Bispecific Anti-PSMA×Anti-CD28 antibodies and uses thereof

Abstract

To provide an anti-CD28 antibody that is safe for use in a pharmaceutical composition and useful in therapeutic settings in which specific targeting and T cell-mediated killing of cells that express the target antigen is desired.SOLUTION: The present invention provides bispecific antigen-binding molecules comprising a first antigen-binding domain that specifically binds human CD28, and a second antigen-binding molecule that specifically binds human PSMA. In certain embodiments, the bispecific antigen- binding molecules of the present invention are capable of inhibiting the growth of tumors expressing PSMA, such as prostate tumors. The antibodies and bispecific antigen-binding molecules of the invention are useful for the treatment of diseases and disorders in which an up-regulated or induced targeted immune response is desired and / or therapeutically beneficial.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] Related Applications This application claims priority to U.S. Provisional Patent Application No. 62 / 688,227, filed June 21, 2018, U.S. Provisional Patent Application No. 62 / 781,930, filed December 19, 2018, U.S. Provisional Patent Application No. 62 / 781,980, filed December 19, 2018, and U.S. Provisional Patent Application No. 62 / 815,878, filed March 8, 2019, the entire contents of which are expressly incorporated herein by reference.

[0002] Sequence Listing This application contains a Sequence Listing that has been submitted electronically in ASCII format and is incorporated herein by reference in its entirety. The above ASCII copy, created on June 20, 2019, is named 10367WO01_118003-45220_SL.TXT and is 48,690 bytes in size.

[0003] The present invention also relates to bispecific antigen-binding molecules that bind to target molecules such as CD28 and PSMA, and methods of their use. [Background technology]

[0004] CD28 is a type I transmembrane protein expressed on the surface of T cells, which has a single extracellular IgV-like domain assembled as a homodimer. CD28 is a receptor for the CD80 (B7.1) and CD86 (B7.2) proteins and is activated by CD80 or CD86 expressed on antigen-presenting cells (APCs). Binding of CD28 to CD80 or CD86 provides a costimulatory signal important for T cell activation and survival. T cell stimulation via CD28 in addition to the T cell receptor (TCR) provides a potent signal for the production of various interleukins. After TCR activation, CD28 enhances cellular signals, such as pathways controlled by the NFκB transcription factor. CD28 costimulation is important for effective T cell activation, including T cell differentiation, proliferation, cytokine release, and cell death.

[0005] Anti-CD28 antibodies have been proposed for therapeutic purposes, including T cell activation. One specific anti-CD28 antibody, TGN1412 (an anti-CD28 superagonist), was used in clinical trials. TGN1412 induced a cytokine storm, which was not predicted in toxicology or ex vivo human PBMC studies. In 2006, six healthy volunteers were intravenously administered TGN1412 (an anti-CD28 superagonist) at a dose of 0.1 mg / kg. Within two hours, all six patients experienced a significant inflammatory response (cytokine storm). Within 16 hours, all patients had multiple organ failure. When the subjects were treated with corticosteroids, cytokine levels returned to normal within two to three days. The starting dose of 0.1 mg / kg in a phase 1 study (related to CRS) was based on 500 times the no-observed-adverse-effect level (NOAEL) of 50 mg / kg in cynomolgus monkeys (Non-Patent Document 1).

[0006] Toxicology studies in cynomolgus monkeys failed to predict the cytokine responses observed in humans.

[0007] PSMA (prostate-specific membrane antigen) / FOLH1 is a well-characterized tumor target. PSMA is a type II transmembrane glycoprotein that is overexpressed in prostate cancer. It is also known as glutamate carboxypeptidase II (GPC). In the normal human prostate, PSMA is associated with the cytoplasm and apical side of the epithelium surrounding the prostatic ducts. Dysplasia and / or neoplastic transformation of prostate tissue results in the translocation of PSMA from the apical membrane to the luminal surface of the ducts. PSMA is constitutively endocytosed and not shed. PSMA is the target of various clinical ADC (antibody-drug conjugate) studies and imaging approaches. PSMA is highly expressed in human prostate adenocarcinoma and coincides with metastases (lymph nodes). In prostate tumors, PSMA expression levels increase depending on the stage and grade of the disease. Transition to androgen-independent prostate cancer ultimately leads to increased expression. Interestingly, PSMA expression has also been reported in the tumor neovasculature of some solid tumors, including subtypes of colon, lung, breast, renal, and bladder cancers.

[0008] PSMA is also expressed in normal tissues. The strongest expression is seen in prostate epithelial cells, duodenum, renal tubular cells, salivary gland, and stellate cells. PSMA is weakly expressed in the fallopian tube and breast, and rarely in the endothelium of the cervix.

[0009] Bispecific antigen-binding molecules that bind to both CD28 and a target antigen (such as PSMA) are useful in therapeutic settings where specific targeting and T cell-mediated killing of cells expressing the target antigen is desired. There is also a need for anti-CD28 antibodies that are safe for use in pharmaceutical compositions. [Prior art documents] [Non-patent literature]

[0010] [Non-Patent Document 1] Suntharalingam, et al., Cytokine Storm in a Phase 1 Trial of the Anti-CD28 Monoclonal Antibody TGN1412, NEJM 355:1018-1028(2006) Summary of the Invention

[0011] In a first aspect, the present invention provides a bispecific antigen-binding molecule that binds to CD28 and a target antigen. According to certain exemplary embodiments, the bispecific antigen-binding molecule binds to CD28 and PSMA, and such bispecific antigen-binding molecules are also referred to herein as "anti-CD28 / anti-PSMA bispecific molecules." The anti-PSMA portion of the anti-CD28 / anti-PSMA bispecific molecule is useful for targeting tumor cells that express PSMA (e.g., prostate tumor cells), and the anti-CD28 portion of the bispecific molecule is useful for activating T cells. The simultaneous binding of PSMA on tumor cells and CD28 on T cells induces directional killing (cytolysis) of the target tumor cells by activated T cells. Therefore, the anti-CD28 / anti-PSMA bispecific molecules of the present invention are useful, inter alia, for treating diseases and disorders associated with or caused by PSMA-expressing tumors (e.g., prostate cancer).

[0012] The bispecific antigen-binding molecule according to this aspect of the present invention comprises a first antigen-binding domain that specifically binds to human CD28 and a second antigen-binding domain that specifically binds to PSMA. The present invention includes anti-CD28 / anti-PSMA bispecific molecules (e.g., bispecific antibodies) in which each antigen-binding domain comprises a heavy chain variable region (HCVR) paired with a light chain variable region (LCVR). In certain exemplary embodiments of the present invention, the anti-CD28 antigen-binding domain and the anti-PSMA antigen-binding domain each comprise a distinct and separate HCVR paired with a common LCVR.

[0013] The present invention provides anti-CD28 / anti-PSMA bispecific molecules, wherein the first antigen-binding domain that specifically binds to CD28 comprises any one of the HCVR amino acid sequences set forth in Table 1. or any of the heavy chain CDR1-CDR2-CDR3 amino acid sequences listed in Table 1. The first antigen-binding domain that specifically binds to CD28 may also comprise any of the LCVR amino acid sequences listed in Table 1. According to certain embodiments, the first antigen-binding domain that specifically binds to CD28 comprises any of the HCVR / LCVR amino acid sequence pairs listed in Table 1. The present invention also provides anti-CD28 / anti-PSMA bispecific molecules, wherein the first antigen-binding domain that specifically binds to CD28 comprises any of the heavy chain CDR1-CDR2-CDR3 amino acid sequences listed in Table 1 and / or any of the light chain CDR1-CDR2-CDR3 amino acid sequences listed in Table 1.

[0014] According to certain embodiments, the present invention provides an anti-CD28 / anti-PSMA bispecific molecule, wherein the first antigen-binding domain that specifically binds to CD28 comprises a heavy chain variable region (HCVR) having an amino acid sequence selected from the group consisting of SEQ ID NOs: 10, 26, and 58, or a substantially similar sequence thereof having at least 90%, at least 95%, at least 98%, or at least 99% sequence identity.

[0015] The present invention also provides an anti-CD28 / anti-PSMA bispecific molecule, wherein the first antigen-binding domain that specifically binds to CD28 comprises a light chain variable region (LCVR) having an amino acid sequence selected from the group consisting of SEQ ID NOs: 18, 42, and 66, or a substantially similar sequence thereof having at least 90%, at least 95%, at least 98%, or at least 99% sequence identity.

[0016] The present invention also provides anti-CD28 / anti-PSMA bispecific molecules, wherein a first antigen-binding domain that specifically binds to CD28 comprises a HCVR and LCVR (HCVR / LCVR) amino acid sequence pair selected from the group consisting of SEQ ID NOs: 10 / 18, 26 / 42, and 58 / 66.

[0017] The present invention also provides an anti-CD28 / anti-PSMA bispecific molecule, wherein the first antigen-binding domain that specifically binds to CD28 comprises a heavy chain CDR3 (HCDR3) domain having an amino acid sequence selected from the group consisting of SEQ ID NOs: 16, 32, and 64, or a substantially similar sequence thereof with at least 90%, at least 95%, at least 98%, or at least 99% sequence identity, and a light chain CDR3 (LCDR3) domain having an amino acid sequence selected from the group consisting of SEQ ID NOs: 24, 48, and 72, or at least and substantially similar sequences having 90%, at least 95%, at least 98%, or at least 99% sequence identity.

[0018] In certain embodiments, the first antigen-binding domain that specifically binds to CD28 comprises an HCDR3 / LCDR3 amino acid sequence pair selected from the group consisting of SEQ ID NOs: 16 / 24, 32 / 48, and 64 / 72.

[0019] The present invention also provides an anti-CD28 / anti-PSMA bispecific antigen-binding molecule, wherein the first antigen-binding domain that specifically binds to CD28 comprises a heavy chain CDR1 (HCDR1) having an amino acid sequence selected from the group consisting of SEQ ID NOs: 12, 28, and 60, or a substantially similar sequence thereof 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 NOs: 14, 30, and 62, or a substantially similar sequence thereof having at least 90%, at least 95%, at least 98%, or at least 99% sequence identity. a light chain CDR1 (LCDR1) domain having an amino acid sequence selected from the group consisting of SEQ ID NOs: 20, 44, and 68, or a substantially similar sequence thereof with at least 90%, at least 95%, at least 98%, or at least 99% sequence identity; and a light chain CDR2 (LCDR2) domain having an amino acid sequence selected from the group consisting of SEQ ID NOs: 22, 46, and 70, or a substantially similar sequence thereof with at least 90%, at least 95%, at least 98%, or at least 99% sequence identity.

[0020] Certain non-limiting exemplary anti-CD28 / anti-PSMA bispecific antigen binding molecules of the present invention comprise a first antigen binding domain that specifically binds to CD28, comprising HCDR1-HCDR2-HCDR3-LCDR1-LCDR2-LCDR3 domains having amino acid sequences selected from the group consisting of SEQ ID NOs: 12-14-16-20-22-24, 28-30-32-44-46-48, and 60-62-64-68-70-72, respectively.

[0021] The present invention also provides anti-CD28 / anti-PSMA bispecific molecules, wherein the second antigen-binding domain that specifically binds to PSMA comprises a heavy chain variable region (HCVR) having an amino acid sequence selected from the group consisting of SEQ ID NOs: 2, 34, and 50, or a substantially similar sequence thereof having at least 90%, at least 95%, at least 98%, or at least 99% sequence identity.

[0022] The present invention also provides anti-CD28 / anti-PSMA bispecific molecules, wherein the second antigen-binding domain that specifically binds to PSMA comprises a light chain variable region (LCVR) having an amino acid sequence selected from the group consisting of SEQ ID NOs: 18, 42, and 66, or a substantially similar sequence thereof having at least 90%, at least 95%, at least 98%, or at least 99% sequence identity.

[0023] The present invention also provides anti-CD28 / anti-PSMA bispecific molecules, wherein the second antigen-binding domain that specifically binds to PSMA comprises a HCVR and LCVR (HCVR / LCVR) amino acid sequence pair selected from the group consisting of SEQ ID NOs: 2 / 18, 34 / 42, and 50 / 66.

[0024] The present invention also provides an anti-CD28 / anti-PSMA bispecific molecule, wherein the second antigen-binding domain that specifically binds to PSMA comprises a heavy chain CDR3 (HCDR3) domain having an amino acid sequence selected from the group consisting of SEQ ID NOs: 8, 40, and 56, or a substantially similar sequence thereof with at least 90%, at least 95%, at least 98%, or at least 99% sequence identity, and a light chain CDR3 (LCDR3) domain having an amino acid sequence selected from the group consisting of SEQ ID NOs: 24, 48, and 72, or a substantially similar sequence thereof with at least 90%, at least 95%, at least 98%, or at least 99% sequence identity.

[0025] In certain embodiments, the second antigen-binding domain that specifically binds to PSMA comprises an HCDR3 / LCDR3 amino acid sequence pair selected from the group consisting of SEQ ID NOs: 8 / 24, 40 / 48, and 56 / 72.

[0026] The present invention also provides an anti-CD28 / anti-PSMA bispecific antigen binding molecule, wherein the second antigen binding domain that specifically binds to PSMA is a heavy chain CDR1 (HCDR1) having an amino acid sequence selected from the group consisting of SEQ ID NOs: 4, 36, and 52, or a substantially similar sequence thereof 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 NOs: 6, 38, and 54, or a substantially similar sequence thereof having at least 90%, at least 95%, at least 98%, or at least 99% sequence identity. a light chain CDR1 (LCDR1) domain having an amino acid sequence selected from the group consisting of SEQ ID NOs: 20, 44, and 68, or a substantially similar sequence thereof with at least 90%, at least 95%, at least 98%, or at least 99% sequence identity; and a light chain CDR2 (LCDR2) domain having an amino acid sequence selected from the group consisting of SEQ ID NOs: 22, 46, and 70, or a substantially similar sequence thereof with at least 90%, at least 95%, at least 98%, or at least 99% sequence identity.

[0027] Certain non-limiting exemplary anti-CD28 / anti-PSMA bispecific antigen binding molecules comprise a second antigen binding domain that specifically binds to PSMA, comprising HCDR1-HCDR2-HCDR3-LCDR1-LCDR2-LCDR3 domains having amino acid sequences selected from the group consisting of SEQ ID NOs: 4-6-8-20-22-24, 36-38-40-44-46-48, and 52-54-56-68-70-72, respectively.

[0028] In a related embodiment, the invention includes an anti-CD28 / anti-PSMA bispecific antigen binding molecule, wherein the second antigen-binding domain that specifically binds PSMA comprises heavy and light chain CDR domains contained within heavy and light chain variable region (HCVR / LCVR) sequences selected from the group consisting of SEQ ID NOs: 2 / 18, 34 / 42, and 50 / 66.

[0029] In another aspect, the present invention provides nucleic acid molecules encoding any of the HCVR, LCVR, or CDR sequences of the anti-CD28 / anti-PSMA bispecific antigen-binding molecules disclosed herein, including nucleic acid molecules comprising a polynucleotide sequence set forth in Table 1 herein, and nucleic acid molecules comprising two or more of the polynucleotide sequences set forth in Table 1, in any functional combination or arrangement thereof. Recombinant expression vectors harboring nucleic acids of the invention, and host cells into which such vectors have been introduced, are also encompassed by the present invention, as are methods of producing the antibodies by culturing the host cells under conditions that allow for antibody production, and recovering the antibodies produced.

[0030] The present invention includes anti-CD28 / anti-PSMA bispecific antigen binding molecules in which any of the aforementioned antigen binding domains that specifically bind CD28 is combined, linked, or otherwise associated with any of the aforementioned antigen binding domains that specifically bind PSMA to form a bispecific antigen binding molecule that binds CD28 and PSMA.

[0031] The present invention includes anti-CD28 / anti-PSMA bispecific antigen binding molecules with modified glycosylation patterns. In some applications, modifications to remove undesired glycosylation sites or antibodies lacking fucose moieties present on the oligosaccharide chains, for example, to enhance antibody-dependent cellular cytotoxicity (ADCC) function, may be useful (Shield (See, e.g., JBC 277:26733, J. et al. (2002) JBC 277:26733). In other applications, galactosylation modifications can be made to modify complement dependent cytotoxicity (CDC).

[0032] In another aspect, the invention provides a pharmaceutical composition comprising an anti-CD28 / anti-PSMA bispecific antigen-binding molecule disclosed herein and a pharmaceutically acceptable carrier. In a related aspect, the invention features a composition that is a combination of an anti-CD28 / anti-PSMA bispecific antigen-binding molecule and a second therapeutic agent. In one embodiment, the second therapeutic agent is any agent that can be advantageously combined with the anti-CD28 / anti-PSMA bispecific antigen-binding molecule. Exemplary agents that can be advantageously combined with the anti-CD28 / anti-PSMA bispecific antigen-binding molecule are discussed in detail elsewhere herein.

[0033] In yet another aspect, the present invention provides therapeutic methods for targeting / killing tumor cells expressing PSMA using the anti-CD28 / anti-PSMA bispecific antigen binding molecules of the present invention, and these therapeutic methods include the anti-CD28 / anti-PSMA bispecific antigen binding molecules of the present invention. It involves administering a therapeutically effective amount of the pharmaceutical composition to a subject in need thereof.

[0034] The present invention also includes the use of an anti-CD28 / anti-PSMA 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 PSMA-expressing cells.

[0035] In yet another aspect, the invention provides a therapeutic method for targeting / killing tumor cells expressing PSMA using the anti-CD28 / anti-PSMA bispecific antigen binding molecule of the invention, wherein the anti-CD28 / anti-PSMA bispecific antigen binding molecule is combined with another anti-tumor bispecific antigen binding molecule that binds CD3 (e.g., anti-CD28 / anti-PSMA combined with an anti-CD3 / anti-PSMA antibody).

[0036] In yet another aspect, the invention provides a therapeutic method for targeting / killing tumor cells expressing PSMA using the anti-CD28 / anti-PSMA bispecific antigen binding molecule of the invention, wherein the anti-CD28 / anti-PSMA bispecific antigen binding molecule is combined with a checkpoint inhibitor targeting PD-1 or CTLA-4 (e.g., anti-CD28 / anti-PSMA is combined with an anti-PD-1 antibody).

[0037] In yet another aspect, the invention provides a therapeutic method for targeting / killing tumor cells expressing PSMA using the anti-CD28 / anti-PSMA bispecific antigen binding molecule of the invention, in combination with other anti-tumor bispecific antigen binding molecules that bind CD3 (e.g., anti-CD28 / anti-PSMA combined with an anti-CD3 / anti-PSMA bispecific antibody) and with checkpoint inhibitors that target PD-1 or CTLA-4 (e.g., anti-CD28 / anti-PSMA combined with an anti-PD-1 antibody).

[0038] Other embodiments will become apparent from review of the following detailed description. [Brief explanation of the drawings]

[0039] [Figure 1] FIG. 1 is a schematic diagram showing the structure of an exemplary anti-CD28 / PSMA bispecific antibody. [Figure 2] Figure 1 is a graph showing tumor growth inhibition in engineered cell lines transduced with costimulatory ligand expression. Three tumor cell lines, B16F10.9, EL4, and MC38, were engineered to express costimulatory ligands or GFP as a control. Figure 2 is a graph showing percent survival. Mice were euthanized when tumors grew beyond 2000 mm3. [Figure 3] FIG. 1 is a schematic diagram showing primary and engineered bioassays for testing anti-PSMA×CD28 bispecific antibodies. [Figure 4A]Figure 1 shows that activation of CD4+ T cells and engineered JRT3.T3 / 1G4 / hCD28 cells was enhanced by anti-hPSMAxhCD28 in the presence of a primary stimulus (REGN2281) and PSMA expressed on HEK293 / hCD20 cells. [Figure 4B] Figure 1 shows that activation of CD4+ T cells and engineered JRT3.T3 / 1G4 / hCD28 cells was enhanced by anti-hPSMAxhCD28 in the presence of a primary stimulus (REGN2281) and PSMA expressed on HEK293 / hCD20 cells. [Figure 5A] Figures 5A-5H show that in PSMA-transfected engineered cell lines, the PSMAxCD28 bispecific antibody enhances T cell activation in the presence of TCR stimulation with an anti-CD20xCD3 bispecific antibody. Figure 5A is a schematic diagram showing the mode of action of the bispecific antibody. [Figure 5B] Figure 5B is a graph showing quantification of the ratio of CD28 fluorescence entering and exiting the immune synapse. PSMA-engineered target cells (HEK293) and human Jurkat T cells were cocultured with fluorescently labeled bispecific antibodies (anti-PSMA x CD28, anti-CD20 x CD3) for 1 hour at 37°C, gently fixed, and stained with anti-CD28. The number of cells analyzed per group is shown in the bar graph. [Figure 5C] Figures 5C and 5D show the proliferation of human T cells cultured with PSMA-engineered target cells. Data shown are mean ± SEM. Data represent at least two experiments. Figure 5C provides a graph showing dose titration of CD20xCD3 in the presence of 0.5 nM hIgG4s isotype control or PSMAxCD28. [Figure 5D] FIG. 5D provides a graph showing dose titration of PSMA×CD28 or the indicated control in the presence of 5 pM hIgG4s isotype control or CD20×CD3. [Figure 5E]Figures 5E-5G are graphs showing proliferation and cytokine release of human T cells cultured with TAA (PSMA or CD20)-engineered target cells as indicated at the top of the panels in the presence of 5 pM hIgG4s isotype control (lower panel of each figure) or CD20xCD3 (upper panel of each figure). Data are mean ± SEM. Data represent at least three experiments. Figure 5E is a graph showing proliferation. [Figure 5F] FIG. 5F is a graph showing IL-2 release. [Figure 5G] FIG. 5G is a graph showing IFNγ release. [Figure 5H] Figure 5H is a graph showing that anti-PSMAxCD28 and anti-PSMAxCD3 bispecific antibodies can simultaneously bind to PSMA-expressing cells. 22RV1 cells were preincubated for 30 minutes at 4°C in flow cytometry buffer (PBS + 1% FBS) containing 20 mg / ml of PSMAxCD3 or 20 mg / ml of an anti-PSMA antibody containing the same anti-PSMA arm as the PSMAxCD28 bispecific antibody. After incubation, the cells were washed with flow cytometry buffer and incubated with 5 mg / ml of PSMAxCD28 directly labeled with Alexa647 for 20 minutes at 4°C. After incubation, the cells were washed, resuspended in flow cytometry buffer, and analyzed by flow cytometry. [Figure 6A] 1 shows that anti-PSMA×CD28 enhanced the cytotoxic efficacy of anti-PSMA×CD3 in the presence of prostate cancer and human or cynomolgus monkey T cells. [Figure 6B] 1 shows that anti-PSMA×CD28 enhanced the cytotoxic efficacy of anti-PSMA×CD3 in the presence of prostate cancer and human or cynomolgus monkey T cells. [Figure 7A]Figures 7A-7H provide graphs showing that in cancer cell lines with endogenous PSMA, the PSMAxCD28 bispecific antibody enhances T cell activation in the presence of TCR stimulation with the PSMAxCD3 bispecific antibody. For Figures 7A-7D, human T cells were cultured for 96 hours with cancer target cells (prostate cancer line C4-2) with endogenous PSMA expression and the indicated bispecific antibody. Figure 7A is a graph showing tumor cell killing. Data shown are the percentage of viable cells. [Figure 7B] FIG. 7B is a graph showing IFNγ release. [Figure 7C] FIG. 7C provides a graph showing the number of CD4 T cells and the frequency of CD25+ cells as a percentage of CD4 T cells. [Figure 7D] FIG. 7D provides a graph showing the number of CD8 T cells and the frequency of CD25+ cells as a percentage of CD8 T cells. [Figure 7E] For Figures 7E-7G, cynomolgus monkey T cells were cultured for 96 hours with cancer cells (prostate cancer line C4-2) with endogenous PSMA expression and the indicated bispecific antibodies. Figure 7E is a graph showing tumor cell killing. Data shown are the percentage of viable cells. [Figure 7F] FIG. 7F provides a graph showing the number of CD4 T cells and the frequency of CD25+ cells as a percentage of CD4 T cells. [Figure 7G] FIG. 7G provides a graph showing the number of CD8 T cells and the frequency of CD25+ cells as a percentage of CD8 T cells. [Figure 7H] FIG. 7H provides a graph showing antibody binding to cellular targets as measured by flow cytometry. [Figure 8] 1 shows a comparison of the properties of TGN superagonists with the properties of the anti-CD28 antibodies of the present invention. [Figure 9A]Figure 9 provides graphs and plots showing that PSMAxCD28 bispecific antibodies enhance antitumor immunity and induced T cell activation by PSMAxCD3. MC38 / hPSMA tumor cells were implanted subcutaneously into hCD3 / hCD28 / hPSMA-humanized mice. Mice were treated with 5 mg / kg of the indicated bispecific antibodies on days 0, 3, and 7. Figure 9A shows tumor volume over time. Values represent the mean ± SEM and are representative of three experiments with 3–7 mice per group. P values were calculated by two-way ANOVA (*, p<0.05; **, p<0.01; ***, p<0.001; and ****, p<0.0001). [Figure 9B] Figure 9B provides graphs and plots showing that PSMAxCD28 bispecific antibodies enhance antitumor immunity and induced T cell activation by PSMAxCD3. MC38 / hPSMA tumor cells were implanted subcutaneously into hCD3 / hCD28 / hPSMA humanized mice. Mice were treated with 5 mg / kg of the indicated bispecific antibodies on days 0, 3, and 7. Figure 9B provides a graph showing the percentage of cells within each cluster for each treatment group (upper panel), overlaid with a viSNE plot of the indicated clusters (lower panel). [Figure 9C] Figure 1 provides a bar graph showing that PSMAxCD28 bispecific antibody enhances cytokine release induced by PSMAxCD3. Mice were bled for serum cytokines 4 hours after administration on day 0. Statistical significance was calculated by one-way ANOVA compared to isotype **p<0.01 and ***p<0.0001. n=7 mice / group. Data are representative of 3 experiments. [Figure 9D] Figure 1 provides a bar graph showing that PSMAxCD28 bispecific antibody enhances cytokine release induced by PSMAxCD3. Mice were bled for serum cytokines 4 hours after administration on day 0. Statistical significance was calculated by one-way ANOVA compared to isotype **p<0.01 and ***p<0.0001. n=7 mice / group. Data are representative of 3 experiments. [Figure 9E]1 provides graphs and plots showing that PSMAxCD28 bispecific antibody enhances T cell activation induced by PSMAxCD3. [Figure 10] 1 shows that PSMAxCD28 bispecific antibody or parental CD28 bivalent antibody did not induce serum cytokine production in CD3 / CD28 / PSMA humanized mice. [Figure 11] 1 shows that CD28 superagonist treatment, but not anti-CD28 antibody mAb14226P2, promoted cytokine responses at 4 hours in PBMC-transplanted NSG mice. [Figure 12] This demonstrates the potentially safer toxicological profile of the costimulatory bispecific antigen binding molecules of the invention. Anti-CD28 x anti-PSMA did not induce a cytokine response, whereas anti-CD3 x anti-PSMA did. [Figure 13A] Compared with PSMAxCD3 and CD28 superagonists, PSMAxCD28 did not induce cytokine production or T cell margination. LLOQ: "Lower limit of quantification." [Figure 13B] Compared with PSMAxCD3 and CD28 superagonists, PSMAxCD28 did not induce cytokine production or T cell margination. LLOQ: "Lower limit of quantification." [Figure 14]Figures 14 and 17 show that TSAxCD28 and PD-1 blockade activate TCR / CD3 and CD28 on T cells at tumor sites. Figures 15, 16, and 18 show that TSAxCD28 bispecific and PD-1 blockade synergistically promote T cell activation in vitro. Figures 14-18 show the conjugates of T cells (Jurkat / PD-1) and target cells (Raji WT) in the presence of a nonblocking PD-1 mAb (NB PD-1 mAb) or a blocker (PD-1 mAb), and a CD20xCD3 bispecific antibody. Bar graphs showing quantification of PD-1 and CD28 localization at the immune synapse are provided. Statistical significance was calculated by unpaired t-test (not significant, ns). To quantify the localization of PD-1 and CD28 at the immune synapse, we imaged the complexes of T cells (Jurkat / PD-1) and target cells (Raji WT) in the presence of a nonblocking PD-1 mAb (NB PD-1 mAb) or a blocker (PD-1 mAb) and an anti-CD20xCD3 bispecific antibody. The PD-1 mAb was directly labeled with Alexa647, the anti-CD20xCD3 bispecific antibody was directly labeled with Alexa488, and the CD28 mAb was directly labeled with PE. Nuclei were stained with Hoechst 33342 (images not shown). [Figure 15] A bar graph showing quantification of PD-1 and CD28 localization at the immune synapse is presented. Statistical significance was calculated using an unpaired t-test (p<0.0001, ****). To quantify PD-1 and CD28 localization at the immune synapse, images were taken of conjugates of T cells (Jurkat / PD-1) and target cells (Raji / PD-L1) in the presence of a nonblocking PD-1 mAb (NB PD-1 mAb) or a blocker (PD-1 mAb) and an anti-CD20xCD3 bispecific antibody. PD-1 mAb was directly labeled with Alexa647 (shown in red), CD20xCD3 was directly labeled with Alexa488 (shown in green), and CD28 mAb was directly labeled with PE (shown in blue). Nuclei were stained with Hoechst 33342 (shown in gray). Dotted lines represent cell outlines drawn based on brightfield images. [Figure 16] The viability (%) of 22RV1-PDL1 cells at 96 hours is shown. [Figure 17] 1 is a graph showing IFNγ levels in the supernatant at 96 hours. [Figure 18] 1 is a graph showing IL-2 release at 96 hours. [Figure 19A] Figures 19A-19D show that expression of the CD28 ligand (CD86) on tumor cells synergizes with anti-PD1 treatment to induce CD8-dependent antitumor immunity. MC38 tumor cells were transduced with ligands for CD28, CD86 (MC38 / CD86), or an empty vector control (MC38 / EV). WT C57BL6 mice were initially implanted with 1 x 10 tumor cells per mouse and treated with 5 mg / kg of PD-1 mAb or rat isotype control on days 0, 3, 7, 10, and 14 after tumor implantation. Figure 19A shows the mean tumor volume over time. Error bars represent + / - SEM. Statistical significance was determined by two-way ANOVA and Tukey's multiple comparison test. [Figure 19B] Figure 19B shows survival over time (percentage of mice bearing tumors less than 2000 mm3) at 60 days post-implantation. Statistical significance was determined by the log-rank (Mantel-Cox) test. [Figure 19C] In Figure 19C, mice were treated with CD8-depleting antibody (CD8 depleted) or isotype control (no depletion). Mean tumor volume over time with CD8 depletion (dotted line) and without depletion (solid line) is shown + / - SEM. Statistical significance was determined by two-way ANOVA and Tukey's multiple comparison test. [Figure 19D] Figure 19D shows secondary tumor engraftment (rechallenge) of tumor-free mice engrafted with MC38 / CD86 and treated with PD1 mAb. In Figures 19A-19D, the data shown are from a single experiment with 10 mice per group. Data represent at least four separate experiments. Statistical significance is indicated (*p<0.05, **p<0.01, ***p<0.001, and ****p<0.0001). [Figure 20A] Figures 20, 21A, 22, 24, and 25 show that PSMAxCD28 synergizes with PD1 mAb treatment to induce antitumor immunity. MC38 / hPSMA tumor cells were implanted subcutaneously into hCD3 / hCD28 / hPSMA mice. PSMAxCD28 bispecific antibody, PD-1 mAb, or rat IgG2a isotype control were administered as monotherapy or in combination via intraperitoneal injection at 5 mg / kg each. Figures 21B and 21C show that the combination of PSMAxCD28 and PD-1 mAb increases the frequency of tumor-specific T cells. Figures 23A and 23B show that PSMAxCD28 synergizes with anti-PD1 treatment to induce intratumoral cytokines, but not splenic or systemic cytokines. Figures 20A-20E show that immediate treatment with anti-PSMA x anti-CD28 antibodies enhances immunity at the tumor site and synergizes with anti-PD-1 antibodies to promote tumor rejection. Figure 20A shows the mean tumor volume over time. Error bars represent + / - SEM. Statistical significance was determined by two-way ANOVA and Tukey's multiple comparison test (***, p<0.001 and ****, p<0.0001). [Figure 20B] Figure 20B shows survival over time (mice bearing tumors greater than 2000 mm were euthanized). Statistical significance was determined by the log-rank (Mantel-Cox) test (**, p<0.01). [Figure 20C] Figure 20C shows the mean tumor volume 21 days after implantation. Error bars represent + / - SEM. Statistical significance was determined by one-way analysis of variance and Holm-Sidak multiple comparison test (****, p<0.0001). [Figure 20D] FIG. 20D shows a table of tumor volumes (mm 3 ) 21 days after implantation. [Figure 20E] FIG. 20E shows a table of the number of tumor-free mice per group from six separate experiments. [Figure 21A] Figure 1 shows secondary tumor challenge of tumor-free mice previously treated with a combination of PSMAxCD28 and PD1 mAb. Error bars represent + / - SEM. Data are representative of four experiments. [Figure 21B] MC38 / PSMA tumor cells were implanted into CD3 / CD28 / PSMA-humanized mice and treated with 5 mg / kg of isotype control, PSMA x CD28, PD1 mAb, or the combination on days 10 and 14 post-implantation. Spleens were harvested on day 17. Splenocytes were cultured overnight in T cell medium containing 10 mg / ml peptide (p15E or OVA) and 2 mg / ml anti-CD28. After overnight incubation, intracellular cytokine staining was performed according to standard procedures. [Figure 21C] Tumor-free mice from anti-CD28×PSMA and anti-PD1, but not anti-CD28×PSMA and anti-CD3×PSMA treatment, showed a rejection of the second tumor re-challenge. [Figure 22A] Figures 22A-22D, 24, and 25 show the delayed / therapeutic treatment regimen (administration indicated by arrows: days 9, 16, and 22 for Figures 22A and 22B, and days 7, 11, and 14 for Figures 24 and 25). Mean tumor volume over time is shown. Error bars represent + / - SEM. Statistical significance was determined by two-way ANOVA and Tukey's multiple comparison test. Data represent triplicate experiments. [Figure 22B] Ex vivo spleen and intratumor cytokines are shown. Points represent data from individual mice. Bars are mean + / - SEM. [Figure 22C] The mean tumor size on day 20 is shown. [Figure 22D] The survival rates of mice receiving different treatments are shown. [Figure 23A] Figure 23A shows that PSMAxCD28 synergizes with anti-PD1 treatment to induce intratumoral, but not splenic or systemic, cytokines. In Figure 23A, data correspond to Figure 22B. Ex vivo splenic and intratumoral cytokines. Points represent data from individual mice. Bars are mean + / - SEM. [Figure 23B]Figure 23B shows that PSMAxCD28 synergizes with anti-PD1 treatment to induce intratumoral cytokines, but not splenic or systemic cytokines. In Figure 23B, CD3 / CD28 / PSMA triple-humanized mice were implanted with MC38 / hPSMA and treated with the indicated antibodies at 5 mg / kg on day 0. Mice were bled and serum was collected 4 hours post-dose. In both Figures 23A and 23B, statistical significance was calculated by one-way ANOVA and Tukey's multiple comparison test. *p<0.05, **p<0.01, ****p<0.0001 [Figure 24] PD1 expression on T cell subsets from draining lymph nodes (dLN), spleen (Sp), and tumor (Tu) is shown. Points represent data from individual mice. Bars are mean + / - SEM. [Figure 25] The frequencies of CD8 T cells in C1 and C2 from the indicated treatment groups are shown. Data in Figures 22B, 24, and 25 are from one experiment each. Figures 22B and 24, n = 4-6 mice / group. Figure 25, n = 10 mice / group. [Figure 26A] Figures 26A-26C, together with Figure 27, show that TAAxCD28 alone or in combination with PD1 therapy does not induce systemic T cell activation compared to CD28 superagonist in cynomolgus monkeys. Cynomolgus monkeys were treated with a single dose of bispecific antibody at the indicated doses (1 mg / kg or 10 mg / kg as indicated). Time (hours) after administration is shown. Figure 26A shows serum cytokines. [Figure 26B] Figure 26B shows relative peripheral blood T cell counts. [Figure 26C] Figure 26C shows the frequency of Ki67+ and ICOS+ T cells (% of CD3). In Figures 26A-26C, values represent the mean + / - SEM. N=3 animals / group. [Figure 27]CD3 / CD28 / PSMA triple-humanized mice were treated with a single dose of antibody (0.25 mg / kg or 2.5 mg / kg, as indicated). Mice were bled 4 hours after administration (day 0) and serum was collected. Statistical significance was calculated using one-way ANOVA and Holm-Sidak multiple comparison test. *p<0.05, **p<0.01, ****p<0.0001 [Figure 28] 1 shows that PSMA×CD28+ / −PD1 did not increase serum cytokines in tumor-bearing mice. [Figure 29A] Figure 27 shows that PSMAxCD28 alone or in combination with PD1 mAb exhibits a safe cytokine profile in vivo in non-tumor-bearing mice. Data corresponds to Figure 27. CD3 / CD28 / PSMA triple-humanized mice were treated with a single dose of antibody (0.25 mg / kg or 2.5 mg / kg, as indicated). Mice were bled and serum collected 4 hours (day 0) and 72 hours (day 3) after administration. Statistical significance was calculated using one-way ANOVA and Holm-Sidak's multiple comparison test. *p<0.05, **p<0.01, ****p<0.0001 [Figure 29B] Figure 27 shows that PSMAxCD28 alone or in combination with PD1 mAb exhibits a safe cytokine profile in vivo in non-tumor-bearing mice. Data corresponds to Figure 27. CD3 / CD28 / PSMA triple-humanized mice were treated with a single dose of antibody (0.25 mg / kg or 2.5 mg / kg, as indicated). Mice were bled and serum collected 4 hours (day 0) and 72 hours (day 3) after administration. Statistical significance was calculated using one-way ANOVA and Holm-Sidak's multiple comparison test. *p<0.05, **p<0.01, ****p<0.0001 DETAILED DESCRIPTION OF THE INVENTION

[0040] Before describing the present invention, it is to be understood that this invention is not limited to the particular methods and experimental conditions described, as such methods and conditions may vary. It is also to be understood that the terminology used herein is for the purpose of describing particular embodiments only, and is not intended to be limiting, as the scope of the present invention will be limited only by the appended claims.

[0041] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art to which this invention belongs.As used herein, the term "about" when used in relation to a specific listed value means that the value can vary by 1% or less from the listed value.For example, when used in the present invention, the expression "about 100" includes 99 and 101, and all values therebetween (for example, 99.1, 99.2, 99.3, 99.4, etc.).

[0042] Any methods and materials similar or equivalent to those described herein may be used in the practice of the present invention. or can be used for testing, the preferred methods and materials are now described.All patents, applications, and non-patent publications mentioned herein are incorporated by reference in their entirety.

[0043] definition As used herein, the term "CD28" refers to an antigen expressed on T cells as a costimulatory receptor. Human CD28 comprises the amino acid sequence set forth in SEQ ID NO: 74 and / or has the amino acid sequence set forth in NCBI accession number NP_006130.1. All references herein to proteins, polypeptides, and protein fragments are intended to refer to the human form of the respective protein, polypeptide, or protein fragment, unless specifically identified as being derived from a non-human species. Thus, the term "CD28" refers to human CD28 unless specifically identified as being derived from a non-human species, e.g., "mouse CD28," "monkey CD28," etc.

[0044] As used herein, "antibodies that bind to CD28" or "anti-CD28 antibodies" include antibodies and antigen-binding fragments thereof that specifically recognize monomeric CD28, as well as antibodies and antigen-binding fragments thereof that specifically recognize dimeric CD28. The antibodies and antigen-binding fragments of the present invention can bind to soluble CD28 and / or cell surface-expressed CD28. Soluble CD28 includes native CD28 protein as well as recombinant CD28 protein variants, such as monomeric and dimeric CD28 constructs, that lack the transmembrane domain or are otherwise not associated with the cell membrane.

[0045] As used herein, the phrase "cell surface-expressed CD28" refers to one or more CD28 proteins(s) expressed on the surface of a cell in vitro or in vivo, where at least a portion of the CD28 protein is exposed to the extracellular side of the cell membrane and is accessible to the antigen-binding portion of an antibody. "Cell surface-expressed CD28" includes CD28 protein in the context of a functional T cell costimulatory receptor within the cell membrane. The phrase "cell surface-expressed CD28" includes CD28 protein expressed as part of a homodimer on the surface of a cell. "Cell surface-expressed CD28" may comprise or consist of CD28 protein expressed on the surface of a cell that normally expresses CD28 protein. Alternatively, "cell surface-expressed CD28" may comprise or consist of CD28 protein expressed on the surface of a cell that does not normally express human CD28 on its surface but has been artificially engineered to express CD28 on its surface.

[0046] As used herein, the phrase "anti-CD28 antibody" includes both monovalent antibodies having a single specificity, and bispecific antibodies comprising a first arm that binds to CD28 and a second arm that binds to a second (target) antigen, wherein the anti-CD28 arm comprises any of the HCVR / LCVR or CDR sequences set forth in Table 1 herein. Examples of anti-CD28 bispecific antibodies are described elsewhere herein. The term "antigen-binding molecule" includes antibodies and antigen-binding fragments of antibodies, including, for example, bispecific antibodies.

[0047] As used herein, the term "antibody" refers to any antigen-binding molecule or molecular complex that contains at least one complementarity-determining region (CDR) that specifically binds to or interacts with a particular antigen (e.g., CD28). The term "antibody" includes immunoglobulin molecules that contain four polypeptide chains, two heavy (H) chains and two light (L) chains, interconnected by disulfide bonds, as well as multimers thereof (e.g., 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 is made up of three domains: C, C, and L. H 1. C H 2, and C H 3 Each light chain comprises a light chain variable region (abbreviated herein as LCVR or VL) and The light chain constant region comprises one domain (C L 1) V H Area and V L The regions are composed of a number of distinct regions interspersed with more conserved regions called framework regions (FR). It can be further subdivided into hypervariable regions called complementarity-determining regions (CDRs). H and V L is in the order of FR1, CDR1, FR2, CDR2, FR3, CDR3, FR4 It consists of three CDRs and four FRs arranged from the amino terminus to the carboxy terminus. In different embodiments of the present invention, the FRs of an anti-CD28 antibody (or its antigen-binding portion) may be identical to the human germline sequence or may be naturally or artificially modified. An amino acid consensus sequence may be defined based on a parallel analysis of two or more CDRs.

[0048] The term "antibody" as used herein also includes antigen-binding fragments of complete antibody molecules. "Antigen-binding portion" of an antibody, "antigen-binding fragment" of an antibody, and similar terms, as used herein, include naturally occurring, enzymatically obtained, synthetic, or genetically engineered polypeptides or glycoproteins that specifically bind antigens to form complexes. Antibody-binding fragments of antibodies can be derived from complete antibody molecules using any suitable standard techniques, such as, for example, proteolytic or recombinant genetic engineering techniques involving the manipulation and expression of DNA encoding antibody variable domains and, optionally, constant domains. 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. DNA can be sequenced and manipulated, for example, chemically or by using molecular biology techniques, to arrange 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, etc.

[0049] Non-limiting examples of antibody-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, and (vii) minimal recognition units consisting of amino acid residues mimicking a hypervariable region of an antibody (e.g., an isolated complementarity-determining region (CDR) such as a CDR3 peptide), or a constrained FR3-CDR3-FR4 peptide. Domain-specific antibodies, single-domain antibodies, domain-deleted antibodies, chimeric antibodies, CDR-grafted antibodies, diabodies, triabodies, tetrabodies, minibodies, nanobodies (e.g., monovalent nanobodies, bivalent nanobodies, etc.), small modular immunopharmaceuticals (SMIPs), and other engineered molecules such as shark variable IgNAR domains are also encompassed by the term "antigen-binding fragment" as used herein.

[0050] An antigen-binding fragment of an antibody will typically contain at least one variable domain. A variable domain may be of any size or amino acid composition and generally contains at least one CDR adjacent to or in-frame with one or more framework sequences. L V bound to the domain H In an antibody-binding fragment having a domain, V H Domains and V L The domains can be arranged relative to each other in any suitable configuration. For example, the variable region is a dimer, with the V H -V H , V H -V L or V L -V L Alternatively, the antigen-binding fragment of an antibody may comprise a dimer of monomeric V H Domain or V L It may include a domain.

[0051] In certain embodiments, an antigen-binding fragment of an antibody may comprise at least one variable domain covalently linked to at least one constant domain. Non-limiting exemplary configurations of variable and constant domains that may be found in an antigen-binding fragment of an antibody of the invention include: (i) V H -C H 1, (ii) V H -C H 2, (iii) V H -C H 3, (iv) V H -C H 1-C H 2. (v) V H -C H 1-C H 2-C H 3. (vi) V H -C H 2 -C H 3, (vii)V H -C L , (viii) V L -C H 1, (ix)V L -CH 2, (x) V L -C H 3. (xi) V L -C H 1-C H 2, (xii)V L -C H 1-C H 2-C H 3. (x iii)V L -C H 2-C H 3, and (xiv) V L -C L The examples listed above include: In any configuration of variable and constant domains, including any of the configurations shown, the variable and constant domains may be directly linked to each other or may be linked by a complete or partial hinge or linker region. A hinge region may consist of at least two (e.g., 5, 10, 15, 20, 40, 60, or more) amino acids that provide a flexible or semi-flexible connection between adjacent variable and / or constant domains in a single polypeptide molecule. Moreover, antigen-binding fragments may be linked to each other and / or to one or more monomeric V H domain Or V L non-covalently bound to the domain (e.g., by disulfide bond(s)) In the present invention, the variable domain configurations may comprise homodimers or heterodimers (or other multimers) of any of the variable and constant domain configurations listed above.

[0052] Like intact antibody molecules, antibody-binding fragments can be monospecific or multispecific (e.g., bispecific). Multispecific antigen-binding fragments of antibodies will typically comprise at least two different variable domains, each capable of specifically binding to a separate antigen or to a different epitope on the same antigen. Any multispecific antibody format, including the exemplary bispecific antibody formats disclosed herein, can be adapted for use in connection with the antigen-binding fragments of antibodies of the present invention, using routine techniques available in the art.

[0053] The antibodies of the present invention can 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 antibodies of the present invention in the presence of complement. "Antibody-dependent cell-mediated cytotoxicity" (ADCC) refers to a cell-mediated reaction in which nonspecific cytotoxic cells expressing Fc receptors (FcRs) (e.g., natural killer (NK) cells, neutrophils, and macrophages) recognize bound antibodies on target cells, thereby resulting in lysis of the target cells. CDC and ADCC can be measured using assays well known and available in the art. (See, e.g., U.S. Patent Nos. 5,500,362 and 5,821,337, and Clynes et al. (1998) Proc. Natl. Acad. Sci. (USA) 95:652-656.) The constant region of an antibody is important in the ability of an antibody to fix complement and mediate cell-dependent cytotoxicity. Thus, the isotype of the antibody can be selected based on whether it is desirable for the antibody to mediate cytotoxicity.

[0054] In certain embodiments of the present invention, the anti-CD28 antibodies (monospecific or bispecific) of the present 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 present invention may include, for example, amino acid residues in the CDRs, particularly CDR3, that are not encoded by human germline immunoglobulin sequences (e.g., mutations introduced by random or site-specific mutagenesis in vitro or by somatic mutation in vivo). However, the term "human antibody," as used herein, is not intended to include antibodies in which CDR sequences derived from the germline of another mammalian species (e.g., a mouse) have been grafted onto human framework sequences.

[0055] The antibodies of the present invention may, in some embodiments, be recombinant human antibodies. As used herein, the term "recombinant human antibody" refers to any human antibody that is prepared, expressed, generated, or isolated by recombinant means, such as antibodies expressed using a recombinant expression vector transfected into a host cell (described below), recombinant combinatorial human antibody libraries, or the like. The term "human antibodies" is intended to include all human antibodies prepared, expressed, produced, or isolated by any other means, including antibodies isolated from a human (described below), from an animal (e.g., a mouse) that is transgenic for human immunoglobulin genes (e.g., Taylor et al. (1992) Nucl. Res. 20:6287-6295), or by any other means, including splicing of human immunoglobulin gene sequences into other DNA sequences. Such recombinant human antibodies have variable and constant regions derived from human germline immunoglobulin sequences. However, in certain embodiments, such recombinant human antibodies have been subjected to in vitro mutagenesis (or, when animals transgenic for human Ig sequences are used, in vivo somatic mutagenesis), thus modifying the V and constant regions of the recombinant antibody. H oh Yobi V L The amino acid sequence of the region is human germline VH and V L Although derived from and related to the sequence are sequences that may not naturally exist within the human antibody germline repertoire in vivo.

[0056] Human antibodies can exist in two forms related to hinge heterogeneity. In the first form, the immunoglobulin molecule contains a stable four-chain construct of approximately 150-160 kDa in which dimers are held together by interchain heavy chain disulfide bonds. In the second form, the dimers are not linked via interchain disulfide bonds, forming approximately 75-80 kDa molecules consisting of covalently linked light and heavy chains (half antibodies). These forms are extremely difficult to separate, even after affinity purification.

[0057] The frequency of occurrence of the second form in various intact IgG isotypes is due to, but not limited to, structural differences associated with the antibody hinge region isotype. A single amino acid substitution in the hinge region of a human IgG4 hinge can significantly reduce the occurrence of the second form to levels typically observed with a human IgG1 hinge (Angal et al. al. (1993) Molecular Immunology 30:105). The present invention encompasses antibodies with one or more mutations in the hinge, CH2, or CH3 regions, which may be desirable, for example, in production, to improve the yield of the desired antibody form.

[0058] The antibody of the present invention may be an isolated antibody. As used herein, an "isolated antibody" refers to an antibody that has been identified and separated and / or recovered from at least one component of its natural environment. For example, an antibody that has been separated or removed from at least one component of an organism, or from a tissue or cell in which it naturally occurs or is naturally produced, is an "isolated antibody" for purposes of the present invention. An isolated antibody also includes an antibody in situ within a recombinant cell. An isolated antibody is an antibody that has undergone at least one purification or isolation step. According to certain embodiments, an isolated antibody may be substantially free of other cellular material and / or chemicals.

[0059] The present invention also includes one-arm antibodies that bind to CD28. As used herein, "one-arm antibody" refers to an antigen-binding molecule comprising a single antibody heavy chain and a single antibody light chain. The one-arm antibodies of the present invention may comprise any of the HCVR / LCVR or CDR amino acid sequences listed in Table 2.

[0060] The anti-CD28 antibodies or antigen-binding domains thereof herein may 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 compared to the corresponding germline sequence from which the antigen-binding protein or antigen-binding domain is derived. Such mutations can be readily ascertained by comparing the amino acid sequences disclosed herein to germline sequences available, for example, from public antibody sequence databases. The present invention includes antibodies, and antigen-binding domains thereof, derived from any of the amino acid sequences disclosed herein, wherein one or more amino acids in one or more framework and / or CDR regions are replaced by the corresponding residue(s) in the germline sequence from which the antibody is derived, or by the corresponding residue(s) in another human germline sequence. The V is mutated to a specific residue or residues, or to a conservative amino acid substitution of the corresponding germline residue(s) (such sequence changes are collectively referred to herein as "germline mutations"). Starting from the heavy and light chain variable region sequences disclosed herein, one skilled in the art can readily produce many antibodies and antibody-binding fragments containing one or more individual germline mutations or combinations thereof. In certain embodiments, V H and / or V L All framework and / or CDR residues within the domain are identical to those from which the antibody is derived. In other embodiments, only certain residues are mutated back to the original germline sequence, e.g., only the mutated residues are found within the first eight amino acids of FR1 or the last eight amino acids of FR4, or only the mutated residues are found in CDR1, CDR2, or CDR3. In other embodiments, one or more of the framework and / or CDR residue(s) are mutated to the corresponding residue(s) in a different germline sequence (i.e., a germline sequence that differs from the germline sequence from which the antibody was originally derived). Furthermore, antibodies of the present invention, or antigen-binding domains thereof, may contain any combination of two or more germline mutations in the framework and / or CDR regions, e.g., certain individual residues are mutated to the corresponding residue in a particular germline sequence, while certain other residues that differ from the original germline sequence are maintained or mutated to the corresponding residue in a different germline sequence. Once obtained, antigen-binding domains containing 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 antagonist or agonist biological properties (as the case may be), reduced immunogenicity, etc. Antibodies or antigen-binding fragments thereof obtained by this general method are encompassed within the scope of the present invention.

[0061] The present invention also includes anti-CD28 antibodies and antigen-binding molecules comprising variants of any of the HCVR, LCVR, and / or CDR amino acid sequences disclosed herein. Exemplary variants included in this aspect of the invention include variants of any of the HCVR, LCVR, and / or CDR amino acid sequences disclosed herein with one or more conservative substitutions. For example, the present invention includes anti-CD28 antibodies and antigen-binding molecules having HCVR, LCVR, and / or CDR amino acid sequences with, for example, 10 or fewer, 8 or fewer, 6 or fewer, or 4 or fewer conservative amino acid substitutions relative to any of the HCVR, LCVR, and / or CDR amino acid sequences listed in Table 1 herein.

[0062] 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. A single antigen may have more than one epitope. Thus, different antibodies may bind to different regions on the antigen and have different biological effects. Epitopes can be either conformational or linear. Conformational epitopes are generated by spatially juxtaposed amino acids from different segments of a linear polypeptide chain. Linear epitopes are generated by adjacent amino acid residues in a polypeptide chain. In certain circumstances, epitopes may include carbohydrate, phosphoryl, or sulfonyl moieties on the antigen.

[0063] The term "substantial identity" or "substantially identical," when referring to a nucleic acid or fragment thereof, indicates that when optimally aligned with appropriate nucleotide insertions or deletions with another nucleic acid (or its complementary strand), there is at least about 95%, more preferably at least about 96%, 97%, 98%, or 99% nucleotide sequence identity of the nucleotide bases as measured by any well-known algorithm of sequence identity, such as FASTA, BLAST, or Gap, as discussed below. A nucleic acid molecule having substantial identity to a reference nucleic acid molecule may, in certain cases, encode a polypeptide having the same or substantially similar amino acid sequence as the polypeptide encoded by the reference nucleic acid molecule. can.

[0064] When applied to polypeptides, the terms "substantial similarity" or "substantially similar" mean that two peptide sequences, when optimally aligned using a program such as GAP or BESTFIT with a predetermined gap weight, share at least 95% sequence identity, and even more preferably at least 98% or 99% sequence identity. Preferably, non-identical residue positions differ by conservative amino acid substitutions. A "conservative amino acid substitution" is one in which an amino acid residue is replaced by another amino acid residue having a side chain (R group) with similar chemical properties (e.g., charge or hydrophobicity). Generally, conservative amino acid substitutions do not substantially alter the functional properties of a protein. When two or more amino acid sequences differ from each other by conservative substitutions, the percent sequence identity or degree of similarity may be adjusted upward to correct for the conservative nature of the substitution. Means for making this adjustment are well known to those of skill in the art. See, e.g., Pearson (1994) Methods Mol. Biol. 24:307-331. Examples of groups of amino acids having 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) basic side chains: lysine, arginine, and histidine, (6) acidic side chains: aspartic acid and glutamic acid, and (7) sulfur-containing side chains: cysteine and methionine. Preferred conservative amino acid substitution groups are valine-leucine-isoleucine, phenylalanine-tyrosine, lysine-arginine, alanine-valine, glutamic acid-aspartic acid, and asparagine-glutamine. Alternatively, a conservative substitution is any change that has a positive value in the PAM250 log-likelihood matrix disclosed in Gonnet et al. (1992) Science 256:1443-1445. A "moderately conservative" substitution is any change that has a non-negative value in the PAM250 log-likelihood matrix.

[0065] Sequence similarity for polypeptides, also known as sequence identity, is typically measured using sequence analysis software. Protein analysis software matches similar sequences using similarity measures assigned to various substitutions, deletions, and other modifications, including conservative amino acid substitutions. For example, GCG software contains programs such as Gap and Bestfit, which can be used with default parameters to determine sequence homology or sequence identity between closely related polypeptides, such as homologous polypeptides from different species of organisms, or between a wild-type protein and its mutant protein. See, for example, GCG version 6.1. Polypeptide sequences can also be compared using FASTA, a program in GCG version 6.1, with default or recommended parameters. FASTA (e.g., FASTA2 and FASTA3) provides alignments and percent sequence identity of the regions of best overlap between the query and search sequences (Pearson (2000) supra). Another preferred algorithm for comparing the sequences of the present invention to a database containing a large number of sequences from different organisms is the computer program BLAST, particularly BLASTP or TBLASTN, using default parameters. See, e.g., Altschul et al. (1990) J. Mol. Biol. 215:403-410 and Altschul et al. (1997) Nucleic Acids Res. 25:3389-402.

[0066] Bispecific antigen binding molecules The antibodies of the present invention may 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 two or more target polypeptides. See, for example, Tutt et al., 1991, J. Immunol. 147:60-69; Kufer et al. al., 2004, Trends Biotechnol. 22:238-244. The anti-CD28 antibodies of the invention can be linked to or co-expressed with another functional molecule, e.g., another peptide or protein. For example, an antibody or fragment thereof can be operatively linked (e.g., by chemical bond, genetic fusion, non-covalent bond, or other means) to one or more other molecular entities, such as another antibody or antibody fragment, to produce a bispecific or multispecific antibody having a second binding specificity.

[0067] The use of the phrase "anti-CD28 antibody" herein is intended to include both monospecific anti-CD28 antibodies and bispecific antibodies comprising a CD28-binding arm and a second arm that binds to a target antigen. Thus, the present invention includes bispecific antibodies in which one immunoglobulin arm binds to human CD28 and the other immunoglobulin arm is specific for a target antigen. The target antigen to which the other arm of a CD28 bispecific antibody binds can be any antigen expressed on or near a cell, tissue, organ, microorganism, or virus against which a targeted immune response is desired. The CD28-binding arm can comprise any of the HCVR / LCVR or CDR amino acid sequences listed in Table 1 herein. In certain embodiments, the CD28-binding arm binds to human CD28 and induces human T cell proliferation.

[0068] In the context of a bispecific antibody of the invention, in which one arm of the antibody binds to CD28 and the other arm binds to a target antigen, the target antigen can be a tumor-associated antigen such as PSMA.

[0069] According to certain exemplary embodiments, the present invention includes bispecific antigen-binding molecules that specifically bind to CD28 and PSMA. Such molecules may be referred to herein as, for example, "anti-CD28 / anti-PSMA" or "anti-CD28xPSMA" or "CD28xPSMA" or "anti-PSMA / anti-CD28" or "anti-PSMAxCD28" or "PSMAxCD28" bispecific molecules, or other similar terms.

[0070] The term "PSMA," as used herein, refers to the human PSMA protein, unless specified as being derived from a non-human species (e.g., "mouse PSMA," "simian PSMA," etc.). The human PSMA protein has the amino acid sequence set forth in SEQ ID NO: 73 and / or has the amino acid sequence set forth in NCBI Accession No. NP_004467.1.

[0071] According to certain exemplary embodiments shown in FIG. 1 , a bispecific antigen-binding molecule (e.g., a bispecific antibody) may have an effector arm and a targeting arm. The effector arm may be a first antigen-binding domain (e.g., an anti-CD28 antibody) that binds to an antigen on an effector cell (e.g., a T cell). The targeting arm may be a second antigen-binding domain (e.g., an anti-PSMA antibody) that binds to an antigen on a target cell (e.g., a tumor cell). According to certain exemplary embodiments, the effector arm binds to CD28, and the targeting arm binds to PSMA. The bispecific anti-CD28 / PSMA can provide a costimulatory signal to effector cells (e.g., T cells). The effector arm is ineffective at stimulating T cells without clustering. Once clustered, the effector arm alone has little effect at stimulating T cells. In combination with the targeting arm, the effector arm stimulates T cells. The tumor-targeting arm may have incomplete tumor specificity. The antigen targeted by the targeting arm (e.g., PSMA) may be expressed on a portion of tumor cells. The specificity of the tumor targeting arm may overlap with that of an anti-CD3 bispecific antigen-binding molecule (e.g., an anti-CD3 / PSMA bispecific antibody). Therefore, it can increase.

[0072] As used herein, the phrase "antigen-binding molecule" refers to a protein, polypeptide, or molecular complex comprising or consisting of at least one complementarity-determining region (CDR) alone or in combination with one or more additional CDRs and / or framework regions (FRs), which specifically binds to a particular antigen. In certain embodiments, the antigen-binding molecule is an antibody or a fragment of an antibody, as those terms are defined elsewhere herein.

[0073] As used herein, the term "bispecific antigen-binding molecule" refers to 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 comprises at least one CDR that specifically binds to a particular antigen, either alone or in combination with one or more additional CDRs and / or FRs. In the context of the present invention, the first antigen-binding domain specifically binds to a first antigen (e.g., CD28), and the second antigen-binding domain specifically binds to a second, distinct antigen (e.g., PSMA).

[0074] In certain exemplary embodiments of the present invention, the bispecific antigen-binding molecule is a bispecific antibody. Each antigen-binding domain of the bispecific antibody comprises a heavy chain variable domain (HCVR) and a light chain variable domain (LCVR). In the context of a bispecific antigen-binding molecule (e.g., a bispecific antibody) comprising a first and a second antigen-binding domain, the CDRs of the first antigen-binding domain may be designated with the prefix "D1", and the CDRs of the second antigen-binding domain may be designated with the prefix "D2". Thus, the CDRs of the first antigen-binding domain may be referred to herein as D1-HCDR1, D1-HCDR2, and D1-HCDR3, and the CDRs of the second antigen-binding domain may be referred to herein as D2-HCDR1, D2-HCDR2, and D2-HCDR3.

[0075] The first and second antigen-binding domains can be directly or indirectly linked to each other to form the bispecific antigen-binding molecule of the present invention. Alternatively, the first and second antigen-binding domains can each be linked to a separate multimerization domain. The association of one multimerization domain with another multimerization domain promotes the association between the two antigen-binding domains, thereby forming the bispecific antigen-binding molecule. As used herein, a "multimerization domain" is any polymer, protein, polypeptide, peptide, or amino acid capable of associating with a second multimerization domain of the same or similar structure or configuration. For example, a multimerization domain can be linked to a second multimerization domain of the same or similar structure or configuration. H 3D A non-limiting example of a multimerization component is a polypeptide comprising the Fc portion of an immunoglobulin (C H 2-C H 3 domains), for example the Fc domain of IgG selected from the isotypes IgG1, IgG2, IgG3, and IgG4, as well as any allotype within each isotype group.

[0076] The bispecific antigen-binding molecules of the present invention typically comprise two multimerization domains, e.g., two Fc domains that are each part of a separate antibody heavy chain. The first and second multimerization domains may be of the same IgG isotype, e.g., IgG1 / IgG1, IgG2 / IgG2, or IgG4 / IgG4. Alternatively, the first and second multimerization domains may be of different IgG isotypes, e.g., IgG1 / IgG2, IgG1 / IgG4, or IgG2 / IgG4.

[0077] In certain embodiments, the multimerization domain is an Fc fragment or an amino acid sequence of 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 multimerization domains include peptides or polypeptides that comprise or consist of leucine zipper, helix-loop-helix, or coiled-coil motifs.

[0078] Any bispecific antibody format or technology can be used to generate the bispecific antigen-binding molecules of the present invention. For example, an antibody or fragment thereof having a first antigen-binding specificity can be operatively linked (e.g., by chemical conjugation, genetic fusion, non-covalent bonding, etc.) to one or more other molecular entities, such as another antibody or antibody fragment having a second antigen-binding specificity, to produce a bispecific antigen-binding molecule. Specific exemplary bispecific formats that can be used in the context of the present invention include, for example, scFv-based or diabody bispecific formats, IgG-scFv fusions, dual variable domain (OVO)-Ig, Quadroma, knobs-into-holes, common light chains (e.g., common light chains with knobs-into-holes), CrossMab, CrossFab, (SEEO) bodies, leucine zippers, Ouobodies, IgG1 / IgG2, dual acting Fab (OAF)-IgG, and Mab. 2 These include, but are not limited to, bispecific formats (as described above). For a review, see, e.g., Klein et al. 2012, mAbs 4:6, 1-11, and the references cited therein).

[0079] In the context of the bispecific antigen-binding molecules of the present invention, the multimerization domain, e.g., the Fc domain, may contain one or more amino acid changes (e.g., insertions, deletions, or substitutions) compared to a wild-type, naturally occurring Fc domain. For example, the present invention includes bispecific antigen-binding molecules containing one or more modifications in the Fc domain that result in a modified Fc domain with modified binding interactions (e.g., enhanced or reduced) between Fc and FcRn. In one embodiment, the bispecific antigen-binding molecule comprises a C H 2 or C HThe FcRn-containing FcRn-binding domain contains modifications in three regions that enhance the affinity of the Fc domain for FcRn in an acidic environment (e.g., in the endosome at a pH range of about 5.5 to about 6.0). Non-limiting examples of such Fc modifications include modifications at positions 250 (e.g., E or Q), 250 and 428 (e.g., L or F), 252 (e.g., LN / FIW or T), 254 (e.g., S or T), and 256 (e.g., S / R / Q / EID or T), or at positions 428 and / or 433 (e.g., UR / S / P / Q or K) and / or 434 (e.g., H / F or V), or at positions 250 and / or 428, or at positions 307 or 308 (e.g., 308F, V308F) and 434. In one embodiment, the modifications include 428L (e.g., M428L) and 434S (e.g., N434S) modifications; 428L, 2591 (e.g., V2591), and 308F (e.g., V308F) modifications; 433K (e.g., H433K) and 434 (e.g., 434Y) modifications; 252, 254, and 256 (e.g., 252Y, 254T, and 256E) modifications, 250Q and 428L modifications (e.g., T250Q and M428L), and 307 and / or 308 modifications (e.g., 308F or 308P).

[0080] The present invention also provides a first C H 3 domain and second IgC H a bispecific antigen-binding molecule comprising three domains, a first and a second IgC H 3 domains contain at least one amino acid The first IgC differs from the second IgC by at least one amino acid difference, and the at least one amino acid difference reduces binding of the bispecific antibody to Protein A compared to a bispecific antibody lacking that amino acid difference. H The 3 domain binds to protein A and the second IgC H The 3 domain contains mutations that reduce or eliminate Protein A binding, such as the H95R modification (according to IMGT exon numbering, H435R in EU numbering). H 3 may further comprise a Y96F modification (by IMGT, Y436F by EU). Additional modifications that may be found within the second CH3 include D16E, L1 for IgG1 antibodies, These include 8M, N44S, K52N, V57M, and V821 (by IMGT; D356E, L358M, N384S, K392N, V397M, and V4221 by EU), for IgG2 antibodies, N44S, K52N, and V821 (by IMGT; N384S, K392N, and V4221 by EU), and for IgG4 antibodies, Q15R, N44S, K52N, V57M, R69K, E79Q, and V821 (by IMGT; Q355R, N384S, K392N, V397M, R409K, E419Q, and V4221 by EU).

[0081] In certain embodiments, the Fc domain can be a chimera that combines Fc sequences from two or more immunoglobulin isotypes. For example, the chimeric Fc domain can be a chimeric Fc domain that combines the Fc sequences of human IgG1, human IgG2, or human IgG4. H C derived from 2 regions H Part or all of the 2 sequence, and C derived from human IgG1, human IgG2, or human IgG4 H 3 The chimeric Fc domain may comprise, from the N-terminus to the C-terminus, a "chimeric hinge" sequence derived from a human IgG1 hinge region, a human IgG2 hinge region, or a human IgG4 hinge region in combination with a "lower hinge" sequence derived from a human IgG1 hinge region, a human IgG2 hinge region, or a human IgG4 hinge region. A specific example of a chimeric Fc domain that may be included in any of the antigen-binding molecules described herein is a chimeric Fc domain that comprises, from the N-terminus to the C-terminus, a "chimeric hinge" sequence derived from a human IgG1 hinge region, a human IgG2 hinge region, or a human IgG4 hinge region in combination with a "lower hinge" sequence derived from a human IgG1 hinge region, a human IgG2 hinge region, or a human IgG4 hinge region. H 1]-[IgG4 upper hinge]-[IgG2 lower hinge]-[IgG4 CH 2]-[IgG4 C H 3]. Any of the antigen-binding molecules described herein. Another example of a chimeric Fc domain that may be included in the method is one that contains, from the N-terminus to the C-terminus, [IgG1 C H 1]-[IgG1 upper hinge]-[IgG2 lower hinge]-[IgG4 C H 2]-[I gG1 C H 3]. Any of the antigen-binding molecules of the present invention can be included. These and other examples of chimeric Fc domains are described in WO2014 / 022540A1, and chimeric Fc domains having these general structural sequences, and variants thereof, can alter Fc receptor binding, which in turn affects Fc effector function.

[0082] Sequence variants 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 and / or CDR regions of the heavy and light chain variable domains compared to the corresponding germline sequences from which the individual antigen-binding domains were derived. Such mutations can be readily identified by comparing the amino acid sequences disclosed herein with germline sequences available, for example, from public antibody sequence databases. The antigen-binding molecules of the present invention may comprise antigen-binding fragments derived from any of the exemplary amino acid sequences disclosed herein, in which one or more amino acids in one or more framework and / or CDR regions are mutated to the corresponding residue(s) in the germline sequence from which the antibody was derived, or to the corresponding residue(s) in another human germline sequence, or to conservative amino acid substitutions of the corresponding germline residue(s) (such sequence changes are collectively referred to herein as "germline mutations"). Starting from the heavy and light chain variable region sequences disclosed herein, one skilled in the art can easily produce numerous antibodies and antibody-binding fragments containing one or more individual germline mutations or combinations thereof. In certain embodiments, V H and / or V LAll framework and / or CDR residues within the domain 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, e.g., only the mutated residues are found within the first 8 amino acids of FR1 or the last 8 amino acids of FR4, or only the mutated residues are found in CDR1, CDR2, or CDR3. In other embodiments, one or more of the framework and / or CDR residue(s) are mutated back to the residues found in a different germline sequence (i.e., the sequence from which the antigen-binding domain was originally derived). The antigen-binding domain is mutated to the corresponding residue(s) in a particular germline sequence (different from the germline sequence from which it was derived). Furthermore, the antigen-binding domain may contain any combination of two or more germline mutations within the framework and / or CDR regions, e.g., certain individual residues are mutated to the corresponding residue in a particular germline sequence, while certain other residues that differ from the original germline sequence are either maintained or mutated to the corresponding residue in a different germline sequence. Once obtained, antigen-binding domains containing 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 antagonist or agonist biological properties (as the case may be), reduced immunogenicity, etc. Bispecific antigen-binding molecules comprising one or more antigen-binding domains obtained in this general manner are encompassed within the scope of the present invention.

[0083] The present invention also includes antigen-binding molecules in which one or both antigen-binding domains comprise variants of any of the HCVR, LCVR, and / or CDR amino acid sequences disclosed herein with one or more conservative substitutions. For example, the present invention includes antigen-binding molecules comprising antigen-binding domains having HCVR, LCVR, and / or CDR amino acid sequences with, for example, 10 or fewer, 8 or fewer, 6 or fewer, or 4 or fewer conservative amino acid substitutions relative to any of the HCVR, LCVR, and / or CDR amino acid sequences disclosed herein. A "conservative amino acid substitution" is one in which an amino acid residue is replaced with another amino acid residue having a side chain (R group) with similar chemical properties (e.g., charge or hydrophobicity). Generally, conservative amino acid substitutions will not substantially alter the functional properties of a protein. Examples of groups of amino acids having 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) basic side chains: lysine, arginine, and histidine, (6) acidic side chains: aspartic acid and glutamic acid, and (7) sulfur-containing side chains: cysteine and methionine. Preferred conservative amino acid substitution groups are valine-leucine-isoleucine, phenylalanine-tyrosine, lysine-arginine, alanine-valine, glutamic acid-aspartic acid, and asparagine-glutamine. Alternatively, a conservative substitution is any change that has a positive value in the PAM250 log-likelihood matrix disclosed in Gonnet et al. (1992) Science 256:1443-1445. A "moderately conservative" substitution is any change that has a non-negative value in the PAM250 log-likelihood matrix.

[0084] The present invention also includes antigen-binding molecules comprising antigen-binding domains having HCVR, LCVR, and / or CDR amino acid sequences substantially identical to any of the HCVR, LCVR, and / or CDR amino acid sequences disclosed herein. When referring to amino acid sequences, the term "substantial identity" or "substantially identical" means that two amino acid sequences share at least 95% sequence identity, and even more preferably at least 98% or 99% sequence identity, when optimally aligned using programs such as GAP or BESTFIT with a predetermined gap weight. Preferably, residue positions that are not identical differ by conservative amino acid substitutions. When two or more amino acid sequences differ from each other by conservative substitutions, the percentage of sequence identity or degree of similarity may be adjusted upward to correct for the conservative nature of the substitutions. 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.

[0085] Sequence similarity for polypeptides, also called sequence identity, is typically measured using sequence analysis software. Protein analysis software can analyze sequences that include conservative amino acid substitutions. Similar sequences are matched using similarity measures assigned to various substitutions, deletions, and other modifications, including substitutions, deletions, and other modifications. For example, GCG software contains programs such as Gap and Bestfit, which can be used with default parameters to determine sequence homology or sequence identity between closely related polypeptides, such as homologous polypeptides from different species of organisms, or between a wild-type protein and its mutant protein. See, for example, GCG version 6.1. Polypeptide sequences can also be compared using FASTA, a program in GCG version 6.1, with default or recommended parameters. FASTA (e.g., FASTA2 and FASTA3) provides alignments and percent sequence identity of the regions of best overlap between the query and search sequences (Pearson (2000) supra). Another preferred algorithm for comparing the sequences of the present invention to a database containing a large number of sequences from different organisms is the computer program BLAST, particularly BLASTP or TBLASTN, using default parameters. See, e.g., Altschul et al. (1990) J. Mol. Biol. 215:403-410 and Altschul et al. (1997) Nucleic Acids Res. 25:3389-402.

[0086] pH dependent binding The present invention includes anti-CD28 / anti-PSMA bispecific antigen-binding molecules with pH-dependent binding properties. For example, an anti-CD28 antibody of the present invention may exhibit reduced binding to CD28 at acidic pH compared to neutral pH. Alternatively, an anti-PSMA antibody of the present invention may exhibit enhanced binding to PSMA at acidic pH compared to neutral pH. The term "acidic pH" includes pH values below about 6.2, e.g., 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 lower. As used herein, the term "neutral pH" refers to a pH of about 7.0 to about 7.4. The expression "neutral pH" includes pH values of about 7.0, 7.05, 7.1, 7.15, 7.2, 7.25, 7.3, 7.35, and 7.4.

[0087] In some cases, "reduced binding at acidic pH compared to neutral pH" refers to the K of an antibody that binds to its antigen at neutral pH. D K value of an antibody that binds to its antigen at acidic pH D For example, an antibody or antigen-binding fragment thereof may have an acidic / neutral K of about 3.0 or greater. D For the purposes of the present invention, when a ratio is presented, an antibody or In certain exemplary embodiments, the acidic / neutral K of an antibody or antigen-binding fragment of the invention can be considered to exhibit "reduced binding to CD28 at acidic pH compared to neutral pH." D The ratios are approximately 3.0, 3.5, 4.0, 4.5, and 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 more.

[0088] Antibodies with pH-dependent binding properties can be obtained, for example, by screening a population of antibodies for reduced (or enhanced) binding to a specific antigen at acidic pH compared to neutral pH. Furthermore, modification of the antigen-binding domain at the amino acid level can produce antibodies with pH-dependent characteristics. For example, by substituting one or more amino acids in the antigen-binding domain (e.g., within the CDR) with histidine residues, an antibody can be obtained that has reduced antigen binding at acidic pH compared to neutral pH.

[0089] Antibodies containing Fc variants According to certain embodiments of the invention, an Fc domain comprising one or more mutations that enhance or decrease antibody binding to the FcRn receptor, e.g., at acidic pH compared to neutral pH. For example, the present invention provides an anti-CD28 / anti-PSMA bispecific antigen-binding molecule comprising the C of the Fc domain. H 2 or C HThe present invention includes antibodies and antigen-binding molecules containing mutations in the FcRn region that increase the affinity of the Fc domain for FcRn in acidic environments (e.g., within endosomes at a pH range of about 5.5 to about 6.0). Such mutations can increase the serum half-life of the antibody when administered to an animal. Non-limiting examples of such Fc modifications include, for example, modifications at positions 250 (e.g., E or Q), 250 and 428 (e.g., L or F), 252 (e.g., L / Y / F / W or T), 254 (e.g., S or T), and 256 (e.g., S / R / Q / E / D or T), or modifications at positions 428 and / or 433 (e.g., H / L / R / S / P / Q or K) and / or 434 (e.g., H / F or Y), or modifications at positions 250 and / or 428, or at positions 307 or 308 (e.g., 308F, V308F), and 434. In one embodiment, the modifications include 428L (e.g., M428L) and 434S (e.g., N434S) modifications; 428L, 259I (e.g., V259I), and 308F (e.g., V308F) modifications; 433K (e.g., H433K) and 434 (e.g., 434Y) modifications; 252, 254, and 256 (e.g., 252Y, 254T, and 256E) modifications; 250Q and 428L modifications (e.g., T250Q and M428L); and 307 and / or 308 modifications (e.g., 308F or 308P).

[0090] For example, the present invention includes anti-CD28 / anti-PSMA bispecific antigen-binding molecules comprising an Fc domain containing one or more pairs or groups of mutations selected from the group consisting of 250Q and 248L (e.g., T250Q and M248L); 252Y, 254T and 256E (e.g., M252Y, S254T and T256E); 428L and 434S (e.g., M428L and N434S); and 433K and 434F (e.g., H433K and N434F). All possible combinations of the foregoing Fc domain mutations, and other mutations in antibody variable domains disclosed herein, are contemplated within the scope of the present invention.

[0091] Biological properties of antibodies and antigen-binding molecules The present invention includes antibodies and antigen-binding fragments thereof that bind to human CD28 and / or PSMA with high affinity. The present invention also includes antibodies and antigen-binding fragments thereof that bind to human CD28 and / or PSMA with moderate or low affinity, depending on the therapeutic situation and the specific targeting properties desired. For example, in the context of a bispecific antigen-binding molecule in which one arm binds to CD28 and another arm binds to a target antigen (e.g., PSMA), it may be desirable for the target antigen-binding arm to bind to the target antigen with high affinity, while the anti-CD28 arm binds to CD28 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 / non-target CD28 binding and the resulting adverse side effects associated therewith.

[0092] According to certain embodiments, the present invention provides a method for the treatment of rhodopsin-related ... D Antibodies that bind to human CD28 at (e.g., 25°C) and antigen binding of the antibodies In certain embodiments, antibodies or antigen-binding fragments of the invention have a K of less than about 150 nM, less than about 130 nM, less than about 120 nM, less than about 100 nM, less than about 50 nM, less than about 80 nM, less than about 60 nM, less than about 40 nM, or less than about 30 nM, as measured by surface plasmon resonance, e.g., using an assay format as defined in Example 3 herein, or a substantially similar assay. D to CD28 In certain embodiments, the antibodies or antigen-binding fragments of the invention have a K of between about 30 nM and about 207 nM. D It binds to CD28.

[0093] The present invention also includes antibodies and antigen-binding fragments thereof that bind to CD28 with a dissociation half-life (t) of greater than about 3.5 minutes as measured by surface plasmon resonance at 25°C or 37°C, e.g., using an assay format as defined in Example 3 herein, or a substantially similar assay. In certain embodiments, antibodies or antigen-binding fragments of the invention bind to CD28 with a t of greater than about 5 minutes, greater than about 10 minutes, greater than about 20 minutes, greater than about 30 minutes, greater than about 40 minutes, greater than about 50 minutes, greater than about 60 minutes, greater than about 70 minutes, greater than about 80 minutes, greater than about 90 minutes, greater than about 100 minutes, greater than about 200 minutes, greater than about 300 minutes, greater than about 400 minutes, greater than about 500 minutes, greater than about 600 minutes, greater than about 700 minutes, greater than about 800 minutes, greater than about 900 minutes, greater than about 1000 minutes, or greater than about 1200 minutes as measured by surface plasmon resonance at 25°C or 37°C using, for example, an assay format as defined in Example 3 herein, or a substantially similar assay.

[0094] The present invention includes bispecific antigen-binding molecules (e.g., bispecific antibodies) capable of simultaneously binding to human CD28 and human PSMA. According to certain embodiments, the bispecific antigen-binding molecules of the present invention specifically interact with cells expressing CD28 and / or PSMA. The extent to which a bispecific antigen-binding molecule binds to cells expressing CD28 and / or PSMA can be assessed by fluorescence-activated cell sorting (FACS), as exemplified in Example 4 herein. For example, the present invention includes bispecific antigen-binding molecules that specifically bind to a human cell line that expresses CD28 but not PSMA (e.g., HEK293 engineered to express CD28) and a human prostate cancer cell line that expresses PSMA but not CD28 (e.g., C4-2). The present invention provides a bispecific antigen-binding molecule that specifically binds to approximately 9.6 x 10 cells / mL of CD28, as determined using a FACS assay as described in Example 4 or a substantially similar assay. -9 ~Approx. 3.5×10 -10 , or lower EC 50 value, the aforementioned cells and The present invention includes a bispecific antigen-binding molecule that binds to either of the cell lines.

[0095] The present invention also provides anti-CD28 / anti-PSMA bispecific antigen binding molecules that induce or increase T cell-mediated killing of tumor cells. For example, the present invention provides anti-CD28 / anti-PSMA bispecific antigen binding molecules that induce or increase T cell-mediated killing of tumor cells with an EC of less than about 78 pM as measured in an in vitro T cell-mediated tumor cell killing assay using an assay format as defined in Example 6 herein (e.g., assessing the extent of C4-2 tumor cell killing by human PBMCs in the presence of anti-CD28 x PSMA antibodies) or a substantially similar assay. 50 In certain embodiments, the antibodies or antigen-binding fragments of the invention have an EC50 of less than about 40 pM, less than about 20 pM, less than about 16 pM, less than about 10 pM, less than about 5.0 pM, less than about 4.0 pM, less than about 3.0 pM, less than about 2.5 pM, less than about 2.0 pM, less than about 1.5 pM, or less than about 1.45 pM, as measured by an in vitro T-cell-mediated tumor cell killing assay using an assay format as defined in Example 6 herein or a substantially similar assay. 50 At these values, it induces T cell-mediated tumor cell killing (e.g., PBMC-mediated killing of C4-2 cells).

[0096] The present invention also provides a compound having an EC of between 1.0 pM and 1000 nM. 50 In certain embodiments, the anti-CD28 / anti-PSMA bispecific antigen binding molecule binds to CD28-expressing human T cells with an EC value between 48 nM and 180 nM. 50 For example, the present invention provides antibodies that bind to human T cells expressing CD28 with an EC value of about 1 pM, about 10 pM, about 100 pM, about 500 pM, about 1 nM, about 2 nM, about 5 nM, about 10 nM, about 20 nM, about 30 nM, about 40 nM, about 50 nM, about 60 nM, about 70 nM, about 80 nM, about 90 nM, about 100 nM, about 200 nM, about 300 nM, about 500 nM, about 800 nM, about 1000 nM, or more. 50 The present invention relates to a bispecific antigen-binding molecule that binds to human T cells expressing CD28 at a specific level.

[0097] The present invention also provides antibodies that (a) induce T cell proliferation in vitro (see, e.g., Example 8 herein), (b) activate T cells and induce upregulation of CD25 and PD-1 in human PBMCs (see, e.g., Example 8 herein), (c) increase human T cell-mediated cytotoxicity against PSMA-expressing cell lines (see, e.g., Example 8 herein), (d) induce naive primate T cell-mediated cytotoxicity against PSMA-expressing cell lines (see, e.g., Example 8 herein), (e) deplete tumor cells in mice (see, e.g., Example 10 herein), and (f) induce tumor cell proliferation in mice. (g) enhance tumor clearance in rhesus monkeys (e.g., Example 10 herein); (g) do not induce cytokine storm (e.g., Example 10 herein); (h) do not induce systemic T cell action in cynomolgus monkeys (e.g., Example 11 herein); (i) enhance the effect of PD-1 blockade on tumor cell killing induced by T cell activation (e.g., Example 13 herein); and (j) enhance the proliferation of memory T cells (e.g., Example 13 herein).

[0098] The present invention includes anti-CD28 / anti-PSMA bispecific antigen-binding molecules that can deplete tumor cells in a subject (see, e.g., Example 8). For example, according to certain embodiments, anti-CD28 / anti-PSMA bispecific antigen-binding molecules are provided, wherein a single administration of the bispecific antigen-binding molecule to a subject (e.g., about 0.1 mg / kg, about 0.08 mg / kg, about 0.06 mg / kg, about 0.04 mg / kg, about 0.04 mg / kg, about 0.02 mg / kg, about 0.01 mg / kg, or less) causes a reduction in the number of tumor cells in the subject.

[0099] Epitope mapping and related techniques The epitopes on CD28 and / or PSMA to which the antigen-binding molecules of the present invention bind may consist of a single contiguous sequence of three or more (e.g., 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, or more) amino acids of the CD28 or PSMA protein. Alternatively, the epitope may consist of multiple non-contiguous amino acids (or amino acid sequences) of CD28 or PSMA. The antibodies of the present invention may interact with amino acids contained within a CD28 monomer or with amino acids on two or more different CD28 chains of a CD28 dimer. 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 a paratope. A single antigen may have two or more epitopes. Thus, different antibodies may bind to different regions of an antigen and have different biological effects. Epitopes can be either conformational or linear. Conformational epitopes are generated by spatially juxtaposed amino acids from different segments of a linear polypeptide chain. Linear epitopes are generated by adjacent amino acid residues in a polypeptide chain. In certain circumstances, epitopes can include sugar, phosphoryl, or sulfonyl moieties on an antigen.

[0100] Various techniques known to those skilled in the art can be used to determine whether an antigen-binding domain of an antibody "interacts with one or more amino acids" within a polypeptide or protein. Exemplary techniques that can be used to determine the epitope or binding domain of a particular antibody or antigen-binding domain include, for example, Antibodies, Harlow and These include routine cross-blocking assays such as those described in Lane (Cold Spring Harbor Press, Cold Spring Harbor, NY), point mutagenesis (e.g., alanine scanning mutagenesis, arginine scanning mutagenesis, etc.), peptide blot analysis (Reineke, 2004, Methods Mol Biol 248:443-463), protease protection, and peptide cleavage analysis. In addition, methods such as epitope excision, epitope extraction, and chemical modification of antigens may be employed. (Tomer, 2000, Protein Science 9:487-496). Another method that can be used to identify amino acids within a polypeptide with which an antibody interacts is hydrogen / deuterium exchange detected by mass spectrometry. Generally speaking, the hydrogen / deuterium exchange method involves deuterium-labeling a protein of interest and then binding an antibody to the deuterium-labeled protein. The protein / antibody complex is then transferred to water, allowing hydrogen-deuterium exchange to occur at all residues except those protected by the antibody (which remain deuterium-labeled). After dissociation of the antibody, the target protein is subjected to protease cleavage and mass spectrometry, thereby revealing the deuterium-labeled residues corresponding 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 can also be used to identify amino acids within a polypeptide with which an antibody interacts.

[0101] The present invention further includes anti-CD28 and anti-PSMA antibodies (e.g., antibodies comprising any of the amino acid sequences as set forth in Table 1 herein) that bind to the same epitope as any of the specific exemplary antibodies described herein. Similarly, the present invention also includes anti-CD28 and / or anti-PSMA antibodies (e.g., antibodies comprising any of the amino acid sequences as set forth in Table 1 herein) that compete with any of the specific exemplary antibodies described herein for binding to CD28 and / or PSMA.

[0102] The present invention also includes bispecific antigen-binding molecules comprising a first antigen-binding domain that specifically binds human CD28 and a second antigen-binding domain that specifically binds human PSMA, wherein the first antigen-binding domain binds to the same epitope on CD28 as any of the specific exemplary CD28-specific antigen-binding domains described herein, and / or the second antigen-binding domain binds to the same epitope on PSMA as any of the specific exemplary PSMA-specific antigen-binding domains described herein.

[0103] Similarly, the present invention also includes bispecific antigen-binding molecules comprising a first antigen-binding domain that specifically binds human CD28 and a second antigen-binding fragment that specifically binds human PSMA, wherein the first antigen-binding domain competes for binding to CD28 with any of the specific exemplary CD28-specific antigen-binding domains described herein, and / or the second antigen-binding domain competes for binding to PSMA with any of the specific exemplary PSMA-specific antigen-binding domains described herein.

[0104] Whether a particular antigen-binding molecule (e.g., an antibody) or its antigen-binding domain binds to the same epitope as a reference antigen-binding molecule of the present invention or competes for binding with a reference antigen-binding molecule of the present invention can be easily determined using routine methods known in the art. For example, to determine whether a test antibody binds to the same epitope on CD28 (or PSMA) as a reference bispecific antigen-binding molecule of the present invention, the reference bispecific molecule is first bound to the CD28 protein (or PSMA protein). The ability of the test antibody to bind to the CD28 (or PSMA) molecule is then evaluated. If the test antibody can bind to CD28 (or PSMA) after saturation binding with the reference bispecific antigen-binding molecule, it can be concluded that the test antibody binds to a different epitope on CD28 (or PSMA) from the reference bispecific antigen-binding molecule. On the other hand, if the test antibody cannot bind to CD28 (or PSMA) after saturation binding with the reference bispecific antigen-binding molecule, the test antibody may bind to the same epitope on CD28 (or PSMA) as the epitope bound by the reference bispecific antigen-binding molecule of the present invention. Next, it is possible to determine whether the observed loss of binding of the test antibody is in fact due to binding to the same epitope as the reference bispecific antigen-binding molecule, or whether steric blocking (or another phenomenon) is responsible for the observed loss of binding. Further routine experiments (e.g., peptide mutations and binding analysis) can be performed to confirm whether or not the two antigen-binding proteins bind to the same (or overlapping) epitope. This type of experiment can be performed using ELISA, RIA, Biacore, flow cytometry, or any other quantitative or qualitative antibody binding assay available in the art. According to certain embodiments of the invention, two antigen-binding proteins bind to the same (or overlapping) epitope if, for example, a 1-fold, 5-fold, 10-fold, 20-fold, or 100-fold excess of one antigen-binding protein inhibits binding of the other by at least 50%, but preferably 75%, 90%, or even 99%, as measured in a competitive binding assay (see, e.g., Junghans et al., J. Immunol. 2010, 102, 103, 104, 105, 106, 107, 108, 109, 110, 111, 112, 113, 114, 115, 116, 117, 118, 119, 120, 121, 122, 123, 124, 125, 126, 127, 128, 129, 130, 131, 132, 133, 134, 135, 136, 137, 138, 139, 140, 141, 142, 143, 144, 145, 146, 147, 148, 149, 150, 151, 152, 153, 154, 155, 156, 157, 158, 159, 160, 161 (See, e.g., et al., Cancer Res. 1990:50:1495-1502). Alternatively, two antigen-binding proteins are considered to bind the same epitope if essentially all amino acid mutations in the antigen that reduce or eliminate binding of one antigen-binding protein also reduce or eliminate binding of the other. Two antigen-binding proteins are considered to have "overlapping epitopes" if only a subset of the amino acid mutations that reduce or eliminate binding of one antigen-binding protein also reduce or eliminate binding of the other.

[0105] To determine whether an antibody or its antigen-binding domain competes with a reference antigen-binding molecule for binding, the above-mentioned binding method is carried out in two ways. In the first way, the reference antigen-binding molecule is bound to CD28 protein (or PSMA protein) under saturating conditions, and then the binding of the test antibody to the CD28 (or PSMA) molecule is evaluated. In the second way, the test antibody is bound to CD28 (or PSMA) molecule under saturating conditions, and then the binding of the reference antigen-binding molecule to the CD28 (or PSMA) molecule is evaluated. In both ways, if only the first (saturating) antigen-binding molecule can bind to the CD28 (or PSMA) molecule, it is concluded that the test antibody and the reference antigen-binding molecule compete for binding to CD28 (or PSMA). As will be recognized by those skilled in the art, an antibody that competes for binding with a reference antigen-binding molecule does not necessarily bind to the same epitope as the reference antibody, but may sterically block the binding of the reference antibody by binding to an overlapping or adjacent epitope.

[0106] Preparation of antigen-binding domains and construction of bispecific molecules Antigen-binding domains specific to a particular antigen can be prepared by any antibody production technique known in the art. Once obtained, two different antigen-binding domains specific to two different antigens (e.g., CD28 and PSMA) can be appropriately positioned relative to one another to produce a bispecific antigen-binding molecule of the present invention using routine methods. (A discussion of exemplary bispecific antibody formats that can be used to construct bispecific antigen-binding molecules of the present invention is provided elsewhere herein.) In certain embodiments, one or more individual components (e.g., heavy and light chains) of a multispecific antigen-binding molecule of the present invention are derived from a chimeric antibody, a humanized antibody, or a fully human antibody. Methods for producing such antibodies are well known in the art. For example, one or more heavy and / or light chains of a bispecific antigen-binding molecule of the present invention can be prepared using VELOCIMMUNE™ technology. Using VELOCIMMUNE™ technology (or any other human antibody generation technology), a high-affinity chimeric antibody against a particular antigen (e.g., CD28 or PSMA) having a human variable region and a mouse constant region is first isolated. Antibodies are characterized and selected for desirable characteristics, including affinity, selectivity, epitope, etc. The mouse constant regions are replaced with the desired human constant regions to produce fully human heavy and / or light chains that can be incorporated into the bispecific antigen-binding molecules of the invention.

[0107] Genetically engineered animals can be used to generate human bispecific antigen-binding molecules. For example, genetically modified mice can be used that are unable to rearrange and express endogenous mouse immunoglobulin light chain variable sequences, and the mice are encoded with human immunoglobulin sequences operably linked to mouse kappa constant genes at the endogenous mouse kappa locus. These genetically engineered mice express only one or two human light chain variable domains encoded by the human light chain variable region gene segments. Such genetically engineered mice can be used to produce fully human bispecific antigen-binding molecules containing two different heavy chains associated with the same light chain containing variable domains derived from one of two different human light chain variable region gene segments. (See, e.g., US2011 / 0195454 for a detailed discussion of such engineered mice and their use to produce bispecific antigen-binding molecules.)

[0108] biological equivalent The present invention encompasses antigen-binding molecules having amino acid sequences that differ from those of the described antibodies but retain the ability to bind to CD28 and / or PSMA. Such variant antibodies contain one or more amino acid additions, deletions, or substitutions compared to the parent sequence, but exhibit essentially equivalent biological activity to that of the described antigen-binding molecules. Similarly, DNA sequences encoding the antibody-binding molecules of the present invention encompass sequences that contain one or more nucleotide additions, deletions, or substitutions compared to the disclosed sequences, but are essentially biologically equivalent to the described antigen-binding molecules of the present invention. Examples of such variant amino acid and DNA sequences are discussed above.

[0109] The present invention includes antigen-binding molecules that are biologically equivalent to any of the exemplary antigen-binding molecules described herein. Two antigen-binding proteins or antibodies are considered to be bioequivalent if, for example, they are pharmaceutical equivalents or pharmaceutical alternatives that do not show significant differences in absorption rate and extent when administered at the same molar dose, either in a single dose or multiple doses, under similar experimental conditions. Some antibodies are considered to be equivalent or pharmaceutical alternatives if their absorption extent is equivalent but their absorption rate is not, and such differences in absorption rate are intentional and reflected in the labeling, and therefore can be considered bioequivalent, for example, they are not essential for achieving effective body drug concentrations in long-term use and are not considered medically significant for the particular drug product studied.

[0110] In one embodiment, two antigen binding proteins are bioequivalent if there are no clinically significant differences in their safety, purity, or efficacy.

[0111] In one embodiment, two antigen binding proteins are bioequivalent if a patient can be switched one or more times compared to therapy continued without switching between the reference product and the biological product without an expected increase in the risk of adverse effects, including clinically significant changes in immunogenicity or reduced efficacy.

[0112] In one embodiment, two antigen binding proteins are biologically equivalent if they both act by a common mechanism or mode of action for a condition or condition of use, to the extent that such mechanism is known.

[0113] Bioequivalence may be demonstrated by in vivo and in vitro methods. Bioequivalence measurements include, for example, (a) in vivo studies in humans or other mammals in which the concentration of an antibody or its metabolites is measured as a function of time in blood, plasma, serum, or other biological fluids, (b) in vitro studies that correlate with and reasonably predict human bioavailability data, (c) in vivo studies in humans or other mammals in which the relevant acute pharmacological effect of the antibody (or its target) is measured as a function of time, and (d) well-controlled clinical trials that establish the safety, efficacy, or bioavailability or bioequivalence of the antibody.

[0114] Biologically equivalent variants of the exemplary bispecific antigen-binding molecules shown herein can be prepared, for example, by producing various substitutions of residues or sequences, or by adding termini not required for biological activity. Biologically equivalent antibodies may be constructed by deleting internal or internal residues or sequences. For example, cysteine residues that are not essential for biological activity can be deleted or substituted with other amino acids to prevent the formation of unnecessary or incorrect intramolecular disulfide bridges during renaturation. In other contexts, biologically equivalent antibodies may include variants of the exemplary bispecific antigen-binding molecules described herein that contain amino acid changes that modify the glycosylation characteristics of the antibody, e.g., mutations that eliminate or remove glycosylation.

[0115] Species selectivity and species cross-reactivity According to certain embodiments, the present invention provides antigen-binding molecules that bind to human CD28 but not to CD28 from other species. The present invention also provides antigen-binding molecules that bind to human PSMA but not to PSMA from other species. The present invention also includes antigen-binding molecules that bind to human CD28 and CD28 from one or more non-human species, and / or antigen-binding molecules that bind to human PSMA and PSMA from one or more non-human species.

[0116] According to certain exemplary embodiments of the present invention, there is provided an antigen-binding molecule that binds to human CD28 and / or human PSMA, and optionally binds to one or more of mouse, rat, guinea pig, hamster, gerbil, pig, cat, dog, rabbit, goat, sheep, cow, horse, camel, cynomolgus monkey, marmoset, rhesus monkey, or chimpanzee CD28 and / or PSMA, or does not bind to these. For example, certain exemplary embodiments of the present invention provide a bispecific antigen-binding molecule comprising a first antigen-binding domain that binds to human CD28 and cynomolgus monkey CD28, and a second antigen-binding domain that specifically binds to human PSMA.

[0117] immunoconjugate The present invention encompasses antigen-binding molecules conjugated to a therapeutic moiety ("immunoconjugate"), such as a cytotoxin, a chemotherapeutic agent, an immunosuppressant, or a radioisotope. Cytotoxic agents include any agent that is harmful to cells. Examples of cytotoxic agents and chemotherapeutic agents suitable for forming immunoconjugates are known in the art (see, for example, WO05 / 103081).

[0118] Therapeutic Formulations and Administration The present invention provides pharmaceutical compositions comprising the antigen-binding molecules of the present invention. The pharmaceutical compositions of the present invention are formulated with suitable carriers, excipients, and other agents that provide improved transfer, delivery, tolerance, etc. Many suitable formulations can be found in formularies known to all pharmaceutical chemists: Remington's Pharmaceutical Sciences, Mack Publishing Company, Easton, PA. These formulations include, for example, powders, pastes, ointments, jellies, waxes, oils, lipids, lipid (cationic or anionic) containing vesicles (such as LIPOFECTIN™, Life Technologies, Carlsbad, CA), DNA conjugates, anhydrous absorption pastes, oil-in-water and water-in-oil emulsions, emulsion carbowax (polyethylene glycol of various 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.

[0119] The dose of the antigen-binding molecule administered to a patient may vary depending on the age and size of the patient, the target disease, pathological condition, route of administration, etc. A preferred dose is typically calculated according to body weight or body surface area. When the bispecific antigen-binding molecule of the present invention is used for the treatment of adult patients, In such cases, it may be advantageous to intravenously administer the bispecific antigen-binding molecules of the present invention in a single dose of typically about 0.01 to about 20 mg / kg body weight, more preferably about 0.02 to about 7, about 0.03 to about 5, or about 0.05 to about 3 mg / kg body weight. The frequency and duration of treatment can be adjusted depending on the severity of the condition. Effective doses and schedules for administering bispecific antigen-binding molecules can be determined empirically; for example, the patient's progress can be monitored by periodic evaluation, and the dose can be adjusted accordingly. Furthermore, interspecies scaling of dosages can be performed using methods well known in the art (e.g., Mordenti et al., 1991, Pharmaceut. Res. 8:1351).

[0120] Various delivery systems, such as encapsulation in liposomes, microparticles, microcapsules, recombinant cells capable of expressing mutant viruses, and receptor-mediated endocytosis, are known and can be used to administer the pharmaceutical compositions of the present invention (see, e.g., Wu et al. (1987) J. Biol. Chem. 262:4429-4432). Methods of introduction include, but are not limited to, intradermal, transdermal, intramuscular, intraperitoneal, intravenous, subcutaneous, intranasal, epidural, and oral routes. The compositions can be administered by any convenient route, for example, by infusion or bolus injection, or absorption through epithelial or mucocutaneous linings (e.g., oral mucosa, rectal, and intestinal mucosa), and can be administered together with other biologically active agents. Administration can be systemic or local.

[0121] The pharmaceutical composition of the present invention can be delivered subcutaneously or intravenously using a standard needle and syringe. In addition, for subcutaneous delivery, a pen delivery device is easily applicable for delivering the pharmaceutical composition of the present invention. Such a pen delivery device can be reusable or disposable. Reusable pen delivery devices generally utilize a replaceable cartridge containing the pharmaceutical composition. Once the pharmaceutical composition in the cartridge has been administered and the cartridge is empty, the empty cartridge can be easily discarded and replaced with a new cartridge containing the pharmaceutical composition. The pen delivery device can then be reused. In disposable pen delivery devices, there is no replaceable cartridge. Rather, the disposable pen delivery device is pre-filled with the pharmaceutical composition held in a reservoir within the device. Once the reservoir is empty of the pharmaceutical composition, the entire device is discarded.

[0122] Numerous reusable pen delivery devices and autoinjector delivery devices have utility in the subcutaneous delivery of the pharmaceutical compositions of the present invention. Examples include the AUTOPEN™ (Owen Mumford, Inc., Woodstock, UK), the DISETRONIC™ pen (Disetronic Medical Systems, Bergdorf, Switzerland), the HUMALOG MIX 75 / 25™ pen, the HUMALOG™ pen, the HUMALIN 70 / 30™ pen (Eli Lilly and Co., Indianapolis, IN), the NOVOPEN™ I, II, and III (Novo Nordisk, Copenhagen, Denmark), the NOVOPEN JUNIOR™ (Novo Nordisk, Copenhagen, Denmark), the BD™ pen (Becton Dickinson, Franklin Lakes, NJ), the OPTIPEN™, the OPTIPEN PRO™, the OPTIPEN IV ... Examples of disposable pen delivery devices that have application in the subcutaneous delivery of pharmaceutical compositions of the present invention include, but are not limited to, the SOLOSTAR™ pen (Sanofi-Aventis), FLEXPEN™ (Novo Nordisk), and KWIKPEN™ (Eli), to name just a few. Lilly), SURECLICK(TM) auto-injector (Amgen, Thousand Oaks, CA), PENLET™ (Haselmeier, Stuttgart, Germany), EPIPEN (Dey, LP), and HUMIRA™ pen (Abbott Labs, Abbott Park IL).

[0123] In certain circumstances, pharmaceutical compositions can be delivered in a sustained-release system. In one embodiment, a pump can be used (see Langer, supra; Sefton, 1987, CRC Crit. Ref. Biomed. Eng. 14:201). In another embodiment, a polymeric material can be used. See Medical Applications of Controlled Release, Langer and Wise (eds.), 1974, CRC Press, Boca Raton, Florida. In yet another embodiment, a sustained-release system can be placed in the vicinity of the target of the composition, thereby requiring only a fraction of the systemic dose (see, e.g., Goodson, 1984, in Medical Applications of Controlled Release, supra, vol. 2, pp. 115-138). Other sustained-release systems are discussed in the review by Langer, 1990, Science 249:1527-1533.

[0124] Injectable preparations may include dosage forms for intravenous injection, subcutaneous injection, intradermal injection, intramuscular injection, drip infusion, etc. These injectable preparations may be prepared by publicly known methods. For example, injectable preparations may be prepared by dissolving, suspending, or emulsifying the above-mentioned antibody or its salt in a sterile aqueous or oily medium conventionally used for injections. Aqueous media for injection include, for example, saline, glucose-containing isotonic solutions, and other adjuvants, which may be used in combination with suitable solubilizers such as alcohols (e.g., ethanol), polyalcohols (e.g., propylene glycol, polyethylene glycol), and nonionic surfactants [e.g., polysorbate 80, HCO-50 (polyoxyethylene (50 mole) adduct of hydrogenated castor oil)]. Oily media include, for example, sesame oil and soybean oil, which may be used in combination with solubilizers such as benzyl benzoate and benzyl alcohol. Injections prepared in this manner are preferably filled into appropriate ampoules.

[0125] Advantageously, the pharmaceutical compositions for oral or parenteral use described above are prepared into a suitable unit dose dosage form to accommodate the dose of the active ingredient. Such unit dose dosage forms include, for example, tablets, pills, capsules, injections (ampoules), suppositories, etc. The amount of antibody contained therein is generally about 5 to about 500 mg per unit dose dosage form, and particularly in the form of injection, the amount of antibody contained therein is preferably about 5 to about 100 mg, and for other dosage forms, about 10 to about 250 mg.

[0126] Therapeutic Uses of Antigen-Binding Molecules The present invention includes methods comprising administering to a subject in need thereof a therapeutic composition comprising an anti-CD28 antibody or a bispecific antigen-binding molecule that specifically binds to CD28 and a target antigen (e.g., PSMA). The therapeutic composition may comprise any of the antibodies or bispecific antigen-binding molecules disclosed herein and a pharmaceutically acceptable carrier or diluent. As used herein, the phrase "subject in need thereof" refers to a human or non-human animal that exhibits one or more symptoms or signs of cancer (e.g., a subject that develops a tumor or a subject that is afflicted with any of the cancers described herein below), or who would otherwise benefit from inhibiting or reducing PSMA activity or depleting PSMA+ cells.

[0127] The antibodies and bispecific antigen-binding molecules of the present invention (and therapeutic compositions comprising same) In particular, they are useful for the treatment of any disease or disorder in which stimulating, activating, and / or targeting immune responses is beneficial. In particular, the anti-CD28 / anti-PSMA bispecific antigen binding molecules of the present invention can be used to treat, prevent, and / or ameliorate any disease or disorder associated with or mediated by PSMA expression or activity, or the proliferation of PSMA+ cells. The mechanism of action achieved by the therapeutic method of the present invention involves the killing of PSMA-expressing cells in the presence of effector cells, such as T cells. PSMA-expressing cells that can be inhibited or killed using the bispecific antigen binding molecules of the present invention include, for example, tumorigenic prostate cells.

[0128] The antigen-binding molecules of the present invention can be used to treat primary and / or metastatic tumors arising from, for example, subtypes of colon, lung, breast, renal, and bladder cancer. According to certain exemplary embodiments, the bispecific antigen-binding molecules of the present invention are used to treat prostate cancer.

[0129] The present invention also includes methods for treating residual cancer in a subject. As used herein, the term "residual cancer" refers to the presence or persistence of one or more cancerous cells in a subject after treatment with an anti-cancer therapy.

[0130] According to certain embodiments, the present invention provides methods for treating a disease or disorder associated with PSMA expression (e.g., prostate cancer), comprising administering one or more of the bispecific antigen-binding molecules described elsewhere herein to a subject after the subject has been shown to be non-responsive to other types of anti-cancer therapy. For example, the present invention includes methods for treating prostate cancer, comprising administering an anti-CD28 / anti-PSMA bispecific antigen-binding molecule to a patient 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 has received standard treatment for patients with cancer, e.g., prostate cancer. In other embodiments, a bispecific antigen-binding molecule of the present invention comprising an IgG4 Fc domain (anti-CD28 / anti-PSMA bispecific antigen-binding molecule) is initially administered to a subject at one or more time points (e.g., to provide robust initial depletion of prostate cancer cells), followed by administration of an equivalent bispecific antigen-binding molecule comprising a different IgG domain, such as an IgG1 Fc domain, at a subsequent time point. It is contemplated that the anti-CD28 / anti-PSMA antibody of the present invention can be used in conjunction with other bispecific antigen-binding molecules, such as anti-PSMA / anti-CD3 bispecific antibodies.It is also contemplated that the bispecific antibody of the present invention can be used in conjunction with checkpoint inhibitors, such as those targeting PD-1 and CTLA-4, and other targets.It may be advantageous to combine two bispecific antibodies that target the same tumor antigen (e.g., PSMA), but one of the bispecific antibodies targets CD3 on T cells, and the other bispecific antibody targets a costimulatory molecule such as CD28.This combination can be used alone to enhance tumor cell killing, or can be used in combination with checkpoint inhibitors.

[0131] Combination Therapies and Formulations The present invention includes compositions and therapeutic formulations comprising any of the exemplary antibodies and bispecific antigen-binding molecules described herein in combination with one or more additional therapeutically active ingredients, and methods of treatment comprising administering such combinations to a subject in need thereof.

[0132] Exemplary additional therapeutic agents that can be combined with or administered in combination with the antigen binding molecules of the present invention include, for example, chemotherapy, radiation therapy, checkpoint inhibitors that target PD-1 (e.g., anti-PD-1 antibodies such as pembrolizumab or nivolumab, see also US 9,987,500), CTLA-4, LAG3, TIM These include costimulatory agonist bivalent antibodies targeting molecules such as GITR, OX40, 4-1BB, CD3x bispecific antibodies (see, e.g., WO2017 / 053856A1, WO2014 / 047231A1, WO2018 / 067331A1, and WO2018 / 058001A1), other antibodies targeting PSMAxCD3 (see, e.g., WO2017 / 023761A1), and other costimulatory CD28x bispecific antibodies.

[0133] Other agents that may be beneficially administered in combination with the antibodies of the invention include, for example, tamoxifen, aromatase inhibitors, and 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, and IL-18, or their respective receptors. Pharmaceutical compositions of the invention (e.g., pharmaceutical compositions comprising the anti-CD28 / anti-PSMA bispecific antigen binding molecules disclosed herein) also include other agents, such as "ICE": ifosfamide (e.g., Ifex®), carboplatin (e.g., Paraplatin®), etoposide (e.g., Etopophos®, Toposar®, VePesid®, VP-16), "DHAP": dexamethasone (e.g., Decadron®), cytarabine (e.g., Cytosar-U), and others. (R), cytosine arabinoside, ara-C), cisplatin (e.g., Platinol®-AQ), and "ESHAP": etoposide (e.g., Etopofos®, Toposar®, VePesid®, VP-16), methylprednisolone (e.g., Medrol®), high-dose cytarabine, cisplatin (e.g., Platinol®-AQ).

[0134] The present invention also includes therapeutic combinations comprising any of the antigen binding molecules referred to herein and one or more inhibitors of VEGF, Ang2, DLL4, EGFR, ErbB2, ErbB3, ErbB4, EGFRvIII, cMet, IGF1R, B-raf, PDGFR-o, PDGFR-I3, FOLH1, PRLR, STEAP1, STEAP2, TMPRSS2, MSLN, CA9, uroplakin, or any of the foregoing cytokines, wherein the inhibitor is an aptamer, antisense molecule, ribozyme, siRNA, peptibody, nanobody, or antibody fragment (e.g., Fab fragment; F(ab')2 fragment; Fd fragment; Fv fragment; sc fragment). Fv; dAb fragments; or other engineered molecules such as diabodies, triabodies, tetrabodies, minibodies, and minimal recognition units). The antigen-binding molecules of the present invention can also be administered in combination with and / or co-formulated with antiviral drugs, antibiotics, analgesics, corticosteroids, and / or NSAIDs. The antigen-binding molecules of the present invention can also be administered as part of a treatment regimen that includes radiation therapy and / or conventional chemotherapy, or treatment with biologic agents, including checkpoint inhibitors or other bispecific antibodies.

[0135] The present invention includes compositions and therapeutic formulations comprising any of the antigen-binding molecules described herein in combination with one or more chemotherapeutic agents. Examples of chemotherapeutic agents include alkylating agents such as thiotepa and cyclophosphamide (Cytoxan™); alkylsulfonates such as busulfan, improsulfan, and piposulfan; aziridines such as benzodopa, carboquone, meturedopa, and uredopa; ethylenimines and methylameramines, including altretamine, triethylenemelamine, triethylenephosphoramide, triethylenethiophosphaolamide, and trimethylolmelamine; nitrates such as chlorambucil, chlornaphazine, chlorophosphamide, estramustine, ifosfamide, mechlorethamine, mechlorethamine oxide hydrochloride, melphalan, nobembine, phenesterine, prednimustine, trofosfamide, and uracil mustard. nitrosoureas such as carmustine, chlorozotocin, fotemustine, lomustine, nimustine, and ranimustine; aclacinomycin, actinomycin, autramycin, azaserine, bleomycin, cactinomycin, calicheamicin, carabicin, carminomycin, carzinophilin, chromomycin, dactinomycin, daunorubicin, detorubicin, 6-diazo-5-oxo-L-norleucine, doxorubicin, epirubicin, esorubicin, idarubicin, marcellomycin, mitomycin, mycobacterial Antibiotics such as cophenolic acid, nogalamycin, olivomycin, peplomycin, potfilomycin, puromycin, chelamycin, rhodrubicin, streptonigrin, streptozocin, tubercidin, ubenimex, zinostatin, and zorubicin; antimetabolites such as methotrexate and 5-fluorouracil (5-FU); folic acid analogs such as denopterin, methotrexate, pteropterin, and trimetrexate; purine analogs such as fludarabine, 6-mercaptopurine, thiamiprine, and thioguanine; ancitabine and azacitidine Pyrimidine analogues such as 6-azauridine, carmofluridine, cytarabine, dideoxyuridine, doxifluridine, enocitabine, and floxuridine; androgens such as calcesterone, dromostanolone propionate, epitiostanol, mepitiostane, and testolactone; antiadrenal agents such as aminoglutethimide, mitotane, and trilostane; folic acid supplements such as floric acid; aceglatone; aldophosphamide glycosides; aminolevulinic acid; amsacrine; bestravcil; bisantrene; edatraxate; defofamine; de Mecolcine; Diaziquone; Eflornithine; Elliptinium acetate; Etoglucide; Gallium nitrate; Hydroxyurea; Lentinan; Lonidamine; Mitoguazone; Mitoxantrone; Mopidamol; Nitracrine; Pentostatin; Fenamet; Pirarubicin; Podophyllic acid; 2-Ethylhydrazide; Procarbazine; PSK (trademark); Razoxane; Sizofiran; Spirogermanium; Tenuazonic acid; Triaziquone; 2,2´,2´´-Trichlorotriethylamine; Urethane; Vindesine; Dacarbazine; Mannomustine; Mitobronitol;Mitolactol; pipobroman; gacytosine; arabinoside ("Ara-C"); cyclophosphamide; thiotepa; taxanes, such as paclitaxel (Taxol™, Bristol-Myers Squibb Oncology, Princeton, NJ) and docetaxel (Taxotere™; Aventis Antony, France); chlorambucil; gemcitabine; 6-thioguanine; mercaptopurine; methotrexate; platinum analogs such as cisplatin and carboplatin; vinblastine; platinum; etoposide (VP-16); ifosfamide; mitomycin C; mitoxantrone; vincristine; vinorelbine; navelbine; novantrone; teniposide; daunomycin; aminopterin; xeloda; ibandronate; CPT-11; the topoisomerase inhibitor RFS2000; difluoromethylornithine (DMFO); retinoic acid; esperamicin; capecitabine; and pharmaceutically acceptable salts, acids, or derivatives of any of the above. Also included within this definition are antihormonal agents that act to regulate or inhibit hormone action on tumors, such as antiestrogens, including, for example, tamoxifen, raloxifene, aromatase-inhibiting 4(5)-imidazoles, 4-hydroxytamoxifen, trioxifene, keoxifene, LY117018, onapristone, and toremifene (Fareston); antiandrogens, such as flutamide, nilutamide, bicalutamide, leuprolide, and goserelin; and pharmaceutically acceptable salts, acids, or derivatives of any of the above.

[0136] The additional therapeutically active ingredient(s) may be administered immediately before, simultaneously with, or immediately after administration of the antigen-binding molecule of the present invention. (For purposes of this disclosure, such administration regimens will be considered to be administration of the antigen-binding molecule "in combination" with the additional therapeutically active ingredient.

[0137] The present invention includes pharmaceutical compositions in which the antigen-binding molecules of the present invention are co-formulated with one or more of the additional therapeutically active ingredient(s) described elsewhere herein.

[0138] Dosing regimen According to certain embodiments of the present invention, multiple doses of an antigen-binding molecule (e.g., an anti-CD28 antibody or a bispecific antigen-binding molecule that specifically binds to PSMA and CD28) can be administered to a subject over a predetermined period of time. The method according to this aspect of the present invention comprises sequentially administering multiple doses of the antigen-binding molecule of the present invention to the subject. As used herein, "sequentially administering" means that each dose of the antigen-binding molecule is administered to the subject at different time points, for example, on different days separated by a predetermined interval (e.g., several hours, several days, weeks, or months). The present invention includes methods comprising sequentially administering to a patient a single initial dose of the antigen-binding molecule, followed by one or more secondary doses of the antigen-binding molecule, and then optionally one or more tertiary doses of the antigen-binding molecule.

[0139] The terms "primary dose," "secondary dose," and "tertiary dose" refer to the time sequence of administration of an antigen-binding molecule of the present invention. Thus, a "primary dose" is a dose administered at the beginning of a treatment regimen (also referred to as a "baseline dose"), a "secondary dose" is a dose administered after the primary dose, and a "tertiary dose" is a dose administered after the secondary dose. The primary, secondary, and tertiary doses may all contain the same amount of antigen-binding molecule, but generally may differ from each other in terms of administration frequency. However, in certain embodiments, the amount of antigen-binding molecule contained in the primary, secondary, and / or tertiary dose differs from each other (e.g., adjusted up or down as appropriate) during the course of treatment. In certain embodiments, two or more doses (e.g., two, three, four, or five) are administered as a "loading dose" at the beginning of a treatment regimen, followed by subsequent doses (e.g., "maintenance doses") administered at a lower frequency.

[0140] In certain exemplary embodiments of the invention, each secondary and / or tertiary dose is 1 to 26 (e.g., 1, 1 1 / 2, 2, 2 1 / 2, 3, 3 1 / 2, 4, 4 1 / 2, 5, 5 1 / 2, 6, 6 1 / 2, 7, 7 1 / 2, 8, 8 1 / 2, 9, 9 1 / 2, 10, 10 1 / 2, 11, 11 1 / 2, 12, 12 1 / 2, 13, The phrase "immediately preceding dose" as used herein refers to a dose of an antigen-binding molecule in a multiple administration series that is administered to a patient prior to administration of the immediately following dose with no intervening doses.

[0141] The method according to this aspect of the invention may include administering any number of secondary and / or tertiary doses of an antigen-binding molecule (e.g., an anti-CD28 antibody or bispecific antigen-binding molecule that specifically binds PSMA and CD28) to the patient. For example, in certain embodiments, only a single secondary dose is administered to the patient. In other embodiments, two or more (e.g., two, three, four, five, six, seven, eight, or more) secondary doses are administered to the patient. Similarly, in certain embodiments, only a single tertiary dose is administered to the patient. In other embodiments, two or more (e.g., two, three, four, five, six, seven, eight, or more) tertiary doses are administered to the patient.

[0142] In embodiments involving multiple secondary doses, each secondary dose may be administered at the same frequency as the other secondary doses. For example, each secondary dose may be administered to the patient 1-2 weeks after the immediately preceding dose. Similarly, in embodiments involving multiple tertiary doses, each tertiary dose may 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 the secondary and / or tertiary doses are administered to the patient may vary over the course of the treatment regimen. The frequency of administration may also be determined based on the needs of the individual patient following clinical testing. Depending on the patient's condition, the dosage may be adjusted by the physician during the course of treatment.

[0143] Diagnostic Uses of Antibodies The bispecific antibodies of the present invention can also be used, for example, for diagnostic purposes, to detect and / or measure CD28 or PSMA or CD28- or PSMA-expressing cells in a sample. For example, an anti-CD28 antibody x PSMA antibody, or a fragment thereof, may be used to diagnose a condition or disease characterized by abnormal expression (e.g., overexpression, underexpression, lack of expression, etc.) of CD28 or PSMA. An exemplary diagnostic assay for CD28 or PSMA may, for example, involve contacting a sample obtained from a patient with an antibody of the present invention, wherein the antibody is labeled with a detectable label or reporter molecule. Alternatively, an unlabeled antibody can be used in diagnostic applications in combination with a secondary antibody that is itself detectably labeled. The detectable label or reporter molecule may be 3 H, 14 C. 32 P, 35 S, or 125The CD28 or PSMA signaling molecule may be a radioisotope such as I, a fluorescent or chemiluminescent moiety such as fluorescein isothiocyanate or rhodamine, or an enzyme such as alkaline phosphatase, β-galactosidase, horseradish peroxidase, or luciferase. Specific exemplary assays that can be used to detect or measure CD28 or PSMA in a sample include enzyme-linked immunosorbent assay (ELISA), radioimmunoassay (RIA), and fluorescence-activated cell sorting (FACS). Samples that can be used in the CD28 or PSMA diagnostic assays of the present invention include any tissue or body fluid sample obtainable from a patient that contains a detectable amount of CD28 or PSMA protein or a fragment thereof under normal or pathological conditions. Generally, the level of CD28 or PSMA in a particular sample obtained from a healthy patient (e.g., a patient not suffering from a disease or condition associated with abnormal CD28 or PSMA levels or activity) is measured to first establish a baseline or standard level of CD28 or PSMA. This baseline level of CD28 or PSMA can then be compared to the level of CD28 or PSMA measured in a sample obtained from an individual suspected of having a disease or condition associated with CD28 or PSMA. [Example]

[0144] The following examples are presented so as to provide those of ordinary skill in the art with a complete disclosure and description of how to make and use the methods and compositions of the present invention, and are not intended to limit the scope of what the inventors regard as the invention. Efforts have been made to ensure accuracy with respect to numbers used (e.g., amounts, temperatures, etc.), but some experimental error and deviation should be accounted for. Unless otherwise indicated, parts are parts by weight, molecular weights are average molecular weights, temperatures are in degrees Celsius, and pressures are at or near atmospheric.

[0145] summary The recent clinical success of tumor-specific antigen (TSA) or tumor-associated antigen (TAA)-CD3 bispecific antibodies (e.g., anti-PSMA × CD3 bispecific antibodies) is transforming cancer immunotherapy. TAA-CD3 antibodies offer a relatively safe and readily available therapeutic solution and may prove to be an important new class of immunotherapy. However, they may not achieve the efficacy seen with CAR-T cell approaches that involve a second costimulatory signal, often mediated by the CD28 intracellular domain. Indeed, CD28-activating antibodies, while incredibly potent, are also toxic in human studies. The terms TSA and TAA may be used interchangeably herein.

[0146] Herein, we report a series of TAA-CD3-mediated chemoattractants (CTAs) that appear to be safe and well tolerated on their own, and that, when combined with TAA-CD3, generate a significantly amplified and highly targeted antitumor response. A new class of bispecific antibody, termed D28, is described. Using a TAA such as PSMA expressed on prostate tumors, TAA×CD28 is shown to enhance T cell activation and significantly increase tumor cell lysis in the presence of tumor targeting and TCR / CD3 stimulation.

[0147] The TAAxCD28 bispecific antibody of the present invention has been shown to significantly enhance antitumor immunity mediated by TAAxCD3-induced T cell activation in syngeneic tumor models. When paired with tumor antigens with restricted expression in normal tissues, CD3 induced little or no peripheral cytokine production alone or in combination with TAAxCD28 in non-human primate and humanized mouse models. Furthermore, TAAxCD28 alone induced minimal or no serum cytokine secretion and T cell activation in cynomolgus monkeys compared with a CD28 superagonist antibody. These results suggest that this novel class of costimulatory bispecific antibody, when combined with a new class of TAA-CD3, may provide a safer, off-the-shelf biologic solution that can significantly enhance the efficacy of conventional TAA-CD3 bispecific antibodies.

[0148] Introduction Monoclonal antibodies have been established as antitumor therapeutics over the past two decades, but their ability to recruit T cells and efficiently exert cytotoxic activity at tumor sites remains limited. The ability of bispecific antibodies for T cell-targeted immunotherapy and subsequent tumor cell killing has previously been demonstrated. Indeed, bispecific antibody platforms aim to recruit immune effector cells by combining anti-CD3 and antitumor target-binding domains. Recently, catumaxomab (EpCAM × CD3) and blinatumomab (CD19 × CD3) received regulatory approval for acute lymphoblastic leukemia, and many other bispecific antibodies are under investigation. Indeed, bispecific antibodies that recognize both the B cell marker CD20 and the CD3 component of the T cell repertoire have been constructed and are currently undergoing clinical investigation for hematological malignancies. Indeed, TAA-CD3 offers a relatively safe, off-the-shelf therapeutic solution that does not require advanced and laborious customization for individual patients. However, while TAA-CD3 bispecific antibodies may prove to be an important new class of immunotherapy, cross-sectional comparisons suggest that they may not achieve the efficacy seen with CAR-T approaches.

[0149] Two chimeric antigen receptor (CAR)-T cell drugs, Kymriiah and Yescarta, recently received FDA approval for hematologic B-cell malignancies, demonstrating the strong potential of this approach in the field of personalized cancer immunotherapy. Both products utilize the CD19 antigen as a tumor-associated antigen (TAA), which, due to its restricted expression on B cells, is an ideal target for T-cell-mediated killing, minimizing off-target toxicity and enhancing antitumor efficacy. However, the high potency of CAR-T cells is associated with side effects such as cytokine release syndrome (CRS) and neurotoxicity. Furthermore, not all patients have yet benefited from treatment, and the number of variables that can affect each patient's clinical outcome is relatively high in both autologous and allogeneic approaches. Furthermore, attempts to target tumor-associated antigens in solid tumors have so far achieved limited success, demonstrating either minimal antitumor activity or severe side effects. Inhibition of the tumor microenvironment, poor access of CAR-T cells throughout tumor tissue, and laborious manufacturing represent some of the current challenges to promising CAR-T cell therapy.

[0150] One of the limitations of current immunotherapeutic treatments is optimally inducing a patient's own immune response against tumor cells through the recognition of specific tumor cells and the induction of cytotoxicity. Effective activation of naive T cells and induction of experienced memory T cell populations requires the delivery of costimulatory signals (signal 2) in addition to antigen-specific stimulation via the TCR / CD3 complex (signal 1). ) are required. Agonism of the costimulatory pathways CD28 and 4-1BB can result in significant enhancement of target cell lysis, which may aid in patient resistance to different immunotherapies. However, the broader role of costimulation is still being evaluated in clinical oncology. Many tumors lack costimulatory receptors, preventing them from further enhancing T cell activation induced by CD3×TAA.

[0151] A blinded screening of costimulatory pathways performed herein by forced expression of costimulatory ligands on a panel of syngeneic tumors established CD28, along with 4-1BB, as one of the most potent costimulatory receptors. Furthermore, a costimulatory-based bispecific antibody that crosslinks CD28 on the surface of T cells was constructed herein using a TAA from prostate tissue to enhance antitumor responses. Disclosed herein are data demonstrating that combination therapy with TAA-CD3 and TAA-CD28 bispecific antibodies efficiently enhances T cell activation and cytotoxicity in the presence of both TAA and TCR stimulation, resulting in enhanced antitumor immunity. Indeed, the data disclosed herein suggest that combining this novel class of bispecific antibody (TAA-CD28) with the emerging TAA-CD3 class may provide a safer, off-the-shelf biologic solution that may approach the efficacy of customized CAR-T cell therapy.

[0152] More specifically, an exemplary PSMAxCD28 bispecific antibody was generated, and the exemplary anti-PSMAxCD28 antibody was demonstrated to enhance PSMAxCD3- or CD20xCD3-induced T cell activation in vitro and safely enhance antitumor efficacy in vivo. In vitro activity was demonstrated by imaging the bispecific antibody localizing to the immune synapse of T cell and target cell complexes, enhancing PSMAxCD3-induced proliferation, cytokine release, and cytotoxicity. In vivo antitumor efficacy was assessed in a syngeneic mouse tumor model. Tumor volume and serum cytokines were monitored over time to indicate response to bispecific antibody treatment. Studies were conducted in cynomolgus monkeys to determine the safety and tolerability of the exemplary PSMAxCD28 antibody of the present invention in non-human primates. Animals were examined for toxicity by clinical observation and blood sample collection to analyze serum cytokines and T cell phenotypes.

[0153] As described in detail below, an exemplary anti-PSMAxCD28 bispecific antibody was generated to enhance TCR / CD3-dependent T cell activation, thus mimicking costimulation (signal 2) provided by professional APCs.

[0154] All procedures were performed in accordance with the National Institutes of Health (NIH) Guide for the Care and Use of Laboratory Animals. The protocol was approved by the Regeneron Pharmaceuticals Animal Care and Use Committee.

[0155] Example 1. Construction of anti-PSMA x CD28 antibody Generation of anti-CD28 antibodies Anti-CD28 antibodies were obtained by immunizing VELOCIMMUNE® mice (i.e., genetically engineered mice containing DNA encoding human immunoglobulin heavy chain and kappa light chain variable regions) with human CD28 fused to the Fc portion of mouse IgG2a, or with cells expressing CD28 or DNA encoding CD28. Antibody immune responses were monitored by CD28-specific immunoassays. When the desired immune response was achieved, splenocytes were harvested and fused with mouse myeloma cells to maintain their viability and form hybridoma cell lines. The hybridoma cell lines were screened and selected to identify cell lines producing CD28-specific antibodies. Using this technique, several anti-CD28 chimeric antibodies (i.e., antibodies with human variable domains and mouse constant domains) were obtained. Furthermore, as described in US2007 / 0280945A1, several This fully human anti-CD28 antibody was isolated directly from antigen-positive B cells without fusion to myeloma cells.

[0156] Certain biological properties of exemplary anti-CD28 antibodies generated according to the methods of this example are described in detail in the Examples set forth below.

[0157] Generation of anti-PSMA antibodies Anti-PSMA antibodies were obtained by immunizing genetically modified mice with human PSMA antigen or by immunizing genetically engineered mice containing DNA encoding human immunoglobulin heavy and kappa light chain variable regions with human PSMA antigen. Alternatively, mice were immunized with human prostate cancer cells (LNCaP, ATTC, Manassas, Virginia, USA) expressing human PSMA (UniProtKB / Swiss-Prot. No. Q04609). Antibody immune responses were monitored by PSMA-specific immunoassays. Once the desired immune response was achieved, splenocytes from each mouse were harvested and either (1) fused with mouse myeloma cells to maintain their viability and form hybridoma cells, which were then screened for PSMA specificity, or (2) B cells were selected using human PSMA (R&D, catalog number 4234-ZN) with an N-terminal 6-His tag as a selection reagent to bind and identify reactive antibodies (antigen-positive B cells) (described in US 2007 / 0280945A1). Chimeric antibodies against PSMA with human variable regions and mouse constant regions were first isolated. The antibodies were characterized and selected for desired characteristics, including affinity, selectivity, etc. If necessary, the mouse constant region was replaced with a desired human constant region, such as wild-type or modified IgG1 or IgG4, to generate fully human anti-PSMA antibodies.

[0158] Generation of a bispecific antibody that binds CD28 and PSMA Bispecific antibodies comprising an anti-PSMA-specific binding domain and an anti-CD28-specific binding domain were constructed using standard methods, with the anti-PSMA antigen-binding domain and the anti-CD28 antigen-binding domain each comprising a distinct HCVR paired with a common LCVR. In some cases, bispecific antibodies were constructed utilizing a heavy chain from an anti-CD28 antibody, a heavy chain from an anti-PSMA antibody, and a common light chain (see Table 1). In certain embodiments, the heavy chain amino acid sequence from the anti-CD28 antibody of the exemplary bispecific antibody (bs16429D) is set forth in SEQ ID NO: 81. The heavy chain amino acid sequence from the anti-PSMA antibody of the exemplary bispecific antibody (bs16429D) is set forth in SEQ ID NO: 82. The common light chain amino acid sequence of bs16429D is set forth in SEQ ID NO: 83.

[0159] The bispecific antibody generated according to this example contains two distinct antigen-binding domains (i.e., binding arms). The first antigen-binding domain contains a heavy chain variable region derived from an anti-CD28 antibody ("CD28-VH"), and the second antigen-binding domain contains a heavy chain variable region derived from an anti-PSMA antibody ("PSMA-VH"). Both anti-PSMA and anti-CD28 share a common light chain. The CD28-VH / PSMA-VH pairing creates an antigen-binding domain that specifically recognizes CD28 on T cells and PSMA on tumor cells.

[0160] Example 2. Amino acid and nucleic acid sequences of heavy and light chain variable regions Table 1 shows the amino acid sequence identifiers for the heavy and light chain variable regions and CDRs of various bispecific antibodies produced according to Example 1. The corresponding nucleic acid sequence identifiers are shown in Table 2. [Table 1] [Table 2]

[0161] Example 3. CD28 and 4-1BB are potent costimulatory receptors To determine which costimulatory receptors are effective in providing the costimulatory signal critical for T cell activation, a blind screen of costimulatory pathways performed by forced expression of costimulatory ligands on a panel of syngeneic tumors (Table 3 and Figure 2) again established CD28, along with 4-1BB, as one of the most potent costimulatory receptors. Table 3 summarizes the number of tumor-free mice in the blind screen. Five mice were tested in each group. The assay was performed with three different tumor cell lines engineered to express seven different costimulatory ligands. Briefly, EL4, MC38, and B16F10.9 tumor cells were engineered to express individual costimulatory ligands by lentiviral transduction. The cells were implanted subcutaneously into WT C57BL6 mice. Tumor growth was measured 18, 24, and 25 days after implantation for EL4, MC38, and B16F10.9 tumor types, respectively. This is the time point when tumor volumes in the control group reached the maximum allowable size (>2000 mm3). Tumor size was measured, and consistent with the data in Table 3, CD28 and 4-1BB were also most effective in reducing tumor size (data not shown). [Table 3]

[0162] Example 4. Surface Plasmon Resonance-Derived Binding Affinity and Rate Constants of Anti-PSMA x CD28 Bispecific Antibodies To determine the binding kinetics of anti-PSMAxCD28 bispecific monoclonal antibodies, surface plasmon resonance-derived binding affinities and rate constants of anti-PSMAxCD28 bispecific antibodies and relevant parent monoclonal antibodies against PSMA and / or CD28 were determined.

[0163] Binding kinetics of anti-PSMA x CD28 bispecific monoclonal antibody to PSMA The equilibrium dissociation constant (K) of 6h.hPSMA (recombinant human PSMA / FOLH1 protein, R&D, Cat. No. 4234-ZN) binding to purified anti-PSMA x CD28 bispecific monoclonal antibody or anti-PSMA bivalent parent monoclonal antibody.D value) is Bi The determination was performed using a real-time surface plasmon resonance biosensor on a Biacore T-200 instrument. The CM5 Biacore sensor surface was derivatized by amine coupling with a monoclonal mouse anti-human Fc antibody and used to capture purified anti-PSMA × CD28 bispecific or anti-PSMA and anti-CD28 parental monoclonal antibodies.

[0164] Biacore binding studies were performed in a buffer consisting of 0.01 M HEPES (pH 7.4), 0.15 M NaCl, 0.5 mM MgCl2, 1.0 mM CaCl2, and 0.05% v / v surfactant P20 (HBS-P++ running buffer). Different concentrations of hPSMA with an N-terminal polyhistidine tag (6h.hPSMA, R&D) were prepared in HBS-P++ running buffer in three-fold serial dilutions ranging from 10 nM to 0.4 nM for the anti-PSMA x CD28 bispecific antibody and the anti-PSMA or anti-CD28 parent monoclonal antibodies.

[0165] Different concentrations of 6h.hPSMA were injected over the monoclonal antibody-captured surface at a flow rate of 50 μL / min. The association of 6h.hPSMA to the captured monoclonal antibody was monitored for 3 min, and the dissociation of 6h.hPSMA in HBS-P++ running buffer was monitored for 10 min. Kinetic binding (k) was calculated by fitting the real-time sensorgrams to a 1:1 binding model using Scrubber 2.0c curve-fitting software (BioLogic Software). a ) rate constant and kinetic dissociation (k d The binding-dissociation equilibrium constant (K D ) and dissociation half-life (t1 / 2) were calculated from the kinetic rate constants as follows: Calculated: K D (M)=k d / k a , and t1 / 2(min)=0.693 / k d / 60

[0166] The binding kinetic parameters of 6h.hPSMA binding to purified monoclonal antibodies at 25°C are shown in Table 4 below. [Table 4]

[0167] The binding kinetic parameters of 6h.hPSMA binding to one purified exemplary monoclonal bispecific antibody at 37°C are shown below in Table 5. As defined by the manufacturer, one RU (response unit) is equal to 1 mm 2 Each represents 1 pg of protein. [Table 5]

[0168] Binding kinetics of anti-PSMA x CD28 bispecific monoclonal antibody to CD28 The equilibrium dissociation constant (K) of hCD28.mmh binding to purified monoclonal antibody D value ) was determined using a real-time surface plasmon resonance biosensor with a Biacore T-200 instrument. The CM5 Biacore sensor surface was derivatized by amine coupling with a monoclonal mouse anti-human Fc antibody to capture purified anti-PSMA x CD28 bispecific monoclonal antibody and anti-PSMA or anti-CD28 parental monoclonal antibodies.

[0169] Different concentrations of hCD28.mmh were injected over the monoclonal antibody-captured surface at a flow rate of 50 μL / min. Association of hCD28.mmh to the captured monoclonal antibody was monitored for 5 min, and dissociation of hCD28.mmh in HBS-P++ running buffer was monitored for 10 min. Kinetic binding (k) was determined by fitting the real-time sensorgrams to a 1:1 binding model using Scrubber 2.0c curve-fitting software. a ) Speed constant number and dynamic dissociation (k dThe binding-dissociation equilibrium constant (K D ) and dissociation half-life (t1 / 2) were calculated from the kinetic rate constants as follows: K D (M)=k d / k a , and t1 / 2(min)=0.693 / k d / 60

[0170] The binding kinetic parameters for hCD28.mmh binding to purified anti-PSMAxCD28 bispecific monoclonal antibodies or anti-CD28 bivalent parental monoclonal antibodies at 25°C are shown in Table 6 below. [Table 6]

[0171] The binding kinetic parameters for hCD28.mmh binding to purified anti-PSMAxCD28 bispecific monoclonal antibodies at 37°C are shown in Table 7 below. [Table 7]

[0172] As shown in Tables 4-7, some anti-CD28 antibodies of the invention bind to CD28 with high affinity. Some anti-PSMA antibodies of the invention bind to PSMA with high affinity. Some anti-PSMA×CD28 bispecific antibodies bind to both CD28 and PSMA with high affinity.

[0173] Example 5. Cell-binding potency and specificity of anti-PSMA x CD28 bispecific monoclonal antibodies to PSMA and CD28 To evaluate the ability of these antibodies (anti-PSMA, anti-CD28, and anti-PSMAxD28 monoclonal antibodies) to specifically bind to cell surface proteins, we developed an in vitro binding assay using cells expressing human PSMA or CD28 on an electrochemiluminescence-based detection platform (MSD). Two studies were performed. One study evaluated the potency and specificity of the monoclonal antibodies for binding to cell surface antigens. These antibodies (Anti-PSMA, anti-CD28, and anti-PSMAxCD28 antibodies) showed specific binding to cell lines expressing human PSMA or human CD28. In supplementary experiments, bridging studies to the IgG4s isotype were performed.

[0174] Methods used to determine the potency and specificity of cell binding of anti-PSMA x CD28 bispecific antibodies to PSMA and CD28 The human epithelial prostate cancer cell line, C4-2 (UroCor), endogenously expresses human PSMA. The HEK293 / hCD28-expressing cell line was engineered by transducing human embryonic kidney cells from ATCC (CRL-1573) with a neomycin-resistant lentiviral construct encoding human CD28 (hCD28 accession number NP_006130.1). To assess binding specificity, these two cell lines were evaluated by fluorescence-activated cell sorting (FACS) in parallel with the parental HEK293HZ cell line (negative for PSMA and CD28). An anti-Feld1 human IgG1 antibody was included as a negative control for IgG detection.

[0175] The above cell lines were treated with Ca 2+ / Mg 2+ (Irvine Scientific, catalog no. The cells were then rinsed once with 1x PBS buffer (free of Ca) and incubated with enzyme-free cell dissociation solution (Millipore, Cat. No. S-004-C) at 37°C for 10 minutes to detach the cells from the flask. 2+ / Mg 2+ (Irvine Scienti An additional wash with 1x PBS containing 1x BSA (Sigma, Cat. No. 9236) was performed. Cells were then counted using a Cellometer™ Auto T4 cell counter (Nexcelom Bioscience, Model #Auto T4). Approximately 10,000 cells per well in cell wash buffer were individually seeded onto a 96-well carbon electrode plate (MULTI-ARRAY High Binding Plate, MSD, Meso Scale Discovery, Cat. No. L15XB-3 / LX11XB-3) and incubated at 37°C for 1 hour to allow cell attachment. Nonspecific binding sites were blocked with 2% BSA (w / v) in PBS (Sigma, Cat. No. A8577-1L) for 1 hour at room temperature.

[0176] Solutions containing anti-PSMA, anti-CD28, anti-PSMA x CD28, or control antibodies in serial dilutions ranging from 1.7 pM to 100 nM, or no antibody, were added to the plate-bound cells and incubated for 1 hour at room temperature. Except where indicated, assays were performed in duplicate. The plates were then washed to remove unbound antibody using an AquaMax2000 plate washer equipped with a cell wash head (MDS Analytical Technologies, Model #2000). Plate-bound antibodies were detected with heavy- and light-chain-specific SULFO-TAG™-conjugated goat polyclonal anti-human IgG antibodies (Jackson ImmunoResearch, Catalog #109-005-088) for 1 hour at room temperature. After washing, the plates were developed using read buffer (MSD, Meso Scale Discovery, Catalog #R92TD-2) according to the manufacturer's recommended procedure, and the luminescent signal was analyzed using a SECTOR Imager 600 (Meso Recordings were made using a Scale Discovery, model #600) instrument.

[0177] Direct binding signals (relative light units, RLU) were analyzed as a function of antibody concentration. Data were fitted to a sigmoidal (four-parameter logistic) dose-response model using GraphPad Prism™ software (GraphPad software version #6). EC50 values, defined as the antibody concentration at which 50% of the maximum binding signal is detected, were determined for binding to HEK293 / hCD28 and C4-2 cells, indicating the potency of each antibody binding to CD28 or PSMA, respectively. Additionally, the ratio of the binding signal of 11.1 nM antibody on HEK293 / hCD28 or C4-2 cells to that on HEK293HZ cells was calculated. This representative concentration was chosen due to its high target cell binding and low background signal on CD28- and PSMA-negative HEK293HZ cells. Antibodies with binding ratios less than 3 were marked as NB in Table 8 below, which is defined as "no specific binding observed under the assay conditions."

[0178] In a supplemental study, a separate experiment was performed using the same protocol as above, but including an IgG4 isotype control (see USSN 15 / 147,791) and a parental CD28 monoclonal antibody. Test solutions contained serial dilutions of anti-PSMA, anti-CD28, anti-PSMA x CD28 bispecific monoclonal antibodies, or control antibodies from 3.4 pM to 200 nM. Binding ratios were calculated at 7.4 nM.

[0179] Results, Summary, and Conclusions The ability of the anti-PSMA x CD28 bispecific antibody to specifically bind to cells expressing either PSMA or CD28 was assessed using an immunobinding assay in comparison with cell lines negative for PSMA or CD28 expression. For comparison, a bivalent parent antibody specific for PSMA or CD28 was included. Dose-dependent binding of the antibody to cells on a 96-well high-binding plate (MSD, Meso Scale Discovery, Cat. No. L15XB-3 / LX11XB-3) at antibody concentrations up to 100 nM was detected using a SULFO-TAG™-conjugated anti-human IgG antibody, and electrochemiluminescence binding signals were recorded on a Sector Imager 600 (MSD). The RLU values of antibody binding to cells were determined. For CD28- or PSMA-expressing cells, EC 50 Values were calculated as a measure of potency. For samples that produced high background signals, the high concentration was excluded from the EC50 value calculation, and the values are indicated with an asterisk in Table 8. Comparison of the binding signal between antibody and HEK293 / hCD28, or C4-2 cells and negative HEK293HZ cells at 11.1 nM, was used to assess antibody binding specificity. Specific binding is defined as an antibody with a 3-fold or greater ratio of binding to CD28- or PSMA-expressing cells compared to HEK293HZ cells at that concentration.

[0180] The binding results are summarized in Table 8. At a concentration of 11.1 nM, three exemplary anti-PSMA x CD28 bispecific antibodies of the invention (bs16429D, bs16430D, and bs16431D) specifically bound to both HEK293-hCD28 and C4-2 cells at ratios 13-31-fold and 4-10-fold higher than HEK293HZ cells, respectively. The potencies of the bispecific antibodies were EC50s of 5.31-9.58 nM for HEK293-hCD28 cells and 0.35-5.24 nM for C4-2 cells. 50 The values were in the range. [Table 8]

[0181] The parental antibodies specifically bound to the cells corresponding to the antigen used to generate them, as indicated in parentheses in the Antibody ID column. EC values for binding to HEK293-hCD28 and C4-2 cells were 50 The values ranged from 1.04–10.2 nM on HEK293-hCD28 cells to 0.738–0.946 nM on C4-2 cells. For antibodies, mAb11838P2 and mAb11810P2 binding values at higher concentrations on C4-2 cells were calculated using the EC 50 It was excluded from the calculation of values. As expected, the IgG control antibody did not specifically bind to CD28 or PSMA-expressing cell lines. Supplementary experiments were performed as described above, but included an IgG4 isotype control and a parent CD28 monoclonal antibody. This data was generated as a bridging study to demonstrate that there was little background signal using the hIgG4s isotype control. The data are summarized in Table 9. As shown in Table 9, the anti-CD28 antibody mAb14226P2 specifically binds to HEK293-hCD28P-3, a human CD28-expressing cell line. The anti-CD28 antibody mAb14226P2 does not bind to cells that do not express human CD28 (HEK293HZ), regardless of whether the cells express PSMA (C4-2). [Table 9]

[0182] Binding of PSMA×CD28 bispecific antibodies to T cells and target cells Experimental procedure Flow cytometry analysis was used to determine the binding of anti-PSMA x CD28 bispecific antibodies to C4-2, 22RV1, RAJI, human, and cynomolgus monkey T cells, followed by detection with phycoerythrin (PE)-labeled anti-human IgG antibody. Briefly, 1 × 10 5 Cells / well were incubated with either PSMAxCD28 bispecific antibody or human IgG4 antibody (USS The cells were incubated for 30 minutes at 4°C with serial dilutions of IgG1 (see N15 / 147,791), which binds to the human antigen without cross-reactivity with human or cynomolgus CD28, ranging from 133 nM to 8.14 pM for human and cynomolgus T cells and from 133 nM to 61 pM for PSMA-expressing cells. After incubation, the cells were washed twice with cold PBS containing 1% filtered FBS, and a PE-conjugated anti-human secondary antibody (Jackson Immunoresearch, catalog no. 709-116-149) was added to the cells and incubated for an additional 30 minutes. Live / dead dye was added to the human and cynomolgus T cell incubations. Wells containing no antibody or only the secondary antibody served as controls.

[0183] After incubation with PSMA-expressing cells, 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.

[0184] After incubation with human or cynomolgus T cells, cells were washed and stained with a cocktail of anti-CD2 (BD, Catalog No. 562638), anti-CD16 (BD, Catalog No. 562874), anti-CD4 (BD, Catalog No. 564305), and anti-CD8 (BD, Catalog No. 563795) antibodies in Brilliant Staining Buffer (BD, Catalog No. 566349) for an additional 20-minute incubation at 4°C. After washing, cells were resuspended in 200 μL of cold PBS (Gibco, Catalog No. 14190-144) containing 1% filtered FBS (TCB, Catalog No. 101), gated into Live / CD2+ / CD4+ / CD16 or Live / CD2+ / CD8+ / CD16 gates, and analyzed by flow cytometry on a BD FACS LSR-Fortessa-X20.

[0185] Results, Summary, and Conclusions: The binding of the PSMAxCD28 bispecific antibody to the surface of human T cells was tested by flow cytometry.

[0186] bs16429D is 4.80 x 10 -8 M bound to all T cells with an EC50 value of 5.09 × 10 -8 M and 5.89 x 10 -8 It bound to both CD4+ and CD8+ T cells with an EC50 value of M.

[0187] bs16431D is 1.80 x 10 -7 It bound weakly to all T cells with an EC50 value of 1.67E-07M and 1.80E-07M, respectively, and to both CD4+ and CD8+ T cells.

[0188] The binding of the PSMAxCD28 bispecific antibody to the surface of PSMA-expressing cell lines was tested by flow cytometry.

[0189] C4-2 is a CaP (prostate cancer) cell line derived from LNCaP (androgen-sensitive human prostate adenocarcinoma cells derived from lymph node metastasis, see Wu et al., Int. J. Cancer, 57:406-412 (1994)) cells. Both bs16429D and bs16431D were 3.87 x 10 -9 M and 1.50 x 10 -9 It bound to C4-2 cells (see Liu et al., 2004, Prostate, 60:98-108) with an EC50 value of M.

[0190] 22RV1 is an epithelial prostate cancer cell line (see In Vitro Cell Dev. Biol. Anim., 1999, 35(7):403-409). Both bs16429D and bs16431D were 3.05 × 10 -9 M and 6.33 x 10 -09 It bound to 22RV1 cells with an EC50 value of M.

[0191] These results are summarized in Tables 10-12. [Table 10] [Table 11] [Table 12]

[0192] Example 6: Primary and engineered bioassays of PSMAxCD28 bispecific antibodies T cell activation is achieved by stimulating the T cell receptor (TCR), which recognizes specific peptides presented by major histocompatibility complex class I or II (MHC1 or MHCII) proteins on antigen-presenting cells (APCs) (Goldrath et al., Selecting and maintaining a diverse T-cell repertoire, Nature 402, 255-262 (1999)). Activated TCRs, in turn, initiate a cascade of signaling events that can be monitored by reporter genes driven by various transcription factors, such as activator protein 1 (AP-1), nuclear factor of activated T cells (NFAT), or nuclear factor kappa-light-chain-enhancer of activated B cells (NFκB). T cell responses are then further refined through the engagement of coreceptors expressed either constitutively or inducibly on T cells, such as CD28, CTLA-4 (cytotoxic T lymphocyte-associated protein 4), PD-1 (programmed cell death protein 1), LAG-3 (lymphocyte activation gene 3), or other molecules (Sharpe et al., The B7-CD28 Superfamily, Nat. Rev. Immunol., 2(2):116-26 (2002)). CD28, a costimulatory molecule, is activated by its endogenous ligands CD80 or CD86 expressed on APCs. Following TCR activation, CD28 enhances cellular signals, such as pathways controlled by the NFκB transcription factor. CD28 co-signaling is important for effective T cell activation, including T cell differentiation, proliferation, cytokine release, and cell death (Smeets et al., NFκB activation induced by T cell receptor / CD28 costimulation is mediated by protein kinase C-Θ, PNAS, 97(7):3394-3399 (2012)).

[0193] To identify antibodies that enhance T cell activity in the presence of both primary stimulation and PSMA target expression, anti-CD28 antibodies and anti-PSMA x CD28 bispecific antibodies were characterized in engineered reporter bioassays and cell-based assays using human primary T cells. The assays evaluate the behavior of anti-PSMA / CD28 bispecific antibodies in the presence and absence of primary stimulation and target expression. A schematic diagram of the assay is shown in Figure 3. The assay was performed to select anti-PSMA x CD28 bispecific antibodies that enhance T cell activity in the presence of primary stimulation and target expression. Therefore, the assays evaluated the behavior of bispecific antibodies in the presence and absence of primary stimulation and target expression.

[0194] 1) Luciferase-based reporter assay: a) Engineering reporter T cells: A Jurkat-derived T cell clone, JRT3.T3.5 (ATCC, #TIB-153), was transduced with an NFκB luciferase reporter construct (NFκB-Luc, SA Biosciences / Qiagen, catalog #CLS-013L). After isolation of a puromycin-resistant clone (JRT3.T3.5 / NFκB-Luc clone 1C2), full-length human TCRα (1G4A - amino acids M1 to S274) and TCRβ fragments were transduced. The cells were further engineered to express the CD8α subunit (hCD8a, amino acids M1 to V235, accession number NP_001139345) and the CD8β subunit (hCD8b, amino acids M1 to K210, accession number P10966) after isolation of a single clone (J.RT3-T3.5 / NFκB-Luc / 1G4AB clone 1D2). A single clone was again generated (J.RT3-T3.5 / NFκB-Luc / 1G4AB / hCD8ab clone 1D5) and further transduced with full-length human CD28 (hCD28-amino acids M1 to S220 from accession number #P10747). Cells were selected for high expression of CD28 and cultured in RPMI + 20% Engineered reporter T cells were maintained in FBS + penicillin / streptomycin / glutamine (P / S / G) + NEAA + NaPyr + 1 μg / mL puromycin + 500 μg / mL G418 + 250 μg / mL hygromycin + 10 μg / mL blasticidin. For faster growth, engineered reporter T cells were maintained in antibiotic-free cell culture medium and used for cell-based luciferase experiments. Reagent information is as follows: RPMI 1640, Irvine Scientific, catalog number 9160; FBS, Seradigm, catalog number 1500-50; penicillin / streptomycin / glutamine 100x (P / S / G), Thermo Fisher Scientific, catalog number 10378-016; non-essential amino acids (NEAA), Irvine Scientific, catalog number 9304; sodium pyruvate (NaPyr), Millipore, catalog number TMS-005-C; puromycin, Sigma, catalog number P8833; geneticin (G418), Thermo Fisher Scientific, catalog number 11811-098; hygromycin; blasticidin.

[0195] b) APC operation: A stable HEK293 cell line (ATCC, #CRL-1573) expressing human CD20 (amino acids M1 to P297 of accession number NP_068769.2) was transduced with human PSMA (amino acids M1 to A750 of accession number Q04609). Human PSMA-positive cells were isolated by fluorescence-activated cell sorting (FACS). The resulting cell line, HEK293 / CD20 / hPSMA, was maintained in DMEM + 10% + P / S / G + NEAA supplemented with 500 μg / mL G418.

[0196] c) T cell / APC stimulation: In this experiment, engineered reporter T cells are stimulated via two bispecific antibodies. The first stimulus is delivered by a T cell-activating bispecific antibody, anti-CD3 x CD20 hIgG4 (see WO 14 / 047231, US9657102, and USSN 14 / 661,334), which targets the CD3 molecule on the engineered reporter T cells and CD20 on HEK293 cells. Here, the first stimulus bypasses the need for TCR activation by natural ligands, specific peptides displayed on MHC molecules. The second stimulus is driven by a CD28 bispecific antibody. This antibody mimics CD28 activation in T cells by CD80 / CD86, a ligand expressed on APCs. Here, the antibody interacts with CD28 on T cells and PSMA on the engineered HEK293 cells, promoting CD28 activation on the engineered reporter T cells. Coactivation of TCR and CD28 enhances the transcriptional activity of NFκB, which in turn induces the production of the reporter gene luciferase.

[0197] d) Luciferase assay setup: RPMI1640 supplemented with 10% FBS and P / S / G was used as the assay medium to prepare cell suspensions and antibody dilutions for screening of anti-PSMAxCD28 bispecific antibodies.

[0198] The day before screening, engineered reporter T cells were plated at 1 x 10 in cell culture medium. 6 cell 100 nM. Three-fold (1:3) serial dilutions of anti-PSMA x CD28 bispecific antibodies and controls were tested in the presence of a constant 50 pM anti-CD20 x CD3 or hIgG4 isotype control. Ten-point dilutions ranged from 15 pM to 100 nM, with no anti-PSMA x CD28 antibody in the final dilution. Reagents were added in the following order: 1) serially diluted antibodies were added to corresponding wells of a 96-well white flat-bottom plate; 2) a fixed concentration of 50 pM anti-CD20 x CD3 or hIgG4 isotype control was added to each well; and 3) 4 x 10 5Resuspend APCs in 1000 cells / mL to a final concentration of 1 x 10 4 4) The reporter T cells were cultured overnight and added to the plate at 2 × 10 cells / well. 6 / m Resuspend in 1 L for a final concentration of 5 x 10 4 Cells / well were added to the plate. After incubation at 37°C / 5% CO2 for 4-6 hours, 100 μL of ONE-Glo (TM) was added. The cells were lysed by adding IgG (Promega, Cat. No. E6051) reagent, and luciferase activity was detected. The emitted light was captured in relative light units (RLU) on a multilabel plate reader Envision (PerkinElmer, Model 2104). All serial dilutions were tested in duplicate.

[0199] Antibody EC 50 Values were determined by fitting the data to a four-parameter logistic equation on a 10-point dose-response curve using GraphPad Prism™ software. Fold induction was calculated using the following equation:

number

[0200] 2) Primary human CD4 + IL-2 Functional Assay Using T Cells: First generation CD4 + T cell / APC functional assays were performed with anti-PSMA×CD28 bispecific antibodies. This assay was developed to assess the effect of CD28 activation on IL-2 production by engagement of CD28.

[0201] a) Isolation of human primary CD4+ T cells: Human peripheral blood mononuclear cells (PBMCs) were isolated from healthy donor leukocyte packs. PBMC isolation was achieved by density gradient centrifugation using 50 mL SepMate™ tubes (StemCell Technologies, Cat. #85450) according to the manufacturer's recommended protocol. Briefly, 15 mL of Ficoll-Paque PLUS was layered onto a 50 mL SepMate tube, followed by the addition of 30 mL of leukocytes diluted 1:2 with D-PBS (Dulbecco's Phosphate Buffered Saline, Irvine Scientific, Cat. #9240). Subsequent steps followed the SepMate manufacturer's protocol. CD4 + T cells, Miltenyi Bi Human CD4 was isolated from PBMCs using the Otec Human CD4 MicroBead Kit (Cat. No. 130-045-101) according to the manufacturer's instructions. + T cells , 10% DMSO (Macron Fine Chemicals, Cat. No. 4948-02) at 5 × 10 per vial 6 Cells were frozen in FBS containing 1000 μg of PBS at a concentration of 1000 μg / ml. .

[0202] b) Primary CD4 cells treated with CD28 antibody + IL-2 release from T cells: In this assay, primary CD4 + T cells express anti-CD20×CD3 bispecific antibodies In combination with HEK293 cells engineered to express PSMA-CD20, the antibody is activated via cross-linking of CD3 on the surface. Binding the CD20 arm of the anti-CD20 x CD3 bispecific antibody to CD20-expressing HEK293 cells promotes CD3 receptor clustering, providing the first signal critical for T cell stimulation. However, to detect quantifiable IL-2 release, costimulation, which can be provided by cross-linking CD28 molecules, is crucial. Here, the bispecific anti-PSMA x CD28 antibody binds to CD4 +CD28 on T cells and PSMA on engineered HEK293 / hCD20 cells The combined TCR and CD28 interact with each other, promoting the clustering and activation of CD28. Engagement of the combined TCR and CD28 enhances IL-2 production, which is released into the cell culture medium. IL-2 is detected and quantified in the cell supernatant using a homogenous, no-wash AlphaLisa kit (PerkinElmer, catalog number AL221).

[0203] Human CD4 previously isolated and frozen from donors + T cells were added to the assay at 50 U Cells were thawed in stimulation medium (X-VIVO 15 cell culture medium (Lonza, catalog #04-418Q) supplemented with 10% FBS, HEPES, NaPyr, NEAA, and 0.01 mM BME (β-mercaptoethanol, Sigma-Aldrich, catalog #M-7522) containing 1 × 10 / mL benzonase nuclease (EMD Millipore, catalog #71205-3). Cells were centrifuged at 1200 rpm for 10 min, resuspended in stimulation medium, and diluted to 1 × 10 5 in a 96-well round-bottom plate at a concentration of cells / well HEK293 cells engineered to express human CD20 alone or in combination with human PSMA were plated at 10 × 10 in primary stimulation medium with 15 μg / mL mitomycin C (Sigma-Aldrich, catalog #M4287). 6 cell / After incubation at 37°C for 1 hour, HEK29 cells were incubated in 5% CO2. 3. Wash the cells three times with D-PBS containing 2% FBS and plate them at 1 × 10 cells per well. 4 Thin CD4 at a final concentration of 1000 + was added to the wells containing the T cells.

[0204] Subsequently, 1:3 serial dilutions of anti-PSMAxCD28 bispecific or control antibodies ranging from 15 pM to 100 nM were added to the wells in the presence of 50 pM anti-CD20xCD3 or hIgG4 isotype control. The end point of the 10-point dilution did not contain anti-PSMAxCD28 or anti-CD28 antibodies. The plates were incubated at 37°C, 5% CO2 for 72 h. After incubation, they were centrifuged to pellet the cells, and 40 μL of the medium supernatant was collected. From this, 5 μL was tested in the human IL-2 Alpha LISA assay according to the manufacturer's protocol. Measurements were taken on a multilabel plate reader, Envision (PerkinElmer, Model 2104). A standard curve of known IL-2 concentrations was used to determine the concentration of IL-2 produced in the assay wells. All serial dilutions were tested in duplicate.

[0205] Antibody EC 50 Values were determined by fitting the data to a four-parameter logistic equation on a 10-point dose-response curve using GraphPad Prism™ software. Fold induction was calculated using the following equation:

number

[0206] Results, Summary, and Conclusions As shown in Figures 4A and 4B, activation of CD4+ T cells (measured by IL-2 release) and engineered JRT3.T3 / 1G4 / hCD28 cells (measured by luciferase activity) was significantly enhanced by primary stimulation (REGN2281 anti-CD20 x CD3) and HEK29 T cell activation was enhanced by hPSMA×hCD28 in the presence of PSMA expressed on hCD3 / hCD20 cells. The CD28 bivalent antibody, mAb14193P2, slightly enhanced T cell activity in the presence of primary stimulation and in engineered bioassays in the absence of primary stimulation. The CD28 superagonist, TGN1412, enhanced T cell activation in both primary and engineered assays in the presence of CD20×CD3 stimulation, although to a lesser extent than the PSMA×CD28 bispecific antibody in the presence of target PSMA and primary stimulation.

[0207] 1) Luciferase-based reporter assay: EC 50 and fold induction values are summarized in Tables 13 and 14 for engineered reporter T cells co-incubated with HEK293 / hCD20 or HEK293 / hCD20 / hPSMA cells in addition to 50 pM of constant hIgG4 isotype control or anti-CD3xCD20 bispecific antibody (T cell stimulatory bispecific antibody). [Table 13] Table 14 shows the EC values for luciferase activity in engineered T cells co-incubated with HEK293 / hCD20 or HEK293 / hCD20 / hPSMA cells and 50 pM of constant hIgG4 isotype control. 50 The values and fold induction results are summarized. [Table 14] Table 14 shows the results for engineered T cells co-incubated with HEK293 / hCD20 or HEK293 / hCD20 / hPSMA cells and 50 pM of constant anti-CD3×CD20. Luciferase activity in EC 50 The values and fold induction results are summarized.

[0208] When T cells and APCs were treated with 50 pM of hIgG4 isotype control, none of the CD28 bispecific antibodies showed an increase in luciferase activity in the absence of TCR stimulation, regardless of the APC lineage used in the assay. Little luciferase activation was observed with one of the parental CD28 antibodies (mAb14226P2) on HEK293 / hCD20 cells (4.76x) and HEK293 / hCD20 / hSPMA cells (3.59x) shown in Table 14.

[0209] In contrast, when cells were treated with 50 pM of anti-CD3 x CD20 bispecific antibody, all three anti-PSMA x CD28 bispecific antibodies, bs16429D, bs16430D, and bs16431D, strongly induced luciferase activity when co-incubated with APCs expressing hPSMA on their surface. Regardless of APC lineage, very low or no activation was observed with the one-arm controls (one-arm mAb14226P2, mAb14193P2, and mAb14216P2). As shown in Table 14, minimal luciferase activation was observed with all three parental CD28 antibodies (mAb14226P2, mAb14193P2, and mAb14216P2).

[0210] 2) Primary human CD4 + IL-2 Functional Assay Using T Cells: The ability of the anti-PSMA x CD28 bispecific antibody to provide costimulation through CD28 on T cells in the absence or presence of PSMA target expression was assessed using functional primary CD4 T cells measuring IL-2 cytokine production. + was assessed by T cell assay.

[0211] EC 50 and fold induction values are calculated using 50 pM of constant hIgG4 isotype control or anti-CD3 x CD20 bispecific antibody (T cell stimulatory bispecific antibody) plus CD4 co-incubated with HEK293 / hCD20 or HEK293 / hCD20 / hPSMA cells. + For T cells, the results are summarized in Table 15.

[0212] As expected, no measurable IL-2 release was observed in wells containing a fixed amount of hIgG4 isotype control due to the absence of primary T cell stimulation.

[0213] In contrast, measurable IL-2 levels were detected in samples treated with anti-CD3xCD20 bispecific antibodies. + T cells were isolated from HEK293 / h When co-incubated with CD20 cells, all CD28 monoclonal antibodies tested, including anti-CD28 antibodies and anti-PSMA x CD28 bispecific antibodies, except for bs16430D and bs16431D, showed an increase in IL-2 levels (Table 15). The parental antibody, mAb14226P2, showed the highest fold induction and an EC of approximately 6 nM. 50 CD4 + T cells were co-immunopreserved with hPSMA-expressing APCs and an anti-CD3 × CD20 bispecific antibody. Upon culture, IL-2 release was detected with all three anti-PSMA x CD28 bispecific antibodies (bs16429D, bs16430D, and bs16431D). Lower IL-2 levels are measured using the one-arm control antibody and parental antibody in the same settings as shown in Table 15. [Table 15] Table 15 shows the primary CD4+ cells co-incubated with HEK293 / hCD20 or HEK293 / hCD20 / hPSMA cells and 50 pm of constant anti-CD3×CD20 antibody. + T IL-2 production from EC 50 Values and fold induction are summarized.

[0214] Example 7: Anti-PSMAxCD28 bispecific antibody enhances T cell activation in the presence of both PSMA and TCR stimulation with anti-CD20xCD3 An exemplary anti-PSMA×CD28 bispecific antibody has been generated and validated by biacore (see, e.g., Examples 1 and 4). A pair of bispecific antibodies of the invention (CD20×CD3 and an exemplary anti-PSMA×CD28 bispecific antibody) was used to induce clustering of T cell and costimulatory receptors on tumor cells by binding to PSMA (CD20 and prostate-specific membrane tumor antigen—PSMA) (Figure 5A). A series of in vitro cell-based assays was performed to demonstrate that PSMA×CD28 binds and activates CD28 in the presence of PSMA expressed on target cells and TCR activation.

[0215] CD28 localization cell line A stable HEK293 cell line (ATCC, #CRL-1573) was used to generate the cell lines HEK293 / hCD20, HEK293 / hPSMA, and HEK293 / hCD20 / hPSMA. To generate the HEK293 / hPSMA cell line, stable transfection was performed using a mammalian vector encoding ubiquitin C promoter-driven PSMA (amino acids M1 to A750 of accession number Q04609) and a neomycin resistance gene. Similarly, the HEK293 / hCD20 cell line was generated using a mammalian vector encoding ubiquitin promoter-driven hCD20 (amino acids M1 to P297 of accession number NP_068769.2). Transfected cells were cultured in 500 μg / ml of Geneticin A to select for stably expressing cell lines.

[0216] For the generation of HEK293 / hCD20 / hPSMA cells, a lentiviral plasmid encoding human PSMA (amino acids M1 to A750 of accession number Q04609) and a neomycin resistance gene was used to transfect HEK293T cells to promote the production of viral particles, which were then used to infect HEK293 / hCD20 cells. Human PSMA-positive cells were isolated by fluorescence-activated cell sorting (FACS). All generated cell lines were cultured in DMEM+10 medium supplemented with 500 μg / mL G418. Jurkat clone E6-1 (ATCC, #TIB-152) was cultured according to the ATCC recommended protocol.

[0217] To generate MC38 / hPSMA cells, HEK293T cells were transfected with a lentiviral plasmid encoding human PSMA (amino acids M1 to A750 of accession number Q04609) and a neomycin resistance gene to promote the production of viral particles, which were then used to infect MC38 / parental cells. Human PSMA-positive cells were isolated by FACS. MC38 / hPSMA cells were maintained in DMEM + 10% FBS + P / S / G + NEAA supplemented with 500 μg / mL G418.

[0218] Amnis Image stream Jurkat T cells and target cells (HEK293 / hPSMA or HEK293 / hPSMA / hCD20) were incubated with CD20×CD3-Alexa488 (REGN2280, 0.5 μg / ml) alone or with PSMA×CD28-Alexa647 (bs16429D, 1 μg / ml) for 1 hour at 37°C. Cells were gently washed twice with FACS buffer (3% FBS, 2 mM EDTA in PBS) and stained with anti-CD28-biotin (REGN1412, 2 μg / ml) for 15 minutes at 4°C, followed by streptavidin-PE-CF594 (BD562284, 1 μg / ml) and Hoechst 33342 (Thermo Fisher H3570, 1 μM) for 15 minutes at 4°C. Cells were washed with PBS and stored in BD stabilizing fixative (BD338036). Images of cells were collected on an Amnis® imaging flow cytometer and analyzed using IDEAS® software. Cells were gated for doublet bright field, doublet nuclei, nuclear focus, single spot count, and singlet CD28. The synaptic region was defined by a valley mask based on nuclear staining. Cells with an incorrect valley mask were excluded by gating the overlapping region between the valley mask and CD28. The ratio of CD28 to intrasynaptic / extrasynaptic regions was calculated using the following formula: Intrasynaptic and extrasynaptic CD28 = intrasynaptic CD28 intensity / (total CD28 intensity - intrasynaptic CD28 intensity)*100%.

[0219] Results, Summary, and Conclusions T cells were co-cultured with engineered target cells overexpressing CD20 and PSMA (HEK293 / hPSMA or HEK293 / hCD20 / hPSMA) and fluorescently labeled bispecific antibodies (CD20xCD3 in green, PSMAxCD28 in red). To determine CD28 localization, cells were fixed and stained with anti-CD28 after 1 hour of incubation at 37°C. Images of T cell and target cell complexes were acquired using an Amnis ImageStream imaging flow cytometer. In the absence of PSMA expression on target cells, the CD20xCD3 bispecific antibody alone induced little or no CD28 clustering on T cells. When PSMA was expressed on target cells, CD20xCD3 localized at the interface of T cell and target cell complexes, forming an immune synapse (IS) where CD28 was localized. PSMAxCD28, together with CD20xCD3, further enhanced CD28 accumulation at the IS. The distribution of CD28 was quantified by calculating the ratio of CD28 staining inside and outside the IS (Fig. 5B).

[0220] It is concluded that PSMA×CD3 and PSMA×CD28 in the presence of PSMA-expressing target cells promote robust CD28 accumulation at the IS, where T cell activation signaling occurs.

[0221] Cytokine release To further investigate the effect of TAA×CD28 on T cell activation, primary human T cells were cultured in vitro. T cell proliferation and IL-2 and IFNγ cytokine release were measured after incubation with dose-titrated CD20×CD3 or PSMA×CD28 bispecific antibodies in cocultures with engineered target cells (HEK293 / hPSMA, HEK293 / hCD20, or HEK293 / hCD20 / hPSMA) overexpressing different TAAs (Figures 5C-5G). Exemplary anti-PSMA×CD28 and anti-PSMA×CD3 bispecific antibodies of the invention did not compete for binding to PSMA-expressing cells, thus confirming that they bind to different epitopes (Figure 5H). PSMA×CD28 was further validated to induce T cell activation in the presence of both CD3 stimulation and PSMA expression on target cells (HEK293 / hCD20 / hPSMA + 5 pM CD20×CD3). T cell activation by the exemplary anti-PSMA×CD28 bispecific antibody was not observed in the absence of PSMA expression on target cells (coculture with HEK293 / hCD20) or in the absence of CD3 stimulation (coculture with HEK293 / PSMA + 5 pM CD20×CD3 or HEK293 / hCD20 / hPSMA + 5 pM isotype control) (Figures 5E-5G). Overall, the exemplary anti-PSMA×CD28 antibody of the invention was demonstrated to promote T cell activation in the presence of PSMA×CD3 and PSMA-expressing target cells, resulting in increased proliferation and cytokine secretion.

[0222] Example 8: Killing of PSMA-expressing cells Two FACS-based cytotoxicity studies were conducted. In the first study, FACS-based cytotoxicity was performed on C4-2 cells in the presence of human peripheral blood mononuclear cells (PBMCs) with or without anti-PSMA×CD28 stimulation. Stimulation was performed in the presence of a fixed concentration of anti-PSMA×CD28 bispecific antibody and serially diluted anti-PSMA×CD3 bispecific antibody. The second study was identical to the first except that cynomolgus monkey PBMCs were used instead of human PBMCs. It was concluded that the PSMA×CD28 bispecific antibody enhanced T cell activation and cytotoxicity against prostate tumor cells in the presence of TCR stimulation with PSMA×CD3.

[0223] Experimental Method Primary human CD4 + Isolation of T cells Human peripheral blood mononuclear cells (PBMCs) were isolated from healthy donor leukocyte packs by density gradient centrifugation using 50 mL SepMate™ tubes according to the manufacturer's protocol. PBMCs were then sequestered in EasySep™ Human CD4 (StemCell Technologies) tubes. + Use a T cell isolation kit CD4 + T cells were isolated from PBMCs. D4 + T cells were added at 50 x 10 per vial. 6 The cells were frozen in FBS containing 10% DMSO at a concentration of 1000 μg / ml.

[0224] Primary human T cell activation assay Previously isolated and frozen human CD4 + T cells were treated with 50 U / ml benzonase nuclear Cells were thawed on the day of the assay in stimulation medium (X-VIVO 15 cell culture medium supplemented with 10% FBS, HEPES, NaPyr, NEAA, and 0.01 μM BME) containing ATP. Cells were centrifuged at 1200 rpm for 10 min, resuspended in stimulation medium, and plated at 1 × 10 per well. 5 HEK29 cells were plated in 96-well round-bottom plates at a concentration of 0.1%. 3 cells (HEK293 / hPSMA, HEK293 / hCD20, or HEK293 / hPSMA / hCD20) were cultured at 10 × 10 6 in primary stimulation medium at a concentration of 15 μL of cells / mL After incubation for 1 hour at 37°C in 5% CO2, HEK293 cells were washed three times with D-PBS containing 2% FBS and plated in one well. 2 x 10 4 Final concentration of cells CD4 + was added to the wells containing the T cells.

[0225] Determine suboptimal CD20xCD3 bispecific antibody concentrations for activating T cells To test for CD20xCD3, CD20xCD3 was serially diluted 1:3 from 1.5 pM to 10 nM in the presence of 500 pM of PSMAxCD28 bispecific antibody or hIgG4s isotype control. A constant 5 pM of CD20xCD3 bispecific antibody or hIgG4 control was selected, and the following antibodies were titrated from 15 pM to 100 nM in 1:3 dilutions: 1) PSMAxCD28, 2) non-targeting xCD28 control, 3) CD28 parent, 4) CD28SA (superagonist), 5) hIgG4 isotype control, and 6) hIgG4s isotype control. The final point of the 10-point dilution did not contain the titrated antibody, but only 5 pM of CD20xCD3 bispecific antibody or hIgG4 control.

[0226] The plates were incubated at 37°C and 5% CO2 for 48 hours, after which they were centrifuged. Cells were pelleted using a 50 μL syringe and 50 μL of media supernatant was collected. From this, 5 μL was tested in the human IL-2 and human IFNγ AlphaLISA assays according to the manufacturer's protocol. Measurements were taken on a Perkin Elmer Envision multilabel plate reader. Standard curves of known IL-2 or IFNγ concentrations were generated to estimate the pg / mL of IL-2 or IFNγ produced in the assay wells. All serial dilutions were tested in duplicate. Pelleted cells were incubated with [methyl-3H]-thymidine, 0.25 μCi / well, for 16 hours at 37°C, 5% CO2. Cells were harvested onto Perkin Elmer Unifilter plates using a Perkin Elmer Unifilter 96 cell harvester. After adding 30 μl of scintillation fluid, the plates were sealed and counts per minute were obtained for each well using a Perkin Elmer TopCount NXT.

[0227] Antibody EC 50Values were determined from a four-parameter logistic equation on a 10-point dose-response curve using GraphPad Prism™ software.

[0228] FACS-based cytotoxicity assay To monitor the killing of PSMA+ cells in the presence of a combination of anti-PSMAxCD3 and anti-PSMAxCD28 antibodies, C4-2 cells were labeled with 1 μM of the fluorescent tracking dye Violet Cell Tracker (Invitrogen, Cat. No. 34557). After labeling, cells were plated overnight at 37°C. Separately, human PBMCs (New York Blood Center) or cynomolgus monkey PBMCs (Covance, Cranford) were used. NJ) to 1×10 6 Adherent macrophages, seeded in supplemented RPMI medium at 100 cells / mL, The cells were incubated overnight at 37°C to enrich for lymphocytes by depleting dendritic cells and some monocytes. The next day, target cells were co-incubated with naive PBMCs (effector / target cell ratio 4:1) depleted of adherent cells and serial dilutions (concentration range: 0.42 nM to 0.1 pM) of anti-PSMA x CD3 bispecific antibody or IgG4 control alone or in combination with a fixed concentration of the PSMA x CD28 costimulatory molecules bs16429D or bs16431D at 2.5 μg / ml (16.7 nM) for 96 h at 37°C.

[0229] After incubation, cells were removed from cell culture plates (trypsin-EDTA (Millipore, catalog no. SM-2004-C) dissociation buffer) and analyzed by FACS on a FACS BD LSRFortessa-X20 (BD).

[0230] For FACS analysis, cells were stained with a far-red cell tracker for viability (Invitrogen) and antibodies directly conjugated to CD2, CD4, CD8, and CD25 (BD). Samples were run with calibration beads for cell counting. To assess the specificity of killing, target cells were gated as the violet cell tracker-positive population. The percent of live target cells was calculated as follows: Percentage of live cells (%) = (R1 / R2) * 100 (where R1 = percentage of live target cells in the presence of antibody). - percentage (%), and R2 = percentage (%) of live target cells in the absence of test antibody. T cell activation was determined by CD2 + / CD4 + or CD2 + / CD8 + ) Activation of T cells (CD25 + T cell numbers were measured by the percentage of carrier T cells. Viable CD4 per ablation bead + or CD8 + The number of cells was determined by counting.

[0231] The levels of cytokines accumulated in the medium were analyzed using the BD Cytometric Bead Array (CBA) Human Th1 / Th2 / Th17 Cytokine Kit according to the manufacturer's protocol.

[0232] For FACS analysis, cells were stained with dead / live near-infrared-reactive (Invitrogen, Cat. No. L34976) dye. Immediately prior to FACS analysis, 500,000 (5 × 10 5 100,000 (1× counting beads) were added to each well. 10 5 To assess the specificity of killing, cells were viably viable. The cells were gated on the dot-labeled population. The percent of the live population was recorded and used to calculate survival.

[0233] T cell activation and upregulation of PD-1 markers were assessed by incubating cells with antibodies directly conjugated to CD2, CD4, CD8, CD25, and PD-1 and reporting the percentage of late-activated (CD25+ / CD8+) T cells and PD-1+ / CD4+ T cells among total T cells (CD2+). The directly conjugated antibodies are as follows: CD2, PE:CD2 (Cl:RPA-2.1), BD, Catalog No. 555327; CD4, PerCP-Cy5.5:CD4 (Cl:OKT-4), Biolegend, Catalog No. 317428; CD8, APC:CD8 (Cl:RPA-T8), Biolegend, Catalog No. 301049; CD25, BV510:CD25 (Cl:M-A251), BD, Catalog No. 563352; and PD-1, PE-Cy7:PD1 (Cl:EH12.2H7), Biolegend, Catalog No. 329918.

[0234] Supernatants from assay wells from the human PBMC assay were assessed for Th1 / Th2 cytokine release using the BD Cytometric Bead Array Human Kit (BD, Cat. No. 560484) according to the manufacturer's protocol.

[0235] Results, Summary, and Conclusions: Figures 6A and 6B show that anti-PSMA×CD28 enhanced the cytotoxic potency of bs13644D (anti-PSMA×CD3) in the presence of human or cynomolgus monkey T cells. The anti-PSMA×CD3 bispecific antibody was tested for its ability to induce naive human T cells to kill human PSMA-expressing target cells, either as a single agent or in the presence of costimulatory anti-PSMA×CD28 bispecific antibody. In the studies shown in Figures 6A and 6B, the anti-PSMA×CD28 antibody was fixed at 2.5 μg / ml. Anti-PSMA×CD3 was serially diluted 1:4. The antibodies were incubated with the cells for 96 hours. Table 16 summarizes the components of the antibodies used in this study. [Table 16]

[0236] We tested whether the exemplary anti-PSMAxCD28 bispecific antibody could enhance T cell activation and cytotoxicity against prostate cancer cells by targeting PSMA. Using FACS-based cytotoxicity and T cell activation assays, we tested the exemplary anti-PSMAxCD28 bispecific antibody in combination with PSMAxCD3 (Figures 6, 7A-7D). Human PBMCs containing T cells were cocultured with C4-2 prostate cancer cells, which endogenously express high levels of PSMA (data not shown). PSMAxCD28 significantly increased the potency of cytotoxicity induced by PSMAxCD3 alone, increasing the EC50 (over a log shift in potency) to 4.3 x 10. -11 to 1.5 x 10 -12 The IFNγ release shifted to a higher level (Figures 6 and 7A). Consistent with the induction of T cell cytotoxicity, PSMA×CD28 enhanced the level of IFNγ release induced by PSMA×CD3 by fourfold (Figure 7B). Similarly, the combination of PSMA×CD28 with PSMA×CD3 increased the number of CD4 and CD8 T cells and the expression of the activation marker CD25 (Figures 7C and 7D). When the exemplary anti-PSMA×CD28 bispecific antibody was combined with a non-targeting CD3 bispecific antibody, no effect on T cell cytotoxicity or activation was observed. Furthermore, to demonstrate that the exemplary anti-PSMA×CD28 bispecific antibody can enhance PSMA×CD3 cytotoxicity and activation of T cells from non-human primates, the same assay was performed using PBMCs from cynomolgus monkeys, with similar results (Figures 6, 7E-7G). These results demonstrated that the exemplary anti-PSMAxCD28 bispecific antibody can potently enhance PSMAxCD3-mediated T cell activation not only through proliferation and cytokine release but also through cytotoxicity. Figure 7H shows that the exemplary anti-PSMAxCD28 bispecific antibody binds to cellular targets as measured by flow cytometry.

[0237] Anti-PSMA x CD3 bispecific antibodies activated human T cells and induced them to deplete C4-2 cells (a human prostate cancer cell line). Target cell killing was observed in the presence of anti-PSMA x CD3 bispecific antibodies, and C4-2 cells were depleted at pM EC 50 The observed target cell lysis was also greater than picomolar EC 50 Second, this was associated with upregulation of CD25+ and PD-1+ T cells over CD2+ T cells (Table 17). The anti-PSMAxCD3 bispecific antibody induced the release of human cytokines. The cytotoxic activity observed with the anti-PSMAxCD3 bispecific antibody as a single agent was enhanced in the presence of a fixed concentration of anti-PSMAxCD28 costimulatory molecule (bispecific antibody) (Table 17).

[0238] In summary, costimulation with the anti-PSMAxCD28 bispecific antibody increased T cell activation, PD-1 upregulation, and cytokine release when compared to that observed with the anti-PSMAxCD3 bispecific antibody as a single agent. Table 16 summarizes the results of experiments using human PBMCs. [Table 17]

[0239] The anti-PSMAxCD3 bispecific antibody was also tested for its ability to induce naive cynomolgus monkey T cells to kill target cells expressing human PSMA, either as a single agent or in the presence of a costimulatory anti-PSMAxCD28 bispecific antibody. At selected antibody titrations, the anti-PSMAxCD3 bispecific antibody activated human T cells but did not induce the T cells to deplete C4-2 cells (Table 18). Costimulation with the anti-PSMAxCD28 antibody increased T cell activation, enhanced cytotoxic activity, and upregulated the PD-1 marker on T cells (Table 18). [Table 18]

[0240] Example 9: Cytokine release from cells As noted elsewhere herein, ex vivo cytokine release assays using human PBMCs in a "soluble" format were unable to predict cytokine release, so a "coated" assay format was developed.

[0241] The anti-PSMA / CD28 bispecific antibodies and anti-CD28 antibodies of the invention were evaluated for their ability to induce human PBMC proliferation and cytokine release from the cells using a coated assay format. In the PBMC proliferation assay, 1 x 10 freshly isolated PBMCs were cultured in a 5% COOH / 2000 PBSC culture medium. 5 For co-culture, mitomycin C was used at 1000 μg / well. 1 x 10 C4-2 cells treated with 4 Antibody dilutions were added at 1000p / well. Plates were wet-coated or dry-coated overnight. Plates were washed prior to adding PBMCs. Supernatants were collected at approximately 54 hours for cytokine analysis by Meso Scale Diagnostics (MSD, Rockville, MD). 3 H-thymidine was added for 18 hours and proliferation was measured.

[0242] The results showed that the soluble anti-PSMA x CD28 bispecific antibody induced proliferation of human PBMCs in the presence of CD3 stimulation. In contrast, the anti-CD28 superagonist induced proliferation of human PBMCs in the absence of CD3 stimulation (data not shown).

[0243] Furthermore, wet-coated bs16429D (PSMA × CD28 “A”) and mAb14226P2 (CD28 parent “A”) were significantly different from wet-coated anti-C Although they induced less cytokine release than the D28 superagonist, bs16431D (PSMAxCD28 "B") and mAb14216P2 (CD28 parental "B") did not induce significantly more cytokine release compared to controls (data not shown).

[0244] The anti-CD28 antibodies of the present invention exhibit several distinct properties compared to the superagonist TGN1412. Conventional CD28 agonist mAbs (non-superagonists) are thought to bind to membrane-distal epitopes, allowing only monovalent Ab binding (Dennehy et al., Cutting Edge: Monovalency of CD28 maintains the antigen dependence of T). CD28 antibodies promote optimal T cell activation in the presence of antigen-specific recognition of peptide-MHC by the TCR. CD28 signaling modulates the threshold for TCR activation, significantly reducing the number of TCR engagements required for effective T cell activation. In contrast, TGN anti-CD28 antibodies possess some superagonist properties (Luhder et al., Topological requirements and signaling properties of T cell-activating, anti-CD28 antibody superagonists, J. of Exp. Med. 197(8):955-966(2003); Riley et al., The CD28 family: a T-cell rheostat for therapeutic control of T-cell activation, Blood 105(1):13-21(2005)). For example, TGN anti-CD28 induces robust T cell proliferation and IL2 production in vitro and in vivo even in the absence of TCR signaling, enhances NF-kB activity more effectively than conventional anti-CD28 / CD3 signaling, induces AP-1 / SRE activation, and binds to CD28 epitopes proximal to the cell surface, allowing bivalent binding.

[0245] Thus, as shown in Figure 8, the CD28 superagonist TGN1412 activated the AP-1 reporter in the absence of CD3 primary stimulation. In contrast, the anti-CD28 antibody of the present invention (parent of PSMAxCD28) minimally activated AP-1 in an AP-1 luciferase bioassay.

[0246] Example 10: In vivo studies of anti-PSMA x CD28 antibodies Syngeneic tumor models Consistent with the in vitro studies in Example 8, combining anti-CD3xPSMA and anti-CD28xPSMA bispecific antibodies targeting tumor antigens enhanced tumor clearance in mouse models. To test the efficacy of the anti-PSMAxCD28 bispecific antibody as monotherapy or in combination with the anti-PSMAxCD3 bispecific antibody, syngeneic tumor experiments were performed in mice expressing human CD28, human CD3, and human PSMA instead of the corresponding mouse genes (hCD3 / hCD28 / hPSMA mice), generated using Velocigene's proprietary technology. It was concluded that the anti-PSMAxCD28 bispecific antibody enhanced antitumor immunity in vivo through anti-PSMAxCD3-induced T cell activation.

[0247] T cell activation is initiated when the T cell receptor (TCR) / CD3 complex binds to a peptide-MHC complex ("signal 1"). Activation is then reinforced by the engagement of a second "costimulatory" receptor, such as the CD28 receptor on the T cell, which binds to its cognate ligand(s) on the target cell ("signal 2"). Recently described CD3-based "bispecific antibodies" replace traditional signal 1 and act by attaching a tumor-specific antigen (TSA) to one arm of the bispecific antibody, allowing the T cell to bind to the tumor cell, and bridging it to the TCR / CD3 in the other arm. Of these TSA x CD3 bispecific antibodies, Although some have shown promising antitumor effects in cancer patients, their activity has yet to be optimized. As described elsewhere herein, the present invention introduces a novel class of bispecific antibodies that mimic signal 2 by crosslinking a second TSA to the costimulatory CD28 receptor on T cells. These bispecific antibodies are referred to as TSA×CD28 bispecific antibodies or anti-TSA / CD28 bispecific antibodies. As described herein, one exemplary antibody of the present invention is specific for a prostate cancer antigen (e.g., PSMA). Unlike CD28 superagonists, which broadly activate T cells and, in certain cases, have resulted in severe toxicity in early clinical trials, these TSA×CD28 bispecific antibodies have shown limited activity and no toxicity when used alone in genetically humanized immunocompetent mouse models or primates. However, when combined with a TSA×CD3 bispecific antibody, the exemplary antibody of the present invention strengthened the artificial synapse between T cells and their target cells, enhanced T cell activation, and significantly improved the antitumor activity of a CD3 bispecific antibody in a syngeneic tumor model. Combining this novel class of CD28 costimulatory bispecific antibodies with the new class of TSAxCD3 bispecific antibodies may provide a well-tolerated "off-the-shelf" antibody therapy with enhanced antitumor efficacy.

[0248] The ability of T cells to recognize and kill cellular targets, such as virus-infected or tumor cells, depends on a coordinated set of interactions. The most important of these is the recognition and binding of target cells by the TCR complex (containing the associated CD3γ, δ, ε, and ζ chains). This interaction is referred to as "signal 1" of T cell activation. The TCR can recognize viral or tumor peptides present in the groove of the MHC protein expressed on the surface of target cells. This binding is usually low affinity; therefore, successful triggering of signal 1 requires the clustering of many TCR complexes along the interface between the T cell and its target cell, known as the immune synapse (J.B. Huppa, M.M. Davis, T-cell-antigen recognition and the immunological synapse. Nat Rev Immunol 3, 973–983 (2003)). T cell activation and proliferation are then further promoted by additional interactions with costimulatory receptors such as CD28 ("Signal 2") (J.H.E. Sensten, Y.A. Helou, G. Chopra, A. Weiss, J.A. Bluestone, CD28 Costimulation: From Mechanism to Therapy. Immunity 44, 973-988 (2016)). T cell activation is enhanced when T cells recognize target cells via the TCR complex and engage Signal 2 via CD28 binding to its cognate ligand(s) (CD80 / B7.1 and / or CD86 / B7.2) on professional antigen-presenting cells or target cells. Similar to Signal 1, CD28-mediated Signal 2 is thought to occur through co-clustering at the immune synapse.

[0249] Conventional monoclonal antibodies targeting tumor-specific antigens (TSAs) have been used as antitumor therapeutic agents for the past 20 years (G. Salles et al., Rituximab in B-Cell Hematologic Malignancies: A Review of 20 Years of Clinical Experience. Adv Ther 34, 2232-2273 (2017); MV Mateos et al., Daratumab plus Bortezomib, Melphalan, and Prednisone for Untreated Myeloma. N Engl J Med 378, 518-528 (2018); W. Eiermann, G. International Herceptin Study, Trastuzumab combined with chemotherapy for the treatment of HER2-positive metastatic breast cancer: pivotal trial data. A nn Oncol 12 Suppl 1, S57-62 (2001); JM Connors et al., Brentuximab Vedotin with Chemotherapy for Stage III or IV Hodgkin's Lymphoma. N Engl J Med 378, 331-344 (2018); V. Dieras et al., Trastuzumab emtansine versus capecitabine plus lapatinib in patients with previously treated HER2-positive advanced breast cancer (EMILIA): a descriptive analysis of final overall survival results from a randomized, open-label, phase 3 trial. Lancet Oncol 18, 732-742 (2017). However, this class of antibodies has limited ability to induce T cell-mediated cytotoxicity and instead acts by promoting antibody-dependent cellular cytotoxicity (ADCC) and / or complement-dependent cytotoxicity (CDC) or by delivering toxins to tumor cells. Recently, a new class of bispecific antibodies (TSA × CD3) has emerged that can efficiently induce T cell-mediated killing of tumor cells by binding T cells to tumor cells and activating the CD3 / TCR complex via a surrogate mechanism (usually via the e chain of CD3), thus mimicking signal 1.An early version of such a bispecific antibody (one arm binding to CD19 on leukemia cells and the other to CD3) recently received regulatory approval for B-cell acute lymphoblastic leukemia (R. Bargou et al., Tumor regression in cancer patients by very low doses of a T cell engaging antibody. Science 321, 974-977 (2008); H. Kantarjian et al., Blinatumomab versus chemotherapy for advanced acute lymphoblastic leukemia. N Engl J Med 376, 836-847 (2017)). Recently, a more advanced version of the bispecific antibody has been shown to have good activity against non-Hodgkin's lymphoma, targeting CD20 on these lymphomas (EJ Smith et al., A novel, native-format bispecific antibody triggering T-cell killing of B cells is robustly active in mouse tumors). models and cynomolgus monkeys.Sci Rep 5,17943(2015)、L.L.Sun et al.,Anti-CD20 / CD3 T cell-dependent bispecific antibody for the treatment of B cell malignancies.Sci Transl Med 7,287ra270(2015)、M.Bacac et al.,CD20-TCB with Obinutuzumab Pretreatment as Next-Generation Treatment of Hematologic Malignancies.Clin Cancer Res 24,4785-4797(2018)、R.Bannerji et al.,Emerging Clinical Activity of REGN1979,an Anti-CD20xAnti-CD3 Bispecific Antibody,in Patients with Relapsed / Refractory Follicular Lymphoma(FL),Diffuse Large B-Cell Lymphoma(DLBCL),and Other B-Cell Non-Hodgkin Lymphoma(B-NHL)Subtypes.American Society of Hematology,(2018)、L.Budde et al.,Mosunetuzumab,a Full-Length Bispecific CD20 / CD3 Antibody,Displays Clinical Activity in Relapsed / Refractory B-Cell Non- Hodgkin Lymphoma(NHL):Interim Safety and Efficacy Results from a Phase 1 Study. American Society of Hematology, (2018)). However, although TSA × CD3 bispecific antibodies have emerged as an important new class of immunotherapy in hematological malignancies, a cross-sectional comparison (E A Zhukovsky, R J Morse, M V Maus, Bispecific antibodies and CARs: generalized immunotherapeutics harnessing T cell redirection. Curr Opin Immunol 40, 24-35 (2016)) suggests that in some cases, they may not achieve the level of efficacy seen with personalized chimeric antigen receptor T cell (CAR-T) therapy.

[0250] One of the reasons for the powerful efficacy of CAR-T therapy is that the chimeric antigen receptor (CAR) is engineered to provide both signal 1 (via a portion of the CD3z cellular domain) and signal 2 (e.g., via a portion of the CD28 cellular domain) upon binding to its target on a tumor cell. Two CAR-T cell therapies have recently received FDA approval for B-cell malignancies, both of which work by binding to and targeting the antigen CD19 (S.S. Neelapu et al., Axicabtagene Ciloleucel CAR T-Cell Therapy in Refractory Large B-Cell Lymphoma. N Engl J Med 377, 2531-2544 (2017); S.J. Chuster et al., Chimeric Antigen Receptor T Cells in Refractory B-Cell Lymphomas. N Engl J Med 377, 2545-2554 (2017)). CAR-T cell approaches have been shown to reduce the risk of serious adverse effects such as cytokine release syndrome (CRS) and neurotoxicity (S.S.Neelapu et al. al., Chimeric antigen receptor T-cell therapy-assessment and management of toxicities. Nat Rev Clin Oncol 15,47-62(2018), J. Gust et al., Endothelial Activation and Blood-Brain Barrier Disruption in Neurotoxicity after Adoptive Immunotherapy with CD19 CAR-T Cells. Cancer Discov 7, 1404-1419 (2017), A. Shimabukuro-Vornhagen et al., Cytokine release syndrome. J Immunother Cancer 6, 56 (2018)), as well as the requirement for highly personalized manufacturing processes and pre-adjustment of chemotherapy regimens (SS Neelapu et al., Axicabtagene Ciloleucel CAR T-Cell Therapy in Refractory Large Cells B Cell Lymphoma. N Engl J Med 377, 2531-2544 (2017), SJ Schuster et al., Chimeric Antigen Receptor T Cells in Refractory B-Cell Lymphomas. N Engl J Med 377, 2545-2554 (2017), P. Salmikangas, N. Kinsella, P. Chamberlain, Chimeric Antigen Receptor T-Cells (CART-Cells) for Cancer Immunotherapy - Moving Target for Industry? Pharm Res 35, 152 (2018), Many patients are not considered suitable candidates.

[0251] The advantage of the TSAxCD3 bispecific antibody as a relatively well-tolerated, "off-the-shelf" therapeutic solution for a broader patient population is that its antitumor activity could be further optimized, particularly in patients with refractory disease. If this could be done without sacrificing tolerability, further improvements could be achieved, perhaps even greater specificity for tumor cells as opposed to normal cells. Toward this end, it was hypothesized that pairing the TSAxCD3 bispecific antibody with a novel class of bispecific antibodies that independently activate signal 2 might offer opportunities for potentially improved efficacy and specificity. Therefore, a second class of bispecific antibodies was designed. These bispecific antibodies could engage either a second epitope on the same tumor-specific antigen or a second, distinct tumor antigen in conjunction with the costimulatory receptor CD28 expressed on T cells (TSAxCD28 bispecific antibody). Combining TSA1xCD3 and TSA2xCD28 would enable directed and enhanced surrogate activation of T cells by triggering both signal 1 and signal 2, with specificity targeted exclusively to tumor cells expressing both epitopes or both antigens, potentially enabling greater antitumor activity along with the opportunity for enhanced specificity.

[0252] Described herein is the generation and testing of TSA x CD28 costimulatory bispecific antibodies (PSMA x CD28, binding to prostate-specific membrane tumor antigen) targeting prostate cancer. Toxicology studies in genetically humanized immunocompetent mice and cynomolgus monkeys demonstrate that these bispecific antibodies exhibit limited activity and no toxicity as single agents. However, these novel costimulatory bispecific antibodies can be effectively combined with a new class of TSA x CD3 bispecific antibodies to enhance antitumor responses in syngeneic tumor models. Collectively, these data suggest that combining this novel class of CD28-based bispecific antibodies (TSA x CD28) with CD3-based bispecific antibodies (TSA x CD3) could provide a well-tolerated, "off-the-shelf" biologic solution with significantly enhanced synergistic antitumor activity.

[0253] material and method In Examples 10 and 13, the following materials and methods were used.

[0254] Syngeneic tumor studies Mice expressing human CD28, human CD3, and human PSMA in place of the corresponding mouse genes were generated using Velocigene® technology (referred to as hCD3 / hCD28 / hPSMA humanized mice) as previously described (Valenzuela (2003), Nat Biotechnol, Jun;21(6):652-9; Crawford et al. 2018, Manuscript in preparation). For each humanized mouse, correct gene targeting in F1H4 (C57BL / 6x129 hybrid) embryonic stem (ES) cell clones was identified by allelic loss assays as previously described (Poueymirou et al. (2007), Nat Biotechnol, Jan;25(1):91-9). Targeted ES cells were injected into 8-cell stage Swiss Webster embryos to generate fully heterozygous F0 mice for homozygous breeding with C57BL / 6N mice (Taconic, Rensselaer, NY). hCD3 / hCD28 / hPSMA mice (4–8 / group, 8–16 weeks old) were injected with 1 × 10 6 MC38 / hPSMA tumor cells Anti-PSMA x CD28 bispecific antibody, anti-PSMA x CD3 bispecific antibody, or human IgG4 isotype control were administered as monotherapy or in combination by intraperitoneal injection of 5 mg / kg on days 0, 3, and 7.

[0255] Tumor growth was monitored over time using caliper measurements of X and Y diameters. Tumor volume was calculated (X*Y*(X / 2)). When tumor size reached 2000 mm 3 When it exceeds The mice were euthanized.

[0256] Measurement of serum cytokine levels in mice Blood was collected by submandibular puncture into microtainer serum tubes (BD365967) at the indicated time points. Cytokine levels were analyzed using the V-plex Human ProInflammatory-10 Plex kit according to the manufacturer's instructions (Meso Scale Diagnostics, Rockville, MA).

[0257] Results, Summary, and Conclusions Expression of human CD28 on T cells and NK cells was verified by FACS (data not shown). Triple-humanized mice were functionally validated using an in vitro T cell proliferation assay (data not shown). PSMA expression was confirmed by QPCR (data not shown). MC38 / hPSMA tumor growth was monitored over time in triple-humanized hCD3 / hCD28 / hPSMA mice (as described above) (Figure 9A). Compared with the isotype control, both PSMAxCD3 and PSMAxCD28 monotherapy treatments significantly inhibited tumor growth (p<0.001 and p<0.0001, respectively). Tumor growth was further significantly inhibited by combination therapy (p<0.00001). Compared with the isotype control, the combination of PSMAxCD3 and PSMAxCD28 bispecific antibodies provided the greatest cytokine production (Figures 9C and 9D). Intratumoral CD8 + To further understand the role of combination therapy on T cell activation status, viSNE Analysis was performed (Figure 9B). viSNE is a tool that maps high-dimensional cytometry data to 2D while preserving high-dimensional structure. Each treatment promoted unique CD8+ T cell clusters in the spleen and tumor. Combination therapy promoted the expansion of activated / memory T cell phenotypes (expressing TCF1, CD1-2, CD127, PD-1, ICOS, KLRG1, and CD38), as shown in cluster 4 of the tumor.

[0258] hCD3 + / + / hPSMA + / + / hCD28+ / + Mouse and hCD3 + / - / hPSMA + / - / hCD28 + / - Studies were conducted to determine serum cytokine (IFNγ, IL-2, IL-6, IL-10, TNFα, IL-4, and IL-5) levels in both mice. Blood was collected on days 0 (4 hours), 3, 7, and 11 after treatment. hCD3 + / + / hPSMA + / + / hCD28 + / + On day 0, except for IL-10 in mice In treatments involving anti-PSMAxCD3 antibodies, the combination of anti-PSMAxCD3 and anti-PSMAxCD28 significantly increased cytokine levels, providing the greatest cytokine release. In contrast, the IgG control and anti-PSMAxCD28 monotherapy did not induce an increase in cytokine production. Cytokine production, with the exception of IL-5 and IL-10, decreased to similar levels in all treatment groups after 3 days. IL-5 production was significantly higher in hCD3. + / + / hPSMA + / + / hCD28 + / + Although the production rate was lower on day 0, it was significantly higher on day 3. hCD3 showed survival but received combination therapy + / - / hPSMA + / - / hCD28 + / - With the mouse was not the case. hCD3 + / + / hPSMA + / + / hCD28 + / + In mice, IL- hCD3 production was similar in all treatment groups on days 0 and 3, but combination therapy resulted in significantly higher levels in the combination group on days 7 and 11. + / - / h PSMA + / - / hCD28 + / -In mice, combination therapy produced significantly more IL-10 on days 0, 3, and 11, but not on day 7. Anti-PSMA x CD3 monotherapy produced significantly more IL-10 on days 0 and 3, but not on days 7 and 11. Anti-PSMA x CD28 monotherapy only produced significantly more IL-10 on day 11 (data not shown).

[0259] As shown in Figure 9A, unlike previous in vitro analyses (see Example 8 above), in which the CD28 bispecific antibody had very limited single-agent activity, the CD28 bispecific antibody in this syngeneic MC38 / hPSMA model had more pronounced activity as a single agent. This suggested that "signal 1" was already activated to some extent in this MC38 model. Consistent with this, MC38 tumor cells express high levels of reactivated endogenous retroviral proteins, such as p15E, and C57BL6 mice express this neoepitope. It has previously been shown that endogenous T cells that recognize and respond to two self-antigens can be generated (J.C.Y.ang, D.Perry-Lalley, The envelope protein of an endogenous murine retrovirus is a tumor-associated T-cell antigen for multiple murine tumors. J Immunother 23, 177-183 (2000); H.J.Zeh, 3rd, D.Perry-Lalley, M.E.Dudley, S.A.Rosenberg, J.C.Y.ang, High avidity CTLs for two self-antigens demonstrate superior in vitro and in vivo antitumor In the MC38 model, intratumoral T cells responding to this p15E neoantigen were readily detected. Thus, the CD28 bispecific antibody in this syngeneic tumor model can promote endogenous TCR / CD3-dependent T cell responses, which can be further enhanced by providing additional "signal 1" activation via the CD3 bispecific antibody.

[0260] To determine the cellular mechanisms underlying the combination therapy, tumor-infiltrating and splenic CD8+ T cells from these experiments were profiled by high-dimensional flow cytometry and an unsupervised clustering approach. Each treatment was found to promote unique CD8+ T cell clusters in the spleen and tumor. As shown in cluster C35, single treatment regimens reduced intratumoral CD8+ T cells with a less activated phenotype (lower ICOS, KLRG1, Ki67, PD1, CD38, and LAG3, Figure 9E). However, as shown in cluster C4, combination therapy significantly promoted the expansion of a more activated / memory T cell phenotype (expressing Tcf1, CD122, CD127, PD1, ICOS, KLRG1, and CD38, Figure 9B).

[0261] Anti-CD28 × anti-PSMA does not induce elevated serum cytokines in CD28 / CD3 / PSMA humanized mice in the absence and presence of tumor. As described elsewhere herein, the superagonist anti-CD28 antibody TGN1412 induced a harmful "cytokine storm" in patients. The anti-CD28 antibodies and anti-CD28 x PSMA bispecific antibodies of the present invention did not cause a cytokine storm. In this study, several antibodies (summarized in Table 19) were administered to hCD3 / hCD28 / hPSMA humanized mice at a dose of 2.5 mg / kg or 0.25 mg / kg. Mice were bled 4 hours and 3 days after antibody administration. As shown in Figure 10, the anti-PSMA x CD28 bispecific antibody or the parent CD28 bivalent antibody did not induce serum cytokine production in CD3 / CD28 / PSMA humanized mice in the absence of tumor. [Table 19] At 2.5 mg / kg, TGN1412 inhibited IL-2, IL-4, IL-5, and It may significantly increase blood levels of TNFα (data not shown).

[0262] Furthermore, as shown in Figure 28, anti-PSMA x CD28 treatment alone or in combination with anti-PD1 did not elevate serum cytokines in tumor-bearing mice. In contrast, anti-PSMA x CD3 treatment, alone or in combination with anti-PD1, increased serum cytokine levels, including TNFα, IL-5, IL-10, IL-2, and IL-4, in tumor-bearing mice. Anti-PSMA x CD3 treatment induced cytokine expression 4 hours after administration. The cytokine elevation did not persist beyond day 7. This study was conducted in hCD3 / hCD28 / hPSMA-humanized mice bearing MC38 / hPSMA tumors. Serum was collected from hCD3 / hCD28 / hPSMA-humanized mice bearing MC38 / hPSMA tumors 4 hours after administration of 5 mg / kg of the indicated antibody or bispecific antibody on day 0 and 4 days later on day 7. Serum cytokines were measured using a 10-plex mouse cytokine MSD kit according to the manufacturer's protocol.

[0263] Furthermore, as shown in Figure 11, CD28 superagonist treatment, but not the anti-CD28 antibody of the present invention, promoted cytokine responses at 4 hours in PBMC-engrafted NSG mice. In the study shown in Figure 11, immunodeficient NSG mice were treated with 5x10 PBMCs from normal healthy donors on day -10. 6 PBMCs were engrafted. On day -1, T cells were Systemic engraftment of the cells was confirmed by staining peripheral blood for human T cell markers. On day 0, mice were intraperitoneally injected with either 50 μg of an IgG4 isotype control antibody, 5 μg or 50 μg of an anti-CD28 superagonist, or 5 μg or 50 μg of the anti-CD28 mAb 14226P2 antibody. Four hours after antibody injection, blood was collected from the animals and serum was prepared. Cytokine concentrations in serum were analyzed by multiplex assay (Meso Scale Discovery V-PLEX kit). The anti-CD28 superagonist increased serum levels of interferon-gamma (IFN-γ), IL-2, IL-6, and TNFα compared with animals treated with the isotype control antibody, whereas no increased cytokine response was observed in the serum of animals treated with anti-CD28 mAb 14226P2. Therefore, bispecific anti-CD28×PSMA potentially has a safer toxicological profile among costimulatory bispecific antigen binding molecules. Bispecific anti-CD28×PSMA did not induce a cytokine response, whereas anti-CD3×PSMA did. As shown in Figure 12, anti-CD3×PSMA, but not anti-CD28×PSMA, induced elevated IFNγ in humanized mice. In this study, serum was collected 4 hours after administration of 5 mg / kg of the indicated bispecific antibody from CD3 / CD28 / PSMA or CD3 / CD28 humanized mice bearing MC38 / hPSMA tumors. Serum cytokines were measured using a 10-plex mouse cytokine MSD kit according to the manufacturer's protocol. Data points represent individual cytokine levels per mouse. Bars on the graphs represent the mean per treatment group. Error bars represent + / - SEM. Statistical significance determined by one-way ANOVA and Holm-Sidak multiple comparison test using isotype-treated mice as control (**, p<0.01).

[0264] Consistent with the data from humanized mice, anti-CD3 × PSMA bispecific antibodies induced elevated CRP and systemic cytokine release in cynomolgus monkeys, a non-human primate. In monkey studies, anti-CD3 × PSMA bispecific antibodies were administered to cynomolgus monkeys at doses of 0.01, 0.1, and 0.5 mg / kg. The monkeys exhibited the following symptoms: vomiting, swelling, red / discolored skin; elevated CRP (approximately 10 vs. 1 mg / dL in control mice); elevated plasma cytokines (IL-6, TNF-α, IFN-γ, IL-2, and MCP); and clinical signs by day 1, including a decrease in absolute T cell counts in the blood (data not shown).

[0265] Furthermore, mice receiving anti-CD28 x PSMA treatment did not exhibit cytokine production or T cell margination compared with anti-PSMA x CD3 or anti-CD28 superagonist treatment (Figures 13A and 13B).

[0266] overview It has long been recognized that T cell activation via the TCR complex ("signal 1") can be significantly enhanced by costimulatory signals ("signal 2"), such as those mediated when the CD28 receptor on the T cell engages its ligands (CD80 / B7.1 and CD86 / B7.2) on the target cell (JH E Sensten, YA Helou, G. Chopra, A. Weiss, J.A. Bluestone, CD28 Costimulation: From Mechanism to Therapy. Immunity 44, 973-988 (2016)). Consistent with the data disclosed herein, the potential for CD28 costimulation to enhance T cell antitumor activity was first demonstrated in studies in which B7 ligands were overexpressed on tumor cells (R.H. Schwartz, "Costimulation of T lymphocytes: the role of CD28, CTLA-4, and B7 / B1 in interleukin-2 production and immunotherapy," Cell 71, 1065-1068 (1992); L. Chen et al., "Costimulation of antitumor immunity by the B7 counterreceptor for the T lymphocyte molecules CD28 and CTLA-4," Cell 71, 1093-1102 (1992)) and demonstrated improved T cell rejection of such B7-expressing tumors. This potentially inspired efforts to evaluate CD28-activating antibodies in human clinical trials. Unfortunately, a 2006 trial of such an antibody (TGN1412) resulted in life-threatening complications in all six human volunteers due to multiple organ failure caused by massive cytokine release syndrome (CRS) (G. Suntharalingam et al., Cytokine storm in a phase 1 trial of the anti-CD28 monoclonal antibody TGN1412. N Engl J Med 355, 1018-1028 (2006)). This catastrophe led to the halt of further trials of CD28 activating antibodies in humans.

[0267] Described herein is a novel class of CD28 costimulatory bispecific antibodies that can significantly and safely enhance antitumor activity by providing a costimulatory "signal 2." While these CD28 bispecific antibodies have limited activity on their own (in the absence of "signal 1"), pairing these CD28 bispecific antibodies with a new class of CD3 bispecific antibodies (or when these CD28 bispecifics are used in situations where an endogenous population of tumor-specific T cells already exists) can significantly enhance antitumor activity in the setting of "signal 1." Described herein is the generation and testing of a TSA x CD28 costimulatory bispecific antibody targeting prostate cancer (PSMA x CD28). In the absence of "signal 1," these CD28 bispecific antibodies were shown to have minimal activity in vitro or in vivo. However, these CD28 bispecifics can be paired with CD3 bispecific antibodies to form an artificial "immune synapse" containing tumor antigens and TCR and CD28 complexes. Furthermore, when paired with an appropriate CD3 bispecific antibody in vitro, these CD28 bispecific antibodies can efficiently and specifically promote T cell activation and tumor cell killing in an antigen-dependent manner. Furthermore, these CD28 bispecific antibodies also efficiently enhance the antitumor activity of CD3 bispecific antibodies in vivo in a tumor antigen-specific manner in syngeneic tumor models. In such models, CD28 bispecific antibodies appear to have minimal single-agent activity unless tumor-specific T cells are already present, and in such settings, they enhance this specific activity in a tumor antigen-dependent manner. Furthermore, TSA×CD Combination therapy of 28 with TSAxCD3 significantly enhances the expansion of intratumoral activated / memory T cell phenotypes in vivo. Finally, toxicology studies in genetically humanized immunocompetent mice and cynomolgus monkeys demonstrate that these bispecific antibodies exhibit limited activity and no toxicity as single agents in direct comparison with conventional CD28-activating antibodies.

[0268] Characterization of human-specific clinical candidates in the field of immuno-oncology is often limited to testing in xenogeneic tumor models using engrafted human immune cells. While these xenogeneic models are very convenient, they have limitations. The mice used in such xenogeneic models do not express the human tumor target in their normal tissues, thereby preventing the evaluation of test agents in the setting of normal tissue expression of the target. Indeed, if the target is also normally expressed at high levels in normal tissues, this may limit antitumor efficacy by diverting the test agent from the tumor and may result in toxicity to these normal tissues, neither of which could be evaluated in xenogeneic models. Additional limitations may involve the activity of engrafted human peripheral blood mononuclear cells (PBMCs) transferred into immunodeficient mice, which may differ from the activity of normal host T cells found in the immunocompetent system. To overcome these limitations and provide a better model for testing human-specific clinical candidates, double and triple genetically humanized mice were created here. In these models, tumor antigens were genetically humanized to allow normal expression in appropriate host tissues, and the CD3 and / or CD28 components were genetically humanized to allow immunocompetent host cells to respond to human-specific clinical candidates. In these genetically humanized, immunocompetent syngeneic animal models, a PSMA tumor-targeted CD28 bispecific antibody was found to enhance the antitumor activity of its corresponding CD3 bispecific counterpart. Similar enhancement of antitumor efficacy by different TSA x CD28 bispecific antibodies across multiple preclinical models suggests that this therapy is robust and not limited to specific tumor models, and may have broader utility as a novel combination target class for immunotherapy. Overall, the findings highlight that TSA x CD28 bispecifics can synergize with TSA x CD3 bispecifics, significantly enhancing the efficacy of well-studied TSA x CD3 bispecifics in a reasonably safe and well-tolerated manner, potentially providing a biologic solution that could justify testing in human clinical trials.

[0269] TSA × CD3 bispecific antibodies represent a promising new class of immunotherapy, but further optimization of antitumor activity will certainly be important in many cases. Similar to the CAR-T approach, which employs chimeric receptors that artificially activate both "signal 1" and "signal 2" to improve antitumor activity (E.A. Zhukovsky, R.J. Morse, M.V. Maus, Bispecific antibodies and CARs: generalized immunotherapeutics harnessing T cell redirection. Curr Opin Immunol 40, 24-35 (2016); S.L. Maude et al., Tisagenlecleucel in Children and Young Adults with B-Cell Lymphoblastic Leukemia. N Engl J Med 378, 439-448 (2018)), we demonstrate the potential benefits of combining a CD3-specific antibody (providing "signal 1") with a CD28 bispecific antibody (providing "signal 2") to enhance antitumor activity. In addition to the practical advantages such an approach offers over CAR-T therapy, it does not require laborious cell therapy preparations that must be individually customized for each patient, and allows patients to accept this cell therapy without the need for preemptive "lymphocyte depletion" via toxic chemotherapy, which often results in side effects (A. Shimabukuro-Vornhagen et al., Cytokine release syndrome. J Immunother Cancer 6, 56 (2018); C.H. June, R.S.O. Connor, O.U. Kawalekar, S. Ghassemi, M.C. Milone, CAR T cell immunotherapy for human cancer. Science 359, 1 361-1365 (2018))—The bispecific approach according to the present invention offers the potential for improved efficacy as well as improved safety and specificity of action. Specifically, by pairing a CD3 bispecific antibody specific for one antigen with a CD28 bispecific antibody specific for a second antigen, it is possible to utilize "combinatorial targeting," with improved efficacy occurring only in tumor cells expressing both antigens, thus focusing T cell killing solely on tumor cells expressing both antigens while limiting "off-target toxicity" in normal tissues expressing only one of the antigens. Collectively, the data disclosed herein suggest that combining a CD28-based bispecific antibody with a CD3-based bispecific antibody may provide a well-tolerated, "off-the-shelf" biologic solution with significantly enhanced and synergistic antitumor activity. Initial testing of this potential in human clinical trials is expected this year.

[0270] Example 11: Cynomolgus Monkey Toxicology Study These studies demonstrated that PSMAxCD28 alone or in combination did not induce systemic T cell effects compared to CD28 superagonists in cynomolgus monkeys. An exemplary anti-PSMAxCD28 bispecific antibody of the invention enhanced TAAxCD3 activation of T cells from cynomolgus monkeys (Example 8, Figures 7E-7G). To determine the safety and tolerability of an exemplary anti-PSMAxCD28 bispecific antibody of the invention, alone or in combination with anti-PSMAxCD3, a single-dose toxicology study was conducted in cynomolgus monkeys. Female or male cynomolgus monkeys were assigned to treatment groups as shown in Table 20. [Table 20]

[0271] Cynomolgus monkey studies were conducted in accordance with IACUC guidelines. Male cynomolgus monkeys (3 per group) received a single dose of each test article via intravenous infusion over approximately 30 minutes (combined treatments were administered as separate infusions totaling 1 hour). Toxicity assessment was based on clinical observations, qualitative food consumption, body weight, neurological examination, vital signs (temperature, heart rate, pulse oximetry, and respiratory rate), and clinical and anatomic pathology. Blood and tissue samples were collected for cytokine analysis, FACS immunophenotyping, histopathology, and toxicokinetic evaluation. CRP levels were analyzed on a Roche Modular P800 system. Cytokines were measured by Meso Scale Diagnostics (MSD, Rockville, MD). For peripheral blood flow cytometry, blood was collected in potassium EDTA tubes, lysed, stained with the indicated antibodies, such as anti-CD3, anti-Ki67, and anti-ICOS (BD Biosciences), and analyzed on a FACS Canto II. Following administration of a single dose of PSMAxCD28 at 1 mg / kg or 10 mg / kg, no significant cytokine release, T cell margination, or T cell activation marker upregulation was observed and did not induce significant cytokine release, T cell margination, or T cell activation marker upregulation (Table 21). [Table 21]

[0272] In contrast, monkeys administered the CD28 superagonist showed significant cytokine release, lymphocyte margination, and T cell activation. Furthermore, these findings were verified using dry-coated and wet-coated human T cell proliferation assays (Example 9). Indeed, the exemplary anti-PSMA x CD28 bispecific antibody and the parent bivalent CD28 antibody were not observed to induce human T cell proliferation compared to the CD28 superagonist antibody. Overall, the exploratory single-dose toxicology studies in monkeys and the in vitro human T cell-based assays suggest that the exemplary anti-PSMA x CD28 bispecific antibody of the present invention is safe and well tolerated.

[0273] As shown in Table 21, the CD28 superagonist TGN1412 caused a modest cytokine release and transient increase in T cells (measured on day 15). This study was conducted in male cynomolgus monkeys, with weekly doses of 5–50 mg / kg administered for 4 weeks. FACS analysis of leukocyte subsets was performed in male cynomolgus monkeys after weekly (days 1, 8, 15, and 22) IV injections of increasing doses (5, 10, 25, and 50 mg / kg) of TGN1412. Slight increases in IL-2, IL-5, IL-6, and IFN-γ were observed during the first 2–24 hours (2–20-fold). There were no substantial changes in IL-4 or TNF-α. An exploratory single-dose monkey toxicology study using anti-CD28×PSMA was conducted to establish the safety and pharmacokinetic profile of the anti-CD28×PSMA bispecific antibody.

[0274] In summary, anti-CD28×PSMA antibodies have been shown to enhance immunity at tumor sites. Anti-CD28×PSMA transformed tumor cells into antigen-presenting cells (APCs). Compared to anti-CD28 superagonists, anti-CD28×PSMA antibodies did not induce T cell proliferation or stimulate cytokine release.

[0275] Example 12: Epitope mapping of mAb14226P2 binding to CD28 by hydrogen-deuterium exchange H / D exchange epitope mapping by mass spectrometry (HDX-MS) was performed to determine the amino acid residues of CD28 (recombinant human CD28, SEQ ID NO: 75, designated as hCD28 ecto(N19-P152).mmh) that interact with the anti-hCD28 monoclonal antibody. A general description of the H / D exchange method is described in Ehring (1999) Analytical Biochemistry 267(2):252-259 and Engen and Smith (2001) Anal. Chem. 73:256A-265A. It has been done.

[0276] HDX-MS experiments were performed on an integrated HDX / MS platform, consisting of a Leaptec HDX PAL system for deuterium labeling and quenching, a Waters Acquity M-Class (auxiliary solvent manager) for sample digestion and loading, a Waters Acquity M-Class (μBinary solvent manager) for analytical gradients, and a Thermo Q Exactive HF mass spectrometer for peptide mass measurement.

[0277] The labeling solution was prepared as PBS buffer (10 mM phosphate buffer) in D2O at pH 7.0. Buffer, 140 mM NaCl, and 3 mM KCl, equivalent to pH 7.4 at 25 °C. For deuterium labeling, 11 μL of CD28.mmH (Regeneron in house protein REGN2011, 127 μM) or CD28.mmH (Ag-Ab complex) premixed with REGN5705 at a 1:0.6 molar ratio was added to 44 μL of DO labeling solution at 20 °C. Samples were incubated in duplicate for various time points (e.g., undeuterated control = 0 s, deuterium-labeled = 5 and 10 min). The deuteration reaction was quenched by adding 55 μL of pre-chilled quench buffer (0.5 M TCEP-HCl, 8 M urea, and 1% formic acid) to each sample over a 5-minute incubation at 20 °C. The quenched samples were then injected into a Waters HDX Manager for online pepsin / Protease XIII digestion. The digested peptides were separated on a C8 column (1.0 mm × 50 mm, NovaBioassays) with a 13-minute gradient from 10% to 32% B (mobile phase A: 0.5% formic acid in water, mobile phase B: 0.1% formic acid in acetonitrile). The eluted peptides were analyzed by Q Exactive HF mass spectrometry in LC-MS / MS or LC-MS mode.

[0278] The LC-MS / MS data of the undeuterated CD28 samples were searched against a database containing CD28 and its randomized sequences using the Byonic search engine (Protein Metrics). Search parameters (ELN) were set as default, with nonspecific enzyme digestion and human glycosylation as common variable modifications. The list of identified peptides was then imported into HDX Workbench software (version 3.3) to calculate the deuterium uptake of each peptide detected by LC-MS from all deuterated samples. For a given peptide, the centroid mass (intensity-weighted average mass) at each time point was used to calculate the deuterium uptake (D) and percentage of deuterium uptake (%D).

number

[0279] A total of 73 peptides from hCD28.mmH (SEQ ID NO: 75) were identified from both hCD28.mmH alone and anti-CD28.mmH in complex with anti-CD28 antibody samples, representing 85.8% sequence coverage of hCD28. Any peptide that showed a percent D-incorporation value that differed by more than 5% was defined as significantly protected. For hCD28.mmH (SEQ ID NO: 75), regions corresponding to amino acids 5-20 (VKQSPMLVAYDNAVNL, SEQ ID NO: 77), 29-38 (FSREFRASLH, SEQ ID NO: 78), 80-84 (YLQNL, SEQ ID NO: 79), and 91-108 (IYFCKIEVMYPPPYLDNE, SEQ ID NO: 80) were significantly protected by anti-CD28 antibodies, with amino acids 91-108 (IYFCKIEVMYPPPYLDNE, SEQ ID NO: 80) representing the primary sequence of CD28. The epitopes were defined as follows: Protection of these residues by anti-CD28 antibodies was confirmed using hCD28.mFc (SEQ ID NO: 76). See also Table 22 below for a summary of the results of this study. [Table 22]

[0280] Example 13: PSMA x CD28 bispecific antibody potently enhances the antitumor effect of PD-1 immunotherapy summary T cell activation is enhanced by the engagement of a second costimulatory receptor ("signal 2") in addition to antigen-specific TCR / CD3 activation ("signal 1"). The goal of this approach is to optimally activate and recruit T cells to detect and kill tumor cells. However, current treatments tend not to efficiently and selectively activate T cells at tumor sites, often fail to achieve durable responses, and / or result in undesirable toxicity. Here, a novel tumor-targeted immunotherapy modality is introduced that combines PD-1 blockade with a bispecific antibody. The bispecific antibody binds a tumor-specific antigen (TSA) (e.g., PSMA) in one arm and the costimulatory receptor CD28 on T cells in the other arm. Indeed, PD-1-PD-L1 signaling inhibition significantly increases the ratio of CD28 accumulated in the immune synapse, allowing the TSA x CD28 bispecific antibody to exert its effects. This combination immunotherapy was validated using a bispecific antibody specific for a prostate antigen (e.g., PSMA). Unlike nonspecific CD28 superagonists, which broadly activate T cells, the TSA×CD28 bispecific antibody was well tolerated when used alone or in combination with a PD-1 blocker in genetically humanized immunocompetent mouse models or primates. Importantly, in the presence of endogenous TCR / CD3 triggers, TSA×CD28 significantly improved the antitumor activity of PD-1 antibodies, which was associated with durable antitumor responses. The combination therapy specifically enhanced intratumoral T cell activation and promoted an effector memory-like T cell phenotype without systemic cytokine secretion in various syngeneic and human tumor xenograft models. Combining this class of CD28 costimulatory bispecific antibodies with clinically validated anti-PD-1 treatments may provide significantly enhanced antitumor efficacy to well-tolerated "off-the-shelf" antibody therapies.

[0281] Introduction Numerous monoclonal antibodies (mAbs) aimed at enhancing T cell activation are in clinical development as antitumor therapeutics (M.K. Callahan, M.A. Postow, J.D. Wolchok, Targeting T Cell Co-receptors for Cancer Therapy. Immunity 44, 1069-1078 (2016)). However, most current therapies are challenged by overcoming the suppressive nature of the tumor microenvironment and therefore fail to generate efficient tumor-specific T cell activation and subsequent tumor cell killing (K.G. Anderson, I.M. Stromnes, P.D. Greenberg, Obstacles Posed by the Tumor Microenvironment to T Cell Activity: A Case for Synergistic Therapies. Cancer Cell 31, 311-325 (2017)).Several blocking mAbs directed against checkpoint inhibitors, such as cytotoxic T-lymphocyte-associated protein 4 (CTLA-4) and programmed cell death 1 (PD-1) / programmed cell death ligand 1 (PD-L1), have been clinically approved for melanoma, renal cell carcinoma, non-small cell lung cancer, and advanced metastatic cutaneous squamous cell carcinoma (J.S. Weber et al., Nivolumab versus chemotherapy in patients with advanced melanoma who progressed after anti-CTLA-4 treatment (CheckMate 037): a randomized, controlled, open-label, phase 3 trial. Lancet Oncol 16, 375-384 (2015); S.L. Topalian et al., Survival, durable tumor remission, and long-term safety in patients with advanced melanoma receiving nivolumab. J Clin Oncol 32, 1020-1030 (2014); M.A. Postow, M.K. Callahan, J.D. Wolchok, Immune Checkpoint Blockade in Cancer Therapy.J Clin Oncol 33,1974-1982(2015), MRMigden et al.,PD-1 Blockade with Cemiplimab in Advanced Cutaneous Squamous-C. ell Carcinoma. N Engl J Med 379, 341-351 (2018)). Blocking PD-1 reverses the disruption of T cell activation, but its efficacy as a single agent is often insufficient to achieve tumor clearance and durable antitumor responses. Combining anti-PD-1 / PD-L1 with anti-CTLA-4 has shown high response rates in certain tumor types, but severe toxicity is often observed (J. Larkin et al., Combined Nivolumab and Ipilimumab or Monotherapy in Untreated Melanoma. N Engl J Med 373, 23-34 (2015), DB Johnson et al.,Fulminant Myocarditis with Combination Immune Checkpoint Blockade.N Engl J Med 375, 1749-1755 (2016); MH Pollack et al., Safety of resuming anti-PD-1 in patients with immune-related adverse events (irAEs) during combined anti-CTLA-4 and anti-PD-1 in metastatic melanoma. Ann Oncol 29, 250-255 (2018); JD Wolchok et al., Nivolumab plus ipilimumab in advanced melanoma. N Engl J Med 369, 122-133 (2013)). As a result, considerable efforts are underway to identify patients likely to respond to checkpoint blockade through the identification of biomarkers that predict response (R. Cristescu et al., Pan-tumor genomic biomarkers for PD-1 checkpoint blockade-based immunotherapy. Science 362, (2018)).Furthermore, combination therapies aimed at improving the efficacy and durability of antitumor responses of PD-1 blockade by combining agonist antibodies that induce costimulatory receptors or by combining it with other modalities such as chemotherapy or radiotherapy are undergoing preclinical and clinical evaluation (S. Hu-Lieskovan, A. Ribas, New Combination Strategies Using Programmed Cell Death 1 / Programmed Cell Death Ligand 1 Checkpoint Inhibitors as a Backbone. Cancer J 23, 10-22 (2017); Y.K. Chae et al., Current landscape and future of dual anti-CTLA4 and PD-1 / PD-L1 blockade immunotherapy in cancer; lessons learned from clinical trials with melanoma and non-small cell lung cancer (NSCLC). J Immunother Cancer 6, 39 (2018); P.S.Chowdhury, K. Chamoto, T. Honjo, Combination). therapy strategies for improving PD-1 blockade efficacy: a new era in cancer immunotherapy.J Intern Med 283,110-120(2018), B.Wang et al.,Combination cancer immunotherapy targeting PD-1 and GITR can rescue CD8(+)T cell dysfunction and maintain memory phenotype.Sci Immunol 3,(2018), S.Chen et al.,Combination of 4-1BB agonist and PD-1 antagonist promotes antitumor effector / memory CD8 T cells in a poorly immunogenic tumor model.Cancer Immunol Res 3,149-160(2015)). Microsatellite instability and high mutational burden gene These mutant peptides are expressed as potential endogenous antigens in certain cancers (KW Mouw, MS Goldberg, PA Konstantinopoulos, ADD´ Andrea, DNA Damage and Repair Biomarkers of Immunotherapy Response. Cancer Discov 7, 675-693 (2017)). T cells recognize these mutant peptides as neoantigens (M. Efremova, F. Finotello, D. Rieder, Z. Trajanoski, Neoantigens Generated by Individual Mutations and Their Role in Cancer Immunity and Immunotherapy. Front Immunol 8, 1679 (2017). However, in isolation, presentation of these antigens is not sufficient to promote robust T cell activation to generate antitumor activity. (S. Spranger, R. Bao, T.F. Gajewski, Melanoma-intrinsic beta-catenin signaling prevents antitumor immunity. Nature 523, 231-235 (2015)). This may be due to the immunosuppressive tumor microenvironment.

[0282] Described herein is a novel immunotherapy modality using a TSA×CD28 bispecific antibody targeting the prostate cancer TSA PSMA×CD28. When combined with a PD-1 blocking antibody, it elicited long-lived antitumor immunity, promoted robust intratumoral T cell activation and T cell memory, and showed no signs of systemic cytokine release in animal tumor models. Toxicology studies in genetically humanized immunocompetent mice and cynomolgus monkeys demonstrated that these bispecific antibodies exhibited no toxicity by themselves or in combination with anti-PD-1 antibodies. Collectively, these data suggest that combining this class of CD28-based bispecific antibodies (TSA×CD28) with PD-1 blockade could provide a well-tolerated, "off-the-shelf" biologic solution with significantly enhanced, specific, and synergistic antitumor activity.

[0283] material and method In Example 13, the following materials and methods were used:

[0284] Test Design One exemplary objective of the present invention was to develop a TSA×CD28 bispecific antibody and demonstrate that TSA×CD28 enhances PD-1-induced T cell activation in vitro and safely potentiates antitumor efficacy in vivo. In vitro activity was demonstrated by imaging the bispecific antibody, PD-1, and CD28 localized at the immune synapse of T cell and target cell complexes and demonstrating enhanced PD-1 T cell cytokine release. In vivo antitumor efficacy was assessed in a syngeneic mouse tumor model. Tumor volume and serum cytokines were monitored over time to indicate response to bispecific antibody treatment. One of the objectives of the cynomolgus monkey study was to determine the safety and tolerability (pharmacological and toxicological profiles) of TSA×CD28 as monotherapy or in combination with PD-1 in non-human primates. Animals were examined for toxicity by clinical observation and blood sample collection to analyze serum cytokines and T cell phenotypes.

[0285] Animal experiments All procedures were performed in accordance with the National Institutes of Health (NIH) Guide for the Care and Use of Laboratory Animals. The protocol was approved by the Regeneron Pharmaceuticals Animal Care and Use Committee.

[0286] cell line Jurkat clone E6-1 (ATCC, #TIB-152), Raji (ATCC, #CCL-86™), HEK293 cell line (ATCC, #CRL-1573), and A-431 (ATCC, #CRL-1555™) were cultured according to the protocol recommended by ATCC. HEK293 / hCD20 cell lines were generated using a mammalian vector encoding ubiquitin promoter-driven hCD20 (amino acids M1 to P297 of accession number NP_068769.2). Transfected cells were cultured in 500 μg / ml of geneticin A (G418) and selected for stably expressing cell lines. To generate hCD80- or hPDL1-expressing cells, human CD80 (288 aa long, accession number NM_005191.4) and neomycin were transfected. HEK293T cells were transfected with a cysteine resistance gene or human PDL1 (290 aa long, accession number NM_14143.4) and a puromycin resistance gene to promote viral particle production, which was then used to infect HEK293 / hCD20 or Raji cells. Human CD80 or PDL1-positive cells were isolated by FACS. Jurkat cells were transduced with NFκB-Luc using lentivirus from Qiagen (catalog number CLS-013L) and a lentiviral plasmid encoding human PD-1 and puromycin resistance genes. All generated cell lines were maintained in DMEM + 10% FBS + P / S / G + NEAA supplemented with 500 μg / mL G418 and / or 0.5 μg / mL puromycin.

[0287] The DU145 / hPSMA cell line was generated by transducing DU145 cells (ATCC, HTB-81) with viral particles produced by HEK293T cells transfected with a lentiviral plasmid encoding human PSMA (amino acids M1 to A750 of accession number Q04609) and a neomycin resistance gene. After infection, cells were cultured in 500 μg / mL of geneticin A (G418) to select for cells stably expressing PSMA. The generated cell line, DU145 / PSMA, was maintained in MEM + 10% FBS + P / S / G containing 500 μg / mL of G418.

[0288] To generate engineered tumor cell lines expressing costimulatory ligands, pLVX lentiviral plasmids encoding mouse CD86 or empty vectors, carrying the EF1a promoter and a puromycin resistance gene (pLVX.EF1a.CD86-puro and pLVX.EF1a.EV-puro, respectively), were used to transfect HEK293T cells to drive viral particle production, which was then used to infect MC38 (National Cancer Institute, Laboratory of Tumor Immunology & Biology). Engineered cell lines expressing CD86 were isolated by fluorescence-activated cell sorting (FACS). Cells were maintained under conditions recommended by ATCC in the presence of 0.5 μg / ml puromycin. The resulting cell lines were designated MC38 / CD86 and MC38 / EV.

[0289] To generate MC38 / hPSMA cells, HEK293T cells were transfected with a lentiviral plasmid encoding human PSMA (amino acids M1 to A750 of accession number Q04609) and a neomycin resistance gene to promote the production of viral particles, which were then used to infect MC38 / parental cells. Human PSMA-positive cells were isolated by FACS. MC38 / hPSMA cells were maintained in DMEM + 10% FBS + P / S / G + NEAA supplemented with 500 μg / mL G418.

[0290] Amnis Image Stream Amnis Image Stream was performed as described in Example 7.

[0291] Isolation of human primary CD3+ T cells T cell isolation was performed as described in Example 8.

[0292] Primary CD3 in MLR reactions with DU145 / PSMA cells + IL-2 release from T cells Out Previously isolated and frozen human CD3 + T cells were treated with 50 U / ml benzonase nuclear Cells were thawed on the day of the assay in stimulation medium (X-VIVO 15 cell culture medium supplemented with 10% FBS, HEPES, NaPyr, NEAA, and 0.01 mM BME) containing ATP. Cells were centrifuged at 1200 rpm for 10 min, resuspended in stimulation medium, and diluted to 1 × 10 5 Cells were seeded into 96-well round-bottom plates at a concentration of 1000 cells / well. DU145 parental cells or DU145 cells engineered to 6 cells / mL After incubation at 37°C for 1 hour in 5% CO2, the mitomycin C-treated cells were transferred to a medium containing 2% FBS. Wash three times with D-PBS and add 5 x 10 cells per well. 4 Final concentration of cells: CD3 +The antibodies were added to wells containing T cells. To prevent potential CD28 agonist activity due to the Fc anchor of the CD28 antibody to the Fc receptor, saturating amounts of nonspecific human IgG antibodies (100 nM each, hIgG1, hIgG4, and hIgG4s) were included in each assay well. Subsequently, PSMA x CD28, nontarget x CD28 control, or hIgG4s isotype control antibodies were titrated from 30 pM to 200 nM in a 1:3 dilution and added to the wells. The end point of the 10-point dilution did not contain the titrated antibody. Because DU145 cells endogenously express PD-L1, the effect of PD-1 inhibition on T cell activity was assessed by adding the PD-1 antagonist REGN2810 to the wells at a constant concentration of 20 nM. A condition without PD-1 inhibition was also included; instead, a 20 nM matched hIgG4 isotype control was used. The plates were incubated at 37°C, 5% CO2 for 72 hours, followed by centrifugation. The cells were detached and pelleted. 50 μL of the culture supernatant was collected, from which 5 μL was tested in the human IL-2 AlphaLISA assay according to the manufacturer's protocol. Measurements were taken on a Perkin Elmer Envision multilabel plate reader. A standard curve of known IL-2 concentrations was generated to estimate pg / mL of IL-2 produced in the assay wells. All serial dilutions were tested in duplicate. EC of the antibody 50 Values were determined from a four-parameter logistic equation on a 10-point dose-response curve using GraphPad Prism™ software.

[0293] FACS-based cytotoxicity assay FACS-based cytotoxicity assays were performed as previously described (Example 8).

[0294] Syngeneic tumor studies MC38 / EV and MC38 / CD86 were cultured according to ATCC guidelines. 1 × 10 6 MC38 / EV or MC38 / CD86 were transfected into C57BL / 6 mice. Tumors were implanted subcutaneously. Mice were treated with 5 mg / kg of PD-1 antibody (RPM1-14, BioXcell) or rat IgG2a isotype control (BioXcell) by intraperitoneal injection on days 0, 3, 7, 10, and 14 after tumor implantation. Tumor size was measured twice weekly using a caliper (Roboz RS-6466). Tumor volume was calculated using the formula X*Y*(X / 2), where Y is the longest dimension and X is the perpendicular dimension. 2000 mm 3 Mice with larger tumors or ulcerated tumors were euthanized by CO2 asphyxiation.

[0295] hCD3 / hCD28 / hPSMA humanized mice were generated as described in Example 10. Antibody treatment was also similar, except that the indicated antibody or bispecific antibody was administered as monotherapy or in combination with intraperitoneal injection at 5 mg / kg on days 0, 7, and 14 (prophylactic treatment) or days 9, 13, and 22 (delayed treatment).

[0296] Ex vivo tissue cytokine analysis On day 29 post-implantation, CD3 / CD28 / PSMA mice bearing MC38 / hPSMA tumors were euthanized by carbon dioxide asphyxiation. Spleens and tumors were collected and stored in culture medium on ice. All subsequent steps were performed on ice or at 4°C unless otherwise specified. Tumors were cut into small pieces, and fragments were processed into a single-cell suspension using a Miltenyi Mouse Tumor Dissociation Kit according to the manufacturer's protocol (Miltenyi 130-096-730). Spleens were processed into a single-cell suspension using a gentle MACS mechanical dissociator (Spleen 4 program) and crushed through a 70-micron filter using the rubber end of a 3 ml syringe. Cells were pelleted by centrifugation at 1200 rpm for 5 minutes. The cell pellet was resuspended in 1 ml of ACK lysis buffer and incubated on ice for 5 minutes to lyse red blood cells. The ACK lysis buffer was quenched with FACS buffer. Cells were pelleted by centrifugation at 1200 rpm for 5 minutes. The cell suspension was resuspended in 1 ml of medium, and 0.2 ml was seeded into a 96-well plate (20–400K tumor cells or 50–70K spleen cells). Cells were incubated overnight at 37°C, and the culture supernatant was collected. Cytokine levels in the tissue culture supernatant were measured using the V-Plex Inflammatory MSD Kit according to the manufacturer's protocol (Meso Scale Diagnostics K15048D-4). The number of cells seeded per well was determined by FACS analysis. Cytokine levels were normalized to the number of cells seeded. Calibration beads were run alongside the cells to accurately measure cell number using the following calculation: Number of cells = (number of input beads x number of cells counted by FACS) / number of beads counted by FACS

[0297] Measurement of serum cytokine levels in mice Measurement of serum cytokine levels in mice was performed as previously described (Example 10).

[0298] Flow cytometry analysis For flow cytometry analysis of in vivo experiments, tumors were harvested, single-cell suspensions were prepared, and red blood cells were lysed using ACK lysis buffer (ThermoFisher Scientific). Live / dead cell discrimination was performed using a Live / Dead Fixable Blue Dead Cell Stain Kit (ThermoFisher Scientific). Samples were acquired on a Symphony (BD Bioscience) and analyzed using Cytobank software (Cytobank, Santa Clara, CA). Analysis was performed with equal numbers of events per sample. Event coverage was determined by the sample with the fewest acquired events. CITRUS analysis from Cytobank was used to automatically cluster T cells based on specific markers.

[0299] Toxicology studies in cynomolgus monkeys Cynomolgus monkey studies were performed as previously described (Example 11).

[0300] result PD-1 checkpoint blockade increases the relative abundance of CD28 within the immune synapse, allowing the TSA × CD28 bispecific antibody to significantly enhance the ability of anti-PD-1 to promote T cell activation in vitro. To test whether a costimulatory bispecific agonist can complement checkpoint blockade, an exemplary PSMAxCD28 bispecific antibody of the invention (bs16429D) was tested for its ability to enhance the efficacy of PD-1 blockade in a TCR / CD3-dependent T cell activation mode. Indeed, efficient T cell activation relies on TCR / CD3 interaction at the "immune synapse" (IS). However, activation signals from both TCR / CD3 and CD28 are directly inhibited by PD-1-Shp-2 phosphorylation following PD-1 / PD-L1 clustering at the synapse (E. Hui et al., T cell costimulatory receptor CD28 is a primary target for PD-1-mediated inhibition.Science 355,1428-1433(2017), JMChemnitz,RVParry,KENichols,CHJune,JLRiley,SHP-1 and SHP-2 associate with immunoreceptor tyrosine-based switch motif of programmed death 1 upon primary Human T cell stimulation, but only receptor ligation prevents T cell activation. J Immunol 173, 945-954 (2004)). To determine the relative localization of CD28 and PD-1 at the IS, we developed an in vitro assay, Amnis Image Stream, as described in Example 7, using Jurkat T cells overexpressing PD-1 and Raji tumor target cells engineered to overexpress PD-L1. We also used a fluorescently labeled bispecific CD20 x CD3 antibody (EJ Smith et al., A novel, native-format bispecific antibody triggering T-cell killing of B-cells are robustly active in mouse tumor models and cynomolgus monkeys.Sci Rep 5,17943 (2015)) was used to replicate peptide-MHC / TCR binding and visualize T cell interactions with target cells forming an IS. Two different fluorescently labeled monoclonal PD-1 antibodies (PD-1 mAbs), a blocker (PD-1 mAb, REGN2810 (E. Burova et al., Characterization of the Anti-PD-1 Antibody REGN2810 and Its Antitumor Activity in Human PD-1 Knock-In Mice. Mol Cancer Ther 16, 861-870 (2017)), and a non-blocker (NB PD-1 mAb) were used to simultaneously block the interaction with PD-L1 and visualize PD-1 localization. We found that in the absence of PD-L1 expression on target cells and in the presence of PD-1 mAb, there was no change in the relative amounts of PD-1 or CD28 at the IS (Figure 14). However, PD-L1 expression on target cells in the presence of non-blocking PD-1 mAb reduced CD28 at the synapse and promoted high PD-1 accumulation at the IS. Conversely, blocking PD-1 In the presence of the mAb, PD-1 localization was significantly reduced, while relative CD28 levels were maintained at the IS. PD-1 and CD28 distribution was quantified by calculating the ratio of antibody staining inside to outside the IS (Figure 15). This data indicates that expression of PD-L1 on target cells enhances PD-1 localization at the IS while reducing CD28. Furthermore, PD-1-blocking mAb reduces the relative amount of PD-1 at the IS and rescues CD28 levels, thereby increasing the relative ratio of CD28 to PD-1.

[0301] Next, we tested whether PSMA×CD28 could enhance the effect of PD-1 blockade on tumor cell killing induced by T cell activation. To this end, we utilized the prostate cancer line 22RV1. 22RV1 cells endogenously express PSMA and have been engineered to express PD-L1 (22RV1 / PD-L1). Due to the lack of TCR / CD3 stimulation in allogeneic T cell responses, PSMA×CD3 (U.S. Patent No. 10,179,819) was used to provide primary stimulation. In cocultures of human peripheral blood mononuclear cells (PBMCs, which contain human T cells) with 22RV1 / PD-L1 cells, PSMA×CD3 alone induced approximately 40% tumor cell killing (Figure 16, diamonds, EC2 of 8E-10). 50 Addition of PD-1 mAb to PSMAxCD3 enhanced tumor cell killing by approximately 55% (Figure 16, closed triangles, 4E-10 EC 50 Interestingly, the combination of PSMAxCD28 and PSMAxCD3 similarly enhanced the depth of tumor cell killing to approximately 55%, enhancing efficacy (Figure 16, filled circles, EC of 6E-11). 50 ), suggesting that PSMA×CD28 can override PD-1 / PD-L1-mediated inhibition. Interestingly, the triple combination of PSMA×CD28, PD-1 mAb, and PSMA×C3 showed the strongest tumor cell killing at approximately 70%, indicating the synergistic effect of this combination (Figure 16, black squares, EC of 7E-11). 50 ). As expected, neither PSMAxCD28 alone, PD-1 mAb, nor the combination of PSMAxCD28 and PD-1 mAb induced tumor cell killing (Figure 16, open symbols). Consistently, the greatest increase in IFNγ release was observed with triple combination treatment (Figure 17).

[0302] Next, we determined the effect of combining PSMA×CD28 with PD-1 mAb on primary human T cell activation in vitro. A mixed lymphocyte reaction (MLR) was employed to recapitulate physiological PD-L1 expression and TCR / CD3 stimulation. In a unidirectional MLR, incompatibility of allogeneic determinants leads to T cell activation, which can be quantified by cytokine production. Here, T cells from healthy donors were incubated with DU145 / PSMA cells, an engineered prostate cancer cell line that endogenously expresses PD-L1 and overexpresses PSMA, and the indicated antibodies (Figure 18). In the presence of DU145 / PSMA cells and T cells, the PSMA×CD28 bispecific antibody resulted in a dose-dependent increase in IL-2 release approximately 3-4-fold over the IgG4 isotype control (Figure 18, circle). Similarly, the addition of 20 nM PD-1 mAb to the MLR assay also increased IL-2 release approximately 3-4 fold over the IgG4 isotype control (Figure 18, triangles). In combination, PSMAxCD28 and 20 nM PD-1 mAb significantly enhanced the activation induced by the PSMAxCD28 bispecific antibody, with maximum IL-2 levels increasing approximately 20-fold over the isotype control (Figure 18, squares), demonstrating that the PSMAxCD28 bispecific antibody in combination with the PD-1 blocking mAb potently and synergistically activated T cells in the presence of tumor cells with endogenous levels of TCR / CD3 activation and PD-L1 blockade.

[0303] Overall, these results demonstrated that PSMA×CD28 bispecific antibodies can potently enhance the ability of PD-1 mAbs to promote T cell activation in the presence of TCR / CD3 signaling (driven by CD3 bispecific antibodies or alloreactivity), resulting in increased cytokine release and killing of PSMA- and PD-L1-expressing tumor cells in vitro.

[0304] Overexpression of the natural CD28 ligand on tumor cells synergizes with PD-1 mAb treatment to induce CD8 T cell-dependent durable antitumor immunity in vivo. To determine whether engagement of CD28 by its natural ligand(s) could enhance the antitumor effects of PD-1 mAb in vivo, MC38 tumor cells were engineered to overexpress CD86, one of CD28's costimulatory ligands (data not shown). Combination of MC38 / CD86 cells with PD-1 mAb treatment significantly inhibited tumor growth (Figure 19A) and resulted in complete tumor regression associated with a robust survival benefit when compared to negative control MC38 cells transfected with an empty vector control (MC38 / EV) (Figure 19B). CD8 + T cell depletion enhances PD-1 mAb therapy in MC38 / CD86 cells completely abolished the antitumor effect induced by the combination with CD8 + The results showed a dependency on T cells (Fig. 19C). Tumor-free mice implanted with 86 cells and treated with PD-1 mAb rejected a second MC38 parental tumor implanted more than 60 days after implantation of the primary tumor, indicating the presence of a T cell memory response (Figure 19D). These data suggest that the synergistic effect of constitutively expressed CD28 ligand and anti-PD-1 therapy may result in durable CD8-dependent antitumor immunity in vivo.

[0305] PSMA×CD28 synergizes with PD-1 mAb treatment to induce antitumor immunity in syngeneic tumor models. Next, we extended these findings to demonstrate the antitumor efficacy of treatment with the TSA x CD28 bispecific antibody alone or in combination with a PD-1 mAb in a syngeneic tumor model. Using the established C57BL6 syngeneic MC38 tumor model described herein, we genetically introduced the hPSMA gene (pLVX.EF1a.hPSMA) into MC38 cells to generate the tumor-specific antigen MC38 / hPSMA described herein. To avoid the possibility that mice might spontaneously reject these syngeneic tumors expressing the introduced human tumor antigen, PSMA was genetically humanized in these mice. Furthermore, the CD3γ-δ-ε and CD28 genes were also humanized using VelociGene technology as previously described (DM Valenzuela et al., High-throughput engineering of the mouse genome coupled with high-resolution expression analysis. Nat Biotechnol 21, 652-659 (2003); WT Poueymirou et al., F0 generation mice fully derived from gene-targeted embryonic stem cells allowing immediate phenotypic analyses. Nat Biotechnol 25, 91-99 (2007)). This allows the bispecific antibody to recognize human tumor antigens (via hCD3 or hCD28) on host T cells and in both normal tissues and tumors (i.e., hPSMA), representing a realistic clinical situation (Y. Kinoshita et al., Expression of prostate-specific membrane antigen in normal and malignant human tissues. World J Surg 30, 628-636 (2006)). In this model, the combination of a PSMAxCD28 bispecific antibody and a PD-1 mAb provided the greatest control of tumor growth, which translated into a robust survival benefit.

[0306] Immediate treatment with a combination of anti-CD28 × anti-PSMA and anti-PD1 synergistically inhibits tumor growth In immediate treatment studies, anti-CD28 x anti-PSMA, anti-PD1, or rat IgG2a isotype control were administered as monotherapy or in combination by intraperitoneal injection at 5 mg / kg on days 0, 7, and 14. Caliper measurements of X and Y diameters were used to monitor tumor growth over time. Tumor volume was calculated (X*Y*(X / 2)). Tumors with a tumor size of 2000 mm 3 When the stool concentration exceeded 100 mg / kg, the mice were euthanized.

[0307] As shown in Figures 20A-20E, when used alone, the anti-CD28 x anti-PSMA antibody not only inhibited tumor growth and increased survival in mice, but also synergized with the anti-PD-1 antibody to promote tumor rejection and further increase survival compared to when used in the absence of the PD-1 antibody.

[0308] Treatment with the combination of CD28xPSMA and PD1 induces long-lived antitumor immunity against secondary tumor challenge. Furthermore, tumor-free mice implanted and treated with a combination of PSMAxCD28 bispecific antibody and PD-1 mAb rejected a second MC38 parental tumor line implanted more than 60 days after primary tumor implantation, demonstrating the generation of immune memory consistent with the observations described herein (Figure 21A). These results suggest that an endogenous antigen-specific TCR signal (signal 1) is generated from peptide-MHC complexes on MC38 / PSMA-implanted tumor cells. MC38 tumor cells express high levels of reactivated endogenous retroviral proteins, such as p15E, and intratumoral T cells in C57BL6 mice are responsive to this p15E antigen. It has been shown that (J.C. Yang, D. Perry-Lalley, The The envelope protein of an endogenous murine retrovirus is a tumor-associated T-cell antigen for multiple murine tumors. J Immunother 23, 177-183 (2000); HJ Zeh, 3rd, D. Perry-Lalley, M.E. Dudley, S.A. Rosenberg, J.C. Yang, High avidity CTLs for two self-antigens demonstrate superior in vitro and in vivo antitumor efficacy. J Immunol 162, 989-994 (1999)). Consistent with this finding, we found that combination therapy with PSMA×CD28 and PD-1 mAb induced peripheral T cells reactive to at least one endogenous antigen, P15E, supporting the generation of antitumor immune memory (Figure 21B). Notably, in a similar experiment in which treatment with PSMAxCD28 in combination with PSMAxCD3 induced MC38 / hPSMA tumor rejection, these tumor-free mice failed to reject a secondary tumor rechallenge (Figure 21C). Together, these data demonstrate that CD28 bispecific antibodies can synergize with PD-1 mAb to promote endogenous TCR / CD3-dependent T cell responses in this PSMA syngeneic tumor model.

[0309] Therapeutic treatment with a combination of anti-CD28 × anti-PSMA and anti-PD1 synergistically inhibits tumor growth Similarly, in a delayed treatment protocol, the combination of PSMAxCD28 and PD-1 mAb on day 10 post-implantation inhibited the growth of established MC38 / hPSMA tumors (Figure 22A) and also resulted in a significant increase in survival (Figure 22C) and a reduction in tumor volume (Figure 22D). Interestingly, tumor-targeted combination therapy selectively increased intratumoral cytokines, as demonstrated by IFNγ (Figure 22B). No splenic or systemic cytokine induction was observed in the same mice (Figures 23A and 23B). Furthermore, expression profiling of T cell activation markers supported this combination treatment, demonstrating that intratumoral PD-1 expression was increased with PSMAxCD28 treatment (Figure 24). To further characterize the T cell subsets responding to the combination treatment, tumor-infiltrating CD8 + T cells in a high-dimensional flow T-cells were profiled 17 days after tumor challenge by cytometry. CITRUS (Cluster Identification, Characterization, and Regression) was used to independently stratify statistically significant distinct T-cell clusters. Importantly, PD-1 blockade resulted in the differentiation of effector (CD44 high, CD62L low) CD8 T-cells expressing high levels of activation / exhaustion markers (PD-1, TIM3, LAG3, Ki67). + It was found that T cells (cluster C1) were expanded. (Figure 25). However, only the combination treatment resulted in the expansion of intratumoral CD8+ cells with a memory-like phenotype (high Tcf1, EOMES, CD62L, intermediate CD122, and CD127) and a less exhausted phenotype (low PD-1, LAG3, TIM3, CD38, KLRG1, higher CD5). + It was possible to promote the expansion of T cells (cluster C2) (Figure 25) ( M. Philip et al., Chromatin states define tumor-specific T cell dysfunction and reprogramming. Nature 545, 452-456 (2017). These data demonstrate that combination therapy of a PSMA x CD28 bispecific antibody with anti-PD-1 promotes robust antitumor immunity associated with activation of intratumoral T cells with a memory-like phenotype and a survival benefit.

[0310] TSA×CD28 alone or in combination with PD-1 mAb does not induce systemic T cell activation compared with CD28 superagonists in cynomolgus monkeys. Early clinical trial data showed that a bivalent CD28-activating antibody called a "CD28 superagonist" (CD28-SA) broadly activated T cells and resulted in severe toxicity associated with cytokine release syndrome (CRS) in a group of healthy volunteers ( G. Suntharalingam et al., Cytokine storm in a phase 1 trial of the anti-CD28 monoclonal antibody TGN1412. N Engl J Med 355, 1018-1028 (2006)). To evaluate the tolerability of the TSAxCD28 bispecific alone or the potential for synergistic pharmacology in combination with a PD-1 mAb, exploratory studies were conducted in genetically engineered triple-humanized mice and cynomolgus monkeys. Three monkeys per treatment group received a single dose of PSMAxCD28 (10 mg / kg) by intravenous infusion, either alone or in combination with a PD-1 mAb (REGN2810) (10 mg / kg). The combination group received a continuous infusion (Table 20, Figures 26A-26C, and 27).

[0311] Additionally, three monkeys per group received a single dose (0.1 mg / kg) of anti-PSMAxCD3 bispecific antibody and a single dose (10 mg / kg) of CD28 superagonist antibody, as described herein (Table 20, Figures 26A-26C, and 27). Toxicity assessment was based on clinical observations, qualitative food consumption, body weight, vital signs (temperature, heart rate, pulse oximetry, and respiratory rate), and clinical and anatomical pathology at the completion of the experiment. Blood samples were collected for cytokine and FACS immunophenotypic analysis. PSMAxCD28 alone or in combination with PD-1 was well tolerated, and all animals survived the study period. No test article-related clinical observations were observed (data not shown). No changes in organ weights were observed, and no gross changes were noted at terminal necropsy (data not shown). Furthermore, no significant cytokine release, T cell margination, or activation was observed (Table 23, Figures 26A-26C). In contrast, monkeys treated with the CD28 "superagonist" alone exhibited significant cytokine release, lymphocyte margination, and T cell activation (Figures 29A and 29B). Massive infiltration of immune cells was observed in the kidneys, brain, and seminal vesicles of animals treated with the CD28 superagonist. In contrast, no significant treatment-related histological changes were observed in animals treated with PSMAxCD28 alone or in combination with PD-1 (data not shown). [Table 23-1] [Table 23-2]

[0312] Consistent with the above studies in cynomolgus monkeys, no cytokine elevation was observed in tumor- or non-tumor-bearing naive triple-humanized mice (hCD3 / hCD28 / hPSMA) administered PSMAxCD28 alone or in combination with PD-1 (Figure 27, and Figures 29A and 29B). In contrast, administration of the CD28 superagonist induced significant increases in IFNγ, TNFα, IL-2, IL-4, and IL-5 at 4 hours post-administration (Figure 27, and Figures 29A and 29B). Consistent with the above results, we have previously shown that TSAxCD28 bispecific antibodies (and the parental bivalent non-superagonist CD28 antibody used to generate these bispecific antibodies) failed to induce human T cell proliferation in FDA-recommended in vitro dry-coating and wet-coating assays, compared with the robust proliferation induced by CD28-SA (R. Stebbings, D. Eastwood, S. Poole, R. Thorpe, After TGN1412: recent developments in cytokine release assays. J Immunotoxicol 10, 75-82 (2013)) (data not shown). Overall, these data suggest that TSAxCD28 bispecific antibodies are well tolerated.

[0313] Furthermore, as previously described (Example 10, Figure 28), anti-PSMAxCD28 treatment alone or in combination with anti-PD1 did not elevate serum cytokines in tumor-bearing mice, whereas anti-PSMAxCD3 treatment, alone or in combination with anti-PD1, elevated serum cytokine levels in tumor-bearing mice.

[0314] Consideration Presented and validated herein is a novel tumor-targeted immunotherapy using a TSA×CD28 bispecific antibody in combination with a PD-1-blocking mAb, which induces long-lived antitumor immunity and promotes robust intratumoral T cell activation in animal tumor models. Toxicology studies in genetically humanized immunocompetent mice and cynomolgus monkeys demonstrated no toxicity of these bispecific antibodies alone or in combination with a PD-1 mAb, suggesting that this therapeutic approach may offer a well-tolerated, "off-the-shelf" biologic solution with significantly enhanced, specific, and synergistic antitumor activity.

[0315] Although checkpoint inhibition by PD-1-blocking mAbs is known to reverse the interruption of T cell activation, their efficacy as single agents is often insufficient to achieve tumor clearance and durable antitumor responses in many cancers. Several approaches to improve response rates to PD-1 blockade are currently being evaluated. Indeed, PD-1 Identification of biomarkers to predict responsiveness to mAbs (R. Cristescu et al., Pan-tumor genomic biomarkers for PD-1 checkpoint blockade-based immunotherapy. Science 362 (2018)), non-tumor targeted combination therapies using PD-1 blockade together with agonist antibodies that trigger costimulatory receptors and improve T cell activation, or together with chemotherapy or radiotherapy, are all currently undergoing preclinical and clinical trials (S. Hu-Lieskovan, A. Ribas, New Combination Strategies Using Programmed Cell Death 1 / Programmed Cell Death Ligand 1 Checkpoint Inhibitors as a Backbone. Cancer J 23, 10-22 (2017), Y.K. Chae et al., Current landscape and future of dual anti-CTLA4 and PD-1 / PD-L1 blockade immunotherapy in cancer;lessons learned from clinical t trials with melanoma and non-small cell lung cancer (NSCLC). J Immunother Cancer 6, 39 (2018); P.S.Chowdhury, K. Chamoto, T. Honjo, Combination therapy strategies for improving PD-1 blockade efficacy: a new era in cancer immunotherapy. J Intern Med 283, 110-120 (2018). However, the challenge is that many of these combinations are often based on the availability of existing drugs and post hoc rationales for combining treatments rather than a truly hypothesis-driven approach, sometimes resulting in worse patient outcomes (M.J.Ahn, J.M.Sun, S.H.Lee, J.S.Ahn, K.Park, EGFR TKI combination with immunotherapy in non-small cell lung cancer. Expert Opin Drug Saf 16, 465-469 (2017)). It is clear that checkpoint inhibition and reactivation of the immune system offers the possibility of long-term remission in many patients (J.S. Weber et al., Nivolumab versus chemotherapy in patients with advanced melanoma who progressed after anti-CTLA-4 treatment (CheckMate 037): a randomized, controlled, open-label, phase 3 trial. Lancet Oncol 16, 375-384 (2015); S.L. Topalian et al., Survival, durable tumor remission, and long-term safety in patients with advanced melanoma receiving nivolumab.J Clin Oncol 32, 1020-1030 (2014); MA Postow, MK Callahan, JD Wolchok, Immune Checkpoint Blockade in Cancer Therapy. J Clin Oncol 33, 1974-1982 (2015); MR Migden et al., PD-1 Blockade with Cemiplimab in Advanced Cutaneous Squamous-Cell Carcinoma. N Engl J Med 379, 341-351 (2018)). Therefore, methods to further improve or enhance T cell activity and promote more durable responses are warranted. Here, we introduce the concept of enhancing T cell signaling and activation using a TSA x CD28 bispecific antibody to improve the antitumor efficacy of PD-1 mAbs. Indeed, this novel combination immunotherapy was validated using tumor targets (e.g., PSMA) and demonstrated that the CD28 costimulatory bispecific antibody synergized with the PD-1 mAb to not only generate robust T cell activation but also provide durable antitumor responses without systemic toxicity. As a result, this tumor-targeted combination therapy may offer considerable advantages over previously described non-targeted approaches.The use of CD28 bispecific antibodies, which do not directly activate CD28 unless clustered on the tumor cell surface, offers the possibility of promoting costimulation only at the tumor site, thus avoiding the systemic toxicity of conventional CD28-activating antibodies (G. Suntharalingam et al., Cytokine storm in a phase 1 trial of the anti-CD28 monoclonal antibody TGN1412. N Engl J Med 355, 1018-1028 (2006)), the toxicity often observed with the combination of CLTA-4 and PD-1 blockade (J. Larkin et al., Combined Nivolumab and Ipilimumab or Monotherapy in Untreated Melanoma. N Engl J Med 373, 23-34 (2015); D.B. Johnson et al., Fulminant Myocarditis with Combination Immune Checkpoint Blockade. N Engl J Med 375, 1749-1755 (2016); M.H. Po. llack et al., Safety of resuming anti-PD-1 in patients with immune-related adverse events (irAEs) during combined anti-CTLA-4 and anti-PD-1 in metastatic melanoma. Ann Oncol 29, 250-255 (2018)), or other costimulatory agonist bivalent antibodies (N.H. Segal et al., Results from an Integrated Safety Analysis of Urelumab, an Agonist Anti-CD137 Monoclonal Antibody. Clin Cancer Res 23, 1929-1936 (2017)). Toxicology studies in genetically humanized immunocompetent mice and cynomolgus monkeys have shown that these bispecific antibodies exhibit no toxicity as single agents or in combination with PD-1 mAbs. The safety profile, together with the enhanced antitumor efficacy of our anti-PSMA×CD28 bispecific antibody with PD-1 mAb in syngeneic models, suggests that this treatment is robust and may have broader utility as a novel combination class of immunotherapy.

[0316] Tumor-targeted approaches are being developed to enhance T cell-mediated tumor cell killing (E. Dahlen, N. Veitonmaki, P. Norlen, Bispecific antibodies in cancer immunotherapy. Ther Adv Vaccines Immunother 6, 3-17 (2018)). Indeed, CD3-based bispecific antibodies represent a new class of antibodies that can efficiently trigger T cell activation by binding T cells to tumor cells and activating TCR / CD3 (E. J Smith et al., A novel, native-format bispecific antibody triggering T cell killing of B cells is robustly active in mouse tumor models and cynomolgus monkeys. Sci Rep 5, 17943 (2015)), thus mimicking the normal "signal 1" signal. However, despite their promising clinical efficacy, CD3 bispecific antibodies may be associated with cytokine release syndrome (CRS) due to their lack of direct T cell activation and tumor-specific specificity (SL Maude, D. Barrett, DT Teachey, S.A. Rupp, Managing cytokine release syndrome associated with novel T cell-engaging therapies. Cancer J 20, 119-122 (2014)). Here, we demonstrate for the first time that combination therapy with a TSA × CD28 bispecific antibody and a PD-1 mAb induces tumor-specific T cell activation associated with long-term memory responses in immunocompetent mouse tumor models.TSA x CD28 bispecific antibodies have limited or no activity in the absence of "signal 1," and PD-1 blockade depends on endogenous antigen-specific T cell responses to tumor peptides (W. Hugo et al., Genomic and Transcriptomic Features of Response to Anti-PD-1 Therapy in Metastatic Melanoma. Cell 165, 35-44 (2016); N. Arizvi et al., Cancer immunology. Mutational landscape determines sensitivity to PD-1 blockade in non-small cell lung cancer). cell lung cancer.Science 348,124-128(2015), JMMehnert et al.,Immune activation and response to pembrolizumab in POLE-mutant endometrial cancer.J Clin Invest 126,2334-2340(2016), DTLe et al.,Mismatch repair deficiency predicts response of Solid tumors to PD-1 blockade. Science 357, 409-413 (2017). Therefore, the "signal 1" provided by endogenous tumor antigens is important for the combination therapy of PSMA×CD28 and PD-1 mAb. This contrasts with CD3 bispecific antibodies, which activate T cells independently of TCR specificity and therefore may not generate long-lived tumor-specific immunity. Indeed, we found that combined treatment with PSMA×CD3 and PSMA×CD28 induced potent antitumor effects but did not induce a strong memory response. Furthermore, MC38 tumor cells express high levels of reactivated endogenous retroviral peptides, such as p15E, and C57BL6 mice have been shown to generate endogenous T cells that recognize and respond to this neoepitope (JC Yang, D. Perry-Lalley, The envelope protein of an endogenous murine retrovirus is a tumor-associated T-cell antigen for multiple murine tumors. J Immunother 23, 177-183 (2000); HJ Zeh, 3rd, D. Perry-Lalley, M. Eudley, S. A. Rosenberg, JC Yang, High avidity CTLs for two self-antigens demonstrate superior in vitro and in vivo antitumor efficacy. J Immunol 162, 989-994 (1999)). In the MC38 model described herein, combination therapy with PSMA×CD28 and PD-1 mAb demonstrated increased numbers of T cells responding to this p15E neoantigen. Furthermore, extensive profiling of tumor-infiltrating T cells revealed that the combination of PSMA×CD28 and PD-1 mAb resulted in less dysfunctional CD8 T cells and promoted a potent intratumor memory T cell phenotype. Thus, CD28 bispecific antibodies, together with PD-1 blockade, promote endogenous TCR / CD3-dependent T cell responses and promote durable antitumor responses.

[0317] The data herein demonstrate that when PD-L1 is expressed by target cells, PD-1 accumulates at the immune synapse, and this accumulation is associated with a reduction in CD28 at the synapse, suggesting that PD-1 can exert T cell inhibition by preventing CD28 localization at the synapse. Furthermore, we found that PD-1 blockade prevented PD-1 synaptic localization while increasing CD28 accumulation at the synapse, and that the TSA×CD28 bispecific antibody significantly enhanced the ability of PD-1 mAb to promote T cell activation. This may be one of the mechanisms by which PD-1 blocking antibodies promote T cell activation. Overall, visualization of PD-1-PD-L1 interaction and / or the localization of PD-1 and CD28 at the immune synapse after PD-1 inhibition can better understand the effects of PD-1 blockade on T cell activation, as well as the synergistic effects between TSA×CD28 and PD-1 mAb at the level of the immune synapse.

[0318] PD-1 mAbs are an important new class of immunotherapy, but further optimization of their antitumor activity will certainly be crucial in many cases. Just as the CAR-T approach employs chimeric receptors that artificially activate both "signal 1" and "signal 2" to improve antitumor activity (E.A.Zhukovsky, R.J.Morse, M.V.Maus, Bispecific antibodies and CARs: generalized immunotherapeutics harnessing T cell redirection. Curr Opin Immunol 40, 24-35 (2016); A.I.Salter et al., Phosphoproteomic analysis of chimeric antigen receptor signaling reveals kinetic and quantitative differences that affect cell function. Sci Signal 11, (2018)), PD-1 blockade can be combined with CD28 bispecific antibodies (" The potential benefit of combining CD28-based bispecific antibodies with a CD28-based mAb (which provides a signal 2) to enhance antitumor activity has been demonstrated. This approach offers several practical advantages over CAR-T therapy, including the elimination of laborious cell therapy preparation, which must be individually customized for each patient, and the elimination of the need to preemptively "lymphodeplete" patients with toxic chemotherapy, which is often associated with adverse effects that prevent patients from accepting cell therapy (CH June, RSO Connor, OU Kawalekar, S. Ghassemi, M.C. Milone, CAR T cell immunotherapy for human cancer. Science 359, 1361-1365 (2018)). This bispecific approach offers the potential for improved efficacy and safety through its specificity of action. Collectively, these data suggest that combining a CD28-based bispecific antibody with a clinically validated PD-1 mAb, such as cemiplimab, could provide a well-tolerated, "off-the-shelf" biologic solution with significantly enhanced and synergistic antitumor activity.

[0319] The present invention is not to be limited in scope by the specific embodiments described herein. Indeed, var...

Claims

1. 1. An isolated bispecific antigen-binding molecule comprising: a) Approximately 10 as measured by surface plasmon resonance at 25°C -6 K less than M D Human CD2 a first antigen-binding domain (D1) that binds to 8; b) about 10 as measured by surface plasmon resonance at 25°C -9 K less than M D And target tumor and a second antigen-binding domain (D2) that specifically binds to human prostate-specific membrane antigen (PSMA) on a cell.

2. The bispecific antigen-binding molecule has a binding affinity of about 10 -6 EC less than M 50 The isolated bispecific antigen-binding molecule of claim 1, which binds to the surface of a human T cell at

3. The bispecific antigen-binding molecule has a binding affinity of about 10 -7 EC less than M 50 The isolated bispecific antigen-binding molecule of claim 1, which binds to the surface of cynomolgus monkey T cells at

4. The bispecific antigen-binding molecule has a binding affinity of about 10 -8 EC less than M 50 The isolated bispecific antigen-binding molecule of claim 1, which binds to the surface of a cell line expressing PSMA.

5. 2. The isolated bispecific antigen-binding molecule of claim 1, wherein the bispecific antigen-binding molecule exhibits a costimulatory effect when used in combination with an anti-PSMA x CD3 bispecific antibody and tested on target cells expressing PSMA.

6. 6. The isolated bispecific antigen-binding molecule of claim 5, wherein the costimulatory effect is demonstrated by one or more of the following: (a) the ability to activate and induce human T cells to kill target cells expressing PSMA, (b) the ability to upregulate PD-1 on T cells, (c) the ability to increase the release of the cytokines IFNγ and TNF from PBMCs, (d) the ability to deplete tumor cells, or (f) the ability to enhance tumor clearance.

7. The costimulatory effect is determined by: (g) activation of NFκB activity in a T cell / APC luciferase-based reporter assay; or (h) activation of primary CD4 + T cell / APC function 7. The isolated bispecific antigen-binding molecule of claim 6, further demonstrated by one or more of the following: measurement of IL-2 cytokine production using an IL-2 cytokine assay.

8. The bispecific antigen-binding molecule has an EC50 binding affinity of about 5.0 nM to about 10 nM as measured by an electrochemiluminescence detection platform. 50 8. The isolated bispecific antigen-binding molecule of claim 1, which specifically binds to cells engineered to express CD28 in vivo.

9. The bispecific antigen-binding molecule has an EC50 binding affinity of about 0.3 nM to about 5.0 nM as measured by an electrochemiluminescence detection platform. 50 8. The isolated bispecific antigen-binding molecule of claim 1, which specifically binds to a human epithelial prostate cancer cell line at 200 ng / mL.

10. The isolated bispecific antigen-binding molecule of any one of claims 1 to 9, wherein the target tumor cell is a prostate cancer cell.

11. the first antigen-binding domain (D1) comprises: a) a heavy chain comprising an amino acid sequence selected from the group consisting of SEQ ID NOs: 10, 34, and 58; three heavy chain complementarity determining regions (HCDR1, HCDR2, and HCDR3) contained within the heavy chain variable region (HCVR); and b) three light chain complementarity determining regions (LCDR1, LCDR2, and LCDR3) contained within a light chain variable region (LCVR) comprising an amino acid sequence selected from the group consisting of SEQ ID NOs: 18, 42, and 66.

12. 12. The isolated bispecific antigen-binding molecule of claim 11, comprising: a) an HCDR1 comprising an amino acid sequence selected from the group consisting of SEQ ID NOs: 12, 36, and 60; an HCDR2 comprising an amino acid sequence selected from the group consisting of SEQ ID NOs: 14, 38, and 62; and an HCDR3 comprising an amino acid sequence selected from the group consisting of SEQ ID NOs: 16, 40, and 64.

13. 13. The isolated bispecific antigen-binding molecule of claim 12, comprising: a) an LCDR1 comprising an amino acid sequence selected from the group consisting of SEQ ID NOs: 20, 44, and 68; an LCDR2 comprising an amino acid sequence selected from the group consisting of SEQ ID NOs: 22, 46, and 70; and an LCDR3 comprising an amino acid sequence selected from the group consisting of SEQ ID NOs: 24, 48, and 72.

14. the first antigen-binding domain comprises:

12. The isolated bispecific antigen-binding molecule of claim 11, comprising: a) a set of HCVR CDRs (HCDR1, HCDR2, HCDR3) comprising the amino acid sequences selected from the group consisting of SEQ ID NOs: 12, 14, 16; 36, 38, 40; 60, 62, and 64; and a set of LCVR CDRs (LCDR1, LCDR2, LCDR3) comprising the amino acid sequences of SEQ ID NOs: 20, 22, 24; 44, 46, 48; and 68, 70, 72.

15. 12. The isolated bispecific antigen-binding molecule of claim 11, wherein the first antigen-binding domain comprises an HCVR / LCVR pair comprising an amino acid sequence selected from the group consisting of SEQ ID NOs: 10 / 18, 34 / 42, and 58 / 66.

16. the second antigen-binding domain comprises: a) three HCDRs contained within the HCVR comprising an amino acid sequence selected from the group consisting of SEQ ID NOs: 2, 26, and 50; and b) three LCDRs comprised within an LCVR comprising an amino acid sequence selected from the group consisting of SEQ ID NOs: 18, 42, and 66.

17. the second antigen-binding domain comprises: a) an HCDR1 comprising an amino acid sequence selected from the group consisting of SEQ ID NOs: 4, 28, and 52; b) an HCDR2 comprising an amino acid sequence selected from the group consisting of SEQ ID NOs: 6, 30, and 54; and c) an HCDR3 comprising an amino acid sequence selected from the group consisting of SEQ ID NOs: 8, 32, and 56.

18. the second antigen-binding domain comprises: a) an LCDR1 comprising an amino acid sequence selected from the group consisting of SEQ ID NOs: 20, 44, and 68; an LCDR2 comprising an amino acid sequence selected from the group consisting of SEQ ID NOs: 22, 46, and 70; and an LCDR3 comprising an amino acid sequence selected from the group consisting of SEQ ID NOs: 24, 48, and 72. and an LCDR3 comprising the sequence:

19. the second antigen-binding domain comprises:

19. The isolated bispecific antigen-binding molecule of claim 18, comprising: a) a set of HCVR CDRs (HCDR1, HCDR2, HCDR3) comprising amino acid sequences selected from the group consisting of SEQ ID NOs: 4, 6, 8; 28, 30, 32; and 52, 54, 56; and a set of LCVR CDRs (LCDR1, LCDR2, LCDR3) comprising amino acid sequences selected from the group consisting of SEQ ID NOs: 20, 22, 24; 44, 46, 48; and 68, 70, 72.

20. a) a first antigen-binding domain comprising HCVR CDRs comprising the amino acid sequences of SEQ ID NOs: 12, 14, and 16, and LCVR CDRs comprising the amino acid sequences of SEQ ID NOs: 20, 22, and 24; and b) a second antigen-binding domain comprising an HCVR CDR comprising the amino acid sequence of SEQ ID NO: 4, 6, or 8, and an LCVR CDR comprising the amino acid sequence of SEQ ID NO: 20, 22, or 24.

21. a) a first antigen-binding domain comprising an HCDR comprising the amino acid sequence of SEQ ID NO: 36, 38, or 40, and an LCDR comprising the amino acid sequence of SEQ ID NO: 44, 46, or 48; and b) a second antigen-binding domain comprising an HCDR comprising the amino acid sequence of SEQ ID NO: 28, 30, 32, and an LCDR comprising the amino acid sequence of SEQ ID NO: 44, 46, 48.

22. a) a first antigen-binding domain comprising an HCDR comprising the amino acid sequence of SEQ ID NO: 60, 62, 64 and an LCDR comprising the amino acid sequence of SEQ ID NO: 68, 70, 72; and b) a second antigen-binding domain comprising an HCDR comprising the amino acid sequence of SEQ ID NO: 52, 54, 56, and an LCDR comprising the amino acid sequence of SEQ ID NO: 68, 70, 72.

23. a) a first antigen-binding domain comprising an HCVR / LCVR pair comprising the amino acid sequence of SEQ ID NO: 10 / 18; and b) a second antigen-binding domain comprising an HCVR / LCVR pair comprising the amino acid sequence of SEQ ID NO: 2 / 18.

24. a) a first antigen-binding domain comprising an HCVR / LCVR pair comprising the amino acid sequences of SEQ ID NOs: 34 / 42; and b) a second antigen-binding domain comprising an HCVR / LCVR pair comprising the amino acid sequences of SEQ ID NOs: 26 / 42.

25. a) a first antigen-binding domain comprising an HCVR / LCVR pair comprising the amino acid sequence of SEQ ID NO: 58 / 66; and b) a second antigen-binding domain comprising an HCVR / LCVR pair comprising the amino acid sequences of SEQ ID NOs: 50 / 66.

26. compete for binding to PSMA or bind to the same epitope on PSMA as the reference antibody 1. An isolated bispecific antigen-binding molecule that binds to a reference antibody, wherein the reference antibody comprises a first antigen-binding domain having an HCVR / LCVR pair comprising the amino acid sequence of SEQ ID NO: 10 / 18, 34 / 42, or 58 / 66, and a second antigen-binding domain having an HCVR / LCVR pair comprising the amino acid sequence of any of SEQ ID NOs: 2 / 18, 26 / 42, or 50 / 66.

27. An isolated bispecific antigen-binding molecule that competes for binding to human CD28 or binds to the same epitope on human CD28 as a reference antibody, wherein the reference antibody comprises a first antigen-binding domain having an HCVR / LCVR pair comprising the amino acid sequence of SEQ ID NO: 10 / 18, 34 / 42, or 58 / 66, and a second antigen-binding domain having an HCVR / LCVR pair comprising the amino acid sequence of any of SEQ ID NOs: 2 / 18, 26 / 42, or 50 / 66.

28. A pharmaceutical composition comprising the bispecific antigen-binding molecule of any one of claims 1 to 27 and a pharmaceutically acceptable carrier or diluent.

29. A nucleic acid comprising a nucleotide sequence encoding the bispecific antibody of any one of claims 1 to 27.

30. 30. An expression vector comprising the nucleic acid of claim 29.

31. A host cell comprising the expression vector of claim 30.

32. 29. A method of inhibiting the growth of a prostate cell tumor in a subject, comprising administering to the subject the isolated bispecific antibody of any one of claims 1 to 27 or the pharmaceutical composition of claim 28.

33. 33. The method of claim 32, further comprising administering a second therapeutic agent.

34. 34. The method of claim 33, wherein the second therapeutic agent comprises an anti-tumor agent, radiation therapy, an antibody drug conjugate, a bispecific antibody conjugated to an anti-tumor agent, a checkpoint inhibitor, or a combination thereof.

35. 29. A method of treating a patient suffering from prostate cancer or another PSMA-expressing cellular malignancy, comprising administering to the subject the isolated bispecific antibody of any one of claims 1-27 or the pharmaceutical composition of claim 28.

36. 36. The method of claim 35, further comprising administering a second therapeutic agent.

37. 37. The method of claim 36, wherein the second therapeutic agent comprises an anti-tumor agent, radiation therapy, an antibody drug conjugate, a bispecific antibody conjugated to an anti-tumor agent, a checkpoint inhibitor, or a combination thereof.

38. 37. The method of any one of claims 33 or 36, wherein the second therapeutic agent is a different bispecific antibody comprising a first antigen-binding domain that binds to the same tumor target antigen and a second antigen-binding domain that binds to CD3 on T cells.

39. 39. The method of any one of claims 33-38, wherein the second therapeutic agent is a checkpoint inhibitor.

40. 40. The method of claim 39, wherein the checkpoint inhibitor targets PD-1 or CTLA4.

41. 41. The method of claim 40, wherein the PD-1 inhibitor is selected from the group consisting of pembrolizumab (Keytruda), nivolumab (Opdivo), atezolizumab (Tenectriq), and cemiplimab.

42. 41. The method of claim 40, wherein the CTLA4 inhibitor is ipilimumab.

43. 43. The method of any one of claims 32 to 42, wherein the method activates T cells and induces them to kill tumor target cells without inducing a cytokine storm.

44. 28. The isolated bispecific antigen-binding molecule of any one of claims 1 to 27, wherein the antigen-binding molecule induces T-cell-mediated cytotoxicity of human prostate cells.

45. A bispecific antigen-binding molecule comprising a first antigen-binding domain that specifically binds to human CD28 and a second antigen-binding domain that specifically binds to human PSMA.

46. The antigen-binding molecule is 1 x 10 -12 M~1 x 10 -6 M's EC 50 At this value, CD28 expression 46. The bispecific antigen-binding molecule of claim 45, which binds to human T cells.

47. The antigen-binding molecule is 1 x 10 -9 M~1 x 10 -6 EC 50 47. The bispecific antigen-binding molecule of claim 46, which binds to human T cells expressing CD28 at a value of

48. The bispecific antigen-binding molecule of any one of claims 45 to 47, wherein the antigen-binding molecule binds to human cells expressing human CD28 and cynomolgus monkey cells expressing cynomolgus monkey CD28.

49. The bispecific antigen-binding molecule of any one of claims 45 to 47, wherein the antigen-binding molecule induces proliferation of human and cynomolgus monkey peripheral blood mononuclear cells (PBMCs) in vitro.

50. The bispecific antigen-binding molecule of any one of claims 45 to 47, wherein the antigen-binding molecule induces cytokine release and CD25 upregulation in human whole blood.

51. The bispecific antigen-binding molecule of any one of claims 45 to 47, wherein the antigen-binding molecule induces T-cell-mediated cytotoxicity of human prostate cells.

52. The bispecific antigen-binding molecule of any one of claims 45 to 51, wherein the first antigen-binding domain that specifically binds to human CD28 comprises the heavy chain complementarity-determining regions (HCDR1, HCDR2, and HCDR3) derived from a heavy chain variable region (HCVR) comprising an amino acid sequence selected from the group consisting of SEQ ID NOs: 10, 34, and 58, and the light chain complementarity-determining regions (LCDR1, LCDR2, and LCDR3) derived from a light chain variable region (LCVR) comprising an amino acid sequence selected from the group consisting of SEQ ID NOs: 18, 42, and 66.

53. The second antigen-binding domain that specifically binds to human PSMA comprises the heavy chain complementarity-determining regions (HCDR1, HCDR2, and HCDR3) derived from a heavy chain variable region (HCVR) comprising SEQ ID NOs: 2, 26, and 50, and the light chain complementarity-determining regions (LCVR) derived from a light chain variable region (LCVR) comprising an amino acid sequence selected from the group consisting of SEQ ID NOs: 18, 42, and 66. (LCDR1, LCDR2, and LCDR3).

54. 52. The bispecific antigen-binding molecule of any one of claims 45 to 51, wherein the first antigen-binding domain that specifically binds to human CD28 comprises three heavy chain complementarity determining regions (HCDR1, HCDR2, and HCDR3) and three light chain complementarity determining regions (LCDR1, LCDR2, and LCDR3), wherein HCDR1 comprises an amino acid sequence selected from the group consisting of SEQ ID NOs: 12, 36, and 60, HCDR2 comprises an amino acid sequence selected from the group consisting of SEQ ID NOs: 14, 38, and 62, and HCDR3 comprises an amino acid sequence selected from the group consisting of SEQ ID NOs: 16, 40, and 64, wherein LCDR1 comprises an amino acid sequence selected from the group consisting of SEQ ID NOs: 20, 44, and 68, wherein LCDR2 comprises an amino acid sequence selected from the group consisting of SEQ ID NOs: 22, 46, and 70, and LCDR3 comprises an amino acid sequence selected from the group consisting of SEQ ID NOs: 24, 48, and 72.

55. 52. The bispecific antigen-binding molecule of claim 45, wherein the second antigen-binding domain that specifically binds to human PSMA comprises three heavy chain complementarity determining regions (HCDR1, HCDR2, and HCDR3) and three light chain complementarity determining regions (LCDR1, LCDR2, and LCDR3), wherein HCDR1 comprises an amino acid sequence selected from the group consisting of SEQ ID NOs: 4, 28, and 52; HCDR2 comprises an amino acid sequence selected from the group consisting of SEQ ID NOs: 6, 30, and 54; HCDR3 comprises an amino acid sequence selected from the group consisting of SEQ ID NOs: 8, 32, and 56; LCDR1 comprises an amino acid sequence selected from the group consisting of SEQ ID NOs: 20, 44, and 68; LCDR2 comprises an amino acid sequence selected from the group consisting of SEQ ID NOs: 22, 46, and 70; and LCDR3 comprises an amino acid sequence selected from the group consisting of SEQ ID NOs: 24, 48, and 72.

56. the first antigen-binding domain that specifically binds to human CD28 comprises three heavy chain complementarity-determining regions (HCDR1, HCDR2, and HCDR3) and three light chain complementarity-determining regions (LCDR1, LCDR2, and LCDR3); and the second antigen-binding domain that specifically binds to human PSMA comprises three heavy chain complementarity-determining regions (HCDR1, HCDR2, and HCDR3) and three light chain complementarity-determining regions (LCDR1, LCDR2, and LCDR3); the first antigen-binding domain comprises an HCDR1 comprising an amino acid sequence selected from the group consisting of SEQ ID NOs: 12, 36, and 60, an HCDR2 comprising an amino acid sequence selected from the group consisting of SEQ ID NOs: 14, 38, and 62, an HCDR3 comprising an amino acid sequence selected from the group consisting of SEQ ID NOs: 16, 40, and 64, an LCDR1 comprising an amino acid sequence selected from the group consisting of SEQ ID NOs: 20, 44, and 68, an LCDR2 being an amino acid sequence selected from the group consisting of SEQ ID NOs: 22, 46, and 70, and an LCDR3 comprising an amino acid sequence selected from the group consisting of SEQ ID NOs: 24, 48, and 72; 52. The bispecific antigen-binding molecule of claim 45, wherein the second antigen-binding domain comprises HCDR1 comprising an amino acid sequence selected from the group consisting of SEQ ID NOs: 4, 28, and 52; HCDR2 comprising an amino acid sequence selected from the group consisting of SEQ ID NOs: 6, 30, and 54; HCDR3 comprising an amino acid sequence selected from the group consisting of SEQ ID NOs: 8, 32, and 56; LDCR1 comprising an amino acid sequence selected from the group consisting of SEQ ID NOs: 20, 44, and 68; LDCR2 comprising an amino acid sequence selected from the group consisting of SEQ ID NOs: 22, 46, and 70; and LCDR3 comprising an amino acid sequence selected from the group consisting of SEQ ID NOs: 24, 48, and 72.

57. The first antigen-binding domain binds to three heavy chain complementarity determinants for binding to human CD28.

52. The bispecific antigen-binding molecule of any one of claims 45 to 51, wherein the bispecific antigen-binding molecule competes with a reference antigen-binding protein comprising three light chain complementarity determining regions (HCDR1, HCDR2, and HCDR3) and three light chain complementarity determining regions (LCDR1, LCDR2, and LCDR3), wherein HCDR1 comprises an amino acid sequence selected from the group consisting of SEQ ID NOs: 12, 36, and 60, HCDR2 comprises an amino acid sequence selected from the group consisting of SEQ ID NOs: 14, 38, and 62, and HCDR3 comprises an amino acid sequence selected from the group consisting of SEQ ID NOs: 16, 40, and 64, wherein LCDR1 comprises an amino acid sequence selected from the group consisting of SEQ ID NOs: 20, 44, and 68, LCDR2 comprises an amino acid sequence selected from the group consisting of SEQ ID NOs: 22, 46, and 70, and LCDR3 comprises an amino acid sequence selected from the group consisting of SEQ ID NOs: 24, 48, and 72.

58. 52. The bispecific antigen-binding molecule of any one of claims 45 to 51, wherein the first antigen-binding domain competes for binding to human CD28 with a reference antigen-binding protein comprising a heavy chain variable region (HCVR) comprising an amino acid sequence selected from the group consisting of SEQ ID NOs: 10, 34, and 58, and a light chain variable region (LCVR) comprising an amino acid sequence selected from the group consisting of SEQ ID NOs: 18, 42, and 66.

59. 52. The bispecific antigen-binding molecule of claim 45, wherein the second antigen-binding domain competes for binding to human PSMA with a reference antigen-binding protein comprising three heavy chain complementarity determining regions (HCDR1, HCDR2, and HCDR3) and three light chain complementarity determining regions (LCDR1, LCDR2, and LCDR3), wherein HCDR1 comprises an amino acid sequence selected from the group consisting of SEQ ID NOs: 4, 28, and 52; HCDR2 comprises an amino acid sequence selected from the group consisting of SEQ ID NOs: 6, 30, and 54; and HCDR3 comprises an amino acid sequence selected from the group consisting of SEQ ID NOs: 8, 32, and 56; LCDR1 comprises an amino acid sequence selected from the group consisting of SEQ ID NOs: 20, 44, and 68; LCDR2 comprises an amino acid sequence selected from the group consisting of SEQ ID NOs: 22, 46, and 70; and LCDR3 comprises an amino acid sequence selected from the group consisting of SEQ ID NOs: 24, 48, and 72.

60. 52. The bispecific antigen-binding molecule of any one of claims 45 to 51, wherein the second antigen-binding domain competes for binding to human PSMA with a reference antigen-binding protein comprising heavy chain variable regions (HCVRs) 2, 26, and 50 and a light chain variable region (LCVR) comprising an amino acid sequence selected from the group consisting of SEQ ID NOs: 18, 42, and 66.

61. 52. The bispecific antigen-binding molecule of any one of claims 45 to 51, wherein the first antigen-binding domain competes for binding to human CD28 with a reference antigen-binding protein comprising a heavy chain variable region (HCVR) comprising an amino acid sequence selected from the group consisting of SEQ ID NOs: 10, 34, and 58, and a light chain variable region (LCVR) comprising an amino acid sequence selected from the group consisting of SEQ ID NOs: 18, 42, and 66, and the second antigen-binding domain competes for binding to human PSMA with a reference antigen-binding protein comprising a heavy chain variable region (HCVR) comprising the amino acid sequence selected from the group consisting of SEQ ID NOs: 2, 26, and 50, and a light chain variable region (LCVR) comprising amino acids selected from the group consisting of SEQ ID NOs: 18, 42, and 66.

62. A pharmaceutical composition comprising the bispecific antigen-binding molecule of any one of claims 45 to 61 and a pharmaceutically acceptable carrier or diluent.

63. 63. A method for treating prostate cancer in a subject, comprising administering to the subject the pharmaceutical composition of claim 62.

64. 42. The method of any one of claims 39-41, further comprising administering a second bispecific antibody, wherein the second bispecific antibody comprises a first antigen-binding domain that binds to human CD3 and a second antigen-binding domain that binds to human PSMA.

Images