EA044194B1 - ANTI-STEAP2 ANTIBODIES, ANTIBODY-DRUG CONJUGATES AND BISPECIFIC ANTIGEN-BINDING MOLECULES THAT BIND STEAP2 AND CD3, AND THEIR APPLICATIONS - Google Patents

Description

Link to sequence list

This application incorporates by reference a sequence listing provided in machine-readable form as file 10296WO01-Sequence.txt, created on September 22, 2017, containing 739,964 bytes.

Field of technology

This invention relates to antibodies, and antigen-binding fragments thereof, that are specific for six-transmembrane epithelial antigen of prostate 2 (STEAP2) and methods of using them. The present invention also relates to bispecific antigen binding molecules that bind STEAP2 and CD3, and methods of using them. This invention also provides antibody-drug conjugates comprising an anti-STEAP2 antibody or fragment thereof and a therapeutic agent (eg, a cytotoxic agent).

State of the art

Six transmembrane epithelial antigen of prostate 2 (STEAP2), also known as STEAP-2, STEAP2 metal reductase, prostate cancer associated protein 1, metastatic prostate cancer activated protein, six transmembrane prostate protein 1 (STAMP1) and 098P4B6, is an integral, six-stranded transmembrane protein that is activated in healthy and malignant prostate cells. STEAP2, which acts as a shuttle between the Golgi complex and the plasma membrane, is a metal reductase that reduces iron and copper, facilitating their entry into the cell. STEAP2 is mainly localized in prostate epithelial cells. STEAP2 is also expressed in healthy heart, brain, pancreas, ovary, skeletal muscle, breast, testis, uterus, kidney, lung, trachea, colon and liver. STEAP2 is overexpressed in cancer tissues, including tumors of the prostate, bladder, cervix, lung, colon, kidney, breast, pancreas, stomach, uterus and ovary (Gomes, I.M. et al., 2012, Mol. Cancer Res. 10:573-587; Challita-Eid-P.M., et al., 2003, WO 03/087306; Emtage, P.C.R., 2005, WO 2005/079490).

CD3 is a homodimeric or heterodimeric antigen expressed on T cells in association with the T cell receptor (TCR) complex and is required for T cell activation. Functional CD3 is formed from the dimeric association of two of four different chains: epsilon, zeta, delta and gamma. CD3 dimeric configurations include gamma/epsilon, delta/epsilon, and zeta/zeta. Anti-CD3 antibodies have been shown to cluster CD3 on T cells, thereby causing T cell activation in a manner similar to TCR engagement by peptide-loaded MHC molecules. Thus, anti-CD3 antibodies have been proposed for therapeutic purposes involving T cell activation. In addition, bispecific antibodies that are capable of binding CD3 and a target antigen have been proposed for therapeutic applications involving targeting T cell immune responses to tissues and cells expressing the target antigen.

Antigen-binding molecules that target STEAP2, including antibody-drug conjugates as well as bispecific antigen-binding molecules that bind both STEAP2 and CD3, will be useful in treatment plans in which specific targeting and T cell-mediated killing of cells that express STEAP2 is desired .

Brief description of the invention

In a first aspect, the present invention provides antibodies and antigen binding fragments thereof that bind to human STEAP2. Antibodies according to this aspect of the invention are useful, inter alia (among other things), for targeting cells expressing STEAP2. The present invention also provides bispecific antibodies and antigen binding fragments thereof that bind human STEAP2 and human CD3. Bispecific antibodies in accordance with this aspect of the invention are useful, inter alia, for targeting T cells expressing CD3 and for promoting T cell activation, inter alia, in conditions where T cell-mediated killing of STEAP2 expressing cells is beneficial or desirable. For example, bispecific antibodies can direct CD3-mediated T cell activation to specific STEAP2-expressing cells, such as prostate tumor cells.

Exemplary anti-STEAP2 antibodies according to this invention are listed in table. 1 and 2 of this document. In table 1 shows the amino acid sequence identifiers of the heavy chain variable regions (HCVR) and light chain variable regions (LCVR), as well as the heavy chain complementarity determining regions (HCDR1, HCDR2 and HCDR3) and the light chain complementarity determining regions (LCDR1, LCDR2 and LCDR3) of illustrative anti -STEP2 antibodies. In table 2 shows the sequence identifiers of the nucleic acid molecules encoding HCVR, LCVR, HCDR1, HCDR2, HCDR3, LCDR1, LCDR2 and LCDR3 of exemplary anti-STEP2 antibodies.

This invention provides antibodies or antigen binding fragments thereof comprising an HCVR comprising an amino acid sequence selected from any of

- 1,044,194 HCVR sequences listed in table. 1, or a substantially similar sequence thereto having at least 90%, at least 95%, at least 98%, or at least 99% sequence identity.

The present invention also provides antibodies or antigen binding fragments thereof comprising an LCVR comprising an amino acid sequence selected from any of the LCVR amino acid sequences listed in Table. 1, or a substantially similar sequence thereto having at least 90%, at least 95%, at least 98%, or at least 99% sequence identity.

The present invention also provides antibodies or antigen binding fragments thereof comprising a pair of amino acid sequences HCVR and LCVR (HCVR/LCVR) comprising any of the amino acid sequences HCVR listed in Table. 1, paired with any of the LCVR amino acid sequences listed in table. 1. In accordance with some embodiments, the present invention provides antibodies or antigen binding fragments thereof comprising the HCVR/LCVR amino acid sequence pair contained in any of the exemplary anti-STEP2 antibodies listed in Table. 1. In some embodiments, the HCVR/LCVR amino acid sequence pair is selected from the group consisting of SEQ ID NO: 250/258 (eg, H2M11162N).

The present invention also provides antibodies or antigen binding fragments thereof comprising a heavy chain CDR1 (HCDR1) comprising an amino acid sequence selected from any of the HCDR1 amino acid sequences listed in Table. 1, or a substantially similar sequence thereof having at least 90%, at least 95%, at least 98%, or at least 99% sequence identity.

The present invention also provides antibodies or antigen binding fragments thereof comprising a heavy chain CDR2 (HCDR2) comprising an amino acid sequence selected from any of the HCDR2 amino acid sequences listed in Table. 1, or a substantially similar sequence thereto having at least 90%, at least 95%, at least 98%, or at least 99% sequence identity.

The present invention also provides antibodies or antigen binding fragments thereof comprising a heavy chain CDR3 (HCDR3) comprising an amino acid sequence selected from any of the HCDR3 amino acid sequences listed in Table. 1, or a substantially similar sequence thereto having at least 90%, at least 95%, at least 98%, or at least 99% sequence identity.

The present invention also provides antibodies or antigen binding fragments thereof comprising a light chain CDR1 (LCDR1) comprising an amino acid sequence selected from any of the LCDR1 amino acid sequences listed in Table. 1, or a substantially similar sequence thereto having at least 90%, at least 95%, at least 98%, or at least 99% sequence identity.

The present invention also provides antibodies or antigen binding fragments thereof comprising a light chain CDR2 (LCDR2) comprising an amino acid sequence selected from any of the LCDR2 amino acid sequences listed in Table. 1, or a substantially similar sequence thereto having at least 90%, at least 95%, at least 98%, or at least 99% sequence identity.

The present invention also provides antibodies or antigen binding fragments thereof comprising a light chain CDR3 (LCDR3) comprising an amino acid sequence selected from any of the LCDR3 amino acid sequences listed in Table. 1 or a substantially similar sequence having at least 90%, at least 95%, at least 98%, or at least 99% sequence identity.

The present invention also provides antibodies or antigen binding fragments thereof comprising a pair of amino acid sequences HCDR3 and LCDR3 (HCDR3/LCDR3) comprising any of the amino acid sequences of HCDR3 listed in Table. 1, paired with any of the LCDR3 amino acid sequences listed in table. 1. In accordance with certain embodiments, the present invention provides antibodies or antigen binding fragments thereof comprising the HCDR3/LCDR3 amino acid sequence pair contained in any of the exemplary anti-STEP2 antibodies listed in Table. 1. In some embodiments, the HCDR3/LCDR3 amino acid sequence pair is selected from the group consisting of SEQ ID NO: 256/264 (eg, H2M11162N).

The present invention also provides antibodies or antigen binding fragments thereof comprising a set of six CDRs (ie, HCDR1-HCDR2-HCDR3-LCDR1-LCDR2-LCDR3) contained in any of the exemplary anti-STEP2 antibodies listed in Table. 1. In some embodiments, the set of amino acid sequences HCDR1-HCDR2-HCDR3-LCDR1-LCDR2-LCDR3 is selected from the group consisting of SEQ ID NO: 252-254-256-260-262-264 (eg, H2M11162N).

In a related embodiment, the present invention provides antibodies or antigen-binding fragments thereof comprising a set of six CDRs (i.e., HCDR1-HCDR2-HCDR3-LCDR1-LCDR2-LCDR3) contained within the amino acid sequence pair HCVR/LCVR, as determined by any of the exemplary anti-STEP2 antibodies listed in Table. 1. For example, the present invention includes antibodies or antigen-binding fragments thereof comprising a set of amino acid sequences HCDR1-HCDR2-HCDR3-LCDR1-LCDR2-LCDR3 contained in a pair of amino acid sequences HCVR/LCVR selected from the group consisting of SEQ ID NO: 250/ 258 (for example, H2M11162N). Methods and techniques for identifying CDRs in HCVR and LCVR amino acid sequences are well known in the art and can be used to identify CDRs in said HCVR and/or LCVR amino acid sequences described herein. Generally accepted rules that can be applied to delimit the CDR include, for example, the Kabat definition, the Chotia definition, and the AbM definition. In general, the Kabat definition is based on sequence variability, the Chotiya definition is based on the location of structural loop regions, and the AbM definition is a compromise between the Kabat and Chotiya approaches. See, for example, Kabat, Sequences of Proteins of Immunological Interest, National Institutes of Health, Bethesda, Md. (1991); Al-Lazikani et al., J. Mol. Biol. 273:927-948 (1997); and Martin et al., Proc. Natl. Acad. Sci. USA 86:9268-9272 (1989). Public databases are also available to identify CDR sequences in an antibody.

The present invention also provides nucleic acid molecules encoding anti-STEP2 antibodies or portions thereof. For example, the present invention provides nucleic acid molecules encoding any of the HCVR amino acid sequences listed in table. 1; in certain embodiments, the nucleic acid molecule comprises a polynucleotide sequence selected from any of the HCVR nucleic acid sequences listed in Table. 2, or a substantially similar sequence having at least 90%, at least 95%, at least 98%, or at least 99% sequence identity.

The present invention also provides nucleic acid molecules encoding any of the LCVR amino acid sequences listed in table. 1; in certain embodiments, the nucleic acid molecule comprises a polynucleotide sequence selected from any of the LCVR nucleic acid sequences listed in Table. 2, or a substantially similar sequence having at least 90%, at least 95%, at least 98%, or at least 99% sequence identity.

The present invention also provides nucleic acid molecules encoding any of the HCDR1 amino acid sequences listed in table. 1; in certain embodiments, the nucleic acid molecule comprises a polynucleotide sequence selected from any of the HCDR1 nucleic acid sequences listed in Table. 2, or a substantially similar sequence having at least 90%, at least 95%, at least 98%, or at least 99% sequence identity.

The present invention also provides nucleic acid molecules encoding any of the HCDR2 amino acid sequences listed in table. 1; in certain embodiments, the nucleic acid molecule comprises a polynucleotide sequence selected from any of the HCDR2 nucleic acid sequences listed in Table. 2, or a substantially similar sequence having at least 90%, at least 95%, at least 98%, or at least 99% sequence identity.

The present invention also provides nucleic acid molecules encoding any of the HCDR3 amino acid sequences listed in table. 1; in certain embodiments, the nucleic acid molecule comprises a polynucleotide sequence selected from any of the HCDR3 nucleic acid sequences listed in Table. 2, or a substantially similar sequence having at least 90%, at least 95%, at least 98%, or at least 99% sequence identity.

The present invention also provides nucleic acid molecules encoding any of the LCDR1 amino acid sequences listed in table. 1; in certain embodiments, the nucleic acid molecule comprises a polynucleotide sequence selected from any of the LCDR1 nucleic acid sequences listed in Table. 2, or a substantially similar sequence having at least 90%, at least 95%, at least 98%, or at least 99% sequence identity.

The present invention also provides nucleic acid molecules encoding any of the LCDR2 amino acid sequences listed in table. 1; in certain embodiments, the nucleic acid molecule comprises a polynucleotide sequence selected from any of the LCDR2 nucleic acid sequences listed in Table. 2, or a substantially similar sequence having at least 90%, at least 95%, at least 98%, or at least 99% sequence identity.

This invention also relates to nucleic acid molecules encoding any of the LCDR3 amino acid sequences listed in table. 1; in certain embodiments, the nucleic acid molecule comprises a polynucleotide sequence selected from any of the LCDR3 nucleic acid sequences listed in Table. 2, or a substantially similar sequence having at least 90%, at least 95%, at least 98%, or at least 99% sequence identity.

The present invention also provides nucleic acid molecules encoding HCVR, wherein HCVR contains a set of three CDRs (i.e., HCDR1-HCDR2-HCDR3), wherein the set of amino acid sequences HCDR1-HCDR2-HCDR3 is as defined by any exemplary anti- STEP2 antibodies listed in table. 1.

The present invention also provides nucleic acid molecules encoding LCVR, wherein LCVR contains a set of three CDRs (i.e., LCDR1-LCDR2-LCDR3), wherein the set of amino acid sequences LCDR1-LCDR2-LCDR3 is as defined by any exemplary anti- STEP2 antibodies listed in table. 1.

The present invention also provides nucleic acid molecules encoding both HCVR and LCVR, wherein HCVR contains the amino acid sequence of any of the HCVR amino acid sequences listed in table. 1, and wherein the LCVR contains the amino acid sequence of any of the LCVR amino acid sequences listed in table. 1. In some embodiments, the nucleic acid molecule comprises a polynucleotide sequence selected from any of the HCVR nucleic acid sequences listed in Table. 2, or a substantially similar sequence thereto having at least 90%, at least 95%, at least 98%, or at least 99% sequence identity and a polynucleotide sequence selected from any LCVR nucleic acid sequence , listed in table. 2, or a substantially similar sequence having at least 90%, at least 95%, at least 98%, or at least 99% sequence identity. In some implementations in accordance with this aspect of the invention, the nucleic acid molecule encodes HCVR and LCVR, wherein both HCVR and LCVR are derived from the same anti-STEP2 antibody specified in table. 1.

The present invention also provides a recombinant expression vector capable of expressing a polypeptide comprising a heavy or light chain variable region of an anti-STEAP2 antibody. For example, the present invention includes recombinant expression vectors containing any of the above-mentioned nucleic acid molecules, i.e. nucleic acid molecules encoding any of the HCVR, LCVR and/or CDR sequences, as indicated in table. 1. Also included within the scope of this invention are the host cells into which such vectors are introduced, as well as methods for producing antibodies or parts thereof by culturing the host cells under conditions that ensure the production of antibodies or antibody fragments, and isolating the antibodies and antibody fragments so obtained way.

This invention includes anti-STEP2 antibodies having a modified glycosylation profile. In some embodiments, modification to remove undesired glycosylation sites or an antibody lacking the fucose moiety present in the oligosaccharide chain may be useful, for example, to enhance the function of antibody-dependent cellular cytotoxicity (ADCC) (see Shield et al. (2002) JBC 277:26733 ). In other applications, galactosylation modification can be made in order to alter complement-dependent cytotoxicity (CDC).

In another aspect, the present invention provides a pharmaceutical composition comprising a recombinant human antibody or fragment thereof that specifically binds STEAP2, and a pharmaceutically acceptable carrier. In a related aspect, the invention provides a composition that is a combination of an anti-STEP2 antibody and a second therapeutic agent. In one embodiment, the second therapeutic agent is any agent that is advantageously combined with an anti-STEP2 antibody.

Additional combination therapies and combination formulations comprising anti-STEP2 antibodies of this invention are described elsewhere herein.

In another aspect, the present invention provides therapeutic methods for targeting/killing tumor cells expressing STEAP2 using an anti-STEP2 antibody of the invention, the therapeutic methods comprising administering a therapeutically effective amount of a pharmaceutical composition comprising an anti-STEP2 antibody of the invention to a subject in need of this. In some cases, anti-STEP2 antibodies (or antigen-binding fragments thereof) may be used to treat prostate cancer or may be modified to be more cytotoxic by methods including, but not limited to, modified Fc domains to increase ADCC (see ., for example, Shield et al (2002) JBC 277: 26733), radioimmunotherapy, antibody-drug conjugates, or other methods for increasing the efficiency of tumor ablation.

The present invention also includes the use of an anti-STEAP2 antibody of the invention in the manufacture of a medicament for the treatment of a disease or disorder (eg, cancer) associated with or caused by STEAP2-expressing cells. In one aspect, the invention provides a compound comprising an aHTu-STEP2 antibody or an antigen binding fragment thereof, or a STEAP2xCD3 bispecific antibody as described herein, for use in medicine. In one aspect, the invention relates to a compound containing an antibody-drug conjugate (ADC), as described herein, for use in medicine.

In yet another aspect, the present invention provides monospecific anti-STEP2 antibodies for diagnostic applications, such as, for example, imaging reagents.

In yet another aspect, the invention provides therapeutic methods for promoting T cell activation using an anti-CD3 antibody or an antigen binding portion of an antibody of the invention, the therapeutic methods comprising administering a therapeutically effective amount of a pharmaceutical composition comprising the antibody.

In another aspect, the present invention provides an isolated antibody or antigen binding fragment thereof that binds STEP2-expressing C4-2 cells with an EC50 of less than 50 nM, as measured by FACS analysis. In another aspect, the present invention provides an isolated antibody or antigen-binding fragment thereof that binds to and is internalized by STEP2-expressing C4-2 cells.

The invention further provides an antibody or antigen-binding fragment that competes for binding to human STEAP2 with a reference antibody containing the amino acid sequence pair HCVR/LCVR, as set forth in Table. 1. In another aspect, the invention provides an antibody or antigen binding fragment that competes for binding to human STEAP2 with a control antibody comprising the amino acid sequence pair HCVR/LCVR selected from the group consisting of SEQ ID NO: 2/10; 18/26; 34/42; 50/58; 66/58; 74/58; 82/58; 90/58; 98/58; 106/114; 122/130; 138/146; 154/162; 170/178; 186/194; 202/210; 218/226; 234/242; 250/258; 266/274; 282/290; 298/306; 314/322; 330/338; 346/354; 362/370; and 378/386.

The invention further provides an antibody or antigen-binding fragment, wherein the antibody or antigen-binding fragment thereof binds to the same epitope on human STEAP2 as the reference antibody comprising the amino acid sequence pair HCVR/LCVR as set forth in Table. 1. In another aspect, the antibody or antigen binding fragment thereof binds to the same epitope on human STEAP2 as the reference antibody comprising the amino acid sequence pair HCVR/LCVR selected from the group consisting of SEQ ID NO: 2/10; 18/26; 34/42; 50/58; 66/58; 74/58; 82/58; 90/58; 98/58; 106/114; 122/130; 138/146; 154/162; 170/178; 186/194; 202/210; 218/226; 234/242; 250/258; 266/274; 282/290; 298/306; 314/322; 330/338; 346/354; 362/370; and 378/386.

The invention also provides an isolated antibody or antigen-binding fragment thereof that binds human STEAP2, wherein the antibody or antigen-binding fragment comprises: heavy chain variable region (HCVR) complementarity determining regions (CDRs) having the amino acid sequence shown in table. 1; and a light chain variable region (LCVR) CDR having an amino acid sequence as shown in table. 1. In another aspect, the isolated antibody or antigen binding fragment comprises a heavy and light chain CDR of a pair of amino acid sequences HCVR/LCVR selected from the group consisting of: 2/10; 18/26; 34/42; 50/58; 66/58; 74/58; 82/58; 90/58; 98/58; 106/114; 122/130; 138/146; 154/162; 170/178; 186/194; 202/210; 218/226; 234/242; 250/258; 266/274; 282/290; 298/306; 314/322; 330/338; 346/354; 362/370; and 378/386. In yet another aspect, the isolated antibody or antigen binding fragment thereof comprises HCDR1-HCDR2-HCDR3-LCDR1-LCDR2-LCDR3 domains, respectively, selected from the group consisting of: SEQ ID NO: 4-6-8-12-14-16; 20-22-24-28-30-32; 36-38-40-44-46-48; 52-54-56-60-62-64; 68-70-72-60-62-64; 76-78-80-60-62-64; 84-8 6-88-60-62-64; 92-94-96-60-62-64; 100-102-104-60-62-64; 108-110-112-116-118-120; 124-12 6-128-132-134-136; 140-142-144-148-150-152; 156-158-160-164-166-168; 172-174-176-180-182-184; 188-190-192-19 6-198-200; 204-206-208-212-214-216; 220-222-224-228-230-232; 236-238-240-244-24 6-248; 252-254-256-2 60-262-264; 268-270-272-276-278-280; 284-286-288-292-294-296; 300-302-304-308-310-312; 316-318-320-324-32 6-328; 332-334-336-340-342-344; 348-350-352-356-358-360; 364-366-368-372-374-376; and 380-382-384-388-390-392.

In another aspect, the invention provides an isolated antibody or antigen binding fragment thereof that binds human STEAP2, wherein the antibody or antigen binding fragment comprises: (a) a heavy chain variable region (HCVR) having an amino acid sequence selected from the group consisting of SEQ ID NO: 2 , 18, 34, 50, 66, 74, 82, 90, 98, 106, 122, 138, 154, 170, 186, 202, 218, 234, 250, 266, 282, 298, 314, 330, 346, 362 , and 378; and (b) a light chain variable region (LCVR) having an amino acid sequence selected from the group consisting of SEQ ID NO: 10; 26; 42; 58; 114; 130; 146; 162; 178; 194; 210; 226, 242; 258; 274; 290; 306; 322; 338; 354; 370; and 386. In a further aspect, the isolated antibody or antigen binding fragment of claim 10, wherein the antibody or antigen binding fragment comprises an HCVR/LCVR amino acid sequence pair selected from the group consisting of: SEQ ID NO: 2/10; 18/26; 34/42; 50/58; 66/58; 74/58; 82/58; 90/58; 98/58; 106/114; 122/130; 138/146; 154/162; 170/178; 186/194; 202/210; 218/226; 234/242; 250/258; 266/274; 282/290; 298/306; 314/322; 330/338; 346/354; 362/370 and 378/386.

In accordance with another aspect, the present invention provides antibody-drug conjugates comprising an anti-STEP2 antibody or antigen-binding moiety and a therapeutic agent (eg, a cytotoxic agent). In some embodiments, the antibody or antigen binding fragment thereof and the cytotoxic agent are covalently linked through a linker, as discussed herein. In various embodiments, the anti-STEAP 2 antibody or antigen binding fragment thereof may be any of the anti-STEAP 2 antibodies or fragments described herein.

In some embodiments, the cytotoxic agent is selected from auristatin, a maytansinoid, tubulisin, a tomaimycin derivative, or a dolastatin derivative. In some cases, the cytotoxic agent is an auristatin selected from MMAE or MMAF, or a maytansinoid selected from DM1 or DM4. In some embodiments, the cytotoxic agent is a maytansinoid having the structure of formula (I) or formula (II), as discussed herein.

In some embodiments, the cytotoxic agent is a maytansinoid having the structure:

In some embodiments, the cytotoxic agent is a maytansinoid having the structure:

o _o N ON.R ^S Nz ^CH ·

H ₃ C' CH ₃ HN. /x.N. Ah μ - O O CH ₃

OSN ₃

C.I.

In some embodiments, the antibody-drug conjugate comprises a STEAP 2 antibody or fragment thereof, and

where § represents a link to an anti-STEAP 2 antibody or fragment thereof.

In some embodiments, the antibody-drug conjugate comprises a STEAP 2 antibody or fragment thereof, and

where * represents a link to an anti-STEAP 2 antibody or fragment thereof.

In some embodiments, the antibody-drug conjugate comprises a STEAP 2 antibody or fragment thereof, and an anti-anti-antibody.

where " represents a linkage to an anti-STEAP 2 antibody or fragment thereof.

- 6 044194

In some embodiments, the association with the antibody or fragment thereof is through the sulfur moiety of the cysteine residue.

In some embodiments, the antibody-drug conjugate comprises an anti-STEAP 2 antibody or fragment thereof, and

o or

o, or a mixture thereof, where it represents a bond with an anti-STEAP 2 antibody or a fragment thereof.

In some embodiments, the association with the antibody or fragment thereof is through the nitrogen moiety of the lysine residue.

In any of the various embodiments of the antibody-drug conjugates described above or herein, the antibody-drug conjugate may contain from 1 to 4 cytotoxic agents per anti-STEAP 2 antibody or fragment thereof.

In accordance with another aspect, the present invention provides bispecific antigen-binding molecules (eg, antibodies) that bind STEAP2 and CD3. Such bispecific antigen binding molecules are also referred to herein as anti-STEP2/anti-CD3 bispecific molecules, anti-CD3/anti-STEP2 bispecific molecules or STEAP2xCD3 bAb (bispecific antibody). The Ahtu-STEP2 portion of the anti-STEP2/anti-CD3 bispecific molecule is useful for targeting cells (eg, tumor cells) that express STEAP2 (eg, prostate tumors), and the anti-CD3 portion of the bispecific molecule is useful for activating T cells. Simultaneous binding of STEAP2 on the tumor cell and CD3 on the T cell facilitates targeted killing (cell lysis) of the target tumor cell by the activated T cell. Accordingly, the anti-STEP2/αanti-CD3 bispecific molecules of the invention are therefore useful for the treatment of diseases and disorders associated with or caused by STEP2-expressing tumors (eg, prostate cancer).

Bispecific antigen binding molecules in accordance with this aspect of the invention comprise a first antigen binding domain that specifically binds human CD3 and a second antigen binding domain that specifically binds STEAP2. The present invention includes anti-STEAP2/anti-CD3 bispecific molecules (eg, bispecific antibodies) in which each antigen binding domain contains a heavy chain variable region (HCVR) paired with a light chain variable region (LCVR). In some illustrative embodiments, the anti-CD3 antigen binding domain and the anti-STEP2 antigen binding domain each comprise a different, defined HCVR paired with a common LCVR. For example, as shown in Example 4 herein, bispecific antibodies were constructed containing a first antigen binding domain that specifically binds CD3, wherein the first antigen binding domain comprises an HCVR derived from an anti-CD3 antibody coupled to an LCVR derived from an anti-STEP2 antibody (eg, the same LCVR that is included in the anti-STEP2 antigen binding domain); and a second antigen binding domain that specifically binds STEAP2, the second antigen binding domain comprising HCVR/LCVR derived from an anti-STEP2 antibody. In other words, in the exemplary molecules described herein, pairing of HCVR from an anti-CD3 antibody with LCVR from an anti-STEAP2 antibody creates an antigen binding domain that specifically binds CD3 (but does not bind STEAP2). In such embodiments, the first and second antigen binding domains comprise different anti-CD3 and α-anti-STEP2 HCVRs, but share a common anti-STEP2 LCVR. In other embodiments, the bispecific antigen binding molecules contain different anti-CD3 and α-anti-STEP2 HCVRs, but have a common LCVR. The amino acid sequence of this LCVR is shown, for example, in SEQ ID NO: 18-90, and the amino acid sequences of the corresponding CDRs (ie, LCDR1-LCDR2-LCDR3) are shown in SEQ ID NO: 1892, 1894 and 1896, respectively . Genetically modified mice can be used to produce fully human bispecific antigen binding molecules containing two different heavy chains that are linked to an identical light chain that contains a variable domain derived from one of two different human light chain variable region gene segments. Alternatively, the heavy chain variable regions can be paired with a single common light chain and expressed recombinantly in host cells. Thus, the antibodies of the invention may contain immunoglobulin heavy chains linked to a single rearranged light chain. In some embodiments, the light chain comprises a variable domain derived from a human Vk1-39 gene segment or a Vk3-20 gene segment. In other embodiments, the light chain comprises a variable domain derived from a human Vk1-39 gene segment rearranged with a human Jk5 gene segment or a human Jk1 gene segment.

This invention provides anti-CD3/anti-STEP2 bispecific molecules, wherein the first antigen binding domain that specifically binds CD3 comprises any of the HCVR amino acid sequences, any of the LCVR amino acid sequences, any of the HCVR/LCVR amino acid sequence pairs, any of the HCVR/LCVR amino acid sequences Heavy chain CDR1-CDR2-CDR3 or any of the light chain amino acid sequences CDR1CDR2-CDR3 as set forth in US Publication 2014/0088295 published March 27, 2014 and PCT/US2016/044732 filed July 29, 2016.

In addition, the present invention provides anti-CD3/anti-STEAP2 bispecific molecules, characterized in that the first antigen binding domain, which specifically binds CD3, contains any of the HCVR amino acid sequences as set forth in Tables 9, 11 and 15 herein. The first antigen binding domain, which specifically binds CD3, may also comprise any of the LCVR amino acid sequences as set forth in Tables 1, 9, 12 and 17 herein. In certain embodiments, the first antigen binding domain that specifically binds CD3 comprises any of the HCVR/LCVR amino acid sequence pairs as set forth in Table. 9, 11, 12, 15 and 17 in this document. The present invention also provides anti-CD3/anti-STEAP2 bispecific molecules, wherein the first antigen binding domain that specifically binds CD3 comprises any of the heavy chain amino acid sequences CDR1-CDR2-CDR3 as set forth in Table. 9, 11, and 15 herein, and/or any of the light chain amino acid sequences CDR1-CDR2-CDR3 as set forth in Tables 1, 9, 12, and 17 herein.

In accordance with some embodiments, the present invention provides anti-CD3/anti-STEP2 bispecific molecules, wherein the first antigen binding domain that specifically binds CD3 comprises a heavy chain variable region (HCVR) having the amino acid sequence as set forth in Table. 9, 11 and 15 herein, or a substantially similar sequence thereto having at least 90%, at least 95%, at least 98%, or at least 99% sequence identity.

The present invention also provides anti-CD3/anti-STEAP2 bispecific molecules, wherein the first antigen binding domain that specifically binds CD3 comprises a light chain variable region (LCVR) having the amino acid sequence as set forth in Table 1. 1, 9, 12 and 17 herein, or a substantially similar sequence thereto having at least 90%, at least 95%, at least 98%, or at least 99% sequence identity.

The present invention also provides anti-CD3/anti-STEAP2 bispecific molecules, wherein the first antigen binding domain that specifically binds CD3 contains a pair of amino acid sequences HCVR and LCVR (HCVR/LCVR) as set forth in Tables 9, 11, 12, 15 and 17 in this document.

The present invention also provides anti-CD3/anti-STEAP2 bispecific molecules, wherein the first antigen binding domain that specifically binds CD3 comprises a heavy chain CDR3 domain (HCDR3) having the amino acid sequence as set forth in Table 1. 9, 11 and 15 herein, or a substantially similar sequence thereto having at least 90%, at least 95%, at least 98%, or at least 99% sequence identity; and a light chain CDR3 domain (LCDR3) having an amino acid sequence as shown in table. 1, 9, 12 and 17 herein, or a substantially similar sequence thereto having at least 90%, at least 95%, at least 98%, or at least 99% sequence identity.

In some embodiments, the first antigen binding domain that specifically binds CD3 comprises the amino acid sequence pair HCDR3/LCDR3, as set forth in

- 8 044194 table. 9, 11, 12, 15 and 17 in this document.

The present invention also provides aHTu-CD3/aHTu-STEAP2 bispecific antigen binding molecules, wherein the first antigen binding domain that specifically binds CD3 comprises a heavy chain CDR1 domain (HCDR1) having the amino acid sequence as set forth in Table 1. 9, 11 and 15 herein, or a substantially similar sequence thereto having at least 90%, at least 95%, at least 98%, or at least 99% sequence identity; heavy chain CDR2 domain (HCDR2) having an amino acid as indicated in table. 9, 11 and 15, or a substantially similar sequence thereof having at least 90%, at least 95%, at least 98%, or at least 99% sequence identity; heavy chain CDR3 domain (HCDR3) having an amino acid as indicated in table. 9, 11 and 15, or a substantially similar sequence thereof having at least 90%, at least 95%, at least 98%, or at least 99% sequence identity; light chain CDR1 (LCDR1) domain having the amino acid sequence as indicated in table. 1, 9, 12 and 17 herein, or a substantially similar sequence thereto having at least 90%, at least 95%, at least 98%, or at least 99% sequence identity; light chain CDR2 domain (LCDR2) having the amino acid sequence as indicated in table. 1, 9, 12 and 17 herein, or a substantially similar sequence having at least 90%, at least 95%, at least 98%, or at least 99% sequence identity; and a light chain CDR3 domain (LCDR3) having an amino acid sequence as shown in table. 1, 9, 12 and 17 herein, or a substantially similar sequence having at least 90%, at least 95%, at least 98%, or at least 99% sequence identity.

Some non-limiting, illustrative anti-CD3/anti-STEAP2 bispecific antigen binding molecules of the invention comprise a first antigen binding domain that specifically binds CD3 comprising the domains HCDR1-HCDR2-HCDR3-LCDR1-LCDR2-LCDR3, respectively, having the amino acid sequences as set forth in Table 1 . 9, 11, 12, 15 and 17 in this document.

The present invention also provides a bispecific antigen binding molecule, wherein the first antigen binding domain, which specifically binds human CD3, comprises heavy chain complementarity determining regions (HCDR1, HCDR2 and HCDR3) from a heavy chain variable region (HCVR) containing the amino acid sequence as set forth in Table 1. 9, table. 11, or table. 15, and the light chain complementarity determining regions (LCDR1, LCDR2 and LCDR3) from the light chain variable region (LCVR) containing the amino acid sequence as shown in table. 1, table. 9, table. 12 or table 17.

In another aspect, the present invention provides a bispecific antigen binding molecule, wherein the first antigen binding domain that specifically binds human CD3 comprises heavy chain complementarity determining regions (HCDR1, HCDR2 and HCDR3) from a heavy chain variable region (HCVR) selected from the group consisting of SEQ ID NOs: 1730, 1762 and 1866, and light chain complementarity determining regions (LCDR1, LCDR2 and LCDR3) from a light chain variable region (LCVR) containing an amino acid sequence selected from the group consisting of SEQ ID NO: 258.

The invention further provides a bispecific antigen binding molecule, wherein the first antigen binding domain, which specifically binds human CD3, contains three heavy chain complementarity determining regions (A1-HCDR1, A1-HCDR2 and A1-HCDR3) and three light chain complementarity determining regions (A1-HCDR3). LCDR1, A1-LCDR2 and A1-LCDR3), wherein A1HCDR1 contains an amino acid sequence selected from the group consisting of SEQ ID NO: 1732, 1764, and 1868; A1-HCDR2 contains an amino acid sequence selected from the group consisting of SEQ ID NO: 1734, 1766, and 1870; A1-HCDR3 contains an amino acid sequence selected from the group consisting of SEQ ID NO: 1736, 1768 and 1872; A1-LCDR1 contains the amino acid sequence SEQ ID NO: 260; A1-LCDR2 contains the amino acid sequence SEQ ID NO: 262; and A1-LCDR3 contains the amino acid sequence of SEQ ID NO: 264.

In yet another aspect, the present invention provides a bispecific antigen binding molecule, wherein the first antigen binding domain that specifically binds human CD3 comprises a heavy and light chain CDR of the HCVR/LCVR amino acid sequence pair selected from the group consisting of: SEQ ID NO: 1730/258 , 1762/258, and 1866/258.

In another aspect, the present invention provides a bispecific antigen binding molecule, wherein the first antigen binding domain that specifically binds human CD3 comprises three heavy chain complementarity determining regions (A1-HCDR1, A1-HCDR2 and A1-HCDR3) and three light chain complementarity determining regions (A1 -LCDR1, A1-LCDR2 and A1-LCDR3), and wherein the second antigen binding domain, which specifically binds human STEAP2, contains three heavy chain complementarity determining regions (A2-HCDR1, A2-HCDR2 and A2-HCDR3) and three light chain complementarity determining regions (A2-LCDR1, A2-LCDR2 and A2-LCDR3); and A1

- 9 044194

HCDR1 contains an amino acid sequence selected from the group consisting of SEQ ID NO: 1732, 1764 and 1868; A1-HCDR2 contains an amino acid sequence selected from the group consisting of SEQ ID NO: 1736, 1768 and 1872; A1-HCDR3 contains an amino acid sequence selected from the group consisting of SEQ ID NO: 1736, 1768 and 1872; A1-LCDR1 contains the amino acid sequence SEQ ID NO: 260; A1-LCDR2 contains the amino acid sequence SEQ ID NO: 262; and A1-LCDR3 contains the amino acid sequence of SEQ ID NO: 264; and wherein A2-HCDR1 contains the amino acid sequence of SEQ ID NO: 252; A2-HCDR2 contains the amino acid sequence SEQ ID NO: 254; A2-HCDR3 contains the amino acid sequence SEQ ID NO: 256; A2-LCDR1 contains the amino acid sequence SEQ ID NO: 260; A2-LCDR2 contains the amino acid sequence SEQ ID NO: 262; and A2-LCDR3 contains the amino acid sequence of SEQ ID NO: 264.

Some non-limiting, illustrative anti-CD3/anti-STEAP2 bispecific antigen binding molecules of the invention comprise a first antigen binding domain that specifically binds CD3, comprising a heavy chain containing variable domain frameworks having an amino acid sequence selected from FR1 (SEQ ID NO: 1903) , FR2 (SEQ ID NO: 1904), FR3 (SEQ ID NO: 1905) and FR4 (SEQ ID NO: 1906).

In several embodiments, exemplary anti-CD3/anti-ETEAR2 bispecific antigen binding molecules of the invention comprise a bispecific antigen binding molecule, wherein the first antigen binding domain that specifically binds human CD3 comprises an HCVR comprising HCDR1-HCDR2-HCDR3 having the amino acid sequences of SEQ ID NO: 1907-1908-1909.

The present invention also provides anti-CD3/anti-STEAP2 bispecific molecules, wherein the second antigen binding domain that specifically binds STEAP2 comprises a heavy chain variable region (HCVR) having an amino acid sequence selected from the group consisting of SEQ ID NO: 2, 18 , 34, 50, 66, 74, 82, 90, 98, 106, 122, 138, 154, 170, 186, 202, 218, 234, 250, 266, 282, 298, 314, 330, 346, 362 and 378 , or a substantially similar sequence thereof having at least 90%, at least 95%, at least 98%, or at least 99% sequence identity.

The present invention also provides anti-CD3/anti-STEAP2 bispecific molecules, wherein the second antigen binding domain that specifically binds STEAP2 comprises a light chain variable region (LCVR) having an amino acid sequence selected from the group consisting of SEQ ID NO: 10; 26; 42; 58; 114; 130; 146; 162; 178; 194; 210; 226, 242; 258; 274; 290; 306; 322; 338; 354; 370; and 386, or a substantially similar sequence thereof, having at least 90%, at least 95%, at least 98%, or at least 99% sequence identity.

The present invention also provides anti-CD3/anti-STEAP2 bispecific molecules, wherein the second antigen binding domain, which specifically binds STEAP2, contains a pair of amino acid sequences HCVR and LCVR (HCVR/LCVR) SEQ ID NO: 250/258.

The present invention also provides anti-CD3/anti-STEAP2 bispecific molecules, wherein the second antigen binding domain that specifically binds STEAP2 comprises a heavy chain CDR3 (HCDR3) domain having the amino acid sequence of SEQ ID NO: 256, or a substantially similar sequence thereof having at least 90%, at least 95%, at least 98%, or at least 99% sequence identity; and a light chain CDR3 (LCDR3) domain having the amino acid sequence of SEQ ID NO: 264, or a substantially similar sequence thereof, having at least 90%, at least 95%, at least 98%, or at least 99% sequence identity.

In some embodiments, the second antigen binding domain that specifically binds STEAP2 comprises the amino acid sequence pair HCDR3/LCDR3 selected from the group consisting of SEQ ID NO: 256/264.

The present invention also provides anti-CD3/anti-STEAP2 bispecific antigen binding molecules, wherein the second antigen binding domain that specifically binds STEAP2 comprises a heavy chain CDR1 domain (HCDR1) having an amino acid sequence selected from the group consisting of SEQ ID NO: 4, 20, 36, 52, 68, 76, 84, 92, 100, 108, 124, 140, 156, 172, 188, 204, 220, 236, 252, 268, 284, 300, 316, 332, 348, 364, and 380, or a substantially similar sequence thereto, having at least 90%, at least 95%, at least 98%, or at least 99% sequence identity; a heavy chain CDR2 (HCDR2) domain having an amino acid sequence selected from the group consisting of SEQ ID NO: 6, 22, 38, 54, 70, 78, 86, 94, 102, 110, 126, 142, 158, 174, 190, 206, 222, 238, 254, 270, 286, 302, 318, 334, 350, 366 and 382, or a substantially similar sequence having at least 90%, at least 95%, at least 98 % or at least 99% sequence identity; a heavy chain CDR3 (HCDR3) domain having an amino acid sequence selected from the group consisting of SEQ ID NO: 8, 24, 40, 56, 72, 80, 88, 96, 104, 112, 128, 144, 160, 176, 182, 208, 224, 240, 256,

- 10 044194

272, 288, 304, 320, 336, 352, 368, and 384, or a substantially similar sequence having at least 90%, at least 95%, at least 98%, or at least 99% sequence identity ; light chain CDR1 (LCDR1) domain having an amino acid sequence selected from the group consisting of SEQ ID NO: 12, 28, 44, 60, 116, 132, 148, 164, 180, 196, 212, 228, 244, 260, 276, 292, 308, 324, 340, 356, 372 and 38 8, or a substantially similar sequence having at least 90%, at least 95%, at least 98%, or at least 99% sequence identity ; and a light chain CDR2 domain (LCDR2) having an amino acid sequence selected from the group consisting of SEQ ID NO: 14, 30, 46, 62, 118, 134, 150, 166, 182, 198, 214, 230, 246, 262 , 278, 294, 310, 326, 342, 358, 374, and 390, or a substantially similar sequence having at least 90%, at least 95%, at least 98%, or at least 99% identity sequences; and a light chain CDR3 (LCDR3) domain having an amino acid sequence selected from the group consisting of SEQ ID NO: 16, 32, 48, 64, 120, 136, 152, 168, 184, 200, 216, 232, 248, 264 , 280, 296, 312, 328, 344, 360, 376, and 392, or a substantially similar sequence having at least 90%, at least 95%, at least 98%, or at least 99% sequence identity .

Some non-limiting, illustrative anti-CD3/αanti-STEAP2 bispecific antigen binding molecules of the invention comprise a second antigen binding domain that specifically binds STEAP2, comprising HCDR1-HCDR2-HCDR3-LCDR1-LCDR2-LCDR3 domains, respectively, having amino acid sequences selected from the group, consisting of: SEQ ID NO: 252-254-256-260-262-264.

In a related embodiment, the invention includes anti-CD3/anti-STEP2 bispecific antigen binding molecules, wherein the second antigen binding domain that specifically binds STEAP2 comprises heavy and light chain CDR domains contained in heavy and light chain variable region (HCVR/LCVR) sequences. selected from the group consisting of SEQ ID NO: 250/258.

In another aspect, the present invention provides a bispecific antigen binding molecule comprising a first antigen binding domain that binds human CD3 and a second antigen binding domain that binds human STEAP2, wherein the second antigen binding domain is derived from an antibody or antigen binding fragment of any of the anti-STEP2 antibodies of the invention. In a further aspect, the invention provides a bispecific antigen binding molecule comprising a first antigen binding domain that specifically binds human CD3 and a second antigen binding domain that specifically binds human STEAP2.

The invention further provides a bispecific antigen-binding molecule that binds human cells expressing human CD3 and cynomolgus cells expressing cynomolgus CD3. In another aspect, a bispecific antigen binding molecule binds human cells expressing human STEAP2.

In another aspect, the present invention provides a bispecific antigen-binding molecule that inhibits tumor growth in immunocompromised mice bearing human prostate cancer xenografts.

In some embodiments, the anti-CD3 antibodies of the invention, antigen binding fragments, and bispecific antibodies thereof were made by replacing amino acid residues of the original in a stepwise manner based on differences between the germline sequence and the sequence of the original antibody.

In some embodiments, the present invention provides a bispecific antigen binding molecule, wherein the second antigen binding domain competes for binding to human STEAP2 with a reference antigen binding protein containing three heavy chain complementarity determining regions (A2-HCDR1, A2-HCDR2 and A2-HCDR3) and three complementarity determining regions light chain regions (A2-LCDR1, A2-LCDR2 and A2-LCDR3), wherein A2-HCDR1 contains an amino acid sequence selected from the group consisting of SEQ ID NO: 252; A2-HCDR2 contains the amino acid sequence SEQ ID NO: 254; A2-HCDR3 contains the amino acid sequence SEQ ID NO: 256; A2-LCDR1 contains the amino acid sequence SEQ ID NO: 260; A2-LCDR2 contains the amino acid sequence SEQ ID NO: 262; and A2-LCDR3 comprises the amino acid sequence of SEQ ID NO: 264. In some embodiments, the present invention provides a bispecific antigen binding molecule, wherein the second antigen binding domain competes for binding to human STEAP2 with a reference antigen binding protein containing a heavy chain variable region (HCVR) containing the amino acid the sequence of SEQ ID NO: 250, and a light chain variable region (LCVR) containing the amino acid sequence of SEQ ID NO: 258.

In some embodiments, the invention provides a bispecific antigen binding molecule, wherein the first antigen binding domain competes for binding to human CD3 with a reference antigen binding protein containing three complementarity determining proteins

- 11 044194 heavy chain regions (A1-HCDR1, A1-HCDR2 and A1-HCDR3) and three light chain complementarity determining regions (A1-LCDR1, A1-LCDR2 and A1-LCDR3), wherein A1-HCDR1 contains an amino acid sequence selected from the group consisting of SEQ ID NOs: 1732, 1764 and 1868; A1-HCDR2 contains an amino acid sequence selected from the group consisting of SEQ ID NO: 1734, 1766 and 1870; A1-HCDR3 contains an amino acid sequence selected from the group consisting of SEQ ID NO: 1736, 1768 and 1872; A1-LCDR1 contains the amino acid sequence SEQ ID NO: 260; A1-LCDR2 contains the amino acid sequence SEQ ID NO: 262; and A1-LCDR3 comprises the amino acid sequence of SEQ ID NO: 264. In some embodiments, the invention provides a bispecific antigen binding molecule, wherein the first antigen binding domain competes for binding to human CD3 with a reference antigen binding protein containing a heavy chain variable region (HCVR) containing the amino acid sequence , selected from the group consisting of SEQ ID NO: 1730, 1762 and 1866, and a light chain variable region (LCVR) containing the amino acid sequence of SEQ ID NO: 258.

In some embodiments, the invention provides a bispecific antigen binding molecule, wherein the first antigen binding domain competes for binding to human CD3 with a reference antigen binding protein containing a heavy chain variable region (HCVR) comprising an amino acid sequence selected from the group consisting of SEQ ID NO: 1730, 1762 and 1866, and a light chain variable region (LCVR) containing the amino acid sequence SEQ ID NO: 258; and wherein the second antigen binding domain competes for binding to human STEAP2 with a reference antigen binding protein comprising a heavy chain variable region (HCVR) comprising the amino acid sequence of SEQ ID NO: 250 and a light chain variable region (LCVR) comprising the amino acid sequence of SEQ ID NO: 250 : 258.

In one aspect, the present invention provides a pharmaceutical composition comprising an anti-STEP2 antigen binding molecule or an anti-STEP2/anti-CD3 bispecific antigen binding molecule and a pharmaceutically acceptable carrier or diluent. The invention also provides a method of treating cancer in a subject, comprising administering to the subject a pharmaceutical composition comprising an anti-STEP2 antigen binding molecule or an anti-STEP2/anti-CD3 bispecific antigen binding molecule, and a pharmaceutically acceptable carrier or diluent. In some embodiments, the cancer is selected from the group consisting of prostate cancer, bladder cancer, cervical cancer, lung cancer, colon cancer, kidney cancer, breast cancer, pancreatic cancer, stomach cancer, uterine cancer, and ovarian cancer. In some cases, the cancer is prostate cancer. In some cases, prostate cancer is castration-resistant prostate cancer.

In another aspect, the present invention provides nucleic acid molecules encoding any of the HCVR, LCVR or CDR sequences of the anti-CD3/anti-STEP2 bispecific antigen binding molecules described herein, including nucleic acid molecules containing polynucleotide sequences as set forth in Table. 2, 10, 13, 14, 16 and 18 herein, as well as nucleic acid molecules containing two or more of the polynucleotide sequences as indicated in table. 2, 10, 13, 14, 16 and 18 in any functional combination or arrangement. Recombinant expression vectors having the nucleic acids of the invention and the host cells into which such vectors have been introduced are also included in the invention, as are methods for producing antibodies by culturing the host cells under antibody production conditions and isolating the resulting antibodies.

This invention includes anti-CD3/anti-STEP2 bispecific antigen-binding molecules, wherein any of the above antigen-binding domains that specifically binds CD3 combines, joins or otherwise binds to any of the above-mentioned antigen-binding domains that specifically binds STEAP2 to form a bispecific antigen-binding molecule. molecule that binds CD3 and STEAP2.

The present invention includes anti-CD3/anti-STEP2 bispecific antigen binding molecules having a modified glycosylation profile. In some embodiments, modification to remove undesired glycosylation sites or an antibody lacking the fucose moiety present in the oligosaccharide chain may be useful, for example, to enhance the function of antibody-dependent cellular cytotoxicity (ADCC) (see Shield et al. (2002) JBC 277:26733 ). In other applications, galactosylation modification can be made in order to alter complement-dependent cytotoxicity (CDC).

In another aspect, the present invention provides a pharmaceutical composition comprising an anti-CD3/anti-STEP2 bispecific antigen binding molecule as described herein and a pharmaceutically acceptable carrier. In a related aspect, the present invention provides a composition that is a combination of an anti-CD3/anti-STEP2 bispecific antigen binding molecule and a second therapeutic agent. In one embodiment, the second

- 12 044194 a therapeutic agent is any agent that is advantageously combined with an anti-Tu-CD3/aHTu-STEP2 bispecific antigen-binding molecule. Exemplary agents that may advantageously be combined with an anti-CD3/anti-STEAP2 bispecific antigen binding molecule are discussed in detail elsewhere herein.

In yet another aspect, the present invention provides therapeutic methods for targeting/killing tumor cells expressing STEAP2 using an anti-CD3/anti-STEAP2 bispecific antigen binding molecule of the invention, the therapeutic methods comprising administering a therapeutically effective amount of a pharmaceutical composition comprising anti-CD3/ an anti-STEP2 bispecific antigen binding molecule according to the invention, to a subject in need thereof.

The present invention also includes the use of the anti-CD3/anti-STEAP2 bispecific antigen binding molecule of the invention in the manufacture of a medicament for the treatment of a disease or disorder associated with or caused by STEP2-expressing cells.

Other embodiments will become apparent from a review of the following detailed description.

Brief description of graphic materials

In fig. Figure 1 depicts the efficacy of H1H7814N-7 in a STEP2-positive prostate cancer xenograft model (SCID mice implanted with C4-2 cells) at a dose of 10, 20, or 40 mg/kg H1H7814N-7 administered on day 13 postimplantation.

In fig. Figure 2 depicts the efficacy of H1H7814N-7 in a STEP2-positive prostate cancer xenograft model (SCID mice implanted with C4-2 cells) at a dose of 20 mg/kg H1H7814N-7 administered on day 14 postimplantation.

In fig. Figure 3 depicts the efficacy of H1H7814N-7 in a STEP2-positive prostate cancer xenograft model (SCID mice implanted with C4-2 cells) at a dose of 150 μg/kg H1H7814N-7 administered on day 17 postimplantation.

In fig. Figure 4 depicts the efficacy of H1H7814N-60 in a STEP2-positive prostate cancer xenograft model (SCID mice implanted with C4-2 cells) at a dose of 2.5 mg/kg (DAR 3.6 TV) of H1H7814N-60 administered on day 29 postimplantation.

In fig. Figure 5 depicts the binding of STEAP2xCD3 bispecific antibodies to Jurkat cells.

In fig. 6 depicts the binding of STEAP2xCD3 bispecific antibodies to a human prostate cancer cell line (PC3) engineered to express a chimeric STEAP2/1 construct.

In fig. 7 and 8 depict the binding of STEAP2xCD3 bispecific antibodies to cynomolgus monkey T cells.

In fig. Figure 9 depicts the induction of human PBMC proliferation by STEAP2xCD3 bispecific antibodies.

In fig. 10 shows the induction of proliferation of cynomolgus monkey PBMCs by STEAP2xCD3 bispecific antibodies.

In fig. 11 depicts depletion of C4-2 cells (target cells bearing STEAP2) in a cytotoxicity assay using representative STEAP2xCD3 bispecific antibodies in the presence of human PBMCs.

In fig. 12 depicts activation of human T cells by exemplary STEAP2xCD3 bispecific antibodies, which correlates with the observed target cell lysis depicted in FIG. eleven.

Detailed description of the invention

Before describing the present invention, it should be understood that the present invention is not limited to the specific methods described and the experimental conditions described, since such methods and conditions may vary. It should also be understood that the terminology used herein is intended to describe specific embodiments of the invention only and is not intended to be limiting, as the scope of the invention will be limited only by the appended claims.

Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by one skilled in the art to which this invention relates. As used herein, the term about, when used in relation to a specific numerical value given, means that the value may differ from the given value by no more than 1%. For example, as used herein, the expression about 100 includes 99 and 101 and all values in between (eg, 99.1, 99.2, 99.3, 99.4, etc.).

Although any methods and materials similar or equivalent to those described herein may be used in the practice or testing of this invention, preferred methods and materials are now described. All references contained herein, including patent applications and publications, are incorporated herein by reference in their entirety for all purposes.

- 13 044194

Definitions.

The expression CD3 as used herein refers to an antigen that is expressed on T cells as part of the multimolecular T cell receptor (TCR), and which consists of a homodimer or heterodimer formed from the association of two of the four chains of the receptor: CD3 epsilon, CD3 delta, CD3 zeta and CD3 gamma. Human CD3 epsilon contains the amino acid sequence shown in SEQ ID NO: 1897; Human CD3 delta contains the amino acid sequence set forth in SEQ ID NO: 1898. All references to proteins, polypeptides, and protein fragments herein are intended to refer to the human variant of the corresponding protein, polypeptide, or protein fragment unless specifically stated to refer to to a non-human species. Thus, the expression CD3 means human CD3 unless specified to be of a non-human species, such as mouse CD3, monkey CD3, etc.

As used herein, the term antibody that binds CD3 or anti-CD3 antibody includes antibodies and antigen binding fragments that specifically recognize one CD3 subunit (e.g., epsilon, delta, gamma, or zeta), as well as antibodies and antigen binding fragments thereof that specifically recognize a dimeric complex of two CD3 subunits (eg CD3 gamma/epsilon, delta/epsilon and zeta/zeta dimers). Antibodies and antigen binding fragments of this invention can bind soluble CD3 and/or cell surface expressed CD3. Soluble CD3 includes natural CD3 proteins as well as variants of recombinant CD3 proteins, such as, for example, monomeric and dimeric CD3 constructs that do not have a transmembrane domain or are otherwise not associated with the cell membrane.

As used herein, the expression CD3 expressed on the cell surface means one or more CD3 protein(s) that/are expressed on the cell surface in vitro or in vivo such that at least a portion of the CD3 protein interacts with the extracellular side of the cell. membrane and is accessible to the antigen-binding portion of the antibody. Cell surface expressed CD3 includes CD3 proteins contained within a functional T cell receptor in the cell membrane. CD3 expression expressed on the cell surface includes CD3 protein expressed as part of a homodimer or heterodimer on the cell surface (eg, CD3 gamma/epsilon, delta/epsilon and zeta/zeta dimers). Cell surface expressed CD3 also includes a CD3 chain (eg, CD3 epsilon, CD3 delta, or CD3 gamma) that is expressed alone, without other types of CD3 chain, on the cell surface. CD3 expressed on the cell surface may include or consist of CD3 protein expressed on the surface of a cell that normally expresses CD3 protein. Alternatively, cell surface expressed CD3 may include or consist of a CD3 protein expressed on the surface of a cell that does not normally express human CD3 on its surface but has been artificially engineered to express CD3 on its surface.

The expression STEAP2 as used herein refers to six-transmembrane prostate epithelial antigen 2. STEAP2 is an integral six-stranded transmembrane protein that is expressed at high levels in prostate epithelial cells and is a cell surface marker for prostate cancer, e.g. that STEAP2 is expressed in significant amounts on the prostate cell line LNCaP (Porkka, et al. Lab Invest 2002, 82:1573-1582). STEAP2 (UniProtKB/Swiss-Prot: Q8NFT2.3) is a 490 amino acid protein encoded by the STEAP2 gene located in the chromosomal region 7q21 in humans, see, for example, the amino acid sequence of human STEAP2 as indicated in SEQ ID NO: 1899.

As used herein, the term antibody that binds STEAP2 or anti-STEP2 antibody includes antibodies and antigen binding fragments that specifically recognize STEAP2.

The term antigen-binding molecule includes antibodies and antigen-binding fragments of antibodies, including, for example, bispecific antibodies.

The term antibody as used herein means any antigen-binding molecule or molecular complex containing at least one complementarity determining region (CDR) that specifically binds to or interacts with a particular antigen (eg, STEAP2 or CD3). The term antibody includes immunoglobulin molecules containing four polypeptide chains, two heavy (H) chains and two light (L) chains linked by disulfide bonds, as well as their multimers (for example, IgM). Each heavy chain contains a heavy chain variable region (abbreviated herein as HCVR or VH) and a heavy chain constant region. The heavy chain constant region contains three domains: _CH 1, _CH 2 and _CH 3. Each light chain contains a light chain variable region (abbreviated herein as LCVR or VL) and a light chain constant region. The light chain constant region contains one domain (CL1). The VH and VL regions can be further subdivided into regions of hypervariability called complementarity determining regions (CDRs), which alternate with more conserved regions called framework regions (FRs). Each V _H and V _L consists

- 14 044194 of three CDRs and four FRs, located from the N-terminus to the C-terminus in the following order: FR1, CDR1, FR2, CDR2, FR3, CDR3, FR4. In various embodiments of the invention, the FR of an anti-STEP2 antibody or an anti-CD3 antibody (or an antigen-binding portion thereof) may be identical to human germline sequences, or may be natural or artificially modified. The amino acid consensus sequence can be determined by parallel analysis of two or more CDRs.

The term antibody as used herein also includes antigen-binding fragments of full-length antibody molecules. The terms antigen binding portion of an antibody, antigen binding fragment of an antibody, and the like as used herein include any naturally occurring, enzymatically produced, synthetic or genetically modified polypeptide or glycoprotein that specifically binds an antigen to form a complex. Antigen-binding antibody fragments can be prepared, for example, from full-length antibody molecules using any suitable standard techniques, such as proteolytic digestion or recombinant genetic engineering techniques involving the manipulation and expression of DNA encoding the variable and optionally constant domains of the antibody. Such DNA is known and/or readily available, for example, from commercial sources, DNA libraries (including, for example, phage antibody libraries), or can be synthesized. The DNA may be sequenced and processed chemically or using molecular biology techniques, for example, to organize one or more variable and/or constant domains into a suitable configuration or to introduce codons, create cysteine residues, modify, add or delete amino acids, and the like.

Non-limiting examples of antigen binding fragments include: (i) Fab fragments; (ii) F(ab) 2-fragments; (iii) Fd fragments; (iv) Fv fragments; (v) single chain Fv (scFv) molecules; (vi) dAb fragments (domain antibody); and (vii) minimal recognition units consisting of amino acid residues that mimic the hypervariable region of an antibody (eg, a dedicated complementarity determining region (CDR) such as the CDR3 peptide) or a conformationally restricted peptide FR3-CDR3-FR4. Other engineered molecules, such as domain-specific antibodies, single-domain antibodies, domain-deleted antibodies, chimeric antibodies, CDR-grafted antibodies, diabodies, tri-bodies, tetrabodies, mini-bodies, nanobodies (e.g., monovalent nanobodies, divalent nanobodies, and the like), immunopharmaceuticals small modular protein (SMIP)-based agents and shark IgNAR variable domains are also encompassed by the expression antigen binding fragment as used herein.

The antigen binding fragment of an antibody typically contains at least one variable domain. The variable domain can be of any size or amino acid composition and typically comprises at least one CDR that is adjacent or in frame with one or more framework sequences. In antigen binding fragments having a VH domain associated with a VL domain, the _VH and _VL domains may be arranged relative to each other in any suitable order. For example, the variable region may be dimeric and contain dimers V _H -V _H , V _H -V _L or V _L -V _L . Alternatively, the antigen binding fragment of the antibody may comprise a monomeric V _H or V _L domain.

In some embodiments, the antigen binding fragment of an antibody may comprise at least one variable domain covalently linked to at least one constant domain. Non-limiting exemplary variable and constant domain configurations that may be found in an antigen binding fragment of an antibody of the present invention include: (i) Vh-Ch1; (ii) Vh-Ch2; (iii) Vh-Ch3; (iv) Vh-Ch1-Ch2; (v) Vh-Ch1-Ch2-Ch3; (vi) Vh-Ch2-Ch3; (vii) Vh-Cl; (viii) Vl-Ch1; (ix) Vl-Ch2; (x) Vl-Ch3; (xi) Vl-Ch1-Ch2; (xii) Vl-Ch1-Ch2-Ch3; (xiii) V _l -C _h 2-C _h 3; and (xiv) V _l -C _l . In any configuration of variable and constant domains, including any of the above exemplary configurations, the variable and constant domains may either be directly linked to each other or may be linked by a full or partial hinge or linker region. The hinge region may consist of at least 2 (eg, 5, 10, 15, 20, 40, 60 or more) amino acids that result in flexible or semi-flexible linkage between adjacent variable and/or constant domains in a single polypeptide molecule. In addition, the antigen binding fragment of an antibody of the present invention may comprise a homodimer or heterodimer (or other multimer) of any of the variable and constant domain configurations listed above in non-covalent association with each other and/or with one or more monomeric V _H or V _L - domains (eg, disulfide bond(s)).

As with full-length antibody molecules, antigen-binding fragments can be monospecific or multispecific (eg, bispecific).

A multispecific antigen binding antibody fragment will typically contain at least two different variable domains, with each variable domain being capable of specifically binding to a different antigen or to a different epitope on the same antigen. Any multispecific antibody format, including the exemplary bispecific antibody formats described herein, can be adapted for use in the context of an antigen binding fragment of an antibody of the present invention using conventional techniques available in the art.

Antibodies of the present invention may function through complement-dependent cytotoxicity (CDC) or antibody-dependent cell-mediated cytotoxicity (ADCC). Complement-dependent cytotoxicity (CDC) refers to the lysis of antigen-expressing cells by the antibody of the invention in the presence of complement.

Antibody-dependent cell-mediated cytotoxicity (ADCC) refers to a cell-mediated reaction in which nonspecific cytotoxic cells that express Fc receptors (FcR) (eg, natural killer (NK) cells, neutrophils, and macrophages) recognize bound antibodies on the target cell and thereby lead to lysis of the target cell. CDC and ADCC can be measured using assays that are well known and available in the art. (See, for example, US Patent Nos. 5500362 and 5821337, and Clynes et al. (1998) Proc. Natl. Acad. Sci. (USA) 95:652-656). The constant region of an antibody is important for the ability of the antibody to fix complement and mediate cell-dependent cytotoxicity. Thus, the antibody isotype can be selected based on whether the antibody is desired to mediate cytotoxicity.

In some embodiments of the present invention, the anti-STEP2 monospecific antibodies or anti-STEAP2/anti-CD3 bispecific antibodies of the invention are human antibodies. The term human antibody, as used herein, is intended to include antibodies having variable and constant regions derived from human germline immunoglobulin sequences. Human antibodies of the invention may include amino acid residues that are not encoded by human germline immunoglobulin sequences (eg, mutations introduced by random or site-specific mutagenesis in vitro or somatic mutation in vivo), for example, in the CDR and, in particular, CDR3. However, the term human antibody, as used herein, is not intended to include antibodies in which CDR sequences derived from the germ line of another mammalian species, such as a mouse, have been grafted into a human framework sequence.

Antibodies of the invention, in some embodiments, may be recombinant human antibodies. The term recombinant human antibody, as used herein, is intended to include all human antibodies that are produced, expressed, created or isolated by recombinant methods, such as antibodies expressed using a recombinant expression vector transfected into a host cell (described further below) , antibodies isolated from a recombinant combinatorial library of human antibodies (described further below), antibodies isolated from an animal (eg, mouse) that is transgenic for human immunoglobulin genes (see, for example, Taylor et al. (1992) Nucl. Acids Res. 20: 6287-6295), or antibodies produced, expressed, created or isolated by any other method that involves splicing human immunoglobulin gene sequences with other DNA sequences. Such recombinant human antibodies have variable and constant regions derived from human germline immunoglobulin sequences. However, in some embodiments, such recombinant human antibodies are subject to in vitro mutagenesis (or, when an animal transgenic for human Ig sequences is used, to somatic mutagenesis in vivo) and thus the amino acid sequences of the V _H and V _L regions of the recombinant antibodies are the sequences , which, although derived from and related to human germline VH and _VL sequences, would not naturally exist in the human antibody germline repertoire in vivo.

Human antibodies can exist in two forms, which are associated with hinge heterogeneity. In one form, the immunoglobulin molecule contains a stable four-chain construct of approximately 150-160 kDa in which the dimers are held together by an interchain heavy chain disulfide bond. In the second form, the dimers are not linked through interchain disulfide bonds, and a molecule of about 75-80 kDa is formed, consisting of a covalently linked light and heavy chain (half-antibody). These forms are extremely difficult to separate, even after affinity purification.

The frequency of occurrence of the second form in different isotypes of intact IgG is due to, but is not limited to, structural differences associated with the hinge region of the antibody isotype. A single amino acid substitution in the hinge region of the human IgG4 hinge can significantly reduce the occurrence of the second form (Angal et al. (1993) Molecular Immunology 30: 105) to levels typically observed using the human IgG1 hinge. The present invention covers antibodies having one or more mutations in the hinge, C _H 2 or C _H 3 region, which may be desirable, for example, in production, to improve the yield of the desired form of the antibody.

The antibodies of the invention may be isolated antibodies. An isolated antibody, as used herein, means an antibody that has been identified and separated and/or isolated from at least one component of its natural environment. For example,

- 16 044194 an antibody that has been isolated or removed from at least one component of the body, or from a tissue or cell in which the antibody naturally exists or is naturally produced, is an isolated antibody for the purposes of this invention. The isolated antibody also includes the antibody in situ in the recombinant cell. Isolated antibodies are antibodies that have been subjected to at least one purification or isolation step. In certain embodiments of the invention, the isolated antibody may be substantially free of other cellular material and/or chemicals.

The present invention also includes single-arm antibodies that bind STEAP2. As used herein, the term single-arm antibody means an antigen-binding molecule comprising one antibody heavy chain and one antibody light chain. Single-arm antibodies according to this invention may contain any of the amino acid sequences HCVR/LCVR or CDR, as indicated in table. 1.

Ahtu-STEAP2 or anti-STEP2/anti-CD3 antibodies described herein may contain one or more amino acid substitutions, insertions and/or deletions in the framework regions and/or CDR regions of the heavy and light chain variable domains compared to the corresponding germline sequences from which the antibodies were derived. Such mutations can be readily detected by comparing the amino acid sequences described herein with germline sequences available, for example, from publicly available antibody databases. This invention includes antibodies and antigen binding fragments thereof that are derived from any of the amino acid sequences described herein, wherein one or more amino acids in one or more framework regions and/or CDR regions are mutated to the corresponding germline sequence residue(s) , from which the antibody was derived, or with the corresponding residue(s) of another human germline sequence, or with a conservative amino acid substitution of the corresponding germline residue(s) (such sequence changes are referred to collectively herein as germline mutations). A person skilled in the art, starting with the heavy and light chain variable region sequences described herein, can readily produce numerous antibodies and antigen binding fragments that contain one or more individual germline mutations or combinations thereof. In some embodiments, all of the framework residues and/or CDR residues in the V _H and/or V _L domains are mutated back to the residues found in the original germline sequence from which the antibody was derived. In other embodiments, only certain residues are mutated back to the original germline sequence, for example, only the mutated residues found in the first 8 amino acids of FR1 or the last 8 amino acids of FR4, or only the mutated residues found in CDR1, CDR2, or CDR3. In other embodiments, one or more framework residue(s) and/or CDR residue(s) are mutated with the corresponding residue(s) of another germline sequence (i.e., a germline sequence that is different from the germline sequence from which the initially obtained antibody). In addition, antibodies of the present invention may contain any combination of two or more germline mutations in framework regions and/or CDR regions, for example, where certain individual residues are mutated to the corresponding residue of a certain germline sequence, while certain other residues that differ from the original germline sequence are retained or mutated with the corresponding residue of another germline sequence. Once produced, antibodies and antigen-binding fragments that contain one or more germline mutations can be readily tested for one or more desirable properties, such as improved binding specificity, improved binding affinity, improved or enhanced antagonistic or agonistic biological properties (as appropriate ), reduced immunogenicity and the like. Antibodies and antigen binding fragments obtained using this general method are covered by this invention.

The invention also includes anti-STEP2 or anti-STEP2/anti-CD3 antibodies containing variants of any of the HCVR, LCVR and/or CDR amino acid sequences described herein having one or more conservative substitutions. For example, the present invention includes anti-STEP2 or anti-STEP2/anti-CD3 antibodies having the amino acid sequences HCVR, LCVR and/or CDR with, for example, 10 or less, 8 or less, 6 or less, 4 or less, and the like conservative amino acid substitutions in relation to any of the amino acid sequences HCVR, LCVR and/or CDR presented in table. 1 in this document or as described in table. 9, 11, 12, 15 and 17 in this document.

The term epitope refers to an antigenic determinant that interacts with a specific antigen-binding site in the variable region of an antibody molecule, known as the paratope. One antigen can have more than one epitope. Thus, different antibodies may bind to different regions on the antigen and may have different biological effects. Epitopes can be conformational or linear. A conformational epitope is produced by spatially juxtaposed amino acids from different segments of a linear polypeptide chain. A linear epitope is an epitope produced by adjacent amino acid residues in a polypeptide chain. In certain circumstances, the epitope may include moieties of saccharides, phosphoryl groups, or sulfonyl groups on the antigen.

The term substantial identity or substantially identical, when referring to a nucleic acid or fragment thereof, indicates that, when optimally aligned with corresponding nucleotide insertions or deletions with another nucleic acid (or its complementary strand), there is at least an amount of nucleotide sequence identity about 95%, and more preferably at least about 96, 97, 98, or 99% of the nucleotide bases, as measured by any known sequence identity algorithm, such as FASTA, BLAST, or Gap, as discussed below. A nucleic acid molecule having substantial identity with a reference nucleic acid molecule may, in some cases, encode a polypeptide having the same or substantially similar amino acid sequence as the polypeptide encoded by the reference nucleic acid molecule.

When applied to polypeptides, the term substantially similar or substantially similar means that two peptide sequences, when optimally aligned, for example using the GAP or BESTFIT programs using standard gap-opening penalties, have at least 95% sequence identity, even more preferably the sequence at least 98% or 99% identical. Preferably, positions of residues that are not identical are distinguished by conservative amino acid substitutions. A conservative amino acid substitution is an amino acid substitution in which an amino acid residue is replaced by another amino acid residue having a side chain (R group) with similar chemical properties (eg, charge or hydrophobicity). In general, a conservative amino acid substitution will not significantly change the functional properties of the protein. In cases where two or more amino acid sequences differ from each other by conservative substitutions, the percentage sequence identity or degree of similarity can be adjusted upward to correct the conservative nature of the substitution. Means for making this adjustment are well known to those skilled in the art. See, for example, Pearson (1994) Methods Mol. Biol. 24: 307-331, incorporated herein by reference. Examples of groups of amino acids with side chains with similar chemical properties include: (1) aliphatic side chains: glycine, alanine, valine, leucine and isoleucine; (2) aliphatic hydroxyl side chains: serine and threonine; (3) amide-containing side chains: asparagine and glutamine; (4) aromatic side chains: phenylalanine, tyrosine and tryptophan; (5) main side chains: lysine, arginine and histidine; (b) acidic side chains: aspartate and glutamate and (7) sulfur-containing side chains - cysteine and methionine. Preferred groups of conservative amino acid substitutions are: valine-leucine-isoleucine, phenylalanine-tyrosine, lysine-arginine, alanine-valine, glutamate-aspartate and asparagine-glutamine. Alternatively, a conservative substitution is any change that has a positive value in the PAM250 log-likelihood matrix described in Gonnet et al. (1992) Science 256: 1443-1445, incorporated herein by reference. A moderately conservative replacement is any change that has a non-negative value in the PAM250 log-likelihood matrix.

Sequence similarity for polypeptides, also called sequence identity, is typically measured using sequence analysis software. Protein analysis software finds similar sequences using similarity measurements assigned to various substitutions, deletions and other modifications, including conserved amino acid substitutions. For example, GCG software contains programs such as Gap and Bestfit that can be used with default parameters to determine sequence homology or sequence identity between related polypeptides, such as homologous polypeptides from different species or between a wild-type protein and its mutein. See, for example, GCG version 6.1. Polypeptide sequences can also be compared using FASTA with standard or recommended parameters, program in GCG version 6.1. FASTA (eg, FASTA2 and FASTA3) provides alignments and percent sequence identity of the regions of best overlap between the query sequence and the search sequence (Pearson (2000) supra). Another preferred algorithm when comparing a sequence of the invention to a database containing a large number of sequences from different organisms is the BLAST computer program, especially BLASTP or TBLASTN, using default parameters. See, for example, Altschul et al. (1990) J. Mol. Biol. 215:403-410 and Altschul et al. (1997) Nucleic Acids Res. 25:3389-402, each of which is incorporated herein by reference.

Germline mutations.

Ahtu-CD3 antibodies described herein contain one or more amino acid substitutions, insertions and/or deletions in the framework regions and/or CDR regions of the severe variable domains.

- 18,044,194 chains compared to the corresponding germline sequences from which the antibodies were derived.

This invention also includes antibodies and antigen binding fragments thereof that are derived from any of the amino acid sequences disclosed herein, wherein one or more amino acids in one or more framework regions and/or CDR regions are mutated to the corresponding germline sequence residue(s) , from which the antibody was derived, or to the corresponding residue(s) of another human germline sequence, or to a conservative amino acid substitution of the corresponding germline residue(s) (such sequence changes are referred to collectively herein as germline mutations) and having a weak or undetectable binding to CD3 antigen. Several such exemplary antibodies that recognize CD3 are described in table. 12 and 18 of this document.

In addition, antibodies of the present invention may contain any combination of two or more germline mutations in framework regions and/or CDR regions, for example, where certain individual residues are mutated to the corresponding residue of a certain germline sequence, while certain other residues that differ from the original germline sequence are retained or mutated with the corresponding residue of another germline sequence. Once produced, antibodies and antigen binding fragments that contain one or more germline mutations can be tested for one or more desired properties, such as improved binding specificity, weak or reduced binding affinity, improved or increased pharmacokinetic properties, reduced immunogenicity, etc. Antibodies and antigen binding fragments obtained by this general method in accordance with the guidelines according to this invention are included in this invention.

The invention also includes anti-CD3 antibodies containing variants of any of the HCVR, LCVR and/or CDR amino acid sequences described herein having one or more conservative substitutions. For example, the present invention includes anti-CD3 antibodies having amino acid sequences HCVR, LCVR and/or CDRs with, for example, 10 or less, 8 or less, 6 or less, 4 or less, and the like. conservative amino acid substitutions in relation to any of the amino acid sequences HCVR, LCVR and/or CDR presented in table. 1, 9, 11, 12, 15 and 17 in this document. The antibodies and bispecific antigen binding molecules of the present invention contain one or more amino acid substitutions, insertions and/or deletions in the framework and/or CDR regions of the heavy and light chain variable domains as compared to the corresponding germline sequences from which the individual antigen binding domains were derived. maintaining or improving the desired weak or undetectable binding to the CD3 antigen. A conservative amino acid substitution is an amino acid substitution in which an amino acid residue is replaced by another amino acid residue having a side chain (R group) with similar chemical properties (eg, charge or hydrophobicity). In general, a conservative amino acid substitution will not significantly alter the functional properties of the protein, that is, the amino acid substitution maintains or improves the desired weak or undetectable binding affinity in the case of anti-CD3 binding molecules. Examples of groups of amino acids with side chains with similar chemical properties include: (1) aliphatic side chains: glycine, alanine, valine, leucine and isoleucine; (2) aliphatic hydroxyl side chains: serine and threonine; (3) amide-containing side chains: asparagine and glutamine; (4) aromatic side chains: phenylalanine, tyrosine and tryptophan; (5) main side chains: lysine, arginine and histidine; (b) acidic side chains: aspartate and glutamate and (7) sulfur-containing side chains - cysteine and methionine. Preferred groups of conservative amino acid substitutions are: valine-leucine-isoleucine, phenylalanine-tyrosine, lysine-arginine, alanine-valine, glutamate-aspartate and asparagine-glutamine. Alternatively, a conservative substitution is any change that has a positive value in the PAM250 log-likelihood matrix described in Gonnet et al. (1992) Science 256: 1443-1445. A moderately conservative replacement is any change that has a non-negative value in the PAM250 log-likelihood matrix.

The invention also includes antigen binding molecules comprising an antigen binding domain with an HCVR and/or CDR amino acid sequence that is substantially identical to any of the HCVR and/or CDR amino acid sequences described herein while maintaining or improving the desired weak affinity for the CD3 antigen. The term substantial identity or substantially identical, when referring to an amino acid sequence, means that two amino acid sequences when optimally aligned, for example by the GAP or BESTFIT programs using default gap weights, share at least 95% sequence identity. even more preferably there is at least 98 or 99% sequence identity. Preferably, positions of residues that are not identical are distinguished by conservative amino acid substitutions. In cases where two or more amino acid sequences differ from each other by conservative substitutions, the percentage sequence identity or degree of similarity can be adjusted to

- 19 044194 on the larger side to correct the conservative nature of the replacement. Means for making this adjustment are well known to those skilled in the art. See, for example, Pearson (1994)

Methods Mol. Biol. 24: 307-331.

Sequence similarity for polypeptides, also called sequence identity, is typically measured using sequence analysis software. Protein analysis software finds similar sequences using similarity measurements assigned to various substitutions, deletions and other modifications, including conserved amino acid substitutions. For example, GCG software contains programs such as Gap and Bestfit that can be used with default parameters to determine sequence homology or sequence identity between related polypeptides, such as homologous polypeptides from different species or between a wild-type protein and its mutein. See, for example, GCG version 6.1. Polypeptide sequences can also be compared using FASTA with standard or recommended parameters, program in GCG version 6.1. FASTA (eg, FASTA2 and FASTA3) provides alignments and percent sequence identity of the regions of best overlap between the query sequence and the search sequence (Pearson (2000) supra). Another preferred algorithm when comparing a sequence of the invention to a database containing a large number of sequences from different organisms is the BLAST computer program, especially BLASTP or TBLASTN, using default parameters. See, for example, Altschul et al. (1990) J. Mol. Biol. 215:403-410 and Altschul et al. (1997) Nucleic Acids Res. 25:3389-402.

Once generated, antigen binding domains that contain one or more germline mutations were tested for reduced binding affinity using one or more in vitro assays. Although antibodies that recognize a particular antigen are typically screened for their intended purpose by testing for high (ie, strong) affinity binding to the antigen, the antibodies of the present invention exhibit weak or undetectable binding. Bispecific antigen binding molecules containing one or more antigen binding domains prepared in this general manner are also included in the present invention and have been found to be advantageous as a therapy for avidity driven tumors.

Unexpected benefits, such as improved pharmacokinetic properties and low patient toxicity, may be realized from the methods described herein.

Binding properties of antibodies.

As used herein, the term binding in the context of binding of an antibody, immunoglobulin, antigen-binding fragment, or Fc-containing protein, or, for example, a predetermined antigen such as a cell surface protein or fragment thereof, generally refers to the interaction or association between at least two objects or molecular structures, such as antibody-antigen interactions.

For example, binding affinity typically corresponds to a K _D value of about 10 ^-7 M or less, for example about 10 ^-8 M or less, for example about 10 ^-9 M or less, when determined, for example, using surface plasmonic technology resonance (SPR) on a BIAcore 3000 instrument using an antigen as a ligand and an antibody, Ig, antibody-binding fragment or Fc-containing protein as an analyte (or antibody to the ligand). Cellular binding strategies such as flow cytometry (FACS) are also commonly used, and FACS data correlates well with other techniques such as competitive radioligand binding and SPR (Benedict, CA, J. Immunol Methods. 1997, 201(2):223 -31;Geuijen, CA, et al. J. Immunol Methods. 2005, 302 (1-2):68-77).

Accordingly, the antibody or antigen-binding protein of the invention binds to a predetermined antigen or cell surface molecule (receptor) having an affinity corresponding to a KD value that is at least ten times lower than its binding affinity to a non-specific antigen (for example, BSA ( bovine serum albumin), casein). In accordance with the present invention, the affinity of an antibody corresponding to a K _D value that is equal to or less than ten times less than that of a nonspecific antigen can be considered as undetectable binding, but such antibody can be paired with a second antigen binding arm to obtain a bispecific antibody according to the invention .

The term K _D (M) refers to the dissociation of the equilibrium constant of a particular antigen-antibody interaction, or the dissociation of the equilibrium constant of binding of an antibody or antibody-binding fragment to an antigen. There is an inverse relationship between KD and binding affinity, so the lower the KD value, the higher, that is, stronger, the affinity. Thus, the terms higher affinity or stronger affinity refer to a higher ability to form an interaction and therefore a lower KD value, and conversely, the terms lower affinity or weaker affinity refer to a lower ability to form an interaction and therefore a larger value K _D . In some circumstances, the higher binding affinity (or _KD ) of a particular molecule (such as an antibody) to its interacting

- 20 044194 by a partner molecule (e.g. antigen X) compared to the binding affinity of a molecule (e.g. antibody) to another interacting partner molecule (e.g. antigen Y) can be expressed as a binding ratio determined by dividing the larger value of K _D ( lower or weaker affinity) by a lower _KD value (higher or stronger affinity), for example, expressed as 5-fold or 10-fold greater binding affinity, as appropriate.

The term kd (c-1 or 1/c) refers to the dissociation rate constant of a particular antigen-antibody interaction, or the dissociation rate constant of an antibody or antibody-binding moiety. The specified value is also referred to as the k _off value.

The term k _a (M-1 x s-1 or 1/M) refers to the rate constant of association of a particular antibody-antigen interaction, or the rate constant of association of an antibody or antibody-binding moiety.

The term KA (M-1 or 1/M) refers to the equilibrium constant of association of a particular antibody-antigen interaction, or the equilibrium constant of association of an antibody or antibody-binding moiety. The associative equilibrium constant is obtained by dividing k _a by kd.

The term EC50 or _EC50 refers to the half maximum effective concentration, which includes the concentration of the antibody that induces a response halfway between the baseline level and the maximum after a specified exposure time. The EC ₅₀ is essentially the concentration of an antibody at which 50% of its maximum effect is observed. In some embodiments, the EC ₅₀ value is equal to the concentration of an antibody of the invention that produces half-maximal binding to cells expressing CD3 or tumor-associated antigen, as determined, for example, by a FACS binding assay. Thus, reduced or weaker binding is observed as the EC ₅₀ or half the maximum effective concentration increases.

In one embodiment, a decrease in binding can be defined as an increase in the EC ₅₀ concentration of an antibody that allows binding to a half-maximal number of target cells.

In another embodiment, the EC ₅₀ value represents the concentration of an antibody of the invention that causes half-maximal depletion of target cells via T cell cytotoxic activity. Thus, increased cytotoxic activity (eg, T cell-mediated tumor cell death) is observed at a reduced EC ₅₀ value or half-maximal effective concentration value.

Bispecific antigen-binding molecules.

Antibodies according to this invention can be monospecific, bispecific or multispecific. Multispecific antibodies may be specific for different epitopes of a single target polypeptide or may contain antigen binding domains specific for more than one target polypeptide. See, for example, Tutt et al., 1991, J. Immunol. 147:60-69; Kufer et al., 2004, Trends Biotechnol. 22:238-244. Monospecific anti-STEP2 antibodies or bispecific anti-STEP2/anti-CD3 antibodies of the present invention may be linked to or co-expressed with another functional molecule, such as another peptide or protein. For example, an antibody or fragment thereof may be operably linked (e.g., by chemical linkage, genetic fusion, non-covalent association, or otherwise) to one or more other molecular entities, such as another antibody or antibody fragment, to produce a bispecific or multispecific antibody with a second or additional binding specificity.

The use of the expression anti-CD3 antibody or anti-STEP2 antibody herein is intended to include both monospecific anti-CD3 or anti-STEP2 antibodies, as well as bispecific antibodies containing a CD3 binding arm and a STEP2 binding arm. Thus, the present invention includes bispecific antibodies in which one immunoglobulin arm binds human CD3 and the other immunoglobulin arm is specific for human STEAP2. The CD3 binding arm may contain any of the amino acid sequences HCVR/LCVR or CDR, as indicated in table. 1, 9, 11, 12, 15 and 17 in this document.

In some embodiments, the CD3 binding arm binds to human CD3 and induces activation of human T cells. In some embodiments, the CD3 binding arm weakly binds to human CD3 and induces activation of human T cells. In other embodiments, the CD3-binding arm weakly binds to human CD3 and induces the killing of tumor-associated antigen-expressing cells in the context of a bispecific or multispecific antibody. In other embodiments, the CD3 binding arm binds or weakly binds to human and cynomolgus CD3, but the binding interaction is not detected by in vitro assays known in the art. The STEP2 binding arm may contain any of the amino acid sequences HCVR/LCVR or CDR, as indicated in table. 1 in this document.

- 21 044194

In some exemplary embodiments, the invention includes bispecific antigen binding molecules that specifically bind CD3 and STEAP2.

Such molecules may be referred to herein as, for example, anti-CD3/anti-STEP2, anti-CD3xSTEAP2 or CD3xSTEAP2 bispecific molecules or other similar terminology (eg, anti-STEP2/anti-CD3).

The term STEAP2 as used herein refers to the human STEAP2 protein unless specified to be from a non-human species (eg, mouse STEAP2, monkey STEAP2, etc.). The human STEAP2 protein has the amino acid sequence shown in SEQ ID NO: 1899.

The above-mentioned bispecific antigen binding molecules that specifically bind CD3 and STEAP2 may comprise an anti-CD3 antigen binding molecule that binds CD3 with weak binding affinity, such as exhibiting a K _D of greater than about 40 nM, as measured by an in vitro binding affinity assay. In some cases, the CD3 binding arm binds CD3 to a _KD or EC ₅₀ greater than about 100 nM, greater than about 200 nM, greater than about 300 nM, greater than about 400 nM, greater than about 500 nM, or greater than about 1 µM (eg, as measured in surface plasmon resonance analysis). In some cases, the first antigen binding domain specifically binds CD3 (eg, human CD3 and/or cynomolgus CD3 with weak or undetectable affinity).

As used herein, the expression antigen binding molecule means a protein, polypeptide, or molecular complex containing or consisting of at least one complementarity determining region (CDR), which alone or in combination with one or more additional CDRs and/or framework regions ( FR) specifically binds to a specific antigen. In certain embodiments, the antigen binding molecule is an antibody or antibody fragment, as those terms are defined elsewhere herein.

As used herein, the expression bispecific antigen binding molecule means a protein, polypeptide, or molecular complex comprising at least a first antigen binding domain and a second antigen binding domain. Each antigen binding domain in a bispecific antigen binding molecule contains at least one CDR that, alone or in combination with one or more additional CDRs and/or FRs, specifically binds to a particular antigen. In the context of the present invention, the first antigen binding domain specifically binds a first antigen (eg, CD3), and the second antigen binding domain specifically binds a second different antigen (eg, STEAP2).

In some illustrative embodiments of the present invention, the bispecific antigen binding molecule is a bispecific antibody. Each antigen binding domain of a bispecific antibody contains a heavy chain variable domain (HCVR) and a light chain variable domain (LCVR). In the context of a bispecific antigen binding molecule comprising a first and a second antigen binding domain (eg, a bispecific antibody), the CDRs of the first antigen binding domain may be denoted by the prefix A1, and the CDRs of the second antigen binding domain may be denoted by the prefix A2. Thus, the CDRs of the first antigen binding domain may be referred to herein as A1-HCDR1, A1-HCDR2 and A1-HCDR3; and the second antigen binding domain CDRs may be referred to herein as A2-HCDR1, A2-HCDR2, and A2-HCDR3.

The first antigen binding domain and the second antigen binding domain can be directly or indirectly linked to each other to form a bispecific antigen binding molecule according to the present invention. Alternatively, the first antigen binding domain and the second antigen binding domain may each be associated with a separate multimerization domain. The association of one multimerizing domain with another multimerizing domain facilitates the association between the two antigen binding domains, thereby forming a bispecific antigen binding molecule. As used herein, the term multimerization domain is any macromolecule, protein, polypeptide, peptide, or amino acid that has the ability to bind to a second multimerization domain of the same or similar structure or structure. For example, the multimerizing domain may be a polypeptide containing an immunoglobulin C _H 3 domain. A non-limiting example of a multimerizing component is the Fc portion of an immunoglobulin (containing the CH2-CH3 domain), for example, the IgG Fc domain selected from the IgG1, IgG2, IgG3 and IgG4 isotypes, as well as any allotype within each isotype group.

The bispecific antigen binding molecules of the present invention typically contain two multimerizing domains, for example two Fc domains, each of which is individually part of a distinct antibody heavy chain. The first and second multimerization domains may be of the same IgG isotype, such as, for example, IgG1/IgG1, IgG2/IgG2, IgG4/IgG4. Alternatively, the first and second multimerizing domains may be of different IgG isotypes, such as, for example, IgG1/IgG2, IgG1/IgG4, IgG2/IgG4, etc.

In some embodiments, the multimerization domain is an Fc fragment

- 22 044194 or an amino acid sequence containing from 1 to about 200 amino acids in length, containing at least one cysteine residue. In other embodiments, the multimerization domain is a cysteine residue or a short cysteine-containing peptide. Other multimerizing domains include peptides or polypeptides containing or consisting of a leucine zipper, a helix-loop motif, or a supercoil motif.

Any bispecific antibody format or technology can be used to create a bispecific antigen binding molecule according to this invention. For example, an antibody or fragment thereof having a first antigen-binding specificity may be operably linked (e.g., by chemical binding, genetic fusion, non-covalent association, or otherwise) to one or more other molecular entities, such as another antibody or antibody fragment having a second antigen binding specificity to obtain a bispecific antigen binding molecule. Specific exemplary bispecific formats that may be used in the context of this invention include, but are not limited to, scFv or diabody based bispecific formats, IgG-scFv hybrids, double variable domain (DVD)-Ig, quadroma, peak-to-valley , common light chain (e.g., conventional peak-to-trough light chain and the like), CrossMab, CrossFab, SEED body (surrogate antibody, antibody containing a chain exchange engineered domain), leucine zipper, Duobody, IgG1 /IgG2, dual action Fab (DAF)-IgG and Mab ² (see, for example, Klein et al. 2012, mAb 4:6, 1-11, and references cited therein, for an overview of the above formats).

In the context of the bispecific antigen binding molecules of the present invention, multimerizing domains, eg, Fc domains, may include one or more amino acid changes (eg, insertions, deletions or substitutions) compared to the wild type, naturally occurring variant of the Fc domain. For example, the invention includes bispecific antigen binding molecules containing one or more modifications in the Fc domain, which results in the modified Fc domain having a modified binding interaction (eg, increased or decreased) between the Fc and FcRn. In one embodiment, the bispecific antigen binding molecule contains a modification at the C _H 2 or C _H 3 region, which modification increases the affinity of the Fc domain for FcRn in an acidic environment (for example, in the endosome, where the pH ranges from about 5.5 to about 6. 0). Non-limiting examples of such Fc modifications include, for example, modification at position 250 (eg, E or Q); 250 and 428 (for example, L or F); 252 (for example, L/Y/F/W or T), 254 (for example, S or T) and 256 (for example, S/R/Q/E/D or T); or modification at position 428 and/or 433 (eg, L/R/S/P/Q or K) and/or 434 (eg, H/F or Y); or modification at position 250 and/or 428; or a modification at position 307 or 308 (eg, 308F, V308F) and 434. In one embodiment of the invention, the modification comprises modification 428L (eg, M428L) and 434S (eg, N434S); modification 428L, 259I (for example, V259I) and 308F (for example, V308F); modification 433K (for example, Н433К) and 434 (for example, 434Y); modification 252, 254 and 256 (for example, 252Y, 254T and 256e); modification 250Q and 428L (for example, T250Q and M428L); and modification 307 and/or 308 (for example, 308F or 308P).

This invention also includes bispecific antigen-binding molecules comprising a first _CH 3 domain and a second _CH 3 domain of Ig3, wherein the first and second _CH 3 Ig domains differ from each other by at least one amino acid, and wherein the difference is at least one amino acid decreases the binding of the bispecific antibody to protein A compared to a bispecific antibody in which there is no amino acid difference. In one embodiment, the first CH3 Ig domain binds Protein A, and the second CH3 Ig domain contains a mutation that reduces or abolishes binding to Protein A, similar to the modification H95R (as per IMGT exon numbering, H435R as per EU numbering). The second _CH 3 may additionally contain the modification Y96F (according to IMGT; Y436F according to EU). See, for example, US Pat. No. 8,586,713. Other modifications that may be found in the second _CH 3 include: D16E, L18M, N44S, K52N, V57M and V82I (according to IMGT; D356E, L358M, N384S, K392N, V397M and V422I according to EU) in case of IgG1 antibodies; N44S, K52N and V82I (according to IMGT; N384S, K392N and V422I according to EU) in the case of IgG2 antibodies; and Q15R, N44S, K52N, V57M, R69K, E79Q and V82I (according to IMGT; Q355R, N384S, K392N, V397M, R409K, E419Q and V422I according to EU) for IgG4 antibodies.

In some embodiments, the Fc domain may be a chimeric, combining Fc sequences derived from more than one immunoglobulin isotype. For example, a chimeric Fc domain may comprise part or all of a _CH 2 sequence derived from the C _H 2 region of human IgG1, human IgG2, or human IgG4, and part or all of a _CH 3 sequence derived from human IgG1, human IgG2, or human IgG4 . The chimeric Fc domain may also contain a chimeric hinge region. For example, a chimeric hinge may comprise an upper hinge sequence derived from a human IgG1, human IgG2, or human IgG4 hinge region in combination with a lower hinge sequence derived from a human IgG1, human IgG2, or human IgG4 hinge region. A specific example of a chimeric Fc domain that can be included in any of the antigen-binding

- 23 044194 molecules set forth herein contain from N- to C-terminus: [IgG4 _CH 1] - [upper hinge IgG4] [lower hinge IgG2] - [CH2 IgG4] - [CH3 IgG4]. Another example of a chimeric Fc domain that can be included in any of the antigen-binding molecules set forth herein contains from N- to C-terminus: [IgG1 C _H 1] - [IgG1 upper hinge] - [IgG2 lower hinge] - [CH2 IgG4] [CHZ IgG1]. These and other examples of chimeric Fc domains that can be included in any of the antigen binding molecules of this invention are described in US Publication No. 2014/0243504, published August 28, 2014, which is incorporated herein in its entirety. Chimeric Fc domains sharing these common structural arrangements and variants thereof may have altered binding to the Fc receptor, which in turn affects Fc effector function.

In some embodiments, the invention provides an antibody heavy chain wherein the heavy chain constant region (CH) contains at least 95%, at least 96%, at least 97%, at least 98% amino acid sequence, at least 99% identical to any of SEQ ID NO: 1911, SEQ ID NO: 1912, SEQ ID NO: 1913, SEQ ID NO: 1914, SEQ ID NO: 1915, SEQ ID NO: 1916, SEQ ID NO: 1917 , SEQ ID NO: 1918, SEQ ID NO: 1919, or SEQ ID NO: 1920. In some embodiments, the heavy chain constant region (CH) comprises an amino acid sequence selected from the group consisting of SEQ ID NO: 1911, SEQ ID NO: 1912, SEQ ID NO: 1913, SEQ ID NO: 1914, SEQ ID NO: 1915, SEQ ID NO: 1916, SEQ ID NO: 1917, SEQ ID NO: 1918, SEQ ID NO: 1919 and SEQ ID NO: 1920.

In other embodiments, the invention provides an antibody heavy chain wherein the Fc domain contains an amino acid sequence that is at least 95%, at least 96%, at least 97%, at least 98%, at least 99% identical any of SEQ ID NO: 1921, SEQ ID NO: 1922, SEQ ID NO: 1923 SEQ ID NO: 1924 SEQ ID NO: 1925, SEQ ID NO: 1926, SEQ ID NO: 1927, SEQ ID NO: 1928, SEQ ID NO: 1929 or SEQ ID NO: 1930. In some embodiments, the Fc domain comprises an amino acid sequence selected from the group consisting of SEQ ID NO: 1921, SEQ ID NO: 1922, SEQ ID NO: 1923 SEQ ID NO: 1924 SEQ ID NO: 1925, SEQ ID NO: 1926, SEQ ID NO: 1927, SEQ ID NO: 1928, SEQ ID NO: 1929 and SEQ ID NO: 1930.

Sequence options.

The antibodies and bispecific antigen binding molecules of the present invention may contain one or more amino acid substitutions, insertions and/or deletions in the framework regions and/or CDR regions of the heavy and light chain variable domains as compared to the corresponding germline sequences from which the individual antigen binding molecules were derived. domains. Such mutations can be readily detected by comparing the amino acid sequences described herein with germline sequences available, for example, from publicly available antibody databases. The antigen binding molecules of the present invention may comprise antigen binding domains that are derived from any of the exemplary amino acid sequences described herein, wherein one or more amino acids in one or more framework regions and/or CDR regions are mutated to the corresponding residue(s) germline sequence from which the antibody was derived, or with the corresponding residue(s) of another human germline sequence, or with a conservative amino acid substitution of the corresponding germline residue(s) (such sequence changes are referred to collectively herein as germline mutations). A person skilled in the art, starting with the heavy and light chain variable region sequences described herein, can readily produce numerous antibodies and antigen binding fragments that contain one or more individual germline mutations or combinations thereof. In some embodiments, all of the framework residues and/or CDR residues in the VH and/or _VL domains are mutated back to the residues found in the original germline sequence from which the antigen binding domain was originally derived. In other embodiments, only certain residues are mutated back to the original germline sequence, for example, only the mutated residues found in the first 8 amino acids of FR1 or the last 8 amino acids of FR4, or only the mutated residues found in CDR1, CDR2, or CDR3. In other embodiments, one or more framework residue(s) and/or CDR residue(s) are mutated with the corresponding residue(s) of another germline sequence (i.e., a germline sequence that is different from the germline sequence from which the originally obtained antigen-binding domain). In addition, antigen binding domains may contain any combination of two or more germline mutations in framework regions and/or CDR regions, for example, where certain individual residues are mutated to the corresponding residue of a certain germline sequence, while certain other residues that differ from the original germline sequence are conserved or mutate with the corresponding residue of another germline sequence. Once prepared, antigen binding domains that contain one or more germline mutations can be readily tested for one or more desired properties, such as improved binding specificity, increased binding affinity, improved or enhanced antagonistic or agonistic biological properties (depending on depending on the circumstances), reduced immunogenicity, etc. Bispecific antigen binding molecules containing one or more antigen binding domains obtained in such a general manner are covered by this invention.

The invention also includes antigen binding molecules, wherein one or both antigen binding domains comprise variants of any of the HCVR, LCVR and/or CDR amino acid sequences described herein having one or more conservative substitutions. For example, the present invention includes antigen binding molecules comprising an antigen binding domain having the amino acid sequences HCVR, LCVR and/or CDR with, for example, 10 or less, 8 or less, 6 or less, 4 or less, and the like of conservative amino acid substitutions relative to any of the HCVR, LCVR and/or CDR amino acid sequences described herein. A conservative amino acid substitution is an amino acid substitution in which an amino acid residue is replaced by another amino acid residue having a side chain (R group) with similar chemical properties (eg, charge or hydrophobicity). In general, a conservative amino acid substitution will not significantly change the functional properties of the protein. Examples of groups of amino acids with side chains with similar chemical properties include: (1) aliphatic side chains: glycine, alanine, valine, leucine and isoleucine; (2) aliphatic-hydroxyl side chains: serine and threonine; (3) amide-containing side chains: asparagine and glutamine; (4) aromatic side chains: phenylalanine, tyrosine and tryptophan; (5) main side chains: lysine, arginine and histidine; (b) acidic side chains: aspartate and glutamate and (7) sulfur-containing side chains - cysteine and methionine. Preferred groups of conservative amino acid substitutions are: valine-leucine-isoleucine, phenylalanine-tyrosine, lysine-arginine, alanine-valine, glutamate-aspartate and asparagine-glutamine. Alternatively, a conservative substitution is any change that has a positive value in the PAM250 log-likelihood matrix described in Gonnet et al. (1992) Science 256: 144 3-1445, incorporated herein by reference. A moderately conservative replacement is any change that has a non-negative value in the PAM250 log-likelihood matrix.

The invention also includes antigen binding molecules comprising an antigen binding domain with an HCVR, LCVR, and/or CDR amino acid sequence that is substantially identical to any of the HCVR, LCVR, and/or CDR amino acid sequences described herein. The term substantial identity or substantially identical, when referring to an amino acid sequence, means that two amino acid sequences when optimally aligned, for example by the GAP or BESTFIT programs using default gap weights, share at least 95% sequence identity. even more preferably at least 98% or 99% sequence identity. Preferably, positions of residues that are not identical are distinguished by conservative amino acid substitutions. In cases where two or more amino acid sequences differ from each other by conservative substitutions, the percentage sequence identity or degree of similarity can be adjusted upward to correct the conservative nature of the substitution. Means for making this adjustment are well known to those skilled in the art. See, for example, Pearson (1994) Methods Mol. Biol. 24: 307-331, incorporated herein by reference.

Sequence similarity for polypeptides, also called sequence identity, is typically measured using sequence analysis software. Protein analysis software finds similar sequences using similarity measurements assigned to various substitutions, deletions and other modifications, including conserved amino acid substitutions. For example, GCG software contains programs such as Gap and Bestfit that can be used with default parameters to determine sequence homology or sequence identity between related polypeptides, such as homologous polypeptides from different species or between a wild-type protein and its mutein. See, for example, GCG version 6.1. Polypeptide sequences can also be compared using FASTA with standard or recommended parameters, program in GCG version 6.1. FASTA (eg, FASTA2 and FASTA3) provides alignments and percent sequence identity of the regions of best overlap between the query sequence and the search sequence (Pearson (2000) supra). Another preferred algorithm when comparing a sequence of the invention to a database containing a large number of sequences from different organisms is the BLAST computer program, especially BLASTP or TBLASTN, using default parameters. See, for example, Altschul et al. (1990) J. Mol. Biol. 215:403-410 and Altschul et al. (1997) Nucleic Acids Res. 25:3389-402, each of which is incorporated herein by reference.

pH-dependent binding.

This invention includes anti-STEP2 antibodies and anti-CD3/anti-STEAP2 bispecific

- 25 044194 antigen-binding molecules, with pH-dependent binding characteristics. For example, the aHTu-STEP2 antibody of the present invention may exhibit reduced binding to STEAP2 at acidic pH compared to neutral pH. Alternatively, the anti-STEP2 antibodies of the invention may exhibit enhanced binding to STEAP2 at acidic pH compared to neutral pH. The expression acidic pH includes pH values less than about 6.2, such as about 6.0, 5.95, 5.9, 5.85, 5.8, 5.75, 5.7, 5.65, 5. 6, 5.55, 5.5, 5.45, 5.4, 5.35, 5.3, 5.25, 5.2, 5.15, 5.1, 5.05, 5.0 or less. As used herein, the expression neutral pH means a pH of about 7.0 to about 7.4. The expression neutral pH includes pH values around 7.0, 7.05, 7.1, 7.15, 7.2, 7.25, 7.3, 7.35 and 7.4.

In some cases, reduced binding...at an acidic pH, compared to a neutral pH, is expressed in terms of the ratio of the _KD value of an antibody binding to its antigen at an acidic pH to the _KD value of an antibody binding to its antigen at a neutral pH (or vice versa ). For example, an antibody or antigen binding fragment thereof may be considered to exhibit reduced binding to STEAP2 at acidic pH compared to neutral pH for purposes of the present invention if the antibody or antigen binding fragment thereof has a K _D at acidic to a K _D at neutral pH of about 3.0 or more. In some illustrative embodiments, the acidic K _D to neutral pH K _D ratio for an antibody or antigen binding fragment of the present invention may be about 3.0, 3.5, 4.0, 4.5, 5.0, 5.5 , 6.0, 6.5, 7.0, 7.5, 8.0, 8.5 9.0, 9.5, 10.0, 10.5, 11.0, 11.5, 12, 0, 12.5, 13.0, 13.5, 14.0, 14.5, 15.0, 20.0, 25.0, 30.0, 40.0, 50.0, 60.0, 70.0, 100.0 or higher.

Antibodies with pH-dependent binding characteristics can be obtained, for example, by screening a population of antibodies for decreased (or increased) binding to a particular antigen at an acidic pH compared to a neutral pH. In addition, modifications of the antigen-binding domain at the amino acid level can produce antibodies with pH-dependent characteristics. For example, by replacing one or more amino acids of the antigen binding domain (eg, in the CDR) with a histidine residue, an antibody with reduced antigen binding at an acidic pH relative to a neutral pH can be obtained.

Antibodies containing Fc variants.

In accordance with some embodiments of the present invention, anti-STEP2 antibodies and anti-CD3/anti-STEAP2 bispecific antigen-binding molecules are provided comprising an Fc domain containing one or more mutations that increase or decrease the binding of the antibody to the FcRn receptor, e.g. pH compared to neutral pH. For example, the present invention includes antibodies containing a mutation in the C _H 2 or C _H 3 region of the Fc domain, wherein the mutation(s) increases the affinity of the Fc domain for FcRn in an acidic environment (for example, in the endosome, where the pH ranges from about 5. 5 to about 6.0). Such mutations may result in an increase in the serum half-life of the antibody when administered to an animal. Non-limiting examples of such Fc modifications include, for example, modification at position 250 (eg, E or Q); 250 and 428 (for example, L or F); 252 (for example, L/Y/F/W or T), 254 (for example, S or T) and 256 (for example, S/R/Q/E/D or T); or modification at position 428 and/or 433 (for example, H/L/R/S/P/Q or K) and/or 434 (for example, H/F or Y); or modification at position 250 and/or 428; or modification at positions 307 or 308 (eg, 308F, V308F) and 434. In one embodiment of the invention, the modification includes modification 428L (eg, M428L) and 434S (eg, N434S); modification 428L, 259I (for example, V259I) and 308F (for example, V308F); modification 4 33K (for example, Н433К) and 434 (for example, 434Y); modification 252, 254 and 256 (for example, 252Y, 254T and 256E); modification 250Q and 428L (for example, T250Q and M428L); and modification 307 and/or 308 (for example, 308F or 308P).

For example, the present invention includes anti-STEP2 antibodies and anti-CD3/anti-STEP2 bispecific antigen-binding molecules containing an Fc domain containing one or more pairs or groups of mutations selected from the group consisting of: 250Q and 248L (for example, T250Q and M248L); 252Y, 254T and 256E (for example, M252Y, S254T and T256E); 428L and 434S (for example, M428L and N434S); and 433K and 434F (for example, Н433К and N434F). All possible combinations of the above Fc domain mutations and other mutations in the variable domains of the antibodies described herein are contemplated within the scope of this invention.

Biological characteristics of antibodies and bispecific antigen-binding molecules.

The present invention includes antibodies and antigen-binding fragments thereof that bind human STEAP2 with high affinity (eg, subnanomolar K _D values).

The present invention also includes anti-CD3/anti-STEAP2 bispecific antigen binding molecules that inhibit tumor growth in immunocompromised mice bearing human prostate cancer xenografts. (see, for example, Example 5).

This invention includes antibodies and antigen-binding fragments thereof that bind human CD3 with high affinity. The invention also includes antibodies and antigen-binding fragments thereof that bind human CD3 with moderate or low affinity, depending on the therapeutic context and the specific target properties that are desired. For example, in the context of a bispecific antigen-binding molecule, in which one arm binds CD3 and

- 26 044194 the other arm binds the target antigen (eg, STEAP2), it may be desirable that the target antigen binding arm binds the target antigen with high affinity, while the antiCD3 arm binds CD3 with only moderate or low affinity. In this way, preferential targeting of the antigen-binding molecule to cells expressing the target antigen can be achieved while avoiding general/off-target CD3 binding and associated undesirable effects.

The present invention includes bispecific antigen binding molecules (eg, bispecific antibodies) that are capable of simultaneously binding to human CD3 and human STEAP2. The binding arm that interacts with cells that express CD3 may have weak or undetectable binding as measured in a suitable in vitro binding assay. The extent to which a bispecific antigen binding molecule binds cells that express CD3 and/or STEAP2 can be assessed using flow cytometry (FACS).

The invention also includes antibodies, antigen binding fragments, and bispecific antibodies thereof that bind to STEP2-expressing cells and cell lines (eg, CA-2 cells) with an _EC50 value between about 1 nM and 50 nM, as determined using a binding assay FACS as described in Example 2, or using a substantially similar analysis. In some embodiments, antibodies, antigen binding fragments, and bispecific antibodies thereof that bind to STEAP2-expressing cells and cell lines (e.g., CA-2 cells) have an EC ₅₀ value of about 50 nM, about 40 nM, about 30 nM, about 20 nM, less than about 15 nM, about 10 nM, about 5 nM, about 4 nM, about 3 nM, or about 2 nM, about 1 nM, as determined using a FACS binding assay as described in Example 2, or using essentially a similar analysis.

The present invention includes antibodies, antigen binding fragments, and bispecific antibodies thereof, which bind human CD3 with weak (ie low) or even undetectable affinity. In accordance with some embodiments, the invention includes antibodies and antigen-binding antibody fragments that bind human CD3 (eg, at 37°C) with a KD of less than about 11 nM, as measured by surface plasmon resonance. In some embodiments, antibodies or antigen binding fragments of the present invention bind CD3 with a KD greater than about 15 nM, greater than about 20 nM, greater than about 25 nM, greater than about 30 nM, greater than about 35 nM, greater than about 40 nM , more than about 45 nM, more than about 50 nM, more than about 55 nM, more than about 60 nM, more than about 65 nM, more than about 70 nM, more than about 75 nM, at least 80 nM, more more than about 90 nM, more than about 100 nM, more than about 110 nM, at least 120 nM, more than about 130 nM, more than about 140 nM, more than about 150 nM, at least 160 nM, more than about 170 nm, more than about 180 nM, more than about 190 nM, more than about 200 nM, more than about 250 nM, more than about 300 nM, more than about 400 nM, more than about 500 nM, or more than about 1 μM or with undetectable affinity, as measured by surface plasmon resonance (eg, mAb capture or antigen capture format), or by a substantially similar assay.

The present invention includes antibodies, antigen binding fragments, and bispecific antibodies thereof that bind monkey (ie, cynomolgus) CD3 with weak (ie, low) or even undetectable affinity.

The present invention includes anti-CD3/anti-STEP2 bispecific antigen binding molecules that bind to and are internalized by human STEAP2 expressing cells (e.g., CA-2 cells) as measured by the assay format as defined in Example 3 herein. or by a substantially similar analysis. The present invention includes anti-CD3/anti-STEP2 bispecific antigen binding molecules that are specific for binding to human STEAP2. In some embodiments, the anti-CD3/anti-STEAP2 bispecific antigen binding molecules of the present invention bind human STEAP-2 transiently expressed in HEK293 cells, as measured by the assay format defined in Example 3 herein, or by a substantially similar analysis. In some embodiments, the anti-CD3/anti-STEP2 bispecific antigen binding molecules of the present invention do not bind human STEAP1, human STEAP2, or human STEAP4 transiently expressed in HEK293 cells, as measured using the assay format defined in Example 3 herein. or by a substantially similar analysis.

The present invention includes anti-CD3/anti-STEAP2 bispecific antigen binding molecules that are capable of inhibiting C4-2 tumor growth (see, for example, Example 5). For example, in some embodiments, anti-CD3/anti-STEP2 bispecific antigen-binding molecules are provided wherein a single administration, e.g., at a dose of about 0.1 mg/kg or about 0.01 mg/kg), causes a decrease in tumor size compared to animals treated with a control isotype bispecific antibody, when measured 46 days after tumor implantation, as detected in the subject using standard caliper measurement methods, for example, as described in Example 5 herein.

- 27 044194

The invention also includes aHTu-STEAP2 antibody-drug conjugates that inhibit tumor growth in in vivo STEAP2-positive prostate cancer xenograft models (see, for example, Example 7, or in a substantially similar assay). In some embodiments, anti-STEP2 antibody-drug conjugates with Compound 7 are provided, wherein a single dose of 10, 20, or 40 mg/kg administered on day 13 after tumor implantation inhibits C4-2 tumor growth in in vivo STEP2-positive xenograft models of prostate cancer. In some embodiments, anti-STEP2 antibody-drug conjugates with Compound 7 are provided, wherein a single dose of 5 or 20 mg/kg administered on day 14 postimplantation inhibits C4-2 tumor growth in in vivo STEP2-positive xenograft models of prostate cancer . In some embodiments, anti-STEP2 antibody-drug conjugates with compound 7 are provided, wherein a single dose of 150 mg/kg administered on day 17 postimplantation inhibits C4-2 tumor growth in in vivo STEP2-positive xenograft models of prostate cancer . In other embodiments, anti-STEP2 antibody-drug conjugates with compound 60 are provided, wherein a single dose of at least 2.5 mg/kg, administered on day 29 postimplantation, inhibits C4-2 tumor growth in in vivo STEP2-positive xenograft models of prostate cancer.

Epitope mapping and related technologies.

The epitope on CD3 and/or STEAP2 to which the antigen-binding molecules of the present invention bind may consist of a single contiguous sequence of 3 or more (e.g., 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 2 0 or more) amino acids of CD3 or STEAP2 protein. Alternatively, the epitope may consist of multiple non-contiguous amino acids (or amino acid sequences) of CD3 or STEAP2. Antibodies of the invention may react with amino acids contained on one CD3 chain (eg, CD3 epsilon, CD3 delta, or CD3 gamma) or may react with amino acids on two or more different CD3 chains. As used herein, the term epitope refers to an antigenic determinant that interacts with a specific antigen-binding site in the variable region of an antibody molecule known as the paratope. One antigen can have more than one epitope. Thus, different antibodies may bind to different regions on the antigen and may have different biological effects. Epitopes can be conformational or linear. A conformational epitope is produced by spatially juxtaposed amino acids from different segments of a linear polypeptide chain. A linear epitope is an epitope produced by adjacent amino acid residues in a polypeptide chain. In certain circumstances, the epitope may include moieties of saccharides, phosphoryl groups, or sulfonyl groups on the antigen.

Various methods known to those skilled in the art can be used to determine whether the antigen binding domain of an antibody interacts with one or more amino acids in a polypeptide or protein. Exemplary methods include, for example, standard epitope cross-competition assays such as those described in Antibodies, Harlow and Lane (Cold Spring Harbor Press, Cold Spring Harb., New York), alanine scanning mutational analysis, peptide blot analysis (Reineke, 2004 , Methods Mol Biol 248: 443–463) and peptide cleavage assay. In addition, methods such as epitope excision, epitope isolation, and chemical modification of antigens can be used (Tomer, 2000, Protein Science 9:487-496). Another method that can be used to identify amino acids within a polypeptide with which the antigen binding domain of an antibody interacts is hydrogen/deuterium exchange detected by mass spectrometry. In general, the hydrogen/deuterium exchange method involves labeling a protein of interest with deuterium, followed by binding of an antibody to the deuterium-labeled protein. The protein/antibody complex is then transferred to water to allow hydrogen-deuterium exchange across all residues except the antibody-protected residues (which remain deuterium-labeled). Once the antibody is dissociated, the target protein is subjected to protease digestion and analysis by mass spectrometry, thereby identifying deuterium-labeled residues that correspond to the specific amino acids with which the antibody interacts. See, for example, Ehring (1999) Analytical Biochemistry 267 (2):252-259; Engen and Smith (2001) Anal. Chem. 73:256A-265A. X-ray crystallography of the antigen/antibody complex can also be used for epitope mapping.

The invention further includes anti-STEP2 antibodies that bind to the same epitope as any of the specific exemplary antibodies described herein (eg, antibodies containing any of the amino acid sequences set forth in Table 1 herein). Likewise, the present invention also includes anti-STEAP2 antibodies that compete for binding to STEAP2 with any of the specific exemplary antibodies described herein (e.g., antibodies containing any of the amino acid sequences as set forth in Table 1 herein) .

The invention also includes bispecific antigen binding molecules comprising a first antigen binding domain that specifically binds human CD3 and/or

- 28 044194

Cynomolgus CD3 with low or detectable binding affinity, and a second antigen binding domain that specifically binds human STEAP2, wherein the first antigen binding domain binds to the same epitope on CD3 as any of the specific exemplary CD3-specific antigen binding domains described herein, and/or wherein the second antigen binding domain binds to the same epitope on STEAP2 as any of the specific exemplary STEP2-specific antigen binding domains described herein.

In addition, the present invention also includes bispecific antigen binding molecules comprising a first antigen binding domain that specifically binds human CD3 and a second antigen binding domain that specifically binds human STEAP2, wherein the first antigen binding domain competes for binding to CD3 with any of the specific exemplary CD3- specific antigen binding domains described herein, and/or in which the second antigen binding domain competes for binding to STEAP2 with any of the specific exemplary STEP2-specific antigen binding domains described herein.

Whether a particular antigen-binding molecule (eg, an antibody) or its antigen-binding domain binds to the same epitope or competes for binding with a reference antigen-binding molecule of the present invention can be readily determined using standard methods known in the art. For example, to determine whether a test antibody binds to the same epitope on STEAP2 (or CD3) as a reference bispecific antigen binding molecule of the present invention, the reference bispecific molecule is first allowed to bind to the STEAP2 protein (or CD3 protein). The ability of the test antibody to bind to the STEAP2 (or CD3) molecule is then assessed. If the test antibody is able to bind to STEAP2 (or CD3) after saturation of binding to the reference bispecific antigen binding molecule, it can be concluded that the test antibody binds to a different epitope of STEAP2 (or CD3) than the reference bispecific antigen binding molecule. On the other hand, if the test antibody is unable to bind to a STEAP2 (or CD3) molecule after saturation of binding to the reference bispecific antigen binding molecule, then the test antibody may bind to the same STEAP2 (or CD3) epitope as the epitope bound by the reference bispecific antigen binding molecule according to invention. Additional standard testing (e.g., peptide mutation and binding assays) can then be performed to confirm whether the observed lack of binding of the test antibody is actually due to binding to the same epitope as the reference bispecific antigen binding molecule or the cause lack of observed binding is steric blocking (or other phenomenon). Experiments of this type can be performed using ELISA, RIA (radioimmunoassay), Biacore, flow cytometry, or any other quantitative or qualitative antibody binding assay available in the art. In accordance with some embodiments of the present invention, two antigen binding proteins bind to the same (or overlapping) epitope if, for example, a 1-, 5-, 10-, 20-, or 100-fold excess of one antigen binding protein inhibits the binding of the other by at least 50%, but preferably 75%, 90% or even 99%, as measured in a competitive binding assay (see, for example, Junghans et al., Cancer Res. 1990: 50: 1495-1502). Alternatively, two antigen binding proteins are considered to bind to the same epitope if substantially all amino acid mutations in the antigen that reduce or eliminate the binding of one antigen binding protein reduce or eliminate the binding of the other. Two antigen binding proteins are considered to have overlapping epitopes if only a subset of amino acid mutations that reduce or eliminate the binding of one antigen binding protein reduce or eliminate the binding of the other.

To determine whether an antibody or its antigen-binding domain competes for binding to a reference antigen-binding molecule, the binding technique described above is performed in two orientations: in the first orientation, the reference antigen-binding molecule is allowed to bind the STEAP2 protein (or CD3 protein) under saturation conditions, followed by an assessment of the binding of the test antibody with the STEAP2 (or CD3) molecule. In the second orientation, the test antibody is allowed to bind to the STEAP2 (or CD3) molecule under saturated conditions, followed by assessment of the binding of the reference antigen binding molecule to the STEAP2 (or CD3) molecule. If in both orientations only the first (saturating) antigen binding molecule is able to bind to the STEAP2 (or CD3) molecule, then it is concluded that the test antibody and the reference antigen binding molecule are competing for binding to STEAP2 (or CD3). As will be appreciated by one skilled in the art, an antibody that competes for binding to a reference antigen binding molecule may not necessarily bind to the same epitope as the reference antibody, but may sterically block binding of the reference antibody by binding to an overlapping or adjacent epitope.

- 29 044194

Preparation of antigen-binding domains and construction of bispecific molecules.

Antigen-binding domains specific for specific antigens can be obtained using any antibody technology known in the art. Once two different antigen binding domains specific for two different antigens (eg, CD3 and STEAP2) are obtained, they can be suitably positioned relative to each other to produce a bispecific antigen binding molecule of the present invention using standard methods. (A discussion of exemplary bispecific antibody formats that can be used to construct bispecific antigen binding molecules according to this invention is provided elsewhere herein.) In certain embodiments, one or more individual components (eg, heavy and light chains) of the multispecific antigen binding molecules of the invention are derived from chimeric, humanized, or fully human antibodies. Methods for producing such antibodies are well known in the art. For example, one or more heavy and/or light chains of the bispecific antigen binding molecules of the present invention can be produced using VELOCIMMUNE™ technology. Using VELOCIMMUNE™ technology (or any other technology that generates human antibodies), chimeric antibodies with high affinity for a specific antigen (eg, CD3 or STEAP2), having a human variable region and a mouse constant region, are initially isolated. Antibodies are characterized and selected for desired characteristics including affinity, selectivity, epitope, etc. The mouse constant regions are replaced with the desired human constant region to generate fully human heavy and/or light chains that can be incorporated into the bispecific antigen binding molecules of the present invention.

Genetically modified animals can be used to produce human bispecific antigen-binding molecules. For example, a genetically modified mouse may be used that is unable to rearrange and express an endogenous mouse immunoglobulin light chain variable sequence, wherein the mouse expresses only one or two human light chain variable domains encoded by human immunoglobulin sequences operably linked to a mouse immunoglobulin constant region gene in endogenous mouse kappa locus. Such genetically modified mice can be used to produce fully human bispecific antigen binding molecules containing two different heavy chains that are linked to an identical light chain that contains a variable domain derived from one of two different human light chain variable region gene segments. (See for example US 2011/0195454). The term fully human refers to an antibody or an antigen binding fragment or immunoglobulin domain thereof comprising an amino acid sequence encoded by DNA derived from a human sequence throughout the entire length of each antibody polypeptide or antigen binding fragment or immunoglobulin domain thereof. In some cases, completely human sequences are obtained from a protein endogenous to humans. In other cases, the entirely human protein or protein sequence contains a chimeric sequence, each component sequence being derived from a human sequence. Without being bound by any particular theory, chimeric proteins or chimeric sequences are generally designed to minimize the formation of immunogenic epitopes at component sequence junctions, for example, compared to any regions or domains of wild-type human immunoglobulin.

Bioequivalents.

The present invention covers antigen binding molecules having amino acid sequences that differ from those of the exemplary molecules described herein, but which retain the ability to bind CD3 and/or STEAP2. Such variant molecules may contain one or more insertions, deletions or substitutions of amino acids relative to the original sequence, but exhibit biological activity that is substantially equivalent to that of the described bispecific antigen binding molecules.

The present invention includes antigen binding molecules that are bioequivalent to any of the exemplary antigen binding molecules described herein. Two antigen-binding proteins or antibodies are considered bioequivalent if, for example, they are pharmaceutical equivalents or pharmaceutical alternatives whose rate and extent of absorption do not show a significant difference when administered at the same molar dose under similar experimental conditions, either in a single dose or in multiple doses. dose. Some antigen binding proteins will be considered equivalents or pharmaceutical alternatives if they are equivalent in their extent of absorption, but not in their rate of absorption, and may still be considered bioequivalent because such differences in rate of absorption are intentional and reflected in the labeling are not essential to achieving effective concentrations of the drug, for example, during long-term use, and are considered medically insignificant for the particular drug product being studied.

- 30 044194

In one embodiment, two antigen binding proteins are bioequivalent if there are no clinically significant differences in their safety, purity, and potency.

In one embodiment, two antigen binding proteins are bioequivalent if the patient can switch one or more times between the reference drug and the biologic without an expected increase in the risk of adverse effects, including a clinically significant change in immunogenicity or reduced efficacy compared to ongoing therapy without such switching.

In one embodiment, two antigen binding proteins are bioequivalent if they both act by a common mechanism or mechanisms of action for the condition or conditions of use, to the extent such mechanisms are known.

Bioequivalence can be demonstrated by in vivo and in vitro methods. Bioequivalence measurements include, for example, (a) an in vivo test in humans or other mammals in which the concentration of an antibody or its metabolites is measured in blood, plasma, serum or other biological fluid as a function of time; (b) an in vitro test that correlates with and reliably predicts in vivo human bioavailability data; (c) an in vivo test in humans or other mammals in which the relevant short-term pharmacological effect of the antibody (or its target) is measured as a function of time; and (d) in a well-controlled clinical study that establishes the safety, efficacy or bioavailability or bioequivalence of the antigen binding protein.

Bioequivalent versions of the illustrative bispecific antigen binding molecules set forth herein can be constructed, for example, by making various residue or sequence substitutions or removing terminal or internal residues or sequences that are not essential for biological activity. For example, cysteine residues that are not essential for biological activity can be removed or replaced with other amino acids to prevent the formation of unnecessary or incorrect intramolecular disulfide bridges upon renaturation. In other contexts, bioequivalent antigen binding proteins may include variants of the illustrative bispecific antigen binding molecules set forth herein containing amino acid substitutions that modify the glycosylation characteristics of the molecules, for example, mutations that eliminate or stop glycosylation.

Species selectivity and species cross-reactivity.

In accordance with some embodiments of the invention, antigen binding molecules are provided that bind to human CD3, but not to CD3 from other species. Also provided are antigen binding molecules that bind to human STEAP2 but not to STEAP2 from other species. The invention also includes antigen-binding molecules that bind to human CD3 and to CD3 of one or more non-human species; and/or antigen-binding molecules that bind to human STEAP2 and STEAP2 of one or more non-human species.

In accordance with some illustrative embodiments of the invention, antigen binding molecules are provided that bind to human CD3 and/or human STEAP2 and may or may not bind, as appropriate, to one or more of mouse, rat, marine CD3 and/or STEAP2. pig, hamster, gerbil, pig, cat, dog, rabbit, goat, sheep, cow, horse, camel, cynomolgus, monkey, rhesus or chimpanzee. For example, a specific exemplary embodiment of the present invention provides bispecific antigen binding molecules comprising a first antigen binding domain that binds human CD3 and cynomolgus CD3, and a second antigen binding domain that specifically binds human STEAP2.

Antibody-drug conjugates (ADCs).

This invention provides antibody-drug conjugates (ADCs) containing an anti-STEP2 antibody or an antigen-binding fragment thereof conjugated to a therapeutic molecule, such as a cytotoxic agent, a chemotherapeutic drug, an immunosuppressant, or a radioisotope. In general terms, an ADC contains: A - [L - P] _y , where A is an antigen binding molecule, for example, an anti-STEP2 antibody or fragment thereof (for example, a fragment containing at least HCDR3 selected from any of the amino acid sequences of HCDR3 listed in Table 1), L is a linker, P is a drug payload or moiety (eg, a cytotoxic agent), and y is an integer from 1 to 30. In various embodiments, the ADC comprises an anti-STEAP2 antibody or an antigen-binding fragment thereof that contains CDRs HCVR and LCVR having the amino acid sequences of SEQ ID NO (eg, SEQ ID NO: 2 and 10) listed in table. 1, or specific HCVR/LCVR pairs (eg, SEQ ID NO: 2/10). In some cases, the anti-STEP2 antibody or fragment thereof contains a CDR with the amino acid sequences SEQ ID NO (for example, SEQ ID NO: 4-6-8-12-14-16) shown in table. 1. In some cases, the anti-STEAP2 antibody or fragment thereof contains HCVR and LCVR,

- 31 044194 having the amino acid sequences of SEQ ID NO (for example, SEQ ID NO: 2 and 10) presented in table. 1, or pairs of specific amino acid sequences (eg, SEQ ID NO: 2/10).

Cytotoxic agents include any agent that is harmful to the growth, viability or reproduction of cells. The antigen-binding molecules or antibodies of the invention deliver these cytotoxic agents, referred to herein as payloads, to target cells. Examples of suitable cytotoxic agents and chemotherapeutic agents for the formation of ADCs are known in the art.

Examples of suitable cytotoxic agents and chemotherapeutic agents that may be conjugated to anti-STEP2 antibodies in accordance with this aspect of the invention include, for example, 1-(2-chloroethyl)-1,2 dimethanesulfonylhydrazide, 1,8-dihydroxy-bicyclo[ 7.3.1]trideca-4, 9-dien-2,6-dien-13-one, 1-dehydrotestosterone, 5-fluorouracil, 6-mercaptopurine, 6-thioguanine, 9-amine-camptothecin, actinomycin D, amanitins, aminopterin , anguidine, anthracycline, anthramycin (AMC), auristatins (monomethyl auristatin E or monomethyl auristatin F), bleomycin, bisulfan, butyric acid, calicheamicins, camptothecin, carminomycin, carmustine, quemadotins, cisplatin, colchicine, combretastatins, cyclophosphamide, cytarabine, cytochalasin B , dactinomycin, daunorubicin, decarbazine, diacetoxypentyldoxorubicin, dibromomannitol, dihydroxyanthracenedione, disorazoles, dolastatin, doxorubicin, duocarmycin, echinomycins, eleutherobines, emetine, epothilones, esperamycin, estramustines, ethidium bromide, etoposide, fluorouracils, g eldanamycins, gramicidin D, glucocorticoids, irinotecans, leptomycins , leurosines, lidocaine, lomustine (CCNU), maytansinoids, mechlorethamine, melphalan, mercatopurines, methopterins, methotrexate, mithramycin, mitomycin, mitoxantrone, N8-acetyl spermidine, podophyllotoxins, procaine, propranolol, pteridines, pyromycin, rhizoxins, streptozotocin , thalisomycins, taxol , tenoposide, tetracaine, thioepa, chlorambucil, tomaimycins, topotecans, tubulisin, vinblastine, vincristine, vindesine, vinorelbine, and derivatives of any of the above.

In some embodiments, the cytotoxic agent that is conjugated to the anti-STEP2 antibody is an auristatin, such as monomethyl auristatin E (MMAE) or monomethyl auristatin F (MMAF), a tubulisin such as TUB-OH or TUB-OMOM, a derivative tomaimycin, a dolastatin derivative, or a maytansinoid such as DM1 or DM4. In some embodiments, the cytotoxic agent is a maytansinoid having the structure of formula (I), including stereoisomers of compounds of formula (I):

o CH ₃ (Formula I) where A represents arylene or heteroarylene.

In some embodiments, A is a divalent benzene, pyridine, naphthalene, or quinolone radical, which is optionally substituted.

In some embodiments, A is arylene.

In some embodiments, A is:

where R ¹ independently represents in each case alkyl, alkenyl, alkynyl, aryl, alkaryl, o

-|-or ^a -^-so ₂ r ^a -I-^r ^a aralkyl, halogen, heteroaryl, heterocycloalkyl, hydroxyl, cyano, nitro, '' or azido where R ^a is alkyl or heteroalkyl;

n is an integer from 0 to 4;

m is an integer from 0 to 3;

p is an integer from 0 to 6; and q is an integer from 0 to 5.

In some embodiments, the compound of formula I is selected from the group consisting of:

- 32 044194

- 33 044194

- 34 044194

In one embodiment, a compound of formula (I) is:

In some embodiments, the maytansinoid of formula (I) is conjugated to an anti-STEP2 antibody or antigen binding fragment thereof via a linker as shown in formula (IA) below:

wherein A is arylene or heteroarylene, as discussed above in connection with formula (I);

L represents a linker;

BA is an anti-STEP2 antibody or an antigen-binding fragment thereof; and k is an integer from 1 to 30.

In various embodiments, L is:

A

4-SP-AA ¹ -AA ² -^where SP is a spacer;

* represents one or more links to an anti-STEP2 antibody or fragment thereof;

AA ¹ is an amino acid; And

AA ² is an amino acid.

In some embodiments, AA ¹ -AA ² is: valine-citrulline, citrulline valine, lysine-phenylalanine, phenylalanine-lysine, valine-asparagine, asparagine-valine, threonine-asparagine, asparagine-threonine, serine-asparagine, asparagine-serine, phenylalanine-asparagine, asparagine-phenylalanine, leucine-asparagine, asparagine-leucine, isoleucine-asparagine, asparagine-isoleucine, glycine-asparagine, asparagine-glycine, glutamic acid-asparagine, asparagine-glutamic acid, citrulline-asparagine, asparagine-citrulline, alanine-asparagine or asparagine-alanine.

In some embodiments, the SP is:

- 35 044194

where ⁵ represents a linkage to an anti-STEP2 antibody or fragment thereof; and b is an integer from 2 to 8.

In other embodiments, L is:

where * represents a link to the aHTU-STEP2 antibody or fragment thereof; and b is an integer from 2 to 8.

In one embodiment, a compound of formula (IA), including a linker that binds to an anti-STEP2 antibody or antigen binding fragment thereof, is:

where represents a link to an anti-STEP2 antibody or fragment thereof. In some cases, this fragment is called compound 10.

In one embodiment, a compound of formula (IA), including a linker that binds to an anti-STEP2 antibody or antigen binding fragment thereof, is:

where * represents a linkage to an anti-STEP2 antibody or fragment thereof. In some cases this fragment is called compound 60.

In some embodiments, the cytotoxic agent is a maytansinoid having the structure of formula (II), including stereoisomers of compounds of formula (II):

basic ₃ ° CH ₃ (Formula II) where A _3a represents an amino acid, a peptide having 2-20 amino acids, alkyl, alkynyl, alkenyl, cycloalkyl, aryl, heteroaryl, heterocyclyl, -CR5R6-, -O-, -C(= О)-, -О-С(=О)-, -С(=О)-О-,

- 36 044194

-OC(=O)-O-, -C(=O)-(CH _x )p _G , -C(=O)-O-(CHx)p1-, -(CH _x )p _G C(=O )-, -(CH _x )р1-С(=О)-О-, -(О-(СН2)р2-)рз-,

-((CH2)р2-О-)рз—, -C(=S)-, -C(=S)-S-, -C(=S)-NH-, -S-C(=S)-, - S-C(=S)-S-, -S-, -SO-, -SO2-, -NR4-, -N(R4)C(=O)-N(R8)-, -N(R)-C( =O)O-, -N(R4)-C(=O)-, -C(=O)-N(R)-, -C(=O)-N(R4)-C(=O)- or -O-C(=O)-NR4-, wherein alkyl, alkynyl, alkenyl, cycloalkyl, aryl, heteroaryl and heterocyclyl are optionally substituted; and p1, p2 and p3 are each independently 0 or an integer from 1 to 100;

x represents 0, 1 or 2;

R4, R5, R6 and _R8 are each independently H or substituted or unsubstituted: alkyl, alkenyl, alkynyl, aryl, heteroaryl or heterocyclyl; And

R4 _a is substituted or unsubstituted: alkyl, alkenyl, alkynyl, aryl, heteroaryl or heterocyclyl.

In some embodiments, the compound of formula (II) is selected from the group consisting of:

O CH ₃ O CH ₃

About sn ₃

In one embodiment, the compound of formula (II) is:

o sn ₃

In some embodiments, the maytansinoid of formula (II) is conjugated to an anti-STEP2 antibody or antigen binding fragment thereof via a linker as shown in formula (IIA) below:

(Formula HA) where BA is an anti-STEP2 antibody or an antigen-binding fragment thereof;

a is an integer from 1 to 30;

Z ₂ has the following structural formula: -Z _2A -Z _2B -Z _2C -Z _2D; where each Z _2A , Z _2B , Z _2C and Z _2D are independently absent, amino acid, peptide having 2-20 amino acids, alkyl, alkynyl, alkenyl, cycloalkyl, aryl, heteroaryl, heterocyclyl, -CR5R6-, -O-, - С(=О)-, -О-С(=О)-, -С(=О)-О-,

- 37 044194

-О-С(=О)-О-, -C(=O)-(CHx)pi, -С(=О)-О-(СНх)р1, -(СНх)р1-С(=О)- , -(CH _x )p ₁ -C(=O)-O-, -(O-(CH ₂ )p ₂ -)pz-,

-((CH2)р2-О-)рз-, -C(=S)-, -C(=S)-S-, -C(=S)-NH-, -S-C(=S)-, - S-C(=S)-S-, -S-, -SO-, -SO2-, -NR-, -N(R4) -c (=O) -N(R8) -, -N(R)-C (=O) 0-, -N(R4)-C(=O)-, -C(=O)-N(R)-, -€(=O)-N(R)-C(=O) -, -O-C (=O)-N (R4) , -OC(=S)-N(R4)-, -C(=S)-N(R4)-, -N=C=S, -N= C=O,

A is a natural or unnatural amino acid or peptide containing 2-20 amino acids;

W is -O-, -S-, -CR5R6-, or -NR4-;

X represents aryl, heteroaryl, cycloalkyl or heterocyclyl, where aryl, heteroaryl, cycloalkyl and heterocyclyl are optionally substituted;

where A ₁ , A ₃ and R ₁ are each independently an amino acid, a peptide having 2-20 amino acids, alkyl, alkynyl, alkenyl, cycloalkyl, aryl, heteroaryl, heterocyclyl, -CR5R6-, -O-, -C(= О)-, -О-С(=О)-, -С(=О)-О-, -О-С(=О)-О-, -С(=О)-(СНх)р1-, - С(=О)-О-(СНх)р1-, -(СНх)р1-С(=О)-, -(СНх)р1-С(=О)-О-, -(О-(СН2)р2 -)р3-, -((СН2)р2-О-)р3-, -C(=S)-, -C(=S)-S-, -SC(=S)-, -C(=S) -NH-, -SC(=S)-S-, -S-, -SO-, -SO2-, -NR4-, -N(R4)-C(=O)-N(R8)-, -N (R4)-C(=O)O-, -N(R4)-C(=O)-, -C(=O)-N(R4)-, -C(=O)-N(R4)- C(=O)-, or -OC(=O)-NR4-, where alkyl, alkynyl, alkenyl, cycloalkyl, aryl, heteroaryl and heterocyclyl are optionally substituted;

R ₁₇ is selected from the group consisting of O, S, NR ₁₈ and CR5R6;

R ₁₈ is selected from the group consisting of H, alkyl, alkynyl, alkenyl, cycloalkyl, aryl, heteroryl, heterocyclyl and acyl, where alkyl, alkynyl, alkenyl, cycloalkyl, aryl, heteroaryl, heterocyclyl and acyl are optionally substituted;

R ₄ , R 5 , R 6 and R ₈ are each independently H or substituted or unsubstituted: alkyl, alkenyl, alkynyl, aryl, heteroaryl or heterocyclyl;

R _4a is substituted or unsubstituted: alkyl, alkenyl, alkynyl, aryl, heteroaryl or heterocyclyl.

p ₁ , p ₂ and p ₃ are each independently 0 or an integer from 1 to 100; and x represents 0, 1 or 2.

In some embodiments of formula (IIA), A is a peptide selected from the group consisting of valine-citrulline, citrulline-valine, lysine-phenylalanine, phenylalanine-lysine, valine-asparagine, asparagine-valine, threonine-asparagine, asparagine- threonine, serine-asparagine, asparagine-serine, phenylalanine-asparagine, asparagine-phenylalanine, leucine-asparagine, asparagine-leucine, isoleucine-asparagine, asparagine-isoleucine, glycine-asparagine, asparagine-glycine, glutamic acid-asparagine, asparagine-glutamic acid, citrulline-asparagine, asparagine-citrulline, alanine-asparagine and asparagine-alanine.

In one embodiment, the compound of formula (IIA) that binds to an anti-STEP2 antibody or antigen binding fragment thereof is:

where “represents a linkage to an anti-STEP2 antibody or fragment thereof. In some cases, this fragment is called Compound 7.

In some embodiments, the cytotoxic agent that is conjugated to the anti-STEP2 antibody or fragment thereof is a pure or substantially pure diastereomer of DM1:

and y is an integer between 1 and 0.

In another embodiment, the ADC contains the structure A - [L - P] _y , in which A is an anti-STEAP2 antibody or antigen binding fragment thereof, and [L - P] is:

- 38 044194

a mixture thereof, and where y is an integer from 1 to 30, and represents a linkage to an anti-STEP2 antibody or fragment thereof.

Other maytansinoid derivatives are described in WO 2014/145090, WO2016/160615, WO 2015/031396, each of which is incorporated herein by reference in its entirety.

In some embodiments, the cytotoxic agent that is conjugated to the anti-STEP2 antibody or fragment thereof is MMAEM or MMAF.

Other cytotoxic agents known in the art are within the scope of this invention, including, for example, protein toxins such as ricin, C. difficile toxin, Pseudomonas exotoxin, diphtheria toxin, botulinum toxin, briodine, saporin, lactox toxins (t ie phytolaccatoxin and phytolacigenin) and others, such as those described in Sapra et al., Pharmacol. & Therapeutics, 2013, 138:452-469.

Cytotoxic agents (payloads) can be linked to an anti-STEP2 antigen-binding molecule or antibody of the invention through a chemical linker that covalently binds the payload compound to the protein molecule (ie, antibody). Exemplary embodiments of specific linkers are discussed above. More generally, and as used herein, the term linker refers to any divalent group or moiety that connects, combines, or binds a binding agent (eg, an antibody or antigen-binding fragment thereof) to a specified payload compound herein. In general, suitable binding agent linkers for the antibody conjugates described herein are those that are stable enough to utilize the half-life of the antibody in the bloodstream, while still being capable of releasing its payload following antigen-mediated internalization of the conjugate. Linkers can be cleavable or non-cleavable. Cleavable linkers are linkers that are cleaved through intracellular metabolism after internalization, such as cleavage by hydrolysis, reduction, or enzymatic reaction. Non-cleavable linkers are linkers that release the attached payload through lysosomal degradation of the antibody after internalization. Suitable linkers include, but are not limited to, acid-labile linkers, hydrolysis-labile linkers, enzymatically cleavable linkers, reducing labile linkers, self-cleavable linkers, and non-cleavable linkers. Suitable linkers also include, but are not limited to, those that are or contain peptides, glucuronides, succinimide thioesters, polyethylene glycol (PEG) moieties, hydrazones, malcaproyl moieties, dipeptide moieties, valine-citrulline moieties, and para-aminobenzyl (PAB) moieties. In some cases, the linker is capable of binding to an antibody or antigen-binding fragment through a lysine residue or a cysteine residue (eg, through cleavage of a disulfide group of the antibody or fragment or through a cysteine residue incorporated into the antibody or fragment). In some cases, the linker is capable of binding to the antibody or fragment through a glutamine residue, including those obtained by transglutaminase-mediated conjugation.

Exemplary linkers that may be used in the context of this invention include linkers that contain or consist of, for example, MC (6-maleimidocaproyl), MCC (maleimidomethyl cyclohexane-1-carboxylate), MP (maleimidopropanoyl), val-cit (valine -citrulline), val-ala (valine-alanine), ala-phe (alanine-phenylalanine), phe-lys (phenylalanine-lysine), dipeptide site in protease-cleavable linker, PAB (p-aminobenzyloxycarbonyl), SPP (N-succinimidyl 4-(2

- 39 044194 pyridylthio)pentanoate), SMCC (N-succinimidyl 4-(N-maleimidomethyl)cyclohexane-1-carboxylate), SIAB (N-succinimidyl (4-iodo-acetyl)aminobenzoate) and their variants and combinations. Additional examples of linkers that can be used in the context of this invention are described, for example, in US patent No. 7754681 and in Ducry, Bioconjugate Chem., 2010, 21: 5-13, and in the references therein, the contents of which are incorporated herein in their entirety by reference in their entirety. In some cases, the linker is or contains a self-cleaving spacer, such as those discussed in Jin, et al., Bioorganic & Medicinal Chemistry, 2012, 20:3465-3469, and Wu, et al., Bioorganic & Medicinal Chemistry, 2016, 24: 2697-2706.

Payloads can be linked to an anti-STEP2 antibody or antigen-binding fragment thereof by attachment at specific amino acids within the antibody or antigen-binding molecule. Exemplary amino acid attachments that may be used in the context of this aspect of the invention include, for example, lysine (see, for example, US Pat. No. 5,208,020; US 2010/0129314; Hollander et al., Bioconjugate Chem., 2008, 19:358-361 ; WO 2005/089808; US patent No. 5714586; US 2013/010154 6; and US 2012/0585592), cysteine (see, for example, US 2007/0258987; WO 2013/055993; WO 2013/055990; WO 2013 /053873 ; WO 2013/053872; WO 2011/130598; US 2013/0101546; and US patent No. 7750116), selenocysteine (see, for example, WO 2008/122039; and Hofer et al., Proc. Natl. Acad. Sci. , USA, 2008, 105:12451-12456), formyl glycine (see, for example, Carrico et al., Nat. Chem. Biol., 2007, 3:321-322; Agarwal et al., Proc. Natl. Acad. Sci., USA, 2013, 110:46-51, and Rabuka et al., Nat. Protocols, 2012, 10:1052-1067), unnatural amino acids (see, for example, WO 2013/068874 and WO 2012/166559 ) and acidic amino acids (see, for example, WO 2012/05982). Linkers can also be conjugated to antigen binding protein by attachment to carbohydrates (see, for example, US 2008/0305497 and Ryan et al., Food & Agriculture Immunol., 2001, 13: 127-130) and disulfide linkers (see, for example, , WO 2013/085925, WO 2010/010324, WO 2011/018611 and Shaunak et al., Nat. Chem. Biol., 2006, 2: 312313).

The drug-to-antibody ratio (DAR) is the average number of drugs conjugated to an antibody or antigen-binding moiety and has an important influence on the efficacy, activity, and pharmacokinetics of an ADC. In various embodiments, the DAR is 1, 2, 3, 4, 5, 6, 7, or 8 drug molecules per antibody. In some embodiments, the DAR is between 1 and 4. In certain embodiments, the DAR is between 2 and 4. In some embodiments, the DAR is between 2 and 3. In some embodiments, the DAR is between 3 and 4. In some embodiments, the DAR is between 1 up to 10, 1 to 20, or 1 to 30 (i.e., 1 to 30 drug molecules per antibody or antigen-binding fragment).

Therapeutic composition and administration.

The present invention provides pharmaceutical compositions containing the antigen binding molecules of the present invention. The pharmaceutical compositions of the invention are formulated with suitable carriers, excipients and other agents that provide improved transport, delivery, tolerability and the like. Many suitable formulations can be found in the formulary known to all pharmaceutical chemists: Remington's Pharmaceutical Sciences, Mack Publishing Company, Easton, PA. These dosage forms include, for example, powders, pastes, ointments, jellies, waxes, oils, lipids, lipid (cationic or anionic) containing vesicles (such as LIPOFECTIN™ Life Technologies, Carlsbad, California), DNA conjugates, anhydrous absorption pastes, emulsions such as oil in water or water in oil, emulsions based on carbowax (polyethylene glycols with different molecular weights), semi-solid gels and semi-solid mixtures containing carbowax. See also Powell et al. Compendium of excipients for parenteral formulations PDA (1998) J Pharm Sci Technol 52:238–311.

The dose of antigen-binding molecule administered to a patient may vary depending on the age and size of the patient, the target disease, pathological conditions, route of administration, and the like. The preferred dose is usually based on body weight or body surface area. When the bispecific antigen binding molecule of the present invention is used therapeutically in an adult patient, it may be appropriate to administer the bispecific antigen binding molecule of the present invention intravenously, typically in a single dose of about 0.01 to about 20 mg/kg body weight, more preferably from about 0.02 to about 7, about 0.03 to about 5, or about 0.05 to about 3 mg/kg body weight. Depending on the severity of the pathological condition, the frequency and duration of treatment can be adjusted. Effective doses and schedules of administration of a bispecific antigen-binding molecule can be determined experimentally; for example, the patient's progress can be monitored through periodic assessment, and the dose adjusted accordingly. In addition, cross-species dose adjustment can be accomplished using methods well known in the art (eg, Mordenti et al., 1991, Pharmaceut. Res. 8:1351).

Various delivery systems are known and can be used for the administration of pharmaceutical

- 40 044194 compositions according to the invention, for example, encapsulation in liposomes, microparticles, microcapsules, recombinant cells capable of expressing mutant viruses, receptor-mediated endocytosis (see, for example, Wu et al., 1987, J. Biol. Chem. 262:4429- 4432). Routes of administration include, but are not limited to, intradermal, intramuscular, intraperitoneal, intravenous, subcutaneous, intranasal, epidural and oral routes of administration. The composition can be administered by any suitable route, for example, by infusion or bolus injection, absorption through epithelial or mucocutaneous membranes (for example, oral mucosa, rectal and intestinal mucosa, etc.) and can be used in conjunction with other biologically active agents. Administration may be systemic or local.

The pharmaceutical composition of this invention can be delivered subcutaneously or intravenously using a standard needle and syringe. Moreover, with respect to subcutaneous delivery, a delivery pen can be easily used to deliver the pharmaceutical composition of the present invention. This syringe pen can be reusable or disposable. A reusable pen syringe uses a replaceable cartridge that contains a pharmaceutical composition. Once all of the pharmaceutical composition within the cartridge has been administered and the cartridge becomes empty, the empty cartridge can be easily removed and replaced with a new cartridge that contains the pharmaceutical composition. The delivery pen can then be reused. The disposable syringe pen does not have a replaceable cartridge. Rather, the disposable delivery pen is filled with a pharmaceutical composition held in a reservoir within the device. Once the reservoir is empty of pharmaceutical composition, the entire device is discarded.

Numerous reusable pens and syringe tubes are useful for subcutaneous delivery of the pharmaceutical composition of this invention. Examples include, but are not limited to, AUTOPEN™ (Owen Mumford, Inc., Woodstock, UK), DISETRONIC™ pen (Disetronic Medical Systems, Bergdorf, Switzerland), HUMALOG MIX 75/25™ pen, HUMALOG™ pen , HUMALIN 70/30™ pen (Eli Lilly and Co., Indianapolis, IN), NOVOPEN™ I, II and III (Novo Nordisk, Copenhagen, Denmark), NOVOPEN JUNIOR™ (Novo Nordisk, Copenhagen, Denmark), syringe -BD™ pen (Becton Dickinson, Franklin Lake, NJ), OPTIPEN™, OPTIPEN PRO™, OPTIPEN STARLET™, and OPTICLIK™ (sanofi-aventis, Frankfurt, Germany), to name a few. Examples of disposable delivery pens that are used for subcutaneous delivery of the pharmaceutical composition of this invention include, but are not limited to, the SOLOSTAR™ pen (sanofi-aventis), FLEXPEN™ (Novo Nordisk) and KWIKPEN™ (Eli Lilly) , SURECLICK™ auto-injector (Amgen, Thousand Oaks, CA), PENLET™ (Haselmeier, Stuttgart, Germany), EPIPEN (Dey, L.P.), and HUMIRATM pen (Abbott Labs, Abbott Park, IL), to name a few them.

In certain situations, the pharmaceutical composition may be provided in a controlled release system. In one embodiment, a pump may be used (see Langer, supra; Sefton, 1987, CRC Crit. Ref. Biomed. Eng. 14:201). Alternatively, polymeric materials may be used; see, Medical Applications of Controlled Release, Langer and Wise (eds.), 1974, CRC Pres., Boca Raton, FL. In yet another embodiment, the controlled release system may be located close to the target of the composition, thus requiring only a fraction of the systemic dose (see, for example, Goodson, 1984, in Medical Applications of Controlled Release, supra, vol. 2, pp. 115-138 ). Other controlled release systems are discussed in Langer, 1990, Science 249: 1527-1533.

Injectable forms may include dosage forms for intravenous, subcutaneous, intradermal and intramuscular injections, drip infusions and the like. These injectable forms can be prepared by generally known methods. For example, injectable forms can be prepared, for example, by dissolving, suspending or emulsifying the antibody or salt thereof described above in a sterile aqueous or oily medium commonly used for injection. As the aqueous injection medium, there are, for example, saline solution, isotonic solution containing glucose and other auxiliaries and the like, which can be used in combination with a corresponding solubilizing agent such as alcohol (for example, ethanol), polyhydric alcohol (for example , propylene glycol, polyethylene glycol), nonionic surfactant [for example, polysorbate 80, HCO-50 (polyoxyethylene (50 mol) adduct of hydrogenated castor oil] and the like. The oil medium used is, for example, sesame oil, soybean oil, etc. the like, which can be used in combination with a solubilizing agent such as benzyl benzoate, benzyl alcohol, etc. The injection thus obtained is preferably filled into a suitable ampoule.

Advantageously, the pharmaceutical compositions for oral or parenteral use described above are prepared in a single dose dosage form adjusted to contain the dose of the active ingredients. Such unit dose dosage forms include, for example, tablets, pills, capsules, injections (ampoules), suppositories and the like. Quantity

- 41 044194 of the above-mentioned antibody contained is usually from 5 to 500 mg per dosage form in a single dose; especially in injection form, it is preferable that the above antibody be contained in an amount of from about 5 to about 100 mg and from about 10 to about 250 mg for other dosage forms.

Therapeutic applications of antigen-binding molecules.

The present invention includes methods comprising administering to a subject in need thereof a therapeutic composition comprising an anti-STEP2 antibody or an antigen binding fragment thereof or a bispecific antigen binding molecule that specifically binds CD3 and STEAP2. The therapeutic composition may contain any of the antibodies or bispecific antigen binding molecules as described herein and a pharmaceutically acceptable carrier or diluent. As used herein, the expression subject in need thereof means a human or non-human animal that exhibits one or more symptoms or signs of cancer (for example, a subject expressing a tumor or suffering from any of the cancers mentioned herein below) , or who would otherwise benefit from inhibition or reduction of STEAP2 activity or depletion of STEAP2+ cells (eg, prostate cancer cells).

The antibodies and bispecific antigen binding molecules of the invention (and therapeutic compositions containing the same) are useful in particular for the treatment of any disease or disorder in which stimulation, activation and/or targeting of the immune response would be beneficial. In particular, anti-STEP2 antibodies or anti-CD3/anti-STEP2 bispecific antigen binding molecules of this invention can be used for the treatment, prevention and/or amelioration of any disease or disorder associated with or mediated by the expression or activity of STEAP2, or the proliferation of STEAP2+ cells. The mechanism of action by which the therapeutic methods of the invention are achieved involves killing cells expressing STEAP2 in the presence of effector cells, for example, through CDC, apoptosis, ADCC, phagocytosis, or a combination of two or more of these mechanisms. Cells expressing STEAP2 that can be inhibited or killed using the bispecific antigen binding molecules of the invention include, for example, prostate tumor cells.

The antigen-binding molecules of this invention can be used to treat, for example, primary and/or metastatic tumors arising in the prostate, bladder, cervix, lung, colon, kidney, breast, pancreas, stomach, uterus and/or ovaries . In certain embodiments, the bispecific antigen binding molecules of the invention are used to treat one or more of the following types of cancer: prostate cancer, bladder cancer, cervical cancer, lung cancer, colon cancer, kidney cancer, breast cancer, pancreatic cancer, stomach cancer, uterine cancer and ovarian cancer. In accordance with some embodiments of the present invention, anti-STEP2 antibodies or anti-STEAP2/anti-CD3 bispecific antibodies are useful for treating a patient suffering from castration-resistant prostate cancer. In other related embodiments, the invention provides methods comprising administering an anti-STEP2 antibody or an anti-CD3/anti-STEP2 bispecific antigen binding molecule, as described herein, to a patient suffering from castration-resistant prostate cancer.

Analytical/diagnostic methods known in the art, such as tumor scanning, etc., can be used to determine whether a patient has a tumor that is castration resistant.

The present invention also includes methods for treating residual tumor in a subject. As used herein, the term residual tumor means the presence or presence of one or more cancer cells in a subject after treatment with anticancer therapy.

In accordance with some aspects, the present invention provides methods for treating a disease or disorder associated with STEAP2 expression (eg, prostate cancer), comprising administering one or more of the anti-STEP2 or bispecific antigen-binding molecules described elsewhere herein to a subject following how the subject was determined to have prostate cancer (eg, castration-resistant prostate cancer). For example, the present invention includes methods of treating prostate cancer comprising administering an anti-STEP2 antibody or an anti-CD3/anti-STEP2 bispecific antigen binding molecule to a patient after 1 day, 2 days, 3 days, 4 days, 5 days, 6 days, 1 week , 2 weeks, 3 weeks, or 4 weeks, 2 months, 4 months, 6 months, 8 months, 1 year or more after the subject received hormonal therapy (eg, androgen deprivation therapy).

Combination therapies and formulations.

The present invention provides methods that include administering a pharmaceutical composition containing any of the exemplary antibodies and bispecific antigen-binding molecules described herein in combination with one or more additional therapeutic agents. Illustrative additional therapeutic agents that can be

- 42 044194 to combine or administer in combination with an antigen binding molecule according to this invention include, for example, an EGFR antagonist (for example, an anti-EGFR antibody [for example, cetuximab or panitumumab] or a small molecule EGFR inhibitor [for example, gefitinib or erlotinib]), an antagonist of another a member of the EGFR family such as Her2/ErbB2, ErbB3 or ErbB4 (for example, an anti-ErbB2, anti-ErbB3 or anti-ErbB4 antibody or a small molecule inhibitor of ErbB2, ErbB3 or ErbB4 activity), an EGFRvIII antagonist (for example, an antibody that specifically binds EGFRvIII), cMET antagonist (eg, anti-cMET antibody), IGF1R antagonist (eg, anti-IGF1R antibody), B-raf inhibitor (eg, vemurafenib, sorafenib, GDC-0879, PLX-4720), PDGFR-α inhibitor (eg, anti -PDGFR-a antibody), PDGFR-β inhibitor (e.g., anti-PDGFR-e antibody), VEGF antagonist (e.g., VEGF decoy, see, e.g., US 7,087,411, (also referred to herein as VEGF inhibitory fusion protein ), anti-VEGF antibody (eg, bevacizumab), small molecule VEGF receptor kinase inhibitor (eg, sunitinib, sorafenib or pazopanib)), DLL4 antagonist (eg, anti-DLL4 antibody described in US 2009/0142354 such as REGN421), antagonist Ang2 (eg, anti-Ang2 antibody described in US 2011/0027286, such as H1H685P), FOLH1 antagonist (PSMA), PRLR antagonist (eg, anti-PRLR antibody), STEAP1 or STEAP2 antagonist (eg, anti-STEAP1 antibody or anti-STEP2 antibody), TMPRSS2 antagonist (e.g. anti-TMPRSS2 antibody), MSLN antagonist (e.g. anti-MSLN antibody), CA9 antagonist (e.g. anti-CA9 antibody), uroplakin antagonist (e.g. anti-uroplakin antibody), etc. Other agents that may be advantageously administered in combination with the antigen binding molecules of the invention include cytokine inhibitors, including small molecule cytokine inhibitors and antibodies that bind to cytokines such as IL-1, IL-2, IL-3, IL-4, IL-5, IL-6, IL-8, IL-9, IL-11, IL-12, IL-13, IL-17, IL-18 or their respective receptors. Pharmaceutical compositions of the present invention (e.g., pharmaceutical compositions containing an anti-CD3/anti-STEAP2 bispecific antigen-binding molecule as described herein) can also be administered as part of a treatment regimen comprising one or more therapeutic combinations selected from ICE: ifosfamide ( eg Ifex®), carboplatin (eg Paraplatin®), etoposide (eg Etopophos®, Toposar®, VePesid®, VP-16); DHAP: dexamethasone (eg, Decadron®), cytarabine (eg, Cytosar-U®, cytosine arabinoside, ara-C), cisplatin (eg, Platinol®-AQ); and ESHAP: etoposide (eg, Etopophos®, Toposar®, VePesid®, VP-16), methylprednisolone (eg, Medrol®), high dose cytarabine, cisplatin (eg, Platinol®-AQ).

This invention also includes therapeutic combinations containing any of the antigen binding molecules mentioned herein and an inhibitor of one or more of VEGF, Ang2, Dll4, EGFR, ErbB2, ErbB3, ErbB4, EGFRvIII, cMet, IGF1R, B-raf, PDGFR- α, PDGFR-β, FOLH1 (PSMA), PRLR, STEAP1, STEAP2, TMPRSS2, MSLN, CA9, uroplakin, or any of the above cytokines, wherein the inhibitor is an aptamer, antisense molecule, ribozyme, siRNA (small interfering RNA), peptide body , nanobody or antibody fragment (for example, Fab fragment; F(ab) ₂ fragment; Fd fragment; Fv fragment; scFv; dAb fragment (domain antibody) or other engineered molecules, for example, diabodies, tribodies, tetrabodies, mini-antibodies and minimum recognition units). The antigen-binding molecules of the invention can also be administered and/or co-formulated in combination with antivirals, antibiotics, analgesics, corticosteroids and/or NSAIDs (non-steroidal anti-inflammatory drug). The antigen-binding molecules of the invention can also be administered as part of a treatment regimen that also includes radiation therapy and/or conventional chemotherapy.

Additional therapeutically active component(s) can be administered immediately before, simultaneously with, or shortly after administration of the antigen binding molecule of the present invention; (for the purposes of this invention, such administration regimens are considered to be administration of an antigen-binding molecule in combination with an additional therapeutically active component).

The present invention includes pharmaceutical compositions in which the antigen binding molecule of the present invention is co-formulated with one or more additional therapeutically active component(s) as described elsewhere herein.

Application schemes.

In accordance with some embodiments of the present invention, multiple doses of an antigen binding molecule (eg, an anti-STEP2 antibody or a bispecific antigen binding molecule that specifically binds STEAP2 and CD3) may be administered to a subject over a period of time. Methods in accordance with this aspect of the invention involve sequentially administering multiple doses of an antigen binding molecule of the invention to a subject. As used herein, sequential administration means that each dose of the antigen binding molecule is administered to a subject at different points in time, eg, on different days separated by a predetermined interval (eg, hours, days, weeks, or months). The present invention includes methods that involve sequentially administering to a patient a single initial dose of an antigen-binding molecule, followed by one or more secondary doses of the antigen-binding molecule, and optionally followed by one or more tertiary doses of the antigen-binding molecule.

The terms initial dose, secondary doses and tertiary doses refer to the time sequence of administration of the antigen binding molecule according to the invention. Thus, the starting dose is the dose that is administered at the beginning of the course of treatment (also called the initial dose); secondary doses are doses administered after the initial dose; and tertiary doses are doses administered after secondary doses. The initial, secondary and tertiary doses may contain the same amount of antigen-binding molecule, but generally may differ from each other in terms of frequency of administration. However, in certain embodiments, the amount of antigen-binding molecule contained in the initial, secondary, and/or tertiary doses differs from each other (eg, adjusted up or down as needed) over the course of treatment. In some embodiments, two or more (eg, 2, 3, 4, or 5) doses are administered at the beginning of the treatment regimen as loading doses, followed by subsequent doses that are administered less frequently (eg, maintenance doses).

In one illustrative embodiment of the present invention, each secondary and/or tertiary dose is administered from 1 to 26 (eg, 1.11/ ₂ , 2, 21/2, 3, 31/2, ₄ , 41/2, 5, 51/ 2, 6, 61/2, 7 _, 71/2, 8, 81/2, 9, 91/2, 10, 101/2, 11, 111/2, 12, 121/2, 13, 131/2, 14, 141/2, 15, 151/2, 16, 161/2, 17, 171/2, 18, 181/2, 19, 191/2, 20, 201/2, 21, 211/2, 22, 221/2, 23, 231/2, 24, 241/2, 25, 251/2, 26, 261/2, or more) weeks after the immediately preceding dose. The phrase immediately preceding dose, as used herein, means, in a multiple administration sequence, the dose of an antigen binding molecule that is administered to a patient before the next dose in the sequence without intervening doses.

The methods of this aspect of the invention may include administering to the patient any number of secondary and/or tertiary doses of an antigen binding molecule (eg, an anti-STEP2 antibody or a bispecific antigen binding molecule that specifically binds STEAP2 and CD3). For example, in certain embodiments, only one secondary dose is administered to the patient. In other embodiments, the patient is administered two or more (eg, 2, 3, 4, 5, 6, 7, 8 or more) secondary doses. Likewise, in certain embodiments, only one tertiary dose is administered to the patient. In other embodiments, the patient is administered two or more (eg, 2, 3, 4, 5, 6, 7, 8 or more) tertiary doses.

In embodiments involving multiple secondary doses, each secondary dose can be administered at the same frequency as the other secondary doses. For example, each secondary dose can be administered to the patient 1-2 weeks after the immediately preceding dose. Likewise, in embodiments of the invention that include multiple tertiary doses, each tertiary dose can be administered at the same frequency as the other tertiary doses. For example, each tertiary dose may be administered to the patient 2-4 weeks after the immediately preceding dose. Alternatively, the frequency with which secondary and/or tertiary doses are administered to the patient may vary during the course of treatment. The frequency of administration may also be adjusted during the course of treatment by the attending physician depending on the needs of the individual patient after clinical examination.

Diagnostic applications of antibodies.

Anti-STEAP2 antibodies according to this invention can also be used to detect and/or measure STEAP2 or STEP2-expressing cells in a sample, for example, for diagnostic purposes. For example, an anti-STEP2 antibody or fragment thereof can be used to diagnose a pathological condition or disease characterized by aberrant expression (eg, overexpression, underexpression, lack of expression, etc.) of STEAP2. Exemplary diagnostic assays for STEAP2 may include, for example, contacting a sample obtained from a patient with an anti-STEP2 antibody of the invention, wherein the anti-STEP2 antibody is labeled with a detectable label or reporter molecule. Alternatively, an unlabeled anti-STEP2 antibody may be used in diagnostic applications in combination with a secondary antibody that is itself labeled for detection. The detectable label or reporter molecule may be a radioisotope such as ³ H, ¹⁴ C, ³² P, ³⁵ S, or ¹²⁵ I; a fluorescent or chemiluminescent moiety such as fluorescein isothiocyanate or rhodamine; or an enzyme such as alkaline phosphatase, beta-galactosidase, horseradish peroxidase or luciferase. Another exemplary diagnostic use of anti-STEP2 antibodies of the invention includes ⁸⁹ Zr-labeled, such as ⁸⁹ Zr-desferrioxamine-labeled, antibody for the purpose of non-invasively identifying and tracking tumor cells in a subject (eg, positron emission tomography (PET) imaging). (See, for example, Tavare, R. et al. Cancer Res. 2016 Jan 1;76 (1):73-82; and Azad, B.B. et al. Oncotarget. 2016 Mar 15;7(11):12344- 58.) Specific exemplary assays that can be used to detect or measure STEAP2 in a sample include enzyme-linked immunosorbent assay (ELISA), radioimmunoassay (RIA), and flow cytometry (FACS).

Samples that can be used in STEAP2 diagnostic assays in accordance with this invention include any tissue or fluid sample that can be obtained from patients

- 44 044194 that which contains detectable amounts of the STEAP2 protein or its fragments in normal conditions or in pathology. Typically, STEAP2 levels in a given sample obtained from a healthy patient (eg, a patient not suffering from a disease or condition associated with abnormal STEAP2 levels or activity) will be measured for the purpose of initially establishing a baseline or standard STEAP2 level. This baseline STEAP2 level can then be compared to STEAP2 levels measured in samples obtained from individuals suspected of having a STEAP2-related disease (eg, a tumor containing STEAP2-expressing cells) or condition.

Examples

The following examples are provided to provide those skilled in the art with complete disclosure and description of how to make and use the methods and compositions of the invention, and are not intended to limit the scope of the invention to what the inventors believe to be their invention. Care has been taken to ensure accuracy in the numbers used (e.g. quantities, temperatures, etc.), but some experimental errors and deviations should be considered. Unless otherwise noted, parts are parts by weight, molecular weight is average molecular weight, temperature is in degrees Celsius, and pressure is at or near atmospheric pressure.

Example 1. Preparation of anti-STEP2 antibodies.

Ahtu-STEAP2 antibodies are produced by immunizing a genetically modified mouse using the human STEAP2 antigen or by immunizing a modified mouse containing DNA encoding the variable regions of the human immunoglobulin heavy chain and light chain using the human STEAP2 antigen.

Genetically modified mice were immunized with hSTEAP2 antigen (SEQ ID: 1899). Following immunization, splenocytes were collected from each mouse and (1) fused with mouse myeloma cells to maintain their viability and generate hybridoma cells and screened for STEAP2 specificity, or (2) B cells were sorted (as described in US 2007/0280945A1) using fragment of human STEAP2 as a sorting reagent that binds and identifies reactive antibodies (antigen-positive B cells).

Initially, chimeric antibodies to STEAP2 were isolated, having a human variable region and a mouse constant region. Antibodies were characterized and selected for desired characteristics including affinity, selectivity, etc. When necessary, mouse constant regions were replaced with the desired human constant region, such as wild type or modified IgG1 or IgG4 constant region, to generate a fully human anti-STEP2 antibody . Although the selected constant region may vary depending on the specific application, the variable region provides high affinity antigen binding and target specific characteristics. Antibody name symbols such as H1H11243N and H1M7804N indicate fully human H1H antibodies or chimeric H1M antibodies with human variable regions/mouse constant regions. Antibodies identified by the hybridoma method are designated by antibody identification numbers ending in N or N2. Antibodies identified by the B cell sorting method are designated by antibody identification numbers ending in P or P2.

Certain biological properties of exemplary anti-STEAP2 antibodies produced in accordance with the methods of this example are described in detail in the examples below.

Sequences of amino acids and nucleic acids of the variable regions of the heavy and light chains of anti-STEP2 antibodies.

In table 1 shows the amino acid sequence identifiers of the variable regions and CDRs of the heavy and light chains of selected anti-STEP2 antibodies according to the invention. The corresponding nucleic acid sequence identifiers are given in Table. 2.

- 45 044194

Table 1

Amino acid sequence identifiers

SEQ ID NO: Antibodies are indicated HCVR HCDR1 HCDR2 HCDR3 LCVR LCDR1 LCDR2 LCDR3 H1H11243N 2 4 6 8 10 12 14 16 Н1Н11878Р 18 20 22 24 26 28 thirty 32 N1N11880R 34 36 38 40 42 44 46 48 Н1Н11888Р2 50 52 54 56 58 60 62 64 Н1Н11892Р2 66 68 70 72 58 60 62 64 Н1Н11893Р2 74 76 78 80 58 60 62 64 N1N11894R2 82 84 86 88 58 60 62 64 N1N11895R2 90 92 94 96 58 60 62 64 N1N11896R2 98 100 102 104 58 60 62 64 N1N11897R2 106 108 110 112 114 116 118 120 Н1Н7968Р 122 124 126 128 130 132 134 136 Н1Н7969Р 138 140 142 144 146 148 150 152 Н1Н7970Р 154 156 158 160 162 164 166 168 Н1Н7971Р 170 172 174 176 178 180 182 184 Н1Н7972Р 186 188 190 192 194 196 198 200 H1M7804N 202 204 206 208 210 212 214 216 H1M7814N 218 220 222 224 226 228 230 232 H1M7832N 234 236 238 240 242 244 246 248 H2M11162N 250 252 254 256 258 260 262 264 H2M11163N 266 268 270 272 274 276 278 280 H2M11164N 282 284 286 288 290 292 294 296 H2M7806N 298 300 302 304 306 308 310 312 H2M7807N 314 316 318 320 322 324 326 328 H2M7809N 330 332 334 336 338 340 342 344 H2M7810N 346 348 350 352 354 356 358 360 H2M7811N 362 364 366 368 370 372 374 376 H2M7812N 378 380 382 384 386 388 390 392

table 2

Nucleic acid sequence identifiers

SEQ ID NO: Antibody designation HCVR HCDR1 HCDR2 HCDR3 LCVR LCDR1 LCDR2 LCDR3 H1H11243N 1 3 5 7 9 eleven 13 15 H1H11878P 17 19 21 23 25 27 29 31

- 46 044194

N1N11880R 33 35 37 39 41 43 45 47 Н1Н11888Р2 49 51 53 55 57 59 61 63 Н1Н11892Р2 65 67 69 71 57 59 61 63 Н1Н11893Р2 73 75 77 79 57 59 61 63 N1N11894R2 81 83 85 87 57 59 61 63 N1N11895R2 89 91 93 95 57 59 61 63 N1N11896R2 97 99 101 103 57 59 61 63 N1N11897R2 105 107 109 111 113 115 117 119 Н1Н7968Р 121 123 125 127 129 131 133 135 Н1Н7969Р 137 139 141 143 145 147 149 151 Н1Н7970Р 153 155 157 159 161 163 165 167 Н1Н7971Р 169 171 173 175 177 179 181 183 Н1Н7972Р 185 187 189 191 193 195 197 199 H1M7804N 201 203 205 207 209 211 213 215 H1M7814N 217 219 221 223 225 227 229 231 H1M7832N 233 235 237 239 241 243 245 247 H2M11162N 249 251 253 255 257 259 261 263 H2M11163N 265 267 269 271 273 275 277 279 H2M11164N 281 283 285 287 289 291 293 295 H2M7806N 297 299 301 303 305 307 309 311 H2M7807N 313 315 317 319 321 323 325 327 H2M7809N 329 331 333 335 337 339 341 343 H2M7810N 345 347 349 351 353 355 357 359 H2M7811N 361 363 365 367 369 371 373 375

Example 2 Antibodies against human STEAP2 selectively bind to cell lines expressing STEAP2 via FACS.

The ability of anti-STEP2 antibodies to selectively bind to cell lines endogenously expressing human six-transmembrane epithelial antigen prostate 2 (STEAP2) was determined by FACS analysis.

Briefly, 1 × 10 ⁵ cells were incubated with 10 μg/ml anti-STEAP2 antibody or isotype control antibody for 30 min on ice in antibody dilution buffer. After one wash with antibody dilution buffer, cells were incubated with 10 μg/ml PE-conjugated anti-human or anti-mouse Fc secondary antibodies for 30 min on ice. After one additional wash, samples were incubated with Cytofix (1% formaldehyde) for 20 min. After the final wash, samples were filtered through a 96-well Pall filter unit, run on a Hypercyt® cytometer, and analyzed in ForeCyt™ (IntelliCyt, Albuquerque, NM). Average fluorescence intensities (AFIs) were expressed as fold changes above unstained levels (background). The average fold above background was determined at antibody concentrations of 100–300 nM. To determine the EC50 of cell binding, mAb concentrations ranged from 300 nM to 5 pM, and EC50 values were determined from a four-parameter logistic equation using a 12-point dose-response curve (GraphPad Prism).

Tables 3A and 3B. FACS binding properties of anti-STEP2 antibodies to STEP2-expressing and STEAP2-negative cell lines.

- 47 044194

Table FOR

Binding properties of selected anti-STEAP2 antibodies to human Fc

STEAP2- expressing S TE AR 2 - negative Cell line HEK293 S4-2 FADU SKBR3 Raji Antibody Average K.V.F. Average K.V.F. Average EC ₅₀ (nM) Average K.V.F. Average K.V.F. Average K.V.F. H1H7809N 31 89 >50 21 37 3 H1H7810N 3 26 BUT 3 3 2 H1H7811N 5 65 >50 3 5 4 H1H7814N 2 96 4 2 2 2 Н1Н7972Р 4 119 9 2 3 59 H1H11162N 10 104 1 3 3 2 H1H11163N 6 90 3 2 3 2 H1H11164N 10 117 2 3 2 2 H1H11243N 3 41 BUT 2 3 2 Н1Н11878Р 5 27 >50 BUT BUT BUT N1N11880R 7 6 5 BUT BUT BUT Н1Н11888Р2 9 6 3 BUT BUT BUT Н1Н11892Р2 3 4 BUT BUT BUT BUT Н1Н11893Р2 3 9 21 BUT BUT BUT N1N11894R2 4 7 BUT BUT BUT BUT N1N11895R2 4 6 16 BUT BUT BUT N1N11896R2 7 6 21 BUT BUT BUT N1N11897R2 16 29 >50 BUT BUT BUT Isotype control -I hlgGl 1 2 BUT 1 2 4 Isotype control--!! hlgGl 1.7 ± 0.5 2 BUT 3 2 BUT anti-human FE 1.1 ± 0.1 1 BUT 1 2 1 Unpainted 1±0 1 BUT 1 1 BUT

K.V.F: fold above background; N/O: not found.

- 48 044194

Table 3 B

Binding properties of aHTu-STEP2 antibodies generated by hybridoma

STEAP2expressing S ΤΕΆΡ2 - negative Cell line HEK293 S4-2 FADU SKBR3 Raji Antibody Average K.V.F. Average K.V.F. Average EC ₅₀ (nM) Average K.V.F. Average K.V.F. Average K.V.F. H1M11243N But 116 23 BUT BUT BUT H1M11249N But eleven BUT But But But H2M11160N But 8 BUT But But But H2M11162N But 365 4.9 But But But H2M11163N But 256 13 But But But H2M11164N But 360 3.7 But But But H2M11166N But 12 BUT But But But H2M11168N But 9 BUT But But But H2M11245N But 18 BUT But But But H2M11246N But eleven BUT But But But H2M11247N But 482 18 But But But H2M11248N But 50 >50 But But But H3M11161N But 5 BUT But But But H3M11165N But 13 BUT But But But H3M11167N But 14 BUT But But But H3M11244N But 2 BUT But But But Isotype control mlgGl But 2 BUT But But But Isotype control mIgG2 But 3 BUT But But But Isotype control mIgG3 But 4 BUT But But But anti-human, FE 1.1 ± 0.1 1 BUT 1 2 1 Unpainted 1±0 1 But 1 1 But

K.V.F: fold above background; N/O: not found.

As the results in table demonstrate. 3A and 3B, several anti-STEP2 antibodies specifically bind to C4-2 prostate adenocarcinoma cell lines with high STEAP2 expression at levels greater than 50-fold above background, with low nM EC50 values, by FACS. Some anti-STEP2 antibodies also bind weakly to HEK293 cells with low STEAP2 expression. Little binding was observed for most anti-STEP2 antibodies on STEP2-negative FADU, SK-BR-3 and Raji cells. This Example illustrates the ability of several anti-STEP2 antibodies according to this invention to specifically and selectively bind to cell lines with high expression of STEAP2.

Example 3 Antibodies against human STEAP2 exhibit strong internalization and specificity for human STEAP2.

The ability of the anti-STEP2 antibodies of this invention to selectively bind to cell lines expressing STEAP2 is described (see Example 2 - FACS binding). The internalization properties of the anti-STEP2 antibodies of the present invention were also then evaluated.

Briefly, 20,000 C4-2 cells were seeded in 9b-well plates coated with PDL. The next day, cells were incubated with anti-human STEAP2 antibodies (10 μg/ml) for 30 min on ice, followed by washing with two PBS (phosphate-buffered saline). Cells were then incubated with alexa488-conjugated anti-hFc Fab secondary antibody for 30 min on ice, followed by two additional washes with PBS (phosphate-buffered saline). Antibodies were allowed to internalize for 1 h at 37°C in internalization buffer (PBS+2% FBS (fetal bovine serum)) or stored at 4°C. Cells were fixed in 4% formaldehyde, nuclei were stained with DRAQ5 (Cell signaling) and imaged on ImageXpress micro XL (Molecular Devices).

Qualitative visual assessment of the total binding intensity and the intensity of antibodies that were internalized into the vesicles was performed and scored according to the following criteria: - (no internalization or binding), + (weak internalization or binding), ++ (moderate internalization or binding), and + ++ (strong internalization or binding).

As the results in table demonstrate. 4, some antibodies showed the potential for strong internalization on the C4-2 cell line. Overall, strong internalization correlated with the highest levels of overall binding intensity.

Selected anti-STEAP2 antibodies were then tested for binding to other members of the human STEAP (h) family (STEAP1, STEAP3 and STEAP4). To evaluate the specificity of anti-STEP2 antibodies, plasmid constructs expressing hSTEAP1, hSTEAP2, hSTEAP3, or hSTEAP4 fused to green fluorescent protein (GFP) were transiently introduced into HEK293 cells using a Lipofectamine 2000-based technique. After 48 h, transiently transfected cells were stained with anti-STEP2 antibodies. -STEP2 antibodies and visualized as described above for internalization assay. Wells with GFP-positive cells that bound anti-STEAP2 antibodies were scored as positive (+), and those that did not bind anti-STEP2 antibodies were scored as negative (-). All antibodies tested bound to hSTEP2-GFP-positive cells but did not bind to STEAP1-GFP, STEAP3-GFP, or STEP4-GFP-positive cells, confirming the specificity of binding to human STEAP2. The binding results are given in table. 5.

Briefly, several anti-STEP2 antibodies of this invention exhibit potent internalization capabilities and are specific binders for human STEAP2.

Table 4

Qualitative assessment of the internalization and overall binding properties of anti-STEP2 antibodies in the C4-2 cell line with high STEAP2 expression

Specificity of anti-STEP2 antibodies: assessment of binding to hSTEAP1, hSTEAP2, hSTEAP3 or hSTEAP4 fused to GFP

HEK293 cells transfected with STEAP/GFP hybrid plasmids Antibodies hSTEAPl HSTEAP2 HSTEAP3 HSTEAP4 N2M7807 - + - - N2M7810 - + - - N2M7811 - + - - N1M7814 - + - - Н1Н7972 - + - -

- 50 044194

Example 4: Generation of bispecific antibodies that bind STEAP2 and CD3.

The present invention provides bispecific antigen binding molecules that bind CD3 and STEAP2; such bispecific antigen binding molecules are also referred to herein as anti-STEP2/αanti-CD3 or anti-STEAP2xCD3 bispecific molecules. The Ahtu-STEP2 portion of the anti-STEAP2/anti-CD3 bispecific molecule is useful for targeting tumor cells that express six-transmembrane epithelial prostate antigen 2 (STEAP2) (STEAP2), and the anti-CD3 portion of the bispecific molecule is useful for activating T cells . Simultaneous binding of STEAP2 on the tumor cell and CD3 on the T cell facilitates targeted killing (cell lysis) of the target tumor cell by the activated T cell.

Bispecific antibodies containing an anti-STEP2-specific binding domain and an anti-CD3-specific binding domain were recombinantly constructed using standard molecular cloning techniques and expressed in CHO cells, with an anti-STEP2 antigen binding domain and an α anti-CD3 antigen binding domain each which contains different, specific HCVRs paired with a common LCVR. In the present bispecific antibodies, molecules were constructed using a heavy chain from an anti-CD3 antibody, a heavy chain from an anti-STEP2 antibody, and a common light chain from an anti-STEP2 antibody, and expressed in CHO cells. In some cases, bispecific antibodies can be constructed using a heavy chain from an anti-CD3 antibody, a heavy chain from an anti-STEP2 antibody, and a light chain from an anti-CD3 antibody, or an antibody light chain that is known to be heterogeneous or bind effectively to various heavy chain arms such as Vk1-39JK5 or Vk3-20JK1.

The bispecific antibodies described in the following examples consist of anti-CD3 binding arms having different binding affinities to the human soluble hCD3ε/δ heterodimeric protein (as described in Example 12 herein); and human STEAP2 (see examples 1-2 above). The exemplary bispecific antibodies were produced using a modified (chimeric) IgG4 Fc domain as disclosed in US Patent Application Publication No. US20140243504A1 published August 28, 2014.

A brief description of the constituent parts of the antigen binding domains of the various engineered anti-STEAP2xCD3 bispecific antibodies, as set out in Table. 6.

Table 6

Construction of STEAP2xCD3 bispecific antibodies

Bispecific Antibody Identifier Anti-STEAR2 Antigen-binding domain Ahth-CD3 Antigen binding domain Common light chain variable region Varna, white region of heavy chains Variable region severe chains BSSTEAP2/CD3-001 H2M11162N (SEQ ID NO:250) CD3-VH-G (SEQ ID NO: 1730) H2M11162N (SEQ ID NO: 258) BSSTEAP2/CD3-002 CD3-VH-G5 (SEQ ID NO: 1762) BSSTEAP2/CD3-003 CD3-VH-G20 (SEQ ID NO:1866) BSSTEAP2/CD3-004 H1H7814N (SEQ ID NO:218) H1H7251P (SEQ ID NO: 1570) H1H7814N (SEQ ID NO: 226) BSSTEAP2/CD3-005 H1H11162 (SEQ ID NO:250) H1H7208P (SEQ ID NO:1490) H1H11162 (SEQ ID NO: 258) BSSTEAP2/CD3-006 CD3-VH-P (SEQ ID NO:1882) BSSTEAP2/CD3-007 H1H7195P (SEQ ID NQ:1450)

- 51 044194

BSSTEAP2/CD3-008 N1N11163 (SEQ ID NO:266) H1H7208P (SEQ ID NO:1490) H1H11163 (SEQ NO:274) ID BSSTEAP2/CD3-009 N1N11164 (SEQ ID NO:282) H1H7208P (SEQ ID NQ:1490) H1H11164 (SEQ NO:290) ID H1H7198P BSSTEAP2/CD3-010 H1H7809N (SEQ ID NQ:330) (SEQ ID NO:1466) H1H7809N (SEQ NO:339) ID BSSTEAP2/CD3-011 H1H7203P (SEQ ID NO:1474)

Light chains listed in table. 6 were common to both the CD3- and STEP2-targeting arms of bispecific antibodies. in table 1 and 2 show the amino acid and nucleic acid sequence identifiers, respectively, for the various heavy chain variable regions and their corresponding CDR, anti-STEP2 arms (ie, HCVR and LCVR are derived from H2M11162N) for the construction of the bispecific antibodies of this example. in table 15 and 16 show the amino acid and nucleic acid sequence identifiers, respectively, for the various heavy chain variable regions and their corresponding CDRs, the anti-CD3 arms of the bispecific antibodies of this example.

Example 5 Anti-STEP2/anti-CD3 bispecific antibodies exhibit potent antitumor efficacy in vivo.

To determine the effectiveness of the exemplary anti-STEAP2/anti-CD3 bispecific antibodies in vivo, studies were conducted in immunocompromised mice bearing prostate cancer xenografts.

Efficacy of anti-STEP2/anti-CD3 bispecific antibodies in human tumor xenograft models

To evaluate the in vivo efficacy of bispecific anti-STEAP2/anti-CD3 in human tumor xenograft studies, NOD scid gamma (NSG) mice (NOD - non-obese diabetes) / SCID - severe combined immunodeficiency) (Jackson Laboratories, Bar Harbor, Maine ) were co-implanted with human peripheral blood mononuclear cells (PBMCs; ReachBio LLC., Seattle, WA) with C4-2 human prostate cancer cells (MD Anderson Cancer Center, Houston, TX), which endogenously express STEAP2.

Briefly, 5.0 × ^{10 6} C4-2 cells were co-implanted subcutaneously (SC) with 1.25 × ^{10 6} human PBMCs in a 50/50 mixture of Matrigel matrix (BD Biosciences, San Jose, CA) into the right flank of male NSG mice. Mice were treated intraperitoneally (ip) on the day of implantation (immediate treatment model) with bispecific anti-STEAP2/anti-CD3 BSSTEAP2/CD3-001, BSSTEAP2/CD3-002 or BSSTEAP2/CD3003 or isotype control at a dose of 0.1 or 0. 01 mg/kg (N=5 mice/group).

Tumor size was measured 2 times a week using a compass, and tumor volume was calculated as volume=(length x width ² )/2. Data are presented as tumor size (mm ³ ) at the end point of the study, 46 days after tumor implantation (Table 7).

Table 7 Efficacy of anti-STEP2/anti-CD3 bispecific antibodies in an immunocompromised xenograft model: immediate dosing

Tumor model/mouse strain Bispecific Antibody Identifier Dose (mg/kg) N Tumor size (mm ³ ) 4 6 days after tumor implantation (mean ± SD) C4-2/NSG BSSTEAP2/CD3-001 0.1 5 18.0 ± 14.0 0.01 5 23.0 ± 220 BSSTEAP2/CD3-002 0.1 5 15.0 + 12.0 0.01 5 17.0 ± 8.0 BSSTEAP2/CD3-003 oh 1 5 19.0 ± 12.0 0.01 5 25.0 ± 21.0 Control bispecific oh 1 5 1020.0 ± 922.0 antibody

- 52 044194

As the results in table demonstrate. 7 BSSTEAP2/CD3-001, BSSTEAP2/CD3-002 and

BSSTEAP2/CD3-003 significantly suppressed tumor growth compared to isotype control when tumor size was measured at study endpoint. Importantly, anti-STEP2/anti-CD3 bispecific antibodies were effective in inhibiting C4-2 tumor growth even at the lowest dose of 0.1 mg/kg.

Example 6. Preparation and characterization of the conjugate.

All monoclonal antibodies were expressed in CHO cells and purified using protein A. An isotype control was also prepared in a similar manner. A non-binding isotype control antibody was derived from an immunological antigen not related to oncology.

Antibody (10 mg/ml) in 50 mM HEPES, 150 mM NaCl, pH 7.5 was treated with 1 mM dithiothreitol at 37°C for 30 min. After gel filtration (G-25, pH 4.5, sodium acetate), the payload compound 7 derivative with maleimide linker (1.2 equivalents/SH group) in DMSO (10 mg/ml) was added to the reduced antibody and the mixture was adjusted to to pH 7.0 with 1 M HEPES (pH 7.4). Compound 7 and methods for preparing the compound are described in PCT Publication No. WO2014/145090, published September 18, 2014, which is incorporated herein by reference in its entirety. After 1 h, the reaction was quenched with an excess of N-ethylmaleimide. The conjugates were purified by size exclusion chromatography and sterile filtered. Protein and payload linker concentrations were determined by UV spectral analysis. Size exclusion HPLC determined that all conjugates used were >95% monomeric, and RP-HPLC determined that the unconjugated linker payload was <0.5%. The output is given in the table. 8 based on determination of protein titer. All conjugated antibodies were analyzed by ultraviolet light for payload linker content in accordance with Hamblett et al., Cancer Res 2004 10 7063. The results are shown in table. 8.

A conjugate containing compound 60 can be prepared using a similar method. Compound 60 and methods for preparing the compound are described in PCT Publication No. WO2016/160615 (Example 20), published October 6, 2016, which is incorporated herein by reference in its entirety. Compound 60 is maytansine-N-methyl-L-alanine-(3-methoxy-4amino)benzamido-Cit-Val-Cap-Mal.

Table 8

Summary of payload (chemotoxic drug) and antibody-drug conjugate parameters

Compound ε 252 nm (ohm ^-1 Μ ^-1 ) ε 280 nm (cm ^-1 M ¹ ) 7 [Mai ta nzine - 3 -N-methyl -L- (S)-alanine- pr opanamidyl-3-N-methylN-[4-(amino-citrulline- valine-hexanamide-6maleimidyl benzyl]carbamate] 50600 8100 Antibody ε 252 nm (ohm ^-1 Μ ^-1 ) ε 280 nm (cm ^-1 M' ¹ ) H1H7814N 110440 212400 Isotype control 75113 218360 Antibody conjugate Payload: antibody (UV) Exit, % H1H7814N-7 2.7 48 Isotype control- 7 3.0 48

Example 7 Anti-STEP2 antibody drug conjugates (ADCs) are potent inhibitors of tumor growth in STEAP2-positive In vivo prostate cancer xenograft models.

A. To determine the in vivo efficacy of anti-STEP2 antibodies conjugated to Compound 7, studies were conducted in immunocompromised mice bearing STEP2-positive prostate cancer xenografts.

For these studies, male SCID mice (Taconic, Hudson, NY) were implanted with cells

- 53 044194

C4-2 endogenously expressing STEAP2. Once tumors reached an average volume of 200-250 mm ³ (~13-17 days), mice were randomized into treatment groups and administered either anti-STEP2-conjugated antibody, non-binding conjugated antibody, or vehicle. In these in vivo studies, antibodies were administered once and tumors were then monitored until the average tumor size reached approximately 1500-2000 mm ³ in the vehicle-only group (~40-50 days). Treatment groups demonstrating efficacy were monitored over a longer period of time (80-110 days).

In an initial study, an exemplary anti-STEP2 antibody conjugated to Compound 7 was tested for effectiveness in reducing C4-2 tumor volume. Mice received a single dose of anti-STEP2 and control ADCs at 10, 20, or 40 mg/kg on day 13 postimplantation. As shown in FIG. 1, H1H7 8 41N-7 (DAR 2.92) strongly inhibited tumor growth at all doses tested. At the highest dose (40 mg/kg), H1H784N-7 effectively reduced tumor size, although the nonbinding control antibody (40 mg/kg) also showed an effect on tumor volume. At all doses tested, H1H784N-7 reduced tumor size more than the control conjugated antibody.

In the second study, anti-STEP2 ADC was administered at a dose of 5 and 20 mg/kg, and a control antibody was administered at a dose of 20 mg/kg on day 14 after implantation. As shown in FIG. 2, H1H7841N-7 (DAR 2.92) strongly inhibited tumor growth at a dose of 20 mg/kg, as in the previous experiment, but showed reduced effectiveness at a dose of 5 mg/kg. The control antibody at 20 mg/kg showed no difference from the vehicle control.

In an additional study, H1H7841N-7 (DAR 2.7) and control antibody were administered in μg/kg drug equivalents based on ADC drug:antibody ratios (DAR). The dose was 150 μg/kg, administered on the 17th day after implantation (Fig. 3). H1H7841N-7 potently inhibited tumor growth at 150 μg/kg, demonstrating tumor regression within 42 days postimplantation and 25 days postinjection. At this point, tumor growth began to resume. Tumor growth with the control antibody at this dose did not differ from the solvent control.

B. In similar studies, male SCID mice (Taconic, Hudson, NY) were implanted with C4-2 cells endogenously expressing STEAP2. An exemplary anti-STEP2 (H1H7814N) antibody conjugated to compound 60 was tested for efficacy in C4-2 tumor regression. Mice received a single dose of anti-STEP2 ADC, isotype control ADC (binds to irrelevant antigen), or vehicle (PBS) at 2.5 mg/kg on day 29 postimplantation. Tumor volume and body weight were recorded on day 0 (day of administration) and on days 4, 6, 8, 12, 14 and 20 after administration. As shown in FIG. 4, H1H7814N-60 (DAR 3.6) strongly inhibited tumor growth at the dose tested, demonstrating tumor regression up to 20 days post-injection (49 days post-implantation). The percentage change in body weight for the test ADC did not exceed -2.01% after 14 days (after administration of H1H7814N-60) compared to mice treated with the control Ab-ADC (antibody-drug conjugate), in which the percentage change in body Macci observed from - 4.02% to - 11.55% after 14 days.

Example 8. Preparation of anti-CD3 antibodies.

Ahtu-CD3 antibodies were produced by immunizing a modified mouse containing DNA encoding the variable regions of the human immunoglobulin kappa heavy and light chain with cells expressing CD3 or with DNA encoding CD3. The antibody response was monitored using a CD3-specific immunoassay. Once the desired immune response was achieved, splenocytes were collected and fused with mouse myeloma cells to maintain their viability and generate hybridoma cell lines. Hybridoma cell lines were screened and selected to identify cell lines that produce CD3-specific antibodies. Using this technique, several chimeric anti-CD3 antibodies (ie, antibodies having human variable domains and mouse constant domains) have been produced. In addition, several fully human anti-CD3 antibodies have been isolated directly from antigen-positive B cells without fusion with myeloma cells, as described in US 2007/0280945A1.

The biological properties of exemplary anti-CD3 antibodies produced in accordance with the methods of this Example are described in detail in the Examples below.

Example 9. Amino acid and nucleic acid sequences of the heavy and light chain variable regions

In table 9 shows the amino acid sequence identifiers of the variable regions and CDRs of the heavy and light chains of selected anti-CD3 antibodies according to the invention. The corresponding nucleic acid sequence identifiers are given in Table. 10. Methods for producing the anti-CD3 antibodies described herein can also be found in US Publication 2014/0088295, published March 27, 2014.

- 54 044194

Amino acid sequence identifiers

Table 9

SEQ ID NO: Antibody designation HCVR HCDR1 HCDR2 HCDR3 LCVR LCDR11 LCDR2 LCDR3 H1H2712N 4 02 404 406 408 410 412 414 416 H1M2692N 418 420 422 424 426 428 430 432 H1M3542N 434 436 438 440 442 444 446 448 H1M3544N 450 452 454 456 458 4 60 4 62 4 64 H1M3549N 466 4 68 4 70 472 474 476 4 78 480 H1M3613N 482 484 486 488 490 492 494 496 H2M2689N 498 500 502 504 506 508 510 512 H2M2690N 514 516 518 520 522 524 526 528 H2M2691N 530 532 534 53 6 538 540 542 544 H2M2704N 546 548 550 552 554 55 6 558 5 60 H2M2705N 562 564 566 568 570 572 574 576 H2M2706N 578 580 582 584 586 588 590 592 H2M2707N 594 596 598 600 602 604 606 608 H2M2708N 610 612 614 616 618 620 622 624 H2M2709N 626 628 630 632 634 63 6 638 640 H2M2710N 642 644 64 6 648 650 652 654 656 H2M2711N 658 660 662 664 666 668 670 672 H2M2774N 674 676 678 680 682 684 68 6 688 H2M2775N 690 692 694 69 6 698 700 702 704 H2M2776N 706 708 710 712 714 716 718 720 H2M2777N 722 724 726 728 730 732 734 736 H2M2778N 738 740 742 744 746 748 750 752 H2M2779N 754 756 758 760 762 764 766 768 H2M2789N 770 772 774 776 778 780 782 784 H2M2862N 786 788 790 792 794 79 6 798 800 H2M2885N 802 804 806 808 810 812 814 816 H2M2886N 818 820 822 824 82 6 828 830 832 H2M3540N 834 836 838 840 842 844 846 848 H2M3541N 850 852 854 856 858 8 60 8 62 8 64 H2M3543N 866 868 870 872 874 876 878 880 H2M3547N 882 884 886 888 890 892 894 896 H2M3548N 898 900 902 904 906 908 910 912 H2M3563N 914 916 918 920 922 924 926 928 Н1Н5751Р 930 932 934 93 6 938 940 942 944 Н1Н5752Р 946 948 950 952 954 95 6 958 9 60 Н1Н5753В 962 964 966 968 970 972 974 976 Н1Н5754В 978 980 982 984 986 988 990 992 Н1Н5755В 994 996 998 1000 1002 1004 1006 1008

- 55 044194

H1H5756B 1010 1012 1014 1016 1018 1020 1022 1024 H1H5757B 1026 1028 1030 1032 1034 1036 1038 1040 H1H5758B 1042 1044 1046 1048 1050 1052 1054 1056 H1H5761P 1058 1060 1062 10 64 1066 10 68 1070 1072 H1H5763P 1074 1076 1078 1080 1082 1084 1086 1088 H1H5764P 1090 1092 1094 1096 1098 1100 1102 1104 H1H5769P 1106 1108 1110 1112 1114 1116 1118 1120 H1H5771P 1122 1124 1126 1128 ISO 1132 1134 1136 H1H5772P 1138 1140 1142 1144 1146 1148 1150 1152 H1H5777P 1154 1156 1158 1160 1162 1164 1166 1168 H1H5778P 1170 1172 1174 1176 1178 1180 1182 1184 H1H5780P 1186 1188 1190 1192 1194 1196 1198 1200 H1H5781P 1202 1204 1206 1208 1210 1212 1214 1216 H1H5782P 1218 1220 1222 1224 1226 1228 1230 1232 H1H5785B 1234 1236 1238 1240 1242 1244 1246 1248 H1H5786B 1250 1252 1254 1256 1258 1260 1262 1264 H1H5788P 1266 1268 1270 1272 1274 1276 1278 1280 H1H5790B 1282 1284 1286 1288 1290 1292 1294 1296 H1H5791B 1298 1300 1302 1304 1306 1308 1310 1312 H1H5792B 1314 1316 1318 1320 1322 1324 1326 1328 H1H5793B 1330 1332 1334 1336 1338 1340 1342 1344 H1H5795B 1346 1348 1350 1352 1354 1356 1358 1360 H1H5796B 1362 1364 1366 1368 1370 1372 1374 1376 H1H5797B 1378 1380 1382 1384 1386 1388 1390 1392 H1H5798B 1394 1396 1398 1400 14 02 1404 1406 1408 H1H5799P 1410 1412 1414 1416 1418 1420 1422 1424 H1H5801B 1426 1428 1430 1432 1434 1436 1438 1440 H1H7194B 1442 1444 1446 1448 1634 1636 1638 1640 H1H7195B 1450 1452 1454 1456 1634 1636 1638 164 0 H1H7196B 1458 14 60 14 62 14 64 1634 1636 1638 164 0 H1H7198B 1466 14 68 1470 1472 1634 1636 1638 164 0

- 56 044194

Н1Н7203В 1474 1476 1478 1480 1634 1636 1638 1640 Н1Н7204В 1482 1484 1486 1488 1634 1636 1638 1640 Н1Н7208В 1490 1492 1494 1496 1634 1636 1638 164 0 Н1Н7211В 1498 1500 1502 1504 1634 1636 1638 164 0 Н1Н7221В 1506 1508 1510 1512 1634 1636 1638 164 0 Н1Н7223В 1514 1516 1518 1520 1634 1636 1638 164 0 Н1Н7226В 1522 1524 1526 1528 1634 1636 1638 164 0 Н1Н7232В 1530 1532 1534 1536 1634 1636 1638 1640 Н1Н7233В 1538 1540 1542 1544 1634 1636 1638 1640 Н1Н7241В 1546 1548 1550 1552 1634 1636 1638 1640 Н1Н7242В 1554 1556 1558 1560 1634 1636 1638 164 0 Н1Н7250В 1562 1564 1566 1568 1634 1636 1638 164 0 Н1Н7251В 1570 1572 1574 1576 1634 1636 1638 164 0 Н1Н7254В 1578 1580 1582 1584 1634 1636 1638 1640 Н1Н7258В 1586 1588 1590 1592 1634 1636 1638 164 0 Н1Н7269В 1594 1596 1598 1600 1634 1636 1638 1640 Н1Н7279В 1602 1604 160 6 1608 1634 1636 1638 164 0 Н1хН7221 G 1610 1612 1614 1616 1634 1636 1638 1640 Н1хН7221 G3 1618 1620 1622 1624 1634 1636 1638 1640 Н1ХН7221 G5 1626 1628 1630 1632 1634 1636 1638 1640

Table 10

Nucleic acid sequence identifiers

SEQ ID NO: Antibody designation HCVR HCDR1 HCDR2 HCDR3 LCDR1 LCDR2 LCDR3 H1H2712N 401 403 4 05 407 409 411 413 415 H1M2692N 417 419 421 423 425 427 429 431 H1M3542N 433 435 437 439 441 443 445 447 H1M3544N 449 451 453 455 457 459 4 61 4 63 H1M3549N 4 65 467 4 69 471 4 73 475 477 479

- 57 044194

H1M3613N 481 483 485 487 489 491 493 495 H2M2689N 497 499 501 503 505 507 509 511 H2M2690N 513 515 517 519 521 523 525 527 H2M2691N 529 531 533 535 537 539 541 543 H2M2704N 545 547 549 551 553 555 557 559 H2M2705N 561 563 5 65 567 5 69 571 573 5 75 H2M2706N 577 579 581 583 585 587 589 591 H2M2707N 593 595 597 599 601 603 605 607 H2M2708N 609 611 613 615 617 619 621 623 H2M2709N 625 627 629 631 633 635 63 7 639 H2M2710N 641 643 645 647 649 651 653 655 H2M2711N 657 659 661 663 665 667 669 671 H2M2774N 673 675 677 679 681 683 685 68 7 H2M2775N 689 691 693 695 697 699 701 703 H2M2776N 705 707 709 711 713 715 717 719 H2M2777N 721 723 725 727 729 731 733 735 H2M2778N 737 739 741 743 745 747 749 751 H2M2779N 753 755 757 759 761 763 765 767 H2M2789N 769 771 773 7 75 777 779 781 783 H2M2862N 785 787 789 791 793 795 797 799 H2M2885N 801 803 805 807 809 811 813 815 H2M2886N 817 819 821 823 825 827 829 831 H2M3540N 833 835 837 839 841 843 845 847 H2M3541N 849 851 853 855 857 859 8 61 8 63 H2M3543N 865 867 8 69 871 873 875 877 879 H2M3547N 881 883 885 887 889 891 893 895 H2M3548N 897 899 901 903 905 907 909 911 H2M3563N 913 915 917 919 921 923 925 927 H1H5751P 929 931 933 935 937 939 941 943 H1H5752P 945 947 949 951 953 955 957 959 H1H5753B 961 963 9 65 967 9 69 971 973 9 75 H1H5754B 977 979 981 983 985 98 7 989 991 H1H5755B 993 995 997 999 1001 1003 1005 1007 H1H5756B 1009 1011 1013 1015 1017 1019 1021 1023

- 58 044194

H1H5757B 1025 1027 1029 1031 1033 1035 103 7 1039 H1H5758B 1041 1043 1045 1047 1049 1051 1053 1055 H1H5761P 1057 1059 1061 1063 1065 1067 1069 1071 H1H5763P 1073 1075 1077 1079 1081 1083 1085 1087 H1H5764P 1089 1091 1093 1095 1097 1099 1101 1103 H1H5769P 1105 1107 1109 1111 1113 1115 1117 1119 H1H5771P 1121 1123 1125 1127 1129 1131 1133 1135 H1H5772P 1137 1139 1141 1143 1145 1147 1149 1151 H1H5777P 1153 1155 1157 1159 1161 1163 1165 1167 H1H5778P 1169 1171 1173 1175 1177 1179 1181 1183 H1H5780P 1185 1187 1189 1191 1193 1195 1197 1199 H1H5781P 1201 1203 1205 1207 1209 1211 1213 1215 H1H5782P 1217 1219 1221 1223 1225 1227 1229 1231 H1H5785B 1233 1235 1237 1239 1241 1243 1245 1247 H1H5786B 1249 1251 1253 1255 1257 1259 1261 1263 H1H5788P 1265 1267 1269 1271 1273 1275 1277 1279 H1H5790B 1281 1283 1285 1287 1289 1291 1293 1295 H1H5791B 1297 1299 1301 1303 1305 1307 1309 1311 H1H5792B 1313 1315 1317 1319 1321 1323 1325 1327 H1H5793B 1329 1331 1333 1335 1337 1339 1341 1343 H1H5795B 1345 1347 1349 1351 1353 1355 1357 1359 H1H5796B 1361 1363 1365 1367 1369 1371 1373 1375 H1H5797B 1377 1379 1381 1383 1385 1387 1389 1391 H1H5798B 1393 1395 1397 1399 1401 14 03 1405 1407 H1H5799P 1409 1411 1413 1415 1417 1419 1421 1423 H1H5801B 1425 1427 1429 1431 1433 1435 1437 1439 H1H7194B 1441 1443 1445 1447 1633 1635 163 7 1639 H1H7195B 1449 1451 1453 1455 1633 1635 1637 1639 H1H7196B 1457 1459 14 61 14 63 1633 1635 163 7 1639 H1H7198B 14 65 1467 14 69 1471 1633 1635 1637 1639 H1H7203B 1473 1475 1477 1479 1633 1635 1637 1639 H1H7204B 1481 1483 1485 1487 1633 1635 1637 1639 H1H7208B 1489 1491 1493 1495 1633 1635 163 7 1639 H1H7211B 1497 1499 1501 1503 1633 1635 1637 1639

- 59 044194

Н1Н7221В 1505 1507 1509 1511 1633 1635 1637 1639 Н1Н7223В 1513 1515 1517 1519 1633 1635 163 7 1639 Н1Н7226В 1521 1523 1525 1527 1633 1635 1637 1639 Н1Н7232В 1529 1531 1533 1535 1633 1635 163 7 1639 Н1Н7233В 1537 1539 1541 1543 1633 1635 163 7 1639 Н1Н7241В 1545 1547 1549 1551 1633 1635 1637 1639 Н1Н7242В 1553 1555 1557 1559 1633 1635 1637 1639 Н1Н7250В 1561 1563 1565 1567 1633 1635 163 7 1639 Н1Н7251В 1569 1571 1573 1575 1633 1635 1637 1639 Н1Н7254В 1577 1579 1581 1583 1633 1635 163 7 1639 Н1Н7258В 1585 1587 1589 1591 1633 1635 1637 1639 Н1Н7269В 1593 1595 1597 1599 1633 1635 163 7 1639 Н1Н7279В 1601 1603 1605 160 7 1633 1635 1637 1639 HlxH7221G 1609 1611 1613 1615 1633 1635 163 7 1639 H1XH7221G 3 1617 1619 1621 1623 1633 1635 1637 1639 H1XH7221G 5 1625 1627 1629 1631 1633 1635 1637 1639

Antibodies are generally referred to herein according to the following nomenclature: an Fc prefix (e.g., HH, H1M, H2M, etc.) followed by a numerical identifier (e.g., 2712, 2692, etc., as shown in Table 1 followed by the suffix P, N, or B. Thus, according to this nomenclature, the antibody may be referred to herein as, for example, H1H2712N, H1M2692N, H2M2689N, etc. Prefixes H1H, H1M, and H2M in antibody designations as used herein, indicate the specific Fc region of the antibody isotype. For example, antibody H1H has a human IgG1 Fc, antibody H1M has a mouse IgG1 Fc, and antibody H2M has a mouse IgG2 Fc (all variable regions are fully human, as indicated by the first 'H in the designation of antibodies).As one skilled in the art will appreciate, an antibody having a particular Fc isotype can be converted into an antibody with a different Fc isotype (e.g., an antibody with a mouse IgG1 Fc can be converted into an antibody with a human IgG4, etc.) , but in any case, variable domains (including CDRs), which are designated by the numerical identifiers given in table. 1 will remain unchanged and the binding properties are expected to be identical or substantially the same regardless of the nature of the Fc domain.

In table 11 and 12 set forth the amino acid sequence identifiers for the heavy chain variable regions (Table 13) and light chain variable regions (Table 14), and their corresponding CDRs, additional anti-CDS HCVR and LCVR applicable in the anti-STEP2 x anti-CDS bispecific antibodies according to the invention.

Table 11 Amino acid sequences of the heavy chain variable region

SEQ ID NO: Heavy chain ID HCVR HCDR1 HCDR2 HCDR3 CD3-VH-AA 1642 1644 164 6 1648 CD3-VH-B 1658 1660 1662 1664 CD3-VH-C 1674 1676 1678 1680 CD3-VH-D 1690 1692 1694 1696 CD3-VH-E 1706 1708 1710 1712 CD3-VH-F* 1721 1722 1723 1724

- 60 044194

Table 12

Amino acid sequences of the light chain variable region

SEQ ID NO: Light chain ID LCVR LCDR1 LCDR2 LCDR3 CD3-VL-AA 1650 1652 1654 1656 CD3-VL-B 1666 1668 1670 1672 CD3-VL-C 1682 1684 1686 1688 CD3-VL-D 1698 1700 1702 1704 CD3-VL-E 1714 1716 1718 1720 CD3-VL-F# 1725 1726 1727 1728

The CD3-VH-F and CD3-VL-F heavy and light chain variable regions were derived from the anti-CD3 antibody designated L2K as described in WO 2004/106380.

In addition, in table. 13 and 14 set forth sequence identifiers for nucleotide sequences encoding heavy chain variable regions (Table 13) and light chain variable regions (Table 14), and their corresponding CDRs, additional anti-CD3 HCVR and LCVR applicable in anti-STEP2 x anti-CD3 bispecific antibodies according to the invention.

Table 13

Nucleotide sequences encoding heavy chain variable region sequences

SEQ ID NO: Heavy chain ID HCVR HCDR1 HCDR2 HCDR3 CD3-VH-AA 1641 1643 1645 1647 CD3-VH-B 1657 1659 1661 1663 CD3-VH-C 1673 1675 1677 1679 CD3-VH-D 1689 1691 1693 1695 CD3-VH-E 1705 1707 1709 1711

Table 14

Nucleotide sequences encoding light chain variable region sequences

SEQ ID NO: Light chain ID LCVR LCDR1 LCDR2 LCDR3 CD3-VL-AA 1649 1651 1653 1655 CD3-VL-B 1665 1667 1669 1671 CD3-VL-C 1681 1683 1685 1687 CD3-VL-D 1697 1699 1 701 1703 CD3-VL-E 1713 1715 1717 1719

Control constructs used in the following examples.

Various control constructs (anti-CD3 antibodies) were included in the following experiments for comparative purposes: OKT-3, a mouse monoclonal antibody against human T-cell surface antigens, is available from the American Type Culture Collection (ATCC) catalog no. CRL-8001; and SP34, a commercially available murine monoclonal antibody obtained from, for example, Biolegend, San Diego, CA (cat. no. 302914) or BD Pharmagen, cat. 55052, reacts against the epsilon chain of the T3 complex on human T-lymphocyte cells.

Example 10. Production of additional anti-CD3 antibodies.

The following procedures were aimed at identifying antibodies that specifically recognized CD3 (T cell coreceptor) as an antigen.

A pool of anti-CD3 antibodies was obtained from a genetically modified mouse. Briefly, mice were immunized with CD3 antigen and B cells were generated that contained multiple human VH rearrangements to express a diverse repertoire of high-affinity antigen-specific antibodies. Antibodies described in table. 15-18 have the same light chain sequence VK1-39JK5 (LCVR presented in SEQ ID NO: 1890).

The generated antibodies were tested for binding to human CD3 antigen and Javanese

- 61 044 194 th macaque in an in vitro binding assay, and, for example, one anti-CD3 antibody was identified: designated CD3-VH-P (HCVR, presented in SEQ ID NO: 1882), among several others that were found to bind to both human and cynomolgus CD3 having EC ₅₀ binding between 1 and 40 nM (or cell binding titration) as determined by FACS titration of Jurkat cells and cynomolgus T cells, respectively. See, for example, the FACS binding experiments described in Example 12 and PCT/US2016/044732, filed July 29, 2016.

The germline amino acid residues of CD3-VH-P were then identified and the antibody, designated CD3-VH-G, was designed to contain only the germline framework regions. Other antibody derivatives have been developed using well-known molecular cloning techniques to replace amino acid residues in a stepwise manner based on differences between the germline sequence and the CD3-VH-P sequence. Each antibody derivative is designated CD3-VH-G. See table. 15.

While CD3-VH-G and some other engineered antibodies retained their binding affinity as seen in FACS assays, several anti-CD3 antibodies in a bispecific format bound to human or cynomolgus CD3 in vitro with weak or unmeasurable binding affinity, such as higher than EC50 100 nM. Binding affinity, binding kinetics and other biological properties to elucidate toxicological and pharmacokinetic (pK) profiles were subsequently further studied for bispecific antibodies containing exemplary anti-CD3 antibodies and were prepared according to the methods of this example.

Example 11: Heavy and light chain variable regions (amino acid sequences and CDR nucleic acid sequences).

In table 15 shows the amino acid sequence identifiers of the variable regions and heavy chain CDRs of selected anti-CD3 antibodies according to the invention. The corresponding nucleic acid sequence identifiers are given in Table. 16.

Amino acid and nucleic acid sequences were determined for each antibody heavy chain sequence. Each antibody heavy chain derived from the germline sequence (SEQ ID NO: 1910) was assigned a G designation for consistent nomenclature. in table 15 shows the amino acid sequence identifiers of the heavy chain variable regions and CDRs of the constructed anti-CD3 antibodies according to the invention. The corresponding nucleic acid sequence identifiers are given in Table. 16. The light chain variable region amino acid and nucleic acid sequence identifiers and CDRs are also listed below in Table. 17 and 18, respectively.

- 62 044194

Table 15

Heavy chain amino acid sequence identifiers

Antibody Designation CD3-VH SEQ ID NO: HCVR CDR1 CDR2 CDR3 CD3-VH-G 1730 1732 1734 1736 CD3-VH-G2 1738 1740 1742 1744 CD3-VH-G3 1746 1748 1750 1752 CD3-VH-G4 1754 1756 1758 1760 CD3-VH-G5 1762 1764 1766 1768 CD3-VH-G8 1770 1772 1774 1776 CD3-VH-G9 1778 1780 1782 1784 CD3-VH-G10 1786 1788 1790 1792 CD3-VH-G11 1794 1796 1798 1800 CD3-VH-G12 1802 1804 1806 1808 CD3-VH-G13 1810 1812 1814 1816 CD3-VH-G14 1818 1820 1822 1824 CD3-VH-G15 1826 1828 1830 1832 CD3-VH-G16 1834 1836 1838 1840 CD3-VH-G17 1842 1844 1846 1848 CD3-VH-G18 1850 1852 1854 1856 CD3-VH-G19 1858 1860 1862 1864 CD3-VH-G20 1866 1868 1870 1872 CD3-VH-G21 1874 1876 1878 1880 CD3-VH-P 1882 1884 1886 1888

Table 16

Heavy chain nucleic acid sequence identifiers

Antibody Designation CD3-VH SEQ ID NO: HCVR CDR1 CDR2 CDR3 CD3-VH-G 1729 1731 1733 1735 CD3-VH-G2 1737 1739 1741 1743 CD3-VH-G3 1745 1747 1749 1751 CD3-VH-G4 1753 1755 1757 1759 CD3-VH-G5 1761 1763 1765 1767 CD3-VH-G8 1769 1771 1773 1775 CD3-VH-G9 1777 1779 1781 1783 CD3-VH-G10 1785 1787 1789 1,791

- 63 044194

Light chain amino acid sequence identifiers

Antibody Designation SEQ ID NO: LCVR CDR1 CDR2 CDR3 VK1-39JK5 1890 1892 1894 1896

Table 18

Light chain nucleic acid sequence identifiers

Antibody Designation SEQ ID NO: LCVR CDR1 CDR2 CDR3 VK1-39JK5 1889 1891 1893 1895

Antibody control 1, designated CD3-L2K, was constructed based on a known anti-CD3 antibody (ie anti-CD3 antibody L2K, as described in WO 2004/106380).

The isotype control antibody referred to in the examples provided herein is an isotype corresponding (modified IgG4) antibody that reacts with an irrelevant antigen, i.e. FelD1 antigen.

Example 12. In vitro and in vivo studies on human monoclonal anti-CD3 antibodies.

In vivo and in vitro studies of human monoclonal anti-CD3 antibodies were performed as described in US publication 2014/0088295, published March 27, 2014, and PCT/US2016/044732, filed July 29, 2016, which are hereby incorporated by links.

Certain human monoclonal anti-CD3 antibodies of the present invention bind soluble heterodimeric CD3 protein, either in antibody-capture or antigen-capture formats, with high affinity. Soluble heterodimeric CD3 protein (hCD3-epsilon/hCD3-delta; SEQ ID NO: 1900/1901) was prepared with either a human Fc tag (hFcΔAdp/hFc; SEQ ID NO: 1931/1932) or a mouse Fc tag (mFcΔAdp/mfc ; SEQ ID NO: 1933/1934). The heterodimeric CD3 protein was purified using the method described in Davis et al. (US2010/0331527).

Certain human monoclonal anti-CD3 antibodies of the invention bound to human T cells and induced T cell proliferation. Some human monoclonal anti-CD3 antibodies of the invention bound to monkey CD2+CD4+ T cells and induced their proliferation. Several human monoclonal anti-CD3 antibodies supported redirected T cell-mediated killing through Fc/FcR interaction in a calcein-based U937 killing assay. The observed killing, which is believed to depend on the interaction of the Fc antibody with the Fc receptor on U937 cells leading to clustering of CD3 on adjacent T cells, was suppressed by the addition of nonspecific human IgG (data not shown).

Example 13. In vitro studies on human bispecific antibodies STEAP2xCD3.

Titration of FACS binding on Jurkat cells, PC3_STEAP2/1 cells, and cynomolgus T cells. Flow cytometry analysis was used to determine the binding of STEAP2xCD3 bispecific antibodies to chimeric Jurkat cells, PC3_STEAP2/1, chimeric T cells, and cynomolgus T cells, followed by detection with phycoerythrin-labeled anti-human IgG antibody. Briefly, ^2x105 cells/well were incubated for 30 min at 4°C with serial dilutions of STEAP2xCD3 bispecific antibodies or a control antibody (human IgG1 antibody that binds feline antigen without cross-reactivity with human STEAP2 or CD3

- 64 044194 or cynomolgus monkey) 66.6 to 0.001 nM After incubation, cells were washed twice with cold PBS containing 1% filtered FBS (fetal bovine serum), and PE-conjugated secondary anti-human antibody was added to the cells and incubated for an additional 30 minutes . Wells containing no antibody or only secondary antibody were used as controls. After incubation, cells were washed, resuspended in 200 μl of cold PBS containing 1% filtered FBS, and analyzed by flow cytometry on a BD FACS Canto II.

Table 19

FACS binding of selected STEAP2xCD3 bispecific antibodies to Jurkat, PC3_STEAP2/1 and cynomolgus T cells

Designation of bispecific antibody FACS Jurkat EK50 [Μ] FACS PC3_STEAP2/1, EK50 [M] Cynomolgus T cells EC50 [M] BSSTEAP2/CD3-0010 1.36E-08 Binding absent Very weak BSSTEAP2/CD3-004 6.88E-10 7.91E-08 1.99E-09 BSSTEAP2/CD3-0011 8.63E-09 Binding absent Binding absent BSSTEAP2/CD3-005 1.41E-08 3.18E-08 Binding absent BSSTEAP2/CD3-001 7.19E-09 3.44E-09 7.27E-09 BSSTEAP2/CD3-006 3.98E-09 1.22E-08 7.99E-09 BSSTEAP2/CD3-007 6.15E-10 5.37E-09 1.73E-08 BSSTEAP2/CD3-008 1.52E-09 6.88E-08 1.66E-08 BSSTEAP2/CD3-009 4.14E-09 4.21E-08 Binding absent

Jurkat cells are derived from a T lymphoblast cell line that expresses human CD3. All bispecific antibodies tested (Table 19 and Fig. 5) bound to Jurkat cells with EC50 ranging from 1.41E-08 M to 6.15E-10 M. PC3 cells, a human prostate cancer cell line, were modified to express the chimeric construct STEAP2/1. Several bispecific antibodies bound to PC3_STEAP2/1 cells, with EC50 ranging from 7.91E-08 M to 3.44E-09 M (Table 19 and Fig. 6).

The binding of STEAP2xCD3 bispecific antibodies to the surface of purified cynomolgus T cells was also tested. Several bispecific antibodies bound to EC50 ranging from 1.73E-08 M to 7.27E-09 M. Control antibodies did not bind to any cell line. See table. 19 and fig. 7 and 8.

T cell proliferation assay: thawed human PBMCs or freshly isolated cynomolgus PBMCs (50,000 cells/well) were incubated in 3× (human, concentration range: 5E-10M to 2.82E-15M; cynomolgus, concentration range: 1E-15M). 09M to 4.57E-13M) serial dilutions of bispecific STEAP2xCD3 or isotype control in complete medium (RPMI supplemented with 10% FBS, 100 U/ml penicillin, 100 μg/ml streptomycin, 292 μg/ml L-glutamine) and fixed concentration (human : 200 ng/ml, cynomolgus: 500 ng/ml) commercial anti-CD28 antibody (Biolegend, cat. no. 302914) in white 96-well flat-bottom plates for 72 hours at 37°C. Isolated monkey PBMCs were obtained from two donors (identified as mk8781M or mk9381M). After incubation, CellTiter Glo® (Promega, cat. no. 7573) was added and luminescence was measured as a measure of cell viability using a VICTOR X5 multichannel plate reader. The cell titer was calculated by dividing the luminescence of stimulated cells by the initial luminescence of unstimulated cells.

All aSTEAP2xaCD3 bispecific antibodies induce proliferation of human PBMCs in the presence of co-stimulatory anti-CD28 antibody (see Table 20 and FIG. 9). PBMCs were incubated with serial dilutions of bispecific antibodies or control antibody and a fixed concentration of anti-CD28 for 72 h, and cell viability was measured in a luminescence live cell detection assay. Proliferation was determined by comparing the luminescence of cells stimulated with bispecific antibodies with cells without antibodies. EC ₅₀ values (defined as the antibody concentration required to generate half-maximal proliferation) ranged from 3.68E-13M to 1.60E-10M. In contrast, the control antibody showed no activity under the same conditions.

- 65 044194

Table 20

Activation of T cell proliferation induced by selected STEAP2xCD3 bispecific antibodies

Bispecific designation antibodies Proliferation of human PBMCs [M] Proliferation of Javanese MCPC macaque [M] (donor) BSSTEAP2/CD3-0010 6.55E-12 7.034E-13 (mk8781M) BSSTEAP2/CD3-004 7.10E-13 [++] (mk8781M) [-] (mk9381M) BSSTEAP2/CD3-011 8.62E-12 3.597E-12 (mk8781M) BSSTEAP2/ CD3-005 2.76E-12 [-] (mk9381M) (mk8781M) BSSTEAP2/ CD3-001 1.60E-10 4.592E-12 (mk9381M) [+] (mk8781M) BSSTEAP2/ CD3-006 8.14E-13 1.532E-11 (mk9381M) [+] (mk8781M) BSSTEAP2/CD3-007 8.26E-11 [+/-] (mk9381M) [+] (mk8781M) BSSTEAP2/ CD3-008 7.76E-12 [+/-] (mk9381M) [+] (mk8781M) BSSTEAP2/CD3-009 3.68E-13 [-] (mk9381M) (mk8781M)

Bispecific antibodies BSSTEAP2/CD3-0010 and BSSTEAP2/CD3-011 also induced proliferation of cynomolgus monkey PBMCs (with donor mk8781M), demonstrating EC ₅₀ 7E-13 and 3.6E-12, respectively. The activity of BSSTEAP2/CD3-004 was donor dependent. Two additional bispecific antibodies, BSSTEAP2/CD3-001 and BSSTEAP2/CD3-006, induced robust proliferation of cynomolgus monkey PBMCs in all donors tested. The EC ₅₀ values using the mk9381M donor were 4.6E-12M and 1.53E-11M, respectively (see Table 20 and FIG. 10).

The activity of BSSTEAP2/CD3-007 and BSSTEAP2/CD3-008 was donor dependent. In contrast, BSSTEAP2/CD3-005, BSSTEAP2/CD3-009 and isotype control showed no activity.

Cytotoxicity assay targeting C4-2 cells in the presence of anti-STEAP2xCD3 bispecific antibodies and human T cells. To monitor specific killing of STEAP2-containing target cells by flow cytometry, C4-2 cells were labeled with 1 μM fluorescent Violet Cell Tracker dye (Life Technologies kit, no. C34557). After labeling, cells were incubated overnight at 37°C. Separately, human PBMCs were seeded in RPMI medium supplemented with 1 x ¹⁰⁶ cells/ml and incubated overnight at 37°C to enrich lymphocytes by depleting adherent macrophages, dendritic cells and some monocytes. The next day, target cells were co-incubated with naïve adherent cell-depleted PBMCs (4:1 effector/target cell ratio) and serial dilutions of STEAP2xCD3 bispecific antibody or IgG1 control antibody (does not bind to STEAP2) (concentration range: 66.7 nM to 0.25 µM) for 48 hours at 37°C. Cells were recovered from cell culture plates using cell dissociation buffer without enzymes and analyzed by FACS. For FACS analysis, cells were stained with dead/live far red cell tracker (Invitrogen). 5 x ¹⁰⁵ counting beads were added to each well immediately prior to FACS analysis. ^1x105 beads were collected for each sample. To assess killing specificity, cells were gated on Violet-labeled living populations. The percentage of the population alive was recorded and used to calculate normalized survival.

T cell activation was assessed by incubating cells with directly conjugated antibodies to CD2 and CD69, and by analyzing the percentage of activated (CD69+) T cells out of total T cells (CD2+). Several anti-STEAP2xCD3 bispecific antibodies have been tested for their ability to induce naïve T cells to kill target cells expressing human STEAP2 (see Table 21 and FIG. 11). All antibodies tested activated and directed human T cells to deplete C4-2 cells (a subline of human prostate adenocarcinoma derived from LnCap cells). Depletion of target cells was observed only in the presence of bispecific antibodies, with C4-2 cells being depleted in a dose-dependent manner with pM EC50 values. In addition, the observed lysis of target cells was associated with an increase in the number of CD69 cells on CD2+ T cells with pM EC50 values (see Table 21 and Fig. 12).

-

Claims (31)

Table 21 Cytotoxicity and T cell activation properties of selected STEAP2xCD3 bispecific antibodies Bispecific antibody Designation Depletion of C4 cells 2, EK50 [M] T cell activation, EC50 [M] BSSTEAP2/CD3-004 3.11E-12 9.40E-13 BSSTEAP2/ CD3-005 7.29E-12 +++ BSSTEAP2/CD3-001 4.68E-12 1.61E-12 BSSTEAP2/ CD3-006 3.93E-12 +++ BSSTEAP2/CD3-007 8.11E-12 +++ BSSTEAP2/CD3-008 4.11E-12 +++ BSSTEAP2/CD3-009 2.73E-12 +++ The present invention should not be limited in scope to the specific embodiments described herein. Indeed, various modifications of the invention in addition to those described herein will become apparent to those skilled in the art from the foregoing description. Such modifications are intended to be covered by the appended claims. CLAIM 1. An antibody-drug conjugate (ADC) comprising an antibody to six-transmembrane prostate epithelial antigen 2 (anti-STEP2 antibody) or an antigen-binding fragment thereof and a cytotoxic agent, wherein said antibody or antigen-binding fragment and said cytotoxic agent are covalently linked through a linker, wherein said antibody or antigen binding fragment comprises complementarity determining regions (CDRs) of a heavy chain variable region containing the amino acid sequence of SEQ ID NO: 218, and a CDR of a light chain variable region containing the amino acid sequence of SEQ ID NO: 226. 2. ADC according to claim 1, characterized in that said anti-STEP2 antibody or antigen-binding fragment contains domains HCDR1-HCDR2-HCDR3-LCDR1-LCDR2-LCDR3, respectively, containing the amino acid sequences SEQ ID NO: 220-222-224-228 -230-232. 3. The ADC according to claim 1, wherein said anti-STEP2 antibody or antigen binding fragment comprises a heavy chain variable region (HCVR) containing the amino acid sequence of SEQ ID NO: 218, and a light chain variable region (LCVR) containing the amino acid sequence SEQ ID NO: 226. 4. An ADC according to any one of claims 1 to 3, which is internalized by cells expressing human STEAP2. 5. ADC according to any one of claims 1-3, characterized in that said antibody is completely human. 6. ADC according to any one of claims 1 to 5, characterized in that the cytotoxic agent is selected from auristatin, maytansinoid, tubulisin, tomaimycin derivative or dolastatin derivative. 7. ADC according to any one of claims 1 to 5, characterized in that the cytotoxic agent is an auristatin selected from MMAE or MMAF, or a maytansinoid selected from DM1 or DM4. 8. ADC according to any one of claims 1 to 5, where the cytotoxic agent is a maytansinoid having the structure: o CH ₃ (Formula I), where A represents arylene or heteroarylene, or - 67 044194 (Formula II) where A _3a represents an amino acid, a peptide having 2-20 amino acids, alkyl, alkynyl, alkenyl, cycloalkyl, aryl, heteroaryl, heterocyclyl, -CR ₅ R ₆ -, -Ο-, -C(=O) -, -О-С(=О)-, -С(=О)-О-, -О-С(=О)-О-, -С(=О)-(СН _x ) _р1 -, -С (=O)-O-(CH _x ) _p1 -, -(CH _x ) _p1 -C(=O)-, -(CH _x ) _p1 -C(=O)-O-, -(O-(CH ₂ ) _р2 -) _р3 -,-((СН ₂ )р ₂ -О-) _р3 -, -C(=S)-, -C(=S)-S-, -C(=S)-NH- , -SC(=S)-, -SC(=S)-S-, -S-, -SO-, -SO ₂ -, -NR4-, -N(R4)-C(=O)-N( R ₈ )-, -N(R4)-C(=O)O-, -N(R4)-C(=O)-, -C(=O)-N(R4)-, -C(=O )-N(R4)-C(=O)- or -OC(=O)-NR4-, where alkyl, alkynyl, alkenyl, cycloalkyl, aryl, heteroaryl and heterocyclyl are optionally substituted; and pl, p2 and p3 are each independently 0 or an integer from 1 to 100; x represents 0, 1 or 2; R4, _R5 , _R6 and _R8 are each independently H or substituted or unsubstituted: alkyl, alkenyl, alkynyl, aryl, heteroaryl or heterocyclyl; And R4 _a is substituted or unsubstituted: alkyl, alkenyl, alkynyl, aryl, heteroaryl or heterocyclyl. 9. ADC according to claim 8, characterized in that the maytansinoid is: 10. ADC according to claim 8, characterized in that the maytansinoid is: I. ADC according to claim 8, characterized in that the maytansinoid is: 12. An ADC according to any one of claims 1 to 5, comprising an anti-8TER2 antibody or an antigen-binding fragment thereof, and where r represents a bond to an anti-8TER2 antibody or an antigen-binding fragment thereof. 13. ADC according to any one of claims 1 to 5, containing an anti-8TER2 antibody or fragment thereof, and -68044194 where “ represents the binding to the anti-STEP2 antibody or its volume. antigen-binding fragment 14. ADC according to any one of claims 1 to 5, containing aHTu-STEP2 antibody or fragments thereof, and where “ represents a link to an anti-STEP2 antibody or an antigen-binding fragment thereof. 15. ADC according to any one of claims 12-14, characterized in that the connection with the antibody or its antigen-binding fragment is carried out through the sulfur component of the cysteine residue. 16. ADC according to any one of claims 1-5, containing an anti-STEP2 fragment, and an antibody or antigen-binding antibody thereof where represents a link to an anti-STEP2 antibody or fragment thereof. 17. ADC according to claim 16, where the connection with the antibody or its antigen-binding fragment is carried out through the nitrogen component of the lysine residue. 18. ADC according to any one of claims 1 to 17, characterized in that the ADC contains from 1 to 4 cytotoxic agents per anti-STEP2 antibody or antigen-binding fragment thereof. 19. A pharmaceutical composition containing an antibody-drug conjugate according to any one of claims 1 to 18 and a pharmaceutically acceptable carrier or diluent, for the treatment of cancer associated with the expression of STEAP-2, selected from the group consisting of prostate cancer, bladder cancer , cervical cancer, lung cancer, colon cancer, kidney cancer, breast cancer, pancreatic cancer, stomach cancer, uterine cancer and ovarian cancer. 20. The pharmaceutical composition according to claim 19, containing an ADC mixture, where the ADC mixture contains: An ADC according to any one of claims 1 to 5, comprising an anti-STEP2 antibody or an antigen-binding fragment thereof and - 69 044194 An ADC according to any one of claims 1 to 5, comprising an aHmu-STEP2 antibody or an antigen-binding fragment thereof and where * represents a linkage to an anti-STEP2 antibody or fragment thereof. 21. Pharmaceutical composition according to claim 19 or 20, characterized in that the cancer is prostate cancer. 22. An isolated antibody or antigen-binding fragment thereof that binds human six-transmembrane epithelial antigen of the prostate 2 (STEAP2), wherein said antibody or antigen-binding fragment comprises a complementarity determining region (CDR) of a heavy chain variable region containing the amino acid sequence of SEQ ID NO: 218, and CDR of the light chain variable region containing the amino acid sequence SEQ ID NO: 22 6. 23. An antibody or an antigen-binding fragment thereof according to claim 22, characterized in that said antibody or antigen-binding fragment contains the domains HCDR1-HCDR2-HCDR3-LCDR1-LCDR2-LCDR3, respectively, containing the amino acid sequences SEQ ID NO: 220-222-224- 228-230-232. 24. The antibody or antigen-binding fragment according to claim 22, characterized in that said antibody or antigen-binding fragment comprises a heavy chain variable region (HCVR) containing the amino acid sequence of SEQ ID NO: 218, and a light chain variable region (LCVR) containing the amino acid sequence SEQ ID NO: 226. 25. An antibody or antigen-binding fragment thereof according to any one of claims 22 to 24, which is internalized by cells expressing human STEAP2. 26. An antibody or antigen-binding fragment thereof according to any one of claims 22-25, characterized in that said antibody or antigen-binding fragment thereof is entirely human. 27. A bispecific antibody or antigen-binding fragment thereof that binds human six-transmembrane epithelial antigen 2 (STEAP2) and human CD3, wherein said antibody or antigen-binding fragment comprises an anti-STEP2 antigen-binding domain containing complementarity determining regions (CDRs) of a heavy chain variable region, containing the amino acid sequence of SEQ ID NO: 218, and a light chain variable region CDR containing the amino acid sequence of SEQ ID NO: 226. 28. A bispecific antibody or an antigen-binding fragment thereof according to claim 27, characterized in that the anti-STEP2 antigen-binding domain contains the domains HCDR1-HCDR2-HCDR3-LCDR1-LCDR2-LCDR3, respectively, containing the amino acid sequences SEQ ID NO: 220-222-224 -228-230-232. 29. The bispecific antibody or antigen-binding fragment thereof according to claim 28, wherein the anti-STEP2 antigen-binding domain comprises a heavy chain variable region (HCVR) containing the amino acid sequence of SEQ ID NO: 218, and a light chain variable region (LCVR) containing amino acid sequence SEQ ID NO: 226. 30. A pharmaceutical composition containing an antibody or an antigen-binding fragment thereof according to any one of claims 22-26 or a bispecific antibody or an antigen-binding fragment thereof according to any one of claims 27-29 and a pharmaceutically acceptable carrier or diluent, for the treatment of cancer associated with the expression of STEAP- 2, wherein the cancer associated with STEAP-2 expression is selected from the group consisting of prostate cancer, bladder cancer, cervical cancer, lung cancer, colon cancer, kidney cancer, breast cancer, pancreatic cancer, stomach cancer , uterine cancer and ovarian cancer. 31. The pharmaceutical composition according to claim 30, characterized in that the cancer is prostate cancer. 32. Use of an ADC according to any one of claims 1 to 18 for the manufacture of a medicinal product intended -