Bispecific anti-PSMA X anti-CD28 antibodies and uses thereof

Abstract

The present invention provides bispecific antigen binding molecules comprising a first antigen binding domain that specifically binds to human CD28 and a second antigen binding molecule that specifically binds to human PSMA. In certain embodiments, the bispecific antigen binding molecules of the present invention are capable of inhibiting the growth of PSMA-expressing tumors, 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 upregulated or induced targeted immune response is desirable and / or therapeutically beneficial.

Description

[0001] This application is a divisional application of Chinese patent application 201980041082.5. The filing date of the original application is June 21, 2019. The name of the invention is “Bispecific anti-PSMA x anti-CD28 antibodies and their uses”.

[0002] CROSS-REFERENCE TO RELATED APPLICATIONS

[0003] This application is related to and claims priority to U.S. Provisional Application No. 62 / 688,227 filed on June 21, 2018, U.S. Provisional Application No. 62 / 781,930 filed on December 19, 2018, U.S. Provisional Application No. 62 / 781,980 filed on December 19, 2018, and U.S. Provisional Application No. 62 / 815,878 filed on March 8, 2019. The entire contents of the foregoing applications are expressly incorporated herein by reference. This application is a divisional application of Chinese patent application 201980041082.5, the application date of the original application is January 29, 2019, and the invention name is "Botulinum neurotoxin that cuts non-neuronal SNAREs".

[0004] Sequence Listing

[0005] This application contains a sequence listing that has been submitted electronically in ASCII format and is hereby incorporated by reference in its entirety. The ASCII copy was created on June 20, 2019, is named 10367WO01_118003-45220_SL.TXT and is 48,690 bytes in size. Technical Field

[0006] The present invention relates to bispecific antigen binding molecules that bind to CD28 and a target molecule such as PSMA and methods of using the same. Background Art

[0007] CD28 is a type I transmembrane protein expressed on the surface of T cells, which has a single extracellular Ig-V-like domain assembled into a homodimer. CD28 is a receptor for CD80 (B7.1) and CD86 (B7.2) proteins and is activated by CD80 or CD86 expressed on APCs. The binding of CD28 to CD80 or CD86 provides a co-stimulatory signal that is important for T cell activation and survival. In addition to the T cell receptor (TCR), T cell stimulation by CD28 provides a potent signal for the production of various interleukins. CD28 strengthens cell signals after TCR activation, such as pathways controlled by NFκB transcription factors. CD28 co-signals are important for effective T cell activation (such as T cell differentiation, proliferation, cytokine release, and cell death).

[0008] Anti-CD28 antibodies have been proposed for therapeutic purposes involving T cell activation. A specific anti-CD28 antibody, TGN1412 (anti-CD28 superagonist), is in clinical trials. TGN1412 induces a cytokine storm that was not predicted in toxicology or ex vivo human PBMC studies. In 2006, six healthy volunteers were injected intravenously with TGN1412 (anti-CD28 superagonist) at a dose of 0.1 mg / kg. Within 2 hours, all six patients had a significant inflammatory response (cytokine storm). Within 16 hours, all patients had multiple organ failure. The subjects were treated with corticosteroids, and cytokine levels returned to normal levels within 2-3 days. In the Phase 1 study (associated with CRS), the starting dose of 0.1 mg / kg was based on a 500-fold multiple of the no observed adverse effect level "NOAEL" of 50 mg / kg in cynomolgus monkeys (Suntharalingam et al., "Cytokine Storm in a Phase 1 Trial of the Anti-CD28 Monoclonal Antibody TGN1412," NEJM 355:1018-1028 (2006)).

[0009] Toxicology studies in cynomolgus monkeys are not predictive of the cytokine responses observed in humans.

[0010] 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 called glutamate carboxypeptidase II (GPC). In normal human prostates, PSMA is associated with the cytoplasm and apical surface of epithelial cells surrounding prostate ducts. Dysplastic and / or tumorigenic transformation of prostate tissue causes PSMA to transfer from the apical membrane to the luminal surface of the duct. PSMA is constitutively endocytosed, and it is not shed. PSMA is the target of various clinical ADC (antibody-drug conjugate) tests and imaging methods. PSMA is highly expressed in human prostate adenocarcinoma and matches metastasis (lymph nodes). In prostate tumors, PSMA expression levels increase according to stage and grade. The 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 colon cancer, lung cancer, breast cancer, kidney cancer, and bladder cancer subtypes).

[0011] PSMA is also expressed in normal tissues. The strongest expression is found in prostate epithelial cells, duodenum, renal tubular cells, salivary glands, and astrocytes. PSMA is weakly expressed in the fallopian tubes and breast, and very rarely in the cervical endothelium.

[0012] Bispecific antigen binding molecules that bind 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. Summary of the invention

[0013] In the 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; 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 can be used to target tumor cells expressing PSMA (e.g., prostate tumor cells), and the anti-CD28 portion of the bispecific molecule can be used to activate T cells. The simultaneous binding of PSMA on tumor cells and CD28 on T cells promotes the directional killing (cell lysis) of targeted tumor cells by activated T cells. Therefore, the anti-CD28 / anti-PSMA bispecific molecules of the present invention are particularly useful for treating diseases and conditions (e.g., prostate cancer) associated with or caused by tumors expressing PSMA.

[0014] The bispecific antigen binding molecules according to this aspect of the invention include a first antigen binding domain that specifically binds to human CD28 and a second antigen binding domain that specifically binds to PSMA. The invention encompasses anti-CD28 / anti-PSMA bispecific molecules (e.g., bispecific antibodies) wherein each antigen binding domain includes a heavy chain variable region (HCVR) paired with a light chain variable region (LCVR). In certain exemplary embodiments of the invention, the anti-CD28 antigen binding domain and the anti-PSMA antigen binding domain each include a different unique HCVR paired with a common LCVR.

[0015] The present invention provides an anti-CD28 / anti-PSMA bispecific molecule, wherein the first antigen binding domain that specifically binds to CD28 comprises any one of the HCVR amino acid sequences shown in Table 1. The first antigen binding domain that specifically binds to CD28 may also comprise any one of the LCVR amino acid sequences shown in Table 1. According to certain embodiments, the first antigen binding domain that specifically binds to CD28 comprises any one of the HCVR / LCVR amino acid sequence pairs shown in Table 1. The present invention also provides an anti-CD28 / anti-PSMA bispecific molecule, wherein the first antigen binding domain that specifically binds to CD28 comprises any one of the heavy chain CDR1-CDR2-CDR3 amino acid sequences shown in Table 1 and / or any one of the light chain CDR1-CDR2-CDR3 amino acid sequences shown in Table 1.

[0016] 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, the substantially similar sequence having at least 90%, at least 95%, at least 98% or at least 99% sequence identity.

[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 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, wherein the substantially similar sequence has at least 90%, at least 95%, at least 98%, or at least 99% sequence identity.

[0018] 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 HCVR and LCVR (HCVR / LCVR) amino acid sequence pair selected from the group consisting of SEQ ID NO: 10 / 18, 26 / 42 and 58 / 66.

[0019] 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, said substantially similar sequence having 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, said substantially similar sequence having at least 90%, at least 95%, at least 98%, or at least 99% sequence identity.

[0020] In certain embodiments, the first antigen binding domain that specifically binds CD28 comprises a HCDR3 / LCDR3 amino acid sequence pair selected from the group consisting of SEQ ID NOs: 16 / 24, 32 / 48, and 64 / 72.

[0021] 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) domain, the heavy chain CDR1 domain having an amino acid sequence selected from the group consisting of SEQ ID NOs: 12, 28 and 60, or a substantially similar sequence thereof, the substantially similar sequence having at least 90%, at least 95%, at least 98% or at least 99% sequence identity; a heavy chain CDR2 (HCDR2) domain, the heavy chain CDR2 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, the substantially similar sequence having at least 90%, at least 95%, at least 98% or at least 99% sequence identity; a light chain CDR1 (LCDR1) domain, the light chain CDR1 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, the substantially similar sequence having at least 90%, at least 95%, at least 98% or at least 99% sequence identity. NO:20, 44 and 68 or a substantially similar sequence thereof, said substantially similar sequence having at least 90%, at least 95%, at least 98% or at least 99% sequence identity; and a light chain CDR2 (LCDR2) domain, said light chain CDR2 domain having an amino acid sequence selected from the group consisting of SEQ ID NO:22, 46 and 70 or a substantially similar sequence thereof, said substantially similar sequence having at least 90%, at least 95%, at least 98% or at least 99% sequence identity.

[0022] Certain non-limiting, exemplary anti-CD28 / anti-PSMA bispecific antigen binding molecules of the invention comprise a first antigen binding domain that specifically binds to CD28 comprising HCDR1-HCDR2-HCDR3-LCDR1-LCDR2-LCDR3 domains, respectively, the first antigen binding domain having an amino acid sequence selected from the group consisting of: SEQ ID NO: 12-14-16-20-22-24; 28-30-32-44-46-48; and 60-62-64-68-70-72.

[0023] 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 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, the substantially similar sequence having at least 90%, at least 95%, at least 98% or at least 99% sequence identity.

[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 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, wherein the substantially similar sequence has at least 90%, at least 95%, at least 98%, or at least 99% sequence identity.

[0025] 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 HCVR and LCVR (HCVR / LCVR) amino acid sequence pair selected from the group consisting of SEQ ID NO: 2 / 18, 34 / 42 and 50 / 66.

[0026] 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, the substantially similar sequence having 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, the substantially similar sequence having at least 90%, at least 95%, at least 98% or at least 99% sequence identity.

[0027] In certain embodiments, the second antigen binding domain that specifically binds PSMA comprises a HCDR3 / LCDR3 amino acid sequence pair selected from the group consisting of SEQ ID NOs: 8 / 24, 40 / 48, and 56 / 72.

[0028] 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 comprises: a heavy chain CDR1 (HCDR1) domain, the heavy chain CDR1 domain having an amino acid sequence selected from the group consisting of SEQ ID NOs: 4, 36 and 52, or a substantially similar sequence thereof, the substantially similar sequence having at least 90%, at least 95%, at least 98% or at least 99% sequence identity; a heavy chain CDR2 (HCDR2) domain, the heavy chain CDR2 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, the substantially similar sequence having at least 90%, at least 95%, at least 98% or at least 99% sequence identity; a light chain CDR1 (LCDR1) domain, the light chain CDR1 domain having an amino acid sequence selected from the group consisting of SEQ ID NOs: 4, 36 and 52, or a substantially similar sequence thereof, the substantially similar sequence having at least 90%, at least 95%, at least 98% or at least 99% sequence identity. NO:20, 44 and 68 or a substantially similar sequence thereof, said substantially similar sequence having at least 90%, at least 95%, at least 98% or at least 99% sequence identity; and a light chain CDR2 (LCDR2) domain, said light chain CDR2 domain having an amino acid sequence selected from the group consisting of SEQ ID NO:22, 46 and 70 or a substantially similar sequence thereof, said substantially similar sequence having at least 90%, at least 95%, at least 98% or at least 99% sequence identity.

[0029] Certain non-limiting, exemplary anti-CD28 / anti-PSMA bispecific antigen binding molecules of the invention comprise a second antigen binding domain that specifically binds to PSMA comprising HCDR1-HCDR2-HCDR3-LCDR1-LCDR2-LCDR3 domains, respectively, the second antigen binding domain having an amino acid sequence selected from the group consisting of: SEQ ID NO: 4-6-8-20-22-24; 36-38-40-44-46-48; and 52-54-56-68-70-72.

[0030] In related embodiments, the present invention comprises an anti-CD28 / anti-PSMA bispecific antigen binding molecule, wherein the second antigen binding domain that specifically binds to PSMA comprises heavy and light chain CDR domains contained within a heavy chain and light chain variable region (HCVR / LCVR) sequence selected from the group consisting of SEQ ID NOs: 2 / 18, 34 / 42, and 50 / 66.

[0031] In another aspect, the present invention provides a nucleic acid molecule encoding any one of the HCVR, LCVR or CDR sequences of the anti-CD28 / anti-PSMA bispecific antigen binding molecules disclosed herein, the nucleic acid molecule comprising: a nucleic acid molecule comprising a polynucleotide sequence as shown in Table 1 herein and a nucleic acid molecule comprising two or more of the polynucleotide sequences as shown in Table 1 (in any functional combination or arrangement thereof). Recombinant expression vectors carrying the nucleic acids of the present invention and host cells into which such vectors have been introduced are also encompassed by the present invention, as are methods of producing antibodies by culturing host cells under conditions that allow for the production of antibodies and recovering the produced antibodies.

[0032] The present invention comprises an anti-CD28 / anti-PSMA bispecific antigen-binding molecule, wherein any one of the above-mentioned antigen-binding domains that specifically bind to CD28 is combined, linked or otherwise associated with any one of the above-mentioned antigen-binding domains that specifically bind to PSMA, thereby forming a bispecific antigen-binding molecule that binds to CD28 and PSMA.

[0033] The present invention comprises anti-CD28 / anti-PSMA bispecific antigen binding molecules with modified glycosylation patterns. In some applications, for example, modifications can be used to remove undesirable glycosylation sites, or antibodies lacking fucose moieties are present on oligosaccharide chains to increase antibody-dependent cellular cytotoxicity (ADCC) function (see Shield et al., (2002) Journal of Biological Chemistry (JBC) 277:26733). In other applications, modifications of galactosylation can be performed to modify complement-dependent cytotoxicity (CDC).

[0034] In another aspect, the invention provides a pharmaceutical composition comprising an anti-CD28 / anti-PSMA bispecific antigen binding molecule as 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 is advantageously combined with an anti-CD28 / anti-PSMA bispecific antigen binding molecule. Exemplary agents that can be advantageously combined with anti-CD28 / anti-PSMA bispecific antigen binding molecules are discussed in detail elsewhere herein.

[0035] In another aspect, the present invention provides a therapeutic method for targeting / killing PSMA-expressing tumor cells using the anti-CD28 / anti-PSMA bispecific antigen binding molecules of the present invention, wherein the therapeutic method comprises administering to a subject in need thereof a therapeutically effective amount of a pharmaceutical composition comprising the anti-CD28 / anti-PSMA bispecific antigen binding molecules of the present invention.

[0036] The present invention also encompasses the use of the anti-CD28 / anti-PSMA bispecific antigen binding molecules of the present invention in the manufacture of a medicament for treating a disease or disorder associated with or caused by PSMA expression.

[0037] In yet another aspect, the present invention provides therapeutic methods for targeting / killing PSMA-expressing tumor cells using the anti-CD28 / anti-PSMA bispecific antigen binding molecules of the present invention, wherein the anti-CD28 / anti-PSMA bispecific antigen binding molecules are combined with other anti-tumor bispecific antigen binding molecules that bind to CD3 (e.g., anti-CD28 / anti-PSMA and anti-CD3 / anti-PSMA antibody combinations).

[0038] In still another aspect, the present invention provides a therapeutic method for targeting / killing PSMA-expressing tumor cells using the anti-CD28 / anti-PSMA bispecific antigen binding molecules of the present invention, wherein the anti-CD28 / anti-PSMA bispecific antigen binding molecules are combined with a checkpoint inhibitor targeting PD-1 or CTLA-4 (e.g., anti-CD28 / anti-PSMA in combination with an anti-PD-1 antibody).

[0039] In still another aspect, the present invention provides a therapeutic method for targeting / killing PSMA-expressing tumor cells using the anti-CD28 / anti-PSMA bispecific antigen binding molecules of the present invention, wherein the anti-CD28 / anti-PSMA bispecific antigen binding molecules are combined with other anti-tumor bispecific antigen binding molecules that bind to CD3 (e.g., anti-CD28 / anti-PSMA in combination with anti-CD3 / anti-PSMA bispecific antibodies) and with a checkpoint inhibitor targeting PD-1 or CTLA-4 (e.g., anti-CD28 / anti-PSMA in combination with anti-PD-1 antibodies).

[0040] Other embodiments will become apparent through reading the following detailed description. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 is a schematic diagram showing the structure of an exemplary anti-CD28 / PSMA bispecific antibody.

[0042] Figure 2is a graph showing tumor growth inhibition in engineered cell lines with introduced co-stimulatory ligand expression. Three tumor cell lines, B16F10.9, EL4 and MC 38, were engineered to express co-stimulatory ligands or GFP as a control. Figure 2 is a graph showing the survival rate percentage (%). When the tumor grows to more than 2000 mm 3 When , the mice were euthanized.

[0043] Figure 3 is a schematic diagram showing the primary and engineered bioassays used to test anti-PSMAxCD28 bispecific antibodies.

[0044] Figure 4A and 4B It is shown that anti-hPSMA x hCD28 enhanced the activation of CD4+ T cells and engineered JRT3.T3 / 1G4 / hCD28 cells in the presence of primary stimulation (REGN2281) and PSMA expressed on HEK293 / hCD20 cells.

[0045] Figures 5A-5H It was shown that in an engineered cell line with introduced PSMA, the PSMAxCD28 bispecific antibody enhanced T cell activation in the presence of TCR stimulation by the anti-CD20xCD3 bispecific antibody.

[0046] Figure 5A is a schematic diagram showing the mode of action of a bispecific antibody.

[0047] Figure 5B Is a graph showing the quantification of the ratio of CD28 fluorescence inside / outside the immune synapse. PSMA-engineered target cells (HEK293) and human Jurkat T cells were co-cultured with fluorescently labeled bispecific antibodies (anti-PSMAxCD28, anti-CD20xCD3) at 37°C for 1 hour, gently fixed and stained with anti-CD28. The number of cells analyzed in each group is indicated on the bar graph.

[0048] Figure 5C and 5D Proliferation of human T cells cultured with PSMA-engineered target cells is shown. Data shown are mean ± SEM. Data represent at least two (2) experiments.

[0049] Figure 5C A graph showing dose titration of CD20xCD3 in the presence of 0.5 nM hIgG4s isotype control or PSMAxCD28 is provided.

[0050] Figure 5DGraphs showing dose titration of PSMAxCD28 or indicated controls in the presence of 5 pM hIgG4s isotype control or CD20xCD3 are provided.

[0051] Figure 5E-5G Graphs showing proliferation and cytokine release of human T cells cultured with TAA (PSMA or CD20) engineered target cells in the presence of 5 pM hIgG4s isotype control (bottom panel of each graph) or CD20xCD3 (top panel of each graph), as indicated in the top panels. Data are mean ± SEM. Data represent at least three (3) experiments.

[0052] Figure 5E is a graph showing proliferation.

[0053] Fig. 5F is a graph showing IL-2 release.

[0054] Figure 5G It is a graph showing IFNγ release.

[0055] Figure 5H is a diagram showing that anti-PSMAxCD28 and anti-PSMAxCD3 bispecific antibodies can bind to PSMA-expressing cells simultaneously. 22RV1 cells were pre-incubated in flow cytometry buffer (PBS + 1% FBS) at 4°C for 30 minutes with 20 mg / ml of PSMAxCD3 or 20 mg / ml of anti-PSMA antibodies, which have anti-PSMA arms similar to the PSMAxCD28 bispecific antibodies. After incubation, 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, cells were washed, resuspended in flow cytometry buffer, and analyzed by flow cytometry.

[0056] Fig. 6A and 6B It was shown that anti-PSMAxCD28 enhanced the cytotoxic efficacy of anti-PSMAxCD3 in the presence of prostate cancer and human or cynomolgus monkey T cells.

[0057] Figures 7A-7H The provided figures show that in cancer cell lines with endogenous PSMA, the PSMAxCD28 bispecific antibody enhances T cell activation in the presence of TCR stimulation by the PSMAxCD3 bispecific antibody. Figures 7A-7D , human T cells were cultured with cancer target cells with endogenous PSMA expression (prostate cancer cell line C4-2) and the indicated bispecific antibodies for 96 h.

[0058] Fig. 7A is a graph showing tumor cell killing. The data shown are the percentage of viable cells.

[0059] Figure 7B It is a graph showing IFNγ release.

[0060] Figure 7C Provides a graph showing CD4 T cell counts and CD25 + Plot of cell frequency as percentage of CD4 T cells.

[0061] Fig.7D Provides a graph showing CD8 T cell counts and CD25 + Plot of cell frequency as percentage of CD8 T cells.

[0062] for Figure 7E-7G , endogenous T cells were cultured with cancer cells with endogenous PSMA expression (prostate cancer cell line C4-2) and the indicated bispecific antibodies for 96 h.

[0063] Fig. 7E is a graph showing tumor cell killing. The data shown are the percentage of viable cells.

[0064] Figure 7F Provides a graph showing CD4 T cell counts and CD25 + Plot of cell frequency as percentage of CD4 T cells.

[0065] Figure 7G Provides a graph showing CD8 T cell counts and CD25 + Plot of cell frequency as percentage of CD8 T cells.

[0066] Figure 7H Graphs showing antibody binding to cellular targets measured by flow cytometry are provided.

[0067] Figure 8 A comparison between the properties of the TGN superagonist and the properties of the anti-CD28 antibodies of the present invention is shown.

[0068] Fig.9A and 9B Figures and drawings showing that PSMAxCD28 bispecific antibodies enhance anti-tumor immunity through PSMAxCD3 and induced T cell activation are provided. MC38 / hPSMA tumor cells were implanted subcutaneously in hCD3 / hCD28 / hPSMA humanized mice. Mice were treated with the indicated 5 mg / kg bispecific antibodies on days 0, 3, and 7.

[0069] Fig.9ATumor volume changes over time are shown. Values ​​represent mean ± SEM and are representative of three (3) experiments with 3-7 mice per group. p values ​​were calculated using two-way ANOVA. (*, p<0.05; **, p<0.01; ***, p<0.001 and ****, p<0.0001).

[0070] Fig. 9B Figures provided show the percentage of cells in each cluster from each treatment group (top panels); overlay of the designated clusters on the viSNE plot (bottom panels).

[0071] Fig. 9C and 9D Bar graphs showing that PSMAxCD28 bispecific antibodies enhance cytokine release induced by PSMAxCD3 are provided. Mice were bled for serum cytokines 4 hours after dosing on day 0. Statistical significance was calculated using one-way ANOVA compared to isotype **p<0.01 and ***p<0.0001, n=7 mice per group. Data represent 3 experiments.

[0072] Fig.9E Graphs and drawings showing that PSMAxCD28 bispecific antibodies enhance T cell activation induced by PSMAxCD3 are provided.

[0073] Fig.10 It was shown that the PSMA x CD28 bispecific antibody or the parental CD28 bivalent antibody did not induce serum cytokine production in CD3 / CD28 / PSMA humanized mice.

[0074] Fig.11 It was shown that CD28 superagonist treatment drove a cytokine response within 4 hours in NSG mice transplanted with PBMCs, whereas the anti-CD28 antibody mAb 14226P2 did not.

[0075] Fig.12 The potentially safer toxicology profile of the co-stimulatory bispecific antigen binding molecules of the invention is shown. Anti-CD28 x anti-PSMA did not elicit a cytokine response, whereas anti-CD3 x anti-PSMA did.

[0076] Fig.13A and 13B It is shown that PSMA x CD28 does not show cytokine production or T cell marginalization compared to PSMA x CD3 and CD28 superagonists. LLOQ: “Lower Limit of Quantitation”.

[0077] Fig.14 and 17TSAxCD28 and PD-1 blockade were shown to activate TCR / CD3 and CD28 on T cells at tumor sites. Fig.15 , 16 and 18 show that TSAxCD28 bispecific and PD-1 blockade synergistically promote T cell activation in vitro. Figure 14-18 In the assay, T cells (Jurkat / PD-1) and target cells (Raji WT) were conjugated in the presence of a non-blocking PD-1 mAb (NB PD-1 mAb) or a blocker (PD-1 mAb) and a CD20xCD3 bispecific.

[0078] Fig.14 Provided is a quantitative bar graph showing the localization of PD-1 and CD28 at the immune synapse. Unpaired t-test was used to calculate statistical significance (not significant, ns). In order to quantitatively localize PD-1 and CD28 at the immune synapse, images of T cells (Jurkat / PD-1) and target cells (Raji WT) conjugates in the presence of non-blocker PD-1mAb (NB PD-1mAb) or blocker (PD-1mAb) and anti-CD20xCD3 bispecific antibodies were taken. PD-1mAbs were directly labeled with Alexa647, anti-CD20xCD3 bispecific antibodies were directly labeled with Alexa488, CD28 mAbs were directly labeled with PE, and cell nuclei were stained with Hoechst33342 (image not shown).

[0079] Fig.15 Provided is a quantitative bar graph showing the location of PD-1 and CD28 at the immune synapse. Unpaired t-test was used to calculate statistical significance (p < 0.0001, ****). In order to quantitatively locate PD-1 and CD28 at the immune synapse, images of T cells (Jurkat / PD-1) and target cells (Raji / PD-L1) conjugates in the presence of non-blocker PD-1mAb (NB PD-1mAb) or blockers (PD-1mAb) and anti-CD20xCD3 bispecific antibodies were taken. PD-1mAbs were directly labeled with Alexa647 (shown in red), CD20xCD3 was directly labeled with Alexa488 (shown in green), CD28 mAb was directly labeled with PE (shown in blue), and the nucleus was stained with Hoechst 33342 (shown in gray). Dashed lines are cell outlines drawn based on bright field images.

[0080] Fig.16 The percent (%) viability of 22RV1-PDL1 cells at 96 hours is shown.

[0081] Fig.17 is a graph showing IFNγ levels in supernatants at 96 hours.

[0082] Fig.18 is a graph showing IL-2 release at 96 hours.

[0083] Figures 19A-19D The results show that expression of 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). Initially, 1x10 6 tumor cells and were treated with PD-1 mAb or rat isotype control at 5 mg / kg on days 0, 3, 7, 10, and 14 after tumor implantation.

[0084] Fig.19A Mean tumor volume changes over time are shown. Error bars represent + / - SEM. Statistical significance was determined using two-way ANOVA and Tukey's multiple comparison test.

[0085] Fig.19B The survival rate is shown over time (tumor <2000mm 3 Statistical significance at day 60 post-implantation was determined using the log-rank (Mantel-Cox) test.

[0086] exist Fig.19C In the Figure, mice were treated with a CD8 depleting antibody (CD8 depletion) or an isotype control (no depletion). Mean tumor volume over time in the case of CD8 depletion (dashed line) and mean tumor volume over time in the case of no depletion (solid line) are shown as + / - SEM. Statistical significance was determined using two-way ANOVA and Tukey's multiple comparison test.

[0087] Fig.19D Shown is a second tumor implantation (re-challenge) of tumor-free mice implanted with MC38 / CD86 and treated with PD1 mAb.

[0088] exist Figures 19A-19D In the Figures, the data shown are from 1 experiment with 10 mice per group. The data represent at least 4 independent experiments. Statistical significance is specified (*p<0.05, **p<0.01, ***p<0.001 and ****p<0.0001).

[0089] Figures 20, 21A, 22, 24 and 25 show that PSMAxCD28 synergizes with PD1 mAb treatment to induce anti-tumor immunity. MC38 / hPSMA tumor cells were implanted subcutaneously in hCD3 / hCD28 / hPSMA mice. PSMAxCD28 bispecific antibody, PD-1 mAb or rat IgG2a isotype control were administered as monotherapy or in combination at a dose of 5 mg / kg each by intraperitoneal injection. Fig. 21B and 21C It was shown that the combination of PSMAxCD28 and PD-1 mAb increased the frequency of tumor-specific T cells. Fig.23A and 23B PSMA x CD28 was shown to synergize with anti-PD1 treatment to induce intratumoral cytokines, but not splenic or systemic cytokines.

[0090] Figures 20A-20E showed that immediate treatment with anti-PSMA x anti-CD28 antibodies enhanced immunity at the tumor site and synergized with anti-PD-1 antibodies to promote tumor rejection. Fig. 20A Mean tumor volume changes over time are shown. Error bars represent + / - SEM. Statistical significance was determined using two-way ANOVA and Tukey's multiple comparison test (***, p<0.001 and ****, p<0.0001). Fig. 20B The survival rate is shown over time (for tumors > 2000 mm 3 The mice were euthanized. The log-rank (Mantel-Cox) test was used to determine statistical significance (**, p<0.01). Fig. 20C Mean tumor volume at day 21 post-implantation is shown. Error bars represent + / - SEM. Statistical significance was determined using one-way ANOVA and Holm-Sidak multiple comparison test (****, p<0.0001). Fig.20D The tumor volume (mm) on day 21 after implantation is shown. 3 ) table. Fig.20E Shown is a table of the number of tumor-free mice per group from 6 independent experiments.

[0091] Fig.21A Shown is a second tumor challenge of tumor-free mice previously treated with a combination of PSMAxCD28 and PD1 mAbs. Error bars represent + / - SEM and data represent 4 experiments.

[0092] Fig. 21BMC38 / PSMA tumor cells implanted into CD3 / CD28 / PSMA humanized mice and treated with 5 mg / kg of isotype control, PSMAxCD28, PD1 mAb or combination on days 10 and 14 post-implantation are shown. 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.

[0093] Fig. 21C It is shown that tumor-free mice from anti-CD28 x PSMA and anti-PD1 treatment, but not anti-CD28 x PSMA and anti-CD3 x PSMA treatment, rejected a second tumor re-challenge.

[0094] Figures 22A-22D , 24 and 25 show delayed / curative treatment regimens (doses are indicated by arrows); Fig.22A and 22B for the 9th, 16th and 22nd days; Fig.24 and 25 They are the 7th, 11th and 14th days.

[0095] Fig.22A Mean tumor volume changes over time are shown. Error bars represent + / - SEM. Statistical significance was determined using two-way ANOVA and Tukey's multiple comparison test. Data represent 3 experiments.

[0096] Fig. 22B Ex vivo spleen cytokines and intratumoral cytokines are shown. Points represent data from a single mouse. Bars are mean + / - SEM.

[0097] Fig. 22C Mean tumor size at day 20 is shown.

[0098] Fig.22D Shown are the survival rates of mice receiving different treatments.

[0099] Fig.23A and 23B It was shown that PSMA x CD28 synergized with anti-PD1 treatment to induce intratumoral cytokines, but not splenic or systemic cytokines.

[0100] exist Fig.23A The data corresponds to Fig. 22B Ex vivo spleen cytokines and intratumoral cytokines. Points represent data from a single mouse. Bars are mean + / - SEM.

[0101] exist Fig. 23BIn the study, CD3 / CD28 / PSMA triple humanized mice were implanted with MC38 / hPSMA and treated with 5 mg / kg of the indicated antibodies on day 0. Mice were bled 4 hours after dosing and serum was collected. Fig.23A and 23B One-way ANOVA and Tukey's multiple comparison test were used to calculate statistical significance. *p<0.05, **p<0.01, ****p<0.0001

[0102] Fig.24 PD1 expression on T cell subsets from draining lymph nodes (dLN), spleen (Sp), and tumor (Tu) is shown. Points represent data from a single mouse. Bars are mean + / - SEM.

[0103] Fig.25 The frequencies of CD8 T cells in C1 and C2 from the indicated treatment groups are shown.

[0104] Fig. 22B , 24 The data in and 25 are from 1 experiment each. Fig. 22B and 24 , n = 4-6 mice per group. Fig.25 , n = 10 mice per group.

[0105] Figures 26A-26C Figures 27 and 28 show that TAAxCD28 alone or in combination with PD1 therapy does not induce systemic T cell activation compared to CD28 superagonists in cynomolgus monkeys. Cynomolgus monkeys were treated with a single dose of bispecific antibody at the indicated dose (1 or 10 mg / kg as indicated). The indicated time (hours) after dosing.

[0106] Fig.26A Serum cytokines are shown.

[0107] Fig.26B Relative peripheral blood T cell counts are shown.

[0108] Fig.26C Ki67 is shown + and ICOS + The frequency of T cells (percentage of CD3).

[0109] exist Figures 26A-26C Values ​​represent mean + / - SEM. N = 3 animals per group.

[0110] Fig. 27CD3 / CD28 / PSMA triple humanized mice were treated with a single dose of antibody (0.25 or 2.5 mg / kg as indicated). Mice were bled 4 hours after dosing (day 0) and serum was collected. Statistical significance was calculated using one-way ANOVA and Holm-Sidak's multiple comparison test. *p<0.05, **p<0.01, ****p<0.0001

[0111] Fig.28 It was shown that PSMA x CD28+ / -PD1 did not elevate serum cytokines in tumor-bearing mice.

[0112] Fig.29A and 29B PSMAxCD28 alone or in combination with PD1 mAb showed a safe cytokine profile in non-tumor bearing mice. Fig. 27 Data in correspondence. CD3 / CD28 / PSMA triple humanized mice were treated with a single dose of antibody (0.25 or 2.5 mg / kg as indicated). Mice were bled 4 hours (day 0) and 72 hours (day 3) after dosing, and serum was collected. 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

[0113] Before describing the present invention, it should be understood that the present invention is not limited to the specific methods and experimental conditions described, because such methods and conditions can vary. It should also be understood that the terminology used herein is only for the purpose of describing specific embodiments, and is not intended to be limiting, because the scope of the present invention is limited only by the appended claims.

[0114] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those generally understood by those of ordinary skill in the art to which the present invention belongs. As used herein, when used to refer to a specific listed value, the term "about" means that the numerical value may differ from the listed value by no more than 1%. For example, as used herein, the expression "about 100" includes 99 and 101 and all values ​​therebetween (e.g., 99.1, 99.2, 99.3, 99.4, etc.).

[0115] Although any methods and materials similar or equivalent to those described herein can be used in the practice or testing of the present invention, the preferred methods and materials are now described. All patents, applications and non-patent publications mentioned in this specification are incorporated herein by reference in their entirety.

[0116] definition

[0117] As used herein, the expression "CD28" refers to an antigen expressed on T cells as a co-stimulatory receptor. Human CD28 includes an amino acid sequence as shown in SEQ ID NO:74, and / or has an amino acid sequence as shown in NCBI accession number NP_006130.1. All references to proteins, polypeptides and protein fragments herein are intended to refer to the human form of the corresponding protein, polypeptide or protein fragment, unless it is explicitly stated that it is from a non-human species. Therefore, the expression "CD28" means human CD28, unless specified as being from a non-human species, for example, "mouse CD28", "monkey CD28" etc.

[0118] 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 CD28 expressed on the surface of cells. Soluble CD28 includes natural CD28 protein as well as recombinant CD28 protein variants, such as, for example, monomeric and dimeric CD28 constructs that lack a transmembrane domain or are otherwise unrelated to the cell membrane.

[0119] As used herein, the expression "CD28 expressed on the cell surface" means one or more CD28 proteins expressed on the cell surface in vitro or in vivo, so that at least a portion of the CD28 protein is exposed to the extracellular side of the cell membrane and can be approached by the antigen binding portion of the antibody."CD28 expressed on the cell surface" includes CD28 proteins contained in the context of functional T cell co-stimulatory receptors in the membrane of the cell. The expression "CD28 expressed on the cell surface" includes CD28 proteins expressed as part of homodimers on the cell surface. "CD28 expressed on the cell surface" may include or consist of CD28 proteins expressed on the surface of cells that normally express CD28 proteins. Alternatively, "CD28 expressed on the cell surface" may include or consist of CD28 proteins expressed on the surface of such a cell, wherein the cell does not normally express human CD28 on its surface, but has been artificially engineered to express CD28 on its surface.

[0120] As used herein, the expression "anti-CD28 antibody" includes monovalent antibodies with a single specificity, as well as bispecific antibodies comprising a first arm that binds CD28 and a second arm that binds a second (target) antigen, wherein the anti-CD28 arm comprises any of the HCVR / LCVR or CDR sequences as shown 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.

[0121] As used herein, the term "antibody" means any antigen binding molecule or molecular complex comprising at least one complementary determining region (CDR) that specifically binds or interacts with a specific antigen (e.g., CD28). The term "antibody" includes immunoglobulin molecules that include four polypeptide chains, two heavy (H) chains and two light (L) chains interconnected by disulfide bonds, and multimers thereof (e.g., IgM). Each heavy chain includes a heavy chain variable region (abbreviated herein as HCVR or VH) and a heavy chain constant region. The heavy chain constant region includes three domains C H 1. C H 2 and C H 3. Each light chain includes a light chain variable region (abbreviated herein as LCVR or VL) and a light chain constant region. The light chain constant region includes a domain (C L 1) You can set V H Area and V L The V region is further subdivided into regions of hypervariability, called complementarity determining regions (CDRs), interspersed with regions that are more conserved, called framework regions (FRs). H and V L It includes three CDRs and four FRs, arranged from the amino terminus to the carboxyl terminus in the following order: FR1, CDR1, FR2, CDR2, FR3, CDR3, FR4. In different embodiments of the present invention, the FRs of the anti-CD28 antibody (or its antigen-binding portion) may be identical to human germline sequences, or may be natural or artificially modified. An amino acid consensus sequence may be defined based on a side-by-side analysis of two or more CDRs.

[0122] As used herein, the term "antibody" also includes the antigen binding fragment of a complete antibody molecule. As used herein, the term "antigen binding portion thereof", "antigen binding fragment" of an antibody, etc., include any naturally occurring, enzymatically obtainable, synthetic or genetically engineered polypeptide or glycoprotein, which specifically binds to an antigen to form a complex. The antigen binding fragment of an antibody can be derived from a complete antibody molecule, for example, using any suitable standard technique (such as proteolytic digestion or recombinant genetic engineering technology, which relates to manipulation and expression of DNA encoding antibody variable domains and optional constant domains). This DNA is known and / or easy to obtain from, for example, commercial sources, DNA libraries (including, for example, phage-antibody libraries), or can be synthesized. DNA can be sequenced and operated by chemical methods or by molecular biological techniques, for example, one or more variable domains and / or constant domains are arranged into a suitable configuration, or codons are introduced, cysteine ​​residues are produced, modified, added or missing amino acids, etc.

[0123] Non-limiting examples of antigen-binding fragments include: (i) Fab fragments; (ii) F(ab')2 fragments; (iii) Fd fragments; (iv) Fv fragments; (v) single-chain Fv (scFv) molecules; (vi) dAb fragments; and (vii) minimal recognition units consisting of amino acid residues that mimic the hypervariable region of an antibody (e.g., isolated complementary determining regions (CDRs), such as CDR3 peptides) or constrained FR3-CDR3-FR4 peptides. As used herein, the expression "antigen-binding fragment" also encompasses other engineered molecules, such as domain-specific antibodies, single domain antibodies, domain-deleted antibodies, chimeric antibodies, CDR-grafted antibodies, diabodies, triabodies, tetrabodies, minibodies, nanobodies (e.g., monovalent nanobodies, divalent nanobodies, etc.), small modular immunopharmaceuticals (SMIPs), and shark variable IgNAR domains.

[0124] The antigen-binding fragment of an antibody generally includes at least one variable domain. The variable domain can be of any size or amino acid composition and will generally include at least one CDR adjacent to or within one or more framework sequences. L Domain-associated V H In the antigen-binding fragment of the structural domain, V H Domain and V L The domains may be positioned relative to each other in any suitable arrangement. For example, the variable region may be a dimer and contain V H -V H 、V H -V L or V L Alternatively, the antigen-binding fragment of the antibody may contain a monomer V H or V L Structural domain.

[0125] In certain embodiments, the antigen-binding fragment of an antibody may contain 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 within the 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) VH -C H 2-C H 3; (vii) V H -C L ; (viii) V L -C H 1; (ix) V L -C H 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; (xiii) V L -C H 2-C H 3; and (xiv) V L -C L . In any configuration of variable domains and constant domains (including any of the exemplary configurations listed above), the variable domains and constant domains may be directly connected to each other or may be connected by a complete or partial hinge or linker region. The hinge region may be composed of at least 2 (e.g., 5, 10, 15, 20, 40, 60 or more) amino acids that result in a flexible or semi-flexible connection between adjacent variable domains and / or constant domains in a single polypeptide molecule. In addition, the antigen-binding fragment may include homodimers or heterodimers (or other multimers) of any of the variable domain and constant domain configurations listed above, which are bound to each other and / or to one or more monomer V H or V L The domains are non-covalently associated (eg, through one or more disulfide bonds).

[0126] As with complete antibody molecules, antigen-binding fragments can be monospecific or multispecific (e.g., bispecific). Multispecific antigen-binding fragments of antibodies typically include at least two different variable domains, each of which is capable of specifically binding to a separate antigen or a different epitope on the same antigen. Conventional techniques available in the art can be used to apply any multispecific antibody format (including the exemplary bispecific antibody formats disclosed herein) to the context of antigen-binding fragments of the antibodies of the present invention.

[0127] The antibodies of the present invention can act by complement dependent cytotoxicity (CDC) or antibody dependent cell-mediated cytotoxicity (ADCC). "Complement dependent cytotoxicity" (CDC) refers to the lysis of cells expressing antigens 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 non-specific cytotoxic cells (e.g., natural killer (NK) cells, neutrophils and macrophages) expressing Fc receptors (FcR) recognize the binding antibodies on target cells, and thereby cause the lysis of target cells. CDC and ADCC can be measured using assays well known in the art and available. (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 for the ability of an antibody to fix complement and mediate cell dependent cytotoxicity. Thus, the isotype of an antibody can be selected based on whether the antibody is desired to mediate cytotoxicity.

[0128] In certain embodiments of the present invention, the anti-CD28 antibodies (monospecific or bispecific) of the present invention are human antibodies. As used herein, the term "human antibody" is intended to include antibodies with variable regions and constant regions derived from human germline immunoglobulin sequences. The human antibodies of the present invention may include amino acid residues that are not encoded by human germline immunoglobulin sequences (e.g., mutations introduced by random or site-specific mutagenesis in vitro or by somatic mutations in vivo), such as in CDR, particularly in CDR3. However, as used herein, the term "human antibody" is not intended to include antibodies in which CDR sequences derived from germline of another mammalian species such as mice have been transplanted to human framework sequences.

[0129] In some embodiments, the antibody of the present invention can be a recombinant human antibody. As used herein, the term "recombinant human antibody" is intended to include all human antibodies prepared, expressed, produced or separated by a recombinant manner, such as antibodies expressed using a recombinant expression vector transfected into a host cell (described further below), antibodies separated from a recombinant combination human antibody library (described further below), antibodies separated from human immunoglobulin gene transgenic animals (e.g., mice) (see, for example, Taylor et al., (1992), "Nucl. Acids Res.", 20: 6287-6295) or antibodies prepared, expressed, produced or separated by any other means, the other means involving splicing human immunoglobulin gene sequences to other DNA sequences. Such recombinant human antibodies have variable regions and constant regions derived from human germline immunoglobulin sequences. However, in certain embodiments, such recombinant human antibodies undergo in vitro mutagenesis (or, when using human Ig sequence transgenic animals, undergo in vivo somatic cell mutagenesis), and therefore the V H Area and V L The amino acid sequence of the region is as follows: H Sequence and V L sequences but may not naturally occur in the human antibody germline repertoire in vivo.

[0130] Human antibodies can exist in two forms related to hinge heterogeneity. In one form, the immunoglobulin molecule includes a stable four-chain construct of about 150-160 kDa, in which the dimer is held together by an interchain heavy chain disulfide bond. In the second form, the dimer is not connected by an interchain disulfide bond and forms a molecule of about 75-80 kDa (half antibody), which includes a covalently coupled light chain and heavy chain. Even after affinity purification, these forms are extremely difficult to separate.

[0131] The frequency of the second form in various complete IgG isotype forms is based on, but not limited to, structural differences associated with the hinge region isotype of the antibody. Single amino acid substitutions in the hinge region of the human IgG4 hinge can significantly reduce the appearance of the second form (Angal et al., (1993), Molecular Immunology 30:105) to the level usually observed with the human IgG1 hinge. The present invention encompasses antibodies having one or more mutations in the hinge region, CH2 region or CH3 region, which mutations may be, for example, desired in production, so as to improve the yield of the desired antibody form.

[0132] The antibody of the present invention can be an isolated antibody. As used herein, "isolated antibody" means 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 naturally occurring or naturally produced tissue or cell of an antibody is an "isolated antibody" for purposes of the present invention. Isolated antibodies also include in situ antibodies within recombinant cells. An isolated antibody is an antibody that has been subjected to at least one purification or separation step. According to certain embodiments, the isolated antibody may be substantially free of other cellular materials and / or chemicals.

[0133] The present invention also includes one-armed antibodies that bind CD28. As used herein, "one-armed antibodies" means an antigen binding molecule comprising a single antibody heavy chain and a single antibody light chain. The one-armed antibodies of the present invention may include any one of the HCVR / LCVR or CDR amino acid sequences shown in Table 2.

[0134] The anti-CD28 antibodies or antigen-binding domains thereof herein may include 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 antigen-binding proteins or antigen-binding domains are derived. Such mutations can be easily determined by comparing the amino acid sequences disclosed herein with germline sequences that can be obtained, for example, from public antibody sequence databases. The present invention includes antibodies and antigen-binding domains 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 mutated to one or more corresponding residues of the germline sequence from which the antibody is derived, or mutated to one or more corresponding residues of another human germline sequence, or mutated to conservative amino acid substitutions of one or more corresponding germline residues (such sequence changes are collectively referred to herein as "germline mutations"). Starting from the heavy and light chain variable region sequences disclosed herein, one of ordinary skill in the art can easily generate many antibodies and antigen-binding fragments comprising one or more individual germline mutations or combinations thereof. In certain embodiments, V H Domain and / or V LThe framework and / or CDR residues in the domain are all mutated back to the residues found in the original germline sequence of the derived antibody. In other embodiments, only some residues are mutated back to the original germline sequence, for example, only the residue after the mutation found in the first 8 amino acids of FR1 or in the last 8 amino acids of FR4, or only the residue after the mutation found in CDR1, CDR2 or CDR3. In other embodiments, one or more mutations in one or more framework and / or CDR residues are one or more corresponding residues of different germline sequences (that is, germline sequences different from the germline sequence of the initial derived antibody). In addition, the antibody of the present invention or its antigen-binding domains can contain any combination of two or more germline mutations in the framework and / or CDR district, for example, wherein some individual residues are mutated into the corresponding residues of a specific germline sequence, and some other residues different from the original germline sequence can remain unchanged or mutate into the corresponding residues of different germline sequences. Once obtained, antibodies or antigen-binding fragments thereof 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 antagonistic or agonistic biological properties (as the case may be), reduced immunogenicity, etc. Antibodies or antigen-binding fragments thereof obtained in this general manner are encompassed within the present invention.

[0135] 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 having 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, relative to any of the HCVR, LCVR and / or CDR amino acid sequences as shown in Table 1 herein, the HCVR, LCVR and / or CDR amino acid sequences having, for example, 10 or less, 8 or less, 6 or less, 4 or less, etc., conservative amino acid substitutions.

[0136] 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 more than one epitope. Thus, different antibodies may bind to different regions on the antigen and may have different biological effects. Epitopes may be conformational or linear. Conformational epitopes are produced by spatially juxtaposed amino acids from different segments of a linear polypeptide chain. A linear epitope is an epitope produced by adjacent amino acid residues in a polypeptide chain. In some cases, an epitope may comprise a portion of a sugar, phosphoryl, or sulfonyl group on the antigen.

[0137] The term "substantial identity" or "substantially identical" when referring to a nucleic acid or a fragment thereof means that the nucleotide sequence identity is at least about 95% of the nucleotide bases when optimally aligned with another nucleic acid (or its complementary strand) by appropriate nucleotide insertions or deletions, and more preferably, at least about 96%, 97%, 98% or 99%, as measured by any well-known sequence identity algorithm such as FASTA, BLAST or Gap, as discussed below. In some cases, a nucleic acid molecule having substantial identity to a reference nucleic acid molecule can encode a polypeptide having an amino acid sequence that is identical or substantially similar to a polypeptide encoded by the reference nucleic acid molecule.

[0138] When applied to polypeptides, the term "substantially similar" or "substantially similar" means that when the two peptide sequences are optimally aligned, such as by using the program GAP or BESTFIT with default gap weights, they share at least 95% sequence identity, and even more preferably, at least 98% or 99% sequence identity. Preferably, the residue positions that are not identical differ by conservative amino acid substitutions. "Conservative amino acid substitutions" are amino acid substitutions in which an amino acid residue is substituted by another amino acid residue having a similar side chain (R group) with chemical properties (e.g., charge or hydrophobicity). Generally, conservative amino acid substitutions do not substantially change the functional properties of the protein. In the case where two or more amino acid sequences differ from each other by conservative substitutions, the sequence identity percentage or degree of similarity can be adjusted upward to correct for the conservative nature of the substitution. Methods for making this adjustment are well known to those skilled in the art. See, for example, Pearson (1994) Methods Mol. Biol. 24: 307-331. Examples of amino acid groups having chemically similar side chains 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 are cysteine ​​and methionine. Preferred conservative amino acid substitutions are: valine-leucine-isoleucine, phenylalanine-tyrosine, lysine-arginine, alanine-valine, glutamic acid-aspartic acid, and asparagine-glutamine. Alternatively, a conservative substitution is any change with a positive value in the PAM250 log-likelihood matrix as disclosed in Gonnet et al. (1992) Science 256: 1443-1445. A "moderately conservative" substitution is any change with a non-negative value in the PAM250 log-likelihood matrix.

[0139] Sequence analysis software is usually used to measure the sequence similarity of polypeptides, also referred to as sequence identity. Protein analysis software uses similarity metrics assigned to various substitutions, deletions and other modifications (comprising conservative amino acid substitutions) to match similar sequences. For example, GCG software contains programs such as Gap and Bestfit, which can be used with default parameters to determine the sequence homology or sequence identity between closely related polypeptides (such as homologous polypeptides from different organism species) or between wild-type proteins and their mutant proteins. See, for example, 6.1 version GCG. FASTA (a program in 6.1 version GCG) using default or recommended parameters can also be used to compare polypeptide sequences. FASTA (for example, FASTA2 and FASTA3) provides comparison and sequence identity percentage of the best overlapping region between query sequence and search sequence (Pearson (2000), supra). When the sequence of the present invention is compared with a database containing a large number of sequences from different organisms, another preferred algorithm is a computer program BLAST using default parameters, especially BLASTP or TBLASTN. See, e.g., Altschul et al. (1990) J. Mol. Biol. 215:403-410 and Altschul et al. (1997) Nucleic Acids Res. 25:3389-402.

[0140] Bispecific Antigen Binding Molecules

[0141] The antibodies of the present invention may be monospecific, bispecific or multispecific. A multispecific antibody may be specific for different epitopes of a target polypeptide, or may contain an antigen binding domain that is specific for more than one target polypeptide. See, for example, Tutt et al., 1991, J. Immunol. 147:60-69; Kufer et al., 2004, Trends Biotechnol. 22:238-244. The anti-CD28 antibodies of the present invention may be linked or co-expressed with another functional molecule (e.g., another peptide or protein). For example, an antibody or fragment thereof may be functionally linked (e.g., by chemical coupling, genetic fusion, non-covalent association or otherwise) to one or more other molecular entities (e.g., another antibody or antibody fragment) to produce a bispecific or multispecific antibody with a second binding specificity.

[0142] The expression "anti-CD28 antibody" used herein is intended to include monospecific anti-CD28 antibodies and bispecific antibodies comprising a CD28 binding arm and a second arm that binds to a target antigen. Therefore, the present invention includes bispecific antibodies in which one arm of the immunoglobulin binds to human CD28 and the other arm of the immunoglobulin is specific for the target antigen. The target antigen bound by the other arm of the CD28 bispecific antibody can be any antigen expressed on or near a cell, tissue, organ, microorganism or virus, and a targeted immune response to the antigen is required. The CD28 binding arm may include any of the HCVR / LCVR or CDR amino acid sequences as shown in Table 1 herein. In certain embodiments, the CD28 binding arm binds to human CD28 and induces human T cell proliferation.

[0143] In the context of the bispecific antibodies of the invention, where one arm of the antibody binds CD28 and the other arm binds a target antigen, the target antigen may be a tumor-associated antigen, such as PSMA.

[0144] According to certain exemplary embodiments, the present invention comprises 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.

[0145] As used herein, the term "PSMA" refers to human PSMA protein, unless specified as being from a non-human species (e.g., "mouse PSMA," "monkey PSMA," etc.). Human PSMA protein has the amino acid sequence shown in SEQ ID NO: 73, and / or has the amino acid sequence shown in NCBI Accession No. NP_004467.1.

[0146] according to Figure 1In certain exemplary embodiments shown, 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. Bispecific anti-CD28 / PSMA may provide a co-stimulatory signal to effector cells (e.g., T cells). The effector arm has no effect of stimulating T cells in the absence of aggregation. When aggregated, a single effector arm has almost no effect of stimulating T cells. The effector arm stimulates T cells by combining with a targeting arm. The tumor targeting arm may have imperfect tumor specificity. The antigen (e.g., PSMA) that is the target of the targeting arm may be expressed on a portion of a tumor cell. The specificity of the tumor targeting arm may be increased by overlapping with a combination of an anti-CD3 bispecific antigen binding molecule (e.g., an anti-CD3 / PSMA bispecific antibody).

[0147] As used herein, the expression "antigen binding molecule" means a protein, polypeptide or molecular complex comprising or consisting of at least one complementary determining region (CDR), which specifically binds to a particular antigen alone or in combination with one or more additional CDRs and / or framework regions (FRs). In certain embodiments, the antigen binding molecule is an antibody or an antibody fragment, such as those terms defined elsewhere herein.

[0148] As used herein, the expression "bispecific antigen binding molecule" means a protein, polypeptide or molecular complex comprising at least a first antigen binding domain and a second antigen binding domain. Each antigen binding domain within a bispecific antigen binding molecule comprises at least one CDR that specifically binds to a specific antigen 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 different antigen (e.g., PSMA).

[0149] In certain exemplary embodiments of the present invention, bispecific antigen binding molecules are bispecific antibodies.Each antigen binding domain of a bispecific antibody includes 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) including a first antigen binding domain and a second antigen binding domain, the CDR of the first antigen binding domain can be represented by the prefix "D1", and the CDR of the second antigen binding domain can be represented by the prefix "D2".Therefore, the CDR of the first antigen binding domain can be referred to as D1-HCDR1, D1-HCDR2 and D1-HCDR3 in this article; and the CDR of the second antigen binding domain can be referred to as D2-HCDR1, D2-HCDR2 and D2-HCDR3 in this article.

[0150] The first antigen-binding domain and the second antigen-binding domain can be directly or indirectly connected to each other to form a bispecific antigen binding molecule of the present invention. Alternatively, the first antigen-binding domain and the second antigen-binding domain can each be connected to a separate multimerization domain. The association of a multimerization domain with another multimerization domain promotes the association between the two antigen-binding domains, thereby forming a bispecific antigen binding molecule. As used herein, a "multimerization domain" is any macromolecule, protein, polypeptide, peptide or amino acid with the ability to associate with a second multimerization domain having the same or similar structure or composition. For example, the multimerization domain can be an immunoglobulin C or a polypeptide. H A non-limiting example of a multimeric component is the Fc portion of an immunoglobulin (including the C H 2-C H 3 domains), e.g., an IgG selected from the group consisting of isotypes IgG1, IgG2, IgG3, and IgG4, and an Fc domain of any isotype within each isotype group.

[0151] The bispecific antigen binding molecules of the present invention generally include two multimerization domains, for example, two Fc domains, which are each independently a part of a separate antibody heavy chain. The first multimerization domain and the second multimerization domain can be the same IgG isotype, such as, for example, IgG1 / lgG1, IgG2 / lgG2, IgG4 / lgG4. Alternatively, the first multimerization domain and the second multimerization domain can be different IgG isotypes, such as, for example, IgG1 / lgG2, IgG1 / lgG4, IgG2 / lgG4, etc.

[0152] 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 peptide containing cysteine. Other multimerization domains include peptides or polypeptides including or consisting of a leucine zipper, a spiral loop motif, or a coiled coil motif.

[0153] Any bispecific antibody form or technology can be used to prepare the bispecific antigen binding molecules of the present invention.For example, an antibody or fragment thereof with a first antigen binding specificity can be functionally connected (e.g., by chemical coupling, genetic fusion, non-covalent association or otherwise) with one or more other molecular entities (e.g., another antibody or antibody fragment with a second antigen binding specificity) to produce a bispecific antigen binding molecule. Specific exemplary bispecific forms that can be used in the context of the present invention include but are not limited to: for example, based on scFv or double antibody bispecific forms, IgG-scFv fusion, dual variable domains (OVO)-Ig, quadroma, knobs-into-holes, common light chains (e.g., common light chains with knobs-into-holes, etc.), CrossMab, CrossFab, (SEEO) body, leucine zipper, Ouobody, IgG1 / lgG2, dual-acting Fab (OAF)-lgG and Mab 2 Bispecific formats (see, e.g., Klein et al. 2012, mAbs 4:6, 1-11, and references cited therein, for a review of the foregoing formats).

[0154] In the context of the bispecific antigen binding molecules of the present invention, the multimerization domain (e.g., Fc domain) may include one or more amino acid changes (e.g., insertions, deletions, or substitutions) compared to the wild-type, naturally occurring form of the Fc domain. For example, the present invention comprises bispecific antigen binding molecules comprising one or more modifications in the Fc domain, thereby generating a modified Fc domain with a modified binding interaction (e.g., enhanced or weakened) between Fc and FcRn. In one embodiment, the bispecific antigen binding molecule comprises C H Zone 2 or C HA modification in region 3, wherein the modification increases the affinity of the Fc domain for FcRn in an acidic environment (e.g., in an endosome with a pH range of about 5.5 to about 6.0). Non-limiting examples of such Fc modifications include, e.g., modification at position 250 (e.g., E or Q); modification at position 250 and position 428 (e.g., L or F); modification at position 252 (e.g., LN / FIW or T), position 254 (e.g., S or T), and position 256 (e.g., S / R / Q / E / EID or T); or modification at position 428 and / or position 433 (e.g., UR / S / P / Q or K) and / or position 434 (e.g., H / F or V); or modification at position 250 and / or position 428; or modification at position 307 or 308 (e.g., 308F, V308F) and position 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).

[0155] The present invention also includes a bispecific antigen binding molecule comprising a first C H 3 domain and the second Ig C H 3 domains, the first and second C H The three domains differ from each other by at least one amino acid, and wherein the at least one amino acid difference reduces binding of the bispecific antibody to protein A compared to a bispecific antibody lacking the amino acid difference. H 3 domains bind protein A, and the second Ig C H The 3 domain contains a mutation that reduces or eliminates protein A binding, such as the H95R modification (by IMGT exon numbering; H435R by EU numbering). H3 may further include modifications Y96F (by IMGT; Y436F by EU). Additional modifications that may be found within the second CH3 include: D16E, L18M, N44S, K52N, V57M and V821 (by IMGT; D356E, L358M, N384S, K392N, V397M and V4221 by EU) in the case of IgG1 antibodies; N44S, K52N and V821 (IMGT; N384S, K392N and V4221 by EU) in the case of IgG2 antibodies; and Q15R, N44S, K52N, V57M, R69K, E79Q and V821 (by IMGT; Q355R, N384S, K392N, V397M, R409K, E419Q and V4221 by EU) in the case of IgG4 antibodies.

[0156] In certain embodiments, the Fc domain may be chimeric, with a combination of Fc sequences derived from more than one immunoglobulin isotype. For example, a chimeric Fc domain may include a region derived from human IgG1, a region derived from human IgG2, or a region derived from human IgG4. H C in Zone 2 H 2 sequence, and C derived from human IgG1, human IgG2 or human IgG4 H 3 sequence. The chimeric Fc domain may also contain a chimeric hinge region. For example, the chimeric hinge may include an "upper hinge" sequence derived from a human IgG1 hinge region, a human IgG2 hinge region, or a human IgG4 hinge region, combined with a "lower hinge" sequence derived from a human IgG1 hinge region, a human IgG2 hinge region, or a human IgG4 hinge region. Specific examples of chimeric Fc domains that may be included in any antigen-binding molecule described herein include, from N-terminus to C-terminus: [IgG4 C H 1]-[lgG4 upper hinge]-[lgG2 lower hinge]-[lgG4 CH2]-[lgG4 C H 3]. Another example of a chimeric Fc domain that can be included in any of the antigen binding molecules set forth herein includes, from N-terminus to C-terminus: [IgG1 C H 1]-[IgG1 upper hinge]-[IgG2 lower hinge]-[IgG4 C H 2]-[IgG1 C H 3]. These and other examples of chimeric Fc domains that may be included in any antigen binding molecule of the invention are described in WO2014 / 022540A1, and chimeric Fc domains and variants thereof having these general structural arrangements may have altered Fc receptor binding, which in turn affects Fc effector function.

[0157] Sequence variants

[0158] Compared to the corresponding germline sequences derived from a single antigen binding domain, the antibodies and bispecific antigen binding molecules of the present invention may include 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. Such mutations can be easily determined by comparing the amino acid sequences disclosed herein with germline sequences that can be obtained from, for example, public antibody sequence databases. The antigen binding molecules of the present invention may include an antigen binding fragment derived from any one of the exemplary amino acid sequences disclosed herein, wherein one or more amino acid mutations in one or more frameworks and / or CDR regions are mutated to one or more corresponding residues of the germline sequence of the derived antibody, or mutated to one or more corresponding residues of another human germline sequence, or mutated to conservative amino acid substitutions of one or more corresponding germline residues (such sequence changes are collectively referred to herein as "germline mutations"). Starting from the heavy and light chain variable region sequences disclosed herein, a person of ordinary skill in the art can easily produce many antibodies and antigen binding fragments including one or more individual germline mutations or combinations thereof. In certain embodiments, V H Domain and / or V L The framework and / or CDR residues in the domain are all mutated back to the residues found in the original germline sequence of the initial derived antigen-binding domains. In other embodiments, only some residues are mutated back to the original germline sequence, for example, only the residue after the mutation found in the first 8 amino acids of FR1 or in the last 8 amino acids of FR4, or only the residue after the mutation found in CDR1, CDR2 or CDR3. In other embodiments, one or more mutations in one or more framework and / or CDR residues are one or more corresponding residues of different germline sequences (that is, germline sequences different from the germline sequences of the initial derived antigen-binding domains). In addition, the antigen-binding domains of the present invention can contain any combination of two or more germline mutations in the framework and / or CDR regions, for example, wherein some individual residues are mutated into the corresponding residues of a specific germline sequence, and some other residues different from the original germline sequence can remain unchanged or mutate into the corresponding residues of different germline sequences. 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 antagonistic or agonistic biological properties (as the case may be), reduced immunogenicity, etc. Bispecific antigen binding molecules comprising antigen binding domains obtained in this general manner are encompassed by the present invention.

[0159] The present invention also includes antigen binding molecules, wherein one or both of the antigen binding domains 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 antigen binding molecules, which include antigen binding domains having HCVR, LCVR and / or CDR amino acid sequences, relative to any of the HCVR, LCVR and / or CDR amino acid sequences disclosed herein, the HCVR, LCVR and / or CDR amino acid sequences have, for example, 10 or less, 8 or less, 6 or less, 4 or less conservative amino acid substitutions. "Conservative amino acid substitutions" are amino acid substitutions in which an amino acid residue is replaced by another amino acid residue having a similar side chain (R group) with chemical properties (e.g., charge or hydrophobicity). Typically, conservative amino acid substitutions do not substantially change the functional properties of the protein. Examples of amino acid groups with chemically similar side chains 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 are cysteine ​​and methionine. Preferred conservative amino acid substitutions are: valine-leucine-isoleucine, phenylalanine-tyrosine, lysine-arginine, alanine-valine, glutamic acid-aspartic acid, and asparagine-glutamine. Alternatively, a conservative substitution is any change with 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.

[0160] The present invention also comprises antigen binding molecules, which include antigen binding domains having HCVR, LCVR and / or CDR amino acid sequences substantially identical to any one of the HCVR, LCVR and / or CDR amino acid sequences disclosed herein. When referring to amino acid sequences, the term "substantially identical" or "substantially identical" means that when the best alignment is performed by programs such as GAP or BESTFIT using default gap weights, the two amino acid sequences share at least 95% sequence identity, and even more preferably, at least 98% or 99% sequence identity. Preferably, the non-identical residue positions differ due to conservative amino acid substitutions. In the case where two or more amino acid sequences differ from each other due to conservative substitutions, the sequence identity percentage or degree of similarity can be adjusted upward to correct the conservative nature of the substitution. Methods for making this adjustment are well known to those skilled in the art. See, for example, Pearson (1994) Methods in Molecular Biology 24: 307-331.

[0161] Sequence analysis software is usually used to measure the sequence similarity of polypeptides, also referred to as sequence identity. Protein analysis software uses similarity metrics assigned to various substitutions, deletions and other modifications (comprising conservative amino acid substitutions) to match similar sequences. For example, GCG software contains programs such as Gap and Bestfit, which can be used with default parameters to determine the sequence homology or sequence identity between closely related polypeptides (such as homologous polypeptides from different organism species) or between wild-type proteins and their mutant proteins. See, for example, 6.1 version GCG. FASTA (a program in 6.1 version GCG) using default or recommended parameters can also be used to compare polypeptide sequences. FASTA (for example, FASTA2 and FASTA3) provides comparison and sequence identity percentage of the best overlapping region between query sequence and search sequence (Pearson (2000), supra). When the sequence of the present invention is compared with a database containing a large number of sequences from different organisms, another preferred algorithm is a computer program BLAST using default parameters, especially BLASTP or TBLASTN. See, e.g., Altschul et al. (1990) J. Molecular Biology 215:403-410 and Altschul et al. (1997) Nucleic Acids Res. 25:3389-402.

[0162] pH-dependent binding

[0163] The present invention includes anti-CD28 / anti-PSMA bispecific antigen binding molecules having pH-dependent binding properties. For example, the anti-CD28 antibodies of the present invention may exhibit reduced binding to CD28 at acidic pH compared to neutral pH. Alternatively, the anti-PSMA antibodies of the present invention may exhibit enhanced binding to PSMA at acidic pH compared to neutral pH. The expression "acidic pH" includes a pH value of less than 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 less). As used herein, the expression "neutral pH" means a pH of about 7.0 to about 7.4. The expression "neutral pH" encompasses pH values ​​of about 7.0, 7.05, 7.1, 7.15, 7.2, 7.25, 7.3, 7.35 and 7.4.

[0164] In certain instances, "reduced binding at acidic pH compared to neutral pH" refers to the K of an antibody that binds to its antigen at acidic pH. D The K value is similar to that of an antibody bound to its antigen at neutral pH. D For example, if an antibody or antigen-binding fragment thereof exhibits an acidic / neutral K of about 3.0 or greater, D For purposes of the present invention, an antibody or antigen-binding fragment thereof may be considered to exhibit "reduced binding to CD28 at acidic pH compared to neutral pH" if the ratio of acidic / neutral K to CD28 is greater than or equal to 1.5. D The ratio may be about 3.0, 3.5, 4.0, 4.5, 5.0, 5.5, 6.0, 6.5, 7.0, 7.5, 8.0, 8.5, 9.0, 9.5, 10.0, 10.5, 11.0, 11.5, 12.0, 12.5, 13.0, 13.5, 14.0, 14.5, 15.0, 20.0, 25.0, 30.0, 40.0, 50.0, 60.0, 70.0, 100.0 or more.

[0165] Antibodies with pH-dependent binding properties can be obtained, for example, by screening for a population of antibodies that have reduced (or enhanced) binding to a specific antigen at an acidic pH compared to a neutral pH. In addition, modifications of the antigen-binding domain at the amino acid level can produce antibodies with pH-dependent properties. For example, by replacing one or more amino acids of an antigen-binding domain (e.g., within a CDR) with histidine residues, antibodies with reduced binding to an antigen at an acidic pH relative to a neutral pH can be obtained.

[0166] Antibodies including Fc variants

[0167] According to certain embodiments of the present invention, there is provided an anti-CD28 / anti-PSMA bispecific antigen binding molecule comprising an Fc domain comprising one or more mutations that enhance or attenuate (e.g., at acidic pH compared to neutral pH) the binding of the antibody to the FcRn receptor. H Zone 2 or C H Region 3 includes mutated antibodies and antigen-binding molecules, wherein one or more mutations increase the affinity of the Fc domain for FcRn in an acidic environment (e.g., in an endosome with a pH range of about 5.5 to about 6.0). When such mutations are administered to an animal, they can result in an increase in the serum half-life of the antibody. Non-limiting examples of such Fc modifications include, e.g., modification at position 250 (e.g., E or Q); modification at position 250 and position 428 (e.g., L or F); modification at position 252 (e.g., L / Y / F / W or T), position 254 (e.g., S or T), and position 256 (e.g., S / R / Q / E / D or T); or modification at position 428 and / or position 433 (e.g., H / L / R / S / P / Q or K) and / or position 434 (e.g., H / F or Y); or modification at position 250 and / or position 428; or modification at position 307 or 308 (e.g., 308F, V308F) and position 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).

[0168] For example, the present invention comprises an anti-CD28 / anti-PSMA bispecific antigen binding molecule comprising an Fc domain comprising one or more mutation pairs or groups 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 within the antibody variable domains disclosed herein are within the scope of the present invention.

[0169] Biological properties of antibodies and antigen-binding molecules

[0170] The present invention comprises antibodies and antigen-binding fragments thereof that bind to human CD28 / anti-PSMA with high affinity. The present invention also comprises antibodies and antigen-binding fragments thereof that bind to human CD28 / anti-PSMA with moderate or low affinity, depending on the desired therapeutic context and specific targeting properties. For example, in the context of a bispecific antigen-binding molecule, in which one arm binds CD28 and the other arm binds a target antigen (e.g., PSMA), it may be desirable that the target antigen-binding arm binds to the target antigen with high affinity, while the anti-CD28 arm binds to CD28 only with 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-targeted CD28 binding and the adverse side effects associated therewith.

[0171] According to certain embodiments, the invention comprises antibodies and antigen-binding fragments of antibodies that have a K of less than about 210 nM. D Binding to human CD28 (e.g., at 25°C) as measured, for example, by surface plasmon resonance using the assay format defined in Example 3 herein. In certain embodiments, the 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. D Binding to CD28 as measured, for example, by surface plasmon resonance using the assay format defined in Example 3 herein or a substantially similar assay. 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 Binds to CD28.

[0172] The present invention also includes antibodies and antigen-binding fragments thereof that bind to CD28, wherein the dissociation half-life (t 1 / 2) is greater than about 3.5 minutes, as measured, for example, by surface plasmon resonance at 25°C or 37°C using an assay format defined in Example 3 herein or a substantially similar assay. In certain embodiments, the antibodies or antigen-binding fragments of the invention bind to CD28, wherein t 1 / 2 is 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, for example, by surface plasmon resonance at 25°C or 37°C using an assay format defined in Example 3 herein or a substantially similar assay.

[0173] The present invention comprises bispecific antigen binding molecules (e.g., bispecific antibodies) that can simultaneously bind 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 bispecific antigen binding molecules bind to cells expressing CD28 and / or PSMA can be assessed by fluorescence activated cell sorting (FACS), as shown in Example 4 herein. For example, the present invention comprises bispecific antigen binding molecules that specifically bind to human cell lines expressing CD28 but not PSMA (e.g., HEK293 engineered to express CD28) and human prostate cancer cell lines expressing PSMA but not CD28 (e.g., C4-2). The present invention comprises bispecific antigen binding molecules that bind to any of the above cells and cell lines, wherein EC 50 The value is about 9.6x10 -9 To about 3.5x10 -10 or less, as determined using a FACS assay as set forth in Example 4, or a substantially similar assay.

[0174] The present invention also provides an anti-CD28 / anti-PSMA bispecific antigen binding molecule that induces or increases T cell-mediated tumor cell killing. For example, the present invention comprises an anti-CD28xPSMA antibody that induces or increases T cell-mediated tumor cell killing, wherein EC 50 Less than about 78 pM, as measured, for example, in an in vitro T cell-mediated tumor cell killing assay using the assay format defined in Example 6 herein (e.g., assessing the extent of killing of C4-2 tumor cells by human PBMCs in the presence of an anti-CD28xPSMA antibody) or a substantially similar assay. In certain embodiments, the antibodies or antigen-binding fragments of the invention induce T cell-mediated tumor cell killing (e.g., PBMC-mediated killing of C4-2 cells), wherein EC 50A value 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, for example, by an in vitro T cell-mediated tumor cell killing assay using the assay format defined in Example 6 herein or a substantially similar assay.

[0175] The present invention also comprises an anti-CD28 / anti-PSMA bispecific antigen binding molecule that binds to human T cells expressing CD28, wherein EC 50 The value is between 1.0 pM and 1000 nM. In certain embodiments, the anti-CD28 / anti-PSMA bispecific antigen binding molecule binds to human T cells expressing CD28, wherein EC 50 The value is between 48 nM and 180 nM. For example, the present invention comprises an anti-CD28 / anti-PSMA bispecific antigen binding molecule that binds to human T cells expressing CD28, wherein EC 50 The values ​​are 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.

[0176] The present invention also comprises an anti-CD28 / anti-PSMA bispecific antigen binding molecule that exhibits one or more properties selected from the group consisting of: (a) inducing T cell proliferation in vitro (see, e.g., Example 8 herein); (b) activating T cells, inducing upregulation of CD25 and PD-1 in human PBMCs (see, e.g., Example 8 herein); (c) increasing human T cell-mediated cytotoxicity against PSMA-expressing cell lines (see, e.g., Example 8 herein); (d) inducing primitive primate T cell-mediated cytotoxicity on PSMA-expressing cell lines; (e) depletion of tumor cells in mice (e.g., Example 10 herein); (f) enhancement of tumor clearance in mice (e.g., Example 10 herein); (g) no induction of cytokine storm (e.g., Example 10 herein); (h) no induction of systemic T cell effects in cynomolgus monkeys (e.g., Example 11 herein); (i) enhancement of the effect of PD-1 blockade on tumor cell killing induced by T cell activation (e.g., Example 13 herein); (j) enhancement of the expansion of memory T cells (e.g., Example 13 herein).

[0177] The present invention comprises anti-CD28 / anti-PSMA bispecific antigen binding molecules capable of depleting 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., at a dose of 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) results in a decrease in the number of tumor cells in the subject.

[0178] Epitope mapping and related technologies

[0179] The epitope on CD28 or PSMA to which the antigen-binding molecules of the present invention bind may consist of a single continuous sequence of 3 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 protein or PSMA protein. Alternatively, the epitope may consist of multiple non-continuous amino acids (or amino acid sequences) of CD28 or PSMA. The antibodies of the present invention may interact with the amino acids contained in the CD28 monomer, or may interact with the amino acids on the two different CD28 chains of the 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 more than one epitope. Therefore, different antibodies may bind to different regions on the antigen and may have different biological effects. An epitope may be conformational or linear. Conformational epitopes are produced by spatial juxtaposition of amino acids from different segments of a linear polypeptide chain. A linear epitope is an epitope produced by adjacent amino acid residues in a polypeptide chain. In some cases, an epitope may comprise a portion of a sugar, phosphoryl or sulfonyl group on the antigen.

[0180] Various techniques known to those of ordinary skill 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, conventional cross-blocking assays, such as those described in: Antibodies ( Antibodies), Harlow and Lane (Cold Spring Harbor Press, Cold Spring Harb., New York (NY)); point mutagenesis (e.g., alanine scanning mutagenesis, arginine scanning mutagenesis, etc.); peptide blot analysis (Reineke, 2004, Methods in Molecular Biology 248:443-463); protease protection; and peptide cleavage analysis. In addition, methods such as epitope excision, epitope extraction, and chemical modification of antigens can be used (Tomer, 2000, Protein Science, 9:487-496). Another method that can be used to identify amino acids within a polypeptide that interacts with an antibody is hydrogen / deuterium exchange detected by mass spectrometry. In general, hydrogen / deuterium exchange methods involve deuterium labeling of the protein of interest and then binding the antibody to the deuterium-labeled protein. Next, the protein / antibody complex is transferred to water to allow hydrogen-deuterium exchange to occur at all residues except the residues protected by the antibody (which remain deuterium-labeled). After dissociation of the antibody, the target protein is subjected to protease cleavage and mass spectrometry analysis to reveal the deuterium-labeled residues corresponding to the specific amino acids that interact with the antibody. See, for example, Ehring (1999) Analytical Biochemistry, 267 (2): 252-259; Engen and Smith (2001) Anal. Chem., 73: 256A-265A. X-ray crystal structure analysis can also be used to identify amino acids in polypeptides that interact with antibodies.

[0181] The present invention further encompasses anti-CD28 and anti-PSMA antibodies that bind to the same epitope as any of the specific exemplary antibodies described herein (e.g., antibodies comprising any of the amino acid sequences shown in Table 1 herein). Likewise, the present invention also encompasses anti-CD28 and / or anti-PSMA antibodies that compete for binding to CD28 and / or PSMA with any of the specific exemplary antibodies described herein (e.g., antibodies comprising any of the amino acid sequences shown in Table 1 herein).

[0182] The present invention also encompasses a bispecific antigen binding molecule comprising a first antigen binding domain that specifically binds to human CD28 and a second antigen binding fragment that specifically binds to 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 wherein 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.

[0183] Likewise, the present invention also encompasses a bispecific antigen binding molecule comprising a first antigen binding domain that specifically binds to human CD28 and a second antigen binding fragment that specifically binds to 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 wherein the second antigen binding domain competes for binding to PSMA with any of the specific exemplary PSMA-specific antigen binding domains described herein.

[0184] By using conventional methods known in the art, it is easy to determine whether a specific antigen binding molecule (e.g., antibody) or its antigen binding domain binds to the same epitope as a reference antigen binding molecule of the present invention, or whether it competes with the reference antigen binding molecule for binding. For example, in order 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 allowed to bind to the CD28 protein (or PSMA protein). Next, the ability of the test antibody to bind to the CD28 (or PSMA) molecule is evaluated. If the test antibody is able to 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 different epitopes of CD28 (or PSMA) with the reference bispecific antigen binding molecule. On the other hand, if the test antibody cannot bind to the CD28 (or PSMA) molecule after saturation binding with the reference bispecific antigen binding molecule, the test antibody can bind to the same epitope of CD28 (or PSMA) as the epitope bound by the reference bispecific antigen binding molecule of the present invention. Then other routine experiments (e.g., peptide mutations and binding analysis) can be performed to confirm whether the observed lack of binding of the test antibody is actually due to binding to the same epitope with the reference bispecific antigen binding molecule, or whether steric blocking (or other phenomena) causes the observed lack of binding. Such experiments can be performed using ELISA, RIA, Biacore, flow cytometry, or any other quantitative or qualitative antibody binding assays available in the art. According to certain embodiments of the present invention, if, for example, 1-fold, 5-fold, 10-fold, 20-fold, or 100-fold excess of one antigen-binding protein inhibits at least 50% of another antigen-binding protein, but preferably, 75%, 90%, or even 99% binding (as measured in a competitive binding assay), then two antigen-binding proteins bind to the same (or overlapping) epitope (see, e.g., Junghans et al., Cancer Res. 1990:50:1495-1502). Alternatively, two antigen binding proteins are considered to bind to the same epitope if substantially all amino acid mutations in the antigen that reduce or eliminate binding of one antigen binding protein also reduce or eliminate binding of the other antigen binding protein. Two antigen binding proteins are considered to have "overlapping epitopes" if only a subset of amino acid mutations that reduce or eliminate binding of one antigen binding protein reduce or eliminate binding of the other antigen binding protein.

[0185] To determine whether an antibody or its antigen binding domain competes with a reference antigen binding molecule for binding, the above binding method is performed in two orientations: In the first orientation, the reference antigen binding molecule is allowed to bind to the CD28 protein (or PSMA protein) under saturation conditions, and then the binding of the test antibody to the CD28 (or PSMA) molecule is evaluated. In the second orientation, the test antibody is allowed to bind to the CD28 (or PSMA) molecule under saturation conditions, and then the binding of the reference antigen binding molecule to the CD28 (or PSMA) molecule is evaluated. In both orientations, if only the first (saturated) antigen binding molecule is able to bind to the CD28 (or PSMA) molecule, it can be concluded that the test antibody competes with the reference antigen binding molecule for binding to CD28 (or PSMA). As will be appreciated by those of ordinary skill in the art, an antibody that competes with a reference antigen binding molecule for binding may not necessarily bind to the same epitope as the reference antibody, but may spatially block the binding of the reference antibody by binding to an overlapping or adjacent epitope.

[0186] Preparation of antigen binding domains and construction of bispecific molecules

[0187] Antigen binding domains with specificity for a particular antigen can be prepared by any antibody production technique known in the art. Once obtained, two different antigen binding domains with specificity for two different antigens (e.g., CD28 and PSMA) can be appropriately arranged relative to each other to produce the bispecific antigen binding molecules of the present invention using conventional methods. (Discussion of exemplary bispecific antibody forms 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 chains and light chains) of the multispecific antigen binding molecules of the present invention are derived from chimeric antibodies, humanized antibodies, or fully human antibodies. Methods for preparing such antibodies are well known in the art. For example, VELOCIMMUNE can be used TM One or more heavy chains and / or light chains of the bispecific antigen-binding molecules of the present invention are prepared using VELOCIMMUNE TM The invention relates to the use of chimeric antibodies with high affinity for a specific antigen (e.g., CD28 or PSMA) having a human variable region and a mouse constant region using the bispecific antibody generation technology (or any other human antibody generation technology). The antibodies are characterized and selected for desired properties, including affinity, selectivity, epitope, etc. The mouse constant region is replaced with the desired human constant region to generate fully human heavy and / or light chains that can be incorporated into the bispecific antigen binding molecules of the invention.

[0188] Human bispecific antigen binding molecules can be manufactured using genetically engineered animals.For example, genetically modified mice that cannot rearrange and express endogenous mouse immunoglobulin light chain variable sequences can be used, wherein the mouse only expresses one or two human light chain variable domains encoded by human immunoglobulin sequences, and the human immunoglobulin sequences are operably connected to the mouse κ constant gene at the endogenous mouse κ locus.This genetically modified mouse can be used to produce a fully human bispecific antigen binding molecule including two different heavy chains, the heavy chain is associated with the same light chain, and the light chain includes a variable domain derived from one of two different human light chain variable region gene segments.(For a detailed discussion of the use of this engineered mouse and its production of bispecific antigen binding molecules, see, for example, US2011 / 0195454).

[0189] Bioequivalence

[0190] The present invention encompasses antigen binding molecules having an amino acid sequence that is different from the amino acid sequence of the antibody but retains the ability to bind to CD28 and / or PSMA. When compared to the parent sequence, such variant molecules include one or more additions, deletions or substitutions of amino acids, but exhibit biological activity substantially the same as the biological activity of the antigen binding molecule. Similarly, the DNA sequences encoding the antigen binding molecules of the present invention encompass such sequences: compared to the disclosed sequences, they include one or more additions, deletions or substitutions of nucleotides, but encode antigen binding molecules that are substantially biologically equivalent to the antigen binding molecules described in the present invention. Examples of such variant amino acids and DNA sequences are discussed above.

[0191] The present invention includes antigen binding molecules that are bioequivalent to any exemplary antigen binding molecules shown herein. If, for example, two antigen binding proteins or antibodies are drug equivalents or drug substitutes that absorption rate and absorption extent do not show significant differences when using (single dose or multiple doses) with identical molar doses under similar experimental conditions, then it is considered that the two antigen binding proteins or antibodies are bioequivalent. If some antibodies are equivalent in absorption extent but their absorption rate is not equivalent, then it is considered that the antibody is an equivalent or drug substitute, and it can be considered that the antibody is bioequivalent, because this intentional and the difference in absorption rate reflected on the label is not necessary for reaching effective body drug concentration when, for example, long-term use, and is considered to be medically insignificant for the specific drug product studied.

[0192] In one embodiment, two antigen binding proteins are bioequivalent if there are no clinically meaningful differences in safety, purity, and potency of the two antigen binding proteins.

[0193] In one embodiment, two antigen binding proteins are bioequivalent if patients can switch between the reference product and the biological product one or more times without an expected increased risk of adverse effects, including clinically significant changes in immunogenicity or reduced effectiveness, compared to continued treatment in the absence of switching between the reference product and the biological product.

[0194] In one embodiment, two antigen binding proteins are bioequivalent if they both act by one or more common mechanisms of action for one or more conditions of use (provided such mechanisms are known).

[0195] Bioequivalence can be demonstrated by in vivo and in vitro methods. Bioequivalence measurements include, for example: (a) in vivo tests performed in humans or other mammals, where the concentration of the antibody or its metabolites is measured over time in blood, plasma, serum or other biological fluids; (b) in vitro tests that are correlated with and reasonably predictive of human in vivo bioavailability data; (c) in vivo tests performed in humans or other mammals, where appropriate acute pharmacological effects of the antibody (or its target) are measured over time; and (d) in well-controlled clinical trials to establish the safety, efficacy or bioavailability or bioequivalence of the antibody.

[0196] Bioequivalent variants of the exemplary bispecific antigen-binding molecules described herein can be constructed by, for example, making various substitutions to residues or sequences or deleting terminal or internal residues or sequences that are not necessary for biological activity. For example, cysteine ​​residues that are not necessary for biological activity can be deleted or replaced with other amino acids to prevent the formation of unnecessary or incorrect intramolecular disulfide bonds during renaturation. In other cases, bioequivalent antibodies can include the exemplary bispecific antigen-binding molecules described herein that include amino acid changes that alter the glycosylation characteristics of the antibody, for example, mutations that eliminate or remove glycosylation.

[0197] Species selectivity and cross-species reactivity

[0198] 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 bind to CD28 from one or more non-human species; and / or antigen binding molecules that bind to human PSMA and bind to PSMA from one or more non-human species.

[0199] According to certain exemplary embodiments of the present invention, an antigen binding molecule is provided, which binds to human CD28 and / or human PSMA, and may bind to or not bind to one or more of mouse, rat, guinea pig, hamster, gerbil, pig, cat, dog, rabbit, goat, sheep, cattle, horse, camel, cynomolgus monkey, marmoset, rhesus monkey or chimpanzee CD28 and / or PSMA, as appropriate. For example, in a specific exemplary embodiment of the present invention, a bispecific antigen binding molecule is provided, which includes 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.

[0200] Immunoconjugates

[0201] The present invention encompasses antigen binding molecules conjugated to a therapeutic moiety ("immunoconjugate"), such as a cytotoxin, a chemotherapeutic drug, an immunosuppressant, or a radioisotope. Cytotoxic agents include any agent that is harmful to cells. Examples of suitable cytotoxic and chemotherapeutic agents for forming immunoconjugates are known in the art (see, e.g., WO 05 / 103081).

[0202] Therapeutic Formulations and Administration

[0203] The present invention provides pharmaceutical compositions comprising 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, Pennsylvania. These formulations include, for example, powders, pastes, ointments, gels, waxes, oils, lipids, lipids (cations or anions) containing vesicles (such as LIPOFECTIN TM , Life Technologies, Carlsbad, CA), DNA conjugates, anhydrous absorption creams, oil-in-water and water-in-oil emulsions, polyethylene glycol emulsions (polyethylene glycol of various molecular weights), semisolid gels, and semisolid mixtures containing polyethylene glycol. See, e.g., Powell et al. "Compendium of excipients for parenteral formulations," PDA (1998), J Pharm Sci Technol, 52: 238-311.

[0204] The dosage of the antigen binding molecules applied to the patient can vary according to the patient's age and body type, target disease, illness, route of administration, etc. The preferred dosage is usually calculated according to body weight or body surface area. When the bispecific antigen binding molecules of the present invention are used for the treatment of adult patients, it may be advantageous to administer the bispecific antigen binding molecules of the present invention intravenously, usually with a single dose of about 0.01 mg / kg to about 20 mg / kg body weight, more preferably, about 0.02 mg / kg to about 7 mg / kg body weight, about 0.03 mg / kg to about 5 mg / kg body weight or about 0.05 mg / kg to about 3 mg / kg body weight. Depending on the severity of the disease, the frequency and duration of treatment can be adjusted. The effective dose and schedule of the bispecific antigen binding molecules can be determined empirically; for example, patient progress can be monitored by regular assessment, and the dosage can be adjusted accordingly. In addition, interspecific analogies of dosage can be performed using methods well known in the art (e.g., Mordenti et al., 1991, "Pharmaceut. Res.", 8: 1351).

[0205] Various delivery systems are known and can be used to administer the pharmaceutical compositions of the present invention, for example, encapsulated in liposomes, microparticles, microcapsules, recombinant cells capable of expressing mutant viruses, receptor-mediated endocytosis (see, for example, Wu et al., 1987, J. Biol. Chem. 262: 4429-4432). Methods of introduction include, but are not limited to, intradermal, intramuscular, intraperitoneal, intravenous, subcutaneous, intranasal, epidural, and oral routes. The composition can be administered by any convenient route, for example, by infusion or bolus injection, by absorption through epithelial or mucocutaneous linings (e.g., oral mucosa, rectal and intestinal mucosa, etc.), and can be administered together with other bioactive agents. Administration can be systemic or local.

[0206] The pharmaceutical compositions of the present invention can be delivered subcutaneously or intravenously with a standard needle and syringe. In addition, for subcutaneous delivery, pen delivery devices are readily applicable to deliver the pharmaceutical compositions of the present invention. Such pen delivery devices can be reusable or disposable. Reusable pen delivery devices typically utilize a replaceable cartridge containing the pharmaceutical composition. Once the entire pharmaceutical composition within 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 a disposable pen delivery device, there is no replaceable cartridge. Instead, the disposable pen delivery device is preloaded with the pharmaceutical composition held in a reservoir within the device. Once the pharmaceutical composition in the reservoir is emptied, the entire device is discarded.

[0207] Many reusable pen delivery devices and auto-injector delivery devices are useful for subcutaneous delivery of the pharmaceutical compositions of the present invention. Examples include, but are not limited to, AUTOPEN TM (Owen Mumford, Inc., Woodstock, UK), DISETRONIC TM pen (Disetronic Medical Systems, Bergdorf, Switzerland), HUMALOGMIX 75 / 25 TM Pen, HUMALOG TM Pen, HUMALIN 70 / 30 TM PEN (Eli Lilly and Co., Indianapolis, Indiana), NOVOPEN TM I, II and III (Novo Nordisk, Copenhagen, Denmark), NOVOPEN JUNIOR TM (Novo Nordisk, Copenhagen, Denmark), BD TM PEN (Becton Dickinson, Franklin Lakes, NJ), OPTIPEN TM 、OPTIPEN PRO TM 、OPTIPEN STARLET TM and OPTICLK TM (Sanofi-Aventis, Frankfurt, Germany), to name a few. Examples of disposable pen delivery devices for subcutaneous delivery of the pharmaceutical compositions of the present invention include, but are not limited to, SOLOSTAR TM Pen (Sanofi), FLEXPEN TM (Novo Nordisk) and KWIKPEN TM (Eli Lilly), SURECLlCK TM Autoinjector (Amgen, Thousand Oaks, CA), PENLET TM (Haselmeier, Stuttgart, Germany), EPIPEN (Dey, LP), and HUMIRA TM PEN (Abbott Labs, Abbott Technology Park, IL), to name a few.

[0208] In some cases, the pharmaceutical composition can be delivered in a controlled release system. In one embodiment, a pump can be used (see Langer, supra; Sefton, 1987, CRC: Biomedical Engineering Review (CRC Crit. Ref. Biomed.) Eng.), 14: 201). In another embodiment, a polymer material can be used; see, Medical Applications of Controlled Release, Langer and Wise (ed.), 1974, CRC Press, Boca Raton, Florida. In yet another embodiment, a controlled release system can be placed near the target of the composition, so only a small portion of the systemic dose is required (see, for example, Goodson, 1984, Medical Applications of Controlled Release, supra, Vol. 2, pp. 115-138). Other controlled release systems are discussed in the review of Langer, 1990, Science 249: 1527-1533.

[0209] Injectable preparations can include dosage forms for intravenous, subcutaneous, intradermal and intramuscular injections, instillations, etc. These injectable preparations can be prepared by known methods. For example, injectable preparations can be prepared by, for example, dissolving, suspending or emulsifying the above-mentioned antibody or its salt in a sterile aqueous medium or a conventional oily medium for injection. As an aqueous medium for injection, there are, for example, physiological saline, isotonic solutions containing glucose and other adjuvants, etc., which can be used in combination with suitable solubilizers (such as alcohol (for example, ethanol), polyols (for example, propylene glycol, polyethylene glycol)), nonionic surfactants [for example, polysorbate 80, HCO-50 (polyoxyethylene (50mol) adducts of hydrogenated castor oil)], etc. As an oily medium, for example, sesame oil, soybean oil, etc. can be used, which can be used in combination with solubilizers such as benzyl benzoate, benzyl alcohol, etc. The injection prepared in this way is preferably filled in a suitable ampoule.

[0210] Advantageously, the pharmaceutical composition for oral or parenteral use is prepared into a unit dosage form suitable for matching a certain dose of active ingredient. Such unit dosage forms include, for example, tablets, pills, capsules, injections (ampoules), suppositories, etc. The amount of the above-mentioned antibody contained is usually about 5 mg to about 500 mg per unit dosage form; especially in the form of injection, the content of the above-mentioned antibody is preferably about 5 mg to about 100 mg, and for other dosage forms, about 10 mg to about 250 mg.

[0211] Therapeutic uses of antigen binding molecules

[0212] The present invention includes methods comprising administering a therapeutic composition to a subject in need thereof, the 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 include any antibody or bispecific antigen binding molecule disclosed herein and a pharmaceutically acceptable carrier or diluent. As used herein, the expression "subject in need thereof" means a human or non-human animal (e.g., a subject expressing a tumor or suffering from any cancer mentioned below) that exhibits one or more symptoms or signs of cancer or that otherwise benefits from inhibition or reduction of PSMA activity or depletion of PSMA+ cells.

[0213] The antibodies and bispecific antigen binding molecules of the present invention (and therapeutic compositions comprising the same) are particularly useful for treating any disease or condition in which stimulation, activation and / or targeting of an immune response would be beneficial. Specifically, 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 condition associated with or mediated by the expression or activity of PSMA or the proliferation of PSMA+ cells. The mechanism of action for achieving the therapeutic methods of the present invention includes killing cells expressing PSMA in the presence of effector cells (e.g., T cells). Cells expressing PSMA that can be inhibited or killed using the bispecific antigen binding molecules of the present invention include, for example, tumorigenic prostate cells.

[0214] The antigen binding molecules of the present invention can be used to treat primary and / or metastatic tumors occurring in, for example, colon cancer, lung cancer, breast cancer, kidney cancer, and bladder cancer subtypes. According to certain exemplary embodiments, the bispecific antigen binding molecules of the present invention are used to treat prostate cancer.

[0215] The present invention also encompasses methods for treating residual cancer in a subject.As used herein, the term "residual cancer" means the presence or persistence of one or more cancer cells in a subject after treatment with an anti-cancer therapy.

[0216] According to certain aspects, the present invention provides a method for treating a disease or condition associated with PSMA expression (e.g., prostate cancer), the method comprising administering to the subject one or more bispecific antigen binding molecules described elsewhere herein after showing that the subject is unresponsive to other types of anti-cancer therapy. For example, the present invention includes a method for treating prostate cancer, the method comprising administering to the patient an anti-CD28 / anti-PSMA bispecific antigen binding molecule 1 day, 2 days, 3 days, 4 days, 5 days, 6 days, 1 week, 2 weeks, 3 weeks, 4 weeks, 2 months, 4 months, 6 months, 8 months, 1 year or more after the subject receives standard care for cancer (e.g., prostate cancer) patients. In other aspects, the subject is initially administered a bispecific antigen binding molecule of the invention (anti-CD28 / anti-PSMA bispecific antigen binding molecule) comprising an IgG4 Fc domain at one or more time points (e.g., to provide a robust initial depletion of prostate cancer cells), followed by administration of an equivalent bispecific antigen binding molecule comprising a different IgG domain (e.g., an IgG1 Fc domain) at a subsequent time point. It is foreseeable that the anti-CD28 / anti-PSMA antibodies of the present invention can be used in combination with other bispecific antigen binding molecules, such as anti-PSMA / anti-CD3 bispecific antibodies. It is also foreseeable that the bispecific antibodies of the present invention will be used in combination with checkpoint inhibitors, for example, checkpoint inhibitors targeting PD-1 and CTLA-4 and other targets. It can be advantageous to combine two bispecific antibodies targeting the same tumor antigen (e.g., PSMA), but one of the bispecific antibodies targets CD3 on T cells and the other bispecific antibody targets co-stimulatory molecules such as CD28. This combination can be used alone to enhance tumor cell killing, or can be used in combination with checkpoint inhibitors.

[0217] Combination therapies and formulations

[0218] The invention encompasses 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 components, as well as therapeutic methods comprising administering such combinations to a subject in need thereof.

[0219] Exemplary additional therapeutic agents that can be combined or administered in combination with the antigen binding molecules of the invention include, for example, chemotherapy, radiotherapy, checkpoint inhibitors targeting PD-1 (e.g., anti-PD-1 antibodies such as pembrolizumab or nivolumab; see also US9,987,500), CTLA-4, LAG3, TIM3, etc., co-stimulatory agonist bivalent antibodies targeting molecules such as GITR, OX40, 4-1BB, etc., CD3x bispecific antibodies (see, e.g., WO2017 / 053856A1, WO2014 / 047231A1, WO2018 / 067331A1, and WO2018 / 058001A1), other antibodies targeting PSMA x CD3 (see, e.g., WO2017 / 023761A1), and other co-stimulatory CD28x bispecific antibodies.

[0220] Other agents that can 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, IL-18, or their respective receptors. The pharmaceutical compositions of the invention (e.g., pharmaceutical compositions comprising anti-CD28 / anti-PSMA bispecific antigen binding molecules as disclosed herein) can also be administered as part of a treatment regimen that includes one or more therapeutic combinations selected from the following: "ICE": ifosfamide (e.g., ), carboplatin (e.g., ), etoposide (e.g., VP-16); "DHAP": dexamethasone (e.g., ), cytarabine (e.g., cytosine arabinoside, ara-C), cisplatin (e.g., -AQ); and "ESHAP": etoposide (e.g., VP-16), methylprednisolone (e.g. ), high-dose cytarabine, cisplatin (e.g. -AQ).

[0221] The present invention also comprises a therapeutic combination, the therapeutic combination comprising any antigen binding molecule mentioned herein and an inhibitor of one or more of VEGF, Ang2, DLL4, EGFR, ErbB2, ErbB3, ErbB4, EGFRvlll, cMet, IGF1R, B-raf, PDGFR-o, PDGFR-I3, FOLH1, PRLR, STEAP1, STEAP2, TMPRSS2, MSLN, CA9, uroplakin or any of the above cytokines, wherein the inhibitor is an aptamer, an antisense molecule, a ribozyme, siRNA, a peptibody, a nanobody or an antibody fragment (e.g., a Fab fragment; a F(ab')2 fragment; a Fd fragment; a Fv fragment; a scFv; a dAb fragment; or other engineered molecules, such as a double antibody, a three-antibody, a four-antibody, a mini-antibody and a minimum recognition unit). The antigen binding molecules of the present invention can also be administered in combination with antiviral agents, antibiotics, analgesics, corticosteroids and / or NSAIDs and / or co-formulated. The antigen binding molecules of the invention may also be administered as part of a treatment regimen that also includes radiation therapy and / or conventional chemotherapy, or treatment with biologics (including checkpoint inhibitors or other bispecific antibodies).

[0222] 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 TM); alkyl sulfonates, such as busulfan, improsulfan, and bosufan; aziridines, such as benzodepa, carboquinone, metodepa, and uredepa; ethyleneimines and methylmelamines, including hexamethylmelamine, triethylenemelamine, triethylenephosphoramide, triethylenethiophosphoramide, and trimethylolomelamine; nitrogen mustards, such as chlorambucil, naphthyl mustard, cholophosphamide, estramustine, ifosfamide, dichloromethane, and chlorambucil. diethylamine, oxychloromethyl diethylamine hydrochloride, melphalan, nebixin, phenylephrine, prednimustine, trofosfamide, uracil nitrogen mustard; nitrosoureas, such as carmustine, chlorozotocin, fotemustine, lomustine, nimustine, ranimustine; antibiotics, such as aclarubicin, actinomycin, anthramycin, azaserine, bleomycin, cactinomycin, calicheamicin, calabikine, carminomycin, carmoglobin, chromomycin, actinomycin D, daunorubicin, detox amycin, 6-diazo-5-oxo-L-norleucine, doxorubicin, epirubicin, esorubicin, demethoxydaunorubicin, mexilomycin, mitomycins, mycophenolic acid, nogamycin, olive mycins, peplomycin, potfiromycin, puromycin, triferric doxorubicin, rhodorubicin, streptozocin, streptozotocin, tuberculocin, ubenimex, netastatin, daunorubicin; antimetabolites, such as methotrexate and 5-fluorouracil (5-FU); folic acid analogs, such as Dimethoate, methotrexate, pteropterin, trimetrexate; purine analogs, such as fludarabine, 6-mercaptopurine, thiopurine, thioguanine; pyrimidine analogs, such as ancitabine, azacitidine, 6-azauridine, carmofur, cytarabine, dideoxyuridine, doxifluridine, enocitabine, floxuridine; androgens, such as caprotestosterone, drostanolone propionate, cyclothiodine, melastosane, testolactone; antiadrenal drugs, such as aminoglutethimide, mitotane, trilostane; folic acid supplements, such as folinic acid (frolinic acid) acid); acetoglucuronide; aldehyde phosphamide glycoside; aminolevulinic acid; amsacrine; bacitracin; bisantrene; edatrexate; defosfamide; demecosine; desacquinone; eflornithine; elitrilonium acetate; etogluc; gallium nitrate; hydroxyurea; lentinan; lonidamine; mitoguanidine; mitoxantrone; mopidarol; diamine nitracrine; pentostatin; methamine nitrogen mustard; pirarubicin; podophyllic acid; 2-ethylhydrazide; procarbazine; PSK TM ; aprotinin; sizoran; spirogermanium; tricine; tricyclic ethylimine quinone; 2,2',2"-trichlorotriethylamine; urethane; vindesine; dacarbazine; mannomustine; dibromomannitol; dibromodulanol; pipobroman; gacytosine; arabinoside ("Ara-C"); cyclophosphamide; thiotepa; taxanes, such as paclitaxel (Taxol TM, Bristol-Myers Squibb Oncology, Princeton, NJ) and docetaxel TM , 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; noantor; teniposide; daunomycin; aminopterin; xeloda; ibandronate; CPT-11; topoisomerase inhibitor RFS 2000; difluoromethylornithine (DMFO); retinoic acid; esperamicins; capecitabine; and pharmaceutically acceptable salts, acids or derivatives of any of the above. This definition also includes antihormonal agents used to modulate or inhibit hormonal effects on tumors, such as: antiestrogens, including, for example, tamoxifen, raloxifene, aromatase inhibiting 4(5)-imidazoles, 4-hydroxytamoxifen, trioxifene, keoxifene, LY117018, onapristone and toremifene (Falutamide); and antiandrogens, such as flutamide, nilutamide, bicalutamide, leuprolide and goserelin; and pharmaceutically acceptable salts, acids or derivatives of any of the above.

[0223] One or more additional therapeutically active components may be administered prior to, simultaneously with, or after administration of the antigen binding molecules of the invention; (for purposes of this disclosure, such administration regimens are considered administration of the antigen binding molecules "in combination" with the additional therapeutically active components).

[0224] The invention encompasses pharmaceutical compositions in which the antigen binding molecules of the invention are co-formulated together with one or more additional therapeutically active ingredients as described elsewhere herein.

[0225] Administration regimen

[0226] According to certain embodiments of the present invention, multiple doses of antigen binding molecules (e.g., anti-CD28 antibodies or bispecific antigen binding molecules that specifically bind PSMA and CD28) can be administered to a subject over a defined time course. The method according to this aspect of the present invention includes sequentially administering multiple doses of the antigen binding molecules of the present invention to a subject. As used herein, "sequential administration" means administering to a subject, each dose of antigen binding molecules at different time points, such as on different dates separated by a predetermined interval (e.g., hours, days, weeks, or months). The present invention includes the following method: it includes sequentially administering a single initial dose of antigen binding molecules to a patient, then administering one or more second doses of antigen binding molecules, and optionally subsequently administering one or more third doses of antigen binding molecules.

[0227] The terms "initial dose", "secondary dose" and "tertiary dose" refer to the time series of the antigen binding molecules of the present invention. Therefore, "initial dose" is the dosage (also referred to as "baseline dose") applied at the beginning of the treatment regimen; "secondary dose" is the dosage applied after the initial dose; and "tertiary dose" is the dosage applied after the second dose. Initial dose, second dose and third dose can all contain the same amount of antigen binding molecules, but may generally be different from each other in terms of frequency of application. However, in certain embodiments, during the course of treatment, the amount of antigen binding molecules contained in initial dose, second dose and / or third dose is different from each other (for example, appropriately adjusted or reduced). In certain embodiments, two or more (for example, 2, 3, 4 or 5) dosages are applied as "loading doses" at the beginning of the treatment regimen, and subsequent doses (for example, "maintenance doses") are then applied at a lower frequency.

[0228] In an exemplary embodiment of the invention, 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, 13 1 / 2, 14, 14 1 / 2, 15, 15 1 / 2, 16, 16 1 / 2, 17, 17 1 / 2, 18, 18 1 / 2, 19, 19 1 / 2, 20, 20 1 / 2, 21, 21 1 / 2, 22, 22 1 / 2, 23, 23 1 / 2, 24, 24 1 / 2, 25, 25 1 / 2, 26, 26 1 Each second dosage and / or the 3rd dosage is administered in 2 weeks (1 / 2 or more). As used herein, phrase "previous dosage" means in a sequence of multiple administrations, the dosage of the antigen binding molecules administered to the patient before the next dosage in the sequence without an intermediate dosage.

[0229] The method according to this aspect of the invention may include administering any number of second doses and / or third 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) to the patient. For example, in certain embodiments, only a single second dose is administered to the patient. In other embodiments, two or more (e.g., 2, 3, 4, 5, 6, 7, 8 or more) second doses are administered to the patient. Similarly, in certain embodiments, only a single third dose is administered to the patient. In other embodiments, two or more (e.g., 2, 3, 4, 5, 6, 7, 8 or more) third doses are administered to the patient.

[0230] In embodiments involving multiple second doses, each second dose can be administered at the same frequency as other second doses. For example, each second dose can be administered to the patient 1 to 2 weeks after the previous dose. Similarly, in embodiments involving multiple third doses, each third dose can be administered at the same frequency as other third doses. For example, each third dose can be administered to the patient 2 to 4 weeks after the previous dose. Alternatively, the frequency of administering the second dose and / or the third dose to the patient can vary during the course of the treatment regimen. The frequency of administration can also be adjusted during the physician's treatment process according to the needs of individual patients after clinical examination.

[0231] Diagnostic Uses of Antibodies

[0232] The bispecific antibodies of the invention can also be used, for example, for diagnostic purposes to detect and / or measure CD28 or PSMA in a sample, or cells expressing CD28 or expressing PSMA. For example, anti-CD28xPSMA antibodies or fragments thereof can be used to diagnose a condition or disease characterized by abnormal expression of CD28 or PSMA (e.g., overexpression, underexpression, lack of expression, etc.). Exemplary diagnostic assays for CD28 or PSMA can include, for example, contacting a sample obtained from a patient with an antibody of the invention, wherein the antibody is labeled with a detectable marker or reporter molecule. Alternatively, an unlabeled antibody can be combined with a secondary antibody that is itself detectably labeled for diagnostic applications. The detectable marker or reporter molecule can be a radioactive isotope, such as 3 H. 14 C. 32 p. 35 S or 125 I; fluorescent or chemiluminescent moieties, such as fluorescein isothiocyanate or rhodamine; or enzymes, 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 CD28 or PSMA diagnostic assays according to the present invention include any tissue or fluid sample that can be obtained from a patient containing a detectable amount of CD28 or PSMA protein or fragments thereof under normal or pathological conditions. Typically, the level of CD28 or PSMA in a specific sample obtained from a healthy patient (e.g., a patient who does not suffer from a disease or condition associated with abnormal CD28 or PSMA levels or activity) is measured to preliminarily establish a baseline or standard level of CD28 or PSMA. This baseline level of CD28 or PSMA can then be compared with the level of CD28 or PSMA measured in a sample obtained from an individual suspected of having a CD28 or PSMA-related disease or condition.

[0233] Examples

[0234] The following examples are presented to provide a complete disclosure and description of how to prepare and use the methods and compositions of the present invention to those of ordinary skill in the art, and are not intended to limit the scope of what the inventors consider to be their invention. Efforts have been made to ensure accuracy with respect to the numerals used (e.g., amounts, temperatures, etc.), but some experimental errors and deviations should be considered. Unless otherwise indicated, parts are parts by weight, molecular weight is average molecular weight, temperature is degrees Celsius, and pressure is atmospheric pressure or near atmospheric pressure.

[0235] summary

[0236] The recent clinical success of tumor-specific antigen (TSA) or tumor-associated antigen (TAA)-CD3 bispecific antibodies (such as anti-PSMAxCD3 bispecific antibodies) is transforming cancer immunotherapy. TAA-CD3 antibodies provide a relatively safe and immediately available therapeutic solution that may prove to be an important new immunotherapy. However, it may not reach the efficacy observed in the CAR-T cell approach, which uses a second co-stimulatory signal that is usually mediated by the CD28 intracellular domain. In fact, CD28 activating antibodies are extremely powerful in human studies, but they are also toxic. The terms TSA and TAA can be used interchangeably in this article.

[0237] This article describes a novel bispecific antibody, termed TAA-CD28, which appears to be safe and well tolerated on its own and produces a greatly amplified and highly targeted anti-tumor response when combined with TAA-CD3. Using TAAs expressed on prostate tumors such as PSMA, it has been found that TAAxCD28 enhances T cell activation and significantly potentiates tumor cell lysis in the presence of tumor targets and TCR / CD3 stimulation.

[0238] It has been demonstrated that the TAAxCD28 bispecific antibodies of the present invention significantly enhance anti-tumor immunity mediated by T cell activation induced by TAAxCD3 in syngeneic tumor models. Pairing CD3 with tumor antigens with limited normal tissue expression, alone or in combination with TAAxCD28, induced little or no peripheral cytokine production in non-human primate and humanized mouse models. In addition, TAAxCD28 alone in cynomolgus monkeys had minimal serum cytokine secretion and T cell activation compared to CD28 superagonist antibodies. Insert content These results suggest that combining this novel co-stimulatory bispecific antibody with the emerging class of TAA-CD3 antibodies could provide a safer off-the-shelf biologic solution that significantly enhances the efficacy of traditional TAA-CD3 bispecific antibodies.

[0239] introduce

[0240] Although monoclonal antibodies have been established as anti-tumor therapeutic agents in the past two decades, their ability to mobilize T cells at the tumor site and effectively exercise their cytotoxic activity is limited. The ability of bispecific antibodies for T cell targeted immunotherapy and subsequent tumor cell killing has been previously demonstrated. In fact, the bispecific antibody platform aims to recruit immune effector cells by combining anti-CD3 with anti-tumor target binding domains. Recently, catumaxomab (EpCAMxCD3) and blinatumomab (CD19xCD3) have obtained regulatory approval for acute lymphoblastic leukemia, and many other bispecific antibodies are under study. In fact, bispecific antibodies that recognize the CD3 components of B cell markers CD20 and T cell repertoires have been constructed, and clinical studies are currently being conducted for hematological malignancies. In fact, TAA-CD3 provides a relatively safe and ready-made treatment solution that does not have to be highly and laboriously customized for individual patients. However, although TAA-CD3 bispecific antibodies may prove to be important new immunotherapies, cross-study comparisons show that they may not be able to achieve the efficacy observed in the CAR-T approach.

[0241] Two chimeric antigen receptor (CAR)-T cell drugs, Kymriiah and Yescarta, have recently received FDA approval for hematological B-cell malignancies, demonstrating the great potential of this approach in the field of personalized cancer immunotherapy. Both products utilize the CD19 antigen as a tumor-associated antigen (TAA), which is an ideal target for T cell-mediated killing due to its restricted expression on B cells, minimizing off-target toxicity and enhancing anti-tumor efficacy. However, the high efficacy of CAR-T cells has been associated with side effects such as cytokine release syndrome (CRS) and neurotoxicity. In addition, all patients have not benefited from treatment, and the number of variables that may affect the clinical outcome of each patient is relatively high in autologous and allogeneic approaches. Further, attempts to target tumor-associated antigens in solid tumors have so far achieved limited success, showing minimal anti-tumor activity or severe side effects. Suppressing the tumor microenvironment, the difficulty of CAR-T cells to enter the entire tumor tissue, and laborious manufacturing are some of the current challenges of promising CAR-T cell therapies.

[0242] One of the limitations of current immunotherapy treatments is the optimal induction of the patient's own immune response to tumor cells through specific tumor cell recognition and cytotoxicity induction. In addition to antigen-specific stimulation through the TCR / CD3 complex (signal 1), costimulatory signals (signal 2) are required for the effective activation of naive T cells and the induction of experienced memory T cell populations. Agonism of the costimulatory pathways CD28 and 4-1BB can provide significant enhancement of target cell lysis, thereby facilitating patient resistance to different immunotherapies. However, the broader role of costimulation in clinical oncology remains to be evaluated. Many tumors lack costimulatory receptors, which prevents them from further enhancing CD3xTAA-induced T activation.

[0243] This paper conducts blind screening of costimulatory pathways by forcibly expressing costimulatory ligands on a group of syngeneic tumors, and the blind screening establishes CD28 as one of the most potent costimulatory receptors together with 4-1BB. Further, this paper constructs a bispecific antibody based on costimulation that bridges CD28 on the surface of T cells, wherein TAA from prostate tissue is used to strengthen anti-tumor response. The data disclosed herein show that in the presence of both TAA and TCR stimulation, the combination therapy of TAA-CD3 and TAA-CD28 bispecific antibodies effectively enhances T cell activation and cytotoxicity, thereby leading to anti-tumor immune enhancement. In fact, the data disclosed herein show that combining this novel bispecific antibody (TAA-CD28) with the emerging TAA-CD3 class antibody can provide a safer off-the-shelf biological product solution that can approach the efficacy of customized CAR-T cell therapy.

[0244] More specifically, exemplary PSMAxCD28 bispecific antibodies were generated and it was demonstrated that exemplary anti-PSMAxCD28 potentiated PSMAxCD3 or CD20xCD3-induced T cell activation in vitro and safely enhanced anti-tumor efficacy in vivo. In vitro activity was demonstrated by images showing bispecific antibodies localized at the immune synapse of T cell and target cell conjugates, PSMAxCD3-induced proliferation enhancement, cytokine release, and cytotoxicity. In vivo anti-tumor efficacy was evaluated in a mouse tumor model (syngeneic). Tumor volume and serum cytokines were monitored over time to show response to bispecific antibody treatment. A study was conducted in cynomolgus monkeys to determine the safety and tolerability of exemplary PSMAxCD28 of the present invention in non-human primates. Toxicity to animals was examined by clinical observation and blood sample collection to analyze serum cytokines and T cell phenotypes.

[0245] As described in detail below, exemplary anti-PSMAxCD28 bispecific antibodies have been generated to enhance TCR / CD3-dependent T cell activation, thereby mimicking the co-stimulation provided by professional APCs (signal 2).

[0246] All procedures were performed in accordance with the NIH Guide for the Care and Use of Laboratory Animals. The protocol was approved by the Institutional Animal Care and Use Committee of Regeneron Pharmaceuticals.

[0247] Example 1. Construction of anti-PSMAxCD28 antibodies

[0248] Generation of anti-CD28 antibodies

[0249] The cells were fused with human CD28 protein fused to the Fc portion of mouse IgG2a, or with cells expressing CD28 or with DNA encoding CD28. Mice (i.e., engineered mice comprising DNA encoding human immunoglobulin heavy chains and kappa light chain variable regions) are immunized to obtain anti-CD28 antibodies. Antibody immune responses are monitored by CD28-specific immunoassays. When the desired immune response is reached, splenocytes are harvested and fused with mouse myeloma cells to maintain their viability and form hybridoma cell lines. Hybridoma cell lines are screened and selected to identify cell lines that produce CD28-specific antibodies. Several anti-CD28 chimeric antibodies (i.e., antibodies with human variable domains and mouse constant domains) have been obtained using this technology. In addition, as described in US2007 / 0280945A1, several fully human anti-CD28 antibodies are isolated directly from antigen-positive B cells without fusion with myeloma cells.

[0250] Certain biological properties of exemplary anti-CD28 antibodies produced according to the methods of this example are described in detail in the Examples set forth below.

[0251] Generation of anti-PSMA antibodies

[0252] Anti-PSMA antibodies were obtained by immunizing genetically modified mice or engineered mice including DNA encoding human immunoglobulin heavy chain and kappa light chain variable regions with human PSMA antigen. Alternatively, mice were immunized with human prostate cancer cells (LNCaP, ATTC, Manassas, VA, USA) expressing human PSMA protein (UniProtKB / Swiss Patent No. Q04609). Antibody immune responses were monitored by PSMA-specific immunoassays. When the desired immune response was achieved, splenocytes were harvested from each mouse and (1) fused with mouse myeloma cells to maintain their viability and form hybridoma cells and screened for PSMA specificity, or (2) B cells were sorted using human PSMA with an N-terminal 6-His tag (R&D, Catalog No. 4234-ZN) as a sorting agent to bind and identify reactive antibodies (antigen-positive B cells) (as described in US2007 / 0280945A1). Initially, chimeric antibodies against PSMA are isolated that have human variable regions and mouse constant regions. The antibodies are characterized and selected for desired properties, including affinity, selectivity, etc. If desired, the mouse constant region is replaced with a desired human constant region (e.g., wild-type or modified IgG1 or IgG4) to generate fully human anti-PSMA antibodies.

[0253] Generation of bispecific antibodies binding CD28 and PSMA

[0254] Bispecific antibodies comprising an anti-PSMA specific binding domain and an anti-CD28 specific binding domain are constructed using standard methods, wherein the anti-PSMA antigen binding domain and the anti-CD28 antigen binding domain each comprise a different unique HCVR paired with a common LCVR. In some cases, the bispecific antibodies are constructed using 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 of the anti-CD28 antibody from an exemplary bispecific antibody (bs16429D) is shown in SEQ ID NO.: 81. The heavy chain amino acid sequence of the anti-PSMA antibody from an exemplary bispecific antibody (bs16429D) is shown in SEQ ID NO.: 82. The common light chain amino acid sequence of bs16429D is shown in SEQ ID NO.: 83.

[0255] The bispecific antibody produced according to this example includes two separate antigen binding domains (i.e., binding arms). The first antigen binding domain includes a heavy chain variable region derived from an anti-CD28 antibody ("CD28-VH"), and the second antigen binding domain includes 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 produces an antigen binding domain that specifically recognizes CD28 on T cells and PSMA on tumor cells.

[0256] Example 2. Heavy and light chain variable region amino acid and nucleic acid sequences

[0257] Table 1 shows the amino acid sequence identifiers of the heavy and light chain variable regions and CDRs of various bispecific antibodies prepared according to Example 1. The corresponding nucleic acid sequence identifiers are listed in Table 2.

[0258] Table 1: Amino Acid Sequence IDs

[0259]

[0260] Table 2: Nucleotide sequence ID

[0261]

[0262]

[0263] Example 3. CD28 and 4-1BB are potent co-stimulatory receptors

[0264] To determine which co-stimulatory receptors are effective in providing co-stimulatory signals important for T cell activation, a blinded screen of co-stimulatory pathways was performed by forced expression of co-stimulatory ligands on a panel of syngeneic tumors (Tables 3 and Figure 2 ) CD28 was again established as one of the most potent co-stimulatory receptors together with 4-1BB. Table 3 summarizes the number of tumor-free mice in the blind screening. Five (5) mice were tested in each group. The assay was performed on three (3) different tumor cell lines, which were engineered to express seven (7) different co-stimulatory ligands. In short, EL4, MC38 and B16F10.9 tumor cells were engineered to express separate co-stimulatory ligands by lentiviral transduction. The cells were implanted subcutaneously in WT C57BL6 mice. Tumor growth was measured on the 18th day, the 24th day and the 25th day after implantation of EL4, MC38 and B16F10.9 tumor types, respectively. This is the time point when the tumor volume from the control group reaches the maximum allowed size (>2000mm3). Tumor size was measured. Consistent with the data in Table 3, CD28 and 4-1BB are also the most effective in reducing tumor size (data not shown).

[0265] Table 3: Tumor growth inhibition in engineered cell lines with introduced co-stimulatory ligand expression

[0266]

[0267] Example 4. Surface Plasmon Resonance-Derived Binding Affinity and Kinetic Constants of Anti-PSMAxCD28 Bispecific Antibodies

[0268] To determine the binding kinetics of the anti-PSMAxCD28 bispecific monoclonal antibodies, surface plasmon resonance-derived binding affinities and kinetic constants of the anti-PSMAxCD28 bispecific and associated parental monoclonal antibodies to PSMA and / or CD28 were determined.

[0269] Binding kinetics of anti-PSMAxCD28 bispecific monoclonal antibody to PSMA

[0270] The equilibrium dissociation constant (K) of 6h.hPSMA (recombinant human PSMA / FOLH1 protein, R&D, catalog number 4234-ZN) binding to purified anti-PSMAxCD28 bispecific monoclonal antibody or anti-PSMA bivalent parental monoclonal antibody was determined using a real-time surface plasmon resonance biosensor using a Biacore T-200 instrument. D The CM5 Biacore sensor surface was derivatized by amine coupling with a monoclonal mouse anti-human Fc antibody to capture purified anti-PSMAxCD28 bispecific antibody or anti-PSMA and anti-CD28 parental monoclonal antibodies.

[0271] This Biacore binding study was 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, 0.05% v / v surfactant P20 (HBS-P++ running buffer). Different concentrations of hPSMA with an N-terminal polyhistidine tag (6h.hPSMA, R&D) ranging from 10 nM to 0.4 nM were prepared in HBS-P++ running buffer and serially diluted 3-fold for anti-PSMAxCD28 bispecific antibodies and anti-PSMA or anti-CD28 parental monoclonal antibodies.

[0272] Different concentrations of 6h.hPSMA were injected over the monoclonal antibody capture surface at a flow rate of 50 μL / min. The association of 6h.hPSMA with the captured monoclonal antibody was monitored for 3 minutes, and the dissociation of 6h.hPSMA in HBS-P++ running buffer was monitored for 10 minutes. The kinetic association (k) was determined by fitting the real-time sensorgrams to a 1:1 binding model using Scrubber 2.0c curve fitting software (BioLogic Software). a ) and dissociation (k d ) rate constant. The binding and dissociation equilibrium constant (K) was calculated based on the following kinetic rate constants: D ) and dissociation half-life (t 1 / 2):

[0273] K D (M) = k d / k a , and t 1 / 2(minutes)=0.693 / k d / 60

[0274] The binding kinetic parameters of hPSMA binding to the purified monoclonal antibody at 25°C for 6 h are shown in Table 4 below.

[0275] Table 4: Biacore binding affinity of monoclonal antibodies to PSMA at 25°C

[0276]

[0277]

[0278] TBD: Not tested

[0279] NB: No binding observed

[0280] The binding kinetic parameters of hPSMA binding to a purified exemplary monoclonal bispecific antibody at 37°C for 6 h are shown in Table 5 below. One (1) RU (response unit) represents one mm 2 1 pg of protein, as defined by the manufacturer.

[0281] Table 5. Biacore binding affinity of monoclonal antibodies to PSMA at 37°C

[0282]

[0283] Binding kinetics of anti-PSMAxCD28 bispecific monoclonal antibody to CD28

[0284] The equilibrium dissociation constant (K) for hCD28.mmh binding to purified monoclonal antibodies was determined using a real-time surface plasmon resonance biosensor using a Biacore T-200 instrument. D The CM5 Biacore sensor surface was derivatized by amine coupling with a monoclonal mouse anti-human Fc antibody to capture purified anti-PSMAxCD28 bispecific mAb and anti-PSMA or anti-CD28 parental mAbs.

[0285] Different concentrations of hCD28.mmh were injected over the mAb capture surface at a flow rate of 50 μl / min. The association of hCD28.mmh with the captured mAb was monitored for 5 min, and the dissociation of hCD28.mmh in HBS-P++ running buffer was monitored for 10 min. The kinetic association (k) was determined by fitting the real-time sensorgrams to a 1:1 binding model using Scrubber 2.0c curve fitting software. a ) and dissociation (k d ) rate constant. The binding and dissociation equilibrium constant (K) was calculated based on the following kinetic rate constants: D ) and dissociation half-life (t 1 / 2):

[0286] K D (M) = k d / k a , and t 1 / 2(minutes)=0.693 / k d / 60

[0287] The binding kinetic parameters for hCD28.mmh binding to purified anti-PSMAxCD28 bispecific monoclonal antibody or anti-CD28 bivalent parental monoclonal antibody at 25°C are shown in Table 6 below.

[0288] Table 6: Biacore binding affinity of monoclonal antibodies to CD28 at 25°C

[0289]

[0290] TBD: Not tested

[0291] NB: No binding observed

[0292] The binding kinetic parameters of hCD28.mmh binding to purified anti-PSMAxCD28 bispecific monoclonal antibody at 37°C are shown in Table 7 below.

[0293] Table 7. Biacore binding affinity of monoclonal antibodies to CD28 at 37°C

[0294]

[0295] As shown in Tables 4 to 7, several anti-CD28 antibodies of the invention bind CD28 with high affinity. Several anti-PSMA antibodies of the invention bind PSMA with high affinity. Several anti-PSMAxCD28 bispecific antibodies bind CD28 and PSMA with high affinity.

[0296] Example 5. Potency and specificity of anti-PSMAxCD28 bispecific monoclonal antibodies for cellular binding of PSMA and CD28

[0297] To evaluate the ability of these antibodies (anti-PSMA, anti-CD28, and anti-PSMAxCD28 monoclonal antibodies) to specifically bind to cell surface proteins, in vitro binding assays were developed using cells expressing human PSMA or CD28 in an electrochemiluminescence-based detection platform (MSD). Two studies were performed. In one study, the potency and specificity of the monoclonal antibodies binding to cell surface antigens were evaluated. These antibodies (anti-PSMA, anti-CD28, and anti-PSMAxCD28 antibodies) showed specific binding to cell lines expressing human PSMA or human CD28. In a complementary experiment, bridging studies with the IgG4s isotype were performed.

[0298] Methods for determining the potency and specificity of cellular binding of anti-PSMAxCD28 bispecific antibodies to PSMA and CD28

[0299] The human epithelial prostate cancer cell line (C4-2 (UroCor)) endogenously expresses human PSMA. The cell line expressing HEK293 / hCD28 was engineered by transducing human embryonic kidney cells (CRL-1573) from ATCC with a neomycin-resistant lentiviral construct encoding human CD28 (hCD28 Accession No. NP_006130.1). To assess the specificity of binding, these two cell lines were evaluated in parallel with the parental HEK293HZ cell line (which is negative for PSMA and CD28) by fluorescence activated cell sorting (FACS). Anti-Feld1 human IgG1 antibody was included as a negative control for IgG detection.

[0300] The above cell lines were treated with Ca-free 2+ / Mg 2+ The cells were washed once with 1x PBS buffer (Irvine Science, catalog number 9240) and incubated with enzyme-free cell dissociation solution (Millipore, accession number S-004-C) at 37°C for 10 minutes to detach the cells from the flask. 2+ / Mg 2+Additional washes were performed with 1x PBS (Irvine Scientific, Cat. No. 9236). TM Cells were counted using an automated T4 cell counter (Nexcelom Bioscience, Model #Automatic T4). Approximately 10,000 cells per well in cell wash buffer were seeded into 96-well carbon electrode plates (Multi-array High Binding Plates, MSD, Meso Scale Discovery, Catalog No. L15XB-3 / LX11XB-3) and incubated at 37°C for 1 hour to allow cell adhesion. Nonspecific binding sites were blocked for 1 hour at room temperature by 2% BSA (w / v) (Sigma, Catalog No. A8577-1L) in PBS.

[0301] Solutions containing anti-PSMA, anti-CD28, anti-PSMAxCD28 antibodies or control antibodies in serial dilutions (ranging from 1.7 pM to 100 nM) or solutions without antibodies were added to the plate-bound cells and incubated for 1 hour at room temperature. Unless otherwise indicated, the analysis was performed in duplicate. The plates were then washed using an AquaMax2000 plate washer with a cell washing head (MDS Analytical Technologies, Model #2000) to remove unbound antibodies. The cells were then washed with SULFO-TAG specific for the heavy and light chains at room temperature. TM Conjugated goat polyclonal anti-human IgG antibody (Jackson ImmunoResearch, catalog number 109-005-088) was used to detect plate-bound antibodies for 1 hour. After washing, the plates were developed with read buffer (MSD, MesoScale Discovery, catalog number R92TD-2) according to the manufacturer's recommended procedures, and luminescent signals were recorded with a SECTOR Imager 600 (Meso Scale Discovery, model #600) instrument.

[0302] Direct binding signal (in relative light units (RLU)) was analyzed as a function of antibody concentration. GraphPad Prism TMSoftware (GraphPad software version #6) was used to fit the data to a sigmoid (four-parameter logistic) dose-response model. EC50 values ​​for binding to HEK293 / hCD28 and C4-2 cells were determined, which were defined as the antibody concentration at which 50% of the maximum binding signal was detected to indicate the potency of each antibody in binding to CD28 or PSMA, respectively. In addition, the ratio of the binding signal of the antibody at 11.1 nM to HEK293 / hCD28 or C4-2 cells to the binding signal of the antibody to HEK293HZ cells was calculated. This representative concentration was selected for high target cell binding but with low background signal on CD28, PSMA-negative HEK293HZ cells. Antibodies with binding ratios less than 3 are marked as NB in ​​Table 8 below. NB is defined as "no specific binding was observed under the assay conditions"

[0303] In a supplemental study, a separate experiment was performed using the same protocol as above, but with the inclusion of an IgG4 isotype control (see USSN 15 / 147,791) and a parental CD28 monoclonal antibody. Test solutions contained anti-PSMA, anti-CD28, anti-PSMAxCD28 bispecific monoclonal antibodies or control antibodies in serial dilutions from 3.4 pM to 200 nM. Binding ratios were calculated at 7.4 nm.

[0304] Results, Summary and Conclusion

[0305] Using an immunobinding assay, the ability of the anti-PSMAxCD28 bispecific antibodies to specifically bind to cells expressing PSMA or CD28 was evaluated compared to cell lines negative for PSMA or CD28 expression. Bivalent parental antibodies specific for PSMA or CD28 were included for comparison. TM The dose-dependent binding of antibodies to cells was detected with anti-human IgG conjugated to 96-well high binding plates (MSD, Meso Scale Discovery, catalog number L15XB-3 / LX11XB-3) at antibody concentrations up to 100 nM, and the binding signal in electrochemiluminescence was recorded on a Sector Imager 600 (MSD). The RLU value of antibody bound to cells was determined. For cells expressing CD28 or PSMA, EC 50The values ​​were calculated as a measure of potency. For samples that produced high background signals, higher concentrations were excluded from the calculation of EC50 values, and the values ​​are indicated with asterisks in Table 8. The binding specificity of the antibodies was assessed by comparing the binding signals of the antibodies to HEK293 / hCD28 or C4-2 cells at 11.1 nM with the binding signals of the antibodies to negative HEK293HZ cells. Specific binding was defined as a 3-fold or greater ratio of antibody binding to cells expressing CD28 or PSMA compared to HEK293HZ cells at this concentration.

[0306] The binding results are summarized in Table 8. At a concentration of 11.1 nM, three exemplary anti-PSMAxCD28 bispecific antibodies of the present invention (bs16429D, bs16430D, and bs16431D) specifically bound to both HEK293-hCD28 and C4-2 cells at ratios 13-31 and 4-10 times higher than those of HEK293HZ cells, respectively. The potency of the bispecific antibodies against HEK293-hCD28 cells ranged from an EC of 5.31 to 9.58 nM. 50 The potency against C4-2 cells was in the range of 0.35-5.24 nM.

[0307] Table 8: Binding of monoclonal antibodies to cell surface antigens

[0308]

[0309]

[0310] NB: non-specific binding, as indicated by a less than 3-fold ratio of cell signal to HEK293HZ at 11.1 nM *: Concentrations above 11.1 nM were removed from the EC due to high background. 50 Exclude from calculation

[0311] #: Samples were run as a single dilution series

[0312] **: Ratio of the cell binding signal (RLU) of 11.1 nM antibody to HEK293-hCD28 relative to HEK293HZ parent

[0313] ***: Ratio of 11.1nM antibody to C4-2 cell binding signal (RLU) relative to HEK293HZ

[0314] The parent antibody specifically binds to cells corresponding to the antigen used to produce the antibody, as indicated in parentheses in the Antibody ID column. 50The values ​​were in the range of 1.04-10.2 nM for HEK293-hCD28 cells and 0.738-0.946 nM for C4-2 cells. For the antibodies, the higher concentrations of mAb11838P2 and mAb11810P2 binding values ​​on C4-2 cells were excluded from the EC 50 The values ​​were excluded from the calculation to compensate for the high background on negative cells. As expected, the IgG control antibody did not specifically bind to cell lines expressing CD28 or PSMA.

[0315] Supplementary experiments were performed as described above, but IgG4 isotype controls and parental CD28 monoclonal antibodies were included. This data was generated as a bridging study to demonstrate that there was a significant background signal using the hIgG4s isotype control. The data are summarized in Table 9. As shown in Table 9, anti-CD28 antibody mAb14226P2 specifically binds to cells expressing human CD28, HEK293-hCD28 P-3. Anti-CD28 antibody mAb14226P2 does not bind to cells that do not express human CD28, regardless of whether the cells express PSMA (C4-2) (HEK293HZ).

[0316] Table 9: Binding of monoclonal antibody isotypes to cell surface antigens

[0317]

[0318] NB: non-specific binding, as indicated by a ratio of less than 3-fold of cell signal to HEK293HZ at 11.1 nM

[0319] *: Ratio of the cell binding signal (RLU) of 7.4 nM antibody to HEK293-hCD28 relative to HEK293HZ parent

[0320] **: Ratio of 7.4 nM antibody to C4-2 cell binding signal (RLU) relative to HEK293HZ

[0321] Binding of PSMAxCD28 bispecific antibody to T cells and target cells

[0322] Experimental Procedure

[0323] Binding of the anti-PSMAxCD28 bispecific antibody to C4-2, 22RV1, RAJI, human, and cynomolgus monkey T cells was determined by flow cytometry analysis and then detected with phycoerythrin (PE)-labeled anti-human IgG antibody. 5Cells / wells were incubated for 30 minutes at 4°C with serial dilutions of PSMAxCD28 bispecific antibodies or human IgG4 antibodies (see USSN 15 / 147,791), which bind to human antigens without cross-reaction to human or cynomolgus CD28, ranging from 133nM to 8.14pM for human and cynomolgus T cells and 133nM to 61pM for PSMA-expressing cells. After incubation, cells were washed twice with cold PBS containing 1% filtered FBS, and PE-conjugated anti-human secondary antibodies (Jackson ImmunoResearch, catalog number 709-116-149) were 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 secondary antibody served as controls.

[0324] Following 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.

[0325] After incubation with human or cynomolgus T cells, cells were washed and stained with a mixture of anti-CD2 (BD, catalog number 562638), anti-CD16 (BD, catalog number 562874), anti-CD4 (BD, catalog number 564305) and anti-CD8 (BD, catalog number 563795) antibodies in bright staining buffer (BD, catalog number 566349) for an additional 20 minutes at 4° C. After washing, cells were resuspended in 200 μL of cold PBS (Gibco, catalog number 14190-144) containing 1% filtered FBS (TCB, catalog number 101), gated in the live / CD2+ / CD4+ / CD16- or live / CD2+ / CD8+ / CD16- gates, and analyzed by flow cytometry on a BD FACS LSR-Fortessa-X20.

[0326] Results, Summary and Conclusion:

[0327] Binding of the PSMAxCD28 bispecific antibody to the surface of human T cells was tested by flow cytometry.

[0328] bs16429D bound to all T cells with an EC50 value of 4.80x10 -8 M. It binds to both CD4+ and CD8+ T cells with EC50 values ​​of 5.09x10 -8 M and 5.89x10 -8 M.

[0329] bs16431D weakly bound to all T cells, with an EC50 value of 1.80x10-7 It weakly binds to CD4+ and CD8+ T cells with EC50 values ​​of 1.67E-07M and 1.80E-07M, respectively.

[0330] Binding of the PSMAxCD28 bispecific antibody to the surface of PSMA-expressing cell lines was tested by flow cytometry.

[0331] C4-2 is a CaP (prostate cancer) cell line derived from LNCaP (derived from lymph node-migrating androgen-sensitive human prostate adenocarcinoma cells; see Wu et al., Int. J. Cancer, 57:406-412 (1994)) cells. Both bs16429D and bs16431D bind to C4-2 cells (see Liu et al., 2004, Prostate, 60:98-108), with EC50 values ​​of 3.87 x 10 -9 M and 1.50x10 -9 M.

[0332] 22RV1 is an epithelial prostate cancer cell line (see In Vitro Cell Dev. Biol. Anim., 1999, 35(7):403-409). Both bs16429D and bs16431D bind to 22RV1 cells, with EC50 values ​​of 3.05x10 -9 M and 6.33x10 -09 M.

[0333] These results are summarized in Tables 10-12.

[0334] Table 10: Binding of anti-PSMAxCD28 bispecific antibodies to human T cells

[0335]

[0336] Table 11: Binding of anti-PSMAxCD28 bispecific antibodies to cynomolgus monkey T cells

[0337]

[0338]

[0339] Table 12: Binding of anti-PSMAxCD28 bispecific antibodies to PSMA-expressing cells

[0340] Antibody ID C4-2 cell FACS[M] 22RV1 cell FACS[M] Raji cell FACS[M] bs16429D 3.87E-09 3.05E-09 No combination bs16431D 1.50E-09 6.33E-09 No combination Isotype Control No combination No combination No combination

[0341] Example 6: Primary and Engineered Bioassays of PSMAxCD28 Bispecific Antibodies

[0342] T cell activation is achieved by stimulating the T cell receptor (TCR), which recognizes specific peptides presented by class I or class II major histocompatibility complex (MHCI or MHCII) proteins on antigen presenting cells (APCs) (Goldrath et al., "Selecting and maintaining a diverse T-cell repertoire", Nature 402, 255-262 (1999)). The activated TCR in turn initiates a series of signal transduction events that can be monitored by reporter genes and driven by various transcription factors such as activating protein 1 (AP-1), nuclear factor of activated T cells (NFAT), or nuclear factor kappa-light-chain enhancer of activated B cells (NFκB). The T cell response is then further refined by engagement of co-receptors expressed 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., "B7-CD28 Superfamily", Nat. Rev. Immunol., 2(2): 116-26 (2002). The co-stimulatory molecule CD28 is activated by endogenous ligands CD80 or CD86 expressed on APCs. CD28 enhances cell signaling after TCR activation, such as pathways controlled by the NFκB transcription factor. CD28 costimulation is important for efficient T cell activation, such as 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,” Proceedings of the National Academy of Sciences of the United States of America (PNAS), 97(7):3394-3399 (2012)).

[0343] To identify antibodies that enhance T cell activity in the presence of primary stimulation and PSMA target expression, anti-CD28 antibodies and anti-PSMAxCD28 bispecific antibodies were characterized in an engineered reporter gene bioassay using human primary T cells and a cell-based assay. The assay evaluates the behavior of the anti-PSMA / CD28 bispecific antibodies in the presence and absence of primary stimulation and in the presence and absence of target expression. A schematic diagram of the assay is shown in Figure 3The assay was performed to select anti-PSMAxCD28 bispecific antibodies that enhance T cell activity in the presence of primary stimulation and target expression. Thus, the assay evaluates the behavior of bispecific antibodies in the presence and absence of primary stimulation and target expression.

[0344] 1) Luciferase-based reporter gene assay:

[0345] a) Engineering of reporter gene T cells:

[0346] Jurkat derived T cell clone JRT3.T3.5 (ATCC, #TIB-153) was transduced with NFKB luciferase reporter construct (NFKB-Luc, SA Biosciences / Qiagen, Cat. No. CLS-013L). After isolation of a puromycin-resistant clone (JRT3.T3.5 / NFκB-Luc clone 1C2), the cells were further engineered to express full-length human TCRα (1G4A - amino acids M1 to S274) and TCRβ subunits (1G4B - amino acids M1 to G311) (Robbins et al., "Single and Dual Amino Acid Substitutions in TCR CDRs Can Enhance Antigen-Specific T Cell Functions", J. Immunol. 180(9):6116-31 (2008)). After isolating a single clone (J.RT3-T3.5 / NFκB-Luc / 1G4AB clone 1D2), the cells were further engineered to express full-length human CD8α (hCD8a - amino acids M1 to V235 of accession number NP_001139345) and human CD8β subunit (hCD8b - amino acids M1 to K210 of accession number P10966). κB-Luc / 1G4AB / hCD8ab clone 1D5), and further transduced with full-length human CD28 (hCD28-amino acid M1 to S220 accession number P10747). Cells were sorted for high CD28 expression and maintained in RPMI+20% FBS+penicillin / streptomycin / glutamine (P / S / G)+NEAA+NaPyr+1μg / mL puromycin+500μg / mL G418+250μg / mL hygromycin+10μg / mL blasticidin. For faster growth, engineered reporter T cells were stored in cell culture medium without antibiotics and used as engineered reporter T cells for cell-based luciferase experiments. The 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.

[0347] b) Engineering of APC:

[0348] 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), and the highly sorted cell line HEK293 / CD20 / hPSMA was maintained in DMEM+10%+P / S / G+NEAA supplemented with 500 μg / mL G418.

[0349] c) T cell / APC stimulation:

[0350] In this experiment, engineered reporter T cells were stimulated by two bispecific antibodies. The first stimulation was delivered by T cell activation bispecific antibody anti-CD3xCD20 hIgG4 (see WO14 / 047231, US9657102 and USSN 14 / 661,334), which targets CD3 molecules on engineered reporter T cells and CD20 on HEK293 cells. Here, the first stimulation bypasses the need to activate TCR by its natural ligand, which is a specific peptide displayed on MHC molecules. The second stimulation is driven by CD28 bispecific antibody. This antibody mimics CD28 activation on T cells by its ligand CD80 / CD86 expressed on APC. Here, antibodies interact with CD28 on T cells and PSMA on engineered HEK293 cells, and drive the activation of CD28 on engineered reporter T cells. The simultaneous activation of TCR and CD28 causes the transcriptional activity of NFκB to increase, which in turn induces the generation of reporter luciferase.

[0351] d) Luciferase assay setup:

[0352] RPMI1640 supplemented with 10% FBS and P / S / G was used as assay medium for preparation of cell suspensions and antibody dilutions to screen anti-PSMA x CD28 bispecific antibodies.

[0353] One day before screening, culture engineered reporter T cells to 1x10 6 cells / mL. Three-fold (1:3) serial dilutions of anti-PSMA x CD28 bispecific antibodies and controls were tested in the presence of a constant 50pM anti-CD20 x CD3 or hIgG4 isotype control. The 10-point dilutions ranged from 15pM to 100nM, with the final dilution containing no anti-PSMA x CD28 antibody. Reagents were added in the following order: 1) serial dilutions of antibody were added to corresponding wells of a 96-well white flat-bottom plate; 2) a fixed concentration of 50pM anti-CD20 x CD3 or hIgG4 isotype control was added to each well; 3) the cells were resuspended to 4x10 5 APCs were added to the plate at a final concentration of 1x10 cells / mL. 4 cells / well; 4) The reporter gene T cells cultured overnight were cultured at 2x10 6 / mL and added to the plate at a final concentration of 5x10 4 cells / well. Add 100 μL ONE-Glo TMThe plates were incubated at 37°C / 5% CO2 for 4-6 hours before adding luciferase (Promega, catalog number E6051) reagent to lyse the cells and detect luciferase activity. The emitted light was captured in relative light units (RLU) on a multi-label plate reader Envision (PerkinElmer, model 2104). All serial dilutions were tested in duplicate.

[0354] By using GraphPad Prism TM The software fits the data to a four-parameter logistic equation on a 10-point dose-response curve to determine the EC of the antibody. 50 Fold induction was calculated using the following equation:

[0355]

[0356] 2) Use primary human CD4 + T cell IL-2 functional assay:

[0357] Developed primary CD4 + T cell / APC functional assay to evaluate the impact of CD28 activation on IL-2 production when engaged with an anti-PSMA x CD28 bispecific antibody.

[0358] a) Isolation of primary human CD4+ T cells:

[0359] Human peripheral blood mononuclear cells (PBMCs) were isolated from healthy donor leukocytes using 50 mL of SepMate TM PBMC separation was completed by density gradient centrifugation in a 50 mL SepMate tube (StemCell Technologies, catalog number 85450). Briefly, 15 mL FicollPaque PLUS was layered into a 50 mL SepMate tube, and then 30 mL of leukocytes diluted 1:2 with PBS (Dulbecco's phosphate buffered saline solution, Irvine Scientific, catalog number 9240) were added. Subsequent steps were performed according to the SepMate manufacturer's protocol. CD4 was then isolated from PBMC using a human CD4 microbead kit (catalog number 130-045-101) from Miltenyi Biotec, according to the manufacturer's instructions. + T cells. The isolated CD4 + T cells were frozen at a concentration of 5 × 10 per vial. 6 The cells were cultured in FBS containing 10% DMSO (Macron Fine Chemicals, Cat. No. 4948-02).

[0360] b) from primary CD4 + IL-2 released from T cells:

[0361] In this assay, an anti-CD20 x CD3 bispecific antibody was used in combination with HEK293 cells engineered to express human CD20 to activate primary CD4 T cells via cross-linking of CD3 on their surface. + T cells. Binding of the CD20 arm of the anti-CD20 x CD3 bispecific antibody to CD20-expressing HEK293 cells drives clustering of the CD3 receptor, providing the first signal important for T cell stimulation. However, in order to detect quantifiable IL-2 release, co-stimulation provided by cross-linking the CD28 molecule is important. Here, the bispecific anti-PSMA x CD28 antibody was coupled to CD4 + CD28 on T cells interacts with PSMA on engineered HEK293 / hCD20 cells and drives activation of aggregated-CD28. The combined engagement of TCR and CD28 results in increased IL-2 production released into the cell culture medium. IL-2 is detected and quantified from cell supernatants using a homogenous, no-rinse AlphaLisa kit (PerkinElmer, catalog number AL221).

[0362] On the day of the assay, human CD4 T cells previously isolated and frozen from a donor were thawed in stimulation medium (X-VIVO 15 cell culture medium (Lonza, catalog number 04-418Q)) supplemented with 10% FBS, HEPES, NaPyr, NEAA, and 0.01 mM BME (β-mercaptoethanol, Sigma-Aldrich, catalog number M-7522) containing 50 U / ml benzonase nuclease (EMD Millipore, catalog number 71205-3). + T cells. The cells were centrifuged at 1200 rpm for 10 minutes, resuspended in stimulation medium, and cultured at 1x10 5 HEK293 cells engineered to express human CD20 alone or in combination with human PSMA were plated at a concentration of 10 × 10 cells / well in a 96-well round bottom plate. HEK293 cells engineered to express human CD20 alone or in combination with human PSMA were treated with 15 μg / mL mitomycin C (Sigma-Aldrich, catalog number M4287) in primary stimulation medium. 6 After incubation at 37°C, 5% CO2 for 1 hour, HEK293 cells were washed three times with D-PBS containing 2% FBS and 1×10 cells / mL were added to each well. 4 The final concentration of cells was added to the CD4 + T cells in the wells.

[0363] Subsequently, a 1:3 serial dilution of anti-PSMA x CD28 bispecific or control antibody ranging from 15 pM to 100 nM was added to the wells in the presence of 50 pM anti-CD20 x CD3 or hIgG4 isotype control. The final point of the 10-point dilution did not contain anti-PSMA x CD28 or anti-CD28 antibody. After the plate was incubated at 37°C, 5% CO2 for 72 hours, it was centrifuged to pellet the cells and 40 μL of the culture supernatant was collected. From this, 5 μL was tested in the human IL-2 AlphaLISA assay according to the manufacturer's protocol. Measurements were obtained on a multi-label plate reader Envision (Perkin Elmer, Model 2104). The IL-2 concentration produced in the assay wells was determined using a standard curve of known IL-2 concentrations. All serial dilutions were tested in duplicate.

[0364] By using GraphPad Prism TM The software fits the data to a four-parameter logistic equation on a 10-point dose-response curve to determine the EC of the antibody. 50 Fold induction was calculated using the following equation:

[0365]

[0366] Results, Summary and Conclusion

[0367] like Figure 4A and 4B As shown, hPSMA x hCD28 enhanced activation of CD4+ T cells (as measured by IL-2 release) and engineered JRT3.T3 / 1G4 / hCD28 cells (as measured by luciferase activity) in the presence of primary stimulation (REGN2281 anti-CD20xCD3) and PSMA expressed on HEK293 / hCD20 cells. The CD28 bivalent antibody (mAb14193P2) slightly enhanced T cell activity in the presence of primary stimulation, but slightly enhanced activity in the engineered bioassay 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 x CD3 stimulation, although to a lesser extent than the PSMA x CD28 bispecific antibody in the presence of target PSMA and primary stimulation.

[0368] 1) Luciferase-based reporter gene assay:

[0369] Tables 13 and 14 summarize the EC values ​​of engineered reporter T cells co-incubated with HEK293 / hCD20 or HEK293 / hCD20 / hPSMA cells in addition to 50 pM constant hIgG4 isotype control or anti-CD3 x CD20 bispecific antibody (T cell stimulating bispecific antibody). 50 and fold induction values.

[0370] Table 13: Luciferase activity in engineered reporter T cells in the absence of TCR stimulation

[0371]

[0372] Table 14 summarizes the EC values ​​for luciferase activity in engineered T cells co-incubated with HEK293 / hCD20 or HEK293 / hCD20 / hPSMA cells and a 50 pM constant hIgG4 isotype control. 50 Values ​​and multiples induce results.

[0373] Table 14: Luciferase activity in engineered reporter T cells in the presence of TCR stimulation

[0374]

[0375]

[0376] Table 14 summarizes the EC values ​​for luciferase activity in engineered T cells co-incubated with HEK293 / hCD20 or HEK293 / hCD20 / hPSMA cells and 50 pM constant anti-CD3xCD20. 50 Values ​​and multiples induce results.

[0377] When T cells and APCs were treated with 50 pM hIgG4 isotype control, no CD28 bispecific antibody showed an increase in luciferase activity in the absence of TCR stimulation, regardless of the APC line used in the assay. Slight luciferase activation was observed in one of the parental CD28 antibodies (mAb14226P2) on HEK293 / hCD20 cells (4.76x) and HEK293 / hCD20 / hSPMA cells (3.59x) as shown in Table 14.

[0378] In contrast, if cells were treated with 50 pM anti-CD3 x CD20 bispecific antibodies, 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. Very low or no activation was observed in the single-arm controls (one arm of mAb14226P2, mAb14193P2, and mAb14216P2), regardless of the APC line. As shown in Table 14, slight luciferase activation was observed in all three parental CD28 antibodies (mAb14226P2, mAb14193P2, and mAb14216P2).

[0379] 2) Use primary human CD4 + T cell IL-2 functional assay:

[0380] Measuring IL-2 cytokine production in functional primary CD4 + The ability of the anti-PSMA x CD28 bispecific antibody to provide co-stimulation via CD28 on T cells in the absence or presence of PSMA target expression was evaluated in a T cell assay.

[0381] Table 15 summarizes the CD4 + T cell EC 50 and fold induction values.

[0382] As expected, no measurable IL-2 release was observed in wells containing a constant amount of hIgG4 isotype control, since primary T cell stimulation was absent.

[0383] In contrast, measurable levels of IL-2 were detected in samples treated with the anti-CD3 x CD20 bispecific antibody. + When T cells were co-incubated with HEK293 / hCD20 cells, all tested CD28 monoclonal antibodies (including anti-CD28 antibodies and anti-PSMA x CD28 bispecific antibodies (except bs16430D and bs16431D)) showed increased IL-2 levels (Table 15). The parent antibody mAb14226P2 showed the highest induction fold and an EC of around 6 nM. 50 When CD4 +When T cells were co-cultured with hPSMA-expressing APCs and anti-CD3 x CD20 bispecific antibodies, IL-2 release was detected in all three anti-PSMAxCD28 bispecific antibodies (bs16429D, bs16430D, and bs16431D). Lower IL-2 levels were measured with their single-armed control antibodies and parental antibodies in the same setting as shown in Table 15.

[0384] Table 15: TCR stimulation from primary CD4 + T cells produce IL-2

[0385]

[0386] Table 15 summarizes the primary CD4 + T cells and ECs producing IL-2 with 50 pm constant anti-CD3xCD20 antibodies 50 Values ​​and multiples induce results.

[0387] Example 7: Anti-PSMAxCD28 Bispecific Antibody Enhances T Cell Activation in the Presence of PSMA and TCR Stimulation with Anti-CD20xCD3

[0388] Exemplary anti-PSMAxCD28 bispecific antibodies have been generated and biacore validated (see, e.g., Examples 1 and 4). A pair of bispecific antibodies (CD20xCD3 and an exemplary anti-PSMAxCD28 bispecific antibody of the invention) was used to induce the aggregation of T cells and co-stimulatory receptors on tumor cells by binding to PSMA (CD20 and prostate-specific membrane tumor antigen-PSMA). Figure 5A To demonstrate that PSMAxCD28 binds and activates CD28 in the presence of PSMA expressed on target cells and TCR activation, a series of cell-based in vitro assays were performed.

[0389] Localization of CD28

[0390] Cell lines

[0391] 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, a mammalian vector encoding PSMA driven by the ubiquitin C promoter (amino acids M1 to A750 of accession number Q04609) and a neomycin resistance gene was used for stable transfection. Similarly, a mammalian vector encoding hCD20 driven by the ubiquitin promoter (amino acids M1 to P297 of accession number NP_068769.2) was used to generate the HEK293 / hCD20 cell line. Transfected cells were cultured in 500 μg / ml of Geneticin A to select for stably expressing cell lines.

[0392] To generate HEK293 / hCD20 / 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 HEK293 / hCD20 cells. Human PSMA-positive cells were isolated by fluorescence activated cell sorting (FACS). All generated cell lines were maintained in DMEM+10% FBS+P / S / G+NEAA supplemented with 500 μg / mL G418. Jurkat clone E6-1 (ATCC, #TIB-152) was cultured according to the ATCC recommended protocol.

[0393] 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 / hPMA was maintained in DMEM+10% FBS+P / S / G+NEAA supplemented with 500 μg / mL G418.

[0394] Amnis Image Stream

[0395] Jurkat T cells and target cells (HEK293 / hPSMA or HEK293 / hPSMA / hCD20) were incubated with CD20xCD3-Alexa488 (REGN2280, 0.5 μg / ml) alone or with PSMAxCD28-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., and then stained with streptavidin-PE-CF594 (BD 562284, 1 μg / ml) and Hoechst 33342 (Thermo Fisher H3570, 1 μM) for 15 minutes at 4° C. The cells were washed with PBS and stored in BD stabilizing fixative (BD 338036). Cell images were collected on an imaging flow cytometer and analyzed using The software was used for analysis. Cells were gated on double-peak bright field, double-peak nucleus, nuclear focus, single point count, and single-peak CD28. The synaptic region was defined by a peak-valley mask based on nuclear staining. Cells with erroneous peak-valley masks were eliminated by gating on the overlapping area between the peak-valley mask and CD28. The ratio of CD28 in / out of the synapse was calculated by the following formula:

[0396] CD28 in / out of synapse = intensity of CD28 in synapse / (total intensity of CD28 - intensity of CD28 in synapse)*100%.

[0397] Results, Summary and Conclusion

[0398] T cells were co-cultured with engineered target cells (HEK293 / hPSMA or HEK293 / hCD20 / hPSMA) overexpressing CD20 and PSMA, and fluorescently labeled bispecific antibodies (green for CD20xCD3 and red for PSMAxCD28). To determine the localization of CD28, cells were fixed and stained with anti-CD28 after incubation for 1 hour at 37°C. Images of T cell and target cell conjugates were obtained using Amnis Image Stream imaging flow cytometry. In the absence of PSMA expression on target cells, the CD20xCD3 bispecific antibody alone induced little or no CD28 aggregation on T cells. When PSMA was expressed on target cells, CD20xCD3 localized to the interface of the T cell and target cell conjugates, and an immune synapse (IS) was formed where CD28 was localized. PSMAxCD28, together with CD20xCD3, further enhanced the accumulation of CD28 at the IS. The distribution of CD28 was quantified by calculating the ratio of CD28 staining inside the IS to CD28 staining outside the IS ( Figure 5B ).

[0399] It has been concluded that PSMAxCD28 drives robust CD28 accumulation at the IS, where T cell activation signaling occurs, in the presence of PSMAxCD3 and PSMA-expressing target cells.

[0400] Cytokine release

[0401] To further investigate the effect of TAAxCD28 on T cell activation, T cell proliferation and the release of IL-2 and IFNγ cytokines were measured after incubation with dose-titrated CD20xCD3 or PSMAxCD28 bispecific antibodies co-cultured with primary human T cells and engineered target cells overexpressing different TAAs (HEK293 / hPSMA, HEK293 / hCD20, or HEK293 / hCD20 / hPSMA). Figure 5C-5G It has been demonstrated that the exemplary anti-PSMAxCD28 antibodies and anti-PSMAxCD3 bispecific antibodies of the invention do not compete for binding to PSMA-expressing cells and therefore bind to different epitopes ( Figure 5H ). It was further confirmed that PSMAxCD28 induced T cell activation in the presence of both CD3 stimulation and PSMA expression on target cells (HEK293 / hCD20 / hPSMA plus 5pM CD20xCD3). In the absence of PSMA expression on target cells (co-cultured with HEK293 / hCD20) or CD3 stimulation (co-cultured with HEK293 / PSMA plus 5pM CD20xCD3 or HEK293 / hCD20 / hPSMA plus 5pM isotype control), no activation of T cells by the exemplary anti-PSMAxCD28 bispecific antibody was observed. ( Figure 5E-5G ). In summary, it was demonstrated that exemplary anti-PSMAxCD28 antibodies of the invention drive T cell activation in the presence of PSMAxCD3 and PSMA-expressing target cells, resulting in increased proliferation and cytokine secretion.

[0402] Example 8: Killing of PSMA-Expressing Cells

[0403] Two FACS-based cytotoxicity studies were performed. In the first study, FACS-based cytotoxicity was performed on C4-2 cells in the presence of human peripheral blood mononuclear cells (PBMCs) in the presence or absence of anti-PSMAxCD28 stimulation. Stimulation was performed in the presence of a fixed concentration of anti-PSMAxCD28 bispecific antibody and serial dilutions of anti-PSMAxCD3 bispecific antibody. The second study was otherwise identical to the first study except that cynomolgus monkey PBMCs were used instead of human PBMCs. It was concluded that the PSMAxCD28 bispecific antibody enhanced T cell activation and cytotoxicity against prostate tumor cells in the presence of TCR stimulation of PSMAxCD3.

[0404] Experimental methods

[0405] Human primary CD4 + T cell isolation

[0406] Human peripheral blood mononuclear cells (PBMCs) were isolated from healthy donor leukocytes using 50 mL of SepMate TM PBMCs were isolated by density gradient centrifugation. TM Human CD4 + T cell isolation kit and follow the manufacturer's instructions to isolate CD4 + T cells. The isolated CD4 + T cells were frozen in FBS containing 10% DMSO at a concentration of 50 × 10 per vial. 6 cells.

[0407] Primary human T cell activation assay

[0408] On the day of the assay, previously isolated and frozen human CD4 + T cells. The cells were centrifuged at 1200 rpm for 10 minutes, resuspended in stimulation medium, and plated at 1x10 5 HEK293 cells (HEK293 / hPSMA, HEK293 / hCD20, or HEK293 / hPSMA / hCD20) were plated in 96-well round-bottom plates at a concentration of 10 × 10 cells / mL in primary stimulation medium. 6After incubation at 37°C, 5% CO2 for 1 hour, HEK293 cells were washed three times with D-PBS containing 2% FBS and plated at 2×10 cells / mL per well. 4 The final concentration of cells was added to the CD4 + T cells in the wells.

[0409] To determine the suboptimal concentration of CD20xCD3 bispecific antibody for activation of T cells, CD20xCD3 was serially diluted 1:3 from 1.5pM to 10nM in the presence of 500pM PSMAxCD28 bispecific or hIgG4s isotype control. A constant 5pM CD20xCD3 bispecific or hIgG4 control was selected, and the following antibodies were titrated from 15pM to 100nM at a 1:3 dilution: 1) PSMA x CD28, 2) non-target x CD28 control, 3) CD28 parent, 4) CD28SA (superagonist), 5) hIgG4 isotype control, and 6) hIgG4s isotype control. The final point of the 10-point dilution contained no titrated antibody, only 5pM CD20xCD3 bispecific or hIgG4 control.

[0410] After the plate was incubated at 37°C, 5% CO2 for 48 hours, it was centrifuged to precipitate cells and 50 μL of culture supernatant was collected. Thus, 5 μL was tested in human IL-2 and human IFNγAlphaLISA assays according to the manufacturer's protocol. These measurements were obtained on PerkinElmer's multi-label plate reader Envision. A standard curve of known IL-2 or IFNγ concentrations was generated to infer the pg / mL of IL-2 or IFNγ produced in the assay wells. All serial dilutions were tested in duplicate. The precipitated cells were incubated with [methyl-3H]-thymidine at 0.25uCi / well at 37°C, 5% CO2 for 16 hours. Cells were collected on PerkinElmer Unifilter plates using PerkinElmer's Unifilter96 cell harvester. After adding 30 μl of scintillation fluid, the plate was sealed and counted per minute for each well obtained using PerkinElmer's TopCount NXT.

[0411] By using GraphPad Prism TM The software determines the EC of the antibody based on a four-parameter logistic equation on a 10-point dose-response curve. 50 value.

[0412] FACS-based cytotoxicity assay

[0413] 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, catalog number 34557). After labeling, cells were plated overnight at 37°C. Separately, human PBMCs (New York Blood Center) or cynomolgus macaque PBMCs (Covance, Cranford, NJ) were plated at 1×10 6 Cells / mL were plated in supplemented RPMI medium and incubated overnight at 37°C to enrich lymphocytes by depleting adherent macrophages, dendritic cells, and some monocytes. The next day, target cells were incubated with primary PBMCs depleted of adherent cells (effector cell / target cell ratio of 4:1) and serial dilutions of anti-PSMAxCD3 bispecific antibody or IgG4 control alone (concentration range: 0.42nM to 0.1pM) or with a fixed concentration of PSMAxCD28 costimulatory molecules bs16429D or bs16431D at 2.5μg / ml (16.7nM) for 96 hours at 37°C.

[0414] After incubation, cells were removed from the cell culture plates in trypsin-EDTA (Millipore, accession number SM-2004-C) dissociation buffer and analyzed by FACS on a FACS BD LSRFortessa-X20 (BD).

[0415] For FACS analysis, cells were stained with Violet Far Red Cell Tracker (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 Violet Cell Tracker positive populations. The percentage 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, and R2 = Percentage of live target cells (%) in the absence of test antibody. CD2 + / CD4 + or CD2 + / CD8 + Activated T cells (CD25 + )T cells were used to measure T cell activation. The viable CD4 + or CD8 + The number of cells is used to measure T cell counts.

[0416] The levels of cytokines accumulated in the culture medium were analyzed using the BD Flow Cytometry Bead Array (CBA) Human Th1 / Th2 / Th17 Cytokine Kit according to the manufacturer's protocol.

[0417] For FACS analysis, cells were stained with Dead / Live Near IR Responder (Invitrogen, Cat. No. L34976) dye. Immediately prior to FACS analysis, 500,000 (5x10 5 ) counting beads. One hundred thousand (1x10 5 ) beads. To assess the specificity of killing, cells were gated on the live Violet-labeled population. The percentage of live population was recorded and used to calculate viability.

[0418] 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 to total T cells (CD2+). The information of directly conjugated antibodies is as follows: CD2, PE:CD2 (Cl:RPA-2.1), BD, catalog number 555327; CD4, PerCP-Cy5.5:CD4 (Cl:OKT-4), Biolegend, catalog number 317428; CD8, APC:CD8 (Cl:RPA-T8), Biolegend, catalog number 301049; CD25, BV510:CD25 (Cl:M-A251), BD, catalog number 563352; and PD-1, PE-Cy7:PD1 (Cl:EH12.2H7), Biolegend, catalog number 329918.

[0419] Supernatants from assay wells of the human PBMC assay were assessed for Th1 / Th2 cytokine release using the BD Flow Cytometry Bead Array Human Kit (BD, Cat. No. 560484) and following the manufacturer's protocol.

[0420] Results, Summary and Conclusion:

[0421] Fig. 6A and 6B It is shown that anti-PSMAxCD28 enhances the cytotoxic potency of bs13644D (anti-PSMAxCD3) in the presence of human or cynomolgus monkey T cells. The ability of the anti-PSMAxCD3 bispecific antibody to induce killing of human PSMA-expressing target cells by naive human T cells was tested as a single agent or in the presence of a co-stimulatory anti-PSMAxCD28 bispecific antibody. Fig. 6A and 6BIn the studies shown, anti-PSMAxCD28 antibodies were fixed at 2.5 μg / ml. Anti-PSMAxCD3 was serially diluted 1:4. Antibodies were incubated with cells for 96 hours. Table 16 summarizes the composition of antibodies used in this study.

[0422] Table 16: Components of bispecific antibodies in cytotoxic potency studies

[0423]

[0424] We tested whether an 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, an exemplary anti-PSMAxCD28 bispecific antibody was tested in combination with PSMAxCD3 (Figures 6, 7A-7D). T cells containing human PBMCs were co-cultured with C4-2 prostate cancer cells that endogenously express high levels of PSMA (data not shown). PSMAxCD28 significantly increased the potency of cytotoxicity induced by PSMAxCD3 alone, increasing the EC50 from 4.3x10 -11 Shift to 1.5x10 -12 (more than a log shift in potency) (Figures 6 and 7A). Consistent with the induction of T cell toxicity, PSMAxCD28 increased the level of IFNγ release induced by PSMAxCD3 by 4-fold ( Figure 7B Similarly, the combination of PSMAxCD28 and PSMAxCD3 increased the number of CD4 and CD8 T cells and the expression of the activation marker CD25 ( Figure 7C and 7D ). When the exemplary anti-PSMAxCD28 bispecific antibody was combined with a non-targeted CD3 bispecific antibody, no effect on T cell cytotoxicity or activation was observed. In addition, to show that the exemplary anti-PSMAxCD28 bispecific antibody can enhance PSMAxCD3 cytotoxicity and activation of T cells from non-human primates, the same assay was performed using PBMCs from cynomolgus monkeys, and similar results were obtained (Figures 6, 7E-7G). These results indicate that the exemplary anti-PSMAxCD28 bispecific antibody can effectively enhance PSMAxCD3-mediated T cell activation not only by proliferation and cytokine release, but also by cytotoxicity. Figure 7H Exemplary anti-PSMAxCD28 bispecific antibodies are shown to bind to cellular targets as measured by flow cytometry.

[0425] The anti-PSMA x CD3 bispecific antibody activated and directed human T cells to deplete C4-2 cells (a human prostate cancer cell line). Target cell killing was observed in the presence of the anti-PSMA x CD3 bispecific antibody, and the EC 50 Killing of C4-2 cells occurred in a dose-dependent manner (Table 17). The observed target cell lysis was associated with upregulation of CD25+ and PD-1+ cells on CD2+ T cells, also in agreement with the picomolar EC 50 The anti-PSMAxCD3 bispecific antibodies induced the release of human cytokines. In the presence of a fixed concentration of anti-PSMAxCD28 co-stimulatory molecules (bispecific antibodies), the cytotoxic activity observed with the anti-PSMAxCD3 bispecific antibodies as single agents was enhanced (Table 17).

[0426] In summary, co-stimulation with the anti-PSMAxCD28 bispecific antibody increased T cell activation, PD-1 upregulation, and cytokine release compared to that observed with the anti-PSMAxCD3 bispecific antibody as a single agent. Table 16 summarizes the results of experiments using human PBMCs.

[0427] Table 17: Effect of anti-PSMAxCD28 and anti-PSMAxCD3 on cytotoxicity of C4-2 cells in the presence of human PBMCs

[0428]

[0429] The ability of the anti-PSMAxCD3 bispecific antibodies to induce killing of human PSMA-expressing target cells by naive cynomolgus monkey T cells as a single agent or in the presence of a co-stimulatory anti-PSMAxCD28 bispecific antibody was also tested. In the selected antibody titrations, the anti-PSMAxCD3 bispecific antibodies activated human T cells but did not induce T cell depletion of C4-2 cells (Table 18). Co-stimulation with the anti-PSMAxCD28 antibody resulted in increased T cell activation, enhanced cytotoxic activity, and upregulation of the PD-1 marker on T cells (Table 18).

[0430] Table 18: Effect of anti-PSMAxCD28 against PSMAxCD3 cytotoxicity on C4-2 cells in the presence of cynomolgus monkey PBMCs

[0431]

[0432] Example 9: Cytokine release from cells

[0433] As described elsewhere herein, in vitro cytokine release assays using a "soluble" format of human PBMCs were unpredictable of cytokine release. Therefore, a "coated" assay format was developed.

[0434] The ability of the anti-PSMA / CD28 bispecific antibodies and anti-CD28 antibodies of the invention to induce proliferation and cytokine release in human PBMCs was evaluated using a coated assay format. For the PBMC proliferation assay, 1×10 5 Use freshly isolated PBMCs / well. For co-culture, use 1x10 4 C4-2 cells treated with mitomycin C were added to each well. Antibody dilutions were wet or dry coated on the plates overnight. Plates were washed before adding PBMCs. Supernatants were collected at approximately 54 hours for cytokine analysis by Meso Scale Diagnostics (MSD, Rockville, MD). 3 H thymidine for 18 hours and proliferation was measured.

[0435] The results showed that soluble anti-PSMA x CD28 bispecific antibody induced proliferation of human PBMCs in the presence of CD3 stimulation. In contrast, anti-CD28 superagonist induced proliferation of human PBMCs in the absence of CD3 stimulation (data not shown).

[0436] Furthermore, wet-coated bs16429D (PSMA x CD28 "A") and mAb14226P2 (CD28 parental "A") induced less cytokine release than wet-coated anti-CD28 superagonists, whereas bs16431D (PSMA x CD28 "B") and mAb14216P2 (CD28 parental "B") did not induce significantly more cytokine release compared to controls (data not shown).

[0437] The anti-CD28 antibodies of the present invention show several different properties compared to the superagonist TGN1412. Conventional CD28 agonistic mAbs (non-superagonists) are thought to bind to membrane distal epitopes, thereby allowing only monovalent Ab binding (Dennehy et al., "Cutting Edge: monovalency of CD28 maintains the antigen dependence of T cell costimulatory responses", Journal of Immunology 176(10):5725-29 (2006)). CD28 antibodies drive optimal T cell activation in the presence of peptide MHC antigen-specific recognition by TCR. CD28 signaling regulates the threshold of TCR activation and significantly reduces the number of TCR engagements required for effective activation of T cells. In contrast, TGN anti-CD28 antibodies have some superagonist properties (Luhder et al., "Topological requirements and signaling properties of T cell-activating, anti-CD28 antibody superagonists", Journal 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 induced potent T cell proliferation and IL2 production in vitro and in vivo, enhanced NF-kB activity and induced AP-1 / SRE activation more effectively than conventional anti-CD28 / CD3 signaling, and bound to a CD28 epitope near the cell surface, allowing bivalent binding.

[0438] Therefore, if Figure 8 As shown, the CD28 superagonist TGN 1412 activated the AP-1 reporter gene in the absence of CD3 primary stimulation. In contrast, the anti-CD28 antibody of the invention (parent of PSMA x CD28) minimally activated AP-1 in the AP-1 luciferase bioassay.

[0439] Example 10: In vivo studies of anti-PSMAxCD28 antibodies

[0440] Syngeneic Tumor Models

[0441] Consistent with the in vitro studies in Example 8, in a mouse model, combining anti-CD3xPSMA with anti-CD28xPSMA bispecific antibodies targeting tumor antigens enhanced tumor clearance. To test the effectiveness of anti-PSMAxCD28 bispecific antibodies as monotherapy or in combination with anti-PSMAxCD3 bispecific antibodies, syngeneic tumor experiments were performed in mice expressing human CD28, human CD3, and human PSMA, in place of the corresponding mouse genes (hCD3 / hCD28 / hPSMA mice) generated using Velocigene's proprietary technology. In conclusion, the anti-PSMAxCD28 bispecific antibody enhanced in vivo anti-tumor immunity through anti-PSMAxCD3-induced T cell activation.

[0442] T cell activation is initiated when the T cell receptor (TCR) / CD3 complex engages a peptide-MHC complex ("signal 1"); activation is then enhanced by engagement of a second "co-stimulatory" receptor (such as the CD28 receptor on the T cell) with one or more of its cognate ligands on the target cell (signal 2). Recently described CD3-based "bispecific antibodies" work by connecting T cells to tumor cells by binding a tumor-specific antigen (TSA) with one arm of the bispecific antibody, and bridging TCR / CD3 with the other arm, instead of the conventional signal 1. Although some of these TSAxCD3 bispecific antibodies have demonstrated promising anti-tumor efficacy in cancer patients, their activity remains to be optimized. As described elsewhere herein, the present invention introduces a novel class of bispecific antibodies that mimic signal 2 by bridging a second TSA to the co-stimulatory CD28 receptor on T cells. These bispecific antibodies are referred to as TSAxCD28 bispecific antibodies or anti-TSA / CD28 bispecific receptors. As described herein, an exemplary antibody of the present invention is specific for a prostate cancer antigen (e.g., PSMA). Unlike CD28 superagonists, which extensively activate T cells and in some cases cause severe toxicity in early clinical trials, these TSAxCD28 bispecific antibodies show limited activity and no toxicity when used alone in genetically humanized immunocompetent mouse models or primates. However, when combined with TSAxCD3 bispecific antibodies, the exemplary antibodies of the present invention enhance artificial synapses between T cells and their target cells, strengthen T cell activation, and significantly improve the anti-tumor activity of CD3 bispecific antibodies in syngeneic tumor models. Combining this class of novel CD28 co-stimulatory bispecific antibodies with an emerging class of TSAxCD3 bispecific antibodies can provide well-tolerated "off-the-shelf" antibody therapies with potentially enhanced anti-tumor efficacy.

[0443] The ability of T cells to recognize and kill their cellular targets (such as virus-infected cells or tumor cells) depends on a series of coordinated interactions. The most important of these is the recognition and binding of the target cell by the TCR complex (the complex contains the associated CD3γ, δ, ε, ζ chains); this interaction is called "signal 1" for T cell activation. TCR can recognize viral or tumor peptides present in the grooves of MHC proteins expressed on the surface of target cells. This binding is usually low affinity; therefore, in order to successfully trigger signal 1, it is important that many TCR complexes gather along the interface between the T cell and its target cell, and this interface is called the immune synapse (JBHuppa, MMDavis, "T cell-antigen recognition and the immunological synapse" "Nature Review of Immunology (Nat Rev Immunol)" 3, 973-983 (2003)). T cell activation and proliferation are then further promoted through additional interactions with costimulatory receptors such as CD28 ("signal 2") (JH Esensten, YA Helou, G. Chopra, A. Weiss, JA Bluestone, "CD28 co-stimulation: From Mechanism to Therapy" Immunity 44, 973-988 (2016)). T cell activation is enhanced when the T cell recognizes the target cell through the TCR complex and engages signal 2 by binding CD28 to one or more cognate ligands (CD80 / B7.1 and / or CD86 / B7.2) on a professional antigen presenting cell or target cell. Like signal 1, CD28-mediated signal 2 is thought to occur by co-aggregation at the immune synapse.

[0444] Over the past two decades, conventional monoclonal antibodies targeting tumor-specific antigens (TSAs) have been used as anti-tumor therapeutics (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., "Daratumumab plus Bortezomib, Melphalan, and Prednisone for Untreated Myeloma." N Engl Med 2014, 14(5):115-123. JMed) 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" Ann Oncol 12 Suppl 1, S57-62 (2001); JM Connors et al., "Brentuximab Vedotin with Chemotherapy for Stage III or IV Hodgkin's Lymphoma." New England Journal of Medicine 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 randomised, open-label, phase 3 trial." Lancet Oncol 18, 732-742 (2017). However, such antibodies have limited ability to induce T cell-mediated cytotoxicity and instead act 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 (TSAxCD3) has emerged that can efficiently trigger T cell-mediated tumor cell killing by ligating T cells to tumor cells and activating the CD3 / TCR complex (usually through the e-chain of CD3), thereby mimicking signal 1. Recently, an early version of this bispecific antibody (one arm binds to CD19 on leukemia cells and the other arm binds to CD3) 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." New England Journal of Medicine 376, 836-847 (2017)). More recently, a more advanced version of the bispecific antibody has been shown to have good activity against non-Hodgkin's lymphomas, which targets CD20 on these lymphomas (EJSmith 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); LL 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,” =Hematologic Malignancies”, Clin Cancer Res, 24, 4785-4797 (2018); R. Bannerji et al., “Emerging Clinical Activity of REGN1979, an Anti-CD20 xAnti-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 TSAxCD3 bispecific antibodies are emerging as an important new class of immunotherapy for hematological malignancies, cross-study comparisons (A. Zhukovsky, RJ Morse, MV Maus, "Bispecific antibodies and CARs: generalized immunotherapeutics harnessing T cell redirection" Curr Opin Immunol 40, 24-35 (2016)) suggest that in some cases, the bispecific antibodies may not achieve the level of efficacy observed in personalized chimeric antigen receptor T cell (CAR-T) therapy. .

[0445] One of the reasons for the powerful efficacy of CAR-T therapy is that the chimeric antigen receptor (CAR) is engineered to provide signal 1 (through part of the CD3z cellular domain) and signal 2 (for example, through part of the CD28 cellular domain) when it binds to its target on tumor cells. Recently, the FDA approved two CAR-T cell therapies for B-cell malignancies, both of which work by binding to and targeting the antigen CD19 (SS Neelapu et al., "Axicabtagene Ciloleucel CAR T-Cell Therapy in Refractory Large B Cell Lymphoma," New England Journal of Medicine 377, 2531-2544 (2017); SS Schuster et al., "Chimeric antigen receptor T cells in refractory B-cell lymphomas," New England Journal of Medicine 377, 2545-2554 (2017)). CAR-T cell approaches may be associated with severe side effects such as cytokine release syndrome (CRS) and neurotoxicity (SS Neelapu et 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)); and due to the highly personalized manufacturing process and the requirement for pre-conditioning chemotherapy regimens (S.S. Neelapu et al., “Axicabtagene Ciloleucel CAR T-Cell Therapy in Refractory Large B Cell Lymphoma” New England Journal of Medicine 377, 2531-2544 (2017); S. J. Schuster et al., “Chimeric antigen receptor T cells in refractory B-cell lymphomas” New England Journal of Medicine 377, 2545-2554 (2017); P. Salmikangas, N. Kinsella, P. Chamberlain, “Chimeric Antigen Receptor T-Cells (CART-Cells) for Cancer Immunotherapy-Moving Target for Industry?” Industry? )" Pharmaceutical Research (PharmRes) 35,152 (2018)), many patients are considered not suitable candidates.

[0446] The advantages of TSAxCD3 bispecific antibodies as relatively well-tolerated and "off-the-shelf" treatment options for a wider patient population will be enhanced if their anti-tumor activity can be further optimized, especially if it can be done without sacrificing tolerability, or even increasing the specificity of tumor cells relative to normal cells. For this purpose, it is assumed that pairing TSAxCD3 bispecific antibodies with a novel class of bispecific antibodies that independently activate signal 2 can provide potential enhanced efficacy as well as the opportunity to enhance specificity. Therefore, a second class of bispecific antibodies was designed. These bispecific antibodies can engage with a second epitope on the same tumor-specific antigen or a second separate tumor antigen, where the co-stimulatory receptor CD28 (TSAxCD28 bispecific antibody) is expressed on T cells. It is inferred that combining TSA1xCD3 with TSA2xCD28 should allow for the guidance and enhancement of alternative activation of T cells by triggering signal 1 and signal 2, with its specificity targeting only tumor cells expressing two epitopes or two antigens, thereby allowing greater anti-tumor activity and the opportunity to increase specificity.

[0447] This article describes the generation and testing of TSAxCD28 co-stimulatory bispecific antibodies (PSMAxCD28, which bind to prostate-specific membrane tumor antigen) targeting prostate cancer. Toxicology studies in genetically humanized immunocompetent mice and cynomolgus monkeys showed that these bispecific antibodies exhibited limited activity and no toxicity as single agents. However, these novel co-stimulatory bispecific antibodies can be effectively combined with an emerging class of TSAxCD3 bispecific antibodies to enhance anti-tumor responses in syngeneic tumor models. Overall, these data suggest that combining this novel CD28-based bispecific antibody (TSAxCD28) with a CD3-based bispecific antibody (TSAxCD3) can provide a well-tolerated "off-the-shelf" biologic solution with significantly enhanced and synergistic anti-tumor activity.

[0448] Materials and Methods

[0449] The following materials and methods were used in Examples 10 and 13.

[0450] Syngeneic Tumor Studies

[0451] As mentioned before, use Mice expressing human CD28, human CD3, and human PSMA (referred to as hCD3 / hCD28 / hPSMA humanized mice) were generated using a technique to replace the corresponding mouse genes (Valenzuela (2003), Nat Biotechnol, Jun;21(6):652-9; Crawford et al., 2018, manuscript in preparation). For each humanized mouse, the correct gene targeting in F1H4 (C57BL / 6×129 hybrid) embryonic stem (ES) cell clones was identified by allelic deletion assay as previously described (Poueymirou et al. (2007), Nat Biotechnol, Jan;25(1):91-9). The targeted ES cells were injected into Swiss Webster embryos at the 8-cell stage to generate fully heterozygous F0 mice for homozygous breeding with C57BL / 6N mice (Taconic, Rensselaer Polytechnic Institute, New York, USA). hCD3 / hCD28 / hPSMA mice (4-8 / group, 8-16 weeks old) were injected subcutaneously with 1x10 6 On days 0, 3, and 7, anti-PSMAxCD28 bispecific antibody, anti-PSMAxCD3 bispecific antibody, or human IgG4 isotype control were administered as monotherapy or combination therapy at 5 mg / kg by intraperitoneal injection.

[0452] Tumor growth was monitored over time using caliper measurements of X and Y diameters. Tumor volume was calculated (X*Y*(X / 2)). 3 When , the mice were euthanized.

[0453] Measurement of serum cytokine levels in mice

[0454] At the indicated time points, blood was collected by submandibular puncture into micro serum collection tubes (BD 365967). Cytokine levels were analyzed using the V-plex Human Pro-Inflammatory-10 Plex Kit according to the manufacturer's instructions (Meso Scale Diagnostics, Rockville, MD).

[0455] Results, Summary and Conclusion

[0456] 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). Growth of MC38 / hPSMA tumors was monitored over time in triple humanized hCD3 / hCD28 / hPSMA mice (described above). Fig.9A ). Both PSMAxCD3 and PSMAxCD28 monotherapy treatments significantly inhibited tumor growth compared to isotype control (p<0.001 and p<0.0001, respectively). Tumor growth inhibition was further significantly achieved by combination therapy (p<0.00001). The combination of PSMAxCD3 and PSMAxCD28 bispecific antibodies provided the greatest cytokine production compared to isotype control ( Fig. 9C and 9D To further understand the effects of combination therapy on intratumoral CD8 + The role of T cell activation status was analyzed by viSNE ( Fig. 9B ). viSNE is a tool for mapping high-dimensional cytometry data to 2D while preserving the high-dimensional structure. Each treatment drove unique CD8+ T cell clusters in the spleen and tumor. Combination therapy drove an expansion of the activated / memory T cell phenotype (expressing TCF1, CD1-2, CD127, PD-1, ICOS, KLRG1, and CD38), as shown in cluster 4 in the tumor.

[0457] Studies were conducted to determine whether hCD3 + / + / hPSMA + / + / hCD28 + / + Mouse and hCD3 + / - / hPSMA + / - / hCD28 + / -Serum cytokine levels (IFNγ, IL-2, IL-6, IL-10, TNFα, IL-4, and IL-5) in mice. Blood was collected on days 0 (4 hours), 3, 7, and 11 after treatment. + / + / hPSMA + / + / hCD28 + / + IL-10 in mice, cytokine levels were significantly increased in treatments containing anti-PSMAxCD3 antibodies at day 0, with the combination therapy using anti-PSMAxCD3 and anti-PSMAxCD28 providing the greatest cytokine release. In contrast, IgG control and anti-PSMAxCD28 monotherapy did not result in increased cytokine production. With the exception of IL-5 and IL-10, cytokine production in all treatment groups declined to similar levels after three (3) days. + / + / hPSMA + / + / hCD28 + / + In mice, IL-5 production was significantly higher on day 3 (although lower than on day 0), but in hCD3 + / - / hPSMA + / - / hCD28 + / - In mice, production was not increased. + / + / hPSMA + / + / hCD28 + / + In mice, IL-10 production was similar in all treatment groups on days 0 and 3, but the combination therapy produced significantly higher levels on days 7 and 11. + / - / hPSMA + / - / hCD28 + / - In mice, the combination therapy produced significantly more IL-10 on days 0, 3, and 11, but not on day 7. Anti-PSMAxCD3 monotherapy produced significantly more IL-10 on days 0 and 3, but not on days 7 and 11. Anti-PSMAxCD28 monotherapy produced significantly more IL-10 only on day 11 (data not shown).

[0458] like Fig.9AAs shown, unlike previous in vitro assays in which the CD28 bispecific antibody had very limited single agent activity (see Example 8 above), the CD28 bispecific antibody in this syngeneic MC38 / hPSMA model had more significant activity as a single agent. This suggests that in this MC38 model, "Signal 1" has been activated to some extent. Consistent with this, it has been shown previously that MC38 tumor cells express high levels of reactivated endogenous retroviral proteins, such as p15E, and that C57BL6 mice can generate endogenous T cells that recognize and respond to this neoepitope (JC Lang, 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 ed., D. Perry-Lalley, MEDudley, SA Rosenberg, JC Lang, “High avidity CTLs for two self-antigens demonstrate superior in vitro and in vivo antitumor efficacy” efficacy)" Journal of Immunology 162, 989-994 (1999)). In fact, it has been demonstrated that in the MC38 model, intratumoral T cells responding to this p15E neoantigen can be readily detected. Therefore, in this syngeneic tumor model, the CD28 bispecific antibody can enhance the endogenous TCR / CD3-dependent T cell response, which can then be further enhanced by providing additional "signal 1" activation through the CD3 bispecific antibody.

[0459] To determine the cellular mechanisms of the combination therapy, tumor-infiltrating cells and splenic CD8+ T cells from these experiments were analyzed by high-dimensional flow cytometry and an unsupervised clustering approach was used. It was found that each treatment drove unique CD8+ T cell clusters in the spleen and tumor. As shown in cluster C35, the single treatment regimen reduced CD8+ T cells with a less activated phenotype within the tumor (lower ICOS, KLRG1, Ki67, PD1, CD38, and LAG3, Fig.9EHowever, the combination therapy significantly drove a more activated / memory T cell phenotype (expression of Tcf1, CD122, CD127, PD1, ICOS, KLRG1, and CD38, Fig. 9B ), as shown in cluster C4.

[0460] Anti-CD28 x anti-PSMA does not induce serum cytokine elevation in CD28 / CD3 / PSMA humanized mice in the absence and presence of tumor

[0461] As described elsewhere herein, the superagonist anti-CD28 antibody TGN1412 induces a "cytokine storm" that is harmful to patients. The anti-CD28 antibodies and anti-CD28xPSMA 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. Blood was drawn from the mice 4 hours and 3 days after antibody administration. Fig.10 As shown, in the absence of tumor, anti-PSMAxCD28 bispecific antibody or parental CD28 bivalent antibody did not induce serum cytokine production in CD3 / CD28 / PSMA humanized mice.

[0462] Table 19: Cytokine storm studies of antibodies in hCD3 / hCD28 / hPSMA humanized mice

[0463] Antibody ID Batch number name REGN1945 REGN1945-L49 hIgG4 mAb14226P2 REGN5705-L1 CD28 mAb REGN2329 (TeGenero) REGN2329-L3 CD28 SA bs16429D bs16429D-L2 PSMAxCD28 BE0146 614216J2B PD1, RMP1-14 bs15780D bs15780D-L2 PSMAxCD3

[0464] At 2.5 mg / kg, TGN1412 could lead to significant increases in the levels of IL-2, IL-4, IL-5 and TNFα in the blood (data not shown).

[0465] In addition, if Fig.28As shown, anti-PSMAxCD28 treatment alone or in combination with anti-PD1 did not elevate serum cytokines in tumor-bearing mice. In contrast, anti-PSMAxCD3 treatment alone or in combination with anti-PD1 increased serum cytokine levels in tumor-bearing mice, such as TNFα, IL-5, IL-10, IL-2, and IL-4. Anti-PSMAxCD3 treatment induced cytokine expression 4 hours after administration. The cytokine elevation was not sustained after day 7. This study was done in MC38 / hPSMA tumor-bearing hCD3 / hCD28 / hPSMA humanized mice. Serum was collected from MC38 / hPSMA tumor-bearing hCD3 / hCD28 / hPSMA humanized mice at day 0 and day 7, 4 hours after administration, and 4 days later with 5 mg / kg of the indicated antibodies or bispecific antibodies. Serum cytokines were measured using a 10-plex mouse cytokine MSD kit following the manufacturer's protocol.

[0466] In addition, if Fig.11 As shown, CD28 superagonist treatment drives cytokine responses within 4 hours in NSG mice transplanted with PBMCs, whereas the anti-CD28 antibodies of the invention do not. Fig.11 In the studies shown, immunodeficient NSG mice were transplanted with 5x10 6 PBMC.On the 1st day, the systemic transplantation of T cells was confirmed by staining human T cell markers in peripheral blood.On the 0th day, 50 μg IgG4 isotype control antibodies, 5 μg or 50 μg anti-CD28 super agonists, or 5 μg or 50 μg anti-CD28 mAb14226P2 antibodies were injected intraperitoneally into mice.Four hours after antibody injection, blood was harvested from animals and serum was prepared.The cytokine concentration in serum was analyzed by multiple determinations (Meso Scale Discovery V-PLEX kit).Although compared with animals treated with isotype control antibodies, anti-CD28 super agonists drive interferon gamma (IFN-γ), IL-2, IL-6, and TNFα serum levels to increase, but no increase in cytokine responses was observed in the serum of animals treated with anti-CD28mAb14226P2.Therefore, in costimulatory bispecific antigen binding molecules, bispecific anti-CD28xPSMA has a potentially safer toxicological profile. The bispecific anti-CD28xPSMA did not elicit a cytokine response, whereas anti-CD3xPSMA did. Fig.12As shown, anti-CD3xPSMA induced elevated IFNγ in humanized mice, whereas anti-CD28xPSMA did not. In this study, serum was collected from MC38 / hPSMA tumor-bearing CD3 / CD28 / PSMA or CD3 / CD28 humanized mice 4 hours after administration of 5 mg / kg of the indicated bispecific antibodies. Serum cytokines were measured using the 10-plex Mouse Cytokine MSD Kit following the manufacturer's protocol. Data points represent individual cytokine levels per mouse. Bars on the graph represent the mean for each treatment group. Error bars represent + / - SEM. Statistical significance was determined using one-way ANOVA and Holm-Sidak multiple comparison test using isotype-treated mice as controls (**, p<0.01)

[0467] Consistent with the humanized mouse data, anti-CD3xPSMA bispecific antibodies lead to elevated CRP and systemic cytokine release in non-human primates (cynomolgus monkeys). In monkey studies, anti-CD3xPSMA bispecific antibodies were administered to cynomolgus monkeys at doses of 0.01, 0.1, and 0.5 mg / kg. Monkeys showed the following symptoms: clinical symptoms on the first day, such as vomiting, hunchback, skin redness / discoloration; elevated CRP (approximately 10 vs 1 mg / dL with control mice); elevated plasma cytokines (IL-6, TNF-α, IFN-γ, IL-2, and MCP); and decreased absolute T cell counts in the blood (data not shown).

[0468] Furthermore, mice treated with anti-CD28xPSMA showed no cytokine production or T cell marginalization compared with anti-PSMAxCD3 or anti-CD28 superagonist treatment ( Fig.13A and 13B ).

[0469] Summarize

[0470] It has long been recognized that T cell activation through the TCR complex (“signal 1”) can be significantly enhanced by co-stimulatory signals, such as the co-stimulatory signals mediated when the CD28 receptor on T cells engages its ligands (CD80 / B7.1 and CD86 / B7.2) on target cells (“signal 2”) (JH Esensten, YA Helou, G. Chopra, A. Weiss, JA Bluestone, “CD28 co-stimulation: from mechanism to therapy” Immunity 44, 973-988 (2016)). Consistent with the data disclosed herein, the potential of CD28-costimulation to enhance T cell antitumor activity was first demonstrated by studies on overexpression of B7 ligands on tumor cells (R. H. Schwartz, "Costimulation of T lymphocytes: the role of CD28, CTLA-4, and B7 / BB1 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 in antitumor immunity." This potential inspired efforts to evaluate CD28-activating antibodies in human trials. Sadly, a 2006 trial of this antibody (TGN1412) resulted in life-threatening complications in all six human volunteers due to multi-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," New England Journal of Medicine 355, 1018-1028 (2006)). This disaster led to the cessation of any further testing of CD28-activating antibodies in humans.

[0471] This article describes a novel class of CD28 co-stimulatory bispecific antibodies that can significantly and safely promote anti-tumor activity by providing co-stimulatory "signal 2". These CD28 bispecific antibodies have limited activity themselves (in the absence of "signal 1"), but can significantly enhance anti-tumor activity in the presence of "signal 1", such as by pairing these CD28 bispecific antibodies with a class of emerging CD3 bispecific antibodies (or if these CD28 bispecific antibodies are used in an environment where an endogenous population of tumor-specific T cells already exists). This article describes the production and testing of TSAxCD28 co-stimulatory bispecific antibodies (PSMAxCD28) targeting prostate cancer. The results show that these CD28 bispecific antibodies have minimal activity in vitro or in vivo in the absence of "signal 1". However, these CD28 bispecific antibodies can be paired with CD3 bispecific antibodies to form artificial "immune synapses" containing tumor antigens and TCR and CD28 complexes. In addition, when paired with appropriate CD3 bispecific antibodies in vitro, these CD28 bispecific antibodies can effectively and specifically promote T cell activation and tumor cell killing in an antigen-dependent manner. In addition, in a syngeneic tumor model, these CD28 bispecific antibodies also effectively enhance the anti-tumor activity of CD3 bispecific antibodies in vivo in a tumor antigen-specific manner; in this model, CD28 bispecific antibodies have minimal single-agent activity unless tumor-specific T cells are already present, and in this environment, the bispecific antibodies appear to enhance this specific activity in a tumor antigen-dependent manner. In addition, TSAxCD28 and TSAxCD3 combination therapies significantly drive the expansion of activated / memory T cell phenotypes in tumors in vivo. Finally, toxicology studies in genetically humanized immunocompetent mice and cynomolgus monkeys showed that these bispecific antibodies exhibited limited activity and no toxicity as single agents in direct comparison with conventional CD28 activating antibodies.

[0472] Generally, in the field of immuno-oncology, the characterization of human-specific clinical candidates is limited to testing in xenogeneic tumor models with transplanted human immune cells. Although these xenogeneic models can be very useful, they have limitations. The mice used in this xenogeneic model do not express human tumor targets in their normal tissues, thus eliminating the assessment of test agents in the normal tissue expression environment of the target. In fact, if the target is usually also expressed at a high level in normal tissues, this may limit the anti-tumor efficacy by transferring the test agent from the tumor, and may produce toxicity to these normal tissues-these toxicities cannot be assessed in xenogeneic models under normal circumstances. Additional restrictions may relate to the activity of transplanted human peripheral blood mononuclear cells (PBMC) transferred to immunodeficient mice, which may be different from the activity of normal host T cells found in the immune active system. In order to overcome these limitations and provide a better model for testing human-specific clinical candidates, double and triple gene humanized mice are produced herein. In these models, tumor antigens were genetically humanized to allow their normal expression in appropriate host tissues, and 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, CD28 bispecific antibodies to the PSMA tumor target were found to enhance the anti-tumor activity of their appropriate CD3 bispecific antibodies. Similar enhancements in anti-tumor efficacy of different TSAxCD28 bispecific antibodies in multiple preclinical models suggest that this treatment paradigm is robust, not limited to specific tumor models, and may have broader utility as a new combination target class for immunotherapy. Overall, these findings highlight that TSAxCD28 bispecific antibodies can synergize with TSAxCD3 bispecific antibodies and may provide a biological solution that significantly enhances the efficacy of well-studied TSAxCD3 bispecific antibodies in a reasonably safe and well-tolerated manner, thus justifying testing in human trials.

[0473] TSAxCD3 bispecific antibodies represent a promising new class of immunotherapies, but further optimization of antitumor activity is certainly important in many cases. Just as the CAR-T approach uses chimeric receptors that artificially activate "signal 1" and "signal 2" to improve their antitumor activity (EA Zhukovsky, RJ Morse, MV Maus, "Bispecific antibodies and CARs: broad immunotherapy using T cell redirection" Current Opinion in Immunology 40, 24-35 (2016); SL ​​Maude et al., "Tisagenlecleucel in Children and Young Adults with B-Cell Lymphoblastic Leukemia" Leukemia” New England Journal of Medicine 378, 439-448 (2018)), now shows the potential benefit of combining a CD3-specific antibody (providing “signal 1”) with a CD28 bispecific antibody (which provides “signal 2”) to enhance anti-tumor activity. In addition to the practical benefits that this approach may have over CAR-T therapy - because it does not require laborious cell therapy preparation that is individually tailored for each patient, nor does it require patients to be pre-treated with “lymphoid depletion” through toxic chemotherapy that is often associated with side effects in order to make them acceptable to this cell therapy (A. Shimabukuro-Vornhagen et al., “Cytokine release syndrome” Journal of Cancer Immunotherapy 6, 56 (2018); CH June, RSO'Connor, OU Kawalekar, S. Ghassemi, MC Milone, “CAR T cell immunotherapy for human cancer” Cancer)" Science 359, 1361-1365 (2018)) - the bispecific approach according to the present invention offers the potential for increased efficacy as well as increased safety and specificity of action. That is, by pairing a CD3 bispecific antibody against one antigen with a CD28 bispecific antibody against a second antigen, the advantage of "combination targeting" can be exploited - the increased efficacy will only occur on tumor cells expressing both antigens, thereby focusing T cell killing only on tumor cells expressing both antigens while limiting "off-target toxicity" in normal tissues expressing only one of the antigens. In summary, the data disclosed herein indicate that combining a CD28-based bispecific antibody with a CD3-based bispecific antibody can provide a well-tolerated "off-the-shelf" biologic solution with significantly enhanced and synergistic anti-tumor activity. This possibility will be initially tested in human trials this year.

[0474] Example 11: Toxicology study in cynomolgus monkeys

[0475] These studies indicate that PSMAxCD28 alone or in combination therapy does not induce systemic T cell effects compared to CD28 superagonists in cynomolgus monkeys. Exemplary anti-PSMAxCD28 bispecific antibodies of the invention potentiate TAAxCD3 activation of cynomolgus monkey T cells (Example 8, Figure 7E-7G To determine the safety and tolerability of exemplary anti-PSMAxCD28 bispecific antibodies of the invention alone or in combination with anti-PSMAxCD3, a single-dose toxicity study was performed in cynomolgus monkeys. Female or male cynomolgus monkeys were assigned to treatment groups as indicated in Table 20.

[0476] Table 20: Exploratory Single-Dose Monkey Toxicology Study with Anti-PSMAxCD28

[0477]

[0478]

[0479] The cynomolgus monkey study was conducted according to the IACUC guidelines. Male cynomolgus monkeys (3 animals / group) received a single dose of each test article by intravenous infusion for approximately 30 minutes (combination therapy was administered as a separate infusion, 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 anatomical pathology. Blood and tissue samples were collected for cytokine analysis, FACS immunophenotyping, histopathology and toxicokinetic assessment. CRP levels were analyzed on the Roche Modular P800 system. Cytokines were measured by Meso Scale Diagnostics (MSD in Rockville, Maryland). For peripheral blood flow cytometry, blood was collected into potassium EDTA tubes, lysed, stained with indicated antibodies such as anti-CD3, anti-Ki67, and anti-ICOS (BD Biosciences), and analyzed using FACS Canto II. No significant cytokine release, T cell marginalization, or upregulation of T cell activation markers was observed after single-dose administration of 1 or 10 mg / kg PSMAxCD28, which did not induce significant cytokine release, T cell marginalization, or upregulation of T cell activation markers (Table 21).

[0480] Table 21: Summary of toxicity studies in cynomolgus monkeys

[0481]

[0482]

[0483] BLQ: below the limit of quantification

[0484] In contrast, significant cytokine release, lymphocyte marginalization, and T cell activation were observed in monkeys administered CD28 superagonists. Further, these findings were validated using dry and wet coated human T cell proliferation assays (Example 9). Indeed, no human T cell proliferation induced by the exemplary anti-PSMAxCD28 bispecific antibodies as well as the parental bivalent CD28 antibody was observed compared to the CD28 superagonist antibody. Overall, the exploratory single-dose toxicology studies in monkeys and in vitro human T cell-based assays indicate that the exemplary anti-PSMAxCD28 bispecific antibodies of the invention are safe and well tolerated.

[0485] As shown in Table 21, the CD28 superagonist TGN1412 causes a moderate release of cytokines and a transient increase in T cells (measured on D15). This study was completed on male cynomolgus monkeys, with weekly dosing of 5 to 50 mg / kg for 4 weeks. After intravenous injection of TGN1412 at increasing doses (5, 10, 25, 50 mg / kg) weekly (d1, 8, 15, 22), FACS analysis of leukocyte subsets was performed on male cynomolgus monkeys. Moderate increases (2 to 20 times) in IL-2, IL-5, IL-6, and IFN-γ were observed within the first 2 to 24 hours. There were no substantial changes in IL-4 and TNF-α. An exploratory single-dose monkey toxicology study using anti-CD28xPSMA was conducted to determine the safety and pharmacokinetic characteristics of anti-CD28xPSMA bispecific antibodies.

[0486] In summary, it has been shown that anti-CD28xPSMA antibodies enhance immunity at tumor sites. Anti-CD28xPSMA transforms tumor cells into antigen presenting cells (APCs). Compared to anti-CD28 superagonists, anti-CD28xPSMA antibodies do not induce T cell proliferation or stimulate cytokine release.

[0487] Example 12: Epitope mapping of mAb 14226P2 binding to CD28 by hydrogen-deuterium exchange

[0488] H / D exchange epitope mapping and mass spectrometry (HDX-MS) were performed to identify the amino acid residues of CD28 (recombinant human CD28, shown as hCD28 extracellular (N19-P152). mmh; SEQ ID NO: 75) that interact with the anti-hCD28 monoclonal antibody. A general description of H / D exchange is shown, for example, in Ehring (1999) Analytical Biochemistry 267(2): 252-259; and Engen and Smith (2001) Analytical Chemistry 73: 256A-265A.

[0489] 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 stage (μBinary Solvent Manager) for sample digestion and loading, a Waters Acquity M stage (μBinary Solvent Manager) for analytical gradients, and a Thermo Q Exactive HF mass spectrometer for peptide mass measurement.

[0490] At pD 7.0, the labeling solution was prepared as PBS buffer (10mM phosphate buffer, 140mM NaCl and 3mM KCl, equivalent to pH 7.4 at 25°C) in D2O. For deuterium labeling, at various time points, 11μCD28.mmH (Regeneron's inhouse protein REGN2011, 127μM) or CD28.mmH premixed with REGN5705 at a 1:0.6 molar ratio (Ag-Ab complex) was incubated with 44μL D2O labeling solution at 20°C in duplicate (e.g., undeuterated control = 0 seconds; deuterium labeled, 5 minutes and 10 minutes). The deuteration reaction was quenched by adding 55μL pre-cooled quenching buffer (0.5M TCEP-HCl, 8M urea and 1% formic acid) to each sample and incubated at 20°C for 5 minutes. The quenched sample was then injected into a Waters HDX Manager for online pepsin / protease XIII digestion. The digested peptides were separated by a C8 column (1.0 mm x 50 mm, NovaBioassays) with a 13-minute gradient from 10%-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.

[0491] The LC-MS / MS data of the undeuterated CD28 samples were searched against a database containing CD28 and its random sequences using the Byonic search engine (Protein Metrics). The search parameters (in ELN) were set to default values ​​using nonspecific enzymatic digestion and human glycosylation as common variable modifications. The identified peptide list was then imported into HDX Workbench software (version 3.3) to calculate the deuterium uptake for 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 the percentage of deuterium uptake (%D).

[0492] Deuterium absorption (D-absorption) = average mass (deuterated) - average mass (undeuterated)

[0493]

[0494] A total of 73 peptides from hCD28.mmH were identified from hCD28.mmH alone and hCD28.mmH mixed with anti-CD28 antibody samples (SEQ ID NO: 75), which represents a sequence coverage of 85.8% for hCD28. Any peptide that exhibited a D-absorption magnitude difference percentage greater than 5% was defined as being 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, wherein amino acids 91-108 (IYFCKIEVMYPPPYLDNE; SEQ ID NO: 80) were defined as the major epitope on CD28. Protection of these residues by anti-CD28 antibodies was confirmed using hCD28.mFc (SEQ ID NO: 76). A summary of the results of this study is also provided in Table 22 below.

[0495] Table 22: Selected CD28.mmH peptides that have significant protective effects when combined with anti-CD28 antibodies

[0496]

[0497]

[0498] Example 13: PSMA x CD28 bispecific antibody effectively enhances the anti-tumor efficacy of PD-1 immunotherapy

[0499] summary

[0500] In addition to antigen-specific TCR / CD3 activation ("signal 1"), T cell activation is enhanced by engaging a second co-stimulatory receptor ("signal 2"). The goal of cancer immunotherapy is to optimally activate and mobilize T cells to detect and kill tumor cells. However, current treatments often fail to effectively and selectively activate T cells at the tumor site, which often fails to achieve a durable response and / or causes undesirable toxicity. This article introduces a novel tumor-targeted immunotherapy model that combines PD-1 inhibition with bispecific antibodies. Bispecific antibodies bind to tumor-specific antigens (TSAs) (e.g., PSMA) with one arm and to co-stimulatory receptor CD28 on T cells with another arm. In fact, PD-1-PD-L1 signaling inhibition significantly increases the ratio of CD28 accumulated in the immune synapse, allowing the TSAxCD28 bispecific antibody to work. This combination immunotherapy was validated using a bispecific antibody specific for prostate antigens (e.g., PSMA). Unlike nonspecific CD28 superagonists that broadly activate T cells, TSAxCD28 bispecific antibodies were well tolerated when used alone or in combination with PD-1 blockade in genetically humanized immunocompetent mouse models or in primates. Importantly, in the presence of endogenous TCR / CD3 triggering, TSAxCD28 significantly improved the antitumor activity of PD-1 antibodies, which was associated with durable antitumor responses. In various syngeneic and human tumor xenograft models, the combination therapy specifically enhanced intratumoral T cell activation, promoting an effector memory-like T cell phenotype without systemic cytokine secretion. Combining such CD28 co-stimulatory bispecific antibodies with clinically validated anti-PD-1 treatments could provide well-tolerated, “off-the-shelf” antibody therapies with significantly enhanced antitumor efficacy.

[0501] introduce

[0502] Many monoclonal antibodies (mAbs) designed to enhance T cell activation are in clinical development as anti-tumor therapeutics (MK Callahan, MA Postow, JD Wolchok, "Targeting T Cell Co-receptors for Cancer Therapy" Immunity 44, 1069-1078 (2016)). However, most current treatments face the challenge of overcoming the suppressive nature of the tumor microenvironment and therefore fail to produce effective tumor-specific T cell activation and subsequent tumor cell killing (KG Anderson, IM Stromnes, PD 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 against checkpoint inhibitors, such as cytotoxic T lymphocyte-associated protein (CTLA-4) and programmed cell death 1 (PD-1) / programmed cell death ligand 1 (PD-L1), have been clinically approved for the treatment of melanoma, renal cell carcinoma, non-small lung cancer, and advanced metastatic cutaneous squamous cell carcinoma (JS Weber et al., “Nivolumab versus chemotherapy in patients with advanced melanoma who progressed after anti-CTLA-4 treatment (CheckMate 037): a randomised, controlled, open-label, phase 3 trial,” Lancet Oncology 16, 375–384 (2015); SLTopalian 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 Magden et al., “PD-1 Blockade with Cemiplimab in Advanced Cutaneous Squamous-Cell Carcinoma,” NEJM 379, 341-351 (2018)). Blocking PD-1 releases interruptions in T cell activation, but its efficacy as a single agent is generally insufficient to achieve tumor clearance and durable anti-tumor responses. When combined, anti-PD-1 / PD-L1 and anti-CTLA-4 do produce high response rates in certain tumor types, but high-grade toxicity is also generally observed (J. Larkin et al., "Combined Nivolumab and Ipilimumab or Monotherapy in Untreated Melanoma" New England Journal of Medicine 373, 23-34 (2015); DB Johnson et al., "Fulminant Myocarditis with Combination Immune Checkpoint Blockade" New England Journal of Medicine 375, 1749-1755 (2016); MHPollack et al., “Safety of resuming anti-PD-1 in patients with immune-related adverse events (irAEs) during combined anti-CTLA-4 and anti-PD1 in metastatic melanoma,” Annals of Oncology 29, 250-255 (2018); JD Wolchok et al., “Nivolumab plus ipilimumab in advanced melanoma,” New England Journal of Medicine 369, 122-133 (2013)). Therefore, considerable efforts are underway to identify patients who are more likely to respond to checkpoint inhibition by identifying biomarkers that predict responsiveness (R. Cristescu et al., "Pan-tumor genomic biomarkers for PD-1 checkpoint blockade-based immunotherapy," Science 362, (2018)). In addition, combination therapies aimed at improving the efficacy and durability of anti-tumor responses of PD-1 blockade by combining agonistic antibodies that trigger co-stimulatory receptors or in combination with other modalities such as chemotherapy or radiation therapy are undergoing preclinical and clinical evaluation (S. Hu-Lieskovan, A. Ribas, "New Combination Strategies Using Programmed Cell Death 1 / Programmed Cell Death Ligand1Checkpoint Inhibitors as a Backbone" Cancer J 23, 10-22 (2017); YKChae 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)” Journal of Cancer Immunotherapy 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. Med) 283, 110-120 (2018); B. Wang et al., "Combination cancer immunotherapy targeting PD-1and 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). The instability and high mutational burden of microsatellite markers generate potential endogenous antigens that are expressed in certain cancers (KW Mouw, MSGoldberg, P.A. Konstantinopoulos, A.D. D'Andrea, "DNA Damage and Repair Biomarkers of Immunotherapy Response" Cancer Discovery 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" Frontiers in Immunology 8, 1679 (2017)). However, in isolation, presentation of these antigens is insufficient to promote robust T cell activation to generate antitumor activity (S. Spranger, R. Bao, TF Gajewski, "Melanoma-intrinsic beta-catenin signaling prevents anti-tumour immunity," Nature 523, 231-235 (2015)). This may be due to the immunosuppressive microenvironment of the tumor.

[0503] This article describes a novel immunotherapy paradigm using TSAxCD28 bispecific antibodies targeting the prostate cancer TSA PSMAxCD28, which, when combined with a PD-1 blocking antibody, induced long-term anti-tumor immunity in animal tumor models and promoted robust intratumoral T cell activation and T cell memory without signs of systemic cytokine release. Toxicology studies in genetically humanized immunocompetent mice and cynomolgus monkeys showed that these bispecific antibodies were non-toxic either by themselves or in combination with anti-PD-1 antibodies. Overall, these data suggest that combining such CD28-based bispecific antibodies (TSAxCD28) with PD-1 inhibition can provide a well-tolerated "off-the-shelf" biologic solution with significantly enhanced, specific, and synergistic anti-tumor activity.

[0504] Materials and Methods

[0505] The following materials and methods were used in Example 13.

[0506] Study Design

[0507] An exemplary purpose of the present invention is to develop TSAxCD28 bispecific antibodies and demonstrate that TSAxCD28 enhances PD-1-induced T cell activation in vitro and safely enhances anti-tumor efficacy in vivo. In vitro activity is demonstrated by showing images of bispecific antibodies PD-1 and CD28, PD-1 T cell cytokine release enhanced at the immune synapse of T cell and target cell conjugates. In vivo anti-tumor efficacy was evaluated in a syngeneic mouse tumor model. Tumor volume and serum cytokines were monitored over time to show responses to bispecific antibody treatment. One purpose of the cynomolgus monkey study is to determine the safety and tolerability (pharmacology and toxicology characteristics) of TSAxCD28 in non-human primates as a monotherapy or in combination with PD-1. Toxicity to animals is examined by clinical observation and blood sample collection to analyze serum cytokines and T cell phenotypes.

[0508] Animal studies

[0509] All procedures were performed in accordance with the NIH Guide for the Care and Use of Laboratory Animals. The protocol was approved by the Institutional Animal Care and Use Committee of Regeneron Pharmaceuticals.

[0510] Cell lines

[0511] 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 ATCC recommended protocol. HEK293 / hCD20 cell lines were generated using a mammalian vector encoding hCD20 (amino acids M1 to P297 of accession number NP_068769.2) driven by the ubiquitin promoter. Transfected cells were cultured in 500 μg / ml of Geneticin A (G418) to select for stably expressing cell lines. To generate cells expressing hCD80 or hPDL1, HEK293T cells were transfected with lentiviral plasmids encoding human CD80 (288aa long; Accession No. NM_005191.4) and neomycin resistance gene or human PDL1 (290aa long; Accession No. NM_14143.4) and puromycin resistance gene, thereby promoting the production of viral particles, which were 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 No. CLS-013L) and lentiviral plasmids 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.

[0512] 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 cells that stably express PSMA. The resulting cell line DU145 / PSMA was maintained in MEM+10% FBS+P / S / G with 500 μg / mL G418.

[0513] To generate tumor cell lines engineered to express co-stimulatory ligands, HEK293T cells were transfected with pLVX lentiviral plasmids with EF1a promoters encoding mouse CD86 and puromycin resistance genes or empty vectors and puromycin resistance genes (pLVX.EF1a.CD86-puro and pLVX.EF1a.EV-puro, respectively), thereby promoting the production of viral particles, which were 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 as MC38 / CD86 and MC38 / EV.

[0514] 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 / hPMA was maintained in DMEM+10% FBS+P / S / G+NEAA supplemented with 500 μg / mL G418.

[0515] Amnis Image Stream

[0516] Amnis image streaming was performed as described in Example 7.

[0517] Isolation of primary human CD3+ T cells

[0518] T cell isolation was performed as described in Example 8.

[0519] In the MLR reaction with DU145 / PSMA cells, IL-2 was expressed from primary CD3 + Released from T cells

[0520] On the day of the assay, previously isolated and frozen human CD3 T cells were thawed in stimulation medium (X-VIVO 15-cell culture medium supplemented with 10% FBS, HEPES, NaPyr, NEAA and 0.01 mM BME) containing 50 U / ml benzonase nuclease. + T cells. The cells were centrifuged at 1200 rpm for 10 minutes, resuspended in stimulation medium, and cultured at 1x10 5DU145 parental cells or DU145 cells engineered to express human PSMA were plated at a concentration of 10 × 10 cells / well in a 96-well round-bottom plate. DU145 parental cells or DU145 cells engineered to express human PSMA were treated with 25 μg / mL mitomycin C in the primary stimulation medium at a concentration of 10 × 10 6 After incubation at 37°C, 5% CO2 for 1 hour, the cells treated with mitomycin C were washed three times with D-PBS containing 2% FBS and 5×10 cells / mL were added to each well. 4 The final concentration of cells was added to the CD3 + T cell wells. To prevent possible CD28 agonistic activity by Fc-anchoring CD28 antibodies to Fc-receptors, saturating amounts of nonspecific human IgG antibodies (100 nM each: hIgG1, hIgG4, and hIgG4s) were included in each assay well. Subsequently, PSMA x CD28, non-target x CD28 controls, or hIgG4s isotype controls, antibodies were titrated from 30 pM to 200 nM at a 1:3 dilution and then added to the wells. The final point of the 10-point dilution did not contain titrated antibodies. Since DU145 cells endogenously express PD-L1, the effect of PD-1 inhibition on T cell activity was assessed by adding a constant 20 nM of PD-1 antagonist REGN2810 to the wells. Conditions in the absence of PD-1 inhibition were also included, in which a 20 nM matching hIgG4 isotype control was used. The plate was incubated at 37 ° C, 5% CO2 for 72 hours, and then centrifuged to precipitate the cells. 50 μL of culture supernatant was collected and, from this, 5 μL was tested in the human IL-2 AlphaLISA assay according to the manufacturer's protocol. These measurements were obtained on a PerkinElmer multi-label plate reader Envision. A standard curve of known IL-2 concentrations was generated to infer the pg / mL of IL-2 produced in the assay wells. All serial dilutions were tested in duplicate. The results were analyzed by using GraphPad Prism TM The software determines the EC of the antibody based on a four-parameter logistic equation on a 10-point dose-response curve. 50 value.

[0521] FACS-based cytotoxicity assay

[0522] FACS-based cytotoxicity assays were performed as described previously (Example 8).

[0523] Syngeneic Tumor Studies

[0524] MC38 / EV and MC38 / CD86 were cultured according to ATCC guidelines. 1x10 6MC38 / EV or MC38 / CD86 were implanted subcutaneously in C57BL / 6 mice. Mice were treated with intraperitoneal injection of 5 mg / kg of PD-1 antibody (RPM1-14, BioXcell) or rat IgG2a isotype control (BioXcell) on days 0, 3, 7, 10 and 14 after tumor implantation. Tumor size was measured twice a week 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 vertical dimension. Tumors larger than 2000mm were quantified by CO2 asphyxiation. 3 Or mice with ulcerated tumors were euthanized.

[0525] hCD3 / hCD28 / hPSMA humanized mice were generated as described in Example 10. Antibody treatment was also similar, except that 5 mg / kg of the indicated antibody or bispecific antibody was administered by intraperitoneal injection as monotherapy or combination therapy on days 0, 7, and 14 (prophylactic treatment) or on days 9, 13, and 22 (delayed treatment).

[0526] In vitro tissue cytokine analysis

[0527] On the 29th day after implantation, MC38 / hPSMA tumor-bearing CD3 / CD28 / PSMA mice were euthanized by carbon dioxide asphyxiation. Spleens and tumors were collected and stored in culture medium on ice. Unless otherwise stated, all the following steps were performed on ice or at 4°C. According to the manufacturing protocol (Miltenyi 130-096-730), the tumor was cut into small pieces using the Miltenyi mouse tumor dissociation kit, and the fragments were processed into a single cell suspension. The spleen was processed into a single cell suspension using a gentle MACS mechanical dissociation (spleen 4 program) and mashed through a 70 micron filter using the rubber end of a 3ml syringe. The cells were precipitated by centrifugation at 1200rpm for 5 minutes. Red blood cells were lysed by resuspending the cell pellet in 1ml ACK lysis buffer and incubating on ice for 5 minutes. The ACK lysis buffer was quenched with FACS buffer. The cells were precipitated by centrifugation at 1200rpm for 5 minutes. The cell suspension was resuspended in 1 ml of culture medium, and 0.2 ml of plates were inoculated in 96-well plates (20-400K tumor cells or 50-70K spleen cells). The cells were incubated overnight at 37 ° C, and the culture supernatant was collected. The cytokine levels in tissue culture supernatants were measured using the V-Plex pro-inflammatory MSD kit according to the manufacturer's protocol (Meso Scale Diagnostics K15048D-4). The number of cells inoculated per well was determined by FACS analysis. The cytokine levels were normalized to the number of cells inoculated on the plate. The calibration beads were run with the cells to accurately measure the number of cells using the following calculations:

[0528] # of cells = (# input beads x # cells counted by FACS) / # beads counted by FACS Measurement of mouse serum cytokine levels

[0529] Measurement of mouse serum cytokine levels was performed as previously described (Example 10).

[0530] Flow cytometric analysis

[0531] For flow cytometry analysis of in vivo experiments, tumors were harvested, single cell suspensions were prepared, and erythrocytes were lysed using ACK lysis buffer (ThermoFisher Scientific). Live / dead fixable blue dead cell staining kit (ThermoFisher Scientific) was used to identify live / dead cells. Samples were obtained on Symphony (BD Biosciences) and analyzed using Cytobank software (Cytobank in Santa Clara, California). The same number of events was used for each sample for analysis. The event range was determined by the sample with the least event obtained. In order to automatically cluster T cells based on specific markers, CITRUS analysis was used for Cytobank.

[0532] Toxicology study in cynomolgus monkeys

[0533] The cynomolgus monkey study was performed as previously described (Example 11).

[0534] result

[0535] PD-1 checkpoint inhibition increases the relative ratio of CD28 within the immune synapse, allowing the TSAxCD28 bispecific antibody to significantly enhance the ability of anti-PD-1 to promote T cell activation in vitro

[0536] To test whether co-stimulatory bispecific agonists can complement checkpoint inhibition, the ability of an exemplary PSMAxCD28 bispecific antibody (bs16429D) of the invention to enhance the effectiveness of PD-1 blockade in a TCR / CD3-dependent T cell activation manner was tested. In fact, effective T cell activation depends on the co-aggregation of TCR / CD3 and CD28 complexes at the "immune synapse" (IS). However, activation signals from TCR / CD3 and CD28 are directly inhibited by PD-1-Shp-2 phosphorylation, followed by PD-1 / PD-L1 accumulation at synapses (E. Hui et al., “T cell costimulatory receptor CD28 is a primary target for PD-1-mediated inhibition,” Science, 355, 1428-1433 (2017); JM Chemnitz, RV Parry, KE Nichols, CH June, JL Riley, “SHP-1and SHP-2associate with immunoreceptor tyrosine-based switch motif of programmed death 1upon primary human T cell stimulation, but only receptor ligationprevents T cell activation” To determine the relative localization of CD28 and PD-1 at the IS, an in vitro assay (Amnis Image Stream as described in Example 7) was developed using Jurkat T cells overexpressing PD-1 and Raji tumor target cells engineered to overexpress PD-L1. A fluorescently labeled bispecific CD20xCD3 antibody (EJ Smith et al., "A novel natural form of bispecific antibody that triggers T cell killing of B cells has robust activity in mouse tumor models and cynomolgus monkeys" Scientific Reports 5, 17943 (2015)) was used to replicate peptide MHC / TCR binding and visualize the interaction of T cells with target cells forming the 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 (NBPD-1 mAb) were used to simultaneously block the interaction with PD-L1 and visualize the localization of PD-1. It was found that in the absence of PD-L1 expression on target cells and in the presence of PD-1 mAbs, there was no change in the relative amount of PD-1 or CD28 at the IS (. Fig.14 ). However, in the presence of non-blocking PD-1 mAb, PD-L1 expression on target cells reduced CD28 at synapses and promoted high PD-1 accumulation at the IS. In contrast, in the presence of blocking PD-1 mAb, PD-1 localization was significantly reduced and relative CD28 levels were maintained at the IS. The distribution of PD-1 and CD28 was quantified by calculating the ratio of antibody staining inside the IS to that outside the IS ( Fig.15 ). This data suggests that PD-L1 expression on target cells enhances PD-1 localization at the IS while reducing CD28. In addition, PD-1 blocking mAb reduces the relative amount of PD-1 at the IS and rescues the level of CD28, thereby increasing the relative ratio of CD28 to PD-1.

[0537] We next tested whether PSMAxCD28 could enhance the effects of PD-1 blockade on tumor cell killing induced by T cell activation. To this end, the prostate cancer line 22RV1 was utilized. 22RV1 cells endogenously express PSMA and are engineered to express PD-L1 (22RV1 / PD-L1). Since there is no TCR / CD3 stimulation caused by an allogeneic T cell response, PSMAxCD3 (U.S. Patent No. 10,179,819) was used to provide the primary stimulation. In co-cultures of human peripheral blood mononuclear cells (PBMCs, containing human T cells) and 22RV1 / PD-L1 cells, PSMAxCD3 alone induced approximately 40% tumor cell killing ( Fig.16 , diamond, EC of 8E-10 50 ). Adding PD-1 mAb to PSMAxCD3 enhanced tumor cell killing to approximately 55% ( Fig.16, solid triangle, EC of 4E-10 50 Interestingly, the combination of PSMAxCD28 and PSMAxCD3 similarly enhanced the depth of tumor cell killing to approximately 55%, with increased potency ( Fig.16 , solid circle, EC of 6E-11 50 ), suggesting that PSMAxCD28 can override PD-1 / PD-L1-mediated inhibition. Interestingly, the triple combination of PSMAxCD28, PD-1 mAb, and PSMAxCD3 had the strongest tumor cell killing effect (approximately 70%), demonstrating the synergistic effect of this combination ( Fig.16 , solid squares, EC of 7E-11 50 As expected, neither PSMAxCD28, PD-1 mAb alone, nor the combination of PSMAxCD28 plus PD-1 mAb induced any tumor cell killing ( Fig.16 , open symbols). Consistently, the greatest increase in IFNγ release was observed with triple combination treatment ( Fig.17 ).

[0538] We next determined the effect of the PSMAxCD28 and PD-1 mAb combination on activation of primary human T cells in vitro. To replicate physiological PD-L1 expression and TCR / CD3 stimulation, a mixed lymphocyte reaction (MLR) was employed. In a one-way 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 ( Fig.18 In the presence of DU145 / PSMA cells and T cells, the PSMAxCD28 bispecific resulted in a dose-dependent increase in IL-2 release of approximately 3-4 fold over the IgG4 isotype control ( Fig.18 , circles). Similarly, the addition of 20 nM PD-1 mAb to the MLR assay also increased IL-2 release by approximately 3-4 fold compared to the IgG4 isotype control ( Fig.18 , triangles). In combination, PSMAxCD28 and 20 nM PD-1 mAb significantly enhanced activation induced by the PSMAxCD28 bispecific antibody, with maximal IL-2 levels increased by approximately 20-fold compared to isotype controls ( Fig.18 , squares), thus demonstrating that the combination of the PSMAxCD28 bispecific antibody and a PD-1 blocking mAb potently and synergistically activates T cells in the presence of tumor cells with endogenous levels of TCR / CD3 activation and PD-L1 inhibition.

[0539] Together, these results suggest that in the presence of TCR / CD3 signaling (driven by CD3 bispecific or alloreactivity), the PSMAxCD28 bispecific antibody can effectively enhance the ability of the PD-1 mAb to promote T cell activation, thereby increasing cytokine release and killing of tumor cells expressing PSMA and PD-L1 in vitro.

[0540] Overexpression of natural CD28 ligands on tumor cells synergizes with PD-1 mAb therapy to induce CD8 T cell-dependent durable antitumor immunity in vivo

[0541] To determine whether engagement of CD28 with one or more of its natural ligands could enhance the antitumor efficacy of PD-1 mAb in vivo, MC38 tumor cells were engineered to overexpress CD86, one of the co-stimulatory ligands of CD28 (data not shown). The combination of MC38 / CD86 cells and PD-1 mAb treatment significantly inhibited tumor growth compared to negative control MC38 cells transfected with an empty vector control (MC38 / EV). Fig.19A ), resulting in complete tumor regression with a robust survival benefit ( Fig.19B During the course of treatment, CD8 + Depletion of T cells completely abolished the antitumor efficacy elicited by combining PD-1 mAb therapy with MC38 / CD86 cells, suggesting that CD8 + T cells are dependent ( Fig.19C Notably, tumor-free mice initially implanted with MC38 / CD86 cells and treated with PD-1 mAb rejected a secondary MC38 parental tumor implanted more than 60 days after the primary tumor implantation, suggesting the presence of a T cell memory response ( Fig.19D ). Thus, these data suggest that the synergistic effect of constitutively expressing CD28 ligands with anti-PD-1 therapy can generate durable CD8-dependent antitumor immunity in vivo.

[0542] PSMAxCD28 synergizes with PD-1 mAb therapy to induce anti-tumor immunity in syngeneic tumor models

[0543] The above findings were then expanded to demonstrate the anti-tumor efficacy of treatment with TSAxCD28 bispecific antibodies alone or in combination with PD-1 mAb in syngeneic tumor models. Using the established C57BL6 syngeneic MC38 tumor model as described herein, the hPSMA gene (pLVX.EF1a.hPSMA) was genetically introduced into MC38 cells to produce the tumor-specific antigen MC38 / hPSMA as described herein. In order to avoid the possibility that mice would spontaneously reject these syngeneic tumors that otherwise express the introduced human tumor antigen, PSMA was genetically humanized in these mice. In addition, as previously described, the CD3γ-δ-ε and CD28 genes were also humanized using VelociGene technology (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," Nature Biotechnol 25, 91-99 (2007)), so that the bispecific antibody will recognize host T cells (via hCD3 or hCD28) as well as human tumor antigens (i.e., hPSMA) in normal tissues and tumors, which represents the actual clinical situation (Y. Kinoshita et al., "Expression of prostate-specific membrane antigen in normal and malignant human tissues" ofprostate-specific membrane antigen in normal and malignant human tissues” World J Surg 30, 628-636 (2006). In this model, the combination of PSMAxCD28 bispecific and PD-1 mAb provided the greatest tumor growth control, which translated into a robust survival benefit.

[0544] Immediate treatment with a combination of anti-CD28 x anti-PSMA and anti-PD1 synergistically inhibits tumor growth

[0545] In the immediate treatment study, 5 mg / kg anti-CD28x anti-PSMA, anti-PD1, or rat IgG2a isotype control was administered by intraperitoneal injection as monotherapy or combination therapy on days 0, 7, and 14. Tumor growth was monitored over time using caliper measurements of X and Y diameters. Tumor volume was calculated (X*Y*(X / 2)). When the tumor size was greater than 2000 mm 3 When , the mice were euthanized.

[0546] like Figures 20A to 20E As shown, anti-CD28 x anti-PSMA antibodies inhibited tumor growth and increased survival in mice when used alone, but also synergized with anti-PD-1 antibodies to promote tumor rejection and increase survival, even better than when used without PD-1 antibodies.

[0547] Treatment with a combination of CD28xPSMA and PD1 induces long-term antitumor immunity against a second tumor attack

[0548] Furthermore, tumor-free mice implanted and treated with the combination of PSMAxCD28 bispecific antibody and PD-1 mAb rejected a secondary MC38 parental tumor line implanted more than 60 days after primary tumor implantation, suggesting the generation of immune memory ( Fig.21A) is consistent with the observations described herein. These results suggest that endogenous antigen-specific TCR signaling (signal 1) is generated by peptide-MHC complexes on MC38 / PSMA-implanted tumor cells. It has been shown that MC38 tumor cells express high levels of reactivated endogenous retroviral proteins, such as p15E, and that intratumoral T cells in C57BL6 mice respond to this p15E antigen (JC Lang, D. Perry-Lalley, "The envelope protein of an endogenous murine retrovirus is a tumor-associated T-cell antigen for multiple murine tumors," Journal of Immunotherapy 23, 177-183 (2000); HJ Zeh, 3rd edition, D. Perry-Lalley, MEDudley, SA Rosenberg, JC Lang, "High-avidity CTLs for two self-antigens exhibit excellent anti-tumor efficacy in vitro and in vivo," Journal of Immunology 162, 989-994 (1999)). Consistent with this finding, PSMAxCD28 and PD-1 mAb combination therapy was found to induce peripheral T cells reactive to at least one endogenous antigen, P15E, thereby supporting the generation of antitumor immune memory ( Fig. 21B Notably, in similar experiments in which MC38 / hPSMA tumor rejection was induced using a combination of PSMAxCD28 and PSMAxCD3, these tumor-free mice were unable to reject secondary tumor rechallenge ( Fig. 21C ). In summary, this data demonstrates that in this PSMA syngeneic tumor model, the CD28 bispecific antibody synergizes with the PD-1 mAb and can enhance endogenous TCR / CD3-dependent T cell responses.

[0549] Therapeutic treatment using a combination of anti-CD28x anti-PSMA and anti-PD1 synergistically inhibits tumor growth

[0550] Similarly, in the delayed treatment regimen, the combination of PSMAxCD28 and PD-1 mAb suppressed the growth of established MC38 / hPSMA tumors 10 days after implantation ( Fig.22A ), and also resulted in a significant increase in survival ( Fig. 22C ) and tumor volume reduction ( Fig.22D Interestingly, tumor-targeted combination therapy selectively increased intratumoral cytokines, as shown by IFNγ ( Fig. 22BNo splenic or systemic cytokine induction was observed in the same mice ( Fig.23A and 23B In addition, profiling of T cell activation markers showed increased intratumoral PD-1 expression during PSMAxCD28 treatment, supporting this combination therapy ( Fig.24 To further characterize the responsive T cell subsets after combination therapy, tumor-infiltrating CD8 + T cells. CITRUS (Cluster Identification, Characterization, and Regression) was used to independently stratify statistically distinct T cell clusters. Importantly, PD-1 blockade was found to expand effector cells (CD44 高 CD62L 低 )CD8 + T cells (cluster C1) Fig.25 However, only the combination therapy was able to drive intratumoral CD8 + Expansion of T cells (cluster C2) 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) ( Fig.25 ) (M. Philip et al., "Chromatin states define tumour-specific T cell dysfunction and reprogramming", Nature 545, 452-456 (2017)). These data suggest that PSMAxCD28 bispecific antibody and anti-PD-1 combination therapy can drive robust anti-tumor immunity associated with activation of intratumoral T cells with a memory-like phenotype and drive survival benefits.

[0551] TSAxCD28 alone or in combination with PD-1 mAb therapy does not induce systemic T cell activation compared with CD28 superagonists in cynomolgus monkeys

[0552] Early clinical trial data showed that bivalent CD28-activating antibodies, known as "CD28 superagonists" (CD28-SA), extensively activated T cells and caused 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 anti-CD28 monoclonal antibody TGN1412" New England Journal of Medicine 355, 1018-1028 (2006)). In order to evaluate the tolerability of TSAxCD28 bispecific antibodies alone, or the potential for synergistic pharmacology when used in combination with PD-1 mAb, exploratory studies were conducted in genetically engineered triple humanized mice and cynomolgus monkeys. Three monkeys in each treatment group received a single dose (10 mg / kg) of PSMAxCD28 alone or in combination with PD-1 mAb (REGN2810) (10 mg / kg) by intravenous infusion (the combination group received continuous infusion) (Table 20, Figures 26A-26C and 27).

[0553] In addition, 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 assessments were based on clinical observations, qualitative food consumption, body weight, vital signs (temperature, heart rate, pulse oximetry, and respiratory rate), clinical and anatomical pathology (after completion of the experiment). Blood samples were collected for cytokine and FACS immunophenotyping. PSMAxCD28 alone or in combination with PD-1 was well tolerated, and all animals survived the study. No clinical observations related to the test article were observed (data not shown). No changes in organ weights were found, nor were any macroscopic changes found at the final necropsy (data not shown). In addition, no significant cytokine release, T cell marginalization, or activation was observed (Tables 23, Figures 26A-26C In contrast, significant cytokine release, lymphocyte marginalization, and T cell activation were observed in monkeys administered the CD28 “superagonist” alone ( Fig.29A and 29B ). Extensive infiltration of immune cells was observed in the kidneys, brain, and seminal vesicles of animals treated with CD28 superagonists. In contrast, no significant histological changes associated with treatment were observed in animals administered PSMAxCD28 alone or in combination with PD-1 (data not shown).

[0554] Table 23: No significant clinical observations or cytokine release observed with anti-PSMAxCD28 alone or in combination with anti-PD-1 in an exploratory single-dose monkey toxicology study

[0555]

[0556] *BLQ: below the limit of quantitation

[0557] Consistent with the above studies in cynomolgus monkeys, no cytokine elevation was observed in tumor-bearing or non-tumor-bearing naive triple humanized mice (hCD3 / hCD28 / hPSMA) dosed with PSMAxCD28 alone or in combination with PD-1 ( Fig. 27 and Fig.29A and 29B In contrast, administration of a CD28 superagonist induced a significant increase in IFNγ, TNFα, IL-2, IL-4, and IL-5 at 4 h post-dose ( Fig. 27 and Fig.29A and 29B ). Consistent with the above results, it has been previously shown that TSAxCD28 bispecific antibodies (as well as parental bivalent non-superagonist CD28 antibodies used to prepare these bispecific antibodies) are unable to induce human T cell proliferation in FDA-recommended in vitro dry and wet coating assays (R. Stebbings, D. Eastwood, S. Poole, R. Thorpe, "After TGN1412: recent developments in cytokine release assays" Journal of Immunotoxicology 10, 75-82 (2013)), compared to the strong proliferation induced by CD28-SA (data not shown). Overall, these data indicate that TSAxCD28 bispecific antibodies are well tolerated.

[0558] In addition, as previously described (Example 10, Fig.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 increased serum cytokine levels in tumor-bearing mice.

[0559] discuss

[0560] This article describes and validates novel tumor-targeted immunotherapy using TSAxCD28 bispecific antibodies in combination with PD-1 blocking mAbs, which induce long-term anti-tumor immunity and promote robust intratumoral T cell activation in animal tumor models. Toxicology studies in genetically humanized immunocompetent mice and cynomolgus monkeys showed that these bispecific antibodies exhibited no toxicity either by themselves or in combination with PD-1 mAbs, suggesting that this therapeutic approach could provide a well-tolerated "off-the-shelf" biologic solution with significantly enhanced, specific, and synergistic anti-tumor activity.

[0561] Checkpoint inhibition using PD-1 blocking mAbs is known to unleash disruption of T cell activation, but its efficacy as a single agent is often insufficient to achieve tumor clearance and durable antitumor responses in many cancers. Several approaches to improve response rates to PD-1 inhibition are currently being evaluated. In fact, biomarkers have been identified to predict responsiveness to PD-1 mAbs (R. Cristescu et al., “Pan-tumor genomic biomarkers for PD-1 checkpoint blockade-based immunotherapy”). =Initial immunotherapy (Science 362, (2018)), non-tumor-targeted combination therapies that use PD-1 inhibition with agonistic antibodies that trigger co-stimulatory receptors to improve T cell activation, or therapies that use them with chemotherapy or radiation therapy are currently undergoing preclinical and clinical testing (S. Hu-Lieskovan, A. Ribas, "New combination strategies using programmed cell death 1 / programmed cell death ligand 1 checkpoint inhibitors as a backbone," Journal of Cancer 23, 10-22 (2017); YK Chae et al., "The current status and future of dual blockade immunotherapy with anti-CTLA4 and PD-1 / PD-L1 in cancer: lessons from clinical trials in melanoma and non-small cell lung cancer (NSCLC)," Journal of Cancer Immunotherapy 6, 39 (2018); P. S. Chowdhury, K. Chamoto, T. Honjo, "Combination therapy strategies to improve the efficacy of PD-1 blockade: a new era in cancer immunotherapy," Journal of Internal Medicine 283, 110-120 (2018)). The challenge, however, is that many of these combinations are often based on the availability of existing drugs and post hoc justification for combination therapy rather than a truly hypothesis-driven approach, which in some cases has led to poorer 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)).Clearly, checkpoint inhibition and reactivation of the immune system offer the potential for long-term remission in many patients (JS Weber et al., “Nivolumab versus chemotherapy for patients with advanced melanoma following anti-CTLA-4 therapy (CheckMate 037): a randomized, controlled, open-label phase 3 trial” The Lancet Oncology 16, 375-384 (2015); SL ​​Topalian et al., “Survival, durable tumor responses, and long-term safety in patients with advanced melanoma receiving nivolumab” Journal of Clinical Oncology 32, 1020-1030 (2014); MA Postow, MK Callahan, JD Wolchok, “Immune checkpoint blockade in cancer therapy” Journal of Clinical Oncology 33, 1974-1982 (2015); MR Magden et al., “PD-1 blockade with cemiplimab in advanced cutaneous squamous cell carcinoma” New England Journal of Medicine 379, 341-351 (2018)), therefore, further methods to improve or enhance T cell activity to promote more durable responses are needed. Here, in order to improve the anti-tumor efficacy of PD-1 mAb, the concept of using TSAxCD28 bispecific antibodies to enhance T cell signaling and activation was introduced. In fact, this novel combination immunotherapy was validated using tumor targets (e.g., PSMA), and it was demonstrated that CD28 co-stimulatory bispecific antibodies synergized with PD-1 mAb not only resulted in robust T cell activation, but also provided durable anti-tumor responses without systemic toxicity. Therefore, this tumor-targeted combination therapy can provide considerable advantages over the previously described non-targeted approaches.The use of CD28 bispecific antibodies, which do not directly activate CD28 unless accumulated on the surface of tumor cells, offers the possibility of promoting co-stimulation only at the tumor site, thereby 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," New England Journal of Medicine 355, 1018-1028 (2006)), i.e., the combined use of CLTA-4 and PD-1 blockade (J. Larkin et al., "Nivolumab and ipilimumab combination therapy or monotherapy in untreated melanoma," New England Journal of Medicine 373, 23-3 4 (2015); DB Johnson et al., “Fulminant Myocarditis with Combination Immune Checkpoint Blockade,” New England Journal of Medicine 375, 1749-1755 (2016); MH Pollack et al., “Safety of Resuming Anti-PD-1 in Patients with Immune-Related Adverse Events (irAEs) During Combination Anti-CTLA-4 and Anti-PD1 in Metastatic Melanoma,” Annals of Oncology 29, 250-255 (2018)) or other co-stimulatory agonist bivalent antibodies (NH Segal et al., “Results from an Integrated Safety Analysis of Urelumab, an Agonist Anti-CD137 Monoclonal Antibody,” Clinical Cancer Research 23, 1929-1936 (2017)). Toxicology studies in genetically humanized immunocompetent mice and cynomolgus monkeys showed that these bispecific antibodies were non-toxic as single agents or in combination with PD-1 mAb. In syngeneic models, the safety profile and the enhanced anti-tumor efficacy of the anti-PSMAxCD28 bispecific antibodies of the present invention with PD-1 mAb indicate that this treatment paradigm is robust and has broader utility as a novel class of immunotherapy combinations.

[0562] In order to enhance T cell-mediated killing of tumor cells, tumor-targeted approaches are being developed (E. Dahlen, N. Veitonmaki, P. Norlen, "Bispecific antibodies in cancer immunotherapy" Ther Adv Vaccines Immunother 6, 3-17 (2018)). In fact, CD3-based bispecific antibodies represent an emerging class of antibodies that can effectively trigger T cell activation by connecting T cells to tumor cells and activating TCR / CD3 (EJ Smith et al., "A novel natural form of bispecific antibody that triggers T cell killing of B cells has robust activity in mouse tumor models and cynomolgus monkeys" Science and Technology Reports 5, 17943 (2015)), thereby mimicking normal "signal 1". However, despite the promising clinical efficacy of CD3 bispecific antibodies, they may be associated with cytokine release syndrome (CRS) due to direct T cell activation and lack of specificity for tumors only (SLMaude, D.Barrett, DTTeachey, SAGrupp, "Managing cytokine release syndrome associated with novel T cell-engaging therapies" Journal of Cancer 20, 119-122 (2014)). Here, it is demonstrated for the first time that TSAxCD28 bispecific antibody and PD-1 mAb combination therapy induces tumor-specific T cell activation associated with long-term memory responses in an immunocompetent mouse tumor model.TSAxCD28 bispecific antibodies have limited or no activity in the absence of "signal 1", and PD-1 blockade depends on the response of endogenous antigen-specific T cells 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); NARizvi et al., "Mutational landscape determines sensitivity to PD-1 blockade in non-small cell lung cancer", Science 348, 124-128 (2015); JM Mehnert et al., "Immune activation and response to pembrolizumab in POLE-mutant endometrial cancer". “Signal 1” provided by endogenous tumor antigens is therefore important for PSMAxCD28 and PD-1 mAb combination therapy. This is in contrast to CD3 bispecific antibodies that activate T cells independently of TCR specificity and therefore may not generate long-term tumor-specific immunity. In fact, it has been found that although PSMAxCD3 and PSMAxCD28 combination therapy induced a strong anti-tumor effect, it did not produce a stro...

Claims

1. An antibody or antigen-binding fragment thereof that binds to human CD28, wherein the antibody or antigen-binding fragment comprises three heavy chain complementary determining regions (HCDR1, HCDR2 and HCDR3) contained in a heavy chain variable region (HCVR), and three light chain complementary determining regions (LCDR1, LCDR2 and LCDR3) contained in a light chain variable region, wherein the HCDR1, HCDR2, HCDR3, LCDR1, LCDR2 and LCDR3 are respectively composed of: (a) SEQ ID NO: 12, 14, 16, 20, 22 and 24; or (b) SEQ ID NO: 28, 30, 32, 44, 46 and 48; or (c) SEQ ID NO: 60, 62, 64, 68, 70 and 72.

2. The antibody or antigen-binding fragment of claim 1, wherein HCDR1, HCDR2, HCDR3, LCDR1, LCDR2 and LCDR3 consist of SEQ ID NOs: 12, 14, 16, 20, 22 and 24, respectively.

3. The antibody or antigen-binding fragment of claim 1, wherein HCDR1, HCDR2, HCDR3, LCDR1, LCDR2 and LCDR3 consist of SEQ ID NOs: 28, 30, 32, 44, 46 and 48, respectively.

4. The antibody or antigen-binding fragment of claim 1, wherein HCDR1, HCDR2, HCDR3, LCDR1, LCDR2 and LCDR3 consist of SEQ ID NOs: 60, 62, 64, 68, 70 and 72, respectively.

5. The antibody or antigen-binding fragment of claim 2, wherein the HCVR consists of the amino acid sequence of SEQ ID NO: 10, and the LCVR consists of the amino acid sequence of SEQ ID NO:

18.

6. The antibody or antigen-binding fragment of claim 3, wherein the HCVR consists of the amino acid sequence of SEQ ID NO: 26, and the LCVR consists of the amino acid sequence of SEQ ID NO:

42.

7. The antibody or antigen-binding fragment of claim 4, wherein the HCVR consists of the amino acid sequence of SEQ ID NO: 58 and the LCVR consists of the amino acid sequence of SEQ ID NO:

66.

8. The antibody or antigen-binding fragment according to any one of claims 1 to 7, which is an antibody.

9. The antibody of claim 8, comprising a human IgG heavy chain constant region.

10. The antibody of claim 9, wherein the heavy chain constant region is of human IgG1 or human IgG4 isotype. The antibody according to claim 8 , which is a bispecific antibody.

12. A pharmaceutical composition comprising the antibody or antigen-binding fragment of any one of claims 1-7.

13. A nucleic acid molecule comprising a nucleotide sequence encoding the antibody or antigen-binding fragment of any one of claims 1 to 7, or A pair of nucleic acid molecules, comprising a nucleotide sequence encoding the HCVR of the antibody or antibody binding fragment of any one of claims 1 to 7 and a nucleotide sequence encoding the LCVR of the antibody or antibody binding fragment of any one of claims 1 to 7, respectively.

14. An expression vector comprising the nucleic acid molecule of claim 13, or A pair of expression vectors, each of which comprises the pair of nucleic acid molecules according to claim 13.

15. A host cell comprising the antibody or antigen-binding fragment of any one of claims 1-7, the nucleic acid molecule or a pair of nucleic acid molecules of claim 13, or the expression vector or a pair of expression vectors of claim 14.

16. A method for producing an anti-CD28 antibody or antigen-binding fragment, comprising culturing the host cell of claim 15 under conditions that allow production of the antibody or antigen-binding fragment, and recovering the produced antibody or antigen-binding fragment.

17. The method of claim 16, further comprising formulating the antibody or antigen-binding fragment into a pharmaceutical composition with a suitable carrier.

18. An anti-CD28 antibody or antigen-binding fragment obtainable by the method of claim 16.

Images