CN111315773A

CN111315773A - Conditionally active binding moieties comprising an Fc region

Info

Publication number: CN111315773A
Application number: CN201880072457.XA
Authority: CN
Inventors: R·B·迪布瑞吉; A·潘沙尔
Original assignee: Maverick Therapeutics Inc
Current assignee: Maverick Therapeutics Inc
Priority date: 2017-09-08
Filing date: 2018-09-06
Publication date: 2020-06-19
Also published as: EP3679068A2; WO2019051122A2; JP2020534811A; WO2019051122A3

Abstract

Provided herein are compositions of conditionally active binding proteins containing an Fc region, and methods for co-expressing and purifying such conditionally active binding proteins. Also provided are methods of treating cancer by administering the conditionally activated binding protein to a patient.

Description

Conditionally active binding moieties comprising an Fc region

Cross Reference to Related Applications

The present application claims the benefit of U.S. provisional application No. 62/555,999 filed on 8.9.2017 and U.S. provisional application No. 62/555,943 filed on 8.9.2017, the disclosures of which are incorporated herein by reference in their entirety.

Background

In various clinical settings, selective destruction of individual cells or specific cell types is often required. For example, the main goal of cancer therapy is to specifically destroy tumor cells while leaving healthy cells and tissues as intact as possible. One such method is by inducing an immune response against the tumor to cause immune effector cells such as Natural Killer (NK) cells or Cytotoxic T Lymphocytes (CTLs) to attack and destroy the tumor cells.

The use of intact monoclonal antibodies (mabs), which provide excellent binding specificity and affinity for tumor-associated antigens, has been successfully applied in the field of cancer therapy and diagnosis. However, the large size of intact mabs, their poor biodistribution and persistence in the blood pool limit their clinical utility. For example, intact antibodies may exhibit specific accumulation in the tumor region. In biodistribution studies, when tumors are examined accurately, the presence of uneven antibody distribution and initial accumulation in the peripheral area is noted. Due to tumor necrosis, uneven antigen distribution and increased interstitial tissue pressure, it is not possible to reach the central part of the tumor with the intact antibody construct. In contrast, smaller antibody fragments show rapid tumor localization, penetrate deeper into the tumor, and are also cleared relatively rapidly from the bloodstream.

Single chain fragments (scFv) derived from the small binding domains of the parent MAb provide better biodistribution than the intact MAb for clinical applications and can target tumor cells more efficiently. Single-chain fragments can be efficiently engineered by bacteria, however, most engineered scfvs have a monovalent structure and exhibit reduced tumor accumulation (e.g., short residence time on tumor cells) and specificity compared to their parent MAb due to the lack of avidity possessed by bivalent compounds.

Despite the favorable properties of scFv, certain features have hindered their full clinical deployment in cancer chemotherapy. Of particular note is the cross-reactivity between diseased and healthy tissue due to the targeting of these agents to cell surface receptors common to both diseased and healthy tissue. ScFv with improved therapeutic index would provide significant advances in the clinical use of these agents. The present invention provides such improved scfvs and methods of making and using the same. The improved scFv of the present invention has the unexpected benefit of overcoming the lack of avidity, as evidenced by the formation of dimeric compounds by individual units.

Disclosure of Invention

In one aspect, provided herein is a heterodimeric protein composition comprising: (a) a first monomer comprising from N-terminus to C-terminus: (i) a first antigen binding domain that binds to a first Tumor Target Antigen (TTA); (ii) a domain linker; (iii) a constrained Fv domain comprising: (1) a variable heavy domain comprising vhCDR1, vhCDR2, and vhCDR 3; (2) a constrained non-cleavable linker; and (3) a variable light domain comprising vlCDR1, vlCDR2, and vlCDR 3; (iv) a first cleavable linker; and (v) a first Fc domain; and (b) a second monomer comprising from N-terminus to C-terminus: (i) a pseudo Fv domain comprising: (1) a pseudo-variable heavy domain; (2) a non-cleavable linker; and (3) a pseudo-variable light domain; (ii) a second cleavable linker; and (iii) a second Fc domain, wherein the first and second Fc domains comprise a knob-in-hole modification, and wherein the constrained Fv domain does not bind human CD3 in the absence of cleavage at the cleavable linker. 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 cleavable linker.

In certain embodiments, the variable heavy chain comprises the amino acid sequence of SEQ ID NO 16. In some cases, the variable heavy domain comprises vhFR1-vhCDR1-vhFR2-vhCDR2-vhFR3-vhCDR3-vhFR 4.

In some embodiments, the variable light domain comprises the amino acid sequence of SEQ ID NO 17. In certain instances, the variable light domain comprises vlFR1-vlCDR1-vl FR2-vlCDR2-vlFR3-vlCDR3-vlFR 4.

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 other embodiments, the second TTA is selected from the group consisting of EGFR and EpCAM. In certain instances, the first TTA and the second TTA are the same. In some cases, the first and second TTAs 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 cleavable linker contains a cleavage site for MMP 9. In some embodiments, the first and/or second cleavable linker contains a cleavage site for a membrane-penetrating peptidase (meprin).

In various embodiments, the heterodimeric protein comprises 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 aspect, provided herein is a heterodimeric protein composition comprising: (a) a first monomer comprising from N-terminus to C-terminus: (i) a first antigen binding domain that binds to a first Tumor Target Antigen (TTA); (ii) a first domain linker; (iii) a first pseudo Fv domain comprising: (1) a variable light domain comprising vlCDR1, vlCDR2 and vlCDR 3; (2) a first cleavable linker; and (3) a pseudo variable heavy domain; and (iv) a first Fc domain; and (b) a second monomer comprising 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; (iii) a second pseudo Fv domain comprising: (1) a variable heavy domain comprising vhCDR1, vhCDR2, and vhCDR 3; (2) a second cleavable linker; and (3) a pseudo-variable light domain; and (iv) a first Fc domain; and wherein the first and second Fc domains comprise a knob-and-hole modification, and wherein the variable light domain of the first pseudo Fv domain and the variable heavy domain of the second pseudo Fv domain do not bind human CD3 in the absence of cleavage at the cleavable 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 comprises a second hinge linker at the N-terminus of the second Fc domain.

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 vlFR1-vlCDR1-vlFR2-vlCDR2-vlFR3-vlCDR3-vlFR 4.

SEQ ID NOs

14, 15, and 21-25.

In some embodiments, the first and/or second cleavable linker contains a cleavage site for MMP 9. In some embodiments, the first and/or second cleavable linker contains a cleavage site for a membrane-penetrating peptidase.

In some embodiments, the heterodimeric protein comprises 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 present invention, provided herein is a heterodimeric protein composition comprising: (a) a first monomer comprising from N-terminus to C-terminus: (i) a first antigen binding domain that binds to a first Tumor Target Antigen (TTA); and (ii) a first Fc domain; and (b) a second monomer comprising from N-terminus to C-terminus: (i) a second antigen-binding domain that binds to a second Tumor Target Antigen (TTA); (ii) a domain linker; (iii) a first pseudo Fv domain comprising: (1) a variable heavy domain comprising vhCDR1, vhCDR2, and vhCDR 3; (2) a first cleavable linker; and (3) a pseudo-variable light domain; (iv) a second Fc domain; (v) a second cleavable linker; (vi) a third antigen binding domain that binds to a third Tumor Target Antigen (TTA); and (vii) a second pseudo Fv domain comprising: (1) a variable light domain comprising vlCDR1, vlCDR2 and vlCDR 3; (2) a third cleavable linker; and (3) a pseudo variable heavy domain; wherein the first and second Fc domains comprise a knob-and-hole modification, and the variable heavy domain of the first pseudo Fv domain and the variable light domain of the second pseudo Fv domain do not bind to human CD3 in the absence of cleavage at a cleavable linker.

In certain embodiments, the variable heavy chain comprises the amino acid sequence of SEQ ID NO 16. In some cases, the variable heavy 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 each case, the variable light domain comprises vlFR1-vlCDR1-vl FR2-vlCDR2-vlFR3-vlCDR3-vlFR 4.

SEQ ID NOs

14, 15, and 21-25.

In some embodiments, the first TTA, the second TTA, and/or the third TTA are 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 instances, the third TTA is EGFR. In other cases, the third TTA is EpCAM.

In certain embodiments, the first and second TTAs or the first and third TTAs or the second and third TTAs or the first, second and third TTAs are the same. In particular embodiments, the first and second TTAs or the first and third TTAs or the second and third TTAs or the first, second and third TTAs are 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 cleavable linker contains a cleavage site for MMP 9. In other embodiments, the first, second and/or third cleavable linker contains a cleavage site for a membrane-penetrating peptidase.

In some embodiments, the heterodimeric protein comprises amino acid sequences of Pro69 and Pro70 or SEQ ID NOS: 6 and 7.

In another aspect, provided herein is a heterodimeric protein composition comprising: (a) a first monomer comprising from N-terminus to C-terminus: (i) a first antigen binding domain that binds to a first Tumor Target Antigen (TTA); (ii) a first Fc domain; (iii) a first cleavable linker; (iv) a second antigen-binding domain that binds to a second Tumor Target Antigen (TTA); (v) a first domain linker; and (vi) a first pseudo Fv domain comprising: (1) a variable light domain comprising vlCDR1, vlCDR2 and vlCDR 3; (2) a second cleavable linker; (3) a pseudo-variable heavy domain; and (b) a second monomer comprising 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; (iii) a second pseudo Fv domain comprising: (1) a variable heavy domain comprising vhCDR1, vhCDR2, and vhCDR 3; (2) a third cleavable linker; and (3) a pseudo-variable light domain; and (iv) a second Fc domain; wherein the first and second Fc domains comprise a knob-and-hole modification, and wherein the variable light domain of the first pseudo Fv domain and the variable heavy domain of the second pseudo Fv domain do not bind human CD3 in the absence of cleavage at a cleavable 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 comprises a second hinge linker at the N-terminus of the second Fc domain.

SEQ ID NOs

14, 15, and 21-25.

In some embodiments, the heterodimeric protein comprises amino acid sequences of Pro67 and Pro71 or

SEQ ID NOS

4 and 8.

Provided herein is a nucleic acid encoding a first monomer of any of the heterodimeric proteins described herein. Also provided is a nucleic acid encoding the second monomer of any of the heterodimeric proteins described herein. In addition, provided herein are expression vectors comprising a nucleic acid encoding a first monomer, expression vectors comprising a nucleic acid encoding a second monomer, or expression vectors comprising a nucleic acid encoding a first monomer and a nucleic acid encoding a second monomer. In some embodiments, provided herein is a host cell comprising any one of the expression vectors disclosed herein.

In one aspect of the invention, there is provided a method of making any of the heterodimeric proteins described herein. The method comprises culturing a host cell described herein under conditions in which the heterodimeric protein is expressed, and recovering the heterodimeric protein.

In one aspect of the invention, there is provided a method of treating cancer comprising administering to a patient any one of the heterodimeric proteins of the invention.

The heterodimeric protein compositions described herein may be referred to as prodrug compositions without being cleaved by a cognate protease.

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

The present application references international published patent application numbers WO2017/156178 filed on 3/8/2017, us provisional application number 62/305,092 filed on 3/8/2016, us provisional application number 62/555,943 filed on 9/8/2017, us provisional application number 62/555,999 filed on 9/8/2017, us provisional application number 62/583,327 filed on 11/15/2017, and us provisional application number 62/587,318 filed on 11/16/2017, the disclosures of which are incorporated herein in their entirety by reference, including the figures, legends, and definitions, and the referenced embodiments.

Drawings

Figure 1 shows an exemplary embodiment of a conditionally activated binding polypeptide in the form of "construct 1" comprising an Fc hole/knob region. Schematic of the Pro37+ Pro36 prodrug construct and the resulting bispecific polypeptide after Enterokinase (EK) cleavage are shown, but other cleavage sites, such as described herein, may also be used. The bispecific polypeptide comprises an sdABD that binds EGFR and an Fv domain that binds CD 3. It should be noted that sdabds that bind other Target Tumor Antigens (TTAs) such as, but not limited to, FOLR1, B7H3, and EpCAM can be used in other embodiments. Furthermore, figure 1 also shows the use of two different protein "tags" at the C-terminus of the Fc domain, which are used to facilitate purification of the heterodimeric proteins of the invention, but these can be removed as will be understood by those skilled in the art.

Figure 2 shows an exemplary embodiment of a conditionally activated binding polypeptide in the form of "construct 2" comprising an Fc hole/knob region. A schematic of the Pro38+ Pro36 prodrug construct is shown, again using the Flag cleavage site of EK, but many embodiments utilize other cleavage sites. It should be noted that sdabds that bind other Target Tumor Antigens (TTAs) such as, but not limited to, FOLR1, B7H3, and EpCAM can be used in other embodiments. Furthermore, figure 2 also shows the use of two different protein "tags" at the C-terminus of the Fc domain, which are used to facilitate purification of the heterodimeric proteins of the invention, but these can be removed as will be understood by those skilled in the art.

Fig. 3A-3B show that some exemplary heterodimeric Fc prodrug constructs described herein exhibit low or lack of conditionality when cleaved with a homologous protease in a TDCC assay. In fig. 3A, Pro36+37 (circles) was not pretreated with EK protease, while cleaved Pro36+37 (squares) was pretreated with EK protease. In fig. 3B, Pro36+38 (circles) was not pretreated with EK protease, while cleaved Pro36+38 (squares) was pretreated with EK protease. Pro214 is a full-length negative control (open squares) and Pro 51 (triangles) is a positive control that does not require protease cleavage for activity.

Fig. 4 shows an exemplary embodiment of a conditionally activated binding polypeptide in the form of "construct 3" comprising an Fc hole/knob region. A schematic of the Pro68+ Pro67 prodrug construct is shown along with the resulting bispecific polypeptide after Enterokinase (EK) cleavage, again using the Flag cleavage site of EK, but many embodiments utilize other cleavage sites. The bispecific polypeptide comprises an sdABD that binds EGFR and an Fv domain that binds CD 3. It should be noted that sdabds that bind other Target Tumor Antigens (TTAs) such as, but not limited to, FOLR1, B7H3, and EpCAM can be used in other embodiments. Furthermore, figure 2 also shows the use of two different protein "tags" at the C-terminus of the Fc domain, which are used to facilitate purification of the heterodimeric proteins of the invention, but these can be removed as will be understood by those skilled in the art.

Fig. 5 shows an exemplary embodiment of a conditionally activated binding polypeptide in the form of "construct 4" comprising an Fc hole/knob region. A schematic of the Pro69+ Pro70 prodrug construct is shown, again using the Flag cleavage site of EK, but many embodiments utilize other cleavage sites. It is noted that sdabds that bind other Target Tumor Antigens (TTAs) such as, but not limited to, FOLR1, H7B3, and EpCAM can be used in other embodiments. Furthermore, figure 5 also shows the use of two different protein "tags" at the C-terminus of the monomeric protein, which are used to facilitate purification of the heterodimeric proteins of the invention, but these can be removed as will be appreciated by those skilled in the art.

Fig. 6 shows an exemplary embodiment of a conditionally activated binding polypeptide in the form of "construct 5" comprising an Fc hole/knob region. A schematic of the Pro71+ Pro67 prodrug construct is shown, again using the Flag cleavage site of EK, but many embodiments utilize other cleavage sites. It is noted that sdabds that bind other Target Tumor Antigens (TTAs) such as, but not limited to, FOLR1, H7B3, and EpCAM can be used in other embodiments. Furthermore, figure 2 also shows the use of two different protein "tags" at the C-terminus of the monomeric protein, which are used to facilitate purification of the heterodimeric proteins of the invention, but these may be removed as will be appreciated by those skilled in the art.

Fig. 7A-7C show that some exemplary heterodimeric Fc prodrug constructs described herein exhibit conditionality when cleaved with a cognate protease in a TDCC assay, but lack high conditionality. In fig. 7A, Pro67+68 (circles) was not pretreated with EK protease, while cleaved Pro67+68 (squares) was pretreated with EK protease. In fig. 7B, Pro69+70 (circles) was not pretreated with EK protease, while cleaved Pro69+70 (squares) was pretreated with EK protease. In fig. 7C, Pro67+71 (circles) was not pretreated with EK protease, while cleaved Pro67+71 (squares) was pretreated with EK protease. Pro214 is a full-length negative control (open squares) and Pro 51 (triangles) is a positive control that does not require protease cleavage for activity.

Fig. 8 shows an exemplary embodiment of a conditionally activated binding polypeptide in the form of "construct 6" comprising an Fc hole/knob region. A schematic of a Pro219+ Pro218 prodrug construct is shown, using MMP9 protease cleavage sites, but other cleavage sites as described herein may also be used. It is noted that sdabds that bind other Target Tumor Antigens (TTAs) such as, but not limited to, FOLR1, H7B3, and EpCAM can be used in other embodiments. Furthermore, figure 8 also shows the use of two different protein "tags" at the C-terminus of the monomeric protein, which are used to facilitate purification of the heterodimeric proteins of the invention, but these can be removed as will be understood by those skilled in the art.

Fig. 9 shows an exemplary embodiment of a conditionally activated binding polypeptide in the form of "construct 7" comprising an Fc hole/knob region. A schematic of a Pro217+ Pro218 prodrug construct is shown, using MMP9 protease cleavage sites, but other cleavage sites as described herein may also be used. It should be noted that other Target Tumor Antigens (TTAs), such as but not limited to, sdabds of EpCAM, can be used in other embodiments. In addition, figure 9 also shows the use of two different protein "tags" at the C-terminus of the monomeric protein, which are used to facilitate purification of the heterodimeric proteins of the invention, but these can be removed as will be understood by those skilled in the art.

Fig. 10A-10B show that the exemplary heterodimeric Fc prodrug constructs described herein exhibit conditional and high potency in a TDCC assay upon cleavage with a cognate protease. In fig. 10A, Pro217+218 (circles) was not pretreated with EK protease, while cleaved Pro217+218 (squares) was pretreated with EK protease. In fig. 10B, Pro218+219 (circles) was not pretreated with EK protease, while cleaved Pro218+219 (squares) was pretreated with EK protease. Pro214 is a full-length negative control (open squares) and Pro 51 (triangles) is a positive control that does not require protease cleavage for activity.

Fig. 11A-11C depict a number of suitable protease cleavage sites. As will be appreciated by those skilled in the art, these cleavage sites may serve as cleavable linkers. In some embodiments, for example, where a more flexible cleavable linker is desired, additional amino acids (typically glycine and serine) may be present at the N-terminus or C-terminus of these cleavage sites.

Figure 12 depicts a number of suitable scFv linkers or domain linkers.

Fig. 13A-13G depict a number of sequences of the present invention. For the antigen binding domain, the CDRs are bold underlined.

Fig. 14A to 14G depict the sequence of the present invention. Linkers are underlined, wherein cleavable linkers are double underlined. CDRs are bold underlined. A slash ("/") depicts a structural domain spacer. C-terminal tags such as Maltose Binding Protein (MBP), (His)10 and

the II tag is in bold, but is optional as described herein, depending on the purification scheme used. Thus, the sequence of figure 14 excluding the C-terminal tag is included within the description herein.

Fig. 15A-15C depict additional Pro219 constructs (fig. 15A) and additional Pro217 constructs (fig. 15B and 15C).

Fig. 16A to 16G depict further sequences of the present invention. For the antigen binding domain, the CDRs are bold underlined. "/" indicates the intersection of the domains, the domain linker is underlined, and the cleavable linker is underlined and italicized. Many constructs contain a histidine tag, which is optional depending on the use.

Detailed Description

I.Introduction to the design reside in

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, and therefore only require the binding ability to activate therapeutic molecules in the vicinity of diseased tissue to avoid off-target interactions. Accordingly, the present invention relates to multivalent conditionally active ("MCE") proteins with many functional protein domains. Generally, one of these domains is an Antigen Binding Domain (ABD) that binds to a Target Tumor Antigen (TTA). The other domain is an ABD that binds a T-cell antigen (such as CD3) under certain conditions, such as when a portion of the ABD is in close proximity to a complementary portion of the ABD to form an anti-CD 3Fv binding domain. That is, the therapeutic molecule is prepared in a "prodrug" -like form in which the CD3 binding domain is inactive until exposed to the tumor environment. To achieve this conditionality, the present invention utilizes "dummy" or "inactive" or "inert" variable domains in several different ways depending on the form, as described herein and shown in the drawings. These are referred to herein as "iVH" and "iVL" domains.

In some embodiments of the invention, the CD3 binding domain ("CD 3 Fv") is in a constrained form in which the linker between the variable heavy and variable light domains that traditionally form the Fv is too short to allow the two domains to bind to each other. In some embodiments, the prodrug polypeptide further comprises a "pseudo Fv domain" in a prodrug (e.g., uncleaved) form. The pseudo Fv domain may comprise a variable heavy domain having standard framework regions but "inert" or "virtual" CDRs (inactive variable heavy domain), a variable light domain having standard framework regions but "inert" or "virtual" CDRs (inactive variable light domain), or both. Thus, the constrained Fv domain binds to the pseudo Fv domain due to the affinity of each framework region. However, due to the "inert" CDRs of the pseudodomains, the resulting ABD will not bind CD3, thereby preventing off-target toxicity. However, in the presence of proteases in or near the tumor, the prodrug construct is cleaved in a manner so as to allow association of the "authentic" variable heavy and variable light domains, thereby triggering active CD3 binding and resulting tumor efficacy.

In other embodiments, in prodrug forms, the prodrug polypeptide comprises two pseudo-Fv domains and one Fc domain attached to each pseudo-Fv domain. The first pseudo Fv domain can comprise an inactive variable heavy domain and an inactive variable light domain, and the second pseudo Fv domain can comprise an active variable heavy domain and an inactive variable light domain. ABD in the prodrug form will bind CD3 in proteolytically inactive tissues. However, in or near the tumor, the protease may cleave the prodrug construct such that the active variable heavy domain and the active variable light domain may associate and bind to CD3, thereby inducing target tumor cytotoxicity.

Thus, the prodrug constructs provided herein comprise heterodimeric IgG Fc regions that form a "knob and hole" ("KIH") conformation. A detailed description of the pestle and mortar concept can be found in, for example, U.S. Pat. nos. 5,731,168 and 7,186,076; and Ridgway et al, Protein Engineering, Design and Selection,1996,9(7): 617-; merchant et al, Nat Biotechnol,1998,16: 677-; and Carter, J.immunological Methods,2001,24(1-2): 7-15. Briefly, a knob can be generated at the CH3 domain interface of a first IgG Fc chain by replacing the smaller one with the larger amino acid side chain (e.g., T366W); and the mortar can be generated in a juxtaposed position at the CH3 interface of the second IgG Fc chain by replacing the larger one with a smaller amino acid side chain (e.g., Y407V). Suitable KIH variants are described below. Thus, a CH3 domain with "one or more knob substitutions" is referred to herein as a "CH 3-knob", and a CD3 domain with "one or more hole substitutions" is referred to herein as a "CH 3-hole", the generic term being "CH 3-KIH" to encompass both, and thus, as will be understood by those of skill in the art, which "side" of an Fc dimer "contains a" knob variant "and which side contains a" knob variant "is not definitive and can vary.

As discussed herein, there are a wide variety of conformations and forms that can be used in the present invention. The conformation of the prodrug construct can take a variety of configurations such that prodrug activation can occur in several general ways, as shown in figure 1 as "construct 1", as shown in figure 2 as "construct 2", as shown in figure 4 as "construct 3", as shown in figure 5 as "construct 4", as shown in figure 6 as "construct 5", as shown in figure 8 as "construct 6", and as shown in figure 9 as "construct 7". These constructs typically rely on the formation of the Fc domain of the heterodimeric Fc structure to allow proper lytic preassociation of the inert binding domain.

In a "construct 1" embodiment, the prodrug construct comprises a first Fc polypeptide comprising a CH2-CH 3-hole polypeptide, a first pseudo Fv domain, and an Antigen Binding Domain (ABD) that can bind to a Target Tumor Antigen (TTA); and a second Fc polypeptide comprising a CH2-CH 3-knob polypeptide, a second pseudo Fv domain, and an Antigen Binding Domain (ABD) that can bind to a Target Tumor Antigen (TTA). A pseudo Fv domain refers to a CD3 binding domain (Fv) that is inactive prior to exposure 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 pseudo Fv domain comprises an active variable light domain (active VL) and an inactive variable heavy domain (inactive VH). In addition, as understood by those skilled in the art, the pseudo Fv domain of "construct 1" may be VH-linker-VL or VL-linker-VH in either direction from N-terminus to C-terminus (e.g., in fig. 1/"construct 1", the pseudo Fv domain is shown as VH-linker-VL, but this may be switched).

In some embodiments of "construct 1", the first Fc polypeptide (from N-terminus to C-terminus) comprises: an antigen-binding domain of a first TTA connected by a domain linker to an active VL domain, the active VL domain being attached by a cleavable linker to an inactive VH domain connected to a second antigen-binding domain connected to a CH2-CH3-KIH polypeptide; and the second Fc polypeptide (from N-terminus to C-terminus) comprises: an antigen binding domain of a second TTA connected by a domain linker to an active VH domain attached by a cleavable linker to an inactive VL domain connected to a CH2-CH3-KIH polypeptide. In some cases, the first Fc polypeptide comprises a CH 3-hole and the second Fc polypeptide comprises a CH 3-knob. Upon cleavage of the cleavable linker of the first Fc polypeptide and cleavage of the cleavable linker of the second Fc polypeptide at or near the tumor site, the active VL of the first Fc polypeptide and the active VH of the second Fc polypeptide may associate and trigger active CD3 binding. In addition to the innate self-assembly of active VH and VL domains, each domain is linked to an antigen binding domain, which in turn is linked to a tumor antigen. Thus, the products of protease cleavage can bind to tumor cells and recruit T cells to the tumor site. In some embodiments, the first TTA and the second TTA are the same tumor antigen. In other embodiments, the first TTA and the second TTA are different tumor antigens.

In some cases, the prodrug construct of "construct 1" has two cleavage sites: one between the active and inactive variable heavy chains of the first Fc polypeptide and the second between the active and inactive variable light chains of the second Fc polypeptide. In some embodiments, both cleavage sites are recognized and cleaved by the same protease. Thus, the two cleavage sites may have the same or substantially the same amino acid sequence. In other embodiments, the two cleavage sites are recognized and cleaved by different proteases. Thus, the two cleavage sites may have different amino acid sequences.

In some embodiments of "construct 2", the prodrug construct comprises a first Fc polypeptide comprising a CH2-CH3-KIH polypeptide, a first pseudo Fv domain comprising an active variable light chain (active VL) and an inactive variable heavy chain (inactive VH), and an Antigen Binding Domain (ABD) that can bind to a Target Tumor Antigen (TTA); and a second Fc polypeptide comprising a CH2-CH3-KIH polypeptide, a second pseudo Fv domain comprising an active variable heavy chain (active VH) and an inactive variable light chain (inactive VL), and an Antigen Binding Domain (ABD) that can bind to a Target Tumor Antigen (TTA). In some cases, the first Fc polypeptide comprises a CH 3-hole and the second Fc polypeptide comprises a CH 3-knob.

In some embodiments of "construct 2", the first Fc polypeptide (from N-terminus to C-terminus) comprises: an antigen binding domain of a first TTA connected by a domain linker to an active VL domain, the active VL domain being attached by a cleavable linker to an inactive VH domain, the inactive VH domain being connected by a domain linker to a CH2-CH3-KIH polypeptide; and the second Fc polypeptide (from N-terminus to C-terminus) comprises: an antigen binding domain of a second TTA connected by a domain linker to an active VH domain, the active VH domain being attached by a cleavable linker to an inactive VL domain, the inactive VL domain being connected by a domain linker to a CH2-CH3-KIH polypeptide. In some cases, the first Fc polypeptide comprises a CH 3-hole and the second Fc polypeptide comprises a CH 3-knob. Upon cleavage of the cleavable linker of the first Fc polypeptide and cleavage of the cleavable linker of the second Fc polypeptide at or near the tumor site, the active VL of the first Fc polypeptide and the active VH of the second Fc polypeptide may associate and trigger active CD3 binding. In addition to the innate self-assembly of active VH and VL domains, each domain is linked to an antigen binding domain, which in turn is linked to a tumor antigen. Thus, the products of protease cleavage can bind to tumor cells and recruit T cells to the tumor site. In some embodiments, the first TTA and the second TTA are the same tumor antigen. 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 lacks the domain linker between the inactive VH domain of the first Fc polypeptide and the CH 3-hole polypeptide and the domain linker between the inactive VL domain of the second Fc polypeptide and the CH 3-knob polypeptide.

Also provided herein is a prodrug construct (e.g., "construct 4") comprising a first Fc polypeptide comprising a CH2-CH3-KIH polypeptide and an Antigen Binding Domain (ABD) that can bind to a Target Tumor Antigen (TTA); and a second Fc polypeptide comprising a CH2-CH3-KIH polypeptide, a first pseudo-Fv domain, a second pseudo-Fv domain, and a second antigen-binding domain that can bind to a Target Tumor Antigen (TTA) and a third antigen-binding domain that can bind to a Target Tumor Antigen (TTA). In some cases, the first Fc polypeptide comprises a CH 3-hole and the second Fc polypeptide comprises a CH 3-knob. 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 may bind to the same tumor antigen. The first and second antigen-binding domains may bind different tumor antigens. The first and third antigen binding domains may bind to the same tumor antigen. The first and third antigen binding domains may bind different tumor antigens. The second and third antigen binding domains may bind to the same tumor antigen. The second and third antigen binding domains may bind different tumor antigens.

In a "construct 4" embodiment, the first Fc polypeptide (from N-terminus to C-terminus) comprises: a first antigen binding domain linked to TTA of a CH2-CH3-KIH polypeptide; and the second Fc polypeptide (from N-terminus to C-terminus) comprises: a second antigen-binding domain of TTA linked by a domain linker to an active VH domain, the active VH domain being attached by a cleavable linker to an inactive VL domain, the inactive VL domain being linked to a CH2-CH3-KIH polypeptide, the CH2-CH3-KIH polypeptide being linked by a cleavable linker to a third antigen-binding domain of TTA, the third antigen-binding domain being linked by a domain linker to an active VL domain, the active VL domain being linked by a cleavable linker to an inactive VH domain. In some cases, the first Fc polypeptide comprises a CH 3-hole and the second Fc polypeptide comprises a CH 3-knob.

In a "construct 5" embodiment, the first Fc polypeptide (from N-terminus to C-terminus) comprises: a first antigen-binding domain of TTA linked to a CH2-CH3-KIH polypeptide, said CH2-CH3-KIH polypeptide being linked to a second antigen-binding domain by a cleavable linker, said second antigen-binding domain being linked to an active VL domain by a domain linker, said active VL domain being linked to an inactive VH domain by a cleavable linker; and the second Fc polypeptide (from N-terminus to C-terminus) comprises: a third antigen-binding domain of TTA connected by a domain linker to an active VH domain connected by a cleavable linker to an inactive VL domain connected to a CH2-CH3-KIH polypeptide. In some cases, the first Fc polypeptide comprises a CH 3-hole and the second Fc polypeptide comprises a CH 3-knob. 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 may bind to the same tumor antigen. The first and second antigen-binding domains may bind different tumor antigens. The first and third antigen binding domains may bind to the same tumor antigen. The first and third antigen binding domains may bind different tumor antigens. The second and third antigen binding domains may bind to the same tumor antigen. The second and third antigen binding domains may bind different tumor antigens.

Provided herein is a prodrug construct (e.g., "construct 6") comprising a first Fc polypeptide comprising a CH2-CH3-KIH polypeptide and a first pseudo-Fv domain comprising a variable heavy domain and a variable light domain having standard framework regions and "inert" or "virtual" CDRs; and a second Fc polypeptide comprising a CH2-CH3-KIH polypeptide, an Antigen Binding Domain (ABD) that can bind to a Target Tumor Antigen (TTA), and a CD3 binding domain in a constrained form, wherein the linker between the variable heavy and variable light domains that traditionally forms an Fv is too short to allow the two domains to bind to each other. In some embodiments, the constrained active Fv domain is covalently attached to the CH2-CH3-KIH polypeptide through a cleavable linker, and the first pseudo Fv domain is covalently attached to the CH2-CH3-KIH polypeptide through a cleavable linker. In some cases, the first Fc polypeptide comprises a CH 3-hole and the second Fc polypeptide comprises a CH 3-knob. In some embodiments, the cleavable linker may be recognized by the same protease. In other embodiments, the cleavable linker may be recognized by a different protease.

In a "construct 6" embodiment, the second Fc polypeptide (from N-terminus to C-terminus) comprises: a first antigen-binding domain of TTA that is linked to a constrained active Fv domain via a domain linker (e.g., to an active variable heavy chain of an active variable light chain via a constrained non-cleavable linker, or to an active variable light chain of an active variable heavy chain via a constrained non-cleavable linker), which is linked to a CH2-CH3-KIH polypeptide via a cleavable linker; and the first Fc polypeptide (from N-terminus to C-terminus) comprises: a pseudo Fv domain (e.g., an inactive variable light domain connected to an inactive variable heavy domain by a non-cleavable linker, or an inactive variable heavy domain connected to an inactive variable light domain by a non-cleavable linker) connected to a CH2-CH3-KIH polypeptide by a cleavable linker or a non-cleavable linker. In some cases, the first Fc polypeptide comprises a CH 3-hole and the second Fc polypeptide comprises a CH 3-knob.

Provided herein is another prodrug construct (e.g., "construct 7"), which is similar to "construct 6". An exemplary embodiment of construct 7 comprises a second Fc polypeptide comprising a CH2-CH3-KIH polypeptide, a first antigen-binding domain (ABD) that can bind to a Target Tumor Antigen (TTA), a second antigen-binding domain (ABD) that can bind to a Target Tumor Antigen (TTA), and a CD3 binding domain in a constrained form, wherein the linker between the variable heavy and light domains that traditionally form an Fv is too short to allow the two domains to bind to each other; and a first Fc polypeptide comprising a CH2-CH3-KIH polypeptide and a first pseudo Fv domain. In some cases, the first Fc polypeptide comprises a CH 3-hole and the second Fc polypeptide comprises a CH 3-knob. In some cases, the first and second antigen-binding domains can bind to the same tumor antigen. In other cases, the first and second antigen-binding domains can bind different tumor antigens.

In the "construct 7" embodiment, the second Fc polypeptide (from N-terminus to C-terminus) comprises: a first antigen-binding domain of TTA that is linked to a constrained active Fv domain (e.g., an active variable heavy chain of an active variable light chain via a constrained non-cleavable linker, or an active variable light chain of an active variable heavy chain via a constrained non-cleavable linker) via a domain linker, which is linked to a second antigen-binding domain linked to a CH2-CH3-KIH polypeptide via a cleavable linker; and the first Fc polypeptide (from N-terminus to C-terminus) comprises: a pseudo Fv domain (e.g., an inactive variable light domain linked to an inactive variable heavy domain by a non-cleavable linker, or an inactive variable heavy domain linked to an inactive variable light domain by a non-cleavable linker) linked to a CH2-CH3-KIH polypeptide by a cleavable or non-cleavable linker. In some cases, the first Fc polypeptide comprises a CH 3-hole and the second Fc polypeptide comprises a CH 3-knob. In some embodiments, the cleavable linker adjacent to the CH2-CH 3-knob is the same cleavable linker adjacent to the CH2-CH 3-hole polypeptide. In other embodiments, the cleavable linker is different.

II.Definition of

In order that this application may be more fully understood, several definitions are set forth below. Such definitions are intended to encompass grammatical equivalents.

The term "COBRA^TM"and" conditional bispecific redirected activation "refers to a bispecific conditionally effective protein with many functional protein domains. In some embodiments, one of the functional domains is an Antigen Binding Domain (ABD) that binds a Target Tumor Antigen (TTA). In certain embodiments, the other domain is an ABD that binds to a T cell antigen under certain conditions. T cell antigens include, but are not limited to, CD 3. The term "semi-COBRA^TMBy "is meant that the variable heavy chain of a half-COBRA can interact with another half-COBRA due to innate self-assembly when concentrated on the surface of target-expressing cells^TM(complementary half-COBRA^TM) Can bind to T cell antigens when associated with variable light chainsConditionally active proteins.

As used herein, "amino acid" and "amino acid identity" mean one of the 20 naturally occurring amino acids or any non-natural analog that may be present at a particular, defined position. In many embodiments, "amino acid" means one of the 20 naturally occurring amino acids. By "protein" herein is meant at least two covalently attached amino acids, including proteins, polypeptides, oligopeptides, and peptides.

"amino acid modification" herein means amino acid substitution, insertion and/or deletion in a polypeptide sequence or alteration of a moiety chemically linked to a protein. For example, the modification may be an altered carbohydrate or PEG structure attached to the protein. For clarity, unless otherwise noted, amino acid modifications are always directed to amino acids encoded by DNA, e.g., 20 amino acids with codons in DNA and RNA. Preferred amino acid modifications herein are substitutions.

By "amino acid substitution" or "substitution" herein is meant the replacement of an amino acid at a particular position in a parent polypeptide sequence with a different amino acid. In particular, in some embodiments, a substitution refers to an amino acid that does not naturally occur at a particular position, or that does not naturally occur in an organism or any organism. For clarity, a protein that has been engineered to alter a nucleic acid coding sequence, but not the starting amino acid (e.g., exchange of CGG (encoding arginine) for CGA (still encoding arginine) to increase expression levels in a host organism) is not an "amino acid substitution"; that is, although a new gene encoding the same protein is produced, if the protein has the same amino acid at the specific position where it starts, the protein 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 a parent polypeptide sequence.

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

As outlined herein, the polypeptides of the invention specifically bind to CD3 and a Target Tumor Antigen (TTA), such as a target cell receptor. By "specifically binds" or "specific for a particular antigen or epitope" is meant binding that is significantly different from non-specific interactions. Specific binding can be measured, for example, by determining the binding of the molecule compared to the binding of a control molecule, which is typically a similarly structured molecule that 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 can be, for example, by a KD for the antigen or epitope of at least about 10^-4M, at least about 10^-5M, at least about 10^-6M, at least about 10^-7M, at least about 10^-8M, at least about 10^-9M, or at least about 10^-10M, at least about 10^-11M, at least about 10^-12M or higher, wherein KD refers to the off-rate of a particular antibody-antigen interaction. Typically, an antibody that specifically binds an antigen has a KD relative to a control molecule of 20, 50, 100, 500, 1000, 5,000, 10,000 or more fold for the antigen or epitope.

Furthermore, specific binding to a particular antigen or epitope can be exhibited, for example, by an antibody having KA or KA for the antigen or epitope that is at least 20, 50, 100, 500, 1000, 5,000, 10,000, or more fold higher than for a control, where KA or KA refers to the association rate of a particular antibody-antigen interaction. Binding affinity is typically measured using a Biacore assay or Octet as known in the art.

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

As used herein, "position" means a position in the sequence of a protein. Positions may be numbered sequentially, or according to a given format, such as the 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 series 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" means a polypeptide comprising VL and VH domains, typically from an antigen binding domain of an antibody. If the Fv domain contains active VH and VL domains, the Fv domain typically forms an "antigen binding domain" or "ABD" (but in some cases, an Fv containing a constrained linker) as discussed herein. As discussed below, Fv domains can be organized in a variety of ways in the present invention, and can be "active" or "inactive," such as in scFv formats, constrained Fv formats, pseudo Fv formats, and the like. It will be appreciated that in the present invention, in some cases, the Fv domain consists of VH and VL domains on a single polypeptide chain, such as that shown in figures 8 and 9, but with a linker constrained such that intramolecular ABD cannot be formed. In these embodiments, two active ABDs are formed after lysis. In some cases, the Fv domain consists of VH and VL domains, one of which is inert such that an intermolecular ABD is formed only upon cleavage.

By "variable region" herein is meant a region of an immunoglobulin comprising one or more Ig domains encoded by essentially any vk, V λ and/or VH genes that constitute the kappa, λ and heavy chain immunoglobulin loci, respectively. Each VH and VL is composed of three hypervariable regions ("complementarity determining regions", "CDRs") and four "framework regions" or "FRs", arranged from amino-terminus to carboxy-terminus in the following order: FR1-CDR1-FR2-CDR2-FR3-CDR3-FR 4. Thus, the VH domain has the structure vhFR1-vhCDR1-vhFR2-vhCDR2-vhFR3-vhCDR3-vhFR4 and the VL domain has the structure vlFR1-vlCDR1-vlF R2-vlCDR2-vlFR3-vlCDR3-vlFR 4. The vhFR region and vlFR region self-assemble to form an Fv domain, as described more fully herein. In general, in the prodrug forms of the invention, there are "constrained Fv domains" in which the VH and VL domains cannot self-associate, and "pseudo Fv domains" in which the CDRs do not form a functional (active) antigen binding domain when self-associated. "

The hypervariable region confers antigen-binding specificity and typically encompasses about amino acid residues 24-34(LCDR 1; "L" represents light chain), 50-56(LCDR2) and 89-97(LCDR3) in the light chain variable region and about 31-35B (HCDR 1; "H" represents heavy chain), 50-65(HCDR2) and 95-102(HCDR3) in the heavy chain variable region; kabat et al, SEQUENCES OFPROTEINS OF IMMUNOLOGICAL INTEREST, 5 th edition Public Health Service, national institutes OF Health, Bethesda, Md. (1991) and/or those residues that form hypervariable loops (e.g., residues 26-32(LCDR1), 50-52(LCDR2) and 91-96(LCDR3) in the light chain variable region and residues 26-32(HCDR1), 53-55(HCDR2) and 96-101(HCDR3) in the heavy chain variable region; Chothia and Lesk (1987) J.mol.biol.196: 901-917. the specific CDRs OF the present invention are described below.

As will be appreciated by those skilled in the art, the exact number and location of CDRs in different numbering systems may vary. However, it is to be understood that disclosure of variable heavy and/or variable light sequences includes disclosure of the relevant (inherent) CDRs. Thus, the disclosure of each variable heavy region is that of a vhCDR (e.g., vhCDR1, vhCDR2, and vhCDR3), and the disclosure of each variable light region is that of a vlCDR (e.g., vlCDR1, vlCDR2, and vlCDR 3).

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

TABLE 1

Throughout this specification, when referring to residues in the variable domain (approximately, residues 1-107 of the light chain variable region and residues 1-113 of the heavy chain variable region), the Kabat numbering system and the EU numbering system of the Fc region are typically used (e.g., Kabat et al, supra (1991)).

The present invention provides a large number of different sets of CDRs. In this case, the "full CDR set" includes three variable light CDRs and three variable heavy CDRs, e.g., vlCDR1, vlCDR2, vlCDR3, vhCDR1, vhCDR2, and vhCDR 3. As will be understood by those skilled in the art, each set of CDRs, i.e., VH and VL CDRs, can be bound to an antigen individually and as a set. For example, in a constrained Fv domain, vhcdrs can, for example, bind to CD3 and vlcdrs can bind to CD3, but in a constrained form they cannot bind to CD 3.

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

The CDRs facilitate the formation of an antigen binding site, or more particularly an epitope binding site. An "epitope" refers to a determinant that interacts with a specific antigen binding site in the variable region called the paratope. Epitopes are groups of molecules such as amino acids or sugar side chains and generally have specific structural characteristics, as well as specific charge characteristics. A single antigen may have more than one epitope.

An epitope may comprise amino acid residues that are directly involved in binding (also referred to as the immunodominant component 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; in other words, the amino acid residues are within the footprint (footprint) of the specific antigen-binding peptide.

Epitopes can be either conformational or linear. Conformational epitopes are generated by spatially juxtaposing amino acids from different segments of a linear polypeptide chain. Linear epitopes are produced by adjacent amino acid residues in a polypeptide chain. Conformational and non-conformational epitopes may differ in that: in the presence of denaturing solvents, the binding to the former is lost without losing the binding to the latter.

Epitopes typically comprise at least 3, and more typically at least 5 or 8-10 amino acids in a unique spatial conformation. Antibodies recognizing the same epitope can be validated in a simple immunoassay showing the ability of one antibody to block the binding of another antibody to the target antigen, e.g., "binning". As outlined below, the present invention includes not only the antigen binding domains and antibodies listed herein, but also those domains and antibodies that compete for binding with the epitope to which the listed antigen binding domain binds.

The variable heavy and variable light domains of the invention may be "active" or "inactive".

As used herein, "inactive VH" ("iVH") and "inactive VL" ("iVL") refer to components of pseudo-Fv domains that, when paired with their cognate VL or VH partners, respectively, form a VH/VL pair that cannot specifically bind to an antigen to which an "active" VH or "active" VL would be able to bind if bound to a non-inactive analogous VL or VH. Exemplary "inactive VH" and "inactive VL" domains are formed by mutation of wild-type VH or VL sequences. Exemplary mutations are within CDR1, CDR2, or CDR3 of VH or VL. Exemplary mutations include placement of a domain linker within CDR2, thereby forming an "inactive VH" or "inactive VL" domain. In contrast, an "active VH" (aVH) or "active VL" (aVL) is a VH or VL that is capable of specifically binding to its target antigen after pairing with its "active" homologous partner, i.e., VL or VH, respectively.

In contrast, as used herein, the term "active" refers to a CD3 binding domain capable of specifically binding to CD 3. This term is used in two cases: (a) when referring to a single member of an Fv binding pair (i.e., VH or VL) having a sequence capable of pairing with its cognate partner and specifically binding to CD 3; and (b) a pair of homologues of a sequence capable of specifically binding to CD3 (i.e., VH and VL). Exemplary "active" VH, VL or VH/VL pairs are wild-type or parental sequences.

"CD-x" refers to a Cluster of Differentiation (CD) protein. In exemplary embodiments, CD-x is selected from those CD proteins that have an effect on the recruitment or activation of T cells in a subject administered a polypeptide construct of the invention. In an exemplary embodiment, CD-x is CD 3.

The term "binding domain" characterizes in relation to the present invention a domain that (specifically) binds to/interacts with/recognizes a given target epitope or a given target site on a target molecule (antigen) such as, respectively: EGFR and CD 3. The structure and function of the target antigen binding domain (recognizing EGFR), and preferably also the structure and/or function of the CD3 binding domain (recognizing CD3) is based on the structure and/or function of an antibody (e.g., a full-length or whole immunoglobulin molecule, including sdABD). According to the present invention, the target antigen binding domain is generally characterized by the presence of three CDRs (commonly referred to in the art as variable heavy domains, although no corresponding light chain CDRs are present) that bind the target tumor antigen. Alternatively, ABD for TTA may comprise three light chain CDRs (i.e., CDR1, CDR2, and CDR3 of the VL region) and/or three heavy chain CDRs (i.e., CDR1, CDR2, and CDR3 of the VH region). The CD3 binding domain preferably also contains at least the minimal structural requirements of an antibody that allows target binding. More preferably, the CD3 binding domain comprises at least three light chain CDRs (i.e., CDR1, CDR2, and CDR3 of the VL region) and/or three heavy chain CDRs (i.e., CDR1, CDR2, and CDR3 of the VH region). It is contemplated that in exemplary embodiments, the target antigen and/or CD3 binding domain is produced by or obtainable by phage display or library screening methods.

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

By "domain linker" herein is meant an amino acid sequence that connects two domains, as outlined herein. The domain linker can be a cleavable linker, a constrained cleavable linker, a non-cleavable linker, a constrained non-cleavable linker, a scFv linker, and the like.

By "hinge linker" herein is meant an amino acid sequence of the hinge region that connects the domain to the Fc domain, as outlined herein. The hinge linker can be a cleavable linker, a constrained cleavable linker, a non-cleavable linker, a constrained non-cleavable linker, a scFv linker, and the like.

By "cleavable linker" ("CL") herein is meant an amino acid sequence that can be cleaved by a protease in a diseased tissue, preferably a human protease, as outlined herein. Cleavable linkers are typically at least 3 amino acids in length, with 4, 5,6, 7, 8, 9, 10, 11, 12, 13, 14, 15 or more amino acids being useful in the present invention, depending on the flexibility desired.

By "non-cleavable linker" ("NCL") herein is meant an amino acid sequence that is not cleavable by a human protease under normal physiological conditions.

By "non-cleavable constrained linker" or "constrained cleavable linker" ("CCL") herein is meant a short polypeptide containing a protease cleavage site (as defined herein) that links two domains as outlined herein in such a way that the two domains cannot significantly interact with each other until after they reside on different polypeptide chains, e.g., post-cleavage. When the CCL links VH and VL domains as defined herein, VH and VL cannot self-assemble to form a functional Fv prior to cleavage due to steric constraints in the intramolecular manner. Upon cleavage by the relevant protease, VH and VL can assemble intermolecularly to form an active antigen-binding domain. Typically, CCL is less than 10 amino acids in length, with 9, 8, 7, 6, 5, and 4 amino acids being useful in the present invention. Generally, protease cleavage sites are typically at least 4+ amino acids in length to confer sufficient specificity, as shown in fig. 11A, 11B, and 11C.

By "non-cleavable constrained linker" ("NCCL") or "constrained non-cleavable linker" ("CNCL") herein is meant a short polypeptide linking two domains as outlined herein in such a way that the two domains cannot significantly interact with each other and are not significantly cleaved by a human protease under physiological conditions.

By "constrained Fv domain" herein is meant a Fv domain comprising an active variable heavy domain and an active variable light domain covalently linked to a constrained linker as outlined herein in a manner such that the active heavy and light variable domains cannot interact intramolecularly to form an active Fv that will bind an antigen, such as CD 3. Thus, a constrained Fv domain is one that is similar to an scFv, but cannot bind antigen due to the presence of a constrained linker.

By "pseudo Fv domain" herein is meant a domain comprising (i) a pseudo or inactive variable heavy domain and a pseudo or inactive variable light domain, (ii) a pseudo or inactive variable heavy domain and an active variable light domain, or (iii) an active variable heavy domain and a pseudo or inactive variable light domain, said domains being connected using a domain linker (which may be cleavable, constrained, non-cleavable, unconstrained, etc.). The VHi and VLi domains of the pseudo Fv domain do not bind to human antigen when associated with each other (VHi/VLi) or when associated with an active VH or VL; thus, the VHi/VLi, VHi/VL and VLi/VH Fv domains do not bind significantly to human proteins, making these domains inert in humans.

By "single chain Fv" or "scFv" herein is meant a Variable Heavy (VH) domain covalently attached to a Variable Light (VL) domain, typically using a scFv linker as discussed herein, to form a scFv or scFv domain. The scFv domains may be located in either orientation from N-terminus to C-terminus (VH-linker-VL or VL-linker-VH).

By "single domain Fv", "sdFv", "single domain antibody" or "sdABD" herein is meant an antigen binding domain with only three CDRs, typically based on camelid antibody technology. See: protein Engineering 9(7), 1129-35 (1994); rev Mol Biotech74:277-302 (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, "protease cleavage domain" refers to a peptide sequence that incorporates "protease cleavage sites" as well as any linkers between the individual protease cleavage sites and between one or more protease cleavage sites and other functional components of the constructs of the invention (e.g., VH, VL, VHi, VLi, one or more target antigen binding domains, half-life extending domains, etc.).

As used herein, "Fc" or "Fc region" or "Fc domain" means a polypeptide comprising an antibody constant region other than a first constant region immunoglobulin domain. For IgG, the Fc domain comprises all or a portion of the hinge region between the immunoglobulin domains C γ 2 and C γ 3(CH2 and CH3) and optionally C γ 1(CH1) and C γ 2(CH 2). In EU numbering of human IgG1, the CH2-CH3 domain comprises amino acids 231 to 447, and the hinge is 216 to 230. Thus, the definition of "Fc domain" includes both amino acids 231-447(CH2-CH3) or 216-447 (hinge domain-CH 2-CH 3).

III.The protein of the present invention

The fusion proteins of the present invention have many different components (generally referred to herein as domains) linked together in a variety of ways. Some of the domains are binding domains that each bind to a target antigen (e.g., TTA or CD 3). When they bind to more than one antigen, they are referred to herein as "multispecific"; for example, the prodrug constructs of the present invention can bind to TTA and CD3, and thus be "bispecific," as shown in figure 1. The proteins of the invention may also have higher specificity; for example, if a first antigen-binding domain binds to EGFR, a second antigen-binding domain binds to EpCAM and an anti-CD 3 binding domain is present, this would be a "trispecific" molecule.

The proteins of the present invention may comprise CD3 antigen binding domains, tumor target antigen binding domains, half-life extending domains, linkers, and the like, arranged in a variety of ways as outlined herein.

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, such that the first tumor target antigen binding domain and the second tumor target antigen binding domain bind to the same tumor target antigen. In certain instances, 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 different tumor target antigens.

The proteins of the invention can be produced by co-expression and co-purification in cells to obtain complementary pairs of proteins that can bind to CD3 and tumor target antigen. In some embodiments, each complementary pair of proteins is purified separately. In some embodiments, each complementary pair of proteins is purified simultaneously or together.

In some embodiments, the expression vector comprises a nucleic acid sequence encoding one protein of a complementary pair of proteins and a nucleic acid sequence encoding the other protein of the complementary pair of proteins. In some embodiments, the host cell comprises such an expression vector. In some cases, such host cells can be cultured in a culture medium under suitable conditions to produce the protein. In some embodiments, the host cell is cultured under suitable conditions to secrete the proteins described herein into the culture medium. In certain embodiments, a culture medium comprising a secreted protein of the invention is purified to obtain a protein of a complementary pair of proteins. Useful purification methods include, but are not limited to, 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.

A.CD3 antigen binding domain

As part of the T cell receptor complex, CD3 is a protein complex comprising a CD3 γ (γ) chain, a CD3 δ (δ) chain, and two CD3 epsilon (epsilon) chains present on the cell surface CD3 associates with both the α (α) and β (β) chains of the T Cell Receptor (TCR) and CD-zeta (ζ) all to make up the T cell receptor complex, the clustering of CD3 on T cells (such as by binding to the Fv domain of CD3) results in T cell activation, similar to T cell receptor engagement, but independent of its clonal typical specificity.

However, as is known in the art, CD3 activation can cause a number of toxic side effects, and thus, the present invention relates to providing active CD3 binding of the polypeptides of the invention only in the presence of tumor cells in which specific proteases are found which then cleave the prodrug polypeptides of the invention to provide an active CD3 binding domain. Thus, in the present invention, the binding of the anti-CD 3Fv domain to CD3 is regulated by a protease cleavage domain that limits the binding of the CD3Fv domain to CD3 only in the microenvironment of diseased cells or tissues with elevated protease levels (e.g., in the tumor microenvironment described herein).

Thus, the invention provides two sets of VH and VL domains, one active set (VH and VL) and one inactive set (VHi and VLi), all four being present in one or more prodrug constructs. The constructs are formatted such that the VH and VL groups are not self-associated, but are associated with inactive partners (e.g., VHi and VL and VLi and VH as shown herein).

There are many suitable active sets of CDRs and/or VH and VL domains known in the art to be suitable for use in the present invention. For example, the CDR and/or VH and VL domains are derived from known anti-CD 3 antibodies, such as, for example, Moluomab-CD 3(OKT3), oxzelizumab (TRX4), tiplizumab (MGA031), vesizumab (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, 35 32, SPv-T3B, 11D8, XIII-141, XIII-46, XIII-87, 12F6, T3/RW2-8C8, WT 3/RW2-4B6, OKT3D, M-T301, 2, F101.01, HT 1-WT 31, and WT 31.

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

Inactive VHi and VLi domains contain "conventional" Framework Regions (FRs) that allow association such that an inactive variable domain will associate with an active variable domain, rendering the pair inactive, e.g. unable to bind CD 3. In one embodiment, the VHi and VLi of the inactive Fv domain are formed when one or both of the inactive domains are present in a complementary pair of constructs. In one embodiment, the VHi and VLi that form inactive Fv domains when one or both of the inactive domains are present are shown in FIG. 13A as construct 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 shown as Pro36 in FIG. 14A (SEQ ID NO:1), Pro67 in FIG. 14B (SEQ ID NO:4), Pro70 in FIG. 14D (SEQ ID NO:7) and Pro218 in FIG. 14E (SEQ ID NO: 10). 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, Pro68(SEQ ID NO:5) in FIG. 14C, Pro70(SEQ ID NO:7) in FIG. 14D, Pro71(SEQ ID NO:8) in FIG. 14D and Pro218(SEQ ID NO:10) in FIG. 14E.

In some embodiments, an inactive VHi domain comprises one or more (e.g., 1,2, 3, 4, 5,6, 7, 8, 9, or more) amino acid modifications (e.g., amino acid insertions, deletions, or substitutions) that, when paired with an active VL domain, render the paired VHi-VL domain incapable of binding a target antigen. In other embodiments, the inactive VLi domain comprises one or more (e.g., 1,2, 3, 4, 5,6, 7, 8, 9, or more) amino acid modifications (e.g., amino acid insertions, deletions, or substitutions) that, when paired with an active VH domain, render the paired VH-VLi domain incapable of binding to the target antigen.

As will be appreciated by those skilled in the art, there are many "inactive" variable domains that can be used in the present invention. Basically, any variable domain having a human framework region that allows self-assembly with another variable domain, regardless of the amino acids in the CDR positions in the variable region, can be used. For clarity, inactive domains are said to comprise CDRs, although technically inactive variable domains do not confer binding capacity.

In some cases, the inactive domain can be engineered to promote selective binding in prodrug form to promote the formation of intramolecular VHi-VL and VH-VLi domains prior to cleavage (e.g., beyond intermolecular pair formation). See, e.g., Igawa et al, Protein Eng.Des.selection 23(8):667-677(2010), hereby expressly incorporated by reference in its entirety, and in particular with regard to the interfacial residue amino acid substitutions.

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 constructs described herein comprise 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 CD3 epsilon when activated by a protease. In some embodiments, the polypeptide constructs described herein comprise two or more domains that specifically bind to CD3 epsilon when activated by a protease.

In some embodiments, the protease cleavage site is between the anti-CD 3 active VH domain and inactive VL domain on the first monomer and prevents them from folding and binding to CD3 on T cells. In some embodiments, the protease cleavage site is between the anti-CD 3 inactive VH domain and the active VL domain on the second monomer and prevents them from folding and binding to CD3 on T cells. Once the protease cleavage site is cleaved by a protease present at the target cell, the anti-CD 3 active VH domain of the first monomer and the anti-CD 3 active VL domain of the second monomer are capable of binding to CD3 on T cells.

In certain embodiments, the CD 3-binding domain of the polypeptide constructs described herein not only exhibits potent CD3 binding affinity to human CD3, but also exhibits excellent cross-reactivity with the corresponding cynomolgus monkey CD3 protein. In some cases, the CD 3-binding domain of the polypeptide construct cross-reacts with CD3 from cynomolgus monkeys. In some cases, the human to cynomolgus monkey KD ratio of CD3 is between 5 and 0.2.

In some embodiments, the CD3 binding domain of the antigen binding protein can be any domain that binds to CD3, including but not limited to domains from monoclonal antibodies, polyclonal antibodies, recombinant antibodies, human antibodies, humanized antibodies. In some cases, it is beneficial for the CD3 binding domain to be derived from the same species in which the antigen binding protein will ultimately be used. For example, for use in humans, it may be beneficial for the CD3 binding domain of an antigen binding protein to comprise human or humanized residues from the antigen binding domain of an 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-CD 3 binding domain comprises one or more (e.g., all three) light chain complementarity determining region 1(LC CDR1), light chain complementarity determining region 2(LC CDR2), and light chain complementarity determining region 3(LC CDR3) of a humanized or human anti-CD 3 binding domain described herein, and/or one or more (e.g., all three) heavy chain complementarity determining region 1(HC 1), heavy chain complementarity determining region 2(HC CDR2), and heavy chain complementarity determining region 3(HC CDR3) of a humanized or human anti-CD 3 binding domain described herein, e.g., a human or humanized anti-CD 3 binding domain comprising one or more (e.g., all three) LC CDRs and one or more (e.g., all three) HC CDRs.

In some embodiments, the humanized or human anti-CD 3 binding domain comprises a humanized or human light chain variable region specific for CD3, wherein the light chain variable region specific for CD3 comprises human or non-human light chain CDRs in a human light chain framework region. In some cases, the light chain framework region is a λ (λ) light chain framework. In other cases, the light chain framework region is a kappa (κ) light chain framework.

In some embodiments, one or more CD3 binding domains are humanized or fully human. In some embodiments, the one or more activated CD3 binding domains bind with a KD of 1000nM or less to CD3 on CD3 expressing cells. In some embodiments, the one or more activated CD3 binding domains bind with a KD of 100nM or less to CD3 on CD3 expressing cells. In some embodiments, the one or more activated CD3 binding domains bind with a KD of 10nM or less to CD3 on CD3 expressing cells. In some embodiments, one or more CD3 binding domains are cross-reactive with cynomolgus monkey CD 3. In some embodiments, one or more CD3 binding domains comprise an amino acid sequence provided herein.

In some embodiments, the humanized or human anti-CD 3 binding domain comprises a humanized or human heavy chain variable region specific for CD3, wherein the heavy chain variable region specific for CD3 comprises human or non-human heavy chain CDRs in a human heavy chain framework region.

In one embodiment, the anti-CD 3 binding domain is an Fv comprising a light chain and a heavy chain having the amino acid sequences provided herein. In one embodiment, the anti-CD 3 binding domain comprises: a light chain variable region comprising an amino acid sequence having at least one, two, or three modifications (e.g., substitutions, insertions, and deletions), but no more than 30, 20, or 10 modifications (e.g., substitutions, insertions, and deletions) of the amino acid sequence of a light chain variable region provided herein, or a sequence having 95% -99% identity to an amino acid sequence provided herein; and/or a heavy chain variable region comprising an amino acid sequence having at least one, two, or three modifications (e.g., substitutions, insertions, and deletions), but no more than 30, 20, or 10 modifications (e.g., substitutions, insertions, and deletions) of the amino acid sequence of a heavy chain variable region provided herein, or a sequence having 95% -99% identity to an amino acid sequence provided herein. In one embodiment, the humanized or human anti-CD 3 binding domain is an scFv and the light chain variable region comprising an amino acid sequence described herein is attached via an scFv linker to the heavy chain variable region comprising an amino acid sequence described herein. The light chain variable region and the heavy chain variable region of the scFv can be, for example, in any of the following orientations: 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 CD 3-binding domain of the antigen-binding protein has an affinity for CD3 on CD 3-expressing cells with a KD of 1000nM or less, 100nM or less, 50nM or less, 20nM or less, 10nM or less, 5nM or less, 1nM or less, or 0.5nM or less. In some embodiments, the CD 3-binding domain of the antigen-binding protein has an affinity for CD3 epsilon with a KD of 1000nM or less, 100nM or less, 50nM or less, 20nM or less, 10nM or less, 5nM or less, 1nM or less, or 0.5nM or less. In other embodiments, the CD 3-binding domain of the antigen-binding protein has a low affinity for CD3, i.e., about 100nM or greater.

Affinity for binding to CD3 can be, for example, by coating the antigen binding protein itself or its CD3 binding domain on an assay plate; displayed on the surface of a microbial cell; the ability of CD3 to bind in solution, as known in the art, is typically determined using Biacore or Octet assays. The binding activity of the antigen binding proteins of the present disclosure themselves or their CD3 binding domains to CD3 can be determined by immobilizing the ligand (e.g., human CD3) or the antigen binding protein itself or its CD3 binding domain to beads, substrates, cells, etc. The additives may be in an appropriate buffer and the binding partners incubated for a period of time at a given temperature. After washing to remove unbound material, bound protein can be released with, for example, SDS, buffers with high pH, or the like, and analyzed, for example, by Surface Plasmon Resonance (SPR).

B.Antigen binding domains against tumor target antigens

In addition to the described CD3 and half-life extending domains, the polypeptide constructs described herein further comprise at least one or at least two or more domains that bind to one or more target antigens or one or more regions on a single target antigen. It is contemplated herein that the polypeptide constructs of the invention will be cleaved at the protease cleavage domain, e.g., in a disease-specific microenvironment or in the blood of a subject, and each target antigen binding domain will bind to a target antigen on a target cell, thereby activating the CD3 binding domain to bind to T cells. In general, TTA binding domains can bind to their target prior to protease cleavage, so they can "wait" on the target cell for activation as T cell adaptors. The at least one target antigen is involved in and/or associated with a disease, disorder or condition. Exemplary target antigens include those associated with a proliferative disease, a neoplastic disease, an inflammatory disease, an immune disorder, an autoimmune disease, an infectious disease, a viral disease, an allergic reaction, a parasitic reaction, a graft-versus-host disease, or a host-versus-graft disease. In some embodiments, the target antigen is a tumor antigen expressed on a tumor cell. Or in some embodiments, the target antigen is associated with a pathogen (such as a virus or bacterium). The at least one target antigen may also be directed against healthy tissue.

In some embodiments, the target antigen is a cell surface molecule, such as a protein, lipid, or polysaccharide. In some embodiments, the target antigen is on a tumor cell, a virus-infected cell, a bacteria-infected cell, an injured red blood cell, an arterial plaque cell, or a fibrotic tissue cell. It is contemplated herein that upon binding more than one target antigen, two inactive CD3 binding domains co-localize on the surface of the target cell and form an active CD3 binding domain. In some embodiments, the antigen binding domain comprises more than one target antigen binding domain to activate an inactive CD3 binding domain in the antigen binding protein. In some embodiments, the antigen binding domain comprises more than one target antigen binding domain to enhance the binding strength to a target cell. In some embodiments, the antigen binding domain comprises more than one target antigen binding domain to enhance the binding strength to a target cell. In some embodiments, more than one antigen binding domain comprises the same antigen binding domain. In some embodiments, more than one antigen binding domain comprises different antigen binding domains. For example, two different antigen binding domains that are known to be doubly expressed in a diseased cell or tissue (e.g., a tumor or cancer cell) can enhance the binding or selectivity of an antigen binding protein for a target.

Polypeptide constructs contemplated herein comprise at least one antigen binding domain, wherein the antigen binding domain binds to at least one target antigen. In some cases, the target antigen is expressed on the surface of a diseased cell or tissue (e.g., a tumor or cancer cell). Target antigens include, but are not limited to, epithelial cell adhesion molecule (EpCAM), Epidermal Growth Factor Receptor (EGFR), human epidermal growth factor receptor 2(HER-2), human epidermal growth factor receptor 3(HER-3), c-Met, folate receptor 1(FOLR1), B7H3(CD276), LY6/PLAUR domain protein 3(LYPD3), and carcinoembryonic antigen (CEA).

Polypeptide constructs disclosed herein also include proteins comprising two antigen binding domains that bind to two different target antigens known to be expressed on diseased 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.

In some embodiments, the binding domain of the target antigen is a 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 variable domain of a nanobody derived from a camel (VHH), in other embodiments, the binding domain of the target antigen is a non-Ig binding domain, i.e., an antibody mimetic such as anticalin, human ubiquitin (affilin), an affibody molecule, an affimer, an avidin (affitin), α, an avimer, a DARPin, a fynomer, a kunitz domain peptide, and an antibody analog (monodies).

In some embodiments, the target cell antigen-binding domain independently comprises a scFv, a VH domain, a VL domain, a non-Ig domain, or a ligand that specifically binds to a target antigen. In some embodiments, the target antigen binding domain specifically binds to a cell surface molecule. In some embodiments, the target antigen binding domain specifically binds to a tumor antigen. In some embodiments, the target antigen binding domain specifically and independently binds to an antigen selected from at least one of: EpCAM, EGFR, HER-2, HER-3, cMet, LyPD3, CEA, and FoIR. In some embodiments, the target antigen binding domain specifically and independently binds to two different antigens, wherein at least one of the antigens is selected from one of EpCAM, EGFR, HER-2, HER-3, cMet, CEA, LyPD3, B7H3, and FOLR 1.

In many embodiments, the Antigen Binding Domain (ABD) of the Target Tumor Antigen (TTA) is based on the single domain antigen binding domain of a camelid single domain antibody (sdABD) (sdABD-TTA). sdABD-TTA has a framework region (as with a classical antibody), and three CDRs, but does not have any heavy chain constant domain. These sdABD-TTAs are generally preferred over TTA-binding scfvs because the intramolecular folds that result in the formation of inactive fvs that do not bind to CD3 are less complex, having fewer VH and VL domains. These sdABD-TTAs can be labeled by the target to which they bind, e.g., sdABD-EGFR is an sdABD that binds to human EGFR, etc.

In some embodiments, the antigen binding domain binds 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 EGFR and has the amino acid sequence set forth in SEQ ID NO. 14 or shown in FIG. 13A as construct I-1 in a humanized version. In other embodiments, the antigen binding domain binds EGFR and has the CDRs and/or variable domains 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 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 EGFR and has the amino acid sequence set forth in SEQ ID NO 21 or shown in FIG. 13A as construct I-8 in a humanized version. In some embodiments, the antigen binding domain binds EGFR and has the CDR and/or variable domain of the sequence set forth in SEQ ID NO:21 or shown in FIG. 13A as construct I-8.

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

In some embodiments, the antigen binding domain binds 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 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 EpCAM and has the CDRs and/or variable domains of the sequences set forth in SEQ ID No. 15 or shown as construct I-2 in fig. 13A.

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

C.Half-life extension

The proteins of the invention optionally comprise a half-life extending domain. Such domains are contemplated to include, but are not limited to, HSA binding domains, Fc regions, small molecules, and other half-life extending domains known in the art.

1.Fc region

Proteins of the invention include Fc domain-fusion proteins that combine the Fc region of an antibody with additional components as outlined herein, including ABD to TTA and Fv domains, typically pseudo-domains as outlined herein.

The term "homodimeric Fc protein" as used herein refers to an Fc protein that is capable of forming a heterodimer with a heterodimer. In some cases, the mortar form may be combined with a disulfide bond or a pair of charged amino acid substitutions to further facilitate heterodimer formation.

In some embodiments, the heterodimeric Fc protein comprises an Fc arm comprising a knob or hole in the Fc region. In other words, the first monomeric Fc arm comprises a pestle, and the second monomeric Fc arm comprises a hole. In embodiments of "construct 6" or "construct 7", the monomeric Fc arm containing an active Fv domain (e.g., an anti-CD 3 variable heavy chain and variable light chain) comprises a CH 3-knob, and the monomeric Fc arm containing a pseudo-Fv domain (e.g., an inactive variable heavy chain and an inactive variable light chain) comprises a CH 3-hole, although this may be reversed. In other embodiments, the monomeric Fc arm containing an active Fv domain comprises a CH 3-hole and the monomeric Fc arm containing a pseudo Fv domain comprises a CH 3-knob.

The amino acid residues that form the pestle are typically 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 residues that form the knob have a small side chain volume, such as alanine, asparagine, aspartic acid, glycine, serine, threonine, or valine. Exemplary amino acid substitutions in the CH3 domain that form the pestle include, but are not limited to, T366W, T366Y, or F405W substitutions.

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

In general, preferred Fc domains for use herein are human IgG domains, and typically IgG1 or IgG 4. In some cases, e.g., when effector function is undesirable, IgG4 is used, and in some cases, the S228P variant is contained in the hinge domain, as this prevents arm exchange.

It is understood that other modifications to the Fc region known in the art to favor heterodimerization are also contemplated and encompassed by the present application.

In some embodiments, the Fc region of the formats described herein comprises a tag at the C-terminus of the Fc, such as, but not limited to, a histidine tag (e.g., (His)6)), a streptavidin tag (e.g., a streptavidin-tag or streptavidin-tag II), or a Maltose Binding Protein (MBP) tag.

In addition, the Fc domain may contain additional amino acid modifications to alter effector function or half-life, as known in the art.

2.Human serum albumin binding domains

Human Serum Albumin (HSA) (molecular weight about 67kDa) is the most abundant protein in plasma, is present at about 50mg/ml (600 μ M), and has a half-life of about 20 days in humans. HSA is used to maintain plasma pH, promote colloidal blood pressure, act as a carrier for many metabolites and fatty acids, and act as the primary drug transporter in plasma.

Non-covalent association with albumin extends the elimination half-life of the short-lived protein. For example, recombinant fusions of albumin binding domains with Fab fragments resulted in 25 and 58 fold reduction in vivo clearance and 26 and 37 fold increase in half-life when administered intravenously to mice and rabbits, respectively, as compared to administration of Fab fragments alone. In another example, when insulin is acylated with a fatty acid to promote association with albumin, a long lasting effect is observed when injected subcutaneously in rabbits or pigs. Together, these studies demonstrate a link between albumin binding and protraction.

In one aspect, the antigen binding proteins described herein comprise a half-life extending domain, e.g., a domain that specifically binds to HSA. In some embodiments, the HSA binding domain of the antigen binding protein can be any domain that binds to HSA, including but not limited to domains from monoclonal antibodies, polyclonal antibodies, recombinant antibodies, human antibodies, humanized antibodies. In some embodiments, the HSA binding domain is a single chain variable fragment (scFv), a single domain antigen binding domain (sdABD) such as a heavy chain variable domain (VH), a light chain variable domain (VL), and a variable domain of a camelid-derived nanobody (VHH), a peptide, ligand, or small molecule specific for HSA. In certain embodiments, the HSA binding domain is from a single domain antibody (sdABD) and comprises a single domain antigen binding domain (sdABD); that is, sdABD is a single variable domain (VHH) that contains three CDRs rather than the standard six CDRs in the Fv of traditional antibodies. In other embodiments, the HSA binding domain is a peptide. In other embodiments, the HSA binding domain is a small molecule. It is contemplated that in some embodiments, the HSA binding domain of the antigen binding protein is relatively small and does not exceed 25kD, does not exceed 20kD, does not exceed 15kD, or does not exceed 10 kD. In certain instances, if the HSA binding domain is a peptide or small molecule, it is 5kD or less.

The half-life extending domain of the antigen binding protein provides altered pharmacodynamics and pharmacokinetics of the antigen binding protein itself. As described above, the half-life extending domain extends the elimination half-life. The half-life extending domain also alters pharmacodynamic properties, including altering tissue distribution, penetration, and diffusion of the antigen binding protein. In some embodiments, the half-life extending domain provides improved tissue (including tumor) targeting, tissue penetration, tissue distribution, tissue in-diffusion, and enhanced efficacy compared to proteins without the half-life extending binding domain. In one embodiment, the method of treatment effectively and efficiently utilizes reduced amounts of antigen binding proteins, resulting in reduced side effects, such as reduced non-tumor cell cytotoxicity.

In addition, features of the half-life extending domain, e.g., the HSA binding domain, include the binding affinity of the HSA binding domain for HSA. The affinity of the HSA binding domain can be selected to target a particular elimination half-life in a particular polypeptide construct. Thus, 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 other embodiments, the HSA binding domain has a low or edge binding affinity. Exemplary binding affinities include KD concentrations of 10nM or less (high), between 10nM and 100nM (medium), and greater than 100nM (low). As described above, the binding affinity for HSA is determined by a known method such as Surface Plasmon Resonance (SPR).

D.Protease cleavage site

As outlined herein, the polypeptide (e.g., protein) compositions of the present invention, and in particular the prodrug constructs, comprise one or more protease cleavage sites, typically located in a cleavable linker.

As described herein, the prodrug constructs of the present invention comprise at least one protease cleavage site comprising an amino acid sequence that is cleaved by at least one protease. In some cases, a protein described herein comprises 1,2, 3, 4, 5,6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20 or more protease cleavage sites that are cleaved by at least one protease. As discussed more fully herein, when more than one protease cleavage site is used in the prodrug construct, the protease cleavage sites may be the same (e.g., multiple sites cleaved by a single protease) or different (two or more cleavage sites 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 may utilize one, two, three, etc.; for example, some constructs may utilize three sites for two different proteases, etc.

The amino acid sequence of the protease cleavage site will depend on the protease targeted. As is known in the art, there are many human proteases that are found in vivo and may be associated with disease states.

Proteases are known to be secreted by some diseased cells and tissues, such as tumor or cancer cells, creating a protease-rich microenvironment (tumor-rich in proteins). In some cases, the subject's blood is rich in proteases. In some cases, cells surrounding the tumor secrete proteases into the tumor microenvironment. Cells surrounding tumors that secrete proteases include, but are not limited to, tumor stromal cells, myofibroblasts, blood cells, mast cells, B cells, NK cells, regulatory T cells, macrophages, cytotoxic T lymphocytes, dendritic cells, mesenchymal stem cells, polymorphonuclear cells, and other cells. In some cases, the protease is present in the blood of the subject, e.g., a protease that targets an amino acid sequence found in a microbial peptide. This feature allows for additional specificity of targeted therapeutics such as antigen binding proteins, since T cells will not be bound by antigen binding proteins except in the protease-rich microenvironment of the targeted cells or tissues.

Proteases include, but are not limited to, serine proteases, cysteine proteases, aspartic proteases, threonine proteases, glutamic proteases, metalloproteinases, asparaginase, serum proteases, cathepsins (e.g., cathepsin B, cathepsin C, cathepsin D, cathepsin E, cathepsin K, cathepsin L, cathepsin S), kallikrein, hK1, hK10, hK15, KLK7, cathepsin B, collagenase type IV, stromelysin, factor XA, chymotrypsin-like protease, trypsin-like protease, elastase-like protease, subtilisin-like protease, actinidin, bromelain, calpain, caspase (e.g., caspase-3), Mir1-CP, papain, HIV-1 protease, HSV protease, CMV protease, rennin, pepsin, proteolytic enzyme, legumain, plasmodium aspartic protease (plasin-3), MMP (endoplasmin kinase), MMP (MMP-III) protease), proteinase (MMP-III) and MMP (MMP) 2-III) protease).

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

E.Joint

As discussed herein, the different domains of the invention are typically linked together using amino acid linkers, which may also confer functionality, including flexibility or rigidity (e.g., steric constraints) and the ability to cleave using in situ proteases. These linkers can be classified in a variety of ways.

The present invention provides "domain linkers" for linking two or more domains (e.g., VH and VL, target tumor antigen binding domains (TTABD, also sometimes referred to herein as "α TTA") (for "anti-TTA") to VH or VL, half-life extending domains to another component, etc. domain linkers can be, for example, non-cleavable linkers (NCL), cleavable linkers ("CL"), Cleavable and Constrained Linkers (CCL), and non-cleavable and constrained linkers (NCCL).

1.Non-cleavable linker

In one embodiment, the domain linker is a non-cleavable linker (NCL). In this embodiment, linkers are used to join domains to retain the functionality of the domains, typically by longer flexible domains that are not cleaved by in situ proteases in the patient. Examples of non-cleavable linkers suitable for ligation to the interior of a domain in a polypeptide of the invention include, but are not limited to, (GS) n, (GGS) n, (GGGS) n (SEQ ID NO:27), (GGSG) n (SEQ ID NO:28), (GGSGG) n (SEQ ID NO:29), or (GGGGS) n (SEQ ID NO:30), wherein n is 1,2, 3, 4, 5,6, 7, 8, 9, or 10.

In some embodiments, the linker does not contain a cleavage site, and is also too short to allow intramolecular self-assembly of the protein domains separated by the linker, and is a "constrained non-cleavable linker" or "CNCL". For example, in Pro219 and Pro217, the active VH and active VL are separated by 8 amino acids ("8 mers") that do not allow intramolecular assembly of the VH and VL into an active antigen binding domain; instead, intermolecular assembly occurs in the case of Pro218 until cleaved by tumor proteases. In some embodiments, the linker is still flexible; for example, (GGGS) n, where n ═ 2. In other embodiments, although generally less preferred, more rigid linkers, such as those including proline or bulky amino acids, may be used.

2.Cleavable linker

All prodrug constructs herein comprise at least one cleavable linker. Thus, in one embodiment, the domain linker is Cleavable (CL), sometimes referred to herein as a "protease cleavage domain" ("PCD"). In this embodiment, the CL contains a protease cleavage site, as outlined herein and as depicted in fig. 11A, 11B, and 11C. In some cases, CL contains only protease cleavage sites. Optionally, depending on the length of the cleavage recognition site, there may be additional several connecting amino acids at one or both of the N-terminus or C-terminus of the CL; for example, 1,2, 3, 4, or 5-8 amino acids may be present at one or both of the N-terminus and C-terminus of the cleavage site.

IV.Expression method

The invention provides nucleic acids and expression vectors and host cells encoding two monomers of the heterodimeric proteins of the invention. One or two expression vectors may be prepared as will be understood by those skilled in the art. That is, a first nucleic acid encoding a first monomer and a second nucleic acid encoding a second monomer can be placed in a single expression vector or in two expression vectors. One or more of the expression vectors are then placed in a host cell, which is grown such that both monomers are expressed. In some cases, although this is generally not preferred, each monomer may be produced in a separate host cell, and the expression products may then be combined to form the heterodimeric prodrug proteins of the invention.

However, most embodiments rely on the use of co-expression of two monomers. That is, provided herein are methods for producing a protein of the invention by co-expressing and co-purifying in a cell (e.g., a host cell) to obtain a first monomeric Fc polypeptide and a second monomeric Fc polypeptide. In some embodiments, complementary pairs of proteins (e.g., a first monomeric Fc polypeptide and a second monomeric Fc polypeptide) are produced in about an equimolar ratio (e.g., about a 1:1 ratio). In other embodiments, the complementary pair of proteins (e.g., the first monomeric Fc polypeptide and the second monomeric Fc polypeptide) are produced in a non-equimolar ratio (e.g., not about a 1:1 ratio). In other words, the methods described herein may be used to obtain a ratio of the first polypeptide to the second polypeptide, such as, 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, 1:90, 1:80, 1:90, 1:1, 1:30, 1:35, etc.

A polynucleotide (or expression vector) encoding a polypeptide can be expressed in a cell in a specific amount to produce a desired amount of the polypeptide. In some embodiments, the amount of the first nucleotide (or first expression vector) encoding the first monomeric Fc polypeptide and the amount of the polynucleotide (or expression vector) encoding the second monomeric Fc polypeptide introduced (e.g., transfected, electroporated, transduced, etc.) into the cell are the same. For example, the first polynucleotide and the second polynucleotide can be introduced into the cell at a ratio of about 1: 1. In other embodiments, the amount of the first nucleotide (or first expression vector) encoding the first monomeric Fc polypeptide and the amount of the second polynucleotide (or expression vector) encoding the second monomeric Fc polypeptide introduced into the cell are different. For example, the first polynucleotide and the second polynucleotide can be introduced into the cell at a ratio such as, 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, and the like.

Expression vectors for polypeptides may comprise one or more components (e.g., promoters, regulatory elements, enhancers, etc.) capable of producing the polypeptide from a cell in a desired ratio. In some cases, the first expression vector of the first monomeric Fc polypeptide comprises a component that increases the expression level of the vector compared to the expression level of the second expression vector of the second polypeptide. In other cases, the second expression vector for the second monomeric Fc polypeptide comprises a component that increases the expression level of the vector compared to the expression level of the first expression vector for the first monomeric Fc polypeptide. In certain instances, the first expression vector for the first monomeric Fc polypeptide comprises components such that the expression level of the vector is the same as the expression level of the second expression vector for the second monomeric Fc polypeptide.

In some cases, the nucleic acids described herein provide for the production of a bispecific conditionally effective protein of the present disclosure, e.g., in a mammal. The nucleotide sequences encoding the first and/or second polypeptides of the disclosure may be operably linked to transcriptional control elements, such as promoters, enhancers, and the like.

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, λ P, and trc. For expression in eukaryotic cells, suitable promoters include, but are not limited to, light and/or heavy chain immunoglobulin gene promoters and enhancer elements; cytomegalovirus immediate early promoter; a herpes simplex virus thymidine kinase promoter; early and late SV40 promoters; a promoter from a retrovirus that is present in the long terminal repeat; EF-1a, mouse metallothionein-I promoter; and various tissue-specific promoters known in the art.

The nucleic acid or nucleotide sequence encoding the protein (e.g., the prodrug construct described herein) may be present in an expression vector and/or a cloning vector. Where a protein (e.g., a prodrug construct) comprises two separate polypeptides, the nucleotide sequences encoding the two polypeptides may be cloned in the same or different vectors. The expression vector may comprise a selectable marker, an origin of replication, and other features that provide for replication and/or maintenance of the vector. Suitable expression vectors include, for example, plasmids, viral vectors, and the like.

Expression vectors typically have convenient restriction sites located near the promoter sequence to provide for insertion of a nucleic acid sequence encoding a heterologous protein. Selectable markers operable in the expression host may be present. Suitable expression vectors include, but are not limited to, viral vectors (e.g., vaccinia virus-based viral vectors; polioviruses; adenoviruses (see, e.g., Li et al, Invest Opthalmol Vis Sci 35: 25432549,1994; Borras et al, Gene Therr 6: 515524,1999; Li and Davidson, PNAS 92: 77007704,1995; Sakamoto et al, H Gene Ther 5: 10881097,1999; WO 94/12649, WO 93/03769; WO 93/19191; WO 94/28938; WO 95/11984 and WO 95/00655); adeno-associated viruses (see, e.g., Ali et al, Hum Gene Ther 9: 8186,1998, Flanne et al, PNAS94: 69166921,1997; Bennett et al, Invest Opthalmol Vis Sci 38:28572863,1997; Jomary et al, Gene Therr 4:683690,1997, Rolling et al, Genem Hur 10:641648,1999; Alnnett et al, Movat Vil scientific 93/09239; Srstein 591594,1996, J.Vir. (1989)63: 3822-3828; mendelson et al, Virol, (1988)166: 154-165; and Flotte et al, PNAS (1993)90: 10613-10617); sv 40; herpes simplex virus; human immunodeficiency virus (see, e.g., Miyoshi et al, PNAS94: 1031923,1997; Takahashi et al, J Virol 73: 78127816,1999); retroviral vectors (e.g., murine leukemia virus, spleen necrosis virus, and vectors derived from rous sarcoma virus, hayweed sarcoma virus, avian leukemia virus, human immunodeficiency virus, myeloproliferative sarcoma virus, and mammary tumor virus), and the like.

The present disclosure provides a mammalian cell modified to produce a protein, such as a prodrug construct of the present disclosure. The polynucleotides described herein can be introduced into mammalian cells using any method known to those skilled 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 (e.g., mouse, rat) cell lines, and the like. Suitable mammalian cell lines include, but are not limited to, HeLa cells (e.g., American Type Culture Collection (ATCC) number CCL-2), CHO cells (e.g., ATCC number CRL9618, CCL61, CRL9096), 293 cells (e.g., ATCC number CRL-1573), Vero cells, NIH 3T3 cells (e.g., ATCC number CRL-1658), Huh-7 cells, BHK cells (e.g., ATCC number CCL10), PC12 cells (ATCC number CRL1721), COS cells, COS-7 cells (ATCC number CRL1651), RAT1 cells, mouse cells (ATCC number CCLI.3), Human Embryonic Kidney (HEK) cells (ATCC number CRL1573), HEK293 cells, expi293 cells, HLHepG2 cells, Hut-78, Jurkat, HL-60, NK cell lines (e.g., NK number CRL, NK92, and NKS), and the like. Suitable host cells for cloning or expressing a vector encoding a target protein include prokaryotic or eukaryotic cells as described herein.

For expression of polypeptides in bacteria, see, e.g., U.S. Pat. nos. 5,648,237, 5,789,199, and 5,840,523. (see also Charlton, Methods in Molecular Biology, Vol.248 (B.K.C.Lo, eds., HumanaPress, Totowa, N.J.,2003), p.245-254, which describes the expression of antibody fragments in E.coli). After expression, the Fc fusion protein can be isolated from the bacterial cell paste as a soluble fraction and can be further purified.

In addition to prokaryotes, eukaryotic microorganisms such as filamentous fungi or yeast are suitable for cloning or expression hosts, including fungal and yeast strains, whose glycosylation pathways are "humanized", resulting in the production of antibodies with partially or fully human glycosylation patterns. See, e.g., Gerngross, Nat Biotech,2004,22: 1409-.

Plant cell cultures may also be used as hosts. See, for example, U.S. Pat. nos. 5,959,177, 6,040,498, 6,420,548, 7,125,978, and 6,417,429.

Host cells suitable for expression of glycosylated proteins are also derived from multicellular organisms (invertebrates and vertebrates). Examples of invertebrate cells include plant cells and insect cells. Various baculoviral strains have been identified that can be used for binding to insect cells, in particular for transfection of Spodoptera frugiperda cells.

In some embodiments, the host cell or stable host cell is selected based on the amount of polypeptide produced and secreted by the cell. The prodrug compositions described herein can be produced and secreted by a host cell or a stable host cell. In some cases, a suitable cell can produce an equimolar ratio (e.g., about a 1:1 ratio) of any one of the first polypeptides and any one of the second polypeptides described herein. In other embodiments, suitable cells produce a non-equimolar ratio (e.g., a ratio other than 1: 1) of any one first monomeric Fc polypeptide and any one second monomeric Fc polypeptide.

V.Exemplary forms of the invention

As will be appreciated by those skilled in the art, heterodimeric protein compositions comprising two monomers that form a prodrug composition can take a wide variety of forms. Importantly, the active variable heavy domain and the active variable light domain each eventually associate with sdABD-TTA upon cleavage. That is, typically one sdABD-TTA is linked to the active variable heavy domain by a non-cleavable domain linker and one sdABD-TTA is linked to the active variable light domain by a non-cleavable domain linker. This ensures that active CD3 ABD can be formed on the surface of tumor cells. Once cleavage occurs and the inactive VH and VL dissociate, aVH and aVL associate intermolecularly to form one or more CD3 ABDs.

For all constructs and formats provided herein, a variety of different components, e.g., sdABD-TTA, cleavage site, aVH and aVL domains, iVH and iVL domains, and Fc domains (such as all depicted in fig. 13) can be "mixed and matched" in each format.

Provided herein are heterodimeric Fc fusion prodrug proteins (see, fig. 1) comprising a first monomeric Fc polypeptide comprising (from N-terminus to C-terminus) an ABD-domain linker for TTA-an anti-CD 3 pseudo-Fv domain (e.g., active VL domain-cleavable linker-inactive VH domain) -an antigen binding domain for GFP-domain linker (hinge linker) -a hole Fc; and a second monomeric Fc polypeptide comprising (from N-terminus to C-terminus) an ABD-domain linker-anti-CD 3 pseudo Fv domain (e.g., active VH domain-cleavable linker-inactive VL domain) -domain linker (hinge linker) -Fc knob for TTA. In some embodiments, the first monomeric Fc polypeptide comprises (from N-terminus to C-terminus) an ABD-domain linker for TTA-an anti-CD 3 pseudo Fv domain (e.g., active VH domain-cleavable linker-inactive VL domain) -an antigen binding domain for GFP-domain linker (hinge linker) -Fc hole; and the second monomeric Fc polypeptide comprises (from N-terminus to C-terminus) an ABD-domain linker-anti-CD 3 pseudo-Fv domain (e.g., active VL domain-cleavable linker-inactive VH domain) -domain linker (hinge linker) -Fc knob for TTA. In some embodiments, the C-terminus of the first monomeric Fc polypeptide comprises a tag, such as but not limited to a (His)10 tag or a streptavidin-II tag, but this is not typically used for the actual prodrug molecule to be administered to a patient. In some embodiments, the C-terminus of the second monomeric Fc polypeptide comprises a tag, such as, but not limited to, a (His)10 tag or a streptavidin-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 having the mortar form; and a second monomeric Fc comprising an sdabd (tta) -NCL-active VH-CL-VLi-NCL-Fc region comprising CH2-CH3 in pestle form. In some embodiments, such heterodimeric Fc fusion prodrug proteins comprise Pro37 and Pro36, as depicted in figure 1. The amino acid sequence of Pro37 is shown in FIG. 14A. The amino acid sequence of Pro37 is shown in FIG. 14A.

Also provided herein are heterodimeric Fc fusion prodrug proteins (see, fig. 2) comprising a first monomeric Fc polypeptide comprising (from N-terminus to C-terminus) an ABD-domain linker-anti-CD 3 pseudo-Fv domain against TTA (e.g., active VL domain-cleavable linker-inactive VH domain) -domain linker (hinge linker) -Fc mortar; and a second monomeric Fc polypeptide comprising an ABD-domain linker-anti-CD 3 pseudo-Fv domain (e.g., active VH domain-cleavable linker-inactive VL domain) -domain linker (hinge linker) -Fc knob for TTA. In some embodiments, the first monomeric Fc polypeptide comprises (from N-terminus to C-terminus) an ABD-domain linker-anti-CD 3 pseudo Fv domain for TTA (e.g., active VH domain-cleavable linker-inactive VL domain) -domain linker (hinge linker) -Fc mortar; and the second monomeric Fc polypeptide comprises (from N-terminus to C-terminus) an ABD-domain linker-anti-CD 3 pseudo-Fv domain (e.g., active VL domain-cleavable linker-inactive VH domain) -domain linker (hinge linker) -Fc knob for TTA. In some embodiments, the C-terminus of the first monomeric Fc polypeptide comprises a tag, such as, but not limited to, a (His)10 tag or a streptavidin-II tag. In some embodiments, the C-terminus of the second monomeric Fc polypeptide comprises a tag, such as, but not limited to, a (His)10 tag or a streptavidin-II tag. In some embodiments, such heterodimeric Fc fusion prodrug proteins comprise Pro38 and Pro36, as depicted in figure 2. The amino acid sequence of Pro36 is depicted in FIG. 14A. The amino acid sequence of Pro38 is depicted in FIG. 14B.

Provided herein are heterodimeric Fc fusion prodrug proteins (see, fig. 4) comprising a first monomeric Fc polypeptide comprising (from N-terminus to C-terminus) an ABD-domain linker-anti-CD 3 pseudo-Fv domain against TTA (e.g., active VL domain-cleavable linker-inactive VH domain) -Fc mortar; and a second monomeric Fc polypeptide comprising (from N-terminus to C-terminus) an ABD-domain linker-anti-CD 3 pseudo Fv domain (e.g., active VH domain-cleavable linker-inactive VL domain) -Fc knob for TTA. In some embodiments, the first monomeric Fc polypeptide comprises (from N-terminus to C-terminus) an ABD-domain linker-anti-CD 3 pseudo Fv domain for TTA (e.g., active VH domain-cleavable linker-inactive VL domain) -Fc mortar; and the second monomeric Fc polypeptide comprises (from N-terminus to C-terminus) an ABD-domain linker-anti-CD 3 pseudo-Fv domain (e.g., active VL domain-cleavable linker-inactive VH domain) -Fc knob for TTA. In some embodiments, the C-terminus of the first monomeric Fc polypeptide comprises a tag, such as, but not limited to, a (His)10 tag or a streptavidin-II tag. In some embodiments, the C-terminus of the second monomeric Fc polypeptide comprises a tag, such as, but not limited to, a (His)10 tag or a streptavidin-II tag. In some embodiments, such heterodimeric Fc fusion prodrug proteins comprise Pro68 and Pro67, as depicted in fig. 4. Pro68 is similar to Pro37, but does not include a sdABD that binds GFP or a domain linker attached to the CH2 domain. Pro67 is similar to Pro36, but does not include a domain linker attached to the CH2 domain. The amino acid sequence of Pro68 is depicted in FIG. 14C. The amino acid sequence of Pro67 is depicted in FIG. 14B.

Provided herein are heterodimeric Fc fusion prodrug proteins (see figure 5) comprising a first monomeric Fc polypeptide comprising (from N-terminus to C-terminus) an ABD-Fc socket for TTA; and a second monomeric Fc polypeptide comprising (from N-terminus to C-terminus) ABD-domain linker-anti-CD 3 pseudo Fv domain for TTA (e.g., active VH domain-cleavable linker-inactive VL domain) -Fc knob-cleavable linker-ABD-domain linker for TTA-anti-CD 3 pseudo Fv domain (e.g., active VL domain-cleavable linker-inactive VH domain). In some embodiments, the first monomeric Fc polypeptide comprises (from N-terminus to C-terminus) an ABD-Fc socket for TTA; and the second monomeric Fc polypeptide comprises (from N-terminus to C-terminus) an ABD-domain linker for TTA-an anti-CD 3 pseudo-Fv domain (e.g., active VL domain-cleavable linker-inactive VH domain) -Fc knob-cleavable linker-ABD-domain linker for TTA-an anti-CD 3 pseudo-Fv domain (e.g., active VH domain-cleavable linker-inactive VL domain). In some embodiments, the C-terminus of the first monomeric Fc polypeptide comprises a tag, such as, but not limited to, a (His)10 tag or a streptavidin-II tag. In some embodiments, the C-terminus of the second monomeric Fc polypeptide comprises a tag, such as, but not limited to, a (His)10 tag or a streptavidin-II tag. In some embodiments, such heterodimeric Fc fusion prodrug proteins comprise Pro69 and Pro70, as depicted in figure 5. Pro70 is similar to Pro67, with the Pro9 construct attached at the C-terminus. The amino acid sequence of Pro69 is depicted in FIG. 14C. The amino acid sequence of Pro70 is depicted in FIG. 14D.

Provided herein are heterodimeric Fc fusion prodrug proteins (see, fig. 6) comprising a first monomeric Fc polypeptide comprising (from N-terminus to C-terminus) ABD-Fc socket-cleavable linker for TTA-ABD-domain linker for TTA-anti-CD 3 pseudo-Fv domain (e.g., active VL domain-cleavable linker-inactive VH domain); and a second monomeric Fc polypeptide comprising (from N-terminus to C-terminus) an ABD-domain linker-anti-CD 3 pseudo Fv domain (e.g., active VH domain-cleavable linker-inactive VL domain) -Fc knob for TTA. In some embodiments, the first monomeric Fc polypeptide comprises (from N-terminus to C-terminus) ABD-Fc mortar-cleavable linker for TTA-ABD-domain linker for TTA-anti-CD 3 pseudo Fv domain (e.g., active VH domain-cleavable linker-inactive VL domain); and the second monomeric Fc polypeptide comprises (from N-terminus to C-terminus) an ABD-domain linker-anti-CD 3 pseudo-Fv domain (e.g., active VL domain-cleavable linker-inactive VH domain) -Fc knob for TTA. In some embodiments, the C-terminus of the first monomeric Fc polypeptide comprises a tag, such as, but not limited to, a (His)10 tag or a streptavidin-II tag. In some embodiments, the C-terminus of the second monomeric Fc polypeptide comprises a tag, such as, but not limited to, a (His)10 tag or a streptavidin-II tag. In some embodiments, such heterodimeric Fc fusion prodrug proteins comprise Pro71 and Pro67, as depicted in figure 6. Pro71 is similar to Pro69, with the Pro9 construct attached at the C-terminus. Pro67 is similar to Pro36, but does not have a domain linker attached to the CH2 domain. The amino acid sequence of Pro71 is depicted in FIG. 14D. The amino acid sequence of Pro67 is depicted in FIG. 14B.

Provided herein are heterodimeric Fc fusion prodrug proteins (see, fig. 8) comprising a first monomeric Fc polypeptide comprising (from N-terminus to C-terminus) an ABD-domain linker-constrained anti-CD 3Fv domain for TTA (e.g., active VH domain-non-cleavable constrained linker (NCCL) -active VL domain) -cleavable linker-Fc knob; and a second monomeric Fc polypeptide comprising (from N-terminus to C-terminus) an anti-CD 3 pseudo-Fv domain (e.g., inactive VL domain-non-cleavable linker-inactive VH domain) -cleavable linker-Fc hole. In some embodiments, the first monomeric Fc polypeptide comprises (from N-terminus to C-terminus) an ABD-domain linker-constrained anti-CD 3Fv domain for TTA (e.g., active VL domain-non-cleavable constrained linker (NCCL) -active VH domain) -cleavable linker-Fc knob; and a second monomeric Fc polypeptide comprising (from N-terminus to C-terminus) an anti-CD 3 pseudo Fv domain (e.g., inactive VL domain-non-cleavable linker-inactive VL domain) -cleavable linker-Fc hole. In some embodiments, the C-terminus of the first monomeric Fc polypeptide comprises a tag, such as, but not limited to, a (His)10 tag or a streptavidin-II tag. In some embodiments, the C-terminus of the second monomeric Fc polypeptide comprises a tag, such as, but not limited to, a (His)10 tag or a streptavidin-II tag. In some embodiments, such heterodimeric Fc fusion prodrug proteins comprise Pro219 and Pro218, as depicted in figure 8. The amino acid sequence of Pro218 is depicted in FIG. 14E. The amino acid sequence of Pro219 is depicted in FIG. 14E.

Provided herein are heterodimeric Fc fusion prodrug proteins (see, fig. 9) comprising a first monomeric Fc polypeptide comprising (from N-terminus to C-terminus) an ABD-domain linker-constrained anti-CD 3Fv domain for TTA (e.g., active VH domain-non-cleavable constrained linker (NCCL) -active VL domain) -ABD-cleavable linker-Fc knob for TTA; and a second monomeric Fc polypeptide comprising (from N-terminus to C-terminus) an anti-CD 3 pseudo-Fv domain (e.g., inactive VL domain-non-cleavable linker-inactive VH domain) -cleavable linker-Fc hole. In some embodiments, the first monomeric Fc polypeptide comprises (from N-terminus to C-terminus) an ABD-domain linker-constrained anti-CD 3Fv domain for TTA (e.g., active VL domain-non-cleavable constrained linker (NCCL) -active VH domain) -ABD-cleavable linker-Fc knob for TTA; and a second monomeric Fc polypeptide comprising (from N-terminus to C-terminus) an anti-CD 3 pseudo-Fv domain (e.g., inactive VH domain-non-cleavable linker-inactive VL domain) -cleavable linker-Fc hole. In some embodiments, the C-terminus of the first monomeric Fc polypeptide comprises a tag, such as, but not limited to, a (His)10 tag or a streptavidin-II tag. In some embodiments, the C-terminus of the second monomeric Fc polypeptide comprises a tag, such as, but not limited to, a (His)10 tag or a streptavidin-II tag. In some embodiments, such heterodimeric Fc fusion prodrug proteins comprise Pro217 and Pro218, as depicted in fig. 9. The amino acid sequence of Pro218 is depicted in FIG. 14E. The amino acid sequence of Pro217 is depicted in FIG. 14E.

The ABD directed against TTA of the heterodimeric Fc prodrug construct may be a single domain antibody that binds to TTA. In some embodiments, the single domain antibody (sdABD) is an sdABD that is directed to EGFR, an sdABD that is directed to EpCAM, an sdABD that is directed to another target tumor antigen. In some embodiments of the heterodimeric Fc fusion prodrug protein, the sdABD of the first monomeric Fc and the sdABD of the second monomeric Fc have identical or substantially identical amino acid sequences. In some embodiments, the sdABD of the first monomeric Fc and the sdABD of the second monomeric Fc bind the same TTA. In other embodiments, the sdABD of the first monomeric Fc and the sdABD of the second monomeric Fc bind different TTAs. In other cases, the sdABD of the first monomeric Fc and the sdABD of the second monomeric Fc have different amino acid sequences. In some embodiments, the sdABD has the CDRs and/or variable domains set forth in SEQ id No. 14 or construct I-1 of fig. 13A. In some embodiments, the sdABD has the CDRs and/or variable domains set forth in SEQ ID No. 15 or construct I-2 of fig. 13A.

In some embodiments of the CD3 binding domain, the active VH and active VL have the sequences of constructs I-3 and I-4, and VHi and VLi have the sequences of constructs I-6 and I-5 of FIG. 13A. In some cases, the pseudo-Fv domain of the first monomeric Fc protein can comprise VL and VHi, linked using a cleavable linker, (N-terminus to C-terminus) VL-linker-VHi or VHi-linker-VL. In some embodiments, the pseudo Fv domain has the structure (N-terminal to C-terminal) vlFR1-vlCDR1-vlFR2-vlCDR2-vlFR3-vlCDR3-vl FR4-CL-vhiFR1-vhi CDR1-vhi FR2-vhi CDR2-vhi FR3-vhi CDR3-vhi FR 4. In other cases, the pseudo Fv domain has the structure (N-terminal to C-terminal) vhiFR1-vhi CDR1-vhi FR2-vhi CDR2-vhi FR3-vhi CDR3-vhi FR4-CL-vlFR 1-vlI CDR1-vlFR 2-vlI CDR2-vlFR 3-vlI CDR3-vlFR 4. In some embodiments, the pseudo Fv domain of the second monomeric Fc protein may comprise VH and VLi connected using a cleavable linker, either a (N-terminus to C-terminus) VH-linker-VLi or VLi-linker-VH. In some cases, the pseudo-Fv domain has the structure (N-terminal to C-terminal) vhFR1-vhCDR1-vhFR2-vhC DR2-vhFR3-vhCDR3-vhFR4-CL-vliFR1-vliCDR1-vliFR2-vliCDR2-vliF R3-vliCDR3-vliFR 4. In other cases, the pseudo-Fv domain has the structure (N-terminal to C-terminal) vliFR1-vliCDR1-vliFR2-vliCDR2-vliFR3-vliCDR3-vliFR4-CL-vhFR1-vhCDR1-vhFR2-vhCDR2-vhFR3-vhCDR3-vhFR 4.

In some embodiments, the present invention provides constrained Fv domains comprising active VH and active VL domains that are covalently attached using a constrained linker (which may be cleavable or non-cleavable, as outlined herein). The constrained linker prevents intramolecular association between VH and VL in the absence of cleavage. Thus, the constrained Fv domain comprises a set of six CDRs contained within the variable domain, wherein VH hcdr1, vhCDR2 and vhCDR3 of VH bind to human CD3, and VL vhCDR1, vlCDR2 and vlCDR3 of VL bind to human CD3, but in the prodrug form (e.g. uncleaved), VH and VL cannot spatially associate to form an active binding domain.

The constrained Fv domain may comprise an active VH and an active VL (VHa and VLa) or an inactive VH and VL (VHi and VLi). As will be appreciated by those skilled in the art, in constrained active Fv domains, the order of VH and VL may be (N-terminal to C-terminal) VH-linker-VL or VL-linker-VH. As outlined herein, a constrained active Fv domain may comprise a VH and VL linked using a non-cleavable linker, in some cases such as those shown as Pro219 in fig. 8 and 14E or Pro217 in fig. 9 and 14E. In this embodiment, the constrained Fv domain has the structure (N-terminal to C-terminal) vhFR1-vhCDR1-vhFR2-vhCDR2-vhFR3-vhCDR3-vhFR4-CCL-vlFR1-vlCDR1-vlFR2-vlCDR2-vlFR3-vlCDR3-vlFR 4. In this embodiment, the CDRs and/or variable domains are those of constructs I-3 and I-4 of FIG. 13A.

VI.Examples

A. Example 1: construction and purification of Pro constructs

Transfection

Each pair of constructs was expressed from a separate expression vector (pcdna3.4 derivative). Equal amounts of plasmid DNA encoding one and a half-COBRA were mixed and transfected into Expi293 cells according to the manufacturer's transfection protocol. 5 days after transfection, conditioned medium was harvested by centrifugation (6000rpm X25') and filtration (0.2uM filter). Protein expression was confirmed by SDS-PAGE. The construct was purified and the final buffer composition was: 25mM citrate, 75mM arginine, 75mM NaCl, 4% sucrose, pH 7. The final preparation was stored at-80 ℃.

Protease cleavage

EK

Recombinant human enterokinase (R & D systems catalog number 1585-SE-010) was used to cleave the heterodimeric Fc prodrug protein described herein. The recombinant protease was activated according to the manufacturer's procedure and prepared at a stock concentration of about 100 mM.

Buffer exchange of test samples (Pro36+37, Pro36+38, Pro67+68, Pro69+70, Pro67+71, Pro217+218, and Pro218+219) to HEPES buffered saline with sodium chloride (25mM HEPES, 50mM NaCl, 2mM CaCl)₂) And incubated overnight at room temperature with the appropriate protease at 10nM final concentration. Cleavage was confirmed by SDS-PAGE.

MMP-9

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

Dilution of rhMMP9 to about 100ug/ml with assay buffer (25uL hMMP9+75uL assay buffer)

Addition of phenylmercuric p-aminoacetate (APMA) from 100mM stock solution in DMSO to a final concentration of 1mM (1uL to 100uL)

Incubation at 37' C for 24 hours

Dilution of MMP9 to 10ng/ul (addition of 900ul assay buffer to 100ul activation solution)

The concentration of activated rhMMP9 was about 100 nM.

Cleavage of constructs for TDCC assay

To lyse the construct, 100ul of a 1mg/ml concentration (10.5uM) protein sample in formulation buffer (25mM citric acid, 75mM L-arginine, 75mM NaCl, 4% sucrose) was supplemented with CaCl2 to 10 mM. Activated rhMMP9 was added to a concentration of 20-35 nM. The samples were incubated overnight (16-20 hours) at room temperature. SDS PAGE (10% -20% TG, TG running buffer, 200v, 1 h) was used to verify the completion of the lysis. The sample is typically 98% lysed.

B. Example 2: t cell-dependent cytotoxicity (TDCC) assay to test the efficacy of activated heterodimeric Fc prodrug proteins

Firefly luciferase-transduced HT-29 cells were grown to approximately 80% confluence and detached with Versene (EDTA-Ca-Mg in 0.48 mMPBS). The cells were centrifuged and resuspended in TDCC medium (5% heat-inactivated FBS in RPMI 1640 with HEPES, GlutaMax, sodium pyruvate, non-essential amino acids and β -mercaptoethanol). The purified human P whole-T cells were thawed, centrifuged and resuspended in TDCC medium.

Co-cultures of HT-29_ Luc cells and T cells were added to 384-well cell culture plates. The serial dilutions of COBRA were then added to the co-culture and incubated

Incubation was performed for 48 hours. Finally, an equal volume of SteadyGlo luciferase assay reagent was added to the plate and incubated for 20 minutes. Plates were read on a Perkin elmer envision with an exposure time of 0.1 s/well. Total luminescence was recorded and data was analyzed on GraphPad Prism 7.

We tested the percentage of specific cytotoxicity induced when activated (cleaved) heterodimeric Fc prodrug protein engages T cells and directs cell lysis to target positive tumor cells using the above assay.

Fig. 3A and 3B show that some exemplary heterodimeric Fc prodrug constructs (such as Pro36+37 and Pro36+38) exhibit low or lack of conditionality when cleaved with a homologous protease in a TDCC assay.

Fig. 7A, 7B, and 7C show that some exemplary heterodimeric Fc prodrug constructs (such as Pro67+68 and Pro69+70) exhibit conditionality upon cleavage with a cognate protease in a TDCC assay, but lack high conditionality.

Fig. 10A and 10B show that exemplary heterodimeric Fc prodrug constructs (such as Pro217+218 and Pro218+219) exhibit conditional and high potency in a TDCC assay upon cleavage with a cognate protease.

The prodrug constructs described herein containing a single domain antibody to EGFR induce cancer cell killing upon protease cleavage in a manner comparable to a fusion protein comprising a single domain antibody to EGFR and an anti-CD 3Fv domain.

All cited references are expressly incorporated herein by reference in their entirety. While specific embodiments of the invention have been described above for purposes of illustration, it will be understood by those skilled in the art that various changes in detail may be made without departing from the invention as described in the following claims.

Claims

1. A heterodimeric protein composition comprising:

(a) a first monomer comprising from N-terminus to C-terminus:

i) a first single domain antigen binding domain (sdABD) (sdABD-TTA) that binds to a first Tumor Target Antigen (TTA);

ii) a domain linker;

iii) a first constrained Fv domain comprising:

1) a variable light domain comprising vlCDR1, vlCDR2 and vlCDR 3;

2) CNCL; and

3) a variable heavy domain comprising vhCDR1, vhCDR2, and vhCDR 3;

iv) a domain linker;

v) a second sdABD-TTA;

vi) a first cleavable linker; and

vii) a first Fc domain; and

(b) a second monomer comprising from N-terminus to C-terminus:

i) a first pseudo-Fv domain comprising:

1) a pseudo-variable light domain;

2) a non-cleavable linker; and

3) a pseudo-variable heavy domain;

ii) a second cleavable linker; and

iii) a second Fc domain;

wherein the first and second Fc domains comprise a knob and hole modification; wherein the first variable heavy domain and the first variable light domain are capable of binding to human CD3, but the constrained Fv domain does not bind to CD 3; wherein the first variable heavy domain and the first pseudo-variable light domain associate intermolecularly to form an inactive Fv; and wherein the first variable light domain and the first pseudo-variable heavy domain associate intermolecularly to form an inactive Fv.

2. A heterodimeric protein composition comprising:

(a) a first monomer comprising from N-terminus to C-terminus:

i)sdABD-TTA；

ii) a domain linker;

iii) a constrained Fv domain comprising:

1) a variable heavy domain comprising vhCDR1, vhCDR2, and vhCDR 3;

2) a constrained non-cleavable linker; and

3) a variable light domain comprising vlCDR1, vlCDR2 and vlCDR 3;

iv) a first cleavable linker; and

v) a first Fc domain; and

(b) a second monomer comprising from N-terminus to C-terminus:

i) a pseudo Fv domain comprising:

1) a pseudo-variable heavy domain;

2) a non-cleavable linker; and

3) a pseudo-variable light domain;

ii) a second cleavable linker; and

iii) a second Fc domain,

3. A heterodimeric protein composition comprising:

(a) a first monomer comprising from N-terminus to C-terminus:

i) a first sdABD-TTA;

ii) a first domain linker;

iii) a first pseudo-Fv domain comprising:

1) a variable light domain comprising vlCDR1, vlCDR2 and vlCDR 3;

2) a first cleavable linker; and

3) a pseudo-variable heavy domain; and

iv) a first Fc domain; and

(b) a second monomer comprising from N-terminus to C-terminus:

i) a second sdABD-TTA;

ii) a second domain linker;

iii) a second pseudo-Fv domain comprising:

1) a variable heavy domain comprising vhCDR1, vhCDR2, and vhCDR 3;

2) a second cleavable linker; and

3) a pseudo-variable light domain; and

iv) a first Fc domain; and is

Wherein the first and second Fc domains comprise a knob-and-hole modification, and wherein the variable light domain of the first pseudo Fv domain and the variable heavy domain of the second pseudo Fv domain do not bind human CD3 in the absence of cleavage at the cleavable linker.

4. A heterodimeric protein composition comprising:

(a) a first monomer comprising from N-terminus to C-terminus:

i) a first sdABD-TTA;

ii) a first Fc domain;

iii) a first cleavable linker;

iv) a second sdABD-TTA;

v) a first domain linker; and

vi) a first pseudo-Fv domain comprising:

1) a variable light domain comprising vlCDR1, vlCDR2 and vlCDR 3;

2) a second cleavable linker;

3) a pseudo-variable heavy domain; and

(b) a second monomer comprising from N-terminus to C-terminus:

i) a third sdABD-TTA;

ii) a second domain linker;

iii) a second pseudo-Fv domain comprising:

1) a variable heavy domain comprising vhCDR1, vhCDR2, and vhCDR 3;

2) a second cleavable linker; and

3) a pseudo-variable light domain; and

iv) a second Fc domain;

5. A heterodimeric protein composition comprising:

(a) a first monomer comprising from N-terminus to C-terminus:

i) a first sdABD-TTA; and

ii) a first Fc domain; and

(b) a second monomer comprising from N-terminus to C-terminus:

i) a second sdABD-TTA;

ii) a domain linker;

iii) a first pseudo-Fv domain comprising:

1) a variable heavy domain comprising vhCDR1, vhCDR2, and vhCDR 3;

2) a first cleavable linker; and

3) a pseudo-variable light domain;

iv) a second Fc domain;

v) a second cleavable linker;

vi) a third sdABD-TTA; and

vii) a second pseudo-Fv domain comprising:

1) a variable light domain comprising vlCDR1, vlCDR2 and vlCDR 3;

2) a third cleavable linker; and

3) a pseudo-variable heavy domain;

wherein the first and second Fc domains comprise a knob-and-hole modification, and the variable heavy domain of the first pseudo Fv domain and the variable light domain of the second pseudo Fv domain do not bind to human CD3 in the absence of cleavage at the cleavable linker.

6. The heterodimeric protein of claim 1 or 2, wherein the first variable heavy domain is N-terminal to the first variable light domain and the pseudo light variable domain is N-terminal to the pseudo variable heavy domain.

7. The heterodimeric protein of claim 1 or 2, wherein the first variable light domain is N-terminal to the first variable heavy domain and the pseudo heavy variable domain is N-terminal to the pseudo light variable domain.

8. The heterodimeric protein of 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.

9. The heterodimeric protein composition of any one of claims 1-8, 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.

10. The heterodimeric protein composition of any one of claims 1 to 9, wherein the TTA is selected from the group consisting of EGFR, FOLR1, H7B3, and EpCAM.

11. The heterodimeric protein composition of any one of claims 1 and 3-10, wherein the first and second sdabds bind to the same TTA.

12. The heterodimeric protein composition of any one of claims 1 and 3-10, wherein the first and second sdabds bind to different TTAs.

13. The heterodimeric protein composition of any one of claims 1-12, wherein the one or more sdabds are selected from the group consisting of SEQ ID NOs 14, 15, 21, 22, 23, 25, 155, 159, 163, and 167.

14. The heterodimeric protein composition of any one of claims 1-13, wherein one of the first and second Fc domains has SEQ ID No. 199 and the other has SEQ ID No. 200.

15. The heterodimeric protein composition of any one of claims to 11, wherein the first and/or second cleavable linker is cleaved by a human protease selected from the group consisting of: MMP2, MMP9, cathepsin S, cathepsin K, cathepsin L, granzyme B, uPA, Kallekriein7, proteolytic enzymes and thrombin.

16. A nucleic acid composition comprising, respectively:

a) a first nucleic acid encoding the first monomer according to any one of claims 1 to 15; and

b) a second nucleic acid encoding the second monomer according to any one of claims 1 to 15.

17. An expression vector composition comprising the first nucleic acid and the second nucleic acid according to claim 16.

18. An expression vector composition comprising:

a) a first expression vector comprising the first nucleic acid according to claim 16; and

b) a second expression vector comprising said second nucleic acid according to claim 16.

19. A host cell comprising the expression vector composition according to claim 17 or 18.

20. A method of making a heterodimeric protein, the method comprising: culturing the host cell according to claim 19 under conditions in which the heterodimeric protein is expressed, and recovering the heterodimeric protein.

21. A method of treating cancer, the method comprising administering the heterodimeric protein according to any one of claims 1 to 15.