Bifunctional molecule targeting PD-L1 and TGF-β

Abstract

Provided is an anti-PD-L1 antibody having excellent activity in blocking the interaction between PD-1 and PD-L1. Also provided is a multifunctional molecule comprising the anti-PD-L1 antibody or a fragment thereof fused to the extracellular structure of the human TGF-β receptor type 2.

Description

[0001] This application is a divisional application of the invention patent application with the title of "Bifunctional Molecule Targeting PD-L1 and TGF-β", application number 202180061312.1, application date July 27, 2021, and entered the Chinese national phase on January 16, 2023. Background Art

[0002] In recent years, the exciting progress of cancer immunotherapy has led to a paradigm shift in oncology. The most remarkable results are T cell-based therapies, including immune checkpoint inhibitors (ICIs), genetically engineered T cells, and bispecific antibodies (BsAbs). T cells are a major class for immune surveillance and tumor eradication, with high specificity and long-term memory. However, in the tumor microenvironment, T cells may become exhausted or tolerant to tumor cells. T cell exhaustion is usually associated with the overexpression of inhibitory receptors, including programmed death receptor-1 (PD-1), cytotoxic T lymphocyte antigen-4 (CTLA-4), lymphocyte activation gene-3 (LAG-3), T cell immunoglobulin domain and mucin domain-3 (TIM-3), IL-10 receptor, and killer immunoglobulin receptor.

[0003] Monoclonal antibody (mAb)-based therapies can counteract these checkpoint molecules, thus removing the brakes that inhibit tumor-infiltrating T cells, and thereby achieving significant clinical benefits in different malignancies. For example, blocking the PD-1 / PD-L1 interaction can enhance immune normalization and enhance the anti-cancer response. However, a significant drawback of PD-1 / PD-L1 blockers is the inconsistency in homogeneous study populations with similar tumor characteristics. In addition, PD-1 / PD-L1 blockade therapy may also cause certain inflammatory side effects in some patients. The limitations of PD-1 / PD-L1 blocker monotherapy and the lack of promising alternatives make it necessary to seek combination therapies that can activate anti-tumor immunity and improve treatment efficacy.

[0004] M7824 (bintrafusp alfa) is a bifunctional protein composed of a monoclonal antibody against programmed death ligand 1 (PD-L1) that is fused to the extracellular domain of the human transforming growth factor-β (TGF-β) receptor II, which acts as a "trap" for all three TGF-β subtypes. The PD-L1 moiety is based on avelumab, which has been approved for the treatment of Merkel cell carcinoma and urothelial carcinoma. However, current clinical data show that the use of M7824 is associated with undesired skin growth, and in a phase II trial of patients with HPV-positive malignancies, the overall response rate was only 35% to 40%. Therefore, there is a need to improve therapies.

[0005] Overview

[0006] In some embodiments, the present disclosure provides bifunctional molecules that target the PD-L1 protein and TGF-β. The disclosed PD-L1 targeting unit consists of an anti-PD-L1 antibody fused to the extracellular domain of the human transforming growth factor-β (TGF-β) receptor II, which serves as a trap for TGF-β. Experimental data indicate that these new bifunctional molecules are more effective than the leading candidate drug M7824 currently in clinical development.

[0007] Accordingly, in one embodiment of the present disclosure, there is provided a multifunctional molecule comprising an anti-PD-L1 (programmed death ligand 1) antibody or fragment thereof and the extracellular domain of human TGF-βRII (type 2 TGF-β receptor), wherein the anti-PD-L1 antibody or fragment thereof is specific for the human PD-L1 protein and comprises a heavy chain variable region (VH) containing VH CDR1, VH CDR2, and VH CDR3, and a light chain variable region (VL) containing VL CDR1, VL CDR2, and VL CDR3, wherein the VH CDR1, VH CDR2, VH CDR3, VL CDR1, VL CDR2, and VL CDR3 comprise the amino acid sequences of SEQ ID NO:7-12, or SEQ ID NO:13-18, respectively, or wherein VH CDR1 comprises SEQ ID NO:19, VH CDR2 comprises SEQ ID NO:20, 91, or 92, VH CDR3 comprises SEQ ID NO:21, VL CDR1 comprises SEQ ID NO:22, VL CDR2 comprises SEQ ID NO:23, and VL CDR3 comprises SEQ ID NO:24 or 93, and wherein the human TGF-βRII extracellular domain comprises the amino acid sequence of SEQ ID NO:72 and is fused to the anti-PD-L1 antibody or fragment thereof.

[0008] In one embodiment, there is provided an anti-PD-L1 (programmed death ligand 1) antibody or fragment thereof that is specific for human PD-L1 protein and includes a heavy chain variable region (VH) containing VH CDR1, VH CDR2, and VH CDR3, and a light chain variable region (VL) containing VL CDR1, VL CDR2, and VL CDR3, wherein the VH CDR1, VH CDR2, VH CDR3, VL CDR1, VL CDR2, and VL CDR3 include the amino acid sequences of SEQ ID NO:7-12, or SEQ ID NO:13-18, respectively, or wherein VH CDR1 includes SEQ ID NO:19, VH CDR2 includes SEQ ID NO:20, 91 or 92, VH CDR3 includes SEQ ID NO:21, VL CDR1 includes SEQ ID NO:22, VL CDR2 includes SEQ ID NO:23, and VL CDR3 includes SEQ ID NO:24 or 93.

[0009] There is also provided a multifunctional molecule that includes an antibody or antigen-binding fragment thereof fused to the N-terminus of the amino acid sequence of SEQ ID NO:72 via a peptide linker, wherein the peptide linker (a) has a length of at least 30 amino acid residues, or (b) has a length of at least 25 amino acid residues and contains an α-helical motif.

[0010] There is also provided the use and method of treating cancer with any molecule of the present disclosure. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] Figure 1 It is shown that 47C6A3, 67F3G7, and 89C10H8 are able to bind to human PD-L1 with high affinity.

[0012] Figure 2 It is shown that 47C6A3, 67F3G7, and 89C10H8 antibodies can effectively bind to PD-L1 expressed on mammalian cells.

[0013] Figure 3 It is shown that 47C6A3, 67F3G7, and 89C10H8 antibodies can bind to cynomolgus monkey PD-L1 with high affinity, but not to rat or mouse PD-L1.

[0014] Figure 4 It is shown that 47C6A3, 67F3G7, and 89C10H8 can effectively inhibit the binding of human PD-L1 to human PD1.

[0015] Figure 5 The binding kinetics of 47C6A3, 67F3G7, and 89C10H8 to recombinant PD-L1 are shown.

[0016] Figure 6A -C shows that the binding efficiency of all tested humanized antibodies to human PD-L1 is comparable to that of chimeric antibodies.

[0017] Figure 7 It shows that the tested humanized antibodies can bind efficiently to PD-L1 expressed on mammalian cells, comparable to chimeric antibodies.

[0018] -C shows that some humanized antibodies can effectively inhibit the binding of human PD-L1 to human PD1.

[0019] Figure 8A -C shows that some humanized antibodies can effectively inhibit the binding of human PD-L1 to human CD80.

[0020] Figure 9A The binding kinetics of LP008-06, LP008-06a, LP008-06a-DA, and LP008-06a-ES to recombinant human PD-L1 were shown.

[0021] Figure 10 The binding kinetics of LP008-02 to human PD-L1 and human TGF-β1 were shown.

[0022] Figure 11 It was shown that LP008-02 and LP008-06a-ES can block the interaction between PD1 and PD-L1 with higher affinity than M7824.

[0023] Figure 12 It was shown that M7824, LP008-02, and LP008-06a-ES can effectively block TGF-β canonical signaling.

[0024] Figure 13 It was shown that LP008-02 and LP008-06a-ES bind to human PD-L1 with high affinity.

[0025] Figure 14 It was shown that LP008-02 and LP008-06a-ES can bind to cynomolgus monkey PD-L1 with higher affinity, but cannot bind to rat PD-L1 or mouse PD-L1.

[0026] Figure 15 It was shown that the binding efficiency of LP008-02 and LP008-06a-ES to human TGF-β is comparable to that of M7824.

[0027] Figure 16It is shown that the binding efficiencies of LP008-02 and LP008-06a-ES to cynomolgus monkey TGF-β, mouse TGF-β, and rat TGF-β are comparable to those of M7824.

[0028] Figure 17 -B shows the drug effects of LP008-02 and LP008-06a-ES in an animal model.

[0029] Figure 18A It is shown that the binding efficiencies of all tested modified bifunctional molecules to human TGF-β are comparable to those of LP008-02-1.

[0030] Figure 19 It is shown that all tested modified bifunctional molecules can effectively block TGF-β canonical signaling.

[0031] Figure 20 It is shown that the binding efficiencies of all tested modified bifunctional molecules to human TGF-β are comparable to those of LP008-02-1.

[0032] Figure 21 It is shown that all tested modified bifunctional molecules can effectively block TGF-β canonical signaling.

[0033] Figure 22 It is shown that the antibodies MPDL3280A, 47C6A3, Hu67F3G7-22, and Hu89C10H8-7 can block the PD1 and PD-L1 interaction with high affinity. Detailed description

[0034] Definition

[0035] It should be noted that the term "a" or "an" entity refers to one or more of that entity; for example, "an antibody" is understood to represent one or more antibodies. Thus, the terms "a" (or "an"), "one or more", and "at least one" are used interchangeably herein.

[0036] As used herein, "antibody" or "antigen-binding polypeptide" refers to a polypeptide or polypeptide complex that specifically recognizes and binds to an antigen. An antibody can be a whole antibody and any of its antigen-binding fragments or single chains. Thus, the term "antibody" includes any protein- or peptide-containing molecule that contains at least a portion of an immunoglobulin molecule that has biological activity for binding to an antigen. Such examples include, but are not limited to, complementarity-determining regions (CDRs) of a heavy or light chain or ligand-binding portions thereof, a heavy chain or heavy chain variable region, a heavy or light chain constant region, a framework (FR) region, or any portion thereof, or at least a portion of a binding protein.

[0037] As used herein, the term "antibody fragment" or "antigen-binding fragment" refers to a part of an antibody, such as F(ab')2, F(ab)2, Fab', Fab, Fv, scFv, etc. Regardless of structure, an antibody fragment binds to the same antigen recognized by the intact antibody. The term "antibody fragment" includes aptamers, spiegelmers, and diabodies. The term "antibody fragment" also includes any synthetic or genetically engineered protein that acts like an antibody by binding to a specific antigen to form a complex.

[0038] The term "antibody" encompasses a wide variety of polypeptide classes that can be distinguished by biochemical methods. Those skilled in the art will appreciate that heavy chains are classified as γ, μ, α, δ, ε, some of which have subclasses (e.g., γ1-γ4). It is the nature of these chains that determines the "class" of the antibody to be IgG, IgM, IgA, IgD, or IgE, respectively. Immunoglobulin subclasses (isotypes), such as IgG1, IgG2, IgG3, IgG4, IgG5, etc., are well-characterized and are known to confer functional specificity. Given the present disclosure, modified versions of each of these classes and isotypes will be readily discernible to those skilled in the art and are thus within the scope of the present disclosure. All immunoglobulin classes are clearly within the scope of the present disclosure, and the following discussion will generally be directed to the IgG class of immunoglobulin molecules. With respect to IgG, a standard immunoglobulin molecule comprises two identical light-chain polypeptides with a molecular weight of approximately 23,000 daltons, and two identical heavy-chain polypeptides with a molecular weight of 53,000-70,000. These four chains are typically linked by disulfide bonds in a "Y" configuration, where the light chains surround the heavy chains, starting from the mouth of the "Y" and continuing through the variable regions.

[0039] "Specifically binds" or "is specific for" generally means that an antibody binds to an epitope through its antigen-binding domain, and such binding requires some complementarity between the antigen-binding domain and the epitope. By this definition, an antibody is said to "specifically bind" an epitope when it binds to that epitope more readily through its antigen-binding domain than to a random, unrelated epitope. As used herein, the term "specificity" is used to define the relative affinity of a particular antibody for binding to a particular epitope. For example, antibody "A" may be considered to have a higher specificity for a given epitope than antibody "B", or antibody "A" may be considered to bind epitope "C" with a higher specificity than it binds to a related epitope "D".

[0040] As used herein, the term "treatment" refers to therapeutic treatment and prophylactic or preventive measures, wherein the purpose is to prevent or slow down (mitigate) an undesired physiological change or disorder, such as cancer progression. Beneficial or desired clinical outcomes include, but are not limited to, alleviation of symptoms, reduction in the degree of the disease, stabilization of the disease state (i.e., not getting worse), delay or slowing of disease progression, improvement or alleviation of the disease state, and remission (whether partial or complete), whether detectable or not. "Treatment" also means an extension of survival as compared to the expected survival of a person not receiving treatment. Persons in need of treatment include those who already have the disease or disorder, as well as those who are predisposed to having the disease or disorder or in need of preventing the disease or disorder.

[0041] "Subject / individual" or "individual" or "animal" or "patient" or "mammal" refers to any subject in need of diagnosis, prognosis, or treatment, particularly mammalian subjects. Mammalian subjects include humans, domestic animals, farm animals, zoo animals, sport animals, or pet animals, such as dogs, cats, guinea pigs, rabbits, rats, mice, horses, cows, dairy cows, etc.

[0042] As used herein, phrases such as "a patient in need of treatment" or "a subject in need of treatment" include subjects, such as mammalian subjects, who would benefit from the administration of an antibody or composition of the present invention, e.g., for detection, diagnostic procedures, and / or treatment.

[0043] Multifunctional molecule

[0044] As demonstrated by the appended experimental examples, the inventors were able to identify a number of bifunctional fusion proteins comprising an anti-PD-L1 unit and a TGF-β targeting unit. For example, as shown in Example 14, in the MC38 mouse model, two tested bifunctional proteins, LP008-02 and LP008-06a-ES, both showed better efficacy than M7824. M7824 is a PD-L1 / TGF-β dual-targeting fusion protein that is currently in a Phase II clinical trial for patients with HPV-positive malignancies. The anti-PD-L1 unit of M7824 is based on avelumab, which is a leading PD-L1 antibody that has been approved for the treatment of Merkel cell carcinoma and urothelial carcinoma. Thus, the excellent performance of the newly disclosed bifunctional proteins is surprising compared to M7824.

[0045] In addition, as shown in Example 12, the bifunctional proteins of the present disclosure have better species specificity. Different from M7824, which also reacts with mouse and rat PD-L1, the new bifunctional proteins bind only to human and cynomolgus monkey PD-L1 in addition to having excellent PD-L1 binding activity.

[0046] Thus, in one embodiment, the present disclosure provides a multifunctional molecule having at least an anti-PD-L1 unit and a TGF-β targeting unit. The anti-PD-L1 unit can include the anti-PD-L1 antibody or fragment of the present disclosure. The TGF-β targeting unit is preferably the extracellular domain of the human transforming growth factor-β (TGF-β) receptor II (TGF-βRII or TGFBR2).

[0047] There are two subtypes of TGF-βRII. Subtype A (NP_001020018.1; SEQ ID NO:70) has a longer extracellular fragment than subtype B (NP_003233.4; SEQ ID NO:71), but they share the same core extracellular domain (ectodomain) (SEQ ID NO:72). Their sequences are provided in Table A below.

[0048] Table A. Sequences related to TGF-βRII (underlined and bold: core extracellular domain; underlined and italic: residues different between subtypes; only underlined: mutations)

[0049]

[0050]

[0051] In some embodiments, the TGF-βRII extracellular domain includes the core extracellular domain (SEQ ID NO:72) and some flanking residues. For example, Variant 1 (SEQ ID NO:61) tested in Examples 8-16 includes an additional 25 residues on the N-terminal side and nine residues on the C-terminal side. Another variant - Variant 2 (SEQ ID NO:73) includes only nine C-terminal flanking residues. Other variants such as Variants 4-7 (SEQ ID NO:75-78) include alternative linkers that replace part of the N-terminal sequence of SEQ ID NO:61.

[0052] In some embodiments, the TGF-βRII extracellular domain does not include the first 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24 or 25 amino acids of SEQ ID NO:61. In some embodiments, the TGF-βRII extracellular domain does not include the last 1, 2, 3, 4, 5, 6, 7, 8 or 9 amino acids of SEQ ID NO:61.

[0053] Another variant - Variant 3 is based on Variant 1, but includes at least an amino acid substitution at the X position within the N-terminal portion (SEQ ID NO:88). These X positions are potential glycosylation sites. Thus, the substitution is made with an amino acid other than K, S, and N. Examples of substitutions are R, A, G, Q, I, L, D, or E, but are not limited thereto.

[0054] In some embodiments, the anti-PD-L1 unit consists of an anti-PD-L1 antibody or fragment thereof as further described below. The antibody or fragment can be in any antibody form, such as a conventional full IgG form, Fab fragment, single-chain fragment, or single-domain antibody, but is not limited thereto. When the antibody or its fragment has a light chain and a separate heavy chain, the extracellular domain of TGF-βRII can be fused to the light chain or the heavy chain. When the antibody or its fragment has a light chain and a heavy chain on a single protein chain (e.g., scFv), the extracellular domain of TGF-βRII can be fused closer to the light chain or the heavy chain.

[0055] In some embodiments, the extracellular domain of TGF-βRII is fused to the N-terminus of the chain of the anti-PD-L1 unit. In some embodiments, the extracellular domain of TGF-βRII is fused to the C-terminus of the chain of the anti-PD-L1 unit. In a preferred embodiment, the extracellular domain of TGF-βRII is fused to the C-terminus of the heavy chain of the anti-PD-L1 unit, optionally via a peptide linker (e.g., SEQ ID NO:60, or one, two, or three GGGGS (SEQ ID NO:86) repeats).

[0056] In some embodiments, the anti-PD-L1 unit includes VH (heavy chain variable region) and VL (light chain variable region). The VH and VL regions include VH CDR1, VH CDR2, VH CDR3, VL CDR1, VL CDR2, and VL CDR3, such as those shown in Tables 1A - 1C.

[0057] In one embodiment, VH CDR1, VH CDR2, VH CDR3, VL CDR1, VL CDR2, and VL CDR3 comprise the sequences of SDYAWN (SEQ ID NO:7), YIIYSGSTSYNPSLKS (SEQ ID NO:8), STMIATNWFAY (SEQ ID NO:9), KASQDVSLAVA (SEQ ID NO:10), WASTRHT (SEQ ID NO:11), and QQHYITPWT (SEQ ID NO:12), respectively. Examples of such VH sequences are provided in SEQ ID NO:25 (mouse) and 26 - 28 (humanized). Examples of such VL sequences are provided in SEQ ID NO:29 (mouse) and 30 (humanized). Exemplary humanized antibodies include antibodies having a VH of SEQ ID NO:26, or 27, or 28 and a VL of SEQ ID NO:30.

[0058] In one embodiment, VH CDR1, VH CDR2, VH CDR3, VL CDR1, VL CDR2, and VL CDR3 comprise the sequences of DFWVS (SEQ ID NO:13), EIYPNSGVSRYNEKFKG (SEQ ID NO:14), YFGYTYWFGY (SEQ ID NO:15), RASKSVSTYMH (SEQ ID NO:16), SASHLES (SEQ ID NO:17), and QQSNELPVT (SEQ ID NO:18), respectively. Examples of such VH sequences are provided in SEQ ID NO:31 (mouse) and 32 - 37 (humanized). Examples of such VL sequences are provided in SEQ ID NO:38 (mouse) and 39 - 43 (humanized). Exemplary humanized antibodies include antibodies having a VH of SEQ ID NO:34 and a VL of SEQ ID NO:39, 40, or 43, an antibody having a VH of SEQ ID NO:35 and a VL of SEQ ID NO:39, or an antibody having a VH of SEQ ID NO:37 and a VL of SEQ ID NO:39. In one embodiment, the humanized antibody comprises a VH of SEQ ID NO:34 and a VL of SEQ ID NO:43.

[0059] In one embodiment, VH CDR1, VH CDR2, VH CDR3, VL CDR1, VL CDR2, and VL CDR3 contain the sequences of NYWMT (SEQ ID NO:19), SITNTGSSTFYPDSVKG (SEQ ID NO:20), DTTIAPFDY (SEQ ID NO:21), KASQNLNEYLN (SEQ ID NO:22), KTNTLQA (SEQ ID NO:23), and SQYNSGNT (SEQ ID NO:24), respectively. Alternatively, VH CDR2 may include SITNTGSSTFYPDAVKG (SEQ ID NO:91) or SITNTGSSTFYPESVKG (SEQ ID NO:92). Alternatively, VL CDR3 may be SQYQSGNT (SEQ ID NO:93).

[0060] In one embodiment, VH CDR1, VH CDR2, VH CDR3, VL CDR1, VL CDR2, and VL CDR3 contain the sequences of NYWMT (SEQ ID NO:19), SITNTGSSTFYPDSVKG (SEQ ID NO:20), DTTIAPFDY (SEQ ID NO:21), KASQNLNEYLN (SEQ ID NO:22), KTNTLQA (SEQ ID NO:23), and SQYNSGNT (SEQ ID NO:24), respectively. In one embodiment, VH CDR1, VH CDR2, VH CDR3, VL CDR1, VL CDR2, and VL CDR3 contain the sequences of NYWMT (SEQ ID NO:19), SITNTGSSTFYPDAVKG (SEQ ID NO:91), DTTIAPFDY (SEQ ID NO:21), KASQNLNEYLN (SEQ ID NO:22), KTNTLQA (SEQ ID NO:23), and SQYNSGNT (SEQ ID NO:24), respectively. In one embodiment, VH CDR1, VH CDR2, VH CDR3, VL CDR1, VL CDR2, and VL CDR3 contain the sequences of NYWMT (SEQ ID NO:19), SITNTGSSTFYPESVKG (SEQ ID NO:92), DTTIAPFDY (SEQ ID NO:21), KASQNLNEYLN (SEQ ID NO:22), KTNTLQA (SEQ ID NO:23), and SQYNSGNT (SEQ ID NO:24), respectively.

[0061] In one embodiment, VH CDR1, VH CDR2, VH CDR3, VL CDR1, VL CDR2, and VL CDR3 comprise the sequences of NYWMT (SEQ ID NO:19), SITNTGSSTFYPDSVKG (SEQ ID NO:20), DTTIAPFDY (SEQ ID NO:21), KASQNLNEYLN (SEQ ID NO:22), KTNTLQA (SEQ ID NO:23), and SQYQSGNT (SEQ ID NO:93), respectively. In one embodiment, VH CDR1, VH CDR2, VH CDR3, VL CDR1, VL CDR2, and VL CDR3 comprise the sequences of NYWMT (SEQ ID NO:19), SITNTGSSTFYPDAVKG (SEQ ID NO:91), DTTIAPFDY (SEQ ID NO:21), KASQNLNEYLN (SEQ ID NO:22), KTNTLQA (SEQ ID NO:23), and SQYQSGNT (SEQ ID NO:93), respectively. In one embodiment, VH CDR1, VH CDR2, VH CDR3, VL CDR1, VL CDR2, and VL CDR3 comprise the sequences of NYWMT (SEQ ID NO:19), SITNTGSSTFYPESVKG (SEQ ID NO:92), DTTIAPFDY (SEQ ID NO:21), KASQNLNEYLN (SEQ ID NO:22), KTNTLQA (SEQ ID NO:23), and SQYQSGNT (SEQ ID NO:93), respectively.

[0062] Examples of such VH sequences are provided in SEQ ID NO:44 (mouse) and 45 - 49 (humanized) and 57 - 58 (humanized). Examples of such VL sequences are provided in SEQ ID NO:50 (mouse) and 51 - 55 (humanized) and 56 (humanized).

[0063] Exemplary humanized antibodies include antibodies having a VH of SEQ ID NO:49 and a VL of SEQ ID NO:52 or 54, or antibodies having a VH of SEQ ID NO:48 and a VH of SEQ ID NO:53 or 54. In one embodiment, the humanized antibody includes a VH of SEQ ID NO:48 and a VL of SEQ ID NO:53. In one embodiment, the humanized antibody includes a VH of SEQ ID NO:48 and a VL of SEQ ID NO:56. In one embodiment, the humanized antibody includes a VH of SEQ ID NO:57 and a VL of SEQ ID NO:56. In one embodiment, the humanized antibody includes a VH of SEQ ID NO:58 and a VL of SEQ ID NO:56.

[0064] In some embodiments, the antibody or fragment thereof further includes a heavy chain constant region (e.g., CH1, CH2, and / or CH3) and / or a light chain constant region (e.g., CL). Exemplary heavy chain constant regions are provided in SEQ ID NO:59, and an exemplary light chain constant region is provided in SEQ ID NO:67 (residues 108 - 214).

[0065] TGF-βRII x antibody fusion

[0066] Tests conducted using different fusion protein designs (e.g., Table 15) indicate that only the core extracellular domain of TGF-βRII (SEQ ID NO:72) is necessary for activity. In addition, the extracellular domain of TGF-βRII should not be directly fused to the antibody. There should be sufficient distance provided by a peptide linker.

[0067] Regarding the extracellular domain, the peptide linker (which can be a completely artificial linker or include a partial extracellular fragment of the N-terminus of the extracellular domain SEQ ID NO:89) should have a minimum length. If the distance is too short, the stability or activity of the fusion protein is reduced. In some embodiments, the minimum length is 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, or 40 amino acid residues. In some embodiments, the linker is not longer than 35, 40, 45, 50, 55, 60, 65, 70, 75, 80, 85, 90, 95, 100, 110, 120, 130, 140, 150, 170, or 200 amino acid residues.

[0068] In some embodiments, the inclusion of flexible linkers, such as one or more G4S (SEQ ID NO:86) units, is useful for the stability and / or activity of multifunctional molecules. In some embodiments, the flexible linker comprises at least 40%, 50%, 60%, 70% or 80% glycine. In some embodiments, the flexible linker comprises one or more serines. In some embodiments, the flexible linker comprises 1, 2, 3, 4, 5 or 6 G4S (SEQ ID NO:86) repeats.

[0069] In some embodiments, it is shown (e.g., in Example 17) that the native N-terminal fragment (IPPHVQKSVNNDMIVTDNNGAVKFP; SEQ ID NO:89) can be replaced with an alternative peptide to increase stability without sacrificing or even enhancing activity. In some embodiments, the alternative peptide differs from SEQ ID NO:89 but has at least 30%, 40%, 50%, 60%, 70%, 80% or 90% sequence identity with SEQ ID NO:99.

[0070] Exemplary alternative peptides are IPPHVQXXVNNDMIVTDNXGAVKFP (SEQ ID NO:88), where X is any amino acid other than K, S or N. In some embodiments, substitutions can be made to remove the rigid dipeptide PP, remove potential cleavage sites QK, N and / or K, include multiple glycine residues to increase flexibility, and / or reduce hydrophobic residues. One such example is TAGHTQTSTGGGAITTGTSGAGHGP (SEQ ID NO:87) or a variant having at least 50%, 55%, 60%, 65%, 70%, 75%, 80%, 85%, 90% or 95% sequence identity with SEQ ID NO:87. In some embodiments, the variant comprises at least 4 Gs, no PP dipeptide, and no more than 3 hydrophobic amino acid residues selected from I, L, M, F, V, W, Y and P. In some embodiments, the variant comprises at least 5 Gs and no more than 1 hydrophobic amino acid residue selected from I, L, M, F, V, W, Y and P.

[0071] In some embodiments, the peptide linker between the antibody or its fragment and the extracellular domain of TGF-βRII (SEQ ID NO:72) comprises a flexible linker. In some embodiments, the peptide linker comprises an alternative peptide of SEQ ID NO:89. In some embodiments, the peptide linker comprises both a flexible linker and an alternative peptide. In some embodiments, the flexible linker is at the N-terminus of the alternative peptide. In some embodiments, the flexible linker is at the C-terminus of the alternative peptide.

[0072] In some embodiments, the multifunctional molecule does not include at least the entire sequence of EEYNTSNPD (SEQ ID NO:90). The multifunctional molecule can have the entire SEQ ID NO:90 removed from the extracellular domain of TGF-βRII. In some embodiments, the multifunctional molecule does not include more than 1, 2, 3, 4, 5, 6, 7, or 8 amino acid residues of EEYNTSNPD (SEQ ID NO:90).

[0073] Antibodies or antigen-binding fragments thereof of the multifunctional molecule can target any antigen. Non-limiting examples are PD-1, PD-1, PD-L1, CTLA-4, LAG-3, CD28, CD122, 4-1BB, TIM3, OX-40, OX40L, CD40, CD40L, LIGHT, ICOS, ICOS-L, GITR, GITR-L, TIGIT, CD27, VISTA, B7H3, B7H4, BTLA, CD4, CD2, CD8, CD47, and CD73. They can also be any of the antibodies or fragments disclosed herein.

[0074] The extracellular domain of TGF-βRII can be fused to any part of the antibody or fragment. In some embodiments, the extracellular domain is fused to the C-terminus of the heavy or light chain of the antibody or fragment. In some embodiments, the extracellular domain is fused to the C-terminus of the Fc fragment of the antibody or fragment.

[0075] Anti-PD-L1 antibodies and fragments

[0076] ]Also provided are anti-PD-L1 antibodies and fragments, which can be used as the anti-PD-L1 unit in multifunctional molecules, bispecific or multispecific antibodies, or alone as a monospecific antibody.

[0077] Exemplary murine anti-PD-L1 antibodies and their humanized and improved versions have been prepared and tested in the accompanying experimental examples. All murine antibodies (47C6A3, 67F3G7, and 89C10H8) and their corresponding humanized versions exhibited excellent binding affinity, cross-reactivity, and effectiveness in inhibiting PD-1 / PD-L1 binding.

[0078] Importantly, compared to MPDL3280A (Atezolizumab), the humanized 67F3G7 and 89C10H8 exhibit higher activity in blocking the interaction between PD-1 and PD-L1 than MPDL3280A (see, for example, Example 18). Additionally, interestingly, all of the tested antibodies of the present disclosure show lower hydrophobicity and lower viscosity than MPDL3280A. Higher hydrophobicity is known to reduce the solubility of proteins. Similarly, high viscosity is an obstacle to the development of high-concentration protein formulations. Thus, these data indicate that the antibodies of the present invention are more suitable for preparing high-concentration antibody formulations.

[0079] Furthermore, the antigen-binding fragments of the antibodies of the present disclosure are included as a unit in a bifunctional fusion protein that also includes a TGF-β targeting unit. In the MC38 mouse model, the resulting bifunctional fusion protein exhibits better efficacy than M7824. M7824 is a PD-L1 / TGF-β bifunctional fusion protein that is currently in a Phase II clinical trial in patients with HPV-positive malignancies. The anti-PD-L1 unit of M7824 is based on avelumab, which is a leading PD-L1 antibody that has been approved for the treatment of Merkel cell carcinoma and urothelial carcinoma. Thus, these data demonstrate the unique advantages of the antibodies of the present disclosure in preparing bifunctional or multifunctional molecules.

[0080] In some embodiments, the anti-PD-L1 antibody or fragment includes a VH (heavy chain variable region) and a VL (light chain variable region). The VH and VL regions include VH CDR1, VH CDR2, VH CDR3, VL CDR1, VL CDR2, and VL CDR3, as shown, for example, in Tables 1A-1C.

[0081] In one embodiment, VH CDR1, VH CDR2, VH CDR3, VL CDR1, VL CDR2, and VL CDR3 respectively comprise the sequences of SDYAWN (SEQ ID NO:7), YIIYSGSTSYNPSLKS (SEQ ID NO:8), STMIATNWFAY (SEQ ID NO:9), KASQDVSLAVA (SEQ ID NO:10), WASTRHT (SEQ ID NO:11), and QQHYITPWT (SEQ ID NO:12). Examples of such VH sequences are provided in SEQ ID NO:25 (mouse) and 26-28 (humanized). Examples of such VL sequences are provided in SEQ ID NO:29 (mouse) and 30 (humanized). Exemplary humanized antibodies include antibodies having a VH of SEQ ID NO:26, or 27, or 28 and a VL of SEQ ID NO:30.

[0082] In one embodiment, VH CDR1, VH CDR2, VH CDR3, VL CDR1, VL CDR2, and VL CDR3 contain the sequences of DFWVS (SEQ ID NO:13), EIYPNSGVSRYNEKFKG (SEQ ID NO:14), YFGYTYWFGY (SEQ ID NO:15), RASKSVSTYMH (SEQ ID NO:16), SASHLES (SEQ ID NO:17), and QQSNELPVT (SEQ ID NO:18), respectively. Examples of such VH sequences are provided in SEQ ID NO:31 (mouse) and 32 - 37 (humanized). Examples of such VL sequences are provided in SEQ ID NO:38 (mouse) and 39 - 43 (humanized). Exemplary humanized antibodies include antibodies having VH of SEQ ID NO:34 and VL of SEQ ID NO:39, 40, or 43, VH of SEQ ID NO:35 and VL of SEQ ID NO:39, or VH of SEQ ID NO:37 and VL of SEQ ID NO:39. In one embodiment, the humanized antibody includes VH of SEQ ID NO:34 and VL of SEQ ID NO:43.

[0083] In one embodiment, VH CDR1, VH CDR2, VH CDR3, VL CDR1, VL CDR2, and VL CDR3 contain the sequences of NYWMT (SEQ ID NO:19), SITNTGSSTFYPDSVKG (SEQ ID NO:20), DTTIAPFDY (SEQ ID NO:21), KASQNLNEYLN (SEQ ID NO:22), KTNTLQA (SEQ ID NO:23), and SQYNSGNT (SEQ ID NO:24), respectively. Alternatively, VH CDR2 may include SITNTGSSTFYPDAVKG (SEQ ID NO:91) or SITNTGSSTFYPESVKG (SEQ ID NO:92). Alternatively, VL CDR3 may be SQYQSGNT (SEQ ID NO:93).

[0084] In one embodiment, VH CDR1, VH CDR2, VH CDR3, VL CDR1, VL CDR2, and VL CDR3 contain the sequences of NYWMT (SEQ ID NO:19), SITNTGSSTFYPDSVKG (SEQ ID NO:20), DTTIAPFDY (SEQ ID NO:21), KASQNLNEYLN (SEQ ID NO:22), KTNTLQA (SEQ ID NO:23), and SQYNSGNT (SEQ ID NO:24), respectively. In one embodiment, VH CDR1, VH CDR2, VH CDR3, VL CDR1, VL CDR2, and VL CDR3 contain the sequences of NYWMT (SEQ ID NO:19), SITNTGSSTFYPDAVKG (SEQ ID NO:91), DTTIAPFDY (SEQ ID NO:21), KASQNLNEYLN (SEQ ID NO:22), KTNTLQA (SEQ ID NO:23), and SQYNSGNT (SEQ ID NO:24), respectively. In one embodiment, VH CDR1, VH CDR2, VH CDR3, VL CDR1, VL CDR2, and VL CDR3 contain the sequences of NYWMT (SEQ ID NO:19), SITNTGSSTFYPESVKG (SEQ ID NO:92), DTTIAPFDY (SEQ ID NO:21), KASQNLNEYLN (SEQ ID NO:22), KTNTLQA (SEQ ID NO:23), and SQYNSGNT (SEQ ID NO:24), respectively.

[0085] In one embodiment, VH CDR1, VH CDR2, VH CDR3, VL CDR1, VL CDR2, and VL CDR3 comprise the sequences of NYWMT (SEQ ID NO:19), SITNTGSSTFYPDSVKG (SEQ ID NO:20), DTTIAPFDY (SEQ ID NO:21), KASQNLNEYLN (SEQ ID NO:22), KTNTLQA (SEQ ID NO:23), and SQYQSGNT (SEQ ID NO:93), respectively. In one embodiment, VH CDR1, VH CDR2, VH CDR3, VL CDR1, VL CDR2, and VL CDR3 comprise the sequences of NYWMT (SEQ ID NO:19), SITNTGSSTFYPDAVKG (SEQ ID NO:91), DTTIAPFDY (SEQ ID NO:21), KASQNLNEYLN (SEQ ID NO:22), KTNTLQA (SEQ ID NO:23), and SQYQSGNT (SEQ ID NO:93), respectively. In one embodiment, VH CDR1, VH CDR2, VH CDR3, VL CDR1, VL CDR2, and VL CDR3 comprise the sequences of NYWMT (SEQ ID NO:19), SITNTGSSTFYPESVKG (SEQ ID NO:92), DTTIAPFDY (SEQ ID NO:21), KASQNLNEYLN (SEQ ID NO:22), KTNTLQA (SEQ ID NO:23), and SQYQSGNT (SEQ ID NO:93), respectively.

[0086] Examples of such VH sequences are provided in SEQ ID NO:44 (mouse) and 45 - 49 (humanized) and 57 - 58 (humanized). Examples of such VL sequences are provided in SEQ ID NO:50 (mouse) and 51 - 55 (humanized) and 56 (humanized).

[0087] Exemplary humanized antibodies include antibodies having a VH of SEQ ID NO:49 and a VL of SEQ ID NO:52 or 54, or antibodies having a VH of SEQ ID NO:48 and a VH of SEQ ID NO:53 or 54. In one embodiment, the humanized antibody includes a VH of SEQ ID NO:48 and a VL of SEQ ID NO:53. In one embodiment, the humanized antibody includes a VH of SEQ ID NO:48 and a VL of SEQ ID NO:56. In one embodiment, the humanized antibody includes a VH of SEQ ID NO:57 and a VL of SEQ ID NO:56. In one embodiment, the humanized antibody includes a VH of SEQ ID NO:58 and a VL of SEQ ID NO:56.

[0088] In some embodiments, the antibody or fragment thereof further includes a heavy chain constant region (e.g., CH1, CH2, and / or CH3) and / or a light chain constant region (e.g., CL). Exemplary heavy chain constant regions are provided in SEQ ID NO:59, and exemplary light chain constant regions are provided in SEQ ID NO:67 (residues 108 - 214).

[0089] Small changes (e.g., addition, deletion, or substitution of one amino acid) are expected to be designed in these CDR sequences, which may retain the activity of the antibody or even improve them. Such modified CDR sequences are referred to as CDR variants. Those of ordinary skill in the art will also understand that the antibodies disclosed herein can be modified such that their amino acid sequences are different from the naturally occurring binding polypeptides from which they are derived. For example, a polypeptide or amino acid sequence derived from a specified protein may be similar to the starting sequence, e.g., having a certain percentage identity, e.g., it may be 60%, 70%, 75%, 80%, 85%, 90%, 95%, 98%, or 99% identical to the starting sequence. In some embodiments, the modified antibody or fragment retains the specified CDR sequence.

[0090] In certain embodiments, the antibody contains an amino acid sequence or one or more moieties that are not normally associated with an antibody. Exemplary modifications are described in more detail below. For example, the antibodies of the present disclosure may contain a flexible linker sequence or may be modified to add a functional moiety (e.g., PEG, drug, toxin, or tag).

[0091] Polynucleotides encoding proteins and methods for preparing proteins

[0092] The present disclosure also provides isolated polynucleotides or nucleic acid molecules encoding the multifunctional proteins, antibodies, variants or derivatives thereof of the present disclosure. The polynucleotides of the present disclosure can encode the entire heavy and light chain variable regions of an antigen-binding polypeptide, its variant or derivative on the same polynucleotide molecule or on an isolated polynucleotide. In addition, the polynucleotides of the present disclosure can encode a part of the heavy and light chain variable regions of an antigen-binding polypeptide, its variant or derivative on the same polynucleotide molecule or on an isolated polynucleotide.

[0093] Methods for preparing antibodies are well known in the art and are described herein. In certain embodiments, both the variable and constant regions of the antigen-binding polypeptides of the present disclosure are fully human. Fully human antibodies can be prepared using techniques described in the art and as described herein. For example, by administering an antigen to a transgenic animal that has been modified to produce such antibodies in response to antigen challenge but whose endogenous loci have been disabled, fully human antibodies against a specific antigen can be prepared. Exemplary techniques that can be used to make such antibodies are described in U.S. Patents 6,150,584, 6,458,592, 6,420,140, the entire contents of which are incorporated herein by reference.

[0094] Cancer treatment

[0095] As described herein, the antibodies, variants or derivatives of the present disclosure can be used in certain therapeutic and diagnostic methods.

[0096] The present invention further relates to multifunctional molecule- or antibody-based therapies, which involve administering the multifunctional molecules and antibodies of the present invention to a patient (such as an animal, mammal and human) to treat one or more diseases or disorders described herein. The therapeutic compounds of the present disclosure include, but are not limited to, the antibodies of the present disclosure (including their variants and derivatives as described herein) and nucleic acids or polynucleotides encoding the antibodies of the present disclosure (including their variants and derivatives as described herein).

[0097] The antibodies of the present invention can also be used to treat or inhibit cancer. PD-L1 can be overexpressed in tumor cells. Tumor-derived PD-L1 can bind to PD-1 on immune cells, thereby restricting anti-tumor T cell immunity. Results using small molecule inhibitors or monoclonal antibodies against PD-L1 in murine tumor models indicate that targeting PD-L1 therapy is an important alternative and realistic approach for effectively controlling tumor growth. As shown in the experimental examples, anti-PD-L1 antibodies activate the adaptive immune response mechanism, which can improve the survival rate of cancer patients.

[0098] Thus, in some embodiments, methods of treating cancer in a patient in need thereof are provided. In one embodiment, the method involves administering to the patient an effective amount of a multifunctional molecule or antibody of the present disclosure. In some embodiments, at least one cancer cell (e.g., stromal cell) in the patient expresses, overexpresses, or is induced to express PD-L1. For example, induction of PD-L1 expression can be accomplished by administering a tumor vaccine or radiotherapy.

[0099] Tumors expressing the PD-L1 protein include those of bladder cancer, non-small cell lung cancer, renal cancer, breast cancer, urethral cancer, colorectal cancer, head and neck cancer, squamous cell carcinoma, Merkel cell carcinoma, gastrointestinal cancer, gastric cancer, esophageal cancer, ovarian cancer, renal cancer, and small cell lung cancer. Accordingly, the antibodies of the present disclosure can be used to treat any one or more of such cancers.

[0100] The present invention also provides cell therapies, such as chimeric antigen receptor (CAR) T cell therapy. Suitable cells can be used, which are contacted with the anti-PD-L1 antibody of the present disclosure (or alternatively engineered to express the anti-PD-L1 antibody of the present disclosure). After such contact or engineering, the cells can be introduced into a cancer patient in need of treatment. The cancer patient may have any type of cancer disclosed herein. The cells (e.g., T cells) can be, for example, tumor-infiltrating T lymphocytes, CD4+ T cells, CD8+ T cells, or a combination thereof, but are not limited thereto.

[0101] In some embodiments, the cells are isolated from the cancer patient themselves. In some examples, the cells are provided by a donor or a cell bank. When the cells are isolated from the cancer patient, unwanted immune responses can be minimized.

[0102] Other diseases or conditions related to increased cell viability that can be treated, prevented, diagnosed, and / or predicted with the antibodies, variants, or derivatives of the present disclosure include, but are not limited to, the progression and / or metastasis of malignancies and related diseases, such as leukemia (including acute leukemia (e.g., acute lymphocytic leukemia, acute myeloid leukemia (including myeloblastic, promyelocytic, myelomonocytic, monocytic, and erythroleukemia)) and chronic leukemia (e.g., chronic myeloid (granulocytic) leukemia and chronic lymphocytic leukemia)), polycythemia vera, lymphoma (e.g., Hodgkin's disease and non-Hodgkin's disease), multiple myeloma, Waldenstrom's macroglobulinemia, heavy chain disease, and solid tumors, including but not limited to sarcomas and carcinomas, such as fibrosarcoma, myxosarcoma, liposarcoma, chondrosarcoma, osteosarcoma, chordoma, angiosarcoma, endotheliosarcoma, lymphangiosarcoma, lymphangioendotheliosarcoma, synovioma, mesothelioma, Ewing's tumor, leiomyosarcoma, rhabdomyosarcoma, colon cancer, pancreatic cancer, breast cancer, thyroid cancer, endometrial cancer, melanoma, prostate cancer, ovarian cancer, prostate cancer, squamous cell carcinoma, basal cell carcinoma, adenocarcinoma, sweat gland carcinoma, sebaceous gland carcinoma, papillary carcinoma, papillary adenocarcinoma, cystadenocarcinoma, medullary carcinoma, bronchogenic carcinoma, renal cell carcinoma, hepatocellular carcinoma, cholangiocarcinoma, choriocarcinoma, seminoma, embryonal carcinoma, Wilm's tumor, cervical cancer, testicular tumor, lung cancer, small cell lung cancer, bladder cancer, epithelial carcinoma, glioma, astrocytoma, medulloblastoma, craniopharyngioma, ependymoma, pinealoma, hemangioblastoma, acoustic neuroma, oligodendroglioma, hemangioma, melanoma, neuroblastoma, and retinoblastoma.

[0103] Composition

[0104] The present invention also provides pharmaceutical compositions. Such compositions comprise an effective amount of an antibody and an acceptable carrier. In some embodiments, the composition further includes a second anti-cancer agent (e.g., an immune checkpoint inhibitor).

[0105] In a specific embodiment, the term "pharmaceutically acceptable" means approved by a federal or state government regulatory agency or listed in the U.S. Pharmacopeia or other generally recognized pharmacopeias for use in animals, particularly humans. In addition, a "pharmaceutically acceptable carrier" is generally a non-toxic solid, semi-solid, or liquid filler, diluent, encapsulating material, or any type of formulation adjuvant.

[0106] The term "carrier" refers to a diluent, adjuvant, excipient, or vehicle with which a therapeutic agent is administered. Such pharmaceutical carriers can be sterile liquids such as water and oils, including those of petroleum, animal, vegetable, or synthetic origin, such as peanut oil, soybean oil, mineral oil, sesame oil, etc. Water is a preferred carrier when the pharmaceutical composition is administered intravenously. Saline solutions and aqueous glucose and glycerol solutions can also be used as liquid carriers, particularly for injectable solutions. Suitable pharmaceutical excipients include starch, glucose, lactose, sucrose, gelatin, malt, rice, flour, chalk, silica, sodium stearate, glycerol monostearate, talc, sodium chloride, dried skim milk, glycerol, propylene, ethylene glycol, water, ethanol, etc. If desired, the composition may also contain minor amounts of wetting or emulsifying agents, or pH buffering agents such as acetates, citrates, or phosphates. Antibacterial agents such as benzyl alcohol or methylparaben; antioxidants such as ascorbic acid or sodium bisulfite; chelating agents such as ethylenediaminetetraacetic acid; and agents for regulating tonicity such as sodium chloride or glucose may also be contemplated. These compositions can take the form of solutions, suspensions, emulsions, tablets, pills, capsules, powders, sustained-release formulations, etc. The compositions can be formulated into suppositories using conventional binders and carriers such as triglycerides. Oral formulations may include standard carriers such as pharmaceutical grade mannitol, lactose, starch, magnesium stearate, saccharin sodium, cellulose, magnesium carbonate, etc. Examples of suitable pharmaceutical carriers are described in Remington's Pharmaceutical Sciences by E.W. Martin, which is incorporated herein by reference. Such compositions will contain a therapeutically effective amount of the antigen-binding polypeptide, preferably in purified form, and a suitable amount of carrier so as to provide the form for proper administration to a patient. The formulations should suit the mode of administration. The parenteral formulations can be encapsulated in ampoules, disposable syringes, or multi-dose vials made of glass or plastic.

[0107] In one embodiment, the composition is formulated, according to conventional procedures, into a pharmaceutical composition suitable for intravenous administration to humans. Generally, compositions for intravenous administration are solutions in sterile isotonic aqueous buffer. If necessary, the composition may also include solubilizing agents and local anesthetics such as lidocaine to relieve the pain at the injection site. Usually, the ingredients are supplied individually or mixed together in unit dose form, for example, as a dry lyophilized powder or an anhydrous concentrate, in a container (such as an ampoule or sachet) indicating the amount of the active agent. If the composition is administered by infusion, an infusion bottle containing sterile pharmaceutical grade water or saline can be used for dispensing. In the case of administering the composition by injection, ampoules of sterile water for injection or saline can be provided so that the ingredients can be mixed before administration. Examples

[0108] Example 1: Generation of a murine monoclonal antibody against human PD-L1

[0109] This example describes the generation of anti-human PD-L1 murine monoclonal antibodies using hybridoma technology.

[0110] Antigens: Human PDL1-Fc protein and human PD-L1 highly expressed on the CHOK1 cell line (PDL1-CHOK1 cell line).

[0111] Immunization: To generate murine monoclonal antibodies targeting human PD-L1, Balb / c mice and Wistar rats were first immunized with the PD-L1-Fc protein. Subsequently, the immunized mice and rats were boosted with the PD-L1-Fc protein and CHO-K1 / PD-L1 stable cells, respectively. To select mice or rats that produce antibodies binding to the PD-L1 protein, the antibody titers of the sera from the immunized mice or rats were evaluated by ELISA. Briefly, microtiter plates were coated with 0.5 μg / ml of human PD-L1 protein in ELISA coating buffer, 100 μl / well overnight at 4°C, and then blocked with 150 μl / well of 1% BSA. Dilutions of sera from immunized mice were added to each well and incubated at 37°C for 1 - 2 hours. The plates were washed with PBS / Tween and then incubated with anti-mouse IgG antibody conjugated to horseradish peroxidase (HRP) or anti-rat IgG antibody conjugated to HRP at 37°C for 1 hour. After washing, the plates were developed with TMB substrate and analyzed with a spectrophotometer at OD 450 nm. After 3 rounds of immunization, the immune responses to the rhPD-L1 protein were also detected by serum ELISA and to the CHO-K1 / PDL-1 stable cell line by FACS, with the CHO-K1 parental cell line as a negative control. After 3 rounds of immunization, mice showing sufficient titers of anti-PDL1-IgG were boosted with 25 μg of human PDL1-Fc protein. The resulting mice were used for fusion. The anti-PD-L1 IgG of the hybridoma supernatants was detected by ELISA.

[0112] Cell fusion: Fusion was performed by electrofusion. The fused cells were placed in 50 96-well plates for each fusion.

[0113] Screening: The supernatants were screened by ELISA against recombinant human (rh) PD-L1-Fc protein and counter-screening antigens. Then, preliminary screening was performed on the CHO-K1 / PD-L1 stable cell line and rhPD-1-Fc protein using receptor-blocking FACS, and confirmation screening was performed on the positive supernatants.

[0114] Subcloning and screening: The positive primary clones from each fusion were subcloned by limiting dilution to ensure that the subclones were derived from a single parental cell. The subclones were screened in the same manner as the primary clones, and the culture supernatants of the positive clones were subjected to additional confirmation screening by affinity ranking.

[0115] Hybridoma clones 47C6A3, 67F3G7, and 89C10H8 were selected for further analysis. The amino acid sequences of the variable regions of 47C6A3, 67F3G7, and 89C10H8 are listed in Table 1 below.

[0116] Table 1. Sequences of the variable regions of 47C6A3, 67F3G7, and 89C10H8

[0117]

[0118] Table 1A. CDR sequences of 47C6A3

[0119] Figure 23 47C6A3 Sequence SEQ ID NO: CDRH1 7 SDYAWN CDRH2 8 YIIYSGSTSYNPSLKS CDRH3 9 STMIATNWFAY CDRL1 10 KASQDVSLAVA CDRL2 11 WASTRHT CDRL3 12

[0120] Table 1B. CDR sequences of 67F3G7

[0121] QQHYITPWT 67F3G7 Sequence SEQ ID NO: CDRH1 13 DFWVS CDRH2 14 EIYPNSGVSRYNEKFKG CDRH3 15 YFGYTYWFGY CDRL1 16 RASKSVSTYMH CDRL2 17 SASHLES CDRL3 18

[0122] Table 1C. CDR sequences of 89C10H8

[0123]

[0124] Example 2: Binding activity to the PD-L1 antigen

[0125] ELISA assay

[0126] To evaluate the binding activity of hybridoma clones 47C6A3, 67F3G7, and 89C10H8, chimeric mAbs from these clones were tested by ELISA.

[0127] Briefly, microtiter plates were coated with 0.5 μg / ml of human PD-L1-Fc protein in PBS, 100 μl / well overnight at 4°C, and then blocked with 150 μl / well of 1% BSA. Triplicate dilutions of the 47C6A3, 67F3G7, and 89C10H8 antibodies starting from 10 μg / ml were added to each well and incubated for 1 hour at 37°C. The plates were washed with PBS / Tween and then incubated with a mouse anti-human IgG Fab antibody conjugated to horseradish peroxidase (HRP) for 30 minutes at 37°C. After washing, the plates were developed with TMB substrate and analyzed with a spectrophotometer at OD 450 nm. As QQSNELPVT shown, 47C6A3, 67F3G7, and 89C10H8 bound to human PD-L1 with high affinity (EC 50 of 47C6A3 = 10.24 ng / ml, EC 50 of 67F3G7 = 10.76 ng / ml, EC 50 of 89C10H8 = 8.112 ng / ml).

[0128] Cell-based binding: Flow cytometry (FACS) was used to evaluate the binding activity of the 47C6A3, 67F3G7, and 89C10H8 chimeric mAbs to CHOK1 cells overexpressing human PD-L1.

[0129] Briefly, PDL1-CHOK1 cells were first incubated with 3-fold serial dilutions of the 47C6A3, 67F3G7, and 89C10H8 chimeric mAbs starting at 100 nM for 40 minutes at 4 °C. After washing with PBS, Alexa 647 AffiniPure Goat Anti-Human IgG (H+L) was added to each well and incubated for 30 minutes at 4 °C. The samples were washed twice with FACS buffer. The mean fluorescence intensity (MFI) of Alexa 647 was evaluated by FACSCanto. As Figure 1 shown, 47C6A3, 67F3G7, and 89C10H8 bound to PDL1-CHOK1 cells with high affinity (EC 50 of 47C6A3 = 0.1476 nM, EC 50 of 67F3G7 = 0.1035 nM, and EC 50 of 89C10H8 = 0.1696 nM).

[0130] Cross-species activity

[0131] ELISA tests were performed to evaluate the binding of the chimeric antibodies to human, mouse, rat, and cynomolgus monkey PD-L1, respectively.

[0132] Briefly, microtiter plates were coated with 0.5 μg / ml of human, mouse, rat, and cynomolgus monkey PD-L1 proteins in PBS, 100 μl / well overnight at 4 °C, and then blocked with 1% BSA at 150 μl / well. Chimeric antibodies diluted three-fold starting from 10 μg / ml were added to each well and incubated for 1 hour at 37 °C. The plates were washed with PBS / Tween and then incubated with mouse anti-human IgG Fab antibody conjugated to horseradish peroxidase (HRP) for 30 minutes at 37 °C. After washing, the plates were developed with TMB substrate and analyzed with a spectrophotometer at OD 450 nm. The 47C6A3, 67F3G7, and 89C10H8 antibodies bound to human and cynomolgus monkey PD-L1 but not to rat and mouse PD-L1 ( Figure 2 and Table 2).

[0133] Table 2. Cross-species activity of 47C6A3, 67F3G7, and 89C10H8

[0134] Figure 3 Human Cynomolgus monkey Rat Mouse EC50 of 47C6A3 10.24 ng / ml 6.336 ng / ml Does not bind Does not bind EC50 of 67F3G7 10.76 ng / ml 6.797 ng / ml Does not bind Does not bind EC50 of 89C10H8 8.112 ng / ml 6.774 ng / ml Does not bind

[0135] Example 3. Blocking the binding of PD-L1 to PD-1 with antibodies

[0136] To evaluate the blocking effects of the 47C6A3, 67F3G7, and 89C10H8 chimeric mAbs on the binding of recombinant human PD-L1 to its receptor PD-1, an ELISA-based receptor blocking assay was employed.

[0137] Briefly, microtiter plates were coated with 0.5 μg / ml of human PD-L1-Fc protein in PBS, 100 μl / well overnight at 4°C, and then blocked with 150 μl / well of 1% BSA. Fifty microliters of biotinylated human PD-1-Fc protein and 50 μl of the 47C6A3, 67F3G7, and 89C10H8 antibodies serially diluted threefold starting from 10 μg / ml were added to each well and incubated at 37°C for 1 hour. The plates were washed with PBS / Tween and then incubated with streptavidin HRP at 37°C for 10 minutes. After washing, the plates were developed with TMB substrate and analyzed with a spectrophotometer at OD 450 nm. As Does not bind shown, 47C6A3, 67F3G7, and 89C10H8 effectively inhibited the binding of human PD-L1 to human PD1 at IC 50 values of 91.18 ng / ml, 139.8 ng / ml, and 129.8 ng / ml, respectively.

[0138] Example 4: Binding affinity of mAbs

[0139] The binding of the 47C6A3, 67F3G7, and 89C10H8 antibodies to recombinant PD-L1 protein (human PD-L1-his tag) was tested by Biacore using the capture method. The 47C6A3, 67F3G7, and 89C10H8 mAbs were captured using a Protein A chip. Serial dilutions of human PD-L1-his tag protein were injected at a flow rate of 30 μl / min for 2 min over the captured antibodies. Antigen dissociation was allowed for 480 - 1500 s. All experiments were performed on a Biacore T200. Data analysis was performed using the Biacore T200 evaluation software. The results are shown in Figure 4 and Table 3 below.

[0140] Table 3. Affinity measured by Biacore

[0141]

[0142] Example 5. Humanization of murine antibodies

[0143] Humanized mAbs were created using the variable region genes of 47C6A3, 67F3G7, and 89C10H8. In the first step of the method, the amino acid sequences of the VH and VL or VK of 47C6A3, 67F3G7, and 89C10H8 were compared with the available databases of human Ig gene sequences to find the human germline Ig gene sequences that were overall the best matches. For the light chain of 47C6A3, human Vk1-4 was the most suitable germline, and for the heavy chain, human VH1-2 was selected as the framework. For the light chain of 67F3G7, the closest human match was the Vk1-39 / JK4 gene, and for the heavy chain, the closest human match was the VH1-2 / JH4-FW4 gene. For the light chain of 89C10H8, the closest human match was the Vk1-17 / JK2 gene, and for the heavy chain, the closest human match was the VH3-21 / JH3 gene.

[0144] For the VL of 47C6A3, human Vk1-4 was the most suitable germline, and for the VH of 47C6A3, human VH1-2 was selected as the framework. Then, the humanized variable domain sequences of 47C6A3 were designed, in which CDRL1, L2, and L3 were grafted onto the framework sequence of the Vk1-4 gene, and CDRH1, H2, and H3 were grafted onto the framework sequence of the VH1-2 gene. Then, 3D models were generated to determine if there were any framework positions where replacing murine amino acids with human amino acids might affect binding and / or CDR conformation. In the case of the heavy chain, R, M, and I in the framework were associated with back mutations.

[0145] Then, the humanized variable domain sequences of 67F3G7 were designed, in which CDRL1, L2, and L3 were grafted onto the framework sequence of the Vk1-39 / JK4 gene, and CDRH1, H2, and H3 were grafted onto the framework structure sequence of the VH1-2 / JH4-FW4 gene. Then, 3D models were generated to determine if there were any framework positions where replacing murine amino acids with human amino acids might affect binding and / or CDR conformation. In the case of the heavy chain, V, K, T, and I in the framework were associated with back mutations. In the case of the light chain, T, V, L, and Q in the framework were associated with back mutations.

[0146] Then, the humanized variable domain sequences of 89C10H8 were designed, in which CDRL1, L2, and L3 were grafted onto the framework sequence of the Vk1-17 / JK2 gene, and CDRH1, H2, and H3 were grafted onto the framework sequence of the VH3-21 / JH3 gene. Then, 3D models were generated to determine if there were any framework positions where replacing murine amino acids with human amino acids might affect binding and / or CDR conformation. In the case of the heavy chain, A, T, I, and S in the framework were associated with back mutations. In the case of the light chain, Y, I, E, and F in the framework were associated with back mutations.

[0147] The amino acid and nucleotide sequences of some humanized antibodies are listed in Table 4 below.

[0148] Table 4. Humanized Antibody Sequences (underlined indicates CDR; bold / italic indicates back mutations)

[0149]

[0150]

[0151]

[0152]

[0153]

[0154] These genes were cloned into the pcDNA3.4 vector and transfected into 293F cells. Antibodies were prepared according to the following table.

[0155] Humanized VH and VL genes were generated by synthesis and then cloned separately into vectors containing the human γ1 and human κ constant domains. Pairing of human VH and human VL yielded 41 humanized antibodies (see Table 5).

[0156] Table 5. Humanized Antibodies and Their VH and VL Regions A. 47C6A3

[0157]

[0158] B. 67F3G7

[0159]

[0160] C. 89C10H8

[0161]

[0162] Example 6: Antigen-Binding Properties of Humanized Antibodies

[0163] Binding to Recombinant Human PD-L1

[0164] To evaluate the antigen-binding activity, the humanized antibodies were tested by ELISA. Briefly, microtiter plates were coated with 0.5 μg / ml of human PD-L1-Fc protein in PBS, 100 μl / well overnight at 4 °C, and then blocked with 200 μl / well of 1% BSA. Triplicate dilutions of the humanized antibodies starting from 10 μg / ml were added to each well and incubated for 1 h at 37 °C. The plates were washed with PBS / Tween and then incubated with mouse anti-human IgG Fab antibody conjugated to horseradish peroxidase (HRP) for 1 h at 37 °C. After washing, the plates were developed with TMB substrate and analyzed with a spectrophotometer at OD 450 nm. As shown in Figure 6, all the humanized antibodies showed comparable binding efficiency to human PD-L1 with the chimeric antibody.

[0165] To explore the binding kinetics of the humanized antibodies, in this example, affinity ranking was performed using Biacore. As shown in Table 6, Hu67F3G7-2, Hu67F3G7-3, Hu67F3G7-5, Hu67F3G7-7, Hu67F3G7-22, Hu89C10H8-4, Hu89C10H8-7, Hu89C10H8-11, and Hu89C10H8-12 showed high affinity, which was comparable to the chimeric antibody.

[0166] Table 6. Affinity ranking of humanized antibodies

[0167]

[0168]

[0169]

[0170] Binding to human PD-L1 overexpressed on mammalian cells

[0171] To evaluate the antigen-binding properties, the binding of the humanized antibodies to PD-L1 overexpressed on mammalian cells was analyzed by FACS. Briefly, PDL1-CHOK1 cells were first incubated with triplicate serial dilutions of the humanized antibodies starting from 15 μg / ml at 4 °C for 40 min. After washing with PBS, Alexa 647 AffiniPure goat anti-human IgG (H+L) antibody was added to each well and incubated at 4 °C for 30 min. The MFI of Alexa 647 was evaluated by FACSCanto. As Figure 5 shown, all the humanized antibodies could bind efficiently to PD-L1 expressed on mammalian cells.

[0172] Full kinetic affinity of humanized antibodies measured by Biacore

[0173] The binding of humanized antibodies to recombinant PD-L1 protein (human PD-L1-his tag) was tested by Biacore using the capture method. Hu47C6A3-1, Hu47C6A3-2, Hu47C6A3-3, Hu67F3G7-2, Hu67F3G7-3, Hu67F3G7-5, Hu67F3G7-7, Hu67F3G7-22, Hu89C10H8-4, Hu89C10H8-7, Hu89C10H8-11 and Hu89C10H8-12 mAbs were captured using a Protein A chip. Serial dilutions of the human PD-L1-his tag protein were injected over the captured antibodies at a flow rate of 30 μl / min for 2 minutes. Antigen dissociation was allowed for 1500 s. All experiments were performed on a Biacore T200. Data analysis was performed using the Biacore T200 evaluation software, and the results are shown in Table 7 below.

[0174] Table 7 Affinity measured by Biacore

[0175]

[0176]

[0177]

[0178] Example 7: Blocking of the binding of PDL1 to PD1 by humanized antibodies

[0179] Receptor blocking assay using recombinant human PD-L1

[0180] Human PD-L1 has two receptors, PD-1 and CD80. To investigate the blocking characteristics of humanized PD-L1 antibodies against these two proteins, a protein-based receptor blocking assay was employed here.

[0181] Briefly, microtiter plates were coated with 0.5 μg / ml of human PD-L1-Fc protein in PBS, 100 μl / well overnight at 4 °C, and then blocked with 150 μl / well of 1% BSA for 2 hours at 37 °C. 50 μl of biotinylated human PD-1-Fc or CD80-Fc protein and 50 μl of a 3-fold dilution of PD-L1 antibody starting at 10 μg / ml were added to each well and incubated for 1 hour at 37 °C. The plates were washed with PBS / Tween and then incubated with streptavidin HRP for 10 minutes at 37 °C. After washing, the plates were developed with TMB substrate and analyzed with a spectrophotometer at OD 450 nm. As shown in Figure 8, Hu47C6A3-1, Hu47C6A3-2, Hu47C6A3-3, Hu67F3G7-2, Hu67F3G7-3, Hu67F3G7-5, Hu67F3G7-7, Hu67F3G7-22, Hu89C10H8-4, Hu89C10H8-7, Hu89C10H8-11, and Hu89C10H8-12 effectively inhibited the binding of human PD-L1 to human PD1. In addition, Hu47C6A3-1, Hu47C6A3-2, Hu47C6A3-3, Hu67F3G7-2, Hu67F3G7-3, Hu67F3G7-5, Hu67F3G7-7, Hu67F3G7-22, Hu89C10H8-4, Hu89C10H8-7, Hu89C10H8-11, and Hu89C10H8-12 effectively inhibited the binding of human PD-L1 to human CD80 in a dose-dependent manner (Figure 9).

[0182] Example 8. Bifunctional Proteins Targeting the PD-L1 and TGF-β Pathways

[0183] In this example, bifunctional recombinant anti-PD-L1 antibody and TGF-βRII fusion proteins were prepared and tested.

[0184] The light chain of the molecule is the light chain of an anti-PDL1 mAb. The heavy chain is a fusion of the heavy chain of an anti-PDL1 mAb with the N-terminus of the soluble extracellular domain of TGF-βRII via a flexible (Gly4Ser)4Gly linker. At the fusion junction, the C-terminal lysine residue of the antibody heavy chain was mutated to alanine to reduce potential proteolytic cleavage.

[0185] In some instances, potential modification sites in the CDRs were mutated to similar amino acids. The sequence of the anti-PD-L1 portion is shown in Table 8 below.

[0186] Table 8. Sequences of the Variable Regions of the Antibody Portion in the Bifunctional Molecule

[0187]

[0188] Table 9. VH / VL of bifunctional molecules

[0189] Figure 7 Bifunctional molecule VH VL LP008-02 02 VH 02VL LP008-06 06VH 06VL LP008-06a 06a VH 06a VL LP008-06a-DA 06a-DA VH 06a VL LP008-06a-ES 06a-ES VH

[0190] In addition to VH, the heavy chain of the bifunctional molecule also includes a constant region (K at the C-terminus mutated to A), a (Gly4Ser)4Gly linker, and the N-terminus of the soluble extracellular domain of TGF-βRII. Its sequence is shown in Table 10.

[0191] Table 10. Other sequences of the heavy chain, and the entire heavy / light chain

[0192]

[0193]

[0194] Example 9: Binding affinity of bifunctional molecules

[0195] The binding of LP008-06, LP008-06a, LP008-06a-DA, and LP008-06a-ES bifunctional molecules to recombinant PD-L1 protein (human PD-L1-his tag) was tested by Biacore using the capture method.

[0196] The bifunctional molecules were captured using a Protein A chip. Serial dilutions of human PD-L1-his tag protein were injected over the captured antibody at a flow rate of 30 μl / min for 2 minutes. Antigen dissociation was allowed for 1500 s. All experiments were performed on a Biacore T200. Data analysis was performed using the Biacore T200 evaluation software. The results are shown in 06a VL and Table 11 below.

[0197] Table 11. Affinity test by Biacore

[0198]

[0199] The binding of LP008-02 to recombinant PD-L1 protein and human TGF-β1 was tested by Biacore using the capture method.

[0200] LP008-02 was captured using a Protein A chip. Serial dilutions of human PD-L1-his tag protein and human TGF-β1 were injected over the captured antibody at a flow rate of 30 μl / min for 2 minutes. PD-L1 dissociation was allowed for 680 s, and TGF-β1 dissociation was allowed for 1000 s. All experiments were performed on a Biacore T200. Data analysis was performed using the Biacore T200 evaluation software. The results are shown in Figure 10 and Table 12 below.

[0201] Table 12. Affinity Test of Biacore

[0202]

[0203] Example 10: Functional Assay of PD-1 / PD-L1 Blockade

[0204] In this example, the activity of the bifunctional molecule in blocking the PD1 / PD-L1 interaction was determined by a bioluminescence cell-based assay.

[0205] In this assay, when PD1 effector cells are co-cultured with PD-L1 target cells, the PD-1 / PD-L1 interaction inhibits TCR signaling and NFAT RE-mediated luminescence. Adding anti-PD-1 or anti-PD-L1 antibodies that block the PD-1 / PD-L1 interaction will release the inhibitory signal and lead to TCR activation and NFAT RE-mediated luminescence.

[0206] As Figure 11 shown, LP008-02 and LP008-06a-ES blocked the PD1 and PD-L1 interaction with much higher activity than M7824 (M7824 EC 50 = 0.8504 nM, LP008-02 EC 50 = 0.3630 nM, LP008-06a-ES EC 50 = 0.4553 nM).

[0207] Example 11: Functional Assay of TGF-β

[0208] This example used a luciferase assay to evaluate the effect of LP008-02 and LP008-06a-ES on classical TGF-β signaling.

[0209] In the presence of recombinant human TGF-β, serial dilutions of M7824 (bifunctional anti-PD-L1 / TGFβ trap fusion protein, see, for example, Knudson et al., Oncoimmunology. 2018;7(5):e1426519), LP008-02 or LP008-06a-ES were incubated with 293 cells transfected with the SBE luciferase reporter for approximately 20 hours.

[0210] As Figure 12As shown, in the TGF-β SBE luciferase reporter assay system constructed in 293 cells, M7824, LP008-02 and LP008-06a-ES blocked TGF-β canonical signaling (IC50 = 0.06687 nM, IC50 = 0.07352 nM, IC50 = 0.07167 nM).

[0211] Example 12: Binding Activity to Human PD-L1

[0212] ELISA of Recombinant Human PD-L1

[0213] To evaluate the binding activity of M7824, LP008-02 and LP008-06a-ES, ELISA tests were performed on the bifunctional molecules.

[0214] Briefly, microtiter plates were coated with 0.5 μg / ml human PD-L1-His protein in PBS, 100 μl / well overnight at 4°C, and then blocked with 150 μl / well of 1% BSA. Serial dilutions of M7824, LP008-02 and LP008-06a-ES starting from 1 μg / ml were added to each well and incubated at 37°C for 1 hour. The plates were washed with PBS / Tween and then incubated with goat anti-human IgG antibody conjugated to horseradish peroxidase (HRP) at 37°C for 30 minutes. After washing, the plates were developed with TMB substrate and analyzed with a spectrophotometer at OD 450 nm.

[0215] As Figure 13 shown, LP008-02 and LP008-06a-ES bound to human PD-L1 with significantly higher activity than M7824 (EC 50 = 11.82 ng / ml and EC 50 = 14.36 ng / ml vs. EC 50 = 23.68 ng / ml).

[0216] Cross-Species Activity

[0217] To evaluate the binding of the bispecific antibody to mouse PD-L1, rat PD-L1, cynomolgus monkey PD-L1, ELISA tests were performed on the antibody.

[0218] Briefly, microtiter plates were coated with mouse, rat, and cynomolgus monkey PD-L1 proteins at 0.5 μg / ml in PBS, 100 μl / well overnight at 4°C, and then blocked with 1% BSA at 150 μl / well. Triplicate dilutions of bispecific antibodies starting from 1 μg / ml were added to each well and incubated at 37°C for 1 hour. The plates were washed with PBS / Tween and then incubated with goat anti-human IgG antibody conjugated to horseradish peroxidase (HRP) at 37°C for 30 minutes. After washing, the plates were developed with TMB substrate and analyzed with a spectrophotometer at OD 450 nm.

[0219] LP008-02 and LP008-06a-ES were able to bind cynomolgus monkey PD-L1 with higher affinity than M7824, but only M7824 was able to bind rat and mouse PD-L1 ( Figure 14 and Table 13).

[0220] Table 13. Cross-species activities of M7824, CZ010-02, and CZ010-06a-ES

[0221] Figure 15 EC50 Cynomolgus monkey Rat M7824 Mouse 25.60 ng / ml 14.25 ng / ml 10.60 ng / ml LP008-02 7.890 ng / ml Does not bind Does not bind LP008-06a-ES 10.92 ng / ml Does not bind

[0222] Example 13: Binding activity to human TGF-β

[0223] ELISA using recombinant human TGF-β

[0224] To evaluate the binding activities of M7824, LP008-02, and LP008-06a-ES to human TGF-β, ELISA tests were performed on these bifunctional molecules.

[0225] Briefly, microtiter plates were coated with human TGF-β protein at 1 μg / ml in PBS, 100 μl / well overnight at 4°C, and then blocked with 1% BSA at 150 μl / well. Triplicate dilutions of M7824, LP008-02, and LP008-06a-ES bifunctional molecules starting from 10 μg / ml were added to each well and incubated at 37°C for 1 hour. The plates were washed with PBS / Tween and then incubated with goat anti-human IgG antibody conjugated to horseradish peroxidase (HRP) at 37°C for 30 minutes. After washing, the plates were developed with TMB substrate and analyzed with a spectrophotometer at OD 450 nm.

[0226] As Does not bind shown, M7824, LP008-02, and LP008-06a-ES all bound human TGF-β with high activity (EC 50 = 43.43 ng / ml, EC 50 = 28.58 ng / ml, EC50 = 39.38 ng / ml).

[0227] Cross-species activity

[0228] To evaluate the binding of the bispecific antibody to TGF-β of mice, rats, and cynomolgus monkeys, an ELISA test was performed on the bifunctional molecule.

[0229] Briefly, microtiter plates were coated with 1 μg / ml of mouse, rat, and cynomolgus monkey TGF-β protein in PBS, 100 μl / well overnight at 4 °C, and then blocked with 150 μl / well of 1% BSA. Serial dilutions of the bispecific antibody starting from 10 μg / ml were added to each well and incubated for 1 hour at 37 °C. The plates were washed with PBS / Tween and then incubated with goat anti-human IgG antibody conjugated to horseradish peroxidase (HRP) for 30 minutes at 37 °C. After washing, the plates were developed with TMB substrate and analyzed with a spectrophotometer at OD 450 nm.

[0230] All tested bifunctional molecules bound to cynomolgus monkey, rat, and mouse TGF-β with high activity ( Figure 16 and Table 14).

[0231] Table 14. Cross-species activity of M7824, LP008-02, and LP008-06a-ES

[0232] Figure 17 EC50 Cynomolgus monkey M7824 Rat Mouse 39.66 ng / ml 123.9 ng / ml 46.17 ng / ml LP008-02 34.41 ng / ml 77.84 ng / ml 34.70 ng / ml LP008-06a-ES 58.35 ng / ml

[0233] Example 14: Efficacy in MC38 tumor mouse model

[0234] This example uses a tumor mouse model to test the in vivo efficacy of the bifunctional molecule.

[0235] MC38 cells expressing human PD-L1 resuspended in PBS were subcutaneously inoculated into the right skin of B-hPD-L1 humanized mice at a concentration of 5 x 10 5 cells in a volume of 0.2 mL. When the average tumor volume reached approximately 55 mm 3 , 24 mice with appropriate individual tumor volumes were selected as a group, and the animals were randomly divided into 4 experimental groups of 6 each according to tumor volume. After injection of anti-mCD20 mAb, total human IgG, M7824, LP008-02, and LP008-06a-ES were administered 3 times a week by intraperitoneal injection. The dose was calculated at 10 μg / g based on the body weight of the experimental animals. The body weight and tumor size of the mice were tested twice a week.

[0236] The results are shown in Figure 18. In these animal models, the bifunctional molecules LP008-02 and LP008-06a-ES showed better efficacy than M7824 in tumor growth inhibition. In addition, animal deaths were observed in both the IgG and M7824 groups, but not in the LP008-02 and LP008-06a-ES groups, indicating that the new bifunctional molecules have better safety.

[0237] Example 15. Modification of Bifunctional Molecules

[0238] In this example, the in vitro potency of certain modified bifunctional molecules (Table 15) in TGF-β functional assays was tested. Some of them included linker sequences with TAGHTQTSTGGGAITTGTSGAGHGP (SEQ ID NO:87), HYP, and / or G4S (SEQ ID NO:86) repeats. These molecules were designated as LP008-02-1 to LP008-02-7, respectively.

[0239] Table 15. Modified Sequence Designs of Linker and TGF-βRII

[0240]

[0241] ELISA of Recombinant Human TGF-β1

[0242] To evaluate the binding activity of the modified LP008-02 bifunctional molecules, these bifunctional molecules were tested by ELISA.

[0243] Briefly, microtiter plates were coated with 1 μg / ml human TGF-β1 protein (Acro, TG1-H4212) in PBS, 100 μl / well overnight at 4°C, and then blocked with 150 μl / well of 1% BSA. Serial dilutions of the modified LP008-02 bifunctional molecules starting from 30 nM were added to each well and incubated at 37°C for 1 hour. The plates were washed with PBS / Tween and then incubated with goat anti-human IgG (H+L) antibody conjugated to horseradish peroxidase (HRP) at 37°C for 30 minutes. After washing, the plates were developed with TMB substrate and analyzed with a spectrophotometer at OD 450 nm.

[0244] As 120.7 ng / ml shown, all modified LP008-02 bifunctional molecules bound to human TGF-β1 with high activity, comparable to LP008-02-1.

[0245] TGF-β Functional Assay

[0246] In the presence of recombinant human TGF-β1, serial dilutions of the modified LP008-02 bifunctional molecule were incubated with 293 cells transfected with the SBE luciferase reporter for approximately 22 hours.

[0247] As 58.26 ng / ml shown, in the TGF-β SBE luciferase reporter assay system constructed in 293 cells, LP008-02-2, LP008-02-3, and LP008-02-4 effectively blocked TGF-β canonical signaling (IC50 = 0.1435 nM, IC50 = 0.1639 nM, IC50 = 0.1882 nM), as compared to LP008-02-1.

[0248] Example 16. Comparison of Bifunctional Molecules

[0249] Molecules 1-7 in Table 15 contain different sequences at the N- and C-termini of the extracellular domain (SEQ ID NO:72). Their stability and activity were tested to evaluate the impact of these sequences.

[0250] Molecule 1 (LP008-02-1) includes the entire extracellular portion of the protein (SEQ ID NO:61), which contains 25 amino acids from the N-terminus of the extracellular domain (IPPHVQKSVNNDMIVTDNNGAVKFP, SEQ ID NO:89, or amino acids 24-48 of subtype B, SEQ ID NO:71) and a C-terminal fragment (EEYNTSNPD, SEQ ID NO:90). In addition, several G4S (SEQ ID NO:86) repeats were added to the linker in this molecule.

[0251] Compared to Molecule 1, Molecule 2 (LP008-02-2) replaced the N-terminal portion of the extracellular domain (amino acids 24-48 of subtype B, SEQ ID NO:89) with an artificial linker TAGHTQTSTGGGAITTGTSGAGHGP (SEQ ID NO:87). This linker was modeled based on SEQ ID NO:89. The changes included: (i) removal of the rigid dipeptide PP, (ii) removal of potential cleavage sites QK, N, and K, (iii) inclusion of multiple glycine residues to increase flexibility, and (iv) partial removal of hydrophobic residues (e.g., only one I was retained). These changes are shown in Table 16 below. Molecule 2 also contains a single G4S unit at the N-terminus.

[0252] Table 16. Artificial Linker

[0253] Figure 19 Figure 20 Name Sequence SEQ ID NO: 89 Original IPPHVQKSVNNDMIVTDNNGAVKFP 87

[0254] Molecule 3 (LP008-02-3) contains a longer G4S linker than Molecule 2. Above Molecule 3, Molecule 4 (LP008-02-4) has a deletion of the C-terminal fragment EEYNTSNPD (SEQ ID NO:90). Molecule 5 (LP008-02-5) replaces the artificial linker SEQ ID NO:87 with the short linker HYP. Molecules 6 (LP008-02-6) and 7 (LP008.02-7) include G4S linkers of different lengths on the N-terminal side of the HYP linker.

[0255] Example 17. Binding Activity and Stability of Bifunctional Molecules

[0256] In this example, SEC-HPLC and CE-SDS were used to evaluate the stability of some modified bifunctional molecules, including LP008-02-1 and four further modified molecules LP008-02-2, LP008-02-3, LP008-02-6, and LP008-02-7.

[0257] Five sequences were expressed in CHO-K1 cells by polyethyleneimine (PEI)-mediated transient transfection, and the supernatants were harvested after 10 days. The bifunctional molecules were purified from the culture supernatants by Protein A and then purified by Superdex 200pg, with a purity level greater than 99% as detected by SEC-HPLC (Table 17).

[0258] Table 17. SEC-HPLC and CE-SDS Results of Test Articles on Day 0

[0259]

[0260]

[0261] To evaluate the binding activity of the modified LP008-02 bifunctional molecules, these bifunctional molecules were tested by ELISA.

[0262] Briefly, microtiter plates were coated with 1 μg / ml of human TGF-β1 protein (Acro, TG1-H4212) in PBS, 100 μl / well overnight at 4°C, and then blocked with 1% BSA at 150 μl / well. Four-fold serial dilutions of the modified LP008-02 bifunctional molecules starting from 30 nM were added to each well and incubated for 1 hour. The plates were washed with PBS / Tween and then incubated with goat anti-human IgG Fc antibody conjugated to horseradish peroxidase (HRP) for 30 minutes. After washing, the plates were incubated with TMB substrate for color development and analyzed by spectrophotometer at OD 450 nm.

[0263] As ModifiedAs shown, all other modified LP008-02 bifunctional molecules bind to human TGF-β1 with high activity, comparable to LP008-02-1.

[0264] To evaluate the effect of the modified LP008-02 bifunctional molecules on classical TGF-β signaling, the modified bifunctional molecules were tested using a luciferase assay. In the presence of recombinant human TGF-β, serial dilutions of the bifunctional molecules were incubated with 293 cells transfected with the SBE luciferase reporter for 24 hours. As TAGHTQTSTGGGAITTGTSGAGHGP shown, in the TGF-β SBE luciferase reporter assay system constructed in 293 cells, like LP008-02-1, LP008-02-2, LP008-02-3, LP008-02-6, and LP008-02-7 effectively blocked classical TGF-β signaling (IC50 = 0.04231 nM, IC50 = 0.0527 nM, IC50 = 0.09616 nM, and IC = 0.1962 nM).

[0265] The bifunctional molecules were dissolved in two buffers respectively for antibody stability detection. Buffer information is as follows: Buffer A: 20 mM sodium acetate, 250 mM sorbitol, 0.02% polysorbate 80, pH 4.9; Buffer B: 20 mM His / HisHCl, 250 mM trehalose, pH 5.4.

[0266] The prepared samples at 3.0 mg / ml were incubated at 40 °C and then detected by SEC-HPLC and CE-SDS on day 0 and day 14 respectively. As shown in Table 18, LP008-02-2, LP008-02-3, LP008-02-6, and LP008-02-7 formulated in Buffer A and Buffer B had higher stability than LP008-02-1 in SEC-HPLC, non-reducing CE-SDS, and reducing CE-SDS.

[0267] Table 18. SEC-HPLC and CE-SDS results of the test samples on day 14

[0268]

[0269]

[0270] Thus, this example demonstrates that the modified bifunctional molecules LP008-02-2, LP008-02-3, LP008.02-6, and LP008-02-7 exhibit activities similar to LP008-02-1, but their stability is significantly higher than that of LP008-02-1. Replacing the N-terminal portion of TGF-βRII in LP008-02-1 (IPPHVQKSVNNDMIVTDNNGAVKFP, SEQ ID NO:89) with an artificial linker (such as TAGHTQTSTGGGAITTGTSGAGHGP (SEQ ID NO:87) or HYP) results in a significant improvement in stability.

[0271] Example 18. High-Concentration Formulations of Anti-PD-L1 Antibodies

[0272] In this example, HIC-HPLC and viscosity tests were used to evaluate the development potential and risks of high-concentration anti-PD-L1 molecule formulations.

[0273] Four anti-PD-L1 molecules were expressed in CHO-K1 or 293F cells by transient transfection. The constant region of the heavy chain was human IgG1 (N297A)-Fc. The purified MPDL3280A (Atezolizumab), 47C6A3, Hu67F3G7-22, and Hu89C10H8-7 antibodies were tested by HIC-HPLC, and the ammonium sulfate concentration corresponding to the hydrophobic elution time was obtained to predict the solubility range of these molecules. As shown in Table 19, the ammonium sulfate concentrations corresponding to the hydrophobic elution time of MPDL3280A, 47C6A3, Hu67F3G7-22, and Hu89C10H8-7 were 0.41 M, 0.78 M, 0.97 M, and 1.10 M, respectively. All newly developed antibodies have lower hydrophobicity than the reference antibody MPDL3280A.

[0274] Table 19. Hydrophobicity of Antibodies from HIC-HPLC Tests

[0275] Figure 21 Figure 22 Sample 0.41 HIC(M) 0.78 MPDL3280A 0.97 47C6A3 1.10

[0276] Then, the activity of anti-PD-L1 antibodies in blocking the PD1 / PD-L1 interaction was measured using a bioluminescence cell-based assay. In this assay, when PD1 effector cells are co-cultured with PD-L1 target cells, the PD-1 / PD-L1 interaction inhibits TCR signaling and NFAT RE-mediated luminescence. Adding anti-PD-1 or anti-PD-L1 antibodies that block the PD-1 / PD-L1 interaction will release the inhibitory signal and result in TCR activation and NFAT-RE-mediated luminescence. As Hu67F3G7-22 Hu89C10H8-7 Figure 23As shown, MPDL3280A, 47C6A3, Hu67F3G7-22 and Hu89C10H8-7 block the PD1 and PD-L1 interaction with rather high activity (MPDL3280A EC 50 = 0.1327 nM, 47C6A3 EC 50 = 0.1501 nM, Hu67F3G7-22 EC 50 = 0.1034 nM, Hu89C10H8-7 EC 50 = 0.2138 nM).

[0277] MPDL3280A and Hu67F3G7-22 with human IgG1 Fc were expressed in CHO-K1 cells by transient transfection. The purified MPDL3280A-hIgG1 Fc and Hu67F3G7-22-hIgG1 Fc antibodies were tested by HIC-HPLC, and the ammonium sulfate concentration corresponding to the hydrophobic wash-off time was obtained to predict the solubility range of the two molecules. As shown in Table 20, the ammonium sulfate concentrations corresponding to the hydrophobic wash-off time of MPDL3280A-hIgG1 Fc and Hu67F3G7-22-hIgG1 Fc were 0.42 M and 0.99 M, respectively. Similarly, for the same Fc fragment, Hu67F3G7-22 showed lower hydrophobicity than MPDL3280A.

[0278] Table 20. Results of HIC-HPLC test

[0279]

[0280] To further confirm the solubility and viscosity characteristics of the antibodies, the two purified candidates were directly concentrated in phosphate buffer (including 60 mM NaCl) by ultrafiltration. During the ultrafiltration process, the concentration, SEC-HPLC and viscosity characteristics were measured at different stages. As shown in Table 21, at similar concentrations, the viscosity of MPDL3280A-hIgG1 Fc was much higher than that of Hu67F3G7-22-hIgG1 Fc. For high-concentration formulations, antibodies with lower viscosity are generally superior to those with higher viscosity. Therefore, as a therapeutic protein, the Hu67F3G7-22 antibody has higher potential than MPDL3280A.

[0281] Table 21. Results of solubility test

[0282]

[0283]

[0284] ***

[0285] The scope of the present disclosure is not limited by the specific embodiments described, which are intended to be illustrative of various aspects of the present disclosure, and any compositions or methods that are functionally equivalent are within the scope of the present disclosure. It will be apparent to those skilled in the art that various modifications and variations can be made to the methods and compositions of the present disclosure without departing from the spirit or scope of the present disclosure. Accordingly, the present disclosure is intended to cover modifications and variations of the present disclosure, provided they are within the scope of the appended claims and their equivalents.

[0286] All publications and patent applications mentioned in this specification are incorporated herein by reference to the same extent as if each individual publication or patent application was specifically and individually indicated to be incorporated by reference.

Claims

1. A multifunctional molecule comprising an anti-PD-L1 (programmed death ligand 1) antibody and the extracellular domain of human TGF-βRII (transforming growth factor-β receptor type II), wherein the anti-PD-L1 antibody comprises a heavy chain containing a heavy chain variable region (VH) and a separate light chain containing a light chain variable region (VL), and the C-terminus of the heavy chain of the anti-PD-L1 antibody is fused to the N-terminus of the extracellular domain of human TGF-βRII through a peptide linker; wherein the anti-PD-L1 antibody comprises a VH containing VH CDR1, VH CDR2 and VH CDR3, and a VL containing VL CDR1, VL CDR2 and VL CDR3, wherein the amino acid sequences of VH CDR1, VH CDR2, VH CDR3, VL CDR1, VL CDR2 and VL CDR3 are SEQ ID NO:13-18 respectively; wherein the amino acid sequence of the extracellular domain of human TGF-βRII is selected from one of SEQ ID NO:61 and 75-78.

2. The multifunctional molecule according to claim 1, wherein: (1) the amino acid sequence of the VH is selected from SEQ ID NO:31, and the amino acid sequence of the VL is selected from SEQ ID NO:38; (2) the amino acid sequence of the VH is selected from SEQ ID NO:33, and the amino acid sequence of the VL is selected from one of SEQ ID NO:39-43; (3) the amino acid sequence of the VH is selected from SEQ ID NO:34, and the amino acid sequence of the VL is selected from one of SEQ ID NO:39-43; (4) the amino acid sequence of the VH is selected from SEQ ID NO:35, and the amino acid sequence of the VL is selected from one of SEQ ID NO:39-42; (5) the amino acid sequence of the VH is selected from SEQ ID NO:36, and the amino acid sequence of the VL is selected from one of SEQ ID NO:39-42; or (6) the amino acid sequence of the VH is selected from SEQ ID NO:37, and the amino acid sequence of the VL is selected from one of SEQ ID NO:39-42.

3. The multifunctional molecule according to claim 2, wherein the amino acid sequence of the VH is selected from SEQ ID NO:34, and the amino acid sequence of the VL is selected from SEQ ID NO:

43.

4. The multifunctional molecule according to any one of claims 1-3, which comprises a light chain containing the VL and a light chain constant region, and a heavy chain containing the VH, a heavy chain constant region, a peptide linker, and the extracellular domain of TGF-βRII.

5. The multifunctional molecule according to claim 4, wherein the amino acid sequence of the heavy chain constant region is SEQ ID NO:59.

Images