HOMODIMER-TYPE BISPECIFIC ANTIBODY AGAINST HER2 AND CD3 AND USE THEREOF

Abstract

A tetravalent, homodimer-type bispecific antibody molecule that simultaneously targets immune effector cell antigen CD3 and human epidermal growth factor receptor 2 (Her2); the bispecific antibody molecule comprises, from in sequence from N-terminus to C-terminus, a first single-chain Fv capable of specifically binding to Her2, a second single-chain Fv capable of specifically binding to CD3, and an Fc fragment; the first and second single-chain Fv are connected by means of a connection peptide, and the second single-chain Fv is connected to the Fc directly fragment or is connected by means of a connection peptide; the Fc fragment does not have effector functions such as CDC, ADCC and ADCP. The bispecific antibody may significantly inhibit or kill tumor cells, and has controlled toxic side effects that may be caused by excessive activation of effector cells. The maximum safe starting dose in preclinical toxicology evaluation tests is significantly higher than other doses having the same target, and no systemic immunotoxicity occurs, suggesting that the drug administration safety window for the bispecific antibody is wide; in addition, said bispecific antibody is a homodimer that does not experience the problem of heavy chain and light chain mismatching; the steps of purification are simple and efficient, expression is high, and the physicochemical and in vivo stability of the antibody are significantly improved.

Claims

1. A bispecific antibody, which is a tetravalent homodimer formed by two identical polypeptide chains that bind to each other by a covalent bond, wherein each of the polypeptide chains comprises a first single-chain Fv that specifically binds to human epidermal growth factor receptor 2, a second single-chain Fv that specifically binds to effector cell antigen CD3, and an Fc fragment in sequence from N-terminus to C-terminus; wherein the first single-chain Fv is linked to the second single-chain Fv by a linker peptide, the second single-chain Fv is linked to the Fc fragment directly or by a linker peptide, and the Fc fragment has no effector functions comprising CDC, ADCC, and ADCP.

2. The bispecific antibody according to claim 1, wherein the first single-chain Fv comprises a VH domain and a VL domain that are linked by a linker peptide which has an amino acid sequence of (GGGGX).sub.n, wherein X comprises Ser or Ala, and n is a natural number from 1 to 5.

3. The bispecific antibody according to claim 1, wherein the first single-chain Fv comprises a VH domain and a VL domain selected from the group consisting of: (i) a VH domain that contains HCDR1, HCDR2 and HCDR3 as shown in SEQ ID NOs: 9, 10 and 11, respectively or having sequences that are at least 80%, 85%, 90%, 92%, 95%, 97%, 98%, 99% or more similar to or have one or more amino acid substitutions than any of SEQ ID NOs: 9, 10 and 11; and a VL domain that contains LCDR1, LCDR2 and LCDR3 as shown in SEQ ID NOs: 12, 13 and 14, respectively or having sequences that are at least 80%, 85%, 90%, 92%, 95%, 97%, 98%, 99% or more similar to or have one or more amino acid substitutions than any of SEQ ID NOs: 12, 13 and 14; (ii) a VH domain that contains HCDR1, HCDR2 and HCDR3 as shown in SEQ ID NOs: 17,18 and 19, respectively or having sequences that are at least 80%, 85%, 90%, 92%, 95%, 97%, 98%, 99% or more similar to or have one or more amino acid substitutions than any of SEQ ID NOs: 17, 18 and 19; and a VL domain that contains LCDR1, LCDR2 and LCDR3 as shown in SEQ ID NOs: 20, 21 and 22, respectively or having sequences that are at least 80%, 85%, 90%, 92%, 95%, 97%, 98%, 99% or more similar to or have one or more amino acid substitutions than any of SEQ ID NOs: 20, 21 and 22; and (iii) a VH domain that contains HCDR1, HCDR2 and HCDR3 as shown in SEQ ID NOs: 25, 26 and 27, respectively or having sequences that are at least 80%, 85%, 90%, 92%, 95%, 97%, 98%, 99% or more similar to or have one or more amino acid substitutions than any of SEQ ID NOs: 25, 26 and 27; and a VL domain that contains LCDR1, LCDR2 and LCDR3 as shown in SEQ ID NOs: 28, 29 and 30, respectively or having sequences that are at least 80%, 85%, 90%, 92%, 95%, 97%, 98%, 99% or more similar to or have one or more amino acid substitutions than any of SEQ ID NOs: 28, 29 and 30.

4. The bispecific antibody according to claim 1, wherein the first single-chain Fv comprises a VH domain and a VL domain selected from the group consisting of: (i) a VH domain comprising an amino acid sequence as shown in SEQ ID NO: 15 or having a sequence that is at least 80%, 85%, 90%, 92%, 95%, 97%, 98%, 99% or more similar to or has one or more amino acid substitutions than SEQ ID NO: 15; and a VL domain comprising an amino acid sequence as shown in SEQ ID NO: 16 or having a sequence that is at least 80%, 85%, 90%, 92%, 95%, 97%, 98%, 99% or more similar to or has one or more amino acid substitutions than SEQ ID NO: 16; (ii) a VH domain comprising an amino acid sequence as shown in SEQ ID NO: 23 or having a sequence that is at least 80%, 85%, 90%, 92%, 95%, 97%, 98%, 99% or more similar to or has one or more amino acid substitutions than SEQ ID NO: 23; and a VL domain comprising an amino acid sequence as shown in SEQ ID NO: 24 or having a sequence that is at least 80%, 85%, 90%, 92%, 95%, 97%, 98%, 99% or more similar to or has one or more amino acid substitutions than SEQ ID NO: 24; and (iii) a VH domain comprising an amino acid sequence as shown in SEQ ID NO: 31 or having a sequence that is at least 80%, 85%, 90%, 92%, 95%, 97%, 98%, 99% or more similar to or has one or more amino acid substitutions than SEQ ID NO: 31; and a VL domain comprising an amino acid sequence as shown in SEQ ID NO: 32 or having a sequence that is at least 80%, 85%, 90%, 92%, 95%, 97%, 98%, 99% or more similar to or has one or more amino acid substitutions than SEQ ID NO: 32.

5. The bispecific antibody according to claim 1, wherein the second single-chain Fv comprises a VH domain and a VL domain that are linked by a linker peptide which has an amino acid sequence of (GGGGX).sub.n, wherein X comprises Ser or Ala, preferably Ser, and n is a natural number from 1 to 5, preferably 3, wherein the second single-chain Fv binds to an effector cell an an EC.sub.50 value greater than about 50 nM, or greater than 100 nM, or greater than 300 nM, or greater than 500 nM in an in vitro binding affinity assay; and wherein, the second single-chain Fv of the bispecific antibody is capable of binding to human CD3 and specifically binding to CD3 of a cynomolgus monkey or a rhesus monkey.

6. (canceled)

7. The bispecific antibody according to claim 1, wherein the VH domain of the second single-chain Fv contains HCDR1, HCDR2 and HCDR3 as shown in SEQ ID NOs: 33, 34 and 35, respectively or having sequences that are at least 80%, 85%, 90%, 92%, 95%, 97%, 98%, 99% or more similar to or have one or more amino acid substitutions than SEQ ID NOs: 33, 34 and 35; and the VL domain of the second single-chain Fv contains LCDR1, LCDR2 and LCDR3 as shown in SEQ ID NOs: 36, 37 and 38, respectively or having sequences that are at least 80%, 85%, 90%, 92%, 95%, 97%, 98%, 99% or more similar to or have one or more amino acid substitutions than SEQ ID NOs: 36, 37 and 38; the VH of the second single-chain Fv contains HCDR1, HCDR2 and HCDR3 as shown in SEQ ID NOs: 41, 42 and 43, respectively or having sequences that are at least 80%, 85%, 90%, 92%, 95%, 97%, 98%, 99% or more similar to or have one or more amino acid substitutions than SEQ ID NOs: 41, 42 and 43; and the VL domain of the second single-chain Fv contains LCDR1, LCDR2 and LCDR3 as shown in SEQ ID NOs: 44, 45 and 46, respectively or having sequences that are at least 80%, 85%, 90%, 92%, 95%, 97%, 98%, 99% or more similar to or have one or more amino acid subsitutions than SEQ ID NOs: 44, 45 and 46.

8. (canceled)

9. The bispecific antibody according to claim 7, wherein the second single-chain Fv specifically binds to CD3; the VH domain of the second single-chain Fv contains an amino acid sequence as shown in SEQ ID NO: 39 or has a sequence that is at least 80%, 85%, 90%, 92%, 95%, 97%, 98%, 99% or more similar to or has one or more amino acid substitutions than SEQ ID NO: 39; and the VL domain of the second single-chain Fv contains an amino acid sequence as shown in SEQ ID NO: 40 or has a sequence that is at least 80%, 85%, 90%, 92%, 95%, 97%, 98%, 99% or more similar to or has one or more amino acid substitutions SEQ ID NO: 40; the second single-chain FV specificallybinds to CD3; the VH domain of the second single-chain Fv contains an amino acid sequence as shown in SEQ ID NO: 47 or has a sequence that is at least 80%, 85%, 90%, 92%, 95%, 97%, 98%, 99% or more similar to or has one or more amino acid substitutions than SEQ ID NO: 47; and the VL domain of the second single-chain Fv contains an amino acid sequence as shown in SEQ ID NO: 48 or has a sequence that is at least 80%, 85%, 90%, 92%, 95%, 97%, 98%, 99% or more similar to or has one or more amino acid substituions than SEQ ID NO: 48.

10. (canceled)

11. The bispecific antibody according to claim 1, wherein the linker peptide that links the first single-chain Fv to the second single-chain Fv consists of a flexible peptide and a rigid peptide; wherein the flexible peptide comprises two or more amino acids, and preferably selected from the following amino acids: Gly(G), Ser(S), Ala(A) and Thr(T); more preferably, the flexible peptide comprises G and S residues; most preferably, an amino acid composition structure of the flexible peptide has a general formula of G.sub.xS.sub.y(GGGGS).sub.z, wherein x, y and z are integers greater than or equal to 0 and x+y+z≥1; the rigid peptide is derived from a full-length sequence consisting of amino acids 118 to 145 at carboxyl terminus of natural human chorionic gonadotropin β-subunit as shown in SEQ ID NO: 49 or a truncated fragment thereof; preferably, the rigid peptide comprises SSSSKAPPPS.

12. The bispecific antibody according to claim 11, wherein the linker peptide comprises an amino acid sequence as shown in SEQ ID NO: 50, 51, 52 or 53.

13. The bispecific antibody according to claim 1, wherein the linker peptide that links the Fc fragment to the second single-chain Fv comprises 1-20 amino acids, and preferably selected from the following amino acids: Gly(G), Ser(S), Ala(A) and Thr(T); preferably Gly (G) and Ser (S); more preferably, the linker peptide consists of (GGGGS)n, wherein n=1, 2, 3 or 4.

14. The bispecific antibody according to claim 1, wherein the Fc fragment comprises a hinge region, a CH2 domain and a CH3 domain from a human immunoglobulin heavy chain constant region; preferably, the Fc fragment is selected from heavy chain constant regions of human IgG1, IgG2, IgG3, IgG4, IgM, IgA1, IgA2, IgD and IgE; more preferably, the Fc fragment is selected from heavy chain constant regions of human IgG1, IgG2, IgG3 and IgG4; further preferably, the Fc fragment is selected from a heavy chain constant region of human IgG1 or IgG4; and the Fc fragment has one or more amino acid substitutions, deletions or additions with reduced or eliminated effector functions comprising ADCP, ADCC and CDC effects than a natural sequence from which the Fc fragment is derived.

15. (canceled)

16. The bispecific antibody according to claim 15, wherein the Fc fragment comprises amino acid substitutions L234A/L235A/P331S that are determined according to an EU numbering system.

17. The bispecific antibody according to claim 16, wherein the Fc fragment further comprises amino acid substitutions, deletions or additions with one or more of the following properties: (i) enhanced binding affinity to a neonatal receptor (FcRn); (ii) reduced or eliminated glycosylation; and (iii) reduced or eliminated charge heterogeneity.

18. The bispecific antibody according to claim 17, wherein the Fc fragment further comprises one or more of amino acid substitutions, deletions or additions as follow: (i) amino acid substitutions M428L, T250Q/M428L, M428L/N434S or M252Y/S254T/T256E determined according to the EU numbering system; (ii) an amino acid substitution N297A determined according to the EU numbering system; and (iii) an amino acid deletion K447 determined according to the EU numbering system.

19. The bispecific antibody according to claim 18, wherein the Fc fragment has an amino acid sequence as shown in SEQ ID NO: 55 that has six amino acid substitutions or replacements L234A/L235A/N297A/P331S/T250Q/M428L determined according to the EU numbering system and a deleted or removed K447 determined according to the EU numbering system compared to the natural sequence from which the Fc fragment is derived.

20. The bispecific antibody according to claim 1, wherein the bispecific antibody comprises an amino acid sequence as follows: (i) a sequence as shown in SEQ ID NO: 8; (ii) a sequence with one or more substitutions, deletions or additions (such as 1, 2, 3, 4 or 5 substitutions, deletions or additions) than the sequence as shown in SEQ ID NO: 8; or (iii) a sequence with at least 80%, at least 85%, at least 90%, at least 91%, at least 92%, at least 93%, at least 94%, at least 95%, at least 96%, at least 97%, at least 98%, at least 99% or 100% sequence identity relative to the sequence as shown in SEQ ID NO: 8.

21. A DNA molecule encoding the bispecific antibody according to claim 1.

22. The DNA molecule according to claim 21, which has a nucleotide sequence as shown in SEQ ID NO: 56.

23-24. (canceled)

25. A pharmaceutical composition, comprising the bispecific antibody according to claim 1 and a pharmaceutically acceptable excipient, carrier or diluent, preferably, the pharmaceutical composition is a solution formulation, preferably, the pharmaceutical composition further comprises a pH regulator, a stabilizer, and a surfactant: preferably, the pH regulator is a citrate buffer or a histidine buffer, the stabilizer is sucrose, and the surfactant is tween˜80: more preferably, the formulation comprises 0.5 mg/mL of the bispecific antibody, 20 mM of citrate or histidine, 8% (w/v) of sucrose and 0.02% (w/v) of PS80: the formulation has a pH of 5.5.

26. (canceled)

27. A method for preparing the bispecific antibody according to claim 1, comprising: (a) obtaining a fusion gene of the bispecific antibody, and constructing an expression vector of the bispecific antibody; (b) transfecting the expression vector into a host cell by a genetic engineering method; (c) culturing the host cell under conditions that allow the bispecific antibody to be generated; (d) separating and purifying the bispecific antibody; wherein the expression vector in step (a) is one or more selected from a plasmid, a bacterium and a virus; preferably, the expression vector is an pCDNA3.4 vector; wherein the host cell into which the constructed vector is transfected by the genetic engineering method in step (b) comprises a prokaryotic cell, a yeast cell or a mammalian cell, such as a CHO cell, an NS0 cell or another mammalian cell, preferably a CHO cell; and wherein the bispecific antibody is separated and purified in step (d) by a conventional immunoglobulin purification method comprising protein A affinity chromatography and ion exchange, hydrophobic chromatography or molecular sieve.

28-30. (canceled)

31. A method for treating, preventing or ameliorating a tumor or an immune disorder or disease in a patient or subject, which comprises administering a therapeutically effective amount of the bispecific antibody of claim 1 to the patient or subject, wherein the tumor comprises breast cancer, prostate cancer, non-small cell lung cancer, bladder cancer, ovarian cancer, gastric cancer, colorectal cancer, esophageal cancer, head and neck squamous cell carcinoma, cervial cancer, pancreatic cancer, testicular cancer, malignant melanoma, and soft tissue cancer.

Description

BRIEF DESCRIPTION OF DRAWINGS

[0231] FIG. 1-1 illustrates configurations of bispecific antibodies AB7K, AB7K4, AB7K5, AB7K6, AB7K7 and AB7K8 as shown in a, b, c, d, e and f, respectively.

[0232] FIG. 1-2 illustrates an expression plasmid map of bispecific antibody AB7K7. The expression plasmid has a full length of 9293 bp and contains nine major gene fragments which are (1) an hCMV promoter, (2) target genes, (3) EMCV IRES, (4) mDHFR screening gene, (5) a Syn discontinuation sequence, (6) an SV40 promoter, (7) Kalamycin resistance gene; (8) an SV40 termination sequence, and (9) a PUC replicon.

[0233] FIG. 1-3 illustrates SEC-HPLC test results of purified samples of AB7K7.

[0234] FIG. 1-4 illustrates SDS-PAGE electrophoresis results of purified samples of AB7K7.

[0235] FIG. 1-5 illustrates SDS-PAGE results of AB7K7 in acceleration experiments at 25° C.

[0236] FIG. 1-6 illustrates SDS-PAGE results of AB7K7 in freeze-thaw experiments.

[0237] FIG. 2-1 illustrates abilities, detected through an FACS, of bispecific antibodies AB7K and AB7K4 to bind to tumor cells BT474.

[0238] FIG. 2-2 illustrates abilities, detected through an FACS, of bispecific antibodies AB7K and AB7K5 to bind to tumor cells BT474.

[0239] FIG. 2-3 illustrates abilities, detected through an FACS, of bispecific antibodies AB7K and AB7K6 to bind to tumor cells BT474.

[0240] FIG. 2-4 illustrates abilities, detected through an FACS, of bispecific antibodies AB7K and AB7K7 to bind to cancer cells BT474.

[0241] FIG. 2-5 illustrates an ability, detected through an FACS, of a bispecific antibody AB7K8 to bind to tumor cells BT474.

[0242] FIG. 2-6 illustrates abilities, detected through an FACS, of bispecific antibodies AB7K and AB7K4 to bind to effector cells CIK.

[0243] FIG. 2-7 illustrates abilities, detected through an FACS, of bispecific antibodies AB7K and AB7K5 to bind to effector cells CIK.

[0244] FIG. 2-8 illustrates an ability, detected through an FACS, of a bispecific antibody AB7K6 to bind to effector cells CIK.

[0245] FIG. 2-9 illustrates abilities, detected through an FACS, of bispecific antibodies AB7K and AB7K7 to bind to effector cells CIK.

[0246] FIG. 2-10 illustrates an ability, detected through an FACS, of a bispecific antibody AB7K8 to bind to effector cells CIK.

[0247] FIG. 2-11 illustrates an ability, detected through an FACS, of a bispecific antibody AB7K to bind to T cells of cynomolgus monkeys.

[0248] FIG. 2-12 illustrates abilities, detected through an ELISA, of five Anti-Her2×CD3 bispecific antibodies to bind to CD3 molecules and Her2 molecules.

[0249] FIG. 2-13 illustrates abilities, detected through a microplate reader, of five Anti-Her2×CD3 bispecific antibodies to activate Jurkat T cells of a reporter gene cell line.

[0250] FIG. 2-14 illustrates structural modeling of a CTP linker and anti-CD3 scFv VH.

[0251] FIG. 2-15 illustrates structural modeling of a GS linker and anti-CD3 scFv VH.

[0252] FIG. 2-16 illustrates a molecular docking model of anti-CD3 scFv and a CD3 epsilon chain.

[0253] FIG. 3-1 illustrates in vivo anti-tumor effects of bispecific antibodies AB7K4 and AB7K7 in a transplanted tumor model in which human CIK cells and HCC1954 cells were co-inoculated subcutaneously in NCG mice.

[0254] FIG. 3-2 illustrates an in vivo anti-tumor effect of a bispecific antibody AB7K7 in a transplanted tumor model in which human CIK cells and human breast cancer cells HCC1954 were co-inoculated subcutaneously in NPG mice.

[0255] FIG. 3-3 illustrates in vivo anti-tumor effects of bispecific antibodies AB7K7 and AB7K8 at different administration frequencies in a transplanted tumor model in which human CIK cells and human breast cancer cells HCC1954 were co-inoculated subcutaneously in NPG mice.

[0256] FIG. 3-4 illustrates an in vivo anti-tumor effect of a bispecific antibody AB7K7 in a transplanted tumor model in which human CIK cells and SK-OV-3 cells were co-inoculated subcutaneously in NPG mice.

[0257] FIG. 3-5 illustrates an in vivo anti-tumor effect of a bispecific antibody AB7K7 in a transplanted tumor model in which human CIK cells and HT-29 cells were co-inoculated subcutaneously in NPG mice.

[0258] FIG. 3-6 illustrates an in vivo anti-tumor effect of a bispecific antibody AB7K7 in a transplanted tumor model in which human breast cancer cells HCC1954 were inoculated subcutaneously in NPG mice with immunology reconstituted by CD34.

[0259] FIG. 3-7 illustrates an in vivo anti-tumor effect of a bispecific antibody AB7K7 in a transplanted tumor model in which human breast cancer cells HCC1954 were inoculated in NPG mice with immunology reconstituted by PBMC.

[0260] FIG. 4-1 illustrates anti-tumor effects of bispecific antibodies AB7K4 and AB7K7 in a transplanted tumor model in which human CIK cells and human Burkkit's lymphoma Raji cells were co-inoculated subcutaneously in NCG mice.

[0261] FIG. 4-2 illustrates an anti-tumor effect of a bispecific antibody AB7K7 in a transplanted tumor model in which human breast cancer cells HCC1954 were inoculated separately in NPG mice.

[0262] FIG. 4-3 illustrates changes in weight of normal cynomolgus monkeys administered with bispecific antibodies AB7K7 and AB7K8 multiple times.

[0263] FIG. 5-1 illustrates concentration-time curves of a bispecific antibody AB7K7 in SD rats by two ELISA methods.

[0264] FIG. 5-2 illustrates concentration-time curves of a bispecific antibody AB7K8 in SD rats by two ELISA methods.

[0265] FIG. 5-3 illustrates concentration-time curves of bispecific antibodies AB7K7 and AB7K8 in cynomolgus monkeys.

[0266] FIG. 5-4 illustrates abilities of bispecific antibodies AB7K, AB7K5 and AB7K7 to bind to FcRn, which are determined at a pH of 6.0.

[0267] FIG. 5-5 illustrates abilities of bispecific antibodies AB7K, AB7K5 and AB7K7 to bind to FcRn, which are determined at a pH of 7.0.

DETAILED DESCRIPTION

[0268] The present disclosure is further described through examples which should not be construed as further limitations. All drawings, all reference documents, and the contents of patents and published patent applications cited in the entire application are expressly incorporated herein by reference.

Example 1 Design and Preparation of Anti-Her2×CD3 Bispecific Antibodies Having Different Structures

1.1 Design of Bispecific Antibodies Having Different Structures

[0269] In order to screen bispecific antibodies having suitable configuration, bispecific antibodies having six different configurations were designed for Her2 and CD3, among which AB7K5, AB7K6, and AB7K8 are single-chain bivalent bispecific antibodies while AB7K, AB7K4, and AB7K7 are double-chain tetravalent bispecific antibodies (see FIG. 1-1), where only AB7K8 is free of Fc fragments. Specifically, the configuration of the bispecific antibodies with the above four configurations and their composition from the N-terminus to the C-terminus as well as their amino acid sequence numbers are shown in Table 1-1. The specific structural composition properties of the six bispecific antibodies are described below:

[0270] Bispecific antibody AB7K consists of an anti-Her2 full-length antibody whose two heavy chains are each linked at the C-terminus to an anti-CD3 scFv domain by a linker peptide (L1). For the amino acid sequence of the intact antibody against Her2 contained in AB7K, reference is made to the sequence of monoclonal antibody Herceptin® (IMGT database INN 7637), wherein AB7K contains an Fc fragment from human IgG1 and has D356E/L358M mutations (EU numbering). The linker peptide L1 consists of a flexible peptide and a rigid peptide, wherein the composition of the flexible peptide is GS(GGGGS)3 and the rigid peptide is SSSSKAPPPSLPSPSRLPGPSDTPILPQ, wherein the composition of the linker peptide L2 between VH and VL of the anti-CD3 scFv is (GGGGS).sub.3.

[0271] Bispecific antibody AB7K4 consists of an anti-Her2 full-length antibody whose two light chains are each linked at the C-terminus to an anti-CD3 scFv domain by a linker peptide (L1). For the amino acid sequence of the heavy chain variable region of the intact antibody against Her2 contained in AB7K4, reference is made to the available region sequence of the monoclonal antibody Herceptin®, and for the light chain amino acid sequence thereof, reference is made to the light chain amino acid sequence of the monoclonal antibody Herceptin® (IMGT database INN 7637). The AB7K4 heavy chain contains an Fc fragment from human IgG1, has multiple amino acid substitutions/replacements, which are L234A, L235A, T250Q, N297A, P331S, and M428L (EU numbering), respectively, and also has a deleted/missed K447 (EU numbering) at the C-terminus of the Fc fragment. The linker peptide L1 consists of a flexible peptide and a rigid peptide, wherein the composition of the flexible peptide is G.sub.2(GGGGS)3 and the rigid peptide is SSSSKAPPPS, wherein the composition of the linker peptide L2 between VH and VL of the anti-CD3 scFv is (GGGGS).sub.3.

[0272] Bispecific antibody AB7K5 consists of an anti-Her2 scFv, an Fc fragment, a linker peptide L2 and an anti-CD3 scFv, which are sequentially connected in series, wherein VH and VL in the anti-Her2 scFv are connected by a linker peptide L1, and VH and VL in the anti-CD3 scFv are connected by a linker peptide L3. For the amino acid sequence of the scFv against Her2 contained in AB7K5, reference is made to the available region sequence of the monoclonal antibody Herceptin®. The AB7K5 contains an Fc fragment from human IgG1 and has multiple amino acid substitutions/replacements, which are C226S, C229S, L234A, L235A, T250Q, N297A, P331S, T366R, L368H, K409T, and M428L (EU numbering), respectively. Mutations at the five sites C226S, C229S, T366R, L368H, and K409T can prevent polymerization between Fc fragments, thereby promoting the formation of a single-chain bivalent bispecific antibody. ADCC and CDC activities are removed from Fc fragments carrying the mutations L234A/L235A/P331S. The mutations T250Q/M428L can enhance the binding affinity of Fc fragments for the receptor FcRn, thereby extending the half-life. The mutation N297A avoids antibody glycosylation and loses the ability to bind FcγRs. In addition, K447 (EU numbering) at the C-terminus of the Fc fragment is deleted/missed, thereby eliminating the charge heterogeneity of the antibody. The linker peptide (L2) consists of a flexible peptide and a rigid peptide, wherein the flexible peptide is G.sub.2(GGGGS).sub.3 and the rigid peptide is SSSSKAPPPS. The composition of the linker peptides L1 and L3 inside each scFv is (GGGGS).sub.3.

[0273] Bispecific antibody AB7K6 consists of an anti-Her2 scFv, a linker peptide L2, an anti-CD3 scFv, and an Fc fragment, which are sequentially connected in series, wherein VH and VL in the anti-Her2 scFv are connected by a linker peptide L1, and VH and VL in the anti-CD3 scFv are connected by a linker peptide L3. The AB7K6 contains an Fc fragment from human IgG1 and has multiple amino acid substitutions/replacements, which are C226S, C229S, L234A, L235A, T250Q, N297A, P331S, T366R, L368H, K409T, and M428L (EU numbering), respectively. Mutations at the five sites C226S, C229S, T366R, L368H, and K409T can prevent polymerization between Fc fragments, thereby promoting the formation of a single-chain bivalent bispecific antibody. ADCC and CDC activities are removed from Fc fragments carrying the mutations L234A/L235A/P331S. The mutations T250Q/M428L can enhance the binding affinity of Fc fragments for the receptor FcRn, thereby extending the half-life. The mutation N297A avoids antibody glycosylation and loses the ability to bind FcγRs. In addition, K447 (EU numbering) at the C-terminus of the Fc fragment is deleted/missed, thereby eliminating the charge heterogeneity of the antibody. The linker peptide (L2) consists of a flexible peptide and a rigid peptide, wherein the flexible peptide is G.sub.2(GGGGS).sub.3 and the rigid peptide is SSSSKAPPPS. The composition of the linker peptides L1 and L3 inside each scFv is (GGGGS).sub.3.

[0274] Bispecific antibody AB7K7 consists of an anti-Her2 scFv, a linker peptide L2, an anti-CD3 scFv, and an Fc fragment, which are sequentially connected in series, wherein VH and VL in the anti-Her2 scFv are connected by a linker peptide L1, and VH and VL in the anti-CD3 scFv are connected by a linker peptide L3. For the amino acid sequence of the scFv against Her2 contained in AB7K7, reference is made to the available region sequence of the monoclonal antibody Herceptin®. The AB7K7 contains an Fc fragment from human IgG1, and has multiple amino acid substitutions/replacements, which were L234A, L235A, T250Q, N297A, P331S, and M428L (EU numbering), respectively, and also has a deleted/missed K447 (EU numbering) at the C-terminus of the Fc fragment. The linker peptide (L2) consists of a flexible peptide and a rigid peptide, wherein the flexible peptide is G.sub.2(GGGGS)3 and the rigid peptide is SSSSKAPPPS. The composition of the linker peptides L1 and L3 inside each scFv is (GGGGS).sub.3.

[0275] Bispecific antibody AB7K8 consists of an anti-Her2 scFv, a linker peptide L2, an anti-CD3 scFv, and a His-tag, which are sequentially connected in series, wherein VH and VL in the anti-Her2 scFv are connected by a linker peptide L1l, and VH and VL in the anti-CD3 scFv are connected by a linker peptide L3. For the amino acid sequence of the scFv against Her2 contained in AB7K8, reference is made to the available region sequence of the monoclonal antibody Herceptin®. AB7K8 is added with a His-tag at the C-terminus of the anti-CD3 scFv to facilitate antibody purification, wherein the composition of the His tag is HHHHHHHH. The linker peptide (L2) consists of a flexible peptide and a rigid peptide, wherein the flexible peptide is G.sub.2(GGGGS).sub.3 and the rigid peptide is SSSSKAPPPS. The composition of the linker peptides L1 and L3 inside each scFv is (GGGGS).sub.3.

[0276] VH and VL amino acid sequences of the anti-CD3 scFv contained in the above six bispecific antibodies are as shown in SEQ ID NO: 247 and SEQ ID NO: 248, respectively, wherein VH and VL are connected to each other by (GGGGS).sub.3. The monoclonal antibody (designated as CD3-3) specifically binds to human and cynomolgus monkey CD3 antigens and has a weak binding affinity for CD3.

TABLE-US-00001 TABLE 1-1 Bispecific antibodies with four different structures against Her2 and CD3 Composition from N-terminus to Amino acid Code Configuration C-terminus sequence No. Single-chain AB7K5 scFv-mFc-scFv [VH-L1-VL].sub.Her2-mFc-L2-[VH-L3-VL].sub.CD3 SEQ ID NO: 1 bivalent bispecific AB7K6 scFv-scFv-mFc [VH-L1-VL].sub.Her2 L2-[VH-L3-VL].sub.CD3-mFc SEQ ID NO: 2 antibody AB7K8 scFv-scFv-His [VH-L1-VL].sub.Her2-L2-[VH-L3-VL].sub.CD3-H.sub.8 SEQ ID NO: 3 tag Double-chain AB7K IgG(H)-scFv [VH-CH].sub.Her2-L1-[VH-L2-VL].sub.CD3 SEQ ID NO: 4 tetravalent [VL-CL].sub.Her2 SEQ ID NO: 5 bispecific AB7K4 IgG(L)-scFv [VH-CH].sub.Her2 SEQ ID NO: 6 antibody [VL-CL].sub.Her2-L1-[VH-L2-VL].sub.CD3 SEQ ID NO: 7 AB7K7 scFv-scFv-mFc [VH-L1-VL].sub.Her2-L2-[VH-L3-VL].sub.CD3-mFc SEQ ID NO: 8 Note: Ln in the table represents the linker peptides between different structural units, wherein n is numbered sequentially in the order of the linker peptides contained between different structural units from the N-terminus to the C-terminus of the bispecific antibody.

1.2 Construction of an Expression Vector of a Bispecific Antibody Molecule

[0277] Genes encoding the preceding five bispecific antibodies were synthesized by conventional molecular biology method, and cDNA encoding the obtained fusion genes were inserted into corresponding restriction endonuclease sites of eukaryotic expression plasmids pCMAB2M modified with PCDNA3.1. The heavy chains and light chains of AB7K and AB7K4 may be constructed into one vector or two different vectors, respectively. For example, a cDNA sequence (as shown in SEQ ID NO: 56) encoding AB7K7 was inserted into the expression plasmid shown in FIG. 1-2, which contains cytomegalovirus early promoter. The promoter is an enhancer required for the high-level expression of foreign genes in mammalian cells. The plasmid pCMAB2M also contains a selective marker so that kanamycin resistance may be present in bacteria, and G418 resistance may be present in mammalian cells. In addition, when host cells are deficient in the expression of DHFR genes, the pCMAB2M expression vector contains mouse dihydrofolate reductase (DHFR) genes so that target genes and the DHFR genes can be co-amplified in the presence of methotrexate (MTX) (see U.S. Pat. No. 4,399,216).

1.3 Expression of Bispecific Antibody Molecules

[0278] The preceding constructed expression plasmids were transfected into a mammalian host cell line to express bispecific antibodies. To maintain stable and high-level expression, the preferred host cell line is a DHFR deficient CHO-cell (see U.S. Pat. No. 4,818,679), and in this Example, the host cell was selected as the CHO-derived cell strain DXB11. A preferred transfection method is electroporation. Other methods, including calcium phosphate co-precipitation and lipofection may also be used. During electroporation, 50 μg of expression vector plasmids DNA were added to 5×10.sup.7 cells in a cuvette with a Gene Pulser Electroporator (Bio-Rad Laboratories, Hercules, Calif.) with an electric field of 300 V and capacitance of 1500 μFd. After two-day transfection, the medium was changed to a growth medium containing 0.6 mg/mL G418. Transfectants were subcloned by the limiting dilution method, and the secretion rate of each cell line was determined by ELISA. Cell strains expressing bispecific antibodies at high levels were screened.

[0279] To achieve the high-level expression of fusion proteins, DHFR genes inhibited by MTX should be used for co-amplification. The transfected fusion protein genes were co-amplified with the DHFR genes in growth media containing MTX with increasing concentrations. Subclones that were positive for DHFR expression were subjected to limiting dilution with gradually increased pressure to screen transfectants capable of growing in media with MTX of up to 6 μM. The secretion rates of the transfectants were determined and cell lines with high foreign protein expression were screened. Cell lines with a secretion rate of greater than about 5 μg/10.sup.6 (millions) cells/24 hours (preferably about 15 μg/10.sup.6 cells/24 hours) were adaptively suspended using a serum-free medium. Cell supernatants were collected and bispecific antibodies were separated and purified.

[0280] Hereinafter, the purification process, stability, in vitro and in vivo biological functions, safety, and pharmacokinetics of the bispecific antibodies of several configurations were evaluated to screen the bispecific antibody of an appropriate configuration.

1.4 Purification Process and Stability Detection of Bispecific Antibodies

[0281] Antibodies are generally purified by a three-step purification strategy: crude purification (sample capture), intermediate purification, and fine purification. In the crude purification stage, the target antibodies are generally captured by affinity chromatography which can effectively remove a large number of impurities such as heterologous proteins, nucleic acids, endotoxins, and viruses from the sample. The intermediate purification is often carried out using hydrophobic chromatography or CHT hydroxyapatite chromatography to remove most of the remaining impurity proteins and polymers. Fine purification is mostly carried out using anion exchange chromatography or gel filtration chromatography (molecular sieve) to remove the small or trace amount of remaining impurity proteins whose nature is similar to the nature of the target antibodies and further to remove contaminants such as HCP and DNA.

[0282] In the present disclosure, the culture supernatant of bispecific antibody AB7K8 fused with His-tag can be crudely purified using a metal chelation affinity chromatography column (e.g., HisTrap FF from GE). The bispecific antibodies AB7K4, AB7K5, AB7K6, AB7K, and AB7K7 containing Fc can be crudely purified using a Protein A/G affinity chromatography column (e.g., Mabselect SURE from GE). The products obtained after the above crude purification are then subjected to the intermediate purification and the fine purification to finally obtain purified target antibodies of high purity and high quality. The preservation buffers for the above bispecific antibodies are then replaced with PBS or other suitable buffers using desalination columns (e.g., HiTrap desalting from GE).

[0283] a) Purification of Double-Chain Tetravalent Bispecific Antibody AB7K7

[0284] Specific purification steps and solutions for such bispecific antibodies of a tetravalent homodimer configuration are illustrated below by using an example of AB7K7.

[0285] The bispecific antibody AB7K7 was purified by three-step chromatography. The three-step chromatography included affinity chromatography, hydrophobic chromatography, and anion exchange chromatography. (The protein purifier used in this example was AKTA pure 25 M from GE, U.S. Reagents used in this example were purchased from Sinopharm Chemical Reagent Co., Ltd and had purity at an analytical grade).

[0286] In a first step, affinity chromatography was performed. Sample capture and concentration and the removal of partial pollutants were performed using an affinity chromatography medium MabSelect Sure from GE or other commercially available affinity media (e.g., Diamond Protein A from Bestchrom). First, chromatography columns were equilibrated with 3-5 column volumes (CVs) of an equilibration buffer (20 mM PB, 140 mM NaCl, pH 7.4) at a linear flow rate of 100-200 cm/h. The clarified fermentation broth was loaded at a linear flow rate of 100-200 cm/h with a load not higher than 20 mg/mL. After loading, the chromatography columns were equilibrated with 3-5 column volumes (CVs) of an equilibration buffer (20 mM PB, 140 mM NaCl, pH 7.4) at a linear flow rate of 100-200 cm/h to remove unbound components. The chromatography columns were rinsed with 3-5 column volumes of decontamination buffer 1 (50 mM NaAc-HAc, 1 M NaCl, pH 5.0) at a linear flow rate of 100-200 cm/h to remove partial pollutants. The chromatography columns were equilibrated with 3-5 column volumes (CVs) of decontamination buffer 2 (50 mM NaAc-HAc, pH 5.0) at a linear flow rate of 100-200 cm/h. The target product was eluted using an elution buffer (40 mM NaAc-HAc, pH 3.5) at a linear flow rate not higher than 100 cm/h and target peaks were collected.

[0287] In a second step, hydrophobic chromatography was performed. Intermediate purification was performed using Butyl HP from Bestchrom or other commercially available hydrophobic chromatography media to reduce the content of polymers. After the target proteins were polymerized, since the polymers and monomers differed in property such as charge characteristics and hydrophobicity, the polymers and the monomers could be separated on the basis of the above differences between them. First, chromatography columns were equilibrated with 3-5 column volumes (CVs) of an equilibration buffer (20 mM PB, 0.3 M (NH.sub.4).sub.2SO.sub.4, pH 7.0) at a linear flow rate of 100-200 cm/h. The target proteins separated through the affinity chromatography in the first step were subjected to conductivity adjustment to 40-50 ms/cm with the solution of 2 M (NH.sub.4).sub.2SO.sub.4 and then loaded with a load controlled to be less than 20 mg/mL. After loading, the chromatography columns were rinsed with 3-5 column volumes (CVs) of an equilibration buffer (20 mM PB, 0.3 M (NH.sub.4).sub.2SO.sub.4, pH 7.0) at a linear flow rate of 100-200 cm/h. Finally, the target proteins were eluted using 3-5 column volumes (CVs) of an elution buffer (20 mM PB, pH 7.0) with gradients of 40%, 80% and 100% at a linear flow rate not higher than 100 cm/h. Eluted fractions were collected and sent for SEC-HPLC, respectively. Target components with the percentage of monomers being greater than 90% were combined for chromatography in the next step.

[0288] In a third step, anion exchange chromatography was performed. Fine purification was performed by using Q-HP from Bestchrom or other commercially available anion exchange chromatography media (e.g., Q HP from GE, Toyopearl GigaCap Q-650 from TOSOH, DEAE Beads 6FF from Smart-Lifesciences, Generik MC-Q from Sepax Technologies, Inc, Fractogel EMD TMAE from Merck, and Q Ceramic HyperD F from Pall) to separate structural variants and further remove pollutants such as HCP and DNA. First, chromatography columns were rinsed with 3-5 column volumes (CVs) of an equilibration buffer (20 mM PB, pH 7.0) at a linear flow rate of 100-200 cm/h. The target proteins separated through the hydroxyapatite chromatography in the second step were loaded and through-flow peaks were collected. After loading, the chromatography columns were rinsed with 3-5 column volumes (CVs) of an equilibration buffer (20 mM PB, pH 7.0) at a linear flow rate of 100-200 cm/h. The through-flow components were collected and sent for the detection of protein content, SEC-HPLC and electrophoresis.

[0289] The SEC-HPLC purity results and SDS-PAGE electrophoresis results of the samples are shown in FIG. 1-3 and FIG. 1-4. The SEC-HPLC results show that the purity of the main peak of the bispecific antibody was more than 95% after three-step chromatography. The band pattern in the SDS-PAGE electrophoresis was as expected, wherein a band was shown at 180 KDa in the non-reducing electrophoresis and a clear single-chain band (90 KDa) was obtained after reduction.

[0290] b) Purification of single-chain bivalent bispecific antibodies AB7K5 and AB7K6

[0291] The bispecific antibody AB7K5 was purified by Protein A affinity chromatography and hydroxyapatite (CHT) chromatography. After the SEC-HPLC test, it was found that the purity of bispecific antibody AB7K5 was low, its yield was not high, and there was a problem of extremely low expression yield.

[0292] For another single-chain bivalent bispecific antibody AB7K6, there also was a problem of high process development difficulty. The bispecific antibody AB7K6 was subjected to two-step purification, that is, Protein A affinity chromatography and molecular sieve chromatography Superdex 200. After the SEC-HPLC test, it was found that it was difficult to quantify the purity of bispecific antibody AB7K6, and there was a significant “shoulder peak” in the main peak; in addition, the expression yield of AB7K6 was very low and very unstable. After 24 hours of standing in a refrigerator at 4° C., it was found that the peak shape in the SEC-HPLC result was changed from two peaks to one main peak, which attributed, presumably, to the conversion from the single-chain structure to the double-chain structure in AB7K6. From the above, it can be seen that the current process development difficulty of AB7K6 is too high to achieve process scale-up and industrialization.

[0293] In summary, AB7K7 had significant advantages over AB7K5 and AB7K6 in terms of process development and had advantages such as high yield, simple and efficient purification methods, and stable downstream processes. The physicochemical stability of AB7K7 in different buffer systems and at different storage conditions was further studied.

[0294] c) Assay on stability of bispecific antibody AB7K7

[0295] The stability of AB7K7 proteins in a citrate buffer system (20 mM citrate, pH 5.5) and a histidine buffer system (20 mM histidine, pH 5.5) was studied, respectively. AB7K7 proteins were stored for four weeks under accelerated conditions at 25° C. for the evaluation of protein stability.

[0296] AB7K7 proteins were transferred to the preceding citrate buffer system (F2) and the histidine buffer system (F3), respectively, with the concentration adjusted to 0.5 mg/mL, wherein 8% sucrose (w/v) and 0.02% PS80 (w/v) were added to both buffer systems as excipients. The above buffer systems were filtered using a 0.22 μm PES membrane needle filter, and then vialed into 2 mL penicillin bottles, respectively, 0.8 mL in each bottle. After the vialing, a stopper was immediately pressed and capped. Samples were placed in different stability chambers according to the schemes in Table 1-2. Samples were taken at each sampling point for detection and analysis, wherein the detection terms included appearance, concentration, purity (detected by SEC-HPLC), HMW%, LMW%, and turbidity (A340) of the sample.

TABLE-US-00002 TABLE 1-2 Stability detection scheme Condition T.sub.0 Sampling point and detection term 40° C. X, Y 1 Week (W) 2 W 4 W X, Y X X, Y 25° C. 1 W 2 W 4 W X, Y X X, Y Freeze-thaw 3 cycles (−70° C./room X, Y temperature) Note: X = appearance, concentration, SEC-HPLC, SDS-PAGE (reducing & non-reducing); Y = turbidity (A340)

[0297] The appearance, concentration, turbidity and SEC-HPLC detection results of two preparations stored for 0-4 weeks at 25° C. are shown in Table 1-3 and Table 1-4, and SDS-PAGE (reducing/non-reducing) results thereof are shown in FIG. 1-5. There was no significant change in the appearance and concentration of the two preparations. In the SEC-HPLC results, the SEC results of the F2 and F3 preparations did not show any significant change. The purity after 4 weeks was 97.9% and 98.2%, respectively. SDS-PAGE (reducing/non-reducing) results were generally consistent with the trend of LMW% results, and F2 and F3 slightly changed.

TABLE-US-00003 TABLE 1-3 Appearance, concentration, and turbidity results in the acceleration test at 25° C. Appearance Concentration Turbidity A340 T0 1W 2W 4W T0 1W 2W 4W T0* 1W 4W F2 Colorless clear liquid without 0.46 0.46 0.45 0.46 0.003 0.004 0.002 F3 visible foreign matter 0.47 0.46 0.45 0.47 0.004 0.003 0.005 *T0 turbidity: the sample to be detected was a sample subjected to 1 cycle of freeze-thaw.

TABLE-US-00004 TABLE 1-4 SEC-HPLC results in the acceleration test at 25° C. SEC-Purity% SEC-HMW% SEC-LMW% T0 1W 2W 4W T0 1W 2W 4W T0 1W 2W 4W F2 97.5 98.3 98.5 97.9 2.5 1.5 1.5 1.5 0 0.3 0.0 0.6 F3 97.7 98.8 98.7 98.2 2.3 1.2 1.3 1.2 0 0.0 0.0 0.6

[0298] To know the unfolding temperature of AB7K7 proteins in the two buffer systems, the Tm (unfolding temperature) and Tmonset (the temperature at which the protein begins to unfold) in the two preparations were measured by DSF and the results are shown in Table 1-5. Both preparations had low Tmonset values, and F2 and F3 had Tmonset values less than 45° C.

TABLE-US-00005 TABLE 1-5 DSF results Tmonset (° C.) Tm1 (° C.) Tm2 (° C.) F2 42.0 46.0 60.5 F3 41.0 45.0 58.0

[0299] The stability of AB7K7 proteins in the above two buffer systems during freeze-thaw (−70° C./room temperature, 3 cycles of freeze-thaw) was studied by performing 3 cycles of freeze-thaw. The preparation and detection solution of the sample were the same as those described above.

[0300] The appearance, concentration, turbidity and SEC-HPLC detection results of samples are shown in Table 1-6, and SDS-PAGE (reducing/non-reducing) results thereof are shown in FIG. 1-6. In the SDS-PAGE (non-reducing) results, there were no significant changes in the results of each detection item of both F2 and F3 preparations subjected to three cycles of freeze-thaw.

TABLE-US-00006 TABLE 1-6 Appearance, concentration, turbidity, and SEC-HPLC results in the freeze-thaw test Appearance Concentration Turbidity A340 SEC-Purity % SEC-HMW % SEC-LMW % T0 FT-3C T0 FT-3C T0 FT-3C T0 FT-3C T0 FT-3C T0 FT-3C F2 Colorless clear Colorless clear 0.46 0.46 0.003 0.005 97.5 98.4 2.5 1.6 0 0.0 F3 liquid without liquid without 0.47 0.48 0.004 0.005 97.7 98.5 2.3 1.4 0 0.2 visible foreign visible foreign matter matter

Example 2 Evaluation of In Vitro Biological Functions of Anti-Her2×CD3 Bispecific Antibodies

[0301] 2.1 Detection of Binding Activities of Bispecific Antibodies to Effector Cells and Target Cells (FACS)

[0302] a) Detection of Binding Activities of Bispecific Antibodies to Her2-Positive Tumor Cells BT-474 by Flow Cytometry

[0303] Tumor cells BT-474 that were positive for Her2 expression (from the cell bank of Chinese Academy of Sciences, Shanghai) were cultured, then digested with 0.25% trypsin, and centrifuged to collect cells. The collected cells were resuspended with 1% PBSB, placed in 96-well plates after the cell density was adjusted to (2×10.sup.6) cells/ml, 100 μl (2×10.sup.5 cells) per well, and blocked for 0.5 hours at 4° C. The blocked cells were centrifuged to discard the supernatant, and a series of diluted bispecific antibodies were added to the cells. The cells were incubated for 1 hour at 4° C., then centrifuged to discard the supernatant, and washed three times using a PBS solution with 1% BSA (PBSB). Diluted AF488-labeled goat anti-human IgG antibodies or murine anti-6×His IgG antibodies were added to the cells, and the cells were incubated for 1 hour at 4° C. in the dark. The obtained cells were centrifuged to discard the supernatant, and washed twice with 1% PBSB, and cells in each well were resuspended with 100 μl of 1% paraformaldehyde (PF). The signal intensity was detected by flow cytometry. The analysis was performed with the average fluorescence intensity as the Y-axis and the antibody concentration as the X-axis through software GraphPad to calculate the EC.sub.50 value for the binding of bispecific antibodies to tumor cells BT-474.

[0304] The results show that bispecific antibodies with different structures had good binding activity to tumor cells overexpressing Her2. FIG. 2-1 to FIG. 2-5 show binding curves of bispecific antibodies with different structures to tumor cells BT-474. As shown in Table 2-1, the EC.sub.50 for the binding of AB7K to tumor cells and the EC.sub.50 for the binding of AB7K4 to tumor cells both were around 5 nM, the EC.sub.50 for the binding of AB7K7 to tumor cells was close to 50 nM, the EC.sub.50 for the binding of AB7K5 to tumor cells and the EC.sub.50 for the binding of AB7K8 to tumor cells both were greater than 100 nM, and the EC.sub.50 for the binding of AB7K6 to tumor cells was greater than 200 nM.

TABLE-US-00007 TABLE 2-1 Detection of abilities of Anti-Her2 × CD3 bispecific antibodies to bind to tumor cells BT474 AB7K AB7K4 AB7K5 AB7K6 AB7K7 AB7K8 EC.sub.50 (nM) 5.009 4.388 125.0 239.9 51.98 125.3

[0305] b) Detection of Binding Activities of Bispecific Antibodies to Human T Cells by FACS

[0306] PBMCs were prepared from fresh human blood by density gradient centrifugation. The prepared PBMCs were resuspended in a 1640 medium containing 10% heat-inactivated FBS, added with 2 μg/ml OKT3 for activation for 24 h, then added with 250 IU/ml IL-2 for amplification for 7 days, to prepare cytokine-induced killer (CIK) cells which were detected by flow cytometry to be positive for CD3 expression on the surface. The to-be-detected samples were prepared and detected in the same manner as in a) of Example 2.1. Cells resuspended with 1% PF were detected on a machine and, with the average fluorescence intensity, analyzed by software OriginPro 8 to calculate the EC.sub.50 value for the binding of each bispecific antibody to human CIK cells.

[0307] The results show that there were great differences among the binding of each bispecific antibody to CIK cells (FIG. 2-6 to FIG. 2-10). As shown in Table 2-2, the EC.sub.50 of AB7K was about 20 nM, which was roughly equal to the EC.sub.50 of AB7K4, the EC.sub.50 of AB7K7 was more than 6 times higher than the EC.sub.50 of AB7K, and the EC.sub.50 of AB7K5, AB7K6 and AB7K8 was more than 10 times higher than the EC.sub.50 of AB7K.

TABLE-US-00008 TABLE 2-2 Detection of abilities of Anti-Her2 × CD3 bispecific antibodies to bind to effector cells CIK AB7K AB7K4 AB7K5 AB7K6 AB7K7 AB7K8 EC.sub.50 (nM) 20.51 19.44 375.2 241.7 132.3 504.1

[0308] c) Detection of Cross-Reactivity of Bispecific Antibodies with Cynomolgus Monkey CIK Cell Membrane CD3 by FACS

[0309] PBMCs were prepared from fresh cynomolgus monkey blood by density gradient centrifugation. The prepared PBMCs were resuspended in a 1640 medium containing 10% heat-inactivated FBS, added with 2 μg/m1 OKT3 for activation for 24 h, then added with 250 IU/ml IL-2 for amplification for 7 days, to prepare cynomolgus monkey CIK cells for use. Human CIK cells and cynomolgus monkey CIK cells were collected by centrifugation, followed by the same test procedure as in the above examples. Cells resuspended with 1% paraformaldehyde solution were detected on a machine and, with the average fluorescence intensity, analyzed by software OriginPro 8 to calculate the EC.sub.50 values for the binding of bispecific antibodies to human CIK cells and the EC.sub.50 values for the binding of bispecific antibodies to cynomolgus monkey CIK cells.

[0310] As shown in FIG. 2-11, the bispecific antibody AB7K bound well to cynomolgus monkey T cells, the ability of AB7K to bind to cynomolgus monkey T cells was roughly equal to the ability of AB7K to bind to human T cells, and the EC.sub.50 for the binding of AB7K to cynomolgus monkey T cells was approximately 26 nM as detected by flow cytometry. Bispecific antibodies AB7K4, AB7K5, AB7K6, AB7K7, and AB7K8 bound specifically to cynomolgus monkey T cells, as did AB7K.

2.2 Detection of Abilities of Bispecific Antibodies to Bind to Antigens

[0311] The binding of bispecific antibodies to soluble CD3 and Her2 was detected by double antigen sandwich ELISA.

[0312] Her2 proteins (SinoBiological, Beijing, Cat. No. 10004-H08H4) were diluted with PBS to a concentration of 0.1 μg/m1 and added to 96-well plates, 100 μl per well. The plates were coated at 4° C. overnight. The plates were then blocked with 1% skimmed milk powder for 1 hour at room temperature. Each bispecific antibody was diluted simultaneously with a 4-fold gradient for a total of 11 concentration gradients. The 96-well plates were then washed with PBST, and then the diluted bispecific antibodies were added. Control wells without antibodies were set. Incubated for 1 hour at room temperature. Unbound bispecific antibodies were washed away with PBST. Biotinylated CD3E and CD3D (ACRO Biosystem, Cat. No. CDD-H82W1) were mixed at 50 ng/ml with streptavdin HRP (BD, Cat. No. 554066), added in the 96-well plates, 100 μl per well, and incubated for 1 hour at room temperature. 96-well plates were washed with PBST, and TMB was added to the plates, 100 μl per wellss. Color development was performed at room temperature for 15 minutes, and then 0.2 M H.sub.2SO.sub.4 was added to stop the color development reaction. The light absorbance values at A450-620 nm were measured by a microplate reader. Analysis was performed by software OriginPro 8, and the EC.sub.50 values for the binding of bispecific antibodies to two antigens were calculated.

[0313] The results show that each bispecific antibody bound specifically to both CD3 and Her2 molecules and exhibited good dose-dependence as the concentration of the antibodies changed (FIG. 2-12). The abilities of several bispecific antibodies to bind to soluble CD3 and Her2 are shown in Table 2-3, with EC.sub.50 values ranging from 0.03 nM to 3.8 nM which differ by two orders of magnitude. AB7K had the best binding activity, binding activities of AB7K4 and AB7K7 differed by one order of magnitude, and AB7K5 and AB7K8 had the weakest binding activity.

TABLE-US-00009 TABLE 2-3 Detection of abilities of Anti-Her2 × CD3 bispecific antibodies to bind to CD3 and Her2 molecules AB7K AB7K4 AB7K5 AB7K7 AB7K8 EC.sub.50 (nM) 0.03128 0.1518 1.004 0.1398 3.815

2.3 Evalution of Abilities of Bispecific Antibodies to Activate T Cells Through Reporter Gene Cell Strains

[0314] Jurkat T cells containing NFAT RE reporter genes (BPS Bioscience, Cat. No. 60621) can overexpress luciferase in the presence of bispecific antibodies and target cells, and the degree of activation of the Jurkat T cells can be quantified by detecting the activity of the luciferase. A four-parameter curve was fitted using the concentration of bispecific antibodies as the X-axis and the fluorescein signal as the Y-axis.

[0315] The test results from FIG. 2-13 show that the monoclonal antibody Herceptin targeting Her2 cannot activate Jurkat T cells. T cells can be activated only in the presence of both antibodies. The ability of each antibody to activate Jurkat T cells is shown in Table 2-4. AB7K4 had the strongest ability to activate T cells, AB7K8 had the weakest ability to activate T cells, and their EC.sub.50 values differed by one order of magnitude.

TABLE-US-00010 TABLE 2-4 Detection of abilities of Anti-Her2 × CD3 bispecific antibodies to a reporter gene cell strain that are Jurkat T cells AB7K AB7K4 AB7K5 AB7K7 AB7K8 Herceptin EC50 (nM) 0.02263 0.01338 0.05357 0.08952 0.1575 0.009907

2.4 Abilities of Bispecific Antibodies to Mediate T Cells to Kill Tumor Cells

[0316] Normally cultured tumor cell lines, including SK-BR-3, MCF-7, HCC1937, NCI-N87, HCC1954 cells (all purchased from the cell bank of Chinese Academy of Sciences, Shanghai), as target cells, were digested with 0.25% trypsin to prepare single-cell suspensions, added to 96-well cell culture plates after the cell density was adjusted to 2×10.sup.5 cells/ml, 100 μl per well, and cultured overnight. The antibodies were diluted according to the test design, and added to the cells, 50 μl per well, while wells without the addition of antibodies were supplemented with the same volume of the medium. Effector cells (human PBMCs or expanded CIK cells) whose number was five times larger than the number of target cells, were then added, 100 μl per well. Control wells were set, and wells without the addition of effector cells were supplemented with the same volume of the medium. After incubation for 48 hours, the supernatant was discarded from the 96-well plates. The cells were then washed three times with PBS, and a complete medium containing 10% CCK-8 was added, 100 μl per well, and the cells were incubated for 4 hours at 37° C. The light absorbance values at A450-620 nm were measured by a microplate reader. Analysis was performed by software OriginPro 8, and the ability of each bispecific antibody to mediate the killing of tumor cells and the ability of the same target monoclonal antibody Herceptin to mediate the killing of tumor cells were calculated and compared.

[0317] The EC.sub.50 values of each bispecific antibody to mediate effector cells to kill tumor cells are shown in Table 2-5. The results show that each bispecific antibody exhibited a very significant killing effect on tumor cells (e.g., SK-BR-3, NCI-N87, and HCC1954) with high expression of Her2 in a dose-dependent manner. Each bispecific antibody, in particular AB7K7, also exhibited a good killing effect on breast cancer cells MCF-7 with low expression of Her2. Each bispecific antibody also had a good killing effect on the Herceptin-resistant cell strain HCC1954 while each bispecific antibody exhibited the killing effect on the cell strain HCC1937 that was negative for Her2 expression (with little expression) only at two highest concentrations.

TABLE-US-00011 TABLE 2-5 EC.sub.50 values of bispecific antibodies to mediate PBMCs to kill different tumor cells EC50 (nM) AB7K7 AB7K8 AB7K5 Herceptin SK-BR-3 ~0.001 ~0.002 ~0.001  — ~0.001 0.011 — 0.067 MCF-7 ~0.005 0.079 0.055 — HCC1937 0.659 ~2.269 1.223 — 0.579 4.011 — >6.667  NCI-N87 0.015 0.034 — 0.129 HCC1954 0.002 0.018 — 0.050 Note: ~means approximately equal to, and — means that no detection is performed.
2.5 Evaluation of the Effect of GS-CTP Linker Peptide on the Ability of Anti-CD3 scFv to Bind to CD3 Molecules by Computer Techniques

[0318] The anti-CD3 scFv VH containing the GS-CTP linker peptide was structurally modeled using computer software and the spatial conformation of molecular docking of anti-CD3 scFv and its antigen CD3 epsilon chain was simulated and predicted.

[0319] The sequence of the GS-CTP linker peptide between anti-Her2 scFv and anti-CD3 scFv in the bispecific antibody AB7K7 is (GGGGGGSGGGGSGGGGSSSSSKAPPPS), wherein the first half of the sequence is a GS-flexible peptide (GGGGGGSGGGGSGGGGS), and the second half is CTP-rigid peptide (SSSSKAPPPS). The rigid CTP portion (SSSSKAPPPS) is connected to the N-terminus of the anti-CD3 scFv VH. Through three-dimensional structural modeling using software phyre2, it is found that the CTP peptide fragment structurally overlays on the CDR1 region of VH of the anti-CD3 scFv (FIG. 2-14), which may hinder or disrupt the binding of the CD3 antibody to its antigen. The VH of the anti-CD3 scFv connected to the GS linker peptide (containing only the GS flexible peptide with the removal of CTP) was subjected to three-dimensional structural modeling using software phyre2, and then it is found that the GS linker peptide is far from the CDR region (FIG. 2-15) and does not affect antigen-antibody binding. Even if the GS linker peptide is close to the CDR region, the GS linker peptide can freely move away from the antigen-antibody binding region due to its own flexibility and thus does not affect antigen-antibody binding.

[0320] Further, the molecular docking between the anti-CD3 scFv and its antigen CD3 epsilon chain was simulated by software Discovery Studio. Since the structure of the double-chain anti-CD3 FV is highly similar to the structure of the anti-CD3 scFv, the structure simulation was performed using the double-chain anti-CD3 FV instead of the anti-CD3 scFv. The simulation results show that the antigen CD3 epsilon chain binds to CDR2 and CDR3 of VH of the anti-CD3 Fv while does not bind to the CDR1 region (FIG. 2-16), which indicates that the CTP overlaying the VH CDR1 region of the anti-CD3 Fv does not interfere with the binding of the anti-CD3 scFv to the antigen. However, given that the CD3 molecule is a complex including one CD3 gamma chain, one CD3 delta chain, and two CD3 epsilon chains, the CD3 molecule, together with the TCR and Zeta chains, constitutes a T-cell receptor complex. Although the CTP peptide fragment covering the VH CDR1 of the anti-CD3 scFv does not directly interfere with the binding of the anti-CD3 scFv to its antigen CD3 epsilon chain, the CTP peptide fragment may indirectly affect the binding of the anti-CD3 scFv to its antigen CD3 epsilon chain by making spatial structural contact with a certain constituent protein of the T-cell receptor complex.

[0321] Due to the presence of CTP covering the VH CDR1 region of the anti-CD3 scFv, the binding affinity of the anti-CD3 scFv for its antigen is greatly diminished so that there is no substantial release of cytokines caused by the overactivation of T cells, thereby avoiding some unnecessary T cell-mediated non-specific killing.

Example 3 Pharmacodynamics Study of Anti-Her2×CD3 Bispecific Antibodies in a Mouse Transplanted Tumor Model

3.1 NCG Mouse Model of Transplanted Tumor Constructed by Subcutaneously Co-Inoculating Human CIK Cells and Human Breast Cancer Cells HCC1954

[0322] Her2-positive human breast cancer cells HCC1954 were selected to study the effect of bispecific antibodies in inhibiting tumor growth in vivo in an NCG mouse model of transplanted tumor constructed by subcutaneously co-inoculating human CIK cells and HCC1954 cells.

[0323] The peripheral blood of a normal human was subjected to density gradient centrifugation (Lymphoprep™, Lymphocytes Separation Medium, STEMCELL) to separate human PBMCs. Then the human PBMCs were resuspended in RPMI-1640 culture medium added with 10% inactivated FBS, and added with OKT3 at a final concentration of 1 μg/mL and human IL-2 at 250 IU/mL. After three days of culture, the human PBMCs were centrifuged at 300 g for 5 minutes, and the medium was changed. The cells were cultured in RPMI-1640 added with 10% inactivated FBS and added with human IL-2 at 250 IU/mL. After that, a fresh medium was then added every 2 days and CIK cells were collected on the tenth day of culture. Female NCG mice at the age of seven to eight weeks (purchased from Jiangsu GemPharmatech Co. Ltd Company) were selected and HCC1954 cells in the logarithmic growth stage were collected. 5×10.sup.6 HCC1954 cells and 5×10.sup.5 CIK cells were mixed and inoculated subcutaneously on the right back of each NCG mouse. One hour later, the mice were randomly divided into seven groups with five mice in each group according to their weights and intraperitoneally administered with corresponding drugs. All positive control groups and PBS control group were administered twice a week for a total of 3 doses, wherein the positive control groups were administered with Herceptin (from Roche) at doses of 1 mg/kg and 3 mg/kg, respectively, and the PBS control group was administered with a PBS solution of the same volume as Herceptin. The treated groups were administered with bispecific antibodies AB7K4 and AB7K7 every day at doses of 0.1 mg/kg and 1 mg/kg, respectively, for a total of 10 doses. The day of administration was recorded as Day 0. The maximum diameter (D) and the minimum diameter (d) of the tumor were measured weekly with an electronic vernier caliper. The volume of the tumor was calculated using the following formula: volume (mm.sup.3)=[D×d.sup.2]/2. The tumor growth inhibition rate (TGI) was calculated for each treated group using the following formula: TGI (%) =(1−volume of the treated group/volume of the control group)×100%.

[0324] As shown in FIG. 3-1, on Day 33 of administration, the average tumor volume of the PBS control group was 1494.61±500.28 mm.sup.3; the average tumor volume of the treated group administrated with Herceptin at a dose of 1 mg/kg was 1327.29±376.65 mm.sup.3; the average tumor volume of the treated group administrated with Herceptin at a dose of 3 mg/kg was 510.49±106.07 mm.sup.3, and the TGI was 65.84%, which was not significantly different from that of the control group. The average tumor volumes of treated groups administrated with AB7K4 at doses of 0.1 mg/kg and 1 mg/kg were 304.10 ±108.50 mm.sup.3 and 79.70 ±58.14 mm.sup.3, respectively, and TGIs thereof were 79.65% and 94.67%, respectively, which were significantly different from that of the PBS control group (P<0.05). The average tumor volumes of treated groups administrated with AB7K7 at doses of 0.1 mg/kg and 1 mg/kg were 385.82 ±95.41 mm.sup.3 and 209.98 ±51.74 mm.sup.3, respectively, and TGIs thereof were 74.19% and 85.95%, respectively, which were significantly different from that of the PBS control group (P<0.05). In summary, the results show that the bispecific antibodies AB7K4 and AB7K7 at different doses could inhibit the growth of tumor cells by activating human immune cells in animals and exhibited great anti-tumor effects; and at the same dose of 1 mg/kg, the anti-tumor effect of the bispecific antibody was better than that of the monoclonal antibody Herceptin.

3.2 NPG Mouse Model of Transplanted Tumor Constructed by Subcutaneously Co-Inoculating Human CIK Cells and Human Breast Cancer Cells HCC1954

[0325] Her2-positive human breast cancer cells HCC1954 were selected to study the inhibiting effect of bispecific antibodies on tumor growth in vivo in an NPG mouse model of transplanted tumor constructed by subcutaneously co-inoculating human CIK cells and human breast cancer cells HCC1954.

[0326] CIK cells were prepared in the method as described in Example 3.1. Female NPG mice at the age of seven to eight weeks (purchased from Beijing Vitalstar Biotechnology Co., Ltd.) were selected and HCC1954 cells in the logarithmic growth stage were collected. 5×10.sup.6 HCC1954 cells and 5×10.sup.5 CIK cells were mixed and inoculated subcutaneously on the right back of each NPG mouse. After 6 days of tumor growth, the mice were randomly divided into three groups with six mice in each group according to the tumor volumes and weights and intraperitoneally administered with corresponding drugs. Specifically, AB7K7 treated groups were administered twice a week at doses of 0.1 mg/kg and 1 mg/kg, respectively, and the control group was administered with a PBS solution of the same volume as AB7K7. The day of administration was recorded as Day 0. The maximum diameter (D) and the minimum diameter (d) of the tumor were measured weekly. The volume (mm.sup.3) of the tumor of each group and the tumor growth inhibition rate (TGI) (%) of each treated group were calculated using the formulas as shown in Example 3.1.

[0327] As shown in FIG. 3-2, on Day 21 of administration, the average tumor volume of the PBS control group was 821.73±201.82 mm.sup.3; the average tumor volume of the treated group administrated with AB7K7 at a dose of 0.1 mg/kg was 435.60±51.04 mm.sup.3, and the TGI was 50.83%, which was not significantly different from that of the control group; the average tumor volume of the treated group administrated with AB7K7 at a dose of 1 mg/kg was 40.98 ±12.64 mm.sup.3, and the TGI was 95.37%, which was significantly different from that of the control group (P<0.01). The above results show that the administration of the bispecific antibody AB7K7 had a good therapeutic effect even when tumors had grown to a certain volume, wherein 50% tumor inhibition effect was achieved at the low dose of 0.1 mg/kg, and there was complete tumor regression in 4 of 6 mice in the treated group at the dose of 1 mg/kg and the tumor volumes in the other 2 mice were both less than 100 mm.sup.3, which was smaller than the tumor volume at the time of grouping (the average tumor volume of this group at the time of grouping was 161.37±18.98 mm.sup.3). Therefore, the bispecific antibody AB7K7 had a great therapeutic effect on tumors.

[0328] In addition, the inhibiting effect of bispecific antibodies AB7K7 and AB7K8 on tumor growth in the above-described transplanted tumor model at two administration frequencies were also studied. CIK cells were prepared in the method as described above. Female NPG mice at the age of seven to eight weeks were selected, and 5×10.sup.6 HCC1954 cells and 5×10.sup.5 CIK cells were mixed and inoculated subcutaneously on the right back of each NPG mouse. One hour later, the mice were randomly divided into six groups with six mice in each group according to their weights and intraperitoneally administered with corresponding drugs. Specifically, the control group and the Herceptin treated group were administered twice a week, wherein Herceptin was administrated at a dose of 3 mg/kg and the control group was administered with a PBS solution of the same volume as Herceptin. The bispecific antibody AB7K7 was administered at a dose of 1 mg/kg and AB7K8 was administered at a dose of 0.7 mg/kg. Two administration frequencies were set for each of the two bispecific antibodies, wherein the QD group was administered once a day for 10 consecutive days and the BIW group was administered twice a week. The day of administration was recorded as Day 0. The maximum diameter (D) and the minimum diameter (d) of the tumor were measured weekly. The tumor volume (mm.sup.3) of each group and the tumor growth inhibition rate (TGI) (%) of each treated group were calculated using the formulas as shown above.

[0329] As shown in FIG. 3-3, on Day 25 of administration, the average tumor volume of the PBS control group was 1588.12±120.46 mm.sup.3; the average tumor volume of the treated group administrated with Herceptin at a dose of 3 mg/kg was 361.72±134.70 mm.sup.3; the average tumor volumes of the QD group and the BIW group administrated with AB7K7 were 260.18±45.96 mm.sup.3 and 239.39±40.62 mm.sup.3, respectively, and TGIs were 83.62% and 84.93%, respectively, which were significantly different from that of the PBS control group (P<0.01); the average tumor volumes of the QD group and the BIW group administrated with AB7K8 were 284.98±26.62 mm.sup.3 and 647.14±118.49 mm.sup.3, respectively, and TGIs were 82.06% and 59.25%, respectively, which were significantly different from that of the PBS control group (P<0.01). As can be seen from the above results, the anti-tumor effect of the bispecific antibody AB7K7 was superior to that of the Herceptin in both the QD group and the BIW group; at equimolar doses, AB7K8 and AB7K7 in the QD group exhibited basically equal tumor inhibiting effects, while the anti-tumor effect of AB7K7 in the BIW group was significantly superior than that of AB7K8, presumably due to the fact that AB7K8 is a bispecific antibody of BiTE configuration with no Fc domain, and therefore AB7K7 has a longer half-life than AB7K8, from which it is anticipated that the clinical administration frequency of AB7K7 is reduced and AB7K7 has a better therapeutic effect.

3.3 NPG Mouse Model of Transplanted Tumor Constructed by Subcutaneously Co-Inoculating Human CIK Cells and Human Ovarian Cancer Cells SK-OV-3

[0330] Her2-positive human ovarian cancer cells SK-OV-3 were selected to study the inhibiting effect of bispecific antibodies on tumor growth in vivo in an NPG mouse model of transplanted tumor constructed by subcutaneously co-inoculating human CIK cells and SK-OV-3 cells.

[0331] The peripheral blood of a normal human was subjected to density gradient centrifugation to separate human PBMCs. Then the human PBMCs were resuspended in McCoy's 5A culture medium added with 10% inactivated FBS, and added with OKT3 at a final concentration of 1 μg/mL and human IL-2 at 250 IU/mL. After three days of culture, the human PBMCs were centrifuged at 300 g for 5 minutes, and the supernatant was discarded. The cells were resuspended in RPMI-1640 added with 10% inactivated FBS and added with 250 IU/mL of human IL-2. After that, a fresh medium was then added every 2 days and CIK cells were collected on the tenth day of culture. Female NPG mice at the age of seven to eight weeks were selected and SK-OV-3 cells (purchased from the cell bank of Chinese Academy of Sciences, Shanghai) in the logarithmic growth stage were collected. 3×10.sup.6 SK-OV-3 cells and 3×10.sup.5 CIK cells were mixed and inoculated subcutaneously on the right back of each NPG mouse. One hour after inoculation, the mice were randomly divided into seven groups with six mice in each group according to their weights and intraperitoneally administered with corresponding drugs. Herceptin and AB7K7 treated groups were administered twice a week at doses of 1 mg/kg, 0.2 mg/kg, and 0.04 mg/kg, respectively, and the control group was administered with a PBS solution of the same volume. The day of administration was recorded as Day 0. The maximum diameter (D) and the minimum diameter (d) of the tumor were measured weekly. The volume (mm.sup.3) of the tumor of each group and the tumor growth inhibition rate (TGI) (%) of each treated group were calculated using the formulas as shown in Example 3.1.

[0332] As shown in FIG. 3-4, on Day 21 of administration, the average tumor volume of the PBS control group was 834.09±45.64 mm.sup.3; the average tumor volumes of the treated groups administrated with Herceptin at doses of 1 mg/kg, 0.2 mg/kg, and 0.04 mg/kg were 644.84±58.22 mm.sup.3, 884.95±38.63 mm.sup.3, and 815.79±78.39 mm.sup.3, respectively; and the tumors in all AB7K7 treated groups were completely regressed. The above results show that in the ovarian cancer SK-OV-3 model, AB7K7 enabled the tumor to completely regress even at a very low dose of 0.04 mg/kg, exhibiting an excellent anti-tumor effect.

3.4 NPG Mouse Model of Transplanted Tumor Constructed by Subcutaneously Co-Inoculating Human CIK Cells and Human Colon Cancer Cells HT-29

[0333] Her2-positive human colon cancer cells HT-29 were selected to study the inhibiting effect of bispecific antibodies on tumor growth in vivo in an NPG mouse model of transplanted tumor constructed by subcutaneously co-inoculating human CIK cells and HT-29 cells.

[0334] CIK cells were prepared in the method as described in Example 3.1. Female NPG mice at the age of seven to eight weeks were selected and HT-29 cells (purchased from the cell bank of Chinese Academy of Sciences, Shanghai) in the logarithmic growth stage were collected. 3×10.sup.6 HT-29 cells and 3×10.sup.6 CIK cells were mixed and inoculated subcutaneously on the right back of each NPG mouse. One hour after inoculation, the mice were randomly divided into five groups with six mice in each group according to their weights and intraperitoneally administered with corresponding drugs. Specifically, Herceptin was administered at a dose of 3 mg/kg, and AB7K7 was administered at doses of 3 mg/kg, 1 mg/kg, and 0.3 mg/kg, respectively. All treated groups were administered twice a week. The control group was administered with a PBS solution of the same volume. The day of administration was recorded as Day 0. The maximum diameter (D) and the minimum diameter (d) of the tumor were measured weekly. The volume (mm.sup.3) of the tumor of each group and the tumor growth inhibition rate (TGI) (%) of each treated group were calculated using the formulas as shown in Example 3.1.

[0335] As shown in FIG. 3-5, on Day 21 of administration, the average tumor volume of the PBS control group was 1880.52±338.26 mm.sup.3; the average tumor volume of the treated group administrated with Herceptin at a dose of 3 mg/kg was 1461.36±177.94 mm.sup.3; the average tumor volumes of the treated groups administrated with AB7K7 at doses of 3 mg/kg, 1 mg/kg, and 0.3 mg/kg were 13.94±7.06 mm.sup.3, 26.31±10.75 mm.sup.3, and 10.47±6.71 mm.sup.3, wherein tumors in four mice in the treated group at a dose of 0.3 mg/kg were completely regressed, tumors in three mice in the treated group at a dose of 1 mg/kg were completely regressed, and tumors in four mice in the treated group at a dose of 3 mg/kg were completely regressed. The above results show that in the colon cancer HT-29 model, Herceptin had few pharmacological effect on this tumor model, whereas AB7K7 exhibited complete tumor regression in mice at all three doses and exhibited excellent anti-tumor effect even at very low doses.

3.5 CD34 Immune-Reconstituted NPG Mouse Model of Transplanted Tumor Constructed by Inoculating Human Breast Cancer Cells HCC1954

[0336] Her2-positive human breast cancer cells HCC1954 were selected to study the inhibiting effect of bispecific antibodies on tumor growth in vivo in a CD34 immune-reconstituted NPG mouse model of transplanted tumor constructed by subcutaneously inoculating human breast cancer cells HCC1954.

[0337] CD34-positive hematopoietic stem cells were enriched from fresh umbilical cord blood using CD34-positive selective magnetic beads (purchased from Miltenyi Biotec, Germany). Female NPG mice at the age of seven to eight weeks (purchased from Beijing Vitalstar Biotechnology Co., Ltd.) were selected and injected with CD34-positive hematopoietic stem cells via the tail vein to reconstitute a human immune system in each mouse. Sixteen weeks later, blood was collected from the orbital venous plexus of mice for flow cytometry, and when the proportion of human CD45 in mice was greater than 15%, it was considered that the immune reconstitution succeeded. HCC1954 cells in the logarithmic growth stage were collected and 5 ×10.sup.6 HCC1954 cells were inoculated subcutaneously on the right back of the mice with successful immune reconstitution. One hour after inoculation, the mice were randomly divided into three groups with six mice in each group according to their weights. The treated groups were intraperitoneally administered with AB7K7 and Herceptin at a dose of 1 mg/kg, and the control group was administered with a PBS solution of the same volume, twice a week for a total of 6 doses. The day of administration was recorded as Day 0. The maximum diameter (D) and the minimum diameter (d) of the tumor were measured weekly. The volume (mm.sup.3) of the tumor of each group and the tumor growth inhibition rate (TGI) (%) of each treated group were calculated using the formulas as shown in Example 3.1.

[0338] As shown in FIG. 3-6, on Day 21 of administration, the average tumor volume of the PBS control group was 475.23±58.82 mm.sup.3; the average tumor volume of the treated group administrated with Herceptin was 293.27±66.35 mm.sup.3, and the TGI was 38.29%, which was not significantly different from that of the control group; the average tumor volume of the treated group administrated with AB7K7 was 0.67 ±0.67 mm.sup.3, and the TGI was 99.86%, meaning that basically all tumors were regressed, which was significantly different from that of the control group (P<0.01). In summary, the above results show that the bispecific antibody AB7K7 had an excellent anti-tumor effect in the CD34 immune-reconstituted model.

3.6 PBMC Immune-Reconstituted NPG Mouse Model of Transplanted Tumor Constructed by Inoculating Human Breast Cancer Cells HCC1954

[0339] Her2-positive HCC1954 cells were selected to study the inhibiting effect of bispecific antibodies on tumor growth in vivo in a PBMC immune-reconstituted NPG mouse model of transplanted tumor constructed by inoculating human breast cancer cells HCC1954.

[0340] The peripheral blood of a normal human was subjected to density gradient centrifugation to separate human PBMCs. Female NPG mice at the age of five to six weeks were selected and intraperitoneally injected with human PBMC cells to reconstitute a human immune system in each mouse. After seven days of PBMC injection, HCC1954 cells in the logarithmic growth stage were collected and 5×10.sup.6 HCC1954 cells were inoculated subcutaneously on the right back of each mouse. After 13 days of PBMC injection, blood was collected from the orbital venous plexus for flow cytometry, and when the proportion of human CD45 in mice was greater than 15%, it was considered that the immune reconstitution succeeded. After 14 days of PBMC injection, the mice with successful immune reconstitution were randomly divided into two groups with six mice in each group according to the tumor volumes and weights. The treated group was intraperitoneally administered with AB7K7 at a dose of 1 mg/kg, and the control group was administered with PBS, three times a week. The day of administration was recorded as Day 0. The maximum diameter (D) and the minimum diameter (d) of the tumor were measured weekly. The volume (mm.sup.3) of the tumor of each group and the tumor growth inhibition rate (TGI) (%) of each treated group were calculated using the formulas as shown in Example 3.1.

[0341] As shown in FIG. 3-7, on Day 23 of administration, the average tumor volume of the PBS control group was 1224.05±224.39 mm.sup.3; the average tumor volume of the treated group administrated with AB7K7 was 32.00±0.00 mm.sup.3, and the TGI was 97.41%, meaning that basically all tumors were regressed, which was significantly different from that of the control group (P<0.001). In summary, the above results show that the bispecific antibody AB7K7 had an excellent anti-tumor effect in the PBMC immune-reconstituted model.

Example 4 Evaluation of the Safety of Anti-Her2×CD3 Bispecific Antibodies

[0342] 4.1 Bispecific Antibodies being Incapable of Mediating Non-Specific Killing on Her2-negative Tumor Cells

[0343] Her2-negative human Burkkit's lymphoma Raji cells were selected to study whether bispecific antibodies can inhibit tumor growth in an NCG mouse model of transplanted tumor constructed by subcutaneously co-inoculating human CIK cells and human Burkkit's lymphoma Raji cells.

[0344] CIK cells were prepared in the method as described in Example 3.1. Female NCG mice at the age of seven to eight weeks were selected and Raji cells (purchased from the cell bank of Chinese Academy of Sciences, Shanghai) in the logarithmic growth stage were collected. 5×10.sup.6 Raji cells and 2×10.sup.6 CIK cells were mixed and inoculated subcutaneously on the right back of each NCG mouse. One hour after inoculation, the mice were randomly divided into three groups with five mice in each group according to their weights. The treated groups were intraperitoneally administered with AB7K4 and AB7K7 at a dose of 1 mg/kg, and the control group was administered with a PBS solution of the same volume, once a day continuously for 10 days. The day of administration was recorded as Day 0. The maximum diameter (D) and the minimum diameter (d) of the tumor were measured weekly. The volume (mm.sup.3) of the tumor of each group and the tumor growth inhibition rate (TGI) (%) of each treated group were calculated using the formulas as shown in Example 3.1.

[0345] As shown in FIG. 4-1, on Day 25 of administration, the average tumor volume of the PBS control group was 2439.88±193.66 mm.sup.3; the average tumor volume of the treated group administrated with AB7K4 was 2408.81±212.44 mm.sup.3; the average tumor volume of the treated group administrated with AB7K7 was 2598.11±289.35 mm.sup.3; and there was no difference between the average tumor volume of each of the two treated groups and the average tumor volume of the control group. In summary, the results show that bispecific antibodies AB7K4 and AB7K7 exhibited no non-specific killing on Her2-negative cell strains, which indicates that AB7K4 and AB7K7 do not mediate T cells to kill non-target tissues in vivo (i.e., specifically dependent on binding of bispecific antibodies to corresponding target antigens), there is no off-target toxicity, and the safety is high.

4.2 Bispecific Antibodies Killing Tumor Cells Depending on the Activation of T Cells

[0346] Her2-positive human breast cancer cells HCC1954 were selected to study whether bispecific antibodies inhibit tumor growth in an NPG mouse model of transplanted tumor constructed by subcutaneously inoculating human breast cancer cells HCC1954.

[0347] Female NPG mice at the age of seven to eight weeks were selected and HCC1954 cells in the logarithmic growth stage were collected. 5×10.sup.6 HCC1954 cells and Matrigel (Corning, Cat. No. 354234) were mixed in a volume ratio of 1:1 and then inoculated subcutaneously on the right back of each NPG mouse. After 6 days of tumor growth, the mice were randomly divided into three groups with six mice in each group according to the tumor volumes and weights. The treated groups were intraperitoneally administered with Herceptin at a dose of 3 mg/kg and AB7K7 at a dose of 1 mg/kg, respectively, and the control group was administered with a PBS solution of the same volume, twice a week. The day of administration was recorded as Day 0. The maximum diameter (D) and the minimum diameter (d) of the tumor were measured weekly. The volume (mm.sup.3) of the tumor of each group and the tumor growth inhibition rate (TGI) (%) of each treated group were calculated using the formulas as shown in Example 3.1.

[0348] As shown in FIG. 4-2, on Day 21 of administration, the average tumor volume of the PBS control group was 1311.35±215.70 mm.sup.3; the average tumor volume of the treated group administrated with Herceptin was 273.98±60.10 mm.sup.3; the average tumor volume of the treated group administrated with AB7K7 was 1243.20±340.31 mm.sup.3, which was not different from the average tumor volume of the control group. In summary, the results show that AB7K7 did not inhibit the growth of HCC1954 subcutaneous tumors in the absence of human immune cells, indicating that bispecific antibody AB7K7 needs to be mediated by immune effector cells so as to kill tumor cells, unlike Herceptin, which primarily depends on FcγR-mediated ADCC or CDC effects to kill tumor cells. Thus, it is proved that Fc variants contained in AB7K7 cannot bind to FcγR, which avoids mediating systemic activation of T cells caused by extensive expression of its receptor FcγR, resulting in higher drug safety.

4.3 Evaluation of Toxicity of Bispecific Antibodies to Normal Cynomolgus Monkeys

[0349] Adult cynomolgus monkeys (purchased from Guangzhou Xiangguan Biotechnology Co., Ltd.) at the age of 3-4 years and with the weight of 3-4 kg were divided into three groups with one monkey in each group, wherein the three groups were a vehicle control group, an AB7K7 treated group and an AB7K8 treated group. The groups were administrated via intravenous drip by a peristaltic pump for 1 hour on Day 0 (D0), Day 7 (D7), Day 21 (D21), and Day 28 (D28), respectively, for a total of four doses, and the drug dose was gradually increased each time. The monkeys were weighed weekly. The dose amount and volume administered are shown in Table 4-1.

TABLE-US-00012 TABLE 4-1 Dosing schedule for cynomolgus monkey acute toxicity evaluation To-be-tested Group drugs name Dose volume Dose amount G1 Vehicle control D0: 5 mL/kg N/A group D7: 5 mL/kg D21: 10 mL/kg D28: 10 mL/kg G2 AB7K7 D0: 5 mL/kg D0: 0.06 mg/kg D7: 5 mL/kg D7: 0.3 mg/kg D21: 10 mL/kg D21: 1.5 mg/kg D28: 10 mL/kg D28: 3 mg/kg G3 AB7K8 D0: 5 mL/kg D0: 0.04 mg/kg D7: 5 mL/kg D7: 0.2 mg/kg D21: 10 mL/kg D21: 1 mg/kg D28: 10 mL/kg D28: 2 mg/kg

[0350] On D0, after administration, cynomolgus monkeys in the AB7K8 treated group exhibited somnolence and pupil contraction and recovered to normal the next day while there was no abnormality in the other groups. On D7, after administration, cynomolgus monkeys in the AB7K7 treated group exhibited vomiting symptoms 2-3 hours after administration and recovered to normal the next day of administration while there was no abnormality in the other groups. On D21, after administration, cynomolgus monkeys in the AB7K7 treated group exhibited symptoms of vomiting food 3 hours after administration and excreted jelly-like feces, cynomolgus monkeys in the AB7K8 treated group exhibited symptoms of vomiting food 1 hour after administration, and cynomolgus monkeys in both groups recovered to normal on the second day after administration; On D28, after administration, cynomolgus monkeys in both AB7K7 treated group and AB7K8 treated group exhibited vomiting symptoms 40 to 50 minutes later and excreted feces 3 hours later, in which jelly-like mucus was found; cynomolgus monkeys in the AB7K7 treated group excreted watery feces with fishy smelling; and 24 hours later, all the animals recovered to normal and ingested normally. The body weight change of cynomolgus monkeys is shown in FIG. 4-3, wherein the arrow represents the administration time. It can be seen that the body weight of each group does not change too much and fluctuates within the normal physiological range.

[0351] The different degree of diarrhea observed in this example may be related to the expression of related receptors in the gut, which is supposed to be caused by the imbalance of chloride ion in the gut caused by the inhibition of heterodimer of Her1/Her2 or Her2/Her3 by bispecific antibodies, which belongs to the extension of pharmacological action and can recover to normal after 24 hours of administration. Cynomolgus monkeys were still well tolerated when administrated with AB7K7 at a high dose of 3 mg/kg. The results of pharmacodynamics test in mice show that AB7K7 at a low dose shows a good anti-tumor effect, which indicates that AB7K7 has a wide treatment window and high safety.

Example 5 Pharmacokinetics Study of Anti-Her2×CD3 Antibodies

5.1 In Vivo Pharmacokinetics Test of Bispecific Antibody AB7K7 in SD Mice

[0352] AB7K7 was administered to four healthy Sprague-Dawley (SD) rats (purchased from Shanghai Salccas Laboratory Animals Co., Ltd.,) via the tail vein at a dose of 1 mg/kg. The blood sampling time points were Hour 1, Hour 3, Hour 6, Hour 24, Hour 72, Hour 96, Hour 120, Hour 168, Hour 216 and Hour 264, respectively. A certain amount of whole blood was taken at each time point, the serum was separated, and then the drug concentration in the serum was detected by two ELISA methods.

[0353] Method I. Plates were coated with the anti-AB7K7 antibody A (Ampsource Biopharma Shanghai Inc., mouse-anti-herceptin) at a concentration of 0.5 AB7K7 was formulated at concentrations of 100 ng/mL, 50 ng/mL, 25 ng/mL, 12.5 ng/mL, 6.25 ng/mL, 3.125 ng/mL and 1.56 ng/mL, separately. Standard curves were established. HRP-labeled anti-AB7K7 antibody B (Ampsource Biopharma Shanghai Inc., anti-herceptin-HRP) was used at a concentration of 1:5000, and developed with TMB. The pharmacokinetics parameters were calculated using software PKSolver. Specific parameters are shown in Table 5-1.

[0354] Method II. The drug concentration in the serum of the SD rats was detected. Plates were coated with the anti-AB7K7 antibody A (Ampsource Biopharma Shanghai Inc., mouse-anti-herceptin) at a concentration of 0.5 AB7K7 was formulated at concentrations of 5 ng/mL, 2.5 ng/mL, 1.25 ng/mL, 0.625 ng/mL, 0.3125 ng/mL, 0.156 ng/mL and 0.078 ng/mL, separately. Standard curves were established. Mouse anti-human IgG Fc-HRP (Ampsource Biopharma Shanghai Inc.) was added at a concentration of 1:5000, and developed with TMB. The pharmacokinetics parameters were calculated using software PKSolver. Specific parameters are shown in Table 5-2.

[0355] FIG. 5-1 shows the blood drug concentration of AB7K7 in the body of rats detected using two different detection methods. It can be seen that the blood drug concentrations obtained by detecting the concentration of AB7K7in blood using two different detection methods were basically the same, and the calculated pharmacokinetics parameters were roughly equivalent, which indicates that AB7K7 can be metabolized in the form of intact molecules in vivo, thereby ensuring the biological function of AB7K7.

TABLE-US-00013 TABLE 5-1 Pharmacokinetics parameters of bispecific antibody AB7K7 in SD rats (Method 1) AUC t.sub.1/2 0-inf_ob Vz_obs Cl_obs AB7K7 (h) (ng/mL*h) (μg)/(ng/mL) (μg)/(ng/mL)/h Pharmacokinetics 42.10 550236.77 0.02351 3.811E−4 parameter

TABLE-US-00014 TABLE 5-2 Pharmacokinetics parameters of bispecific antibody AB7K7 in SD rats (Method 2) AUC t½ 0-inf_ob Vz_obs Cl_obs AB7K7 (h) (ng/mL*h) (μg)/(ng/mL) (μg)/(ng/mL)/h Pharmacokinetics 41.02 706126.89 0.01720 2.899E−4 parameter

5.2 In Vivo Pharmacokinetics Test on Bispecific Antibody AB7K7 in NPG Model Mice

[0356] NPG mice (purchased from Beijing Vitalstar Biotechnology Co., Ltd.) were inoculated with HCC1954 cells (purchased from the Institute of Cells, Chinese Academy of Sciences) one week before administration, with an inoculum density of 3.5×10.sup.6/mouse. CIK cells were resuscitated two days before administration, cultured for 24 hours and then collected and injected intravenously into mice. The mice were randomly divided into three groups with four mice in each group. The three treated groups were administrated at doses of 0.3 mg/kg, 1 mg/kg and 3 mg/kg, respectively. The blood sampling time points were Hour 1, Hour 3, Hour 6, Hour 24, Hour 48, Hour 72, Hour 96, Hour 120, Hour 168, Hour 216 and Hour 264, respectively. A certain amount of whole blood was taken at each time point, the serum was separated, and then the drug concentration in the serum was detected by ELISA.

[0357] Plates were coated with the anti-AB7K7 antibody A (Ampsource Biopharma Shanghai Inc., mouse-anti-herceptin) at a concentration of 0.5 AB7K7 was formulated at concentrations of 100 ng/mL, 50 ng/mL, 25 ng/mL, 12.5 ng/mL, 6.25 ng/mL, 3.125 ng/mL and 1.56 ng/mL, separately. Standard curves were established. HRP-labeled anti-AB7K7 antibody B (Ampsource Biopharma Shanghai Inc., mouse-anti-herceptin) was used at a concentration of 1:5000, and developed with TMB. The pharmacokinetics parameters were calculated using software PKSolver. Specific parameters are shown in Table 5-3.

[0358] It can be seen from Table 5-3 that the pharmacokinetics parameters of AB7K7 in NPG model mice were not significantly different from those in SD rats.

TABLE-US-00015 TABLE 5-3 Pharmacokinetics parameters of bispecific antibody AB7K7 in NPG model mice Parameter t.sub.1/2 AUC0-inf_obs Vz_ob Cl_obs Group (h) (ng/mL*h) (μg)/(ng/mL) (μg)/(ng/mL)/h 0.3 mg/kg IV.sup.  39.54 79932.39 0.004872 8.968E−05 1 mg/kg IV 42.70 597036.63 0.002461 3.996E−05 3 mg/kg IV 46.03 2171649.41 0.002292 3.469E−05

5.3 In Vivo Pharmacokinetics Test on Bispecific Antibody AB7K8 in SD Rats

[0359] AB7K8 was administered to three healthy SD rats via the tail vein at doses of 1 mg/kg and 3 mg/kg, respectively. The blood sampling time points were Hour 0.25, Hour 0.5, Hour 1, Hour 2, Hour 3, Hour 4, Hour 5, and Hour 7, respectively. A certain amount of whole blood was taken at each time point, the serum was separated, and then the drug concentration in the serum was detected by ELISA.

[0360] Plates were coated with the anti-AB7K8 antibody C (Ampsource Biopharma Shanghai Inc., mouse-anti-herceptin) at a concentration of 2.5 AB7K8 was formulated at concentrations of 25 ng/mL, 12.5 ng/mL, 6.25 ng/mL, 3.125 ng/mL, 1.56 ng/mL, and 0.78 ng/mL, separately. Standard curves were established. HRP-labeled anti-his antibody (Ampsource Biopharma Shanghai Inc.) was used at a concentration of 1:5000, and developed with TMB. The pharmacokinetics parameters were calculated using software PKSolver. Specific parameters are shown in Table 5-4.

[0361] The pharmacokinetics parameter T.sub.1/2 of AB7K8 were almost the same at two doses, which indicates that AB7K8 showed linear metabolic kinetics in SD rats. Since AB7K8 does not contain Fc, T.sub.1/2 of AB7K8 is very short, about twenty times shorter than T.sub.1/2 of AB7K7.

TABLE-US-00016 TABLE 5-4 Pharmacokinetics parameters of bispecific antibody AB7K8 in SD rats AUC t.sub.1/2 0-inf_ob Vz_obs Cl_obs AB7K8 (h) (ng/mL*h) (μg)/(ng/mL) (μg)/(ng/mL)/h 1 mg/kg IV 2.27 4623.14 0.17082 0.05191 3 mg/kg IV 1.98 20608.77 0.10220 0.03579

5.4 In Vivo Pharmacokinetics Test on Bispecific Antibody AB7K in SD Rats

[0362] AB7K was administered to four healthy SD rats via the tail vein at a dose of 0.8 mg/kg. The blood sampling time points were Hour 2, Hour 24, Hour 48, Hour 72, Hour 96, Hour 120, Hour 144, Hour 168, Hour 216 and Hour 264, respectively. A certain amount of whole blood was taken at each time point, the serum was separated, and then the drug concentration in the serum was detected by two ELISA methods.

[0363] Method I. Plates were coated with the anti-AB7K antibody A (Ampsource Biopharma Shanghai Inc., mouse-anti-herceptin) at a concentration of 1 AB7K was formulated at concentrations of 20 ng/mL, 10 ng/mL, 5 ng/mL, 2.5 ng/mL, 1.25 ng/mL, 0.625 ng/mL and 0.3125 ng/mL, separately. Standard curves were established. 25 ng/mL of biotin-labeled human CD3E&CD3D (Acro, Cat. No. CDD-H82W0) was added, incubated for 1 hour, and after that, HRP-labeled streptavidin (BD Pharmingen, Cat. No. 554066) diluted at a factor of 1:500 was added, and developed with TMB. The pharmacokinetics parameters were calculated using software PKSolver. Specific parameters are shown in Table 5-5.

[0364] Method II. Plates were coated with the anti-AB7K antibody A (Ampsource Biopharma Shanghai Inc., mouse-anti-herceptin) at a concentration of 1 AB7K was formulated at concentrations of 20 ng/mL, 10 ng/mL, 5 ng/mL, 2.5 ng/mL, 1.25 ng/mL, 0.625 ng/mL and 0.3125 ng/mL, separately. Standard curves were established. Mouse anti-human IgG Fc-HRP (diluted at 1:10000) (Ampsource Biopharma Shanghai Inc.) was added, incubated in an incubator for 1 hour, and developed with TMB.

[0365] FIG. 5-2 shows the blood drug concentration of AB7K in rats by using two different detection methods. From the results, it is showed that the difference between the two detection methods is large. The concentrations of the first two points (2h, 1D) on the curve were close, but after the next day, the concentrations detected by the two methods differed greatly, which is supposed to be caused by the fact that the linkage peptide between the heavy chains of anti-CD3 scFv and anti-Her2 antibodies was broken. AB7K is structurally unstable in vivo and thus can't play its biological function, while the improved AB7K7 can metabolize in complete form in vivo and thus can play its biological function normally.

TABLE-US-00017 TABLE 5-5 Pharmacokinetics parameters of bispecific antibody AB7K AUC t.sub.1/2 0-inf_ob Vz_obs Cl_obs AB7K (h) (ng/mL*h) (μg)/(ng/mL) (μg)/(ng/mL)/h Pharmacokinetics 60.47 1022788.69 0.01726 1.985E−4 parameter

5.5 In Vivo Pharmacokinetics Test on Bispecific Antibodies AB7K7 and AB7K8 in Cynomolgus Monkeys

[0366] Female cynomolgus monkeys (purchased from Guangzhou Xiangguan Biotechnology Co., Ltd.) with the weight of 3-4 kg were divided into three groups with one monkey in each group. The first group (G1-1) was a blank control group; the second group (G2-1) was an AB7K7 treated group administrated at a dose of 0.3 mg/kg; and the third group (G3-1) was an AB7K8 treated group administrated at a dose of 0.2 mg/kg. The blood sampling time points were Minute 15,

[0367] Hour 1, Hour 3, Hour 6, Hour 24, Hour 48, Hour 72, Hour 96, Hour 144, Hour 192, Hour 240 and Hour 288, respectively, a total of 13 time points. Serum was collected from blood and frozen at −80° C. The concentration of the drug in serum was determined by ELISA.

[0368] Plates were coated with the anti-AB7K7 antibody A (Ampsource Biopharma Shanghai Inc., mouse-anti-herceptin) at a concentration of 0.5 AB7K7 was formulated at concentrations of 100 ng/mL, 50 ng/mL, 25 ng/mL, 12.5 ng/mL, 6.25 ng/mL, 3.125 ng/mL and 1.56 ng/mL, separately. Standard curves were established. HRP-labeled anti-AB7K7 antibody B (Ampsource Biopharma Shanghai Inc., mouse-anti-herceptin) was used at a concentration of 1:5000, and developed with TMB. The pharmacokinetics parameters were calculated using software PKSolver. Specific parameters are shown in Table 5-6.

[0369] FIG. 5-3 shows the blood drug concentration of AB7K7 in rats. T.sub.1/2 of AB7K7 in the normal cynomolgus monkey was only about eight hours. Pharmacokinetics parameters of AB7K8 could not be calculated due to too few points on the concentration-time curve of AB7K8. However, it can be seen from the concentration-time curve that the half-life of AB7K7 in the normal cynomolgus monkey was much longer than the half-life of AB7K8.

TABLE-US-00018 TABLE 5-6 Pharmacokinetics parameters of bispecific antibody AB7K7 in cynomolgus monkeys AUC t.sub.1/2 0-inf_obs Vz_obs Cl_obs AB7K7 (h) (ng/mL*h) (μg)/(ng/mL) (μg)/(ng/mL)/h Pharmacokinetics 7.95 87995.48 0.1563 0.01364 parameter

5.6 Evaluation of Abilities of Bispecific Antibodies to Bind to FcRn by ELISA

[0370] Each bispecific antibody was diluted with the PBS solution to a concentration of 10 μg/ml and added to 96-well plates, 100 μl per well. The plates were coated at 4° C. overnight. The plates were then blocked with 1% skimmed milk powder for 1 hour at room temperature. Biotin-labeled FcRn proteins (ACRO Biosystem, Cat. No. FCM-H8286) were diluted using diluents at pH of 6.0 and 7.0, respectively, with a 4-fold gradient for a total of 11 concentration gradients. The 96-well plates were then washed with PBST of the same pH, and then each of the bispecific antibodies diluted with the diluent at the same pH was added. Control wells without antibodies were set. Incubated for 1 hour at room temperature. Plates were washed with the

[0371] PBST solution of the same pH, streptavidin-HRP (BD, Cat. No. 554066) was added to 96-well plates, 100 μl per well, and the plates were incubated for 0.5 hours at room temperature. 96-well plates were washed with PBST, and TMB was added to the plates, 100 μl per wells. Color development was performed at room temperature for 15 minutes, and then 0.2 M H.sub.2SO.sub.4 was added to stop the color development reaction. The light absorbance values at A450-620 nm were measured by a microplate reader. Analysis was performed by software OriginPro 8, and the EC.sub.50 values for the binding of bispecific antibodies to FcRn were calculated.

[0372] The results show that the ability of each antibody to bind to FcRn was different under different pH conditions, and it is analyzed in conjunction with data of in vivo PK that the half-life of bispecific antibody AB7K7 was longer than the half-life of AB7K but shorter than the half-life of Herceptin, which may be more favorable for clinical application (FIGS. 5-4 and 5-5). Table 5-7 and Table 5-8 show the detection results of the ability of each antibody to bind to FcRn at pHs 6.0 and 7.0, respectively.

TABLE-US-00019 TABLE 5-7 Detection of abilities of bispecific antibodies AB7K, AB7K5 and AB7K7 to bind to FcRn at pH 6.0 Herceptin AB7K AB7K5 AB7K7 EC50 (μg/ml) 2.591 0.8027 1.706 0.4630

TABLE-US-00020 TABLE 5-8 Detection of abilities of bispecific antibodies AB7K, AB7K5 and AB7K7 to bind to FcRn at pH 7.0 Herceptin AB7K AB7K5 AB7K7 EC50 (μg/ml) ~287.1 1.651 13.43 4.838

[0373] Based on the results of the preceding researches on six anti-Her2×CD3 bispecific antibodies from multiple aspects, it can be determined that such bispecific antibodies with the configuration of scFv1-scFv2-Fc as AB7K7 are easy to be prepared and simple and efficient to be purified and has good stability during preparation and storage. More favorably, such bispecific antibodies have the significant advantages of weak non-specific killing effects on normal cells, controlled toxic side effects that may be caused by over-activation of effector cells, and good druggability.

[0374] Some preferred amino acid sequences of the VH domain and its complementarity determining regions (HCDR1, HCDR2 and HCDR3) and amino acid sequences of the VL domain and its complementarity determining regions (LCDR1, LCDR2 and LCDR3) of a first single-chain Fv against Her2 are listed in Table 5-9, wherein amino acid residues contained in the CDRs are defined according to the Kabat rule. The amino acid composition of the linker peptide between VH and VL of anti-Her2 scFv is (GGGGS)n, wherein n=1, 2, 3, 4 or 5.

TABLE-US-00021 TABLE 5-9 Amino acid sequences of the anti-Her2  ScFv contained in the bispecific  antibody and its CDRs Her2 SEQ ID NO: 9 HCDR1 DTYIH SEQ ID NO: 10 HCDR2 RIYPTNGYTRYADSVKG SEQ ID NO: 11 HCDR3 WGGDGFYAMDY SEQ ID NO: 12 LCDR1 RASQDVNTAVA SEQ ID NO: 13 LCDR2 SASFLYS SEQ ID NO: 14 LCDR3 QQHYTTPPT SEQ ID NO: 15 VH EVQLVESGGGLVQPGGSLRLSCAASGFNIK DTYIHWVRQAPGKGLEWVARIYPTNGYTRY ADSVKGRFTISADTSKNTAYLQMNSLRAED TAVYYCSRWGGDGFYAMDYWGQGTLVTVSS SEQ ID NO: 16 VL DIQMTQSPSSLSASVGDRVTITCRASQDVN TAVAWYQQKPGKAPKLLIYSASFLYSGVPS RFSGSRSGTDFTLTISSLQPEDFATYYCQQ HYTTPPTFGQGTKVEIK SEQ ID NO: 17 HCDR1 DYTMD SEQ ID NO: 18 HCDR2 DVNPNSGGSIYNQRFKG SEQ ID NO: 19 HCDR3 NLGPSFYFDY SEQ ID NO: 20 LCDR1 KASQDVSIGVA SEQ ID NO: 21 LCDR2 SASYRYT SEQ ID NO: 22 LCDR3 QQYYIYPYT SEQ ID NO: 23 VH EVQLVESGGGLVQPGGSLRLSCAASGFTFT DYTMDWVRQAPGKGLEWVADVNPNSGGSIY NQRFKGRFTLSVDRSKNTLYLQMNSLRAED TAVYYCARNLGPSFYFDYWGQGTLVTVSS SEQ ID NO: 24 VL DIQMTQSPSSLSASVGDRVTITCKASQDVS IGVAWYQQKPGKAPKLLIYSASYRYTGVPS RFSGSGSGTDFTLTISSLQPEDFATYYCQQ YYIYPYTFGQGTKVEIK SEQ ID NO: 25 HCDR1 DTYIH SEQ ID NO: 26 HCDR2 RIYPTNGYTRYDPKFQD SEQ ID NO: 27 HCDR3 WGGDGFYAMDY SEQ ID NO: 28 LCDR1 KASQDVNTAVA SEQ ID NO: 29 LCDR2 SASFRYT SEQ ID NO: 30 LCDR3 QQHYTTPPT SEQ ID NO: 31 VH QVQLQQSGPELVKPGASLKLSCTASGFNIK DTYIHWVKQRPEQGLEWIGRIYPTNGYTRY DPKFQDKATITADTSSNTAYLQVSRLTSED TAVYYCSRWGGDGFYAMDYWGQGASVTVSS SEQ ID NO: 32 VL DIVMTQSHKFMSTSVGDRVSITCKASQDVN TAVAWYQQKPGHSPKLLIYSASFRYTGVPD RFTGSRSGTDFTFTISSVQAEDLAVYYCQQ HYTTPPTFGGGTKVEIK

[0375] The anti-CD3 scFv binds to an effector cell at an EC.sub.50 value greater than about 50 nM, or greater than 100 nM, or greater than 300 nM, or greater than 500 nM in an in vitro FACS binding assay; more preferably, the second single-chain Fv of the bispecific antibody is capable of binding to human CD3 and specifically binding to CD3 of a cynomolgus monkey or a rhesus monkey.

[0376] Some preferred amino acid sequences of the VH domain and its complementarity determining regions (HCDR1, HCDR2 and HCDR3) and amino acid sequences of the VL domain and its complementarity determining regions (LCDR1, LCDR2 and LCDR3) of the anti-CD3 scFv are listed in Table 5-10, wherein amino acid residues contained in the CDRs are defined according to the Kabat rule. The amino acid composition of the linker peptide between VH and VL of the anti-CD3 scFv is (GGGGS)n, where n =1, 2, 3, 4 or 5.

TABLE-US-00022 TABLE 5-10 Amino acid sequences of the anti-CD3 ScFv  contained in the bispecific antibody and its CDRs CD3-3 SEQ ID NO: 33 HCDR1 TYAMN SEQ ID NO: 34 HCDR2 RIRSKYNNYATYYADSVKD SEQ ID NO: 35 HCDR3 HGNFGNSYVSWFAY SEQ ID NO: 36 LCDR1 RSSTGAVTTSNYAN SEQ ID NO: 37 LCDR2 GTNKRAP SEQ ID NO: 38 LCDR3 ALWYSNLWV SEQ ID NO: 39 VH EVQLLESGGGLVQPGGSLKLSCAASGFT FNTYAMNWVRQAPGKGLEWVARIRSKYN NYATYYADSVKDRFTISRDDSKNTAYLQ MNNLKTEDTAVYYCVRHGNFGNSYVSWF AYWGQGTLVTVSS SEQ ID NO: 40 VL ELVVTQEPSLTVSPGGTVTLTCRSSTGA VTTSNYANWVQQKPGQAPRGLIGGTNKR APGTPARFSGSLLGGKAALTLSGVQPED EAEYYCALWYSNLWVFGGGTKLTVL CD3-4 SEQ ID NO: 41 HCDR1 KYAMN SEQ ID NO: 42 HCDR2 RIRSKYNNYATYYADSVKD SEQ ID NO: 43 HCDR3 HGNFGNSYISYWAY SEQ ID NO: 44 LCDR1 GSSTGAVTSGYYPN SEQ ID NO: 45 LCDR2 GTKFLAP SEQ ID NO: 46 LCDR3 ALWYSNRWV SEQ ID NO: 47 VH EVQLLESGGGLVQPGGSLKLSCAASGFT FNKYAMNWVRQAPGKGLEWVARIRSKYN NYATYYADSVKDRFTISRDDSKNTAYLQ MNNLKTEDTAVYYCVRHGNFGNSYISYW AYWGQGTLVTVSS SEQ ID NO: 48 VL ELVVTQEPSLTVSPGGTVTLTCGSSTGA VTSGYYPNWVQQKPGQAPRGLIGGTKFL APGTPARFSGSLLGGKAALTLSGVQPED EAEYYCALWYSNRWVFGGGTKLTVL

[0377] The linker peptide linking the anti-Her2 scFv and the anti-CD3 scFv consists of a flexible peptide and a rigid peptide; preferably, the amino acid composition structure of the flexible peptide has a general formula of G.sub.xS.sub.y(GGGGS).sub.z, wherein x, y and z are integers greater than or equal to 0 and x+y+z≥1. The rigid peptide is derived from a full-length sequence (as shown in SEQ ID NO: 49) consisting of amino acids 118 to 145 at the carboxyl terminus of natural human chorionic gonadotropin β-subunit or a truncated fragment thereof; preferably, the composition of the CTP rigid peptide is SSSSKAPPPS (CTP.sup.1). Some preferred amino acid sequences of the linker peptide linking the anti-Her2 scFv and the anti-CD3 scFv are listed in Table 5-11.

TABLE-US-00023 TABLE 5-11 Amino acid sequences of the linker peptide  linking the anti-Her2 scFv and the  anti-CD3 scFv SEQ ID NO: 50 G.sub.2(GGGGS).sub.3CTP.sup.1 GGGGGGSGGGGSGGGGS SSSSKAPPPS SEQ ID NO: 51 (GGGGS).sub.3CTP.sup.1 GGGGSGGGGSGGGGSSS SSKAPPPS SEQ ID NO: 52 GS(GGGGS).sub.2CTP.sup.1 GSGGGGSGGGGSSSSSK APPPS SEQ ID NO: 53 (GGGGS).sub.1CTP.sup.4 GGGGSSSSSKAPPPSLP SPSRLPGPSDTPILPQ

[0378] The Fc fragment is linked to the anti-CD3 scFv directly or by a linker peptide, and the linker peptide includes 1-20 amino acids, and preferably selected from the following amino acids: Gly(G), Ser(S), Ala(A) and Thr(T), more preferably selected from Gly (G) and Ser (S), most preferably, the linker peptide consists of (GGGGS)n, wherein n=1, 2, 3 or 4.

[0379] The Fc fragment is preferably selected from heavy chain constant regions of human IgG1, IgG2, IgG3 and IgG4, and more particularly selected from heavy chain constant regions of human IgG1 or IgG4; and Fc is mutated to modify the properties of the bispecific antibody molecule, e.g., to show reduced affinity to at least one of human FcγRs (FcγRI, FcγRIIa or FcγRIIIa) and C1q, a reduced effector cell function, or a reduced complement function. In addition, the Fc fragment may also contain amino acid substitutions that change one or more other characteristics (such as an ability of binding to an FcRn receptor, the glycosylation of the antibody or the charge heterogeneity of the antibody).

[0380] Some amino acid sequences of the Fc fragment with one or more amino acid mutations are listed in Table 5-12.

TABLE-US-00024 TABLE 5-12 Amino acid sequences of Fc from human IgG Amino acid sequence of constant region of  IgG1 Fc (L234A, L235A) mutant (EU numbering) SEQ ID  DKTHTCPPCP APEAAGGPSV FLFPPKPKDT LMISRTPEVT NO: 54 CVVVDVSHED PEVKFNWYVD GVEVHNAKTK PREEQYNSTY RVVSVLTVLH QDWLNGKEYK CKVSNKALPA PIEKTISKAK GQPREPQVYT LPPSREEMTK NQVSLTCLVK GFYPSDIAVE WESNGQPENN YKTTPPVLDS DGSFFLYSKL TVDKSRWQQG NVFSCSVMHE ALHNHYTQKS LSLSPGK Amino acid sequence of constant region of IgG1 (L234A, L235A, T250Q, N297A, P331S, M428L, K447.sup.-)  mutant (EU numbering) SEQ ID  DKTHTCPPCP APEAAGGPSV FLFPPKPKDQ LMISRTPEVT NO: 55 CVVVDVSHED PEVKFNWYVD GVEVHNAKTK PREEQYASTY RVVSVLTVLH QDWLNGKEYK CKVSNKALPA SIEKTISKAK GQPREPQVYT LPPSRDELTK NQVSLTCLVK GFYPSDIAVE WESNGQPENN YKTTPPVLDS DGSFFLYSKL TVDKSRWQQG NVFSCSVLHE ALHNHYTQKS LSLSPG

[0381] All the publications mentioned in the present disclosure are incorporated herein by reference as if each publication is separately incorporated herein by reference. In addition, it should be understood that those skilled in the art, who have read the disclosure, can make various changes or modifications to the present disclosure, and these equivalent forms fall within the scope of the appended claims.