Anti-CD28×anti-ENPP3 antibody

Abstract

This specification provides novel anti-CD28 × anti-ENPP3 antibodies and methods for using such antibodies for the treatment of ENPP3-related cancers. The target anti-CD28 × anti-ENPP3 antibodies can agonist-conjugate CD28 costimulatory molecules on T cells and ENPP3 on tumor cells. Thus, such antibodies selectively enhance antitumor activity at tumor sites while minimizing peripheral toxicity. The target antibodies provided herein are particularly useful in combination with other anticancer therapies, for example, bispecific antibodies for the treatment of ENPP3-related cancers.

Description

[Technical Field]

[0001] Claim of priority This application claims priority and interest in U.S. Provisional Application No. 63 / 496,367, filed April 14, 2023, the contents of which are incorporated herein by reference in their entirety. [Background technology]

[0002] Antibody-based therapies have been used to successfully treat a variety of diseases, including cancer. An increasingly widespread approach being explored is the manipulation of a single immunoglobulin molecule that co-engages with two different antigens. Such alternative antibody forms that engage with two different antigens are often referred to as bispecific antibodies. One particular approach for bispecific antibodies is that the bispecific antibody is CD3 + The goal is to design a first binding domain that engages with CD3 and a second binding domain that engages with an antigen associated with cancer cells or an antigen upregulated on cancer cells (e.g., ENPP3) to redirect T cells and destroy cancer cells.

[0003] However, TILs lose their cytotoxic capacity over time due to upregulation of inhibitory immune checkpoints. While checkpoint blockade has demonstrated increased clinical response rates compared to other treatment options, many patients still fail to achieve a response to checkpoint blockade. Engagement of costimulatory receptors on TILs could provide a positive signal that could overcome the negative signaling of immune checkpoints. Preclinical and clinical studies of agonist costimulatory receptor antibodies have indeed demonstrated that costimulatory receptor agonism can lead to superior antitumor responses by activating T cells to attack tumor cells.

[0004] Improving antitumor activity by specifically destroying tumor cells while minimizing peripheral toxicity is also important for cancer therapy. In this context, it is crucial that costimulatory signals are delivered only to T cells in the presence of target tumor cells. However, costimulatory receptor agonism with single-specific full-length antibodies is likely not discriminatory with respect to TILs, peripheral T cells, or autoantigen-reactive T cells that contribute to autoimmune toxicity.

[0005] Ectonucleotide pyrophosphatase / phosphodiesterase family member 3 (ENPP3) is a type II transmembrane glycoprotein belonging to the nucleotide pyrophosphatase / phosphodiesterase family. These proteins are involved in the hydrolysis of extracellular nucleotides and also possess ATPase and ATP pyrophosphatase activity. ENPP3 is overexpressed in a variety of cancers, including most renal cell carcinomas and some hepatic cancers. Because ENPP3 is overexpressed on a variety of tumors, it is a candidate for targeted therapy development. While immunotherapies targeting ENPP3 are being attempted, novel immune response-enhancing compositions are still needed to treat ENPP3-related cancers. [Overview of the project]

[0006] This specification provides novel anti-CD28 × anti-ENPP3 antibodies and methods for using such antibodies for the treatment of ENPP3-related cancers. The target anti-CD28 × anti-ENPP3 antibodies can agonist-conjugate CD28 costimulatory molecules on T cells and ENPP3 on tumor cells. Thus, such antibodies selectively enhance antitumor activity at tumor sites while minimizing peripheral toxicity. The target antibodies provided herein are particularly useful in combination with other anticancer therapies, for example, bispecific antibodies for the treatment of ENPP3-related cancers.

[0007] In a first aspect, the herein provides an anti-CD28 × anti-ENPP3 heterodimer antibody comprising a) a first monomer, b) a second monomer, and c) a light chain. The first monomer comprises i) a single-stranded variable fragment (scFv) and ii) a first Fc domain, the scFv being covalently bound to the N-terminus of the first Fc domain using a domain linker. The second monomer comprises VH1-CH1-hinge-CH2-CH3 from the N-terminus to the C-terminus, where VH1 is the first variable heavy domain and CH2-CH3 is the second Fc domain. The light chain comprises VL1-CL from the N-terminus to the C-terminus, where VL1 is the first variable light domain and CL is the constant light domain. The scFv comprises a second VH domain (VH2), an scFv linker, and a second variable light domain (VL2). VH1 and VL1 together form a first antigen-binding domain (ABD), and VH2 and VL2 together form a second ABD. Furthermore, one of the first and second ABDs is a CD28-binding domain, and the other of the first and second ABDs is an ectonucleotide pyrophosphatase / phosphodiesterase 3 (ENPP3)-binding domain.

[0008] In some embodiments, the scFv includes a VH2-scFv linker-VL2 from the N-terminus to the C-terminus.

[0009] In exemplary embodiments, the first ABD is an ENPP3 binding domain, and the second ABD is a CD28 binding domain. In some embodiments, VH1 and VL1 are VH and VL of any of the ENPP3 binding domains in Figure 22 or their variants. In some embodiments, VH2 and VL2 are selected from one of the following: 1) VH and VL of any of the CD28 binding domains in Figures 15, 18, and 21 or their variants, or 2) (i) VH or its variant in Figure 15 or 16 and (ii) VL or its variant in Figure 15 or 17.

[0010] In some embodiments, the first Fc domain and the second Fc domain are variant Fc domains.

[0011] In some embodiments, the first and second Fc domains include a set of heterodimerized scuba riants selected from the following heterodimerized variants:S364K / E357Q:L368D / K370S;S364K:L368D / K370S;S364K:L368E / K370S;D401K:T411E / K360E / Q362E; and T366W:T366S / L368A / Y407V, numbered according to EU numbering. In some embodiments, the first and second Fc domains include the heterodimerized scuba riant S364K / E357Q:L368D / K370S, numbered according to EU numbering.

[0012] In certain embodiments, the first and second Fc domains each include one or more attenuation variants. In exemplary embodiments, the one or more attenuation variants include E233P / L234V / L235A / G236del / S267K, the numbering following EU numbering.

[0013] In some embodiments, one of the first or second monomer further comprises one or more pI variants. In some embodiments, the CH1-hinge-CH2-CH3 of the second monomer comprises pI variants N208D / Q295E / N384D / Q418E / N421D, the numbering following EU numbering.

[0014] In some embodiments, the CH1-hinge-CH2-CH3 of the second monomer contains the amino acid variant E233P / L234V / L235A / G236del / S267K / L368D / K370S / N208D / Q295E / N384D / Q418E / N421D, and the first Fc domain contains the amino acid variant E233P / L234V / L235A / G236del / S267K / S364K / E357Q, with the numbering following EU numbering. In exemplary embodiments, the first and second variant Fc domains each further contain the amino acid variant 428L / 434S.

[0015] In some embodiments, the scFv linker is GKPGSGKPGSGKPGSGKPGS.

[0016] In another embodiment, provided herein is an anti-CD28 × anti-ENPP3 heterodimer antibody comprising a) a first monomer, b) a second monomer, c) a first light chain, and d) a second light chain. The first monomer comprises VH1-CH1-first domain linker-scFv-second domain linker-CH2-CH3 from N-terminus to C-terminus, where VH1 is the first variable heavy domain and CH2-CH3 is the first Fc domain. The second monomer comprises VH1-CH1-hinge-CH2-CH3 from N-terminus to C-terminus, where CH2-CH3 is the second Fc domain. The first and second light chains each comprise VL1-CL from N-terminus to C-terminus, where VL1 is the first variable light domain and CL is the constant light domain. scFv comprises a second VH domain (VH2), an scFv linker, and a second variable light domain (VL2). The VH1 of the first monomer and the VL1 of the first light chain, and the VH1 of the second monomer and the VL1 of the second light chain, each form a first antigen-binding domain (ABD), while VH2 and VL2 form a second ABD. Furthermore, the first ABD is an ectonucleotide pyrophosphatase / phosphodiesterase 3 (ENPP3) binding domain and the second ABD is a CD28 binding domain, or the first ABD is a CD28 binding domain and the second ABD is an ENPP3 binding domain.

[0017] In some embodiments, the scFv includes a VH2-scFv linker-VL2 from the N-terminus to the C-terminus.

[0018] In exemplary embodiments, the first ABD is an ENPP3 binding domain, and the second ABD is a CD28 binding domain. In some embodiments, VH1 and VL1 are selected from one of the following: 1) VH and VL of any of the ENPP3 binding domains in Figure 22 or their variants. In some embodiments, VH2 and VL2 are selected from one of the following: 1) VH and VL of any of the CD28 binding domains in Figures 15, 18, and 21 or their variants, or 2) (i) VH or its variant in Figure 15 or 16 and (ii) VL or its variant in Figure 15 or 17.

[0019] In some embodiments, the first Fc domain and the second Fc domain are variant Fc domains.

[0020] In some embodiments, the first and second Fc domains include a set of heterodimerized scuba riants selected from the following heterodimerized variants:S364K / E357Q:L368D / K370S;S364K:L368D / K370S;S364K:L368E / K370S;D401K:T411E / K360E / Q362E; and T366W:T366S / L368A / Y407V, numbered according to EU numbering. In some embodiments, the first and second Fc domains include the heterodimerized scuba riant S364K / E357Q:L368D / K370S, numbered according to EU numbering.

[0021] In certain embodiments, the first and second Fc domains each include one or more attenuation variants. In exemplary embodiments, the one or more attenuation variants include E233P / L234V / L235A / G236del / S267K, the numbering following EU numbering.

[0022] In some embodiments, one of the first or second monomer further comprises one or more pI variants. In some embodiments, the CH1-hinge-CH2-CH3 of the second monomer comprises the pI variant N208D / Q295E / N384D / Q418E / N421D, and the numbering follows the EU numbering.

[0023] In some embodiments, the CH1-hinge-CH2-CH3 of the second monomer comprises the amino acid variants E233P / L234V / L235A / G236del / S267K / L368D / K370S / N208D / Q295E / N384D / Q418E / N421D, and the first Fc domain comprises the amino acid variants E233P / L234V / L235A / G236del / S267K / S364K / E357Q, and the numbering follows the EU numbering. In an exemplary embodiment, the first and second variant Fc domains each further comprise the amino acid variant 428L / 434S.

[0024] In some embodiments, the scFv linker is GKPGSGKPGSGKPGSGKPGS.

[0025] In another aspect, provided herein is an anti-CD28×anti-ENPP3 heterodimeric antibody comprising a) a first monomer, b) a second monomer, c) a first light chain, and d) a second light chain. The first monomer comprises, from the N-terminus to the C-terminus, VH1-CH1-hinge-CH2-CH3-domain linker-scFv, where VH1 is the first variable heavy domain and CH2-CH3 is the first Fc domain. The second monomer comprises, from the N-terminus to the C-terminus, VH1-CH1-hinge-CH2-CH3, where VH1 is the first variable heavy domain and CH2-CH3 is the second Fc domain. The first and second light chains each comprise, from the N-terminus to the C-terminus, VL1-CL, where VL1 is the first variable light domain and CL is the constant light domain. The scFv comprises a second VH domain (VH2), an scFv linker, and a second variable light domain (VL2). The VH1 of the first monomer and the VL1 of the first light chain and the VH1 of the second monomer and the VL1 of the second light chain each form a first antigen-binding domain (ABD), and VH2 and VL2 form a second ABD. Further, the first ABD is an ectonucleotide pyrophosphatase / phosphodiesterase 3 (ENPP3) binding domain, and the second ABD is a CD28 binding domain, or the first ABD is a CD28 binding domain and the second ABD is an ENPP3 binding domain.

[0026] In some embodiments, the scFv comprises, from the N-terminus to the C-terminus, VH2-scFv linker-VL2. In some embodiments, the scFv comprises, from the N-terminus to the C-terminus, VL2-scFv linker-VH2.

[0027] In exemplary embodiments, the first ABD is an ENPP3 binding domain, and the second ABD is a CD28 binding domain. In some embodiments, VH1 and VL1 are selected from one of the following: 1) VH and VL of any of the ENPP3 binding domains in Figure 22 or their variants. In some embodiments, VH2 and VL2 are selected from one of the following: 1) VH and VL of any of the CD28 binding domains in Figures 15, 18, and 21 or their variants, or 2) (i) VH or its variant in Figure 15 or 16 and (ii) VL or its variant in Figure 15 or 17.

[0028] In some embodiments, the first Fc domain and the second Fc domain are variant Fc domains.

[0029] In some embodiments, the first and second Fc domains include a set of heterodimerized scuba riants selected from the following heterodimerized variants:S364K / E357Q:L368D / K370S;S364K:L368D / K370S;S364K:L368E / K370S;D401K:T411E / K360E / Q362E; and T366W:T366S / L368A / Y407V, numbered according to EU numbering. In some embodiments, the first and second Fc domains include the heterodimerized scuba riant S364K / E357Q:L368D / K370S, numbered according to EU numbering.

[0030] In certain embodiments, the first and second Fc domains each include one or more attenuation variants. In exemplary embodiments, the one or more attenuation variants include E233P / L234V / L235A / G236del / S267K, the numbering following EU numbering.

[0031] In some embodiments, one of the first or second monomer further comprises one or more pI variants. In some embodiments, the CH1-hinge-CH2-CH3 of the second monomer comprises pI variants N208D / Q295E / N384D / Q418E / N421D, the numbering following EU numbering.

[0032] In some embodiments, the CH1-hinge-CH2-CH3 of the second monomer contains the amino acid variant E233P / L234V / L235A / G236del / S267K / L368D / K370S / N208D / Q295E / N384D / Q418E / N421D, and the first Fc domain contains the amino acid variant E233P / L234V / L235A / G236del / S267K / S364K / E357Q, with the numbering following EU numbering. In exemplary embodiments, the first and second variant Fc domains each further contain the amino acid variant 428L / 434S. In some embodiments, the scFv linker is GKPGSGKPGSGKPGSGKPGS.

[0033] In another embodiment, the herein provides a heterodimer antibody comprising a) a first monomer, b) a second monomer, c) a first light chain, and d) a second light chain. The first monomer comprises VH1-CH1-domain linker-VH1-CH1-hinge-CH2-CH3 from N-terminus to C-terminus, where each of VH1 is a first variable weight domain and CH2-CH3 is a first Fc domain. The second monomer comprises scFv-domain linker-CH2-CH3 from N-terminus to C-terminus, where VH1 is a first variable weight domain and CH2-CH3 is a second Fc domain. The first and second light chains each comprise VL1-CL from N-terminus to C-terminus, where VL1 is a first variable light domain and CL is a constant light domain. scFv comprises a second VH domain (VH2), an scFv linker, and a second variable light domain (VL2). The VH1 of the first monomer and the VL1 of the first light chain, and the VH1 of the second monomer and the VL1 of the second light chain, each form a first antigen-binding domain (ABD), while VH2 and VL2 form a second ABD. Furthermore, the first ABD is an ectonucleotide pyrophosphatase / phosphodiesterase 3 (ENPP3) binding domain and the second ABD is a CD28 binding domain, or the first ABD is a CD28 binding domain and the second ABD is an ENPP3 binding domain.

[0034] In some embodiments, the first and second Fc domains include a set of heterodimerized scuba riants selected from the following heterodimerized variants:S364K / E357Q:L368D / K370S;S364K:L368D / K370S;S364K:L368E / K370S;D401K:T411E / K360E / Q362E; and T366W:T366S / L368A / Y407V, numbered according to EU numbering. In some embodiments, the first and second Fc domains include the heterodimerized scuba riant S364K / E357Q:L368D / K370S, numbered according to EU numbering.

[0035] In certain embodiments, the first and second Fc domains each include one or more attenuation variants. In exemplary embodiments, the one or more attenuation variants include E233P / L234V / L235A / G236del / S267K, the numbering following EU numbering.

[0036] In some embodiments, one of the first or second monomer further comprises one or more pI variants. In some embodiments, the CH1-hinge-CH2-CH3 of the second monomer comprises pI variants N208D / Q295E / N384D / Q418E / N421D, the numbering following EU numbering.

[0037] In some embodiments, the CH1-hinge-CH2-CH3 of the second monomer contains the amino acid variant E233P / L234V / L235A / G236del / S267K / L368D / K370S / N208D / Q295E / N384D / Q418E / N421D, and the first Fc domain contains the amino acid variant E233P / L234V / L235A / G236del / S267K / S364K / E357Q, with the numbering following EU numbering. In exemplary embodiments, the first and second variant Fc domains each further contain the amino acid variant 428L / 434S.

[0038] In some embodiments, the scFv linker is GKPGSGKPGSGKPGSGKPGS.

[0039] In another embodiment, provided herein is a bispecific antibody comprising an ENPP3-binding domain comprising a)i) a first variable heavy domain (VH1) and ii) a first variable light domain (VL1), and an anti-CD28-binding domain comprising b)i) a second variable heavy domain (VH2) and ii) a second variable light domain (VL2). In some embodiments, VH1 and VL1 are selected from one of the following: 1) VH and VL or variants thereof of any of the ENPP3-binding domains in Figure 22. In some embodiments, VH2 and VL2 are selected from one of the following: 1) VH and VL or variants thereof of any of the CD28-binding domains in Figures 15, 18, and 21, or 2) (i) VH or a variant thereof in Figure 15 or 16 and (ii) VL or a variant thereof in Figure 15 or 17.

[0040] In some embodiments of the bispecific antibody, the first Fc domain and the second Fc domain are variant Fc domains, respectively.

[0041] In some embodiments of the bispecific antibody, the first and second Fc domains include a set of heterodimerized scuba riants selected from the following heterodimerized variants:S364K / E357Q:L368D / K370S;S364K:L368D / K370S;S364K:L368E / K370S;D401K:T411E / K360E / Q362E; and T366W:T366S / L368A / Y407V, numbered according to EU numbering. In some embodiments, the first and second Fc domains include the heterodimerized scuba riant S364K / E357Q:L368D / K370S, numbered according to EU numbering.

[0042] In certain embodiments, the first and second Fc domains each include one or more attenuation variants. In exemplary embodiments, the one or more attenuation variants include E233P / L234V / L235A / G236del / S267K, the numbering following EU numbering.

[0043] In some embodiments, one of the first or second monomer further comprises one or more pI variants. In some embodiments, the CH1-hinge-CH2-CH3 of the second monomer comprises pI variants N208D / Q295E / N384D / Q418E / N421D, the numbering following EU numbering.

[0044] Furthermore, this specification also provides nucleic acid compositions comprising nucleic acids encoding antibodies described herein, expression vector compositions comprising such nucleic acids, host cells for producing antibodies comprising the expression vector compositions, and methods for producing antibodies.

[0045] In another embodiment, the foregoing provides a method for treating ENPP3-associated cancer in a patient requiring treatment for ENPP3-associated cancer, the method comprising administering to the patient an anti-CD28 × anti-ENPP3 bispecific antibody as described herein.

[0046] In another embodiment, the foregoing provides a method for treating ENPP3-associated cancer in a patient requiring treatment for ENPP3-associated cancer, the method comprising administering to the patient an anti-CD28 × anti-ENPP3 bispecific antibody and an anti-CD3 × anti-ENPP3 bispecific antibody as described herein. [Brief explanation of the drawing]

[0047] [Figure 1A] The sequences of human and mouse CD28 are shown. [Figure 1B] The sequence of CD28 from a cynomolgus monkey is shown. Such CD28 is useful for developing cross-reactive CD28 antigen-binding domains to facilitate clinical development. [Figure 2A] To facilitate clinical development, the antigenic sequences of several antigens used in the present invention, including those of both humans and cynomolgus monkeys, are shown to facilitate the development of antigen-binding domains that bind to both humans and cynomolgus monkeys. [Figure 2B]To facilitate clinical development, the antigenic sequences of several antigens used in the present invention, including those of both humans and cynomolgus monkeys, are shown to facilitate the development of antigen-binding domains that bind to both humans and cynomolgus monkeys. [Figure 3A] This shows useful pairs of heterodimerized variant sets (including skew and pI variants). [Figure 3B] This shows useful pairs of heterodimerized variant sets (including skew and pI variants). [Figure 3C] This shows useful pairs of heterodimerized variant sets (including skew and pI variants). [Figure 3D] This shows useful pairs of heterodimerized variant sets (including skew and pI variants). [Figure 3E] This shows useful pairs of heterodimerized variant sets (including skew and pI variants). [Figure 3F] Useful pairs of heterodimerization variant sets (including skew and pI variants) are shown. Figure 3F shows variants in which there is no variant corresponding to "monomer 2". Such variants are pI variants and can be used alone for either monomer of a bispecific antibody (e.g., ENPP3×CD28 bsAb), or may be included in the non-scFv side in a form that utilizes scFv as a component, for example, and a suitable charged scFv linker can be used for the second monomer that utilizes scFv as the CD28 binding domain. Preferred charged linkers are shown in Figure 6. The heterodimer yield (%) and CH3 Tm (°C) of preferred Fc heterodimerization variants have been previously described (see, for example, Figure 8 of U.S. Patent Application No. 2019 / 0248898). [Figure 4]A list of isosteric variant antibody constant regions and their respective substitutions is shown. pI_(-) indicates a lower pI variant, while pI_(+) indicates a higher pI variant. These variants can be combined arbitrarily and independently with other variants, including the heterodimerized variants outlined herein. [Figure 5] The document exhibits useful attenuation variants (also referred to as "knockout" or "KO" variants) that reduce FcγR binding. In some embodiments, such attenuation variants are present in the Fc domains of both monomers of the target antibody described herein. In other embodiments, the attenuation variant is present in only one variant's Fc domain. [Figure 6] As described herein, numerous charged scFv linkers are used to increase or decrease the pI of a heterodimer bispecific antibody (e.g., ENPP3×CD28 bsAb) that utilizes one or more scFv components. (+H) positive linkers are particularly utilized herein, especially in conjunction with the anti-CD28 VL and VH sequences shown herein. A single-charged, single-prior-art scFv linker is referred to as "Whitlow" from Whitlow et al., Protein Engineering 6(8):989-995 (1993). It should be noted that this linker was used to reduce aggregation and improve proteolytic stability in scFv. Such charged scFv linkers may be used in any of the target antibody forms disclosed herein that include scFv (e.g., 1+1 Fab-scFv-Fc and 2+1 Fab2-scFv-Fc forms). [Figure 7] Numerous exemplary domain linkers are shown. In some embodiments, these linkers are used to link a single-stranded Fv to an Fc chain. In some embodiments, these linkers can be combined in any orientation. For example, the GGGGS linker can be combined with a “bottom half-hinge” linker at the N-terminus or C-terminus. [Figure 8]This invention presents a bispecific antibody platform particularly useful for ENPP3×CD28 bsAb. While the platform is shown in relation to the 1+1 Fab-scFv-Fc format, it can also be applied to the use of other bispecific antibody formats. [Figure 9A] Based on human IgG1, the sequences of several useful heterodimer ENPP3×CD28 bsAb skeletons, which do not contain cytokine sequences, are shown. [Figure 9B] Based on human IgG1, the sequences of several useful heterodimer ENPP3×CD28 bsAb skeletons, which do not contain cytokine sequences, are shown. [Figure 9C] Based on human IgG1, the sequences of several useful heterodimer ENPP3×CD28 bsAb skeletons, which do not contain cytokine sequences, are shown. [Figure 9D]Based on human IgG1, the sequences of several useful heterodimer ENPP3×CD28 bsAb skeletons, which do not contain cytokine sequences, are shown. Heterodimer Fc skeleton 1, based on human IgG1 (356E / 358M allotype), includes the L368D / K370S scuba riant and Q295E / N384D / Q418E / N421D pI variant on the first heterodimer Fc chain, the S364K / E357Q scuba riant on the second heterodimer Fc chain, and the E233P / L234V / L235A / G236del / S267K attenuation variant on both chains. The heterodimer Fc skeleton 2 is based on human IgG1 (356E / 358M allotype) and includes the L368D / K370S scuba riant and Q295E / N384D / Q418E / N421D pI variant on the first heterodimer Fc chain, the S364K scuba riant on the second heterodimer Fc chain, and the E233P / L234V / L235A / G236del / S267K attenuation variant on both chains. The heterodimer Fc skeleton 3 is based on human IgG1 (356E / 358M allotype) and includes the L368E / K370S scuba riant and Q295E / N384D / Q418E / N421D pI variant on the first heterodimer Fc chain, the S364K scuba riant on the second heterodimer Fc chain, and the E233P / L234V / L235A / G236del / S267K attenuation variant on both chains. The heterodimer Fc skeleton 4 is based on human IgG1 (356E / 358M allotype) and includes the K360E / Q362E / T411E scuba riant and Q295E / N384D / Q418E / N421D pI variant on the first heterodimer Fc chain, the D401K scuba riant on the second heterodimer Fc chain, and the E233P / L234V / L235A / G236del / S267K attenuation variant on both chains.The heterodimer Fc skeleton 5 is based on human IgG1 (356D / 358L allotype) and includes the L368D / K370S scuba riant and Q295E / N384D / Q418E / N421D pI variant on the first heterodimer Fc chain, the S364K / E357Q scuba riant on the second heterodimer Fc chain, and the E233P / L234V / L235A / G236del / S267K attenuation variant on both chains. The heterodimer Fc skeleton 6 is based on human IgG1 (356E / 358M allotype) and includes the L368D / K370S scuba riant and Q295E / N384D / Q418E / N421D pI variant on the first heterodimer Fc chain, the S364K / E357Q scuba riant on the second heterodimer Fc chain, and the E233P / L234V / L235A / G236del / S267K attenuation variant and the N297A variant that removes glycosylation on both chains. The heterodimer Fc skeleton 7 is based on human IgG1 (356E / 358M allotype) and includes the L368D / K370S scuba riant and Q295E / N384D / Q418E / N421D pI variant on the first heterodimer Fc chain, the S364K / E357Q scuba riant on the second heterodimer Fc chain, and the E233P / L234V / L235A / G236del / S267K decrement variant and the N297S variant that removes glycosylation on both chains. The heterodimer Fc skeleton 8 is based on human IgG4 and includes the L368D / K370S scuba riant and Q295E / N384D / Q418E / N421D pI variants on the first heterodimer Fc chain, the S364K / E357Q scuba riant on the second heterodimer Fc chain, and the S228P (according to EU numbering, S241P in Kabat) variant which reduces Fab arm exchange (as known in the art) on both chains. The heterodimer Fc skeleton 9 is based on human IgG2 and includes the L368D / K370S scuba riant and Q295E / N384D / Q418E / N421D pI variants on the first heterodimer Fc chain, and the S364K / E357Q scuba riant on the second heterodimer Fc chain.The heterodimer Fc skeleton 10 is based on human IgG2 and includes the L368D / K370S scuba riant and Q295E / N384D / Q418E / N421D pI variant on the first heterodimer Fc chain, the S364K / E357Q scuba riant on the second heterodimer Fc chain, and the S267K attenuation variant on both chains. The heterodimer Fc skeleton 11 is based on human IgG1 (356E / 358M allotype) and includes the L368D / K370S scuba riant and Q295E / N384D / Q418E / N421D pI variant on the first heterodimer Fc chain, the S364K / E357Q scuba riant on the second heterodimer Fc chain, and the E233P / L234V / L235A / G236del / S267K attenuation variant and the M428L / N434S Xtend variant on both chains. The heterodimer Fc skeleton 12 is based on human IgG1 (356E / 358M allotype) and includes the L368D / K370S scuba rianto on the first heterodimer Fc chain, the S364K / E357Q scuba rianto and P217R / P229R / N276K pI variant on the second heterodimer Fc chain, and the E233P / L234V / L235A / G236del / S267K attenuation variant on both chains. The heterodimer Fc skeleton 13, based on human IgG1 (356D / 358L allotype), includes the L368D / K370S scuba riant and Q295E / N384D / Q418E / N421D pI variant on the first heterodimer Fc chain, the S364K / E357Q scuba riant on the second heterodimer Fc chain, and the E233P / L234V / L235A / G236del / S267K attenuation variant and M428L / N434S Xtend variant on both chains. The heterodimer Fc skeleton 14, based on human IgG1 (356E / 358M allotype), includes the L368D / K370S scuba riant and Q295E / N384D / Q418E / N421D pI variant on the first heterodimer Fc chain, the S364K / E357Q scuba riant on the second heterodimer Fc chain, and the E233P / L234V / L235A / G236del / S267K attenuation variant and the M428L / N434A Xtend variant on both chains.The heterodimer Fc skeleton 15 is based on human IgG1 (356D / 358L allotype) and includes the L368D / K370S scuba riant and Q295E / N384D / Q418E / N421D pI variant on the first heterodimer Fc chain, the S364K / E357Q scuba riant on the second heterodimer Fc chain, and the E233P / L234V / L235A / G236del / S267K attenuation variant and the M428L / N434A Xtend variant on both chains. Sequences that are 90, 95, 98, and 99% identical (as defined herein) to the described sequences, and / or contain 1, 2, 3, 4, 5, 6, 7, 8, 9, or 10 or more additional amino acid substitutions (as understood by those skilled in the art, compared to parent human IgG1 (or IgG2 or IgG4 depending on the skeleton), compared to the “parent” in the figure which already contains numerous amino acid modifications) are included in each of these skeletons. That is, the described skeletons may contain additional amino acid modifications (usually amino acid substitutions) in addition to or as alternatives to the skew, pI, and attenuation variants contained within the skeleton in this figure. Furthermore, the skeletons shown herein may include deletions of C-terminal glycine (K446_) and / or lysine (K447_). C-terminal glycine and / or lysine deletions may be intentionally manipulated to reduce heterogeneity or in relation to a predetermined bispecific form such as the mAb-scFv form. Furthermore, C-terminal glycine and / or lysine deletions may occur naturally, for example, during production and storage. [Figure 10] Exemplary sequences of the heterodimer ENPP3×CD28 bsAb skeleton for use in the 2+1 mAb-scFv form are shown. The form shown here is based on the heterodimer Fc skeleton 1 shown in Figure 9, but further including G446_ on monomer 1(-) and G446_ / K447_ on monomer 2(+). It should be noted that any of the additional skeletons shown in Figure 9 can be applied to use in the 2+1 mAb-scFv form, whether or not they include K447_ on one or both strands. It should be noted that these sequences may further include the M428L / N434S variant. [Figure 11]The sequence of "CH1" used in an embodiment of ENPP3×CD28 bsAb is shown. [Figure 12] This shows the arrangement of "hinge" used in an embodiment of ENPP3×CD28 bsAb. [Figure 13] This shows the constant domain of the congeneral light chain used in ENPP3×CD28 bsAb, which utilizes the Fab binding domain. [Figure 14A] The present invention presents a bispecific form. It presents a "1+1 Fab-scFv-Fc" form having a first Fab arm that binds to a first antigen and a second scFv arm that binds to a second antigen. The 1+1 Fab-scFv-Fc form comprises a first monomer containing a first heavy chain variable region (VH1) covalently (optionally via a linker) attached to the N-terminus of a first heterodimer Fc skeleton, a second monomer containing a single-stranded Fv covalently (optionally via a linker) attached to the N-terminus of a second corresponding heterodimer Fc skeleton, and a third monomer containing a light chain variable region covalently attached to a light chain constant domain (the light chain variable region is complementary to VH1). [Figure 14B] The present invention presents a bispecific form. It presents a "2+1 Fab2-scFv-Fc" form having a first Fab arm and a second Fab-scFv arm, where Fab binds to the first antigen and scFv binds to the second antigen. The 2+1 Fab2-scFv-Fc form comprises a first monomer containing a first heavy chain variable region (VH1) covalently bonded (optionally via a linker) to the N-terminus of a first heterodimer Fc skeleton, a second monomer containing VH1 covalently bonded (optionally via a linker) to a single-stranded Fv covalently bonded (optionally via a linker) to the N-terminus of a second corresponding heterodimer Fc skeleton, and a third monomer containing a light chain variable region covalently bonded to a light chain constant domain (the light chain variable region is complementary to VH1). [Figure 14C]The present invention presents a bispecific form. It presents a "1+1 common light chain" or "1+1 CLC" form having a first Fc containing a first Fab arm that binds to a first antigen and a second Fc containing a second Fab arm that binds to a second antigen. The 1+1 CLC form comprises a first monomer containing VH1-CH1-hinge-CH2-CH3, a second monomer containing VH2-CH1-hinge-CH2-CH3, and a third monomer containing VL-CL. VL pairs with VH1 to form a binding domain having a first antigen-binding specificity, and VL pairs with VH2 to form a binding domain having a second antigen-binding specificity. [Figure 14D] The present invention presents a bispecific form. It presents a "2+1 common light chain" or "2+1 CLC" form having a first Fc containing two Fab arms that bind to a first antigen and a second Fc containing one Fab arm that binds to a second antigen. The 2+1 CLC form comprises a first monomer containing VH1-CH1-hinge-VH1-CH1-hinge-CH2-CH3, a second monomer containing VH2-CH1-hinge-CH2-CH3, and a third monomer containing VL-CL. VL pairs with the first and second VH1 to form a binding domain having a first antigen-binding specificity, and VL pairs with VH2 to form a binding domain having a second antigen-binding specificity. [Figure 14E] The present invention presents a bispecific form. It presents a "2+1 mAb-scFv" form having a first Fc containing an N-terminal Fab arm that binds to a first antigen, and a second Fc containing an N-terminal Fab arm that binds to a first antigen and a C-terminal scFv that binds to a second antigen. The 2+1 mAb-scFv form comprises a first monomer containing VH1-CH1-hinge-CH2-CH3, a second monomer containing VH1-CH1-hinge-CH2-CH3-scFv, and a third monomer containing VL-CL. VL pairs with the first and second VH1 to form a binding domain having binding specificity for the first antigen. [Figure 14F]The present invention presents a bispecific form. It presents a "2+1 Fab2-Fc×scFv-Fc" form having a first Fc domain containing two N-terminal Fab arms that bind to a first antigen and a second Fc domain containing an N-terminal scFv that binds to a second antigen. The 2+1 Fab2-Fc×scFv-Fc form comprises a first monomer containing VH1-CH1-VH2-CH1-hinge-CH2-CH3, a second monomer containing scFv-domain linker-CH2-CH3, and third and fourth monomers containing VL-CL. VL pairs with VH1 and VH2 to form a binding domain having binding specificity to the first antigen. [Figure 14G] This shows the dual-specific form of the present invention. It is a dual scFv, an additional dual-specific form. [Figure 14H] The present invention presents a dual-specific form. An additional dual-specific form is a 1-arm scFv-mAb. [Figure 14I] The present invention presents a bispecific form. An additional bispecific form is scFv-mAb. [Figure 14J] This presents a bispecific form of the present invention. It is a bispecific mAb of an additional bispecific form. [Figure 14K] The present invention presents a dual-specific form. An additional dual-specific form is a 1-arm central-scFv. [Figure 14L] The present invention presents a bispecific form. An additional bispecific form is mAb-Fv. [Figure 14M] The present invention presents a bispecific form. The additional bispecific form is the central-Fv. [Figure 14N] This presents a dual-specific form of the present invention. It is a trident of additional dual-specific forms. [Figure 15]The variable heavy chain and variable light chain sequences of 1A7, an exemplary phage-derived CD28-binding domain, are shown, as well as the sequence of XENP28428, an anti-CD28 mAb based on the IgG1 scaffold with 1A7 and the E233P / L234V / L235A / G236del / S267K attenuation variant. CDRs are indicated by underlines, and slashes indicate the boundary(s) between the variable region and the constant domain. As is the case with all sequences described herein that contain CDRs, the precise location of the CDR may vary slightly depending on the numbering used, as shown in Table 2. Therefore, this specification includes not only underlined CDRs but also CDRs contained within VH and VL domains using other numbering systems. Furthermore, for all sequences in the figures, these VH and VL sequences may be used in either scFv or Fab format. [Figure 16A] The sequence of the affinity-optimized variable weight domain derived from anti-CD28 clone 1A7 is shown. [Figure 16B] The sequence of the affinity-optimized variable weight domain derived from anti-CD28 clone 1A7 is shown. [Figure 16C] The sequence of the affinity-optimized variable weight domain derived from anti-CD28 clone 1A7 is shown. [Figure 16D] The sequence of the affinity-optimized variable weight domain derived from anti-CD28 clone 1A7 is shown. [Figure 16E] The sequence of the affinity-optimized variable weight domain derived from anti-CD28 clone 1A7 is shown. [Figure 16F] The sequence of the affinity-optimized variable weight domain derived from anti-CD28 clone 1A7 is shown. It should be noted that the variable weight domain can pair with any of the other variable light domains shown in Figures 15, 17, and 18, including sequence numbers XXX-YYY. [Figure 17A] The sequence of the affinity-optimized variable light domain derived from anti-CD28 clone 1A7 is shown. [Figure 17B] The sequence of the affinity-optimized variable light domain derived from anti-CD28 clone 1A7 is shown. [Figure 17C]The sequence of the affinity-optimized variable light domain derived from anti-CD28 clone 1A7 is shown. [Figure 17D] The sequence of the affinity-optimized variable light domain derived from anti-CD28 clone 1A7 is shown. [Figure 17E] The sequence of the affinity-optimized variable light domain derived from anti-CD28 clone 1A7 is shown. [Figure 17F] The sequence of the affinity-optimized variable light domain derived from anti-CD28 clone 1A7 is shown. [Figure 17G] The sequence of the affinity-optimized variable light domain derived from anti-CD28 clone 1A7 is shown. [Figure 17H] The sequence of the affinity-optimized variable light domain derived from anti-CD28 clone 1A7 is shown. [Figure 17I] The sequence of the affinity-optimized variable light domain derived from anti-CD28 clone 1A7 is shown. It should be noted that the variable heavy domain can pair with any of the other variable light domains shown in Figures 15, 16, and 18, including sequence numbers XXX-YYY. [Figure 18A] The following shows an exemplary affinity-optimized sequence of 1A7 VH / VL pairs. [Figure 18B] The following shows an exemplary affinity-optimized sequence of 1A7 VH / VL pairs. [Figure 18C] Exemplary affinity-optimized 1A7 VH / VL pair sequences are shown. It should be noted that these pairs can be formalized as Fab or scFv. Furthermore, in scFv form, these pairs can be formalized with VHVL orientation or VLVH orientation. [Figure 19A] The consensus framework region (FR) and complementarity determination region (CDR) (in Kabat) of the variable weight and variable light domain variants of anti-CD28 clone 1A7 are shown. [Figure 19B] The consensus framework region (FR) and complementarity determination region (CDR) (in Kabat) of the variable weight and variable light domain variants of anti-CD28 clone 1A7 are shown. [Figure 20]This shows the binding affinity in relation to the exemplary affinity operation 1A7VH / VL pair and scFv (in relation to the 1+1 Fab-scFv-FcbsAb form). [Figure 21A] The variable heavy chain sequence and variable light chain sequence of the additional CD28-binding domain used in ENPP3×CD28 of the present invention are shown. [Figure 21B] The variable heavy chain sequence and variable light chain sequence of the additional CD28-binding domain used in ENPP3×CD28 of the present invention are shown. [Figure 21C] The variable heavy chain sequence and variable light chain sequence of the additional CD28-binding domain used in ENPP3×CD28 of the present invention are shown. [Figure 21D] The variable heavy chain sequence and variable light chain sequence of the additional CD28-binding domain used in ENPP3×CD28 of the present invention are shown. [Figure 21E] The variable heavy chain sequence and variable light chain sequence of the additional CD28-binding domain used in ENPP3×CD28 of the present invention are shown. [Figure 21F] The variable heavy chain sequence and variable light chain sequence of the additional CD28-binding domain used in ENPP3×CD28 of the present invention are shown. [Figure 21G] The variable heavy chain sequence and variable light chain sequence of the additional CD28-binding domain used in ENPP3×CD28 of the present invention are shown. [Figure 21H] The variable heavy chain and variable light chain sequences of the additional CD28-binding domains used in ENPP3×CD28 of the present invention are shown. As is the case with all sequences described herein that contain CDRs, the precise location of the CDR may vary slightly depending on the numbering used, as shown in Table 2. Therefore, this specification includes not only the underlined CDRs but also CDRs contained within the VH and VL domains using other numbering systems. Furthermore, for all sequences in the figures, these VH and VL sequences may be used in either scFv or Fab format. [Figure 22A]The following shows exemplary variable heavy chain and variable light chain sequences of the ENPP3 binding domain used in ENPP3×CD28 bsAb of the present invention. [Figure 22B] The following shows exemplary variable heavy chain and variable light chain sequences of the ENPP3 binding domain used in ENPP3×CD28 bsAb of the present invention. [Figure 22C] The following shows exemplary variable heavy chain and variable light chain sequences of the ENPP3 binding domain used in ENPP3×CD28 bsAb of the present invention. [Figure 22D] The following shows exemplary variable heavy chain and variable light chain sequences of the ENPP3 binding domain used in ENPP3×CD28 bsAb of the present invention. [Figure 22E] The following shows exemplary variable heavy chain and variable light chain sequences of the ENPP3 binding domain used in ENPP3×CD28 bsAb of the present invention. [Figure 22F] The following shows exemplary variable heavy chain and variable light chain sequences of the ENPP3 binding domain used in ENPP3×CD28 bsAb of the present invention. [Figure 22G] The following shows exemplary variable heavy chain and variable light chain sequences of the ENPP3 binding domain used in ENPP3×CD28 bsAb of the present invention. [Figure 22H] The following shows exemplary variable heavy chain and variable light chain sequences of the ENPP3 binding domain used in ENPP3×CD28 bsAb of the present invention. [Figure 22I] The following shows exemplary variable heavy chain and variable light chain sequences of the ENPP3 binding domain used in ENPP3×CD28 bsAb of the present invention. [Figure 22J] The following shows exemplary variable heavy chain and variable light chain sequences of the ENPP3 binding domain used in ENPP3×CD28 bsAb of the present invention. [Figure 22K]The variable heavy chain and variable light chain sequences of the exemplary ENPP3 binding domain used in ENPP3×CD28 bsAb of the present invention are shown. As is the case with all sequences described herein that contain a CDR, the location of the CDR may be determined according to the numbering scheme shown in Table 2. Furthermore, for all sequences in the figure, these VH and VL sequences may be used in either scFv or Fab format. [Figure 23] A) Classical T cell / APC interaction, and B) Reproduction of classical T cell / APC interaction by combining CD3 bispecific antibody and CD28 bispecific antibody. In classical T cell / APC interaction, there is a first signal (signal 1) provided by TCR reactivity with peptide-MHC and a second signal (signal 2) provided by crosslinking of CD28 by CD80 / CD86 expressed on APC, which together fully activate the T cell. In contrast, treatment with CD3 bispecificity provides only the first signal. The CD28 signal may be provided by CD28 bispecificity with the intention of promoting activation and proliferation via CD28 costimulation. In some embodiments, TAA1 and TAA2 may be different antigens. In some embodiments, TAA1 and TAA2 may be the same antigen but may be different epitopes. In some embodiments, TAA1 and TAA2 may be the same antigen and the same epitope. [Figure 24A] The sequence of ENPP3×CD28 bsAb, an example of the 1+1 Fab-scFv-Fc format, is shown. [Figure 24B] The sequence of ENPP3×CD28 bsAb, an example of the 1+1 Fab-scFv-Fc format, is shown. [Figure 24C] The sequence of ENPP3×CD28 bsAb, an example of the 1+1 Fab-scFv-Fc format, is shown. [Figure 24D] The sequence of ENPP3×CD28 bsAb, an example of the 1+1 Fab-scFv-Fc format, is shown. [Figure 24E]The sequence of ENPP3×CD28 bsAb, an example of the 1+1 Fab-scFv-Fc format, is shown. [Figure 24F] The sequence of ENPP3×CD28 bsAb, an example of the 1+1 Fab-scFv-Fc format, is shown. [Figure 24G] The sequence of ENPP3×CD28 bsAb, an example of the 1+1 Fab-scFv-Fc format, is shown. [Figure 24H] The sequence of an exemplary ENPP3×CD28 bsAb in the 1+1 Fab-scFv-Fc format is shown. Slashes indicate the boundary(s) between the variable domain and other domains, e.g., the domain linker and the constant domain. In addition, the nomenclature convention indicates the orientation of the scFv from the N-terminus to the C-terminus. As is the case with all sequences described herein that contain a CDR, the location of the CDR may be determined according to the numbering scheme shown in Table 2. It should be noted that ENPP3×CD28 bsAb may utilize variable, Fc, and constant domain sequences that are 90, 95, 98, and 99% identical (as defined herein) and / or contain 1, 2, 3, 4, 5, 6, 7, 8, 9, or 10 amino acid substitutions. In addition, each sequence outlined herein may or may not contain the M428L / N434S variant in one or preferably both Fc domains, and the inclusion of M428L / N434S results in a longer half-life in serum. [Figure 25A] The following is an exemplary sequence of an anti-CD3 binding domain suitable for use in a CD3 bispecific antibody that can be combined with the CD28 bispecific antibody of the present invention. [Figure 25B] The following is an exemplary sequence of an anti-CD3 binding domain suitable for use in a CD3 bispecific antibody that can be combined with the CD28 bispecific antibody of the present invention. [Figure 25C] The following is an exemplary sequence of an anti-CD3 binding domain suitable for use in a CD3 bispecific antibody that can be combined with the CD28 bispecific antibody of the present invention. [Figure 25D]The following is an exemplary sequence of an anti-CD3 binding domain suitable for use in a CD3 bispecific antibody that can be combined with the CD28 bispecific antibody of the present invention. [Figure 25E] The following is an exemplary sequence of an anti-CD3 binding domain suitable for use in a CD3 bispecific antibody that can be combined with the CD28 bispecific antibody of the present invention. [Figure 25F] The following is a sequence of an exemplary anti-CD3 binding domain suitable for use in a CD3 bispecific antibody that can be combined with the CD28 bispecific antibody of the present invention. The CDR is indicated by an underline, the scFv linker by a double underline (in the sequence, the scFv linker is the positively charged scFv(GKPGS)4 linker (SEQ ID NO: XXX), but as will be understood to those skilled in the art, this linker can be replaced by other linkers, including uncharged or negatively charged linkers, some of which are shown in Figure 6), and a slash indicates the boundary(s) of a variable domain. In addition, the nomenclature indicates the orientation of the scFv from the N-terminus to the C-terminus. As is true for all sequences described herein that contain a CDR, the precise location of the CDR may vary slightly depending on the numbering used, as shown in Table 2; therefore, this specification includes not only underlined CDRs but also CDRs contained within the VH and VL domains using other numbering systems. Furthermore, for all sequences shown in the figure, these VH and VL sequences can be used in either scFv or Fab format. [Figure 26] This shows the induction of A) IL2 secretion and B) IFNγ secretion from purified T cells incubated with RXF-393 (ENPP3+) tumor cells and exemplary ENPP3×CD3 bsAb at 1 μg / ml and exemplary ENPP3×CD28 bsAb at escalating doses. [Figure 27] This shows IL-2 release when T cells purified from 14 unique T cell donors were incubated with TUHR10TKB ENPP3+ target cells in a 1:1 E:T ratio and treated with 10 μg / ml ENPP3×CD28 bsAb in combination with 1 μg / ml exemplary ENPP3×CD3. [Figure 28]This study demonstrates that ENPP3×CD28 enhances target cell killing by VMRC-RCW target cells via (A)ENPP3×CD3 and (B)B7H3×CD3. [Figure 29] This study demonstrates that ENPP3×CD28 amplifies the induction of BCIXL in RTCCs mediated by (A) ENPP3×CD3 and (B) B7H3×CD3. [Figure 30] This study demonstrates that ENPP3×CD28 increases T cell cytotoxicity in CD3×ENPP3-mediated RTCC, as measured by granzyme B and CD107a. [Figure 31A] This study demonstrates that ENPP3 increases T cell activation and PD1 upregulation in both A) ENPP3×CD3-mediated RTCC and B) B7H3×CD3-mediated RTCC. [Figure 31B] This study demonstrates that ENPP3 increases T cell activation and PD1 upregulation in both A) ENPP3×CD3-mediated RTCC and B) B7H3×CD3-mediated RTCC. [Figure 32] This study demonstrates the upregulation of intracellular IFNg by ENPP3×CD28 in RTCC mediated by ENPP3×CD3. [Figure 33] This demonstrates the induction of IL2 secretion from purified T cells incubated with VMRC-RCW tumor cells (approximately 13,000 ENPP3 antigens) with a 1:1 E:T ratio, as well as with exemplary ENPP3×CD3 at 1 μg / ml and various forms of escalating doses of ENPP3×CD28 bsAb. [Figure 34] This demonstrates the induction of IL2 secretion from purified T cells incubated with VMRC-RCW tumor cells (approximately 13,000 ENPP3 antigens) with a 1:1 E:T ratio, as well as with exemplary ENPP3×CD3 at 1 μg / ml and escalating doses of ENPP3×CD28 bsAb (possessing various ENPP3 and CD28 binding domains, all in 1+1 Fab-scFv-Fc form). [Figure 35]XENP46666 demonstrates that in antigen-specific RTCC assays, XENP46666 enhances IL2 release after 24 hours by 2-3 times in T cells + ENPP3 + pp65-MDA-MB-231 target cells compared to target cells alone. [Figure 36] The antitumor activity of ENPP3×CD28 bsAb XENP46666 and XENP46667, when combined with ENPP3×CD3 bsAb, is demonstrated. [Figure 37] Pharmacokinetic data for XENP46666 and its surrogate XENP46674 in cynomolgus monkeys are shown. [Figure 38] This demonstrates enhanced cell killing via EpCAM×CD3 by ENPP3×CD28. [Figure 39] This shows enhancement of IFNg release via EpCAM×CD3 by ENPP3×CD28. [Figure 40A] This shows ENPP3×CD28 bsAb in the format scFv, with one arm at the center. [Figure 40B] This shows ENPP3×CD28 bsAb in the format scFv, with one arm at the center. [Figure 41A] This shows ENPP3×CD28 bsAb in 2+1 mAb-scFv format. [Figure 41B] This shows ENPP3×CD28 bsAb in 2+1 mAb-scFv format. [Figure 41C] This shows ENPP3×CD28 bsAb in 2+1 mAb-scFv format. [Figure 41D] This shows ENPP3×CD28 bsAb in 2+1 mAb-scFv format. [Figure 41E] This shows ENPP3×CD28 bsAb in 2+1 mAb-scFv format. [Figure 41F] This shows ENPP3×CD28 bsAb in 2+1 mAb-scFv format. [Figure 42A] This shows ENPP3×CD28 bsAb in 2+1 Fab2-scFv-Fc format. [Figure 42B] This shows ENPP3×CD28 bsAb in 2+1 Fab2-scFv-Fc format. [Figure 42C] This shows ENPP3×CD28 bsAb in 2+1 Fab2-scFv-Fc format. [Figure 42D] This shows ENPP3×CD28 bsAb in 2+1 Fab2-scFv-Fc format. [Figure 42E] This shows ENPP3×CD28 bsAb in 2+1 Fab2-scFv-Fc format. [Figure 42F] This shows ENPP3×CD28 bsAb in 2+1 Fab2-scFv-Fc format. [Figure 43] This shows increased thermal stability of disulfide-stabilized 1A7 scFv. [Figure 44] The CD28 affinity of additional 1A7 scFv variants, with or without disulfide stabilization (indicated as [SS]), is shown, obtained from different experiments in association with different CD28 multispecific antibodies. All scFvs shown below were performed using scFv linker (GKPGS) 4, but any scFv linker (including that in Figure 5) may also be used. Furthermore, the scFv may be in either the VH-scFv linker-VL or VL-scFv linker-VH direction. [Figure 45A] An exemplary 1A7-based scFv is shown, in which cysteine ​​was manipulated for disulfide stabilization. [Figure 45B] An exemplary 1A7-based scFv is shown, in which cysteine ​​was manipulated for disulfide stabilization. [Figure 45C] An exemplary 1A7-based scFv is shown, in which cysteine ​​was manipulated for disulfide stabilization. [Figure 45D] An exemplary 1A7-based scFv is shown, in which cysteine ​​was manipulated for disulfide stabilization. [Figure 45E]An exemplary 1A7-based scFv is shown, in which cysteine ​​is manipulated for disulfide stabilization. All scFvs shown below use the scFv linker (GKPGS) 4, but any scFv linker (including that shown in Figure 5) may be used. Furthermore, the scFv can be in either the VH-scFv linker-VL or VL-scFv linker-VH direction; however, only the former is illustrated. [Figure 46] This shows the steady-state (SS) and kinetic (1:1 Langmuir) fitting KD values ​​of the ENPP3-binding domain for human and cynomolgus monkey ENPP3. [Figure 47A] This document shows a novel 1A7 VH sequence manipulated to pair with the IGKV1-39 germline sequence (although it may also pair with any 1A7 VL sequence), as well as the consensus framework region (FR) and complementarity determination region (CDR). [Figure 47B] This document shows a novel 1A7 VH sequence manipulated to pair with the IGKV1-39 germline sequence (although it may also pair with any 1A7 VL sequence), as well as the consensus framework region (FR) and complementarity determination region (CDR). [Figure 47C] This document shows a novel 1A7 VH sequence manipulated to pair with the IGKV1-39 germline sequence (although it may also pair with any 1A7 VL sequence), as well as the consensus framework region (FR) and complementarity determination region (CDR). [Figure 48] This study demonstrates the affinity of a novel 1A7 variant paired with IGKV1-39 human germline VL to human CD28. [Modes for carrying out the invention]

[0048] I. Overview Ectonucleotide pyrophosphatase / phosphodiesterase family member 3 (ENPP3) is a type II transmembrane glycoprotein belonging to the nucleotide pyrophosphatase / phosphodiesterase family. These proteins are involved in the hydrolysis of extracellular nucleotides and also possess ATPase and ATP pyrophosphatase activity. ENPP3 is overexpressed in various cancers, including most renal cell carcinomas and some liver cancers. Because ENPP3 is overexpressed in a variety of tumors, it is a candidate for targeted therapy development.

[0049] T cell activation in cancer treatment is a widely studied topic. T cells require multiple signals for complete activation and differentiation. As shown in Figure 23A, signal 1, facilitated by the recognition of peptide-MHC (pMHC) complexes by the T cell receptor (TCR), is absolutely essential for T cell activation. Signal 2, which acts synergistically with and amplifies signal 1, is typically provided by the interaction of CD28 itself with CD80 and CD86, which are ligands for CD28. CD28 engagement alone is typically inactive, but when combined with signal 1 activation, it promotes additional activation, survival, and proliferation signals (including IL-2 secretion). Since CD80 and CD86 are spontaneously expressed only by professional antigen-presenting cells (APCs), the degree of CD28 costimulation in the tumor environment can vary considerably. Therefore, the present invention relates to a novel class of tumor-targeted anti-CD28 × anti-ENPP3 bispecific antibodies, wherein the CD80 / CD86 engagement of CD28 mimics the CD80 / CD86 engagement of CD28, thereby providing an artificial source of signal 2 (Figure 23B). Notably, signal 1 may be provided by the innate TCR:pMHC recognition of tumor cells, or by a combination of CD28 bispecificity and CD3 bispecificity that can mimic signal 1 (e.g., anti-CD3 × anti-ENPP3).

[0050] Accordingly, provided herein are novel anti-CD28 × anti-ENPP3 (also referred to as "αCD28 × αENPP3" and sometimes "CD28 × ENPP3") bispecific antibodies and methods for using such antibodies for the treatment of ENPP3-associated cancers. In many cases, these bispecific antibodies are heterodimers. The αCD28 × αENPP3 antibodies of interest can agonist bind to CD28 costimulatory molecules on T cells and target ENPP3 on ENPP3-expressing tumor cells. Thus, such antibodies selectively enhance antitumor activity at ENPP3-expressing tumor sites while minimizing peripheral toxicity. The antibodies of interest provided herein are particularly useful for enhancing antitumor activity, whether used alone as monotherapy or in combination with other anticancer therapies more fully described herein.

[0051] Accordingly, in one embodiment, what is provided herein is a heterodimer antibody that binds to two different antigens, for example, the antibody is "bispecific" in that it binds to two different target antigens, generally CD28 and ENPP3, as described below. These heterodimer antibodies can bind to each of the target antigens in either a monovalent (e.g., having a single antigen-binding domain, e.g., a pair of variable heavy and variable light domains) or a bivalent (having two antigen-binding domains, each independently binding to the antigen). In some embodiments, the heterodimer antibodies provided herein contain one CD28-binding domain and one ENPP3-binding domain (e.g., a heterodimer antibody of the "1+1 Fab-scFv-Fc" form described herein, and thus bispecific and bivalent). In other embodiments, the heterodimer antibodies provided herein contain one CD28-binding domain and two ENPP3-binding domains (e.g., heterodimer antibodies of the “2+1 Fab2-scFv-Fc”, “2+1 mAb-scFv”, and “2+1 Fab2-Fc×scFv-Fc” forms described herein, which are thus bispecific but trivalent because they contain three antigen-binding domains (ABD)). The heterodimer antibodies provided herein are based on the use of different monomers containing amino acid substitutions (i.e., “scuba riant”) that “bias” towards heterodimer formation rather than homodimer formation, as will be more thoroughly outlined below. In some embodiments, the heterodimer antibodies are also combined with a purification variant (e.g., “pI variant”) that allows for the simple purification of heterodimers from homodimers, as will also be outlined below. The heterodimer bispecific antibodies provided typically rely on the use of engineered or variant Fc domains that can be self-constructed in generative cells to generate heterodimer proteins, and on methods for generating and purifying such heterodimer proteins.

[0052] II. Nomenclature The nomenclature for specific antigen-binding domains (e.g., ENPP3 and CD28-binding domains) uses the format "Hx.xx_Ly.yy", where the numbers are unique identifiers for specific variable-chain sequences. For example, the CD28-binding domain "1A7[CD28]_H1_L1" (Figure 15) contains a variable heavy domain, H1, and a variable light domain, L1. When these sequences are used as scFv, the name "H1_L1" indicates that the binding domain contains a variable heavy domain "H1" combined with a variable light domain "L1", with an N-to-C-terminal VH-linker-VL orientation. A molecule having the same sequences of heavy and light variable domains but in the reverse order (N-to-C-terminal VL-linker-VH orientation) would be designated "L1_H1". Similarly, different constructs can "mix and match" heavy and light chains, as is evident from the sequence listings and figures.

[0053] III. Definition To ensure that this application can be fully understood, some definitions are provided below. Such definitions are intended to encompass grammatical equivalents.

[0054] In this specification, “CD28,” “differentiation antigen group 28,” and “Tp44” (e.g., Genebank accession numbers NP_001230006 (human), NP_001230007 (human), NP_006130 (human), and NP_031668 (mouse)) refer to B7 receptors expressed on T cells that provide costimulatory signals necessary for T cell activation and survival. T cell stimulation via CD28, in addition to the T cell receptor (TCR), provides a potent signal for the production of various interleukins. CD28 is a receptor for the CD80 (B7.1) and CD86 (B7.2) proteins. CD28 contains an intercellular domain with a YMNM motif important for the recruitment of SH2 domain-containing proteins, particularly PI3K. CD28 also contains two proline-rich motifs that can bind SH3-containing proteins. An exemplary CD28 sequence is shown in Figure 1. Unless otherwise stated, references to CD28 refer to human CD28 sequences.

[0055] In this specification, "ENPP3," or "Ectonucleotide pyrophosphatase / phosphodiesterase family member 3," or "B10," or "CD203c," or "NPP3," or "PD-IBETA," or "PDNP3" refers to type II transmembrane glycoproteins belonging to the nucleotide pyrophosphatase / phosphodiesterase family. These proteins are involved in the hydrolysis of extracellular nucleotides and also possess ATPase and ATP pyrophosphatase activity. The ENPP3 sequence is shown, for example, in Figure 2. ENPP3 is overexpressed in various cancers, including renal cancer.

[0056] In this specification, "reduction" means a decrease or removal of activity. Therefore, for example, "reduction of FcγR binding" means that the Fc region amino acid variant has less than 50% of the initial binding compared to an Fc region that does not contain the particular variant, with a loss of activity of more than 70-80-90-95-98% being preferred, and is usually below the binding level detectable in Biacore, SPR, or BLI assays. Those particularly used in FcγR binding reduction are shown in Figure 5, and these are usually attached to both monomers.

[0057] "ADCC," or antibody-dependent cell-mediated cytotoxicity, as used herein, refers to a cell-mediated reaction in which nonspecific cytotoxic cells expressing FcγR recognize a bound antibody on a target cell, subsequently causing lysis of the target cell. ADCC correlates with binding to FcγRIIIa; increased binding to FcγRIIIa results in increased ADCC activity.

[0058] When used herein, "ADCP" or "antibody-dependent cell-mediated phagocytosis" refers to a cell-mediated response in which nonspecific phagocytic cells expressing FcγR recognize bound antibodies on target cells, and subsequently cause phagocytosis of the target cells.

[0059] As used herein, the term “antibody” is used in a general sense. The antibodies provided herein may take many forms as described herein, including conventional antibodies and antibody derivatives, fragments, and mimetic compounds described herein.

[0060] Conventional immunoglobulin (Ig) antibodies are "Y"-shaped tetramers. Each tetramer typically consists of two identical polypeptide chain pairs, each pair containing one "light chain" monomer (typically with a molecular weight of about 25 kDa) and one "heavy chain" monomer (typically with a molecular weight of about 50–70 kDa).

[0061] Other useful antibody formats include, but are not limited to, the “1+1 Fab-scFv-Fc,” “2+1 Fab2-scFv-Fc,” “2+1 mAb-scFv,” and “2+1 Fab2-FcxscFv-Fc” formats provided herein (see, for example, Figure 14). Additional useful antibody formats include, but are not limited to, the “1+1 common light chain,” “2+1 common light chain,” “mAb-Fv,” “mAb-scFv,” “central-Fv,” “1-arm scFv-mAb,” “scFv-mAb,” “double scFv,” and “trident” format antibodies (Figure 14). See also US20180127501A1 (which is incorporated herein by reference, particularly the relevant parts concerning antibody formats) (see, for example, Figure 2 of US20180127501A1).

[0062] Antibody heavy chains typically include a variable weight (VH) domain containing vhCDR1-3 and an Fc domain containing CH2-CH3 monomers. In some embodiments, antibody heavy chains also include hinge and CH1 domains. Conventional antibody heavy chains are monomers organized from N to C-terminus: VH-CH1-hinge-CH2-CH3. The CH1-hinge-CH2-CH3 are collectively referred to as the "constant domain" or "constant region" of the heavy chain of antibodies in which five different categories or "isotypes" exist: IgA, IgD, IgG, IgE, and IgM.

[0063] In some embodiments, the antibodies provided herein include a constant IgG isotype domain, which has several subclasses, including but not limited to IgG1, IgG2, IgG3, and IgG4. In the IgG subclasses of immunoglobulins, several immunoglobulin domains are present in the heavy chain. As used herein, “immunoglobulin (Ig) domain” means a region of immunoglobulin having a distinctive tertiary structure. Heavy chain domains, including constant heavy (CH) domains and hinge domains, are of interest in the present invention. In relation to IgG antibodies, each IgG isotype has three CH domains. Thus, in relation to IgG, the “CH” domains are as follows: “CH1” points to positions 118-215 according to the EU index in Kabat; “Hinge” points to positions 216-230 according to the EU index in Kabat; “CH2” points to positions 231-340 according to the EU index in Kabat; and “CH3” points to positions 341-447 according to the EU index in Kabat. As shown in Figure 1, the exact numbering and arrangement of heavy chain domains may differ between different numbering systems. As shown herein and described below, pI variants may be in one or more CH regions and in the hinge regions discussed below.

[0064] It should be noted that IgG1 has different allotypes with polymorphisms at 356(D or E) and 358(L or M). The sequences shown herein use the 356E / 358M allotype, but other allotypes are included herein. That is, any sequence containing the IgG1 Fc domain included herein may have 356D / 358L, which replaces the 356E / 358M allotype. It should also be understood that therapeutic antibodies may also contain isotype and / or subclass hybrids. For example, as shown in U.S. Publication 2009 / 0163699 (incorporated by reference), the antibody in some embodiments contains a human IgG1 / G2 hybrid.

[0065] When used herein, “Fc,” “Fc region,” or “Fc domain” refers to a polypeptide containing the constant region of an antibody, in some examples all or part of the first constant region immunoglobulin domain (e.g., CH1), and in some cases optionally all or part of the hinge. In the case of IgG, the Fc domain includes the immunoglobulin domains CH2 and CH3 (Cγ2 and Cγ3), and optionally all or part of the hinge region between CH1 (Cγ1) and CH2 (Cγ2). Thus, in some cases, the Fc domain includes CH2-CH3 and hinge-CH2-CH3 from N to C-terminus. In some embodiments, the Fc domain is from IgG1, IgG2, IgG3, or IgG4, with IgG1 hinge-CH2-CH3 and IgG4 hinge-CH2-CH3 being particularly utilized in many embodiments. Also, in the case of the human IgG1 Fc domain, the hinge may include the C220S amino acid substitution. Furthermore, in the case of the human IgG4 Fc domain, the hinge may include the S228P amino acid substitution. While the boundaries of the Fc region can vary, the human IgG heavy chain Fc region is typically defined as containing residues E216, C226, or A231 at its carboxyl terminus, and the numbering follows the EU index in Kabat. In some embodiments, amino acid modifications are made to the Fc region to alter binding to one or more FcγR or FcRn, as more fully described below.

[0066] In this specification, “heavy chain constant region” means the CH1-hinge-CH2-CH3 portion of an antibody (or fragment thereof), excluding the variable heavy domain; in the EU numbering of human IgG1, this corresponds to amino acids 118-447. In this specification, “heavy chain constant region fragment” means a heavy chain constant region containing fewer amino acids from either or both of the N and C-terminuses, but still retaining the ability to form dimers with another heavy chain constant region.

[0067] Another type of heavy chain domain is the hinge region. In this specification, “hinge,” “hinge region,” “antibody hinge region,” or “hinge domain” refers to a flexible polypeptide containing amino acids between the first and second constant domains of an antibody. Structurally, the IgG CH1 domain terminates at EU position 215, and the IgG CH2 domain begins at residue EU position 231. Thus, in the case of IgG, the antibody hinge is defined herein as containing positions 216 (E216 in IgG1) to 230 (P230 in IgG1), with numbering following the EU index in Kabat. In some cases, “hinge fragments” containing fewer amino acids at either or both of the N and C-terminuses of the hinge domain are used. As described herein, pI variants can also be made from hinge regions. Many of the antibodies herein have at least one cysteine ​​at position 220, following the EU numbering (hinge region), replaced with serine. Typically, this modification is on the “scFv monomer” side (when the 1+1 or 2+1 form is used) for most of the sequences shown herein, but it can also be on the “Fab monomer” side, or both, to reduce disulfide formation. One or both of these replaced cysteines (C220S) are specifically included in the sequences shown herein.

[0068] As will be understood by those skilled in the art, the precise numbering and arrangement of heavy chain constant region domains (i.e., CH1, hinge, CH2, and CH3 domains) may differ between different numbering systems. A useful comparison of heavy constant region numbering according to EU and Kabat is provided below; please refer to Edelman et al., 1969, Proc Natl Acad Sci USA 63:78-85 and Kabat et al., 1991, Sequences of Proteins of Immunological Interest, 5th Ed., United States Public Health Service, National Institutes of Health, Bethesda (incorporated by reference throughout).

[0069] [Table 1]

[0070] Antibody light chains typically contain two domains: a variable light domain (VL) containing light chain CDR vlCDR1-3, and a constant light chain region (often referred to as CL or Cκ). Antibody light chains are typically sequenced VL-CL from the N-terminus to the C-terminus.

[0071] In this specification, “antigen-binding domain” or “ABD” means a set of six complementarity-determining regions (CDRs) that, when present as part of a polypeptide sequence, specifically bind to the target antigens discussed herein (e.g., ENPP3 or CD28). As is known in the art, these CDRs generally exist as a first set of variable weight CDRs (vhCDR or VHCDR) and a second set of variable light CDRs (vlCDR or VLCDR), each containing three CDRs: vhCDR1, vhCDR2, vhCDR3 variable weight CDRs and vlCDR1, vlCDR2, vlCDR3, and vhCDR3 variable light CDRs. The CDRs reside in variable weight domains (vhCDR1-3) and variable light domains (vlCDR1-3). The variable weight domains and variable light domains form the Fv region.

[0072] The present invention provides a number of different CDR sets. In this case, a “complete CDR set” includes three variable light CDRs and three variable heavy CDRs, for example, vlCDR1, vlCDR2, vlCDR3, vhCDR1, vhCDR2, and vhCDR3. These may each be part of a larger variable light or variable heavy domain. Furthermore, as will be more fully outlined herein, the variable heavy and variable light domains may be on separate polypeptide chains when heavy and light chains are used (e.g., when Fab is used), or on a single polypeptide chain in the case of scFv sequences.

[0073] As will be understood by those skilled in the art, the exact numbering and arrangement of CDRs may differ between different numbering systems. However, it should be understood that the disclosure of variable weight and / or variable light sequences includes the disclosure of the associated (intrinsic) CDRs. Thus, the disclosure of each variable weight region is a disclosure of vhCDRs (e.g., vhCDR1, vhCDR2, and vhCDR3), and the disclosure of each variable light region is a disclosure of vlCDRs (e.g., vlCDR1, vlCDR2, and vlCDR3). A useful comparison of CDR numbering is as follows: See Lafranc et al., Dev.Comp.Immunol.27(1):55-77(2003):

[0074] [Table 2]

[0075] Throughout this specification, the Kabat numbering system is generally used when referring to residues in the variable domain (approximately residues 1-107 in the light chain variable region and residues 1-113 in the heavy chain variable region), while the EU numbering system is used for the Fc region (e.g., Kabat et al. (cited above) (1991)).

[0076] CDRs contribute to the antigen-binding domain and antigen binding of antibodies, or more specifically, the formation of epitope-binding sites. An "epitope" refers to a determinant that interacts with a specific antigen-binding site in the variable region of an antibody molecule, also known as a paratope. Epitopes are groups of molecules, such as amino acids or sugar side chains, and typically possess specific structural and electrical characteristics. A single antigen may have multiple epitopes.

[0077] An epitope may include amino acid residues directly involved in binding (also known as the immunodominant component of the epitope) and other amino acid residues not directly involved in binding, such as amino acid residues that are effectively blocked by a specific antigen-binding peptide (i.e., those amino acid residues are within the footprint of the specific antigen-binding peptide).

[0078] Epitopes can be either conformational or linear. Conformational epitopes are generated by spatially juxtaposed amino acids from different segments of a linear polypeptide chain. Linear epitopes are generated by adjacent amino acid residues in a polypeptide chain. Conformational and non-conformational epitopes can be distinguished in that binding to the former, rather than the latter, is lost in the presence of a denaturing solvent.

[0079] An epitope typically contains at least three, more commonly five or eight to ten, amino acids in its own spatial conformation. Antibodies that recognize the same epitope can be identified in a simple immunoassay, e.g., "binning," in which one antibody demonstrates its ability to block the binding of another antibody to a target antigen. As outlined below, the present invention includes not only the antigen-binding domains and antibodies listed herein, but also those that compete for binding to the epitopes bound by the listed antigen-binding domains.

[0080] In some embodiments, the six CDRs of the antigen-binding domain are provided by variable heavy and variable light domains. In the "Fab" form, the set of six CDRs consists of two distinct polypeptide sequences: a variable heavy domain (vh or VH; containing vhCDR1, vhCDR2, and vhCDR3) and a variable light domain (vl or VL; containing vlCDR1, vlCDR2, and vlCDR3), where the C-terminus of the vh domain is bound to the N-terminus of the CH1 domain of the heavy chain, and the C-terminus of the vl domain is bound to the N-terminus of the constant light domain (thus forming the light chain). In the scFv form, the vh and vl domains are covalently linked to a single polypeptide sequence, typically via the use of a linker outlined herein ("scFv linker"), which can be either vh-linker-vl or vl-linker-vh (starting from the N-terminus), with the former being the preferred configuration (each side may include an arbitrary domain linker, depending on the form used). Typically, the C-terminus of the scFv domain is linked to the N-terminus of all or part of the hinge in the second monomer.

[0081] When used herein, “variable region” or “variable domain” means a region of immunoglobulin that includes one or more Ig domains substantially encoded by any of the Vκ, Vλ, and / or VH genes constituting kappa, lambda, and each heavy chain immunoglobulin locus, and contains a CDR that confers antigen specificity. Thus, a “variable heavy domain” pairs with a “variable light domain” to form an antigen-binding domain (“ABD”). Each variable domain also includes three hypervariable regions (“complementarity-determining regions,” “CDRs”) (vhCDR1, vhCDR2, and vhCDR3 in the case of a variable heavy domain and vlCDR1, vlCDR2, and vlCDR3 in the case of a variable light domain) and four framework (FR) regions, arranged from the amino terminus to the carboxyl terminus in the following order: FR1-CDR1-FR2-CDR2-FR3-CDR3-FR4.

[0082] "Fab" or "Fab region," as used herein, typically refers to an antibody region on two different polypeptide chains containing VH, CH1, VL, and CL immunoglobulin domains (e.g., VH-CH1 on one chain and VL-CL on the other). Fab may refer to this isolated region or to this region in relation to the bispecific antibody of the present invention. In relation to Fab, Fab includes the Fv region in addition to the CH1 and CL domains.

[0083] When used herein, “Fv,” “Fv fragment,” or “Fv region” means an antibody region containing the VL and VH domains. The Fv region can be formalized as both Fab (two distinct polypeptides, including the constant region typically outlined above, as discussed above) and single-chain Fv (scFv) (where the vl and vh domains are contained in a single peptide conjugated to a linker, typically as discussed herein).

[0084] In this specification, “single-stranded Fv” or “scFv” means a variable weight domain covalently linked to a variable light domain using an scFv linker, typically as discussed herein, to form an scFv or scFv domain. The scFv domain can be oriented either N-to-C-terminus (vh-linker-vl or vl-linker-vh). In sequences shown in the sequence listings and figures, the order of the vh and vl domains is included in the name. For example, H.X_L.Y means that the N-to-C-terminus is vh-linker-vl, and L.Y_H.X means vl-linker-vh.

[0085] Some embodiments of the target antibodies provided herein include at least one scFv domain, which typically contains a variable heavy domain and a variable light domain that are not naturally present but are linked together by an scFv linker. As outlined herein, scFv domains are typically oriented from N to C-terminus as VH-scFv linker-VL, whereas this can be reversed to VL-scFv linker-VH for any scFv domain (or one constructed using vh and vl sequences from Fab) using any linker at one or both ends, depending on the form.

[0086] In this specification, “modification” or “variant” means an amino acid substitution, insertion, and / or deletion in a polypeptide sequence or an alteration to a site chemically linked to a protein. For example, a modification may be an altered carbohydrate or PEG structure linked to a protein. In this specification, “amino acid modification” means an amino acid substitution, insertion, and / or deletion in a polypeptide sequence. For clarity, unless otherwise stated, amino acid modification always refers to amino acids encoded by DNA, e.g., the 20 amino acids that have codons in DNA and RNA.

[0087] In this specification, “amino acid substitution” or “substitution” means the replacement of an amino acid at a specific position in the parent polypeptide sequence with a different amino acid. In particular, in some embodiments, a substitution is a substitution of an amino acid at a specific position that is not found in nature and does not exist naturally in any organism, including within a living organism. For example, substitution E272Y refers to a variant polypeptide, in this case the Fc variant, in which glutamic acid at position 272 is replaced with tyrosine. For clarity, a protein that is manipulated to alter the nucleic acid coding sequence but not alter the starting amino acid (e.g., replacing CGG (encoding arginine) with CGA (still encoding arginine) to increase the host organism’s expression level) is not an “amino acid substitution”; that is, even though it is the generation of a new gene encoding the same protein, if the protein has the same amino acid at the specific position where it starts, it is not an amino acid substitution.

[0088] As used herein, "amino acid insertion" or "insertion" means the addition of an amino acid sequence at a specific position in the parent polypeptide sequence. For example, -233E or 233E specifies the insertion of glutamic acid after position 233 and before position 234. Also, -233ADE or A233ADE specifies the insertion of AlaAspGlu after position 233 and before position 234.

[0089] "Amino acid deletion" or "deletion," as used herein, means the removal of an amino acid sequence at a specific position in the parent polypeptide sequence. For example, E233- or E233#, E233() or E233del specify the deletion of glutamic acid at position 233. Also, EDA233- or EDA233# specifies the deletion of the sequence GluAspAla that begins at position 233.

[0090] "Variant protein," "protein variant," or "variant" as used herein means a protein that differs from that of the parent protein by at least one amino acid modification. A protein variant has at least one amino acid modification compared to the parent protein, but not so many that the variant protein cannot be aligned with the parent protein using an alignment program such as those described below. Typically, a variant protein (e.g., a variant Fc domain outlined herein) is at least 75, 80, 85, 90, 91, 92, 93, 94, 95, 96, 97, 98, or 99% identical to the parent protein using an alignment program such as BLAST, as described below.

[0091] When used herein, "variant" also refers to a specific amino acid modification that confers a particular function (e.g., "heterodimerated variant," "pI variant," "decay variant," etc.).

[0092] As described below, in some embodiments, the parent polypeptide, e.g., the Fc parent polypeptide, is a multiple constant domain or Fc region from a human wild-type sequence, e.g., IgG1, IgG2, IgG3, or IgG4. However, a human sequence having a variant may also function as a “parent polypeptide,” e.g., the IgG1 / 2 hybrid described in U.S. Publication 2006 / 0134105. The protein variant sequences herein preferably have at least about 80% identity with the parent protein sequence, most preferably at least about 90% identity, and more preferably at least about 95-98-99% identity. Therefore, as used herein, “antibody variant” or “variant antibody” means an antibody that differs from the parent antibody by at least one amino acid modification; “IgG variant” or “variant IgG” means an antibody that differs from the parent IgG (which, in this case, is also often derived from a human IgG sequence) by at least one amino acid modification; and “immunoglobulin variant” or “variant immunoglobulin” means an immunoglobulin sequence that differs from the parent immunoglobulin sequence by at least one amino acid modification. “Fc variant” or “variant Fc” means a protein that contains amino acid modifications in its Fc domain compared to the Fc domain of human IgG1, IgG2, or IgG4.

[0093] "Fc variant" or "variant Fc," as used herein, means a protein containing amino acid modifications in the Fc domain. Modifications may be additions, deletions, or substitutions. Fc variants are defined according to the amino acid modifications that constitute them. For example, N434S or 434S is an Fc variant having a serine substitution at position 434 compared to the parent Fc polypeptide, and the numbering follows the EU index. Similarly, M428L / N434S is an Fc variant having the substitutions M428L and N434S compared to the parent Fc polypeptide. WT amino acid identity may not be specified, in which case the aforementioned variant is referred to as 428L / 434S. It should be noted that the order in which substitutions are provided is arbitrary (i.e., for example, 428L / 434S is the same Fc variant as 434S / 428L). For all positions discussed herein relating to antibodies or their derivatives and fragments (e.g., Fc domains), unless otherwise noted, amino acid numbering follows the EU index. The “EU index,” “EU index in Kabat,” or “EU numbering” scheme refers to the numbering of EU antibodies (Edelman et al., 1969, Proc Natl Acad Sci USA 63:78-85 (incorporated herein by reference)). Modifications may be additions, deletions, or substitutions.

[0094] Typically, a variant Fc domain has at least approximately 80, 85, 90, 95, 97, 98, or 99 percent identity with the corresponding parental human IgG Fc domain (using the identity algorithms discussed below (one embodiment utilizes the BLAST algorithm as known in the art using default parameters)). Alternatively, a variant Fc domain may have 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 11, 12, 13, 14, 15, 16, 17, 18, 19, or 20 or more amino acid modifications compared to the parental Fc domain. Alternatively, the variant Fc domain may have up to 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 11, 12, 13, 14, 15, 16, 17, 18, 19, or 20 amino acid modifications compared to the parent Fc domain. Furthermore, as discussed herein, the variant Fc domains described herein still retain the ability to form dimers with other Fc domains when measured using known techniques described herein, such as non-denaturing gel electrophoresis.

[0095] As used herein, "protein" means at least two covalently linked amino acids, and includes proteins, polypeptides, oligopeptides, and peptides. Furthermore, polypeptides constituting the antibodies of the present invention may include synthetic derivatization, glycosylation, PEGylation, cyclic substitution, cyclization, linking to other molecules, fusion to proteins or protein domains, and addition of peptide tags or labels to one or more side chains or terminals.

[0096] As used herein, "residue" means the position in a protein and its associated amino acid identity. For example, asparagine 297 (also referred to as Asn297 or N297) is the residue at position 297 in the human antibody IgG1.

[0097] As used herein, "IgG subclass modification" or "isotype modification" refers to an amino acid modification that converts a certain amino acid in one IgG isotype to a corresponding amino acid in a differently aligned IgG isotype. For example, since IgG1 contains tyrosine at EU position 296 and IgG2 contains phenylalanine, the F296Y substitution in IgG2 is considered an IgG subclass modification.

[0098] As used herein, “modifications not found in nature” means non-isomorphic amino acid modifications. For example, since none of the human IgGs contain serine at position 434, the substitution 434S in IgG1, IgG2, IgG3, or IgG4 (or their hybrids) is considered a modification not found in nature.

[0099] As used herein, "amino acid" and "amino acid identity" mean one of the 20 naturally occurring amino acids encoded by DNA and RNA.

[0100] As used herein, "effector function" means a biochemical event resulting from the interaction between an antibody Fc region and an Fc receptor or ligand. Effector functions include, but are not limited to, ADCC, ADCP, and CDC.

[0101] As used herein, "IgG Fc ligand" means a molecule, preferably a polypeptide, from any organism that binds to the Fc region of an IgG antibody to form an Fc / Fc ligand complex. Fc ligands include, but are not limited to, FcγRI, FcγRII, FcγRIII, FcRn, C1q, C3, mannan-binding lectins, mannose receptors, staphylococcal protein A, streptococcal protein G, and viral FcγR. Fc ligands also include Fc receptor homologs (FcRH), a family of Fc receptors homologous to FcγR (Davis et al., 2002, Immunological Reviews 190:123-136 (integrated as a whole by reference)). Fc ligands may include undiscovered molecules that bind Fc. Specific IgG Fc ligands are FcRn and Fc gamma receptors. As used herein, "Fc ligand" means a molecule, preferably a polypeptide, from any organism that binds to the Fc region of an antibody in order to form an Fc / Fc ligand complex.

[0102] When used herein, “Fc gamma receptor,” “FcγR,” or “Fc gamma R” means any member of the family of proteins that ligate to the Fc region of an IgG antibody and are encoded by the FcγR gene. In humans, this family includes, but is not limited to, FcγRI(CD64) including isoforms FcγRIa, FcγRIb, and FcγRIc; FcγRII(CD32) including isoform FcγRIIa (including allotypes H131 and R131), FcγRIIb (including FcγRIIb-1 and FcγRIIb-2), and FcγRIIc; and FcγRIII(CD16) including isoforms FcγRIIIa (including allotypes V158 and F158) and FcγRIIIb (including allotypes FcγRIIb-NA1 and FcγRIIb-NA2) (Jefferis et al., 2002, Immunol Lett 82:57-65 (incorporated by reference as a whole)), as well as any undiscovered human FcγR or FcγR isoform or allotype. FcγR can originate from any organism, including but not limited to humans, mice, rats, rabbits, and monkeys. Mouse FcγR includes, but is not limited to, FcγRI(CD64), FcγRII(CD32), FcγRIII(CD16), and FcγRIII-2(CD16-2), as well as any undiscovered mouse FcγR or FcγR isoform or allotype.

[0103] When used herein, “FcRn” or “fetal Fc receptor” means a protein that conjugates the Fc region of an IgG antibody and is at least partially encoded by the FcRn gene. FcRn may be from any organism, including but not limited to humans, mice, rats, rabbits, and monkeys. As is known in the art, a functional FcRn protein comprises two polypeptides, often referred to as a heavy chain and a light chain. The light chain is beta-2-microglobulin, and the heavy chain is encoded by the FcRn gene. Unless otherwise described herein, FcRn or FcRn protein refers to the complex of the FcRn heavy chain and beta-2-microglobulin. Various FcRn variants are used to increase binding to the FcRn receptor and, in some cases, to increase the serum half-life. “FcRn variant” is an amino acid modification that contributes to increased binding to the FcRn receptor, and preferred FcRn variants are shown below.

[0104] Where used herein, “parent polypeptide” means the initiating polypeptide that is later modified to produce a variant. The parent polypeptide may be a naturally occurring polypeptide, or a variant or engineered version of a naturally occurring polypeptide. Thus, where used herein, “parent immunoglobulin” means an unmodified immunoglobulin polypeptide that is modified to produce a variant, and “parent antibody” means an unmodified antibody that is modified to produce a variant antibody. It should be noted that “parent antibody” includes known commercially available recombinantly produced antibodies, as outlined below. In this context, “parent Fc domain” is relative to the described variant; thus, “variant human IgG1 Fc domain” is compared to the parent Fc domain of human IgG1, “variant human IgG4 Fc domain” is compared to the parent Fc domain of human IgG4, and so on.

[0105] As used herein, "position" means a location in the sequence of a protein. Positions may be numbered sequentially or in an established format, for example, according to the EU index for numbering antibody domains (e.g., CH1, CH2, CH3, or hinge domains).

[0106] As used herein, "target antigen" means a molecule that is specifically bound by an antigen-binding domain, which includes a variable region of a given antibody.

[0107] In the context of the monomers of the heterodimer antibodies of the present invention as used herein, “strandiness” means maintaining the ability to “pair” to form a heterodimer by incorporating the heterodimerizing variant into each monomer, similar to how two strands of DNA “pair.” For example, when manipulating monomer A with a certain pI variant (e.g., to increase the pI), a stereovariant that is a “charge pair” and can be similarly utilized does not interfere with the pI variant, and for example, a charge variant that increases the pI will be placed on the same “strand” or “monomer” to maintain the functionality of both. Similarly, with regard to “skew” variants provided in pairs of sets, as will be more thoroughly outlined below, those skilled in the art will consider the pI when determining which strand or monomer one set of the pair will be placed on, so that pI separation is also maximized using the pI of the skew.

[0108] As used herein, "target cell" means a cell that expresses a target antigen.

[0109] In relation to generating the bispecific antibody according to the present invention, "host cell" means a cell that contains exogenous nucleic acids encoding the components of the bispecific antibody and is capable of expressing the bispecific antibody under suitable conditions. Suitable host cells are discussed below.

[0110] In this specification, “wild-type” or “WT” means a naturally occurring amino acid or nucleotide sequence, including allele variations. WT proteins have an amino acid or nucleotide sequence that is not intentionally modified.

[0111] This specification provides numerous antibody domains (e.g., Fc domains) that have sequence identity with human antibody domains. Sequence identity between two similar sequences (e.g., antibody variable domains) can be determined using the following methods: Smith, TF & Waterman, MS (1981) “Comparison Of Biosequences,” Adv.Appl.Math.2:482 [Local homology algorithm]; Needleman, SB & Wunsch, CD (1970) “A General Method Applicable To The Search For Similarities In The Amino Acid Sequence Of Two Proteins,” J.Mol.Biol.48:443 [Homologous alignment algorithm]; Pearson, WR & Lipman, DJ (1988) “Improved Tools For Biological Sequence Comparison,” Proc.Natl.Acad.Sci.(USA)85:2444 [Similarity search method]; or Altschul, SF et al, (1990) “Basic Local Alignment Search Tool,” Sequence identity can be measured by algorithms such as the "BLAST" algorithm (see https: / / blast.ncbi.nlm.nih.gov / Blast.cgi), as described in J.Mol.Biol.215:403-10. When using any of the aforementioned algorithms, default parameters (regarding window length, gap penalty, etc.) are used. In one embodiment, sequence identity is performed using the BLAST algorithm with default parameters.

[0112] The antibodies of the present invention are typically isolated or recombinant. "Isolated", when used to describe the various polypeptides disclosed herein, means a polypeptide that has been identified, separated and / or recovered from the cells or cell culture in which it was expressed. Usually, an isolated polypeptide will be prepared by at least one purification step. An "isolated antibody" refers to an antibody that substantially lacks other antibodies having different antigen specificities. "Recombinant" means that the antibodies are produced using recombinant nucleic acid techniques in an exogenous host cell and that they can also be isolated.

[0113] "Specific binding" to or "specifically binds to" or "specific for" a particular antigen or epitope means a binding that is clearly distinct from non-specific interactions. Specific binding can be measured, for example, by determining the binding of a molecule compared to the binding of a control molecule, which is typically a molecule of a similar structure that does not have binding activity. For example, specific binding can be determined by competition with a control molecule similar to the target.

[0114] Specific binding to a particular antigen or epitope is at least about 10 -4 M, at least about 10 -5 M, at least about 10 -6 M, at least about 10 -7 M, at least about 10 -8 M, at least about 10 -9 M, alternatively at least about 10 -10 M, at least about 10 -11 M, at least about  10 -12 M, or can be shown by an antibody having a KD for an antigen or epitope greater than that, where KD refers to the dissociation rate of a particular antibody-antigen interaction. Typically, an antibody that specifically binds an antigen will have a KD that is 20, 50, 100, 500, 1000, 5,000, 10,000 times higher or more than that of a control molecule for the antigen or epitope.

[0115] Furthermore, specific binding to a particular antigen or epitope can be demonstrated by antibodies with a KA or Ka constant for the antigen or epitope that is at least 20, 50, 100, 500, 1,000, 5,000, or more than 10,000 times compared to the control, where KA or Ka refers to the binding constant of a particular antibody-antigen interaction. Binding affinity is typically measured using Biacore, SPR, or BLI assays.

[0116] IV. Anti-CD28 x anti-ENPP3 antibody In one embodiment, a novel anti-CD28 × anti-ENPP3 antibody is provided herein. In some embodiments, the anti-CD28 × anti-ENPP3 antibody described herein can agonist-conjugate to CD28 costimulatory molecules on T cells and to ENPP3 on tumor cells. Such antibodies selectively enhance antitumor activity at ENPP3-related tumor sites while minimizing peripheral toxicity. The target antibodies provided herein are particularly useful in combination with other anticancer therapies, for example, including bispecific antibodies for the treatment of ENPP3-related cancers.

[0117] The anti-CD28 × anti-ENPP3 antibody is polyvalent and comprises at least two antigen-binding domains (ABDs), where at least one antigen-binding domain is a CD28-binding domain and at least one antigen-binding domain is an ENPP3-binding domain. The anti-CD28 × anti-ENPP3 antibody of interest may contain any suitable CD28-binding domains and ENPP3-binding domains, including, for example, the CD28-binding domains and ENPP3-binding domains provided herein.

[0118] The antigen-binding domains provided herein generally include variable heavy domains (VH) having VH-CDR1, VH-CDR-2, and VH-CDR-3, and variable light domains (VL) having VL-CDR1, VL-CDR-2, and VL-CDR-3.

[0119] Furthermore, as discussed above, the numbering used to identify CDRs in sequence listings and figures is Kabat, but as shown in Table 2, different numbering can be used, which changes the amino acid sequence of the CDR.

[0120] Further variants can be created for all of the variable weight domains and variable light domains listed herein. As outlined herein, in some embodiments, a set of six CDRs may have 0, 1, 2, 3, 4, or 5 amino acid modifications (including amino acid substitutions used for specific applications), in addition to variations in the framework regions of the variable weight domains and variable light domains, provided that the framework (excluding the CDRs) retains at least approximately 80, 85, 90, 95, or 99% identity with a human germline sequence selected from those listed in Figure 1 of U.S. Patent No. 7,657,380 (the figure and legend are incorporated in their entirety by reference). Thus, for example, identical CDRs described herein can be combined with different framework sequences of human germline sequences, provided that the framework regions retain at least 80, 85, or 90% identity with a human germline sequence selected from those listed in Figure 1 of U.S. Patent No. 7,657,380. Alternatively, a CDR may have amino acid modifications (for example, one, two, three, four, or five amino acid modifications in a set of CDRs (i.e., a CDR may be modified as long as the total number of changes in a set of six CDRs is less than six amino acid modifications, and any combination of CDRs to be modified is arbitrary, for example, one change in vlCDR1, two changes in vhCDR2, and zero changes in vhCDR3)) as well as changes in the framework region, provided that the framework region maintains at least 80, 85, or 95-99% identity with respect to a human germline sequence selected from those enumerated in Figure 1 of U.S. Patent No. 7,657,380).

[0121] As will be understood by those skilled in the art, any set of six CDR or VH and VL domains may be in scFv or Fab form, which is then appended to the heavy and light chain constant domains, the heavy chain constant domain containing variants (including within the Fc domain in addition to within the CH1 domain).

[0122] In addition, in embodiments in which the target antibody contains scFv, the scFv may be oriented from the N-terminus to the C-terminus of a VH-scFv linker-VL or a VL-scFv linker-VH. In some forms, one or more ABDs are generally Fabs containing a VH domain on one protein chain (generally as a component of the heavy chain) and a VL domain on another protein chain (generally as a component of the light chain). Exemplary scFv linkers used in the target antibody are shown in Figure 6.

[0123] In some embodiments, the anti-CD28 × anti-ENPP3 antibody is a bispecific antibody. In some embodiments, the anti-CD28 × anti-ENPP3 antibody is a bivalent antibody. In some embodiments, the anti-CD28 × anti-ENPP3 antibody is a trivalent antibody. In some embodiments, the anti-CD28 × anti-ENPP3 antibody is a bispecific bivalent antibody. In some embodiments, the anti-CD28 × anti-ENPP3 antibody contains one CD28-binding domain and one ENPP3-binding domain. In exemplary embodiments, the anti-CD28 × anti-ENPP3 antibody is a bispecific trivalent antibody. In some embodiments, the anti-CD28 × anti-ENPP3 antibody contains one CD28-binding domain and two ENPP3-binding domains.

[0124] The anti-CD28 × anti-ENPP3 antibodies provided herein may be any useful form, including, for example, standard immunoglobulins, as well as the “1+1 Fab-scFv-Fc,” “2+1 mAb-scFv,” “2+1 Fab2-scFv-Fc,” and “2+1 Fab2-Fc × scFv-Fc” forms described herein (Figure 14). Additional useful forms include, but are not limited to, the “mAb-Fv,” “central-Fv,” “1-arm scFv-mAb,” “scFv-mAb,” “double scFv,” and “trident” forms provided herein (e.g., Figure 14). See also US20180127501A1 (which is incorporated herein by reference, particularly the relevant parts concerning antibody forms) (e.g., see Figure 2). In some embodiments, the anti-CD28 × anti-ENPP3 antibody is a heterodimer bispecific antibody containing a variant Fc domain having one of the heterodimerized scuba riant, pI variant, and / or attenuation variant described herein. See, for example, Figure 8.

[0125] Unless otherwise specified herein, the order of names in the antigen list does not confer structure. That is, in the anti-ENPP3 × anti-CD28 1+1 Fab-scFv-Fc antibody, scFv may bind to either ENPP3 or CD28. However, in some cases, the order indicates the structure.

[0126] The anti-CD28 × anti-ENPP3 antibodies provided herein further comprise different antibody domains. As described herein and known in the art, the antibodies described herein comprise different domains within the heavy and light chains, which may overlap. These domains include, but are not limited to, Fc domains, CH1 domains, CH2 domains, CH3 domains, hinge domains, heavy constant domains (CH1-hinge-Fc domain or CH1-hinge-CH2-CH3), variable heavy domains, variable light domains, light constant domains, Fab domains, and scFv domains.

[0127] As shown herein, there are numerous suitable linkers that can be used to covalently link described domains (e.g., scFv, Fab, Fc domains, VH domains, VL domains, etc.) containing conventional peptide bonds, generated by recombinant technology (for use as either domain linkers or scFv linkers). An exemplary linker for linking domains of a target antibody is shown in Figure 7. In some embodiments, the linker peptide may primarily contain the following amino acid residues: Gly, Ser, Ala, or Thr. The linker peptide should have a length appropriate for linking two molecules in such a way that they assume the correct conformation to each other so that they retain the desired activity. In one embodiment, the linker is about 1 to 50 amino acids long, preferably about 1 to 30 amino acids long. In one embodiment, a linker of 1 to 20 amino acids long may be used, and about 5 to about 10 amino acids are utilized in some embodiments. Useful linkers include, for example, glycine-serine polymers, glycine-alanine polymers, alanine-serine polymers, and other flexible linkers, including (GS)n, (GSGGS)n, (GGGGS)n, and (GGGS)n (wherein n is an integer of at least 1 (and generally 3 to 4)). Alternatively, a variety of non-proteinaceous polymers, including but not limited to polyethylene glycol (PEG), polypropylene glycol, polyoxyalkylene, or copolymers of polyethylene glycol and polypropylene glycol, may be used as linkers.

[0128] Other linker sequences may include any sequence of any length from the CL / CH1 domain, but not all residues of the CL / CH1 domain, for example, the first 5-12 amino acid residues of the CL / CH1 domain. Linkers may originate from immunoglobulin light chains, e.g., Cκ or Cλ. Linkers may originate from immunoglobulin heavy chains of any isotype, including, for example, Cγ1, Cγ2, Cγ3, Cγ4, Cα1, Cα2, Cδ, Cε, and Cμ. Linker sequences may also originate from other proteins, e.g., Ig-like proteins (e.g., TCR, FcR, KIR), hinge region-derived sequences, and other native sequences from other proteins.

[0129] In some embodiments, the linker is a “domain linker” used to ligate any two domains outlined herein. For example, in the 2+1 Fab2-scFv-Fc form, there may be a domain linker that ligates the C-terminus of the CH1 domain of Fab to the N-terminus of scFv, and another arbitrary domain linker that ligates the C-terminus of scFv to the CH2 domain (although in many embodiments, a hinge is used as this domain linker). Any suitable linker may be used, but many embodiments utilize an arbitrary peptide sequence that enables recombination of two domains having sufficient length and flexibility to allow each domain to retain its biological function, in addition to a glycine-serine polymer as a domain linker, including, for example, (GS)n, (GSGGS)n, (GGGGS)n, and (GGGS)n (wherein n is an integer of at least 1 (and generally 3-4-5)). In some cases, charged domain linkers, as used in some embodiments of the scFv linker, may be used, with regard to the “strandiness” outlined below. An example of a useful domain linker is shown in Figure 7.

[0130] In some embodiments, the linker is an scFv linker used to covalently link the VH and VL domains, as discussed herein. Often, the scFv linker is a charged scFv linker, some of which are shown in Figure 6. Provided herein, so as therein, are charged scFv linkers for facilitating separation in pI between a first monomer and a second monomer. That is, by incorporating either a positive or negative charged scFv linker (or both; in the case of a skeleton using scFv on different monomers), monomers containing a charged linker can alter pI without causing further alteration to the Fc domain. These charged linkers can be substituted within any scFv containing a standard linker. In this case as well, as will be understood by those skilled in the art, the charged scFv linker is used on the correct “chain” or monomer according to the desired alteration of pI. For example, to produce a 1+1 Fab-scFv-Fc heterodimer antibody as discussed herein, the original pI of the Fv region for each of the desired antigen-binding domains is calculated, one is selected to produce the scFv, and a positive or negative linker is selected depending on the pI. Charged domain linkers can also be used to increase the pI separation of monomers according to the present invention, and thus those shown in Figure 7 can be used in any embodiment of this specification in which a linker is utilized.

[0131] An exemplary anti-CD28×anti-ENPP3 antibody is shown, for example, in Figure 24. During cell culture generation of the anti-CD28×anti-ENPP3 antibodies provided herein, a C-terminal lysine residue or a C-terminal lysine and glycine residue may be cleaved from the heavy chain monomer, thereby resulting in a C-terminus "cut-off" variant. See, for example, Jiang et al., Journal of Pharmaceutical Sciences 105:2066-2072 (2016). Thus, in some embodiments provided herein, the anti-CD28×anti-ENPP3 antibody is a variant of one of the anti-CD28×anti-ENPP3 antibodies and includes a deletion of a C-terminal lysine (-K) terminus or a C-terminal lysine and glycine (-GK) residue in one or both Fc domains of the anti-CD28×anti-ENPP3 antibodies described herein. In some embodiments, the deletion is G446del and / or K447del (EU numbering).

[0132] The embodiments of the anti-CD28 × anti-ENPP3 antibody are described in further detail below.

[0133] A. CD28 binding domain The anti-CD28 × anti-ENPP3 antibodies provided herein include at least one CD28-binding domain. The anti-CD28 × anti-ENPP3 antibodies provided herein may include any preferred CD28-binding domain. In exemplary embodiments, the CD28-binding domain is an agonist CD28 ABD that advantageously provides T-cell costimulatory activity.

[0134] As will be understood by those skilled in the art, a suitable CD28-binding domain may comprise a set of six CDRs, either as the underlined CDRs shown in the figure, or as CDRs identified using other alignments within the variable weight (VH) domain and variable light domain (VL) sequences of the CD28-binding domain shown in Figures 15, 18, and 21, when different numbering schemes are used as described herein and shown in Table 2. Additional VH and VL sequences of exemplary CD28-binding domains that can be used in the target antibody are shown in Figures 16 and 17. A suitable CD28 ABD may also include these sequences and the entire VH and VL sequences shown in the figure, used as scFv or Fab.

[0135] In one embodiment, the CD28 antigen-binding domain includes, but is not limited to, those shown in Figures 15, 18, and 21, any six CDRs of any of the CD28-binding domains described herein (i.e., vhCDR1-3 and vlCDR1-3). In some embodiments, the CD28 ABD that binds to human CD28 is the following CD28 One of the ABDs: 1A7[CD28]_H1L1, 1A7[CD28]_H1.1_L1, 1A7[CD28]_H1_L1.71, 1A7[CD28]_H1.1_L1.71, 1A7[CD28]_H1.14_L1, 1A7[CD28]_H1.14_L1.71, CD28.3[CD28]_H0L0, hCD28.3[CD28]_H1L1, 5.11A1[CD28]_H0L0, TGN1412_H1L1, 341VL 34[CD28]_H1L1, 341VL36[CD28]_H1L1, 281VL4[CD28]_H1L1, HuTN228[CD28]_H1L1, PV1[CD28]_H0L0, m9.3[CD28]_H0L0, hu9.3[CD28]_H1L1, 9G2[CD28]_H0L0, 9G2[CD28]_H1L1, 2F10A3.140[CD28]_H1L1, and TN228[CD28]_H4L2 (Figures 15, 18, and 21). In some embodiments, the CD28 ABD includes a VH / VL pair selected from VH and VL shown in Figures 15, 16, and 17.

[0136] In addition to the parent CDR sets disclosed in Figures and Sequence Listings that form the ABD for CD28, variant CD28 ABDs are provided herein that have a CDR containing at least one modification of the CD28 ABD CDRs disclosed herein (e.g., Figures 15, 18, and 21, and Sequence Listings). In one embodiment, the CD28 ABD of the target anti-CD28 × anti-ENPP3 antibody includes a set of six CDRs having 1, 2, 3, 4, 5, 6, 7, 8, 9, and 10 amino acid modifications, compared to the six CDRs of the CD28 ABDs described herein, including Figures and Sequence Listings. In an exemplary embodiment, the CD28 ABD of the target anti-CD28 × anti-ENPP3 antibody includes the following CD28 When compared to one of the six CDRs in ABD, it includes a set of six CDRs with 1, 2, 3, 4, 5, 6, 7, 8, 9, and 10 amino acid modifications: 1A7[CD28]_H1L1, 1A7[CD28]_H1.1_L1, 1A7[CD28]_H1_L1.71, 1A7[CD28]_H1.1_L1.71, 1A7[CD28]_H1.14_L1, 1A7[CD28]_H1.14_L1.71, CD28.3[CD28]_H0L0, hCD28.3[CD28]_H1L1, 5.11A1[CD28]_H0L0, TGN1412 _H1L1, 341VL34[CD28]_H1L1, 341VL36[CD28]_H1L1, 281VL4[CD28]_H1L1, HuTN228[CD28]_H1L1, PV1[CD28]_H0L0, m9.3[CD28]_H0L0, hu9.3[CD28]_H1L1, 9G2[CD28]_H0L0, 9G2[CD28]_H1L1, 2F10A3.140[CD28]_H1L1, and TN228[CD28]_H4L2 (Figures 15, 18, and 21). In certain embodiments, the CD28 of the target anti-CD28 × anti-ENPP3 antibody ABD can bind to the CD28 antigen when measured by at least one of the following assays: Biacore, surface plasmon resonance (SPR), flow cytometry, and / or BLI (biolayer interferometry, e.g., Octet assay) (the latter being particularly used in many embodiments). In certain embodiments, CD28 ABD can bind to human CD28 antigen (see Figure 1).

[0137] In some embodiments, the CD28 ABD of the anti-CD28 × anti-ENPP3 antibody of interest comprises six CDRs that are at least 90, 95, 97, 98, or 99% identical to six CDRs of the CD28 ABD described herein, including in the figures and sequence listings. In exemplary embodiments, the CD28 ABD of the anti-CD28 × anti-ENPP3 antibody of interest comprises the following CD28 1A7[CD28]_H1L1, 1A7[CD28]_H1.1_L1, 1A7[CD28]_H1_L1.71, 1A7[CD28]_H1.1_L1.71, 1A7[CD28]_H1.14_L1, 1A7[CD28]_H1.14_L1.71, CD28.3[CD28]_H0L0, hCD28.3[CD28]_H1L1, 5.11A1[CD28]_H0 L0, TGN1412_H1L1, 341VL34[CD28]_H1L1, 341VL36[CD28]_H1L1, 281VL4[CD28]_H1L1, HuTN228[CD28]_H1L1, PV1[CD28]_H0L0, m9.3[CD28]_H0L0, hu9.3[CD28]_H1L1, 9G2[CD28]_H0L0, 9G2[CD28]_H1L1, 2F10A3.140[CD28]_H1L1, and TN228[CD28]_H4L2 (Figures 15, 18, and 21). In certain embodiments, CD28 ABD can bind to CD28 when measured by at least one of Biacore, surface plasmon resonance (SPR), flow cytometry, and / or BLI (biolayer interferometry, e.g., Octet assay) assays (the latter being particularly used in many embodiments). In certain embodiments, CD28 ABD can bind to human CD28 antigen (see Figure 1).

[0138] In another exemplary embodiment, the CD28 ABD of the target anti-CD28 × anti-ENPP3 antibody comprises one of the variable weight (VH) domains and variable light (VL) domains of the CD28 ABDs described herein, including in the figures and sequence listings. In the exemplary embodiment, the CD28 ABD is the following CD28 One of the ABDs: 1A7[CD28]_H1L1, 1A7[CD28]_H1.1_L1, 1A7[CD28]_H1_L1.71, 1A7[CD28]_H1.1_L1.71, 1A7[CD28]_H1.14_L1, 1A7[CD28]_H1.14_L1.71, CD28.3[CD28]_H0L0, hCD28.3[CD28]_H1L1, 5.11A1[CD28]_H0L0, TGN1412_H1L1, 341VL 34[CD28]_H1L1, 341VL36[CD28]_H1L1, 281VL4[CD28]_H1L1, HuTN228[CD28]_H1L1, PV1[CD28]_H0L0, m9.3[CD28]_H0L0, hu9.3[CD28]_H1L1, 9G2[CD28]_H0L0, 9G2[CD28]_H1L1, 2F10A3.140[CD28]_H1L1, and TN228[CD28]_H4L2 (Figures 15, 18, and 21). In some embodiments, the CD28 ABD includes pairs of VH and VL selected from VH and VL shown in Figures 15, 16, and 17.

[0139] In some embodiments, the anti-CD28 × anti-ENPP3 antibody comprises a CD28 ABD including a variable heavy domain and / or variable light domain, which are variants of the CD28 ABD VH and VL domains disclosed herein. In one embodiment, the variant VH domain and / or VL domain has 1, 2, 3, 4, 5, 6, 7, 8, 9, or 10 amino acid changes from the VH and / or VL domains of the CD28 ABD described herein, including figures and sequence listings. In exemplary embodiments, the variant VH domain and / or VL domain is the following CD28 One of the ABD has 1, 2, 3, 4, 5, 6, 7, 8, 9, or 10 amino acid changes from the VH and / or VL domain: 1A7[CD28]_H1L1, 1A7[CD28]_H1.1_L1, 1A7[CD28]_H1_L1.71, 1A7[CD28]_H1.1_L1.71, 1A7[CD28]_H1.14_L1, 1A7[CD28]_H1.14_L1.71, CD28.3[CD28]_H0L0, hCD28.3[CD28]_H1L1, 5.11A1[CD28] _H0L0, TGN1412_H1L1, 341VL34[CD28]_H1L1, 341VL36[CD28]_H1L1, 281VL4[CD28]_H1L1, HuTN228[CD28]_H1L1, PV1[CD28]_H0L0, m9.3[CD28]_H0L0, hu9.3[CD28]_H1L1, 9G2[CD28]_H0L0, 9G2[CD28]_H1L1, 2F10A3.140[CD28]_H1L1, and TN228[CD28]_H4L2 (Figures 15, 18, and 21). In some embodiments, the variation is located within the VH domain shown in Figures 15-18 and 21. In some embodiments, the changes are located within the VL domain shown in Figures 15-18 and 21. In some embodiments, the changes are located within the VH and VL domains shown in Figures 15-18 and 21. In some embodiments, one or more amino acid changes are located within the VH and / or VL framework regions (FR1, FR2, FR3, and / or FR4). In some embodiments, one or more amino acid changes are located within one or more CDRs.In certain embodiments, CD28 ABD can bind to CD28 when measured by at least one of Biacore, surface plasmon resonance (SPR), flow cytometry, and / or BLI (biolayer interferometry, e.g., Octet assay) assays (the latter being particularly used in many embodiments). In certain embodiments, CD28 ABD can bind to human CD28 antigen (see Figure 1).

[0140] In one embodiment, the variant VH and / or VL domains are at least 90, 95, 97, 98, or 99% identical to the VH and / or VL of CD28 ABD described herein, including in the figures and sequence listings. In an exemplary embodiment, the variant VH and / or VL domains are the following CD28 These are at least 90, 95, 97, 98, or 99% identical to one of the VH and / or VL of ABD: 1A7[CD28]_H1L1, 1A7[CD28]_H1.1_L1, 1A7[CD28]_H1_L1.71, 1A7[CD28]_H1.1_L1.71, 1A7[CD28]_H1.14_L1, 1A7[CD28]_H1.14_L1.71, CD28.3[CD28]_H0L0, hCD28.3[CD28]_H1L1, 5.11A1[CD28]_H0L0 , TGN1412_H1L1, 341VL34[CD28]_H1L1, 341VL36[CD28]_H1L1, 281VL4[CD28]_H1L1, HuTN228[CD28]_H1L1, PV1[CD28]_H0L0, m9.3[CD28]_H0L0, hu9.3[CD28]_H1L1, 9G2[CD28]_H0L0, 9G2[CD28]_H1L1, 2F10A3.140[CD28]_H1L1, and TN228[CD28]_H4L2 (Figures 15, 18, and 21). In some embodiments, the CD28 ABD contains VH which is at least 90, 95, 97, 98, or 99% identical to the VH domain shown in Figures 15-18 and 21. In some embodiments, the CD28 ABD contains VL which is at least 90, 95, 97, 98, or 99% identical to the VL domain shown in Figures 15-18 and 21. In some embodiments, the CD28 ABD contains VH and VL which are at least 90, 95, 97, 98, or 99% identical to the VH and VL domains shown in Figures 15-18 and 21. In certain embodiments, the CD28 ABD can bind to CD28 when measured by at least one of Biacore, surface plasmon resonance (SPR), flow cytometry, and / or BLI (biolayer interferometry, e.g., Octet assay) assays (the latter being particularly utilized in many embodiments).In certain embodiments, the CD28 ABD can bind to the human CD28 antigen (see Figure 1).

[0141] In some embodiments, the CD28-binding domain of the target anti-CD28 × anti-ENPP3 antibody includes VH, which contains one of the sequences VHCDR1-3 or HFR1-4 shown in Figure 19A. In some embodiments, the CD28-binding domain includes VL, which contains one of the sequences VLCDR1-3 or LFR1-4 shown in Figure 19B.

[0142] In some embodiments, the anti-CD28 × anti-ENPP3 antibody includes a CD28-binding domain comprising VH and VL selected from the following:

[0143] (i) VH containing vhCDR1, vhCDR2, and vhCDR3, each having the amino acid sequences of vhCDR1, vhCDR2, and vhCDR3 of any VH or variant thereof of the CD28 binding domain shown in Figures 15, 18, and 21; and (ii) VL containing vlCDR1, vlCDR2, and vlCDR3, each having the amino acid sequences of vlCDR1, vlCDR2, and vlCDR3 of any VL or variant thereof of the CD28 binding domain shown in Figures 15, 18, and 21; or

[0144] (i) a VH containing vhCDR1, vhCDR2, and vhCDR3, each having the amino acid sequences of vhCDR1, vhCDR2, and vhCDR3 of the VH or its variant shown in Figure 15 or 16, and (ii) a VL containing vlCDR1, vlCDR2, and vlCDR3, each having the amino acid sequences of vlCDR1, vlCDR2, and vlCDR3 of the VL or its variant shown in Figure 15 or 17.

[0145] In some embodiments, the anti-CD28 × anti-ENPP3 antibody includes a CD28-binding domain comprising VH and VL selected from the following:

[0146] (i) VH having the amino acid sequence of any VH or a variant thereof of the CD28 binding domain shown in Figures 15, 18, and 21, and (ii) VL having the amino acid sequence of any VL or a variant thereof of the CD28 binding domain shown in Figures 15, 18, and 21; or

[0147] (i) VH having the amino acid sequence of VH or a variant thereof shown in Figure 15 or 16, and (ii) VL having the amino acid sequence of VL or a variant thereof shown in Figure 15 or 17.

[0148] B.ENPP3 binding domain The anti-CD28 × anti-ENPP3 antibodies provided herein contain at least one ENPP3-binding domain. A target antibody containing such an ENPP3 antigen-binding domain (e.g., an anti-ENPP3 × anti-CD3 bispecific antibody) favorably targets cells expressing higher levels of ENPP3 than cells expressing normal levels of ENPP3 (e.g., normal cells).

[0149] In some embodiments, the ENPP3 ABD of the anti-CD28 × anti-ENPP3 antibody comprises a set of six CDRs, either as the underlined CDRs as shown in the sequence listing and Figure 22, or as the CDRs identified using other alignments within the variable weight (VH) domain and variable light domain (VL) domain sequences as shown in Figure 22 and the sequence listing, when a different numbering scheme is used as described herein and shown in Table 2 (see Table 2).

[0150] In exemplary embodiments, the ENPP3 ABD of the anti-CD28 × anti-ENPP3 antibody comprises one of the variable weight (VH) domains and variable light (VL) domains of the ENPP3 ABDs described herein, including figures and sequence listings. In exemplary embodiments, the ENPP3 ABD of the anti-CD28 × anti-ENPP3 antibody is the ENPP3 ABD shown in Figure 22.

[0151] In addition to the parent ENPP3 variable heavy domain and variable light domain disclosed herein, ENPP3 ABDs are provided herein that include variable heavy domains and / or variable light domains which are variants of the ENPP3 ABD VH and VL domains disclosed herein. In one embodiment, the variant VH domain and / or VL domain has 1, 2, 3, 4, 5, 6, 7, 8, 9, or 10 amino acid changes from the VH and / or VL domains of ENPP3 ABD described herein, including figures and sequence listings. In exemplary embodiments, the variant VH domain and / or VL domain has 1, 2, 3, 4, 5, 6, 7, 8, 9, or 10 amino acid changes from the VH and / or VL domains of ENPP3 ABD shown in Figure 22. In some embodiments, the changes are located within the VH domain shown in Figure 22. In some embodiments, the changes are located within the VL domain shown in Figure 22. In some embodiments, the changes are located within the VH and VL domains shown in Figure 22. In some embodiments, one or more amino acid changes are located within the VH and / or VL framework regions (FR1, FR2, FR3, and / or FR4). In some embodiments, one or more amino acid changes are located within one or more of the vhCDR1-3 and / or vlCDR1-3. In some embodiments, the ENPP3 variant includes the vhCDR1-3 and / or vlCDR1-3 of any of the ENPP3 ABDs in Figure 22. In certain embodiments, the ENPP3 ABD of an anti-CD28 × anti-ENPP3 antibody is capable of binding to ENPP3 when measured by at least one of the following assays: Biacore, surface plasmon resonance (SPR), flow cytometry, and / or BLI (biolayer interferometry, e.g., Octet assay) (the latter being particularly utilized in many embodiments). In certain embodiments, the ENPP3 ABD is capable of binding to the human ENPP3 antigen (Figure 2).

[0152] In one embodiment, the variant VH and / or VL domains are at least 90, 95, 97, 98, or 99% identical to the VH and / or VL of the ENPP3 ABD described herein, including in the Figures and Sequence Listings. In an exemplary embodiment, the variant VH and / or VL domains are at least 90, 95, 97, 98, or 99% identical to the VH and / or VL of the ENPP3 ABD shown in Figure 22. In some embodiments, the ENPP3 ABD includes VH which is at least 90, 95, 97, 98, or 99% identical to the VH domain shown in Figure 22. In some embodiments, the ENPP3 ABD includes VL which is at least 90, 95, 97, 98, or 99% identical to the VL domain shown in Figure 22. In some embodiments, the ENPP3 ABD includes VH and VL which are at least 90, 95, 97, 98, or 99% identical to the VH and VL domains shown in Figure 22. In some embodiments, the ENPP3 ABD comprises any six CDRs (vhCDR1-3 and vlCDR1-3) of the ENPP3 ABD shown in Figure 22. In certain embodiments, the ENPP3 ABD of anti-CD28 × anti-ENPP3 antibody is capable of binding to ENPP3 when measured by at least one of the following: Biacore, surface plasmon resonance (SPR), flow cytometry, and / or BLI (biolayer interferometry, e.g., Octet assay) assays (the latter being particularly utilized in many embodiments). In certain embodiments, the ENPP3 ABD is capable of binding to the human ENPP3 antigen (Figure 2).

[0153] In some embodiments, the anti-CD28 × anti-ENPP3 antibody is a bivalent antibody (e.g., a 1+1 Fab-scFv-Fc type antibody) containing one ENPP3-binding domain. In other embodiments, the anti-CD28 × anti-ENPP3 antibody is a trivalent antibody (e.g., a 2+1 mAb-scFv, 2+1 Fab2-scFv-Fc, and 2+1 Fab2-FcxscFv-Fc type antibody) containing two ENPP3-binding domains.

[0154] C. Chimeric antibodies and humanized antibodies In certain embodiments, the target antibodies provided herein include heavy chain variable regions from specific germline heavy chain immunoglobulin genes and / or light chain variable regions from specific germline light chain immunoglobulin genes. For example, such antibodies include, or may consist of, human antibodies containing heavy chain or light chain variable regions that are "products of" or "derived from" a specific germline sequence. Human antibodies that are "products of" or "derived from" a human germline immunoglobulin sequence can be identified, for example, by comparing the amino acid sequence of the human antibody with the amino acid sequence of the human germline immunoglobulin and selecting the human germline immunoglobulin sequence that is closest in sequence to (i.e., maximum % identity) the sequence of the human antibody (using the methods outlined herein). Human antibodies that are "products of" or "derived from" a specific human germline immunoglobulin sequence may contain amino acid differences when compared to the germline sequence, for example, by the intentional introduction of naturally occurring somatic mutations or site-directed mutations. However, humanized antibodies are typically at least 90% identical to the amino acid sequence encoded by human germline immunoglobulin genes and contain amino acid residues that identify the antibody as derived from a human sequence when compared to germline immunoglobulin amino acid sequences of other species (e.g., mouse germline sequences). In given cases, a humanized antibody may be at least 95, 96, 97, 98, or 99% identical in amino acid sequence to the amino acid sequence encoded by a germline immunoglobulin gene, or even at least 96%, 97%, 98, or 99%. Typically, a humanized antibody derived from a particular human germline sequence will show a difference of 10 to 20 amino acids or less from the amino acid sequence encoded by the human germline immunoglobulin gene (before the introduction of any skew, pI, and attenuation variants as used herein (i.e., the number of variants is usually small before the introduction of the variants of the present invention)).In certain cases, the humanized antibody may exhibit a difference of 5 or fewer amino acids, or further 4, 3, 2, or 1 or fewer, from the amino acid sequence encoded by the germline immunoglobulin gene (in this case as well, before the introduction of any skew, pI, and attenuation variants as specified herein (i.e., the number of variants is usually small before the introduction of the variants of the present invention)).

[0155] In one embodiment, the parent antibody is affinity-matured as known in the art. Structure-based methods may be used for humanization and affinity maturation, for example, as described in USSN 11 / 004,590. Humanization and / or affinity maturation of the antibody variable region may be performed using selection-based methods, including, but not limited to, those described in Wu et al., 1999, J.Mol.Biol.294:151-162; Baca et al., 1997, J.Biol.Chem.272(16):10678-10684; Rosok et al., 1996, J.Biol.Chem.271(37):22611-22618; Rader et al., 1998, Proc.Natl.Acad.Sci.USA 95:8910-8915; Krauss et al., 2003, Protein Engineering 16(10):753-759 (all incorporated by reference). Other humanization methods, including but not limited to those described in USSN09 / 810,510; Tan et al.,2002,J.Immunol.169:1119-1125;De Pascalis et al.,2002,J.Immunol.169:3076-3084 (all incorporated by reference), may involve grafting only a portion of the CDR.

[0156] D. Heterodimer antibodies In exemplary embodiments, the anti-CD28×anti-ENPP3 antibody provided herein is a heterodimer bispecific antibody comprising two variant Fc domain sequences. Such variant Fc domains include amino acid modifications to facilitate the self-construction and / or purification of the heterodimer antibody.

[0157] An ongoing challenge in antibody technology is the desire for "bispecific" antibodies that bind simultaneously to two different antigens, typically by bringing the different antigens into proximity and resulting in novel functionalities and therapies. These antibodies are usually produced by incorporating genes for each heavy and light chain into the host cell. This typically results in the formation of the desired heterodimer (AB), as well as two homodimers (AA and BB (without the light chain heterodimer problem)). However, a major obstacle in the formation of bispecific antibodies is the bias towards the formation of the desired heterodimer antibody relative to the formation of the homodimer, and / or the difficulty in purifying the heterodimer antibody from the homodimer.

[0158] Numerous mechanisms exist that can be used to generate target heterodimer antibodies. Furthermore, as will be understood by those skilled in the art, these different mechanisms can be combined to ensure high heterodimerization. Amino acid modifications that facilitate the generation and purification of heterodimers are usually collectively referred to as “heterodimerizing variants.” As will be discussed below, heterodimerizing variants include “skew” variants (e.g., the “knob and hole” and “charge pair” variants described below) as well as “pI variants” that enable the purification of heterodimers from homodimers. Useful mechanisms for heterodimerization, as described in U.S. Patent No. US9,605,084 (which in whole and in particular is incorporated herein by reference with respect to the discussion of heterodimerization variants below), include “knobs and holes” ("KIH") as described in U.S. Patent No. US9,605,084, “electrostatic steering” or “charge pairs” as described in U.S. Patent No. US9,605,084, pI variants as described in U.S. Patent No. US9,605,084, and additional general Fc variants outlined in U.S. Patent No. US9,605,084 and below.

[0159] Heterodimerized variants useful for the formation and purification of target heterodimerized antibodies (e.g., bispecific antibodies) are discussed in more detail below.

[0160] 1. Scubarian In some embodiments, heterodimer antibodies include scubarians, which are one or more amino acid modifications in the first Fc domain (A) and / or the second Fc domain (B) that prioritize the formation of an Fc heterodimer (an Fc dimer (AB) containing the first and second Fc domains) rather than an Fc homodimer (an Fc dimer (AA or BB) containing two first Fc domains or two second Fc domains). Preferred scubarians are shown in Figure 29, and in Figures 3 and 8, of U.S. Publication No. 2016 / 0355608 (which is incorporated herein by reference in whole and in particular with respect to its disclosure of scubarians).

[0161] One particular type of scubarian is usually referred to in the art as “knob and hole” when referring to amino acid manipulations that produce steric effects that favor heterodimerization and not homodimerization, as described in USSN 61 / 596,846, Ridgway et al., Protein Engineering 9(7):617(1996); Atwell et al., J.Mol.Biol.1997 270:26; U.S. Patent No. 8,216,805 (all of these are incorporated herein by reference in whole and in particular with respect to the disclosure of “knob and hole” mutations). This is sometimes referred to herein as “steric variant.” The figure identifies numerous “monomer A-monomer B” pairs that depend on “knob and hole.” Furthermore, as described by Merchant et al., Nature Biotech. 16:677 (1998), these “knob and hole” mutations can be combined with disulfide bonds to further favor the formation of Fc heterodimers.

[0162] Another method used in the formation of heterodimers is sometimes referred to as “electrostatic steering,” as described in Gunasekaran et al., J. Biol. Chem. 285(25):19637 (2010) (the whole of which is incorporated herein by reference). This is sometimes referred to herein as “charge pairing.” In this embodiment, electrostatics is used to bias towards the formation of heterodimers. As those skilled in the art will understand, these can also affect pI and thus purification, and thus in some cases can be considered pI variants. However, these are classified as “scuba riants” because they were generated to force heterodimerization and were not used as purification tools. These include, but are not limited to, D221E / P228E / L368E paired with D221R / P228R / K409R (for example, these are "corresponding monomer sets") and C220E / P228E / 368E paired with C220R / E224R / P228R / K409R.

[0163] In some embodiments, scubarians favorably and simultaneously prioritize heterodimerization based on both a “knob and hole” mechanism and an “electrostatic steering” mechanism. In some embodiments, a heterodimer antibody contains one or more sets of such heterodimerized scubarians. These variants are provided in “pairs” of “sets,” that is, one set of the pair is incorporated into a first monomer, and the other set of the pair is incorporated into a second monomer. It should be noted that these sets do not necessarily behave as “knob in hole” variants having a one-to-one correspondence between residues on one monomer and residues on the other monomer. In other words, pairs of these sets can rather form interfaces between the two monomers, thereby promoting heterodimerization and suppressing homodimerization, making it possible to raise the proportion of heterodimers spontaneously formed under biological conditions to over 90% (25% homodimer A / A: 50% heterodimer A / B: 25% homodimer B / B) rather than the expected 50%. Exemplary heterodimerizing "skewed" variants are shown in Figures 3 and 8. Such "skewed" variants include, but are not limited to, S364K / E357Q:L368D / K370S;L368D / K370S:S364K;L368E / K370S:S364K;T411T / E360E / Q362E:D401K;L368D / K370S:S364K / E357L;K370S:S364K / E357Q (EU numbering).

[0164] In exemplary embodiments, the heterodimer antibody includes the S364K / E357Q:L368D / K370S;L368D / K370S:S364K;L368E / K370S:S364K;T411T / E360E / Q362E:D401K;L368D / K370S:S364K / E357L;K370S:S364K / E357Q; or T366S / L368A / Y407V:T366W (optionally containing a cross-linked disulfide, T366S / L368A / Y407V / Y349C:T366W / S354C) "skew" variant amino acid substitution set (EU numbering). In exemplary embodiments, the heterodimer antibody contains the "S364K / E357Q:L368D / K370S" amino acid substitution set. From a nomenclature standpoint, the pair "S364K / E357Q:L368D / K370S" means that one monomer contains an Fc domain with amino acid substitutions S364K and E357Q, and the other monomer contains an Fc domain with amino acid substitutions L368D and K370S; as described above, the "strandiness" of these pairs depends on the initiation pI.

[0165] In some embodiments, the scuba riants provided herein may, optionally and independently, be incorporated into one or both of the first and second Fc domains of a heterodimer antibody along with any other modifications, including but not limited to other scuba riants (see, for example, Figure 37 of U.S. Publication No. 2012 / 0149876 (which is incorporated herein by reference in particular for its disclosure of scuba riants)), pI variants, isomorphic variants, FcRn variants, attenuation variants, and the like. Furthermore, individual modifications may also, independently and optionally, be included in or excluded from the heterodimer antibody in question.

[0166] In some embodiments, the scuba rians outlined herein may be optionally and independently incorporated into one or both heavy chain monomers together with any pI variant (or other variants, e.g., Fc variant, FcRn variant, etc.), and may be independently and optionally included in or excluded from the heterodimer antibody of interest.

[0167] 2. Purified variant In some embodiments, the heterodimer antibody includes a purified variant that advantageously allows for the separation of the heterodimer protein (e.g., anti-CD28 × anti-ENPP3 bispecific antibody) from the homodimer protein.

[0168] Several fundamental mechanisms exist that can facilitate the purification of heterodimer antibodies. For example, modifications to one or both of antibody heavy chain monomers A and B, such that each monomer has a different pI, enable isoelectric purification of heterodimer AB antibodies from monomer AA and BB proteins. Alternatively, several skeletal configurations, such as the "1+1 Fab-scFv-Fc" configuration and the "2+1 Fab2-scFv-Fc" configuration, enable size-based separation. As mentioned above, it is also possible to "distort" heterodimer formation relative to homodimers using scuba riants. Therefore, combinations of heterodimerizing scuba riants and purification variants are particularly useful for the heterodimer antibodies provided herein.

[0169] Furthermore, as will be outlined more thoroughly below, depending on the form of the heterodimeric antibody, purified variants contained in the constant region and / or Fc domain and / or domain linker of the monomer may be used. In some embodiments, the heterodimeric antibody includes additional modifications for alternative functionality, such as Fc, FcRn, and KO variants, which may also produce pI changes.

[0170] In some embodiments, the heterodimer antibodies provided herein include at least one monomer having one or more modifications (i.e., “pI variants”) that alter the pI of the monomer. Generally, as understood by those skilled in the art, there are two general categories of pI variants: those that increase the pI of the protein (basic changes) and those that decrease the pI of the protein (acidic changes). As described herein, all combinations of these variants are possible: one monomer may be wild-type, or a variant that does not exhibit a significantly different pI from the wild-type, while the other may be either more basic or more acidic. Alternatively, each monomer may be modified to be one more basic and the other more acidic.

[0171] Depending on the form of the heterodimer antibody, the pI variant may be contained within the constant and / or Fc domain of the monomer, or a charged linker (either a domain linker or an scFv linker) may be used. That is, antibody forms utilizing scFv(s), e.g., the "1+1 Fab-scFv-Fc" form, may include a charged scFv linker (either positive or negative) to provide a further increase in pI for purification purposes. As will be understood by those skilled in the art, some 1+1 Fab-scFv-Fc and 2+1 Fab2-scFv-Fc forms are useful simply with a charged scFv linker and without additional pI adjustment, but the present invention also provides pI variants located in one or both monomers and / or in the charged domain linker. Furthermore, additional amino acid manipulation for alternative functionality may also confer pI changes, e.g., Fc, FcRn, and KO variants.

[0172] In target heterodimer antibodies that utilize pI as a separation mechanism to enable the purification of heterodimer proteins, amino acid variants are introduced into one or both monomeric polypeptides. That is, the pI of one monomer (referred herein as "monomer A" for simplicity) may be manipulated separately from monomer B, or both monomer A and B changes may be altered by increasing pI for monomer A and decreasing pI for monomer B. As will be more thoroughly outlined below, pI changes in either or both monomers may be achieved by removing or adding charged residues (e.g., a neutral amino acid is replaced by a positively or negatively charged amino acid residue (e.g., glycine to glutamic acid)), changing a charged residue from positive or negative to the opposite charge (e.g., aspartic acid to lysine), or changing a charged residue to a neutral residue (e.g., loss of charge; lysine to serine). Numerous of these variants are shown in Figures 3 and 4.

[0173] Therefore, in some embodiments, the target heterodimer antibody includes amino acid modifications in the constant region that alter the isoelectric point (pI) of at least one (if not both) monomer of the dimer protein in order to form a “pI antibody” by incorporating an amino acid substitution (“pI variant” or “pI substitution”) in one or both monomers. As shown herein, if the pIs of the two monomers differ by only 0.1 pH units (0.2, 0.3, 0.4, and 0.5 or more are all utilized in the present invention), separation of heterodimers from two homodimers can be achieved.

[0174] As will be understood by those skilled in the art, the number of pI variants present in each or both monomers(s) to obtain good separation will depend in part on the starting pIs of the components, e.g., in the form 1+1 Fab-scFv-Fc, 2+1 Fab2-scFv-Fc, 1+1 CLC, and 2+1 CLC, on the scFv(1+1 Fab-scFv-Fc, 2+1 Fab2-scFv-Fc) and Fab(s)(s) of the target(s). That is, the Fv sequences of the two target antigens are calculated and then the decision is made to determine which monomers to operate on or in which "direction" (e.g., more positive or more negative). As is known in the art, different Fvs will have different starting pIs to utilize in the present invention. Typically, as outlined herein, the pIs are operated to result in a total pI difference of at least about 0.1 log (preferably 0.2 to 0.5 as outlined herein) between each monomer.

[0175] When pI variants are used to achieve heterodimerization, a more modular approach is provided for designing and purifying antibody-containing bispecific proteins by utilizing the constant region(s) of the heavy chain(s). Thus, in some embodiments, heterodimerizing variants (including skew and pI heterodimerizing variants) are not included in the variable region, so that each individual antibody must be manipulated. Also, in some embodiments, the potential for immunogenicity due to pI variants is significantly reduced by introducing pI variants from different IgG isotypes so that pI can be altered without introducing significant immunogenicity. Thus, an additional problem to be solved is the elucidation of low pI constant domains with a high human sequence content, e.g., minimizing or avoiding non-human residues at any particular position. As an alternative to or in addition to isomorphic substitutions, the potential for immunogenicity due to pI variants is significantly reduced by utilizing isosteric substitutions (e.g., Asn to Asp; and Gln to Glu).

[0176] As discussed below, the secondary benefits that may arise from this pI manipulation are also the extension of serum half-life and increased FcRn binding. Specifically, as described in U.S. Publication No. US2012 / 0028304 (which is incorporated in its entirety by reference), reducing the pI of antibody constant domains (including those found in antibodies and Fc fusions) can result in longer serum retention in vivo. These pI variants for increased serum half-life also facilitate pI alteration for purification.

[0177] Furthermore, it should be noted that pI variants provide additional benefits for the analysis and quality control processes of bispecific antibodies, as their ability to exclude, minimize, and distinguish homodimers is important. Similarly, the ability to reliably test the reproducibility of heterodimeric antibody production is important.

[0178] Typically, the most commonly used embodiments rely on a set of variants, combining a scuba riant that promotes heterodimerization rather than homodimerization with a pI variant that increases the pI difference between the two monomers, making it easier to purify the heterodimer from the homodimer.

[0179] Exemplary combinations of pI variants are shown in Figures 3 and 4, and Figure 30, of U.S. Publication No. 2016 / 0355608 (all of which are incorporated herein by reference in whole and in particular for the disclosure of pI variants). Preferred combinations of pI variants are shown in Figures 3 and 4. As outlined herein and shown in the figures, these variations are shown relative to IgG1, but all isotypes and isotype hybrids can be modified in the same way. R133E and R133Q can also be used when the heavy chain constant domain is derived from IgG2-4.

[0180] In one embodiment, a preferred combination of pI variants has one monomer (negative Fab side) containing the has208D / 295E / 384D / 418E / 421D variant (N208D / Q295E / N384D / Q418E / N421D compared to human IgG1) and a second monomer (positive scFv side) containing a positively charged scFv linker containing (GKPGS)4 (SEQ ID NO: XXX). However, as will be understood by those skilled in the art, the first monomer contains a CH1 domain including position 208. Therefore, in constructs that do not contain the CH1 domain (for example, antibodies that do not utilize the CH1 domain in one of their domains), the preferred negative pI variant Fc set includes the 295E / 384D / 418E / 421D variants (Q295E / N384D / Q418E / N421D compared to human IgG1).

[0181] Therefore, in some embodiments, one monomer has the set of substitutions shown in Figure 8, and the other monomer has a charged linker (either a charged scFv linker (since the monomer contains scFv) or a charged domain linker (which can be selected from those shown in Figure 6, as the form indicates)).

[0182] In some embodiments, modifications are made at the hinge of the Fc domain, including positions 216, 217, 218, 219, 220, 221, 222, 223, 224, 225, 226, 227, 228, 229, and 230, based on EU numbering. Thus, pI mutations and especially substitutions may occur at one or more of positions 216-230, with 1, 2, 3, 4, or 5 mutations being utilized. In this case as well, all possible combinations are intended alone or in conjunction with other pI variants in other domains.

[0183] Specific substitutions used to reduce the pI of the hinge domain include, but are not limited to, deletions at position 221, non-native valine or threonine at position 222, deletion at position 223, non-native glutamate at position 224, deletion at position 225, deletion at position 235, and deletion or non-native alanine at position 236. In some cases, only pI substitutions occur in the hinge domain, while in other cases, these substitutions are added in any combination to other pI variants in other domains.

[0184] In some embodiments, mutations may occur in the CH2 region, including positions 233, 234, 235, 236, 274, 296, 300, 309, 320, 322, 326, 327, 334, and 339, based on EU numbering. It should be noted that changes at positions 233–236 may occur (along with 327A) to increase effector function in the IgG2 backbone. In this case as well, all possible combinations of these 14 positions may occur; for example, antibodies provided herein may include variant Fc domains having 1, 2, 3, 4, 5, 6, 7, 8, 9, or 10 CH2 pI substitutions.

[0185] Specific substitutions used to reduce the pI of the CH2 domain include, but are not limited to, non-native glutamine or glutamate at position 274, non-native phenylalanine at position 296, non-native phenylalanine at position 300, non-native valine at position 309, non-native glutamate at position 320, non-native glutamate at position 322, non-native glutamate at position 326, non-native glycine at position 327, non-native glutamate at position 334, non-native threonine at position 339, and all possible combinations within the CH2 and other domains.

[0186] In this embodiment, modifications may be independently and optionally selected from positions 355, 359, 362, 384, 389, 392, 397, 418, 419, 444, and 447 (EU numbering) of the CH3 domain. Specific substitutions used to reduce the pI of the CH3 domain include, but are not limited to, non-native glutamine or glutamate at position 355, non-native serine at position 384, non-native asparagine or glutamate at position 392, non-native methionine at position 397, non-native glutamate at position 419, non-native glutamate at position 359, non-native glutamate at position 362, non-native glutamate at position 389, non-native glutamate at position 418, non-native glutamate at position 444, and deletion or non-native aspartate at position 447.

[0187] In some embodiments, anti-CD28 × anti-ENPP3 antibodies include an amino acid substitution in one of their Fc domains that reduces binding to protein A. Such purified variants produce heterodimers with asymmetric binding to protein A, which can be used to separate the heterodimers from the homodimer population by a pH gradient. Exemplary purified amino acid substitutions that reduce binding to protein A include, but are not limited to, H435R and Y436F (IgG1 CH3 domain, EU numbering). See, for example, US2010331527 (which is incorporated by reference in whole and in particular for relevant disclosures regarding Fc domain modifications for reducing protein A binding).

[0188] 3. Isomorphic Variants Furthermore, many embodiments of the target heterodimer antibodies rely on the “introduction” of pI amino acids at specific positions from one IgG isotype to another, thus reducing or eliminating the possibility of unwanted immunogenicity being introduced into the variant. Numerous of these are shown in Figure 21 of U.S. Publication 2014 / 0370013 (incorporated herein by reference). That is, IgG1 is a common isotype for therapeutic antibodies for a variety of reasons, including high effector function. However, the polyconstant region of IgG1 has a higher pI than that of IgG2 (8.10 vs. 7.31). By introducing IgG2 residues at specific positions into the IgG1 backbone, the pI of the resulting monomer is reduced (or increased), additionally exhibiting a longer serum half-life. For example, IgG1 has glycine at position 137 (pI 5.97), while IgG2 has glutamic acid (pI 3.22); introducing glutamic acid would affect the pI of the resulting protein. As described below, numerous amino acid substitutions are usually required to significantly affect the pI of variant antibodies. However, it should be noted that even changes in the IgG2 molecule can lead to an increase in serum half-life, as discussed below.

[0189] In other embodiments, as described further below, non-isotypic amino acid changes are made to reduce the overall charge state of the resulting protein (for example, by changing higher pI amino acids to lower pI amino acids) or to allow structural modifications for stability or other reasons.

[0190] Furthermore, significant changes can be observed in each monomer of the heterodimer by manipulating the pI of both the heavy and light steady domains. As discussed herein, by making the pI of the two monomers differ by at least 0.5, separation by ion-exchange chromatography, isoelectric focusing, or other isoelectric-sensitive methods may be possible.

[0191] 4. Calculation of pI The pI of each monomer of the antibodies provided herein may depend on the pI of the variant heavy chain constant domain as well as the pI of the total monomer including the variant heavy chain constant domain and its fusion partner. Therefore, in some embodiments, the pI change is calculated based on the variant heavy chain constant domain using the chart in Figure 19 of U.S. Publication 2014 / 0370013. As discussed herein, which monomer to manipulate is typically determined by the Fv and the pI inherent in the skeletal region. Alternatively, the pI of each monomer may be compared.

[0192] 5. pI variants that also confer better FcRn in vivo binding. If a pI variant reduces monomeric pI, the pI variant may have the added benefit of improving serum retention in vivo.

[0193] Although still under investigation, the Fc region is thought to have a longer half-life in vivo because binding to FcRn at pH 6 in the endosome sequesters the Fc (Ghetie and Ward, 1997 Immunol Today. 18(12):592-598 (incorporated herein by reference in its entirety)). The endosomal compartment then recycles the Fc to the cell surface. When the compartment opens to the extracellular space, the higher pH of about 7.4 induces release of the Fc back into the blood. In mice, Dall’ Acqua et al. showed that an Fc variant with increased FcRn binding at pH 6 and pH 7.4 actually had reduced serum concentrations and the same half-life as wild-type Fc (Dall’ Acqua et al. 2002, J. Immunol. 169:5171-5180 (incorporated herein by reference in its entirety)). The increased affinity of Fc for FcRn at pH 7.4 is thought to impede release of the Fc back into the blood. Therefore, Fc variants that would increase the in vivo half-life of Fc would ideally increase FcRn binding at lower pH while still allowing release of the Fc at higher pH. The amino acid histidine changes its charge state in the pH range of 6.0-7.4. Therefore, it is not surprising to find His residues at important positions in the Fc / FcRn complex.

[0194] Recently, it has been suggested that antibodies with variable regions having a lower isoelectric point may also have a longer serum half-life (Igawa et al., 2010 PEDS. 23(5):385-392 (incorporated herein by reference in its entirety)). However, this mechanism is still not well understood. Moreover, the variable regions are different for each antibody. Constant region variants with reduced pI and extended half-life would provide a more modular approach to improving the pharmacokinetic properties of the antibodies described herein.

[0195] E. Additional Fc Variants for Additional Functionality In addition to the heterodimerization variants discussed above, for a variety of reasons, including but not limited to modifying binding to one or more FcγR receptors, modified binding to the FcRn receptor, etc., as discussed below, there are a number of useful Fc amino acid modifications that can be made.

[0196] Thus, the antibodies (heterodimers and homodimers) provided herein can include such amino acid modifications, with or without the heterodimerization variants (e.g., pI variants and steric variants) outlined herein. Each set of variants can be independently and optionally included in or excluded from any particular heterodimeric protein.

[0197] 1. FcγR and FcRn Variants Therefore, there are numerous useful Fc substitutions that can be performed to alter the binding of one or more FcγR receptors. In a given embodiment, the target antibody contains modifications (i.e., "FcγR variants") that alter the binding of one or more FcγR receptors. Substitutions resulting in increased and decreased binding may be useful. For example, increased binding to FcγRIIIa is known to typically result in increased ADCC (antibody-dependent cell-mediated cytotoxicity; a cell-mediated reaction in which nonspecific cytotoxic cells expressing FcγR recognize the bound antibody on target cells, subsequently causing lysis of the target cells). Similarly, decreased binding to FcγRIIb (an inhibitory receptor) may be equally beneficial in several situations. The amino acid substitutions used in the target antibodies include those listed in U.S. Patent Nos. 8,188,321 (particularly Figure 41) and 8,084,582, and U.S. Publication Nos. 20060235208 and 20070148170 (all of which are expressly incorporated herein by reference with respect to the variants disclosed therein that affect them in whole and in particular to Fcγ receptor binding). Specific variants used include, but are not limited to, 236A, 239D, 239E, 332E, 332D, 239D / 332E, 267D, 267E, 328F, 267E / 328F, 236A / 332E, 239D / 332E / 330Y, 239D, 332E / 330L, 243A, 243L, 264A, 264V, and 299T. Such modifications may be present in one or both Fc domains of the target antibody.

[0198] In some embodiments, the target antibody includes one or more Fc modifications that increase the serum half-life. These include, but are not limited to, 434S, 434A, 428L, 308F, 259I, 428L / 434S, 259I / 308F, 436I / 428L, 436I or V / 434S, 436V / 428L, 259I / 308F / 428L, and M252Y / S254T / T256E, as specifically disclosed in USSN12 / 341,769 (the entirety of which is incorporated herein by reference), Fc substitutions utilized for increased binding to the FcRn receptor and increased serum half-life. Such modifications may be contained in one or both Fc domains of the target antibody.

[0199] 2. Decline variant In some embodiments, heterodimer antibodies include one or more modifications that reduce or eliminate the normal binding of the Fc domain to one or more or all of the Fcγ receptors (e.g., FcγR1, FcγRIIa, FcγRIIb, FcγRIIIa, etc.) in order to avoid additional mechanisms of action. Such modifications are referred to as “FcγR reduction variants” or “Fc knockout (FcKO or KO)” variants. In these embodiments, for some therapeutic applications, it is desirable to reduce or eliminate the normal binding of the Fc domain to one or more or all of the Fcγ receptors (e.g., FcγR1, FcγRIIa, FcγRIIb, FcγRIIIa, etc.) in order to avoid additional mechanisms of action. That is, for example, in many embodiments, particularly in the use of bispecific antibodies that monovalently bind CD28, it is usually desirable to reduce FcγRIIIa binding in order to eliminate or significantly reduce ADCC activity. In some embodiments of the target antibodies described herein, at least one of the Fc domains contains one or more Fcγ receptor attenuation variants. In some embodiments of the target antibodies described herein, both of the Fc domains contain one or more Fcγ receptor attenuation variants. These attenuation variants are shown in Figure 5, each of which may be included or excluded independently and optionally. A preferred embodiment utilizes attenuation variants selected from the group consisting of L234A / L235A / D265S, G236R / L328R, E233P / L234V / L235A / G236del / S239K, E233P / L234V / L235A / G236del / S267K, E233P / L234V / L235A / G236del / S239K / A327G, E233P / L234V / L235A / G236del / S267K / A327G, and E233P / L234V / L235A / G236del. It should be noted that the diminishing variants referred to herein diminish FcγR binding but do not typically diminish FcRn binding.

[0200] As is known in the art, the Fc domain of human IgG1 has the highest binding to the Fcγ receptor, and therefore the reduced-binding variant can be used when the constant domain (or Fc domain) in the heterodimeric antibody backbone is IgG1. Alternatives to, or in addition to, the reduced-binding variant in the IgG1 background, mutations at glycosylation site 297 (usually to A or S) can significantly reduce binding to FcγRIIIa, for example. Human IgG2 and IgG4 have naturally reduced binding to the Fcγ receptor, and therefore their backbones can be used with or without the reduced-binding variant.

[0201] F. Combination of heterodimer and Fc variant As will be understood by those skilled in the art, all of the listed heterodimerized variants (including skewed and / or purified variants) can be combined in any way, independently and arbitrarily, as long as they retain their "strandiness" or "monomer partitioning." Furthermore, all of these variants can be combined with any of the heterodimerized forms.

[0202] While the most commonly used embodiments for the pI variant are shown in the figure, other combinations can be generated according to fundamental rules for modifying the pI difference between the two monomers to facilitate purification.

[0203] Furthermore, both heterodimerized variants (skewed and purified variants) can be independently and optionally combined with Fc-decaying variants, Fc variants, and FcRn variants, as generally outlined herein.

[0204] Exemplary combinations of variants included in some embodiments of heterodimer antibodies in the form of 1+1 Fab-scFv-Fc, 2+1 mAb-Fc, 2+1 Fab2-scFv-Fc, and 2+1 Fab2-Fc×scFv-Fc are shown in Figure 8. In some embodiments, the heterodimer antibodies include the variant combinations shown in Figure 8.

[0205] G. Useful antibody formats As will be understood by those skilled in the art and will be discussed more thoroughly below, the heterodimer bispecific antibodies provided herein can take on several different configurations, as schematically shown in Figure 14.

[0206] As will be understood by those skilled in the art, the heterodimer form of the present invention may have different valencies and may be bispecific. That is, the heterodimer antibody of the present invention may be bivalent and bispecific, or trivalent and bispecific, where the first antigen is bound by two binding domains and the second antigen is bound by a second binding domain. As outlined herein, when CD28 is one of the target antigens, it is preferable that CD28 binds only in a monovalent state.

[0207] The present invention utilizes a CD28-binding domain in combination with an ENPP3-binding domain. As will be understood by those skilled in the art, any combination of anti-CD28 CDRs, anti-CD28 variable light domains and variable heavy domains, Fab and scFv or their variants shown in any of the figures (see in particular Figures 15-18 and 21) may be used. Similarly, regardless of which of the CDRs, variable light domains and variable heavy domains, Fab and scFv or their variants shown in any of the figures (see, for example, Figure 22) may be used, any of the ENPP3 antigen-binding domains may be used and may be combined in any combination, either independently or arbitrarily.

[0208] 1.1+1 Fab-scFv-Fc format One heterodimer antibody form particularly utilized in the anti-CD28 × anti-ENPP3 antibodies provided herein is the “1+1 Fab-scFv-Fc” or “bottle opener” form, as shown in Figure 14A. The 1+1 Fab-scFv-Fc form antibody contains a first monomer which is a “standard” heavy chain (VH1-CH1-hinge-CH2-CH3), where VH1 is the first variable heavy domain and CH2-CH3 is the first Fc domain. The 1+1 Fab-scFv-Fc also contains a light chain which contains a first variable light domain VL1 and a constant light domain CL. The light chain interacts with VH1-CH1 of the first monomer to form a first antigen-binding domain which is Fab. The second monomer of the antibody contains a second binding domain which is a single-stranded Fv (defined hereafter as “scFv”) and a second Fc domain. The scFv comprises a second variable heavy domain (VH2) and a second variable light domain (VL2), where VH2 is bound to VL2 using a potentially charged scFv linker (see, e.g., Figure 6). The scFv is bound to the heavy chain using a domain linker (see, e.g., Figure 7). The two monomers are brought about by the use of amino acid variants (e.g., heterodimerizing variants discussed above) in constant regions (e.g., Fc domain, CH1 domain, and / or hinge region) that promote the formation of a heterodimer antibody, as will be described more thoroughly below. This structure is roughly visually similar to a bottle opener and is therefore sometimes referred to herein as the “bottle opener” form. In some embodiments, the 1+1 Fab-scFv-Fc form antibody is a bivalent antibody.

[0209] The "1+1 Fab-scFv-Fc" format offers several unique advantages. As is known in the art, antibody analogs that rely on two scFv constructs often have stability and aggregation problems, which can be mitigated in this invention by the addition of "standard" heavy-chain and light-chain pairings. Furthermore, in contrast to formats that rely on two heavy chains and two light chains, there is no problem with incorrect pairing of heavy and light chains (e.g., heavy 1 pairing with light 2).

[0210] In some embodiments of the 1+1 Fab-scFv-Fc antibody, one of the first or second antigen-binding domains is a CD28-binding domain, and the other binding domain is an ENPP3-binding domain. In some embodiments of 1+1 Fab-scFv-Fc, it is an scFv that binds to CD28 and a Fab that binds to ENPP3. An exemplary anti-CD28 × anti-ENPP3 bispecific antibody of the 1+1 Fab-scFv-Fc form is shown in Figure 24.

[0211] In some embodiments, the first and second Fc domains of a 1+1 Fab-scFv-Fc antibody are variant Fc domains containing heterodimerized scuba riants (e.g., the set of amino acid substitutions shown in Figures 3 and 8). Particularly useful heterodimerized scuba riants include S364K / E357Q:L368D / K370S;L368D / K370S:S364K;L368E / K370S:S364K;T411T / E360E / Q362E:D401K;L368D / K370S:S364K / E357L;K370S:S364K / E357Q;T366S / L368A / Y407V:T366W and T366S / L368A / Y407V / Y349C:T366W / S354C (EU numbering)). In an exemplary embodiment, one of the first or second variant Fc domains contains the heterodimerized scuba riant L368D / K370S, and the other of the first or second variant Fc domains contains the heterodimerized scuba riant S364K / E357Q, with numbering following EU numbering.

[0212] In some embodiments, the variant Fc domain includes a de-evolution variant (including the one shown in Figure 5). In some embodiments, each of the first and second variant Fc domains includes a de-evolution variant E233P / L234V / L235A / G236_ / S267K, numbered according to EU numbering.

[0213] In some embodiments, the steady-state domain (CH1-hinge-CH2-CH3) of the first monomer includes pI variants (including those shown in Figures 3 and 4). In exemplary embodiments, the steady-state domain (CH1-hinge-CH2-CH3) of the first monomer includes pI variants N208D / Q295E / N384D / Q418E / N421D, numbered according to EU numbering.

[0214] In exemplary embodiments, the 1+1 Fab-scFv-Fc form antibody contains the amino acid modification combination shown in Figure 8. In such embodiments, the CH1-hinge-CH2-CH3 of the first monomer contains the amino acid variant L368D / K370S / N208D / Q295E / N384D / Q418E / N421D / E233P / L234V / L235A / G236del / S267K, and the second Fc domain contains the amino acid variant S364K / E357Q / E233P / L234V / L235A / G236del / S267K, with the numbering following EU numbering.

[0215] In some embodiments, the scFv of the 1+1 Fab-scFv-Fc antibody provided herein includes a charged scFv linker (including the one shown in Figure 6). In some embodiments, the 1+1 Fab-scFv-Fc antibody provided herein includes the FcRn variant M428L / N434S, which is numbered according to EU numbering.

[0216] In the exemplary 1+1 Fab-scFv-Fc format antibody of the embodiments, the first Fc domain contains the heterodimerization variant L368D / K370S, the second Fc domain contains the heterodimerization variant S364K / E357Q, each of the first and second Fc domains contains the hyporesponsive variant E233P / L234V / L235A / G236_ / S267K, and the constant domain (CH1-hinge-CH2-CH3) of the first monomer contains the pI variant N208D / Q295E / N384D / Q418E / N421D, with the numbering following the EU numbering. In some embodiments, the scFv of the 1+1 Fab-scFv-Fc format antibody provided herein contains a (GKPGS)4 charged scFv linker. In some embodiments, the 1+1 Fab-scFv-Fc format antibody provided herein contains the FcRn variant M428L / N434S, with the numbering following the EU numbering. In some embodiments, the scFv of the 1+1 Fab-scFv-Fc format antibody provided herein contains a charged scFv linker (including those shown in Figure 6).

[0217] The target 1+1 Fab-scFv-Fc antibody may contain any suitable CD28-binding domain, including any of the CD28-binding domains provided herein. In some embodiments, the CD28-binding domain is one of the following CD28-binding domains or a variant thereof: 1A7[CD28]_H1L1, 1A7[CD28]_H1.1_L1, 1A7[CD28]_H1_L1.71, 1A7[CD28]_H1.1_L1.71, 1A7[CD28]_H1.14_L1, 1A7[CD28]_H1.14_L1.71, CD28.3[CD28]_H0L0, hCD28.3[CD28]_H1L1, 5.11A1[CD28]_H 0L0, TGN1412_H1L1, 341VL34[CD28]_H1L1, 341VL36[CD28]_H1L1, 281VL4[CD28]_H1L1, HuTN228[CD28]_H1L1, PV1[CD28]_H0L0, m9.3[CD28]_H0L0, hu9.3[CD28]_H1L1, 9G2[CD28]_H0L0, 9G2[CD28]_H1L1, 2F10A3.140[CD28]_H1L1, and TN228[CD28]_H4L2 (Figures 15, 18, and 21). Exemplary additional VH and VL sequences of the CD28-binding domain that can be used with the target 1+1 Fab-scFv-Fc type antibody are shown in Figures 15, 16, and 17.

[0218] In some embodiments of the 1+1 Fab-scFv-Fc format, the anti-CD28 ABD has VH and VL domains selected from the following:

[0219] (i) VH containing vhCDR1, vhCDR2, and vhCDR3, each having the amino acid sequences of vhCDR1, vhCDR2, and vhCDR3 of any VH or variant thereof of the CD28 binding domain shown in Figures 15, 18, and 21; and (ii) VL containing vlCDR1, vlCDR2, and vlCDR3, each having the amino acid sequences of vlCDR1, vlCDR2, and vlCDR3 of any VL or variant thereof of the CD28 binding domain shown in Figures 15, 18, and 21; or

[0220] (i) a VH containing vhCDR1, vhCDR2, and vhCDR3, each having the amino acid sequences of vhCDR1, vhCDR2, and vhCDR3 of the VH or its variant shown in Figure 15 or 16, and (ii) a VL containing vlCDR1, vlCDR2, and vlCDR3, each having the amino acid sequences of vlCDR1, vlCDR2, and vlCDR3 of the VL or its variant shown in Figure 15 or 17.

[0221] In some embodiments of the 1+1 Fab-scFv-Fc format, the anti-CD28 ABD has VH and VL domains selected from the following:

[0222] (i) VH having the amino acid sequence of any VH or a variant thereof of the CD28 binding domain shown in Figures 15, 18, and 21, and (ii) VL having the amino acid sequence of any VL or a variant thereof of the CD28 binding domain shown in Figures 15, 18, and 21; or

[0223] (i) VH having the amino acid sequence of VH or a variant thereof shown in Figure 15 or 16, and (ii) VL having the amino acid sequence of VL or a variant thereof shown in Figure 15 or 17.

[0224] The target 1+1 Fab-scFv-Fc antibody may contain any suitable ENPP3-binding domain, including any of the ENPP3-binding domains provided herein. Exemplary ENPP3-binding domains that can be used in the target 1+1 Fab-scFv-Fc antibody are shown in Figure 22.

[0225] In some embodiments of the 1+1 Fab-scFv-Fc form, the ENPP3 ABD comprises (i) a VH having the amino acid sequence of any one of the ENPP3 binding domains shown in Figure 22 or a variant thereof, and (ii) a VL having the amino acid sequence of the ENPP3 binding domain or a variant thereof, as shown in Figure 22.

[0226] Figure 9 shows several exemplary Fc domain sequences useful in 1+1 Fab-scFv-Fc type antibodies. The “monomer 1” sequence shown in Figure 9 typically refers to the Fc domain of the “Fab-Fc heavy chain,” and the “monomer 2” sequence refers to the Fc domain of the “scFv-Fc heavy chain.” Figures 11 and 12 also provide exemplary CH1-hinge domains, CH1 domains, and hinge domains that may be included in the first or second monomer of the 1+1 Fab-scFv-Fc form. Furthermore, Figure 13 provides useful CL sequences that can be used in this form.

[0227] 2.2+1 mAb-scFv format One heterodimeric antibody form particularly used in the target bispecific anti-CD28 × anti-ENPP3 antibody is the 2+1 mAb-scFv form shown in Figure 14E. This antibody form contains three antigen-binding domains, namely two Fab moieties and an scFv bound to one C-terminus of the heavy chain. In some embodiments of this form, each Fab moiety binds to ENPP3 (in this case, human ENPP3), and the "additional" scFv domain binds to CD28. In other words, this mAb-scFv form is a trivalent antibody.

[0228] In these embodiments, the first chain or monomer comprises VH1-CH1-hinge-CH2-CH3 from N to the C-terminus, and the second monomer comprises VH1-CH1-hinge-CH2-CH3-domain linker-scFv domain from N to the C-terminus, where the scFv domain comprises a second VH(VH2), a second VL(VL2), and an scFv linker. For all scFv domains in this specification, the scFv domain may be VH2-scFv linker-VL2 or VL2-scFv linker-VH2 in either orientation from N to the C-terminus. Therefore, the second monomer may contain, from N to C-terminus, VH1-CH1-hinge-CH2-CH3-domain linker-VH2-scFv linker-VL2 or VH1-CH1-hinge-CH2-CH3-domain linker-VL2-scFv linker-VH2. The composition also contains a light chain VL1-CL. In embodiments, this form contains two identical light chains (VL1-CL). In this form, VH1 is a first variable weight (VH) domain, VL1 is a first variable light (VL) domain, VH2 is a second variable weight domain, and VL2 is a second variable light domain. In some embodiments, the form comprises two identical common light chains (VL1-CL, each), each of which associates with one of the first and second monomer VH1-CH1 to form two identical Fabs (i.e., the first antigen-binding domain). In some embodiments, scFv is the second antigen-binding domain. In some embodiments, the first ABD binds to human ENPP3, and the second ABD binds to human CD28.

[0229] In some embodiments, the first and second Fc domains of a 2+1 mAb-scFv antibody are variant Fc domains containing heterodimerized scuba riants (e.g., the set of amino acid substitutions shown in Figures 3 and 8). Particularly useful heterodimerized scuba riants include S364K / E357Q:L368D / K370S;L368D / K370S:S364K;L368E / K370S:S364K;T411T / E360E / Q362E:D401K;L368D / K370S:S364K / E357L;K370S:S364K / E357Q;T366S / L368A / Y407V:T366W and T366S / L368A / Y407V / Y349C:T366W / S354C (EU numbering)). In an exemplary embodiment, one of the first or second variant Fc domains contains the heterodimerized scuba riant L368D / K370S, and the other of the first or second variant Fc domains contains the heterodimerized scuba riant S364K / E357Q, with numbering following EU numbering.

[0230] In some embodiments, the variant Fc domain includes a de-evolution variant (including the one shown in Figure 5). In some embodiments, each of the first and second variant Fc domains includes a de-evolution variant E233P / L234V / L235A / G236_ / S267K, numbered according to EU numbering.

[0231] In some embodiments, the steady-state domain (CH1-hinge-CH2-CH3) of the first monomer includes pI variants (including those shown in Figures 3 and 4). In exemplary embodiments, the steady-state domain (CH1-hinge-CH2-CH3) of the first monomer includes pI variants N208D / Q295E / N384D / Q418E / N421D, numbered according to EU numbering.

[0232] In some embodiments, the scFv of the 2+1 mAb-scFv antibody provided herein includes a charged scFv linker (including the one shown in Figure 6). In some embodiments, the 2+1 mAb-scFv antibody provided herein includes the FcRn variant M428L / N434S, which is numbered according to EU numbering.

[0233] In exemplary embodiments, the 2+1 mAb-scFv form antibody includes the amino acid modification combinations shown in Figure 8. In such embodiments, the first variant Fc domain contains the heterodimerized scuba rianto L368D / K370S, the second variant Fc domain contains the heterodimerized scuba rianto S364K / E357Q, each of the first and second variant Fc domains contains the diminished variant E233P / L234V / L235A / G236_ / S267K, and the constant domain (CH1-hinge-CH2-CH3) of the first monomer contains the pI variant N208D / Q295E / N384D / Q418E / N421D, with numbering following EU numbering. In some embodiments, the scFv of the 2+1 mAb-scFv antibody provided herein includes a (GKPGS) 4-charged scFv linker. In some embodiments, the 2+1 mAb-scFv antibody provided herein includes the FcRn variant M428L / N434S, with numbering according to EU numbering.

[0234] In some embodiments, the scFv of the second monomer of a 2+1 Fab2-scFv-Fc antibody is CD28-conjugated, and the VH1 of the first and second monomers and the VL1 of the common light chain each form a binding domain that binds to ENPP3. The 2+1 mAb-scFv antibody may include any preferred ENPP3-binding domain and CD28-binding domain, including the ENPP3-binding domain and CD28-binding domain provided herein, as well as any of the associated VH and VL or their variants (see, for example, Figures 15-18 and 21).

[0235] In some embodiments of the 2+1 mAb-scFv format, the anti-CD28 ABD has VH and VL domains selected from the following:

[0236] (i) VH containing vhCDR1, vhCDR2, and vhCDR3, each having the amino acid sequences of vhCDR1, vhCDR2, and vhCDR3 of any VH or variant thereof of the CD28 binding domain shown in Figures 15, 18, and 21; and (ii) VL containing vlCDR1, vlCDR2, and vlCDR3, each having the amino acid sequences of vlCDR1, vlCDR2, and vlCDR3 of any VL or variant thereof of the CD28 binding domain shown in Figures 15, 18, and 21; or

[0237] (i) a VH containing vhCDR1, vhCDR2, and vhCDR3, each having the amino acid sequences of vhCDR1, vhCDR2, and vhCDR3 of the VH or its variant shown in Figure 15 or 16, and (ii) a VL containing vlCDR1, vlCDR2, and vlCDR3, each having the amino acid sequences of vlCDR1, vlCDR2, and vlCDR3 of the VL or its variant shown in Figure 15 or 17.

[0238] In some embodiments of the 2+1 mAb-scFv format, the anti-CD28 ABD has VH and VL domains selected from the following:

[0239] (i) VH having the amino acid sequence of any VH or a variant thereof of the CD28 binding domain shown in Figures 15, 18, and 21, and (ii) VL having the amino acid sequence of any VL or a variant thereof of the CD28 binding domain shown in Figures 15, 18, and 21; or

[0240] (i) VH having the amino acid sequence of VH or a variant thereof shown in Figure 15 or 16, and (ii) VL having the amino acid sequence of VL or a variant thereof shown in Figure 15 or 17.

[0241] In some embodiments of the 2+1 mAb-scFv format, each ENPP3 ABD has (i) a VH having the amino acid sequence of any one of the ENPP3 binding domains shown in Figure 22 or a variant thereof, and (ii) a VL having the amino acid sequence of the ENPP3 binding domain or a variant thereof, as shown in Figure 22.

[0242] Figure 10 shows several exemplary Fc domain sequences useful in 2+1 mAb-scFv format antibodies. Figures 11 and 12 also provide exemplary CH1-hinge domains, CH1 domains, and hinge domains that may be included in the first or second monomer of the 2+1 mAb-scFv format. Furthermore, Figure 13 provides useful CL sequences that may be used in this format.

[0243] Figure 41 shows an exemplary sequence of ENPP3×CD28 bsAb in the 2+1 mAb-scFv format.

[0244] 3.2+1 Fab2-scFv-Fc format One heterodimeric antibody form particularly utilized in the anti-CD28 × anti-ENPP3 antibodies provided herein is the 2+1 Fab2-scFv-Fc form (also referred to as the “central-scFv form”), shown in Figure 14B. This antibody form includes three antigen-binding domains: two Fab moieties and an scFv inserted between the VH-CH1 and CH2-CH3 regions of one of the monomers. In some embodiments of this form, each Fab moiety binds to ENPP3, and the “additional” scFv domain binds to CD28. In some embodiments, the 2+1 Fab2-scFv-Fc form antibody is a trivalent antibody.

[0245] In some embodiments of the 2+1 Fab2-scFv-Fc form, the first monomer includes a standard heavy chain (i.e., VH1-CH1-hinge-CH2-CH3), where VH1 is the first variable weight domain and CH2-CH3 is the first Fc domain. The second monomer includes another first variable weight domain (VH1), a CH1 domain (and an optional hinge), a second Fc domain, and an scFv containing an scFv variable light domain (VL2), an scFv linker, and an scFv variable weight domain (VH2). The scFv is covalently linked between the C-terminus of the CH1 domain of the second monomer and the N-terminus of the second Fc domain using an arbitrary domain linker (VH1-CH1-[arbitrary linker]-VH2-scFv linker-VH2-[arbitrary linker]-CH2-CH3, or the opposite orientation for scFv: VH1-CH1-[arbitrary linker]-VL2-scFv linker-VH2-[arbitrary linker]-CH2-CH3). The arbitrary linker can be any suitable peptide linker, including, for example, the domain linker shown in Figure 7. This embodiment further utilizes a common light chain comprising a variable light domain (VL1) and a constant light domain (CL). In some embodiments, the form comprises two identical common light chains (VL1-CL, respectively), each of which associates with one of the VH1-CH1 domains of the first and second monomers to form two identical Fabs. In some embodiments, each identical Fab is coupled to ENPP3, and scFv is coupled to CD28. As with many of the embodiments herein, these constructs may include, as desired and as described herein, scuba riants, pI variants, attenuation variants, additional Fc variants, etc.

[0246] In some embodiments, the second and second Fc domains of the 2+1 Fab2-scFv-Fc antibody are variant Fc domains containing heterodimerized scuba riants (e.g., the set of amino acid substitutions shown in Figures 3 and 8). Particularly useful heterodimerized scuba riants include S364K / E357Q:L368D / K370S;L368D / K370S:S364K;L368E / K370S:S364K;T411T / E360E / Q362E:D401K;L368D / K370S:S364K / E357L;K370S:S364K / E357Q;T366S / L368A / Y407V:T366W and T366S / L368A / Y407V / Y349C:T366W / S354C (EU numbering)). In an exemplary embodiment, one of the first or second variant Fc domains contains the heterodimerized scuba riant L368D / K370S, and the other of the first or second variant Fc domains contains the heterodimerized scuba riant S364K / E357Q, with numbering following EU numbering.

[0247] In some embodiments, the variant Fc domain includes a de-evolution variant (including the one shown in Figure 5). In some embodiments, each of the first and second variant Fc domains includes a de-evolution variant E233P / L234V / L235A / G236_ / S267K, numbered according to EU numbering.

[0248] In some embodiments, the steady-state domain (CH1-hinge-CH2-CH3) of the first monomer includes pI variants (including those shown in Figures 3 and 4). In exemplary embodiments, the steady-state domain (CH1-hinge-CH2-CH3) of the first monomer includes pI variants N208D / Q295E / N384D / Q418E / N421D, numbered according to EU numbering.

[0249] In some embodiments, the scFv of the 2+1 Fab2-scFv-Fc antibody provided herein includes a charged scFv linker (including the one shown in Figure 6). In some embodiments, the 2+1 Fab2-scFv-Fc antibody provided herein includes the FcRn variant M428L / N434S, which is numbered according to EU numbering.

[0250] In exemplary embodiments, the 2+1 Fab2-scFv-Fc form antibody includes the amino acid modification combinations shown in Figure 8. In such embodiments, the first variant Fc domain contains the heterodimerized scuba rianto L368D / K370S, the second variant Fc domain contains the heterodimerized scuba rianto S364K / E357Q, each of the first and second variant Fc domains contains the diminished variant E233P / L234V / L235A / G236_ / S267K, and the constant domain (CH1-hinge-CH2-CH3) of the first monomer contains the pI variant N208D / Q295E / N384D / Q418E / N421D, with numbering following EU numbering. In some embodiments, the scFv of the 2+1 Fab2-scFv-Fc antibody provided herein includes a (GKPGS)4-charged scFv linker. In some embodiments, the 2+1 Fab2-scFv-Fc antibody provided herein includes the FcRn variant M428L / N434S, with numbering according to EU numbering.

[0251] In some embodiments, the CH1-hinge-CH2-CH3 of the first monomer contains the amino acid variant L368D / K370S / N208D / Q295E / N384D / Q418E / N421D / E233P / L234V / L235A / G236del / S267K, and the second Fc domain contains the amino acid variant S364K / E357Q / E233P / L234V / L235A / G236del / S267K, with the numbering following EU numbering.

[0252] In some embodiments, the scFv of the second monomer of the 2+1 Fab2-scFv-Fc antibody is CD28-binding, and the VH1 of the first and second monomers and the VL1 of the common light chain each form a binding domain that binds to ENPP3. The 2+1 Fab2-scFv-Fc antibody of interest may contain any preferred CD28-binding domain, including any of the CD28-binding domains provided herein. In some embodiments, the CD28 binding domain is one of the following CD28 binding domains or a variant thereof: 1A7[CD28]_H1L1, 1A7[CD28]_H1.1_L1, 1A7[CD28]_H1_L1.71, 1A7[CD28]_H1.1_L1.71, 1A7[CD28]_H1.14_L1, 1A7[CD28]_H1.14_L1.71, CD28.3[CD28]_H0L0, hCD28.3[CD28]_H1L1, 5.11A1[CD28]_H 0L0, TGN1412_H1L1, 341VL34[CD28]_H1L1, 341VL36[CD28]_H1L1, 281VL4[CD28]_H1L1, HuTN228[CD28]_H1L1, PV1[CD28]_H0L0, m9.3[CD28]_H0L0, hu9.3[CD28]_H1L1, 9G2[CD28]_H0L0, 9G2[CD28]_H1L1, 2F10A3.140[CD28]_H1L1, and TN228[CD28]_H4L2 (Figures 15, 18, and 21).

[0253] In some embodiments of the 2+1 Fab2-scFv-Fc format, the anti-CD28 ABD has VH and VL domains selected from the following:

[0254] (i) VH containing vhCDR1, vhCDR2, and vhCDR3, each having the amino acid sequences of vhCDR1, vhCDR2, and vhCDR3 of any VH or variant thereof of the CD28 binding domain shown in Figures 15, 18, and 21; and (ii) VL containing vlCDR1, vlCDR2, and vlCDR3, each having the amino acid sequences of vlCDR1, vlCDR2, and vlCDR3 of any VL or variant thereof of the CD28 binding domain shown in Figures 15, 18, and 21; or

[0255] (i) a VH containing vhCDR1, vhCDR2, and vhCDR3, each having the amino acid sequences of vhCDR1, vhCDR2, and vhCDR3 of the VH or its variant shown in Figure 15 or 16, and (ii) a VL containing vlCDR1, vlCDR2, and vlCDR3, each having the amino acid sequences of vlCDR1, vlCDR2, and vlCDR3 of the VL or its variant shown in Figure 15 or 17.

[0256] In some embodiments of the 2+1 Fab2-scFv-Fc format, the anti-CD28 ABD has VH and VL domains selected from the following:

[0257] (i) VH having the amino acid sequence of any VH or a variant thereof of the CD28 binding domain shown in Figures 15, 18, and 21, and (ii) VL having the amino acid sequence of any VL or a variant thereof of the CD28 binding domain shown in Figures 15, 18, and 21; or

[0258] (i) VH having the amino acid sequence of VH or a variant thereof shown in Figure 15 or 16, and (ii) VL having the amino acid sequence of VL or a variant thereof shown in Figure 15 or 17.

[0259] In some embodiments, the VH1 of the first and second monomers and the VL1 of the common light chain of the 2+1 Fab2-scFv-Fc antibody each form a binding domain that binds to ENPP3. The 2+1 Fab2-scFv-Fc antibody of interest may include any suitable ENPP3 binding domain, including any of the ENPP3 binding domains provided herein. Exemplary ENPP3 binding domains that can be used in the 2+1 Fab2-scFv-Fc antibody of interest are shown in Figure 22.

[0260] In some embodiments of the 2+1 Fab2-scFv-Fc form, each ENPP3 ABD comprises (i) a VH having the amino acid sequence of any one of the VH or a variant of the ENPP3 binding domain shown in Figure 22, and (ii) a VL having the amino acid sequence of the VL or a variant of the ENPP3 binding domain shown in Figure 22.

[0261] Figure 9 shows several exemplary Fc domain sequences useful in 2+1 Fab2-scFv-Fc type antibodies. Figures 11 and 12 also provide exemplary CH1-hinge domains, CH1 domains, and hinge domains that may be included in the second or third monomer of the 2+1 Fab2-scFv-Fc form. Furthermore, Figure 13 provides useful CL sequences that may be used in this form.

[0262] Figure 42 shows an exemplary sequence of ENPP3×CD28 bsAb in the 2+1 Fab2-scFv-Fc format.

[0263] 4.2+1 Fab2-Fc×scFv-Fc format One heterodimer antibody configuration particularly utilized in the target anti-CD28 × anti-ENPP3 antibodies provided herein is the 2+1 Fab2-Fc × scFv-Fc configuration (also known as the “stacked bottle opener”) shown in Figure 14F. This configuration comprises a first monomer, a second monomer, and a common light chain. The first monomer contains VH1-CH1-domain linker-VH1-CH1-hinge-CH2-CH3 from the N-terminus to the C-terminus, where VH1 is the first variable weight domain and CH2-CH3 is the first variant Fc domain. The domain linker can be any useful domain linker (see, for example, Figure 7). The second monomer contains a single-stranded Fv ("scFv") covalently linked to the second variant Fc domain by the domain linker (scFv-domain linker-CH2-CH3). The common light chain comprises VL1-CL, where VL1 is the first variable light domain. The second monomer scFv comprises a second variable heavy domain (VH2) linked to the second variable light domain (VL2) by an scFv linker. In some embodiments, this form comprises two identical common light chains (VL1-CL). In this embodiment, the two VH1-CH1 of the first monomer each interact with the common light chain to form two identical Fabs, which are the first antigen-binding domains, and VH2 and VL2 form the second antigen-binding domain. In some embodiments, each of the first antigen-binding domains is an ENPP3-binding domain, and the second antigen-binding domain is a CD28-binding domain. As with many of the embodiments herein, these constructs may include scuba riants, pI variants, attenuation variants, additional Fc variants, etc., as desired and as described herein.

[0264] In some embodiments, the second and second Fc domains of the 2+1 Fab2-Fc×scFv-Fc antibody are variant Fc domains containing heterodimerized scuba riants (e.g., the set of amino acid substitutions shown in Figures 3 and 8). Particularly useful heterodimerized scuba riants include S364K / E357Q:L368D / K370S;L368D / K370S:S364K;L368E / K370S:S364K;T411T / E360E / Q362E:D401K;L368D / K370S:S364K / E357L;K370S:S364K / E357Q;T366S / L368A / Y407V:T366W and T366S / L368A / Y407V / Y349C:T366W / S354C (EU numbering)). In an exemplary embodiment, one of the first or second variant Fc domains contains the heterodimerized scuba riant L368D / K370S, and the other of the first or second variant Fc domains contains the heterodimerized scuba riant S364K / E357Q, with numbering following EU numbering.

[0265] In some embodiments, the variant Fc domain includes a de-evolution variant (including the one shown in Figure 5). In some embodiments, each of the first and second variant Fc domains includes a de-evolution variant E233P / L234V / L235A / G236_ / S267K, numbered according to EU numbering.

[0266] In some embodiments, the steady-state domain (CH1-hinge-CH2-CH3) of the first monomer includes pI variants (including those shown in Figures 3 and 4). In exemplary embodiments, the steady-state domain (CH1-hinge-CH2-CH3) of the first monomer includes pI variants N208D / Q295E / N384D / Q418E / N421D, numbered according to EU numbering.

[0267] In some embodiments, the 2+1 Fab2-Fc×scFv-Fc antibody provided herein includes a charged scFv linker (including the one shown in Figure 6). In some embodiments, the 2+1 Fab2-scFv-Fc antibody provided herein includes the FcRn variant M428L / N434S, which is numbered according to EU numbering.

[0268] In exemplary embodiments, the 2+1 Fab2-Fc×scFv-Fc antibody includes the amino acid modification combinations shown in Figure 8. In such embodiments, the first variant Fc domain contains the heterodimerized scuba rianto L368D / K370S, the second variant Fc domain contains the heterodimerized scuba rianto S364K / E357Q, each of the first and second variant Fc domains contains the diminished variant E233P / L234V / L235A / G236_ / S267K, and the constant domain (CH1-hinge-CH2-CH3) of the first monomer contains the pI variant N208D / Q295E / N384D / Q418E / N421D, with numbering following EU numbering. In some embodiments, the scFv of the 2+1 Fab2-scFv-Fc antibody provided herein includes a (GKPGS) 4-charged scFv linker (SEQ ID NO: 24). In some embodiments, the 2+1 Fab2-scFv-Fc antibody provided herein includes the FcRn variant M428L / N434S, which is numbered according to EU numbering.

[0269] In some embodiments, the CH1-hinge-CH2-CH3 of the first monomer contains the amino acid variant L368D / K370S / N208D / Q295E / N384D / Q418E / N421D / E233P / L234V / L235A / G236del / S267K, and the second Fc domain contains the amino acid variant S364K / E357Q / E233P / L234V / L235A / G236del / S267K, with the numbering following EU numbering.

[0270] In some embodiments, the scFv of the second monomer of the 2+1 Fab2-Fc×scFv-Fc antibody is CD28-binding, and the VH1 of the first and second monomers and the VL1 of the common light chain each form a binding domain that binds to ENPP3. The 2+1 Fab2-scFv-Fc antibody of interest may contain any preferred CD28-binding domain, including any of the CD28-binding domains provided herein. In some embodiments, the CD28 binding domain is one of the following CD28 binding domains or a variant thereof: 1A7[CD28]_H1L1, 1A7[CD28]_H1.1_L1, 1A7[CD28]_H1_L1.71, 1A7[CD28]_H1.1_L1.71, 1A7[CD28]_H1.14_L1, 1A7[CD28]_H1.14_L1.71, CD28.3[CD28]_H0L0, hCD28.3[CD28]_H1L1, 5.11A1[CD28]_H 0L0, TGN1412_H1L1, 341VL34[CD28]_H1L1, 341VL36[CD28]_H1L1, 281VL4[CD28]_H1L1, HuTN228[CD28]_H1L1, PV1[CD28]_H0L0, m9.3[CD28]_H0L0, hu9.3[CD28]_H1L1, 9G2[CD28]_H0L0, 9G2[CD28]_H1L1, 2F10A3.140[CD28]_H1L1, and TN228[CD28]_H4L2 (Figures 15, 18, and 21).

[0271] In some embodiments of the 2+1 Fab2-Fc×scFv-Fc format, the anti-CD28 ABD has VH and VL domains selected from the following:

[0272] (i) VH containing vhCDR1, vhCDR2, and vhCDR3, each having the amino acid sequences of vhCDR1, vhCDR2, and vhCDR3 of any VH or variant thereof of the CD28 binding domain shown in Figures 15, 18, and 21; and (ii) VL containing vlCDR1, vlCDR2, and vlCDR3, each having the amino acid sequences of vlCDR1, vlCDR2, and vlCDR3 of any VL or variant thereof of the CD28 binding domain shown in Figures 15, 18, and 21; or

[0273] (i) a VH containing vhCDR1, vhCDR2, and vhCDR3, each having the amino acid sequences of vhCDR1, vhCDR2, and vhCDR3 of the VH or its variant shown in Figure 15 or 16, and (ii) a VL containing vlCDR1, vlCDR2, and vlCDR3, each having the amino acid sequences of vlCDR1, vlCDR2, and vlCDR3 of the VL or its variant shown in Figure 15 or 17.

[0274] In some embodiments of 2+1 Fab2-Fc×scFv-Fc, the anti-CD28 ABD has VH and VL domains selected from the following:

[0275] (i) VH having the amino acid sequence of any VH or a variant thereof of the CD28 binding domain shown in Figures 15, 18, and 21, and (ii) VL having the amino acid sequence of any VL or a variant thereof of the CD28 binding domain shown in Figures 15, 18, and 21; or

[0276] (i) VH having the amino acid sequence of VH or a variant thereof shown in Figure 15 or 16, and (ii) VL having the amino acid sequence of VL or a variant thereof shown in Figure 15 or 17.

[0277] In some embodiments, the VH1 of the first and second monomers and the VL1 of the common light chain of the 2+1 Fab2-Fc×scFv-Fc antibody each form a binding domain that binds to ENPP3. The 2+1 Fab2-Fc×scFv-Fc antibody of interest may include any suitable ENPP3 binding domain, including any of the ENPP3 binding domains provided herein. Exemplary ENPP3 binding domains that can be used in the 2+1 Fab2-Fc×scFv-Fc antibody of interest are shown in Figure 22.

[0278] In some embodiments of the 2+1 Fab2-Fc×scFv-Fc form, each ENPP3 ABD contains a VH having the amino acid sequence of one of the ENPP3 binding domains or a variant thereof, as shown in Figure 22.

[0279] Figure 9 shows several exemplary Fc domain sequences useful in 2+1 Fab2-Fc×scFv-Fc format antibodies. Figures 11 and 12 also provide exemplary CH1-hinge domains, CH1 domains, and hinge domains that may be included in the second or third monomer of the 2+1 Fab2-Fc×scFv-Fc format. Furthermore, Figure 13 provides useful CL sequences that may be used in this format.

[0280] 5.1+1 CLC format One heterodimer antibody form particularly utilized in the anti-CD28 × anti-ENPP3 antibodies provided herein is the "1+1 common light chain" or "1+1 CLC" form, shown in Figure 14C. The 1+1 CLC form antibody comprises a first monomer containing VH1-CH1-hinge-CH2-CH3 (where VH1 is the first variable weight domain and CH2-CH3 is the first Fc domain), a second monomer containing VH2-CH1-hinge-CH2-CH3 (where VH2 is the second variable weight domain and CH2-CH3 is the second Fc domain), and a third monomer "common light chain" containing VL-CL (where VL is the common variable light domain and CL is the constant light domain). In such embodiments, VL pairs with VH1 to form a first binding domain having a first antigen-binding specificity, and VL pairs with VH2 to form a second binding domain having a second antigen-binding specificity. In some embodiments, the 1+1 CLC-type antibody is a bivalent antibody.

[0281] In some embodiments, the first and second Fc domains of the 1+1 CLC form are variant Fc domains containing heterodimerized scuba riants (e.g., the set of amino acid substitutions shown in Figures 3 and 8). Particularly useful heterodimerized scuba riants include S364K / E357Q:L368D / K370S;L368D / K370S:S364K;L368E / K370S:S364K;T411T / E360E / Q362E:D401K;L368D / K370S:S364K / E357L;K370S:S364K / E357Q;T366S / L368A / Y407V:T366W and T366S / L368A / Y407V / Y349C:T366W / S354C (EU numbering)). In an exemplary embodiment, one of the first or second variant Fc domains contains the heterodimerized scuba riant L368D / K370S, and the other of the first or second variant Fc domains contains the heterodimerized scuba riant S364K / E357Q, with numbering following EU numbering.

[0282] In some embodiments, the variant Fc domain includes a de-evolution variant (including the one shown in Figure 5). In some embodiments, each of the first and second variant Fc domains includes a de-evolution variant E233P / L234V / L235A / G236_ / S267K, numbered according to EU numbering.

[0283] In some embodiments, the steady-state domain (CH1-hinge-CH2-CH3) of the first or second monomer includes a pI variant (including those shown in Figure 4). In exemplary embodiments, the steady-state domain (CH1-hinge-CH2-CH3) of the first or second monomer includes pI variants N208D / Q295E / N384D / Q418E / N421D, numbered according to EU numbering.

[0284] In some embodiments, the 1+1 CLC-type antibodies provided herein include the FcRn variant M428L / N434S, which is numbered according to EU numbering.

[0285] In exemplary embodiments, the first variant Fc domain contains the heterodimerized scuba rianto L368D / K370S, the second variant Fc domain contains the heterodimerized scuba rianto S364K / E357Q, each of the first and second variant Fc domains contains the diminished variant E233P / L234V / L235A / G236_ / S267K, and the steady-state domain (CH1-hinge-CH2-CH3) of the first monomer contains the pI variant N208D / Q295E / N384D / Q418E / N421D, with numbering following EU numbering.

[0286] In some embodiments, the CH1-hinge-CH2-CH3 of the first monomer contains the amino acid variant L368D / K370S / N208D / Q295E / N384D / Q418E / N421D / E233P / L234V / L235A / G236del / S267K, and the second Fc domain contains the amino acid variant S364K / E357Q / E233P / L234V / L235A / G236del / S267K, with the numbering following EU numbering.

[0287] In some embodiments, the 1+1 CLC-type antibody provided herein further comprises the FcRn variant M428L / N434S, the numbering of which follows EU numbering.

[0288] In some embodiments, one of the first or second binding domains binds to CD28, and the other binding domain binds to ENPP3. The 1+1 CLC-type antibody of interest may contain any preferred CD28-binding domain and ENPP3-binding domain, including either or a variant thereof provided herein (see, for example, Figures 15-18, 21, and 22).

[0289] 6.2+1 CLC format Another heterodimer antibody form particularly utilized in the CD28× anti-ENPP3 antibodies provided herein is the "2+1 common light chain" or "2+1 CLC" form, shown in Figure 14D. The 2+1 CLC form comprises a first monomer containing VH1-CH1-linker-VH1-CH1-hinge-CH2-CH3 (where VH1 is the first variable weight domain and CH2-CH3 is the first Fc domain), a second monomer containing VH2-CH1-hinge-CH2-CH3 (where VH2 is the second variable weight domain and CH2-CH3 is the second Fc domain), and a third monomer containing a "common light chain" VL-CL (where VL is the common variable light domain and CL is the constant light domain). VL pairs with each of the VH1 molecules of the first monomer to form two first binding domains, each having a first antigen-binding specificity, and VL pairs with VH2 to form a second binding domain having a second antigen-binding specificity. The linker of the first monomer can be any preferred linker, including any one or a combination thereof of the domain linkers shown in Figure 7. In some embodiments, the 2+1 CLC-type antibody is a trivalent antibody.

[0290] In some embodiments, the second and second Fc domains of the 2+1 CLC form are variant Fc domains containing heterodimerized scuba riants (e.g., the set of amino acid substitutions shown in Figures 3 and 8). Particularly useful heterodimerized scuba riants include S364K / E357Q:L368D / K370S;L368D / K370S:S364K;L368E / K370S:S364K;T411T / E360E / Q362E:D401K;L368D / K370S:S364K / E357L;K370S:S364K / E357Q;T366S / L368A / Y407V:T366W and T366S / L368A / Y407V / Y349C:T366W / S354C (EU numbering)). In an exemplary embodiment, one of the first or second variant Fc domains contains the heterodimerized scuba riant L368D / K370S, and the other of the first or second variant Fc domains contains the heterodimerized scuba riant S364K / E357Q, with numbering following EU numbering.

[0291] In some embodiments, the variant Fc domain includes a de-evolution variant (including the one shown in Figure 5). In some embodiments, each of the first and second variant Fc domains includes a de-evolution variant E233P / L234V / L235A / G236_ / S267K, numbered according to EU numbering.

[0292] In some embodiments, the steady-state domain (CH1-hinge-CH2-CH3) of the first or second monomer includes a pI variant (including those shown in Figure 4). In exemplary embodiments, the steady-state domain (CH1-hinge-CH2-CH3) of the first or second monomer includes pI variants N208D / Q295E / N384D / Q418E / N421D, numbered according to EU numbering.

[0293] In some embodiments, the 2+1 CLC-type antibodies provided herein further include the FcRn variant M428L / N434S, which is numbered according to EU numbering.

[0294] In exemplary embodiments, the first variant Fc domain contains the heterodimerized scuba rianto L368D / K370S, the second variant Fc domain contains the heterodimerized scuba rianto S364K / E357Q, each of the first and second variant Fc domains contains the diminished variant E233P / L234V / L235A / G236_ / S267K, and the constant domain (CH1-hinge-CH2-CH3) of the first monomer contains the pI variant N208D / Q295E / N384D / Q418E / N421D, numbered according to EU numbering. In some embodiments, the 2+1 CLC-type antibody provided herein further contains the FcRn variant M428L / N434S, numbered according to EU numbering.

[0295] In some embodiments, the CH1-hinge-CH2-CH3 of the second monomer contains the amino acid variant L368D / K370S / N208D / Q295E / N384D / Q418E / N421D / E233P / L234V / L235A / G236del / S267K, and the first Fc domain contains the amino acid variant S364K / E357Q / E233P / L234V / L235A / G236del / S267K, with the numbering following EU numbering.

[0296] In some embodiments, each of the two first binding domains binds to tumor ENPP3, and the second binding domain binds to CD28. The 2+1 CLC-type antibody of interest may contain any preferred CD28-binding domain and ENPP3 domain, including either the CD28-binding domain and the ENPP3-binding domain or a variant thereof (see, for example, Figures 15-18, 21, and 22).

[0297] 7. Dual scFv format One heterodimer antibody form particularly used in the target bispecific anti-CD28 × anti-ENPP3 antibody is known in the art and is the bispecific scFv form shown in Figure 14G. In this embodiment, the heterodimer bispecific antibody consists of two scFv-Fc monomers (both in either the (vh-scFv linker-vl-[any domain linker]-CH2-CH3) form or the (vl-scFv linker-vh-[any domain linker]-CH2-CH3) form, or one monomer in one orientation and the other in the other orientation).

[0298] In this case, all ABDs are in scFv format. A dual scFv anti-CD28 × anti-ENPP3 antibody may contain any suitable ENPP3-binding domain and CD28-binding domain, including either the ENPP3-binding domain or the CD28-binding domain provided herein.

[0299] In addition, the Fc domain in the double scFv form is a scuba rian (for example, a set of amino acid substitutions shown in Figures 3 and 8, with particularly useful scuba rians being S364K / E357Q:L368D / K370S;L368D / K370S:S364K;L368E / K370S:S364K;T411T / E360E / Q362E:D401K;L368D / K370S:S364K / E357 The linker is selected from the group consisting of L, K370S:S364K / E357Q, T366S / L368A / Y407V:T366W, and T366S / L368A / Y407V / Y349C:T366W / S354C), optionally includes a de-escalating variant (including those shown in Figure 5), optionally includes a charged scFv linker (including those shown in Figure 6), and the heavy chain includes a pI variant (including those shown in Figure 4).

[0300] In some embodiments, the dual scFv form includes a scuba riant, a pI variant, and a decay variant. Thus, in some embodiments, a) a first monomer (VH1-scFv linker-VL1-[any domain linker]-CH2-CH3 or VL1-scFv linker-VH1-[any domain linker]-CH2-C) containing a scuba riant S364K / E357Q, a decay variant E233P / L234V / L235A / G236del / S267K, and an scFv that binds the first antigen. H3) and b) include forms comprising a first monomer (VH1-scFv linker-VL1-[any domain linker]-CH2-CH3 or VL1-scFv linker-VH1-[any domain linker]-CH2-CH3) containing the scuba antigen L368D / K370S, the diminished variant E233P / L234V / L235A / G236del / S267K, and scFv for binding a second antigen. pI variants may be as outlined herein, but the most common are charged scFv linkers with opposite charges for each monomer. FcRn variants, in particular 428L / 434S, may optionally be included.

[0301] The dual scFv format may include any preferred ENPP3-binding domain and CD28-binding domain, including the ENPP3-binding domain and CD28-binding domain provided herein, as well as any of the associated VH and VL or their variants (see, for example, Figures 15-18, 21, and 22).

[0302] 8.1 Arm Center - scFv One heterodimer antibody form particularly utilized in the anti-CD28 × anti-ENPP3 antibodies provided herein is the one-arm central-scFv form shown in Figure 14K. In this embodiment, one monomer simply contains an Fc domain, while the other monomer contains a Fab domain (first antigen-binding domain), an scFv domain (second antigen-binding domain), and an Fc domain, with the scFv domain inserted between the two Fc domains.

[0303] In this embodiment, one monomer comprises a first heavy chain containing a first variable heavy domain, a CH1 domain, and an Fc domain, and scFv comprises an scFv variable light domain, an scFv linker, and an scFv variable heavy domain. scFv is covalently linked between the C-terminus of the CH1 domain of the heavy steady domain and the N-terminus of the first Fc domain using a domain linker in one of the orientations, i.e., VH1-CH1-[any domain linker]-VH2-scFv linker-VL2-[any domain linker]-CH2-CH3 or VH1-CH1-[any domain linker]-VL2-scFv linker-VH2-[any domain linker]-CH2-CH3. The second monomer comprises an Fc domain (CH2-CH3). This embodiment further utilizes a light chain containing a variable light domain and a steady light domain, which associates with the heavy chain to form a Fab. As with many of the embodiments herein, these constructs may include, as desired and as described herein, scuba riants, pI variants, attenuation variants, additional Fc variants, etc.

[0304] The 1-arm central-scFv format may include any preferred ENPP3-binding domain and CD28-binding domain, including the ENPP3-binding domain and CD28-binding domain provided herein, as well as any of the associated VH and VL or their variants (see, for example, Figures 15-18, 21, and 22).

[0305] Figure 40 shows an exemplary array of ENPP3×CD28 bsAb in the 1-arm central-scFv format.

[0306] 9.1 ARM scFv-mAb format One heterodimeric antibody form particularly utilized in the anti-CD28 × anti-ENPP3 antibodies provided herein is the one-arm mAb-scFv form shown in Figure 14H. This form comprises: 1) a first monomer containing an "empty" Fc domain; 2) a second monomer containing a first variable weight domain (VH), an scFv domain (second antigen-binding domain), and an Fc domain (the scFv domain is bound to the N-terminus of the first variable weight domain); and 3) a light chain containing a first variable light domain and a constant light domain. The first variable weight domain and the first variable light domain form the first antigen-binding domain, and scFv is the second antigen-binding domain. In this form, one of the first and second antigen-binding domains binds to CD28, and the other antigen-binding domain binds to ENPP3. As with many of the embodiments herein, these constructs may include, as desired and as described herein, scuba riants, pI variants, attenuation variants, additional Fc variants, etc.

[0307] A one-arm scFv-mAb antibody may contain any preferred ENPP3-binding domain and CD28-binding domain, including the ENPP3-binding domain and CD28-binding domain provided herein, as well as any associated VH and VL or their variants (see, for example, Figures 15-18, 21, and 22).

[0308] 10.scFv-mAb format One heterodimer antibody form particularly utilized in the anti-CD28 × anti-ENPP3 antibodies provided herein is the mAb-scFv form shown in Figure 14I. In this embodiment, the form relies on the use of N-terminal conjugation of scFv to one of the monomers, thereby forming a third antigen-binding domain, where the Fab portions of the two monomers each bind to one target and the "added" scFv domain binds to a different target.

[0309] In this embodiment, the first monomer comprises a first heavy chain (including a variable heavy domain and a steady domain), and the scFv variable light domain, scFv linker and scFv variable heavy domain are in either orientation ((vh1-scFv linker-vl1-[any domain linker]-vh2-CH1-hinge-CH2-CH3) or (for scFv with the opposite orientation) ((vl1-scFv linker-vh1-[any domain linker]-vh2-CH1-hinge-CH2-CH3)) It has an N-terminal covalent bond scFv. The second monomer contains a heavy chain VH2-CH1-hinge-CH2-CH3. This embodiment further utilizes a common light chain containing a variable light domain and a constant light domain, which associates with the heavy chain to form two identical Fabs. As with many of the embodiments herein, these constructs include scuba variants, pI variants, decay variants, additional Fc variants, etc., as desired and as described herein.

[0310] The scFv-mAb antibody format may include any preferred ENPP3-binding domain and CD28-binding domain, including the ENPP3-binding domain and CD28-binding domain provided herein, as well as any associated VH and VL or their variants (see, for example, Figures 15-18, 21, and 22).

[0311] 11.mAb-Fv format One heterodimer antibody form particularly utilized in the anti-CD28 × anti-ENPP3 antibodies provided herein is the mAb-Fv form (Figure 14L). In this embodiment, the form relies on the use of C-terminal attachment of an "additional" variable heavy domain to one monomer and C-terminal attachment of an "additional" variable light domain to the other monomer, thereby forming a third antigen-binding domain (i.e., an "additional" Fv domain), where the Fab portions of the two monomers bind to CD28 and the "additional" Fv domain binds to ENPP3.

[0312] In this embodiment, the first monomer comprises a first heavy chain comprising a first variable heavy domain and a first steady heavy domain containing a first Fc domain, and having a first variable light domain covalently bonded to the C-terminus of the first Fc domain using a domain linker (vh1-CH1-hinge-CH2-CH3-[any linker]-vl2). The second monomer comprises a second variable heavy domain, a second steady heavy domain containing a second Fc domain, and a third variable heavy domain covalently bonded to the C-terminus of the second Fc domain using a domain linker (vh1-CH1-hinge-CH2-CH3-[any linker]-vh2). This embodiment further utilizes a common light chain comprising a variable light domain and a steady light domain, which associates with the heavy chain to form two identical Fabs containing two identical Fvs. The two C-terminally bonded variable domains constitute an “additional” third Fv. As with many of the embodiments herein, these constructs may include, as desired and as described herein, scuba riants, pI variants, attenuation variants, additional Fc variants, etc.

[0313] The mAb-Fv format may include any preferred ENPP3-binding domain and CD28-binding domain, including the ENPP3-binding domain and CD28-binding domain provided herein, as well as any of the associated VH and VL or their variants (see, for example, Figures 15-18, 21, and 22).

[0314] 12.Central-Fv format One heterodimer antibody form particularly utilized in the anti-CD28 × anti-ENPP3 antibody provided herein is the central-scFv form shown in Figure 14M. In this embodiment, the form relies on the use of an inserted Fv domain, thereby forming an “additional” third antigen-binding domain, where the Fab portions of the two monomers bind to ENPP3 and the “additional” central-Fv domain binds to CD28. The Fv domain is inserted between the Fc domain and the CH1-Fv region of the monomer, thereby providing a third antigen-binding domain, and each monomer contains the Fv component (for example, one monomer contains a variable weight domain of the “additional” central Fv domain and the other contains a variable light domain).

[0315] In this embodiment, one monomer comprises a first heavy chain containing a first variable heavy domain, a CH1 domain, and an Fc domain and an additional variable light domain. The additional variable light domain is covalently linked between the C-terminus of the CH1 domain of the heavy steady domain and the N-terminus of the first Fc domain using a domain linker (vh1-CH1-[any linker]-vh2-hinge-CH2-CH3). The other monomer comprises a first heavy chain containing a first variable heavy domain, a CH1 domain, and an Fc domain and an additional variable heavy domain (vh1-CH1-[any linker]-vh2-hinge-CH2-CH3). The additional variable heavy domain is covalently linked between the C-terminus of the CH1 domain of the heavy chain steady domain and the N-terminus of the first Fc domain using a domain linker. This embodiment utilizes a common light chain containing a variable light domain and a steady light domain, which associates with the heavy chain to form two identical Fabs that each bind to ENPP3. Additional variable heavy domains and additional variable light domains form an “additional” central Fv that binds to CD28. As with many of the embodiments herein, these constructs include, optionally and as described herein, scuba variants, pI variants, attenuation variants, additional Fc variants, etc.

[0316] The central-Fv format may include any preferred ENPP3-binding domain and CD28-binding domain, including the ENPP3-binding domain and CD28-binding domain provided herein, as well as any of the associated VH and VL or their variants (see, for example, Figures 15-18, 21, and 22).

[0317] 13. Non-heterodimer bispecific antibodies As will be understood by those skilled in the art, the anti-CD28 × anti-ENPP3 antibodies provided herein may also include a non-heterodimerative bispecific form (Figure 14J). In this form, the anti-CD28 × anti-ENPP3 comprises: 1) a first monomer containing VH1-CH1-hinge-CH2-CH3; 2) a second monomer containing VH2-CH1-hinge-CH2-CH3; 3) a first light chain containing VL1-CL; and 4) a second light chain containing VL2-CL. In such embodiments, VH1 and VL1 form a first antigen-binding domain, and VH2 and VL2 form a second antigen-binding domain. One of the first or second antigen-binding domains binds to CD28, and the other antigen-binding domain binds to ENPP3.

[0318] Anti-CD28 × anti-ENPP3 antibodies in the form of non-heterodimerary bispecific antibodies may include any preferred ENPP3-binding domain and CD28-binding domain, including the ENPP3-binding domain and CD28-binding domain provided herein, as well as either the associated VH and VL or their variants (see, for example, Figures 15-18, 21, and 22).

[0319] 14. Trident format In some embodiments, the anti-CD28×anti-ENPP3 antibodies provided herein are in the “trident” form, as commonly described in WO2015 / 184203 (the whole and in particular the figures, legend, definitions, and sequences of the “heterodimerization-promoting domain” or “HPD” containing the “K-coil” and “E-coil” sequences are incorporated herein by explicit reference). The trident relies on the use of two different HPDs that associate to form a heterodimer structure as structural components. See Figure 14N. In this embodiment, the trident form comprises a “conventional” heavy and light chain (e.g., VH1-CH1-hinge-CH2-CH3 and VL1-CL), a third chain (VH2-(linker)-VL3-HPD1) containing a first “diabody-type binding domain” or “DART®”, and a fourth chain (VH3-(linker)-(linker)-VL2-HPD2) containing a second DART®. VH1 and VL1 form the first ABD, VH2 and VL2 form the second ABD, and VH3 and VL3 form the third ABD. In some cases, as shown in Figure 14P, the second and third ABDs bind to the same antigen.

[0320] The Trident-Fv format may include any preferred ENPP3-binding domain and CD28-binding domain, including the ENPP3-binding domain and CD28-binding domain provided herein, as well as any of the associated VH and VL or their variants (see, for example, Figures 15-18, 21, and 22).

[0321] V. Nucleic acids In another embodiment, what is provided herein is a nucleic acid composition encoding an anti-CD28 × anti-ENPP3 antibody provided herein. The nucleic acid composition may refer to one or more polynucleotides.

[0322] As will be understood by those skilled in the art, nucleic acid compositions will depend on the form and backbone of the heterodimeric protein. Therefore, for example, if the form requires three amino acid sequences for 1+1 Fab-scFv-Fc, 2+1 mAb-scFv, 2+1 Fab2-scFv-Fc, and 2+1 Fab2-Fc×scFv-Fc forms, then three polynucleotides may be incorporated into one or more expression vectors for expression. In exemplary embodiments, each polynucleotide is incorporated into a different expression vector.

[0323] As is known in the art, the nucleic acids encoding the binding domains and antibody components disclosed herein may be incorporated into an expression vector as is known in the art, and depending on the host cell used to form the heterodimeric antibody of the present invention. Typically, the nucleic acids are functionally ligated to any number of regulatory elements (promoters, origins of replication, selection markers, ribosome binding sites, inducers, etc.). The expression vector may be extrachromosomal or embedded.

[0324] The polynucleotide and / or expression vector of the present invention is then transformed into any number of different types of host cells well known in the art, including mammalian, bacterial, yeast, insect, and / or fungal cells (mammalian cells (e.g., CHO cells) are used in many embodiments).

[0325] In some embodiments, each polynucleotide encoding a monomer is contained within a single expression vector, typically under different or the same promoter control. In certain embodiments of use in the present invention, each of these polynucleotides is contained in a different expression vector. Different vector ratios can be used to induce heterodimerization, as shown herein and in US62 / 025,931 (incorporated herein by reference). Surprisingly, the protein contains the first monomer:second monomer:light chain (in many embodiments herein having three polypeptides including a heterodimeric antibody) in a 1:1:2 ratio, while these are not the ratios that yield the best results.

[0326] The antibodies provided herein are produced by culturing host cells containing an expression vector(s), as is well known in the art. Once produced, conventional antibody purification steps, including an ion-exchange chromatography step, are performed. As discussed herein, by making the pIs of the two monomers differ by at least 0.5, separation by ion-exchange chromatography, isoelectric focusing, or other isoelectric-sensitive methods may be possible. That is, the inclusion of pI substitution modifies the isoelectric point (pI) of each monomer so that each monomer has a different pI and the heterodimer also has a distinguishable pI, thus facilitating isoelectric purification (e.g., anion-exchange column, cation-exchange column) of the "1+1 Fab-scFv-Fc" heterodimer. These substitutions are also useful in determining and monitoring any contaminated double scFv-Fc and mAb homodimers after purification (e.g., IEF gel, cIEF, and analytical IEX column).

[0327] VI. Biological and biochemical functions of anti-CD28 × anti-ENPP3 antibodies Typically, the anti-CD28 × anti-ENPP3 antibodies described herein are administered to patients with ENPP3-associated cancers, and their efficacy is evaluated using the numerous methods described herein. Therefore, standard efficacy assays, such as assessment of cancer burden, tumor size, and the presence or extent of metastasis, may be performed, and cancer immunotherapy may also be evaluated based on an assessment of the immune status. This can be done using numerous methods, including both in vitro and in vivo assays.

[0328] A. Antibody composition for in vivo administration Embodiments of the present invention relate to pharmaceutical compositions comprising one of the anti-CD28 × anti-ENPP3 antibodies described herein and a pharmaceutically acceptable carrier. The anti-CD28 × anti-ENPP3 antibody formulations described herein are prepared for storage in the form of lyophilized formulations or aqueous solutions by mixing the antibody of desired purity with any pharmaceutically acceptable carrier, excipient, or stabilizer (Remington's Pharmaceutical Sciences 16th edition, Osol, A. Ed.

[1980] ). Acceptable carriers, excipients, or stabilizers are non-toxic to the recipient at the dosage and concentration used and include buffers, e.g., phosphoric acid, citrate, and other organic acids; antioxidants, including ascorbic acid and methionine; preservatives (e.g., octadecyldimethylbenzylammonium chloride; hexamethonium chloride; benzalkonium chloride, benzethonium chloride; phenol, butyl alcohol, or benzyl alcohol; alkylparabens, e.g., methylparaben or propylparaben; catechol; resorcinol; cyclohexanol; 3-pentanol; and m-cresol); low molecular weight (less than about 10 residues) polypeptides; proteins, e.g. These include serum albumin, gelatin, or immunoglobulin; hydrophilic polymers, e.g., polyvinylpyrrolidone; amino acids, e.g., glycine, glutamine, asparagine, histidine, arginine, or lysine; monosaccharides, disaccharides, and other carbohydrates, including glucose, mannose, or dextrin; chelating agents, e.g., EDTA; sugars, e.g., sucrose, mannitol, trehalose, or sorbitol; salt-forming counterions, e.g., sodium; metal complexes (e.g., Zn-protein complexes); and / or nonionic surfactants, e.g., TWEEN®, PLURONICS®, or polyethylene glycol (PEG).

[0329] VII. Treatment The compositions of the present invention, once prepared, are utilized in numerous oncological applications, particularly when used in conjunction with anticancer therapies such as antitumor bispecific antibodies, by treating cancer, typically by enhancing the immune response (e.g., T cell activation and proliferation). In some embodiments, the antibodies provided herein enhance the immune response (e.g., T cell activation and proliferation) by providing agonist-like costimulation to T cells in the tumor microenvironment expressing ENPP3.

[0330] In some embodiments, the anti-CD28 × anti-ENPP3 bispecific antibody provided herein is administered in conjunction with an antitumor therapy, for example, an antitumor-associated antigen (TAA) bispecific antibody.

[0331] A. Anti-CD28 x anti-ENPP3 / anti-ENPP3 bispecific antibody In some embodiments, the anti-CD28 × anti-ENPP3 antibody provided herein is administered together with an anti-ENPP3 bispecific antibody, which is a T cell engagement bispecific antibody, such as one that binds to human CD3.

[0332] In classical T cell / APC interactions, there are two signals: a first signal (signal 1) provided by TCR reactivity with peptide-MHC, and a second signal (signal 2) provided by crosslinking of CD28 by CD80 / CD86 expressed on APCs. These two signals work together to fully activate T cells (see Figure 23A). In contrast, treatment with a CD3 bispecific antibody targeting tumor-associated antigens (TAAs) (i.e., an anti-CD3 × anti-ENPP3 bispecific antibody) provides only the first signal.

[0333] Without being bound by any particular theory of operation, the anti-CD28 × anti-ENPP3 bispecific antibody provided herein is thought to be able to enhance the antitumor response of the anti-CD3 × anti-ENPP3 bispecific antibody through CD28 co-stimulation (see Figure 23B). Thus, in one embodiment, what is provided herein is a method for treating ENPP3-related cancer in a patient by administering to the patient an anti-CD3 × anti-ENPP3 bispecific antibody and the anti-CD28 × anti-ENPP3 bispecific antibody provided herein.

[0334] Anti-CD3 × anti-ENPP3 antibodies useful for producing “Signal 1” when combined with the target anti-CD28 × anti-ENPP3 antibody include those having either a CD3-binding domain or an ENPP3-binding domain (see, for example, Figures 22 and 25) provided herein. Suitable antibody forms for such anti-CD3 × anti-ENPP3 antibodies include, but are not limited to, the antibody forms described herein (see, for example, Figure 14). In some embodiments, the anti-CD3 × anti-ENPP3 antibody and the anti-CD28 × anti-ENPP3 antibody used in combination bind to the same ENPP3 epitope. In some embodiments, the anti-CD3 × anti-ENPP3 antibody and the anti-CD28 × anti-ENPP3 antibody used in combination bind to different ENPP3 epitopes.

[0335] B. Dosage Modalities The antibodies provided herein are administered to subjects by known methods, for example, intravenous administration as a bolus or by continuous infusion over a predetermined period.

[0336] C. Treatment Modalities In the method of the present invention, the therapy is used to provide a positive therapeutic response with respect to a disease or condition.

[0337] A “positive therapeutic response” is intended to be an improvement in the disease or condition, and / or an improvement in symptoms associated with the disease or condition. For example, a positive therapeutic response would refer to one or more of the following improvements in the disease: (1) a reduction in the number of neoplastic cells; (2) an increase in neoplastic cell death; (3) an inhibition of neoplastic cell survival; (5) an inhibition of tumor growth (i.e., slowing it down to some extent, preferably stopping it); (6) an increased patient survival rate; and (7) some relief of one or more symptoms associated with the disease or condition.

[0338] A positive therapeutic response in any given disease or condition can be determined by standardized response criteria specific to that disease or condition. Tumor response can be assessed for changes in tumor morphology (i.e., overall tumor load, tumor size, etc.) using screening techniques such as magnetic resonance imaging (MRI) scans, X-ray imaging, computed tomography (CT) scans, bone scan imaging, endoscopy, and tumor biopsy sampling including bone marrow aspiration (BMA) and counting of circulating tumor cells.

[0339] In addition to these positive therapeutic responses, those receiving therapy may experience beneficial effects such as improvement in disease-related symptoms.

[0340] The treatment according to the present invention includes the “therapeutic effective dose” of the drug used. The “therapeutic effective dose” refers to the amount that is effective in the dosage and duration required to achieve the desired therapeutic outcome.

[0341] The therapeutically effective dose may vary depending on factors such as the individual's disease state, age, sex, and weight, as well as the drug's ability to induce the desired response in the individual. The therapeutically effective dose is also defined as the amount at which the therapeutically beneficial effect outweighs any toxic or adverse effects of the antibody or antibody portion.

[0342] The "therapeutic effective dose" for tumor therapy can also be measured by its ability to stabilize disease progression. The ability of a compound to inhibit cancer can be evaluated in animal model systems to predict efficacy in human tumors.

[0343] Alternatively, this property of a composition can be assessed by testing the compound's ability to inhibit cell growth or induce apoptosis using in vitro assays known to those skilled in the art. A therapeutically effective amount of a therapeutic compound may reduce tumor size or otherwise improve symptoms in the subject. Those skilled in the art may determine such an amount based on factors such as the size of the subject, the severity of the subject's symptoms, and the specific composition or route of administration selected.

[0344] The drug regimen may be adjusted to provide the optimal desired response (e.g., therapeutic response). For example, a single bolus may be administered, several divided doses may be administered over time, or the dose may be proportionally reduced or increased as indicated by the requirements of the therapeutic situation. Parenteral compositions may be formulated in drug unit form for ease of administration and uniformity of dosage. When used herein, drug unit form refers to a physically distinct unit suitable as a unit dose for the subject being treated; each unit contains a predetermined amount of the active compound calculated to produce the desired therapeutic effect in relation to the required pharmaceutical carrier.

[0345] The specifications for the drug unit form of the present invention are determined by and directly depend on (a) the unique characteristics of the active compound and the specific therapeutic effect to be achieved, and (b) the limitations inherent in the art to formulate such an active compound for highly sensitive treatment in an individual.

[0346] The efficient dosage and administration regimen for the bispecific antibody used in the present invention depend on the disease or condition being treated and can be determined by those skilled in the art. [Examples]

[0347] Example 1: CD28-binding domain The sequences of human, mouse, and cynomolgus monkey CD28 are shown in Figure 1. These are useful for developing cross-reactive CD28 antigen-binding domains that facilitate clinical development.

[0348] 1A: Novel CD28-bound domain An approach considered to avoid the superagonism associated with TGN1412 was to generate novel CD28-binding domains with lower affinity binding to CD28 than TGN1412 and / or binding to different CD28 epitopes than TGN1412. One means of generating such novel CD28-binding domains was to pan a new in-house phage library against CD28.

[0349] 1A(a): Phage-derived clone 1A7 It should be noted that this phage library utilizes diverse human germline VL introduced into LCDR3. The amino acid sequence of an exemplary phage-derived clone 1A7 is shown in Figure 15.

[0350] To optimize CD28 bsAb, numerous 1A7 affinity variants were developed by manipulating VH variants (exemplary sequences shown as sequence numbers XXX-YYY in Figure 16), VL variants (exemplary sequences shown as sequence numbers XXX-YYY in Figure 17 and Figure 19), the consensus sequence shown in Figure 19, and combinations thereof (exemplary sequences shown in Figure 18). The monovalent affinity of the exemplary variants in relation to scFv is shown in Figure 20.

[0351] Surprisingly, the variable light domain of clone 1A7 was found to differ from the parent germline IGKV1-39VL by only one amino acid in LCDR3. However, despite this single-amino acid difference, the VH of 1A7 did not adequately pair with the parent germline VL, resulting in reduced binding. For example, 1A7_H1.14_L1 showed an affinity of approximately 200 nM, while 1A7_H1.14_IGKV1-39 showed an affinity of over 1 μM. However, for 1A7 VH to be usable in a common light chain bispecific form and to be able to pair with other binding domains that utilize the IGKV1-39 variable light domain, it is crucial that 1A7 VH pairs with IGKV1-39. Therefore, we generated affinity-optimized phage libraries focusing solely on substitutions in the variable heavy domain of 1A7. The amino acid sequences of exemplary affinity-optimized 1A7 variable domains for pairing with IGKV1-39 are shown in Figure 47 and sequence numbers: XXX-YYY. The monovalent KD of the affinity-optimized 1A7Fab variant was determined using Octet, and the data is shown in Figure 48.

[0352] After identifying these additional 1A7 VH variants, preferred variants were paired with the original 1A7_L1 and L1.71 VL to generate further affinity variants (in relation to scFv). The KD of these additional variants is shown in Figure 44. Finally, disulfide-stabilized scFv may be used to stabilize the scFv-containing form. This may improve the behavior in high-concentration formulations, enhance thermal stability, and prevent Fab-LC intercalation. Disulfide-stabilized scFv can be achieved by including cysteine ​​substitutions in VH and VL to form a disulfide bond between two cysteines. An exemplary disulfide-stabilized scFv is shown in Figure 45, with its KD shown in Figure 44. Furthermore, as shown in Figure 43, disulfide stabilization improved the melting temperature by 5.7°C.

[0353] 1B: Additional CD28 binding domains The sequences of additional CD28-binding domains that may be used in the ENPP3×CD28 bsAb of the present invention are shown in Figure 21.

[0354] Example 2: ENPP3 binding domain The sequences of human, mouse, and cynomolgus monkey ENPP3 are shown in Figure 2. These are useful for developing cross-reactive ENPP3 antigen-binding domains that facilitate clinical development. The sequences of ENPP3-binding domains that can be used in the ENPP3×CD28 bsAb of the present invention are shown in Figure 22. The KD values ​​of specific ENPP3-binding domains for both human ENPP3 and cynomolgus monkey ENPP3 were measured using Biacore. Divalent anti-ENPP3 IgG1 mAbs were diluted with HBS-EP+ containing 0.5% BSA. The anti-ENPP3 mAbs were captured at 10 nM for 10 seconds on an anti-human Fc chip prepared by amine coupling with 50 μg / ml Affinipure goat anti-human Fc mAb. Analytes of human ENPP3 (Acro Biosystem EN3-H52H4) or cynomolgus monkey ENPP3 were flowed at 30 μl / min at 37°C with an association time of 5 minutes and a dissociation time of 5 minutes. The results are shown in Figure 46.

[0355] Example 3: Operation of ENPP3×CD28 bsAb T cells require multiple signals for complete activation and differentiation. Signal 1, facilitated by the recognition of peptide-MHC (pMHC) complexes by the T cell receptor (TCR), is absolutely essential for T cell activation. Signal 2, which acts synergistically with and amplifies signal 1, is typically provided by the interaction of CD28 itself with CD80 and CD86, which are ligands for CD28. CD28 engagement alone is typically inactive, but when combined with signal 1 activation, it promotes additional activation, survival, and proliferation signals (including IL-2 secretion) (see Figure 23). Since CD80 and CD86 are spontaneously expressed only by professional antigen-presenting cells (APCs), the degree of CD28 costimulation in the tumor environment can vary considerably. By creating this novel class of tumor-targeted CD28 bispecific antibodies, it is possible to mimic CD28's CD80 / CD86 engagement and provide an artificial source of signal 2. Notably, the signal may be provided by the tumor cell's innate TCR:pMHC recognition, or by a combination of CD28 and CD3 bispecificity (which can mimic signal 1). With these concepts in mind, we devised ENPP3×CD28 bsAb and explored many forms of use. A schematic diagram is outlined in Figure 14.

[0356] 3A:1+1 Fab-scFv-Fc format One exemplary form utilizing the Fab domain and scFv is the 1+1 Fab-scFv-Fc form (schematically shown in Figure 14A), which includes a first monomer containing a single-stranded Fv ("scFv") with first antigen-binding specificity covalently bound to a first heterodimer Fc domain, i.e., scFv-domain linker-CH2-CH3; a second monomer containing a heavy chain, i.e., VH-CH1-hinge-CH2-CH3 (where CH2-CH3 is a second heterodimer Fc domain complementary to the first heterodimer Fc domain); and a light chain (LC) separately transfected so that the Fab domain with second antigen-binding specificity is formed together with a variable heavy domain. This (and other) bispecific forms can utilize any number of heterodimerization approaches known in the art, in combination with any number of approaches for purifying heterodimers (including those shown in Figure 3) from contaminated homodimers. To link the VH and VL domains of scFv, any number of linkers known in the art can be used. Finally, it may be useful to maximize the serum half-life of bsAb, and any number of half-life extension variants known in the art can also be used for these bsAb.

[0357] In particular, 1+1 Fab-scFv-Fc bsAb can utilize skeleton 1 or 11 in Figure 9. This skeleton utilizes the heterodimer Fc variant L368D / K370S (on HC):S364K / E357Q (on scFv-Fc). The HC side further includes the pI variant N208D / Q295E / N384D / Q418E / N421D to increase the negative charge of the heavy chain. scFv utilizes a positively charged (GKPGS)4 linker between the VH and VL domains to increase the positive charge of the scFv-Fc chain. In summary, these two approaches enable the easy purification of heterodimers from contaminated homodimers. The FcγR reduction variants used in this platform are E233P / L234V / L235A / G236_ / S267K substitutions in both HC and scFv-Fc monomers. In some cases, bsAb includes the M428L / N434S half-life extension variant. The sequence of an exemplary ENPP3×CD28 bsAb in the 1+1 Fab-scFv-Fc form is shown in Figure 24.

[0358] 3B: 1-arm central - SCFV format Another exemplary form utilizing the Fab domain and scFv is the one-arm central-scFv (or “half-stack”) form shown in Figure 14K. In this embodiment, one monomer simply contains an Fc domain, while the other monomer contains a Fab domain (first antigen-binding domain), an scFv domain (second antigen-binding domain), and an Fc domain, with the scFv domain inserted between the Fc domains.

[0359] In this embodiment, one monomer comprises a first heavy chain containing a first variable heavy domain, a CH1 domain, and an Fc domain, and scFv comprises an scFv variable light domain, an scFv linker, and an scFv variable heavy domain. scFv is covalently linked between the C-terminus of the CH1 domain of the heavy steady domain and the N-terminus of the first Fc domain using a domain linker in one of the orientations, i.e., VH1-CH1-[any domain linker]-VH2-scFv linker-VL2-[any domain linker]-CH2-CH3 or VH1-CH1-[any domain linker]-VL2-scFv linker-VH2-[any domain linker]-CH2-CH3. The second monomer comprises an Fc domain (CH2-CH3). This embodiment further utilizes a light chain containing a variable light domain and a steady light domain, which associates with the heavy chain to form a Fab. As with many of the embodiments herein, these constructs may include, as desired and as described herein, scuba riants, pI variants, attenuation variants, additional Fc variants, etc.

[0360] The one-arm central-scFv format may include any preferred ENPP3-binding domain and CD28-binding domain, including the ENPP3-binding domain and CD28-binding domain provided herein, as well as any of the associated VH and VL or their variants (see, for example, Figures 15-18, 21, and 22). An exemplary ENPP3×CD28 bsAb of the one-arm central-scFv format is shown in Figure 40.

[0361] 3C:2+1 mAb-scFv format Another exemplary form utilizing the Fab domain and scFv is the 2+1 mAb-scFv form (schematically shown in Figure 14E), which includes a first monomer containing a first heavy chain covalently linked to a single-stranded Fv ("scFv") having a first antigen-binding specificity, i.e., VH-CH1-hinge-CH2-CH3-domain linker-scFv (where CH2-CH3 is the first heterodimer Fc domain), a second monomer containing the heavy chain, i.e., VH-CH1-hinge-CH2-CH3 (where CH2-CH3 is the second heterodimer Fc domain complementary to the first heterodimer Fc domain), and a light chain (LC) separately transfected so that a Fab domain having a second antigen-binding specificity is formed together with two VH domains. In particular, the 2+1 Fab-scFv-Fc bsAb can utilize the skeleton shown in Figure 10. The skeleton utilizes the L368D / K370S (on HC):S364K / E357Q (on HC-scFv side) heterodimer Fc variant. HC further includes the pI variant N208D / Q295E / N384D / Q418E / N421D to increase the negative charge of the heavy chain. scFv utilizes a positively charged (GKPGS) 4-linker between the VH and VL domains to increase the positive charge of the HC-scFv chain. In summary, these two approaches enable the easy purification of heterodimers from contaminated homodimers. The FcγR reduction variant used in this platform is the E233P / L234V / L235A / G236del / S267K substitution in both the HC and scFv-Fc monomers. In some cases, bsAb includes the M428L / N434S half-life extended variant. An exemplary sequence of ENPP3×CD28 bsAb in the 2+1 mAb-scFv format is shown in Figure 41.

[0362] 3D:2+1 Fab2-scFv-Fc format Another exemplary form utilizing the Fab domain and scFv is the 2+1 Fab2-scFv-Fc form (schematically shown in Figure 14B), which may also be referred to as the 2+1 central scFv-Fc form. The 2+1 Fab2-scFv-Fc includes a first monomer containing a heavy chain, namely VH-CH1-hinge-CH2-CH3 (where CH2-CH3 is the first heterodimer Fc domain), a second monomer containing VH-CH1-linker-scFv-linker-CH2-CH3 (where CH2-CH3 is the second heterodimer Fc domain complementary to the first heterodimer Fc domain, and scFv has the first antigen-binding specificity), and a light chain (LC) that is transfected separately so that a Fab domain having the second antigen-binding specificity is formed together with two VH domains. Figure 42 shows an exemplary sequence of ENPP3×CD28 bsAb in the 2+1 Fab2-scFv-Fc format.

[0363] While the above provides a particularly useful framework, the bispecific form can utilize any number of heterodimerization approaches known in the art, combined with any number of approaches for purifying heterodimers (including those shown in Figure 3) from contaminated homodimers. Any number of linkers known in the art can be used to ligate the VH and VL domains of scFv. Finally, it may be useful to maximize the serum half-life of bsAb, and any number of half-life extension variants known in the art can also be used for these bsAb. Despite the bsAb form, CD28 bispecific antibodies are monovalent to CD28 and incorporate Fc variants manipulated to eliminate FcγR binding to avoid potential superagonism. Such Fc variants include those shown in Figure 5.

[0364] Example 4: ENPP3×CD28 bsAb can be effectively combined with CD3 bsAb. As described in Example 3, ENPP3×CD28 bsAb is intended to be combined with a CD3 bsAb. Such a CD3 bsAb may utilize the CD3 binding domain shown in Figure 24. For example, ENPP3×CD28 bsAb can be combined with ENPP3×CD3 bsAb (see, for example, U.S. Patent No. 11,472,890 and WO2020 / 180726). Alternatively, ENPP3×CD28 bsAb may be combined with CD3 bsAb targeting other antigens.

[0365] To investigate, RXF-393 (ENPP3 + Tumor cells were mixed with purified T cells and incubated with escalating doses of exemplary ENPP3×CD28 bsAb XENP44587 (sequence shown in Figure 24) and 1 μg / ml of exemplary ENPP3×CD3 bsAb. Cell supernatants were assayed for IL-2 and IFNγ by MSD (Meso Scale Discovery, Rockville, Md.). The data shown in Figure 26 demonstrate that each of the ENPP3×CD28 bsAbs dose-dependently enhances ENPP3×CD3 bsAb-driven IL-2 secretion.

[0366] Example 5: Further investigation of RTCC enhancement via ENPP3×CD3 and B7H3×CD3 in ENPP3×CD28 Donor PBMCs (80,000 cells, 2:1 E:T) recovered overnight were incubated with or without CFSE-labeled VMRC-RCW (40,000 cells) with semi-logarithmic serial dilutions of exemplary B7H3×CD3 or ENPP3×CD3, and with or without 1 ug / mL of XENP44587 ENPP3(ENPP3-A)×CD28(1A7 H1.14_L1.71, 37nM) in 1+1 Fab-scFv-Fc format. VMRC-RCW target cells had an ENPP3 antigen-binding ability of approximately 10K–20K and were representative of papillary renal cell carcinoma. After 48 hours, cells were stained with surface antibodies to identify cell subsets. Cell viability was measured by Zombie Aqua. Results are shown in Figures 28–32. PBMC + VMRC-RCW alone refers to PBMCs incubated in R10 medium with CFSE-labeled VMRC-RCW in a 2:1 E:T ratio, without CD3 bsAb and without 1 ug / mL CD28 bsAb. 1 ug / mL ENPP3×CD28 refers to PBMCs incubated in R10 medium with CFSE-labeled VMRC-RCW in a 2:1 E:T ratio, without CD3 bsAb and with 1 ug / mL CD28 bsAb. As shown in Figure 28, ENPP3×CD28 bsAb enhances target cell killing via ENPP3×CD3 (Figure 28A) and B7H3×CD3 (Figure 28B). As shown in Figure 29, ENPP3×CD28 bsAb amplifies BclXL induction in CD3 bsAb-mediated RTCC. As shown in Figure 30, ENPP3×CD28 bsAb increases T cell cytotoxicity in CD3 bsAb-mediated RTCC, as measured by granzyme B and CD107a. As shown in Figure 31, ENPP3×CD28 bsAb increases T cell activation and PD1 upregulation in CD3 bsAb-mediated RTCC. As shown in Figure 32, ENPP3×CD28 upregulates intracellular IFNg in ENPP3×CD3-mediated RTCC.

[0367] Example 6: Drag adjustment of ENPP3×CD28 bsAb As described in Example 3, various forms of ENPP3×CD28 bsAb are attempted. Furthermore, as described in Examples 1 & 2, to investigate the effect of different forms and different binding domains of ENPP3×CD28 bsAb on L2 release ability, experiments were conducted in which 8,000 adherent VMRC-RCW cells were incubated overnight with 8,000 T cells in a 1:1 E:T ratio with 3-fold dilution of ENPP3×CD28 and 1 ug / mL of ENPP3×CD3. After incubation at 37°C for 24 hours, the supernatant was collected for MSD analysis. T cells were isolated using the StemCell EasySep® Human T Cell Enrichment Kit (catalog number 19051). The experimental results are shown in Figure 33. The figure demonstrates that the 1+1 Fab-scFv-Fc configuration exhibits the strongest IL2 induction ability compared to 2+1 mAb-scFv, 2+1 Fab2-scFv-Fc (also referred to as 2+1 central-scFv), or 1-arm central-scFv (also referred to as the "half-stack" configuration). In particular, XENP46666, in the 1+1 Fab-scFv-Fc configuration and possessing both an ENPP3-A ENPP3 binding domain and an H1.129_L1 SS (disulfide-stabilized) CD28 binding domain, showed the strongest IL2 induction compared to other binding domains in the same 1+1 Fab-scFv-Fc configuration, as shown in Figure 34.

[0368] In additional experiments, T cells from 14 unique donors were isolated using the StemCell EasySep® Human T Cell Enrichment Kit (catalog no. 19051), and TUHR10TKB cells were seeded in a 1:1 E:T ratio (8,000 T cells:8,000 target cells) with or without exemplary ENPP3×CD3 at 1 ug / mL and ENPP3×CD28 at 10 ug / mL. TUHR10TKB cells had an ENPP3 antigen-binding capacity of approximately 52K, which approximates target cells of renal clear cell carcinoma. After incubation at 37°C for 24 hours, the supernatant was collected for MSD analysis. As shown in Figure 27, XENP4666 again showed significantly higher IL-2 release levels compared to the “half-stack” format or to the same format with the AN1 ENPP3 binding domain.

[0369] Furthermore, another experiment showed that ENPP3×CD28 bsAb enhances RTCC and IFNg production via EpCAM×CD3. Overnight-recovered donor PBMCs (80,000 cells, 2:1 E:T) were incubated with or without CFSE-labeled VMRC-RCW (40,000 cells) with semi-logarithmic serial dilutions of exemplary EpCAM×CD3 and with or without 1 ug / mL of ENPP3×CD28 test. After 48 hours, cells were stained with surface antibodies to identify cell subsets. Cell viability was measured by Zombie Aqua. As shown in Figures 38 and 39, the ENPP3-A ENPP3-binding domain of XENP46666 was superior to the AN1 ENPP3-binding domain of XENP46658 in enhancing target cell killing and INFg production.

[0370] Example 7: XENP46666 enhances IL2 and IFNy release in antigen-specific RTCC assays. For further investigation, pp65-MDA-MB-231 cells were transfected to express ENPP3 with an antigen-binding capacity exceeding 500K. Next, 20,000 ENPP3+ pp65-MDA-MB-231 cells were incubated overnight and then incubated with 400,000 purified T cells from a CMV+ HLA-A*02:01 donor and 3-fold serial dilutions of XENP46666 at a 20:1 E:T ratio. CMV-reactive T cells were purified using the StemCell EasySep® Human T Cell Enrichment Kit (catalog no. 19051). After 24 hours at 37°C, the supernatant was collected for MSD analysis of cytokines. As shown in Figure 35, XENP46666 enhanced IL2 release 2-3-fold compared to T cells + target cells alone.

[0371] Example 8: XENP46666 and XENP46667 exhibit antitumor effects in combination with ENPP3 × CD3 in a mouse model. To test the in vivo antitumor effect of ENPP3×CD28 bsAb, 150 female NSG-DKO mice were given 1×10⁶ bsAb per mouse on day 7. 6 Each RXF-393 cell was intradermally inoculated with 0.1 mL of fluid. On day 1, caliber measurements were performed to divide the mice into 13 groups. On day 0, 5 × 10⁶ mice were inoculated. 6 Each huPBMC was intraperitoneally transplanted in a volume of 0.5 ml, and then dose #1 of the test substance shown in Figure 36 was administered intraperitoneally in a volume of 0.1 ml. As shown, both XENP46666 and XENP46667 showed stronger antitumor activity when administered in combination with ENPP3×xCD3 compared to ENPP3×xCD3 alone.

[0372] Example 9: XENP46666 and the alternative XENP46674 are well tolerated in cynomolgus monkeys. When cynomolgus monkeys were administered XENP46666 or XENP46674 (used as a substitute for cyno due to its strong binding to cyno ENPP3, as shown in Figure 46), no associated deaths or changes in clinical findings, qualitative diet intake, body weight, or hematological parameters were observed. As shown in Figure 37, the half-life ranged from 8.59 to 10.6 days, and C max and AUC 0-last The measured values ​​were dose-proportional across all dose levels.

Claims

1. A heterodimer antibody, a) The first monomer, i) Single-stranded variable fragments (scFv); and ii) The first monomer comprising the first Fc domain, wherein the scFv is covalently bonded to the N-terminus of the first Fc domain using a domain linker; b) The second monomer comprising VH1-CH1-hinge-CH2-CH3 from the N-terminus to the C-terminus, wherein VH1 is a first variable weight domain and CH2-CH3 is a second Fc domain; and c) A light chain comprising VL1-CL from the N-terminus to the C-terminus, wherein VL1 is a first variable light domain and CL is a constant light domain, The scFv includes a second VH domain (VH2), an scFv linker, and a second variable light domain (VL2), The VH1 and VL1 together form a first antigen-binding domain (ABD), and the VH2 and VL2 together form a second ABD. The heterodimer antibody wherein one of the first ABD and the second ABD is a CD28 binding domain, and the other of the first ABD and the second ABD is an ectonucleotide pyrophosphatase / phosphodiesterase 3 (ENPP3) binding domain.

2. The heterodimer antibody according to claim 1, wherein the scFv comprises a VH2-scFv linker-VL2 from the N-terminus to the C-terminus.

3. The heterodimer antibody according to claim 1, wherein the scFv comprises VL2-scFv linker-VH2 from the N-terminus to the C-terminus.

4. The heterodimer antibody according to any one of claims 1 to 3, wherein the first ABD is the ENPP3 binding domain and the second ABD is the CD28 binding domain.

5. The heterodimer antibody according to claim 4, wherein VH1 and VL1 are VH and VL or variants of any of the ENPP3 binding domains shown in Figure 22.

6. The heterodimer antibody according to claim 4 or 5, wherein VH2 and VL2 are selected from one of the following: (1) VH and VL of any of the CD28 binding domains in Figures 15, 18, and 21, or their variants, (2) (i) VH or a variant thereof as shown in Figure 15 or 16 and (ii) VL or a variant thereof as shown in Figure 15 or 17.

7. The heterodimer antibody according to any one of claims 1 to 6, wherein the first Fc domain and the second Fc domain are each variant Fc domains.

8. The heterodimer antibody according to claim 7, wherein the first and second Fc domains comprise a set of heterodimerized scuba arians selected from the following heterodimerized variants: S364K / E357Q:L368D / K370S; S364K:L368D / K370S; S364K:L368E / K370S; D401K:T411E / K360E / Q362E; and T366W:T366S / L368A / Y407V, the numbering according to EU numbering.

9. The heterodimer antibody according to claim 8, wherein the first and second Fc domains comprise the heterodimerized scuba rianto S364K / E357Q:L368D / K370S, and the numbering follows EU numbering.

10. A heterodimer antibody according to any one of claims 7 to 9, wherein each of the first and second Fc domains comprises one or more attenuation variants.

11. The heterodimer antibody according to claim 10, wherein one or more attenuation variants include E233P / L234V / L235A / G236del / S267K, and the numbering follows EU numbering.

12. A heterodimer antibody according to any one of claims 7 to 11, wherein one of the first or second monomers further comprises one or more pI variants.

13. The heterodimer antibody according to claim 12, wherein the CH1-hinge-CH2-CH3 of the second monomer contains the pI variant N208D / Q295E / N384D / Q418E / N421D, and the numbering follows EU numbering.

14. The CH1-hinge-CH2-CH3 of the second monomer contains the amino acid variant E233P / L234V / L235A / G236del / S267K / L368D / K370S / N208D / Q295E / N384D / Q418E / N421D, The first Fc domain contains the amino acid variant E233P / L234V / L235A / G236del / S267K / S364K / E357Q, The heterodimer antibody according to any one of claims 7 to 13, wherein the numbering follows EU numbering.

15. The heterodimer antibody according to any one of claims 8 to 14, wherein the first and second variant Fc domains each further comprise the amino acid variant 428L / 434S.

16. The heterodimer antibody according to any one of claims 1 to 15, wherein the scFv linker is GKPGSGKPGSGKPGSGKPGS.

17. A heterodimer antibody, a) A first monomer comprising VH1-CH1-first domain linker-scFv-second domain linker-CH2-CH3 from the N-terminus to the C-terminus, The first monomer, wherein VH1 is a first variable weight domain and CH2-CH3 is a first Fc domain; b) The second monomer comprising VH1-CH1-hinge-CH2-CH3 from the N-terminus to the C-terminus, wherein CH2-CH3 is a second Fc domain; c) A first light chain comprising VL1-CL from the N-terminus to the C-terminus, wherein VL1 is a first variable light domain and CL is a constant light domain; and d) A second light chain comprising VL1-CL extending from the N-terminus to the C-terminus, wherein VL1 is a first variable light domain and CL is a constant light domain, The scFv includes a second VH domain (VH2), an scFv linker, and a second variable light domain (VL2), The VH1 of the first monomer and the VL1 of the first light chain and the VH1 of the second monomer and the VL1 of the second light chain each form a first antigen-binding domain (ABD), and the VH2 and VL2 form a second ABD. The heterodimer antibody wherein the first ABD is an ectonucleotide pyrophosphatase / phosphodiesterase 3 (ENPP3) binding domain and the second ABD is a CD28 binding domain, or the first ABD is a CD28 binding domain and the second ABD is an ENPP3 binding domain.

18. The heterodimer antibody according to claim 17, wherein the scFv comprises a VH2-scFv linker-VL2 from the N-terminus to the C-terminus.

19. The heterodimer antibody according to claim 17, wherein the scFv comprises VL2-scFv linker-VH2 from the N-terminus to the C-terminus.

20. The heterodimer antibody according to any one of claims 17 to 19, wherein the first ABD is the ENPP3 binding domain and the second ABD is the CD28 binding domain.

21. The heterodimer antibody according to claim 20, wherein VH1 and VL1 are VH and VL of either of the ENPP3 binding domains in Figure 22.

22. The heterodimer antibody according to claim 20 or 21, wherein VH2 and VL2 are selected from one of the following: (1) VH and VL of any of the CD28 binding domains in Figures 15, 18, and 21, or their variants, (2) (i) VH or a variant thereof as shown in Figure 15 or 16 and (ii) VL or a variant thereof as shown in Figure 15 or 17.

23. The heterodimer antibody according to any one of claims 17 to 22, wherein the first Fc domain and the second Fc domain are each variant Fc domains.

24. The heterodimer antibody according to claim 23, wherein the first and second Fc domains comprise a set of heterodimerized scuba arians selected from the following heterodimerized variants: S364K / E357Q:L368D / K370S; S364K:L368D / K370S; S364K:L368E / K370S; D401K:T411E / K360E / Q362E; and T366W:T366S / L368A / Y407V, the numbering according to EU numbering.

25. The heterodimer antibody according to claim 24, wherein the first and second Fc domains comprise the heterodimerized scuba rianto S364K / E357Q:L368D / K370S, and the numbering follows EU numbering.

26. A heterodimer antibody according to any one of claims 23 to 25, wherein each of the first and second Fc domains comprises one or more attenuation variants.

27. The heterodimer antibody according to claim 26, wherein one or more attenuation variants include E233P / L234V / L235A / G236del / S267K, and the numbering follows EU numbering.

28. A heterodimer antibody according to any one of claims 23 to 27, wherein one of the first or second monomers further comprises one or more pI variants.

29. The heterodimer antibody according to claim 28, wherein the CH1-hinge-CH2-CH3 of the second monomer contains the pI variant N208D / Q295E / N384D / Q418E / N421D, and the numbering follows EU numbering.

30. The CH1-hinge-CH2-CH3 of the second monomer contains the amino acid variant E233P / L234V / L235A / G236del / S267K / L368D / K370S / N208D / Q295E / N384D / Q418E / N421D, The first Fc domain contains the amino acid variant E233P / L234V / L235A / G236del / S267K / S364K / E357Q, The heterodimer antibody according to any one of claims 23 to 29, wherein the numbering follows EU numbering.

31. The heterodimer antibody according to any one of claims 24 to 30, wherein the first and second variant Fc domains each further comprise the amino acid variant 428L / 434S.

32. The heterodimer antibody according to any one of claims 17 to 31, wherein the scFv linker is GKPGSGKPGSGKPGSGKPGS.

33. A heterodimer antibody, a) A first monomer comprising VH1-CH1-hinge-CH2-CH3-domain linker-scFv from the N-terminus to the C-terminus, The first monomer, wherein VH1 is a first variable weight domain and CH2-CH3 is a first Fc domain; b) The second monomer comprising VH1-CH1-hinge-CH2-CH3 from the N-terminus to the C-terminus, wherein VH1 is a first variable weight domain and CH2-CH3 is a second Fc domain; c) A first light chain comprising VL1-CL from the N-terminus to the C-terminus, wherein VL1 is a first variable light domain and CL is a constant light domain; and d) A second light chain comprising VL1-CL extending from the N-terminus to the C-terminus, wherein VL1 is a first variable light domain and CL is a constant light domain, The scFv includes a second VH domain (VH2), an scFv linker, and a second variable light domain (VL2), The VH1 of the first monomer and the VL1 of the first light chain and the VH1 of the second monomer and the VL1 of the second light chain each form a first antigen-binding domain (ABD), and the VH2 and VL2 form a second ABD. The heterodimer antibody wherein the first ABD is an ectonucleotide pyrophosphatase / phosphodiesterase 3 (ENPP3) binding domain and the second ABD is a CD28 binding domain, or the first ABD is a CD28 binding domain and the second ABD is an ENPP3 binding domain.

34. The heterodimer antibody according to claim 33, wherein the scFv comprises a VH2-scFv linker-VL2 from the N-terminus to the C-terminus.

35. The heterodimer antibody according to claim 33, wherein the scFv comprises VL2-scFv linker-VH2 from the N-terminus to the C-terminus.

36. The heterodimer antibody according to any one of claims 33 to 35, wherein the first ABD is the ENPP3 binding domain and the second ABD is the CD28 binding domain.

37. The heterodimer antibody according to claim 36, wherein VH1 and VL1 are VH and VL or variants of any of the ENPP3 binding domains shown in Figure 22.

38. The heterodimer antibody according to claim 36 or 37, wherein VH2 and VL2 are selected from one of the following: (1) VH and VL of any of the CD28 binding domains in Figures 15, 18, and 21, or their variants, (2) (i) VH or a variant thereof as shown in Figure 15 or 16 and (ii) VL or a variant thereof as shown in Figure 15 or 17.

39. The heterodimer antibody according to any one of claims 33 to 38, wherein the first Fc domain and the second Fc domain are each variant Fc domains.

40. The heterodimer antibody according to claim 39, wherein the first and second Fc domains comprise a set of heterodimerized scuba arians selected from the following heterodimerized variants: S364K / E357Q:L368D / K370S; S364K:L368D / K370S; S364K:L368E / K370S; D401K:T411E / K360E / Q362E; and T366W:T366S / L368A / Y407V, the numbering according to EU numbering.

41. The heterodimer antibody according to claim 40, wherein the first and second Fc domains comprise the heterodimerized scuba rianto S364K / E357Q:L368D / K370S, and the numbering follows EU numbering.

42. A heterodimer antibody according to any one of claims 39 to 41, wherein each of the first and second Fc domains comprises one or more attenuation variants.

43. The heterodimer antibody according to claim 42, wherein one or more attenuation variants include E233P / L234V / L235A / G236del / S267K, and the numbering follows EU numbering.

44. A heterodimer antibody according to any one of claims 39 to 43, wherein one of the first or second monomers further comprises one or more pI variants.

45. The heterodimer antibody according to claim 44, wherein the CH1-hinge-CH2-CH3 of the second monomer contains the pI variant N208D / Q295E / N384D / Q418E / N421D, and the numbering follows EU numbering.

46. The CH1-hinge-CH2-CH3 of the second monomer contains the amino acid variant E233P / L234V / L235A / G236del / S267K / L368D / K370S / N208D / Q295E / N384D / Q418E / N421D, The first Fc domain contains the amino acid variant E233P / L234V / L235A / G236del / S267K / S364K / E357Q, The heterodimer antibody according to any one of claims 39 to 45, wherein the numbering follows EU numbering.

47. The heterodimer antibody according to any one of claims 40 to 46, wherein the first and second variant Fc domains each further comprise the amino acid variant 428L / 434S.

48. The heterodimer antibody according to any one of claims 33 to 47, wherein the scFv linker is GKPGSGKPGSGKPGSGKPGS.

49. A heterodimer antibody, a) A first monomer comprising VH1-CH1-domain linker-VH1-CH1-hinge-CH2-CH3 from the N-terminus to the C-terminus, Each of the VH1 is a first variable weight domain, and CH2-CH3 is a first Fc domain, in the first monomer; b) The second monomer comprising scFv-domain linker-CH2-CH3 from the N-terminus to the C-terminus, wherein VH1 is a first variable weight domain and CH2-CH3 is a second Fc domain; c) A first light chain comprising VL1-CL from the N-terminus to the C-terminus, wherein VL1 is a first variable light domain and CL is a constant light domain; and d) A second light chain comprising VL1-CL extending from the N-terminus to the C-terminus, wherein VL1 is a first variable light domain and CL is a constant light domain, The scFv includes a second VH domain (VH2), an scFv linker, and a second variable light domain (VL2), The VH1 of the first monomer and the VL1 of the first light chain and the VH1 of the second monomer and the VL1 of the second light chain each form a first antigen-binding domain (ABD), and the VH2 and VL2 form a second ABD. The heterodimer antibody wherein the first ABD is an ectonucleotide pyrophosphatase / phosphodiesterase 3 (ENPP3) binding domain and the second ABD is a CD28 binding domain, or the first ABD is a CD28 binding domain and the second ABD is an ENPP3 binding domain.

50. The heterodimer antibody according to claim 49, wherein the scFv comprises a VH2-scFv linker-VL2 from the N-terminus to the C-terminus.

51. The heterodimer antibody according to claim 49, wherein the scFv comprises VL2-scFv linker-VH2 from the N-terminus to the C-terminus.

52. The heterodimer antibody according to any one of claims 49 to 51, wherein the first ABD is the ENPP3 binding domain and the second ABD is the CD28 binding domain.

53. The heterodimer antibody according to claim 52, wherein VH1 and VL1 are VH and VL or variants of any of the ENPP3 binding domains in Figure 22.

54. The heterodimer antibody according to claim 52 or 53, wherein VH2 and VL2 are selected from one of the following: (1) VH and VL of any of the CD28 binding domains in Figures 15, 18, and 21, or their variants, (2) (i) VH or a variant thereof as shown in Figure 15 or 16 and (ii) VL or a variant thereof as shown in Figure 15 or 17.

55. A heterodimer antibody according to any one of claims 49 to 54, wherein the first Fc domain and the second Fc domain are each variant Fc domains.

56. The heterodimer antibody according to claim 55, wherein the first and second Fc domains comprise a set of heterodimerized scuba arians selected from the following heterodimerized variants: S364K / E357Q:L368D / K370S; S364K:L368D / K370S; S364K:L368E / K370S; D401K:T411E / K360E / Q362E; and T366W:T366S / L368A / Y407V, the numbering according to EU numbering.

57. The heterodimer antibody according to claim 56, wherein the first and second Fc domains comprise the heterodimerized scuba riant S364K / E357Q:L368D / K370S, and the numbering follows EU numbering.

58. A heterodimer antibody according to any one of claims 55 to 57, wherein each of the first and second Fc domains comprises one or more attenuation variants.

59. The heterodimer antibody according to claim 58, wherein one or more attenuation variants include E233P / L234V / L235A / G236del / S267K, and the numbering follows EU numbering.

60. A heterodimer antibody according to any one of claims 55 to 59, wherein one of the first or second monomers further comprises one or more pI variants.

61. The heterodimer antibody according to claim 60, wherein the CH1-hinge-CH2-CH3 of the second monomer contains the pI variant N208D / Q295E / N384D / Q418E / N421D, and the numbering follows EU numbering.

62. The CH1-hinge-CH2-CH3 of the second monomer contains the amino acid variant E233P / L234V / L235A / G236del / S267K / L368D / K370S / N208D / Q295E / N384D / Q418E / N421D, The first Fc domain contains the amino acid variant E233P / L234V / L235A / G236del / S267K / S364K / E357Q, The heterodimer antibody according to any one of claims 55 to 61, wherein the numbering follows EU numbering.

63. The heterodimer antibody according to any one of claims 56 to 62, wherein the first and second variant Fc domains each further comprise the amino acid variant 428L / 434S.

64. The heterodimer antibody according to any one of claims 49 to 63, wherein the scFv linker is GKPGSGKPGSGKPGSGKPGS.

65. a) i) a first variable heavy domain (VH1) and ii) a first variable light domain (VL1) comprising an embryonic antigen-associated cell adhesion molecule 5 (ENPP3) binding domain, b) A bispecific antibody comprising an anti-CD28 binding domain containing i) a second variable heavy domain (VH2) and ii) a second variable light domain (VL2).

66. The bispecific antibody according to claim 65, wherein VH1 and VL1 are VH and VL or variants of either of the ENPP3 binding domains shown in Figure 22.

67. The bispecific antibody according to claim 65 or 66, wherein VH2 and VL2 are selected from one of the following: (1) VH and VL of any of the CD28 binding domains in Figures 15, 18, and 21, or their variants, (2) (i) VH or a variant thereof as shown in Figure 15 or 16 and (ii) VL or a variant thereof as shown in Figure 15 or 17.

68. The bispecific antibody according to any one of claims 65 to 67, wherein the bispecific antibody further comprises a first Fc domain and a second Fc domain.

69. The bispecific antibody according to claim 68, wherein the first and second Fc domains comprise a set of heterodimerized scuba ariants selected from the following heterodimerized variants: S364K / E357Q:L368D / K370S; S364K:L368D / K370S; S364K:L368E / K370S; D401K:T411E / K360E / Q362E; and T366W:T366S / L368A / Y407V, the numbering according to EU numbering.

70. The bispecific antibody according to claim 69, wherein the first and second Fc domains comprise the heterodimerized scuba rianto S364K / E357Q:L368D / K370S, and the numbering follows EU numbering.

71. A bispecific antibody according to any one of claims 68 to 70, wherein each of the first and second Fc domains comprises one or more attenuation variants.

72. The bispecific antibody according to claim 71, wherein one or more attenuation variants include E233P / L234V / L235A / G236del / S267K, and the numbering follows EU numbering.

73. The bispecific antibody according to any one of claims 68 to 72, wherein one of the first or second monomers further comprises one or more pI variants.

74. The bispecific antibody according to claim 73, wherein the pI variant is N208D / Q295E / N384D / Q418E / N421D, and the numbering follows EU numbering.

75. a) A first nucleic acid encoding a first monomer according to any one of claims 1 to 64; b) A second nucleic acid encoding the second monomer according to any one of claims 1 to 64; and c) A nucleic acid composition comprising a third nucleic acid encoding a light chain according to any one of claims 1 to 64.

76. a) A first expression vector comprising the first nucleic acid according to claim 75; b) A second expression vector comprising the second nucleic acid described in claim 75; and c) An expression vector composition comprising a third expression vector each containing the third nucleic acid described in claim 75.

77. A host cell comprising the expression vector composition according to claim 76.

78. A method for producing a heterodimer antibody according to any one of claims 1 to 64, comprising culturing a host cell according to claim 77 under conditions in which the heterodimer antibody is expressed, and recovering the heterodimer antibody.

79. A method for treating ENPP3-related cancer in a patient requiring treatment for ENPP3-related cancer, comprising administering a heterodimer antibody according to any one of claims 1 to 64 to the patient.

80. A method for treating ENPP3-related cancer in a patient requiring treatment for ENPP3-related cancer, comprising administering to the patient a heterodimer antibody and an anti-CD3 × anti-ENPP3 bispecific antibody according to any one of claims 1 to 64.

81. A method for treating ENPP3-related cancer in a patient requiring treatment for ENPP3-related cancer, comprising administering a bispecific antibody according to any one of claims 65 to 74 to the patient.

82. A method for treating ENPP3-related cancer in a patient requiring treatment for ENPP3-related cancer, comprising administering to the patient a bispecific antibody according to any one of claims 65 to 74 and an anti-CD3 × anti-ENPP3 bispecific antibody.

Images