Bifunctional molecules targeting PD-L1 and TGF-beta

Abstract

Provided is an anti-PD-L1 antibody which has excellent activity in blocking the interaction between PD-1 and PD-L1. A multifunctional molecule comprising the anti-PD-L1 antibody or fragment thereof fused to an extracellular structure of human TGF-beta receptor type 2 is also provided.

Description

[0001] This application is a divisional application of the invention patent application which entered the Chinese national stage on January 16, 2023, with application number 202180061312.1, application date on July 27, 2021, and invention name “Bifunctional molecules targeting PD-L1 and TGF-β”. Background Art

[0002] Exciting advances in cancer immunotherapy in recent years have led to a paradigm shift in oncology. The most striking results are T cell-based therapies, including immune checkpoint inhibitors (ICIs), genetically engineered T cells, and bispecific antibodies (BsAbs). T cells are a large class of immune surveillance and tumor eradication, with high specificity and long-term memory. However, in the tumor microenvironment, T cells may be exhausted or become tolerant to tumor cells. T cell exhaustion is often associated with 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, thereby removing the brakes that inhibit tumor-infiltrating T cells, resulting in significant clinical benefits in different malignancies. For example, blocking the PD-1 / PD-L1 interaction can enhance immune normalization and enhance anti-cancer responses. However, a significant drawback of PD-1 / PD-L1 blockade is the inconsistency in homogenous study populations with similar tumor characteristics. In addition, PD-1 / PD-L1 blockade therapy may also lead to certain inflammatory side effects in some patients. The limitations of PD-1 / PD-L1 blockade monotherapy and the lack of promising alternatives make it necessary to seek combination therapy approaches that can activate anti-tumor immunity and improve therapeutic efficacy.

[0004] M7824 (bintrafusp alfa) is a bifunctional protein consisting of a monoclonal antibody against programmed death ligand 1 (PD-L1) fused to the extracellular domain of human transforming growth factor-β (TGF-β) receptor II, which acts as a "trap" for all three TGF-β isoforms. The PD-L1 portion 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 unwanted skin growths, and in a phase II trial in patients with HPV-positive malignancies, the overall response rate was only 35% to 40%. Therefore, improved therapies are needed.

[0005] Overview

[0006] In some embodiments, the present disclosure provides bifunctional molecules targeting 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 human transforming growth factor-β (TGF-β) receptor II, which serves as a trap for TGF-β. Experimental data show that these new bifunctional molecules are more effective than the current lead candidate drug M7824 in clinical development.

[0007] Therefore, according to one embodiment of the present disclosure, a multifunctional molecule is provided, which includes an anti-PD-L1 (programmed death ligand 1) antibody or a fragment thereof and an extracellular domain of human TGF-βRII (TGF-β receptor type 2), wherein the anti-PD-L1 antibody or a fragment thereof is specific to human PD-L1 protein, and includes a heavy chain variable region (VH) comprising VH CDR1, VH CDR2 and VH CDR3, and a light chain variable region (VL) comprising 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 sequence 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, and VL CDR1 includes SEQ ID NO: 22. NO:22, VL CDR2 includes SEQ ID NO:23, and VL CDR3 includes SEQ ID NO:24 or 93, wherein the human TGF-βRII extracellular domain includes the amino acid sequence of SEQ ID NO:72 and is fused to the anti-PD-L1 antibody or a fragment thereof.

[0008] In one embodiment, an anti-PD-L1 (programmed death-ligand 1) antibody or a fragment thereof is provided, which is specific to human PD-L1 protein and includes a heavy chain variable region (VH) comprising VH CDR1, VH CDR2 and VH CDR3, and a light chain variable region (VL) comprising VL CDR1, VL CDR2 and VL CDR3, wherein the VH CDR1, VH CDR2, VHCDR3, VL CDR1, VL CDR2 and VL CDR3 include the amino acid sequences of SEQ ID NOs: 7-12, or SEQ ID NOs: 13-18, respectively, or wherein VH CDR1 includes SEQ ID NO: 19, VH CDR2 includes SEQ ID NOs: 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] Also provided is a multifunctional molecule comprising an antibody or an 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) is at least 30 amino acid residues in length, or (b) is at least 25 amino acid residues in length and comprises an α-helical motif.

[0010] Also provided are uses and methods of treating cancer with any of the molecules disclosed herein. BRIEF DESCRIPTION OF THE DRAWINGS

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

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

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

[0014] Figure 4 It was 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] Fig. 6A -C shows that all tested humanized antibodies bind to human PD-L1 with comparable efficiency as the chimeric antibody.

[0017] Figure 7 The humanized antibodies tested were shown to bind to PD-L1 expressed on mammalian cells with high efficiency, comparable to the chimeric antibodies.

[0018] Fig. 8A -C shows that some humanized antibodies can effectively inhibit the binding of human PD-L1 to human PD1.

[0019] Fig.9A -C shows that some humanized antibodies can effectively inhibit the binding of human PD-L1 to human CD80.

[0020] Fig.10 The binding kinetics of LP008-06, LP008-06a, LP008-06a-DA, and LP008-06a-ES to recombinant human PD-L1 are shown.

[0021] Fig.11 The binding kinetics of LP008-02 to human PD-L1 and human TGF-β1 are shown.

[0022] Fig.12 It was shown that LP008-02 and LP008-06a-ES can block PD1 and PD-L1 interaction with higher affinity than M7824.

[0023] Fig.13 It showed that M7824, LP008-02 and LP008-06a-ES could effectively block TGF-β classical signaling.

[0024] Fig.14 LP008-02 and LP008-06a-ES were shown to bind human PD-L1 with high affinity.

[0025] Fig.15 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] Fig.16 The results showed that the binding efficiency of LP008-02 and LP008-06a-ES to human TGF-β was comparable to that of M7824.

[0027] Fig.17It was shown that LP008-02 and LP008-06a-ES bind to cynomolgus monkey TGF-β, mouse TGF-β, and rat TGF-β with comparable efficiency to M7824.

[0028] Fig.18A -B shows the drug effects of LP008-02 and LP008-06a-ES in animal models.

[0029] Fig.19 All modified bifunctional molecules tested showed comparable binding efficiency to human TGF-β as LP008-02-1.

[0030] Fig. 20 It was shown that all modified bifunctional molecules tested could effectively block TGF-β canonical signaling.

[0031] Fig.21 All modified bifunctional molecules tested showed comparable binding efficiency to human TGF-β as LP008-02-1.

[0032] Fig. 22 It was shown that all modified bifunctional molecules tested could effectively block TGF-β canonical signaling.

[0033] Fig.23 It was shown that antibodies MPDL3280A, 47C6A3, Hu67F3G7-22, and Hu89C10H8-7 can block PD1 and PD-L1 interaction with high affinity. Detailed description

[0035] definition

[0036] It should be noted that the term "a" or "an / an" entity refers to one or more of the entity; for example, "an antibody" is understood to represent one or more antibodies. Therefore, the terms "a" (or "a"), "one or more" and "at least one" can be used interchangeably herein.

[0037] As used herein, "antibody" or "antigen-binding polypeptide" refers to a polypeptide or polypeptide complex that specifically recognizes and binds to an antigen. The antibody can be a complete antibody and any antigen-binding fragment or single chain thereof. Therefore, the term "antibody" includes any protein- or peptide-containing molecule that comprises at least a portion of an immunoglobulin molecule that has a biological activity of binding to an antigen. Such examples include, but are not limited to, a heavy or light chain complementary determining region (CDR) or a ligand-binding portion thereof, a heavy 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.

[0038] As used herein, the term "antibody fragment" or "antigen-binding fragment" is a portion of an antibody, such as F(ab') 2 、F(ab) 2 , Fab', Fab, Fv, scFv, etc. Regardless of the structure, antibody fragments bind 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.

[0039] The term "antibody" includes a wide variety of polypeptide classes that can be distinguished biochemically. Those skilled in the art will appreciate that heavy chains are classified as gamma, mu, alpha, delta, epsilon, with some subclasses (e.g., gamma 1-gamma 4). It is the nature of this chain that determines the "class" of an antibody, IgG, IgM, IgA, IgG, or IgE, respectively. Immunoglobulin subclasses (isotypes), such as IgG 1 IgG 2 IgG 3 IgG 4 IgG 5 etc., are well characterized and are known to confer functional specificity. In view of the present disclosure, modified versions of each of these classes and isotypes are readily discernible to those skilled in the art and, therefore, are 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 regard 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. The four chains are typically connected by disulfide bonds in a "Y" configuration, with the light chains surrounding the heavy chains, starting at the mouth of the "Y" and continuing through the variable region.

[0040] "Specific binding" or "specific for..." generally means that the antibody binds to the epitope through its antigen binding domain, and that such binding requires some complementarity between the antigen binding domain and the epitope. According to this definition, an antibody is said to "specifically bind" to an epitope when it binds to the epitope more easily through its antigen binding domain than to a random, unrelated epitope. The term "specificity" is used herein to qualify the relative affinity with which a particular antibody binds 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 to epitope "C" with a higher specificity than it does to a related epitope "D".

[0041] As used herein, the term "treatment" refers to therapeutic treatment and preventive or preventive measures, wherein the purpose is to prevent or slow down (mitigate) undesirable physiological changes or disorders, such as cancer progression. Beneficial or desired clinical results include, but are not limited to, symptom relief, disease severity relief, stable disease state (i.e., no deterioration), delay or slowing of disease progression, improvement or alleviation of disease state, and relief (whether partial or complete), whether detectable or undetectable. "Treatment" also means extending survival compared to the expected survival without treatment. People in need of treatment include people who already have the disease or condition, as well as people who are susceptible to the disease or condition or people who need to prevent the disease or condition.

[0042] "Subject" or "individual" or "animal" or "patient" or "mammal" refers to any subject, especially a mammalian subject, for whom diagnosis, prognosis or treatment is desired. Mammalian subjects include humans, livestock, farm animals, zoo animals, sports animals or pet animals, such as dogs, cats, guinea pigs, rabbits, rats, mice, horses, cows, dairy cows, etc.

[0043] As used herein, phrases such as "to a patient in need of treatment" or "a subject in need of treatment" include subjects, such as mammalian subjects, who would benefit from administering an antibody or composition of the invention, e.g., for detection, diagnostic procedures and / or therapy.

[0044] Multifunctional molecules

[0045] As demonstrated in the attached experimental examples, the inventors were able to identify many bifunctional fusion proteins including an anti-PD-L1 unit and a TGF-β targeting unit. For example, as shown in Example 14, in the MC38 mouse model, both tested bifunctional proteins LP008-02 and LP008-06a-ES showed better efficacy than M7824. M7824 is a PD-L1 / TGF-β dual-targeting fusion protein that is currently undergoing a Phase II clinical trial in patients with HPV-positive malignancies. The anti-PD-L1 unit of M7824 is based on avelumab, a leading PD-L1 antibody that has been approved for the treatment of Merkel cell carcinoma and urothelial carcinoma. Therefore, the superior performance of the newly disclosed bifunctional protein compared to M7824 is surprising.

[0046] In addition, as shown in Example 12, the bifunctional protein of the present disclosure has better species specificity. Unlike M7824, which also reacts with mouse and rat PD-L1, the new bifunctional protein, in addition to having excellent PD-L1 binding activity, also only binds to human and cynomolgus monkey PD-L1.

[0047] Therefore, 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 may include an anti-PD-L1 antibody or fragment of the present disclosure. The TGF-β targeting unit is preferably the extracellular domain of human transforming growth factor-β (TGF-β) receptor II (TGF-βRII or TGFBR2).

[0048] There are two subtypes of TGF-βRII. Subtype A (NP_001020018.1; SEQ ID NO: 70) has a longer extracellular segment 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.

[0049] Table A. Sequences related to TGF-βRII (underlined and bold: core ectodomain; underlined and italicized: residues that differ between isoforms; underlined only: mutations)

[0050]

[0051]

[0052] 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.

[0053] 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.

[0054] Another variant, variant 3, is based on variant 1, but includes at least amino acid substitutions at positions X within the N-terminal portion (SEQ ID NO: 88). These X positions are potential glycosylation sites. Therefore, substitutions are made with amino acids other than K, S and N. Examples of substitutions are R, A, G, Q, I, L, D or E, but are not limited thereto.

[0055] 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, a Fab fragment, a single chain fragment, or a single domain antibody, but is not limited thereto. When the antibody or fragment thereof has a light chain and a separate heavy chain, the TGF-βRII extracellular domain can be fused to the light chain or the heavy chain. When the antibody or fragment thereof has a light chain and a heavy chain on a single protein chain (e.g., scFv), the TGF-βRII extracellular domain can be fused to be closer to the light chain or the heavy chain.

[0056] In some embodiments, the TGF-βRII extracellular domain is fused to the N-terminus of the chain of the anti-PD-L1 unit. In some embodiments, the TGF-βRII extracellular domain is fused to the C-terminus of the chain of the anti-PD-L1 unit. In a preferred embodiment, the TGF-βRII extracellular domain 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 repeats of GGGGS (SEQ ID NO: 86)).

[0057] 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.

[0058] In one embodiment, VH CDR1, VH CDR2, VH CDR3, VL CDR1, VL CDR2 and VL CDR3 respectively comprise the sequence 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.

[0059] 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, having a VH of SEQ ID NO: 35 and a VL of SEQ ID NO: 39, or 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.

[0060] In one embodiment, VH CDR1, VH CDR2, VH CDR3, VL CDR1, VL CDR2 and VL CDR3 respectively comprise the sequence 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). 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).

[0061] In one embodiment, VH CDR1, VH CDR2, VH CDR3, VL CDR1, VL CDR2 and VL CDR3 comprise the sequence 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 comprise the sequence 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 comprise the sequence 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.

[0062] In one embodiment, VH CDR1, VH CDR2, VH CDR3, VL CDR1, VL CDR2 and VL CDR3 comprise the sequence 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 sequence 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 sequence 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.

[0063] 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).

[0064] 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 comprises a VH of SEQ ID NO: 48 and a VL of SEQ ID NO: 53. In one embodiment, the humanized antibody comprises a VH of SEQ ID NO: 48 and a VL of SEQ ID NO: 56. In one embodiment, the humanized antibody comprises a VH of SEQ ID NO: 57 and a VL of SEQ ID NO: 56. In one embodiment, the humanized antibody comprises a VH of SEQ ID NO: 58 and a VL of SEQ ID NO: 56.

[0065] In some embodiments, the antibody or fragment thereof further comprises a heavy chain constant region (e.g., CH1, CH2 and / or CH3) and / or a light chain constant region (e.g., CL). An exemplary heavy chain constant region is provided in SEQ ID NO: 59, and an exemplary light chain constant region is provided in SEQ ID NO: 67 (residues 108-214).

[0066] TGF-βRII x Antibody Fusion

[0067] Tests using different fusion protein designs (e.g., Table 15) showed that only the core extracellular domain of TGF-βRII (SEQ ID NO: 72) is required for activity. In addition, the extracellular domain of TGF-βRII should not be fused directly to the antibody. There should be sufficient distance, provided by the peptide linker.

[0068] About extracellular domain, peptide linker (it can be completely artificial linker, or comprise the part extracellular fragment SEQ ID NO of N-terminal of extracellular domain:89) should have minimum length.If distance is too short, then the stability or activity of fusion protein is reduced.In some embodiments, 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 no 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.

[0069] In some embodiments, the inclusion of a flexible linker, such as one or more G4S (SEQ ID NO: 86) units, is useful for the stability and / or activity of the multifunctional molecule. 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.

[0070] In some embodiments, it is shown (e.g., 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 improving activity. In some embodiments, the alternative peptide is different from SEQ ID NO: 89, but has at least 30%, 40%, 50%, 60%, 70%, 80% or 90% sequence identity with SEQ ID NO: 99.

[0071] An exemplary replacement peptide is IPPHVQXXVNNDMIVTDNXGAVKFP (SEQ ID NO: 88), wherein X is any amino acid except 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 to SEQ ID NO: 87. In some embodiments, the variant includes at least 4 Gs, no PP dipeptide, no more than 3 hydrophobic amino acid residues selected from I, L, M, F, V, W, Y and P. In some embodiments, the variant includes at least 5 Gs and no more than 1 hydrophobic amino acid residue selected from I, L, M, F, V, W, Y, and P.

[0072] In some embodiments, the peptide linker between the antibody or fragment thereof and the extracellular domain of TGF-βRII (SEQ ID NO: 72) comprises a flexible linker. In some embodiments, the peptide linker comprises a surrogate peptide of SEQ ID NO: 89. In some embodiments, the peptide linker comprises both a flexible linker and a surrogate peptide. In some embodiments, the flexible linker is at the N-terminus of the surrogate peptide. In some embodiments, the flexible linker is at the C-terminus of the surrogate peptide.

[0073] In some embodiments, the multifunctional molecule does not include at least the entire sequence of EEYNTSNPD (SEQ ID NO: 90). The multifunctional molecule may 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).

[0074] The antibody or antigen-binding fragment 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, ICOSL, GITR, GITRL, TIGIT, CD27, VISTA, B7H3, B7H4, BTLA, CD4, CD2, CD8, CD47 and CD73. They can also be any antibody or fragment disclosed herein.

[0075] 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 chain 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.

[0076] Anti-PD-L1 antibodies and fragments

[0077] Also provided are anti-PD-L1 antibodies and fragments that can be used as multifunctional molecules, as anti-PD-L1 units in bispecific or multispecific antibodies, or alone in monospecific antibodies.

[0078] Exemplary mouse anti-PD-L1 antibodies and their humanized and modified antibodies have been prepared and tested in the accompanying experimental examples. All mouse antibodies (47C6A3, 67F3G7 and 89C10H8) and their corresponding humanized versions showed excellent binding affinity, cross-reactivity and effectiveness in inhibiting PD-1 / PD-L1 binding.

[0079] Importantly, compared to MPDL3280A (Atezolizumab), humanized 67F3G7 and 89C10H8 showed higher activity than MPDL3280A in blocking the interaction between PD-1 and PD-L1 (see, e.g., Example 18). In addition, it is interesting that all tested antibodies disclosed herein show lower hydrophobicity and lower viscosity than MPDL3280A. It is known that higher hydrophobicity reduces the solubility of proteins. Similarly, high viscosity is also an obstacle to the development of high-concentration protein formulations. Therefore, these data show that the antibodies of the present invention are more suitable for preparing high-concentration antibody formulations.

[0080] In addition, the antigen-binding fragment of the antibody disclosed herein is included as a unit in a bifunctional fusion protein, which also includes a TGF-β targeting unit. In the MC38 mouse model, the resulting bifunctional fusion protein showed better efficacy than M7824. M7824 is a PD-L1 / TGF-β dual-targeting fusion protein that is currently undergoing a Phase II clinical trial in patients with HPV-positive malignancies. The anti-PD-L1 unit of M7824 is based on avelumab, a leading PD-L1 antibody that has been approved for the treatment of Merkel cell carcinoma and urothelial carcinoma. Therefore, these data demonstrate the unique advantages of the antibodies disclosed herein in the preparation of bifunctional or multifunctional molecules.

[0081] 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, such as those shown in Tables 1A-1C.

[0082] In one embodiment, VH CDR1, VH CDR2, VH CDR3, VL CDR1, VL CDR2 and VL CDR3 respectively comprise the sequence 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.

[0083] 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, having a VH of SEQ ID NO: 35 and a VL of SEQ ID NO: 39, or 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.

[0084] In one embodiment, VH CDR1, VH CDR2, VH CDR3, VL CDR1, VL CDR2 and VL CDR3 respectively comprise the sequence 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). 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).

[0085] In one embodiment, VH CDR1, VH CDR2, VH CDR3, VL CDR1, VL CDR2 and VL CDR3 comprise the sequence 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 comprise the sequence 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 comprise the sequence 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.

[0086] In one embodiment, VH CDR1, VH CDR2, VH CDR3, VL CDR1, VL CDR2 and VL CDR3 comprise the sequence 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 sequence 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 sequence 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.

[0087] 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).

[0088] 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 comprises a VH of SEQ ID NO: 48 and a VL of SEQ ID NO: 53. In one embodiment, the humanized antibody comprises a VH of SEQ ID NO: 48 and a VL of SEQ ID NO: 56. In one embodiment, the humanized antibody comprises a VH of SEQ ID NO: 57 and a VL of SEQ ID NO: 56. In one embodiment, the humanized antibody comprises a VH of SEQ ID NO: 58 and a VL of SEQ ID NO: 56.

[0089] In some embodiments, the antibody or fragment thereof further comprises a heavy chain constant region (e.g., CH1, CH2 and / or CH3) and / or a light chain constant region (e.g., CL). An exemplary heavy chain constant region is provided in SEQ ID NO: 59, and an exemplary light chain constant region is provided in SEQ ID NO: 67 (residues 108-214).

[0090] It is expected that small changes (e.g., addition, deletion or substitution of an amino acid) can be designed in these CDR sequences, which can retain the activity of the antibody or even improve them. Such modified CDR sequences are referred to as CDR variants. It will also be appreciated by those of ordinary skill in the art that the antibodies disclosed herein can be modified so 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 can be similar to the starting sequence, for example, having a certain percentage of identity, for example, it can 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.

[0091] In certain embodiments, the antibody comprises 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, an antibody of the present disclosure may comprise a flexible linker sequence, or may be modified to add a functional moiety (e.g., PEG, a drug, a toxin, or a label).

[0092] Polynucleotide encoding protein and method for preparing protein

[0093] The present disclosure also provides isolated polynucleotides or nucleic acid molecules encoding multifunctional proteins, antibodies, variants or derivatives thereof of the present disclosure. The polynucleotides of the present disclosure can encode the entire heavy chain and light chain variable region of antigen-binding polypeptides, variants or derivatives thereof on the same polynucleotide molecule or on isolated polynucleotides. In addition, the polynucleotides of the present disclosure can encode a portion of the heavy chain and light chain variable region of antigen-binding polypeptides, variants or derivatives thereof on the same polynucleotide molecule or on isolated polynucleotides.

[0094] Methods for preparing antibodies are well known in the art and described herein. In certain embodiments, the variable regions and constant regions of the antigen-binding polypeptides disclosed herein are fully human. Fully human antibodies can be prepared using techniques described in the art and as described herein. For example, fully human antibodies against a specific antigen can be prepared by administering an antigen to a transgenic animal that has been modified to produce such an antibody in response to an antigenic challenge, but whose endogenous loci have been disabled. Exemplary techniques for making 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.

[0095] Cancer treatment

[0096] As described herein, the antibodies, variants or derivatives of the present disclosure may be used in certain therapeutic and diagnostic methods.

[0097] The present invention further relates to multifunctional molecule or antibody-based treatments, which involve administering the multifunctional molecules and antibodies of the present invention to patients (e.g., animals, mammals, and humans) to treat one or more diseases or conditions described herein. Therapeutic compounds of the present invention include, but are not limited to, antibodies of the present invention (including variants and derivatives thereof as described herein) and nucleic acids or polynucleotides encoding antibodies of the present invention (including variants and derivatives thereof as described herein).

[0098] 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 limiting anti-tumor T cell immunity. The results of using small molecule inhibitors or monoclonal antibodies against PD-L1 in mouse tumor models show that targeted PD-L1 therapy is an important alternative and realistic method for effectively controlling tumor growth. As shown in the experimental examples, anti-PD-L1 antibodies activate adaptive immune response mechanisms, which can improve the survival rate of cancer patients.

[0099] Therefore, in some embodiments, a method of treating cancer in a patient in need is provided. In one embodiment, the method requires 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., a 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 radiation therapy.

[0100] Tumors expressing PD-L1 protein include bladder cancer, non-small cell lung cancer, kidney 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. Therefore, the antibodies of the present disclosure can be used to treat any one or more of these cancers.

[0101] The present invention also provides cell therapy, such as chimeric antigen receptor (CAR) T cell therapy. Suitable cells can be used, which are contacted with the anti-PD-L1 antibodies disclosed herein (or alternatively engineered to express the anti-PD-L1 antibodies disclosed herein). After such contact or engineering, the cells can be introduced into cancer patients in need of treatment. Cancer patients may suffer from any type of cancer disclosed herein. Cells (such as 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.

[0102] In some embodiments, cells are isolated from cancer patients themselves. In some embodiments, cells are provided by donors or cell banks. When cells are isolated from cancer patients, undesirable immune responses can be minimized.

[0103] Other diseases or conditions associated with increased cell survival that can be treated, prevented, diagnosed and / or predicted with the antibodies of the present disclosure, or variants or derivatives thereof, include, but are not limited to, progression, and / or metastasis of malignancies and related diseases, such as leukemias (including acute leukemias (e.g., acute lymphocytic leukemia, acute myeloid leukemia (including myeloblastic, promyelocytic, myelomonocytic, monocytic and erythroleukemia)) and chronic leukemias (e.g., chronic myeloid (granulocytic) leukemia and chronic lymphocytic leukemia)), polycythemia vera, lymphomas (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 fibrosarcomas, myxosarcoma, liposarcoma, chondrosarcoma, osteoblastic sarcoma, Tumors, 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, liver cancer, bile duct carcinoma, choriocarcinoma, seminoma, embryonal carcinoma, Wilm's tumor, cervical cancer, testicular tumors, 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.

[0104] Composition

[0105] The present invention also provides a pharmaceutical composition. Such a composition comprises an effective amount of an antibody and an acceptable carrier. In some embodiments, the composition further comprises a second anticancer agent (e.g., an immune checkpoint inhibitor).

[0106] In a specific embodiment, the term "pharmaceutically acceptable" means approved by a federal or state regulatory agency, or listed in the U.S. Pharmacopeia or other generally recognized pharmacopeia for use in animals, especially 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 auxiliary.

[0107] The term "carrier" refers to a diluent, adjuvant, excipient or carrier used together with the therapeutic agent. This pharmaceutical carrier can be a sterile liquid, such as water and oil, including liquids from petroleum, animal, plant or synthetic sources, such as peanut oil, soybean oil, mineral oil, sesame oil, etc. When the pharmaceutical composition is administered intravenously, water is a preferred carrier. 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 gel, sodium stearate, glyceryl monostearate, talc, sodium chloride, dry skim milk, glycerol, propylene, ethylene glycol, water, ethanol, etc. If desired, the composition can also contain a small amount of wetting agent or emulsifier, or pH buffer, such as acetate, citrate or phosphate. Antimicrobial agents such as benzyl alcohol or methyl parabens; antioxidants such as ascorbic acid or sodium bisulfite; chelating agents such as ethylenediaminetetraacetic acid; and agents for adjusting tension such as sodium chloride or glucose can also be envisioned. These compositions can take the form of solutions, suspensions, emulsions, tablets, pills, capsules, powders, sustained-release formulations, and the like. The composition can be formulated as a suppository with conventional binders and carriers such as triglycerides. Oral formulations may include standard carriers such as pharmaceutical grade mannitol, lactose, starch, magnesium stearate, sodium saccharin, cellulose, magnesium carbonate, and the like. Examples of suitable pharmaceutical carriers are described in Remington's Pharmaceutical Sciences by EW Martin, which is incorporated herein by reference. Such compositions will contain a therapeutically effective amount of an antigen-binding polypeptide, preferably in purified form, and an appropriate amount of a carrier to provide a form of appropriate administration to the patient. The formulation should conform to the mode of administration. The parent formulation can be packaged in an ampoule, a disposable syringe, or a multi-dose vial made of glass or plastic.

[0108] In one embodiment, the composition is formulated into a pharmaceutical composition suitable for intravenous administration to a human being according to conventional procedures. Typically, the composition for intravenous administration is a solution in a sterile isotonic aqueous buffer. If necessary, the composition may also include a solubilizing agent and a local anesthetic, such as lidocaine, to alleviate pain at the injection site. Typically, the composition is supplied separately or mixed together in a unit dose form, for example, as a dry lyophilized powder or anhydrous concentrate, in a container (such as an ampoule or a pouch) indicating the amount of the active agent. If the composition is administered by infusion, an infusion bottle containing sterile pharmaceutical grade water or saline may be used for distribution. In the case of a composition administered by injection, an ampoule of sterile water for injection or saline may be provided so that the composition may be mixed before administration. Example

[0109] Example 1: Generation of mouse monoclonal antibodies against human PD-L1

[0110] This example describes the generation of anti-human PD-L1 mouse monoclonal antibodies using hybridoma technology.

[0111] Antigens: human PDL1-Fc protein and human PD-L1 highly expressed on the CHOK1 cell line (PDL1-CHOK1 cell line).

[0112] Immunization: To generate mouse monoclonal antibodies targeting human PD-L1, Balb / c mice and Wistar rats were first immunized with PD-L1-Fc protein. Immunized mice and rats were subsequently enhanced with PD-L1-Fc protein and CHO-K1 / PD-L1 stable cells, respectively. To select mice or rats that produce antibodies that bind to PD-L1 protein, the antibody titer of sera from immunized mice or mice was evaluated by ELISA. Briefly, microtiter plates were coated with 0.5 μg / ml human PD-L1 protein in ELISA coating buffer, 100 μl / well at 4°C overnight, and then blocked with 150 μl / well of 1% BSA. Serum dilutions 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 matrix and analyzed with a spectrophotometer at OD 450nm. After 3 rounds of immunization, the immune response of the CHO-K1 / PDL-1 stable cell line was also detected by serum ELISA for rhPD-L1 protein and FACS, with the CHO-K1 parental cell line as a negative control. After 3 rounds of immunization, mice showing anti-PDL1-IgG with sufficient titers were boosted with 25μg human PDL1-Fc protein. The resulting mice were used for fusion. Hybridoma supernatants were tested for anti-PD-L1 IgG by ELISA.

[0113] Cell fusion: Fusion was performed by electrofusion. Fused cells were plated in 50 96-well plates for each fusion.

[0114] Screening: Supernatants were screened by ELISA against recombinant human (rh) PD-L1-Fc protein and counter-screening antigen. Then, positive supernatants were screened for confirmation by receptor blocking FACS against CHO-K1 / PD-L1 stable cell line and rhPD-1-Fc protein.

[0115] Subcloning and screening: Positive primary clones from each fusion were subcloned by limiting dilution to ensure that the subclones were derived from a single parental cell. Subclones were screened in the same manner as the primary clones, and culture supernatants of positive clones were additionally confirmed by affinity sorting.

[0116] 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.

[0117] Table 1. Sequences of the variable regions of 47C6A3, 67F3G7 and 89C10H8

[0118]

[0119] Table 1A. CDR sequences of 47C6A3

[0120] 47C6A3 sequence SEQ ID NO: CDRH1 SDYAWN 7 CDRH2 YIIYSGSTSYNPSLKS 8 CDRH3 STMIATNWFAY 9 CDRL1 KASQDVSLAVA 10 CDRL2 WASTRHT 11 CDRL3 QQHYITPWT 12

[0121] Table 1B. CDR sequences of 67F3G7

[0122] 67F3G7 sequence SEQ ID NO: CDRH1 DFWVS 13 CDRH2 EIYPNSGVSRYNEKFKG 14 CDRH3 YFYTYWFGY 15 CDRL1 RASKSVSTYMH 16 CDRL2 SASHLES 17 CDRL3 QQSNELPVT 18

[0123] Table 1C. CDR sequences of 89C10H8

[0124]

[0125] Example 2: Binding activity to PD-L1 antigen

[0126] ELISA test

[0127] To evaluate the binding activity of hybridoma clones 47C6A3, 67F3G7, and 89C10H8, chimeric mAbs from these clones were tested in ELISA.

[0128] Briefly, microtiter plates were coated with 0.5 μg / ml human PD-L1-Fc protein in PBS, 100 μl / well at 4°C overnight, and then blocked with 150 μl / well of 1% BSA. Three-fold dilutions of 47C6A3, 67F3G7, and 89C10H8 antibodies starting at 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 mouse anti-human IgG Fab antibodies 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 450nm. Figure 1 As shown, 47C6A3, 67F3G7, and 89C10H8 bind to human PD-L1 with high affinity (EC 50 =10.24 ng / ml, EC of 67F3G7 50 =10.76ng / ml, EC of 89C10H8 50 =8.112ng / ml).

[0129] Cell-based binding: FACS was used to evaluate the binding activity of 47C6A3, 67F3G7, and 89C10H8 chimeric mAbs to CHOK1 cells overexpressing human PD-L1.

[0130] Briefly, PDL1-CHOK1 cells were first incubated with 3-fold serial dilutions of 47C6A3, 67F3G7, and 89C10H8 chimeric mAbs starting at 100 nM at 4°C for 40 min. 647AffiniPure goat anti-human IgG (H+L) was added to each well and incubated at 4°C for 30 min. Samples were washed twice with FACS buffer. Alexa Fluor was evaluated by FACSCanto The mean fluorescence intensity (MFI) of 647. Figure 2 As shown, 47C6A3, 67F3G7, and 89C10H8 bound to PDL1-CHOK1 cells with high affinity (EC 50 =0.1476 nM, EC of 67F3G7 50 =0.1035nM, EC of 89C10H8 50 =0.1696nM).

[0131] Cross-species activity

[0132] ELISA tests were performed to evaluate the binding of chimeric antibodies to human, mouse, rat and cynomolgus monkey PD-L1, respectively.

[0133] Briefly, microtiter plates were coated with 0.5 μg / ml of human, mouse, rat, and cynomolgus PD-L1 proteins in PBS, 100 μl / well overnight at 4°C, and then blocked with 150 μl / well of 1% BSA. Chimeric antibodies diluted three times 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 mouse anti-human IgG Fab antibodies 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 450nm. 47C6A3, 67F3G7, and 89C10H8 antibodies bind to human and cynomolgus PD-L1 but not to rat and mouse PD-L1 ( Figure 3 and Table 2 ).

[0134] Table 2. Cross-species activities of 47C6A3, 67F3G7 and 89C10H8

[0135] people Cynomolgus monkey Rat Mouse EC50 of 47C6A3 10.24ng / ml 6.336ng / ml No combination No combination EC50 of 67F3G7 10.76ng / ml 6.797ng / ml No combination No combination EC50 of 89C10H8 8.112ng / ml 6.774ng / ml No combination No combination .

[0136] Example 3. Blocking the binding of PD-L1 to PD-1 by antibodies

[0137] To evaluate the blocking effect of 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 used.

[0138] Briefly, microtiter plates were coated with 0.5 μg / ml human PD-L1-Fc protein in PBS, 100 μl / well at 4°C overnight, and then blocked with 150 μl / well 1% BSA. 50 μl of biotinylated human PD-1-Fc protein and 50 μl of 47C6A3, 67F3G7 and 89C10H8 antibodies diluted 3 times 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 for 10 minutes at 37°C. After washing, the plates were developed with TMB substrate and analyzed with a spectrophotometer at OD 450nm. Figure 4 As shown, 47C6A3, 67F3G7 and 89C10H8 had IC values ​​of 91.18 ng / ml, 139.8 ng / ml and 129.8 ng / ml, respectively. 50 Effectively inhibits the binding of human PD-L1 to human PD1.

[0139] Example 4: Binding affinity of mAbs

[0140] The binding of 47C6A3, 67F3G7 and 89C10H8 antibodies to recombinant PD-L1 protein (human PD-L1-his tag) was tested with Biacore using the capture method. 47C6A3, 67F3G7 and 89C10H8 mAbs were captured using a protein A chip. Serial dilutions of human PD-L1-his tag protein were injected over the captured antibodies at a flow rate of 30 μl / min for 2 min. 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. Results are shown in Figure 5 and Table 3 below.

[0141] Table 3. Affinity measured by Biacore

[0142]

[0143] Example 5. Humanization of mouse antibodies

[0144] Humanized mAbs were created using 47C6A3, 67F3G7, and 89C10H8 variable region genes. 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 to available databases of human Ig gene sequences to find the best overall matching human germline Ig gene sequences. For the light chain of 47C6A3, human Vk1-4 was the most suitable germline, while 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, while 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, while for the heavy chain, the closest human match was the VH3-21 / JH3 gene.

[0145] For the VL of 47C6A3, human Vk1-4 is the most suitable germline, while for the VH of 47C6A3, human VH1-2 is selected as the framework. The humanized variable domain sequence of 47C6A3 was then 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. A 3D model was then generated to determine whether there were any framework positions where the mouse amino acids were replaced with human amino acids that might affect binding and / or CDR conformation. In the case of the heavy chain, R, M and I in the framework are associated with reverse mutations.

[0146] Then design the humanized variable domain sequence of 67F3G7, wherein CDRL1, L2 and L3 are grafted onto the framework sequence of Vk1-39 / JK4 gene, and CDRH1, H2 and H3 are grafted onto the framework sequence of VH1-2 / JH4-FW4 gene. Then generate 3D model to determine whether there is any framework position that may affect binding and / or CDR conformation by replacing mouse amino acid with human amino acid. In the case of heavy chain, V, K, T and I in the framework are related to reverse mutation. In the case of light chain, T, V, L and Q in the framework are related to reverse mutation.

[0147] The humanized variable domain sequence of 89C10H8 was then 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. A 3D model was then generated to determine if there were any framework positions where replacing mouse amino acids with human amino acids could affect binding and / or CDR conformation. In the case of the heavy chain, A, T, I and S in the framework were associated with reverse mutations. In the case of the light chain, Y, I, E and F in the framework were associated with reverse mutations.

[0148] The amino acid and nucleotide sequences of some humanized antibodies are listed in Table 4 below.

[0149] Table 4. Humanized antibody sequences (underlined CDRs; bold / italics indicate reverse mutations)

[0150]

[0151]

[0152]

[0153] These genes were cloned into pcDNA3.4 vector and transfected into 293F cells. Antibodies were prepared according to the table below.

[0154] Humanized VH and VL genes were generated synthetically and then cloned into vectors containing human γ1 and human κ constant domains, respectively. Pairing of human VH and human VL generated 41 humanized antibodies (see Table 5).

[0155] Table 5. Humanized antibodies and their VH and VL regions A.47C6A3

[0156]

[0157] B.67F3G7

[0158]

[0159] C.89C10H8

[0160]

[0161] Example 6: Antigen Binding Properties of Humanized Antibodies

[0162] Binding to recombinant human PD-L1

[0163] To evaluate antigen binding activity, humanized antibodies were subjected to ELISA testing. Briefly, microtiter plates were coated with 0.5 μg / ml human PD-L1-Fc protein in PBS, 100 μl / well at 4 ° C overnight, and then blocked with 200 μl / well 1% BSA. Three-fold dilutions of humanized antibodies starting at 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 mouse anti-human IgGFab antibodies conjugated to horseradish peroxidase (HRP) at 37 ° C for 1 hour. After washing, the plates were developed with TMB substrate and analyzed with a spectrophotometer at OD 450nm. As shown in Figure 6, all humanized antibodies showed binding efficiency to human PD-L1 comparable to that of chimeric antibodies.

[0164] In order to explore the binding kinetics of humanized antibodies, this example is sorted by affinity 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 show high affinity, which is comparable to chimeric antibodies.

[0165] Table 6. Affinity ranking of humanized antibodies

[0166]

[0167]

[0168] Binds to human PD-L1 overexpressed on mammalian cells

[0169] To evaluate the antigen binding properties, the binding of humanized antibodies to PD-L1 overexpressed on mammalian cells was analyzed by FACS. Briefly, PDL1-CHOK1 cells were first incubated with 3-fold serial dilutions of humanized antibodies starting at 15 μg / ml for 40 min at 4°C. After washing with PBS, Alexa Fluor 647AffiniPure goat anti-human IgG (H+L) antibody was added to each well and incubated at 4°C for 30 min. Alexa Fluor was evaluated by FACSCanto MFI of 647. Figure 7 As shown, all humanized antibodies can efficiently bind to PD-L1 expressed on mammalian cells.

[0170] Full kinetic affinity of humanized antibodies measured by Biacore

[0171] The binding of humanized antibodies to recombinant PD-L1 protein (human PD-L1-his tag) was tested by Biacore using a 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 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 Biacore T200 evaluation software, and the results are shown in Table 7 below.

[0172] Table 7 Affinity measured by Biacore

[0173]

[0174]

[0175] Example 7: Blocking of the binding of PDL1 to PD1 by humanized antibodies

[0176] Receptor blockade trial using recombinant human PD-L1

[0177] Human PD-L1 has two receptors, PD-1 and CD80. In order to explore the blocking properties of humanized PD-L1 antibodies against these two proteins, a protein-based receptor blocking assay was used here.

[0178] Briefly, microtiter plates were coated with 0.5 μg / ml human PD-L1-Fc protein in PBS, 100 μl / well at 4°C overnight, and then blocked with 150 μl / well 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 at 37°C for 1 hour. 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 450nm. As shown in FIG. 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 ).

[0179] Example 8. Bifunctional protein targeting PD-L1 and TGF-β pathways

[0180] In this example, a bifunctional recombinant anti-PD-L1 antibody and TGF-βRII fusion protein were prepared and tested.

[0181] The light chain of the molecule is the light chain of anti-PDL1 mAb. The heavy chain is the heavy chain of anti-PDL1 mAb. 4 Ser) 4 Fusion of a Gly linker to the N-terminus of the soluble extracellular domain of TGF-βRII. At the fusion junction, the C-terminal lysine residue of the antibody heavy chain was mutated to alanine to reduce potential proteolytic cleavage.

[0182] In some examples, potential modification sites in the CDRs are mutated to similar amino acids. The sequences of the anti-PD-L1 portions are shown in Table 8 below.

[0183] Table 8. Sequences of the variable regions of the antibody portion of the bifunctional molecule

[0184]

[0185] Table 9. VH / VL of bifunctional molecules

[0186] Bifunctional molecules VH V L LP008-02 02VH 02VL LP008-06 06VH 06VL LP008-06a 06a VH 06l LP008-06a-DA 06a-DA VH 06l LP008-06a-ES 06a-ES VH 06l .

[0187] In addition to VH, the heavy chain of the bifunctional molecule also includes a constant region (C-terminal K mutated to A), (Gly 4 Ser) 4 Gly linker and the N-terminus of the soluble extracellular domain of TGF-βRII. Its sequence is shown in Table 10.

[0188] Table 10. Other sequences of heavy chains, and entire heavy / light chains

[0189]

[0190]

[0191] Example 9: Binding affinity of bifunctional molecules

[0192] 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.

[0193] Bifunctional molecules were captured using a protein A chip. Serial dilutions of human PD-L1-his tagged 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. Results are shown in Fig.10 and Table 11 below.

[0194] Table 11. Affinity test by Biacore

[0195]

[0196] The binding of LP008-02 to recombinant PD-L1 protein and human TGF-β1 was tested by Biacore using a capture method.

[0197] LP008-02 was captured using a protein A chip. Serial dilutions of human PD-L1-his tagged protein and human TGF-β1 were injected over the captured antibody at a flow rate of 30 μl / min for 2 minutes. PD-L1 was allowed to dissociate for 680 s, and TGF-β1 was allowed to dissociate for 1000 s. All experiments were performed on a Biacore T200. Data analysis was performed using the Biacore T200 evaluation software. Results are shown in Fig.11 and Table 12 below.

[0198] Table 12. Affinity test by Biacore

[0199]

[0200] Example 10: Functional assay of PD-1 / PD-L1 blockade

[0201] In this example, the activity of the bifunctional molecules in blocking PD1 / PD-L1 interaction was determined by a bioluminescent cell-based assay.

[0202] 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. Addition of 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.

[0203] like Fig.12 As shown, LP008-02 and LP008-06a-ES blocked the interaction between PD1 and PD-L1 with much higher activity than M7824 (M7824 EC 50 =0.8504nM, LP008-02 EC 50 =0.3630nM, LP008-06a-ES EC 50 =0.4553nM).

[0204] Example 11: Functional assay of TGF-β

[0205] This example uses a luciferase assay to evaluate the effects of LP008-02 and LP008-06a-ES on canonical TGF-β signaling.

[0206] Serial dilutions of M7824 (a bifunctional anti-PD-L1 / TGFβ trap fusion protein, see, e.g., 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 in the presence of recombinant human TGF-β for approximately 20 hours.

[0207] like Fig.13 As shown, in the TGF-β SBE luciferase reporter assay system constructed in 293 cells, M7824, LP008-02 and LP008-06a-ES blocked TGF-β classical signaling (IC50 = 0.06687 nM, IC50 = 0.07352 nM, IC50 = 0.07167 nM).

[0208] Example 12: Binding activity to human PD-L1

[0209] ELISA for recombinant human PD-L1

[0210] To evaluate the binding activity of M7824, LP008-02, and LP008-06a-ES, ELISA tests were performed on the bifunctional molecules.

[0211] Briefly, microtiter plates were coated with 0.5 μg / ml human PD-L1-His protein in PBS, 100 μl / well at 4°C overnight, and then blocked with 150 μl / well 1% BSA. Three-fold dilutions of M7824, LP008-02, and LP008-06a-ES starting at 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 antibodies 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 450nm.

[0212] like Fig.14 As shown, LP008-02 and LP008-06a-ES bind to human PD-L1 with significantly higher activity than M7824 (EC 50 =11.82ng / ml and EC 50 =14.36ng / ml vs.EC 50 =23.68ng / ml).

[0213] Cross-species activity

[0214] To evaluate the binding of the bispecific antibodies to mouse PD-L1, rat PD-L1, and cynomolgus monkey PD-L1, the antibodies were tested using ELISA.

[0215] Briefly, microtiter plates were coated with 0.5 μg / ml mouse, rat and cynomolgus PD-L1 proteins in PBS, 100 μl / well at 4°C overnight, and then blocked with 150 μl / well 1% BSA. Three-fold dilutions of bispecific antibodies starting at 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 antibodies 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 450nm.

[0216] 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 ( Fig.15 and Table 13).

[0217] Table 13. Cross-species activity of M7824, CZ010-02 and CZ010-06a-ES

[0218] EC50 Cynomolgus monkey Rat Mouse M7824 25.60ng / ml 14.25ng / ml 10.60ng / ml LP008-02 7.890ng / ml No combination No combination LP008-06a-ES 10.92ng / ml No combination No combination .

[0219] Example 13: Binding activity to human TGF-β

[0220] ELISA using recombinant human TGF-β

[0221] To evaluate the binding activity of M7824, LP008-02, and LP008-06a-ES to human TGF-β, ELISA tests were performed on these bifunctional molecules.

[0222] 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 150 μl / well 1% BSA. Three-fold dilutions of M7824, LP008-02, and LP008-06a-ES bifunctional molecules starting at 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 antibodies 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 450nm.

[0223] like Fig.16As shown, M7824, LP008-02 and LP008-06a-ES all bind to human TGF-β (EC 50 =43.43ng / ml, EC 50 =28.58ng / ml, EC 50 =39.38ng / ml).

[0224] Cross-species activity

[0225] To evaluate the binding of the bispecific antibodies to mouse, rat, and cynomolgus monkey TGF-β, ELISA tests were performed on the bifunctional molecules.

[0226] Briefly, microtiter plates were coated with mouse, rat and cynomolgus monkey TGF-β proteins at 1 μg / ml in PBS, 100 μl / well at 4°C overnight, and then blocked with 150 μl / well 1% BSA. Three-fold dilutions of bispecific antibodies starting at 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 antibodies 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 450nm.

[0227] All tested bifunctional molecules bound cynomolgus monkey, rat and mouse TGF-β with high activity ( Fig.17 and Table 14).

[0228] Table 14. Cross-species activity of M7824, LP008-02 and LP008-06a-ES

[0229] EC50 Cynomolgus monkey Rat Mouse M7824 39.66ng / ml 123.9ng / ml 46.17ng / ml LP008-02 34.41ng / ml 77.84ng / ml 34.70ng / ml LP008-06a-ES 58.35ng / ml 120.7ng / ml 58.26ng / ml .

[0230] Example 14: Efficacy in MC38 tumor mouse model

[0231] This example uses a tumor mouse model to test the in vivo efficacy of the bifunctional molecule.

[0232] MC38 cells expressing human PD-L1 were resuspended in PBS at a volume of 0.2 mL at 5 x 10 5 The concentration of cells was subcutaneously inoculated into the right skin of B-hPD-L1 humanized mice. When the average tumor volume reached about 55 mm 3When 24 mice with appropriate individual tumor volumes were selected as groups, the animals were randomly divided into 4 experimental groups of 6 mice each according to the tumor volume. Total human IgG, M7824, LP008-02, and LP008-06a-ES were administered 3 times a week by intraperitoneal injection after injection of anti-mCD20 mAb. The dose was calculated at 10 μg / g according to the body weight of the experimental animals. The body weight and tumor size of the mice were tested twice a week.

[0233] 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.

[0234] Example 15. Modification of bifunctional molecules

[0235] This example tests the in vitro efficacy of certain modified bifunctional molecules (Table 15) on TGF-β functional assays. Some of them include repeating linker sequences of TAGHTQTSTGGGAITTGTSGAGHGP (SEQ ID NO: 87), HYP and / or G4S (SEQ ID NO: 86). These molecules are referred to as LP008-02-1 to LP008-02-7, respectively.

[0236] Table 15. Modified sequence design of linker and TGF-βRII

[0237]

[0239] ELISA for recombinant human TGF-β1

[0240] To evaluate the binding activity of the modified LP008-02 bifunctional molecules, these bifunctional molecules were tested using ELISA.

[0241] 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 1% BSA. Three-fold dilutions of the modified LP008-02 bifunctional molecule 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 with 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 450nm.

[0242] like Fig.19 As shown, all modified LP008-02 bifunctional molecules bound to human TGF-β1 with high activity, comparable to LP008-02-1.

[0243] TGF-β functional assay

[0244] Serial dilutions of the modified LP008-02 bifunctional molecule were incubated with 293 cells transfected with the SBE luciferase reporter in the presence of recombinant human TGF-β1 for approximately 22 hours.

[0245] like Fig. 20 As 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 classical signaling (IC50 = 0.1435 nM, IC50 = 0.1639 nM, IC50 = 0.1882 nM), compared with LP008-02-1.

[0246] Example 16. Comparison of bifunctional molecules

[0247] Molecules 1-7 of 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.

[0248] 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 isoform B, SEQ ID NO: 71) and a C-terminal fragment (EEYNTSNPD, SEQ ID NO: 90). In addition, this molecule has several G4S (SEQ ID NO: 86) repeats added to the linker.

[0249] Compared to molecule 1, molecule 2 (LP008-02-2) replaces 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). The linker is modeled based on SEQ ID NO: 89. Changes include: (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 is retained). These changes are shown in Table 16 below. Molecule 2 also contains a single G4S unit at the N-terminus.

[0250] Table 16. Artificial connectors

[0251] name sequence SEQ ID NO: original IPPHVQKSVNNDMIVTDNNGAVKFP 89 Modified TAGHTQTSTGGGAITTGTSGAGHGP 87 .

[0252] 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 a 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.

[0253] Example 17. Binding activity and stability of bifunctional molecules

[0254] This example uses SEC-HPLC and CE-SDS 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.

[0255] The five sequences were expressed in CHO-K1 cells by polyethyleneimine (PEI)-mediated transient transfection and the supernatant was harvested after 10 days. The bifunctional molecules were purified from the culture supernatant by protein A and then by Superdex 200 pg, with a purity level greater than 99% as detected by SEC-HPLC (Table 17).

[0256] Table 17. SEC-HPLC and CE-SDS results of test articles on day 0

[0257]

[0258]

[0259] To evaluate the binding activity of the modified LP008-02 bifunctional molecules, these bifunctional molecules were tested by ELISA.

[0260] Briefly, microtiter plates were coated with 1 μg / ml human TGF-β1 protein (Acro, TG1-H4212) in PBS, 100 μl / well at 4°C overnight, and then blocked with 150 μl / well 1% BSA. Four-fold serial dilutions of the modified LP008-02 bifunctional molecule 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 with 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.

[0261] like Fig.21 As shown, all other modified LP008-02 bifunctional molecules bound to human TGF-β1 with high activity, comparable to LP008-02-1.

[0262] To evaluate the effect of the modified LP008-02 bifunctional molecule on canonical TGF-β signaling, the modified bifunctional molecule was tested using a luciferase assay. Serial dilutions of the bifunctional molecule were incubated with 293 cells transfected with the SBE luciferase reporter for 24 h in the presence of recombinant human TGF-β. Fig. 22 As 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 TGF-β classical signaling (IC50 = 0.04231nM, IC50 = 0.0527nM, IC50 = 0.09616nM, and IC50 = 0.1962nM).

[0263] The bifunctional molecules were dissolved in two buffers for antibody stability testing. The 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.

[0264] The prepared samples at 3.0 mg / ml were incubated at 40°C and then tested 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 were more stable than LP008-02-1 in SEC-HPLC, non-reduced CE-SDS, and reduced CE-SDS.

[0265] Table 18. SEC-HPLC and CE-SDS results of test articles on day 14

[0266]

[0267]

[0268] Thus, this example shows that the modified bifunctional molecules LP008-02-2, LP008-02-3, LP008.02-6 and LP008-02-7 exhibit similar activity to LP008-02-1, but their stability is significantly higher than LP008-021. Replacing the N-terminal portion of TGF-βRII in LP008-02-1 (IPPHVQKSVNNDMIVTDNNGAVKFP, SEQ ID NO: 89) with an artificial linker (e.g., TAGHTQTSTGGGAITTGTSGAGHGP (SEQ ID NO: 87) or HYP) results in significantly improved stability.

[0269] Example 18. High concentration formulation of anti-PD-L1 antibody

[0270] This example uses HIC-HPLC and viscosity testing to evaluate the development potential and risk of high-concentration anti-PD-L1 molecule formulations.

[0271] Four anti-PD-L1 molecules were expressed in CHO-K1 or 293F cells by transient transfection. The constant region of the heavy chain is human IgG1 (N297A) -Fc. 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 0.41M, 0.78M, 0.97M and 1.10M, respectively. All newly developed antibodies have lower hydrophobicity than the reference antibody MPDL3280A.

[0272] Table 19. Antibody hydrophobicity from HIC-HPLC testing

[0273] sample HIC(M) MPDL3280A 0.41 47C6A3 0.78 Hu67F3G7-22 0.97 Hu89C10H8-7 1.10 .

[0274] The activity of anti-PD-L1 antibodies in blocking PD1 / PD-L1 interaction was then measured using a bioluminescent 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. Addition of anti-PD-1 or anti-PD-L1 antibodies that block PD-1 / PD-L1 interaction will release the inhibitory signal and lead to TCR activation and NFAT-RE-mediated luminescence. Fig.23 As shown, MPDL3280A, 47C6A3, Hu67F3G7-22, and Hu89C10H8-7 blocked the interaction between PD1 and PD-L1 with considerable activity (MPDL3280AEC 50 =0.1327nM, 47C6A3 EC 50 =0.1501nM, Hu67F3G7-22 EC 50 =0.1034nM, Hu89C10H8-7 EC 50 =0.2138nM).

[0275] MPDL3280A and Hu67F3G7-22 with human IgG1 Fc were expressed in CHO-K1 cells by transient transfection. Purified MPDL3280A-hIgG1 Fc and Hu67F3G7-22-hIgG1 Fc 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 the two molecules. As shown in Table 20, the ammonium sulfate concentration corresponding to the hydrophobic elution time of MPDL3280A-hIgG1 Fc and Hu67F3G7-22-hIgG1 Fc was 0.42M and 0.99M, respectively. Similarly, for the same Fc fragment, Hu67F3G7-22 exhibits a lower hydrophobicity than MPDL3280A.

[0276] Table 20. Results of HIC-HPLC test

[0277]

[0278] In order to further confirm the solubility and viscosity characteristics of the antibody, two purified candidates were directly concentrated in phosphate buffer (including 60mM NaCl) by ultrafiltration. During ultrafiltration, 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 is much higher than that of Hu 67F3G7-22-hIgG1 Fc. For high concentration preparations, antibodies with lower viscosity are generally superior to antibodies with higher viscosity. Therefore, as a therapeutic protein, Hu67F3G7-22 antibody has a higher potential than MPDL3280A.

[0279] Table 21. Results of solubility tests

[0280]

[0281]

[0282] ***

[0283] The scope of the present disclosure is not limited by the specific embodiments described, which are intended to be a single illustration of the 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 clear to those skilled in the art that various modifications and changes may be made to the methods and compositions of the present disclosure without departing from the spirit or scope of the present disclosure. Therefore, the present disclosure is intended to cover modifications and changes of the present disclosure as long as they are within the scope of the appended claims and their equivalents.

[0284] All publications and patent applications mentioned in this specification are herein incorporated 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. An anti-PD-L1 (programmed death ligand 1) antibody or a fragment thereof, which is specific to human PD-L1 protein and comprises a heavy chain variable region (VH) comprising VH CDR1, VH CDR2 and VH CDR3, and a light chain variable region (VL) comprising VL CDR1, VL CDR2 and VLCDR3, wherein the VH CDR1, VH CDR2, VH CDR3, VL CDR1, VL CDR2 and VL CDR3 comprise the amino acid sequences of SEQ ID NOs: 13-18, respectively.

2. The anti-PD-L1 antibody or fragment thereof according to claim 1, wherein the VH comprises an amino acid sequence selected from the group consisting of SEQ ID NOs: 31-37, and the VL comprises an amino acid sequence selected from the group consisting of SEQ ID NOs: 38-43.

3. The anti-PD-L1 antibody or fragment thereof according to claim 2, wherein the VH comprises the amino acid sequence of SEQ ID NO: 34, and the VL comprises the amino acid sequence of SEQ ID NO:

43.

4. A cell comprising one or more polynucleotides encoding the anti-PD-L1 antibody or fragment thereof according to any one of claims 1 to 3.

5. One or more polynucleotides encoding the anti-PD-L1 antibody or fragment thereof according to any one of claims 1 to 3.

6. A composition comprising the anti-PD-L1 antibody or fragment thereof according to any one of claims 1 to 3 and a pharmaceutically suitable carrier.

7. Use of the anti-PD-L1 antibody or fragment thereof according to any one of claims 1 to 3 for preparing a medicament for treating cancer.

8. The use according to claim 7, wherein the cancer is a solid tumor.

9. The method of claim 7 or 8, wherein the cancer is selected from bladder cancer, liver cancer, colon cancer, rectal cancer, endometrial cancer, leukemia, lymphoma, pancreatic cancer, small cell lung cancer, non-small cell lung cancer, breast cancer, urethral cancer, head and neck cancer, gastrointestinal cancer, gastric cancer, esophageal cancer, ovarian cancer, kidney cancer, melanoma, prostate cancer and thyroid cancer.

Images