivo potency of GPC3 ERY8-2 and GPC3 ERY10-1, which demonstrated cytotoxic activity comparable to or greater than GPC3 BiTE when assessed using the in vitro assay described in section (1), was determined. GPC3-expressing cells from the human lung cancer cell line PC-10 were mixed with human PBMCs and then transplanted into severe combined immunodeficient NOD mice. Mice were treated with GPC3 ERY8-2 or GPC3 ERY10-1 (referred to as the "pre-mix" model).

In particular, the evaluation of the effectiveness of GPC3 ERY8-2 on the "pre-mix" model using PC-10 was carried out as follows. PBMCs were isolated from the blood of healthy donors. NK cells were isolated from PBMCs using CD56 MicroBeads for human blood separation (MCAS Miltenyi biotec). The cell line of human squamous cell carcinoma of the lung PC-10 (Immuno-Biological Laboratories Co., Ltd.) (5 × 10 ⁶ cells), human PBMC without NK cells (4.5 × 10 ⁶ cells) and (artificial) basal Matrigel Matrix membrane (BD) and then transplanted subcutaneously into the inguinal region of NOD severe combined immunodeficiency mice (CLEA Japan Inc.; 7-week-old females). The day of transplantation was designated as day 0. On the day before transplantation, anti-asialo-ganglioside M1 (asialo-GM1) antibody (Wako Pure Chemical Industries) was injected intraperitoneally at 0.2 mg/individual. 2 hours after transplantation, GPC ERY8-2 was administered intraperitoneally at a rate of 30 µg/individual. GPC ERY8-2 was administered a total of five times during the period from 0 to 4 days.

In addition, the evaluation of the effectiveness of GPC3 ERY10-1 on the "pre-mix" model using PC-10 was carried out as follows. PBMCs were isolated from the blood of healthy donors. NK cells were isolated from PBMCs using CD56 MicroBeads for human blood separation (MCAS Miltenyi biotec). The cell line of human squamous cell carcinoma of the lung PC-10 (Immuno-Biological Laboratories Co., Ltd.) (5 × 10 ⁶ cells), human PBMC without NK cells (4.5 × 10 ⁶ cells) and Matrigel Matrix basement membrane were mixed (BD) and then transplanted subcutaneously into the groin of NOD severe combined immunodeficiency mice (CLEA Japan Inc.; 7-week-old females). The day of transplantation was designated as day 0. On the day before transplantation, mice were intraperitoneally injected with anti-asialo-GM1 antibody (Wako Pure Chemical Industries) at a rate of 0.2 mg/individual. 2 hours after transplantation, GPC ERY10-1 was administered intraperitoneally at a rate of 30 μg/individual. GPC ERY10-1 was administered a total of 13 times during the time periods from 0 to 4 days, from 7 to 11 days and from 14 to 16 days.

The result demonstrates that in the groups treated with GPC3 ERY8-2 and GPC3 ERY10-1, tumor growth was significantly suppressed compared to the solvent-treated (PBS) group (FIGS. 6 and 7).

In addition, the in vivo efficacy of GPC3 ERY10-1 was also evaluated using an alternative model. In particular, T cells were grown by culturing human PBMCs in vitro and then introduced into severe combined immunodeficient NOD mice that had advanced tumors that had developed after PC-10 transplantation. Mice were treated with GPC3 ERY10-1 (designated T cell transfer model).

In particular, the evaluation of the effectiveness of GPC3 ERY10-1 based on the T-cell transfer model using the PC-10 cell line was carried out as follows. Culture expansion of T cells was carried out using a human cell activation/expansion kit (MACS Miltenyi biotec) and PBMC isolated from the blood of healthy volunteers. A PC-10 human squamous cell lung carcinoma cell line (Immuno-Biological Laboratories Co., Ltd.) (1 x 10 ⁷ cells) was mixed with Matrigel Matrix basement membrane (BD) and then transplanted subcutaneously into the groin of severe combined immunodeficient mice. NOD lines (CLEA Japan Inc.; 7 week old females). The day of transplantation was designated as day 0. Day 1 before transplantation and on days 6, 8, 12, 16 and 20, mice were intraperitoneally injected with anti-asialo-GM1 antibody (Wako Pure Chemical Industries) at a rate of 0.2 mg/individual. On day 6 after transplantation, mice were grouped by tumor size and body weight, and then T cells obtained by cultural expansion as described above were transplanted at 1 x 10 ⁷ cells/individual into the abdominal cavity. 2 hours after transplantation, GPC ERY10-1 was administered intraperitoneally at a rate of 30 μg/individual. GPC ERY10-1 was administered a total of 5 times on days 7, 8, 12, 16 and 17.

The result indicates that when using this model, the GPC3 ERY10-1 treated group also had a pronounced antitumor effect compared to the solvent treated group (PBS) (Fig. 8).

The data described above demonstrate that a series of molecules in which a single anti-CD3 chain epsilon scFv is added to an IgG scaffold having a silent Fc has a pronounced antitumor activity in vivo .

(3) Assessment of retention (retention) in plasma

To resolve the issue of whether molecules such as GPC3 ERY8-2, GPC3 ERY9-1 and GPC3 ERY10-1 have a significantly longer plasma half-life compared to GPC3 BiTE, GPC3 ERY9-1 and GPC3 ERY10-1 were administered. at the rate of 30 μg/individual to mice with severe combined immunodeficiency of the NOD line, which were not transplanted with cancer cells, and their plasma concentrations were determined over time.

Specifically, a pharmacokinetic (PK) analysis was performed as follows. GPC3 ERY9-1 and GPC3 ERY10-1 were intraperitoneally administered to severe combined immunodeficient NOD NOD mice (CLEA Japan Inc.; 8-week-old females) at a rate of 30 μg/individual. Blood samples were collected from the buccal vein of mice using a hematocrit capillary (Terumo) at 15 minutes, 2 hours, 1 day, 2 days and 7 days after administration. Plasma was obtained from blood.

ERY9-1 GPC3 and ERY10-1 GPC3 were appropriately diluted and added to GPC3-expressing Ba/F3 (GPC3/BaF) cells or Ba/F3 (CD3/BaF) human CD3 epsilon chain-expressing cells to allow GPC3 ERY9- 1 or GPC3 ERY10-1 with GPC3/BaF and CD3/BaF. After washing these cells, a FITC-labeled secondary antibody was added for additional interaction. After washing said cells, the fluorescence intensity of the label on the cells was assessed using an Epics XL flow cytometer (Beckman Coulter) to obtain a calibration curve for each antibody.

During a certain period of time, blood samples were collected from mice that were treated with GPC3 ERY9-1 or GPC3 ERY10-1. Plasma was obtained from the blood and diluted accordingly. Using a method similar to that described above, which was used to obtain calibration curves, plasma samples were subjected to interaction with GPC3/BaF or CD3/BaF, determining the amount of GPC3 ERY9-1 and GPC3 ERY10-1 in plasma associated with each cell. Based on the measured values, the plasma concentration of each antibody was calculated and calibration curves were generated according to the method described above.

The result indicates that the concentration of both ERY9-1 GPC3 and ERY10-1 GPC3 in blood was maintained at a level exceeding 10nM 2 days after administration (FIGS. 9 and 10). These data indicate that molecules such as GPC3 ERY9-1 and GPC3 ERY10-1 have a significantly longer plasma half-life compared to BiTE.

(4) Effect of Silent Fc on Cancer Antigen Independent Cytokine Induction

(4-1) Construction of GPC3 ERY15-1 having an FcgR-binding Fc

To address the question of whether molecules such as GPC3 ERY8-2, GPC3 ERY9-1, and GPC3 ERY10-1 can cause cancer antigen-independent cytokine induction, GPC3 ERY15-1 having an FcgR-binding Fc was engineered (Fig. 17K ).

Specifically, similarly to the method described above, PCR was used using primers containing appropriate additional sequences, and a method well known to those skilled in the art, such as the method using QuikChange Site-Directed Mutagenesis Kit (Stratagene) to generate expression vectors into which a polynucleotide is inserted encoding GPC3 ERY15-1_Hh (SEQ ID NO: 47; mature sequence that does not contain 19 amino-terminal amino acids serving as a signal sequence) or GPC3 ERY15-1_Hk (SEQ ID NO: 48; mature a sequence that does not contain 19 amino acids at the amino terminus serving as a signal sequence).

Expression vectors for GPC3 ERY15-1_Hh (SEQ ID NO: 47; mature sequence that lacks 19 amino-terminal amino acids serving as signal sequence), GPC3 ERY15-1_Hk (SEQ ID NO: 48; mature sequence that lacks 19 amino-terminal amino acids serving as a signal sequence) and ERY7_L GPC3 were co-introduced into FreeStyle293-F cells to transiently express ERY15-1 GPC3. The resulting culture supernatant was loaded onto an Anti FLAG M2 column (Sigma). After washing, the column was eluted with 0.1 mg/ml FLAG peptide (Sigma). The fraction containing GPC3 ERY15-1 was applied to a HisTrap HP column (GE Healthcare). After washing, the column was eluted with an imidazole concentration gradient. The fraction containing the generated molecule was concentrated by ultrafiltration. The fraction containing GPC3 ERY15-1 was then loaded onto a Superdex 200 column (GE Healthcare). From the eluate, only the GPC3 ERY15-1 monomer fraction was collected, yielding a purified GPC3 ERY15-1 molecule.

(4-2) Analysis of the ability to cause cancer antigen-independent cytokine induction

The ability of GPC3 ERY15-1 to cause cancer antigen-independent cytokine induction was compared with the ability of GPC3 BiTE, GPC3 ERY9-1, GPC3 ERY10-1 and catumaxomab. Using the method described above, PBMC were obtained from the blood of healthy volunteers. 50 µl of each antibody, adjusted to 40 nM, was added to 50 µl of human PBMC suspension (2 × 10 ⁵ cells/well), and then 100 µl of 10% FBS/D-MEM was added to them. The reaction mixture was incubated in an atmosphere containing 5% nitrogen dioxide gas at 37°C. After incubation for 72 hours, the culture supernatant was collected and the cytokines that were secreted into the culture supernatant were quantified using C ytometric Beads Array (CBA) technology using a human Th1/Th2/ Th17 kit (BD). The analysis was carried out in triplicate using the method specified in the attached protocol.

It was found that GPC3 ERY15-1 and catumaxomab, which have an FcgR-binding Fc, had a pronounced ability to induce cytokines. In contrast, the GPC3 BiTE antibody, which lacked Fc, and GPC3 ERY9-1 and GPC3 ERY10-1, which carried a silent Fc, were not found to induce cytokines (FIG. 11). This result suggests that molecules that have a silent Fc, such as GPC3 ERY8-2, GPC3 ERY9-1 and GPC3 ERY10-1, are highly safe molecules that do not have the ability to induce cytokines in a cancer antigen-independent manner.

Example 4 Design and evaluation of GPC3 ERY18 L1, L2, L3, L4 and S1

Molecules having a CD3 binding domain different in structure from scFv were evaluated. GPC3 ERY18 (Fig. 17K) was constructed in which the VH and VL domains of an anti-CD3 antibody were fused to the C-terminus of two IgG H chains against the GPC3 cancer antigen. With this design, a series of molecules (GPC3 ERY18 L1, L2, L3 and L4) were created that had 1-4 linker units (Gly-Gly-Gly-Gly-Ser) at the junction. Simultaneously, another molecule (GPC3 ERY18 S1) was constructed, in which the amino acids at the appropriate positions were replaced by Cys, which allowed the introduction of a disulfide bridge.

In particular, a method known in the art, such as PCR using primers containing, as in the method described above, the appropriate additional sequences, was carried out to create a series of expression vectors into which a polynucleotide encoding GPC3 ERY18 L1_Hh (SEQ ID NO : 49; mature sequence that does not contain 19 amino acids serving as a signal sequence at the amino terminus), GPC3 ERY18 L1_Hk (SEQ ID NO: 50; mature sequence that does not contain 19 amino acids serving as a signal sequence at the amino terminus), GPC3 ERY18 L2_Hh (SEQ ID NO: 51; mature sequence that does not contain 19 amino-terminal amino acids serving as a signal sequence), GPC3 ERY18 L2_Hk (SEQ ID NO: 52; mature sequence that does not contain 19 amino-terminal amino acids serving as as a signal sequence), GPC3 ERY18 L3_Hh (SEQ ID NO: 53; mature sequence which th does not contain 19 amino acids serving as a signal sequence at the amino terminus), GPC3 ERY18 L3_Hk (SEQ ID NO: 54; mature sequence that lacks amino-terminal 19 amino acids serving as signal sequence), GPC3 ERY18 L4_Hh (SEQ ID NO: 55; mature sequence that lacks amino-terminal 19 amino acids serving as signal sequence), GPC3 ERY18 L4_Hk ( SEQ ID NO: 56; mature sequence that does not contain 19 amino-terminal amino acids serving as a signal sequence), GPC3 ERY18 S1_Hh (SEQ ID NO: 57; mature sequence that does not contain 19 amino-terminal amino acids serving as a signal sequence ) or GPC3 ERY18 S1_Hk (SEQ ID NO: 58; mature sequence that does not contain 19 amino acids at the amino terminus serving as a signal sequence).

For transient expression of the engineered molecules, the following combinations of expression vectors were introduced into FreeStyle293-F cells.

G. Engineered Molecule: GPC3 ERY18 L1

Expression vectors: GPC3 ERY18 L1_Hh (SEQ ID NO: 49; mature sequence that lacks amino-terminal 19 amino acids serving as a signal sequence), GPC3 ERY18 L1_Hk (SEQ ID NO: 50; mature sequence that lacks amino-terminal 19 amino acids serving as a signal sequence) and GPC3 ERY7 L.

H. Engineered molecule: GPC3 ERY18 L2

Expression vectors: GPC3 ERY18 L2_Hh (SEQ ID NO: 51; mature sequence that lacks amino-terminal 19 amino acids serving as a signal sequence), GPC3 ERY18 L2_Hk (SEQ ID NO: 52; mature sequence that lacks amino-terminal 19 amino acids serving as a signal sequence) and GPC3 ERY7 L.

I. Engineered molecule: GPC3 ERY18 L3

Expression vectors: GPC3 ERY18 L3_Hh (SEQ ID NO: 53; mature sequence that lacks amino-terminal 19 amino acids serving as a signal sequence), GPC3 ERY18 L3_Hk (SEQ ID NO: 54; mature sequence that lacks amino-terminal 19 amino acids serving as a signal sequence) and GPC3 ERY7 L.

J. Engineered molecule: GPC3 ERY18 L4

Expression vectors: GPC3 ERY18 L4_Hh (SEQ ID NO: 55; mature sequence that does not contain 19 amino-terminal amino acids serving as a signal sequence), GPC3 ERY18 L4_Hk (SEQ ID NO: 56; mature sequence that does not contain 19 amino-terminal amino acids serving as a signal sequence) and GPC3 ERY7 L.

L. Engineered molecule: GPC3 ERY18 S1

Expression vectors: GPC3 ERY18 S1_Hh (SEQ ID NO: 57; mature sequence that lacks amino-terminal 19 amino acids serving as a signal sequence), GPC3 ERY18 S1_Hk (SEQ ID NO: 58; mature sequence that lacks amino-terminal 19 amino acids serving as a signal sequence) and GPC3 ERY7 L.

The resulting culture supernatant was loaded onto an Anti FLAG M2 column (Sigma). After washing, the column was eluted with 0.1 mg/ml FLAG peptide (Sigma). The fraction containing the generated molecule was loaded onto a HisTrap HP column (GE Healthcare). After washing, the column was eluted with an imidazole concentration gradient. The fraction containing the generated molecule was concentrated by ultrafiltration. The fraction was then applied to a Superdex 200 column (GE Healthcare). Only the monomer fraction was collected from the eluate, yielding each purified molecule created.

For each of the GPC3 ERY18 L1, GPC3 ERY18L2, GPC3 ERY18L3, GPC3 ERY18L4 and GPC3 ERY18S1 molecules, cytotoxic activity was evaluated in vitro (FIGS. 12 and 13). The result indicates that all molecules except GPC3 ERY18 L1 had activity comparable to that of GPC3 ERY10-1. This result demonstrates that molecules that have a structure different from that of scFv have comparable cytotoxic activity. The structure in which both the VH and VL domains of an anti-CD3 antibody are linked to the C-terminus of the two anti-cancer antigen IgG (GPC3) H chains is likely to be involved in the stabilization of the polypeptide complexes of the present invention.

Example 5 Design and Evaluation of GPC3 ERY19-3

Molecules having a Fab-like CD3 binding domain were then examined. GPC3 ERY19-3 was constructed (FIG. 17M) in which the VH and CH1 domains and the VL and CL domains of an anti-CD3 antibody were all linked to the C-terminus of the two H chains of an anti-cancer antigen (GPC3) IgG-type antibody. . In particular, a method known in the art, such as PCR, was carried out using primers containing, as in the method described above, the appropriate additional sequences, creating expression vectors into which a polynucleotide encoding GPC3 ERY19-3_Hh (SEQ ID NO: 59; a mature sequence that lacks 19 amino acids serving as a signal sequence at the amino terminus) or GPC3 ERY19-3_Hk (SEQ ID NO: 60; a mature sequence that lacks 19 amino acids serving as a signal sequence at the amino terminus).

Expression vectors for GPC3 ERY19-3_Hh (SEQ ID NO: 59; mature sequence that does not contain 19 amino acids serving as a signal sequence at the amino terminus), GPC3 ERY19-3_Hk (SEQ ID NO: 60; mature sequence that does not contain at the amino terminus amino-terminus of 19 amino acids serving as a signal sequence) and ERY7_L GPC3 were co-introduced into FreeStyle293-F cells to transiently express ERY19-3 GPC3. The resulting culture supernatant was applied to a HiTrap rProtein A FF recombinant protein A column (GE Healthcare). After washing, the column was eluted with acid. The fraction containing GPC3 ERY19-3 was concentrated by ultrafiltration and then applied to a Superdex 200 column (GE Healthcare). From the eluate, only the GPC3 ERY19-3 monomer fraction was collected, yielding a purified GPC3 ERY19-3 molecule.

The cytotoxic activity of the GPC3 ERY19-3 molecule was evaluated in vitro . The result demonstrated that the molecule had comparable activity to GPC3 BiTE (FIG. 14). The Fab-like CD3 binding domain appears to be involved in stabilizing the molecules of the polypeptide complexes of the present invention.

Example 6 Preparation of Polypeptide Complexes Using Only a Purification Step on Protein A by Introducing a Mutation in the CH3 Domain of GPC3 ERY 10-1

(1) Basic principles

In GPC3 ERY10-1, obtained according to the method described in example 3, the CH3 domain has a ridge-trough structure. The desired GPC3 ERY10-1 molecule, in which two H chains are heteromerically associated, was purified using two types of affinity-based purification using a His tag and a FLAG tag attached to the C-terminus of each H chain. If the ERY10-1 GPC3 molecule is obtained as a pharmaceutical agent, then first, protein A chromatography is performed on the supernatant of a culture of ERY10-1 GPC3 expressing cells to purify the polypeptide complex having an Fc domain. This step is followed by an additional chromatographic purification step using His-tag affinity chromatography and FLAG-tag affinity chromatography. This increases the cost of the cleaning process. Thus, in the example below, molecular modifications were explored that allow the desired GPC3 ERY10-1 molecule having two heteromerically associated H chains to be purified by protein A chromatography alone without the use of a His tag and a FLAG tag.

In particular, modifications leading to the elimination of protein A binding in one of the two H chains were investigated. As a result of these modifications, in which there was a homomeric association of non-protein A binding H chains, the molecule could not bind to protein A and therefore passed through the protein A column during chromatography. On the other hand, a molecule in which a non-protein A binding H chain has heteromerically associated with a protein A binding H chain and a molecule in which a protein A binding H chain has homomerically associated can be separated by protein A chromatography. based on differences in affinity for protein A. However, in the Fc domain of an antibody, the protein A binding site overlaps with the FcRn binding site, which is critical to the plasma retention time of the antibody. Thus, it is only necessary to selectively reduce protein A binding activity while maintaining FcRn binding activity. The replacement of His at position 435 (EU numbering) with Arg was identified as such a modification. The combination of this mutation with the mutations described in WO 2006/106905 (replacing Asp at position 356 (EU numbering) in one of the H chains with Lys and Lys at position 439 (EU numbering) in the other H chain with Glu), that enhance the heteromeric association of two H chains was evaluated as to whether this could facilitate the purification of polypeptide complexes such as GPC3 ERY10-1 using protein A chromatography alone.

(2) Construction of expression vectors carrying antibody genes and expression of corresponding antibodies

To obtain the variable region of the H chain of the antibody, a gene encoding GC33(2)H (variable region of the H chain of the antibody to human glypican, SEQ ID NO: 61; a mature sequence that does not contain 19 amino acids at the amino terminus serving as a signal sequence) was constructed. ), using a method known to those skilled in the art. Similarly, a gene encoding GC33-k0 (anti-human glypican antibody L chain, SEQ ID NO: 62; mature sequence that does not contain 19 amino-terminal amino acids serving as a signal sequence) was constructed to obtain an antibody L chain, with using a method known to those skilled in the art. The genes described below were then constructed to obtain the antibody H chain constant region using a method known to those skilled in the art.

M. Engineered molecule: LALA-G1d

LALA-G1d (SEQ ID NO: 63; mature sequence which does not contain 19 amino acids at the amino terminus serving as a signal sequence) in which Leu at positions 234 and 235 (EU numbering) are replaced by Ala, Asn at position 297 is replaced on Ala, and Gly and Lys at the C-terminus are removed from the IgG1 sequence.

H. Engineered molecule: LALA-G1d-CD3

LALA-G1d-CD3 (SEQ ID NO: 64; a mature sequence that does not contain the 19 amino-terminal amino acids serving as a signal sequence) in which CD3 scFv (anti-human CD3 antibody H chain variable region and L chain variable region anti-human CD3 antibodies are linked via a polypeptide linker) linked to the C-terminus of LALA-G1d (SEQ ID NO: 63; mature sequence that does not contain 19 amino acids at the amino terminus serving as a signal sequence).

A. Engineered molecule: LALA-G3S3E-G1d

LALA-G3S3E-G1d (SEQ ID NO: 65; a mature sequence that lacks the 19 amino-terminal amino acids serving as a signal sequence) in which His at position 435 (EU numbering) is replaced by Arg and Lys at position 439 ( EU numbering) is replaced by Glu in the sequence LALA-G1d (SEQ ID NO: 63; mature sequence that does not contain 19 amino acids at the amino terminus serving as a signal sequence).

P. Engineered molecule: LALA-S3K-G1d-CD3

LALA-S3K-G1d-CD3 (SEQ ID NO: 66; mature sequence that does not contain 19 amino acids at the amino terminus serving as a signal sequence) in which Asp at position 356 (EU numbering) is replaced by Lys in the sequence LALA- G1d-CD3 (SEQ ID NO: 64; mature sequence that does not contain 19 amino acids at the amino terminus serving as a signal sequence).

The NTA1L and NTA1R genes for the anti-human GPC3 H chains were constructed by adding LALA-G1d-CD3 or LALA-G1d downstream of GC33(2)H, respectively. And the NTA2L and NTA2R genes of the anti-human GPC3 H chains were constructed by adding LALA-S3K-G1d-CD3 or LALA-G3S3E-G1d downstream of GC33(2)H, respectively.

Expression vectors for NTA1L, NTA1R, NTA2L, NTA2R (H chain) and GC33-k0 (L chain) were constructed by inserting each gene into an expression vector suitable for animal cells. These vectors were combined according to the method described below and introduced into FreeStyle293 cells (Invitrogen) using a method known to those skilled in the art to transiently express the polypeptide complexes described below. As indicated below, the polypeptide complexes were designated by the names of the introduced genes, combined in the following order [first H chain/second H chain/L chain].

NTA1L/NTA1R/GC33-k0

NTA2L/NTA2R/GC33-k0

(3) Purification of expressed samples and evaluation of their ability to form heteromeric complexes

The cell supernatant of FreeStyle293 cells (hereinafter referred to as CM) containing the following polypeptide complex was used as a sample.

NTA1L/NTA1R/GC33-k0

NTA2L/NTA2R/GC33-k0

CM was filtered through a 0.22 µm filter and applied to a column packed with rProtein A-Sepharose Fast Flow (GE Healthcare) equilibrated with D-PBS. Wash steps 1 and 2 and elution step 1 were performed using the buffers shown in Table 1. The amount of CM applied was adjusted so that the amount of antibody applied was 20 mg/ml resin. The eluted fractions were collected and analyzed by gel filtration to identify components in the fractions.

Table 1

balance D-PFR. Wash 1 1mM sodium acetate, 150mM NaCl, pH 6.5 Wash 2 0.3mM HCl, 150mM NaCl, pH 3.7 Elution 1 2mM HCl, pH 2.7

The results of the gel filtration analysis of each eluted fraction are shown in FIG. 15 and in Table 2. The indicated values represent the area of the eluted peak in percent. When NTA1L/NTA1R/GC33-k0 or NTA2L/NTA2R/GC33-k0 were expressed, the homomeric anti-GPC3 antibody (NTA1L/GC33-k0 or NTA2L/GC33-k0) could hardly be detected in CM. At the same time, the content of the homomeric antibody to GPC3 (NTA2R/GC33-k0) in CM was only about 2% when NTA2L/NTA2R/GC33-k0 was expressed, while its content in CM was about 76% when NTA1L/NTA1R/ was expressed. GC33-k0. This result demonstrates that when His at position 435 (EU numbering) is changed to Arg, and to ensure efficient formation of a heteromeric molecule containing the corresponding H chains. Asp at position 356 (EU numbering) in the polypeptide sequence of one of the H chains is replaced by Lys, and Lys at position 439 (EU numbering) in the polypeptide sequence of the other H chain is replaced by Glu, then heteromeric polypeptide complexes having the same the same molecular form as ERY10-1 GPC3 can be efficiently purified, reaching a purity of more than 98%, with only one protein A purification step.

table 2

Homomeric anti-CD3 antibody heteromeric antibody Homomeric anti-GPC3 antibody NTA1L/NTA1R/GC33-k0 0.7 23.5 75.8 NTA2L/NTA2R/GC33-k0 - 98.2 1.8

Example 7: Design and evaluation of GPC3 ERY 17-2 and GPC3 ERY 17-3

(1) Construction of GPC3 ERY 17-2 and GPC3 ERY 17-3

Then, a molecule was created by using anti-cancer antigen (GPC3) IgG as a scaffold and replacing one of the Fab with a CD3 epsilon chain binding domain. As in the embodiments described above, the Fc of the IgG scaffold was a silent Fc with reduced FcgR (Fcγ receptor) binding activity. For the CD3 epsilon chain binding domains, the VH and VL domains of the anti-CD3 epsilon chain Fab were altered to generate ERY17-2 GPC3 (Fig. 19A), and the CH1 and CL domains were altered to generate ERY17-3 GPC3 ( Fig. 19B).

In particular, using a method well known to those skilled in the art, such as PCR using primers containing the same appropriate additional sequences as in the method described above, a series of expression vectors were constructed into which a polynucleotide encoding ERY17-2_Hh was inserted. (SEQ ID NO: 73; mature sequence that does not contain 19 amino-terminal amino acids serving as a signal sequence), ERY17-2_L (SEQ ID NO: 74; mature sequence that does not contain 19 amino-terminal amino acids serving as a signal sequence sequence), ERY17-3_Hh (SEQ ID NO: 75; mature sequence that does not contain amino-terminal 19 amino acids serving as a signal sequence) or ERY17-3_L (SEQ ID NO: 76; mature sequence that does not contain amino-terminal 19 amino acids serving as a signal sequence).

For short-term expression of each generated molecule, the following combinations of expression vectors were introduced into FreeStyle293-F cells.

P. Engineered molecule: GPC3 ERY17-2

Polypeptides encoded by polynucleotides inserted into expression vectors: GPC3 ERY8-2_Hk, GPC3 ERY7_L, ERY17-2_Hh (SEQ ID NO: 73; mature sequence that does not contain 19 amino acids at the amino terminus serving as a signal sequence) and ERY17-2_L ( SEQ ID NO: 74; mature sequence that does not contain 19 amino acids at the amino terminus serving as a signal sequence).

C. Engineered molecule: GPC3 ERY17-3

Polypeptides encoded by polynucleotides inserted into expression vectors: GPC3 ERY8-2_Hk, GPC3 ERY7_L, ERY17-3_Hh (SEQ ID NO: 75; mature sequence that does not contain 19 amino acids at the amino terminus serving as a signal sequence) and ERY17-3_L ( SEQ ID NO: 76; mature sequence that does not contain 19 amino acids at the amino terminus serving as a signal sequence).

(2) Purification of GPC3 ERY 17-2 and GPC3 ERY 17-3

The resulting culture supernatant was applied to an Anti FLAG M2 column (Sigma). After washing, the column was eluted with 0.1 mg/ml FLAG peptide (Sigma). The fraction containing the generated molecule was loaded onto a HisTrap HP column (GE Healthcare). After washing, the column was eluted with an imidazole concentration gradient. The fraction containing the generated molecule was concentrated by ultrafiltration. The fraction was then applied to a Superdex 200 column (GE Healthcare). Only the monomeric fraction was collected from the eluate, yielding each purified molecule created.

(3) Cytotoxic activity of GPC3 ERY 17-2 and GPC3 ERY 17-3

For GPC3 ERY 17-2 and GPC3 ERY 17-3, in vitro cytotoxic activity was evaluated (FIG. 20). The result demonstrated that both molecules had a markedly higher cytotoxic activity than GPC3 BiTE. Thus, the present invention demonstrated for the first time that molecules in which anti-cancer antigen IgG was used as a scaffold and one of the Fab was replaced by a CD3 epsilon chain binding domain had a cytotoxic activity comparable to or higher than that of BiTE.

(4) GPC3 ERY17-2 efficacy evaluation based on T-cell transfer model using PC-10

GPC3 ERY17-2, which demonstrated cytotoxic activity comparable to or greater than that of GPC3 BiTE in an in vitro assay , was evaluated for in vivo efficacy based on a T cell transfer model using PC-10 cells. In particular, the evaluation of the effectiveness of GPC3 ERY17-2 based on the model of T-cell transfer using PC-10 cells was carried out as follows. Culture expansion of T cells was carried out using a human cell activation/expansion kit (MACS Miltenyi biotec) and PBMC isolated from the blood of healthy volunteers. A PC-10 human squamous cell lung carcinoma cell line (Immuno-Biological Laboratories Co., Ltd.) (1 x 10 ⁷ cells) was mixed with Matrigel Matrix basement membrane (BD) and then transplanted subcutaneously into the groin of severe combined immunodeficient mice. NOD lines (CLEA Japan Inc.; 7 week old females). The day of transplantation was designated as day 0. Day 1 before transplantation and on days 13, 17, 21 and 25, mice were intraperitoneally injected with anti-asialo-GM1 antibody (Wako Pure Chemical Industries) at a rate of 0.2 mg/individual. On day 13 after transplantation, mice were grouped according to tumor size and body weight. On day 14 after transplantation, T cells obtained by cultural expansion as described above were transplanted at a rate of 3×10 ⁷ cells/individual into the abdominal cavity. 2 hours after transplantation, GPC ERY17-2 was administered intravenously at a rate of 30 µg/individual. GPC ERY10-1 was administered a total of 5 times on days 14, 15, 16, 17 and 18.

The result showed that, using this model, there was also a pronounced anti-cancer effect in the ERY17-2 GPC3 treated group compared to the solvent treated group (FIG. 21).

The data described above demonstrated that molecules in which anticancer IgG was used as a scaffold and one of the Fab was replaced with a CD3 epsilon binding domain had a pronounced anticancer effect in vivo .

Example 8 Design and Evaluation of GPC3 ERY17-2-M20

(1) Construction of GPC3 ERY17-2-M20

Next, a molecule was designed that retains the desired activity even after changes in the CD3 epsilon binding domain. Created GPC3 ERY17-2-M20 (Fig. 19A), which was changed the sequence of the domain that binds the CD3 epsilon chain. Specifically, using an anti-CD3 antibody (M20) expression vector as a template, using a method well known to those skilled in the art, such as PCR using primers containing the same appropriate additional sequences as in the method described above, a series of expression vectors into which a polynucleotide encoding ERY17-2-M20_Hh (SEQ ID NO: 77; mature sequence that does not contain 19 amino acids serving as a signal sequence) or ERY17-2-M20_L (SEQ ID NO : 78; mature sequence that does not contain 19 amino acids at the amino terminus serving as a signal sequence).

(2) Cleaning GPC3 ERY17-2-M20

Expression vectors for GPC3 ERY8-2_Hk, GPC3 ERY7_L, ERY17-2-M20_Hh (SEQ ID NO: 77) and ERY17-2-M20_L (SEQ ID NO: 78) were co-introduced into FreeStyle293-F cells to transiently express GPC3 ERY17-2 -M20. The resulting culture supernatant was filtered through a 0.22 μm filter and then applied to a balanced column packed with rProtein A-Sepharose Fast Flow (GE Healthcare). Purified GPC3 ERY17-2-M20 was obtained after wash steps 1 and 2 and elution steps 1 using the buffers shown in Table 3.

Table 3

Balancing FreeStyle293 Expression Medium (Invitrogen) 1% Pen Strep (Invitrogen) Wash 1 1mM sodium acetate, 150mM NaCl, pH 6.5 Wash 2 0.3mM HCl, 150mM NaCl, pH 3.7 Elution 1 2mM HCl, pH 2.7

(3) Cytotoxic activity of GPC3 ERY17-2-M20

The in vitro cytotoxic activity of ERY17-2-M20 GPC3 was evaluated and the cytotoxic activity was found to be comparable to that of ERY17-2 GPC3 (FIG. 22). These data demonstrate that even molecules with an altered sequence of the CD3 epsilon chain binding domain have comparable cytotoxic activity.

Example 9 Design and evaluation of EpCAM ERY17-2 and EpCAM ERY17-3

(1) Construction of EpCAM ERY17-2 and EpCAM ERY17-3

Molecules were then designed that targeted a different cancer antigen but retained the desired activity. EpCAM ERY17-2 (FIG. 19A), in which the anti-GPC3 Fab in GPC3 ERY17-2 was replaced with an anti-EpCAM Fab, and EpCAM ERY17-3 (Fig. 19B), in which the anti-GPC3 Fab in GPC3 ERY17 were obtained -3 has been replaced with anti-EpCAM Fab. Specifically, using an expression vector for an anti-EpCAM antibody as a template, a series of expression lines were constructed using a method well known to those skilled in the art such as PCR using primers containing the same appropriate additional sequences as in the method described above. vectors into which a polynucleotide encoding EpCAM ERY17_Hk (SEQ ID NO: 79; mature sequence that does not contain the 19 amino-terminal amino acids serving as a signal sequence) or EpCAM ERY17_L (SEQ ID NO: 80; mature sequence that does not contains 19 amino acids at the amino terminus serving as a signal sequence).

For short-term expression of each generated molecule, the following combinations of expression vectors were introduced into FreeStyle293-F cells.

T. Engineered molecule: EpCAM ERY17-2

Polypeptides encoded by polynucleotides inserted into expression vectors: EpCAM ERY17_Hk (SEQ ID NO: 79; mature sequence that does not contain 19 amino-terminal amino acids serving as a signal sequence), EpCAM ERY17_L (SEQ ID NO: 80; mature sequence that does not contain 19 amino acids serving as a signal sequence at the amino terminus), ERY17-2_Hh and ERY17-2_L

U. Engineered molecule: EpCAM ERY17-3

Polypeptides encoded by polynucleotides inserted into expression vectors: EpCAM ERY17_Hk, EpCAM ERY17_L, ERY17-3_Hh and ERY17-3_L.

(2) Cleaning EpCAM ERY17-2 and EpCAM ERY17-3

The resulting culture supernatant was loaded onto an Anti FLAG M2 column (Sigma). After washing, the column was eluted with 0.1 mg/ml FLAG peptide (Sigma). The fraction containing the generated molecule was loaded onto a HisTrap HP column (GE Healthcare). After washing, the column was eluted with an imidazole concentration gradient. The fraction containing the generated molecule was concentrated by ultrafiltration. The fraction was then applied to a Superdex 200 column (GE Healthcare). Only the monomer fraction was collected from the eluate, yielding each purified molecule created.

(3) Cytotoxic activity of EpCAM ERY17-2 and EpCAM ERY17-3

For EpCAM ERY17-2 and EpCAM ERY17-3, in vitro cytotoxic activity was evaluated and both molecules were found to have strong cytotoxic activity (FIG. 23). Thus, the present invention demonstrated that molecules in which anti-cancer antigen IgG was used as a scaffold and one of the Fab was replaced by a CD3 epsilon chain-binding domain had cytotoxic activity even when the type of cancer antigen was changed.

Example 10 Construction and Evaluation of Bispecific Antibodies with Modulated Association at the CH1/CL Interface

(1) Creation of a bispecific antibody

By introducing mutations into each of the CH1 and CL domains of a bispecific antibody and thereby modulating association at the CH1/CL interface using repulsive electrical charges at the CH1/CL interface, conditions can be created for a specific association between H α-chain and L-chain of an anti-GPC3 antibody and between the H-chain and L-chain of an anti-CD3 antibody. In order to modulate association at the CH1/CL interface using electric charge repulsion, amino acid residues in CH1 of H chains or in CL of L chains were changed to Lys, which is a positively charged amino acid, or to Glu, which is a negatively charged amino acid. charged amino acid.

(2) Construction of expression vectors carrying antibody genes and expression of corresponding antibodies

A bispecific antibody (Fig. 24A) was generated by modulating association at the CH1/CL interface of anti-CD3 antibody M12 (H chain, SEQ ID NO: 81; L chain, SEQ ID NO: 82) and anti-GPC3 antibody GC33(2) (H-chain, SEQ ID NO: 83; L-chain, SEQ ID NO: 84) and further introduction of ridge-depression (KiH) modifications therein (WO 1996/027011; Ridgway J.B. et al., Protein Engineering 9, 1996, 617-621, Merchant A.M. et al., Nat Biotechnol 16,1998, 677-681) in order to avoid association of the H chains with each other. A control bispecific antibody was also constructed (FIG. 24B) into which neither association modulation at the CH1/CL interface nor ridge-hole (KiH) modifications were introduced. In particular, using a method known to those skilled in the art, expression vectors were constructed carrying as an insert a polynucleotide encoding M12_TH2h (SEQ ID NO: 85), in which several amino acids in the CH1 H chain of the M12 antibody (SEQ ID NO: 81 ) are replaced by Lys, or M12_TL17 (SEQ ID NO: 86), in which several amino acids in the CL L chain (SEQ ID NO: 82) are replaced by Glu. Similarly, using a method known to those skilled in the art, expression vectors were constructed carrying as an insert a polynucleotide encoding GC33(2)_TH13k (SEQ ID NO: 87) or GC33(2)_TH15k (SEQ ID NO: 88), in which several amino acids in the CH1 H chain of the antibody GC33(2) (SEQ ID NO: 83) are replaced by Glu, or GC33(2)_TL16 (SEQ ID NO: 89) or GC33(2)_TL19 (SEQ ID NO: 90 ), in which several amino acids in the CL L chain (SEQ ID NO: 84) are replaced by Lys.

For short-term expression of each generated molecule, the following combinations of expression vectors were introduced into FreeStyle293-F cells.

F. Engineered molecule: GM1

Expression vectors: M12_TH2h (SEQ ID NO: 85), M12_TL17 (SEQ ID NO: 86), GC33(2)_TH13k (SEQ ID NO: 87) and GC33(2)_TL16 (SEQ ID NO: 89)

X. Engineered: GM2

Expression vectors: M12_TH2h (SEQ ID NO: 85), M12_TL17 (SEQ ID NO: 86), GC33(2)_TH15k (SEQ ID NO: 88) and GC33(2)_TL19 (SEQ ID NO: 90)

C. Engineered molecule: GM0

Expression vectors: M12 H chain (SEQ ID NO: 81), M12 L chain (SEQ ID NO: 82), GC33(2) H chain (SEQ ID NO: 83), and L chain of GC33(2) (SEQ ID NO: 84).

From the resulting cell supernatant, antibodies were purified using a method known to those skilled in the art using an rProtein A Sepharose ^™ Fast Flow column (GE Healthcare).

(3) Cytotoxic activity of GM1, GM2 and GM0

The cytotoxic activity of GM1, GM2, and GM0 polypeptide complexes was evaluated in vitro . The result demonstrates that GM1 and GM2 had comparable cytotoxic activity, and this activity is higher than that of GM0 (FIG. 25). Thus, the present invention demonstrates that the combination of CH1/CL association modulation and KiH modifications allows efficient production of bispecific antibodies.

Example 11 Design and Evaluation of EGFR ERY17-2

(1) Construction of EGFR ERY17-2

In addition, a molecule with the desired activity was created, the target of which is another cancer antigen. EGFR ERY17-2 was constructed (FIG. 19A) by replacing the anti-GPC3 Fab of the GPC3 ERY17-2 antibody with an anti-EGFR Fab. Specifically, using an anti-EGFR antibody expression vector as a template, a series of expression vectors were constructed using a method well known to those skilled in the art such as PCR using primers containing the same appropriate additional sequences as in the method described above. into which a polynucleotide encoding EGFR ERY17_Hk (SEQ ID NO: 91; a mature sequence that does not contain the 19 amino-terminal amino acids serving as a signal sequence) or EGFR ERY17_L (SEQ ID NO: 92: a mature sequence that does not contain at the amino-terminus of 19 amino acids serving as a signal sequence).

For transient expression of each engineered molecule, the following combinations of expression vectors were introduced into FreeStyle293-F cells.

H. Engineered molecule: EGFR ERY17-2

Polypeptides encoded by polynucleotides inserted into expression vectors: EGFR ERY17_Hk (SEQ ID NO: 91: mature sequence which does not contain 19 amino acids serving as a signal sequence at the amino terminus), EGFR ERY17_L (SEQ ID NO: 92: mature sequence which does not contain 19 amino acids serving as a signal sequence at the amino terminus), ERY17-2_Hh and ERY17-2_L.

(2) Purification of EGFR ERY17-2

The resulting culture supernatant was loaded onto an Anti FLAG M2 column (Sigma). After washing, the column was eluted with 0.1 mg/ml FLAG peptide (Sigma). The fraction containing the generated molecule was loaded onto a HisTrap HP column (GE Healthcare). After washing, the column was eluted with an imidazole concentration gradient. The fraction containing the generated molecule was concentrated by ultrafiltration. The fraction was then applied to a Superdex 200 column (GE Healthcare). Only the monomer fraction was collected from the eluate, yielding each purified molecule created.

(3) Cytotoxic activity of EGFR ERY17-2

For EGFR ERY17-2, cytotoxic activity was evaluated in vitro , and in this case, a strong cytotoxic activity was detected (Fig. 26). Thus, the present invention demonstrates that molecules in which anti-cancer antigen IgG served as a scaffold and one of the Fab was replaced by a CD3 epsilon binding domain had cytotoxic activity, even when the cancer antigen representing GPC3 or EpCAM were additionally replaced with another type of antigen.

Industrial Applicability

The present invention provides novel polypeptide complexes that retain the potent anti-cancer activity of BiTE and have a long blood half-life, as well as very high safety, as evidenced by no induction of a cancer antigen-independent cytokine storm or the like. By replacing the antigen-binding domain of the polypeptide complex of the present invention, therapeutic agents that contain the polypeptide complex as an active ingredient to induce cellular cytotoxicity can target and damage various cells, including cancer cells. Thus, various types of cancer can be treated or prevented. This makes it possible to carry out the required treatment, which is highly safe and convenient, and to reduce the physical burden on patients.

Claims (46)

1. A polypeptide complex with cell-mediated cytotoxicity, which includes: (1) a cancer antigen-binding domain, wherein said antigen is not CD3; (2) any of the Fc domains of SEQ ID NOs: 23, 24, 25, and 26 in which a mutation has occurred in the amino acid(s) forming the Fc domain that has reduced binding activity for the human Fcγ receptor compared to a control polypeptide complex comprising the Fc domain of a wild-type IgG1, wild-type IgG2, wild-type IgG3, or wild-type IgG4 monoclonal antibody of the same isotype; and (3) a T cell receptor complex binding domain, wherein the T cell receptor complex binding domain is a CD3 binding domain, wherein both the cancer antigen binding domain and the T cell receptor complex binding domain are , including both the antibody light chain variable region (VL) and the antibody heavy chain variable region (VH), wherein the polypeptide complex is any of the complexes selected from the group consisting of: (i) a polypeptide complex comprising two cancer antigen-binding domains and one or two T-cell receptor complex-binding domains, wherein the cancer antigen-binding domains are individually linked to either of the two polypeptides forming an Fc domain, and the domain α, which binds the T cell receptor complex, is associated with either or both of the CH3 regions forming the Fc domain; (ii) a polypeptide complex comprising two cancer antigen binding domains and one or two T cell receptor complex binding domains, wherein the cancer antigen binding domains are individually linked to either of the two polypeptides forming the Fc domain, and the domain, a T-cell receptor complex-binding complex associated with either or both cancer antigen-binding domains; and (iii) a polypeptide complex that includes one cancer antigen binding domain and one T cell receptor complex binding domain, wherein the cancer antigen binding domain is linked to one of the polypeptides forming the Fc domain and the receptor complex binding domain T-cells, associated with another of the polypeptides that form the Fc-domain. 2. The polypeptide complex of claim 1, wherein the Fc domain is a human Fc domain. 3. A polypeptide complex according to claim 1 or 2, wherein the Fc domain exhibits reduced binding activity to the human Fcγ receptor, FcγI, FcγIIIA and/or FcγIIIB. 4. The polypeptide complex of claim 3, wherein the CD3 binding domain is Fv. 5. The polypeptide complex of claim 3, wherein the CD3 binding domain is a Fab. 6. The polypeptide complex of claim 3, wherein the CD3 binding domain is scFv. 7. The polypeptide complex according to any one of paragraphs. 1-6, wherein the T cell receptor complex binding domain is a monovalent domain. 8. The polypeptide complex of claim 1(iii), wherein electrical charges are controlled such that the VH region of the cancer antigen binding domain is assembled to the VL region of the cancer antigen binding domain or the VH region of the cancer antigen binding domain. T-cell receptor complex, is assembled to the VL region of the T-cell receptor complex-binding domain. 9. The polypeptide complex of claim 8, wherein the amino acid residue in the VH region of the T cell receptor complex binding domain has the same type of electrical charge as the amino acid residue in the VL region of the cancer antigen binding domain. 10. The polypeptide complex of claim 8, wherein the amino acid residue in the VH region of the cancer antigen binding domain has the same type of electrical charge as the amino acid residue in the VL region of the T cell receptor complex binding domain. 11. The polypeptide complex of claim 8, wherein the amino acid residue in the VH region of the T cell receptor complex binding domain has the same electrical charge as the amino acid residue in the VL region of the cancer antigen binding domain and the amino acid residue in The VH region of the cancer antigen binding domain has the same type of electrical charge as the amino acid residue in the VL region of the T cell receptor complex binding domain. 12. A polypeptide complex according to claim 9 or 11, wherein the amino acid residue in the VH region of the T cell receptor complex binding domain has an electrical charge of the opposite type to that of the amino acid residue in the VL region of the T cell receptor complex binding domain. 13. A polypeptide complex according to claim 10 or 11, wherein the amino acid residue in the VH region of the cancer antigen binding domain has an electrical charge of the opposite type to that of the amino acid residue in the VL region of the cancer antigen binding domain. 14. The polypeptide complex according to claim 12 or 13, in which the amino acid residue with the opposite electric charge is selected from amino acid residues of any of the groups: (X) glutamic acid (E) and aspartic acid (D); or (Y) lysine (K), arginine (R) and histidine (H). 15. The polypeptide complex according to any one of paragraphs. 1-14, wherein the Fc domain has reduced Fcγ receptor binding activity for all human Fcγ receptors, FcγI, FcγIIA, FcγIIB, FcγIIIA and FcγIIIB. 16. The polypeptide complex according to any one of paragraphs. 1-15, wherein the Fc domain includes any of the following amino acids: the amino acid sequence at positions 118 - 260 (EU numbering) is the sequence disclosed in SEQ ID NO: 24; or the amino acid sequence at positions 261 - 447 (EU numbering) is the sequence disclosed in SEQ ID NO: 26. 17. The polypeptide complex according to any one of paragraphs. 1-15, in which the amino acids forming the Fc domain include a mutation at any of the following positions: 220, 226, 229, 231, 232, 233, 234, 235, 236, 237, 238, 239, 240, 264, 265 , 266, 267, 269, 270, 295, 296, 297, 298, 299, 300, 325, 327, 328, 329, 330, 331 and 332 (EU numbering). 18. The polypeptide complex according to claim 17, in which the Fc domain includes a mutation of the amino acids forming the Fc domain of the sequence SEQ ID NO: 23. 19. The polypeptide complex of claim 18, wherein the Fc domain is an Fc domain comprising an amino acid substitution at positions 233, 234, 235, 236, 237, 327, 330, or 331 (EU numbering) with an amino acid at the corresponding position (EU numbering). EU) corresponding IgG2 or IgG4. 20. The polypeptide complex of claim 18, wherein the Fc domain includes an amino acid mutation at positions 234, 235, or 297 (EU numbering). 21. A polypeptide complex according to claim 20, wherein the amino acid(s) at position 234, 235 and/or 297 are substituted(s) with alanine. 22. The polypeptide complex according to any one of paragraphs. 1-21, in which the sequences of the two polypeptides forming the Fc domain differ from each other. 23. The polypeptide complex according to any one of paragraphs. 1-22, where among the amino acid residues of one of the two polypeptides forming the Fc domain, the amino acid at position 349 is replaced by cysteine and the amino acid at position 366 (EU numbering) is replaced by tryptophan; and among the amino acid residues of another polypeptide, the amino acid at position 356 is changed to cysteine, the amino acid at position 366 is changed to serine, the amino acid at position 368 is changed to alanine, and the amino acid at position 407 (EU numbering) is changed to valine. 24. The polypeptide complex according to any one of paragraphs. 1-22, where among the amino acid residues of one of the two polypeptides forming the Fc domain, the amino acid at position 356 (EU numbering) is replaced by lysine, in the other polypeptide, the amino acid at position 439 (EU numbering) is replaced by glutamic acid; and among the amino acid residues of either of the two polypeptides, the amino acid at position 435 (EU numbering) is changed to arginine. 25. A polypeptide complex according to claim 23 or 24, wherein the GK sequence is omitted from the C-terminus of the two polypeptides forming the Fc domain. 26. A polypeptide complex according to p. l (i) or 1 (ii), in which the domains that bind cancer antigens bind to the same epitope. 27. A polypeptide complex according to claim 26, wherein the same epitope is present in a protein comprising the amino acid sequence of SEQ ID NO: 2. 28. The polypeptide complex of claim 26, wherein the same epitope is present in a protein comprising the amino acid sequence of SEQ ID NO: 4. 29. The polypeptide complex according to p. l (i) or 1 (ii), where the cancer antigen binding domains bind to different epitopes. 30. The polypeptide complex of claim 29, wherein various epitopes are present in a protein comprising the amino acid sequence of SEQ ID NO: 2. 31. The polypeptide complex of claim 29, wherein various epitopes are present in a protein comprising the amino acid sequence of SEQ ID NO: 4. 32. Polynucleotide encoding a polypeptide complex according to any one of paragraphs. 1-31. 33. The vector for the expression of the polypeptide complex according to any one of paragraphs. 1-31, including the polynucleotide according to claim 32. 34. Cell for obtaining a polypeptide complex according to any one of paragraphs. 1-31, including the vector according to claim 33. 35. The method of obtaining a polypeptide complex according to any one of paragraphs. 1-31, including culturing cells according to claim 34 and isolating the polypeptide complex from the cell culture supernatant. 36. Pharmaceutical composition for the induction of cell-mediated cytotoxicity, containing an effective amount of the active polypeptide complex according to any one of paragraphs. 1-31. 37. The pharmaceutical composition of claim 36, wherein the pharmaceutical composition for inducing cell-mediated cytotoxicity is an anti-cancer pharmaceutical composition. 38. The pharmaceutical composition according to claim 37, wherein the cancer is liver cancer or lung cancer.

Images