Anti-CD137 antibodies and methods of use thereof

Abstract

The present disclosure provides antigen-binding fragments thereof that bind to human CD137, multispecific antibodies that recognize CD137 as one antigen and at least one other antigen, pharmaceutical compositions comprising CD137 antibodies, and uses of the antibodies, multispecific antibodies or compositions to treat diseases such as cancer.

Description

[Technical field]

[0001] Disclosed herein are antibodies that specifically bind to both human and Macaca fascicularis CD137 (TNF receptor superfamily member 9 (TNFRSF9)) without cross-reactivity with other human TNF receptor members, as well as isolated nucleic acids, vectors and host cells. These antibodies can be used to construct multispecific antibodies with other modalities such as tumor-associated antigens, immune checkpoints, immune stimulators, etc. Finally, the anti-CD137 antibodies disclosed herein can be used to treat various cancers. [Background technology]

[0002] CD137, also known as TNF receptor superfamily member 9 (TNFRSF9), ILA, or 4-1BB, is a member of the TNF receptor superfamily and plays an important role in the clonal proliferation, survival, and development of T cells. CD137 is a 30 kDa type I membrane glycoprotein with an extracellular domain that contains four cysteine-rich pseudorepeats (CRDs), a short helical transmembrane domain, and a cytoplasmic signaling domain (Kwon et al., (1989) Proc Natl Acad Sci USA, 86, 1963-7).

[0003] CD137 is a marker for activated CD4 + and CD8 +It is expressed on various cell populations, including T cells, regulatory T cells (Tregs), dendritic cells (DCs), monocytes, mast cells, eosinophils, and tumor endothelial cells. Activation of CD137 plays an important role in the activation and survival of CD8+ T cells (Lee et al., (2002) J Immunol, 169, 4882-8; Pulle et al., (2006) J Immunol, 176, 2739-48). It maintains and enhances effector functions and induces Th1 cytokine production (Bartkowiak et al., (2015) Front Oncol, 5, 117; Shuford et al., (1997) J Exp Med, 186, 47-55). CD137 signaling leads to increased expression of pro-survival molecules through NF-κB pathway activation upon binding to its sole ligand, CD137 ligand (CD137L, 4-1BBL, or TNFSF9) (Wang et al. (2009) Immunol Rev, 229, 192-215).

[0004] Several studies have demonstrated that engagement of CD137, either by CD137L or agonistic antibodies, can inhibit tumor growth by promoting T cell activity (Dubrot et al., (2010) Cancer Immunol Immunother, 59, 1223-33; Gauttier et al., (2014) Int J Cancer, 135, 2857-67; Sallin et al., (2014) Cancer Immunol Immunother, 63, 947-58; McMillin et al., (2006) Hum Gene Ther, 17, 798-806). The effect of CD137 activation on inhibition of activation-induced cell death (AICD) has been demonstrated both in vitro (Hurtado et al., (1997) J Immunol, 158, 2600-9) and in vivo (Takahashi et al., (1999) J Immunol, 162, 5037-40).

[0005] Both CD4+ and CD8+ T cells have been shown to respond to CD137 stimulation, but the enhancement of T cell function appears to be greater in CD8+ cells (Shuford et al., (1997) J Exp Med, 186, 47-55; Gramaglia et al., (2000) Eur J Immunol, 30, 392-402).

[0006] In addition, CD137 agonists can synergize with several immunomodulatory agents, including CpG, TRAIL, CD40, OX40, DR5, PD-1 / PD-L1, CTLA4, Tim3, IL-2, and IL-12 (Taraban et al., (2002) Eur J Immunol, 32, 3617-27; Curran et al., (2011) PLoS One, 6, e19499; Gray et al., (2008) Eur J Immunol, 38, 2499-511; Wei et al., (2013) PLoS One, 8, e84927; Guo et al., (2013) J Transl Med, 11, 215; Kwong et al., (2013) Cancer Res, 73, 1547-58; Lee et al., (2013) Cancer Res, 73, 1547-58; Lee et al., (2013) Cancer Res, 73, 1547-58). al., (2004) J Immunother, 27, 201-10). CD137 agonists have also been demonstrated to ameliorate autoimmunity in animal models of lupus, collagen-induced arthritis, and experimental autoimmune encephalomyelitis (Vinay et al., (2006) J Immunol, 177, 5708-17).

[0007] Several CD137 antibodies are in clinical development. Urelumab (BMS-66513) is a fully human non-ligand blocking IgG4 antibody developed by Bristol-Myers Squibb. Several phase I and II trials in various indications are currently ongoing. Severe hepatotoxicity (grade IV hepatitis) has been observed in phase I and II trials (NCT00309023, NCT00612664, NCT01471210) with urelumab (Segal et al., (2017) Clin Cancer Res, 23, 1929-1936: Timmerman et al., (2020) Am J Hematol, 95, 510-520; Chester et al., Blood, 131, 49-57). Utomirumab is a ligand-blocking IgG2 antibody with reduced toxicity and few grade III-IV side effects, with no dose-limiting toxicities reported at doses up to 10 mg / kg (Fisher et al., (2012) Cancer Immunol Immunother, 61, 1721-33; Segal et al., (2018) Clin Cancer Res, 24, 1816-1823; Gopal et al., (2020) Clin Cancer Res, 26, 2524-2534).

[0008] It has been shown that the production of interleukin-27 by activation of CD137 on liver myeloid cells is essential for hepatotoxicity (Bartkowiak et al., (2018) Clin Cancer Res, 24, 1138-1151). To avoid systemic toxicity due to CD137 activation while maintaining good antitumor efficacy in the tumor microenvironment, tumor-associated antigen (TAA)-directed CD137 multispecific antibodies can reduce toxicity. Without being bound to any one mechanism of action, CD137xTAA multispecific antibodies crosslink CD137 receptors only when TAA is also present, causing stimulation of immune cells in the tumor microenvironment. Therefore, CD137xTAA multispecific antibodies can significantly reduce the possibility of systemic toxicity.

[0009] There is no approved therapeutic antibody against CD137, and there is still an unmet medical need for therapies targeting CD137. The present disclosure includes specific antibodies and antibody fragments against human CD137. Furthermore, the CD137 VH domain fragments disclosed herein can be used to construct multispecific antibodies with other modalities such as TAA, immune checkpoints or immune stimulators. CD137 antibodies, alone or in combination with other modality antibodies, may be used to treat or prevent cancer, autoimmune disease, or infectious disease. Summary of the Invention [Means for solving the problem]

[0010] The present disclosure is directed to anti-CD137 antibodies and antigen-binding antibody fragments thereof that specifically bind to CD137.

[0011] In one embodiment, the disclosure provides an antibody, or antigen-binding fragment thereof, that binds to human CD137.

[0012] The present disclosure encompasses the following embodiments.

[0013] An antibody or antigen-binding antibody fragment thereof that specifically binds to human CD137.

[0014] An antibody antigen-binding fragment that specifically binds to human CD137, (i) a heavy chain variable region comprising (a) an HCDR1 (heavy chain complementarity determining region 1) of SEQ ID NO: 14, (b) an HCDR2 of SEQ ID NO: 29, and (c) an HCDR3 of SEQ ID NO: 30; (ii) a heavy chain variable region comprising (a) an HCDR1 of SEQ ID NO: 14, (b) an HCDR2 of SEQ ID NO: 22, and (c) an HCDR3 of SEQ ID NO: 16; (iii) a heavy chain variable region comprising (a) an HCDR1 of SEQ ID NO: 14, (b) an HCDR2 of SEQ ID NO: 15, and (c) an HCDR3 of SEQ ID NO: 16; or (vi) a heavy chain variable region comprising (a) an HCDR1 of SEQ ID NO: 4, (b) an HCDR2 of SEQ ID NO: 5, and (c) an HCDR3 of SEQ ID NO: 6. The antibody antigen-binding fragment comprising:

[0015] (i) a heavy chain variable region (VH) comprising an amino acid sequence at least 90, 91, 92, 93, 94, 95, 96, 97, 98, or 99% identical to SEQ ID NO: 33; (ii) a heavy chain variable region (VH) comprising an amino acid sequence at least 90, 91, 92, 93, 94, 95, 96, 97, 98 or 99% identical to SEQ ID NO:24; (iii) a heavy chain variable region (VH) comprising an amino acid sequence at least 90, 91, 92, 93, 94, 95, 96, 97, 98, or 99% identical to SEQ ID NO:19; (iv) a heavy chain variable region (VH) comprising an amino acid sequence at least 90, 91, 92, 93, 94, 95, 96, 97, 98 or 99% identical to SEQ ID NO:9; or (v) a heavy chain variable region (VH) comprising an amino acid sequence at least 90, 91, 92, 93, 94, 95, 96, 97, 98 or 99% identical to SEQ ID NO: 103 The antibody antigen-binding fragment comprising:

[0016] The antibody antigen-binding fragment, wherein 1, 2, 3, 4, 5, 6, 7, 8, 9, or 10 amino acids within SEQ ID NO: 33, 24, 19, 9, or 103 are inserted, deleted, or substituted.

[0017] (i) a heavy chain variable region (VH) comprising SEQ ID NO: 33; (ii) a heavy chain variable region (VH) comprising SEQ ID NO: 24; (iii) a heavy chain variable region (VH) comprising SEQ ID NO: 19; (iv) a heavy chain variable region (VH) comprising SEQ ID NO:9; (iv) a heavy chain variable region (VH) comprising SEQ ID NO: 103 The antibody antigen-binding fragment comprising:

[0018] Heavy chain (scFv), heavy chain Fab fragment, heavy chain Fab' fragment, or heavy chain F(ab') 2 The antibody antigen-binding fragment, which is a fragment.

[0019] (i) a heavy chain variable region comprising (a) an HCDR1 (heavy chain complementarity determining region 1) of SEQ ID NO: 14, (b) an HCDR2 of SEQ ID NO: 29, and (c) an HCDR3 of SEQ ID NO: 30; (ii) a heavy chain variable region comprising (a) an HCDR1 of SEQ ID NO: 14, (b) an HCDR2 of SEQ ID NO: 22, and (c) an HCDR3 of SEQ ID NO: 16; (iii) a heavy chain variable region comprising (a) an HCDR1 of SEQ ID NO: 14, (b) an HCDR2 of SEQ ID NO: 15, and (c) an HCDR3 of SEQ ID NO: 16; or (iv) at least a first antigen-binding domain that specifically binds to human CD137, comprising a heavy chain variable region comprising (a) an HCDR1 of SEQ ID NO: 4, (b) an HCDR2 of SEQ ID NO: 5, and (c) an HCDR3 of SEQ ID NO: 6; and A multispecific antibody comprising at least a second antigen-binding domain that specifically binds to a human tumor-associated antigen (TAA).

[0020] The first antigen-binding domain comprises: (i) a heavy chain variable region (VH) comprising SEQ ID NO: 33; (ii) a heavy chain variable region (VH) comprising SEQ ID NO: 24; (iii) a heavy chain variable region (VH) comprising SEQ ID NO: 19; (iv) a heavy chain variable region (VH) comprising SEQ ID NO:9; (iv) a heavy chain variable region (VH) comprising SEQ ID NO: 103; The multispecific antibody comprises at least a second antigen-binding domain that specifically binds to a human tumor-associated antigen (TAA).

[0021] The multispecific antibody is a bispecific antibody.

[0022] The bispecific antibody, wherein the bispecific is in a 1+1 format.

[0023] The bispecific antibody, wherein the bispecific is in a 1+2 format.

[0024] The bispecific antibody, wherein the bispecific is in a 2+2 format.

[0025] The bispecific antibody, wherein the bispecific is in a 2+2 format.

[0026] The bispecific antibody, wherein the linker is any of the sequences of SEQ ID NO: 239 to SEQ ID NO: 280.

[0027] The bispecific antibody, wherein the linker is SEQ ID NO: 246.

[0028] The bispecific antibody, wherein the linker is SEQ ID NO: 251.

[0029] The antibody or the antibody fragment has antibody-dependent cellular cytotoxicity (ADCC) or complement-dependent cytotoxicity (CDC).

[0030] The antibody or antibody fragment thereof, which has reduced glycosylation, no glycosylation or is hypofucosylated.

[0031] The antibody or antibody fragment comprising increased bisecting GlcNac structures.

[0032] The antibody or antibody fragment, wherein the Fc domain is IgG1.

[0033] The antibody or the antibody fragment, wherein the Fc domain is IgG4.

[0034] 16. A pharmaceutical composition comprising the antibody or antibody fragment according to any one of claims 1 to 15, further comprising a pharma- ceutically acceptable carrier.

[0035] A method of treating cancer comprising administering to a patient in need thereof an effective amount of said antibody or said antibody fragment.

[0036] The method, wherein the cancer is gastric cancer, colon cancer, pancreatic cancer, breast cancer, head and neck cancer, kidney cancer, liver cancer, lung cancer, small cell lung cancer, non-small cell lung cancer, ovarian cancer, skin cancer, mesothelioma, lymphoma, leukemia, myeloma and sarcoma.

[0037] The above methods, wherein the antibody or antibody fragment is administered in combination with another therapeutic agent.

[0038] The method, wherein the therapeutic agent is paclitaxel or a paclitaxel agent, docetaxel, carboplatin, topotecan, cisplatin, irinotecan, doxorubicin, lenalidomide, or 5-azacytidine.

[0039] The method, wherein the therapeutic agent is an anti-PD-1 antibody.

[0040] An isolated nucleic acid encoding said antibody or said antibody fragment.

[0041] A vector comprising the nucleic acid.

[0042] A host cell comprising said nucleic acid or said vector.

[0043] A method of producing said antibody or said antibody fragment, comprising culturing said host cell and recovering the antibody or antibody fragment from the culture.

[0044] In one embodiment, the antibody or antigen-binding fragment thereof comprises one or more complementarity determining regions (CDRs) comprising an amino acid sequence selected from the group consisting of SEQ ID NO:4, SEQ ID NO:5, SEQ ID NO:6, SEQ ID NO:14, SEQ ID NO:15, SEQ ID NO:16, SEQ ID NO:22, SEQ ID NO:29, and SEQ ID NO:30.

[0045] In another embodiment, the antibody or antigen-binding fragment thereof comprises a heavy chain variable region comprising one or more heavy chain complementarity determining regions (HCDRs) comprising an amino acid sequence selected from the group consisting of SEQ ID NO:4, SEQ ID NO:5, SEQ ID NO:6, SEQ ID NO:14, SEQ ID NO:15, SEQ ID NO:16, SEQ ID NO:22, SEQ ID NO:29, and SEQ ID NO:30.

[0046] In another embodiment, the antibody or antigen-binding fragment thereof comprises a heavy chain variable region comprising three heavy chain complementarity determining regions (HCDRs): HCDR1 comprising the amino acid sequence of SEQ ID NO:4 or SEQ ID NO:14, HCDR2 comprising the amino acid sequence of SEQ ID NO:5, SEQ ID NO:15, SEQ ID NO:22 or SEQ ID NO:29, and HCDR3 comprising the amino acid sequence of SEQ ID NO:6, SEQ ID NO:16 or SEQ ID NO:30.

[0047] In another embodiment, the antibody or antigen-binding fragment thereof comprises a heavy chain variable region comprising three heavy chain complementarity determining regions: HCDR1 comprising the amino acid sequence of SEQ ID NO: 4, HCDR2 comprising the amino acid sequence of SEQ ID NO: 5, and HCDR3 comprising the amino acid sequence of SEQ ID NO: 6; or three heavy chain complementarity determining regions: HCDR1 comprising the amino acid sequence of SEQ ID NO: 14, HCDR2 comprising the amino acid sequence of SEQ ID NO: 15, and HCDR3 comprising the amino acid sequence of SEQ ID NO: 16; or three heavy chain complementarity determining regions (HCDRs): HCDR1 comprising the amino acid sequence of SEQ ID NO: 14, HCDR2 comprising the amino acid sequence of SEQ ID NO: 22, and HCDR3 comprising the amino acid sequence of SEQ ID NO: 16; or three heavy chain complementarity determining regions (HCDRs): HCDR1 comprising the amino acid sequence of SEQ ID NO: 14, HCDR2 comprising the amino acid sequence of SEQ ID NO: 29, and HCDR3 comprising the amino acid sequence of SEQ ID NO: 30.

[0048] In one embodiment, an antibody or antigen-binding fragment thereof of the present disclosure comprises: (a) a heavy chain variable region having an amino acid sequence of SEQ ID NO:33, SEQ ID NO:19, SEQ ID NO:24, or SEQ ID NO:103, or an amino acid sequence at least 95%, 96%, 97%, 98% or 99% identical to any one of SEQ ID NO:33, SEQ ID NO:19, SEQ ID NO:24, or SEQ ID NO:103.

[0049] In another embodiment, an antibody or antigen-binding fragment thereof of the present disclosure comprises a heavy chain variable region comprising the amino acid sequence of SEQ ID NO:9, SEQ ID NO:19, SEQ ID NO:24, SEQ ID NO:33, or SEQ ID NO:103, or an amino acid sequence having one, two, or three amino acid substitutions in the amino acid sequence of SEQ ID NO:9, SEQ ID NO:19, SEQ ID NO:24, SEQ ID NO:33, or SEQ ID NO:103.

[0050] In one embodiment, an antibody or antigen-binding fragment thereof of the present disclosure comprises: (a) a heavy chain variable region comprising the amino acid sequence of SEQ ID NO:9; (b) a heavy chain variable region comprising the amino acid sequence of SEQ ID NO: 19; (c) a heavy chain variable region comprising the amino acid sequence of SEQ ID NO: 24; (c) a heavy chain variable region comprising the amino acid sequence of SEQ ID NO: 33; (d) a heavy chain variable region comprising the amino acid sequence of SEQ ID NO: 103 Includes.

[0051] In one embodiment, the antibodies of the disclosure are of the IgG1, IgG2, IgG3, or IgG4 isotype. In more specific embodiments, the antibodies of the disclosure comprise a wild-type human IgG1 (also called human IgG1wt or huIgG1) or IgG2 Fc domain. In another embodiment, the antibodies of the disclosure comprise a human IgG4 Fc domain with S228P and / or R409K substitutions (according to the EU numbering system).

[0052] In one embodiment, the antibody of the disclosure is administered at a concentration of 1×10 -6 M~1×10 -10 Binding affinity of M (K D In another embodiment, the antibody of the disclosure binds to CD137 at about 1×10 -6 M, about 1 x 10 -7 M, about 1 x 10 -8 M, about 1 x 10 -9 M, or approximately 1 x 10 -10 Binding affinity of M (K D ) and binds to CD137.

[0053] In another embodiment, the anti-human CD137 antibodies of the present disclosure exhibit cross-species binding activity for cynomolgus monkey CD137.

[0054] In one embodiment, the antibodies of the disclosure have potent Fc-mediated effector function: the antibodies mediate antibody-dependent cellular cytotoxicity (ADCC) against target cells expressing CD137.

[0055] The present disclosure relates to an isolated nucleic acid comprising a nucleotide sequence encoding the amino acid sequence of an antibody or antigen-binding fragment. In one embodiment, the isolated nucleic acid comprises a VH nucleotide sequence of SEQ ID NO: 10, SEQ ID NO: 20, SEQ ID NO: 25, SEQ ID NO: 34, or SEQ ID NO: 104, or a nucleotide sequence having at least 95%, 96%, 97%, 98%, or 99% identity to SEQ ID NO: 10, SEQ ID NO: 20, SEQ ID NO: 25, SEQ ID NO: 34, or SEQ ID NO: 104, and encodes the VH region of an antibody or antigen-binding fragment of the present disclosure.

[0056] In another aspect, the present disclosure relates to a pharmaceutical composition comprising a CD137 antibody or antigen-binding fragment thereof, and optionally a pharma- ceutically acceptable excipient.

[0057] In yet another aspect, the present disclosure relates to a method of treating a disease in a subject, comprising administering a therapeutically effective amount of a CD137 antibody or antigen-binding antibody fragment thereof, or a CD137 antibody pharmaceutical composition to a subject in need thereof, hi another embodiment, the disease treated by the antibody or antigen-binding fragment is cancer.

[0058] The present disclosure relates to the use of antibodies or antigen-binding antibody fragments thereof, or CD137 antibody pharmaceutical compositions to treat diseases such as cancer. [Brief description of the drawings]

[0059] [Figure 1A]FIG. 1 shows an overview of human anti-huCD137 VH domain antibodies identified from each sub-library. [Figure 1B] Graph showing phylogenetic tree of human anti-huCD137 VH domain antibodies from each sub-library. VH sequences of candidate anti-huCD137 VH domain antibodies were aligned using DNASTAR's Megaalign™ software. Sequence homology is displayed in the phylogenetic tree. [Diagram 2] (A) is a schematic diagram of the human Fc-fused VH antibody format (VH-Fc). The VH domain antibody was fused to the N-terminus of an inactive Fc (no FcγR binding) with a G4S linker in between. (B) shows representative screening results using supernatants containing VH-Fc protein. (C) demonstrates that one of the clones, BGA-4712, can stimulate IL-2 production in Hut78 / huCD137 cells in a dose-dependent manner. [Diagram 3] A-C are representative binding profiles of anti-huCD137 VH domain antibody BGA-4712. A shows the affinity measurement of purified human anti-huCD137 VH domain antibody BGA-4712 by surface plasmon resonance (SPR). B shows the measurement of human anti-huCD137 VH domain antibody BGA-4712 binding by flow cytometry. C shows blocking of human anti-huCD137 VH domain antibody BGA-4712 by huCD137 ligand (human CD137 ligand-ECD-mIgG2a fusion protein) interaction. Binding of purified human anti-huCD137 VH domain antibody BGA-4712 to CD137-expressing Hut78 / huCD137 cells (Hut78 / huCD137) was measured by flow cytometry. [Figure 4] 5A-D are schematic diagrams of exemplary CD137xCEA multispecific antibody formats. [Diagram 5]Figures A to B show a comparison of cell binding of CD137xCEA multispecific antibodies by flow cytometry: Figure A shows binding to CEA-expressing cells CT26 / CEA; Figure B shows binding to CD137-expressing cells Hut78 / huCD137. [Figure 6] Figures A-B demonstrate that A-CD137xCEA stimulates PBMCs to produce IFN-γ in the presence of CEA+ tumor cells. Figure A shows that A-CD137xCEA, a CD137xCEA multispecific antibody, induces the CD137-expressing cell line Hut78 / huCD137 to produce Il-2. Figure B shows that A-CD137xCEA, a CD137xCEA multispecific antibody, induces human peripheral blood mononuclear cells (PBMCs) to produce IFN-γ in a dose-dependent manner. [Figure 7] FIG. 13: Binding of the optimized A-CD137xCEA BGA-4712 variant to CD137 expressing cells Hut78 / huCD137 by flow cytometry. [Figure 8] FIG. 1 shows CD137 activation in a PBMC-based cytokine release assay, demonstrating that function of A-CD137xCEA-M3 is maintained following removal of the PTM sites. [Figure 9] A is a schematic diagram of phage display CH3 fusion of anti-huCD137VH domain antibody. The VH domain antibody is fused to the C-terminus of CH3, mimicking the bispecific format A-CD137xCEA. B is a diagram demonstrating by ELISA that the supernatant containing CH3-BGA-4712-M3 can bind to CD137 (human CD137-ECD mIgG2a). BGA-4712 (VH-Fc) is used as a positive control and huIgG is used as a negative control. CH3-BGA-4712-M3 with W47G or W47F or W47Y mutations in the VH region could not bind to pre-coated human CD137-ECD-mIgG2a. [Figure 10] FIG. 1 shows sequence logos of the CDR regions of BGA-4712-M3 after four rounds of selection. [Figure 11] FIG. 1 shows a binding assay of anti-huCD137 VH domain antibody BGA-5623 by flow cytometry, demonstrating that binding to CD137 is improved after affinity maturation. [Figure 12] (A-B) Epitope mapping of human anti-huCD137 VH domain antibody BGA-5623. (A) Representative screening results in a cell-based binding assay. Expression of huCD137 variants was monitored by urelumab analogs. (B) BGA-5623 binding of purified huCD137 variants. [Figure 13] FIG. 1 shows molecular modelling of the huCD137 monomer. [Figure 14] (A) demonstrates that CD137 ligands compete with the human anti-huCD137 VH domain antibody BGA-5623 by ELISA, and (B) demonstrates that the CD137xCEA multispecific antibody BGA-5623 can reduce the CD137 / CD137 ligand interaction in a cell-based ligand competition assay. [Figure 15] FIG. 13 demonstrates the lack of off-target binding of BGA-5623 to other TNF receptor family members by ELISA. [Figure 16] AB show affinity measurements of purified BGA-7556 mutants by surface plasmon resonance (SPR). [Figure 17] FIG. 1 is a schematic diagram of the CD137xCEA multispecific antibody format to investigate other parameters such as module ratio that may affect CD137 activation in vitro. [Figure 18] FIG. 1 demonstrates that the bispecific antibody A-41A11 / 41A11 with a modular ratio of 2:4 is able to activate CD137 regardless of the presence or absence of CEA+ tumor cells. [Figure 19] FIG. 1 is a schematic diagram of the CD137xCEA multispecific antibody format for investigating other parameters such as Fc function and module orientation that may influence CD137 activation in vitro. [Figure 20] FIG. 1 demonstrates that only the CD137xCEA multispecific antibodies studied stimulate PBMCs to produce IFN-γ in the presence of CEA+ tumor cells. [Figure 21] FIG. 1 demonstrates that linker length has minimal effect on CD137 activation in vitro in the presence of CEA+ tumor cells. [Figure 22] FIG. 1 shows that format A-BGA-5623, but not A-IgG1-BGA-5623, induces a significant inhibition of tumor growth in vivo. [Figure 23] FIG. 1 is a schematic diagram of the designed tumor-targeting TAA×CD137 multispecific antibody format. [Figure 24] AB show the binding of BE-146 to MKN45 cells (A) and Hut78 / huCD137 cells (B). CD show the binding of BE-830 to HepG2 cells (C) and Hut78 / huCD137 cells (D). [Diagram 25] Figures A-C demonstrate that the CEA x CD137 multispecific antibody BE-146 induces IL-2 and IFN-γ release from human PBMCs. A is a schematic of CD137 activation by costimulating huPBMCs with BE-146 cells and HEK293 / OS8 cells in the presence of MKN45 cells. B-C show that BE-146 can induce IL-2 (B) and IFN-γ (C) from human PBMCs. PBMCs from two donors were tested. Results are shown as mean ± SD of duplicates. [Figure 26] Figures A-B demonstrate that the CEA x CD137 multispecific antibody BE-146 induces IL-2 and IFN-γ release from human T cells. Figure B shows that BE-146 can induce IL-2 and IFN-γ (C) from human PBMCs. PBMCs from two donors were tested. Results are presented as mean ± SD of duplicates. [Figure 27]Figures A-B demonstrate that the CEAxCD137-induced response is CEA-dependent. A shows that BE-146 was able to induce significant IL-2 and IFN-γ release from PBMCs against CEA-overexpressing HEK293 cells (B), but not against non-CEA-transduced HEK293 cells. PBMCs from three donors were tested. Results are shown as mean ± SD of duplicates. [Figure 28] A-B show that CD137xCEA-induced responses are not significantly blocked by recombinant soluble CEA. Results show that BE-146-induced IL-2 (A) and IFN-γ (B) release from PBMCs was not significantly blocked by soluble CEA at 50 ng / ml or 500 ng / ml. PBMCs from two donors were tested. Results are presented as mean ± SD of duplicates. [Figure 29] FIG. 1 shows that BE-146 and urelumab analogs induce significant inhibition of tumor growth. [Diagram 30] FIG. 1 shows that the combination of BE-146 with anti-PD-1 induces a significantly increased anti-tumor effect. [Diagram 31] BE-146 has no hepatotoxicity in vivo. High doses of urelumab analogs, but not BE-146, induced significant increases in alanine transaminase (ALT) and aspartate aminotransferase (AST) concentrations and increased inflammatory cell infiltration in the liver. [Diagram 32] Figure 1 shows that Claudin6×CD137 (BE-268) induces IFN-γ release from human PBMC. Cancer cell lines with different Claudin6 expression levels were used. PA-1 has high Claudin6 expression, Bewo has intermediate Claudin6 expression, and MKN45 has negative Claudin6 expression. BE-268-induced IFN-γ release correlates with the expression level of Claudin6. [Diagram 33]Figure 2: Trop2xCD137 (BE-907) induces IFN-γ release from human PBMC. PBMC from two donors were tested. Results are presented as mean ± SD of duplicates. [Diagram 34] 5A-B demonstrate that the GPC3×CD137 multispecific antibody BE-830 induces significant cytokine release, such as IL-2 (A) and IFN-γ (B), from human PBMC in the presence of GPC3. [Diagram 35] FIG. 1 shows that BE-830 and urelumab analogs induce significant inhibition of tumor growth. [Diagram 36] Figure 1 shows partial competitive binding of VHH(BGA-5623) to CD137L for CD137. The crystal structure of VHH(BGA-5623) / CD137 was superimposed with the CD137L / CD137 complex (PDB:6MGP) via CD137. CD137, CD137L, and VHH are colored black, white, and gray, respectively. [Figure 37] Figure 1 shows that CDR3 of VHH(BGA-5623) undergoes dramatic conformational changes upon CD137 binding. CD137-binding VHH(BGA-5623) in black is overlaid with apoVHH(BGA-5623) in white. [Figure 38] Figure 1 shows atomic interactions on the binding surface of the VHH(BGA-5623) / CD137 complex. The binding interface between VHH(BGA-5623) and CD137 identifies certain key residues of BGA-5623 (paratope residues) and CD137 (epitope residues). The CRD1 and CRD2 domains of CD137 are shown in grey paint covered by a white transparent surface. The paratope residues are coloured black. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0060] definition Unless specifically defined elsewhere in this document, all other technical and scientific terms used herein have the meanings commonly understood by those of ordinary skill in the art.

[0061] As used in this specification, including the appended claims, singular words such as "a," "an," and "the" include references to their corresponding plurals unless the context clearly dictates otherwise.

[0062] The term "or" is used to mean and is used synonymously with the term "and / or," unless context clearly dictates otherwise.

[0063] The term "anti-cancer agent" as used herein refers to any agent that can be used to treat a cell proliferative disorder such as cancer, including, but not limited to, cytotoxic agents, chemotherapeutic agents, radiotherapy and radiotherapeutic agents, targeted anti-cancer agents, and immunotherapeutic agents.

[0064] The term "CD137" or "TNFRSF9", "ILA" or "41BB" refers to the amino acid sequence of human CD137 (SEQ ID NO: 47), which can also be found under Accession No. Q07011 (TNR9_HUMAN) or U03397. The nucleic acid sequence of CD137 is set forth in SEQ ID NO: 48.

[0065] The terms "administration", "administering", "treating" and "treatment" as used herein, when applied to an animal, human, experimental subject, cell, tissue, organ, or biological fluid, refer to the contact of an exogenous pharmaceutical, therapeutic, diagnostic, or composition to the animal, human, subject, cell, tissue, organ, or biological fluid. Treatment of a cell encompasses contact of a reagent to the cell as well as contact of a reagent to a fluid, which in turn contacts the cell. The terms "administration" and "treatment" also refer to in vitro and ex vivo treatment, e.g., of a cell with a reagent, diagnostic, binding compound, or with another cell. The term "subject" herein includes any organism, preferably an animal, more preferably a mammal (e.g., rat, mouse, dog, cat, rabbit), and most preferably a human. In one aspect, treating any disease or disorder refers to alleviating the disease or disorder (i.e., delaying or preventing or reducing the onset of the disease or at least one of its clinical symptoms). In another embodiment, "treat", "treating", or "treatment" refers to alleviating or improving at least one physical parameter, including those that may not be discernible by the patient. In yet another embodiment, "treat", "treating", or "treatment" refers to modulating a disease or disorder, either physically (e.g., stabilizing a discernible symptom), physiologically (e.g., stabilizing a physical parameter), or both. In yet another embodiment, "treat", "treating", or "treatment" refers to preventing or delaying the onset or development or progression of a disease or disorder.

[0066] The term "subject" in the context of this disclosure is a mammal, e.g., a primate, preferably a higher primate, e.g., a human (e.g., a patient having or at risk of having a disorder described herein).

[0067] The term "affinity" as used herein refers to the strength of interaction between an antibody and an antigen. Within an antigen, the variable region of an antibody interacts with the antigen at many sites through non-covalent forces. In general, the more interactions, the stronger the affinity.

[0068] The term "antibody" as used herein refers to a polypeptide of the immunoglobulin family that can non-covalently, reversibly, and specifically bind to a corresponding antigen. For example, naturally occurring IgG antibodies are tetramers that contain at least two heavy (H) chains and two light (L) chains interconnected by disulfide bonds. Each heavy chain consists of a heavy chain variable region (abbreviated herein as VH) and a heavy chain constant region. The heavy chain constant region consists of three domains, CH1, CH2, and CH3. Each light chain consists of a light chain variable region (abbreviated herein as VL) and a light chain constant region. The light chain constant region consists of one domain, CL. The VH and VL regions can be further subdivided into regions of hypervariability, termed complementarity determining regions (CDRs), interspersed with more conserved regions, termed framework regions (FRs). Each VH and VL is composed of three CDRs and four framework regions (FRs), arranged from the amino terminus to the carboxy terminus in the following order: FR1, CDR1, FR2, CDR2, FR3, CDR3, and FR4. The variable regions of the heavy and light chains contain a binding domain that interacts with an antigen. The constant region of the antibody can mediate the binding of the immunoglobulin to host tissues or factors, including various cells of the immune system (e.g., effector cells) and the first component (C1q) of the classical complement system.

[0069] The term "antibody" includes, but is not limited to, monoclonal antibodies, human antibodies, humanized antibodies, chimeric antibodies, and anti-idiotypic (anti-Id) antibodies. Antibodies can be of any isotype / class (e.g., IgG, IgE, IgM, IgD, IgA, and IgY) or subclass (e.g., IgG1, IgG2, IgG3, IgG4, IgA1, and IgA2).

[0070] In some embodiments, the anti-CD137 antibody comprises at least one antigen binding site, at least the variable region. In some embodiments, the anti-CD137 antibody comprises an antigen binding fragment from a CD137 antibody described herein. In some embodiments, the anti-CD137 antibody is isolated or recombinant.

[0071] The term "monoclonal antibody" or "mAb" or "Mab" as used herein refers to a population of substantially homogenous antibodies, i.e., the antibody molecules in the population are identical in amino acid sequence, except for possible naturally occurring mutations that may be present in minor amounts. In contrast, conventional (polyclonal) antibody preparations typically contain a large number of different antibodies that contain different amino acid sequences within the variable domains, particularly the complementarity determining regions (CDRs), which are often specific for different epitopes. The modifier "monoclonal" indicates the character of the antibody as being obtained from a population of substantially homogenous antibodies and is not to be construed as requiring production of the antibody by any particular method. Monoclonal antibodies (mAbs) can be obtained by methods known to those skilled in the art. See, for example, Kohler et al., Nature 1975 256:495-497; US Pat. No. 4,376,110; Ausubel et al., CURRENT PROTOCOLS IN MOLECULAR BIOLOGY 1992; Harlow et al., ANTIBODIES: A LABORATORY MANUAL, Cold Spring Harbor Laboratory 1988; and Colligan et al., CURRENT PROTOCOLS IN IMMUNOLOGY 1993. The antibodies disclosed herein can be of any immunoglobulin class, including IgG, IgM, IgD, IgE, IgA, and any subclass thereof, such as IgG1, IgG2, IgG3, IgG4, etc. Hybridomas producing monoclonal antibodies can be cultured in vitro or in vivo. High titer monoclonal antibodies can be obtained by in vivo production by intraperitoneally injecting cells from individual hybridomas into mice (such as primed Balb / c mice) to produce ascites fluid containing high concentrations of the desired antibodies. Monoclonal antibodies of isotype IgM or IgG can be purified from such ascites fluid or culture supernatants using column chromatography methods well known to those skilled in the art.

[0072] Generally, the basic antibody structural unit comprises a tetramer. Each tetramer comprises two identical pairs of polypeptide chains, each pair comprising one "light chain" (about 25 kDa) and one "heavy chain" (about 50-70 kDa). The amino-terminal portion of each chain comprises a variable region of about 100-110 or more amino acids in length that is primarily responsible for antigen recognition. The carboxy-terminal portion of the heavy chain may define a constant region that is primarily responsible for effector function. Human light chains are typically classified as kappa light chains and lambda light chains. Furthermore, human heavy chains are typically classified as α, δ, ε, γ, or μ, and the antibody isotypes are defined as IgA, IgD, IgE, IgG, and IgM, respectively. Within the light and heavy chains, the variable and constant regions are linked by a "J" region of about 12 or more amino acids, and the heavy chain also comprises a "D" region of about 10 or more amino acids.

[0073] The variable regions of each light / heavy chain (VL / VH) pair form the antibody binding site. Thus, an intact antibody generally has two binding sites. Except for bifunctional or bispecific antibodies, the two binding sites are generally identical in primary sequence.

[0074] Typically, both heavy and light chain variable domains contain three hypervariable regions, also called "complementarity determining regions (CDRs)", located between relatively conserved framework regions (FRs). The CDRs are usually aligned by the framework regions and allow binding to a specific epitope. Generally, from N-terminus to C-terminus, both light chain variable domains and heavy chain variable domains contain FR-1 (or FR1), CDR-1 (or CDR1), FR-2 (FR2), CDR-2 (CDR2), FR-3 (or FR3), CDR-3 (CDR3), and FR-4 (or FR4). The locations of CDRs and framework regions can be determined using various definitions well known in the art, such as Kabat, Chothia, AbM and IMGT (e.g., Johnson et al., Nucleic Acids Res., 29:205-206 (2001); Chothia and Lesk, J. Mol. Biol., 196:901-917 (1987); Chothia et al., Nature, 342:877-883 (1989); Chothia et al., J. Mol. Biol., 227:799-817 (1992); Al-Lazikani et al., J. Mol. Biol., 273:927-748 (1997) ImMunoGenTics (IMGT) numbering (Lefranc, M.-P., The Immunologist, 7, 132-136 (1999); Lefranc, M.-P. et al., Dev. Comp. Immunol., 27, 55-77 (2003) (see "IMGT" numbering scheme)).Definitions of antigen-binding sites are also described in: Ruiz et al., Nucleic Acids Res., 28:219-221 (2000); and Lefranc, MP, Nucleic Acids Res., 29:207-209 (2001); MacCallum et al., J. Mol. Biol., 262:732-745 (1996); and Martin et al., Proc. Natl. Acad. Sci. USA, 86:9268-9272 (1989); Martin et al., Methods Enzymol., 203:121-153 (1991); and Rees et al., In Sternberg MJE (ed.), Protein Structure Prediction, Oxford University Press, Oxford, 141-172 (1996). For example, in Kabat, the CDR amino acid residues of the heavy chain variable domain (VH) are numbered 31-35 (HCDR1), 50-65 (HCDR2), and 95-102 (HCDR3), and the CDR amino acid residues of the light chain variable domain (VL) are numbered 24-34 (LCDR1), 50-56 (LCDR2), and 89-97 (LCDR3). In Chothia, the CDR amino acids of the VH are numbered 26-32 (HCDR1), 52-56 (HCDR2), and 95-102 (HCDR3), and the amino acid residues of the VL are numbered 26-32 (LCDR1), 50-52 (LCDR2), and 91-96 (LCDR3). A combination of the Kabat and Chothia CDR definitions number the CDRs as follows: 26-35 (HCDR1), 50-65 (HCDR2), and 95-102 (HCDR3) for human VH, and 24-34 (LCDR1), 50-56 (LCDR2), and 89-97 (LCDR3) for human VL. In IMGT, the VH CDR amino acid residues are numbered approximately 26-35 (HCDR1), 51-57 (HCDR2), and 93-102 (HCDR3), and the VL CDR amino acid residues are numbered approximately 27-32 (LCDR1), 50-52 (LCDR2), and 89-97 (LCDR3) (Kabat numbering).In IMGT, the CDR regions of an antibody can be determined using the program IMGT / DomainGap Align.

[0075] The term "hypervariable region" refers to the amino acid residues of an antibody which are involved in antigen binding. A hypervariable region comprises amino acid residues from the "CDRs" (e.g., LCDR1, LCDR2, and LCDR3 of the light chain variable domain and HCDR1, HCDR2, and HCDR3 of the heavy chain variable domain). See Kabat et al., (1991) Sequences of Proteins of Immunological Interest, 5th Ed. Public Health Service, National Institutes of Health, Bethesda, Md., which define antibody CDR regions by sequence. See also Chothia and Lesk (1987) J. Mol. Biol. 196:901-917, which define antibody CDR regions by structure. The term "framework" or "FR" residues refers to variable domain residues other than the hypervariable region residues defined herein as CDR residues.

[0076] Unless otherwise indicated, "antigen-binding fragment" refers to an antigen-binding fragment of an antibody, i.e., an antibody fragment that retains the ability to specifically bind to the antigen bound by the full-length antibody, e.g., a fragment that retains one or more CDR regions. Examples of antigen-binding fragments include, but are not limited to, Fab, Fab', F(ab')2, and Fv fragments; bispecific antibodies; linear antibodies; single-chain antibody molecules, e.g., single-chain Fv (ScFv); nanobodies and multispecific antibodies formed from antibody fragments.

[0077] As used herein, an antibody "specifically binds" to a target protein means that the antibody exhibits preferential binding to its target compared to other proteins, but this specificity does not require absolute binding specificity. An antibody "specifically binds" or "selectively binds" is used in the context of describing the interaction between an antigen (e.g., a protein) and an antibody or antigen-binding antibody fragment, and refers to a binding reaction that determines the presence of the antigen in a heterogeneous population of proteins and other biologics, such as a biological sample, blood, serum, plasma, or tissue sample. Thus, under certain specified immunoassay conditions, an antibody or antigen-binding fragment thereof specifically binds to a particular antigen at least 2-fold stronger compared to background levels and does not specifically bind in significant amounts to other antigens present in the sample. In one aspect, under specified immunoassay conditions, an antibody or antigen-binding fragment thereof specifically binds to a particular antigen at least 10-fold stronger compared to background levels of binding and does not specifically bind in significant amounts to other antigens present in the sample.

[0078] As used herein, the term "human antibody" refers to an antibody that contains only human immunoglobulin protein sequences. A human antibody may contain mouse glycosylation if produced in a mouse, a mouse cell, or a hybridoma derived from a mouse cell. Similarly, a "mouse antibody" or a "rat antibody" refers to an antibody that contains only mouse or rat immunoglobulin protein sequences, respectively.

[0079] The term "humanized" or "humanized antibody" refers to forms of antibodies that contain sequences derived from non-human (e.g., murine) and human antibodies. Such antibodies contain minimal sequences derived from non-human immunoglobulins. Generally, a humanized antibody contains substantially all of at least one, and typically two, variable domains, with all or substantially all of the hypervariable loops corresponding to those of a non-human immunoglobulin, and all or substantially all of the FR regions being those of a human immunoglobulin sequence. A humanized antibody optionally also contains at least a portion of an immunoglobulin constant region (Fc), typically at least a portion of a human immunoglobulin constant region (Fc). When necessary to distinguish the humanized antibody from a rodent parent antibody, the prefixes "hum," "hu," "Hu," or "h" are added to the name of the antibody clone. Humanized forms of rodent antibodies generally contain the same CDR sequences as the rodent parent antibody, but can contain certain amino acid substitutions to increase affinity, increase the stability of the humanized antibody, remove post-translational modifications, or for other reasons.

[0080] The term "corresponding human germline sequence" refers to a nucleic acid sequence encoding a human variable region amino acid sequence or subsequence that shares the highest amino acid sequence identity determined with a reference variable region amino acid sequence or subsequence compared to all other known variable region amino acid sequences encoded by human germline immunoglobulin variable region sequences. Corresponding human germline sequence can also refer to a human variable region amino acid sequence or subsequence that has the highest amino acid sequence identity with a reference variable region amino acid sequence or subsequence compared to all other evaluated variable region amino acid sequences. Corresponding human germline sequence can be framework regions only, complementarity determining regions only, framework and complementarity determining regions, variable segments (as defined above), or other combinations of sequences or subsequences that include variable regions. Sequence identity can be determined using methods described herein, such as aligning the two sequences using BLAST, ALIGN, or another alignment algorithm known in the art. The corresponding human germline nucleic acid or amino acid sequence can have at least about 90%, 91, 92%, 93%, 94%, 95%, 96%, 97%, 98%, 99%, or 100% sequence identity with the nucleic acid or amino acid sequence of the reference variable region. Furthermore, if the antibody contains a constant region, the constant region is also derived from such a human sequence, e.g., a human germline sequence, or a mutated version of a human germline sequence, or an antibody containing a consensus framework sequence derived from human framework sequence analysis, e.g., as described in Knappik et al., J. Mol. Biol. 296:57-86, 2000.

[0081] “Equilibrium dissociation constant (K D The term "dissociation rate constant (kd, time -1 ) as the binding rate constant (ka, time -1 , M -1 The equilibrium dissociation constant can be measured using any method known in the art. Antibodies of the present disclosure generally have an equilibrium dissociation constant of about 10 -7 Less than M or 10 -8 Less than m, e.g., about 10-9 Less than M or 10 -10 M or less, and in some embodiments, about 10 -11 Less than M, 10 -12 Less than M or 10 -13 M

[0082] The term "cancer" or "tumor" as used herein has the broadest meaning as understood in the art and refers to a physiological condition in a mammal that is typically characterized by unregulated cell proliferation. In the context of this disclosure, cancer is not limited to a particular type or location.

[0083] In the context of the present disclosure, when referring to an amino acid sequence, the term "conservative substitution" refers to the replacement of an original amino acid with a new amino acid that does not substantially alter the chemical, physical and / or functional properties of the antibody or fragment, such as its binding affinity to CD137. Common conservative changes of amino acids are well known in the art.

[0084] An example of an algorithm suitable for determining percent sequence identity and sequence similarity is the BLAST algorithm described in Altschul et al, Nuc. Acids Res. 25:3389-3402, 1977; and Altschul et al., J. Mol. Biol. 215:403-410, 1990, respectively. Software for performing BLAST analysis is publicly available through the National Center for Biotechnology Information. This algorithm involves first identifying high-scoring sequence pairs (HSPs) by identifying short words of length W in the query sequence that match or meet some positive threshold score T when aligned with words of the same length in database sequences. T is referred to as the neighborhood word score threshold. These initial neighborhood word hits serve as a value for initiating searches to find longer HSPs that contain them. The word hits are extended in both directions along each sequence for as far as the cumulative alignment score can be increased. Cumulative scores are calculated using, for nucleotide sequences, the parameters M (reward score for a pair of matching residues, always >0) and N (penalty score for mismatching residues, always <0). For amino acid sequences, a scoring matrix is ​​used to calculate the cumulative score. Extension of the word hits in each direction is halted when the cumulative alignment score falls by an amount X from its maximum achieved value, or when the cumulative score becomes zero or below zero due to the accumulation of one or more negative scoring residue alignments, or when either end of the sequence is reached. The BLAST algorithm parameters W, T, and X determine the sensitivity and speed of the alignment. The BLASTN program (for nucleotide sequences) uses as defaults a wordlength (W) of 11, an expectation (E) of 10, M=5, N=4, and a comparison of both strands.For amino acid sequences, the BLAST program uses as defaults a wordlength of 3, and an expectation (E) of 10, and the BLOSUM62 scoring matrix (see Henikoff and Henikoff, (1989) Proc. Natl. Acad. Sci. USA 89:10915) of 50. Alignment (B), expectation (E) of 10, M=5, N=-4, and a comparison of both strands.

[0085] BLAST algorithm also performs statistical analysis of the similarity between two sequences (see, for example, Karlin and Altschul, Proc. Natl. Acad. Sci. USA 90:5873-5787, 1993). One measure of similarity provided by BLAST algorithm is the minimum total probability (P(N)), which provides an indication of the probability that a match between two nucleotide or amino acid sequences will occur by chance. For example, a nucleic acid is considered to be similar to a reference sequence when the minimum total probability when comparing test nucleic acid with reference nucleic acid is less than about 0.2, more preferably less than about 0.01, and most preferably less than about 0.001.

[0086] The percent identity between two amino acid sequences can also be determined using the algorithm built into the ALIGN program (version 2.0) using a PAM120 weight residue table, a gap length penalty of 12, and a gap penalty of 4 (E. Meyers and W. Miller, Comput. Appl. Biosci. 4:11-17, (1988)). Additionally, the percent identity between two amino acid sequences can be determined using the algorithm built into the GAP program of the GCG software package using either a BLOSUM62 matrix or a PAM250 matrix, and gap weights of 16, 14, 12, 10, 8, 6, or 4 and length weights of 1, 2, 3, 4, 5, or 6 (Needleman and Wunsch, J. Mol. Biol. 48:444-453, (1970)).

[0087] The term "nucleic acid" is used interchangeably herein with the term "polynucleotide" and refers to deoxyribonucleotides or ribonucleotides and polymers thereof in either single-stranded or double-stranded form. The term encompasses nucleic acids containing known nucleotide analogs or modified backbone residues or linkages, which are synthetic, naturally occurring, and non-naturally occurring, have similar binding properties as the reference nucleic acid, and are metabolized in a manner similar to the reference nucleotide. Examples of such analogs include, but are not limited to, phosphorothioates, phosphoramidates, methyl phosphonates, chiral methyl phosphonates, 2-O-methyl ribonucleotides, and peptide nucleic acids (PNAs).

[0088] The term "operably linked" in the context of nucleic acids refers to a functional relationship between two or more polynucleotide (e.g., DNA) segments. Typically, it refers to the functional relationship of a transcriptional regulatory sequence to a transcriptional sequence. For example, a promoter or enhancer sequence is operably linked to a coding sequence if it stimulates or regulates the transcription of the coding sequence in an appropriate host cell or other expression system. Generally, promoter transcriptional regulatory sequences that are operably linked to a transcriptional sequence are physically contiguous to the transcriptional sequence, i.e., they are cis-acting. However, some transcriptional regulatory sequences, such as enhancers, do not need to be physically contiguous or located in close proximity to the coding sequences whose transcription they enhance.

[0089] In some embodiments, the present disclosure provides compositions, e.g., pharma- ceutically acceptable compositions, comprising at least an anti-CD137 binding antibody described herein, formulated with at least one pharma- ceutically acceptable excipient. As used herein, the term "pharma- ceutically acceptable excipient" includes any and all solvents, dispersion media, isotonicity agents, and absorption delaying agents, etc., that are physiologically compatible. The excipient may be suitable for intravenous, intramuscular, subcutaneous, parenteral, rectal, spinal, or epidermal administration (e.g., by injection or infusion).

[0090] The compositions disclosed herein can be in various forms. These include, for example, liquid, semi-solid and solid dosage forms, such as liquid solutions (e.g., injection and infusion solutions), dispersions or suspensions, liposomes, and suppositories. The suitable form depends on the intended mode of administration and therapeutic application. Exemplary suitable compositions are in the form of injection or infusion solutions. One suitable mode of administration is parenteral (e.g., intravenous, subcutaneous, intraperitoneal, intramuscular). In some embodiments, the antibody is administered by intravenous infusion or injection. In certain embodiments, the antibody is administered by intramuscular or subcutaneous injection.

[0091] The term "therapeutically effective amount" as used herein refers to an amount of an antibody that, when administered to a subject for treating a disease or at least one of the clinical symptoms of a disease or disorder, is sufficient to achieve such treatment for the disease, disorder, or condition. A "therapeutically effective amount" may vary with the antibody, the disease, disorder, and / or symptoms of the disease or disorder, the severity of the disease, disorder, and / or symptoms of the disease or disorder, the age of the subject being treated, and / or the weight of the subject being treated. The appropriate amount in any given case may be apparent to one skilled in the art or can be determined by routine experimentation. In the case of combination therapy, a "therapeutically effective amount" refers to the total amount of the combination for effective treatment of a disease, disorder, or condition.

[0092] The term "combination therapy" refers to the administration of two or more therapeutic agents to treat a therapeutic condition or disorder described in this disclosure. Such administration includes the simultaneous administration of these therapeutic agents in a substantially simultaneous manner. Such administration also includes the simultaneous administration in multiple containers or in separate containers (e.g., capsules, powders, and liquids) for each active ingredient. The powders and / or liquids can be reconstituted or diluted to the desired dose before administration. Furthermore, such administration also includes the sequential use of various therapeutic agents, either at about the same time or at different times. In either case, the treatment regimen provides the beneficial effects of the drug combination in treating the condition or disorder described herein.

[0093] As used herein, the phrase "in combination with" means that the anti-CD137 antibody, antigen-binding fragment, or anti-CD137-containing multispecific antibody is administered to the subject simultaneously with, immediately before, or immediately after administration of an additional therapeutic agent. In certain embodiments, the anti-CD137 antibody, antigen-binding fragment, or anti-CD137-containing multispecific antibody is administered as a co-formulation with the additional therapeutic agent.

[0094] DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS The present disclosure provides antibodies, antigen-binding fragments, or anti-CD137-containing multivalent antibodies that specifically bind to human CD137. Additionally, the present disclosure provides antibodies that have desirable pharmacokinetic properties and other desirable attributes and thus can be used to reduce the likelihood of cancer or to treat cancer. The present disclosure further provides pharmaceutical compositions comprising the antibodies or antigen-binding fragments, as well as methods of making and using such pharmaceutical compositions, for the prevention and treatment of cancer and related disorders.

[0095] Anti-CD137 antibody The present disclosure provides an antibody or antigen-binding fragment thereof that specifically binds to CD137. The antibody or antigen-binding fragment of the present disclosure includes, but is not limited to, an antibody or antigen-binding fragment thereof generated as described below.

[0096] The present disclosure provides an antibody or antigen-binding fragment that specifically binds to CD137, wherein the antibody or antibody fragment (e.g., antigen-binding fragment) comprises a VH domain comprising the amino acid sequence of SEQ ID NO:9, SEQ ID NO:19, SEQ ID NO:24, SEQ ID NO:33, or SEQ ID NO:103 (Table 1). The present disclosure also provides an antibody or antigen-binding fragment that specifically binds to CD137, wherein the antibody or antigen-binding fragment comprises an HCDR (heavy chain complementarity determining region) comprising the amino acid sequence of any one of the HCDRs listed in Table 1. In one aspect, the present disclosure provides an antibody or antigen-binding fragment that specifically binds to CD137, wherein the antibody comprises one, two, three, or more HCDRs (or alternatively consists of one, two, three, or more HCDRs) comprising the amino acid sequence of any of the HCDRs listed in Table 1.

[0097] Other antibodies or antigen-binding fragments of the present disclosure contain amino acids in the CDR regions that are altered but have at least 60%, 70%, 80%, 90%, 95% or 99% identity with the CDR regions disclosed in Table 1. In some embodiments, this includes amino acid changes in which no more than 1, 2, 3, 4 or 5 amino acids are altered in the CDR regions when compared to the CDR regions shown in the sequences set forth in Table 1.

[0098] Other antibodies of the disclosure include those in which the amino acids, or the nucleic acids encoding the amino acids, have been altered but have at least 60%, 70%, 80%, 90%, 95% or 99% identity to the sequences set forth in Table 1. In some embodiments, this includes changes in the amino acid sequence in which no more than 1, 2, 3, 4 or 5 amino acids are altered in the variable regions while retaining substantially the same therapeutic activity when compared to the variable regions set forth in the sequences set forth in Table 1.

[0099] The present disclosure also provides nucleic acid sequences encoding the VH domain and full-length heavy chains of antibodies that specifically bind to CD137. Such nucleic acid sequences can be optimized for expression in mammalian cells.

[0100] [Table 1-1]

[0101] [Table 1-2]

[0102] [Table 1-3]

[0103] [Table 1-4]

[0104] Identification of epitopes and antibodies that bind to the same epitopes The present disclosure provides antibodies and antigen-binding fragments thereof that bind to an epitope of human CD137. In certain embodiments, the antibodies and antigen-binding fragments can bind to the same epitope of CD137.

[0105] The present disclosure also provides antibodies and antigen-binding fragments thereof that bind to the same epitope as the anti-CD137 antibodies described in Table 1. Thus, additional antibodies and antigen-binding fragments thereof can be identified based on their ability to cross-compete (e.g., competitively inhibit binding in a statistically significant manner) with other antibodies in a binding assay. The ability of a test antibody to inhibit the binding of an antibody and antigen-binding fragment thereof of the present disclosure to CD137 demonstrates that the test antibody can compete with that antibody or antigen-binding fragment thereof for binding to CD137. Without being bound to any one theory, such antibodies can bind to the same or related (e.g., structurally similar or spatially proximal) epitope on CD137 as the competing antibody or antigen-binding fragment thereof. In certain embodiments, an antibody that binds to the same epitope on CD137 as an antibody or antigen-binding fragment thereof of the present disclosure is a human or humanized monoclonal antibody. Such human or humanized monoclonal antibodies can be prepared and isolated as described herein.

[0106] Change of FC area In yet another embodiment, the Fc region is modified by replacing at least one amino acid residue with a different amino acid residue to modify the effector function of the antibody. For example, one or more amino acids can be replaced with a different amino acid residue such that the antibody has a modified affinity for an effector ligand but retains the antigen-binding ability of the parent antibody. The effector ligand to which affinity is modified can be, for example, an Fc receptor or the C1 component of complement. This approach is described, for example, in U.S. Patent Nos. 5,624,821 and 5,648,260 by Winter et al.

[0107] In another embodiment, one or more amino acid residues can be replaced with one or more different amino acid residues such that the antibody has altered C1q binding and / or reduced or eliminated complement dependent cytotoxicity (CDC). This approach is described, for example, in U.S. Patent No. 6,194,551 by Idusogie et al.

[0108] In yet another embodiment, one or more amino acid residues are altered to modify the antibody's ability to fix complement. This approach is described, for example, in Publication WO94 / 29351 by Bodmer et al. In certain embodiments, one or more amino acids of the antibody or antigen-binding fragment thereof of the present disclosure are replaced by one or more allotypic amino acid residues for the IgG1 subclass and the kappa isotype. The allotypic amino acid residues include, but are not limited to, the constant regions of the heavy chains of the IgG1, IgG2, and IgG3 subclasses, and the constant regions of the light chains of the kappa isotype described by Jefferis et al., MAbs1:332-338 (2009).

[0109] In another embodiment, the Fc region is modified to increase the ability of the antibody to mediate antibody-dependent cellular cytotoxicity (ADCC) and / or to increase the affinity of the antibody for the Fcγ receptor by modifying one or more amino acids. This approach is described, for example, in publication WO00 / 42072 by Presta. Furthermore, the binding sites on human IgG1 for FcγRI, FcγRII, FcγRIII and FcRn have been mapped and mutants with improved binding have been described (Shields et al., J. Biol. Chem. 276:6591-6604, 2001).

[0110] In yet another embodiment, the glycosylation of the CD137 antibody or antigen-binding fragment is modified. For example, an aglycosylated antibody can be generated (i.e., the antibody lacks or has reduced glycosylation). The glycosylation can be altered to, for example, increase the affinity of the antibody for an "antigen". Such carbohydrate modification can be achieved, for example, by altering one or more sites of glycosylation within the antibody sequence. For example, one or more amino acid substitutions can be made that result in the elimination of one or more variable region framework glycosylation sites, thereby eliminating glycosylation at that site. Such aglycosylation can increase the affinity of the antibody for the antigen. Such an approach is described, for example, in U.S. Patent Nos. 5,714,350 and 6,350,861 by Co et al.

[0111] Additionally or alternatively, antibodies can be made with modified types of glycosylation, such as hypofucosylated antibodies with reduced amounts of fucosyl residues or antibodies with increased bisecting GlcNac structures. Such modified glycosylation patterns have been demonstrated to increase the ADCC ability of antibodies. Such carbohydrate modifications can be achieved, for example, by expressing the antibody in a host cell with a modified glycosylation pathway. Cells with modified glycosylation pathways have been described in the art and can be used as host cells to express recombinant antibodies to produce antibodies with modified glycosylation. For example, EP 1,176,195 by Hang et al. describes cell lines with a functionally disrupted FUT8 gene encoding fucosyltransferase, such that antibodies expressed in such cell lines exhibit hypofucosylation. Publication WO 03 / 035835 by Presta describes a mutant CHO cell line, Lecl3 cells, that has a reduced ability to attach fucose to Asn(297)-linked carbohydrates, which also results in hypofucosylation of antibodies expressed in the host cells (see also Shields et al., (2002) J. Biol. Chem. 277:26733-26740). WO 99 / 54342 by Umana et al. describes cell lines engineered to express glycoprotein-modifying glycosyltransferases (e.g., beta(1,4)-N-acetylglucosaminyltransferase III (GnTIII)) such that antibodies expressed in the engineered cell lines exhibit an increase in bisecting GlcNac structures, resulting in increased ADCC activity of the antibodies (see also Umana et al., Nat. Biotech. 17:176-180, 1999).

[0112] In another embodiment, when reduced ADCC is desired, human antibody subclass IgG4 has been shown in many previous reports to have only moderate ADCC and little CDC effector function (Moore et al., 2010 MAbs, 2:181-189). However, native IgG4 has been found to be less stable under stress conditions such as in acidic buffer or at elevated temperature (Angal, 1993 Mol Immunol, 30:105-108; Dall'Acqua et al, 1998 Biochemistry, 37:9266-9273; Aalberse et al., 2002 Immunol, 105:9-19). Reduced ADCC can be achieved by operably binding the antibody to an IgG4 Fc engineered with a combination of modifications that reduce FcγR binding or C1q binding activity, thereby reducing or eliminating ADCC and CDC effector function. Considering the physicochemical properties of antibodies as biological drugs, one of the more undesirable intrinsic properties of IgG4 is the dynamic separation of its two heavy chains in solution to form half antibodies, which generates bispecific antibodies in vivo through a process called "Fab arm exchange" (Van der Neut Kolfschoten M, et al., 2007 Science, 317: 1554-157). Mutation of serine to proline at position 228 (EU numbering system) appeared to be inhibitory to IgG4 heavy chain separation (Angal, 1993 Mol Immunol, 30: 105-108; Aalberse et al., 2002 Immunol, 105: 9-19).Some of the amino acid residues in the hinge and gamma Fc regions have been reported to affect antibody interaction with Fcγ receptors (Chappel et al., 1991 Proc. Natl. Acad. Sci. USA, 88:9036-9040; Mukherjee et al., 1995 FASEB J, 9:115-119; Armour et al., 1999 Eur J Immunol, 29:2613-2624; Clynes et al, 2000 Nature Medicine, 6:443-446; Arnold, 2007 Annu Rev Immunol, 25:21-50). Furthermore, some IgG4 isoforms occurring rarely in the human population can also induce different physicochemical properties (Brusco et al., 1998 Eur J Immunogenet, 25:349-55; Aalberse et al., 2002 Immunol, 105:9-19). To generate CD137 antibodies with low ADCC and CDC but good stability, it is possible to modify the hinge and Fc regions of human IgG4 and introduce several modifications. These modified IgG4 Fc molecules can be found in SEQ ID NOs: 83-88 of US Patent No. 8,735,553 by Li et al.

[0113] CD137 antibody production Anti-CD137 antibodies, antigen-binding fragments, and multispecific antibodies can be produced by any means known in the art, including, but not limited to, recombinant expression of antibody tetramers, chemical synthesis, and enzymatic digestion, while full-length monoclonal antibodies can be obtained, for example, by hybridoma or recombinant production. Recombinant expression can be derived from any suitable host cell known in the art, for example, mammalian host cells, bacterial host cells, yeast host cells, insect host cells, etc.

[0114] The disclosure further provides polynucleotides encoding the antibodies described herein, e.g., polynucleotides encoding a heavy or light chain variable region or segment comprising a complementarity determining region as described herein. In some embodiments, the polynucleotide encoding the heavy chain variable region has at least 85%, 89%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, 99%, or 100% nucleic acid sequence identity to a polynucleotide selected from the group consisting of SEQ ID NO:10, SEQ ID NO:20, SEQ ID NO:25, SEQ ID NO:33, or SEQ ID NO:104.

[0115] The polynucleotides of the disclosure can encode the variable region sequences of an anti-CD137 antibody. The polynucleotides of the disclosure can also encode both the variable and constant regions of the antibody. Some of the polynucleotide sequences encode a polypeptide that includes both the heavy and light chain variable regions of one of the exemplified anti-CD137 antibodies.

[0116] Also provided in the present disclosure are expression vectors and host cells for producing anti-CD137 antibodies. The choice of expression vector depends on the intended host cell in which the vector is expressed. Typically, the expression vector contains a promoter and other regulatory sequences (e.g., enhancers) operably linked to the polynucleotide encoding the anti-CD137 antibody chain or antigen-binding fragment. In some embodiments, an inducible promoter is used to prevent expression of the inserted sequence except when under the control of an inducing condition. Inducible promoters include, for example, arabinose, lacZ, metallothionein promoter, or heat shock promoter. Cultures of transformed organisms can be grown under non-inducing conditions without biasing the population in favor of coding sequences whose expression products are better tolerated by the host cell. In addition to promoters, other regulatory elements may be required or desired for efficient expression of anti-CD137 antibodies or antigen-binding fragments. These elements typically include an ATG initiation codon and adjacent ribosome binding sites or other sequences. In addition, the efficiency of expression can be increased by including enhancers appropriate for the cell system being used (see, e.g., Scharf et al., Results Probl. Cell Differ. 20:125, 1994; and Bittner et al., Meth. Enzymol., 153:516, 1987). For example, the SV40 enhancer or CMV enhancer can be used to increase expression in mammalian host cells.

[0117] Host cells for carrying and expressing anti-CD137 antibody chains can be prokaryotic or eukaryotic. E. coli is one prokaryotic host useful for cloning and expressing the polynucleotides of the present disclosure. Other microbial hosts suitable for use include bacilli, such as Bacillus subtilis, and other enterobacteria, such as Salmonella, Serratia, and various Pseudomonas species. In these prokaryotic hosts, expression vectors can also be made, which typically contain expression control sequences compatible with the host cell (e.g., origin of replication). In addition, there will be any number of different known promoters, such as the lactose promoter system, the tryptophan (trp) promoter system, the beta-lactamase promoter system, or a promoter system from phage lambda. The promoters typically control expression, optionally with an operator sequence, and have ribosome binding site sequences, etc., for initiating and completing transcription and translation. Other microbes, such as yeast, can also be used to express anti-CD137 polypeptides. Insect cells in combination with baculovirus vectors can also be used.

[0118] In other embodiments, mammalian host cells are used to express and produce the anti-CD137 polypeptides of the present disclosure. For example, they can be hybridoma cell lines expressing endogenous immunoglobulin genes, or mammalian cell lines with exogenous expression vectors. These include any normal mortal or normal or abnormal immortal animal or human cells. For example, several suitable host cell lines capable of secreting intact immunoglobulins have been developed, including CHO cell lines, various COS cell lines, HEK293 cells, myeloma cell lines, transformed B cells, and hybridomas. The use of mammalian tissue cell cultures to express polypeptides is generally discussed, for example, in Winnacker, From Genes to Clones, VCH Publishers, NY, NY, 1987. Expression vectors for mammalian host cells can include expression control sequences such as origins of replication, promoters, and enhancers (see, for example, Queen et al., Immunol. Rev. 89:49-68, 1986), as well as necessary processing information sites such as ribosome binding sites, RNA splice sites, polyadenylation sites, and transcription terminator sequences. These expression vectors usually contain promoters derived from mammalian genes or mammalian viruses. Suitable promoters can be constitutive, cell type-specific, stage-specific, and / or modulatable or regulatable. Useful promoters include, but are not limited to, metallothionein promoter, constitutive adenovirus major late promoter, dexamethasone-inducible MMTV promoter, SV40 promoter, MRP polIII promoter, constitutive MPSV promoter, tetracycline-inducible CMV promoter (such as human immediate early CMV promoter), constitutive CMV promoter, and promoter-enhancer combinations known in the art.

[0119] CD137 multispecific antibody In one embodiment, the anti-CD137 antibodies disclosed herein can be incorporated into an anti-CD137xTAA multispecific antibody, where the TAA is an antibody or fragment thereof against any human tumor-associated antigen (TAA). The antibody molecule is a multispecific antibody molecule, e.g., comprising multiple antigen-binding domains, where at least one antigen-binding domain sequence specifically binds to CD137 and a second antigen-binding domain sequence specifically binds to a TAA. In one embodiment, the multispecific antibody comprises a third, fourth, or fifth antigen-binding domain. In one embodiment, the multispecific antibody is a bispecific antibody, trispecific antibody, or tetraspecific antibody. In each instance, the multispecific antibody comprises at least one anti-CD137 antigen-binding domain and at least one anti-TAA antigen-binding domain.

[0120] In one embodiment, a multispecific antibody is a bispecific antibody. As used herein, a bispecific antibody specifically binds to only two antigens. A bispecific antibody comprises a first antigen-binding domain that specifically binds to CD137 and a second antigen-binding domain that specifically binds to a TAA. This includes bispecific antibodies that comprise a heavy chain variable domain that specifically binds to CD137 and a heavy chain variable domain and a light chain variable domain that specifically bind to a TAA. In another embodiment, a bispecific antibody comprises an antigen-binding fragment of an antibody that specifically binds to CD137 and an antigen-binding fragment that specifically binds to a TAA. In bispecific antibodies that comprise an antigen-binding fragment, the antigen-binding fragment can be a Fab, F(ab')2, Fv, or a single chain Fv (ScFv) or scFv.

[0121] Previous work (Coloma and Morrison Nature Biotech. 15:159-163 (1997)) described tetravalent bispecific antibodies engineered by fusing DNA encoding a single-chain anti-dansyl antibody Fv (scFv) after the C-terminus (CH3-scFv) or after the hinge (hinge-scFv) of an IgG3 anti-dansyl antibody. The present disclosure provides multivalent antibodies (e.g., tetravalent antibodies) having at least two antigen-binding domains, which may be readily produced by recombinant expression of nucleic acids encoding antibody polypeptide chains. Multivalent antibodies herein contain 3 to 8, preferably 4, antigen-binding domains that specifically bind to at least two antigens.

[0122] The present disclosure provides a bispecific tetravalent antibody comprising VD1-CL-(X1)n-VD2-CH1-Fc or VD1-CH-(X1)n-VD2-CL-Fc, where VD1 is a first variable domain, VD2 is a second variable domain, Fc is one polypeptide chain of the Fc region, CH or CL is a constant heavy chain domain or constant light chain domain, and (X1)n is a linker of at least two amino acids.

[0123] In one embodiment, the bispecific tetravalent antibody may be a multimer of four polypeptide chains, where each of the two heavy chains comprises a first VH domain (VH1), a first CH1 domain, a second VH domain (VH2), and an Fc region comprising a second CH1, a hinge, a CH2, and a CH3, and each of the two light chains comprises a first VL domain (VL1), a first CL region, a second VL domain (VL2), and a second CL region. In another embodiment, the bispecific tetravalent antibody may comprise multiple antibody Fab fragments linked together to a single Fc domain. For example, Fab1 may be linked to Fab2 via a polypeptide linker, where Fab2 comprises a Fab, one CH1 domain of the hinge region, then the CH2 and CH3 of the Fc domain. For example, an anti-TAA Fab can be linked via a linker from the CL domain of the anti-TAA Fab to the VH domain of the anti-CD137 Fab, and from the CH1 domain, hinge region, CH2 domain, and CH3 domain of the anti-CD137 Fab. In another example, an anti-CD137 Fab can be linked via a linker from the CL domain of the anti-CD137 Fab to the VH domain of the anti-TAA Fab, and from the CH1 domain, hinge region, CH2 domain, and CH3 domain of the anti-TAA Fab.

[0124] Linker It is also understood that the domains and / or regions of the polypeptide chains of a multispecific antibody can be separated by linker regions of various lengths. In some embodiments, the epitope binding domains are separated from each other, from the CL, CH1, hinge, CH2, CH3, or the entire Fc region by a linker region. For example, a linker region such as VL1-CL-(linker)VH2-CH1 can contain a random assortment of amino acids or a limited set of amino acids. Such linker regions can be flexible or rigid (see US2009 / 0155275).

[0125] Multispecific antibodies have been developed through dimerization devices such as leucine zippers (Kostelny et al., J. Immunol. 1992 148:1547-53; de Kruifetal J. Biol. Chem. 1996 271:7630-4) and Ig C / CH1 domains (Muller et al., FEBS Lett. 422:259-64), with or without the use of flexible linkers; bispecific antibodies (Holliger et al., (1993) Proc. Nat. Acad. Sci. USA. 1998 90:6444-8; Zhu et al., Bio / Technology (NY) 1996 14:192-6); Fab-scFv fusions (Schoonjans et al., J. Immunol. 2000 165:7050-7); and miniantibody formats (Packet al., Biochemistry 1992.31:1579-84; Packet et al., Bio / Technology 1993 11:1271-7) by genetically fusing two single-chain Fv (scFv) or Fab fragments (Mallender et al., J. Biol. Chem. 1994 269:199-206; Macket et al., Proc. Natl. Acad. Sci. USA 1995 92:7021-5; Zapata Protein Eng. 1995 8.1057-62).

[0126] A multispecific antibody disclosed herein comprises a linker region of at least 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 25, 30, 35, 40, 45, 50, 55, 60, 65, 70, 75, or more amino acid residues between one or more of its epitope binding domains, CL domain, CH1 domain, hinge region, CH2 domain, CH3 domain, or Fc region. In some embodiments, the amino acids glycine and serine comprise the amino acids within the linker region. In another embodiment, the linker is selected from the group consisting of GS (SEQ ID NO:239), GGS (SEQ ID NO:240), GSG (SEQ ID NO:241), SGG (SEQ ID NO:242), GGG (SEQ ID NO:243), GGGS (SEQ ID NO:244), SGGG (SEQ ID NO:245), GGGGS (SEQ ID NO:246), GGGGSGS (SEQ ID NO:247), GGGGSGS (SEQ ID NO:248), GGGGSGGS (SEQ ID NO:249), GGGGSGGGGS (SEQ ID NO:250), GGGSGGGGSGGGGS (SEQ ID NO:251), A KTTPKLEEGEFSEAR (SEQ ID NO: 252), AKTTPKLEEGEFSEARV (SEQ ID NO: 253), AKTTPKLGG (SEQ ID NO: 254), SAKTTPKLGG (SEQ ID NO: 255), AKTTPKLEEGEFSEARV (SEQ ID NO: 256), SAKTTP (SEQ ID NO: 257), SAKTTPKLGG (SEQ ID NO: 258), RADAAP (SEQ ID NO: 259), RADAAPTVS (SEQ ID NO: 260), RADAAAAAGGPGS (SEQ ID NO: 261), RADAAAA(G 4 S) 4(SEQ ID NO: 262), SAKTTP (SEQ ID NO: 263), SAKTTPKLGG (SEQ ID NO: 264), SAKTTPKLEEGEFSEARV (SEQ ID NO: 265), ADAAP (SEQ ID NO: 266), ADAAPTVSIFPP (SEQ ID NO: 267), TVAAP (SEQ ID NO: 268), TVAAPSVFIFPP (SEQ ID NO: 269), QPKAAP (SEQ ID NO: 270), QPKAAPSVTLFPP (SEQ ID NO: 271), AKTTPP (SEQ ID NO: 272), AKTTPPSV TPLAP (SEQ ID NO: 273), AKTTAP (SEQ ID NO: 274), AKTTAPSVYPLAP (SEQ ID NO: 275), ASTKGP (SEQ ID NO: 276), ASTKGPSVFPLAP (SEQ ID NO: 277), GENKVEYAPALMALS (SEQ ID NO: 278), GPAKELTPLKEAKVS (SEQ ID NO: 279), and GHEAAAVMQVQYPAS (SEQ ID NO: 280), or any combination thereof (see WO 2007 / 024715).

[0127] Dimerization specific amino acids In one embodiment, the multispecific antibody comprises at least one dimerization-specific amino acid change. The dimerization-specific amino acid change results in "knobs into holes" interactions and increases the assembly of correct multispecific antibodies. The dimerization-specific amino acid can be in the CH1 domain or the CL domain or a combination thereof. The dimerization-specific amino acid is used to pair the CH1 domain with another CH1 domain (CH1-CH1) and the CL domain with another CL domain (CL-CL) and can be found at least in the disclosures of WO2014082179, WO2015181805 and WO2017059551. The dimerization-specific amino acid can be in the Fc domain or combined with the dimerization-specific amino acid in the CH1 or CL domain.

[0128] Detection and Diagnostic Methods The antibody or antigen-binding fragment of the present disclosure is useful in various applications, including, but not limited to, a method for detecting CD137. In one embodiment, the antibody or antigen-binding fragment is useful for detecting the presence of CD137 in a biological sample. As used herein, the term "detecting" includes quantitative or qualitative detection. In certain embodiments, the biological sample includes cells or tissues. In other embodiments, such tissues include normal tissues and / or cancerous tissues that express CD137 at a higher level than other tissues.

[0129] In one embodiment, the present disclosure provides a method for detecting the presence of CD137 in biological sample.In a particular embodiment, the method comprises contacting biological sample with anti-CD137 antibody or its antigen-binding fragment under conditions that allow antibody to bind to antigen, and detecting whether complex is formed between antibody and antigen.Biological sample can include, but is not limited to, urine, tissue, sputum, or blood sample.

[0130] Also included are methods of diagnosing disorders associated with expression of CD137. In certain embodiments, the methods include contacting a test cell with an anti-CD137 antibody or antigen-binding fragment thereof, determining the expression level (either quantitatively or qualitatively) of CD137 expressed by the test cell by detecting binding of the anti-CD137 antibody or antigen-binding fragment thereof to a CD137 polypeptide, and comparing the expression level by the test cell to the expression level of CD137 in a control cell (e.g., a normal cell or a non-CD137-expressing cell of the same tissue origin as the test cell), where a higher level of CD137 expression in the test cell compared to the control cell indicates the presence of a disorder associated with expression of CD137.

[0131] Treatment method The antibody or antigen-binding fragment of the present disclosure is useful in various applications, including but not limited to, methods for treating CD137-related disorders or diseases.In one embodiment, the CD137-related disorders or diseases is cancer.In the case of CD137xTAA multispecific antibody, the cancer is specific to TAA, and CD137 acts to recruit immune cells to TAA-expressing tumors.

[0132] In one embodiment, the present disclosure provides a method for treating cancer.In a particular embodiment, the method comprises administering an effective amount of an anti-CD137 antibody, its antigen-binding fragment, or a CD137-containing multispecific antibody to a patient in need of the treatment.Cancer includes, but is not limited to, gastric cancer, colon cancer, pancreatic cancer, breast cancer, head and neck cancer, kidney cancer, liver cancer, lung cancer, small cell lung cancer, non-small cell lung cancer, ovarian cancer, skin cancer, mesothelioma, lymphoma, leukemia, myeloma, and sarcoma.

[0133] The anti-CD137 antibodies or antigen-binding fragments disclosed herein can be administered by any suitable means, including parenteral, intrapulmonary, and intranasal, including localized therapeutic, intralesional, or intratumoral administration, if desired. Parenteral infusions include intramuscular, intravenous, intraarterial, intraperitoneal, or subcutaneous administration. Dosing can be by any suitable route, for example, injection, such as intravenous or subcutaneous infusion, depending in part on whether administration is brief or chronic. Various dosing schedules are contemplated herein, including, but not limited to, single or multiple administrations at various times, bolus administration, and pulse infusion.

[0134] The antibodies or antigen-binding fragments of the present disclosure can be formulated, dosed, and administered in a manner consistent with good medical practice. Factors to consider in this context include the particular disorder being treated, the particular mammal being treated, the clinical condition of the individual patient, the cause of the disorder, the site of delivery of the agent, the method of administration, the scheduling of administration, and other factors known to medical practitioners. The antibodies are optionally, but not necessarily, formulated with one or more agents currently used to prevent or treat the disorder in question. The effective amount of such other agents will depend on the amount of antibody present in the formulation, the type of disorder or treatment, and other factors discussed above. These are generally used in the same dosages and by any route of administration as described herein, or about 1-99% of the dosages described herein, or any dosages empirically / clinically determined to be appropriate, by any route.

[0135] The appropriate dosage of the antibody, antigen-binding fragment thereof, or multispecific antibody of the present disclosure for the prevention or treatment of a disease will vary depending on the type of disease being treated, the type of antibody, the severity and course of the disease, whether the antibody is administered for prophylactic or therapeutic purposes, previous therapies, the patient's clinical history and response to the antibody, and the judgment of the attending physician. The antibody is suitably administered to the patient at one time or over a series of treatments. Depending on the type and severity of the disease, for example, about 1 μg / kg to 100 mg / kg of antibody may be an initial candidate dosage for administration to the patient, whether by one or more separate administrations or by continuous infusion. One typical daily dosage may range from about 1 μg / kg to 100 mg / kg or more, depending on the factors mentioned above. For repeated administration over several or more days, depending on the condition, treatment will generally be sustained until a desired suppression of disease symptoms occurs. Such doses may be administered intermittently, for example, every week or every three weeks (e.g., so that the patient receives from about 2 to about 20 doses). An initial high loading dose, followed by one or more lower doses, can be administered. However, other dosage regimes may be useful, and the progress of this therapy is easily monitored by conventional techniques and assays.

[0136] Combination therapy In one embodiment, the CD137 antibodies, antigen-binding fragments thereof, or multispecific antibodies of the present disclosure can be used in combination with other therapeutic agents. Other therapeutic agents that can be used with the CD137 antibodies of the present disclosure include chemotherapeutic agents (e.g., paclitaxel or paclitaxel agents (e.g., Abraxane®), docetaxel, carboplatin, topotecan, cisplatin, irinotecan, doxorubicin, lenalidomide, 5-azacytidine, ifosfamide, oxaliplatin, pemetrexed disodium, cyclophosphamide, etoposide, dextran, cyclophosphamide ... vincristine, fludarabine, vincristine, bendamustine, chlorambucil, busulfan, gemcitabine, melphalan, pentostatin, mitoxantrone, pemetrexed disodium), tyrosine kinase inhibitors (e.g., EGFR inhibitors (e.g., erlotinib), multikinase inhibitors (e.g., MGCD265, RGB-286638), CD20 targeting agents (e.g., rituximab, ofatumumab, RO507 2759, LFB-R603), CD52 targeting agents (e.g., alemtuzumab), prednisolone, darbepoetin alfa, lenalidomine, Bcl-2 inhibitors (e.g., oblimersen sodium), Aurora kinase inhibitors (e.g., MLN8237, TAK-901), proteasome inhibitors (e.g., bortezomib), CD19 targeting agents (e.g., MEDI-551, MOR208), MEK inhibitors (e.g., ABT -348), JAK-2 inhibitors (e.g., INCB018424), mTOR inhibitors (e.g., temsirolimus, everolimus), BCR / ABL inhibitors (e.g., imatinib), ET-A receptor antagonists (e.g., ZD4054), TRAIL receptor 2 (TR-2) agonists (e.g., CS-1008), EGEN-001, Polo-like kinase 1 inhibitors (e.g., BI672).

[0137] The anti-CD137xTAA antibody of the present disclosure can be used in combination with other therapeutic agents, such as other immune checkpoint antibodies. Such immune checkpoint antibodies can include anti-PD1 antibodies. Anti-PD1 antibodies can include, but are not limited to, tislelizumab, pembrolizumab, or nivolumab. Tislelizumab is disclosed in US 8,735,553. Pembrolizumab (formerly known as MK-3475), disclosed in US 8,354,509 and US 8,900,587, is a humanized IgG4-K immunoglobulin that targets the PD1 receptor and inhibits the binding of the PD1 receptor ligands PD-L1 and PD-L2. Pembrolizumab has been approved for the indications of metastatic melanoma and metastatic non-small cell lung cancer (NSCLC), and is under clinical investigation for the treatment of head and neck squamous cell carcinoma (HNSCC) and refractory Hodgkin's lymphoma (cHL). Nivolumab (disclosed by Bristol-Meyers Squibb) is a fully human IgG4-K monoclonal antibody. Nivolumab (clone 5C4) is disclosed in U.S. Patent No. 8,008,449 and WO2006 / 121168. Nivolumab is approved for the treatment of melanoma, lung cancer, kidney cancer, and Hodgkin's lymphoma.

[0138] Pharmaceutical Compositions and Formulations Also provided are compositions, including pharmaceutical preparations, that include anti-CD137 antibodies, their antigen-binding fragments, multispecific antibodies, or polynucleotides that include sequences that code for anti-CD137 antibodies, their antigen-binding fragments, or multispecific antibodies. In certain embodiments, the compositions include one or more CD137 antibodies or their antigen-binding fragments that bind to CD137, or one or more polynucleotides that include sequences that code for one or more CD137 antibodies or their antigen-binding fragments that bind to CD137. These compositions can further include suitable carriers, such as pharma- ceutically acceptable excipients, including buffers, that are well known in the art.

[0139] Pharmaceutical formulations of the anti-CD137 antibodies or antigen-binding fragments thereof described herein are prepared in the form of lyophilized formulations or aqueous solutions by mixing such antibodies or antigen-binding fragments having the desired degree of purity with one or more optional pharma- ceutically acceptable carriers (Remington's Pharmaceutical Sciences 16th edition, Osol, A. Ed. (1980)). Pharmaceutically acceptable carriers are generally non-toxic to recipients at the dosages and concentrations used, and include, but are not limited to, buffers such as phosphate, citric acid, and other organic acids, antioxidants including ascorbic acid and methionine, preservatives (octadecyldimethylbenzylammonium chloride, hexamethonium chloride, benzalkonium chloride, benzethonium chloride, phenol, butyl or benzyl alcohol, alkyl parabens such as methyl or propyl paraben, catechol, resorcinol, cyclohexanol, 3-pentanol, and m-cresol), low molecular weight (approximately Polypeptides (less than 10 residues), proteins, such as serum albumin, gelatin, or immunoglobulins, hydrophilic polymers such as polyvinylpyrrolidone, amino acids such as glycine, glutamine, asparagine, histidine, arginine, or lysine, monosaccharides, disaccharides, and other carbohydrates including glucose, mannose, or dextrins, chelating agents such as EDTA, sugars such as sucrose, mannitol, trehalose, or sorbitol, salt-forming counterions such as sodium, metal complexes (e.g., Zn-protein complexes), and / or non-ionic surfactants such as polyethylene glycol (PEG). Exemplary pharmaceutically acceptable carriers herein further include interstitial drug dispersion agents, such as soluble neutral active hyaluronidase glycoproteins (sHASEGPs), e.g., human soluble PH-20 hyaluronidase glycoproteins, e.g., rHuPH20 (HYLENEX®, Baxter International, Inc.). Certain exemplary sHASEGPs and methods of use, including rHuPH20, are described in U.S. Pat. Nos. 7,871,607 and 2006 / 0104968.In one embodiment, the sHASEGP is combined with one or more additional glycosaminoglycanases, such as chondroitinases.

[0140] Exemplary lyophilized antibody formulations are described in US Patent No. 6,267,958. Aqueous antibody formulations include those described in US Patent No. 6,171,586 and WO2006 / 044908, the latter formulations containing histidine-acetate buffer. Sustained release formulations can be prepared. Suitable examples of sustained release preparations include semipermeable matrices of solid hydrophobic polymers containing antibody, these matrices being in the form of shaped articles, such as films or microcapsules. The formulations used for in vivo administration are generally sterile. Sterility can be easily achieved, for example, by filtration through sterile filtration membranes. EXAMPLES

[0141] Example 1. Construction of a synthetic human VH antibody repertoire The synthetic library was constructed essentially using germline 3-23 (SEQ ID NO: 45 and 46). Randomization of the heavy chain CDRs (HCDRs) was performed by combinatorial mutagenesis using degenerate oligonucleotides (Table 2). Randomization of the HCDR1 and HCDR2 regions was performed via multiple site-directed mutagenesis by polymerase chain reaction as described by Meetei (Meetei et al., (1998) Anal Biochem, 264, 288-91; Meetei et al., (2002) Methods Mol Biol, 182, 95-102). For the CDR3 region, degenerate oligonucleotides of different lengths from 8 to 14 (Kabat definition) were synthesized (Invitrogen) and diversity was introduced by splice overlap extension PCR. The PCR products after the mutagenesis step were double digested with NcoI / NotI and ligated into the phagemid vector pCANTAB-5E. The repertoire was then transformed into Escherichia coli TG1 bacteria and verified by DNA Sanger sequencing of random clones (>96 clones analyzed).Phages were purified directly from the culture supernatant by two precipitations with PEG / NaCl after a rescue step using KM13 helper phage.

[0142] The library was designed to mimic the amino acid distribution commonly observed in the human repertoire, especially in the HCDR1 and HCDR2 regions. It has been demonstrated that the introduction of negatively charged amino acids at the CDR1 position not only significantly improves colloidal stability but also improves the expression and purification yield of human VH domains (Dudgeon et al., (2013) Protein Eng Des Sel, 26, 671-4; Dudgeon et al., (2012) Proc Natl Acad Sci USA, 109, 10879-84). Therefore, two different degenerate oligonucleotides were designed for HCDR1 randomization containing a high percentage of negatively charged amino acids. For HCDR3 diversification, NNY and NNK were used to obtain maximum repertoire diversity (Table 2). Furthermore, individual sub-libraries with defined HCDR3 lengths (Kabat definition) were constructed (Table 3). After transformation into E. coli bacteria, the total size of the sub-libraries was 1.38 × 10 11 A library of

[0143] [Table 2-1]

[0144] [Table 2-2]

[0145] Diverse positions are shown in bold (encoded amino acids (codons in brackets)): ABN, 25% Ile, 16.67% Arg, Ser, 8.33% Met, and 33.33% Thr; TTB, 66.67% Phe, and 33.34% Leu; NMC 12.5% ​​Ala, 12.5% ​​Asp, 12.5% ​​His, 12.5% ​​Asn, 12.5% ​​Pro, 12.5% ​​Ser. , 12.5% ​​Thr, 12.5% ​​Tyr; VAH, 11.11% Glu, Lys, Gln and 22.22% Asp, Asn, His; DTR, 16.67% Ile, 33.33% Leu, 16.67% Met and 33.33% Val; RSC, 25% Ala, 25% Gly, 25% Ser and 25% Thr; WBK, 8.33% Cys, Phe, Ile, Leu, Met, Arg, Trp, 25% Ser, and 16.67% Thr; RRH, 8.33% Glu, Lys, Arg, 16.66% Asp, Gln, Ser, and 25% Gly; RNC, 12.5% ​​Ala, 12.5% ​​Asp, 12.5% ​​Gly, 12.5% ​​Ile, 12.5% ​​Gln, 12.5% ​​Arg, 12.5% ​​Ser, and 12.5% ​​Thr; NTC, 25% Ph e, 25%Leu, 25%Val, 25%Ile; GAB, 66.67%Asp, 33.33%Glu; WWC, 25%Phe, 25%ILe, 25%Gln, 25%Tyr; NNY , 6.25%Ala, Cys, Asp, Phe, Gly, His, ILe, Leu, Gln, Phe, Arg, Thr, Val, Tyr and 12.5%Ser; NNK total 20AAs.

[0146] [Table 3]

[0147] Example 2. Production of recombinant proteins and stable cell lines CD137 recombinant protein for phage campaigns and binding assays To discover VH domain antibodies against CD137 with cross-binding to human and Macaca mulatta CD137 but without off-target binding to other human TNF receptor members, several recombinant proteins were designed and expressed for phage panning and screening (see Table 4). The cDNA coding region of full-length human CD137 (SEQ ID NO: 47) was ordered based on the CD137 GenBank sequence (Accession No: NM_001561.4, gene available from Sinobio, Catalog No. HG10041-M). Human CD137 ligand (TNFSF9) (SEQ ID NO: 57) was ordered based on (Accession No: NM_003811.3, gene available from Sinobio, Catalog No. HG15693-G). Monkey (Macaca mulatta) CD137 (SEQ ID NO: 63) was ordered according to (Accession No: NM_001266128.1, gene available from Genscript, catalog number OMb00270). Full-length human CD40 (SEQ ID NO: 69) was ordered according to (Accession No: NM_001250.4, gene available from Sinobio, catalog number HG10774-M). OX40 (SEQ ID NO: 75) was ordered according to (Accession No: NM_003327.2, gene available from Sinobio, catalog number HG10481-UT). Briefly, the coding regions of the extracellular domain (ECD) consisting of amino acids (AA) 24-183 of huCD137 (SEQ ID NO: 49), the coding region of the ECD consisting of AA 71-254 of human CD137 ligand (SEQ ID NO: 59), the coding region of the ECD consisting of AA 24-186 of cynoCD137 (SEQ ID NO: 65), and the coding region of the ECD consisting of AA 1-194 of human CD40 (SEQ ID NO: 71) were PCR amplified. The coding region of mIgG2a Fc (SEQ ID NO: 55) was PCR amplified and then combined with the ECDs of human CD137, human CD137 ligand, monkey CD137, or human CD40 by overlap PCR to generate mIgG2a Fc fusion proteins.The PCR products were then cloned into a pcDNA3.1-based expression vector (Invitrogen, Carlsbad, CA, USA) to obtain five recombinant mIgG2a Fc fusion protein expression plasmids, human CD137 ECD-mIgG2a, human CD137 ligand-mIgG2a, cyno CD137 ECD-mIgG2a, and human CD40 ECD-mIgG2a. Alternatively, the coding region of the ECD consisting of AA24-183 (SEQ ID NO: 49) of huCD137 (SEQ ID NO: 47) and the coding region of the ECD consisting of AA1-216 of human OX40 (SEQ ID NO: 77) were also cloned into a pcDNA3.1-based expression vector (Invitrogen, Carlsbad, CA, USA) fused with a 6xHis tag at the C-terminus to obtain human CD137-his and human OX40-his, respectively. For the production of recombinant fusion proteins, plasmids were transiently transfected into a HEK293-based mammalian cell expression system (developed in-house) and incubated in a CO2-free medium with a rotating shaker. 2 The cells were cultured in an incubator for 5–7 days. The supernatant containing the recombinant proteins was collected and removed by centrifugation. The recombinant proteins were purified using a Protein A column (catalog no. 17127901, GE Life Sciences) or Ni-NTA agarose (catalog no. R90115, Invitrogen). All recombinant proteins were dialyzed against phosphate-buffered saline (PBS) and stored in small aliquots in a -80°C freezer.

[0148] Stable expression in cell lines To establish a stable cell line expressing full-length human CD137 (huCD137), the huCD137 sequence was cloned into the retroviral vector pFB-Neo (cat. no. 217561, Agilent, USA). Dual-tropic retroviral vectors were generated according to a previous protocol (Zhang, et al., (2005) Blood, 106, 1544-1551). The vector containing huCD137 was transduced into Hut78 cells (ATCC, TIB-161) or NK92-mi cells (ATCC, CRL-2408) to generate huCD137-expressing cell lines, Hut78 / huCD137 or NK92-mi / huCD137. The huCD137-expressing cell lines were selected by culturing in medium containing 10% FBS with G418, and then verified by FACS.

[0149] [Table 4-1]

[0150] [Table 4-2]

[0151] [Table 4-3]

[0152] [Table 4-4]

[0153] [Table 4-5]

[0154] [Table 4-6]

[0155] [Table 4-7]

[0156] [Table 4-8]

[0157] [Table 4-9]

[0158] [Table 4-10]

[0159] [Table 4-11]

[0160] [Table 4-12]

[0161] [Table 4-13]

[0162] [Table 4-14]

[0163] Example 3. Generation of anti-huCD137 VH domain antibodies Phage display panning and screening Phage display selection was performed by phage display using standard protocols (Silacci et al., (2005) Proteomics, 5, 2340-50; Zhao et al., (2014) PLoS One, 9, e111339). Briefly, in rounds 1 and 2, 10 μg / ml of immobilized human CD137 ECD-mIgG2a in immunotubes (cat. no. 470319, ThermoFisher) was used. In rounds 3 and 4, Hut78 / huCD137 cells were used for selection. The immunotubes were blocked with 5% milk powder (w / v) in PBS supplemented with 1% Tween 20 (MPBST) for 1 h. After washing with PBST (PBS buffer supplemented with 0.05% Tween 20), 5 × 10 IgG from each sublibrary was added to the immunotubes. 12 (Round 1) or 5×10 11 Phages (round 2) were depleted with human CD40 ECD-mIgG2a in MPBST for 1 h and then incubated with antigen for 1 h. For the third and fourth rounds of selection, cell panning was performed using Hut78 / huCD137 cells (round 3) and HEK293 (ATCC, CRL-1573) cells as depleted cells. After washing with PBST, bound phages were eluted with 100 mM triethylamine (Sigma-Aldrich). Eluted phages were used to infect mid-logarithmic phase E. coli TG1 bacteria and plated on TYE agar plates supplemented with 2% glucose and 100 μg / ml ampicillin. After four rounds of selection, individual clones were selected and phage-containing supernatants were prepared using standard protocols. Anti-huCD137 VH domain antibodies were screened using phage ELISA and FACS.

[0164] For phage ELISA, Maxisorp immunoplates were coated with antigen and blocked with 5% milk powder (w / v) in PBS buffer. Phage supernatants were blocked with MPBST for 30 min and added to the wells of the ELISA plate for 1 h. After washing with PBST, bound phages were detected using HRP-conjugated anti-M13 antibody (GE Healthcare) and 3,3',5,5'-tetramethylbenzidine substrate (cat. no. 00-4201-56, eBioscience, USA). ELISA-positive clones were further verified by flow cytometry using Hut78 / huCD137 cells. CD137-expressing cells (10 5 Cells / well) were incubated with ELISA-positive phage supernatants and subsequently bound with Alexa Fluro-647-labeled anti-M13 antibody (GE Healthcare). Cell fluorescence was quantified using a flow cytometer (Guava easyCyte™ 8HT, Merck-Millipore, USA).

[0165] Clones that showed positive signals in the FACS screening and bound to both huCD137 and cynoCD137, but not to huOX40 or huCD40, were selected and sequenced. Approximately 76 unique sequences were identified from 93 positive clones, and most of the clones were derived from two sublibraries, HC-10 and HC-13 (Figures 1A-1B).

[0166] Expression and purification of Fc-fused VH antibodies The VH sequences were analyzed by comparing sequence homology and grouped based on sequence similarity. Complementarity determining regions (CDRs) were defined by sequence annotation and internet-based sequence analysis based on Kabat (Wu and Kabat (1970) J. Exp. Med. 132:211-250) and IMGT (Lefranc (1999) Nucleic Acids Research 27:209-212) systems. The amino acid and DNA sequences of two representative top clones BGA-7207 and BGA-4712 are shown in Table 5 below. After sequence confirmation by SPR and binding curve analysis, anti-huCD137 VH domain antibodies were constructed in human Fc fused VH antibody format (VH-Fc) using an in-house developed expression vector. As shown in Figure 2A, the VH domain antibodies were fused to the N-terminus of human Fc via a G4S (SEQ ID NO: 246) linker in between. An Fc null version (inactive Fc with no FcγR binding) of human IgG1 (SEQ ID NO: 85) was used. Expression and preparation of Fc fusion VH antibodies was achieved by transfection into 293G cells and purification using a Protein A column (Cat. No. 17543802, GE Life Sciences). Purified antibodies were concentrated to 0.5-5 mg / mL in PBS and stored in aliquots at -80°C in a freezer.

[0167] [Table 5-1]

[0168] [Table 5-2]

[0169] [Table 5-3]

[0170] Example 4. Functional screening of anti-huCD137 VH domain antibodies Using the supernatant containing VH-Fc protein, functional screening was first applied to the selected anti-huCD137 VH domain antibodies with strong agonism. Briefly, 96-well white / clear bottom plates (Thermo Fisher) were pre-incubated with 3 μg / ml anti-hu CD3 (Invitrogen, Cat. No. 16-0037-85) at 50 μl / well for 5 min, then washed with PBS buffer. Next, 5 × 10 Hut78 / huCD137 cells were cultured in a 5% PBS buffer. 5 Cells were resuspended at 1000 cells / ml and plated directly onto pre-coated plates at 50 μl / well (25,000 cells per well). Supernatants containing the various VH-Fc proteins were mixed with the cells. Alternatively, for purified VH domain antibodies with Fc fusions, a dose titration of purified VH-Fc protein preparations was added in duplicate at 25, 5, 1, 0.2, 0.04, 0.008 or 0.0016 μg / ml at 50 μl / well. Goat anti-hu IgG (H&L) polystyrene particles (6.46 um) (Cat. No. HUP-60-5, Spherotech) were added as a cross-linker. The assay plates were incubated overnight at 37°C and the concentration of IL-2 was measured after 24 hours. Data was plotted as the fold increase of IL-2 compared to the concentration in wells containing medium alone. Figure 2B shows representative screening results using supernatants containing VH-Fc proteins, demonstrating that one of the clones, BGA-4712, can stimulate IL-2 production in Hut78 / huCD137 cells in a dose-dependent manner (Figure 2C).

[0171] Example 5. Characterization of purified anti-huCD137 VH domain antibodies Characterization of purified antibodies by ELSIA For antigen ELISA, Maxisorp immunoplates were coated with antigen and blocked with 3% BSA (w / v) in PBS buffer (blocking buffer). Monoclonal VH domain antibodies were blocked with blocking buffer for 30 min and added to the wells of the ELISA plate for 1 h. After washing with PBST, bound antibodies were detected using an HRP-conjugated anti-human IgG antibody (Sigma, A0170) and 3,3',5,5'-tetramethylbenzidine substrate (cat. no. 00-4201-56, eBioscience, USA). All selected clones were shown to cross-react with cynoCD137 and not bind to human OX40 ECD and human CD40 ECD.

[0172] Characterization of purified antibodies by SPR analysis Characterization of anti-huCD137 VH domain antibodies was performed by SPR assay using a BIAcore™ T-200 (GE Life Sciences). Briefly, anti-human IgG (Fc) antibodies were immobilized on an activated CM5 biosensor chip (catalog no. BR100839, GE Life Sciences). Anti-huCD137 domain antibodies were flowed over the chip surface and captured by anti-human IgG (Fc) antibodies. Then, serial dilutions (6.0 nM to 2150 nM) of human CD137 ECD-mIgG2a were flowed over the chip surface and the change in surface plasmon resonance signal was analyzed to estimate the association rate (k) using a one-to-one Langmuir binding model (BIA Evaluation Software, GE Life Sciences). on ) and dissociation rate (k off The equilibrium dissociation constant (K D ) as the ratio k off / k on It was calculated as:

[0173] Characterization of purified antibodies by flow cytometry Flow cytometry shows that human CD137 + Expressing cells (10 5Cells / well) were incubated with various concentrations of purified VH domain antibodies, and then bound with Alexa Fluro-647-labeled anti-hu IgG Fc antibody (catalog number 409320, BioLegend, USA). The fluorescence of cells was quantified using a flow cytometer (Guava easyCyte™ 8HT, Merck-Millipore, USA). Ligand competition was also applied to the flow cytometry-based assay. Briefly, Hut78 / huCD137 was incubated with Fc-fused VH domain antibodies (VH-Fc) in the presence of serially diluted human CD137 ligand-mIgG2a, followed by detection with Alexa Fluro-647-labeled anti-hu IgG Fc antibody (catalog number 409320, BioLegend, USA).

[0174] Selected VH domain antibodies were then characterized for affinity, cell binding, and ligand competition. SPR studies, FACS analysis, and ligand competition results of one representative top clone, BGA-4712, are shown in Figure 3.

[0175] Example 6. Construction of CD137xCEA multispecific antibody using anti-CD137 VH domain antibody BGA-4712 and anti-CEA antibody To explore CD137-based mechanisms of action (MOA) that are potentially more effective than single antibody therapy, multiple multispecific formats utilizing anti-huCD137 VH domain antibodies have been constructed and tested. Here, multiple formats have been employed to create CD137-based T cell engagers (TCEs), where the first antigen-binding domain targets a tumor-associated antigen (TAA) and the second antigen-binding domain targets the CD137 activating receptor. For example, the first antigen-binding domain of the anti-CEA antibody (SEQ ID NO: 87 and 89) was used to pair with the second antigen-binding domain of the anti-huCD137 VH domain antibody BGA-4712 (SEQ ID NO: 17) in a specifically defined format (Table 6) as shown below. For this construct, an inactive Fc was used (SEQ ID NO: 85). Expression and preparation of these multispecific antibodies was achieved by transfection into 293G cells and purification using a Protein A column (catalog no. 17543802, GE Life Sciences). The purified antibodies were concentrated to 0.5-5 mg / mL in PBS and stored in aliquots at -80°C in a freezer.

[0176] Format A (A-CD137xCEA) Format A provides a symmetric IgG-like multispecific molecule with a FabxVH configuration. As shown in Figure 4A, the anti-huCD137 VH domain antibody BGA-4712 was fused to the C-terminus of the Fc (CH3 domain) of an anti-CEA antibody with one G4S linker between them (SEQ ID NOs: 89 and 91).

[0177] Format B (B-CD137xCEA) Format B also provides a symmetric IgG-like multispecific molecule with a FabxVH configuration: the anti-huCD137 VH domain antibody BGA-4712 was fused to the c-terminus of the light chain (Cκ) of an anti-CEA antibody with one G4S linker between them (SEQ ID NOs: 87 and 93) as shown in Figure 4B.

[0178] Format C (C-CD137xCEA) Format C provides a symmetric VH antibody-like multispecific molecule with a Fab×VH configuration. As shown in Figure 4C, the Fab region of an anti-CEA antibody was fused to the N-terminus of the VH of the anti-huCD137 VH domain antibody BGA-4712 with one G4S linker between them (SEQ ID NOs: 89 and 95).

[0179] Format D (D-CD137xCEA) Format D also provides a symmetric IgG-like multispecific molecule with a FabxVH configuration: the anti-huCD137 VH domain antibody BGA-4712 was fused to the N-terminus of the heavy chain (Vh) of an anti-CEA antibody with one G4S linker between them (SEQ ID NOs: 89 and 97) as shown in Figure 4D.

[0180] The yields and biochemical properties of the various CD137xCEA multispecific antibodies are summarized in Table 7. For the two molecules A-CD137xCEA and D-CD137xCEA, both are over 95% monomeric based on the SEC-HPLC profiles (Table 7). Flow cytometry-based assays demonstrate that format A has very little loss of affinity of the anti-CEA arm, whereas format D has a significant loss of affinity of the anti-CEA arm (Figure 5A). It also demonstrates that format A has a loss of affinity of the CD137 arm, while format D has little or no effect (Figure 5D).

[0181] [Table 6-1]

[0182] [Table 6-2]

[0183] [Table 6-3]

[0184] [Table 6-4]

[0185] [Table 6-5]

[0186] [Table 6-6]

[0187] [Table 6-7]

[0188] [Table 6-8]

[0189] [Table 6-9]

[0190] [Table 7]

[0191] Example 7. CD137-based multispecific antibody A-CD137xCEA activates CD137 in a TAA (tumor-associated antigen)-dependent manner CD137-based multispecific antibodies induce CD137 activation in CD137-expressing cells CEA + CD137 in response to stimulation by tumor cells +To test the ability of CD137-based multispecific antibodies to induce cellular responses, Hut78 / huCD137 was used to test CD137 activation. CEA-expressing CT26 (CT26 / CEA) cells were generated by retroviral transduction into CT26 cells (ATCC CRL-2638) according to a previously described protocol (Zhang et al., Blood. 2005 106(5):1544-51). Hut78 / huCD137 cells were co-cultured overnight with CT26 / CEA or CT26 (CEA-negative) cells in the presence of CD137xCEA multispecific constructs in OKT3-precoated 96-well plates, and interleukin-2 (IL-2) was measured as an indicator of CD137 activation in Hut78 / huCD137 cells. As shown in Figure 6A, A-CD137xCEA increased the expression of CEA + In the presence of CT26 / CEA cells, Hut78 / huCD137 cells were induced to secrete IL-2 in a dose-dependent manner. The induction of IL-2 was enhanced by CEA. + This was not observed in the absence of CT26 / CEA cells.

[0192] CD137-based multispecific antibodies induce CD137 activation in human peripheral blood mononuclear cells (PBMCs) Human peripheral blood mononuclear cells (PBMCs) were isolated from whole blood of healthy donors by Ficoll (Histopaque-1077, Sigma-St. Louis MO) separation. OS8-expressing HEK293 (HEK293 / OS8) cells were generated by retroviral transduction into HEK293 (ATCC CRL-1573) according to a previously described protocol (Zhang et al., Blood. 2005 106(5):1544-51). CD137xCEA polyspecific antibody was used to express CEA. + To examine whether T cells can be activated in the presence of tumor cells, PBMCs (2 × 10 5A-CD137xCEA (100 / well) were co-cultured with HEK293 / OS8 and CT26 / CEA cells for 48 h in the presence of CD137xCEA polyspecific antibody. Activation of CD137 by CD137xCEA polyspecific antibody was determined by measuring IFN-γ in PBMCs. The results showed that A-CD137xCEA could induce significant CD137 activation in PBMCs in the presence of CEA-expressing cells (Figure 6B).

[0193] Example 8. Removal of post-translational modifications Based on the BGA-4712 sequence, the multispecific construct A-CD137xCEA was engineered by introducing mutations in the HCDR and framework regions to remove post-translational modification (PTM) sites and improve expression. Substitutions included amino acid changes F28R, M29T, V35M, D62E, S75A, and N84S in the BGA-4712 VH region. As a result of the operation, A-CD137xCEA-M1 (M28T, V34M, D62E, S75A, and N84S), A-CD137xCEA-M2 (F27R, M28T, V35M, D62E, S75A, and N84S), A-CD137xCEA-M3 (M28T, D62E, S75A, and N84S), A-CD137xCEA-M4 (F27R, M28T, D62E, S75A, and N84S), A-CD137xCEA-M5 ( A-CD137xCEA-M6 (F27R, M28T, V35M, S75A, and N84S), A-CD137xCEA-M7 (M28T, S75A, and N84S), and A-CD137xCEA-M8 (F27R, M28T, S75A, and N84S) were obtained, and all antibodies had similar binding specificity to the parent A-CD137xCEA, and none of the changes disrupted binding. While maintaining specificity, amino acid composition and expression level were also taken into consideration. All mutations were made using primers containing mutations at specific positions and a site-directed mutagenesis kit (catalog number FM111-02, TransGen, Beijing, China). The desired mutations were confirmed by sequence analysis. These A-CD137xCEA mutants were tested in binding (Table 8) and functional assays as described in Example 6. For affinity measurements, A-CD137xCEA was captured by anti-human Fc surface and used in an affinity assay based on surface plasmon resonance (SPR) technology. The results of the binding profiles of the variants determined by SPR are summarized in Table 8. The binding of all the above variants to Hut78 / huCD137 cells was also confirmed (Figure 8). The results showed that A-CD137xCEA-M3 (SEQ ID NO: 101-102) in BGA-4712-M3 (SEQ ID NO: 24-25) was comparable to the parent antibody BGA-4712 (SEQ ID NO: 19-20).The sequence of BGA-4712-M3 is disclosed in Table 9. It has also been demonstrated that A-CD137xCEA-M3 was able to induce CD137 activation in a PBMC-based cytokine release assay as described above in Example 7 (Figure 8). These results indicate that A-CD137xCEA-M3 is capable of inducing CD137 activation in a PBMC-based cytokine release assay as described above in Example 7 (Figure 8). + It was further confirmed that IL-2 could be induced in a dose-dependent manner in the presence of CT26 / CEA cells, whereas in the absence of CEA-expressing cells, no induction of IL-2 was observed, and thus no activation of CD137 was observed (Figure 8).

[0194] [Table 8]

[0195] [Table 9-1]

[0196] [Table 9-2]

[0197] [Table 9-3]

[0198] Example 9. Camelization Heavy chain antibodies (VHH) are a type of single domain antibody that lack the CH1 domain and can be generated from camels and llamas (Harmsen et al., (2007) Appl Microbiol Biotechnol, 77, 13-22; Kastelic et al., (2009) J Immunol Methods, 350, 54-62).

[0199] VHHs are the smallest (about 120 amino acids) antibody fragments capable of binding to antigens. In addition to their smaller size, VHHs are usually more stable and soluble than conventional antibodies. These single domain antibodies can therefore serve as modular building blocks for bispecific and multispecific constructs (Els Conrath et al., (2001) J Biol Chem, 276, 7346-50). A "camelization" strategy has been developed to generate autonomous human VH domain antibodies (aVHs) that have favorable properties on isolated human VH domains (Riechmann et al., (1999) J Immunol Methods, 231, 25-38). It is generally believed that a series of substitutions in the germline (Val37 to Phe / Tyr, Gly44 to Glu, Leu45 to Arg, and Trp47 to Gly / Leu, Trp103 to Arg / Gly) contributed to these highly desirable properties (Vincke et al., (2009) J Biol Chem, 284, 3273-84; Nguyen et al., (2000) Embo J, 19, 921-30; Kunz et al., (2018) Sci Rep, 8, 7934). Other attempts to generate binders based on isolated human VH domains have been attempted with success (Jespers et al., (2004a) Nat Biotechnol, 22, 1161-5; Jespers et al., (2004b) J Mol Biol, 337, 893-903; Barthelemy et al., (2008) J Biol Chem, 283, 3639-54). Although many human family consensus domains readily aggregate when expressed alone, it has been demonstrated that human VH3 has more favorable properties (Ewert et al., (2002) Biochemistry, 41, 3628-36; Ewert et al., (2003) BJ Mol Biol, 325, 531-53).

[0200] To improve the biochemical and biophysical properties of the selected VH domain antibodies, a "camelization" strategy was applied based on BGA-4712-M3 (SEQ ID NO: 24-25). Amino acid composition, thermal stability (Tm), surface hydrophobicity, and isoelectronic point (pI) were taken into consideration while maintaining functional activity. Substitutions were made mainly in framework 2 (FW2) and framework 4 (FW4), e.g., V37F or Y, G44 / E, L45 / R or G or Y, and W47 / G or S or F or L or R or Y, W103 / R (Kabat definition, Table 10). Mutants were expressed in both Fc-fused VH and A-CD137xCEA multispecific antibody formats as previously described in Example 6. Substitutions without significant affinity loss were identified (Table 11). The W103R change in BGA-4712-M3 was demonstrated to exhibit improved solubility, non-specific binding, and yield, and was named BGA-7556. Collectively, these results demonstrated that BGA-7556 (SEQ ID NOs: 103-104) in format A-CD137xCEA (SEQ ID NOs: 107-108) is highly similar in binding affinity to the parent antibody BGA-4712 (SEQ ID NOs: 19-20). The sequence information of BGA-7556 is listed in Table 12.

[0201] [Table 10]

[0202] [Table 11]

[0203] [Table 12-1]

[0204] [Table 12-2]

[0205] [Table 12-3]

[0206] Example 10. Generation of affinity maturation libraries To further explore potentially effective CD137-based mechanisms of action (MOAs), we aimed to generate affinity matured BGA-4712-M3 variants with improved druggability by phage display. Furthermore, we reasoned that an affinity maturation library using the format of BGA-4712-M3 fused to the c-terminus of the CH3 domain (SEQ ID NOs: 109-110) would be most likely to yield affinity matured variants without CH3 interference, as shown in Table 13 (Figure 9). The structure of the library was previously described. Briefly, the phagemid vector pCANTAB 5E (GE Healthcare) was used by standard molecular biology techniques to construct a phagemid designed to display CH3-G4S(linker)-BGA-4712-M3 on the surface of M13 bacteriophage as a fusion with the N-terminus of a fragment of the minor coat protein of gene 3. The gene 3 sequence was preceded by an amber stop codon to allow expression of a dimer of the CH3 fusion protein directly from the phagemid clone. Using the phagemid as a template, 2.0 × 10 8 A phage display library containing 5.0 × 10 unique members was constructed. All three CDRs were randomized, with each CDR having a maximum of one mutation in each clone, except for HCDR3, which could have two simultaneous mutations. Each position was randomized with the NNK codon (IUPAC code), which codes for any amino acid or an amber stop codon. The library design of the combined VH included 5.0 × 10 6There was a potential diversity of 100 unique full-length clones, with the expected distribution of clones with 0, 1, 2, and 3 mutations, respectively, being approximately 0.02%, 1.1%, 17%, and 82%. A minority of heavy chain clones were expected to have 4 mutations due to primer design in the HCDR3 region. Randomization of the HCDR1, HCDR2, and HCDR3 regions was performed via multiple site-directed mutagenesis by polymerase chain reaction and via splice overlap extension PCR, as described by Meetei et al., (1998) Anal Biochem, 264, 288-91; Meetei et al., (2002) Methods Mol Biol, 182, 95-102. The resulting fragment was then gel purified and ligated with pCANTAB 5E after NcoI / NotI digestion. The purified ligation was transformed into TG1 bacteria by electroporation. Sequencing of 48 clones from each library confirmed randomization at each position (not shown), although due to limited sampling depth, not all amino acid mutations were observed at every position. Over 61% of the libraries contained full-length randomized clones, representing 2.0 × 10 clones generated even with moderate incorporation bias in oligonucleotide synthesis and library construction. 8 independent clones were sufficient to cover all the potential diversity of the design.

[0207] [Table 13]

[0208] Example 11. Generation of affinity matured BGA-4712-M3 variants Library Selection and Screening Generation of affinity matured BGA-4712-M3 variants was performed by phage display using standard protocols (Silacci et al., (2005) Proteomics, 5, 2340-50; Zhao et al., (2014) PLoS One, 9, e111339). In the first and second rounds of selection, heat denaturation selection (Jespers et al., (2004a) Nat Biotechnol, 22, 1161-5) was performed against immobilized human CD137 ECD-mIgG2a or human CD137 ECD-his in immunotubes (cat. no. 470319, ThermoFisher). Briefly, immunotubes were pre-coated with human CD137 ECD-mIgG2a or human CD137 ECD-HIS (10ug / ml in PBS) overnight at 4°C. The affinity maturation library was heated at 70°C for 10 min and then cooled to 4°C for 30 min. The heat-denatured phage library was incubated with the pre-coated immunotubes for 1 h. In the third and fourth rounds of selection, cell panning was performed using Hut78 / huCD137 cells with HEK293 cells as depletion cells. After four rounds of selection, individual clones were selected and phage-containing supernatants were prepared as described in standard protocols. ELISA-positive clones were sequenced and mutation sites were analyzed.

[0209] Analysis of CDR mutation frequency The frequency of mutations in each HCDR after four rounds of selection was relatively high. The mutation rates were 78.13% in HCDR1, 93.75% in HCDR2, and 96.85% in HCDR3. HCDR1 contained a more diverse array of changes. Residue 29 was mutated from Leu to Ile in 11.45% of clones and from Leu to Val in 30.21%. Positions 26, 28, 30, and 31 did not have high mutation frequencies and no obvious pattern included large hydrophobic and polar residues such as Tyr, Phe, Thr, and Asn. In HCDR2, mutations occurred in 90.63% of clones for F55. Residue 55 was mutated from Phe to Asn (22.18%), Phe to Lys (22.18%), Phe to Ser (11.46%), and Phe to Gln (9.38%) of clones. There were also a few clones containing changes to other residues such as Leu, Gly, Tyr, Thr, and His. HCDR3 underwent changes at three sites in at least 90% of the clones, two of which had additional mutations in approximately 50% of the clones. Residue 109 was mutated from Phe to a similar hydrophobic residue such as Tyr or Trp. Residue 99 was mutated from Val to Tyr (15.63%) and from Val to Ile (28.13%). Residue 110 was changed from Tyr to Thr (55.20%) and from Tyr to Leu (7.29%) of the clones. Figure 10 shows the sequences of the HCDR regions after four rounds of selection.

[0210] Expression of selected mutants All mutations were introduced into BGA-7556 (SEQ ID NO: 103-104) to generate affinity matured variants, except for BGA-3386. Of BGA-3386, mutations were introduced into BGA-4712-M3 (SEQ ID NO: 24-25). All variants were expressed in format A (A-CD137xCEA) as both monoclonal antibodies (VH-Fc) and their corresponding multispecific antibodies, as described in Examples 5 and 6. Purified antibodies were concentrated to 0.5-10 mg / mL in PBS and stored in aliquots in a -80°C freezer.

[0211] Characterization of selected mutants Affinity comparison of affinity matured clones was performed by SPR assay using BIAcore™ T-200 (GE Life Sciences) and flow cytometry as described in Example 5. Briefly, anti-human IgG (Fc) antibody was immobilized on an activated CM5 biosensor chip (catalog no. BR100839, GE Life Sciences). Anti-huCD137 monoclonal or multispecific antibodies were flowed over the chip surface and captured by anti-human IgG (Fc) antibody. Then, serial dilutions (6.0 nM to 2150 nM) of human CD137 ECD-mIgG2a were flowed over the chip surface and the change in surface plasmon resonance signal was analyzed to determine the association rate (k) using a one-to-one Langmuir binding model (BIA Evaluation Software, GE Life Sciences). on ) and dissociation rate (k off The equilibrium dissociation constant (K D ) as the ratio k off / k on Flow cytometry was performed using CD137 expressing cells (10 5 Cells / well) were incubated with various concentrations of purified antibodies and then bound with Alexa Fluro-647-labeled anti-hu IgG Fc antibody (Cat. No. 409320, BioLegend, USA). The fluorescence of the cells was quantified using a flow cytometer (Guava easyCyte™ 8HT, Merck-Millipore, USA). In addition, the affinity of the multispecific antibodies to human CEA was also measured by SPR assay using recombinant CEA protein produced in-house. Binding to CEA-expressing cells was also confirmed by flow cytometry. The sequence information is shown in Table 16, and the results of the SPR-determined binding profile of anti-huCD137 antibodies are summarized in Tables 14 and 15. Three variants with different affinities to human CD137 were selected for further characterization.

[0212] [Table 14]

[0213]

Table 15

[0214]

Table 16-1

[0215]

Table 16-2

[0216]

Table 16-3

[0217]

Table 16-4

[0218]

Table 16-5

[0219]

Table 16-6

[0220]

Table 16-7

[0221]

Table 16-8

[0222]

Table 16-9

[0223]

Table 16-10

[0224]

Table 16-11

[0225]

Table 16-12

[0226]

Table 16-13

[0227]

Table 16-14

[0228]

Table 16-15

[0229]

Table 16-16

[0230]

Table 16-17

[0231]

Table 16-18

[0232]

Table 16-19

[0233]

Table 16-20

[0234]

Table 16-21

[0235]

Table 16-22

[0236]

Table 16-23

[0237]

Table 16-24

[0238]

Table 16-25

[0239]

Table 16-26

[0240]

Table 16-27

[0241]

Table 16-28

[0242]

Table 16-29

[0243]

Table 16-30

[0244]

Table 16-31

[0245]

Table 16-32

[0246]

Table 16-33

[0247]

Table 16-34

[0248]

Table 16-35

[0249]

Table 16-36

[0250]

Table 16-37

[0251]

Table 16-38

[0252]

Table 16-39

[0253]

Table 16-40

[0254] [Table 16-41]

[0255] [Table 16-42]

[0256] [Table 16-43]

[0257] [Table 16-44]

[0258] [Table 16-45]

[0259] [Table 16-46]

[0260] Example 13. Binding profile of anti-CD137 antibody BGA-5623 BGA-5623 was generated with a human IgG1 Fc fusion and characterized for its binding kinetics by SPR assay using a BIAcore™ T-200 (GE Life Sciences). Briefly, anti-human IgG (Fc) antibody was immobilized on an activated CM5 biosensor chip (catalog no. BR100839, GE Life Sciences). Anti-huCD137 domain antibody was flowed over the chip surface and captured by anti-human IgG (Fc) antibody. Then, serial dilutions (6.0 nM to 2150 nM) of human CD137 ECD-mIgG2a or cyno CD137 ECD-mIgG2a were flowed over the chip surface and the change in surface plasmon resonance signal was analyzed to determine the association rate (k ) using a one-to-one Langmuir binding model (BIA Evaluation Software, GE Life Sciences). on ) and dissociation rate (k off The equilibrium dissociation constant (K D ) as the ratio k off / k on The binding activity of the anti-huCD137 VH domain antibodies to native huCD137 on live cells was evaluated by transfecting Hut78 cells to overexpress human CD137. Hut78 / huCD137-expressing live cells were seeded in 96-well plates and incubated with serially diluted anti-huCD137 VH domain antibodies. Goat anti-human IgG was used as the secondary antibody to detect antibody binding to the cell surface. EC of dose-dependent binding to human native CD137 50 Values ​​were determined by fitting the dose-response data to a four-parameter logistic model using GraphPad Prism™. As shown in Figure 26, BGA-5623 demonstrated high binding affinity to native CD137 on live cells.

[0261] [Table 17]

[0262] [Table 18-1]

[0263] [Table 18-2]

[0264] Example 14. Epitope mapping of BGA-5623 To characterize the binding epitope of BGA-5623, 17 amino acid residues of human CD137 were individually mutated to alanine to generate 17 single-mutation huCD137 mutants based on information from the previously reported crystal structure of CD137 (Bitra et al., (2018) J Biol Chem, 293, 9958-9969; Chin et al., (2018) Nat Commun, 9, 4679).

[0265] CD137 mutants were transiently expressed in HEK293 cells (ATCC CRL-1573) together with wild-type CD137. Their recognition and binding by BGA-5623 was analyzed by flow cytometry. Urelumab analogs (SEQ ID NOs: 199-202), generated in-house by using the publicly available sequence of urelumab, were used in the same assay to monitor the expression of CD137 mutants. In this assay, human CD137 or CD137 mutant expressing cells (10 5Cells / well) were incubated with 2 μg / ml of purified BGA-5623-mutFc (Fc fusion VH Ab) or urelumab analogs, then bound with Alexa Fluro-647 labeled anti-hu IgG Fc antibody (cat. no. 409320, BioLegend, USA). Fluorescence of cells was quantified using a flow cytometer (Guava easyCyte™ 8HT, Merck-Millipore, USA). All results were normalized using the average fluorescence reading of wild type CD137 binding signal as a standard. To simplify data analysis, if the FACS binding signal of an antibody against a particular mutant CD137 was reduced to 25% or less, the amino acid at that site was considered to be important for the epitope. As shown in FIG. 12A, the epitope of BGA-5623 has the important residues for binding to CD137 amino acids F36, I44, P47, P49 and S52.

[0266] To further explore the BGA-5623 epitope, human CD137 ECD mutants with single AA substitutions were expressed and purified and prepared for ELISA. In addition, utomirumab analog antibodies (SEQ ID NOs: 203-206) were generated in-house using the publicly available utomirumab sequence. CD137 mutants and wild-type CD137 were analyzed by direct ELISA for binding by BGA-5623. Briefly, 50 ng each of wild-type or mutant CD137 was coated onto an ELISA plate. After blocking, 100 μl of BGA-5623-mutFc, urelumab analog, or utomirumab analog antibody at a concentration of 2 μg / ml was added to the plate, and the binding signal of each antibody was detected by HRP-conjugated secondary antibody. In ELISA binding assays using wild-type or mutant huCD137, amino acids F36A, P47A, and P49A significantly impaired the binding of BGA-5623 to CD137 (Figure 12A-B). Changes at amino acid F36A only slightly reduced the binding of urelumab or utomilumab analogs, which may indicate that F36A plays an important role in the conformational integrity of CD137. In contrast, changes at amino acids P47A or P49A did not disrupt the binding of urelumab or utomilumab analogs to CD137, indicating that BGA-5623, urelumab analogs, or utomilumab have different epitopes. This data indicated that amino acids F36A, P47A, and P49A are important residues in the epitope of antibody BGA-5623. Molecular modeling of CD137, shown in FIG. 13, indicates that amino acids F36A, P47A and P49A are close to each other when CD137 is in its folded conformation, but are located in two distinct domains of CD137, CRD1 and CRD2.

[0267] [Table 19-1]

[0268] [Table 19-2]

[0269] [Table 19-3]

[0270] [Table 19-4]

[0271] Example 15. Ligand competition Human CD137 binds to its major ligand, human CD137 ligand (CD137L), with a weak affinity with an approximate Kd of three orders of magnitude M (Chin et al., (2018) Nat Commun, 9, 4679). The epitope mapping results of Example 14 above indicate that amino acid residues F36A, P47A and P49A of CD137 are important amino acid residues that constitute part of the epitope of the BGA-5623 antibody. Furthermore, the ligand binds to CD137 along the entire length of the receptor CRD-2 and the A2 motif of CRD-3, and the interface between the receptor and the ligand is mainly mediated by hydrogen bonds and van der Waals interactions (Bitra et al., (2018) J Biol Chem, 293, 9958-9969). Based on this data, it was hypothesized that the BGA-5623 antibody can block the CD137 / CD137 ligand interaction. BGA-5623 was generated using a human IgG4 Fc fusion. For the CD137 ligand competition ELISA, Maxisorp immunoplates were coated with human CD137 ECD-mIgG2a and blocked with 3% BSA (w / v) in PBS buffer (blocking buffer). The VH domain antibody BGA-5623 was blocked for 30 min with blocking buffer and added to the wells of the ELISA plate for 1 h in the presence of serially diluted human CD137 ligand ECD-mIgG2a. After washing with PBST, bound antibody was detected using an HRP-conjugated anti-human IgG antibody (Sigma, A0170) and 3,3',5,5'-tetramethylbenzidine substrate (cat. no. 00-4201-56, eBioscience, USA) (Figure 14A). In a flow cytometric assay of CD137 ligand competition, the CD137 stably transduced cell line Hut78 / huCD137 was incubated with human CD137 ligand ECD-mIgG2a in the presence of serial dilutions of BGA-5623 (IgG4) followed by detection with goat anti-mouse IgG-APC (Figure 14B). As shown in Figure 14, BGA-5623 competes with the CD137 ligand and reduces the CD137 / CD137 ligand interaction.

[0272] Example 16. Assessment of off-target specificity The off-target specificity of BGA-5623 was evaluated by ELISA. Antigen ELISA was performed as described above in Example 5. TNF receptor family members, such as TNFRSF1A (CD120a) (Cat. No. 10872-H08H, Sino Biological, China), TNFRSF1B (CD120b) (Cat. No. 10417-H08H1, Sino Biological, China), TNFRSF4 (OX40) (SEQ ID NO: 77), TNFRSF5 (CD40) (SEQ ID NO: 71), TNFRSF7 (CD27) (Cat. No. 10039-H08B1, Sino Biological, China), TNFRSF9 (CD137) (SEQ ID NO: 49) and TNFRSF18 (GITR) (Cat. No. 13643-H08H, Sino Biological, China) were coated in 96-well plates at a concentration of 10 μg / ml overnight at 4°C. BGA-5623 fused to wild type IgG1Fc (SEQ ID NO: 195) was added. As shown in Figure 15, no binding to other TNF receptor family members was observed.

[0273] Example 17. Characterization of BGA-4712 variants in multispecific antibody format A-CD137xCEA with different affinities Three BGA-4712 variants (BGA-2164, BGA-6468, and BGA-9442) with high, medium, and low affinity were selected for potency comparison. SPR studies and FACS analysis were performed as described above in Example 5 and shown in Table 20 and Figure 16. Flow cytometry revealed that human CD137 + or human CEA + Expressing cells (10 5Cells / well) were incubated with various concentrations of purified VH domain antibodies and then bound with Alexa Fluro-647 labeled anti-hu IgG Fc antibody (cat. no. 409320, BioLegend, USA). The fluorescence of the cells was quantified using a flow cytometer (Guava easyCyte™ 8HT, Merck-Millipore, USA). There was no significant difference in CEA binding between the three multispecific antibodies tested, but as expected, different binding affinities to human CD137 were observed by flow cytometry. Next, the PBMC-based cytokine release assay described above in Example 7 was applied to evaluate the potency of these BGA-4712 variants with different affinities in the multispecific antibody format A-CD137xCEA. As shown in Table 20, the CD137 activation induced by these variants is proportional to the increased affinity of the CD137 arm. Considering other biophysical properties such as non-specific binding and aggregation in SEC-HPLC, BGA-5623 was selected for further characterization and investigation of parameters that may affect CD137 activation.

[0274] [Table 20]

[0275] Example 18. Parameters that can affect in vitro CD137 activation The above data showed that in addition to affinity, receptor density and epitope location on CD137 as well as molecular format, there are also other important parameters that can significantly affect cytokine release (Il-2 and IFN-γ), such as module ratio, module orientation, linker length and Fc function. Therefore, to inform the rational design of CD137-based multispecific antibodies, we adopted a systematic approach to investigate how these parameters affect CD137 agonism. Expression and preparation of these multispecific antibodies were performed as described in Example 6.

[0276] First, CD137xCEA multispecific antibody variants with different module ratios such as 2:4, 1:1, 1:2 were constructed, namely BE-718 (A-BGA-5623-BGA-5623) (SEQ ID NOs: 207 and 89), BE-942 (ZW1+1) (SEQ ID NOs: 211, 213, and 215), which is a 1+1 configuration of BGA-5623, and BE-755 (ZW1+2) (SEQ ID NOs: 211, 213, and 217), which is a 1+2 configuration of BGA-5623 (Figure 17). For the antibody constructs with the designation "ZW", an inactive Fc was used for these multispecific antibodies and the FabxVH configuration was assembled using Zymeworks' Azymetric™ platform. In this configuration, ZW1 mutations (chain A: T350V / L351Y / F405A / Y407V; chain B: T350V / T366L / K392L / T394W) were introduced into the CH3 domain of the heavy chain to allow efficient heterodimer formation (Von Kreudenstein et al., (2013) Mabs 5(5):646-54). For BE-189 (A-BGA-5623) (SEQ ID NOs: 167 and 89), which represents a multispecific antibody with a module ratio of 2:2, it was possible to investigate how the module ratio affects cytokine release. As described in Example 7 above, the high CEA expressing cell line CT26 / CEA was cultured with PBMCs (2×10 5 BE-189 (format A-BGA-5623) activates CD137 in a CEA-dependent manner. In contrast, BE-718 (format A-BGA-5623-BGA-5623) activates CD137 in the absence of CEA-expressing cells.

[0277] Next, we investigated how the orientation of the modules and Fc function affect the activation of CD137. In this experiment, we constructed BE-740 (A-IgG1-BGA-5623) (SEQ ID NOs: 209 and 89), which is exactly the same format as A-BGA-5623 except that wild-type IgG1 Fc was used to replace the inactive Fc. We also constructed BE-562 (E-muFc-BGA-5623) (SEQ ID NOs: 219 and 89) and BE-375 (E-IgG1-BGA-5623) (SEQ ID NOs: 221 and 89), respectively. As shown in Figure 19, these two multispecific antibodies share the same pair of anti-CEA antibody and anti-huCD137 VH domains (CEA and BGA-5623) as A-BGA-5623 and A-IgG1-BGA-5623, but in the opposite orientation. A PBMC-based cytokine release assay was used to quantify the potency of CD137 activation, as described in Example 7. Based on the in vitro results, A-BGA-5623 and A-IgG1-BGA-5623 were demonstrated to be more potent in CD137 activation than E-muFc-BGA-5623 and E-IgG1-BGA-5623. Furthermore, based on this experiment, Fc function appears to have minimal impact on CD137 activation (Figure 20).

[0278] Finally, the linker linking the Fc and VH domain antibodies was evaluated for its effect on CD137 activation. A-(G4S)3-BGA-5623 (BE-244) (SEQ ID NOs: 223 and 89) and G4S were (G4S) 3 (SEQ ID NO: 251). Again, potency was compared using a PBMC-based cytokine release assay. As shown in Figure 21 and Table 21, linker length has minimal effect on CD137 activation.

[0279] [Table 21]

[0280] [Table 22-1]

[0281]

Table 22-2

[0282]

Table 22-3

[0283]

Table 22-4

[0284]

Table 22-5

[0285]

Table 22-6

[0286]

Table 22-7

[0287]

Table 22-8

[0288]

Table 22-9

[0289]

Table 22-10

[0290]

Table 22-11

[0291] [Table 22-12]

[0292] [Table 22-13]

[0293] Example 19. In vivo efficacy of single agent CD137xCEA multispecific antibodies To measure the in vivo efficacy of the CD137×CEA multispecific antibodies BE-189 and BE-740 against CEA+ tumor cells, CT26 / CEA cells (1×10 6 ) was subcutaneously injected into humanized CD137 mice on a BALB / c background. BE-189 (3 mg / kg), BE-740 (3 mg / kg), anti-CEA Ab (SEQ ID NOs: 87 and 89) (3 mg / kg), urelumab analog (3 mg / kg) or vehicle control was administered twice weekly starting on the day of tumor injection (6 mice per group). Compared to the vehicle control, BE-189 (A-BGA-5623) and urelumab analog induced significant inhibition of tumor growth (P<0.001), as shown in FIG. 22.

[0294] Example 20. Design of tumor-targeted CD137 agonists Agonistic anti-huCD137 antibodies have shown toxicity in clinical settings, which may indicate that systemic FcγR cross-linking is not ideal for CD137 activation. The aim was to achieve potent CD137 stimulation specifically at the tumor site without systemic CD137 activation for a wide range of cancers. To overcome the dependency of FcγR cross-linking, a TAA×CD137 multispecific antibody was generated with the following characteristics shown in FIG. 23. This specific construct contained a modular ratio of 2:2 IgG fusion-like multispecific antibody format, a bivalent F(ab')2 fragment that binds to a TAA, e.g., CEA, GPC3, Claudin6 or Trop2, a VH domain fragment with a fusion at the C-terminus of CH3 that binds to huCD137, and a null version of the Fc of huIgG1 that has no FcγR binding but retains FcRn binding. The yield and biochemical properties of the generated tumor-targeted CD137 agonists are summarized in Table 23, and sequence information is shown in Table 24. To construct CD137xCEA, we selected variants of anti-CEA antibodies with improved biophysical properties and higher yields. Figures 24A-D show representative cell binding results for CD137xCEA (Figures 24A-B) and Glypican3 (GPC3)xCD137 (Figures 24C-D).

[0295] [Table 23]

[0296] [Table 24-1]

[0297] [Table 24-2]

[0298] [Table 24-3]

[0299] [Table 24-4]

[0300] [Table 24-5]

[0301] [Table 24-6]

[0302] [Table 24-7]

[0303] [Table 24-8]

[0304] [Table 24-9]

[0305] Example 21. CD137xCEA induces T cell activation in a CEA-dependent manner The functionality of the CD137xCEA multispecific antibody BE-146 was evaluated in various in vitro experiments. First, human peripheral blood mononuclear cells (PBMCs) from healthy donors were used to activate human T cells with CD137xCEA and HEK293 / OS8, which provided the first stimulatory signal. PBMCs were isolated from whole blood of healthy donors by Ficoll (Histopaque-1077, Sigma-St. Louis MO) separation. CD137xCEA inhibited the CEA + To examine whether the presence of tumor cells could induce cytokine release from human PBMCs, PBMCs (1x10 5 / well) CEA + MKN45 cells (2x10 5 / well) and HEK293 / OS8 (1x10 5PBMCs were co-cultured with CD137xCEA (100 / well) cells in 96-well V-bottom plates for 2 days. IL-2 and IFN-γ release from PBMCs was measured by ELISA, which is shown diagrammatically in Figure 25A. The results showed that CD137xCEA was able to induce significant cytokine release (Figures 25B-C). PBMCs from two donors were tested. Results are shown as mean ± SD of duplicates.

[0306] We next investigated whether CD137xCEA could enhance antigen-specific CD8+ T cell function. Human peripheral blood mononuclear cells (PBMCs) were isolated from whole blood of healthy donors by Ficoll (Histopaque-1077, Sigma-St. Louis MO) separation. T cells were isolated using a Human Pan T cell isolation kit (Miltenyi, catalog no. 130-096-535). To investigate whether BE-146 could induce cytokine release from human T cells in the presence of CEA+ tumor cells, T cells (1x10 5 / well) with CEA + MKN45 cells (2x10 5 / well) and HEK293 / OS8 (1x10 5 T cells were co-cultured with 100% T cells (100% IL-2 / well) in 96-well V-bottom plates for 2 days (Figure 26A). IL-2 and IFN-γ release from T cells was measured by ELISA. The results showed that the multispecific antibody BE-146 could induce significant IL-2 (Figure 26B) and IFN-γ (Figure 26C) release.

[0307] We next investigated whether CD137xCEA could trigger a CEA-dependent response. Human peripheral blood mononuclear cells (PBMCs) were isolated from whole blood of healthy donors by Ficoll (Histopaque-1077, Sigma-St. Louis MO) separation. To investigate whether CD137xCEA-induced cytokine release from human PBMCs was CEA-dependent, PBMCs (1x10 5 / well) and HEK293 or CEA-overexpressing HEK293 cells (HEK293 / CEA) (1x10 5 / well), and HEK293 / OS8 (1x10 5The cells were co-cultured with CEA-overexpressing HEK293 cells (1 / well) in 96-well V-bottom plates for 2 days. IL-2 and IFN-γ release from PBMCs was measured by ELISA. The results showed that the multispecific antibody BE-146 was able to induce significant IL-2 and IFN-γ release from PBMCs against CEA-overexpressing HEK293 cells, but not against non-CEA-transduced HEK293 cells (Figures 27A-B).

[0308] Furthermore, a series of experiments were performed to examine whether the response induced by the CD137xCEA construct could be blocked by soluble CEA. Human peripheral blood mononuclear cells (PBMCs) were isolated from whole blood of healthy donors by Ficoll (Histopaque-1077, Sigma-St. Louis MO) separation. To examine whether BE-146-induced cytokine release from human PBMCs could be blocked by soluble CEA, PBMCs (1x10 5 / well) MKN45 (1x10 5 / well) and HEK293 / OS8 (1x10 5 PBMCs were co-cultured with 1000μg / well (100μg / well) cells in 96-well V-bottom plates for 2 days. IL-2 and IFN-γ release from PBMCs was measured by ELISA. Results showed that the multispecific antibody BE-146 induced IL-2 (Figure 28A) and IFN-γ (Figure 28B) release from PBMCs, and this release was not significantly blocked by 50ng / ml or 500ng / ml soluble CEA. Only very high concentrations of CEA (5000ng / ml) led to a decrease.

[0309] These data indicate that CD137xCEA in the presence of CD3ε or T cell receptor stimulation induces potent CEA-dependent T cell activation.

[0310] Example 22. CD137xCEA reduces tumors in vivo To measure the in vivo efficacy of the CD137×CEA multispecific antibody BE-146 against CEA+ tumor cells, MC38 / CEA cells (1×10 6 ) was injected subcutaneously into humanized CD137 mice on a C57BL / 6 background. Mice developed tumors with a mean tumor volume of approximately 100 mm on day 5 post-injection. 3 Patients were randomized when they reached a mean age of 18. BE-146 (0.5 mg / kg) or urelumab analog (0.5 mg / kg) or vehicle control was administered once a week starting on day 5. Compared to vehicle control, both BE-146 and urelumab analog induced significant inhibition of tumor growth (P<0.001) (Figure 29).

[0311] Example 23. Combination treatment with anti-PD-1 antibody and CD137xCEA induces increased tumor regression CEA + To investigate the therapeutic effect of the combination of the CEA × CD137 multispecific antibody BE-146 and an anti-PD-1 antibody on tumor cells, CT26 / CEA cells (1 × 10 6 ) was injected subcutaneously into humanized CD137 mice on a BALB / c background. Mice developed tumors with a mean tumor volume of approximately 100 mm on day 4 post-injection. 3 Patients were randomized when they reached a median age of 18. BE-146 (0.6 mg / kg), anti-PD-1 antibody (0.3 mg / kg), or a combination of both were administered once a week starting on day 4. Compared to vehicle control or single agent treatment, the combination of BE-146 with anti-PD-1 significantly increased the anti-tumor effect (Figure 30).

[0312] Example 24. CD137xCEA does not induce hepatotoxicity in vivo BE-146 or urelumab analog antibody (30 mg / kg) was injected into humanized CD137 mice on a C57BL / 6 background, once a week, for three injections. Blood was collected on day 22 and analyzed by blood biochemistry tests. Compared with vehicle controls, high doses of urelumab analog induced significant increases in alanine transaminase (ALT) and aspartate aminotransferase (AST) concentrations, indicative of hepatotoxicity, but BE-146 did not. Furthermore, microscopic changes of increased inflammatory cells were observed in liver tissues from the urelumab analog treatment group, but no significant microscopic changes were observed in the BE-146 treatment group (Figure 31). Thus, CD137xTAA is a promising combination partner for cancer immunotherapy without hepatotoxicity, such as checkpoint inhibitors and T cell engagers.

[0313] Example 25. Other CD137xTAA combinations induce T cell activation in a TAA target-dependent manner Using the same concept on Claudin6+ and Trop2+ tumor cells, we tested other CD137xTAA combinations. Using PBMCs from healthy donors, we activated human T cells with Claudin6×CD137 (BE-268) or Trop2×CD137 (BE-907), and HEK293 / OS8 providing the readout. In the case of BE-268, we used different Claudin6 expression levels (1×10 4Cancer cell lines with BE-268 and BE-907 (cells / well) were co-cultured for 2 days in 96-well U-bottom plates (Corning™ Costar™ 9018). PA-1 cells, which have high Claudin6 expression, and Bewo cells, which have moderate Claudin6 expression, were purchased from ATCC. MKN45 cells, which are negative for Claudin6 expression, were purchased from JCRB cell bank. To assay for BE-907, Trop2-expressing cell line Mc38 / Trop2 was generated according to a previously described protocol (Zhang et al., Blood. 2005 106(5):1544-51). IFN-γ release from T cells was measured by ELISA. The results showed that both BE-268 and BE-907 could induce significant IFN-γ release (Figures 32 and 33). These results showed that, as expected, BE-268 and BE-907 induced potent T cell activation in a Claudin6- and Trop2-dependent manner.

[0314] Example 26. CD137xGPC3 induces T cell activation in a GPC3-dependent manner A further TAA tested was Glypican3 (GPC3). GPC3+ tumor cells were generated. Human peripheral blood mononuclear cells (PBMCs) were obtained from healthy donors and human T cells were activated with CD137xGPC3 (BE-830) and OS8 to provide the readout, as described above. To make the system very simple, the GPC3+ cell lines HepG2 OS8, Huh7 OS8-HiBit and Hep3B OS8-HiBit expressing OS8 were generated by retroviral transduction into HepG2 (ATCC, HB8065), Huh7 (JCRB, JCRB0403) and Hep3B (ATCC, HB8064) according to a previously described protocol (Zhang et al., Blood. 2005 106(5):1544-51). In addition, we generated SK-HEP-1 (ATCC, HTB-52) with OS8 (SK-HEP-1 OS8-HiBit), which is negative for GPC3 expression and was used as a negative control. Co-culture was performed in the presence of BE-830 at the indicated concentrations (0.0001-10 μg / ml) at an E:T ratio of 2:1 for 2 days, and IFN-γ and IL-2 were measured by commercial ELISA kits. As shown in Figure 34, in human HCC cell lines with GPC3 expression, BE-830 exhibited similar functional effects in inducing IFN-γ release in co-cultured PBMCs, regardless of the expression level of GPC3. BE-830 appears to stimulate relatively high IL-2 production in PBMCs co-cultured with GPC3-highly expressing HepG2 cells. These data indicated that the multispecific antibody BE-830 is functionally active, similar to the CEA construct BE-146.

[0315] Example 27. BE-830 reduces tumors in vivo To measure the in vivo efficacy of BE-830 against GPC3+ tumor cells, Hepa1-6T / hGPC3 cells (1 × 10 7) was injected subcutaneously into humanized CD137 mice on a C57BL / 6 background. BE-830 (0.5, 3 or 10 mpk) and vehicle control were given twice a week starting on the day of tumor injection (5 mice per group). Compared to vehicle, BE-830 at a dose of 0.5 mg / kg showed tumor inhibition that was statistically significantly different (P<0.01) from the vehicle control (Figure 35).

[0316] Example 28. Resolution of crystal structure To better understand how the anti-CD137 single domain antibody arm has high affinity for CD137 and can achieve robust agonism of the CD137 / CD137L interaction, we determined the crystal structure of VHH (BGA-5623) in complex with CD137.

[0317] Expression, purification and crystallization of CD137 and VHH (BGA-5623) Human CD137 ectodomain containing the four CRDs (1-4; amino acids 24-162) with C121S, N138D and N149Q mutations was expressed in HEK293G cells. cDNA encoding CD137 was cloned into a pMAX vector with an N-terminal secretion sequence and a C-terminal TEV cleavage site followed by an Fc tag. Culture supernatants containing secreted CD137-Fc fusion protein were mixed with Mab Select Sure™ resin (GE Healthcare Life Sciences) for 3 h at 4°C. Proteins were washed with a buffer containing 20 mM Tris-HCl pH 8.0, 150 mM NaCl, then eluted with 50 mM acetic acid (pH value adjusted to 3.5 with 5 M NaOH) and finally neutralized with 1 / 10 CV 1.0 M Tris-HCl pH 8.0. The eluted protein was mixed with TEV protease (10:1 molar ratio) and dialyzed against buffer (20 mM Tris-HCl, pH 8.0, 100 mM NaCl) overnight at 4° C. The mixture was loaded onto a Ni-NTA column (Qiagen) and Mab Select Sure resin to remove the TEV protease and Fc tag, and the flow-through was then further purified by size-exclusion chromatography in buffer (20 mM Tris pH 8.0, 100 mM NaCl) using a HiLoad 16 / 600 Superdex™ 75 pg column (GE Healthcare Life Sciences).

[0318] The DNA sequence encoding the VHH (BGA-5623) was cloned into the PET21a vector with an N-terminal HIS-MBP tag followed by a TEV protease site. Protein expression in Shuffle T7 was induced with 1 mM IPTG at 18 °C for 16 h at an OD600 of 0.6–1.0. Cells were harvested by centrifugation at 7,000 g for 10 min. The cell pellet was lysed in lysis buffer (50 mM Na 3 PO 4The proteins were resuspended in lysis buffer (pH 7.0, 300 mM NaCl) and lysed under sonication on ice. The lysate was then centrifuged at 48,000g for 30 min at 4°C. The supernatant was mixed with Talon resin and batch processed at 4°C for 3 h. The resin was washed with lysis buffer containing 5 mM imidazole and the proteins were further eluted in lysis buffer containing 100 mM imidazole. The eluate was mixed with TEV protease (10:1 molar ratio) and dialyzed overnight at 4°C against buffer (20 mM Tris-HCl, pH 8.0, 100 mM NaCl). The mixture was loaded onto a Talon column to remove the TEV protease and HIS-MBP tag, and the flow-through was then further purified by size-exclusion chromatography in buffer (20 mM Tris pH 8.0, 100 mM NaCl) using a HiLoad 16 / 600 Superdex™ 75pg column (GE Healthcare Life Sciences).

[0319] Purified CD137 was mixed with excess purified VHH(BGA-5623) (molar ratio 1:1.5) to generate CD137 / VHH(BGA-5623) complex. The complex was then further purified by gel filtration in buffer (20 mM Tris pH 8.0, 100 mM NaCl) using a HiLoad 16 / 600 Superdex™ 75pg column (GE Healthcare Life Sciences). CD137 / VHH(BGA-5623) complex (10 mg / ml) was diluted with 0.6 M Li 2 SO4, 0.01M NiCl 2 The apoVHH (BGA-5623) was crystallized in 0.1 M Tris pH 9.0. The crystals were cryoprotected by stepwise addition of 5% D-(+)-sucrose to a final concentration of 20% and flash frozen in liquid nitrogen. 4 ) 2 SO 4 The crystals were cryoprotected in 7% glycerol and flash frozen in liquid nitrogen. X-ray diffraction data were collected at beamline BL45XU of the Spring-8 Synchrotron Radiation Facility (Hyogo, Japan).

[0320] Data collection and structure solution X-ray diffraction data were collected under cryogenic cooling conditions at 100 Kelvin at beamline BL45XU equipped with the ZOO (Hirata, K., et al., Acta Crystallogr D Struct Biol, 2019.75(Pt2):138-150) automated data collection system at the Spring-8 Synchrotron Radiation Facility (Hyogo, Japan). Diffraction images were processed with the integrated data processing software KAMO (Yamashita, et al., Acta Crystallogr D Struct Biol, 2018.74(Pt5):441-449) using XDS (Kabsch W., Acta Crystallor D Biol Crystallogr, 2010.66(Pt2):125-32). The structures of human CD137 (PDB:6MGP) and VHH model (PDB:4U3X) were used as search models. An initial solution was found using the molecular replacement program PHASER (McCoy et al., Phaser crystallographic software. J Appl Crystallogr, 2007. 40 (Pt4): 658-674). The model was then iteratively built manually using the program COOT (Emsley et al., Acta Crystallogr D Biol Crystallogr, 2004. 60 (Pt12 Pt1): 2126-32) and refined using PHENIX (Adams et al., Acta Crystallogr D Biol Crystallogr, 2010. 66 (Pt2): 213-21). The final model was refined to acceptable R and R free values ​​and Ramachandran statistics (calculated by Molprobity). Data processing and refinement statistics are shown in Table 25.

[0321] [Table 25-1]

[0322] [Table 25-2]

[0323] Structure of VHH (BGA-5623) bound to human CD137 VHH (BGA-5623) complexed with CD137 was expressed as I4 1 It crystallized in space group BGA-5623, with one complex in the asymmetric unit, and diffracted to 2.58 Å. The structure of VHH(BGA-5623) bound to human CD137 shows that VHH(BGA-5623) partially sterically interfaces with CD137L binding (Figure 36). The buried surface area between VHH(BGA-5623) and CD137 is approximately 571 Å. 2VHH(BGA-5623) interactions are clustered around the CD137 CRD2 domain. These interactions are mainly mediated by VHH(BGA-5623) CDR2 and CDR3, which make more extensive contacts with CD137. VHH(BGA-5623) CDR1 does not make direct contacts with CD137, but CDR3 undergoes a dramatic conformational change from an unstructured loop to a β-sheet upon CD137 binding (Figure 37). VHH(BGA-5623) CDR2 Leu52, Tyr58 make contacts with CD137 residues Pro50, Asn51. VHH(BGA-5623) CDR3 residues Gly100A, Gly100B, Val100C, Thr100D, and Phe100E contact CD137 residues Phe36, Pro47, Pro49, Arg60, Cys62, and Ile64. In addition, FR2 Leu45 and Trp47 contact CD137 residues Pro47, Cys48, Pro49, and Pro50, which contribute significantly to CD137 binding. VHH(BGA-5623) interacts with CD137 using a combination of hydrogen bonds and hydrophobic interactions. For example, FR2 Trp47 forms strong hydrophobic contacts with CD137 residues Pro47, Cys48, Pro49, and Pro50. CDR3 residue Phe100E forms hydrophobic interactions with CD137 residues Phe36 and Pro47. FR2 residue Trp47 and CDR3 residue Gly100A form one hydrogen bond with CD137 residues Pro47 and Ile64, respectively. CDR3 residue Val100C forms two hydrogen bonds with CD137 residue Cys62 (Figure 38).

[0324] Based on the crystal structure of the VHH(BGA-5623) / CD137 complex, the residues of CD137 that VHH(BGA-5623) contacts (i.e., the epitope residues of CD137 that VHH binds) and the residues of VHH(BGA-5623) that CD137 contacts (i.e., the paratope residues of VHH that CD137 contacts) were determined. Table 26 below shows the residues of CD137 and VHH(BGA-5623) that they contact, assessed using a contact distance stringency of 3.7 Å, the point at which van der Waals (non-polar) interaction forces are at their highest.

[0325]

Table 26

Claims

1. An antibody or antigen-binding fragment thereof capable of binding to human CD137, (i) a heavy chain variable region comprising (a) an HCDR1 (heavy chain complementarity determining region 1) of SEQ ID NO: 14, (b) an HCDR2 of SEQ ID NO: 29, and (c) an HCDR3 of SEQ ID NO: 30; (ii) a heavy chain variable region comprising (a) an HCDR1 of SEQ ID NO: 14, (b) an HCDR2 of SEQ ID NO: 22, and (c) an HCDR3 of SEQ ID NO: 16; (iii) a heavy chain variable region comprising (a) an HCDR1 of SEQ ID NO: 14, (b) an HCDR2 of SEQ ID NO: 15, and (c) an HCDR3 of SEQ ID NO: 16; or (iv) a heavy chain variable region comprising (a) an HCDR1 of SEQ ID NO: 4, (b) an HCDR2 of SEQ ID NO: 5, and (c) an HCDR3 of SEQ ID NO:

6. An antibody or antigen-binding fragment thereof comprising:

2. An antibody or antigen-binding fragment thereof, (i) a heavy chain variable region (VH) comprising an amino acid sequence at least 90, 91, 92, 93, 94, 95, 96, 97, 98, or 99% identical to SEQ ID NO: 33; (ii) a heavy chain variable region (VH) comprising an amino acid sequence at least 90, 91, 92, 93, 94, 95, 96, 97, 98, or 99% identical to SEQ ID NO:24; (iii) a heavy chain variable region (VH) comprising an amino acid sequence at least 90, 91, 92, 93, 94, 95, 96, 97, 98, or 99% identical to SEQ ID NO: 19; (iv) a heavy chain variable region (VH) comprising an amino acid sequence at least 90, 91, 92, 93, 94, 95, 96, 97, 98, or 99% identical to SEQ ID NO:9; or (v) a heavy chain variable region (VH) comprising an amino acid sequence at least 90, 91, 92, 93, 94, 95, 96, 97, 98, or 99% identical to SEQ ID NO:

103. The antibody or antigen-binding fragment thereof of claim 1 .

3. The antibody or antigen-binding fragment thereof of claim 2, wherein 1, 2, 3, 4, 5, 6, 7, 8, 9, or 10 amino acids within SEQ ID NO: 33, 24, 19, 9, or 103 have been inserted, deleted, or substituted.

4. An antibody or antigen-binding fragment thereof comprising: (i) a heavy chain variable region (VH) comprising SEQ ID NO: 33; (ii) a heavy chain variable region (VH) comprising SEQ ID NO:24; (iii) a heavy chain variable region (VH) comprising SEQ ID NO: 19; (iv) a heavy chain variable region (VH) comprising SEQ ID NO:9; or (v) a heavy chain variable region (VH) comprising SEQ ID NO: 103 The antibody or antigen-binding fragment thereof of claim 1 .

5. The antibody or antigen-binding fragment thereof is a heavy chain (scFv), a heavy chain Fab fragment, a heavy chain Fab' fragment, or a heavy chain F(ab') 2 The antibody or antigen-binding fragment thereof according to any one of claims 1 to 4, which is a fragment.

6. (i) a heavy chain variable region comprising (a) an HCDR1 (heavy chain complementarity determining region 1) of SEQ ID NO: 14, (b) an HCDR2 of SEQ ID NO: 29, and (c) an HCDR3 of SEQ ID NO: 30; (ii) a heavy chain variable region comprising (a) an HCDR1 of SEQ ID NO: 14, (b) an HCDR2 of SEQ ID NO: 22, and (c) an HCDR3 of SEQ ID NO: 16; (iii) a heavy chain variable region comprising (a) an HCDR1 of SEQ ID NO: 14, (b) an HCDR2 of SEQ ID NO: 15, and (c) an HCDR3 of SEQ ID NO: 16; or (vi) a heavy chain variable region comprising (a) an HCDR1 of SEQ ID NO: 4, (b) an HCDR2 of SEQ ID NO: 5, and (c) an HCDR3 of SEQ ID NO:

6. at least a first antigen-binding domain capable of binding to human CD137, comprising A multispecific antibody comprising at least a second antigen-binding domain that specifically binds to a human tumor-associated antigen (TAA).

7. The first antigen-binding domain comprises: (i) a heavy chain variable region (VH) comprising SEQ ID NO: 33; (ii) a heavy chain variable region (VH) comprising SEQ ID NO:24; (iii) a heavy chain variable region (VH) comprising SEQ ID NO: 19; (iv) a heavy chain variable region (VH) comprising SEQ ID NO:9; or (v) a heavy chain variable region (VH) comprising SEQ ID NO: 103; The multispecific antibody of claim 6, comprising at least a second antigen-binding domain that specifically binds to a human tumor-associated antigen (TAA).

8. The multispecific antibody of claim 7, wherein the multispecific antibody is a bispecific antibody.

9. The multispecific antibody of claim 8, wherein the bispecific is in a 1+1 format, a 1+2 format, or a 2+2 format.

10. The multispecific antibody according to any one of claims 6 to 9, further comprising a linker, said linker comprising the amino acid sequence of any one of SEQ ID NOs: 239 to 280.

11. The multispecific antibody of claim 10, wherein the linker comprises the amino acid sequence of SEQ ID NO: 246 or SEQ ID NO:

251.

12. The antibody or antigen-binding fragment thereof of claim 1 or claim 6, wherein the antibody or antigen-binding fragment thereof has antibody-dependent cellular cytotoxicity (ADCC) or complement-dependent cytotoxicity (CDC), has reduced, no or hypofucosylated glycosylation, or contains increased bisecting GlcNac structures.

13. The antibody or antigen-binding fragment thereof of claim 1 or claim 6, wherein the Fc domain is IgG1.

14. The antibody or antigen-binding fragment thereof of claim 13, wherein the IgG1 has reduced ADCC.

15. A pharmaceutical composition comprising the antibody or antigen-binding fragment thereof described in claim 1 or claim 6 and a pharma- ceutically acceptable carrier.

16. A pharmaceutical composition for treating cancer comprising an antibody or antigen-binding fragment thereof described in claim 1 or claim 6.

17. 17. The pharmaceutical composition of claim 16, wherein the cancer is gastric cancer, colon cancer, pancreatic cancer, breast cancer, head and neck cancer, kidney cancer, liver cancer, lung cancer, small cell lung cancer, non-small cell lung cancer, ovarian cancer, skin cancer, mesothelioma, lymphoma, leukemia, myeloma, or sarcoma.

18. 10. An isolated nucleic acid encoding an antibody or antigen-binding fragment thereof according to claim 1 or claim 6.

19. A vector comprising the nucleic acid of claim 18.

20. A host cell comprising the nucleic acid of claim 18.

21. 21. A method for producing an antibody or antigen-binding fragment thereof, comprising culturing the host cell of claim 20 and recovering the antibody or antigen-binding fragment thereof from the culture.