Modified antigen binding polypeptide constructs and uses thereof

CN106661121BActive Publication Date: 2026-11-03ZYMEWORKS BC INC
View PDF 18 Cites 0 Cited by

Patent Information

Application Number
CN201580036617.1
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Priority Date
2015-04-28
Filing Date
2015-05-29
Publication Date
2026-11-03
Estimated Expiration
2035-05-29

AI Technical Summary

Technical Problem

Existing technologies struggle to effectively address the mixed pairing of heavy and light chains in bispecific antibodies, leading to difficulties in the preparation of bispecific therapeutic agents, particularly in terms of uniform pairing and biological efficacy.

Method used

By introducing amino acid modifications into specific domains of the antibody heavy and light chains, their pairing process is optimized, so that antibody heavy chain H1 preferentially pairs with light chain L1, and antibody heavy chain H2 preferentially pairs with light chain L2, forming a stable heterodimer. The dimerization stability is further improved by modifying the Fc domain.

Benefits of technology

This method enables the preparation of highly pure and efficient bispecific antibodies, improves the accuracy and thermal stability of heavy-light chain pairing, and enhances the affinity and biological efficacy of antigen binding.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN106661121B_ABST
    Figure CN106661121B_ABST
Patent Text Reader

Abstract

The present invention provides a heterodimer pair that can comprise a first heterodimer and a second heterodimer, wherein each heterodimer comprises an immunoglobulin heavy chain or fragment thereof and an immunoglobulin light chain or fragment thereof. At least one of the heterodimers can comprise one or more amino acid modifications in the CH1 and / or CL domains, one or more amino acid modifications in the VH and / or VL domains, or a combination thereof. The modified amino acids can be part of the interface between the light and heavy chains, and are typically modified to create a preferential pairing between each heavy chain and a desired light chain, such that when the two heavy chains and two light chains of the heterodimer pair are co-expressed in a cell, the heavy chain of the first heterodimer preferentially pairs with one of the light chains but not the other. Likewise, the heavy chain of the second heterodimer preferentially pairs with the second light chain but not the first.
Need to check novelty before this filing date? Find Prior Art

Description

[0001] Cross-referencing of related patent applications

[0002] This patent application claims priority to U.S. Provisional Application No. 62 / 003,663, filed May 28, 2014, and U.S. Provisional Application No. 62 / 154,055, filed April 28, 2015, which are incorporated herein by reference in their entirety for all purposes.

[0003] This patent application relates to PCT / CA2013 / 050914, filed November 28, 2013; U.S. Provisional Application No. 61 / 730,906, filed November 28, 2012; U.S. Provisional Application No. 61 / 761,641, filed February 6, 2013; U.S. Provisional Application No. 61 / 818,874, filed May 2, 2013; and U.S. Provisional Application No. 61 / 869,200, filed August 23, 2013, the entire disclosure of each of these patent applications is hereby incorporated by reference in its entirety for all purposes.

[0004] sequence list

[0005] This patent application includes a sequence list submitted electronically in ASCII format, which is incorporated herein by reference in its entirety. The ASCII copy was created on May 29, 2015, named 97993-945204(000110PC)_SL.txt, and is 27,012 bytes in size.

[0006] background

[0007] Bispecific antibodies are capable of binding to two distinct epitopes. These epitopes can be on the same antigen, or each epitope can be on a different antigen. This characteristic of bispecific antibodies makes them a compelling tool for a variety of therapeutic applications where the antibody has the therapeutic benefit of targeting or recruiting more than one molecule in the treatment of a disease. One approach to forming bispecific antibodies involves the simultaneous expression of two distinct antibody heavy chains and two distinct antibody light chains. Properly forming a wild-type-like form of bispecific antibody remains challenging because the antibody heavy chains involve binding to the antibody light chains in a relatively mixed manner. As a result of this mixed pairing, the simultaneous expression of two antibody heavy chains and two antibody light chains naturally leads to mixed heavy-light chain pairing. This mismatch remains a major challenge in the preparation of bispecific therapeutics, where uniform pairing is a fundamental requirement for good processability and biological efficacy.

[0008] Various methods for preparing bispecific antibodies have been described, in which a specific antibody light chain or fragment pairs with a specific antibody heavy chain or fragment. A review of various methods for addressing this problem can be found in Klein et al., (2012) mAbs 4:6, 1-11. International patent application No. PCT / EP2011 / 056388 (WO 2011 / 131746) describes an in vitro method for preparing heterodimeric proteins in which an asymmetric mutation is introduced into the CH3 region of two monospecific initiating proteins to drive a directional “Fab arm” or “half-molecule” exchange between two monospecific IgG4 or IgG4-like antibodies upon incubation under reducing conditions.

[0009] Schaefer et al. (Roche Diagnostics GmbH) described a method for assembling two heavy chains and two light chains derived from two existing antibodies into a human bivalent bispecific IgG antibody without the use of artificial linkers (PNAS(2011)108(27):11187-11192). This method involves exchanging the heavy chain and light chain domains within half of the antigen-binding fragment (Fab) of the bispecific antibody.

[0010] Strop et al. (Rinat-Pfizer Inc.) described a method for preparing stable bispecific antibodies by expressing and purifying two antibodies of interest separately, and then mixing them together under specified redox conditions (J.Mol.Biol.(2012)420:204-19).

[0011] Zhu et al. (Genentech) engineered mutations in the VL / VH interface of a biantibody construct (composed of antibody fragments with variable domains completely lacking constant domains) and prepared heterodimeric biantibodies (Protein Science (1997) 6: 781-788). Similarly, Igawa et al. (Chugai) also engineered mutations in the VL / VH interface of single-chain biantibodies to facilitate selective expression of the biantibody and inhibit its structural isomerization (Protein Engineering, Design & Selection (2010) 23: 667-677).

[0012] US Patent Publication No. 2009 / 0182127 (Novo Nordisk, Inc.) describes the preparation of bispecific antibodies by modifying amino acid residues at the Fc interface and CH1:CL interface of the light chain-heavy chain pair, which reduces the ability of one pair of light chains to interact with the other pair of heavy chains.

[0013] US Patent Publication No. 2014 / 0370020 (Chugai) describes regulating the association between antibody regions by replacing amino acids present at the interface between the CH1 and CL regions with charged amino acids.

[0014] Overview

[0015] This article describes isolated antigen-binding polypeptide constructs comprising at least a first heterodimer and a second heterodimer, wherein the first heterodimer comprises a first immunoglobulin heavy chain polypeptide sequence (H1) and a first immunoglobulin light chain polypeptide sequence (L1); and the second heterodimer comprises a second immunoglobulin heavy chain polypeptide sequence (H2) and a second immunoglobulin light chain polypeptide sequence (L2), wherein at least one of the H1 or L1 sequences of the first heterodimer is different from the corresponding H2 or L2 sequence of the second heterodimer, and wherein H1 and H2 each comprise at least a heavy chain variable domain (V H structural domains) and heavy-chain constant structural domains (C H1 L1 and L2 each contain at least one light chain variable structural domain (V). L (structural domain) and light chain constant structural domain (C L ); and at least one of H1, H2, L1 and L2 comprises at least one amino acid modification of at least one constant domain and / or at least one variable domain, wherein H1 is preferentially paired with L1 compared with L2, and H2 is preferentially paired with L2 compared with L1.

[0016] In some respects, the construct also includes a heterodimer Fc, which contains at least two C atoms. H3 The sequence, wherein the Fc is coupled to a first heterodimer and a second heterodimer via or without one or more connectors, wherein the dimerized C H3 The sequence has a melting temperature (Tm) of about 68 °C or higher (as determined by differential scanning calorimetry (DSC)) and the construct is bispecific.

[0017] In some respects, at least one amino acid modification is selected from at least one amino acid modification shown in the table or examples.

[0018] In some respects, when H1, H2, L1 and L2 are co-expressed in cells or mammalian cells, or when H1, H2, L1 and L2 are co-expressed in cell-free expression systems, or when H1, H2, L1 and L2 are co-generated, or when H1, H2, L1 and L2 are co-generated via redox generation, H1 preferentially pairs with L1 compared to L2, and H2 preferentially pairs with L2 compared to L1.

[0019] In some respects, at least one of H1, H2, L1, and L2 contains V. Hand / or V L At least one amino acid modification of the domain and C H1 and / or C L At least one amino acid of the domain is modified such that H1 preferentially pairs with L1 compared to L2, and / or H2 preferentially pairs with L2 compared to L1.

[0020] In some respects, if H1 contains C H1 If at least one amino acid in the domain is modified, then at least one of L1 and L2 contains C. L At least one amino acid modification in the domain; and / or if H1 contains V H If at least one amino acid in the domain is modified, then at least one of L1 and L2 contains V. L At least one amino acid modification in the domain.

[0021] In some aspects, H1, L1, H2 and / or L2 contain at least 1, 2, 3, 4, 5, 6, 7, 8, 9 or 10 amino acid mutations. In some aspects, at least one of H1, H2, L1 and L2 contains at least 2, 3, 4, 5, 6, 7, 8, 9 or 10 amino acid modifications of at least one constant domain and / or at least one variable domain.

[0022] In some respects, when both L1 and L2 are co-expressed with at least one of H1 and H2, the ratio of the relative pairing of at least one of the H1-L1 and H2-L2 heterodimer pairs to their respective corresponding H1-L2 or H2-L1 heterodimer pairs is greater than 50%, 51%, 52%, 53%, 54%, 55%, 56%, 57%, 58%, 59%, 60%, 61%, 62%, 63%, 64%, 65%, 66%, 67%, 68%, 69%, 70%, 71%, 72%, 73%. %, 74%, 75%, 76%, 77%, 78%, 79%, 80%, 81%, 82%, 83%, 84%, 85%, 86%, 87%, 88%, 89%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, or 99%, and wherein the relative pairing of the modified H1-L1 or H2-L2 heterodimer pairs is greater than the respective relative pairing observed in the corresponding H1-L1 or H2-L2 heterodimer pairs without at least one amino acid modification.

[0023] In some aspects, the thermal stability (as determined by DSF) of at least one of the first and second heterodimers, measured by melting temperature (Tm), is within about 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, or 10 °C of the Tm of the corresponding heterodimer without at least one amino acid modification. In some aspects, the thermal stability (as determined by DSF) of each heterodimer containing at least one amino acid modification, measured by melting temperature (Tm), is within about 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, or 10 °C of the Tm of the corresponding heterodimer without at least one amino acid modification. In some embodiments, the thermal stability (as determined by DSF) of each heterodimer containing at least one amino acid modification, measured by melting temperature (Tm), is within about 0, 1, 2, or 3 °C of the Tm of the corresponding heterodimer without at least one amino acid modification.

[0024] In some respects, the affinity of each heterodimer for the antigen it binds to is within approximately 1, 2, 3, 4, 5, 6, 7, 8, 9, 10 times that of the affinity of the respective unmodified heterodimer for the same antigen (as determined by surface plasmon resonance (SPR) or FACS).

[0025] In some aspects, at least one of H1 and L1 contains at least one domain that includes at least one amino acid modification, producing greater amino acid spatial complementarity when H1 and L1 are not paired compared to L2. In some aspects, at least one of H2 and L2 contains at least one domain that includes at least one amino acid modification, producing greater amino acid spatial complementarity when H2 and L2 are paired compared to L1. In some aspects, at least one of H1 and L1 contains at least one domain that includes at least one amino acid modification, producing greater electrostatic complementarity between charged amino acids when H1 and L1 are paired compared to L2. In some aspects, at least one of H2 and L2 contains at least one domain that includes at least one amino acid modification, producing greater electrostatic complementarity between charged amino acids when H2 and L2 are paired compared to L1.

[0026] In some respects, at least one amino acid modification is a set of mutations shown in at least one of the tables or examples.

[0027] In some respects, the construct also contains Fc, which contains at least two Cs. H3 A sequence in which the Fc is coupled to a first heterodimer and a second heterodimer, or not, through one or more connectors.

[0028] In some aspects, the Fc is a human Fc, human IgG1 Fc, human IgA Fc, human IgG Fc, human IgD Fc, human IgE Fc, human IgM Fc, human IgG2 Fc, human IgG3 Fc, or human IgG4 Fc. In some aspects, the Fc is a heterodimer Fc. In some aspects, the Fc contains C. H3 One or more modifications in at least one of the sequences. In some respects, this dimerization C H3 The sequence has a melting temperature (Tm) of about 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 77.5, 78, 79, 80, 81, 82, 83, 84 or 85 °C or higher (as determined by DSC). In some aspects, during preparation, the Fc is a heterodimer formed with a purity greater than about 75%, 76%, 77%, 78%, 79%, 80%, 81%, 82%, 83%, 84%, 85%, 86%, 87%, 88%, 89%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98% or 99%; or wherein in When expressed or expressed via a single cell, the Fc is formed as a heterodimer with a purity greater than about 75%, 76%, 77%, 78%, 79%, 80%, 81%, 82%, 83%, 84%, 85%, 86%, 87%, 88%, 89%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, or 99%. In some aspects, the Fc contains C H3 One or more modifications in at least one of the sequences contribute to the formation of a heterodimer Fc with stability equivalent to the wild-type homodimer Fc. In some aspects, the Fc also contains at least one C H2 Sequence. In some respects, the C of this Fc H2 The sequence contains one or more modifications. In some respects, the Fc contains one or more modifications that facilitate the selective binding of the Fc-γ receptor.

[0029] In some implementations, the Fc includes:

[0030] i) A heterodimer IgG1 Fc having the modification L351Y_F405A_Y407V in the first Fc polypeptide and the modification T366L_K392M_T394W in the second Fc polypeptide;

[0031] ii) Heterodimer IgG1Fc having the modification L351Y_F405A_Y407V in the first Fc polypeptide and the modification T366L_K392L_T394W in the second Fc polypeptide;

[0032] iii) Heterodimer IgG1Fc having the modification T350V_L351Y_F405A_Y407V in the first Fc polypeptide and the modification T350V_T366L_K392L_T394W in the second Fc polypeptide;

[0033] iv) Heterodimeric IgG1Fc having the modification T350V_L351Y_F405A_Y407V in the first Fc polypeptide and the modification T350V_T366L_K392M_T394W in the second Fc polypeptide; or

[0034] v) A heterodimer IgG1Fc having the modification T350V_L351Y_S400E_F405A_Y407V in the first Fc polypeptide and the modification T350V_T366L_N390R_K392M_T394W in the second Fc polypeptide.

[0035] In some aspects, the Fc is coupled to a heterodimer via one or more adapters, or wherein the Fc is coupled to H1 and H2 via one or more adapters. In some aspects, the one or more adapters are one or more peptide adapters. In some aspects, the one or more adapters comprise one or more antibody hinge regions. In some aspects, the one or more adapters comprise one or more IgG1 hinge regions. In some aspects, the one or more adapters comprise one or more modifications. In some aspects, one or more modifications to the one or more adapters contribute to the selective binding of the Fc-γ receptor.

[0036] In some respects, at least one amino acid modification is at least one amino acid mutation, or at least one amino acid modification is at least one amino acid substitution.

[0037] In some respects, the sequences of each of H1, H2, L1, and L2 are derived from human sequences.

[0038] In some respects, the construct is multispecific or bispecific. In other respects, the construct is multivalent or divalent.

[0039] In some respects, the heterodimers described herein preferentially pair to form bispecific antibodies. For example, in some embodiments, the heavy chain polypeptide sequences H1 and H2 comprise a full-length heavy chain sequence containing a heavy chain constant domain (C). H1 (structural domain), C H2 Domain and C H3Domains. In some embodiments, the percentage of correctly paired heavy and light chains in the bispecific antibody (e.g., H1-L1:H2-L2) is greater than 50%, 51%, 52%, 53%, 54%, 55%, 56%, 57%, 58%, 59%, 60%, 61%, 62%, 63%, 64%, 65%, 66%, 67%, 68%, 69%, 70%, 71%, 72%, 73%, 74%, 75%, 76%, 77%, 78%, 79%, 80%, 81%, 82%, 83%, 84%, 85%, 86%, 87%, 88%, 89%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, or 99%.

[0040] This document also describes isolated polynucleotides or isolated polynucleotide groups comprising at least one sequence encoding the construct or heavy or light chain described herein. In some aspects, the polynucleotide or polynucleotide group is cDNA. This document also describes vectors or vector groups comprising one or more of the polynucleotides or polynucleotide groups described herein. In some aspects, the vector or vector group is selected from plasmids, polycistronic vectors, viral vectors, non-free mammalian vectors, expression vectors, and recombinant expression vectors.

[0041] This article also describes isolated cells containing the polynucleotides or polynucleotide sequences described herein, or the vectors or vector sequences described herein. In some respects, the cells are hybridomas, Chinese hamster ovary (CHO) cells, or HEK293 cells.

[0042] This article also describes pharmaceutical compositions comprising the constructs described herein and pharmaceutically acceptable carriers. In some aspects, the composition further comprises one or more substances selected from buffers, antioxidants, low molecular weight molecules, pharmaceuticals, proteins, amino acids, sugars, lipids, chelating agents, stabilizers, and excipients.

[0043] This article also describes the use of the constructs or pharmaceutical compositions described herein for the treatment of diseases or conditions, or cancer or vascular diseases in subjects, or for use in the manufacture of pharmaceuticals.

[0044] This article also describes methods for treating subjects with diseases or conditions or cancer or vascular diseases, which include administering the constructs or compositions described herein to the subjects.

[0045] This document also describes a method for obtaining the construct described herein from a host cell culture, the method comprising the steps of: (a) obtaining a host cell culture containing at least one host cell that contains one or more nucleic acid sequences encoding the construct; and (b) recovering the construct from the host cell culture.

[0046] This document also describes a method for obtaining the construct described herein, comprising the following steps: (a) obtaining H1, L1, H2, or L2; (b) allowing H1 to be paired with L1 preferentially compared to L2, and H2 to be paired with L2 preferentially compared to L1; and (c) obtaining the construct.

[0047] This document also describes a method for preparing the construct described herein, comprising: obtaining a polynucleotide or polynucleotide set encoding at least one construct; determining an optimal ratio for introducing each of the polynucleotides or polynucleotide sets into at least one host cell, wherein the optimal ratio is determined by evaluating the amount of H1-L1 and H2-L2 heterodimer pairs formed during H1, L1, H2, or L2 expression relative to mismatched H1-L2 and H2-L1 heterodimer pairs formed during H1, L1, H2, or L2 expression; selecting a preferred optimal ratio, wherein transfection of at least one host cell with the polynucleotides or polynucleotide set at the preferred optimal ratio results in expression of the construct; transfecting at least one host cell with the polynucleotides or polynucleotide set at the optimal ratio; and culturing at least one host cell to express the construct.

[0048] In some respects, the selection of the optimal ratio is evaluated through transfection in a transient transfection system. In some respects, transfection of at least one host cell with a preferred optimal ratio of polynucleotides or polynucleotide groups results in optimal expression of the construct. In some respects, the construct contains an Fc containing at least two C... H3 The Fc sequence, wherein the Fc is coupled to a first heterodimer and a second heterodimer, with or without one or more linkers. In some aspects, the Fc is a heterodimer, optionally containing one or more amino acid modifications.

[0049] This document also describes a computer-readable storage medium for storing a dataset containing data representing complementary mutations in a first heterodimer and a second heterodimer, the first heterodimer comprising a first immunoglobulin heavy chain polypeptide sequence (H1) and a first immunoglobulin light chain polypeptide sequence (L1); and the second heterodimer comprising a second immunoglobulin heavy chain polypeptide sequence (H2) and a second immunoglobulin light chain polypeptide sequence (L2), wherein H1 and H2 each contain at least a heavy chain variable domain (V H structural domains) and heavy-chain constant structural domains (C H1 (structural domains); where L1 and L2 each contain at least a light chain variable structural domain (V L (structural domain) and light chain constant structural domain (C LThe dataset of complementary mutations contains data representing those mutations or subsets of those mutations listed in the table or embodiments; and computer-executable code for determining the probability that H1 is preferentially paired with L1 compared to L2 and / or that H2 is preferentially paired with L2 compared to L1.

[0050] This document also describes a computer-implemented method for determining preferred pairings, comprising: obtaining a dataset containing data representing complementary mutations in a first heterodimer and a second heterodimer, the first heterodimer comprising a first immunoglobulin heavy chain polypeptide sequence (H1) and a first immunoglobulin light chain polypeptide sequence (L1); and the second heterodimer comprising a second immunoglobulin heavy chain polypeptide sequence (H2) and a second immunoglobulin light chain polypeptide sequence (L2), wherein H1 and H2 each comprise at least a heavy chain variable domain (V H structural domains) and heavy-chain constant structural domains (C H1 (structural domains); where L1 and L2 each contain at least a light chain variable structural domain (V L (structural domain) and light chain constant structural domain (C L The method includes a structural domain, wherein the dataset of complementary mutations contains data representing those mutations or subsets of those mutations listed in the table or examples; and a computer processor determines the probability that H1 preferentially pairs with L1 compared to L2 and / or H2 preferentially pairs with L2 compared to L1. In some aspects, the method also includes preparing the constructs described herein.

[0051] This article also describes a method for preparing a bispecific antigen-binding polypeptide construct comprising a first heterodimer and a second heterodimer. The first heterodimer comprises a first immunoglobulin heavy chain polypeptide sequence (H1) and a first immunoglobulin light chain polypeptide sequence (L1) from a first single-specific antigen-binding polypeptide; and the second heterodimer comprises a second immunoglobulin heavy chain polypeptide sequence (H2) and a second immunoglobulin light chain polypeptide sequence (L2) from a second single-specific antigen-binding polypeptide, wherein H1 and H2 each comprise at least a heavy chain variable domain (V). H structural domains) and heavy-chain constant structural domains (C H1 (structural domains); where L1 and L2 each contain at least a light chain variable structural domain (V L (structural domain) and light chain constant structural domain (C L The method includes: introducing one or more complementary mutations from the dataset described herein into a first heterodimer and / or a second heterodimer; and co-expressing the first heterodimer and the second heterodimer in at least one host cell to prepare an expression product containing a bispecific construct.

[0052] In some aspects, the method further includes determining the amount of the bispecific construct in the expression product relative to other peptide products to select a preferred subset of complementary mutants. In some aspects, the bispecific construct is generated with a purity greater than 70% (e.g., greater than 75%, 80%, 85%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, or 99%) relative to other peptide products. In some aspects, the dataset is the dataset described herein. In some aspects, the method further includes the step of adding additional amino acid modifications to at least one of H1, H2, L1, or L2 to improve the purity of the bispecific construct relative to other peptide products. In some aspects, the construct comprises an Fc containing at least two Cs. H3 The sequence wherein the Fc is coupled to a first heterodimer and a second heterodimer, with or without one or more linkers. In some aspects, the Fc is a heterodimer, optionally containing one or more amino acid modifications. In some aspects, the antigen-binding polypeptide is an antibody, Fab, or scFv.

[0053] In some embodiments of this construct, H1 and / or H2 contains at least one or a group of amino acid modifications at L124, K145, D146, Q179, and S186, and L1 and / or L2 contains at least one or a group of amino acid modifications at Q124, S131, V133, Q160, S176, T178, and T180. For example, in some embodiments, H1 and / or H2 contains at least one or a group of amino acid modifications selected from L124R, L124E, K145M, K145T, D146N, Q179E, Q179K, S186R, and S186K, and L1 and / or L2 contains at least one or a group of amino acid modifications selected from Q124E, S131R, S131K, V133G, Q160E, S176R, S176D, T178D, T178E, and T180E. In some embodiments, H1 contains an amino acid modification selected from L124E, K145M, K145T, and Q179E or combinations thereof; L1 contains an amino acid modification selected from S131R, S131K, V133G, and S176R or combinations thereof; H2 contains an amino acid modification selected from L124R, D146N, Q179K, S186R, and S186K or combinations thereof; and L2 contains an amino acid modification selected from Q124E, V133G, Q160E, S176D, T178D, T178E, and T180E or combinations thereof. In some embodiments, H1 contains amino acid modifications L124E, K145T, and Q179E; L1 contains amino acid modifications S131K, V133G, and S176R; H2 contains amino acid modifications L124R and S186R; and L2 contains amino acid modifications V133G, S176D, and T178D.

[0054] In some embodiments of the construct, H1 and / or H2 contain at least one or a group of amino acid modifications at L124, L143, K145, D146, Q179, and S186, and L1 and / or L2 contain at least one or a group of amino acid modifications at Q124, V133, Q160, S176, T178, and T180. In some embodiments, H1 and / or H2 comprises at least one or a group of amino acid modifications selected from L124E, L124R, L143E, L143D, K145T, K145M, D146N, Q179K, S186R, and S186K; and L1 and / or L2 comprises at least one or a group of amino acid modifications selected from Q124K, Q124E, V133G, Q160K, S176R, S176D, T178E, T178K, T178R, T178D, and T180E. In some embodiments, H1 contains an amino acid modification selected from L124E, L143E, L143D, K145T, and K145M or combinations thereof; L1 contains an amino acid modification selected from Q124K, V133G, Q160K, S176R, T178K, and T178R or combinations thereof; H2 contains an amino acid modification selected from L124R, D146N, Q179K, S186R, and S186K or combinations thereof; and L2 contains an amino acid modification selected from Q124E, V133G, S176D, T178E, T178D, and T180E or combinations thereof. In some embodiments, H1 comprises amino acid modifications L124E, L143E, and K145T; L1 comprises amino acid modifications Q124K, V133G, and S176R; H2 comprises amino acid modifications L124R and Q179K; and L2 comprises amino acid modifications V133G, S176D, and T178E. In some embodiments, H1 comprises amino acid modifications L124E, L143E, and K145T; L1 comprises amino acid modifications Q124K, V133G, and S176R; H2 comprises amino acid modifications L124R and S186R; and L2 comprises amino acid modifications V133G, S176D, and T178D.

[0055] In some embodiments of the construct, H1 and / or H2 comprises at least one or a group of amino acid modifications at Q39, L45, L124, L143, F122 and H172, and L1 and / or L2 comprises at least one or a group of amino acid modifications at Q38, P44, Q124, S131, V133, N137, S174, S176 and T178. In some embodiments, H1 and / or H2 comprises at least one or a group of amino acid modifications selected from Q39E, Q39R, L45P, F122C, L124E, L124R, L143F, H172T, and H172R; and L1 and / or L2 comprises at least one or a group of amino acid modifications selected from Q38R, Q38E, P44F, Q124C, S131T, S131E, V133G, N137K, S174R, S176R, S176K, S176D, T178Y, and T178D. In some embodiments, H1 contains an amino acid modification selected from Q39E, L45P, F122C, L124E, L143F, H172T, and H172R or combinations thereof; L1 contains an amino acid modification selected from Q38R, P44F, Q124C, S131T, V133G, N137K, S174R, S176R, S176K, and T178Y or combinations thereof; H2 contains an amino acid modification selected from Q39R, L124R, and H172R or combinations thereof; and L2 contains an amino acid modification selected from Q38E, S131E, V133G, S176D, and T178D or combinations thereof. In some embodiments, H1 comprises amino acid modifications Q39E and L124E; L1 comprises amino acid modifications Q38R, V133G, and S176R; H2 comprises amino acid modifications Q39R and L124R; and L2 comprises amino acid modifications Q38E, V133G, and S176D. In some embodiments, H1 comprises amino acid modifications L45P and L124E; L1 comprises amino acid modifications P44F, V133G, and S176R; H2 comprises amino acid modification L124R; and L2 comprises amino acid modifications V133G, S176D, and T178D. In some embodiments, H1 comprises amino acid modifications L124E and L143F; L1 comprises amino acid modifications V133G and S176R; H2 comprises amino acid modification L124R; and L2 comprises amino acid modifications V133G, S176D, and T178D. In some embodiments, H1 contains amino acid modifications F122C and L124E; L1 contains amino acid modifications Q124C, V133G and S176R; H2 contains amino acid modification L124R; and L2 contains amino acid modifications V133G and S176D.In some embodiments, H1 contains amino acid modifications L124E and H172T; L1 contains amino acid modifications V133G, N137K, S174R, and S176R; H2 contains amino acid modifications L124R and H172R; and L2 contains amino acid modifications V133G, S176D, and T178D.

[0056] In some embodiments of the construct, H1 and / or H2 contains at least one or a group of amino acid modifications at L124, A125, H172, and K228, and L1 and / or L2 contains at least one or a group of amino acid modifications at S121, V133, N137, S174, S176, and T178. In some embodiments, H1 and / or H2 contains at least one or a group of amino acid modifications selected from L124E, L124R, A125S, A125R, H172R, H172T, and K228D; and (ii) L1 and / or L2 contains at least one or a group of amino acid modifications selected from S121K, V133G, N137K, S174R, S176K, S176R, S176D, and T178D. In some embodiments, H1 comprises an amino acid modification selected from L124E, A125S, H172R, and K228D, or combinations thereof; L1 comprises an amino acid modification selected from S121K, V133G, and S176R, or combinations thereof; H2 comprises an amino acid modification selected from L124R, A125R, and H172T, or combinations thereof; and L2 comprises an amino acid modification selected from V133G, N137K, S174R, S176D, and T178D, or combinations thereof. In some embodiments, H1 comprises amino acid modifications L124E and K228D; L1 comprises amino acid modifications S121K, V133G, and S176R; H2 comprises amino acid modifications L124R and A125R; and L2 comprises amino acid modifications V133G and S176D. In some embodiments, H1 contains amino acid modifications L124E and H172R; L1 contains amino acid modifications V133G and S176R; H2 contains amino acid modifications L124R and H172T; and L2 contains amino acid modifications V133G, S174R, and S176D.

[0057] In some embodiments of this construct, H1 and / or H2 contains at least one or a group of amino acid modifications at L124, A139, and V190, and L1 and / or L2 contains at least one or a group of amino acid modifications at F116, V133, L135, and S176. In some embodiments, H1 and / or H2 contains at least one or a group of amino acid modifications selected from L124E, L124R, A139W, A139G, and V190A; and L1 and / or L2 contains at least one or a group of amino acid modifications selected from F116A, V133G, L135V, L135W, S176R, and S176D. In some embodiments, H1 comprises an amino acid modification selected from L124E and A139W or combinations thereof; L1 comprises an amino acid modification selected from F116A, V133G, L135V and S176R or combinations thereof; H2 comprises an amino acid modification selected from L124R, A139G and V190A or combinations thereof; and L2 comprises an amino acid modification selected from V133G, L135W and S176D or combinations thereof. In some embodiments, H1 comprises amino acid modifications L124E and A139W; L1 comprises amino acid modifications F116A, V133G, L135V and S176R; H2 comprises amino acid modifications L124R, A139G and V190A; and L2 comprises amino acid modifications V133G, L135W and S176D.

[0058] In some embodiments of this construct, H1 and / or H2 comprises at least one or a group of amino acid modifications at Q39, L45, K145, H172, Q179, and S186, and L1 and / or L2 comprises at least one or a group of amino acid modifications at Q38, P44, Q124, S131, Q160, T180, and C214. In some embodiments, H1 and / or H2 comprises at least one or a group of amino acid modifications selected from Q39E, Q39R, L45P, K145T, H172R, Q179E, and S186R; and L1 and / or L2 comprises at least one or a group of amino acid modifications selected from Q38R, Q38E, P44F, Q124E, S131K, Q160E, T180E, and C214S. In some embodiments, H1 contains amino acid modifications selected from Q39E, L45P, K145T, H172R, and Q179E, or combinations thereof; L1 contains amino acid modifications selected from Q38R, P44F, and S131K, or combinations thereof; H2 contains amino acid modifications selected from Q39R, H172R, and S186R, or combinations thereof; and L2 contains amino acid modifications selected from Q38E, Q124E, Q160E, T180E, and C214S, or combinations thereof. In some embodiments, H1 contains amino acid modifications of Q39E, K145T, and Q179E; L1 contains amino acid modifications of Q38R and S131K; H2 contains amino acid modifications of Q39R and S186R; and L2 contains amino acid modifications of Q38E, Q124E, Q160E, and T180E. In some embodiments, H1 contains amino acid modifications L45P, K145T, H172R, and Q179E; L1 contains amino acid modifications P44F and S131K; H2 contains amino acid modifications H172R and S186R; and L2 contains amino acid modifications Q124E, Q160E, and T180E.

[0059] In some embodiments of this construct, H1 and / or H2 contains at least one or a group of amino acid modifications at A139, L143, K145, Q179, and V190, and L1 and / or L2 contains at least one or a group of amino acid modifications at F116, Q124, L135, Q160, T178, and T180. In some embodiments, H1 and / or H2 contains at least one or a group of amino acid modifications selected from A139W, A139G, L143E, K145T, Q179E, Q179K, and V190A; and L1 and / or L2 contains at least one or a group of amino acid modifications selected from F116A, Q124R, Q124E, L135V, L135W, Q160E, T178R, and T180E. In some embodiments, H1 contains amino acid modifications selected from A139W, L143E, K145T, and Q179E, or combinations thereof; L1 contains amino acid modifications selected from F116A, Q124R, L135V, and T178R, or combinations thereof; H2 contains amino acid modifications selected from A139G, Q179K, and V190A, or combinations thereof; and L2 contains amino acid modifications selected from Q124E, L135W, Q160E, and T180E, or combinations thereof. In some embodiments, H1 contains amino acid modifications A139W, L143E, K145T, and Q179E; L1 contains amino acid modifications F116A, Q124R, L135V, and T178R; H2 contains amino acid modifications Q179K; and L2 contains amino acid modifications Q124E, L135W, Q160E, and T180E.

[0060] In some embodiments of this construct, H1 and / or H2 comprises at least one or a group of amino acid modifications at Q39, L143, K145, D146, H172, and Q179, and L1 and / or L2 comprises at least one or a group of amino acid modifications at Q38, Q124, Q160, T178, and T180. In some embodiments, H1 and / or H2 comprises at least one or a group of amino acid modifications selected from Q39E, Q39R, L143E, K145T, D146G, H172R, Q179E, and Q179K; and L1 and / or L2 comprises at least one or a group of amino acid modifications selected from Q38R, Q38E, Q124R, Q124E, Q160K, Q160E, T178R, and T180E. In some embodiments, H1 contains an amino acid modification selected from Q39E, L143E, K145T, H172R, and Q179E or combinations thereof; L1 contains an amino acid modification selected from Q38R, Q124R, Q160K, and T178R or combinations thereof; H2 contains an amino acid modification selected from Q39R, H172R, and Q179K or combinations thereof; and L2 contains an amino acid modification selected from Q38E, Q124E, D146G, Q160E, and T180E or combinations thereof. In some embodiments, H1 comprises amino acid modifications Q39E, L143E, K145T, and Q179E; L1 comprises amino acid modifications Q38R, Q124R, Q160K, and T178R; H2 comprises amino acid modifications Q39R, H172R, and Q179K; and L2 comprises amino acid modifications Q38E, Q124E, Q160E, and T180E.

[0061] In some embodiments of the construct, H1 and / or H2 comprises at least one or a group of amino acid modifications at L45, L143, K145, D146, H172 and Q179, and L1 and / or L2 comprises at least one or a group of amino acid modifications at Q38, P44, Q124, N137, Q160, S174, T178, T180 and C214. In some embodiments, H1 and / or H2 comprises at least one or a group of amino acid modifications selected from L45P, L143E, K145T, D146G, H172R, H172T, Q179E, and Q179K; and (ii) L1 and / or L2 comprises at least one or a group of amino acid modifications selected from Q38E, P44F, Q124R, Q124E, N137K, Q160K, Q160E, S174R, T178R, T180E, and C214S. In some embodiments, H1 contains an amino acid modification selected from L45P, L143E, K145T, H172R, and Q179E or combinations thereof; L1 contains an amino acid modification selected from P44F, Q124R, Q160K, and T178R or combinations thereof; H2 contains an amino acid modification selected from D146G, H172R, H172T, and Q179K or combinations thereof; and L2 contains an amino acid modification selected from Q38E, Q124E, N137K, Q160E, S174R, T180E, and C214S or combinations thereof. In some embodiments, H1 comprises amino acid modifications L45P, L143E, and K145T; L1 comprises amino acid modifications P44F, Q124R, Q160K, and T178R; H2 comprises amino acid modifications D146G and Q179K; and L2 comprises amino acid modifications Q38E, Q124E, Q160E, and T180E. In some embodiments, H1 comprises amino acid modifications L143E, K145T, and H172R; L1 comprises amino acid modifications Q124R, Q160K, and T178R; H2 comprises amino acid modifications H172T and Q179K; and L2 comprises amino acid modifications Q124E, Q160E, N137K, S174R, and T180E.

[0062] In some embodiments of this construct, H1 and / or H2 comprises at least one or a group of amino acid modifications at L124, L143, K145, and Q179, and L1 and / or L2 comprises at least one or a group of amino acid modifications at Q124, S131, V133, S176, T178, and T180. In some embodiments, H1 and / or H2 comprises at least one or a group of amino acid modifications selected from L124W, L124A, L143E, L143F, K145T, Q179E, and Q179K; and L1 and / or L2 comprises at least one or a group of amino acid modifications selected from Q124R, Q124K, Q124E, S131K, V133A, V133W, S176T, T178R, T178L, T178E, and T180E. In some embodiments, H1 contains an amino acid modification selected from L124W, L143E, K145T, and Q179E or combinations thereof; L1 contains an amino acid modification selected from Q124R, Q124K, S131K, V133A, S176T, T178R, and T178L or combinations thereof; H2 contains an amino acid modification selected from L124A, L143F, and Q179K or combinations thereof; and L2 contains an amino acid modification selected from Q124E, V133W, S176T, T178L, T178E, and T180E or combinations thereof. In some embodiments, H1 comprises amino acid modifications L124W, L143E, K145T, and Q179E; L1 comprises amino acid modifications Q124R, V133A, S176T, and T178R; H2 comprises amino acid modifications L124A, L143F, and Q179K; and L2 comprises amino acid modifications Q124E, V133W, S176T, T178L, and T180E.

[0063] In some embodiments of the construct, H1 and / or H2 contain at least one or a group of amino acid modifications at A139, L143, K145, Q179 and S186, and L1 and / or L2 contain at least one or a group of amino acid modifications at F116, Q124, V133, Q160, T178 and T180. In some embodiments, H1 and / or H2 comprises at least one or a group of amino acid modifications selected from A139C, L143E, L143D, L143R, L143K, K145T, Q179E, Q179D, Q179R, Q179K, S186K, and S186R; and L1 and / or L2 comprises at least one or a group of amino acid modifications selected from F116C, Q124R, Q124K, Q124E, V133E, V133D, Q160K, Q160E, T178R, T178K, T178E, and T180E. In some embodiments, H1 contains an amino acid modification selected from A139C, L143E, L143D, K145T, Q179E and Q179D or combinations thereof; L1 contains an amino acid modification selected from F116C, Q124R, Q124K, Q160K, T178R and T178K or combinations thereof; H2 contains an amino acid modification selected from L143R, L143K, Q179R, Q179K, S186K and S186R or combinations thereof; and L2 contains an amino acid modification selected from Q124E, V133E, V133D, Q160E, T178E and T180E or combinations thereof. In some embodiments, H1 comprises amino acid modifications A139C, L143E, K145T, and Q179E; L1 comprises amino acid modifications F116C, Q124R, and T178R; H2 comprises amino acid modification Q179K; and L2 comprises amino acid modifications Q124E, Q160E, and T180E. In some embodiments, H1 comprises amino acid modifications L143E, K145T, and Q179E; L1 comprises amino acid modifications Q124R and T178R; H2 comprises amino acid modification S186K; and L2 comprises amino acid modifications Q124E, Q160E, and T178E. In some embodiments, H1 comprises amino acid modifications L143E, K145T, and Q179E; L1 comprises amino acid modifications Q124R and T178R; H2 comprises amino acid modification L143R; and L2 comprises amino acid modifications Q124E and V133E.

[0064] In some embodiments of the construct, H1 and / or H2 comprises at least one or a group of amino acid modifications at L124, L143, K145, D146, Q179, S186, and S188, and L1 and / or L2 comprises at least one or a group of amino acid modifications at Q124, S131, V133, Q160, S176, T178, and T180. In some embodiments, H1 and / or H2 comprises at least one or a group of amino acid modifications selected from L124A, L143A, L143R, L143E, L143K, K145T, D146G, Q179R, Q179E, Q179K, S186R, S186K, and S188L; and L1 and / or L2 comprises at least one or a group of amino acid modifications selected from Q124R, Q124E, S131E, S131T, V133Y, V133W, V133E, V133D, Q160E, Q160K, Q160M, S176L, T178R, T178E, T178F, T178Y, and T180E. In some embodiments, H1 contains an amino acid modification selected from L143E, K145T, Q179E and S188L or combinations thereof; L1 contains an amino acid modification selected from Q124R, Q160K and T178R or combinations thereof; H2 contains an amino acid modification selected from L124A, L143A, L143R, L143K, D146G, Q179R, Q179K, S186R and S186K or combinations thereof; and L2 contains an amino acid modification selected from Q124E, S131E, S131T, V133Y, V133W, V133E, V133D, Q160E, Q160M, S176L, T178E, T178F, T178Y and T180E or combinations thereof. In some embodiments, H1 comprises amino acid modifications of L143E, K145T, Q179E, and S188L; L1 comprises amino acid modifications of Q124R and T178R; H2 comprises amino acid modification of S186K; and L2 comprises amino acid modifications of Q124E, S176L, and T180E. In some embodiments, H1 comprises amino acid modifications of L143E, K145T, Q179E, and S188L; L1 comprises amino acid modifications of Q124R and T178R; H2 comprises amino acid modification of S186K; and L2 comprises amino acid modifications of Q124E, S131T, T178Y, and T180E. In some embodiments, H1 comprises amino acid modifications of L143E and K145T; L1 comprises amino acid modifications of Q124R, Q160K, and T178R; H2 comprises amino acid modification of S186K; and L2 comprises amino acid modification of S131E.In some embodiments, H1 contains amino acid modifications L143E and K145T; L1 contains amino acid modification Q124R; H2 contains amino acid modification L143R; and L2 contains amino acid modifications Q124E and V133E.

[0065] In some embodiments of this construct, H1 contains at least one or a group of amino acid modifications at F122 and C233, and L1 contains at least one or a group of amino acid modifications at Q124 and C214. In some embodiments, H1 contains at least one or a group of amino acid modifications selected from F122C and C233S; and L1 contains at least one or a group of amino acid modifications selected from Q124C and C214S. In some embodiments, H1 contains amino acid modifications selected from F122C and C233S or combinations thereof; L1 contains amino acid modifications selected from Q124C and C214S or combinations thereof; H2 contains wild-type or unmodified amino acid sequences; and L2 contains wild-type or unmodified amino acid sequences. In some embodiments, H1 contains amino acid modifications F122C and C233S; L1 contains amino acid modifications Q124C and C214S; H2 contains wild-type or unmodified amino acid sequences; and L2 contains wild-type or unmodified amino acid sequences.

[0066] In some implementations, the construct includes SMCA designs selected from the tables herein: 9561-9095_1, 9561-9095_2, 9121-9373_1, 9121-9373_2, 9116-9349_1, 9116-9349_2, 9134-9521_1, 9134-9521_2, 9286-9402_1, 9286-9402_2. Amino acid modifications of 9667-9830_1, 9667-9830_2, 9696-9848_1, 9696-9848_2, 9060-9756_1, 9060-9756_2, 9682-9740_1, 9682-9740_2, 9049-9759_1, 9049-9759_2, 9820-9823_1, and 9820-9823_2. In some implementations, the construct comprises amino acid modifications of SMCA designs 9327-6054_1, 9815-9825_1, 9815-9825_2, 9587-9735_1, 9587-9735_2, 3522_1, 3522_2, 3519_1, and 3519_2 selected from the tables herein.

[0067] In some embodiments, H1 and / or H2 do not contain the amino acid modification at position Q179. In some embodiments, H1 does not contain the amino acid modification Q179E and / or H2 does not contain the amino acid modification Q179K. In some embodiments, L1 does not contain the amino acid modification at position S131. In one embodiment, L1 does not contain the amino acid modification S131K. In some embodiments, L2 does not contain the amino acid modification at position T180. In one embodiment, L2 does not contain the amino acid modification T180E. In some embodiments, the construct does not contain such a combination of amino acid modifications where H1 contains Q179E, L1 contains S131K, H2 contains Q179K, and L2 contains T180E.

[0068] In some embodiments, H1 does not contain amino acid modifications at positions Q39 and / or Q179. In some embodiments, H1 does not contain amino acid modifications Q39E and / or Q179E. In some embodiments, L1 does not contain amino acid modifications at position Q160. In one embodiment, L1 does not contain amino acid modifications Q160K. In some embodiments, H2 does not contain amino acid modifications at position Q179. In one embodiment, H2 does not contain amino acid modifications Q179K. In some embodiments, L2 does not contain amino acid modifications at positions Q38, Q160, and / or T180. In one embodiment, L2 does not contain amino acid modifications Q38E, Q160E, and / or T180E. In some embodiments, the construct does not contain such a combination of amino acid modifications where H1 contains Q39E and / or Q179E, L1 contains Q160K, H2 contains Q179K, and L2 contains Q38E, Q160E, and / or T180E. For example, in some embodiments, the construct does not contain such amino acid modification combinations, wherein: (i) H1 contains Q179E, L1 contains Q160K, H2 contains Q179K, and L2 contains Q160E and T180E; (ii) H1 contains Q39E and Q179E, L1 contains Q160K, H2 contains Q179K, and L2 contains Q38E, Q160E, and T180E; or (iii) H1 contains Q39E, L1 contains Q160K, H2 contains Q179K, and L2 contains Q38E, Q160E, and T180E.

[0069] In some embodiments, H1 does not contain an amino acid modification at position Q179. In some embodiments, H1 does not contain amino acid modifications Q179K or Q179E. In some embodiments, L1 does not contain amino acid modifications at positions Q160 and / or T180. In one embodiment, L1 does not contain amino acid modifications Q160E, Q160K, and / or T180E. In some embodiments, H2 does not contain an amino acid modification at position Q179. In one embodiment, H2 does not contain amino acid modifications Q179K or Q179E. In some embodiments, L2 does not contain amino acid modifications at positions Q160 and / or T180. In one embodiment, L2 does not contain amino acid modifications Q160K, Q160E, and / or T180E. In some embodiments, the construct does not contain such a combination of amino acid modifications, wherein H1 contains Q179K or Q179E, L1 contains Q160E, Q160K and / or T180E, H2 contains Q179K or Q179E, and L2 contains Q160K, Q160E and / or T180E.

[0070] In some embodiments, H1 and / or H2 do not contain the amino acid modification at position Q179. In some embodiments, H1 does not contain the amino acid modification Q179K and / or H2 does not contain the amino acid modification Q179E. In some embodiments, L1 does not contain the amino acid modification at position T180. In one embodiment, L1 does not contain the amino acid modification T180E. In some embodiments, L2 does not contain the amino acid modification at position S131. In one embodiment, L2 does not contain the amino acid modification S131K. In some embodiments, the construct does not contain such a combination of amino acid modifications where H1 contains Q179K, L1 contains T180E, H2 contains Q179E, and L2 contains S131K.

[0071] In some embodiments, H1 does not contain the amino acid modification at position Q179. In some embodiments, H1 does not contain the amino acid modification Q179E. In some embodiments, L1 does not contain the amino acid modification at position Q160. In one embodiment, L1 does not contain the amino acid modification Q160K. In some embodiments, H2 does not contain the amino acid modification at position Q179. In one embodiment, H2 does not contain the amino acid modification Q179K. In some embodiments, L2 does not contain the amino acid modification at position T180. In one embodiment, L2 does not contain the amino acid modification T180E. In some embodiments, the construct does not contain such a combination of amino acid modifications where H1 contains Q179E, L1 contains Q160K, H2 contains Q179K, and L2 contains T180E.

[0072] In some embodiments, H1 does not contain the amino acid modification at position A139. In some embodiments, H1 does not contain the amino acid modification A139C. In some embodiments, L1 does not contain the amino acid modification at position F116. In one embodiment, L1 does not contain the amino acid modification F116C. In some embodiments, the construct does not contain such a combination of amino acid modifications where H1 contains A139C and L1 contains F116C.

[0073] In some embodiments, the construct does not contain a native disulfide bond between the heavy and light chains. For example, in some embodiments, the cysteine ​​residue at position 214 of L1 and / or L2 is modified with another amino acid. In some embodiments, L1 and / or L2 contains the amino acid modification C214S. In some embodiments, the cysteine ​​residue at position 233 of H1 and / or H2 is modified with another amino acid. In one embodiment, H1 and / or H2 contains the amino acid modification C233S.

[0074] The implementation scheme described in this article is applicable to both Fab-based and full-antibody constructs. Brief description of the attached diagram

[0076] Figure 1 shows the alignment of the D3H44 heavy and light chain amino acid sequences with the variable, constant, and J regions of a typical human germline sequence (Note: *Sequence identity). Figure 1A The human VH germline subpopulations are shown (one representative sequence is displayed for each family). Sequence identity is based on the alignment of D3H44 with VH3 and IGHJ3*02. Figure 1B Human κVL germline subgroups are shown (one representative sequence is displayed for each family). Sequence identity is based on the alignment of D3H44 with VKI and IGKJ1*01. Figure 1C The human λVL germline subpopulations are shown (one representative sequence is displayed for each family). Sequence identity is based on the alignment of D3H44 with VL1 and IGLJ1*01. Figure 1D The human CH1 allele sequence is shown. Figure 1E The human κ and λ allele sequences are shown.

[0077] Figure 2 A flowchart is shown for identifying key interface residues and for computational modeling for designing using preferred heavy-light chain pairings.

[0078] Figure 3Exemplary sets of chains H1, L1, H2, and L2 are shown, designed such that H1 preferentially pairs with L1 compared to L2, and H2 preferentially pairs with L2 compared to L1. An animated representation of the 3D crystal structure of the variable region heavy and light chain interfaces is provided. The introduction of abrupt changes at the interfaces enables the preferential formation of specific pairs H1-L1 and H2-L2 with electrostatic and spatial complementarity, respectively. On the other hand, incorrect pairings result in unfavorable spatial and electrostatic mismatches, leading to a reduced pairing tendency and decreased stability.

[0079] Figure 4 This diagram presents a high-level schematic overview of the engineering requirements for forming bispecific mabs (monoclonal antibodies) and the analytical requirements for heavy-light chain pair quantification. The design requirements for high-purity (i.e., HL-associated with few or no mismatches) engineered bispecific mabs can be achieved by rationally engineering (through the introduction of specific amino acid mutations) the preferential pairing of two distinct heavy chains with their distinct homologous light chains. This process is schematically illustrated; here, H1 is engineered to preferentially pair with L1 rather than L2. Similarly, H2 is engineered to preferentially pair with L2 rather than L1. Experimental screening for bispecific mab design requires analyses capable of simultaneously quantifying H1-L1:H1-L2 and H2-L2:H2-L1. These analytical requirements can be simplified by the assumption that each bispecific Fab arm can be independently engineered. In this case, the analysis only needs to quantify H1-L1:H1-L2 or H2-L2:H2-L1, without the need for simultaneous quantification of both.

[0080] Figure 5 A schematic diagram is provided illustrating how to determine the tags and preferred pairings of heavy and light chains. In this diagram, circles represent cells transfected with three constructs. The expression product is secreted from the cells, causing the supernatant (SPNT) to flow through the detection device (in this case, an SPR chip). Quantitative estimation of the preferred pairing of the heavy chain with the two light chains can be performed based on the detection levels of two different tags fused to the two light chains that are competing for heavy chain pairing.

[0081] Figure 6 Box plots are shown, illustrating the pairings for each cluster: the average LCCA performance of mismatched Fab heterodimers is at least 86:14.

[0082] Figure 7 The diagram shows representative UPLC-SEC plots of A) WT Fab heterodimers and B) Fab heterodimers of representative designs (H1L1 Fab components of LCCA designs 9735, 9737 and 9740).

[0083] Figure 8This diagram illustrates the expected heavy chain association products when two distinct light chains are co-expressed with two distinct heavy chains in cells. Preferred pairings were assessed using SMCA (monoclonal antibody competition assay).

[0084] Figure 9 illustrates the preference / strand utilization bias in the D3H44 / trastuzumab, D3H44 / cetuximab, and trastuzumab / cetuximab bispecific systems. Strand utilization of different strands was assessed by LC-MS. The x-axis represents the H1:H2:L1:L2 DNA ratio, and the Y-axis shows the corresponding percentage of each strand in different transfection experiments. In the balanced system, all H and L strands accounted for 25%. A preference for one light strand was observed in all bispecific systems.

[0085] Figure 10 Representative UPLC-SEC spectra of WT heterodimer and engineered heterodimer antibodies are shown. Figure 10 a and 10b relate to D3H44 / trastuzumab WT and 9060-9756_1, respectively. Figure 10 c and 10d involve D3H44 / cetuximab WT and 9820-9823_1, respectively. Figure 10 e and 10f involve trastuzumab / cetuximab WT and 9696-9848_1, respectively.

[0086] Figure 11 shows a box plot of the percentage change for correctly paired Fab components versus all those using the same heavy chain and mismatched Fab components (H1:L1 versus all H1 species competing with D3H44 / trastuzumab and D3H44 / cetuximab relative to wild type; H2:L2 versus all H2 species competing with trastuzumab / cetuximab relative to wild type) and the percentage change for the required bispecific antibody relative to wild type competing for each cluster of engineered bispecific antibody samples. The percentage change for correctly paired Fab components versus all those using the same heavy chain and clusters and mismatched Fab components for each system is shown, where a) D3H44 / trastuzumab, c) D3H44 / cetuximab, and e) trastuzumab / cetuximab. The percentage change for the required bispecific antibody relative to wild type and clusters for each system is shown, where b) D3H44 / trastuzumab, d) D3H44 / cetuximab, and f) trastuzumab / cetuximab. In all bispecific systems, the percentage variation of correctly paired Fab components with all those utilizing the same heavy chains and clusters, and mismatched Fab components is shown in the figure. Figure 11g The percentage change in required bispecific antibodies relative to wild-type and cluster is shown in the figure. Figure 11h Note that the reported values ​​also include projected variations in engineered bispecific antibody samples for which the corresponding wild-type constructs were not evaluated via SMCA.

[0087] Figure 12 A method for preparing bispecific antibodies using the specific mutation-pair library provided herein is illustrated.

[0088] Detailed Explanation

[0089] This article provides antigen-binding peptide constructs (also known as heterodimer pairs) that may comprise a first heterodimer and a second heterodimer, wherein each heterodimer comprises an immunoglobulin heavy chain or a fragment thereof and an immunoglobulin light chain. The two heterodimers may comprise one or more amino acid modifications in the constant domain 1 (CH1) of the immunoglobulin heavy chain and one or more amino acid modifications in the constant domain (CL) of the immunoglobulin light chain; one or more amino acid modifications in the variable domain (VH) of the immunoglobulin heavy chain and one or more amino acid modifications in the variable domain (VL) of the immunoglobulin light chain; or a combination of the aforementioned amino acid modifications and both the constant and variable domains of the heavy and light chains. The modified amino acids are typically part of the interface between the light and heavy chains and are modified to generate preferential pairing between each heavy chain and the desired light chain, such that the heavy chain of the first heterodimer preferentially pairs with one light chain rather than the other. Similarly, the heavy chain of the second heterodimer may preferentially pair with the second light chain rather than the first light chain.

[0090] As described above, the specific combinations of amino acid modifications described herein facilitate preferential pairing of the heavy chain with specific light chains, thereby enabling the bispecific monoclonal antibody (Mab) to be expressed with negligible or limited mismatches and minimizing the need for the desired heterodimer to purify unwanted or mismatched products. The heterodimer can exhibit thermostability equivalent to that of the heterodimer without amino acid modifications, and can also exhibit antigen-binding affinity equivalent to that of the heterodimer without amino acid modifications. The design of first and second heterodimers can be used to generate bispecific antibodies targeting two different therapeutic targets or targeting two different epitopes (overlapping or non-overlapping) in the same antigen.

[0091] This article also provides a method for preparing heterodimer pairs.

[0092] definition

[0093] 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 the claimed subject matter pertains. If multiple definitions exist for any term herein, this section shall prevail. When citing URLs or other such identifiers or addresses, it should be understood that such identifiers are subject to change, and specific information on the Internet is constantly evolving, but equivalent information can be obtained by searching the Internet. Such citations attest to their availability and public dissemination.

[0094] It should be understood that the above general description and the following detailed description are merely exemplary and illustrative, and are not limited to any of the claimed subjects. In this patent application, unless otherwise specifically indicated, the use of the singular includes the plural as well.

[0095] In specific embodiments of the invention, any concentration range, percentage range, ratio range, or integer range should be understood to include any integer value within the listed range, and (if applicable) fractions thereof (e.g., one-tenth and one-hundredth of an integer), unless otherwise specified. As used herein, “about” means ±10% of an indicated range, value, sequence, or structure, unless otherwise specified. It should be understood that, as used herein, the terms “an” and “a” mean “one or more” of the listed components, unless the context otherwise indicates or requires. The use of alternative terms (e.g., “or”) should be understood to mean one, both, or any combination of alternative forms. As used herein, the terms “comprising” and “including” may be used as synonyms. Furthermore, it should be understood that this patent application discloses various combinations of single-chain polypeptides or immunoglobulin constructs derived from the structures and substituents described herein, to the extent that each single-chain polypeptide or heterodimer appears to be individually represented. Therefore, the selection of specific components forming a single-chain polypeptide or heterodimer is within the scope of this disclosure.

[0096] The subheadings used herein are for organizational purposes only and should not be construed as limiting the subject matter. All documents or portions thereof referenced in this patent application, including but not limited to patents, patent applications, articles, books, manuals, and discussions, are expressly incorporated herein by reference in their entirety for any purpose.

[0097] It should be understood that the methods and compositions described herein are not limited to the specific methods, protocols, cell lines, constructs, and reagents described herein, and therefore can vary. It should also be understood that the terminology used herein is for descriptive purposes only and is not intended to limit the scope of the methods and compositions described herein, which is limited only by the appended claims.

[0098] All disclosures and patents mentioned herein are incorporated herein by reference for descriptive and disclosure purposes, such as the constructs and methods described in the disclosures, which may be used in conjunction with the methods, compositions, and compounds described herein. The disclosures discussed herein are provided separately to reflect prior disclosure prior to the filing date of this patent application. This document should not be construed as allowing the inventors described herein to have prior knowledge of such disclosures in light of the foregoing invention or for any other reason.

[0099] In this patent application, the use of amino acid names and atom names (e.g., N, O, C, etc.) is as defined in Protein Data Bank (PDB) (www.pdb.org), which are based on IUPAC nomenclature and symbolism for Amino Acids and Peptides, Eur. J. Biochem., 138, 9-37 (1984) and its revision Eur. J. Biochem., 152, 1 (1985)). The term "amino acid residue" is primarily intended to refer to the amino acid residues contained in the group consisting of 20 naturally occurring amino acids: alanine (Ala or A), cysteine ​​(Cys or C), aspartic acid (Asp or D), glutamic acid (Glu or E), phenylalanine (Phe or F), glycine (Gly or G), histidine (His or H), isoleucine (Ile or I), lysine (Lys or K), leucine (Leu or L), methionine (Met or M), asparagine (Asn or N), proline (Pro or P), glutamine (Gln or Q), arginine (Arg or R), serine (Ser or S), threonine (Thr or T), valine (Val or V), tryptophan (Trp or W), and tyrosine (Tyr or Y).

[0100] The terms “polypeptide,” “peptide,” and “protein” are used interchangeably herein and refer to a polymer of amino acid residues. That is, the description of a polypeptide also applies to the description of a peptide and the description of a protein, and vice versa. The term applies to naturally occurring amino acid polymers as well as amino acid polymers in which one or more amino acid residues are non-naturally encoded amino acids. As used herein, the term covers amino acid chains of any length, including full-length proteins in which amino acid residues are linked by covalent peptide bonds.

[0101] The terms "nucleotide sequence" or "nucleic acid sequence" are intended to refer to a continuous segment of two or more nucleotide molecules. Nucleotide sequences can be genomic, cDNA, RNA, semi-synthetic or synthetically derived, or any combination thereof.

[0102] The terms “cell,” “host cell,” “cell line,” and “cell culture” are used interchangeably in this document, and all such terms should be understood to include progeny produced by cell growth or culture. “Transformation” and “transfection” are used interchangeably and refer to the process of introducing a nucleic acid sequence into a cell.

[0103] The term "amino acid" refers to naturally occurring and non-naturally occurring amino acids, as well as amino acid analogs and amino acid mimics that function in a manner similar to naturally occurring amino acids. Naturally encoded amino acids are 20 common amino acids (alanine, arginine, asparagine, aspartic acid, cysteine, glutamine, glutamic acid, glycine, histidine, isoleucine, leucine, lysine, methionine, phenylalanine, proline, serine, threonine, tryptophan, tyrosine, and valine), as well as pyrrolidone and selenocysteine. Amino acid analogs are compounds that have the same basic chemical structure as naturally occurring amino acids, i.e., carbon bonded to hydrogen, a carboxyl group, an amino group, and an R group, such as homoserine, ortholeucine, methionine sulfoxide, and methionine methylsulfonium. These analogs have modified R groups (such as ortholeucine) or modified peptide backbones, but retain the same basic chemical structure as naturally occurring amino acids. The reference to amino acids includes, for example, naturally occurring proteinaceous L-amino acids; chemically modified amino acids, such as amino acid variants and derivatives; naturally occurring non-proteinaceous amino acids, such as alanine, ornithine, etc.; and chemically synthesized compounds having properties known in the art to be characteristic of amino acids. Examples of non-naturally occurring amino acids include, but are not limited to, -methyl amino acids (e.g., methylalanine), D-amino acids, histidine-like amino acids (e.g., 2-amino-histidine, hydroxy-histidine, homohistidine), amino acids having an additional methylene group in the side chain (“homo” amino acids), and amino acids in which the carboxylic acid functional group in the side chain is replaced by a sulfonic acid group (e.g., sulfoalanine). It is advantageous to incorporate non-natural amino acids (including synthetic non-natural amino acids, substituted amino acids, or one or more D-amino acids) into the proteins of the present invention in a variety of different ways. D-amino acid-containing peptides, etc., exhibit increased in vitro or in vivo stability than their L-amino acid-containing counterparts. Therefore, the construction of peptides, etc., incorporating D-amino acids can be particularly useful when greater intracellular stability is desired or required. More specifically, D-peptides and the like are resistant to endogenous peptidases and proteases, thus providing improved molecular bioavailability and extended lifespan in vivo when such properties are desired. Furthermore, D-peptides and the like cannot be efficiently processed for presentation to helper T cells restricted by major histocompatibility complexes (MHCs) of type II, therefore, they are unlikely to elicit a humoral immune response throughout the organism.

[0104] Amino acids are represented in this article using the well-known three-letter symbols recommended by the IUPAC-IUB Biochemical Nomenclature Commission or by a one-letter symbol. Similarly, nucleotides can be represented by a widely accepted single-letter code.

[0105] The term "conserved variant" applies to both amino acid and nucleic acid sequences. For a specific nucleic acid sequence, a "conserved variant" refers to a nucleic acid that encodes the same or substantially the same amino acid sequence, or a substantially identical sequence in which the nucleic acid does not encode an amino acid sequence. Due to the degeneracy of the genetic code, many functionally identical nucleic acids encode any given protein. For example, codons GCA, GCC, GCG, and GCU all encode the amino acid alanine. Therefore, at each position of alanine specified by the codon, the codon can be changed to any corresponding codon without altering the encoded polypeptide. Such nucleic acid variants are "silent variants," a type of conserved variant. Each nucleic acid sequence encoding a polypeptide described herein also describes each possible silent variant of the nucleic acid. Those skilled in the art will recognize that each codon in a nucleic acid (except AUG and TGG, where AUG is typically the only codon for methionine and TGG is typically the only codon for tryptophan) can be modified to obtain a functionally identical molecule. Therefore, each silent variant of a nucleic acid encoding a polypeptide is implicit in each of the stated sequences.

[0106] Regarding amino acid sequences, those skilled in the art will recognize that a single substitution, deletion, or addition (that is, alteration, addition, or deletion of a single amino acid or a few amino acids in the coding sequence) of a nucleic acid, peptide, polypeptide, or protein sequence is a "conserved variant" where the change results in the deletion, addition, or substitution of an amino acid for a chemically similar amino acid. A table of conserved substitutions provides functionally similar amino acids, which is known to those skilled in the art. Such conserved variants also include, and do not exclude, the polymorphic variants, interspecies homologues, and alleles of the present invention.

[0107] A table of conserved substitutions provides functionally similar amino acids, as is known to those skilled in the art. Each of the following eight groups contains amino acids that are conserved substitutions for each other:

[0108] Alanine (A), glycine (G);

[0109] Aspartic acid (D), glutamic acid (E);

[0110] Asparagine (N), glutamine (Q);

[0111] Arginine (R), lysine (K);

[0112] Isoleucine (I), leucine (L), methionine (M), valine (V);

[0113] Phenylalanine (F), tyrosine (Y), tryptophan (W);

[0114] Serine (S), threonine (T); and

[0115] Cysteine ​​(C), Methionine (M)

[0116] (See, for example, Creighton, Proteins: Structures and Molecular Properties (W H Freeman & Co.; 2nd edition (December 1993)).

[0117] In the context of two or more nucleic acid or polypeptide sequences, the term "identical" or percentage "identity" refers to two or more identical sequences or subsequences. When performing a maximum correspondence comparison and alignment on a comparison window or specified region, a sequence is considered "substantially identical" (i.e., possessing at least about 50%, 55%, 60%, 65%, 70%, 75%, 80%, 85%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, or 99% identity with the specified region) if such identity is determined by manual alignment and visual inspection, using one of the following sequence comparison algorithms (or other algorithms available to those skilled in the art). This definition also relates to complementary sequences of the test sequence. Identity may be present in regions of at least about 50 amino acids or nucleotides in length, or in regions of 75-100 amino acids or nucleotides in length, or, if not specified, throughout the entire polynucleotide or polypeptide sequence. Polynucleotides encoding the polypeptides of the present invention (including homologues from species other than humans) can be obtained by a method comprising the steps of: screening a library using a labeled probe having a polynucleotide sequence or fragment thereof of the present invention under strict hybridization conditions, and isolating full-length cDNA and a genomic clone containing said polynucleotide sequence. Such hybridization techniques are well known to those skilled in the art.

[0118] If the amino acid sequence of a derivative or variant of a polypeptide shares at least 50% identity with the 100-amino acid sequence of the original peptide, the derivative or variant is considered to share "homology" with the peptide or is "homologous" to the peptide. In some embodiments, the derivative or variant shares at least 75% identity with the number of amino acid residues of the peptide or peptide fragment identical to the derivative. In some embodiments, the derivative or variant shares at least 85% identity with the number of amino acid residues of the peptide or peptide fragment identical to the derivative. In some embodiments, the amino acid sequence of the derivative shares at least 90% identity with the number of amino acid residues of the peptide or peptide fragment identical to the derivative. In some embodiments, the amino acid sequence of the derivative shares at least 95% identity with the number of amino acid residues of the peptide or peptide fragment identical to the derivative. In some embodiments, the derivative or variant shares at least 99% identity with the number of amino acid residues of the peptide or peptide fragment identical to the derivative.

[0119] As used herein, “isolated” polypeptide or construct means a construct or polypeptide that has been identified, isolated, and / or recovered from components of its native cell culture environment. Contaminating components of its native environment are substances that typically hinder the diagnostic or therapeutic use of heteropolymers and may include enzymes, hormones, and other proteinaceous or non-proteinaceous solutes.

[0120] In some embodiments, as used herein, the “isolated” antigen-binding polypeptide constructs described herein include heterodimer pairs or “isolated” heterodimer pairs, including heterodimers or heterodimer pairs identified and isolated and / or recovered from components of their native cell culture environment. Contaminating components of their native environment are substances that impede the diagnostic or therapeutic use of the heterodimers or antigen-binding polypeptide constructs and may include enzymes, hormones, and other proteinaceous or non-proteinaceous solutes.

[0121] Heterodimers and antigen-binding peptide constructs and heterodimers are typically purified to be substantially homogeneous. The phrases “substantially homogeneous,” “substantially homogeneous form,” and “substantially homogeneous” are used to indicate products substantially free of byproducts derived from undesirable peptide combinations (e.g., homodimers). In this context, the species of interest are heterodimers containing H1 and L1 (H1-L1) or H2 and L2 (H2-L2). Contaminants include heterodimers containing H1 and L2 (H1-L2) or H2 and L1 (H2-L1) or containing H1 and L1 or H2 and L2 (regardless of whether the Fab moieties are correctly paired or mismatched). In terms of purity, substantially homogeneous means that the amount of byproducts does not exceed 10%, for example, the total LC-MS intensity of all species present in the mixture is less than 5%, less than 1%, or less than 0.5%, where this percentage reflects the results of mass spectrometry analysis.

[0122] The phrase “selective (or specific) hybridization” means that when the sequence is present in a complex mixture (including but not limited to total cellular or library DNA or RNA), under strict hybridization conditions the molecule binds, pairs, or hybridizes only with a specific nucleotide sequence.

[0123] The terminology used by those skilled in the art of antibody technology has its own meaning as it is understood in the art, unless otherwise expressly defined herein. Antibodies are known to have variable regions, hinge regions, and constant structural domains. The structure and function of immunoglobulins are described, for example, in Harlow et al., Antibodies: A Laboratory Manual, Chapter 14 (Cold Spring Harbor Laboratory, Cold Spring Harbor, 1988).

[0124] As used herein, the terms “antibody” and “immunoglobulin” or “antigen-binding polypeptide construct” are used interchangeably. An “antigen-binding polypeptide construct” is a polypeptide encoded essentially by one or more immunoglobulin genes, or one or more segments thereof, that specifically binds to an analyte (antigen). Recognized immunoglobulin genes include the κ, λ, α, γ, δ, ε, and μ constant region genes, as well as numerous immunoglobulin variable region genes. Light chains are classified as κ or λ. Heavy chains are classified as γ, μ, α, δ, or ε, which in turn define immunoglobulin isotypes IgG, IgM, IgA, IgD, and IgE, respectively. Additionally, antibodies may belong to one of several subtypes; for example, IgG may belong to the IgG1, IgG2, IgG3, or IgG4 subclass.

[0125] An exemplary immunoglobulin (antibody) structural unit consists of two pairs of polypeptide chains, each pair having one "light" chain (approximately 25 kDa) and one "heavy" chain (approximately 50-70 kDa). The term "light chain" includes a full-length light chain and fragments thereof with sufficient variable region sequence to confer binding specificity. The full-length light chain includes a variable region domain (VL) and a constant region domain (CL). The variable region domain of the light chain is located at the amino terminus of the polypeptide. The light chain includes a κ chain and a λ chain. The term "heavy chain" includes a full-length heavy chain and fragments thereof with sufficient variable region sequence to confer binding specificity. The full-length heavy chain includes a variable region domain (VH) and three constant region domains (CH1, CH2, and CH3). The VH domain is located at the amino terminus of the polypeptide, and the CH domains are located at the carboxyl terminus, with CH3 being closest to the carboxyl terminus of the polypeptide. The heavy chain can belong to any isotype, including IgG (including IgG1, IgG2, IgG3, and IgG4 subclasses), IgA (including IgA1 and IgA2 subclasses), IgM, and IgE. The term "variable region" or "variable domain" refers to a portion of the antibody's light and / or heavy chains, typically responsible for antigen recognition, and usually comprises approximately 120 to 130 amino acids from the N-terminus of the heavy chain (VH) and approximately 100 to 110 N-terminal amino acids of the light chain (VL). The "complementarity-determining region" or "CDR" is an amino acid sequence that contributes to antigen-binding specificity and affinity. The "framework" region (FR) helps maintain the proper conformation of the CDR to facilitate binding between the antigen-binding region and the antigen. Structurally, the framework region may be located between CDRs of the antibody. Variable regions typically exhibit the same general formula structure of a relatively conserved framework region (FR) linked by three hypervariable CDRs. The CDRs of the two chains in each pair are typically aligned through the framework region, which allows binding to specific epitopes. From the N-terminus to the C-terminus, both the light and heavy chain variable regions typically contain FR1, CDR1, FR2, CDR2, FR3, CDR3, and FR4 domains. The amino acid assignment of each domain is generally according to the definition in the Kabat Sequences of Proteins of Immunological Interest (National Institutes of Health, Bethesda, Md. (1987 and 1991)) unless otherwise specified. In some embodiments, the antigen-binding peptide construct includes at least one immunoglobulin domain linked to a therapeutic peptide of IgG, IgM, IgA, IgD, or IgE. In some embodiments, the immunoglobulin domains included in the antigen-binding peptide constructs provided herein are derived from immunoglobulin-based constructs, such as biantibodies or monomers. In some embodiments, the antigen-binding peptide constructs described herein include at least one immunoglobulin domain derived from a heavy chain antibody, such as an alpaca antibody.In some embodiments, the antigen-binding polypeptide constructs provided herein comprise at least one immunoglobulin domain from mammalian antibodies such as bovine antibodies, human antibodies, alpaca antibodies (single-domain and non-single-domain), rodent antibodies, humanized antibodies, non-humanized antibodies, mouse antibodies, or any chimeric antibody. In some embodiments, the antigen-binding polypeptide constructs provided herein comprise at least one immunoglobulin domain from antibodies generated from a synthetic library.

[0126] "Bispecific," "dual-specific," or "bifunctional" antigen-binding proteins or antibodies are hybrid antigen-binding proteins with two distinct antigen-binding sites. Bispecific antigen-binding proteins and antibodies are a type of multispecific antigen-binding protein antibody. The two binding sites of a bispecific antigen-binding protein or antibody will bind to two different epitopes, which may be present on the same or different molecular targets. A "multispecific antigen-binding protein" or "multispecific antibody" is an antibody that targets more than one antigen or epitope. A "bivalent antigen-binding protein" or "bivalent antibody" contains two antigen-binding sites. In some cases, the two binding sites have the same antigen specificity. Bivalent antigen-binding proteins and bivalent antibodies can be bispecific, see below. In some embodiments, bivalent antibodies other than "multispecific" or "multifunctional" antibodies are generally understood to have the same binding site.

[0127] The term "preferred pairing" is used herein to describe the pairing of a first polypeptide with a second polypeptide, such as in antigen-binding polypeptide constructs and heterodimer pairs described herein, involving immunoglobulin heavy and light chains. Therefore, "preferred pairing" refers to the preferred pairing of a first polypeptide with a second polypeptide when pairing occurs simultaneously with one or more other distinct polypeptides. Preferred pairing typically results from modifications (e.g., amino acid modifications) of one or both of the first and second polypeptides. Generally, preferred pairing produces the most abundant dimer after pairing. It is known in the art that if an immunoglobulin heavy chain (H1) is co-expressed with two distinct immunoglobulin light chains (L1 and L2), it will pair with both light chains with statistically equal probability, producing an approximately 50:50 mixture of H1 / L1 pairings and H1 / L2 pairings. In this context, when H1 is co-expressed with both L1 and L2, if the amount of the H1-L1 heavy-light heterodimer is greater than the amount of the H1-L2 heterodimer, "preferred pairing" will occur, for example, between H1 and L1. Therefore, in this case, H1 preferentially pairs with L1 compared to L2.

[0128] However, in the context of generating wild-type bispecific antibodies from two starting antibody systems, it is also known in the art that, in some cases, there is an inherent preference when the light chain of one antibody system preferentially pairs with the heavy chains of both antibody systems. Therefore, when determining the design strength in the context of a bispecific antigen-binding construct, it may be necessary to evaluate the degree of pairing of the design compared to that in the wild-type system. Thus, in one embodiment, a design is considered to exhibit preferential pairing if the amount of the desired bispecific antibody is greater than the amount obtained in the wild-type system. In another embodiment, a design is considered to exhibit preferential pairing if the amount of pairing in the weaker arm of the antibody is greater than the amount present in the wild-type system.

[0129] The antibody heavy chain pairs with the antibody light chain and meets or contacts each other at one or more "interfaces". An "interface" comprises one or more "contacting" amino acid residues of a first polypeptide that interact with one or more "contacting" amino acid residues of a second polypeptide. For example, interfaces exist between CH3 polypeptide sequences with dimerized CH3 domains, between the CH1 domain of the heavy chain and the CL domain of the light chain, and between the VH domain of the heavy chain and the VL domain of the light chain. The "interface" may be derived from IgG antibodies, such as those derived from human IgG1 antibodies.

[0130] As used herein, the term “amino acid modification” includes, but is not limited to, amino acid mutations, insertions, deletions, substitutions, chemical modifications, physical modifications, and rearrangements.

[0131] Antigen-binding peptide constructs and heterodimer pairs

[0132] The antigen-binding polypeptide constructs described herein may comprise a first heterodimer and a second heterodimer; each heterodimer is obtained by pairing an immunoglobulin heavy chain with an immunoglobulin light chain. The structure and organization of the constant and variable domains of the immunoglobulin heavy and light chains are well known in the art. An immunoglobulin heavy chain typically comprises one variable (VH) domain and three constant domains CH1, CH2, and CH3. An immunoglobulin light chain typically comprises one variable (VL) domain and one constant (CL) domain. Various modifications can be made to these typical forms.

[0133] The antigen-binding polypeptide construct and heterodimer pair described herein may comprise a first heterodimer and a second heterodimer, each heterodimer comprising an immunoglobulin / antibody heavy chain or a fragment thereof having at least VH and CH1 domains, and an immunoglobulin / antibody light chain having VL and CL domains. In one embodiment, the two heterodimers of the heterodimer pair and the antigen-binding polypeptide construct comprise a full-length immunoglobulin heavy chain. In another embodiment, the two heterodimers of the heterodimer pair or the antigen-binding polypeptide construct comprise a fragment of an immunoglobulin heavy chain (including at least VH and CH1 domains). In one embodiment, the two heterodimers of the heterodimer pair comprise an N-terminal fragment of an immunoglobulin heavy chain (including at least VH and CH1 domains). In another embodiment, the two heterodimers of the heterodimer pair comprise a C-terminal fragment of an immunoglobulin heavy chain (including at least VH and CH1 domains).

[0134] Each heterodimer in a heterodimer pair can specifically bind to an antigen or epitope. In one embodiment, the immunoglobulin heavy chain and immunoglobulin light chain of each heterodimer are derived from or engineered by one or more modifications of a known antibody, such as a therapeutic antibody. A therapeutic antibody is an antibody used to treat a disease or condition in mammals that have or are prone to having a disease or condition. Suitable therapeutic antibodies from each heterodimer source include, but are not limited to, abagovomab, adalimumab, alemtuzumab, aurograb, bapineuzumab, basiliximab, belimumab, bevacizumab, briakinumab, canakinumab, catumaxomab, pegylated certolizumab, cetuximab, daclizumab, denosumab, efalizumab, galiximab, gemtuzumab ozogamicin, golimumab, and ibritumomab. Tiuxetan, infliximab, ipilimumab, lumiliximab, mepolizumab, motavizumab, muromonab, mycograb, natalizumab, nimotuzumab, ocrelizumab, ofatumumab, omalizumab, palivizumab, and more. Panitumumab, pertuzumab, ranibizumab, reslizumab, rituximab, teplizumab, tocilizumab / atlizumab, tositumomab, trastuzumab, Proxinium™, Rencarex™, ustekinumab, and zalutumumab.

[0135] In one embodiment, the immunoglobulin heavy chain and / or immunoglobulin light chain of each heterodimer are derived from or engineered by one or more modifications of the antibody binding the molecule, which includes, but is not limited to, the following protein list, and subunits, domains, motifs, and epitopes belonging to the following protein list: renin; growth hormone, including human growth hormone and bovine growth hormone; growth hormone-releasing factor; parathyroid hormone; thyroid-stimulating hormone; lipoproteins; α-1-antitrypsin; insulin A-chain; insulin B-chain; proinsulin; follicle-stimulating hormone; calcitonin; luteinizing hormone; glucagon; coagulation factors such as factor VII, factor VIIIC, factor IX, tissue factor (TF), and Winwib factor; anticoagulation factors such as protein C; atrial diuretic sodium factor; pulmonary surfactant; plasminogen activator, such as urokinase or human urine or tissue plasminogen activator (t-PA); bufotin; thrombin; hematopoietic growth factor; tumor necrosis factor-α and-β; enkephalin; RANTES (normal T cell expression and secretion factors regulated upon activation); human macrophage inflammatory protein (MIP-1-α); serum albumin such as human serum albumin; Müllerian duct inhibitory substance; relaxin A-chain; relaxin B-chain; pro-relaxin; mouse gonadotropin-related peptide; microbial proteins such as β-lactamase; DNase; IgE; cytotoxic T-lymphocyte-associated antigen (CTLA), such as CTLA-4; inhibin; activin Vascular endothelial growth factor (VEGF); receptors for hormones or growth factors, such as EGFR, VEGFR; interferons such as α-interferon (α-IFN), β-interferon (β-IFN), and γ-interferon (γ-IFN); protein A or D; rheumatoid factor; neurotrophic factors such as bone-derived neurotrophic factor (BDNF), neurotrophic protein-3, -4, -5, or -6 (NT-3, NT-4, NT-5, or NT-6) or nerve growth factor; platelet-derived growth factor (PDGF); fibroblast growth factors such as AFGF and PFGF; epidermal growth factor (EGF); transforming growth factor (TGF) such as TGF-α and TGF-β, including TGF-1, TG F-2, TGF-3, TGF-4, or TGF-5; insulin-like growth factor-I and-II (IGF-I and IGF-II); des(1-3)-IGF-I (brain IGF-I), insulin-like growth factor binding protein; CD proteins such as CD2, CD3, CD4, CD8, CD11a, CD14, CD18, CD19, CD20, CD22, CD23, CD25, CD33, CD34, CD40, CD40L, CD52, CD63, CD64, CD80, and CD147; erythropoietin; bone formation inducing factor; immunotoxins; bone morphogenetic protein (BMP); interferons such as interferon-α, -β, and -γ;Colony-stimulating factors (CSFs), such as M-CSF, GM-CSF, and G-CSF; interleukins (ILs), such as IL-1 to IL-13; TNF-α, superoxide dismutase; T-cell receptors; surface membrane proteins; decay accelerators; viral antigens, such as parts of the AIDS envelope, such as gp120; transport proteins; homing receptors; addressins; regulatory proteins; cell adhesion molecules such as LFA-1, Mac... 1. p150.95, VLA-4, ICAM-1, ICAM-3 and VCAM, a4 / p7 integrins and (Xv / p3 integrins including or subunits thereof, integrin α subunits such as CD49a, CD49b, CD49c, CD49d, CD49e, CD49f, α7, α8, α9, αD, CD11a, CD11b, CD51, CD11c, CD41, αIIb, αIELb; integrin β subunits such as CD29, CD 18. CD61, CD104, β5, β6, β7, and β8; integrin subunit combinations including but not limited to αVβ3, αVβ5, and α4β7; members of the apoptosis pathway; IgE; blood group antigens; flk2 / flt3 receptors; obesity (OB) receptors; mp1 receptors; CTLA-4; protein C; Eph receptors such as EphA2, EphA4, and EphB2; human leukocyte antigens (HLA) such as HLA-DR; complement proteins such as complement receptors CR1 and C1Rq and other complement factors such as C3 and C5; glycoprotein receptors such as GpIb.α, GPIIb / IIIa, and CD200; and fragments of any of the polypeptides listed above.

[0136] In one embodiment, the immunoglobulin heavy and / or light chains of each heterodimer are derived from or engineered by one or more modifications of an antibody that specifically binds to a cancer antigen, including but not limited to ALK receptor (pleiotropic growth factor receptor), pleiotropic growth factor, KS 1 / 4 pan-cancer antigen; ovarian cancer antigen (CA125); prostatic acid phosphate; prostate-specific antigen (PSA); melanoma-associated antigen p97; melanoma antigen gp75; high molecular weight melanoma antigen (HMW-MAA); prostate-specific membrane antigen; carcinoembryonic antigen (CEA); polymorphic epithelial mucin antigen; human milk fat globule antigen; and colorectal tumor-associated antigens such as CEA, TAG-72, C017-1A, and GICA. 19-9, CTA-1 and LEA; Burkitt lymphoma antigen-38.13; CD19; human B-lymphoma antigen-CD20; CD33; melanoma-specific antigens such as ganglioside GD2, ganglioside GD3, ganglioside GM2 and ganglioside GM3; tumor-specific transplantable cell surface antigen (TSTA); viral tumor antigens, including T-antigens of DNA tumor viruses and envelope antigens of RNA tumor viruses; carcinoembryonic antigen (CEA) and alpha-fetoprotein (AFP) such as CEA of the colon. 514 Carcinoembryonic trophoblast glycoproteins and bladder tumor carcinoembryonic antigens; differentiation antigens such as human lung cancer antigens L6 and L20; fibrosarcoma antigens; human leukemia T-cell antigen-Gp37; glycomimetic proteins; sphingolipids; breast cancer antigens such as EGFR (epidermal growth factor receptor); NY-BR-16; NY-BR-16 and HER2 antigen (p185HER2); polymorphic epithelial mucin (PEM); malignant human lymphocyte antigen-APO-1; differentiation antigens, such as I present in fetal erythrocytes. Antigens; Proendoderm I antigen present on adult erythrocytes; Preimplantation embryo; I (Ma) present in gastric adenocarcinoma; M18, M39 present in mammary epithelium; SSEA-1 present in bone marrow cells; VEP8; VEP9; Myl; Va4-D5; D156-22 present in colorectal cancer; TRA-1-85 (blood type H); SCP-1 present in testicular and ovarian cancer; C14 present in colon adenocarcinoma; F3 present in lung adenocarcinoma; AH6 present in gastric cancer; Y hapten; present Ley in embryonic carcinoma cells; TL5 (blood type A); EGF receptor present in A431 cells; E1 series present in pancreatic cancer (blood type B); FC10.2 present in embryonic carcinoma cells; gastric adenocarcinoma antigen; CO-514 present in adenocarcinoma (blood type Lea); NS-10 present in adenocarcinoma; CO-43 (blood type Leb); G49 EGF receptor present in A431 cells; MH2 present in colon adenocarcinoma (blood type ALeb / Ley); 19 present in colon cancer.9; gastric cancer mucin; T5A7 present in bone marrow cells; R24 present in melanoma; 4.2, GD3, D1.1, OFA-1, GM2, OFA-2, GD2 and M1:22:25:8 present in embryonic cancer cells; and SSEA-3 and SSEA-4 present in 4- to 8-cell embryos; cutaneous T-cell lymphoma antigen; MART-1 antigen; Sialy Tn (STn) antigen; colon cancer antigen NY-CO-45; lung cancer antigen NY-LU-12valiant A; Adenocarcinoma antigen ART1; Paraneoplastic brain-testis cancer antigens (tumor neuron antigen MA2; paraneoplastic neuronal antigen); Neural tumor ventral antigen 2 (NOVA2); Hepatocellular carcinoma antigen gene 520; Tumor-associated antigen CO-029; Tumor-associated antigens MAGE-C1 (cancer / testis antigen CT7), MAGE-B1 (MAGE-XP antigen), MAGE-B2 (DAM6), MAGE-2, MAGE-4-a, MAGE-4-b, and MAGE-X2; Cancer-testis antigen (NY-EOS-1); and fragments of any of the peptides listed above.

[0137] Human antibodies can be classified into isotypes, including IgG, IgA, IgE, IgM, and IgD. In one embodiment, Fc is derived from the IgG isotype. In another embodiment, Fc is derived from the IgA isotype. In yet another embodiment, Fc is derived from the IgE isotype. In yet another embodiment, Fc is derived from the IgM isotype. In yet another embodiment, Fc is derived from the IgD isotype.

[0138] Human IgG antibodies can also be divided into subclasses IgG1, IgG2, IgG3, and IgG4. Therefore, in some implementations, it is envisioned that the Fc can be derived from the IgG1, IgG2, IgG3, or IgG4 subclass of the antibody.

[0139] Each heterodimer in a heterodimer pair can specifically bind to an antigen or epitope. In one embodiment, each heterodimer in the heterodimer pair binds to the same epitope. In another embodiment, the first heterodimer in the heterodimer pair specifically binds to an epitope on an antigen, and the second heterodimer in the heterodimer pair specifically binds to a different epitope on the same antigen. In yet another embodiment, the first heterodimer in the heterodimer pair specifically binds to an epitope on a first antigen, and the second heterodimer in the heterodimer pair specifically binds to an epitope on a second antigen different from the first antigen. For example, in one embodiment, the first heterodimer specifically binds to tissue factor, while the second heterodimer specifically binds to antigen Her2 (ErbB2), and vice versa. In an alternative embodiment, the first heterodimer specifically binds to tissue factor, while the second heterodimer specifically binds to EGFR, and vice versa. In another embodiment, the first heterodimer specifically binds to EGFR, while the second heterodimer specifically binds to antigen Her2, and vice versa. In another embodiment, the first heterodimer specifically binds to the aforementioned molecule or cancer antigen. In another embodiment, the second heterodimer specifically binds to the aforementioned molecule or cancer antigen.

[0140] As described above, in some embodiments, the immunoglobulin heavy chain and immunoglobulin light chain of each heterodimer are derived from or engineered from one or more modifications of a known therapeutic antibody or an antibody that binds to various target molecules or cancer antigens. The amino acid and nucleotide sequences of many such molecules are readily available (see, for example, GenBank: AJ308087.1 (humanized anti-human tissue factor antibody D3H44 light chain variable region and CL domain); GenBank: AJ308086.1 (humanized anti-human tissue factor antibody D3H44 heavy chain variable region and CH1 domain); GenBank: HC359025.1 (pertuzumab Fab light chain gene module); GenBank: HC359024.1 (pertuzumab Fab). The sequences described in this article are from the NCBI website and were last updated on November 28, 2012. Each sequence is incorporated herein by reference in its entirety for all purposes. (Heavy chain gene module); GenBank: GM685465.1 (Trastuzumab (=Herceptin) - Wild type; Light chain); GenBank: GM685463.1 (Trastuzumab (=Herceptin) - Wild type; Heavy chain); GenBank: GM685466.1 (Trastuzumab (=Herceptin) - GC-optimized light chain); and GenBank: GM685464.1 (Trastuzumab (=Herceptin) - GC-optimized heavy chain). The amino acid and nucleotide sequences of cetuximab are also known in the art; see, for example, the Drug Bank website, accession number DB00002, supported by the Canadian Institutes of Health Research, Alberta Innovates-Health Solutions and The Metabolomics Innovation Centre (TMIC).

[0141] In some aspects, the isolated antigen-binding polypeptide construct comprises an amino acid sequence having at least 80%, 85%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, 99%, or 100% identity with the amino acid sequence or fragment thereof shown in the tables or registries disclosed herein. In some aspects, the isolated antigen-binding polypeptide construct comprises an amino acid sequence encoded by a polynucleotide having at least 80%, 85%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, 99%, or 100% identity with the nucleotide sequence or fragment thereof shown in the tables or registries disclosed herein.

[0142] Amino acid modifications of immunoglobulin heavy and light chains

[0143] At least one of the heterodimers in the heterodimer pair may contain one or more amino acid modifications of the immunoglobulin heavy chain and / or immunoglobulin light chain, such that the heavy chain of the first heterodimer preferentially pairs with one of the light chains rather than the other. Similarly, the heavy chain of the second heterodimer may preferentially pair with the second light chain rather than the first light chain. This preferential pairing of one heavy chain with one of the two light chains may be based on a design set comprising one immunoglobulin heavy chain and two immunoglobulin light chains (referred to as the LCCA design set), wherein when the immunoglobulin heavy chain is co-expressed with two immunoglobulin light chains, the immunoglobulin heavy chain preferentially pairs with one of the two immunoglobulin light chains rather than the other. Therefore, the LCCA design set may comprise one immunoglobulin heavy chain, a first immunoglobulin light chain, and a second immunoglobulin light chain.

[0144] In one embodiment, one or more amino acid modifications include one or more amino acid substitutions.

[0145] In one embodiment, the preferred pairings illustrated in the LCCA design set are established by modifying one or more amino acids that form part of the interface between the light and heavy chains. In another embodiment, the preferred pairings illustrated in the LCCA design set are established by modifying one or more amino acids of at least one of the CH1 domain of the immunoglobulin heavy chain, the CL domain of the first immunoglobulin light chain, and the CL domain of the second immunoglobulin light chain.

[0146] In one embodiment, one or more amino acid modifications are restricted to conserved framework residues of variable (VH, VL) and constant (CH1, CL) domains, as indicated by the Kabat numbering of the residues. For example, Almagro [Frontiers In Bioscience (2008) 13: 1619-1633] provides the definition of framework residues according to the Kabat, Chotia, and IMGT numbering schemes.

[0147] In one embodiment, at least one of the heterodimers comprises one or more mutations that introduce complementary immunoglobulin heavy and light chains. Complementarity at the heavy and light chain interface can be achieved based on spatial and hydrophobic contacts, electrostatic / charge interactions, or a combination of multiple interactions. Complementarity between protein surfaces has been extensively described in the literature as lock-and-key, button-and-loop, protrusions and grooves, donor and acceptor, etc., all of which suggest a structural and chemical matching nature between two interacting surfaces. In one embodiment, at least one of the heterodimers comprises one or more mutations that introduce mutations in the immunoglobulin heavy and light chains, introducing novel hydrogen bonds between the light and heavy chains at the interface. In one embodiment, at least one of the heterodimers comprises one or more mutations that introduce mutations in the immunoglobulin heavy and light chains, introducing novel salt bridges between the light and heavy chains at the interface.

[0148] Non-limiting examples of suitable LCCA design sets are described in the embodiments, tables, and figures. In one embodiment, the LCCA design set comprises an immunoglobulin heavy chain having at least one amino acid modification in the CH1 domain, a first immunoglobulin light chain having at least one amino acid modification in the CL domain, and a second immunoglobulin light chain without any amino acid modification in the CL domain. In another embodiment, the LCCA design set comprises an immunoglobulin heavy chain having at least one amino acid modification in the CH1 domain, a first immunoglobulin light chain having at least one amino acid modification in the CL domain, and a second immunoglobulin light chain having at least one amino acid modification in the CL domain. In yet another embodiment, the LCCA design set comprises an immunoglobulin heavy chain having at least one amino acid modification in the CH1 domain, a first immunoglobulin light chain having at least two amino acid modifications in the CL domain, and a second immunoglobulin light chain having at least two amino acid modifications in the CL domain. In yet another embodiment, the LCCA design set comprises an immunoglobulin heavy chain having at least one amino acid modification in the CH1 domain, a first immunoglobulin light chain having at least two amino acid modifications in the CL domain, and a second immunoglobulin light chain having at least one amino acid modification in the CL domain.

[0149] In one embodiment, the LCCA design set includes an immunoglobulin heavy chain without amino acid modifications in the CH1 domain, a first immunoglobulin light chain without amino acid modifications in the CL domain, and a second immunoglobulin light chain with at least one amino acid modification in the CL domain. In another embodiment, the LCCA design set includes an immunoglobulin heavy chain without amino acid modifications in the CH1 domain, a first immunoglobulin light chain without amino acid modifications in the CL domain, and a second immunoglobulin light chain with at least two amino acid modifications in the CL domain.

[0150] In one embodiment, the LCCA design set includes an immunoglobulin heavy chain having at least two amino acid modifications in the CH1 domain, a first immunoglobulin light chain without amino acid modifications in the CL domain, and a second immunoglobulin light chain having at least one amino acid modification in the CL domain. In another embodiment, the LCCA design set includes an immunoglobulin heavy chain having at least two amino acid modifications in the CH1 domain, a first immunoglobulin light chain having at least one amino acid modification in the CL domain, and a second immunoglobulin light chain having at least two amino acid modifications in the CL domain. In yet another embodiment, the LCCA design set includes an immunoglobulin heavy chain having at least two amino acid modifications in the CH1 domain, a first immunoglobulin light chain having at least two amino acid modifications in the CL domain, and a second immunoglobulin light chain having at least two amino acid modifications in the CL domain. In one embodiment, the LCCA design set includes an immunoglobulin heavy chain having at least two amino acid modifications in the CH1 domain, a first immunoglobulin light chain having at least three amino acid modifications in the CL domain, and a second immunoglobulin light chain having at least two amino acid modifications in the CL domain.

[0151] In one embodiment, the LCCA design set includes an immunoglobulin heavy chain having at least three amino acid modifications in the CH1 domain, a first immunoglobulin light chain without amino acid modifications in the CL domain, and a second immunoglobulin light chain having at least one amino acid modification in the CL domain. In another embodiment, the LCCA design set includes an immunoglobulin heavy chain having at least three amino acid modifications in the CH1 domain, a first immunoglobulin light chain having at least one amino acid modification in the CL domain, and a second immunoglobulin light chain having at least one amino acid modification in the CL domain. In yet another embodiment, the LCCA design set includes an immunoglobulin heavy chain having at least three amino acid modifications in the CH1 domain, a first immunoglobulin light chain having at least three amino acid modifications in the CL domain, and a second immunoglobulin light chain having at least two amino acid modifications in the CL domain. In yet another embodiment, the LCCA design set includes an immunoglobulin heavy chain having at least three amino acid modifications in the CH1 domain, a first immunoglobulin light chain having at least four amino acid modifications in the CL domain, and a second immunoglobulin light chain having at least three amino acid modifications in the CL domain. In one implementation, the preferred pairings shown in the LCCA design set are established by modifying one or more amino acids of at least one of the VH domain of the immunoglobulin heavy chain, the VL domain of the first immunoglobulin light chain, and the VL domain of the second immunoglobulin light chain. Non-limiting examples of suitable LCCA design sets are shown in the following tables and examples.

[0152] In one embodiment, the LCCA design set includes an immunoglobulin heavy chain without amino acid modifications in the VH domain, a first immunoglobulin light chain without amino acid modifications in the VL domain, and a second immunoglobulin light chain with at least one amino acid modification in the VL domain. In another embodiment, the LCCA design set includes an immunoglobulin heavy chain without amino acid modifications in the VH domain, a first immunoglobulin light chain without amino acid modifications in the VL domain, and a second immunoglobulin light chain with at least two amino acid modifications in the VL domain.

[0153] In one embodiment, the LCCA design set includes an immunoglobulin heavy chain having at least one amino acid modification in the VH domain, a first immunoglobulin light chain without amino acid modification in the VL domain, and a second immunoglobulin light chain having at least one amino acid modification in the VL domain. In another embodiment, the LCCA design set includes an immunoglobulin heavy chain having at least one amino acid modification in the VH domain, a first immunoglobulin light chain having at least one amino acid modification in the VL domain, and a second immunoglobulin light chain having at least one amino acid modification in the VL domain. In yet another embodiment, the LCCA design set includes an immunoglobulin heavy chain having at least one amino acid modification in the VH domain, a first immunoglobulin light chain having at least two amino acid modifications in the VL domain, and a second immunoglobulin light chain having at least two amino acid modifications in the VL domain.

[0154] In one embodiment, the LCCA design set includes an immunoglobulin heavy chain having at least two amino acid modifications in its VH domain, a first immunoglobulin light chain without amino acid modifications in its VL domain, and a second immunoglobulin light chain having at least one amino acid modification in its VL domain. In another embodiment, the LCCA design set includes an immunoglobulin heavy chain having at least two amino acid modifications in its VH domain, a first immunoglobulin light chain having at least two amino acid modifications in its VL domain, and a second immunoglobulin light chain having at least one amino acid modification in its VL domain. In yet another embodiment, the LCCA design set includes an immunoglobulin heavy chain having at least two amino acid modifications in its VH domain, a first immunoglobulin light chain having at least one amino acid modification in its VL domain, and a second immunoglobulin light chain having at least one amino acid modification in its VL domain.

[0155] In one embodiment, the LCCA design set can be combined to provide a combination comprising two distinct immunoglobulin heavy chains (H1 and H2) and two distinct immunoglobulin light chains (L1 and L2), wherein when H1, H2, L1, and L2 are co-expressed, H1 preferentially pairs with L1, and H2 preferentially pairs with L2. In some embodiments, the amino acid modifications described herein are within the context of a bispecific antibody construct. For example, the design set described herein can be incorporated into a full-length immunoglobulin heavy chain such that the full-length heavy chain preferentially pairs with the immunoglobulin light chain. In some embodiments, the full-length immunoglobulin heavy chain contains amino acid modifications that facilitate dimerization in the Fc region, as described in the examples.

[0156] The transferability of specific amino acid modifications described in this article to other antibodies:

[0157] While the specific amino acid modifications of the immunoglobulin heavy and light chains described above in conjunction with the D3H44 anti-tissue factor extracellular domain antibodies trastuzumab and cetuximab immunoglobulin heavy and light chains are described below, for the purposes of this document, it is envisioned and illustrated (see Examples, Figures and Tables) that these amino acid modifications can be transferred to other immunoglobulin heavy and light chains to produce a similar manner in which one immunoglobulin heavy chain preferentially pairs with one of two immunoglobulin light chains.

[0158] The VH:VL and CH1:CL interface residues at the interface between the heavy and light chains of immunoglobulins are relatively conserved (Padlan et al., 1986, Mol. Immunol. 23(9): 951-960). This sequence conservation is a result of evolutionary constraints, increasing the likelihood of forming functionally activated antibody-binding domains during combined pairing of light and heavy chains. This sequence conservation allows sequence modifications (driving preferential pairing) in the specific instance of D3H44 described above to be transferred to other heavy and light chain heterodimers, yielding approximately equivalent results in terms of preferential pairing, as this region exhibits high sequence conservation in antibodies; additionally, when sequence differences occur, these differences are typically located distal to the CH1:CL interface. This is particularly true for the CH1 and CL domains. However, some sequence variations exist at the antigen-binding site in terms of CDR (complementarity-determining region) loop residues (and length), particularly CDR-H3. Therefore, in one embodiment, the heterodimer pair described herein comprises a heterodimer wherein, when the amino acid sequence of the antigen-binding site is significantly different from that of the D3H44 antibody, at least one heterodimer comprises one or more amino acid modifications in the VH and / or VL domains located distal to the CDR ring. In another embodiment, the heterodimer pair described herein comprises a heterodimer wherein, when the amino acid sequence of the antigen-binding site is substantially identical to that of the D3H44 antibody, at least one heterodimer comprises one or more amino acid modifications in the VH and / or VL domains located proximal to or distal to the CDR ring.

[0159] In one embodiment, the amino acid modifications described herein can be transferred to the immunoglobulin heavy and light chains of the antibody, based on human or humanized IgG1 / κ. Non-limiting examples of such IgG1 / κ chains include oflambumab (for humans) or trastuzumab, pertuzumab or bevacizumab (for humanized).

[0160] In another embodiment, the amino acid modifications described herein can be transferred to the immunoglobulin heavy and light chains of the antibody using commonly used VH and VL subsets. Non-limiting examples of such antibodies include pertuzumab.

[0161] In one embodiment, the amino acid modifications described herein can be transferred to the heavy and light chains of immunoglobulins of antibodies with a framework close to the germline. Examples of such antibodies include obinutuzumab.

[0162] In one embodiment, the amino acid modifications described herein can be transferred to the immunoglobulin heavy and light chains of an antibody where the VH:VL domain angle is close to the average value of the observed heavy and light chain pairs. Examples of this type of antibody include, but are not limited to, pertuzumab. In another embodiment, the amino acid modifications described herein can be transferred to the immunoglobulin heavy and light chains of an antibody having typical CL and CH1 domains. Suitable examples of such antibodies include, but are not limited to, trastuzumab.

[0163] In some implementations, a subset of the amino acid modifications described herein are used for variable domains in the antigen-binding constructs provided above.

[0164] Examples, figures, and tables show that amino acid modifications (e.g., within one or more Fab fragments containing variable and constant regions) can be transferred to other immunoglobulin heavy and light chains, resulting in a similar manner in which one immunoglobulin heavy chain preferentially pairs with one of two immunoglobulin light chains.

[0165] Priority pairing

[0166] As described above, at least one heterodimer of the antigen-binding construct / heterodimer pair described herein may contain one or more amino acid modifications to its immunoglobulin heavy chain and / or immunoglobulin light chain, such that the heavy chain of one heterodimer, for example H1, preferentially pairs with one light chain, for example L1, rather than another light chain, L2, and the heavy chain of the other heterodimer, H2, preferentially pairs with light chain L2 rather than light chain L1. In other words, the desired preferred pairing is considered to be between H1 and L1 and between H2 and L2. Preferred pairing is considered to occur, for example, between H1 and L1, when H1 is combined with L1 and L2, if the yield of the H1-L1 heterodimer is greater than the yield of the mismatched H1-L2 heterodimer. Similarly, relative to the respective pairings of the corresponding H1-L1 pairs with H2-L2 pairs that do not contain one or more amino acid modifications, when H2 is combined with L1 and L2, if the yield of the H2-L2 heterodimer is greater than the yield of the mismatched H2-L1 heterodimer, then preferential pairing is considered to have occurred between H2 and L2. In this context, heterodimers containing H1 and L1 (H1-L1) or H2 and L2 (H2-L2) are referred to herein as preferentially paired, correctly paired, specifically paired, or desired heterodimers, while heterodimers containing H1 and L2 (H1-L2) or H2 and L1 (H2-L1) are referred to herein as mismatched heterodimers. The mutation set corresponding to the two heavy chains and the two light chains implies selective pairing of H1-L1 and H2-L2; this mutation set is called the design set.

[0167] Therefore, in one embodiment, when one immunoglobulin heavy chain of the heterodimer is co-expressed with two immunoglobulin light chains, the required relative yield of the heterodimer is greater than 55%. In another embodiment, when one immunoglobulin heavy chain of the heterodimer is co-expressed with two immunoglobulin light chains, the required relative yield of the heterodimer is greater than 60%. In another embodiment, when one immunoglobulin heavy chain of the heterodimer is co-expressed with two immunoglobulin light chains, the required relative yield of the heterodimer is greater than 70%. In another embodiment, when one immunoglobulin heavy chain of the heterodimer is co-expressed with two immunoglobulin light chains, the required relative yield of the heterodimer is greater than 80%. In another embodiment, when one immunoglobulin heavy chain of the heterodimer is co-expressed with two immunoglobulin light chains, the required relative yield of the heterodimer is greater than 90%. In yet another embodiment, when one immunoglobulin heavy chain of the heterodimer is co-expressed with two immunoglobulin light chains, the required relative yield of the heterodimer is greater than 95%.

[0168] In the above examples, when H1 is co-expressed with L1 and L2, if the required amount of H1-L1 heterodimer is greater than the amount of mismatched H1-L2 heterodimer, then preferential pairing is considered to occur between H1 and L1. Similarly, when H2 is co-expressed with L1 and L2, if the required amount of H2-L2 heterodimer is greater than the amount of mismatched H2-L2 heterodimer, then preferential pairing is considered to occur between H2 and L2. Therefore, in one embodiment, when one immunoglobulin heavy chain of the heterodimer is co-expressed with two immunoglobulin light chains, the ratio of the required heterodimer to the mismatched heterodimer is greater than 1.25:1. In one embodiment, when one immunoglobulin heavy chain of the heterodimer is co-expressed with two immunoglobulin light chains, the ratio of the required heterodimer to the mismatched heterodimer is greater than 1.5:1. In another embodiment, when one immunoglobulin heavy chain of the heterodimer is co-expressed with two immunoglobulin light chains, the required ratio of heterodimer to mismatched heterodimer is greater than 2:1. In another embodiment, when one immunoglobulin heavy chain of the heterodimer is co-expressed with two immunoglobulin light chains, the required ratio of heterodimer to mismatched heterodimer is greater than 3:1. In another embodiment, when one immunoglobulin heavy chain of the heterodimer is co-expressed with two immunoglobulin light chains, the required ratio of heterodimer to mismatched heterodimer is greater than 5:1. In another embodiment, when one immunoglobulin heavy chain of the heterodimer is co-expressed with two immunoglobulin light chains, the required ratio of heterodimer to mismatched heterodimer is greater than 10:1. In another embodiment, when one immunoglobulin heavy chain of the heterodimer is co-expressed with two immunoglobulin light chains, the required ratio of heterodimer to mismatched heterodimer is greater than 25:1. In another implementation, when one immunoglobulin heavy chain of the heterodimer is co-expressed with two immunoglobulin light chains, the required ratio of heterodimer to mismatched heterodimer is greater than 50:1.

[0169] In some embodiments, the heterodimers described herein preferentially pair to form bispecific antibodies. In some embodiments, the construct comprises heterodimers that preferentially pair to form bispecific antibodies selected from D3H44 / trastuzumab, D3H44 / cetuximab, and trastuzumab / cetuximab. In some embodiments, the bispecific antibody comprises the amino acid modifications described in Tables 28a-28c.

[0170] In some embodiments, two full-length heavy chain constructs are co-expressed with two distinct light chain constructs to obtain ten possible antibody species: H1-L1:H1-L1, H1-L2:H1-L2, H1-L1:H1-L2, H2-L1:H2-L1, H2-L2:H2-L2, H2-L1:H2-L2, H1-L1:H2-L1, H1-L2:H2-L2, H1-L2:H2-L1, and H1-L1:H2-L2. The H1-L1:H2-L2 species is considered the correctly paired bispecific antibody species. In some embodiments, the DNA ratio is selected to obtain the maximum number of correctly paired bispecific antibody species. For example, in some embodiments, the H1:H2:L1:L2 ratio is 15:15:53:17. In some implementations, the ratio of H1:H2:L1:L2 is 15:15:17:53.

[0171] In some embodiments, the percentage of correctly paired bispecific species relative to all species is at least 40%, 45%, 50%, 55%, 60%, 65%, 70%, 75%, 80%, 85%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, 99%, or 100% (see, for example, Tables 29a-29c and 30a-30c). In some embodiments, the percentage of correctly paired bispecific species is greater than the percentage of correctly paired bispecific species obtained by co-expression of wild-type H1, H2, L1, and L2 without the amino acid modifications described in Tables 28a-28c. In some implementations, the percentage of correctly paired bispecific species is reduced by at least 5%, 10%, 15%, 20%, 30%, 40%, 50%, 60%, 70%, or 75% compared to the percentage of correctly paired bispecific species obtained by co-expression of wild-type H1, H2, L1, and L2 without the amino acid modifications described in Tables 28a-28c.

[0172] Thermal stability of heterodimers

[0173] In addition to facilitating preferential pairing, the selection of amino acid substitutions ensures that mutations do not destabilize the Fab heterodimer. Therefore, in most cases, the stability measurements of the Fab heterodimer are very close to those of wild-type Fab (e.g., within 3°C of wild-type Fab).

[0174] Therefore, in some embodiments, each heterodimer of the heterodimer pair described herein has a thermal stability equivalent to that of a heterodimer comprising the same immunoglobulin heavy and light chains but without amino acid modifications containing the CH1, CL, VH, or VL domains described herein. In one embodiment, thermal stability is determined by melting temperature or Tm. Therefore, in one embodiment, the thermal stability of the heterodimer described herein is within about 10°C of a heterodimer comprising the same immunoglobulin heavy and light chains but without amino acid modifications containing the CH1, CL, VH, or VL domains described herein. Therefore, in one embodiment, the thermal stability of the heterodimer described herein is within about 5°C of a heterodimer comprising the same immunoglobulin heavy and light chains but without amino acid modifications containing the CH1, CL, VH, or VL domains described herein. In another embodiment, the thermal stability of the heterodimer described herein is within about 3°C ​​of a heterodimer comprising the same immunoglobulin heavy and light chains but without amino acid modifications containing the CH1, CL, VH, or VL domains described herein. In another embodiment, the thermostability of the heterodimer described herein is within about 2°C of a heterodimer containing the same immunoglobulin heavy and light chains but without amino acid modifications of the CH1, CL, VH, or VL domains described herein. In another embodiment, the thermostability of the heterodimer described herein is within about 1.5°C of a heterodimer containing the same immunoglobulin heavy and light chains but without amino acid modifications of the CH1, CL, VH, or VL domains described herein. In another embodiment, the thermostability of the heterodimer described herein is within about 1°C of a heterodimer containing the same immunoglobulin heavy and light chains but without amino acid modifications of the CH1, CL, VH, or VL domains described herein. In another embodiment, the thermostability of the heterodimer described herein is within about 0.5°C of a heterodimer containing the same immunoglobulin heavy and light chains but without amino acid modifications of the CH1, CL, VH, or VL domains described herein. In another embodiment, the thermostability of the heterodimer described herein is within about 0.25°C of the heterodimer containing the same immunoglobulin heavy and light chains but without amino acid modifications of the CH1, CL, VH, or VL domains described herein.

[0175] Furthermore, in some embodiments, the thermal stability of the heterodimers described herein is unexpectedly improved (i.e., increased) compared to heterodimers containing the same immunoglobulin heavy and light chains but without amino acid modifications of the CH1, CL, VH, or VL domains described herein. Therefore, in one embodiment, the thermal stability of the heterodimer described herein is increased by approximately 0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9, 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, 3.0, 3.1, 3.2, 3.3, 3.4, 3.5, 3.6, 3.7, 3.8, 3.9, 4.0, 4.5, 5.0 °C or more compared to heterodimers containing the same immunoglobulin heavy and light chains but without the amino acid modifications of the CH1, CL, VH, or VL domains described herein.

[0176] Affinity of heterodimers to antigens

[0177] In one embodiment, each heterodimer of the heterodimer pair has the same or equivalent affinity for its respective antigen as a heterodimer comprising the same immunoglobulin heavy and light chains but without amino acid modifications of the CH1, CL, VH, or VL domains as described herein. In one embodiment, the affinity of the heterodimer of the heterodimer pair for its respective antigen is within about 50 times that of a heterodimer comprising the same immunoglobulin heavy and light chains but without amino acid modifications of the CH1, CL, VH, or VL domains as described herein. In one embodiment, the affinity of the heterodimer of the heterodimer pair for its respective antigen is within about 25 times that of a heterodimer comprising the same immunoglobulin heavy and light chains but without amino acid modifications of the CH1, CL, VH, or VL domains as described herein. In one embodiment, the heterodimers of the heterodimer pair exhibit an affinity for their respective antigens that is within about 10 times that of heterodimers containing the same immunoglobulin heavy and light chains but without amino acid modifications of the CH1, CL, VH, or VL domains as described herein. In another embodiment, the heterodimers of the heterodimer pair exhibit an affinity for their respective antigens that is within about 5 times that of heterodimers containing the same immunoglobulin heavy and light chains but without amino acid modifications of the CH1, CL, VH, or VL domains as described herein. In yet another embodiment, the heterodimers of the heterodimer pair exhibit an affinity for their respective antigens that is within about 2.5 times that of heterodimers containing the same immunoglobulin heavy and light chains but without amino acid modifications of the CH1, CL, VH, or VL domains as described herein. In another embodiment, the heterodimer of the heterodimer pair has an affinity for its respective antigen that is approximately twice that of a heterodimer containing the same immunoglobulin heavy and light chains but without amino acid modifications of the CH1, CL, VH, or VL domains as described herein. In another embodiment, the heterodimer of the heterodimer pair has an affinity for its respective antigen that is approximately 1.5 times that of a heterodimer containing the same immunoglobulin heavy and light chains but without amino acid modifications of the CH1, CL, VH, or VL domains as described herein. In another embodiment, the heterodimer of the heterodimer pair has an affinity for its respective antigen that is approximately the same as that of a heterodimer containing the same immunoglobulin heavy and light chains but without amino acid modifications of the CH1, CL, VH, or VL domains as described herein.

[0178] Other optional modifications

[0179] In one embodiment, the immunoglobulin heavy and light chains of the heterodimer pair described herein may be further modified (i.e., through covalent linkage of various types of molecules) such that the covalent linkage does not impede the preferential pairing between the heavy and light chains, or affect the heterodimer's ability to bind its antigen, or affect its stability. Such modifications include, for example, but not limited to, glycosylation, acetylation, polyethylene glycolation, phosphorylation, amidation, derivatization via known protecting / capping groups, protease cleavage, linkage to cellular ligands or other proteins, etc. Any variety of chemical modifications can be performed using known techniques, including but not limited to specific chemical cleavage, acetylation, formylation, and the metabolic synthesis of tunicamycin.

[0180] In another embodiment, the immunoglobulin heavy and light chains of the heterodimer pair described herein may be conjugated (directly or indirectly) to a therapeutic or pharmaceutical moiety that modifies a given biological response. The therapeutic or pharmaceutical moiety should not be construed as limited to classical chemotherapeutic agents. For example, the pharmaceutical moiety may be a protein or peptide having the desired biological activity. Such proteins may include, for example, toxins such as abrinogen, ricin A, onconase (or another cytotoxic RNase), Pseudomonas exotoxin, cholera toxin, or diphtheria toxin; proteins such as tumor necrosis factor, alpha-interferon, beta-interferon, nerve growth factor, platelet-derived growth factor, tissue plasminogen activator, apoptosis agents such as TNF-α, TNF-β, AIM I (see International Patent Publication No. WO 97 / 33899), AIM II (see International Patent Publication No. WO 97 / 34911), Fas ligand (Takahashi et al., 1994, J. Immunol., 6:1567), and VEGI (see International Patent Publication No. WO 97 / 34911). 99 / 23105); thrombogenic agents or antithrombogenic agents such as angiostatin or endostatin; or biological response modulators such as lymphokines (e.g., interleukin-1 (“IL-1”), interleukin-2 (“IL-2”), interleukin-6 (“IL-6”), granulocyte-macrophage colony-stimulating factor (“GM-CSF”) and granulocyte colony-stimulating factor (“G-CSF”)) or growth factors (e.g., growth hormone (“GH”)).

[0181] In alternative embodiments, the antibody may be conjugated to the therapeutic portion, such as a radioactive material or a macrocyclic chelating agent for conjugating radioactive metal ions (see examples of radioactive materials above). In some embodiments, the macrocyclic chelating agent is 1,4,7,10-tetraazacyclododecane-N,N′,N″,N″-tetraacetic acid (DOTA), which can be linked to the antibody via a linker molecule. Such linker molecules are well known in the art and are described in Denardo et al., 1998, Clin Cancer Res. 4: 2483; Peterson et al., 1999, Bioconjug. Chem. 10: 553; and Zimmerman et al., 1999, Nucl. Med. Biol. 26: 943.

[0182] In some embodiments, the heterodimeric immunoglobulin heavy and light chains are expressed as a fusion protein containing a tag that facilitates purification and / or testing. As described herein, a “tag” is any added amino acid series that may be present at the C-terminus, N-terminus, or interior of a protein to aid in protein identification or purification. Suitable tags include, but are not limited to, tags known to those skilled in the art for purification and / or testing, such as albumin-binding domains (ABD), His tags, FLAG tags, glutathione S-transferases, hemagglutinin (HA), and maltose-binding proteins. Such tagged proteins may also be engineered to include cleavage sites, such as thrombin, enterokinase, or factor X cleavage sites, to facilitate tag removal before, during, or after purification.

[0183] In some embodiments, one or more cysteine ​​residues at the bottom of the Fab domain of the light chain (position 214, Kabat number) and the heavy chain (position 233, Kabat number) that form interchain disulfide bonds may be modified to serine or alanine or noncysteine ​​or different amino acids.

[0184] It is conceivable that additional amino acid modifications could be made to the immunoglobulin heavy chain to improve the preferential pairing level and / or the thermal stability of heterodimer pairs. For example, additional amino acid modifications could be made to the Fc domain of the immunoglobulin heavy chain to drive preferential pairing between heterodimer pairs and homodimer pairs. Such amino acid modifications are known in the art and include, for example, those described in U.S. Patent Publication No. 2012 / 0149876. Alternatively, alternative strategies for driving preferential pairing between heterodimer pairs and homodimer pairs could be utilized, such as, for example, "button-in," charged residues with ion-interacting properties, and chain exchange engineered domain (SEED) technology. The latter strategies are described in the art and reviewed in Klein et al., ibid. Further discussion of the Fc domain follows.

[0185] Fc structural domain

[0186] In embodiments where the antigen-binding polypeptide construct comprises a full-length immunoglobulin heavy chain, the construct will contain an Fc. In some aspects, the Fc contains at least one or two CH3 domain sequences. In some aspects, when the antigen-binding polypeptide construct comprises a heterodimer (containing only the Fab region of the heavy chain), the Fc is linked to a first heterodimer and / or a second heterodimer, with or without one or more linkers. In some aspects, the Fc is a human Fc. In some aspects, the Fc is a human IgG or IgG1 Fc. In some aspects, the Fc is a heterodimeric Fc. In some aspects, the Fc contains at least one or two CH2 domain sequences.

[0187] In some respects, the Fc contains C H3 Modifications of at least one or more of the domain sequences. In some respects, the Fc contains C H2 Modification of at least one or more of the domain sequences. In some respects, Fc is a single peptide. In other respects, Fc is multiple peptides, such as two peptides.

[0188] In some respects, the Fc contains C H3 One or more modifications of at least one of the sequences. In some respects, the Fc contains C H2 One or more modifications in at least one of the sequences. In some respects, Fc is a single peptide. In other respects, Fc is multiple peptides, such as two peptides.

[0189] In some respects, Fc is the Fc described in patent application PCT / CA2011 / 001238 filed on November 4, 2011, or PCT / CA2012 / 050780 filed on November 2, 2012, the entire disclosure of each patent application of which is incorporated herein by reference in its entirety for all purposes.

[0190] In some respects, the constructs described herein comprise a heterodimer Fc containing an asymmetrically modified CH3 domain. The heterodimer Fc may contain two heavy-chain constant domain polypeptides: a first heavy-chain polypeptide and a second heavy-chain polypeptide, which are interchangeable, provided that Fc contains one first heavy-chain polypeptide and one second heavy-chain polypeptide. Generally, the first heavy-chain polypeptide contains a first CH3 sequence, and the second heavy-chain polypeptide contains a second CH3 sequence.

[0191] When two CH3 sequences dimerize, those sequences typically produce a heterodimer Fc rather than a homodimer, where one or more amino acid modifications are introduced asymmetrically. As used herein, “asymmetric amino acid modification” refers to any modification in which the amino acid at a specific position on the first CH3 sequence differs from the amino acid at the same position on the second CH3 sequence, and the first and second CH3 sequences preferentially pair to form a heterodimer rather than a homodimer. This heterodimerization can be the result of modifying only one of the two amino acids at the same corresponding amino acid position on each sequence; or the result of modifying both amino acids at the same corresponding position on each of the first and second CH3 sequences. The first and second CH3 sequences of the heterodimer Fc may contain one or more asymmetric amino acid modifications.

[0192] Table X provides the amino acid sequences of the human IgG1 Fc sequence corresponding to amino acids 231 to 447 of the full-length human IgG1 heavy chain. The CH3 sequence contains amino acids 341-447 of the full-length human IgG1 heavy chain.

[0193] Typically, the Fc may comprise two adjacent heavy chain sequences (A and B) capable of dimerization. In some aspects, one or both sequences of the Fc include one or more mutations or modifications at the following positions: L351, F405, Y407, T366, K392, T394, T350, S400, and / or N390 (using EU numbers). In some aspects, the Fc includes the mutant sequences shown in Table X. In some aspects, the Fc includes mutations of variant 1A-B. In some aspects, the Fc includes mutations of variant 2A-B. In some aspects, the Fc includes mutations of variant 3A-B. In some aspects, the Fc includes mutations of variant 4A-B. In some aspects, the Fc includes mutations of variant 5A-B.

[0194]

[0195] The first and second CH3 sequences may contain the amino acid mutations described herein for binding the full-length human IgG1 heavy chain, amino acids 231 to 447. In one embodiment, the heterodimer Fc comprises a modified CH3 domain, wherein the first CH3 sequence has amino acid modifications at positions F405 and Y407, and the second CH3 sequence has an amino acid modification at position T394. In one embodiment, the heterodimer Fc comprises a modified CH3 domain, wherein the first CH3 sequence has one or more amino acid modifications selected from L351Y, F405A, and Y407V, and the second CH3 sequence has one or more amino acid modifications selected from T366L, T366I, K392L, K392M, and T394W.

[0196] In one embodiment, the heterodimer Fc comprises a modified CH3 domain, wherein a first CH3 sequence has amino acid modifications at positions L351, F405, and Y407, a second CH3 sequence has amino acid modifications at positions T366, K392, and T394, one of the first or second CH3 sequences further comprises an amino acid modification at position Q347, and the other CH3 sequences further comprise an amino acid modification at position K360. In another embodiment, the heterodimer Fc comprises a modified CH3 domain, wherein a first CH3 sequence has amino acid modifications at positions L351, F405, and Y407, a second CH3 sequence has amino acid modifications at positions T366, K392, and T394, one of the first or second CH3 sequences further comprises an amino acid modification at position Q347, the other CH3 sequences further comprise an amino acid modification at position K360, and one or both of the CH3 sequences further comprise an amino acid modification T350V.

[0197] In one embodiment, the heterodimer Fc comprises a modified CH3 domain, wherein a first CH3 sequence has amino acid modifications at positions L351, F405, and Y407, a second CH3 sequence has amino acid modifications at positions T366, K392, and T394, one of the first and second CH3 sequences further comprises amino acid modifications D399R or D399K, and other CH3 sequences comprise one or more of T411E, T411D, K409E, K409D, K392E, and K392D. In another embodiment, the heterodimer Fc comprises a modified CH3 domain, wherein the first CH3 sequence has amino acid modifications at positions L351, F405, and Y407, the second CH3 sequence has amino acid modifications at positions T366, K392, and T394, one of the first and second CH3 sequences further comprises the amino acid modification D399R or D399K, and the other CH3 sequences comprise one or more of T411E, T411D, K409E, K409D, K392E, and K392D, and one or both of the CH3 sequences further comprise the amino acid modification T350V.

[0198] In one embodiment, the heterodimer Fc comprises a modified CH3 domain, wherein a first CH3 sequence has amino acid modifications at positions L351, F405, and Y407, and a second CH3 sequence has amino acid modifications at positions T366, K392, and T394, wherein one or both of the CH3 sequences further comprise the amino acid modification T350V.

[0199] In one embodiment, the heterodimer Fc includes a modified CH3 domain comprising the following amino acid modifications, wherein "A" represents an amino acid modification of a first CH3 sequence and "B" represents an amino acid modification of a second CH3 sequence: A: L351Y_F405A_Y407V, B: T366L_K392M_T394W, A: L351Y_F405A_Y407V, B: T366L_K392L_T394W, A: T35 0V_L351Y_F405A_Y407V, B: T350V_T366L_K392L_T394W, A: T350V_L351Y_F405A_Y407V, B: T350V_T 366L_K392M_T394W, A: T350V_L351Y_S400E_F405A_Y407V and / or B: T350V_T366L_N390R_K392M_T394W.

[0200] One or more asymmetric amino acid modifications may facilitate the formation of a heterodimeric Fc domain, wherein the heterodimeric CH3 domain has stability equivalent to that of the wild-type homodimeric CH3 domain. In one embodiment, one or more asymmetric amino acid modifications facilitate the formation of a heterodimeric Fc domain, wherein the heterodimeric Fc domain has stability equivalent to that of the wild-type homodimeric Fc domain. In one embodiment, one or more asymmetric amino acid modifications facilitate the formation of a heterodimeric Fc domain, wherein the heterodimeric Fc domain exhibits stability observed in differential scanning calorimetry studies by melting temperature (Tm), and wherein the melting temperature is within 4 °C of the observed corresponding symmetric wild-type homodimeric Fc domain. In some aspects, Fc contains C H3 One or more modifications in at least one of the sequences contribute to the formation of a heterodimer Fc with stability equivalent to that of the wild-type homodimer Fc.

[0201] In one embodiment, the stability of the CH3 domain can be assessed, for example, by determining the melting temperature of the CH3 domain using differential scanning calorimetry (DSC). Therefore, in another embodiment, the CH3 domain has a melting temperature of about 68°C or higher. In another embodiment, the CH3 domain has a melting temperature of about 70°C or higher. In another embodiment, the CH3 domain has a melting temperature of about 72°C or higher. In another embodiment, the CH3 domain has a melting temperature of about 73°C or higher. In another embodiment, the CH3 domain has a melting temperature of about 75°C or higher. In yet another embodiment, the CH3 domain has a melting temperature of about 78°C or higher. In some aspects, dimerized C H3The sequence has a melting temperature (Tm) of approximately 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 77.5, 78, 79, 80, 81, 82, 83, 84 or 85 °C or higher.

[0202] In some embodiments, the heterodimer Fc containing the modified CH3 sequence can be formed with at least about 75% purity compared to the homodimer Fc in the expression product. In another embodiment, the heterodimer Fc is formed with a purity greater than about 80%. In another embodiment, the heterodimer Fc is formed with a purity greater than about 85%. In another embodiment, the heterodimer Fc is formed with a purity greater than about 90%. In another embodiment, the heterodimer Fc is formed with a purity greater than about 95%. In another embodiment, the heterodimer Fc is formed with a purity greater than about 97%. In some respects, Fc is expressed as a heterodimer with a purity greater than about 75%, 76%, 77%, 78%, 79%, 80%, 81%, 82%, 83%, 84%, 85%, 86%, 87%, 88%, 89%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, or 99%. In other respects, Fc is expressed by a single cell as a heterodimer with a purity greater than about 75%, 76%, 77%, 78%, 79%, 80%, 81%, 82%, 83%, 84%, 85%, 86%, 87%, 88%, 89%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, or 99%.

[0203] Another method for modifying monomeric Fc peptides to facilitate the formation of heterodimer Fc is described in International Patent Publication No. WO 96 / 027011 (Newt), Gunasekaran et al. (Gunasekaran K. et al. (2010) J Biol Chem. 285, 19637-46, electrostatic design to achieve selective heterodimerization), Davis et al. (Davis, JH. et al. (2010) Prot Eng Des Sel; 23(4): 195-202, strand exchange engineered domain (SEED) technology), and Labrijn et al. [Efficient generation of stable bispecific IgG1 by controlled Fab-arm exchange. Labrijn AF, Meesters JI, de Goeij BE, van den Bremer ET, Neijssen J, van Kampen MD, Strumane K, Verploegen S, Kundu A, Gramer MJ, van Berkel PH, van de Winkel As described in [JG, Schuurman J, Parren PW. Proc Natl Acad Sci US A. 2013 Mar 26; 110(13): 5145-50]. In some embodiments, the isolated constructs described herein include antigen-binding constructs that bind antigens; and dimer Fc peptide constructs that, relative to antigen-binding constructs that do not include the same Fc peptide, have superior biophysical properties such as stability and are easy to prepare. Many mutations in the heavy chain sequence of Fc are known in the art to selectively alter the affinity of antibody Fc for different Fcγ receptors. In some aspects, the Fc contains one or more modifications that contribute to the selective binding of Fc-γ receptors.

[0204] The CH2 domain is amino acids 231-340 of the sequence shown in Table X. Exemplary mutations are as follows:

[0205] S298A / E333A / K334A, S298A / E333A / K334A / K326A (Lu Y, Verne s JM, Chiang N et al., J Immunol Methods., 28 February 2011, 365(1-2): 132-41); F243L / R292P / Y300L / V305I / P396L, F243L / R292P / Y300L / L235V / P396L (Stavenhagen JB, Gorlatov S, Tuaillon N et al., Cancer Res., 15 September 2007, 67(18): 8882-90; Nordstrom JL, Gorlatov S, Zhang W et al., Breast Cancer Res., November 30, 2011, 13(6): R123); F243L (Stewart R, Thom G, Levens M et al., Protein Eng Des S el., September 2011, 24(9): 671-8.), S298A / E333A / K334A (Shields RL, Namenuk AK, Hong K et al., J Biol Chem., March 2, 2001, 276(9): 6591-604); S239D / I332E / A330L, S239D / I332E (Lazar GA, Dang W, Karki S et al., Proc Natl Acad Sci US A., March 14, 2006, 103(11):4005-10); S239D / S267E, S267E / L328F (Chu SY, Vostiar I, KarkiS et al., Mol Immunol., September 2008, 45(15): 3926-33); S239D / D265S / S298A / I332E, S239E / S298A / K 326A / A327H, G237F / S298A / A330L / I332E, S239D / I332E / S298A, S239D / K326E / A330L / The mutations listed in I332E / S298A, G236A / S239D / D270L / I332E, S239E / S267E / H268D, L234F / S267E / N325L, G237F / V266L / S267D, and other mutations listed in WO2011 / 120134 and WO2011 / 120135 are incorporated herein by reference.The mutations are listed on page 283 of Therapeutic Antibody Engineering (by William R. Strohland and Lila M. Strohl, Woodhead Publishing series in Biomedicine No. 11, ISBN 1907568 37 9, October 2012).

[0206] In some embodiments, the CH2 domain contains one or more asymmetric amino acid modifications. In some embodiments, the CH2 domain contains one or more asymmetric amino acid modifications that facilitate the selective binding of FcγR. In some embodiments, the CH2 domain allows for the isolation and purification of the isolated constructs described herein.

[0207] FcRn combined with PK parameters

[0208] As is known in the art, binding to FcRn enables endocytosed antibodies to recirculate from the endosome back into the bloodstream (Raghavan et al., 1996, Annu Rev Cell Dev Biol 12: 181-220; Ghetie et al., 2000, Annu Rev Immunol 18: 739-766). This process, combined with renal filtration (due to the large molecular weight of full-length molecules), results in a favorable antibody serum half-life ranging from one to three weeks. Fc binding to FcRn also plays a crucial role in antibody transport. Therefore, in one embodiment, the construct of the present invention is capable of binding to FcRn.

[0209] Further modifications to improve effector function.

[0210] In some implementations, the constructs described herein can be modified to improve their effector functions. Such modifications are known in the art and include de-algylation or engineering the Fc moiety of the antibody to enhance its affinity for activated receptors, primarily FCGR3a for ADCC and C1q for CDC. Table Y below summarizes various designs reported in the literature for engineered effector functions.

[0211]

[0212] Therefore, in one embodiment, the construct described herein may include a dimer Fc comprising one or more amino acid modifications, said amino acid modifications as described in the table above, conferring improved effector function. In another embodiment, the construct may be de-algylated to improve effector function.

[0213] connector

[0214] The constructs described herein may include one or more heterodimers operatively coupled to the Fc described herein. In some aspects, the Fc is coupled to the one or more heterodimers, with or without one or more connectors. In some aspects, the Fc is directly coupled to the one or more heterodimers. In some aspects, the Fc is coupled to the one or more heterodimers via one or more connectors. In some aspects, the Fc is coupled to the heavy chain of each heterodimer via a connector.

[0215] In some aspects, the one or more adapters are one or more peptide adapters. In some aspects, the one or more adapters include one or more antibody hinge regions. In some aspects, the one or more adapters include one or more IgG1 hinge regions.

[0216] Preparation method of heterodimer pairs

[0217] As described above, the heterodimer pair described herein may comprise a first heterodimer and a second heterodimer, each heterodimer comprising an immunoglobulin heavy chain or fragment thereof having at least VH and CH1 domains, and an immunoglobulin light chain having VL and CL domains. The immunoglobulin heavy chain and immunoglobulin light chain of the heterodimer can be prepared using recombinant DNA techniques known in the art. Standard techniques such as those described in, for example, Sambrook and Russell, *Molecular Cloning: A Laboratory Manual* (Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY, 3rd ed., 2001); Sambrook et al., *Molecular Cloning: A Laboratory Manual* (Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY, 2nd ed., 1989); *Short Protocols in Molecular Biology* (Ausubel et al., John Wiley and Sons, New York, 4th ed., 1999); and Glick and Pasternak, *Molecular Biotechnology: Principles and Applications of Recombinant DNA* (ASM Press, Washington, DC, 2nd ed., 1998), can be used for recombinant nucleic acid methods, nucleic acid synthesis, cell culture, transgenic incorporation, and recombinant protein expression. Alternatively, the heterodimers and heterodimer pairs described herein can be chemically synthesized.

[0218] The nucleic acid and amino acid sequences of the immunoglobulin heavy and light chains of the antibodies that derive heterodimers are known in the art or can be determined using readily available nucleic acid and / or protein sequencing methods. Methods for fusing the tag genes described herein into the immunoglobulin heavy and / or light chains are known in the art, and some of these methods are described below and in the examples.

[0219] For example, methods for expressing and co-expressing immunoglobulin heavy and light chains in host cells are well known in the art. Furthermore, methods for labeling heavy and / or light chains using recombinant DNA technology are also well known in the art. Expression vectors and host cells suitable for heavy and light chain expression are also well known in the art, as described below.

[0220] Bispecific antibody preparation methods that do not rely solely on the use of monoclonal or transient cell lines expressing all four chains are known in the art (Gramer et al., (2013) mAbs 5, 962; Strop et al., (2012) J Mol Biol 420, 204). These methods rely solely on post-generative arm exchange under redox conditions involving two pairs of light and heavy chains involved in bispecific antibody formation (redox generation). In this method, the H1:L1 and H2:L2 pairs can be expressed in two different cell lines to independently generate two Fab arms. Subsequently, the two Fab arms are mixed under selected redox conditions to achieve the reassociation of the two distinct heavy chains H1 and H2 to form a bispecific antibody comprising the L1:H1:H2:L2 chain. It is conceivable to prepare bispecific antibodies using the design libraries / datasets described herein using redox generation methods or modified forms of these methods.

[0221] In some embodiments, a cell-free protein expression system is used to co-express peptides (e.g., heavy and light chain peptides) without the use of living cells. Instead, all components required for DNA transcription into RNA and RNA translation into protein (e.g., ribosomes, tRNA, enzymes, cofactors, amino acids) are provided in solution for in vitro use. In some embodiments, in vitro expression requires (1) a genetic template (mRNA or DNA) encoding heavy and light chain peptides and (2) a reaction solution containing the necessary transcription and translation molecular machinery. In some embodiments, cell extracts essentially provide the components of the reaction solution, such as: RNA polymerase for mRNA transcription, ribosomes for peptide translation, tRNA, amino acids, enzyme cofactors, energy sources, and key cellular components for proper protein folding. Cell-free protein expression systems can be prepared using lysates derived from bacterial cells, yeast cells, insect cells, plant cells, mammalian cells, human cells, or combinations thereof. Such cell lysates provide the correct composition and proportion of enzymes and structural units required for translation. In some embodiments, the cell membrane is removed, leaving only the cytosol and organelle components of the cell.

[0222] Many cell-free protein expression systems are known in the art, as described by Carlson et al., (2012) Biotechnol. Adv. 30: 1185-1194. For example, cell-free protein expression systems can be derived from prokaryotic or eukaryotic cells. Examples of prokaryotic cell-free expression systems include those derived from *Escherichia coli* (E. coli). Eukaryotic cell-free protein expression systems can be derived from extracts such as rabbit reticulocytes, wheat germ, and insect cells. Such prokaryotic and eukaryotic cell-free protein expression systems are commercially available from companies such as Roche, Invitrogen, Qiagen, and Novagen. Those skilled in the art can readily select a suitable cell-free protein expression system that can generate peptides (e.g., heavy and light chain peptides) that can pair with each other. Additionally, cell-free protein expression systems can be supplemented with molecular chaperones (e.g., BiP) and isomerases (e.g., disulfide isomerases) to improve the efficiency of IgG folding.

[0223] In some implementations, cell-free expression systems are used to co-express heavy and light chain polypeptides derived from a DNA template (transcription and translation) or an mRNA template (translation only).

[0224] Vector and host cell

[0225] Recombinant expression of heavy and light chains requires the construction of expression vectors containing polynucleotides encoding either the heavy or light chain (e.g., antibodies or fusion proteins). Once the polynucleotide encoding the heavy or light chain is obtained, the vectors used to generate the heavy or light chain can be produced using recombinant DNA techniques well known in the art. Therefore, this document describes methods for preparing proteins by expressing polynucleotides containing nucleotide sequences encoding either the heavy or light chain. Methods well known to those skilled in the art can be used to construct expression vectors containing heavy or light chain coding sequences and appropriate transcriptional and translational control signals. These methods include, for example, in vitro recombinant DNA techniques, synthetic techniques, and in vivo genetic recombination. Therefore, the present invention provides reproducible vectors containing nucleotide sequences encoding either the heavy or light chain operatively linked to a promoter.

[0226] The expression vector is transferred into host cells using conventional techniques, and the transfected cells are then cultured using conventional techniques to generate the modified heavy or light chain used in the method of this invention. In a specific embodiment, the heavy and light chains used in this method are co-expressed in host cells to express the whole immunoglobulin molecule, as detailed below.

[0227] Many host expression vector systems can be used to express modified heavy and light chains. Such host expression systems represent tools for generating and subsequently purifying the coding sequences of interest, and also represent cells that can express modified heavy and light chains in situ upon transformation or transfection with appropriate nucleotide coding sequences. These host expression systems include, but are not limited to, microorganisms, such as bacteria transformed with recombinant phage DNA, plasmid DNA, or coliform DNA expression vectors containing modified heavy and light chain coding sequences (e.g., *Escherichia coli* and *Bacillus subtilis*); yeast transformed with recombinant yeast expression vectors containing modified heavy and light chain coding sequences (e.g., *Saccharomyces* and *Pichia*); insect cell systems transfected with recombinant viral expression vectors containing modified heavy and light chain coding sequences (e.g., baculoviruses); and cells transfected with recombinant viral expression vectors containing modified heavy and light chain coding sequences. The expression vectors may contain recombinant viral expression vectors (e.g., cauliflower mosaic virus, CaMV; tobacco mosaic virus, TMV) or plant cell systems transformed with recombinant plasmid expression vectors containing modified heavy and light chain coding sequences (e.g., Ti plasmid); or mammalian cell systems (e.g., COS, CHO, BHK, HEK-293, NSO, and 3T3 cells) containing promoters derived from mammalian cell genomes (e.g., metallothionein promoters) or promoters derived from mammalian viruses (e.g., adenovirus late promoters; vaccinia virus 7.5K promoters). In some embodiments, bacterial cells such as *Escherichia coli* or eukaryotic cells are used to express modified heavy and light chains, which are recombinant antibody or fusion protein molecules. For example, mammalian cells (such as Chinese hamster ovary cells (CHO)) and binding vectors (such as major immediate early gene promoter elements derived from human cell cytomegalovirus) are efficient antibody expression systems (Foecking et al., 1986, Gene 45:101; and Cockett et al., 1990, Bio / Technology 8:2). In specific embodiments, the expression of the nucleotide sequences encoding the immunoglobulin heavy and light chains of each heterodimer is regulated by constitutive promoters, inducible promoters, or tissue-specific promoters.

[0228] In mammalian host cells, numerous virus-based expression systems are available. In cases where adenovirus is used as the expression vector, the modified heavy and light chain coding sequences of interest can be linked to the adenoviral transcription / translation control complex, such as a late promoter and a triple leader sequence. This chimeric gene can then be inserted into the adenoviral genome via in vitro or in vivo recombination. Insertion into non-essential regions of the viral genome (e.g., regions E1 or E3) will produce a viable recombinant virus capable of expressing the modified heavy and light chains in an infected host (e.g., see Logan & Shenk, 1984, Proc. Natl. Acad. Sci. USA 81:355-359). Specific initiation signals can also be necessary for the efficient translation of the inserted antibody-coding sequence. These signals include the ATG start codon and adjacent sequences. Furthermore, the start codon must be in phase with the reading frame of the desired coding sequence to ensure the translation of the entire inserted sequence. These exogenous translation control signals and start codons can have many sources, both natural and synthetic. The efficiency of expression can be improved by including appropriate transcriptional enhancer elements, transcription terminators, etc. (see, for example, Bittner et al., 1987, Methods in Enzymol. 153: 516-544).

[0229] The expression of the heavy and light chains of heterodimer immunoglobulins can be controlled by any promoter or enhancer element known in the art. Promoters that can be used to control the expression of genes encoding modified heavy and light chains (e.g., antibodies or fusion proteins) include, but are not limited to, the early promoter region of SV40 (Bernoist and Chambon, 1981, Nature 290: 304-310), promoters contained in the 3′ long terminal repeat sequence of Laureus sarcoma virus (Yamamoto et al., 1980, Cell 22: 787-797), the herpes thymidine kinase promoter (Wagner et al., 1981, Proc. Natl. Acad. Sci. USA 78. 1441-1445), regulatory sequences of metallothionein genes (Brinster et al., 1982, Nature 296: 39-42), and the tetracycline (Tet) promoter (Gossen et al., 1995, Proc. Nat. Acad. Sci. USA). 89:5547-5551); prokaryotic expression vectors such as β-lactamase promoters (Villa-Kamaroff et al., 1978, Proc. Natl. Acad. Sci. USA 75:3727-3731) or tac promoters (DeBoer et al., 1983, Proc. Natl. Acad. Sci. USA 80:21-25; see also "Useful proteins from recombinant bacteria" in Scientific American, 1980, 242:74-94); plant expression vectors containing the alkaloid synthase promoter region (Herrera-Estrella et al., Nature 303:209-213) or cauliflower mosaic virus 35S RNA promoters (Gardner et al., 1981, Nucl. Acids Res. 9: 2871) and the promoter of the photosynthetic enzyme ribulose-1,6-bisphosphate carboxylase (Herrera-Estrella et al., 1984, Nature 310: 115-120); promoter elements of yeast or other fungi such as Gal4 promoter, ADC (alcohol dehydrogenase) promoter, PGK (glycerol phosphate kinase) promoter, alkaline phosphatase promoter, and the following tissue-specific animal transcriptional control regions used in transgenic animals: the control region of the elastase I gene active in pancreatic acinar cells (Swift et al., 1984, Cell 38: 639-646; Ornitz et al., 1986, Cold Spring Harbor Symp. Quant. Biol.).50:399-409; MacDonald, 1987, Hepatology 7:425-515); Insulin gene control region active in pancreatic β cells (Hanahan, 1985, Nature 315:115-122); Immunoglobulin gene control region active in lymphocytes (Grosschedl et al., 1984, Cell 38:647-658; Adames et al., 1985, Nature 318:533-538; Alexander et al., 1987, Mol. Cell. Biol. 7:1436-1444); Mouse mammary tumor virus control region active in testes, mammary glands, lymphocytes, and mast cells (Leder et al., 1986, Cell 45:485-495); Albumin gene control region active in the liver (Pinkert et al., 1987, Genes and Devel. 1: 268-276), the active α-fetoglobulin gene control region in the liver (Krumlauf et al., 1985, Mol. Cell. Biol. 5: 1639-1648; Hammer et al., 1987, Science 235: 53-58); the active α-1-antitrypsin gene control region in the liver (Kelsey et al., 1987, Genes and Devel. 1: 161-171); the active β-globulin gene control region in bone marrow cells (Mogram et al., 1985, Nature 315: 338-340; Kollias et al., 1986, Cell 46: 89-94); and the active myelin basic protein gene control region in oligodendrocytes in the brain (Readhead et al., 1987, Cell 235: 53-58). 48:703-712); Myosin light chain-2 gene control region active in skeletal muscle (Sani, 1985, Nature 314:283-286); Neuron-specific enolase (NSE) active in neuronal cells (Morelli et al., 1999, Gen.Virol. 80:571-83); Brain-derived neurotrophic factor (BDNF) gene control region active in neuronal cells (Tabuchi et al., 1998, Biochem. Biophysic. Res. Com. 253:818-823); Glial fibrillary acidic protein (GFAP) promoter active in astrocytes (Gomes et al., 1999, Braz J Med Biol Res 32(5):619-631; Morelli et al., 1999, Gen.Virol.80:571-83) and the region controlling the active gonadotropin-releasing hormone gene in the hypothalamus (Mason et al., 1986, Science 234:1372-1378).

[0230] Furthermore, host cell lines can be selectively regulated to modulate the expression of the inserted sequence or to modify and process the gene product in a desired specific manner. Expression from certain promoters can be enhanced in the presence of certain inducers; thus, the expression of genetically engineered fusion proteins can be controlled. Moreover, different host cells possess characteristics and specific mechanisms for translation and post-translational processing and modification (e.g., protein glycosylation, phosphorylation). Appropriate cell lines or host systems can be selected to ensure the desired modification and processing of the expressed exogenous protein. For example, expression in a bacterial system will produce an unglycosylated product, while expression in yeast will produce a glycosylated product. Eukaryotic host cells with appropriate cellular machinery for the processing (e.g., glycosylation and phosphorylation) of the primary transcript of the gene product can be used.These mammalian host cells include, but are not limited to, CHO, VERY, BHK, HeLa, COS, MDCK, HEK-293, 3T3, WI38, and NSO, especially neuronal cell lines such as SK-N-AS, SK-N-FI, SK-N-DZ human neuroblastoma (Sugimoto et al., 1984, J. Natl. Cancer Inst. 73: 51-57), SK-N-SH human neuroblastoma (Biochim. Biophys. Acta, 1982, 704: 450-460), Daoy human cerebellar medulloblastoma (He et al., 1992, Cancer Res. 52: 1144-1148), DBTRG-05MG glioblastoma cells (Kruse et al., 1992, In Vitro Cell. Dev. Biol. 28A: 609-614), and IMR-32 human neuroblastoma (Cancer Res., 1970, 30: 2110-2118), 1321N1 human astrocytoma (Proc. Natl. Acad. Sci. USA, 1977, 74: 4816), MOG-G-CCM human astrocytoma (Br. J. Cancer, 1984, 49: 269), U87MG human glioblastoma-astrocytoma (Acta Pathol. Microbiol. Scand., 1968, 74: 465-486), A172 human glioblastoma (Olopade et al., 1992, Cancer Res. 52: 2523-2529), C6 rat glioma cells (Benda et al., 1968, Science Neuroblastoma in mice (Proc. Natl. Acad. Sci. USA, 1970, 65: 129-136), neuroblastoma in NB41A3 mice (Proc. Natl. Acad. Sci. USA, 1962, 48: 1184-1190), choroid plexus in sheep (Bolin et al., 1994, J. Virol. Methods 48: 211-221), normal astrocytes in cats (G355-5, PG-4, Haapala et al., 1985, J. Virol. 53: 827-833), ferret brain (Mpf, Trowbridge et al., 1982, In Vitro 18: 952-960), and normal cell lines such as CTX. Normal cerebral cortex of TNA2 rats (Radany et al., 1992, Proc. Natl. Acad. Sci. USA 89: 6467-6471), such as CRL7030 and Hs578Bst.Furthermore, different vector / host expression systems can affect the processing response to varying degrees.

[0231] For the long-term, high-yield production of recombinant proteins, stable expression is generally preferred. For example, cell lines can be engineered to stably express the modified heavy and light chains (e.g., antibodies or fusion proteins) of the present invention. Host cells can be transformed with DNA controlled by appropriate expression control elements (e.g., promoters, enhancers, sequences, transcription terminators, polyadenylation sites, etc.) and selective markers, without using expression vectors containing viral origins of replication. After the introduction of exogenous DNA, engineered cells are allowed to grow in enrichment media for 1–2 days, then switched to selection media. Selective markers in the recombinant plasmid confer selective resistance, allowing cells to stably integrate the plasmid into the chromosome and grow to form a focal point, which can then be cloned and expanded into cell lines.

[0232] Many selection systems can be used, including but not limited to herpes simplex virus thymidine kinase (Wigler et al., 1977, Cell 11:223), hypoxanthine-guanine phosphoribosyltransferase (Szybalska & Szybalski, 1962, Proc. Natl. Acad. Sci. USA 48:2026), and adenine phosphoribosyltransferase (Lowy et al., 1980, Cell 22:817) genes that can be used for tk-, hgprt-, or aprt- cells, respectively. In addition, antimetabolite resistance can be used as the basis for gene selection for dhfr (conferring methotrexate resistance) (Wigler et al., 1980, Natl. Acad. Sci. USA 77: 3567; O′Hare et al., 1981, Proc. Natl. Acad. Sci. USA 78: 1527); gpt (conferring mycophenolic acid resistance) (Mulligan & Berg, 1981, Proc. Natl. Acad. Sci. USA 78: 2072); neo (conferring aminoglycoside G-418 resistance) (Colberre-Garapin et al., 1981, J. Mol. Biol. 150: 1); and hygro (conferring hygromycin resistance) (Santerre et al., 1984, Gene 30: 147).

[0233] Co-expression of heavy and light chains

[0234] The heterodimer pairs of immunoglobulin heavy and light chains described herein can be co-expressed in mammalian cells, as described above. In one embodiment, one heavy chain is co-expressed with two distinct light chains using the aforementioned LCCA design set, wherein the heavy chain preferentially pairs with one of the two light chains. In another embodiment, two distinct heavy chains are co-expressed with two distinct light chains, wherein each heavy chain preferentially pairs with one light chain.

[0235] Testing of heterodimer pairs

[0236] As described above, at least one heterodimer of the heterodimer pair described herein may include one or more amino acid modifications to its immunoglobulin heavy chain and / or immunoglobulin light chain, such that when the two distinct heavy chains and two distinct light chains of the heterodimer pair are co-expressed in mammalian cells, the heavy chain of the first heterodimer preferentially pairs with one light chain rather than the other. Similarly, the heavy chain of the second heterodimer preferentially pairs with the second light chain rather than the first light chain. The degree of preferential pairing can be assessed, for example, using the methods described below. The affinity of each heterodimer of the heterodimer pair for its respective antigen can be tested as described below. The thermal stability of each heterodimer of the heterodimer pair can also be tested as described below.

[0237] Methods for determining preferred pairing

[0238] LCCA

[0239] In one embodiment, the preferential pairing between immunoglobulin heavy and light chains is determined by performing light chain competition analysis (LCCA). Various embodiments of LCCA are described in co-owned patent application PCT / US2013 / 063306, filed October 3, 2013, which is incorporated herein by reference in its entirety for all purposes. This method allows for quantitative analysis of heavy chain pairing with specific light chains in a mixture of co-expressed proteins and can be used to determine whether a particular immunoglobulin heavy chain selectively associates with one of two immunoglobulin light chains when both are co-expressed. The method is briefly described as follows: at least one heavy chain and two distinct light chains are co-expressed in a cell such that the heavy chain is the limiting ratio of the pairing reactants; optionally, secreted proteins are isolated from the cells; immunoglobulin light chain polypeptides that bind the heavy chain are isolated from the remaining secreted proteins to generate isolated heavy chain pairing fractions; the amount of each distinct light chain in the isolated heavy chain fractions is determined; and the relative amount of each distinct light chain in the isolated heavy chain fractions is analyzed to determine the ability of at least one heavy chain to selectively pair with one light chain.

[0240] This method offers reasonable throughput and is stable (i.e., insensitive to small changes in operation, such as user or flow rate) and accurate. It provides sensitive analysis that can detect small changes in protein sequences. Mixed protein-protein, domain-domain, and chain-chain interactions with large surface areas often require multiple mutations (replacements) to introduce selectivity. Protein products do not require separation and purification, allowing for more efficient screening. Further details regarding embodiments of this method are described in the examples.

[0241] Alternative assays with preferred pairing

[0242] Alternative methods for detecting preferred pairings include quantifying the relative heterodimer populations comprising each light chain using LC-MS (liquid chromatography-mass spectrometry), identifying each distinct species using differences in their molecular weights. Antigen activity assays can also be used to quantify the relative heterodimer populations comprising each light chain, and the binding extent determined thereby (relative to a control) can be used to estimate the respective heterodimer populations.

[0243] Other methods, such as SMCA, are described in the embodiments, figures, and tables.

[0244] thermal stability

[0245] The thermal stability of heterodimers can be determined according to methods known in the art. The melting temperature of each heterodimer is a characterization of its thermal stability. The melting temperature of heterodimers can be determined using techniques such as differential scanning calorimetry (Chen et al., (2003) Pharm Res 20: 1952-60; Ghirlando et al., (1999) Immunol Lett 68: 47-52). Alternatively, the thermal stability of heterodimers can be determined using circular dichroism (Murray et al., (2002) J. Chromatogr Sci 40: 343-9).

[0246] Affinity of antigens

[0247] The binding affinity of the heterodimer to its respective antigen and the dissociation rate of the interaction can be determined by competitive binding assays according to methods well known in the art. An example of a competitive binding assay is radioimmunoassay, which involves incubating a labeled antigen (e.g., 3H or 125I with the molecule of interest, such as the heterodimer of the present invention, in an increasing amount of unlabeled antigen, and detecting the molecules bound to the labeled ligand. The affinity and dissociation rate of the heterodimer of the present invention to the antigen can be determined from saturation data obtained from Scatchard analysis.

[0248] The kinetic parameters of the heterodimers described herein can also be determined using surface plasmon resonance (SPR)-based analyses known in the art (e.g., BIAcore kinetic analysis). For reviews of SPR-based techniques, see Mullet et al., 2000, Methods 22: 77-91; Dong et al., 2002, Review in Mol. Biotech., 82: 303-23; Fivash et al., 1998, Current Opinion in Biotechnology 9: 97-101; Rich et al., 2000, Current Opinion in Biotechnology 11: 54-61. Furthermore, the method of this invention contemplates any SPR instrument and SPR-based method for determining protein-protein interactions described in U.S. Patent Nos. 6,373,577, 6,289,286, 5,322,798, 5,341,215, and 6,268,125. As is known in the art, FACS can also be used to determine affinity.

[0249] Libraries designed using bispecific antibody mutation sets generate bispecific antibodies from Mab1 and Mab2.

[0250] In one embodiment, a bispecific antibody mutation design set is described, designed to selectively form bispecific antibodies from two typical antibodies, Mab1 and Mab2 (composed of antigen-binding fragments Fab1 and Fab2, respectively). This design set consists of homologous mutations corresponding to Fab1, Fab2, and Fc, respectively. In one embodiment, the design set library is represented by the design sets included in any of Tables 5, 12, or 15 through 17. In the presence of competing light and heavy chains of Fab2, mutations are introduced at the interface between the light and heavy chains of Fab1 to achieve selective pairing between the two specific chains. Selective pairing is achieved by introducing favorable complementary mutations in the two specific light and heavy chains based on spatial, hydrophobic, or electrostatic complementarity between certain hotspot framework residues at the interface, while also involving unfavorable interfacial interactions between these mutated residues and non-specific chain pairs. In each design set, selective pairing mutations may also be introduced at the interface between the light and heavy chains of Fab2 in the presence of competing light and heavy chains of Fab1 to achieve selective pairing between these two specific chains. This mutation aims to reduce mismatches between the light chain of Fab1 and the heavy chain of Fab2 (and vice versa). Introducing the mutation into the Fc interface allows for selective pairing of the heavy chains, resulting in an asymmetric antibody molecule containing two distinct heavy chains. Engineering at certain interfacial residue sites of the antibody's light and heavy chains can often lead to detrimental effects, such as loss of antigen-binding affinity, stability, solubility, and aggregation tendency. Many related properties can be affected, such as Kon and Koff rates, melting temperature (Tm), and stability under stress states such as acid, alkali, oxidation, freeze / thaw, stirring, and pressure. This is typically influenced by the complementarity-determining region (CDR) of the antibody of interest. Assuming that the CDRs of different antibodies are generally not identical, the effect of the mutation design set on the aforementioned properties may not be the same across all antibodies. A method is provided here to create bispecific antibodies with significantly higher purity relative to other contaminated, mispaired antibody-like structures, thus yielding any two usable antibodies, Mab1 and Mab2. After introducing homologous mutations into each mutant design set, the light and heavy chains of Mab1 and Mab2 are co-expressed, and the expressed antibody products are analyzed and screened to estimate the purity of the preferred bispecific antibody relative to other Mab-like species expressed in the protein product. In some embodiments, the analytical screening step may be based on LC-MS technology. In some embodiments, the analytical screening step may be based on charge separation techniques such as capillary isoelectric focusing (cIEF) or chromatography. An example of a screening technique based on the SMCA step is provided in Example 9. In some embodiments, significant purity of the bispecific antibody is defined as greater than 70% of all obtained Mab-like species in the expressed protein product. In some embodiments, significant purity of the bispecific antibody is defined as greater than 90% of all obtained Mab-like species in the expressed protein product.The steps for preparing and selecting bispecific Mab design sets to obtain Mab1 and Mab2 are as follows. Figure 12 It is shown schematically.

[0251] Pharmaceutical Composition

[0252] This invention also provides pharmaceutical compositions comprising the heterodimers or heterodimer pairs described herein. Such compositions comprise a therapeutically effective amount of the heterodimer or heterodimer pair and a pharmaceutically acceptable carrier. In specific embodiments, the term "pharmaceuticalally acceptable" means approved by a federal regulatory agency or state government or listed in the United States Pharmacopeia or other well-known pharmacopoeia for use in animals, more specifically in humans. The term "carrier" refers to a diluent, excipient, vehicle, or medium administered with the treatment. Such pharmaceutical carriers can be sterile liquids, such as water and oils, including those of petroleum, animal, plant, or synthetic origin, such as peanut oil, soybean oil, mineral oil, sesame oil, etc. Water is a preferred carrier when the pharmaceutical composition is administered intravenously. Saline solutions, dextran solutions, and glycerol solutions can also be used as liquid carriers, especially for injection. Suitable pharmaceutical excipients include starch, glucose, lactose, sucrose, gelatin, malt, rice flour, chalk, silica gel, sodium stearate, glyceryl monostearate, talc, sodium chloride, skim milk powder, glycerin, propylene glycol, water, ethanol, etc. If desired, the composition may also contain trace amounts of wetting or emulsifying agents or pH buffers. These compositions can be in the form of solutions, suspensions, emulsions, tablets, pills, capsules, powders, sustained-release formulations, etc. The composition can be formulated as suppositories with conventional base materials and carriers such as triglycerides. Oral formulations may include standard carriers such as pharmaceutical-grade mannitol, lactose, starch, magnesium stearate, sodium saccharin, cellulose, magnesium carbonate, etc. Examples of suitable pharmaceutical carriers are described by EW Martin at "Remington's Pharmaceutical Sciences". Such compositions will contain a therapeutically effective amount of the compound (preferably in pure form) and an appropriate amount of carrier to obtain a form suitable for administration to the patient. The formulation should be suitable for the administration mode.

[0253] In some embodiments, compositions comprising heterodimers or heterodimer pairs are formulated according to conventional procedures into pharmaceutical compositions suitable for intravenous administration to humans. Typically, intravenously administered compositions are solutions dissolved in sterile isotonic buffer solutions. Where necessary, the composition may also include a solubilizer and a local anesthetic such as lidocaine to relieve pain at the injection site. Generally, the ingredients are provided individually or in combination in unit dosage forms, such as lyophilized powders or anhydrous concentrates in sealed containers, such as ampoules or pouches indicating the amount of activator. When the composition is administered by infusion, it can be prepared using infusion bottles containing pharmaceutical-grade sterile water or saline. When the composition is administered by injection, ampoules of sterile water or saline for injection can be provided so that the ingredients can be mixed prior to administration.

[0254] In some embodiments, the compositions described herein may be formulated as neutral or salt forms. Pharmaceutically acceptable salts include those that form with anions, such as those derived from hydrochloric acid, phosphoric acid, acetic acid, oxalic acid, tartaric acid, etc., and those that form with cations, such as those derived from sodium, potassium, ammonium, calcium, ferric hydroxide, isopropylamine, triethylamine, 2-ethylaminoethanol, histidine, procaine, etc.

[0255] The amounts of the compositions described herein will be effectively used for the treatment, inhibition, prevention, and / or therapeutic activity of proteins associated with abnormal expression of related diseases or conditions, and these amounts can be determined using standard clinical techniques. Additionally, in vitro analysis may optionally be used to help identify the optimal dose range. The precise dosage used in the formulation also depends on the route of administration and the severity of the disease or condition, and should be determined based on the judgment of the practicing physician and the individual patient's situation. Effective doses are inferred from dose-response curves derived from in vitro or animal model testing systems.

[0256] Use of heterodimer pairs

[0257] As described above, the heterodimer pair described herein may comprise a first heterodimer and a second heterodimer, wherein the immunoglobulin heavy chain and / or immunoglobulin light chain of each heterodimer contains one or more modifications of a known antibody or known antibody that is a known therapeutic antibody or binding molecule. Therefore, it is envisioned that heterodimers containing these antibody modifications could be used to treat or prevent the same diseases, conditions, or infections that can be treated with known therapeutic antibodies or known antibodies.

[0258] In another embodiment, the heterodimer pairs described herein can also be advantageously combined with other therapeutic agents known in the art for the treatment or prevention of cancer, autoimmune diseases, inflammatory conditions, or infectious diseases. In specific embodiments, the heterodimer pairs described herein can be used in combination with monoclonal or chimeric antibodies, lymphokines, or hematopoietic growth factors (such as, for example, IL-2, IL-3, and IL-7) to increase the number or activity of effector cells interacting with the molecules, and to enhance the immune response. The heterodimer pairs described herein can also be advantageously combined with one or more drugs, such as, for example, anticancer agents, anti-inflammatory agents, or antiviral agents, for the treatment of diseases, conditions, or infections.

[0259] Reagent test kit

[0260] The present invention also provides a kit comprising one or more heterodimer pairs. Individual components of the kit may be packaged in separate containers, and such containers may also include notices prescribed by government agencies regulating the manufacture, use, or sale of the drug or biological product, indicating that approval has been obtained from the manufacturing, using, or selling agency. The kit may optionally include instructions or guidance outlining the method of use or administration of the heterodimer pairs.

[0261] When one or more components of the kit are provided as a solution, such as an aqueous solution or a sterile aqueous solution, the container means itself to be an inhaler, syringe, pipette, eye dropper or other similar device that can be used to apply or administer the solution to a subject and to mix with other components of the kit.

[0262] The components of this kit may also be provided in dried or lyophilized form, and the kit may additionally contain a suitable solvent for reconstitution of the lyophilized components. Regardless of the number or type of containers, the kit of the present invention may also include an instrument to assist in the administration of the composition to a subject. Such an instrument may be an inhaler, nasal spray device, syringe, pipette, surgical forceps, measuring spoon, eye dropper, or similar medically approved delivery tool.

[0263] Computer implementation

[0264] In one embodiment, the computer includes at least one processor connected to a chipset. Also connected to the chipset are memory, storage devices, a keyboard, a graphics adapter, a pointing device, and a network adapter. A display is connected to the graphics adapter. In one embodiment, the functionality of the chipset is provided by a storage control center and an I / O control center. In another embodiment, the memory is directly connected to the processor instead of the chipset.

[0265] Storage devices are any devices capable of holding data, such as hard disk drives, CD-ROMs, DVDs, or solid-state storage devices. Memory holds the instructions and data used by the processor. Pointing devices can be mice, trackballs, or other types of pointing devices, and are used in combination with keyboards to input data into the computer system. Graphics adapters display images and other information on a monitor. Network adapters connect the computer system to a local area network (LAN) or a wide area network (WAN).

[0266] As is known in the art, a computer may have components different from those described above and / or others. Furthermore, a computer may lack certain components. Additionally, storage devices may be local to the computer and / or remote (such as being included within a storage area network (SAN)).

[0267] As is known in the art, a computer is capable of executing computer program modules to provide the functions described herein. As used herein, the term "module" refers to computer program logic used to provide a specified function. Therefore, modules can be implemented in hardware, firmware, and / or software. In one embodiment, the program module is stored on a storage device, loaded into memory, and executed by a processor. It should be understood that the examples and embodiments described herein are merely illustrative, and various modifications or variations thereto will be apparent to those skilled in the art and are included within the spirit and scope of this patent application and the appended claims. Example

[0268] The following are examples of specific embodiments of the invention. These embodiments are provided for illustrative purposes only and are not intended to limit the scope of the invention in any way. Efforts have been made to ensure the accuracy of the figures used (e.g., quantities, temperatures, etc.), and some experimental errors and deviations should, of course, be allowed.

[0269] Unless otherwise specified, the practice of this invention will utilize conventional methods of protein chemistry, biochemistry, recombinant DNA techniques, and pharmacology in the art. Such techniques are well explained in the literature. See, for example, TECreighton, Proteins: Structures and Molecular Properties (WH Freeman and Company, 1993); ALLehninger, Biochemistry (Worth Publishers, Inc., latest edition); Sambrook et al., Molecular Cloning: A Laboratory Manual (2nd edition, 1989); Methods In Enzymology (edited by S. Colowick and N. Kaplan, Academic Press, Inc.); Remington's Pharmaceutical Sciences, 18th edition (Easton, Pennsylvania: Mack Publishing Company, 1990); Carey and Sundberg, Advanced Organic Chemistry, 3rd edition (Plenum Press), Volumes A and B (1992).

[0270] Example 1: Molecular modeling and computer-guided Fab interface engineering modification

[0271] Methods guided by structural and computational molecular modeling are used to generate libraries of heavy and light chain mutant designs that can be screened across a range of other antibodies (Abs) or fragments thereof to identify mutations exhibiting the desired specificity in the antibody of interest. A design strategy prioritizing engineered heavy chain (H)-light chain (L) pairings involves first identifying a representative Fab (i.e., D3H44).

[0272] As shown in Table 1, the key criteria for this Fab were: it was humanized / humanized, possessed the commonly used VH and VL subsets, and contained minimal frame region mutations. Furthermore, structural considerations included: the VH:VL domain angle should be close to the average value of the observed antibodies. After selecting Fab D3H44, a two-dimensional approach was used for computer simulation analysis of the Fab interface to identify and understand residues important for the interaction between the heavy and light chains.

[0273] The first method involves a global analysis of sequence conservation between the variable and constant interfaces of the Fabry disease, which is performed by sequence and structure alignment of known antibodies. The alignment of constant and variable domain sequences for various antibody subpopulations is shown in Figure 1. Figure 1A A comparison of representative human VH germline subpopulations is shown. Figure 1B A comparison of representative human κVL germline subpopulations is shown. Figure 1C A comparison of representative human λVL germline subpopulations is shown. Figure 1D The alignment of the human CH1 allele sequence is shown. Figure 1E The alignment of human κ and λ allele sequences is shown. The second method involves the analysis of the D3H44 crystal structure interface using numerous molecular modeling tools, such as... Figure 2 As shown (e.g., ResidueContacts) TM The results of these analyses identify a list of hotspot sites for HL pairings preferred for engineering modification. The hotspot sites identified by this analysis are listed in Table 2. These sites and amino acids are primarily framework residues (except for a few in the CDR3 ring) and are also largely conserved in the λL chain. The amino acid and Kabat numbers of the parent D3H44 sequence are provided in Tables 3a-3b.

[0274] Then, potential mutations at hotspot sites and adjacent hotspot sites of interest in the 3D crystal structure were induced by computer simulation mutagenesis and Zymepack. TM Stacking / modeling simulation and validation. Zymepack TM Zymepack is a software package that takes an input structure and a set of mutations, alters the types of residues in the input structure based on the provided mutations, and generates a new structure that approximates the physical structure of the mutant protein. Additionally, Zymepack assesses the properties of mutant proteins by calculating numerous quantitative metrics. These metrics include measurements of spatial and electrostatic complementarity, which are correlated with the stability, binding affinity, or heterodimer specificity of the mutant protein.

[0275] Figure 3 This paper provides a subset of hotspot locations at the heavy-chain and light-chain interfaces in variable structural domains and demonstrates how mutations can be introduced at these interface locations to promote selective strand pairing while repelling the formation of incorrect strand pairs. Zymepack is included. TMThe computational method models spatial complementarity and calculates it based on energy factors such as van der Waals stacking, hole effects, and close contact of hydrophobic groups. Similarly, electrostatic interaction energies are modeled and evaluated based on Coulomb interactions between charges, hydrogen bonding, and desolvation effects. Simulations calculate relative spatial and electrostatic scores by introducing both preferred heavy-chain / light-chain pair models (such as H1:L1 (or H2:L2)) and incorrect pair models (such as H1:L2 (and H2:L1)) obtained from mutations of interest. This allows determining whether a particular set of mutations results in favorable energies, i.e., preferred (specific) heavy-chain / light-chain pairs with greater spatial and / or electrostatic complementarity relative to incorrect (non-specific) pairs. Calculated spatial and electrostatic energies are components of the free energy associated with light-chain / heavy-chain pairing. Therefore, greater spatial and electrostatic complementarity is a characterization of a larger change in free energy associated with specific pairings relative to non-specific pairings. Greater spatial or electrostatic complementarity results in preferred (selective) pairings of specific heavy-chain and light-chain pairs (relative to non-specific pairs).

[0276] Example 2: Design Selection and Description

[0277] The method described in Example 1 is used to design heavy-light chain heterodimer pairs (e.g., H1-L1 and H2-L2) that exhibit selective or preferential pairing. The heterodimer pairing design, referred to as the “design” or “design set,” includes a set of substitutions on the H1, L1, H2, or L2 chain that promotes preferential pairing (Table 5). The design set is initially tested with an “LCCA design” (Table 4), in which one heavy chain is co-expressed with two light chains to evaluate relative pairing. The Kabat numbering system is used in conjunction with Tables 3a and 3b to identify amino acid substitutions.

[0278] The design library described in Table 30 of International Patent Application No. PCT / CA2013 / 050914 was used as a starting point for identifying some of the LCCA designs shown in Table 4 and the design set shown in Table 5. Some designs in Tables 4 and 5 are new, independent designs. Table 6 shows the core designs, along with their associated unique identifiers. Most designs only span the constant region, while a few designs also include modifications to the variable region. These designs were proposed to further drive pairing specificity while also benefiting the transferability of other antibody systems.

[0279] For the derived designs, the design library described in Table 30 of International Patent Application No. PCT / CA2013 / 050914 was used as the starting point. These designs were clustered by structural similarity and ranked according to the strength of pairing specificity, their effect on antigen binding, and their stability as determined by differential scanning calorimetry (DSC). The derived designs were then combined (see examples in Table 7) and / or optimized (see examples in Tables 8 and 9) to obtain the final designs. For the combinations, at least one design exhibited high pairing specificity, while the others exhibited a range of favorable pairing specificities. All designs selected for combination and / or optimization exhibited no / minimum effect on antigen binding and no / minimum effect on melting temperature (Tm).

[0280] Independent designs are tested individually (classified as Independent under the Design Type column in Table 5) and in combination with derived designs (classified as Independent / Combined under the Design Type column in Table 5; see also the examples in Table 10).

[0281] The designs were stacked into a molecular model of D3H44 and the metrics were calculated (as described in Example 1). The optimal design was then selected based on risk (potential impact on stability and immunogenicity) and impact (the recommended strength considering the drive pair specificity). These optimal designs are shown in Table 5.

[0282] Example 3: Preparation of Fab constructs encoding the D3H44IgG heavy chain and the D3H44IgG light chain.

[0283] Wild-type Fab heavy and light chains of the anti-tissue factor antibody D3H44 were prepared as described below. The D3H44 Fab light chain (AJ308087.1) and heavy chain (AJ308086.1) sequences were obtained from GenBank (Table 3c). Genes were synthesized and codon-optimized for mammalian expression. The light chain vector insert sequence, consisting of a 5'-EcoRI cleavage site-HLA-A signal peptide-HA or FLAG tag-light chain Ig clone-"TGA terminator"-BamH1 cleavage site-3', was ligated into the pTT5 vector (Durocher Y et al., Nucl. Acids Res. 2002; 30, No. 2e9). The resulting vector + insert sequence was sequenced to confirm the correct reading frame and encoding DNA sequence. Similarly, the heavy chain vector insert sequence, consisting of a 5'-EcoR1 cleavage site-HLA-A signal peptide-heavy chain clone (terminated at T238; see Table 3a)-ABD2-His6 tag-TGA termination-BamH1 cleavage site-3', was ligated to the pTT5 vector (ABD; albumin-binding domain). The resulting vector + insert sequence was also sequenced to confirm the correct reading frame and encoding DNA sequence. Various Fab D3H44 constructs containing amino acid substitutions of the design set were generated via gene synthesis or site-directed mutagenesis (Braman J, Papworth C & Greener A., ​​Methods Mol. Biol. (1996) 57: 31-44).

[0284] Heavy and light chains were labeled at the C- and N-termini, respectively, to facilitate the evaluation of preferred pairings via competitive analysis-SPR screening. The ABD2-His6 heavy chain tag specificity allowed HL complexes to be captured on the anti-his-tagged SPR chip surface, while the FLAG and HA light chain tags allowed for the quantification of relative L1 and L2 populations.

[0285] Example 4: Preferred pairing evaluation of Fab heterodimers containing constant domain modifications or combinations of constant and variable domain modifications of the D3H44IgG light and / or heavy chains.

[0286] Constructs encoding the D3H44IgG heavy and light chains in Fab form, containing amino acid modifications according to the LCCA design set in Table 12, were prepared as described in Example 3. The ability of the constructs to preferentially pair to form the desired H1-L1 heterodimer within the LCCA design set (H1, L1, L2) was determined using light chain competition analysis (LCCA).

[0287] LCCA quantifies the relative pairing of one heavy chain with at least two unique light chains, and can be summarized as follows. One D3H44 heavy chain Fab construct is co-expressed with two unique D3H44 light chain Fab constructs, and the relative light chain pairing specificity (e.g., H1-L1:H1-L2) is determined by competitive analysis-SPR screening, repeated. The LCCA screening ratio is distorted by reducing the amount of L1 (designed to preferentially pair with the H chain) relative to L2 (e.g., L1:L2 = 1:3, by weight), while maintaining a limited amount of heavy chain (i.e., H1:L1+L2 = 1:3) to identify strong drivers. The amount of each heterodimer formed (i.e., H1-L1 and H1-L2) is determined as follows: the heavy chain is bound to the SPR chip by his-tag pull-down, and then the amount of each light chain tag (HA or FLAG) is detected using these tag-specific antibodies. Subsequently, the selection of heterodimer hits was verified by light chain competition analysis, whereby the L1:L2 DNA ratio varied between 1:3 and 1:9 during transfection, while maintaining a limited amount of heavy chain. It was also noted that in the D3H44 system, light chain tags (HA or FLAG) did not affect LCCA pairing (see Example 10 of International Patent Application No. PCT / CA2013 / 050914). A schematic representation of the analytical design is shown below. Figure 4 As shown. Figure 5 This demonstrates how to label heavy and light chains and how to evaluate preferred pairings. Experimental details for LCCA are provided below.

[0288] Transfection method

[0289] The LCCA design, comprising one heavy chain and two light chains, was prepared as described in Example 3 and transfected into CHO-3E7 cells as follows. CHO-3E7 cells were cultured at 1.7–2 x 10⁻⁶ cells per cell line. 6 FreeStyle at a density of cells / ml at 37°C, supplemented with 4 mM glutamine and 0.1% Koliphor P188 (Sigma#K4894). TM Cells were cultured in F17 medium (Invitrogen CAT#A-1383501). A total of 2 μg of DNA was transfected into a 2 ml total volume using PEI-pro (Polyplus transfection#115-375) at a DNA:PEI ratio of 1:2.5. Twenty-four hours later, the DNA-PEI mixture was added, and cells were transferred to 32°C. On day 7, non-reducing SDS-PAGE analysis was performed, followed by Coomassie blue staining to observe bands and assess expression in the supernatant. The H:L ratios are shown in Table 11.

[0290] Competition Analysis SPR Method

[0291] In the LCCA design, the degree to which the D3H44 light chain is preferentially paired with the D3H44 heavy chain is assessed using SPR-based readings of unique epitope tags located at the N-terminus of each light chain.

[0292] Surface plasmon resonance (SPR) was supplied by Bio-Rad Laboratories (Canada) Ltd. (Mississauga, ON). GLC sensor chips, Biorad ProteOn amine coupling kit (1-ethyl-3-(3-dimethylaminopropyl)carbodiimide hydrochloride (EDC), N-hydroxysulfosuccinimide (sNHS), and ethanolamine), and 10 mM sodium acetate buffer were purchased from Bio-Rad Laboratories (Canada) Ltd. (Mississauga, ON). 4-(2-hydroxyethyl)-1-piperazine ethanesulfonic acid (HEPES) buffer, ethylenediaminetetraacetic acid (EDTA), and NaCl were purchased from Sigma-Aldrich (Oakville, ON). 10% Tween 20 solution was purchased from Teknova (Hollister, CA).

[0293] SPR mutual sensing analysis. All surface plasmon resonance analyses were performed using a BioRad ProteOn XPR36 instrument (Bio-Rad Laboratories (Canada) Ltd. (Mississauga, ON)) and PBST run buffer (PBSTeknova Inc., containing 0.05% Tween 20) at 25°C. Anti-5-His capture surfaces were generated using a GLM sensor chip, which was activated by injecting 100 μL / min of a 1:5 diluted standard BioRad sNHS / EDC solution along the analyte (horizontal) direction for 140 s. Immediately after activation, 25 μg / mL of anti-5-His antibody (Qiagen Inc.) dissolved in 10 mM NaOAc pH 4.5 was injected along the analyte (vertical) direction at a flow rate of 25 μL / min until approximately 3000 resonance units (RUs) were immobilized. The remaining activation groups were quenched by injecting 1M ethanolamine at 100 μL / min along the analyte direction for 140 s, which also ensured the generation of a simulated activation midpoint for blank reference.

[0294] Screening for heterodimers binding to anti-FLAG (Sigma Inc.) and anti-HA (Roche Inc.) monoclonal antibodies was performed in two steps: the heterodimers were indirectly captured on the anti-5-His surface along the ligand direction, followed by injection of anti-FLAG and anti-HA in the analyte direction. First, PBST was injected once for 30 s at 100 μL / min along the ligand direction to stabilize the baseline. For each heterodimer capture, the unpurified heterodimer in cell culture medium was diluted to 4% in PBST. One to five heterodimers or controls (i.e., controls containing 100% HA-light chain or 100% FLAG-light chain) were simultaneously injected into a single ligand channel at a flow rate of 25 μL / min for 240 s, resulting in a saturated heterodimer capture of approximately 300 to 400 RU on the anti-5-His surface. If desired, the first ligand channel was left empty as a blank control. Immediately following the heterodimer capture step, two buffer injections were performed along the analyte direction to stabilize the baseline. Then, 5 nM anti-FLAG and 5 nM anti-HA were injected twice at 50 μL / min for 120 s each, with a 180 s dissociation phase, resulting in a set of binding sensitograms with buffer references for each captured heterodimer. Tissue factor (TF) antigen bound to the heterodimer was injected into the last remaining analyte channel as an activity control. The heterodimers were regenerated by a 100 μL / min pulse of 0.85% phosphate for 18 s to prepare the anti-5-His surface for the next injection cycle. The sensitograms were compared and analyzed using ProteinOn Manager software v3.0 with double references of buffer blank injections and midpoints.

[0295] result

[0296] The LCCA results are shown in Tables 12, 13a, and 14a. Note that in Tables 13 and 14, the “unique identifier” may not correspond exactly to Table 5, as the two unique identifiers that make up the LCCA can be in either direction ((set #H1L1L2 - set #H2L2L1) or (set #H2L2L1 - set #H1L1L2)). The evaluation of preferred pairings for each LCCA design is shown in the last three columns of Table 12. The same data also includes the design pair ranges in columns 5, 6, and 8 or 10, 11, and 13 of Tables 13a and 14a. Each unique H1, L1, and L2 mutation (LCCA design) group is assigned a unique number or 'set #' (e.g., 9567 or 9087). When the data is presented in H1L1 H2 L2 form (Fab pair form or design set), such design sets are represented by a “unique identifier” that includes the set numbers of the two constituent LCCAs (e.g., 9567-9087). Note that most LCCA experiments were performed on constructs containing interchain Fab disulfide bonds (H / C233-L / C214, Kabat numbers) located in constant structural domains. In Tables 13(a and b) and 14(a and b), two complementary LCCA sets (H1, L1, L2 and H2, L2, L1) are represented as Fab pairs to highlight the success of the specially designed pairings relative to preferred pairings. The presence of tags (L chain: HA and FLAG, H chain: ABD2-His6) does not affect the expected ~50%:50% neutral pairing of D3H44WT.

[0297] In the table, the reported LCCA data are median values ​​in ratio form (H1-L1:H1-L2 and H2-L2:H2-L1) (normalized to a 1:1 L1:L2 DNA ratio). Furthermore, the LCCA data are normalized to 100% because the total L1 and L2 amounts observed for some variants differed significantly from 100%. This inconsistency in the total light chain percentage is believed to be partly due to variable nonspecific binding during the initial heterodimer capture on the SPR chip. Since the LCCA experiments were performed at two different L1:L2 DNA ratios (L1:L2 1:3 and 1:9, respectively), two LCCA normalized ratios are listed in the table. It is noted that LCCA data for some experiments were not reported because the obtained experimental data did not meet the inclusion criteria (e.g., Fab captured on the SPR chip was less than 100, or the total LCCA amounts for L1 and L2 fell outside the range of 60 to 140).

[0298] Table 12 lists all LCCA designs (530) for which data were available. Considering two L1:L2 DNA ratios of 1:3 and 1:9, 490 (92.5%) of the 530 LCCA designs had at least 60% correct pairing (L1:L2 DNA ratio normalized to 1:1). The remaining LCCA designs included predominantly neutral designs (32 / 530 or 6.0%) and a small number producing inconsistent results (8 / 530 or 1.5%). The designs shown in Table 12 are primarily electrostatic (based on specific drives utilizing hydrogen bonds or charge-charge interactions), and some designs also include spatial complementarity and / or interchain covalent disulfide bonds. Some designs also include mutations to form new disulfide bonds in the absence of natural interchain disulfide bonds (formed via H / C233-L / C214).

[0299] For the two heterodimers in the design set, Tables 13(a and b) and 14(a and b) list 447 designs that provide LCCA data. Tables 13a and 14a show that the computer simulation design method described in Example 1 achieves preferred pairings of H1-L1 instead of H1-L2 and H2-L2 instead of H2-L1 in many different groups of designs and their variations.

[0300] Tables 13 (a and b) list those designs where the average LCCA performance of the paired: mismatched Fab heterodimers (the median average of the L1:L2 ratios normalized to 1:1 for H1-L1:H1-L2 and H2-L2:H2-L1) is at least 86:14, while Tables 14 (a and b) list those designs where the average LCCA performance of the paired: mismatched Fab heterodimers is less than 86:14. The performance of each LCCA was normalized to 100% and a 1:1 L1:L2 DNA ratio (as described in the example above), and was described by scalar values ​​((ln(r1 / f1) or ln(r2 / f2)) (where r1 and r2 correspond to the median H1L1:H1L2 and H2L2:H2L1 at the experimental ratios, respectively, and f1 and f2 correspond to their respective experimental ratios) and the ratio of paired:mismatched Fab heterodimers. Each design also had an associated average LCCA performance scalar value (0.5(ln(r1 / f1) + ln(r2 / f2))), which was also normalized. The L1:L2 DNA ratio was set to 100% and 1:1 (as described in the example above). Furthermore, the scalar range of each LCCA design is shown (LCCA1 and LCCA2 correspond to H1L1:H1L2 and H2L2:H2L1 experiments, respectively). Of the 447 Mab designs, 354 (79.2%) exhibited an average LCCA performance of at least 86:14 (Tables 13a and b). The designs within Tables 13 (a and b) are also characterized into 13 clusters based on design similarity. Designs within each cluster are arranged from highest to lowest average LCCA performance scalar value.

[0301] In addition, the LCCA data in Table 13a are also available. Figure 7 The diagram is shown in Chinese. Figure 7Box plots are shown, illustrating the pairings for each cluster: mismatched Fab heterodimers have an average LCCA performance of at least 86:14. The bottom of each box represents the first quartile (Q1), which is the median average LCCA performance between the minimum and median values, such that values ​​less than the first quartile represent the smallest 25% of the data. The horizontal bars inside the boxes represent the second quartile, which is the median average LCCA performance of the cluster. The top of each box represents the third quartile (Q3), which is the median average LCCA performance between the maximum and median values, such that values ​​greater than the third quartile represent the largest 25% of the data. The interquartile range is the difference between Q3 and Q1. Boxes extending vertically in both directions represent the data range of those values ​​within Q1 - (1.5 * IQR) or Q3 + (1.5 * IQR). The horizontal bars covering the boxes represent the maximum and minimum values ​​within that range. Data outside the box plot and box whiskers are identified as outliers. Moderate outliers are indicated by points (differences from Q1 or Q3 ranging from 1.5*IQR to 3*IQR), while extreme outliers are indicated by plus signs (differences from Q1 or Q3 greater than 3*IQR).

[0302] Example 5: Scale-up of biophysical characterization

[0303] Scale up (typically to 20 ml) the correctly paired heterodimers shown in Table 5 with the unique identifier set, and purify as described below to test thermostability and antigen binding. The heavy and light chains of each heterodimer were expressed in 20 ml of CHO-3E7 cell culture. CHO-3E7 cells were expressed at 1.7–2 x 10⁻⁶ cells per ... 6 FreeStyle at a density of cells / ml at 37°C, supplemented with 4 mM glutamine and 0.1% Koliphor P188 (Sigma#K4894). TM Cells were cultured in F17 medium (Invitrogen cat#A-1383501). A total of 20 μg of DNA was transfected into a 20 ml total volume using PEI-pro (Polyplus cat#115-375) at a DNA:PEI ratio of 1:2.5. Twenty-four hours later, the DNA-PEI mixture was added, and the cells were transferred to 32°C.

[0304] Seven days after transfection, cells were centrifuged, and heterodimers were purified from the supernatant using high-throughput nickel affinity chromatography, as described below. The supernatant was rinsed in equilibration buffer (calcium, magnesium, and phenol red free of HyClone). TM #SH30028.02) was diluted in Dulbecco phosphate-buffered saline (DPBS) to 20-25% of the cell culture supernatant, and then mixed with... Ni-NTA resin (Thermo Scientific #PT-88222) (also pre-equilibrated with equilibration buffer) was stirred and incubated for 12 hours. The resin was then collected by centrifugation, transferred to a 96-well sintered plate, washed three times with equilibration buffer, and then... Elution with elution buffer (Sigma-Aldrich #H5413).

[0305] After purification, heterodimer expression was assessed using a Caliper LabChip GXII (Perkin Elmer #760499) via High Throughput Protein Express (HT Protein Express) analysis. The procedure was performed according to the HT Protein Express LabChip User Guide, Edition 2, LabChip GXII User Manual, with the following modifications: 2 μl or 5 μl of heterodimer sample (concentration range 5–2000 ng / μl) and 7 μl of HT Protein Express Sample Buffer (Perkin Elmer #760328) were added to individual wells in a 96-well plate (BioRad #HSP9601). The heterodimer sample was then denatured at 70 °C for 15 min. The LabChip instrument was run using the HT Protein Express Chip (Perkin Elmer #760499) and Ab-200 analysis settings. After use, clean the chip with MilliQ water and store it at 4°C.

[0306] Example 6: Thermal stability determination of Fab heterodimer by DSF.

[0307] To assess thermal stability, differential scanning fluorescence (DSF) was used as a high-throughput method to screen all correctly paired heterodimers from wild-type, unmodified heavy-light chain pairs. The heterodimers were prepared as described in Example 5.

[0308] Determination of thermal stability

[0309] The thermal stability of all heterodimer pairs was determined using DSF as described below. Each heterodimer was purified as described in Example 5 and diluted to 0.5 mg / mL in DPBS (HyClone Cat#SH30028.02). For most samples, Sypro Orange gel staining working stock (Life Technologies Cat#S-6650) was prepared by diluting 4 μL of Sypro Orange gel staining to 2 mL of DPBS. DSF samples were prepared by adding 14 μL of 0.5 mg / mL protein to 60 μL of diluted Sypro Orange gel staining working stock. However, for proteins less than 0.5 mg / mL, each DSF sample was prepared by adding 14 μL of undiluted protein to 60 μL of Sypro Orange dye working stock (i.e., diluted to 1:1500 in DPBS). DSF analysis was then performed repeatedly on 20 μl aliquots using a Rotor-Gene 6000 qPCR instrument (QiaGen Inc). Each sample was scanned at 1°C intervals from 30°C to 94°C, with a 10-second equilibration period between each step and a 30-second initial wait time. A 470 nm excitation filter and a 610 nm emission filter were used, with a gain of 9. The maximum value of the first derivative of the denaturation curve was used as Tm, and data were analyzed using Rotor-Gene 6000 software. The remaining DSF samples were prepared and analyzed similarly using the following modifications without altering the Tm determination: 1) a working stock was prepared by diluting 1 μL of Sypro Orange gel staining to 2 ml of DPBS; 2) 30 μl aliquots were analyzed; and 3) a gain of 10 was used.

[0310] The DSF results are shown in Tables 12, 13b, and 14b. The thermal stability of H1:L1Fab (DSF values ​​and DSF value variations compared to wild type) within the LCCA design range is shown in columns 3 and 4 of Table 12. The same DSF values ​​are also included in the design pair ranges in columns 7 and 8 of Tables 13b and 14b. For each Fab heterodimer in which repetition is performed, the reported Tm value is the median. Comparison of the Tm values ​​of the Fab heterodimer relative to the wild-type Fab heterodimer (wild-type Fab constructs containing the HA tag, with a median Tm of 81.0 °C) is reported in the H1L1_dTm_dsf column. Note that for new Fab heterodimers lacking native interchain disulfide bonds (between H chain C233 and L chain C214), the H1L1_dTm_dsf values ​​for the corresponding wild-type Fab not identified as lacking native interchain disulfide bonds are not evaluated. It was also noted that some Tm values ​​for Fab heterodimers (17 / 230 or 7.4% of Fab heterodimers) were not reported due to the quality of the corresponding experiments (e.g., low yield, low intensity, partially blocked peaks, and variability greater than 1°C between repeating sequences in Fab heterodimers). For some of these Fab heterodimers, predicted Tm values ​​corresponding to similar Fab heterodimer Tm values ​​were reported, differing only in the presence / absence or identification of linked L-chain tags (HA or FLAG). For the predicted Tm values, the corresponding wild-type Tm value (81.2°C) was the median obtained from all wild-type Fab heterodimer constructs (i.e., Fab constructs containing HA or FLAG tags). HA or FLAG tags did not significantly affect the Tm values ​​of wild-type Fab heterodimers. In summary, Fab heterodimers exhibited similar Tm values ​​compared to WT. Among Fab heterodimers containing natural interchain disulfide bonds and for which DSF data was available, 93% (195 / 209) of the Fab heterodimers showed a temperature reduction of 3°C or less relative to WT. Furthermore, the most affected Fab heterodimers showed a temperature reduction of 6.5°C relative to WT. Table 12 lists the LCCA designs arranged in descending order of Tm.

[0311] In addition, thirteen amino acid substitutions that substantially improved the stability of the Fab heterodimer were identified (see Table 34). Stable mutations were identified after comparing Fab heterodimers including stable mutations with similar Fab heterodimers that did not contain stable mutations. Stable heavy chain mutations included A125R, H172R, L143F, Q179D, Q179E, Q39R, S188L, and V190F. Stable light chain mutations included Q124E, Q124R, Q160F, S176L, and T180E. In summary, stable mutations increased stability by 0.4 °C to 2.1 °C. Heavy chain stable mutations A125R, H172R, L143F, Q179D, Q179E, Q39R, S188L, and V190F increased stability by 0.4℃ to 0.6℃, 0.4℃ to 2.1℃, 0.4℃, 0.5℃ to 0.6℃, 0.5℃ to 0.8℃, 1.1℃ to 1.6℃, 0.4℃ to 1.2℃, and 1℃, respectively. Light chain stable mutations Q124E, Q124R, Q160F, S176L, and T180E increased stability by 0.4℃ to 0.5℃, 0.8℃ to 0.9℃, 0.6℃, 0.4℃ to 1.0℃, and 0.5℃, respectively.

[0312] Example 7: Determination of affinity for Fab heterodimer antigen.

[0313] The ability of Fab heterodimers to bind tissue factor was assessed to determine whether amino acid substitutions affect the ability of heterodimers to bind antigens. The affinity of each Fab heterodimer for tissue factor was determined by SPR as described below.

[0314] SPR supplier. GLC sensor chip, Biorad ProteOn amine coupling kit (EDC, sNHS, and ethanolamine), and 10 mM sodium acetate buffer were purchased from Bio-Rad Laboratories (Canada) Ltd. (Mississauga, ON). PBS run buffer containing 0.05% Tween 20 (PBST) was purchased from Teknoca Inc. (Hollister, CA).

[0315] Fab heterodimers were batched. Purified Fab heterodimers were tested in three batches, A, B, and C. Batches A and B were stored at 4°C for approximately one month before SPR analysis, while batch C, containing purified Fab heterodimers, was stored at 4°C for approximately two months before SPR analysis. The Fab heterodimers in batch C are indicated in Table 12 by a "+" sign following the corresponding KD value.

[0316] All surface plasmon resonance analyses were performed using a BioRad ProteOn XPR36 instrument (Bio-Rad Laboratories (Canada) Ltd. (Mississauga, ON)) and PBST run buffer at 25°C. Anti-pentahis capture surfaces were generated using a GLC sensor chip activated by injecting 100 μL / min of a 1:5 diluted standard BioRads NHS / EDC solution along the analyte direction (horizontal) for 140 s. Immediately after activation, 25 μg / mL of anti-pentahis antibody (Qiagen Inc.) dissolved in 10 mM NaOAc pH 4.5 was injected along the analyte direction (vertical) at a flow rate of 25 μL / min until approximately 3000 resonance units (RUs) were immobilized. The remaining activated groups were quenched by injecting 1 M ethanolamine at 100 μL / min along the analyte direction for 140 s, which also ensured the generation of a simulated activation midpoint for blank reference.

[0317] Screening for Fab heterodimers binding to the TF antigen was performed in two steps: Fab heterodimers were indirectly captured onto the surface of the anti-5-His antibody along the ligand direction, followed by simultaneous injection of five concentrations of purified antigen and one buffer blank as a dual reference along the analyte direction. First, buffer was injected once for 30 seconds at 100 μL / min along the ligand direction to stabilize the baseline. One to five variants or controls were dissolved in PBST at a concentration of 3.4 μg / ml and simultaneously injected through a single ligand channel at a flow rate of 25 μL / min for 240 seconds. For batches A and B, this resulted in an average capture of approximately 1000 RU on the anti-5-His surface, and for batch C, an average capture of approximately 600 RU on the anti-5-His surface. If desired, the first ligand channel was left empty as a blank control. Immediately following the capture step, two buffer injections were performed along the analyte direction at 100 μL / min for 30 s each to stabilize the baseline. Then, 60 nM, 20 nM, 6.7 nM, 2.2 nM, and 0.74 nM antigens (TFs) and a buffer blank were simultaneously injected at 50 μL / min for 120 s, with a dissociation phase of 600 s. The captured antibody surface was regenerated by a pulse of 0.85% phosphate at 100 μL / min for 18 s, followed by two 18-s regeneration cycles in preparation for the next injection cycle. Sensing maps were compared using a double reference with buffer blank injections and midpoints, and the resulting sensing maps were analyzed using ProteOn Manager software v3.1. The double-referenced sensing maps were fitted to a 1:1 binding model. The Rmax value for each antigen was normalized to the antibody capture level for each variant and compared to a 100% control.

[0318] The antigen affinity (KD) values ​​for Fab heterodimer samples are reported in Tables 12, 13b, and 14b. The KD values ​​(KD, range of KD values, and variation of the median KD value compared to wild type) for H1:L1 Fab within the LCCA design range are shown in columns 5, 6, and 7 of Table 12, respectively. The same KD values ​​are also included in the design range in columns 3 (KD of H1-L1 Fab heterodimer), 4 (variation of KD of H1-L1 Fab heterodimer compared to wild type), 5 (KD of H2-L2 Fab heterodimer), and 6 (variation of KD of H2-L2 Fab heterodimer compared to wild type) of Tables 13b and 14b. Only Fab heterodimer samples exhibiting Fab heterodimer capture of at least 100 RU were measured. The reference wild-type KD (0.157 nM) reflects the median KD of wild-type Fab heterodimers in which the light chain contains the FLAG tag. Wild-type Fab heterodimers (containing either the FLAG or HA tag) exhibit similar KD values, resulting in a 2.6-fold difference between the observed maximum and minimum values. Tables 12, 13b, and 14b show the KD differences relative to wild-type antigen-binding affinity, calculated using -(log(KD)design - log(KD)wild-type), where positive values ​​indicate a decrease in KD value for antigen-binding affinity of the Fab heterodimer compared to the wild-type, while negative values ​​indicate an increase in KD value. It is noted that some Fab heterodimers lack measurable KD values. In some of these cases, Fab heterodimers were evaluated, but SPR experiments showed reduced Fab heterodimer capture (i.e., less than 100 RU), making accurate KD value determination impossible. For those Fab heterodimers exhibiting similarity to other Fab heterodimers (i.e., differing only in the presence / absence or identification of linked L-chain tags (HA or FLAG)), predicted KD values ​​corresponding to similar Fab heterodimer KD values ​​are provided (as shown in Tables 12, 13b, and 14b). The corresponding predicted wild-type KD value (0.15 nM) is the median obtained from all wild-type Fab heterodimer constructs (i.e., Fab constructs containing HA or FLAG tags). In summary, the results show that correctly paired heterodimers (from a design perspective) exhibit antigen-binding affinity similar to wild-type (within approximately 2.3 times the reference wild-type affinity).

[0319] Example 8. Ultra-high performance liquid chromatography-size exclusion chromatography (UPLC-SEC) chromatograms of wild-type labeled D3H44 heterodimers and preferred paired heterodimers.

[0320] Wild-type D3H44 heterodimers (one heavy chain and one light chain) with a C-terminal ABD2-His6 tag on the heavy chain and an N-terminal tag (FLAG in one construct and HA in the other) on the light chain were expressed and purified according to methods known in the art and similar to those described in Example 5. As described in Example 5, preferred or correctly paired heterodimers were amplified and purified separately via His tag affinity purification.

[0321] UPLC-SEC was performed using a Waters BEH200SEC column (2.5 mL, 4.6 x 150 mm, stainless steel, 1.7 μm particles) set at 30 °C and mounted on a Waters Acquity UPLC system with a PDA detector. Run times were 7 min, with a total injection volume of 2.8 mL of Hyclone DPBS / modified calcium-magnesium run buffer (part No. SH30028.02) at a flow rate of 0.4 mL / min. Elution was monitored by UV absorbance in the 200–400 nm range, and chromatograms were extracted at 280 nm. Peak integration was performed using Empower 3 software.

[0322] Figure 6 The UPLC-SEC spectra of representative WT Fab heterodimer pairs (with FLAG tags on the L chain) and representative design Fab heterodimer pairs (H1L1 Fab components of LCCA designs 9735, 9737, and 9740) are shown. Generally, design Fab heterodimer pairs exhibit similar UPLC-SEC spectra compared to WT.

[0323] Example 9: Evaluation of preferential pairing of heterodimers in a co-expression set containing constant domains or constant and variable domain modifications in the form of bispecific antibodies.

[0324] The heterodimer designs were evaluated to determine whether they also allowed for preferential pairing in bispecific antibody forms. In this example, to facilitate heterodimerization of unique heavy chains, the Fc region of the full-length heavy chain of each heterodimer was asymmetrically modified such that one heavy chain contained mutants T350V, L351Y, F405A, and Y407V, and the other heavy chain contained mutants T350V, T366L, K392L, and T394W (EU number).

[0325] Construction of the construct:

[0326] Heterodimeric designs were tested across the following range of bispecific antibodies: a) D3H44 / trastuzumab, b) D3H44 / cetuximab, and c) trastuzumab / cetuximab. Note that D3H44 is a human antibody, trastuzumab is a humanized antibody, and cetuximab is a chimeric antibody composed of human IgG1 and the mouse Fv region. Constructs encoding the amino acid-modified heavy and light chains of D3H44, trastuzumab, and cetuximab IgG were prepared as described below, according to the design. The basic DNA sequences of the heavy and light chains of D3H44, trastuzumab, and cetuximab are shown in Table 3C. The light chain sequences of D3H44, trastuzumab, and cetuximab were prepared as described in Example 3, except that some sequences lacked tags while others contained FLAG or HA tags. The D3H44, trastuzumab, and cetuximab heavy chain sequences were prepared as described in Example 3, except that the full-length heavy chains were created from an attached IgG1*01 DNA sequence encoding the hinge-CH2-CH3 domain and modified to promote heterodimerization at the C-terminus of the CH1 domain of the Fab heavy chain. Notably, typical C-terminal lysine residues of the heavy chain were removed to prevent LC-MS signal heterogeneity due to C-terminal lysine cleavage (Lawrence W. Dick Jr. et al., Biotechnol. Bioeng. (2008) 100: 1132-43).

[0327] Analytical Form (SMCA)

[0328] The ability of a heterodimer co-expression set to preferentially pair and form bispecific antibodies was evaluated as described below. This analysis was based on the co-expression of four chains (H1 and L1 chains of one antibody, and H2 and L2 chains of another antibody), and the presence of correctly formed bispecific antibodies was detected using mass spectrometry (LC-MS). Figure 8 A schematic diagram is provided illustrating four starting polypeptide chains and the possible products obtained from the co-expression of these starting polypeptide chains in the absence of preferential pairing between heavy and light chains and heterodimer pairs. Co-expression of two full-length heavy chain constructs with two unique light chain constructs yields ten possible antibody species: H1-L1∶H1-L1, H1-L2∶H1-L2, H1-L1∶H1-L2, H2-L1∶H2-L1, H2-L2∶H2-L2, H2-L1∶H2-L2, H1-L1∶H2-L1, H1-L2∶H2-L2, H1-L2∶H2-L1, and H1-L1∶H2-L2. The H1-L1∶H2-L2 species is a correctly paired bispecific antibody (see [link]). Figure 8Following pA purification and deglycosylation, the relative pairing specificity of the preferred species H1-L1:H2-L2 relative to the amounts of other species was determined using LC-MS. If possible, the chains were left unlabeled to obtain all Mab and half-antibody species that were at least 50 Da different from each other. When poor quality ruled out, an N-terminal tag (HA or FLAG) was added to the light chain to provide sufficient inter-species quality differentiation.

[0329] This analysis involves the expression and screening of bispecific antibodies, a process known as SMCA.

[0330] Mass spectrometry

[0331] After purification of protein A and non-denaturing deglycosylation, mass spectrometry was used to assess the preferential pairing of the D3H44 light chain and D3H44 heavy chain in the co-expression pool. Since the D3H44 / trastuzumab heterodimer contains only Fc N-linked glycans, the system was treated with only one enzyme, N-glycosidase F (PNGase-F). The purified sample was deglycosylated with PNGase F as follows: 0.2 U PNGase F / μg antibody was dissolved in 50 mM Tris-HCl pH 7.0 and incubated overnight at 37°C, resulting in a final protein concentration of 0.5 mg / mL. For the D3H44 / cetuximab and trastuzumab / cetuximab systems, because the Fab region of cetuximab contains additional N-linked glycans, the system was treated with N-glycosidase F in addition to multiple exoglycosidases. Typically, a mixture of four enzymes is used for this purpose: N-glycosidase F, β-galactosidase (Prozyme), β-N-acetylglucosidase (NewEngland Biolabs), and neuraminidase. N-glycosidase F removes Fc N-linked glycans, while exoglycosidase trims FabN-linked glycans into a homogeneous core structure M3F (GlcNAc2Man3Fuc1). As described below, the purified sample was deglycosylated using a mixture of four enzymes: 0.2 U PNGase F / μg antibody, 0.002 U α-neuraminidase / μg antibody, 0.0001 U β-galactosidase / μg antibody, and 0.2 U β-N-acetylglucosidase / μg antibody dissolved in 50 mM Tris-HCl pH 7.0 and incubated overnight at 37°C, resulting in a final protein concentration of 0.5 mg / mL. However, in some cases, treatment with three enzymes (N-glycosidase F, β-galactosidase, and neuraminidase) is preferred to avoid mass overlap of sample components in LC-MS analysis. In these cases, Fab glycan is reduced to the slightly larger structure G0F (Man3GlcNAc2Fuc1GlcNAc2). The purified sample is deglycosylated using a mixture of three enzymes at the same concentrations and conditions as described for a mixture of four enzymes. After deglycosylation, the sample is stored at 4°C before LC-MS analysis.

[0332] Deglycosylated protein samples were analyzed by full LC-MS using an Agilent 1100 HPLC system (via an Ion Maxelectrospray ion source (ThermoFisher) connected to an LTQ-Orbitrap XL mass spectrometer (ThermoFisherScientific)). Samples (5 μg) were injected into a 2.1 x 30 mm Poros R2 reversed-phase column (Applied Biosystems) and resolved using the following gradient conditions: 0–3 min: 20% solvent B; 3–6 min: 20–90% solvent B; 6–7 min: 90–20% solvent B; 7–9 min: 20% solvent B. Solvent A was a degassed 0.1% formic acid aqueous solution, and solvent B was a degassed acetonitrile. The flow rate was 3 mL / min. A split was applied post-column, with 100 μL introduced into the electrospray interface. The column was heated to 82.5 °C, and the solvent was heated to 80 °C before the column to improve protein peak shape. The LTQ-Orbitrap XL was calibrated using ThermoFisher Scientific's LTQ PositiveIon ESI calibration solution (caffeine, MRFA, and Ultramark 1621) and tuned using 10 mg / m³ LCSI solution. The cone voltage (source fragmentation setting) was 40 V, the FT resolution was 7,500, and the scan range was m / z 400–4,000. The LTQ-Orbitrap XL was tuned for optimal detection of large proteins (>50 kDa).

[0333] The MaxEnt 1 module of MassLynx (Waters) instrument control and data analysis software will be used to analyze whole antibodies (m / z 2000-3800) and half antibodies (m / z 1400-2000). In short, the raw protein LC-MS data is first opened in the QualBrower spectral view module of Xcalibur (ThermoScientific) and converted to a MassLynx-compatible format using the Databridge file conversion program provided by Waters. The converted protein spectra are then viewed in the MassLynx spectral module, and deconvolution is performed using MaxEnt 1. The abundance of different antibody species in each sample is determined directly from the resulting molecular weight spectra.

[0334] Representative design of SMCA analysis

[0335] A total of 25 representative designs with high average LCCA performance values ​​were selected from clusters 1 to 12 for testing in SMCA form. Representative designs were selected based on corresponding design sets occupying similar spaces, using similar drivers and sharing similar mutations. At least one representative design was selected from each cluster. Some clusters were represented by one representative design (i.e., clusters 1, 5, 7, 8, and 10). The remaining clusters had more than one representative design because these clusters were large (i.e., cluster 2) or consisted of small clusters (i.e., clusters 3, 4, 6, 9, 11, and 12). Although the designs within each cluster shared sequence similarity, the differences between the sub-clusters within a cluster were at least one set of driver mutations. For clusters containing sub-clusters, additional representative designs were selected from each sub-cluster.

[0336] The amino acid substitutions for each cluster are listed in Tables 15 to 27, with the corresponding representative amino acid substitutions for each cluster / subcluster indicated. For cluster 1, only one design (9134-9521) is selected to represent the cluster, as these designs utilize similar electrostatic drives occupying similar spaces (see Table 15). For all members of this cluster, H1 is designed to allow negatively charged substitutions (L124E and Q179E) to form salt bridges with positively charged substitutions in L1 (S176R and S131K or S131R). H2 is designed to allow positively charged substitutions (L124R and Q179K or S186K) to form salt bridges with negatively charged substitutions in L2 (S176D and T178D or T178E and / or T180E). Mismatches between H1L2 and H2L1 are repulsive, primarily due to electrostatic repulsion.

[0337] For cluster 2, two representative designs (9279-9518 and 9286-9402) are selected to represent the large cluster (see Table 16). The designs within this cluster utilize similar electrostatic drives occupying similar spaces. For all members of this cluster, H1 is designed to allow negatively charged substitutions (L124E and L143E or L143D) to form salt bridges with positively charged substitutions in L1 (S176R and combinations of (Q124K and / or T178K) or (Q124K and Q160K)). H2 is designed to allow positively charged substitutions (L124R and Q179K or S186K or S186R) to form salt bridges with negatively charged substitutions in L2 (S176D and T178D or T178E and / or T180E). The mismatches between H1L2 and H2L1 are repulsive, primarily due to electrostatic repulsion.

[0338] For cluster 3, five representative designs (9338-9748, 9815-9825, 6054-9327, 9066-9335, and 9121-9373) are selected to represent each of the five subclusters (see Table 17). All members of this cluster utilize similar electrostatic drives on H1 (L124E), L1 (S176R), H2 (L124R), and L2 (S176D), which allow the formation of salt bridges in preferentially paired heterodimers, while mispairing is repulsive, primarily due to electrostatic repulsion. To represent those designs that primarily utilize those constant-region drives, design 6054-9327 is selected to represent this subcluster. In addition to these electrostatic interactions, one subcluster also includes variable-region space drives (H1L45P and L1P44F), therefore a representative design including this variable-region drive is selected to represent this subcluster (9338-9748). Another subcluster also includes variable-region electrostatic driving in two Fab heterodimers (H1Q39E, L1Q38R, H2Q39R, L2Q38E), and therefore a representative design including this variable-region driving is chosen to represent this subcluster (9815-9825). Furthermore, a subcluster consisting of one member, thus a representative design (9066-9335), includes an engineered disulfide bond between H1F122C and L1Q124C. The remaining subclusters, represented as 9121-9373, primarily utilize constant-region driving with additional substitutions (H172T in H1 and S174R in L1) to slightly modify the constant-region driving interaction of H1L1, while also probing the effect of H172R in HC2.

[0339] For cluster 4, two representative designs (9168-9342 and 9118-6098) are selected to represent each of the two subclusters (see Table 18). All members of this cluster utilize similar electrostatic drives on H1 (L124E), L1 (S176R or S176K), H2 (L124R), and L2 (S176D), which allow the formation of salt bridges in preferentially paired heterodimers, while mispairing is repulsive, primarily due to electrostatic repulsion. One subcluster, represented by 9118-6098, primarily uses the shared electrostatic drive for preferential pairing, while the other subclusters, represented by 9168-9342, also utilize substitutions of H1 (K228D) and L1 (S121K) that allow the formation of additional salt bridges.

[0340] Cluster 5, represented by the unique identifier 9116-9349, consists of only one member (see Table 19). This design utilizes two electrostatic drives on H1 (L124E), L1 (S176R), H2 (L124R), and L2 (S176D), and spatial drives on H1 (A139W), L1 (F116A_V133G_L135V), H2 (A139G_V190A), and L2 (V133G_L135W). Therefore, for preferentially paired heterodimers, charged displacement allows for the formation of salt bridges. For mismatched Fab heterodimers, their formation is repulsive due to electrostatic repulsion and additional spatial effects.

[0341] For cluster 6, two representative designs are selected to represent each of the two sub-clusters (see Table 20). All members of this cluster utilize similar electrostatic drives in the constant region (Q179E on H1, S131K on L1, S186R on H2, and Q124E, Q160E, and T180E on L2), which allow the formation of salt bridges in preferentially paired heterodimers, while mispairing is repulsive, primarily due to electrostatic repulsion. Furthermore, this sub-cluster is also composed of different variable region drives. One sub-cluster, represented by unique identifiers 9814-9828, utilizes electrostatic drives in the variable region (Q39E on H1, Q38R on L1, Q39R on H2, and Q38E on L2). The other sub-clusters utilize variable region spatial drives composed of L45P in H1 and P44F in L1. Therefore, for this sub-cluster, mispairing is also repulsive due to the introduced spatial effects. Note that this cluster is represented by a design derived from the unique identifier 9745-9075, the only difference being the absence of Q38E on L2.

[0342] For cluster 7, only one design (9060-9756) is selected to represent the cluster because these designs utilize similar electrostatic and spatial actuations (see Table 21). Common electrostatic actuations include L143E and Q179E on H1, Q124R on L1, Q179K on H2, and Q124E, Q160E, and T180E on L2. Common spatial actuations include A139W on H1, F116A-L135V on L1, and L135W on L2. Therefore, for preferred heterodimers, charged displacement in the Fab region allows for the formation of salt bridges. For mismatched heterodimers, their formation is repulsive due to electrostatic repulsion and additional spatial effects.

[0343] For cluster 8, only one design (9820-9823) is selected to represent the cluster because these designs utilize similar electrostatic actuation (see Table 22). In the variable region, Q39E on H1, Q38R on L1, Q39R on H2, and Q38E on L2 are utilized. In the constant region, L143E on H1, Q124R, Q160K, and T178R on L1, Q179K on H2, and Q124E, Q160E, and T180E on L2 are utilized. For preferred-paired heterodimers, charged displacement in the Fab region allows for the formation of salt bridges, while for mismatched heterodimers, their formation is repulsive, primarily due to electrostatic repulsion.

[0344] For cluster 9, two representative designs are selected to represent each of the two subclusters (see Table 23). All members of this cluster utilize similar electrostatic drives in the constant region (L143E on H1, Q124R on L1, Q179K on H2, and Q124E, Q160E, and T180E on L2), which allow the formation of salt bridges in preferentially paired heterodimers, while mispairing is repulsive, primarily due to electrostatic repulsion. Furthermore, the subclusters differ in the presence / absence of variable region drives (L45P on H1 and P44F on L1). Therefore, for subclusters containing variable region drives, mispairing is also repulsive due to the introduced spatial effects. Representative designs for this subcluster including variable region drives are derived from unique identifiers 9751-9065, which differ only in the absence of Q38E on L2. Representative designs for subclusters lacking variable region substitutions are 9611-9077.

[0345] For cluster 10, only one design (9561-9095) is selected to represent the cluster because these designs utilize similar electrostatic and spatial drives occupying similar spaces (see Table 24). Common electrostatic drives include L143E and Q179E on H1, Q124R, Q124K, or S131K at similar positions on L1, Q179K on H2, and Q124E and T180E on L2. Common spatial drives include L124W on H1, V133A on L1, and V133W on L2. Therefore, for preferred heterodimers, charged displacement in the Fab region allows for the formation of salt bridges. For mismatched heterodimers, their formation is repulsive due to electrostatic repulsion and additional spatial effects.

[0346] For cluster 11, three designs (9049-9759, 9682-9740, and 9667-9830) are selected to represent each of the three sub-clusters (see Table 25). All members of this cluster utilize electrostatic displacement to drive the preferential pairing of heterodimers. Therefore, for preferentially paired heterodimers, charged displacement in the Fab region allows the formation of salt bridges. For mismatched heterodimers, their formation is repulsive, primarily due to electrostatic repulsion. For the sub-clusters represented by the unique identifiers 9667-9830, the common displacements include negatively charged displacements on H1 (L143E or L143D and Q179E or Q179D), positively charged displacements on L1 (T178R or T178K), positively charged displacements on H2 (S186K or S186R or Q179K or Q179R), and negatively charged displacements on L2 (Q124E). Another subcluster, represented by a single member of this subcluster (unique identifiers 9049-9759), also contains substitutions for disulfide bonds used to form engineered modifications. The remaining subclusters, represented by unique identifiers 9682-9740, utilize H1 and L1, similar to the other two subclusters; however, they utilize different constant regions H2 and L2. H2 utilizes L143R or L143K, while L2 utilizes V133E or V133D, except for the Q124E substitution (shared with the other two subclusters).

[0347] For cluster 12, four designs (9696-9848, 9986-9978, 9692-9846, and 9587-9735) are selected to represent each of the four sub-clusters (see Table 26). All members of this cluster utilize electrostatic substitution to drive preferential pairing of heterodimers. Some members also utilize spatial driving. The sub-clusters represented by the unique identifiers 9696-9848 utilize both electrostatic and spatial driving. The common electrostatic substitution within this sub-cluster consists of L143E on H1, Q124R and T178R on L1, S186K or S186R or Q179K or Q179R at similar positions on H2, and Q124E and T180E on L2. The shared spatial permutations within this subcluster consist of S188L on H1 and S176L or V133Y or V133W on L2; for designs utilizing V133Y or V133W on L2, L143A or L124A also exists on H2 to accommodate large volumetric abrupt changes. For subclusters represented by unique identifiers 9692-9846, similar electrostatic driving is used compared to subclusters represented by unique identifiers 9696-9848; for some members, the permutation T178E at a similar position is used instead of T180E on L2. Furthermore, subsets of this subcluster also utilize the permutation T178Y or T178F at a similar position, instead of the similar spatial driving of S176L on L2. Subclusters represented by unique identifiers 9986-9978 utilize only electrostatic driving to drive preferred pairing. Similar shared permutations are used for H1, L1, and H2; however, a different L2 permutation (S131E) is used. The remaining subclusters, identified by unique identifiers 9587-9735, utilize similar electrostatic drives on H1 and L1 (the difference being that T178R on L1 is not used for all members within this subcluster); however, different electrostatic drives are used for H2 (L143R or L143K) and L2 (Q124E and V133E or Q124E and V133D). Two members within this subcluster also utilize similar spatial drives, which are constituted by S188L on H1 and S176L on L2. In summary, for preferred heterodimers, charged substitution in the Fab region allows for the formation of salt bridges. For mismatched heterodimers, their formation is repulsive due to electrostatic repulsion. Furthermore, for designs also utilizing spatial drives, their formation is also repulsive due to steric effects. Cluster 13 consists of one member, 9122-9371 (see Table 27). This design utilizes an engineered disulfide bond between H1F122C and L1Q124C as a covalently driven preferred pairing of heterodimers. Furthermore, since this design lacks native interchain disulfide bonds, disulfide bond formation was confirmed by non-reducing and reducing SDS-PAGE gel electrophoresis. This design was not tested in SMCA form; however, in the presence of native interchain disulfide bonds, the engineered disulfide bonds were tested in combination with another constant-region driven design (cluster 3, representative designs 9066-9335).

[0348] Transfection method

[0349] As described below, a co-expression set containing two heavy chain and two light chain constructs was transfected into CHO-3E7 cells. CHO-3E7 cells were cultured at a concentration of 1.7–2 x 10⁻⁶ cells / cells. 6 Cells were cultured at a density of [number] cells / ml in FreeStyle™ F17 medium (Invitrogen cat#A-1383501) supplemented with 4 mM glutamine and 0.1% Pluronic F-68 (Invitrogen cat#24040-032) at 37°C. A total of 50 μg of DNA was transfected into a 50 ml total volume using PEI-pro (Polyplus cat#115-010) at a DNA:PEI ratio of 1:2.5. Twenty-four hours later, the DNA-PEI mixture was added, and the cells were transferred to 32°C and incubated for 7 days, after which they were harvested. The medium was collected by centrifugation and vacuum filtered using a Sterelip 0.2 μM filter. The filtered medium was then purified using Protein A MabSelect SuRe resin (GE Healthcare #17-5438-02) as described below. The filtered culture medium was applied to a column previously equilibrated with PBS (Hyclone DPBS / modified, calcium-free, magnesium-free, #SH-300028.02). The heterodimeric antibody species were then washed with PBS and eluted with 100 mM citrate at pH 3.6 using an Amicon Ultra 15 filter Ultracel 10K (Millipore #SCGP00525). The buffer was then exchanged with PBS, and the sample was assessed by calipers before deglycosylation and LC-MS.

[0350] To assess the inherent bispecific system preference in wild-type bispecific antibody systems (where the light chain of one system preferentially binds to the heavy chains of both antibody systems), a set of H1:H2:L1:L2 DNA ratios were tested in CHO expression. These ratios attempt to compensate for natural differences in expression levels and / or intrinsic pairing preferences between the heavy and light chains of the two different antibodies. For all bispecific antibody systems, preference was observed among all tested ratios (Figure 9). For the D3H44 / trastuzumab system, preference for trastuzumab, i.e., the D3H44 heavy chain preferentially pairing with the trastuzumab light chain (see Figure 9). Figure 9a For D3H44 / cetuximab, a preference for cetuximab was observed, i.e., the D3H44 heavy chain preferentially paired with the cetuximab light chain (see [link to study]). Figure 9bFor the trastuzumab / cetuximab system, a preference for trastuzumab was observed, i.e., the cetuximab heavy chain preferentially paired with the trastuzumab light chain (see [link to study]). Figure 9c ).

[0351] For each of the 25 representative designs within each bispecific antibody system tested, the H1:H2:L1:L2 DNA ratio used was the ratio of the corresponding wild-type bispecific system that produced a large number of bispecific antibody species while having a small amount of half-antibody (see Tables 32a, b, and c). For the D3H44 / trastuzumab system, the ratio used was 15(H1), 15(H2), 53(L1), and 17(L1), where H1 and L1 refer to D3H44, and H2 and L2 refer to trastuzumab. For the trastuzumab / cetuximab system, the ratio used was 15(H1), 15(H2), 17(L1), and 53(L2), where H1 and L1 refer to trastuzumab, and H2 and L2 refer to cetuximab. For the D3H44 / cetuximab system, the ratios used are 15(H1), 15(H2), 53(L1), and 17(L2), where H1 and L1 refer to D3H44, and H2 and L2 refer to cetuximab.

[0352] Furthermore, for the D3H44 / cetuximab and trastuzumab / cetuximab bispecific systems, a two-directional testing design was employed, such that in one direction, substitutions present on H1L1 and H2L2 were tested on antibody 1 (Ab1) and antibody 2 (Ab2) of the bispecific antibody system, respectively, and in the other “flipped” direction, substitutions present on H1L1 and H2L2 were tested on Ab2 and Ab1, respectively (see Tables 28a and b). The “_1” appended to the unique identifier indicates a design in which the heavy chain and associated light chain substitutions that yield a stronger LCCA preferred pairing result (see Table 13a) are placed on antibodies in which the heavy chain weakly competes for its associated light chain, rather than the light chain of other antibodies. The “_2” appended to the unique identifier indicates the opposite “flipped” direction: the heavy chain and associated light chain substitutions that yield a stronger LCCA preferred pairing result (see Table 13a) are placed on antibodies in which the heavy chain strongly competes for its associated light chain, rather than the light chain of other antibodies. For the D3H44 / trastuzumab system, the test design was only performed along the "_1" direction (see Table 28c).

[0353] SMCA Results

[0354] The D3H44 / trastuzumab system was treated with only one enzyme (PNGase-F) and completely deglycosylated. For multi-enzyme treatments, the linker sugar in the Fab region was typically truncated to a core M3F (treated with four enzymes) or G0F (treated with three enzymes). In summary, in most cases, deglycosylation treatment yielded the ability to identify all possible different species as identified by LC-MS. In many cases, each species was represented by a single LC-MS peak. Exceptions include side peaks that may correspond to the desired bispecific species (possibly due to inhomogeneities in adduct or leader peptide cleavage); however, these side peaks are not considered in the contribution of bispecific species due to their ambiguity. Furthermore, some designs within the D3H44 / cetuximab (3519_1, 3522_1) and trastuzumab / cetuximab (9748-9338_1) systems required multiple peaks to account for species arising from fluctuations in the linked high mannose content. All these designs introduced glycosylation sites in the cetuximab light chain. It should be noted that in some cases, due to the low quality separation between species (i.e., differences less than 50 Da), it is impossible to distinguish some small species (including less than 5% of all species). Furthermore, the desired bispecific species H1-L1_H2-L2 often cannot be distinguished experimentally from the mismatched type H1-L2_H2-L1 by LC / MS. Therefore, when bispecific content is reported in a table, the absence of mismatched species of that type cannot be completely ruled out. However, the observed very low levels of species such as H1-L2_H1-L2 and H2-L1_H2-L1, as well as H1-L2 and H2-L1 haptens, are indicative of only trace amounts (if any) of bispecific species impurities.

[0355] LC-MS data are provided in Tables 29a, 29b, and 29c. For comparison, wild-type data are also provided in Tables 33a, 33b, and 33c, indicated as "NA" in the "SMCA Unique Identifier" column and the "Cluster" column. All three bispecific wild-type systems exhibited a distorted preference, making one light chain preferred over two heavy chains (see Table 33 and Figure 9). Furthermore, at least in the trastuzumab / cetuximab system, tag position also appears to have a significant impact on H1L1 and H2L2 pairing. Therefore, to assess the impact of the design on the transferability and percentage of the desired bispecific species relative to the wild type, comparisons were made with the corresponding wild-type bispecific constructs at the same H1:H2:L1:L2 DNA ratio, and reported in the columns “% H1L1 pairing change (all H1 species), relative to wild type,” “% H2L2 pairing change (all H2 species), relative to wild type,” and “% H1:H2:L1:L2 change, relative to wild type” (considering only all antibody species) (see Table 29). Note that for the assessment of % H1L1 pairing (all H1 species) or % H2L2 pairing (all H2 species), pairing in the Fab region was assessed for all species. When the corresponding wild-type bispecific construct was not assessed by SMCA, comparisons were made with similar wild-type constructs. Estimates are indicated by “***” after the reported value. Similar wild-type constructs for comparison were selected as described below. To assess transferability, each wild-type construct was represented by a SMCA experiment (performed at different ratios) showing the highest “%H1L1 and %H2L2 pairing (all species)”. To assess the impact of the design on the percentage of desired bispecific species to wild-type, each wild-type construct was represented by a SMCA experiment (performed at different ratios) showing the highest %H1:H2:L1:L2 (considering only all antibody species). In both cases, the median value was selected as the wild-type value for comparison among all wild-type constructs within the bispecific system.

[0356] For each design, transferability was assessed by recording the overall H:L pairings relative to the wild type in all species, particularly the increase in %H1L1 / all H1 species and / or %H2L2 / all H2 species. Furthermore, the impact on the percentage of desired bispecific species was evaluated, focusing on whole antibody species comparisons only, as haptens (if present) can be removed / minimized by preparative SEC or by additional H1:H2:L1:L2 DNA titration. Tables 30a, b, and c show that preparative SEC can be effectively used for the removal / minimization of hapten species. Tables 32a, b, and c show that the percentage of hapten species can also be manipulated using various DNA titration ratios during transfection.

[0357] For the D3H44 / cetuximab system (Table 29a), transfers were made in all designs except one (9327-6054_2), as assessed by H1L1 pairings relative to wild-type in all species. Most designs (except 9327-6054_2 and 9134-9521_2) also showed an increased percentage of bispecific antibodies required when considering only all antibody species. Furthermore, except for one non-transfer design (9327-6054_2), the designs reduced the number of initially mismatched antibody species (H1H2L2L2) observed in wild-type. Additionally, designs transferred in two directions, except for 9327-6054_2 and its corresponding design 9327-6054_1, with most designs showing similar effective H:L pairings in both directions.

[0358] For the D3H44 / trastuzumab system (Table 29b), all designs were transferred, as assessed by H1L1 pairing relative to wild-type in all species. Furthermore, all designs showed an increased percentage of the desired bispecific antibody (when considering only all antibody species). Additionally, most designs significantly reduced the number of initially mismatched antibody species (H1H2L2L2) observed in wild-type. However, it is noted that data for 9611-9077_1 were not reported due to lack of expression (Table 28c).

[0359] For the trastuzumab / cetuximab system (Table 29c), at least 35 of the 49 designs showed transferability, as assessed by H2L2 pairings across all H2 species (positive values ​​in the "% H2L2 pairing variation (all H2 species) relative to wild type" column). Designs that appeared to be non-transferable included 9279-9518_2, 3522_2, 9815-9825_2, 9327-6054_2, 9118-6098_2, 9748-9338_2, 9692-9846_2, 9587-9735_2, 9814-9828_2, 3519_2, 9986-9978_2, 9168-9342_2, and 9066-9335_1 (negative values ​​in the "% H2L2 pairing variation (all H2 species), relative to wild type" column); however, designs in other orientations showed transferability (note that 9279-9518_1 was not tested due to lack of samples). All designs showing transferability exhibited a reduction in the percentage of initially mismatched antibody species (H1H2L1L1) observed in wild-type experiments. Furthermore, in the design of the transfer, when only all antibody species are considered, only two designs (9134-9521_1 and 9279-9518_2) showed a percentage reduction in the required bispecific antibody compared to the wild type.

[0360] Generally, designs with increased H:L pairing of most weakly competitive antibodies reduce the percentage of bispecific antibody required (considering only whole antibodies). In terms of orientation, most designs in the "_1" orientation showed similar or better transferability of H:L pairing compared to the "_2" orientation (exceptions were primarily observed in the trastuzumab / cetuximab system). Furthermore, Table 35a lists those designs that transferred in both orientations across all three tested bispecific systems (D3H44 / cetuximab, D3H44 / trastuzumab, and trastuzumab / cetuximab). Table 35b lists those designs that transferred in one orientation across all three bispecific systems (D3H44 / cetuximab, D3H44 / trastuzumab, and trastuzumab / cetuximab), and those that transferred in the other orientation across only one bispecific system. Additionally, in the specified orientation, the same mutation was present in both the heavy chain and the weakly competing homologous light chain in all three bispecific systems, and the light chain utilization was at least greater than 10%.

[0361] Regarding the transferability and performance of clusters, in the case of the D3H44 / trastuzumab bispecific system, all members within all clusters exhibited transferability (see [link to relevant documentation]). Figure 11a ), and the required percentage increase in bispecific antibodies (considering only whole antibodies) (see Figure 11b Regarding the D3H44 / cetuximab bispecific system, all clusters showed transferability, with only one member in cluster 3 showing reduced H1L1 pairing compared to wild-type in all H1 species (see [link]). Figure 11c Additionally, all clusters include members exhibiting an increased percentage of the desired bispecific antibody relative to the wild type (considering only whole antibodies); however, three clusters (clusters 1, 3, and 4) also include members exhibiting a decreased percentage of the desired bispecific antibody relative to the wild type (considering only whole antibodies) (see [link to relevant documentation]). Figure 11d For the trastuzumab / cetuximab bispecific system, all clusters include variants that exhibit design transferability; however, only a few clusters (1, 5, 7, 8, 10, 11) include variants in which all of their respective members exhibit transferability (see [link to relevant documentation]). Figure 11e Furthermore, all clusters include members exhibiting the desired percentage increase in bispecific antibodies relative to wild-type (considering only whole antibodies) (see [link to relevant documentation]). Figure 11f For those clusters where all members showed transferability, all members within clusters 5, 7, 8, 10, and 11 also showed an increased percentage of the required bispecific antibody relative to the wild type (considering only whole antibodies).

[0362] In summary, when all three bispecific systems are considered holistically, all members within clusters 1, 5, 7, 8, 10, and 11 exhibit transferability (see [link to relevant documentation]). Figure 11gClusters 5, 7, 8, 10, and 11 include members that exhibit an increased percentage of the desired bispecific antibody relative to the wild type (considering only whole antibodies) (see [link to relevant documentation]). Figure 11h ).

[0363] In summary, designs with increased H:L pairing for most weakly competitive antibodies reduced the percentage of bispecific antibodies required (considering only whole antibodies). In terms of orientation, most designs in the "_1" orientation showed similar or better transferability of H:L pairing comparisons compared to the "_2" orientation (exceptions were primarily observed in the trastuzumab / cetuximab system).

[0364] Example 10: Preparative size exclusion chromatography (SEC) of selected SMCA bispecific heterodimer antibody and parental Mab for biophysical characterization.

[0365] A subset of SMCA samples were selected for further biophysical characterization. Most of these SMCA samples typically exhibited high pairing (greater than ~80% pairing, in the H1L1+H2L2 / all species column) and low abundance (less than -30%, considering all hapten species). Preparative SEC was performed as described below. [The following text appears to be unrelated and possibly a separate excerpt:] ...using equipment installed at Pharmacia (GE Healthcare) Heterodimeric antibody samples were separated using a Superdex 200 10 / 300 GL (GE Healthcare) column on the Purifier system. Heterodimeric antibody samples (0.3–0.5 ml) dissolved in PBS (Hyclone DPBS / modified, calcium-free, magnesium-free, Cat no. SH-300028.02) were manually loaded into 0.5 ml loops (filled with PBS). The samples were then automatically injected into the column and eluted using a 1CV column at 0.5 ml / min. OD was monitored. 280 Proteins were eluted and 0.5 ml fractions were collected. For each SMCA sample, fractions containing the main peak were pooled and further biophysically characterized.

[0366] Example 11: Preferred pairing evaluation of bispecific heterodimers in antibody form after preparative size exclusion chromatography

[0367] Following preparative SEC, as described in Example 9, the preferred pairing of bispecific heterodimeric antibodies in the selected samples was analyzed using LC-MS. All these samples showed an increase in the percentage of desired bispecific antibody species and a decrease in the percentage of half-antibody species (Tables 29 and 30).

[0368] Example 12: Thermal stability of SMCA bispecific heterodimer antibody.

[0369] Following the preparative SEC, the thermostability of the selected SMCA bispecific heterodimer antibody was determined and compared with that of parental D3H44 and trastuzumab monoclonal antibodies, as well as cetuximab single-arm antibody. Generally, a single-arm antibody refers to a construct consisting of a full-length heavy chain, a truncated heavy chain lacking the Fab region (including C233S substitution), and a light chain, wherein the heavy chain heterodimerization is achieved as described in Example 9.

[0370] Determination of thermal stability

[0371] The thermostability of the bispecific heterodimer antibody and wild-type control was determined using differential scanning calorimetry (DSC) as described below. Following preparative SEC treatment, 400 μL samples, primarily dissolved in PBS at concentrations of 0.2 mg / mL or 0.4 mg / mL, were analyzed using a VP-Capillary DSC (GE Healthcare). At the start of each DSC run, five buffer blank injections were performed to stabilize the baseline, followed by buffer injections, and then each sample injection was used as a reference. Each sample was scanned from 20°C to 100°C at a rate of 60°C / hour with low feedback, 8-s filtration, 5-min preTstat, and 70 psi nitrogen pressure. The resulting thermograms were compared and analyzed using Origin 7 software.

[0372] The results are shown in Tables 31a, b, and c. The wild-type Fab Tm values ​​reported in the tables were obtained from the homodimer antibodies of D3H44 (79°C) and trastuzumab (81°C), and the single-arm antibody of cetuximab (72°C). For the WT D3H44 / cetuximab and trastuzumab / cetuximab heterodimer antibodies, only two peaks corresponding to Fab Tm were observed. No distinct peaks were observed for CH2 (due to overlap with cetuximab Fab) or CH3 (due to overlap with the Tm values ​​of D3H44 and trastuzumab Fab). For the WT D3H44 / trastuzumab heterodimer antibody, since the Tm values ​​of the two Fabs of D3H44 and trastuzumab are similar, the peak at 81°C may correspond to both Fabs, while the peak at approximately 72°C may correspond to CH2.

[0373] In Tables 31a, b, and c, only the Tm values ​​for the peaks corresponding to the two Fab peaks are reported, unless otherwise specified. It is also noted that, for some heterodimer samples, the protein concentration was very low (less than 0.4 mg / mL), resulting in increased baseline noise. Therefore, in the D3H44 / trastuzumab system, some samples produced DSC curves with low peak intensities, making it difficult to distinguish the CH2 peak from the potentially unstable Fab peak. In these cases, Tm values ​​at 70–72 °C are also reported (Table 31a). In summary, most heterodimer antibodies exhibit thermal stability similar to their corresponding wild-type molecules (3 °C or less). Furthermore, most heterodimer antibodies do not exhibit additional peaks, suggesting significant instability of the CH2 or CH3 peaks. One exception is the heterodimer antibody 9611-9077_2 engineered from the trastuzumab / cetuximab system, which exhibited an additional peak at 60 °C, possibly due to CH2 instability.

[0374] Example 13: Determination of antigen affinity of bispecific heterodimeric antibodies

[0375] The ability of bispecific antibodies to bind to relevant antigens was assessed to determine whether amino acid substitutions affected antigen binding. Antigen-binding affinity was determined by SPR as described below.

[0376] SPR Biosensing Analysis

[0377] EDC: 1-Ethyl-3-(3-dimethylaminopropyl)carbodiimide hydrochloride; NHS: N-hydroxysuccinimide; SPR: Surface plasmon resonance; EDTA: ethylenediaminetetraacetic acid; TF: tissue factor; EGFR ECD: extracellular domain of epidermal growth factor receptor; Her2ECD: extracellular domain of human epidermal growth factor receptor 2.

[0378] SPR supplier. Series S sensor chip CM5, Biacore amine coupling kit (NHS, EDC, and 1M ethanolamine), and 10mM sodium acetate buffer were purchased from GE Healthcare Life Science (Mississauga, ON). Recombinant Her2 extracellular domain (ECD) protein was purchased from eBioscience (San Diego, CA). PBS running buffer containing 1% Tween20 (PBST) was purchased from Teknova Inc. (Hollister, CA). Goat polyclonal anti-human Fc antibody was purchased from Jackson ImmunoResearch Laboratories Inc. (West Grove, PA). EDTA was purchased from Bioshop (Burlington, ON).

[0379] All surface plasmon resonance (SPR) analyses were performed using a Biacore T200 SPR instrument (GE Healthcare Life Science, Mississauga, ON) and PBST run buffer (with 0.5 M EDTA stock solution added to achieve a final concentration of 3.4 mM) at 25°C. Anti-human Fc capture surfaces were generated using the Series S sensor chip CM5 with default parameters via the Immobilization Wizard in the Biacore T200 control software (set to target 2000 resonance units (RU)). Screening for antibody variants targeting Her2ECD, TF, or EGFR ECD antigens was performed in two steps: the antibody variants were indirectly captured onto the surface of the anti-human Fc antibody flow cell, followed by injection of five concentrations of purified antigen, followed by kinetic analysis using a single-cycle kinetic method. Variants or controls used for capture were injected at a flow rate of 10 μL / min for 60 s in a single flow cell at 1 μg / mL. Typically, approximately 50 to 100 RU were captured on the anti-human Fc surface. The first flow cell was left empty as a blank control. Immediately after this capture step, five concentrations of antigen (5 nM, 2.5 nM, 1.25 nM, 0.63 nM, and 0.31 nM for TF or EGFR ECD antigen, or 40 nm, 20 nm, 10 nm, 5 nm, and 2.5 nm for Her2ECD antigen) were sequentially injected into all four flow cells at 100 μL / min for 180 s, followed by a dissociation phase of 300 s for EGFR ECD, 1800 s for Her2ECD, and 3600 s for TF. The captured antibody surface was regenerated by injecting 10 mM glycine at pH 1.5 at 30 μL / min for 120 s to prepare for the next injection cycle. For each analyte injection, at least two injections of simulated buffer were performed as a reference. The resulting single-cycle kinetic sensor plots were analyzed using Biacore T200 BiaEvaluation software and fitted to a 1:1 binding model.

[0380] Antigen affinity of heterodimeric antibodies was assessed in conjunction with their respective wild-type controls: Mab of D3H44, trastuzumab OAA, and cetuximab OAA. Antigen affinity of wild-type bispecific antibodies was also obtained; however, SPR capture of WT bispecific antibodies can be heterogeneous (e.g., involving the capture of mismatched heterodimers), thus interfering with KD assays (see Tables 31a and c). For heterodimeric antibodies measuring antigen binding in the D3H44 / cetuximab system, antigen affinity was similar to the corresponding WT controls (see Table 31b). For most heterodimeric antibodies measuring antigen binding in the D3H44 / trastuzumab and trastuzumab / cetuximab systems, antigen affinity was similar to the corresponding WT controls (see Tables 31a and c). Exceptions include eleven engineered antibodies that did not exhibit Her2 binding. In the D3H44 / trastuzumab and trastuzumab / cetuximab systems, no HER2 binding was observed in the six engineered heterodimeric antibodies 9049-9759_1, 9682-9740_1, and 3522_1. Furthermore, regarding the trastuzumab / cetuximab system, five additional engineered antibodies 9696-9848_1, 9561-9095_2, 9611-9077_2, 9286-9402_2, and 9060-9756_2 also lacked HER2 binding. Ten of these eleven engineered antibodies shared constant region mutations on the H chain (L143E_K145T) and L chain (Q124R_T178R). Another engineered antibody, 9286-9402_2, shares the same constant region mutation on the H chain (L143E_K145T) and similar mutations on the L chain (Q124K and S176R).

[0381] Example 14: Ultra-high performance liquid chromatography-size exclusion chromatography (UPLC-SEC) chromatograms of engineered heterodimer antibodies, wild-type heterodimer antibodies, and homodimer antibodies.

[0382] Following preparative SEC of engineered heterodimeric antibodies and control wild-type bispecific and homodimeric antibodies, UPLC-SEC was performed using a Waters BEH200 SEC column (2.5 mL, 4.6 x 150 mm, stainless steel, 1.7 μm particles) set at 30 °C and mounted on a Waters Acquity UPLC system with a PDA detector. Run times were 7 min, with a total injection volume of 2.8 mL per run. Run buffer consisted of PBS and 0.02% polysorbate 20 or 20 mM NaPO4, 50 mM KCl, 0.02% polysorbate 20, and 10% acetonitrile, pH 7, at a flow rate of 0.4 mL / min. Elution was monitored by UV absorbance in the 210–400 nm range, and chromatograms were extracted at 280 nm. Peak integration was performed using Empower 3 software.

[0383] Figure 10 UPLC-SEC profiles of representative engineered heterodimer antibodies and representative wild-type heterodimer antibodies are shown. In most cases, engineered heterodimer antibodies exhibit UPLC-SEC profiles similar to their corresponding wild-type heterodimer antibodies, with average monomer percentages of 99.18%, 98.70%, and 98.77% for D3H44 / trastuzumab, D3H44 / cetuximab, and trastuzumab / cetuximab, respectively (see Tables 31a, 31b, and 31c).

[0384] While the invention has been particularly shown and described in conjunction with preferred embodiments and various alternative embodiments, those skilled in the art will understand that various changes may be made to its form and details without departing from the spirit and scope of the invention.

[0385] All references, published patents, patent publications and serial numbers disclosed in this article are incorporated herein by reference in their entirety for all purposes.

[0386] sheet

[0387]

[0388]

[0389]

[0390] Table 1: Key Criteria for the Fab Model

[0391]

[0392] Table 2: Hotspot amino acid positions at the interface between the heavy and light chains in D3H44 (a typical Fab contains a κ light chain).

[0393] Heavy chain* Light chain* V37 Y36 Q39 Q38 L45 P44 W47 L89 F100 F98 W103 F116 L124 F118 A139 V133 F174 L135

[0394] *Kabat number

[0395]

[0396]

[0397]

[0398]

[0399]

[0400]

[0401]

[0402]

[0403]

[0404]

[0405]

[0406]

[0407]

[0408]

[0409]

[0410]

[0411]

[0412]

[0413]

[0414]

[0415]

[0416]

[0417]

[0418] Table 4. LCCA designs with modifications of one immunoglobulin heavy chain and / or two immunoglobulin light chains, wherein H1 preferentially pairs with L1.

[0419]

[0420]

[0421]

[0422]

[0423]

[0424]

[0425]

[0426]

[0427]

[0428]

[0429]

[0430] *Kabat number; WT refers to wild-type immunoglobulin chains without amino acid mutations.

[0431] **Assign a set number or "unique identifier" to each unique H1, L1, and L2 mutation set (LCCA form).

[0432]

[0433]

[0434]

[0435]

[0436]

[0437]

[0438]

[0439]

[0440]

[0441]

[0442]

[0443]

[0444]

[0445]

[0446]

[0447]

[0448]

[0449]

[0450]

[0451]

[0452]

[0453]

[0454]

[0455]

[0456]

[0457]

[0458]

[0459]

[0460]

[0461]

[0462]

[0463]

[0464]

[0465]

[0466]

[0467] Table 11: H1:L1:L2 DNA ratio used for light chain competition analysis and testing

[0468]

[0469] ^Additional DNA: AKTdd pTT22 refers to a vector containing a constitutively active protein kinase B mutation (dominant-positive AKT mutation); ssDNA refers to salmon sperm DNA.

[0470] Table 12. Performance, stability, and antigen binding evaluation of LCCA-designed heterodimers, listed in descending order of DSF values.

[0471]

[0472]

[0473]

[0474]

[0475]

[0476]

[0477]

[0478]

[0479]

[0480]

[0481]

[0482]

[0483]

[0484]

[0485]

[0486] * indicates an estimate of other Fab heterodimers derived from L-chain tags (HA or FLAG) that differ only in the presence or absence of linkages.

[0487] The values ​​are derived from LCCA experiments of 333 (H1), 250 (L1), and 749 (L2).

[0488] The values ​​were derived from LCCA experiments of 333 (H1), 100 (L1), and 899 (L2).

[0489] ND indicates that the data is unavailable.

[0490] Table 13a. LCCA performance of designs that meet the average performance criterion of LCCA for correctly paired: mismatched Fab heterodimers of 86:14.

[0491]

[0492]

[0493]

[0494]

[0495]

[0496]

[0497]

[0498]

[0499]

[0500]

[0501]

[0502]

[0503]

[0504] *Values ​​were obtained from LCCA experiments (performed at a 1:3 L1:L2 DNA ratio) and normalized to a 1:1 L1:L2 DNA ratio.

[0505] The values ​​were obtained from LCCA experiments (performed at a 1:9 L1:L2 DNA ratio) and normalized to a 1:1 L1:L2 DNA ratio.

[0506] ***A "unique identifier" consists of two unique identifiers that make up the LCCA along the direction of (set #H1L1L2-set #H2L2L1) or (set #H2L2L1-set #H1L1L2).

[0507] Table 13b. Stability and antigen binding assessment of designs meeting the LCCA average performance criterion of correct pairing: mismatched Fab heterodimers of 86:14.

[0508]

[0509]

[0510]

[0511]

[0512]

[0513]

[0514]

[0515]

[0516] * indicates an estimate of other Fab heterodimers derived from L-chain tags (HA or FLAG) that differ only in the presence or absence of linkages.

[0517] **A "unique identifier" consists of two unique identifiers that make up the LCCA along either the direction of (set #H1L1L2-set #H2L2L1) or (set #H2L2L1-set #H1L1L2).

[0518] Table 14a. Performance below correct pairing: LCCA performance of designs with an average LCCA standard of 86:14 mismatched Fab heterodimers.

[0519]

[0520]

[0521]

[0522]

[0523]

[0524] *Values ​​were obtained from LCCA experiments (performed at a 1:3 L1:L2 DNA ratio) and normalized to a 1:1 L1:L2 DNA ratio.

[0525] The values ​​were obtained from LCCA experiments (performed at a 1:9 L1:L2 DNA ratio) and normalized to a 1:1 L1:L2 DNA ratio.

[0526] ***A "unique identifier" consists of two unique identifiers that make up the LCCA along the direction of (set #H1L1L2-set #H2L2L1) or (set #H2L2L1-set #H1L1L2).

[0527]

[0528]

[0529]

[0530]

[0531]

[0532]

[0533]

[0534]

[0535]

[0536]

[0537]

[0538]

[0539]

[0540]

[0541]

[0542]

[0543]

[0544]

[0545]

[0546]

[0547]

[0548]

[0549]

[0550]

[0551]

[0552]

[0553]

[0554]

[0555]

[0556]

[0557]

[0558]

[0559]

[0560]

[0561]

[0562]

[0563]

[0564]

[0565]

[0566]

[0567]

[0568]

[0569]

[0570]

[0571]

[0572]

[0573]

[0574]

[0575]

[0576]

[0577]

[0578]

[0579]

[0580]

[0581]

[0582]

[0583]

[0584]

[0585]

[0586]

[0587]

[0588]

[0589]

[0590]

[0591]

[0592]

[0593]

[0594]

[0595]

[0596]

[0597]

[0598]

[0599]

[0600]

[0601]

[0602]

[0603]

[0604]

[0605]

[0606]

[0607]

[0608]

[0609]

[0610]

[0611]

[0612]

[0613]

[0614]

[0615]

Claims

1. A separate antigen-binding polypeptide construct comprising at least a first heterodimer and a second heterodimer, The first heterodimer comprises a first human or humanized immunoglobulin heavy chain G polypeptide sequence, namely H1, and a first human or humanized immunoglobulin k light chain polypeptide sequence, namely L1; and the second heterodimer comprises a second human or humanized immunoglobulin heavy chain G polypeptide sequence, namely H2, and a second human or humanized immunoglobulin k light chain polypeptide sequence, namely L2, wherein... At least one of the H1 or L1 sequences of the first heterodimer is different from the corresponding H2 or L2 sequence of the second heterodimer, and wherein H1 and H2 each contain at least a heavy-chain variable structural domain (V H (structural domain) and heavy-chain constant structural domain C H1 Structural domain, wherein C H1 The domain is a variant of the sequence shown by residues A118–V215 of SEQ ID NO: 4; L1 and L2 each contain at least a light chain variable structural domain (V L (structural domain) and light chain constant structural domain (C L (structural domain), wherein C L The domain is a variant of the sequence shown by residues R108–C214 of SEQ ID NO: 1; and Wherein the C of H1 and H2 H1 The structural domains and the C of L1 and L2 L The domain contains a set of amino acid modifications that promote preferential pairing of H1 with L1 compared to L2, and promote preferential pairing of H2 with L2 compared to L1, wherein: A) i) The amino acid substitutions in H1 are 139C, 143E, 145T, and 179E; the amino acid substitutions in L1 are 116C, 124R, and 178R; the amino acid substitution in H2 is 179K; and the amino acid substitutions in L2 are 124E, 160E, and 180E; or ii) The amino acid substitutions in H1 are 143E, 145T, and 179D; the amino acid substitutions in L1 are 124R and 178K; the amino acid substitution in H2 is 179K; and the amino acid substitutions in L2 are 124E, 160E, and 180E; or B) i) The amino acid substitutions in H1 are 124W, 143E, 145T and 179E; the amino acid substitutions in L1 are 124R, 133A, 176T and 178R; the amino acid substitutions in H2 are 124A, 143F and 179K; and the amino acid substitutions in L2 are 124E, 133W, 176T, 178L and 180E. ii) The amino acid substitutions in H1 are 124W, 143E, 145T and 179E; the amino acid substitutions in L1 are 124K, 133A, 176T and 178R; the amino acid substitutions in H2 are 124A, 143F and 179K; and the amino acid substitutions in L2 are 124E, 133W, 176T, 178L and 180E. iii) The amino acid substitutions in H1 are 124W, 143E, 145T and 179E; the amino acid substitutions in L1 are 131K, 133A, 176T and 178L; the amino acid substitutions in H2 are 124A and 179K; and the amino acid substitutions in L2 are 124E, 133W, 176T, 178L and 180E. iv) The amino acid substitutions in H1 are 124W, 143E, 145T and 179E; the amino acid substitutions in L1 are 131K, 133A, 176T and 178L; the amino acid substitutions in H2 are 124A, 143F and 179K; and the amino acid substitutions in L2 are 124E, 133W, 176T, 178L and 180E. The amino acid substitutions in H1 are 124W, 143E, 145T and 179E; the amino acid substitutions in L1 are 124R, 133A, 176T and 178R; the amino acid substitutions in H2 are 124A, 143F and 179K; and the amino acid substitutions in L2 are 124E, 133W, 176T, 178E and 180E. vi) The amino acid substitutions in H1 are 124W, 143E, 145T and 179E; the amino acid substitutions in L1 are 124K, 133A, 176T and 178R; the amino acid substitutions in H2 are 124A, 143F and 179K; and the amino acid substitutions in L2 are 124E, 133W, 176T, 178E and 180E. The amino acid residues are numbered according to the Kabat numbering system.

2. The construct according to claim 1, wherein when H1, H2, L1 and L2 are co-expressed in cells, or when H1, H2, L1 and L2 are co-expressed in a cell-free expression system, or when H1, H2, L1 and L2 are co-generated, H1 preferentially pairs with L1 compared to L2, and H2 preferentially pairs with L2 compared to L1.

3. The construct according to claim 2, wherein when both L1 and L2 are co-expressed with at least one of H1 and H2, the relative pairing of the modified H1-L1 or H2-L2 heterodimer pair is greater than the respective relative pairing observed in the corresponding H1-L1 or H2-L2 heterodimer pair without the amino acid modification group.

4. The construct of claim 1, wherein the construct further comprises Fc, the Fc comprising a first CH3 sequence and a second CH3 sequence, wherein the first CH3 sequence is connected to the first heterodimer via or without one or more connectors, and the second CH3 sequence is connected to the second heterodimer via or without one or more connectors.

5. The construct according to claim 4, wherein the Fc is a human Fc.

6. The construct according to claim 5, wherein the Fc is a heterodimer Fc.

7. The construct of claim 4, wherein the Fc comprises one or more modifications of at least one of the CH3 sequences, the modifications contributing to the formation of a heterodimer Fc with stability equivalent to that of the wild-type homodimer Fc.

8. The construct of claim 7, wherein the Fc comprises: i) A heterodimer IgG1 Fc having modification 351Y_405A_407V in the first Fc polypeptide and modification 366L_392M_394W in the second Fc polypeptide; ii) Heterodimeric IgG1 Fc having modification 351Y_405A_407V in the first Fc polypeptide and modification 366L_392L_394W in the second Fc polypeptide; iii) Heterodimeric IgG1 Fc having the modification 350V_351Y_405A_407V in the first Fc polypeptide and the modification 350V_366L_392L_394W in the second Fc polypeptide; iv) Heterodimeric IgG1 Fc having the modification 350V_351Y_405A_407V in the first Fc polypeptide and the modification 350V_366L_392M_394W in the second Fc polypeptide; or v) A heterodimer IgG1 Fc having the modification 350V_351Y_400E_405A_407V in the first Fc polypeptide and the modification 350V_366L_390R_392M_394W in the second Fc polypeptide; The residues in the IgG1 Fc are numbered according to the EU numbering system.

9. The construct of claim 4, wherein the Fc further comprises at least one CH2 sequence.

10. The construct of claim 9, wherein the CH2 sequence of the Fc comprises one or more modifications.

11. The construct of claim 10, wherein the Fc comprises one or more modifications to facilitate selective binding of the Fc-γ receptor.

12. The construct of claim 4, wherein the Fc is coupled to the heterodimer via one or more connectors, or wherein the Fc is coupled to H1 and H2 via one or more connectors.

13. The construct of claim 12, wherein the one or more adapters are one or more peptide adapters.

14. The construct of claim 12, wherein the one or more connectors comprise one or more antibody hinge regions.

15. The construct of claim 12, wherein the one or more connectors comprise one or more IgG1 hinge regions.

16. The construct of claim 13, wherein the one or more connectors comprise one or more modifications.

17. The construct of claim 16, wherein the one or more modifications contribute to the selective binding of the Fc-γ receptor.

18. The construct according to any one of claims 1-17, wherein the construct is multispecific or bispecific.

19. The construct according to any one of claims 1-17, wherein the construct is multivalent or bivalent.

20. The construct according to any one of claims 1-17, wherein the construct is conjugated to a therapeutic or pharmaceutical portion.

21. An isolated polynucleotide or isolated polynucleotide group comprising at least one construct sequence encoding any one of claims 1-19.

22. The isolated polynucleotide or isolated polynucleotide group according to claim 21, wherein the polynucleotide or polynucleotide group is cDNA.

23. A vector or group of vectors comprising one or more of the polynucleotides or groups of polynucleotides as claimed in claim 21 or 22.

24. The vector or vector group according to claim 23, wherein the vector is selected from plasmids, polycistronic vectors, viral vectors, and non-free mammalian vectors.

25. An isolated cell comprising a polynucleotide or polynucleotide group according to claim 21 or 22, or a vector or vector group according to claim 23 or 24.

26. The isolated cells according to claim 25, wherein the cells are yeast cells, bacterial cells, hybridomas, Chinese hamster ovary (CHO) cells, or HEK293 cells.

27. A pharmaceutical composition comprising a construct according to any one of claims 1-20 and a pharmaceutically acceptable carrier.

28. The pharmaceutical composition of claim 27 further comprises one or more substances selected from buffer solutions, antioxidants, low molecular weight molecules, pharmaceuticals, proteins, amino acids, sugars, lipids, chelating agents, stabilizers, and excipients.

29. A method for obtaining the construct according to any one of claims 1 to 20 from a host cell culture, the method comprising the steps of: (a) Obtaining a host cell culture comprising at least one host cell, said host cell containing one or more nucleic acid sequences encoding the construct; and (b) The construct is recovered from the host cell culture.

30. A method for obtaining a construct according to any one of claims 1 to 20, the method comprising the following steps: (a) Obtain H1, L1, H2, or L2; (b) H1 is allowed to pair with L1 with priority over L2, and H2 is allowed to pair with L2 with priority over L1; as well as (c) Obtain the constructed object.

Citation Information

Patent Citations

  • Stable Heterodimeric Antibody Design with Mutations in the Fc Domain

    US20120149876A1

  • Antigen-binding molecule having regulated conjugation between heavy-chain and light-chain

    US20140370020A1

  • Method for carrying out surface plasmon resonance measurement and sensor for use in the method

    US5322798A

  • Method and apparatus for detecting the presence and / or concentration of biomolecules

    US5341215A

  • Analytical apparatus

    US6268125B1