RU2812861C1 - Library of concentration-dependent ion binding molecules - Google Patents

Abstract

FIELD: biotechnology.

SUBSTANCE: libraries for screening for an antibody that binds to an antigen of interest in a pH-dependent manner. Libraries are essentially composed of multiple sequence-differing antibodies, where the antigen binding domain in each antibody contains at least one histidine, which modifies the antigen binding activity of the antibody depending on pH conditions. Compositions for antibody screening containing a plurality of polynucleotide molecules encoding antibodies, or a plurality of vectors, each of which contains these polynucleotide molecules are also described. In addition, a method for obtaining a library for screening an antibody, a method for selecting an antibody whose antigen-binding activity varies depending on pH conditions, a method for isolating a polynucleotide encoding the said antibody, and variants of the method for producing the said antibody are disclosed.
EFFECT: libraries for screening an antibody are disclosed.

39 cl, 21 dwg, 55 tbl, 31 ex

Description

TECHNICAL FIELD

RELATED APPLICATION

The present application claims priority to Japanese Patent Application No. 2011-218006 (filed September 30, 2011) and 2012-123479 (filed May 30, 2012), the contents of which are incorporated herein by reference in their entirety.

FIELD OF THE INVENTION

The present invention relates to a library of antigen-binding molecules whose antigen-binding activity varies depending on ion concentration conditions, a method for preparing the library, a method for selecting such antigen-binding molecules, a method for producing such antigen-binding molecules, and a pharmaceutical composition containing such an antigen-binding molecule.

BACKGROUND OF THE ART

Antibodies have attracted attention as pharmaceutical agents due to their high stability in plasma and low adverse reactions. Among other things, many IgG antibody-based drugs are already in production, and a large number of antibody-based drugs are in the development stage (Non-Patent Documents 1 and 2). Meanwhile, various technologies applicable to second-generation antibody drugs have been developed. For example, methods for improving effector functions, antigen binding ability, pharmacokinetics or stability, or reducing the risk of immunogenicity are described (Non-Patent Document 3). Possible problems with such antibody drugs are the difficulty of obtaining drugs for subcutaneous administration (since antibody drugs are usually administered in very high doses), high cost of production, etc. Methods for improving the pharmacokinetics of antibodies and methods for increasing the affinity of antibodies for their antigens can be used to reduce the doses of antibody-based drugs.

As a method for improving the pharmacokinetics of antibodies, artificial substitution of amino acids in constant regions has been described (Non-Patent Documents 4 and 5). The previously described affinity maturation, a method of enhancing antigen-binding ability and antigen-neutralizing ability (Non-Patent Document 6), involves mutation of amino acids, for example, in CDR regions of variable regions, to thereby achieve enhanced antigen-binding activity. Such enhancement of antigen-binding ability may improve biological activity in vitro or reduce dosages and may further improve the effectiveness of the drug in vivo (Non-Patent Document 7).

The amount of antigen that can be neutralized by one antibody molecule depends on affinity. Higher affinity allows the antibody to neutralize less of the antigen. The affinity of an antibody can be increased in various ways (Non-Patent Document 6). An antibody capable of covalently binding to an antigen with unlimited affinity is capable of neutralizing, per molecule, one molecule of antigen (or two antigens in the case of a divalent antibody). However, prior methods have a stoichiometric limitation of the neutralization reaction to one molecule of antigen (or two molecules of antigen in the case of a divalent antibody) per antibody molecule and are unable to completely neutralize the antigen when using an amount of antibody less than the amount of antigen. Briefly, there is a limitation on the affinity enhancing effect (Non-Patent Document 9). The inherent duration of the neutralizing effect of a neutralizing antibody requires administration of the antibody in an amount greater than the amount of antigen produced in vivo over a period of time. The only method mentioned above to improve antibody pharmacokinetics or affinity maturation is not sufficient to reduce the required antibody doses. In this regard, a single antibody must neutralize multiple antigens to maintain the antigen neutralizing effect for a period of time of interest in an amount below the amount of the antigen.

To achieve this objective, an antibody that binds to an antigen in a pH-dependent manner has recently been described as a new approach (Patent Document 1). This document describes that a histidine residue is introduced into an antigen-binding molecule to produce an antibody with pH-dependent antigen binding, the properties of which vary between neutral pH and acidic pH conditions. This pH-dependent antigen-binding antibody binds strongly to antigen under neutral conditions in plasma and dissociates from antigen under acidic conditions in the endosome. Thus, an antibody with pH-dependent antigen binding can dissociate from the antigen in the endosome. The pH-dependent antigen-binding antibody, thus dissociated from the antigen, can bind to the antigen after being recycled back into the plasma via FcRn. This allows one antibody to bind to multiple antigens multiple times.

Antigens have a very short plasma retention time compared to antibodies that are recycled by binding to FcRn. Antibody-antigen complexes in the case of antibodies having a long plasma half-life (long plasma retention time) and those antigens having a short plasma half-life (short plasma retention time) have the same long plasma retention time as and antibodies. Thus, binding of an antigen to an antibody is likely to prolong its retention time in plasma and increase the concentration of antigen in plasma. In this case, even increasing the affinity of the antibody for the antigen cannot accelerate the clearance of the antigen from the plasma. Thus, the pH-dependent antigen binding antibody mentioned above is also more effective as an approach for accelerating the clearance of antigen from plasma than conventional antibodies (Patent Document 1).

Thus, a pH-dependent antigen binding antibody can bind to multiple antigens through a single antibody molecule, accelerating the clearance of antigens from plasma compared to conventional antibodies, and itself has effects that cannot be achieved with conventional antibodies. An amino acid in an existing antibody sequence can be replaced to thereby impart pH-dependent antigen binding activity to it. Meanwhile, to obtain such a new antibody, a method for producing antibodies from immunized animals or a method for producing antibodies from a human antibody library can be used, but it has possible limitations as described below.

A method that involves immunizing non-human animals may provide pH-dependent antibody binding, but rarely can provide pH-dependent antigen binding antibodies against different types of antigens in a short period of time or selectively provide antibodies that specifically bind to particular epitopes. Alternatively, the antibody can be obtained by enriching a library of human antibodies with pH-dependent antigen binding ability as an indicator. However, it is generally known that the frequency of occurrence of histidine residues in human antibody variable regions (registered in the Kabat database) is not high, as can be seen from 5.9% for heavy chain CDR1, 1.4% for heavy chain CDR2, 1 .6% for heavy chain CDR3, 1.5% for heavy chain CDR1, 0.5% for heavy chain CDR2, and 2.2% for light chain CDR3, indicating that the human antibody library contains only a very small number of sequences , which may have the ability to bind antigen in a pH-dependent manner. Thus, there is a need for an antibody library that has an increased frequency of histidine at antigen-binding sites and is enriched in sequences that have the ability to pH-dependent antigen binding.

Effects such as accelerating the clearance of antigen from plasma can be achieved if the antibody is made to depend on a factor (other than pH) for antigen-binding ability that differs between plasma and early endosome conditions.

Below is a list of literature in accordance with the present invention.

Bibliography

Patent literature

Patent Document 1: WO 2009125825

Non-patent document

Non-Patent Document 1: Janice M Reichert, Clark J Rosensweig, Laura B Faden & Matthew C Dewitz, Monoclonal antibody successes in the clinic, Nat. Biotechnol. (2005) 23, 1073-1078.

Non-patent document 2: Pavlou AK, Belsey MJ., The therapeutic antibodies market to 2008, Eur J Pharm Biopharm. (2005) 59(3), 389-396.

Non-patent document 3: Kim SJ, Park Y, Hong HJ., Antibody engineering for the development of therapeutic antibodies, Mol Cells. (2005) 20 (1), 17-29.

Non-patent document 4: Hinton PR, Xiong JM, Johlfs MG, Tang MT, Keller S, Tsurushita N., An engineered human IgG1 antibody with longer serum half-life, J Immunol. (2006) 176(1), 346-356.

Non-patent document 5: Ghetie V, Popov S, Borvak J, Radu C, Matesoi D, Medesan C, Ober RJ, Ward ES., Increasing the serum persistence of an IgG fragment by random mutagenesis, Nat. Biotechnol. (1997) 15 (7), 637-640.

Non-patent document 6: Rajpal A, Beyaz N, Haber L, Cappuccilli G, Yee H, Bhatt RR, Takeuchi T, Lerner RA, Crea R., A general method for greatly improving the affinity of antibodies by using combinatorial libraries, Proc. Natl. Acad. Sci. USA. (2005) 102(24), 8466-8471.

Non-Patent Document 7: Wu H, Pfarr DS, Johnson S, Brewah YA, Woods RM, Patel NK, White WI, Young JF, Kiener PA. Development of Motavizumab, an Ultra-potent Antibody for the Prevention of Respiratory Syncytial Virus Infection in the Upper and Lower Respiratory Tract, J. Mol. Biol. (2007) 368, 652-665.

Non-patent document 8: Hanson CV, Nishiyama Y, Paul S. Catalytic antibodies and their applications. Curr Opin Biotechnol, (2005) 16(6), 631-6.

Non-Patent Document 9: Rathanaswami P, Roalstad S, Roskos L, Su QJ, Lackie S, Babcook J. Demonstration of an in vivo generated sub-picomolar affinity fully human monoclonal antibody to interleukin-8, Biochem. Biophys. Res. Commun. (2005) 334(4), 1004-13.

SUMMARY OF THE INVENTION

TECHNICAL PROBLEM

The present invention has been made in view of such a situation, and the object of the present invention is to provide a library essentially consisting of a plurality of antigen-binding molecules differing from each other in sequence, wherein the antigen-binding domain in each of the antigen-binding molecules contains at least one amino acid residue that modifies the antigen-binding activity of an antigen-binding molecule depending on ion concentration conditions, a composition containing a plurality of polynucleotide molecules, each of which encodes antigen-binding molecules, a composition containing a plurality of vectors, each of which contains polynucleotide molecules, a method for selecting antigen-binding molecules, a method for isolating polynucleotide molecules, a method for producing antigen-binding molecules molecules and a pharmaceutical composition containing any of the antigen-binding molecules.

SOLUTION

The inventors of the present invention have extensively studied a library containing a plurality of antigen-binding molecules differing from each other in sequence, wherein the antigen-binding domain in each of the antigen-binding molecules contains at least one amino acid residue that modifies the antigen-binding activity of the antigen-binding molecule depending on environmental differences in vivo . As a result, the present inventors focused on differences in ion concentration, in particular calcium ion concentration, between plasma and early endosome, or the pH of these environments, and discovered that the use of antigen binding molecules having calcium-dependent or pH-dependent antigen binding activity provides accelerating the cellular uptake of antigens by antigen-binding molecules, and obtaining a library essentially consisting of antigen-binding molecules that reduce the concentration of antigen in plasma.

In particular, the present invention relates to a library consisting essentially of a plurality of antigen binding molecules differing from each other in sequence, wherein the antigen binding domain in each of the antigen binding molecules contains at least one amino acid residue that modifies the antigen binding activity of the antigen binding molecule depending on concentration conditions ions, to a composition containing a plurality of polynucleotide molecules, each of which encodes an antigen-binding molecule, a composition containing a plurality of vectors, each of which contains polynucleotide molecules, to a method for selecting antigen-binding molecules, to a method for isolating polynucleotide molecules, to a method for producing antigen-binding molecules, to pharmaceutical a composition containing any of the antigen-binding molecules, etc. More specifically, the present invention relates to the following:

[1] A library consisting essentially of a plurality of antigen binding molecules differing from each other in sequence, wherein the antigen binding domain in each of the antigen binding molecules contains at least one amino acid residue that modifies the antigen binding activity of the antigen binding molecule depending on ion concentration conditions.

[2] Library according to [1], where the ion concentration is the concentration of calcium ions.

[3] The library according to [2], where the amino acid residue is contained in the antigen binding domain of the heavy chain of the antigen binding molecule.

[4] The library according to [3], wherein the antigen binding domain in the heavy chain is a variable region of the heavy chain.

[5] Library according to [4], where the amino acid residue is contained in the CDR3 of the heavy chain variable region.

[6] The library according to any one of [2]-[5], wherein the amino acid residue is located at any one or more of positions 95, 96, 100a and 101, as defined by Kabat numbering, in the CDR3 of the heavy chain.

[7] The library according to any one of [2]-[6], wherein the amino acid sequence, excluding the specified amino acid residue, contains an amino acid sequence of a naive sequence.

[8] A library according to any one of [3]-[7], wherein the light chain variable region of the antigen binding molecule comprises a naïve sequence amino acid sequence.

[9] A library according to [2], where the amino acid residue is contained in the antigen binding domain of the light chain of the antigen binding molecule.

[10] The library according to [9], wherein the antigen binding domain in the light chain is a light chain variable region.

[11] Library according to [10], where the amino acid residue is located in CDR1 of the light chain variable region.

[12] The library according to [11], wherein the amino acid residue is located at any one or more of positions 30, 31 and 32, defined by Kabat numbering, in CDR1.

[13] A library according to any one of [10]-[12], wherein the amino acid residue is located in CDR2 of the light chain variable region.

[14] The library according to [13], where the amino acid residue is located at position 50, determined by Kabat numbering, in CDR2 of the light chain.

[15] A library according to any one of [10]-[14], wherein the amino acid residue is located in CDR3 of the light chain.

[16] The library according to [15], where the amino acid residue is located at position 92, determined by Kabat numbering, in CDR3 of the light chain.

[17] The library according to any one of [2] and [9]-[16], wherein the light chain framework region of the antigen binding molecule comprises a germline framework sequence.

[18] The library according to any one of [2] and [9]-[17], wherein the heavy chain variable region of the antigen binding molecule comprises a naïve sequence amino acid sequence.

[19] A library according to any one of [1]-[18], wherein the amino acid residue forms a calcium-binding motif.

[20] The library as per [19], wherein the calcium binding motif is any calcium binding motif selected from cadherin domain, EF hand, C2 domain, Gla domain, C-type lectin, A domain, annexin, thrombospondin type 3 domain , EGF-like domain, Vk5 domain, the domain represented by SEQ ID NO: 10, and the domain represented by SEQ ID NO: 11.

[21] A library according to any one of [2] to [20], wherein the amino acid residue is an amino acid having a metal chelating effect.

[22] The library according to [21], wherein the amino acid having the metal chelating effect is any one or more amino acids selected from serine, threonine, asparagine, glutamine, aspartic acid and glutamic acid.

[23] Library according to [1], where the ion concentration conditions are the pH conditions.

[24] Library according to [23], where the amino acid residue is located in the antigen binding domain of the heavy chain of the antigen binding molecule.

[25] The library according to [24], wherein the heavy chain antigen binding domain is a heavy chain variable region.

[26] The library according to [25], where the amino acid residue is located at any one or more of positions 27, 31, 32, 33, 35, 50, 52, 53, 55, 57, 58, 59, 61, 62, 95, 96, 97, 98, 99, 100a, 100b, 100d, 100f, 100h, 102 and 107, as defined by Kabat numbering, in the heavy chain variable region.

[27] The library according to [26], where the amino acid sequence, excluding the amino acid residue at any one or more of positions 27, 31, 32, 33, 35, 50, 52, 53, 55, 57, 58, 59, 61, 62, 95, 96, 97, 98, 99, 100a, 100b, 100d, 100f, 100h, 102 and 107, as defined by Kabat numbering, in the heavy chain variable region contains the amino acid sequence of the naïve sequence.

[28] A library according to any one of [23]-[27], wherein the light chain variable region of the antigen binding molecule contains a germline sequence.

[29] Library according to [23], where the amino acid residue is located in the antigen binding domain of the light chain of the antigen binding molecule.

[30] The library according to [29], wherein the antigen binding domain in the light chain is a light chain variable region.

[31] The library according to [30], where the amino acid residue is located at any one or more of positions 24, 27, 28, 30, 31, 32, 34, 50, 51, 52, 53, 54, 55, 56, 89, 90, 91, 92, 93, 94 and 95a, defined by Kabat numbering, in the light chain variable region.

[32] Library according to [30] or [31], where the amino acid residue is located in CDR1 of the light chain variable region.

[33] The library according to [32], wherein the amino acid residue is located at any one or more of positions 24, 27, 28, 30, 31, 32 and 34, as defined by Kabat numbering, in the light chain CDR1.

[34] A library according to any one of [30]-[33], wherein the amino acid residue is located in CDR2 of the light chain.

[35] The library according to [34], wherein the amino acid residue is located at any one or more of positions 50, 51, 52, 53, 54, 55 and 56, as defined by Kabat numbering, in the CDR2 of the light chain.

[36] A library according to any one of [30]-[35], wherein the amino acid residue is located in CDR3 of the light chain.

[37] The library according to [36], wherein the amino acid residue is located at any one or more of positions 89, 90, 91, 92, 93, 94 and 95a, as defined by Kabat numbering, in the CDR3 of the light chain.

[38] A library according to any one of [29]-[37], wherein the light chain framework region comprises a germline framework sequence.

[39] A library according to any one of [29]-[38], wherein the heavy chain variable region has a naive sequence.

[40] A library according to any one of [23]-[39], wherein the amino acid residue is an amino acid having a side chain pKa of 4.0 to 8.0.

[41] A library according to any one of [23]-[40], wherein the amino acid residue is glutamic acid.

[42] A library according to any one of [23]-[39], wherein the amino acid residue is an amino acid having a side chain pKa of 5.5 to 7.0.

[43] A library according to any one of [23]-[40] and [42], wherein the amino acid residue is histidine.

[44] A library consisting essentially of a plurality of fusion polypeptides, each of which contains antigen-binding molecules according to any one of [1]-[43], wherein each of the fusion polypeptides is the product of a fusion of the heavy chain variable region of an antigen-binding molecule and at least part viral envelope protein.

[45] The library according to [44], wherein the viral envelope protein is selected from the group consisting of pIII protein, pVIII major envelope protein, pVII, pIX, Soc, Hoc, gpD, pv1 and variants thereof.

[46] A composition comprising a plurality of polynucleotide molecules, each of which encodes antigen binding molecules differing from each other in sequence according to any of [1]-[43] or fusion polypeptides differing from each other in sequence according to [44] or [45] ].

[47] A composition containing a plurality of vectors, each of which contains a plurality of polynucleotide molecules according to [46] in a functionally linked state.

[48] The composition according to [47], wherein the vectors are expression vectors that can be replicated.

[49] The composition according to [48], wherein each of the expression vectors that can be replicated is an expression vector in which the polynucleotide is operably linked to a promoter region selected from the group consisting of the lacZ promoter system, the phoA alkaline phosphatase (AP) promoter , bacteriophage λPL promoter (temperature sensitive promoter), tac promoter, tryptophan promoter, pBAD promoter and bacteriophage T7 promoter.

[50] The composition according to [48] or [49], wherein each of the expression vectors that can be replicated is an M13, f1, fd or Pf3 phage or a derivative thereof, or a lambdoid phage or a derivative thereof.

[51] A composition containing a plurality of viruses, each of which contains vectors according to any one of [47]-[50].

[52] A composition containing a plurality of viruses, each of which exhibits on its surface antigen-binding molecules differing from each other in sequence according to any of [1]-[43] or fusion polypeptides differing from each other in sequence according to [44] or [45].

[53] A library containing antigen-binding molecules differing from each other in sequence according to any of [1]-[43] or fusion polypeptides differing from each other in sequence according to [44] or [45], where the library has from 1× 10 ⁶ to 1×10 ¹⁴ different variable region sequences.

[54] A library according to [53], where the library has 1×10 ⁸ or more different variable region sequences.

[55] A method of producing a library substantially consisting of a plurality of antigen-binding molecules differing from each other in sequence, the method comprising producing a plurality of antigen-binding molecules designed such that the antigen-binding domain in each of the antigen-binding molecules contains at least one amino acid residue that changes antigen-binding activity of an antigen-binding molecule depending on ion concentration conditions.

[56] The production method according to [55], wherein the antigen binding molecules are antigen binding molecules according to any one of [2] to [43].

[57] The production method according to [55] or [56], wherein the heavy chain variable region of each of the antigen binding molecules is fused to at least a portion of the viral envelope protein.

[58] The production method according to any one of [55]-[57], wherein the viral envelope protein is selected from the group consisting of pIII protein, pVIII core protein, pVII, pIX, Soc, Hoc, gpD, pv1 and variants thereof.

[59] A method for selecting an antigen-binding molecule whose antigen-binding activity varies depending on ion concentration conditions, the method comprising the steps of:

a) obtaining a library consisting essentially of antigen binding molecules differing from each other in sequence according to any of [1]-[43] or fusion polypeptides differing from each other in sequence according to [44] or [45];

b) contacting the library with antigens under two or more different ion concentration conditions;

c) sorting from the library a subpopulation of antigen-binding molecules whose antigen-binding activity varies depending on ion concentration conditions; And

d) isolating each antigen-binding molecule whose antigen-binding activity varies with ion concentration conditions from the subpopulation sorted in step c).

[60] A method for isolating a polynucleotide encoding an antigen-binding molecule, the antigen-binding activity of which varies depending on ion concentration conditions, the method comprising the steps of:

a) obtaining a library containing a plurality of expression vectors that can be replicated, each of which contains in an operably linked state a plurality of polynucleotides, each of which encodes antigen binding molecules differing from each other in sequence, according to any of [1]-[43] or fused polypeptides differing from each other in sequence according to [44] or [45];

b) allowing a plurality of viruses, each transformed by expression vectors contained in the library, to express on their surface antigen-binding molecules or fusion polypeptides that differ from each other in sequence, encoded by the polynucleotides;

c) bringing multiple viruses into contact with antigens under two or more different ion concentration conditions;

d) sorting from the library a subpopulation of viruses whose antigen-binding activity varies depending on ion concentration conditions;

e) isolating each virus whose antigen-binding activity varies with ion concentration conditions from the subpopulation of viruses sorted in step d); And

f) isolating polynucleotides from the isolated virus.

[61] The method according to [60], where steps c) and d) are further repeated at least once.

[62] The method according to any one of [59]-[61], wherein the ion concentration is the concentration of calcium ions.

[63] The method according to [62], wherein an antigen binding molecule is selected which has lower antigen binding activity under conditions of low calcium concentration than under conditions of high calcium concentration.

[64] The method according to [63], wherein the low calcium concentration conditions are 0.1 μM to 30 μM conditions.

[65] The method according to [63] or [64], wherein the high calcium concentration conditions are those of 100 μM to 10 mM.

[66] The method according to any one of [59]-[61], wherein the ion concentration conditions are pH conditions.

[67] The method according to [66], wherein an antigen binding molecule is selected which has lower antigen binding activity under acidic pH conditions than under neutral pH conditions.

[68] The method according to [66], wherein acidic pH conditions are pH conditions between 4.0 and 6.5.

[69] The method according to [67] or [68], where the neutral pH conditions are pH conditions between 6.7 and 10.0.

[70] A method for producing an antigen-binding molecule, the antigen-binding activity of which varies depending on ion concentration conditions, the method comprising the steps of:

a) obtaining a library containing a plurality of expression vectors that can be replicated, each of which contains in an operably linked state a plurality of polynucleotides, each of which encodes antigen binding molecules differing from each other in sequence, according to any of [1]-[43] or fused polypeptides differing from each other in sequence according to [44] or [45];

b) allowing a plurality of viruses, each transformed by expression vectors contained in the library, to express on their surface antigen-binding molecules or fusion polypeptides that differ from each other in sequence, encoded by the polynucleotides;

c) bringing multiple viruses into contact with antigens under two or more different ion concentration conditions;

d) sorting from the library a subpopulation of viruses whose antigen-binding activity varies depending on ion concentration conditions;

e) isolating each virus whose antigen-binding activity varies with ion concentration conditions from the subpopulation of viruses sorted in step d);

f) isolating polynucleotides from the isolated virus;

g) culturing a host cell transfected with a vector having an operably linked insert of the isolated polynucleotides; And

h) collecting antigen-binding molecules from the cell cultures cultured in step g).

[71] A method for producing an antigen-binding molecule, the antigen-binding activity of which varies depending on ion concentration conditions, the method comprising the steps of:

a) obtaining a library containing a plurality of expression vectors that can be replicated, each of which contains in an operably linked state a plurality of polynucleotides, each of which encodes antigen binding molecules differing from each other in sequence, according to any of [1]-[43] or fused polypeptides differing from each other in sequence according to [44] or [45];

b) allowing a plurality of viruses, each transformed by expression vectors contained in the library, to express on their surface antigen-binding molecules or fusion polypeptides that differ from each other in sequence, encoded by the polynucleotides;

c) bringing multiple viruses into contact with antigens under two or more different ion concentration conditions;

d) sorting from the library a subpopulation of viruses whose antigen-binding activity varies depending on ion concentration conditions;

e) isolating each virus whose antigen-binding activity varies with ion concentration conditions from the subpopulation of viruses sorted in step d);

f) isolating polynucleotides from the isolated virus;

g) linking the isolated polynucleotides in frame to a polynucleotide encoding the antibody constant region;

h) culturing a host cell transfected with a vector having an operably linked insert of polynucleotides linked in step g); And

i) isolating antigen-binding molecules from cell cultures cultured in step h).

[72] The production method according to [70] or [71], where steps c) and d) are further repeated at least once.

[73] The production method according to any one of [70] to [72], wherein the ion concentration is the concentration of calcium ions.

[74] The production method according to [73], wherein an antigen binding molecule is selected which has lower antigen binding activity under low calcium concentration conditions than under high calcium concentration conditions.

[75] The production method according to [74], wherein the low calcium concentration conditions are those of 0.1 μM to 30 μM.

[76] The production method according to [74] or [75], wherein the high calcium concentration conditions are those of 100 μM to 10 mM.

[77] The production method according to any one of [70] to [72], wherein the ion concentration conditions are pH conditions.

[78] The production method according to [77], wherein an antigen binding molecule is selected which has lower antigen binding activity under acidic pH conditions than under neutral pH conditions.

[79] The production method according to [78], wherein the acidic pH conditions are pH conditions of 4.0 to 6.5.

[80] The production method according to [78] or [79], wherein the conditions of neutral pH values are those of pH 6.7 to 10.0.

[81] An antigen-binding molecule obtained by the production method according to any one of [70]-[80].

[82] A pharmaceutical composition containing an antigen-binding molecule according to [81] or a modified form thereof.

Brief description of drawings

[Fig. 1] In FIG. 1 is a graph showing the relationship between the amino acid distribution (indicated as library) for the sequence information of approximately 132 clones isolated from E. coli transformed with a library of antibody genes that bind antigens in a pH-dependent manner and the model amino acid distribution (indicated as model). The x-axis represents the position of the amino acid, determined by Kabat numbering. The y-axis represents the proportion of each amino acid in the distribution.

[Fig. 2] In FIG. 2 shows sensorgrams for the anti-IL-6R antibody (tocilizumab), 6RpH#01 antibody, 6RpH#02 antibody and 6RpH#03 antibody at pH 7.4. The x-axis represents time. The ordinate axis represents the RU values.

[Fig. 3] In FIG. 3 shows sensorgrams for the anti-IL-6R antibody (tocilizumab), 6RpH#01 antibody, 6RpH#02 antibody and 6RpH#03 antibody at pH 6.0. The x-axis represents time. The ordinate axis represents the RU values.

[Fig. 4A] In FIG. 4A is a diagram showing the interaction profile between a pH-dependent binding antibody and its antigen in plasma (pH 7.4) and endosome (pH 6.0).

[Fig. 4B] In FIG. 4B is a diagram showing the interaction profile between a calcium-dependent binding antibody and its antigen in plasma (2 mM Ca ²⁺ ) and in the endosome (3 μM Ca ²⁺ ).

[Fig. 4C] In FIG. 4C is a diagram showing the interaction profile between a pH- and calcium-dependent binding antibody and its antigen in plasma (2 mM Ca ²⁺ ) and in the endosome (3 μM Ca ²⁺ ).

[Fig. 5] In FIG. 5 shows ion exchange chromatograms for an antibody containing the human Vk5-2 sequence and an antibody containing the hVk5-2_L65 sequence modified from the human Vk5-2 sequence at the glycosylation sequence region. The solid line corresponds to the chromatogram for an antibody containing the human Vk5-2 sequence (heavy chain: CIM_H (SEQ ID NO: 4) and light chain: hVk5-2 (SEQ ID NO: 1, fused to SEQ ID NO: 26)). The dotted line corresponds to the chromatogram for an antibody having the sequence hVk5-2_L65 (heavy chain: CIM_H (SEQ ID NO: 4) and light chain: hVk5-2_L65 (SEQ ID NO: 5)).

[Fig. 6] In FIG. 6 shows ion exchange chromatograms for an antibody containing the sequence LfVk1_Ca (heavy chain: GC_H (SEQ ID NO: 48) and light chain: LfVk1_Ca (SEQ ID NO: 43)), and an antibody containing a sequence modified from the LfVk1_Ca sequence by replacing the Asp residue (D) on the Ala residue (A), after storage at 5°C (solid line) or after storage at 50°C (dashed line). The highest peak in each ion exchange chromatogram after storage at 5°C is defined as the main peak. In the chart, the y-axis is normalized to the main peak.

[Fig. 7] In FIG. 7 shows ion exchange chromatograms for an antibody containing the sequence LfVk1_Ca (heavy chain: GC_H (SEQ ID NO: 48) and light chain: LfVk1_Ca (SEQ ID NO: 43)), and an antibody containing the sequence LfVk1_Ca6 (heavy chain: GC_H (SEQ ID NO: 48) and light chain: LfVk1_Ca6 (SEQ ID NO: 49)), modified from the sequence LfVk1_Ca by replacing the Asp (D) residue at position 30 (defined by Kabat numbering) with a Ser (S) residue after storage at 5° C (solid line) or after storage at 50°C (dashed line). The highest peak in each ion exchange chromatogram after storage at 5°C is defined as the main peak. In the chart, the y-axis is normalized to the main peak.

[Fig. 8] In FIG. 8 is a graph showing the relationship between the amino acid distribution (listed as library) for sequence information of approximately 290 clones isolated from E. coli transformed with a library of antibody genes that bind antigens in a Ca-dependent manner and the model amino acid distribution (listed as model). The x-axis represents the position of the amino acid, determined by Kabat numbering. The y-axis represents the proportion of each amino acid in the distribution.

[Fig. 9] In FIG. Figure 9 shows sensorgrams for the antibody against IL-6R (tocilizumab), antibody 6RC1IgG_010, antibody 6RC1IgG_012 and antibody 6RC1IgG_019 under conditions of high concentration of calcium ions (1.2 mM). The x-axis represents time. The ordinate axis represents the RU values.

[Fig. 10] In FIG. 10 shows sensorgrams for the antibody against IL-6R (tocilizumab), antibody 6RC1IgG_010, antibody 6RC1IgG_012 and antibody 6RC1IgG_019 under conditions of low calcium ion concentration (3 μM). The x-axis represents time. The ordinate axis represents the RU values.

[Fig. 11] In FIG. 11 shows the heavy chain CDR3 structure of the Fab fragment of antibody 6RL#9 determined by X-ray diffraction analysis. In fig. 11(i) is a diagram showing a heavy chain CDR3 with a crystal structure obtained under crystallization conditions in the presence of calcium ions. In fig. 11(ii) is a diagram showing a heavy chain CDR3 with a crystal structure obtained under crystallization conditions in the absence of calcium ions.

[Fig. 12] In FIG. 12 is a sensogram showing the interaction between anti-human IgA antibody and human IgA at 1.2 mM Ca ²⁺ and at 3 μM Ca ²⁺ using Biacore.

[Fig. 13] In FIG. 13 is a graph showing the interaction between anti-human glypican 3 antibody and recombinant human glypican 3 at 1.2 mM Ca ²⁺ and at 3 μM Ca ²⁺ using ELISA.

[Fig. 14] In FIG. 14 is a sensorgram showing the interaction between anti-mouse IgA antibody and mouse IgA at pH 7.4 and at pH 5.8 using Biacore. The solid line corresponds to results at approximately pH 7.4. The dotted line corresponds to results at approximately pH 5.8.

[Fig. 15] In FIG. 15 is a sensorgram showing the interaction between anti-human HMGB1 antibody and human HMGB1 at pH 7.4 and at pH 5.8 using Biacore. The solid line corresponds to results at approximately pH 7.4. The dotted line corresponds to results at approximately pH 5.8.

[Fig. 16] In FIG. 16 is a graph showing the plasma concentrations of H54/L28-IgG1 antibody, FH4-IgG1 antibody, and 6RL#9-IgG1 antibody in a normal mouse.

[Fig. 17] In FIG. 17 is a graph showing the concentration of soluble human IL-6 receptor (hsIL-6R) in the plasma of a normal mouse administered with H54/L28-IgG1 antibody, FH4-IgG1 antibody, or 6RL#9-IgG1 antibody.

[Fig. 18] In FIG. 18 is a graph showing the plasma concentrations of antibody H54/L28-N434W, antibody FH4-N434W and antibody 6RL#9-N434W in a normal mouse.

[Fig. 19] In FIG. 19 is a graph showing the concentration of soluble human IL-6 receptor (hsIL-6R) in the plasma of a normal mouse administered with the H54/L28-N434W antibody, the FH4-N434W antibody, and the 6RL#9-N434W antibody.

DESCRIPTION OF IMPLEMENTATION OPTIONS

The present invention relates to a library essentially consisting of a plurality of antigen binding molecules differing from each other in sequence, wherein the antigen binding activity of each antigen binding molecule varies depending on ion concentration conditions and the like. The present invention also relates to a new systemic method for producing a library essentially consisting of a plurality of antigen binding molecules differing from each other in sequence, wherein the antigen binding activity of each antigen binding molecule varies depending on conditions of metal ion concentration and/or hydrogen ion concentration. Such a library can be used as a combinatorial library that assists in the selection and/or screening of a synthetic clone of an antigen binding molecule with a desired activity, such as binding affinity and avidity, suitable for, for example, metal ion concentration and/or hydrogen ion concentration conditions.

These libraries are useful for identifying the polypeptide sequence of an antigen-binding molecule that can interact with any of the various types of target antigens. For example, a library containing polypeptides of a variety of antigen binding molecules of the present invention expressed by phage display is particularly useful for selecting and/or screening for an antigen binding molecule of interest. The present invention also relates to an efficient, high-performance automatic system for this purpose. The method of the present invention can provide an antigen-binding molecule that binds to a target antigen in a condition-dependent manner. Moreover, the present invention relates to a pharmaceutical composition containing an antigen binding molecule as an active ingredient.

DEFINITION

Amino acid

Each amino acid is indicated herein by a one-letter code or a three-letter code, or both, for example, as indicated by Ala/A, Leu/L, Arg/R, Lys/K, Asn/N, Met/M, Asp/D , Phe/F, Cys/C, Pro/P, Gln/Q, Ser/S, Glu/E, Thr/T, Gly/G, Trp/W, His/H, Tyr/Y, Ile/I and Val /V.

EU numbering and Kabat numbering

In accordance with the method used in the present invention, the amino acid positions assigned to the CDR and FR of the antibody are determined according to the Kabat method (Sequences of Proteins of Immunological Interest, National Institute of Health, Bethesda, Md., 1987 and 1991). When the antigen binding molecule described herein is an antibody or an antigen binding fragment, amino acids in the variable regions are indicated in accordance with Kabat numbering, and amino acids in the constant regions are indicated in accordance with EU numbering corresponding to Kabat amino acid positions.

Amino acid modification

Amino acids in the amino acid sequences of antigen-binding molecules can be modified using a suitably selected method known in the art, such as site-directed mutagenesis (Kunkel et al., Proc. Natl. Acad. Sci. USA (1985) 82, 488-492 ) or overlap PCR. Also, amino acids can be replaced with unnatural amino acids using a variety of modification methods known in the art (Annu. Rev. Biophys. Biomol. Struct. (2006) 35, 225-249; and Proc. Natl. Acad. Sci. U.S.A. (2003) 100 (11), 6353-6357). For example, it is also preferable to use a tRNA cell-free translation system (Clover Direct (Protein Express, an R&D oriented company)) containing a non-natural amino acid associated with an amber suppressor tRNA complementary to the UAG codon, which is a stop codon. -codon. In addition, an expression can be used to indicate an amino acid modification that uses the single letter codes of the amino acids before and after the modification before and after the number corresponding to a particular position. For example, the P238D modification used to add an amino acid substitution to the Fc region contained in the constant region of an antibody corresponds to the substitution of Pro at position 238, defined by EU numbering, with Asp. In particular, the number corresponds to the position of the amino acid as determined by EU numbering; the one-letter amino acid code preceding the number corresponds to the amino acid before the substitution; and the one-letter amino acid code following the number corresponds to the amino acid after the substitution.

And/or

The term “and/or” described herein includes each combination represented by “and” and “or”, respectively. In particular, for example, the expression "amino acids 33, 55, and/or 9 6 replaced" includes the following amino acid modification options: (a) position 33, (b) position 55, (c) position 96, (d) positions 33 and 55 , (e) regulations 33 and 96, (f) regulations 55 and 96, and (g) regulations 33, 55 and 96.

Antigen binding molecule

The term "antigen binding molecule" as used herein is used to refer to a molecule containing an antigen binding domain in the broadest sense and specifically includes various molecular forms so long as these forms exhibit antigen binding activity. Examples of a molecule comprising an antigen binding domain coupled to an FcRn binding domain include antibodies. Antibodies may include simple monoclonal antibodies (including agonist and antagonist antibodies), human antibodies, humanized antibodies, chimeric antibodies, and the like. Alternatively, a fragment of such an antibody may be used. Preferred examples of the fragment may include antigen binding domains and antigen binding fragments (eg Fab, F(ab')2, scFv and Fv). The antigen binding molecule of the present invention may also include scaffold molecules contained in an antigen binding domain design library containing only partial structures of existing stable conformations (eg, an α/β barrel protein structure) used as scaffolds.

The "antigen binding domain" described herein can be a domain having any structure as long as the domain used binds to the antigen of interest. Preferred examples of such a domain include the variable regions of the heavy and light chains of antibodies, an in-vivo membrane protein-derived module called domain A of approximately 35 amino acids in size contained in avimer (WO 2004044011 and WO 2005040229), adnectin containing a 10Fn3 domain as a binding a domain protein derived from the fibronectin glycoprotein expressed on cell membranes (WO 2002032925), an affinity body containing an IgG-binding domain scaffold consisting of a trihelix bundle of 58 amino acids of protein A (WO 1995001937), DARPin (modeled ankyrin repeat proteins) , which are ankyrin repeat (AR) regions exposed on the surface of molecules, each of which has a structure of 33 amino acid residues arranged into a subunit of a turn, two antiparallel helices and a loop (WO 2002020565), an ankylin having four loop regions connecting eight antiparallel chains bent towards a central axis at one end of a barrel structure highly conserved among lipocalin molecules such as neutrophil gelatin-associated lipocalin (NGAL) (WO 2003029462), and a concave region in the internal structure of parallel layers in a horseshoe-shaped stack formed by repeating leucine-rich repeat (LRR) modules of the variable lymphocyte receptor (VLR), which does not have an immunoglobulin structure, as observed for the acquired immune system of jawless vertebrates such as lamprey and hagfish (WO 2008016854). Preferred examples of the antigen binding domain of the present invention include antigen binding domains containing variable regions of the heavy and light chains of antibodies.

The term "antibody" as used herein refers to a naturally occurring immunoglobulin or an immunoglobulin produced by partial or total synthesis. The antibody can be isolated from natural sources (eg, plasma and serum) in which the antibody occurs naturally, or from culture supernatants of antibody-producing hybridomas. Alternatively, antibodies can be partially or fully synthesized using an approach such as genetic recombination. Preferred examples of antibodies include immunoglobulin isotypes and subclasses of these isotypes. It is known that human immunoglobulins have 9 classes (isotypes): IgG1, IgG2, IgG3, IgG4, IgA1, IgA2, IgD, IgE and IgM. Of these isotypes, the antibody of the present invention may include IgG1, IgG2, IgG3 and IgG4. In Sequences of proteins of immunological interest, NIH Publication No. 91-3242, a variety of allotypic sequences are described, due to polymorphism, as constant regions of human IgG1, human IgG2, human IgG3 and human IgG4, any of which can be used within the scope of the present invention. In particular, human IgG1 may have a sequence with DEL or EEM as the amino acid sequence at positions 356-358, defined by EU numbering. In Sequences of proteins of immunological interest, NIH Publication No. 91-3242, a variety of allotypic sequences due to polymorphism are described as human IgK (kappa) and human IgL7 (lambda) constant regions, any of which can be used within the scope of the present invention. An antibody having the desired binding activity is prepared in a manner well known to those skilled in the art.

The antibody can be produced as a polyclonal or monoclonal antibody using methods known in the art. As the monoclonal antibody, it is preferable to obtain a monoclonal antibody derived from a mammal. Mammalian-derived monoclonal antibody includes, for example, antibodies produced by hybridomas and antibodies produced by host cells transformed with expression vectors containing antibody genes using a genetic engineering approach.

Hybridomas producing monoclonal antibodies can be obtained using a method known in the art. In particular, mammals are immunized with sensitizing antigens in accordance with a conventional immunization method. The resulting immunocytes are fused with parent cells known in the art using a conventional cell fusion method. Further, these fusion cells can be screened for monoclonal antibody producing cells using a conventional screening method to select for hybridomas producing antibodies against sensitizing antigens.

Mammals to be immunized with sensitizing antigens are not limited to any particular animal and are preferably selected for use in cell fusion based on compatibility with the parent cells. In general, it is preferable to use rodents, for example mice, rats, hamsters or rabbits, or other mammals such as monkeys.

These animals are immunized with sensitizing antigens according to a method known in the art. For example, a general method may involve immunizing mammals with sensitizing antigens by intraperitoneal or subcutaneous injection. In particular, sensitizing antigens diluted with PBS (phosphate-buffered saline), saline, etc., with an appropriate degree of dilution, are mixed, if desired, with a conventional adjuvant, for example, Freund's complete adjuvant, and emulsified, and then the resulting an emulsion of sensitizing antigens is administered to mammals several times at intervals of 4 to 21 days. Also, a suitable carrier can be used for immunization with sensitizing antigens. In particular, when low molecular weight partial peptides are used as sensitizing antigens, it is desirable to use sensitizing antigen peptides coupled to carrier proteins such as albumin or keyhole limpet hemocyanin for immunization.

Hybridomas that produce antibodies against the desired polypeptide can also be obtained by DNA immunization, as described below. DNA immunization refers to a method of stimulating immunity that involves: immunizing animals by administering vector DNAs that are engineered in a form capable of expressing genes encoding an antigenic protein in immunized animals; and stimulation of animal immunity through in vivo expression of sensitizing antigens. DNA immunization can be expected to be superior in the following ways relative to common immunization methods that involve immunizing animals by administering protein antigens:

- when the antigen is membrane proteins, DNA immunization can provide stimulation of the immune system while maintaining the structures of membrane proteins; And

- DNA immunization eliminates the need for purification of immunizing antigens.

Mammalian myeloma cells are used in cell fusion with immunocytes. Myeloma cells preferably have a suitable selectable marker for screening. A selectable marker refers to a trait that may (or may not) persist under specific culture conditions. For example, hypoxanthine guanine phosphoribosyltransferase deficiency (hereinafter abbreviated as HGPRT deficiency) or thymidine kinase deficiency (hereinafter abbreviated as TK deficiency) is known in the art as a selectable marker. Cells deficient in HGPRT or TK are hypoxanthine-aminopterin-thymidine sensitive (hereinafter abbreviated as HAT sensitive). HAT-sensitive cells die in the HAT-selective environment because the cells do not synthesize DNA. In contrast, these cells, when fused with normal cells, become capable of growing even in a selective HAT environment, since the fused cells can continue DNA synthesis using the normal cell rescue pathway.

Selection of cells deficient in HGPRT or TK can be carried out in media containing 6-thioguanine or 8-azaguanine (hereinafter abbreviated as 8AG) or 5'-bromodeoxyuridine, respectively. Normal cells die when these pyrimidine analogues are incorporated into their DNA. In contrast, cells deficient in these enzymes can survive in a selective environment because these cells cannot incorporate pyrimidine analogues into their DNA. In addition, a selectable marker called the G418 resistance marker confers resistance to the antibiotic 2-deoxystreptamine (a gentamicin analogue) through the neomycin resistance gene. Various myeloma cells are known in the art that are suitable for cell fusion.

For example, P3 (P3x63Ag8,653) (J. Immunol. (1979) 123 (4), 1548-1550), P3x63Ag8U.1 (Current Topics in Microbiology and Immunology (1978) 81, 1 -7), NS-1 (C. Eur. J. Immunol. (1976) 6 (7), 511-519), MPC-11 (Cell (1976) 8 (3), 405-415), SP2/0 (Nature (1978) 276 (5685), 269-270), FO (J. Immunol. Methods (1980) 35 (1-2), 1-21), S194/5.XX0.BU.1 (J. Exp Med. (1978) 148 (1), 313-323) or R210 (Nature (1979) 277 (5692), 131-133).

In general, cell fusion between immunocytes and myeloma cells is carried out according to a method known in the art, for example, the method of Kohler and Milstein et al. (Methods Enzymol. (1981) 73, 3-46).

More specifically, cell fusion can be carried out, for example, in a conventional culture medium in the presence of a cell fusion accelerator. For example, polyethylene glycol (PEG) or Japanese hemagglutinating virus (HVJ) can be used as a fusion accelerator. In addition, if desired, an auxiliary substance such as dimethyl sulfoxide is added and used to improve the fusion efficiency.

The ratio between the immunocytes and myeloma cells used can be chosen arbitrarily. For example, the number of immunocytes is preferably selected to be 1 to 10 times the number of myeloma cells. As the culture medium for cell fusion, for example, RPMI1640 or MEM media suitable for growing a myeloma cell line, or any other conventional culture medium for use in a cell culture of the above type, can be used, and can be further supplemented with a serum supplemented solution such as fetal calf serum (FCS).

For cell fusion, immunocytes and myeloma cells are mixed well in predetermined quantities in the culture medium, and to them is usually added a solution of PEG (for example, having an average molecular weight of the order of 1000 to 6000), preheated to approximately 37°C, in a concentration of 30 to 60% (w/v). The mixed solution is stirred gently to form the desired fused cells of interest (hybridoma). The appropriate culture medium illustrated above is then sequentially added to the cell suspensions and the supernatant is removed by centrifugation. This procedure can be repeated to thereby remove cell fusion agents, etc., which are detrimental to the growth of hybridomas.

The hybridomas thus obtained can be cultured for selection using a conventional selection medium, for example, HAT medium (culture medium containing hypoxanthine, aminopterin and thymidine). Culture using HAT medium can be continued long enough (usually several days to several weeks) to ensure the death of cells (unfused cells) other than the desired hybridomas. Hybridomas producing the desired antibody are then screened for single clones and cloned as single clones by the usual limiting dilution method.

The hybridomas thus obtained can be selected using a selection medium suitable for the selectable marker contained in the myeloma cells used for cell fusion. For example, selection of cells deficient in HGPRT or TK can be done by culturing in HAT medium (a culture medium containing hypoxanthine, aminopterin and thymidine). Particularly in the case of HAT-sensitive myeloma cells used for cell fusion, only cells successfully fused with normal cells are able to selectively grow in HAT media. Culture using HAT medium is continued for a time long enough to ensure the death of cells (unfused cells) other than the desired hybridomas. In particular, to select the desired hybridomas, cultivation can generally be carried out over a period of several days to several weeks. Hybridomas producing the desired antibody are then screened for single clones and cloned as single clones by the usual limiting dilution method.

Screening of the desired antibody and cloning as single clones can preferably be carried out using an antigen-antibody reaction-based screening method known in the art. Screening of such a monoclonal antibody can be carried out, for example, using FACS (fluorescence-activated cell sorting). FACS refers to a system that can analyze cells exposed to fluorescent antibodies using laser light and measure the fluorescence emitted by individual cells, thereby analyzing the binding of antibodies to the cell surface.

Alternatively, the antibody can be assessed for its binding activity to immobilized antigens based on ELISA principles. For example, antigens are immobilized on the wells of an ELISA plate. The hybridoma culture supernatant is brought into contact with antigens in the wells to detect antibody bound to the antigen. In the case of a mouse-derived monoclonal antibody, the antigen-bound antibody can be detected using an anti-mouse immunoglobulin antibody. These screened hybridomas that produce the desired antibody and have antigen-binding ability can be cloned by limiting dilution techniques or the like. Monoclonal antibody-producing hybridomas thus obtained can be subcloned in conventional culture medium. Hybridomas can also be stored for long periods of time in liquid nitrogen.

Hybridomas can be cultured in the usual way. The desired monoclonal antibody can be obtained from their culture supernatant. Alternatively, hybridomas can be administered to compatible mammals and grown, and a monoclonal antibody can be produced from their ascites fluid. The first of these methods is suitable for obtaining highly pure antibodies.

It is also preferable to use antibodies encoded by antibody genes cloned from antibody-producing cells such as hybridomas. The cloned antibody genes are inserted into suitable vectors and transferred to hosts for expression of the antibodies encoded by the genes. Methods for isolating antibody genes, inserting them into vectors and transforming host cells have already been developed, for example by Vandamme et al. (Eur. J. Biochem. (1990) 192 (3), 767-775). A method for producing recombinant antibodies is also known in the art, as follows.

For example, cDNAs encoding the variable regions (V regions) of an antibody of interest are obtained from hybridoma cells that produce the antibody. For this purpose, total RNA is usually first extracted from hybridomas. For example, the following methods can be used to extract mRNA from cells:

- ultracentrifugation method with guanidine (Biochemistry (1979) 18 (24), 5294-5299), and

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

The extracted mRNA can be purified using an mRNA purification kit (GE Healthcare Bio-Sciences Corp.) or the like. Alternatively, a kit for direct extraction of total mRNA from cells, such as the QuickPrep mRNA purification kit (GE Healthcare Bio-Sciences Corp.), is also commercially available. Using such a kit, it is possible to obtain mRNA from hybridomas. From the resulting mRNAs, cDNAs encoding the antibody V region can be synthesized using reverse transcriptase. cDNA can be synthesized using an AMV reverse transcriptase first-strand cDNA synthesis kit (Seikagaku Corp.) or the like. Alternatively, the SMART RACE cDNA Amplification Kit (Clontech Laboratories, Inc.) and 5'-RACE PCR (Proc. Natl. Acad. Sci. USA (1988) 85 (23), 8998-9002; and Nucleic Acids Res (1989) 17 (8), 2919-2932). During such cDNA synthesis, the appropriate restriction sites described below can be added to both ends of the cDNA.

cDNA fragments of interest are purified from the resulting PCR products and then linked to vector DNAs. The recombinant vectors thus obtained are transferred into E. coli, etc., and then colony selection is carried out. A recombinant vector can then be produced from E. coli that have formed colonies. The presence or absence of the cDNA nucleotide sequence of interest in the recombinant vector is then confirmed by a method known in the art, such as dideoxynucleotide chain termination.

Genes encoding the variable region are conveniently obtained by the 5'-RACE method using primers to amplify the variable region genes. First, cDNA is synthesized using RNA extracted from hybridoma cells as a template to obtain a cDNA library for 5'-RACE. To synthesize a cDNA library for 5'-RACE, an available kit such as the SMART RACE cDNA Amplification Kit is appropriately used.

Antibody genes are amplified by PCR using the resulting 5'-RACE cDNA library as a template. Primers for amplifying mouse antibody genes can be designed based on antibody gene sequences known in the art. These primers have a nucleotide sequence that differs for each immunoglobulin subclass. Thus, the subclass of the antibody of interest is desirably determined in advance using a commercially available kit, such as the Iso Strip Mouse Monoclonal Antibody Isotyping Kit (Roche Diagnostics K.K.).

In particular, primers capable of amplifying the genes encoding the γ1, γ2a, γ2b and γ3 heavy chains and the κ and λ light chains can be used, for example, for the purpose of obtaining genes encoding mouse IgG. Primers that anneal to portions corresponding to constant regions near variable regions are typically used as 3' primers for amplification of IgG variable region genes. On the other hand, the primers included in the 5' RACE cDNA library preparation kit are used as 5' primers.

PCR products amplified in this manner can be used to reconstruct immunoglobulins consisting of heavy and light chains in combination. The reconstituted immunoglobulins can be screened for the desired antibody with their antigen binding activity as an indicator. For example, for the purpose of producing an antibody against an antigen, more preferably, the antibody specifically binds to the antigen. An antibody that binds to an antigen can be screened, for example, by the following steps:

(1) bringing into contact antibodies containing V regions encoded by cDNAs derived from hybridomas with antigens;

(2) detection of antigen-antibody binding; And

(3) selection of antigen-binding antibody.

Antigen-antibody binding detection is carried out in a manner known in the art. In particular, antigen-antibody binding can be detected using an approach such as FACS or ELISA described above.

After each cDNA encoding an antibody V region of interest is obtained, the cDNA is digested with restriction enzymes that recognize restriction sites embedded at both ends of the cDNA. Preferably, restriction enzymes recognize and cleave a nucleotide sequence that occurs at low frequency in the nucleotide sequences constituting antibody genes. More preferably, restriction enzymes cleave the inserted regions to form sticky ends to insert one copy of the cleaved fragment in the correct direction into the vector. cDNAs encoding the antibody V region cleaved in this manner can be inserted into appropriate expression vectors to produce antibody expression vectors. In this case, genes encoding the antibody constant region (C region) are fused in frame with genes encoding the V region to produce chimeric antibodies. In this context, chimeric antibodies refer to antibodies containing constant and variable regions of different origins. Thus, the chimeric antibody of the present invention also covers heterogeneous (eg, mouse-human) chimeric antibodies as well as homogeneous human-human chimeric antibodies. V region genes can be inserted into expression vectors pre-containing constant region genes, constructing expression vectors for chimeric antibodies. In particular, for example, recognition sequences for restriction enzymes that cleave V region genes can be suitably located on the 5' side of expression vectors containing DNA encoding constant regions (C regions) of the desired antibody. The resulting expression vectors and V region genes, digested with the same combination of restriction enzymes, are fused in frame to each other to construct chimeric antibody expression vectors.

To produce the desired antibody, the antibody gene can be inserted into expression vectors such that the gene is operably linked to control sequences. Control sequences for antibody expression include, for example, enhancers and promoters. A signal sequence may also be added to the N-terminus of the antibody for extracellular secretion of the expressed antibody. For example, a peptide having the amino acid sequence MGWSCIILFLVATATGVHS (SEQ ID NO: 13) can be used as a signal sequence. Any other suitable signal sequence may be added to the antibody. The expressed polypeptide is cleaved at the C-terminus of the signal sequence, and the cleaved polypeptide can be secreted extracellularly as a mature polypeptide. Using these expression vectors, suitable host cells can be transformed to produce recombinant cells expressing DNA encoding the desired antibody.

To express antibody genes, the DNA encoding the heavy chain (H chain) and the DNA encoding the light chain (L chain) are separately inserted into different expression vectors. The same host cell can be cotransfected with a vector into which a heavy chain has been inserted and a vector into which a light chain has been inserted, thereby allowing the expression of antibody molecules containing the H and L chains. Alternatively, the DNAs encoding the heavy chain and the light chain can be inserted into a single expression vector, which can then be used to transform a host cell (see WO 1994011523).

Many combinations of host cells and expression vectors are known in the art to produce antibodies by transferring isolated antibody genes into suitable hosts. All of these expression systems can be used to isolate the antigen binding molecules of the present invention. In the case of using eukaryotic cells as host cells, animal, plant or fungal cells can suitably be used. Specifically, examples of animal cells may include the following cells:

(1) mammalian cells such as CHO (Chinese hamster ovary cell line), COS (monkey kidney cell line), myeloma cells (Sp2/O, NS0, etc.), BHK (baby hamster kidney cell line), HEK293 (human embryonic kidney cell line with cleaved adenoviral (Ad)5 DNA), PER.C6 cells (human embryonic retinal cell line transformed with adenovirus type 5 (Ad5) E1A and E1B genes), Hela and Vero (Current Protocols in Protein Science , May, 2001, Unit 5.9, Table 5.9.1);

(2) amphibian cells such as Xenopus oocytes; And

(3) insect cells such as sf9, sf21 and Tn5.

Alternatively, expression systems for antibody genes using cells derived from the genus Nicotiana (eg, Nicotiana tabacum) are known in the art as plant cells. Cultured callus cells can be suitably used to transform plant cells.

The following cells can be used as fungal cells:

- cells derived from yeasts of the genus Saccharomyces (for example, Saccharomyces cerevisiae) and the genus Pichia (for example, Pichia pastoris), and

- cells derived from filamentous fungi of the genus Aspergillus (for example, Aspergillus niger).

Expression systems for antibody genes using prokaryotic cells are also known in the art. In the case of using, for example, bacterial cells, cells of bacteria such as E. coli and Bacillus subtilis can be suitably used. Expression vectors containing the antibody gene of interest are transferred into these cells by transformation. The transformed cells are cultured in vitro, and the desired antibody can be produced from the resulting transformed cell cultures.

In addition to host cells, transgenic animals can be used to produce recombinant antibodies. In particular, the desired antibody can be obtained from animals transfected with the gene encoding the antibody. For example, antibody genes can be inserted in frame into genes encoding proteins specifically produced in milk, constructing fusion genes. As the protein secreted in milk, for example, goat β-casein and the like can be used. DNA fragments containing gene fusions having an antibody gene insert are injected into goat embryos, which are in turn implanted into female goats. From the milk produced by transgenic goats (or their progeny) born to embryo-implanted goats, the desired antibody can be produced as a fusion protein with milk protein. In addition, hormones can be administered to transgenic goats to increase the amount of milk containing the desired antibody produced by transgenic goats (Bio/Technology (1994), 12 (7), 699-702).

In the case of administering the antigen binding molecule described herein to a human, the antigen binding domain for the molecule can be suitably selected as an antigen binding domain derived from a genetically recombinant antibody that is artificially constructed, for example, to reduce the heteroantigenicity of a human. Genetically recombinant antibody includes, for example, humanized antibodies. These engineered antibodies are suitably produced using a method known in the art.

Each antibody variable region used to provide an antigen binding domain in an antigen binding molecule described herein typically consists of three complementarity determining regions (CDRs) flanked by four framework regions (FRs). CDRs are regions that essentially determine the binding specificity of an antibody. CDRs have a variety of amino acid sequences. On the other hand, FRs are largely formed by amino acid sequences that are highly identical even among antibodies with different binding specificities. Thus, in general, the binding specificity of a particular antibody can be grafted into other antibodies by grafting CDRs.

Humanized antibodies are also called reconstituted human antibodies. In particular, what is known in the art is, for example, a humanized antibody consisting of a human antibody into which the CDRs of a non-human animal (eg, mouse) antibody are transplanted. General gene recombination approaches are also known to produce humanized antibodies. In particular, for example, overlap PCR is known in the art as a method for transplanting mouse antibody CDRs into human FRs. In overlapping PCR, a nucleotide sequence encoding each CDR of the mouse antibody to be transplanted is added to the primers for the synthesis of human FR antibodies. Primers are prepared for each of the four FRs. For transplantation of mouse CDRs into human FRs, it is generally considered advantageous to select human FRs that are highly identical to mouse FRs to preserve the functions of the CDRs. In particular, it is generally preferable to use human FRs containing amino acid sequences highly identical to the amino acid sequences of the FRs adjacent to the mouse CDRs to be transplanted.

The nucleotide sequences to be linked are designed such that the sequences are linked in frame with each other. Nucleotide sequences encoding human FR are synthesized individually using their corresponding primers. The resulting products contain DNA encoding a mouse CDR fused to each human FR encoding sequence. The nucleotide sequences encoding the mouse CDRs are designed such that the nucleotide sequence in each product overlaps with another nucleotide sequence. The overlapping CDR regions in the products synthesized from human antibody genes as templates are then annealed to each other for a complementary strand synthesis reaction. Using this reaction, human FR sequences are linked via mouse CDR sequences.

Finally, the full-length V region gene sequence, containing the linked three CDRs and four FRs, is amplified using primers, each of which anneals to its 5' and 3' ends and has an associated recognition sequence for the corresponding restriction enzyme. The DNA thus obtained and the DNA encoding the C region of a human antibody can be inserted into expression vectors so that the DNAs are fused in frame to obtain humanized antibody expression vectors. These insert vectors are transferred into hosts to produce recombinant cells. The recombinant cells are then cultured to express DNA encoding the humanized antibody to produce humanized antibodies in cultured cell cultures (EP 239400 and WO 1996002576).

The humanized antibodies thus obtained can be assessed for their antigen binding activity using a quantitative or qualitative assay, thereby selecting suitable human antibody FRs that allow the CDR to form a suitable antigen binding site when coupled through the CDR. If necessary, the amino acid residue(s) of FR can be replaced so that the CDRs of the resulting reconstituted human antibody form a suitable antigen binding site. For example, the amino acid sequence of FR can be mutated using the PCR method used to transplant a mouse CDR into a human FR. In particular, mutation of a partial nucleotide sequence can be introduced into primers that anneal to the FR nucleotide sequence. The FR nucleotide sequence synthesized using such primers contains the mutation introduced in this way. Such antibody variants having the amino acid(s) replaced can be assessed for their antigen binding activity using the same assay as described above, thereby selecting FR sequence variants having the desired properties (Sato et al., Cancer Res (1993) 53, 851-856).

Alternatively, the desired human antibody can be produced by DNA immunization using transgenic animals having all human antibody gene repertoires (see WO 1993012227, WO 1992003918, WO 1994002602, WO 1994025585, WO 1996034096 and WO 1996033735), as immunized animals.

In addition, a method for producing human antibodies by panning using human antibody libraries is also known. For example, human antibody V regions are expressed as a single chain antibody (scFv) on the surface of phages by a phage display method. Phage expressing the antigen-binding scFv can be selected. The gene of the selected phage can be analyzed to determine the DNA sequences encoding the V regions of the human antibody binding to the antigen. Once the DNA sequence of the antigen binding scFv has been determined, the V region sequences can be fused in frame to the C region sequences of the desired human antibody and then inserted into appropriate expression vectors to produce expression vectors. Expression vectors are transferred into preferred expression cells as illustrated above. The genes encoding human antibodies are expressed by cells to produce human antibodies. These methods are already known in the art (see WO 1992001047, WO 19992020791, WO 1993006213, WO 1993011236, WO 1993019172, WO 1995001438 and WO 1995015388).

An “antigen binding domain” as described herein refers to a region that specifically binds to part or all of an antigen in a complementary manner. Examples of an antigen binding domain may include domains having an antibody antigen binding domain. Examples of an antibody antigen binding domain may include CDRs and variable regions. When a CDR is used as the antigen binding domain of an antibody, the antigen binding domain may contain all of the 6 CDRs contained in the antibody, or it may contain one or two or more of these CDRs. Each CDR contained as an antibody binding region may have, for example, a deletion, substitution, insertion and/or insertion of amino acids, or a portion of the CDR may be used provided that the contained CDR has antigen binding activity. An antibody directed against an antigen that has a large molecular weight can only bind to a specific site on the antigen. This particular region is called an epitope. The antigen binding domain may be formed by one or more antibody variable domains. Preferably, the antigen binding domain comprises an antibody light chain variable region (VL) and an antibody heavy chain variable region (VH). Preferred examples of such an antigen binding domain include "scFv (single chain Fv)", "single chain antibody", "Fv", "scFv2 (single chain Fv2)", "Fab", "dianbody", "linear antibody" and "F(ab') 2".

An "antibody variable region" as used herein refers to the region that is contained in each of the light and heavy chains of an antibody molecule and contains the amino acid sequences of the complementarity determining regions (CDRs; i.e., CDR1, CDR2, and CDR3) and framework regions (FR). VH is the heavy chain variable region. VL is the light chain variable region. In accordance with the method used in the present invention, the amino acid positions assigned to the CDR and FR are determined in accordance with the Kabat method (Sequence of Proteins of Immunological Interest, National Institute of Health, Bethesda, Md., 1987 and 1991). In the present specification, amino acids in the antibody or antigen-binding fragment are also numbered according to Kabat numbering corresponding to Kabat amino acid positions.

The term “complementarity determining regions (CDRs; ie, CDR1, CDR2, and CDR3)” as used herein refers to amino acid residues in the variable region of an antibody whose existence is required for antigen binding. Each variable region contains three CDR regions, generally referred to as CDR1, CDR2 and CDR3. Each complementarity determining region may contain amino acid residues from a “complementarity determining region” as described by Kabat (i.e., residues 24-34 (CDR1), 50-56 (CDR2) and 89-97 (CDR3) in the light chain variable region and residues 31-35 (CDR1), 50-65 (CDR2) and 95-102 (CDR3) in the heavy chain variable region; Kabat et al., Sequences of Proteins of Immunological Interest, 5th edition, Public Health Service, National Institute of Health , Bethesda, MD. (1991)) and/or residues from the “hypervariable loop” (i.e., residues 26-32 (CDR1), 50-52 (CDR2) and 91-96 (CDR3) in the light chain variable region and residues 26-32 (CDR1), 53-55 (CDR2) and 96-101 (CDR3) in the heavy chain variable region; Chothia and Lesk, J. Mol. Biol. (1987) 196, 901-917). In some cases, each complementarity determining region may contain amino acids from both the CDR region and the hypervariable loop determined by the Kabat method.

The term "Fab" fragment includes the light chain variable and constant regions and the heavy chain variable and first constant region (CH1). The F(ab')2 antibody fragment contains a pair of Fab fragments, typically covalently linked at regions near their C-terminal ends between cysteine residues in the hinge region. Other chemical linkages of antibody fragments to which the present invention relates are also known in the art.

The term "single chain Fv" or "scFv" antibody fragment includes the VH and VL regions of the antibody, which in turn form a single polypeptide chain. Typically, the Fv polypeptide also contains a polypeptide linker between the VH and VL regions. This linker allows the formation of the structure desired for antigen binding by the scFv. scFv are reviewed, for example, in Pluckthun, The Pharmacology of Monoclonal Antibodies (1994) Vol. 113, 269-315 (Rosenburg and Moore ed., Springer-Verlag, New York).

The term "diantibody" refers to a small antibody fragment having two antigen-binding sites. This antibody fragment contains a heavy chain variable region (VH) linked to a light chain variable region (VL) in one polypeptide chain (VH and VL). A linker that is too short to pair these two regions on the same polypeptide chain is used to force these regions to pair with their complementarity determining regions on another polypeptide chain so that two antigen binding sites are formed. The diantibody is described in detail, for example, in patent documents such as European Patent No. 404097 and WO 1993011161 and non-patent documents such as Holliger et al., Proc. Natl. Acad. Sci. USA (1993) 90, 6444-6448.

The term "linear antibody" refers to the antibody described in Zapata et al., Protein Eng. (1995) 8(10), 1057-1062. Briefly, this antibody contains a pair of tandem Fd segments (VH-CH1-VH-CH1), which form a pair of antigen binding domains with a complementary light chain polypeptide. A linear antibody may be bispecific or monospecific.

Antigen

The “antigen” described herein is not limited to a particular structure, as long as the antigen contains an epitope to which the antigen binding domain binds. On the other hand, the antigen may be an inorganic substance or it may be an organic substance. Examples of antigen may include the following molecules: 17-IA, 4-1 BB, 4Dc, 6-keto-PGF1a, 8-iso-PGF2a, 8-oxo-dG, adenosine A1 receptor, A33, ACE, ACE-2, activin, activin A, activin AB, activin B, activin C, activin RIA, activin RIA ALK-2, activin RIB ALK-4, activin RIIA, activin RIIB, ADAM, ADAM10, ADAM12, ADAM15, ADAM17/TACE, ADAM8, ADAM9, ADAMTS , ADAMTS4, ADAMTS5, addressin, aFGF, ALCAM, ALK, ALK-1, ALK-7, alpha-1-antitrypsin, alpha-V/beta-1 antagonist, ANG, Ang, APAF-1, APE, APJ, APP, APRIL, AR, ARC, ART, artemin, anti-Id, ASPARTIC, atrial natriuretic factor, integrin av/b3, Ax1, b2M, B7-1, B7-2, B7-H, B-lymphocyte stimulator (BlyS), BACE , BACE-1, Bad, BAFF, BAFF-R, Bag-1, BAK, Bax, BSA-1, BCAM, Bc1, BCMA, BDNF, b-ECGF, bFGF, BID, Bik, BIM, BLC, BL-CAM , BLK, BMP, BMP-2 BMP-2a, osteogenin BMP-3, BMP-4 BMP-2b, BMP-5, BMP-6 Vgr-1, BMP-7 (OP-1), BMP-8 (BMP- 8a, OP-2), BMPR, BMPR-IA (ALK-3), BMPR-IB (ALK-6), BRK-2, RPK-1, BMPR-II (BRK-3), BMP, b-NGF, BOK, bombesin, bone-derived neurotrophic factor, BPDE, BPDE-DNA, BTC, complement factor 3 (C3), C3a, C4, C5, C5a, C10, CA125, CAD-8, calcitonin, cAMP, carcinoembryonic antigen (CEA), associated with malignant tumor antigens, cathepsin A, cathepsin B, cathepsin C/DPPI, cathepsin D, cathepsin E, cathepsin H, cathepsin L, cathepsin O, cathepsin S, cathepsin V, cathepsin X/Z/P, CBL, CCI, CCK2, CCL, CCL1, CCL11, CCL12, CCL13, CCL14, CCL15, CCL16, CCL17, CCL18, CCL19, CCL2, CCL20, CCL21, CCL22, CCL23, CCL24, CCL25, CCL26, CCL27, CCL28, CCL3, CCL4, CCL5, CCL6, CCL7, CCL8, CCL9/10, CCR, CCR1, CCR10, CCR10, CCR2, CCR3, CCR4, CCR5, CCR6, CCR7, CCR8, CCR9, CD1, CD2, CD3, CD3E, CD4, CD5, CD6, CD7, CD8, CD10, CD11a, CD11b, CD11c, CD13, CD14, CD15, CD16, CD18, CD19, CD20, CD21, CD22, CD23, CD25, CD27L, CD28, CD29, CD30, CD30L, CD32, CD33 (p67 protein), CD34, CD38, CD40, CD40L, CD44, CD45, CD46, CD49a, CD52, CD54, CD55, CD56, CD61, CD64, CD66e, CD74, CD80 (B7-1), CD89, CD95, CD123, CD137, CD138, CD140a, CD146 , CD147, CD148, CD152, CD164, CEACAM5, CFTR, cGMP, CINC, botulinum toxin, Clostridium perfringens toxin, CKb8-l, CLC, CMV, CMV UL, CNTF, CNTN-1, COX, C-Ret, CRG-2 , CT-1, CTACK, CTGF, CTLA-4, CX3CL1, CX3CR1, CXCL, CXCL1, CXCL2, CXCL3, CXCL4, CXCL5, CXCL6, CXCL7, CXCL8, CXCL9, CXCL10, CXCL11, CXCL12, CXCL13, CXCL14, CXCL15, CXCL16 , CXCR, CXCR1, CXCR2, CXCR3, CXCR4, CXCR5, CXCR6, cytokeratin tumor-associated antigens, DAN, DCC, DcR3, DC-SIGN, decay accelerating factor, des-(1-3)-IGF-I (IGF- 1 brain), Dhh, digoxin, DNAM-1, DNAse, Dpp, DPPIV/CD2 6, Dtk, ECAD, EDA, EDA-A1, EDA-A2, EDAR, EGF, EGFR (ErbB-1), EMA , EMMPRIN, ENA, endothelin receptor, enkephalinase, eNOS, Eot, eotaxin 1, EpCAM, ephrin B2/EphB4, EPO, ERCC, E-selectin, ET-1, factor IIa, factor VII, factor VIIIc, fibroblast activating factor IX protein (FAP), Fas, FcRl, FEN-1, ferritin, FGF, FGF-19, FGF-2, FGF3, FGF-8, FGFR, FGFR-3, fibrin, FL, FLIP, Flt-3, Flt-4 , follicle-stimulating hormone, fractalkine, FZD1, FZD2, FZD3, FZD4, FZD5, FZD6, FZD7, FZD8, FZD9, FZD10, G250, Gas6, GCP-2, GCSF, GD2, GD3, GDF, GDF-1, GDF-3 ( Vgr-2), GDF-5 (BMP-14, CDMP-1), GDF-6 (BMP-13, CDMP-2), GDF-7 (BMP-12, CDMP-3), GDF-8 (myostatin) , GDF-9, GDF-15 (MIC-1), GDNF, GDNF, GFAP, GFRa-1, GFR-alpha 1, GFR-alpha 2, GFR-alpha 3, GITR, glucagon, Glut4, glycoprotein IIb/IIIa ( GPIIb/IIIa), GM-CSF, gpl30, gp72, GRO, growth hormone releasing factor, hapten (NP-cap or NIP-cap), HB-EGF, HCC, HCMV gB envelope glycoprotein, HCMV gH envelope glycoprotein, HCMV UL , hematopoietic growth factor (HGF), Hep B gp120, heparanase, Her2, Her2/neu (ErbB-2), Her3 (ErbB-3), Her4 (ErbB-4), glycoprotein gB of herpes simplex virus (HSV), glycoprotein gD HSV, HGFA, high molecular weight melanoma-associated antigen (HMW-MAA), HIV gp120, V3 loop gp 120 IIIB HIV, HLA, HLA-DR, HM1.24, HMFG PEM, HRG, Hrk, human cardiac myosin, human cytomegalovirus (HCMV) ), human growth hormone (HGH), HVEM, I-309, IAP, ICAM, ICAM-1, ICAM-3, ICE, ICOS, IFNg, Ig, IgA receptor, IgE, IGF, IGF-binding protein, IGF-1R , IGFBP, IGF-I, IGF-II, IL, IL-1, IL-1R, IL-2, IL-2R, IL-4, IL-4R, IL-5, IL-5R, IL-6, IL -6R, IL-8, IL-9, IL-10, IL-12, IL-13, IL-15, IL-18, IL-18R, IL-23, interferon (INF)-alpha, INF-beta, INF-gamma, inhibin, iNOS, insulin A chain, insulin B chain, insulin-like growth factor 1, integrin alpha 2, integrin alpha 3, integrin alpha 4, integrin alpha 4/beta 1, integrin alpha 4/beta 7, integrin alpha 5 (alpha V), integrin alpha 5/beta 1, integrin alpha 5/beta 3, integrin alpha 6, integrin beta 1, integrin beta 2, interferon-gamma, IP-10, I-TAC, JE, kallikrein 2, kallikrein 5, kallikrein 6, kallikrein 11, kallikrein 12, kallikrein 14, kallikrein 15, kallikrein L1, kallikrein L2, kallikrein L3, kallikrein L4, KC, KDR, keratinocyte growth factor (KGF), laminin 5, LAMP, LAP, LAP ( TGF-1), latent TGF-1, latent TGF-1 bpl, LBP, LDGF, LECT2, lefty, Lewis-Y antigen, Lewis-Y related antigen, LFA-1, LFA-3, Lfo, LIF, LIGHT, lipoprotein , LIX, LKN, Lptn, L-selectin, LT-a, LT-b, LTB4, LTBP-1, lung surface protein, luteinizing hormone, lymphotoxin-beta receptor, Mac-1, MAdCAM, MAG, MAP2, MARC, MCAM , MSAM, MSC-2, MCP, M-CSF, MDC, Mer, metalloproteinases, MGDF receptor, MGMT, MHC (HLA-DR), MIF, MIG, MIP, MIP-1-alpha, MK, MMAC1, MMP, MMP -1, MMP-10, MMP-11, MMP-12, MMP-13, MMP-14, MMP-15, MMP-2, MMP-24, MMP-3, MMP-7, MMP-8, MMP-9 , MPIF, Mpo, MSK, MSP, mucin (Mucl), MUC18, Mullerian inhibitory substance, Mug, MuSK, NAIP, NAP, NCAD, N-C adherin, NCA 90, NCAM, NCAM, neprilysin, neurotrophin-3, -4 or - 6, neuroturin, nerve growth factor (NGF), NGFR, NGF-beta, nNOS, NO, NOS, Npn, NRG-3, NT, NTN, OB, OGG1, OPG, OPN, OSM, OX40L, OX40R, pl50, p95 , PADPr, parathyroid hormone, PARC, PARP, PBR, PBSF, PCAD, P-cadherin, PCNA, PDGF, PDGF, PDK-1, PECAM, PEM, PF4, PGE, PGF, PGI2, PGJ2, PIN, PLA2, placental alkaline phosphatase (PLAP), P1GF, PLP, PP14, proinsulin, prorelaxin, protein C, PS, PSA, PSCA, prostate-specific membrane antigen (PSMA), PTEN, PTHrp, Ptk, PTN, R51, RANK, RANKL, RANTES, RANTES, A -relaxin chain, relaxin B chain, renin, respiratory syncytial virus (RSV) F, RSV Fgp, Ret, rheumatoid factor, RLIP76, RPA2, RSK, S100, SCF/KL, SDF-1, SERINE, serum albumin, sFRP -3, Shh, SIGIRR, SK-1, SLAM, SLPI, SMAC, SMDF, SMOH, SOD, SPARC, Stat, STEAP, STEAP-II, TACE, TACI, TAG-72 (tumor-associated glycoprotein-72), TARC , TCA-3, T cell receptor (e.g. T cell receptor alpha/beta), TdT, TECK, TEM1, TEM5, TEM7, TEM8, TERT, testis PLAP-like alkaline phosphatase, TfR, TGF, TGF-alpha, TGF-beta, panspecific TGF-beta, TGF-beta RI (ALK-5), TGF-beta RII, TGF-beta RIIb, TGF-beta RIII, TGF-beta 1, TGF-beta 2, TGF-beta 3, TGF -beta 4, TGF-beta 5, thrombin, thymic Ck-1, thyroid stimulating hormone, Tie, TIMP, TIQ, tissue factor, TMEFF2, Tmpo, TMPRSS2, TNF, TNF-alpha, TNF-alpha/beta, TNF-beta 2 , TNFc, TNF-RI, TNF-RII, TNFRSF10A(TRAIL R1 Apo-2, DR4), TNFRSF10B (TRAIL R2 DR5, KILLER, TRICK-2A, TRICK-B), TNFRSF10C (TRAIL R3 DcR1, LIT, TRID), TNFRSF10D (TRAIL R4 DcR2, TRUNDD), TNFRSF11A (RANK ODF R, TRANCER), TNFRSF1IB (OPG OCIF, TR1), TNFRSF12 (TWEAK R FN14), TNFRSF13B (TACI), TNFRSF13C (BAFF R), TNFRSF14 (HVEM ATAR, HveA , LIGHT R, TR2), TNFRSF16 (NGFR p75NTR), TNFRSF17 (BCMA), TNFRSF18 (GITR AITR), TNFRSF19 (TROY TAJ, TRADE), TNFRSF19L (RELT), TNFRSF1A (TNF RI CD120a, p55-60), TNFRSF1B ( TNF RII CD120b, p75-80), TNFRSF2 6 (TNFRH3), TNFRSF3 (LTbR TNF RIII, TNFC R), TNFRSF4 (OX40 ACT35, TXGP1 R), TNFRSF5 (CD40 p50), TNFRSF6 (Fas Apo-1, APT1, CD95 ), TNFRSF6B (DcR3 M68, TR6), TNFRSF7 (CD27), TNFRSF8 (CD30), TNFRSF9 (4-1BB CD137, I LA), TNFRSF21 (DR6), TNFRSF22 (DcTRAIL R2 TNFRH2), TNFRST23 (DcTRAIL R1 TNFRH1), TNFRSF25 (DR3 Apo-3, LARD, TR-3, TRAMP, WSL-1), TNFSF10 (TRAIL Apo-2 ligand, TL2), TNFSF11 (TRANCE/RANK ODF ligand, OPG ligand), TNFSF12 (TWEAK Apo-3 ligand , ligand DR3), TNFSF13 (APRIL TALL2), TNFSF13B (BAFF BLYS, TALL1, THANK, TNFSF20), TNFSF14 (ligand LIGHT HVEM, LTg), TNFSF15 (TL1A/VEGI), TNFSF18 (ligand GITR ligand AITR, TL6), TNFSF1A (TNF-a connectin, DIF, TNFSF2), TNFSF1B (TNF-b LTa, TNFSF1), TNFSF3 (LTb TNFC, p33), TNFSF4 (OX40 ligand gp34, TXGP1), TNFSF5 (CD40 ligand CD154, gp39, HIGM1, IMD3, TRAP), TNFSF6 (Fas ligand of Apo-1 ligand, APT1 ligand), TNFSF7 (CD27 ligand of CD70), TNFSF8 (CD30 ligand of CD153), TNFSF9 (4-1BB ligand of CD137 ligand), TP-1, t-PA , Tro, TRAIL, TRAIL R, TRAIL-R1, TRAIL-R2, TRANCE, transferrin receptor, TRF, Trk, TROP-2, TSG, TSLP, tumor-associated antigen CA125, tumor-associated antigen expressing Lewis-Y-related carbohydrate , TWEAK, TXB2, Ung, uPAR, uPAR-1, urokinase, VCAM, VCAM-1, VECAD, VE-cadherin, VE-cadherin-2, VEFGR-1 (flt-1), VEGF, VEGFR, VEGFR-3 ( flt-4), VEGI, VIM, viral antigens, VLA, VLA-1, VLA-4, integrin VNR, von Willebrand factor, WIF-1, WNT1, WNT2, WNT2B/13, WNT3, WNT3A, WNT4, WNT5A, WNT5B , WNT6, WNT7A, WNT7B, WNT8A, WNT8B, WNT9A, WNT9A, WNT9B, WNT10A, WNT10B, WNT11, WNT16, XCL1, XCL2, XCR1, XCR1, XEDAR, XIAP, XPD, HMGB1, IgA, Aβ, CD81, CD97, CD98 , DDR1, DKK1, EREG, Hsp90, IL-17/IL-17R, IL-20/IL-20R, oxidized LDL, PCSK9, prekallikrein, RON, TMEM16F, SOD1, chromogranin A, chromogranin B, tau protein, VAP1, high molecular weight kininogen, IL-31, IL-31R, Nav1.1, Navl.2, Navl.3, Nav1.4, Nav1.5, Nav1.6, Nav1.7, Nav1.8, Nav1.9, EPCR, C1, C1q, C1r, C1s, C2, C2a, C2b, C3, C3a, C3b, C4, C4a, C4b, C5, C5a, C5b, C6, C7, C8, C9, factor B, factor D, factor H, properdin, sclerostin , fibrinogen, fibrin, prothrombin, thrombin, tissue factor, factor V, factor Va, factor VII, factor VIIa, factor VIII, factor VIIIa, factor IX, factor IXa, factor X, factor Xa, factor XI, factor XIa, factor XII , factor XIIa, factor XIII, factor XIIIa, TFPI, antithrombin III, EPCR, thrombomodulin, TAPI, tPA, plasminogen, plasmin, PAI-1, PAI-2, GPC3, syndecan-1, syndecan-2, syndecan-3, syndecan -4, LPA, S1P, acetylcholine receptor, AdipoR1, AdipoR2, ADP ribosyl cyclase-1, alpha-4/beta-7 integrin, alpha-5/beta-1 integrin, alpha-v/beta-6 integrin, alpha-v integrin /beta-1, angiopoietin ligand-2, Angpt12, anthrax, cadherin, carbonic anhydrase-IX, CD105, CD155, CD158a, CD37, CD49b, CD51, CD70, CD72, claudin 18, Clostridium difficile toxin, CS1, ligand 4 delta -like protein, DHICA oxidase, Dickkopf-1 ligand, dipeptidyl peptidase IV, EPOR, RSV F protein, factor Ia, FasL, folate receptor alpha, glucagon receptor, glucagon-related peptide receptor 1, glutamate carboxypeptidase II, GMCSFR, E2 glycoprotein hepatitis C virus, hepcidin, IL-17 receptor, IL-22 receptor, IL-23 receptor, IL-3 receptor, Kit tyrosine kinase, leucine-rich alpha-2-glycoprotein 1 (LRG1), lysosphingolipid receptor, membrane glycoprotein OX2, mesothelin , MET, MICA, MUC-16, myelin-associated glycoprotein, neuropilin-1, neuropilin-2, Nogo receptor, PLXNA1, PLXNA2, PLXNA3, PLXNA4A, PLXNA4B, PLXNB1, PLXNB2, PLXNB3, PLXNC1, PLXND1, programmed cell death ligand 1 , PC9 proprotein convertase, P-selectin glycoprotein ligand-1, RAGE, reticulon 4, RF, RON-8, SEMA3A, SEMA3B, SEMA3C, SEMA3D, SEMA3E, SEMA3F, SEMA3G, SEMA4A, SEMA4B, SEMA4C, SEMA4D, SEMA4F, SEMA4G, SEMA5A, SEMA5B, SEMA6A, SEMA6B, SEMA6C, SEMA6D, SEMA7A, Shiga-like toxin II, sphingosine-1-phosphate receptor-1, ST2, staphylococcal lipoteichoic acid, tenascin, TG2, thymic stromal lymphopoietin receptor, TNF superfamily receptor 12A, NMB transmembrane glycoprotein , TREM-1, TREM-2, trophoblast glycoprotein, TSH receptor, TTR, tubulin, ULBP2 and receptors for hormones or growth factors. Other examples of antigen may include soluble receptor molecules described above that exist in body fluids in vivo without being anchored to cells.

An epitope, which means an antigenic determinant found in an antigen, means a region on an antigen to which an antigen binding domain in an antigen binding molecule described herein binds. Thus, for example, an epitope may be determined by its structure. Alternatively, the epitope is determined by the antigen binding activity of the antigen binding molecule that recognizes the epitope. An epitope in an antigenic peptide or polypeptide may be defined by the amino acid residues that make up the epitope. Alternatively, the sugar chain epitope may be determined by the particular structure of the sugar chain.

A linear epitope refers to an epitope containing an epitope that is recognized through its primary amino acid sequence. A linear epitope typically contains at least 3 and most often at least 5, such as about 8 to about 10 or 6 to 20 amino acids, in a unique sequence.

In contrast to a linear epitope, a conformational epitope refers to an epitope that is contained in a primary amino acid sequence containing a component other than one specific component of the epitope being recognized (eg, an epitope whose primary amino acid sequence may not be recognized by the antibody that detects the epitope). A conformational epitope may contain an increased number of amino acids compared to a linear epitope. The antibody recognizes a conformational epitope by recognizing the three-dimensional structure of the antigenic peptide or protein. For example, a protein molecule can be folded to form a three-dimensional structure. In such a case, certain amino acids and/or polypeptide backbone forming the conformational epitope are arranged in parallel to allow the antibody to recognize the epitope. The conformation of the epitope is determined by methods including, for example, but not limited to, x-ray crystallography, two-dimensional nuclear magnetic resonance spectroscopy and site-specific spin labeling and electron paramagnetic resonance spectroscopy. See, for example, Epitope Mapping Protocols in Methods in Molecular Biology (1996), Vol. 66, Morris ed.

Gene recombination approach

The term "codon set" refers to the set of different nucleotide triplet sequences that are used to encode the desired amino acid. One set of oligonucleotides includes sequences that reflect every possible combination of nucleotide triplets provided by a set of codons and encode the desired group of amino acids. Such a set of oligonucleotides can be synthesized, for example, using a solid-phase method. The IUB provides a standard codon naming system. This code is known in the art. The codon set is usually indicated using three capital letters, for example, NNK, NNS, DVK or DVD.

Code IUB

G: guanine

A: adenine

T: thymine

C: cytosine

R (A or G)

Y (C or T)

M (A or C)

K (G or T)

S (C or G)

W (A or T)

N (A, S or T)

B (C, G or T)

V (A, C or G)

D (A, G or T)

N (A, C, G or T).

For example, in the DVK codon set, D represents an A, G, or T nucleotide; V represents A, G or C; and K is G or T. This set of codons represents 18 different codons and can code for the amino acids Ala, Trp, Tyr, Lys, Thr, Asn, Lys, Ser, Arg, Asp, Glu, Gly and Cys.

An oligonucleotide having a "degenerate" nucleotide at a particular position is designed in a manner known in the field to which the present invention relates (eg, Garrard and Henner, Gene (1993) 128, 103-109). A set of oligonucleotides having this particular type of codon pattern can be synthesized using a commercially available nucleic acid synthesis apparatus (available, for example, from Applied Biosystems, Inc., Foster City, CA), or it can be obtained as commercially available products (for example , Life Technologies, Rockville, MD). Thus, a set of synthetic oligonucleotides having a specific set of codons usually contains many oligonucleotides that differ in sequence. In a non-limiting aspect of the present invention, the oligonucleotides may contain, for example, restriction enzyme sites suitable for cloning.

The terms "cell", "cell system" and "cell culture" are used synonymously herein. Such designations may include each descendant cell or cell system. Thus, for example, the terms “transformant” and “transformed cell” include primary target cells and cultures derived from them, regardless of the number of passages. It should also be understood that each descendant may not have exactly identical DNA contained within them, due to intentional or accidental mutation. These terms may include a descendant variant having substantially the same functions or biological activity as the originally transformed cells screened for. If another designation is intended, it will be obvious from the context.

The term "control sequence" as used to refer to the expression of a coding sequence refers to the DNA nucleotide sequence necessary for the expression of an operably linked coding sequence in a particular host organism. Control sequences useful in prokaryotes, for example, include promoters, optional operator sequences, ribosome binding sites, and likely other sequences that remain to be explored. The use of a promoter, polyadenylation signal and enhancer is known in the art for the expression of a coding sequence in eukaryotic cells.

The expression "operably linked" in relation to a nucleic acid means that the nucleic acid has a functional relationship with another nucleic acid sequence. For example, a presequence or secretory leader DNA is operably linked to the DNA of a polypeptide when it is expressed as a precursor protein involved in the secretion of the polypeptide. A promoter or enhancer, when it influences the transcription of a coding sequence, is operably linked to the sequence. Alternatively, the ribosome binding site, when positioned to facilitate translation, is operably linked to the coding sequence. Typically, the expression "operably linked" means that the linked DNA sequences are arranged sequentially and that the secretory leader sequence, for example, is present in sequential order in the reading frame. However, the enhancer does not have to be located sequentially. This connection is achieved by ligation at appropriate ligation sites. In the absence of such regions, a synthetic oligonucleotide adapter or linker is used in accordance with standard practice. Alternatively, bound nucleic acids can be obtained using the overlap PCR approach described below.

"Ligation" refers to a method of forming a phosphodiester bond between two nucleic acid fragments. To ligate two fragments, the ends of the fragments must be compatible with each other. In some cases, these ends are compatible immediately after endonuclease cleavage. However, compatibility for ligation requires that the sticky ends, typically formed by endonuclease cleavage, first be converted to blunt ends. For conversion to blunt ends, each DNA is treated with approximately 10 units of Klenow fragment of DNA polymerase I or DNA polymerase T4 at 15°C for at least 15 minutes in the presence of four deoxyribonucleotide triphosphates in an appropriate buffer solution. The DNA is then purified by phenol-chloroform extraction and ethanol precipitation or silica purification. The DNA fragments to be fused are added in equimolar quantities to the solution. This solution contains ATP and ligase buffer, as well as approximately 10 units of ligase (eg, T4 DNA ligase) per 0.5 μg of DNA. When DNA is bound to a vector, the vector is first linearized through the cleavage action of an appropriate restriction endonuclease. The linearized fragment is then treated with bacterial alkaline phosphatase or calf intestinal phosphatase, thereby preventing the fragment from self-ligating during the ligation step.

The term "coat protein" refers to a protein, at least a portion of which is present on the surface of the viral particle. In terms of functionality, the envelope protein is a random protein that binds to viral particles during virus construction in host cells and maintains its bound state until the virus infects other host cells. The coat protein may be a major coat protein or may be a minor coat protein. The minor envelope protein is typically an envelope protein present in the viral capsid in an amount of preferably at least about 5, more preferably at least about 7, even more preferably at least about 10 or more copies of the protein per virion. The major envelope protein may be present in tens, hundreds, or thousands of copies per virion. Examples of the major coat protein include the p8 protein of filamentous phage.

The term "detection limit" for a chemical entity, such as an inorganic entity, an organic entity, or an organism in a particular assay, refers to the minimum concentration of the entity detected above the background level for the assay. For example, in phage ELISA, the “detection limit” for a particular phage exhibiting a particular antigen-binding fragment refers to the phage concentration at which the particular phage generates more ELISA signals than the signals generated by a control phage that does not exhibit the antigen-binding fragment.

The term "phage display" refers to an approach whereby polypeptide variants are displayed as fusion proteins with at least a portion of the coat proteins on the surface of phage particles, for example filamentous phages. Phage display is useful because a large library of randomized protein variants can be quickly and efficiently screened for sequence binding to a target antigen with high affinity. Phage display of peptide and protein libraries is used to screen millions of polypeptides for polypeptides with specific binding properties. The polyvalent phage display method is used to display small random peptides and small proteins by fusion with gene III or gene VIII of filamentous phage (Wells and Lowman, Curr. Opin. Struct. Biol. (1992) 3, 355-362; and references cited in specified document). Monovalent phage display involves fusing a library of proteins or peptides to gene III or a portion thereof, and expressing the fusion proteins at low levels in the presence of wild-type gene III protein such that each phage particle displays one copy or no copies of the fusion proteins. Monivalent phages have a lower avidity effect than multivalent phages and are thus screened based on endogenous ligand affinity using phagemid vectors that facilitate DNA manipulation (Lowman and Wells, Methods: A Companion to Methods in Enzymology ( 1991) 3, 205-216).

“Phagemid” refers to a plasmid vector having a bacterial origin of replication, such as Co1E1, and a copy of the bacteriophage intergenic region. As the phagemid, any bacteriophage known in the art can suitably be used, including, for example, filamentous bacteriophages and lambdoid bacteriophages. Typically, the plasmid additionally contains a selective marker for antibiotic resistance. DNA fragments cloned into these vectors can be grown as plasmids. When cells transformed with these vectors possess all the genes necessary to produce phage particles, the plasmid replication profile shifts to a rolling circle replication profile, producing single-strand copies of the plasmid DNA and packaging the phage particles. Phagmid can produce infectious or non-infectious phage particles. The term includes phagemids containing a phage coat protein gene or fragment thereof linked to a heterologous polypeptide gene by gene fusion such that the heterologous peptide is exposed on the surface of the phage particle.

The term "phage vector" means a double-stranded replicative bacteriophage that contains a heterologous gene and is capable of replicating. The phage vector has a phage origin of replication that allows phage replication and the formation of phage particles. Preferably the phage is a filamentous bacteriophage, for example phage M13, f1, fd or Pf3 or a derivative thereof, or a lambdoid phage, for example lambda, 21, phi80, phi81, 82, 424, 434, or any other phage or derivative thereof.

The term "oligonucleotide" refers to a short single-stranded or double-stranded polydeoxynucleotide that is chemically synthesized by a method known in the art (e.g., phosphotriester, phosphite, or phosphoramidite chemistry using a solid-phase approach, such as the approach described in EP266032; or the method via a deoxynucleotide H intermediate -phosphonate, described in Froeshler et al., Nucl. Acids. Res. (1986) 14, 5399-5407). Other methods for synthesizing oligonucleotides include polymerase chain reaction, described below, and other automated methods with primers and solid-support oligonucleotide synthesis. All of these methods are described in Engels et al., Agnew. Chem. Int. Ed. Engl. (1989) 28, 716-734. These methods are used provided that the entire nucleic acid sequence of the gene is known, and provided that a nucleic acid sequence complementary to the coding strand is available. Alternatively, possible nucleic acid sequences can be suitably predicted using known and preferred residues encoding each amino acid residue if the desired amino acid sequence is known. The oligonucleotide can be purified using polyacrylamide gels or molecular sieve columns or by precipitation.

The terms "fusion protein" and "fusion polypeptide" refer to a polypeptide having two segments covalently linked to each other. This polypeptide has various features defined by these segments. These features may represent in each case, for example, a biological property such as in vitro or in vivo activity. Alternatively, these features may be in each case a chemical or physical property, such as binding to a target antigen or catalysis of a reaction. These two segments can be linked either directly through a single peptide bond or through a peptide linker containing one or more amino acid residues. Typically, these two segments and the linker are present in the same reading frame. Preferably, the two polypeptide segments are derived from heterologous or different polypeptides.

The term "heterologous DNA" refers to random DNA that is transferred into host cells. The DNA can come from a variety of sources, including genomic DNA, cDNAs, synthetic DNAs, and fusions or combinations thereof. The DNA may include DNA from the same cells or the same cell type as the host or recipient cells, or DNA from a different cell type, such as a mammal or a plant. The DNA may optionally contain a marker or selection gene, such as an antibiotic resistance gene or a temperature tolerance gene, etc.

The term "highly diverse position" as used herein refers to an amino acid position in the light and heavy chain variable regions having several different amino acids present at that position when amino acid sequences of known and/or natural antibodies or antigen binding fragments are compared . Positions with a high degree of diversity are usually found in CDR areas. In one aspect, the position of high diversity in known and/or natural antibodies is effectively determined based on data provided in Kabat, Sequences of Proteins of Immunological Interest (National Institutes of Health, Bethesda, Md., 1987 and 1991). Many databases (http://vbase.mrc-cpe.cam.ac.uk/ and http://www.bioinf.org.uk/abs/index.html) on the Internet provide an extensive collection of many light and heavy sequences human circuits and their alignment. Information about these sequences and their alignment is useful for determining the position with a high degree of diversity in accordance with the present invention. In accordance with the present invention, an amino acid position is considered to be a position with a high degree of diversity if the amino acid position has a diversity of preferably from about 2 to about 20, preferably from about 3 to about 19, preferably from about 4 to about 18, preferably from 5 to 17, preferably 6 to 16, preferably 7 to 15, preferably 8 to 14, preferably 9 to 13, or preferably 10 to 12 possible amino acid residues at this position. In some embodiments, the amino acid position may have a diversity of preferably at least about 2, preferably at least about 4, preferably at least about 6, preferably at least about 8, preferably about 10, preferably about 12 possible different amino acid residues. In the present description, such amino acid residues are also referred to as non-strict residues. The term "non-random codon set" refers to a set of codons encoding a selected amino acid that partially (preferably completely) satisfies the amino acid selection criteria described herein. The term "random codon set" as used herein refers to a codon set having a combination of codons encoding a random amino acid selected from 20 amino acids (Ala/A, Leu/L, Arg/R, Lys/K, Asn/N , Met/M, Asp/D, Phe/F, Cys/C, Pro/P, Gln/Q, Ser/S, Glu/E, Thr/T, Gly/G, Trp/W, His/H, Tyr /Y, Ile/I and Val/V).

Library

In accordance with one aspect, the present invention relates to a library consisting essentially of a plurality of antigen binding molecules differing from each other in sequence, wherein the antigen binding domain in each of the antigen binding molecules contains at least one amino acid residue that modifies the antigen binding activity of the antigen binding molecule depending on ion concentration conditions. Preferable examples of ion concentration include metal ion concentration and hydrogen ion concentration.

The term “library” as used herein refers to a plurality of antigen binding molecules, a plurality of fusion polypeptides each containing antigen binding molecules, or nucleic acids or polynucleotides each encoding a different sequence. A plurality of antigen-binding molecules contained in a library, or a plurality of fusion polypeptides, each of which contains antigen-binding molecules, do not have the same sequence, and are antigen-binding molecules that differ from each other in sequence, or fusion polypeptides, each of which contains these antigen-binding molecules molecules.

The term "metal ion" as used herein refers to an ion of an element belonging to any of Group I, including the alkali metals except hydrogen and the copper family, Group II, including the alkaline earth metals and the zinc family, Group III, excluding boron , Group IV, excluding carbon and silicon, Group VIII, including the iron family and platinum family, and Subgroup A of Groups V, VI and VII, or a metal element such as antimony, bismuth or polonium. Metal atoms tend to release valence electrons, becoming cations. This property is called ionization tendency. Metals that have a high tendency to ionize are reported to be highly reactive.

Examples of the metal ion preferred in the present invention include calcium ions. Calcium ions are involved in the regulation of a variety of vital events, including contraction of muscles such as skeletal muscle, smooth muscle and cardiac muscle, activation (eg movement and phagocytosis) of leukocytes, activation (eg deformation and secretion) of platelets, activation of lymphocytes, activation (eg, histamine secretion) from mast cells, cellular responses mediated by the catecholamine receptor α or acetylcholine receptor, exocytosis, release of transmitters from neuronal terminals, and axonal flow in neurons. For example, troponin C, calmodulin, parvalbumin, and myosin light chain are known as intracellular calcium ion receptors, which have multiple calcium ion binding sites and may have originated from a common source in terms of molecular evolution. A large number of their binding motifs are also known. Well-known binding motifs are, for example, the cadherin domain, the EF hand contained in calmodulin, the C2 domain contained in protein kinase C, the Gla domain contained in coagulation factor IX, the C-type lectin contained in the asialoglycoprotein receptor or the mannose-binding receptor , domain A contained in the LDL receptor, annexin, thrombospondin type 3 domain, and EGF-like domain.

When the metal ion according to the present invention is a calcium ion, examples of calcium ion concentration conditions include low calcium ion concentration conditions and high calcium ion concentration conditions. The expression “binding activity varies depending on calcium ion concentration conditions” means that the antigen binding activity of each antigen binding molecule varies depending on the difference between low calcium ion concentration conditions and high calcium ion concentration conditions. Examples of this include higher antigen binding activity of an antigen binding molecule under conditions of high calcium ion concentration than under conditions of low calcium ion concentration. Another example of this involves higher antigen binding activity of an antigen binding molecule under conditions of low calcium ion concentration than under conditions of high calcium ion concentration.

As used herein, the high calcium ion concentration is not particularly limited to a single numerical value, and may preferably be a concentration selected from the range of 100 μM to 10 mM. In another aspect, the high calcium ion concentration may be a concentration selected from the range of 200 μM to 5 mM. In another aspect, the high calcium ion concentration may also be a concentration selected from the range of 500 μM to 2.5 mM. In an alternative aspect, the high calcium ion concentration may also be a concentration selected from the range of 200 μM to 2 mM. In addition, this concentration may also be a concentration selected from the range of 400 μM to 1.5 mM. Particularly preferred examples thereof include concentrations selected from the range of 500 μM to 2.5 mM, which are close to the plasma calcium ion concentration in vivo.

In the present description, the low calcium ion concentration is not specifically limited to a single numerical value, and preferably it may be a concentration selected from the range of 0.1 μM to 30 μM. In another aspect, the low calcium ion concentration may be a concentration selected from the range of 0.2 μM to 20 μM. In another aspect, the low calcium ion concentration may also be a concentration selected from the range of 0.5 μM to 10 μM. In an alternative aspect, the low calcium ion concentration may also be a concentration selected from the range of 1 μM to 5 μM. In addition, this concentration may also be a concentration selected from the range of 2 μM to 4 μM. Particularly preferred examples thereof include concentrations selected from the range of 1 μM to 5 μM, which are close to the concentrations of ionized calcium in the early endosome in vivo.

For the purposes of the present invention, the expression "the antigen binding activity of an antigen binding molecule is lower under low calcium ion concentration conditions than under high calcium ion concentration conditions" means that the antigen binding activity of the antigen binding molecule at a calcium ion concentration selected from the range of 0.1 μM to 30 µM is weaker than the calcium ion concentration selected from the range from 100 µM to 10 mM. This expression preferably means that the antigen binding activity of the antigen binding molecule at a calcium ion concentration selected from the range of 0.5 μM to 10 μM is weaker than at a calcium ion concentration selected from the range from 200 μM to 5 mM. Particularly preferably, the expression means that the antigen binding activity at the early endosome calcium ion concentration in vivo is weaker than at the plasma calcium ion concentration in vivo. In particular, this means that the antigen binding activity of the antigen binding molecule at a calcium ion concentration selected from the range of 1 μM to 5 μM is weaker than at a calcium ion concentration selected from the range from 500 μM to 2.5 mM.

Determining whether the antigen-binding activity of an antigen-binding molecule varies depending on metal ion concentration conditions can be performed using an assay method known in the art. For example, the antigen binding activity of an antigen binding molecule is compared between low calcium ion concentration conditions and high calcium ion concentration conditions to confirm that the antigen binding activity of the antigen binding molecule changes to a higher level under high calcium ion concentration conditions than under low calcium ion concentration conditions.

In the context of the present invention, the expression "the antigen binding activity of an antigen binding molecule is lower under conditions of low calcium ion concentration than under conditions of high calcium ion concentration" can be expressed as "the antigen binding activity of an antigen binding molecule is higher under conditions of high calcium ion concentration than under conditions low concentration of calcium ions." In the context of the present invention, the expression "antigen binding activity is weaker under low calcium ion concentration conditions than under high calcium ion concentration conditions" is also described as "antigen binding activity is weaker under low calcium ion concentration conditions than under high calcium ion concentration conditions." In addition, the expression “antigen binding activity under low calcium ion concentration conditions is reduced relative to antigen binding activity under high calcium ion concentration conditions” is also described as “antigen binding activity under low calcium ion concentration conditions is weakened relative to antigen binding activity under high calcium ion concentration conditions.”

Conditions other than calcium ion concentration for the antigen binding activity assay may be appropriately selected by those skilled in the art without particular limitation. Antigen binding activity can be analyzed under conditions such as HEPES buffer and 37°C. Antigen binding activity can also be assayed using, for example, Biacore (GE Healthcare Bio-Sciences Corp.). In an antigen binding activity assay, an antigen binding molecule can be assessed for its binding ability against, for example, soluble antigens by injecting the antigens as an analyte onto a chip on which the antigen binding molecule is immobilized. Alternatively, an antigen-binding molecule can be assessed for its binding ability against, for example, membrane antigens, by injecting the molecule as an analyte onto an antigen-immobilized chip.

In the antigen binding molecule of the present invention, the ratio between the antigen binding activity under low calcium ion concentration conditions and the antigen binding activity under high calcium ion concentration conditions is not particularly limited, as long as the antigen binding activity under low calcium ion concentration conditions is weaker than that under high calcium ion concentration conditions calcium ions. The ratio of the dissociation constant KD under conditions of low calcium ion concentration to KD under conditions of high calcium ion concentration (KD (3 μM Ca)/KD (2 mM Ca)) for antigens is preferably 2 or higher, more preferably 10 or higher, even more preferably 40 or higher. The upper limit of the KD (3 μM Ca)/KD (2 mM Ca) ratio is not particularly limited, and it may be any value including 400, 1000, 10000, etc. provided that the resulting antigen-binding molecule can be technically prepared by those skilled in the art. Alternatively, the ratio may be determined by the value KD (3 μM Ca)/KD (1.2 mM Ca). In particular, the ratio KD (3 μM Ca)/KD (1.2 mM Ca) is 2 or higher, more preferably 10 or higher, even more preferably 40 or higher. The upper limit of the ratio KD (3 μM Ca)/KD (1.2 mM Ca) is not particularly limited, and may be a value including 400, 1000, 10000, etc. provided that the resulting antigen-binding molecule can be technically prepared by those skilled in the art.

The dissociation constant KD can be used as a measure of antigen binding activity for soluble antigens. Alternatively, for membrane antigens, the apparent KD (apparent dissociation constant) can be used. KD (dissociation constant) and apparent KD (apparent dissociation constant) can be measured in a manner well known to those skilled in the art, for example, using a Biacore (GE Healthcare Bio-Sciences Corp.), a Scatchard plot, or a flow cytometer.

Alternatively, it is preferable to use, for example, the dissociation rate constant kd as another indicator that indicates the relationship between the antigen binding activity under low calcium concentration conditions and the antigen binding activity under high calcium concentration conditions for the antigen binding molecule of the present invention. In the case of using kd (dissociation rate constant) instead of KD (dissociation rate constant) as the ratio of binding activities, the ratio of kd (dissociation rate constant) under low calcium ion concentration conditions to kd (dissociation rate constant) under high calcium concentration conditions (kd (in conditions of low calcium ion concentration)/kd (under conditions of high calcium ion concentration)) for antigens is preferably 2 or higher, more preferably 5 or higher, even more preferably 10 or higher, even more preferably 30 or higher. The upper limit of the ratio kd (under low calcium ion concentration conditions)/kd (under high calcium ion concentration conditions) is not particularly limited, and it may be any value including 50, 100, 200, etc., as long as the resulting antigen-binding the molecule can be prepared in accordance with the technical common sense of those skilled in the art.

The dissociation rate constant kd can be used as a measure of antigen binding activity for soluble antigens. Alternatively, for membrane antigens, the apparent kd (apparent dissociation rate constant) can be used. kd (dissociation rate constant) and apparent kd (apparent dissociation rate constant) can be measured in a manner well known to those skilled in the art, for example, using a Biacore (GE Healthcare Bio-Sciences Corp.) or a flow cytometer. In the context of the present invention, the antigen-binding activity of the antigen-binding molecule at various calcium ion concentrations is preferably analyzed, provided that the other calcium concentration is kept constant.

Within the scope of the present invention, the concentration conditions for the proton, i.e. the atomic nucleus of hydrogen, is used synonymously with the terms of the hydrogen index (pH). When the active mass of hydrogen ions in an aqueous solution is given as a _H +, the pH is defined as -log10a _H +. In an aqueous solution having low ionic strength (for example, below 10 ^-3 ), a _H + is practically equal to the ionic strength of hydrogen. For example, the ionic product of water at 25°C and a pressure of 1 atmosphere is Kw=a _H +a _OH =10 ^-14 and, thus, a _H +=a _OH =10 ^-7 for pure water. In this case, pH 7 corresponds to a neutral aqueous solution; A pH less than 7 corresponds to an acidic aqueous solution; and a pH greater than 7 corresponds to an alkaline aqueous solution.

When using pH conditions as hydrogen ion concentration conditions according to the present invention, examples of pH conditions include acidic pH conditions and neutral pH conditions. The expression "binding activity changes depending on pH conditions" means that the antigen binding activity of each antigen binding molecule changes depending on the differences between acidic pH conditions and neutral pH conditions. Examples of this include higher antigen binding activity of an antigen binding molecule under neutral pH conditions than under acidic pH conditions. Another example of this involves higher antigen binding activity of an antigen binding molecule under acidic pH conditions than under neutral pH conditions.

In the present description, the neutral pH values are not particularly limited to a single numerical value, and preferably can be selected from the range of pH 6.7 to pH 10.0. In another aspect, neutral pH values can be selected from the range of pH 6.7 to pH 9.5. In another aspect, neutral pH values can be selected from the range of pH 7.0 to pH 9.0. In an alternative aspect, this pH value may be selected from the range of pH 7.0 to pH 8.0. Particularly preferred examples thereof include a pH value of 7.4, which is close to the pH value of plasma (blood) in vivo.

In the present description, the acidic pH value is not particularly limited to a single numerical value, and preferably can be selected from the range of pH 4.0 to pH 6.5. In another aspect, the acidic pH value may be selected from a range of pH 4.5 to pH 6.5. In another aspect of the acid, the pH value may be selected from a range of pH 5.0 to pH 6.5. In an alternative aspect, this pH value may be selected from the range of pH 5.5 to pH 6.5. Particularly preferred examples thereof include a pH value of 5.8, which is close to the pH value of the early endosome in vivo.

For the purposes of the present invention, the expression "the antigen binding activity of an antigen binding molecule is lower under acidic pH conditions than under neutral pH conditions" means that the antigen binding activity of the antigen binding molecule at a pH value selected from the range of pH 4.0 to pH 6. 5 is weaker than at a pH value selected from the range of pH 6.7 to pH 10.0. This expression preferably means that the antigen binding activity of the antigen binding molecule at a pH value selected from the range of pH 4.5 to pH 6.5 is weaker than at a pH value selected from the range of pH 6.7 to pH 9.5 . More preferably, the expression means that the antigen binding activity of the antigen binding molecule at a pH value selected from the range of pH 5.0 to pH 6.5 is weaker than at a pH value selected from the range of pH 7.0 to pH 9, 0. In addition, the expression preferably means that the antigen binding activity of the antigen binding molecule at a pH value selected from the range of pH 5.5 to pH 6.5 is weaker than at a pH selected from the range of pH 7.0 to pH 8, 0. Particularly preferably, this expression means that the antigen binding activity at the pH value in the early endosome in vivo is weaker than at the pH value in the plasma in vivo. In particular, this means that the antigen binding activity of the antigen binding molecule at pH 5.8 is weaker than at pH 7.4.

Determining whether the antigen binding activity of an antigen binding molecule changes depending on pH conditions can be determined, for example, in accordance with the binding activity assay method under different pH conditions described above. For example, the antigen binding activity of an antigen binding molecule is compared between acidic pH conditions and neutral pH conditions to confirm that the antigen binding activity of the antigen binding molecule changes to a higher level under neutral pH conditions than under acidic pH conditions.

In the context of the present invention, the expression "the antigen-binding activity of an antigen-binding molecule is lower under acidic pH conditions than under neutral pH conditions" may be expressed as "the antigen-binding activity of an antigen-binding molecule is higher under neutral pH conditions than under acidic pH conditions." pH". Within the scope of the present invention, the expression "antigen binding activity is lower under acidic pH conditions than under neutral pH conditions" is also described as "antigen binding activity is weaker under acidic pH conditions than under neutral pH conditions." In addition, the expression "antigen-binding activity under acidic pH conditions is reduced relative to antigen-binding activity under neutral pH conditions" is also described as "antigen-binding activity under acidic pH conditions is weakened relative to antigen-binding activity under neutral pH conditions."

When the metal ion is, for example, a calcium ion, the type of amino acid residue that changes the antigen-binding activity of the antigen-binding molecule depending on the ion concentration conditions as described above is not limited, as long as the amino acid forms a calcium-binding motif. The calcium binding motif is well known to those skilled in the art and has been described in detail (eg, Springer et al., Cell (2000) 102, 275-277; Kawasaki and Kretsinger, Protein Prof. (1995) 2, 305-490; Moncrief et al. ., J. Mol. Evol. (1990) 30, 522-562; Chauvaux et al., Biochem. J. (1990) 265, 261-265; Bairoch and Cox, FEBS Lett. (1990) 269, 454-456 ; Davis, New Biol. (1990) 2, 410-419; Schaefer et al., Genomics (1995) 25, 638-643; Economou et al., EMBO J. (1990) 9, 349-354; and Wurzburg et al., Structure (2006) 14, 6, 1049-1058). In particular, the antigen binding molecule of the present invention may contain any calcium binding motif known in the art, including C-type lectins such as ASGPR, CD23, MBR and DC-SIGN. Other preferred examples of such a calcium-binding motif may include a calcium-binding motif contained in the Vk5 domain of an antibody light chain variable region having a germline sequence such as Vk5-2, as described below.

As an alternative example, an amino acid having a metal chelating effect can preferably be used as an amino acid residue that changes the antigen-binding activity of an antigen-binding molecule depending on calcium ion concentration conditions. Preferable examples of the amino acid having a metal chelating effect include serine (Ser (S)), threonine (Thr (T)), asparagine (Asn (N)), glutamine (Gln (Q)), aspartic acid (Asp (D)) and glutamic acid (Glu(E)).

The position of the amino acid residue contained in the antigen binding domain is not limited to a particular position and may be any position in the variable region of the heavy and light chains constituting the antigen binding domain, provided that the resulting amino acid residue changes the antigen binding activity of the antigen binding molecule depending on the calcium ion concentration conditions . In one aspect, the present invention provides a library consisting essentially of antigen binding molecules differing from each other in sequence, wherein an amino acid residue that modifies the antigen binding activity of the antigen binding molecule depending on calcium ion concentration conditions is contained in the heavy chain antigen binding domain. In another aspect, the present invention relates to a library consisting essentially of antigen binding molecules that differ from each other in sequence, wherein the amino acid residue is contained in CDR3 of the heavy chain. In an alternative aspect, the present invention provides a library consisting essentially of antigen binding molecules differing from each other in sequence, wherein the amino acid residue is located at any one or more of positions 95, 96, 100a, and 101, as defined by Kabat numbering, in CDR3 heavy chain.

In one aspect, the present invention relates to a library consisting essentially of antigen binding molecules differing from each other in sequence, wherein an amino acid residue that modifies the antigen binding activity of the antigen binding molecule depending on calcium ion concentration conditions is located in the light chain antigen binding domain. In another aspect, the present invention relates to a library consisting essentially of antigen binding molecules that differ from each other in sequence, where the amino acid residue is located in CDR1 of the light chain. In an alternative aspect, the present invention relates to a library consisting essentially of antigen binding molecules differing from each other in sequence, wherein the amino acid residue is located at any one or more of positions 30, 31 and 32, as defined by Kabat numbering, in the CDR1 of the light chain .

In an alternative aspect, the present invention provides a library consisting essentially of antigen binding molecules that differ from each other in sequence, where the amino acid residue is located in CDR2 of the light chain. In another alternative aspect, the present invention provides a library consisting essentially of antigen binding molecules differing from each other in sequence, wherein the amino acid residue is located at position 50, determined by Kabat numbering, in CDR2 of the light chain.

In an alternative aspect, the present invention relates to a library consisting essentially of antigen binding molecules that differ from each other in sequence, where the amino acid residue is located in CDR3 of the light chain. In another alternative aspect, the present invention provides a library consisting essentially of antigen binding molecules differing from each other in sequence, wherein the amino acid residue is located at position 92, defined by Kabat numbering, in CDR3 of the light chain.

In a different aspect, the present invention also provides a library consisting essentially of antigen binding molecules differing from each other in sequence, wherein the amino acid residue is located in two or three CDRs selected from the light chain CDR1, CDR2 and CDR3 described above. Furthermore, the present invention relates to a library consisting essentially of antigen binding molecules differing from each other in sequence, wherein the amino acid residue is located at any one or more of positions 30, 31, 32, 50 and 92, defined by Kabat numbering, in light chain.

In a particularly preferred embodiment, it is desirable that the framework sequence in the light chain and/or heavy chain variable regions of each antigen binding molecule has a human germline framework sequence. Thus, in one aspect of the present invention, an antigen binding molecule of the present invention having framework sequences, all of which are fully human sequences, is likely to elicit little or no immunogenic response when administered to a human (eg, to treat a disease). In this context, the expression “comprising a germline sequence” in accordance with the present invention means that at least a portion of the framework sequence of the present invention is identical to a portion of a human germline sequence. For example, the heavy chain FR2 sequence in the antigen binding molecule of the present invention may be a sequence consisting of a combination of heavy chain FR2 sequences from a variety of different human germline framework sequences. Such an antigen binding molecule is also included in the “germline sequence containing” antigen binding molecule of the present invention.

Preferred examples of the framework region include sequences of currently known fully human framework regions listed on a website such as V-Base (http://vbase.mrc-cpe.cam.ac.uk/). Any of the sequences of these framework regions can suitably be used as the germline sequence contained in the antigen binding molecule of the present invention. Germline sequences can be classified based on their analogy (Tomlinson et al., J. Mol. Biol. (1992) 227, 776-798; Williams and Winter, Eur. J. Immunol. (1993) 23, 1456-1461; and Cox et al., Nat. Genetics (1994) 7, 162-168). The preferred germline sequence can be suitably selected from Vκ classified into 7 subgroups, Vλ classified into 10 subgroups, and VH classified into 7 subgroups.

Preferred examples of fully human VH sequences include, but are not limited to, VH1 subgroup VH sequences (eg, VH1-2, VH1-3, VH1-8, VH1-18, VH1-24, VH1-45, VH1-46, VH1- 58 and VH1-69), VH2 subgroups (for example, VH2-5, VH2-26 and VH2-70), VH3 subgroups (VH3-7, VH3-9, VH3-11, VH3-13, VH3-15, VH3- 16, VH3-20, VH3-21, VH3-23, VH3-30, VH3-33, VH3-35, VH3-38, VH3-43, VH3-48, VH3-49, VH3-53, VH3-64, VH3-66, VH3-72, VH3-73 and VH3-74), subgroups VH4 (VH4-4, VH4-28, VH4-31, VH4-34, VH4-39, VH4-59 and VH4-61), subgroups VH5 (VH5-51), VH6 subgroups (VH6-1) and VH7 subgroups (VH7-4 and VH7-81). These sequences are also described in the public literature (Matsuda et al., J. Exp. Med. (1998) 188, 1973-1975), etc. Those skilled in the art can appropriately design the antigen binding molecule of the present invention based on the sequence information. Preferably, any other fully human framework region or subregion of the framework region may be used.

Preferred examples of fully human VK sequences include, but are not limited to, A20, A30, L1, L4, L5, L8, L9, L11, L12, L14, L15, L18, L19, L22, L23, L24, O2, O4, O8 , O12, O14 and O18, belonging to the subgroup Vk1, A1, A2, A3, A5, A7, A17, A18, A19, A23, O1 and O11, belonging to the subgroup Vk2, A11, A27, L2, L6, L10, L16, L20 and L25 belonging to the Vk3 subgroup, B3 belonging to the Vk4 subgroup, B2 belonging to the Vk5 subgroup (also referred to herein as Vk5-2), and A10, A14 and A2b belonging to the VK6 subgroup (Kawasaki et al., Eur. J. Immunol. (2001) 31, 1017-1028; Schable and Zachau, Biol. Chem. Hoppe Seyler (1993) 374, 1001-1022; and Brensing-Kuppers et al., Gene (1997) 191, 173-181).

Preferred examples of fully human VL sequences include, but are not limited to, Vl-2, Vl-3, Vl-4, Vl-5, Vl-7, Vl-9, Vl-11, Vl-13, Vl-16, Vl -17, V1-18, Vl-19, V1-20 and V1-22, belonging to the subgroup VL1, V2-1, V2-6, V2-7, V2-8, V2-11, V2-13, V2-14 , V2-15, V2-17 and V2-19, belonging to the subgroup VL2, V3-2, V3-3 and V3-4, belonging to the subgroup VL3, V4-1, V4-2, V4-3, V4-4 and V4 -6, belonging to the VL4 subgroup, and V5-1, V5-2, V5-4 and V5-6, belonging to the VL5 subgroup (Kawasaki et al., Genome Res. (1997) 7, 250-261).

These framework sequences differ from each other by one or more amino acid residues. These framework sequences can be used in conjunction with "at least one amino acid residue that modifies the antigen-binding activity of the antigen-binding molecule depending on the ion concentration conditions" of the present invention. Examples of fully human framework regions used in conjunction with "at least one amino acid residue that modifies the antigen binding activity of the antigen binding molecule depending on ion concentration conditions" of the present invention include, but are not limited to, KOL, NEWM, REI, EU, TUR , TEI, LAY and ROM (eg, Kabat et al., (1991); and Wu et al., J. Exp. Med. (1970) 132, 211-250).

Although the present invention is not bound by any particular theory, it is expected that the use of a germline sequence will prevent an adverse immune response in virtually all individuals, likely due in part to the following: somatic mutations are common in immunoglobulin variable regions as a result of the affinity maturation step that exists during a normal immune response. These mutations occur primarily near CDRs that have hypervariable sequence, but also affect residues in framework regions. These framework mutations are not present in germline genes and are not unlikely to be immunogenic in patients. On the other hand, normal human populations are exposed to a large number of framework sequences expressed by germline genes. As a result of immunological tolerance, these germline framework regions are assumed to be low or non-immunogenic to patients. To maximize the potential for immunological tolerance, the germline sequence can be selected from functional germline gene clusters where variable region encoding genes are typically present.

The antigen-binding molecule containing "at least one amino acid residue that changes the antigen-binding activity of the antigen-binding molecule depending on the ion concentration conditions" of the present invention in the framework sequence can be prepared by appropriately selecting a method known in the art, such as site-directed mutagenesis (Kunkel et al., Proc. Natl. Acad. Sci. USA (1985) 82, 488-492) or overlap PCR.

For example, light chain variable regions selected as framework sequences precontaining "at least one amino acid residue that modifies the antigen binding activity of the antigen binding molecule depending on ion concentration conditions" can be combined with heavy chain variable regions prepared as a randomized library variable region sequences, obtaining a library of the present invention containing a plurality of antigen-binding molecules differing from each other in sequence. When the ion concentration is a calcium ion concentration, non-limiting examples of such a library preferably include a library containing in combination the light chain variable region sequence shown in SEQ ID NO: 1 (Vk5-2) and the heavy chain variable regions obtained as a randomized library variable region sequences. As a preferred example, the light chain variable region sequence described in SEQ ID NO: 1 (Vk5-2), as well as Vk5-2 variant 1 represented by SEQ ID NO: 2, or Vk5-2 variant 2 represented by SEQ ID NO : 3 described below are suitably used as a light chain variable region sequence containing a Vk5 domain present in the light chain variable region of an antibody having a germline sequence such as Vk5-2. The Vk5-2 domain containing "at least one amino acid residue that modifies the antigen-binding activity of an antigen-binding molecule depending on ion concentration conditions" found in these molecules can be used as the calcium-binding motif-containing domain of the present invention. In a non-limiting aspect of the present invention, at least one amino acid that forms a calcium-binding motif can be used as "at least one amino acid residue that modifies the antigen-binding activity of an antigen-binding molecule depending on ion concentration conditions" of the present invention.

Also, light chain variable region sequences selected as framework sequences pre-containing “at least one amino acid residue that modifies the antigen-binding activity of the antigen-binding molecule depending on ion concentration conditions” can be designed to contain a variety of amino acids as residues different from the specified amino acid residue. For the purposes of the present invention, such residues are referred to as non-strict residues. The number and positions of non-strict residues are not limited to any particular aspect, provided that the antigen binding activity of the antigen binding molecule of the present invention varies depending on the ion concentration conditions. In particular, each of the heavy chain and/or light chain CDR and/or FR sequences may contain one or more non-stringent residues. When the ion concentration is, for example, calcium ion concentration, non-limiting examples of non-strict residues included in the light chain variable region sequence set forth in SEQ ID NO: 1 (Vk5-2) include the amino acid residues described in Tables 13, 14, 17 and 18.

Alternatively, light chain variable regions containing "at least one amino acid residue that modifies the antigen binding activity of the antigen binding molecule depending on ion concentration conditions" can be combined with heavy chain variable regions as a randomized variable region library to produce the library of the present invention , containing many antigen-binding molecules that differ from each other in sequence. When the calcium ion concentration is a calcium ion concentration, non-limiting examples of such a library preferably include a library containing in combination light chain variable region sequences derived from the germline sequence SEQ ID NO: 6 (Vk1), SEQ ID NO: 7 (Vk2), SEQ ID NO: 8 (Vk3), SEQ ID NO: 9 (Vk4), or the like. by replacing a specific residue with "at least one amino acid residue that alters the antigen binding activity of the antigen binding molecule depending on the calcium ion concentration conditions", and the heavy chain variable regions obtained as a randomized library of variable region sequences. Non-limiting examples of the specified amino acid residue include the amino acid residues contained in the light chain CDR1. Other non-limiting examples of amino acid residue include amino acid residues contained in the light chain CDR2. The following non-limiting examples of the specified amino acid residue include the amino acid residues contained in the light chain CDR3.

Non-limiting examples of the amino acid residue contained in the light chain CDR1 as described above include amino acid residues 30, 31 and/or 32, as defined by Kabat numbering, in the CDR1 light chain variable region. Non-limiting examples of the amino acid residue contained in the light chain CDR2 include amino acid residue 50, defined by Kabat numbering, in the light chain variable region of CDR2. Non-limiting examples of the amino acid residue contained in the light chain CDR3 include amino acid residue 92, defined by Kabat numbering, in the light chain variable region of CDR3. These amino acid residues may be present alone or in combination of two or more of these amino acids, provided that the amino acid residues form calcium-binding motifs and/or the antigen-binding activity of the antigen-binding molecule varies depending on calcium ion concentration conditions. For example, troponin C, calmodulin, parvalbumin, and myosin light chain are known to have multiple calcium ion binding sites and are likely to derive from a common source in terms of molecular evolution. Light chain CDR1, CDR2 and/or CDR3 can be designed to contain any of their binding motifs. For example, the cadherin domain, the EF hand contained in calmodulin, the C2 domain contained in protein kinase C, the Gla domain contained in coagulation factor IX, the C-type lectin contained in the asialoglycoprotein or mannose-binding receptor can be suitably used for these purposes. receptor, domain A contained in the LDL receptor, annexin, thrombospondin type 3 domain and EGF-like domain.

In the case of combining light chain variable regions containing "at least one amino acid residue that modifies the antigen-binding activity of the antigen-binding molecule depending on ion concentration conditions" with heavy chain variable regions obtained as a randomized library of variable region sequences, the variable region sequence light chains can be designed to contain loose residues, as in the case described above. The number and positions of non-strict residues are not limited to any particular aspect, provided that the antigen binding activity of the antigen binding molecule of the present invention varies depending on the ion concentration conditions. In particular, each of the heavy chain and/or light chain CDR and/or FR sequences may contain one or more non-stringent residues. When the concentration is, for example, the concentration of calcium ions, non-limiting examples of non-strict residues introduced into the light chain variable region sequences include the amino acid residues described in Tables 13, 14, 17 and 18.

Preferred examples of heavy chain variable regions for combination therewith include a randomized library of variable regions. A randomized library of variable regions is obtained by appropriately combining methods known in the art. In a non-limiting aspect of the present invention, the randomized variable region library may preferably be an immune library constructed from antibody genes derived from lymphocytes of animals immunized with specific antigens, patients with an infectious disease, people having increased antibody titers in the blood as a result of vaccination, patients with a malignant tumor or autoimmune diseases.

In a non-limiting aspect of the present invention, also preferably, a synthetic library obtained by replacing the CDR sequences of V genes of genomic DNA or reconstructed functional V genes with a synthetic set of oligonucleotides containing sequences encoding a set of codons having an appropriate length can be used as a randomized library of variable regions. In this case, only CDR3 sequences can be replaced with them, since diversity is observed in the heavy chain CDR3 gene sequence. The amino acid diversity that provides the variable regions of antigen binding molecules is based on the diversity imparted to the amino acid residues at positions exposed on the surface of the antigen binding molecules. Surface exposed positions refer to positions that are believed to be surface exposed and/or accessible to antigens based on the structures, set of structures, and/or modeled structures of antigen binding molecules, and are typically located in their CDRs. Preferably, the positions exposed on the surface are determined using a computer program, such as the Insight II program (Accelrys Inc.), and coordinates from three-dimensional models of antigen-binding molecules. The positions exposed on the surface can be determined using an algorithm known in the art (e.g., Lee and Richards, J. Mol. Biol. (1971) 55, 379-400; and Connolly, J. Appl. Cryst. (1983) 16, 54 8-558). Positions exposed on the surface can be determined using software suitable for protein modeling and 3D structure information obtained for antibodies. Preferred examples of software that can be used for such a purpose include SYBYL biopolymer module software (Tripos Associates Inc.). Typically and preferably, the "size" of the probes used for the calculation is set to approximately 1.4 angstroms or lower in radius values when the algorithm requires the user to enter size parameters. A method for determining surface exposed areas and areas using personal computer software is described in Pacios, Comput. Chem. (1994) 18 (4), 377-386, and J. Mol. Model. (1995) 1, 46-53.

In another non-limiting aspect of the present invention, a naive library consisting of naive sequences, which are unbiased sets of antibody sequences constructed from antibody genes derived from lymphocytes of healthy humans, is used as a randomized variable region library (Gejima et al., Human Antibodies (2002). ) 11, 121-129; and Cardoso et al., Scand. J. Immunol. (2000) 51, 337-344). An amino acid sequence containing a naive sequence, as described within the scope of the present invention, refers to an amino acid sequence obtained from such a naive library.

In one aspect of the present invention, heavy chain variable regions selected as framework sequences precontaining "at least one amino acid residue that modifies the antigen binding activity of the antigen binding molecule depending on ion concentration conditions" can be combined with light chain variable regions obtained in as a randomized library of variable region sequences to obtain a library of the present invention containing a plurality of antigen binding molecules differing from each other in sequence. When the ion concentration is a calcium ion concentration, non-limiting examples of such a library preferably include a library containing in combination the heavy chain variable region sequence described in SEQ ID NO: 10 (6RL#9H-IgG1) or SEQ ID NO: 11 (6KC4-1 #85H-IgG1), and light chain variable regions obtained as a randomized variable region sequence library. Alternatively, the library can be prepared using suitably selected light chain variable regions having a germline sequence instead of light chain variable regions generated as a randomized variable region sequence library. Preferred examples thereof include a library containing in combination the heavy chain variable region sequence shown in SEQ ID NO: 10 (6RL#9H-IgG1) or SEQ ID NO: 11 (6KC4-1#85H-IgG1) and light chain variable regions , having a germline sequence.

Also, heavy chain variable region sequences selected as framework sequences pre-containing “at least one amino acid residue that modifies the antigen-binding activity of the antigen-binding molecule depending on ion concentration conditions” can be designed to contain non-strict residues. The number and positions of the non-strict residues are not limited to any particular aspect, provided that the antigen binding activity of the antigen binding molecule of the present invention varies depending on the calcium ion concentration conditions. In particular, each of the heavy chain and/or light chain CDR and/or FR sequences may contain one or more non-stringent residues. When the ion concentration is, for example, calcium ion concentration, non-limiting examples of non-strict residues included in the heavy chain variable region sequence described in SEQ ID NO: 10 (6RL#9H-IgG1) include all amino acid residues of the heavy chain CDR1 and CDR2, as well as heavy chain CDR3 amino acid residues excluding positions 95, 96 and/or 100a. Alternatively, non-limiting examples of non-strict residues included in the heavy chain variable region sequence set forth in SEQ ID NO: 11 (6KC4-l#85H-IgG1) include all of the heavy chain CDR1 and CDR2 amino acid residues, as well as the heavy chain CDR3 amino acid residues, with the exception of provisions 95 and/or 101.

Alternatively, heavy chain variable regions containing "at least one amino acid residue that modifies the antigen-binding activity of the antigen-binding molecule depending on ion concentration conditions" can be combined with light chain variable regions obtained as a randomized library of variable region sequences, or variable light chain regions having a germline sequence to obtain a library of the present invention containing a plurality of antigen binding molecules differing from each other in sequence. When the ion concentration is a calcium ion concentration, non-limiting examples of such a library preferably include a library containing in combination heavy chain variable region sequences formed from heavy chain variable regions by replacing a particular residue with at least one amino acid residue that alters the antigen binding activity of the antigen binding molecule depending on the calcium ion concentration conditions", and light chain variable regions obtained as a randomized variable region sequence library, or light chain variable regions having a germline sequence. Non-limiting examples of the specified amino acid residue include amino acid residues contained in the CDR1 of the heavy chain. Other non-limiting examples of the specified amino acid residue include amino acid residues contained in the CDR2 of the heavy chain. The following non-limiting examples of the specified amino acid residue include amino acid residues contained in the CDR3 of the heavy chain. Non-limiting examples of the amino acid residue contained in the heavy chain CDR3 include amino acids at positions 95, 96, 100a and/or 101, as defined by Kabat numbering, in the heavy chain variable region of CDR3. These amino acid residues may be present alone or in combination of two or more of these amino acids, provided that the amino acid residues form calcium-binding motifs and/or the antigen-binding activity of the antigen-binding molecule varies depending on calcium ion concentration conditions.

In the case of combining heavy chain variable regions containing "at least one amino acid residue that modifies the antigen binding activity of the antigen binding molecule depending on ion concentration conditions" with light chain variable regions obtained as a random library of variable region sequences, or variable light chain regions having a germline sequence, the sequences of the heavy chain variable regions can be designed to contain loose residues, as in the case described above. The number and positions of non-strict residues are not limited to any particular aspect, provided that the antigen binding activity of the antigen binding molecule of the present invention varies depending on the ion concentration conditions. In particular, each of the heavy chain CDR and/or FR sequences may contain one or more non-stringent residues. Also, as the amino acid sequences of the heavy chain variable region CDR1, CDR2 and/or CDR3, a randomized library of variable regions can be used, preferably excluding "an amino acid residue that changes the antigen-binding activity of the antigen-binding molecule depending on the ion concentration conditions." When germline sequences are used as light chain variable regions, non-limiting examples thereof may include the germline sequences SEQ ID NO: 6 (Vk1), SEQ ID NO: 7 (Vk2), SEQ ID NO: 8 (Vk3) and SEQ ID NO: 9 (Vk4).

As the amino acid residue that changes the antigen-binding activity of the antigen-binding molecule depending on the calcium ion concentration conditions, any amino acid residue can preferably be used, provided that the amino acid residue forms a calcium-binding motif. Specific examples of such an amino acid residue include amino acids having electron-donating properties. Preferred examples of amino acids having electron-donating properties include serine, threonine, asparagine, glutamine, aspartic acid and glutamic acid.

In one aspect of the present invention, light chain variable regions containing "at least one amino acid residue that modifies the antigen binding activity of the antigen binding molecule depending on hydrogen ion concentration conditions" can also be combined with heavy chain variable regions prepared as a randomized library variable region sequences to obtain a library of the present invention containing a plurality of antigen binding molecules differing from each other in sequence.

Non-limiting examples of the specified amino acid residue include the amino acid residues contained in the light chain CDR1. Other non-limiting examples of the specified amino acid residue include the amino acid residues contained in the light chain CDR2. The following non-limiting examples of the specified amino acid residue include the amino acid residues contained in the light chain CDR3.

Non-limiting examples of the amino acid residue contained in the light chain CDR1 as described above include amino acid residues 24, 27, 28, 30, 31, 32 and/or 34, as defined by Kabat numbering, in the CDR1 light chain variable region. Non-limiting examples of the amino acid residue contained in the light chain CDR2 include amino acid residues 50, 51, 52, 53, 54, 55 and/or 56, as defined by Kabat numbering, in the CDR2 light chain variable region. Non-limiting examples of the amino acid residue contained in the light chain CDR3 include amino acid residues 89, 90, 91, 92, 93, 94 and/or 95a, as defined by Kabat numbering, in the CDR3 light chain variable region. These amino acid residues may be present alone or in combination of two or more of these amino acids, provided that the antigen binding activity of the antigen binding molecule containing the amino acid residue(s) varies depending on hydrogen ion concentration conditions.

In the case of combining light chain variable regions containing "at least one amino acid residue that modifies the antigen-binding activity of the antigen-binding molecule depending on hydrogen ion concentration conditions" with heavy chain variable regions obtained as a randomized library of variable region sequences, the sequence The light chain variable regions can be designed to contain non-strict residues, as in the case described above. The number and positions of non-strict residues are not limited to any particular aspect, provided that the antigen binding activity of the antigen binding molecule of the present invention varies depending on the ion concentration conditions. In particular, each of the heavy chain and/or light chain CDR and/or FR sequences may contain one or more non-stringent residues. Non-limiting examples of non-strict residues introduced into the light chain variable region sequences include the amino acid residues described in Tables 4 and 5. As a non-limiting example, the Vk1 germline sequence (SEQ ID NO: 6) may be preferably used as the light chain variable region amino acid sequences ), Vk2 (SEQ ID NO: 7), Vk3 (SEQ ID NO: 8), Vk4 (SEQ ID NO: 9), etc., except for an amino acid residue that changes the antigen-binding activity of the antigen-binding molecule depending on the concentration conditions hydrogen ions, or weak residues.

Preferred examples of heavy chain variable regions to be combined therewith include a randomized library of variable regions. A randomized library of variable regions is obtained by appropriately combining methods known in the art. In a non-limiting aspect of the present invention, the randomized variable region library may preferably be an immune library constructed from antibody genes derived from lymphocytes of animals immunized with specific antigens, patients with an infectious disease, people having increased antibody titers in the blood as a result of vaccination, patients with a malignant tumor or autoimmune diseases.

In a non-limiting aspect of the present invention, also preferably, a synthetic library obtained by replacing the CDR sequences of V genes of genomic DNA or reconstructed functional V genes with a synthetic set of oligonucleotides containing sequences encoding a set of codons having the appropriate length as and in the case described above.

In another non-limiting aspect of the present invention, a naive library consisting of naive sequences, which are unbiased sets of antibody sequences constructed from antibody genes derived from lymphocytes of healthy humans, is used as a randomized variable region library (Gejima et al., Human Antibodies (2002) ) 11, 121-129; and Cardoso et al., Scand. J. Immunol. (2000) 51, 337-344).

In an alternative aspect of the present invention, heavy chain variable regions containing "at least one amino acid residue that modifies the antigen binding activity of the antigen binding molecule depending on hydrogen ion concentration conditions" can be combined with light chain variable regions obtained as a randomized sequence library variable regions, or light chain variable regions having a germline sequence, to obtain a library of the present invention containing a plurality of antigen binding molecules differing from each other in sequence. Non-limiting examples of such a library preferably include a library containing in combination heavy chain variable region sequences formed from heavy chain variable regions by replacing a particular residue with "at least one amino acid residue that alters the antigen binding activity of the antigen binding molecule depending on hydrogen ion concentration conditions" and light chain variable regions obtained as a randomized variable region sequence library, or light chain variable regions having a germline sequence. Non-limiting examples of the specified amino acid residue include amino acid residues contained in the CDR1 of the heavy chain. Other non-limiting examples of the specified amino acid residue include amino acid residues contained in the CDR2 of the heavy chain. The following non-limiting examples of amino acid residue include amino acid residues contained in the heavy chain CDR3.

Non-limiting examples of the amino acid residue contained in the heavy chain CDR1 include amino acid residues 27, 31, 32, 33 and/or 35, as defined by Kabat numbering, in the CDR1 heavy chain variable region. Non-limiting examples of the amino acid residue contained in the heavy chain CDR2 include amino acid residues 50, 52, 53, 55, 57, 58, 59, 61 and/or 62, as defined by Kabat numbering, in the CDR2 heavy chain variable region. Non-limiting examples of the amino acid residue contained in the heavy chain CDR3 include amino acid residues 95, 96, 97, 98, 99, 100a, 100b, 100d, 100f, 100h, 102 and/or 107, as defined by Kabat numbering, in the heavy chain variable region CDR3 chains. These amino acid residues may be present alone or in combination of two or more of these amino acids, provided that the antigen binding activity of the antigen binding molecule containing the amino acid residue(s) varies depending on hydrogen ion concentration conditions.

In the case of combining heavy chain variable regions containing "at least one amino acid residue that modifies the antigen binding activity of the antigen binding molecule depending on hydrogen ion concentration conditions" with light chain variable regions obtained as a randomized library of variable region sequences, or With light chain variable regions having a germline sequence, the heavy chain variable region sequences can be engineered to contain loose residues as in the case described above. The number and positions of non-strict residues are not limited to any particular aspect, provided that the antigen binding activity of the antigen binding molecule of the present invention varies depending on the hydrogen ion concentration conditions. In particular, each of the heavy chain CDR and/or FR sequences may contain one or more non-stringent residues. Also, the randomized variable region library may preferably be used as the amino acid sequences of the heavy chain variable region CDR1, CDR2 and/or CDR3, excluding "an amino acid residue that changes the antigen binding activity of the antigen binding molecule depending on hydrogen ion concentration conditions." When germline sequences are used as light chain variable regions, non-limiting examples thereof may include the germline sequences SEQ ID NO: 6 (Vk1), SEQ ID NO: 7 (Vk2), SEQ ID NO: 8 (Vk3) and SEQ ID NO: 9 (Vk4).

As the amino acid residue that changes the antigen-binding activity of the antigen-binding molecule depending on the hydrogen ion concentration conditions, any amino acid residue can preferably be used. Specific examples of such an amino acid residue include amino acids having a side chain pKa of 4.0 to 8.0. Preferred examples of amino acids having such electron-donating properties include natural amino acids such as histidine and glutamic acid, as well as non-natural amino acids such as histidine analogues (US 20090035836), m-NO2-Tyr (pKa: 7.45), 3.5 -Br2-Tyr (pKa: 7.21), and 3,5-I2-Tyr (pKa: 7.38) (Bioorg. Med. Chem. (2003) 11 (17), 3761-2768). Particularly preferred examples of the amino acid residue include amino acids having a side chain pKa of 5.5 to 7.0. Preferred examples of amino acids having such electron-donating properties include histidine.

Amino acids in the antigen binding domains can be modified by an appropriately selected method known in the art, such as site-directed mutagenesis (Kunkel et al., Proc. Natl. Acad. Sci. USA (1985) 82, 488-492) or overlap PCR. Also, amino acids can be replaced with unnatural amino acids using a variety of modification methods known in the art (Annu. Rev. Biophys. Biomol. Struct. (2006) 35, 225-249; and Proc. Natl. Acad. Sci. U.S.A. (2003) 100 (11), 6353-6357). For example, it is also preferable to use a tRNA cell-free translation system (Clover Direct (Protein Express, an R&D oriented company)) containing a non-natural amino acid linked to an amber suppressor tRNA complementary to the UAG codon, which is a stop codon. -codon.

Preferred examples of heavy chain variable regions for combination therewith include a randomized library of variable regions. A randomized library of variable regions is obtained by appropriately combining methods known in the art. In a non-limiting aspect of the present invention, the randomized variable region library may preferably be an immune library constructed from antibody genes derived from lymphocytes of animals immunized with specific antigens, patients with an infectious disease, people having increased antibody titers in the blood as a result of vaccination, patients with a malignant tumor or autoimmune diseases.

In a non-limiting aspect of the present invention, also preferably, a synthetic library obtained by replacing the CDR sequences of V genes of genomic DNA or reconstructed functional V genes with a synthetic set of oligonucleotides containing sequences encoding a set of codons having an appropriate length can be used as a randomized library of variable regions. In this case, they can only replace CDR3 sequences, since diversity is observed in the heavy chain CDR3 gene sequence. The amino acid diversity that provides the variable regions of antigen binding molecules is based on the diversity imparted to the amino acid residues at positions exposed on the surface of the antigen binding molecules. Surface exposed positions refer to positions that are believed to be surface exposed and/or accessible to antigens based on the structures, set of structures, and/or modeled structures of antigen binding molecules, and are typically located in their CDRs. Preferably, the positions exposed on the surface are determined using a computer program such as Insight II (Accelrys Inc.) and coordinates from three-dimensional models of antigen binding molecules. The positions exposed on the surface can be determined using an algorithm known in the art (e.g., Lee and Richards, J. Mol. Biol. (1971) 55, 379-400; and Connolly, J. Appl. Cryst. (1983) 16, 54 8-558). Positions exposed on the surface can be determined using software suitable for protein modeling and 3D structure information obtained for antibodies. Preferred examples of software that can be used for such a purpose include SYBYL biopolymer module software (Tripos Associates Inc.). Typically and preferably, the "size" of the probes used for the calculation is set to approximately 1.4 angstroms or lower in radius values when the algorithm requires the user to enter size parameters. A method for determining surface exposed areas and areas using personal computer software is described in Pacios, Comput. Chem. (1994) 18 (4), 377-386, and J. Mol. Model. (1995) 1, 46-53.

In another non-limiting aspect of the present invention, a naive library consisting of naive sequences, which are unbiased sets of antibody sequences constructed from antibody genes derived from lymphocytes of healthy humans, is used as a randomized variable region library (Gejima et al., Human Antibodies (2002). ) 11, 121-129; and Cardoso et al., Scand. J. Immunol. (2000) 51, 337-344).

As the amino acid other than the amino acid residue that changes the antigen-binding activity of the antigen-binding molecule depending on the ion concentration conditions, any amino acid can be used as long as the antigen-binding molecule of the present invention binds to antigens. As a preferred example, antibody phage display library technology known in the art may be used (eg, J.D. Marks et al., J. Mol. Biol. (1991) 222, 581-597). In particular, using a phage library, amino acid sequences are selected, with the exception of an amino acid residue that changes the antigen binding activity of the antigen binding molecule depending on the ion concentration conditions.

The term “sequence-differing” in a plurality of antigen-binding molecules differing in sequence as described herein means that the individual antigen-binding molecules in the library have different sequences. In particular, the number of different sequences in the library reflects the number of independent clones whose sequences differ in the library, and is also called the “library size”. The library size for a typical phage display library is from 10 ⁶ to 10 ¹² and can be expanded to 10 ¹⁴ using a method known in the art, such as the ribosome display method. However, the actual number of phage particles used in panning selection is typically 10-10,000 times the library size. This redundancy factor, also called the “number of library equivalents,” reflects the fact that 10 to 10,000 individual clones may have the same amino acid sequence. Thus, the term "sequence different" as used herein means that the individual antigen binding molecules in the library, except for the number of equivalents in the library, have different sequences, and more specifically means that the library has 10 ⁶ - 10 ¹⁴ , preferably 10 ⁷ -10 ¹² , more preferably 10 ⁸ -10 ¹¹ , particularly preferably 10 ⁸ -10 ¹⁰ antigen-binding molecules differing from each other in sequence.

The term "multiple" in a library of substantially multiple antigen binding molecules as described herein generally refers to a set of two or more types of substances, such as, for example, an antigen binding molecule, a fusion polypeptide, a polynucleotide molecule, a vector, or a virus the present invention. For example, if two or more substances differ from each other in a particular way, the substances exist as two or more types. Examples thereof may include amino acid variants observed at a particular amino acid position in an amino acid sequence. For example, two or more antigen binding molecules of the present invention having substantially the same, preferably identical, sequences except for loose residues or excluding specific amino acid variants at surface exposed amino acid positions with a high degree of diversity are considered multiple antigen binding molecules of the present invention. In another example, two or more polynucleotide molecules of the present invention having substantially the same, preferably identical sequences, except for bases encoding loose residues, or excluding bases encoding specific amino acid variants at surface exposed amino acid positions with a high degree of diversity, are considered to be a variety of polynucleotide molecules of the present invention.

The term "consisting essentially of" in a library consisting essentially of a plurality of antigen-binding molecules as described within the scope of the present invention reflects the number of antigen-binding molecules whose antigen-binding activity varies, depending on ion concentration conditions, among independent clones whose sequences differ, in library. In particular, preferably the library has at least 10 ⁴ antigen binding molecules that have such binding activity. More preferably, the present invention relates to a library having at least 10 ⁵ antigen-binding molecules that have such binding activity. Even more preferably, the present invention relates to a library having at least 10 ⁶ antigen binding molecules that have such binding activity. Particularly preferably, the present invention relates to a library having at least 10 ⁷ antigen binding molecules that have such binding activity. Also preferably, the present invention relates to a library having at least 10 ⁸ antigen binding molecules that have such binding activity. In other words, the term may preferably be indicated by the ratio of antigen-binding molecules whose antigen-binding activity differs depending on ion concentration conditions and the number of independent clones whose sequence differs in the library. In particular, the present invention relates to a library containing antigen-binding molecules that have such binding activity in a ratio of from 0.1% to 80%, preferably from 0.5% to 60%, more preferably from 1% to 40%, even more preferably from 2% to 20%, especially preferably from 4% to 10%, based on the number of independent sequence-different clones in the library. Fusion polypeptides, polynucleotide molecules or vectors can also be indicated by the number of molecules or the ratio relative to all molecules, as in the case described above. Viruses can also be indicated by the number of individual viruses or the ratio relative to all individual viruses, as in the case described above.

Fusion polypeptide containing an antigen-binding molecule

In one embodiment of the present invention, a fusion of an antigen binding molecule of the present invention and a heterologous polypeptide can be produced. In one embodiment, the antigen binding molecule of the present invention can be fused to at least a portion of a viral envelope protein selected from the group consisting, for example, of the viral envelope proteins pIII, pVIII, pVII, pIX, Soc, Hoc, gpD and pVI and their variants to form a fusion polypeptide.

In one embodiment, the antigen binding molecule of the present invention may be a ScFv, a Fab fragment, F(ab)2, or F(ab')2. In another embodiment, the present invention relates to a library consisting essentially of a plurality of fusion molecules differing from each other in sequence, each of the fusion molecules comprising any of the antigen binding molecules and a heterologous polypeptide. In particular, the present invention relates to a library consisting essentially of a plurality of fusion molecules differing from each other in sequence, each of the fusion proteins comprising any of these antigen-binding molecules and at least a portion of a viral envelope protein selected from the group consisting of, for example , from the viral envelope proteins pIII, pVIII, pVII, pIX, Soc, Hoc, gpD and pVI and their variants. The antigen binding molecule of the present invention may further comprise a dimerization domain. In one embodiment, the dimerization domain may be located between the variable region of the antibody heavy or light chain and at least a portion of the viral envelope protein. This dimerization domain may contain at least one dimerization sequence and/or a sequence containing one or more cysteine residues. This dimerization domain may preferably be associated with the C-terminus of the variable region or constant region of the heavy chain. The dimerization domain may have different structures depending on whether the antibody variable region is produced as a fusion protein component with a viral envelope protein component (there is no amber stop codon after the dimerization domain), or depending on whether the antibody variable region is produced mainly without the presence of a viral envelope protein component (for example, an amber stop codon after the dimerization domain is present). When the antibody variable region is produced substantially as a fusion protein with a viral envelope protein component, divalent display is accomplished by one or more disulfide bonds and/or one dimerization sequence. Preferably, the antigen binding molecule of the present invention has a dimerization domain containing both a cysteine residue and a dimerization sequence when the antibody variable region is produced substantially without fusion to a viral envelope protein component (eg, an amber stop codon is present). In one embodiment, the F(ab)2 heavy chains dimerize in a dimerization domain that does not contain a hinge region. This dimerization domain may contain, for example, a leucine zipper sequence known in the art, such as the sequence GCN4: GRMKQLEDKVEELLSKNYHLENEVARLKKLVGERG (SEQ ID NO: 12).

Polynucleotide Encoding an Antigen Binding Molecule A set of oligonucleotides or primers for use in obtaining a polynucleotide encoding each antigen binding molecule can be synthesized using a standard method. For example, one set of oligonucleotides containing sequences that include every possible combination of nucleotide triplets provided by a set of codons and encode the desired group of amino acids can be synthesized by a solid-phase method. Thus, a set of synthetic oligonucleotides having a specific set of codons, as a rule, contains many oligonucleotides that differ from each other in sequence. This difference is provided by the set of codons in the entire sequence. The synthesis of oligonucleotides having selected "degenerate" nucleotides at certain positions is known in the art. Such a set of nucleotides having this particular type of codon pattern can be synthesized using a commercially available nucleic acid synthesis apparatus (Applied Biosystems, Inc.) or can be obtained as commercially available products (eg, Life Technologies). As used herein, oligonucleotides have a sequence that allows hybridization to a variable domain nucleic acid template and may also contain restriction enzyme sites suitable for cloning.

The library can be constructed using sets of upstream and downstream oligonucleotides. Each set of oligonucleotides has a plurality of oligonucleotides having different sequences defined by the set of codons provided in the oligonucleotide sequence. Sets of upstream and downstream oligonucleotides can be used, together with variable domain template nucleic acid sequences, to produce a “library” of PCR products. The PCR products can be fused to other related or unrelated nucleic acid sequences, for example, nucleic acid sequences encoding viral envelope protein components and dimerization domains, using a conventional molecular biology approach. Thus, such a set of oligonucleotides may be called a "nucleic acid cassette".

The sequence of each PCR primer contains one or more sets of codons designed for low-stringency residues with a high degree of diversity exposed on the surface of the antigen-binding molecule. As described above, each set of codons is a collection of different sequences of nucleotide triplets that are used to encode the desired amino acid variants. Also, the set of oligonucleotides has a sequence of sufficient length to hybridize with the template nucleic acid and may optionally have restriction sites. DNA templates are formed using vectors derived from bacteriophage M13 vectors or vectors containing the single-stranded phage origin of replication described by Viera et al. (Meth. Enzymol. (1987) 153, 3). Thus, the DNA to be mutated is inserted into one of these vectors to form a single-stranded template. Preparation of the single-stranded template is described in Sambrook et al.

Methods for incorporating selected amino acids into antigen-binding molecules, as described in a non-limiting embodiment of the present invention, are already known in the art. Some of these methods are described in the present description. For example, "at least one amino acid residue that modifies the antigen-binding activity of the antigen-binding molecule depending on ion concentration conditions" or loose residues with a high degree of diversity can be exposed on the surface of at least one antigen-binding molecule and/or introduced into it using the Kunkel method , provided by Kunkel et al. (Methods Enzymol. (1987) 154, 367-382) to obtain a library. Such a method can be appropriately selected in the case of obtaining, for example, a library of the present invention containing a plurality of antigen-binding molecules differing from each other in sequence by combining light chain variable regions containing "at least one amino acid residue that modifies the antigen-binding activity antigen-binding molecule depending on ion concentration conditions" of the present invention, with heavy chain variable regions obtained as a randomized library of variable region sequences. When the metal ion is a calcium ion, non-limiting examples of such a case include the generation of light chain variable region sequences derived from the germline sequence SEQ ID NO: 6 (Vk1), SEQ ID NO: 7 (Vk2), SEQ ID NO: 8 (Vk3 ), SEQ ID NO: 9 (Vk4), etc. by replacing a particular residue with “at least one amino acid residue that alters the antigen-binding activity of the antigen-binding molecule depending on calcium ion concentration conditions.” The method as described above can be chosen to obtain such light chain variable regions.

In this case, the sets of oligonucleotides can be used as primers in a PCR reaction using the variable region nucleic acid template sequences as templates to produce nucleic acid cassettes. Each template nucleic acid sequence used is preferably a sequence that encodes any region in the light or heavy chain of an immunoglobulin and contains a given nucleic acid sequence (ie, a nucleic acid sequence encoding the amino acid to be replaced). Referring to the above example, a nucleic acid encoding a variable region in the germline sequence SEQ ID NO: 6 (Vk1), SEQ ID NO: 7 (Vk2), SEQ ID NO: 8 (Vk3), SEQ ID NO: 9 (Vk4), etc. The variable region nucleic acid template sequence encodes at least a portion of the variable region and encodes at least one CDR. In some cases, the variable region nucleic acid template sequence encodes a plurality of CDRs. Upstream and downstream regions of the variable region template nucleic acid sequence may be targets for hybridization of members of the upstream and downstream oligonucleotide primer sets. The first oligonucleotide of the upstream primer set can hybridize to the first nucleic acid strand, while the second oligonucleotide of the downstream primer set can hybridize to the second nucleic acid strand. Oligonucleotide primers can be designed to contain one or more sets of codons and hybridize to a portion of the variable region nucleic acid template sequence. The use of these oligonucleotides in PCR can introduce two or more sets of codons into the PCR product (ie, nucleic acid cassette). Oligonucleotide primers that hybridize to regions in the nucleic acid sequence encoding an antibody variable region may include regions encoding CDR residues to be replaced by other amino acids.

As described herein, also as a method for introducing selected amino acids into antigen-binding molecules, overlap PCR can be suitably selected as described for non-limiting embodiments of the present invention (WO 1993003151). “At least one amino acid residue that modifies the antigen-binding activity of an antigen-binding molecule depending on ion concentration conditions” or loose residues with a high degree of diversity can be exposed on the surface of and/or introduced into the at least one antigen-binding molecule to produce a library. The sets of upstream and downstream oligonucleotides used in overlapping PCR may contain a framework sequence of sufficient length to be hybridized to a variable region nucleic acid template sequence, together with a sequence encoding “at least one amino acid residue that modifies the antigen binding activity of the antigen binding molecule depending on ion concentration conditions" or loose residues with a high degree of diversity.

For example, sets of oligonucleotides are prepared encoding sets of codons for "at least one amino acid residue that modifies the antigen-binding activity of the antigen-binding molecule depending on ion concentration conditions" and additional non-stringent residues to produce polynucleotide molecules, each of which encodes light chain variable regions, into which an amino acid residue is introduced, as described in a non-limiting embodiment of the present invention. A framework sequence of sufficient length to hybridize to a variable region nucleic acid template sequence is linked upstream or downstream of the oligonucleotide arrays.

In the case of the introduction of an amino acid residue or additional non-strict residues, for example, in CDR2 and CDR3 of the light chain variable region, an oligonucleotide of sufficient length encoding the amino acid sequence FR2 adjacent to CDR2 is linked to the 5' ends of the set of oligonucleotides, while the oligonucleotide encoding the amino acid sequence of FR3 adjacent to CDR2 is linked to their 3' ends to produce primers. In addition, an oligonucleotide of sufficient length encoding the amino acid sequence of FR4 adjacent to CDR3 is linked to the 5' ends of a set of oligonucleotides complementary to the above set of oligonucleotides, while an oligonucleotide encoding the amino acid sequence of FR3 adjacent to CDR3 is linked to its 3 '-ends to obtain complementary primers. When the oligonucleotide encoding FR3 amino acids in the primers and the oligonucleotide encoding FR3 amino acids in the complementary primers contain overlapping sequences of sufficient length to hybridize to each other, light chain variable regions containing that amino acid residue or additional non-strict residues introduced into CDR2 and CDR3 can be obtained by a PCR reaction under conditions that do not involve the template sequence. When these oligonucleotides do not contain such overlapping sequences, light chain variable regions containing this amino acid residue or additional non-stringent residues introduced in CDR2 and CDR3 can likewise be generated by a PCR reaction using, for example, the variable region nucleic acid of the germline sequence SEQ ID NO: 6 (Vk1), SEQ ID NO: 7 (Vk2), SEQ ID NO: 8 (Vk3), SEQ ID NO: 9 (Vk4), etc. as a matrix sequence.

In the case of obtaining, for example, a library containing light chain variable regions having a CDR3 with an inserted amino acid residue or additional non-strict residues and randomized CDR1 and CDR2, an oligonucleotide of sufficient length encoding the amino acid sequence FR4 adjacent to CDR3 is linked to the 5'- ends of a set of oligonucleotides complementary to the specified set of oligonucleotides, while an oligonucleotide encoding the amino acid sequence of FR3 adjacent to CDR3 is linked to its 3' ends to obtain primers. Similarly, a library of polynucleotides encoding light chain variable regions having CDR3 with an inserted amino acid residue or additional non-stringent residues, and randomized CDR1 and CDR2, can be similarly obtained by PCR using the resulting primers, primers having the nucleotide sequence encoding FR1 , and a library containing randomized variable region nucleic acids (eg, naive sequence) as template sequences. These methods for producing a light chain polynucleotide library containing "at least one amino acid residue that modifies the antigen binding activity of the antigen binding molecule depending on ion concentration conditions" and/or non-stringent residues are also suitably selected for producing a heavy chain polynucleotide library, in which contains an amino acid residue and additional non-strict residues.

Alternatively, sets of upstream and downstream oligonucleotides can be synthesized such that each oligonucleotide sequence contains restriction sites. These restriction sites facilitate the incorporation of the nucleic acid cassette (ie, the product of the PCR reaction) into expression vectors having additional antibody sequences. Preferably, the restriction sites are designed to facilitate cloning of the nucleic acid cassette without introducing foreign nucleic acid sequences or without removing the original CDR nucleic acid sequences or framework regions.

The resulting nucleic acid cassette can be cloned into an appropriate vector to express part or all of the light or heavy chain sequence constituting the antigen binding molecule of the present invention. In accordance with the method described in detail within the scope of the present invention, the nucleic acid cassette is cloned into a vector that allows the production of partial or complete light or heavy chain sequence fused to all or part of the viral envelope protein (ie, formation of a fusion protein). Such a vector, as described below, may be an expression vector designed such that the particles or cells display the fusion protein on their surface. For such a purpose, several types of vectors are prepared and used to practice the present invention. Phagmid vectors are vectors preferably used within the scope of the present invention. As is well known to those skilled in the art, phagemid vectors typically may contain various components, including control sequences (eg, promoters and signal sequences), genes for phenotypic selection, origins of replication regions, and other necessary components.

In a non-limiting embodiment, to express specific amino acid variants in combination, the nucleic acid cassette may encode all or part of a heavy or light chain variable region. For example, in the case of designing a nucleic acid cassette such that the heavy chain CDR3 contains "at least one amino acid residue that modifies the antigen binding activity of the antigen binding molecule depending on ion concentration conditions" and/or non-stringent residues, as described in a non-limiting aspect of the present invention, Nucleic acid cassettes having heavy chain CDR3 sequence diversity can be prepared and then linked to nucleic acid cassettes encoding random variable regions, such as naïve sequences. As in a library, nucleic acid cassettes can be inserted into expression vectors containing additional antibody sequences, such as whole or partial light and heavy chain variable or constant regions, to produce antigen binding molecules containing these amino acid variants or combinations of amino acid variants. These additional sequences in the antigen binding molecules can be fused to other nucleic acid sequences, for example, sequences encoding a component of the viral envelope protein. As a result, fusion proteins can be obtained.

Vector

One non-limiting aspect of the present invention includes an expression vector that can be replicated containing a nucleic acid sequence encoding a fusion gene, wherein the fusion gene encodes a fusion protein comprising one antigen binding molecule variable region or one antigen binding molecule variable region and one constant region fused to the entire protein viral envelope or part thereof. One non-limiting aspect of the present invention also includes a library of diverse expression vectors that can be replicated, each of which contains a plurality of fusion genes, each of which encodes a plurality of different fusion proteins containing variable regions with diverse sequences formed, as described above, in a variety of antigen binding molecules , differing from each other in sequence. Vectors may contain various components and are preferably designed so that the variable region genes of the antigen binding molecules can be transferred between different vectors and/or the fusion proteins can be displayed in different formats.

Preferred examples of vectors include phage vectors. Phage vectors have an origin of replication that allows phage replication and formation of the phage particle. Preferred examples of phage may include filamentous bacteriophages, including phages M13, f1, fd and Pf3 and their derivatives, and lambdoid phages, including phage lambda, 21, phi80, phi81, 82, 424, 434 and other phages and their derivatives.

Examples of viral envelope protein include infectious protein PIII, major envelope protein PVIII, pVII, pIX, Soc (T4), Hoc (T4), gpD (bacteriophage λ), and bacteriophage minor envelope protein 6 (pVI) (filamentous phage (J. Immunol Methods (1999) 231 (1-2), 39-51); and a variant of the basic protein of bacteriophage M13 (P8) (Protein Sci. (2000) 9 (4), 647-654). The fusion protein is displayed on the surface of the phage. Suitable phage systems include helper phage M13KO7, M13R408, M13-VCS and Phi X 174, pJuFo phage system (J. Virol. (2001) 75 (15), 7107-7113), hyperphage (Nat. Biotechnol. (2001) 19 ( 1), 75-78) and KM13 (Fold Des. (1998) 3 (5), 321-328). The preferred helper phage is M13KO7. The preferred coat protein is the M13 phage gene III coat protein. The preferred host is E. coli, and a protease-deficient strain of E. coli. For example, vectors such as fth1 vectors (Nucleic Acids Res. (2001) 29 (10) e50) may be suitable for expressing the fusion protein.

The expression vector may also contain a secretory signal sequence fused to DNA encoding a heavy or light chain variable region of an antigen binding molecule or fragment thereof. This sequence is typically located on the 5' side of the gene encoding the fusion protein and is thus transcribed at the N-terminus of the fusion protein. However, in some cases it has been demonstrated that the signal sequence is located at a position different from the position 5' side of the gene encoding the protein to be secreted. This sequence directs the protein to which the sequence is linked through the inner membrane of the bacterial cell. The DNA encoding the signal sequence is obtained as a restriction endonuclease cleavable fragment from any gene encoding a protein having the signal sequence. For example, the gene encoding LamB or OmpF (Wong et al., Gene (1983), 68 (2), 193-203), MalE or PhoA and other genes can be used as suitable prokaryotic signal sequences. A preferred prokaryotic signal sequence for practicing the present invention is the E. coli thermostable enterotoxin II (STII) signal sequence described by Chang et al. (Gene (1987) 55 (2-3), 189-196), pelB or malE.

The vector typically contains a promoter as a control sequence that drives expression of the fusion protein. Promoters commonly used in prokaryotic vectors include the lacZ promoter system, the alkaline phosphatase phoA (Ap) promoter, the bacteriophage λPL promoter (temperature-sensitive promoter), the tac promoter (a hybrid trp-lac promoter that is regulated by the lac repressor), the tryptophan promoter, the pBAD and bacteriophage T7 promoter. Promoters are reviewed in Sambrook et al. (higher). Although these promoters are the most commonly used, other suitable microbial promoters can be used similarly.

The vector may also contain other nucleic acid sequences, such as sequences encoding gD tags, c-Myc epitopes, polyhistidine tags, fluorescent proteins (e.g., GFP), or a β-galactosidase protein useful for detecting and purifying a surface-expressed fusion protein phage or cells. For example, a gD tag encoding a nucleic acid sequence also provides positive or negative selection of cells or viruses expressing the fusion protein. In some embodiments, the gD tag is preferably fused to an antigen-binding variable region molecule not fused to a viral envelope protein component. For example, a nucleic acid sequence encoding a polyhistidine tag is useful for identifying a fusion protein containing a variable region of an antigen-binding molecule that binds a specific antigen using an immunohistochemical approach. Such a tag useful for detecting antigen binding may be fused to a variable region of an antigen binding molecule not fused to a viral envelope protein component or a variable region of an antigen binding molecule fused to a viral envelope protein component.

Preferred examples of other useful components in the vector used to practice the present invention include genes for phenotypic selection. A typical gene for phenotypic selection is one that encodes a protein that confers antibiotic resistance to host cells. As such an example, the ampicillin resistance gene (ampr) and the tetracycline resistance gene (tetr) can be preferably used.

The vector may also contain a nucleic acid sequence containing unique restriction sites and a suppressible stop codon. The unique restriction sites are useful for transferring the antigen binding molecule variable region gene between different vectors and expression systems and are particularly useful for the production of a full length antigen binding molecule or antigen binding fragment by cell culture.

The suppressible stop codon is useful for regulating the level of expression of the fusion protein and facilitates the purification of a soluble fragment of the antigen-binding molecule. For example, an amber stop codon can be translated into Gln in a supE host capable of phage display, while in a non-supE host that codon is interpreted as a stop codon, allowing the production of a soluble fragment of the antigen-binding molecule unfused to the phage coat protein . These synthetic sequences can be fused to genes encoding one or more antigen binding molecule variable regions in a vector.

Preferably, a vector may be used that allows the nucleic acid encoding the antigen binding molecule sequence of interest to be readily removed from the vector and placed into another vector. For example, suitable restriction sites can be incorporated into the vector to facilitate extraction of the nucleic acid sequence encoding the antigen binding molecule of the present invention or a variable region thereof. Restriction sequences are typically selected as unique sequences in the vector to facilitate efficient cutting and ligation into new vectors. The antigen binding molecule or variable region thereof can then be expressed from the vectors as a molecule having a structure free of fused foreign sequences, such as a viral envelope protein or other sequence tags.

The DNA encoding the termination or stop codon can be inserted between the nucleic acid encoding the variable region or constant region of the antigen binding molecule (gene 1) and the nucleic acid encoding the viral envelope protein component (gene 2). Such termination codons include UAG (amber), UAA (ochre) and UGA (opal) (Davis et al., Microbiology (1980), p. 237, 245-247 and 374, Harper & Row, New York). A termination codon or stop codon expressed in wild-type host cells results in the synthesis of a gene 1 protein product unrelated to the gene 2 protein. However, the fusion protein is synthesized in detectable amounts when grown in suppressor host cells. Such suppressor host cells are known in the art and have been described as E. coli suppressor gene strains (Bullock et al., BioTechniques (1987) 5, 376-379), etc. Such a termination codon can be inserted into the mRNA encoding the fusion polypeptide.

The suppressible codon can be inserted between a first gene encoding a variable or constant region of an antigen binding molecule and a second gene encoding at least a portion of a phage coat protein. Alternatively, a suppressible termination codon can be inserted adjacent to the fusion site by replacing the last amino acid triplet in the variable region of the antigen binding molecule or the first amino acid triplet in the phage coat protein. A suppressible termination codon can be inserted at or after a position corresponding to the C-terminus of the dimerization domain. Replication of the plasmid containing the suppression codon in host suppressor cells provides a fusion polypeptide containing the polypeptide and the coat protein in a detectable amount. Thus, replication of the plasmid in non-suppressor host cells terminates translation on the integrated suppressible triplet UAG, UAA or UGA and thus results in synthesis of the variable region of the antigen-binding molecule substantially without fusion with the phage coat protein. Thus, the variable region of the antigen binding molecule expressed in non-suppressor cells is secreted from the host cells after synthesis due to the absence of a phage coat fusion protein to be anchored to the host membrane.

The variable or constant regions of the light chain and/or heavy chain of the antigen binding molecule can be further fused to a peptide sequence that allows one or more fusion polypeptides to interact on a viral particle or cell surface. This peptide sequence is referred to herein as a “dimerization domain.” The dimerization domain may contain at least one or more dimerization sequences or at least one sequence containing a cysteine residue, or both. Suitable dimerization sequences include dimerization sequences for proteins having amphipathic α-helices that contain regularly spaced hydrophobic residues and allow dimer formation by interaction of the hydrophobic residues of each protein. Such proteins and protein parts contain, for example, leucine zipper regions. The dimerization domain may also contain one or more cysteine residues (eg, provided by the antibody hinge region sequence contained in the dimerization domain). Cysteine residues mediate dimerization by forming one or more disulfide bonds. In one embodiment, where the stop codon is located downstream of the dimerization domain coding sequence, the dimerization domain contains at least one cysteine residue. The dimerization domain is preferably located between the variable or constant domain of the antibody and the viral envelope protein component.

In some cases, the vector encodes, for example, a phage polypeptide of a single antigen-binding molecule in single-chain form containing heavy and light chain variable regions fused to a coat protein. In these cases, the vector is considered "monocistronic", such that a single transcript is expressed under the control of a specific promoter. Examples of such a vector include vectors that use an alkaline phosphatase (AP) promoter or a Tac promoter to drive the expression of a monocistronic sequence encoding a light chain variable region (VL domain) and a heavy chain variable region (VH domain), between which a linker peptide is located. . Such a cistron sequence is linked at its 5' end to the E. coli malE or thermostable enterotoxin II (STII) signal sequence and at its 3' end to the whole or part of the viral envelope protein (eg pIII protein). In some embodiments, the vector may further comprise a dimerization domain coding sequence (eg, a leucine zipper) as shown in SEQ ID NO: 12 at its 3' end between the second variable region sequence and the viral envelope protein sequence.

In other cases, the heavy and light chain variable regions may be expressed as separate polypeptides. Such a vector is “bicistronic” and allows the expression of individual transcripts. In this vector, a suitable promoter, such as the tac or PhoA promoter, can be used to drive expression of the bicistronic mRNA. The first cistron, encoding, for example, the light chain variable and constant regions, is linked at its 5' end to the E. coli malE, pelB or thermostable enterotoxin II (STII) signal sequence and at its 3' end to a nucleic acid sequence encoding gD-tag. The second cistron, encoding, for example, the variable region and constant region CH1 of the heavy chain, is linked at its 5' end to the signal sequence of malE or thermostable enterotoxin II (STII) of E. coli and at its 3' end to the entire viral envelope protein or part of it.

In a vector that produces a bicistronic mRNA and provides F(ab')2-pIII display, a suitable promoter, such as the tac or PhoA (AP) promoter, drives the expression of the first cistron, which encodes the light chain variable and constant regions, and is operably linked to it 5'-end with the signal sequence of malE or heat-stable enterotoxin II (STII) of E. coli and at its 3'-end with a nucleic acid sequence encoding a gD tag. The second cistron encodes, for example, the heavy chain variable and constant regions and is operably linked at its 5' end to the E. coli malE or thermostable enterotoxin II (STII) signal sequence and at its 3' end to a sequence encoding a dimerization domain containing an IgG hinge region sequence and a leucine zipper sequence followed by at least a portion of a viral envelope protein.

Fusion polypeptide display

The antigen binding molecule variable region fusion polypeptide can be displayed in various forms on the surface of a cell, virus or phagemid particle. These forms include single chain Fv fragments (scFv), F(ab) fragments, and multivalent forms of these fragments. The polyvalent forms are preferably a dimer of ScFv, Fab or F(ab'), which are referred to herein as (ScFv)2, F(ab)2 and F(ab')2, respectively. Display of multivalent forms is preferable, perhaps because displayed multivalent forms typically provide identification of low affinity clones and/or have multiple antigen binding sites that allow for more efficient selection of rare clones during breeding.

Methods for displaying fusion polypeptides containing antibody fragments on the surface of bacteriophages are known in the art and are described, for example, in WO 1992001047 and in the present description. Other similar methods are described in WO 1992020791, WO 1993006213, WO 1993011236 and 1993019172. Those skilled in the art can make appropriate use of these methods. Other public literature (H.R. Hoogenboom & G. Winter (1992) J. Mol. Biol. 227, 381-388, WO 1993006213 and WO 1993011236) describes the identification of antibodies using artificially rearranged variable region gene repertoires against various surface-exposed antigens phages.

When a display vector is constructed in the form of an scFv, the vector contains nucleic acid sequences encoding the light and heavy chain variable regions of the antigen binding molecule. Typically, a nucleic acid sequence encoding the heavy chain variable region of an antigen binding molecule is fused to a nucleic acid sequence encoding a viral envelope protein component. A nucleic acid sequence encoding a light chain variable region of an antigen binding molecule is linked to a nucleic acid sequence encoding a heavy chain variable region of an antigen binding molecule through a nucleic acid sequence encoding a peptide linker. The peptide linker typically contains from about 5 to 15 amino acids. Optionally, an additional sequence encoding, for example, a tag useful for purification or detection, can be fused to the 3' end of a nucleic acid sequence encoding the light chain variable region of an antigen binding molecule or a nucleic acid sequence encoding the heavy chain variable region of an antigen binding molecule, or with both of them.

When a display vector is constructed in the form of F(ab), the vector contains nucleic acid sequences encoding the variable and constant regions of the antigen binding molecule. A nucleic acid sequence encoding a light chain variable region is fused to a nucleic acid sequence encoding a light chain constant region. A nucleic acid sequence encoding the heavy chain variable region of an antigen binding molecule is fused to a nucleic acid sequence encoding the heavy chain CH1 constant region. Typically, the nucleic acid sequence encoding the heavy chain variable and constant regions is fused to a nucleic acid sequence encoding all or part of the viral envelope protein. The heavy chain variable and constant regions are preferably expressed as a fusion product with at least a portion of the viral envelope protein, while the light chain variable and constant regions are expressed separately from the heavy chain-viral envelope protein fusion protein. The heavy and light chains can associate with each other through a covalent bond or a non-covalent bond. Optionally, an additional sequence encoding, for example, a polypeptide tag useful for purification or detection may be fused to the 3' end of a nucleic acid sequence encoding the light chain constant region of an antigen binding molecule or a nucleic acid sequence encoding the heavy chain constant region of an antigen binding molecule , or both of them.

Transfer of the vector into the host cell

Vectors constructed in this manner are transferred into host cells for amplification and/or expression. Vectors can be transferred into cells using transformation methods known in the art, including electroporation, calcium phosphate precipitation, and the like. When the vectors are infectious particles such as viruses, the vectors themselves enter the host cells. The fusion proteins are displayed on the surface of phage particles by transfecting host cells with replicable expression vectors having polynucleotide inserts encoding the protein and producing phage particles using an approach known in the art.

Expression vectors that can be replicated can be transferred into host cells using various methods. In a non-limiting embodiment, vectors can be transferred into cells by electroporation, as described in WO 2000106717. Cells are cultured at 37°C, optionally for about 6-4 8 hours (or until the OD at 600 nm reaches 0.6 -0.8) in standard culture medium. Next, the culture medium is centrifuged and the culture supernatant is removed (eg, by transfusion). During the initial purification step, the cell pellet is preferably resuspended in a buffer solution (eg, 1.0 mM HEPES (pH 7.4)). Next, the suspension is centrifuged again to remove the supernatant. The resulting cell sediment is resuspended in glycerol, diluted, for example, to 5-20% v/v. The suspension is again centrifuged to remove the supernatant, thereby obtaining a cell pellet. The resulting cell pellet is resuspended in water or diluted glycerol. Based on the measurement result of the obtained cell concentration, the final cell concentration is adjusted to the desired concentration using water or diluted glycerol.

Examples of preferred receptor cells include E. coli strain SS320, which is capable of responding to electroporation (Sidhu et al., Methods Enzymol. (2000) 328, 333-363). E. coli strain SS320 was obtained by crossing MCI 0 61 cells with XL1-BLUE cells under conditions sufficient to transfer the fertility episome (F'-plasmid) from XL1-BLUE to MC1061 cells. E. coli strain SS320 has been deposited at ATCC (10801 University Boulevard, Manassas, Virginia) under deposit number 98795. Within the framework of the present invention, any F' episome can be used that allows phage replication in this strain. Suitable episome can be obtained from strains deposited in ATCC or can be obtained as commercially available products (TGI, CJ23 6, CSH18, DHF', ER2738, JM101, JM103, JM105, JM107, JM109, JM110, KS1000, XL1-BLUE, 71-18, etc.).

Using higher concentrations of DNA (approximately 10-fold) during electroporation increases the frequency of transformation and increases the amount of DNA transforming host cells. The use of high concentrations of cells also increases efficiency (by approximately 10 times). The increased amount of DNA transferred can provide a library that has greater diversity and a greater number of independent clones that differ in sequence. Transformed cells are usually selected based on the presence or absence of growth on media containing the antibiotic.

A method for selecting an antigen-binding molecule whose binding activity varies depending on conditions, and a method for isolating a polynucleotide encoding the molecule

Methods for using phage display to identify antigen-binding molecules that exhibit desired binding activity against antigens have already been developed in the art, also including variously modified methods thereof. In one non-limiting aspect, suitable host cells are transformed with a library of replicable expression vectors designed to contain transcriptional regulatory elements operably linked to fusion genes encoding the fusion polypeptides. The transformed cells are cultured to form phage particles displaying the fusion polypeptides on their surface. A "binding molecule selection" step is then carried out, which involves selecting and sorting by bringing the recombinant phage particles into contact with target antigens so that at least a portion of a subpopulation of particles binds to the antigens, in order to increase the content of antigen-bound particles in the subpopulation relatively antigen-associated particles during selection. For another round of screening under different or the same reaction conditions, the sorted subpopulation is amplified, for example, by infecting a subpopulation of fresh host cells, such as XL1-Blue cells. Next, the step of bringing the thus amplified subpopulation of the resulting antigen-binding molecules into contact with antigens at different ion concentrations is carried out to screen for antigen-binding molecules that bind to antigens under the desired conditions. This step of bringing the subpopulation into contact with antigens at different ion concentrations can be carried out as an initial step in the selection method. The combination of the "selection of binding molecules" step and the step of selecting binding molecules whose binding activity changes under different ion concentration conditions can be suitably modified. Also, the selection and sorting stage can be carried out as many times as desired. These antigen-binding molecules, which bind to antigens under various ion concentration conditions, are useful as therapeutic drugs capable of rapidly eliminating pathogenic antigens from organisms when administered to the organisms.

Fusion polypeptides of variable regions or parts thereof containing an amino acid that modifies the antigen-binding activity of an antigen-binding molecule depending on ion concentration conditions, and/or non-stringent residues in accordance with the present invention are displayed on the surface of phages, phagemid particles or cells. Further, members comprising a subpopulation of fusion polypeptides can be selected and/or sorted with respect to their ability to bind antigens at different ion concentrations. The method for selecting, sorting and screening fusion polypeptides also includes a method for selecting, sorting and screening on common proteins having affinity for variable regions of antigen binding molecules, such as the L protein or labeled antibodies capable of binding to antigen binding molecules or fragments thereof displayed on phages. This method can be used to increase the size of the library and the number of library equivalents that display correctly folded antigen binding molecules (or fusion polypeptides containing molecules).

For the selection, sorting and screening described above, two main strategies can be used. The first strategy is a solid support method, plate screening or immobilized antigen screening. The second strategy is the solution binding method.

In the "solid support method", antigens can be coupled to a suitable solid or semi-solid matrix known in the art, for example, agarose beads, acrylamide beads, glass beads, cellulose, various acrylic copolymers, hydroxyalkyl methacrylate gels, polyacrylic and polymethacrylic copolymers, nylon, and neutral and ionic carriers. Attachment of antigens to the matrix can be achieved in the manner described in Methods in Enzymology (1976) 44 or other methods known in the art.

After attachment of the antigen to the matrix, the immobilized antigens are contacted with a library expressing the fusion polypeptides under conditions suitable for binding at least one subpopulation of phage particles to the immobilized antigens. Typically, conditions including pH, ionic strength, temperature, etc. are chosen to mimic physiological conditions. Bound particles ("binding elements") with immobilized antigens are separated from particles not bound to the target by washing with water. Washing conditions can be adjusted to remove all but the high affinity binders. Binding elements can be dissociated from the immobilized target in various ways. These methods include, for example, competitive dissociation using wild-type ligands (eg, excess antigens), changes in pH and/or ionic strength, and methods known in the art. Typically, binding elements can be eluted from the affinity matrix using an appropriate elution material, such as acid (eg, 0.1 M HCl) or ligands. Elution at increased ligand concentrations can elute exposed binding molecules with higher affinity.

Suitable host cells can be infected with isolated binding elements, i.e. viral particles (and, if necessary, helper phages, for example when the viral particles are phagemid particles) to re-amplify the binding elements in the cells. The host cells thus obtained are cultured under conditions suitable for amplification of particles that exhibit the desired fusion polypeptide. Next, the phage particles are collected and the selection step is repeated one or more times until the antigen-binding elements increase in number to a significant extent. Selection and screening can be carried out as many times as desired. One selection or screening method may involve isolating binding elements that bind to common affinity proteins, such as the L protein or antibodies against polypeptide tags present in exposed polypeptides, for example, antibodies against gD protein or polyhistidine tags.

In one aspect of the present invention, a liquid phase screening method called a "solution binding method" can be suitably used. The solution binding method can be used to identify improved binding elements from a random library aimed at increasing the binding activity of the desired binding clone or group of clones. This method involves contacting a plurality of polypeptides, eg polypeptides displayed on phage or phagemid particles (library), with antigens labeled or fused to tag molecules. Biotin or other specific binding molecules can be used as a tag. The stringency of the solution phase can be varied by using gradually decreasing concentrations of labeled or fused antigens in the first phase of the binding solution. To further increase stringency, binding to the first antigens labeled with label molecules or fused to labeled label molecules in the first solution phase may suitably be followed by additional contact with a second solution phase having a high concentration of antigens unlabeled with label molecules or unfused to labeled label molecules . In this case, antigens unlabeled with tag molecules or unfused with labeled tag molecules are typically used in an amount ranging from 100 to 1000 times the amount of labeled target in the second phase. The incubation time for the first phase of the solution ranges from a few minutes to 1-2 hours or more until equilibrium is reached. Since binding elements having a high association rate tend to bind to the target quickly in this first phase, contacting conditions can be selected to shorten the binding time. The incubation time and temperature of the second phase can be varied to increase hardness. This variation in incubation conditions provides a selection bias for binding elements that dissociate from the target at a slow rate (dissociation rate). After contacting the plurality of polypeptides displayed on the phage or phagemid particles with antigens, the phage or phagemid particles bound to the antigens labeled with tag molecules or fused to the labeled tag molecules are separated from the unbound phage or phagemid particles. Phage or phagemid particle-antigen mixtures are recovered from the solution phase by contacting the phage or phagemid particles with antigens for a short period of time (e.g., 2 to 5 minutes), which allows for binding to antigens labeled with tag molecules or fused to labeled molecules marks. The initial concentration of antigens labeled with tag molecules or fused to labeled tag molecules ranges from about 0.1 nM to about 1000 nM. The particles can be eluted from the mixture and then grown during the next round of screening. Preferably, multiple rounds of screening are repeated using a lower concentration of antigens labeled with tag molecules or fused to labeled tag molecules in each round of screening. As described below in the Examples section, for example, streptavidin-coated beads can suitably be used in a solution coupling method using biotin as the labeled label molecule.

The solid support method and the solution binding method can suitably be carried out separately or in combination to isolate binding elements having the desired feature. After two, three or four rounds of selection and screening against antigens, a subpopulation selected to identify specific binding elements having the desired property or trait is typically screened against individual clones. Preferably, the screening process can be carried out using an automated system that provides high throughput screening of the library.

Once binding elements have been identified based on antigen binding, nucleic acids can be extracted from the binding elements. The extracted DNA is then used to directly transform E. coli host cells. Alternatively, their coding sequences can be amplified, for example, by PCR using appropriate primers, and then sequenced by a typical sequencing method. Further, DNA binding elements encoding the variable region can be inserted into restriction enzyme digested vectors to express the encoded antigen binding molecules.

To select and screen phage particles expressing an antigen-binding molecule of the present invention, the antigen-binding activity of which varies depending on ion concentration conditions, or a fusion polypeptide thereof, a subpopulation of phage particles whose antigen-binding activity varies is sorted by varying the conditions for bringing the immobilized antigens into contact with a library containing phage particles expressing antigen-binding molecules or fusion polypeptides.

Taking calcium ion as a preferred example of a metal ion, examples of calcium ion concentration conditions include low calcium ion concentration conditions and high calcium ion concentration conditions. The expression "binding activity varies depending on calcium ion concentration conditions" means that the antigen binding activity of each antigen binding molecule changes depending on the difference between low calcium ion concentration conditions and high calcium ion concentration conditions. Examples of this include higher antigen binding activity of an antigen binding molecule under conditions of high calcium ion concentration than under conditions of low calcium ion concentration. Another example of this involves higher antigen binding activity of an antigen binding molecule under conditions of low calcium ion concentration than under conditions of high calcium ion concentration.

In the present description, the high calcium ion concentration is not specifically limited to a single numerical value, and may preferably be a concentration selected from the range of 100 μM to 10 mM. In another aspect, the high calcium ion concentration may be a concentration selected from the range of 200 μM to 5 mM. In another aspect, the high calcium ion concentration may also be a concentration selected from the range of 500 μM to 2.5 mM. In an alternative aspect, the high calcium ion concentration may also be a concentration selected from the range of 200 μM to 2 mM. In addition, this concentration may also be a concentration selected from the range of 400 μM to 1.5 mM. Particularly preferred examples thereof include concentrations selected from the range of 500 μM to 2.5 mM, which are close to in vivo plasma calcium ion concentrations.

In the present description, the low calcium ion concentration is not particularly limited to a single numerical value, and may preferably be a concentration selected from the range of 0.1 μM to 30 μM. In another aspect, the low calcium ion concentration may be a concentration selected from the range of 0.2 μM to 20 μM. In another aspect, the low calcium ion concentration may also be a concentration selected from the range of 0.5 μM to 10 μM. In an alternative aspect, the low calcium ion concentration may also be a concentration selected from the range of 1 μM to 5 μM. In addition, this concentration may also be a concentration selected from the range of 2 μM to 4 μM. Particularly preferred examples thereof include concentrations selected from the range of 1 μM to 5 μM, which are close to the concentrations of ionized calcium in the early endosome in vivo.

For the purposes of the present invention, the expression "the antigen binding activity of an antigen binding molecule is lower under low calcium ion concentration conditions than under high calcium ion concentration conditions" means that the antigen binding activity of the antigen binding molecule at a calcium ion concentration selected from the range of 0.1 μM to 30 µM is weaker than the calcium ion concentration selected from the range from 100 µM to 10 mM. This expression preferably means that the antigen binding activity of the antigen binding molecule at a calcium ion concentration selected from the range of 0.5 μM to 10 μM is weaker than at a calcium ion concentration selected from the range from 200 μM to 5 mM. Particularly preferably, this expression means that the antigen binding activity at the early endosome calcium ion concentration in vivo is weaker than at the plasma calcium ion concentration in vivo. In particular, this means that the antigen binding activity of the antigen binding molecule at a calcium ion concentration selected from the range of 1 μM to 5 μM is weaker than at a calcium ion concentration selected from the range from 500 μM to 2.5 mM.

To select and screen, for example, phage particles expressing antigen-binding molecules whose antigen-binding activity is higher under high-calcium conditions than under low-calcium conditions, or fusion polypeptides thereof, a subpopulation of phage particles that bind to immobilized antigens is first screened under conditions of high calcium concentration. The sorted subpopulation (a library containing phage particles expressing antigen-binding molecules or fusion polypeptides) is then brought into contact with immobilized antigens under conditions of low calcium ion concentration. Phage particles that do not bind to immobilized antigens under conditions of low calcium ion concentration are sorted and separated into the supernatant or subsequent washes. The high calcium concentration condition and the low calcium concentration condition can be suitably selected from the ranges described above. In a non-limiting aspect, the subpopulation can be sorted based on the differences between antigen binding activity at a calcium ion concentration selected from the range of 0.1 μM to 30 μM, and antigen binding activity at a calcium ion concentration selected from the range from 100 μM to 10 mM. In another non-limiting aspect, the subpopulation can be sorted based on the differences between antigen binding activity at a calcium ion concentration selected from the range of 0.5 μM to 10 μM, and antigen binding activity at a calcium ion concentration selected from the range from 200 μM to 5 mM. In an alternative, non-limiting aspect, the subpopulation can be sorted based on differences between antigen binding activity at an early endosome calcium ion concentration in vivo and antigen binding activity at an in vivo plasma calcium ion concentration, in particular, differences between antigen binding activity at a calcium ion concentration selected from 1 μM to 5 μM, and antigen binding activity at a calcium ion concentration selected from the range of 500 μM to 2.5 mM.

Taking the hydrogen ion concentration as a preferred example of the ion concentration, examples of hydrogen ion concentration conditions include acidic pH conditions and neutral pH conditions. The expression "binding activity varies depending on pH conditions" means that the antigen binding activity of each antigen binding molecule changes depending on the differences between acidic pH values and neutral pH values. Examples of this include higher antigen binding activity of the antigen binding molecule at neutral pH values than at acidic pH values. Another example of this involves the higher antigen binding activity of an antigen binding molecule at acidic pH values than at neutral pH values.

In the present description, the neutral pH value is not particularly limited to a single numerical value, but preferably can be selected from the range of pH 6.7 to pH 10.0. In another aspect, the neutral pH value may be selected from the range of pH 6.7 to pH 9.5. In another aspect, the neutral pH value may be selected from the range of pH 7.0 to pH 9.0. In an alternative aspect, this pH value may be selected from the range of pH 7.0 to pH 8.0. Particularly preferred examples thereof include a pH value of 7.4, which is similar to the pH found in plasma (blood) in vivo.

In the present description, the acidic pH value is not particularly limited to a single numerical value, and preferably can be selected from the range of pH 4.0 to pH 6.5. In another aspect, the acidic pH value may be selected from the range of pH 4.5 to pH 6.5. In another aspect, the acidic pH value may be selected from a range of pH 5.0 to pH 6.5. In an alternative aspect, this pH value may be selected from the range of pH 5.5 to pH 6.5. Particularly preferred examples include a pH value of 5.8, which is close to the ionized calcium concentration in the early endosome in vivo.

For the purposes of the present invention, the expression "the antigen binding activity of an antigen binding molecule is lower under neutral pH conditions than under acidic pH conditions" means that the antigen binding activity of the antigen binding molecule at a pH value selected from the range of pH 4.0 to pH 6.5 , is weaker than at a pH value selected from the range of pH 6.7 to pH 10.0. This expression preferably means that the antigen binding activity of the antigen binding molecule at a pH value selected from the range of pH 4.5 to pH 6.5 is weaker than at a pH value selected from the range of pH 6.7 to pH 9.5 . More preferably, the expression means that the antigen binding activity of the antigen binding molecule at a pH value selected from the range of pH 5.0 to pH 6.5 is weaker than at a pH value selected from the range of pH 7.0 to pH 9, 0. In addition, the expression preferably means that the antigen binding activity of the antigen binding molecule at a pH value selected from the range of pH 5.5 to pH 6.5 is weaker than at a pH value selected from the range of pH 7.0 to pH 8 ,0. Particularly preferably, this expression means that the antigen binding activity at the pH in the early endosome in vivo is weaker than at the pH in the plasma in vivo. In particular, this means that the antigen binding activity of the antigen binding molecule at pH 5.8 is weaker than at pH 7.4.

Determination of whether the antigen binding activity of an antigen binding molecule changes depending on pH conditions can be carried out, for example, in accordance with the method of analyzing binding activity under different pH conditions described above. For example, the antigen binding activity of an antigen binding molecule is compared between acidic pH conditions and neutral pH conditions to confirm that the antigen binding activity of the antigen binding molecule changes to a higher level under neutral pH conditions than under acidic pH conditions.

To select and screen, for example, phage particles expressing antigen-binding molecules whose antigen-binding activity is higher under neutral pH conditions than under acidic pH conditions, or their fusion polypeptides, first screen the subpopulation of phage particles that bind immobilized antigens under conditions neutral pH values. The sorted subpopulation (a library containing phage particles expressing antigen-binding molecules or fusion polypeptides) is then brought into contact with immobilized antigens under acidic pH conditions. Phage particles that do not bind to immobilized antigens under acidic pH conditions are sorted and separated into the supernatant or subsequent washes. Neutral pH conditions and acidic pH conditions can be suitably selected from the ranges described above. In a non-limiting aspect, the subpopulation can be sorted based on the differences between antigen binding activity at a pH value selected from the range of pH 4.0 to pH 6.5 and antigen binding activity at a pH value selected from pH 6.7 to pH 10.0 . In another non-limiting aspect, the subpopulation can be sorted based on the differences between antigen binding activity at a pH value selected from the range of pH 4.5 to pH 6.5 and antigen binding activity at a pH value selected from pH 6.7 to pH 9. 5. In an alternative non-limiting aspect, the subpopulation can be sorted based on differences between antigen binding activity at a pH value selected from the range of pH 5.0 to pH 6.5 and antigen binding activity at a pH value selected from the range pH 7.0 to pH 9 ,0. In addition, the subpopulation can be sorted based on the differences between antigen binding activity at a pH value selected from the range of pH 5.5 to pH 6.5 and antigen binding activity at a pH value selected from pH 7.0 to pH 8.0 . In a further alternative non-limiting aspect, the subpopulation can be sorted based on differences between antigen binding activity at early endosome pH in vivo and antigen binding activity at plasma pH in vivo, particularly differences between antigen binding activity at pH 5.8 and antigen binding activity at pH 7 ,4.

Method for producing an antigen-binding molecule whose binding activity varies depending on conditions

In a non-limiting aspect of the present invention, a polynucleotide encoding a thus selected antigen binding molecule, the binding activity of which varies depending on conditions, is isolated and then inserted into an appropriate expression vector. Specifically, after each cDNA encoding an antigen binding molecule variable region of interest is obtained, the cDNA is digested with restriction enzymes that recognize restriction sites embedded at both ends of the cDNA. Preferably, restriction enzymes recognize and cleave a nucleotide sequence that occurs at low frequency in the nucleotide sequences constituting the genes for antigen-binding molecules. Moreover, preferably, restriction enzymes cleave the sites embedded therein to form sticky ends to insert one copy of the cleaved fragment in the correct direction into the vector. cDNAs encoding the variable region of an antigen-binding molecule, thus cleaved, can be inserted into appropriate expression vectors to obtain expression vectors for the antigen-binding molecule of the present invention. In this case, the genes encoding the constant region (C region) of the antibody can be fused in frame with the genes encoding the variable region.

To produce the desired antigen binding molecule, a polynucleotide encoding the antigen binding molecule is inserted into expression vectors in a form operably linked to control sequences. Control sequences include, for example, enhancers and promoters. Also, a suitable signal sequence for extracellular secretion of the expressed antigen-binding molecule can be linked to its amino terminus. For example, a peptide having the amino acid sequence MGWSCIILFLVATATGVHS (SEQ ID NO: 13) can be used as a signal sequence. Any suitable signal sequence may be attached. The expressed polypeptide is cleaved at the C-terminus of the signal sequence, and the cleaved polypeptide can be secreted extracellularly as a mature polypeptide. Suitable host cells can be transformed with these expression vectors to produce recombinant cells expressing a polynucleotide encoding the desired antigen binding molecule.

To express a polynucleotide encoding an antigen binding molecule, a polynucleotide encoding a heavy chain and a polynucleotide encoding a light chain of an antigen binding molecule are separately inserted into different expression vectors. The same host cell can be cotransfected with a vector into which a heavy chain has been inserted and a vector into which a light chain has been inserted, thereby allowing the expression of antigen-binding molecules containing both heavy and light chains. Alternatively, the polynucleotides encoding the heavy chain and the light chain can be inserted into a single expression vector, which is then transformed into a host cell (see WO 1994011523).

Many combinations of host cells and expression vectors are known in the art to produce an antigen binding molecule by transferring an isolated polynucleotide encoding the antigen binding molecule into appropriate hosts. All of these expression systems can be used to isolate the antigen binding molecule of the present invention. In the case of using eukaryotic cells as host cells, animal, plant or fungal cells can be suitably used. Specifically, examples of animal cells may include the following cells:

(1) mammalian cells such as CHO (Chinese hamster ovary cell line), COS (monkey kidney cell line), myeloma cells (Sp2/0, NS0, etc.), BHK (baby hamster cell line), HEK293 (a human embryonic kidney cell line with cleaved adenovirus DNA (Ad)5), PER.C6 cells (a human embryonic retinal cell line transformed with the E1A and E1B genes of adenovirus type 5 (Ad5)), Hela and Vero (Current Protocols in Protein Science (May 2001, section 5.9, table 5.9.1));

(2) amphibian cells such as Xenopus oocytes; And

(3) insect cells such as sf9, sf21 and Tn5.

Alternatively, antibody gene expression systems using cells derived from the genus Nicotiana (eg, Nicotiana tabacum) as plant cells are known in the art. Cultured callus cells can be suitably used to transform plant cells.

The following cells can be used as fungal cells:

- cells derived from yeasts of the genus Saccharomyces (for example, Saccharomyces cerevisiae) and the genus Pichia (for example, Pichia pastoris), and

- cells derived from filamentous fungi of the genus Aspergillus (for example, Aspergillus niger).

Also known in the art are expression systems for polynucleotides encoding antigen binding molecules using prokaryotic cells. In the case of using, for example, bacterial cells, cells of bacteria such as E. coli and Bacillus subtilis can be suitably used. Expression vectors containing a polynucleotide of interest encoding an antigen-binding molecule are transferred into these cells by transformation. The transformed cells are cultured in vitro, and the desired antigen-binding molecule can be obtained from the resulting transformed cell cultures.

In addition to host cells, transgenic animals can be used to produce a recombinant antigen-binding molecule. In particular, the desired antigen binding molecule can be obtained from animals transfected with a polynucleotide encoding the antigen binding molecule. For example, a polynucleotide encoding an antigen-binding molecule can be inserted in frame into genes encoding proteins specifically produced in milk, constructing fusion genes. As proteins secreted in milk, for example, goat β-casein and the like can be used. DNA fragments containing fusion genes having an insertion of a polynucleotide encoding an antigen-binding molecule are injected into goat embryos, which are in turn implanted into female goats. From the milk produced by transgenic goats (or their progeny) born to embryo-implanted goats, the desired antigen-binding molecule can be produced as a fusion protein with milk protein. In addition, hormones can be administered to transgenic goats to increase the amount of milk containing the desired antigen-binding molecule produced by transgenic goats (Bio/Technology (1994), 12 (7), 699-702).

In the case of administering the antigen binding molecule described herein to a human, the antigen binding domain for the molecule can be suitably selected as an antigen binding domain derived from a genetically recombinant antibody that is artificially constructed, for example, to reduce the heteroantigenicity of a human. Genetically recombinant antibody includes, for example, humanized antibodies. These engineered antibodies are suitably produced using a method known in the art.

Each antigen binding molecule variable region used to provide an antigen binding domain in an antigen binding molecule described herein typically consists of three complementarity determining regions (CDRs) flanked by four framework regions (FRs). CDRs are regions that essentially determine the binding specificity of an antigen binding molecule. CDRs have a variety of amino acid sequences. On the other hand, FRs are largely formed by amino acid sequences that are highly identical even among antigen-binding molecules with different binding specificities. Thus, in general, the binding specificity of a particular antigen-binding molecule can be transplanted into other antigen-binding molecules by grafting CDRs.

Humanized antibodies are also called reconstituted human antibodies. In particular, what is known in the art is, for example, a humanized antibody consisting of a human antibody into which the CDRs of a non-human animal (eg, mouse) antibody are transplanted. General gene recombination approaches are also known to produce humanized antibodies. In particular, for example, overlap PCR is known in the art as a method for transplanting mouse antibody CDRs into human FRs. In overlapping PCR, a nucleotide sequence encoding each CDR of the mouse antibody to be transplanted is added to the primers for the synthesis of human FR antibodies. Primers are prepared for each of the four FRs. For transplantation of mouse CDRs into human FRs, it is generally considered advantageous to select human FRs that are highly identical to mouse FRs to preserve the functions of the CDRs. In particular, it is generally preferable to use human FRs containing amino acid sequences highly identical to the amino acid sequences of the FRs adjacent to the mouse CDRs to be transplanted.

The nucleotide sequences to be linked are designed such that the sequences are linked in frame with each other. Nucleotide sequences encoding human FR are synthesized individually using their corresponding primers. The resulting products contain DNA encoding a mouse CDR fused to each human FR encoding sequence. The nucleotide sequences encoding the mouse CDRs are designed such that the nucleotide sequence in each product overlaps with another nucleotide sequence. The overlapping CDR regions in the products synthesized from human antibody genes as templates are then annealed to each other for a complementary strand synthesis reaction. Using this reaction, human FR sequences are linked via mouse CDR sequences.

Finally, the full-length V region gene sequence, containing the linked three CDRs and four FRs, is amplified using primers, each of which anneals to its 5' and 3' ends and has an associated recognition sequence for the corresponding restriction enzyme. The DNA thus obtained and the DNA encoding the C region of a human antibody can be inserted into expression vectors so as to fuse these DNAs in frame to obtain humanized antibody expression vectors. These insert vectors are transferred into hosts to produce recombinant cells. The recombinant cells are then cultured to express DNA encoding the humanized antibody to produce humanized antibodies in cultured cell cultures (EP 239400 and WO 1996002576).

The humanized antibodies thus obtained can be assessed for their antigen binding activity using a quantitative or qualitative assay, thereby selecting suitable human antibody FRs that allow the CDRs to form a suitable antigen binding site when bound through these CDRs. If necessary, the amino acid residue(s) of FR can be replaced so that the CDRs of the resulting reconstituted human antibody form a suitable antigen binding site. For example, the amino acid sequence of FR can be mutated using the PCR method used to transplant a mouse CDR into a human FR. In particular, mutation of a partial nucleotide sequence can be introduced into primers that anneal to the FR nucleotide sequence. The FR nucleotide sequence synthesized using such primers contains the mutation introduced in this way. Such antibody variants having the amino acid(s) replaced can be assessed for their antigen binding activity using the same assay as described above, thereby selecting FR sequence variants having the desired properties (Sato et al., Cancer Res (1993) 53, 851-856).

In a non-limiting aspect of the present invention, a modified form of the thus isolated polynucleotide encoding a selected antigen binding molecule, the binding activity of which varies depending on conditions, is then inserted into appropriate expression vectors. One preferred example of such a modified form includes a humanized form of a polynucleotide sequence encoding an antigen binding molecule of the present invention found by screening using, as a randomized variable region library, a synthetic library or an immune library derived from a non-human animal as a source. The humanized form of the antigen-binding molecule can be prepared by selecting a method similar to the method for producing the humanized antibody described above.

In other aspects, preferred examples of a modified form include modified forms of an isolated polynucleotide sequence that is modified to provide increased antigen binding affinity (affinity maturation) of an antigen binding molecule of the present invention found by screening using a synthetic library or a naive library as a randomized variable region library . Such modified forms can be obtained using various affinity maturation methods known in the art, including CDR mutagenesis (Yang et al., J. Mol. Biol. (1995) 254, 392-403), chain shuffling (Marks et al., Bio/Technology (1992) 10, 779-783), use of E. coli mutator strains (Low et al., J. Mol. Biol. (1996) 250, 359-368), DNA shuffling (Patten et al., Curr. Opin. Biotechnol. (1997) 8, 724-733), phage display (Thompson et al., J. Mol. Biol. (1996) 256, 77-88) and sex PCR (Clameri et al., Nature ( 1998) 391, 288-291).

Examples of the antigen binding molecule obtained by the production method of the present invention include antigen binding molecules containing an FcRn (particularly human FcRn) binding domain as described above. The FcRn (particularly human FcRn) binding domain can be used in various modified forms. In one aspect, preferred examples of a modified form of the present invention also include a polynucleotide encoding an antigen binding molecule having a heavy chain, wherein a polynucleotide encoding a selected antigen binding molecule whose binding activity varies depending on conditions is linked in frame to a polynucleotide encoding such a modified the shape of the FcRn (specifically human FcRn) binding domain.

In a non-limiting aspect of the present invention, preferred examples of an FcRn (particularly human FcRn) binding domain include the constant regions of antibodies such as IgG1 corresponding to SEQ ID NO: 14 (AAC82527.1 with Ala added at the N-terminus), IgG2 corresponding to SEQ ID NO: 15 (AAB59393.1 with Ala added at N-terminus), IgG3 corresponding to SEQ ID NO: 16 (CAA27268.1), and IgG4 corresponding to SEQ ID NO: 17 (AAB59394.1 with Ala added at N -end). The relatively long retention time of IgG molecules in plasma (slow disappearance from plasma) is due to the functions of FcRn, known as the rescue receptor, for IgG molecules. IgG molecules taken up into the endosome via pinocytosis bind to FcRn expressed in the endosome under the acidic conditions of the endosome. IgG molecules that are not bound to FcRn are removed to the lysosome and degraded therein, while FcRn-bound IgG molecules migrate to the cell surface and dissociate from FcRn under neutral conditions in the plasma, returning to the plasma.

The “antigen-binding molecule whose antigen-binding activity varies with calcium ion concentration conditions” of the present invention is strongly associated with the antigen under conditions of high plasma calcium concentration and dissociates from the antigen under conditions of low endosomal calcium ion concentration. Thus, this antigen-binding molecule can dissociate from the antigen in the endosome. Thus, the “antigen-binding molecule whose antigen-binding activity varies with calcium ion concentration conditions” of the present invention, which is strongly associated with an antigen under conditions of high plasma calcium concentration and dissociates from the antigen under conditions of low endosomal calcium ion concentration, dissociates from the antigen and can re-associate with the antigen after recycling into plasma by FcRn. This allows one antigen-binding molecule to bind multiple antigens multiple times. Also, antigen bound to an antigen-binding molecule dissociates from it in the endosome and thus is not recycled into the plasma. This ensures the cellular uptake of antigens by an antigen-binding molecule. Administration of an antigen-binding molecule may accelerate the disappearance of the antigen and reduce the concentration of the antigen in the plasma.

The "antigen-binding molecule whose antigen-binding activity varies with pH conditions" of the present invention is strongly associated with the antigen under neutral pH conditions in the plasma and dissociates from the antigen under acidic conditions in the endosome. Thus, this antigen-binding molecule can dissociate from the antigen in the endosome. "An antigen-binding molecule whose antigen-binding activity varies with pH conditions" of the present invention, which is strongly associated with an antigen under neutral plasma pH conditions and dissociates from the antigen under acidic conditions in the endosome, thereby dissociating from the antigen and being able to re-associate with antigen after recycling into plasma using FcRn. This allows one antigen-binding molecule to bind multiple antigens multiple times. Also, antigen bound to an antigen-binding molecule dissociates from it in the endosome and thus is not recycled into the plasma. This ensures the cellular uptake of antigens by an antigen-binding molecule. Administration of an antigen-binding molecule may accelerate the disappearance of the antigen and reduce the concentration of the antigen in the plasma.

The ability to bind to FcRn under conditions of high plasma calcium concentration can be imparted to an “antigen-binding molecule whose antigen-binding activity varies depending on calcium ion concentration conditions” of the present invention, which is strongly associated with the antigen under conditions of high plasma calcium concentration and dissociates from the antigen under conditions of low concentration of calcium ions in the endosome. This ensures cellular uptake of the antigen-binding molecule-antigen complex. Thus, administration of an antigen-binding molecule can accelerate the disappearance of antigen from plasma and reduce the concentration of antigen in plasma.

The ability to bind to FcRn under neutral plasma pH conditions can be imparted to an “antigen-binding molecule whose antigen-binding activity varies with pH conditions” of the present invention, which is strongly associated with the antigen under neutral plasma pH conditions and dissociates from the antigen under acidic conditions pH values in the endosome. This accelerates the cellular uptake of the antigen-binding molecule-antigen complex. Thus, administration of an antigen-binding molecule can accelerate the disappearance of antigen from plasma and reduce the concentration of antigen in plasma.

A conventional Fc region antibody does not have FcRn binding activity under neutral plasma pH conditions. Conventional antibody and antibody-antigen complex are thus taken up into cells through nonspecific endocytosis and transported to the cell surface by binding to FcRn under acidic pH conditions in the endosome. Because FcRn is responsible for the transport of antibody from the endosome to the cell surface, it is believed that some FcRn receptors also exist on the cell surface. The antibody dissociates from FcRn under neutral pH conditions at the cell surface and is thus recycled into the plasma.

Considering the in vivo kinetics of an antibody containing an Fc region (FcRn-binding domain), an “antigen-binding molecule whose antigen-binding activity varies depending on calcium ion concentration conditions” of the present invention, which strongly associates with an antigen under conditions of high plasma calcium concentration and dissociates from the antigen under conditions of low concentration of calcium ions in the endosome, associates with the antigen in the plasma and dissociates from the bound antigen in the endosome. Thus, the rate of antigen disappearance is considered to be equal to the rate of cellular uptake through nonspecific endocytosis. If the binding of an antigen by an antigen-binding molecule is insufficiently dependent on the concentration of calcium ions, antigens that have not dissociated in the endosome are recycled into the plasma. On the contrary, if the binding of an antigen by an antigen-binding molecule is sufficiently dependent on the concentration of calcium ions, the rate of disappearance of the antigen is determined by the rate of cellular uptake through nonspecific endocytosis.

An "antigen-binding molecule whose antigen-binding activity varies with pH conditions" of the present invention, which is strongly associated with an antigen under neutral pH conditions in the plasma and dissociates from the antigen under acidic pH conditions in the endosome, associates with an antigen in the plasma and dissociates from bound antigen in the endosome. Thus, the rate of antigen disappearance is considered to be equal to the rate of cellular uptake through nonspecific endocytosis. In the case of insufficient pH dependence of antigen binding by an antigen-binding molecule, antigens that have not dissociated in the endosome are recycled into the plasma. In contrast, if the binding of an antigen by an antigen-binding molecule is sufficiently pH dependent, the rate of antigen disappearance is determined by the rate of cellular uptake through nonspecific endocytosis.

Thus, an antigen-binding molecule containing an FcRn-binding domain can be imparted with the ability to bind to FcRn at neutral pH values. The resulting antigen-binding molecule is taken up into cells in an FcRn-dependent manner through binding to FcRn present on the cell surface. The rate of FcRn-mediated cellular uptake is higher than the rate of cellular uptake through nonspecific endocytosis. In this regard, imparting the ability to bind to FcRn at neutral pH may further increase the rate of antigen disappearance from plasma. In particular, an antigen-binding molecule that has the ability to bind to FcRn at neutral pH delivers antigen to cells faster than a conventional Fc-region-containing antigen-binding molecule (which does not have FcRn-binding activity at neutral plasma pH) and dissociates from the antigen at endosome. The dissociated antigen-binding molecule is recycled to the cell surface or plasma, where the molecule re-associates with the antigen, again leading to FcRn-mediated cellular uptake. The greater ability to bind to FcRn at neutral pH may increase the turnover rate of this cycle and thus increase the rate at which the antigen disappears from the plasma. The antigen-binding activity of an antigen-binding molecule at acidic pH values can be reduced relative to the antigen-binding activity at neutral pH values, thereby further enhancing the effectiveness. An increased number of turns of this cycle per antigen-binding molecule will possibly allow one antigen-binding molecule to contact a higher number of antigens.

The antigen binding molecule of the present invention may comprise an antigen binding domain and an FcRn (particularly human FcRn) binding domain. The FcRn binding domain neither influences antigen binding nor is it dependent on the type of antigen, as is also clear from the mechanism mentioned above. The antigen binding activity (binding ability) of an antigen binding molecule under conditions of low calcium ion concentration is reduced relative to the antigen binding activity under conditions of high calcium ion concentration, and/or the FcRn binding activity of this molecule at neutral plasma pH is increased. This can accelerate the cellular uptake of antigens by the antigen-binding molecule and accelerate the rate of disappearance of antigens.

As described above, the antigen binding molecule of the present invention may contain an antigen binding domain and an FcRn binding domain, which neither affects antigen binding nor depends on the type of antigen, as is also understood from the mechanism mentioned above. The antigen binding activity (binding capacity) of an antigen binding molecule under acidic pH conditions is reduced relative to the antigen binding activity under neutral pH conditions, and/or the FcRn binding activity of the molecule under neutral pH conditions in plasma is increased. This can accelerate the cellular uptake of antigens by the antigen-binding molecule and accelerate the rate of disappearance of antigens.

The FcRn binding activity of an FcRn binding domain (eg, an antibody constant region) can be assayed in a manner well known to those skilled in the art, as mentioned above in the binding activity paragraph. Conditions other than pH can be determined accordingly by those skilled in the art. The antigen binding activity and human FcRn binding activity of an antigen binding molecule can be assessed, for example, based on KD (dissociation constant), apparent KD (apparent dissociation constant), kd (dissociation rate), or apparent kd (apparent dissociation rate). These parameters can be measured in a manner well known to those skilled in the art, for example, using a Biacore (GE Healthcare Bio-Sciences Corp.), a Scatchard plot, or a flow cytometer.

Conditions other than pH for assaying the human FcRn binding activity of the FcRn binding domain (eg, the constant region of an antibody) can be appropriately selected by those skilled in the art without particular limitation. Human FcRn binding activity can be assayed under conditions of, for example, MES buffer and 37°C, as described in WO 2009125825. Also, human FcRn binding activity of an FcRn binding domain (eg, antibody constant region) can be assayed in a manner well known to those skilled in the art. for example, using Biacore (GE Healthcare Bio-Sciences Corp.). In a human FcRn binding activity assay, an FcRn binding domain (eg, an antibody constant region) can be assessed for its binding ability by injecting the human FcRn analyte onto a chip immobilized with an FcRn binding domain or an antigen binding molecule of the present invention containing an FcRn binding domain. , or by injecting an assayed FcRn binding domain or an antigen binding molecule of the present invention containing an FcRn binding domain onto a human FcRn immobilized chip.

Neutral pH values, as conditions under which an FcRn-binding domain (eg, an antibody constant region) or an antigen-binding molecule comprising an FcRn-binding domain has FcRn-binding activity, generally means conditions between pH 6.7 and pH 10.0. The neutral pH value is preferably a range indicated by any pH value from pH 7.0 to pH 8.0, and is preferably selected from pH 7.0, 7.1, 7.2, 7.3, 7.4, 7.5, 7.6, 7.7, 7.8, 7.9 and 8.0, especially preferably pH 7.4, which is close to the pH in plasma (blood) in vivo. The binding affinity of the human FcRn binding domain for human FcRn at pH 7.4 may be difficult to assess due to its low binding affinity. In this case, instead of pH 7.4, pH 7.0 can be used. For the purposes of the present invention, acidic pH as the conditions under which a human FcRn binding domain or an antigen binding molecule comprising a human FcRn binding domain has FcRn binding activity generally means pH 4.0 to pH 6.5. Acidic pH preferably means pH 5.5 to pH 6.5, and particularly preferably means pH 5.8 to pH 6.0, which is close to the pH in the early endosome in vivo. The temperature used in the assay conditions may be any temperature from 10° C. to 50° C. at which a human FcRn binding domain or an antigen binding molecule comprising a human FcRn binding domain is assessed for its human FcRn binding affinity. Preferably, a temperature of 15°C to 40°C is used to determine the human FcRn binding affinity of a human FcRn binding domain or an antigen binding molecule comprising a human FcRn binding domain. More preferably, any temperature between 20° C. and 35° C. is also used to determine the human FcRn binding affinity of a human FcRn binding domain or an antigen binding molecule comprising a human FcRn binding domain, for example any of 20, 21, 22, 23 , 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34 and 35°C. In one aspect of the present invention, one non-limiting example is a temperature of 2-5°C.

According to Journal of Immunology (2009) 182: 7663-7671, the binding activity of natural human IgG1 to human FcRn in the acidic pH range (pH 6.0) is KD 1.7 μM, but this activity is practically undetectable at neutral pH values. Thus, in a preferred aspect, it is possible to screen for an antigen binding molecule of the present invention having human FcRn binding activity at acidic pH values and at neutral pH values, which includes an antigen binding molecule having a human FcRn binding activity of a KD of 20 μM or greater at acidic pH values. pH values and human FcRn binding activity equivalent to or greater than the human FcRn binding activity of natural IgG at neutral pH values. In a more preferred aspect, an antigen binding molecule can be screened for having a human FcRn binding activity of 2.0 μM or greater at acidic pH and a human FcRn binding activity of 40 μM or greater at neutral pH. In a further preferred aspect, it is possible to screen for an antigen binding molecule having a human FcRn binding activity of 0.5 μM or greater at acidic pH and a human FcRn binding activity of 15 μM or greater at neutral pH. These KD values are determined by the method described in The Journal of Immunology (2009) 182: 7663-7671 (which involves immobilizing antigen binding molecules onto a chip and injecting a human FcRn analyte onto it).

Within the scope of the present invention, the FcRn binding domain preferably has human FcRn binding activity at acidic pH values and at neutral pH values. An FcRn binding domain naturally having human FcRn binding activity at acidic pH and at neutral pH can be directly used as the domain. When the domain has no or weak human FcRn binding activity at acidic pH and/or neutral pH, the amino acid(s) in the antigen binding molecule can be modified to produce an FcRn binding domain having the desired human FcRn binding activity. Alternatively, the amino acid(s) in the human FcRn binding domain may preferably be modified to provide an FcRn binding domain having the desired human FcRn binding activity at acidic pH and/or at neutral pH. Also, the amino acid(s) in the FcRn binding domain that naturally has human FcRn binding activity at acidic pH and/or neutral pH values can be modified to provide an FcRn binding domain having the desired human FcRn binding activity. Such an amino acid modification that imparts a desired binding activity to the human FcRn binding domain can be identified by comparing the human FcRn binding activity at acidic pH and/or at neutral pH before and after the amino acid modification. Those skilled in the art can appropriately carry out the amino acid modification using an approach known in the art.

As used herein, the term "modification of amino acid(s)" or "amino acid modification" in an FcRn binding domain includes a modification to an amino acid sequence different from the amino acid sequence of the original FcRn binding domain. Any FcRn binding domain can be used as the parent domain, provided that the resulting modified form of the parent FcRn binding domain can bind to human FcRn at acidic pH and/or at neutral pH (thus, the parent FcRn is not necessarily required to -binding domain had human FcRn binding activity under acidic pH and/or neutral pH conditions). Preferred examples of the original FcRn binding domain include the constant regions of IgG antibodies, i.e. natural constant regions corresponding to any of SEQ ID NO: 14-17. Alternatively, as the modified FcRn binding domain of the present invention, an FcRn binding domain further modified from an already modified FcRn binding domain can be preferably used as the original FcRn binding domain. The parent FcRn binding domain may mean the polypeptide itself, a composition containing the parent FcRn binding domain, or an amino acid sequence encoding the parent FcRn binding domain. The original FcRn binding domain may include the FcRn binding domains of IgG antibodies known in the art, produced by recombination discussed in the antibody paragraph. The original FcRn binding domain is not limited in its origin, and can be derived from any non-human animal or human organism. Preferred examples of a random organism include one selected from mice, rats, guinea pigs, hamsters, gerbils, cats, rabbits, dogs, goats, sheep, cattle, horses, camels and non-human primates. In another aspect, the original FcRn-binding domain can be obtained from cynomolgus monkey, marmoset, rhesus monkey, chimpanzee or human. Preferably, the parent Fc region can be derived from human IgG1, although the parent FcRn binding domain of the present invention is not limited to a particular class of IgG. This means that a human IgG1, IgG2, IgG3 or IgG4 Fc region can be used as the source Fc region. In Sequences of Proteins of Immunological Interest, NIH Publication No. 91-3242, a variety of allotypic sequences inherent in polymorphism are described as the Fc regions of human IgG1, human IgG2, human IgG3 and human IgG4, any of which can be used within the scope of the present invention. In particular, human IgG1 may have a sequence with DEL or EEM as the amino acid sequence at positions 356-358, defined by EU numbering. Likewise, for the purposes of the present invention, this means that the starting FcRn binding domain can preferably be an FcRn binding domain of an arbitrary class or subclass of IgG from an arbitrary organism. Examples of naturally occurring IgG variants or manipulated forms thereof are described in the public literature (Curr. Opin. Biotechnol. (2009) 20 (6), 685-91; Curr. Opin. Immunol. (2008) 20 (4), 460- 470; Protein Eng. Des. Sel. (2010) 23 (4), 195-202; WO 2009086320, WO 2008092117, WO 2007041635 and WO 2006105338), although the variants or manipulated forms of the present invention are not limited to those described in these documents

The amino acid(s) in the FcRn binding domain involved in FcRn binding are suitably mutated to impart the ability to bind FcRn (particularly human FcRn) at neutral pH to an antigen binding molecule containing an FcRn binding domain (particularly FcRn person). In the case of using a constant region of an IgG antibody molecule as an FcRn binding domain, examples of such a modified FcRn binding domain include constant regions derived from natural IgG constant regions by modifying amino acids at positions 221-225, 227, 228, 230, 232 , 233-241, 243-252, 254-260, 262-272, 274, 276, 278-289, 291-312, 315-320, 324, 325, 327-339, 341, 343, 345, 360, 362 , 370, 375-378, 380, 382, 385-387, 389, 396, 414, 416, 423, 424, 426-438, 440 and 442, identified by EU numbering, to amino acids other than the corresponding naturally occurring amino acids . More specifically, examples thereof include modified forms of constant regions containing Pro at amino acid position 256, Lys at amino acid position 280, Thr at amino acid position 339, His at amino acid position 385, Leu at amino acid position 428, and/or Trp, Tyr, Phe, Ala or His at amino acid position 434 (all defined by EU numbering). The use of these modified forms can enhance the activity of the Fc region of IgG immunoglobulin in relation to human FcRn binding at neutral pH values.

Alternatively, a modified form capable of binding to human FcRn more strongly at acidic pH values than the natural IgG constant region can be suitably used. Such a modified form can be suitably selected for its ability to also bind strongly to human FcRn at neutral pH and used within the scope of the present invention. Examples of such a modified form include constant regions derived from natural IgG constant regions by modifying the amino acids identified by EU numbering to amino acids other than the corresponding naturally occurring amino acids. As an example of such amino acid modification, modified constant regions containing the amino acids listed in Table 1 can preferably be used.

Examples of particularly preferred amino acid modification include a modification that replaces amino acids at positions 237, 238, 239, 248, 250, 252, 254, 255, 256, 257, 258, 265, 270, 286, 289, 297, 298, 303, 305 , 307, 308, 309, 311, 312, 314, 315, 317, 325, 332, 334, 360, 376, 380, 382, 384, 385, 386, 387, 389, 424, 428, 433, 434 and 4 36 , defined by EU numbering, by amino acids different from the corresponding amino acids in the constant region of natural IgG. At least one amino acid selected from these amino acids can be modified into another amino acid, thereby enhancing the activity of the FcRn binding domain to bind human FcRn at neutral pH.

Examples of particularly preferred modification include modification of the natural IgG constant region, which results in:

Met at amino acid position 237,

Ala at amino acid position 238,

Lys at amino acid position 239,

Ile at amino acid position 248,

Ala, Phe, He, Met, Gin, Ser, Val, Trp or Tyr at amino acid position 250,

Phe, Trp or Tyr at amino acid position 252,

Thr at amino acid position 254,

Glu at amino acid position 255,

Asp, Glu or Gin at amino acid position 256,

Ala, Gly, Ile, Leu, Met, Asn, Ser, Thr or Val at amino acid position 257,

His at amino acid position 258,

Ala at amino acid position 265,

Phe at amino acid position 270,

Ala or Glu at amino acid position 286,

His at amino acid position 289,

Ala at amino acid position 297,

Gly at amino acid position 298,

Ala at amino acid position 303,

Ala at amino acid position 305,

Ala, Asp, Phe, Gly, His, lie, Lys, Leu, Met, Asn, Pro, Gin, Arg, Ser, Val, Trp or Tyr at amino acid position 307,

Ala, Phe, Ile, Leu, Met, Pro, Gin or Thr at amino acid position 308,

Ala, Asp, Glu, Pro or Arg at amino acid position 309,

Ala, His, or Ile at amino acid position 311,

Ala or His at amino acid position 312,

Lys or Arg at amino acid position 314,

Ala or His at amino acid position 315,

Ala at amino acid position 317,

Gly at amino acid position 325,

Val at amino acid position 332,

Leu at amino acid position 334,

His at amino acid position 360,

Ala at amino acid position 376,

Ala at amino acid position 380,

Ala at amino acid position 382,

Ala at amino acid position 384,

Asp or His at amino acid position 385,

Pro at amino acid position 386,

Glu at amino acid position 387,

Ala or Ser at amino acid position 389,

Ala at amino acid position 424,

Ala, Asp, Phe, Gly, His, lie, Lys, Leu, Asn, Pro, Gin, Ser, Thr, Val, Trp or TYr at amino acid position 428,

Lys at amino acid position 433,

Ala, Phe, His, Ser, Trp, or Tyr at amino acid position 434, and

His at amino acid position 436 (all defined by EU numbering).

The number of amino acids to be modified is not particularly limited. Only one amino acid may be modified, or two or more amino acids may be modified. Examples of combinations of two or more amino acids to be modified include combinations as shown in Table 2.

Examples of modification include one or more mutations, for example, a mutation that replaces an amino acid(s) in the original Fc region with amino acid residue(s) different from it, an insertion of one or more amino acid residues into the amino acid sequence of the original Fc region, and deletion of one or more amino acids from the amino acid sequence of the original Fc region. Preferably, the amino acid sequence of the Fc region so modified comprises an amino acid sequence comprising at least a non-natural portion of the Fc region. Such a variant inevitably has less than 100% sequence identity with the original Fc region. In a preferred embodiment, the variant has an amino acid sequence with about 75% to less than 100% sequence identity or similarity, more preferably with about 80% to less than 100%, even more preferably with about 85% to less than 100%, even more preferably with about 90% to less than 100%, most preferably with about 95% to less than 100% sequence identity or similarity to the amino acid sequence of the parent Fc region. In a non-limiting aspect of the present invention, the original Fc region and the modified Fc region of the present invention differ in at least one amino acid. The difference of an amino acid between the original Fc region and the modified Fc region may preferably be determined by the difference of the amino acid at a specified position of its amino acid residue, determined in particular by the EU numbering system.

The amino acid(s) in the Fc region can be modified by an appropriately selected method known in the art, such as site-directed mutagenesis (Kunkel et al., Proc. Natl. Acad. Sci. USA (1985) 82, 488- 492) or overlap PCR. Also, the amino acid(s) can be replaced with unnatural amino acids using a variety of modification methods known in the art (Annu. Rev. Biophys. Biomol. Struct. (2006) 35, 225-249; and Proc. Natl. Acad. Sci. U.S.A. (2003) 100(11), 6353-6357). For example, it is also preferable to use a tRNA cell-free translation system (Clover Direct (Protein Express, an R&D oriented company)) containing a non-natural amino acid associated with an amber suppressor tRNA complementary to the UAG codon, which is a stop codon. -codon.

In one aspect of the modified form of the present invention, a polynucleotide encoding an antigen binding molecule having a heavy chain is produced, in which the polynucleotide encoding a modified FcRn binding domain having the amino acid(s) thereby mutated is linked in frame to the polynucleotide , encoding a selected antigen-binding molecule whose binding activity varies depending on conditions.

The present invention relates to a method for producing an antigen binding molecule, comprising isolating the antigen binding molecule from cell cultures transfected with a vector having an operably linked insert in which a polynucleotide encoding an FcRn binding domain is linked in frame to a polynucleotide isolated from a virus of the present invention. The present invention also relates to a method for producing an antigen-binding molecule, comprising isolating the antigen-binding molecule from cultures of cells transfected with a vector having an operably linked insert, in which a polynucleotide encoding an FcRn-binding domain, previously operably linked to the vector, is linked in frame to the polynucleotide isolated from the virus of the present invention.

Pharmaceutical composition

Although the present invention is not bound by any particular theory, for example, the number of antigens that can be bound per antigen binding molecule is increased and the disappearance of the antigen from the plasma is accelerated as a result of accelerating cellular uptake in the body into which the antigen binding molecule containing the antigen binding a domain whose antigen-binding activity varies with ion concentration conditions such that antigen-binding activity is lower under acidic pH conditions than under neutral pH conditions, and which also contains an FcRn-binding domain (e.g., the constant region of an antibody) , which has human FcRn binding activity under neutral pH conditions. Perhaps the reason for this is the following.

When an antibody that binds a membrane antigen, for example, is introduced as an antigen-binding molecule into the body, the antibody associates with the antigen and is captured, along with the antigen (while maintaining its antigen-binding state), into an intracellular endosome by internalization. The antibody then migrates to the lysosome while maintaining its antigen-binding state, and the antigen-antibody complex is degraded by the lysosome. Internalization-mediated disappearance from plasma is called antigen-dependent disappearance and has been described for a variety of antibody molecules (Drug Discov Today (2006) 11 (1-2), 81-88). After one IgG antibody molecule binds divalently to antigens, that one antibody molecule in a state bound to antigen molecules is internalized and degraded in that state in the lysosome. Thus, one molecule of a conventional IgG antibody cannot bind to three or more antigen molecules. For example, one IgG antibody molecule with neutralizing activity cannot neutralize three or more antigen molecules.

The relatively long retention time of IgG molecules in plasma (slow disappearance) is characteristic of the functions of human FcRn, known as the rescue receptor for IgG molecules. IgG molecules taken up into the endosome via pinocytosis bind to human FcRn expressed in the endosome under the acidic conditions of the endosome. Then the IgG molecules that do not bind to human FcRn migrate to the lysosome and are degraded there. On the other hand, human FcRn-bound IgG molecules migrate to the cell surface. IgG molecules dissociate from human FcRn under neutral conditions in plasma and are thus recycled into plasma.

When the antigen-binding molecule is a soluble antigen-binding antibody, the antibody introduced into the body associates with the antigen and is then taken up into cells while maintaining its antigen-bound state. Most antibodies captured in this way into cells associate with FcRn in the endosome and then migrate to the cell surface. These antibodies dissociate from human FcRn under neutral conditions in plasma and are thus released from cells. However, an antibody containing a conventional antigen-binding domain, the antigen-binding activity of which does not change depending on ion concentration conditions (eg, pH), is released from cells while maintaining its antigen-bound state. Thus, this antibody cannot re-associate with the antigen. Thus, as with an antibody that binds a membrane antigen, one molecule of a conventional IgG antibody, whose antigen-binding activity does not change depending on ion concentration conditions (eg, pH), cannot bind to three or more antigen molecules.

An antibody with pH-dependent antigen binding that is strongly associated with the antigen under neutral pH conditions in the plasma and dissociates from the antigen under acidic pH conditions in the endosome (an antibody capable of binding to the antigen under neutral pH conditions and dissociating from the antigen under acidic conditions pH values) or an antibody with calcium ion concentration-dependent binding, which is strongly associated with the antigen under conditions of high concentration of calcium ions in the plasma and dissociates from the antigen under conditions of low concentration of calcium ions in the endosome (antibody capable of binding to the antigen under conditions of high ion concentration calcium and be separated from the antigen under conditions of low calcium ion concentration) can dissociate from the antigen in the endosome. An antibody with pH-dependent antigen binding or an antibody with calcium ion concentration-dependent antigen binding, thus dissociated from the antigen, can re-associate with the antigen after recycling into plasma by FcRn. This allows one antibody molecule to bind to many antigen molecules multiple times. Also, the antigen bound to the antigen-binding molecule dissociates from the antibody in the endosome and is thus degraded in the lysosome without recycling into the plasma. Introduction of such an antigen-binding molecule into the body can accelerate the cellular uptake of antigens and reduce the concentration of antigen in the plasma.

The ability to bind to human FcRn under neutral pH conditions (pH 7.4) can be conferred on an antibody with pH-dependent antigen binding that is strongly associated with the antigen under neutral pH conditions in plasma and dissociates from the antigen under acidic pH conditions in the endosome ( an antibody capable of binding to an antigen under neutral pH conditions and dissociating from the antigen under acidic pH conditions), or an antibody with calcium ion concentration-dependent antigen binding that is strongly associated with the antigen under conditions of high plasma calcium ion concentration and dissociates from the antigen under conditions of low concentration of calcium ions in the endosome (an antibody that can bind to the antigen under conditions of high concentration of calcium ions and separate from the antigen under conditions of low concentration of calcium ions). The resulting antigen-binding molecule further accelerates the cellular uptake of antigens bound to it. Introduction of such an antigen-binding molecule into the body can accelerate the disappearance of the antigen and reduce the concentration of the antigen in the plasma. A conventional antibody, which has neither pH-dependent antigen binding nor calcium-dependent antigen binding, and its antibody-antigen complex are taken up into cells through nonspecific endocytosis and transported to the cell surface by binding to FcRn under acidic conditions. pH in the endosome. The antibody dissociates from FcRn under neutral pH conditions at the cell surface and is thus recycled into the plasma. When an antibody that binds to an antigen in a sufficiently pH-dependent manner (i.e., which associates with the antigen under neutral pH conditions and dissociates from it under acidic pH conditions) or in a sufficiently calcium ion concentration-dependent manner (i.e., which associates with the antigen under conditions of high concentration of calcium ions and dissociates from it under conditions of low concentration of calcium ions), associates with the antigen in the plasma and dissociates from the associated antigen in the endosome, the rate of disappearance of the antigen is considered equal to the rate of cellular uptake of the antibody and its antibody-antibody complex antigen through nonspecific endocytosis. If the binding of an antibody to an antigen is insufficiently dependent on pH or its dependence on the concentration of calcium ions, antigens that have not dissociated in the endosome are recycled into the plasma along with the antibody. In contrast, if antigen binding is sufficiently pH dependent, the rate of antigen disappearance is determined by the rate of cellular uptake through nonspecific endocytosis. Because FcRn is responsible for the transport of antibodies from the endosome to the cell surface, it is believed that some FcRn receptors also exist on the cell surface.

Typically, IgG immunoglobulin, which is a form of antigen-binding molecule, has little or no FcRn binding activity at neutral pH. The present inventors hypothesized that an IgG immunoglobulin having FcRn binding activity at neutral pH is capable of binding to FcRn present on the cell surface and being taken up into cells in an FcRn dependent manner by binding to FcRn present on the cell surface. The rate of FcRn-mediated cellular uptake is faster than the rate of cellular uptake through nonspecific endocytosis. This indicates that imparting the ability to bind to FcRn at neutral pH may further accelerate the rate of antigen disappearance by the antigen-binding molecule. In particular, an antigen-binding molecule, which has the ability to bind to FcRn at neutral pH, delivers the antigen into cells faster than conventional (natural human) IgG immunoglobulin and dissociates from the antigen in the endosome. The dissociated antigen-binding molecule is recycled to the cell surface or plasma, where the molecule re-associates with the antigen, again leading to FcRn-mediated cellular uptake. The higher ability to bind to FcRn at neutral pH may increase the turnover rate of this cycle and thus accelerate the rate of disappearance of the antigen from the plasma. The antigen-binding activity of the antigen-binding molecule at acidic pH values can be reduced relative to the antigen-binding activity at neutral pH values, thereby further increasing the rate of disappearance of the antigen from the plasma. The increased turnover rate of this cycle may result in an increased number of cycles, which is likely to allow one antigen binding molecule to contact a higher number of antigen molecules. The antigen binding molecule of the present invention may comprise an antigen binding domain and an FcRn binding domain. The FcRn binding domain neither influences antigen binding nor is it dependent on the type of antigen, as is also clear from the mechanism mentioned above. The antigen binding activity (binding ability) of an antigen binding molecule under conditions of low ion concentration, such as conditions of acidic pH or low concentration of calcium ions, is reduced relative to the antigen binding activity under conditions of concentration of ions, such as conditions of neutral pH or high concentration of calcium ions, and/ or the FcRn binding activity of this molecule is enhanced at plasma pH. This can accelerate the cellular uptake of antigens by the antigen-binding molecule and accelerate the rate of disappearance of antigens. Thus, the antigen-binding molecule of the present invention can be expected to exhibit improved effects compared to conventional therapeutic antibodies, for example, in reducing adverse reactions caused by antigens, increasing the dose of the antibody, and improving the kinetics of the antibody in vivo.

In particular, the present invention relates to a pharmaceutical composition containing an antigen binding molecule of the present invention, an antigen binding molecule isolated by the screening method of the present invention, or an antigen binding molecule produced by the production method of the present invention. The antigen-binding molecule of the present invention or the antigen-binding molecule obtained by the production method of the present invention is suitable as a pharmaceutical composition because its administration provides a high effect of reducing the concentration of antigen in plasma compared with conventional antigen-binding molecules. The pharmaceutical composition of the present invention may contain a pharmaceutically acceptable carrier.

In the context of the present invention, a pharmaceutical composition generally relates to an agent for the treatment or prevention of a disease, or examination, or diagnosis.

The pharmaceutical composition of the present invention can be formulated using a method generally known to those skilled in the art. For example, the pharmaceutical composition of the present invention can be used parenterally in the form of an injection in a sterile solution or suspension with water or any other pharmaceutically acceptable solution. For example, the active ingredient may be suitably combined with pharmacologically acceptable carriers or vehicles, in particular sterile water or saline, vegetable oil, emulsifier, suspending agent, surfactant, stabilizer, flavoring agent, excipient, carrier, antiseptic, binder , and so on. and mixed with them in unit dosage form required in accordance with generally accepted pharmaceutical practice to obtain the drug. The amount of active ingredient in these preparations is selected to administer a suitable volume within the prescribed range.

Sterile injectable compositions can be formulated in accordance with standard pharmaceutical practice using a carrier such as distilled water for injection. Examples of injectable aqueous solutions include saline and isotonic solutions containing glucose or other adjuvants (eg, D-sorbitol, D-mannose, D-mannitol, and sodium chloride). In combination with them, a suitable solubilizer can be used, for example, an alcohol (ethanol, etc.), a polyalcohol (propylene glycol, polyethylene glycol, etc.), or a nonionic surfactant (polysorbate 80(TM), HCO-50 , etc.).

Examples of oil solutions include sesame oil and soybean oil. In combination, benzyl benzoate and/or benzyl alcohol can be used as solubilizers. These injectable solutions can be mixed with a buffer (eg, phosphate-buffered saline and sodium acetate-buffered solution), a demulcent (eg, procaine hydrochloride), a stabilizer (eg, benzyl alcohol and phenol), and an antioxidant. The resulting injections are usually filled into suitable ampoules.

The pharmaceutical composition of the present invention is preferably administered by the parenteral route. For example, the composition is administered in an injection, transnasal, transpulmonary, or transdermal dosage form. The composition can be administered systemically or locally, for example, by intravenous injection, intramuscular injection, intraperitoneal injection or subcutaneous injection.

The route of administration can be suitably selected depending on the age and symptoms of the patient. The unit dose of the pharmaceutical composition containing the antigen binding molecule may be selected from the range, for example, from 0.0001 mg to 1000 mg per kg of body weight. Alternatively, the dose may be selected from a range of, for example, 0.001 to 100,000 mg per patient, although the dose of the present invention is not necessarily limited to these numbers. The dose and route of administration vary depending on body weight, age, symptoms, etc. patient. Those skilled in the art can select the appropriate dosage and route of administration based on these conditions.

Amino acids contained in the amino acid sequences described within the scope of the present invention may undergo post-translational modification (eg, modification of N-terminal glutamine to pyroglutamic acid by pyroglutamylation, which is well known to those skilled in the art). Even such forms having post-translationally modified amino acids are naturally included in the amino acid sequences described within the scope of the present invention.

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

As used herein, the aspect represented by the term "comprising" covers the aspect represented by the term "essentially consisting of" and the aspect represented by the term "consisting of".

Examples

In the following, the present invention is described more specifically with the help of examples. However, the present invention is not limited to these examples.

[Example 1] Construction of an Antibody Library with pH-Dependent Binding

(1-1) Method for producing an antibody with pH-dependent binding

WO 2009125825 describes that a histidine is introduced into an antigen-binding molecule to produce an antibody with pH-dependent antigen binding, the property of which varies between the neutral pH and acidic pH regions. The disclosed pH-dependent binding antibody is prepared by a modification that replaces part of the amino acid sequence of the desired antigen-binding molecule with histidine. A possible method for more efficiently producing an antibody with pH-dependent binding without first obtaining an antigen-binding molecule to be modified involves introducing histidine into the variable regions (more preferably, positions that may be involved in antigen binding) and obtaining an antigen-binding molecule that binds to the desired antigen, from the resulting collection of antigen-binding molecules (called His library). The antigen binding molecule obtained from the His library has a higher frequency of histidine than conventional antibody libraries, indicating that antigen binding molecules having the desired property can be efficiently produced.

(1-2) Construction of a collection (His library) of antibody molecules containing histidine residues in variable regions that allow efficient production of a binding antibody that binds to an antigen in a pH-dependent manner

First, histidine insertion positions were selected for the His library. WO 2009125825 describes that antibodies with pH-dependent antigen binding were obtained by replacing amino acid residues in the sequences of anti-IL-6 receptor antibody, anti-IL-6 antibody and anti-IL-31 receptor antibody with histidine. In addition, an antibody against egg white lysozyme (FEBS Letter 11483, 309, 1, 85-88) and an antibody against hepcidin (WO 2009139822), which have the ability to bind antigen in a pH-dependent manner, were obtained by replacing amino acids in the amino acid sequences of antigen-binding molecules with histidine Table 3 shows the positions of histidine in anti-IL-6 receptor antibody, anti-IL-6 antibody, anti-IL-31 receptor antibody, anti-egg white lysozyme antibody and anti-hepcidin antibody. The positions presented in Table 3 may serve as candidate positions that can control antigen-antibody binding. In addition to the positions presented in Table 3, positions that are highly accessible to the antigen were also considered suitable as histidine insertion positions.

In the His library constructed using heavy and light chain variable regions, the heavy chain variable regions used were human antibody sequences, while a histidine residue was introduced into the light chain variable regions. The positions given above and the positions that may be involved in antigen binding, i.e. positions 30, 32, 50, 53, 91, 92 and 93 (defined by Kabat numbering; Kabat E.A. et al., 1991. Sequence of Proteins of Immunological Interest. NIH) in the light chains were selected as histidine contribution positions for the His library. Also, the Vk1 sequence was selected as the light chain variable region template sequence for introducing histidine. Multiple amino acids were allowed to appear at any given position in the template sequence to increase the diversity of antigen-binding molecules composing the library. Surface-exposed positions in the variable regions that are likely to interact with antigens were selected as the position at which a plurality of amino acids appeared. In particular, positions 30, 31, 32, 34, 50, 53, 91, 92, 93, 94 and 96 were chosen as such non-strict residues (defined by Kabat numbering; Kabat E.A. et al. 1991. Sequences of Proteins of Immunological Interest. NIH) in light chains.

Next, the types and occurrence of appearing amino acid residues were selected. Amino acids of non-strict residues were analyzed with respect to their frequency of occurrence in the hVk1 and hVk3 sequences registered in the Kabat database (KABAT, E.A. et al.: "Sequences of Proteins of Immunological interest", vol. 91, 1991, NIH PUBLICATION). Based on the analysis results, the types of amino acids appearing in the His library were selected from amino acids with a high frequency of occurrence at each position. This technique also selected amino acids that were confirmed to have a low frequency of occurrence based on the analysis results to prevent imbalance of amino acid properties. The frequency of occurrence of the selected amino acids was selected based on the results of the analysis of the Kabat database.

Taking into account the amino acids chosen in this way and their frequency of occurrence, two His libraries were constructed: His library 1, fixed to the condition that each CDR contained one histidine residue without exception; and His library 2, which was more focused on sequence diversity than His library 1. The detailed construction of His library 1 and His library 2 is presented in Tables 4 and 5 (“Position” in each table reflects Kabat numbering). The frequency of occurrence of the amino acids described in Tables 4 and 5 may exclude Ser(S) at position 94, determined by Kabat numbering, in the case of Asn(N) at position 92.

[Example 2] Preparation of a human antibody phage display library (His library 1) to produce an antibody that binds to an antigen in a pH-dependent manner

A gene library of antibody heavy chain variable regions was amplified by PCR using poly-A RNA obtained from human PBMC, commercially available human poly-A RNA, and the like. as a matrix. An antibody light chain variable region gene library constructed as His library 1 described in Example 1 was amplified using PCR. Combinations of sequences in the antibody heavy chain variable region gene library and the antibody light chain variable region gene library thus obtained were inserted into phagemid vectors to construct a human antibody phage display library displaying Fab domains consisting of human antibody sequences. The construction method was performed in accordance with Methods Mol Biol. (2002) 178, 87-100. In constructing this library, we used a phage display library sequence that contained a linker portion for binding the phagemid Fab and phage pIII protein, and a trypsin cleavage sequence inserted between the N2 and CT domains of the helper phage pIII protein. Portions of antibody genes isolated from E. coli transformed with the antibody gene library were sequenced, yielding sequence information for approximately 132 clones. In fig. Figure 1 shows the model distribution of amino acids and the distribution of amino acids in the confirmed sequence. A library containing a variety of sequences corresponding to the model amino acid distribution was constructed.

[Example 3] Preparation of an antibody that binds to IL-6R in a pH-dependent manner

(3-1) Obtaining an antibody fragment that binds an antigen in a pH-dependent manner from a library using bead panning

The first round of screening of the constructed His library 1 was carried out by increasing the content of only antibody fragments with the ability to bind antigen (IL-6R).

Phages were obtained from E. coli containing an engineered phagemid for phage display. 2.5 M NaCl/10% PEG was added to E. coli cultures that produced phages, and the pool of phages thus precipitated was diluted with TBS to obtain a phage library solution. BSA and CaCl ₂ were added to the solution with the phage library (final concentration: 4% BSA and 1.2 mM calcium ions). The panning method was carried out in accordance with the conventional panning method using antigens immobilized on magnetic beads (J. Immunol. Methods. (2008) 332 (1-2), 2-9; J. Immunol. Methods. (2001) 247 (1 -2), 191-203; Biotechnol. Prog. (2002) 18 (2) 212-20; and Mol. Cell Proteomics (2003) 2 (2), 61-9). The magnetic beads used were neutravidin-coated beads (Sera-Mag SpeedBeads NeutrAvidin-coated) or streptavidin-coated beads (Dynabeads M-280 Streptavidin).

Specifically, 250 pmol of biotin-labeled antigens were added to the resulting phage library solution and thereby brought into contact with the phage library solution at room temperature for 60 minutes. BSA-blocked magnetic beads were added, and then the antigen-phage complexes were bound to the magnetic beads at room temperature for 15 minutes. The beads were washed three times with 1 ml of 1.2 mM CaCl ₂ /TBS (TBS containing 1.2 mM CaCl ₂ and 0.1% Tween 20), and then washed additionally two times with 1 ml of 1.2 mM CaCl ₂ /TBS ( pH 7.6). After adding 0.5 ml of 1 mg/ml trypsin, the beads were suspended at room temperature for 15 minutes, and immediately thereafter the beads were separated using a magnetic stand to obtain a phage solution. The resulting solution with phages was added to 10 ml of E. coli strain ER2738 at the logarithmic growth phase (OD600: 0.4-0.7). The E. coli strain was infected with phages by gently rotating the strain culture at 37°C for 1 hour. Infected E. coli were inoculated onto a 225 mm×225 mm plate. Next, phages were collected from the inoculated E. coli cultures to obtain a solution with a phage library.

In the second and subsequent rounds of panning, the content of phages with antigen-binding ability or pH-dependent binding ability as an indicator was increased. Specifically, 40 pmol of biotin-labeled antigens were added to the resulting phage library solution and thereby brought into contact with the phage library solution at room temperature for 60 minutes. BSA-blocked magnetic beads were added, and then the antigen-phage complexes were bound to the magnetic beads at room temperature for 15 minutes. The beads were washed several times with 1 ml of 1.2 mM CaCl ₂ /TBST and 1.2 mM CaCl ₂ /TBS. To increase the content with antigen binding capacity as an indicator, beads supplemented with 0.5 ml of 1 mg/ml trypsin were suspended at room temperature for 15 minutes, and immediately thereafter the beads were separated using a magnetic stand to obtain a phage solution. To increase the content with pH-dependent antigen binding capacity as an indicator, beads supplemented with 0.1 ml of 50 mM MES/1.2 mM CaCl ₂ /150 mM NaCl (pH 5.5) were suspended at room temperature. Immediately after this, the granules were separated using a magnetic stand to obtain a new solution with phages. Addition of 5 μl of 100 mg/ml trypsin to the resulting phage solution digested the pIII proteins (helper phage-derived pIII proteins) of non-Fab phages to eliminate the ability of non-Fab phages to infect E. coli. The resulting phages were added to 10 ml of E. coli strain ER2738 at logarithmic growth phase (OD600: 0.4-0.7). The E. coli strain was infected with phages by gently rotating the strain culture at 37°C for 1 hour. Infected E. coli were inoculated onto a 225 mm×225 mm plate. Next, phages were collected from the inoculated E. coli cultures to obtain a solution with a phage library. This panning with antigen-binding ability or pH-dependent binding ability as indicators was performed on everyone for 2 rounds.

(3-2) Evaluation by phage ELISA

Phage-containing culture supernatant was prepared according to a conventional method (Methods Mol. Biol. (2002) 178, 133-145) from each single colony of E. coli obtained using the method described above.

After addition of BSA and CaCl ₂ (final concentration: 4% BSA and 1.2 mM calcium ions), the phage-containing culture supernatant was subjected to ELISA using the following procedure: a StreptaWell 96 microtiter plate (F. Hoffmann-La Roche Ltd.) was covered for nights with 100 μl of PBS containing biotin-labeled antigens. Each well of the plate was washed with PBST (PBS containing 0.1% Tween 20) to remove unbound antigens. The well was then blocked with 250 μl of 4% BSA-TBS for 1 hour or more. After removal of 4% BSA-TBS, the resulting supernatant was added to each well and the plate was allowed to stand at 37°C for 1 hour to allow phage-exposed antibodies to bind to the antigens contained in each well. Each well was washed with 1.2 mM CaCl ₂ /TBST, and 1.2 mM CaCl ₂ /TBS (pH 7.6) or 1.2 mM CaCl ₂ /TBS (pH 5.5) was added. The plates were left to stand at 37°C for 30 minutes for incubation. After washing with 1.2 mM CaCl ₂ /TBST, HRP-conjugated anti-M13 antibodies (Amersham Pharmacia Biotech Inc.) diluted in TBS containing 4% BSA and 1.2 mM ionized calcium were added to each well. The plate was incubated for 1 hour. After washing with 1.2 mM CaCl ₂ /TBST, plain TMB solution (ZYMED Laboratories, Inc.) was added to each well. The chromogenic reaction of the solution in each well was completed by adding sulfuric acid. Color development was then analyzed based on absorbance at 450 nm.

17 clones were ELISA-positive in an antigen-specific manner as a result of phage ELISA for 96 clones (after 2 rounds of panning) that were enriched with antigen-binding capacity as an indicator. Thus, samples obtained in 3 rounds of penning were analyzed. On the other hand, 70 clones were ELISA-positive by phage ELISA for 94 clones (after 2 rounds of panning) that were enriched with pH-dependent antigen binding capacity as an indicator. Thus, these samples obtained in 2 rounds of panning were analyzed.

The genes of the phage ELISA clones were amplified using specific primers and then analyzed for their nucleotide sequences.

The results of the phage ELISA and sequence analysis are presented in Table 6 below.

Antibodies capable of pH-dependent antigen binding were obtained in a similar manner from a naïve human antibody phage display library. 13 types of antibodies with pH-dependent binding were obtained by evaluating 88 clones that were enriched with antigen-binding ability as an indicator. Also, 27 types of pH-dependent binding antibodies were obtained by evaluating 83 clones that were enriched with pH-dependent antigen binding ability as an indicator.

These results demonstrated that His library 1 produced more variant clones with pH-dependent antigen binding capabilities than the naïve human antibody phage display library.

(3-3) Antibody expression and purification

The gene from each clone assessed to have pH-dependent antigen binding by phage ELISA was inserted into plasmids for expression in animal cells. Antibody expression was performed as follows: Human embryonic kidney-derived Freestyle 293-F line (Invitrogen Corp.) was suspended in Freestyle 293 Expression Medium (Invitrogen Corp.). The suspension having a cell density of 1.33×10 ⁶ cells/ml was inoculated at a concentration of 3 ml/well into a 6-well plate. The resulting plasmids were transferred into cells by lipofection. Cells were cultured for 4 days in a CO ₂ incubator (37°C, 8% CO ₂ , 90 rpm). Antibodies were purified from the resulting culture supernatant in a manner well known to those skilled in the art using rProtein A Sepharose(TM) Fast Flow (Amersham Biosciences, Inc.). The absorbance of the purified antibody solution was measured at 280 nm using a spectrophotometer. The antibody concentration was calculated from the resulting measurement using the extinction coefficient calculated by PACE (Protein Science (1995) 4, 2411-2423).

(3-4) Evaluation of the resulting antibody for its ability to pH-dependently bind to the human IL-6 receptor

To evaluate the pH dependence of the human IL-6 receptor binding activity of the antibodies 6RpH#01 (heavy chain: SEQ ID NO: 18 and light chain: SEQ ID NO: 19), 6RpH#02 (heavy chain: SEQ ID NO: 20 and light chain: SEQ ID NO: 21), and 6RpH#03 (heavy chain: SEQ ID NO: 22 and light chain: SEQ ID NO: 23) obtained in step (3-3), these antibodies were analyzed for their interaction with human IL-6 receptors using Biacore T100 (GE Healthcare Bio-Sciences Corp.). Tocilizumab (heavy chain: SEQ ID NO: 24 and light chain: SEQ ID NO: 25) was used as a control antibody lacking pH-dependent binding activity to human IL-6 receptors. Interactions in the antigen-antibody reaction were analyzed in solutions with pH 7.4 and pH 6.0 as neutral pH and acidic pH conditions, respectively. Approximately 300 RU of each antibody of interest was captured on a Protein A/G (Invitrogen Corp.) immobilized CM5 sensor chip (GE Healthcare Bio-Sciences Corp.), in the appropriate amount by the amine coupling method. Two types of buffer solutions were used as running buffers: 20 mM ACES, 150 mM NaCl, 0.05% (w/v) Tween 20, and 1.2 mM CaCl ₂ (pH 7.4); and 20 mM ACES, 150 mM NaCl, 0.05% (w/v) Tween 20, and 1.2 mM CaCl ₂ (pH 6.0). Human IL-6 receptors were also diluted using each of these buffers. All analysis was performed at 37°C.

In an antigen-antibody interaction assay using tocilizumab control antibody, 6RpH#01 antibody, 6RpH#02 antibody, and 6RpH#03 antibody, diluted IL-6 receptor solution or empty running buffer was injected at a flow rate of 5 μL/min for 3 minutes for IL-6 receptors to interact with the tocilizumab antibody, 6RpH#01 antibody, 6RpH#02 antibody, or 6RpH#03 antibody captured on the sensor chip. It was then injected with 10 mM glycine-HCl (pH 1.5) at a flow rate of 30 μL/min for 30 seconds to regenerate the sensor chip.

In fig. Figure 2 shows sensorgrams of antibodies analyzed at pH 7.4 using the method described above. In fig. Figure 3 shows sensograms of antibodies under conditions of pH 6.0, obtained in a similar way.

As a result, it was found that the IL6 receptor binding ability of the 6RpH#01 antibody, the 6RpH#02 antibody, and the 6RpH#03 antibody was significantly reduced when the pH of the buffer was changed from pH 7.4 to pH 6.0.

[Example 4] Preparation of a human antibody phage display library (His library 2) to produce an antibody that binds to an antigen in a pH-dependent manner

A gene library of antibody heavy chain variable regions was amplified by PCR using poly-A RNA obtained from human PBMC, commercially available human poly-A RNA, and the like. as a matrix. To increase the frequency of occurrence of antibodies having the ability to pH-dependent antigen binding, as described in example 1, parts of the light chains with variable regions of the antibody are designed so that in these parts of the light chains the frequency of occurrence of histidine residues in areas likely to act as a reduction site increases into contact with the antigen. The light chain variable region library is contacted such that highly abundant amino acids determined from amino acid frequency information in natural human antibodies are evenly distributed as amino acid residues other than the histidine-embedded residues among the non-stringent residues. The library can be obtained by engaging a commercial authorized company for its synthesis, etc. The sequence combinations of the antibody heavy chain variable region gene library and the antibody light chain variable region gene library thus obtained are inserted into phagemid vectors. A human antibody phage display library displaying Fab domains consisting of human antibody sequences is constructed according to a method known in the art (Methods Mol Biol. (2002) 178, 87-100). Portions of antibody genes isolated from E. coli transformed with the antibody gene library were sequenced according to the method described in Example 2.

[Example 5] Preparation of an antibody that binds to IL-6R in a pH-dependent manner

(5-1) Obtaining an antibody fragment that binds an antigen in a pH-dependent manner from a library using bead panning

The first round of screening of the constructed His library 2 is carried out by increasing the content of only antibody fragments with the ability to bind antigen (IL-6R).

Phages are produced from E. coli containing an engineered phagemid for phage display. 2.5 M NaCl/10% PEG is added to E. coli cultures that produce phages, and the aggregate of phages thus precipitated is diluted with TBS to obtain a phage library solution. BSA and skim milk as a blocking agent are then added to the phage library solution. The panning method is carried out in accordance with the generally accepted panning method using antigens immobilized on magnetic beads (J. Immunol. Methods. (2008) 332 (1-2), 2-9; J. Immunol. Methods. (2001) 247 (1 -2), 191-203; Biotechnol. Prog. (2002) 18 (2) 212-20; and Mol. Cell Proteomics (2003) 2 (2), 61-9). The magnetic beads used are neutravidin-coated beads (Sera-Mag SpeedBeads NeutrAvidin-coated) or streptavidin-coated beads (Dynabeads M-280 Streptavidin).

Specifically, 250 pmol of biotin-labeled antigens are added to the resulting phage library solution and thereby brought into contact with the phage library solution at room temperature for 60 minutes. Magnetic beads blocked with a blocking agent are added, and then the antigen-phage complexes are bound to the magnetic beads at room temperature for 15 minutes. The beads are washed three times with 1 ml TBST and then washed an additional two times with 1 ml TBS. After adding magnetic beads blocked with a blocking agent, the antigen-phage complexes are bound to the magnetic beads at room temperature for 15 minutes. The beads are washed three times with 1 ml TBST and then washed an additional two times with 1 ml TBS. After adding 0.5 ml of 1 mg/ml trypsin, the beads are suspended at room temperature for 15 minutes, and immediately thereafter the beads are separated using a magnetic stand to obtain a solution with phages. The collected phage solution is added to 10 ml of E. coli strain ER2738 at logarithmic growth phase (OD600: 0.4-0.7). The E. coli strain is infected with phages by gently rotating the strain culture at 37°C for 1 hour. Infected E. coli are inoculated onto a 225 mm×225 mm plate. Next, phages are collected from the inoculated E. coli cultures to obtain a solution with a phage library.

In the second and subsequent rounds of panning, the content of phages with antigen-binding ability or pH-dependent binding ability as an indicator is increased. Specifically, 40 pmol of biotin-labeled antigens are added to the resulting phage library solution and thereby brought into contact with the phage library solution at room temperature for 60 minutes. BSA or skim milk blocked magnetic beads are added, and then the antigen-phage complexes are bound to the magnetic beads at room temperature for 15 minutes. The beads are washed with 1 ml TBST and TBS. To increase the content with antigen binding capacity as an indicator, beads supplemented with 0.5 ml of 1 mg/ml trypsin are suspended at room temperature for 15 minutes, and immediately thereafter the beads are separated using a magnetic stand to obtain a phage solution. To increase the content with pH-dependent antigen binding capacity as an indicator, beads supplemented with 0.1 ml of 50 mM MES/1.2 mM CaCl ₂ /150 mM NaCl (pH 5.5) were suspended at room temperature. Immediately after this, the granules are separated using a magnetic stand to obtain a solution with phages. Adding 5 μl of 100 mg/ml trypsin to the collected phage solution cleaves the pIII proteins (helper phage-derived pIII proteins) of non-Fab phages to eliminate the ability of non-Fab phages to infect E. coli. The collected phages are added to 10 ml of E. coli strain ER2738 at logarithmic growth phase (OD600: 0.4-0.7). The E. coli strain is infected with phages by gently rotating the strain culture at 37°C for 1 hour. Infected E. coli are inoculated onto a 225 mm x 22 5 mm plate. Next, phages are collected from the inoculated E. coli cultures to obtain a solution with a phage library. This panning with antigen-binding ability or pH-dependent binding ability as indicators is repeated several times.

(5-2) Evaluation by phage ELISA

A phage-containing culture supernatant is prepared according to a conventional method (Methods Mol. Biol. (2002) 178, 133-145) from each single colony of E. coli obtained using the method described above.

After addition of BSA and CaCl ₂ , the phage-containing culture supernatant was subjected to ELISA using the following procedure: a StreptaWell 96 microtiter plate (F. Hoffmann-La Roche Ltd.) was coated overnight with 100 μl of PBS containing biotin-labeled antigens. Each well of the plate is washed with PBST to remove unbound antigens. The well is then blocked with 250 μl of 4% BSA-TBS for 1 hour or more. After removing 4% BSA-TBS, the resulting supernatant was added to each well and the plate was allowed to stand at 37°C for 1 hour to allow phage-exposed antibodies to bind to the antigens contained in each well. Each well is washed with 1.2 mM CaCl ₂ /TBST, and 1.2 mM CaCl ₂ /TBS (pH 7.6) or 1.2 mM CaCl ₂ /TBS (pH 5.5) is added. The plate is left to stand at 37°C for 30 minutes to incubate. After washing with 1.2 mM CaCl ₂ /TBST, HRP-conjugated anti-M13 antibody (Amersham Pharmacia Biotech Inc.) diluted in TBS containing 4% BSA and 1.2 mM ionized calcium was added to each well. The plate is incubated for 1 hour. After washing with 1.2 mM CaCl ₂ /TBST, plain TMB solution (ZYMED Laboratories, Inc.) was added to each well. The chromogenic reaction of the solution in each well is completed by adding sulfuric acid. Color development is then analyzed based on absorbance at 450 nm.

The genes of antibody fragments assessed to have pH-dependent antigen binding capacity by phage ELISA are amplified as a template using specific primers and then analyzed for their nucleotide sequences.

(5-3) Antibody expression and purification

The gene from each clone assessed to have pH-dependent antigen binding by phage ELISA is inserted into plasmids for expression in animal cells. Antibody expression is carried out as follows: the human embryonic kidney-derived Freestyle 293-F line (Invitrogen Corp.) is suspended in Freestyle 293 Expression Medium (Invitrogen Corp.). The suspension having a cell density of 1.33×10 ⁶ cells/ml is inoculated at a concentration of 3 ml/well into a 6-well plate. The resulting plasmids are transferred into cells by lipofection. Cells are cultured for 4 days in a CO ₂ incubator (37°C, 8% CO ₂ , 90 rpm). Antibodies are purified from the resulting culture supernatant in a manner well known to those skilled in the art using rProtein A Sepharose(TM) Fast Flow (Amersham Biosciences, Inc.). The absorbance of the purified antibody solution is measured at 280 nm using a spectrophotometer. The antibody concentration is calculated from the resulting measurement using the extinction coefficient calculated by PACE (Protein Science (1995) 4, 2411-2423).

(5-4) Evaluation of the resulting antibody for its ability to pH-dependently bind to the human IL-6 receptor

To evaluate the pH dependence of the human IL-6 receptor binding activity of the antibodies prepared in Example 5, the antibodies were assayed for their interaction with human IL-6 receptors using Biacore T100 (GE Healthcare Bio-Sciences Corp.). Tocilizumab (heavy chain: SEQ ID NO: 24 and light chain: SEQ ID NO: 25) was used as a control antibody lacking pH-dependent binding activity to human IL-6 receptors. The interaction in the antigen-antibody reaction was analyzed in solutions with pH 7.4 and pH 6.0 as neutral pH and acidic pH conditions, respectively. Each antibody of interest is captured on a CM5 sensor chip (GE Healthcare Bio-Sciences Corp.) immobilized with Protein A/G (Invitrogen Corp.), in the appropriate amount by amine coupling. Two types of buffer solutions are used as running buffers: 20 mM ACES, 150 mM NaCl, 0.05% (w/v) Tween 20, and 1.2 mM CaCl ₂ (pH 7.4); and 20 mM ACES, 150 mM NaCl, 0.05% (w/v) Tween 20, and 1.2 mM CaCl ₂ (pH 6.0). Human IL-6 receptors are also diluted using each of these buffers. All analysis is carried out at 37°C.

In an antigen-antibody interaction assay using the control antibody tocilizumab and the antibodies prepared in Example 5, a diluted IL-6 receptor solution or empty running buffer is injected to allow the IL-6 receptors to interact with the antibodies immobilized on the sensor chip. It is then injected with 10 mM glycine-HCl (pH 1.5) at a flow rate of 30 μL/min for 30 seconds to regenerate the sensor chip. Sensograms under conditions of pH 6.0 are also obtained in a similar way.

[Example 6] Search for human germline sequence binding to calcium ion

(6-1) Antibody that binds to antigen in a calcium-dependent manner

An antibody that binds to an antigen in a calcium-dependent manner (calcium-dependent antigen binding antibody) is an antibody whose interaction with an antigen varies depending on the concentration of calcium ions. Because calcium-dependent antigen binding antibody is thought to bind to antigen via calcium ions, the amino acids that constitute antigenic epitopes are negatively charged amino acids capable of chelating calcium ions or amino acids that can serve as hydrogen bond acceptors (Figure 4B ). Due to the properties of such epitope-constituting amino acids, a calcium-dependent antigen-binding antibody is able to target an epitope different from that of the pH-dependent antigen-binding molecule produced by insertion of histidine residues, as shown in FIG. 4A. In addition, the use of antigen-binding molecules having the property of binding to antigens in both a calcium-dependent and pH-dependent manner, as shown in FIG. 4C can provide antigen binding molecules capable of individually targeting a variety of epitopes having a wide range of properties. This indicates that an antibody with calcium-dependent antigen binding can be efficiently produced by constructing a set of molecules containing calcium-binding motifs (Ca library) and obtaining antigen-binding molecules from the specified set of molecules.

(6-2) Obtaining the human germline sequence

A possible example of a set of molecules containing calcium-binding motifs is a set of antibodies as such molecules. In other words, a possible example of a Ca library is an antibody library containing calcium binding motifs.

None of the previously described antibodies containing human germline sequences bind to calcium ions. Thus, to determine whether antibodies containing human germline sequences bind to calcium ions, germline DNA sequences of antibodies containing human germline sequences were cloned using, as a template, cDNAs derived from fetal spleen poly-RNA human (Clontech Laboratories, Inc.). The cloned DNA fragments were inserted into expression vectors for animal cells. The resulting expression vectors were sequenced in a manner well known to those skilled in the art. SEQ ID NO of amino acid sequences encoded by specific nucleotide sequences are presented in Table 7. Polynucleotide encoding sequence SEQ ID NO: 6 (Vk1), SEQ ID NO: 7 (Vk2), SEQ ID NO: 8 (Vk3), SEQ ID NO : 9 (Vk4) or SEQ ID NO: 1 (Vk5) was linked by PCR to a polynucleotide encoding the natural kappa chain constant region (SEQ ID NO: 26). The resulting DNA fragments were individually inserted into vectors for expression in animal cells. Also, a heavy chain variable region polynucleotide encoding the sequence SEQ ID NO: 27 (Vk1), SEQ ID NO: 28 (Vk2), SEQ ID NO: 29 (Vk3) or SEQ ID NO: 30 (Vk4) was linked by PCR to the polynucleotide , encoding IgG1 lacking the 2 C-terminal amino acids of SEQ ID NO: 14. The resulting DNA fragments were individually inserted into vectors for expression in animal cells. The sequences of the resulting modified forms were confirmed in a manner well known to those skilled in the art. In this context, human Vk1 is also referred to as hVk1; Human Vk2 is also referred to as hVk2; Human Vk3 is also referred to as hVk3; and human Vk4 are also referred to as hVk4.

(6-3) Antibody expression and purification

Animal cell expression vectors containing the insertion of a DNA fragment encoding each of the 5 types of human germline sequences thus obtained were transfected into animal cells. Antibody expression was performed as follows: Human embryonic kidney-derived Freestyle 293-F line (Invitrogen Corp.) was suspended in Freestyle 293 Expression Medium (Invitrogen Corp.). The suspension having a cell density of 1.33×10 ⁶ cells/ml was inoculated at a concentration of 3 ml/well into a 6-well plate. The resulting plasmids were transferred into cells by lipofection. Cells were cultured for 4 days in a CO ₂ incubator (37°C, 8% CO ₂ , 90 rpm). Antibodies were purified from the resulting culture supernatant in a manner well known to those skilled in the art using rProtein A Sepharose(TM) Fast Flow (Amersham Biosciences, Inc.). The absorbance of the purified antibody solution was measured at 280 nm using a spectrophotometer. The antibody concentration was calculated from the resulting measurement using the extinction coefficient calculated by PACE (Protein Science (1995) 4, 2411-2423).

(6-4) Evaluation of an antibody containing a human germline sequence for its calcium ion binding activity

Purified antibodies were evaluated for their calcium ion binding activity. The average thermal denaturation temperature (Tm) was measured using differential scanning calorimetry (DSC) (MicroCal VP-Capillary DSC, MicroCal) as an indicator to evaluate the binding of calcium ions by antibodies. The average thermal denaturation temperature (Tm) serves as an indicator of stability and becomes higher when the protein is stabilized through calcium ion binding, compared to the average thermal denaturation temperature (Tm) of a non-calcium ion-bound protein (J. Biol. Chem. (2008) ) 283, 37, 25140-25149). To assess the binding activity of the antibody to calcium ions, the change in the Tm value of each antibody was assessed depending on the change in the concentration of calcium ions in the antibody solution. Purified antibodies were dialyzed (EasySEP, Tomy Seiko Co., Ltd.) against a solution containing 20 mM Tris-HCl, 150 mM NaCl and 2 mM CaCl ₂ (pH 7.4) or containing 20 mM Tris-HCl, 150 mM NaCl and 3 µM CaCl ₂ (pH 7.4) as external solution. Each antibody solution was adjusted to approximately 0.1 mg/ml using the dialysis solution and subjected to DSC analysis as a test substance at 20° C. to 115° C. with the temperature rise rate set at 240° C./h. Table 8 presents the average thermal denaturation temperature (Tm) for each antibody Fab domain, calculated from the resulting DSC denaturation curve.

As a result, the value of the Tm Fab domain of an antibody containing the sequence hVk1, hVk2, hVk3 or hVk4 did not change depending on the concentration of calcium ions in the solution containing the Fab domain. In contrast, the Tm value of the Fab domain of the antibody containing the hVk5 sequence varied depending on the calcium ion concentration in the antibody solution containing the Fab domain, indicating that the hVk5 sequence binds to calcium ions.

(6-5) Evaluation of the hVk5-2 sequence for calcium binding

In addition to Vk5-2 (SEQ ID NO: 1, fused with SEQ ID NO: 26), Vk5-2 variant 1 (SEQ ID NO: 2) and Vk5-2 variant 2 (SEQ ID NO: 3), classified as Vk5-2, according to example 6(6-2). These variants were also evaluated for their calcium binding properties. DNA fragments of Vk5-2, Vk5-2 variant 1 and Vk5-2 variant 2 were individually inserted into animal cell expression vectors. The resulting expression vectors were sequenced in a manner well known to those skilled in the art. Animal cells were cotransfected with each animal cell expression vector having a DNA fragment insert of Vk5-2, Vk5-2 variant 1 or Vk5-2 variant 2, and an animal cell expression vector having a DNA insert encoding a CIM_H heavy chain (SEQ ID NO : 4), for expression, in the manner described in example 6(6-3). The resulting antibodies were purified. Purified antibodies were evaluated for their calcium ion binding activity. Purified antibodies were dialyzed (EasySEP, Tomy Seiko Co., Ltd.) against a solution containing 20 mM Tris-HCl, 150 mM NaCl, and 2 mM CaCl ₂ (pH 7.5) or containing 20 mM Tris-HCl and 150 mM NaCl (pH 7.5) (the latter solution is listed as "calcium ion concentration": 0 mM in Table 9) as the external solution. Each antibody solution was adjusted to approximately 0.1 mg/ml with dialysis solution and subjected to DSC analysis as a test substance at 20° C. to 115° C. with the temperature rise rate set at 240° C./h. Table 9 shows the average thermal denaturation temperature (Tm) of each antibody Fab domain calculated from the resulting DSC denaturation curve.

As a result, the Tm value for the Fab domain of an antibody containing the Vk5-2 sequence, Vk5-2 variant 1, or Vk5-2 variant 2 varied depending on the calcium ion concentration in the Fab domain-containing antibody solution, indicating that an antibody having a sequence classified as Vk5-2 binds to calcium ions.

[Example 7] Human Vk5 (hVk5) sequence evaluation

(7-1) Sequence hVk5

In the Kabat database, only the hVk5-2 sequence is registered as the hVk5 sequence. In what follows, hVk5 and hVk5-2 will be used synonymously. WO 2010136598 describes that the relative abundance of the hVk5-2 sequence among germline sequences is 0.4%. Other reports also state that the relative abundance of the hVk5-2 sequence among germline sequences ranges from 0 to 0.06% (J. Mol. Biol. (2000) 296, 57-86; and Proc. Natl. Acad. Sci. (2009) 106, 48, 20216-20221). Because the hVk5-2 sequence has a low frequency of occurrence among germline sequences, as described above, obtaining calcium-binding antibodies from an antibody library consisting of human germline sequences or from B cells obtained by immunizing mice expressing human antibodies has proven to be ineffective. This may provide rationale for constructing a Ca library containing human hVk5-2 sequences. However, previously described synthetic antibody libraries (WO 2010105256 or WO 2010136598) did not include hVk5 sequences. In addition, the physicochemical properties of the hVk5-2 sequence have not been described and its utility was unknown.

(7-2) Construction, expression, and purification of the unglycosylated hVk5-2 sequence

The hVk5-2 sequence has a potential N-glycosylated amino acid sequence at position 20 (defined by Kabat numbering). From the point of view of homogeneity of the substance, it is desirable to avoid proteins being glycosylated, since sugar chains attached to proteins cause heterogeneity. Thus, the modified form of hVk5-2_L65 (SEQ ID NO: 5) was obtained by replacing the Asn (N) residue at position 20 (defined by Kabat numbering) with a Thr (T) residue. Amino acid substitution was performed in a manner well known to those skilled in the art using the QuikChange site-directed mutagenesis kit (Stratagene Corp.). DNA encoding a modified form of hVk5-2_L65 was inserted into vectors for expression in animal cells. Animal cells were cotransfected with the resulting animal cell expression vector having a modified hVk5-2_L65 DNA insert and an animal cell expression vector having a DNA insert encoding the CIM_H heavy chain (SEQ ID NO: 4) for expression in the manner described in Example 6. An antibody containing hVk5-2_L65 and CIM_H was expressed by transfected animal cells and purified by the method described in Example 6.

(7-3) Evaluation of an antibody containing the non-glycosylated hVk5-2 sequence with respect to its physicochemical properties

Analysis of whether the resulting antibody containing the modified hVk5-2_L65 sequence was reduced in heterogeneity relative to the antibody containing the original hVk5-2 sequence that was modified was performed using ion exchange chromatography. Ion exchange chromatography was carried out using the method shown in Table 10. The analysis results demonstrated that, as shown in FIG. 5, hVk5-2_L65, modified at the glycosylation site from the original hVk5-2 sequence, has less heterogeneity than the original hVk5-2 sequence.

Next, the antibody containing the hVk5-2_L65 sequence having reduced heterogeneity was evaluated for its ability to bind calcium ions in the manner described in Example 6. As a result, as shown in Table 11, the Tm value of the Fab domain of the antibody containing hVk5-2_L65 , modified at the glycosylation site, also varied depending on changes in the concentration of calcium ions in the antibody solution. This showed that calcium ions bind to the Fab domain of an antibody containing hVk5-2_L65 modified at the glycosylation site.

[Example 8] Evaluation of an antibody molecule containing the hVk5-2 CDR sequence for its calcium ion binding activity

(8-1) Preparation, expression and purification of engineered antibody containing hVk5-2 CDR sequence

The hVk5-2_L65 sequence is a sequence modified from the human Vk5-2 sequence by an amino acid substitution at the glycosylation site in the framework region. Example 7 shows that calcium ions bind even to an antibody containing a sequence modified at the glycosylation site. Generally, germline sequences have been desirable as framework sequences from an immunogenicity standpoint. Thus, it was investigated whether antibody framework sequences could be replaced with non-glycosylated germline framework sequences while maintaining calcium binding activity of the antibody.

A polynucleotide encoding a sequence modified from a chemically synthesized hVk5-2 sequence by replacing its framework sequences with hVk1, hVk2, hVk3 or hVk4 sequences (the modified sequences were constructed as CaVk1 (SEQ ID NO: 31), CaVk2 (SEQ ID NO: 32), CaVk3 (SEQ ID NO: 33) and CaVk4 (SEQ ID NO: 34), respectively) were linked by PCR to a polynucleotide encoding the natural kappa chain constant region (SEQ ID NO: 26). The resulting DNA fragments were individually inserted into vectors for expression in animal cells. The sequences of the resulting modified forms were confirmed in a manner well known to those skilled in the art. Animal cells were cotransfected with each plasmid thus obtained and a plasmid having a polynucleotide insert encoding CIM_H (SEQ ID NO: 4) in the manner described in Example 6. The desired antibody molecules thus expressed were purified from the culture fluid of the transfected animal cells.

(8-2) Evaluation of an engineered antibody containing the hVk5-2 CDR sequence for its calcium ion binding activity

Engineered antibodies containing framework sequences from a germline sequence (hVk1, hVk2, hVk3 or hVk4) other than the hVk5-2 sequence and a CDR sequence from the hVk5-2 sequence were evaluated for their ability to bind calcium ions in the manner described in the example. 6. The evaluation results are presented in Table 12. It was shown that the Tm value for the Fab domain of each antibody obtained varies depending on changes in the concentration of calcium ions in the antibody solution. These results demonstrated that an antibody containing framework sequences other than the hVk5-2 framework sequences also binds to calcium ions.

As is further apparent from the results, the thermal denaturation temperature (Tm), an indicator of thermal stability, of the Fab domain of each antibody modified to contain framework sequences from a germline sequence (hVk1, hVk2, hVk3, or hVk4) other than the hVk5 sequence -2, and the CDR sequence from the hVk5-2 sequence was higher than that of the Fab domain of the antibody containing the original hVk5-2 sequence that was modified. From this result, it was revealed that the antibody containing the framework sequences of hVk1, hVk2, hVk3 or hVk4 and the CDR sequences of hVk5-2 is a molecule that has calcium ion binding property and is also excellent in terms of thermal stability.

[Example 9] Identification of the calcium ion binding site present in the human germline sequence hVk5-2

(9-1) Construction of the mutation site in the CDR sequence of hVk5-2

As described in Example 8, antibodies containing light chains with CDR domains of the hVk5-2 sequence inserted into different germline sequence framework sequences were also shown to bind to calcium ions. This result indicated that the calcium ion binding site present in hVk5-2 was located in the CDR. Examples of amino acids that bind to calcium ions, i.e. chelating calcium ions include negatively charged amino acids and amino acids that can serve as hydrogen bond acceptors. Thus, antibodies containing the hVk5-2 sequence variant, obtained by mutation of the hVk5-2 sequence by replacing Asp (D) and/or Glu (E) residues in the CDR sequences with Ala (A) residues, were evaluated for their ability to bind to ions calcium.

(9-2) Preparation of the hVk5-2 sequence variant with Ala substitution and expression and purification of antibodies

Antibody molecules were produced that contained modified light chains with Ala residues instead of Asp and/or Glu residues present in the hVk5-2 CDR sequences. As described in Example 7, the non-glycosylated modified form of hVk5-2_L65 retained calcium ion binding and was thus found to be equivalent to the hVk5-2 sequence in terms of calcium ion binding properties. In this example, an amino acid change was performed with hVk5-2_L65 as the template sequence. The resulting modified forms are presented in Table 13. Amino acid substitution was carried out in a manner well known to those skilled in the art using the QuikChange site-directed mutagenesis kit (Stratagene Corp.), In fusion Advantage PCR cloning kit (Takara Bio Inc. ) and so on. Expression vectors for light chains modified by amino acid substitution were constructed.

The resulting expression vectors were sequenced in a manner well known to those skilled in the art. Human embryonic kidney-derived cell line HEK293H (Invitrogen Corp.) or Freestyle 293 cells (Invitrogen Corp.) were transiently cotransfected with the resulting expression vector for each modified light chain and the heavy chain expression vector CIM_H (SEQ ID NO: 4) for antibody expression . Each antibody was purified from the resulting culture supernatant by a method well known to those skilled in the art using rProtein A Sepharose(TM) Fast Flow (GE Healthcare Bio-Sciences Corp.). The absorbance of the purified antibody solution was measured at 280 nm using a spectrophotometer. The antibody concentration was calculated from the resulting measurement using the extinction coefficient calculated by PACE (Protein Science (1995) 4, 2411-2423).

(9-3) Evaluation of an antibody containing the hVk5-2 sequence variant with an Ala substitution for its calcium ion binding activity

Determination of whether the resulting purified antibodies bind to calcium ions was carried out in the manner described in example 6. The results are presented in table 14. The Tm values for the Fab domains of some antibodies did not change depending on the concentration of calcium ions in antibody solutions, as a result of Asp replacement and/or Glu present in the CDR sequences of the hVk5-2 sequence to Ala residues that were unable to participate in the binding or chelation of calcium ions. Substitution sites that did not cause a change in Tm values even when replaced with Ala (positions 32 and 92 (defined by Kabat numbering)) were shown to be particularly important for antibody binding of calcium ions.

[Example 10] Evaluation of an Antibody Containing an hVk1 Sequence Having a Calcium Ion Binding Motif

(10-1) Preparation of the hVk1 sequence having a calcium ion binding motif and expression and purification of antibodies

The results for calcium binding activity of the Ala substitution variants described in Example 9 indicated that the Asp and Glu residues of the CDR sequences in the hVk5-2 sequence are important for calcium binding. Thus, antibodies produced by inserting only residues 30, 31, 32, 50 and 92 (defined by Kabat numbering) into the variable region sequence of a different germline sequence were evaluated for their ability to bind calcium ions. Specifically, a modified form of LfVk1_Ca (SEQ ID NO: 43) was obtained by replacing residues 30, 31, 32, 50 and 92 (identified by Kabat numbering) in the hVk1 human germline sequence with residues 30, 31, 32, 50 and 92 (defined by Kabat numbering) in the sequence hVk5-2. The presence or absence of calcium binding ability was determined against an antibody containing the hVk1 sequence with only these five residues inserted therein, derived from the hVk5-2 sequence. The production of the modified form was carried out analogously to Example 9. The thus obtained modified LfVk1_Ca or LfVk1 light chain (SEQ ID NO: 44), containing the hVk1 light chain sequence, was coexpressed with the CIM_H heavy chain (SEQ ID NO: 4). Expression and purification of antibodies were carried out similarly to example 9.

(10-2) Evaluation of an antibody containing a human hVk1 sequence having a calcium ion binding motif for its calcium ion binding activity

Determination of whether the purified antibodies thus obtained bind to calcium ions was carried out in the manner described in example 6. The results are presented in table 15. The value of the Tm Fab domain of an antibody containing LfVk1 having the hVk1 sequence did not change depending on the change concentration of calcium ions in the antibody solution. In contrast, the Tm value of the LfVk1_Ca-containing antibody sequence changed by 1°C or more depending on the change in calcium concentration in the antibody solution, indicating that the LfVk1_Ca-containing antibody binds to calcium. These results demonstrated that not the entire hVk5-2 CDR sequence is required for calcium ion binding and that only the residues introduced to construct the LfVk1_Ca sequence are sufficient for binding.

(10-3) Construction, expression, and purification of the nondegradable LfVk1_Ca sequence

In Example 10(10-1), a modified form of LfVk1_Ca (SEQ ID NO: 43) was obtained by replacing residues 30, 31, 32, 50 and 92 (identified by Kabat numbering) in the human germline sequence hVk1 with residues 30, 31. 32, 50, and 92 (defined by Kabat numbering) in the hVk5-2 sequence and was shown to bind calcium ions. This may provide rationale for constructing a Ca library containing LfVk1_Ca sequences. Since the physicochemical properties of the LfVk1_Ca sequence have not been described, its utility is unknown. The LfVk1_Ca sequence contained Asp at positions 30, 31 and 32 (defined by Kabat numbering) and contained in its CDR1 sequence the Asp-Asp sequence, which is reported to be degradable under acidic conditions (J. Pharm. Biomed. Anal. (2008 ) 47 (1), 23-30). From a storage stability point of view, it is desirable to avoid degradation under acidic conditions. Thus, modified forms of LfVk1_Ca1 (SEQ ID NO: 45), LfVk1_Ca2 (SEQ ID NO: 46) and LfVk1_Ca3 (SEQ ID NO: 47) were obtained by replacing the degradable Asp (D) residues with Ala (A) residues. Amino acid substitutions were performed in a manner well known to those skilled in the art using the QuikChange site-directed mutagenesis kit (Stratagene Corp.). The DNA encoding each form was inserted into vectors for expression in animal cells. Animal cells were cotransfected with the resulting animal cell expression vector having the DNA insert of each modified form and the animal cell expression vector having the DNA insert encoding the GC_H heavy chain (SEQ ID NO: 48) for expression in the manner described in Example 6 Antibodies expressed in transfected animal cells were purified as described in Example 6.

(10-4) Stability assessment of an antibody containing a non-degradable LfVk1_Ca sequence

An assessment of whether the degradation of each antibody prepared according to Example 10(10-3) in solution at pH 6.0 was reduced compared to an antibody containing the original LfVk1_Ca sequence modified was performed by comparing the heterogeneity of the antibodies after thermal acceleration. Each antibody was dialyzed overnight against a solution containing 20 mM histidine-HCl and 150 mM NaCl (pH 6.0) at 4°C. The dialyzed antibody was adjusted to 0.5 mg/ml and stored at 5°C or 50°C for 3 days. Each antibody thus stored was subjected to ion exchange chromatography in the manner described in Example 7. The results of the analysis demonstrated that, as shown in FIG. 6, LfVk1_Ca1 modified at the degradation site has less heterogeneity than the original LfVk1_Ca sequence, and its degradation under thermal acceleration is significantly reduced. These results demonstrated that degradation occurs at the Asp(D) residue located at position 30 in the LfVk1_Ca sequence and can be avoided by amino acid modification.

(10-5) Preparation of the LVk1_Ca light chain sequence non-degradable at the Asp30 residue and expression and purification of antibodies

The results of inhibition of degradation of the variant with the Ala substitution described in example 10(10-4) showed that degradation under acidic conditions occurs at the Asp (D) residue at position 30 (defined by Kabat numbering) in the CDR sequence of the LfVk1_Ca sequence and can be inhibit by replacing residue 30 (identified by Kabat numbering) with another amino acid (in example 10(10-4), the Ala(A) residue). Thus, the sequence (designated LfVk1_Ca6; SEQ ID NO: 49) obtained by replacing residue 30 (defined by Kabat numbering) with the Ser(S) residue, a typical residue capable of chelating calcium ions, was evaluated for the presence or absence of inhibition degradation. Preparation of the modified form was carried out in the same manner as in Example 10. The resulting modified light chain sequence LfVk1_Ca6 or light chain sequence LfVk1_Ca was coexpressed with the heavy chain GC_H (SEQ ID NO: 48). Expression and purification of the antibody were carried out similarly to Example 10.

(10-6) Evaluation of the light chain sequence LVk1_Ca, non-degradable at the Asp30 residue

The storage stability of the purified antibodies thus obtained under acidic conditions was determined by the method described in example 10(10-4). The results demonstrated that, as shown in FIG. 7, the degradation of the antibody containing the LfVk1_Ca6 sequence was inhibited to a greater extent than the degradation of the antibody containing the original LfVk1_Ca sequence.

In addition, determination of whether the antibody containing the LfVk1_Ca sequence and the antibody containing the LfVk1_Ca6 sequence binds to calcium ions was carried out in the manner described in Example 6. The results are presented in Table 16. Tm values for the Fab domains of the antibody containing the LfVk1_Ca sequence and antibodies containing the non-degradable sequence LfVk1_Ca6, in each case changed by 1°C or more depending on the calcium concentration in the antibody solutions.

[Example 11] Construction of a set (Ca-library) of antibody molecules containing calcium ion binding motifs in variable regions that allow the efficient production of a binding antibody that binds to an antigen in a calcium ion concentration-dependent manner

Preferred examples of calcium binding motifs include the hVk5-2 sequence and its CDR sequences, and further selected residues 30, 31, 32, 50 and 92 (defined by Kabat numbering). In addition, the EF hand motif (calmodulin, etc.) of the calcium-binding protein and the C-type lectin (ASGPR, etc.) also correspond to calcium-binding motifs.

The Ca library is constructed using heavy and light chain variable regions. The heavy chain variable regions used were human antibody sequences, while calcium binding motifs were inserted into the light chain variable regions. The hVk1 sequence was chosen as the heavy chain variable region template sequence to introduce the calcium-binding motif. As shown in Example 10, an antibody containing the LfVk1_Ca sequence, in which the hVk5-2 CDR sequence was introduced into the hVk1 sequence as a calcium binding motif, was shown to bind to calcium ions. In the template sequence, different amino acids were allowed to appear at each given position to expand the diversity of antigen-binding molecules composing the library. Surface exposed positions in the variable regions that are likely to interact with antigens were selected as the position for the appearance of multiple amino acids. In particular, positions 30, 31, 32, 34, 50, 53, 91, 92, 93, 94 and 96 (defined by Kabat numbering) were selected as such non-strict residues.

Next, the types and occurrence of amino acid residues to appear were selected. Amino acids of non-strict residues were analyzed with respect to their frequency of occurrence in the hVk1 and hVk3 sequences registered in the Kabat database (KAVAT, E.A. et al.: "Sequences of Proteins of Immunological interest", vol. 91, 1991, NIH PUBLICATION). Based on the analysis results, the types of amino acids appearing in the Ca library were selected from amino acids with a high frequency of occurrence at each position. This technique also selected amino acids that were confirmed to have a low frequency of occurrence based on the analysis results to prevent imbalance of amino acid properties. The frequency of occurrence of the selected amino acids was selected based on the results of the analysis of the Kabat database.

Considering the amino acids thus selected and their frequency of occurrence, the following Ca library was constructed: A Ca library that was emphasized on sequence diversity so that it contained calcium-binding motifs and a variety of amino acids at each residue other than the motifs. The detailed construction of the Ca-library is presented in Tables 17 and 18 (“Position” in each table reflects the Kabat numbering). The frequency of occurrence of the amino acids described in Tables 17 and 18 may include Leu (L) instead of Ser (S) at position 94, determined by Kabat numbering, in the case of Asn (N) at position 92.

[Example 12] Obtaining the Ca library

A gene library of antibody heavy chain variable regions was amplified by PCR using poly-A RNA obtained from human PBMC, commercially available human poly-A RNA, and the like. as a matrix. The light chain portions of the antibody variable regions were designed to increase the frequency of antibodies that retained calcium binding motifs and were able to bind antigens in a calcium concentration dependent manner as described in Example 11. An antibody light chain variable region library was constructed according to information on the frequency of occurrence of amino acids in natural human antibodies (KAVAT, E.A. et al.: "Sequences of Proteins of Immunological interest", vol. 91, 1991, NIH PUBLICATION), so that amino acids with a high frequency of occurrence in natural antibody sequences human were evenly distributed as amino acid residues other than those with an integrated calcium-binding motif among the non-strict residues. Combinations of sequences in the antibody heavy chain variable region gene library and in the antibody light chain variable region gene library thus obtained were inserted into phagemid vectors. A human antibody phage display library was constructed displaying Fab domains consisting of human antibody sequences (Methods Mol Biol. (2002) 178, 87-100).

Portions of antibody genes isolated from E. coli transformed with the antibody gene library were sequenced according to the method described in Example 2. FIG. Figure 8 shows the distribution of amino acids in the sequences of 290 types of clones obtained in this way, and the model distribution of amino acids.

[Example 13] Evaluation of a molecule contained in a Ca library for its calcium ion binding activity

(13-1) Calcium ion binding activity of a molecule contained in the Ca library

Because the hVk5-2 sequence shown to bind calcium ions has a low frequency of occurrence among germline sequences, as shown in Example 7, obtaining calcium-binding antibodies from an antibody library consisting of human germline sequences or B- cells obtained by immunizing mice expressing human antibodies was ineffective. Thus, a Ca library was constructed according to Example 12. The constructed Ca library was evaluated for the presence or absence of clones exhibiting calcium binding.

(13-2) Antibody expression and purification

The gene of each clone contained in the Ca library was inserted into plasmids for expression in animal cells. Antibody expression was performed as follows: Human embryonic kidney-derived Freestyle 293-F line (Invitrogen Corp.) was suspended in Freestyle 293 Expression Medium (Invitrogen Corp.). The suspension having a cell density of 1.33×10 ⁶ cells/ml was inoculated at a concentration of 3 ml/well into a 6-well plate. The resulting plasmids were transferred into cells by lipofection. Cells were cultured for 4 days in a _CO2 incubator (37°C, 8% _CO2 , 90 rpm). Antibodies were purified from the resulting culture supernatant in a manner well known to those skilled in the art using rProtein A Sepharose(TM) Fast Flow (Amersham Biosciences, Inc.). The absorbance of the purified antibody solution was measured at 280 nm using a spectrophotometer. The antibody concentration was calculated from the resulting measurement using the extinction coefficient calculated by PACE (Protein Science (1995) 4, 2411-2423).

(13-3) Evaluation of the resulting antibody with respect to its binding of calcium ions

Determination of whether the purified antibodies thus obtained bound to calcium ions was carried out in the manner described in Example 6. The results are presented in Table 19. The Tm values for the Fab domains of a variety of antibodies contained in the Ca library varied depending on the concentration calcium ions, indicating that the Ca library contains molecules that bind to calcium ions.

[Example 14] Preparation of an antibody that binds to the IL-6 receptor in a Ca-dependent manner

(14-1) Obtaining an antibody fragment that binds to an antigen in a Ca-dependent manner from a library by bead panning

The first round of screening of the constructed Ca library was carried out by increasing the content of only antibody fragments with the ability to bind antigen (IL-6R).

Phages were obtained from E. coli containing an engineered phagemid for phage display. 2.5 M NaCl/10% PEG was added to E. coli cultures that produced phages, and the pool of phages thus precipitated was diluted with TBS to obtain a phage library solution. BSA and CaCl ₂ were added to the solution with the phage library (final concentration: 4% BSA and 1.2 mM calcium ions). The panning method was carried out in accordance with the conventional panning method using antigens immobilized on magnetic beads (J. Immunol. Methods. (2008) 332 (1-2), 2-9; J. Immunol. Methods. (2001) 247 (1 -2), 191-203; Biotechnol. Prog. (2002) 18 (2) 212-20; and Mol. Cell Proteomics (2003) 2 (2), 61-9). The magnetic beads used were neutravidin-coated beads (Sera-Mag SpeedBeads NeutrAvidin-coated) or streptavidin-coated beads (Dynabeads M-280 Streptavidin).

Specifically, 250 pmol of biotin-labeled antigens were added to the resulting phage library solution and thereby brought into contact with the phage library solution at room temperature for 60 minutes. BSA-blocked magnetic beads were added, and then the antigen-phage complexes were bound to the magnetic beads at room temperature for 15 minutes. The beads were washed three times with 1 ml 1.2 mM CaCl ₂ /TBST (TBST containing 1.2 mM CaCl ₂ ), and then washed additionally two times with 1 ml 1.2 mM CaCl ₂ /TBS (TBS containing 1.2 mM _CaCl2 ). After adding 0.5 ml of 1 mg/ml trypsin, the beads were suspended at room temperature for 15 minutes, and immediately thereafter the beads were separated using a magnetic stand to collect the phage solution. The collected phage solution was added to 10 ml of E. coli strain ER2738 at logarithmic growth phase (OD600: 0.4-0.7). The E. coli strain was infected with phages by gently rotating the strain culture at 37°C for 1 hour. Infected E. coli were inoculated onto a 225 mm × 225 mm dish. Next, phages were collected from the inoculated E. coli cultures to obtain a solution with a phage library.

In the second round of panning, the content of phages with antigen-binding ability or Ca-dependent binding ability as an indicator was increased.

Specifically, to increase the content with antigen-binding ability as an indicator, 40 pmol of biotin-labeled antigens were added to the resulting phage library solution and thereby brought into contact with the phage library solution at room temperature for 60 minutes. BSA-blocked magnetic beads were added, and then the antigen-phage complexes were bound to the magnetic beads at room temperature for 15 minutes. The beads were washed several times with 1 ml of 1.2 mM CaCl ₂ /TBST and 1.2 mM CaCl ₂ /TBS. The beads, supplemented with 0.5 ml of 1 mg/ml trypsin, were then suspended at room temperature for 15 minutes, and immediately thereafter the beads were separated using a magnetic stand to collect the phage solution. Addition of 5 μl of 100 mg/ml trypsin to the collected phage solution digested the pIII proteins (helper phage-derived pIII proteins) of non-Fab phages to eliminate the ability of non-Fab phages to infect E. coli. The collected phage solution was added to 10 ml of E. coli strain ER2738 at logarithmic growth phase (OD600: 0.4-0.7). The E. coli strain was infected with phages by gently rotating the strain culture at 37°C for 1 hour. Infected E. coli were inoculated onto a 225 mm × 225 mm dish. Next, phages were collected from the inoculated E. coli cultures to obtain a solution with a phage library.

To increase the content with Ca-dependent binding ability as an indicator, 40 pmol of biotin-labeled antigens were added to the resulting phage library solution and thereby brought into contact with the phage library solution at room temperature for 60 minutes. BSA-blocked magnetic beads were added, and then the antigen-phage complexes were bound to the magnetic beads at room temperature for 15 minutes. The beads were washed with 1 ml of 1.2 mM CaCl ₂ /TBST and 1.2 mM CaCl ₂ /TBS. The beads supplemented with 0.1 ml of 2 mM EDTA/TBS (TBS containing 2 mM EDTA) were then suspended at room temperature. Immediately after this, the granules were separated using a magnetic stand to collect the solution with phages. Addition of 5 μl of 100 mg/ml trypsin to the collected phage solution digested the pIII proteins (helper phage-derived pIII proteins) of non-Fab phages to eliminate the ability of non-Fab phages to infect E. coli. The collected phage solution was added to 10 ml of E. coli strain ER2738 at logarithmic growth phase (OD600: 0.4-0.7). The E. coli strain was infected with phages by gently rotating the strain culture at 37°C for 1 hour. Infected E. coli were inoculated onto a 225 mm × 225 mm dish. Next, phages were collected from the inoculated E. coli cultures to obtain a solution with a phage library.

(14-2) Evaluation by phage ELISA

Phage-containing culture supernatant was prepared according to a conventional method (Methods Mol. Biol. (2002) 178, 133-145) from each single colony of E. coli obtained using the method described above.

After addition of BSA and CaCl ₂ , the phage-containing culture supernatant was subjected to ELISA using the following procedure: a StreptaWell 96 microtiter plate (F. Hoffmann-La Roche Ltd.) was coated overnight with 100 μl of PBS containing biotin-labeled antigens. Each well of the plate was washed with PBST to remove unbound antigens. The well was then blocked with 250 μl of 4% BSA-TBS for 1 hour or more. After removal of 4% BSA-TBS, the resulting supernatant was added to each well and the plate was allowed to stand at 37°C for 1 hour to allow phage-exposed antibodies to bind to the antigens contained in each well. Each well was washed with 1.2 mM CaCl ₂ /TBST, and 1.2 mM CaCl ₂ /TBS or 1 mM EDTA/TBS was added. The plate was left to stand at 37°C for 30 minutes for incubation. After washing with 1.2 mM CaCl ₂ /TBST, HRP-conjugated anti-M13 antibodies (Amersham Pharmacia Biotech Inc.) diluted with 1.2 mM ionized calcium TBS were added to each well. The plate was incubated for 1 hour. After washing with 1.2 mM CaCl ₂ /TBST, plain TMB solution (ZYMED Laboratories, Inc.) was added to each well. The chromogenic reaction of the solution in each well was completed by adding sulfuric acid. Color development was then analyzed based on absorbance at 450 nm.

The genes of the phage ELISA clones were amplified using specific primers and then analyzed for their nucleotide sequences. The results of the phage ELISA and sequence analysis are presented below in Table 20.

(14-3) Antibody expression and purification

The gene of each clone assessed as having Ca-dependent antigen binding by phage ELISA was inserted into plasmids for expression in animal cells. Antibody expression was performed as follows: Human embryonic kidney-derived Freestyle 293-F line (Invitrogen Corp.) was suspended in Freestyle 293 Expression Medium (Invitrogen Corp.). The suspension having a cell density of 1.33×10 ⁶ cells/ml was inoculated at a concentration of 3 ml/well into a 6-well plate. The resulting plasmids were transferred into cells by lipofection. Cells were cultured for 4 days in a CO ₂ incubator (37°C, 8% CO ₂ , 90 rpm). Antibodies were purified from the resulting culture supernatant in a manner well known to those skilled in the art using rProtein A Sepharose(TM) Fast Flow (Amersham Biosciences, Inc.). The absorbance of the purified antibody solution was measured at 280 nm using a spectrophotometer. The antibody concentration was calculated from the resulting measurement using the extinction coefficient calculated by PACE (Protein Science (1995) 4, 2411-2423).

(14-4) Evaluation of the resulting antibody with respect to its Ca-dependent binding ability to the human IL-6 receptor

To evaluate the Ca dependence of the human IL-6 receptor binding activity of antibodies 6RC1IgG_010 (heavy chain: SEQ ID NO: 72 and light chain: SEQ ID NO: 73), 6RC1IgG_012 (heavy chain: SEQ ID NO: 74 and light chain: SEQ ID NO: 75) and 6RC1IgG_019 (heavy chain: SEQ ID NO: 76 and light chain: SEQ ID NO: 77) prepared according to Example 14, these antibodies were analyzed for their interaction with human IL-6 receptors using Biacore T100 (GE Healthcare Bio-Sciences Corp.). Tocilizumab (heavy chain: SEQ ID NO: 24 and light chain: SEQ ID NO: 25) was used as a control antibody lacking Ca-dependent binding activity to human IL-6 receptors. The interaction was analyzed in solutions with calcium ion concentrations of 1.2 mM and 3 μM as high calcium ion concentration and low calcium ion concentration conditions, respectively. Each antibody of interest was captured on a CM5 sensor chip (GE Healthcare Bio-Sciences Corp.) immobilized with Protein A/G (Invitrogen Corp.), in the appropriate amount using the amine coupling method. Two types of running buffers were used: 20 mM ACES, 150 mM NaCl, 0.05% (w/v) Tween 20, and 1.2 mM CaCl ₂ (pH 7.4); and 20 mM ACES, 150 mM NaCl, 0.05% (w/v) Tween 20, and 3 μM CaCl ₂ (pH 7.4). Human IL-6 receptors were also diluted using each of these buffers. These assays were all performed at 37°C.

In an antigen-antibody interaction assay using tocilizumab control antibody, 6RC1IgG_010 antibody, 6RC1IgG_012 antibody, and 6RC1IgG_019 antibody, diluted IL-6 receptor solution or empty running buffer was injected at a flow rate of 5 μL/min for 3 minutes to interact IL-6 receptors. 6 with tocilizumab antibody, 6RC1IgG_010 antibody, 6RC1IgG_012 antibody or 6RC1IgG_019 antibody captured on the sensor chip. It was then injected with 10 mM glycine-HCl (pH 1.5) at a flow rate of 30 μL/min for 30 seconds to regenerate the sensor chip.

In fig. 9 shows sensorgrams for antibodies analyzed at high calcium ion concentrations using this method.

Sensograms for the tocilizumab antibody, 6RC1IgG_010 antibody, 6RC1IgG_012 antibody and 6RC1IgG_019 antibody under low calcium ion concentration conditions were also obtained in a similar manner. In fig. Figure 10 shows sensograms of antibodies at low concentrations of calcium ions.

As a result, it was found that the ability to bind the IL-6 receptor of antibodies 6RC1IgG_010, antibodies 6RC1IgG_012 and antibodies 6RC1IgG_019 decreased significantly when the concentration of calcium ions in the buffer changed from 1.2 mM to 3 μM.

[Example 15] Preparation of an antibody that binds to the IL-6 receptor in a Ca-dependent manner from a human antibody library using phage display technology

(15-1) Preparation of a phage display library of naive human antibodies

A human antibody phage display library, consisting of a variety of phages displaying Fab domains having different human antibody sequences, was constructed in accordance with a method well known to those skilled in the art using poly-A RNA obtained from human PBMC, commercially available poly-A RNA. And human RNA, etc. as a matrix.

(15-2) Obtaining an antibody fragment that binds to an antigen in a Ca-dependent manner from a library by bead panning

The first round of screening of the constructed naive human antibody phage display library was carried out by increasing the content of only antibody fragments with the ability to bind antigen (IL-6R), or by increasing the content of antibody fragments with the ability to Ca concentration-dependent antigen binding (IL-6 receptor). Increasing the content of antibody fragments with the ability for Ca concentration-dependent antigen binding (IL-6 receptor) as an indicator was carried out by eluting phages from a phage library bound to IL-6 receptors in the presence of Ca ions using EDTA Ca ion chelating agents. The antigens used were biotin-labeled IL-6 receptors.

Phages were obtained from E. coli containing an engineered phagemid for phage display. 2.5 M NaCl/10% PEG was added to E. coli cultures that produced phages, and the pool of phages thus precipitated was diluted with TBS to obtain a phage library solution. Next, BSA and CaCl ₂ were added to the solution with the phage library (final concentration: 4% BSA and 1.2 mM calcium ions). The panning method was carried out in accordance with the conventional panning method using antigens immobilized on magnetic beads (J. Immunol. Methods. (2008) 332 (1-2), 2-9; J. Immunol. Methods. (2001) 247 (1 -2), 191-203; Biotechnol. Prog. (2002) 18 (2) 212-20; and Mol. Cell Proteomics (2003) 2 (2), 61-9). The magnetic beads used were neutravidin-coated beads (Sera-Mag SpeedBeads NeutrAvidin-coated) or streptavidin-coated beads (Dynabeads M-280 Streptavidin).

Specifically, 250 pmol of biotin-labeled antigens were added to the resulting phage library solution and thereby brought into contact with the phage library solution at room temperature for 60 minutes. BSA-blocked magnetic beads were added, and then the antigen-phage complexes were bound to the magnetic beads at room temperature for 15 minutes. The beads were washed once with 1 ml of 1.2 mM CaCl ₂ /TBS (TBS containing 1.2 mM CaCl ₂ ). The phage solution was then collected by elution according to the general method to increase the content of antibody fragments having the ability to bind the IL-6 receptor, or by elution from beads suspended in 2 mM EDTA/TBS (TBS containing 2 mM EDTA) to increase the content antibody fragments with the ability for Ca concentration-dependent binding of the IL-6 receptor as an indicator. The collected phage solution was added to 10 ml of E. coli TG1 strain at logarithmic growth phase (OD600: 0.4-0.7). The E. coli strain was infected with phages by gently rotating the strain culture at 37°C for 1 hour. Infected E. coli were inoculated onto a 225 mm × 225 mm dish. Next, phages were collected from the inoculated E. coli cultures to obtain a solution with a phage library.

In the second and subsequent rounds of panning, the content of phages with Ca-dependent binding ability as an indicator was increased. Specifically, 40 pmol of biotin-labeled antigens were added to the resulting phage library solution and thereby brought into contact with the phage library solution at room temperature for 60 minutes. BSA-blocked magnetic beads were added, and then the antigen-phage complexes were bound to the magnetic beads at room temperature for 15 minutes. The beads were washed with 1 ml of 1.2 mM CaCl ₂ /TBST and 1.2 mM CaCl ₂ /TBS. The beads supplemented with 0.1 mL of 2 mM EDTA/TBS were then suspended at room temperature. Immediately after this, the granules were separated using a magnetic stand to collect the solution with phages. The collected phage solution was added to 10 ml of E. coli TG1 strain at logarithmic growth phase (OD600: 0.4-0.7). The E. coli strain was infected with phages by gently rotating the strain culture at 37°C for 1 hour. Infected E. coli were inoculated onto a 225 mm × 225 mm dish. Next, phages were collected from the inoculated E. coli cultures to obtain a solution with a phage library. Panning with Ca-dependent binding ability as an indicator was repeated several times.

(15-3) Evaluation by phage ELISA

Phage-containing culture supernatant was prepared according to a conventional method (Methods Mol. Biol. (2002) 178, 133-145) from each single colony of E. coli obtained using the method described above.

After addition of BSA and CaCl ₂ (final concentration: 4% BSA and 1.2 mM calcium ions), the phage-containing culture supernatant was subjected to ELISA using the following procedure: a StreptaWell 96 microtiter plate (F. Hoffmann-La Roche Ltd.) was covered for nights with 100 μl of PBS containing biotin-labeled antigens. Each well of the plate was washed with PBST to remove unbound antigens. The well was then blocked with 250 μl of 4% BSA-TBS for 1 hour or more. After removal of 4% BSA-TBS, the resulting supernatant was added to each well and the plate was allowed to stand at 37°C for 1 hour to allow phage-exposed antibodies to bind to the antigens contained in each well. Each well was washed with 1.2 mM CaCl ₂ /TBST, and 1.2 mM CaCl ₂ /TBS or 1 mM EDTA/TBS was added. The plate was left to stand at 37°C for 30 minutes for incubation. After washing with 1.2 mM CaCl ₂ /TBST, HRP-conjugated anti-M13 antibodies (Amersham Pharmacia Biotech Inc.) diluted in TBS containing 4% BSA and 1.2 mM ionized calcium were added to each well (all reported at final concentrations ). The plate was incubated for 1 hour. After washing with 1.2 mM CaCl ₂ /TBST, plain TMB solution (ZYMED Laboratories, Inc.) was added to each well. The chromogenic reaction of the solution in each well was completed by adding sulfuric acid. Color development was then analyzed based on absorbance at 450 nm.

Genes of antibody fragments assessed to have Ca-dependent antigen binding capacity by phage ELISA were amplified as templates using specific primers and then analyzed for their nucleotide sequences.

(15-4) Antibody expression and purification

The gene of each clone assessed as having Ca-dependent antigen binding by phage ELISA was inserted into plasmids for expression in animal cells. Antibody expression was performed as follows: Human embryonic kidney-derived Freestyle 293-F line (Invitrogen Corp.) was suspended in Freestyle 293 Expression Medium (Invitrogen Corp.). The suspension having a cell density of 1.33×10 ⁶ cells/ml was inoculated at a concentration of 3 ml/well into a 6-well plate. The resulting plasmids were transferred into cells by lipofection. Cells were cultured for 4 days in a _CO2 incubator (37°C, 8% _CO2 , 90 rpm). Antibodies were purified from the resulting culture supernatant in a manner well known to those skilled in the art using rProtein A Sepharose(TM) Fast Flow (Amersham Biosciences, Inc.). The absorbance of the purified antibody solution was measured at 280 nm using a spectrophotometer. The antibody concentration was calculated from the resulting measurement using the extinction coefficient calculated by PACE (Protein Science (1995) 4, 2411-2423).

[Example 16] Evaluation of the resulting antibody with respect to its Ca-dependent binding ability to the human IL-6 receptor

To evaluate the Ca dependence of the human IL-6 receptor binding activity of the antibodies 6RL#9-IgG1 (heavy chain: SEQ ID NO: 78 (sequence SEQ ID NO: 10 related to the IgG1-derived constant region) and light chain: SEQ ID NO : 79) and FH4-IgG1 (heavy chain: SEQ ID NO: 80 and light chain: SEQ ID NO: 81) prepared according to Example 15, these antibodies were subjected to kinetic analysis regarding their antigen-antibody reaction with human IL-6 receptors using Biacore T100 (GE Healthcare Bio-Sciences Corp.). H54/L28-IgG1 described in WO 2009125825 (heavy chain: SEQ ID NO: 82 and light chain: SEQ ID NO: 83) was used as a control antibody lacking Ca-dependent binding activity to human IL-6 receptors. Kinetic analysis of the antigen-antibody reaction was performed in solutions with calcium ion concentrations of 2 mM and 3 μM as high calcium ion concentration and low calcium ion concentration conditions, respectively. Each antibody of interest was captured on a CM4 sensor chip.

(GE Healthcare Bio-Sciences Corp.) with Protein A (Invitrogen Corp.) immobilized on it, in an appropriate amount by the amine addition method. Two types of buffer solutions were used as running buffers: 10 mM ACES, 150 mM NaCl, 0.05% (w/v) Tween 20, and 2 mM CaCl ₂ (pH 7.4); and 10 mM ACES, 150 mM NaCl, 0.05% (w/v) Tween 20, and 3 μcol/L CaCl ₂ (pH 7.4). Human IL-6 receptors were also diluted using each of these buffers. All analysis was performed at 37°C.

In the antigen-antibody reaction kinetic assay using the H54/L28-IgG1 antibody, diluted IL-6 receptor solution or empty running buffer was injected at a flow rate of 20 μL/min for 3 minutes to interact the IL-6 receptors with the H54/L28- antibody. IgG1 captured on the sensor chip. The running buffer was then injected at a flow rate of 20 μl/min for 10 minutes, and the dissociation of IL-6 receptors was observed. It was then injected with 10 mM glycine-HCl (pH 1.5) at a flow rate of 30 μL/min for 30 seconds to regenerate the sensor chip. The association rate constant ka (1/M s) and the dissociation rate constant kd (1/s) were calculated as kinetic parameters from the sensorgram obtained in the assay. These values were used to calculate the dissociation constant KD (M) of the H54/L28-IgG1 antibody against the human IL-6 receptor. Each parameter was calculated using Biacore T100 Evaluation Software (GE Healthcare Bio-Sciences Corp.).

In an antigen-antibody reaction kinetic assay using FH4-IgG1 antibody and 6RL#9-IgG1 antibody, diluted IL-6 receptor solution or empty running buffer was injected at a flow rate of 5 μL/min for 15 minutes to interact IL-6 receptors with antibody FH4-IgG1 and antibody 6RL#9-IgG1 captured on the sensor chip. It was then injected with 10 mM glycine-HCl (pH 1.5) at a flow rate of 30 μL/min for 30 seconds to regenerate the sensor chip. The dissociation constant KD (M) was calculated using the steady state affinity model from the sensorgrams obtained in the assay. Each parameter was calculated using Biacore T100 Evaluation Software (GE Healthcare Bio-Sciences Corp.).

Table 21 shows the dissociation constant KD for each IL-6 receptor antibody determined by this method in the presence of 2 mM CaCl ₂ .

The KD of the H54/L28-IgG1 antibody under conditions of a Ca concentration of 3 μM can be calculated in the same way as in the presence of Ca at a concentration of 2 mM. It was revealed that the FH4-IgG1 antibody and the 6RL#9-IgG1 antibody almost did not bind to the IL-6 receptor under conditions of a Ca concentration of 3 μM. Thus, KD is difficult to calculate using the above method. Instead, the KD values of these antibodies under 3 μM Ca concentration conditions can be calculated according to expression 1 (Biacore T100 Software Handbook, BR-1006-48, AE 01/2007).

[Expression 1]

Req = C×Rmax/(KD+C)+RI

In the above expression 1, each character is defined as follows:

Req (RU): Steady state binding levels

Rmax (RU): Ability of a surface to bind an analyte

RI (RU): Cumulative contribution of refractive index in a sample

C (M): Concentration of analyte

KD (M): Equilibrium dissociation constant

Table 22 presents the results of an approximate assessment of the dissociation constant KD of each antibody in relation to the IL-6 receptor at a Ca concentration of 3 µmol/l in accordance with expression 1. In table 22, Req, Rmax, RI and C correspond to the values hypothetically assumed based on analysis results.

As a result, the FH4-IgG1 antibody and the 6RL#9-IgG1 antibody were predicted to have IL-6 receptor KD values that were increased by approximately 60-fold and approximately 120-fold, respectively (affinities decreased by 60-fold and 120-fold or more ) when the concentration of CaCl ₂ in the buffer decreases from 2 mM to 3 μM.

Table 23 summarizes the KD values of three types of antibodies: H54/L28-IgG1, FH4-IgG1 and 6RL#9-IgG1, in the presence of 2 mM CaCl ₂ and in the presence of 3 μM CaCl ₂ , and the dependence of the KD values on Ca.

It was found that the H54/L28-IgG1 antibody did not differ in IL-6 receptor binding depending on differences in Ca concentration. In contrast, a significant decrease in IL-6 receptor binding was observed for the FH4-IgG1 antibody and the 6RL#9-IgG1 antibody under low Ca concentration conditions (Table 23).

[Example 17] Evaluation of the resulting antibody regarding its binding of calcium ions

Next, the average thermal denaturation temperature (Tm) was measured using differential scanning calorimetry (DSC) (MicroCa1 VP-Capillary DSC, MicroCal) as an indicator to evaluate the binding of calcium ions by antibodies. The average thermal denaturation temperature (Tm) serves as an indicator of stability and becomes higher when the protein is stabilized through the binding of calcium ions, compared to the average thermal denaturation temperature (Tm) of a non-calcium ion-bound protein (J. Biol. Spet. (2008 ) 283, 37, 25140-25149). The change in the Tm value of each antibody was assessed depending on the change in the concentration of calcium ions in the antibody solution to assess the binding activity of the antibody with calcium ions. Purified antibodies were dialyzed (EasySEP, Tomy Seiko Co., Ltd.) against a solution containing 20 mM Tris-HCl, 150 mM NaCl and 2 mM CaCl ₂ (pH 7.4) or containing 20 mM Tris-HCl, 150 mM NaCl , and 3 µM CaCl ₂ (pH 7.4) as an external solution. Each antibody solution was adjusted to approximately 0.1 mg/ml antibody using the dialysis solution and subjected to DSC analysis as a test substance at 20°C to 115°C with the temperature rise rate set at 240°C/h . Table 24 presents the average thermal denaturation temperature (Tm) for each Fab domain of the antibody, calculated from the resulting denaturation curve (DSC).

The results presented in Table 24 showed that the Tm values for the Fab domains of antibody FH4-IgG1 and antibody 6RL#9-IgG1, which have calcium-dependent binding ability, varied depending on changes in the concentration of calcium ions, while the value Tm for the Fab domain of the H54/L28-IgG1 antibody, which does not have the ability for calcium-dependent binding, did not change depending on changes in the concentration of calcium ions. This change in Tm values for the Fab domains of the FH4-IgG1 antibody and the 6RL#9-IgG1 antibody indicates that the Fab domains were stabilized through the binding of calcium ions to these antibodies. These results demonstrated that antibody FH4-IgG1 and antibody 6RL#9-IgG1 bind to calcium ions, while antibody H54/L28-IgG1 does not bind to calcium ions.

[Example 18] Identification of calcium ion binding site in antibody 6RL#9 using X-ray diffraction analysis (18-1) X-ray diffraction analysis

As shown in Example 17, thermal denaturation temperature Tm analysis showed that antibody 6RL#9 bound to calcium ions. However, the site through which antibody 6RL#9 bound to calcium ions was unpredictable. Thus, the residue responsible for interaction with calcium ions was identified in the sequence of antibody 6RL#9 using an X-ray diffraction approach.

(18-2) Expression and purification of antibody 6RL#9

Antibody 6RL#9, expressed for X-ray crystallography analysis, was purified. Specifically, animal cells were transiently transfected with expression plasmids in animal cells prepared to provide appropriate expression of the heavy chain (SEQ ID NO: 78) and light chain (SEQ ID NO: 79) of antibody 6RL#9. The resulting plasmids were transfected by lipofection into 800 ml cells of the human embryonic kidney cell line Freestyle 293-F (Invitrogen Corp.), suspended to a final cell density of 1×10 ⁶ cells/ml in Freestyle 293 Expression Medium (Invitrogen Corp. .). Cells transfected with plasmids were cultured for 5 days in a CO ₂ incubator (37°C, 8% CO ₂ , 90 rpm). Antibodies were purified from the resulting culture supernatant in a manner well known to those skilled in the art using rProtein A Sepharose(TM) Fast Flow (Amersham Biosciences, Inc.). The absorbance of the purified antibody solution was measured at 280 nm using a spectrophotometer. The antibody concentration was calculated from the resulting measurement using the extinction coefficient calculated by PACE (Protein Science (1995) 4, 2411-2423).

(18-3) Purification of Fab fragment from antibody 6RL#9

Antibody 6RL#9 was concentrated to 21 mg/ml using an ultrafiltration membrane with a molecular weight cutoff of 10,000 MWCO. The antibody was diluted to 5 mg/ml using 4 mM L-cysteine/5 mM EDTA/20 mM sodium phosphate buffer (pH 6.5), yielding 2.5 ml of antibody sample. After adding 0.125 mg papain (Roche Applied Science), the sample was mixed and then left to stand at 35°C for 2 hours. The sample left standing in this manner was then supplemented with one tablet of Protease Inhibitor Cocktail Mini, EDTA-Free (Roche Applied Science), dissolved in 10 ml of 25 mM MES buffer solution (pH 6), and left to stand on ice to complete the protease reaction. reactions with papain. The sample was then added to a 1-mL HiTrap SP HP cation exchange column (GE Healthcare Bio-Sciences Corp.) (equilibrated with 25 mM MES buffer solution (pH b)), below which a 1-mL HiTrap MabSelect carrier protein A column was connected in series Sure (GE Healthcare Bio-Sciences Corp.). The purified fraction of the Fab fragment of antibody 6RL#9 was obtained by eluting on a linear concentration gradient of NaCl up to 300 mM in this buffer solution. Next, the resulting purified fraction was concentrated to approximately 0.8 ml using a 5000 MWCO ultrafiltration membrane. The concentrate was added to a Superdex 200 10/300 GL gel filtration column (GE Healthcare Bio-Sciences Corp.) equilibrated with 100 mM HEPES buffer (pH 8) containing 50 mM NaCl. The purified Fab fragment of antibody 6RL#9 was eluted from the column for crystallization using the specified buffer solution. All actions on the column were carried out at a low temperature, ranging from 6 to 7.5°C.

(18-4) Crystallization of the Fab fragment of antibody 6RL#9 in the presence of calcium ions

Seed crystals of Fab fragment 6RL#9 were prepared in advance according to generally accepted conditions. Next, the purified Fab fragment of antibody 6RL#9 was adjusted by adding CaCl ₂ to 5 mM and concentrated to 12 mg/ml using a 5000 MWCO ultrafiltration membrane. The sample thus concentrated was then crystallized by the hanging drop method via vapor diffusion. A 100 mM HEPES buffer solution (pH 7.5) containing 20-29% PEG4000 was used as a reservoir solution. Seed crystals were crushed in 100 mM HEPES buffer solution (pH 7.5) containing 29% PEG4000 and 5 mM CaCl ₂ and diluted 100-10000 times, and 0.2 μl of each solution of this dilution series was added to the mixed solution 0. 8 µl of reservoir solution and 0.8 µl of concentrated sample on a coverslip to obtain crystallization drops. The crystallization drops were left to stand at 20°C for 2 days to 3 days. X-ray diffraction data were determined on the resulting crystals, similar to thin plates.

(18-5) Crystallization of the Fab fragment of antibody 6RL#9 in the absence of calcium ions

Purified Fab fragment of antibody 6RL#9 was concentrated to 15 mg/ml using a 5000 MWCO ultrafiltration membrane. The sample thus concentrated was then crystallized by the hanging drop method via vapor diffusion. A 100 mM HEPES buffer solution (pH 7.5) containing 18-25% PEG4000 was used as a reservoir solution. Crystals of the Fab fragment of antibody 6RL#9, obtained in the presence of Ca, were crushed in a 100 mM HEPES buffer solution (pH 7.5) containing 25% PEG4000, and diluted 100-10000 times, and 0.2 μl of each solution in this series dilutions were added to a mixed solution of 0.8 μL reservoir solution and 0.8 μL concentrated sample on a coverslip to obtain crystallization droplets. The crystallization drops were left to stand at 20°C for 2 days to 3 days. X-ray diffraction data were determined on the resulting crystals, similar to thin plates.

(18-6) Obtaining X-ray diffraction data on a crystal of the Fab fragment of antibody 6RL#9 obtained in the presence of calcium ions

One of the single crystals (obtained in the presence of Ca) of the Fab fragment of antibody 6RL#9, immersed in a 100 mM HEPES buffer solution (pH 7.5) containing 35% PEG4000 and 5 mM CaCl ₂ was scooped out, along with the external solution, using very small pin in the form of a nylon loop and frozen in liquid nitrogen. Frozen crystal X-ray diffraction data were analyzed using the BL-17A beamline from Photon Factory, Institute Materials Structure Science, High Energy Accelerator Research Organization (KEK). During the analysis, the frozen crystal was kept in a nitrogen stream at -178°C at all times to maintain its frozen state. A total of 180 diffraction images were acquired with a 1° crystal rotation for each image using a Quantum 315r CCD detector (Area Detector Systems Corporation (ADSC)) equipped with a beamline. Determination of the lattice constant, indexing of diffraction spots, and processing of diffraction data were carried out using the Xia2 program (CCP4 Software Suite), XDS Package (Wolfgang Kabsch), and Scala (CCP4 Software Suite). Finally, diffraction intensity data were obtained with a resolution down to 2.2 angstroms. This crystal belonged to the space group P2 ₁ 2 ₁ 2 ₁ and had lattice constants a=45.47 angstroms, b=79.8b angstroms, c=116.25 angstroms, α=90°, β=90° and γ=90 °.

(18-7) Obtaining X-ray diffraction data on a crystal of the Fab fragment of antibody 6RL#9 obtained in the absence of calcium ions

One of the single crystals (obtained in the presence of Ca) of the Fab fragment of antibody 6RL#9, immersed in a 100 mM HEPES buffer solution (pH 7.5) containing 35% PEG4000, was scooped out, along with the external solution, using a very small pin in the form nylon loop and frozen in liquid nitrogen. Frozen crystal X-ray diffraction data were analyzed using the BL-5A beamline from Photon Factory, Institute Materials Structure Science, High Energy Accelerator Research Organization (KEK). During the analysis, the frozen crystal was kept in a nitrogen stream at -178°C at all times to maintain its frozen state. A total of 180 diffraction images were acquired with a 1° crystal rotation for each image using a Quantum 210r CCD detector (Area Detector Systems Corporation (ADSC)) equipped with a beam line. Determination of the lattice constant, indexing of diffraction spots, and processing of diffraction data were carried out using the Xia2 program (CCP4 Software Suite), XDS Package (Wolfgang Kabsch), and Scala (CCP4 Software Suite). Finally, diffraction intensity data were obtained with resolution down to 2.3 angstroms. This crystal, which was the same type of crystal as the crystal obtained in the presence of Ca, belonged to the space group P2 ₁ 2 ₁ 2 ₁ and had lattice constants a=45.40 angstroms, b=79.63 angstroms, c= 116.07 angstroms, α=90°, β=90° and γ=90°.

(18-8) Structural analysis of a crystal of the Fab fragment of antibody 6RL#9 obtained in the presence of calcium ions

The crystal structure of the Fab fragment of antibody 6RL#9, obtained in the presence of Ca, was determined by the molecular replacement method using the Phaser program (CCP4 Software Suite). The number of molecules in the asymmetric element was estimated based on the size of the resulting crystal lattice and the molecular weight of the Fab fragment of antibody 6RL#9. Based on the homology of the primary sequence, amino acid residues at positions 112-220 in the Ive chain, positions 116-218 of the B chain, identified from the structural coordinate with PDB code: 1ZA6, were selected as model molecules to search for the CL and CH1 regions. Next, amino acid residues at positions 1–115 in chain B, extracted from the structural coordinate with PDB code: 1ZA6, were selected as a model molecule to search for the VH region. Finally, amino acid residues at positions 3-147 in the light chain, identified from the structural coordinate with PDB code: 2A9M, were selected as a model molecule to search for the VL region. From this order, the orientation and position of each model molecule were determined to be searched in the crystal lattice using the rotation and translation functions to obtain the initial structural model of the Fab fragment of antibody 6RL#9. The original structural model was subjected to rigid frame refinement by shifting each of the VH, VL, CH1 and CL domains to obtain a crystallographic confidence factor R of 46.9% and a free R value of 48.6% for reflectance data 25-3 .0 angstrom. In addition, the model was refined using model correction in the Coot iteration program (Paul Emsley) based on electron density maps with 2Fo-Fc and Fo-Fc coefficients calculated using structure refinement using Refmac5 (CCP4 Software Suite), an experimentally determined structural factor Fo, structure factor Fc calculated from this model, and phase. Final detailing was carried out using the Refmac5 program (CCP4 Software Suite) by including Ca ions and water molecules into the model based on electron density maps with coefficients 2Fo-Fc and Fo-Fc. Finally, a crystallographic confidence factor R of 20.0% and a free R value of 27.9% were obtained for the 3440 atom model using 21020 reflectance data at 25-2.2 angstrom resolution.

(18-9) Structural analysis of a crystal of the Fab fragment of antibody 6RL#9 obtained in the absence of calcium ions

The crystal structure of the Fab fragment of antibody 6RL#9 obtained in the absence of Ca was determined using a crystal structure that was the same type as the structure obtained in the presence of Ca. Water molecules and calcium ions Ca were excluded from the structural coordinates of the crystal of the Fab fragment of antibody 6RL#9, obtained in the presence of Ca, followed by detailing the rigid frame with the displacement of each of the VH, VL, CH1 and CL domains, obtaining a crystallographic confidence factor R, of 30.3%, and a free R value of 31.7% for reflectance data at 25-3.0 angstroms. In addition, the model was refined using model correction in the Coot iteration program (Paul Emsley) based on electron density maps with 2Fo-Fc and Fo-Fc coefficients calculated using structure refinement using Refmac5 (CCP4 Software Suite), an experimentally determined structural factor Fo, structure factor Fc calculated from this model, and phase. Final detailing was carried out using the Refmac5 program (CCP4 Software Suite) by including water molecules in the model based on electron density maps with coefficients 2Fo-Fc and Fo-Fc. Finally, a crystallographic confidence factor R of 20.9% and a free R value of 27.7% were obtained for the 3351 atom model using 18357 reflectance data at 25-2.3 angstrom resolution.

(18-10) Comparison of X-ray diffraction data between crystals of the Fab fragment of antibody 6RL#9 obtained in the presence of Ca and in the absence of Ca

As a result of a structural comparison of crystals of the Fab fragment of antibody 6RL#9 obtained in the presence of Ca and in the absence of Ca, a significant change in CDR3 of the heavy chain was observed. In fig. 11 shows the CDR3 structure of the heavy chain of the Fab fragment of antibody 6RL#9, determined by X-ray diffraction analysis. In particular, a calcium ion was present in the central region of the CDR3 loop portion of the heavy chain in a crystal of the Fab fragment of antibody 6RL#9 obtained in the presence of Ca. It was proposed that the calcium ion interacts with amino acid residues 95, 96 and 100a (defined by Kabat numbering) in the heavy chain CDR3. This suggests that in the presence of Ca, the heavy chain CDR3 loop, which is important for antigen binding, is stabilized by calcium binding, adopting a structure optimal for antigen binding. No previous reports have shown that calcium binds to CDR3 of the antibody heavy chain. This calcium binding antibody heavy chain CDR3 structure is a novel structure.

The calcium-binding motifs present in the heavy chain CDR3, which were identified from the structure of the Fab fragment of antibody 6RL#9, can also serve as new factors for constructing the Ca library, as described in Example 11. Although the calcium-binding motifs were introduced into light chain variable region in Example 11, another possible library contains, for example, the heavy chain CDR3 of antibody 6RL#9 and contains loose residues in other CDRs, including the light chain CDRs.

[Example 19] Preparation of an antibody that binds to the IL-6 receptor in a Ca-dependent manner from a human antibody library using phage display technology

(19-1) Preparation of a phage display library of naive human antibodies

A human antibody phage display library, consisting of a variety of phages displaying Fab domains having different human antibody sequences, was constructed in accordance with a method well known to those skilled in the art using poly-A RNA obtained from human PBMC, commercially available poly-A RNA. And human RNA, etc. as a matrix.

(19-2) Obtaining an antibody fragment that binds to an antigen in a Ca-dependent manner from a library by bead panning

The first round of screening of the constructed naive human antibody phage display library was carried out by increasing the content of only antibody fragments with the ability to bind antigen (IL-6R). The antigens used were biotin-labeled IL-6.

Phages were obtained from E. coli containing an engineered phagemid for phage display. 2.5 M NaCl/10% PEG was added to E. coli cultures that produced phages, and the pool of phages thus precipitated was diluted with TBS to obtain a phage library solution. Next, BSA and CaCl ₂ were added to the solution with the phage library (final concentration: 4% BSA and 1.2 mM calcium ions). The panning method was carried out in accordance with the conventional panning method using antigens immobilized on magnetic beads (J. Immunol. Methods. (2008) 332 (1-2), 2-9; J. Immunol. Methods. (2001) 247 (1 -2), 191-203; Biotechnol. Prog. (2002) 18 (2) 212-20; and Mol. Cell Proteomics (2003) 2 (2), 61-9). The magnetic beads used were neutravidin-coated beads (Sera-Mag SpeedBeads NeutrAvidin-coated) or streptavidin-coated beads (Dynabeads M-280 Streptavidin).

Specifically, 250 pmol of biotin-labeled antigens were added to the resulting phage library solution and thereby brought into contact with the phage library solution at room temperature for 60 minutes. BSA-blocked magnetic beads were added, and then the antigen-phage complexes were bound to the magnetic beads at room temperature for 15 minutes. The beads were washed three times with 1 ml 1.2 mM CaCl ₂ /TBST (TBST containing 1.2 mM CaCl ₂ ), and then washed additionally two times with 1 ml 1.2 mM CaCl ₂ /TBS (TBS containing 1.2 mM _CaCl2 ). After adding 0.5 ml of 1 mg/ml trypsin, the beads were suspended at room temperature for 15 minutes, and immediately thereafter the beads were separated using a magnetic stand to collect the phage solution. The collected phage solution was added to 10 ml of E. coli TG1 strain at logarithmic growth phase (OD600: 0.4-0.7). The E. coli strain was infected with phages by gently rotating the strain culture at 37°C for 1 hour. Infected E. coli were inoculated onto a 225 mm × 225 mm dish. Next, phages were collected from the inoculated E. coli cultures to obtain a solution with a phage library.

In the second and subsequent rounds of panning, the content of phages with Ca-dependent binding ability as an indicator was increased. Specifically, 40 pmol of biotin-labeled antigens were added to the resulting phage library solution and thereby brought into contact with the phage library solution at room temperature for 60 minutes. BSA-blocked magnetic beads were added, and then the antigen-phage complexes were bound to the magnetic beads at room temperature for 15 minutes. The beads were washed with 1 ml of 1.2 mM CaCl ₂ /TBST and 1.2 mM CaCl ₂ /TBS. The beads supplemented with 0.1 mL of 2 mM EDTA/TBS were then suspended at room temperature. Immediately after this, the granules were separated using a magnetic stand to collect the solution with phages. Addition of 5 μl of 100 mg/ml trypsin to the collected phage solution digested the pIII proteins (helper phage-derived pIII proteins) of non-Fab phages to eliminate the ability of non-Fab phages to infect E. coli. Phages collected from the trypsin-treated phage solution were added to 10 ml of E. coli strain TG1 at logarithmic growth phase (OD600: 0.4-0.7). The E. coli strain was infected with phages by gently rotating the strain culture at 37°C for 1 hour. Infected E. coli were inoculated onto a 225 mm × 225 mm dish. Next, phages were collected from the inoculated E. coli cultures to obtain a solution with a phage library. This panning, with Ca-dependent binding ability as an indicator, was carried out for a total of 3 rounds. (19-3) Evaluation by phage ELISA

Phage-containing culture supernatant was prepared according to a conventional method (Methods Mol. Biol. (2002) 178, 133-145) from each single colony of E. coli obtained using the method described above.

After addition of BSA and CaCl ₂ (final concentration: 4% BSA and 1.2 mM calcium ions), the phage-containing culture supernatant was subjected to ELISA using the following procedure: a StreptaWell 96 microtiter plate (F. Hoffmann-La Roche Ltd.) was covered for nights with 100 μl of PBS containing biotin-labeled antigens. Each well of the plate was washed with PBST to remove unbound antigens. The well was then blocked with 250 μl of 4% BSA-TBS for 1 hour or more. After removal of 4% BSA-TBS, the resulting supernatant was added to each well and the plate was allowed to stand at 37°C for 1 hour to allow phage-exposed antibodies to bind to the antigens contained in each well. Each well was washed with 1.2 mM CaCl ₂ /TBST, and 1.2 mM CaCl ₂ /TBS or 1 mM EDTA/TBS was added. The plate was left to stand at 37°C for 30 minutes for incubation. After washing with 1.2 mM CaCl ₂ /TBST, HRP-conjugated anti-M13 antibodies (Amersham Pharmacia Biotech Inc.) diluted in TBS containing 4% BSA and 1.2 mM ionized calcium were added to each well (all reported at final concentrations ). The plate was incubated for 1 hour. After washing with 1.2 mM CaCl ₂ /TBST, plain TMB solution (ZYMED Laboratories, Inc.) was added to the wells. The chromogenic reaction of the solution in each well was completed by adding sulfuric acid. Color development was then analyzed based on absorbance at 450 nm.

Antibody 6KC4-1#85, which has Ca-dependent binding to IL-6, was generated by ELISA using 96 isolated clones. The antibody fragment gene assessed to have Ca-dependent antigen binding capacity by phage ELISA was amplified as a template using specific primers and then analyzed for its nucleotide sequence. The heavy chain variable region sequence of the antibody 6KC4-1#85 is shown in SEQ ID NO: 11, and its light chain variable region sequence is shown in SEQ ID NO: 84. A polynucleotide encoding the heavy chain variable region (SEQ ID NO: 11) of the 6KC4 antibody -1#85 was linked by PCR to a polynucleotide encoding a sequence derived from IgG1. The resulting DNA fragment was inserted into expression vectors in animal cells to construct vectors that allow expression of the heavy chain corresponding to SEQ ID NO: 85. The polynucleotide encoding the light chain variable region (SEQ ID NO: 84) of antibody 6KC4-1#85 was bound by PCR with a polynucleotide encoding the natural constant region of the kappa chain (SEQ ID NO: 26). The resulting DNA fragment was inserted into vectors for expression in animal cells. The sequence of the resulting modified form was confirmed in a manner generally known to those skilled in the art. The sequence of the resulting modified form was confirmed in a manner generally known to those skilled in the art. (19-4)

Antibody expression and purification

The gene from clone 6KC4-1#85, assessed as having Ca-dependent antigen binding by phage ELISA, was inserted into plasmids for expression in animal cells. Antibody expression was performed as follows: Human embryonic kidney-derived Freestyle 293-F line (Invitrogen Corp.) was suspended in Freestyle 293 Expression Medium (Invitrogen Corp.). The suspension having a cell density of 1.33×10 ⁶ cells/ml was inoculated at a concentration of 3 ml/well into a 6-well plate. The resulting plasmids were transferred into cells by lipofection. Cells were cultured for 4 days in a CO ₂ incubator (37°C, 8% CO ₂ , 90 rpm). Antibodies were purified from the resulting culture supernatant in a manner well known to those skilled in the art using rProtein A Sepharose(TM) Fast Flow (Amersham Biosciences, Inc.). The absorbance of the purified antibody solution was measured at 280 nm using a spectrophotometer. The antibody concentration was calculated from the resulting measurement using the extinction coefficient calculated by PACE (Protein Science (1995) 4, 2411-2423).

[Example 20] Evaluation of Antibody 6KC4-1#85 for its Calcium Ion Binding

The calcium-dependent antigen binding antibody 6KC4-1#85, obtained from an antibody library, was evaluated for its calcium binding. An assessment of whether the measured Tm value changed at different ionized calcium concentrations was assessed in the manner described in Example 6.

Table 25 shows the Tm value for the Fab domain of antibody 6KC4-1#85. As shown in Table 25, the Tm value of the Fab domain of antibody 6KC4-1#85 varied with calcium ion concentration, demonstrating that antibody 6KC4-1#85 binds calcium.

[Example 21] Identification of calcium ion binding site in antibody 6KC4-1#85

As shown in Example 20, antibody 6KC4-1#85 was shown to bind to calcium ions, but did not possess the calcium-binding motifs identified in the hVk5-2 sequence study. Thus, to confirm that calcium ions were bound to the heavy or light chain of antibody 6KC4-1#85, or both, engineered antibodies containing either the heavy or light chain replaced by the corresponding chain of the anti-glypican 3 antibody (heavy chain sequence GC_H (SEQ ID NO: 48) and the light chain sequence GC_L (SEQ ID NO: 86)) incapable of binding to calcium ions were evaluated for their calcium ion binding. Table 26 shows the Tm values of the engineered antibodies measured in accordance with the method presented in Example 6. As a result, the Tm value of the engineered antibody having the heavy chain of the antibody 6KC4-1#85 varied depending on the calcium ion concentration, indicating that that antibody 6KC4-1#85 binds to calcium through its heavy chain.

To further identify the residue through which the heavy chain of the antibody 6KC4-1#85 binds to calcium ions, modified heavy chains were prepared (6_H1-11 (SEQ ID NO: 87), 6_H1-12 (SEQ ID NO: 88), 6_H1-13 (SEQ ID NO: 89), 6_H1-14 (SEQ ID NO: 90), and 6_H1-15 (SEQ ID NO: 91)) or modified light chains (6_L1-5 (SEQ ID NO: 92) and 6_L1-6 (SEQ ID NO: 93)) by replacing the Asp (D) residues present in the CDR of antibody 6KC4-1#85 with Ala (A) residues, which were unable to participate in the binding or chelation of calcium ions. The engineered antibodies were purified according to the method described in Example 19 from animal cell cultures transfected with expression vectors containing the engineered antibody genes. Calcium binding of the purified engineered antibodies was assayed in the manner described in Example 6. The results of the assay are presented in Table 27. As shown in Table 27, antibody 6KC4-1#85 lost its calcium binding ability when residue 95 or 101 (defined by Kabat numbering) was replaced. in the heavy chain CDR3 to an Ala residue, indicating that these residues are important for calcium binding. The calcium binding motifs present near the base of the heavy chain CDR3 loop in antibody 6KC4-1#85, which were identified based on the calcium binding property of the engineered antibody 6KC4-1#85, may also serve as novel factors for Ca library construction. as described in Example 11. Although calcium binding motifs were introduced into the light chain variable region in Example 11, another possible library contains, for example, the heavy chain CDR3 of the antibody 6KC4-1#85 and contains loose residues in other CDRs, including the light chain CDR.

[Example 22] Preparation of an antibody that binds to human IgA in a Ca-dependent manner

(22-1) Preparation of MRA-hIgA, GC-hIgA, and biotinylated human IgA-Fc

MRA-hIgA (heavy chain: SEQ ID NO: 97 and light chain: SEQ ID NO: 25), GC-hIgA (heavy chain: SEQ ID NO: 98 and light chain: SEQ ID NO: 99), and biotinylated IgA- Fc (also referred to as biotin-labeled hIgA-Fc; hIgA_CH2-CH3-Avitag: SEQ ID NO: 100) was prepared as human IgA as follows:

(22-1-1) Preparation of MRA-hIgA

Recombinant human IgA MRA-hIgA (hereinafter referred to as MRA-hIgA) was prepared as follows: expressed hIgA containing H(WT)-IgA1 (SEQ ID NO: 97) and L(WT) (SEQ ID NO: 25) was purified by the method , well known to those skilled in the art, using ion exchange chromatography and gel filtration.

(22-1-2) Preparation of GC-hIgA

Recombinant human IgA GC-hIgA was prepared as follows: a gene fragment encoding GC-hIgA (heavy chain: SEQ ID NO: 98 and light chain: SEQ ID NO: 99) was inserted into a vector for expression in animal cells. Freestyle 293 (Invitrogen Corp.) was cotransfected with the constructed plasmid vector and the gene encoding EBNA1 to be expressed using 293Fectin (Invitrogen Corp.). Cells transfected with these genes were then cultured at 37°C for 6 days in an atmosphere of 8% CO2 to secrete GC-hIgA proteins into the culture supernatant.

GC-hIgA-containing cell cultures were filtered through a 0.22-μm bottle neck filter to obtain culture supernatant. The culture supernatant was diluted with 20 mM Tris-HCl (pH 8.0) and applied to a HiTrap Q HP (GE Healthcare Bio-Sciences Corp.) equilibrated with this solution, and then GC-hIgA was eluted on a NaCl concentration gradient. Associates were then removed by gel filtration using Superdex 200 and buffer exchanged to 20 mM His-HCl and 150 mM NaCl (pH 6.0) to obtain purified GC-hIgA.

(22-1-3) Preparation of biotin-labeled hIgA-Fc

To add biotin to the C-terminus of the protein of interest (human IgA-Fc), a gene fragment encoding a specific sequence (Avitag sequence) for biotin ligase-mediated biotinylation was added downstream of the gene fragment encoding the human IgA-Fc region. A gene fragment encoding human IgA protein linked to the Avitag sequence (hIgA_CH2-CH3-Avitag (SEQ ID NO: 100)) was inserted into vectors for expression in animal cells. Freestyle 293 (Invitrogen Corp.) was transfected with constructed plasmid vectors using 293Fectin (Invitrogen Corp.). This transfection was performed simultaneously with the gene encoding EBNA1 to be expressed and the gene encoding the biotin ligase (BirA) to be expressed for biotinylation to tag the protein with biotin. Cells transfected with these genes according to the procedures described above were cultured at 37°C for 6 days in an atmosphere of 8% CO ₂ to secrete the protein of interest into the culture supernatant.

Cell cultures containing human IgA-Fc of interest were filtered through a 0.22-μm bottleneck filter to obtain culture supernatant. The culture supernatant was diluted with 20 mM Tris-HCl (pH 7.4) and applied to a HiTrap Q HP (GE Healthcare Bio-Sciences Corp.) equilibrated with this solution, and then the human IgA-Fc of interest was eluted on a NaCl concentration gradient. The HiTrap Q HP eluate was diluted with 50 mM Tris-HCl (pH 8.0) and applied to a SoftLink Avidin column equilibrated with this solution, followed by elution with 5 mM biotin, 150 mM NaCl, and 50 mM Tris-HCl (pH 8. 0). Associates were then removed by gel filtration using Superdex 200 and buffer exchanged to 20 mM His-HCl and 150 mM NaCl (pH 6.0) to obtain purified human IgA-Fc.

(22-2) Preparation of an antibody fragment that binds an antigen in a Ca-dependent manner from a library using bead panning

The first round of screening of the constructed Ca-dependent antigen binding library (Ca-library) was performed with antigen binding ability (human IgA-Fc) as an indicator.

Phages were obtained from E. coli containing an engineered phagemid for phage display. 2.5 M NaCl/10% PEG was added to E. coli cultures that produced phages, and the pool of phages thus precipitated was diluted with TBS to obtain a phage library solution. Next, BSA or skim milk and CaCl ₂ (final concentration: 4% BSA and 1.2 mM calcium ions; or 3% skim milk and 1.2 mM calcium ions) were added to the solution with the phage library, obtaining a solution with a blocked phage library. The panning method was carried out in accordance with the conventional panning method using antigens immobilized on magnetic beads (J. Immunol. Methods. (2008) 332 (1-2), 2-9; J. Immunol. Methods. (2001) 247 (1 -2), 191-203; Biotechnol. Prog. (2002) 18 (2) 212-20; and Mol. Cell Proteomics (2003) 2 (2), 61-9). The magnetic beads used were neutravidin-coated beads (Sera-Mag SpeedBeads NeutrAvidin-coated) or streptavidin-coated beads (Dynabeads M-280 Streptavidin).

Specifically, 250 pmol of biotin-labeled antigens (biotin-labeled IgA-Fc) was added to the resulting phage library solution and thereby brought into contact with the phage library solution at room temperature for 60 minutes. BSA or skim milk blocked magnetic beads were added, and then the antigen-phage complexes were bound to the magnetic beads at room temperature for 15 minutes. The beads were washed with 1 ml of 1.2 mM CaCl ₂ /TBST (TBS containing 1.2 mM CaCl ₂ and 0.1% Tween 20) and 1 ml of 1.2 mM CaCl ₂ /TBS (TBS containing 1.2 mM CaCl ₂ ). After adding 0.5 ml of 1 mg/ml trypsin, the beads were suspended at room temperature for 15 minutes, and immediately thereafter the beads were separated using a magnetic stand to collect the phage solution. The collected phage solution was added to 10 ml of E. coli strain ER2738 at logarithmic growth phase (OD600: 0.4-0.7). The E. coli strain was infected with phages by gently rotating the strain culture at 37°C for 1 hour. Infected E. coli were inoculated onto a 225 mm × 225 mm dish. Next, phages were collected from the inoculated E. coli cultures to obtain a solution with a phage library.

In the second and third rounds of panning, the content of phages was increased with the ability of Ca ion concentration-dependent antigen binding as an indicator. Specifically, 40 pmol of biotin-labeled antigens were added to the phage library solution obtained by blocking in the same manner as in the first round of penning, and thereby brought into contact with the phage library solution at room temperature for 60 minutes. BSA-blocked magnetic beads were added, and then the antigen-phage complexes were bound to the magnetic beads at room temperature for 15 minutes. The beads were washed with 1 ml of 1.2 mM CaCl ₂ /TBST and 1.2 mM CaCl ₂ /TBS. The beads supplemented with 0.1 ml of 2 mM EDTA/TBS (TBS containing 2 mM EDTA) were then suspended at room temperature. Immediately after this, the granules were separated using a magnetic stand to collect the solution with phages. Addition of 5 μl of 100 mg/ml trypsin to the collected phage solution digested the pIII proteins (helper phage-derived pIII proteins) of non-Fab phages to eliminate the ability of non-Fab phages to infect E. coli. The collected phage solution was added to 10 ml of E. coli strain ER2738 at logarithmic growth phase (OD600: 0.4-0.7). The E. coli strain was infected with phages by gently rotating the strain culture at 37°C for 1 hour. Infected E. coli were inoculated onto a 225 mm × 225 mm dish. Next, phages were collected from the inoculated E. coli cultures to obtain a solution with a phage library.

(22-3) Screening for Human IgA Binding Antibody Using Biacore

Antibody fragment genes extracted from E. coli-derived phagemids obtained after completion of the second round of panning were inserted into vectors for expression in animal cells. Antibody expression was performed as follows: the human embryonic kidney-derived cell line Freestyle 293-F (Invitrogen Corp.) was suspended in Freestyle 293 Expression Medium (Invitrogen Corp.). The suspension having a cell density of 2.63×10 ⁵ cells/ml was inoculated at a concentration of 190 μl/well into a 9b-well plate. The resulting plasmids were transferred into cells by lipofection. Cells were cultured for 4 in a CO ₂ incubator (37°C, 8% CO ₂ ).

The culture supernatant obtained as described above was used to assay GC-hIgA binding capacity using Biacore A100. Antibodies in the culture supernatant were immobilized onto a CM5 sensor chip (GE Healthcare Bio-Sciences Corp.) immobilized with Protein A (Invitrogen Corp.) in a suitable amount using the amine coupling method. The running buffer was a buffer solution containing 20 mM ACES, 150 mM NaCl, 0.05% (w/v) Tween 20, and 1.2 mM CaCl ₂ (pH 7.4). GC-hIgA was also diluted in this buffer. All analysis was performed at 25°C.

Antibody genes assessed to have the ability to bind GC-hIgA by the Biacore IgG binding ability assay were amplified using specific primers from the expression vectors in the animal cells used to express them and then analyzed for their nucleotide sequences.

(22-4) Screening of IgA binding antibody using phage ELISA

Phage-containing culture supernatant was obtained by phage culture according to the conventional method (Methods Mol. Biol. (2002) 178, 133-145) from each single colony of E. coli obtained after the third round of panning carried out at stage (22-2 ). After addition of BSA and CaCl ₂ (final concentration: 4% BSA and 1.2 mM calcium ions), the phage-containing culture supernatant was subjected to ELISA using the following procedure: a StreptaWell 96 microtiter plate (F. Hoffmann-La Roche Ltd.) was covered for overnight, 100 μl of PBS containing biotin-labeled IgA-Fc. Each well of the plate was washed with PBST to remove unbound antigens. The well was then blocked with 250 μl of 4% BSA-TBS for 1 hour or more. After removing 4% BSA-TBS, the resulting supernatant was added to each well and the plate was left at 37°C for 1 hour to allow phage-exposed antibodies to bind to the IgA-Fc contained in each well. Each well was washed with 1.2 mM CaCl ₂ /TBST, and 1.2 mM CaCl ₂ /TBS or 1 mM EDTA/TBS was added. The plate was left to stand at 37°C for 30 minutes for incubation. After washing with 1.2 mM CaCl ₂ /TBST, HRP-conjugated anti-M13 antibodies (Amersham Pharmacia Biotech Inc.) diluted in TBS containing 4% BSA and 1.2 mM ionized calcium were added to each well (all reported at final concentrations ). The plate was incubated for 1 hour. After washing with 1.2 mM CaCl ₂ /TBST, plain TMB solution (ZYMED Laboratories, Inc.) was added to the wells. The chromogenic reaction of the solution in each well was completed by adding sulfuric acid. Color development was then analyzed based on absorbance at 450 nm.

Clones assessed as having Ca2-dependent IgA-Fc binding capacity by phage ELISA were analyzed for the nucleotide sequences of their antibody fragment genes.

(22-5) Obtaining an antibody fragment that binds an antigen in a Ca-dependent manner from a library using bead panning

The first round of screening of the constructed Ca-dependent antigen binding library (Ca-library) was performed with antigen binding ability (human IgA-Fc) as an indicator.

Phages were obtained from E. coli containing an engineered phagemid for phage display. 2.5 M NaCl/10% PEG was added to E. coli cultures that produced phages, and the pool of phages thus precipitated was diluted with TBS to obtain a phage library solution. Next, BSA or skim milk and CaCl ₂ (final concentration: 4% BSA and 1.2 mM calcium ions; or 3% skim milk and 1.2 mM calcium ions) were added to the solution with the phage library, obtaining a solution with a blocked phage library. In the first round of panning, the panning method was carried out in accordance with the conventional panning method using antigens immobilized on magnetic beads (J. Immunol. Methods. (2008) 332 (1-2), 2-9; J. Immunol. Methods. (2001 ) 247 (1-2), 191-203; Biotechnol. Prog. (2002) 18 (2) 212-20; and Mol. Cell Proteomics (2003) 2 (2), 61-9). The magnetic beads used were neutravidin-coated beads (Sera-Mag SpeedBeads NeutrAvidin-coated) or streptavidin-coated beads (Dynabeads M-280 Streptavidin).

Specifically, 250 pmol of biotin-labeled antigens (biotin-labeled IgA-Fc) was added to the resulting phage library solution and thereby brought into contact with the phage library solution at room temperature for 60 minutes. Skim milk-blocked magnetic beads were added, and then the antigen-phage complexes were bound to the magnetic beads at room temperature for 15 minutes. The beads were washed with 1 ml 1.2 mM CaCl ₂ /TBST (TBS containing 1.2 mM CaCl ₂ and 0.1% Tween 20) and 1 ml 1.2 mM CaCl ₂ /TBS (TBS containing 1.2 mM CaCl ₂ ). After adding 0.5 ml of 1 mg/ml trypsin, the beads were suspended at room temperature for 15 minutes, and immediately thereafter the beads were separated using a magnetic stand to collect the phage solution. The collected phage solution was added to 10 ml of E. coli strain ER2738 at logarithmic growth phase (OD600: 0.4-0.7). The E. coli strain was infected with phages by gently rotating the strain culture at 37°C for 1 hour. Infected E. coli were inoculated onto a 225 mm × 225 mm dish. Next, phages were collected from the inoculated E. coli cultures to obtain a solution with a phage library.

The second round of panning was performed using antigens immobilized on the plate. Specifically, biotin-labeled antigens were added at a concentration of 5 pmol/well to a streptavidin-coated 10-well plate (StreptaWell, F. Hoffmann-La Roche Ltd.) and contacted at room temperature for 60 minutes. The plate was then washed three times with TBST (TBS containing 0.1% Tween 20), resulting in a plate with antigen immobilized on it. A phage library blocked with skim milk containing 1.2 mM Ca was added to it and thereby brought into contact with antigens at room temperature for 60 minutes. The plate was washed three times with 1.2 mM CaCl ₂ /TBST using a plate washer (Skan WASHER, SKARON). The plate was then further immersed in 2 L of 1.2 mM CaCl ₂ /TBST and gently shaken for 60 minutes. The phages in each well, supplemented with 0.1 ml of 2 mM EDTA/TBS (TBS containing 2 mM EDTA), were suspended at room temperature, and then the phage solution was collected. Addition of 5 μl of 100 mg/ml trypsin to the collected phage solution digested the pIII proteins (helper phage-derived pIII proteins) of non-Fab phages to eliminate the ability of non-Fab phages to infect E. coli. The collected phage solution was added to 10 ml of E. coli strain ER2738 at logarithmic growth phase (OD600: 0.4-0.7). The E. coli strain was infected with phages by gently rotating the strain culture at 37°C for 1 hour. Infected E. coli were inoculated onto a 225 mm × 225 mm dish.

(22-6) Screening of hIgA-binding antibody using phage ELISA

Phage-containing culture supernatant was obtained by phage culture according to the conventional method (Methods Mol. Biol. (2002) 178, 133-145) from each single colony of E. coli obtained after panning. Phage ELISA was performed as described for step (22-4). Genes from clones assessed to have Ca-dependent antigen binding were analyzed for their nucleotide sequences, then inserted into animal cell expression vectors and expressed as antibodies, which were then purified.

[Example 23] Evaluation of the resulting antibody with respect to its ability to Ca-dependent binding to human IgA

(23-1) Expression and purification of the resulting anti-human IgA antibody

Among the resulting antibodies assessed as having the ability to bind human IgA in Example 22, antibodies GA1-IgG1 (obtained in step (22-3); heavy chain: SEQ ID NO: 101 and light chain: SEQ ID NO: 102), GA2- IgG1 (obtained at the stage

(22-4); heavy chain: SEQ ID NO: 103 and light chain: SEQ ID NO: 104), GA3-IgG1 (prepared in step (22-6); heavy chain: SEQ ID NO: 105 and light chain: SEQ ID NO: 106) , and GA4-IgG1

(obtained at step (22-3); heavy chain: SEQ ID NO: 107 and light chain: SEQ ID NO: 108) were expressed using the following method and then purified: line Freestyle 293-F, derived from human embryonic kidney, (Invitrogen Corp.) was suspended in Freestyle 293 Expression Medium (Invitrogen Corp.). The suspension having a cell density of 1.33×10 ⁶ cells/ml was inoculated at a concentration of 3 ml/well into a 6-well plate. The resulting plasmids were transfected into cells by lipofection. The cells were cultured for 4 days in a CO ₂ incubator (37°C, 8% CO ₂ , 90 rpm. Antibodies were purified from the resulting culture supernatant in a manner well known to those skilled in the art, using rProtein A Sepharose(TM) Fast Flow (Amersham Biosciences, Inc.).

The absorbance of the purified antibody solution was measured at 280 nm using a spectrophotometer. The antibody concentration was calculated from the resulting measurement using the extinction coefficient calculated by PACE (Protein Science (1995) 4, 2411-2423).

(23-2) Evaluation of the resulting antibody with respect to its Ca-dependent IgA binding ability

Antibodies (GA1-IgG1, GA2-IgG1, GA3-IgG1 and GA4-IgG1) obtained in step (23-1) were assessed for their human IgA binding activity (dissociation constant K _D (M)) using Biacore T200 ( GE Healthcare Bio-Sciences Corp.). Binding activity was assayed using 0.05% Tween 20, 20 mmol/L ACES and 150 mmol/L NaCl supplemented with 3 µM or 1.2 mM CaCl ₂ (pH 7.4 and pH 5.8) or 0.05 % Tween 20, 20 mmol/L ACES and 150 mmol/L NaCl (pH 8.0) with the addition of 0.1 μM or 10 mM CaCl ₂ as running buffer.

Each antibody was coupled to recombinant protein A/G (Thermo Fisher Scientific K.K.) immobilized in the appropriate amount onto a CM5 sensor chip (GE Healthcare Bio-Sciences Corp.) using the amine coupling method.

An appropriate concentration of MRA-hIgA (described in step (22-1)) was injected onto it as an analyte and reacted with the antibody on the sensor chip. Then, 10 mmol/L glycine-HCl (pH 1.5) was injected to regenerate the sensor chip. The analysis was carried out at 37°C. The dissociation constant K _D (M) was calculated using curve fitting analysis and equilibrium value analysis based on the results of the analysis using Biacore T200 Evaluation Software (GE Healthcare Bio-Sciences Corp.). The results are presented in Table 28. The obtained sensorgrams are also presented in Fig. 12. As clearly shown, GA2-IgG1, GA3-IgG1 and GA4-IgG1 bind strongly to human IgA at a Ca ²⁺ concentration of 1.2 mM, but weakly bind to human IgA at a Ca 2+ concentration of 3 μM.

Although antibodies whose interaction with antigens (IL6 receptors) varied depending on the concentration of Ca ions were obtained from the Ca library according to Example 14, it was discovered that antibodies that bind to antigens in a manner dependent on the concentration of Ca ions can be obtained not only using IL6 receptors, but also using human IgA.

[Example 24] Preparation of an antibody that binds to human glypican 3 (GPC3) in a Ca-dependent manner

(24-1) Preparation of glypican 3 people

Recombinant human glypican 3 (hereinafter referred to as GPC3) for use as an antigen was prepared as follows: the culture supernatant was collected from CHO cells stably transfected with plasmids to express the transmembrane region-free amino acid sequence of human glypican 3 linked to 6 histidine residues (SEQ ID NO: 109). The resulting culture supernatant was purified by ion exchange chromatography, then subjected to His tag affinity purification, and purified by gel filtration to obtain GPC3. GPC3 was labeled with biotin using EZ-Link NHS-PEG4-Biotin (Thermo Fisher Scientific K.K.) to generate biotin-labeled GPC3.

(24-2) Preparation of an antibody fragment that binds an antigen in a Ca-dependent manner from a library using bead panning

The first round of screening of the constructed antibody library with Ca-dependent GPC3 binding was performed with antigen binding ability (GPC3) as an indicator.

Phages were obtained from E. coli containing an engineered phagemid for phage display. 2.5 M NaCl/10% PEG was added to E. coli cultures that produced phages, and the pool of phages thus precipitated was diluted with TBS to obtain a phage library solution. Next, BSA or skim milk and CaCl ₂ (final concentration: 4% BSA and 1.2 mM calcium ions; or 3% skim milk and 1.2 mM calcium ions) were added to the solution with the phage library, obtaining a solution with a blocked phage library. The panning method was carried out in accordance with the conventional panning method using antigens immobilized on magnetic beads (J. Immunol. Methods. (2008) 332 (1-2), 2-9; J. Immunol. Methods. (2001) 247 (1 -2), 191-203; Biotechnol. Prog. (2002) 18 (2) 212-20; and Mol. Cell Proteomics (2003) 2 (2), 61-9). The magnetic beads used were neutravidin-coated beads (Sera-Mag SpeedBeads NeutrAvidin-coated) or streptavidin-coated beads (Dynabeads M-280 Streptavidin).

Specifically, 250 pmol of biotin-labeled antigens (biotin-labeled GPC3) was added to the resulting phage library solution and thereby brought into contact with the phage library solution at room temperature for 60 minutes. BSA or skim milk blocked magnetic beads were added, and then the antigen-phage complexes were bound to the magnetic beads at room temperature for 15 minutes. The beads were washed with 1 ml of 1.2 mm CaCl ₂ /TBST (TBS containing 1.2 mm CaCl ₂ ) and 1 ml of 1.2 mm CaCl ₂ /TBS (TBS containing 1.2 mm CaCl ₂ ). After adding 0.5 ml of 1 mg/ml trypsin, the beads were suspended at room temperature for 15 minutes, and immediately thereafter the beads were separated using a magnetic stand to collect the phage solution. The collected phage solution was added to 10 ml of E. coli strain ER2738 at logarithmic growth phase (OD600: 0.4-0.7). The E. coli strain was infected with phages by gently rotating the strain culture at 37°C for 1 hour. Infected E. coli were inoculated onto a 225 mm × 225 mm dish. Next, phages were collected from the inoculated E. coli cultures to obtain a solution with a phage library.

In the second and subsequent rounds of panning, the content of phages with Ca-dependent antigen binding ability as an indicator was increased. Specifically, 40 pmol of biotin-labeled antigens were added to the phage library solution obtained by blocking in the same manner as in the first round of penning, and thereby brought into contact with the phage library solution at room temperature for 60 minutes. BSA or skim milk blocked magnetic beads were added, and then the antigen-phage complexes were bound to the magnetic beads at room temperature for 15 minutes. The beads were washed with 1 ml of 1.2 mM CaCl ₂ /TBST and 1.2 mM CaCl ₂ /TBS. The beads supplemented with 0.1 ml of 2 mM EDTA/TBS (TBS containing 2 mM EDTA) were then suspended at room temperature. Immediately after this, the granules were separated using a magnetic stand to collect the solution containing phages. Addition of 5 μl of 100 mg/ml trypsin to the collected phage solution digested the pIII proteins (helper phage-derived pIII proteins) of non-Fab phages to eliminate the ability of non-Fab phages to infect E. coli. The collected phage solution was added to 10 ml of E. coli strain ER2738 at logarithmic growth phase (OD600: 0.4-0.7). The E. coli strain was infected with phages by gently rotating the strain culture at 37°C for 1 hour. Infected E. coli were inoculated onto a 225 mm × 225 mm dish. Next, phages were collected from the inoculated E. coli cultures to obtain a solution with a phage library.

(24-3) Screening of GPC3 binding antibody using phage ELISA

Phage-containing culture supernatant was obtained by phage culture according to the conventional method (Methods Mol. Biol. (2002) 178, 133-145) from each single colony of E. coli obtained after the second and third rounds of panning carried out in the manner described above.

After addition of BSA and CaCl ₂ (final concentration: 4% BSA and 1.2 mM calcium ions), the phage-containing culture supernatant was subjected to ELISA using the following procedure: a StreptaWell 96 microtiter plate (F. Hoffmann-La Roche Ltd.) was covered for nights with 100 μl of PBS containing biotin-labeled antigens. Each well of the plate was washed with PBST (PBS containing 0.1% Tween 20) to remove unbound antigens. The well was then blocked with 250 μl of 4% BSA-TBS for 1 hour or more. After removal of 4% BSA-TBS, the resulting supernatant was added to each well and the plate was allowed to stand at 37°C for 1 hour to allow phage-exposed antibodies to bind to the antigens contained in each well. Each well was washed with 1.2 mM CaCl ₂ /TBST, and 1.2 mM CaCl ₂ /TBS or 1 mM EDTA/TBS was added. The plate was left to stand at 37°C for 30 minutes for incubation. After washing with 1.2 mM CaCl ₂ /TBST, HRP-conjugated anti-M13 antibodies (Amersham Pharmacia Biotech Inc.) diluted in TBS containing 4% BSA and 1.2 mM ionized calcium were added to each well (all reported at final concentrations ). The plate was incubated for 1 hour. After washing with 1.2 mM CaCl ₂ /TBST, plain TMB solution (ZYMED Laboratories, Inc.) was added to the wells. The chromogenic reaction of the solution in each well was completed by adding sulfuric acid. Color development was then analyzed based on absorbance at 450 nm.

Antibody fragment genes assessed to have Ca-dependent antigen binding capacity by phage ELISA were amplified as a template using specific primers and then analyzed for their nucleotide sequences.

(24-4) Expression and purification of human GPC3-binding antibody

The four antibody genes assessed to have Ca-dependent antigen binding by phage ELISA, i.e. CSCM-01_005 (heavy chain sequence: 110 and light chain sequence: 111), CSCM-01_009 (heavy chain sequence: 112 and light chain sequence: 113), CSCM-01_015 (heavy chain sequence: 114 and light chain sequence: 115) , and CSCM-01_023 (heavy chain sequence: 116 and light chain sequence: 117), and anti-human GPC3 antibodies GC-IgG1 (heavy chain sequence: 118 and light chain sequence: 119) as a control were separately inserted into expression plasmids in animal cells. Antibody expression was performed as follows: Human embryonic kidney-derived Freestyle 293-F line (Invitrogen Corp.) was suspended in Freestyle 293 Expression Medium (Invitrogen Corp.). The suspension having a cell density of 1.33×10 ⁶ cells/ml was inoculated at a concentration of 3 ml/well into a 6-well plate. The resulting plasmids were transferred into cells by lipofection. Cells were cultured for 4 days in a CO ₂ incubator (37°C, 8% CO ₂ , 90 rpm). Antibodies were purified from the resulting culture supernatant in a manner well known to those skilled in the art using rProtein A Sepharose(TM) Fast Flow (Amersham Biosciences, Inc.). The absorbance of the purified antibody solution was measured at 280 nm using a spectrophotometer. The antibody concentration was calculated from the resulting measurement using the extinction coefficient calculated by PACE (Protein Science (1995) 4, 2411-2423). GC-IgG1 antibody was purified as described above from the culture supernatant of CHO cells stably expressing GC-IgG1 antibody, and its concentration was calculated.

(24-5) Evaluation of the resulting antibody for Ca-dependent binding to human GPC3

The resulting antibodies were subjected to ELISA using the following procedure: a StreptaWell 96 microtiter plate (F. Hoffmann-La Roche Ltd.) was coated overnight with 100 μl of PBS containing biotin-labeled antigens. Each well of the plate was washed with ACES buffer (10 mM ACES, 150 mM NaCl, 100 mM CaCl ₂ and 0.05% Tween 20 (pH 7.4)) to remove unbound antigens. The well was then blocked with 250 μl ACES buffer containing 2% BSA for 1 hour or more. After removing the ACES buffer containing 2% BSA, 100 μl of each of the 4-fold serial dilutions of purified IgG, pre-diluted from 10 μg/ml, was added to each well and the plate was allowed to stand for 1 hour to allow the IgG to bind to the antigens contained in each hole.

Each well was washed with ACES buffer, and “10 mM ACES, 150 mM NaCl and 1.2 mM CaCl ₂ (pH 7.4)”, “10 mM ACES, 150 mM NaCl and 3 µM CaCl ₂ (pH 7.4)” were added. ", "10 mM ACES, 150 mM NaCl and 1.2 mM CaCl ₂ (pH 5.8)" or "10 mM ACES, 150 mM NaCl and 3 µM CaCl ₂ (pH 5.8)". The plate was left to stand at 37°C for 30 minutes for incubation. After washing with ACES buffer, HRP-conjugated anti-human IgG antibodies were added to each well.

(BioSource International, Inc.) diluted in ACES buffer containing 2% BSA. The plate was incubated for 1 hour. After washing with ACES buffer, plain TMB solution (ZYMED Laboratories, Inc.) was added to the wells. The chromogenic reaction of the solution in each well was completed by adding sulfuric acid. Color development was then analyzed based on absorbance at 450 nm.

The results of the analysis are presented in Fig. 13. Absorbance was constant for GC-IgG1, regardless of calcium ion concentration, while absorbance was significantly lower for CSCM-01_005, CSCM-01_009, CSCM-01_015 and CSCM-01_023 at 3 µM calcium ion concentration (low concentration calcium ions) than at a calcium ion concentration of 1.2 mM (high calcium ion concentration). From these results, it was shown that CSCM-01_005, CSCM-01_009, CSCM-01_015, and CSCM-01_023 have antigen binding property that varies with calcium ion concentration, indicating that calcium-dependent antibody may also be obtained against glypican 3 people.

[Example 25] Preparation of an antibody that binds to mouse IgA in a pH-dependent manner

(25-1) Preparation of GC-mIgA and biotinylated mouse IgA-Fc

GC-mIgA (heavy chain: SEQ ID NO: 120 and light chain: SEQ ID NO: 121) and biotinylated mouse IgA-Fc (also referred to as biotin-labeled mIgA-Fc; mIgA_CH2-CH3-Avitag: SEQ ID NO: 122) prepared as mouse IgA as follows:

(25-1-1) Preparation of GC-mIgA

Recombinant mouse IgA GC-mIgA was prepared as follows: a gene fragment encoding GC-mIgA (heavy chain: SEQ ID NO: 120 and light chain: SEQ ID NO: 121) was inserted into a vector for expression in animal cells. Freestyle 293 (Invitrogen Corp.) was cotransfected with the constructed plasmid vector and the gene encoding EBNA1 to be expressed using 293Fectin (Invitrogen Corp.). Cells transfected with these genes were then cultured at 37°C for 4 days in an atmosphere of 8% CO ₂ to secrete GC-mIgA proteins into the culture supernatant.

GC-mIgA-containing cell cultures were filtered through a 0.22-μm bottle neck filter to obtain a culture supernatant. The culture supernatant was diluted in 20 mM Tris-HCl (pH 8.0) and applied to a HiTrap Q HP (GE Healthcare Bio-Sciences Corp. .), equilibrated with this solution, and then GC-mIgA was eluted on a NaCl concentration gradient. Associates were then removed by gel filtration using Superdex 200 and buffer exchanged to 20 mM His-HCl and 150 mM NaCl (pH 6.0) to obtain purified GC-mIgA.

(25-1-2) Preparation of biotin-labeled mIgA-Fc

To add biotin to the C-terminus of the protein of interest (mouse IgA-Fc), a gene fragment encoding the specific sequence (Avitag sequence) for biotin ligase-mediated biotinylation was appended downstream of the gene fragment encoding the mouse IgA-Fc region. A gene fragment encoding the mouse IgA protein linked to the Avitag sequence (mIgA_CH2-CH3-Avitag (SEQ ID NO: 122)) was inserted into vectors for expression in animal cells. Freestyle 293 (Invitrogen Corp.) was transfected with constructed plasmid vectors using 293Fectin (Invitrogen Corp.). This transfection was performed simultaneously with the gene encoding EBNA1 to be expressed and the gene encoding the biotin ligase (BirA) to be expressed for biotinylation to tag the protein with biotin. Cells transfected with these genes according to the procedures described above were cultured at 37°C for 6 days in an atmosphere of 8% CO ₂ to secrete the protein of interest into the culture supernatant.

Cell cultures containing mouse IgA-Fc of interest were filtered through a 0.22-μm bottleneck filter to obtain culture supernatant. The culture supernatant was diluted with 20 mM Tris-HCl (pH 7.4) and applied to a HiTrap Q HP (GE Healthcare Bio-Sciences Corp.) equilibrated with this solution, and then the mouse IgA-Fc of interest was eluted on a NaCl concentration gradient. Next, the HiTrap Q HP eluate was diluted with 50 mM Tris-HCl (pH 8.0) and applied to a SoftLink Avidin column equilibrated with this solution, and then the mouse IgA-Fc of interest was eluted with 5 mM biotin, 150 mM NaCl, and 50 mM Tris-HCl (pH 8.0). Associates were then removed by gel filtration using Superdex 200 and the resulting mouse IgA-Fc buffer was replaced with 20 mM His-HCl and 150 mM NaCl (pH 6.0) to obtain purified mouse IgA-Fc.

(25-2) Preparation of an antibody fragment that binds an antigen in a pH-dependent manner from a His library using bead panning

The first round of screening of the constructed pH-dependent antigen binding library (His library 1) was performed with antigen binding ability (mouse IgA-Fc) as an indicator.

Phages were obtained from E. coli containing an engineered phagemid for phage display. 2.5 M NaCl/10% PEG was added to E. coli cultures that produced phages, and the pool of phages thus precipitated was diluted with TBS to obtain a phage library solution. Next, skim milk (final concentration: 3% skim milk) was added to the phage library solution to obtain a phage library blocked solution. The panning method was carried out in accordance with the conventional panning method using antigens immobilized on magnetic beads (J. Immunol. Methods. (2008) 332 (1-2), 2-9; J. Immunol. Methods. (2001) 247 (1 -2), 191-203; Biotechnol. Prog. (2002) 18 (2) 212-20; and Mol. Cell Proteomics (2003) 2 (2), 61-9). The magnetic beads used were neutravidin-coated beads (Sera-Mag SpeedBeads NeutrAvidin-coated) or streptavidin-coated beads (Dynabeads M-280 Streptavidin).

Specifically, biotin-labeled antigens (biotin-labeled mIgA-Fc) were added to the resulting phage library solution and thereby brought into contact with the phage library solution at room temperature for 60 minutes. Biotin-labeled antigens were used in panning at 250 pmol for the first round, 40 pmol for the second round, and 10 pmol for the third round. After adding milk-blocked magnetic beads, antigen-phage complexes were bound to the magnetic beads at room temperature for 15 minutes. The beads were washed with 1 ml 1.2 mM CaCl ₂ /TBST (TBS containing 1.2 mM CaCl ₂ and 0.1% Tween 20 (pH 7.6)) and 1 ml 1.2 mM CaCl ₂ /TBS (TBS, containing 1.2 mM CaCl ₂ (pH 7.6)). After adding 0.5 ml of 1 mg/ml trypsin, the beads were suspended at room temperature for 15 minutes, and immediately thereafter the beads were separated using a magnetic stand to collect the phage solution. The collected phage solution was added to 10 ml of E. coli strain ER2738 at logarithmic growth phase (OD600: 0.4-0.7). The E. coli strain was infected with phages by gently rotating the strain culture at 37°C for 1 hour. Infected E. coli were inoculated onto a 225 mm × 225 mm dish. Next, phages were collected from the inoculated E. coli cultures to obtain a solution with a phage library. This panning was carried out for only 3 rounds.

(25-3) Evaluation of a mouse IgA-binding antibody for its binding ability using Biacore

Antibody fragment genes extracted from E. coli-derived phagemids obtained after completion of the third round of panning were inserted into vectors for expression in animal cells. Antibody expression was performed as follows: the human embryonic kidney-derived cell line Freestyle 293-F (Invitrogen Corp.) was suspended in Freestyle 293 Expression Medium (Invitrogen Corp.). The suspension having a cell density of 2.63×10 ⁵ cells/ml was inoculated at a concentration of 190 μl/well into a 96-well plate. The resulting plasmids were transferred into cells by lipofection. Cells were cultured for 4 in a CO ₂ incubator (37°C, 8% CO ₂ ).

The culture supernatant obtained as described above was used for GC-mIgA binding capacity assay using Biacore A100. Antibodies in the culture supernatant were captured on a CM5 sensor chip (GE Healthcare Bio-Sciences Corp.) with protein A/G (Invitrogen Corp.) immobilized on it in a suitable amount by the amine coupling method. Two types of buffer solutions were used as running buffers: 20 mM ACES, 150 mM NaCl, 0.05% (w/v) Tween 20 and 1.2 mM CaCl ₂ (pH 7.4); and 20 mM ACES, 150 mM NaCl, 0.05% (w/v) Tween 20, and 1.2 mM CaCl ₂ (pH 5.8). Interactions in the antigen-antibody reaction were analyzed under neutral pH and acidic pH conditions. GC-mIgA was also diluted with each of these buffers and injected at a flow rate of 10 μl/min for 60 seconds to interact the antigen with the antibodies captured on the sensor chip. Then, 10 mM glycine-HCl (pH 1.5) was injected at a flow rate of 30 μm/min for 30 seconds to regenerate the sensor chip. All analysis was performed at 25°C.

As a result of binding assessment using Biacore, mIAP1 B3-3_#024 (heavy chain: SEQ ID NO: 123 and light chain: SEQ ID NO: 124), mIAP1 B3-3_#130 (heavy chain: SEQ ID NO: 125 and light chain: SEQ ID NO: 126), and mIAP1 B3-3_#230 (heavy chain: SEQ ID NO: 127 and light chain: SEQ ID NO: 128) were obtained as antibodies assessed as having pH-dependent binding ability with GC-mIgA. In fig. Figure 14 shows sensograms for these antibodies at pH 7.4 and pH 5.8. It was found that the ability of these antibodies to bind mouse IgA decreased when the pH of the buffer was changed from pH 7.4 to pH 5.8.

The genes for these antibodies from the animal cell expression vectors used for their expression were analyzed for their nucleotide sequences.

Although antibodies whose interaction with antigens (IL-6R) varied depending on pH were obtained from His library 1 in Example 3, it was discovered that antibodies that bind to antigens in a pH-dependent manner can be obtained not only from using IL-6R, but also using mouse IgA.

[Example 26] Preparation of an antibody that binds to human HMGB1 in a pH-dependent manner

(26-1) Preparation of Human HMGB1 and Biotinylated Human HMGB1 Human HMGB1 (SEQ ID NO: 129) and biotinylated human HMGB1-Avi (also referred to as biotin-labeled hHMGB1; hHMGB1-Avitag: SEQ ID NO: 130) were prepared as follows:

(26-1-1) Preparation of human HMGB1

Recombinant human HMGB1 hHMGB1 was prepared as follows: a gene fragment encoding hHMGB1 (SEQ ID NO: 129) was inserted into an expression vector in animal cells. Freestyle 293 (Invitrogen Corp.) was cotransfected with the constructed plasmid vector and the gene encoding EBNA1 to be expressed using 293Fectin (Invitrogen Corp.). Cells transfected with these genes were then cultured at 37°C in an atmosphere of 8% CO ₂ to secrete hHMGB1 proteins into the culture supernatant. The thus obtained culture supernatant was applied to HiTrap SP Sepharose HP (GE Healthcare Bio-Sciences Corp.), pre-equilibrated with PBS , and the column was washed with PBS, and then the proteins adsorbed on the column were eluted on a linear sodium chloride concentration gradient. After equilibration with 20 mM histidine-HCl (pH 5.0), the eluted fraction containing hHMGB1, diluted 3-fold with this buffer solution, was applied to the column. The column was washed with this buffer, and then the proteins adsorbed on the column were eluted on a linear sodium chloride concentration gradient. The hHMGB1-containing fraction was concentrated using an ultrafiltration membrane and then applied to a Superdex 200 column (GE Healthcare Bio-Sciences Corp.) equilibrated with 300 mM NaCl and 20 mM histidine (pH 6.0). The purified hHMGBl fraction was obtained by separation using the same buffer solution.

(26-1-2) Preparation of biotin-labeled human HMGB1

To add biotin to the C-terminus of the protein of interest (human HMGB1), a gene fragment encoding a specific sequence (Avitag sequence) for biotin ligase-mediated biotinylation was appended downstream of the gene fragment encoding the human HMGB1 region. A gene fragment encoding the human HMGB1 protein linked to the Avitag sequence (hHMGB1-Avitag (SEQ ID NO: 130)) was inserted into vectors for expression in animal cells. Freestyle 2 93 (Invitrogen Corp.) was transfected with constructed plasmid vectors using 293Fectin (Invitrogen Corp.). This transfection was performed simultaneously with the gene encoding EBNA1 to be expressed and the gene encoding the biotin ligase (BirA) to be expressed for biotinylation to tag the protein with biotin. Cells transfected with these genes according to the procedures described above were cultured at 37°C in an atmosphere of 8% CO ₂ to secrete the protein of interest into the culture supernatant.

The culture supernatant thus obtained was applied to HiTrap SP Sepharose HP (GE Healthcare Bio-Sciences Corp.) pre-equilibrated with PBS, and the column was washed with PBS, and then the proteins adsorbed on the column were eluted on a linear sodium chloride concentration gradient. The fraction containing HMGB1-Avi diluted 2-fold in 100 mM Tris-HCl (pH 7.4) was applied to a SoftLink Soft Release Avidin Resin column (Promega Corp.) equilibrated with TBS. The column was washed with TBS, and then the proteins adsorbed on the column were eluted using TBS (pH 8.0) containing 5 mM biotin. The fraction containing HMGB1-Avi was concentrated through an ultrafiltration membrane and then applied to a Superdex 200 column (GE Healthcare Bio-Sciences Corp.) equilibrated with 300 mM NaCl and 20 mM histidine (pH 6.0). The purified biotinylated HMGB1-Avi fraction was obtained by separation using the same buffer solution.

(26-2) Preparation of an antibody fragment that binds an antigen in a pH-dependent manner from a His library using bead panning

The first round of screening of the constructed pH-dependent antigen binding library (His library 1) was performed with antigen binding ability (human HMGB1) as an indicator.

Phages were obtained from E. coli containing an engineered phagemid for phage display. 2.5 M NaCl/10% PEG was added to E. coli cultures that produced phages, and the pool of phages thus precipitated was diluted with TBS to obtain a phage library solution. Next, BSA and CaCl ₂ (final concentration: 4% BSA and 1.2 mM calcium ions) were added to the solution with the phage library, obtaining a solution with a blocked phage library. The panning method was carried out in accordance with the conventional panning method using antigens immobilized on magnetic beads (J. Immunol. Methods. (2008) 332 (1-2), 2-9; J. Immunol. Methods. (2001) 247 (1 -2), 191-203; Biotechnol. Prog. (2002) 18 (2) 212-20; and Mol. Cell Proteomics (2003) 2 (2), 61-9). The magnetic beads used were neutravidin-coated beads (Sera-Mag SpeedBeads NeutrAvidin-coated) or streptavidin-coated beads (Dynabeads M-280 Streptavidin).

Specifically, 250 pmol of biotin-labeled antigens (biotin-labeled hHMGB1) was added to the resulting phage library solution and thereby brought into contact with the phage library solution at room temperature for 60 minutes. BSA-blocked magnetic beads were added, and then the antigen-phage complexes were bound to the magnetic beads at room temperature for 15 minutes. The beads were washed with 1 ml 1.2 mM CaCl ₂ /TBST (50 mM Tris, 300 mM NaCl, 1.2 mM CaCl ₂ and 0.1% Tween 20 (pH 7.6)) and 1 ml 1.2 mM CaCl ₂ /TBS (50 mM Tris, 300 mM NaCl, and 1.2 mM CaCl ₂ (pH 7.6)). After adding 0.5 ml of 1 mg/ml trypsin, the beads were suspended at room temperature for 15 minutes, and immediately thereafter the beads were separated using a magnetic stand to collect the phage solution. The collected phage solution was added to 10 ml of E. coli strain ER2738 at logarithmic growth phase (OD600: 0.4-0.7). The E. coli strain was infected with phages by gently rotating the strain culture at 37°C for 1 hour. Infected E. coli were inoculated onto a 225 mm × 225 mm dish. Next, phages were collected from the inoculated E. coli cultures to obtain a solution with a phage library.

In the second and subsequent rounds of panning, the content of phages with antigen-binding ability or pH-dependent binding ability as an indicator was increased. Specifically, 40 pmol of biotin-labeled antigens were added to the resulting phage library solution and thereby brought into contact with the phage library solution at room temperature for 60 minutes. BSA-blocked magnetic beads were added, and then the antigen-phage complexes were bound to the magnetic beads at room temperature for 15 minutes. The beads were washed several times with 1 ml of 1.2 mM CaCl ₂ /TBST and 1.2 mM CaCl ₂ /TBS. To increase the content with antigen binding capacity as an indicator, beads supplemented with 0.5 ml of 1 mg/ml trypsin were suspended at room temperature for 15 minutes, and immediately thereafter the beads were separated using a magnetic stand to collect the phage solution. To increase the content with pH-dependent antigen binding capacity as an indicator, 0.4 ml (round 2) or 0.1 ml (round 3 or beyond) 50 mM MES/1.2 mM CaCl ₂ /150 mM was added to the beads NaCl (pH 5.5) and the resulting granules were suspended at room temperature. Then the granules were separated using a magnetic stand, obtaining a solution with phages. Adding 20 µl (round 2) or 5 µl (round 3 or beyond) 100 mg/ml trypsin to the collected phage solution digested the pIII proteins (helper phage-derived pIII proteins) of non-Fab-exposing phages to eliminate the ability of non-Fab-exposing phages infect with E. coli. The collected phages were added to 10 ml of E. coli strain ER2738 at logarithmic growth phase (OD600: 0.4-0.7). The E. coli strain was infected with phages by gently rotating the strain culture at 37°C for 1 hour. Infected E. coli were inoculated onto a 225 mm x 225 mm plate. Next, phages were collected from the inoculated E. coli cultures to obtain a solution with a phage library. This panning, with antigen-binding ability or pH-dependent binding ability as an indicator, was carried out for a total of 3 rounds.

(26-3) Screening for human HMGB1 binding using phage ELISA

Phage-containing culture supernatant was obtained by phage culture according to the conventional method (Methods Mol. Biol. (2002) 178, 133-145) from each single colony of E. coli obtained after the third and fourth rounds of panning carried out at step (26 -2). After addition of BSA and CaCl ₂ (final concentration: 4% BSA and 1.2 mM calcium ions), the phage-containing culture supernatant was subjected to ELISA using the following procedure: a StreptaWell 96 microtiter plate (F. Hoffmann-La Roche Ltd.) was covered for four hours or more than 100 μl of PBS containing biotin-labeled hHMGBl. Each well of the plate was washed with PBST to remove unbound antigens. The well was then blocked with 250 μl of 4% BSA-TBS (50 mM Tris, 300 mM NaCl, 2.5 mM CaCl ₂ and 4% BSA) for 1 hour or more. After removing 4% BSA-TBS, the resulting supernatant was added to each well and the plate was left at 37°C for 1 hour to allow phage-exposed antibodies to bind to the HMGB1 contained in each well. Each well was washed with 1.2 mM CaCl ₂ /TBST (pH 7.6) and added 1.2 mM CaCl ₂ /TBS (pH 7.6) or 1.2 mM CaCl ₂ /150 mM NaCl/50 mM MES (pH 5.5). The plate was left to stand at 37°C for 30 minutes for incubation. After washing with 1.2 mM CaCl ₂ /TBST, HRP-conjugated anti-M13 antibody (Amersham Pharmacia Biotech Inc.) diluted in 4% BSA-TBS was added to each well. The plate was incubated for 1 hour. After washing with 1.2 mM CaCl ₂ /TBST (pH 7.6), plain TMB solution (Invitrogen Corp.) was added to the wells. The chromogenic reaction of the solution in each well was completed by adding sulfuric acid. Color development was then analyzed based on absorbance at 450 nm.

Antibody fragment genes contained in clones assessed to have pH-varying ability to bind human HMGB1 based on phage ELISA results were analyzed for their nucleotide sequences.

(26-4) Expression and purification of an antibody that binds to human HMGB1

The genes of the three antibodies assessed as having pH-dependent antigen binding by phage ELISA, i.e. HM_3_2_R_017 (heavy chain: SEQ ID NO: 131 and light chain: SEQ ID NO: 132), HM_3_2_R_054 (heavy chain: SEQ ID NO: 133 and light chain: SEQ ID NO: 134) and HM_4_1_R_001 (heavy chain: SEQ ID NO: 134) : 135 and light chain: SEQ ID NO: 136) were individually inserted into plasmids for expression in animal cells. Expression of these antibodies was carried out in the following manner: the human embryonic kidney-derived line Freestyle 2 93-F (Invitrogen Corp.) was suspended in Freestyle 293 Expression Medium (Invitrogen Corp.). The suspension having a cell density of 1.33×10 ⁶ cells/ml was inoculated at a concentration of 3 ml/well into a 6-well plate. The resulting plasmids were transferred into cells by lipofection. Cells were cultured for 4 days in a CO ₂ incubator (37°C, 8% CO ₂ , 90 rpm). Antibodies were purified from the resulting culture supernatant in a manner well known to those skilled in the art using rProtein A Sepharose(TM) Fast Flow (Amersham Biosciences, Inc.). The absorbance of the purified antibody solution was measured at 280 nm using a spectrophotometer. The antibody concentration was calculated from the resulting measurement using the extinction coefficient calculated by PACE (Protein Science (1995) 4, 2411-2423).

(26-5) Evaluation of the resulting antibody for its pH-dependent binding ability to human HMGB1

To evaluate the pH dependence of the hHMGBl binding activity of antibodies HM_3_2_R_017 (heavy chain: SEQ ID NO: 131 and light chain: SEQ ID NO: 132), HM_3_2_R_054 (heavy chain: SEQ ID NO: 133 and light chain: SEQ ID NO: 134) and HM_4_1_R_001 (heavy chain: SEQ ID NO: 135 and light chain: SEQ ID NO: 136) obtained in step (26-4), these antibodies were analyzed for their interaction with hHMGB1 using Biacore T100 (GE Healthcare Bio-Sciences Corp.). Antigen-antibody interactions were analyzed in pH 7.4 and pH 5.8 solutions as neutral pH and acidic pH conditions, respectively. Approximately 200 RU of each antibody of interest was captured on a CM4 sensor chip (GE Healthcare Bio-Sciences Corp.) with Protein A/G (Pierce Biotechnology Inc.) immobilized thereon in an appropriate amount by the amine coupling method. Two types of buffer solutions were used as running buffers: 20 mM ACES, 150 mM NaCl, 0.05% (w/v) Tween 20 and 1.2 mM CaCl ₂ (pH 7.4); and 20 mM ACES, 150 mM NaCl, 0.05% (w/v) Tween 20, and 1.2 mM CaCl ₂ (pH 5.8). hHMGB1 was also diluted in each of these buffers. All analysis was performed at 25°C.

In the antigen-antibody interaction assay using antibody HM_3_2_R_017, antibody HM_3_2_R_054, and antibody HM_4_1_R_001, diluted hHMGB1 solution or empty running buffer was injected at a flow rate of 10 μL/min for 60 seconds to capture the antigen onto the sensor chip. Antibody HM_3_2_R_017, antibody HM_3_2_R_054 and antibody HM_4_1_R_001 were reacted with hHMGBl. Then, 10 mM glycine-HCl (pH 1.5) was injected at a flow rate of 30 μL/min for 30 seconds to regenerate the sensor chip. In fig. 15 shows sensograms of these antibodies analyzed as described above at pH 7.4 and pH 5.8.

From these results, it was revealed that the ability of antibody HM_3_2_R_017, antibody HM_3_2_R_054 and antibody HM_4_1_R_001 to bind hHMGB1 decreased when the pH of the buffer was changed from pH 7.4 to pH 5.8, indicating that antibodies with pH-dependent binding can also be produced against human HMGB1.

[Reference Example 1] Evaluation of a Ca-dependent binding antibody for its effect on plasma antigen retention time using a normal mouse

(1-1) In vivo test using a normal mouse hsIL-6R (soluble human IL-6 receptor; prepared according to Reference Example 4) was administered separately or simultaneously with each anti-human IL-6 receptor antibody to each normal mouse (C57BL/ 6J, Charles River Laboratories Japan, Inc.). hsIL-6R and anti-human IL-6 receptor antibody were then evaluated for their in vivo pharmacokinetics. A solution of hsIL-6R (5 μg/ml) or a mixed solution of hsIL-6R and anti-human IL-6 receptor antibody was administered in a single dose of 10 ml/kg into the tail vein. The anti-human IL-6 receptor antibody used was H54/L28-IgG1, 6RL#9-IgG1 or FH4-IgG1, described above.

The mixed solution had a fixed hsIL-6R concentration of 5 μg/ml, but had an anti-human IL-6 receptor antibody concentration that varied among antibodies: 0.1 mg/ml for H54/L28-IgG1 and 10 mg/ml for 6RL#9- IgG1 and FH4-IgG1. In this case, an excess amount of anti-human IL-6 receptor antibody was present in the mixed solution to be sufficient to bind hsIL-6R. Thus, a large majority of hsIL-6R antigens appeared to be bound by antibodies. Blood was taken 15 minutes, 7 hours, 1 day, 2 days, 4 days, 7 days, 14 days, 21 days and 28 days after administration. The collected blood was immediately centrifuged at 12,000 rpm at 4°C for 15 minutes to obtain plasma. The separated plasma was stored in a freezer set at -20°C or lower until used in the assay.

(1-2) Measurement of anti-human IL-6 receptor antibody concentration in normal mouse plasma by ELISA

The concentration of anti-human IL-6 receptor antibody in mouse plasma was determined by ELISA. First, the F(ab')2 fragment anti-human IgG antibody (γ-chain specific) (Sigma-Aldrich Corp.) was dispensed into Nunc-Immuno Plate, MaxiSoup (Nalge nunc International) and allowed to stand overnight at 4 °C to obtain plates with anti-human IgG antibody in the solid phase. Calibration curve samples having plasma concentrations of 0.64, 0.32, 0.16, 0.08, 0.04, 0.02, or 0.01 μg/mL and mouse plasma assay samples diluted in 100 times or more, spread into a plate with anti-human IgG antibody on the solid phase, and then incubated at 25°C for 1 hour. The plate was then allowed to react with biotinylated anti-human IL-6R antibody (R&D Systems, Inc.) at 25°C for 1 hour. Then the reaction was carried out with streptavidin-PolyHRP80 (Stereospecific Detection Technologies GmbH) at 25°C for 0.5 hour. TMB One Component HRP Microwell Substrate (BioFX Laboratories Inc.) was used as a substrate in the chromogenic reaction. The chromogenic reaction was stopped by adding 1 N sulfuric acid (Showa Chemical Industry Co., Ltd.). The absorbance of the color solution was then measured at 450 nm using a microplate reader. Plasma antibody concentrations in mice were calculated from the absorbance values of the calibration curve using SOFTmax PRO analytical software (Molecular Devices, LLC). In fig. 16 shows the change in the concentrations of antibodies H54/L28-IgG1, 6RL#9-IgG1 and FH4-IgG1 measured by this method in the plasma of normal mice after intravenous administration.

(1-3) Measurement of hsIL-6R concentration in plasma by electrochemiluminescence

The concentration of hsIL-6R in mouse plasma was determined by electrochemiluminescence. hsIL-6R calibration curve samples prepared at 2000, 1000, 500, 250, 125, 62.5, or 31.25 pg/mL and a mouse plasma assay sample diluted 50-fold or higher were reacted at 4 °C overnight with a mixed solution of anti-human IL-6R monoclonal antibody (R&D Systems, Inc.), ruthenium-labeled with SULFO-TAG NHS Ester (Meso Scale Discovery), biotinylated anti-human IL-6R antibody (R&D Systems, Inc. .) and tocilizumab solution (heavy chain SEQ ID NO: 24, light chain SEQ ID NO: 25). The free Ca concentration in the samples was reduced to allow virtually all of the hsIL-6R antigens in the samples to dissociate from 6RL#9-IgG1 or FH4-IgG1 and then associate with the added tocilizumab. For this purpose, the assay buffer contained 10 mM EDTA. The reaction solution was then dispensed into an MA400 PR Streptavidin Plate (Meso Scale Discovery). After a subsequent reaction at 25°C for 1 hour, each well of the plate was washed, and then buffer T (×4) (Meso Scale Discovery) was dispensed into each well. Immediately thereafter, the reaction solution was analyzed using a SECTOR PR 400 reader (Meso Scale Discovery). The concentration of hsIL-6R was calculated from the response of the calibration curve using SOFTmax PRO analysis software (Molecular Devices, LLC). In fig. 17 shows the change in hsIL-6R concentration measured by this method in the plasma of normal mice after intravenous administration.

As a result, hsIL-6R alone was cleared from the blood quickly, while its simultaneous administration with the conventional H54/L28-IgG1 antibody, which does not have Ca-dependent binding of hsIL-6R, significantly prolonged the clearance of hsIL-6R. In contrast, its co-administration with the antibody 6RL#9-IgG1 or FH4-IgG1, which has 100-fold or higher Ca-dependent hsIL-6R binding, significantly accelerated the clearance of hsIL-6R. Antibodies 6RL#9-IgG1 and FH4-IgG1, simultaneously administered with hsIL-6R, reduced the concentration of hsIL-6R in plasma after 1 day by 3–9 times and 2 times, respectively, compared with H54/L28-IgG1, simultaneously administered with them. This demonstrated that calcium-dependent binding antibody could accelerate the clearance of antigens from plasma.

[Reference Example 2] Study on improving the antigen clearance accelerating effect of an antibody with Ca-dependent antigen binding (antibody production)

(2-1) Binding of FcRn to IgG antibody

The IgG antibody has a longer plasma retention time as a result of FcRn binding. Binding between IgG and FcRn is observed only under acidic conditions (pH 6.0), and is practically undetectable under neutral conditions (pH 7.4). The IgG antibody is taken up into cells in a nonspecific manner, but is recycled to the cell surface by binding to endosomal FcRn under the acidic conditions of endosomes, and dissociates from FcRn under neutral plasma conditions. IgG that has lost FcRn binding under acidic conditions as a result of mutation in the Fc region is no longer recycled into the plasma from the endosome, resulting in a significantly reduced plasma retention time of the antibody.

Previously described methods for increasing the plasma retention time of IgG antibodies involve increasing FcRn binding under acidic conditions. An amino acid substitution is introduced into the Fc region of an IgG antibody to increase its binding to FcRn under acidic conditions, thereby increasing the efficiency of recycling of the IgG antibody into plasma from the endosome. As a result, the retention time of the IgG antibody in plasma increases. When making a substitution, it is considered important not to increase binding to FcRn under neutral conditions. It is believed that the plasma retention time of such an IgG antibody binding to FcRn under neutral conditions is rather reduced, since the IgG antibody can be recycled to the cell surface through FcRn binding under acidic conditions in the endosome, but it cannot dissociate from FcRn under neutral conditions. conditions in plasma, nor recycled into plasma.

For example, as described in Dall' Acqua et al. (J. Immunol. (2002) 169 (9), 5171-5180), an IgG1 antibody that is rendered capable of binding to mouse FcRn under neutral conditions (pH 7.4) due to amino acid substitution is reported to exhibit impaired retention time in plasma when administered to mice. As described by Yeung et al. (J. Immunol. (2009) 182 (12), 7663-7671), Datta-Mannan et al. (J. Biol. Chem. (2007) 282 (3), 1709-1717), and Dall' Acqua et al. (J. Immunol. (2002) 169 (9), 5171-5180), it was confirmed that an engineered IgG1 antibody having an improved ability to bind to human FcRn under acidic conditions (pH 6.0) as a result of amino acid substitution also binds with human FcRn under neutral conditions (pH 7.4). As reported, the antibody administered to a cynomolgus monkey showed no improvement in plasma retention time or change in plasma retention time. Conventional antibody engineering techniques to improve antibody function have focused on increasing the plasma retention time of antibodies by enhancing binding to human FcRn under acidic conditions without enhancing binding to human FcRn under neutral conditions (pH 7.4). In particular, none of the previous reports has mentioned the benefits of an IgG1 antibody having enhanced binding to human FcRn under neutral conditions (pH 7.4) due to an amino acid substitution in the Fc region.

Antibodies with Ca-dependent antigen binding are highly beneficial due to their effects of accelerating the clearance of soluble antigens and binding a single antibody molecule to soluble antigens multiple times. A method of enhancing FcRn binding under neutral conditions (pH 7.4) was investigated to further improve this effect of accelerating antigen clearance.

(2-2) Preparation of a Ca-dependent human IL-6 receptor binding antibody having FcRn binding activity under neutral conditions

The amino acids in the Fc regions of FH4-IgG1 and 6RL#9-IgG1, which have calcium-dependent antigen-binding ability, and H54/L28-IgG1, used as a control that does not have calcium-dependent antigen-binding ability, were mutated to give modified forms that bind FcRn under neutral conditions (pH 7.4). The amino acid mutation was introduced by PCR using a method well known to those skilled in the art. Specifically, FH4-N434W (heavy chain: SEQ ID NO: 94 and light chain: SEQ ID NO: N81), 6RL#9-N434W (heavy chain: SEQ ID NO: 95 and light chain: SEQ ID NO: 79) and H54/L28-N434W (heavy chain: SEQ ID NO: 96 and light chain: SEQ ID NO: 83) were prepared by replacing the Asn at amino acid 434 defined by EU numbering in the IgG1 heavy chain constant region with Trp. Animal cell expression vectors having a polynucleotide insert encoding each of these amino acid substitution variants were prepared using the QuikChange site-directed mutagenesis kit (Stratagene Corp.) according to the method described in the manual supplied with the kit. Expression and purification of the antibody, and measurement of concentration were carried out in the manner described in Example 15.

[Reference example 3]

Evaluation of a Ca-dependent binding antibody for its extinction accelerating effect using a normal mouse

(3-1) In vivo test using a normal mouse hsIL-6R (soluble human IL-6 receptor; prepared according to Reference Example 4) was administered separately or simultaneously with each anti-human IL-6 receptor antibody to each normal mouse (C57BL/6J mouse , Charles River Laboratories Japan, Inc.). hsIL-6R and anti-human IL-6 receptor antibody were then evaluated for their in vivo kinetics. A solution of hsIL-6R (5 μg/ml) or a mixed solution of hsIL-6R and anti-human IL-6 receptor antibody was administered in a single dose of 10 ml/kg into the tail vein. The anti-human IL-6 receptor antibody used was H54/L28-N434W, 6RL#9-N434W, or FH4-N434W, described above.

The mixed solution had a fixed hsIL-6R concentration of 5 μg/ml, but had an anti-human IL-6 receptor antibody concentration that varied among antibody types: 0.042 mg/ml for H54/L28-N434W, 0.55 mg/ml for 6RL#9 -N434W and 1 mg/ml for FH4-N434W. In this case, anti-human IL-6 receptor antibody was present in excess amount sufficient for hsIL-6R. Thus, a large majority of hsIL-6R antigens appeared to be bound by antibodies. Blood was taken 15 minutes, 7 hours, 1 day, 2 days, 4 days, 7 days, 14 days, 21 days and 28 days after administration. The collected blood was immediately centrifuged at 12,000 rpm at 4°C for 15 minutes to obtain plasma. The separated plasma was stored in a freezer set at -20°C or lower until used in the assay.

(3-2) Measurement of anti-human IL-6 receptor antibody concentration in normal mouse plasma by ELISA

The concentration of anti-human IL-6 receptor antibody in mouse plasma was measured using ELISA in the same manner as in Reference Example 1. FIG. 18 shows the change in the concentrations of antibodies H54/L28-N434W, 6RL#9-N434W and FH4-N434W measured by this method, in the plasma of normal mice after intravenous administration.

(3-3) Measurement of hsIL-6R concentration in plasma by electrochemiluminescence

The concentration of hsIL-6R in mouse plasma was determined by electrochemiluminescence. hsIL-6R calibration curve samples prepared at 2000, 1000, 500, 250, 125, 62.5, or 31.25 pg/mL and mouse plasma assay samples diluted 50-fold or higher were reacted at 4 °C overnight with a mixed solution of anti-human IL-6R monoclonal antibody (R&D Systems, Inc.), ruthenium-labeled with SULFO-TAG NHS Ester (Meso Scale Discovery) and biotinylated anti-human IL-6R antibody (R&D Systems, Inc. .). The free Ca concentration in the samples was reduced to allow virtually all of the hsIL-6R antigens in the samples to dissociate from 6RL#9-N434W or FH4-N434W to be in the free form. For this purpose, the assay buffer contained 10 mM EDTA. The reaction solution was then dispensed into an MA400 PR Streptavidin Plate (Meso Scale Discovery). After a subsequent reaction at 25°C for 1 hour, each well of the plate was washed, and then buffer T (×4) (Meso Scale Discovery) was dispensed into each well. Immediately thereafter, the reaction solution was analyzed using a SECTOR PR 400 reader (Meso Scale Discovery). The concentration of hsIL-6R was calculated from the response of the calibration curve using SOFTmax PRO analysis software (Molecular Devices, LLC). In fig. 19 shows the change in hsIL-6R concentration measured by this method in the plasma of normal mice after intravenous administration.

As a result, simultaneous administration of hsIL-6R with the antibody H54/L28-N434W, which has FcRn binding activity at pH 7.4, but not Ca-dependent binding activity with hsIL-6R, significantly slowed down the disappearance of hsIL-6R compared with the administration of hsIL-6R separately. In contrast, coadministration with antibody 6RL#9-N434W or antibody FH4-N434W, which has 100-fold or more Ca-dependent hsIL-6R binding as well as FcRn binding at pH 7.4, accelerated the disappearance of hsIL-6R, compared with with administration of hsIL-6R alone. Antibody 6RL#9-N434W and antibody FH4-N434W, administered simultaneously with hsIL-6R, reduced the concentration of hsIL-6R in plasma by 3 times and 8 times within one day after administration, respectively, compared with hsIL-6R administered alone . These results demonstrated that a calcium-dependent antigen-binding antibody with FcRn-binding activity at pH 7.4 could further accelerate the clearance of antigens from plasma.

It was confirmed that the 6RL#9-IgG1 antibody or the FH4-IgG1 antibody, which has 100-fold or more Ca-dependent hsIL-6R binding activity, has the effect of increasing the disappearance of hsIL-6R, compared with the H54/L28-IgG1 antibody. lacking Ca-dependent binding of hsIL-6R. It was confirmed that antibody 6RL#9-N434W or antibody FH4-N434W, which has 100-fold or more Ca-dependent hsIL-6R binding activity and also FcRn binding at pH 7.4, accelerates the disappearance of hsIL-6R compared with hsIL -6R, introduced separately. These data indicate that, as with pH-dependent antibody binding, antibody with Ca-dependent antibody binding dissociates from the antigen in the endosome.

[Reference Example 4] Preparation of Soluble Human IL-6 Receptor (hsIL-6R)

Recombinant human IL-6 receptor antigen was prepared as follows: a CHO cell line stably expressing soluble human IL-6 receptor (hereinafter referred to as hsIL-6R), consisting of the N-terminal amino acid sequence at positions 1 to 357, as described by Mullberg et al. (J. Immunol. (1994) 152, 4958-4968) were constructed in a manner well known to those skilled in the art. The expression line was cultured to express hsIL-6R. hsIL-6R was purified from the resulting culture supernatant by Blue Sepharose 6 FF column chromatography and gel filtration column chromatography. The fraction eluted as the main peak in the final step was used as the final purified product.

[Reference Example 5] NMR analysis of an antibody containing a human hVkl sequence having a calcium ion binding motif for its calcium ion binding activity

(5-1) Antibody expression and purification

Antibody containing LfVk1_Ca and antibody containing LfVk1 were expressed for NMR analysis and purified. Specifically, animal cells were transiently transfected with animal cell expression plasmids prepared to allow expression of the heavy chain (SEQ ID NO: 24) and light chain (SEQ ID NO: 43) of an antibody containing LfVk1_Ca (also referred to as LfVk1_Ca antibody ). Also, animal cells were transiently transfected with animal cell expression plasmids prepared to provide appropriate expression of the heavy chain (SEQ ID NO: 24) and light chain (SEQ ID NO: 44) of an antibody containing LfVk1 (also referred to as LfVk1 antibody ). Amino acids for labeling were added to 100 ml of a cell suspension of the human embryonic kidney-derived cell line, Freestyle 293-F (Invitrogen Corp.), suspended at a final cell density of 1 x 10 ⁶ cells/ml in Freestyle 293 Expression Medium (Invitrogen Corp.). In particular, for the Asp/Glu/Gln/Asn label, L-aspartic acid - ¹³ C ₄ , ¹⁵ N (10 mg), L-glutamic acid - ¹³ C ₅ , ¹⁵ N (2.5 mg), L-glutamine - ¹³ C ₅ , ¹⁵ N ₂ (60 mg), L-asparagine- ¹³ C ₄ , ¹⁵ N ₂ ⋅H ₂ O (2.5 mg) and β-chloro-L-alanine (6 mg) were suspended in 10 ml water, and the resulting solution was filtered through a 0.22-μm filter and then added to the cell suspension. For Leu label, L-leucine- ^15N (30 mg) and β-chloro-L-alanine (6 mg) were suspended in 10 ml water and the resulting solution was filtered through a 0.22 μm filter and then added to the suspension cells. The resulting plasmids were transfected into cells using the lipofection method. Cells transfected with plasmids were cultured for 5 days in a CO ₂ incubator (37°C, 8% CO ₂ , 90 rpm). Antibodies were purified from the resulting culture supernatant in a manner well known to those skilled in the art using rProtein A Sepharose(TM) Fast Flow (Amersham Biosciences, Inc.). The absorbance of the purified antibody solution was measured at 280 nm using a spectrophotometer. The antibody concentration was calculated from the resulting measurement using the extinction coefficient calculated by PACE (Protein Science (1995) 4, 2411-2423).

(5-2) Preparation of Fab fragment

Each antibody was concentrated to 8.2-11.8 mg/ml using an ultrafiltration membrane having a molecular weight cutoff of 30,000 MWCO. The antibody was diluted to 8 mg/ml with a mixture of 1 mM L-cysteine, 2 mM EDTA and 50 mM acetic acid/125 mM tris buffer solution (pH 6.8) to obtain the antibody sample. After adding papain (Roche Applied Science) to 1/240 of each antibody, the sample was mixed and then allowed to stand at 37°C for 1 hour. The sample thus retained was added to Gly-Gly-Tyr-Arg peptide-conjugated (Sigma-Aldrich Corp.) l-ml HiTrap NHS-activated HP (GE Healthcare Bio-Sciences Corp.) (equilibrated with 50 mM acetic acid/ 125 mM tris buffer solution (pH 6.8)) connected in series to the following 1 ml HiTrap MabSelect Sure protein A carrier column (GE Healthcare Bio-Sciences Corp.). Activated papain was removed by the upstream Gly-Gly-Tyr-Arg peptide, while Fc fragments and uncleaved antibodies were removed by the downstream carrier protein A column, yielding a purified Fab fraction. 10 μM cysteine protease inhibitor E64 (Sigma-Aldrich Corp.) was added to the Fab fraction to prevent activation of the inactive papain contained in the Fab fraction. All actions on the column were carried out at room temperature from 20 to 25°C.

(5-3) Obtaining samples for NMR of Fab fragments of antibodies LfVk1_Ca and LfVk1

Each antibody solution was concentrated to 0.5 ml by centrifugation using an MWCO 5000 Vivaspin ultrafilter (Sartorius). Next, a diafiltration dish was placed in the ultrafilter, and the buffer solution was replaced with an NMR buffer solution: 5 mM d-BisTris, 20 mM NaCl, 0.001% (w/v) NaN ₃ , and 5% (v/v) ₂ H ₂ O (pH 7.0) (pH adjusted using NaOH or HCl) (5 ml of buffer solution was added to a diafiltration dish, the contents of which were concentrated to 0.5 ml by centrifugation; this action was repeated three times). As a result, the resulting solution was concentrated to 0.25 ml. Finally, the ultrafilter was thoroughly washed with NMR buffer solution. The wash liquids were combined with the concentrate to obtain 420 μl of LfVk1_Ca antibody solution and 270 μl of LfVkl antibody solution. At this stage, the pH of each solution was reconfirmed and adjusted, if necessary, to pH 7.0 with NaOH or HCl. Absorbance was measured at 280 nm using a Nanodrop UV meter (Thermo Fisher Scientific KK). Each Fab fragment was quantified using the molar extinction coefficient at 280 nm set to 70,000 M ^-1 ⋅cm ^-1 and therefore the amount was determined to be 0.12 mM for Leu-labeled antibody LfVk1_Ca and Leu-labeled antibody LfVk1 and for the Asp/Glu/Asn/Gln-labeled LfVk1_Ca antibody and the Asp/Glu/Asn/Gln-labeled LfVk1 antibody, the amount is 0.24 mM. Of these samples, each LfVk1_Ca antibody was placed into an NMR sample tube (Shigemi Co., Ltd.) with a diameter of 5 mm, while each LfVk1 antibody was placed into a symmetrical microsample aqueous solution tube (Shigemi Co., Ltd.) with a diameter 5 mm using a Pasteur pipette. In the Ca ²⁺ titration experiment with the LfVk1_Ca antibody, CaCl ₂ solutions were sequentially added to the antibody solution in amounts of 1, 2, 5, 10, and 20 molar equivalents of Ca ²⁺ relative to the antibody. The CaCl ₂ solutions used for addition were prepared as 10, 20, 50 and 100 mM CaCl ₂ solutions dissolved in NMR buffer. The required amounts of CaCl ₂ solutions were directly added in volumes ranging from 3 to 10 μl to the antibody solutions placed in NMR sample tubes using a custom microsyringe (Ito Corp.) where the syringe part protruded from the finished part. Sample tubes were mixed using a vortex mixer and then centrifuged in a manual centrifuge (Shimadzu Corp.).

(5-3) NMR analysis to detect signals of amide groups in Fab fragments of antibodies LfVk1_Ca and LfVk1_Ca

NMR analysis was performed using a DRX750 NMR spectroscope (Bruker Biospin KK) equipped with a TCI CryoProbe. The temperature was set to 307 K (34°C) (gas flow: 535 l/h). ¹ H- ¹⁵ N HSQC was used to detect amide group signals in NMR analysis. The analysis method used ¹ H- ¹⁵ N FHSQC, involving the simultaneous ¹³ C-elimination of α-carbon and carbonyl carbon during the evolution of ¹⁵ N, and a group of 3-9-19 pulses to quench the aqueous solvent signals. To control them, we used a pulse program included as a standard by the manufacturer (Bruker Biospin KK). The NMR analysis conditions were as follows: spectral width: 12019 Hz (f2) and 1976 Hz (f1), and number of data points: 2048 (f2) and 128 (f1). Topspin 3.0 (Bruker Biospin KK) was used for data processing. The data processing conditions were as follows: for both f2 and f1, the data was multiplied based on the window function shifted sine (QSINE)-bell and padded with zeros to double the number of data points, followed by a Fourier transform. Chemical shifts of the signals were calculated using Sparky NMR analysis software (UCSF).

(5-4) Distribution of NMR signal of main chain amide groups 8 0% of the NMR signals of main chain amide groups in the Fab fragment of tocilizumab (heavy chain: SEQ ID NO: 24 and light chain: SEQ ID NO: 25) have already been distributed (data not published). The amino acid sequence of the Fab fragment of the LfVk1_Ca antibody is the same as the amino acid sequence of the Fab fragment of tocilizumab, except for part of the light chain CDR1, CDR2 and CDR3 and amino acid residues 73 and 83 of the light chain. Because the NMR signals of the same portions of the amino acid sequences of these antibodies have identical or similar chemical shifts, the distribution information in tocilizumab could be applied to the LfVk1_Ca antibody. Distribution of light chain residues 11, (33), (46), (47), (54), (78), 125, 135, 136, 154, 175, 179, 181 and 201 and residues 18, 46, 64, 71 , 81, 83, 114, 144, 147, 165, 176, 181, 184, and 195 heavy chain could be applied to the Leu-labeled sample. Numbers without parentheses correspond to residue numbers for which the distribution could be applied because their chemical shifts are identical to those of tocilizumab. The numbers in parentheses correspond to the residue numbers for which the distribution could be applied because their chemical shifts were similar to those of tocilizumab and there were no other signals with similar chemical shifts. For the Asp/Glu/Asn/Gln-labeled sample, four signals were again identified in LfVk1_Ca by spectral comparison between the LfVk1_Ca antibody and the LfVk1 antibody. These signals were successfully classified as signals originating from any 4 of the 5 residues contributed as Ca ²⁺ -binding motifs, i.e. Asp30, Asp31, Asp32, Asp92 and Glu50 in the light chain, the sequences of which differ between antibodies, among the residues Asp, Glu, Asn and Gin.

(5-5) Identification of the Ca ²⁺ -binding site on the LfVk1_Ca antibody

Signals with altered chemical shifts were extracted by comparing the ¹ H- ¹⁵ N HSQC spectra between the Fab fragment of the LfVk1_Ca antibody, without the addition of Ca ²⁺ , and the Fab fragment supplemented with 20 molar equivalents of Ca ²⁺ . The results for Leu-labeled samples demonstrated that light chain Leu33 is involved in binding, while other Leu residues are not involved in binding. Results for Asp/Glu/Asn/Gln-labeled samples demonstrated that any 4 of the 5 residues (Asp30, Asp31, Asp32, Asp92, and Glu50 of the light chain) are involved in binding, while the other residues, Asp, Glu, Asn, and Gln, with the exception of 1 residue, are not involved in binding. From these results, amino acids in at least the light chain CDR1 in the amino acid sequence inserted as Ca ²⁺ binding motifs, as well as one or both of the light chain CDR2 and CDR3, were identified to be involved in Ca ²⁺ binding . This was consistent with the results confirmed in Example 15, indicating that it is important for calcium ion binding that 4 of residues 30, 31, 32, 50 and 92 (defined by Kabat numbering) are amino acids of the sequence hVk5-2.

(5-6) Calculation of the Ca ²⁺ dissociation constant in a titration experiment

¹ H- ¹⁵ N HSQC spectra were used, which were obtained when the Ca ²⁺ concentration was 0, 1, 2, 5, 10 or 20 molar equivalents relative to the Fab fragment of the LfVk1_Ca antibody. The ¹ H or ¹⁵ N chemical shift in the light chain Leu33 signal identified as the binding site was plotted as the ordinate, while the above molar Ca ²⁺ equivalents were plotted as the abscissa. The data was fitted to a function corresponding to the following expression 2 using Gnuplot plotting software:

[Expression 2]

f(x)=s×[1-0.5/a×{(a×x+a+Kd)-((a×x+a+Kd) ² -4×x×a ² ) ^0.5 } +t×[0.5/a×{(a×x+a+Kd) - ((a×x+a+Kd) ² -4×x×a ² ) ^0.5 }

In the function corresponding to expression 2, s and t correspond to the chemical shift [ppm] in the absence of bound Ca ²⁺ and the expected chemical shift [ppm] in the presence of bound Ca ²⁺ at saturation, respectively; a corresponds to the concentration [M] of the Fab fragment of the antibody; Kd corresponds to the dissociation constant; and x corresponds to the molar equivalent of Ca ²⁺ added to the Fab fragment of the antibody. For approximation, s, t and Kd were used as approximation parameters. As a result, from a chemical shift of ¹ N an estimate of Kd=7.1×10 ^-5 [M] was obtained, and from a chemical shift of ¹⁵ N an estimate of Kd=5.9×10 ^-5 [M] was obtained.

Claims (93)

1. A library for screening an antibody that binds to an antigen of interest in a pH-dependent manner, essentially consisting of a plurality of sequence-differing antibodies, wherein the antigen binding domain in each antibody contains at least one histidine that modifies the antigen binding activity of the antibody depending on pH conditions, and where the histidine is at any one or more positions selected from the group 27, 31, 32, 33, 35, 50, 52, 53, 55, 57, 58, 59, 61, 62, 95, 96, 97, 98 , 99, 100a, 100b, 100d, 100f, 100h, 102 and 107, defined according to Kabat numbering, in the heavy chain variable region and at positions selected from the group 24, 27, 28, 30, 31, 32, 34, 50 , 51, 52, 53, 54, 55, 56, 89, 90, 91, 92, 93, 94 and 95a, defined according to Kabat numbering, in the light chain variable region. 2. The library according to claim 1, where the histidine is located at any one or more positions selected from the group 27, 31, 32, 33, 35, 50, 52, 53, 55, 57, 58, 59, 61, 62 , 95, 96, 97, 98, 99, 100a, 100b, 100d, 100f, 100h, 102 and 107, as defined by Kabat numbering, in the heavy chain variable region. 3. The library according to claim 2, where the amino acid sequence contains a naive sequence, excluding a histidine located at any one or more positions selected from the group 27, 31, 32, 33, 35, 50, 52, 53, 55, 57, 58, 59, 61,62, 95, 96, 97, 98, 99, 100a, 100b, 100d, 100f, 100h, 102, and 107, as defined by Kabat numbering, in the heavy chain variable region. 4. The library according to any one of claims 1 to 3, wherein the antibody light chain variable region comprises a germline sequence. 5. The library according to claim 1, where the histidine is located at any one or more positions selected from the group 24, 27, 28, 30, 31, 32, 34, 50, 51, 52, 53, 54, 55, 56 , 89, 90, 91, 92, 93, 94 and 95a, defined by Kabat numbering, in the light chain variable region. 6. The library according to claim 5, wherein the histidine is located in CDR1 of the light chain variable region. 7. The library of claim 6, wherein the histidine is located at any one or more positions selected from the group 24, 27, 28, 30, 31, 32 and 34, as defined by Kabat numbering, in the CDR1 of the light chain. 8. The library according to claim 5, where histidine is located in CDR2 of the light chain. 9. The library of claim 8, wherein the histidine is located at any one or more positions selected from the group 50, 51, 52, 53, 54, 55 and 56, as defined by Kabat numbering, in the CDR2 of the light chain. 10. The library according to claim 5, where the histidine is located in CDR3 of the light chain. 11. The library of claim 10, wherein the histidine is located at any one or more positions selected from the group 89, 90, 91, 92, 93, 94 and 95a, defined according to Kabat numbering, in the CDR3 of the light chain. 12. Library according to any one of paragraphs. 5-11, wherein the light chain framework region comprises a germline framework sequence. 13. The library according to any one of claims 5 to 12, wherein the heavy chain variable region contains a naive sequence. 14. A library for screening an antibody that binds to an antigen of interest in a pH-dependent manner, essentially consisting of a plurality of fusion polypeptides, each containing different antibody sequences, wherein each of the fusion polypeptides is a variable region fusion product an antibody heavy chain and at least a portion of a viral envelope protein, wherein the antigen binding domain in each antibody contains at least one histidine that modifies the antigen binding activity of the antibody depending on pH conditions, and where the histidine is located at any of the positions selected from the group 27, 31, 32, 33, 35, 50, 52, 53, 55, 57, 58, 59, 61, 62, 95, 96, 97, 98, 99, 100a , 100b, 100d, 100f, 100h, 102 and 107, defined according to Kabat numbering, in the heavy chain variable region and at positions selected from the group 24, 27, 28, 30, 31, 32, 34, 50, 51, 52 , 53, 54, 55, 56, 89, 90, 91, 92, 93, 94 and 95a, defined according to Kabat numbering, in the light chain variable region. 15. The library of claim 14, wherein the viral envelope protein is selected from the group consisting of pIII protein, major envelope protein pVIII, pVII, pIX, Soc, Hoc, gpD, pv1, and variants thereof. 16. A composition for screening an antibody that binds to an antigen of interest in a pH-dependent manner, comprising a plurality of polynucleotide molecules, each of which encodes a different sequence of an antibody or a different sequence of fusion polypeptides, each of which contains an antibody, wherein each of the fusion polypeptides is a fusion product of an antibody heavy chain variable region and at least a portion of a viral envelope protein, wherein the antigen binding domain in each of the antibodies contains at least one histidine, which modifies the antigen binding activity of the antibody depending on pH conditions, and where the histidine is located at any of the positions selected from the group 27, 31, 32, 33, 35, 50, 52, 53, 55, 57, 58, 59, 61, 62, 95, 96, 97, 98, 99, 100a , 100b, 100d, 100f, 100h, 102 and 107, defined according to Kabat numbering, in the heavy chain variable region and at positions selected from the group 24, 27, 28, 30, 31, 32, 34, 50, 51, 52 , 53, 54, 55, 56, 89, 90, 91, 92, 93, 94 and 95a, defined according to Kabat numbering, in the light chain variable region. 17. A composition for screening an antibody that binds to an antigen of interest in a pH-dependent manner, comprising a plurality of vectors, each of which contains a plurality of polynucleotide molecules in an operably linked state, wherein each of the polynucleotide molecules encodes antibodies that differ in sequence from each other, or fusion polypeptides, each of which contains different antibody sequences, wherein each of the fusion polypeptides is a fusion product of an antibody heavy chain variable region and at least a portion of a viral envelope protein, wherein the antigen binding domain in each of the antibodies contains at least one histidine, which modifies the antigen binding activity of the antibody depending on pH conditions, and where the histidine is located at any of the positions selected from the group 27, 31, 32, 33, 35, 50, 52, 53, 55, 57, 58, 59, 61, 62, 95, 96, 97, 98, 99, 100a , 100b, 100d, 100f, 100h, 102 and 107, defined according to Kabat numbering, in the heavy chain variable region and at positions selected from the group 24, 27, 28, 30, 31, 32, 34, 50, 51, 52 , 53, 54, 55, 56, 89, 90, 91, 92, 93, 94 and 95a, defined according to Kabat numbering, in the light chain variable region. 18. The composition of claim 17, wherein the vectors are expression vectors that can be replicated. 19. The composition of claim 18, wherein each of the replicable expression vectors is an expression vector in which the polynucleotide is operably linked to a promoter region selected from the group consisting of a lacZ promoter system, a phoA alkaline phosphatase (Ap) promoter, bacteriophage λPL promoter (temperature sensitive promoter), tac promoter, tryptophan promoter, pBAD promoter and bacteriophage T7 promoter. 20. The composition according to claim 18 or 19, wherein each of the expression vectors that can be replicated is an M13, f1, fd or Pf3 phage or a derivative thereof or a lambdoid phage or a derivative thereof. 21. A composition for screening an antibody that binds to an antigen of interest in a pH-dependent manner, comprising a plurality of viruses, each of which contains vectors, each of the vectors contains a plurality of polynucleotide molecules in a functionally linked state, wherein each of the polynucleotide molecules encodes antibodies that differ in sequence from each other or fusion polypeptides, each of which contains antibodies that differ in sequence from each other, wherein each of the fusion polypeptides is a fusion product of an antibody heavy chain variable region and at least a portion of a viral envelope protein, wherein the antigen binding domain in each of the antibodies contains at least one histidine, which modifies the antigen binding activity of the antibody depending on pH conditions, and where the histidine is located at any of the positions selected from the group 27, 31, 32, 33, 35, 50, 52, 53, 55, 57, 58, 59, 61, 62, 95, 96, 97, 98, 99, 100a , 100b, 100d, 100f, 100h, 102 and 107, defined according to Kabat numbering, in the heavy chain variable region and at positions selected from the group 24, 27, 28, 30, 31, 32, 34, 50, 51, 52 , 53, 54, 55, 56, 89, 90, 91, 92, 93, 94 and 95a, defined according to Kabat numbering, in the light chain variable region. 22. A composition for screening an antibody that binds to an antigen of interest in a pH-dependent manner, comprising a plurality of viruses, each of which displays on its surface sequence-different antibodies or fusion polypeptides, each of which contains antibodies different from each other. friend by sequence, wherein each of the fusion polypeptides is a fusion product of an antibody heavy chain variable region and at least a portion of a viral envelope protein, wherein the antigen binding domain in each of the antibodies contains at least one histidine, which changes the antigen binding activity of the antibody depending on pH conditions, and where the histidine is located at any of the positions selected from the group 27, 31, 32, 33, 35, 50, 52, 53, 55, 57, 58, 59, 61, 62, 95, 96, 97, 98, 99, 100a , 100b, 100d, 100f, 100h, 102 and 107, defined according to Kabat numbering, in the heavy chain variable region and at positions selected from the group 24, 27, 28, 30, 31, 32, 34, 50, 51, 52 , 53, 54, 55, 56, 89, 90, 91, 92, 93, 94 and 95a, defined according to Kabat numbering, in the light chain variable region. 23. The library according to any one of claims 1 to 15, where the library has from 1*10 ⁶ to 1*10 ¹⁴ different variable region sequences. 24. The library according to claim 23, where the library has 1*10 ⁸ or more different variable region sequences. 25. A method of producing a library for screening an antibody that binds to an antigen of interest in a pH-dependent manner, essentially consisting of a plurality of sequence-differing antibodies, the method comprising producing a plurality of antibodies designed such that the antigen binding domain in each of the antibodies contains at least one histidine, which changes the antigen-binding activity of the antibody depending on pH conditions, where the histidine is located at any of the positions selected from the group 27, 31, 32, 33, 35, 50, 52, 53, 55, 57, 58, 59, 61, 62, 95, 96, 97, 98, 99, 100a , 100b, 100d, 100f, 100h, 102 and 107, defined according to Kabat numbering, in the heavy chain variable region and at positions selected from the group 24, 27, 28, 30, 31, 32, 34, 50, 51, 52 , 53, 54, 55, 56, 89, 90, 91, 92, 93, 94 and 95a, defined according to Kabat numbering, in the light chain variable region. 26. The method of production according to claim 25, wherein the variable region of the heavy chain of each of the antibodies is fused to at least a portion of the viral envelope protein. 27. The method of production according to claim 26, wherein the viral envelope protein is selected from the group consisting of pIII protein, the main envelope protein pVIII, pVII, pIX, Soc, Hoc, gpD, pv1 and variants thereof. 28. A method for selecting an antibody whose antigen-binding activity varies depending on pH conditions, the method comprising the steps of: a) obtaining the library according to any of claims 1-15; b) contacting the library with antigens under two or more different pH conditions; c) sorting from the library a subpopulation of antibodies whose antigen-binding activity varies depending on pH conditions; And d) isolating each antibody whose antigen-binding activity varies with pH conditions from the subpopulation sorted in step c). 29. A method for isolating a polynucleotide encoding an antibody, the antigen-binding activity of which varies depending on pH conditions, the method comprising the steps of: a) obtaining a library containing a plurality of expression vectors that can be replicated, each of which contains in an operably linked state a plurality of polynucleotides, each of which encodes a different sequence of antibodies or a different sequence of fusion polypeptides, each of which contains antibodies, wherein each of the fusion polypeptides is a fusion product of an antibody heavy chain variable region and at least a portion of a viral envelope protein, wherein the antigen binding domain in each of the antibodies contains at least one histidine, which modifies the antigen binding activity of the antibody depending on pH conditions, and where the histidine is located at any of the positions selected from the group 27, 31, 32, 33, 35, 50, 52, 53, 55, 57, 58, 59, 61, 62, 95, 96, 97, 98, 99, 100a , 100b, 100d, 100f, 100h, 102 and 107, defined according to Kabat numbering, in the heavy chain variable region and at positions selected from the group 24, 27, 28, 30, 31, 32, 34, 50, 51, 52 , 53, 54, 55, 56, 89, 90, 91, 92, 93, 94 and 95a, defined according to Kabat numbering, in the light chain variable region; b) allowing a plurality of viruses, each transformed by expression vectors contained in the library, to express on their surface antibodies encoded by polynucleotides or fusion polypeptides differing from each other in sequence; c) bringing multiple viruses into contact with antigens under two or more different pH conditions; d) sorting from the library a subpopulation of viruses whose antigen-binding activity varies depending on pH conditions; e) isolating each virus whose antigen-binding activity varies with pH conditions from the subpopulation of viruses sorted in step d); And f) isolating polynucleotides from the isolated virus. 30. The method according to claim 29, where steps c) and d) are additionally repeated at least once. 31. Method according to any one of paragraphs. 28-30, where an antibody having lower antigen binding activity is selected under more acidic pH conditions than at neutral pH conditions. 32. The method of claim 31, wherein the acidic pH conditions are pH conditions between 4.0 and 6.5. 33. The method of claim 31 or 32, wherein the neutral pH conditions are pH conditions between 6.7 and 10.0. 34. A method for producing an antibody, the antigen-binding activity of which varies depending on pH conditions, the method comprising the steps of: a) obtaining a library containing a plurality of expression vectors that can be replicated, each of which contains in an operably linked state a plurality of polynucleotides, each of which encodes a different sequence of antibodies or fusion polypeptides, each of which contains a different sequence of antibodies; wherein each of the fusion polypeptides is a fusion product of an antibody heavy chain variable region and at least a portion of a viral envelope protein, wherein the antigen binding domain in each of the antibodies contains at least one histidine, which modifies the antigen binding activity of the antibody depending on pH conditions, and where the histidine is located at any of the positions selected from the group 27, 31, 32, 33, 35, 50, 52, 53, 55, 57, 58, 59, 61, 62, 95, 96, 97, 98, 99, 100a , 100b, 100d, 100f, 100h, 102 and 107, defined according to Kabat numbering, in the heavy chain variable region and at positions selected from the group 24, 27, 28, 30, 31, 32, 34, 50, 51, 52 , 53, 54, 55, 56, 89, 90, 91, 92, 93, 94 and 95a, defined according to Kabat numbering, in the light chain variable region; b) allowing a plurality of viruses, each transformed by expression vectors contained in the library, to express on their surface sequence-different antibodies or fusion polypeptides, each of which is encoded by polynucleotides; c) bringing multiple viruses into contact with antigens under two or more different pH conditions; d) sorting from the library a subpopulation of viruses whose antigen-binding activity varies depending on pH conditions; e) isolating each virus whose antigen-binding activity varies with pH conditions from the subpopulation of viruses sorted in step d); f) isolating polynucleotides from the isolated virus; g) culturing a host cell transfected with a vector having an operably linked insert of the isolated polynucleotides; And h) collecting antibodies from the cell cultures cultured in step g). 35. A method for producing antibodies, the antigen-binding activity of which varies depending on pH conditions, the method comprising the steps of: a) obtaining a library containing a plurality of expression vectors that can be replicated, each of which contains, in an operably linked state, a plurality of polynucleotides, each of which encodes a different sequence of antibodies or a different sequence of fusion polypeptides, wherein each of the fusion polypeptides is a fusion product of an antibody heavy chain variable region and at least a portion of a viral envelope protein, wherein the antigen binding domain in each of the antibodies contains at least one histidine, which modifies the antigen binding activity of the antibody depending on pH conditions, and where the histidine is located at any of the positions selected from the group 27, 31, 32, 33, 35, 50, 52, 53, 55, 57, 58, 59, 61, 62, 95, 96, 97, 98, 99, 100a , 100b, 100d, 100f, 100h, 102 and 107, defined according to Kabat numbering, in the heavy chain variable region and positions selected from the group 24, 27, 28, 30, 31, 32, 34, 50, 51, 52, 53, 54, 55, 56, 89, 90, 91, 92, 93, 94 and 95a, defined according to Kabat numbering, in the light chain variable region; b) allowing a plurality of viruses, each transformed by expression vectors contained in the library, to express on their surface different antibody sequences or fusion polypeptides, each of which is encoded by polynucleotides; c) bringing multiple viruses into contact with antigens under two or more different pH conditions; d) sorting from the library a subpopulation of viruses whose antigen-binding activity varies depending on pH conditions; e) isolating each virus whose antigen-binding activity varies with pH conditions from the subpopulation of viruses sorted in step d); f) isolating polynucleotides from the isolated virus; g) linking the isolated polynucleotides in frame to a polynucleotide encoding the constant region of the antibody; h) culturing a host cell transfected with a vector having an operably linked insert of polynucleotides linked in step g); And i) isolating antibodies from cell cultures cultured in step h). 36. The production method according to claim 34 or 35, where steps c) and d) are additionally repeated at least once. 37. Method of production according to paragraphs. 34-36, wherein an antigen binding molecule is selected which has lower antigen binding activity under acidic pH conditions than under neutral pH conditions. 38. The production method according to claim 37, wherein the acidic pH conditions are pH conditions of 4.0 to 6.5. 39. The production method according to claim 37 or 38, wherein the neutral pH conditions are those between pH 6.7 and 10.0.

Images