Heterodimeric antibodies that bind ENPP3 and CD3

Abstract

To provide heterodimeric antibodies that bind ENPP3 and CD3.SOLUTION: The present invention is directed to antibodies, including novel antigen binding domains and heterodimeric antibodies, that bind Ectonucleotide pyrophosphatase / phosphodiesterase family member 3 (ENPP3).SELECTED DRAWING: Figure 15A

Description

[Technical Field]

[0001] Priority claim This application claims priority to U.S. Provisional Application Nos. 62 / 812,922, filed March 1, 2019, and 62 / 929,687, filed November 1, 2019, which are incorporated herein by reference in their entireties. [Background technology]

[0002] Antibody-based therapies have been successfully used to treat a variety of diseases, including cancer. An increasingly popular avenue being explored is the engineering of a single immunoglobulin molecule that simultaneously links two different antigens. Such alternative antibody formats that link two different antigens are often referred to as bispecific antibodies. Because the considerable diversity of antibody variable regions (Fvs) allows for the generation of Fvs that recognize virtually any molecule, a typical approach to generating bispecific antibodies is to introduce a new variable region into an antibody.

[0003] A particularly useful approach for bispecific antibodies is to engineer a first binding domain that binds to CD3 and a second binding domain that binds to an antigen associated with or upregulated on cancer cells, such that the bispecific antibody binds to CD3. + The goal is to redirect T cells to destroy cancer cells. Ectonucleotide pyrophosphatase / phosphodiesterase family member 3 (ENPP3) has previously been reported to be highly expressed in renal cell carcinoma and minimally expressed in normal tissues. Given this, anti-ENPP3 antibodies are believed to be useful, for example, for localizing antitumor therapeutic agents (e.g., chemotherapeutic agents and T cells) to such ENPP3-expressing tumors. CD3 + Provided herein are novel bispecific antibodies against CD3 and ENPP3 that are capable of localizing effector T cells to ENPP3-expressing tumors. Summary of the Invention

[0004] Thus, ENPP3 antigen-binding domains and anti-ENPP3 antibodies are provided herein (e.g., bispecific antibodies).

[0005] In one embodiment, the following ENPP3 binding domains are used: AN1[ENPP3]H1L1, AN1[ENPP3]H1 L1.33, AN1[ENPP3]H1 L1.77, AN1[ENPP3]H1.8 L1, AN1[ENPP3]H1.8 L1.33, AN1[ENPP3]H1 L1.77, H16-7.213, H16-9.69, H16-1.52, Ha16-1(1)23, H16-9.44, H16-1.67, Ha1 6-1(3,5)36, H16-1.86, H16-9.10, H16-9.33, H16-1.68, Ha16-1(1)1, Ha1

[0023] Provided herein are compositions comprising an ectonucleotide pyrophosphatase / phosphodiesterase family member 3 (ENPP3) binding domain comprising variable heavy chain complementarity determining regions 1-3 (vhCDR1-3) and variable light chain complementarity determining regions (vlCDR1-3) of any of Ha16-1(3,5)18, Ha16-1(2,4)4, Ha16-1(3,5)56, H16-7.8, H16-1.93, Ha1 6-1(3,5)27.1, H16-1.61, H16-1(3,5)5, ​​H16-7.200, H16-1(3,5)42, H1 6-9.65, Ha1-1(3,5)19, and Ha16-1.80 (Figures 12, 13A-13B, and 14A-14I). In some embodiments, vhCDR1-3 and vlCDR1-3 are selected from the vhCDR1-3 and vlCDR1-3 sequences of the ENPP3 binding domain provided in Figures 12, 13A-13B, and 14A-14I.

[0006] In another embodiment, the following ENPP3 binding domains are used: AN1[ENPP3]H1L1, AN1[ENPP3]H1 L1.33, AN1[ENPP3]H1 L1.77, AN1[ENPP3]H1.8 L1, AN1[ENPP3]H1.8 L1.33, AN1[ENPP3]H1 L1.77, H16-7.213, H16-9.69, H16-1.52, Ha16-1(1)23, H16-9.44, H16-1.67, Ha1 6-1(3,5)36, H16-1.86, H16-9.10, H16-9.33, H16-1.68, Ha16-1(1)1, Ha1

[0023] Provided herein are compositions comprising an ectonucleotide pyrophosphatase / phosphodiesterase family member 3 (ENPP3) binding domain comprising the variable heavy chain domain and the variable light chain domain of any of Ha16-1(3,5)18, Ha16-1(2,4)4, Ha16-1(3,5)56, H16-7.8, H16-1.93, Ha1 6-1(3,5)27.1, H16-1.61, H16-1(3,5)5, ​​H16-7.200, H16-1(3,5)42, H1 6-9.65, Ha1-1(3,5)19, and Ha16-1.80 (Figures 12, 13A-13B, and 14A-14I).

[0007] In another aspect, the present invention provides the following ENPP3 binding domains: AN1[ENPP3]H1L1, AN1[ENPP3]H1 L1.33, AN1[ENPP3]H1 L1.77, AN1[ENPP3]H1.8 L1, AN1[ENPP3]H1.8 L1.33, AN1[ENPP3]H1 L1.77, H16-7.213, H16-9.69, H16-1.52, Ha16-1(1)23, H16-9.44, H16-1.67, Ha1 6-1(3,5)36, H16-1.86, H16-9.10, H16-9.33, H16-1.68, Ha16-1(1)1, Ha1

[0023] Compositions are provided that include an ectonucleotide pyrophosphatase / phosphodiesterase family member 3 (ENPP3) binding domain selected from Ha16-1(3,5)18, Ha16-1(2,4)4, Ha16-1(3,5)56, H16-7.8, H16-1.93, Ha16-1(3,5)27.1, H16-1.61, H16-1(3,5)5, ​​H16-7.200, H16-1(3,5)42, H16-9.65, Ha1-1(3,5)19, and Ha16-1.80 (Figures 12, 13A-13B, and 14A-14I).

[0008] In another aspect, the present invention provides a nucleic acid composition comprising: a) a first nucleic acid encoding a variable heavy chain domain comprising variable heavy chain complementarity determining regions 1 to 3 (vhCDR1 to 3) of an ENPP3-binding domain; and b) a second nucleic acid encoding a variable light chain domain comprising variable light chain complementarity determining regions 1 to 3 (vlCDR1 to 3) of an ENPP3-binding domain, wherein the ENPP3-binding domain is selected from the group consisting of the following ENPP3-binding domains: AN1[ENPP3]H1L1, AN1[ENPP3]H1 L1.33, AN1[ENPP3]H1 L1.77, AN1[ENPP3]H1.8 L1, AN1[ENPP3]H1.8 L1.33, AN1[ENPP3]H1 L1.77, H16-7.213, H16-9.69, H16-1.52, Ha16-1(1)23, H16-9.44, H16-1.67, Ha1 6-1(3,5)36, H16-1.86, H16-9.10, H16-9.33, H16-1.68, Ha16-1(1)1, Ha1 6-1(3,5)18, Ha16-1(2,4)4, Ha16-1(3,5)56, H16-7.8, H16-1.93, Ha1 6-1(3,5)27.1, H16-1.61, H16-1(3,5)5, ​​H16-7.200, H16-1(3,5)42, H1 In some embodiments, the vhCDR1-3 and vlCDR1-3 are selected from the vhCDR1-3 and vlCDR1-3 sequences provided in Figures 12, 13A-13B, and 14A-14I.

[0009] In another aspect, the present invention provides a nucleic acid composition comprising: a) a first nucleic acid encoding a variable heavy chain domain comprising the variable heavy chain domain of an ENPP3-binding domain; and b) a second nucleic acid encoding a variable light chain domain comprising the variable light chain domain of an ENPP3-binding domain, wherein the ENPP3-binding domain is selected from the group consisting of the following ENPP3-binding domains: AN1[ENPP3]H1L1, AN1[ENPP3]H1 L1.33, AN1[ENPP3]H1 L1.77, AN1[ENPP3]H1.8 L1, AN1[ENPP3]H1.8 L1.33, AN1[ENPP3]H1 L1.77, H16-7.213, H16-9.69, H16-1.52, Ha16-1(1)23, H16-9.44, H16-1.67, Ha1 Any one of H16-1(3,5)36, H16-1.86, H16-9.10, H16-9.33, H16-1.68, Ha16-1(1)1, Ha1 6-1(3,5)18, Ha16-1(2,4)4, Ha16-1(3,5)56, H16-7.8, H16-1.93, Ha1 6-1(3,5)27.1, H16-1.61, H16-1(3,5)5, ​​H16-7.200, H16-1(3,5)42, H1 6-9.65, Ha1-1(3,5)19, and Ha16-1.80 (Figures 12, 13A-13B, and 14A-14I).

[0010] In some embodiments, the invention provides an expression vector composition comprising: a) a first expression vector comprising a first nucleic acid; and b) a second expression vector comprising a second nucleic acid. In further embodiments, the invention provides a host cell comprising the expression vector composition.

[0011] In some embodiments, the invention provides a method for producing an ectonucleotide pyrophosphatase / phosphodiesterase family member 3 (ENPP3) binding domain, comprising culturing a host cell under conditions in which the ENPP3 binding domain is expressed, and recovering the ENPP3 binding domain.

[0012] In one aspect, the present invention provides an anti-ENPP3 antibody comprising an ectonucleotide pyrophosphatase / phosphodiesterase family member 3 (ENPP3) binding domain, wherein the ENPP3 binding domain is selected from the group consisting of the following ENPP3 binding domains: AN1[ENPP3]H1L1, AN1[ENPP3]H1 L1.33, AN1[ENPP3]H1 L1.77, AN1[ENPP3]H1.8 L1, AN1[ENPP3]H1.8 L1.33, AN1[ENPP3]H1 L1.77, H16-7.213, H16-9.69, H16-1.52, Ha16-1(1)23, H16-9.44, H16-1.67, Ha1 6-1(3,5)36, H16-1.86, H16-9.10, H16-9.33, H16-1.68, Ha16-1(1)1, Ha1 6-1(3,5)18, Ha16-1(2,4)4, Ha16-1(3,5)56, H16-7.8, H16-1.93, Ha1 6-1(3,5)27.1, H16-1.61, H16-1(3,5)5, ​​H16-7.200, H16-1(3,5)42, H1 In some embodiments, the variable heavy chain complementarity determining regions (vhCDRs) 1-3 and variable light chain complementarity determining regions (vlCDRs) 1-3 of any of Ha1-1.6-9.65, Ha1-1(3,5)19, and Ha16-1.80 (Figures 12, 13A-13B, and 14A-14I) are selected from the vhCDRs 1-3 and vlCDRs 1-3 of any of the following ENPP3-binding domains in Figures 12, 13A-13B, and 14A-14I:

[0013] In another aspect, the present invention provides an anti-ENPP3 antibody comprising an ectonucleotide pyrophosphatase / phosphodiesterase family member 3 (ENPP3) binding domain, wherein the ENPP3 binding domain is selected from the group consisting of the following ENPP3 binding domains: AN1[ENPP3]H1L1, AN1[ENPP3]H1 L1.33, AN1[ENPP3]H1 L1.77, AN1[ENPP3]H1.8 L1, AN1[ENPP3]H1.8 L1.33, AN1[ENPP3]H1 L1.77, H16-7.213, H16-9.69, H16-1.52, Ha16-1(1)23, H16-9.44, H16-1.67, Ha1 H16-1(3,5)42, H1 6-9.65, H1-1(3,5)19, and H16-1.80 (Figures 12, 13A-13B, and 14A-14I).

[0014] In another embodiment, the following ENPP3 binding domains are used: AN1[ENPP3]H1L1, AN1[ENPP3]H1 L1.33, AN1[ENPP3]H1 L1.77, AN1[ENPP3]H1.8 L1, AN1[ENPP3]H1.8 L1.33, AN1[ENPP3]H1 L1.77, H16-7.213, H16-9.69, H16-1.52, Ha16-1(1)23, H16-9.44, H16-1.67, Ha1 6-1(3,5)36, H16-1.86, H16-9.10, H16-9.33, H16-1.68, Ha16-1(1)1, Ha1

[0023] Provided herein are anti-ENPP3 antibodies comprising an ectonucleotide pyrophosphatase / phosphodiesterase family member 3 (ENPP3) binding domain selected from any one of Ha16-1(3,5)18, Ha16-1(2,4)4, Ha16-1(3,5)56, H16-7.8, H16-1.93, Ha1 6-1(3,5)27.1, H16-1.61, H16-1(3,5)5, ​​H16-7.200, H16-1(3,5)42, H1 6-9.65, Ha1-1(3,5)19, and Ha16-1.80 (Figures 12, 13A-13B, and 14A-14I).

[0015] In some embodiments, the antibody comprises a) a first monomer comprising a first antigen-binding domain and a first constant domain, and b) a second monomer comprising a second antigen-binding domain and a second constant domain, wherein either the first antigen-binding domain or the second antigen-binding domain is an ENPP3-binding domain. In further embodiments, the first antigen-binding domain and the second antigen-binding domain bind to different antigens. In further embodiments, the first antigen-binding domain is an ENPP3-binding domain and the second antigen-binding domain is a CD3-binding domain. In further embodiments, the CD3 binding domain comprises vhCDRs 1-3, and vlCDRs 1-3 of any of the following CD3 binding domains: H1.30_L1.47, H1.32_L1.47, H1.89_L1.47, H1.90_L1.47, H1.33_L1.47, H1.31_L1.47, L1.47_H1.30, L1.47_H1.30, L1.47_H1.32, L1.47_H1.89, L1.47_H1.90, L1.47_H1.33, and L1.47_H1.31 (Figures 10A-10F). In a further embodiment, the vhCDR1-3 and vlCDR1-3 of the CD3 binding domain are selected from vhCDR1-3 and vlCDR1-3 of Figures 10A-10F.

[0016] In some embodiments, the CD3 binding domain comprises the variable heavy and light domains of any of the following CD3 binding domains: H1.30_L1.47, H1.32_L1.47, H1.89_L1.47, H1.90_L1.47, H1.33_L1.47, H1.31_L1.47, L1.47_H1.30, L1.47_H1.30, L1.47_H1.32, L1.47_H1.89, L1.47_H1.90, L1.47_H1.33, and L1.47_H1.31 (Figures 10A-10F).

[0017] In some embodiments, the CD3 binding domain is an anti-CD3 scFv.

[0018] In some embodiments, the first and second constant domains each comprise a CH2-CH3.

[0019] In some embodiments, the first and second constant domains each comprise CH1-hinge-CH2-CH3.

[0020] In some embodiments, the first and second constant domains are each variant constant domains.

[0021] In some embodiments, the first and second monomers comprise a set of heterodimerization variants, which are any one of the variants shown in Figures 1A-1E. In some embodiments, the set of heterodimerization variants comprises one of the following sets of variants: S364K / E357Q:L368D / K370S; S364K:L368D / K370S; S364K:L368E / K370S; D401K:T411E / K360E / Q362E; and T366W:T366S / L368A / Y407V.

[0022] In some embodiments, the first and second monomers each further comprise a deletion variant, hi further embodiments, the deletion variant is E233P / L234V / L235A / G236del / S267K.

[0023] In some embodiments, at least one of the first or second monomers further comprises a pI variant. In further embodiments, the pI variant is N208D / Q295E / N384D / Q418E / N421D. In some embodiments, the scFv comprises a charged scFv linker.

[0024] In some embodiments, the invention provides a nucleic acid composition comprising a nucleic acid encoding an anti-ENPP3. In some embodiments, the composition comprises a nucleic acid encoding a first and a second monomer. In some embodiments, the invention provides an expression vector comprising the nucleic acid. In further embodiments, the invention provides a host cell transformed with the expression vector.

[0025] In some embodiments, the present invention provides a method for producing an anti-ENPP3 antibody according to any one of claims B1 to B21, comprising culturing a host cell according to claim B25 under conditions in which the anti-ENPP3 antibody is expressed, and recovering the anti-ENPP3 antibody. In some embodiments, the present invention provides a method for treating cancer, comprising administering the antibody to a patient in need thereof.

[0026] In another aspect, the invention provides a heterodimeric antibody comprising: a) a first monomer comprising i) an anti-CD3 scFv comprising a first variable light chain domain, an scFv linker and a first variable heavy chain domain; and ii) a first Fc domain, wherein the scFv is covalently linked to the N-terminus of the first Fc domain using a domain linker; b) a second monomer comprising a VH2-CH1-hinge-CH2-CH3 monomer, wherein VH is a second variable heavy chain domain and CH2-CH3 is a second Fc domain; and c) a light chain comprising a second variable light chain domain, wherein the second variable heavy chain domain and the second variable light chain domain form an ENPP3-binding domain.

[0027] In some embodiments, the ENPP3 binding domain is any of the following ENPP3 binding domains: AN1[ENPP3]H1L1, AN1[ENPP3]H1 L1.33, AN1[ENPP3]H1 L1.77, AN1[ENPP3]H1.8 L1, AN1[ENPP3]H1.8 L1.33, AN1[ENPP3]H1 L1.77, H16-7.213, H16-9.69, H16-1.52, Ha16-1(1)23, H16-9.44, H16-1.67, Ha1 6-1(3,5)36, H16-1.86, H16-9.10, H16-9.33, H16-1.68, Ha16-1(1)1, Ha1 and Ha16-1(3,5)18, Ha16-1(2,4)4, Ha16-1(3,5)56, H16-7.8, H16-1.93, Ha1 6-1(3,5)27.1, H16-1.61, H16-1(3,5)5, ​​H16-7.200, H16-1(3,5)42, H1 6-9.65, Ha1-1(3,5)19, and Ha16-1.80 (Figures 12, 13A-13B, and 14A-14I).

[0028] In some embodiments, the vhCDR1-3 and vlCDR1-3 of the ENPP3 binding domain are selected from the vhCDR1-3 and vlCDR1-3 sequences of the ENPP3 binding domain provided in Figures 12, 13A-13B, and 14A-14I.

[0029] In some embodiments, the second heavy variable domain is selected from the group consisting of the following ENPP3 binding domains: AN1[ENPP3]H1L1, AN1[ENPP3]H1 L1.33, AN1[ENPP3]H1 L1.77, AN1[ENPP3]H1.8 L1, AN1[ENPP3]H1.8 L1.33, AN1[ENPP3]H1 L1.77, H16-7.213, H16-9.69, H16-1.52, Ha16-1(1)23, H16-9.44, H16-1.67, Ha1 6-1(3,5)36, H16-1.86, H16-9.10, H16-9.33, H16-1.68, Ha16-1(1)1, Ha1 6-1(3,5)18, Ha16-1(2,4)4, Ha16-1(3,5)56, H16-7.8, H16-1.93, Ha1 6-1(3,5)27.1, H16-1.61, H16-1(3,5)5, ​​H16-7.200, H16-1(3,5)42, H1 The heavy variable domains of the AN1[ENPP3]H1L1, AN1[ENPP3]H1L1.33, AN1[ENPP3]H1L1.77, AN1[ENPP3]H1.8L1, AN1[ENPP3]H1.8L1.33, AN1[ENPP3]H1L1.77, H16-7.213, H16-9.69, H16-1.52, Ha16-1(1)23, H16-9.44, H16- H16-1.67, Ha16-1(3,5)36, H16-1.86, H16-9.10, H16-9.33, H16-1.68, Ha16-1(1)1, Ha16-1(3,5)18, Ha16-1(2,4)4, Ha16-1(3,5)56, H16-7.8, H16-1.93, Ha16-1(3,5)27.1, H16-1.61, H16-1(3,5)5, ​​H16-7.200, H16-1(3,5)42, H16-9.65, Ha1-1(3,5)19, and Ha16-1.80 (Figures 12, 13A-13B, and 14A-14I).

[0030] In further embodiments, the anti-CD3 binding scFv comprises vhCDR1-3 and vlCDR1-3 of any of the following CD3 binding domains: H1.30_L1.47, H1.32_L1.47, H1.89_L1.47, H1.90_L1.47, H1.33_L1.47, H1.31_L1.47, L1.47_H1.30, L1.47_H1.30, L1.47_H1.32, L1.47_H1.89, L1.47_H1.90, L1.47_H1.33, and L1.47_H1.31 (Figures 10A-10F).

[0031] In some embodiments, the vhCDR1-3 and vlCDR1-3 of the anti-CD3 scFv are selected from the vhCDR1-3 and vlCDR1-3 of Figures 10A-10F.

[0032] In some embodiments, the anti-CD3 scFv comprises the variable heavy and light domains of any of the following CD3 binding domains: H1.30_L1.47, H1.32_L1.47, H1.89_L1.47, H1.90_L1.47, H1.33_L1.47, H1.31_L1.47, L1.47_H1.30, L1.47_H1.30, L1.47_H1.32, L1.47_H1.89, L1.47_H1.90, L1.47_H1.33, and L1.47_H1.31 (Figures 10A-10F).

[0033] In some embodiments, the first variable light chain domain is covalently linked to the N-terminus of the first Fc domain using a domain linker.

[0034] In some embodiments, the first variable heavy domain is covalently linked to the N-terminus of the first Fc domain using a domain linker.

[0035] In some embodiments, the scFv linker is a charged scFv linker.

[0036] In some embodiments, the first and second Fc domains are variant Fc domains.

[0037] In some embodiments, the first and second monomers comprise a set of heterodimerization variants selected from any of the heterodimers or variants of Figures 1A-1E. In some embodiments, the set of heterodimerization variants selected is derived from the following: S364K / E357Q:L368D / K370S; S364K:L368D / K370S; S364K:L368E / K370S; D401K:T411E / K360E / Q362E; and T366W:T366S / L368A / Y407V, where numbering is according to EU numbering.

[0038] In some embodiments, the first and second monomers further comprise a deletion variant, hi some embodiments, the deletion variant is E233P / L234V / L235A / G236del / S267K, where numbering is according to EU numbering.

[0039] In some embodiments, one of the first or second monomers comprises a pI variant.

[0040] In some embodiments, the pI variant is N208D / Q295E / N384D / Q418E / N421D, where numbering is according to EU numbering.

[0041] In some embodiments, the first monomer comprises the amino acid variants S364K / E357Q / E233P / L234V / L235A / G236del / S267K and the second monomer comprises the amino acid variants L368D / K370S / N208D / Q295E / N384D / Q418E / N421D / E233P / L234V / L235A / G236del / S267K, where numbering is according to EU numbering.

[0042] In some embodiments, the scFv linker is a charged scFv linker having the amino acid sequence (GKPGS)4.

[0043] In some embodiments, the first and second monomers each further comprise the amino acid variant 428 / 434S.

[0044] In some embodiments, the heterodimeric antibodies include the following heterodimeric antibodies: XENP24804, XENP26820, XENP28287, XENP28925, XENP29516, XENP30262, XENP26821, XENP29436, XENP28390, XENP29463, and XENP30263.

[0045] In another aspect, the invention provides a heterodimeric antibody comprising: a) a first monomer comprising, from N-terminus to C-terminus, scFv-linker-CH2-CH3, where the scFv is an anti-CD3 scFv and CH2-CH3 is a first Fc domain; b) a second monomer comprising, from N-terminus to C-terminus, VH-CH1-hinge-CH2-CH3, where CH2-CH3 is a second Fc domain; and c) a light chain comprising VL-CL, wherein the first variant Fc domain is an amino acid sequence. the first and second variant Fc domains each comprise the amino acid variants E233P / L234V / L235A / G236del / S267K; the hinge-CH2-CH3 of the second monomer comprises the amino acid variants N208D / Q295E / N384D / Q418E / N421D; and the VH and VL are AN1[ENPP3]H1L1 and AN1[ENPP3]H1, respectively. L1.33, AN1[ENPP3]H1 L1.77, AN1[ENPP3]H1.8 L1, AN1[ENPP3]H1.8 L1.33, AN1[ENPP3]H1 L1.77, H16-7.213, H16-9.69, H16-1.52, Ha16-1(1)23, H16-9.44, H16-1.67, Ha1 6-1(3,5)36, H16-1.86, H16-9.10, H16-9.33, H16-1.68, Ha16-1(1)1, Ha1 6-1(3,5)18, Ha16-1(2,4)4, Ha16-1(3,5)56, H16-7.8, H16-1.93, Ha1 The ENPP3-binding domain comprises the variable heavy and light chain domains of an ENPP3-binding domain selected from H16-1(3,5)27.1, H16-1.61, H16-1(3,5)5, ​​H16-7.200, H16-1(3,5)42, H16-9.65, Ha1-1(3,5)19, and Ha16-1.80 (Figures 12, 13A-13B, and 14A-14I), and the anti-CD3 scFvs are H1.30_L1.47, H1.32_L1.47, H1.89_L1.47, H1.90_L1.47, H1.33_L1.47, H1.31_L1.47, and L1.47_H1.and a variable heavy chain domain and a variable light chain domain of a CD3 binding domain selected from L1.47_H1.30, L1.47_H1.30, L1.47_H1.32, L1.47_H1.89, L1.47_H1.90, L1.47_H1.33, and L1.47_H1.31 (Figures 10A-10F), where numbering is according to EU numbering.

[0046] In some embodiments, the scFv comprises a charged scFv linker having the amino acid sequence (GKPGS)4.

[0047] In some embodiments, the first and second variant Fc domains each further comprise the amino acid variant 428 / 434S, where numbering is according to EU numbering.

[0048] In some embodiments, the invention provides a nucleic acid composition comprising a first and second monomer and a nucleic acid encoding a light chain of an antibody.

[0049] In some embodiments, the invention provides an expression vector comprising the nucleic acid, hi some embodiments, the invention provides a host cell transformed with the expression vector.

[0050] In some embodiments, the present invention provides a method of treating an ENPP3-associated cancer comprising administering to a patient in need thereof any one of the antibodies provided herein.

[0051] In another aspect, the invention provides a antibody comprising: a) a first monomer comprising, from N-terminus to C-terminus, VH1-CH1-linker1-scFv-linker2-CH2-CH3, where VH1 is a first variable heavy domain, scFv is an anti-CD3 scFv, linker1 and linker2 are a first domain linker and a second domain linker, respectively, and CH2-CH3 is a first Fc domain; and b) a first monomer comprising, from N-terminus to C-terminus, VH1-CH1-linker1-scFv-linker2-CH2-CH3, where VH1 is a first variable heavy domain, scFv is an anti-CD3 scFv, linker1 and linker2 are a first domain linker and a second domain linker, respectively, and CH2-CH3 is a first Fc domain. and c) a second monomer comprising VH2-CH1-hinge-CH2-CH3, where VH2 is a second variable heavy chain domain and CH2-CH3 is a second Fc domain; and c) a common light chain comprising a variable light chain domain, wherein the first variable heavy chain domain and the variable light chain domain form a first ENPP3-binding domain, and the second variable heavy chain domain and the variable light chain domain form a second ENPP3-binding domain.

[0052] In some embodiments, the first and second ENPP3 binding domains are each one of the following ENPP3 binding domains: AN1[ENPP3]H1L1, AN1[ENPP3]H1 L1.33, AN1[ENPP3]H1 L1.77, AN1[ENPP3]H1.8 L1, AN1[ENPP3]H1.8 L1.33, AN1[ENPP3]H1 L1.77, H16-7.213, H16-9.69, H16-1.52, Ha16-1(1)23, H16-9.44, H16-1.67, Ha1 6-1(3,5)36, H16-1.86, H16-9.10, H16-9.33, H16-1.68, Ha16-1(1)1, Ha1 and Ha16-1(3,5)18, Ha16-1(2,4)4, Ha16-1(3,5)56, H16-7.8, H16-1.93, Ha1 6-1(3,5)27.1, H16-1.61, H16-1(3,5)5, ​​H16-7.200, H16-1(3,5)42, H1 6-9.65, Ha1-1(3,5)19, and Ha16-1.80 (Figures 12, 13A-13B, and 14A-14I).

[0053] In some embodiments, the vhCDR1-3 and vlCDR1-3 of the first and second ENPP3 binding domains are selected from the vhCDR1-3 and vlCDR1-3 provided in Figures 14 and 45.

[0054] In some embodiments, the first and second variable heavy chain domains each comprise a variable heavy chain domain of an ENPP3 binding domain, and the first and second variable light chain domains each comprise a variable light chain domain of an ENPP3 binding domain, wherein the ENPP3 binding domain is selected from the group consisting of the following ENPP3 binding domains: AN1[ENPP3]H1L1, AN1[ENPP3]H1 L1.33, AN1[ENPP3]H1 L1.77, AN1[ENPP3]H1.8 L1, AN1[ENPP3]H1.8 L1.33, AN1[ENPP3]H1 L1.77, H16-7.213, H16-9.69, H16-1.52, Ha16-1(1)23, H16-9.44, H16-1.67, Ha1 6-1(3,5)36, H16-1.86, H16-9.10, H16-9.33, H16-1.68, Ha16-1(1)1, Ha1 6-1(3,5)18, Ha16-1(2,4)4, Ha16-1(3,5)56, H16-7.8, H16-1.93, Ha1 6-1(3,5)27.1, H16-1.61, H16-1(3,5)5, ​​H16-7.200, H16-1(3,5)42, H1 6-9.65, Ha1-1(3,5)19, and Ha16-1.80 (Figures 12, 13A-13B, and 14A-14I).

[0055] In some embodiments, the scFv comprises vhCDR1-3 and vlCDR1-3 of any of the following CD3 binding domains: H1.30_L1.47, H1.32_L1.47, H1.89_L1.47, H1.90_L1.47, H1.33_L1.47, H1.31_L1.47, L1.47_H1.30, L1.47_H1.30, L1.47_H1.32, L1.47_H1.89, L1.47_H1.90, L1.47_H1.33, and L1.47_H1.31 (Figures 10A-10F).

[0056] In some embodiments, the vhCDR1-3 and vlCDR1-3 of the scFv are selected from the vhCDR1-3 and vlCDR1-3 of Figures 10A-10F.

[0057] In some embodiments, the scFv comprises the variable heavy and light domains of any of the following CD3 binding domains: H1.30_L1.47, H1.32_L1.47, H1.89_L1.47, H1.90_L1.47, H1.33_L1.47, H1.31_L1.47, L1.47_H1.30, L1.47_H1.30, L1.47_H1.32, L1.47_H1.89, L1.47_H1.90, L1.47_H1.33, and L1.47_H1.31 (Figures 10A-10F).

[0058] In some embodiments, the scFv comprises an scFv variable heavy chain domain, an scFv variable light chain domain, and an scFv linker connecting the scFv variable heavy chain domain and the scFv variable light chain domain.

[0059] In some embodiments, the scFv variable heavy chain domain is linked to the C-terminus of the CH1 of the first monomer using a first domain linker, and the scFv variable light chain domain is covalently linked to the N-terminus of the first Fc domain using a second Fc domain linker.

[0060] In some embodiments, the scFv variable light chain domain is linked to the C-terminus of the CH1 of the first monomer using a first domain linker, and the scFv variable heavy chain domain is covalently linked to the N-terminus of the first Fc domain using a second Fc domain linker.

[0061] In some embodiments, the scFv linker is a charged scFv linker.

[0062] In some embodiments, the first and second Fc domains are variant Fc domains.

[0063] In some embodiments, the first and second monomers comprise a set of heterodimerization variants selected from those shown in Figures 1A-1E.

[0064] In some embodiments, the set of selected heterodimerization variants is derived from the following: S364K / E357Q:L368D / K370S; S364K:L368D / K370S; S364K:L368E / K370S; D401K:T411E / K360E / Q362E; and T366W:T366S / L368A / Y407V, where numbering is according to EU numbering.

[0065] In some embodiments, the first and second monomers further comprise deletion variants.

[0066] In a further embodiment, the deletion variant is E233P / L234V / L235A / G236del / S267K, where numbering is according to EU numbering.

[0067] In some embodiments, one of the first or second monomers further comprises a pI variant.

[0068] In some embodiments, the pI variant is N208D / Q295E / N384D / Q418E / N421D, where numbering is according to EU numbering.

[0069] In some embodiments, the first variant Fc domain comprises the amino acid variants S364K / E357Q / E233P / L234V / L235A / G236del / S267K and the second variant Fc domain comprises the amino acid variants L368D / K370S / N208D / Q295E / N384D / Q418E / N421D / E233P / L234V / L235A / G236del / S267K, where numbering is according to EU numbering.

[0070] In some embodiments, the scFv linker is a charged scFv linker having the amino acid sequence (GKPGS)4.

[0071] In some embodiments, the first and second variant Fc domains each further comprise the amino acid variant 428 / 434S, where numbering is according to EU numbering.

[0072] In some embodiments, the heterodimeric antibodies include the following heterodimeric antibodies: XENP29437, XENP29520, XENP30264, XENP26822, XENP28438, XENP29438, XENP29467, XENP30469, XENP30470, XENP30819, XENP30821, XENP31148, XENP31149, XENP31150, XENP31419, and XENP31471.

[0073] In another embodiment, the heterodimeric antibody comprises: a) a first monomer comprising, from N-terminus to C-terminus, VH1-CH1-linker1-scFv-linker2-CH2-CH3, where scFv is an anti-CD3 scFv and CH2-CH3 is a first Fc domain; b) a second monomer comprising, from N-terminus to C-terminus, VH1-CH1-hinge-CH2-CH3, where CH2-CH3 is a second Fc domain; and c) a common light chain comprising VL-CL, wherein the first variant Fc domain the first and second variant Fc domains each comprise the amino acid variants E233P / L234V / L235A / G236del / S267K; the hinge-CH2-CH3 of the second monomer comprises the amino acid variants N208D / Q295E / N384D / Q418E / N421D; and the VH and VL of the second monomer comprise the amino acid variants AN1[ENPP3]H1L1, AN1[ENPP3]H1L2, AN1[ENPP3]H1L3, AN1[ENPP3]H1L4, AN1[ENPP3]H1L5, AN1[ENPP3]H1L6, AN1[ENPP3]H1L7, AN1[ENPP3]H1L8, AN1[ENPP3]H1L9, AN1[ENPP3]H1L1, AN1[ENPP3]H1L1, AN1[ENPP3]H1L1, AN1[ENPP3]H1L2, AN1[ENPP3]H1L3, AN1[ENPP3]H1L4, AN1[ENPP3]H1L5, AN1[ENPP3]H1L6, AN1[ENPP3]H1L7, AN1[ENPP3]H1L8, AN1[ENPP3]H1L9, AN1[ENPP3]H1L9, AN1[ENPP3]H1L1 ... L1.33, AN1[ENPP3]H1 L1.77, AN1[ENPP3]H1.8 L1, AN1[ENPP3]H1.8 L1.33, AN1[ENPP3]H1 L1.77, H16-7.213, H16-9.69, H16-1.52, Ha16-1(1)23, H16-9.44, H16-1.67, Ha1 6-1(3,5)36, H16-1.86, H16-9.10, H16-9.33, H16-1.68, Ha16-1(1)1, Ha1 6-1(3,5)18, Ha16-1(2,4)4, Ha16-1(3,5)56, H16-7.8, H16-1.93, Ha1 The anti-CD3 scFvs comprise the variable heavy and light chain domains of the ENPP3-binding domain selected from H1.6-1(3,5)27.1, H16-1.61, H16-1(3,5)5, ​​H16-7.200, H16-1(3,5)42, H1 6-9.65, Ha1-1(3,5)19, and Ha16-1.80 (Figures 12, 13A-13B, and 14A-14I), and the anti-CD3 scFvs are H1.30_L1.47, H1.32_L1.47, H1.89_L1.47, H1.90_L1.47, H1.33_L1.47, H1.31_L1.47, L1.47_H1.30, and L1.47_H1.and a variable heavy chain domain and a variable light chain domain of a CD3 binding domain selected from L1.47_H1.30, L1.47_H1.32, L1.47_H1.89, L1.47_H1.90, L1.47_H1.33, and L1.47_H1.31 (Figures 10A-10F), where numbering is according to EU numbering.

[0074] In some embodiments, the scFv comprises a charged scFv linker having the amino acid sequence (GKPGS)4.

[0075] In some embodiments, the first and second variant Fc domains each further comprise the amino acid variant 428 / 434S.

[0076] In some embodiments, the first and second monomers of the antibody and a common light chain. In some embodiments, the invention provides an expression vector comprising the nucleic acid. In some embodiments, the invention provides a host cell transformed with the expression vector. In some embodiments, the invention provides a method for treating ENPP3-associated cancer, comprising administering the antibody to a patient in need thereof.

[0077] In another aspect, the present invention provides heterodimeric antibodies comprising the following heterodimeric antibodies: XENP24804, XENP26820, XENP28287, XENP28925, XENP29516, XENP30262, XENP26821, XENP29436, XENP28390, XENP29463, and XENP30263.

[0078] In another aspect, the present invention provides heterodimeric antibodies, including the following heterodimeric antibodies: XENP29437, XENP29520, XENP30264, XENP26822, XENP28438, XENP29438, XENP29467, XENP30469, XENP30470, XENP30819, XENP30821, XENP31148, XENP31149, XENP31150, XENP31419, and XENP31471. In some embodiments, the present invention provides nucleic acid compositions comprising nucleic acids encoding the heterodimeric antibodies. In some embodiments, the present invention provides expression vectors comprising the nucleic acids. In further embodiments, the present invention provides host cells transformed with the expression vectors.

[0079] In some embodiments, the present invention provides a method of treating an ENPP3-associated cancer comprising administering to a patient in need thereof any one of the heterodimeric antibodies provided herein. [Brief explanation of the drawings]

[0080] [Figure 1A] Useful pairs for a set of Fc heterodimerization variants (including scubariant and PI variants) are shown. There are variants for which there is no corresponding "monomer 2" variant; these are pi variants that can be used alone in either monomer. [Figure 1B] Useful pairs for a set of Fc heterodimerization variants (including scubariant and PI variants) are shown. There are variants for which there is no corresponding "monomer 2" variant; these are pi variants that can be used alone in either monomer. [Figure 1C] Useful pairs for a set of Fc heterodimerization variants (including scubariant and PI variants) are shown. There are variants for which there is no corresponding "monomer 2" variant; these are pi variants that can be used alone in either monomer. [Figure 1D]Useful pairs for a set of Fc heterodimerization variants (including scubariant and PI variants) are shown. There are variants for which there is no corresponding "monomer 2" variant; these are pi variants that can be used alone in either monomer. [Figure 1E] Useful pairs for a set of Fc heterodimerization variants (including scubariant and PI variants) are shown. There are variants for which there is no corresponding "monomer 2" variant; these are pi variants that can be used alone in either monomer. [Figure 2] A list of equivalent variant antibody constant regions and their respective substitutions is provided below. pI_(-) indicates a low pI variant and pI_(+) indicates a high pI variant. These can be optionally and independently combined with other heterodimerization variants of the antibodies described herein (as well as other variants as outlined herein). [Figure 3] Useful deletion variants that eliminate FcγR binding are shown (often referred to as "knockout" or "KO" variants). Generally, deletion variants are found in both monomers, although in some cases they may be found in only one monomer. [Figure 4] Particularly useful embodiments of the "non-Fv" components of the antibodies described herein are provided. [Figure 5]As described herein, several charged scFv linkers are shown for use in increasing or decreasing the pI of the subject heterodimeric bsAbs that utilize one or more scFvs as building blocks. (+H) Positive linkers are particularly used herein, particularly with the anti-CD3 VL and VH sequences shown herein. A single prior art scFv linker bearing a single charge 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 increase proteolytic stability in scFvs. Such charged scFv linkers can be used in any of the subject antibody formats disclosed herein that include scFvs (e.g., 1+1 Fab-scFv-Fc and 2+1 Fab2-scFv-Fc formats). [Figure 6] Some exemplary domain linkers are shown. In some embodiments, these linkers find use in linking a single-chain Fv to an Fc chain. In some embodiments, these linkers can be combined. For example, a GGGGS linker can be combined with a "half-hinge" linker. [Figure 7A]

[0039] Figure 1 shows the sequences of several useful 1+1 Fab-scFv-Fc bispecific antibody format heavy chain scaffolds based on human IgG1, without the Fv sequences (e.g., VH for the scFv and Fab sides). Scaffold 1 is based on human IgG1 (356E / 358M allotype) and contains the S364K / E357Q:L368D / K370S scaffold variant, C220S on the chain with the S364K / E357Q scaffold variant, N208D / Q295E / N384D / Q418E / N421D pI variant on the chain with the L368D / K370S scaffold variant, and E233P / L234V / L235A / G236del / S267K deletion variants on both chains. Scaffold 2 is based on human IgG1 (356E / 358M allotype) and contains the S364K:L368D / K370S scuba variant, C220S on the chain with the S364K scuba variant, N208D / Q295E / N384D / Q418E / N421D pI variants on the chain with the L368D / K370S scuba variant, and E233P / L234V / L235A / G236del / S267K deletion variants on both chains. Scaffold 3 is based on human IgG1 (356E / 358M allotype) and contains the S364K:L368E / K370S scuba variant, C220S on the chain with the S364K scuba variant, N208D / Q295E / N384D / Q418E / N421D pI variants on the chain with the L368E / K370S scuba variant, and E233P / L234V / L235A / G236del / S267K deletion variants on both chains. Scaffold 4 is based on human IgG1 (356E / 358M allotype) and contains the D401K:K360E / Q362E / T411E scubariant, C220S on the chain with the D401K scubariant, N208D / Q295E / N384D / Q418E / N421D pI variants on the chain with the K360E / Q362E / T411E scubariant, and E233P / L234V / L235A / G236del / S267K deletion variants on both chains.Scaffold 5 is based on human IgG1 (356D / 358L allotype) and contains the S364K / E357Q:L368D / K370S scuba variant, C220S on the chain with the S364K / E357Q scuba variant, N208D / Q295E / N384D / Q418E / N421D pI variants on the chain with the L368D / K370S scuba variant, and E233P / L234V / L235A / G236del / S267K deletion variants on both chains. Scaffold 6 is based on human IgG1 (356E / 358M allotype) and contains the S364K / E357Q:L368D / K370S scubariant, C220S on the chain with the S364K / E357Q scubariant, N208D / Q295E / N384D / Q418E / N421D pI variants on the chain with the L368D / K370S scubariant, E233P / L234V / L235A / G236del / S267K deletion variants on both chains, and the N297A variant on both chains. Scaffold 7 is identical to 6 except for the N297S mutation. Scaffold 8 is based on human IgG4 and contains the S364K / E357Q:L368D / K370S scuba variant, the N208D / Q295E / N384D / Q418E / N421D pI variant on the chain with the L368D / K370S scuba variant, and the S228P (EU numbering, which is S241P in Kabat) variant on both chains, which eliminates Fab arm exchange, as known in the art. Scaffold 9 is based on human IgG2 and contains the S364K / E357Q:L368D / K370S scuba variant, the N208D / Q295E / N384D / Q418E / N421D pI variant on the chain with the L368D / K370S scuba variant. Scaffold 10 is based on human IgG2 and contains the S364K / E357Q:L368D / K370S scubariant, the N208D / Q295E / N384D / Q418E / N421D pI variants on the chain with the L368D / K370S scubariant, and the S267K variant on both chains. Scaffold 11 is identical to Scaffold 1 except that it contains the M428L / N434S Xtend mutations.Scaffold 12 is based on human IgG1 (356E / 358M allotype) and includes the P217R / P229R / N276K pI variant on one chain with the S364K / E357Q:L368D / K370S scuba variant, the C220S, and the S364K / E357Q scuba variant, and the E233P / L234V / L235A / G236del / S267K deletion variant on both chains. Included in each of these scaffolds are sequences that are 90, 95, 98, and 99% identical (as defined herein) to the listed sequence and / or contain 1, 2, 3, 4, 5, 6, 7, 8, 9, or 10 additional amino acid substitutions (compared to the "parent" in the figure, which already contains some amino acid modifications compared to the parent human IgG1 (or IgG2 or IgG4, depending on the scaffold), as will be appreciated by those skilled in the art). That is, the recited backbone may contain additional amino acid modifications (generally amino acid substitutions) in addition to the scuba variants, pI variants and deletion variants contained within the backbone of this figure. [Figure 7B]

[0039] Figure 1 shows the sequences of several useful 1+1 Fab-scFv-Fc bispecific antibody format heavy chain scaffolds based on human IgG1, without the Fv sequences (e.g., VH for the scFv and Fab sides). Scaffold 1 is based on human IgG1 (356E / 358M allotype) and contains the S364K / E357Q:L368D / K370S scaffold variant, C220S on the chain with the S364K / E357Q scaffold variant, N208D / Q295E / N384D / Q418E / N421D pI variant on the chain with the L368D / K370S scaffold variant, and E233P / L234V / L235A / G236del / S267K deletion variants on both chains. Scaffold 2 is based on human IgG1 (356E / 358M allotype) and contains the S364K:L368D / K370S scuba variant, C220S on the chain with the S364K scuba variant, N208D / Q295E / N384D / Q418E / N421D pI variants on the chain with the L368D / K370S scuba variant, and E233P / L234V / L235A / G236del / S267K deletion variants on both chains. Scaffold 3 is based on human IgG1 (356E / 358M allotype) and contains the S364K:L368E / K370S scuba variant, C220S on the chain with the S364K scuba variant, N208D / Q295E / N384D / Q418E / N421D pI variants on the chain with the L368E / K370S scuba variant, and E233P / L234V / L235A / G236del / S267K deletion variants on both chains. Scaffold 4 is based on human IgG1 (356E / 358M allotype) and contains the D401K:K360E / Q362E / T411E scubariant, C220S on the chain with the D401K scubariant, N208D / Q295E / N384D / Q418E / N421D pI variants on the chain with the K360E / Q362E / T411E scubariant, and E233P / L234V / L235A / G236del / S267K deletion variants on both chains.Scaffold 5 is based on human IgG1 (356D / 358L allotype) and contains the S364K / E357Q:L368D / K370S scuba variant, C220S on the chain with the S364K / E357Q scuba variant, N208D / Q295E / N384D / Q418E / N421D pI variants on the chain with the L368D / K370S scuba variant, and E233P / L234V / L235A / G236del / S267K deletion variants on both chains. Scaffold 6 is based on human IgG1 (356E / 358M allotype) and contains the S364K / E357Q:L368D / K370S scubariant, C220S on the chain with the S364K / E357Q scubariant, N208D / Q295E / N384D / Q418E / N421D pI variants on the chain with the L368D / K370S scubariant, E233P / L234V / L235A / G236del / S267K deletion variants on both chains, and the N297A variant on both chains. Scaffold 7 is identical to 6 except for the N297S mutation. Scaffold 8 is based on human IgG4 and contains the S364K / E357Q:L368D / K370S scuba variant, the N208D / Q295E / N384D / Q418E / N421D pI variant on the chain with the L368D / K370S scuba variant, and the S228P (EU numbering, which is S241P in Kabat) variant on both chains, which eliminates Fab arm exchange, as known in the art. Scaffold 9 is based on human IgG2 and contains the S364K / E357Q:L368D / K370S scuba variant, the N208D / Q295E / N384D / Q418E / N421D pI variant on the chain with the L368D / K370S scuba variant. Scaffold 10 is based on human IgG2 and contains the S364K / E357Q:L368D / K370S scubariant, the N208D / Q295E / N384D / Q418E / N421D pI variants on the chain with the L368D / K370S scubariant, and the S267K variant on both chains. Scaffold 11 is identical to Scaffold 1 except that it contains the M428L / N434S Xtend mutations.Scaffold 12 is based on human IgG1 (356E / 358M allotype) and includes the P217R / P229R / N276K pI variant on one chain with the S364K / E357Q:L368D / K370S scuba variant, the C220S, and the S364K / E357Q scuba variant, and the E233P / L234V / L235A / G236del / S267K deletion variant on both chains. Included in each of these scaffolds are sequences that are 90, 95, 98, and 99% identical (as defined herein) to the listed sequence and / or contain 1, 2, 3, 4, 5, 6, 7, 8, 9, or 10 additional amino acid substitutions (compared to the "parent" in the figure, which already contains some amino acid modifications compared to the parent human IgG1 (or IgG2 or IgG4, depending on the scaffold), as will be appreciated by those skilled in the art). That is, the recited backbone may contain additional amino acid modifications (generally amino acid substitutions) in addition to the scuba variants, pI variants and deletion variants contained within the backbone of this figure. [Figure 7C]

[0039] Figure 1 shows the sequences of several useful 1+1 Fab-scFv-Fc bispecific antibody format heavy chain scaffolds based on human IgG1, without the Fv sequences (e.g., VH for the scFv and Fab sides). Scaffold 1 is based on human IgG1 (356E / 358M allotype) and contains the S364K / E357Q:L368D / K370S scaffold variant, C220S on the chain with the S364K / E357Q scaffold variant, N208D / Q295E / N384D / Q418E / N421D pI variant on the chain with the L368D / K370S scaffold variant, and E233P / L234V / L235A / G236del / S267K deletion variants on both chains. Scaffold 2 is based on human IgG1 (356E / 358M allotype) and contains the S364K:L368D / K370S scuba variant, C220S on the chain with the S364K scuba variant, N208D / Q295E / N384D / Q418E / N421D pI variants on the chain with the L368D / K370S scuba variant, and E233P / L234V / L235A / G236del / S267K deletion variants on both chains. Scaffold 3 is based on human IgG1 (356E / 358M allotype) and contains the S364K:L368E / K370S scuba variant, C220S on the chain with the S364K scuba variant, N208D / Q295E / N384D / Q418E / N421D pI variants on the chain with the L368E / K370S scuba variant, and E233P / L234V / L235A / G236del / S267K deletion variants on both chains. Scaffold 4 is based on human IgG1 (356E / 358M allotype) and contains the D401K:K360E / Q362E / T411E scubariant, C220S on the chain with the D401K scubariant, N208D / Q295E / N384D / Q418E / N421D pI variants on the chain with the K360E / Q362E / T411E scubariant, and E233P / L234V / L235A / G236del / S267K deletion variants on both chains.Scaffold 5 is based on human IgG1 (356D / 358L allotype) and contains the S364K / E357Q:L368D / K370S scuba variant, C220S on the chain with the S364K / E357Q scuba variant, N208D / Q295E / N384D / Q418E / N421D pI variants on the chain with the L368D / K370S scuba variant, and E233P / L234V / L235A / G236del / S267K deletion variants on both chains. Scaffold 6 is based on human IgG1 (356E / 358M allotype) and contains the S364K / E357Q:L368D / K370S scubariant, C220S on the chain with the S364K / E357Q scubariant, N208D / Q295E / N384D / Q418E / N421D pI variants on the chain with the L368D / K370S scubariant, E233P / L234V / L235A / G236del / S267K deletion variants on both chains, and the N297A variant on both chains. Scaffold 7 is identical to 6 except for the N297S mutation. Scaffold 8 is based on human IgG4 and contains the S364K / E357Q:L368D / K370S scuba variant, the N208D / Q295E / N384D / Q418E / N421D pI variant on the chain with the L368D / K370S scuba variant, and the S228P (EU numbering, which is S241P in Kabat) variant on both chains, which eliminates Fab arm exchange, as known in the art. Scaffold 9 is based on human IgG2 and contains the S364K / E357Q:L368D / K370S scuba variant, the N208D / Q295E / N384D / Q418E / N421D pI variant on the chain with the L368D / K370S scuba variant. Scaffold 10 is based on human IgG2 and contains the S364K / E357Q:L368D / K370S scubariant, the N208D / Q295E / N384D / Q418E / N421D pI variants on the chain with the L368D / K370S scubariant, and the S267K variant on both chains. Scaffold 11 is identical to Scaffold 1 except that it contains the M428L / N434S Xtend mutations.Scaffold 12 is based on human IgG1 (356E / 358M allotype) and includes the P217R / P229R / N276K pI variant on one chain with the S364K / E357Q:L368D / K370S scuba variant, the C220S, and the S364K / E357Q scuba variant, and the E233P / L234V / L235A / G236del / S267K deletion variant on both chains. Included in each of these scaffolds are sequences that are 90, 95, 98, and 99% identical (as defined herein) to the listed sequence and / or contain 1, 2, 3, 4, 5, 6, 7, 8, 9, or 10 additional amino acid substitutions (compared to the "parent" in the figure, which already contains some amino acid modifications compared to the parent human IgG1 (or IgG2 or IgG4, depending on the scaffold), as will be appreciated by those skilled in the art). That is, the recited backbone may contain additional amino acid modifications (generally amino acid substitutions) in addition to the scuba variants, pI variants and deletion variants contained within the backbone of this figure. [Figure 7D]

[0039] Figure 1 shows the sequences of several useful 1+1 Fab-scFv-Fc bispecific antibody format heavy chain scaffolds based on human IgG1, without the Fv sequences (e.g., VH for the scFv and Fab sides). Scaffold 1 is based on human IgG1 (356E / 358M allotype) and contains the S364K / E357Q:L368D / K370S scaffold variant, C220S on the chain with the S364K / E357Q scaffold variant, N208D / Q295E / N384D / Q418E / N421D pI variant on the chain with the L368D / K370S scaffold variant, and E233P / L234V / L235A / G236del / S267K deletion variants on both chains. Scaffold 2 is based on human IgG1 (356E / 358M allotype) and contains the S364K:L368D / K370S scuba variant, C220S on the chain with the S364K scuba variant, N208D / Q295E / N384D / Q418E / N421D pI variants on the chain with the L368D / K370S scuba variant, and E233P / L234V / L235A / G236del / S267K deletion variants on both chains. Scaffold 3 is based on human IgG1 (356E / 358M allotype) and contains the S364K:L368E / K370S scuba variant, C220S on the chain with the S364K scuba variant, N208D / Q295E / N384D / Q418E / N421D pI variants on the chain with the L368E / K370S scuba variant, and E233P / L234V / L235A / G236del / S267K deletion variants on both chains. Scaffold 4 is based on human IgG1 (356E / 358M allotype) and contains the D401K:K360E / Q362E / T411E scubariant, C220S on the chain with the D401K scubariant, N208D / Q295E / N384D / Q418E / N421D pI variants on the chain with the K360E / Q362E / T411E scubariant, and E233P / L234V / L235A / G236del / S267K deletion variants on both chains.Scaffold 5 is based on human IgG1 (356D / 358L allotype) and contains the S364K / E357Q:L368D / K370S scuba variant, C220S on the chain with the S364K / E357Q scuba variant, N208D / Q295E / N384D / Q418E / N421D pI variants on the chain with the L368D / K370S scuba variant, and E233P / L234V / L235A / G236del / S267K deletion variants on both chains. Scaffold 6 is based on human IgG1 (356E / 358M allotype) and contains the S364K / E357Q:L368D / K370S scubariant, C220S on the chain with the S364K / E357Q scubariant, N208D / Q295E / N384D / Q418E / N421D pI variants on the chain with the L368D / K370S scubariant, E233P / L234V / L235A / G236del / S267K deletion variants on both chains, and the N297A variant on both chains. Scaffold 7 is identical to 6 except for the N297S mutation. Scaffold 8 is based on human IgG4 and contains the S364K / E357Q:L368D / K370S scuba variant, the N208D / Q295E / N384D / Q418E / N421D pI variant on the chain with the L368D / K370S scuba variant, and the S228P (EU numbering, which is S241P in Kabat) variant on both chains, which eliminates Fab arm exchange, as known in the art. Scaffold 9 is based on human IgG2 and contains the S364K / E357Q:L368D / K370S scuba variant, the N208D / Q295E / N384D / Q418E / N421D pI variant on the chain with the L368D / K370S scuba variant. Scaffold 10 is based on human IgG2 and contains the S364K / E357Q:L368D / K370S scubariant, the N208D / Q295E / N384D / Q418E / N421D pI variants on the chain with the L368D / K370S scubariant, and the S267K variant on both chains. Scaffold 11 is identical to Scaffold 1 except that it contains the M428L / N434S Xtend mutations.Scaffold 12 is based on human IgG1 (356E / 358M allotype) and includes the P217R / P229R / N276K pI variant on one chain with the S364K / E357Q:L368D / K370S scuba variant, the C220S, and the S364K / E357Q scuba variant, and the E233P / L234V / L235A / G236del / S267K deletion variant on both chains. Included in each of these scaffolds are sequences that are 90, 95, 98, and 99% identical (as defined herein) to the listed sequence and / or contain 1, 2, 3, 4, 5, 6, 7, 8, 9, or 10 additional amino acid substitutions (compared to the "parent" in the figure, which already contains some amino acid modifications compared to the parent human IgG1 (or IgG2 or IgG4, depending on the scaffold), as will be appreciated by those skilled in the art). That is, the recited backbone may contain additional amino acid modifications (generally amino acid substitutions) in addition to the scuba variants, pI variants and deletion variants contained within the backbone of this figure. [Figure 8A]The following shows the sequences of some useful 2+Fab2-scFv-Fc bispecific antibody format heavy chain scaffolds based on human IgG1, without the Fv sequences (e.g., VH for the scFv and Fab sides): Scaffold 1 is based on human IgG1 (356E / 358M allotype) and contains the S364K / E357Q:L368D / K370S scaffold variant, the N208D / Q295E / N384D / Q418E / N421D pI variant on the chain with the L368D / K370S scaffold variant, and the E233P / L234V / L235A / G236del / S267K deletion variant on both chains. Scaffold 2 is based on human IgG1 (356E / 358M allotype) and contains the S364K:L368D / K370S scuba variant, the N208D / Q295E / N384D / Q418E / N421D pI variants on the chain with the L368D / K370S scuba variant, and the E233P / L234V / L235A / G236del / S267K deletion variants on both chains. Scaffold 3 is based on human IgG1 (356E / 358M allotype) and contains the S364K:L368E / K370S scuba variant, N208D / Q295E / N384D / Q418E / N421D pI variants on the chain with the L368E / K370S scuba variant, and E233P / L234V / L235A / G236del / S267K deletion variants on both chains. Scaffold 4 is based on human IgG1 (356E / 358M allotype) and contains the D401K:K360E / Q362E / T411E scubariant, N208D / Q295E / N384D / Q418E / N421D pI variants on the chain with the K360E / Q362E / T411E scubariant, and E233P / L234V / L235A / G236del / S267K deletion variants on both chains. Scaffold 5 is based on human IgG1 (356D / 358L allotype) and contains the S364K / E357Q:L368D / K370S scuba variant, N208D / Q295E / N384D / Q418E / N421D pI variants on the chain with the L368D / K370S scuba variant, and E233P / L234V / L235A / G236del / S267K deletion variants on both chains.Scaffold 6 is based on human IgG1 (356E / 358M allotype) and contains the S364K / E357Q:L368D / K370S scuba variant, the N208D / Q295E / N384D / Q418E / N421D pI variant on the chain with the L368D / K370S scuba variant, and the E233P / L234V / L235A / G236del / S267K deletion variant on both chains, as well as the N297A variant on both chains. Scaffold 7 is identical to 6 except for the N297S mutation. Scaffold 8 is identical to Scaffold 1 except for the M428L / N434S Xtend mutation. Scaffold 9 is based on human IgG1 (356E / 358M allotype) and includes the S364K / E357Q:L368D / K370S scuba variant, the P217R / P229R / N276K pI variant on one chain with the S364K / E357Q scuba variant, and the E233P / L234V / L235A / G236del / S267K deletion variant on both chains. Included in each of these scaffolds are sequences that are 90, 95, 98, and 99% identical (as defined herein) to the listed sequence and / or contain 1, 2, 3, 4, 5, 6, 7, 8, 9, or 10 additional amino acid substitutions (compared to the "parent" in the figure, which already contains some amino acid modifications compared to the parent human IgG1 (or IgG2 or IgG4, depending on the scaffold), as will be understood by those skilled in the art). That is, the recited backbone may contain additional amino acid modifications (generally amino acid substitutions) in addition to the scuba variants, pI variants and deletion variants contained within the backbone of this figure. [Figure 8B]The following shows the sequences of some useful 2+Fab2-scFv-Fc bispecific antibody format heavy chain scaffolds based on human IgG1, without the Fv sequences (e.g., VH for the scFv and Fab sides): Scaffold 1 is based on human IgG1 (356E / 358M allotype) and contains the S364K / E357Q:L368D / K370S scaffold variant, the N208D / Q295E / N384D / Q418E / N421D pI variant on the chain with the L368D / K370S scaffold variant, and the E233P / L234V / L235A / G236del / S267K deletion variant on both chains. Scaffold 2 is based on human IgG1 (356E / 358M allotype) and contains the S364K:L368D / K370S scuba variant, the N208D / Q295E / N384D / Q418E / N421D pI variants on the chain with the L368D / K370S scuba variant, and the E233P / L234V / L235A / G236del / S267K deletion variants on both chains. Scaffold 3 is based on human IgG1 (356E / 358M allotype) and contains the S364K:L368E / K370S scuba variant, N208D / Q295E / N384D / Q418E / N421D pI variants on the chain with the L368E / K370S scuba variant, and E233P / L234V / L235A / G236del / S267K deletion variants on both chains. Scaffold 4 is based on human IgG1 (356E / 358M allotype) and contains the D401K:K360E / Q362E / T411E scubariant, N208D / Q295E / N384D / Q418E / N421D pI variants on the chain with the K360E / Q362E / T411E scubariant, and E233P / L234V / L235A / G236del / S267K deletion variants on both chains. Scaffold 5 is based on human IgG1 (356D / 358L allotype) and contains the S364K / E357Q:L368D / K370S scuba variant, N208D / Q295E / N384D / Q418E / N421D pI variants on the chain with the L368D / K370S scuba variant, and E233P / L234V / L235A / G236del / S267K deletion variants on both chains.Scaffold 6 is based on human IgG1 (356E / 358M allotype) and contains the S364K / E357Q:L368D / K370S scuba variant, the N208D / Q295E / N384D / Q418E / N421D pI variant on the chain with the L368D / K370S scuba variant, and the E233P / L234V / L235A / G236del / S267K deletion variant on both chains, as well as the N297A variant on both chains. Scaffold 7 is identical to 6 except for the N297S mutation. Scaffold 8 is identical to Scaffold 1 except for the M428L / N434S Xtend mutation. Scaffold 9 is based on human IgG1 (356E / 358M allotype) and includes the S364K / E357Q:L368D / K370S scuba variant, the P217R / P229R / N276K pI variant on one chain with the S364K / E357Q scuba variant, and the E233P / L234V / L235A / G236del / S267K deletion variant on both chains. Included in each of these scaffolds are sequences that are 90, 95, 98, and 99% identical (as defined herein) to the listed sequence and / or contain 1, 2, 3, 4, 5, 6, 7, 8, 9, or 10 additional amino acid substitutions (compared to the "parent" in the figure, which already contains some amino acid modifications compared to the parent human IgG1 (or IgG2 or IgG4, depending on the scaffold), as will be understood by those skilled in the art). That is, the recited backbone may contain additional amino acid modifications (generally amino acid substitutions) in addition to the scuba variants, pI variants and deletion variants contained within the backbone of this figure. [Figure 8C]The following shows the sequences of some useful 2+Fab2-scFv-Fc bispecific antibody format heavy chain scaffolds based on human IgG1, without the Fv sequences (e.g., VH for the scFv and Fab sides): Scaffold 1 is based on human IgG1 (356E / 358M allotype) and contains the S364K / E357Q:L368D / K370S scaffold variant, the N208D / Q295E / N384D / Q418E / N421D pI variant on the chain with the L368D / K370S scaffold variant, and the E233P / L234V / L235A / G236del / S267K deletion variant on both chains. Scaffold 2 is based on human IgG1 (356E / 358M allotype) and contains the S364K:L368D / K370S scuba variant, the N208D / Q295E / N384D / Q418E / N421D pI variants on the chain with the L368D / K370S scuba variant, and the E233P / L234V / L235A / G236del / S267K deletion variants on both chains. Scaffold 3 is based on human IgG1 (356E / 358M allotype) and contains the S364K:L368E / K370S scuba variant, N208D / Q295E / N384D / Q418E / N421D pI variants on the chain with the L368E / K370S scuba variant, and E233P / L234V / L235A / G236del / S267K deletion variants on both chains. Scaffold 4 is based on human IgG1 (356E / 358M allotype) and contains the D401K:K360E / Q362E / T411E scubariant, N208D / Q295E / N384D / Q418E / N421D pI variants on the chain with the K360E / Q362E / T411E scubariant, and E233P / L234V / L235A / G236del / S267K deletion variants on both chains. Scaffold 5 is based on human IgG1 (356D / 358L allotype) and contains the S364K / E357Q:L368D / K370S scuba variant, N208D / Q295E / N384D / Q418E / N421D pI variants on the chain with the L368D / K370S scuba variant, and E233P / L234V / L235A / G236del / S267K deletion variants on both chains.Scaffold 6 is based on human IgG1 (356E / 358M allotype) and contains the S364K / E357Q:L368D / K370S scuba variant, the N208D / Q295E / N384D / Q418E / N421D pI variant on the chain with the L368D / K370S scuba variant, and the E233P / L234V / L235A / G236del / S267K deletion variant on both chains, as well as the N297A variant on both chains. Scaffold 7 is identical to 6 except for the N297S mutation. Scaffold 8 is identical to Scaffold 1 except for the M428L / N434S Xtend mutation. Scaffold 9 is based on human IgG1 (356E / 358M allotype) and includes the S364K / E357Q:L368D / K370S scuba variant, the P217R / P229R / N276K pI variant on one chain with the S364K / E357Q scuba variant, and the E233P / L234V / L235A / G236del / S267K deletion variant on both chains. Included in each of these scaffolds are sequences that are 90, 95, 98, and 99% identical (as defined herein) to the listed sequence and / or contain 1, 2, 3, 4, 5, 6, 7, 8, 9, or 10 additional amino acid substitutions (compared to the "parent" in the figure, which already contains some amino acid modifications compared to the parent human IgG1 (or IgG2 or IgG4, depending on the scaffold), as will be understood by those skilled in the art). That is, the recited backbone may contain additional amino acid modifications (generally amino acid substitutions) in addition to the scuba variants, pI variants and deletion variants contained within the backbone of this figure. [Figure 9] Some useful constant light chain domain scaffold sequences are shown, based on human IgG1, without including Fv sequences (e.g., scFv or Fab). Included herein are constant light chain scaffold sequences that are 90, 95, 98, and 99% identical (as defined herein) to the listed sequences and / or contain 1, 2, 3, 4, 5, 6, 7, 8, 9, or 10 additional amino acid modifications. [Figure 10A]The sequence of an exemplary anti-CD3 scFv suitable for use in the bispecific antibodies described herein is shown. The CDRs are underlined, the scFv linker is double underlined (in the sequence, the scFv linker is a positively charged scFv(GKPGS)4 linker (SEQ ID NO:XXX), although one of skill in the art will understand that this linker can be replaced with other linkers, including uncharged or negatively charged linkers, some of which are shown in Figure 5), and the slash indicates the boundary of the variable domain. Furthermore, the nomenclature indicates the orientation of the scFv from N-terminus to C-terminus. As applies to all sequences described herein and containing CDRs herein, the exact identification of the CDR positions may vary slightly depending on the numbering used, as shown in Table 2; therefore, not only the underlined CDRs but also CDRs contained within VH and VL domains using other numbering systems are included herein. Furthermore, with respect to all sequences in the figure, these VH and VL sequences can be used in either scFv or Fab format. [Figure 10B] The sequence of an exemplary anti-CD3 scFv suitable for use in the bispecific antibodies described herein is shown. The CDRs are underlined, the scFv linker is double underlined (in the sequence, the scFv linker is a positively charged scFv(GKPGS)4 linker (SEQ ID NO:XXX), although one of skill in the art will understand that this linker can be replaced with other linkers, including uncharged or negatively charged linkers, some of which are shown in Figure 5), and the slash indicates the boundary of the variable domain. Furthermore, the nomenclature indicates the orientation of the scFv from N-terminus to C-terminus. As applies to all sequences described herein and containing CDRs herein, the exact identification of the CDR positions may vary slightly depending on the numbering used, as shown in Table 2; therefore, not only the underlined CDRs but also CDRs contained within VH and VL domains using other numbering systems are included herein. Furthermore, with respect to all sequences in the figure, these VH and VL sequences can be used in either scFv or Fab format. [Figure 10C] The sequence of an exemplary anti-CD3 scFv suitable for use in the bispecific antibodies described herein is shown. The CDRs are underlined, the scFv linker is double underlined (in the sequence, the scFv linker is a positively charged scFv(GKPGS)4 linker (SEQ ID NO:XXX), although one of skill in the art will understand that this linker can be replaced with other linkers, including uncharged or negatively charged linkers, some of which are shown in Figure 5), and the slash indicates the boundary of the variable domain. Furthermore, the nomenclature indicates the orientation of the scFv from N-terminus to C-terminus. As applies to all sequences described herein and containing CDRs herein, the exact identification of the CDR positions may vary slightly depending on the numbering used, as shown in Table 2; therefore, not only the underlined CDRs but also CDRs contained within VH and VL domains using other numbering systems are included herein. Furthermore, with respect to all sequences in the figure, these VH and VL sequences can be used in either scFv or Fab format. [Figure 10D] The sequence of an exemplary anti-CD3 scFv suitable for use in the bispecific antibodies described herein is shown. The CDRs are underlined, the scFv linker is double underlined (in the sequence, the scFv linker is a positively charged scFv(GKPGS)4 linker (SEQ ID NO:XXX), although one of skill in the art will understand that this linker can be replaced with other linkers, including uncharged or negatively charged linkers, some of which are shown in Figure 5), and the slash indicates the boundary of the variable domain. Furthermore, the nomenclature indicates the orientation of the scFv from N-terminus to C-terminus. As applies to all sequences described herein and containing CDRs herein, the exact identification of the CDR positions may vary slightly depending on the numbering used, as shown in Table 2; therefore, not only the underlined CDRs but also CDRs contained within VH and VL domains using other numbering systems are included herein. Furthermore, with respect to all sequences in the figure, these VH and VL sequences can be used in either scFv or Fab format. [Figure 10E] The sequence of an exemplary anti-CD3 scFv suitable for use in the bispecific antibodies described herein is shown. The CDRs are underlined, the scFv linker is double underlined (in the sequence, the scFv linker is a positively charged scFv(GKPGS)4 linker (SEQ ID NO:XXX), although one of skill in the art will understand that this linker can be replaced with other linkers, including uncharged or negatively charged linkers, some of which are shown in Figure 5), and the slash indicates the boundary of the variable domain. Furthermore, the nomenclature indicates the orientation of the scFv from N-terminus to C-terminus. As applies to all sequences described herein and containing CDRs herein, the exact identification of the CDR positions may vary slightly depending on the numbering used, as shown in Table 2; therefore, not only the underlined CDRs but also CDRs contained within VH and VL domains using other numbering systems are included herein. Furthermore, with respect to all sequences in the figure, these VH and VL sequences can be used in either scFv or Fab format. [Figure 10F]The sequence of an exemplary anti-CD3 scFv suitable for use in the bispecific antibodies described herein is shown. The CDRs are underlined, the scFv linker is double underlined (in the sequence, the scFv linker is a positively charged scFv(GKPGS)4 linker (SEQ ID NO:XXX), although one of skill in the art will understand that this linker can be replaced with other linkers, including uncharged or negatively charged linkers, some of which are shown in Figure 5), and the slash indicates the boundary of the variable domain. Furthermore, the nomenclature indicates the orientation of the scFv from N-terminus to C-terminus. As applies to all sequences described herein and containing CDRs herein, the exact identification of the CDR positions may vary slightly depending on the numbering used, as shown in Table 2; therefore, not only the underlined CDRs but also CDRs contained within VH and VL domains using other numbering systems are included herein. Furthermore, with respect to all sequences in the figure, these VH and VL sequences can be used in either scFv or Fab format. [Figure 11A] The antigen sequences of some of the antigens used in the antibodies described herein are provided, including both human and cynomolgus monkey, to facilitate the development of antigen-binding domains that bind to both, making them amenable to clinical development. [Figure 11B] The antigen sequences of some of the antigens used in the antibodies described herein are provided, including both human and cynomolgus monkey, to facilitate the development of antigen-binding domains that bind to both, making them amenable to clinical development. [Figure 12]The variable heavy and variable light chain sequences of an exemplary humanized ENPP3 binding domain, designated AN1 herein, and the sequence of XENP28278, an anti-ENPP3 mAb based on AN1 and an IgG1 backbone, are shown, each with an E233P / L234V / L235A / G236del / S267K deletion variant. CDRs are underlined, and a slash indicates the boundary between the variable and constant domains. As is true for all sequences described herein and containing CDRs, the exact identification of CDR positions may vary slightly depending on the numbering used, as shown in Table 2; therefore, not only the underlined CDRs but also CDRs contained within VH and VL domains using other numbering systems are included herein. Furthermore, for all sequences in the figure, these VH and VL sequences can be used in either scFv or Fab formats. [Figure 13A] The variable heavy and variable light chain sequences of AN1 variants designed for improved purification and / or modulation of ENPP3 binding affinity and / or potency are shown. The CDRs are underlined, and a slash indicates the boundary between the variable region and the constant domain. As is true for all sequences described herein and containing CDRs herein, the exact identification of the CDR positions may vary slightly depending on the numbering used, as shown in Figure 12 ; therefore, not only the underlined CDRs but also CDRs contained within VH and VL domains using other numbering systems are included herein. Furthermore, for all sequences in the figure, these VH and VL sequences can be used in either scFv or Fab formats. Furthermore, each of the variable heavy chain domains shown herein can be paired with any other αENPP3 variable light chain domain, and each of the variable light chain domains shown herein can be paired with any other αENPP3 variable heavy chain domain. [Figure 13B]The variable heavy and variable light chain sequences of AN1 variants designed for improved purification and / or modulation of ENPP3 binding affinity and / or potency are shown. The CDRs are underlined, and a slash indicates the boundary between the variable region and the constant domain. As is true for all sequences described herein and containing CDRs herein, the exact identification of the CDR positions may vary slightly depending on the numbering used, as shown in Figure 12 ; therefore, not only the underlined CDRs but also CDRs contained within VH and VL domains using other numbering systems are included herein. Furthermore, for all sequences in the figure, these VH and VL sequences can be used in either scFv or Fab formats. Furthermore, each of the variable heavy chain domains shown herein can be paired with any other αENPP3 variable light chain domain, and each of the variable light chain domains shown herein can be paired with any other αENPP3 variable heavy chain domain. [Figure 14A] The variable regions of additional ENPP3 antigen-binding domains that can be used in αENPP3×αCD3 antibodies are shown. The CDRs are underlined. As is true for all sequences described herein and containing CDRs herein, the exact identification of the CDR positions may vary slightly depending on the numbering used, as shown in FIG. 12; therefore, not only the underlined CDRs but also CDRs contained within VH and VL domains using other numbering systems are included herein. Furthermore, for all sequences in the figures, these VH and VL sequences can be used in either scFv or Fab formats. [Figure 14B]The variable regions of additional ENPP3 antigen-binding domains that can be used in αENPP3×αCD3 antibodies are shown. The CDRs are underlined. As is true for all sequences described herein and containing CDRs herein, the exact identification of the CDR positions may vary slightly depending on the numbering used, as shown in FIG. 12; therefore, not only the underlined CDRs but also CDRs contained within VH and VL domains using other numbering systems are included herein. Furthermore, for all sequences in the figures, these VH and VL sequences can be used in either scFv or Fab formats. [Figure 14C] The variable regions of additional ENPP3 antigen-binding domains that can be used in αENPP3×αCD3 antibodies are shown. The CDRs are underlined. As is true for all sequences described herein and containing CDRs herein, the exact identification of the CDR positions may vary slightly depending on the numbering used, as shown in FIG. 12; therefore, not only the underlined CDRs but also CDRs contained within VH and VL domains using other numbering systems are included herein. Furthermore, for all sequences in the figures, these VH and VL sequences can be used in either scFv or Fab formats. [Figure 14D] The variable regions of additional ENPP3 antigen-binding domains that can be used in αENPP3×αCD3 antibodies are shown. The CDRs are underlined. As is true for all sequences described herein and containing CDRs herein, the exact identification of the CDR positions may vary slightly depending on the numbering used, as shown in FIG. 12; therefore, not only the underlined CDRs but also CDRs contained within VH and VL domains using other numbering systems are included herein. Furthermore, for all sequences in the figures, these VH and VL sequences can be used in either scFv or Fab formats. [Figure 14E]The variable regions of additional ENPP3 antigen-binding domains that can be used in αENPP3×αCD3 antibodies are shown. The CDRs are underlined. As is true for all sequences described herein and containing CDRs herein, the exact identification of the CDR positions may vary slightly depending on the numbering used, as shown in FIG. 12; therefore, not only the underlined CDRs but also CDRs contained within VH and VL domains using other numbering systems are included herein. Furthermore, for all sequences in the figures, these VH and VL sequences can be used in either scFv or Fab formats. [Figure 14F] The variable regions of additional ENPP3 antigen-binding domains that can be used in αENPP3×αCD3 antibodies are shown. The CDRs are underlined. As is true for all sequences described herein and containing CDRs herein, the exact identification of the CDR positions may vary slightly depending on the numbering used, as shown in FIG. 12; therefore, not only the underlined CDRs but also CDRs contained within VH and VL domains using other numbering systems are included herein. Furthermore, for all sequences in the figures, these VH and VL sequences can be used in either scFv or Fab formats. [Figure 14G] The variable regions of additional ENPP3 antigen-binding domains that can be used in αENPP3×αCD3 antibodies are shown. The CDRs are underlined. As is true for all sequences described herein and containing CDRs herein, the exact identification of the CDR positions may vary slightly depending on the numbering used, as shown in FIG. 12; therefore, not only the underlined CDRs but also CDRs contained within VH and VL domains using other numbering systems are included herein. Furthermore, for all sequences in the figures, these VH and VL sequences can be used in either scFv or Fab formats. [Figure 14H]The variable regions of additional ENPP3 antigen-binding domains that can be used in αENPP3×αCD3 antibodies are shown. The CDRs are underlined. As is true for all sequences described herein and containing CDRs herein, the exact identification of the CDR positions may vary slightly depending on the numbering used, as shown in FIG. 12; therefore, not only the underlined CDRs but also CDRs contained within VH and VL domains using other numbering systems are included herein. Furthermore, for all sequences in the figures, these VH and VL sequences can be used in either scFv or Fab formats. [Figure 14I] The variable regions of additional ENPP3 antigen-binding domains that can be used in αENPP3×αCD3 antibodies are shown. The CDRs are underlined. As is true for all sequences described herein and containing CDRs herein, the exact identification of the CDR positions may vary slightly depending on the numbering used, as shown in FIG. 12; therefore, not only the underlined CDRs but also CDRs contained within VH and VL domains using other numbering systems are included herein. Furthermore, for all sequences in the figures, these VH and VL sequences can be used in either scFv or Fab formats. [Figure 15A] Several formats of the antibodies described herein are shown. Figure 15A shows a "1+1 Fab-scFv-Fc" format, in which the first arm contains an ENPP3-binding Fab and the second arm contains a CD3-binding scFv. Figure 30B shows a "2+1 Fab2-scFv-Fc" format, in which the first arm contains an ENPP3-binding Fab and the second arm contains a Fab and scFv, where the Fab binds to ENPP3 and the scFv binds to CD3. [Figure 15B]Several formats of the antibodies described herein are shown. Figure 15A shows a "1+1 Fab-scFv-Fc" format, in which the first arm contains an ENPP3-binding Fab and the second arm contains a CD3-binding scFv. Figure 30B shows a "2+1 Fab2-scFv-Fc" format, in which the first arm contains an ENPP3-binding Fab and the second arm contains a Fab and scFv, where the Fab binds to ENPP3 and the scFv binds to CD3. [Figure 16] 1 shows the amino acid sequence of a control anti-RSV x high CD3 bispecific antibody in bottle opener format (Fab-scFv-Fc). The antibody is named with the Fab variable region 1 and the scFv variable region 2 separated by a dash. The CDRs are underlined, and slashes indicate the boundaries of the variable regions. The scFv domains have a VH-scFv linker-VL orientation (N-terminus to C-terminus), although this can be reversed. Additionally, each sequence outlined herein can include or exclude the M428L / N434S variant in one or, preferably, both Fc domains, which results in a longer half-life in serum. [Figure 17A] The sequence of an exemplary αENPP3×αCD3 bsAb is shown in 1+1 Fab-scFv-Fc format and includes the H1.30_L1.47 anti-CD3 scFv (also known as CD3 High [VHVL]). CDRs are underlined, and slashes indicate boundaries between the variable region and other chain components (e.g., constant regions and domain linkers). Note that αENPP3×αCD3 bsAbs can utilize variable region, Fc region, 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. Additionally, each sequence outlined herein can include or exclude the M428L / N434S variant in one or, preferably, both Fc domains, which confers a longer half-life in serum. [Figure 17B]The sequence of an exemplary αENPP3×αCD3 bsAb is shown in 1+1 Fab-scFv-Fc format and includes the H1.30_L1.47 anti-CD3 scFv (also known as CD3 High [VHVL]). CDRs are underlined, and slashes indicate boundaries between the variable region and other chain components (e.g., constant regions and domain linkers). Note that αENPP3×αCD3 bsAbs can utilize variable region, Fc region, 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. Additionally, each sequence outlined herein can include or exclude the M428L / N434S variant in one or, preferably, both Fc domains, which confers a longer half-life in serum. [Figure 17C] The sequence of an exemplary αENPP3×αCD3 bsAb is shown in 1+1 Fab-scFv-Fc format and includes the H1.30_L1.47 anti-CD3 scFv (also known as CD3 High [VHVL]). CDRs are underlined, and slashes indicate boundaries between the variable region and other chain components (e.g., constant regions and domain linkers). Note that αENPP3×αCD3 bsAbs can utilize variable region, Fc region, 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. Additionally, each sequence outlined herein can include or exclude the M428L / N434S variant in one or, preferably, both Fc domains, which confers a longer half-life in serum. [Figure 18A]The sequence of an exemplary αENPP3×αCD3 bsAb is shown in 1+1 Fab-scFv-Fc format and includes the H1.32_L1.47 anti-CD3 scFv (also known as CD3 High-Int#1 [VHVL]). CDRs are underlined, and slashes indicate boundaries between the variable region and other chain components (e.g., constant regions and domain linkers). Note that αENPP3×αCD3 bsAbs can utilize variable regions, Fc regions, and constant domains 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. Additionally, each sequence outlined herein can include or exclude the M428L / N434S variant in one or, preferably, both Fc domains, which confers a longer half-life in serum. [Figure 18B] The sequence of an exemplary αENPP3×αCD3 bsAb is shown in 1+1 Fab-scFv-Fc format and includes the H1.32_L1.47 anti-CD3 scFv (also known as CD3 High-Int#1 [VHVL]). CDRs are underlined, and slashes indicate boundaries between the variable region and other chain components (e.g., constant regions and domain linkers). Note that αENPP3×αCD3 bsAbs can utilize variable regions, Fc regions, and constant domains 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. Additionally, each sequence outlined herein can include or exclude the M428L / N434S variant in one or, preferably, both Fc domains, which confers a longer half-life in serum. [Figure 18C]The sequence of an exemplary αENPP3×αCD3 bsAb is shown in 1+1 Fab-scFv-Fc format and includes the H1.32_L1.47 anti-CD3 scFv (also known as CD3 High-Int#1 [VHVL]). CDRs are underlined, and slashes indicate boundaries between the variable region and other chain components (e.g., constant regions and domain linkers). Note that αENPP3×αCD3 bsAbs can utilize variable regions, Fc regions, and constant domains 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. Additionally, each sequence outlined herein can include or exclude the M428L / N434S variant in one or, preferably, both Fc domains, which confers a longer half-life in serum. [Figure 19A] The sequence of an exemplary αENPP3×αCD3 bsAb is shown in a 2+1 Fab2-scFv-Fc format and includes the H1.30_L1.47 anti-CD3 scFv (also known as CD3 High [VHVL]). CDRs are underlined, and slashes indicate boundaries between the variable region and other chain components (e.g., constant regions and domain linkers). Note that αENPP3×αCD3 bsAbs can utilize variable regions, Fc regions, and constant domains 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. Additionally, each sequence outlined herein can include or exclude the M428L / N434S variant in one or, preferably, both Fc domains, which confers a longer half-life in serum. [Figure 19B]The sequence of an exemplary αENPP3×αCD3 bsAb is shown in a 2+1 Fab2-scFv-Fc format and includes the H1.30_L1.47 anti-CD3 scFv (also known as CD3 High [VHVL]). CDRs are underlined, and slashes indicate boundaries between the variable region and other chain components (e.g., constant regions and domain linkers). Note that αENPP3×αCD3 bsAbs can utilize variable regions, Fc regions, and constant domains 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. Additionally, each sequence outlined herein can include or exclude the M428L / N434S variant in one or, preferably, both Fc domains, which confers a longer half-life in serum. [Figure 19C] The sequence of an exemplary αENPP3×αCD3 bsAb is shown in a 2+1 Fab2-scFv-Fc format and includes the H1.30_L1.47 anti-CD3 scFv (also known as CD3 High [VHVL]). CDRs are underlined, and slashes indicate boundaries between the variable region and other chain components (e.g., constant regions and domain linkers). Note that αENPP3×αCD3 bsAbs can utilize variable regions, Fc regions, and constant domains 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. Additionally, each sequence outlined herein can include or exclude the M428L / N434S variant in one or, preferably, both Fc domains, which confers a longer half-life in serum. [Figure 20A]The sequence of an exemplary αENPP3×αCD3 bsAb is shown, comprising the H1.32_L1.47 anti-CD3 scFv (also known as CD3 High-Int#1 [VHVL]) in a 2+1 Fab2-scFv-Fc format. CDRs are underlined, and slashes indicate boundaries between the variable region and other chain components (e.g., constant regions and domain linkers). Note that αENPP3×αCD3 bsAbs can utilize variable regions, Fc regions, and constant domains 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. Additionally, each sequence outlined herein can include or exclude the M428L / N434S variant in one or, preferably, both Fc domains, which confers a longer half-life in serum. [Figure 20B] The sequence of an exemplary αENPP3×αCD3 bsAb is shown, comprising the H1.32_L1.47 anti-CD3 scFv (also known as CD3 High-Int#1 [VHVL]) in a 2+1 Fab2-scFv-Fc format. CDRs are underlined, and slashes indicate boundaries between the variable region and other chain components (e.g., constant regions and domain linkers). Note that αENPP3×αCD3 bsAbs can utilize variable regions, Fc regions, and constant domains 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. Additionally, each sequence outlined herein can include or exclude the M428L / N434S variant in one or, preferably, both Fc domains, which confers a longer half-life in serum. [Figure 20C]The sequence of an exemplary αENPP3×αCD3 bsAb is shown, comprising the H1.32_L1.47 anti-CD3 scFv (also known as CD3 High-Int#1 [VHVL]) in a 2+1 Fab2-scFv-Fc format. CDRs are underlined, and slashes indicate boundaries between the variable region and other chain components (e.g., constant regions and domain linkers). Note that αENPP3×αCD3 bsAbs can utilize variable regions, Fc regions, and constant domains 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. Additionally, each sequence outlined herein can include or exclude the M428L / N434S variant in one or, preferably, both Fc domains, which confers a longer half-life in serum. [Figure 20D] The sequence of an exemplary αENPP3×αCD3 bsAb is shown, comprising the H1.32_L1.47 anti-CD3 scFv (also known as CD3 High-Int#1 [VHVL]) in a 2+1 Fab2-scFv-Fc format. CDRs are underlined, and slashes indicate boundaries between the variable region and other chain components (e.g., constant regions and domain linkers). Note that αENPP3×αCD3 bsAbs can utilize variable regions, Fc regions, and constant domains 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. Additionally, each sequence outlined herein can include or exclude the M428L / N434S variant in one or, preferably, both Fc domains, which confers a longer half-life in serum. [Figure 21]The sequence of an exemplary αENPP3×αCD3 bsAb is shown in a 2+1 Fab2-scFv-Fc format and includes the L1.47_H1.30 anti-CD3 scFv (also known as CD3 High [VLVH]). CDRs are underlined, and slashes indicate boundaries between the variable region and other chain components (e.g., constant regions and domain linkers). Note that αENPP3×αCD3 bsAbs can utilize variable regions, Fc regions, and constant domains 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. Additionally, each sequence outlined herein can include or exclude the M428L / N434S variant in one or, preferably, both Fc domains, which confers a longer half-life in serum. [Figure 22A] The sequence of an exemplary αENPP3×αCD3 bsAb is shown, comprising the L1.47_H1.32 anti-CD3 scFv (also known as CD3 High-Int#1 [VLVH]) in a 2+1 Fab2-scFv-Fc format. CDRs are underlined, and slashes indicate boundaries between the variable region and other chain components (e.g., constant regions and domain linkers). Note that αENPP3×αCD3 bsAbs can utilize variable regions, Fc regions, and constant domains 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. Additionally, each sequence outlined herein can include or exclude the M428L / N434S variant in one or, preferably, both Fc domains, which results in a longer half-life in serum. [Figure 22B]The sequence of an exemplary αENPP3×αCD3 bsAb is shown, comprising the L1.47_H1.32 anti-CD3 scFv (also known as CD3 High-Int#1 [VLVH]) in a 2+1 Fab2-scFv-Fc format. CDRs are underlined, and slashes indicate boundaries between the variable region and other chain components (e.g., constant regions and domain linkers). Note that αENPP3×αCD3 bsAbs can utilize variable regions, Fc regions, and constant domains 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. Additionally, each sequence outlined herein can include or exclude the M428L / N434S variant in one or, preferably, both Fc domains, which results in a longer half-life in serum. [Figure 22C] The sequence of an exemplary αENPP3×αCD3 bsAb is shown, comprising the L1.47_H1.32 anti-CD3 scFv (also known as CD3 High-Int#1 [VLVH]) in a 2+1 Fab2-scFv-Fc format. CDRs are underlined, and slashes indicate boundaries between the variable region and other chain components (e.g., constant regions and domain linkers). Note that αENPP3×αCD3 bsAbs can utilize variable regions, Fc regions, and constant domains 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. Additionally, each sequence outlined herein can include or exclude the M428L / N434S variant in one or, preferably, both Fc domains, which results in a longer half-life in serum. [Figure 23A]The sequence of an exemplary αENPP3×αCD3 bsAb is shown, comprising the L1.47_H1.89 anti-CD3 scFv (also known as CD3 High-Int#2 [VLVH]) in a 2+1 Fab2-scFv-Fc format. CDRs are underlined, and slashes indicate boundaries between the variable region and other chain components (e.g., constant regions and domain linkers). Note that αENPP3×αCD3 bsAbs can utilize variable regions, Fc regions, and constant domains 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. Additionally, each sequence outlined herein can include or exclude the M428L / N434S variant in one or, preferably, both Fc domains, which results in a longer half-life in serum. [Figure 23B] The sequence of an exemplary αENPP3×αCD3 bsAb is shown, comprising the L1.47_H1.89 anti-CD3 scFv (also known as CD3 High-Int#2 [VLVH]) in a 2+1 Fab2-scFv-Fc format. CDRs are underlined, and slashes indicate boundaries between the variable region and other chain components (e.g., constant regions and domain linkers). Note that αENPP3×αCD3 bsAbs can utilize variable regions, Fc regions, and constant domains 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. Additionally, each sequence outlined herein can include or exclude the M428L / N434S variant in one or, preferably, both Fc domains, which results in a longer half-life in serum. [Figure 23C]The sequence of an exemplary αENPP3×αCD3 bsAb is shown, comprising the L1.47_H1.89 anti-CD3 scFv (also known as CD3 High-Int#2 [VLVH]) in a 2+1 Fab2-scFv-Fc format. CDRs are underlined, and slashes indicate boundaries between the variable region and other chain components (e.g., constant regions and domain linkers). Note that αENPP3×αCD3 bsAbs can utilize variable regions, Fc regions, and constant domains 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. Additionally, each sequence outlined herein can include or exclude the M428L / N434S variant in one or, preferably, both Fc domains, which results in a longer half-life in serum. [Figure 23D] The sequence of an exemplary αENPP3×αCD3 bsAb is shown, comprising the L1.47_H1.89 anti-CD3 scFv (also known as CD3 High-Int#2 [VLVH]) in a 2+1 Fab2-scFv-Fc format. CDRs are underlined, and slashes indicate boundaries between the variable region and other chain components (e.g., constant regions and domain linkers). Note that αENPP3×αCD3 bsAbs can utilize variable regions, Fc regions, and constant domains 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. Additionally, each sequence outlined herein can include or exclude the M428L / N434S variant in one or, preferably, both Fc domains, which results in a longer half-life in serum. [Figure 23E]The sequence of an exemplary αENPP3×αCD3 bsAb is shown, comprising the L1.47_H1.89 anti-CD3 scFv (also known as CD3 High-Int#2 [VLVH]) in a 2+1 Fab2-scFv-Fc format. CDRs are underlined, and slashes indicate boundaries between the variable region and other chain components (e.g., constant regions and domain linkers). Note that αENPP3×αCD3 bsAbs can utilize variable regions, Fc regions, and constant domains 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. Additionally, each sequence outlined herein can include or exclude the M428L / N434S variant in one or, preferably, both Fc domains, which results in a longer half-life in serum. [Figure 24A] Figure 1 shows induction of RTCC in CFSE-labeled KU812 cells, as indicated by A) the decrease in the number of CFSE+ KU812 cells and B) the percentage of CFSE+ KU812 cells stained with Zombie Aqua, after 24 hours of incubation of CFSE-labeled KU812 human PBMCs (10:1 effector to target cell ratio) with αENPP3 x αCD3 bispecific antibodies (XENP26820, XENP26821, XENP28287, and XENP28390). Controls used were αRSV x αCD3 bispecific antibody (XENP13245), effector and target cells only, and target cells only. Taken together, the data show that the prototype αENPP3 × αCD3 bsAb induced redirected T cell cytotoxicity (RTCC) on KU812 cells in a dose-dependent manner, that CD3 binding affinity correlated with RTCC potency (i.e., bsAbs containing CD3 High induced RTCC more potently than bsAbs containing CD3 High-Int#1), and that bsAbs containing the AN1-binding domain induced RTCC more potently than bsAbs containing the H16-7.8-binding domain. [Figure 24B]Figure 1 shows induction of RTCC in CFSE-labeled KU812 cells, as indicated by A) the decrease in the number of CFSE+ KU812 cells and B) the percentage of CFSE+ KU812 cells stained with Zombie Aqua, after 24 hours of incubation of CFSE-labeled KU812 human PBMCs (10:1 effector to target cell ratio) with αENPP3 x αCD3 bispecific antibodies (XENP26820, XENP26821, XENP28287, and XENP28390). Controls used were αRSV x αCD3 bispecific antibody (XENP13245), effector and target cells only, and target cells only. Taken together, the data show that the prototype αENPP3 × αCD3 bsAb induced redirected T cell cytotoxicity (RTCC) on KU812 cells in a dose-dependent manner, that CD3 binding affinity correlated with RTCC potency (i.e., bsAbs containing CD3 High induced RTCC more potently than bsAbs containing CD3 High-Int#1), and that bsAbs containing the AN1-binding domain induced RTCC more potently than bsAbs containing the H16-7.8-binding domain. [Figure 25A]Figure 1 shows CD4+ T cell activation as indicated by A) CD107a MFI on CD4+ T cells, B) CD25 MFI on CD4+ T cells, and C) CD69 MFI on CD4+ T cells after 24 hours of incubation of CFSE-labeled KU812 human PBMCs (10:1 effector to target cell ratio) with αENPP3 x αCD3 bispecific antibodies (XENP26820, XENP26821, XENP28287, and XENP28390). Controls used were αRSV x αCD3 bispecific antibody (XENP13245), effector and target cells only, and target cells only. Consistent with the RTCC data, αENPP3×αCD3 bsAb induced CD4+ T cell activation in a dose-dependent manner, CD3 binding affinity correlated with activation potency (i.e., bsAbs containing CD3 High induced CD4+ T cell activation more potently than bsAbs containing CD3 High-Int#1), and bsAbs containing the AN1-binding domain induced CD4+ T cell activation more potently than bsAbs containing the H16-7.8-binding domain. [Figure 25B] Figure 1 shows CD4+ T cell activation as indicated by A) CD107a MFI on CD4+ T cells, B) CD25 MFI on CD4+ T cells, and C) CD69 MFI on CD4+ T cells after 24 hours of incubation of CFSE-labeled KU812 human PBMCs (10:1 effector to target cell ratio) with αENPP3 x αCD3 bispecific antibodies (XENP26820, XENP26821, XENP28287, and XENP28390). Controls used were αRSV x αCD3 bispecific antibody (XENP13245), effector and target cells only, and target cells only. Consistent with the RTCC data, αENPP3×αCD3 bsAb induced CD4+ T cell activation in a dose-dependent manner, CD3 binding affinity correlated with activation potency (i.e., bsAbs containing CD3 High induced CD4+ T cell activation more potently than bsAbs containing CD3 High-Int#1), and bsAbs containing the AN1-binding domain induced CD4+ T cell activation more potently than bsAbs containing the H16-7.8-binding domain. [Figure 25C] Figure 1 shows CD4+ T cell activation as indicated by A) CD107a MFI on CD4+ T cells, B) CD25 MFI on CD4+ T cells, and C) CD69 MFI on CD4+ T cells after 24 hours of incubation of CFSE-labeled KU812 human PBMCs (10:1 effector to target cell ratio) with αENPP3 x αCD3 bispecific antibodies (XENP26820, XENP26821, XENP28287, and XENP28390). Controls used were αRSV x αCD3 bispecific antibody (XENP13245), effector and target cells only, and target cells only. Consistent with the RTCC data, αENPP3×αCD3 bsAb induced CD4+ T cell activation in a dose-dependent manner, CD3 binding affinity correlated with activation potency (i.e., bsAbs containing CD3 High induced CD4+ T cell activation more potently than bsAbs containing CD3 High-Int#1), and bsAbs containing the AN1-binding domain induced CD4+ T cell activation more potently than bsAbs containing the H16-7.8-binding domain. [Figure 26A] Activated CD8+ T cells as indicated by A) CD107a MFI on CD8+ T cells, B) CD25 MFI on CD8+ T cells, and C) CD69 MFI on CD8+ T cells after 24 hours of incubation with CFSE-labeled KU812 human PBMCs (10:1 effector to target cell ratio) and αENPP3 x αCD3 bispecific antibodies (XENP26820, XENP26821, XENP28287, and XENP28390). Controls used were αRSV x αCD3 bispecific antibody (XENP13245), effector and target cells only, and target cells only. Consistent with the RTCC data, the αENPP3×αCD3 bsAb induced CD8+ T cell activation in a dose-dependent manner, CD3 binding affinity correlated with activation potency (i.e., bsAbs containing CD3 High induced CD8+ T cell activation more potently than bsAbs containing CD3 High-Int#1), and bsAbs containing the AN1-binding domain induced CD8+ T cell activation more potently than bsAbs containing the H16-7.8-binding domain. [Figure 26B] Activated CD8+ T cells as indicated by A) CD107a MFI on CD8+ T cells, B) CD25 MFI on CD8+ T cells, and C) CD69 MFI on CD8+ T cells after 24 hours of incubation with CFSE-labeled KU812 human PBMCs (10:1 effector to target cell ratio) and αENPP3 x αCD3 bispecific antibodies (XENP26820, XENP26821, XENP28287, and XENP28390). Controls used were αRSV x αCD3 bispecific antibody (XENP13245), effector and target cells only, and target cells only. Consistent with the RTCC data, the αENPP3×αCD3 bsAb induced CD8+ T cell activation in a dose-dependent manner, CD3 binding affinity correlated with activation potency (i.e., bsAbs containing CD3 High induced CD8+ T cell activation more potently than bsAbs containing CD3 High-Int#1), and bsAbs containing the AN1-binding domain induced CD8+ T cell activation more potently than bsAbs containing the H16-7.8-binding domain. [Figure 26C]Activated CD8+ T cells as indicated by A) CD107a MFI on CD8+ T cells, B) CD25 MFI on CD8+ T cells, and C) CD69 MFI on CD8+ T cells after 24 hours of incubation with CFSE-labeled KU812 human PBMCs (10:1 effector to target cell ratio) and αENPP3 x αCD3 bispecific antibodies (XENP26820, XENP26821, XENP28287, and XENP28390). Controls used were αRSV x αCD3 bispecific antibody (XENP13245), effector and target cells only, and target cells only. Consistent with the RTCC data, the αENPP3×αCD3 bsAb induced CD8+ T cell activation in a dose-dependent manner, CD3 binding affinity correlated with activation potency (i.e., bsAbs containing CD3 High induced CD8+ T cell activation more potently than bsAbs containing CD3 High-Int#1), and bsAbs containing the AN1-binding domain induced CD8+ T cell activation more potently than bsAbs containing the H16-7.8-binding domain. [Figure 27A]Figure 1 shows induction of RTCC in CFSE-labeled RXF393 cells, as indicated by A) the decrease in the number of CFSE+ RXF393 cells and B) the percentage of CFSE+ RXF393 cells stained with Zombie Aqua, after 24 hours of incubation of CFSE-labeled RXF393 with human PBMCs (effector:target cell ratio of 20:1) and αENPP3 x αCD3 bispecific antibodies (XENP26820, XENP26821, XENP28287, and XENP28390). Controls used were αRSV x αCD3 bispecific antibody (XENP13245), effector and target cells only, and target cells only. Consistent with the data for KU812 cells, the data show that the prototype αENPP3×αCD3 bsAb induced redirected T cell cytotoxicity (RTCC) on RXF393 cells in a dose-dependent manner, CD3 binding affinity correlated with RTCC potency (i.e., bsAbs containing CD3 High induced RTCC more potently than bsAbs containing CD3 High-Int#1), and bsAbs containing the AN1-binding domain induced RTCC more potently than bsAbs containing the H16-7.8-binding domain. [Figure 27B]Figure 1 shows induction of RTCC in CFSE-labeled RXF393 cells, as indicated by A) the decrease in the number of CFSE+ RXF393 cells and B) the percentage of CFSE+ RXF393 cells stained with Zombie Aqua, after 24 hours of incubation of CFSE-labeled RXF393 with human PBMCs (effector:target cell ratio of 20:1) and αENPP3 x αCD3 bispecific antibodies (XENP26820, XENP26821, XENP28287, and XENP28390). Controls used were αRSV x αCD3 bispecific antibody (XENP13245), effector and target cells only, and target cells only. Consistent with the data for KU812 cells, the data show that the prototype αENPP3×αCD3 bsAb induced redirected T cell cytotoxicity (RTCC) on RXF393 cells in a dose-dependent manner, CD3 binding affinity correlated with RTCC potency (i.e., bsAbs containing CD3 High induced RTCC more potently than bsAbs containing CD3 High-Int#1), and bsAbs containing the AN1-binding domain induced RTCC more potently than bsAbs containing the H16-7.8-binding domain. [Figure 28A]Figure 1 shows CD4+ T cell activation as indicated by A) CD107a MFI on CD4+ T cells, B) CD25 MFI on CD4+ T cells, and C) CD69 MFI on CD4+ T cells after 24 h incubation of CFSE-labeled RXF393 human PBMCs (effector to target cell ratio of 20:1) with αENPP3 x αCD3 bispecific antibodies (XENP26820, XENP26821, XENP28287, and XENP28390). Controls used were αRSV x αCD3 bispecific antibody (XENP13245), effector and target cells only, and target cells only. Consistent with the RTCC data, the αENPP3×αCD3 bsAb induced CD4+ T cell activation in a dose-dependent manner, CD3 binding affinity correlated with activation potency (i.e., bsAbs containing CD3 High induced CD4+ T cell activation more potently than bsAbs containing CD3 High-Int#1), and bsAbs containing the AN1-binding domain induced CD4+ T cell activation more potently than bsAbs containing the H16-7.8-binding domain. [Figure 28B] Figure 1 shows CD4+ T cell activation as indicated by A) CD107a MFI on CD4+ T cells, B) CD25 MFI on CD4+ T cells, and C) CD69 MFI on CD4+ T cells after 24 h incubation of CFSE-labeled RXF393 human PBMCs (effector to target cell ratio of 20:1) with αENPP3 x αCD3 bispecific antibodies (XENP26820, XENP26821, XENP28287, and XENP28390). Controls used were αRSV x αCD3 bispecific antibody (XENP13245), effector and target cells only, and target cells only. Consistent with the RTCC data, the αENPP3×αCD3 bsAb induced CD4+ T cell activation in a dose-dependent manner, CD3 binding affinity correlated with activation potency (i.e., bsAbs containing CD3 High induced CD4+ T cell activation more potently than bsAbs containing CD3 High-Int#1), and bsAbs containing the AN1-binding domain induced CD4+ T cell activation more potently than bsAbs containing the H16-7.8-binding domain. [Figure 28C] Figure 1 shows CD4+ T cell activation as indicated by A) CD107a MFI on CD4+ T cells, B) CD25 MFI on CD4+ T cells, and C) CD69 MFI on CD4+ T cells after 24 h incubation of CFSE-labeled RXF393 human PBMCs (effector to target cell ratio of 20:1) with αENPP3 x αCD3 bispecific antibodies (XENP26820, XENP26821, XENP28287, and XENP28390). Controls used were αRSV x αCD3 bispecific antibody (XENP13245), effector and target cells only, and target cells only. Consistent with the RTCC data, the αENPP3×αCD3 bsAb induced CD4+ T cell activation in a dose-dependent manner, CD3 binding affinity correlated with activation potency (i.e., bsAbs containing CD3 High induced CD4+ T cell activation more potently than bsAbs containing CD3 High-Int#1), and bsAbs containing the AN1-binding domain induced CD4+ T cell activation more potently than bsAbs containing the H16-7.8-binding domain. [Figure 29A] Activated CD8+ T cells as indicated by A) CD107a MFI on CD8+ T cells, B) CD25 MFI on CD8+ T cells, and C) CD69 MFI on CD8+ T cells after 24 hours of incubation of CFSE-labeled RXF393 human PBMCs (20:1 effector to target cell ratio) with αENPP3 x αCD3 bispecific antibodies (XENP26820, XENP26821, XENP28287, and XENP28390). Controls used were αRSV x αCD3 bispecific antibody (XENP13245), effector and target cells only, and target cells only. Consistent with the RTCC data, the αENPP3×αCD3 bsAb induced CD8+ T cell activation in a dose-dependent manner, CD3 binding affinity correlated with activation potency (i.e., bsAbs containing CD3High induced CD8+ T cell activation more potently than bsAbs containing CD3High-Int#1), and bsAbs containing the AN1-binding domain induced CD8+ T cell activation more potently than bsAbs containing the H16-7.8-binding domain. [Figure 29B] Activated CD8+ T cells as indicated by A) CD107a MFI on CD8+ T cells, B) CD25 MFI on CD8+ T cells, and C) CD69 MFI on CD8+ T cells after 24 hours of incubation of CFSE-labeled RXF393 human PBMCs (20:1 effector to target cell ratio) with αENPP3 x αCD3 bispecific antibodies (XENP26820, XENP26821, XENP28287, and XENP28390). Controls used were αRSV x αCD3 bispecific antibody (XENP13245), effector and target cells only, and target cells only. Consistent with the RTCC data, the αENPP3×αCD3 bsAb induced CD8+ T cell activation in a dose-dependent manner, CD3 binding affinity correlated with activation potency (i.e., bsAbs containing CD3High induced CD8+ T cell activation more potently than bsAbs containing CD3High-Int#1), and bsAbs containing the AN1-binding domain induced CD8+ T cell activation more potently than bsAbs containing the H16-7.8-binding domain. [Figure 29C]Activated CD8+ T cells as indicated by A) CD107a MFI on CD8+ T cells, B) CD25 MFI on CD8+ T cells, and C) CD69 MFI on CD8+ T cells after 24 hours of incubation of CFSE-labeled RXF393 human PBMCs (20:1 effector to target cell ratio) with αENPP3 x αCD3 bispecific antibodies (XENP26820, XENP26821, XENP28287, and XENP28390). Controls used were αRSV x αCD3 bispecific antibody (XENP13245), effector and target cells only, and target cells only. Consistent with the RTCC data, the αENPP3×αCD3 bsAb induced CD8+ T cell activation in a dose-dependent manner, CD3 binding affinity correlated with activation potency (i.e., bsAbs containing CD3High induced CD8+ T cell activation more potently than bsAbs containing CD3High-Int#1), and bsAbs containing the AN1-binding domain induced CD8+ T cell activation more potently than bsAbs containing the H16-7.8-binding domain. [Figure 30] A) Chromatograms showing the purity and homogeneity of XENP28287 purification part 2 (protein A chromatography followed by cation exchange chromatography) and peaks B and BC isolated from the cation exchange separation shown in Figure 30A (as well as pre-purified material), B) analytical size exclusion chromatography with multi-angle light scattering (aSEC-MALS) and C) analytical cation exchange chromatography (aCIEX). Figure 30B also shows the molecular weights of the protein species in the peaks as determined by multi-angle light scattering. [Figure 31]A) Chromatograms showing the purity and homogeneity of XENP28925 purification part 2 (protein A chromatography followed by cation exchange chromatography) and peak B isolated from the cation exchange separation shown in Figure 31A (as well as pre-purified material), B) analytical size exclusion chromatography with multi-angle light scattering (aSEC-MALS), and C) analytical cation exchange chromatography (aCIEX). Figure 31B also shows the molecular weight of the protein species in the peak as determined by multi-angle light scattering. [Figure 32] A) Chromatogram showing part 2 of the purification of XENP31149 (protein A chromatography followed by cation exchange chromatography), and B) chromatogram showing the identity of peaks A and B isolated from the cation exchange separation shown in Figure XA (as well as material pre-purified by analytical size-exclusion chromatography with multi-angle light scattering (aSEC-MALS)). [Figure 33] A) Chromatogram showing part 2 of the purification of XENP31419 (protein A chromatography followed by cation exchange chromatography), and B) chromatogram showing the identity of peaks A and B isolated from the cation exchange separation shown in Figure XA (as well as material pre-purified by analytical size-exclusion chromatography with multi-angle light scattering (aSEC-MALS)). [Figure 34] Induction of RTCC in CFSE-labeled KU812 (solid line, ENPP3 high) or CFSE-labeled RCC4 (dashed line, ENPP3 low) cells, as indicated by the percentage of CFSE+ cells stained with Zombie Aqua, is shown after 18 hours of incubation of CFSE-labeled target cells with human PBMCs (effector:target cell ratio of 10:1) and the αENPP3 × αCD3 bispecific antibody XENP28925. [Figure 35] Binding of affinity engineered αENPP3×αCD31+1bsAb to ENPP3-high KU812 cells is shown. [Figure 36]Figure 1 shows the induction of RTCC in CFSE-labeled KU812 (solid line, ENPP3 high) or CFSE-labeled RCC4 (dashed line, ENPP3 low) cells, as indicated by the percentage of CFSE+ cells stained with Zombie Aqua, after 42 hours of incubation of CFSE-labeled target cells with human PBMCs (effector-to-target cell ratio of 10:1) and the αENPP3 x αCD3 bispecific antibodies XENP28925 (WT, high ENPP3 binding), XENP29516 (intermediate ENPP3 binding), or XENP30262 (low ENPP3 binding). The data show that both XENP29516 and XENP30262 were substantially less potent at inducing RTCC in ENPP3-low RCC4 cells compared to XENP28925, and that RTCC potency correlates with binding potency as described above. XENP29516 and XENP30262 also showed less potent induction of RTCC in ENPP3-high cells. [Figure 37A] Figure 1 shows induction of A) IFNγ, B) IL-6, and C) TNFα release by human PBMCs incubated for 18 hours with KU812 cells (10:1 effector-to-target cell ratio) and the αENPP3 x αCD3 bispecific antibodies XENP28925 (CD3 High) or XENP29436 (CD3 High-Int#1). The data show that XENP29436 is substantially less potent at inducing cytokine release than XENP28925. [Figure 37B] Figure 1 shows induction of A) IFNγ, B) IL-6, and C) TNFα release by human PBMCs incubated for 18 hours with KU812 cells (10:1 effector-to-target cell ratio) and the αENPP3 x αCD3 bispecific antibodies XENP28925 (CD3 High) or XENP29436 (CD3 High-Int#1). The data show that XENP29436 is substantially less potent at inducing cytokine release than XENP28925. [Figure 37C]Figure 1 shows induction of A) IFNγ, B) IL-6, and C) TNFα release by human PBMCs incubated for 18 hours with KU812 cells (10:1 effector-to-target cell ratio) and the αENPP3 x αCD3 bispecific antibodies XENP28925 (CD3 High) or XENP29436 (CD3 High-Int#1). The data show that XENP29436 is substantially less potent at inducing cytokine release than XENP28925. [Figure 38] Figure 1 shows induction of IFNγ release by human PBMCs incubated for 18 hours with RCC4 cells (10:1 effector-to-target cell ratio) and the αENPP3 x αCD3 bispecific antibodies XENP28925 (CD3 High) or XENP29436 (CD3 High-Int#1). The data show that XENP29436 demonstrated negligible induction of cytokine release compared to XENP28925 in the presence of ENPP3-low RCC4 cells. [Figure 39] Figure 1 shows the induction of RTCC in CFSE-labeled KU812 (solid line, ENPP3 high) or CFSE-labeled RCC4 (dashed line, RCC4 low) cells, as indicated by the percentage of CFSE+ cells stained with Zombie Aqua, after 42 hours of incubation of CFSE-labeled target cells with human PBMCs (effector-to-target cell ratio of 10:1) and the αENPP3 x αCD3 bispecific antibodies XENP28925 (CD3 high) or XENP29436 (CD3 high-Int#1). The data demonstrate that XENP29436 is substantially less potent at inducing RTCC in ENPP3 low cells compared to XENP28925, but that XENP29436 also demonstrated reduced potency in inducing RTCC in ENPP3 high cells. [Figure 40]Figure 1 shows the induction of IFNγ release by human PBMCs incubated with KU812 cells (10:1 effector-to-target cell ratio) and the αENPP3 x αCD3 bispecific antibodies XENP28925 (ENPP3 High; CD3 High), XENP29436 (ENPP3 High; CD3 High-Int#1), XENP29518 (ENPP3 Intermediate; CD3 High), XENP29463 (ENPP3 Intermediate; CD3 High-Int#1), XENP30262 (ENPP3 Low; CD3 High), or XENP30263 (ENPP3 Low; CD3 High-Int#1). The data show that reducing the avidity of CD3 or ENPP3 reduces the induction of cytokine release. Notably, reducing the avidity of CD3 and ENPP3 further reduces the induction of cytokine release. [Figure 41] Figure 1 shows induction of RTCC in CFSE-labeled KU812 (solid line, ENPP3 high) or CFSE-labeled RCC4 (dashed line, ENPP3 low) cells, as indicated by the percentage of CFSE+ cells stained with Zombie Aqua, after 42 hours of incubation with CFSE-labeled target cells (10:1 effector-to-target ratio) and the αENPP3 × αCD3 bispecific antibodies XENP28925 (WT, high ENPP3 binding; CD3 high; monovalent ENPP3 binding), XENP29516 (intermediate ENPP3 binding; CD3 high; monovalent ENPP3 binding), or XENP29520 (intermediate ENPP3 binding; CD3 high; bivalent ENPP3 binding). The data show that bivalent binding (with intermediate ENPP3 binding) maintained the reduced RTCC potency in ENPP3-low cells but restored RTCC potency in ENPP3-high cells close to that shown by XENP28925. [Figure 42]Figure 1 shows the induction of RTCC in CFSE-labeled KU812 (solid line, ENPP3 high) or CFSE-labeled RCC4 (dashed line, ENPP3 low) cells, as indicated by the percentage of CFSE cells stained with Zombie Aqua, after 42 hours of incubation of CFSE-labeled target cells with human PBMCs (10:1 effector-to-target cell ratio) and the αENPP3 × αCD3 bispecific antibodies XENP28925 (WT, high ENPP3 binding; CD3 high; monovalent ENPP3 binding), XENP30262 (low ENPP3 binding; CD3 high; monovalent ENPP3 binding), or XENP30264 (low ENPP3 binding; CD3 high; bivalent ENPP3 binding). The data show that bivalent binding (with low ENPP3 binding) further reduced RTCC potency in ENPP3 low cells and restored some RTCC potency in ENPP3 high cells. [Figure 43] Figure 1 shows the induction of RTCC in CFSE-labeled KU812 cells, as indicated by the percentage of CFSE+ cells stained with Zombie Aqua, after 44 hours of incubation of CFSE-labeled target cells with human PBMCs (10:1 effector-to-target ratio) and the αENPP3 × αCD3 bispecific antibodies XENP28925 (CD3 High; monovalent ENPP3 binding), XENP29437 (CD3 High; bivalent ENPP3 binding), XENP29436 (CD3 High-Int#1; monovalent ENPP3 binding), or XENP29438 (CD3 High-Int#1; bivalent ENPP3 binding). Unexpectedly, XENP29438 failed to induce RTCC in KU812 cells. [Figure 44]Figure 1 shows induction of RTCC with CFSE-labeled KU812 (solid line, ENPP3 high) or CFSE-labeled RCC4 (dashed line, ENPP3 low), as indicated by the percentage of CFSE+ cells stained with Zombie Aqua, after 24 hours of incubation with CFSE-labeled target cells (10:1 effector-to-target ratio) of human PBMCs and the αENPP3 x αCD3 bispecific antibodies XENP294377 (CD3 High VH / VL; bivalent ENPP3 binding), XENP30469 (CD3 High VL / VH; bivalent ENPP3 binding), XENP29428 (CD3 High-Int#1 VH / VL; bivalent ENPP3 binding), or XENP30470 (CD3 High-Int#2 VL / VH; bivalent ENPP3 binding). The data showed that swapping the orientation of the variable heavy and light domains of CD3 High-Int#1 scFv restored its activity in the context of a 2+1 Fab2-scFv-Fc bsAb format (XENP29438 vs. XENP30470). Swapping the orientation of the variable heavy and light domains of CD3 High scFv allowed for a much more modest improvement in RTCC potency in the context of a 2+1 Fab2-scFv-Fc bsAb format (XENP29437 vs. XENP30469). [Figure 45]CFSE-labeled target cells were human PBMCs (effector-to-target cell ratio of 10:1) and the αENPP3 × αCD3 bispecific antibodies XENP29520 (CD3 High [VH / VL]; bivalent ENPP3 intermediate binding), XENP30819 (CD3 High-Int#1 [VL / VHL]; bivalent ENPP3 intermediate binding), XENP31149 (CD3 High-Int#2 [VL / VHL]; bivalent ENPP3 intermediate binding), XENP30264 (CD3 High [VH / VL] bivalent ENPP3 low binding), XENP30821 (CD3 High-Int#1 [VL / VHL]; bivalent ENPP3 low binding), or XENP31150 (CD3 Figure 1 shows the induction of RTCCs with CFSE-labeled KU812 (solid line, ENPP3 high) or CFSE-labeled RCC4 (dashed line, ENPP3 low) after 40 hours of incubation with High-Int#2 [VL / VHL]; bivalent ENPP3 low binding), as indicated by the percentage of CFSE+ cells stained with Zombie Aqua. [Figure 46] The sequences of XENP16432, an anti-PD-1 mAb based on nivolumab and an IgG1 backbone with E233P / L234V / L235A / G236del / S267K deletion variants, and XENP21461 (pembrolizumab) are shown. [Figure 47] Figure 1 shows the time course of tumor volume (determined by caliper measurement) in NSG mice implanted with KU812 and huPBMCs treated with PBS, XENP16432 (a bivalent anti-PD-1 mAb), or exemplary αENPP3×αCD3 bsAbs (XENP30819, XENP30821, or XENP31419) alone or in combination with XENP16432. Each αENPP3×αCD3 bsAb was able to enhance the allogeneic anti-tumor effect of T cells against KU812 cells at low and / or high doses and was successfully combined with PD-1 blockade. [Figure 48A]Figure 1 shows the proliferation of A) CD45+ lymphocytes, B) CD8+ T cells, and C) CD4+ T cells by day 14 in the blood of NSG mice engrafted with KU812 and huPBMCs treated with PBS, XENP16432 (a bivalent anti-PD-1 mAb), or exemplary αENPP3xαCD3 2+1 bsAbs (XENP30819, XENP30821, or XENP31419) alone or in combination with XENP16432. In all cases, lymphocyte proliferation was enhanced when combined with PD-1 blockade. [Figure 48B] Figure 1 shows the proliferation of A) CD45+ lymphocytes, B) CD8+ T cells, and C) CD4+ T cells by day 14 in the blood of NSG mice engrafted with KU812 and huPBMCs treated with PBS, XENP16432 (a bivalent anti-PD-1 mAb), or exemplary αENPP3xαCD3 2+1 bsAbs (XENP30819, XENP30821, or XENP31419) alone or in combination with XENP16432. In all cases, lymphocyte proliferation was enhanced when combined with PD-1 blockade. [Figure 48C] Figure 1 shows the proliferation of A) CD45+ lymphocytes, B) CD8+ T cells, and C) CD4+ T cells by day 14 in the blood of NSG mice engrafted with KU812 and huPBMCs treated with PBS, XENP16432 (a bivalent anti-PD-1 mAb), or exemplary αENPP3xαCD3 2+1 bsAbs (XENP30819, XENP30821, or XENP31419) alone or in combination with XENP16432. In all cases, lymphocyte proliferation was enhanced when combined with PD-1 blockade. [Figure 49]Figure 1 shows the time course of tumor volume (determined by caliper measurement) in NSG mice implanted with RXF-393 and huPBMCs treated with PBS, XENP16432 (a bivalent anti-PD-1 mAb), or exemplary αENPP3×αCD3 bsAbs (XENP30819, or XENP31419) alone or in combination with XENP16432. Each αENPP3×αCD3 bsAb was able to enhance the allogeneic antitumor effect of T cells against KU812 cells at low, medium, and / or high doses and was successfully combined with PD-1 blockade. [Figure 50A] Figure 1 shows the time course of tumor volume (determined by caliper measurement) in individual KU812 and huPBMC-implanted NSG mice treated with A) PBS, B) XENP16432 (a bivalent anti-PD-1 mAb), or exemplary αENPP3×αCD3 bsAbs (XENP30819, XENP30821, or XENP31419) alone or in combination with XENP16432. Each αENPP3×αCD3 bsAb was able to enhance the allogeneic anti-tumor effect of T cells against KU812 cells at low and / or high doses and successfully combined with PD-1 blockade. [Figure 50B] Figure 1 shows the time course of tumor volume (determined by caliper measurement) in individual KU812 and huPBMC-implanted NSG mice treated with A) PBS, B) XENP16432 (a bivalent anti-PD-1 mAb), or exemplary αENPP3×αCD3 bsAbs (XENP30819, XENP30821, or XENP31419) alone or in combination with XENP16432. Each αENPP3×αCD3 bsAb was able to enhance the allogeneic anti-tumor effect of T cells against KU812 cells at low and / or high doses and successfully combined with PD-1 blockade. [Figure 50C]Figure 1 shows the time course of tumor volume (determined by caliper measurement) in individual KU812 and huPBMC-implanted NSG mice treated with A) PBS, B) XENP16432 (a bivalent anti-PD-1 mAb), or exemplary αENPP3×αCD3 bsAbs (XENP30819, XENP30821, or XENP31419) alone or in combination with XENP16432. Each αENPP3×αCD3 bsAb was able to enhance the allogeneic anti-tumor effect of T cells against KU812 cells at low and / or high doses and successfully combined with PD-1 blockade. [Figure 50D] Figure 1 shows the time course of tumor volume (determined by caliper measurement) in individual KU812 and huPBMC-implanted NSG mice treated with A) PBS, B) XENP16432 (a bivalent anti-PD-1 mAb), or exemplary αENPP3×αCD3 bsAbs (XENP30819, XENP30821, or XENP31419) alone or in combination with XENP16432. Each αENPP3×αCD3 bsAb was able to enhance the allogeneic anti-tumor effect of T cells against KU812 cells at low and / or high doses and successfully combined with PD-1 blockade. [Figure 50E] Figure 1 shows the time course of tumor volume (determined by caliper measurement) in individual KU812 and huPBMC-implanted NSG mice treated with A) PBS, B) XENP16432 (a bivalent anti-PD-1 mAb), or exemplary αENPP3×αCD3 bsAbs (XENP30819, XENP30821, or XENP31419) alone or in combination with XENP16432. Each αENPP3×αCD3 bsAb was able to enhance the allogeneic anti-tumor effect of T cells against KU812 cells at low and / or high doses and successfully combined with PD-1 blockade. [Figure 50F]Figure 1 shows the time course of tumor volume (determined by caliper measurement) in individual KU812 and huPBMC-implanted NSG mice treated with A) PBS, B) XENP16432 (a bivalent anti-PD-1 mAb), or exemplary αENPP3×αCD3 bsAbs (XENP30819, XENP30821, or XENP31419) alone or in combination with XENP16432. Each αENPP3×αCD3 bsAb was able to enhance the allogeneic anti-tumor effect of T cells against KU812 cells at low and / or high doses and successfully combined with PD-1 blockade. [Figure 50G] Figure 1 shows the time course of tumor volume (determined by caliper measurement) in individual KU812 and huPBMC-implanted NSG mice treated with A) PBS, B) XENP16432 (a bivalent anti-PD-1 mAb), or exemplary αENPP3×αCD3 bsAbs (XENP30819, XENP30821, or XENP31419) alone or in combination with XENP16432. Each αENPP3×αCD3 bsAb was able to enhance the allogeneic anti-tumor effect of T cells against KU812 cells at low and / or high doses and successfully combined with PD-1 blockade. [Figure 50H] Figure 1 shows the time course of tumor volume (determined by caliper measurement) in individual KU812 and huPBMC-implanted NSG mice treated with A) PBS, B) XENP16432 (a bivalent anti-PD-1 mAb), or exemplary αENPP3×αCD3 bsAbs (XENP30819, XENP30821, or XENP31419) alone or in combination with XENP16432. Each αENPP3×αCD3 bsAb was able to enhance the allogeneic anti-tumor effect of T cells against KU812 cells at low and / or high doses and successfully combined with PD-1 blockade. [Figure 50I]Figure 1 shows the time course of tumor volume (determined by caliper measurement) in individual KU812 and huPBMC-implanted NSG mice treated with A) PBS, B) XENP16432 (a bivalent anti-PD-1 mAb), or exemplary αENPP3×αCD3 bsAbs (XENP30819, XENP30821, or XENP31419) alone or in combination with XENP16432. Each αENPP3×αCD3 bsAb was able to enhance the allogeneic anti-tumor effect of T cells against KU812 cells at low and / or high doses and successfully combined with PD-1 blockade. [Figure 50J] Figure 1 shows the time course of tumor volume (determined by caliper measurement) in individual KU812 and huPBMC-implanted NSG mice treated with A) PBS, B) XENP16432 (a bivalent anti-PD-1 mAb), or exemplary αENPP3×αCD3 bsAbs (XENP30819, XENP30821, or XENP31419) alone or in combination with XENP16432. Each αENPP3×αCD3 bsAb was able to enhance the allogeneic anti-tumor effect of T cells against KU812 cells at low and / or high doses and successfully combined with PD-1 blockade. [Figure 50K] Figure 1 shows the time course of tumor volume (determined by caliper measurement) in individual KU812 and huPBMC-implanted NSG mice treated with A) PBS, B) XENP16432 (a bivalent anti-PD-1 mAb), or exemplary αENPP3×αCD3 bsAbs (XENP30819, XENP30821, or XENP31419) alone or in combination with XENP16432. Each αENPP3×αCD3 bsAb was able to enhance the allogeneic anti-tumor effect of T cells against KU812 cells at low and / or high doses and successfully combined with PD-1 blockade. [Figure 50L]Figure 1 shows the time course of tumor volume (determined by caliper measurement) in individual KU812 and huPBMC-implanted NSG mice treated with A) PBS, B) XENP16432 (a bivalent anti-PD-1 mAb), or exemplary αENPP3×αCD3 bsAbs (XENP30819, XENP30821, or XENP31419) alone or in combination with XENP16432. Each αENPP3×αCD3 bsAb was able to enhance the allogeneic anti-tumor effect of T cells against KU812 cells at low and / or high doses and successfully combined with PD-1 blockade. [Figure 50M] Figure 1 shows the time course of tumor volume (determined by caliper measurement) in individual KU812 and huPBMC-implanted NSG mice treated with A) PBS, B) XENP16432 (a bivalent anti-PD-1 mAb), or exemplary αENPP3×αCD3 bsAbs (XENP30819, XENP30821, or XENP31419) alone or in combination with XENP16432. Each αENPP3×αCD3 bsAb was able to enhance the allogeneic anti-tumor effect of T cells against KU812 cells at low and / or high doses and successfully combined with PD-1 blockade. [Figure 50N] Figure 1 shows the time course of tumor volume (determined by caliper measurement) in individual KU812 and huPBMC-implanted NSG mice treated with A) PBS, B) XENP16432 (a bivalent anti-PD-1 mAb), or exemplary αENPP3×αCD3 bsAbs (XENP30819, XENP30821, or XENP31419) alone or in combination with XENP16432. Each αENPP3×αCD3 bsAb was able to enhance the allogeneic anti-tumor effect of T cells against KU812 cells at low and / or high doses and successfully combined with PD-1 blockade. [Figure 51A]Figure 1 shows the time course of tumor volume (determined by caliper measurement) in individual RXF-393 and huPBMC-implanted NSG mice treated with A) PBS, B) XENP16432 (a bivalent anti-PD-1 mAb), or exemplary αENPP3×αCD3 bsAbs (XENP30819, or XENP31419) alone or in combination with XENP16432. Each αENPP3×αCD3 bsAb was able to enhance the allogeneic antitumor effect of T cells against KU812 cells at low, medium, and / or high doses and successfully combined with PD-1 blockade. [Figure 51B] Figure 1 shows the time course of tumor volume (determined by caliper measurement) in individual RXF-393 and huPBMC-implanted NSG mice treated with A) PBS, B) XENP16432 (a bivalent anti-PD-1 mAb), or exemplary αENPP3×αCD3 bsAbs (XENP30819, or XENP31419) alone or in combination with XENP16432. Each αENPP3×αCD3 bsAb was able to enhance the allogeneic antitumor effect of T cells against KU812 cells at low, medium, and / or high doses and successfully combined with PD-1 blockade. [Figure 51C] Figure 1 shows the time course of tumor volume (determined by caliper measurement) in individual RXF-393 and huPBMC-implanted NSG mice treated with A) PBS, B) XENP16432 (a bivalent anti-PD-1 mAb), or exemplary αENPP3×αCD3 bsAbs (XENP30819, or XENP31419) alone or in combination with XENP16432. Each αENPP3×αCD3 bsAb was able to enhance the allogeneic antitumor effect of T cells against KU812 cells at low, medium, and / or high doses and successfully combined with PD-1 blockade. [Figure 51D]Figure 1 shows the time course of tumor volume (determined by caliper measurement) in individual RXF-393 and huPBMC-implanted NSG mice treated with A) PBS, B) XENP16432 (a bivalent anti-PD-1 mAb), or exemplary αENPP3×αCD3 bsAbs (XENP30819, or XENP31419) alone or in combination with XENP16432. Each αENPP3×αCD3 bsAb was able to enhance the allogeneic antitumor effect of T cells against KU812 cells at low, medium, and / or high doses and successfully combined with PD-1 blockade. [Figure 51E] Figure 1 shows the time course of tumor volume (determined by caliper measurement) in individual RXF-393 and huPBMC-implanted NSG mice treated with A) PBS, B) XENP16432 (a bivalent anti-PD-1 mAb), or exemplary αENPP3×αCD3 bsAbs (XENP30819, or XENP31419) alone or in combination with XENP16432. Each αENPP3×αCD3 bsAb was able to enhance the allogeneic antitumor effect of T cells against KU812 cells at low, medium, and / or high doses and successfully combined with PD-1 blockade. [Figure 51F] Figure 1 shows the time course of tumor volume (determined by caliper measurement) in individual RXF-393 and huPBMC-implanted NSG mice treated with A) PBS, B) XENP16432 (a bivalent anti-PD-1 mAb), or exemplary αENPP3×αCD3 bsAbs (XENP30819, or XENP31419) alone or in combination with XENP16432. Each αENPP3×αCD3 bsAb was able to enhance the allogeneic antitumor effect of T cells against KU812 cells at low, medium, and / or high doses and successfully combined with PD-1 blockade. [Figure 51G]Figure 1 shows the time course of tumor volume (determined by caliper measurement) in individual RXF-393 and huPBMC-implanted NSG mice treated with A) PBS, B) XENP16432 (a bivalent anti-PD-1 mAb), or exemplary αENPP3×αCD3 bsAbs (XENP30819, or XENP31419) alone or in combination with XENP16432. Each αENPP3×αCD3 bsAb was able to enhance the allogeneic antitumor effect of T cells against KU812 cells at low, medium, and / or high doses and successfully combined with PD-1 blockade. [Figure 51H] Figure 1 shows the time course of tumor volume (determined by caliper measurement) in individual RXF-393 and huPBMC-implanted NSG mice treated with A) PBS, B) XENP16432 (a bivalent anti-PD-1 mAb), or exemplary αENPP3×αCD3 bsAbs (XENP30819, or XENP31419) alone or in combination with XENP16432. Each αENPP3×αCD3 bsAb was able to enhance the allogeneic antitumor effect of T cells against KU812 cells at low, medium, and / or high doses and successfully combined with PD-1 blockade. [Figure 51I] Figure 1 shows the time course of tumor volume (determined by caliper measurement) in individual RXF-393 and huPBMC-implanted NSG mice treated with A) PBS, B) XENP16432 (a bivalent anti-PD-1 mAb), or exemplary αENPP3×αCD3 bsAbs (XENP30819, or XENP31419) alone or in combination with XENP16432. Each αENPP3×αCD3 bsAb was able to enhance the allogeneic antitumor effect of T cells against KU812 cells at low, medium, and / or high doses and successfully combined with PD-1 blockade. [Figure 51J]Figure 1 shows the time course of tumor volume (determined by caliper measurement) in individual RXF-393 and huPBMC-implanted NSG mice treated with A) PBS, B) XENP16432 (a bivalent anti-PD-1 mAb), or exemplary αENPP3×αCD3 bsAbs (XENP30819, or XENP31419) alone or in combination with XENP16432. Each αENPP3×αCD3 bsAb was able to enhance the allogeneic antitumor effect of T cells against KU812 cells at low, medium, and / or high doses and successfully combined with PD-1 blockade. [Figure 51K] Figure 1 shows the time course of tumor volume (determined by caliper measurement) in individual RXF-393 and huPBMC-implanted NSG mice treated with A) PBS, B) XENP16432 (a bivalent anti-PD-1 mAb), or exemplary αENPP3×αCD3 bsAbs (XENP30819, or XENP31419) alone or in combination with XENP16432. Each αENPP3×αCD3 bsAb was able to enhance the allogeneic antitumor effect of T cells against KU812 cells at low, medium, and / or high doses and successfully combined with PD-1 blockade. [Figure 51L] Figure 1 shows the time course of tumor volume (determined by caliper measurement) in individual RXF-393 and huPBMC-implanted NSG mice treated with A) PBS, B) XENP16432 (a bivalent anti-PD-1 mAb), or exemplary αENPP3×αCD3 bsAbs (XENP30819, or XENP31419) alone or in combination with XENP16432. Each αENPP3×αCD3 bsAb was able to enhance the allogeneic antitumor effect of T cells against KU812 cells at low, medium, and / or high doses and successfully combined with PD-1 blockade. [Figure 52A]Several formats for use with the anti-ENPP3 x anti-CD3 bispecific antibodies disclosed herein are shown. The first is the "1 + 1 Fab-scFv-Fc" format (also known as the "bottle opener" or "triple F" format), which has a first antigen-binding domain that is a Fab domain and a second antigen-binding domain that is an scFv domain (Figure 1A). Additionally, "mAb-Fv," "mAb-scFv," "2 + 1 Fab2-scFv-Fc" (also known as the "central scFv" or "central-scFv" format), "central-Fv," "one-arm central-scFv," "one-scFv-mAb," "scFv-mAb," "dual-scFv," "trident," and non-heterodimeric bispecific formats are all shown. The scFv domains shown in Figure 49 can be, from N- to C-terminus, either variable heavy chain-(optional linker)-variable light chain or variable light chain-(optional linker)-variable heavy chain. Furthermore, in the case of one-arm scFv-mAbs, the scFv can be attached to either the N-terminus of the heavy chain monomer or the N-terminus of the light chain. In certain embodiments, "anti-antigen 1" in Figure 52 refers to the ENPP3-binding domain. In certain embodiments, "anti-antigen 1" in Figure 52 refers to the CD3-binding domain. In certain embodiments, "anti-antigen 2" in Figure 52 refers to the ENPP3-binding domain. In certain embodiments, "anti-antigen 2" in Figure 52 refers to the CD3-binding domain. In some embodiments, "anti-antigen 1" in Figure 52 refers to the ENPP3-binding domain and "anti-antigen 2" in Figure 52 refers to the CD3-binding domain. In some embodiments, "anti-antigen 1" in Figure 52 refers to the CD3-binding domain and "anti-antigen 2" in Figure 52 refers to the ENPP3-binding domain. Any of the disclosed ENPP3-binding domains and CD3-binding domains can be included in the bispecific format of Figure 52. [Figure 52B]Several formats for use with the anti-ENPP3 x anti-CD3 bispecific antibodies disclosed herein are shown. The first is the "1 + 1 Fab-scFv-Fc" format (also known as the "bottle opener" or "triple F" format), which has a first antigen-binding domain that is a Fab domain and a second antigen-binding domain that is an scFv domain (Figure 1A). Additionally, "mAb-Fv," "mAb-scFv," "2 + 1 Fab2-scFv-Fc" (also known as the "central scFv" or "central-scFv" format), "central-Fv," "one-arm central-scFv," "one-scFv-mAb," "scFv-mAb," "dual-scFv," "trident," and non-heterodimeric bispecific formats are all shown. The scFv domains shown in Figure 49 can be, from N- to C-terminus, either variable heavy chain-(optional linker)-variable light chain or variable light chain-(optional linker)-variable heavy chain. Furthermore, in the case of one-arm scFv-mAbs, the scFv can be attached to either the N-terminus of the heavy chain monomer or the N-terminus of the light chain. In certain embodiments, "anti-antigen 1" in Figure 52 refers to the ENPP3-binding domain. In certain embodiments, "anti-antigen 1" in Figure 52 refers to the CD3-binding domain. In certain embodiments, "anti-antigen 2" in Figure 52 refers to the ENPP3-binding domain. In certain embodiments, "anti-antigen 2" in Figure 52 refers to the CD3-binding domain. In some embodiments, "anti-antigen 1" in Figure 52 refers to the ENPP3-binding domain and "anti-antigen 2" in Figure 52 refers to the CD3-binding domain. In some embodiments, "anti-antigen 1" in Figure 52 refers to the CD3-binding domain and "anti-antigen 2" in Figure 52 refers to the ENPP3-binding domain. Any of the disclosed ENPP3-binding domains and CD3-binding domains can be included in the bispecific format of Figure 52. [Figure 52C]Several formats for use with the anti-ENPP3 x anti-CD3 bispecific antibodies disclosed herein are shown. The first is the "1 + 1 Fab-scFv-Fc" format (also known as the "bottle opener" or "triple F" format), which has a first antigen-binding domain that is a Fab domain and a second antigen-binding domain that is an scFv domain (Figure 1A). Additionally, "mAb-Fv," "mAb-scFv," "2 + 1 Fab2-scFv-Fc" (also known as the "central scFv" or "central-scFv" format), "central-Fv," "one-arm central-scFv," "one-scFv-mAb," "scFv-mAb," "dual-scFv," "trident," and non-heterodimeric bispecific formats are all shown. The scFv domains shown in Figure 49 can be, from N- to C-terminus, either variable heavy chain-(optional linker)-variable light chain or variable light chain-(optional linker)-variable heavy chain. Furthermore, in the case of one-arm scFv-mAbs, the scFv can be attached to either the N-terminus of the heavy chain monomer or the N-terminus of the light chain. In certain embodiments, "anti-antigen 1" in Figure 52 refers to the ENPP3-binding domain. In certain embodiments, "anti-antigen 1" in Figure 52 refers to the CD3-binding domain. In certain embodiments, "anti-antigen 2" in Figure 52 refers to the ENPP3-binding domain. In certain embodiments, "anti-antigen 2" in Figure 52 refers to the CD3-binding domain. In some embodiments, "anti-antigen 1" in Figure 52 refers to the ENPP3-binding domain and "anti-antigen 2" in Figure 52 refers to the CD3-binding domain. In some embodiments, "anti-antigen 1" in Figure 52 refers to the CD3-binding domain and "anti-antigen 2" in Figure 52 refers to the ENPP3-binding domain. Any of the disclosed ENPP3-binding domains and CD3-binding domains can be included in the bispecific format of Figure 52. [Figure 52D]Several formats for use with the anti-ENPP3 x anti-CD3 bispecific antibodies disclosed herein are shown. The first is the "1 + 1 Fab-scFv-Fc" format (also known as the "bottle opener" or "triple F" format), which has a first antigen-binding domain that is a Fab domain and a second antigen-binding domain that is an scFv domain (Figure 1A). Additionally, "mAb-Fv," "mAb-scFv," "2 + 1 Fab2-scFv-Fc" (also known as the "central scFv" or "central-scFv" format), "central-Fv," "one-arm central-scFv," "one-scFv-mAb," "scFv-mAb," "dual-scFv," "trident," and non-heterodimeric bispecific formats are all shown. The scFv domains shown in Figure 49 can be, from N- to C-terminus, either variable heavy chain-(optional linker)-variable light chain or variable light chain-(optional linker)-variable heavy chain. Furthermore, in the case of one-arm scFv-mAbs, the scFv can be attached to either the N-terminus of the heavy chain monomer or the N-terminus of the light chain. In certain embodiments, "anti-antigen 1" in Figure 52 refers to the ENPP3-binding domain. In certain embodiments, "anti-antigen 1" in Figure 52 refers to the CD3-binding domain. In certain embodiments, "anti-antigen 2" in Figure 52 refers to the ENPP3-binding domain. In certain embodiments, "anti-antigen 2" in Figure 52 refers to the CD3-binding domain. In some embodiments, "anti-antigen 1" in Figure 52 refers to the ENPP3-binding domain and "anti-antigen 2" in Figure 52 refers to the CD3-binding domain. In some embodiments, "anti-antigen 1" in Figure 52 refers to the CD3-binding domain and "anti-antigen 2" in Figure 52 refers to the ENPP3-binding domain. Any of the disclosed ENPP3-binding domains and CD3-binding domains can be included in the bispecific format of Figure 52. [Figure 52E]Several formats for use with the anti-ENPP3 x anti-CD3 bispecific antibodies disclosed herein are shown. The first is the "1 + 1 Fab-scFv-Fc" format (also known as the "bottle opener" or "triple F" format), which has a first antigen-binding domain that is a Fab domain and a second antigen-binding domain that is an scFv domain (Figure 1A). Additionally, "mAb-Fv," "mAb-scFv," "2 + 1 Fab2-scFv-Fc" (also known as the "central scFv" or "central-scFv" format), "central-Fv," "one-arm central-scFv," "one-scFv-mAb," "scFv-mAb," "dual-scFv," "trident," and non-heterodimeric bispecific formats are all shown. The scFv domains shown in Figure 49 can be, from N- to C-terminus, either variable heavy chain-(optional linker)-variable light chain or variable light chain-(optional linker)-variable heavy chain. Furthermore, in the case of one-arm scFv-mAbs, the scFv can be attached to either the N-terminus of the heavy chain monomer or the N-terminus of the light chain. In certain embodiments, "anti-antigen 1" in Figure 52 refers to the ENPP3-binding domain. In certain embodiments, "anti-antigen 1" in Figure 52 refers to the CD3-binding domain. In certain embodiments, "anti-antigen 2" in Figure 52 refers to the ENPP3-binding domain. In certain embodiments, "anti-antigen 2" in Figure 52 refers to the CD3-binding domain. In some embodiments, "anti-antigen 1" in Figure 52 refers to the ENPP3-binding domain and "anti-antigen 2" in Figure 52 refers to the CD3-binding domain. In some embodiments, "anti-antigen 1" in Figure 52 refers to the CD3-binding domain and "anti-antigen 2" in Figure 52 refers to the ENPP3-binding domain. Any of the disclosed ENPP3-binding domains and CD3-binding domains can be included in the bispecific format of Figure 52. [Figure 52F]Several formats for use with the anti-ENPP3 x anti-CD3 bispecific antibodies disclosed herein are shown. The first is the "1 + 1 Fab-scFv-Fc" format (also known as the "bottle opener" or "triple F" format), which has a first antigen-binding domain that is a Fab domain and a second antigen-binding domain that is an scFv domain (Figure 1A). Additionally, "mAb-Fv," "mAb-scFv," "2 + 1 Fab2-scFv-Fc" (also known as the "central scFv" or "central-scFv" format), "central-Fv," "one-arm central-scFv," "one-scFv-mAb," "scFv-mAb," "dual-scFv," "trident," and non-heterodimeric bispecific formats are all shown. The scFv domains shown in Figure 49 can be, from N- to C-terminus, either variable heavy chain-(optional linker)-variable light chain or variable light chain-(optional linker)-variable heavy chain. Furthermore, in the case of one-arm scFv-mAbs, the scFv can be attached to either the N-terminus of the heavy chain monomer or the N-terminus of the light chain. In certain embodiments, "anti-antigen 1" in Figure 52 refers to the ENPP3-binding domain. In certain embodiments, "anti-antigen 1" in Figure 52 refers to the CD3-binding domain. In certain embodiments, "anti-antigen 2" in Figure 52 refers to the ENPP3-binding domain. In certain embodiments, "anti-antigen 2" in Figure 52 refers to the CD3-binding domain. In some embodiments, "anti-antigen 1" in Figure 52 refers to the ENPP3-binding domain and "anti-antigen 2" in Figure 52 refers to the CD3-binding domain. In some embodiments, "anti-antigen 1" in Figure 52 refers to the CD3-binding domain and "anti-antigen 2" in Figure 52 refers to the ENPP3-binding domain. Any of the disclosed ENPP3-binding domains and CD3-binding domains can be included in the bispecific format of Figure 52. [Figure 52G]Several formats for use with the anti-ENPP3 x anti-CD3 bispecific antibodies disclosed herein are shown. The first is the "1 + 1 Fab-scFv-Fc" format (also known as the "bottle opener" or "triple F" format), which has a first antigen-binding domain that is a Fab domain and a second antigen-binding domain that is an scFv domain (Figure 1A). Additionally, "mAb-Fv," "mAb-scFv," "2 + 1 Fab2-scFv-Fc" (also known as the "central scFv" or "central-scFv" format), "central-Fv," "one-arm central-scFv," "one-scFv-mAb," "scFv-mAb," "dual-scFv," "trident," and non-heterodimeric bispecific formats are all shown. The scFv domains shown in Figure 49 can be, from N- to C-terminus, either variable heavy chain-(optional linker)-variable light chain or variable light chain-(optional linker)-variable heavy chain. Furthermore, in the case of one-arm scFv-mAbs, the scFv can be attached to either the N-terminus of the heavy chain monomer or the N-terminus of the light chain. In certain embodiments, "anti-antigen 1" in Figure 52 refers to the ENPP3-binding domain. In certain embodiments, "anti-antigen 1" in Figure 52 refers to the CD3-binding domain. In certain embodiments, "anti-antigen 2" in Figure 52 refers to the ENPP3-binding domain. In certain embodiments, "anti-antigen 2" in Figure 52 refers to the CD3-binding domain. In some embodiments, "anti-antigen 1" in Figure 52 refers to the ENPP3-binding domain and "anti-antigen 2" in Figure 52 refers to the CD3-binding domain. In some embodiments, "anti-antigen 1" in Figure 52 refers to the CD3-binding domain and "anti-antigen 2" in Figure 52 refers to the ENPP3-binding domain. Any of the disclosed ENPP3-binding domains and CD3-binding domains can be included in the bispecific format of Figure 52. [Figure 52H]Several formats for use with the anti-ENPP3 x anti-CD3 bispecific antibodies disclosed herein are shown. The first is the "1 + 1 Fab-scFv-Fc" format (also known as the "bottle opener" or "triple F" format), which has a first antigen-binding domain that is a Fab domain and a second antigen-binding domain that is an scFv domain (Figure 1A). Additionally, "mAb-Fv," "mAb-scFv," "2 + 1 Fab2-scFv-Fc" (also known as the "central scFv" or "central-scFv" format), "central-Fv," "one-arm central-scFv," "one-scFv-mAb," "scFv-mAb," "dual-scFv," "trident," and non-heterodimeric bispecific formats are all shown. The scFv domains shown in Figure 49 can be, from N- to C-terminus, either variable heavy chain-(optional linker)-variable light chain or variable light chain-(optional linker)-variable heavy chain. Furthermore, in the case of one-arm scFv-mAbs, the scFv can be attached to either the N-terminus of the heavy chain monomer or the N-terminus of the light chain. In certain embodiments, "anti-antigen 1" in Figure 52 refers to the ENPP3-binding domain. In certain embodiments, "anti-antigen 1" in Figure 52 refers to the CD3-binding domain. In certain embodiments, "anti-antigen 2" in Figure 52 refers to the ENPP3-binding domain. In certain embodiments, "anti-antigen 2" in Figure 52 refers to the CD3-binding domain. In some embodiments, "anti-antigen 1" in Figure 52 refers to the ENPP3-binding domain and "anti-antigen 2" in Figure 52 refers to the CD3-binding domain. In some embodiments, "anti-antigen 1" in Figure 52 refers to the CD3-binding domain and "anti-antigen 2" in Figure 52 refers to the ENPP3-binding domain. Any of the disclosed ENPP3-binding domains and CD3-binding domains can be included in the bispecific format of Figure 52. [Figure 52I]Several formats for use with the anti-ENPP3 x anti-CD3 bispecific antibodies disclosed herein are shown. The first is the "1 + 1 Fab-scFv-Fc" format (also known as the "bottle opener" or "triple F" format), which has a first antigen-binding domain that is a Fab domain and a second antigen-binding domain that is an scFv domain (Figure 1A). Additionally, "mAb-Fv," "mAb-scFv," "2 + 1 Fab2-scFv-Fc" (also known as the "central scFv" or "central-scFv" format), "central-Fv," "one-arm central-scFv," "one-scFv-mAb," "scFv-mAb," "dual-scFv," "trident," and non-heterodimeric bispecific formats are all shown. The scFv domains shown in Figure 49 can be, from N- to C-terminus, either variable heavy chain-(optional linker)-variable light chain or variable light chain-(optional linker)-variable heavy chain. Furthermore, in the case of one-arm scFv-mAbs, the scFv can be attached to either the N-terminus of the heavy chain monomer or the N-terminus of the light chain. In certain embodiments, "anti-antigen 1" in Figure 52 refers to the ENPP3-binding domain. In certain embodiments, "anti-antigen 1" in Figure 52 refers to the CD3-binding domain. In certain embodiments, "anti-antigen 2" in Figure 52 refers to the ENPP3-binding domain. In certain embodiments, "anti-antigen 2" in Figure 52 refers to the CD3-binding domain. In some embodiments, "anti-antigen 1" in Figure 52 refers to the ENPP3-binding domain and "anti-antigen 2" in Figure 52 refers to the CD3-binding domain. In some embodiments, "anti-antigen 1" in Figure 52 refers to the CD3-binding domain and "anti-antigen 2" in Figure 52 refers to the ENPP3-binding domain. Any of the disclosed ENPP3-binding domains and CD3-binding domains can be included in the bispecific format of Figure 52. [Figure 52J]Several formats for use with the anti-ENPP3 x anti-CD3 bispecific antibodies disclosed herein are shown. The first is the "1 + 1 Fab-scFv-Fc" format (also known as the "bottle opener" or "triple F" format), which has a first antigen-binding domain that is a Fab domain and a second antigen-binding domain that is an scFv domain (Figure 1A). Additionally, "mAb-Fv," "mAb-scFv," "2 + 1 Fab2-scFv-Fc" (also known as the "central scFv" or "central-scFv" format), "central-Fv," "one-arm central-scFv," "one-scFv-mAb," "scFv-mAb," "dual-scFv," "trident," and non-heterodimeric bispecific formats are all shown. The scFv domains shown in Figure 49 can be, from N- to C-terminus, either variable heavy chain-(optional linker)-variable light chain or variable light chain-(optional linker)-variable heavy chain. Furthermore, in the case of one-arm scFv-mAbs, the scFv can be attached to either the N-terminus of the heavy chain monomer or the N-terminus of the light chain. In certain embodiments, "anti-antigen 1" in Figure 52 refers to the ENPP3-binding domain. In certain embodiments, "anti-antigen 1" in Figure 52 refers to the CD3-binding domain. In certain embodiments, "anti-antigen 2" in Figure 52 refers to the ENPP3-binding domain. In certain embodiments, "anti-antigen 2" in Figure 52 refers to the CD3-binding domain. In some embodiments, "anti-antigen 1" in Figure 52 refers to the ENPP3-binding domain and "anti-antigen 2" in Figure 52 refers to the CD3-binding domain. In some embodiments, "anti-antigen 1" in Figure 52 refers to the CD3-binding domain and "anti-antigen 2" in Figure 52 refers to the ENPP3-binding domain. Any of the disclosed ENPP3-binding domains and CD3-binding domains can be included in the bispecific format of Figure 52. [Figure 52K]Several formats for use with the anti-ENPP3 x anti-CD3 bispecific antibodies disclosed herein are shown. The first is the "1 + 1 Fab-scFv-Fc" format (also known as the "bottle opener" or "triple F" format), which has a first antigen-binding domain that is a Fab domain and a second antigen-binding domain that is an scFv domain (Figure 1A). Additionally, "mAb-Fv," "mAb-scFv," "2 + 1 Fab2-scFv-Fc" (also known as the "central scFv" or "central-scFv" format), "central-Fv," "one-arm central-scFv," "one-scFv-mAb," "scFv-mAb," "dual-scFv," "trident," and non-heterodimeric bispecific formats are all shown. The scFv domains shown in Figure 49 can be, from N- to C-terminus, either variable heavy chain-(optional linker)-variable light chain or variable light chain-(optional linker)-variable heavy chain. Furthermore, in the case of one-arm scFv-mAbs, the scFv can be attached to either the N-terminus of the heavy chain monomer or the N-terminus of the light chain. In certain embodiments, "anti-antigen 1" in Figure 52 refers to the ENPP3-binding domain. In certain embodiments, "anti-antigen 1" in Figure 52 refers to the CD3-binding domain. In certain embodiments, "anti-antigen 2" in Figure 52 refers to the ENPP3-binding domain. In certain embodiments, "anti-antigen 2" in Figure 52 refers to the CD3-binding domain. In some embodiments, "anti-antigen 1" in Figure 52 refers to the ENPP3-binding domain and "anti-antigen 2" in Figure 52 refers to the CD3-binding domain. In some embodiments, "anti-antigen 1" in Figure 52 refers to the CD3-binding domain and "anti-antigen 2" in Figure 52 refers to the ENPP3-binding domain. Any of the disclosed ENPP3-binding domains and CD3-binding domains can be included in the bispecific format of Figure 52. [Figure 53]Schematics of heterodimeric Fc proteins described herein are provided, including 2:1 Fab2-scFv-Fc, 1:1 Fab=scFv-Fc, Y / Z-Fc (e.g., non-targeted interleukin-Fc), anti-XxY / ZF (e.g., targeted interleukin-Fc), and one-arm Fc proteins. [Figure 54] We provide a structural model of the CH3-CH3 interface constructed using MOE based on Protein Data Bank entry 3AVE. A new set of Fc substitutions can achieve over 95% heterodimer yield with little change in thermal stability. [Figure 55] The isosteric substitutions used to minimize effects on tertiary structure are shown. The difference in the isoelectric points of the Fc region allows or facilitates the direct purification of the Fc heterodimer. [Figure 56] Figure 1 shows that hinge and CH2 substitutions abolish FcγR binding. [Figure 57A] Figures 57A-C show that the 2:1 Fab2-scFv-Fc format enables targeting of tumor antigens with low density on normal cells. Tuning the valence of the TAA and the affinity of the TAA / CD3 allows for selective cytotoxicity of cell lines that mimic cancerous tissues and normal tissues (high / low antigen density). The tuned 2:1 bispecific antibody reduces interference from soluble antigens and also reduces cytokine release. Figure 57A shows that tuning the valence of the FAP and the affinity of the FAP / CD3 allows for selective cytotoxicity of cell lines that mimic cancerous tissues and normal tissues (high / low antigen density). XENP23535 represents a tuned 1:1 format targeting FAP. XENP25967 represents a tuned 2:1 format targeting FAP. [Figure 57B]We demonstrate that tuning the valency of SSTR2 and the affinity of SSTR2 / CD3 allows for selective cytotoxicity of cell lines mimicking cancerous and normal tissues (high / low antigen density). XENP18087 represents a tuned 1:1 format targeting SSTR2. XENP30458 represents a tuned 2:1 format targeting SSTR2. [Figure 57C] We demonstrate that tuning the valency of ENPP3 and the affinity of ENPP3 / CD3 enables selective cytotoxicity of cell lines mimicking cancerous and normal tissues (high / low antigen density). XENP28925 represents a tuned 1:1 format targeting ENPP3. XENP31149 represents a tuned 2:1 format targeting ENPP3. [Figure 58] The advantages of research-scale production of heterodimeric Fc proteins using the methods described herein are demonstrated, which are useful for the direct production of heterodimeric Fc proteins. [Figure 59] We demonstrate that establishment of stable cell lines yields clones with high titers and a high percentage of heterodimers. Top clones have shake flask yields of 1-2 g / L and approximately 90% heterodimer content. Data were obtained after only a standard Protein A purification step. [Figure 60] Induction of RTCC in A549 cells transfected with SSTR2 (high, medium, low density) by A) XENP18087 or B) XENP30458. [Figure 61] Figure 1 shows A) reduction in target cell numbers and release of B) IL-6, C) TNFα, D) IFNγ, and E) IL-1β by effector cells after 48 hours of incubation of CFSE-labeled [SSTR2+]COR-L279 target cells with human PBMCs (effector:target ratio of 20:1) in the presence of XENP18087 or XENP30458. [Figure 62A]Sequences of exemplary 1:1 and 2:1 tailored TAA x CD3 bispecifics described herein are shown. Anti-TAA (e.g., anti-FAP, anti-SSTR2, and anti-ENPP3) components, such as variable regions, anti-CD3 components (variable region, constant / Fc region), and linkers are shown. The linkers are double underlined (although, as will be understood by those skilled in the art, other linkers can be used), and a slash ( / ) indicates the boundary between the variable region, constant / Fc region, and linker. CDRs are underlined. In some embodiments, the 1:1 format TAA x CD3 bispecific is XENP23535, XENP18087, or XENP28925. In some embodiments, the 2:1 format TAA x CD3 bispecific is XENP25967, XENP30458, or XENP31149. [Figure 62B] Sequences of exemplary 1:1 and 2:1 tailored TAA x CD3 bispecifics described herein are shown. Anti-TAA (e.g., anti-FAP, anti-SSTR2, and anti-ENPP3) components, such as variable regions, anti-CD3 components (variable region, constant / Fc region), and linkers are shown. The linkers are double underlined (although, as will be understood by those skilled in the art, other linkers can be used), and a slash ( / ) indicates the boundary between the variable region, constant / Fc region, and linker. CDRs are underlined. In some embodiments, the 1:1 format TAA x CD3 bispecific is XENP23535, XENP18087, or XENP28925. In some embodiments, the 2:1 format TAA x CD3 bispecific is XENP25967, XENP30458, or XENP31149. [Figure 62C]Sequences of exemplary 1:1 and 2:1 tailored TAA x CD3 bispecifics described herein are shown. Anti-TAA (e.g., anti-FAP, anti-SSTR2, and anti-ENPP3) components, such as variable regions, anti-CD3 components (variable region, constant / Fc region), and linkers are shown. The linkers are double underlined (although, as will be understood by those skilled in the art, other linkers can be used), and a slash ( / ) indicates the boundary between the variable region, constant / Fc region, and linker. CDRs are underlined. In some embodiments, the 1:1 format TAA x CD3 bispecific is XENP23535, XENP18087, or XENP28925. In some embodiments, the 2:1 format TAA x CD3 bispecific is XENP25967, XENP30458, or XENP31149. [Figure 62D] Sequences of exemplary 1:1 and 2:1 tailored TAA x CD3 bispecifics described herein are shown. Anti-TAA (e.g., anti-FAP, anti-SSTR2, and anti-ENPP3) components, such as variable regions, anti-CD3 components (variable region, constant / Fc region), and linkers are shown. The linkers are double underlined (although, as will be understood by those skilled in the art, other linkers can be used), and a slash ( / ) indicates the boundary between the variable region, constant / Fc region, and linker. CDRs are underlined. In some embodiments, the 1:1 format TAA x CD3 bispecific is XENP23535, XENP18087, or XENP28925. In some embodiments, the 2:1 format TAA x CD3 bispecific is XENP25967, XENP30458, or XENP31149. [Figure 63] The sequence for the SSTR2 binding domain [αSSTR2]_H1.24_L1.30 is shown. DETAILED DESCRIPTION OF THE INVENTION

[0081] I. Overview Anti-bispecific antibodies that co-link CD3 and tumor antigen targets are used to redirect T cells to attack and lyse target tumor cells. Examples include the Bite® and DART formats, which monovalently link CD3 and tumor antigens. While CD3-targeting approaches have shown considerable promise, a common side effect of such therapies is the associated cytokine production, often resulting in toxic cytokine release syndrome. Because the anti-CD3 binding domain of bispecific antibodies binds to all T cells, a highly cytokine-producing CD4+ T cell subset is recruited. Furthermore, CD4+ T cell subsets include regulatory T cells, whose recruitment and proliferation can lead to immunosuppression and negatively impact long-term tumor suppression. Furthermore, these formats do not contain an Fc domain and exhibit a very short serum half-life in patients.

[0082] Provided herein are novel anti-CD3 x anti-ENPP3 (also referred to as anti-ENPP3 x anti-CD3, αCD3 x αENPP3, or αENPP3 x αCD3) heterodimeric bispecific antibodies and methods of using such antibodies for the treatment of cancer. In particular, provided herein are anti-CD3, anti-ENPP3 bispecific antibodies in various formats, such as those shown in Figures 15A and 15B. These bispecific antibodies are useful for the treatment of cancer, particularly cancers with increased ENPP3 expression, such as renal cell carcinoma. Such antibodies can be used to target CD3+ effector T cells to ENPP3+ tumors, thereby enabling them to attack and lyse ENPP3+ tumors.

[0083] Furthermore, in some embodiments, the present disclosure provides bispecific antibodies with different binding affinities for human CD3 that can alter or reduce potential side effects of anti-CD3 therapy. That is, in some embodiments, the antibodies described herein are "strong" or "high affinity" binders for CD3 (e.g., heavy chain and light variable domains designated as H1.30_L1.47 (optionally including a charged linker as needed)), and antibody constructs are provided that include an anti-CD3 antigen-binding domain that also binds to ENPP3. In other embodiments, the antibodies described herein are "light" or "low affinity" binders for CD3. Additional embodiments provide antibody constructs that include an anti-CD3 antigen-binding domain with intermediate or "moderate" affinity for CD3 that also binds to ENPP3. While numerous anti-CD3 antigen-binding domains (ABDs) can be used, particularly useful embodiments use six different anti-CD3 ABDs, which can be used in two scFv orientations as described herein. Affinity is typically measured using a Biacore assay.

[0084] It should be understood that the "high, medium, and low" anti-CD3 sequences provided herein can be used in a variety of heterodimerization formats, as shown in Figures 15A and 15B. Generally, due to potential side effects of T cell recruitment, exemplary embodiments utilize formats that only monovalently bind to CD3, as shown in Figures 15A and 15B; in the formats shown herein, it is the CD3 ABD that is an scFv, as described more fully herein. In contrast, the subject bispecific antibodies can bind ENPP3 either monovalently (e.g., Figure 15A) or bivalently (e.g., Figure 15B).

[0085] The present specification provides compositions containing an ENPP3-binding domain, including antibodies having such an ENPP3-binding domain (e.g., ENPP3 x CD3 bispecific antibodies). The subject antibodies containing such an ENPP3-binding domain advantageously induce a variety of different immune responses depending on the specific ENPP3-binding domain used. For example, the subject antibodies exhibit selectivity for cells with different ENPP3 expression, potency against ENPP3-expressing cells, ability to induce cytokine release, and sensitivity to soluble ENPP3. Such ENPP3-binding domains and related antibodies are used, for example, in the treatment of ENPP3-associated cancers.

[0086] Thus, in one aspect, heterodimeric antibodies that bind to two different antigens are provided herein; for example, the antibodies are "bispecific" in that they bind to two different target antigens, generally ENPP3 and CD3, as described herein. These heterodimeric antibodies can bind to these target antigens either monovalently (e.g., with a single antigen-binding domain, such as a pair of variable heavy and variable light domains) or bivalently (with two antigen-binding domains, each of which independently binds to an antigen). In some embodiments, the heterodimeric antibodies provided herein comprise one CD3-binding domain and one ENPP3-binding domain (e.g., heterodimeric antibodies in the "1+1 Fab-scFv-Fc" format described herein). In other embodiments, the heterodimeric antibodies provided herein comprise one CD3-binding domain and two ENPP3-binding domains (e.g., heterodimeric antibodies in the "2+1 Fab2-scFv-Fc" format described herein). The heterodimeric antibodies provided herein are based on the use of distinct monomers that contain amino acid substitutions on the homodimer that "skew" heterodimer formation, as outlined more fully below, and are linked to "pI variants" that allow for simple purification of the heterodimer away from the homodimer, as also outlined below. The heterodimeric bispecific antibodies provided generally rely on the use of engineered or variant Fc domains that can self-assemble in the production cell to produce the heterodimeric protein, and methods for generating and purifying such heterodimeric proteins.

[0087] II. Nomenclature The antibodies provided herein are listed in several different formats. In some cases, each monomer of a particular antibody is given a unique "XENP" number, although longer sequences may contain shorter ones, as understood in the art. For example, the "scFv-Fc" monomer of a 1+1 Fab-scFv-Fc format antibody may have a first XENP number, but the scFv domain itself has a different XENP number. Because some molecules have three polypeptides, the XENP number is used in conjunction with the components to designate them. Thus, the 2+1 Fab2-scFv-Fc format molecule XENP29520 contains three sequences (see Figure 19A): 1) a "Fab-Fc heavy chain" monomer; 2) a "Fab-scFv-Fc heavy chain" monomer; and 3) a "light chain" monomer, or equivalents, which one of skill in the art would readily identify through sequence alignment. These XENP numbers are found in the sequence listing as well as in the identifiers and figures. Furthermore, a single molecule containing three components generates multiple sequence identifiers. For example, a Fab listing includes the complete heavy chain sequence, the variable heavy chain domain sequence, and three CDRs of the variable heavy chain domain sequence, as well as the complete light chain sequence, the variable light chain domain sequence, and three CDRs of the variable light chain domain sequence. A Fab-scFv-Fc monomer includes the full-length sequence, the variable heavy chain domain sequence, three heavy chain CDR sequences, and an scFv sequence (including the scFv variable heavy chain domain sequence, the scFv variable light chain domain sequence, and the scFv linker). Note that while a single charged scFv linker (+H) is used in some molecules with scFv domains herein, other linkers can be used. Furthermore, the nomenclature for specific antigen-binding domains (e.g., ENPP3 and CD3-binding domains) uses a format of "Hx.xx_Ly.yy," where the numbers are unique identifiers for the specific variable chain sequences. Thus, the variable domain of the Fab side of the CD3 binding domain AN1[ENPP3]H1L1 (e.g., Figure 12) is "H1L1," indicating that the variable heavy chain domain H1 is combined with the light chain domain L1.When these sequences are used as scFvs, the designation "H1L1" indicates that the variable heavy chain domain H1 is combined with the light chain domain L1, in a VH-linker-VL orientation, going N- to C-terminus. A molecule with the same sequences of heavy and light variable domains but in the reverse order (VL-linker-VH orientation, going N- to C-terminus) would be designated "L1_H1.1." Similarly, different constructs can be "mix and match" heavy and light chains, as will be apparent from the sequence listing and figures.

[0088] III. Definition In order to facilitate a more complete understanding of this application, certain definitions are provided below. Such definitions are intended to encompass grammatical equivalents.

[0089] "ENPP3" or "ectonucleotide pyrophosphatase / phosphodiesterase family member 3" (e.g., Genebank Accession No. NP005012.2) herein refers to a protein that belongs to a series of ectoenzymes that are linked to the hydrolysis of extracellular nucleotides. The ENPP3 sequence is shown, for example, in Figures 11A and 11B. ENPP3 is expressed in certain cancers, including renal cell carcinoma.

[0090] As used herein, "elimination" refers to the reduction or elimination of activity. Thus, for example, "ablation of FcγR binding" means that an Fc region amino acid variant has less than 50% of the starting binding activity compared to an Fc region not containing the particular variant, preferably with 70-80-90-95-98% or more loss of activity, and generally with activity below detectable binding levels in Biacore, SPR, or BLI assays. Particularly useful in ablation of FcγR binding are those shown in Figure 5, which are typically added to both monomers.

[0091] As used herein, "ADCC" or "antibody-dependent cell-mediated cytotoxicity" refers to a cell-mediated reaction in which nonspecific cytotoxic cells expressing FcγR recognize bound antibody on target cells and subsequently cause lysis of the target cells. ADCC correlates with binding to FcγRIIIa, and increased binding to FcγRIIIa results in increased ADCC activity.

[0092] As used herein, "ADCP" or "antibody-dependent cell-mediated phagocytosis" refers to a cell-mediated reaction in which non-specific phagocytes expressing FcγR recognize bound antibody on target cells and subsequently cause phagocytosis of the target cells.

[0093] As used herein, the term "antibody" is used generically. Antibodies according to the present invention can take several formats as described herein, including conventional antibodies as well as antibody derivatives, fragments, and mimetics as described herein.

[0094] Conventional immunoglobulin (Ig) antibodies are "Y"-shaped tetramers. Each tetramer typically consists of two identical pairs of polypeptide chains, each pair having one "light" chain monomer (typically having a molecular weight of about 25 kDa) and one "heavy" chain monomer (typically having a molecular weight of about 50-70 kDa).

[0095] Other useful antibody formats include, but are not limited to, the 1+1 Fab-scFv-Fc and 2+1 Fab-scFv-Fc antibody formats described herein, as well as antibodies in the "mAb-Fv", "mAb-scFv", "central-Fv", "one-arm scFv-mAb", "scFv-mAb", "dual-scFv", and "trident" formats, as shown in Figure 49.

[0096] An antibody heavy chain typically comprises a variable heavy (VH) domain containing vhCDR1-3 and an Fc domain containing CH2-CH3 monomers. In some embodiments, an antibody heavy chain comprises a hinge and CH1 domain. 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 heavy chain "constant domain" or "constant region," although there are five different categories or "isotypes" of antibodies: IgA, IgD, IgG, IgE, and IgM. Thus, as used herein, "isotype" refers to any subclass of immunoglobulin defined by the chemical and antigenic characteristics of their constant region. It should be understood that therapeutic antibodies may also comprise hybrids of isotypes and / or subclasses. For example, as set forth in US Patent Application Publication No. 2009 / 0163699, which is incorporated by reference, the antibodies described herein include the use of human IgG1 / G2 hybrids.

[0097] In some embodiments, the antibodies provided herein comprise an IgG isotype constant domain, with several subclasses, including, but not limited to, IgG1, IgG2, IgG3, and IgG4. The IgG subclass of immunoglobulins contains several immunoglobulin domains in the heavy chain. As used herein, "immunoglobulin (Ig) domain" refers to a region of an immunoglobulin with distinct tertiary structures. The heavy chain domain, including the constant heavy chain CH domain and hinge domain, is of interest in the antibodies described herein. In the context of IgG antibodies, IgG isotypes each have three CH regions. Thus, the "CH" domains in the context of IgG are as follows: "CH1" refers to positions 118-220 according to the EU index as per Kabat; "CH2" refers to positions 237-340 according to the EU index as per Kabat; and "CH3" refers to positions 341-447 according to the EU index as per Kabat. As indicated herein and described below, pI variants may exist in one or more of the CH region and, as discussed below, the hinge region.

[0098] It should be noted that IgG1 has different allotypes with polymorphisms at 356 (D or E) and 358 (L or M). While the sequences depicted herein use the 356D / 358M allotype, other allotypes are encompassed herein. That is, any sequence comprising an IgG1 Fc domain encompassed herein can have the 356E / 358L allotype instead of the 356D / 358M allotype. It should be understood that therapeutic antibodies can also comprise isotype and / or subclass hybrids. For example, as shown in U.S. Patent Application Publication No. 2009 / 0163699, incorporated by reference, the present antibodies in some embodiments comprise an IgG1 / IgG2 hybrid.

[0099] As used herein, "Fc" or "Fc region" or "Fc domain" refers to a polypeptide comprising the constant region of an antibody, in some cases excluding all or a portion of the first constant region immunoglobulin domain (e.g., CH1), and in some cases, optionally including all or a portion of the hinge. In the case of an IgG, the Fc domain comprises immunoglobulin domains CH2 and CH3 (Cγ2 and Cγ3), and optionally all or a portion of the hinge region between CH1 (Cγ1) and CH2 (Cγ2). Thus, in some cases, the Fc domain comprises, from N- to C-terminus, CH2-CH3 and hinge-CH2-CH3. In some embodiments, the Fc domain is from IgG1, IgG2, IgG3, or IgG4, with IgG1 hinge-CH2-CH3 and IgG4 hinge-CH2-CH3 finding particular use in many embodiments. Additionally, in the case of human IgG1 Fc domains, the hinge frequently comprises a C220S amino acid substitution. Additionally, in the case of human IgG4 Fc domains, the hinge frequently contains a S228P amino acid substitution. Although the boundaries of the Fc region may vary, the human IgG heavy chain Fc region is usually defined to include residues E216, C226, or A231 at its carboxy-terminus, where numbering is according to the EU index as in Kabat. In some embodiments, amino acid modifications are made to the Fc region to, for example, alter binding to one or more FcγRs or FcRn, as described in more detail below.

[0100] As used herein, "heavy chain constant region" refers to the CH1-hinge-CH2-CH3 portion of an antibody (or fragment thereof), excluding the variable heavy domain, which in EU numbering for human IgG1 corresponds to amino acids 118 to 447. As used herein, "heavy chain constant region fragment" refers to a heavy chain constant region having fewer amino acids from either or both of the N-terminus and C-terminus, but still retaining the ability to form a dimer with another heavy chain constant region.

[0101] Another type of Ig domain in the heavy chain is the hinge region. As used herein, "hinge" or "hinge region" or "antibody hinge region" or "hinge domain" refers to the flexible polypeptide comprising the amino acids between the first and second constant domains of an antibody. Structurally, the IgG CH1 domain ends at EU position 215, and the IgG CH2 domain begins at EU position 231. Thus, for IgG, the antibody hinge is defined herein to include positions 216 (E216 in IgG1) to 230 (P230 in IgG1), numbered according to the EU index as in Kabat. In some cases, "hinge fragments" are used, which contain fewer amino acids at either or both the N-terminus and C-terminus of the hinge domain. As described herein, pI variants can also be generated in the hinge region. Many of the antibodies herein have at least one cysteine ​​at position 220 according to EU numbering (hinge region) replaced with serine. Generally, this modification is on the "scFv monomer" side for most of the sequences shown herein, but can also be on the "Fab monomer" side to reduce disulfide formation, or both. Specifically included within the sequences herein are those in which one or both of these cysteines have been replaced (C220S).

[0102] As will be understood by those skilled in the art, the exact numbering and arrangement of heavy constant region domains may vary in different numbering systems. A useful comparison of EU and Kabat heavy constant region numbering is as follows: 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 in its entirety). [Table 1]

[0103] Antibody light chains generally contain two domains: a variable light domain (VL) containing the light chain CDRs v1CDR1-3, and a constant light domain (often referred to as CL or Cκ). Antibody light chains are usually organized from the N-terminus to the C-terminus: VL-CL.

[0104] As used herein, the term "antigen-binding domain" or "ABD" refers to a set of six complementarity-determining regions (CDRs) that, when present as part of a polypeptide sequence, specifically binds to a target antigen (e.g., ENPP3 or CD3) as discussed herein. As known in the art, these CDRs generally exist as a first set of variable heavy chain CDRs (vhCDRs or VH CDRs) and a second set of variable light chain CDRs (vlCDRs or VL CDRs), each of which contains three CDRs: vhCDR1, vhCDR2, and vhCDR3 variable heavy chain CDRs, and vlCDR1, vlCDR2, and vlCDR3 variable light chain CDRs. The CDRs are present in the variable heavy chain domain (vhCDR1-3) and the variable light chain domain (vlCDR1-3). The variable heavy chain domain and the variable light chain domain from the Fv region.

[0105] The antibodies described herein provide a number of different CDR sets. In this case, a "complete CDR set" includes three variable light chain CDRs and three variable heavy chain CDRs, e.g., vlCDR1, vlCDR2, vlCDR3, vhCDR1, vhCDR2, and vhCDR3. These may be part of a larger variable light chain or variable heavy chain domain, respectively. Furthermore, as more fully outlined herein, the variable heavy and variable light chain domains may be on separate polypeptide chains when heavy and light chains are used (e.g., when Fabs are used), or on a single polypeptide chain in the case of scFv sequences.

[0106] As will be understood by those skilled in the art, the exact numbering and arrangement of CDRs may vary in different numbering systems. However, it should be understood that the disclosure of a variable heavy chain and / or variable light chain sequence includes the disclosure of the associated (unique) CDRs. Thus, the disclosure of each variable heavy region is the disclosure of the vhCDRs (e.g., vhCDR1, vhCDR2, and vhCDR3), and the disclosure of each variable light region is the disclosure of the 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). [Table 2]

[0107] Throughout this specification, the Kabat numbering system is generally used when referring to residues within the variable domain (approximately residues 1-107 of the light chain variable region and residues 1-113 of the heavy chain variable region), and the EU numbering system is for the Fc region (see, e.g., Kabat et al., supra (1991)).

[0108] CDRs contribute to the formation of antigen binding, or more specifically, the formation of the antigen-binding domain and epitope binding site of an antibody. "Epitope" refers to the determinant that interacts with the specific antigen-binding site in the variable region of an antibody molecule, known as paratope. An epitope is a grouping of molecules, such as amino acids or sugar side chains, and usually has specific structural characteristics and specific charge characteristics. A single antigen may have two or more epitopes.

[0109] An epitope can include amino acid residues that are directly involved in binding (also referred to as the immunodominant components of the epitope) and other amino acid residues that are not directly involved in binding, e.g., amino acid residues that are effectively blocked by a specific antigen-binding peptide, in other words, amino acid residues that lie within the footprint of a specific antigen-binding peptide.

[0110] Epitopes can be either conformational or linear. Conformational epitopes are generated by the spatial juxtaposition of amino acids from different segments of a linear polypeptide chain. Linear epitopes are generated by adjacent amino acid residues within a polypeptide chain. Conformational and nonconformational epitopes can be distinguished in that the binding to the former, but not the latter, is lost in the presence of denaturing solvents.

[0111] An epitope typically comprises at least three, more usually at least five, or 8-10 amino acids in a unique spatial conformation. Antibodies that recognize the same epitope can be verified in a simple immunoassay demonstrating the ability of one antibody to block the binding of another antibody to a target antigen, e.g., "binning." As outlined below, the present disclosure includes not only the recited antigen-binding domains and antibodies herein, but also those that compete for binding to the epitope bound by the recited antigen-binding domains.

[0112] In some embodiments, the six CDRs of an antigen-binding domain are contributed by a variable heavy chain and a variable light chain domain. In the "Fab" format, the set of six CDRs is provided by two different polypeptide sequences: the variable heavy chain domain (vh or V H , including vhCDR1, vhCDR2 and vhCDR3) and a variable light chain domain (vl or V L, vlCDR1, vlCDR2, and vlCDR3) are contributed, of which the C-terminus of the vh domain is attached to the N-terminus of the CH1 domain of the heavy chain and the C-terminus of the vl domain is attached to the N-terminus of the constant light chain domain (thus forming the light chain). In the scFv format, the vh and vl domains are generally covalently joined into a single polypeptide sequence by the use of a linker (an "scFv linker"), as outlined herein, which (starting from the N-terminus) can be either vh-linker-vl or vl-linker-vh, with the former generally preferred (with optional domain linkers on either side, depending on the format used (e.g., from Figure 1 )). Generally, the C-terminus of the scFv domain is attached to the N-terminus of the hinge of the second monomer.

[0113] As used herein, a "variable region" or "variable domain" refers to a region of an immunoglobulin that includes one or more Ig domains substantially encoded by any of the Vκ, Vλ, and / or VH genes that constitute the kappa, lambda, and heavy chain immunoglobulin loci, respectively, and contains CDRs that confer antigen specificity. Thus, a "variable heavy chain domain" pairs with a "variable light chain domain" to form an antigen-binding domain ("ABD"). Each variable domain further includes three hypervariable regions ("complementarity-determining regions," "CDRs") (VHCDR1, VHCDR2, and VHCDR3 in the variable heavy chain domain, and VLCDR1, VLCDR2, and VLCDR3 in the variable light chain domain) and four framework (FR) regions, arranged from the amino terminus to the carboxy terminus in the following order: FR1-CDR1-FR2-CDR2-FR3-CDR3-FR4. The hypervariable regions generally comprise amino acid residues from about amino acid residues 24-34 (LCDR1, "L" represents the light chain), 50-56 (LCDR2), and 89-97 (LCDR3) in the light chain variable region, and from about amino acid residues 31-35B (HCDR1, "H" represents the heavy chain), 50-65 (HCDR2), and 95-102 (HCDR3) in the heavy chain variable region; Kabat et al., SEQUENCES OF PROTEINS OF IMMUNOLOGICAL INTEREST, 5th Ed. Public Health Service, National Institutes of Health Health, Bethesda, Md. (1991) and / or those residues that form the hypervariable loops (e.g., residues 26-32 (LCDR1), 50-52 (LCDR2), and 91-96 (LCDR3) in the light chain variable region and 26-32 (HCDR1), 53-55 (HCDR2), and 96-101 (HCDR3) in the heavy chain variable region; Chothia and Lesk (1987) J. Mol. Biol. 196:901-917). Specific CDRs of the present invention are set forth in Table 2.

[0114] As used herein, "Fab" or "Fab region" generally refers to a polypeptide comprising the VH, CH1, VL, and CL immunoglobulin domains of two different polypeptide chains (e.g., VH-CH1 of one chain, VL-CL of the other chain). Fab can refer to this region alone or in the context of the bispecific antibodies described herein. In the context of Fab, Fab includes the Fv region in addition to the CH1 and CL domains.

[0115] As used herein, "Fv" or "Fv fragment" or "Fv region" refers to a polypeptide comprising the VL and VH domains of an ABD. The Fv region can be formatted as both a Fab (which, as noted above, are two distinct polypeptides that also generally comprise a constant region as outlined above) and an scFv, where the VL and VH domains combine (generally with a linker as discussed herein) to form the scFv.

[0116] As used herein, "single-chain Fv" or "scFv" generally refers to a variable heavy domain covalently linked to a variable light domain, forming an scFv or scFv domain using an scFv linker as discussed herein. The scFv domains may be in either orientation, from N-terminus to C-terminus (VH-linker-VL or VL-linker-VH). In the sequences shown in the sequence listing and figures, the order of the VH and VL domains is indicated in the name. For example, H.X_L.Y means, from N-terminus to C-terminus, VH-linker-VL, and L.Y_H.X means VL-linker-VH.

[0117] The subject antibody embodiments provided herein comprise at least one scFv domain, which, while not naturally occurring, generally comprises a variable heavy domain and a variable light domain linked together by an scFv linker. As outlined herein, the scFv domains are generally oriented N- to C-terminus as VH-scFv linker-VL, although this can be reversed for any scFv domain (or domain constructed using Fab-derived VH and VL sequences) to VL-scFv linker-VH, with optional linkers at either or both ends depending on the format.

[0118] As used herein, "modification" refers to the substitution, insertion, and / or deletion of an amino acid in a polypeptide sequence, or to a modification to a moiety chemically linked to a protein. For example, the modification may be a modified carbohydrate or PEG structure attached to a protein. As used herein, "amino acid modification" refers to the substitution, insertion, and / or deletion of an amino acid in a polypeptide sequence. For clarity, unless otherwise specified, the amino acid modification is always to an amino acid encoded by DNA, e.g., the 20 amino acids that have codons in DNA and RNA.

[0119] As used herein, "amino acid substitution" or "substitution" refers to the replacement of an amino acid at a particular position in a parent polypeptide sequence with a different amino acid. In particular, in some embodiments, the substitution is for an amino acid that does not naturally occur at the particular position (either not naturally occurring in an organism or not occurring in any organism). For example, the substitution E272Y refers to a variant polypeptide, in this case an Fc variant, in which glutamic acid at position 272 is replaced with tyrosine. For clarity, a protein engineered to alter a nucleic acid coding sequence but not the starting amino acid (e.g., replacing CGG (which encodes arginine) with CGA (which still encodes arginine) to increase host organism expression levels) is not an "amino acid substitution." That is, if a protein has the same amino acid at the particular position where it starts, despite the creation of a new gene encoding the same protein, it is not an amino acid substitution.

[0120] As used herein, "amino acid insertion" or "insertion" refers to the addition of an amino acid sequence at a specific position in a parent polypeptide sequence. For example, -233E or 233E indicates the insertion of glutamic acid after position 233 and before position 234. Furthermore, -233ADE or A233ADE indicates the insertion of AlaAspGlu after position 233 and before position 234.

[0121] As used herein, "amino acid deletion" or "deletion" refers to the removal of an amino acid sequence at a particular position in a parent polypeptide sequence. For example, E233- or E233#, E233() or E233del indicates the deletion of glutamic acid at position 233. Additionally, EDA233- or EDA233# indicates the deletion of the sequence GluAspAla beginning at position 233.

[0122] As used herein, "variant protein" or "protein variant" or "variant" refers to a protein that differs from that of a parent protein based on 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 does not align with the parent protein using an alignment program such as those described below. Generally, the variant proteins outlined herein (such as variant Fc domains) are generally 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. As used herein, "variant" also refers to specific amino acid modifications that confer specific functions (e.g., "heterodimerization variants," "pI variants," "deletion variants," etc.).

[0123] As described below, in some embodiments, the parent polypeptide, e.g., an Fc parent polypeptide, is a human wild-type sequence, such as a heavy chain constant domain or Fc region from IgG1, IgG2, IgG3, or IgG4; however, human sequences with variants can also serve as "parent polypeptides," including, for example, the IgG1 / 2 hybrids of U.S. Publication No. 2006 / 0134105. The sequence of a protein variant herein preferably has at least about 80% identity, most preferably at least about 90% identity, and more preferably at least about 95-98-99% identity to the parent protein sequence. Thus, as used herein, "antibody variant" or "variant antibody" refers to an antibody that differs from the parent antibody by at least one amino acid modification; as used herein, "IgG variant" or "variant IgG" refers to an antibody that differs from the parent IgG (again, often derived from human IgG) by at least one amino acid modification; and as used herein, "immunoglobulin variant" or "variant immunoglobulin" refers to an immunoglobulin sequence that differs from that of the parent immunoglobulin sequence by at least one amino acid modification. As used herein, "Fc variant" or "variant Fc" refers to a protein that contains amino acid modifications compared to the Fc domain of human IgG1, IgG2, or IgG4.

[0124] As used herein, "Fc variant" or "variant Fc" refers to a protein containing an amino acid modification in the Fc domain. The modification can be an addition, deletion, or substitution. Fc variants are defined according to the amino acid modification that constitutes them. Thus, for example, N434S or 434S is an Fc variant with a substituted serine at position 434 relative to the parent Fc polypeptide, where the numbering is according to the EU index. Similarly, M428L / N434S defines an Fc variant with the substitutions M428L and N434S relative to the parent Fc polypeptide. The identity of the WT amino acids may not be specified, in which case the variant is referred to as 428L / 434S. It is noted that the substitutions may be provided in any order; i.e., for example, 428L / 434S is the same Fc variant as 434S / 428L. For all positions discussed in this invention relating to antibodies or derivatives and fragments thereof (e.g., Fc domains), unless otherwise specified, the numbering of amino acid positions is according to the EU index. The "EU index" or "Kabat-like EU index" or "EU numbering" scheme refers to the numbering of the EU antibody (Edelman et al., 1969, Proc Natl Acad Sci USA 63:78-85, hereby incorporated by reference in its entirety).

[0125] Generally, the variant Fc domain will have at least about 80, 85, 90, 95, 97, 98, or 99 percent identity to the corresponding parent human IgG Fc domain (using an identity algorithm discussed below (in one embodiment the BLAST algorithm known in the art) with default parameters). Alternatively, the 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 amino acid modifications relative to the parent 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. Additionally, as discussed herein, the variant Fc domains described herein still retain the ability to form dimers with another Fc domain as measured using known techniques described herein, such as non-denaturing gel electrophoresis.

[0126] As used herein, "protein" refers to at least two covalently attached amino acids, including proteins, polypeptides, oligopeptides, and peptides. Additionally, the polypeptides from which the antibodies described herein are made may include one or more side chain or terminus synthetic derivatizations, glycosylation, PEGylation, circular permutation, cyclization, linkers to other molecules, fusions to proteins or protein domains, and addition of peptide tags or labels.

[0127] "Residue," as used herein, refers to a position in a protein and its associated amino acid identity. For example, asparagine 297 (also called Asn297 or N297) is the residue at position 297 in the human antibody IgG1.

[0128] As used herein, "IgG subclass modification" or "isotype modification" refers to an amino acid modification that converts one amino acid of one IgG isotype to the corresponding amino acid of a different, matching IgG isotype. For example, because IgG1 contains tyrosine and IgG2 contains phenylalanine at EU position 296, an F296Y substitution in IgG2 is considered to be an IgG subclass modification.

[0129] As used herein, a "non-naturally occurring modification" refers to an amino acid modification that is not isotypic. For example, the substitution 434S in IgG1, IgG2, IgG3, or IgG4 (or hybrids thereof) is considered to be a non-naturally occurring modification because none of the human IgGs contain serine at position 434.

[0130] As used herein, "amino acid" and "amino acid identity" refer to one of the 20 naturally occurring amino acids encoded by DNA and RNA.

[0131] As used herein, "effector function" refers to a biochemical event that results from the interaction of an antibody Fc region with an Fc receptor or ligand. Effector functions include, but are not limited to, ADCC, ADCP, and CDC.

[0132] As used herein, "IgG Fc ligand" refers to any biologically derived molecule, preferably a polypeptide, 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 lectin, mannose receptor, 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, incorporated by reference in its entirety). Fc ligands may also include undiscovered molecules that bind to Fc. Specific IgG Fc ligands are FcRn and Fc gamma receptors. As used herein, "Fc ligand" refers to any biologically derived molecule, preferably a polypeptide, that binds to the Fc region of an antibody to form an Fc / Fc ligand complex.

[0133] As used herein, "Fc gamma receptor," "FcγR," or "Fc gamma R" refers to any member of a family of proteins that binds to the Fc region of an IgG antibody and is encoded by the FcγR gene. In humans, this family includes FcγRI (CD64), which includes the isoforms FcγRIa, FcγRIb, and FcγRIc; FcγRII (CD32), which includes the isoforms FcγRIIa (including the allotypes H131 and R131), FcγRIIb (including FcγRIIb-1 and FcγRIIb-2), and FcγRIIc; and FcγRIII (CD16), which includes the isoforms FcγRIIIa (including the allotypes V158 and F158), and FcγRIIIb (including the allotypes FcγRIIb-NA1 and FcγRIIb-NA2) (Jefferis et al., 2002, Immunol Lett 82:57-65, entirely incorporated by reference), and any undiscovered human FcγR or FcγR isoform or allotype. FcγRs can be derived from any organism, including, but not limited to, humans, mice, rats, rabbits, and monkeys. Mouse FcγRs include, but are not limited to, FcγRI (CD64), FcγRII (CD32), FcγRIII (CD16), and FcγRIII-2 (CD16-2), as well as any unidentified mouse FcγR or FcγR isoform or allotype.

[0134] As used herein, "FcRn" or "neonatal Fc receptor" refers to a protein that binds to the Fc region of an IgG antibody and is at least partially encoded by the FcRn gene. FcRn can be derived from any organism, including, but not limited to, humans, mice, rats, rabbits, and monkeys. As known in the art, a functional FcRn protein comprises two polypeptides, often referred to as heavy and light chains. The light chain is β-2-microglobulin, and the heavy chain is encoded by the FcRn gene. Unless otherwise specified herein, FcRn or FcRn protein refers to the complex of the FcRn heavy chain and β-2-microglobulin. Various FcRn variants are available that are used to increase binding to the FcRn receptor and, in some cases, to increase serum half-life. "FcRn variants" are those that increase binding to FcRn. Suitable FcRn variants are listed below.

[0135] As used herein, "parent polypeptide" refers to a starting polypeptide that is subsequently modified to generate a variant. A parent polypeptide may be a naturally occurring polypeptide or a variant or engineered version of a naturally occurring polypeptide. Thus, as used herein, "parent immunoglobulin" refers to an unmodified immunoglobulin polypeptide that is modified to generate a variant, and "parent antibody" refers to an unmodified antibody that is modified to generate a variant antibody. It should be noted that "parent antibody" includes known, commercially available, recombinantly produced antibodies, as outlined below. In this context, the "parent Fc domain" is relative to the recited variant; thus, a "variant human IgG1 Fc domain" is compared to the parent Fc domain of human IgG1, a "variant human IgG4 Fc domain" is compared to the parent Fc domain of human IgG4, etc.

[0136] As used herein, "position" means a location in the sequence of a protein. Positions may be numbered sequentially or according to established formats, such as the EU index for antibody numbering.

[0137] "Target antigen," as used herein, means a molecule that is specifically bound by an antibody binding domain comprising the variable region of a given antibody.

[0138] "Strandedness," in reference to the monomers of heterodimeric antibodies described herein, refers to the incorporation of heterodimerization mutations into each monomer such that they retain the ability to "match" and form heterodimers, similar to "matching" double-stranded DNA. For example, if several pI variants are engineered into monomer A (e.g., to increase the pI), similarly exploitable "charge-paired" steric mutations do not interfere with the pI variants; for example, the pI-increasing charge mutations can be placed in the same "strand" or "monomer," retaining both functions. Similarly, for "skewed" variants that result in paired sets, as outlined in more detail below, one skilled in the art will consider pI when determining which strand or monomer one set of the pair will reside in; thus, pI separation is also maximized using the skewed pI.

[0139] "Target cell," as used herein, means a cell that expresses a target antigen.

[0140] By "host cell" in the context of producing bispecific antibodies using the antibodies described herein is meant a cell that contains exogenous nucleic acid encoding the components of the bispecific antibody and is capable of expressing the bispecific antibody under suitable conditions. Suitable host cells are described below.

[0141] As used herein, "wild-type or WT" refers to an amino acid or nucleotide sequence found in nature, including allelic variations. A WT protein has an amino acid or nucleotide sequence that has not been intentionally modified.

[0142] The present specification provides several antigen-binding domains that share sequence identity with human antibody domains. Sequence identity between two similar sequences (such as antibody variable domains) can be determined using methods such as Smith, T. F. & Waterman, M. S. (1981) "Comparison of Biosequences," Adv. Appl. Math. 2:482 (local homology algorithm), Needleman, S. B. & Wunsch, C. D. (1970) "A General Method Applicable to the Search for Similarities in the Amino Acid Sequence of Two Proteins," J. Mol. Biol. 48:443 (homology alignment algorithm), Pearson, W. R. & Lipman, D. J. (1988) "Improved Tools for Biological Sequence Comparison," Proc. Natl. Acad. Sci. (USA) 85:2444 (search for similarity method), or Altschul, S. F. et al. (1990) "Basic Local Alignment Search Method." Tool, "J. Mol. Biol. 215:403-10, the "BLAST" algorithm (see https: / / blast.ncbi.nlm.nih.gov / Blast.cgi). When using any of the aforementioned algorithms, default parameters (window length, gap penalty, etc.) are used. In one embodiment, sequence identity is measured using the BLAST algorithm using default parameters.

[0143] The antibodies described herein are generally isolated or recombinant. When used to describe various polypeptides disclosed herein, "isolated" refers to a polypeptide that has been identified and separated and / or recovered from the cell or cell culture in which it is expressed. Typically, an isolated polypeptide will be prepared by at least one purification step. An "isolated antibody" refers to an antibody that is substantially free of other antibodies with different antigen specificities. "Recombinant" refers to antibodies that are produced using recombinant nucleic acid technology in an exogenous host cell and may also be isolated.

[0144] "Specific binding" or "specifically binds to" or "specific for" a particular antigen or epitope refers to binding that is measurably different 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 generally a molecule of similar structure that does not have binding activity. For example, specific binding can be determined by competition between the target and a similar control molecule.

[0145] Specific binding to a particular antigen or epitope is, for example, 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 A specific binding activity can be exhibited by an antibody having a KD of M or greater, where KD refers to the dissociation rate of a particular antibody-antigen interaction. Typically, an antibody that specifically binds to an antigen will have a KD that is 20-fold, 50-fold, 100-fold, 500-fold, 1000-fold, 5,000-fold, 10,000-fold, or more greater than that of a control molecule relative to the antigen or epitope.

[0146] Specific binding to a particular antigen or epitope can also be exhibited, for example, by an antibody having a K or K for that antigen or epitope that is at least 20-fold, 50-fold, 100-fold, 500-fold, 1000-fold, 5,000-fold, 10,000-fold, or more greater than that of a control, where K or K refers to the association rate of a particular antibody-antigen interaction. Binding affinity is typically measured using Biacore, SPR, or BLI assays.

[0147] IV. ENPP3-binding domain In one aspect, the present invention provides compositions comprising ENPP3 antigen binding domains (ABDs), including anti-ENPP3 antibodies, and the like.The target antibodies comprising such ENPP3 antigen binding domains (ABDs) (e.g., anti-ENPP3 x anti-CD3 bispecific antibodies) advantageously induce a variety of different immune responses (see Examples 5 and 6).These ENPP3 binding domains and related antibodies are used, for example, to treat ENPP3-related cancers.

[0148] As will be appreciated by those of skill in the art, a suitable ENPP3 binding domain can comprise a set of six CDRs as depicted in the Sequence Listing and Figures 12, 13A-13B, and 14A-14I, as identified using underlined or, if a different numbering scheme is used as described herein and shown in Table 2, other alignments within the variable heavy (VH) and variable light (VL) domain sequences as depicted in Figures 12, 13A-13B, and 14A-14I and the Sequence Listing (see Table 2). A suitable ENPP3 ABD can also comprise these sequences and the entire VH and VL sequences as depicted in the figures, used as an scFv or Fab domain.

[0149] In one embodiment, the ENPP3 antigen binding domain comprises the six CDRs (i.e., vhCDR1-3 and vlCDR1-3) of the ENPP3 ABD described herein, including the Figures and Sequence Listing. In exemplary embodiments, the ENPP3 ABD comprises the following ENPP3 binding domains: AN1[ENPP3]H1L1, AN1[ENPP3]H1 L1.33, AN1[ENPP3]H1 L1.77, AN1[ENPP3]H1.8 L1, AN1[ENPP3]H1.8 L1.33, AN1[ENPP3]H1 L1.77, H16-7.213, H16-9.69, H16-1.52, Ha16-1(1)23, H16-9.44, H16-1.67, Ha1 6-1(3,5)36, H16-1.86, H16-9.10, H16-9.33, H16-1.68, Ha16-1(1)1, Ha1 6-1(3,5)18, Ha16-1(2,4)4, Ha16-1(3,5)56, H16-7.8, H16-1.93, Ha1 6-1(3,5)27.1, H16-1.61, H16-1(3,5)5, ​​H16-7.200, H16-1(3,5)42, H1 6-9.65, Ha1-1(3,5)19, and Ha16-1.80 ( Figures 12 , 13A–13B , and 14A–14I ).

[0150] Provided herein are variant ENPP3 ABDs having CDRs that include at least one modification of an ENPP3 ABD CDR disclosed herein in addition to the parent CDR set disclosed in the Figures and Sequence Listing that form the ABD for ENPP3. In one embodiment, the ENPP3 ABD includes a set of six CDRs that have 1, 2, 3, 4, 5, 6, 7, 8, 9, 10 amino acid modifications compared to the six CDRs of the ENPP3 ABD described herein, including the Figures and Sequence Listing. In exemplary embodiments, the ENPP3 ABDs are the following ENPP3 ABDs: AN1[ENPP3]H1L1, AN1[ENPP3]H1 L1.33, AN1[ENPP3]H1 L1.77, AN1[ENPP3]H1.8 L1, AN1[ENPP3]H1.8 L1.33, AN1[ENPP3]H1 L1.77, H16-7.213, H16-9.69, H16-1.52, Ha16-1(1)23, H16-9.44, H16-1.67, Ha1 6-1(3,5)36, H16-1.86, H16-9.10, H16-9.33, H16-1.68, Ha16-1(1)1, Ha1 H16-1(3,5)19, and H16-1.80 (Figures 12, 13A-13B, and 14A-14I). In certain embodiments, the variant ENPP3 ABD is capable of binding to the ENPP3 antigen as measured by at least one of Biacore, surface plasmon resonance (SPR), and / or BLI (biolayer interferometry, e.g., Octet assay) assays, the latter of which finds particular use in many embodiments. In certain embodiments, the ENPP3 ABD is capable of binding to the human ENPP3 antigen (see Example 5).

[0151] In one embodiment, the ENPP3 ABD comprises six CDRs that are at least 90, 95, 97, 98, or 99% identical to the six CDRs of an ENPP3 ABD described herein, including the figures and sequence listing. In exemplary embodiments, the ENPP3 ABDs are the following ENPP3 ABDs: AN1[ENPP3]H1L1, AN1[ENPP3]H1 L1.33, AN1[ENPP3]H1 L1.77, AN1[ENPP3]H1.8 L1, AN1[ENPP3]H1.8 L1.33, AN1[ENPP3]H1 L1.77, H16-7.213, H16-9.69, H16-1.52, Ha16-1(1)23, H16-9.44, H16-1.67, Ha1 6-1(3,5)36, H16-1.86, H16-9.10, H16-9.33, H16-1.68, Ha16-1(1)1, Ha1 H16-1(3,5)19, and Ha16-1.80 (Figures 12, 13A-13B, and 14A-14I). In certain embodiments, the ENPP3 ABD is capable of binding to the ENPP3 antigen as measured by at least one of Biacore, surface plasmon resonance (SPR), and / or BLI (biolayer interferometry, e.g., Octet assay) assays, the latter of which finds particular use in many embodiments. In certain embodiments, the ENPP3 ABD is capable of binding to human ENPP3 antigen (see Figure 2).

[0152] In another exemplary embodiment, the ENPP3 ABD comprises the variable heavy (VH) domain and variable light (VL) domain of any one of the ENPP3 ABDs described herein, including the figures and sequence listing. In exemplary embodiments, the ENPP3 ABD comprises the following ENPP3 binding domains: AN1[ENPP3]H1L1, AN1[ENPP3]H1 L1.33, AN1[ENPP3]H1 L1.77, AN1[ENPP3]H1.8 L1, AN1[ENPP3]H1.8 L1.33, AN1[ENPP3]H1 L1.77, H16-7.213, H16-9.69, H16-1.52, Ha16-1(1)23, H16-9.44, H16-1.67, Ha1 6-1(3,5)36, H16-1.86, H16-9.10, H16-9.33, H16-1.68, Ha16-1(1)1, Ha1 6-1(3,5)18, Ha16-1(2,4)4, Ha16-1(3,5)56, H16-7.8, H16-1.93, Ha1 6-1(3,5)27.1, H16-1.61, H16-1(3,5)5, ​​H16-7.200, H16-1(3,5)42, H1 6-9.65, Ha1-1(3,5)19, and Ha16-1.80 ( Figures 12 , 13A–13B , and 14A–14I ).

[0153] In addition to the parent ENPP3 variable heavy and light domains disclosed herein, provided herein are ENPP3 ABDs that comprise variable heavy and / or variable light domains that are variants of the ENPP3 ABD VH and VL domains disclosed herein. In one embodiment, the variant VH and / or VL domains have 1, 2, 3, 4, 5, 6, 7, 8, 9, or 10 amino acid changes from the VH and / or VL domains of an ENPP3 ABD described herein, including in the Figures and Sequence Listing. In exemplary embodiments, the variant VH and / or VL domains are selected from the following ENPP3 ABDs: AN1[ENPP3]H1L1, AN1[ENPP3]H1 L1.33, AN1[ENPP3]H1 L1.77, AN1[ENPP3]H1.8 L1, AN1[ENPP3]H1.8 L1.33, AN1[ENPP3]H1 L1.77, H16-7.213, H16-9.69, H16-1.52, Ha16-1(1)23, H16-9.44, H16-1.67, Ha1 6-1(3,5)36, H16-1.86, H16-9.10, H16-9.33, H16-1.68, Ha16-1(1)1, Ha1 H16-1(3,5)18, H16-1(2,4)4, H16-1(3,5)56, H16-7.8, H16-1.93, H16-1(3,5)27.1, H16-1.61, H16-1(3,5)5, ​​H16-7.200, H16-1(3,5)42, H16-9.65, H16-1(3,5)19, and H16-1.80 (Figures 12, 13A-13B, and 14A-14I). In certain embodiments, the ENPP3 ABD is capable of binding to ENPP3 as measured by at least one of Biacore, surface plasmon resonance (SPR), and / or BLI (biolayer interferometry, e.g., Octet assay) assays, the latter of which finds particular use in many embodiments. In certain embodiments, the ENPP3 ABD is capable of binding to the human ENPP3 antigen (see Example 5).

[0154] 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 an ENPP3 ABD described herein, including the Figures and Sequence Listing. In exemplary embodiments, the variant VH and / or VL domains comprise the following ENPP3 ABDs: AN1[ENPP3]H1L1, AN1[ENPP3]H1 L1.33, AN1[ENPP3]H1 L1.77, AN1[ENPP3]H1.8 L1, AN1[ENPP3]H1.8 L1.33, AN1[ENPP3]H1 L1.77, H16-7.213, H16-9.69, H16-1.52, Ha16-1(1)23, H16-9.44, H16-1.67, Ha1 6-1(3,5)36, H16-1.86, H16-9.10, H16-9.33, H16-1.68, Ha16-1(1)1, Ha1 and / or VL of one of Ha16-1(3,5)18, Ha16-1(2,4)4, Ha16-1(3,5)56, H16-7.8, H16-1.93, Ha16-1(3,5)27.1, H16-1.61, H16-1(3,5)5, ​​H16-7.200, H16-1(3,5)42, H16-9.65, Ha1-1(3,5)19, and Ha16-1.80 (Figures 12, 13A-13B, and 14A-14I). In certain embodiments, the ENPP3 ABD is capable of binding to ENPP3 as measured by at least one of Biacore, surface plasmon resonance (SPR), and / or BLI (biolayer interferometry, e.g., Octet assay) assays, the latter of which finds particular use in many embodiments. In certain embodiments, the ENPP3 ABD is capable of binding to the human ENPP3 antigen (see Example 5).

[0155] V. Antibodies In one aspect, provided herein is an antibody that binds to ENPP3 (e.g., an anti-ENPP3 antibody). In certain embodiments, the antibody binds to human ENPP3 (FIG. 11A). Anti-ENPP3 antibodies of interest include monospecific ENPP3 antibodies and multispecific (e.g., bispecific) anti-ENPP3 antibodies. In certain embodiments, the anti-ENPP3 antibody is formatted according to any one of the antibody formats shown in FIGS. 15A, 15B, and 52A-52K.

[0156] In some embodiments, the subject compositions comprise an ENPP3-binding domain. In some embodiments, the compositions comprise an antibody having an ENPP3-binding domain. The antibodies provided herein comprise one, two, three, four, five, or more ENPP3-binding domains. In certain embodiments, the ENPP3-binding domain comprises any one of the vhCDR1, vhCDR2, vhCDR3, vlCDR1, vlCDR2, and vlCDR3 sequences of an ENPP3-binding domain selected from those depicted in Figures 12, 13A-13B, and 14A-14I. In some embodiments, the ENPP3-binding domain comprises the underlined vhCDR1, vhCDR2, vhCDR3, vlCDR1, vlCDR2, and vlCDR3 sequences of an ENPP3-binding domain selected from those depicted in Figures 12, 13A-13B, and 14A-14I. In some embodiments, the ENPP3 binding domain comprises a variable heavy domain and a variable light domain of an ENPP3 binding domain selected from those shown in Figures 12, 13A-13B, and 14A-14I. The ENPP3 binding domains shown in Figures 12, 13A-13B, and 14A-14I include AN1[ENPP3]H1L1, AN1[ENPP3]H1L1.33, AN1[ENPP3]H1L1.77, AN1[ENPP3]H1.8L1, AN1[ENPP3]H1.8L1.33, AN1[ENPP3]H1L1.77, H16-7.213, H16-9.69, H16-1.52, Ha16-1(1)23, H16-9.44, H16-1.67, Ha16-1 These include (3,5)36, H16-1.86, H16-9.10, H16-9.33, H16-1.68, Ha16-1(1)1, Ha16-1(3,5)18, Ha16-1(2,4)4, Ha16-1(3,5)56, H16-7.8, H16-1.93, Ha16-1(3,5)27.1, H16-1.61, H16-1(3,5)5, ​​H16-7.200, H16-1(3,5)42, H16-9.65, Ha1-1(3,5)19, and Ha16-1.80.

[0157] In one aspect, provided herein are bispecific antibodies that bind to ENPP3 and CD3 in various formats, as outlined below and generally shown in Figures 15A and 15B. These bispecific heterodimeric antibodies comprise an ENPP3-binding domain. In certain embodiments, the ENPP3-binding domain comprises the vhCDR1, vhCDR2, vhCDR3, vlCDR1, vlCDR2, and vlCDR3 sequences of an ENPP3-binding domain selected from the group consisting of those depicted in Figures 12, 13A-13B, and 14A-14I. In some embodiments, the ENPP3-binding domain comprises the underlined VHCDR1, VHCDR2, VHCDR3, VLCDR1, VLCDR2, and VLCDR3 sequences of an ENPP3-binding domain selected from those depicted in Figures 12, 13A-13B, and 14A-14I.

[0158] These bispecific heterodimeric antibodies bind to ENPP3 and CD3. Such antibodies contain a CD3-binding domain and at least one ENPP3-binding domain. Any suitable ENPP3-binding domain can be included in the anti-ENPP3X anti-CD3 bispecific antibody. In some embodiments, the anti-ENPP3X anti-CD3 bispecific antibody contains one, two, three, four, or more ENPP3-binding domains, including, but not limited to, those shown in Figures 12, 13A-13B, and 14A-14I. In certain embodiments, the anti-ENPP3X anti-CD3 antibody contains an ENPP3-binding domain containing the VHCDR1, VHCDR2, VHCDR3, VLCDR1, VLCDR2, and VLCDR3 sequences of an ENPP3-binding domain selected from the group consisting of those shown in Figures 12, 13A-13B, and 14A-14I. In some embodiments, an anti-ENPP3X anti-CD3 antibody comprises an ENPP3 binding domain comprising the underlined VHCDR1, VHCDR2, VHCDR3, VLCDR1, VLCDR2, and VLCDR3 sequences of an ENPP3 binding domain selected from the group consisting of those depicted in Figures 12, 13A-13B, and 14A-14I. In some embodiments, an anti-ENPP3X anti-CD3 antibody comprises an ENPP3 binding domain comprising the variable heavy and variable light domains of an ENPP3 binding domain selected from the group consisting of those depicted in Figures 12, 13A-13B, and 14A-14I. In exemplary embodiments, an anti-ENPP3X anti-CD3 antibody comprises an anti-ENPP3 AN1[ENPP3]_H1L1 binding domain.

[0159] The anti-ENPP3x anti-CD3 antibodies provided herein can comprise any suitable CD3-binding domain. In certain embodiments, an anti-ENPP3X anti-CD3 antibody comprises a CD3-binding domain comprising the VHCDR1, VHCDR2, VHCDR3, VLCDR1, VLCDR2, and VLCDR3 sequences of a CD3-binding domain selected from the group consisting of those depicted in Figures 10A-10F. In some embodiments, an anti-ENPP3X anti-CD3 antibody comprises a CD3-binding domain comprising the underlined VHCDR1, VHCDR2, VHCDR3, VLCDR1, VLCDR2, and VLCDR3 sequences of a CD3-binding domain selected from the group consisting of those depicted in Figures 10A-10F. In some embodiments, an anti-ENPP3X anti-CD3 antibody comprises a CD3-binding domain comprising the variable heavy chain domain and variable light chain domain of a CD3-binding domain selected from the group consisting of those depicted in Figures 10A-10F. In some embodiments, the CD3 binding domain is selected from anti-CD3 H1.30_L1.47, anti-CD3 H1.32_L1.47; anti-CD3H1.89_L1.48; anti-CD3H1.90_L1.47; anti-CD3H1.33_L1.47; and anti-CD3H1.31_L1.47. As outlined herein, these anti-CD3 antigen binding domains (CD3-ABD) can be used in either orientation in scFv format (e.g., from N- to C-terminus: VH-scFv linker-VL or VL-scFv linker-VH).

[0160] The antibodies provided herein comprise different antibody domains. As described herein and known in the art, the heterodimeric antibodies described herein comprise different domains within the heavy and light chains, which may also overlap. These domains include, but are not limited to, the Fc domain, CH1 domain, CH2 domain, CH3 domain, hinge domain, heavy chain constant domain (CH1-hinge-Fc domain or CH1-hinge-CH2-CH3), variable heavy chain domain, variable light chain domain, light chain constant domain, Fab domain, and scFv domain.

[0161] As provided herein, there are several suitable linkers (used as either domain linkers or scFv linkers) that can be used to covalently link (including traditional peptide bonds produced recombinantly) the listed domains (e.g., scFv, Fab, Fc domains, etc.). Exemplary linkers for attaching the domains of an antibody of interest to one another are shown in Figure 6. In some embodiments, the linker peptide may primarily comprise the following amino acid residues: Gly, Ser, Ala, or Thr. The linker peptide should be long enough to join the two molecules in a manner that allows them to assume the correct conformation relative to one another so that they retain the desired activity. In one embodiment, the linker is about 1 to 50 amino acids in length, preferably about 1 to 30 amino acids in length. In one embodiment, linkers of 1 to 20 amino acids in length may be used, with some embodiments using linkers of about 5 to about 10 amino acids. Useful linkers include, for example, glycine-serine polymers, glycine-alanine polymers, alanine-serine polymers, including (GS), (GSGGS), (GGGGS), and (GGGS), where n is an integer of at least 1 (and typically 3-4), as well as other flexible linkers, some of which are shown in Figures 5 and 6. Alternatively, a variety of non-proteinaceous polymers may be useful as linkers, including, but not limited to, polyethylene glycol (PEG), polypropylene glycol, polyoxyalkylenes, or copolymers of polyethylene glycol and polypropylene glycol.

[0162] Other linker sequences include any sequence of the CL / CH1 domain of any length, but may not include all residues of the CL / CH1 domain, such as the first 5-12 amino acid residues of the CL / CH1 domain. Linkers can be derived from immunoglobulin light chains, such as Cκ or Cλ. Linkers can be derived from immunoglobulin heavy chains of any isotype, including Cγ1, Cγ2, Cγ3, Cγ4, Cα1, Cα2, Cδ, Cε, and Cμ. Linker sequences can also be derived from other proteins, such as Ig-like proteins (e.g., TCR, FcR, KIR), hinge region-derived sequences, and other naturally occurring sequences from other proteins.

[0163] In some embodiments, the linker is a "domain linker" used to link together any two domains outlined herein. For example, in Figure 15B, there may be a domain linker connecting the C-terminus of the CH1 domain of the Fab to the N-terminus of the scFv, and another optional domain linker connects the C-terminus of the scFv to the CH2 domain (although in many embodiments, a hinge is used as this domain linker). While any suitable linker can be used, many embodiments utilize glycine-serine polymers as domain linkers, including, for example, (GS)n, (GSGGS)n, (GGGGS)n, and (GGGS)n, where n is an integer of at least 1 (generally 3-4-5), as well as any peptide sequence that allows for the recombinant attachment of two domains with sufficient length and flexibility so that each domain retains its biological function. In some cases, charged domain linkers can be used, as used in some embodiments of scFv linkers, with attention to "strandedness," as outlined below. Exemplary useful domain linkers are shown in Figure 6.

[0164] With particular reference to domain linkers used to connect the scFv domain to the Fc domain in a "2+1" format, there are several domain linkers that find particular use, including "Full Hinge C220S Variant," "Flexible Half Hinge," "Charged Half Hinge 1," and "Charged Half Hinge 2," as shown in Figure 6.

[0165] In some embodiments, the linker is an "scFv linker" used to covalently link the VH and VL domains discussed herein. In many cases, the scFv linker is a charged scFv linker, some of which are shown in Figure 5. Thus, in some embodiments, the antibodies described herein further provide a charged scFv linker to facilitate pI separation between the first and second monomers. That is, by incorporating either a positively or negatively charged scFv linker (or both, in the case of scaffolds that use scFvs on different monomers), this allows for the pI of the monomer containing the charged linker to be altered without further altering the Fc domain. These charged linkers can be substituted into any scFv containing a standard linker. Again, as will be understood by those skilled in the art, the charged scFv linker is used on the correct "chain" or monomer depending on the desired change in pI. For example, as discussed herein, to generate a 1+1 Fab-scFv-Fc format heterodimeric antibody, the original pI of the Fv region for each of the desired antigen-binding domains is calculated and one is selected to generate the scFv, and depending on the pI, either a positive or negative linker is selected.

[0166] Charged domain linkers can also be used to increase the pI separation of the antibody monomers described herein, and thus those included in Figure 5 can be used in any embodiment herein in which a linker is utilized.

[0167] In particular, the format shown in Figures 15A and 15B is an antibody commonly referred to as a "heterodimeric antibody," meaning that the protein has at least two related Fc sequences self-assembled into a heterodimeric Fc domain and at least two Fv regions, whether as Fab or as scFv.

[0168] The provided ENPP3-binding domains can be included in any useful antibody format, including, for example, standard immunoglobulins, as well as the 1+1 Fab-scFv-Fc and 2+1 Fab2-scFv-Fv formats provided herein. Other useful antibody formats include, but are not limited to, antibodies in the "mAb-Fv," "mAb-scFv," "central-Fv," "one-arm scFv-mAb," "scFv-mAb," "dual-scFv," and "trident" formats, as shown in Figures 52A-52K.

[0169] In some embodiments, the subject antibodies comprise one or more of the ENPP3 ABDs provided herein. In some embodiments, the antibodies comprise one ENPP3 ABD. In other embodiments, the antibodies comprise two ENPP3 ABDs. In exemplary embodiments, the ENPP3 ABD comprises the following ENPP3 binding domains: AN1[ENPP3]H1L1, AN1[ENPP3]H1 L1.33, AN1[ENPP3]H1 L1.77, AN1[ENPP3]H1.8 L1, AN1[ENPP3]H1.8 L1.33, AN1[ENPP3]H1 L1.77, H16-7.213, H16-9.69, H16-1.52, Ha16-1(1)23, H16-9.44, H16-1.67, Ha1 6-1(3,5)36, H16-1.86, H16-9.10, H16-9.33, H16-1.68, Ha16-1(1)1, Ha1 and H16-1(3,5)18, H16-1(2,4)4, H16-1(3,5)56, H16-7.8, H16-1.93, H16-1(3,5)27.1, H16-1.61, H16-1(3,5)5, ​​H16-7.200, H16-1(3,5)42, H16-9.65, H16-1(3,5)19, and H16-1.80 (Figures 12, 13A-13B, and 14A-14I).In exemplary embodiments, the ENPO3 ABD comprises the following ENPP3 binding domains: AN1[ENPP3]H1L1, AN1[ENPP3]H1 L1.33, AN1[ENPP3]H1 L1.77, AN1[ENPP3]H1.8 L1, AN1[ENPP3]H1.8 L1.33, AN1[ENPP3]H1 L1.77, H16-7.213, H16-9.69, H16-1.52, Ha16-1(1)23, H16-9.44, H16-1.67, Ha1 6-1(3,5)36, H16-1.86, H16-9.10, H16-9.33, H16-1.68, Ha16-1(1)1, Ha1 6-1(3,5)18, Ha16-1(2,4)4, Ha16-1(3,5)56, H16-7.8, H16-1.93, Ha1 6-1(3,5)27.1, H16-1.61, H16-1(3,5)5, ​​H16-7.200, H16-1(3,5)42, H1 6-9.65, Ha1-1(3,5)19, and Ha16-1.80 ( Figures 12 , 13A–13B , and 14A–14I ).

[0170] In exemplary embodiments, the antibody is a bispecific antibody comprising one or two ENPP3 ABDs, including any of the ENPP3 ABDs provided herein. Such ENPP3 ABD-containing bispecific antibodies include, for example, 1+1 Fab-scFv-Fc and 2+1 Fab2-scFv-Fc bispecific format antibodies. In exemplary embodiments, the ENPP3 ABD is one of the following B7H3 ABDs: AN1[ENPP3]H1L1, AN1[ENPP3]H1 L1.33, AN1[ENPP3]H1 L1.77, AN1[ENPP3]H1.8 L1, AN1[ENPP3]H1.8 L1.33, AN1[ENPP3]H1 L1.77, H16-7.213, H16-9.69, H16-1.52, Ha16-1(1)23, H16-9.44, H16-1.67, Ha1 6-1(3,5)36, H16-1.86, H16-9.10, H16-9.33, H16-1.68, Ha16-1(1)1, Ha1 6-1(3,5)18, Ha16-1(2,4)4, Ha16-1(3,5)56, H16-7.8, H16-1.93, Ha1 6-1(3,5)27.1, H16-1.61, H16-1(3,5)5, ​​H16-7.200, H16-1(3,5)42, H1 6-9.65, Ha1-1(3,5)19, and Ha16-1.80 (Figures 12, 13A-13B, and 14A-14I). In exemplary embodiments, the ENPP3-binding domain is a Fab. In some embodiments, such bispecific antibodies are heterodimeric bispecific antibodies comprising any of the heterodimerization variants, pI variants, and / or deletion variants described herein.

[0171] A. Chimeric and Humanized Antibodies In certain embodiments, the antibodies described herein comprise a heavy chain variable region derived from a particular germline heavy chain immunoglobulin gene and / or a light chain variable region derived from a particular germline light chain immunoglobulin gene. For example, such antibodies may comprise or consist of a human antibody comprising a heavy or light chain variable region that is the "product of" or "derived from" a particular germline sequence. A human antibody that is the "product of" or "derived from" a human germline immunoglobulin sequence can be identified as such (using the methods outlined herein) by comparing the amino acid sequence of the human antibody with the amino acid sequences of human germline immunoglobulins and selecting the human germline immunoglobulin that is closest in sequence to the human antibody (i.e., the highest percent identity). A human antibody that is the "product of" or "derived from" a particular human germline immunoglobulin sequence may contain amino acid differences compared to the germline sequence, for example, due to naturally occurring somatic mutations or the intentional introduction of site-specific mutations. However, a humanized antibody typically is at least 90% identical in amino acid sequence to the amino acid sequence encoded by a human germline immunoglobulin gene and includes amino acid residues that identify the antibody as derived from a human sequence when compared to the germline immunoglobulin amino acid sequence of another species (e.g., a murine germline sequence). In certain 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 the germline immunoglobulin gene, or even at least 96%, 97%, 98%, or even 99% identical. Typically, a humanized antibody derived from a particular human germline sequence will exhibit no more than 10-20 amino acid differences from the amino acid sequence encoded by the human germline immunoglobulin gene (before the introduction of any skew, pI, and deletion mutations herein, i.e., before the introduction of the variants described herein, the number of variants is generally small).In certain cases, a humanized antibody may display no more than 5 amino acids, or even no more than 4, 3, 2, or 1 amino acid difference from the amino acid sequence encoded by the germline immunoglobulin gene (again, the number of variants is generally small before the introduction of any skew, pI, and deletion mutations herein, i.e., before the introduction of the variants described herein).

[0172] In one embodiment, the parent antibody has been affinity matured as known in the art. Structure-based methods can be used for humanization and affinity maturation, for example, as described in USSN 11 / 004,590. Selection-based methods may be used to humanize and / or affinity mature antibody variable regions, 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 of which are incorporated by reference in their entirety. Other humanization methods may involve grafting only portions of the CDRs, including, but not limited to, those described in USSN 09 / 810,510; Tan et al., 2002, J. Immunol. 169:1119-1125; De Pascalis et al., 2002, J. Immunol. 169:3076-3084, all of which are incorporated by reference in their entirety.

[0173] B. Heterodimeric Antibodies In exemplary embodiments, the bispecific antibodies provided herein are heterodimeric bispecific antibodies comprising two variant Fc domain sequences, where such variant Fc domains comprise amino acid modifications to facilitate self-assembly and / or purification of the heterodimeric antibody.

[0174] A continuing challenge in antibody technology is the need for "bispecific" antibodies that can simultaneously bind to two different antigens, generally bringing the different antigens into close proximity and providing new functions and new therapeutic approaches. These antibodies are typically generated by including genes for each heavy and light chain in a host cell. This typically results in the formation of two homodimers (AA and BB (not including the issue of light chain heterodimers)) as well as the desired heterodimer (AB). However, a major obstacle in bispecific antibody formation is the difficulty of biasing heterodimer formation over homodimer formation and / or purifying heterodimeric antibodies away from homodimers.

[0175] There are many mechanisms that can be used to generate heterodimeric antibodies of interest. Moreover, as will be appreciated by those skilled in the art, these different mechanisms can be combined to ensure high heterodimerization. Amino acid modifications that facilitate heterodimer generation and purification are collectively referred to as "heterodimerization variants." As discussed below, heterodimerization variants include "skew" variants (e.g., "knob-and-hole" and "charge-pair" variants, discussed below) and "pI variants" that allow for purification of heterodimers from homodimers. As generally described in U.S. Pat. No. 9,605,084, which is incorporated herein by reference in its entirety, and specifically as follows for the discussion of heterodimerization variants, mechanisms useful for heterodimerization include "knob-and-hole" ("KIH") as described in U.S. Pat. No. 9,605,084, "electrostatic steering" or "charge pair" as described in U.S. Pat. No. 9,605,084, pI variants as described in U.S. Pat. No. 9,605,084, and general additional Fc variants as outlined in U.S. Pat. No. 9,605,084 and below.

[0176] Heterodimerization variants useful for forming and purifying the subject heterodimeric antibodies (e.g., bispecific antibodies) are discussed in further detail below.

[0177] 1. Scuba liant In some embodiments, the heterodimeric antibody comprises a scuba variant, which is one or more amino acid modifications in the first Fc domain (A) and / or the second Fc domain (B), that promote the formation of Fc dimers comprising the first and second Fc domains (AB) over Fc homodimers (Fc dimers comprising two of the first Fc domains or two of the second Fc domains; AA or BB). Suitable scuba variants are included in Figure 29 of U.S. Patent Application Publication No. 2016 / 0355608, which is incorporated herein by reference in its entirety and, particularly, for its disclosure of scuba variants, as well as in Figures 1A-1E and 4.

[0178] One mechanism, commonly referred to in the art as "knobs and holes," refers to amino acid engineering that results in steric effects favoring heterodimer formation and disfavoring homodimer formation, and can optionally be used; this is often referred to as "knobs and holes," 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; and USSN 8,216,805, all of which are incorporated herein by reference in their entireties. These diagrams identify several "monomer A-monomer B" pairs that rely on "knobs and holes." Furthermore, as described in Merchant et al., Nature Biotech. 16:677 (1998), these "knobs and holes" mutations can be combined with disulfide bond modification to skew formation toward heterodimerization.

[0179] An additional mechanism used to generate heterodimers is sometimes referred to as "electrostatic steering," as described in Gunasekaran et al., J. Biol. Chem. 285(25):19637 (2010), which is incorporated herein by reference in its entirety. This is sometimes referred to herein as "charge pairing." In this embodiment, electrostatics are used to skew formation toward heterodimerization. As will be appreciated by those skilled in the art, these may also affect pI, i.e., purification, and therefore may in some cases be considered pI variants. However, because these were generated to force heterodimerization and were not used as a purification tool, they are classified as "steric variants." These include, but are not limited to, D221E / P228E / L368E paired with D221R / P228R / K409R (e.g., these are the corresponding set of "monomers"), and C220E / P228E / 368E paired with C220R / E224R / P228R / K409R.

[0180] In some embodiments, scubariants advantageously and simultaneously promote heterodimerization based on both the "knob-and-hole" and "electrostatic steering" mechanisms. In some embodiments, heterodimeric antibodies contain one or more sets of such heterodimerizing scubariants. These variants are "pairs" of "sets." That is, pairs from one set are incorporated into a first monomer, and pairs from the other set are incorporated into a second monomer. It should be noted that these sets do not necessarily behave as "knob-and-hole" variants, but rather have a one-to-one correspondence between residues in one monomer and residues in the other. That is, these pairs of sets may instead form an interface between two monomers that promotes heterodimer formation but not homodimer formation, resulting in a rate of spontaneously forming heterodimers under biological conditions exceeding 90% instead of the expected 50% (25% homodimer A / A: 50% heterodimer A / B: 25% homodimer B / B). Exemplary heterodimerization "skew" variants are shown in Figure 4. In exemplary embodiments, the heterodimeric antibody comprises the following "skew" variant amino acid substitution set: 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 including a bridging disulfide, T366S / L368A / Y407V / Y349C:T366W / S354C). In an exemplary embodiment, the heterodimeric antibody comprises the "S364K / E357Q:L368D / K370S" amino acid substitution set. In terms of nomenclature, the pair "S364K / E357Q:L368D / K370S" means that one of the monomers comprises an Fc domain comprising the amino acid substitutions S364K and E357Q, and the other monomer comprises an Fc domain comprising the amino acid substitutions L368D and K370S. As noted above, the "twistiness" of these pairs depends on the starting pI.

[0181] In some embodiments, the scubariants provided herein are independently incorporated into one or both of the first and second Fc domains of an IL-15-Fc fusion protein along with other modifications to one or both of the first and second Fc domains of a heterodimeric antibody, including, but not limited to, other scubariants (see, e.g., Figure 37 of U.S. Patent Application Publication No. 2012 / 0149876, particularly incorporated herein by reference for its disclosure of scubariants), pI variants, isotype variants, FcRn variants, deletion variants, etc. Furthermore, individual modifications can also independently and optionally include or exclude the heterodimeric antibody.

[0182] The additional monomer A and monomer B variants can be optionally and independently combined in any amount with other variants, such as the pI variants outlined herein or other conformational variants shown in Figure 37 of US2012 / 0149876, all of whose figures, legends and SEQ ID NOs, are expressly incorporated herein by reference.

[0183] In some embodiments, the conformational mutations outlined herein can optionally and independently incorporate any pI mutations (or other mutations such as Fc mutations, FcRn mutations, etc.) into one or both monomers and can independently optionally be included or excluded from the antibody proteins described herein.

[0184] A list of suitable scuba variants can be found in Figures 1A-1E, and Figure 4 shows some pairs of particular utility in many embodiments. Of particular use in many embodiments are pairs in the set including, but not limited to: S364K / E357Q:L368D / K370S; L368D / K370S:S364K; L368E / K370S:S364K; T411T / E360E / Q362E:D401K; L368D / K370S:S364K / E357L and K370S:S364K / E357Q. In terms of nomenclature, the pair "S364K / E357Q:L368D / K370S" means that one monomer has the double variant set S364K / E357Q and the other has the double variant set L368D / K370S.

[0185] 2. Heterodimer pI (isoelectric point) variants In some embodiments, the heterodimeric antibody comprises a purification variant that advantageously allows for separation of the heterodimeric antibody (e.g., an anti-ENPP3 x anti-CD3 bispecific antibody) from homodimeric proteins.

[0186] There are several basic mechanisms that can facilitate the purification of heterodimeric antibodies. For example, modifications to one or both of the antibody heavy chain monomers A and B so that each monomer has a different pI allow for isoelectric purification of the heterodimeric AB antibody from the monomeric AA and BB proteins. Alternatively, some scaffold formats, such as the "1+1 Fab-scFv-Fc" and "2+1 Fab2-scFv-Fc" formats, also allow for size-based separation. As noted above, it is also possible to "skew" heterodimer formation over homodimers using scuba variants. Thus, the combination of heterodimerization scuba variants and pI variants finds particular use in the heterodimeric antibodies provided herein.

[0187] Additionally, depending on the heterodimeric antibody format, pI variants contained within the monomeric constant regions and / or Fc domains, and / or domain linkers can be used, as outlined more fully below. In some embodiments, heterodimeric antibodies contain additional modifications for alternative functions, such as Fc, FcRn, and KO variants, that can also create pI changes.

[0188] In some embodiments, the subject heterodimeric antibodies provided herein include at least one monomer having one or more modifications that alter the pI of the monomer (i.e., a "pI variant"). Generally, as will be understood by those of skill in the art, there are two general categories of pI variants: those that increase the pI of a protein (a basic change) and those that decrease the pI of a protein (an acidic change). As described herein, all combinations of these variants can be made: one monomer can be wild-type or a variant that does not exhibit a pI that is significantly different from wild-type, and the other can be either more basic or more acidic. Alternatively, each monomer can be altered, one to be more basic and one to be more acidic.

[0189] Depending on the heterodimeric antibody format, pI mutations can be included within the constant and / or Fc domains of the monomers, or either charged linkers, domain linkers, or scFv linkers can be used. That is, antibody formats utilizing scFvs, such as "1+1 Fab-scFv-Fc," can include a charged scFv linker (either positive or negative) that provides an additional pI boost for purification purposes. As will be appreciated by those skilled in the art, the antibodies described herein also provide pI variants and / or charged domain linkers in one or both of the monomers, although some 1+1 Fab-scFv-Fc formats are useful with only a charged scFv linker, without additional pI adjustment. Additionally, additional amino acid engineering for alternative functionality can also confer pI changes, such as Fc, FcRn, and KO variants.

[0190] In subject heterodimeric antibodies that utilize pI as a separation mechanism to enable purification of the heterodimeric protein, amino acid variants are introduced into one or both of the monomer polypeptides. That is, the pI of one of the monomers (referred to herein for simplicity as simply "monomer A") can be engineered away from monomer B, or the pI of both monomers A and B can be varied, increasing the pI of monomer A and decreasing the pI of monomer B. As outlined more fully below, pI changes in either or both monomers can be achieved by removing or adding a charged residue (e.g., substituting a neutral amino acid with a positively or negatively charged amino acid residue, e.g., glycine to glutamic acid), by changing a charged residue from positive or negative to the opposite charge (e.g., aspartic acid to lysine), or by changing a charged residue to a neutral residue (e.g., eliminating the charge, lysine to serine). Some of these variants are shown in Figures 3 and 4.

[0191] Thus, in some embodiments, the subject heterodimeric antibodies comprise an amino acid modification in the constant region that alters the isoelectric point (pI) of at least one, if not both, of the monomers of the dimeric protein by incorporating an amino acid substitution (a "pI variant" or "pI substitution") into one or both of the monomers. As demonstrated herein, separation of the heterodimer from the two homodimers can be achieved when the pIs of the two monomers differ by as little as 0.1 pH units, with differences of 0.2, 0.3, 0.4, and 0.5 or more being used in the antibodies described herein.

[0192] As will be understood by those skilled in the art, the number of pI mutations to be included in each or both monomers to obtain good separation will depend in part on the starting pIs of the components, e.g., the starting pIs of the scFv and Fab of interest in the 1+1 Fab-scFv-Fc and 2+1 Fab2-scFv-Fc formats. That is, to determine which monomers to engineer or in which "direction" (e.g., more positive or more negative), the Fv sequences of the two target antigens are calculated and a decision made from there. As is known in the art, different Fvs will have different starting pIs to be utilized in the antibodies described herein. Generally, as outlined herein, pIs are engineered to result in an overall pI difference for each monomer of at least about 0.1 log, with 0.2-0.5 being preferred, as outlined herein.

[0193] The use of heavy chain constant regions provides a more modular approach to designing and purifying bispecific proteins, including antibodies, when pI variants are used to achieve heterodimerization. Thus, in some embodiments, heterodimerization mutations (including skew and pI heterodimerization variants) are not included in the variable region, and therefore each individual antibody must be engineered. Additionally, in some embodiments, the potential for immunogenicity resulting from pI variants is significantly reduced by incorporating pI variants from different IgG isotypes to alter the pI without introducing significant immunogenicity. Therefore, a further 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 specific positions. Alternatively or in addition to isotype substitutions, the potential for immunogenicity resulting from pI variants is significantly reduced by utilizing isosteric substitutions (e.g., Asn to Asp and Gln to Glu).

[0194] As discussed below, potential collateral benefits of this pI engineering are also increased serum half-life and increased FcRn ligation. That is, as described in U.S. Patent Application Publication No. US2012 / 0028304 (incorporated herein by reference in its entirety), lowering 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 extending serum half-life also facilitate pI changes for purification.

[0195] Furthermore, pI variants offer additional benefits to the analytical and quality control processes of bispecific antibodies, as they offer significant capabilities for eliminating, minimizing, and distinguishing between homodimers when they are present. Similarly, the ability to reliably test the reproducibility of heterodimeric antibody production is important.

[0196] Generally, the particular embodiment used relies on a set of variants including a scuba variant, which, in combination with a pI variant that increases the pI difference between the two monomers, favors heterodimer formation in preference to homodimer formation, facilitating purification of the heterodimer by removal of the homodimer.

[0197] Exemplary combinations of pI variants are shown in Figures 4 and 5 and Figure 30 of U.S. Patent Application Publication No. 2016 / 0355608, all of which are incorporated herein by reference in their entirety, particularly for their disclosure of pI variants. Preferred combinations of pI variants are shown in Figures 1 and 2. As outlined herein and shown in the figures, these changes are shown relative to IgG1, but all isotypes can be modified in this manner, as can isotype hybrids. R133E and R133Q can also be used when the heavy chain constant domain is derived from IgG2-4.

[0198] In one embodiment, a preferred combination of pI variants has one monomer (negative Fab side) comprising the 208D / 295E / 384D / 418E / 421D variants (N208D / Q295E / N384D / Q418E / N421D when compared to human IgG1) and a second monomer (positive scFv side) comprising a positively charged scFv linker comprising (GKPGS)4 (SEQ ID NO:XX). However, as will be understood by those skilled in the art, the first monomer comprises a CH1 domain comprising position 208. Thus, in constructs that do not comprise a CH1 domain (e.g., in the case of antibodies that do not utilize a CH1 domain in one of their domains), a preferred negative pI mutated Fc set comprises the 295E / 384D / 418E / 421D mutations (Q295E / N384D / Q418E / N421D when compared to human IgG1).

[0199] Thus, in some embodiments, one monomer has a set of substitutions from Figure 2 and the other monomer has a charged linker (which can be selected from those shown in Figure 5, either in the format of an scFv or a charged scFv linker, as that monomer comprises a charged domain linker, as shown in the format).

[0200] In some embodiments, modifications occur in 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 variants, particularly substitutions, can be made at one or more of positions 216-230, with 1, 2, 3, 4, or 5 mutations finding use. Similarly, all possible combinations are contemplated, alone or with other pI variants in other domains.

[0201] Particular substitutions used to reduce the pI of the hinge domain include, but are not limited to, a deletion at position 221, a non-native valine or threonine at position 222, a deletion at position 223, a non-native glutamic acid at position 224, a deletion at position 225, a deletion at position 235, and a deletion or non-native alanine at position 236. In some cases, only pI substitutions are made in the hinge domain, while in other instances, these substitutions are added in any combination to other pI variants in other domains.

[0202] In some embodiments, mutations can be made within the CH2 region, including positions 233, 234, 235, 236, 274, 296, 300, 309, 320, 322, 326, 327, 334, and 339, based on EU numbering. Note that changes at 233-236 can be made to increase effector function (along with 327A) in an IgG2 backbone. Similarly, all possible combinations of these 14 positions can be made; for example, = can include variant Fc domains with 1, 2, 3, 4, 5, 6, 7, 8, 9, or 10 CH2 pI substitutions.

[0203] Specific substitutions used to reduce the pI of the CH2 domain include, but are not limited to, unnatural glutamine or glutamic acid at position 274, unnatural phenylalanine at position 296, unnatural phenylalanine at position 300, unnatural valine at position 309, unnatural glutamic acid at position 320, unnatural glutamic acid at position 322, unnatural glutamic acid at position 326, unnatural glycine at position 327, unnatural glutamic acid at position 334, unnatural threonine at position 339, and all possible combinations within CH2 and with other domains.

[0204] 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) in the CH3 region. Specific substitutions used to reduce the pI of the CH3 domain include, but are not limited to, an unnatural glutamine or glutamic acid at position 355, an unnatural serine at position 384, an unnatural asparagine or glutamic acid at position 392, an unnatural methionine at position 397, an unnatural glutamic acid at position 419, an unnatural glutamic acid at position 359, an unnatural glutamic acid at position 362, an unnatural glutamic acid at position 389, an unnatural glutamic acid at position 418, an unnatural glutamic acid at position 444, and a deletion or unnatural aspartic acid at position 447.

[0205] In general, as will be understood by those of skill in the art, there are two general categories of pI variants: those that increase the pI of a protein (basic changes) and those that decrease the pI of a protein (acidic changes). As described herein, all combinations of these variants can be made: one monomer can be wild-type or a variant that does not exhibit a pI that is significantly different from wild-type, and the other can be either more basic or more acidic. Alternatively, each monomer can be changed, one to be more basic and one to be more acidic.

[0206] Preferred combinations of pI variants are shown in Figure 4. As outlined herein and shown in the figure, these changes are shown relative to IgG1, but all isotypes can be modified in this manner, as can isotype hybrids. R133E and R133Q can also be used when the heavy chain constant domain is derived from IgG2-4.

[0207] In one embodiment, for example in the format of Figures 15A and 15B, a preferred combination of pI variants has one monomer (negative Fab side) comprising the 208D / 295E / 384D / 418E / 421D variants (N208D / Q295E / N384D / Q418E / N421D for human IgG1) and a second monomer (positive scFv side) comprising a positively charged scFv linker comprising (GKPGS)4 (SEQ ID NO:XXX). However, as will be appreciated by those skilled in the art, the first monomer comprises a CH1 domain comprising position 208. Thus, in constructs that do not include a CH1 domain (e.g., in the case of heterodimeric antibodies that do not utilize a CH1 domain in one of the domains, e.g., in dual scFv or "one-arm" formats as shown in Figure 42B, C, or D), a preferred negative pI variant Fc set includes the 295E / 384D / 418E / 421D variant (Q295E / N384D / Q418E / N421D for human IgG1).

[0208] Thus, in some embodiments, one monomer has a set of substitutions from Figure 4 and the other monomer has a charged linker (which can be selected from those shown in Figure 5, either in the format of an scFv or a charged scFv linker, as that monomer comprises a charged domain linker, as shown in the format).

[0209] 3. Isotype variants Furthermore, many embodiments of the antibodies described herein rely on the "import" of pI amino acids at specific positions from one IgG isotype to another, thus reducing or eliminating the possibility of introducing undesirable immunogenicity into the variant. Some of these are shown in Figure 21 of U.S. Patent Application Publication No. 2014 / 0370013, which is incorporated herein by reference. Specifically, IgG1 is a common isotype for therapeutic antibodies for a variety of reasons, including enhanced effector function. However, the heavy constant region of IgG1 has a higher pI than that of IgG2 (8.10 vs. 7.31). By introducing IgG2 residues into the IgG1 backbone at specific positions, the pI of the resulting monomer is lowered (or increased), and it also exhibits a longer serum half-life. For example, IgG1 has a glycine (pI 5.97) at position 137, while IgG2 has a glutamic acid (pI 3.22), and the introduction of glutamic acid affects the pI of the resulting protein. As described below, several amino acid substitutions are generally required to significantly affect the pI of the variant antibody, however, it should be noted that even changes within the IgG2 molecule can increase serum half-life, as discussed below.

[0210] In other embodiments, non-isotypic amino acid changes are made to reduce the overall charge state of the resulting protein (e.g., by changing from a high pI amino acid to a low pI amino acid) or to allow for structural tuning for stability, etc., as described in more detail below.

[0211] Additionally, by pI engineering both the heavy and light constant domains, significant changes can be seen in each monomer of the heterodimer. As discussed herein, a difference in pI of at least 0.5 between the two monomers can allow for separation by ion exchange chromatography or isoelectric focusing, or other methods sensitive to isoelectric point.

[0212] 4. Calculate pI The pI of each monomer depends on the pI of the variant heavy chain constant domain and the pI of all monomers, including the variant heavy chain constant domain and fusion partner. Thus, in some embodiments, the pI change is calculated based on the variant heavy chain constant domain using the chart in Figure 19 of US Patent Application Publication No. 2014 / 0370013. As discussed herein, the specific pI of the Fv and scaffold that determines which monomers to engineer is generally determined by the specific pI of the Fv and scaffold. Alternatively, the pI of each monomer can be compared.

[0213] 5. pI variants that also confer better in vivo binding to FcRn If the pI variants decrease the pI of the monomer, they may have the additional advantage of improving serum retention in vivo.

[0214] Although still under investigation, it is believed that the Fc region has a longer half-life in vivo because binding to FcRn at pH 6 in endosomes sequesters the Fc (Ghetie and Ward, 1997 Immunol Today. 18(12):592-598, incorporated by reference in its entirety). The endosomal compartment then recycles the Fc to the cell surface. Once the compartment opens to the extracellular space, a higher pH of approximately 7.4 induces the release of Fc into the blood. In mice, Dall'Acqua et al. showed that Fc variants with increased FcRn binding at pH 6 and pH 7.4 actually had reduced serum concentrations and half-lives identical to wild-type Fc (Dall'Acqua et al. 2002, J. Immunol. 169:5171-5180, incorporated by reference in its entirety). The increased affinity of Fc for FcRn at pH 7.4 is thought to prevent the release of Fc into the blood. Therefore, Fc mutations that increase the in vivo half-life of Fc would ideally increase FcRn binding at lower pH while still allowing Fc release at higher pH. The amino acid histidine changes its charge state in the pH range of 6.0 to 7.4. Therefore, it is not surprising to find His residues at key positions in the Fc / FcRn complex.

[0215] Recently, it has been suggested that antibodies with variable regions with lower isoelectric points may also have longer serum half-lives (Igawa et al., 2010 PEDS. 23(5):385-392, incorporated herein by reference in its entirety). However, the mechanism behind this is still unclear. Furthermore, variable regions vary from antibody to antibody. As described herein, constant region variants with reduced pI and extended half-lives may provide a more modular approach to improving the pharmacokinetic properties of antibodies.

[0216] C. Additional Fc variants for additional functionality In addition to the heterodimerization variants outlined above, there are several useful Fc amino acid modifications that can be made for a variety of reasons, including, but not limited to, altering binding to one or more FcγR receptors, altered binding to FcRn, as outlined below.

[0217] Thus, the antibodies (heterodimers, as well as homodimers) provided herein can contain such amino acid modifications with or without the heterodimerization variants (e.g., pI variants and conformational variants) outlined herein, each set of variants independently and optionally can be included or excluded from a particular heterodimeric protein.

[0218] 1. FcγR variants There are several useful Fc substitutions that can be made to alter binding to one or more FcγR receptors. In certain embodiments, the subject antibody contains a modification that alters binding to one or more FcγR receptors (i.e., "FcγR variants"). Substitutions that result in increased binding as well as decreased binding can be useful. For example, increased binding to FcγRIIIa is generally known to result in increased ADCC (antibody-dependent cell-mediated cytotoxicity, i.e., a cell-mediated reaction in which nonspecific cytotoxic cells expressing FcγR recognize ligated antibodies on target cells and subsequently cause lysis of the target cells). Similarly, decreased binding to FcγRIIb (an inhibitory receptor) can also be beneficial under some circumstances. Amino acid substitutions used in the antibodies described herein include those listed in U.S. Patent Nos. 8,188,321 (particularly Figure 41) and 8,084,582, and U.S. Patent Published Application Nos. 20060235208 and 20070148170, all of which are expressly incorporated herein by reference in their entirety, particularly with respect to the variants disclosed therein. Particular 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.

[0219] Furthermore, as specifically disclosed in U.S. Ser. No. 12 / 341,769, which is incorporated herein by reference in its entirety, there are additional Fc substitutions that are used to increase binding to the FcRn receptor and increase serum half-life, including, but not limited to, 434S, 434A, 428L, 308F, 259I, 428L / 434S, 259I / 308F, 436I / 428L, 436I or V / 434S, 436V / 428L, and 259I / 308F / 428L. Such modifications may be included in one or both Fc domains of the subject antibodies.

[0220] 2. Elimination Variant Similarly, another category of functional variants is "FcγR-depleted 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 Fcγ receptors (e.g., FcγR1, FcγRIIa, FcγRIIb, FcγRIIIa, etc.) to avoid additional mechanisms of action. That is, for example, in many embodiments, particularly in the use of bispecific antibodies that monovalently bind to CD3, it is generally desirable to eliminate FcγRIIIa binding to eliminate or significantly reduce ADCC activity, and one of the Fc domains comprises one or more Fcγ receptor-depleted variants. These deletion variants are shown in Figure 14, and each can be independently optionally included or excluded in preferred embodiments using a deletion variant selected from the group consisting of 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. Note that the deletion variants referred to herein eliminate FcγR binding, but not FcRn binding.

[0221] As is known in the art, the Fc domain of human IgG1 has the highest binding to Fcγ receptors, and therefore, when the constant domain (or Fc domain) of the scaffold of a heterodimeric antibody is IgG1, deletion variants can be used. Alternatively, or in addition to deletion variants of the IgG1 background, mutation at glycosylation position 297 (generally to A or S) can, for example, significantly eliminate binding to FcγRIIIa. Because human IgG2 and IgG4 have naturally reduced binding to Fcγ receptors, these scaffolds can be used with or without deletion variants.

[0222] D. Combination of Heterodimers and Fc Variants As will be appreciated by those skilled in the art, all of the listed heterodimerization variants (including scubariants and / or pI variants) can be optionally and independently combined, so long as they maintain their "strandedness" or "monomer split." In some embodiments, the heterodimeric antibodies provided herein comprise a combination of heterodimerization scubariants, isosteric pI substitutions, and FcKO variants, as shown in Figure 4. Furthermore, all of these variants can be combined in any of the heterodimerization formats.

[0223] In the case of pI variants, while a specifically used embodiment is shown in the figure, other combinations can be generated following the basic rule of varying the pI difference between the two monomers to facilitate purification.

[0224] Additionally, any of the heterodimerization variants, skews, and pIs can be independently and optionally combined with Fc-depleted variants, Fc variants, FcRn variants, as generally outlined herein.

[0225] Exemplary combinations of variants included in some embodiments of heterodimeric 1+1 Fab-scFv-Fc and 2+1 Fab2-scFv-Fc format antibodies are included in Figure 4. In certain embodiments, the antibody is a heterodimeric 1+1 Fab-scFv-Fc or 2+1 Fab2-scFv-Fc format antibody, as shown in Figures 15A and 15B.

[0226] E. Anti-ENPP3 x anti-CD3 bispecific antibody In another aspect, provided herein are anti-ENPP3 x anti-CD3 (also referred to herein as "αENPP3 x αCD3") bispecific antibodies. Such antibodies comprise at least one ENPP3-binding domain and at least one CD3-binding domain. In some embodiments, the bispecific αENPP3 x αCD3 described herein provided an immune response selectively at tumor sites expressing ENPP3.

[0227] Note that unless indicated herein, the order of the named antigen listing does not confer structure, i.e., an ENPP3XCD3 1+1Fab-scFv-Fc antibody may bind scFv to ENPP3 or CD3, but in some cases the order as indicated specifies the structure.

[0228] As outlined in more detail herein, these combinations of ABDs can be in a variety of formats, generally with one ABD being in Fab format and the other in scFv format, as outlined below. Exemplary formats for use in the bispecific antibodies provided herein include the 1+1 Fab-scFv-Fc and 2+1 Fab2-scFv-Fv formats (see, e.g., Figures 15A and 15B). Other useful antibody formats include, but are not limited to, antibodies in the "mAb-Fv," "mAb-scFv," "central-Fv," "one-arm scFv-mAb," "scFv-mAb," "dual-scFv," and "trident" formats, as shown in Figures 52A-52K.

[0229] Furthermore, one of the ABDs will generally comprise an scFv as outlined herein in an N- to C-terminal orientation of VH-scFv linker-VL or VL-scFv linker-VH, and depending on the format, one or both of the other ABDs will generally be a Fab comprising a VH domain (generally as part of a heavy chain) on one protein chain and a VL (generally as part of a light chain) on another protein chain.

[0230] As will be appreciated by those skilled in the art, any set of six CDRs or VH and VL domains can be made into scFv or Fab format, which are then added to heavy and light chain constant domains, where the heavy chain constant domains contain mutations (including within the CH1 and Fc domains). The scFv sequences included in the sequence listing utilize specific charged linkers, although uncharged or other charged linkers can be used, as outlined herein, including those shown in Figures 5 and 6.

[0231] Furthermore, as noted above, the numbering used in the sequence listing for identifying the CDRs is Kabat, however, different numbering can be used, which would alter the amino acid sequence of the CDRs as shown in Table 2.

[0232] Further variants can be created for all variable heavy and light chain domains described herein. As outlined herein, in some embodiments, the set of six CDRs has 0, 1, 2, 3, 4, or 5 amino acid modifications (depending on the particular amino acid substitution used), and the framework regions of the variable heavy and light chain domains can have variations in the framework regions, as long as the framework (excluding the CDRs) retains at least about 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, which incorporates the figure and legend herein by reference in its entirety. Thus, for example, the same CDRs described herein can be combined with different framework sequences derived from human germline sequences, as long as 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, the CDRs can have amino acid modifications (e.g., 1, 2, 3, 4, or 5 amino acid modifications in the set of CDRs (i.e., the CDRs can have any combination of CDR modifications, e.g., 1 modification in vlCDR1, 2 modifications in vhCDR2, no modifications in vhCDR3, etc., as long as the total number of modifications in the set of 6 CDRs is less than 6 amino acid modifications)), as well as modifications in the framework regions, as long as the framework regions retain at least 80, 85, or 90% identity to a human germline sequence selected from those listed in Figure 1 of U.S. Pat. No. 7,657,380.

[0233] Anti-ENPP3 x anti-CD3 bispecific antibodies can comprise any suitable CD3 ABD, including those described herein (see, e.g., Figures 10A-10F). In some embodiments, the CD3 ABD of an anti-ENPP3 x anti-CD3 bispecific antibody comprises the variable heavy and light domains of a CD3 ABD provided herein, including those set forth in Figures 10A-10F and the Sequence Listing. In some embodiments, the CD3 ABD comprises the variable heavy and light domains of one of the following CD3 binding domains: H1.30_L1.47, H1.32_L1.47, H1.89_L1.47, H1.90_L1.47, H1.33_L1.47, H1.31_L1.47, L1.47_H1.30, L1.47_H1.30, L1.47_H1.32, L1.47_H1.89, L1.47_H1.90, L1.47_H1.33, and L1.47_H1.31 (Figures 10A-10F). In exemplary embodiments, the CD3 ABD is one of the following CD3 binding domains: H1.30_L1.47, H1.32_L1.47, H1.89_L1.47, H1.90_L1.47, H1.33_L1.47, H1.31_L1.47, L1.47_H1.30, L1.47_H1.30, L1.47_H1.32, L1.47_H1.89, L1.47_H1.90, L1.47_H1.33, and L1.47_H1.31 (Figures 10A-10F) or variants thereof. Anti-ENPP3 x anti-CD3 bispecific antibodies can comprise any suitable ENPP3 ABD, including those described herein (see, e.g., Figures 12, 13A-13B, and 14A-14I). In some embodiments, the ENPP3 ABD of an anti-ENPP3 x anti-CD3 bispecific antibody comprises the variable heavy and light domains of an ENPP3 ABD provided herein, including those set forth in Figures 12, 13A-13B, and 14A-14I and in the Sequence Listing. In some embodiments, the ENPP3 ABD is one of the following ENPP3 ABDs:AN1[ENPP3]H1L1, AN1[ENPP3]H1 L1.33, AN1[ENPP3]H1 L1.77, AN1[ENPP3]H1.8 L1, AN1[ENPP3]H1.8 L1.33, AN1[ENPP3]H1 L1.77, H16-7.213, H16-9.69, H16-1.52, Ha16-1(1)23, H16-9.44, H16-1.67, Ha1 6-1(3,5)36, H16-1.86, H16-9.10, H16-9.33, H16-1.68, Ha16-1(1)1, Ha1 and H16-1(3,5)18, H16-1(2,4)4, H16-1(3,5)56, H16-7.8, H16-1.93, H16-1(3,5)27.1, H16-1.61, H16-1(3,5)5, ​​H16-7.200, H16-1(3,5)42, H16-9.65, H16-1(3,5)19, and H16-1.80 (Figures 12, 13A-13B, and 14A-14I).In an exemplary embodiment, the ENPP3 ABD is one of the following ENPP3 ABDs: AN1[ENPP3]H1L1, AN1[ENPP3]H1 L1.33, AN1[ENPP3]H1 L1.77, AN1[ENPP3]H1.8 L1, AN1[ENPP3]H1.8 L1.33, AN1[ENPP3]H1 L1.77, H16-7.213, H16-9.69, H16-1.52, Ha16-1(1)23, H16-9.44, H16-1.67, Ha1 6-1(3,5)36, H16-1.86, H16-9.10, H16-9.33, H16-1.68, Ha16-1(1)1, Ha1 6-1(3,5)18, Ha16-1(2,4)4, Ha16-1(3,5)56, H16-7.8, H16-1.93, Ha1 6-1(3,5)27.1, H16-1.61, H16-1(3,5)5, ​​H16-7.200, H16-1(3,5)42, H1 6-9.65, Ha1-1(3,5)19, and Ha16-1.80 ( Figures 12 , 13A–13B , and 14A–14I ).

[0234] F. Anti-SSTR2 x anti-CD3 bispecific antibody In another aspect, provided herein is an anti-SStr2 x anti-CD3 (also referred to herein as "αSSTR2 x αCD3") bispecific antibody. Such an antibody comprises at least one SStr2-binding domain and at least one CD3-binding domain. In some embodiments, the bispecific αSSTR2 x αCD3 provided an immune response selectively at tumor sites expressing SSTR2.

[0235] Note that unless indicated herein, the order of the named antigen listing does not confer structure, i.e., SSTR2XCD3 1+1 Fab-scFv-Fc antibody may have scFv binding to SSTR2 or CD3, but in some cases the order as indicated specifies the structure.

[0236] As outlined in more detail herein, these combinations of ABDs can be in a variety of formats, generally with one ABD being in Fab format and the other in scFv format, as outlined below. Exemplary formats for use in the bispecific antibodies provided herein include the 1+1 Fab-scFv-Fc and 2+1 Fab2-scFv-Fv formats (see, e.g., Figures 15A and 15B). Other useful antibody formats include, but are not limited to, antibodies in the "mAb-Fv," "mAb-scFv," "central-Fv," "one-arm scFv-mAb," "scFv-mAb," "dual-scFv," and "trident" formats, as shown in Figures 52A-52K.

[0237] Furthermore, one of the ABDs will generally comprise an scFv as outlined herein in an N- to C-terminal orientation of VH-scFv linker-VL or VL-scFv linker-VH, and depending on the format, one or both of the other ABDs will generally be a Fab comprising a VH domain (generally as part of a heavy chain) on one protein chain and a VL (generally as part of a light chain) on another protein chain.

[0238] As will be appreciated by those skilled in the art, any set of six CDRs or VH and VL domains can be made into scFv or Fab format, which are then added to heavy and light chain constant domains, where the heavy chain constant domains contain mutations (including within the CH1 and Fc domains). The scFv sequences included in the sequence listing utilize specific charged linkers, although uncharged or other charged linkers can be used, as outlined herein, including those shown in Figures 5 and 6.

[0239] Furthermore, as noted above, the numbering used in the sequence listing for identifying the CDRs is Kabat, however, different numbering can be used, which would alter the amino acid sequence of the CDRs as shown in Table 2.

[0240] Further variants can be created for all variable heavy and light chain domains described herein. As outlined herein, in some embodiments, the set of six CDRs has 0, 1, 2, 3, 4, or 5 amino acid modifications (depending on the particular amino acid substitution used), and the framework regions of the variable heavy and light chain domains can have variations in the framework regions, as long as the framework (excluding the CDRs) retains at least about 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, which incorporates the figure and legend herein by reference in its entirety. Thus, for example, the same CDRs described herein can be combined with different framework sequences derived from human germline sequences, as long as 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, the CDRs can have amino acid modifications (e.g., 1, 2, 3, 4, or 5 amino acid modifications in the set of CDRs (i.e., the CDRs can have any combination of CDR modifications, e.g., 1 modification in vlCDR1, 2 modifications in vhCDR2, no modifications in vhCDR3, etc., as long as the total number of modifications in the set of 6 CDRs is less than 6 amino acid modifications)), as well as modifications in the framework regions, as long as the framework regions retain at least 80, 85, or 90% identity to a human germline sequence selected from those listed in Figure 1 of U.S. Pat. No. 7,657,380.

[0241] Anti-SSTR2 x anti-CD3 bispecific antibodies can comprise any suitable CD3 ABD, including those described herein (see, e.g., Figures 10A-10F). In some embodiments, the CD3 ABD of an anti-SSTR2 x anti-CD3 bispecific antibody comprises the variable heavy and light domains of a CD3 ABD provided herein, including those set forth in Figures 10A-10F and the Sequence Listing. In some embodiments, the CD3 ABD comprises the variable heavy and light domains of one of the following CD3 binding domains: H1.30_L1.47, H1.32_L1.47, H1.89_L1.47, H1.90_L1.47, H1.33_L1.47, H1.31_L1.47, L1.47_H1.30, L1.47_H1.30, L1.47_H1.32, L1.47_H1.89, L1.47_H1.90, L1.47_H1.33, and L1.47_H1.31 (Figures 10A-10F). In exemplary embodiments, the CD3 ABD is one of the following CD3 binding domains: H1.30_L1.47, H1.32_L1.47, H1.89_L1.47, H1.90_L1.47, H1.33_L1.47, H1.31_L1.47, L1.47_H1.30, L1.47_H1.30, L1.47_H1.32, L1.47_H1.89, L1.47_H1.90, L1.47_H1.33, and L1.47_H1.31 (Figures 10A-10F) or variants thereof. The anti-SSTR2 x anti-CD3 bispecific antibody may comprise the variable heavy and light domains of [αSSTR2]H1.24_L1.30 (FIG. 63), or variants thereof.

[0242] G. Useful Formats of the Invention As will be appreciated by those of skill in the art and described in more detail below, the bispecific heterodimeric antibodies provided herein can adopt a wide variety of configurations, generally as shown in Figure 1. Some figures depict "single-ended" configurations, with one type of specificity in one "arm" of the molecule and a different specificity in the other "arm." Other figures depict "dual-ended" configurations, with at least one type of specificity in the "top" of the molecule and one or more different specificities in the "bottom" of the molecule. Thus, some embodiments are directed to novel immunoglobulin compositions that simultaneously link different first and second antigens.

[0243] As will be understood by those skilled in the art, the heterodimeric formats of the present invention may have different valencies and may be bispecific. That is, the heterodimeric antibodies of the antibodies described herein may be bivalent and bispecific, with one target tumor antigen (e.g., CD3) bound by one binding domain and the other target tumor antigen (e.g., ENPP3) bound by a second binding domain. Heterodimeric antibodies may also be trivalent and bispecific, with a first antigen bound by two binding domains and a second antigen bound by a second binding domain. As outlined herein, if CD3 is one of the target antigens, it is preferable that CD3 bind only monovalently to reduce potential side effects.

[0244] The antibodies described herein utilize an anti-CD3 antigen-binding domain in combination with an anti-ENPP3 binding domain. As will be appreciated by those skilled in the art, any collection of anti-CD3 CDRs, anti-CD3 variable light and heavy domains, Fabs, and scFvs, as shown in any of the figures, can be used. Similarly, any of the anti-ENPP3 antigen-binding domains can be used, and any of the CDRs, variable light and heavy domains, Fabs, and scFvs, as shown in any of the figures (e.g., Figures 12, 13A-13B, and 14A-14I), can be used in any combination, independently.

[0245] 1.1+1 Fab-scFv-Fc format One heterodimeric scaffold that finds particular use with the antibodies described herein is the "1+1 Fab-scFv-Fc" or "bottle opener" format, as shown in Figure 15A, with an exemplary combination of a CD3-binding domain and a tumor target antigen (ENPP3)-binding domain. In this embodiment, one heavy chain monomer of the antibody comprises a single-chain Fv ("scFv," as defined below) and an Fc domain. The scFv comprises a variable heavy domain (VH1) and a variable light domain (VL1), with VH1 linked to VL1 using a reversible scFv linker (see, e.g., Figure 5). The scFv is linked to the heavy chain using a domain linker (see, e.g., Figure 6). The other heavy chain monomer is a "normal" heavy chain (VH-CH1-hinge-CH2-CH3). The 1+1 Fab-scFv-Fc also comprises a light chain that interacts with VH-CH1 to form Fab. This structure is sometimes referred to herein as the "bottle opener" (format) due to its rough visual resemblance to a bottle opener. The two heavy chain monomers are held together by the use of amino acid variants (e.g., the heterodimerization variants described above) in the constant regions (e.g., the Fc domain, CH1 domain, and / or hinge region) that promote the formation of heterodimeric antibodies, as described in more detail below.

[0246] The current "1+1 Fab-scFv-Fc" format offers several distinct advantages. As is known in the art, antibody analogs that rely on two scFv constructs often have stability and aggregation issues that can be mitigated in the antibodies described herein by the addition of "normal" heavy and light chain pairing. Furthermore, in contrast to formats that rely on two heavy chains and two light chains, there is no problem with heavy and light chain mispairing (e.g., heavy chain 1 pairing with light chain 2).

[0247] Many of the embodiments outlined herein generally rely on 1+1 Fab-scFv-Fc or "bottle opener" format antibodies, which comprise a first monomer comprising an scFv, which comprises variable heavy and variable light chain domains covalently linked using an scFv linker (often, but not always, charged), and the scFv is covalently linked to a first Fc domain, usually via a domain linker. The domain linker may be charged or uncharged, and may be exogenous or endogenous (e.g., all or part of a native hinge domain). Any suitable linker can be used to attach the scFv to the N-terminus of the first Fc domain. In some embodiments, the domain linker is selected from the domain linkers in Figure 6. The second monomer of the 1+1 Fab-scFv-Fc or "bottle opener" format is a heavy chain, and the composition further comprises a light chain.

[0248] Generally, in many preferred embodiments, the scFv is the CD3-binding domain and the Fab forms the ENPP3-binding domain. An exemplary anti-ENPP3 x anti-CD3 bispecific antibody in a 1+1 Fab-scFv-Fc format is shown in Figure 15A. An exemplary anti-ENPP3 x anti-CD3 bispecific antibody in a 1+1 Fab-scFv-Fc format is shown in Figures 17A-17C and 18A-18C.

[0249] Additionally, the Fc domains of the antibodies described herein generally comprise scubariant sequences (e.g., sets of amino acid substitutions such as those shown in Figures 3 and 9, with particularly useful scubariant sequences being 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), optionally a deletion variant (including those shown in Figure 3), optionally a charged scFv linker (including those shown in Figure 5), and the heavy chain comprises a pI variant (including those shown in Figure 4).

[0250] In certain embodiments, the 1+1 Fab-scFv-Fc scaffold format comprises a first monomer comprising an scFv domain linker-CH2-CH3 monomer, a second monomer comprising a first variable heavy domain-CH1-hinge-CH2-CH3, and a third monomer comprising a first variable light domain. In some embodiments, the CH2-CH3 of the first monomer is a first variant Fc domain, and the CH2-CH3 of the second monomer is a second variant Fc domain. In some embodiments, the scFv comprises an scFv variable heavy domain and an scFv variable light domain that form a CD3-binding moiety. In certain embodiments, the scFv variable heavy domain and the scFv variable light domain are covalently linked using an scFv linker (which is often, but not always, charged). See, e.g., Figure 5. In some embodiments, the first variable heavy domain and the first variable light domain form an ENPP3-binding domain. Particularly useful ENPP3 and CD3 combinations for use in the 1+1 Fab-scFv-Fc ENPP3 x CD3 bispecific antibody format are disclosed in Figures 17A-17C and Figures 18A-18C and include ENPP3 H16-1.93 x CD3 H1.30 L1.47, ENPP3 H16-7.8 x CD3 H1.30 L1.47, ENPP3 AN1[ENPP3]H1L1 x CD3 H1.30 L1.47, ENPP3 AN1[ENPP3]H1.8 L1 x CD3 H1.30 L1.47, ENPP3 AN1[ENPP3]H1.8 L1 x CD3 H1.30 L1.47, ENPP3 AN1[ENPP3]H1.8 L1. x CD3 H1.30 L1.47, and ENPP3 AN1[ENPP3]H1.8 L1.33 x CD2 H1.30 L1.47, and ENPP3 H1.8 L1.77×CD3 H.130 L1.47. In some embodiments, the 1+1 Fab-scFv-Fc format comprises a scFv variant, a pI variant, and a deletion variant.Thus, some embodiments are 1+1 Fab-scFv-Fc formats comprising: a) a first monomer ("scFv monomer") comprising a charged scFv linker (in some embodiments the +H sequence of Figure 5 is preferred), scFv variants S364K / E357Q, deletion variants E233P / L234V / L235A / G236del / S267K, and an scFv that binds CD3 as outlined herein; and b) a second monomer ("scFv monomer") comprising a scFv linker that binds CD3 as outlined herein; and c) a second monomer ("scFv monomer") comprising a charged scFv linker (in some embodiments the +H sequence of Figure 5 is preferred), scFv variants S364K / E357Q, deletion variants E233P / L234V / L235A / G236del / S267K, and an scFv that binds CD3 as outlined herein. and a 1+1 Fab-scFv-Fc format comprising: a second monomer ("Fab monomer") comprising the pI variants N208D / Q295E / N384D / Q418E / N421D, the deletion variants E233P / L234V / L235A / G236del / S267K, and a variable heavy domain that, together with the variable light domain, constitutes an Fv that binds to a second antigen as outlined herein; and c) a light chain comprising a variable light domain (VL) and a constant light domain (CL), where numbering is according to EU numbering. The variable heavy domain and the variable light domain constitute the ENPP3-binding moiety. CD3 binding domain sequences that find particular use in these embodiments include, but are not limited to, H1.30_L1.47, H1.32_L1.47, H1.89_L1.47, H1.90_L1.47, H1.33_L1.47, H1.31_L1.47, L1.47_H1.30, L1.47_H1.30, L1.47_H1.32, L1.47_H1.89, L1.47_H1.90, L1.47_H1.33, and L1.47_H1.31 and those shown in Figures 10A-10F.ENPP3 binding domain sequences of particular use in these embodiments include AN1[ENPP3]H1L1, AN1[ENPP3]H1 L1.33, AN1[ENPP3]H1 L1.77, AN1[ENPP3]H1.8 L1, AN1[ENPP3]H1.8 L1.33, AN1[ENPP3]H1 L1.77, H16-7.213, H16-9.69, H16-1.52, Ha16-1(1)23, H16-9.44, H16-1.67, Ha1 6-1(3,5)36, H16-1.86, H16-9.10, H16-9.33, H16-1.68, Ha16-1(1)1, Ha1 Examples of such fragments include, but are not limited to, Ha16-1(3,5)18, Ha16-1(2,4)4, Ha16-1(3,5)56, H16-7.8, H16-1.93, Ha1 6-1(3,5)27.1, H16-1.61, H16-1(3,5)5, ​​H16-7.200, H16-1(3,5)42, H1 6-9.65, Ha1-1(3,5)19, and Ha16-1.80 (Figures 12, 13A-13B, and 14A-14I). Particularly useful combinations of ENPP3 and CD3 sequences for use with 1+1 Fab2-scFv-Fc format antibodies include, for example, ENPP3 H16-1.93 x CD3 H1.30 L1.47, ENPP3 H16-7.8 x CD3 H1.30 L1.47, ENPP3 AN1[ENPP3]H1L1 x CD3 H1.30 L1.47, ENPP3 AN1[ENPP3]H1.8 L1 x CD3 H1.30 L1.47, ENPP3 AN1[ENPP3]H1.8 L1.33 x CD3 H1.30 L1.47, and ENPP3 H1.8 L1.77 x CD3 H.130 L1.47.

[0251] Exemplary variable heavy and light chain domains of scFvs that bind CD3 are included in Figures 10A-10F. Exemplary variable heavy and light chain domains of Fvs that bind ENPP3 are included in Figures 12, 13A-13B, and 14A-14I. In an exemplary embodiment, the ENPP3 binding domain of the 1+1 Fab-scFv-Fc ENPP3 x CD3 bispecific antibody is selected from the group consisting of the following ENPP3 binding domains: AN1[ENPP3]H1L1, AN1[ENPP3]H1 L1.33, AN1[ENPP3]H1 L1.77, AN1[ENPP3]H1.8 L1, AN1[ENPP3]H1.8 L1.33, AN1[ENPP3]H1 L1.77, H16-7.213, H16-9.69, H16-1.52, Ha16-1(1)23, H16-9.44, H16-1.67, Ha1 The VH and VL sequences contained one of the following: H16-1(3,5)36, H16-1.86, H16-9.10, H16-9.33, H16-1.68, Ha16-1(1)1, Ha1 6-1(3,5)18, Ha16-1(2,4)4, Ha16-1(3,5)56, H16-7.8, H16-1.93, Ha1 6-1(3,5)27.1, H16-1.61, H16-1(3,5)5, ​​H16-7.200, H16-1(3,5)42, H1 6-9.65, Ha1-1(3,5)19, and Ha16-1.80 (Figures 12, 13A-13B, and 14A-14I). In one embodiment, the CD3 binding domain of the 1+1 Fab-scFv-Fc ENPP3 x CD3 bispecific antibody comprises a VH and a VL of one of the following CD3 binding domains: H1.30_L1.47, H1.32_L1.47, H1.89_L1.47, H1.90_L1.47, H1.33_L1.47, H1.31_L1.47, L1.47_H1.30, L1.47_H1.30, L1.47_H1.32, L1.47_H1.89, L1.47_H1.90, L1.47_H1.33, and L1.47_H1.31 (Figures 10A-10F).Particularly useful ENPP3 and CD3 combinations for use in the 1+1 Fab-scFv-Fc ENPP3×CD3 bispecific antibody format are disclosed in Figures 17A-17C and 18A-18C and include ENPP3 H16-1.93×CD3 H1.30 L1.47, ENPP3 H16-7.8×CD3 H1.30 L1.47, ENPP3 AN1[ENPP3]H1L1×CD3 H1.30 L1.47, ENPP3 AN1[ENPP3]H1.8 L1×CD3 H1.30 L1.47, ENPP3 AN1[ENPP3]H1.8 L1.33×CD3 H1.30 L1.47, and ENPP3 H1.8 L1.77×CD3 H.130 L1.47.

[0252] In some embodiments, the 1+1 Fab-scFv-Fc format comprises a scFv variant, a pI variant, a deletion variant, and an FcRn variant. Thus, some embodiments are 1+1 Fab-scFv-Fc formats comprising: a) a first monomer ("scFv monomer") comprising a charged scFv linker (in some embodiments the +H sequence of Figure 6 is preferred), a scFv variant S364K / E357Q, a deletion variant E233P / L234V / L235A / G236del / S267K, an FcRn variant M428L / N434S, and an scFv that binds to CD3 as outlined herein; and a) a second monomer ("Fab monomer") comprising a variable heavy domain, and b) a light chain comprising a variable light domain (VL) and a constant light domain (CL), where numbering is according to EU numbering. The variable heavy domain and the variable light domain constitute the ENPP3-binding domain. CD3 binding domain sequences that find particular use in these embodiments include, but are not limited to, H1.30_L1.47, H1.32_L1.47, H1.89_L1.47, H1.90_L1.47, H1.33_L1.47, H1.31_L1.47, L1.47_H1.30, L1.47_H1.30, L1.47_H1.32, L1.47_H1.89, L1.47_H1.90, L1.47_H1.33, and L1.47_H1.31 and those shown in Figures 10A-10F.ENPP3 binding domain sequences of particular use in these embodiments include AN1[ENPP3]H1L1, AN1[ENPP3]H1 L1.33, AN1[ENPP3]H1 L1.77, AN1[ENPP3]H1.8 L1, AN1[ENPP3]H1.8 L1.33, AN1[ENPP3]H1 L1.77, H16-7.213, H16-9.69, H16-1.52, Ha16-1(1)23, H16-9.44, H16-1.67, Ha1 6-1(3,5)36, H16-1.86, H16-9.10, H16-9.33, H16-1.68, Ha16-1(1)1, Ha1 6-1(3,5)18, Ha16-1(2,4)4, Ha16-1(3,5)56, H16-7.8, H16-1.93, Ha1 6-1(3,5)27.1, H16-1.61, H16-1(3,5)5, ​​H16-7.200, H16-1(3,5)42, H1 6-9.65, Ha1-1(3,5)19, and Ha16-1.80 (as shown in Figures 12, 13A-13B, and 14A-14I). Specific useful ENPP3 and CD3 sequence combinations for use in antibodies in the 1+1 Fab-scFv-Fc format are disclosed, for example, in Figures 17A-17C and 18A-18C, and include ENPP3 H16-1.93 x CD3 H1.30 L1.47, ENPP3 H16-7.8 x CD3 H1.30 L1.47, ENPP3 AN1[ENPP3]H1L1 x CD3 H1.30 L1.47, ENPP3 AN1[ENPP3]H1.8 L1 x CD3 H1.30 L1.47, ENPP3 AN1[ENPP3]H1.8 L1.33 x CD3 H1.30 L1.47, and ENPP3 H1.8 L1.77 x CD3 H.130 L1.47.

[0253] Figures 7A-7D show some exemplary Fc domain sequences useful in 1+1 Fab-scFv-Fc format antibodies. The "monomer 1" sequence shown in Figures 7A-7D generally 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." Additionally, Figure 9 shows useful CL sequences that can be used in this format.

[0254] In some embodiments, any of the VH and VL sequences shown herein (including all VH and VL sequences shown in the figures and sequence listing, including those directed against ENPP3) can be added as the "Fab side" to the bottle opener scaffold format of Figures 7A-7D using any of the anti-CD3 scFv sequences shown in the figures and sequence listing.

[0255] For bottle opener scaffold 1 from Figure 7A (optionally including the 428L / 434S variant), CD binding domain sequences that find particular use in these embodiments include, but are not limited to, CD3 binding domains anti-CD3 H1.30_L1.47, anti-CD3 H1.32_L1.47, anti-CD3 H1.89_L1.47, anti-CD3 H1.90_L1.47, anti-CD3 H1.33_L1.47 and anti-CD3 H1.31_L1.47, and those shown in Figures 10A-10F attached as the scFv side of the scaffolds shown in Figures 7A-7D.

[0256] Particularly useful ENPP3 and CD3 sequence combinations (optionally including the 428L / 434S variant) to use are disclosed in Figures 17A-17C and Figures 18A-18C.

[0257] 2. mAb-Fv One heterodimeric scaffold of particular use in the antibodies described herein is the mAb-Fv format. In this embodiment, the format relies on the use of an "additional" variable heavy chain domain C-terminally attached to one monomer and an "additional" variable light chain domain C-terminally attached to the other monomer, thereby forming a third antigen-binding domain, with the Fab portions of the two monomers binding to ENPP3 and the "additional" scFv domain binding to CD3.

[0258] In this embodiment, the first monomer comprises a first heavy chain comprising a first constant heavy chain domain comprising a first variable heavy chain domain and a first Fc domain, and has a first variable light chain domain covalently linked to the C-terminus of the first Fc domain using a domain linker (VH1-CH1-hinge-CH2-CH3-[optional linker]-VL2). The second monomer comprises a second variable heavy domain of a second constant heavy domain comprising a second Fc domain, and a third variable heavy domain covalently linked to the C-terminus of the second Fc domain using a domain linker (vh1-CH1-hinge-CH2-CH3-[optional linker]-VH2. The two C-terminally linked variable domains constitute an Fv that binds to CD3 (as it is less desirable to have bivalent CD3 binding). This embodiment further utilizes a common light chain comprising a variable light domain and a constant light domain, which associates with the heavy chains to form two identical Fabs that link to ENPP3. For many of the embodiments herein, these constructs include scubariant variants, pI variants, deletion variants, additional Fc variants, etc., as desired and described herein.

[0259] The antibodies described herein are provided in a mAb-Fv format in which the CD3-binding domain sequences are shown in Figures 10A-10F. The antibodies described herein are provided in a mAb-Fv format in which the ENPP3-binding domain sequences are shown in Figures 12, 13A-13B, and 14A-14I.

[0260] Additionally, the Fc domain of the mAb-Fv format may be modified with a scubariant (e.g., a set of amino acid substitutions as shown in Figures 3 and 8, where particularly useful scubariant sequences are S364K / E357Q:L368D / K370S, L368D / K370S:S364K, L368E / K370S:S364K, T411T / E360E / Q362E:D401K, L368D / K370S:S364K / E357Q, L368D / K370S:S364K, T411T / E360E / Q362E:D401K, L368D / K370S:S364K / E357Q, L368D / K370S:S364K, L368E / K370S:S364K, T411T / E360E / Q362E:D401K, L368D / K370S:S364K / E357Q, L368D / K370S:S364K / E357Q, L368D / K370S:S364K, L368E ... 7L, K370S:S364K / E357Q, T366S / L368A / Y407V:T366W and T366S / L368A / Y407V / Y349C:T366W / S354C), optionally a deletion variant (including those shown in Figure 3), optionally a charged scFv linker (including those shown in Figure 5), and the heavy chain comprises a pI variant (including those shown in Figure 2).

[0261] In some embodiments, the mAb-Fv format includes scubariant, pI variants, and deletion variants. Thus, some embodiments include a mAb-Fv format comprising: a) a first monomer comprising scubariant S364K / E357Q, deletion variants E233P / L234V / L235A / G236del / S267K, and a first variable heavy chain domain and a second variable heavy chain domain that together with the first variable light chain domain of the light chain constitute an Fv that binds to ENPP3; and b) a first monomer comprising scubariant L368D / K370S, pI variants N208D / Q295E / N384D / Q4 a second monomer comprising a first variable heavy chain domain which together with the first variable light chain domain constitutes an Fv that binds ENPP3 as outlined herein and a second variable light chain domain which together with the second variable heavy chain domain forms an Fv that binds CD3 (ABD); and c) a light chain comprising the first variable light chain domain and a constant light chain domain.

[0262] In some embodiments, the mAb-Fv format includes scubariant, pI variants, deletion variants, and FcRn variants. Thus, some embodiments are mAb-Fv formats comprising: a) a first monomer comprising scubariant S364K / E357Q, deletion variants E233P / L234V / L235A / G236del / S267K, FcRn variants M428L / N434S, and a first variable heavy chain domain that together with a first variable light chain domain of the light chain constitutes an Fv that binds to an antigen, and a second variable heavy chain domain; and b) a first monomer comprising scubariant L368D / K370S, pI variants N208D / Q295E / N384D / Q418 and c) a light chain comprising the first variable light chain domain and a constant light chain domain.

[0263] 3. mAb-scFv One heterodimeric scaffold particularly useful in the antibodies described herein is the mAb-scFv format. In this embodiment, the format relies on the C-terminal attachment of an scFv to one of the monomers, thereby forming a third antigen-binding domain, with the Fab portions of two monomers binding to ENPP3 and the "additional" scFv domain binding to CD3. Thus, the first monomer comprises a first heavy chain (comprising a variable heavy domain and a constant domain) and has a covalently attached scFv at its C-terminus, comprising an scFv variable light domain, an scFv linker, and an scFv variable heavy domain, in either orientation (VH1-CH1-hinge-CH2-CH3-[optional linker]-VH2-scFv linker-VL2 or VH1-CH1-hinge-CH2-CH3-[optional linker]-VL2-scFv linker-VH2). This embodiment further utilizes a common light chain comprising a variable light domain and a constant light domain, which associates with the heavy chain to form two identical Fabs that link to ENPP3. For many of the embodiments herein, these constructs include scuba variants, pI variants, deletion variants, additional Fc variants, etc., as desired and described herein.

[0264] The antibodies described herein are provided in a mAb-scFv format, with CD binding domain sequences shown in Figures 10A-10F and ENPP3 binding domain sequences shown in Figures 12, 13A-13B, and 14A-14I.

[0265] Additionally, the Fc domain of the central scFv format can be modified with scFv variants (e.g., a set of amino acid substitutions as shown in Figure 1, particularly useful scFv variants include S364K / E357Q:L368D / K370S, L368D / K370S:S364K, L368E / K370S:S364K, T411T / E360E / Q362E:D401K, L368D ... / E357L, K370S:S364K / E357Q, T366S / L368A / Y407V:T366W and T366S / L368A / Y407V / Y349C:T366W / S354C), optionally a deletion variant (including those shown in Figure 3), optionally a charged scFv linker (including those shown in Figure 5), and the heavy chain comprises a pI variant (including those shown in Figure 2).

[0266] In some embodiments, the mAb-scFv format comprises scubariant, pI variants, and deletion variants. Thus, some embodiments are mAb-scFv formats comprising: a) a first monomer comprising scubariant S364K / E357Q, deletion variants E233P / L234V / L235A / G236del / S267K, and a variable heavy chain domain that together with the variable light chain domain of the common light chain constitutes an Fv that binds to ENPP3 as outlined herein, and an scFv domain that binds to CD3; and b) a first monomer comprising scubariant L368D / and c) a common light chain comprising a variable light domain and a constant light domain; and a mAb-scFv format comprising: a second monomer comprising a variable heavy domain comprising K370S, pI variants N208D / Q295E / N384D / Q418E / N421D, deletion variants E233P / L234V / L235A / G236del / S267K, and the variable light domain of the common light chain to constitute an Fv that binds to ENPP3 as outlined herein;

[0267] In some embodiments, the mAb-scFv format comprises scFv variants, pI variants, deletion variants, and FcRn variants. Thus, some embodiments provide a mAb-scFv format comprising: a) a first monomer comprising scFv variants S364K / E357Q, deletion variants E233P / L234V / L235A / G236del / S267K, FcRn variant M428L / N434S, and a variable heavy chain domain that together with the variable light chain domain of the common light chain constitutes an Fv that binds to ENPP3 as outlined herein, and an scFv domain that binds to CD3; and b) a first monomer comprising scFv variants L368D / and c) a common light chain comprising a variable light domain and a constant light domain; and a mAb-scFv format comprising: a second monomer comprising a variable heavy domain comprising K370S, pI variants N208D / Q295E / N384D / Q418E / N421D, deletion variants E233P / L234V / L235A / G236del / S267K, FcRn variant M428L / N434S, and the variable light domain of the common light chain that together constitute an Fv that binds to ENPP3 as outlined herein.

[0268] 4.2+1Fab2-scFv-Fc Format One heterodimeric scaffold that finds particular use with the antibodies described herein is the "2+1 Fab2-scFv-Fc" format (also referred to as the "central-scFv format" in previous related filings) shown in Figure 15B with an exemplary combination of a CD3-binding domain and two tumor-targeting antigen (ENPP3)-binding domains. In this embodiment, the format relies on the use of an inserted scFv domain, thereby forming a third antigen-binding domain, with the Fab portions of two monomers binding to ENPP3 and an "additional" scFv domain binding to CD3. The scFv domain is inserted between the Fc domain and CH1-Fv region of one of the monomers, thereby providing the third antigen-binding domain. As described, ENPP3 x CD3 bispecific antibodies with a 2+1 Fab2-scFv-Fc format are potent at inducing redirected T cell cytotoxicity in cellular environments expressing low levels of ENPP3. Furthermore, as shown in the examples, ENPP3 x CD3 bispecific antibodies with a 2+1 Fab2-scFv-Fc format exhibit a wide variety of distinct properties depending on the ENPP3 and / or CD3 binding domains used, allowing for "fine-tuning" of immune responses. For example, such antibodies exhibit selectivity for cells with different ENPP3 expression, potency against ENPP3-expressing cells, ability to induce cytokine release, and sensitivity to soluble ENPP3. These ENPP3 antibodies are used, for example, to treat ENPP3-associated cancers.

[0269] In this embodiment, one monomer comprises a first heavy chain comprising a first variable heavy domain, a CH1 domain (and optional hinge), and an Fc domain, and an scFv comprising an scFv variable light domain, an scFv linker, and an scFv variable heavy domain. The scFv is covalently linked between the C-terminus of the CH1 domain of the heavy chain constant domain and the N-terminus of the first Fc domain using an optional domain linker (VH1-CH1-[optional linker]-VH2-scFv linker-VL2-[optional linker including hinge]-CH2-CH3, or in the opposite orientation relative to the scFv, VH1-CH1-[optional linker]-VL2-scFv linker-VH2-[optional linker including hinge]-CH2-CH3). The optional linker can be any suitable peptide linker, including, for example, the domain linkers contained in FIG. 6. In some embodiments, the optional linker is a hinge or a fragment thereof. The other monomer is a standard Fab side (e.g., VH1-CH1-hinge-CH2-CH3). This embodiment further utilizes a common light chain comprising a variable light domain and a constant light domain, which associates with the heavy chain to form two identical Fabs that bind to ENPP3. For many of the embodiments herein, these constructs include scuba variants, pI variants, deletion variants, additional Fc variants, etc., as desired and described herein.

[0270] In one embodiment, the 2+1 Fab2-scFv-Fc format antibody comprises an scFv having the VH and VL of the CD3-binding domain sequences shown in Figures 10A-10F or in the Sequence Listing. In one embodiment, the 2+1 Fab2-scFv-Fc format antibody comprises two scFvs having the VH and VL of the ENPP3-binding domain sequences shown in Figures 12, 13A-13B, and 14A-14I or in the Sequence Listing. In exemplary embodiments, the ENPP3 binding domain of the 2+1 Fab-scFv-Fc ENPP3 x CD3 bispecific antibody is selected from the group consisting of the following ENPP3 binding domains: AN1[ENPP3]H1L1, AN1[ENPP3]H1 L1.33, AN1[ENPP3]H1 L1.77, AN1[ENPP3]H1.8 L1, AN1[ENPP3]H1.8 L1.33, AN1[ENPP3]H1 L1.77, H16-7.213, H16-9.69, H16-1.52, Ha16-1(1)23, H16-9.44, H16-1.67, Ha1 The VH and VL sequences contained one of the following: H16-1(3,5)36, H16-1.86, H16-9.10, H16-9.33, H16-1.68, Ha16-1(1)1, Ha1 6-1(3,5)18, Ha16-1(2,4)4, Ha16-1(3,5)56, H16-7.8, H16-1.93, Ha1 6-1(3,5)27.1, H16-1.61, H16-1(3,5)5, ​​H16-7.200, H16-1(3,5)42, H1 6-9.65, Ha1-1(3,5)19, and Ha16-1.80 (Figures 12, 13A-13B, and 14A-14I). In one embodiment, the CD3 binding domain of the 2+1 Fab-scFv-Fc ENPP3 x CD3 bispecific antibody compr...

Claims

1. A pharmaceutical composition for treating an ectonucleotide pyrophosphatase / phosphodiesterase family member 3 (ENPP3)-associated cancer in a patient in need thereof, the pharmaceutical composition comprising an anti-CD3 x anti-ENPP3 antibody, the antibody comprising: a. a first monomer comprising SEQ ID NO:531; b. a second monomer comprising SEQ ID NO:532, and c. comprising a light chain comprising SEQ ID NO: 533; Pharmaceutical compositions.

2. The pharmaceutical composition described in claim 1, further comprising an antibody that is a checkpoint inhibitor.

3. The pharmaceutical composition described in claim 2, wherein the checkpoint inhibitor antibody is an antibody-PD1 antibody.

Images