PD-L1xCD28 bispecific antibodies for immune checkpoint dependent t cell activation

Abstract

The present invention relates to PD-L1xCD28 bispecific antibodies that act as immune checkpoint inhibitors by binding and blocking PD-L1 (thus preventing it from binding to PD-1 expressed on T cells) and can further deliver a co-stimulatory signal to T cells by aggregating and competitively binding to CD28.

Description

[0001] CROSS-REFERENCE TO RELATED APPLICATIONS

[0002] This application claims the benefit of and priority to U.S. Provisional Application No. 63 / 418,264, filed on October 21, 2022, and U.S. Provisional Application No. 63 / 446,987, filed on February 20, 2023, each of which is incorporated herein by reference in its entirety. Technical Field

[0003] The present invention relates to fully human PD-L1xCD28 bispecific antibodies (bsAbs) in the form of κλ bodies that can block the PD-1 / PD-L1 interaction while providing co-stimulatory T cell signal 2.

[0004] Reference to electronic sequence listing

[0005] The contents of the electronic Sequence Listing (NOVI_052_001WO_SeqList_ST26.xml; size: 148,790 bytes; and creation date: October 12, 2023) are incorporated herein by reference in their entirety. Background Art

[0006] Over the past few years, new methods have been developed to stimulate the body's own immune cells to better attack and kill cancer cells. An example of a successful cancer immunotherapy is a monoclonal antibody that can block so-called immune checkpoints. Currently approved immune checkpoint inhibitors (ICIs) block CTLA-4 (e.g., ipilimumab, sold under the trade name Yervoy), PD-1 (e.g., pembrolizumab, sold under the trade name Keytruda; and cemiplizumab, sold under the trade name Libtayo), and PD-L1 (e.g., atezolizumab, sold under the trade name Tecentriq). Durable anti-tumor responses can be achieved in a variety of cancer types using ICIs. Unfortunately, responses are limited to subsets of patients, and many cancer types are known to be inherently resistant to ICI monotherapy.

[0007] Other approved cancer immunotherapies include T cell bispecific antibodies—which bridge T cells to target cells expressing tumor-associated antigens (TAAs) via the CD3 receptor on T cells—and chimeric antigen receptor (CAR) T cells. Although very good anti-tumor responses have been observed in T cell bispecific or CAR T cell-based therapies for hematological malignancies, these approaches have not yet achieved real breakthroughs in solid cancers, leaving many cancer patients without treatment options.

[0008] T cell co-stimulatory bispecific antibodies are a new class of therapeutic agents that can elicit anti-tumor responses, especially when combined with T cell bispecific antibodies or immune checkpoint inhibitors (ICIs). Preclinical studies have demonstrated the benefits of adding co-stimulatory CD28 bsAbs for the treatment of solid tumors, namely enhancing the efficacy of bispecific T cell engagers (Correnti et al., 2018; Skokos et al., 2020) or PD-(L)1 checkpoint inhibitors (Waite et al., 2020). They work by providing co-stimulatory signals to T cells within the tumor microenvironment. The specificity of CD28 co-stimulatory bsAbs is provided by a targeting anti-tumor associated antigen (TAA) arm that is paired with a so-called effector arm, specifically an agonist anti-CD28 arm. To date, several TAAxCD28bsAbs have been described (Correnti et al., 2018; and WO2019246514, WO2020132066, WO2020198009, WO2020127618, WO2020132024, WO2021259890, and WO2022040482), some of which are actively tested in early clinical trials (ClinicalTrials.gov identifiers NCT03972657, NCT04590326, NCT04626635, NCT05219513, or NCT05585034).

[0009] To further enhance the antitumor activity of CD28 bsAbs, the anti-TAA targeting arm can be replaced by an antibody arm with intrinsic therapeutic activity. Figure 1 As shown, there are several advantages to using an anti-PD-L1 arm because: (1) monovalent blockade of PD-L1 still prevents PD-1 engagement on T cells; (2) PD-L1 expressed on tumor cells, immune cells, or stromal cells inhibits T cells through PD-1 engagement, which negatively regulates CD28 co-stimulatory signals; and (3) PD-L1 expressed on APCs sequesters CD80 and prevents it from co-stimulating CD28 on T cells.

[0010] This type of PD-L1xCD28 bsAb can (1) bridge PD-L1 + (1) prevent PD-L1 from APCs from interacting with PD-1 on T cells in the tumor microenvironment and in draining lymph nodes; (2) prevent PD-L1 from DCs from sequestering CD80, thereby facilitating CD80 / CD28 interactions; and (3) prevent PD-L1 on stromal cells and immune cells from interacting with PD-1 on T cells.

[0011] Therefore, new antibodies and therapeutic agents that can dually target CD28 and PD-L1 are needed. Summary of the Invention

[0012] The present invention provides immune checkpoint-driven costimulatory bispecific antibodies. In some aspects, the antibody is a bispecific antibody having a first antigen-binding domain that binds to PD-L1 and a second binding domain that binds to CD28, referred to herein as a PD-L1xCD28 bsAb.

[0013] The PD-L1xCD28 bsAbs have a common heavy chain having: a complementarity determining region 1 (CDR1) comprising the amino acid sequence of (SEQ ID NO: 6); a complementarity determining region 2 (CDR2) comprising the amino acid sequence of (SEQ ID NO: 7); and a complementarity determining region 3 (CDR3) comprising the amino acid sequence of (SEQ ID NO: 8).

[0014] In some aspects, the PD-L1xCD28 bsAb has a first light chain variable region comprising: a CDR1 comprising the amino acid sequence of SEQ ID NO: 13, a CDR2 comprising the amino acid sequence of SEQ ID NO: 14, and a CDR3 comprising the amino acid sequence of SEQ ID NO: 15 [S79];

[0015] In some aspects, the PD-L1xCD28 bsAb has a second light chain variable region comprising: a CDR1 comprising the amino acid sequence of SEQ ID NO: 18, a CDR2 comprising the amino acid sequence of SEQ ID NO: 19, and a CDR3 comprising the amino acid sequence of SEQ ID NO: 20 [AI3]; or a CDR1 comprising the amino acid sequence of SEQ ID NO: 23, a CDR2 comprising the amino acid sequence of SEQ ID NO: 24, and a CDR3 comprising the amino acid sequence of SEQ ID NO: 25 [AI13];

[0016] In some aspects, the first heavy chain variable region and the second heavy chain variable region comprise the amino acid sequence of SEQ ID NO: 10. In some aspects, the first heavy chain and the second heavy chain comprise the amino acid sequence of SEQ ID NO: 11 or SEQ ID NO: 12.

[0017] In some aspects, the first light chain variable region is SEQ ID NO:16.

[0018] In other aspects, the first light chain has the amino acid sequence of SEQ ID NO:17.

[0019] In some aspects, the bispecific antibody has a second light chain variable region of SEQ ID NO:21; SEQ ID NO:26.

[0020] In other aspects, the bispecific antibody has a second light chain of ID NO:22; SEQ ID NO:27.

[0021] The present invention also includes compositions of any of the bispecific antibodies disclosed herein. Optionally, the composition further comprises a CD3xCEA bispecific antibody having two identical heavy chains comprising the amino acid sequence of SEQ ID NO: 3, a first light chain having the amino acid sequence of SEQ ID NO: 4, and a second light chain having the amino acid sequence of SEQ ID NO: 5.

[0022] In some aspects, the bispecific antibody has a first light chain that is a kappa light chain and a second light chain that is a lambda light chain.

[0023] In another aspect, a bispecific antibody has a first light chain that is a lambda light chain and a second light chain that is a kappa light chain.

[0024] In some embodiments, a portion of the first light chain is of the κ type, and at least a portion of the second light chain is of the λ type. In some embodiments, the first light chain comprises at least a κ constant region. In some embodiments, the first light chain further comprises a κ variable region. In some embodiments, the first light chain further comprises a λ variable region.

[0025] In some embodiments, the second light chain comprises at least a lambda constant region. In some embodiments, the second light chain further comprises a lambda variable region. In some embodiments, the second light chain further comprises a kappa variable region.

[0026] In some embodiments, the first light chain comprises a kappa constant region and a kappa variable region, and wherein the second light chain comprises a lambda constant region and a lambda variable region.

[0027] Optionally, the bispecific antibody has an Fc domain comprising one or more amino acid substitutions that reduce binding to an activating Fc receptor and / or reduce effector function.

[0028] For example, the bispecific antibody has L234A and L235A substitutions. Additionally, the bispecific antibody has P329A, P329G, or P329R substitutions.

[0029] Bispecific antibodies have an IgG isotype. Bispecific antibodies are human antibodies. Bispecific antibodies can achieve PD-L1-dependent T cell activation. In some embodiments, the immunostimulation of the bispecific antibody occurs within and / or at the tumor. In some embodiments, the immunostimulation of the bispecific antibody occurs outside the tumor. In some embodiments, the immunostimulation of the bispecific antibody occurs in lymphoid organs or systems.

[0030] The present disclosure provides a composition comprising the bispecific antibody described herein and a pharmaceutically acceptable carrier.

[0031] The present disclosure provides a method of reducing tumor cell proliferation and / or enhancing tumor cell killing, the method comprising contacting cells with a composition comprising a bispecific antibody described herein. The present disclosure provides a method of treating cancer in a subject, the method comprising administering to the subject a composition comprising a bispecific antibody described herein.

[0032] In some aspects, the present disclosure provides for the use of compositions comprising the bispecific antibodies described herein for treating, preventing, or delaying the progression of a lesion. In some embodiments, the lesion is cancer. In some embodiments, the cancer is a solid tumor. In some embodiments, the solid tumor is or is derived from breast cancer, ovarian cancer, head and neck cancer, bladder cancer, melanoma, mesothelioma, colorectal cancer, bile duct cancer, pancreatic cancer, lung cancer, leiomyoma, leiomyosarcoma, renal cancer, glioma, glioblastoma, endometrial cancer, esophageal cancer, biliary tract cancer, gastric cancer, prostate cancer, or a combination thereof.

[0033] The present invention also includes an antibody having an antigen binding domain that binds to CD28; wherein the antigen binding domain has: a heavy chain variable region comprising: a complementarity determining region 1 (CDR1) comprising the amino acid sequence of (SEQ ID NO:6); a complementarity determining region 2 (CDR2) comprising the amino acid sequence of (SEQ ID NO:7); and a complementarity determining region 3 (CDR3) comprising the amino acid sequence of (SEQ ID NO:8); and a light chain variable region comprising: a CDR1 comprising the amino acid sequence of SEQ ID NO:18, a CDR2 comprising the amino acid sequence of SEQ ID NO:19; and a CDR3 comprising the amino acid sequence of SEQ ID NO:20; [AI3]; or a CDR1 comprising the amino acid sequence of SEQ ID NO:23, a CDR2 comprising the amino acid sequence of SEQ ID NO:24, and a CDR3 comprising the amino acid sequence of SEQ ID NO:25; [AI13].

[0034] Antibodies are F(ab) fragments, F(ab')2 fragments, and Fv fragments or single-chain Fv fragments. Antibodies are monospecific and monovalent.

[0035] 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. Although methods and materials similar or equivalent to those described herein can be used to practice the present invention, suitable methods and materials are described below. All publications, patent applications, patents and other references mentioned herein are expressly incorporated by reference in their entirety. In the event of a conflict, this specification (including definitions) shall prevail. In addition, the materials, methods and embodiments described herein are merely illustrative and are not intended to be limiting.

[0036] Other features and advantages of the invention will be apparent from and encompassed by the following detailed description and claims. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1 .Mechanism of action achieved by PD-L1xCD28 bsAb.

[0038] Figure 2A-2C Concentration-dependent binding of AI3S79 / N to PD-L1-expressing HPAC cells (preactivated with IFNγ) (A), CD28-expressing Jurkat cells (B), and CD28 and PD-L1 double-negative TIB153 cells (C).

[0039] Figure 3 Concentration-dependent blockade of the PD-1 / PD-L1 interaction, as measured by reporter cell bioassays, indicates that stronger blockade of the PD-1 / PD-L1 axis results in higher luminescence in the reporter cells. AI3S79 / N blocks the PD-1 / PD-L1 axis and, due to its agonist CD28 arm, also enhances the reporter system, resulting in higher RLUs than simple PD-L1 blockers.

[0040] Figure 4 T cell-dependent cytotoxicity of HPAC cells induced by AI3S79 / N when combined with the CEAxCD3 bsAb. Effector cells (healthy donor PBMCs) were co-cultured with PD-L1 / CEA double-positive HPAC cells for 6 days at varying E:T ratios (10:1, 3:1, 1:1, and 1:3 from left to right). AI3S79 / N synergizes with CEAxCD3 to kill HPAC cells. In the absence of signal 1 (no CEAxCD3 bsAb added), no killing was induced. A dose range of CEAxCD3 was observed at a constant CD28 bsAb concentration (2.5 μg / mL).

[0041] Figure 5 .exist Figure 4 Upregulation of the T cell activation marker CD25 on human CD4+ and CD8+ T cells (top and bottom rows, respectively) harvested at the end of the TDCC experiment is shown. The combination of CEAxCD3 and PD-L1xCD28 bsAbs induces stronger CD4+ and CD8+ T cell activation than CEAxCD3 monotherapy.

[0042] Figure 6 .exist Figure 4 Proliferation of human CD4+ and CD8+ T cells (top and bottom rows, respectively) harvested at the end of the TDCC experiment is shown. The combination of CEAxCD3 and PD-L1xCD28 bsAbs potently increased the percentage of proliferating CD4+ and CD8+ T cells compared to CEAxCD3 monotherapy.

[0043] Figures 7A-7F In vivo efficacy study of AI3S79 / N in the subcutaneous syngeneic colorectal cancer model MC38-hPD-L1 in CD28 HuGEMM mice. (A) Experimental design of the in vivo efficacy study. (B) Mean tumor volume of the different treatment groups. (C) Comparison of tumor volume at study termination (day 28 after treatment initiation). For tumors growing over 3000 mm 3 For all tumors, the tumor volume was set to 3000 mm 3 Tumor growth in individual mice (D, E, and F) in the vehicle, atezolizumab, or AI3S79 / N groups, respectively. In contrast to atezolizumab, which only delayed tumor growth on average, AI3S79 / N as a single agent resulted in tumor regression in all mice, with 6 of 10 mice being marked as tumor-free at study termination.

[0044] Figure 8A-8B (A) According to Figure 7A The experimental treatment regimen was used to calculate the weight of mice treated with AI3S79 / N (adjusted for tumor). (B) Figure 7A Cytokine secretion in mice treated with AI3S79 / N 6 hours after the first injection following the experimental treatment regimen. AI3S79 / N was well tolerated in immunocompetent huCD28 mice.

[0045] Figure 9 Mean tumor volume of mice that were re-challenged. Figure 7A Surviving mice from the indicated experiments were re-challenged with tumors. Re-challenged = surviving mice (n=8), untreated. Naive mice (n=8), untreated. AI3S79 / N induces immune memory against MC38.

[0046] Figures 10A-10B IL-2 release from fresh PBMCs isolated from three donors or PBMCs pre-cultured at high density after treatment with TGN1412 (A) or AI3S79 / N (B). In contrast to TGN1412, AI3S79 / N is not a superagonist.

[0047] Figures 11A-11B (A) Protocol of the established T cell and DC mixed lymphocyte reaction (MLR). (B) IL-2 release induced by PD-(L)1 blocking antibodies, including AI3S79 / N and anti-PD-(L)1 blocking antibodies such as nivolumab and atezolizumab.

[0048] Figure 12 Body weight (as a percentage of day 0) of irradiated and PBMC-transplanted mice treated with TGN1412 or dose-response AI3S79 / N. AI3S79 / N is well tolerated in a mouse model sensitive to CD28-mediated CRS.

[0049] Figure 13 Proliferation of CD4+ and CD8+ T cells cultured on plates coated using either the wet or dry coating procedure. Each point represents the mean proliferation obtained using healthy PBMC donors. In contrast to TGN1412, AI3S79 / N is not a superagonist.

[0050] Figures 14A-14B (A) Killing of NLV-loaded tumor cells and (B) corresponding expansion of NLV-specific T cells induced by AI3S79 / N dose response (right panel). Two donors are shown. AI3S79 / N induces killing of NLV-loaded tumor cells and expansion of NLV-specific T cells in a dose-dependent manner.

[0051] Figures 15A-15C In vivo efficacy study of the combination of AI3S79 / N and CEAxCD3 TCE in treating HPAC tumors subcutaneously implanted in PBMC-humanized NOG mice. (A) Experimental design for the in vivo efficacy study. (B) Mean tumor volume across treatment groups. (C) Tumor growth in individual mice. Because two PBMC donors (D415 and D417) were used for humanization, mice are specifically identified as humanized with D415 (dashed line) or D417 (solid line). AI3S79 / N synergistically induces TV regression or arrest with TCE.

[0052] Figures 16A-16BIn vivo efficacy study of the combination of AI3S79 / N and CEAxCD3 TCE in treating HPAF-II subcutaneous tumors in fully humanized BRGSF-HIS mice. (A) Mean tumor volume across treatment groups. (B) Tumor growth in individual mice. Asterisks (*) represent animals euthanized due to tumor ulceration (rather than tumor volume endpoints). AI3S79 / N synergizes with TCE to reduce tumor progression.

[0053] Figure 17 AI3S79 / N concentrations after single (0.5 mg / kg or 10 mg / kg) and repeated (2 x 10 mg / kg) IV administration in monkeys (semi-logarithmic scale). The black dashed line indicates the LLOQ value (0.02 μg / mL).

[0054] Figure 18 Serum cytokine concentrations in cynomolgus monkeys after single or repeated (IV) administration of AI3S79 / N. AI3S79 / N administration resulted in only a mild and transient release of IL-6.

[0055] Figure 19 CRP levels in animals administered AI3S79 / N by single or repeated injections (n=2) AI3S79 / N treatment induced a mild and transient release of CRP. DETAILED DESCRIPTION

[0056] The present invention is based on bispecific antibodies (bsAbs) capable of immune checkpoint-dependent T cell activation and tumor cell killing. Specifically, the present invention is based on bsAbs co-engaging with immune checkpoints such as PD-L1, which are expressed on the surface of tumor cells to mediate CD28 aggregation, and thus PD-L1-mediated T cell activation.

[0057] The bsAbs of the present invention are characterized by a single agonist CD28 antigen-binding domain for monovalent co-stimulation of CD28, and a second antigen-binding domain capable of specifically and monovalently binding to PD-L1 (which prevents PD-L1 from engaging PD-1 expressed on T cells).

[0058] CD28

[0059] CD28 is a key co-stimulatory receptor expressed on the surface of T lymphocytes. It belongs to a subfamily of co-stimulatory molecules characterized by an extracellular variable immunoglobulin-like domain. Other members of the molecule family include CTLA-4, ICOS, PD-1, and BTLA.

[0060] In humans, CD28 is expressed as a disulfide-linked homodimer at the cell surface of T lymphocytes and is present in approximately 80% of human CD4+ T cells and 50% of CD8+ T cells.

[0061] Despite the lack of intrinsic enzymatic activity, binding of CD28 to its ligand results in specific phosphorylation and transcriptional signaling, ultimately leading to metabolic changes and the production of key cytokines, chemokines, and survival signals that are crucial for the long-term expansion and differentiation of T cells.

[0062] The main ligands for CD28 are CD80 (B7.1) and CD86 (B7.2), which are primarily expressed on the surface of professional antigen-presenting cells (APCs). CD80 and CD86 differ in their expression patterns, multimeric states, and functions. Because CD28 and CTLA-4 are highly homologous, they compete for the same ligands. However, since CTLA-4 binds to these ligands with a higher affinity than CD28, CTLA-4 competes with CD28 for ligands and ultimately inhibits T cell responses.

[0063] Several anti-CD28 monoclonal antibodies have been proposed for therapeutic targeting of CD28. A subset of identified anti-CD28 antibodies, termed superagonist (SA) antibodies, was found to induce full activation of primary resting T cells through CD28 aggregation at the T cell surface even in the absence of TCR ligation (signal 1). However, the first human study of one of these SA anti-CD28 antibodies, TGN1412, produced severe inflammatory responses and chronic organ failure in all healthy volunteers treated. These adverse events were attributed to a cytokine storm, which was not predicted by both in vivo and in vitro preclinical safety studies.

[0064] To minimize the side effects of anti-CD28 monoclonal antibodies, Holliger et al. proposed B7 x anti-tumor-associated antigen (TAA) fusion proteins (Holliger et al., 1999). These fusion proteins were found to be as effective as anti-CD28 monoclonal antibodies but with greater specificity, i.e., the B7 x anti-CEA bispecific fusion protein could activate T cells only in the presence of cells expressing CEA.

[0065] Based on the same principle, CD28 bsAbs cannot aggregate CD28 on the surface of T cells alone, but need to engage a second target on the surface of another cell. Therefore, CD28 bispecific antibodies cannot costimulate T cells alone.

[0066] In the context of the present invention, an agonist anti-CD28 binding domain is paired with an anti-PD-L1 binding domain, resulting in a molecule capable of bridging T cells to cells expressing PD-L1.

[0067] Furthermore, even in the presence of PD-L1-positive cells that allow CD28 to accumulate on the T cell surface, the full cytotoxic potential of T cells can only be unleashed in the presence of primary T cell stimulation via the TCR. This is in contrast to the bivalent superagonist CD28 monoclonal antibody described above.

[0068] Preclinical studies have demonstrated the benefits of adding a co-stimulatory tumor-associated antigen xCD28 (TAAxCD28) bsAb to treat solid tumors, enhancing the efficacy of bispecific T cell engagers or PD-(L)1 checkpoint inhibitors. Examples of agonist TAAxCD28 bsAbs are described in WO2019246514, WO2020198009, WO2020132066, WO2020132024, WO2020127618, WO2021259890, WO2021155071, and WO22040482, some of which are currently being tested in clinical trials (ClinicalTrials.gov identifiers: NCT04590326, NCT03972657, NCT04626635, NCT05219513, NCT05585034).

[0069] PD-L1

[0070] Programmed cell death ligand-1 (PD-L1), also known as B7-H1 and CD274, is a transmembrane protein that is constitutively expressed on hematopoietic and non-hematopoietic healthy tissues. It can also be expressed on tumor cells and tumor stroma. In cancer, expression of the inhibitory receptor PD-1 is considered a hallmark of exhausted T cells, which exhibit a dysfunctional phenotype due to persistent antigenic and inflammatory stimulation. In addition, it has been shown that upregulation of PD-L1 in the tumor microenvironment allows tumors to evade the host immune system by interacting with PD-1 on T cells. Multiple studies have reported that PD-L1 is expressed in a variety of tumor tissues, on tumor cells or immune-infiltrating cells, or on both. In patients, blocking the interaction of PD-1 with PD-L1 using monoclonal antibodies has been shown to be a successful therapy for a range of cancer indications and is widely believed to enhance anti-tumor T cell responses by reversing or preventing the onset of T cell exhaustion and also by promoting the expansion of T cells during T cell priming in tumor-draining lymph nodes. However, despite significant improvements in patient outcomes achieved with PD-1 / PD-L1 checkpoint inhibitors, durable responses to these therapies are observed in only a minority of patients, and intrinsic or acquired resistance is common.

[0071] Bispecific antibodies

[0072] The bsAb antibodies according to the present invention can be generated de novo or can be engineered from existing monospecific CD28 and PD-L1 antibodies.

[0073] The bsAbs of the present invention can be based on any of the different antibody formats previously described. Generally speaking, IgG-like formats are preferred because they provide advantageous properties, such as long half-life and potentially reduced immunogenicity, but any other molecular bispecific format can also be used in the present invention.

[0074] The heavy and light chain amino acid sequences of the antibodies are identified by their United States Adopted Drug Names (USANs available, for example, through the American Medical Association at https: / / www.ama-assn.org / or via CAS Registry Numbers).

[0075] Monospecific CD28 and PD-L1 binding variable domains can be selected from, for example, phage display libraries, wherein the phage is engineered to express human immunoglobulins or portions thereof such as Fab, single-chain variable fragments (scFv), or unpaired or paired antibody variable regions, and subsequently engineered into a bispecific format. The CD28 and PD-L1 variable domains can be isolated, for example, from phage display libraries that express antibody heavy and light chain variable regions as fusion proteins fused to the capsid protein pill at the surface of phage M13.

[0076] The antibody library was screened for antibodies that bind to CD28 and PD-L1, and the positive clones obtained were further characterized. Such phage display methods for isolating human antibodies are well known in the art. See, for example, U.S. Patent Nos. 5,223,409; 5,403,484; and 5,571,698, 5,427,908, 5,580,717, 5,969,108, 6,172,197, 5,885,793; 6,521,404; 6,544,731; 6,555,313; 6,582,915 and 6,593,081. The de novo variable regions obtained were engineered into bispecific formats using methods known in the art and described herein.

[0077] In addition, the bispecific antibodies of the present invention can be prepared using techniques including those disclosed in WO 2012 / 023053, filed on August 16, 2011, the contents of which are hereby incorporated by reference in their entirety. The methods described in WO 2012 / 023053 produce bispecific antibodies that are structurally identical to human immunoglobulins. This type of molecule consists of two copies of a unique heavy chain polypeptide: a first light chain variable region fused to a constant kappa domain, and a second light chain variable region fused to a constant lambda domain. Each binding site exhibits a different antigenic specificity, which is determined by both the heavy and light chains. The light chain variable region can be of the lambda or kappa family and is preferably fused to the lambda and kappa constant domains, respectively. This is preferred to avoid creating non-natural polypeptide linkages.

[0078] However, it is also possible to obtain bispecific antibodies of the present invention by fusing a kappa light chain variable domain to a constant lambda domain to obtain a first specificity, and fusing a lambda light chain variable domain to a constant kappa domain to obtain a second specificity. The bispecific antibodies described in WO 2012 / 023053 are referred to as IgG kappa lambda antibodies or "kappa lambda bodies," a new fully human bispecific IgG format. This kappa lambda body format allows affinity purification of bispecific antibodies that are indistinguishable from standard IgG molecules whose characteristics are indistinguishable from those of standard monoclonal antibodies, and is therefore advantageous compared to previous formats.

[0079] In addition to the above methods, the bispecific antibodies of the present invention can be produced in vitro in a cell-free environment by introducing asymmetric mutations in the CH3 regions of two monospecific homodimeric antibodies according to the methods described in International Patent Publication No. WO2011 / 131746, and forming a bispecific heterodimeric antibody from the two parent monospecific homodimeric antibodies under reducing conditions that allow disulfide bond isomerization. In these methods, the first monospecific bivalent antibody and the second monospecific bivalent antibody are engineered to have certain substitutions at the CH3 domain that promote heterodimer stability; the antibodies are incubated together under reducing conditions sufficient to allow cysteines in the hinge region to undergo disulfide bond isomerization; thereby generating a bispecific antibody by Fab arm exchange.

[0080] The antibodies of the present invention have two or more antigen binding domains and are bispecific. The bispecific antibodies of the present invention include antibodies having a full-length antibody structure or a partial-length antibody structure such as Fab.

[0081] As used herein, "full-length antibody" refers to an antibody having two full-length antibody heavy chains and two full-length antibody light chains. A full-length antibody heavy chain (HC) is composed of the well-known heavy chain variable domain and constant domains VH, CH1, CH2, and CH3. A full-length antibody light chain (LC) is composed of the well-known light chain variable domain and constant domains VL and CL. A full-length antibody may lack a C-terminal lysine (K) in one or both heavy chains.

[0082] The term "Fab arm" or "half molecule" refers to a heavy chain-light chain pair that specifically binds to an antigen.

[0083] The full-length bispecific antibodies of the present invention can be produced, for example, by introducing substitutions at the heavy chain CH3 interface in each half molecule in vitro to favor heterodimer formation of two antibody half molecules with different specificities, or by co-expression, using Fab arm exchange (or half-molecule exchange) between two monospecific bivalent antibodies. The Fab arm exchange reaction is the result of a disulfide bond isomerization reaction and dissociation-association of the CH3 domains. The heavy chain disulfide bonds in the hinge region of the parent monospecific antibodies are reduced. The resulting free cysteine residues of one parent monospecific antibody form inter-heavy chain disulfide bonds with cysteine residues of the second parent monospecific antibody molecule, while the CH3 domains of the parent antibodies are released and reformed by dissociation-association. The CH3 domains of the Fab arms can be engineered to favor heterodimerization over homodimerization. The resulting product is a bispecific antibody having two Fab arms or half molecules, each of which binds to a different epitope.

[0084] As used herein, "homodimerization" refers to the interaction of two heavy chains having the same CH3 amino acid sequence. As used herein, "homodimer" refers to an antibody having two heavy chains having the same CH3 amino acid sequence.

[0085] As used herein, "heterodimerization" refers to the interaction of two heavy chains with different CH3 amino acid sequences. As used herein, "heterodimer" refers to an antibody with two heavy chains comprising different CH3 amino acid sequences.

[0086] A "knob-in-hole" strategy (see, e.g., PCT International Publication WO 2006 / 028936) can be used to generate full-length bispecific antibodies. Briefly, selected amino acids that form the interface of the CH3 domains in human IgG can be mutated at positions that affect the interaction of the CH3 domains to promote heterodimer formation. Amino acids with small side chains (holes) are introduced into the heavy chain of an antibody that specifically binds to a first antigen, and amino acids with large side chains (knobs) are introduced into the heavy chain of an antibody that specifically binds to a second antigen. Upon co-expression of the two antibodies, heterodimers are formed due to the preferential interaction of the heavy chain with the "hole" with the heavy chain with the "knob." Exemplary CH3 substitution pairs that form a knob and hole (represented as modified position in the first CH3 domain of the first heavy chain / modified position in the second CH3 domain of the second heavy chain): T366Y / F405A, T366W / F405W, F405W / Y407A, T394W / Y407T, T394S / Y407A, T366W / T394S, F405W / T394S, and T366W / T366S_L368A_Y407V.

[0087] Other strategies can be used, such as using electrostatic interactions to promote heavy chain heterodimerization by replacing positively charged residues at one CH3 surface and negatively charged residues at the second CH3 surface, as described in US Patent Publication Nos. 2010 / 0015133, 2009 / 0182127, 2010 / 028637, or 2011 / 0123532. In other strategies, heterodimerization can be promoted by the following substitutions (represented by the modified position in the first CH3 domain of the first heavy chain / the modified position in the second CH3 domain of the second heavy chain): L351Y_F405A_Y407V / T394W, T366I_K392M_T394W / F405A_Y407V, T366L_K392M_T394W / F405A_Y407V, L351Y _Y407A / T366A_K409F, L351Y_Y407A / T366V_K409F, Y407A / T366A_K409F or T350V_L351Y_F405A_Y407V / T350V_T366L_K392L_T394W, as described in U.S. Patent Publication No. US2012 / 0149876 or U.S. Patent Publication No. US2013 / 0195849.

[0088] Exemplary anti-cell surface antibodies that can be used to engineer bispecific molecules include, for example, anti-tumor associated antigen antibodies known in the art, such as pertuzumab and trastuzumab (HER-2); cetuximab, nexitozumab, panitumumab, and ervantumab (EGFR); labetuzumab and cerbituximab (CEA); ameliorate (mesothelin); cotrastuzumab (glypican 3); atezolizumab, avelumab, and durvalumab (PD-L1); blinatumomab (CD19); brentuximab (CD30); daratumumab (CD38); gemtuzumab (CD33); tositumomab (CD22), or obinutuzumab, ocrelizumab, ofatumumab, rituximab, and ibritumomab tiuxetan (CD20).

[0089] Exemplary bispecific antibodies that bind CD28 and PD-L1

[0090] The bispecific antibodies of the present invention have one antigen-binding region specific for PD-L1 and a second antigen-binding region specific for CD28. In other words, the bispecific antibodies are monovalent for PD-L1 and CD28. The bispecific antibodies share a common heavy chain.

[0091] In some embodiments, the heavy chain is a native heavy chain (i.e., without any mutations). In some embodiments, the heavy chain comprises a mutation relative to a native heavy chain. In some embodiments, the heavy chain is an IgG14 type containing different mutations to minimize effector function. Optionally, the bispecific antibody has different types of light chains. For example, one light chain is a kappa light chain, and the other light chain is a lambda light chain (i.e., k1-body). Different light chains allow the bispecific antibody to be easily purified using kappa and lambda selection resins.

[0092] Exemplary PD-L1 antibodies from which the PD-L1 antigen-binding region can be derived include S8 antibody, S9 antibody, S37 antibody, S14 antibody, S15 antibody, S17 antibody, S57 antibody, S58 antibody, S28 antibody, S30 antibody, S94 antibody, S23 antibody, S46 antibody, S71 antibody, and S79 antibody. Exemplary CD28 antibodies from which the CD28 antigen-binding region can be derived include AI3 antibody, AI13 antibody, AI5 antibody, AI7 antibody, AI8 antibody, AI9 antibody, AI10 antibody, AI11 antibody, AI12 antibody, AI14 antibody, AI15 antibody, AI16 antibody, AI17 antibody, AI18 antibody, AI19 antibody, AI20 antibody, AI21 antibody, AI22 antibody, and AI23 antibody. Thus, reference to an antibody may be described as, for example, "S79xAI3" or "AI3xS79" or "AI3S79" or "S79AI3" to identify the first and second antigen binding domains.

[0093] In some embodiments, the S79×AI3 bispecific antibody has a heavy chain comprising: a complementarity determining region 1 (CDR1) comprising the amino acid sequence of (SEQ ID NO:6); a complementarity determining region 2 (CDR2) comprising the amino acid sequence of (SEQ ID NO:7); and a complementarity determining region 3 (CDR3) comprising the amino acid sequence of (SEQ ID NO:8), a lambda light chain variable region comprising: a CDR1 comprising the amino acid sequence of SEQ ID NO:13; a CDR2 comprising the amino acid sequence of SEQ ID NO:14; and a CDR3 comprising the amino acid sequence of SEQ ID NO:15, and a kappa light chain comprising: a CDR1 comprising the amino acid sequence of SEQ ID NO:18, a CDR2 comprising the amino acid sequence of SEQ ID NO:19, and a CDR3 comprising the amino acid sequence of SEQ ID NO:20.

[0094] In some embodiments, the S79×AI3 bispecific antibody has a heavy chain variable region comprising the amino acid sequence of SEQ ID NO:9, a lambda light chain variable region comprising the amino acid sequence of SEQ ID NO:16, and a kappa light chain variable region comprising the amino acid sequence of SEQ ID NO:21.

[0095] In some embodiments, the S79×AI3 bispecific antibody has: a heavy chain variable region and a constant region comprising an amino acid sequence selected from SEQ ID NO: 10, SEQ ID NO: 11, or SEQ ID NO: 12, a lambda light chain comprising the amino acid sequence of SEQ ID NO: 17, and a kappa light chain comprising the amino acid sequence of SEQ ID NO: 22.

[0096] In some embodiments, the S79×AI13 bispecific antibody has a heavy chain comprising: a complementarity determining region 1 (CDR1) comprising the amino acid sequence of (SEQ ID NO:6); a complementarity determining region 2 (CDR2) comprising the amino acid sequence of (SEQ ID NO:7); and a complementarity determining region 3 (CDR3) comprising the amino acid sequence of (SEQ ID NO:8), a lambda light chain variable region comprising: a CDR1 comprising the amino acid sequence of SEQ ID NO:13; a CDR2 comprising the amino acid sequence of SEQ ID NO:14; and a CDR3 comprising the amino acid sequence of SEQ ID NO:15, and a kappa light chain comprising: a CDR1 comprising the amino acid sequence of SEQ ID NO:23, a CDR2 comprising the amino acid sequence of SEQ ID NO:24, and a CDR3 comprising the amino acid sequence of SEQ ID NO:25.

[0097] In some embodiments, the S79×AI13 bispecific antibody has a heavy chain variable region comprising the amino acid sequence of SEQ ID NO:9, a lambda light chain variable region comprising the amino acid sequence of SEQ ID NO:16, and a kappa light chain variable region comprising the amino acid sequence of SEQ ID NO:26.

[0098] In some embodiments, the S79×AI13 bispecific antibody has: a heavy chain variable region and a constant region comprising an amino acid sequence selected from SEQ ID NO: 10, SEQ ID NO: 11 or SEQ ID NO: 12, a λ light chain comprising the amino acid sequence of SEQ ID NO: 17, and a κ light chain comprising the amino acid sequence of SEQ ID NO: 27.

[0099] Table 1. Exemplary amino acid sequences of the present invention

[0100]

[0101]

[0102]

[0103]

[0104]

[0105]

[0106]

[0107]

[0108]

[0109]

[0110]

[0111]

[0112]

[0113]

[0114]

[0115] definition

[0116] As used herein, including the appended claims, singular forms such as "a," "an," and "the" include plural referents unless the context clearly dictates otherwise.

[0117] As used herein, “TGN1412” refers to a superagonist (SA) anti-huCD28 antibody of the human IgG4 isotype as described in WO2006050949, which comprises the amino acid sequences of SEQ ID NOs: 1 and 2.

[0118] As used herein, "CEAxCD3" refers to the CEAxCD3 bispecific κλ body originally described in WO2021053587 (the entirety of which is hereby incorporated by reference), and comprises the common heavy chain of SEQ ID NO:3, the κ light chain of SEQ ID NO:4, and the λ light chain of SEQ ID NO:5.

[0119] As used herein, "S79" refers to a high-affinity, κλ-compatible (common virtual heavy chain) anti-human PD-L1 blocking antibody with human / cynomolgus monkey / mouse cross-reactivity. It was originally described in WO2022200389 (incorporated by reference in its entirety) and comprises the amino acid sequences of SEQ ID NOs: 9 and 16 as variable heavy chains (VH) and variable light chains (VL), respectively.

[0120] As used herein, "AI3" and "AI13" refer to anti-human CD28 agonist antibodies with κλ body compatibility (common virtual heavy chain) with human / cynomolgus monkey cross-reactivity. They were originally described in WO2023170474 (the entire text of which is hereby incorporated by reference herein). AI3 comprises the amino acid sequences of SEQ ID NOs: 9 and 21 as variable heavy chains (VH) and variable light chains (VL), respectively, while AI13 comprises the amino acid sequences of SEQ ID NOs: 9 and 26 as variable heavy chains (VH) and variable light chains (VL), respectively.

[0121] As used herein, " / N" refers to a set of mutations (Leu234Ala+Leu235Ala+Pro329Ala) introduced into the human IgG1 Fc portion of a given antibody to abolish Fc-mediated effector function (ie, LALAPA).

[0122] As used in accordance with this disclosure, unless otherwise indicated, the following terms shall be understood to have the following meanings:

[0123] As used herein, the term "antibody" refers to immunoglobulin molecules and immunologically active portions of immunoglobulin (Ig) molecules, i.e., molecules that contain an antigen binding site that specifically binds (immunoreacts with) an antigen. "Specifically binds" or "immunoreacts with" or "immunospecifically binds" means that the antibody reacts with one or more antigenic determinants of the desired antigen and does not react with other polypeptides or with a much lower affinity (K). d >10 -6 ) binding. Antibodies include, but are not limited to, polyclonal antibodies, monoclonal antibodies, chimeric antibodies, dAbs (domain antibodies), single-chain, ab 、F ab′ and F (ab')2 Fragments, scFv and F ab Expression library. Antibodies with high affinity have an affinity of about 0.01 nM-25 nM.

[0124] It is known that the basic antibody structural unit comprises a tetramer. Each tetramer is composed of two pairs of identical polypeptide chains, each pair having a "light" chain (about 25kDa) and a "heavy" chain (about 50kDa-70kDa). The amino terminal portion of each chain includes a variable region of about 100 to 110 or more amino acids that is primarily responsible for antigen recognition. The carboxyl terminal portion of each chain defines a constant region that is primarily responsible for effector function. Generally speaking, antibody molecules obtained from humans relate to any one of the categories IgG, IgM, IgA, IgE, and IgD, which differ from each other by the nature of the heavy chain present in the molecule. Certain categories also have subclasses, such as IgG1, IgG2, etc. In addition, in humans, the light chain can be a kappa chain or a lambda chain.

[0125] As used herein, the term "monoclonal antibody" (MAb) or "monoclonal antibody composition" refers to a population of antibody molecules containing only one molecular species of antibody molecules consisting of a unique light chain gene product and a unique heavy chain gene product. Specifically, the complementarity determining regions (CDRs) of the monoclonal antibodies are identical in all molecules of the population. MAbs contain an antigen binding site that is capable of immunoreacting with a specific epitope of an antigen characterized by a unique binding affinity for the antigen.

[0126] The term "antigen binding region" or "antigen binding site" or "binding portion" refers to the portion of an immunoglobulin molecule that participates in antigen binding. The antigen binding site is formed by the amino acid residues of the N-terminal variable ("V") regions of the heavy ("H") and light ("L") chains. Three highly distinct stretches within the V regions of the heavy and light chains (called "hypervariable regions") are interposed between more conserved flanking stretches (called "framework regions" or "FRs"). Thus, the term "FR" refers to the amino acid sequence that naturally occurs between and adjacent to the hypervariable regions in immunoglobulins. In an antibody molecule, the three hypervariable regions of the light chain and the three hypervariable regions of the heavy chain are arranged relative to each other in three-dimensional space to form an antigen-binding surface. The antigen-binding surface is complementary to the three-dimensional surface of the bound antigen, and the three hypervariable regions of each of the heavy and light chains are referred to as "complementarity-determining regions" or "CDRs." Various methods for numbering the amino acid sequences of antibodies and identifying the complementary determining regions are known in the art. For example, the Kabat numbering system (see Kabat, EA et al., Sequences of Protein of immunological interest, 5th ed., US Department of Health and Human Services, US Government Printing Office (1991)) or the IMGT numbering system (see International ImMunoGeneTics Available online: http: / / www.imgt.org / ). The IMGT numbering system is routinely used and is considered a reliable and accurate system in the art to determine amino acid positions in coding sequences, alignment of alleles, and to easily compare sequences in immunoglobulins (IG) and T-cell receptors (TR) from all vertebrate species. The accuracy and consistency of IMGT data are based on the IMGT-ONTOLOGY, the first and so far only ontology for immunogenetics and immunoinformatics (see Lefranc. MP et al., Biomolecules, 2014 Dec; 4(4), 1102-1139). IMGT tools and databases operate against the IMGT reference catalog constructed from large sequence repositories. In the IMGT system, exon delimitations are taken into account to define IG V-DOMAIN and IG C-DOMAIN, where appropriate. Therefore, in order to make more sequences available to the IMGT database, those skilled in the art can and "have" reliably use the IMGT exon numbering system to determine amino acid positions in coding sequences and for allele alignment. Additionally, the correspondence between IMGT unique numbers and other numbers (i.e., Kabat) can be found in the IMGT scientific chart (see Lefranc. MP et al., Biomolecules, December 2014; 4(4), 1102-1139).

[0127] The term "hypervariable region" or "variable region" refers to the amino acid residues of an antibody that are generally responsible for antigen binding. The hypervariable region generally comprises amino acid residues from a "complementarity determining region" or "CDR" (e.g., when numbered according to the Kabat numbering system, V L around residues 24-34 (LI), 50-56 (L2), and 89-97 (L3), as well as V H around 31-35 (H1), 50-65 (H2), and 95-102 (H3) in the middle; Kabat et al., Sequences of Proteins of Immunological Interest, 5th ed., Public Health Service, National Institutes of Health, Bethesda, Md. (1991)); and / or those residues from a "hypervariable loop" (e.g., V when numbered according to the Chothia numbering system, V L Residues 24-34 (LI), 50-56 (L2), and 89-97 (L3) in V H26-32 (H1), 52-56 (H2), and 95-101 (H3) in the SEQ ID NO: 1; Chothia and Lesk, J. Mol. Biol. 196:901-917 (1987)); and / or those residues from the "hypervariable loop" VCDRs (e.g., V L Residues 27-38 (LI), 56-65 (L2), and 105-120 (L3) in V H 27-38 (H1), 56-65 (H2), and 105-120 (H3) in

[0015] ; Lefranc, MP et al., Nucl. Acids Res. 27:209-212 (1999); Ruiz, M. et al., Nucl. Acids Res. 28:219-221 (2000). Optionally, when numbered according to AHo, the antibody has a symmetric insertion at one or more of the following points: V L 28, 36 (LI), 63, 74-75 (L2) and 123 (L3), as well as V H 28, 36 (HI), 63, 74-75 (H2), and 123 (H3) in; Honneger, A. and Plunkthun, A., J. Mol. Biol. 309:657-670 (2001)).

[0128] As used herein, the term "epitope" includes any protein determinant that can specifically bind to an immunoglobulin, scFv or T cell receptor. The term "epitope" includes any protein determinant that can specifically bind to an immunoglobulin or T cell receptor. Epitope determinants are generally composed of chemically active surface groups (such as amino acids or sugar side chains) of molecules and generally have specific three-dimensional structural characteristics and specific charge characteristics. For example, antibodies can be produced against the N-terminal or C-terminal peptide of a polypeptide. According to the design, an antibody or a single antibody arm can specifically bind to an antigen when the dissociation constant is ≤1 μM; for example, ≤100 nM, preferably ≤10 nM and more preferably ≤1 nM. In some embodiments, an antibody or a single antibody arm can bind to an antigen with a dissociation constant greater than 20 nM.

[0129] As used herein, the terms "immunobinding" and "immunobinding properties" refer to the type of non-covalent interaction that occurs between an immunoglobulin molecule and the antigen to which the immunoglobulin is specific. The strength or affinity of the immunobinding interaction can be measured by the dissociation constant (K) of the interaction. d ) indicates that K dThe smaller the value, the greater the affinity. The immunological binding properties of a selected polypeptide can be quantified using methods well known in the art. One such method requires measuring the rates of formation and dissociation of the antigen binding site / antigen complex, where these rates depend on the concentration of the complex partner, the affinity of the interaction, and geometric parameters that affect the rates equally in both directions. Thus, the "binding rate constant" (K on ) and the “dissociation rate constant” (K off ) can be determined by calculating the concentration and the actual association and dissociation rates. (See Nature 361: 186-87 (1993)). off / K on The ratio of α and β cancels out all parameters not related to affinity and is equal to the dissociation constant K. d (See generally, Davies et al. (1990) Annual Rev Biochem 59:439-473). As measured by an assay such as a radioligand binding assay or similar assays known to those skilled in the art. When the equilibrium binding constant (K d )≤1 μM, such as ≤100 nM, preferably ≤10 nM and more preferably ≤1 nM, the antibody or single antibody arm of the present invention specifically binds to its target. In some embodiments, the antibody or single antibody arm can bind to the antigen with a dissociation constant greater than 20 nM.

[0130] As used herein, the term "isolated polynucleotide" shall mean a polynucleotide of genomic, cDNA, or synthetic origin, or some combination thereof, which, by virtue of its origin, is (1) not associated with all or a portion of a polynucleotide in which the "isolated polynucleotide" occurs in nature, (2) is operably linked to a polynucleotide that does not occur in nature, or (3) does not occur in nature as part of a larger sequence. Polynucleotides according to the present invention include nucleic acid molecules encoding heavy chain immunoglobulin molecules and nucleic acid molecules encoding light chain immunoglobulin molecules as described herein.

[0131] As used herein, the term "isolated protein" means a protein of cDNA, recombinant RNA, or synthetic origin, or some combination thereof, which, by virtue of its origin or source of derivation, (1) is not associated with proteins found in nature, (2) is free of other proteins from the same source, e.g., free of marine proteins, (3) is expressed by cells from a different species, or (4) does not occur in nature.

[0132] The term "polypeptide" is used as a general term herein to refer to analogs of native proteins, fragments or polypeptide sequences. Therefore, native protein fragments and analogs are species of the genus Polypeptide. Polypeptides according to the present invention include heavy chain immunoglobulin molecules and light chain immunoglobulin molecules as described herein, as well as antibody molecules formed by a combination comprising a heavy chain immunoglobulin molecule and a light chain immunoglobulin molecule (such as a kappa light chain immunoglobulin molecule), conversely, antibody molecules formed by a combination comprising a light chain immunoglobulin molecule (such as a kappa light chain immunoglobulin molecule) and a heavy chain immunoglobulin molecule, and fragments and analogs thereof.

[0133] As used herein, the term "naturally occurring" as applied to an object refers to the fact that an object can be found in nature. For example, a polypeptide or polynucleotide sequence that is present in an organism (including viruses) and can be isolated from a source in nature and has not been intentionally modified by humans in the laboratory or otherwise is naturally occurring.

[0134] As used herein, the term "operably linked" refers to the position of the components so described as being in a relationship permitting them to function in their intended manner. A control sequence "operably linked" to a coding sequence is linked in such a way that expression of the coding sequence is achieved under conditions compatible with the control sequences.

[0135] As used herein, the term "control sequence" refers to polynucleotide sequences that are necessary for the expression and processing of the coding sequences to which they are connected. The nature of such control sequences is different depending on the host organism in the prokaryote. Such control sequences typically include promoters, ribosome binding sites, and transcription termination sequences in eukaryotes. Generally, such control sequences include promoters and transcription termination sequences. The term "control sequence" is intended to include at least all components necessary for expression and processing, and may also include additional components whose presence is advantageous, such as leader sequences and fusion partner sequences. The term "polynucleotide" as used herein means a polymer of nucleotides of at least 10 bases in length, i.e., a modified form of ribonucleotides or deoxynucleotides or any type of nucleotide. The term includes DNA in single-stranded and double-stranded forms.

[0136] As used herein, twenty conventional amino acids and their abbreviations follow conventional usage. See Immunology-A Synthesis (2nd edition, ES Golub and DR Gren, eds., Sinauer Associates, Sunderland Mass. (1991)). Twenty conventional amino acids, non-natural amino acids such as α-amino acids, α-disubstituted amino acids, N-alkyl amino acids, lactic acid and other unconventional amino acid stereoisomers (e.g., D-amino acids) may also be suitable components of the polypeptides of the present invention. Examples of unconventional amino acids include: 4-hydroxyproline, γ-carboxyglutamate, ε-N,N,N-trimethyllysine, ε-N-acetyllysine, O-phosphoserine, N-acetylserine, N-formylmethionine, 3-methylhistidine, 5-hydroxylysine, σ-N-methylarginine, and other similar amino acids and imino acids (e.g., 4-hydroxyproline). In the polypeptide symbols used herein, according to standard usage and convention, the left-hand direction is the amino terminal direction and the right-hand direction is the carboxyl terminal direction.

[0137] When applied to polypeptides, the term "substantially identical" means that two peptide sequences, when optimally aligned, such as by the programs GAP or BESTFIT using default gap weights, share at least 80% sequence identity, preferably at least 90% sequence identity, more preferably at least 95% sequence identity, and most preferably at least 99% sequence identity.

[0138] Preferably, residue positions that are not identical differ by conservative amino acid substitutions.

[0139] Conservative amino acid substitutions refer to the interchangeability of residues with similar side chains. For example, a group of amino acids with aliphatic side chains is glycine, alanine, valine, leucine, and isoleucine; a group of amino acids with aliphatic-hydroxy side chains is serine and threonine; a group of amino acids with amide-containing side chains is asparagine and glutamine; a group of amino acids with aromatic side chains is phenylalanine, tyrosine, and tryptophan; a group of amino acids with basic side chains is lysine, arginine, and histidine; and a group of amino acids with sulfur-containing side chains is cysteine and methionine. Preferred conservative amino acid substitution groups are: valine-leucine-isoleucine, phenylalanine-tyrosine, lysine-arginine, alanine-valine, glutamic acid-aspartic acid, and asparagine-glutamine.

[0140] As discussed herein, minor changes in the amino acid sequence of an antibody or immunoglobulin molecule are contemplated as being encompassed by the present invention, provided that the change in the amino acid sequence remains at least 75%, more preferably at least 80%, 90%, 95%, and most preferably 99%. In particular, conservative amino acid substitutions are contemplated. Conservative substitutions are substitutions that occur within a family of amino acids that are related in their side chains. Genetically encoded amino acids are generally divided into families: (1) acidic amino acids are aspartic acid, glutamic acid; (2) basic amino acids are lysine, arginine, histidine; (3) non-polar amino acids are alanine, valine, leucine, isoleucine, proline, phenylalanine, methionine, tryptophan, and (4) uncharged polar amino acids are glycine, asparagine, glutamine, cysteine, serine, threonine, tyrosine. Hydrophilic amino acids include arginine, asparagine, aspartic acid, glutamine, glutamic acid, histidine, lysine, serine, and threonine. Hydrophobic amino acids include alanine, cysteine, isoleucine, leucine, methionine, phenylalanine, proline, tryptophan, tyrosine, and valine. Other families of amino acids include (i) serine and threonine, which are aliphatic-hydroxyl families; (ii) asparagine and glutamine, which are amide-containing families; (iii) alanine, valine, leucine, and isoleucine, which are aliphatic families; and (iv) phenylalanine, tryptophan, and tyrosine, which are aromatic families. For example, it is reasonable to expect that replacing leucine with isoleucine or valine alone, replacing aspartic acid with glutamic acid alone, replacing threonine with serine alone, or similar replacement of an amino acid with a structurally related amino acid will not have a significant effect on the binding or properties of the resulting molecule, particularly if the replacement does not involve an amino acid within a framework site. Whether an amino acid change results in a functional peptide can be easily determined by measuring the specific activity of the polypeptide derivative. Assays are described in detail herein. One of ordinary skill in the art can readily prepare fragments or analogs of antibodies or immunoglobulin molecules. The preferred amino terminus and carboxyl terminus of the fragment or analog are present near the boundary of the functional domain. Structural domains and functional domains can be identified by comparing nucleotide and / or amino acid sequence data with public or proprietary sequence databases. Preferably, a computerized comparative method is used to identify sequence motifs or predicted protein conformation domains present in other proteins of known structure and / or function. Methods for identifying protein sequences that fold into known three-dimensional structures are known. Bowie et al., Science 253:164 (1991). Therefore, the aforementioned examples show that those skilled in the art can identify sequence motifs and structural conformations that can be used to define structural domains and functional domains according to the present invention.

[0141] Preferred amino acid substitutions are those that: (1) reduce susceptibility to proteolysis, (2) reduce susceptibility to oxidation, (3) alter binding affinity for forming protein complexes, (4) alter binding affinity, and (5) confer or modify other physicochemical or functional properties of such analogs. Analogs may include various muteins of sequences other than the naturally occurring peptide sequence. For example, single or multiple amino acid substitutions (preferably conservative amino acid substitutions) can be made in the naturally occurring sequence, preferably in portions of the polypeptide outside of the domains forming intermolecular contacts. A conservative amino acid substitution should not substantially alter the structural characteristics of the parent sequence (e.g., the replacing amino acid should not tend to disrupt helices present in the parent sequence, or disrupt other types of secondary structure that characterize the parent sequence). Examples of art-recognized polypeptide secondary and tertiary structure are described in Proteins, Structures and Molecular Principles (Creighton, ed., WH Freeman and Company, New York (1984)); Introduction to Protein Structure (C. Branden and J. Tooze, eds., Garland Publishing, New York, NY (1991)); and Thornton et al., Nature 354:105 (1991).

[0142] As used herein, the term "label" or "labeled" refers to the incorporation of a detectable marker, for example, by incorporation of a radiolabeled amino acid or a polypeptide linked to a biotinyl moiety, which can be detected by labeled avidin (e.g., streptavidin containing a fluorescent marker or enzymatic activity, which can be detected optically or calorimetrically). In some cases, the label or marker can also be a therapeutic agent. Various methods for labeling polypeptides and glycoproteins can be used and are known in the art. Examples of labels for polypeptides include, but are not limited to, the following: radioisotopes or radionuclides (e.g., 3 H. 14 C. 15 N. 35 S. 90 Y. 99 Tc, 111 In, 125 I. 131I), fluorescent labels (e.g., FITC, rhodamine, lanthanide fluorophores), enzyme labels (e.g., horseradish peroxidase, β-galactosidase, luciferase, alkaline phosphatase), chemiluminescent labels, biotinyl groups, predetermined polypeptide epitopes recognized by secondary reporters (e.g., leucine zipper pair sequences, binding sites of secondary antibodies, metal binding domains, epitope tags). In some embodiments, the labels are connected by spacer arms of various lengths to reduce potential steric hindrance. As used herein, the term "agent or drug" refers to a compound or composition that can induce a desired therapeutic effect when properly administered to a patient.

[0143] Other chemical terms are used herein according to conventional usage in the art, as exemplified by The McGraw-Hill Dictionary of Chemical Terms (Parker, S. ed., McGraw-Hill, San Francisco (1985)).

[0144] As used herein, "substantially pure" means that the target species is the predominant species present (i.e., on a molar basis, it is more abundant than any other individual species in the composition), and preferably, a substantially purified fraction is a composition in which the target species comprises at least about 50% (on a molar basis) of all macromolecular species present.

[0145] Generally, a substantially pure composition will contain greater than about 80%, more preferably greater than about 85%, 90%, 95%, and 99% of all macromolecular species present in the composition. Most preferably, the target species is purified to essential homogeneity (contaminant species cannot be detected in the composition by conventional detection methods), wherein the composition consists essentially of a single macromolecular species.

[0146] The term patient includes human and veterinary subjects.

[0147] Antibody

[0148] A variety of procedures known in the art can be used to generate polyclonal or monoclonal antibodies to a given target, such as CD47, a tumor-associated antigen, or other target, or to derivatives, fragments, analogs, homologs, or orthologs thereof (see, e.g., Antibodies: A Laboratory Manual, Harlow E and Lane D, 1988, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY, incorporated herein by reference).

[0149] Antibodies are purified by well-known techniques such as affinity chromatography using protein A or protein G, primarily to provide the IgG fraction of immune serum. Subsequently, or alternatively, the specific antigen or epitope thereof that is the target of the immunoglobulin being sought can be fixed to a column to purify the immunospecific antibodies by immunoaffinity chromatography. The purification of immunoglobulins is discussed, for example, by D. Wilkinson (The Scientist, published by The Scientist, Inc., Philadelphia PA, Vol. 14, No. 8 (April 17, 2000), pp. 25-28).

[0150] In some embodiments, the antibodies of the present invention are monoclonal antibodies. For example, monoclonal antibodies can be produced using the procedures described in the Examples provided herein. Antibodies can also be produced, for example, by immunizing BALB / c mice with a combination of cell transfectants that express high levels of a given target on their surface. Hybridomas produced by myeloma / B cell fusions are then screened for reactivity to the selected target.

[0151] Monoclonal antibodies are prepared, for example, using a hybridoma method such as that described by Kohler and Milstein, Nature, 256:495 (1975). In the hybridoma method, mice, hamsters, or other appropriate host animals are typically immunized with an immunizing agent to elicit lymphocytes that produce or are capable of producing antibodies that will specifically bind to the immunizing agent. Alternatively, lymphocytes can be immunized in vitro.

[0152] Immunizing agents typically include protein antigens, fragments thereof, or fusion proteins thereof. Generally, if cells of human origin are desired, peripheral blood lymphocytes are used, or if non-human mammalian origin is desired, spleen cells or lymph node cells are used. Lymphocytes are then fused with an immortalized cell line using a suitable fusing agent such as polyethylene glycol to form hybridoma cells (Goding, MonoclonalAntibodies:PrinciplesandPractice , Academic Press (1986), pp. 59-103). Immortalized cell lines are typically transformed mammalian cells, particularly myeloma cells of rodent, bovine, and human origin. Typically, rat or mouse myeloma cell lines are used. Hybridoma cells can be cultured in a suitable culture medium that preferably contains one or more substances that inhibit the growth or survival of unfused immortalized cells. For example, if the parental cells lack hypoxanthine guanine phosphoribosyltransferase (HGPRT or HPRT), the culture medium for hybridomas will typically contain hypoxanthine, aminopterin, and thymidine ("HAT medium"), which prevent the growth of HGPRT-deficient cells.

[0153] Preferred immortalized cell lines are cell lines that effectively fuse, support antibodies to be stably expressed at high levels by the selected cells producing the antibodies, and are sensitive to culture medium such as HAT culture medium. A more preferred immortalized cell line is a mouse myeloma cell line, which can be obtained from, for example, the Salk Institute Cell Distribution Center in San Diego, California, and the American Type Culture Collection in Manassas, Virginia. Human myeloma cell lines and mouse-human heteromyeloma cell lines for producing monoclonal antibodies have also been described. (See Kozbor, J. Immunol., 133:3001 (1984); Brodeur et al., Monoclonal Antibody Production Techniques and Applications, Marcel Dekker, Inc., New York (1987), pp. 51-63)).

[0154] Can then measure the presence of monoclonal antibodies for antigen in the culture medium of culture hybridoma cells.Preferably, by immunoprecipitation or by in vitro binding assay, such as radioimmunoassay (RIA) or enzyme-linked immunosorbent assay (ELISA) determine the binding specificity of the monoclonal antibodies produced by hybridoma cells.Such technology and determination are known in the art.The binding affinity of monoclonal antibodies can be determined, for example, by Munson and Pollard, Anal.Biochem., 107:Scatchard analysis in 220 (1980).In addition, in the therapeutic application of monoclonal antibodies, it is important to identify antibodies with high specificity and high binding affinity to the target antigen.

[0155] After the desired hybridoma cells are identified, the clones can be subcloned by limiting dilution procedures and cultured by standard methods. (See Goding, MonoclonalAntibodies:Principles and Practice , Academic Press (1986), pp. 59-103). Suitable culture media for this purpose include, for example, Dulbecco's modified Eagle's medium and RPMI-1640 medium. Alternatively, hybridoma cells can be grown in vivo in mammals as ascites.

[0156] The monoclonal antibodies secreted by the subclones can be isolated or purified from the culture medium or ascites fluid by conventional immunoglobulin purification procedures such as protein A-Sepharose, hydroxylapatite chromatography, gel electrophoresis, dialysis, or affinity chromatography.

[0157] Monoclonal antibodies can also be prepared by recombinant DNA methods such as the methods described in U.S. Patent No. 4,816,567. The DNA encoding the monoclonal antibodies of the present invention can be easily separated and sequenced using conventional procedures (e.g., by using oligonucleotide probes that can specifically bind to the genes encoding the heavy and light chains of mouse antibodies). The hybridoma cells of the present invention are used as a preferred source of such DNA. After separation, the DNA can be placed in an expression vector and then transfected into host cells that do not otherwise produce immunoglobulins, such as monkey COS cells, Chinese hamster ovary (CHO) cells, or myeloma cells, to obtain the synthesis of monoclonal antibodies in recombinant host cells. It is also possible, for example, to replace the homologous mouse sequence with the coding sequence of the human heavy and light chain constant domains (see U.S. Patent No. 4,816,567; Morrison, Nature 368, 812-13 (1994)) or to modify the DNA by covalently linking all or part of the coding sequence of a non-immunoglobulin polypeptide to an immunoglobulin coding sequence. Such non-immunoglobulin polypeptides can be substituted for the constant domains of an antibody of the invention, or can be substituted for the variable domains of one antigen-binding site of an antibody of the invention to create a chimeric bivalent antibody.

[0158] The monoclonal antibodies of the present invention include humanized antibodies or human antibodies. These antibodies are suitable for administration to humans without causing an immune response to the administered immunoglobulin. The humanized form of an antibody is a chimeric immunoglobulin, an immunoglobulin chain or its fragments (such as Fv, Fab, Fab', F(ab')2 or other antigen-binding subsequences of an antibody), which are mainly composed of the sequence of a human immunoglobulin and contain a minimal sequence derived from a non-human immunoglobulin. For example, by following the method of Winter and colleagues (Jones et al., Nature, 321:522-525 (1986); Riechmann et al., Nature, 332:323-327 (1988); Verhoeyen et al., Science, 239:1534-1536 (1988)), humanization is performed by replacing the corresponding sequence of a human antibody with a rodent CDR or CDR sequence. (See also U.S. Patent No. 5,225,539). In some cases, the Fv framework residues of human immunoglobulin are replaced by corresponding non-human residues.Humanized antibodies also include residues that are not present in the receptor antibody or the CDR or framework sequences of the import. Generally speaking, humanized antibodies include substantially all at least one, typically two variable domains, wherein all or substantially all of the CDR regions correspond to the CDR regions of non-human immunoglobulins and all or substantially all of the framework regions are those of human immunoglobulin consensus sequences.Humanized antibodies also optimally include at least a portion of an immunoglobulin constant region (Fc), typically at least a portion of a human immunoglobulin (Jones et al., 1986; Riechmann et al., 1988; and Presta, Curr.Op.Struct.Biol., 2:593-596 (1992)).

[0159] A fully human antibody is an antibody molecule in which the entire sequence of both the light and heavy chains (including the CDRs) is derived from human genes. Such antibodies are referred to herein as "human antibodies" or "fully human antibodies." Monoclonal antibodies can be produced using trioma technology; human B cell hybridoma technology (see Kozbor et al., 1983 Immunol Today 4:72); and EBV hybridoma technology to produce monoclonal antibodies (see Cole et al., 1985, M ONOCLONAL A NTIBODIES AND C ANCER T HERAPY, Alan R. Liss, Inc., pp. 77-96). Prepared by using human hybridomas (see Cote et al., 1983, Proc Natl Acad Sci USA 80: 2026-2030) or by transforming human B cells in vitro with Epstein-Barr virus (see Cole et al., 1985, M ONOCLONAL A NTIBODIES AND C ANCER T HERAPY , Alan R. Liss, Inc., pp. 77-96) can produce and utilize monoclonal antibodies.

[0160] In addition, other techniques (including phage display libraries) can also be used to produce human antibodies. (See Hoogenboom and Winter, J. Mol. Biol., 227:381 (1991); Marks et al., J. Mol. Biol., 222:581 (1991)). Similarly, human antibodies can be prepared by introducing human immunoglobulin loci into transgenic animals (e.g., mice) in which endogenous immunoglobulin genes have been partially or completely inactivated. After challenge, human antibody production is observed that is highly similar to that observed in humans in all aspects (including gene rearrangement, assembly, and antibody repertoire). This approach is described, for example, in U.S. Pat. Nos. 5,545,807; 5,545,806; 5,569,825; 5,625,126; 5,633,425; 5,661,016 and in Marks et al., Bio / Technology 10, 779-783 (1992); Lonberg et al., Nature 368, 856-859 (1994); Morrison, Nature 368, 812-13 (1994); Fishwild et al., Nature Biotechnology 14, 845-51 (1996); Neuberger, Nature Biotechnology 14, 826 (1996); and Lonberg and Huszar, Intern. Rev. Immunol. 13, 65-93 (1995).

[0161] Human antibodies can also be produced using transgenic non-human animals that have been modified to produce fully human antibodies in response to an attack with an antigen rather than the animal's endogenous antibodies. (See PCT Publication WO94 / 02602). The endogenous genes encoding heavy and light chain immunoglobulins in the non-human host have been disabled, and the active loci encoding human heavy and light chain immunoglobulins are inserted into the host's genome. For example, yeast artificial chromosomes containing necessary human DNA fragments are used to incorporate human genes. Animals that provide all desired modifications are then obtained as offspring by hybridizing intermediate transgenic animals that contain less than all modifications. An example of such a non-human animal is a so-called Xenomouse, as disclosed in PCT Publications WO 96 / 33735 and WO96 / 34096. TM The animal produces B cells that secrete fully human immunoglobulins. Antibodies can be obtained directly from the animal after immunization with the immunogen of interest, for example as a preparation of polyclonal antibodies, or alternatively from immortalized B cells derived from the animal, such as hybridomas that produce monoclonal antibodies. In addition, genes encoding immunoglobulins with human variable regions can be recovered and expressed to directly obtain antibodies, or they can be further modified to obtain antibody analogs such as single-chain Fv (scFv) molecules.

[0162] An example of a method for generating a non-human host (using mice as an example) lacking endogenous immunoglobulin heavy chain expression is disclosed in U.S. Patent No. 5,939,598. The non-human host can be obtained by the following method, which comprises: deleting a J segment gene from at least one endogenous heavy chain locus in embryonic stem cells to prevent rearrangement of the locus and the formation of transcripts of the rearranged immunoglobulin heavy chain locus, wherein the deletion is achieved by a targeting vector containing a gene encoding a selective marker; and generating a transgenic mouse from the embryonic stem cell, wherein the somatic cells and germ cells of the transgenic mouse contain a gene encoding a selective marker.

[0163] U.S. Patent No. 5,916,771 discloses a method for producing an antibody of interest (such as a human antibody). The method comprises introducing an expression vector containing a nucleotide sequence encoding a heavy chain into a cultured mammalian host cell, introducing an expression vector containing a nucleotide sequence encoding a light chain into another mammalian host cell, and fusing the two cells to form a hybrid cell. The hybrid cell expresses an antibody containing both heavy and light chains.

[0164] In a further refinement of this procedure, a method for identifying clinically relevant epitopes on an immunogen and a related method for selecting antibodies that specifically bind to the relevant epitopes with high affinity are disclosed in PCT Publication WO 99 / 53049.

[0165] The antibody can be expressed by a vector containing a DNA fragment encoding the above-mentioned single-chain antibody.

[0166] These may include vectors, liposomes, naked DNA, adjuvant-assisted DNA, gene guns, catheters, and the like. Vectors include: chemical conjugates, such as those described in WO 93 / 64701, having a targeting moiety (e.g., a ligand for a cell surface receptor) and a nucleic acid binding moiety (e.g., polylysine); viral vectors (e.g., DNA or RNA viral vectors); fusion proteins, such as those described in PCT / US95 / 02140 (WO 95 / 22618), containing a targeting moiety (e.g., an antibody specific for a target cell) and a nucleic acid binding moiety (e.g., protamine); plasmids; phages, and the like. Vectors may be chromosomal, non-chromosomal, or synthetic.

[0167] Preferred vectors include viral vectors, fusion proteins, and chemical conjugates. Retroviral vectors include Moloney murine leukemia virus. DNA viral vectors are preferred. These vectors include poxvirus vectors such as orthopoxvirus or fowlpoxvirus vectors, herpesvirus vectors such as herpes simplex virus type I (HSV) vectors (see Geller, AI et al., J. Neurochem, 64:487 (1995); Lim, F. et al., DNA Cloning: Mammalian Systems, D. Glover ed. (Oxford Univ. Press, Oxford England) (1995); Geller, AI et al., Proc Natl. Acad. Sci.: USA 90:7603 (1993); Geller, AI et al., Proc Natl. Acad. Sci USA 87:1149 (1990), adenovirus vectors (see LeGal LaSalle et al., Science, 259:988 (1993); Davidson et al., Nat. Genet. 3:219 (1993); Yang et al., J. Virol. 69:2004 (1995) and adeno-associated virus vectors (see Kaplitt, MG et al., Nat. Genet. 8:148 (1994).

[0168] Poxvirus vectors introduce genes into the cytoplasm. Fowlpoxvirus vectors only result in short-term expression of nucleic acids. Preferably, adenovirus vectors, adeno-associated virus vectors and herpes simplex virus (HSV) vectors are used to introduce nucleic acids into nerve cells. Adenovirus vectors result in shorter-term expression (about 2 months) than adeno-associated virus (about 4 months), which is shorter than HSV vectors. The specific vector selected will depend on the target cell and the condition being treated. Standard techniques such as infection, transfection, transduction or transformation can be used to introduce the gene. Examples of gene transfer modes include, for example, naked DNA, CaPO4 precipitation, DEAE dextran, electroporation, protoplast fusion, lipofection, cell microinjection and viral vectors.

[0169] The vector can be used to target substantially any desired target cell. For example, a vector (e.g., adenovirus, HSV) can be guided to a desired location using stereotactic injection. In addition, particles can be delivered by intracerebroventricular (icv) infusion using a micropump infusion system such as the SynchroMed infusion system. It has also been shown that a method based on bulk flow (referred to as convection) is effective in delivering macromolecules to extended areas of the brain and can be used to deliver vectors to target cells. (See Bobo et al., Proc. Natl. Acad. Sci. USA 91: 2076-2080 (1994); Morrison et al., Am. J. Physiol. 266: 292-305 (1994)). Other methods that can be used include catheter, intravenous, parenteral, intraperitoneal and subcutaneous injections, as well as oral or other known routes of administration.

[0170] Bispecific antibodies are antibodies that have binding specificities for at least two different antigens. In this case, one binding specificity is for a target such as CD28 or any fragment thereof. The second binding target is any other antigen, and advantageously is a cell surface protein or receptor or receptor subunit.

[0171] Methods for preparing bispecific antibodies are known in the art. Traditionally, the recombinant production of bispecific antibodies is based on the coexpression of two heavy chain immunoglobulin / light chain pairs, wherein the two heavy chains have different specificities (Milstein and Cuello, Nature, 305:537-539 (1983)). Due to the random distribution of heavy and light chains of immunoglobulins, these hybridomas (quadroma) produce a potential mixture of ten different antibody molecules, wherein only one mixture has the correct bispecific structure. The purification of the correct molecule is usually completed by affinity chromatography steps. Similar procedures are disclosed in WO 93 / 08829 and Traunecker et al., EMBO J., 10:3655-3659 (1991), disclosed on May 13, 1993.

[0172] The bispecific and / or monospecific antibodies of the present invention can be prepared using any of a variety of art-recognized techniques, including those disclosed in co-pending application WO 2012 / 023053, filed on August 16, 2011, the contents of which are hereby incorporated by reference in their entirety. The methods described in WO 2012 / 023053 produce bispecific antibodies that are structurally identical to human immunoglobulins. This type of molecule consists of two copies of a unique heavy chain polypeptide: a first light chain variable region fused to a constant kappa domain, and a second light chain variable region fused to a constant lambda domain. Each binding site exhibits a different antigen specificity, determined by both the heavy and light chains. The light chain variable region can be of the lambda or kappa family and is preferably fused to the lambda and kappa constant domains, respectively. This is preferred to avoid creating non-natural polypeptide linkages. However, it is also possible to obtain the bispecific antibodies of the present invention by fusing a kappa light chain variable domain to a constant lambda domain to obtain a first specificity, and a lambda light chain variable domain to a constant kappa domain to obtain a second specificity. The bispecific antibodies described in WO 2012 / 023053 are referred to as IgGκλ antibodies or "κλ bodies," a new, fully human bispecific IgG format. This κλ body format allows affinity purification of bispecific antibodies that are indistinguishable from standard IgG molecules, whose characteristics are indistinguishable from those of standard monoclonal antibodies, and is therefore advantageous compared to previous formats.

[0173] A basic step of the method is to identify two antibody Fv regions (each consisting of a variable light chain domain and a variable heavy chain domain) with different antigen specificities but sharing the same heavy chain variable domain. Many methods for producing monoclonal antibodies and fragments thereof have been described. (See, for example, Antibodies: A Laboratory Manual, Harlow E and Lane D, 1988, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY, incorporated herein by reference). Fully human antibodies are antibody molecules in which the sequences of both light and heavy chains (including CDR 1 and 2) are all from human genes. The CDR3 region can be human-derived or designed by synthetic means. Such antibodies are referred to herein as "human antibodies" or "fully human antibodies." Human monoclonal antibodies can be produced using trioma technology; human B cell hybridoma technology (see Kozbor et al., 1983 Immunol Today 4: 72); and EBV hybridoma technology to produce human monoclonal antibodies (see Cole et al., 1985, M ONOCLONAL A NTIBODIES AND C ANCER THERAPY , Alan R. Liss, Inc., pp. 77-96). By using human hybridomas (see Cote et al., 1983, Proc Natl Acad Sci USA 80: 2026-2030) or by transforming human B cells in vitro with Epstein-Barr virus (see Cole et al., 1985, M ONOCLONAL A NTIBODIES AND C ANCER T HERAPY , Alan R. Liss, Inc., pp. 77-96) can produce and use human monoclonal antibodies.

[0174] Monoclonal antibodies are produced, for example, by immunizing an animal with a target antigen or an immunogenic fragment, derivative, or variant thereof. Alternatively, an animal is immunized with cells transfected with a vector containing a nucleic acid molecule encoding the target antigen, such that the target antigen is expressed and associates with the surface of the transfected cells. Various techniques for producing xenogeneic non-human animals are well known in the art. For example, see U.S. Patent Nos. 6,075,181 and 6,150,584, which are hereby incorporated by reference in their entirety.

[0175] Alternatively, antibodies are obtained by screening libraries containing antibody or antigen binding domain sequences for binding to the target antigen. The libraries are prepared, for example, in phage with phage coat proteins expressed on the surface of assembled phage particles and protein or peptide fusions of the encoding DNA sequences contained within these phage particles (i.e., "phage display libraries").

[0176] The hybridomas produced by myeloma / B cell fusion are then screened for their reactivity to the target antigen. Monoclonal antibodies are prepared, for example, using a hybridoma method such as that described by Kohler and Milstein, Nature, 256:495 (1975). In the hybridoma method, mice, hamsters, or other appropriate host animals are typically immunized with an immunizing agent to induce lymphocytes to produce or be capable of producing antibodies that will specifically bind to the immunizing agent. Alternatively, lymphocytes can be immunized in vitro.

[0177] Although not strictly impossible, it is very unlikely that accidental identification has the same heavy chain variable domain but different antibodies for different antigens.In fact, in most cases, the heavy chain mainly contributes to the antigen binding surface, and is also the most variable in the sequence.Particularly, the CDR3 on the heavy chain is the most diverse CDR in sequence, length and structure.Therefore, two antibodies that are specific to different antigens almost always carry different heavy chain variable domains.

[0178] The method disclosed in co-pending application WO 2012 / 023053 overcomes this limitation and greatly promotes the separation of antibodies with the same heavy chain variable domain by using an antibody library in which the heavy chain variable domain is the same for all library members, so diversity is limited to the light chain variable domain. Such libraries are described in, for example, co-pending applications WO 2010 / 135558 and WO 2011 / 084255, each of which is hereby incorporated by reference in its entirety. However, since the light chain variable domain is co-expressed with the heavy chain variable domain, both domains can contribute to antigen binding. In order to further promote this process, an antibody library containing the same heavy chain variable domain and a variety of lambda variable light chains or kappa variable light chains can be used in parallel to select antibodies against different antigens in vitro. This method can identify two antibodies with a common heavy chain but one antibody carrying a lambda light chain variable domain and another antibody carrying a kappa light chain variable domain, which can be used as structural units for producing bispecific antibodies in the form of complete immunoglobulins of the present invention. The bispecific antibodies of the present invention can be of different isotypes, and their Fc portions can be modified to change the binding properties to different Fc receptors, and in this way modify the effector functions of the antibodies and their pharmacokinetic properties. Many methods for modifying the Fc portion have been described, and these methods are applicable to the antibodies of the present invention. (See, for example, Strohl, WR Curr Opin Biotechnol 2009 (6): 685-91; U.S. Patent No. 6,528,624; PCT / US2009 / 0191199 filed on January 9, 2009). The methods of the present invention can also be used to produce bispecific antibodies and antibody mixtures in the form of F(ab')2 that lack the Fc portion.

[0179] A common heavy chain and two different light chains are co-expressed in a single cell to allow assembly of the bispecific antibodies of the present invention. If all polypeptides are expressed at the same level and assemble equally well to form immunoglobulin molecules, the ratio of monospecific (same light chain) to bispecific (two different light chains) should be 50%. However, different light chains may be expressed at different levels and / or not assemble with the same efficiency. Therefore, means of regulating the relative expression of different polypeptides are used to compensate for their inherent expression characteristics or different tendencies to assemble with the common heavy chain. Such regulation can be achieved by the use of promoter strength, internal ribosome entry sites (IRES) characterized by different efficiencies, or other types of regulatory elements that can act at the transcriptional or translational level and act on mRNA stability. Different promoters of different strengths can include CMV (immediate early cytomegalovirus promoter); EF1-1α (human elongation factor 1α subunit promoter); Ubc (human ubiquitin C promoter); SV40 (simian virus 40 promoter). Different IRESs from mammalian and viral sources have also been described. (See, for example, Hellen CU and Sarnow P, Genes Dev 2001 15: 1593–612). These IRESs can vary greatly in length and ribosome recruitment efficiency. In addition, it is possible to further regulate activity by introducing multiple copies of IRES (Stephen et al., 2000 Proc Natl Acad Sci USA 97: 1536-1541). Regulation of expression can also be achieved by multiple consecutive transfections of cells to increase the copy number of each gene expressing one or the other light chain and thus changing its relative expression. The examples provided herein show that controlling the relative expression of different chains is crucial for maximizing the assembly and overall yield of bispecific antibodies.

[0180] The co-expression of heavy chain and two light chains produces a mixture of three different antibodies into the cell culture supernatant: two monospecific bivalent antibodies and a bispecific bivalent antibody. The latter must be purified from the mixture to obtain the molecule of interest. The method described herein greatly facilitates the purification procedure by using affinity chromatography media such as CaptureSelect Fabκ and CaptureSelect Fabλ affinity matrices (BAC BV, Holland) that specifically interact with κ or λ light chain constant domains. This multi-step affinity chromatography purification method is effective and is generally applicable to the antibody of the present invention. This is in stark contrast to the specific purification method that must be developed and optimized for each bispecific antibody derived from four source hybridomas or other cell lines expressing antibody mixtures. In fact, if the biochemical characteristics of different antibodies in the mixture are similar, it may be challenging or impossible to separate them using standard chromatographic techniques such as ion exchange chromatography.

[0181] Other suitable purification methods include those disclosed in co-pending application PCT / IB2012 / 003028, filed October 19, 2012, published as WO 2013 / 088259, the contents of which are hereby incorporated by reference in their entirety.

[0182] In other embodiments of producing bispecific antibodies, the antibody variable domains (antibody-antigen binding sites) with the desired binding specificity can be fused with immunoglobulin constant domain sequences. Preferably, the immunoglobulin heavy chain constant domain is fused, and the immunoglobulin heavy chain constant domain comprises at least a portion of a hinge region, a CH2 region, and a CH3 region. Preferably, in at least one fusion, there is a first heavy chain constant region (CH1) containing a light chain binding site. The DNA encoding the immunoglobulin heavy chain fusions and (if necessary) immunoglobulin light chain is inserted into a separate expression vector and co-transfected into a suitable host organism. For further details on producing bispecific antibodies, see, for example, Suresh et al., Methods in Enzymology, 121: 210 (1986).

[0183] In another embodiment, the interface engineering between a pair of antibody molecules can be used to maximize the heterodimer percentage ratio of reclaiming from recombinant cell culture.Preferred interface comprises at least a portion of the CH3 district of antibody constant domain.In this method, one or more small amino acid side chains from the interface of the first antibody molecule are replaced by larger side chain (for example tyrosine or tryptophan).By replacing large amino acid side chain with less amino acid side chain (for example alanine or threonine), on the interface of the second antibody molecule, produce compensation " cavity " of the same or similar size with large side chain.This provides a kind of for heterodimer productive rate being increased to surpass other unwanted end products such as homodimer mechanism.

[0184] The literature has described techniques for generating bispecific antibodies from antibody fragments. For example, bispecific antibodies can be prepared using chemical bonds. The resulting bispecific antibodies can be used as reagents for selectively immobilizing enzymes.

[0185] Various techniques for preparing and isolating bispecific antibody fragments directly from recombinant cell culture have also been described. For example, bispecific antibodies have been produced using leucine zippers. Kostelny et al., J. Immunol. 148(5): 1547-1553 (1992). Leucine zipper peptides from Fos and Jun proteins are linked to the Fab' portions of two different antibodies by gene fusion. Antibody homodimers are reduced to form monomers at the hinge region and then reoxidized to form antibody heterodimers. This method can also be used to produce antibody homodimers. The "bivalent antibody" technology described by Hollinger et al., Proc. Natl. Acad. Sci. USA 90: 6444-6448 (1993) has provided an alternative mechanism for preparing bispecific antibody fragments. These fragments contain a light chain variable domain (V) linked to the light chain variable domain (V) by a linker. L ) of the heavy chain variable domain (V H ), the linker is too short to allow pairing between the two domains on the same chain. H and V L The domain is forced to align with the complementary V L and V H The domains pair to form two antigen-binding sites. Another strategy for preparing bispecific antibody fragments by using single-chain Fv (sFv) dimers has also been reported. See Gruber et al., J. Immunol. 152: 5368 (1994).

[0186] Antibodies with more than two valencies are contemplated. For example, trispecific antibodies can be prepared. Tutt et al., J. Immunol. 147:60 (1991).

[0187] Exemplary bispecific antibodies can bind to two different epitopes, at least one of which is derived from a protein antigen of the present invention. Alternatively, the anti-antigen arm of the immunoglobulin molecule can be combined with an arm that binds to a triggering molecule on a leukocyte, such as a T cell receptor molecule (e.g., CD2, CD3, CD28, or B7) or an Fc receptor (FcγR) of IgG, such as FcγRI (CD64), FcγRII (CD32), and FcγRIII (CD16), to focus cellular defense mechanisms on cells expressing specific antigens. Bispecific antibodies can also be used to guide cytotoxic agents to cells expressing specific antigens. These antibodies have an antigen-binding arm and an arm that binds a cytotoxic agent or a radionuclide chelator, such as EOTUBE, DPTA, DOTA, or TETA. Another bispecific antibody of interest binds to a protein antigen as described herein and further binds to tissue factor (TF).

[0188] Heteroconjugate antibodies are also within the scope of the present invention. Heteroconjugate antibodies are composed of two covalently linked antibodies. For example, it has been proposed that such antibodies target immune system cells to unwanted cells (see U.S. Patent No. 4,676,980), and are used to treat HIV infection (see WO 91 / 00360; WO 92 / 200373; EP 03089). It is expected that antibodies can be prepared in vitro using known methods in synthetic protein chemistry (including methods involving cross-linking agents). For example, immunotoxins can be constructed using disulfide exchange reactions or by forming thioether bonds. Examples of suitable reagents for this purpose include iminothiolates and methyl-4-mercaptobutyrylimidates and, for example, those disclosed in U.S. Patent No. 4,676,980.

[0189] It may be necessary to modify the antibodies of the present invention in terms of effector function to enhance, for example, the effectiveness of antibodies in treating cancer and / or other diseases and conditions associated with abnormal CD28 expression and / or activity. For example, cysteine residues can be introduced into the Fc region, thereby allowing the formation of interchain disulfide bonds in this region. The resulting homodimeric antibodies can have improved internalization capacity and / or increased complement-mediated cell killing and antibody-dependent cellular cytotoxicity (ADCC). (See Caron et al., J. Exp Med., 176: 1191-1195 (1992) and Shopes, J. Immunol., 148: 2918-2922 (1992)). Alternatively, antibodies can be engineered to have dual Fc regions, therefore having enhanced complement lysis and ADCC capabilities. (See Stevenson et al., Anti-Cancer Drug Design, 3: 219-230 (1989)).

[0190] The present invention also relates to immunoconjugates comprising an antibody conjugated to a cytotoxic agent such as a toxin (e.g., an enzymatically active toxin of bacterial, fungal, plant, or animal origin, or fragments thereof) or a radioactive isotope (ie, a radioconjugate).

[0191] Enzymatically active toxins and fragments thereof that can be used include diphtheria A chain, nonbinding active fragments of diphtheria toxin, exotoxin A chain (from Pseudomonas aeruginosa), ricin A chain, abrin A chain, modeccin A chain, α-sarcin, Aleurites fordii proteins, dianthin proteins, Phytolaca americana proteins (PAPI, PAPII, and PAP-S), momordica charantia inhibitor, curcin, crotin, sapaonaria officinalis inhibitor, gelonin, mitogellin, restrictocin, phenomycin, enomycin, and trichothecenes. A variety of radionuclides can be used to produce radioconjugated antibodies. Examples include 212 Bi, 131 I. 131 In, 90 Y and 186 Re.

[0192] Conjugates of antibodies and cytotoxic agents are prepared using a variety of bifunctional protein coupling agents such as N-succinimidyl-3-(2-pyridyldithio) propionate (SPDP), iminothiolane (IT), bifunctional derivatives of imidoesters (such as dimethyl adipimidate hydrochloride), active esters (such as disuccinimidyl suberate), aldehydes (such as glutaraldehyde), bis-azido compounds (such as bis(p-azidobenzoyl)hexanediamine), bis-diazonium derivatives (such as bis(p-diazoniumbenzoyl)-ethylenediamine), diisocyanates (such as toluene 2,6-diisocyanate), and bis-active fluorine compounds (such as 1,5-difluoro-2,4-dinitrobenzene). For example, ricin immunotoxins can be prepared as described in Vitetta et al., Science 238:1098 (1987). Carbon-14 labeled 1-isothiocyanatobenzyl-3-methyldiethylenetriaminepentaacetic acid (MX-DTPA) is an exemplary chelating agent for conjugating radionucleotides to antibodies (see WO 94 / 11026).

[0193] One of ordinary skill in the art will recognize that a wide variety of possible moieties can be coupled to the resulting antibodies of the present invention (see, e.g., "Conjugate Vaccines," in Contributions to Microbiology and Immunology, JM Cruse and R.E. Lewis, Jr. (eds.), Carger Press, New York (1989), the entire contents of which are incorporated herein by reference).

[0194] Coupling can be achieved by any chemical reaction that will combine two molecules, as long as the antibody and another part retain their corresponding activity. This key can include many chemical mechanisms, such as covalent bonding, affinity bonding, embedding, coordinate bonding and complexing. However, preferred bonding is covalent bonding. Covalent bonding can be achieved by direct condensation of existing side chains or by incorporating external bridging molecules. Many divalent or multivalent linking agents can be used for coupling protein molecules (such as antibodies of the present invention) with other molecules. For example, representative coupling agents can include organic compounds such as thioesters, carbodiimides, succinimide esters, diisocyanates, glutaraldehyde, diazobenzene and hexamethylenediamine. This list is not intended to exhaustively list various types of coupling agents known in the art, but as an example of more common coupling agents. (See Killen and Lindstrom, Jour. Immun. 133: 1335-2549 (1984); Jansen et al., Immunological Reviews 62: 185-216 (1982); and Vitetta et al., Science 238: 1098 (1987).

[0195] Preferred linkers are described in the literature. (See, e.g., Ramakrishnan, S. et al., Cancer Res. 44:201-208 (1984) describing the use of MBS (M-maleimidobenzoyl-N-hydroxysuccinimide ester). See also U.S. Pat. No. 5,030,719, which describes the use of haloacetic acid hydrazide derivatives coupled to antibodies via oligopeptide linkers. Particularly preferred linkers include: (i) EDC (1-ethyl-3-(3-dimethylamino-propyl) carbodiimide hydrochloride); (ii) SMPT (4-succinimidyloxycarbonyl-α-methyl-α-(2-pyridyl-dithio)-toluene (Pierce Chem. Co., catalog number (21558G); (iii) SPDP (succinimidyl-6[3-(2-pyridyldithio)propionamido]hexanoate (Pierce Chem. Co., catalog number (21558G)). Chem. Co., catalog number 21651G); (iv) sulfo-LC-SPDP (sulfosuccinimidyl 6[3-(2-pyridyldithio)-propionamido]hexanoate (Pierce Chem. Co., catalog number 2165-G); and (v) sulfo-NHS conjugated with EDC (N-hydroxysulfo-succinimide: Pierce Chem. Co., catalog number 24510).

[0196] The above-mentioned linkers contain components with different properties, thus producing conjugates with different physicochemical properties. For example, sulfo-NHS esters of alkyl carboxylates are more stable than sulfo-NHS esters of aromatic carboxylates. Linkers containing NHS-esters are less soluble than sulfo-NHS esters. In addition, the linker SMPT contains sterically hindered disulfide bonds and can form conjugates with increased stability. Generally speaking, disulfide bonds are less stable than other bonds because they cleave in vitro, resulting in less usable conjugates. Sulfo-NHS can specifically enhance the stability of carbodiimide coupling. When used in combination with sulfo-NHS, carbodiimide coupling (such as EDC) forms esters that are more resistant to hydrolysis than carbodiimide coupling reactions alone.

[0197] The antibodies disclosed herein can also be formulated into immunoliposomes. Liposomes containing antibodies are prepared by methods known in the art, such as Epstein et al., Proc. Natl. Acad. Sci. USA, 82: 3688 (1985); Hwang et al., Proc. Natl Acad. Sci. USA, 77: 4030 (1980); and U.S. Patent Nos. 4,485,045 and 4,544,545. Liposomes with extended circulation time are disclosed in U.S. Patent No. 5,013,556.

[0198] Particularly useful liposomes can be produced by reverse phase evaporation using a lipid composition comprising phosphatidylcholine, cholesterol, and PEG-derived phosphatidylethanolamine (PEG-PE). The liposomes are extruded through a filter of defined pore size to produce liposomes with the desired diameter. The Fab' fragments of the antibodies of the present invention can be conjugated to the liposomes by a disulfide exchange reaction as described in Martin et al., J. Biol. Chem., 257: 286-288 (1982).

[0199] How to use

[0200] It will be understood that the administration of the therapeutic entities according to the present invention will be administered with suitable carriers, excipients and other agents that are incorporated into the formulation to provide improved transfer, delivery, drug resistance, etc. Many suitable formulations can be found in the formulary known to all pharmaceutical chemists: Remington's pharmaceutical sciences (15th ed., Mack Publishing Company, Easton, PA (1975)), particularly Chapter 87 therein by Blaug, Seymour. These formulations include, for example, powders, pastes, ointments, jellies, waxes, oils, lipids, vesicles containing (e.g., Lipofectin), and the like. TM) lipids (cationic or anionic), DNA conjugates, anhydrous absorption pastes, oil-in-water and water-in-oil emulsions, emulsions carbowax (polyethylene glycol of various molecular weights), semisolid gels, and semisolid mixtures containing carbowax. Any of the foregoing mixtures may be suitable for use in the treatment and therapy according to the present invention, provided that the active ingredient in the formulation is not inactivated by the formulation and that the formulation is physiologically compatible and tolerable with the route of administration. See also Baldrick P. "Pharmaceutical excipient development: the need for preclinical guidance." Regul. Toxicol Pharmacol. 32(2):210-8 (2000); Wang W. "Lyophilization and development of solid protein pharmaceuticals." Int. J. Pharm. 203(1-2):1-60 (2000); Charman WN "Lipids, lipophilic drugs, and oral drug delivery-some emerging concepts." J Pharm Sci. 89(8):967-78 (2000); Powell et al. "Compendium of excipients for parenteral formulations" PDA J Pharm Sci Technol. 52:238-311 (1998) and citations therein for additional information on formulations / excipients and carriers familiar to pharmaceutical chemists.

[0201] The therapeutic formulations of the present invention comprise antibodies of the present invention for treating or alleviating symptoms associated with cancer, such as, by way of non-limiting example, leukemia, lymphoma, breast cancer, colon cancer, ovarian cancer, bladder cancer, prostate cancer, glioma, lung and bronchial cancer, colorectal cancer, pancreatic cancer, esophageal cancer, liver cancer, urinary bladder cancer, kidney cancer and renal pelvis cancer, oral and pharyngeal cancer, uterine corpus cancer, and / or melanoma. The present invention also provides methods for treating or alleviating symptoms associated with cancer. The treatment regimen is performed by identifying a subject, e.g., a human patient having or at risk of developing cancer, using standard methods.

[0202] Therapeutic effectiveness is determined in conjunction with any known method for diagnosing or treating a particular immune-related disorder. Alleviation of one or more symptoms of the immune-related disorder indicates that the antibody provides a clinical benefit.

[0203] Methods for screening antibodies with the desired specificity include, but are not limited to, enzyme-linked immunosorbent assay (ELISA) and other immune-mediated techniques known in the art.

[0204] Antibodies (or fragments thereof) directed against targets such as CD28, PD-L1, or a combination thereof can be used in methods known in the art related to the localization and / or quantification of these targets, for example, for measuring the levels of these targets in appropriate physiological samples, for diagnostic methods, for protein imaging, etc.). In a given embodiment, an antibody containing an antibody-derived antigen-binding domain specific for any of these targets, or derivatives, fragments, analogs, or homologs thereof, is used as a pharmacologically active compound (hereinafter referred to as a "therapeutic agent").

[0205] The antibodies of the present invention can be used to isolate specific targets using standard techniques such as immunoaffinity, chromatography or immunoprecipitation. The antibodies of the present invention (or fragments thereof) can be used to diagnose and monitor protein levels in tissues as part of a clinical testing program, for example, to determine the efficacy of a given treatment regimen. Detection can be facilitated by coupling (i.e., physically linking) the antibody to a detectable substance. Examples of detectable substances include various enzymes, prosthetic groups, fluorescent materials, luminescent materials, bioluminescent materials, and radioactive materials. Examples of suitable enzymes include horseradish peroxidase, alkaline phosphatase, β-galactosidase, or acetylcholinesterase; examples of suitable prosthetic group complexes include streptavidin / biotin and avidin / biotin; examples of suitable fluorescent materials include umbelliferone, fluorescein, fluorescein isothiocyanate, rhodamine, dichlorotriazinylaminofluorescein, dansyl chloride, or phycoerythrin; examples of luminescent materials include luminol; examples of bioluminescent materials include luciferase, luciferin, and aequorin, and examples of suitable radioactive materials include 125 I. 131 I. 35 S or 3 H.

[0206] The antibodies of the present invention, including polyclonal, monoclonal, humanized and fully human antibodies, can be used as therapeutic agents. Such agents are generally used to treat or prevent diseases or lesions associated with abnormal expression or activation of a given target in a subject. An antibody preparation, preferably an antibody preparation with high specificity and high affinity for its target antigen, is administered to a subject, and the effect is generally produced due to the binding of the antibody preparation to the target. The administration of the antibody can eliminate or inhibit or interfere with the signal transduction function of the target. The administration of the antibody can eliminate or inhibit or interfere with the binding of the target to its naturally bound endogenous ligand. The administration of the antibody can activate or stimulate or enhance the signal transduction function of the target.

[0207] The therapeutically effective amount of the antibody of the present invention generally relates to the amount required to achieve the therapeutic goal. As mentioned above, this may be the binding interaction between the antibody and its target antigen, and in some cases, this binding interaction interferes with the function of the target. In some embodiments, the administration of the antibody can activate or stimulate or enhance the signal transduction function of the target. In some embodiments, the antibody can eliminate or inhibit or interfere with the combination of the endogenous ligand of the target and its natural combination, and can activate or stimulate or enhance the signal transduction function of another target. The amount to be administered will further depend on the binding affinity of the antibody to its specific antigen, and will also depend on the rate at which the administered antibody is exhausted from the free volume of other subjects to which it is administered. The common range of the therapeutically effective dose of the antibody or antibody fragment of the present invention can be (as a non-limiting example) about 0.1 mg / kg body weight to about 50 mg / kg body weight. Common administration frequency can be in the range of, for example, twice a day to once a week.

[0208] The antibodies or fragments thereof of the present invention can be administered in the form of pharmaceutical compositions to treat a variety of diseases and conditions. The principles and considerations involved in preparing such compositions, as well as guidance for component selection, are provided in, for example, Remington: The Science And Practice Of Pharmacy, 19th ed. (Alfonso R. Gennaro et al., eds.), Mack Pub. Co., Easton, Pa.: 1995; Drug Absorption Enhancement: Concepts, Possibilities, Limitations, And Trends, Harwood Academic Publishers, Langhorne, Pa., 1994; and Peptide And Protein Drug Delivery (Advances In Parenteral Sciences, Vol. 4), 1991, M. Dekker, New York.

[0209] When using an antibody fragment, it is preferred that the smallest inhibitory fragment specifically binds to the binding domain of the target protein. For example, based on the variable region sequence of the antibody, a peptide molecule that retains the ability to bind to the target protein sequence can be designed. Such peptides can be chemically synthesized and / or produced by recombinant DNA technology. (See, for example, Marasco et al., Proc. Natl. Acad. Sci. USA, 90: 7889-7893 (1993)). The preparation may also contain more than one active compound as required for the specific indication being treated, preferably compounds with complementary activities that do not adversely affect each other. Alternatively or in addition, the composition may include agents that enhance its function, such as cytotoxic agents, cytokines, chemotherapeutic agents or growth inhibitors. Such molecules are suitably present in an amount effective for the intended purpose.

[0210] The active ingredient can also be embedded in microcapsules prepared, for example, by coacervation techniques or by interfacial polymerization, such as hydroxymethylcellulose or gelatin microcapsules and poly(methyl methacrylate) microcapsules in colloidal drug delivery systems (e.g., liposomes, albumin microspheres, microemulsions, nanoparticles and nanocapsules) or in macroemulsions, respectively.

[0211] Formulations intended for in vivo administration must be sterile. This is readily accomplished by filtration through sterile filtration membranes.

[0212] Sustained-release formulations can be prepared. Suitable examples of sustained-release formulations include semipermeable matrices of solid hydrophobic polymers containing the antibody, which are in the form of shaped articles such as films or microcapsules. Examples of sustained-release matrices include polyesters, hydrogels (e.g., poly(2-hydroxyethyl-methacrylate) or poly(vinyl alcohol)), polylactides (U.S. Patent No. 3,773,919), copolymers of L-glutamic acid and gamma ethyl-L-glutamate, non-degradable ethylene-vinyl acetate, degradable lactic acid-glycolic acid copolymers such as LUPRON DEPOT®, and the like. TM (injectable microspheres composed of lactic acid-glycolic acid copolymer and leuprolide acetate) and poly-D-(-)-3-hydroxybutyric acid. While polymers such as ethylene-vinyl acetate and lactic acid-glycolic acid can release molecules for more than 100 days, certain hydrogels release proteins over shorter time periods.

[0213] The antibodies according to the present invention can be used as reagents for detecting the presence of a given target (or a protein fragment thereof) in a sample. In some embodiments, the antibody contains a detectable label. The antibody is polyclonal, or more preferably monoclonal. The use of whole antibodies or fragments thereof (e.g., F ab , scFv or F (ab)2). The term "labeled" about a probe or antibody is intended to encompass direct labeling of a probe or antibody by coupling a detectable substance (i.e., physical connection) to the probe or antibody, as well as indirect labeling of the probe or antibody by reactivity with another reagent of direct labeling. Examples of indirect labeling include detection of a primary antibody using a fluorescently labeled secondary antibody and end-labeling of a DNA probe with biotin so that the probe can be detected with a fluorescently labeled streptavidin protein. The term "biological sample" is intended to include tissues, cells, and biological fluids separated from a subject, as well as tissues, cells, and fluids present in a subject. Therefore, the use of the term "biological sample" includes blood and blood fractions or components, including serum, plasma, or lymph. That is, the detection method of the present invention can be used for detecting analyte mRNA, protein, or genomic DNA in a biological sample in vitro and in vivo. For example, the in vitro technique for detecting analyte mRNA includes Northern hybridization and in situ hybridization. The in vitro technique for detecting analyte protein includes enzyme-linked immunosorbent assay (ELISA), western blotting, immunoprecipitation, and immunofluorescence. The in vitro technique for detecting analyte genomic DNA includes Southern hybridization. Procedures for performing immunoassays are described, for example, in "ELISA: Theory and Practice: Methods in Molecular Biology," Vol. 42, J.R. Crowther (ed.), Human Press, Totowa, N.J., 1995; "Immunoassay," E. Diamandis and T. Christopoulus, Academic Press, Inc., San Diego, Calif., 1996; and "Practice and Theory of Enzyme Immunoassays," P. Tijssen, Elsevier Science Publishers, Amsterdam, 1985. In addition, in vivo techniques for detecting analyte proteins include introducing labeled anti-analyte protein antibodies into a subject. For example, the antibody can be labeled with a radioactive marker, the presence and location of which in the subject can be detected by standard imaging techniques.

[0214] Pharmaceutical composition

[0215] The antibodies of the present invention (also referred to herein as "active compounds") and their derivatives, fragments, analogs, and homologs can be incorporated into pharmaceutical compositions suitable for administration. Such compositions typically comprise an antibody and a pharmaceutically acceptable carrier. As used herein, the term "pharmaceutically acceptable carrier" is intended to include any and all solvents, dispersion media, coatings, antibacterial and antifungal agents, isotonic and absorption delaying agents, etc. that are compatible with pharmaceutical administration. Suitable carriers are described in the latest edition of Remington's Pharmaceutical Sciences, a standard reference text in this field, which is incorporated herein by reference. Preferred examples of such carriers or diluents include, but are not limited to, water, saline, Ringer's solution, dextrose solution, and 5% human serum albumin. Liposomes and non-aqueous vehicles such as fixed oils may also be used. The use of such media and agents for pharmaceutically active substances is well known in the art. Unless any conventional media or agents are incompatible with the active compound, their use in the composition is contemplated. Supplementary active compounds may also be incorporated into the composition.

[0216] The pharmaceutical composition of the present invention is formulated to be compatible with its intended route of administration. Examples of routes of administration include parenteral, such as intravenous, intradermal, subcutaneous, oral (e.g., inhalation), transdermal (i.e., topical), transmucosal, and rectal administration. Solutions or suspensions for parenteral, intradermal, or subcutaneous administration may include the following components: sterile diluents such as water for injection, saline solutions, fixed oils, polyethylene glycols, glycerol, propylene glycol, or other synthetic solvents; antibacterial agents such as benzyl alcohol or methyl parabens; antioxidants such as ascorbic acid, sodium bisulfite; chelating agents such as ethylenediaminetetraacetic acid (EDTA); buffers such as acetates, citrates, or phosphates, and agents for regulating tension such as sodium chloride or dextrose. The pH value can be adjusted with acids or bases such as hydrochloric acid or sodium hydroxide. Parenteral preparations can be encapsulated in ampoules, disposable syringes, or multiple-dose vials made of glass or plastic.

[0217] Pharmaceutical compositions suitable for injectable use include sterile aqueous solutions (water soluble) or dispersions and sterile powders for the extemporaneous preparation of sterile injectable solutions or dispersions. For intravenous administration, suitable carriers include physiological saline, bacteriostatic water, Cremophor EL TM(BASF, Parsippany, NJ) or phosphate buffered saline (PBS). In all cases, the composition must be sterile and fluidity should be such that it is easy to inject. It must be stable under the conditions of manufacture and storage and must be preserved to prevent contamination by microorganisms such as bacteria and fungi. It must be stable under the conditions of manufacture and storage and must be preserved to prevent contamination by microorganisms such as bacteria and fungi. The carrier can be a solvent or dispersion medium, which contains, for example, water, ethanol, polyols (such as glycerol, propylene glycol and liquid polyethylene glycol, etc.) and suitable mixtures thereof. Appropriate fluidity can be maintained, for example, by using a coating such as lecithin, by maintaining the desired particle size in the case of a dispersion, and by using a surfactant. Prevention of microbial action can be achieved by various antibacterial and antifungal agents, such as parabens, chlorobutanol, phenol, ascorbic acid, thimerosal and the like. In many cases, it is preferred that an isotonic agent, such as a sugar, a polyol (such as mannitol, sorbitol), or sodium chloride, be included in the composition. Prolonged absorption of the injectable compositions can be brought about by including in the composition an agent that delays absorption, for example, aluminum monostearate and gelatin.

[0218] Sterile injectable solutions can be prepared by incorporating the required amount of the active compound into an appropriate solvent along with one or a combination of the ingredients listed above, as needed, followed by sterilization by filtration. Generally, dispersions are prepared by incorporating the active compound into a sterile vehicle containing a basic dispersion medium and the desired other ingredients from those listed above. In the case of sterile powders for the preparation of sterile injectable solutions, vacuum drying and freeze drying are used to produce a powder of the active ingredient plus any additional desired ingredients from a previously sterile-filtered solution thereof.

[0219] Oral compositions typically contain an inert diluent or edible carrier. They can be encapsulated in gelatin capsules or compressed into tablets. For the purpose of oral therapeutic administration, the active compound can be mixed with excipients and used in the form of tablets, lozenges, or capsules. Oral compositions can also be prepared using a fluid carrier used as a mouthwash, wherein the compound in the fluid carrier is orally administered and rinsed in the mouth, then spit out or swallowed. Pharmaceutically compatible binders and / or adjuvant materials may be included as part of the composition. Tablets, pills, capsules, lozenges, etc. may contain any of the following ingredients or compounds with similar properties: binders such as microcrystalline cellulose, tragacanth gum, or gelatin; excipients such as starch or lactose; disintegrants such as alginic acid, Primogel, or corn starch; lubricants such as magnesium stearate or Sterotes; glidants such as colloidal silicon dioxide; sweeteners such as sucrose or saccharin; or flavorings such as peppermint, methyl salicylate, or orange flavoring.

[0220] For administration by inhalation, the compounds are delivered in the form of an aerosol spray from a pressurized container or dispenser which contains a suitable propellant, eg, a gas such as carbon dioxide, or a nebulizer.

[0221] Systemic administration can also be carried out by transmucosal or transdermal means. For transmucosal or transdermal administration, penetrants suitable for the barrier to be penetrated are used in the formulation. Such penetrants are well known in the art and include, for example, detergents, bile salts, and fusidic acid derivatives for transmucosal administration. Transmucosal administration can be accomplished by the use of nasal sprays or suppositories. For transdermal administration, the active compound is formulated into ointments, salves, gels, or creams as are well known in the art.

[0222] The compounds may also be prepared in the form of suppositories (eg, with conventional suppository bases such as cocoa butter and other glycerides) or retention enemas for rectal delivery.

[0223] In one embodiment, the active compound is prepared with carriers that protect the compound from being rapidly removed from the body, such as controlled release formulations, including implants and microcapsule delivery systems. Biodegradable biocompatible polymers such as ethylene vinyl acetate, polyanhydrides, polyglycolic acid, collagen, polyorthoesters, and polylactic acid can be used. The method for preparing such preparations will be apparent to those skilled in the art. These materials can also be commercially available from Alza Corporation and Nova Pharmaceuticals, Inc. Liposomal suspensions (including liposomes targeted to infected cells with monoclonal antibodies directed against viral antigens) can also be used as pharmaceutically acceptable carriers. These liposomal suspensions can be prepared according to methods known to those skilled in the art, for example, as described in U.S. Patent number 4,522,811.

[0224] For ease of administration and uniformity of dosage, it is particularly advantageous to formulate oral or parenteral compositions in dosage unit form. As used herein, dosage unit form refers to physically discrete units suitable as unitary dosages for the subject to be treated; each unit contains a predetermined quantity of active compound calculated to produce the desired therapeutic effect in association with the required pharmaceutical carrier. The specifications for the dosage unit forms of the present invention are dictated by and directly dependent on the unique characteristics of the active compound and the specific therapeutic effect to be achieved, as well as the limitations inherent in the art of compounding such active compounds for treating individuals.

[0225] The pharmaceutical compositions can be included in a container, pack, or dispenser together with instructions for administration.

[0226] The present invention will be further described in the following examples, which do not limit the scope of the invention described in the claims.

[0227] Example

[0228] Example 1: Expression and purification of bispecific antibodies carrying λ and κ light chains

[0229] The simultaneous expression of one heavy chain and two light chains in the same cell can result in the assembly of three different antibodies. Simultaneous expression can be achieved in different ways, such as transfection of multiple vectors expressing one of the chains to be co-expressed or by using vectors driving the expression of multiple genes.

[0230] Here, the two light chains were cloned into the vector pNoviκHλ, which was previously generated to allow co-expression of one heavy chain, one kappa light chain, and one lambda light chain, as described in US20120184716 and WO2012023053, each of which is hereby incorporated by reference in its entirety. The expression of these three genes is driven by the human cytomegalovirus promoter (hCMV), and the vector also contains a glutamine synthetase gene (GS), which enables selection and establishment of stable cell lines. The common VH and VL genes of anti-CD28 IgG and anti-PD-L1 IgG were cloned into the vector pNoviκHλ for transient expression in mammalian cells. Expi293 cells were cultured in suspension in appropriate Erlenmeyer flasks with appropriate cell numbers and culture medium volumes. Plasmid DNA was transfected into Expi293 cells using PEI. Antibody concentrations in the supernatant of transfected cells were measured using an Octet RED96 during production. The supernatant was harvested 5 to 7 days after transfection, depending on the antibody concentration, and clarified by filtration after adding diatomaceous earth (Sartorius). Purification was based on a three-step purification process. First, the CaptureSelect TM FcXL affinity matrix (Thermo Fisher Scientific) was washed with PBS and then added to the clarified supernatant. After incubation overnight at +4°C and 20rpm, the supernatant was centrifuged at 2000g for 10 minutes, the flow-through was saved, and the resin was washed twice with PBS. The resin was then transferred to an Amicon Pro column and a solution containing 50mM glycine at pH 3.5 was used for elution. Several elution fractions were generated, neutralized with Tris-HCl pH 7.4 and combined. The pool containing total human IgG (bispecific antibody and two monospecific antibodies) was quantified using a NanoDrop spectrophotometer (NanoDrop Technologies). A small aliquot was saved for further analysis, and the remaining sample was incubated at room temperature and 20rpm with an appropriate volume of CaptureSelect TMThe cells were incubated with KappaXL affinity matrix (Thermo Fisher Scientific) for 30 minutes. Resin recovery and washing, elution, and neutralization steps were performed as described above. The final affinity purification step was performed using the same process as the kappa purification step using CaptureSelect TM Alternatively, purification was based on a two-step purification process using only CaptureSelect TM KappaXL Affinity Matrix and CaptureSelect TM Lambda Fab affinity matrix. All elution fractions were pooled and desalted against His-NaCl pH 6.0 formulation buffer using a 50 kDa Amicon Ultra centrifugal filter unit (Merck Millipore). The final product was quantified using Nanodrop.

[0231] Purified bispecific antibodies were analyzed by electrophoresis under denaturing and reducing conditions using an Agilent 2100 bioanalyzer with a Protein 80 kit as described by the manufacturer (Agilent Technologies). Aggregation levels were determined by SEC-UPLC. All samples were tested for endotoxin contamination using a Limulus amebocyte lysate (LAL; Charles River Laboratories) test. Table 2 summarizes the generated PD-L1xCD28κλ bodies.

[0232] Table 2: Chains and nomenclature used to generate the κλ body set of the present invention. AI represents the anti-human CD28 arm, S represents the anti-human PD-L1 arm " / N" represents a heavy chain with LALAPA mutation. .

[0233]

[0234] Example 2: In vitro characterization of PD-L1xCD28 bispecific antibody

[0235] To demonstrate the binding of PD-L1xCD28κλ to target cells ( Figure 1 ), performed a series of flow cytometry-based experiments.

[0236] Examples of cells that can be used include PD-L1 positive cell lines such as the pancreatic cancer epithelial cell line HPAC, CD28 positive cell lines such as leukemia Jurkat T cells, and PD-L1 and CD28 double negative cell lines such as leukemia TIB153 cells.

[0237] Harvest cells, check for viability and count. 200'000 cells were incubated at 4°C for 15 minutes with increasing concentrations of antibodies diluted in FACS buffer (PBS 2% BSA, 0.1% NaN3). The cells were washed twice with cold FACS buffer and incubated for 15 minutes at 4°C with appropriate anti-human IgG secondary antibodies. The cells were washed twice with cold FACS buffer and resuspended in 150 μL FACS buffer with compatible viability markers. The binding of antibodies to living cells was measured by flow cytometry using the Cytoflex platform (Beckman Coulter). FlowJo TM Data were analyzed using BD Biosciences v10 software (BD Life Sciences), and dose-response binding curves were plotted using GraphPad Prism 9 software.

[0238] The binding curves of the exemplary PD-L1xCD28 bsAb AI3S79 / N of the present invention obtained using HPAC, Jurkat and TIB153 cells are shown in FIG. Figure 2A 、 Figure 2B and Figure 2C middle.

[0239] AI3S79 / N and expression of PD-L1( Figure 2A ) or CD28( Figure 2B ) cells. The absence of binding signals on TIB-153 indicates that all binding arms of the present invention are specific for the designated target ( Figure 2C ).

[0240] pass The affinity of AI3S79 / N for CD28, PD-L1, Fcγ receptor (FcγR), and neonatal Fc receptor (FcRn) from human, cynomolgus monkey, and mouse species was determined by surface plasmon resonance (Tables 3, 4, 5, 6, and 7).

[0241] Table 3: Affinity of AI3S79 / N for human, cynomolgus monkey, and mouse CD28 and PD-L1 targets

[0242]

[0243] KD (nM) ± SD (nM), n = 2

[0244] The extracellular domains of human and cynomolgus monkey CD28 are 100% identical, so the KD values of AI3S79 / N for human and cynomolgus monkey CD28 are equal (Table 3). The data show that AI3S79 / N has cross-reactivity to cynomolgus monkey CD28 (KD for both human and cynomolgus monkey CD28 is about 54nM) and PD-L1 (KD for humans is about 0.3nM and for cynomolgus monkeys is about 1.8nM). The anti-PD-L1 arm of AI3S79 / N also has cross-reactivity to mouse PD-L1 (KD for humans is about 0.3nM and for mice is about 0.79nM). However, the anti-CD28 arm of AI3S79 / N has no cross-reactivity to mouse CD28. In fact, it shows poor affinity for mouse CD28 and is therefore considered to have no cross-reactivity.

[0245] As shown in Table 4, AI3S79 / N exhibited comparable affinity for human and cynomolgus monkey FcRn (KD of approximately 7.1 nM for human and approximately 8.6 nM for cynomolgus monkey).

[0246] Table 4: Affinity of AI3S79 / N for human, cynomolgus monkey and mouse FcRn

[0247]

[0248] KD (nM) ± SD (nM), n = 2

[0249] As shown in Table 5, AI3S79 / N bound to human CD64 with poor affinity (KD of approximately 804 nM) and did not bind to other human FcγRs (CD32a R167, CD32a H167, CD32b, CD16a V158, CD16a F158, CD16b).

[0250] Table 5: Affinity of AI3S79 / N for human FcγR

[0251]

[0252] h: human; KD (nM) ± SD (nM), n = 1 (highest tested concentration: 2000 nM)

[0253] The LALAPA mutation introduced into AI3S79 / N renders the molecule Fc-silent. The absence of AI3S79 / N binding to FcγRs prevents Fc-mediated effector functions (i.e., ADCP or ADCC).

[0254] As shown in Table 6, AI3S79 / N binds to cynomolgus monkey CD64 with poor affinity (KD of approximately 183 nM) and does not bind to other cynomolgus monkey FcγRs (CD32a, CD32b, CD16). As in humans, the LALAPA mutation in AI3S79 / N is expected to prevent Fc-mediated effector functions (i.e., ADCP, ADCC, and CDC) in cynomolgus monkeys.

[0255] Table 6: Affinity of AI3S79 / N for cynomolgus monkey FcγR

[0256] cyCD64 cyCD32a cyCD32b cyCD16 183 No binding No binding No binding

[0257] cy: cynomolgus monkey; KD (nM) ± SD (nM), n = 1 (highest tested concentration: 2000 nM)

[0258] As shown in Table 7, AI3S79 / N did not bind to any of the mouse FcγRs tested (CD64, CD32b, CD16).

[0259] Table 7: Affinity of AI3S79 / N for mouse FcγR

[0260] mCD64 mCD32b mCD16 No binding No binding No binding

[0261] KD (nM) ± SD (nM), n = 1 (highest tested concentration: 2000 nM)

[0262] Example 3: PD-1 / PD-L1 Blockade Bioassay

[0263] The ability of AI3S79 / N to block the PD-1 / PD-L1 interaction was assessed using the PD-1 / PD-L1 blocking bioassay, a biologically relevant MOA-based assay that measures the potency of antibodies designed to block the PD-1 / PD-L1 interaction (Promega, J1250).

[0264] The kit consists of two cell lines: (1) artificial antigen-presenting cells (aAPCs) based on engineered CHO-K1 cells that express PD-L1 and proteins designed to activate the cognate TCR in an antigen-independent manner at the cell surface; and (2) Jurkat T cells that stably express human PD-1 and NFAT-induced luciferase. When these two cell types are co-cultured, the PD-1 / PD-L1 interaction inhibits TCR signaling and NFAT-mediated luciferase activity. Adding antibodies that block PD-1 or PD-L1 releases inhibitory signals and leads to TCR signaling and NFAT-mediated luciferase activity.

[0265] like Figure 3As shown, because the reporter cells also express CD28 on the cell surface, AI3S79 / N not only triggers luminescence by blocking PD-1 / PD-L1 engagement, but also further enhances luminescence through CD28 co-stimulation, resulting in a higher RLU signal than a simple PD-(L)1 blocker (i.e., atezolizumab or a bivalent PD-L1 mAb based on the anti-PD-L1 arm S79). Due to the lack of a PD-(L)1 blocking arm, the CD28 monovalent bsAb cannot induce the reporter system alone and therefore does not trigger a luminescent signal.

[0266] This reporter gene assay demonstrated that, in the presence of T cell signal 1 (provided by aAPCs) and PD-L-1, the PD-L1xCD28 bispecific antibody of the present invention can enhance T cell responses by inhibiting the PD-1 / PD-L1 interaction while further delivering the co-stimulatory signal 2 to T cells.

[0267] Example 4: T cell-dependent cytotoxicity (TDCC) mediated by PD-L1xCD28 bispecific antibody

[0268] TDCC of PD-L1 and CEA double-positive cell lines

[0269] Using human PBMCs as effector cells, T cell-dependent cytotoxicity (TDCC) of PD-L1 / CEA double-positive cell lines induced by the combination of the PD-L1xCD28 bispecific antibody of the present invention and CEAxCD3 bsAb was evaluated.

[0270] After washing twice with PBS, target cells were detached using trypsin or cell dissociation solution. After a centrifugation step, cells were resuspended in assay medium, adjusted to the desired concentration, and seeded in 96-well plates.

[0271] Effector cells were grown using Lymphoprep TM SepMate buffer (Stemcell Technologies) TM Human peripheral blood mononuclear cells (PBMCs) were isolated from buffy coats derived from healthy human donors using Stemcell Technologies.

[0272] For the TDCC assay, PBMCs were added to target cells at different final E:T ratios (10:1, 3:1, 1:1, and 1:3). A dose range of CEAxCD3 and a fixed dose of the PD-L1xCD28 antibody of the present invention (2.5 μg / mL) were added to pre-seeded target cells and effector cells. As a negative control, single-agent AI3S79 / N (no CEAxCD3 = no T cell signal 1) was used. After incubation for 6 days at 37°C and 5% CO2, the cells were cultured using Promega's CellTiter- (G7570) Quantify the number of viable adherent cells in culture to assess target cell killing. TDCC curves at each E:T ratio were plotted using GraphPad Prism9 ( Figure 4 ).

[0273] PD-L1xCD28 bsAb synergizes with CEAxCD3 bsAb to kill PD-L1 / CEA double-positive HPAC target cells, especially at lower E:T ratios ( Figure 4 , right panel). While CD3 bsAbs consistently induced target cell killing at an optimal E:T ratio (i.e., 10:1), their activity decreased as the E:T ratio decreased, until, as a single agent, they were completely unable to induce target cell killing (i.e., at an E:T ratio of 1:3). However, addition of AI3S79 / N enhanced CD3 bsAb activity along the E:T ratio gradient, with up to 50% specific lysis observed even at the unfavorable 1:3 E:T ratio. Importantly, in the absence of CEAxCD3, the PD-L1xCD28 bsAb alone did not induce killing, highlighting the importance of primary T cell stimulation (signal 1) for PD-L1xCD28 bsAb activity.

[0274] After killing of PD-L1 / CEA-expressing tumor cells induced by the combination of CEAxCD3 and PD-L1xCD28 bsAbs Upregulation of T cell activation markers

[0275] Killing of CEA-positive tumor cells induced by CEAxCD3 bsAb is based on T cell activation. The activation state of T cells can be further increased by CD28 co-stimulation. Therefore, using antibodies that recognize specific T cell activation markers such as CD25 (a late activation marker), the ability of PD-L1xCD28κλ bodies to enhance T cell activation in the presence of appropriate signals was quantified by flow cytometry.

[0276] In order to assess the activation state of T cells at the end of the killing assay (described in detail in Example 4a), the following procedure was applied: floating cells (which include both CD4+ T cells and CD8+ T cells) were transferred to a new V-bottom 96-well plate. The supernatant was removed by centrifugation, and the cells were washed twice with cold FACS buffer (PBS2% BSA, 0.1% NaN3), then incubated with Fc blocking reagent (BD Biosciences) at 4°C for 15 minutes. After washing twice with FACS buffer, the cells were incubated with the following antibodies at 4°C for 15 minutes: anti-CD8-PerCP-Cy5.5 (BioLegend), anti-CD25-PE (BioLegend) and anti-CD4-APC (ThermoFisher). The cells were washed and analyzed by flow cytometry using a Cytoflex platform (Beckman Coulter). Flow cytometry was performed using FlowJo TM Data were analyzed using v10 software (BD Life Sciences). Figure 4 The quantitative results of T cell activation in the TDCC experiment are shown as Figure 5 shown.

[0277] The late activation marker CD25 was quantified on the surface of both CD4+ T cells and CD8+ T cells at different E:T ratios ( Figure 5 T cell activation was measured using the top and bottom rows. The combination of CEAxCD3 and AI3S79 / N activated both CD4+ and CD8+ T cells to a greater extent than single treatment with CEAxCD3, with CD25 staining being much brighter with the combination treatment (up to 47-fold higher MFI signal for the combination compared to CD3 bsAb alone). The best synergy between CD3 and CD28 bsAbs was observed at high E:T ratios, but even at an unfavorable 1:3 E:T ratio, T cells exhibited brighter CD25 staining (at the highest antibody concentration, the MFI for the combination was 13-fold higher than for CD3 bsAb alone).

[0278] In the presence of CEAxCD3 bsAb, CD28 costimulation mediated by PD-L1xCD28 bsAb has a significant effect on T cell proliferation. The impact of colonization .

[0279] The ability of AI3S79 / N to enhance the effects of CEAxCD3 in inducing T cell proliferation in the presence of PD-L1 / CEA-positive tumor target cells was analyzed. Freshly isolated human PBMCs were stained with the CellTraceViolet Cell Proliferation Kit (ThermoFischer Scientific) according to the manufacturer's instructions, washed, and co-cultured with target cells at varying E:T ratios in the presence of a dose range of CEAxCD3 and a fixed dose of 2.5 μg / mL PD-L1xCD28 bsAb. After co-culture, effector cells were harvested, washed, stained with appropriate viability markers to exclude dead cells, and stained with anti-CD4-APC (ThermoFischer, 17-0049-41) and anti-CD8-PerCP-Cy5.5 (BioLegend, 301032) to identify populations of interest. The proliferation rate of T cells was calculated by measuring the intensity level of CellTrace Violet staining on live CD4+ T cells or CD8+ T cells using flow cytometry using CytoFLEX (Beckman Coulter). Data were evaluated by FlowJo software and plotted using GraphPad Prism ( Figure 6 )

[0280] The percentages of proliferating CD4+ T cells and CD8+ T cells at different E:T ratios are shown in Figure 6 The top and bottom rows show the ability of the CEAxCD3 bsAb to induce T cell proliferation, which was enhanced by the addition of AI3S79 / N, with CD4 T cells being more positively affected by the combination treatment than CD8 T cells. Optimal T cell proliferation was observed at high E:T ratios, depending on T cell activation state, but even at an unfavorable 1:3 E:T ratio, T cells exhibited a higher proliferative potential when treated with the combination (up to 35-fold higher percentage of proliferative T cells compared to CD3 bsAb alone).

[0281] Example 5: In vivo efficacy of AI3S79 / N in treating MC38-hPD-L1 transplanted into CD28 HuGEMM mice Efficacy Research

[0282] This efficacy study was designed to evaluate the antitumor effect of AI3S79 / N as a monotherapy. Figure 7ABriefly, MC38-hPD-L1 HuCELLs (MC38 cells that constitutively express human PD-L1 on the cell surface; Crown Bioscience) were subcutaneously transplanted into CD28 HuGEMM mice (an immunocompetent chimeric mouse model engineered to express humanized CD28 rather than mouse CD28; Crown Bioscience). When the average tumor volume approached 100 mm3 (7 days after transplantation), mice were randomized and treated weekly with the following treatments for a total of 3 injections ( Figure 7A ) : Vehicle, atezolizumab [5 mg / kg, bivalent anti-PD-L1 antibody] or AI3S79 / N [10 mg / kg, PD-L1xCD28, monovalent engagement of each target]. Tumors were measured twice weekly by digital calipers until the end of the experiment (tumor volume = 3000 mm3). Tumor volume was calculated using the formula (length × width2) × 0.5.

[0283] In contrast to atezolizumab, which produced only limited tumor growth inhibition compared to vehicle control, AI3S79 / N resulted in tumor regression in all mice ( Figure 7B Comparison of tumor volumes at study termination (day 28 after treatment initiation) highlighted statistically greater antitumor activity of AI3S79 / N compared with atezolizumab ( Figure 7C The vehicle ( Figure 7D ), Atezolizumab ( Figure 7E ) and AI3S79 / N( Figure 7F ) of individual mice treated with AI3S79 / N. Notably, 6 of the 10 mice treated with AI3S79 / N were considered tumor-free at study termination (day 28), with two additional mice having small regressed tumors. Normal weight gain and no clinical signs of toxicity were observed after the treatment cycle ( Figure 8A Cytokine secretion was measured 6 hours after the first injection of vehicle, atezolizumab (5 mg / kg), or AI3S79 / N (10 mg / kg). Figure 8B ), and low cytokine concentrations were detected in mice treated with AI3S79 / N, consistent with results measured in mice treated with atezolizumab, indicating that the immune system was not overactivated. Overall, AI3S79 / N was well tolerated in immunocompetent huCD28 mice, where AI3S79 / N served as a fully mouse cross-reactive antibody.

[0284] Because most AI3S79 / N-treated mice survived primary tumor challenge with MC38-huPD-L1 (8 / 10 mice) (experimental protocol shown in Figure 7ATherefore, surviving mice were re-challenged with MC38 WT cells on the opposite flank on day 59 ( Figure 9 In contrast to naive mice that all succumbed to MC38WT tumors, none of the re-challenged mice developed tumors, indicating that AI3S79 / N induces immune memory against MC38 cells.

[0285] Example 6: AI3S79 / N is not a CD28 superagonist

[0286] It has been shown in WO2023170474 that the AI3 anti-CD28 arm is not a superagonist. To exclude superagonist activity in the context of the bsAbs of the present invention, the ability of AI3S79 / N to induce T cell proliferation or T cell-mediated cytokine release in the absence of signal 1 was tested in two different in vitro safety assays.

[0287] Wet- and dry-plate-coated T cell proliferation assays (Stebbings assay)

[0288] Methods were adapted from Stebbings et al. (2007). Briefly, 96-well polypropylene plates were coated overnight with antibodies diluted to 10 μg / mL in PBS: 100 μL of antibody solution at 4°C (wet coating) or 50 μL of antibody solution at room temperature, unsealed and in a Class II laminar flow cabinet (to allow buffer evaporation = dry coating). After either coating procedure, the plates were washed twice with PBS. In parallel, PBMCs isolated from buffy coats obtained from healthy donors were stained using the CellTrace Violet Cell Proliferation Kit (ThermoFischer Scientific) according to the manufacturer's instructions. 100,000 stained PBMC cells were added to a 96-well plate in a final volume of 200 μL / well and incubated at 37°C + 5% CO2 for 6 days. The cells were then harvested and stained for flow cytometric evaluation using anti-CD4-APC (ThermoFischer, 17-0049-41) and anti-CD8-PerCP-Cy5.5 (BioLegend, 301032), as described in detail in Examples 2 and 4c. The proliferation rate of live CD4+ T cells and CD8+ T cells was calculated by measuring the level of CellTrace Violet staining using a CytoFLEX (Beckman Coulter) flow cytometer, and the results of the two coating procedures were evaluated by FlowJo software. The anti-CD3 antibody OKT3 and the CD28 SA antibody TGN1412 served as positive controls, while the background proliferation rate of T cells was determined in the presence of an isotype control antibody.

[0289] like Figure 13As shown, the TGN1412 analog induced proliferation of both CD4+ and CD8+ T cells under both wet and dry coating conditions, confirming its super-agonistic characteristics. In contrast, the AI3 anti-CD28 arm did not induce proliferation of resting CD4+ and CD8+ T cells under both experimental conditions (wet and dry coating), neither as a monoclonal antibody (mAbAI3 / N) nor as part of the PD-L1xCD28 bsAb of the present invention (AI3S79 / N).

[0290] To increase the sensitivity of T cell responses, pre-culture PBMCs at high cell density (reconstitution protocol)

[0291] To confirm the lack of superagonism of AI3N79 / N in a more physiologically relevant setting, AI3S79 / N ( et al., 2011). PBMCs from healthy donors were first pre-cultured at high density (HDP) to induce functional maturation of both monocytes and T cells and then cultured at normal density in the presence of soluble antibodies.

[0292] Under these conditions, TGN1412 analogs induced dose-dependent secretion of IL-2 from all three donors tested, although the extent of induction varied, reflecting donor variability ( Figure 10A As expected, the superagonist activity of TGN1412 did not occur if the PBMCs were not subjected to a pre-culture step at high density (fresh PBMCs). Importantly, AI3S79 / N did not induce any IL-2 secretion, even when added to HDP PBMCs ( Figure 10B ).

[0293] Example 7: T cell activation in mixed lymphocyte reactions in the presence of T cell superantigens

[0294] The mixed lymphocyte reaction (MLR) is an in vitro assay in which immune cells from two individuals are co-cultured to trigger "non-self" recognition required for allogeneic T cell activation and proliferation. In this assay, immune checkpoint inhibitors (ICIs) such as anti-PD1 or anti-PD-L1 mAbs enhance the MLR as measured by an increase in cytokine secretion. To evaluate the ability of AI3S79 / N to enhance T cell responses in the presence of appropriate stimuli, a variant of the MLR was developed in which CD4+ T cells (responders) and monocyte-derived DCs (stimulators) were co-cultured in the presence of the Staphylococcal aureus T cell superantigen SEA ( Figure 11A ).

[0295] As expected, both atezolizumab and nivolumab enhanced IL-2 secretion in such an MLR assay ( Figure 11B). The curves for the two antibodies are different and may reflect the fact that atezolizumab only blocks the PD-L1 / PD-1 interaction, while nivolumab, by binding to PD-1, can block both PD-L1 / PD-1 and PD-L2 / PD-1 interactions. S79 mAb, a bivalent monoclonal antibody carrying the same anti-PD-L1 arm as present in AI3S79 / N, induced IL-2 secretion at levels comparable to atezolizumab. Notably, AI3S79 / N was the most active molecule in this assay, as it not only blocks PD-L1 but also provides a co-stimulatory signal to CD4+ T cells, leading to better T cell activation ( Figure 11B ).

[0296] Example 8: In vivo in a humanized mouse model susceptible to CD28-mediated cytokine release syndrome Safety Assessment

[0297] A humanized mouse model sensitive to CD28-mediated CRS was used to evaluate the in vivo safety of AI3S79 / N. Briefly, non-obese diabetic (NOD) scidγ (NSG)-major histocompatibility complex (MHC) I / II double knockout (KO) mice were irradiated and transplanted with human PBMCs previously selected to be sensitive to an anti-CD28 superagonist Ab. Six days later, the mice were administered the Ab. As a positive control, a TGN1412 analog was used. As a negative control, mice were injected with phosphate-buffered saline (PBS) vehicle. Mice were observed daily, and weight monitoring and CRS score assessment were performed. Mice that reached endpoint based on body weight (>20%) or CRS score (>3) were euthanized. Figure 12 As shown, these mice tolerated treatment with a dose range of AI3S79 / N, in contrast to mice that received 2 mg / kg TGN1412, all of which showed significant weight loss on day 6 after drug administration.

[0298] Example 9: In vitro single agent activity of AI3S79 / N in the CMV recall assay

[0299] To evaluate the single-agent activity of AI3S79 / N in a physiologically relevant in vitro setting, tumor cells were artificially loaded with a library of CMV-derived peptides, including NLV peptides (HLA-A*02-restricted epitopes derived from the CMV low matrix protein pp65), in the assay. These tumor cells were then co-cultured with PBMCs from donors known to contain NLV-specific T cells. When these T cells again encountered NLV peptides presented in the context of MHC molecules, they were activated, began to proliferate and specifically killed NLV-loaded target cells, a process expected to be amplified by the PD-L1xCD28 bsAb by providing signal 2 for T cell activation. A luminescent cell viability assay measures remaining live target cells to quantify killing, while fluorescently labeled HLA-A*02-NLV tetramers are used to quantify the expansion of NLV-specific T cells.

[0300] The PD-L1-positive MDA-MB-231 cell line was first loaded with a pool of CMV-derived peptides and then co-cultured with two different PBMC donors containing CMV-reactive T cells. For both donors, dose-dependent killing of NLV-loaded target cells by AI3S79 / N was observed ( Figure 14A Correspondingly, a dose-dependent expansion of NLV-specific T cells induced by AI3S79 / N was observed ( Figure 14B Importantly, when target cells were not loaded with CMV peptide and therefore unable to provide T cell signal 1, they were not killed and T cells did not proliferate in the presence of AI3S79 / N (non-CMV loaded controls on both panels).

[0301] Example 10: In vivo efficacy study of combination with T cell engager (TCE) in a TCE resistance model

[0302] Combination of AI3S79 / N and CEAxCD3 TCE in the treatment of HPAC tumors transplanted into PBMC-humanized NOG mice In vivo efficacy studies

[0303] The synergistic effect between AI3S79 / N and CEAxCD3 observed in vitro (Example 4) was confirmed in vivo in a model in which the human pancreatic cancer epithelial cell line HPAC was subcutaneously transplanted into PBMC humanized NOG mice ( Figure 15A ). In short: 3×10 6 HPAC cells were subcutaneously transplanted into NOG mice (Taconic). When the average tumor volume reached 100 mm 3 When 10×10 6 The mice were humanized with PBMC cells. Two donors were used, and each group ultimately contained 4 mice from each donor, for a total of 8 mice per group. Treatment was performed on day 15 when the average tumor volume approached 300 mm 3 The treatment period was 24 hours and consisted of three injections of vehicle, CEAxCD3 (3 mg / kg), AI3S79 / N (5 mg / kg), or a combination of CEAxCD3 and AI3S79 / N (administered at 3 mg / kg or 1 mg / kg, respectively) on days 15, 21, and 28. Tumors were measured by digital caliper three times per week until the endpoint of the experiment (tumor volume = 1500 mm 3 ) or GvHD symptoms. Use the formula (length × width 2 )×0.5 to calculate the tumor volume.

[0304] The experiment had to be stopped due to the onset of GvHD symptoms on day 31. Before this, and starting from day 24, the mean tumor volume of the combination group began to decrease ( Figure 15B ), and tumor regression or stasis was observed in all mice treated with the combination ( Figure 15C ), indicating that AI3S79 / N synergizes with CEAxCD3 to induce a strong anti-tumor response.

[0305] The combination of AI3S79 / N and CEAxCD3 TCE is effective in treating HPAF-II tumors in fully humanized BRGSF-HIS mice. In vivo efficacy studies

[0306] The synergistic effect between AI3S79 / N and CEAxCD3 was further evaluated in a second model in which the human pancreatic cancer cell line HPAF-II was subcutaneously transplanted into fully humanized BRGSF-HIS mice. 6 CEA+ / PD-L1+HPAF-II cells were transplanted into hFlt-3L enhanced BRGSF-HIS mice (genOway). Each group contained 2 mice for each of the 6 hematopoietic stem cell donors used, with a total of 12 mice in each group. When the average tumor volume reached 65mm3 (8th day after transplantation), treatment was started. Mice received 3 doses of vehicle, CEAxCD3 (10 mg / kg) or CEAxCD3+AI3S79 / N (10 mg / kg+10 mg / kg) combination on days 8, 11 and 14. Tumors were measured by digital calipers 2-3 times a week until the end of the experiment (tumor volume = 1500mm 3 ). Use the formula (length × width 2 )×0.5 to calculate the tumor volume.

[0307] In contrast to CEAxCD3 alone, which showed no antitumor activity in this model, the CEAxCD3+AI3S79 / N combination resulted in reduced tumor progression ( Figure 16A ). Tumor growth kinetics in individual mice were as follows Figure 16B shown.

[0308] Example 11: Single-dose and repeated-dose PK and drug tolerance studies in cynomolgus monkeys

[0309] A study was conducted in cynomolgus monkeys to investigate the PK and tolerability of AI3S79 / N after single and repeated intravenous injections. The study included three groups: (a) a single intravenous injection of AI3S79 / N at 0.5 mg / kg; (b) a single intravenous injection of AI3S79 / N at 10 mg / kg; and (c) repeated (n=2) intravenous injections of 10 mg / kg of AI3S79 / N, followed by histopathological examination.

[0310] Pharmacokinetics

[0311] Serum levels of AI3S79 / N were quantified in the serum of cynomolgus macaques using a validated universal pharmacokinetic assay based on Meso-Scale Discovery (MSD) technology. Briefly, biotinylated anti-human CH2 was coated on a streptavidin MSD plate to capture AI3S79 / N from the samples. Detection was performed by SulfoTag-conjugated anti-human CH2. Signals were acquired on a Meso Sector S600 instrument, and concentrations were extrapolated based on a standard curve for AI3S79 / N. Following quantification of AI3S79 / N in the samples, pharmacokinetic data were evaluated according to standard non-compartmental analysis using SAS software, version 9.4.

[0312] The concentration-time curve of AI3S79 / N obtained after bioanalytical testing is shown in Figure 2. Figure 17 AI3S79 / N PK was close to dose proportional, with mean C values significantly higher after 10 mg / kg compared to 0.5 mg / kg. max , AUC inf and AUC 0-168h The value is 13 to 23 times higher.

[0313] Accumulation of AI3S79 / N was observed after repeated injections at 10 mg / kg, as the mean C values on day 8 were significantly higher in monkeys treated twice compared to day 1. max and AUC 0-168h The values were 1.4 and 1.6 times higher, respectively.

[0314] Drug resistance

[0315] Single and repeated intravenous administration of AI3S79 / N were well tolerated. No deaths or abnormal clinical signs were observed throughout the study. No treatment-related changes were reported in body weight, food intake, clinical lesions, or autopsies.

[0316] Serum cytokines (IFNγ, TNFα, IL-2, IL-6, IL-8, and IL-10) were quantified using MesoScale Discovery (MSD) in samples collected from selected groups throughout the study. Figure 18 No increases in IFNγ, TNFα, IL-2, IL-8, and IL-10 levels were observed following single or repeated treatment with both dose levels of AI3S79 / N. A transient increase in IL-6 was observed at 4 or 8 hours after the first dose. For all three monkeys, the increase in IL-6 levels after the second dose of AI3S79 / N was more modest than after the first dose.

[0317] C-reactive protein (CRP) levels were measured in the serum of treated monkeys by ELISA (Life Diagnostics Inc.). After the first injection, a transient minimal increase in CRP was observed in most animals, starting 8 hours after administration, reaching a peak at 24 hours, and resolving after approximately 4-5 days. In animals that received two doses, a similar transient CRP increase was observed in only one animal after the second injection, while the other animal had a delayed and reduced maximum level, indicating that the AI3S79 / N-related increase in this marker was reduced after repeated injections ( Figure 19 ).

[0318] Overall, changes in serum biomarkers were limited to IL-6 and CRP, and the magnitude of the changes was mild and transient: levels in all animals returned to pre-dose levels within 24-48 hours (IL-6) or 2-3 days (CRP) without anti-inflammatory medication.

[0319] In conclusion, single and repeated intravenous administration of AI3S79 / N were well tolerated. No abnormalities in clinical signs and food intake were observed, and there were no treatment-related meaningful changes in body weight, clinical lesions, and anatomic lesions.

[0320] Other implementation plans

[0321] While the invention has been described in conjunction with its detailed description, the foregoing description is intended to illustrate and not limit the scope of the invention, which is defined by the scope of the appended claims. Other aspects, advantages and modifications are within the scope of the following claims.

[0322] References

[0323] Correnti,Colin E.,George S.Laszlo,Willem J.de van der Schueren,ColinD.Godwin,Ashok Bandaranayake,Melanie A.Busch,Chelsea J.Gudgeon,et al.2018.‘Simultaneous Multiple Interaction T-Cell Engaging(SMITE)BispecificAntibodies Overcome Bispecific T-Cell Engager(BiTE)Resistance via CD28 Co-Stimulation’.Leukemia 32(5):1239–43.https: / / doi.org / 10.1038 / s41375-018-0014-3.

[0324] Holliger,Philipp,Oliver Manzke,Mary Span,Robert Hawkins,BerndFleischmann,Liu Qinghua,Jürgen Wolf,Volker Diehl,Olivier Cochet,and GregWinter.1999.‘Carcinoembryonic Antigen(CEA)-Specific T-Cell Activation inColon Carcinoma Induced by Anti-CD3×Anti-CEA Bispecific Diabodies and B7×Anti-CEA Bispecific Fusion Proteins’.Cancer Research 59(12):2909–16.

[0325] Paula S.,Susanne Berr,Elita Avota,Shin-Young Na,ManuelaBattaglia,Ineke ten Berge,Hermann Einsele,and Thomas Hünig.2011.‘Precultureof PBMCs at High Cell Density Increases Sensitivity of T-Cell Responses,Revealing Cytokine Release by CD28 Superagonist TGN1412’.Blood118(26):6772–82.https: / / doi.org / 10.1182 / blood-2010-12-319780.

[0326] Skokos,Dimitris,Janelle C.Waite,Lauric Haber,Alison Crawford,AynurHermann,Erica Ullman,Rabih Slim,et al.2020.‘A Class of Costimulatory CD28-Bispecific Antibodies That Enhance the Antitumor Activity of CD3-BispecificAntibodies’.Science Translational Medicine 12(525):eaaw7888.https: / / doi.org / 10.1126 / scitranslmed.aaw7888.

[0327] Stebbings,Richard,Lucy Findlay,Cherry Edwards,David Eastwood,ChrisBird,David North,Yogesh Mistry,et al.2007.‘“Cytokine Storm”in the PhaseITrial of Monoclonal Antibody TGN1412:Better Understanding the CausestoImprove PreClinical Testing of Immunotherapeutics’.The JournalofImmunology179(5):3325–31.https: / / doi.org / 10.4049 / jimmunol.179.5.3325.

[0328] Waite,Janelle C.,Bei Wang,Lauric Haber,Aynur Hermann,Erica Ullman,Xuan Ye,Drew Dudgeon,et al.2020.‘Tumor-Targeted CD28 BispecificAntibodiesEnhance the Antitumor Efficacy of PD-1 Immunotherapy’.ScienceTranslationalMedicine 12(549):eaba2325.https: / / doi.org / 10.1126 / scitranslmed.aba2325.

Claims

1. A bispecific antibody, comprising: A first antigen-binding domain that binds to programmed death-ligand 1 (PD-L1); wherein the first antigen-binding domain comprises: i. a first heavy chain variable region comprising: a complementarity determining region 1 (CDR1) comprising the amino acid sequence of SEQ ID NO: 6, a complementarity determining region 2 (CDR2) comprising the amino acid sequence of SEQ ID NO: 7, and a complementarity determining region 3 (CDR3) comprising the amino acid sequence of SEQ ID NO: 8; and ii. a first light chain variable region, the first light chain variable region comprising: a CDR1 comprising the amino acid sequence of SEQ ID NO: 13, a CDR2 comprising the amino acid sequence of SEQ ID NO: 14, and a CDR3 comprising the amino acid sequence of SEQ ID NO: 15; and a second antigen binding domain that binds to CD28, wherein the second antigen binding domain comprises: i. a second heavy chain variable region comprising a CDR1 comprising the amino acid sequence of SEQ ID NO: 6, a CDR2 comprising the amino acid sequence of SEQ ID NO: 7, and a CDR3 comprising the amino acid sequence of SEQ ID NO:

8.

2. The bispecific antibody of claim 1, wherein the second antigen-binding domain comprises: A second light chain variable region, wherein the second light chain variable region comprises: a) CDR1 comprising the amino acid sequence of SEQ ID NO: 18; a CDR2 comprising the amino acid sequence of SEQ ID NO: 19; and a CDR3 comprising the amino acid sequence of SEQ ID NO: 20; or b) a CDR1 comprising the amino acid sequence of SEQ ID NO: 23; a CDR2 comprising the amino acid sequence of SEQ ID NO: 24; and A CDR3 comprising the amino acid sequence of SEQ ID NO:

25.

3. The bispecific antibody according to any one of the preceding claims, wherein the first heavy chain variable region and the second heavy chain variable region comprise the amino acid sequence of SEQ ID NO:

9. 4 . The bispecific antibody according to claim 1 , wherein the first heavy chain and the second heavy chain comprise the amino acid sequence of SEQ ID NO: 11 or SEQ ID NO:

12. 5 . The bispecific antibody according to claim 1 , wherein the first light chain variable region of portion ii comprises the amino acid sequence of SEQ ID NO:

16. The bispecific antibody according to claim 1 , wherein the first light chain of portion ii comprises the amino acid sequence of SEQ ID NO:

17.

7. The bispecific antibody according to claim 2, wherein 1. The second light chain variable region of part (a) comprises the amino acid sequence of SEQ ID NO: 21; and 2. The second light chain variable region of part (b) comprises the amino acid sequence of SEQ ID NO:

26.

8. The bispecific antibody according to claim 2, wherein 1. The second light chain of part (a) comprises the amino acid sequence of SEQ ID NO: 22; and 2. The second light chain of part (b) comprises the amino acid sequence of SEQ ID NO:

27.

9. The bispecific antibody according to any one of the preceding claims, wherein the first light chain is a kappa light chain and the second light chain is a lambda light chain.

10. The bispecific antibody according to any one of the preceding claims, wherein the first light chain is a lambda light chain and the second light chain is a kappa light chain.

11. The bispecific antibody of claim 1, wherein the bispecific antibody has an Fc domain comprising one or more amino acid substitutions that reduce binding to an activating Fc receptor and / or reduce effector function.

12. The bispecific antibody of claim 11, wherein the amino acid substitutions comprise L234A and L235A substitutions.

13. The bispecific antibody of claim 11, wherein the amino acid substitution comprises a P329A, a P329G, or a P329R substitution.

14. The bispecific antibody according to any one of the preceding claims, wherein the antibody has an IgG isotype.

15. The bispecific antibody according to any one of the preceding claims, wherein the antibody is a human antibody.

16. The bispecific antibody according to any one of the preceding claims, wherein the composition allows for PD-L1-dependent T cell activation.

17. A composition comprising the bispecific antibody according to any one of the preceding claims.

18. A method of reducing tumor cell proliferation and / or enhancing tumor cell killing, the method comprising contacting the cells with a composition according to claim 17.

19. A method of treating cancer in a subject, the method comprising administering to the subject the composition of claim 17.

Images