JP2020534811A - Conditionally activated binding moiety containing the Fc region - Google Patents

Abstract

A composition of a conditionally activated binding protein containing an Fc region and a method for co-expressing and purifying such a conditionally activated binding protein are provided herein. .. Also provided is a method of treating cancer by administering to a patient a conditionally activated binding protein.

Description

Cross-reference to related applications This application claims the interests of US Provisional Patent Application No. 62 / 555,999 filed September 8, 2017, and US Provisional Patent Application No. 62 / 555,943 filed September 8, 2017. However, these disclosures are incorporated herein by reference in their entirety.

Selective destruction of individual cells or specific cell types is often desirable in a variety of clinical environments. For example, the primary goal of cancer therapy is to leave healthy cells and tissues as intact and intact as possible while specifically destroying tumor cells. One such method is by inducing an immune response against the tumor, causing immune effector cells such as natural killer (NK) cells or cytotoxic T lymphocytes (CTL) to attack and destroy the tumor cells.

The use of intact monoclonal antibodies (mAbs), which provide excellent binding specificity and affinity for tumor-related antigens, has been successfully applied in the field of cancer treatment and diagnosis. However, the large size of intact MAbs, their inadequate biodistribution and long persistence in blood pools have limited the clinical use of intact MAbs. For example, intact antibodies may show a particular accumulation within the tumor area. In in vivo distribution studies, heterogeneous antibody distribution with primary accumulation in the surrounding region is prominent when investigating tumors precisely. Due to tumor necrosis, heterogeneous antibody distribution, and increased stromal tissue pressure, it is not possible for intact antibody constructs to reach the central part of the tumor. In contrast, smaller antibody fragments exhibit rapid tumor localization, penetrate deeper into the tumor, and are removed from the bloodstream relatively quickly.

Single chain fragments (scFv) derived from the small binding domain of the parent MAb result in better biodistribution than intact MAbs for clinical use and can target tumor cells more efficiently. Single chain fragments can be efficiently manipulated from bacteria, but most manipulated scFvs have a monovalent structure and with their parent MAb due to the lack of binding force experienced by divalent compounds. It exhibits relatively reduced tumor accumulation, eg, short residence time on tumor cells, and specificity.

Despite the favorable properties of scFv, certain features prevent their full clinical development in cancer chemotherapy. Of particular note is their interreactivity between affected and healthy tissues due to the targeting of these agents to cell surface receptors that are common to both affected and healthy tissues. ScFv with an improved therapeutic index brings significant advances in the clinical utility of these agents. The present invention provides such an improved scFv and a method for producing and using it. The improved scFv of the present invention has the unexpected benefit of overcoming the lack of binding force demonstrated by a single unit by forming dimeric compounds.

In one aspect, (a) from the N-terminus to the C-terminus, (i) a first antigen-binding domain that binds to a first tumor target antigen (TTA), (ii) a domain linker, and (iii) (1). A constrained Fv domain, including a variable heavy domain containing vhCDR1, vhCDR2, and vhCDR3, (2) a constrained non-cleavable linker, and (3) a variable light domain containing vlCDR1, vlCDR2, and vlCDR3, and (iv) first. Cleavable linker of, (v) first Fc domain, first monomer, (b) from N-terminus to C-terminus, (i) (1) pseudovariable heavy domain, (2) non- A second monomer comprising a pseudo-Fv domain, including a cleaving linker and (3) a pseudovariable light domain, (iii) a second cleaving linker, and (iii) a second Fc domain. , The first Fc domain and the second Fc domain contain a knob-in-hole modification, and the constrained Fv domain does not bind to human CD3 in the absence of cleavage in the cleaving linker, heterodimer. Body protein compositions are provided herein. In some embodiments, the first monomer further comprises a second antigen binding domain that binds to a second tumor target antigen (TTA) at the N-terminus of the first cleaving linker.

In certain embodiments, the variable heavy chain comprises the amino acid sequence of SEQ ID NO: 16. In some cases, the variable weight domain comprises vhFR1-vhCDR1-vhFR2-vhCDR2-vhFR3-vhCDR3-vhFR4.

In some embodiments, the variable light domain comprises the amino acid sequence of SEQ ID NO: 17. In certain cases, the variable light domain comprises vrFR1-vrCDR1-vrFR2-vrCDR2-vrFR3-vrCDR3-vrFR4.

In certain embodiments, the pseudo-heavy domain comprises the amino acid sequence of SEQ ID NO: 19. In some embodiments, the pseudo-light domain comprises the amino acid sequence of SEQ ID NO: 18.

In some embodiments, the first TTA is selected from the group consisting of EGFR and EpCAM. In another embodiment, the second TTA is selected from the group consisting of EGFR and EpCAM. In certain cases, the first TTA and the second TTA are the same. In certain cases, the first TTA and the second TTA are different.

In some embodiments, the first antigen binding domain comprises the amino acid sequence of any one of SEQ ID NOs: 14, 15, and 21-25. In other embodiments, the second antigen binding domain comprises the amino acid sequence of any one of SEQ ID NOs: 14, 15, and 21-25.

In some embodiments, the first and / or second cleaving linker contains a cleavage site for MMP9. In some embodiments, the first and / or second cleaving linker contains a cleavage site for mepurin.

In various embodiments, the heterodimer protein comprises the amino acid sequences of Pro217 and Pro218, Pro219 and Pro218, SEQ ID NOs: 9 and 10, or SEQ ID NOs: 10 and 11.

In another embodiment, (a) from the N-terminus to the C-terminus, (i) a first antigen-binding domain that binds to a first tumor target antigen (TTA), (ii) a first domain linker, and (iii). A first pseudo-Fv domain, including (1) a variable light domain comprising vlCDR1, vlcDR2, and vlCDR3, (2) a first cleaving linker, and (3) a pseudovariable heavy domain, and (iv) a th. A first monomer comprising one Fc domain, (b) from N-terminus to C-terminus, (i) a second antigen-binding domain that binds to a second tumor target antigen (TTA), and (ii). A second domain linker, including (iii) (1) a variable heavy domain containing vhCDR1, vhCDR2, and vhCDR3, (2) a second cleaving linker, and (3) a pseudovariable light domain. Contains a second monomer, including a pseudo-Fv domain of, and (iv) a first Fc domain, the first Fc domain and the second Fc domain containing a knob-in-hole modification. A heterodimeric protein composition in which the variable light domain of the first pseudo-Fv domain and the variable heavy domain of the second pseudo-Fv domain do not bind to human CD3 in the absence of cleavage in the cleaving linker is described herein. Provided in the book.

In certain embodiments, the first monomer further comprises a first hinge linker at the N-terminus of the first Fc domain. In some embodiments, the second monomer further comprises a second hinge linker at the N-terminus of the second Fc domain. In various embodiments, the first monomer comprises a first hinge linker at the N-terminus of the first Fc domain and the second monomer has a second hinge linker at the N-terminus of the second Fc domain. Including.

In certain embodiments, the variable heavy chain comprises the amino acid sequence of SEQ ID NO: 16. In some cases, the variable weight domain comprises vhFR1-vhCDR1-vhFR2-vhCDR2-vhFR3-vhCDR3-vhFR4.

In some embodiments, the variable light domain comprises the amino acid sequence of SEQ ID NO: 17. In some cases, the variable light domain comprises vrFR1-vrCDR1-vrFR2-vrCDR2-vrFR3-vrCDR3-vrFR4.

In certain embodiments, the pseudo-heavy domain comprises the amino acid sequence of SEQ ID NO: 19. In some embodiments, the pseudo-light domain comprises the amino acid sequence of SEQ ID NO: 18.

In some embodiments, the first TTA is selected from the group consisting of EGFR and EpCAM. In another embodiment, the second TTA is selected from the group consisting of EGFR and EpCAM. In certain cases, the first TTA and the second TTA are the same. In certain cases, the first TTA and the second TTA are different.

In some embodiments, the first antigen binding domain comprises the amino acid sequence of any one of SEQ ID NOs: 14, 15, and 21-25. In other embodiments, the second antigen binding domain comprises the amino acid sequence of any one of SEQ ID NOs: 14, 15, and 21-25.

In some embodiments, the first and / or second cleaving linker contains a cleavage site for MMP9. In some embodiments, the first and / or second cleaving linker contains a cleavage site for mepurin.

In some embodiments, the heterodimer protein comprises the amino acid sequences of Pro36 and Pro37, Pro36 and Pro38, Pro67 and Pro68, SEQ ID NOs: 1 and 2, SEQ ID NOs: 1 and 3, or SEQ ID NOs: 4 and 5.

In yet another aspect of the invention, (a) from the N-terminus to the C-terminus, (i) a first antigen-binding domain that binds to a first tumor target antigen (TTA), and (ii) a first Fc domain. A first monomer containing, (b) from the N-terminus to the C-terminus, (i) a second antigen-binding domain that binds to a second tumor target antigen (TTA), and (ii) a domain linker. , (Iii) (1) a variable weight domain comprising vhCDR1, vhCDR2, and vhCDR3, (2) a first cleaving linker, and (3) a pseudovariable light domain, and (1) a pseudoFv domain. iv) A second Fc domain, (v) a second cleaving linker, (vi) a third antigen-binding domain that binds to a third tumor target antigen (TTA), and (vi) (1) vrCDR1. A second monomer, including a second pseudo-Fv domain, which comprises, vrCDR2, and a variable light domain comprising vlCDR3, (2) a third cleaving linker, and (3) a pseudovariable heavy domain. The first Fc domain and the second Fc domain contain a knob-in-hole modification, and the variable heavy domain of the first pseudo-Fv domain and the variable light domain of the second pseudo-Fv domain are cleavable. Provided herein are heterodimer protein compositions that do not bind to human CD3 in the absence of cleavage at the linker.

In certain embodiments, the first monomer further comprises a first hinge linker at the N-terminus of the first Fc domain. In some embodiments, the second monomer further comprises a second hinge linker at the N-terminus of the second Fc domain. In various embodiments, the first monomer comprises a first hinge linker at the N-terminus of the first Fc domain and the second monomer has a second hinge linker at the N-terminus of the second Fc domain. Including.

In certain embodiments, the variable heavy chain comprises the amino acid sequence of SEQ ID NO: 16. In some cases, the variable weight domain comprises vhFR1-vhCDR1-vhFR2-vhCDR2-vhFR3-vhCDR3-vhFR4.

In some embodiments, the variable light domain comprises the amino acid sequence of SEQ ID NO: 17. In various examples, the variable light domain comprises vrFR1-vrCDR1-vrFR2-vrCDR2-vrFR3-vrCDR3-vrFR4.

In certain embodiments, the pseudo-heavy domain comprises the amino acid sequence of SEQ ID NO: 19. In some embodiments, the pseudo-light domain comprises the amino acid sequence of SEQ ID NO: 18.

In some embodiments, the first TTA is selected from the group consisting of EGFR and EpCAM. In another embodiment, the second TTA is selected from the group consisting of EGFR and EpCAM. In certain cases, the first TTA and the second TTA are the same. In certain cases, the first TTA and the second TTA are different.

In some embodiments, the first antigen binding domain comprises the amino acid sequence of any one of SEQ ID NOs: 14, 15, and 21-25. In other embodiments, the second antigen binding domain comprises the amino acid sequence of any one of SEQ ID NOs: 14, 15, and 21-25.

In some embodiments, the first TTA, the second TTA, and / or the third TTA is selected from the group consisting of EGFR and EpCAM. In some cases, the first TTA is EGFR. In other cases, the first TTA is EpCAM. In some cases, the second TTA is EGFR. In other cases, the second TTA is EpCAM. In certain cases, the third TTA is EGFR. In other cases, the third TTA is EpCAM.

In certain embodiments, a first TTA and a second TTA, or a first TTA and a third TTA, or a second TTA and a third TTA, or a first TTA, a second TTA, And the third TTA is the same. In certain embodiments, the first TTA and the second TTA, or the first TTA and the third TTA, or the second TTA and the third TTA, or the first TTA, the second TTA, and The third TTA is different.

In various embodiments, the first antigen binding domain comprises the amino acid sequence of any one of SEQ ID NOs: 14, 15, and 21-25. In some embodiments, the second antigen binding domain comprises the amino acid sequence of any one of SEQ ID NOs: 14, 15, and 21-25. In various embodiments, the third antigen binding domain comprises the amino acid sequence of any one of SEQ ID NOs: 14, 15, and 21-25.

In some embodiments, the first, second, and / or third cleavage linkers contain a cleavage site for MMP9. In other embodiments, the first, second, and / or third cleaving linkers contain a cleavage site for mepurin.

In some embodiments, the heterodimer protein comprises the amino acid sequences of Pro69 and Pro70, or sequences 6 and 7.

In another embodiment, (a) from the N-terminus to the C-terminus, (i) a first antigen-binding domain that binds to a first tumor target antigen (TTA), (ii) a first Fc domain, and (iii). ) A first cleaving linker, (iv) a second antigen-binding domain that binds to a second tumor target antigen (TTA), (v) a first domain linker, and (vi) (1) vrCDR1, A first monomer and (b) containing a first sham Fv domain, including a variable light domain comprising vulCDR2 and blCDR3, (2) a second cleaving linker, and (3) a sham variable heavy domain. ) From N-terminus to C-terminus, (i) a second antigen-binding domain that binds to a second tumor target antigen (TTA), (ii) a second domain linker, and (iii) (1) vhCDR1, vhCDR2. A second pseudo-Fv domain and (iv) a second Fc domain, including a variable heavy domain containing vhCDR3, (2) a third cleaving linker, and (3) a pseudo-variable light domain. Including a second monomer, the first Fc domain and the second Fc domain contain a knob-in-hole modification, the variable light domain of the first pseudo-Fv domain and the second pseudo-Fv domain. Provided herein are heterodimer protein compositions in which the variable heavy antigen of the terminus does not bind to human CD3 in the absence of cleavage in the cleaving linker.

In some embodiments, the first monomer further comprises a first hinge linker at the N-terminus of the first Fc domain. In some embodiments, the second monomer further comprises a second hinge linker at the N-terminus of the second Fc domain. In various embodiments, the first monomer comprises a first hinge linker at the N-terminus of the first Fc domain and the second monomer has a second hinge linker at the N-terminus of the second Fc domain. Including.

In certain embodiments, the variable heavy chain comprises the amino acid sequence of SEQ ID NO: 16. In some cases, the variable weight domain comprises vhFR1-vhCDR1-vhFR2-vhCDR2-vhFR3-vhCDR3-vhFR4.

In some embodiments, the variable light domain comprises the amino acid sequence of SEQ ID NO: 17. In various examples, the variable light domain comprises vrFR1-vrCDR1-vrFR2-vrCDR2-vrFR3-vrCDR3-vrFR4.

In certain embodiments, the pseudo-heavy domain comprises the amino acid sequence of SEQ ID NO: 19. In some embodiments, the pseudo-light domain comprises the amino acid sequence of SEQ ID NO: 18.

In some embodiments, the first TTA is selected from the group consisting of EGFR and EpCAM. In another embodiment, the second TTA is selected from the group consisting of EGFR and EpCAM. In certain cases, the first TTA and the second TTA are the same. In certain cases, the first TTA and the second TTA are different.

In some embodiments, the first antigen binding domain comprises the amino acid sequence of any one of SEQ ID NOs: 14, 15, and 21-25. In other embodiments, the second antigen binding domain comprises the amino acid sequence of any one of SEQ ID NOs: 14, 15, and 21-25.

In some embodiments, the first and / or second cleaving linker contains a cleavage site for MMP9. In some embodiments, the first and / or second cleaving linker contains a cleavage site for mepurin.

In some embodiments, the heterodimer protein comprises the amino acid sequences of Pro67 and Pro71, or sequences 4 and 8.

Nucleic acids encoding the first monomer of any one of the heterodimer proteins described herein are provided herein. Nucleic acids encoding a second monomer of any one of the heterodimer proteins described herein are also provided herein. In addition, an expression vector containing a nucleic acid encoding a first monomer, an expression vector containing a nucleic acid encoding a second monomer, or an expression containing a nucleic acid encoding a first monomer and a nucleic acid encoding a second monomer. Vectors are provided herein. In some embodiments, a host cell comprising any one of the expression vectors disclosed herein is provided herein.

One aspect of the invention provides a method of making any one of the heterodimer proteins described herein. The method comprises culturing the host cells described herein under conditions that express the heterodimer protein and recovering the heterodimer protein.

One aspect of the invention provides a method of treating cancer, comprising administering to a patient any one of the heterodimeric proteins of the invention.

The heterodimer protein compositions described herein can be referred to as prodrug compositions in the absence of cleavage by cognate proteases.

In another aspect, a method of treating cancer in a human subject in need thereof, comprising administering any of the prodrug compositions described herein, is provided herein.

This application is filed in International Patent Application No. WO2017 / 156178 filed on March 8, 2017, US Provisional Patent Application No. 62 / 305,092 filed on March 8, 2016, and filed on September 8, 2017. US Provisional Patent Application No. 62 / 555,943, US Provisional Patent Application No. 62 / 555,999 filed September 8, 2017, US Provisional Patent Application No. 62 / 583,327 filed November 15, 2017 No., and US Provisional Patent Application No. 62 / 587,318, filed November 16, 2017, and these disclosures, including drawings, legends, and definitions, and all listed embodiments, are by reference. All of them are incorporated herein.

An exemplary embodiment of a conditionally activated binding polypeptide of the "construct 1" form comprising an Fc hole / knob region is shown. Other cleavage sites as described herein can be used, but outline the Pro37 + Pro36 prodrug construct and the bispecific polypeptide obtained after enterokinase (EK) cleavage. The bispecific polypeptide comprises an sdABD that binds to EGFR and an Fv domain that binds to CD3. Note that, but not limited to, sdABD that binds to other target tumor antigens (TTAs) such as FOLR1, B7H3, and EpCAM may be used in other embodiments. In addition, FIG. 1 shows the use of two different proteins "tags" at the C-terminus of the Fc domain, which facilitated purification of the heterodimer proteins of the invention, as understood by those of skill in the art. These can be removed as such. An exemplary embodiment of a conditionally activated binding polypeptide of the "construct 2" form comprising an Fc hole / knob region is shown. Although many embodiments utilize other cleavage sites, the Flag cleavage site of EK is used again to outline the Pro38 + Pro36 prodrug construct. Note that, but not limited to, sdABD that binds to other target tumor antigens (TTAs) such as FOLR1, B7H3, and EpCAM may be used in other embodiments. In addition, FIG. 2 shows the use of two different proteins "tags" at the C-terminus of the Fc domain, which facilitated purification of the heterodimer proteins of the invention, as understood by those of skill in the art. These can be removed as such. It shows that some of the exemplary heterodimer Fc prodrug constructs described herein exhibited low conditioning or lack of conditioning when cleaved by a cognate protease in the TDCC assay. In FIG. 3A, Pro36 + 37 (circles) were not pretreated with EK protease, but cleaved Pro36 + 37 (squares) were pretreated. In FIG. 3B, Pro36 + 38 (circles) were not pretreated with EK protease, but cleaved Pro36 + 38 (squares) were pretreated. Pro214 is a full-length negative control (white square) and Pro51 (triangle) is a positive control that does not require protease cleavage for activity. An exemplary embodiment of a conditionally activated binding polypeptide of the "construct 3" form comprising an Fc hole / knob region is shown. Although many embodiments utilize other cleavage sites, the Flag cleavage site of EK is used again to outline the Pro68 + Pro67 prodrug construct and the bispecific polypeptide obtained after enterokinase (EK) cleavage. Is done. The bispecific polypeptide comprises an sdABD that binds to EGFR and an Fv domain that binds to CD3. Note that, but not limited to, sdABD that binds to other target tumor antigens (TTAs) such as FOLR1, B7H3, and EpCAM may be used in other embodiments. In addition, FIG. 2 shows the use of two different proteins "tags" at the C-terminus of the Fc domain, which facilitated purification of the heterodimer proteins of the invention, as understood by those of skill in the art. These can be removed as such. An exemplary embodiment of a conditionally activated binding polypeptide of the "construct 4" form comprising an Fc hole / knob region is shown. Although many embodiments utilize other cleavage sites, the Flag cleavage site of EK is used again to outline the Pro69 + Pro70 prodrug construct. Note that, but not limited to, sdABD that binds to other target tumor antigens (TTAs) such as FOLR1, H7B3, and EpCAM may be used in other embodiments. In addition, FIG. 5 shows the use of two different proteins "tags" at the C-terminus of the monomeric protein, which facilitated purification of the heterodimer proteins of the invention, as understood by those of skill in the art. These can be removed as such. An exemplary embodiment of a conditionally activated binding polypeptide of the "construct 5" form comprising an Fc hole / knob region is shown. Although many embodiments utilize other cleavage sites, the Flag cleavage site of EK is used again to outline the Pro71 + Pro67 prodrug construct. Note that, but not limited to, sdABD that binds to other target tumor antigens (TTAs) such as FOLR1, H7B3, and EpCAM may be used in other embodiments. In addition, FIG. 2 shows the use of two different proteins "tags" at the C-terminus of the monomeric protein, which facilitated purification of the heterodimer proteins of the invention, as understood by those of skill in the art. These can be removed as such. Some exemplary heterodimer Fc prodrug constructs described herein showed conditioning when cleaved by a cognate protease in the TDCC assay, but showed a lack of high activity. .. In FIG. 7A, Pro67 + 68 (circles) were not pretreated with EK protease, but cleaved Pro67 + 68 (squares) were pretreated. Some exemplary heterodimer Fc prodrug constructs described herein showed conditioning when cleaved by a cognate protease in the TDCC assay, but showed a lack of high activity. .. In FIG. 7B, Pro69 + 70 (circles) were not pretreated with EK protease, but cleaved Pro69 + 70 (squares) were pretreated. Some exemplary heterodimer Fc prodrug constructs described herein showed conditioning when cleaved by a cognate protease in the TDCC assay, but showed a lack of high activity. .. In FIG. 7C, Pro67 + 71 (circle) was not pretreated with EK protease, but cleaved Pro67 + 71 (square) was pretreated. Pro214 is a full-length negative control (white square) and Pro51 (triangle) is a positive control that does not require protease cleavage for activity. An exemplary embodiment of a conditionally activated binding polypeptide of the "construct 6" form comprising an Fc hole / knob region is shown. Other cleavage sites described herein can be used as well, but the MMP9 protease cleavage site is used to outline the Pro219 + Pro218 prodrug construct. Note that, but not limited to, sdABD that binds to other target tumor antigens (TTAs) such as FOLR1, H7B3, and EpCAM may be used in other embodiments. In addition, FIG. 8 shows the use of two different proteins "tags" at the C-terminus of the monomeric protein, which facilitated purification of the heterodimer proteins of the invention, as understood by those of skill in the art. These can be removed as such. An exemplary embodiment of a conditionally activated binding polypeptide of the "construct 7" form comprising an Fc hole / knob region is shown. Other cleavage sites described herein can be used as well, but the MMP9 protease cleavage site is used to outline the Pro217 + Pro218 prodrug construct. Note that, but not limited to, sdABD that binds to other target tumor antigens (TTAs), such as EpCAM, may be used in other embodiments. In addition, FIG. 9 shows the use of two different proteins "tags" at the C-terminus of the monomeric protein, which facilitated purification of the heterodimer proteins of the invention, as understood by those of skill in the art. These can be removed as such. It is shown that the exemplary heterodimer Fc prodrug constructs described herein exhibited conditioning and high potency when cleaved by a cognate protease in the TDCC assay. In FIG. 10A, Pro217 + 218 (circle) was not pretreated with EK protease, but cleaved Pro217 + 218 (square) was pretreated. Pro214 is a full-length negative control (white square) and Pro51 (triangle) is a positive control that does not require protease cleavage for activity. In FIG. 10B, Pro218 + 219 (circles) were not pretreated with EK protease, but cleaved Pro218 + 219 (squares) were pretreated. Pro214 is a full-length negative control (white square) and Pro51 (triangle) is a positive control that does not require protease cleavage for activity. It shows some suitable protease cleavage sites. As will be appreciated by those skilled in the art, these cleavage sites can be used as cleavage linkers. In some embodiments, for example, if a more flexible cleaving linker is required, there are additional amino acids (generally glycine and serine) at either or both of the N-terminus and C-terminus of these cleavage sites. Can be done. It shows some suitable protease cleavage sites. As will be appreciated by those skilled in the art, these cleavage sites can be used as cleavage linkers. In some embodiments, for example, if a more flexible cleaving linker is required, there are additional amino acids (generally glycine and serine) at either or both of the N-terminus and C-terminus of these cleavage sites. Can be done. It shows some suitable protease cleavage sites. As will be appreciated by those skilled in the art, these cleavage sites can be used as cleavage linkers. In some embodiments, for example, if a more flexible cleaving linker is required, there are additional amino acids (generally glycine and serine) at either or both of the N-terminus and C-terminus of these cleavage sites. Can be done. Many suitable scFv linkers or domain linkers are shown. Many sequences of the present invention are shown. For antigen-binding domains, the CDRs are bold and underlined. Many sequences of the present invention are shown. For antigen-binding domains, the CDRs are bold and underlined. Many sequences of the present invention are shown. For antigen-binding domains, the CDRs are bold and underlined. Many sequences of the present invention are shown. For antigen-binding domains, the CDRs are bold and underlined. Many sequences of the present invention are shown. For antigen-binding domains, the CDRs are bold and underlined. Many sequences of the present invention are shown. For antigen-binding domains, the CDRs are bold and underlined. Many sequences of the present invention are shown. For antigen-binding domains, the CDRs are bold and underlined. The sequence of the present invention is shown. Linkers are underlined and cleavable linkers are double underlined. The CDRs are bold and underlined. The diagonal line (“/”) indicates the domain separator. C-terminal tags such as maltose-binding protein (MBP), (His) 10, and Strep-Tag® II tags are in bold, depending on the purification scheme used, as outlined herein. Is optional. Therefore, the sequence of FIG. 14 excluding the C-terminal tag is included within the description herein. The sequence of the present invention is shown. Linkers are underlined and cleavable linkers are double underlined. The CDRs are bold and underlined. The diagonal line (“/”) indicates the domain separator. C-terminal tags such as maltose-binding protein (MBP), (His) 10, and Strep-Tag® II tags are in bold, depending on the purification scheme used, as outlined herein. Is optional. Therefore, the sequence of FIG. 14 excluding the C-terminal tag is included within the description herein. The sequence of the present invention is shown. Linkers are underlined and cleavable linkers are double underlined. The CDRs are bold and underlined. The diagonal line (“/”) indicates the domain separator. C-terminal tags such as maltose-binding protein (MBP), (His) 10, and Strep-Tag® II tags are in bold, depending on the purification scheme used, as outlined herein. Is optional. Therefore, the sequence of FIG. 14 excluding the C-terminal tag is included within the description herein. The sequence of the present invention is shown. Linkers are underlined and cleavable linkers are double underlined. The CDRs are bold and underlined. The diagonal line (“/”) indicates the domain separator. C-terminal tags such as maltose-binding protein (MBP), (His) 10, and Strep-Tag® II tags are in bold, depending on the purification scheme used, as outlined herein. Is optional. Therefore, the sequence of FIG. 14 excluding the C-terminal tag is included within the description herein. The sequence of the present invention is shown. Linkers are underlined and cleavable linkers are double underlined. The CDRs are bold and underlined. The diagonal line (“/”) indicates the domain separator. C-terminal tags such as maltose-binding protein (MBP), (His) 10, and Strep-Tag® II tags are in bold, depending on the purification scheme used, as outlined herein. Is optional. Therefore, the sequence of FIG. 14 excluding the C-terminal tag is included within the description herein. The sequence of the present invention is shown. Linkers are underlined and cleavable linkers are double underlined. The CDRs are bold and underlined. The diagonal line (“/”) indicates the domain separator. C-terminal tags such as maltose-binding protein (MBP), (His) 10, and Strep-Tag® II tags are in bold, depending on the purification scheme used, as outlined herein. Is optional. Therefore, the sequence of FIG. 14 excluding the C-terminal tag is included within the description herein. The sequence of the present invention is shown. Linkers are underlined and cleavable linkers are double underlined. The CDRs are bold and underlined. The diagonal line (“/”) indicates the domain separator. C-terminal tags such as maltose-binding protein (MBP), (His) 10, and Strep-Tag® II tags are in bold, depending on the purification scheme used, as outlined herein. Is optional. Therefore, the sequence of FIG. 14 excluding the C-terminal tag is included within the description herein. An additional Pro219 construct (FIG. 15A) and an additional Pro217 construct (FIGS. 15B and 15C) are shown. An additional Pro219 construct (FIG. 15A) and an additional Pro217 construct (FIGS. 15B and 15C) are shown. An additional Pro219 construct (FIG. 15A) and an additional Pro217 construct (FIGS. 15B and 15C) are shown. An additional sequence of the present invention is shown. For antigen-binding domains, the CDRs are bold and underlined. The "/" indicates the intersection of domains, the domain linker is underlined and the cleaving linker is underlined and italicized. Many of the constructs contain histidine tags, but are optional depending on their use. An additional sequence of the present invention is shown. For antigen-binding domains, the CDRs are bold and underlined. The "/" indicates the intersection of domains, the domain linker is underlined and the cleaving linker is underlined and italicized. Many of the constructs contain histidine tags, but are optional depending on their use. An additional sequence of the present invention is shown. For antigen-binding domains, the CDRs are bold and underlined. The "/" indicates the intersection of domains, the domain linker is underlined and the cleaving linker is underlined and italicized. Many of the constructs contain histidine tags, but are optional depending on their use. An additional sequence of the present invention is shown. For antigen-binding domains, the CDRs are bold and underlined. The "/" indicates the intersection of domains, the domain linker is underlined and the cleaving linker is underlined and italicized. Many of the constructs contain histidine tags, but are optional depending on their use. An additional sequence of the present invention is shown. For antigen-binding domains, the CDRs are bold and underlined. The "/" indicates the intersection of domains, the domain linker is underlined and the cleaving linker is underlined and italicized. Many of the constructs contain histidine tags, but are optional depending on their use. An additional sequence of the present invention is shown. For antigen-binding domains, the CDRs are bold and underlined. The "/" indicates the intersection of domains, the domain linker is underlined and the cleaving linker is underlined and italicized. Many of the constructs contain histidine tags, but are optional depending on their use. An additional sequence of the present invention is shown. For antigen-binding domains, the CDRs are bold and underlined. The "/" indicates the intersection of domains, the domain linker is underlined and the cleaving linker is underlined and italicized. Many of the constructs contain histidine tags, but are optional depending on their use.

I. Introduction The present invention relates to methods of reducing the toxicity and side effects of bispecific antibodies (including antibody-like functional proteins) that bind to important physiological targets such as CD3 and tumor antigens. Many antigen-binding proteins, such as antibodies, can have significant "on-target / off-tumor" side effects, thus activating only the binding capacity of therapeutic molecules near the affected tissue and off-target interactions. Need to be avoided. Accordingly, the present invention relates to polyvalent, conditionally effective (“MCE”) proteins having many functional protein domains. In general, one of these domains is an antigen-binding domain (ABD) that binds to a target tumor antigen (TTA). Another domain is an ABD that binds to a T cell antigen, such as CD3, to form an anti-CD3 Fv binding domain under certain conditions, for example, when the portion of ABD is near the complementary portion of ABD. .. That is, the therapeutic molecule is made in a "prodrug" -like form in which the CD3 binding domain is inactive until exposed to the tumor environment. To achieve this conditioning, the invention is "false" or "inactive" or "inactive" in several different ways, depending on its form, as described herein and shown in the drawings. Use a variable domain. These are referred to herein as "iVH" and "iVL" domains.

In some embodiments of the invention, the CD3 binding domain (“CD3 Fv”) is a constrained form, and the linker between the variable heavy domain and the variable light domain traditionally forming Fv is two domains. Too short to allow them to bind to each other. In some embodiments, in prodrug (eg, uncleaved) form, the prodrug polypeptide also comprises a "pseudo-Fv domain". Pseudo-Fv domains are standard framework regions, but variable weight domains with "inactive" or "dummy" CDRs (inactive variable weight domains), standard framework regions, but "inactive". , Or a variable light domain with a "dummy" CDR (inactive variable light domain), or both. Therefore, the constrained Fv domain binds to the pseudo Fv domain by the affinity of each framework region. However, due to the "inactive" CDRs of the pseudodomain, the resulting ABD does not bind to CD3 and thus prevents off-target toxicity. However, in the presence of proteases in or near the tumor, the prodrug construct is cleaved in such a way that the "true" variable heavy domain and variable light domain can associate, thus activating CD3 binding and resulting tumor efficacy. Causes sex.

In other embodiments, in the prodrug form, the prodrug polypeptide comprises two pseudoFv domains and an Fc domain linked to each pseudoFv domain. The first pseudo-Fv domain may include an inactive variable heavy domain and an active variable light domain, and the second pseudo Fv domain may include an active variable heavy domain and an inactive variable light domain. The prodrug form of ABD does not bind to CD3 in proteolytically inactive tissues. However, within or near the tumor, the protease can cleave the prodrug construct, whereby the active variable heavy domain and the active variable light domain can associate and bind to CD3, thereby allowing the target tumor cells to cleave. Induces cytotoxicity.

Thus, the prodrug constructs provided herein include a heterodimer IgG Fc region that forms a "knob-in-hole" ("KIH") conformation. A detailed description of the knob-in-hole concept is described, for example, in US Pat. Nos. 5,731,168 and 7,186,076, as well as Ridgway et al. , Protein Engineering, Design and Selection, 1996, 9 (7): 617-621, Atwell et al. , J Mol Biol, 1997, 270 (1): 26-35, Merchant et al. , Nat Biotechnology, 1998, 16: 677-681, and Carter, J. et al. It can be found in Immunological Methods, 2001, 24 (1-2): 7-15. Briefly, knobs can be created at the CH3 domain interface of the first IgG Fc chain by replacing the smaller amino acid side chains with larger ones (eg, T366W), and holes are the larger amino acid side chains. Can be made in parallel positions at the CH3 interface of the second IgG Fc chain by replacing with a smaller one (eg, Y407V). Suitable KIH variants are described below. Therefore, the CH3 domain having "knob substitution (s)" is referred to as "CH3-knob" in the present specification, and the CH3 domain having "hole substitution (s)" is referred to as "CH3-" in the present specification. It is uncertain which "side" of the Fc dimer contains the "Hole variant" and which contains the "Knob variant", referred to as "Hole" and as will be understood by those skilled in the art. Since it can change, its general name is "CH3-KIH", which covers both.

As discussed herein, there are various conformations and forms used in the present invention. The conformation of the prodrug construct can have a wide variety of configurations such that prodrug activation can occur in some common ways, as "construction 1" in FIG. 1 and as "construction 2" in FIG. , In FIG. 4 as "Structure 3", in FIG. 5 as "Structure 4", in FIG. 6 as "Structure 5", in FIG. 8 as "Structure 6", and in FIG. 9 as "Structure 7". These constructs generally rely on the Fc domain to form the heterodimeric Fc structure, allowing proper pre-cleavage association of the inactive binding domain.

In the embodiment of "Structure 1", the prodrug construct comprises a CH2-CH3-hole polypeptide, a first sham Fv domain, and an antigen binding domain (ABD) capable of binding to a target tumor antigen (TTA). A second Fc polypeptide comprising one Fc polypeptide, a CH2-CH3-nobu polypeptide, a second pseudoFv domain, and an antigen binding domain (ABD) capable of binding to a target tumor antigen (TTA). including. Pseudo-Fv domains refer to CD3-binding domains (Fv) that are inactive until exposed to the tumor environment. In this embodiment, the pseudo-Fv domain comprises an active variable heavy domain (active VH) and an inactive variable light domain (inactive VL). In other embodiments, the pseudoFv domain comprises an active variable light domain (active VL) and an inactive variable heavy domain (inactive VH). Moreover, as will be appreciated by those skilled in the art, the pseudo-Fv domain of "Structure 1" is either VH-linker-VL or VL-linker-VH in either orientation, from N-terminus to C-terminus. (For example, in FIG. 1 / "Structure 1", the pseudo-Fv domain is shown as VH-linker-VL, which can be switched).

In some embodiments of "Structure 1", the first Fc polypeptide (from N-terminus to C-terminus) is not linked to a second antigen binding domain linked to the CH2-CH3-KIH polypeptide. The second Fc polypeptide comprises an antigen-binding domain of the first TTA linked via a domain linker to the active VL domain attached to the active VH domain via a cleaving linker, and the second Fc polypeptide (from N-terminus to C-terminus). , Includes an antigen-binding domain of a second TTA linked via a domain linker to an active VH domain attached via a cleaving linker to an inactive VL domain linked to a CH2-CH3-KIH polypeptide. In some cases, the first Fc polypeptide contains CH3-holes and the second Fc polypeptide contains CH3-knobs. Cleavage of the cleaving linker of the first Fc polypeptide at or near the tumor site and cleavage of the cleaving linker of the second Fc polypeptide results in the active VL of the first Fc polypeptide and the second Fc polypeptide. Active VHs can associate and cause active CD3 binding. In addition to the unique self-assembly of active VH and VL domains, each domain is linked to an antigen-binding domain for tumor antigens. As such, the protease cleavage product can bind to the tumor cells and recruit T cells to the tumor site. In some embodiments, the first TTA and the second TTA are the same tumor antigens. In other embodiments, the first TTA and the second TTA are different tumor antigens.

In some cases, the prodrug construct of "Structure 1" has two cleavage sites, one between the active variable light chain and the inactive variable heavy chain of the first Fc polypeptide, two. The eye lies between the active variable heavy chain and the inactive variable light chain of the second Fc polypeptide. In some embodiments, two cleavage sites are recognized and cleaved by the same protease. As such, the two cleavage sites can have the same or substantially the same amino acid sequence. In other embodiments, two cleavage sites are recognized and cleaved by different proteases. Therefore, the two cleavage sites may have different amino acid sequences.

In some embodiments of "Structure 2", the prodrug construct is a first sham containing a CH2-CH3-KIH polypeptide, an active variable light chain (active VL) and an inactive variable heavy chain (inactive VH). A first Fc polypeptide comprising an Fv domain and an antigen binding domain (ABD) capable of binding to a target tumor antigen (TTA), and a CH2-CH3-KIH polypeptide, active variable heavy chain (active VH) and inactive. It comprises a second pseudoFv domain containing a variable light chain (inactive VL) and a second Fc polypeptide containing an antigen binding domain (ABD) capable of binding to a target tumor antigen (TTA). In some cases, the first Fc polypeptide contains CH3-holes and the second Fc polypeptide contains CH3-knobs.

In some embodiments of "Structure 2", the first Fc polypeptide (from N-terminus to C-terminus) is cleaved into an inactive VH domain linked to the CH2-CH3-KIH polypeptide via a domain linker. The second Fc polypeptide (from N-terminus to C-terminus) contains the antigen-binding domain of the first TTA linked via a domain linker to the active VL domain attached via a sex linker, CH2-CH3-. It contains an antigen-binding domain of a second TTA linked via a domain linker to an active VH domain attached to an inactive VL domain linked to a KIH polypeptide via a domain linker via a cleaving linker. In some cases, the first Fc polypeptide contains CH3-holes and the second Fc polypeptide contains CH3-knobs. Cleavage of the cleaving linker of the first Fc polypeptide at or near the tumor site and cleavage of the cleaving linker of the second Fc polypeptide results in the active VL of the first Fc polypeptide and the second Fc polypeptide. Active VHs can associate and cause active CD3 binding. In addition to the unique self-assembly of active VH and VL domains, each domain is linked to an antigen-binding domain for tumor antigens. As such, the protease cleavage product can bind to the tumor cells and recruit T cells to the tumor site. In some embodiments, the first TTA and the second TTA are the same tumor antigens. In other embodiments, the first TTA and the second TTA are different tumor antigens.

The prodrug of "construct 3" is similar to "construct 2", but is a domain linker between the inactive VH domain of the first Fc polypeptide and the CH3-hole polypeptide, and a second Fc polypeptide. There is a lack of domain linker between the inactive VL domain of CH3-nobu polypeptide.

A first Fc polypeptide comprising a CH2-CH3-KIH polypeptide and an antigen-binding domain (ABD) capable of binding to a target tumor antigen (TTA), and a CH2-CH3-KIH polypeptide, first pseudoFv domain, first. With a second Fc polypeptide comprising two pseudoFv domains, a second antigen binding domain capable of binding to the target tumor antigen (TTA), and a third antigen binding domain capable of binding to the target tumor antigen (TTA). Also provided herein are prodrug constructs (eg, "construct 4") that include. In some cases, the first Fc polypeptide contains CH3-holes and the second Fc polypeptide contains CH3-knobs. In some embodiments, the first, second, and / or third antigen-binding domains may bind to the same tumor antigen. In other embodiments, the first, second, and / or third antigen-binding domains are different tumor antigens. The first and second antigen-binding domains can bind to the same tumor antigen. The first and second antigen-binding domains can bind to different tumor antigens. The first and third antigen-binding domains can bind to the same tumor antigen. The first and third antigen-binding domains can bind to different tumor antigens. The second and third antigen-binding domains can bind to the same tumor antigen. The second and third antigen-binding domains can bind to different tumor antigens.

In the embodiment of "Structure 4", the first Fc polypeptide (from N-terminus to C-terminus) comprises the first antigen binding domain of TTA linked to the CH2-CH3-KIH polypeptide and the second. The Fc polypeptide (from N-terminus to C-terminus) is cleaved into the third antigen-binding domain of the TTA linked via the domain linker to the active VL domain linked to the inactive VH domain via the cleaving linker. Second antigen binding of TTA ligated via a domain linker to an active VH domain attached via a cleaving linker to an inactive VL domain linked to a CH2-CH3-KIH polypeptide linked via a linker. Includes domain. In some cases, the first Fc polypeptide contains CH3-holes and the second Fc polypeptide contains CH3-knobs.

In the embodiment of "Construct 5", the first Fc polypeptide (from N-terminus to C-terminus) is linked to the inactive VH domain via a cleaving linker to the active VL domain via a domain linker. The second Fc polypeptide comprises the first antigen-binding domain of the TTA linked to the CH2-CH3-KIH polypeptide linked to the second antigen-binding domain via a cleaving linker (from the N-terminus). A third antigen-binding domain of TTA linked via a domain linker to an active VH domain linked to an inactive VL domain linked to a CH2-CH3-KIH polypeptide via a cleaving linker (at the C-terminus). including. In some cases, the first Fc polypeptide contains CH3-holes and the second Fc polypeptide contains CH3-knobs. In some embodiments, the first, second, and / or third antigen-binding domains may bind to the same tumor antigen. In other embodiments, the first, second, and / or third antigen-binding domains are different tumor antigens. The first and second antigen-binding domains can bind to the same tumor antigen. The first and second antigen-binding domains can bind to different tumor antigens. The first and third antigen-binding domains can bind to the same tumor antigen. The first and third antigen-binding domains can bind to different tumor antigens. The second and third antigen-binding domains can bind to the same tumor antigen. The second and third antigen-binding domains can bind to different tumor antigens.

A first including a CH2-CH3-KIH polypeptide and a first pseudo-Fc polypeptide, which comprises a standard framework region and a variable heavy domain and a variable light domain having an "inactive" or "dummy" CDR. A second Fc polypeptide comprising a Fc peptide, a CH2-CH3-KIH polypeptide, an antigen binding domain (ABD) capable of binding to a target tumor antigen (TTA), and a constrained form of the CD3 binding domain. A prodrug construct comprising (eg, "construct 6") is provided herein so that a linker between a variable heavy domain and a variable light domain traditionally forming an Fv allows the two domains to bind to each other. Too short to In some embodiments, the constrained active Fv domain is covalently attached to the CH2-CH3-KIH polypeptide via a cleaving linker and the first pseudoFv domain is covalently linked to CH2-CH3 via a cleaving linker. -Covalently attached to the KIH polypeptide. In some cases, the first Fc polypeptide contains CH3-holes and the second Fc polypeptide contains CH3-knobs. In some embodiments, the cleaving linker can be recognized by the same protease. In other embodiments, the cleaving linker can be recognized by different proteases.

In the embodiment of "Structure 6", the second Fc polypeptide (from N-terminus to C-terminus) is a constrained active Fv domain (eg,) linked to the CH2-CH3-KIH polypeptide via a cleaving linker. , An active variable heavy chain linked to an active variable light chain via a constrained non-cleavable linker, or an active variable light chain linked to an active variable heavy chain via a constrained non-cleavable linker) It contains the first antigen-binding domain of TTA linked via a linker, and the first Fc polypeptide (from N-terminus to C-terminus) is a cleaved or non-cleavable linker to the CH2-CH3-KIH polypeptide. Pseudo-Fv domain linked via (eg, an inactive variable light domain linked to an inactive variable heavy domain via a non-cleavable linker, or an inactive variable light domain linked via a non-cleavable linker. Inactive variable heavy domain) is included. In some cases, the first Fc polypeptide contains CH3-holes and the second Fc polypeptide contains CH3-knobs.

Another prodrug construct similar to "Structure 6" (eg, "Structure 7") is provided herein. An exemplary embodiment of construct 7 is a CH2-CH3-KIH polypeptide, a first antigen binding domain (ABD) capable of binding to a target tumor antigen (TTA), a second capable of binding to a target tumor antigen (TTA). The linker between the second Fc polypeptide and the variable heavy domain and the variable light domain (conventionally forming Fv), including the antigen binding domain (ABD) of, and the constrained form of the CD3 binding domain, is two domains. Too short to allow them to bind to each other), including a CH2-CH3-KIH polypeptide and a first Fc polypeptide, including a first pseudoFv domain. In some cases, the first Fc polypeptide contains CH3-holes and the second Fc polypeptide contains CH3-knobs. In some cases, the first and second antigen-binding domains may bind to the same tumor antigen. In other cases, the first and second antigen-binding domains can bind to different tumor antigens.

In the embodiment of "Structure 7", the second Fc polypeptide (from N-terminus to C-terminus) is domained to a second antigen binding domain linked to the CH2-CH3-KIH polypeptide via a cleaving linker. A constrained active Fv domain linked via a linker (eg, via an active variable heavy chain linked via a constrained non-cleavable linker, or via a non-cleavable linker bound to an active variable heavy chain. The first Fc polypeptide (from N-terminus to C-terminus) comprises the first antigen-binding domain of TTA linked via a domain linker to the linked active variable light chain), CH2-CH3-KIH poly. A pseudo-Fv domain linked to a peptide via a cleaved or non-cleavable linker (eg, an inactive variable light domain or an inactive variable light domain linked to an inactive variable heavy domain via a non-cleavable linker). Inactive variable heavy domain linked via a non-cleavable linker). In some cases, the first Fc polypeptide contains CH3-holes and the second Fc polypeptide contains CH3-knobs. In some embodiments, the cleaving linker adjacent to the CH2-CH3-nobu polypeptide is the same cleaving linker adjacent to the CH2-CH3-hole polypeptide. In other embodiments, the cleaving linker is different.

II. Definitions To help this application be more fully understood, some definitions are given below. Such definitions are intended to include grammatical equivalents.

The terms "COBRA ™" and "conditional bispecific redirection activation" refer to bispecific, conditionally effective proteins with many functional protein domains. In some embodiments, one of the functional domains is an antigen binding domain (ABD) that binds to a target tumor antigen (TTA). In one particular embodiment, another domain is an ABD that binds to a T cell antigen under certain conditions. Examples of the T cell antigen include, but are not limited to, CD3. The term "semi-COBRA ™" refers to another semi-COBRA ™ (complementary semi-COBRA ™) in which the variable heavy chain of a half-COBRA is separated by its own self-assembly when concentrated on the surface of target-expressing cells. ) Refers to a conditionally effective protein capable of binding to a T cell antigen if it can be associated with a variable light chain.

As used herein, "amino acid" and "amino acid identity" are one of 20 naturally occurring amino acids or any unnatural analogs that may be present at a particular defined position. Means one. In many embodiments, "amino acid" means one of 20 naturally occurring amino acids. As used herein, "protein" means at least two covalently linked amino acids, including proteins, polypeptides, oligopeptides, and peptides.

As used herein, "amino acid modification" means an amino acid substitution, insertion, and / or deletion in a polypeptide sequence, or modification to a portion chemically linked to a protein. For example, the modification can be a modified carbohydrate or PEG structure attached to the protein. For clarity, amino acid modifications are always for amino acids encoded by DNA, such as 20 amino acids with codons in DNA and RNA, unless otherwise noted. A preferred amino acid modification herein is a substitution.

As used herein, "amino acid substitution" or "substitution" means the replacement of an amino acid at a particular position in the parent polypeptide sequence with a different amino acid sequence. Specifically, in some embodiments, the substitution is for an amino acid that is not naturally occurring at a particular position, either in that organism or in any organism. For clarity, alter the nucleic acid coding sequence but not the starting amino acid (eg, replace CGG (encoding arginine) with CGA (still encoding arginine) to increase host organism expression levels. ) The engineered protein is not an "amino acid substitution", that is, if the protein has the same amino acid at the particular position from which it originated, despite the generation of a new gene encoding the same protein. It is not an amino acid substitution.

As used herein, "amino acid insertion" or "insertion" means the addition of an amino acid sequence at a particular position in the parent polypeptide sequence.

As used herein, "amino acid deletion" or "deletion" means the removal of an amino acid sequence at a particular position in the parent polypeptide sequence.

The polypeptides of the invention specifically bind to CD3 and target target tumor antigens (TTAs), such as target cell receptors, as outlined herein. "Specific binding" to a particular antigen or epitope or "specifically binding" to it or "specific to" it means a binding that is measurable different from a non-specific interaction. Specific binding can be measured, for example, by determining the binding of the molecule as compared to the binding of a control molecule, which is a molecule of similar structure that generally does not have binding activity. For example, specific binding can be determined by competition with a control molecule that is similar to the target.

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. KD can be indicated by an antibody having KD against an antigen or epitope of about ^10-9 M, or at least about ^10-10 M, at least about ^10-11 M, at least about ^10-12 M, or higher, where the KD is specific. Refers to the dissociation rate of antigen-antibody interaction. Typically, an antibody that specifically binds to an antigen is 20, 50, 100, 500, 1000, 5,000, 10,000 times, or more relative to the antigen or epitope. Has a KD that is.

Also, specific binding to a particular antigen or epitope is, for example, at least 20, 50, 100, 500, 1000, 5,000, 10,000 times, or more relative to the epitope as compared to a control. Can be indicated by an antibody having KA or Ka of an antigen or epitope, KA or Ka refers to the rate of association of a particular antibody-antigen interaction. Binding affinity is generally measured using the Biacore assay or Octet as is known in the art.

As used herein, "parent polypeptide" or "precursor polypeptide" (including Fc parent or precursor) means a polypeptide that is later modified to produce a variant. The parent polypeptide may be a naturally occurring polypeptide, or a variant or engineered form of a naturally occurring polypeptide. The parent polypeptide can refer to the polypeptide itself, a composition comprising the parent polypeptide, or an amino acid sequence encoding the same. Thus, as used herein, "parent Fc polypeptide" means an unmodified Fc polypeptide that is modified to produce a variant, generally the human IgG Fc domain as defined herein. As used herein, "parent antibody" means an unmodified antibody that is modified to produce a variant antibody.

As used herein, "position" means a location in a protein sequence. Positions can be numbered sequentially or according to an established format, eg, EU index for antibody numbering.

As used herein, "target antigen" means a molecule that is specifically bound by the variable region of a given antibody. The target antigen may be a protein, carbohydrate, lipid, or other compound. A wide range of suitable exemplary target antigens are described herein.

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

As used herein, "Fv" or "Fv domain" or "Fv region" generally means a polypeptide comprising the VL and VH domains of the antigen binding domain from an antibody. Fv domains usually form an "antigen binding domain" or "ABD" as discussed herein when they contain active VH and VL domains (as may be discussed below). Fv, Fv domains containing constraint linkers can be organized in many ways in the present invention, although they can be "active" or "inactive" in, for example, scFv form, constrained Fv form, pseudoFv form, etc.) .. In the present invention, in some cases, the Fv domain is composed of VH and VL domains on a single polypeptide chain, as shown in FIGS. 8 and 9, but cannot form an intramolecular ABD. It should be understood that it has a constrained linker. In these embodiments, the two active ABDs are formed after cleavage. In some cases, the Fv domain is composed of VH and VL domains, one of which is inactive, thereby forming an intermolecular ABD only after cleavage.

A "variable domain" herein is one or more substantially encoded by any of the Vκ, Vλ, and / or VH genes that make up the kappa, lambda, and heavy chain immunoglobulin loci, respectively. Means the region of immunoglobulin containing the Ig domain of. Each VH and VL is located in the following order, FR1-CDR1-FR2-CDR2-FR3-CDR3-FR4, from the amino terminus to the carboxy terminus with three hypervariable regions (“complementarity determining regions”, “CDRs””. ) And four "framework regions" or "FRs". Thus, the VH domain has the structure vhFR1-vhCDR1-vhFR2-vhCDR2-vhFR3-vhCDR3-vhFR4 and the VL domain has the structure blFR1-vrCDR1-vrFR2-vrCDR2-vrFR3-vrCDR3-vrFR4. As more fully described herein, the vhFR and vrFR regions self-assemble to form an Fv domain. In general, in the prodrug form of the invention, a "restraint Fv domain" in which the VH and VL domains cannot self-associate, and a "pseudo-Fv domain" that does not form a functional (active) antigen binding domain when the CDRs self-associate. "Exists.

The hypervariable region provides antigen binding specificity, generally approximately amino acid residues 24-34 (LCDR1, "L" indicates light chain), 50-56 (LCDR2), and 89-89 within the light chain variable region. 97 (LCDR3), as well as approximately 31-35B (HCDR1, "H" indicates heavy chain), 50-65 (HCDR2), and 95-102 (HCDR3) (Kabat et al., SEQENCES) within the heavy chain variable region. OF PROTEINS OF IMMUNOLOGICAL INTEREST, 5th Ed. Public Health Service, National Institutes of Health, Bethesda, Md. (1991)), and / or hypervariable loops in the light chain variable region (eg, R 12. , 50-52 (LCDR2), and 91-96 (LCDR3), and residues forming 26-32 (HCDR1), 53-55 (HCDR2), and 96-101 (HCDR3) in the heavy chain variable region. Includes amino acid residues from Chothia and Lesk (1987) J. Mol. Biol. 196: 901-917). Specific CDRs of the invention are described below.

As will be appreciated by those skilled in the art, the exact numbering and placement of CDRs may vary by different numbering systems. However, it should be understood that disclosure of variable heavy sequences and / or variable light sequences includes disclosure of related (unique) CDRs. Therefore, the disclosure of each variable weight region is the disclosure of vhCDRs (eg, vhCDR1, vhCDR2, and vhCDR3), and the disclosure of each variable light region is the disclosure of vlCDRs (eg, vlCDR1, vlCDR2, and vlCDR3).

Useful comparisons of CDR numbering are as follows: Lafranc et al. , Dev. Comp. Immunol. 27 (1): 55-77 (2003).
Table 1

Throughout the specification, the Kabat numbering system is generally used to refer to residues in the variable domain (approximately residues 1-107 in the light chain variable region and residues 1-113 in the heavy chain variable region) and Fc. An EU numbering system is used for the region (eg, Kabat et al., Supra (1991)).

The present invention provides a number of different CDR sets. In this case, the "complete CDR set" includes three variable light CDRs and three variable weight CDRs, such as vlCDR1, vlCDR2, vlCDR3, vhCDR1, vhCDR2, and vhCDR3. As will be appreciated by those skilled in the art, each set of CDRs, VH and VL CDRs, can both bind to the antigen individually and as a set. For example, in a constrained Fv domain, vhCDRs can bind to, for example, CD3s, and vrCDRs can bind to CD3s, but in a constrained form they cannot bind to CD3s.

These CDRs can be part of each of the larger variable light or variable heavy domains. In addition, as more fully outlined herein, variable heavy and variable light domains can be on separate polypeptide chains or, in the case of scFv sequences, on a single polypeptide chain.

CDRs contribute to the formation of antigen-binding sites, or more specifically, epitope-binding sites. "Epitope" refers to a determinant that interacts with a specific antigen binding site within a variable region known as a paratope. Epitopes are a population of molecules, such as amino acids or sugar side chains, and usually have specific structural and specific charge properties. A single antigen can have more than one epitope.

Epitopes are amino acid residues that are directly involved in binding (also called immunodominant constituents of the epitope) and other amino acid residues that are not directly involved in binding, such as amino acid residues that are effectively blocked by a specific antigen-binding peptide. It may contain a group, in other words, the amino acid residue is within the footprint of the specific antigen-binding peptide.

The epitope can be either conformational or linear. The conformational epitope is produced by spatially juxtaposed amino acids from different segments of the linear polypeptide chain. Linear epitopes are produced by adjacent amino acid residues in the polypeptide chain. Conformational and non-conformational epitopes can be distinguished in that their binding to the former is lost in the presence of a denaturing solvent, while the latter is not.

Epitopes typically contain at least 3, usually at least 5 or 8-10 amino acids in a unique spatial conformation. Antibodies that recognize the same epitope can be validated by simple immunoassays, such as "binning," which show the ability of one antibody to block the binding of another antibody to the target antigen. As outlined below, the invention includes not only the antigen-binding domains and antibodies listed herein, but also those that compete for binding to epitopes bound by the listed antigen-binding domains.

The variable heavy domain and variable light domain of the present invention can be "active" or "inactive".

As used herein, "inert VH" ("iVH") and "inert VL" ("iVL") are "inactive VL" ("iVL") when paired with their homologous VL or VH partners, respectively. A pseudo-Fv domain that forms the resulting VH / VL pair that does not specifically bind to the antigen to which the "active" VH or "active" VL would bind if bound to a similar VL or VH that is not "inactive". Refers to the constituents of. The exemplary "inactive VH" and "inactive VL" domains are formed by mutation of the wild-type VH or VL sequence. An exemplary mutation is within CDR1, CDR2, or CDR3 of VH or VL. An exemplary mutation involves placing a domain linker within CDR2, thereby forming an "inactive VH" or "inactive VL" domain. In contrast, "active VH" (aVH) or "active VL" (aVL) can specifically bind to its target antigen when paired with its "active" homologous partner, VL or VH, respectively. Is what it is.

In contrast, as used herein, the term "activity" refers to a CD3 binding domain that is capable of specifically binding to CD3. The term is a single member of an Fv binding pair (ie, VH or VL) that is in two contexts, (a) a sequence capable of pairing with its homologous partner and specifically binding to CD3. ), And (b) a homologous pair of sequences capable of specifically binding to CD3 (ie, VH and VL). An exemplary "active" VH, VL, or VH / VL pair is wild-type or parental sequence.

"CD-x" refers to a differentiated cluster (CD) protein. In an exemplary embodiment, CD-x is selected from CD proteins involved in the recruitment or activation of T cells in subjects to which the polypeptide constructs of the invention have been administered. In an exemplary embodiment, CD-x is CD3.

The term "binding domain" is used in the context of the present invention to (specifically) bind to / interact with a given target epitope or given target site on a target molecule (antigen), such as EGFR and CD3, respectively. Characterize the domain that acts / recognizes it. The structure and function of the target antigen binding domain (recognizing EGFR), and preferably the structure and / or function of the CD3 binding domain (recognizing CD3), is that of an antibody, eg, a full-length or total immunoglobulin molecule, including sdABD. Based on structure and / or function. According to the present invention, the target antigen binding domain is generally characterized by the presence of three CDRs that bind to the target tumor antigen (generally referred to in the art as variable weight domains, but the corresponding light chain CDRs are present. do not do). Alternatively, the ABD for the TTA comprises three light chain CDRs (ie, CDR1, CDR2, and CDR3 in the VL region) and / or three heavy chain CDRs (ie, CDR1, CDR2, and CDR3 in the VH region). Can be done. The CD3 binding domain preferably also includes at least the minimum structural requirements of the antibody to allow target binding. More preferably, the CD3 binding domain is at least three light chain CDRs (ie, CDR1, CDR2, and CDR3 in the VL region) and / or three heavy chain CDRs (ie, CDR1, CDR2, and CDR3 in the VH region). including. In exemplary embodiments, it is envisioned that the target antigen and / or CD3 binding domain can be produced or obtained by phage display or library screening methods.

As used herein, "domain" means a functional protein sequence as outlined herein. Domains of the invention include tumor target antigen binding domains (TTA domains), variable heavy domains, variable light domains, linker domains, and half-life extension domains.

As used herein, "domain linker" means an amino acid sequence that binds two domains as outlined herein. Domain linkers can be cleavable linkers, constrained cleavable linkers, non-cleavable linkers, constrained non-cleavable linkers, scFv linkers and the like.

As used herein, the term "hinge linker" means an amino acid sequence that binds a domain as outlined herein to the hinge region of an Fc domain. Hinge linkers can be cleavable linkers, constrained cleavable linkers, non-cleavable linkers, constrained non-cleavable linkers, scFv linkers and the like.

As used herein, the term "cleavable linker" ("CL") means an amino acid sequence that can be cleaved by a protease, preferably a human protease in the affected tissue, as outlined herein. Cleavable linkers are generally at least 3 amino acids long and have 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15 or more, depending on the flexibility required. Amino acids are used in the present invention.

As used herein, the term "non-cleavable linker" ("NCL") means an amino acid sequence that cannot be cleaved by a human protease under normal physiological conditions.

The "cleavable constraint linker" or "restricted cleavage linker" ("CCL") herein is significant to each other until two domains are present on different polypeptide chains, eg, after cleavage. Means a short polypeptide containing a protease cleavage site (as defined herein) that binds two domains as outlined herein in a manner that cannot interact with. When CCL binds to VH and VL domains as defined herein, VH and VL cannot self-assemble before cleavage to form functional Fv due to intramolecular steric hindrance. Upon cleavage by the relevant protease, VH and VL can aggregate to form an active antigen binding domain between the molecules. In general, CCL is less than 10 amino acids long and 9, 8, 7, 6, 5, and 4 amino acids are used in the present invention. In general, protease cleavage sites are generally at least 4+ amino acid lengths to give sufficient specificity, as shown in FIGS. 11A, 11B, and 11C.

The "non-cleavable constraint linker" ("NCCL") or "constrained non-cleavable linker" ("CNCL") herein is a method in which two domains cannot significantly interact with each other. It means a short polypeptide that binds two domains as outlined herein and is not significantly cleaved by human proteases under physiological conditions.

The term "restraint Fv domain" as used herein is a method in which an active heavy variable domain and a light variable domain cannot interact in a molecule to form an active Fv that binds to an antigen such as CD3. Means an Fv domain containing an active variable heavy domain and an active variable light domain covalently linked to a constraint linker as outlined in. Thus, the constrained Fv domain is similar to scFv, but is not capable of binding to the antigen due to the presence of the constrained linker.

The "pseudo-Fv domain" herein is defined as (i) a pseudo or inactive variable weight domain linked using a domain linker (which can be cleavable, constrained, non-cleavable, unconstrained, etc.). It means a pseudo or inactive variable light domain, (ii) a pseudo or inactive variable heavy domain and an active variable light domain, or (iii) a domain including an active variable heavy domain and a false or inactive variable light domain. The VHi and VLi domains of the pseudo-Fv domain do not bind to human antigens either when associated with each other (VHi / VLi) or when associated with active VH or VL, and thus VHi / VLi, VHi / VL. , And the VLi / VH Fv domains are not clearly bound to human proteins, thereby making these domains inactive in the human body.

The "single chain Fv" or "scFv" herein is covalently linked to a variable light (VL) domain using scFv as generally discussed herein to form a scFv or scFv domain. , Means variable weight (VH) domain. The scFv domain can be oriented either from the N-terminus to the C-terminus (VH-linker-VL or VL-linker-VH).

As used herein, "single domain Fv," "sdFv," "single domain antibody," or "sdABD" means an antigen-binding domain that has only three CDRs, generally based on camel antibody technology. See Protein Engineering 9 (7): 1129-35 (1994), Rev Mol Biotech 74: 277-302 (2001), Ann Rev Biochem 82: 775-97 (2013).

"Protease cleavage site" refers to an amino acid sequence that is recognized and cleaved by a protease. Suitable protease cleavage sites are outlined below.

As used herein, a "protease cleavage domain" is a "protease cleavage site", as well as between individual protease cleavage sites, and the protease cleavage site (s) and other functional components of the constructs of the invention. Refers to a peptide sequence that incorporates any linker between (eg, VH, VL, VHi, VLi, target antigen binding domain (s), extended half-life domain, etc.).

As used herein, "Fc" or "Fc region" or "Fc domain" means a polypeptide comprising a constant region of an antibody excluding a first constant region immunoglobulin domain. In the case of IgG, the Fc domain comprises the immunoglobulin domains Cγ2 and Cγ3 (CH2 and CH3) and optionally all or part of the hinge region between Cγ1 (CH1) and Cγ2 (CH2). In EU numbering of human IgG1, the CH2-CH3 domain contains amino acids 231-447 and the hinges are 216-230. Therefore, the definition of "Fc domain" includes both amino acids 231-447 (CH2-CH3) or 216-447 (hinge domain-CH2-CH3).

III. Proteins of the Invention The proteins of the invention have many different constituents, which are linked together in various ways, commonly referred to herein as domains. Some of the domains are binding domains, each of which binds to a target antigen (eg, TTA or eg, CD3). They are referred to herein as "multispecific" because they bind to more than one antigen, for example, the prodrug constructs of the invention can bind to TTA and CD3 and are therefore shown in FIG. It is "bispecific" as it is. The proteins of the invention can also have higher specificity, eg, the first antigen binding domain binds to EGFR, the second antigen binding domain binds to EpCAM, and an anti-CD3 binding domain is present. If so, this is a "trispecific" molecule.

The proteins of the invention can include CD3 antigen binding domains arranged in various ways as outlined herein, such as tumor target antigen binding domains, half-life extension domains, linkers and the like.

In some embodiments, the first protein comprises a first tumor target antigen binding domain and the second protein comprises a second tumor target antigen binding domain thereby comprising a first tumor target antigen. The binding domain and the second tumor targeting antigen binding domain bind to the same tumor targeting antigen. In certain cases, the first tumor target antigen domain and the second tumor target antigen domain bind to different epitopes, regions, or portions of the same tumor target antigen. In some cases, the first tumor target antigen domain and the second tumor target antigen domain bind to different tumor target antigens.

The proteins of the invention can be produced by co-expression and co-purification in cells to give complementary pairs of proteins capable of binding to CD3 and tumor target antigens. In some embodiments, each of the complementary pair of proteins is purified separately. In some embodiments, each of the complementary pair of proteins is purified simultaneously or incidentally.

In some embodiments, the expression vector comprises a nucleic acid sequence encoding one protein of a complementary protein and a nucleic acid sequence encoding the other protein of the complementary protein. In some embodiments, the host cell comprises such an expression vector. In some cases, such host cells can be cultured in culture medium under suitable conditions to produce proteins. In some embodiments, the host cells can be cultured under suitable conditions to secrete the proteins described herein into the culture medium. In certain embodiments, the culture medium containing the secretory protein of the invention is purified to obtain a complementary pair of protein proteins. Useful purification methods include protein A chromatography, protein G chromatography, heparin binding, reverse phase chromatography, HIC chromatography, CHT chromatography, affinity chromatography, anion exchange chromatography, cation exchange chromatography, size. Exclusion chromatography and the like, but not limited to these.

A. CD3 Antigen Binding Domain The specificity of the T cell response is mediated by the recognition of the antigen (major histocompatibility complex, presented in the context of MHC) by the T cell receptor complex. As part of the T cell receptor complex, CD3 is a protein complex that comprises a CD3γ (gamma) chain, a CD3δ (delta) chain, and two CD3ε (epsilon) chains present on the cell surface. CD3 contains a T cell receptor complex that associates with all of the α (alpha) and β (beta) chains of the T cell receptor (TCR) and CD-ζ (zeta). Clustering of CD3 on T cells, such as by the Fv domain that binds to CD3, results in T cell activation similar to the involvement of T cell receptors, but independent of the specificity typical of the clone.

However, as is known in the art, CD3 activation can cause many toxic side effects, and therefore the invention later cleaves the prodrug polypeptide of the invention to activate the CD3 binding domain. It relates to providing the active CD3 binding of the polypeptides of the invention only in the presence of tumor cells, where specific proteases are found. Thus, in the present invention, binding of the anti-CD3 Fv domain to CD3 is only in the microenvironment of affected cells or tissues with high levels of protease, eg, in the tumor microenvironment as described herein. It is regulated by a protease cleavage domain that limits the binding of the Fv domain to CD3.

Accordingly, the present invention provides two sets of VH and VL domains, an active set (VH and VL), and an inactive set (VHi and VLi), all four in the prodrug construct (s). Exists. The construct is formalized so that the VH and VL sets cannot self-associate, but rather associate with an inactive partner, such as VHi and VL and VLi and VH as shown herein.

There are many suitable active CDR sets known in the art and / or VH and VL domains used in the present invention. For example, the CDR and / or VH and VL domains are, for example, muromonab-CD3 (OKT3), otelixizumab (TRX4), teplizumab (MGA031), bisirizumab (Nuvion), SP34 or I2C, TR-66 or X35-3, VIT3, BMA030 (BW264 / 56), CLB-T3 / 3, CRIS7, YTH12.5, F111-409, CLB-T3.4.2, TR-66, WT32, SPv-T3b, 11D8, XIII-141, XIII-46 , XIII-87, 12F6, T3 / RW2-8C8, T3 / RW2-4B6, OKT3D, M-T301, SMC2, F101.01, UCHT-1, and WT-31 derived from known anti-CD3 antibodies.

In some embodiments, the VH and VL sequences that form the active Fv domain that binds to human CD3 are shown in FIG. 13A as construct I-3 (SEQ ID NO: 16) and construct I-4 (SEQ ID NO: 17).

The inactive VHi and VLi domains contain a "regular" framework region (FR) that allows association, whereby the inactive variable domain associates with the active variable domain and inactivates the pair. For example, it prevents it from binding to CD3. In one embodiment, VHi and VLi form an inactive Fv domain when one or both of the inactive domains are present in a complementary construct pair. In one embodiment, the VHi and VLi forming the inactive Fv domain in the presence of one or both of the inactive domains are shown in FIG. 13A for constructs I-5 (SEQ ID NO: 18) and construct I-6 (SEQ ID NO: 19). ).

In some embodiments, the amino acid sequence of the inactive VLi domain is as Pro36 (SEQ ID NO: 1) in FIG. 14A, Pro67 (SEQ ID NO: 4) in FIG. 14B, Pro70 (SEQ ID NO: 7) in FIG. 14D, and It is shown as Pro218 (SEQ ID NO: 10) in FIG. 14E. In some embodiments, the amino acid sequence of the inactive VHi domain is shown as Pro37 (SEQ ID NO: 2) in FIG. 14A, Pro38 (SEQ ID NO: 3) in FIG. 14B, and Pro68 (SEQ ID NO: 5) in FIG. 14C. It is shown as Pro70 (SEQ ID NO: 7) in 14D, as Pro71 (SEQ ID NO: 8) in FIG. 14D, and as Pro218 (SEQ ID NO: 10) in FIG. 14E.

In some embodiments, the inactive VHi domain prevents the paired VHi-VL domain from binding to the target antigen when paired with the active VL domain, such as one or more, eg, 1. Includes 2, 3, 4, 5, 6, 7, 8, 9, or more amino acid modifications (eg, amino acid insertions, deletions, or substitutions). In other embodiments, the inactive VLi domain prevents the paired VH-VLi from binding to the target antigen when paired with the active VH domain, such as one or more, eg, 1, 2, Includes 3, 4, 5, 6, 7, 8, 9, or more amino acid modifications (eg, amino acid insertions, deletions, or substitutions).

As will be appreciated by those skilled in the art, there are many "inactive" variable domains used in the present invention. Basically, no matter what amino acid is at the CDR position within the variable region, any variable domain with a human framework region that allows self-assembly with another variable domain can be used. For clarity, inactive domains are said to contain CDRs, but technically, inactive variable domains do not confer binding capacity.

In some cases, the inert domain promotes selective binding in a prodrug form to facilitate the formation of intramolecular VHi-VL and VH-VLi domains prior to cleavage (eg, rather than intermolecular pairing). Can be manipulated. For example, by reference, Igawa et al., Which is incorporated herein by reference in its entirety, particularly with respect to interfacial residue amino acid substitutions. , Protein Eng. Des. See Selection 23 (8): 667-677 (2010).

In one aspect, the polypeptide constructs described herein comprise a domain that specifically binds to CD3 when activated by a protease. In one aspect, the polypeptide construct described herein comprises two or more domains that specifically bind to human CD3 when activated by a protease. In some embodiments, the polypeptide constructs described herein comprise two or more domains that specifically bind to CD3ε when activated by a protease. In some embodiments, the polypeptide constructs described herein comprise two or more domains that specifically bind to CD3ε when activated by a protease.

In some embodiments, the protease cleavage site lies between the anti-CD3 active VH domain and the inactive VL domain on the first monomer, preventing them from folding and binding to CD3 on T cells. In some embodiments, the protease cleavage site lies between the anti-CD3 inactive VH domain and the active VL domain on the second monomer, preventing them from folding and binding to CD3 on T cells. When the protease cleavage site is cleaved by a protease present in the target cell, the anti-CD3 active VH domain of the first monomer and the anti-CD3 active VL domain of the second monomer can bind to CD3 on T cells. it can.

In certain embodiments, the CD3 binding domains of the polypeptide constructs described herein not only exhibit strong CD3 binding affinity with human CD3, but also have excellent cross-reactivity with the respective cynomolgus monkey CD3 protein. Also shown. In some cases, the CD3 binding domain of the polypeptide construct is cross-reactive with CD3 from cynomolgus monkeys. In certain cases, the human: kanikuiza KD ratio to CD3 is 5 to 0.2.

In some embodiments, the CD3 binding domain of the antigen-binding protein includes, but is not limited to, domains from monoclonal antibodies, polyclonal antibodies, recombinant antibodies, human antibodies, humanized antibodies. Can be. In some cases, it is beneficial that the CD3 binding domain is derived from the same species in which the antigen binding protein is ultimately used. For example, for use in humans, it may be beneficial that the CD3 binding domain of the antigen binding protein comprises a human or humanized residue from the antigen binding domain of the antibody or antibody fragment.

Thus, in one aspect, the antigen binding domain comprises a humanized or human binding domain. In one embodiment, the humanized or human anti-CD3 binding domains are one or more (eg, all three) light chain complementarity determining regions 1 (eg, all three) of the humanized or human anti-CD3 binding domains described herein. LC CDR1), light chain complementarity determining regions 2 (LC CDR2), and light chain complementarity determining regions 3 (LC CDR3), and / or one or more of the humanized or human anti-CD3 binding domains described herein. Heavy chain complementarity determining regions 1 (HC CDR1), heavy chain complementarity determining regions 2 (HC CDR2), and heavy chain complementarity determining regions 3 (HC CDR3), eg, one. Includes humanized or human anti-CD3 binding domains comprising, for example, all three LC CDRs and one or more, eg, all three HC CDRs.

In some embodiments, the humanized or human anti-CD3 binding domain comprises a CD3-specific humanized or human light chain variable region, and the CD3-specific light chain variable region is a human light chain framework region. Contains human or non-human light chain CDRs within. In certain cases, the light chain framework region is the λ (lambda) light chain framework. In other cases, the light chain framework region is the κ (kappa) light chain framework.

In some embodiments, the one or more CD3 binding domains are humanized or fully human. In some embodiments, one or more activated CD3 binding domains have less than 1000 nM KD binding to CD3 on CD3 expressing cells. In some embodiments, one or more activated CD3 binding domains have less than 100 nM KD binding to CD3 on CD3 expressing cells. In some embodiments, one or more activated CD3 binding domains have less than 10 nM KD binding to CD3 on CD3 expressing cells. In some embodiments, one or more CD3 binding domains are cross-reactive with cynomolgus monkey CD3. In some embodiments, one or more CD3 binding domains include the amino acid sequences provided herein.

In some embodiments, the humanized or human anti-CD3 binding domain comprises a CD3-specific humanized or human heavy chain variable region, and the CD3-specific heavy chain variable region is a human heavy chain framework region. Includes human or non-human heavy chain CDRs within.

In one embodiment, the anti-CD3 binding domain is an Fv comprising the light and heavy chains of the amino acid sequences provided herein. In embodiments, the anti-CD3 binding domain is at least one, two, or three modifications (eg, substitutions, insertions, and deletions) of the amino acid sequence of the light chain variable region provided herein. , 20, or 10 or less amino acid sequences (eg, substitutions, insertions, and deletions), or sequences having 95-99% identity with the amino acid sequences provided herein. At least one, two, or three modifications (eg, substitutions, insertions, and deletions) of the amino acid sequences of the chain variable regions and / or the heavy chain variable regions provided herein, but 30, 20, Alternatively, a heavy chain variable region comprising an amino acid sequence having 10 or less modifications (eg, substitutions, insertions, and deletions), or a sequence having 95-99% identity with the amino acid sequences provided herein. including. In one embodiment, the humanized or human anti-CD3 binding domain is scFv and the light chain variable region comprising the amino acid sequence described herein is via the scFv linker to the amino acid sequence described herein. It is attached to the heavy chain variable region including. The light and heavy chain variable regions of scFv can be, for example, either the following orientation, light chain variable region-scFv linker-heavy chain variable region or heavy chain variable region-scFv linker-light chain variable region. ..

In some embodiments, the CD3 binding domain of the antigen binding protein has an affinity for CD3 on CD3 expressing cells, with KD of 1000 nM or less, 100 nM or less, 50 nM or less, 20 nM or less, 10 nM or less, 5 nM or less. It is 1 nM or less, or 0.5 nM or less. In some embodiments, the CD3 binding domain of the antigen binding protein has an affinity for CD3ε and the KD is 1000 nM or less, 100 nM or less, 50 nM or less, 20 nM or less, 10 nM or less, 5 nM or less, 1 nM or less, or 0. It is 5 nM or less. In a further embodiment, the CD3 binding domain of the antigen binding protein has a low affinity for CD3, ie about 100 nM or more.

The binding affinity for CD3 is, for example, as is known in the art, microbial cells in which the antigen binding protein itself or its CD3 combined domain is coated on an assay plate, generally using the Biacore or Octet assay. It can be determined by the ability to bind to CD3, such as in solution, presented on the surface. The binding activity of the antigen-binding protein itself or its CD3 binding domain to CD3 is assayed by immobilizing a ligand (eg, CD3) or the antigen-binding protein itself or its CD3 binding domain on beads, substrates, cells and the like. obtain. The drug can be added to a suitable buffer and the binding partner can be incubated for a period of time at a given temperature. After washing to remove unbound material, the bound protein can be released, for example, in a high pH SDS buffer, and analyzed, for example, by surface plasmon resonance (SPR).

B. Antigen Binding Domains for Tumor Target Antigens In addition to the CD3 and half-life extension domains described, the polypeptide constructs described herein are also one or more target antigens, or one or more on a single target antigen. It also includes at least one, or at least two, or more domains that bind to the region. As used herein, the polypeptide constructs of the invention are cleaved, for example, in a disease-specific microenvironment or in the blood of a subject in a protease cleavage domain, and each target antigen binding domain is a target antigen on a target cell. It is intended to bind to and thereby activate the CD3 binding domain to bind to T cells. In general, TTA-binding domains can bind to their targets prior to protease cleavage, and thus they can "wait" on target cells to be activated as T cell engagers. At least one target antigen is involved in and / or is associated with a disease, disorder, or condition. Exemplary target antigens are proliferative, neoplastic, inflammatory, immunological, autoimmune, infectious, viral, allergic, parasitic, implant-to-host, or host. Includes those associated with anti-transplant disease. In some embodiments, the target antigen is a tumor antigen expressed on tumor cells. Alternatively, in some embodiments, the target antigen is associated with a pathogen such as a virus or bacterium. At least one target antigen can also be directed to healthy tissue.

In some embodiments, the target antigen is a cell surface molecule such as a protein, lipid, or polypeptide. In some embodiments, the target antigen is on tumor cells, virus-infected cells, bacterial-infected cells, damaged red blood cells, arterial plaque cells, or fibrous tissue cells. It is contemplated herein that upon binding to two or more target antigens, the two inactive CD3 binding domains will co-localize and form an active CD3 binding domain on the surface of the target cell. In some embodiments, the antigen binding protein comprises two or more target antigen binding domains to activate the inactive CD3 binding domain in the antigen binding protein. In some embodiments, the antigen-binding protein comprises two or more target antigen-binding domains in order to enhance the binding strength to the target cell. In some embodiments, the antigen-binding protein comprises two or more target antigen-binding domains in order to enhance the binding strength to the target cell. In some embodiments, the two or more antigen binding domains include the same antigen binding domain. In some embodiments, the two or more antigen binding domains include different antigen binding domains. For example, two different antigen-binding domains known to be doubly expressed in affected cells or tissues, such as tumors or cancer cells, enhance the binding or selectivity of the antigen-binding protein for the target. obtain.

The polypeptide constructs contemplated herein include at least one antigen binding domain, which binds to at least one target antigen. The target antigen is optionally expressed on the surface of affected cells or tissues, such as tumors or cancer cells. Target antigens include epidermal growth factor receptor (EpCAM), epidermal growth factor receptor (EGFR), human epidermal growth factor receptor 2 (HER-2), human epidermal growth factor receptor 3 (HER-3), c- Examples include, but are not limited to, the LY6 / PLAYR domain (LYPD3) containing Met, folic acid receptor 1 (FOLR1), B7H3 (CD276), and 3 carcinoembryonic antigen (CEA).

The polypeptide constructs disclosed herein also include proteins containing two antigen binding domains that bind to two different target antigens known to be expressed on affected cells or tissues. Exemplary antigen-binding domain pairs include, but are not limited to, EGFR / CEA, EpCAM / CEA, EGFR / EpCAM, and HER-2 / HER-3.

The design of the polypeptide constructs described herein is limited to, but not limited to, a binding domain to one or more target antigens, a monoclonal antibody, a polyclonal antibody, a recombinant antibody, a human antibody, a humanized antibody. Allows flexibility in being able to be any kind of binding domain, including domains from. In some embodiments, the binding domain to the target antigen is a variable single chain variable fragment (scFv), a single domain antibody such as a heavy chain variable domain (VH), a light chain variable domain (VL), and a camel-derived Nanobody. It is a domain (VHH). In other embodiments, the binding domain to the target antigen is a non-Ig binding domain, i.e. an antibody mimetic, eg, anticarine, affylin, affibody molecule, affimer, affitin, alphabody, avimer, DARPins, finomer, knitz domain peptide, And monobody. In a further embodiment, the binding domain to one or more target antigens is a receptor domain, lectin, or peptide that binds to or is associated with one or more antigens.

In some embodiments, the target cell antigen binding domain independently comprises a scFv, VH domain, VL domain, non-Ig domain, or ligand that specifically binds to the target antigen. In some embodiments, the target antigen-binding domain specifically binds to cell surface molecules. In some embodiments, the target antigen-binding domain specifically binds to the tumor antigen. In some embodiments, the target antigen binding domain is specific and independent of the antigen selected from at least one of EpCAM, EGFR, HER-2, HER-3, cMet, LyPD3, CEA, and FoIR. And combine. In some embodiments, the target antigen-binding domain specifically and independently binds to two different antigens, with at least one of these antigens being EpCAM, EGFR, HER-2, HER-3, It is selected from one of cMet, CEA, LyPD3, B7H3, and FOLR1.

In many embodiments, the antigen-binding domain (ABD) to the target tumor antigen (TTA) is a single-domain antigen-binding domain (sdABD-TTA) based on a camel single-domain antibody (sdABD). sdABD-TTA has a conventional antibody-like framework region, as well as three CDRs, but no heavy chain constant domain. These sdABD-TTAs are generally preferred over scFvs that bind to TTAs, as the intramolecular folding that results in the formation of inactive Fv that does not bind to CD3 is less complex and has fewer VH and VL domains. These sdABD-TTAs can be labeled by the target to which they bind. For example, sdABD-EGFR is sdABD that binds to human EGFR and the like.

In some embodiments, the antigen binding domain binds to EGFR and has the amino acid sequence set forth in SEQ ID NO: 14 or shown as construct I-1 in FIG. 13A. In some embodiments, the antigen binding domain binds to EGFR and has a humanized version of the amino acid sequence set forth in SEQ ID NO: 14 or shown as construct I-1 in FIG. 13A. In other embodiments, the antigen binding domain binds to EGFR and has the CDR and / or variable domain of the sequence set forth in SEQ ID NO: 14 or shown as construct I-1 in FIG. 13A.

In other embodiments, the antigen binding domain binds to EGFR and has the amino acid sequence set forth in SEQ ID NO: 21 or shown as construct I-8 in FIG. 13A. In other embodiments, the antigen binding domain binds to EGFR and has a humanized version of the amino acid sequence set forth in SEQ ID NO: 21 or shown as construct I-8 in FIG. 13A. In some embodiments, the antigen binding domain binds to EGFR and has the CDR and / or variable domain of the sequence set forth in SEQ ID NO: 21 or shown as construct I-8 in FIG. 13A.

In some embodiments, the antigen binding domain binds to EGFR and has the amino acid sequence set forth in SEQ ID NO: 22 or shown as construct I-11 in FIG. 13A. In other embodiments, the antigen binding domain binds to EGFR and has the CDR and / or variable domain of the sequence set forth in SEQ ID NO: 22 or shown as construct I-11 in FIG. 13A. In certain embodiments, the antigen binding domain binds to EGFR and has the amino acid sequence set forth in SEQ ID NO: 23 or shown as construct I-12 in FIG. 13B. In various embodiments, the antigen binding domain binds to EGFR and has the CDR and / or variable domain of the sequence set forth in SEQ ID NO: 23 or shown as construct I-12 in FIG. 13B.

In some embodiments, the antigen binding domain binds to EpCAM and has the amino acid sequence set forth in SEQ ID NO: 15 or shown as construct I-2 in FIG. 13A. In other embodiments, the antigen binding domain binds to EpCAM and has a humanized version of the amino acid sequence set forth in SEQ ID NO: 15 or shown as construct I-2 in FIG. 13A. In some embodiments, the antigen binding domain binds to EpCAM and has the CDR and / or variable domain of the sequence set forth in SEQ ID NO: 15 or shown as construct I-2 in FIG. 13A.

In some embodiments, the pre-cleavage protein of the protease cleavage domain is less than about 100 kDa. In some embodiments, the protein after cleavage of the protease cleavage domain is about 25-about 75 kDa. In some embodiments, the protein prior to protease cleavage has a size that exceeds the renal threshold for first pass clearance. In some embodiments, the protein before protease cleavage has an elimination half-life of at least about 50 hours. In some embodiments, the protein before protease cleavage has an elimination half-life of at least about 100 hours. In some embodiments, the protein has increased tissue penetration as compared to IgG for the same target antigen. In some embodiments, the protein has an increased tissue distribution as compared to IgG for the same target antigen.

C. Half-life extension The proteins of the invention optionally include a half-life extension domain. It is contemplated that such domains include, but are not limited to, HSA binding domains, Fc domains, small molecules, and other half-life extension domains known in the art.

1. 1. Fc Regions The proteins of the invention include Fc domain fusion proteins that combine the Fc region of an antibody with additional components outlined herein, the ABD-TTA and Fv domains, generally the pseudodomains outlined herein. including.

The knob-in-hole form of the heterodimer Fc protein described herein refers to an amino acid substitution (s) that results in a "stereoscopic effect" that favors heterodimer formation over homodimer formation. .. In some cases, the knob-in-hole form may be combined with a charged amino acid-substituted disulfide bond or pair to further favor the heterodimer form.

In some embodiments, the heterodimer Fc protein comprises an Fc arm containing either a knob or a hole in the Fc region. In other words, the first monomer Fc arm includes a knob and the second monomer Fc arm includes a hole. In the embodiment of "construct 6" or "construct 7", the monomeric Fv arm containing the active Fv domain (eg, anti-CD3 variable heavy chain and variable light chain) comprises a CH3-knob and a pseudo Fv domain (eg, eg). Monomer Fc arms containing (inert variable heavy chain and inactive variable light chain) contain CH3-holes, and vice versa. In other embodiments, the monomeric Fc arm containing the active Fv domain comprises a CH3-hole and the monomeric Fc arm containing a pseudoFv domain comprises a CH3-knob.

Amino acid residues for the formation of knobs are generally naturally occurring amino acid residues and are selected from arginine (R), phenylalanine (F), tyrosine (Y), and tryptophan (W). In some preferred embodiments, the amino acid residues are tryptophan and tyrosine. In one embodiment, the original residue for the formation of the knob has a small side chain volume such as alanine, asparagine, aspartic acid, glycine, serine, threonine, or valine. Exemplary amino acid substitutions in the CH3 domain for forming knobs include, but are not limited to, T366W, T366Y, or F405W substitutions.

The amino acid residue for the formation of holes is usually a naturally occurring amino acid residue and is selected from alanine (A), serine (S), threonine (T), and valine (V). In some preferred embodiments, the original residue for hole formation has a large side chain volume such as tyrosine, arginine, phenylalanine, or tryptophan. Exemplary amino acid substitutions in the CH3 domain to generate holes include, but are not limited to, T366S, L368A, F405A, Y407A, Y407T, and Y407V substitutions. In certain embodiments, the knob comprises a T366W substitution and the hole comprises a T366S / L368A / Y407V substitution.

In general, the preferred Fc domain for use herein is the human IgG domain, generally either IgG1 or IgG4. In some cases, for example, when effector function is undesirable, IgG4 is used, and in some cases it contains an S228P variant in the hinge domain to prevent arm replacement.

It is understood that other modifications to Fc regions known in the art that promote heterodimerization are also contemplated and incorporated by this application.

In some embodiments, the Fc region of the form described herein is, but is not limited to, a histidine tag (eg, (His) 6) at the C-terminus of the Fc, a streptavidin tag (eg, a strep tag or). Includes tags such as Strepttag II), or Maltose-binding protein (MBP) tags.

In addition, the Fc domain contains additional amino acid modifications to modify effector function or half-life, as is known in the art.

2. 2. Human Serum Albumin Binding Domain Human serum albumin (HSA) (molecular weight about 67 kDa) is the most abundant protein in plasma present at about 50 mg / mL (600 μM) and has a half-life of about 20 days in humans. HSA serves to maintain plasma pH, contributes to colloidal blood pressure, functions as a carrier for many metabolites and fatty acids, and serves as a major drug transport protein in plasma.

Non-covalent associations with albumin prolong the elimination half-life of short-lived proteins. For example, recombinant fusion of albumin-binding domains to Fab fragments reduced in vivo clearance by 25-58 fold and 26-fold when administered intravenously to mice and rabbits, respectively, compared to administration of Fab fragments alone. It resulted in a half-life extension of ~ 37 times. In another example, when insulin was acylated with fatty acids to promote association with albumin, a sustained effect was observed when injected subcutaneously into rabbits or pigs. Together, these studies demonstrated a link between albumin binding and sustained action.

In one aspect, the antigen-binding proteins described herein include a half-life extension domain, eg, a domain that specifically binds to HSA. In some embodiments, the HSA-binding domain of an antigen-binding protein includes, but is not limited to, domains from monoclonal antibodies, polyclonal antibodies, recombinant antibodies, human antibodies, humanized antibodies, but any domain that binds to HSA. Can be. In some embodiments, the HSA binding domain is a single chain variable fragment (scFv), a single domain antigen binding domain (sdABD), eg, a camel-derived Nanobody, peptide, ligand, or small molecule heavy chain specific for HSA. Variable domains (VH), light chain variable domains (VL), and variable domains (VHH). In certain embodiments, the HSA binding domain is derived from a single domain antibody (sdABD) and comprises a single domain antigen binding domain (sdABD), i.e., sdABD is the standard six in Fv of conventional antibodies. It is a single variable domain (VHH) containing three CDRs rather than a CDR. In other embodiments, the HSA binding domain is a peptide. In a further embodiment, the HSA binding domain is a small molecule. It is contemplated that the HSA binding domain of the antigen-binding protein is fairly small, and in some embodiments, 25 kD or less, 20 kD or less, 15 kD or less, or 10 kD or less. In certain cases, the HSA binding domain, if it is a peptide or small molecule, is 5 kD or less.

The half-life extension domain of an antigen-binding protein provides modified pharmacodynamics and pharmacokinetics of the antigen-binding protein itself. As mentioned above, the extended half-life domain extends the elimination half-life. The half-life extension domain also alters pharmacological properties, including alterations in tissue distribution, penetration, and diffusion of antigen-binding proteins. In some embodiments, the half-life prolongation domain has improved tissue (including tumor) targeting, tissue penetration, tissue distribution, tissue diffusion, and compared to proteins without a half-life prolongation binding domain. Provides enhanced efficacy. In one embodiment, the therapy effectively and efficiently utilizes reduced amounts of antigen-binding protein, resulting in reduced side effects such as reduced cytotoxicity of non-tumor cells.

In addition, features of extended half-life domains, such as HSA binding domains, include the binding affinity of HSA binding domains for HSA. The affinity of the HSA binding domain can be selected to target a particular elimination half-life in a particular polypeptide construct. Therefore, in some embodiments, the HSA binding domain has a high binding affinity. In other embodiments, the HSA binding domain has a moderate binding affinity. In yet other embodiments, the HSA binding domain has low or slight binding affinity. Exemplary binding affinities include KD concentrations of 10 nM or less (high), 10 nM to 100 nM (medium), and more than 100 nM (low). As mentioned above, the binding affinity for HSA is determined by known methods such as surface plasmon resonance (SPR).

D. Protease Cleavage Sites The polypeptide (eg, protein) compositions of the invention, in particular the prodrug constructs, generally include one or more protease cleavage sites present within a cleaving linker, as outlined herein.

As described herein, the prodrug construct of the invention comprises at least one protease cleavage site comprising an amino acid sequence that is cleaved by at least one protease. In some cases, the proteins described herein are cleaved by at least one protease 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, Includes 15, 16, 17, 18, 19, 20 or more protease cleavage sites. As more fully discussed herein, when two or more protease cleavage sites are used in prodrug construction, they may be the same (eg, multiple sites cleaved by a single protease). ), May be different (two or more cleavage sites are cleaved by at least two different proteases). As will be appreciated by those skilled in the art, constructs containing three or more protease cleavage sites can utilize one, two, three, etc., for example, some constructs for two different proteases. Three parts can be used.

The amino acid sequence of the protease cleavage site depends on the targeted protease. As is known in the art, there are many human proteases found in the body that can be associated with disease states.

Proteases are known to be secreted by several affected cells and tissues, such as tumors or cancer cells, to create a protease-rich or protease-rich microenvironment. In some cases, the blood of interest is rich in proteases. In some cases, the cells surrounding the tumor secrete proteases into the tumor microenvironment. The cells surrounding the tumor that secrete protease include tumor stromal cells, myofibroblasts, blood cells, mast cells, B cells, NK cells, regulatory T cells, macrophages, cytotoxic T cells, and dendritic cells. , Membrane stem cells, polymorphonuclear cells, and other cells, but not limited to these. In some cases, proteases, eg, proteases that target the amino acid sequences found in microbial peptides, are present in the blood of interest. This feature allows targeted therapeutic agents such as antigen-binding proteins to have additional specificity because T cells are not bound by antigen-binding proteins, except in the protease-rich microenvironment of the target cells or tissues. ..

Proteases are proteins that, in some cases, cleave proteins in a sequence-specific manner. Proteases include serine protease, cysteine protease, aspartic acid protease, threonine protease, glutamate protease, metalloprotease, asparagine peptide lyase, serum protease, and Cathesin (eg, Cathesin B, Cathesin C, Cathesin D, Cathesin E, Cathesin K, Cathesin K, Cathesin K, Cathesin. L, CathesinS), kalikrein, hK1, hK10, hK15, KLK7, GranzymeB, plasmin, collagenase, type IV collagenase, stromerycin, factor Xa, chymotrypsin-like protease, trypsin-like protease, elastase-like protease, subtilisin-like protease, actinidin. , Carpine, Caspase (eg Caspase-3), Mira1-CP, papain, HIV-1 protease, HSV protease, CMV protease, chymosin, renin, pepsin, matliptase, regmine, plasmepsin, nepenthesin, metalloexopeptidase, metalloend Peptidase, matrix metalloprotease (MMP), MMP1, MMP2, MMP3, MMP8, MMP9, MMP13, MMP11, MMP14, mepurin, urokinase-type plasminogen activator (uPA), enterokinase, prostate-specific antigens (PSA, hK3) , Interleukin-1β converting enzyme, thrombin, FAP (FAP-α), dipeptidyl peptidase, and dipeptidyl peptidase IV (DPPIV / CD26), but are not limited thereto.

Some suitable proteases and protease cleavage sequences are described as SEQ ID NOs: 21-28 and 29-88 and are shown in FIGS. 11A, 11B, 11C, 13B, and 14A-14F.

E. Linkers As discussed herein, the different domains of the invention are generally linked together using an amino acid linker, which is also functional, including flexibility or inflexibility (eg, steric hindrance). , As well as the ability to be cleaved using in-situ proteases. These linkers can be classified in many ways.

The present invention provides a "domain linker" used to bind two or more domains (eg, a target tumor antigen binding domain for VH and VL, VH or VL (TTABD, "αTTA" herein. (Sometimes referred to (as opposed to "anti-TTA")), such as a half-life extension domain for another component). Domain linkers can be, for example, non-cleavable linkers (NCLs), cleavable linkers (“CL”), cleavable and constrained linkers (CCL), and non-cleavable and constrained linkers (NCCL). In some embodiments, the constraint linker binds the two domains as outlined herein in a manner that prevents the two domains from interacting significantly with each other and is human under physiological conditions. A short polypeptide of less than 10 amino acids (eg, 9, 8, 7, 6, 5, or 4 amino acids) that is not significantly cleaved by the protease. In general, protease cleavage sites are generally at least 4+ amino acid lengths to give sufficient specificity, as shown in FIGS. 11A-11C.

1. 1. Non-cleavable linker In one embodiment, the domain linker is a non-cleavable linker (NCL). In this embodiment, the linker is generally used to bind to the domain and maintain the functionality of the domain through a longer flexible domain that is not cleaved by the in-situ protease in the patient. Examples of internal non-cleavable linkers suitable for binding domains in the polypeptides of the invention are (GS) n, (GGS) n, (GGGS) n (SEQ ID NO: 27), (GGSG) n (sequence). Number 28), (GGSGG) n (SEQ ID NO: 29), or (GGGGS) n (SEQ ID NO: 30) (n is 1, 2, 3, 4, 5, 6, 7, 8, 9, or 10). ), But is not limited to these.

In some embodiments, the linker does not contain a cleavage site and is too short to allow the protein domains separated by the linker to self-assemble within the molecule, a "restraint uncleavable linker" or It is "CNCL". For example, in Pro219 and Pro217, active VH and active VL are separated by 8 amino acids (“8mer”) that do not allow VH and VL to self-assemble into the active antigen binding domain, and instead are tumor proteases. Intermolecular assembly with Pro218 occurs instead until cleavage by. In some embodiments, the linker is still flexible, eg, (GGGS) n (n = 2). In other embodiments, more rigid linkers, such as those containing proline or bulky amino acids, may be used, although generally less preferred.

2. 2. Cleavable Linkers All of the prodrug constructs herein include at least one cleaving linker. Thus, in one embodiment, the domain linker is cleaveable (CL) and is sometimes referred to herein as a "protease cleaving domain"("PCD"). In this embodiment, the CL contains a protease cleavage site, as outlined herein and as shown in FIGS. 11A, 11B, and 11C. In some cases, CL contains only protease cleavage sites. Optionally, there may be several more linking amino acids at either or both of the N-terminus and C-terminus of CL, depending on the length of the cleavage site, eg, either the N-terminus or the C-terminus of the cleavage site. Or both may have 1, 2, 3, 4, or 5-8 amino acids.

IV. Expression Methods The present invention provides nucleic acids encoding two monomers of the heterodimer protein of the present invention, as well as expression vectors and host cells. As will be appreciated by those skilled in the art, one or two expression vectors can be made. That is, the first nucleic acid encoding the first monomer and the second nucleic acid encoding the second monomer can be placed in a single expression vector or two expression vectors. The expression vector (s) are then placed in a host cell that is grown to express the two monomers. In some cases, which is generally not preferred, each monomer can be produced in a separate host cell and then the expression products are combined to form the heterodimer prodrug protein of the invention.

However, most embodiments depend on the use of co-expression of the two monomers. That is, a method for producing a protein of the invention by co-expression and co-purification in a cell (eg, a host cell) to obtain a first monomeric Fc polypeptide and a second monomeric Fc polypeptide is described herein. Provided in the book. In some embodiments, complementary paired proteins (eg, first monomeric Fc polypeptide and second monomeric Fc polypeptide) are produced in approximately equimolar ratios (eg, approximately 1: 1 ratio). In other embodiments, complementary paired proteins (eg, first monomeric Fc polypeptide and second monomeric Fc polypeptide) are produced in non-equal molar ratios (eg, not approximately 1: 1 ratio). In other words, using the methods described herein, for example, but not limited to 100: 1, 95: 1, 90: 1, 85: 1, 80: 1, 75: 1, 70: 1, 65: 1, 60: 1, 55: 1, 50: 1, 45: 1, 40: 1, 35: 1, 30: 1, 25: 1, 20: 1, 15: 1 10: 1, 9: 1. , 8: 1, 7: 1, 6: 1, 5, 1, 4: 1, 3: 1, 2: 1, 1: 2, 1: 3, 1: 4, 1: 5, 1: 6, 1, : 7, 1: 8, 1: 9, 1:10, 1:15, 1:20, 1:25, 1:30, 1:35, 1:40, 1:45, 1:50, 1:55 , 1:60, 1:65, 1:70, 1:75, 1:80, 1:85, 1:90, 1:95, 1: 100 and the like with the first polypeptide and the second polypeptide. The ratio of can be obtained.

A polynucleotide (or expression vector) encoding a specific amount of polypeptide can be expressed intracellularly to produce the desired amount of polypeptide. In some embodiments, the amount of the first polynucleotide (or first expression vector) encoding the first monomeric Fc polypeptide and the amount introduced into the cell (eg, transfected, electroporated). , Transduced, etc.) The amount of polynucleotide (or expression vector) encoding the second monomeric Fc polypeptide is the same. For example, the first polynucleotide and the second polynucleotide can be introduced into cells in a ratio of approximately 1: 1. In other embodiments, the amount of the first polynucleotide (or first expression vector) encoding the first monomeric Fc polypeptide and the second encoding the second monomeric Fc polypeptide introduced into the cell. The amount of polynucleotide (or expression vector) varies. For example, the first and second polynucleotides are, for example, but not limited to, 50: 1, 45: 1, 40: 1, 35: 1, 30: 1, 25: 1, 20: 1, 15 :. 1 10: 1, 9: 1, 8: 1, 7: 1, 6: 1, 5: 1, 4: 1, 3: 1, 2: 1, 1: 2, 1: 3, 1: 4, 1 : 5, 1: 6, 1: 7, 1: 8, 1: 9, 1:10, 1:15, 1:20, 1:25, 1:30, 1:35, 1:40, 1:45 , 1:50, etc., can be introduced into cells.

An expression vector of a polypeptide can include one or more components that allow cells to produce the polypeptide in the desired ratio (eg, promoters, regulators, enhancers, etc.). In some cases, the first expression vector of the first monomeric Fc polypeptide contains components that increase the expression level of the vector as compared to the expression level of the second expression vector of the second polypeptide. In other cases, the second expression vector of the second monomeric Fc polypeptide is a component that increases the expression level of the vector as compared to the expression level of the first expression vector of the first monomeric Fc polypeptide. including. In certain cases, the first expression vector of the first monomeric Fc polypeptide is configured such that the expression level of the vector is the same as the expression level of the second expression vector of the second monomeric Fc polypeptide. Contains ingredients.

In some cases, the nucleic acids described herein provide, for example, the production of conditionally effective proteins of the bispecificity of the present disclosure in mammalian cells. The nucleotide sequences encoding the first and / or second polypeptides of the present disclosure can be operably linked to transcriptional control elements such as promoters and enhancers.

Suitable promoter and enhancer elements are known in the art. For expression in bacterial cells, suitable promoters include, but are not limited to, lacI, lacZ, T3, T7, gpt, λP, and trc. For expression in eukaryotic cells, suitable promoters include light and / or heavy chain immunoglobulin gene promoters and enhancer elements; cytomegalovirus earliest promoters; simple herpesvirus thymidine kinase promoters; early and late SV40 promoters; retro Promoters present in the terminal repeat sequence from the virus; EF-1a, mouse metallothionein-I promoter; and various tissue-specific promoters known in the art are included, but not limited to.

Nucleic acid or nucleotide sequences encoding proteins, such as the prodrug constructs described herein, can be present in expression and / or cloning vectors. If a protein, eg, a prodrug construct, contains two separate polypeptides, the nucleotide sequences encoding the two polypeptides can be cloned in the same or separate vectors. The expression vector may contain selectable markers, the origin of replication, and other features that result in replication and / or maintenance of the vector. Suitable expression vectors include, for example, plasmids, viral vectors and the like.

Expression vectors generally have a convenient restriction site located near the promoter sequence to result in the insertion of a nucleic acid sequence encoding a heterologous protein. There may be selectable markers that function in the expression host. Suitable expression vectors include viral vectors (eg, viral vectors based on vaccinia virus, poliovirus, adenovirus (eg, Li et al., Invest Opthalmol Vis Sci 35: 2543 2549, 1994, Borras et al., Gene Ther). 6: 515 524,1999, Virus Davidson, PNAS 92: 7700 7704, 1995, Sakamoto et al., H Gene The 5: 1088 1097, 1999, WO94 / 12649, WO93 / 03769, WO93 / 19191, WO94 / 28 WO95 / 111984, and WO95 / 00655); Adeno-related viruses (eg, Ali et al., Hum Gene Ther 9: 81 86, 1998, French et al., PNAS 94: 6916 6921, 1997, Bennett et al. , Invest Opthermol Vis Sci 38: 2857 2863, 1997, Jomaly et al., Gene Ther 4: 683 690, 1997, Rolling et al., Hum Gene Ther 10: 641 648, 1999, Ali et al. : 591 594, 1996, Srivastava in WO93 / 09239, Samulski et al., J. Vir. (1989) 63: 3822-3828, Mendelson et al., Virus. (1988) 166: 154-165, and Flotte et al. , PNAS (1993) 90: 10613-10617); SV40; Simple herpesvirus; Human immunodeficiency virus (eg, Miyoshi et al., PNAS 94: 10319 23, 1997, Takahashi et al., J Virus 73: 7812 7816, 1999); Retrovirus vectors (eg, mouse leukemia virus, splenic necrosis virus, and retrovirus-derived vectors, such as Laus sarcoma virus, Harvey Memoroma virus, avian sarcoma virus, human immunodeficiency virus, myeloproliferative sarcoma virus, and breast tumor virus), but are not limited thereto.

The present disclosure provides mammalian cells that are modified to produce proteins, eg, the prodrug constructs of the present disclosure. The polynucleotides described herein can be introduced into mammalian cells using any method known to those of skill in the art, such as, but not limited to, transfection, electroporation, viral infection, and the like.

Suitable mammalian cells include primary cells and immortalized cell lines. Suitable mammalian cell lines include human cell lines, non-human primate cell lines, rodent (eg, mouse, rat) cell lines and the like. Suitable mammalian cell lines include HeLa cells (eg, American Type Culture Collection (ATCC) No. CCL-2), CHO cells (eg, ATCC Nos. CRL9618, CCL61, CRL9096), 293 cells (eg, ATCC No. CRL-1573), Vero cells, NIH 3T3 cells (eg, ATCC No. CRL-1658), Huh-7 cells, BHK cells (eg, ATCC No. CCL10), PC12 cells (ATCC No. CRL1721), COS cells, COS-7 cells (ATCC No. CRL1651), RAT1 cells, mouse L cells (ATCC No. CCLI.3), human fetal kidney (HEK) cells (ATCC No. CRL1573), HEK293 cells, expi293 cells, HLHepG2 cells, Examples include, but are not limited to, Hut-78, Jurkat, HL-60, NK cell lines (eg, NKL, NK92, and YTS). Suitable host cells for cloning or expression of the target protein coding vector include prokaryotic or eukaryotic cells described herein.

For expression of polypeptides in bacteria, see, for example, US Pat. Nos. 5,648,237, 5,789,199, and 5,840,523. (Charlton, Methods in Molecular Biology, Vol. 248 (BKC Lo, ed., Humana Press, Totowa, N.J., 2003), which describes the expression of antibody fragments in E. coli), See also pp.245-254). After expression, the Fc fusion protein may be isolated from the bacterial cell paste in the soluble fraction and can be further purified.

In addition to prokaryotes, eukaryotic microorganisms such as filamentous fungi or yeast are suitable cloning or expression hosts, their glycosylation pathways are "humanized" and have a partial or complete human glycosylation pattern. Includes fungi and yeast strains that result in the production of antibodies. For example, Gerngross, Nat Biotech, 2004, 22: 1409-1414 and Li et al. , Nat Biotechnology, 2006, 24: 210-215.

Plant cell cultures can also be utilized as hosts. See U.S. Pat. Nos. 5,959,177, 6,040,498, 6,420,548, 7,125,978, and 6,417,429. ..

Suitable host cells for the expression of glycosylated proteins are also derived from multicellular organisms (invertebrates and vertebrates). Examples of invertebrate cells include plant cells and insect cells. Many baculovirus strains have been identified and they can be used in conjunction with insect cells, especially for transfection of Prodenia frugiperda cells.

In some embodiments, the host cell or stable host cell line is selected according to the amount of polypeptide produced and secreted by the cell. A host cell or stable host cell line can produce and secrete the prodrug compositions described herein. In some cases, suitable cells are any one of the first polypeptides and any of the second polypeptides described herein in an equimolar ratio (eg, approximately 1: 1 ratio). Can produce one or one. In other embodiments, the preferred cells are those of any one of the first monomeric Fc polypeptides and the second monomeric Fc polypeptide in an unequal molar ratio (eg, a ratio different from 1: 1). Produces any one of them.

V. Illustrative Forms of the Invention As will be appreciated by those skilled in the art, heterodimer protein compositions comprising the two monomers forming the prodrug composition can take a wide variety of forms. Importantly, the variable active domain and the variable active domain were each associated with sdABD-TTA after cleavage. That is, generally one sdABD-TTA is ligated to the active variable heavy domain via a non-cleavable domain linker and one sdABD-TTA is ligated to an active variable light domain via a non-cleavable domain linker. This ensures that active CD3 ABD can form on the surface of tumor cells. When cleavage occurs and the inactive VH and VL dissociate, the aVH and aVL associate between the molecules to form one or more active CD3 ABDs.

For all of the constructs and forms provided herein, many different components such as sdABD-TTA, cleavage sites, aVH and aVL domains, iVH and iVL domains, and Fc domains, such as all shown in FIG. , Can be "mixed and matched" in each form.

GFP-domain linker (hinge linker) -TTA-domain linker for Fc holes-anti-CD3 pseudo-Fv domain (eg, active VL domain-cleavable linker-inactive VH domain) -antigen binding domain (from N-terminus to C-terminus) A first monomeric Fc polypeptide comprising ABD against, and a TTA-domain linker-anti-CD3 pseudo-Fv domain (eg, active VH domain-cleavable linker-inactive VL domain) -domain (from N-terminus to C-terminus). A heterodimeric Fc fusion prodrug protein (see FIG. 1) comprising a second monomeric Fc polypeptide comprising an ABD for a linker (hinge linker) -Fc knob is provided herein. In some embodiments, the first monomeric Fc polypeptide (from N-terminus to C-terminus) has a TTA-domain linker-anti-CD3 pseudo-Fv domain (eg, activity) against GFP-domain linker (hinge linker) -Fc hole The second monomeric Fc polypeptide comprises ABD for the VH domain-cleavable linker-inactive VL domain) -antigen binding domain and the second monomeric Fc polypeptide (from N-terminus to C-terminus) is the TTA-domain linker-anti-CD3 pseudoFv domain (eg Includes ABD for active VL domain-cleavage linker-inactive VH domain) -domain linker (hinge linker) -Fc knob. In some embodiments, the C-terminus of the first monomeric Fc polypeptide comprises, but is not limited to, a tag such as (His) 10 tag or Strept-II tag, which is generally administered to the patient in practice. Not used for prodrug molecules. In some embodiments, the C-terminus of the second monomeric Fc polypeptide comprises, but is not limited to, a tag such as (His) 10 tag or Strept-II tag. In some embodiments, the prodrug construct comprises a first monomeric Fc comprising an sdABD (TTA) -NCL-active VL-CL-VHi-sdABD-NCL-Fc region comprising CH2-CH3 in Hall form and a knob. Includes a second monomer Fc containing an sdABD (TTA) -NCL-active VH-CL-VLi-NCL-Fc region containing CH2-CH3 in the form. In some embodiments, such heterodimer Fc fusion prodrug proteins include Pro37 and Pro36, as shown in FIG. The amino acid sequence of Pro37 is shown in FIG. 14A. The amino acid sequence of Pro37 is shown in FIG. 14A.

Heterodimer Fc fusion prodrug proteins (see FIG. 2) are also provided herein (from N-terminus to C-terminus) of the TTA domain linker-anti-CD3 pseudoFv domain (eg, active VL domain-). Cleavable linker-inactive VH domain) -domain linker (hinge linker) -a first monomeric Fc polypeptide containing ABD for the Fc hole and a TTA-domain linker-anti-CD3 pseudoFv domain (eg, active VH domain-). A second monomeric Fc polypeptide, which comprises an ABD for a cleaving linker-inert VL domain) -domain linker (hinge linker) -Fc knob. In some embodiments, the first monomeric Fc polypeptide (from N-terminus to C-terminus) is a TTA-domain linker anti-CD3 pseudoFv domain (eg, active VH domain-cleavable linker-inactive VL domain)-. The second monomeric Fc polypeptide, which comprises an ABD for a domain linker (hinge linker) -Fc hole, is a TTA-domain linker-anti-CD3 pseudo-Fv domain (eg, active VL domain-cleavable) (from N-terminus to C-terminus). Includes ABD for linker-inert VH domain) -domain linker (hinge linker) -Fc knob. In some embodiments, the C-terminus of the first monomeric Fc polypeptide comprises, but is not limited to, a tag such as (His) 10 tag or Strept-II tag. In some embodiments, the C-terminus of the second monomeric Fc polypeptide comprises, but is not limited to, a tag such as (His) 10 tag or Strept-II tag. In some embodiments, such heterodimer Fc fusion prodrug proteins include Pro38 and Pro36, as shown in FIG. The amino acid sequence of Pro36 is shown in FIG. 14A. The amino acid sequence of Pro38 is shown in FIG. 14B.

With a first monomeric Fc polypeptide comprising ABD for the TTA-domain linker-anti-CD3 pseudo-Fv domain (eg, active VL domain-cleavable linker-inactive VH domain) -Fc hole (from N-terminus to C-terminus) , TTA-domain linker-anti-CD3 pseudo-Fv domain (eg, active VH domain-cleavable linker-inactive VL domain) -ABD for Fc knob (from N-terminus to C-terminus), second monomeric Fc polypeptide And, a heterodimer Fc fusion prodrug protein (see FIG. 4), is provided herein. In some embodiments, the first monomeric Fc polypeptide (from N-terminus to C-terminus) is a TTA-domain linker anti-CD3 pseudo-Fv domain (eg, active VH domain-cleavable linker-inactive VL domain)-. The second monomeric Fc polypeptide, which comprises ABD for the Fc hole, is a TTA-domain linker-anti-CD3 pseudo-Fv domain (eg, active VL domain-cleavable linker-inactive VH domain) (from N-terminus to C-terminus). -Includes ABD for the Fc knob. In some embodiments, the C-terminus of the first monomeric Fc polypeptide comprises, but is not limited to, a tag such as (His) 10 tag or Strept-II tag. In some embodiments, the C-terminus of the second monomeric Fc polypeptide comprises, but is not limited to, a tag such as (His) 10 tag or Strept-II tag. In some embodiments, such heterodimer Fc fusion prodrug proteins include Pro68 and Pro67, as shown in FIG. Pro68 is similar to Pro37 but does not contain a domain linker attached to the sdABD or CH2 domain that binds to GFP. Pro67 is similar to Pro36 but does not contain a domain linker attached to the CH2 domain. The amino acid sequence of Pro68 is shown in FIG. 14C. The amino acid sequence of Pro67 is shown in FIG. 14B.

A first monomeric Fc polypeptide containing an ABD for the TTA-Fc hole (from N-terminus to C-terminus) and a TTA-domain linker-anti-CD3 pseudo-Fv domain (eg, active VL domain) from N-terminus to C-terminus. Includes TTA-domain linker-anti-CD3 pseudo-Fv domain (eg, active VH domain-cleavable linker-inactive VL domain) -Fc knob-cleavable linker-ABD for ABD , A second monomeric Fc polypeptide, and a heterodimeric Fc fusion prodrug protein (see FIG. 5) are provided herein. In some embodiments, the first monomeric Fc polypeptide comprises ABD for the TTA-Fc hole (from N-terminus to C-terminus) and the second monomer Fc polypeptide (from N-terminus to C-terminus). TTA-domain linker-anti-CD3 pseudo-Fv domain (eg, active VH domain-cleavable linker-inactive) against TTA-domain linker-anti-CD3 pseudo-Fv domain (eg, active VL domain-cleavable linker-inactive VH domain) VL domain) -Fc knob-Cleaving linker-Includes ABD for ABD. In some embodiments, the C-terminus of the first monomeric Fc polypeptide comprises a tag such as a (His) 10 tag or a Strept-II tag. In some embodiments, the C-terminus of the second monomeric Fc polypeptide comprises, but is not limited to, a tag such as (His) 10 tag or Strept-II tag. In some embodiments, such heterodimer Fc fusion prodrug proteins include Pro69 and Pro70, as shown in FIG. Pro70 is similar to Pro67, with the Pro9 construct attached to the C-terminus. The amino acid sequence of Pro69 is shown in FIG. 14C. The amino acid sequence of Pro70 is shown in FIG. 14D.

Includes TTA-Fchole for TTA-domain linker-anti-CD3 pseudoFv domain (eg, active VL domain-cleavable linker-inactive VH domain) -ABD for cleaving linker-ABD (from N-terminus to C-terminus). Includes a first monomeric Fc polypeptide and ABD for the TTA-domain linker-anti-CD3 pseudo-Fv domain (eg, active VH domain-cleavable linker-inactive VL domain) -Fc knob (from N-terminus to C-terminus) , A second monomeric Fc polypeptide, and a heterodimeric Fc fusion prodrug protein (see FIG. 6) are provided herein. In some embodiments, the first monomeric Fc polypeptide is relative to the TTA-domain linker anti-CD3 pseudo-Fv domain (eg, active VH domain-cleavable linker-inactive VL domain) (from N-terminus to C-terminus). The second monomeric Fc polypeptide comprises ABD for TTA-Fc hole-cleavable linker-ABD and the second monomeric Fc polypeptide (from N-terminus to C-terminus) is a TTA-domain linker-anti-CD3 pseudoFv domain (eg, active VL domain-cleavage). Includes ABD for sex linker-inert VH domain) -Fc knob. In some embodiments, the C-terminus of the first monomeric Fc polypeptide comprises, but is not limited to, a tag such as (His) 10 tag or Strept-II tag. In some embodiments, the C-terminus of the second monomeric Fc polypeptide comprises, but is not limited to, a tag such as (His) 10 tag or Strept-II tag. In some embodiments, such heterodimer Fc fusion prodrug proteins include Pro71 and Pro67, as shown in FIG. Pro71 is similar to Pro69, with the Pro9 construct attached to the C-terminus. Pro67 is similar to Pro36, but lacks the domain linker attached to the CH2 domain. The amino acid sequence of Pro71 is shown in FIG. 14D. The amino acid sequence of Pro67 is shown in FIG. 14B.

TTA-domain linker-restraint anti-CD3Fv domain (eg, active VH domain-non-cleavage restraint linker (NCCL) -active VL domain) -cleavable linker-containing ABD for the Fc knob (from N-terminus to C-terminus). Containing a monomeric Fc polypeptide of 1 and an anti-CD3 pseudoFv domain (eg, inactive VL domain-non-cleaving linker-inactive VH domain) -cleavable linker-Fc hole (from N-terminus to C-terminus). A heterodimeric Fc fusion prodrug protein (see FIG. 8) comprising two monomeric Fc polypeptides is provided herein. In some embodiments, the first monomeric Fc polypeptide (from N-terminus to C-terminus) is a TTA-domain linker-restraint anti-CD3Fv domain (eg, active VL domain-non-cleavage restraint linker (NCCL) -activity. VH domain) -Cleaving linker-Contains ABD for the Fc knob, and the second monomeric Fc polypeptide is an anti-CD3 pseudo-Fv domain (eg, inactive VH domain-non-cleaving linker-) (from N-terminus to C-terminus). Includes inactive VL domain) -cleavable linker-Fc hole. In some embodiments, the C-terminus of the first monomeric Fc polypeptide comprises, but is not limited to, a tag such as (His) 10 tag or Strept-II tag. In some embodiments, the C-terminus of the second monomeric Fc polypeptide comprises, but is not limited to, a tag such as (His) 10 tag or Strept-II tag. In some embodiments, such heterodimer Fc fusion prodrug proteins include Pro219 and Pro218, as shown in FIG. The amino acid sequence of Pro218 is shown in FIG. 14E. The amino acid sequence of Pro219 is shown in FIG. 14E.

(From N-terminus to C-terminus) TTA-Cleaving Linker-TTA-Domain Linker for Fc Knob-Restricted Anti-CD3Fv Domain (eg Active VH Domain-Non-Cleaving Constraint Linker (NCCL) -Active VL Domain) -ABD for ABD A first monomeric Fc polypeptide comprising, and an anti-CD3 pseudo-Fv domain (eg, inactive VL domain-inert linker-inert VH domain) -cleavable linker-Fc hole (from N-terminus to C-terminus). A heterodimeric Fc fusion prodrug protein (see FIG. 9) comprising a second monomeric Fc polypeptide, comprising:, is provided herein. In some embodiments, the first monomeric Fc polypeptide (from N-terminus to C-terminus) is a TTA-domain linker-restraint anti-CD3Fv domain (eg, active VL domain-non) to the TTA-cleavage linker-Fc knob. The cleaving restraint linker (NCCL) -active VH domain) -contains ABD against ABD, and the second monomeric Fc polypeptide is an anti-CD3 pseudo-Fv domain (eg, inactive VH domain-non) (from N-terminus to C-terminus). Includes cleaving linker-inactive VL domain) -cleaving linker-Fc hole. In some embodiments, the C-terminus of the first monomeric Fc polypeptide comprises, but is not limited to, a tag such as (His) 10 tag or Strept-II tag. In some embodiments, the C-terminus of the second monomeric Fc polypeptide comprises, but is not limited to, a tag such as (His) 10 tag or Strept-II tag. In some embodiments, such heterodimer Fc fusion prodrug proteins include Pro217 and Pro218, as shown in FIG. The amino acid sequence of Pro218 is shown in FIG. 14E. The amino acid sequence of Pro217 is shown in FIG. 14E.

The ABD for the TTA of the heterodimer Fc prodrug construct can be a single domain antibody that binds to the TTA. In some embodiments, the single domain antibody (sdABD) is sdABD for EGFR, sdABD for EpCAM, and sdABD for another target tumor antigen. In some embodiments of the heterodimer Fc fusion prodrug protein, the sdABD of the first monomeric Fc and the sdABD of the second monomeric Fc have the same or substantially the same amino acid sequence. In some embodiments, the sdABD of the first monomer Fc and the sdABD of the second monomer Fc bind to the same TTA. In other embodiments, the sdABD of the first monomer Fc and the sdABD of the second monomer Fc bind to different TTAs. In other cases, the sdABD of the first monomer Fc and the sdABD of the second monomer Fc have different amino acid sequences. In some embodiments, the sdABD has a variable domain set forth in CDR and / or SEQ ID NO: 14 or construct I-1 of FIG. 13A. In some embodiments, the sdABD has a variable domain set forth in CDR and / or SEQ ID NO: 15 or construct I-2 of FIG. 13A.

In some embodiments of the CD3 binding domain, active VH and active VL have the sequences of constructs I-3 and I-4 of FIG. 13A, and VHi and VLi have the sequences of constructs I-6 and I-5. Has. In some cases, the pseudoFv domain of the first monomeric Fc protein is VL linked using either a cleaving linker, VL-linker-VHi (from N-terminus to C-terminus) or VHi-linker-VL. And VHi may be included. In some embodiments, the pseudoFv domain is (from N-terminus to C-terminus) vrFR1-vr1-vr1-vrFR2-vrCDR2-vrFR3-vrCDR3-vrFR4-CL-vhiFR1-vhiCDR1-vhiFR2-vhiCDR2-vhiFR3-vhiCDR3-vhiFR4. Has a structure. In other cases, the pseudoFv domain has the structure of vhiFR1-vhiCDR1-vhiFR2-vhiCDR2-vhiFR3-vhiCDR3-vhiFR4-CL-vrFR1-vrCDR1-vrFR2-vrCDR2-vrFR3-vrCDR3-vrFR4 (from N-terminus to C-terminus). Have. In some embodiments, the pseudoFv domain of the second monomeric Fc protein is linked using a cleaving linker, VH-linker-VLi or VLi-linker-VH (from N-terminus to C-terminus). And VLi may be included. In some cases, the pseudoFv domain has the structure of vhFR1-vhCDR1-vhFR2-vhCDR2-vhFR3-vhCDR3-vhFR4-CL-vliFR1-vliCDR1-vliFR2-vliCDR2-vliFR3-vliCDR3-vliFR4 (from N-terminus to C-terminus). .. In other cases, the pseudoFv domain has the structure of vliFR1-vliCDR1-vliFR2-vliCDR2-vliFR3-vliCDR3-vliFR4-CL-vhFR1-vhCDR1-vhFR2-vhCDR2-vhFR3-vhCDR3-vhFR4 (from N-terminus to C-terminus). Have.

In some embodiments, the present invention provides covalently linked active VH and VL domains using constraint linkers, which can be cleaved or non-cleavable as outlined herein. Provide a binding Fv domain, including. The restraint linker prevents intramolecular association between VH and VL in the absence of cleavage. Thus, the constrained Fv domain contains a set of 6 CDRs contained within a variable domain, the VH vhCDR1, vhCDR2, and vhCDR3 bind to human CD3, and the VL lvCDR1, lvCDR2, and lvCDR3 are human CD3. However, in the prodrug form (eg, uncleaved), VH and VL cannot sterically associate to form an active binding domain.

The constrained Fv domain may include active VH and active VL (VHa and VLa) or inactive VH and VL (VHi and VLi). As will be appreciated by those skilled in the art, the order of VH and VL within the restraining active Fv domain can be either VH-linker-VL or VL-linker-VH (from N-terminus to C-terminus). As outlined herein, the binding active Fv domain is ligated using a non-cleavable linker, such as that shown as Pro219 in FIGS. 8 and 14E, or as Pro217 in FIGS. 9 and 14E. VH and VL can be included. In this embodiment, the constrained Fv domain has vhFR1-vhCDR1-vhFR2-vhCDR2-vhFR3-vhCDR3-vhFR4-CCL-vrFR1-vrCDR1-vrFR2-vrCDR2-vrFR3-vrCDR3-vrFR4 (from N-terminus to C-terminus). In this embodiment, the CDR and / or variable domain is that of constructs I-3 and I-4 of FIG. 13A.

A. Example 1: Construction and Purification of Pro Constructs Transfection Each construct pair was expressed from a separate expression vector (pcdna 3.4 derivative). Equal amounts of plasmid DNA encoding a pair of semi-COBRA were mixed and transfected into Expi293 cells according to the manufacturer's transfection protocol. Conditional medium was collected 5 days after transfection by centrifugation (6000 rpm x 25') and filtration (0.2 uM filter). Protein expression was confirmed by SDS-PAGE. The construct was purified and the final buffer composition was 25 mM citrate, 75 mM arginine, 75 mM NaCl, 4% sucrose, pH 7. The final preparation was stored at −80 ° C.

Protease cleavage EK
Recombinant human enterokinase (R & D Systems Catalog No. 1585-SE-010) was used to cleave the heterodimer Fc prodrug protein described herein. The recombinant protease was activated according to the manufacturer's procedure and prepared at a stock concentration of approximately 100 mM.

Test samples (Pro36 + 37, Pro36 + 38, Pro67 + 68, Pro69 + 70, Pro67 + 71, Pro217 + 218, and Pro218 + 219) were buffered, replaced with HEPES buffered physiological saline containing calcium chloride (25 mM HEPES, 50 mM NaCl, 2 mM CaCl ₂ ) overnight at room temperature. Incubated with the appropriate protease at a final concentration of 10 nM. Cleavage was confirmed by SDS-PAGE.

MMP-9
Activation of MMP9: Recombinant human MMP9 was activated according to the following protocol. Recombinant human MMP-9 (R & D # 911-MP-010) is 0.44 mg / mL (4.7 uM). P-Aminophenylmercury acetate (APMA) (Sigma) is prepared at a stock concentration of 100 mM in DMSO. The assay buffer was 50 mM Tris (pH 7.5), 10 mM CaCl2, 150 mM NaCl, 0.05% Brij-35.

-Dilute rhMMP9 with assay buffer to approximately 100 ug / ml (25 ul hMMP9 + 75 uL assay buffer)

P-Aminophenylmercury acetate (APMA) from a stock of 100 mM in −DMSO is added to a final concentration of 1 mM (1 uL-100 uL).

Incubate at −37 ° C. for 24 hours

-Dilute MMP9 to 10 ng / ul (add 900 ul assay buffer to 100 ul of activation solution)

The concentration of activated rhMMP9 is about 100 nM.

Cleavage of constructs for TDCC assay To cleave constructs, at a concentration of 1 mg / mL (10.5 uM) in formulation buffer (25 mM citric acid, 75 mM L-arginine, 75 mM NaCl, 4% sucrose). A maximum of 10 mM CaCl2 was supplied to a 100 ul protein sample. Activated rhMMP9 was added to a concentration of 20-35 nM. Samples were incubated overnight (16-20 hours) at room temperature. Cleavage integrity was verified using SDS PAGE (10-20% TG, TG electrophoresis buffer, 200 v, 1 hour). The sample was typically 98% cleaved.

B. Example 2: T cell-dependent cytotoxic effect (TDCC) assay to test the efficacy of activated heterodimer Fc prodrug protein Firefly luciferase transduced HT-29 cells are grown to approximately 80% density. , Versene (0.48 mM EDTA in PBS-Ca-Mg). Cells were centrifuged and resuspended in TDCC medium (5% heat-inactivated FBS in RPMI 1640 containing HEPES, GlutaMax, sodium pyruvate, non-essential amino acids, and β-mercaptoethanol). Purified human Pan-T cells were thawed, centrifuged and resuspended in TDCC medium.

A co-culture of HT-29_Luc cells and T cells was added to a 384-well cell culture plate. The serially diluted COBRA was then added to the co-culture and incubated at 37 ° C. for 48 hours. Finally, an equal amount of SteadyGlo luciferase assay reagent was added to the plate and incubated for 20 minutes. Plates were read on a PerkinElmer Envision with an exposure time of 0.1 s / well. Total luminescence was recorded and the data was analyzed on GraphPad Prism 7.

Specific cytotoxicity induced when the activated (cleavage) heterodimer Fc prodrug protein engages T cells and directs cell lysis in the direction of the target positive tumor cells using the above assay. The percentage was tested.

3A and 3B show that some exemplary heterodimer Fc prodrug constructs, such as Pro36 + 37 and Pro36 + 38, presented a lack of conditioning when cleaved by cognate proteases in the TDCC assay.

7A, 7B, and 7C presented conditioning when several exemplary heterodimer Fc prodrug constructs, such as Pro67 + 68 and Pro69 + 70, were cleaved by cognate proteases in the TDCC assay. Indicates that high activity was lacking.

10A and 10B show that exemplary heterodimer Fc prodrug constructs such as Pro217 + 218 and Pro218 + 219 presented conditioning and high potency when cleaved by cognate proteases in the TDCC assay.

The prodrug constructs described herein that contain a single domain antibody against EGFR that induced killing of cancer cells upon cleavage of the protease in a manner comparable to the fusion protein are single domain antibody against EGFR and anti-CD3 Fv. Includes domain.

All cited references are expressly incorporated herein by reference in their entirety. Although certain embodiments of the present invention have been described above for purposes of illustration, it is possible that many modifications of the details can be made without departing from the invention as described in the appended claims. , Will be understood by those skilled in the art.

Claims (21)

Heterodimer protein composition,
(A) From N-terminus to C-terminus,
i) A first single domain antigen binding domain (sdABD) that binds to a first tumor target antigen (TTA) (sdABD-TTA),
ii) Domain linker and
iii)
1) A variable light domain containing vlCDR1, vlCDR2, and vlCDR3,
A first constrained Fv domain, including 2) CNCL, and 3) a variable weight domain containing vhCDR1, vhCDR2, and vhCDR3.
iv) Domain linker and
v) With the second sdABD-TTA,
vi) With the first cleaving linker,
vi) With the first monomer, including the first Fc domain,
(B) From N-terminus to C-terminus,
i)
1) False variable light domain,
A first pseudoFv domain, including a non-cleavable linker and 3) a pseudovariable heavy domain.
ii) With a second cleaving linker,
iii) Containing a second monomer, including a second Fc domain,
The first Fc domain and the second Fc domain include a knob-in-hole modification, and the first variable heavy domain and the first variable light domain can bind to human CD3. However, the constrained Fv domain does not bind to CD3, and the first variable heavy domain and the first pseudo-variable light domain associate between molecules to form an inactive Fv, and the first variable A heterodimeric protein composition in which the light domain and the first pseudovariable domain are associated between molecules to form an inactive Fv. Heterodimer protein composition,
(A) From N-terminus to C-terminus,
i) sdABD-TTA and
ii) Domain linker and
iii)
1) Variable weight domains including vhCDR1, vhCDR2, and vhCDR3,
A constrained Fv domain and a constrained Fv domain, including 2) a constrained non-cleavable linker, and 3) a variable light domain containing vlcDR1, vlcDR2, and vlcDR3.
iv) With the first cleaving linker,
v) With the first monomer, including the first Fc domain,
(B) From N-terminus to C-terminus,
i)
1) False variable heavy domain,
Pseudo-Fv domains, including 2) non-cleavable linkers, and 3) pseudo-variable light domains.
ii) With a second cleaving linker,
iii) Containing a second monomer, including a second Fc domain,
The first Fc domain and the second Fc domain include a knob-in-hole modification, and the first variable heavy domain and the first variable light domain can bind to human CD3. However, the constrained Fv domain does not bind to CD3, and the first variable heavy domain and the first pseudo-variable light domain associate between molecules to form an inactive Fv, and the first variable A heterodimeric protein composition in which the light domain and the first pseudovariable domain are associated between molecules to form an inactive Fv. Heterodimer protein composition,
(A) From N-terminus to C-terminus,
i) With the first sdABD-TTA,
ii) With the first domain linker,
iii)
1) A variable light domain containing vlCDR1, vlCDR2, and vlCDR3,
A first pseudo-Fv domain, including a first cleavable linker, and 3) a pseudo-variable heavy domain.
iv) With the first monomer, including the first Fc domain,
(B) From N-terminus to C-terminus,
i) With the second sdABD-TTA,
ii) With a second domain linker,
iii)
1) Variable weight domains including vhCDR1, vhCDR2, and vhCDR3,
A second pseudo-Fv domain, including a second cleavable linker, and 3) a pseudo-variable light domain.
iv) Containing a second monomer, including a first Fc domain,
The first Fc domain and the second Fc domain include a knob-in-hole modification, and the variable light domain of the first pseudoFv domain and the variable weight domain of the second pseudoFv domain are said. A heterodimeric protein composition that does not bind to human CD3 in the absence of cleavage in a cleaving linker. Heterodimer protein composition,
(A) From N-terminus to C-terminus,
i) With the first sdABD-TTA,
ii) The first Fc domain and
iii) With the first cleaving linker,
iv) With the second sdABD-TTA,
v) With the first domain linker,
vi)
1) A variable light domain containing vlCDR1, vlCDR2, and vlCDR3,
2) Second cleavable linker,
3) A first monomer, including a first pseudo-Fv domain, including a pseudo-variable heavy domain,
(B) From N-terminus to C-terminus,
i) With the third sdABD-TTA,
ii) With a second domain linker,
iii)
1) Variable weight domains including vhCDR1, vhCDR2, and vhCDR3,
A second pseudo-Fv domain, including a second cleavable linker, and 3) a pseudo-variable light domain.
iv) Containing a second monomer, including a second Fc domain,
The first Fc domain and the second Fc domain include a knob-in-hole modification, and the variable light domain of the first pseudoFv domain and the variable heavy domain of the second pseudoFv domain are A heterodimer protein composition that does not bind to human CD3 in the absence of cleavage in the cleaving linker. Heterodimer protein composition,
(A) From N-terminus to C-terminus,
i) With the first sdABD-TTA,
ii) With the first monomer, including the first Fc domain,
(B) From N-terminus to C-terminus,
i) With the second sdABD-TTA,
ii) Domain linker and
iii)
1) Variable weight domains including vhCDR1, vhCDR2, and vhCDR3,
A first pseudo-Fv domain, including a first cleavable linker, and 3) a pseudo-variable light domain.
iv) With the second Fc domain,
v) With a second cleaving linker,
vi) With the third sdABD-TTA,
vii)
1) A variable light domain containing vlCDR1, vlCDR2, and vlCDR3,
2) a second monomer, including a second pseudo-Fv domain, which comprises a third cleavable linker, and 3) a pseudovariable heavy domain.
The first Fc domain and the second Fc domain include a knob-in-hole modification, and the variable heavy domain of the first pseudoFv domain and the variable light domain of the second pseudoFv domain are A heterodimer protein composition that does not bind to human CD3 in the absence of cleavage in the cleaving linker. The heterodimer according to claim 1 or 2, wherein the first variable heavy domain is the N-terminal to the first variable light domain, and the pseudo-light variable domain is the N-terminal to the false variable heavy domain. Dimer protein. The heterodimer according to claim 1 or 2, wherein the first variable light domain is the N-terminal to the first variable heavy domain, and the pseudo-weight variable domain is the N-terminal to the false light variable domain. Dimer protein. The heterodimer protein according to any one of claims 1 to 7, wherein the variable heavy chain comprises the amino acid sequence of SEQ ID NO: 16 and the variable light domain comprises the amino acid sequence of SEQ ID NO: 17. The heterodimer protein composition according to any one of claims 1 to 8, wherein the pseudo-heavy domain comprises the amino acid sequence of SEQ ID NO: 18, and the pseudo-light domain comprises the amino acid sequence of SEQ ID NO: 19. The heterodimer protein composition according to any one of claims 1 to 9, wherein the TTA is selected from the group consisting of EGFR, FOLR1, H7B3, and EpCAM. The heterodimer protein composition according to any one of claims 1 and 3 to 10, wherein the first and second sdABD bind to the same TTA. The heterodimer protein composition according to any one of claims 1 and 3 to 10, wherein the first and second sdABD bind to different TTAs. The hetero according to any one of claims 1 to 12, wherein the sdABD (s) is selected from the group consisting of SEQ ID NOs: 14, 15, 21, 22, 23, 25, 155, 159, 163, and 167. Dimeric protein composition. The heterodimer protein composition according to any one of claims 1 to 13, wherein one of the first and second Fc domains has SEQ ID NO: 199 and the other has SEQ ID NO: 200. The first and / or second cleaving linker is cleaved by a human protease selected from the group consisting of MMP2, MMP9, cathepsin S, cathepsin K, cathepsin L, GranzymeB, uPA, Kallecryin7, matreptase, and thrombin. The heterodimer protein composition according to any one of claims 11 to 11. Nucleic acid composition
a) A first nucleic acid encoding the first monomer according to any one of claims 1 to 15.
b) A nucleic acid composition comprising a second nucleic acid encoding the second monomer according to any one of claims 1 to 15, respectively. An expression vector composition comprising the first and second nucleic acids according to claim 16. An expression vector composition
a) A first expression vector containing the first nucleic acid according to claim 16.
b) An expression vector composition comprising a second expression vector comprising the second nucleic acid according to claim 16. A host cell comprising the expression vector composition according to claim 17 or 18. A method for producing a heterodimer protein, wherein the host cell according to claim 19 is cultured under the condition of expressing the heterodimer protein, and the heterodimer protein is recovered. , Including, methods. A method for treating cancer, which comprises administering the heterodimer protein according to any one of claims 1 to 15.

Images