Method for producing heterodimeric antibodies

Abstract

The present invention relates to a method for producing a heterodimeric antibody. TIFF2022513999000018.tif43146

Description

[Technical field]

[0001] CROSS-REFERENCE TO RELATED APPLICATIONS This application claims the benefit of U.S. Provisional Patent Application No. 62 / 781,180, filed December 18, 2018, the entire contents of which are incorporated herein by reference.

[0002] FIELD OF THEINVENTION The present invention relates to a method for producing a heterodimeric antibody.

[0003] Sequence Listing This application contains a Sequence Listing submitted via EFS-Web, the entire contents of which are incorporated herein by reference. The ASCII text file, created on December 10, 2019, is named JBI6029WOPCT1_ST25.txt and is 21 kilobytes in size. [Background technology]

[0004] 2. Background of the Invention Monoclonal antibodies have proven successful as therapeutic molecules, especially in the treatment of cancer. Bispecific antibodies are expected to increase the potency and efficacy of monoclonal antibody therapy, as they may be used to target drugs or toxic compounds to target cells, redirect effector mechanisms to disease-related sites, or increase specificity for tumor cells, for example by binding to one or more target molecules expressed on tumor cells. Moreover, by combining the specificities of two monoclonal antibodies into one, bispecific antibodies may be involved in a greater number of mechanisms of action.

[0005] Although various formats of bispecific antibodies and methods for their production have been described, challenges remain in producing bispecific antibodies on a large scale by optimizing the manufacturing process. Summary of the Invention

[0006] BRIEF SUMMARY OF THE INVETION The present invention provides a method for producing a heterodimeric antibody, comprising the steps of: providing a first homodimeric antibody comprising a first Fc region of an immunoglobulin comprising a first CH3 region and a second homodimeric antibody comprising a second Fc region of an immunoglobulin comprising a second CH3 region, wherein the amino acid sequences of the first CH3 region and the second CH3 region are different and the amino acid sequences of the first CH3 region and the second CH3 region are such that a heterodimeric interaction between the first CH3 region and the second CH3 region is stronger than a homodimeric interaction between the first CH3 regions or a homodimeric interaction between the second CH3 regions; combining the first homodimeric antibody and the second homodimeric antibody into a mixture; and A step of incubating the mixture in the presence of a reducing agent and ambient dissolved oxygen (DO2) to produce a heterodimeric antibody, wherein the method lacks one or more of the following steps during production of the heterodimeric antibody: measuring the percent (%) DO2 in the mixture, controlling the % DO2 in the mixture, or adding oxygen to the mixture.

[0007] The present invention also provides a method for producing a heterodimeric antibody, comprising the steps of: providing a first homodimeric antibody comprising a first Fc region of an immunoglobulin comprising a first CH3 region and a second homodimeric antibody comprising a second Fc region of an immunoglobulin comprising a second CH3 region, wherein the amino acid sequences of the first CH3 region and the second CH3 region are different and the amino acid sequences of the first CH3 region and the second CH3 region are such that a heterodimeric interaction between the first CH3 region and the second CH3 region is stronger than a homodimeric interaction between the first CH3 regions or a homodimeric interaction between the second CH3 regions; combining the first homodimeric antibody and the second homodimeric antibody into a mixture; incubating the mixture in the presence of a reducing agent; and A step of removing the reducing agent to produce a heterodimeric antibody, wherein the percent (%) of dissolved oxygen (DO2) is controlled to be about 30% or less during the steps of incubating the mixture in the presence of the reducing agent, removing the reducing agent to produce a heterodimeric antibody, or incubating the mixture in the presence of the reducing agent and removing the reducing agent to produce a heterodimeric antibody. [The present invention 1001] A method for producing a heterodimeric antibody, comprising the steps of: a) providing a first homodimeric antibody comprising a first Fc region of an immunoglobulin comprising a first CH3 region and a second homodimeric antibody comprising a second Fc region of an immunoglobulin comprising a second CH3 region, wherein the amino acid sequences of the first CH3 region and the second CH3 region are different and the amino acid sequences of the first CH3 region and the second CH3 region are such that a heterodimeric interaction between the first CH3 region and the second CH3 region is stronger than a homodimeric interaction between the first CH3 regions or a homodimeric interaction between the second CH3 regions; b) combining the first homodimeric antibody and the second homodimeric antibody into a mixture; and c) The mixture is treated with a reducing agent and ambient dissolved oxygen (DO 2 ), to produce a heterodimeric antibody, the method comprising: incubating a mixture of the antibody and a soluble form of the antibody in the presence of a soluble form of the antibody; 2 measuring the %DO in the mixture; 2 or adding oxygen to the mixture. [The present invention 1002] %DO in the mixture 2 is about 30% or more in step 1b). [The present invention 1003] During the production of heterodimeric antibodies, %DO 2 The method of any one of claims 1001 to 1002, wherein the rate of release is about 10% to about 90%. [The present invention 1004] The method of any of claims 1001 to 1003, wherein in step 1b), the first homodimeric antibody and the second homodimeric antibody are combined in a molar ratio of about 1:1 to about 1:2 to form a mixture. [The present invention 1005] The method of any of claims 1001 to 1003, wherein in step 1b) the first homodimeric antibody and the second homodimeric antibody are combined to form a mixture in a molar ratio of about 1.05:1. [The present invention 1006] The method according to any one of claims 1001 to 1005, wherein the total concentration of immunoglobulins in the mixture is about 8 g / L to about 13 g / L, for example, about 10.5 g / L. [The present invention 1007] The method of any one of claims 1001 to 1006, wherein said mixture is incubated in the presence of said reducing agent for about 10 minutes or more. [The present invention 1008] The method of claim 10, wherein said mixture is incubated in the presence of said reducing agent for about 10 minutes to about 24 hours. [The present invention 1009] The method according to any one of claims 1001 to 1008, wherein said reducing agent is 2-mercaptoethylamine (2-MEA), a chemical derivative of 2-MEA, L-cysteine ​​or D-cysteine. [The present invention 1010] The method of claim 1009, wherein the concentration of 2-MEA in said mixture is about 20 mM to 40 mM, for example about 35 mM. [The present invention 1011] The method of any one of claims 1001 to 1010, wherein the mixture comprises a buffer. [The present invention 1012] The method of any one of claims 1 to 11, wherein the buffer comprises a sodium acetate buffer. [The present invention 1013] The method of claim 1012, wherein the buffer further comprises NaCl. [The present invention 1014] The method of any one of claims 1 to 3, wherein said buffer comprises about 100 mM sodium acetate and about 30 mM NaCl. [The present invention 1015] The method of any one of claims 1 to 4, wherein the pH of said buffer is about 7.3. [The present invention 1016] The method according to any one of claims 1001 to 1015, further comprising the step of removing said reducing agent from said mixture. [The present invention 1017] The method of claim 1016, wherein the reducing agent is removed by filtration. [The present invention 1018] The method of claim 1017, wherein the filtration is diafiltration. [The present invention 1019] The method of any of claims 1001 to 1018, wherein the first homodimeric antibody and the second homodimeric antibody are of the IgG1, IgG2 or IgG4 isotype. [The present invention 1020] The method of the present invention, wherein the first CH3 domain and the second CH3 domain comprise the following mutations when compared to the wild type IgG1 of SEQ ID NO: 1, the wild type IgG2 of SEQ ID NO: 2, or the wild type IgG4 of SEQ ID NO: 3: F405L / K409R, wild type / F405L_R409K, T350I_K370T_F405L / K409R, K370W / K409R, D399AFGHILMNRSTVWY / K4 09R, T366ADEFGHILMQVY / K409R, L368ADEGHNRSTVQ / K409AGRH, D399FHKRQ / K409AGRH, F405IKLSTVW / K4 09AGRH, Y407LWQ / K409AGRH, T366Y / F405A, T366W / F405W, F405W / Y407A, T394W / Y407T, T394S / Y407A, T 366W / T394S, F405W / T394S, T366W / T366S_L368A_Y407V, L351Y_F405A_Y407V / T394W, T366I_K392M_T39 4W / F405A_Y407V, T366L_K392M_T394W / F405A_Y407V, L351Y_Y407A / T366A_K409F, L351Y_Y407A / T366 V_K409F, Y407A / T366A_K409F, T350V_L351Y_F405A_Y407V / T350V_T366L_K392L_T394W, K409D / D399K, K409E / D399R, K409D_K360D / D399K_E356K, K409D_K360D / D399E_E356K, K409D_K370D / D399K_E357K, K409D_K370D / D399E_E357K, K409D_K392D / D399K_E356K_E357K, or K409D_K392D / D399E_E356K_E357K. [The present invention 1021] The method of the present invention 1020, wherein the first Fc region and / or the second Fc region comprises one or more mutations that modulate binding of the first Fc region and / or the second Fc region to Fcγ receptors (FcγR), FcRn, or protein A, compared to a wild-type IgG1 of SEQ ID NO: 1, a wild-type IgG2 of SEQ ID NO: 2, or a wild-type IgG4 of SEQ ID NO: 3. [The present invention 1022] The method of the present invention 1021, wherein the FcγR is FcγRI, FcγRIIa, FcγRIIb or FcγRIII. [The present invention 1023] The one or more substitutions that modulate binding of the first Fc region and / or the second Fc region to Fcγ are L234A_L235A, F234A_L235A, S228P_F234A_L235A, S228P_L234A_L235A, V234A_G237A_P238S_H268A_V309L_A330S_P331S, V23 4A_G237A, H268Q_V309L_A330S_P331S, S267E_L328F, L234F_L235E_D265A, L234A_L235A_G237A_P238S_H268A_A330S_P331S, or S228P_F234A_L235A_G237A_P238S. [The present invention 1024] 1021. The method of claim 1021, wherein the one or more substitutions that modulate binding of the first Fc region and / or the second Fc region to FcRn are M428L_N434S, M252Y_S254T_T256E, T250Q_M428L, N434A and T307A_E380A_N434A, H435A, P257I_N434H, D376V_N434H, M252Y_S254T_T256E_H433K_N434F, T308P_N434A, or H435R. [The present invention 1025] 1021. The method of claim 1021, wherein the one or more substitutions that modulate binding of the first Fc region and / or the second Fc region to Protein A are Q311R, Q311K, T307P_L309Q, T307P_V309Q, T307P_L309Q_Q311R, T307P_V309Q_Q311R, H435R, or H435R_Y436F. [The present invention 1026] Any of the methods of claims 1001 to 1025, wherein the step of producing a heterodimeric antibody is carried out under conditions in accordance with GMP. [The present invention 1027] The method of claim 1026, wherein the step of producing a heterodimeric antibody is carried out during the manufacture of a drug substance comprising the heterodimeric antibody. [The present invention 1028] The method of any of claims 1001 to 1027, wherein said heterodimeric antibody is a bispecific antibody. [The present invention 1029] The method of claim 1028, wherein said bispecific antibody binds to CD3, CD123, BCMA, EGFR, or c-Met, or any combination thereof. [The present invention 1030] The method according to any one of claims 1001 to 1029, wherein a mass ratio of said reducing agent to total immunoglobulins in said mixture is from about 1.0 to about 5.0. [The present invention 1031] The method of claim 1030, wherein said mass ratio is from about 1.4 to about 3.8. [The present invention 1032] The method of claim 1030, wherein said mass ratio is from about 3.3 to about 4.4. [The present invention 1033] A method for producing a heterodimeric antibody, comprising the steps of: a) providing a first homodimeric antibody comprising a first Fc region of an immunoglobulin comprising a first CH3 region and a second homodimeric antibody comprising a second Fc region of an immunoglobulin comprising a second CH3 region, wherein the amino acid sequences of the first CH3 region and the second CH3 region are different and the amino acid sequences of the first CH3 region and the second CH3 region are such that a heterodimeric interaction between the first CH3 region and the second CH3 region is stronger than a homodimeric interaction between the first CH3 regions or a homodimeric interaction between the second CH3 regions; b) combining the first homodimeric antibody and the second homodimeric antibody into a mixture; c) incubating the mixture in the presence of a reducing agent; and d) removing the reducing agent to produce a heterodimeric antibody, wherein dissolved oxygen (DO 2 ) is controlled to be about 30% or less. [The present invention 1034] %DO of the mixture 2 is about 25% or less, 20% or less, about 15% or less, about 10% or less, about 9% or less, about 8% or less, about 7% or less, about 6% or less, about 5% or less, about 4% or less, about 3% or less, about 2% or less, or about 1% or less. [The present invention 1035] %DO 2 The method of claim 1034, wherein the is controlled by an overlay of nitrogen. [The present invention 1036] The method of claim 1034 or 1035, wherein in step b) the first homodimeric antibody and the second homodimeric antibody are combined to form a mixture in a molar ratio of about 1:1 to about 1:2. [The present invention 1037] The method of any one of claims 1034 to 1035, wherein in step b) the first homodimeric antibody and the second homodimeric antibody are combined into the mixture in a molar ratio of about 1.05:1. [The present invention 1038] The method of any of claims 1033 to 1037, wherein the total concentration of immunoglobulins in the mixture is about 10.5 g / L. [The present invention 1039] The method of any of claims 1033 to 1038, wherein said mixture is incubated in the presence of said reducing agent for about 10 minutes or more. [The present invention 1040] The method of any of claims 1033 to 1039, wherein the mixture is incubated in the presence of the reducing agent for about 10 minutes to about 24 hours. [The present invention 1041] The method according to any one of claims 1033 to 1040, wherein the reducing agent is 2-mercaptoethylamine (2-MEA), a chemical derivative of 2-MEA, L-cysteine ​​or D-cysteine. [The present invention 1042] The method of claim 1041, wherein the concentration of 2-MEA in said mixture is about 35 mM. [The present invention 1043] The method of any one of claims 1033 to 1042, wherein the mixture comprises a buffer. [The present invention 1044] The method of claim 1043, wherein the buffer comprises a sodium acetate buffer. [The present invention 1045] The method of claim 1044, wherein the buffer further comprises NaCl. [The present invention 1046] The method of any one of claims 1 to 45, wherein said buffer comprises about 100 mM sodium acetate and about 30 mM NaCl. [The present invention 1047] The method of claim 1046, wherein the pH of said buffer is about 7.3. [The present invention 1048] The method according to any one of claims 1033 to 1047, further comprising the step of removing said reducing agent from said mixture. [The present invention 1049] The process of claim 1048, wherein the reducing agent is removed by filtration. [The present invention 1050] The method of claim 1049, wherein the filtration is diafiltration. [The present invention 1051] The method of any of claims 1033 to 1050, wherein the first homodimeric antibody and the second homodimeric antibody are of the IgG1, IgG2 or IgG4 isotype. [The present invention 1052] The method of the present invention, wherein the first CH3 domain and the second CH3 domain comprise the following mutations when compared to the wild type IgG1 of SEQ ID NO: 1, the wild type IgG2 of SEQ ID NO: 2, or the wild type IgG4 of SEQ ID NO: 3: F405L / K409R, wild type / F405L_R409K, T350I_K370T_F405L / K409R, K370W / K409R, D399AFGHILMNRSTVWY / K4 09R, T366ADEFGHILMQVY / K409R, L368ADEGHNRSTVQ / K409AGRH, D399FHKRQ / K409AGRH, F405IKLSTVW / K4 09AGRH, Y407LWQ / K409AGRH, T366Y / F405A, T366W / F405W, F405W / Y407A, T394W / Y407T, T394S / Y407A, T 366W / T394S, F405W / T394S, T366W / T366S_L368A_Y407V, L351Y_F405A_Y407V / T394W, T366I_K392M_T39 4W / F405A_Y407V, T366L_K392M_T394W / F405A_Y407V, L351Y_Y407A / T366A_K409F, L351Y_Y407A / T366 V_K409F, Y407A / T366A_K409F, T350V_L351Y_F405A_Y407V / T350V_T366L_K392L_T394W, K409D / D399K, K409E / D399R, K409D_K360D / D399K_E356K, K409D_K360D / D399E_E356K, K409D_K370D / D399K_E357K, K409D_K370D / D399E_E357K, K409D_K392D / D399K_E356K_E357K, or K409D_K392D / D399E_E356K_E357K. [The present invention 1053] The method of the present invention 1052, wherein the first Fc region and / or the second Fc region comprises one or more mutations that modulate binding of the first Fc region and / or the second Fc region to Fcγ receptors (FcγR), FcRn, or protein A, compared to a wild-type IgG1 of SEQ ID NO: 1, a wild-type IgG2 of SEQ ID NO: 2, or a wild-type IgG4 of SEQ ID NO: 3. [The present invention 1054] The method of the present invention 1053, wherein the FcγR is FcγRI, FcγRIIa, FcγRIIb or FcγRIII. [The present invention 1055] The one or more substitutions that modulate binding of the first Fc region and / or the second Fc region to Fcγ are L234A_L235A, F234A_L235A, S228P_F234A_L235A, S228P_L234A_L235A, V234A_G237A_P238S_H268A_V309L_A330S_P331S, V23 4A_G237A, H268Q_V309L_A330S_P331S, S267E_L328F, L234F_L235E_D265A, L234A_L235A_G237A_P238S_H268A_A330S_P331S, or S228P_F234A_L235A_G237A_P238S. [The present invention 1056] 1053. The method of claim 1053, wherein the one or more substitutions that modulate binding of the first Fc region and / or the second Fc region to FcRn are M428L_N434S, M252Y_S254T_T256E, T250Q_M428L, N434A and T307A_E380A_N434A, H435A, P257I_N434H, D376V_N434H, M252Y_S254T_T256E_H433K_N434F, T308P_N434A, or H435R. [The present invention 1057] 1053. The method of claim 1053, wherein the one or more substitutions that modulate binding of the first Fc region and / or the second Fc region to Protein A are Q311R, Q311K, T307P_L309Q, T307P_V309Q, T307P_L309Q_Q311R, T307P_V309Q_Q311R, H435R, or H435R_Y436F. [The present invention 1058] The method according to any one of claims 1033 to 1057, wherein the step of producing a heterodimeric antibody is carried out under conditions in accordance with GMP. [The present invention 1059] The method according to any one of claims 1033 to 1057, wherein the step of producing the heterodimeric antibody is carried out during the production of a drug substance comprising the heterodimeric antibody. [The present invention 1060] The method of any of claims 1033 to 1059, wherein said heterodimeric antibody is a bispecific antibody. [The present invention 1061] The method of claim 1060, wherein said bispecific antibody binds to CD3, CD123, BCMA, EGFR, or c-Met, or any combination thereof. [The present invention 1062] The method according to any one of claims 1033 to 1061, wherein a mass ratio of said reducing agent to total immunoglobulins in said mixture is from about 1.0 to about 5.0. [The present invention 1063] The method of claim 1062, wherein the mass ratio is from about 1.4 to about 3.8. [The present invention 1064] The method of claim 1062, wherein the mass ratio is from about 3.3 to about 4.4. [The present invention 1065] 10. An isolated bispecific antibody produced by the method of the present invention. [The present invention 1066] A bispecific antibody of the present invention 1065 that binds to EGFR and c-Met. [The present invention 1067] A bispecific antibody of the present invention 1065 or 1066, comprising a first heavy chain (HC1) of SEQ ID NO: 4, a first light chain (LC1) of SEQ ID NO: 5, a second heavy chain (HC2) of SEQ ID NO: 6, and a second light chain (LC2) of SEQ ID NO: 7. [The present invention 1068] A bispecific antibody of the present invention which binds to CD3. [The present invention 1069] An isolated bispecific antibody produced by the method of the present invention. [The present invention 1070] A bispecific antibody of the present invention 1069, which binds to EGFR and c-Met. [The present invention 1071] A bispecific antibody of the present invention 1069 or 1070, comprising a first heavy chain (HC1) of SEQ ID NO: 4, a first light chain (LC1) of SEQ ID NO: 5, a second heavy chain (HC2) of SEQ ID NO: 6, and a second light chain (LC2) of SEQ ID NO: 7. [The present invention 1072] A bispecific antibody of the present invention that binds to CD3. [Brief description of the drawings]

[0008] [Figure 1] FIG. 1 shows an overview of Fab arm exchange. [Diagram 2] FIG. 2 shows the process steps during reduction and ultrafiltration / diafiltration (UF / DF) during the production of a bispecific antibody using Fab arm exchange. [Diagram 3] FIG. 3 shows the molecular details of the thiol-disulfide exchange reaction. [Figure 4] FIG. 4 shows an overview of the reactions that can occur during Fab arm exchange in the presence of 2-MEA. [Diagram 5] FIG. 5 is a schematic diagram showing the UF / DF setup and the location of the DO2 and pH sensors, where RmV=relative millivolts. [Figure 6] FIG. 6 shows the percent (%) DO2 in UF / DF measured in the retentate, permeate and injection lines during the production of bispecific antibody A at low DO2 conditions in UF / DF. As: air saturated. [Figure 7] FIG. 7 shows the percent (%) DO2 in UF / DF measured in the retentate, permeate and injection lines during the production of bispecific antibody B at low DO2 conditions in UF / DF. As: saturated air. [Figure 8]FIG. 8 shows percent (%) DO2 over time during reduction at low DO2 conditions (Low DO Reduction) or at ambient DO2 (Target Reduction). As: saturated air. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0009] Detailed Description of the Invention definition All publications, including but not limited to patents and patent applications, cited in this specification are herein incorporated by reference as if fully set forth.

[0010] It is to be understood that the terms used herein are for the purpose of describing particular embodiments only and are not intended to be limiting. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0011] Although any methods and materials similar or equivalent to those described herein can be used in the practice for testing of the present invention, exemplary materials and methods are described herein. In describing and claiming the present invention, the following terminology will be used.

[0012] As used in this specification and the appended claims, the singular forms "a," "an," and "the" include plural referents unless the content clearly dictates otherwise. Thus, for example, reference to "a cell" includes a combination of two or more cells, and the like.

[0013] The transitional terms "comprising," "consisting essentially of," and "consisting of" are intended to connote their generally accepted meanings in patent language; i.e., (i) "comprising" is synonymous with "including," "containing," or "characterized by," is inclusive or open-ended, and does not exclude additional unrecited elements or method steps; (ii) "consisting of" excludes elements, steps, or ingredients not specified in the claim; and (iii) "consisting essentially of" limits the scope of the claim to the specified materials or steps and those "which do not materially affect the basic and novel characteristics" of the claimed invention. Additionally, an embodiment recited with the phrase "comprising" (or its equivalent) provides as an embodiment the embodiment recited independently with the phrases "consisting of" and "consisting essentially of."

[0014] An "antibody" refers to an immunoglobulin molecule having two heavy chains (HC) and two light chains (LC) interconnected by disulfide bonds. Each heavy chain is composed of a heavy chain variable region (VH) and a heavy chain constant region, which is divided into the regions CH1, hinge, CH2 and CH3. Each light chain is composed of a light chain variable region (VL) and a light chain constant region (CL). VH and VL can be further subdivided into regions of hypervariability called complementarity determining regions (CDRs), which are interposed by framework regions (FRs). Each VH and VL is composed of three CDRs and four FR segments, arranged from the amino terminus to the carboxy terminus in the following order: FR1, CDR1, FR2, CDR2, FR3, CDR3, and FR4. Antibodies include monoclonal antibodies, e.g., murine, human, humanized and chimeric antibodies, bispecific or multispecific antibodies.

[0015] A "complementarity determining region (CDR)" is an antibody region that binds to an antigen. VH has three CDRs (HCDR1, HCDR2, and HCDR3), and VL has three CDRs (LCDR1, LCDR2, and LCDR3). CDRs can be defined using a variety of delineation methods, such as those used by Kabat (Wu et al. (1970) J Exp Med 132: 211-50) (Kabat et al., Sequences of Proteins of Immunological Interest, 5th Ed. Public Health Service, National Institutes of Health, Bethesda, Md., 1991), Chothia (Chothia et al. (1987) J Mol Biol 196: 901-17), IMGT (Lefranc et al. (2003) Dev Comp Immunol 27: 55-77) and AbM (Martin and Thornton (1996) J Bmol Biol 263: 800-15). The correspondence between the various boundary delineations and numbering of variable regions has been described (see, e.g., Lefranc et al. (2003) Dev Comp Immunol 27: 55-77; Honegger and Pluckthun, J Mol Biol (2001) 309:657-70; International ImMunoGeneTics (IMGT) database; web resource, http: / / www_imgt_org). CDR boundaries can be delineated using available programs such as abYsis by UCL Business PLC. The terms "CDR", "HCDR1", "HCDR2", "HCDR3", "LCDR1", "LCDR2" and "LCDR3" as used herein include CDRs defined by any of the methods described above, Kabat, Chothia, IMGT or AbM, unless expressly stated herein.

[0016] Immunoglobulins are assigned to five major classes, IgA, IgD, IgE, IgG, and IgM, depending on the amino acid sequence of the heavy chain constant region. IgA and IgG are further subclassified as isotypes IgA1, IgA2, IgG1, IgG2, IgG3, and IgG4. Antibody light chains of vertebrate species are assigned to one of two clearly distinct types, kappa (κ) and lambda (λ), based on the amino acid sequence of their constant domain.

[0017] "Antigen-binding fragment" refers to a portion of an immunoglobulin molecule that retains the antigen-binding properties of the parent full-length antibody. Exemplary antigen-binding fragments are heavy chain complementarity determining regions (HCDRs) 1, 2 and / or 3, light chain complementarity determining regions (LCDRs) 1, 2 and / or 3, VH, VL, VH and VL, Fab, F(ab')2, Fd and Fv fragments, and domain antibodies (dAbs) consisting of either one VH domain or one VL domain. The VH and VL domains can be linked together via synthetic linkers to form various types of single chain antibody designs, where the VH / VL domains pair intramolecularly, or, when the VH and VL domains are expressed by separate chains, pair intermolecularly to form a monovalent antigen-binding site, such as a single chain Fv (scFv) or diabody; e.g., as described in International Patent Publication Nos. (Int. Pat. Publ. No.) WO1998 / 44001; Int. Pat. Publ. No. WO1988 / 01649; Int. Pat. Publ. No. WO1994 / 13804; Int. Pat. Publ. No. WO1992 / 01047.

[0018] "Monoclonal antibody" refers to an antibody obtained from a substantially homogeneous population of antibody molecules, i.e., the individual antibodies constituting the population are identical except for possible well-known changes such as removal of the C-terminal lysine from the antibody heavy chain, or post-translational modifications such as isomerization or deamidation of amino acids, oxidation of methionine, deamidation of asparagine or glutamine. A monoclonal antibody usually binds to one antigenic epitope. A bispecific monoclonal antibody binds to two different antigenic epitopes. A monoclonal antibody may have heterogeneous glycosylation within the antibody population. A monoclonal antibody may be monospecific or multispecific, such as bispecific, monovalent, bivalent or multivalent.

[0019] "Fab arm" refers to one heavy-light chain pair of an antibody.

[0020] "Homodimerization" refers to the interaction of two heavy chains that have identical CH3 amino acid sequences.

[0021] A "homodimer" refers to an antibody having two heavy chains with identical CH3 amino acid sequences.

[0022] "Heterodimerization" refers to the interaction of two heavy chains that have non-identical CH3 amino acid sequences.

[0023] A "heterodimer" refers to an antibody having two heavy chains which differ in the amino acid sequence of their CH3 regions by one or more amino acids.

[0024] "Fc region" or "Fc domain" refers to the region of an antibody that includes at least a portion of the hinge region, the CH2 region, and the CH3 region. The Fc region can be generated by digestion of an antibody with papain, or pepsing, and the Fc region is the resulting fragment that includes one or both CH2-CH3 regions of the immunoglobulin and a portion of the hinge region. The constant domains of the antibody heavy chain define the antibody's isotype, e.g., IgG1, IgG2, IgG3, IgG4, IgA1, IgA2, IgE. The Fc region, together with cell surface Fc receptors and proteins of the complement system, mediates the effector functions of the antibody.

[0025] "CH1 region" or "CH1 domain" refers to the CH1 region of an immunoglobulin. The CH1 region of a human IgG1 antibody corresponds to amino acid residues 118-215. However, the CH1 region may be any of the other antibody isotypes described herein.

[0026] "CH2 region" or "CH2 domain" refers to the CH2 region of an immunoglobulin. The CH2 region of a human IgG1 antibody corresponds to amino acid residues 231-340. However, the CH2 region may be any of the other antibody isotypes described herein.

[0027] "CH3 region" or "CH3 domain" refers to the CH3 region of an immunoglobulin. The CH3 region of a human IgG1 antibody corresponds to amino acid residues 341-446. However, the CH3 region may be any of the other antibody isotypes described herein.

[0028] "Hinge" or "hinge region" refers to the hinge region of an immunoglobulin. The hinge region of a human IgG1 antibody generally corresponds to amino acids 216-230 according to the EU numbering system. The "hinge" is also considered to include additional residues referred to as the upper and lower hinge regions, for example including amino acid residues 216-239.

[0029] "Mixture" refers to an aqueous solution of two or more antibodies.

[0030] A "reducing agent" refers to an agent capable of reducing the interchain disulfide bond in the hinge region of an antibody.

[0031] "GMP-compliant conditions" refers to manufacturing under Good Manufacturing Practice (CGMP) regulations enforced by the FDA. CGMP provides a system to ensure proper design, monitoring, and control of manufacturing processes and facilities. Compliance with CGMP regulations ensures the identity, strength, quality, and purity of drug products by requiring drug manufacturers to properly control manufacturing operations. This includes establishing a strong quality control system, obtaining raw materials of appropriate quality, establishing robust work procedures, detecting and investigating deviations in product quality, and maintaining reliable testing laboratories. This formal control system of pharmaceutical companies, if properly implemented, helps prevent instances of contamination, mix-ups, bias, adulteration, and errors. This ensures that drug products meet quality standards.

[0032] "Drug substance" or "DS" means a substance or mixture of substances that is intended to be used in the manufacture of a medicinal product and that, when used in the manufacture of a drug, is an active ingredient of the drug. Such a substance is intended to provide pharmacological activity or other direct effect in the diagnosis, cure, mitigation, treatment, or prevention of disease, or to affect the structure or function of the body.

[0033] "Drug product" or "DP" generally, but not necessarily, refers to a finished dosage form, e.g., tablet, capsule, or solution, that contains a pharma- ceutical active ingredient (e.g., drug substance) along with inactive ingredients.

[0034] "Reference product" refers to an approved biological product to which a biosimilar product is compared. A reference product is, inter alia, one that has been approved based on a full complement of safety and efficacy data and has been approved in at least one of the United States, Europe, or Japan.

[0035] "Biosimilar" (approved reference product / biological product) refers to a biological product that is highly similar to the reference product, despite minor differences in clinically inactive ingredients, and that has no clinically meaningful differences in safety, purity, and potency between the biosimilar and the reference product, based on data obtained from (a) analytical studies showing that the biological product is highly similar to the reference product, despite minor differences in clinically inactive ingredients; (b) animal studies (including evaluation of toxicity); and / or (c) clinical trials (including evaluation of immunogenicity and pharmacokinetics or pharmacodynamics) sufficient to demonstrate safety, purity, and potency under one or more appropriate conditions of use (for which the reference product is licensed and intended to be used, and for which licensing of the biosimilar is sought). A biosimilar may be an interchangeable product that can be dispensed in place of the reference product at a pharmacy without the intervention of a prescribing physician. To meet the additional criterion of "interchangeability," a biosimilar is expected to produce the same clinical outcome in a given patient as the reference product, and, when the biosimilar is administered multiple times to an individual, the risks of safety or reduced efficacy of alternating or replacing the biosimilar with the reference product are not greater than the risks of using the reference product without such alternation or replacement. The biosimilar utilizes the same mechanism of action for the proposed conditions of use, to the extent that the mechanism of action is known in the reference product. The proposed, labeled, recommended, or suggested conditions of use for the biosimilar must have been previously approved for the reference product. The route of administration, dosage form, and / or strength of the biosimilar must be the same as those of the reference product, and the biosimilar must be manufactured, processed, packaged, or stored in facilities that meet standards designed to ensure that the biosimilar remains safe, pure, and potent. A biosimilar may contain minor changes in amino acid sequence, such as N- or C-terminal truncations, that are not expected to alter the performance of the biosimilar compared to the reference product. The reference product may be one approved in at least one of the United States, Europe, or Japan.

[0036] "Isolated" refers to a homogenous population of molecules (e.g., synthetic polynucleotides or proteins such as antibodies) that have been substantially separated and / or purified from other components of the system in which the molecule is produced, e.g., a recombinant cell, as well as proteins that have been subjected to at least one purification or separation step. An "isolated antibody" refers to an antibody that is substantially free of other cellular material and / or chemicals and includes antibodies isolated to greater degrees of purity, e.g., 80%, 81%, 82%, 83%, 84%, 85%, 86%, 87%, 88%, 89%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, 99% or 100% purity.

[0037] "Humanized antibody" refers to an antibody in which at least one CDR is derived from a non-human species and at least one framework is derived from a human immunoglobulin sequence. Humanized antibodies may contain substitutions in the framework, i.e., the framework may not be an exact copy of an expressed human immunoglobulin or human immunoglobulin germline gene sequence.

[0038] "Human antibody" refers to an antibody that is optimized to produce minimal immune responses when administered to a human subject. The variable regions of a human antibody are derived from human immunoglobulin sequences. If the human antibody contains a constant region or a portion of a constant region, the constant region is also derived from a human immunoglobulin sequence. If the variable region of a human antibody is obtained from a system that uses human germline immunoglobulin genes or rearranged immunoglobulin genes, the human antibody contains heavy and light chain variable regions "derived" from sequences of human origin. Exemplary such systems are human immunoglobulin gene libraries displayed on phage, and transgenic non-human animals such as mice or rats carrying human immunoglobulin loci. "Human antibodies" usually contain amino acid differences when compared to immunoglobulins expressed in humans, due to differences between the systems used to obtain the human antibodies and the human immunoglobulin loci, the introduction of somatic mutations, or the deliberate introduction of substitutions into the framework or CDRs or both. Typically, a "human antibody" is an amino acid sequence that is at least about 80%, 81%, 82%, 83%, 84%, 85%, 86%, 87%, 88%, 89%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98% or 99% identical to an amino acid sequence encoded by a human germline or rearranged immunoglobulin gene. In some cases, a "human antibody" may comprise a consensus framework sequence obtained from human framework sequence analysis (e.g., as described in Knappik et al., (2000) J Mol Biol 296:57-86) or a synthetic HCDR3 incorporated into a library of human immunoglobulin genes displayed on phage (e.g., as described in Shi et al., (2010) J Mol Biol 397:385-96, and WO2009 / 085462). Antibodies in which at least one CDR is derived from a non-human species are not included in the definition of "human antibody."

[0039] "Recombinant" refers to DNA, antibodies, and other proteins prepared, expressed, created, or isolated by recombinant means, where segments from different sources are joined to result in the recombinant DNA, antibody, or protein. "Recombinant antibodies" include all antibodies prepared, expressed, created, or isolated by recombinant means, including, for example, antibodies isolated from animals (e.g., mice) that are transgenic or transchromosomal for human immunoglobulin genes, or hybridomas made therefrom (discussed further below), antibodies isolated from host cells transformed to express antibodies, antibodies isolated from recombinant combinatorial antibody libraries, and antibodies prepared, expressed, created, or isolated by other means, including splicing human immunoglobulin gene sequences into other DNA sequences, or antibodies made in vitro using Fab arm exchange, such as bispecific antibodies.

[0040] "Bispecific" refers to an antibody that specifically binds to two different antigens or two different epitopes within the same antigen. Bispecific antibodies may have cross-reactivity to other related antigens or may bind to an epitope shared between two or more different antigens.

[0041] "Multispecific" refers to an antibody that specifically binds to at least two different antigens or at least two different epitopes within the same antigen. A multispecific antibody may, for example, bind to two, three, four, or five different antigens or different epitopes within the same antigen.

[0042] "About" means within an acceptable error range of a particular value as determined by one of ordinary skill in the art, which depends in part on how the value is measured or determined, i.e., the limitations of the measurement system. Unless otherwise specified in the Examples or elsewhere herein in connection with a particular assay, result, or embodiment, "about" means within one standard deviation, or up to a 5% range, whichever is greater, per practice in the art.

[0043] "Variant" refers to a polypeptide or polynucleotide that differs from a reference polypeptide or polynucleotide by one or more modifications, such as, for example, a substitution, insertion, deletion, or the like.

[0044] "Mutation" refers to an artificial or naturally occurring change in a polypeptide or polynucleotide sequence compared to a reference sequence. The change may be one or more amino acid or polynucleotide substitutions, insertions or deletions.

[0045] The numbering of amino acid residues in antibody constant regions throughout this specification is according to the EU index as set forth in Kabat et al., Sequences of Proteins of Immunological Interest, 5th Ed. Public Health Service, National Institutes of Health, Bethesda, MD. (1991), unless otherwise stated. Numbering of antibody constant chains can be found, for example, on the ImMunoGeneTics website, IMGT Web Resources, IMGT Scientific Charts.

[0046] The mutations in the CH3 region described herein are represented as the modification position(s) in the first CH3 domain of the first heavy chain / the modification position(s) in the second CH3 domain of the second heavy chain. For example, F405L / K409R means that the first CH3 region has an F405L mutation and the second CH3 region has a K09R mutation. L351Y_F405A_Y407V / T394W means that the first CH3 region has an L351Y, F40FA and Y407V mutation and the second CH3 region has a T394W mutation. D399FHKRQ / K409AGRH refers to the mutations where D399 can be replaced with F, H, K, R or Q, and K409 can be replaced with A, G, R or H.

[0047] Standard one-letter and three-letter amino acid codes, as shown in Table 1, are used herein.

[0048] [Table 1]

[0049] Methods of the Invention The present invention provides a method for producing heterodimeric antibodies using Fab arm exchange. The present invention is based, at least in part, on the realization that contrary to what has been disclosed in the literature (see, for example, US2014 / 0303356), oxygen is not required for the reformation of disulfide bonds after reduction in order to form stable heterodimeric antibodies during Fab arm exchange, thus providing an alternative approach for process control in the large-scale production of heterodimeric antibodies (e.g., bispecific antibodies).

[0050] The present invention provides a method for producing a heterodimeric antibody, comprising the steps of: providing a first homodimeric antibody comprising a first Fc region of an immunoglobulin comprising a first CH3 region and a second homodimeric antibody comprising a second Fc region of an immunoglobulin comprising a second CH3 region, wherein the amino acid sequences of the first CH3 region and the second CH3 region are different and the amino acid sequences of the first CH3 region and the second CH3 region are such that a heterodimeric interaction between the first CH3 region and the second CH3 region is stronger than a homodimeric interaction between the first CH3 regions or a homodimeric interaction between the second CH3 regions; combining the first homodimeric antibody and the second homodimeric antibody into a mixture; and A step of incubating the mixture in the presence of a reducing agent and ambient dissolved oxygen (DO2) to produce a heterodimeric antibody, wherein the method lacks one or more of the following steps during production of the heterodimeric antibody: measuring the percent (%) DO2 in the mixture, controlling the % DO2 in the mixture, or adding oxygen to the mixture.

[0051] In some embodiments, during production of the heterodimeric antibody, the %DO2 in the mixture is about 10% to about 90%.

[0052] In some embodiments, the %DO2 in the mixture is greater than or equal to about 30% in the mixture prior to adding the denaturing agent.

[0053] The present invention also provides a method for producing a heterodimeric antibody, comprising the steps of: providing a first homodimeric antibody comprising a first Fc region of an immunoglobulin comprising a first CH3 region and a second homodimeric antibody comprising a second Fc region of an immunoglobulin comprising a second CH3 region, wherein the amino acid sequences of the first CH3 region and the second CH3 region are different and the amino acid sequences of the first CH3 region and the second CH3 region are such that a heterodimeric interaction between the first CH3 region and the second CH3 region is stronger than a homodimeric interaction between the first CH3 regions or a homodimeric interaction between the second CH3 regions; combining the first homodimeric antibody and the second homodimeric antibody into a mixture; incubating the mixture in the presence of a reducing agent; and A step of removing the reducing agent to produce a heterodimeric antibody, wherein in step c), step d), or both steps c) and d), the percentage (%) of dissolved oxygen (DO2) is controlled to be about 30% or less.

[0054] In some embodiments, the %DO2 of the mixture is about 25% or less, 20% or less, about 15% or less, about 10% or less, about 9% or less, about 8% or less, about 7% or less, about 6% or less, about 5% or less, about 4% or less, about 3% or less, about 2% or less, or about 1% or less.

[0055] In some embodiments, %DO2 is controlled by displacing oxygen from the mixture by overlaying the mixture with an inert gas, such as nitrogen. The concentration of dissolved oxygen can be monitored by known methods, such as by using a dissolved oxygen probe.

[0056] "Stronger" in terms of the heterodimeric interaction between the first and second CH3 regions may be more than two times stronger, e.g., more than three times stronger, more than four times stronger, or more than five times stronger than the strongest homodimeric interaction between the first CH3 regions or the homodimeric interaction between the second CH3 regions. The strength of the interaction between the CH3 domains can be measured using mass spectrometry. In an exemplary assay, constructs containing the first and second CH3 regions, or both containing CH2 regions, are made using standard molecular biology techniques. Samples containing the first CH3 domain, the second CH3 domain, or the first and second CH3 domains are prepared and buffer exchanged into 100 mM ammonium acetate pH 7 using a 10 kDa MWCO spin filter column. Load an aliquot (~1 µL) of the serially diluted sample (20 µM-25 nM; monomer equivalent) into a gold-plated borosilicate capillary and analyze it with an LCT mass spectrometer (Waters). s is defined as the area of ​​the monomer peak as a function of the area of ​​all peaks in the spectrum (M s / (M s +D s ), where D s = dimer signal). Monomer concentration at equilibrium [M] eq is M s ·[M]0, where [M]0 is the total protein concentration in terms of monomers. The dimer concentration at equilibrium [D] eq ([M]0-[M] eq ) / 2. Then, the K D [D] eq Against [M] eq 2 It is extracted from the slope of the plot of

[0057] In some embodiments, the first homodimeric antibody and the second homodimeric antibody are combined into a mixture in a molar ratio of about 1:1.

[0058] In some embodiments, the first homodimeric antibody and the second homodimeric antibody are combined in a mixture in a molar ratio of about 1.05:1.

[0059] In some embodiments, the first homodimeric antibody and the second homodimeric antibody are combined into a mixture at a molar ratio of about 1:1.03 to about 1:2. In some embodiments, the first homodimeric antibody and the second homodimeric antibody are combined into a mixture at a molar ratio of about 1:1.05 to 1:1.5. In some embodiments, the first homodimeric antibody and the second homodimeric antibody are combined into a mixture at a molar ratio of about 1:1.1 to 1:1.5. In some embodiments, the first homodimeric antibody and the second homodimeric antibody are combined into a mixture at a molar ratio of about 1:1.1 to 1:1.4. In some embodiments, the first homodimeric antibody and the second homodimeric antibody are combined into a mixture at a molar ratio of about 1:1.15 to 1:1.35. In some embodiments, the first homodimeric antibody and the second homodimeric antibody are combined into a mixture at a molar ratio of about 1:1.2 to 1:1.3.

[0060] In some embodiments, the total concentration of immunoglobulins in the mixture is about 1 g / L to about 70 g / L. In some embodiments, the total concentration of immunoglobulins in the mixture is about 8 g / L to about 50 g / L. In some embodiments, the total concentration of immunoglobulins in the mixture is about 8 g / L to about 13 g / L. In some embodiments, the total concentration of immunoglobulins in the mixture is about 9.0 g / L. In some embodiments, the total concentration of immunoglobulins in the mixture is about 9.5 g / L. In some embodiments, the total concentration of immunoglobulins in the mixture is about 10 g / L. In some embodiments, the total concentration of immunoglobulins in the mixture is about 10.5 g / L. In some embodiments, the total concentration of immunoglobulins in the mixture is about 11.0 g / L. In some embodiments, the total concentration of immunoglobulins in the mixture is about 11.5 g / L. In some embodiments, the total concentration of immunoglobulins in the mixture is about 12.0 g / L. In some embodiments, the total concentration of immunoglobulins in the mixture is about 12.5 g / L. In some embodiments, the total concentration of immunoglobulins in the mixture is about 13.0 g / L. In some embodiments, the total concentration of immunoglobulins in the mixture is about 13.5 g / L. In some embodiments, the total concentration of immunoglobulins in the mixture is about 14.0 g / L. In some embodiments, the total concentration of immunoglobulins in the mixture is about 14.5 g / L. In some embodiments, the total concentration of immunoglobulins in the mixture is about 15.0 g / L. In some embodiments, the total concentration of immunoglobulins in the mixture is about 20.0 g / L. In some embodiments, the total concentration of immunoglobulins in the mixture is about 25.0 g / L. In some embodiments, the total concentration of immunoglobulins in the mixture is about 30.0 g / L. In some embodiments, the total concentration of immunoglobulins in the mixture is about 35.0 g / L. In some embodiments, the total concentration of immunoglobulins in the mixture is about 40.0 g / L. In some embodiments, the total concentration of immunoglobulins in the mixture is about 45.0 g / L. In some embodiments, the total concentration of immunoglobulins in the mixture is about 50.0 g / L. In some embodiments, the total concentration of immunoglobulins in the mixture is about 55.0 g / L. In some embodiments, the total concentration of immunoglobulins in the mixture is about 60.0 g / L. In some embodiments, the total concentration of immunoglobulins in the mixture is about 65.0 g / L.In one embodiment, the total concentration of immunoglobulin in the mixture is about 70.0 g / L.

[0061] In some embodiments, the mixture is incubated in the presence of the reducing agent for about 10 minutes or more. In some embodiments, the mixture is incubated in the presence of the reducing agent for about 10 minutes to about 30 hours. In some embodiments, the mixture is incubated in the presence of the reducing agent for about 10 minutes to about 24 hours. In some embodiments, the mixture is incubated in the presence of the reducing agent for about 15 minutes. In some embodiments, the mixture is incubated in the presence of the reducing agent for about 20 minutes. In some embodiments, the mixture is incubated in the presence of the reducing agent for about 30 minutes. In some embodiments, the mixture is incubated in the presence of the reducing agent for about 40 minutes. In some embodiments, the mixture is incubated in the presence of the reducing agent for about 50 minutes. In some embodiments, the mixture is incubated in the presence of the reducing agent for about 1 hour. In some embodiments, the mixture is incubated in the presence of the reducing agent for about 2 hours. In some embodiments, the mixture is incubated in the presence of the reducing agent for about 3 hours. In some embodiments, the mixture is incubated in the presence of the reducing agent for about 4 hours. In some embodiments, the mixture is incubated in the presence of the reducing agent for about 4 hours. In some embodiments, the mixture is incubated in the presence of the reducing agent for about 5 hours. In some embodiments, the mixture is incubated in the presence of the reducing agent for about 6 hours. In some embodiments, the mixture is incubated in the presence of a reducing agent for about 7 hours. In some embodiments, the mixture is incubated in the presence of a reducing agent for about 8 hours. In some embodiments, the mixture is incubated in the presence of a reducing agent for about 9 hours. In some embodiments, the mixture is incubated in the presence of a reducing agent for about 10 hours. In some embodiments, the mixture is incubated in the presence of a reducing agent for about 11 hours. In some embodiments, the mixture is incubated in the presence of a reducing agent for about 12 hours. In some embodiments, the mixture is incubated in the presence of a reducing agent for about 13 hours. In some embodiments, the mixture is incubated in the presence of a reducing agent for about 14 hours. In some embodiments, the mixture is incubated in the presence of a reducing agent for about 15 hours. In some embodiments, the mixture is incubated in the presence of a reducing agent for about 16 hours. In some embodiments, the mixture is incubated in the presence of a reducing agent for about 17 hours. In some embodiments, the mixture is incubated in the presence of a reducing agent for about 18 hours. In some embodiments, the mixture is incubated in the presence of a reducing agent for about 19 hours. In some embodiments, the mixture is incubated in the presence of a reducing agent for about 20 hours.In some embodiments, the mixture is incubated in the presence of the reducing agent for about 21 hours. In some embodiments, the mixture is incubated in the presence of the reducing agent for about 22 hours. In some embodiments, the mixture is incubated in the presence of the reducing agent for about 23 hours. In some embodiments, the mixture is incubated in the presence of the reducing agent for about 24 hours. In some embodiments, the mixture is incubated in the presence of the reducing agent for about 25 hours. In some embodiments, the mixture is incubated in the presence of the reducing agent for about 26 hours. In some embodiments, the mixture is incubated in the presence of the reducing agent for about 27 hours. In some embodiments, the mixture is incubated in the presence of the reducing agent for about 28 hours. In some embodiments, the mixture is incubated in the presence of the reducing agent for about 29 hours. In some embodiments, the mixture is incubated in the presence of the reducing agent for about 30 hours.

[0062] In some embodiments, the reducing agent is 2-mercaptoethylamine (2-MEA). In some embodiments, the reducing agent is a chemical derivative of 2-MEA. In some embodiments, the reducing agent is L-cysteine. In some embodiments, the reducing agent is D-cysteine. In some embodiments, the reducing agent is glutathione. In some embodiments, the reducing agent is tris(2-carboxyethyl)phosphine.

[0063] In some embodiments, the concentration of the reducing agent in the mixture is about 0.1 mM to about 1 M. In some embodiments, the concentration of the reducing agent in the mixture is about 1.0 mM to about 500 mM. In some embodiments, the concentration of the reducing agent in the mixture is about 5.0 mM to about 100 mM. In some embodiments, the concentration of the reducing agent in the mixture is about 10 mM. In some embodiments, the concentration of the reducing agent in the mixture is about 15 mM. In some embodiments, the concentration of the reducing agent in the mixture is about 20 mM. In some embodiments, the concentration of the reducing agent in the mixture is about 25 mM. In some embodiments, the concentration of the reducing agent in the mixture is about 30 mM. In some embodiments, the concentration of the reducing agent in the mixture is about 35 mM. In some embodiments, the concentration of the reducing agent in the mixture is about 40 mM. In some embodiments, the concentration of the reducing agent in the mixture is about 50 mM. In some embodiments, the concentration of the reducing agent in the mixture is about 60 mM. In some embodiments, the concentration of the reducing agent in the mixture is about 70 mM. In some embodiments, the concentration of the reducing agent in the mixture is about 80 mM. In some embodiments, the concentration of the reducing agent in the mixture is about 90 mM. In some embodiments, the concentration of the reducing agent in the mixture is about 100 mM.

[0064] In some embodiments, the concentration of 2-MEA in the mixture is about 10 mM to about 100 mM. In some embodiments, the concentration of 2-MEA in the mixture is about 20 mM to about 90 mM. In some embodiments, the concentration of 2-MEA in the mixture is about 20 mM to about 80 mM. In some embodiments, the concentration of 2-MEA in the mixture is about 20 mM to about 70 mM. In some embodiments, the concentration of 2-MEA in the mixture is about 20 mM to about 60 mM. In some embodiments, the concentration of 2-MEA in the mixture is about 20 mM to about 50 mM. In some embodiments, the concentration of 2-MEA in the mixture is about 20 mM to about 40 mM. In some embodiments, the concentration of 2-MEA in the mixture is about 20 mM. In some embodiments, the concentration of 2-MEA in the mixture is about 25 mM. In some embodiments, the concentration of 2-MEA in the mixture is about 30 mM. In some embodiments, the concentration of 2-MEA in the mixture is about 35 mM. In some embodiments, the concentration of 2-MEA in the mixture is about 40 mM.

[0065] In some embodiments, the mass ratio of the total immunoglobulins in the mixture to the total reducing agent is about 1.0 to about 5.0. In some embodiments, the mass ratio of the total immunoglobulins in the mixture to the total reducing agent is about 1.4 to about 3.8. In some embodiments, the mass ratio of the total immunoglobulins in the mixture to the total reducing agent is about 1.4 to about 3.5. In some embodiments, the mass ratio is about 1.8 to about 3.8. In some embodiments, the mass ratio is about 2.3 to about 3.0. In some embodiments, the mass ratio is about 1.6 to about 2.1. In some embodiments, the mass ratio is about 1.0, 1.1, 1.2, 1.3, 1.4, 1.5, 1.6, 1.7, 1.8, 1.9, 2.0, 2.1, 2.2, 2.3, 2.4, 2.5, 2.6, 2.7, 2.8, 2.9, 3.0, 3.1, 3.2, 3.3, 3.4, 3.5, 3.6, 3.7, 3.8, 3.9, 4.0, 4.1, 4.2, 4.3, 4.4, 4.5, 4.6, 4.7, 4.8, 4.9, or 5.0. "Mass ratio" refers to the ratio of total immunoglobulin in grams per liter to total reducing agent in grams per liter. "Total immunoglobulin" refers to the amount of two parent antibodies in the mixture at the start of the reduction process.

[0066] In some embodiments, the mixture comprises a buffer. In some embodiments, the buffer comprises a sodium acetate buffer. In some embodiments, the buffer further comprises NaCl. In some embodiments, the buffer comprises about 100 mM sodium acetate and about 30 mM NaCl. In some embodiments, the pH of the buffer is about 7.3. Other buffers, such as 1× Dulbecco's phosphate buffered saline (DPBS), sodium phosphate buffer, potassium phosphate buffer, Tris buffer, histidine buffer, or citrate buffer, can also be used.

[0067] In some embodiments, the method further comprises removing the reducing agent from the mixture.

[0068] In some embodiments, the reducing agent is removed by filtration.

[0069] In some embodiments, the filtration is diafiltration.

[0070] In some embodiments, the first homodimeric antibody and the second homodimeric antibody are of the IgG1, IgG2 or IgG4 isotype.

[0071] In some embodiments, the first homodimeric antibody and the second homodimeric antibody are of the IgG1 isotype. In some embodiments, the first homodimeric antibody and the second homodimeric antibody are of the IgG2 isotype. In some embodiments, the first homodimeric antibody and the second homodimeric antibody are of the IgG4 isotype.

[0072] In some embodiments, compared to the wild-type IgG1 of SEQ ID NO: 1, the first CH3 domain and the second CH3 domain contain the following mutations: F405L / K409R, T350I_K370T_F405L / K409R, K370W / K409R, D399AFGHILMNRSTVWY / K409R, T366ADEFGHILMQVY / K409R, L368ADEGHNRSTVQ / K409AGRH, D399FHKRQ ... T366ADEFGHILMQVY / K4 K409AGRH, F405IKLSTVW / K409AGRH, Y407LWQ / K409AGRH, T366Y / F405A, T366W / F405W, F405W / Y407A, T394W / Y407T, T394S / Y407A, T366W / T394S, F405W / T394S, T366W / T366S_L368A_Y407V, L351Y_F405A_Y407V / T394W, T366I_K392M_T394W / F405A_Y407V, T366L_K392M_T394W / F405A_Y407V, L351Y_Y407A / T366A_K409F, L351Y _Y407A / T366V_K409F, Y407A / T366A_K409F, T350V_L351Y_F405A_Y407V / T350V_T366L_K392L_T394W, K409 D / D399K, K409E / D399R, K409D_K360D / D399K_E356K, K409D_K360D / D399E_E356K, K409D_K370D / D399K_E35 7K, K409D_K370D / D399E_E357K, K409D_K392D / D399K_E356K_E357K, or K409D_K392D / D399E_E356K_E357K. Mutations can also be introduced into wild type IgG2 of SEQ ID NO: 2 or wild type IgG4 of SEQ ID NO: 3, where, as is well known, the wild type residue at a particular mutation site may not correspond to the IgG1 residue due to sequence differences between wild type IgG1, IgG2 and IgG4. For example, the IgG4 mutation wild type / F405L_R409K corresponds to the IgG1 mutation F405L / K409R and the IgG4 mutation R409D / D399K corresponds to the IgG1 mutation K409D / D399K.The mutations are compared to the reference wild type IgG1 of SEQ ID NO:1, IgG2 of SEQ ID NO:2, and IgG4 of SEQ ID NO:3.

[0073] In some embodiments, the first Fc region and / or the second Fc region comprises one or more mutations that modulate binding of the first Fc region and / or the second Fc region to Fcγ receptors (FcγR), FcRn, or protein A, compared to a wild-type IgG1 of SEQ ID NO: 1, a wild-type IgG2 of SEQ ID NO: 2, or a wild-type IgG4 of SEQ ID NO: 3. TIFF0007645799000002.tif104130

[0074] In some embodiments, the FcγR is FcγRI, FcγRIIa, FcγRIIb, or FcγRIII.

[0075] In some embodiments, the one or more substitutions that modulate binding of the first Fc region and / or the second Fc region to Fcγ are selected from the group consisting of L234A_L235A, F234A_L235A, S228P_F234A_L235A, S228P_L234A_L235A, V234A_G237A_P238S_H268A_V309L_A330S _P331S, V234A_G237A, H268Q_V309L_A330S_P331S, S267E_L328F, L234F_L235E_D265A, L234A_L235A_G237A_P238S_H268A_A330S_P331S, or S228P_F234A_L235A_G237A_P238S.

[0076] In some embodiments, the one or more substitutions that modulate binding of the first Fc region and / or second Fc region to FcRn are M428L_N434S, M252Y_S254T_T256E, T250Q_M428L, N434A and T307A_E380A_N434A, H435A, P257I_N434H, D376V_N434H, M252Y_S254T_T256E_H433K_N434F, T308P_N434A, or H435R.

[0077] In some embodiments, the one or more substitutions that modulate binding of the first Fc region and / or the second Fc region to Protein A are Q311R, Q311K, T307P_L309Q, T307P_V309Q, T307P_L309Q_Q311R, T307P_V309Q_Q311R, H435R, or H435R_Y436F.

[0078] In some embodiments, the process for producing the heterodimeric antibody is carried out under GMP compliant conditions.

[0079] In some aspects, the step of generating the heterodimeric antibody is performed during the manufacture of a drug substance that includes the heterodimeric antibody.

[0080] In some aspects, the heterodimeric antibody is a bispecific antibody.

[0081] In some embodiments, the bispecific antibody binds to CD3, BCMA or CD123.

[0082] In some embodiments, the process of generating heterodimeric antibodies is carried out during the manufacture of an innovator pharmaceutical product.

[0083] In some embodiments, the process of generating heterodimeric antibodies is carried out during the manufacture of generic drugs.

[0084] Fc region mutations that promote Fab arm exchange and heterodimer formation Heterodimeric antibodies can be generated using Fab arm exchange, in which one heavy chain and attached light chain (half arm) of one parent homodimeric antibody is exchanged for one heavy chain and attached light chain of another parent homodimeric antibody to form a heterodimeric antibody consisting of two heavy chains and two attached light chains (van der Neut Kolfschoten et al., (2007) Science 317:1554-1557).

[0085] The two homodimeric parent antibodies are designed with asymmetric mutations in the CH3 region that favor Fab arm exchange and heterodimeric antibody formation upon reduction and reformation of disulfide bridges in the antibody hinge region. An illustration of the Fab arm exchange reaction is shown in Figure 1. When a reducing agent is introduced into a mixture of the two parent homodimeric antibodies, the half arms dissociate and the asymmetric CH3 mutations favor the reformation of the heterodimeric antibody. The reformation of disulfide bridges between the half arms then stabilizes the formed heterodimeric antibody (Gramer et al. (2013) MAbs 5:962-973).

[0086] In the method of the present invention, any mutation of the CH3 region that promotes CH3 heterodimer formation, as described herein, can be used. For the modification of the CH3 region that promotes heterodimer formation, several approaches are known. Usually, in all such approaches, the first CH3 region and the second CH3 region are modified in a complementary manner, so that each CH3 region (or the heavy chain that contains it) can no longer form a homodimer with itself, but is forced to form a heterodimer with the other CH3 region that is complementarily modified (the first CH3 region and the second CH3 region form a heterodimer, and homodimers cannot be formed between the two first CH3 regions or between the two second CH3 regions).

[0087] CH3 mutations that favor Fab arm exchange include Duobody® mutations (Genmab), Knob-in-Hole mutations (Genentech), electrostatically matched mutations (Chugai, Amgen, NovoNordisk, Oncomed), SEEDbody (Strand Exchange Engineered Domain body) (EMD Serono), and other asymmetric mutations (Zymeworks, etc.).

[0088] Duobody® mutations (Genmab) are disclosed, for example, in US 9,150,663 and US 2014 / 0303356 and include the mutations F405L / K409R, wild type / F405L_R409K, T350I_K370T_F405L / K409R, K370W / K409R, D399AFGHILMNRSTVWY / K409R, T366ADEFGHILMQVY / K409R, L368ADEGHNRSTVQ / K409AGRH, D399FHKRQ / K409AGRH, F405IKLSTVW / K409AGRH, and Y407LWQ / K409AGRH.

[0089] Knob-in-hole mutations are disclosed, for example, in WO 1996 / 027011, and involve mutations at the interface of the CH3 domains; in which an amino acid with a small side chain (the hole) is introduced into the first CH3 domain and an amino acid with a large side chain (the knob) is introduced into the second CH3 domain, resulting in preferential interactions between the first and second CH3 domains. Exemplary CH3 domain mutations that form knobs and holes include T366Y / F405A, T366W / F405W, F405W / Y407A, T394W / Y407T, T394S / Y407A, T366W / T394S, F405W / T394S, T366W / T366S_L368A_Y407V, etc.

[0090] Heavy chain heterodimer formation can be promoted using electrostatic interactions by substituting positively charged residues in the first CH3 region and negatively charged residues in the second CH3 region, as described in US2010 / 0015133, US2009 / 0182127, US2010 / 028637 or US2011 / 0123532.

[0091] Other asymmetric mutations that can be used to promote heavy chain heterodimer formation include L351Y_F405A_Y407V / T394W, T366I_K392M_T394W / F405A_Y407V, T366L ... You have M_T394W / F405A_Y407V, L351Y_Y407A / T366A_K409F, L351Y_Y407A / T366V_K409F, Y407A / T366A_K409F, or T350V_L351Y_F405A_Y407V / T350V_T366L_K392L_T394W.

[0092] SEEDbody mutations involve the substitution of selected IgG residues with IgA residues to promote heavy chain heterodimer formation, as described in US20070287170.

[0093] Other exemplary mutations that can be used include R409D_K370E / D399K_E357K, S354C_T366W / Y349C_T366S_L368A_Y407V, Y349C_T366W / S354C_T366S_L368A_Y407V, T366K / L351D, L351K / Y349E, L351K / Y349D, L351K / L 368E, L351Y_Y407A / T366A_K409F, L351Y_Y407A / T366V_K409F, K392D / D399K, K392D / E356K, K253E_D282K_K322D / D239K_E240K_K292D, K392D_K409D / D356K_D399K, and are described in WO2007 / 147901, WO 2011 / 143545, WO2013157954, WO2013096291 and US2018 / 0118849.

[0094] These different approaches for Fab arm exchange can be combined with various bispecific antibody formats, including, for example, VH / VL engineering such as a VH / VL domain swap, a CH1 / CL domain swap, the use of a common light chain as described in WO98050431, or the use of a tethered light chain, including an inside-out tethered light chain as described in US9062120.

[0095] Further engineering of heterodimeric antibodies Fc Engineering In addition to mutations in the CH3 region that promote heterodimer formation, the antibodies used in the methods of the invention can contain mutations in the Fc region that modulate the effector function or half-life of the antibody. Additionally, the antibodies used in the methods of the invention can also contain Fc mutations that modulate binding of the antibody to Protein A, thus facilitating purification of the antibody.

[0096] Positions in Fc that may be mutated to modulate the half-life (e.g., binding to FcRn) of the antibody include positions 250, 252, 253, 254, 256, 257, 307, 376, 380, 428, 434 and 435. Exemplary mutations that can be made alone or in combination are the following mutations: T250Q, M252Y, I253A, S254T, T256E, P257I, T307A, D376V, E380A, M428L, H433K, N434S, N434A, N434H, N434F, H435A and H435R. Exemplary mutations that can be made, either alone or in combination, to increase the half-life of an antibody are M428L / N434S, M252Y / S254T / T256E, T250Q / M428L, N434A, and T307A / E380A / N434A mutations. Exemplary mutations that can be made, either alone or in combination, to decrease the half-life of an antibody are H435A, P257I / N434H, D376V / N434H, M252Y / S254T / T256E / H433K / N434F, T308P / N434A, and H435R.

[0097] The Fc region can be mutated to reduce antibody binding to activating Fcγ receptors (FcγRs) and reduce Fc effector functions, such as, for example, C1q binding, complement-dependent cytotoxicity (CDC), antibody-dependent cellular cytotoxicity (ADCC), or antibody-dependent cellular phagocytosis (ADCP).

[0098] Positions in Fc that can be mutated to reduce antibody binding to activating FcγRs and subsequently reduce effector function include positions 214, 233, 234, 235, 236, 237, 238, 265, 267, 268, 270, 295, 297, 309, 327, 328, 329, 330, 331 and 365. Exemplary mutations that can be made alone or in combination are the K214T, E233P, L234V, L234A, G236 deletion, V234A, F234A, L235A, G237A, P238A, P238S, D265A, S267E, H268A, H268Q, Q268A, N297A, A327Q, P329A, D270A, Q295A, V309L, A327S, L328F, A330S and P331S mutations in IgG1, IgG2, IgG3 or IgG4. Exemplary combinations of mutations that result in antibodies with reduced ADCC are L234A / L235A in IgG1, V234A / G237A / P238S / H268A / V309L / A330S / P331S in IgG2, F234A / L235A in IgG4, S228P / F234A / L235A in IgG4, N297A in all Ig isotypes, V234A / G237A in IgG2, K214T / E233P / L234V / L235A / G236 deletion / A327G / P331 in IgG1, and K214T / E233P / L234V / L235A / G236 deletion / A327G / P331 in IgG1. These mutations are A / D365E / L358M in IgG2, H268Q / V309L / A330S / P331S in IgG2, S267E / L328F in IgG1, L234F / L235E / D265A in IgG1, L234A / L235A / G237A / P238S / H268A / A330S / P331S in IgG1, S228P / F234A / L235A / G237A / P238S in IgG4, and S228P / F234A / L235A / G236 deletion / G237A / P238S in IgG4. Hybrid IgG2 / 4 Fc domains can also be used, such as an Fc that contains residues 117-260 from IgG2 and residues 261-447 from IgG4.

[0099] An exemplary mutation that results in an antibody with reduced CDC is the K322A mutation.

[0100] To increase the stability of IgG4, the well-known S228P mutation may be introduced into IgG4 antibodies.

[0101] The Fc region can include mutations that enhance binding of the antibody to Fcγ receptors (FcγRs) and enhance Fc effector functions, such as, for example, C1q binding, complement-dependent cytotoxicity (CDC), antibody-dependent cellular cytotoxicity (ADCC), and / or antibody-dependent phagocytosis (ADCP).

[0102] Positions in Fc that can be mutated to improve binding of the antibody to activating FcγRs and increase the effector function of the antibody include positions 236, 239, 243, 256, 290, 292, 298, 300, 305, 312, 326, 330, 332, 333, 334, 345, 360, 339, 378, 396, or 430 (residue numbering according to the EU index). Exemplary mutations that can be made alone or in combination are G236A, S239D, F243L, T256A, K290A, R292P, S298A, Y300L, V305L, K326A, A330K, I332E, E333A, K334A, A339T, and P396L. Exemplary combinations resulting in antibodies with increased ADCC or ADCP are S239D / I332E, S298A / E333A / K334A, F243L / R292P / Y300L, F243L / R292P / Y300L / P396L, F243L / R292P / Y300L / V305I / P396L and G236A / S239D / I332E in IgG1.

[0103] Fc positions that can be mutated to increase CDC of an antibody include positions 267, 268, 324, 326, 333, 345 and 430. Exemplary mutations that can be made alone or in combination are S267E, F1268F, S324T, K326A, K326W, E333A, E345K, E345Q, E345R, E345Y, E430S, E430F and E430T. Exemplary combination mutations that result in antibodies with increased CDC are K326A / E333A, K326W / E333A, H268F / S324T, S267E / H268F, S267E / S324T and S267E / H268F / S324T in IgG1.

[0104] "Antibody-dependent cellular cytotoxicity", "antibody-dependent cell-mediated cytotoxicity" or "ADCC" is a mechanism of cell death induction that depends on the interaction of antibody-coated target cells with lytic activity effector cells, such as natural killer cells (NK), monocytes, macrophages, and neutrophils, via Fcγ receptors (FcγR) expressed on the effector cells. For example, NK cells express FcγRIIIa, whereas monocytes express FcγRI, FcγRII, and FcγRIIIa. The ADCC activity of an antibody can be assessed by an in vitro assay using cells expressing the protein to which the antibody binds as target cells and NK cells as effector cells. Cytolysis can be detected by the release of label (e.g., radioactive substrates, fluorescent dyes, or natural intracellular proteins) from lysed cells. In an exemplary assay, target cells are used at a ratio of 1 target cell to 4 effector cells. Target cells are pre-labeled with BATDA and combined with effector cells and test antibodies. Samples are incubated for 2 hours and cytolysis is measured by measuring BATDA released into the supernatant. Data are normalized to maximal cytotoxicity using 0.67% Triton X-100 (Sigma Aldrich) and minimal controls measured by spontaneous release of BATDA from target cells in the absence of antibody.

[0105] "Antibody-dependent cellular phagocytosis" ("ADCP") refers to a mechanism for the elimination of antibody-coated target cells by internalization by phagocytes, such as macrophages or dendritic cells. ADCP can be assessed by using monocyte-derived macrophages as effector cells and cells engineered to express a protein that is bound by the antibody and also express GFP or other marker molecules as target cells. In an exemplary assay, the effector:target cell ratio can be, for example, 4:1. Effector cells can be incubated with target cells for 4 hours in the presence or absence of an antibody of the invention. After incubation, cells can be detached using accutase. Macrophages are identified with anti-CD11b and anti-CD14 antibodies conjugated to fluorescent labels and CD11 isoforms are eluted using standard methods. + CD14 + % phagocytosis can be measured based on % GFP fluorescence of macrophages.

[0106] "Complement-dependent cytotoxicity" or "CDC" refers to a mechanism of inducing cell death in which the Fc effector domain of a target-bound antibody binds and activates complement component C1q, which in turn activates the complement cascade leading to the death of the target cell. Complement activation also leads to the deposition of complement components on the surface of target cells, which promote CDC by binding to complement receptors (e.g., CR3) on leukocytes. To measure CDC on cells, for example, 1 × 10 Daudi cells are cultured in RPMI-B (RPMI supplemented with 1% BSA). 5 Cells / well (50 μL / well) are plated, 50 μL of test antibody is added to the well at a final concentration of 0-100 μg / mL, the reaction is incubated for 15 minutes at room temperature, 11 μL of pooled human serum is added to the well, and the reaction is incubated for 45 minutes at 37° C. Percent (%) lysed cells can be detected as % propidium iodide stained cells by FACS assay using standard methods.

[0107] Binding of antibodies to FcγR or FcRn can be assessed using flow cytometry in cells engineered to express each receptor. In an exemplary binding assay, 2 × 10 antibodies are plated per well in a 96-well plate. 5 Cells are seeded and blocked with BSA Stain Buffer (BD Biosciences, San Jose, USA) for 30 min at 4°C. Cells are incubated with test antibodies on ice for 1.5 h at 4°C. After washing twice with BSA Stain Buffer, cells are incubated with R-PE-labeled anti-human IgG secondary antibody (Jackson Immunoresearch Laboratories) for 45 min at 4°C. Cells are washed twice with Stain Buffer and then resuspended in 150 μL of Stain Buffer containing DRAQ7 Live / Dead Stain Reagent (Cell Signaling Technology, Danvers, USA) diluted 1:200. PE and DRAQ7 signals of stained cells are detected using B2 and B4 channels, respectively, on a Miltenyi MACSQuant flow cytometer (Miltenyi Biotec, Auburn, USA). Live cells are gated by DRAQ7 exclusion and the geometric mean fluorescent signal of at least 10,000 viable events collected is measured. Analysis is performed using FlowJo software (Tree Star). Data are plotted as mean fluorescence signal versus the log of antibody concentration. Nonlinear regression analysis is performed.

[0108] "Enhancing" or "enhancing" refers to an enhancement of effector function (e.g., ADCC, CDC and / or ADCP) or enhanced binding to Fcγ receptors (FcγR) or FcRn of an antibody of the invention having at least one mutation in the Fc region compared to a parent antibody without the mutation. An "enhancement" can be about a 10%, 20%, 30%, 40%, 50%, 60%, 70%, 80%, 90%, 100% or more enhancement, or a statistically significant enhancement.

[0109] By "reducing" or "reduction" is meant a reduction in effector function (e.g., ADCC, CDC and / or ADCP) or reduced binding to Fcγ receptors (FcγR) or FcRn of an antibody of the invention with at least one mutation in the Fc region compared to a parent antibody without the mutation. A "reduction" can be about a 10%, 20%, 30%, 40%, 50%, 60%, 70%, 80%, 90%, 100% or more reduction, or a statistically significant reduction.

[0110] By "modulate" is meant either enhancing or decreasing the effector function (e.g., ADCC, CDC and / or ADCP) or enhancing or decreasing binding to Fcγ receptors (FcγR) or FcRn of an antibody of the invention that has at least one mutation in the Fc region compared to a parent antibody that does not have the mutation.

[0111] The antibodies used in the methods of the invention can be introduced with mutations that modulate the binding of the antibodies to Protein A. To generate and purify full-length bispecific antibodies, it is necessary to efficiently separate the bispecific antibodies from excess parent and / or intermediate molecules. Bispecific antibodies that have asymmetric (e.g., only one heavy chain) Fc mutations that modulate Protein A binding can therefore be purified from the parent antibodies based on their different elution profiles from Protein A affinity columns. Exemplary mutations that can be introduced into the antibodies used in the methods of the invention to modulate Protein A binding are Q311R, Q311K, T307P_L309Q, T307P_V309Q, T307P_L309Q_Q311R, T307P_V309Q_Q311R, H435R, or H435R_Y436F.

[0112] Glycoengineering The ability of the antibodies used in the methods of the invention to induce ADCC can be enhanced by modifying their oligosaccharide components. Human IgG1 is N-glycosylated at Asn297, with the majority of glycans in the well-known biantennary G0, G0F, G1, G1F, G2, or G2F forms. Antibodies produced by non-recombinant CHO cells usually have a glycan fucose content of at least about 85%. Removal of the core fucose from biantennary complex oligosaccharides attached to the Fc region enhances the ADCC of the antibody through improved FcγRIIIa binding, without altering antigen binding or CDC activity. Such mAbs can be achieved using various methods reported to lead to successful expression of relatively highly defucosylated antibodies with biantennary complex-type Fc oligosaccharides, such as controlling the culture osmolarity, using the mutant CHO line Lec13 as a host cell line, using the mutant CHO line EB66 as a host cell line, using the rat hybridoma cell line YB2 / 0 as a host cell line, introducing small interfering RNA specific for the α1,6-fucosyltransferase (FUT8) gene, or co-expressing β-1,4-N-acetylglucosaminyltransferase III with Golgi α-mannosidase II or the potent α-mannosidase I inhibitor kifunensine.

[0113] The heterodimeric antibodies used in the methods of the present invention can have a biantennary glycan structure with a fucose content of about 0% to about 15%, e.g., 15%, 14%, 13%, 12%, 11%, 10%, 9%, 8%, 7%, 6%, 5%, 4%, 3%, 2%, 1% or 0%.

[0114] The heterodimeric antibodies used in the methods of the present invention can have biantennary glycan structures with a fucose content of about 50%, 40%, 45%, 40%, 35%, 30%, 25%, 20%, 15%, 14%, 13%, 12%, 11%, 10%, 9%, 8%, 7%, 6%, 5%, 4%, 3%, 2%, 1% or 0%.

[0115] "Fucose content" means the amount of fucose monosaccharide in the glycan at Asn297. The relative amount of fucose is the percentage of fucose-containing structures compared to all glycan structures. These can be characterized and quantified in multiple ways, for example, 1) MALDI-TOF of samples (e.g., complex, hybrid, oligo and high mannose structures) treated with N-glycosidase F; 2) enzymatic release of the Asn297 glycan followed by derivatization and detection / quantification by HPLC with fluorescence detection (UPLC) and / or HPLC-MS (UPLC-MS); 3) intact protein analysis of native or reduced mAbs in the presence or absence of treatment of the Asn297 glycan with Endo S or other enzymes (cleaving between the first and second GlcNAc monosaccharides leaving a fucose attached to the first GlcNAc); 4) enzymatic digestion (e.g., trypsin or endopeptidase Lys-C) of the mAb into its constituent peptides followed by separation, detection and quantification by HPLC-MS (UPLC-MS); or 5) PNGase at Asn297. Separation of mAb oligosaccharides from mAb proteins by specific enzymatic deglycosylation with F is used. The released oligosaccharides can be labeled with fluorophores and separated and identified by various complementary techniques that allow detailed characterization of glycan structures by matrix-assisted laser desorption / ionization (MALDI) mass spectrometry by comparison of experimental and theoretical masses, measurement of the degree of sialylation by ion-exchange HPLC (GlycoSep C), separation and quantification of oligosaccharide forms according to hydrophilicity criteria by normal-phase HPLC (GlycoSep N), and separation and quantification of oligosaccharides by high-performance capillary electrophoresis-laser-induced fluorescence (HPCE-LIF).

[0116] "Low fucose" or "low fucose content" refers to an antibody having a fucose content of about 0% to 15%.

[0117] "Normal fucose" or "normal fucose content" refers to an antibody having a fucose content of about 50% or more, typically about 60%, 70%, 80% or more, or 85% or more.

[0118] The heterodimeric antibody used in the method of the present invention may be post-translationally modified by processes such as glycosylation, isomerization, deglycosylation, or non-natural covalent modifications, such as the addition of polyethylene glycol moieties (PEGylation) and lipidation. Such modifications may occur in vivo or in vitro. For example, the heterodimeric antibody used in the method of the present invention may be conjugated to polyethylene glycol (PEGylation) to improve its pharmacokinetic profile. Conjugation may be performed by techniques known to those skilled in the art. Conjugation of therapeutic antibodies with PEG has been shown to enhance pharmacodynamics without interfering with function.

[0119] C-Terminal Lysine The heterodimeric antibody used in the method of the present invention may have its C-terminal lysine (CTL) partially removed during production. During production, extracellular Zn is removed as described in US Patent Application Publication No. US20140273092. 2+ , EDTA or EDTA-Fe 3+ By controlling the concentration of, it is possible to control the removal of CTLs to less than the maximum level. The CTL content in an antibody can be measured using known methods.

[0120] The heterodimeric antibody used in the methods of the present invention may have a C-terminal lysine content of about 10% to about 90%, about 20% to about 80%, about 40% to about 70%, about 55% to about 70%, or about 60%.

[0121] The heterodimeric antibodies used in the methods of the present invention may have a C-terminal lysine content of about 0%, 5%, 10%, 20%, 30%, 40%, 50%, 60%, 70%, 80%, 90%, or 100%.

[0122] Antibody allotypes The heterodimeric antibody used in the method of the present invention can be of any allotype. Allotype is not expected to affect Fab arm exchange. The immunogenicity of therapeutic antibodies is associated with an increased risk of infusion reactions and a reduced durability of therapeutic effect (Baert et al., (2003) N Engl J Med 348:602-08). The extent to which a therapeutic antibody induces an immune response in the host may be determined in part by the antibody allotype (Stickler et al., (2011) Genes and Immunity 12:213-21). Antibody allotypes are associated with changes in the amino acid sequence at specific positions of the antibody constant region sequence. Table 2 shows selected IgG1, IgG2 and IgG4 allotypes.

[0123] [Table 2]

[0124] target antigen The method of the present invention can be used to generate heterodimeric antibodies of any specificity using Fab arm exchange, since the reduction and reformation of disulfide bonds during production is not expected to be affected by the VH / VL regions of the antibody, as also demonstrated herein. A heterodimeric antibody may bind to one target antigen with two different epitopes or to two or more target antigens, depending on the specificity of each VH / VL pair. When a heterodimeric antibody binds to two target antigens, the target antigens may be on the same cell or on two different cells. The target antigens may be tumor-associated antigens, antigens involved in inflammation, antigens involved in regulating the activity of T cells or B cells, or generally any target for which it is desired to modulate the biological activity or reduce the number of cells expressing the target.

[0125] Exemplary antigens to which the heterodimeric antibodies used in the methods of the present invention may bind include: It can be one or more of the following: TIFF0007645799000004.tif194142TIFF0007645799000005.tif212137TIFF0007645799000006.tif138133.

[0126] In some embodiments, the heterodimeric antibody used in the methods of the present invention binds to CD3.

[0127] In some embodiments, the heterodimeric antibody used in the methods of the present invention binds to CD3 and a tumor-associated antigen (TAA).

[0128] In some embodiments, the heterodimeric antibody used in the methods of the invention binds to CD123 and CD3 and to BCMA and CD3.

[0129] In some embodiments, the heterodimeric antibody used in the methods of the present invention binds to EGFR and c-Met.

[0130] In some embodiments, a heterodimeric antibody that binds EGFR and c-Met comprises a first heavy chain (HC1) of SEQ ID NO:4, a first light chain (LC1) of SEQ ID NO:5, a second heavy chain (HC2) of SEQ ID NO:6, and a second light chain (LC2) of SEQ ID NO:7.

[0131] The present invention also provides bispecific antibodies produced by the methods of the present invention.

[0132] The present invention also provides a bispecific antibody that binds to EGFR and c-Met produced by the method of the present invention.

[0133] The invention also provides a bispecific antibody that binds EGFR and c-Met, comprising a first heavy chain (HC1) of SEQ ID NO:4, a first light chain (LC1) of SEQ ID NO:5, a second heavy chain (HC2) of SEQ ID NO:6, and a second light chain (LC2) of SEQ ID NO:7, produced by the methods of the invention.

[0134] Preparation of homodimeric and heterodimeric antibodies The first homodimeric antibody and the second homodimeric antibody can be produced together or separately in a culture vessel such as a bioreactor. The first homodimeric antibody and the second homodimeric antibody are produced by co-expression in a host cell, or by using separate host cells that produce each homodimeric antibody. In the latter case, the host cells can be of the same or different origin.

[0135] "Host cell" refers to a cell into which one or more vectors expressing at least one antibody heavy chain and one antibody light chain have been introduced. "Host cell" refers not only to the particular subject cell, but also to the progeny of such a cell and to stable cell lines made from the particular subject cell. Because certain modifications may occur in successive generations due to mutation or environmental influences, such progeny may not be identical to the parent cell, but are still included within the scope of the term "host cell" as used herein.

[0136] Such host cells can be eukaryotic, prokaryotic, plant or archaeal cells. Escherichia coli, bacilli, such as Bacillus subtilis, and other Enterobacteriaceae, such as Salmonella, Serratia, and various Pseudomonas species, are examples of prokaryotic host cells. Other microorganisms, such as yeast, are also useful for expression. Saccharomyces (e.g., S. cerevisiae) and Pichia are examples of suitable yeast host cells. Exemplary eukaryotic cells can be of mammalian, insect, avian or other animal origin. Mammalian eukaryotic cells include immortalized cells such as hybridoma or myeloma cell lines, e.g., SP2 / 0 (American Type Culture Collection (ATCC), Manassas, VA, CRL-1581), NS0 (European Collection of Cell Cultures (ECACC), Salisbury, Wiltshire, UK, ECACC No. 85110503), FO (ATCC CRL-1646) and Ag653 (ATCC CRL-1580) mouse cell lines. An exemplary human myeloma cell line is U266 (ATTC CRL-TIB-196). Other useful cell lines include those derived from Chinese hamster ovary (CHO) cells, e.g., CHOK1SV (Lonza Biologics, Walkersville, MD), Potelligent® CHOK2SV (Lonza), CHO-K1 (ATCC CRL-61) or DG44.

[0137] Exemplary vectors that can be used to express one or more antibody heavy and light chains include pWLneo, pSV2cat, pOG44, PXR1, pSG (Stratagene), pSVK3, pBPV, pMSG and pSVL (Pharmacia), pEE6.4 (Lonza) and pEE12.4 (Lonza), as well as vectors used in Lonza's Xceed system.

[0138] The first and second homodimeric antibodies may be produced by expressing them in separate host cells. In this case, the first and second homodimeric antibodies may be purified before mixing them in the presence of a reducing agent to initiate Fab arm exchange. Purification may be achieved using known methods such as protein A chromatography. Alternatively, culture media containing the first and second homodimeric antibodies may be mixed and a reducing agent added thereto to initiate Fab arm exchange. When the first and second homodimeric antibodies are produced by co-expression, these antibodies may also be purified using protein A chromatography.

[0139] Fab arm exchange to generate heterodimeric antibodies is described herein. Process conditions such as the concentration of the first homodimeric antibody and the second homodimeric antibody, the reducing agent, the reducing agent concentration, the reduction time, the dissolved oxygen concentration, the temperature at which the reaction is performed, and the buffer used are optimized to generate heterodimeric antibodies in high yield and purity as described herein.

[0140] After Fab arm exchange, the resulting heterodimeric antibody can be further purified. Purification methods include a combination of protein A, protein G chromatography, other methods of affinity chromatography, such as affinity chromatography based on antigen binding or binding to anti-idiotypic antibodies, thioaffinity, ion exchange, hydrophobic interaction, hydroxyapatite chromatography, and other mixed mode resins. Further methods can also be used, such as precipitation with salt or polyethylene glycol, to obtain purified heterodimeric antibody.

[0141] Equipment suitable for the process of the method of the invention is known. Expression of the homodimeric antibody by the host cells can take place in a reaction vessel, typically a bioreactor, for example. The reduction and oxidation steps can take place in the same bioreactor as the expression of the first homodimeric antibody and / or the second homodimeric antibody, or in separate reaction vessels. The reaction vessels and supporting process piping can be disposable or reusable and made of standard materials (plastic, glass, stainless steel, etc.). The reaction vessels can be equipped with mixing, sparging, headspace gassing, temperature control, and / or monitored with probes to measure temperature, weight / volume, pH, dissolved oxygen (DO), and redox potential. All such techniques are common within standard unit operations of manufacturing plants and are well known to those skilled in the art.

[0142] Having described the invention in general terms, aspects of the invention will be further disclosed in the following examples; these examples should not be construed as limiting the scope of the claims. EXAMPLES

[0143] Example 1. General Method cSDS Capillary sodium dodecyl sulfate-polyacrylamide gel electrophoresis separates proteins based on molecular weight. A commercially available capillary electrophoresis system with a fused silica capillary in a temperature-controlled cartridge was used for the analysis. The test article was mixed with an internal standard and an alkylating agent and heated for a defined time and temperature to denature the proteins. The sample was then injected electrokinetically and analyzed by applying a larger electric field for a predefined period of time. Detection was performed by absorbance at 220 nm, and purity was determined by calculating the corrected peak area of ​​the main peak relative to all other peaks in the electropherogram after the internal standard.

[0144] IEX-HPLC The percentage of bispecific antibodies remaining relative to the parent homodimeric antibodies was determined by ion-exchange high performance liquid chromatography (IEX-HPLC). Ion-exchange chromatography utilizes differences in the surface charge of molecules to achieve separation. IEX-HPLC was performed using a Thermo Scientific Propac WCX-10 (4 mm ID x 150 mm, packed with 10 μm particles). Separation of species was achieved using salt and / or pH gradients. As the gradient concentration increases, the overall surface interaction of each antibody with the column is neutralized and elution is detected at 280 nm. The relative amounts of each IgG were determined by comparing the peak area counts.

[0145] HIC-HPLC Hydrophobic interaction chromatography separates proteins based on their hydrophobicity. Proteins are retained on the column under high salt conditions and are eluted by decreasing the salt concentration. High salt buffer conditions reduce protein solvation, exposing hydrophobic regions and facilitating binding to the HIC stationary phase. Samples were injected onto a Tosoh TSKgel Butyl-NPR (4.6 mm x 10 cm, 2.5 μm nonporous resin-based beads) column and eluted with an ammonium sulfate gradient. Absorbance at 280 nm was monitored and compared to a reference standard injection to identify peaks.

[0146] RP-HPLC (detection of cystamine) Reversed-phase high-performance liquid chromatography separates solute molecules in the mobile phase based on hydrophobic binding interactions with the stationary phase medium. Taking advantage of the difference in hydrophobicity between 2-MEA and oxidized dimeric cystamine, they were analyzed by RP-HPLC and quantified by comparing the measured values ​​to a standard curve. A Phenomenex Luna 18 (4.6 mm × 150 mm, 3 μm) column with a hexanesulfonate / acetonitrile mobile phase system was used to separate the two species with a step gradient of increasing organic solvent. These species were detected by monitoring the absorbance at 220 nm (2-MEA) and 255 nm (cystamine), and the peak areas were compared to their relative standard curves.

[0147] General antibody production schemes and scales for mono- and bispecifics (done at 200ml scale) Each monospecific antibody was thawed from an individual mammalian cell bank, grown, and expressed separately in a bioreactor. The production bioreactor was supplemented with enriched medium and harvested before significant cell death. The bioreactor was harvested by centrifugation and / or filtration, and the monospecific antibodies were captured separately by affinity Protein A chromatography. The captured monospecific antibodies were then obtained and stored frozen until combined at the FAE stage.

[0148] Example 2. Process improvements during the production of bispecific antibodies using Fab arm exchange Bispecific antibodies were generated using Fab arm exchange by introducing asymmetric mutations into the CH3 domains of two homodimeric parent antibodies, followed by reduction and reformation of the intermolecular disulfide bonds. The asymmetric CH3 domain mutations result in the preferential formation of heterodimeric bispecific antibodies over the homodimeric parent antibodies.

[0149] To ensure the robustness of the process during manufacturing, experiments were designed to understand the possible thresholds and ranges of various parameters during manufacturing. It was shown that the reformation of disulfide bonds in antibodies is dependent on the presence of oxygen and free metals, since EDTA and oxygen-free conditions inhibited the reoxidation of antibody cysteines to disulfide bonds (US2014 / 0303356). Therefore, we initiated a study to examine the need to control the amount of dissolved oxygen (DO2) and metals during the manufacturing of bispecific antibodies with Fab-arm exchange.

[0150] FIG. 2 shows the process steps of reduction and reformation of disulfide bonds during production of bispecific antibodies by Fab arm exchange.

[0151] Theoretically, the reformation of antibody disulfide bonds after reduction with a reducing agent such as 2-mercaptoethylamine (2-MEA) can occur by various pathways. For example, thiol-disulfide exchange can consist of the reduction and reformation of disulfide bonds. For thiol-disulfide exchange to proceed, a reducing agent such as 2-MEA must be able to convert its thiolate (RS - ) form. Figure 3 shows the molecular details of the thiol-disulfide exchange reaction. To reduce the disulfide bonds of the parent antibody, a thiolate anion (circled in Figure 3) attacks the sulfur atom of the disulfide bond, displacing one of the sulfur atoms and forming a new disulfide bond with the 2-MEA thiolate.

[0152] Using the Henderson-Hasselbalch equation shown below, 2-MEA thiol (SH) and thiolate anion (S - ) is a function of pH and the pK of the 2-MEA thiol. a It depends heavily on. TIFF0007645799000007.tif10128

[0153] Thus, reduction of disulfide bonds and global Fab arm exchange become inhibited at low pH, where the protonated thiol form of 2-MEA is favored over its deprotonated thiolate anion form. At higher pH, the equilibrium shifts toward the thiolate, allowing reduction of disulfide bonds and thiol-disulfide exchange.

[0154] An overview of the reactions that can occur during Fab arm exchange in the presence of 2-MEA is shown in Figure 4. During Fab arm exchange, removal of reducing agents such as 2-MEA by buffer exchange (in UF / DF) allows the reformation of disulfide bonds of heterodimeric bispecific antibodies, but the disulfide reformation process occurs through multiple reaction pathways. In the presence of oxygen, oxygen can react with protonated thiolates to promote the reformation of disulfide bonds of heterodimeric bispecific antibodies, as shown in reaction 2 of Figure 4. The presence of oxygen can also deplete residual 2-MEA thiols in the solution to form cystamine dimers and water, as shown in reaction 4 of Figure 4. Therefore, to improve the efficiency and yield of bispecific antibodies, it is necessary to use a sufficiently high concentration of 2-MEA during Fab arm exchange.

[0155] However, thiolate disulfide exchange and disulfide bond reformation may still be possible in the absence of oxygen. Under oxygen-poor conditions, as the concentration of 2-MEA decreases during the buffer exchange process, the antibody thiol groups and the 2-aminoethyl disulfide Fab intermediate will proceed in the reverse direction (reaction 1 in Figure 4) to simultaneously generate 2-MEA thiolate and reform the disulfide bonds of the bispecific heterodimeric antibody. As more 2-MEA is removed during the buffer exchange process, this reaction will continue to proceed in the reverse direction.

[0156] Example 3. Production of bispecific antibody A under low DO2 conditions in UF / DF Experimental Methods In this experiment, ambient DO2 levels were present during reduction and up to 23 hours after 2-MEA addition, after which a nitrogen overlay was performed to minimize %DO2 during UF / DF. This approach reflected the presence of oxygen during the reduction step and the absence of oxygen during the disulfide bond reformation step.

[0157] Bispecific antibody A binds BCMA and CD3, is of IgG4 isotype, and has S228P, F234A, L235A substitutions in both heavy chains ("PAA substitutions"), with F at position 405 and R at position 409 in one heavy chain and L at position 405 and K at position 409 in the other heavy chain to promote heterodimer formation. These parent antibodies were named p1A-IgG4PAAF405R409 and p2A-IgG4PAAL405K409.

[0158] The parental antibodies p1A-IgG4PAAF405R409 and p2A-IgG4PAAL405K409 were harvested from the cell culture bioreactor and purified by Protein A affinity chromatography.

[0159] Solutions were prepared using p1A-IgG4PAAF405R409 and p2A-IgG4PAAL405K409 at the indicated molar ratio or 1:1.06, with deliberate limiting of one of the parent antibodies. The mixture was then adjusted to pH 7.3 with 101 mM sodium acetate, 105 mM Tris base, pH 7.3, and diluted to a total IgG concentration of 10.5 g / L. A stock solution of 50 mM sodium acetate, 800 mM 2-MEA pH 5.0 was added to the parent mAb mixture. The final reduction buffer composition before UF / DF was approximately 100 mM sodium acetate, approximately 35 mM 2-MEA, approximately 30 mM NaCl, pH 7.3. The reduced parent mAb solution was incubated at 24°C for 23 hours and then transferred to the UF / DF retentate vessel. Before ultrafiltration (UF) and diafiltration (DF) were started, a nitrogen overlay was provided to the retentate vessel via the vessel headspace to deplete DO2 during the reformation of disulfide bridges and was maintained throughout the UF and DF processes. The diafiltration buffer (100 mM Tris-acetate, 30 mM NaCl pH 7.5) was sparged with nitrogen during processing to prevent DO2 addition via the diafiltration buffer. An in-line DO2 sensor was used to ensure that the retentate DO2 had reached a level of less than 5% DO2 and that the diafiltration buffer had reached a level of less than 1% DO2 before the start of the UF and DF steps.

[0160] Throughout the UF and DF processes, DO2 levels were measured in the feed inlet line, retentate line, and permeate line during UF / DF and were all maintained at <4% DO2. Diafiltration buffer was maintained at a level of <1% DO2. The UF / DF system was maintained at a cross-flow rate of 54 mL / min with a target transmembrane pressure of 14.5 psi for 11 diafiltration volumes (DV). All processes after incubation were performed at room temperature (18-22 °C). After completion of UF / DF, the collected retentate was quenched with 1.0 M acetic acid, adjusted to pH 5.0 to minimize further disulfide bond formation, and stored at 2-8 °C. Additionally, no further buffer recovery flushes were performed after completion of UF / DF to prevent possible oxidation of the material due to the introduction of additional diafiltration buffer.

[0161] The UF / DF setup allowed for DO2 to be measured at three locations during processing. In-line monitoring of %DO2 in the feed inject line, retentate line, and permeate line is shown in Figure 5. When DO2 was lowered, the levels remained below 5% in the permeate and below 3% in the injector / retentate. Table 3 provides an overview of the UF / DF process parameters and performance.

[0162] [Table 3]

[0163] Table 4 summarizes the offline %DO2, pH, and conductivity measurements before and after the addition of 2-MEA. Figure 6 shows the percent (%)DO2 in the UF / DF measured in the retentate, permeate, and injection lines. After nitrogen overlay on the retentate during recirculation, the %DO2 stabilized over time to 3.2% in the permeate, 1.5% in the retentate, and 1.4% in the feed.

[0164] Non-reducing cSDS analysis was used to determine the integrity of the disulfide bonds of the formed heterodimeric bispecific antibodies (measured as % purity of heterodimeric antibody in non-reducing cSDS). The purity of Antibody A was determined to be 97.59%.

[0165] This study demonstrated that bispecific antibody A was formed with high levels of disulfide bond formation (reflected by % purity in non-reduced cSDS) under ambient %DO2 (ranging from 89 to 13%) and low %DO2 environments (<4%DO2 in UF / DF) during reduction. Previously, it was thought that DO2 was required to stabilize bispecific antibodies by oxidizing thiols to disulfide bonds.

[0166] [Table 4]

[0167] It was calculated that the level of oxygen present in this study was not sufficient to contribute to the oxidation of thiols in UF / DF. The following calculations showed that a DO2 concentration of greater than 30% in solution was required for thiol oxidation to occur using the reaction represented by reaction 2 in Figure 4.

[0168] In this study, the following assumptions were made to calculate the theoretical %DO2 saturated solution that would oxidize all thiols in the amount of antibody present: · IgG solution: 10g / L mol SS (disulfide bonds reduced and oxidized) / mol mAb 1 / 2 mole O2 / mol SS (when oxygen is needed to form a disulfide bond from a thiol, see reaction 2 in Figure 4) Solubility of oxygen in air-saturated water = 7 ppm (g O2 / 106 g solution) = 100% saturation mAb MW: 150kDa

[0169] If 4 disulfides / mol mAb are to be oxidized and O2 is required for oxidation of all thiols, then to calculate the % dissolved DO2 concentration to oxidize a 10 mg / mL antibody solution: 1. (10mg mAb / mL solution) * (moles mAb / 150000g mAb) * (2moles SS / moles mAb) * (1 / 2moles O2 / moles SS) * (1000mL / 1L) (1g / 1000mg) = 6.67 x 10 -4 Molar O2 / L solution 2. (6.67×10 -5 moles O2 / L solution) * (31.998g O2 / 1 mole O2) * (1L / 1000g) = 2.13 x 10 -6 g O2 / g solution 3. (2.13×10 -6 g O2 / g solution) * (10 6 g solution)=2.13g O2 / 10 6 g solution=2.13ppm 4. Air-saturated water: approx. 7g O2 / 10 6 g solution=7ppm 5. 2.13ppm / 7ppm=30% O2 saturated solution is required

[0170] At operating temperature, 2% DO2 was equivalent to 0.192 mg / mL or 6.00 μM, 3% DO2 was equivalent to 0.288 mg / L or 8.99 μM, 4% DO2 was equivalent to 0.384 mg / L or 11.29 μM, 5% DO2 was equivalent to 0.480 mg / mL or 14.99 μM, and 30% DO2 was equivalent to 2.88 mg / mL or 90 μM.

[0171] This indicates that there was an oxygen-independent chemical reaction pathway that allowed the reformation of disulfide bonds in the absence of oxygen during the UF / DF process. When 2-MEA was removed from solution, the reverse reaction of reaction 1 in Figure 4 could drive the reformation of disulfide bridges in the heterodimeric bispecific antibody.

[0172] Conclusion: Maintaining DO2 levels at <4% during UF / DF did not significantly affect disulfide bond formation. Upon completion of UF / DF, the purity of the bispecific antibody was >97%.

[0173] Example 4. Production of Bispecific Antibody B at Low DO2 in UF / DF In this experiment, ambient DO2 levels were present during reduction and up to 23 hours after 2-MEA addition, after which a nitrogen overlay was performed to minimize %DO2 during UF / DF. This approach reflected the presence of oxygen during the reduction step and the absence of oxygen during disulfide bond reformation.

[0174] Bispecific antibody B binds to CD123 and CD3, is of IgG4 isotype, and contains S228P, F234A, L235A substitutions ("PAA substitutions") in both heavy chains, with F at position 405 and R at position 409 in one heavy chain and L at position 405 and K at position 409 in the other heavy chain to promote heterodimer formation. These parent antibodies were named p1B-IgG4PAAF405R409 and p2B-IgG4PAAL405K409.

[0175] The parental antibodies p1B-IgG4PAAF405R409 and p2B-IgG4PAAL405K409 were harvested from the cell culture bioreactor and purified by Protein A affinity chromatography.

[0176] Solutions of p1B-IgG4PAAF405R409 and p2B-IgG4PAAL405K409 were prepared, adjusted to pH 7.3 with 101 mM sodium acetate, 105 mM Tris base, pH 7.3, and diluted to a total IgG concentration of 10.5 g / L. A stock solution of 50 mM sodium acetate, 800 mM 2-MEA pH 5.0 was added to the parent mixture. The final reduction buffer composition before UF / DF was about 100 mM sodium acetate, about 35 mM 2-MEA, about 30 mM NaCl, pH 7.3. The reduced parent solution was incubated at 24° C. for 23 hours and then transferred to the UF / DF retentate vessel. Before ultrafiltration (UF) and diafiltration (DF) were started, a nitrogen overlay was supplied to the retentate vessel via the vessel headspace to deplete DO2 during the reformation of disulfide bridges, and was maintained throughout the UF and DF processes. The diafiltration buffer (100 mM Tris-acetate, 30 mM NaCl pH 7.5) was sparged with nitrogen during processing to prevent the addition of DO2 via the diafiltration buffer. An in-line DO2 sensor was used to ensure that the retentate had reached a level of less than 5% DO2 and the diafiltration buffer had reached a level of less than 1% before the start of the UF and DF steps.

[0177] Throughout the UF and DF processes, DO2 levels were measured in the feed inlet line, retentate line, and permeate line during UF / DF and were all maintained at <4% DO2. Diafiltration buffer was maintained at a level of <1% DO2. The UF / DF system was maintained at a cross-flow rate of 42 mL / min with a target transmembrane pressure of 15.0 psi for 11 diafiltration volumes (DV). All processes after incubation were performed at room temperature (18-22 °C). After completion of UF / DF, a 3 mL sample was taken and immediately stored at -70 °C with unadjusted pH of 7.5 to prevent further oxidation. The remaining bulk of the collected retentate was quenched with 1.0 M acetic acid and adjusted to pH 5.0 to minimize further disulfide bond formation. The bulk material was then stored at -70 °C. To prevent possible oxidation of the material by the introduction of additional diafiltration buffer, no further buffer recovery flushes were performed after completion of UF / DF, although yields were lower than typically observed.

[0178] The UF / DF setup allowed for DO2 to be measured at three locations during processing. In-line monitoring of %DO2 in the feed inlet line, retentate line, and permeate line is shown in Figure 5. When DO2 was lowered, the levels remained below 5% in the permeate and below 3% in the inlet / retentate.

[0179] Table 5 summarizes the UF / DF process parameters. Figure 7 shows the %DO2 in the UF / DF measured in the retentate, permeate, and injection lines. Table 6 summarizes the offline %DO2, pH, and conductivity measurements.

[0180] [Table 5]

[0181] [Table 6]

[0182] The integrity of the disulfide bonds of the formed bispecific antibodies was measured using non-reducing cSDS analysis for both the frozen samples at pH 7.5 and the bulk material at pH 5.0. The purity of the bispecific antibody preparation at pH 7.5 was measured to be 97.16%, and the purity of the antibody preparation at pH 5.0 was measured to be 97.27%.

[0183] This study demonstrated that bispecific antibody B was formed with high levels of disulfide bond formation under ambient %DO2 (ranging from 83 to 24%) during reduction and low %DO2 environment (<4%DO2) during UF / DF.

[0184] Both this study and the study described in Example 2 demonstrated that oxygen is not required during the UF / DF step to reform disulfide bridges during Fab arm exchange.

[0185] Conclusion: Maintaining DO2 levels below 4% during the UF / DF step of producing bispecific antibody B using Fab arm exchange did not significantly affect disulfide bond formation. Upon completion of UF / DF, non-reduced cSDS demonstrated % purity >97%.

[0186] Example 5. Production of Bispecific Antibody A at Low DO in the Presence of EDTA in UF / DF During the production of bispecific antibodies, there is an opportunity for free metal ions to be introduced into the manufacturing process. During the preparation of buffers from raw materials and leaching from metal components, these free metal ions may become involved in the redox reaction of Fab arm exchange. To investigate the possible influence of trace metals on Fab arm exchange, the addition of EDTA during Fab arm exchange was investigated. The addition of EDTA sequestered free metal ions that may be present in the solution, which may catalyze the oxidation reaction of disulfide bond reformation.

[0187] A solution was prepared using p1A-IgG4PAAF405R409 and p2A-IgG4PAAL405K409 in a molar ratio of 1.05:1.00. This mixture was then adjusted to pH 7.3 with 101 mM sodium acetate, 105 mM Tris base, pH 7.3, and diluted to a total IgG concentration of 10.5 g / L. A stock solution of 50 mM sodium acetate, 800 mM 2-MEA pH 5.0 was added to the parent mAb mixture. The final reduction buffer composition before UF / DF was approximately 100 mM sodium acetate, approximately 35 mM 2-MEA, approximately 30 mM NaCl, pH 7.3. This reduced parent solution was incubated at 24°C for 23.5 hours. At the completion of the incubation, a stock solution of 500 mM EDTA, pH 8.0, was added to the reduced parent mixture to target 2 mM EDTA in order to chelate any free metal ions that may have been present in the solution prior to the start of UF / DF. EDTA was also added to the diafiltration buffer to target 100 mM Tris-acetate, 30 mM NaCl, 2 mM EDTA, pH 7.5. This was done to ensure that no additional free metal ions were introduced during the UF / DF buffer exchange. No EDTA was added during the reduction step to allow for oxidation of disulfides and cystamine formation that may occur naturally during this step.

[0188] The reduced parent solution containing EDTA was then transferred to the UF / DF retentate vessel and a nitrogen overlay was provided to the retentate vessel via its headspace to deplete DO2 from the retentate prior to the start of UF and maintained throughout the UF and DF processes. The diafiltration buffer was sparged with nitrogen during processing to prevent DO2 addition via the diafiltration buffer. An in-line DO2 sensor confirmed that the retentate DO2 level reached a level of less than 2% DO2 and the diafiltration buffer reached a DO2 level of less than 1% DO2 prior to the start of the UF and DF steps.

[0189] Throughout the UF and DF process, DO2 levels were measured in the feed inlet line, retentate line, and permeate line during UF / DF and were all maintained at <4% DO2. Diafiltration buffer was maintained at a level of <1% DO2. The UF / DF system was maintained at a cross-flow rate of 58 mL / min with a target transmembrane pressure of 14.5 psi for 11 diafiltration volumes (DV). All processes after incubation were performed at room temperature (18-22 °C). After completion of UF / DF, a 9 mL sample was taken and immediately stored at -70 °C, unadjusted to pH 7.5, to prevent further oxidation. The remaining bulk of the collected retentate was quenched with 1.0 M acetic acid and adjusted to pH 5.0 to minimize further disulfide bond formation. The bulk bispecific antibody sample at pH 5.0 was subsequently stored at -70 °C. After completion of UF / DF, no buffer recovery flush was performed to prevent possible oxidation of the material due to the introduction of additional diafiltration buffer. Table 7 summarizes the UF / DF process parameters and performance.

[0190] [Table 7]

[0191] Non-reducing cSDS analysis was used to measure the integrity of the disulfide bonds of the bispecific antibodies formed for both the frozen samples at pH 7.5 and the bulk material at pH 5.0. The purity of the bispecific antibody preparation at pH 7.5 was determined to be 97.44% and the purity of the bispecific antibody preparation at pH 5.0 was determined to be 97.39%.

[0192] This study demonstrated that bispecific antibody A was formed with high levels of disulfide bond formation under low %DO2 environment and minimal available free metal ions during UF / DF. This study demonstrated that free metal ions to catalyze the oxidation reaction may not be required during UF / DF for Fab arm exchange.

[0193] Conclusion: Experimental results show that the addition of EDTA to the reduced parent mixture prior to UF / DF in Fab arm exchange does not affect disulfide bond formation. Upon completion of UF / DF, non-reduced cSDS resulted in % purity >97%.

[0194] Example 6. Comparison of Fab arm exchange under ambient and low DO2 conditions during reduction and UF / DF This study was designed to evaluate the hypothesis that some DO2-mediated oxidation may occur during 2-MEA reduction.

[0195] In this study, Fab arm exchange was performed using parental antibodies p1B-IgG4PAAF405R409 and p2B-IgG4PAAL405K409. A mixture of p1B-IgG4PAAF405R409 and p2B-IgG4PAAL405K409 in solution was prepared, adjusted to pH 7.3 with 101 mM sodium acetate, 105 mM Tris base, pH 7.3, and diluted to a total IgG concentration of 10.5 g / L. The mixture containing the parental antibodies was separated into two separate containers. In one container, an overlay of nitrogen was supplied to the container via the headspace to deplete DO2 from the parent mixture until %DO2<5% was achieved. The other parent mixture container was left under ambient air conditions as a control for the reduction reaction. A stock solution of 50 mM sodium acetate, 800 mM 2-MEA pH 5.0 was added to each parent mixture. The final reduction buffer composition prior to UF / DF was approximately 100 mM sodium acetate, approximately 35 mM 2-MEA, approximately 30 mM NaCl, pH 7.3. Both reduced parent solutions were incubated at 24 °C for 24 hours and the %DO2 of both vessels was measured; the results are shown in Figure 8. The low %DO2 reduced vessel was maintained under a nitrogen overlay throughout the incubation and remained at <2%DO2 over the 24 hours. The vessel that was under ambient air saw an initial %DO2 drop to approximately 40%DO2 after the addition of 2-MEA, but then the %DO2 gradually increased linearly back to >90%DO2. A summary of the UF / DF process parameters and performance is shown in Table 8.

[0196] [Table 8]

[0197] During reduction, samples were taken from both vessels via a sampling port on the vessel. The port was flushed before each sample was taken and the sample was immediately quenched by adding 10% v / v cystamine RP-HPLC mobile phase, 20 mM hexanesulfonate, pH 2.0 buffer; each sample was ensured to have its pH lowered to <5 before analysis. All time points were analyzed for disulfide bond integrity by non-reduced cSDS and for residual 2-MEA / cystamine concentration by RP-HPLC.

[0198] Analysis of residual 2-MEA / cystamine required additional sample preparation, including removal of the antibody with a 5 kDa molecular weight retention filter. The % purity by non-reduced cSDS and residual 2-MEA / cystamine results are reported in Table 9.

[0199] Upon completion of the 24-hour reduction incubation, the low %DO2 container was transferred to the UF / DF system. The retentate container continued to be supplied with an overlay of nitrogen via its headspace to continue depleting DO2 from the retentate and to maintain the nitrogen overlay throughout the UF and DF process. The diafiltration buffer was also sparged with nitrogen during processing to prevent DO2 from being added via the buffer. An in-line DO2 sensor was used to ensure that the retentate had reached a level of less than 5% DO2 and the diafiltration buffer had reached a level of less than 1% DO2 before the start of the UF and DF steps.

[0200] Throughout the UF and DF processes, DO2 levels were measured in the feed inlet line, retentate line, and permeate line during UF / DF, all of which were maintained at <4% DO2. Diafiltration buffer was maintained at a level of <1% DO2. The UF / DF system was maintained at a cross-flow rate of 55 mL / min with a target transmembrane pressure of 14.5 psi for 12 diafiltration volumes (DV) using a buffer of 100 mM Tris-acetate, 30 mM NaCl, pH 7.5. All processes after incubation were performed at room temperature (18-22 °C). After completion of UF / DF, the collected retentate was quenched with 1.0 M acetic acid and adjusted to pH 5.0 to minimize further disulfide bond formation. The bulk material was subsequently stored at -70 °C. No further buffer recovery flushes were performed after completion of UF / DF to prevent possible oxidation of the material due to the introduction of additional diafiltration buffer.

[0201] Disulfide integrity by non-reduced cSDS and residual 2-MEA / cysteamine by RP-HPLC were measured at various time points throughout reduction and UF / DF in both the low DO2% vessel and the ambient DO2 control vessel.

[0202] A summary of the data is shown in Table 9. In this experiment, only the Fab arm exchange reaction under oxygen-depleted conditions was continued throughout UF / DF. Residual 2-MEA and cystamine were measured only during reduction in both Fab arm exchange conditions; however, samples were taken during UF / DF under oxygen-depleted conditions to measure possible disulfide bond reformation by non-reducing cSDS.

[0203] [Table 9] * The first t=0 sample was taken immediately after the 2-MEA spike of the parent mixture. Sample time points were taken to measure the rate of disulfide bond reduction and reformation during both reduction and UF / DF.

[0204] Conclusion: The data support the hypothesis that an oxygen-independent pathway mediates the formation of disulfide bonds in the absence of oxygen during reduction and UF / DF.

[0205] When Fab arm exchange was performed under oxygen-depleted conditions, minimal cystamine formation and disulfide bond reformation was observed during reduction. By the completion of 24 h incubation in the presence of 2-MEA, only 1.6 mM cystamine and 14.74% intact antibody were observed. Nevertheless, treatment of the antibody sample with UF / DF revealed increasing levels of disulfide bond reformation as 2-MEA was removed. At the end of the DV (DV12), the formed pH 5.0 bispecific antibody preparation was measured to be 90.59% pure.

[0206] At ambient DO2, reduction of disulfide bonds occurred rapidly, with approximately 86% of the parent antibody being reduced 30 min after the start of reduction. Reformation of disulfide bonds was observed for the remainder of the reduction process, even though residual 2-MEA continued to be present in the solution. At the completion of the reduction step, 71% of the antibody had intact disulfide bonds. The formation of cystamine dimers during the reduction incubation, as well as the increasing levels of bispecific antibody formation, indicated that oxygen was being consumed by reactions 4 and 2, respectively, shown in Figure 4. This was supported by RP-HPLC data, which showed a significant depletion of 2-MEA and its conversion to the dimeric compound cystamine (reaction 4 in Figure 4).

[0207] This study revealed that DO2 during the reduction step catalyzes the formation of HC-HC and LC-HC interchain disulfide bonds of IgG but is not essential for the formation of intact bispecific antibodies, as disulfide bonds reform even under oxygen-deficient conditions. Removal of 2-MEA during diafiltration, even in the absence of oxygen, led to the reformation of disulfide bonds by the reverse reaction of reaction 1 in Figure 4.

[0208] Example 7. Generation of bispecific EGFR / c-Met antibodies using Fab arm exchange under ambient DO2 utilizing high concentrations of parental antibodies This study was designed to evaluate the hypothesis that higher protein concentrations in the homodimer mixture and various mass ratios of reducing agent to protein would result in oxidized heterodimeric antibodies.

[0209] In this study, a bispecific EGFR / cMet antibody was used, which comprises a first heavy chain (HC1) of SEQ ID NO: 4, a first light chain (LC) of SEQ ID NO: 5, a second HC (HC2) of SEQ ID NO: 6, and a second LC (LC2) of SEQ ID NO: 7.

[0210] In this study, Fab arm exchange was performed using parental EGFR and c-Met antibodies, both of IgG1 isotype. Mixtures of filtered and neutralized Protein A eluates of parental antibodies in solution were prepared, adjusted to pH 7.9 using 209.5 mM sodium acetate, 300 mM NaCl, pH 7.9, and diluted to a total IgG concentration of 10–35 g / L. A stock solution of 50 mM sodium acetate, 800 mM 2-MEA pH 5.0 was added to each parent mix to achieve a target reducing agent concentration of 35, 50, or 100 mM. Both reduced parent solutions were incubated for 3 hours at either controlled temperatures of 21.5 °C, 24 °C, or left at ambient room temperature. After 3 hours, the incubated mixtures were ultrafiltered and diafiltered against 10 mM Tris, 7.8 mM acetate, pH 7.5. A summary of the operating conditions and analytical results is shown in Table 10. The current study demonstrated oxidized heterodimers under conditions ranging from 21.5 to 24 °C, reducing agent concentrations from 35 to 100 mM, and homodimer mixture protein concentrations from 10 to 35 g / L. Bispecific antibody integrity and purity were assessed by both NR-cSDS and IHC-HPLC using the methods described in Example 1.

[0211] [Table 10] <h2 style=";text-align:left;direction:ltr">

[0212] <h2 style=";text-align:left;direction:ltr"> <h2 style=";text-align:left;direction:ltr"> TIFF0007645799000016.tif199130

Claims

1. A method for producing a heterodimeric antibody, comprising the steps of: a) providing a first homodimeric antibody comprising a first Fc region of an immunoglobulin comprising a first CH3 region and a second homodimeric antibody comprising a second Fc region of an immunoglobulin comprising a second CH3 region, wherein the amino acid sequences of the first CH3 region and the second CH3 region are different and the amino acid sequences of the first CH3 region and the second CH3 region are such that a heterodimeric interaction between the first CH3 region and the second CH3 region is stronger than a homodimeric interaction between the first CH3 regions or a homodimeric interaction between the second CH3 regions; b) combining the first homodimeric antibody and the second homodimeric antibody into a mixture; c) incubating the mixture in the presence of a reducing agent; and d) removing the reducing agent, thereby producing a heterodimeric antibody, wherein in step d), dissolved oxygen (DO 2 ) is controlled to be 15% or less.

2. In steps c) and d), %DO 2 The method of claim 1, wherein the ratio of the saturation temperature to the saturation temperature is controlled to be 15% or less.

3. In step d), the % DO of the mixture 2 is 10% or less, 9% or less, 8% or less, 7% or less, 6% or less, 5% or less, 4% or less, 3% or less, 2% or less, or 1% or less.

4. In step d), the % DO of the mixture 2 The method of any one of claims 1 to 3, wherein the β-amino acid concentration is 4% or less, 3% or less, 2% or less, or 1% or less.

5. %DO 2 The method according to any one of claims 1 to 4, wherein is controlled by an overlay of nitrogen.

6. The method of any one of claims 1 to 5, wherein in step b) the first homodimeric antibody and the second homodimeric antibody are combined into a mixture in a molar ratio of 1:1 to 1:

2.

7. The method of any one of claims 1 to 5, wherein in step b) the first homodimeric antibody and the second homodimeric antibody are combined into a mixture in a molar ratio of 1.05:

1.

8. The method according to any one of claims 1 to 7, wherein the total concentration of immunoglobulins in the mixture is 10.5 g / L.

9. The method of any one of claims 1 to 8, wherein the mixture is incubated in the presence of the reducing agent for 10 minutes or more, such as for 10 minutes to 24 hours.

10. The method according to any one of claims 1 to 9, wherein the reducing agent is 2-mercaptoethylamine (2-MEA), L-cysteine ​​or D-cysteine.

11. The method of claim 10, wherein the concentration of 2-MEA in the mixture is 35 mM.

12. The method according to any one of claims 1 to 11, wherein the mixture comprises a buffer, such as a buffer comprising a sodium acetate buffer, and optionally the buffer further comprises NaCl.

13. 13. The method of claim 12, wherein the buffer comprises 100 mM sodium acetate and 30 mM NaCl.

14. 14. The method of claim 12 or 13, wherein the pH of the buffer is 7.

3.

15. removing said reducing agent from said mixture, optionally wherein said reducing agent is removed by filtration, e.g., diafiltration; The method according to any one of claims 1 to 14, comprising:

16. The method of any one of claims 1 to 15, wherein the first homodimeric antibody and the second homodimeric antibody are of the IgG1, IgG2 or IgG4 isotype.

17. 17. The method of claim 16, wherein the first CH3 domain and the second CH3 domain comprise the following mutations when compared to the wild type IgG1 of SEQ ID NO: 1, the wild type IgG2 of SEQ ID NO: 2, or the wild type IgG4 of SEQ ID NO: 3: F405L / K409R, wild type / F405L_R409K, T350I_K370T_F405L / K409R, K370W / K409R, D399AFGHILMNRSTVWY / K4 09R, T366ADEFGHILMQVY / K409R, L368ADEGHNRSTVQ / K409AGRH, D399FHKRQ / K409AGRH, F405IKLSTVW / K4 09AGRH, Y407LWQ / K409AGRH, T366Y / F405A, T366W / F405W, F405W / Y407A, T394W / Y407T, T394S / Y407A, T 366W / T394S, F405W / T394S, T366W / T366S_L368A_Y407V, L351Y_F405A_Y407V / T394W, T366I_K392M_T39 4W / F405A_Y407V, T366L_K392M_T394W / F405A_Y407V, L351Y_Y407A / T366A_K409F, L351Y_Y407A / T366 V_K409F, Y407A / T366A_K409F, T350V_L351Y_F405A_Y407V / T350V_T366L_K392L_T394W, K409D / D399K, K409E / D399R, K409D_K360D / D399K_E356K, K409D_K360D / D399E_E356K, K409D_K370D / D399K_E357K, K409D_K370D / D399E_E357K, K409D_K392D / D399K_E356K_E357K, or K409D_K392D / D399E_E356K_E357K.

18. The method of claim 16 or 17, wherein the first Fc region and / or the second Fc region comprises one or more mutations that modulate binding of the first Fc region and / or the second Fc region to an Fcγ receptor (FcγR), FcRn, or protein A, compared to a wild-type IgG1 of SEQ ID NO: 1, a wild-type IgG2 of SEQ ID NO: 2, or a wild-type IgG4 of SEQ ID NO: 3, and optionally the FcγR is FcγRI, FcγRIIa, FcγRIIb, or FcγRIII.

19. The one or more substitutions that modulate binding of the first Fc region and / or the second Fc region to Fcγ are selected from the group consisting of L234A_L235A, F234A_L235A, S228P_F234A_L235A, S228P_L234A_L235A, V234A_G237A_P238S_H268A_V309L_A330S_P331S , V234A_G237A, H268Q_V309L_A330S_P331S, S267E_L328F, L234F_L235E_D265A, L234A_L235A_G237A_P238S_H268A_A330S_P331S, or S228P_F234A_L235A_G237A_P238S, or the one or more substitutions that modulate binding of the first Fc region and / or the second Fc region to FcRn are M428L_N434S, M252Y_S254T_T256E, T250Q_M428L, N434A and T307A_E380A_N434A, H435A, P257I_N434H, D376V_N434H, M252Y_S254T_T256E_H433K_N434F, T308P_N434A, or H435R; or the one or more substitutions that modulate binding of the first Fc region and / or the second Fc region to Protein A are Q311R, Q311K, T307P_L309Q, T307P_V309Q, T307P_L309Q_Q311R, T307P_V309Q_Q311R, H435R, or H435R_Y436F; The method of claim 18.

20. The method according to any one of claims 1 to 19, wherein the step of producing the heterodimeric antibody is carried out under GMP-compliant conditions.

21. The method according to any one of claims 1 to 19, wherein the step of producing the heterodimeric antibody is carried out during the manufacture of a drug substance comprising the heterodimeric antibody.

22. The method according to any one of claims 1 to 21, wherein the heterodimeric antibody is a bispecific antibody.

23. 23. The method of claim 22, wherein the bispecific antibody binds to CD3, CD123, BCMA, EGFR, or c-Met, or any combination thereof.

24. 24. The method according to any one of claims 1 to 23, wherein the mass ratio of total immunoglobulins in the mixture to total reducing agent is between 1.0 and 5.

0.

25. 25. The method of claim 24, wherein the mass ratio is between 1.4 and 3.

8.

26. 25. The method of claim 24, wherein the mass ratio is between 3.3 and 4.4.

Images