Multispecific antibody protein and uses thereof
Multispecific antibody fusion proteins with tailored amino acid sequences and linkers, including mutated SIRP IgV domains, address the challenges of assembling PD-L1/PD-1, VEGF, and CD47 targets, enhancing therapeutic efficacy by simultaneous antigen binding and overcoming CD47 expression issues.
Patent Information
- Authority / Receiving Office
- WO · WO
- Patent Type
- Applications
- Current Assignee / Owner
- Filing Date
- 2025-09-24
- Publication Date
- 2026-04-02
AI Technical Summary
Developing multispecific antibody fusion proteins targeting PD-L1/PD-1, VEGF, and CD47 is challenging due to the complexity of assembling multiple monospecific binding components, stability issues, and the ubiquitous expression of CD47 across normal tissues, which complicates therapeutic development and safety.
The construction of multispecific antibody fusion proteins with specific amino acid sequences and linkers, including mutated SIRP IgV domains for enhanced CD47 binding, particularly at acidic tumor pH, to facilitate simultaneous binding to PD-L1/PD-1, VEGF, and CD47.
The multispecific antibody fusion proteins demonstrate improved biophysical properties and antigen binding activities, potentially offering enhanced therapeutic effects against cancer by blocking multiple immune checkpoint pathways.
Smart Images

Figure PCTCN2025123644-FTAPPB-I100001 
Figure PCTCN2025123644-FTAPPB-I100002 
Figure PCTCN2025123644-FTAPPB-I100003
Abstract
Description
Multispecific antibody protein and uses thereofCROSS REFERENCE TO RELATED APPLICATIONS
[0001] This application claims the benefit of, and priority to, PCT Application No. PCT / CN2024 / 120935, filed on September 25th, 2024. The content of these applications is incorporated herein by reference in their entirety for all purposes.TECHNICAL FIELD
[0002] The present disclosure generally relates to multispecific antibody protein and the uses thereof.BACKGROUNDPD-1 and PD-L1
[0003] Programmed cell death 1 (PD-1) and its ligand Programmed death-ligand 1 (PD-L1) are identified as a key pair of immune checkpoint that negatively regulates immune response. PD-1 is a cell surface receptor and inhibitor of both adaptive and innate immune responses, and is expressed on T cells. Particularly PD-1 is highly expressed on T cells in tumors. PD-L1 is often expressed on cell surface of tumor cells, and the binding of tumor cell PD-L1 to PD-1 on T cells activates an inhibitory signal, resulting in decreased T-cell activity and anti-tumor immunity (Parvez et al., Front Immunol, 2023) . Blocking the interaction of PD-1 and PD-L1 by either anti-PD-1 antibody or anti-PD-L1 antibody has yielded notable therapeutic results in diverse tumor types. VEGF and VEGFR
[0004] VEGF (vascular endothelial growth factor) are dimeric glycoproteins consisting of five distinct isoforms: VEGF-A, VEGF-B, VEGF-C, VEGF-D, and placenta growth factor (PlGF) . VEGF plays pivotal roles in regulating tumor angiogenesis (formation of new blood vessels) and physiological vascular function (Apte et al., Cell, 2019) . VEGF is also shown to excert immune-suppressive activity in tumor microenvironment (TME) through a direct effect on multiple immune cell types involved in tumor immunity (Apte et al., Cell, 2019) . Bevacizumab (Avastin) , an FDA approved anti-VEGF monoclonal antibody drug, has shown therapeutic effects in multiple cancer types by inhibiting the biological activities of VEGF. Another protein drug targeting VEGF, Aflibercept, was approved in United State and Europe for treatment of wet macular degeneration under tradename Eylea, and for metastatic colorectal cancer as Zaltrap.
[0005] VEGF receptors (VEGFR) are cognitive receptors for VEGF and have three subtypes: VEGFR-1, VEGFR-2 and VEGFR-3, all having an extracellular portion consisting of 7 immunoglobulin (Ig) -like domains, a single transmembrane region and an intracellular portion. Particularly, Ig-like domains 2 in VEGFR-1 and Ig-like domains 3 in VEGFR-2 are associated with ligand-binding site and ligand-binding specificity, respectively. The binding of VEGF-A to VEGFR-2 is known to account for the majority of the angiogenic stimulatory signal observed in vivo (Apte et al., Cell, 2019) . Ramucirumab (Cyramza) , an FDA approved anti-VEGFR2 monoclonal antibody drug, has shown therapeutic effects in several cancer types by inhibiting the biological activities of VEGFR2. CD47
[0006] The protein CD47 (cluster of differentiation 47) functions as a major anti-phagocytic signal that inhibits phagocytosis of the cells expressing CD47 through interacting with the Signal Regulatory Protein (SIRP) receptor, most notably SIRPα, on phagocytes such as macrophages and triggering an "anti-phagocytic" signal. CD47 is ubiquitously expressed in normal tissues and cells, and plays an important role in sparing the normal cells from phagocytosis. Disease cells such as cancer cells, often with upregulation of CD47 expression, however hijack this mechanism to escape normal immune control and clearance by phagocytes. CD47 has been shown to be highly expressed and associated with adverse prognosis in a wide variety of cancers (Zhang et al., Front Immunol, 2020) . Blocking the CD47 “anti-phagocytic” signal by CD47 binders, together with co-presence of sufficient “pro-phagocytic” signal, is shown to promote phagocytosis of a variety of CD47-expressing cancer cells, and presents an attractive therapeutic strategy to treat a broad range of cancers.
[0007] However, development of anti-CD47 therapeutics is hampered by the ubiquitous expression of CD47 across normal tissues and cells, which present safety risk to CD47-expressing normal tissues and also a huge antigen sink sequestering the anti-CD47 therapeutics from diseased tissue. Multispecific antibody fusion protein against PD-L1 / PD-1 and VEGF, or against PD-L1 / PD-1, VEGF and CD47
[0008] While monospecific antibody and fusion protein therapeutics against PD-L1 / PD-1 or VEGF / VEGFR have showed notable therapeutic effects in multiple cancer types, blocking of both the PD-L1 / PD-1 signaling and VEGF / VEGFR signaling in tumor through either combination therapy or bispecific therapeutics against both PD-L1 / PD-1 and VEGF / VEGFR are pursued for further improving therapeutic efficacy. There are now many monospecific antibodies against PD-L1 / PD-1 developed and also many monospecific antibodies or fusion protein against VEGF / VEGFR developed. However, developing bispecific antibody fusion protein against both PD-L1 / PD-1 and VEGF by assembling monospecific anti-PD-L1 / PD-1 domain and monospecific anti-VEGF / VEGFR domain into a bispedific antibody fusion protein presents significant challenges. There are a vast number of variables to explore in order to potentially identify which specific anti-PD-L1 / PD-1 domain among the many monospecific anti-PD-L1 / PD-1 domains is compatible with a specific anti-VEGF / VEGFR domain among the many monospecific anti-VEGF / VEGFR domains for physically assembling together into an antibody fusion protein, and vice versa. Additionally, there are a large number of possible structural formats to explore to assemble two monospecific protein components, in order to identify which structural format is feasible or optimal for biophysical and functional properties of the bispecific antibody fusion protein. Also assembling two monospecific antibody into a bispecific antibody fusion protein often requires conversion of the Fab domain of one monospecific antibody into a single-chain variable fragment (scFv) , which is well recognized for being prone to aggregation and often requires optimization of a number of variables such as orientation (VH-linker-VL or VL-linker-VH) as well as linker composition and length for proper stability and affinity as well as for compatibility in linking to an antibody protein. The producibility, yield, purity, stability, affinity and other biophysical and pharmaceutical properties of an assembled bispecific protein is also difficult to predict, even ifthe properties of each of the monospecific components are known. Even if an assembled bispecific protein shows good biophysical properties, it’s unknown whether it could indeed simultaneously bind to both antigens.
[0009] Further blocking CD47 / SIRP signaling along with blocking of both PD-L1 / PD-1 and VEGF / VEGFR signaling through trispecific antibody fusion protein against PD-L1 / PD-1, VEGF and CD47 may potentially show further enhanced anti-cancer therapeutic effects. However, developing trispecific antibody fusion protein presents even bigger challenges than developing bispecific antibody fusion protein. The magnitude of the number of possible assemblies of 3 monospecific binding components is much higher than that of assembling 2 monospecific binding components. Also, it’s more challenging to enable the 3 monospecific binding components in a trispecific antibody fusion protein to simultanesouly bind all the 3 target antigens than to simultaneously bind 2 target antigens in a bispecific antibody fusion protein. What’s more, the ubiquitous expression of CD47 across normal tissues and cells, presenting both broad safety risk to CD47-expressing normal tissues and a huge antigen sink, further complicates the development trispecific antibody fusion protein against CD47.
[0010] Through extensive and creative efforts, the present invention constructed and tested a variety of bispecific antibody fusion protein against PD-L1 / PD-1 and VEGF as well as trispecfic antibody fusion against PD-L1 / PD-1, VEGF and CD47 using various monospecific binding components and structural formats, and identified specific bispecific antibody fusion protein against PD-L1 / PD-1 and VEGF as well as trispecfic antibody fusion against PD-L1 / PD-1, VEGF and CD47 with good and / or surprising properties and activities. Furthermore for trispecific antibody fusion protein against PD-L1 / PD-1, VEGF and CD47 using SIRP IgV domain as the CD47-binding component, inventive mutation was further introduced into the wild type SIRP IgV domain to enhance CD47-binding and / or enable acidic pH-sensitive binding to facilitate preferential binding of CD47 in tumors with acidic pH in order to address the issue of ubiquitous expression CD47 across normal tissues with physiological pH.SUMMARY
[0011] In one aspect, the present disclosure provides multispecific antibody fusion protein against PD-L1 / PD-1 and VEGF. In another aspect, the present disclosure provides multispecific antibody fusion protein against PD-L1 / PD-1, VEGF and additionally CD47.
[0012] In one aspect, a multispecific molecule comprising: 1) a backbone antibody that binds to PD-L1 or PD-1; and 2) a single-chain domain that binds to vascular endothelial growth factor (VEGF) , wherein the VEGF-binding domain comprises an amino acid sequence selected from the group consisting of SEQ ID NO: 17 to 30; and 3) the anti-VEGF single-chain domain is linked through a linker to the N-terminal or C-terminal of the heavy chain or light chain of the anti-PD-L1 / PD-1 backbone antibody. In some embodiments, the anti-PD-L1 / PD-1 backbone antibody of the multispecific molecule comprises a VH amino acid sequence of SEQ ID NO: 32 and a VL amino acid sequence of SEQ ID NO: 33, a VH sequence of SEQ ID NO: 34 and a VL sequence of SEQ ID NO: 35, aVH sequence of SEQ ID NO: 36 and a VL sequence of SEQ ID NO: 37, a VH sequence of SEQ ID NO: 38 and a VL sequence of SEQ ID NO: 39, a VH sequence of SEQ ID NO: 40 and a VL sequence of SEQ ID NO: 41, or a VH sequence of SEQ ID NO: 42 and a VL sequence of SEQ ID NO: 43. In some embodiment, the linker linking the anti-VEGF single-chain domain to the backbone anti-PD-L1 / PD-1 antibody of the multispecific molecule comprises an amino acid sequence selected from the group consisting of SEQ ID NO: 9-14. In one embodiment, the linker linking the anti-VEGF single-chain domain to the backbone anti-PD-L1 / PD-1 antibody comprises preferably an amino acid sequence of SEQ ID NO: 13 when the anti-VEGF single-chain domain is linked to the C-terminal of the light chain of the backbone anti-PD-L1 / PD-1 antibody.
[0013] In one aspect, a multispecific molecule comprising: 1) a backbone antibody that binds to PD-L1 or PD-1; and 2) a single-chain domain that binds to vascular endothelial growth factor (VEGF) , wherein the VEGF-binding domain comprises an amino acid sequence selected from the group consisting of SEQ ID NO: 17 to 30 and 451 to 453; and 3) the anti-VEGF single-chain domain is linked through a linker to the N-terminal or C-terminal of the heavy chain or light chain of the anti-PD-L1 / PD-1 backbone antibody. In some embodiments, the anti-PD-L1 / PD-1 backbone antibody of the multispecific molecule comprises a VH amino acid sequence of SEQ ID NO: 32 and a VL amino acid sequence of SEQ ID NO: 33, a VH sequence of SEQ ID NO: 34 and a VL sequence of SEQ ID NO: 35, a VH sequence of SEQ ID NO: 36 and a VL sequence of SEQ ID NO: 37, a VH sequence of SEQ ID NO: 38 and a VL sequence of SEQ ID NO: 39, a VH sequence of SEQ ID NO: 40 and a VL sequence of SEQ ID NO: 41, or a VH sequence of SEQ ID NO: 42 and a VL sequence of SEQ ID NO: 43. In some embodiment, the linker linking the anti-VEGF single-chain domain to the backbone anti-PD-L1 / PD-1 antibody of the multispecific molecule comprises an amino acid sequence selected from the group consisting of SEQ ID NO: 9-14, 472-477 and 478-480, wherein SEQ ID NO: 478 comprises GGG, SEQ ID NO: 479 comprises GG and SEQ ID NO: 480 comprises G. In one embodiment, the linker linking the anti-VEGF single-chain domain to the backbone anti-PD-L1 / PD-1 antibody comprises preferably an amino acid sequence of SEQ ID NO: 13 when the anti-VEGF single-chain domain is linked to the C-terminal of the light chain of the backbone anti-PD-L1 / PD-1 antibody.
[0014] In some embodiment, the anti-PD-L1 backbone antibody of the multispecific molecule comprises a VH of SEQ ID NO: 32 and a VL amino acid sequence of SEQ ID NO: 33, and the anti-VEGF single-chain domain comprises an amino acid sequence of SEQ ID NO: 17, 18, or 451, wherein the anti-VEGF single-chain domain is linked through a linker preferably to the C-terminal of the heavy chain of the anti-PD-L1 antibody, wherein the resultant antibody fusion protein exhibits better biophysical property and / or antigen binding activity than that when the anti-VEGF single-chain domain is linked through a linker to the N-terminal of the heavy chain of the anti-PD-L1 antibody. In some embodiment, the anti-PD-L1 backbone antibody of the multispecific molecule comprises a VH of SEQ ID NO: 34 and a VL amino acid sequence of SEQ ID NO: 35, and the anti-VEGF single-chain domain comprises an amino acid sequence of SEQ ID NO: 17, 18, or 451, wherein the anti-VEGF single-chain domain is linked through a linker preferably to the C-terminal of the heavy chain of the anti-PD-L1 antibody, wherein the resultant antibody fusion protein exhibits better biophysical property and / or antigen binding activity than that when the anti-VEGF single-chain domain is linked through a linker to the N-terminal of the heavy chain of the anti-PD-L1 antibody. In a further embodiment, the anti-VEGF single-chain domain of the two foreging embodiments comprises an amino acid sequence of SEQ ID NO: 17. In further embodiment, the linker in the foregoing embodiments linking the anti-VEGF single chain domain to C-terminal of the heavy chain of the anti-PD-L1 antibody comprises 15 amino acids or less, wherein a shorter linker maintains the biophysical property and antigen binding acitivty similarly as a longer linker and wherein a shorter linker favorably increase the formation of crosslinking multimer complex between the antibody fusion protein and dimeric VEGF antigen than a longer linker. In certain embodiment, the linker comprises 14 amino acids or less. In certain embodiment, the linker comprises preferably 10 amino acids or less. In certain embodiment, the linker comprises 9 amino acids or less. In certain embodiment, the linker comprises 8 amino acids or less. In certain embodiment, the linker comprises 8 amino acids. In certain embodiment, the linker comprises 7 amino acids or less. In certain embodiment, the linker comprises 6 amino acids or less. In certain embodiment, the linker comprises 5 amino acids or less. In certain embodiment, the linker comprises 5 amino acids. In certain embodiment, the linker comprises an amino acid sequence selected from the group consisting of SEQ ID NO: 9-10 and 472-480. In certain embodiment, the linker comprises an amino acid sequence of SEQ ID NO: 473 or 476.
[0015] In some embodiment, the anti-VEGF single-chain domain comprises an amino acid sequence of SEQ ID NO: 25 or 26, and the anti-VEGF single-chain domain is linked through a linker preferably to the C-terminal of the heavy chain of the anti-PD-L1 / PD-1 antibody.
[0016] In some embodiment, the anti-VEGF single-chain domain comprises an amino acid sequence of SEQ ID NO: 27, 28, 29 or 30, the anti-PD-L1 / PD-1 backbone antibody comprises a VH amino acid sequence of SEQ ID NO: 32 and a VL amino acid sequence of SEQ ID NO: 33. In some embodiment, the anti-VEGF single-chain domain comprises an amino acid sequence of SEQ ID NO: 27, 28, 29 or 30, and the anti-PD-L1 / PD-1 backbone antibody comprises a VH amino acid sequence of SEQ ID NO: 34 and a VL amino acid sequence of SEQ ID NO: 35. In some embodiment, the anti-VEGF single-chain domain comprises an amino acid sequence of SEQ ID NO: 27, 28, 29 or 30, and the anti-PD-L1 / PD-1 backbone antibody comprises a VH amino acid sequence of SEQ ID NO: 38 and a VL amino acid sequence of SEQ ID NO: 39. In some embodiment, the anti-VEGF single-chain domain comprises an amino acid sequence of SEQ ID NO: 27, 28, 29 or 30, and the anti-PD-L1 / PD-1 backbone antibody comprises a VH amino acid sequence of SEQ ID NO: 40 and a VL amino acid sequence of SEQ ID NO: 41. In some embodiment, the anti-VEGF single-chain domain comprises an amino acid sequence of SEQ ID NO: 27, 28, 29 or 30, and the anti-PD-L1 / PD-1 backbone antibody comprises a VH amino acid sequence of SEQ ID NO: 42 and a VL amino acid sequence of SEQ ID NO: 43.
[0017] In another aspect, a multispecific molecule comprising: 1) a backbone antibody that binds to VEGF; 2) asingle-chain domain that binds to PD-L1 or PD-1; and 3) the anti-PD-L1 / PD-1 single-chain domain is linked through a linker to the N-terminal or C-terminal of the heavy chain or light chain of the anti-VEGF backbone antibody. In some embodiment, the anti-VEGF backbone antibody comprises a VH sequence of SEQ ID NO: 15 and a VL sequence of SEQ ID NO: 16, and the anti-PD-L1 / PD-1 single-chain domain comprises an amino acid sequence selected from the group consisting of SEQ ID NO: 46 to 54. In some embodiment, the anti-VEGF backbone antibody comprises a VH sequence of SEQ ID NO: 15 and a VL sequence of SEQ ID NO: 16, and the anti-PD-L1 / PD-1 single-chain domain comprises an amino acid sequence selected from the group consisting of SEQ ID NO: 46 to 54 and 455. In some embodiment, the anti-PD-L1 / PD-1 single-chain domain comprises an amino acid sequence of SEQ ID NO: 455 and is preferably linked to the C-terminal of the heavy chain of the anti-VEGF backbone antibody, wherein the resultant antibody fusion protein exhibits better biophysical property and / or antigen binding activity than that when the anti-PD-L1 single-chain domain is linked through a linker to the N-terminal of the heavy chain of the anti-VEGF antibody. In one embodiment, the anti-PD-L1 / PD-1 single-chain domain comprises an amino acid sequence of SEQ ID NO: 46. In one embodiment, the anti-PD-L1 / PD-1 single-chain domain comprises an amino acid sequence of SEQ ID NO: 47. In one embodiment, the anti-PD-L1 / PD-1 single-chain domain comprises an amino acid sequence of SEQ ID NO: 50. In one embodiment, the anti-PD-L1 / PD-1 single-chain domain comprises an amino acid sequence of SEQ ID NO: 51. In one embodiment, the anti-PD-L1 / PD-1 single-chain domain comprises an amino acid sequence of SEQ ID NO: 52. In one embodiment, the anti-PD-L1 / PD-1 single-chain domain comprises an amino acid sequence of SEQ ID NO: 53.In some embodiment, the linker linking the anti-PD-L1 / PD-1 single-chain domain to the backbone anti-VEGF antibody of the multispecific molecule comprises an amino acid sequence selected from the group consisting of SEQ ID NO: 9-14. In some embodiment, the linker linking the anti-PD-L1 / PD-1 single-chain domain to the backbone anti-VEGF antibody of the multispecific molecule comprises an amino acid sequence selected from the group consisting of SEQ ID NO: 9-14 and 472-480. In one embodiment, the linker comprises amino acid sequence of SEQ ID NO: 11.
[0018] In another aspect, the multispecific molecule comprising a backbone anti-PD-L1 / PD-1 antibody and anti-VEGF single-chain domain, further comprises a third domain that binds to CD47, wherein the CD47-binding domain is linked through a linker to the backbone anti-PD-L1 / PD-1 antibody or the anti-VEGF single-chain domain. In some embodiment, the CD47-binding domain comprises a SIRP IgV monomer comprising an amino acid sequence selected from the group consisting of SEQ ID NO: 57 to 75. In some embodiment, the CD47-binding domain comprises a SIRP IgV multimer polypeptide comprising two SIRP IgV monomers linked in tandem through a linker of SEQ ID NO: 11 wherein each of the SIRP IgV monomers comprises an amino acid sequence selected from the group consisting of SEQ ID NO: 57 to 75. In one embodiment, the CD47-binding SIRP IgV monomer comprises an amino acid sequence of SEQ ID NO: 58. In one embodiment, the CD47-binding SIRP IgV monomer comprises an amino acid sequence of SEQ ID NO: 57. In one embodiment, the CD47-binding SIRP IgV monomer comprises an amino acid sequence of SEQ ID NO: 72.
[0019] In some embodiment, the CD47-binding domain comprises a SIRPαor SIRPβIgV monomer comprising a mutation of I31Y or I31W, wherein the SIRP IgV variant monomer comprising the mutation exhibits enhanced binding to CD47 than the parent SIRP IgV monomer without the mutation. In some embodiment, the CD47-binding domain comprises a SIRPγmonomer comprising a mutation of L31Y or L31W, wherein the SIRP IgV variant monomer comprising the mutation exhibits enhanced binding to CD47 than the parent SIRP IgV monomer without the mutation. In some embodiment, the CD47-binding SIRP IgV monomer comprises a mutation of R59H, wherein the SIRP IgV variant monomer comprising the mutation exhibits enhanced binding to CD47 than the parent SIRP IgV monomer without the mutation. In some embodiment, the CD47-binding SIRP IgV monomer comprises a mutation of R59H, R59H+I31Y, R59H+I31W, R59H+L31Y or R59H+L31W, wherein the SIRP IgV variant monomer comprising the mutation exhibits enhanced binding to CD47 than the parent SIRP IgV monomer without the mutation. In some embodiment, the backbone anti-PD-L1 / PD-1 antibody of the multispecific molecule comprising the CD47-binding domain comprises a Fc of human IgG4 (SEQ ID NO: 1) , IgG2 (SEQ ID NO: 2) or silent IgG1 (SEQ ID NO: 4, 5, 6 or 7) . In a preferered embodiment, the backbone anti-PD-L1 / PD-1 antibody comprises a Fc of human IgG4 (SEQ ID NO: 1) .
[0020] In some embodiment, the CD47-binding domain comprises a SIRPαor SIRPβIgV monomer comprising a mutation selected from the group consisting of K53H, R69H, K68H, Q52H, I31E, I31D, Q52H+K68H, K53H+K68H, R69H+K68H, K53H+Q52H, R69H+Q52H, K53H+I31E, R69H+I31E, K68H+I31E, Q52H+I31E, K53H+K68H+Q52H, R69H+K68H+Q52H, Q52H+K68H+I31E, K53H+I31D, R69H+I31D, K68H+I31D, Q52H+I31D, and Q52H+K68H+I31D, wherein the SIRP IgV variant monomer comprising the mutation exhibits higher binding to CD47 at an acidic pH than at physiological pH, wherein the acidic pH is<7.0 and the physiological pH is between 7.2 and 7.5. In some embodiment, the CD47-binding domain comprises a SIRPγIgV monomer comprising a mutation selected from the group consisting of K53H, R69H, K68H, Q52H, L31E, L31D, Q52H+K68H, K53H+K68H, R69H+K68H, K53H+Q52H, R69H+Q52H, K53H+L31E, R69H+L31E, K68H+L31E, Q52H+L31E, K53H+K68H+Q52H, R69H+K68H+Q52H, Q52H+K68H+L31E, K53H+L31D, R69H+L31D, K68H+L31D, Q52H+L31D, Q52H+K68H+L31D, K53H+N101D, K53H+N101D+L37Q, K53H+N101D+L31E and K53H+N101D+L31D, wherein the SIRP IgV variant monomer comprising the mutation exhibits higher binding to CD47 at an acidic pH than at physiological pH, wherein the acidic pH is<7.0 and the physiological pH is between 7.2 and 7.5. In one embodiment, the CD47-binding domain comprises a SIRP IgV monomer comprising a mutation of K53H. In one embodiment, the CD47-binding domain comprises a SIRP IgV monomer comprising a mutation of R69H. In one embodiment, the CD47-binding domain comprises a SIRP IgV monomer comprising a mutation of Q52H. In one embodiment, the CD47-binding domain comprises a SIRP IgV monomer comprising a mutation of K68H. In one embodiment, the CD47-binding domain comprises a SIRP IgV monomer comprising a mutation of Q52H+K68H. In one embodiment, the CD47-binding domain comprises a SIRP IgV monomer comprising a mutation of K53H+I31E or K53H+L31E. In one embodiment, the CD47-binding domain comprises a SIRP IgV monomer comprising a mutation of R69H+I31E or R69H+L31E. In one embodiment, the CD47-binding domain comprises a SIRP IgV monomer comprising a mutation of Q52H+I31E or Q52H+L31E. In one embodiment, the CD47-binding domain comprises a SIRP IgV monomer comprising a mutation of K68H+I31E or K68H+L31E. In one embodiment, the CD47-binding domain comprises a SIRP IgV monomer comprising a mutation of Q52H+K68H+I31E or Q52H+K68H+L31E. In one embodiment, the CD47-binding SIRPγIgV monomer comprises a mutation of K53H+N101D. In one embodiment, the CD47-binding SIRPγ IgV monomer comprises a mutation of K53H+L37Q. In one embodiment, the CD47-binding SIRPγIgV monomer comprises a mutation of K53H+N101D+L37Q. In one embodiment, the CD47-binding SIRPγIgV monomer comprises a mutation of K53H+N101D+L31E. In one embodiment, the CD47-binding SIRPγIgV monomer comprises a mutation of K53H+N101D+L31D. In some embodiment, the backbone anti-PD-L1 / PD-1 antibody of the multispecific molecule comprising the CD47-binding domain, comprises a Fc of human IgG4 (SEQ ID NO: 1) , active IgG1 (SEQ ID NO: 8) , silent IgG1 (SEQ ID NO: 4, 5, 6 or 7) or IgG2 (SEQ ID NO: 2) . In one embodiment, the backbone anti-PD-L1 / PD-1 antibody comprises a Fc of human IgG4 (SEQ ID NO: 1) . In one embodiment, the backbone antibody comprises a Fc of active human IgG1 (SEQ ID NO: 8) preferably when the backbone antibody is an anti-PD-L1 antibody.
[0021] In some embodiment, the CD47-binding domain of the multispecific molecule is linked through a linker to the N-terminal of the light chain of the backbone anti-PD-L1 / PD-1 antibody and the anti-VEGF single-chain domain is linked through a linker to the C-terminal of the light chain or heavy chain of the backbone anti-PD-L1 / PD-1 antibody. In some embodiment, the CD47-binding domain is linked through a linker to the N-terminal of the heavy chain of the backbone anti-PD-L1 / PD-1 antibody and the anti-VEGF single-chain domain is linked through a linker to the C-terminal of the light chain or heavy chain of the backbone anti-PD-L1 / PD-1 antibody. In some embodiment, the CD47-binding domain of the multispecific molecule is linked through a linker to the N-terminal of the light chain of the backbone anti-PD-L1 / PD-1 antibody and the anti-VEGF single-chain domain is linked through a linker to the N-terminal of the heavy chain of the backbone anti-PD-L1 / PD-1 antibody. In some embodiment, the CD47-binding domain of the multispecific molecule is linked through a linker to the N-terminal of the heavy chain of the backbone anti-PD-L1 / PD-1 antibody and the anti-VEGF single-chain domain is linked through a linker to the N-terminal of the light chain of the backbone anti-PD-L1 / PD-1 antibody. In some embodiment, the linker linking the CD47-binding domain to the anti-PD-L1 / PD-1 antibody or the anti-VEGF single-chain domain comprises an amino acid sequence of SEQ ID NO: 9-14. In one preferred embodiment, the linker comprises an amino acid sequence of SEQ ID NO: 11.
[0022] In some embodiment, the anti-PD-L1 / PD-1 backbone antibody of the multispecific molecule comprising the CD47-binding domain comprises a VH amino acid sequence of SEQ ID NO: 32 and a VL amino acid sequence of SEQ ID NO: 33, and the anti-VEGF single-chain domain comprises an amino acid sequence of SEQ ID NO: 17, 18, 24, 25, 26, 27, 28, 29, 30 or 451. In some embodiment, the anti-PD-L1 / PD-1 backbone antibody of the multispecific molecule comprising the CD47-binding domain comprises a VH amino acid sequence of SEQ ID NO: 34 and a VL amino acid sequence of SEQ ID NO: 35, and the anti-VEGF single-chain domain comprises an amino acid sequence of SEQ ID NO: 17, 18, 24, 25, 26, 27, 28, 29, 30 or 451. In some embodiment, the anti-PD-L1 / PD-1 backbone antibody of the multispecific molecule comprising the CD47-binding domain comprises a VH amino acid sequence of SEQ ID NO: 38 and a VL amino acid sequence of SEQ ID NO: 39, and the anti-VEGF single-chain domain comprises an amino acid sequence of SEQ ID NO: 17, 18, 24, 25, 26, 27, 28, 29, 30 or 451. In some embodiment, the anti-PD-L1 / PD-1 backbone antibody of the multispecific molecule comprising the CD47-binding domain comprises a VH amino acid sequence of SEQ ID NO: 40 and a VL amino acid sequence of SEQ ID NO: 41, and the anti-VEGF single-chain domain comprises an amino acid sequence of SEQ ID NO: 17, 18, 24, 25, 26, 27, 28, 29, 30 or 451. In a further embodiment, the anti-VEGF single-chain domain of the multispecific molecule comprising the CD47-binding domain and the anti-PD-L1 / PD-1 antibody of the foregoing embodiments comprises an amino acid sequence of SEQ ID NO: 17.
[0023] In another aspect, the multispecific molecule comprising a backbone anti-VEGF antibody and an anti-PD-L1 / PD-1 single-chain domain, further comprises a third domain that binds to CD47, wherein the CD47-binding domain is linked through a linker to the backbone anti-VEGF antibody or the anti-PD-L1 / PD-1 single-chain domain. In some embodiment, the CD47-binding domain of the multispecific molecule comprises a SIRP IgV monomer comprising an amino acid sequence selected from the group consisting of SEQ ID NO: 57 to 75. In some embodiment, the CD47-binding domain of the multispecific molecule comprises a SIRP IgV multimer polypeptide comprising two SIRP IgV monomers linked in tandem through a linker of SEQ ID NO: 11 wherein each of the SIRP IgV monomers comprises an amino acid sequence selected from the group consisting of SEQ ID NO: 57 to 75. In one embodiment, the CD47-binding domain comprises an amino acid sequence of SEQ ID NO: 58. In one embodiment, the CD47-binding SIRP IgV monomer comprises an amino acid sequence of SEQ ID NO: 57. In one embodiment, the CD47-binding SIRP IgV monomer comprises an amino acid sequence of SEQ ID NO: 72.
[0024] In some embodiment, the CD47-binding domain comprises a SIRPαor SIRPβIgV monomer comprising a mutation of I31Y or I31W, wherein the SIRP IgV variant monomer comprising the mutation exhibits enhanced binding to CD47 than the parent SIRP IgV monomer without the mutation. In some embodiment, the CD47-binding domain comprises a SIRPγmonomer comprising a mutation of L31Y or L31W, wherein the SIRP IgV variant monomer comprising the mutation exhibits enhanced binding to CD47 than the parent SIRP IgV monomer without the mutation. In some embodiment, the CD47-binding SIRP IgV monomer comprises a mutation of R59H, wherein the SIRP IgV variant monomer comprising the mutation exhibits enhanced binding to CD47 than the parent SIRP IgV monomer without the mutation. In some embodiment, the CD47-binding SIRP IgV monomer comprises a mutation of R59H, R59H+I31Y, R59H+I31W, R59H+L31Y or R59H+L31W, wherein the SIRP IgV variant monomer comprising the mutation exhibits enhanced binding to CD47 than the parent SIRP IgV monomer without the mutation. In some embodiment, the backbone anti-VEGF antibody of the multispecific molecule comprises a Fc of human IgG4 (SEQ ID NO: 1) , IgG2 (SEQ ID NO: 2) or silent IgG1 (SEQ ID NO: 4, 5, 6 or 7) . In one preferred embodiment, the backbone anti-VEGF antibody comprises a Fc of human IgG4 (SEQ ID NO: 1) .
[0025] In some embodiment, the CD47-binding domain comprises a SIRPαor SIRPβIgV monomer comprising a mutation selected from the group consisting of K53H, R69H, K68H, Q52H, I31E, I31D, Q52H+K68H, K53H+K68H, R69H+K68H, K53H+Q52H, R69H+Q52H, K53H+I31E, R69H+I31E, K68H+I31E, Q52H+I31E, K53H+K68H+Q52H, R69H+K68H+Q52H, Q52H+K68H+I31E, K53H+I31D, R69H+I31D, K68H+I31D, Q52H+I31D, and Q52H+K68H+I31D, wherein the SIRP IgV variant monomer comprising the mutation exhibits higher binding to CD47 at an acidic pH than at physiological pH, wherein the acidic pH is<7.0 and the physiological pH is between 7.2 and 7.5. In some embodiment, the CD47-binding domain comprises a SIRPγIgV monomer comprising a mutation selected from the group consisting of K53H, R69H, K68H, Q52H, L31E, L31D, Q52H+K68H, K53H+K68H, R69H+K68H, K53H+Q52H, R69H+Q52H, K53H+L31E, R69H+L31E, K68H+L31E, Q52H+L31E, K53H+K68H+Q52H, R69H+K68H+Q52H, Q52H+K68H+L31E, K53H+L31D, R69H+L31D, K68H+L31D, Q52H+L31D, Q52H+K68H+L31D, K53H+N101D, K53H+N101D+L37Q, K53H+N101D+L31E and K53H+N101D+L31D, wherein the SIRP IgV variant monomer comprising the mutation exhibits higher binding to CD47 at an acidic pH than at physiological pH, wherein the acidic pH is<7.0 and the physiological pH is between 7.2 and 7.5. In one embodiment, the CD47-binding domain comprises a SIRP IgV monomer comprising a mutation of K53H. In one embodiment, the CD47-binding domain comprises a SIRP IgV monomer comprising a mutation of R69H. In one embodiment, the CD47-binding domain comprises a SIRP IgV monomer comprising a mutation of Q52H. In one embodiment, the CD47-binding domain comprises a SIRP IgV monomer comprising a mutation of K68H. In one embodiment, the CD47-binding domain comprises a SIRP IgV monomer comprising a mutation of Q52H+K68H. In one embodiment, the CD47-binding domain comprises a SIRP IgV monomer comprising a mutation of K53H+I31E or K53H+L31E. In one embodiment, the CD47-binding domain comprises a SIRP IgV monomer comprising a mutation of R69H+I31E or R69H+L31E. In one embodiment, the CD47-binding domain comprises a SIRP IgV monomer comprising a mutation of Q52H+I31E or Q52H+L31E. In one embodiment, the CD47-binding domain comprises a SIRP IgV monomer comprising a mutation of K68H+I31E or K68H+L31E. In one embodiment, the CD47-binding domain comprises a SIRP IgV monomer comprising a mutation of Q52H+K68H+I31E or Q52H+K68H+L31E. In one embodiment, the CD47-binding SIRPγIgV monomer comprises a mutation of K53H+N101D. In one embodiment, the CD47-binding SIRPγ IgV monomer comprises a mutation of K53H+L37Q. In one embodiment, the CD47-binding SIRPγIgV monomer comprises a mutation of K53H+N101D+L37Q. In one embodiment, the CD47-binding SIRPγIgV monomer comprises a mutation of K53H+N101D+L31E. In one embodiment, the CD47-binding SIRPγIgV monomer comprises a mutation of K53H+N101D+L31D. In some embodiment, the backbone anti-VEGF antibody of the multispecific molecule comprising the CD47-binding domain, comprises a Fc of human IgG4 (SEQ ID NO: 1) , active IgG1 (SEQ ID NO: 8) , silent IgG1 (SEQ ID NO: 4, 5, 6 or 7) or IgG2 (SEQ ID NO: 2) . In one embodiment, the backbone anti-VEGF antibody comprises a Fc of human IgG4 (SEQ ID NO: 1) . In one embodiment, the backbone anti-VEGF antibody comprises a Fc of active human IgG1 (SEQ ID NO: 8) when the anti-PD-L1 / PD-1 single-chain domain is an anti-PD-L1 domain.
[0026] In some embodiment, the CD47-binding domain of the multispecific molecule is linked through a linker to the N-terminal of the light chain of the backbone anti-VEGF antibody and the anti-PD-L1 / PD-1 single-chain domain is linked through a linker to the C-terminal of the heavy chain of the backbone anti-VEGF antibody. In some embodiment, the CD47-binding domain of the multispecific molecule is linked through a linker to the N-terminal of the heavy chain of the backbone anti-VEGF antibody and the anti-PD-L1 / PD-1 single-chain domain is linked through a linker to the C-terminal of the heavy chain of the backbone anti-VEGF antibody. In some embodiment, the CD47-binding domain of the multispecific molecule is linked through a linker to the C-terminal of the heavy chain of the backbone anti-VEGF antibody and the CD47-binding domain is linked through a linker to the C-terminal of the anti-PD-L1 / PD-1 single-chain domain. In some embodiment, the CD47-binding domain of the multispecific molecule is linked through a linker to the C-terminal of the heavy chain of the backbone anti-VEGF antibody and the anti-PD-L1 / PD-1 single-chain domain is linked through a linker to the C-terminal of the CD47-binding SIRP IgV domain. In some embodiment, the CD47-binding domain of the multispecific molecule is linked through a linker to the N-terminal of the heavy chain of the backbone anti-VEGF antibody and the CD47-binding domain is linked through a linker to the C-terminal of the heavy chain of the backbone anti-VEGF antibody. In some embodiment, the CD47-binding domain of the multispecific molecule is linked through a linker to the N-terminal of the light chain of the backbone anti-VEGF antibody and the CD47-binding domain is linked through a linker to the C-terminal of the heavy chain of the backbone anti-VEGF antibody. In some embodiment, the CD47-binding domain of the multispecific molecule is linked through a linker to the N-terminal of the light chain of the backbone anti-VEGF antibody and the anti-PD-L1 / PD-1 single-chain domain is linked through a linker to the N-terminal of the heavy chain of the backbone anti-VEGF antibody. In some embodiment, the CD47-binding domain of the multispecific molecule is linked through a linker to the N-terminal of the heavy chain of the backbone anti-VEGF antibody and the anti-PD-L1 / PD-1 single-chain domain is linked through a linker to the N-terminal of the light chain of the backbone anti-VEGF antibody. In some embodiment, the linker linking the CD47-binding domain to the anti-VEGF antibody or the anti-PD-L1 / PD-1 single-chain domain comprises an amino acid sequence selected from the group consisting of SEQ ID NO: 9-14 and 472-480. In one embodiment, the linker comprises an amino acid sequence of SEQ ID NO: 11.
[0027] In some embodiment, the anti-PD-L1 / PD-1 single-chain domain of the multispecific molecule comprises an amino acid sequence of 46. In some embodiment, the anti-PD-L1 / PD-1 single-chain domain of the multispecific molecule comprises an amino acid sequence of SEQ ID NO: 47. In some embodiment, the anti-PD-L1 / PD-1 single-chain domain of the multispecific molecule comprises an amino acid sequence of SEQ ID NO: 48. In some embodiment, the anti-PD-L1 / PD-1 single-chain domain of the multispecific molecule comprises an amino acid sequence of SEQ ID NO: 49. In some embodiment, the anti-PD-L1 / PD-1 single-chain domain of the multispecific molecule comprises an amino acid sequence of SEQ ID NO: 50. In some embodiment, the anti-PD-L1 / PD-1 single-chain domain of the multispecific molecule comprises an amino acid sequence of SEQ ID NO: 51. In some embodiment, the anti-PD-L1 / PD-1 single-chain domain of the multispecific molecule comprises an amino acid sequence of SEQ ID NO: 52. In some embodiment, the anti-PD-L1 / PD-1 single-chain domain of the multispecific molecule comprises an amino acid sequence of SEQ ID NO: 53. In some embodiment, the anti-PD-L1 / PD-1 single-chain domain of the multispecific molecule comprises an amino acid sequence of SEQ ID NO: 54.
[0028] In another aspect, a multispecific molecule comprising: 1) a CD47-binding domain, wherein the CD47-binding domain comprises a SIRP IgV monomer or a SIRP IgV multimer polypeptide comprising two SIRP IgV monomers linked in tandem through a linker of SEQ ID NO: 11, wherein the SIRP IgV monomer comprising an amino acid sequence selected from the group consisting of SEQ ID NO: 57 to 75; and 2) a second domain that binds to VEGF; and 3) a third domain that binds to PD-L1 or PD-1. In some embodiment, the CD47-binding SIRP IgV monomer of the multispecific molecule is a SIRPαor SIRPβIgV monomer comprising a mutation of I31Y or I31W. In some embodiment, the CD47-binding SIRP IgV monomer of the multispecific molecule is a SIRPγIgV monomer comprising a mutation of L31Y or L31W. In some embodiment, the CD47-binding SIRP IgV monomer of the multispecific molecule is a SIRPαor SIRPβIgV monomer comprising a mutation selected from the group consisting of K53H, R69H, K68H, Q52H, I31E, I31D, Q52H+K68H, K53H+K68H, R69H+K68H, K53H+Q52H, R69H+Q52H, K53H+I31E, R69H+I31E, K68H+I31E, Q52H+I31E, K53H+K68H+Q52H, R69H+K68H+Q52H, Q52H+K68H+I31E, K53H+I31D, R69H+I31D, K68H+I31D, Q52H+I31D, and Q52H+K68H+I31D. In some embodiment, the CD47-binding SIRP IgV monomer of the multispecific molecule is a SIRPγIgV monomer comprising a mutation selected from the group consisting of K53H, R69H, K68H, Q52H, L31E, L31D, Q52H+K68H, K53H+K68H, R69H+K68H, K53H+Q52H, R69H+Q52H, K53H+L31E, R69H+L31E, K68H+L31E, Q52H+L31E, K53H+K68H+Q52H, R69H+K68H+Q52H, Q52H+K68H+L31E, K53H+L31D, R69H+L31D, K68H+L31D, Q52H+L31D, Q52H+K68H+L31D, K53H+N101D, K53H+N101D+L37Q, K53H+N101D+L31E and K53H+N101D+L31D. In one embodiment, the CD47-binding domain comprises a SIRP IgV monomer comprising a mutation of K53H. In one embodiment, the CD47-binding domain comprises a SIRP IgV monomer comprising a mutation of R69H. In one embodiment, the CD47-binding domain comprises a SIRP IgV monomer comprising a mutation of Q52H. In one embodiment, the CD47-binding domain comprises a SIRP IgV monomer comprising a mutation of K68H. In one embodiment, the CD47-binding domain comprises a SIRP IgV monomer comprising a mutation of Q52H+K68H. In one embodiment, the CD47-binding domain comprises a SIRP IgV monomer comprising a mutation of K53H+I31E or K53H+L31E. In one embodiment, the CD47-binding domain comprises a SIRP IgV monomer comprising a mutation of R69H+I31E or R69H+L31E. In one embodiment, the CD47-binding domain comprises a SIRP IgV monomer comprising a mutation of Q52H+I31E or Q52H+L31E. In one embodiment, the CD47-binding domain comprises a SIRP IgV monomer comprising a mutation of K68H+I31E or K68H+L31E. In one embodiment, the CD47-binding domain comprises a SIRP IgV monomer comprising a mutation of Q52H+K68H+I31E or Q52H+K68H+L31E. In one embodiment, the CD47-binding SIRPγIgV monomer comprises a mutation of K53H+N101D. In one embodiment, the CD47-binding SIRPγIgV monomer comprises a mutation of K53H+L37Q. In one embodiment, the CD47-binding SIRPγIgV monomer comprises a mutation of K53H+N101D+L37Q. In one embodiment, the CD47-binding SIRPγIgV monomer comprises a mutation of K53H+N101D+L31E. In one embodiment, the CD47-binding SIRPγIgV monomer comprises a mutation of K53H+N101D+L31D.
[0029] In some embodiment, the second domain of the multispecific molecule comprising the CD47-binding domain is an anti-VEGF single-chain domain comprising an amino acid sequence selected from the group consisting of SEQ ID NO: 17 to 30 and 451 to 453, and the third domain of the multispecific molecule is an anti-PD-L1 / PD-1 antibody comprising a VH amino acid sequence of SEQ ID NO: 32 and a VL amino acid sequence of SEQ ID NO: 33, A VH sequence of SEQ ID NO: 34 and a VL sequence of SEQ ID NO: 35, a VH sequence of SEQ ID NO: 36 and a VL sequence of SEQ ID NO: 37, a VH sequence of SEQ ID NO: 38 and a VL sequence of SEQ ID NO: 39, A VH sequence of SEQ ID NO: 40 and a VL sequence of SEQ ID NO: 41, or a VH sequence of SEQ ID NO: 42 and a VL sequence of SEQ ID NO: 43, wherein both the CD47-binding domain and the anti-VEGF single-chain domain are linked through a linker to the anti-PD-L1 / PD-1 antibody. In one embodiment, the second domain is an anti-VEGF single-chain domain comprising an amino acid sequence of SEQ ID NO: 17. In one embodiment, the second domain is an anti-VEGF single-chain domain comprising an amino acid sequence of SEQ ID NO: 18. In one embodiment, the second domain is an anti-VEGF single-chain domain comprising an amino acid sequence of SEQ ID NO: 26. In one embodiment, the second domain is an anti-VEGF single-chain domain comprising an amino acid sequence of SEQ ID NO: 27. In one embodiment, the second domain is an anti-VEGF single-chain domain comprising an amino acid sequence of SEQ ID NO: 28. In one embodiment, the second domain is an anti-VEGF single-chain domain comprising an amino acid sequence of SEQ ID NO: 29. In one embodiment, the second domain is an anti-VEGF single-chain domain comprising an amino acid sequence of SEQ ID NO: 30. In one embodiment, the third domain is an anti-PD-L1 / PD-1 antibody comprising a VH amino acid sequence of SEQ ID NO: 32 and a VL amino acid sequence of SEQ ID NO: 33. In one embodiment, the third domain is an anti-PD-L1 / PD-1 antibody comprising a VH amino acid sequence of SEQ ID NO: 34 and a VL amino acid sequence of SEQ ID NO: 35. In one embodiment, the third domain is an anti-PD-L1 / PD-1 antibody comprising a set of VH and VL amino acid sequence of SEQ ID NO: 36 and 37. In one embodiment, the third domain is an anti-PD-L1 / PD-1 antibody comprising a VH amino acid sequence of SEQ ID NO: 38 and a VL amino acid sequence of SEQ ID NO: 39. In one embodiment, the third domain is an anti-PD-L1 / PD-1 antibody comprising a set of VH and VL amino acid sequence of SEQ ID NO: 40 and 41. In one embodiment, the third domain is an anti-PD-L1 / PD-1 antibody comprising a VH amino acid sequence of SEQ ID NO: 42 and a VL amino acid sequence of SEQ ID NO: 43.
[0030] In some embodiment, the linker linking the CD47-binding domain or the anti-VEGF single-chain domain to the anti-PD-L1 / PD-1 antibody of the multispecific molecule comprises an amino acid sequence selected from the group consisting of SEQ ID NO: 9-14 and 472-480. In one preferred embodiment, the linker comprises an amino acid sequence of SEQ ID NO: 11. In some embodiment, the anti-VEGF single-chain domain is linked through a linker to the N-terminal or C-terminal of the heavy chain of the anti-PD-L1 / PD-1 antibody, and the CD47-binding domain is linked through a linker to the N-terminal or C-terminal of the light chain of the anti-PD-L1 / PD-1 antibody. In some embodiment, the anti-VEGF single-chain domain is linked through a linker to the N-terminal or C-terminal of the light chain of the anti-PD-L1 / PD-1 antibody, and the CD47-binding domain is linked through a linker to the N-terminal or C-terminal of the heavy chain of the anti-PD-L1 / PD-1 antibody. In some embodiment, the CD47-binding domain is linked through a linker to the N-terminal of the light chain of the backbone anti-PD-L1 / PD-1 antibody and the anti-VEGF single-chain domain is linked through a linker to the C-terminal of the light chain or heavy chain of the backbone anti-PD-L1 / PD-1 antibody. In some embodiment, the CD47-binding domain is linked through a linker to the N-terminal of the heavy chain of the backbone anti-PD-L1 / PD-1 antibody and the anti-VEGF single-chain domain is linked through a linker to the C-terminal of the light chain or heavy chain of the backbone anti-PD-L1 / PD-1 antibody. In some embodiment, the CD47-binding domain is linked through a linker to the N-terminal of the light chain of the backbone anti-PD-L1 / PD-1 antibody and the anti-VEGF single-chain domain is linked through a linker to the N-terminal of the heavy chain of the backbone anti-PD-L1 / PD-1 antibody. In some embodiment, the CD47-binding domain is linked through a linker to the N-terminal of the heavy chain of the backbone anti-PD-L1 / PD-1 antibody and the anti-VEGF single-chain domain is linked through a linker to the N-terminal of the light chain of the backbone anti-PD-L1 / PD-1 antibody.
[0031] In some embodiment, the backbone anti-PD-L1 / PD-1 antibody of the multispecific molecule comprising the CD47-binding domain, comprises a Fc of human IgG4 (SEQ ID NO: 1) , active IgG1 (SEQ ID NO: 8) , silent IgG1 (SEQ ID NO: 4, 5, 6 or 7) or IgG2 (SEQ ID NO: 2) . In one embodiment, the backbone anti-PD-L1 / PD-1 antibody comprises a Fc of human IgG4 (SEQ ID NO: 1) . In one embodiment, the backbone anti-PD-L1 / PD-1 antibody comprises a Fc of active human IgG1 (SEQ ID NO: 8) when the backbone antibody is an anti-PD-L1 antibody.
[0032] In some embodiment, the second domain of the multispecific molecule comprising the CD47-binding domain is an anti-VEGF antibody comprising a VH sequence of SEQ ID NO: 15 and a VL sequence of SEQ ID NO: 16, and the third domain is an anti-PD-L1 / PD-1 single-chain domain comprising an amino acid sequence selected from the group consisting of SEQ ID NO: 46 to 54 and 455, wherein both the CD47-binding domain and the anti-PD-L1 / PD-1 single-chain domain are linked through a linker to the anti-VEGF antibody. In one embodiment, the third domain is an anti-PD-L1 / PD-1 single-chain domain comprising an amino acid sequence of 46. In one embodiment, the third domain is an anti-PD-L1 / PD-1 single-chain domain comprising an amino acid sequence of SEQ ID NO: 47. In one embodiment, the third domain is an anti-PD-L1 / PD-1 single-chain domain comprising an amino acid sequence of SEQ ID NO: 48. In one embodiment, the third domain is an anti-PD-L1 / PD-1 single-chain domain comprising an amino acid sequence of SEQ ID NO: 49. In one embodiment, the third domain is an anti-PD-L1 / PD-1 single-chain domain comprising an amino acid sequence of SEQ ID NO: 50.In one embodiment, the third domain is an anti-PD-L1 / PD-1 single-chain domain comprising an amino acid sequence of SEQ ID NO: 51. In one embodiment, the third domain is an anti-PD-L1 / PD-1 single-chain domain comprising an amino acid sequence of SEQ ID NO: 52. In one embodiment, the third domain is an anti-PD-L1 / PD-1 single-chain domain comprising an amino acid sequence of SEQ ID NO: 53. In one embodiment, the third domain is an anti-PD-L1 / PD-1 single-chain domain comprising an amino acid sequence of SEQ ID NO: 54.
[0033] In some embodiment, the linker linking the CD47-binding domain or the anti-PD-L1 / PD-1 single-chain domain to the anti-VEGF antibody comprises an amino acid sequence selected from the group consisting of SEQ ID NO: 9-14 and 472-480. In one preferred embodiment, the linker comprises an amino acid sequence of SEQ ID NO: 11. In some embodiment, the CD47-binding domain of the multispecific molecule is linked through a linker to the N-terminal of the light chain of the backbone anti-VEGF antibody and the anti-PD-L1 / PD-1 single-chain domain is linked through a linker to the C-terminal of the heavy chain of the backbone anti-VEGF antibody. In some embodiment, the CD47-binding domain of the multispecific molecule is linked through a linker to the N-terminal of the heavy chain of the backbone anti-VEGF antibody and the anti-PD-L1 / PD-1 single-chain domain is linked through a linker to the C-terminal of the heavy chain of the backbone anti-VEGF antibody. In some embodiment, the anti-PD-L1 / PD-1 single-chain domain of the multispecific molecule is linked through a linker to the C-terminal of the heavy chain of the backbone anti-VEGF antibody and the CD47-binding domain is linked through a linker to the C-terminal of the anti-PD-L1 / PD-1 single-chain domain. In some embodiment, the CD47-binding domain of the multispecific molecule is linked through a linker to the C-terminal of the heavy chain of the backbone anti-VEGF antibody and the anti-PD-L1 / PD-1 single-chain domain is linked through a linker to the C-terminal of the CD47-binding SIRP IgV domain. In some embodiment, the anti-PD-L1 / PD-1 single-chain domain of the multispecific molecule is linked through a linker to the N-terminal of the heavy chain of the backbone anti-VEGF antibody and the CD47-binding domain is linked through a linker to the C-terminal of the heavy chain of the backbone anti-VEGF antibody. In some embodiment, the anti-PD-L1 / PD-1 single-chain domain of the multispecific molecule is linked through a linker to the N-terminal of the light chain of the backbone anti-VEGF antibody and the CD47-binding domain is linked through a linker to the C-terminal of the heavy chain of the backbone anti-VEGF antibody. In some embodiment, the CD47-binding domain of the multispecific molecule is linked through a linker to the N-terminal of the light chain of the backbone anti-VEGF antibody and the anti-PD-L1 / PD-1 single-chain domain is linked through a linker to the N-terminal of the heavy chain of the backbone anti-VEGF antibody.
[0034] In some embodiment, the backbone anti-VEGF antibody of the multispecific molecule comprises a Fc of human IgG4 (SEQ ID NO: 1) , silent IgG1 (SEQ ID NO: 4, 5, 6 or 7) , IgG2 (SEQ ID NO: 2) or active IgG1 (SEQ ID NO: 8) . In one embodiment, the backbone anti-VEGF antibody comprises a Fc of human IgG4 (SEQ ID NO: 1) . In one embodiment, the backbone anti-VEGF antibody comprises a Fc of silent human IgG1 (SEQ ID NO: 5) . In one preferred embodiment, the backbone anti-VEGF antibody comprises a Fc of active human IgG1 (SEQ ID NO: 8) when the anti-PD-L1 / PD-1 single-chain domain is an anti-PD-L1 single-chain domain.
[0035] In some embodiment, a nucleic acid or nucleic acids comprise a sequence encoding a polypeptide of the multispecific molecule of present disclosure. In additional embodiments, the nucleic acid is a DNA or RNA. In another embodiment, a vector or vectors (e.g. cloning vector, and expression vector) comprise the foregoing nucleic acid or nucleic acids. In additional embodiments, the vector comprises a plasmid and / or a viral vector. In some embodiments, a host cell comprises one or more of the foregoing vectors. In some embodiments, a process for production of a polypeptide or protein of present disclosure, comprising culturing the host cell with the foregoing vector or vectors and isolating the polypeptide or protein.
[0036] In some embodiment, a pharmaceutical composition comprises a multispecific molecule of present disclosure, and a pharmaceutically acceptable carrier.
[0037] In some embodiments, a kit for diagnosis or treatment, said kit comprises a multispecific molecule of present disclosure, or the pharmaceutical composition thereof, and instruction for using it for diagnosis or treatment.
[0038] In some embodiment, provided herein a method of treating cancer in a mammal comprising administering an effective amount of a multispecific molecule of present disclosure, a nucleic acid of present disclosure, a vector of present disclosure, and / or the pharmaceutical composition, to a subjuet, e.g. a mammal, in need thereof. In some embodiments, the cancer for treatment by the multispecific molecule comprises ovarian cancer, endometrial cancer, uterine cancer, cervical cancer, breast cancer, gastric cancer, esophageal cancer, lung cancer, head and neck cancer, colorectal cancer, liver cancer, bone cancer, sarcoma, osteosarcoma, brain cancer, multiple myeloma, acute myeloid leukemia (AML) , myelodysplastic syndrome, adrenal cancer, anal cancer, bile duct cancer, bladder cancer, basal cell skin cancer, chondrosarcoma, Ewing’s sarcoma, eye cancer, gallbladder cancer, gastrointestinal carcinoid tumors, gastrointestinal stromal tumor (GIST) , glioma, glioblastoma, hepatocellular cancer, kaposi sarcoma, kidney cancer, laryngeal and hypopharyngeal cancer, leiomyosarcoma, non-small cell lung cancer (NSCLC) , small cell lung cancer (SCLC) , lung carcinoid tumor, mesothelioma, Merkel cell carcinoma, melanoma, nasal cavity and paranasal sinus cancer, nasopharyngeal cancer, neuroendocrine cancer, neuroblastoma, oral cavity and oropharyngeal cancer, pancreatic cancer, penile cancer, pituitary tumors, prostate cancer, renal cancer, retinoblastoma, rhabdomyosarcoma, salivary gland cancer, squamous cell skin cancer, small intestine cancer, stomach cancer, testicular cancer, thymus cancer, thyroid cancer, uterine sarcoma, urinary tract cancer, vaginal cancer, vulvar cancer, Wilms’ tumor, diffuse large B-cell lymphoma (DLBCL) , follicular lymphoma, hairy-cell leukemia, Hodgkin’s lymphoma, leukemia, acute lymphocytic leukemia (ALL) , chronic lymphocytic leukemia (CLL) , chronic myeloid leukemia (CML) , chronic myelomonocytic leukemia (CMML) , lymphoma, Mantle cell lymphoma (MCL) , marginal zone lymphoma (MZL) , myeloma, Non-Hodgkin lymphoma (NHL) , plasmacytoma, or Waldenstrom macroglobulinemia.
[0039] In some embodiment, the present disclosure provides a method of combination therapy in human comprising administering a therapeutically effective amount of the multispecific molecule of present invention, and a therapeutically effective amount of another therapy. In some embodiments, another therapy comprises administration of one or more of chemotherapy, molecular targeted therapy, biologics therapy, immunotherapy and radiotherapy. INCORPORATION BY REFERENCE:
[0040] All publications, literature, patents, and patent applications mentioned in this disclosure, including in all tables and figures, are herein incorporated by reference as if each individual publication, patent, or patent application was specifically and individually indicated to be incorporated by reference.
[0041] For illustrative purposes, the principles of the present invention are described by referencing various exemplary embodiments. Although certain embodiments of the invention are specifically described herein, one of ordinary skill in the art will readily recognize that the same principles are equally applicable to, and can be employed in other systems and methods. Before explaining the disclosed embodiments of the present invention in detail, it is to be understood that the invention is not limited in its application to the details of any particular embodiment shown. Additionally, the terminology used herein is for the purpose of description and not of limitation. Furthermore, although certain methods are described with reference to steps that are presented herein in a certain order, in many instances, these steps may be performed in any order as may be appreciated by one skilled in the art; the novel method is therefore not limited to the particular arrangement of steps disclosed herein.BRIEF DESCRIPTION OF THE DRAWINGS
[0042] Some embodiments of the invention are herein described, by way of example only, with reference to the accompanying drawings. With specific reference now to the drawings in detail, it is stressed that the particulars shown are by way of example and for purposes of illustrative discussion of embodiments of the invention. In this regard, the description taken with the drawings makes apparent to those skilled in the art how embodiments of the invention may be practiced.
[0043] FIG. 1: illustration of exemplary structural format variants FV-0 to FV-9.
[0044] FIG. 2: illustration of exemplary structural format variants FV-10 to FV-20.
[0045] FIG. 3: illustration of exemplary structural format variants FV-21 to FV-30.
[0046] FIG. 4: illustration of exemplary structural format variants FV-31 to FV-38.
[0047] FIG. 5: illustration of exemplary structural format variants FV-39 to FV-44.
[0048] FIG. 6: FIG. 6A is alignment of human SIRPαvariant IgV domains, including 4 preceding amino acids from the signal peptide sequence. The residues that show difference among the variants are shadowed and marked as green (the same as SIRPαV1) , blue (the same as SIRPαV2) and orange (unique to the specific variant) . FIG. 6B is alignment of IgV domains of human SIRPα (representative SIRPαV1 and V2) , SIRPβ1 and its representative mutants, SIRPβ2and its representative mutants, as well as SIRPγand its representative mutants.
[0049] FIG. 7: FIG. 7A, ELISA binding result of VEGFR-R1D2 domain-comprising bispecific antibody fusion protein against VEGF; FIG. 7B, ELISA binding result of VEGFR-R1D2 domain-comprising bispecific antibody fusion protein against PD-L1; FIG. 7C, ELISA result of simultanesous binding of PD-L1 and VEGF by VEGFR-R1D2 domain-comprising bispecific antibody fusion protein; FIG. 7D-F, ELISA binding result of various bevacizumab scFv-comprising bispecific antibody fusion protein against VEGF.
[0050] FIG. 8: ELISA binding result of bispecific antibody protein against immobilized PD-L1.
[0051] FIG. 9: ELISA binding result of bi / trispecific antibody protein against immobilized VEGF.
[0052] FIG. 10: ELISA binding result of bi / trispecific antibody protein against immobilized PD-L1.
[0053] FIG. 11: ELISA binding result of bi / trispecific antibody protein against immobilized CD47.
[0054] FIG. 12: FACS binding result of bi / trispecific antibody protein against PD-L1+CD47+HT-1080 cells.
[0055] FIG. 13: ELISA result for simultanesous binding of PD-L1, VEGF and CD47 by trispecific antibody protein. 96-well plates were first coated with PD-L1 antigen, and then incubated with the trispecific antibody protein tested, and after wash, then incubate with the VEGF antigen with 6x His-tag, CD47 antigen with 6x His-tag or a combination of the same VEGF+CD47 antigen, and after wash, then detect with anti-His-HRP.
[0056] FIG. 14: FIG. 14A shows FACS binding on Raji cells of wild type and variant SIRPαV2 IgV-IgG1 Fc fusion protein at physiological pH 7.2. FIG. 14B shows FACS binding on Raji cells of wild type and variant SIRPαV2 IgV-IgG1 Fc fusion protein at an acidic pH 6.0. FIG. 14C-D show FACS binding on Raji cells of wild type and variant SIRPαV2 IgV-IgG1 Fc fusion protein at both physiological and acidic pH.
[0057] FIG. 15: FIG. 15A shows FACS binding on Raji cells of wild type SIRPαV2 IgV-IgG1 Fc fusion protein SIN-300 and variant SIRPαV2 IgV-IgG1 Fc fusion proteinIN-301 and SIN-304 at an acidic pH and physiological pH.FIG. 15B shows FACS binding on Raji cells of wild type SIRPαV2 IgV-IgG1 Fc fusion protein SIN-300 and variant SIRPαV2 IgV-IgG1 Fc fusion protein SIN-302 and SIN-303 at an acidic pH and physiological pH. FIG. 15C shows FACS binding on SK-OV-3 cells of wild type and variant SIRPαV2 IgV-IgG1 Fc fusion protein at an acidic pH and physiological pH.
[0058] FIG. 16: FIG. 16A shows ELISA of wild type and variant SIRPαV2 IgV-IgG1 Fc fusion protein with human CD47 protein at physiological pH 7.2. SIRPαV2 SIRPαV2 IgV-IgG1 Fc fusion protein were immobilized and recombinant human CD47 protein with His-tag was used to detect monovalent binding with the SIRPαV2 IgV-IgG1 Fc fusion proteinSIRPαV2. FIG. 16B shows ELISA of wild type and variant SIRPαV2 IgV-IgG1 Fc fusion protein with human CD47 protein at an acidic pH 6.0. SIRPαV2 IgV-IgG1 Fc fusion protein were immobilized and recombinant human CD47 protein with His-tag was used to detect monovalent binding with the SIRPαV2 IgV-IgG1 Fc fusion proteinSIRPαV2. FIG. 16C shows ELISA of wild type and variant SIRPαV2 IgV-IgG1 Fc fusion protein against immobilized human CD47 protein at an acidic pH 6.0 and physiological pH 7.3. Recombinant human CD47 protein was immobilized and the wild type and variant SIRPαV2 IgV-IgG1 Fc fusion protein comprising bivalent or monovalent SIRPαV2 IgV domain were used to detect binding with the immobilized CD47.
[0059] FIG. 17: FIG. 17A shows FACS binding on primary human platelets of wild type and variant SIRPαV2 IgV-IgG1 Fc fusion protein at physiological pH 7.3. FIG. 17B shows FACS binding on primary human platelets of wild type and variant SIRPαV2 IgV-IgG1 Fc fusion protein at physiological pH 7.3 at high concentrations. FIG. 17C shows FACS binding on primary human T cells of wild type and variant SIRPαV2 IgV-IgG1 Fc fusion protein at physiological pH 7.3 and acidic pH 6.0.
[0060] FIG. 18: FIG. 18A shows FACS binding on Raji cells of wild type and variant SIRPαV1 IgV-IgG1 Fc fusion protein at an acidic pH and physiological pH. FIG. 18B shows FACS binding on Raji cells of wild type and variant SIRPαV8 IgV-IgG1 Fc fusion protein at an acidic pH and physiological pH. FIG. 18C and 18E show FACS binding on Raji cells of wild type and variant SIRPγIgV-IgG1 Fc fusion protein at an acidic pH and physiological pH. FIG. 18D shows FACS binding on Raji cells of parent and variant SIPRβ2 H101D IgV-IgG1 Fc fusion protein at an acidic pH and physiological pH.
[0061] FIG. 19: FIG. 19A shows ADCP activity of wild type and variant SIRPαV2 IgV-IgG1 Fc fusion protein at physiological pH 7.3 using ADCP Jurkat reporter assay against Raji cells. FIG. 19B shows ADCP activity of wild type and variant SIRPαV2 IgV-IgG1 Fc fusion protein at pH 6.5 using ADCP Jurkat reporter assay against Raji cells.
[0062] FIG. 20: FIG. 20A shows ADCC activity of wild type and variant SIRPαV2 IgV-IgG1 Fc fusion protein at physiological pH 7.2 using ADCC Jurkat reporter assay against Raji cells. FIG. 20B shows ADCC activity of wild type and variant SIRPαV2 IgV-IgG1 Fc fusion protein at an acidic pH 6.0 using ADCC Jurkat reporter assay against Raji cells.
[0063] FIG. 21: FIG. 21A shows FACS binding on Raji cells of IgG1 Fc fusion protein of wild type and variant SIRPαV2 IgV with combo mutations (SIN-332 and SIN-333) at an acidic pH and physiological pH. FIG. 21B shows FACS binding on Raji cells of IgG1 Fc fusion protein of wild type and variant SIRPαV2 IgV with combo mutations (SIN-335, SIN-336 and SIN-337) at an acidic pH and physiological pH. FIG. 21C shows FACS binding on Raji cells of IgG1 Fc fusion protein of wild type and variant SIRPαV2 IgV with combo mutations (SIN-330) at an acidic pH and physiological pH.
[0064] FIG. 22A-E: shows FACS binding on Raji cells of exemplary additional single-mutation variant SIRPαV2 IgV-IgG1 Fc fusion protein at an acidic pH and physiological pH.
[0065] FIG. 23: FIG. 23A-B show FACS binding on Raji cells of IgG1 Fc fusion protein of wild type and variant SIRPαV2 IgV with combo mutation of K53H+Q37H at an acidic pH and physiological pH compared to single K53H backbone mutation; FIG. 23B shows FACS binding on Raji cells of IgG1 Fc fusion protein of wild type and variant SIRPαV2 IgV with combo mutation of R69H+Q37H at an acidic pH and physiological pH compared to single R69H backbone mutation. FIG. 23C-D show FACS binding on Raji cells of IgG1 Fc fusion protein of wild type and variant SIRPαV2 IgV comprising combo mutation of K53H+I31E, or R69H+I31E, or Q52H+I31E, or K68H+I31E at an acidic pH and physiological pH compared to counterpart SIRPαV2 IgV protein comprising single backbone mutation of K53H, R69H, Q52H, or K68H respectively.
[0066] FIG. 24: FIG. 24A shows ELISA CD47-binding result of bispecific antibody fusion protein with tetravalent SIRP IgV variant monomers versus protein with bivalent SIRP IgV variant monomers of the same mutation at acidic pH versus physiological pH. FIG. 24B-C shows Raji cell FACS binding results of bispecific antibody fusion protein with tetravalent SIRP IgV variant monomers versus protein with bivalent SIRP IgV variant monomers of the same mutation at acidic pH versus physiological pH.
[0067] FIG. 25: ELISA CD47-binding result of trispecific antibody fusion protein with tetravalent SIRP IgV variant monomers at acidic pH versus physiological pH.DETAILED DESCRIPTION
[0068] Described herein directs to multispecific antibody fusion protein against PD-L1 / PD-1 and VEGF, or against PD-L1 / PD-1, VEGF and additionally CD47, and the uses thereof. Definition
[0069] As used herein, the singular form "a" , "an" and "the" include plural references unless the context clearly dictates otherwise. For example, the term “a substitution" or “at least one substitution" may include a plurality of substitutions, and the term “a domain” or “the domain” may include a plurality of domains.
[0070] As used herein, the term “about” or “approximately” refers to a quantity, level, value, number, frequency, percentage, dimension, size, amount, weight, or length that varies by as much as 30, 25, 20, 25, 10, 9, 8, 7, 6, 5, 4, 3,2, or 1%to a reference quantity, level, value, number, frequency, percentage, dimension, size, amount, weight, or length. In particular embodiments, the terms “about” or “approximately” when preceding a numerical value indicates the value plus or minus a range of 15%, 10%, 5%, or 1%.
[0071] As used herein, the terms "comprise” , "comprises” , and “comprising” , mean to imply the inclusion of a stated step or element or group of steps or elements, but not the exclusion of any other step or element or group of steps or elements. For example, “a variant comprises a substitution” may include further substitutions.
[0072] Reference throughout this disclosure to “one embodiment, ” “an embodiment, ” “some embodiment” , “aparticular embodiment, ” “a related embodiment, ” “a certain embodiment, ” “an additional embodiment, ” or “afurther embodiment, ” or combinations thereof means that a particular feature, structure, or characteristic described in connection with the embodiment is included in at least one embodiment of the present disclosure. Thus, the appearances of the foregoing phrases in various places throughout this specification are not necessarily all referring to the same embodiment. Furthermore, the particular features, structures, or characteristics may be combined in any suitable manner in one or more embodiments.
[0073] As used herein, the term “fusion protein” means a protein comprising at least two fragments that are not naturally fused in the same protein.
[0074] As used herein, the term “multispecific” refers to specificity against two or more different targets.
[0075] As used herein, the term “format variant” , or “FV” in abbreviation, of a protein refers to the structural configuration of the domain components of the protein, as the exemplary format variants (FVs) illustrated in FIG. 1 to 5.
[0076] As used herein, the term “linked” for fusion protein is intended to mean that the two polypeptide fragments are joined into one polypeptide such that the amino acid sequences of the two polypeptide fragments remain in-frame separately.
[0077] The term "amino acid" as used herein refers to any organic compound that contains an amino group (-NH2) and a carboxyl group (-COOH) , preferably either as free groups or alternatively after condensation as part of peptide bonds. The "twenty naturally encoded polypeptide-forming alpha-amino acids" are understood in the art and refer to: alanine (ala or A) , arginine (arg or R) , asparagine (asn or N) , aspartic acid (asp or D) , cysteine (cys or C) , gluatamic acid (glu or E) , glutamine (gin or Q) , glycine (gly or G) , histidine (his or H) , isoleucine (ile or I) , leucine (leu or L) , lysine (lys or K) , methionine (met or M) , phenylalanine (phe or F) , proline (pro or P) , serine (ser or S) , threonine (thr or T) , tryptophan (tip or W) , tyrosine (tyr or Y) , and valine (val or V) .
[0078] The term “antibody” as used herein encompasses any immunoglobulin, monoclonal antibody, polyclonal antibody, multispecific antibody, or bispecific (bivalent) antibody that binds to a specific antigen. A native intact antibody comprises two heavy chains and two light chains. Each heavy chain consists of a variable region ( “VH” ) and a first, second, and third constant region (CH1, CH2 and CH3) , while each light chain consists of a variable region ( “VL” ) and a constant region (CL) . Mammalian heavy chains are classified asα, δ, ε, γ, andμ, and mammalian light chains are classified asλorκ. The variable regions of the light and heavy chains are responsible for antigen binding. The variable regions in both chains generally contain three highly variable loops called the complementarity determining regions (CDRs) (light (L) chain CDRs including LCDR1, LCDR2, and LCDR3, heavy (H) chain CDRs including HCDR1, HCDR2, HCDR3) . The three CDRs are interposed between flanking stretches known as framework regions (FRs) , which are more highly conserved than the CDRs and form a scaffold to support the hypervariable loops. Each VHand VL comprises four FRs, and the CDRs and FRs are arranged from amino terminus to carboxy terminus in the order: FR1, CDR1, FR2, CDR2, FR3, CDR3, FR4. The constant regions of the heavy and light chains are not involved in antigen binding, but exhibit various effector functions. Antibodies are assigned to classes based on the amino acid sequence of the constant region of their heavy chain. The boundaries of the amino acid sequence of a particular CDR may be defined or identified by the conventions of any of the well-known systems, including IMGT numering, Kabat numbering and Chothia numbering, as described in Dev. Comp. Immunol., 27, 55-77 (IMGT numbering system) , Kabat et al. (1991) , Sequences of Protein of Immunological Interest, 5th ed., Public Health Service, National Institutes of Health, Bethesda, MD (Kabat numbering system) , Al-Lazikani et al., (1997) , JMB 273, 927-948 (Chothia numbering system) , Lefranc et al., (2003) , or combined system. Throughout this disclosure, wherein CDR sequences for an antibody are refered to, they are preferably defined by the IMGT CDR definition system, unless specified otherwise. However, it will be obvious to those skilled in the art to define the CDR sequences of an antibody from one definition system to another definition system, for example, from an IMGT CDR definition to a Kabat CDR defintion, or from an IMGT CDR definition to a Chothia CDR definition. Thus, wherein a CDR sequence for an antibody is provided based on a specific CDR definition, for example in IMGT CDR definition in this disclosure, it is understood that the CDR sequence is provided as one exemplary CDR sequence using the IMGT CDR definition for illustration and the disclosure does also imply and include any alternative CDR sequence defined using a different CDR definition, as it will be obvious for those skilled in the art to determine the alternative CDR sequences of the antibody using the other definition systems, based on the exemplary CDR sequence provided using the exemplary IMGT CDR definition.
[0079] In some embodiments, the antibody is an antigen-binding moiety. Antigen-binding moiety as used herein refers to an antibody fragment formed from a portion of an antibody comprising one or more CDRs, or any other antibody fragment that binds to an antigen but does not comprise an intact native antibody structure. Examples of antigen-binding moiety include, without limitation, a variable domain, a variable region, a diabody, a Fab, aFab', a F (ab') 2, an Fv fragment, a disulphide stabilized Fv fragment (dsFv) , a (dsFv) 2, a bispecific dsFv (dsFv-dsFv') , a disulphide stabilized diabody (ds diabody) , a multispecific antibody, a camelized single domain antibody, a nanobody, a domain antibody, and a bivalent domain antibody. An antigen-binding moiety is capable of binding to the same antigen to which the parent antibody binds. In certain embodiments, an antigen-binding moiety may comprise one or more CDRs from a particular human antibody grafted to a framework region from one or more different human antibodies. For more and detailed formats of antigen-binding moiety are described in Spiess et al, 2015 (Supra) , and Brinkman et al., mAbs, 9 (2) , pp. 182–212 (2017) , which are incorporated herein by reference.
[0080] The term “variable domain” or “variable region” with respect to an antibody as used herein refers to an antibody variable region or a fragment thereof comprising one or more CDRs. Although a variable domain or region may comprise an intact variable region (such as VH or VL) , it is also possible to comprise less than an intact variable region yet still retain the capability of binding to an antigen or forming an antigen-binding site.
[0081] A Fab fragment consists of a monovalent antigen-binding fragment of an antibody molecule, and can be produced by digestion of a whole antibody molecule with the enzyme papain, to yield a fragment consisting of an intact light chain and a portion of a heavy chain. A Fab' fragment of an antibody molecule can be obtained by treating a whole antibody molecule with pepsin, followed by reduction, to yield a molecule consisting of an intact light chain and a portion of a heavy chain. Two Fab' fragments are obtained per antibody molecule treated in this manner.
[0082] An Fv fragment is defined as a genetically engineered fragment containing the variable region of a light chain and the variable region of a heavy chain expressed as two chains.
[0083] A Fd chain refers to a fragment of antibody heavy chain comprising VH and CH1.
[0084] “Single-chain Fv antibody” or “scFv” refers to an engineered antibody consisting of a light chain variable region and a heavy chain variable region connected to one another directly or via a peptide linker sequence (Huston JS et al. Proc Natl Acad Sci USA, 85: 5879 (1988) ) .
[0085] As used herein, domain antibodies (dAbs) are the smallest known antigen-binding fragments of antibodies. VH dAb and VL dAb are the robust variable regions of the heavy and light chains of immunoglobulins (VH and VL respectively) .
[0086] As used herein, a VHH antibody (or nanobody) is the antigen binding fragment of heavy chain only antibodies. Heavy chain only antibodies (HcAb) are naturally produced by camelids and sharks. The antigen binding portion of the HcAb is comprised of the VHH fragment.
[0087] As used herein, a “single-chain domain” refers to an antigen-binding domain with one single polypeptide. Single-chain domain includes but not limits to scFv, VHH, VH dAb, VL dAb and et al.
[0088] As used herein, “anti-PD-L1 / PD-1 antibody” refers to an antibody that binds to PD-L1 or PD-1. “anti-PD-L1 / PD-1 single-chain domain” refers to a single-chain domain that binds to PD-L1 or PD-1. “anti-VEGF antibody” refers to an antibody that binds to VEGF. “anti-VEGF single-chain domain” refers to a single-chain domain that binds to VEGF.
[0089] As used herein, a “backbone” antibody of an antibody fusion protein refers to the antibody part comprising a heavy chain and a light chain to which an additional domain is linked to.
[0090] “Fc” with regard to an antibody refers to that portion of the antibody consisting of the second (CH2) and third (CH3) constant regions of a first heavy chain bound to the second and third constant regions of a second heavy chain via disulphide bonding. The Fc portion of the antibody is responsible for various effector functions such as ADCC, ADCP and CDC, but does not function in antigen binding. “CH2 constant region, ” which is also referred to as “CH2 domain, ” as used herein refers to the portion of a heavy chain molecule that extends, e.g., from about amino acid 244 to amino acid 360 of an IgG antibody using conventional numbering schemes (amino acids 244 to 360, Kabat numbering system; and amino acids 231-340, EU numbering system; see Kabat, E., et al., U.S. Department of Health and Human Services, (1983) ) . The “CH3 constant region, ” which is also referred to as “CH3 domain, ” extends from the CH2 domain to the C-terminus of the IgG molecule and comprises approximately 108 amino acids. Certain immunoglobulin classes, e.g., IgM, further include a CH4 region.
[0091] “Hinge region” in terms of an antibody includes the portion of a heavy chain molecule that joins the CH1 domain to the CH2 domain. This hinge region is flexible, thus allowing the two N-terminus antigen binding regions to move independently. In some embodiment, the hinge region comprises about amino acids 234 to 243 of Kabat numbering system. In some embodiment, the hing region comprises about amino acids 226 to 243 of Kabat numbering system. In some embodiment, a Fc comprising the hinge region comprising about amino acids 234 to 243 of Kabat numbering system.
[0092] An “antigen” or “Ag” as used herein refers to a compound, composition, peptide, polypeptide, protein, hapten, or substance that can stimulate the production of antibodies or a T cell response in cell culture or in an animal, including compositions (such as one that includes a cancer-specific protein) that are added to a cell culture (such as a hybridoma) , or injected or absorbed into an animal. An antigen reacts with the products of specific humoral or cellular immunity (such as an antibody) , including those induced by heterologous antigens. The term “fusion” or “fused” when used with respect to amino acid sequences (e.g. peptide, polypeptide, or protein) refers to combination of two or more amino acid sequences, for example by chemical bonding or recombinant means, into a single amino acid sequence that does not exist naturally. A fusion amino acid sequence may be produced by genetic recombination of two encoding polynucleotide sequences, and can be expressed by a method of introducing a construct containing the recombinant polynucleotides into a host cell.
[0093] The term "binding" as used herein refers to interaction of a binding domain with an antigen with the interaction depending upon the presence of a particular structure (e.g., an antigenic determinant or epitope) on the antigen. For example, an antibody variable region or Fv recognizes and binds to a specific protein structure rather than to protein generally. As used herein, the term "specifically binding" or "binding specifically" means that a binding domain binds to or associates with more frequently, more rapidly, with greater duration and / or with greater affinity with a particular antigen than with other protein. For example, an antibody variable region or Fv specifically binds to its antigen with greater affinity, avidity, more readily, and / or with greater duration than it binds to other antigens. For another example, an antibody variable region or Fv binds to a cell surface protein (antigen) with materially greater affinity than it does to related protein or other cell surface protein or to antigens commonly recognized by polyreactive natural antibodies (i.e., by naturally occurring antibodies known to bind a variety of antigens naturally found in humans) . However, "specifically binding" does not necessarily require exclusive binding or non-detectable binding of another antigen, this is meant by the term "selective binding" .
[0094] The term “affinity” as used herein refers to the strength of the sum total of noncovalent interactions between a single binding site of a molecule (e.g., an antibody) and its binding partner (e.g., an antigen) . Unless indicated otherwise, as used herein, “binding affinity” refers to intrinsic binding affinity which reflects a 1: 1 interaction between members of a binding pair (e.g., antibody and antigen) . The affinity of a molecule X for its partner Y can generally be represented by the dissociation constant (Kd) .
[0095] The term “avidity” refers to the strength of the sum total of noncovalent interactions of two or more antigen binding sites and their binding partner (e.g., an antigen) .
[0096] The term “valency” refers to the number of a specific domain that a molecule consists of. For example, “monovalent” “bivalent” , “trivalent” or “tetravalent” refers to respectively one, two, three or four such specific domains comprised by the multispecific molecule in total.
[0097] The term “identity, ” as used herein, refers to a relationship between the sequences of two or more polypeptide molecules or two or more nucleic acid molecules, as determined by aligning and comparing the sequences. “Percent identity” means the percent of identical residues between the amino acids or nucleotides in the compared molecules and is calculated based on the size of the smallest of the multispecific molecules being compared. For these calculations, gaps in alignments (if any) are preferably addressed by a particular mathematical model or computer program (i.e., an “algorithm” ) . Methods that can be used to calculate the identity of the aligned nucleic acids or polypeptides include those described in Computational Molecular Biology, (Lesk, A.M., ed. ) , 1988, New York: Oxford University Press; Biocomputing Informatics and Genome Projects, (Smith, D.W., ed. ) , 1993, New York: Academic Press; Computer Analysis of Sequence Data, Part I, (Griffin, A.M., and Griffin, H. G., eds. ) , 1994, New Jersey: Humana Press; von Heinje, G., 1987, Sequence Analysis in Molecular Biology, New York: Academic Press; Sequence Analysis Primer, (Gribskov, M. and Devereux, J., eds. ) , 1991, New York: M. Stockton Press; and Carillo et al, 1988, SIAMJ. Applied Math. 48: 1073.
[0098] The term “effector functions” as used herein refer to those biological activities attributable to the Fc region of an antibody, which vary with the antibody isotype. Examples of antibody effector functions include: Clq binding and complement dependent cytotoxicity (CDC) ; Fc receptor binding and antibody-dependent cell-mediated cytotoxicity (ADCC) ; Fc receptor binding and antibody-dependent cell-mediated phagocytosis (ADCP) ; down regulation of cell surface receptors (e.g. B cell receptor) ; activation of cells expressing Fc receptor (e.g. B cells, macrophages, dendritic cells) , induction of secretion of cytokines and / or chemokines by cells expressing Fc receptor (e.g. B cells, macrophages, dendritic cells) .
[0099] As used herein, a “silent Fc” refers to an Fc with no or weak effector functions. A “active Fc” refers to an Fc with active effector functions such as ADCC, ADCP and / or CDC.
[0100] The term “phagocytosis” refers to a process by which a substance with size (such as a cell, a fragment of cell, a microbe, or a particle) is engulfed and internalized by a cell. The term “phagocyte” refers to a cell that is capable of phagocytosis.
[0101] As used herein, the term “SIRP” means “Signal regulatory protein” , including SIRPα, SIRPβ, and SIRPγ, which are a family of transmembrane glycoprotein with three extracellular Ig-like domains, including one IgV domain at the N-terminal followed with two IgC domains. The IgV extracellular domain serves as the direct CD47-binding domain of the SIRP protein (Barclay et al., Nat Rev Immunol, 2006) . As used herein, the term “SIRP IgV” means the IgV extracellular domain of the SIRP protein, including the IgV extracellular domain of SIRPα, SIRPβ, and SIRPγ.
[0102] SIRPαcomprises 10 members, namely SIRPαV1 to V10, whose IgV domains (SEQ ID NO: 57 to 65) all bind to CD47. SIRPαV1, V2 and V8 are reported to be most prevalent SIRPαvariants in human (Voets et al., J Immunother Cancer, 2019) . The IgV domain sequences of human SIRPαV1 and SIRPαV2 comprise 118 and 119 amino acids respectively, and they differ in 13 amino acids (as shown in FIG. 6) (Hatherley et al., JBiol Chem, 2014) . Notably, the IgV sequences of human SIRPαV5, V6 and V9 variants each only differ in one amino acid from that of human SIRPαV1, and the IgV sequence of human SIRPαV3 and V7 variants each only differ in one amino acid from that of human SIRPαV2, while human SIRPαV10 only differs in one amino acid in the signal peptide sequence from that of human SIRPαV2 (as shown in FIG. 6) (Hatherley et al., J Biol Chem, 2014) . Meanwhile, human SIRPαV4 and SIRPαV8 variants differ from both human SIRPαV1 and SIRPαV2, comprising a mixture of the 13 amino acid differing between human SIRPαV1 and SIRPαV2 (as shown in FIG. 6) ) (Hatherley et al., JBiol Chem, 2014) .
[0103] SIRPβcomprises SIRPβ1 (NP_006056.2, SEQ ID NO: 55) and SIRPβ2 (NP_001129316.1, SEQ ID NO: 56)(Hatherley et al., Mol Cell, 2008) . The native IgV domain of SIRPβ1 shows no detectable CD47 binding, but mutation of 2 amino acids of SIRPβ1 to the amino acids at the corresponding positions of SIRPαV2 (M27V+M37Q as shown in FIG. 6) results in CD47-binding, with additional mutations that could further increases its CD47 binding, as described in the prior arts (Hatherley et al., Mol Cell, 2008, Lee et al., JImmunol, 2007, Liu et al., JMol Biol, 2007) , hereby incorporated by reference. Similarly, mutation of 1 amino acid in the IgV domain of SIRPβ2 to the amino acid at the corresponding position of SIRPαV2 (H101D as shown in FIG. 6) results in CD47-binding of the variant SIRPβ2 IgV domain (Hatherley et al., Mol Cell, 2008) .
[0104] SIRPγ (NP_061026.2, SEQ ID NO: 72) binds to CD47 about ten times weaker than SIRPα, however mutation of the 2 differing amino acids in SIRPγIgV domain to the amino acid at corresponding positions of SIRPαV2 (L37Q and / or N101D as shown in FIG. 6) increases CD47-binding of the SIRPγIgV domain (Hatherley et al., Mol Cell, 2008) (US9845345B2) . See Table 1 for such SIRPγsequences (SEQ ID NO: 73 to 75) . Table 1. The amino acid sequences of SIRP IgV *WT: wild type
[0105] As used herein, the term “SIRP IgV monomer” means one single IgV extracellular domain of a SIRP protein, including one single IgV extracellular domain of SIRPα, SIRPβ, or SIRPγ. A wild type SIRP IgV monomer of SIRPα, SIRPβ, or SIRPγcomprises an amino acid sequence selected from the group consisting of SEQ ID NO: 55 to 65 and 72. Optionally, a SIRP IgV monomer may additionally comprise one or both of the two IgC domains of the full extracellular domain of a SIRPα, SIRPβ, or SIRPγprotein.
[0106] As used herein, the term “SIRP IgV multimer polypeptide” refers to a polypeptide comprising two or more SIRP IgV monomers serially linked together in the polypeptide preferably with a linker between the two neighboring SIRP IgV monomers.
[0107] As used herein, the term “SIRP IgV domain” refers to a structural or a spatial domain that comprises at least one SIRP IgV monomer. A SIRP IgV domain may comprise one or more SIRP IgV monomers, and the SIRP IgV monomers may be linked together in one polypeptide, or separately located in different polypeptides.
[0108] The term "microenvironment" as used herein means any portion or region of a tissue, organ or body that has constant or temporal, physical or chemical differences from other regions of the tissue, organ or regions of the body. For tumors, the term “tumor microenvironment” as used herein refers to the environment in which a tumor exists, which is the non-cellular area within the tumor and the area directly outside the tumorous tissue but does not pertain to the intracellular compartment of the cancer cell itself. The tumor and the tumor microenvironment are closely related and interact constantly. A tumor can change its microenvironment, and the microenvironment can affect how a tumor grows and spreads. Typically, the tumor microenvironment has a low pH in the range of 5.0 to 6.8, or in the range of 5.8 to 6.8, or in the range of 6.2-6.8. The tumor microenvironment has been discussed in (Gillies et al., J Magn Reson Imaging, 2002) , hereby incorporated by reference here. The term “non-tumor microenvironment” refers to a microenvironment at a site other than a tumor.
[0109] The term “cytotoxic agent” as used herein refers to a substance that inhibits or prevents a cellular function and / or causes cell death or destruction. The term “cytostatic agent” as used herein refers to a compound or composition which arrests growth of a cell either in vitro or in vivo. Thus, a cytostatic agent may be one which significantly reduces the percentage of cells in S phase. The term “chemotherapeutic agent” as used herein refers to a chemical compound useful in the treatment of cancer.
[0110] The term “vector” as used herein refers to a nucleic acid molecule capable of propagating another nucleic acid to which it is linked. The term includes the vector as a self-replicating nucleic acid structure as well as the vector incorporated into the genome of a host cell into which it has been introduced. Certain vectors are capable of directing the expression of nucleic acids to which they are operatively linked. Such vectors are referred to herein as “expression vectors. ”
[0111] The terms “host cell, ” “host cell line, ” and “host cell culture” as used herein are used interchangeably and refer to cells into which exogenous nucleic acid has been introduced, including the progeny of such cells.
[0112] As used herein, the term “subject” is used interchangeably with “patient” and may be a mammal who is in need of prevention or treatment of cancer, such as primates (for example, humans) , companion animals (for example, dogs and cats) , livestock (for example, cows, pigs, horses, sheep, and goats) , and laboratory animals (for example, rats, mice, and guinea pigs) . In an embodiment of the present disclosure, the subject is a human.
[0113] As used herein, the term “treatment” generally means obtaining a desired pharmacological and / or physiological effect. The effect may be therapeutic in terms of partially or completely curing a disease and / or adverse effect attributed to the disease. Desirable therapeutic effects include, but are not limited to, prevention of onset or recurrence of disease, alleviation of symptoms, diminishment of any direct or indirect pathological consequences of the disease, prevention of metastasis, decreasing the rate of disease progression, amelioration or slowing of the disease state, and remission or improved prognosis. Preferably, the “treatment” may refer to medical intervention of a disease or disorder that has already developed.
[0114] As used herein, the term “prevention” relates to a prophylactic treatment, that is, to a measure or procedure, the purpose of which is to prevent, rather than to cure a disease. “Prevention” means that a desired pharmacological and / or physiological effect is obtained which is prophylactic in terms of completely or partially preventing a disease or symptom thereof. As used herein, “preventing or treating cancer” may include inhibiting proliferation, survival, metastasis, recurrence, or therapy resistance of cancer. Such a method may comprise a step of administering the immune cells of the present disclosure to a subject in need of prevention or treatment of cancer. Accordingly, there is provided a use of a composition that comprises the immune cells as an active ingredient, for preventing or treating cancer.
[0115] As used herein, the term “administration” means providing a substanceto a subject to achieve a prophylactic or therapeutic purpose (for example, prevention or treatment of cancer) .
[0116] As used herein, the term “cancer” refers to a physiological condition that is typically characterized by unregulated cell growth in mammals. The cancer to be prevented or treated in the present disclosure may include, depending on the site of occurrence, colorectal cancer, small intestine cancer, rectal cancer, colon cancer, thyroid cancer, endocrine adenocarcinoma, oral cancer, tongue cancer, pharyngeal cancer, laryngeal cancer, esophageal cancer, cervical cancer, uterine cancer, fallopian tube cancer, ovarian cancer, brain cancer, head and neck cancer, lung cancer, lymph gland cancer, gallbladder cancer, bladder cancer, kidney cancer, liver cancer, pancreatic cancer, prostate cancer, skin cancer (or melanoma) , breast cancer, stomach cancer, bone cancer, blood cancer, and the like. However, any cancer can be included therein as long as it expresses an antigen protein on the surface of cancer cells. In an embodiment, the cancer may include at least any one selected from the group consisting of optionally colorectal cancer, rectal cancer, colon cancer, thyroid cancer, oral cancer, pharyngeal cancer, laryngeal cancer, cervical cancer, brain cancer, lung cancer, ovarian cancer, bladder cancer, kidney cancer, liver cancer, pancreatic cancer, prostate cancer, skin cancer, tongue cancer, breast cancer, uterine cancer, stomach cancer, bone cancer, and blood cancer. In another embodiment, the cancer may be a solid cancer.
[0117] The term “pharmaceutically acceptable, ” as used herein, means that the vehicle, diluent, excipient and / or salts thereof, are chemically and / or physically is compatible with other ingredients in the formulation, and the physiologically compatible with the recipient.
[0118] As used herein, the term “a pharmaceutically acceptable carrier and / or excipient” refers to a carrier and / or excipient pharmacologically and / or physiologically compatible with a subject and an active agent, which is well known in the art (see, e.g., Remington's Pharmaceutical Sciences. Edited by Gennaro AR, 19th ed. Pennsylvania: Mack Publishing Company, 1995) , and includes, but is not limited to pH adjuster, surfactant, adjuvant and ionic strength enhancer. For example, the pH adjuster includes, but is not limited to, phosphate buffer; the surfactant includes, but is not limited to, cationic, anionic, or non-ionic surfactant, e.g., Tween-80; the ionic strength enhancer includes, but is not limited to, sodium chloride.
[0119] The term “therapeutically effective amount” of a therapeutic agent or treatment is meant a sufficient amount of therapeutic agent or treatment to have a therapeutic effect in the subject treated, at a reasonable benefit / risk ratio applicable to any medical treatment. It will be understood, however, that the total daily usage of therapeutic agent or treatment will be decided by the attending physician within the scope of sound medical judgment. The specific therapeutically effective dose level for any particular patient will depend upon a variety of factors including the disorder being treated and the severity of the disorder; activity of the specific antibody employed; the specific composition employed, the age, body weight, general health, sex and diet of the patient; the time of administration, route of administration, and rate of excretion of the specific antibody employed; the duration of the treatment; drugs used in combination or coincidental with the specific antibody employed; and like factors well known in the medical arts. For example, it is well known within the skill of the art to start doses of the compound at levels lower than those required to achieve the desired therapeutic effect and to gradually increase the dosage until the desired effect is achieved. Exemplary Embodiments Embodiment Set 1: Construction of bispecific antibody fusion protein against PD-L1 / PD-1 and VEGF using VEGFR domains
[0120] To construct antibody fusion protein against PD-L1 / PD-1 and VEGF, the second extracellular Ig-like domain of human VEGFR1 (VEGFR-R1D2, SEQ ID NO: 17) or a chimeric receptor of the second extracellular Ig-like domain of human VEGFR1 and the third extracellular Ig-like domain of human VEGFR2 (VEGFR-R1D2-R2D3, SEQ ID NO: 18) is fused to the N-terminal or C-terminal of the heavy chain or light chain of an anti-PD-L1 / PD-1 antibody. Exemplary bispecific antibody fusion protein PV3 to PV30 and PV113 to PV118 are illustrated in Table 2 (see next page) . DNAs encoding the full heavy chain polypeptides and the full light chain polypeptides of PV1 to PV30 and PV113 to PV118 are optimized for expression in Homo sapiens, synthesized and cloned into the same pcDNA3.4 mammalian expression vector using standard molecular cloning techniques, as the heavy chain expression vector and light chain expression vector respectively. The antibody fusion protein are then produced through transient co-expression of the heavy chain and light chain vectors at 2: 3 ratio in HEK293 cells and single step of Protein A chromatography purification. The purified antibodies are analyzed by SDS-PAGE and SEC-HPLC.
[0121] In contrast to what reported in US11407832B2 and US12269885B2 that shows preferably linking the VEGFR-R1D2 domain to the N-terminal of the heavy chain of two different anti-PD-L1 antibody (IMM25 and 900339) gives better or at least comparable biophysical property and VEGF binding activity of the resultant anti-PD-L1xVEGF antibody fusion protein than linking the same VEGFR-R1D2 domain to the C-terminal of the heavy chain of IMM25 and 900339 respectively (see US11407832B2 and US12269885B2) , linking the same VEGFR-R1D2 domain to the C-terminal of the heavy chain of the anti-PD-L1 antibody PV1 or PV2 surprisingly gives better biophysical property and VEGF binding activity of the resultant anti-PD-L1 x VEGF antibody fusion protein than linking the same VEGFR-R1D2 domain to the N-terminal of the heavy chain of PV1 or PV2. In addition, in contrast to what reported in US11407832B2 and US12269885B2 that shows when linking the VEGFR1-R1D2 domain to the C-terminal of the heavy chain of the anti-PD-L1 antibody IMM25 and 900339, alinker of 10 amino acids in length compromises the biophysical property and VEGF binding of the resultant antibody fusion protein, a linker of≤10 amino acids in length, such as 8 amino acids or 5 amino acids, still maintains comparable biophysical property and / or VEGF binding activity of the resultant antibody fusion protein when linking the VEGFR1-R1D2 domain to the C-terminal of the heavy chain of the anti-PD-L1 antibody PV1 or PV2 specifically. Moreover comparing to a longer linker, a shorter linker linking the VEGFR-R1D2 domain to the C-terminal of two heavy chains of an antibody is able to favorably increase the formation of multimer complex between the resultant antibody fusion protein and dimeric VEGF. Embodiment Set 2: Construction of bispecific antibody fusion protein against PD-L1 / PD-1 and VEGF using ranibizumab / bevacizuamb derived anti-VEGF domains
[0122] In addition to VEGFR domain, anti-VEGF scFv derived from the anti-VEGF antibody ranibizumab or bevacizumab are tested to construct bispecific antibody fusion protein against PD-L1 / PD-1 and VEGF, by linking anti-VEGF scFv to the C-terminal or N-terminal of the heavy chain or light chain of an anti-PD-L1 / PD-1 antibody. Exemplary bispecific antibody fusion protein PV31 to PV77 and PV119 to PV121 are illustrated in Table 3. Aternatively, bispecific antibody fusion protein against PD-L1 / PD-1 and VEGF are constructed by linking single-chain anti-PD-L1 / PD-1 domain to the C-terminal or N-terminal of the heavy chain or light chain of the anti-VEGF antibody bevacizumab. Exemplary bispecific antibody fusion protein PV78 to PV88 and PV122 to PV126 are illustrated in Table 4. Similarly as in Embodiment Set 1, DNAs encoding the full heavy chain polypeptides and the full light chain polypeptides of PV31 to PV88 and PV119 to PV126 are optimized for expression in Homo sapiens, synthesized and cloned into the same pcDNA3.4 mammalian expression vector using standard molecular cloning techniques, as the heavy chain expression vector and light chain expression vector respectively. The antibody fusion protein are then produced through transient co-expression of the heavy chain and light chain vectors at 2: 3 ratio in HEK293 cells and single step of Protein A chromatography purification. The purified antibodies are analyzed by SDS-PAGE and SEC-HPLC.
[0123] Among the various configurations of bevacizumab-derived scFv and ranibizumab-derived scFv tested, antibody fusion protein comprising the ranibizumab scFv1 (SEQ ID NO: 19) , ranibizumab scFv2 (SEQ ID NO: 21) , bevacizumab scFv1 (SEQ ID NO: 20) or bevacizumab scFv2 (SEQ ID NO: 22) as listed in Table 3 show poor production yield and modest purity. However, the ranibizumab scFv3 (SEQ ID NO: 25) , the stapled bevacizumab scFv3 (SEQ ID NO: 23) and the stapled bevacizumab scFv4 (SEQ ID NO: 24) show improved production yield and high purity. The bevacizumab scFv5 (SEQ ID NO: 26, especially linked at C-terminal of antibody heavy chain) and bevacizumab scFv6 (SEQ ID NO: 453) also show some improved production yield.
[0124] Among the various configurations of BMS936559-derived scFv tested, antibody fusion protein comprising the anti-PD-L1 scFv1 (SEQ ID NO: 44) as listed in Table 4 shows negligible production yield. The anti-PD-L1 scFv2 (SEQ ID NO: 45) shows low production yield and purity, and the anti-PD-L1 scFv3 (SEQ ID NO: 46) shows some improved yield. Interesting, the anti-PD-L1 scFv8 (SEQ ID NO: 455) shows high purity but low yield when linked to the N-terminal of antibody heavy chain, however shows both good production yield and purity when linked to the C-terminal of antibody heavy chain. Meanwhile, it binds to PD-L1 when linked to either the C-terminal or N-terminal of antibody heavy chain. Table 2. Construction of bispecific antibody fusion protein against PD-L1 / PD-1 and VEGF using VEGFR domains Table 3. Construction of bispecific antibody fusion protein against PD-L1 / PD-1 and VEGF using ranibizumab / bevacizuamb derived anti-VEGF scFv domain Table 4. Construction of bispecific antibody fusion protein against PD-L1 / PD-1 and VEGF using bevacizumab as backbone antibody Embodiment Set 3. Construction of bispecific antibody fusion protein against PD-L1 / PD-1 and VEGF using non-bevacizuamb derived single chain anti-VEGF domains
[0125] Additional anti-VEGF single-chain domains including VL dAb (SEQ ID NO: 27) , VH dAb (SEQ ID NO: 28) , VHH (SEQ ID NO: 29) and non-bevacizumab / ranibizumab-derived scFv (SEQ ID NO: 30) are also tested to construct bispecific antibody fusion protein against PD-L1 / PD-1 and VEGF, by linking the anti-VEGF single-chain domain to the C-terminal or N-terminal of the heavy chain or light chain of an anti-PD-L1 / PD-1 antibody. Exemplary bispecific antibody fusion protein PV89 to PV112 are illustrated in Table 5 (see next page) . Similarly as in Embodiment Set 1 and 2, DNAs encoding the full heavy chain polypeptides and the full light chain polypeptides of PV89 to PV112 are optimized for expression in Homo sapiens, synthesized and cloned into the same pcDNA3.4 mammalian expression vector using standard molecular cloning techniques, as the heavy chain expression vector and light chain expression vector respectively. The antibody fusion protein are then produced through transient co-expression of the heavy chain and light chain vectors at 2: 3 ratio in HEK293 cells and single step of Protein A chromatography purification. The purified antibodies are analyzed by SDS-PAGE and SEC-HPLC. Embodiment Set 4. Construction of trispecific antibody fusion protein against PD-L1 / PD-1, VEGF and CD47
[0126] On the basis of the bispecific antibody fusion protein against PD-L1 / PD-1 and VEGF constructed in Embodiment Set 1-3 above, trispecific antibody fusion protein against PD-L1 / PD-1, VEGF and CD47 are further constructed by linking additional CD47-binding SIRP IgV domain to a bispecific antibody fusion protein against PD-L1 / PD-1 and VEGF in various formats, as illustrated in format variants FV10 to FV38 (FIG. 2 to 4) . Preferably, the trispecific antibody fusion protein comprises a Fc of human IgG4 type or silent human IgG1 type. A CD47-binding SIRP IgV domain comprises an IgV extracellular domain derived from a Signal-regulatory protein (SIRP) family protein selected from the group consisting of SIRPα, SIRPβand SIRPγ. A CD47-binding SIRP IgV domain comprises an amino acid sequence selected from the group consisting of SEQ ID NO: 57-75. The wild type SIRPαV2 IgV domain is used to construct exemplary trispecific antibody fusion protein PVC1 to PVC34 and PVC43 to PVC66 as illustrated in Table 6 and 7 respectively. Similarly as in Embodiment Set 1, 2 and 3, DNAs encoding the full heavy chain polypeptides and the full light chain polypeptides of PVC1 to PVC34 and PVC43 to PVC66 are optimized for expression in Homo sapiens, synthesized and cloned into the same pcDNA3.4 mammalian expression vector using standard molecular cloning techniques, as the heavy chain expression vector and light chain expression vector respectively. The antibody fusion protein are then produced through transient co-expression of the heavy chain and light chain vectors at 2: 3 ratio in HEK293 cells and single step of Protein A chromatography purification. The purified antibodies are analyzed by SDS-PAGE and SEC-HPLC. Table 5. Construction of bispecific antibody fusion protein against PD-L1 / PD-1 and VEGF using additional non-bevacizuamb derived anti-VEGF single-chain domain Table 6. Construction of trispecific antibody fusion protein against PD-L1 / PD-1, VEGF and CD47 using anti-PD-L1 / PD-1 antibody as backbone antibody Table 7. Construction of trispecific antibody fusion protein against PD-L1 / PD-1, VEGF and CD47 using anti-VEGF antibody as backbone antibody Embodiment Set 5. Construction of trispecific antibody fusion protein against PD-L1 / PD-1, VEGF and CD47 with SIRP variants with enhanced CD47-binding
[0127] On the basis of the trispecific antibody fusion protein against PD-L1 / PD-1, VEGF and CD47 constructed in Embodiment Set 4 above, additional mutation is introduced into the SIRP IgV domain of the trispecific antibody fusion protein that further enhances the binding of the SIRP IgV variant domain to CD47 over the parent SIRP IgV domain without the mutation. Such SIRP IgV variant domain of the trispecific antibody fusion protein comprises a SIRPαIgV domain or SIRPβIgV domain comprising one or more mutation selected from the group consisting of I31Y, I31W, R59H and N80A, or comprises a SIRPγIgV domain comprising one or more mutation selected from the group consisting of L31Y, L31W and R59H. Alternatively, such SIRP IgV variant domain of the trispecific antibody fusion protein comprises a SIRPα IgV domain comprising a mutation of V6I+A27I+I31F+E47V+K53R+E54S+H56P+L66T+V92I, or a mutation of I31L+E47Q+E54D+R77N+V92I, or a mutation of I31F+K53R+E54S+H56P+S66T+N80A, or a mutation of I31L+V33I+E47V+K53R+E54N+V63I+K68R+R77K. As an example, a SIRPαV2 IgV variant domain comprising a mutation of I31Y or I31W are used to construct exemplary trispecific antibody fusion protein PVC67 to PVC108, as illustrated in Table 8. Exemplary trispecific antibody fusion protein comprising SIRPαV2 IgV variant domain comprising a mutation of R59H, R59H+I31Y or R59H+I31W can also be straightforwardly constructed by replacing the mutation of I31Y or I31W with R59H, R59H+I31Y or R59H+I31W in the SIRPαV2 IgV variant domain of the exemplary trispecific antibody fusion protein PVC67 to PVC108 illustrated in Table 8. Similarly as in Embodiment Set 1, 2 and 3, DNAs encoding the full heavy chain polypeptides and the full light chain polypeptides of PVC67 to PVC108 are optimized for expression in Homo sapiens, synthesized and cloned into the same pcDNA3.4 mammalian expression vector using standard molecular cloning techniques, as the heavy chain expression vector and light chain expression vector respectively. The antibody fusion protein are then produced through transient co-expression of the heavy chain and light chain vectors at 2: 3 ratio in HEK293 cells and single step of Protein A chromatography purification. The purified antibodies are analyzed by SDS-PAGE and SEC-HPLC. Table 8. Construction of trispecific antibody fusion protein against PD-L1 / PD-1, VEGF and CD47 with SIRP variants with enhanced CD47-binding 8A 8B Embodiment Set 6. Construction of trispecific antibody fusion protein against PD-L1 / PD-1, VEGF and CD47 with SIRP variants with higher binding at an acidic pH than at physiological pH
[0128] On the basis of the trispecific antibody fusion protein against PD-L1 / PD-1, VEGF and CD47 constructed in Embodiment Set 4 above, additional mutation is introduced into the SIRP IgV domain of the trispecific antibody fusion protein that further renders the SIRP IgV variant domain with higher binding to CD47 at an acidic pH than at a physiological pH (pH 7.3 to 7.5) . Such SIRP IgV variant domain of the trispecific antibody fusion protein comprises a SIRPαIgV domain or SIRPβIgV domain comprising a mutation selected from the group consisting of K53H, R69H, K68H, Q52H, I31E, I31D, Q52H+K68H, K53H+K68H, R69H+K68H, K53H+Q52H, R69H+Q52H, K53H+I31E, R69H+I31E, K68H+I31E, Q52H+I31E, K53H+K68H+Q52H, R69H+K68H+Q52H, Q52H+K68H+I31E, K53H+I31D, R69H+I31D, K68H+I31D, Q52H+I31D, and Q52H+K68H+I31D, or comprises a SIRPγIgV domain comprising a mutation selected from the group consisting of K53H, R69H, K68H, Q52H, L31E, L31D, Q52H+K68H, K53H+K68H, R69H+K68H, K53H+Q52H, R69H+Q52H, K53H+L31E, R69H+L31E, K68H+L31E, Q52H+L31E, K53H+K68H+Q52H, R69H+K68H+Q52H, Q52H+K68H+L31E, K53H+L31D, R69H+ L31D, K68H+L31D, Q52H+L31D, Q52H+K68H+L31D, K53H+N101D, K53H+N101D+L37Q, K53H+N101D+L31E and K53H+N101D+L31D. Preferably, such a trispecific antibody fusion protein comprises a Fc of active human IgG1 type or human IgG4 type. As an example, a SIRPαV2 IgV variant domain comprising a mutation of K53H, R69H, Q52H, K68H, K53H+I31E or R69H+I31E are used to construct exemplary trispecific antibody fusion protein PVC109 to PVC186, as illustrated in Table 9. Similarly as in Embodiment Set 1, 2 and 3, DNAs encoding the full heavy chain polypeptides and the full light chain polypeptides of PVC109 to PVC186 are optimized for expression in Homo sapiens, synthesized and cloned into the same pcDNA3.4 mammalian expression vector using standard molecular cloning techniques, as the heavy chain expression vector and light chain expression vector respectively. The antibody fusion protein are then produced through transient co-expression of the heavy chain and light chain vectors at 2: 3 ratio in HEK293 cells and single step of Protein A chromatography purification. The purified antibodies are analyzed by SDS-PAGE and SEC-HPLC. Table 9. Construction of trispecific antibody fusion protein against PD-L1 / PD-1, VEGF and CD47 with SIRP variants with higher binding at acidic pH than at physiological pH. 9A 9B EXAMPLES
[0129] The following examples are included to further describe some embodiments of the present disclosure. The examples are illustrative, but not limiting the scope of the disclosure. Other suitable modifications and adaptations of the variety of conditions and parameters normally encountered in the field, and which are obvious to those skilled in the art, are within the scope of the disclosure. General Methods 1) Construction of the plasmids
[0130] The DNA encoding the protein sequences is optimized for expression in Homo sapiens, synthesized and cloned into the pcDNA3.4 mammalian expression vector using standard molecular cloning techniques. For protein without Fc, the protein polypeptide is cloned into a pcDNA3.4 expression vector with a 6x His-tag added to the C-terminal of the protein polypeptide for purification through Ni-NTA column. 2) . Protein expression and purification
[0131] The plasmids encoding the polypeptides of a given protein is co-transfected into HEK293 cells usually using polyethylenimine (PEI) . For antibody protein with heterodimeric knobs-into-holes Fc, a1: 1 ratio of knob: hole heavy chain and a 3: 2 ratio of light chain: heavy chain are usually used for the plasmids co-transfection. HEK293 cells are grown in a humidified incubator at 37℃ with 8%CO2 for 5-7 days after transfection. The cells are pelleted by centrifugation at 8000 rpm for 5 minutes and the supernatant is filtered through a 0.2μm membrane. Pre-assembled Protein A resin column or Ni-NTA resin column is first equilibrated with 1x PBS and the cell supernatant is then loaded through the column. Protein A resin column is washed with 1x PBS and Ni-NTA resin column is washed with 2.5mM imidazole buffer. The bound protein is then eluted from the column with acidic citrate elution buffer (pH 3.4) for Protein A resin column or 250mM imidazole buffer for Ni-NTA resin column. The collected protein is dialyzed into 1x PBS. 3) ELISA assay
[0132] ELISA assays are performed using 96-well ELISA plate coated with the capture protein in 1x PBS / well at 4℃overnight. After washing with 1x PBS+0.05%Tween 20 (PBST) , the plate is blocked with 1x PBS+5%milk for 2 hours at room temperature (RT) . After washing the plate with 1x PBST, serially diluted concentrations of the analyte in 1x PBS+1%milk are added and incubated for 60 minutes at RT. After washing with 1x PBST, 1: 4000 diluted anti-His-HRP (Proteintech, CAT#HRP-66005) or anti-human IgG Fc-HRP (Abcam, CAT#ab97225) in 1x PBS+1%milk is added and incubated for 60 minutes at RT. After washing with 1x PBST, 30 μl / well tetramethylbenzidine (TMB) substrate solution is added and incubated in dark at room temperature for several minutes. 30μl / well H2SO4 is then added to stop the reaction, and OD450 is read out at 450 nm using a microplate reader and data are analyzed by Graphpad Prism 9. 4) . FACS assay
[0133] Target cells of interest are grown into proper confluence and then harvested according to standard procedures. The harvested cells are centrifuged and the cell pellet is resuspended into proper concentration in the cell culture media and incubated at 37℃for 30 minutes. Serially diluted test article is then added into the cells to designated final concentration and incubated at 4℃for 30 minutes. The cells are then washed for three times, and fluorescence labeled secondary antibody (anti-human IgG Fc, Abcam#, ab98596, 1: 300) is added into the cells and incubated at 4℃for 30 min. After wash, the cells are resuspended in FACS buffer and analyzed by Flow Cytometry. Example 1. Construction of bispecific antibody fusion protein against PD-L1 and VEGF using VEGFR domains
[0134] To construct antibody fusion protein against PD-L1 and VEGF, the second extracellular Ig-like domain of human VEGFR1 (VEGFR-R1D2) was fused to the N-terminal or C-terminal of the heavy chain of anti-PD-L1 antibodies. First, two bispecific antibody fusion protein PV3 and PV4were constructed by linking VEGFR-R1D2 (SEQ ID NO: 17) to the N-terminal of the heavy chain of anti-PD-L1 antibody PV1 (analog of BMS-936559) and PV2 (analog of Durvalumab) respectively through a G4Sx4 linker, as detailed in Table 10. DNAs encoding the full heavy chain polypeptides and the full light chain polypeptides of PV1, PV2, PV3 and PV4 were optimized for expression in Homo sapiens, synthesized and cloned into the same pcDNA3.4 mammalian expression vector using standard molecular cloning techniques, as the heavy chain expression vector and light chain expression vector respectively. The antibody fusion protein were then produced through transient co-expression of the heavy chain and light chain vectors at 2: 3 ratio in HEK293 cells and single step of Protein A chromatography purification. The purified antibodies were analyzed by SDS-PAGE and SEC-HPLC. Both the anti-PD-L1 antibody alone PV1 and PV2 showed good yield of 100mg / L and 120mg / L respectively, however the antibody fusion protein PV4 showed surprisingly low yield of only 16mg / L, whereas PV3 showed a high yield of 213mg / L. Meanwhile, all four protein showed>95%purity in SEC-HPLC analysis.
[0135] Additional bispecific antibody fusion protein PV7 was also constructed by linking VEGFR-R1D2 (SEQ ID NO: 17) to the C-terminal of the heavy chain of anti-PD-L1 antibody PV1 through a G4Sx3 linker of 15 amino acid length. PV7 was produced with high yield and purity (>95%SEC-HPLC purity) , similar to that of PV3 in which the VEGFR-R1D2 was linked to the N-terminal of the heavy chain of PV1 (Table 10) . Also as reported in US12269885B2, when linking the same VEGFR-R1D2 to the C-terminal of the heavy chain of a different anti-PD-L1 antibody through a G3Sx4-GS linker of 14 amino acid length to produce the bispecific antibody fusion protein 900388, high purity (>95%SEC-HPLC purity) was shown, similar to that of the bispecific antibody fusion protein 900387 in which the same VEGFR-R1D2 was linked to the N-terminal of the heavy chain of the same anti-PD-L1 antibody. However as reported in US11407832B2, when linking the same VEGFR-R1D2 to the C-terminal of the heavy chain of another anti-PD-L1 antibody IMM25 through a shorter G4Sx2 linker of 10 amino acid length to produce the bispecific antibody fusion protein IMM25011, a lower purity (only~85%SEC-HPLC purity) was shown than that of the bispecific antibody fusion protein IMM2510 produced by linking the same VEGFR-R1D2 to the N-terminal of the heavy chain of the same anti-PD-L1 antibody. Moreover, IMM25011 showed lower binding to VEGF than IMM2510, whereas 900388 showed similar binding to VEGF as 900387 (see US11407832B2 and US12269885B2) , indicating both biophysical and functional compromise of the antibody fusion protein when a linker less than 14 amino acids in length was used to linking VEGF-R1D2 to the C-terminal of an antibody heavy chain. The VEGFR-R1D2 domain was also linked to the C-terminal of the heavy chain of PV1 through shorter linkers of 8 amino acids (G4Sx1-GGG) or only 5 amino acids (G4Sx1) to produce the bispecific antibody fusion protein PV113 and PV114. Surprisingly, PV113 and PV114 showed similar high purity and yield as PV7 (Table 10) . Also surprisingly, PV113 and PV114 showed comparable binding to VEGF as PV7 despite a linker as short as 8 or 5 amino acids only. These results surprisingly showed the unexpected property of the anti-PD-L1 antibody PV1 in tolerating a much shorter linker for linking VEGFR domain to the C-terminal of its heavy chain without notable compromise in its biophysical property or binding to VEGF. Similarly, the VEGFR-R1D2 domain was also linked to the C-terminal of the heavy chain of PV2 through a short linker of 8 amino acids (G4Sx1-GGG) to produce the bispecific antibody fusion protein PV115. Different from the low yield shown by PV4 linking the same VEGFR-R1D2 domain to the N-terminal of PV2 heavy chain, PV115 was surprisingly produced with a high purity and much higher yield than PV4 (Table 10) . These result indicated that for the antibody PV2, it’s not only more compatible to link the VEGFR-R1D2 domain to the C-terminal versus the N-terminal of its heavy chain, but also is tolerant of a shorter linker of≤10 amino acids in length for linking the VEGFR-R1D2 domain to the C-terminal of its heavy chain.
[0136] The anti-VEGF x PD-1 bispecific antibody AK112 is reported to form multimer complex with dimeric VEGF through its two VEGF-binding Fab arms, which greatly increases its avidity binding to PD-1 (see (Zhong et al., iScience, 2025) ) . Similar phenomenon may also occur to bispecific antibody fusion protein comprising bivalent VEGFR-R1D2 domains. However, the length of the linker linking the VEGFR-R1D2 domain to the C-terminal of the two heavy chains of an antibody may affect the tendency of the antibody fusion protein to form multimer complex with dimeric VEGF. To test this, PV7, PV113 and PV114 with different length of linker linking VEGFR-R1D2 to the C-terminal of the two heavy chains of anti-PD-L1 antibody PV1, were subjected to SEC-HPLC analysis, either as a single agent or as a mixture after incubation with VEGF165 dimer at 1: 1 molar ratio for 1 hour at room temperature, to determine the presence of multimer complex. At a 1: 1 molar ratio, the bivalent PV7, PV113 and PV114 may form large crosslinking multimer complex with the dimeric VEGF165. Indeed, the SEC-HPLC analysis of the mixture of PV7, PV113 or PV114 with VEGF165 dimer showed a large complex peak at 7.4 minute retention time versus a monomer peak at 9.7 minute retention time for PV7, PV113 or PV114 alone. Notably, the percentage of the large complex peak increased from 17%for PV7 to 23%for PV113 and 41%for PV114, indicating more crosslinking multimer complex formation as the length of the linker linking VEGFR-R1D2 to the antibody heavy chain C-terminal decreased from 15 amino acids in PV7 to 8 amino acids in PV113 and to 5 amino acids in PV114. While a linker of<14 amino acids in length is shown to compromise biophysical property and VEGF binding when linking VEGFR domain to the heavy chain C-terminal of the anti-PD-L1 antibody IMM25 (see US11407832B2) , these results identified that a linker with<14 amino acids in length not only is effective for linking a VEGFR domain to the heavy chain C-terminal of the anti-PD-L1 antibody PV1 without comprising the biophysical property and VEGF binding of the antibody fusion protein, but also favorably increases the ability of the antibody fusion protein to form multimer complex with dimeric VEGF.
[0137] Additionally, the native amino acid sequence of VEGFR-R1D2 fragment was also extended to different length to produce the bispecific antibody fusion protein PV117 and PV118 (Table 10) . When the VEGFR-R1D2 fragment was extended 8 additional amino acids of its native sequence at its N-terminal (SEQ ID NO: 451) in PV117, PV117 was produced at similar high yield and purity as PV113 and PV114 (Table 10) , however its binding to VEGF165 was reduced compared to PV113 and PV114. Meanwhile, when the VEGFR-R1D2 fragment was extended 15 additional amino acids of its native sequence at its N-terminal (SEQ ID NO: 452) in PV118, both the yield and purity of PV118 production became substantially lower (Table 10) . Table 10. Production of bispecific antibody fusion protein against PD-L1 and VEGF using VEGFR domains
[0138] An ELISA assay was performed to test the bispecific antibody fusion protein’s VEGF binding activity. The ELISA assays was performed using 96-well ELISA plate coated with the test protein at 10μg / ml in 1x PBS / well at 4℃overnight. After washing with 1x PBS+0.05%Tween 20 (PBST) , the plate was blocked with 1x PBS+5%milk for 2 hours at room temperature (RT) . After washing the plate with 1x PBST, 8-fold serially diluted concentrations of VEGF165 protein with C-terminal His-tag in 1x PBS+1%milk starting at 8μg / ml were added and incubated for 60 minutes at RT. After washing with 1x PBST, 1: 4000 diluted anti-His tag-HRP (Proteintech, CAT#HRP-66005) in 1x PBS+1%milk was added and incubated for 60 minutes at RT. After washing with 1x PBST, 30μl / well tetramethylbenzidine (TMB) substrate solution was added and incubated in dark at room temperature for several minutes. 30μl / well H2SO4 was then added to stop the reaction, and OD450 was read out at 450 nm using a microplate reader and data were analyzed by Graphpad Prism. The ELISA result in FIG. 7A showed that PV7 exhibited binding to VEGF comparable to that of bevacizumab, but surprisingly PV3 exhibited significantly weaker binding to VEGF than that of PV7. Thus, contrary to weaker VEGF binding shown for IMM25011 linking VEGFR-R1D2 to the heavy chain C-terminal of the anti-PD-L1 antibody IMM25 than that of IMM2510 linking VEGFR-R1D2 to the heavy chain N-terminal of IMM25 (see US11407832B2) , surprisingly stronger VEGF binding was shown for PV7 linking VEGFR-R1D2 to the heavy chain C-terminal of the anti-PD-L1 antibody PV1 than that of PV3 linking VEGFR-R1D2 to the heavy chain N-terminal of PV1. Similarly, while PV4 linking VEGFR-R1D2 to the heavy chain N-terminal of the anti-PD-L1 antibody PV2 showed higher VEGF binding than that of PV3, it’s still weaker than that by linking the VEGFR-R1D2 domain to the heavy chain C-terminal.
[0139] Additional ELISA was also performed to test the bispecific antibody fusion protein’s PD-L1 binding activity. In this ELISA, 96-well ELISA plate was coated with human PD-L1 ECD (comprising amino acid F19-R238 of accession#NP_054862) protein with a C-terminal STREP Tag-II at 5μg / ml in 1x PBS / well at 4℃overnight. After washing with 1x PBS+0.05%Tween 20 (PBST) , the plate was blocked with 1x PBS+5%milk for 2 hours at room temperature (RT) . After washing the plate with 1x PBST, 8-fold serially diluted concentrations of test protein in 1x PBS+1%milk were added and incubated for 60 minutes at RT. After washing with 1x PBST, 1: 8000 diluted Goat-Anti-Human-IgG-Fc-HRP (Jackson Immuno, CAT#109-035-008) in 1x PBS +1%milk was added and incubated for 60 minutes at RT. After washing with 1x PBST, 30μl / well TMB substrate solution was added and incubated in dark at room temperature for several minutes. 30μl / well H2SO4 was then added to stop the reaction, and OD450 was read out at 450 nm using a microplate reader and data were analyzed by Graphpad Prism. The ELISA result in FIG. 7B showed that PV7 exhibited binding to PD-L1 comparable to that of the parent anti-PD-L1 antibody PV1, but PV3 exhibited significantly lower binding to PD-L1 than that of PV7. Thus together, these results showed that PV7 linking VEGFR-R1D2 to the heavy chain C-terminal of the anti-PD-L1 antibody PV1 exhibited stronger binindg to both VEGF and PD-L1 than that of PV3 linking VEGFR-R1D2 to the heavy chain N-terminal of PV1.
[0140] To further test simultanesous binding of PD-L1 and VEGF, another ELISA was performed. In this ELISA, 96-well ELISA plate was first coated with human PD-L1 ECD as done above. After washing and blocking, fixed concentrations (33.3nM) of test protein in 1x PBS+1%milk were added and incubated for 60 minutes at RT. After washing the plate with 1x PBST, 8-fold serially diluted concentrations of VEGF165 protein with C-terminal His-tag in 1x PBS+1%milk starting at 16μg / ml were added and incubated for 60 minutes at RT. After washing with 1x PBST, 1: 4000 diluted anti-His tag-HRP (Proteintech, CAT#HRP-66005) in 1x PBS+1%milk was added and incubated for 60 minutes at RT. After washing with 1x PBST, 30μl / well TMB substrate solution was added and incubated in dark at room temperature for several minutes. 30μl / well H2SO4 was then added to stop the reaction, and OD450 was read out at 450 nm using a microplate reader and data were analyzed by Graphpad Prism. The ELISA results in FIG. 7C showed that PV7 exhibited robust simultaneous binding to PD-L1 and VEGF, but PV3 exhibited only very weak simultaneous binding to PD-L1 and VEGF, confirming the much better binding to both PD-L1 and VEGF by PV7 linking the VEGFR-R1D2 domain to the C-terminal of the PV1 heavy chain than by PV3 linking the VEGFR-R1D2 domain to the N-terminal of the PV1 heavy chain. The protein PV4, which linked the VEGFR-R1D2 domain to the N-terminal of the PV2 heavy chain, exhibited quite higher simultaneous binding to PD-L1 and VEGF than that of PV3, but it’s still much weaker than that by linking the VEGFR-R1D2 domain to the C-terminal of the heavy chain (FIG. 7C) . Also, changing the VEGF-binding domain from VEGFR-R1D2 (SEQ ID NO: 17) in PV3 to VEGFR-R1D2-R2D3 (SEQ ID NO: 18) in PV11 significantly increased the simultaneous binding to PD-L1 and VEGF, but it’s still much weaker than that of PV7 by just linking the same VEGFR-R1D2 domain to the C-terminal of the heavy chain instead (FIG. 7C) . Meanwhile, the result in FIG. 7C showed that PV113 and PV114 exhibited simultanesous binding to PD-L1 and VEGF similar to that of PV7, confirming that a shorter linker less than 10 amino acids (8 amino acids in PV113 and 5 amino acids in PV114) linking the VEGFR-R1D2 domain to the C-terminal of the PV1 heavy chain, compared to a longer linker (e.g. 15 amino acids in PV7) , not only maintained the high production yield and purity of the antibody fusion protein, but also maintained comparable binding to PD-L1 and VEGF. On the other hand, PV117 showed significantly reduced simultaneous binding to PD-L1 and VEGF than PV7, PV113 and PV114, while PV118 exhibted substantially reduced simultaneous binding to PD-L1 and VEGF than PV7, PV113 and PV114 (FIG. 7C) , indicating further extending the VEGFR-R1D2 domain at its native N-terminal sequence (extending 8 amino acids in PV117 and 15 amino acids in PV118) was detrimental to binding activity of the domain. Example 2. Construction of bispecific antibody fusion protein against PD-L1 and VEGF using bevacizuamb derived anti-VEGF domains
[0141] In addition to VEGFR domain, bispecific antibody against PD-L1 and VEGF were also constructed, similarly as in Example 1, by linking anti-VEGF scFv derived from the anti-VEGF antibody ranibizumab or bevacizumab to the C-terminal of the heavy chain of the anti-PD-L1 antibody PV1 comprising VH and VL of SEQ ID NO: 32 and 33 respectively, as detailed in Table 11. A ranibizumab-derived anti-VEGF scFv (VL-G4Sx3-G-VH, SEQ ID NO: 19) was reported and characterized in the patent US8541203B2. The same ranibizumab-derived anti-VEGF scFv (SEQ ID NO: 19) was fused to the heavy chain C-terminal of the anti-PD-L1 antibody through a G4Sx4 linker to generate PV31. Given bevacizumab and ranibizumab are humanized from the same parent mouse anti-VEGF antibody, a bevacizumab-derived anti-VEGF scFv in the same configuration (VL-G4Sx3-G-VH, SEQ ID NO: 20) was fused to the heavy chain C-terminal of the anti-PD-L1 antibody through a G4Sx4 linker to generate PV32. DNAs encoding the full heavy chain polypeptides and the full light chain polypeptides of PV31 and PV32 were optimized for expression in Homo sapiens, synthesized and cloned into the same pcDNA3.4 mammalian expression vector using standard molecular cloning techniques, and the two protein were then produced through transient co-expression of the heavy chain and light chain vectors at 2: 3 ratio in HEK293 cells and single step of Protein A chromatography purification, as similarly done in Example 1. The purified antibody protein were analyzed by SDS-PAGE and SEC-HPLC. However, the two bispecific antibody PV31 and PV32 showed only very low yield of 0.7 mg / L and 1.5 mg / L respectively, and also low purity of 86.76%and 71.89%respectively in SEC-HPLC analysis (Table 11) . To improve the purity and yield, an additional disulfide bond was engineered into the anti-VEGF scFv of PV31 and PV32 to generate PV33 and PV34 respectively, as detailed in Table 11 below. However, the yield of PV33 and PV34 remained very low (1.33mg / L and 0.65mg / L respectively. In attempt to improve the yield, the linker in the anti-VEGF scFv of PV32 was next replaced by a novel stapled linker to generate PV35 (Table 11 below) and see if such a stapled linker could improve the yield and purity of the scFv-antibody fusion protein. The yield of PV35 appeared better than PV32 but still remained very low (6mg / L) , while the purity of PV35 became high at 99.21%in SEC-HPLC analysis. Thus, the anti-VEGF scFv in PV35 was further mutated to generate PV36 (Table 11 below) and see if changing the cap sequence of the kappa VL fragment of the scFv to a lambda VL cap sequence could increase the yield of the fusion protein. Fortunately, PV36 showed higher yield of 14.7mg / L versus 6mg / L for PV35, with also high purity of 99.4%in SEC-HPLC analysis. However, the yield of PV36 was still modest and much lower than the yield of the backbone anti-PD-L1 antibody PV1. Next, the Fab from bevacizumab was instead linked to the heavy chain C-terminal of the anti-PD-L1 antibody PV1 to generate PV37. However, the yield of PV37 also remained modest at 20.7mg / L but with high purity of 97%in SEC-HPLC analysis. Furthermore, a crossmab Fab from bevacizumab was linked to the N-terminal of the heavy chain of PV1 to generate PV121, which showed similar yield as PV37 in production but lower purity (Table 11) .
[0142] Additional configuration of ranibizumab-derived scFv and bevacizumab-derived scFv in VH-G4Sx4-VL with an engineered disulfide bond (SEQ ID NO: 25 and 26 respectively) were also tested by linking them to the C-terminal of the heavy chain of PV1 to produce the bispecific antibody fusion protein PV38 and PV39, both of which however still had low production yield (Table 11) . Interestingly, when the same bevacizumab scFv (SEQ ID NO: 26) was linked to the N-terminal of the heavy chain of PV1 to produce the bispecific antibody fusion protein PV119, the production yield further reduced to negligible (Table 11) . However, when linking a further engineered bevacizumab scFv with two engineered disulfide bonds (SEQ ID NO: 453) to the N-terminal of the heavy chain of PV1 to produce the protein PV120, the production yield was improved but still low (Table 11) . Table 11. Production of bispecific antibody fusion protein against PD-L1 and VEGF using VEGFR domains
[0143] Also, an ELISA assay was performed to test the binding activity of the bispecific antibody fusion protein generated. The ELISA assays was performed using 96-well ELISA plate coated with 21ng / well VEGF165 (Sino Biologics, CAT#11066-HNAH) in 1x PBS / well at 4℃overnight. After washing with 1x PBS+0.05%Tween 20 (PBST) , the plate was blocked with 1x PBS+5%milk for 2 hours at room temperature (RT) . After washing the plate with 1x PBST, 3-fold serially diluted concentrations of PV31, PV32 and PV37 etc. in 1x PBS+1%milk starting at 30 nM were added and incubated for 60 minutes at RT. After washing with 1x PBST, 1: 4000 diluted anti-human IgG Fc-HRP (Abcam, CAT#ab97225) in 1x PBS+1%milk was added and incubated for 60 minutes at RT. After washing with 1x PBST, 30μl / well TMB substrate solution was added and incubated in dark at room temperature for several minutes. 30μl / well H2SO4 was then added to stop the reaction, and OD450 was read out at 450 nm using a microplate reader and data were analyzed by Graphpad Prism. Surprisingly, PV37 showed no binding to VEGF165 in the ELISA assay (FIG. 7 D) . Similarly, PV31 and PV32 showed minimal and no binding to VEGF165 in the ELISA assay (FIG. 7 D) . These surprising results indicate that the ranibizuamb-based scFv (VL-G4Sx3-VH, SEQ ID NO: 19) reported in US8541203B2 as well as the bevacizumab Fab are actually inactive in binding to VEGF when linked to the C-terminal of the antibody heavy chain.
[0144] In contrast, PV36 showed robust binding to VEGF in another ELISA (FIG. 7E) , indicating engineering the bevacizumab scFv not only improved its biophysical property but also its functional binding to VEGF. PV121 also showed binding with VEGF (FIG. 7F) , indicating a bevacizumab-derived Fab need be preferably linked to the N-terminal, instead of the C-terminal, of a heavy chain to maintain its binding to VEGF. Meanwhile, PV39 with a bevacizumab-derived scFv (SEQ ID NO: 26) linked to the C-terminal of the heavy chain of PV1, also showed binding to VEGF in ELISA (FIG. 7F) , indicating that different from bevacizumab-derived Fab and similar to PV36, a bevacizumab-derived scFv linked to the C-terminal of a heavy chain could retain its binding to VEGF. On the other hand, PV120 showed no binding to VEGF (FIG. 7F) . Example 3. Construction of bispecific antibody fusion protein against PD-L1 and VEGF using bevacizuamb as backbone
[0145] Bispecific antibody fusion protein against PD-L1 and VEGF were also constructed by linking an anti-PD-L1 scFv to the anti-VEGF antibody bevacizumab. As an example, two scFv derived from anti-PD-L1 antibody BMS936559 were constructed from N-terminal to C-terminal as VH-G4Sx3 linker-VL (SEQ ID NO: 44) and VL-G4Sx3 linker-VH (SEQ ID NO: 45) , and linked to the N-terminal of the bevacizumab heavy chain to construct the bispecific antibody fusion protein PV78 and PV79 (Table 12) . DNAs encoding the full heavy chain polypeptides and the full light chain polypeptides of PV78 and PV79 were optimized for expression in Homo sapiens, synthesized and cloned into the same pcDNA3.4 mammalian expression vector using standard molecular cloning techniques, and the two protein were then produced through transient co-expression of the heavy chain and light chain vectors at 2: 3 ratio in HEK293 cells and single step of Protein A chromatography purification, as similarly done in Example 1. However, no protein was detected from repeated expression effort for PV78, whereas PV79 showed expression but at a low yield of 3 mg / L with also a low SEC-HPLC purity of 68%after a single step protein A purification.
[0146] Furthermore, a BMS936559-derived anti-PD-L1 scFv in VL-G4Sx4-VH configuration with one or two engineered disulfide bonds (SEQ ID NO: 46 and 455 respectively) was linked to the N-terminal of the bevacizumab light chain to construct and produce the bispecific antibody fusion protein PV122 and PV123 (Table 12) . PV122 showed low production yield (Table 12) . Different from the bevacizumab-derived scFv, adding an additional engineered disulfide bond into the BMS936559-derived scFv did not improve production but actually further reduced the yield to be negligible (Table 12) . Then a BMS936559-derived stapled scFv in the configuration of VL-linker-VH (SEQ ID NO: 456) was linked to the N-terminal of bevacizumab light chain or to the C-terminal of bevacizumab heavy chain to construct and produce the bispecific antibody fusion protein PV124 and PV125 respectively. PV124 showed high purity but still low yield. In contrast, PV125 showed much higher yield and also>90%SEC-HPLC purity (Table 12) . These result indicated that a stapled scFv derived from BMS936559, when specifically linked to the heavy chain C-terminal of bevacizumab, generated an antibody fusion protein with good production yield and purity. Furthermore, a BMS936559-derived crossmab Fab was linked to the N-terminal of bevacizumab light chain to construct and produce PV126, which however also had low production yield (Table 12) .
[0147] An ELISA assay was also performed to test the binding activity of PV79 against PD-L1 and determine if the ati-PD-L1 scFv in PV79 is functional. The ELISA assays was performed using 96-well ELISA plate coated with 150ng / well PD-L1 (ACROBiosystem, CAT#PD-1-H5282) in 1x PBS / well at 4℃overnight. After washing with 1x PBS+0.05%Tween 20 (PBST) , the plate is blocked with 1x PBS+5%milk for 2 hours at room temperature (RT) . After washing the plate with 1x PBST, 3-fold serially diluted concentrations of PV79 and also the parent anti-PD-L1 antibody PV1 (BMS936559 analog) in 1x PBS+1%milk starting at 3.3 nM were added and incubated for 60 minutes at RT. After washing with 1x PBST, 1: 4000 diluted anti-human IgG Fc-HRP (Abcam, CAT#ab97225) in 1x PBS+1%milk is added and incubated for 60 minutes at RT. After washing with 1x PBST, 30μl / well tetramethylbenzidine (TMB) substrate solution is added and incubated in dark at room temperature for several minutes. 30μl / well H2SO4 is then added to stop the reaction, and OD450 is read out at 450 nm using a microplate reader and data are analyzed by Graphpad Prism. The ELISA result in FIG. 8 showed good binding of PV79 to PD-L1, although modestly weaker than that of the parent anti-PD-L1 antibody. Table 12. Production of bispecific antibody fusion protein against PD-L1 and VEGF using bevacizumab as backbone antibody Example 4. Construction of bispecific antibody fusion protein against PD-L1 / PD-1 and VEGF using non-bevacizuamb derived single chain anti-VEGF domains
[0148] Additional anti-VEGF single-chain domains including VL dAb (SEQ ID NO: 27) , VH dAb (SEQ ID NO: 28) , VHH (SEQ ID NO: 29) and non-bevacizumab / ranibizumab-derived scFv (SEQ ID NO: 30) were tested to construct bispecific antibody fusion protein against PD-L1 and VEGF. Also, a new format wherein the anti-VEGF single chain domains were linked to the C-terminal of the light chain of the anti-PD-L1 antibody was tested. As detailed in Table 13, the anti-VEGF single chain domains were each linked to the C-terminal of the light chain of the anti-PD-L1 antibody PV1 through a G4Sx6 linker to construct the bispecific antibody fusion protein PV89 to PV92, which were produced and analyzed by SDS-PAGE and SEC-HPLC analysis, similarly as done in Example 1-3. The results in Table 13 showed that the PV91 protein comprising the anti-VEGF VHH domain (SEQ ID NO: 29) had a very low yield of only 2 mg / L, and the PV92 protein comprising the anti-VEGF scFv domain (SEQ ID NO: 30) had a modest yield of 16.7 mg / L. In contrast, the PV89 and PV90 protein comprising the anti-VEGF VL dAb domain and VH dAb domain respectively, showed better yield of 35 mg / L and 46.7 mg / L respectively. Meanwhile, all the 4 protein showed good SEC-HPLC purity of 87-97%after a single step Protein A purification.
[0149] Similarly, a bispecific antibody fusion protein PV105 against PD-1 and VEGF was constructed by linking the anti-VEGF VL dAb (SEQ ID NO: 27) to the C-terminal of the light chain of the anti-PD-1 antibody nivolumab through a G4Sx6 linker as illustrated in Table 13, which was produced and analyzed by SDS-PAGE and SEC-HPLC analysis, similarly as done in Example 1 and 2. The result showed that a yield of 22.6 mg / L and SEC-HPLC purity of 95.9%after a single step protein A chromatograph purification. Table 13. Production of bispecific antibody fusion protein against PD-L1 / PD-1 and VEGF using non-bevacizuamb derived single chain anti-VEGF domains Example 5. Construction of trispecific antibody fusion protein against PD-L1 / PD-1, VEGF and CD47
[0150] On the basis of the bispecific antibody fusion protein against PD-L1 and VEGF constructed in Example 1-3 above, trispecific antibody fusion protein against PD-L1, VEGF and CD47 were further constructed by linking CD47-binding wild type SIRPαV2 IgV domain to a bispecific antibody fusion protein against PD-L1 and VEGF in various formats (FIG. 1-4) . Such exemplary trispecific antibody fusion protein PVC1, PVC5-9, PVC14-16, PVC22-24, PVC26 and PVC31-32 were constructed as illustrated in Table 14 and were produced and analyzed by SDS-PAGE and SEC-HPLC analysis, similarly as done in Example 1-4. A monospecific wild type SIRPαV2 IgV domain-Fc fusion protein SIN-300 and a bispecific antibody protein against PD-L1 and C47 by linking the wild type SIRPαV2 IgV domain to the light chain N-terminal of the parent anti-PD-L1 antibody PV1, were also similarly produced as benchmark protein. The results showed generally good yield and high SEC-HPLC purity after a single step Protein A chromatography purification, except the protein PVC15 and PVC23 showed only a 61%SEC-HPLC purity in this result.
[0151] Similarly, a trispecific antibody fusion protein PVC34 against PD-1, VEGF and CD47 was constructed by linking CD47-binding wild type SIRPαV2 IgV domain and the anti-VEGF VL dAb (SEQ ID NO: 27) to the N-terminal and C-terminal respectively of the light chain of the anti-PD-1 antibody Nivolumab, as illustrated in Table 14. The protein was produced and analyzed by SDS-PAGE and SEC-HPLC analysis, similarly as done in Example 1-4. The result showed that a yield of 84 mg / L and SEC-HPLC purity of 93.7%after a single step protein A chromatograph purification.
[0152] Additionally, a trispecific antibody fusion protein PVC62 against PD-L1, VEGF and CD47 was constructed by linking CD47-binding wild type SIRPαV2 IgV domain and the anti-PD-L1 VHH (SEQ ID NO: 50) to the N-terminal of the light chain and heavy chain respectively of the anti-VEGF antibody bevacizuamb, as illustrated in Table 15. The protein was produced and analyzed by SDS-PAGE and SEC-HPLC analysis, similarly as done in Example 1-4. The result showed that a yield of 50.7 mg / L and SEC-HPLC purity of 98.7%after a single step protein A chromatograph purification. Table 14. Production of trispecific antibody fusion protein against PD-L1 / PD-1, VEGF and CD47 Table 15. Production of trispecific antibody fusion protein against PD-L1, VEGF and CD47 using bevacizumab as backbone antibody Example 6. Analyzing binding activity of the bispecific / trispecific antibody fusion protein against VEGF through ELISA assay
[0153] An ELISA assay was performed to test the binding activity against VEGF165 of the bispecific / trispecific antibody fusion protein generated in Example 4-5, which comprised different anti-VEGF single-chain domains and also different structural formats of protein using the same anti-VEGF domain. The ELISA assays was performed using 96-well ELISA plate coated with 21ng / well VEGF165 (Sino Biologics, CAT#11066-HNAH) in 1x PBS / well at 4℃overnight. After washing with 1x PBS+0.05%Tween 20 (PBST) , the plate is blocked with 1x PBS+5%milk for 2 hours at room temperature (RT) . After washing the plate with 1x PBST, 3-fold serially diluted concentrations of the bispecific / trispecific antibody protein (Table 13-15) and also a bevacizumab analog as benchmark in 1x PBS+1%milk starting at 30 nM were added and incubated for 60 minutes at RT. After washing with 1x PBST, 1: 4000 diluted anti-human IgG Fc-HRP (Abcam, CAT#ab97225) in 1x PBS+1%milk is added and incubated for 60 minutes at RT. After washing with 1x PBST, 30μl / well tetramethylbenzidine (TMB) substrate solution is added and incubated in dark at room temperature for several minutes. 30μl / well H2SO4 is then added to stop the reaction, and OD450 is read out at 450 nm using a microplate reader and data are analyzed by Graphpad Prism.
[0154] The ELISA results in FIG. 9 showed a range of different apparent binding affinity to VEGF165 among the tested bispecific / trispecific antibody proteins. Specifically, different apparent affinity was shown among antibody protein having the same structural format but different anti-VEGF domains. For example, the bi / trispecific antibody protein with the anti-VEGF VL dAb domain (SEQ ID NO: 27) generally showed the highest binding to VEGF165 (e.g. PVC22, PV89 and PVC14) , whereas the anti-VEGF scFv domain (SEQ ID NO: 30) generally showed among the lowest binding to VEGF165 (e.g. PVC9 and PV92) (FIG. 9) . Different apparent affinity was also shown among the bi / trispecific antibody protein having the same anti-VEGF domain but different structural formats. Also, the structural format showing the highest binding varied for different anti-VEGF domains. For example, the antibody protein PVC22, PV89, PVC14 and PVC5 had the same anti-VEGF VL dAb domain (SEQ ID NO: 27) but different structural formats of FV-19, FV-4, FV-11 and FV-12 respectively (FIG. 9) , however the 4 protein showed different binding to VEGF165 in the order of PVC22 (EC50=7.08E9 nM) >PV89 (EC50=8.6E3 nM) >PVC14 (3.7E2 nM) >PVC5 (EC50=6.7E2 nM) ≥bevacizumab analog (EC50=0.469 nM) (FIG. 9) . Also, the antibody protein PVC31, PV90, PVC23, PVC6 and PVC15 had the same anti-VEGF VH dAb domain (SEQ ID NO: 28) but different structural formats of FV-17, FV-4, FV-19, FV-12 and FV-11 respectively (FIG. 9) , however the 5 protein showed different binding to VEGF165 in the order of PVC31 (EC50=0.077 nM) >PV90 (EC50=0.084 nM) >PVC23 (0.114 nM) >bevacizumab analog (EC50=0.469 nM) >PVC6 (EC50=0.144 nM) ≈PVC15 (EC50=0.214 nM) (FIG. 9) . Similarly, the antibody protein PVC32, PV24, PVC16, PVC7 and PV91 had the same anti-VEGF VHH domain (SEQ ID NO: 29) but different structural formats of FV-17, FV-19, FV-11, FV-12 and FV-4 respectively (FIG. 9) , however the 5 protein showed different binding to VEGF165 in the order of PVC32 (EC50=0.104 nM) >PVC24 (EC50=0.194 nM) >bevacizumab analog (EC50=0.469 nM) ≥PVC16 (EC50=0.533 nM) ≥PVC7 (EC50=0.536 nM) >PV91 (4.83 nM) (FIG. 9) . Additionally, the antibody protein PV92 and PVC9 had the same anti-VEGF scFv domain (SEQ ID NO: 30) but different structural formats of FV-4 and FV-12 respectively (FIG. 9) , however the 2 protein showed different binding to VEGF165 in the order of bevacizumab analog (EC50=0.469 nM) >PV92 (EC50=1.036 nM) >PVC9 (EC50=2.952 nM) (FIG. 9) .
[0155] Meanwhile, the trispecific antibody protein PVC62 in structural format FV-27 using the anti-VEGF antibody bevacizumab as backbone antibody surprisingly showed similar binding to VEGF165 as the bevacizumab analog (FIG. 9) , despite both an anti-PD-L1 VHH domain and anti-CD47 SIRPαV2 IgV domain were linked to the N-terminal of the heavy chain and light chain respectively of the bevacizumab analog with more expected steric hindrance to VEGF binding of the bevacizumab analog. Example 7. Analyzing binding activity of the bispecific / trispecific antibody fusion protein against PD-L1 through ELISA assay
[0156] ELISA assay was performed to test the binding activity against PD-L1 of the bispecific / trispecific antibody fusion protein generated in Example 4-5, which comprised different anti-VEGF single-chain domains and different structural formats but mostly with the same anti-PD-L1 backbone antibody PV1. The ELISA assays was performed using 96-well ELISA plate coated with 150ng / well PD-L1 protein (ACROBiosystem, CAT#PD-1-H5282) in 1x PBS / well at 4℃overnight. After washing with 1x PBS+0.05%Tween 20 (PBST) , the plate is blocked with 1x PBS+5%milk for 2 hours at room temperature (RT) . After washing the plate with 1x PBST, 3-fold serially diluted concentrations of the bi / trispecific antibody protein, and also the parent anti-PD-L1 antibody PV1 (BMS936559 analog) and a bispecific antibody fusion protein PC1 against PD-L1 and CD47 by linking wild type SIRPαV2 IgV domain to the N-terminal of the light chain of the anti-PD-L1 antibody PV1 as benchmark, in 1x PBS+1%milk starting at 30 nM were added and incubated for 60 minutes at RT. After washing with 1x PBST, 1: 4000 diluted anti-human IgG Fc-HRP (Abcam, CAT#ab97225) in 1x PBS+1%milk is added and incubated for 60 minutes at RT. After washing with 1x PBST, 30μl / well tetramethylbenzidine (TMB) substrate solution is added and incubated in dark at room temperature for several minutes. 30μl / well H2SO4 is then added to stop the reaction, and OD450 is read out at 450 nm using a microplate reader and data are analyzed by Graphpad Prism.
[0157] The ELISA results in FIG. 10 showed most of the various bi / trispecific antibody protein with the same anti-PD-L1 backbone antibody PV1 exhibited largely close binding to PD-L1 similar to that of the parent anti-PD-L1 antibody PV1, despite of different structural formats. However, the bispecific antibody protein PV91 in structural format FV-4 exhibited much lower binding to PD-L1 than the other bi / trispecific antibody protein and the parent anti-PD-L1 antibody PV1 (FIG. 10A) . Also, the trispecific antibody protein PVC8 and PVC7 in structural format FV-12 and the bispecific antibody protein PV92 in structural format FV-4 exhibited some lower binding to PD-L1 than the other bi / trispecific antibody protein tested and also the parent anti-PD-L1 antibody PV1 (FIG. 10) . In contrast, the benchmark protein PC1 and PV1 showed similar binding to PD-L1, and the trispecific antibody protein PVC32 and PVC31 in structural format FV-17 as well as PVC24 in format FV-19 and PVC6 in format FV-12 surprisingly exhibited even some higher binding to PD-L1 than that of PV1 and / or PC1 (FIG. 10B) . Notably, the trispecific antibody protein PVC32 and PVC31 had both anti-VEGF domain and SIRPαV2 IgV domain linked to the N-terminal of the heavy chain and light chain respectively of the anti-PD-L1 backbone antibody and would have expected more crowded steric hindrance with lower binding to PD-L1 than the PV1 and PC1 antibody protein, but actually showed surprisingly higher binding to PD-L1. Similarly, the trispecific antibody protein PVC62 having anti-PD-L1 VHH (from the parent anti-PD-L1 protein KN035) and SIRPαV2 IgV domain linked to the N-terminal of the heavy chain and light chain respectively of the anti-VEGF antibody bevacizumab surprisingly showed actually higher binding to PD-L1 than the parent anti-PD-L1 antibody KN035 analog (FIG. 10C) .
[0158] Additionally, similar to that the VEGFR-R1D2-comprising anti-PD-L1 x VEGF bispecific antibody fusion protein PV7 showed higher binding to PD-L1 than PV3 (FIG. 7B) , PV7’s counterpart trispecific antibody fusion protein PVC1 also showed higher binding to immobilized PD-L1 than PV3’s counterpart trispecific antibody fuson protein PVC26 in another ELISA assay. Furthermore, PVC1 also showed higher binding to CD47 than PVC26 in an ELISA assay. Meanwhile, after mixing PVC1 or PVC26 with dimeric VEGF165 at 1: 2 molar ratio, both PVC1 and PVC26 showed largerly increased binding to immobilized PD-L1 than PVC1 alone or PVC26 alone respectively in an ELISA assay, indicating that PVC1 and PVC26 bound to VEGF and PD-L1 simultanesously and its binding with dimeric VEGF led to formation of multimer complex with increased avidity binding to PD-L1. Example 8. Analyzing binding activity of the bispecific / trispecific antibody fusion protein against CD47 through ELISA assay
[0159] ELISA assay was also performed to test the binding activity against CD47 of the trispecific antibody fusion protein generated in Example 4-5, which comprised different anti-VEGF single-chain domains and different structural formats but with the same wild type CD47-binding SIRPαV2 IgV domain. The ELISA assays was performed using 96-well ELISA plate coated with 120ng / well CD47 protein (ABclonal, CAT#RP01306) in 1x PBS / well at 4℃overnight. After washing with 1x PBS+0.05%Tween 20 (PBST) , the plate is blocked with 1x PBS+5%milk for 2 hours at room temperature (RT) . After washing the plate with 1x PBST, 4-fold serially diluted concentrations of the trispecific antibody protein, and also the bispecific antibody fusion protein PC1 linking wild type SIRPαV2 IgV domain to the light chain N-terminal of the anti-PD-L1 antibody PV1 and monospecific wild type SIRPαV2 IgV domain-Fc fusion protein SIN-300 as benchmark, in 1x PBS+1%milk starting at 50 nM were added and incubated for 60 minutes at RT. After washing with 1x PBST, 1: 4000 diluted anti-human IgG Fc-HRP (Abcam, CAT#ab97225) in 1x PBS+1%milk is added and incubated for 60 minutes at RT. After washing with 1x PBST, 30μl / well tetramethylbenzidine (TMB) substrate solution is added and incubated in dark at room temperature for several minutes. 30μl / well H2SO4 is then added to stop the reaction, and OD450 is read out at 450 nm using a microplate reader and data are analyzed by Graphpad Prism.
[0160] The ELISA results in FIG. 11 showed most of the various trispecific antibody protein with the same CD47-binding SIRPαV2 IgV domain exhibited largely close binding to CD47 similar to that of the benchmark protein PC1 and SIN-300, despite of different structural formats. However, the trispecific antibody protein PVC7 and PVC5 in structural format FV-12 exhibited lower binding to CD47 than the other trispecific antibody protein and the benchmark protein PC1 and SIN-300 in the ELISA result (FIG. 11B) . The trispecific protein PVC15 in structural format FV-11 also showed some lower binding to CD47 than the benchmark protein PC1 and SIN-300 in the ELISA result (FIG. 11A) . Given PVC7, PVC5 and PVC15 had the same fusion of wild type SIRPαV2 IgV domain to the light chain N-terminal of the same anti-PD-L1 backbone antibody PV1 as the benchmark protein PC1, except an additional anti-VEGF domain was linked to the C-terminal of the heavy chain (for PVC7 and PVC5) or light chain (for PVC15) of the anti-PD-L1 backbone antibody PV1, the lower CD47-binding shown by PVC7, PVC5 and PVC15 than PC1 was unexpected. In contrast, the trispecific protein PVC32 and PVC31 had both SIRPαV2 IgV domain and anti-VEGF domain linked to the N-terminal of the light chain and heavy chain respectively of the anti-PD-L1 backbone antibody and would have expected more crowded steric hindrance with lower binding to CD47 than the benchmark PC1 and SIN-300 protein, but actually showed similar binding to CD47.
[0161] Similarly, the trispecific antibody protein PVC62 having SIRPαV2 IgV domain and anti-PD-L1 VHH linked to the N-terminal of the light chain and heavy chain respectively of the anti-VEGF antibody bevacizumab also surprisingly showed similar binding to CD47 as the benchmark protein PC1 and SIN-300 (FIG. 11B) . Example 9. Analyzing co-binding of PD-L1 and CD47 of the bispecific / trispecific antibody fusion protein by FACS assay
[0162] For the trispecific antibody protein against PD-L1, VEGF and CD47, PD-L1 and CD47 locate on the target cell membrane, whereas VEGF is soluble protein. To examine co-binding of PD-L1 and CD47 on cell surface by the trispecific antibody fusion protein generated in Example 4-5, a FACS binding assay on HT-1080 cells that express both PD-L1 and CD47 on cell surface was performed. HT-1080 cells were grown in RPMI 1640 medium with 10%FBS and harvested at leogrithmic growth period with over 90%viability, and seeded at 1x10^5 cells / well in 96-well plate. After incubation at 37’ C for 30 min, 4-fold serially diluted concentrations of the trispecific antibody protein, and also the benchnmark protein PC1, SIN-300, PV1 and KN035, in FACS buffer (1x PBS+1%FBS) starting at 1280 nM were added and incubated at 4℃for 30 minutes. The cells were then washed with FACS buffer for three times, and 1: 1000 diluted PE-labeled goat anti-human IgG Fc secondary antibody (Ebioscience, CAT#12-4998-82) was added into the cells and incubated at 4℃for 40 min. After wash, the cells were resuspended in FACS buffer and analyzed by Flow Cytometry.
[0163] The FACS results in FIG. 12 showed that all the trispecific antibody protein tested exhibited higher binding than that of the benchmark PV1 that binds to PD-L1 only and the benchmark SIN-300 that binds to CD47 only, indicating co-binding of PD-L1 and CD47 on HT-1080 cells by the trispecific antibody protein tested. At the higher concentrations, most of the trispecific antibody protein showed higher bining MFI than the sum of the binding MFI of PV1 and SIN-300 at the same concentrations, indicating synergistic increase from simultanesous binding of PD-L1 and CD47 by the trispecific antibody protein. Although the different trispecific antibody protein share the same PD-L1 binding backbone antibody and the same CD47-binding SIRPαV2 IgV domain, the overall co-binding strength to PD-L1 and CD47 varied significantly among the trispecific antibody protein tested as shown in FIG. 12, with PVC15, PVC5 and PVC23 showing among the lowest overall binding strength and PVC16, PVC31 and PVC32 et al. showing among the higest overall binding strength in the FACS result in FIG. 12.
[0164] Similarly, the FACS result in FIG. 12A showed that the trispecific antibody protein PVC62 having SIRPαV2 IgV domain and anti-PD-L1 VHH linked to the N-terminal of the light chain and heavy chain respectively of the anti-VEGF antibody bevacizumab also exhibited synergistic co-binding of CD47 and PD-L1 higher than the sum of the binding of the CD47-binding only benchmark SIN-300 and the PD-L1-binding only benmark KN035 at the same concentrations. Example 10. Analyzing simultaneous binding of PD-L1, VEGF and CD47 of the trispecific antibody fusion protein through ELISA assay.
[0165] Based on the ELISA and FACS results in Example 6-9, a number of tested trispecific antibody protein including PVC14 and PVC16 in format FV-11, PVC6 in format FV-12, PVC22 in format FV-19, PVC31 in FV-17 and PVC62 in format FV-27 exhibited among the best individual binding to VEGF, PD-L1 and / or CD47. To test if the trispecific antibody protein could simultanesouly bind to all the three target antigens, an ELISA assay was designed and first performed on 4 selected samples. In this ELISA assay, a96-well plate was first coated with 150ng / well PD-L1 (ACROBiosystem, CAT#PD-1-H5282) which comprises a Twin-Strep tag but no His-tag. After blocking and wash, the PD-L1 coated wells were then incubated with the trispecific antibody protein PVC14, PVC22, PVC31 or PVC62 resepectively all at a fixed concentration of 30 nM in 1x PBS+1%milk for 60 minutes at RT. The wells were then washed with 1x PBST and incubated with the different target antigen in parallel: 1) VEGF165 protein with a 6x His-tag (ABclonal, CAT#RP01150) at 4-fold serial dilution of 8 concentrations starting at 16μg / ml in 1x PBS+1%milk for detecting simultaneous binding of PD-L1 and VEGF, or 2) CD47 protein with a 6x His-tag (ABclonal, CAT#RP01306) at 4-fold serial dilution of 8 concentrations starting at 16μg / ml in 1x PBS+1%milk for detecting simultaneous binding of PD-L1 and CD47, or 3) a combination of the same VEGF165 protein+CD47 protein in 1x PBS+1%milk at 4-fold serial dilution of 8 concentrations for VEGF165 and CD47, starting at 16μg / ml VEGF165+16μg / ml CD47, for detecting simultaneous binding of PD-L1, VEGF and CD47. After incubation for 60 minutes at RT, the wells were washed with 1x PBST and incubated with 1: 4000 diluted anti-6xHis tag-HRP (Proteintech, CAT#HRP-66005) in 1x PBS +1%milk for 60 minutes at RT. After washing with 1x PBST, 30μl / well tetramethylbenzidine (TMB) substrate solution is added and incubated in dark at room temperature for several minutes. 30μl / well H2SO4 is then added to stop the reaction, and OD450 is read out at 450 nm using a microplate reader and data are analyzed by Graphpad Prism.
[0166] The ELISA results in FIG. 13 showed all the four trispecific antibody protein tested exhitied simultaneous binding of PD-L1 and VEGF, or simultanesous binding of PD-L1 and CD47. However, for the trispecific antibody protein PVC22, the wells incubated with the combination of VEGF and CD47 antigen in the ELISA assay exhibited the same binding curve as the wells incubated with the CD47 antigen only (FIG. 13A) , indicating no simultanesous binding of PD-L1, VEGF and CD47 at the concentrations tested. Similarly, for the trispecific antibody protein PVC62, the wells incubated with the combination of VEGF and CD47 antigen exhibited only a mildly stronger binding cure than the wells incubated with the CD47 antigen only (FIG. 13B) , indicating only weak simultanesous binding of PD-L1, VEGF and CD47 at the concentrations tested. In contrast, for the trispecific antibody protein PVC14 and PVC31, the wells incubated with the combination of VEGF and CD47 antigen exhibited clearly stronger binding curve than the wells incubated with either the CD47 antigen only or the VEGF antigen only (FIG. 13C-D) , indicating clear simultanesous binding of PD-L1, VEGF and CD47 at the concentrations tested. Example 11. Generation of SIRP IgV domain variants with higher binding to CD47 at an acidic pH than at physiological pH
[0167] The IgV extracellular domain of wild type human SIRPαV2 (SIRPαV2 IgV WT domain, SEQ ID NO: 58) was first selected as the template for generating variants with differential binding to human CD47 (NCBI accession#: NP_001768.1) at an acidic pH than at physiological pH. Eleven initial candidate residues of SIRPαV2 IgV WT were selected, including position V33, Q37, N51, Q52, K53, K68, R69, E70, M72, K96 and G97 of SEQ ID NO: 58, for generating single histidine substitution variant. DNA constructs encoding fusion protein of wild type SIRPαV2 IgV (SEQ ID NO: 58) or SEQ ID NO: 58 derived variant SIRPαV2 IgV with single histidine substitution of V33H, Q37H, N51H, Q52H, K53H, K68H, R69H, E70H, M72H, K96H or G97H directly fused to the N-terminal of human IgG1 Fc (SEQ ID NO: 8) , were generated through gene synthesis and cloned into pcDNA3.4 mammalian expression vector. The SIRPαV2 IgV-IgG1 Fc fusion proteins were then produced through transient expression of the vector in HEK293 cells and single step of Protein A chromatography purification. The proteins were analyzed by SDS-PAGE and SEC-HPLC. As shown in Table 16 below, all the variant SIRPαV2 IgV-IgG1 Fc fusions were expressed and purified with high yield and purity as good as the wild type SIRPαV2 IgV-IgG1 Fc protein, through a single step Protein A chromatography purification. Table 16. Production of wild type and variant SIRPαV2 IgV-IgG1 Fc fusion proteins. Example 12. Binding activities of variant SIRPαV2 IgV-IgG1 Fc fusion proteins measured by FACS
[0168] The binding activities of the SIRPαV2 IgV-IgG1 Fc fusion proteins from Example 11 to Raji cancer cells were measured by fluorescence activated cell sorting (FACS) at pH 7.2 which represents the lower boundary of physiological pH range (pH 7.2 to 7.5) and at an acidic pH 6.0. Serially diluted SIRPαV2 IgV-IgG1 Fc fusion proteins in 1x PBS+2%FBS at pH 7.2 or 6.0 were added at final concentrations of 1280 nM, 640 nM, 160 nM, 40 nM, 10 nM, 2.5 nM, 0.625 nM, 0.15625 nM, 0.0391 nM, 0.0098 nM, 0.0024 nM and 0.0006 nM, and incubated with Raji cells in 96-well plate (2x105 / well) at 4℃for 30 minutes. After washing the cells with 1x PBS+2%FBS at pH 7.2 or 6.0, FITC-labelled anti-human IgG Fc secondary antibody (Abcam#, ab98596, 1: 300) were added and incubated at 4℃for 30 minutes at pH 7.2 or 6.0. After washing, the Raji cells were then analyzed by FACS in 1x PBS+2%FBS at pH 7.2 or 6.0.
[0169] The FACS results in FIG. 14A show that the variant SIRPαV2 IgV-IgG1 Fc fusion protein SIN-306 (M72H) showed no change in binding than wild type SIRPαV2 IgV-IgG1 Fc (SIN-300) at pH 7.2. However, the variant protein SIN-301 (K53H) , SIN-302 (R69H) , SIN-303 (Q52H) , SIN-304 (K68H) and SIN-305 (K96H) all showed reduced binding to Raji cells at pH 7.2 than wild type SIRPαV2 IgV-IgG1 Fc fusion protein SIN-300 which has the exact same amino acid sequence as clinically tested TTI-621 (a wild type SIRPαV2 IgV-IgG1 Fc fusion protein with the same amino acid sequence of SIN-300) . SIN-301 (K53H) , SIN-302 (R69H) and SIN-305 (K96H) show particularly weak binding at pH 7.2. On the other hand, the results in FIG. 14B show that at pH 6.0, SIN-306 (M72H) showed still no change in binding from wild type SIRPαV2 IgV-IgG1 Fc (SIN-300) , whereas SIN-305 (K96H) showed even lower and minimal binding at pH 6.0 than at pH 7.3. However, SIN-301 (K53H) , SIN-302 (R69H) , SIN-303 (Q52H) and SIN-304 (K68H) all showed enhanced binding to Raji cells at pH 6.0. Meanwhile, the results in FIG. 14C and D show that: 1) the variant SIRPαV2 IgV-IgG1 Fc fusion protein SIN-339 (V33H) abolished binding at both physiological pH 7.2 and acidic pH 6.0; 2) Similar to SIN-306 (M72H) , SIN-341 (Q37H) showed no binding difference from the wild type SIRPαV2 IgV-IgG1 Fc (SIN-300) at both physiological pH 7.2 and acidic pH 6.0. This is unexpected, as a Q37H mutation of SIRPαV2 IgV was reported to mildly increase binding to CD47-high and low cells (see US11021694) ; 3) SIN-342 (N51H) showed mildly lower binding than wild type SIN-300 at both physiological pH 7.2 and acidic pH 6.0, but also showed similar binding at both both physiological pH 7.2 and acidic pH 6.0 as similar to wild type SIN-300. Together these results of the 11 variants tested identified the histidine substitution K53H, R69H, Q52H or K68H of SIRPαV2 IgV (SEQ ID NO: 58)rendered higher binding to CD47 at an acidic pH than at physiological pH.
[0170] To further characterize the pH-sensitive CD47 binding of SIN-301 (K53H) , SIN-302 (R69H) , SIN-303 (Q52H) and SIN-304 (K68H) , FACS analysis of their binding to Raji cells at pH 7.3, pH 6.5 and pH 6.0 was performed. The results in FIG. 15A and 15B show that binding of wild type SIRPαV2 IgV-IgG1 Fc (SIN-300) to Raji cells didn’ t significantly change at the different pH, but the binding to Raji cells increased as the pH decreased from pH 7.3 to pH 6.5 and then to pH 6.0 for SIN-301 (K53H) and SIN-302 (R69H) . Additionally, FACS analysis of SIN-301 (K53H) and SIN-302 (R69H) on another cell line SKOV3 expressing lower level of CD47 than Raji (Wang et al., J Immunother Cancer, 2020, Golubovskaya et al., Cancers (Basel) , 2017) was performed. The results in FIG. 15C show that, similar to the results in Raji cells, SIN-301 (K53H) and SIN-302 (R69H) also exhibited higher binding to SKOV3 cells at an acidic pH 6.0 than the weak binding at physiological pH 7.3. Example 13. Binding activities of variant SIRPαV2 IgV-IgG1 Fc fusion proteins measured by ELISA
[0171] The binding of the variant SIRPαV2 IgV to CD47 protein was further measured by ELISA at pH 7.2 and at pH 6.0.96-well ELISA plate was coated with 8μg / mL SIRPαV2 IgV-IgG1 Fc fusion protein SIN-300 (wild type) , SIN-301 (K53H) , SIN-302 (R69H) , SIN-303 (Q52H) or SIN-304 (K68H) in 30μl 1x PBS / well at 4℃overnight. After washing with 1x PBS+0.05%Tween 20 (PBST) , the plate was blocked with 1x PBS+5%milk for 2 hours at room temperature (RT) . After washing the plate with 1x PBST, 4-fold serially diluted concentrations of His-tagged CD47 extracellular domain (ECD) protein (Abclonal, CAT#RP01306) at 200 nM, 50 nM, 12.5 nM, 3.125 nM, 0.781 nM, 0.195 nM, 0.049nM and 0.012 nM in 1x PBS+1%milk at pH 7.2 or 6.0, were added (30 μl / well) and incubated for 60 minutes at RT. After washing with 1x PBST, 1: 4000 diluted anti-His-HRP (Proteintech, CAT#HRP-66005) in 1x PBS+1%milk at pH 7.2 or 6.0 was added and incubated for 60 minutes at RT.After washing with 1x PBST, 30μl / well tetramethylbenzidine (TMB) substrate solution was added and incubated in dark at room temperature for several minutes. 30μl / well 2M H2SO4 was then added to stop the reaction, and OD450 was read out at 450 nm using a SpectraMax 190 microplate reader and data were analyzed by Graphpad Prism 9. Because monovalent his-tag CD47 ECD protein was used to dected coated SIRPαV2 IgV-IgG1 Fc fusion proteins, the binding reflects monovalent affinity between the SIRPαV2 IgV variants and CD47.
[0172] Similar to FACS results in Example 12 above, the ELISA results in FIG. 16A show that SIN-301 (K53H) , SIN-302 (R69H) , SIN-303 (Q52H) and SIN-304 (K68H) all exhibited reduced binding to CD47 than wild type SIRPαV2 IgV-IgG1 Fc (SIN-300) at pH 7.2. SIN-302 (R69H) particularly showed lowest and minimal binding to CD47 at pH 7.2 and SIN-301 (K53H) showed a large reduction in binding to CD47 at pH 7.2. SIN-303 (Q52H) and SIN-304 (K68H) showed similarly smaller reduction in binding to CD47 at pH 7.2. Meanwhile, the results in FIG. 16B show that SIN-302 (R69H) regained binding to CD47 at pH 6.0 and SIN-301 (K53H) showed increased binding to CD47 that’s close to the binding of wild type SIRPαV2 IgV-IgG1 Fc (SIN-300) at pH 6.0. Both SIN-303 (Q52H) and SIN-304 (K68H) showed similar binding to CD47 as that of wild type SIRPαV2 IgV-IgG1 Fc(SIN-300) at pH 6.0.
[0173] To benchmark the binding significance of SIN-301 at physiological pH, another ELISA was similarly done, but using the His-tagged CD47 extracellular domain (ECD) protein as immobilized ligand and fusion proteins comprising bivalent or monovalent SIRPαV2 IgV domain as analyte, to also detect bivalent avidity effect of the binding. The ELISA assays was performed using 96-well ELISA plate coated with 120ng / well CD47 protein (ABclonal, CAT#RP01306) in 1x PBS / well at 4℃overnight. After washing with 1x PBS+0.05%Tween 20 (PBST) , the plate was blocked with 1x PBS+5%milk for 2 hours at room temperature (RT) . After washing the plate with 1x PBST, 6-fold serially diluted concentrations of the tested protein starting at 250 nM in 1x PBS+1%milk at pH 7.3 or pH 6.0, were added and incubated for 60 minutes at RT. After washing with 1x PBST at pH 7.3 or 6.0, 1: 4000 diluted anti-human IgG Fc-HRP (Abcam, CAT#ab97225) in 1x PBS+1%milk at pH 7.3 or 6.0 was added and incubated for 60 minutes at RT. After washing with 1x PBST, 30μl / well tetramethylbenzidine (TMB) substrate solution is added and incubated in dark at room temperature for several minutes. 30μl / well H2SO4 is then added to stop the reaction, and OD450 is read out at 450 nm using a microplate reader and data are analyzed by Graphpad Prism.
[0174] The results in FIG. 16C show the benchmark protein MP-24 comprising monovalent wild type SIRPαV2 IgV domain exhibited much reduced binding to CD47 than the benchmark bivalent protein SIN-300 (TTI-621 analog) comprising bivalent wild type SIRPαV2 IgV domains. Meanwhile, the CD47 binding of both SIN-301 (bivalent SIRPαV2 IgV K53H variant-IgG1 Fc fusion protein) and SIN-302 (bivalent SIRPαV2 IgV R69H variant-IgG1 Fc fusion protein) were even weaker at physiological pH 7.3 than the benchmark monovalent wild type binding of MP-24 (FIG. 16C) . At an acidic pH 6.0, SIN-301 exhibited stronger binding than, while SIN-302 exhibited binding close to, that of the benchmark monovalent wild type binding of MP-24 (FIG. 16C) . In contrast, the CD47 binding of SIN-304 (bivalent SIRPαV2 IgV K68H-IgG1 Fc fusion protein) was stronger than the benchmark monovalent binding of MP-24 at physiological pH 7.3 and was further increased close to the level of the benchmark bivalent binding of SIN-300 at an acidic pH 6.0. Monovalent wild type SIRPαV2 IgV domain in tumor targeting bispecific antibody fusion protein configured in similar format as MP-24, is shown to be effective in promoting phagocytosis of target tumor cells with no significant effect on RBCs and platelets and better safety profile than bivalent wild type SIRPαV2 IgV-IgG1 Fc protein (Liu et al., Cell Rep, 2018) . Given SIN-301 and SIN-302 showed lower CD47 binding at physiological pH but stronger or similar CD47 binding at an acidic pH than monovalent MP-24, the results in FIG. 16C indicate an even better safety and / or activity profile for bivalent SIRPαV2 IgV domains with K53H or R69H mutation than monovalent wild type SIRPαV2 IgV domain.
[0175] Together with the FACS results in Example 12, these results show that the mutation of R69H or K53H largely reduces SIRPαV2 IgV binding to CD47 at physiological pH to much weaker binding than wild type SIRPαV2 IgV, but restores significant binding to CD47 at an acidic pH. Similarly, the mutation of Q52H or K68H also significantly reduces SIRPαV2 IgV binding to CD47 at physiological pH, although to a less extent than R69H or K53H mutation, and restores significant binding to CD47 at an acidic pH that’s stronger than the binding of R69H or K53H mutation at an acidic pH. Such a binding profile of SIRP IgV domain with R69H, K53H, Q52H or K68H mutation is desirable to reduce or minimize binding to the ubiquitous CD47-expressing normal tissues and cells, but allow binding toCD47-expressing disease cells in an acidic disease microenvironment, such as CD47-expressing tumor cells in solid tumors. Example 14. Binding activities of variant SIRPαV2 IgV-IgG1 Fc fusion protein to human primary platelets measured by FACS
[0176] Thrombocytopenia was reported as a most frequent adverse event for both TTI-621 (a wild type SIRPαV2 IgV-IgG1 Fc fusion protein with the same amino acid sequence of SIN-300) and TTI-622 (a wild type SIRPαV2 IgV-IgG41 Fc fusion protein) in clinical trials and an initial dose limiting toxicity for TTI-621 (Trillium Therapeutics, Inc. R&D Day presentation, 4 / 28 / 2021, (Ansell et al., Clin Cancer Res, 2021) and Blood 138 (2021) 2448-2451) . These indicate significant binding of wild type SIRPαV2 IgV to platelets despite its minimal binding to red blood cells (RBCs) . The binding of the variant SIRPαV2 IgV-IgG1 Fc fusion protein to platelets at physiological pH 7.3 were thus tested by FACS.
[0177] Human primary platelets were seeded into 96-well plate at 1x106 / well and incubated with Fc blocking antibody at room temperature. After washing the platelets with 1x PBS+2%FBS pH 7.3, three-fold serially diluted SIRPαV2 IgV-IgG1 Fc fusion protein in 1x PBS+2%FBS at pH 7.3 were added at final concentrations of 1280 nM, 426.67 nM, 142.22 nM, 47.41 nM, 15.80 nM and 5.27 nM, and incubated with the platelets at 4℃for 30 minutes. In addition, a magrolimab (Hu5F9) analog protein was generated and used as a benchmark control. Also, a monovalent wild type SIRPαV2 IgV-IgG1 Fc protein (MP-24) and monovalent variant K53H SIRPαV2 IgV-IgG1 Fc protein (MP-25) were generated and used for the FACS binding. After washing the platelets with 1x PBS+2%FBS at pH 7.3, FITC-labelled anti-human IgG Fc secondary antibody (Abcam, CAT#ab6854) were added and incubated at 4℃for 30 minutes at pH 7.3. After washing, the platelets were then analyzed by FACS in 1x PBS+2%FBS at pH 7.3.
[0178] The FACS results in FIG. 17A show that the homodimeric bivalent wild type SIRPαV2 IgV-IgG1 Fc fusion protein SIN-300 exhibited significant binding to platelets that plateaued after 47 nM, but its peak binding MFI to platelets was less than 1 / 3 of that of the Hu5F9 analog. Meanwhile, the benchmark protein MP-24 comprising monovalent wild type SIRPαV2 IgV domain showed largely reduced binding to platelets than the bivalent SIN-300. Notably, the bivalent variant protein SIN-301 (K53H) and SIN-302 (R69H) as well as the protein MP-25 comprising monovalent SIRPαV2 IgV K53H domain all showed negligible binding to human platelets at pH 7.3. Meanwhile, the bivalent variant SIN-303 (Q52H) showed similar weak binding to platelets as monovalent MP-24 at pH 7.3. In addition, the bivalent variant SIN-302 (R69H) and SIN-301 (K53H) continued to show negligible or very weak binding to human primary platelets at physiological pH 7.3, even at higher concentrations up to 5,125 nM or 410μg / ml (FIG. 17B) tested, which reaches around the reported Week 1 peak serum concentration of the highest 18.0 mg / kg IV dose of TTI-622 tested in clinical trials (Trillium Therapeutics, Inc. R&D Day presentation, 4 / 28 / 2021) . These results indicate negligible or very weak platelet binding for the variant protein SIN-302 (R69H) and SIN-301 (K53H) and also weakened platelet binding for the variant protein SIN-303 (Q52H) and SIN-304 (K68H) at physiological pH to avoid or reduce thrombocytopenia in human patients, even at high clinical doses. Example 15. Binding activities of variant SIRPαV2 IgV-IgG1 Fc fusion protein to human primary T cells measured by FACS
[0179] T cells show high surface CD47 expression among human primary cells with CD47 expression density reported at 83, 477 surface CD47 receptors per cell, versus 34, 439 per cell for platelets (Puro et al., Mol Cancer Ther, 2020) . The binding of the variant SIRPαV2 IgV-IgG1 Fc fusion protein to human primary T cells at physiological pH 7.3 were thus also tested by FACS.
[0180] Human primary T cells were seeded into 96-well plate at 1x105 / well in RPMI-1640 medium at pH 7.3 or 6.0, and incubated at 37℃for 30 minutes. After washing the T cells, serially diluted SIRPαV2 IgV-IgG1 Fc fusion protein were added at final concentrations of 2560 nM, 1280 nM, 640 nM and 160 nM at pH 7.3 or 6.0, and incubated with the T cells at 4℃for 30 minutes. In addition, a magrolimab (Hu5F9) analog protein was used as a benchmark control. After washing the cells with 1x PBS+2%FBS at pH 7.3 or 6.0, PE-labelled anti-human IgG Fc secondary antibody (Abcam, CAT#ab98596) were added and incubated at 4℃for 30 minutes at pH 7.3. After washing, the T cells were then analyzed by FACS in FACS buffer at pH 7.3 or 6.0.
[0181] The FACS results in FIG. 17C show that the homodimeric bivalent wild type SIRPαV2 IgV-IgG1 Fc fusion protein SIN-300 exhibited significant binding to human T cells at both pH 7.3 and 6.0, but its peak binding MFI to human T cells was less than 1 / 2 of that of the Hu5F9 analog. Consistent with the binding result in platelets above, the variant protein SIN-302 (R69H) showed negligible binding to human T cells at pH 7.3 even at high concentration of 2,560 nM, whereas the variant protein SIN-301 (K53H) showed mild binding to human T cells at pH 7.3 at high concentrations above 1,000 nM. At an acidic pH 6.0, SIN-302 (R69H) still shows minimal binding to human T cells even at high concentration of 2, 560 nM, while SIN-301 (K53H) showed increased binding to T cells at an acidic pH 6.0, reaching similar binding MFI as that of wild type SIN-300 at the high concentration of 2,560 nM. These results indicate negligible or weak T cell binding for variant protein SIN-302 (R69H) and SIN-301 (K53H) at physiological pH, thus allowing SIRP IgV domain with R69H and / or K53H mutation to be used as a binding domain for CD47-targeting T cell engagers or chimeric antigen receptor (CAR) T cell therapy or other engineered T cell therapy that avoid or reduce fratricide of CD47-positive T cells. Example 16. Generation and test of variant IgV of other SIRP family subtypes for higher binding to CD47 at an acidic pH than at physiological pH
[0182] As shown in FIG. 6A and 6B, the residues Q52, K53, K68 and R69 of SIRPαV2 IgV (SEQ ID NO: 58) are conserved across all the 10 subtypes of SIRPα (SEQ ID NO: 3 to 11) as well as SIPRβand SIPRγ. Thus, the same histidine substitution of K53H, R69H, Q52H or K68H can be applied to the IgV of SIRPαV1 (SEQ ID NO: 57) , SIRPαV8 (SEQ ID NO: 64) , SIPRβ2-D (SEQ ID NO: 71) , SIPRγ (SEQ ID NO: 72) , SIPRγ-Q (SEQ ID NO: 73) , SIPRγ-D (SEQ ID NO: 74) and SIPRγ-QD (SEQ ID NO: 75) as templates to generate variant SIRP IgV on each of these templates respectively. Directly fusing these variant SIRP IgV domains to the N-terminal of human IgG1 Fc (SEQ ID NO: 8) leads to generation of the SIRP IgV-IgG1 Fc fusion protein as listed in Table 17 below, similarly as the generation of variant SIRPαV2-IgG1 Fc fusion protein detailed in Example 11. It’s also noted that among the 10 SIRPαsubtypes, SIRPαV5, SIRPαV6 and SIRPαV9 each differ from SIRPαV1 in only one residue outside the CD47-binding area, SIRPαV3, SIRPαV7 and SIRPαV10 each differ from SIRPαV2 in only one residue, while SIRPαV4 and SIRPαV8 both comprising a mixture of the 13 amino acid differing between human SIRPαV1 and SIRPαV2 (FIG. 6A) . Importantly, the differing residues among the SIRPαsubtypes locate outside their CD47-binding area and do not affect binding to CD47, as described in (Hatherley et al., JBiol Chem, 2014) . Thus, the variant examples of SIRPαV1, SIRPαV2 and SIRPαV8 tested shall represent all the other SIRPαsubtypes as well.
[0183] The variant SIRP IgV-IgG1 Fc fusion protein as listed in Table 17 were selected for analysis of binding to Raji cells by FACS, similarly as detailed for SIRPαV2-IgG1 Fc fusion protein in Example 12. The FACS results in FIG. 18A show that wild type SIRPαV1 IgV-IgG1 Fc fusion protein SIN-308 exhibited mildly weaker binding than wild type SIRPαV2 IgV-IgG1 Fc fusion protein SIN-300, but overall the mutation of K53H, R69H, Q52H or K68H similarly reduced SIRPαV1 IgV’s binding to CD47 at physiological pH 7.3 but regained higher binding at an acidic pH 6.0. Meanwhile, the FACS results in FIG. 18B show that wild type SIRPαV8 IgV-IgG1 Fc fusion protein SIN-313 exhibited comparable binding to wild type SIRPαV2 IgV-IgG1 Fc fusion protein SIN-300, and overall the mutation of K53H, R69H, Q52H or K68H also reduced SIRPαV8 IgV’s binding to CD47 at physiological pH 7.3 but regained higher binding at an acidic pH 6.0.
[0184] The FACS results in FIG. 18C show that wild type SIRPγIgV-IgG1 Fc fusion protein SIN-318 exhibited much weaker binding than wild type SIRPαV2 IgV-IgG1 Fc fusion protein SIN-300, consistent with prior report (Hatherley et al., Mol Cell, 2008) . Similarly, overall the mutation of K53H, R69H, Q52H or K68H reduced SIRPγIgV’s binding to CD47 at physiological pH 7.3 and regained higher binding at an acidic pH 6.0 than at pH 7.3, although to a lesser extent than the reduction of binding at pH 7.3 and regaining of binding at pH 6.0 of SIRPαV2 IgV, probably because wild type SIRPγIgV starts with already quite low binding to CD47. Additionally mutating the 2 differing amino acids in SIRPγIgV domain to the amino acids in corresponding positions of SIRPαV2, namely L37Q and N101D as shown in FIG. 6B, is indicated to increase SIRPγ’s binding to CD47. However, only mild increase in binding was observed comparing SIRPγIgV protein with R69H+L37Q+N101D mutation (SIN-324) to SIRPγIgV protein with R69H only (SIN-320) (FIG. 18C) . In contrast, SIRPγIgV protein with K53H+L37Q+N101D mutation (SIN-323) exhibited largely increased binding to CD47 than SIRPγIgV protein with K53H mutation only (SIN-319) at an acidic pH, with also large increase in binding at physiological pH but surprisingly much less than the increase at acidic pH (FIG. 18C) . Notably, the SIRPγIgV protein with K53H+L37Q+N101D mutation reached several-fold higher in binding MFI at higher concentrations than even wild type SIRPαV2 IgV-IgG1 Fc protein (SIN-300) (FIG. 18C) . Similarly, the SIRPγIgV protein with K53H+N101D mutation (SIN-361) also showed largely increased binding to CD47 than SIRPγIgV protein with K53H mutation only (SIN-319) at an acidic pH, but more surprisingly with no significant increase in binding observed at physiological pH (FIG. 18E) . On the other hand, the SIRPγIgV protein with K53H+L37Q mutation (SIN-356) showed poor expression and low purity with only 46.51%purity in SEC-HPLC assay (versus 98.73%SEC-HPLC purity for SIN-361) . This however also revealed superior expression and biophysical quality of SIRPγIgV protein with the K53H+N101D mutation than the K53H+L37Q mutation.
[0185] The FACS results in FIG. 18D show that the SIPRβ2 IgV-IgG1 Fc fusion protein with a single mutation of H101D in the SIPRβ2 IgV domain (SIN-325) gained binding to CD47 similar to that of wild type SIRPαV2 IgV-IgG1 Fc fusion protein (SIN-300) , and overall the mutation of K53H, R69H, Q52H or K68H reduced SIPRβ2 H101D IgV’s binding to CD47 at physiological pH 7.3 and regained higher binding at an acidic pH 6.0. Notably, the mutations appeared to result in reduction of CD47 binding of SIPRβ2 H101D IgV at pH 7.3 to a larger extent than that of SIRPαV2 IgV, with R69H and K53H appearing to abolish SIPRβ2 H101D IgV’s binding to CD47 at physiological pH 7.3. At pH 6.0, the binding of SIPRβ2 H101D IgV with K53H, R69H, Q52H or K68H also appeared to be weaker than that of SIRPαV2 IgV with the same mutation.
[0186] Altogether these results indicate that histidine substitution mutation of the conserved residue Q52, K53, K68 or R69 results in reduced CD47 binding at physiological pH but regains higher binding at an acidic pH across all subtypes of SIRPα, SIPRβand SIRPγ. Table 17. List of variant SIRPα, SIPRβand SIPRγIgV-IgG1 Fc fusion protein Example 17. Antibody dependent cell-mediated phagocytosis (ADCP) activities of variant SIRPαV2 IgV-IgG1 Fc fusion protein measured by ADCP Jurkat reporter assay
[0187] To functionally test the pH-dependent binding of SIRPαV2 IgV variant protein, the wild type and variant SIRPαV2 IgV-IgG1 Fc fusion protein were subjected to an ADCP Jurkat reporter assay using a Jurkat-NFAT-Luc2-CD32a-R167 ADCP reporter cell line (KYinno, CAT#KC-1524) , which expresses CD32a-R167 receptor to transmit ADCP actviation signal and activate NFAT-Luc2 luciferase reporter gene, giving rise to chemiluminescence after substrate addition. The ADCP activity of the test articles can be determined by measuring the signal of luminescence using a luciferase assay kit. Raji cells were used as target cells and plated at 20,000 / well in 96-well plate in RPMI 1640 medium at pH 7.3 or 6.5. Five-fold serially diluted concentrations of wild type and variant SIRPαV2 IgV-IgG1 Fc fusion protein at pH 7.3 or 6.5were added to Raji cells and incubated at 37℃for 30 minutes. The Jurkat-NFAT-Luc2-CD32a-R167 ADCP reporter cell at pH 7.3 or 6.5 were then added to the mixture at 100,000 / well and incubated at 37℃for 6 hours. Bright-Glo reagent was then added to the cells and luminescence signal was read out using a multi-mode microplate reader. ADCP activity was calculated as the fold of luminescence strength of a test article over that of buffer negative control (Raji cells+Jurkat report cells+buffer control without test article) .
[0188] The results in FIG. 19A show that at physiological pH 7.3, wild type protein SIN-300 exhibited robust ADCP activity, but SIN-301 (K53H) and SIN-302 (R69H) exhibited minimal ADCP activity less than 1 / 6 of that of SIN-300 at high concentration, while SIN-302 (Q52H) and SIN-304 (K68H) exhibited significant ADCP activity but weaker than that of SIN-300. At an acidic pH 6.5, the results in FIG. 19B show that ADCP activity readout of the assay was largely reduced overall at pH 6.5, likely due to suppressed activity of the Jurkat reporter cells at an acidic pH, so that the peak fold of wild type SIN-300 at pH 6.5 is only~1 / 6 of that at pH 7.3. However, the activity fold of SIN-301 at pH 6.5 reached~60%of the peak level of wild type SIN-300 at high concentration, and the activity fold of SIN-302 at pH 6.5 also reached~40%of the peak level of wild type SIN-300 at high concentration. Similarly, SIN-303 and SIN-304 showed activity closer to that of wild type SIN-300 at pH 6.5. These functional ADCP results are consistent with the reduced binding of the variant SIRPαV2 IgV protein to CD47 at physiological pH but higher binding at an acidic pH. Example 18. Antibody dependent cell-mediated cytotoxicity (ADCC) activities of variant SIRPαV2 IgV-IgG1 Fc fusion protein measured by ADCC Jurkat reporter assay
[0189] To further functionally test the pH-dependent binding of SIRPαV2 IgV variant protein, the wild type (SIN-300) and variant SIRPαV2 IgV-IgG1 Fc fusion protein (SIN-301 and SIN-302) were subjected to an ADCC Jurkat reporter assay using a Jurkat-NFAT-Luc2-CD16 V158 ADCC reporter cell line, which expresses CD16-V158 receptor to transmit ADCC activation signal and activate NFAT-Luc2 luciferase reporter gene, giving rise to chemiluminescence after substrate addition. The ADCC activity of the test articles can be determined by measuring the signal of luminescence using a luciferase assay kit. Raji cells were used as target cells and plated at 20,000 / well in 96-well plate in RPMI 1640 medium at pH 7.3 or 6.5. Five-fold serially diluted concentrations of wild type and variant SIRPαV2 IgV-IgG1 Fc fusion protein at pH 7.2 or 6.0 were added to Raji cells and incubated at 37℃for 30 minutes. The Jurkat-NFAT-Luc2-CD16 V158 ADCC reporter cell at pH 7.2 or 6.0 were then added to the mixture at 100,000 / well and incubated at 37℃for 6 hours. Bright-Glo reagent was then added to the cells and luminescence signal was read out using a multi-mode microplate reader.
[0190] The results in FIG. 20A show that at physiological pH 7.2, wild type protein SIN-300 exhibited robust ADCC activity, but SIN-301 (K53H) and SIN-302 (R69H) exhibited negligible ADCC activity. At an acidic pH 6.0, the results in FIG. 20B show that ADCC activity readout of the assay was largely reduced overall, likely due to suppressed activity of the Jurkat reporter cells at an acidic pH, so that the peak activity of wild type SIN-300 at pH 6.0 was only~1 / 3 of that at pH 7.2. However, the ADCC activity of SIN-301 at pH 6.0 reached~65%of the peak level of wild type SIN-300 at high concentration, and the activity of SIN-302 at pH 6.0 also reached~40%of the peak level of wild type SIN-300 at high concentration. These functional ADCC results are consistent with the ADCP results in Example 17 and further functionally validate the reduced binding of the variant SIRPαV2 IgV protein to CD47 at physiological pH but higher binding at an acidic pH. Example 19. Generation and binding activities of variant SIRPαV2 IgV with combo mutations measured by FACS
[0191] Combination of single mutations are then tested to identify positive combo mutations that further reduce the binding at physiological pH, and / or further increase the binding of the variant protein at an acidic pH. SIRPαV2 IgV-IgG1 Fc fusion protein with various combo mutations comprising two or more mutations selected from K53H, R69H, Q52H and K68H on the wild type SIRPαV2 IgV template (SEQ ID NO: 58) were designed. Initially, exemplary protein with double mutations of SIRPαV2 IgV domain, as listed in Table 18 below, were produced similarly as done in Example 11 and analyzed for binding to CD47 expressing Raji cells by FACS similarly as done in Example 12. The same double mutations of other SIRPα, SIRPβor SIRPγIgV domains can also be produced and analyzed for CD47 binding similarly.
[0192] The results in FIG. 21A-C show that all the combo mutations exhibited further reduced binding at physiological pH 7.2 than the corresponding single mutation component, which is desirable to further reduce or minimize binding to the ubiquitous CD47-expressing normal tissues and cells. Meanwhile at an acidic pH 6.0, except the variant SIN-337 (K53H+R69H) that showed no binding, all the combo mutation variants SIN-330 (Q52H+K68H) , SIN-332 (K53H+Q52H) , SIN-333 (K53H+K68H) , SIN-335 (R69H+Q52H) and SIN-336 (R69H+K68H) , continued to exhibit higher binding than at pH 7.2. Surprisingly, the higher binding at an acidic pH of the two single mutations didn’ t cooperate to increase binding of the double mutation at an acidic pH 6.0 than the two single mutations, but instead decreased the binding at an acidic pH 6.0. Table 18. List of variant SIRPαV2 IgV-IgG1 Fc fusion protein with double mutations produced
[0193] Next, combination of mutations at additional residues of SIRP IgV domain on top of a backbone mutation of K53H, R69H, Q52H, K68H, Q52H+K68H, K53H+K68H, R69H+K68H, K53H+Q52H, R69H+Q52H, K53H+K68H+Q52H, or R69H+K68H+Q52H, as listed in Table 19 below, are designed to further increase CD47 binding at an acidic pH, and / or further reduce CD47 binding at physiological pH. In one category, candidate residues located at the interaction interface between SIRP IgV domain and CD47 are selected, as they may directly affect SIRP IgV domain's binding to CD47 and / or the backbone mutations' interaction with CD47, based on the crystal structure of SIRPαV2 IgV-CD47 complex (PDB: 2JJS) . In another category, additional candidate residues located outside the direct interaction interface are selected for their potential to indirectly affect SIRP IgV domain's binding to CD47 through affecting the SIRP IgV domain's confirmation / structure and / or the backbone mutations'interaction with CD47, based on the crystal structure of SIRPαV2 IgV-CD47 complex (PDB: 2JJS) . Exemplary SIRPαV2 IgV-IgG1 Fc fusion protein (configured as in format FV-39) with single mutation at selected additional residue I31, P35, R40, R46, R59, S66, K96, S98, P99 or K104 of the wild type SIRPαV2 IgV (SEQ ID NO: 58) , as shown in Table 19 below, were produced similarly as done in Example 11, and analyzed for binding to CD47-expressing Raji cells by FACS assay similarly as done in Example 12. The same mutations at the corresponding positions of other SIRPα, SIRPβor SIRPγIgV domains can also be produced and analyzed for CD47 binding similarly, with one additional residue in position numbering after residue 100 in SIRPβ, SIRPγand some SIRPαIgV (e.g. SEQ ID NO: 55-57, 60-62, and 65-75) than in SIRPαV2 IgV (SEQ ID NO: 58) . For example, as shown in FIG. 6A and 6B, K104 in SIRPαV2 IgV is K105 correspondingly in SIRPαV1, SIRPαV4, SIRPαV5, SIRPαV6, SIRPαV9, SIRPβand SIRPγ.
[0194] As shown in the results in Table 20, these proteins were well expressed. The results in FIG. 22A-E, together with the results in FIG. 14A-D, show that the additional single mutations of the parent SIRPαV2 IgV (SEQ ID NO: 58) tested exhibit several types of effects on the SIRP IgV domain’s binding to CD47, as summarized as subgroups in Table 19: 1) Subgroup 1 mutations such as I31E in SIN-307 and I31D in SIN-513, similar to the mutation of K53H, R69H, Q52H and K68H, exhibited higher binding to CD47 at an acidic pH than at physiological pH (FIG. 22A) . Notably, I31E reduced CD47 binding of SIRP IgV at physiological pH without notable impact at an acidic pH (FIG. 22A) , a preferred property also applicable to help further lower CD47-binding at physiological pH when combined with the backbone mutation (K53H, R69H, Q52H, K68H or Q52H+K68H) . Meanwhile, I31D also largely reduced CD47 binding at physiological pH but with less reduction at an acidic pH (FIG. 22A) . Similar to the effect in SIRPαIgV, the corresponding mutation of I31E or I31D in SIRPβIgV and the corresponding mutation of L31E or L31D in SIRPγIgV can similarly lead to higher binding to CD47 at an acidic pH than at physiological pH. Furthermore, the result in FIG. 23C showed that the SIRPαV2 IgV protein (SIN-514) comprising both K53H and I31E mutation exhibited further lowered binding to CD47 at physiological pH but with similar binding at an acidic pH than SIRPαV2 IgV protein (SIN-301) comprising K53H only. Additionally, the results in FIG. 23C-D also showed that with addition of the I31E mutation, SIRPαV2 IgV protein comprising both R69H+I31E mutation (SIN-515) , or Q52H+I31E mutation (SIN-517) , or K68H+I31E mutation (SIN-518) also exhibited further lowed binding to CD47 at physiological pH than SIRPαV2 IgV protein comprising R69H only (SIN-302) , Q52H only (SIN-303) or K68H only (SIN-304) respectively, while maintaining higher binding to CD47 at an acidic pH than at physiological pH. Similarly, addition of the I31D mutation to the mutation of K53H, R69H, Q52H or K68H can further lower binding of the SIRP IgV to CD47 at physiological pH, while maintaining higher binding to CD47 at an acidic pH than at physiological pH. Also similarly, addition of the mutation of L31E or L31D to the mutation of K53H, K53H+N101D or K53H+N101D+L37Q in SIRPγIgV can further lower the binding to CD47 at physiological pH. 2) Subgroup 2 mutations, such as in SIN-348 (R40H) , SIN-342 (N51H) , SIN-343 (S66H) , SIN-354 (K96R) , SIN-345 (G97H) , SIN-346 (S98H) , SIN-347 (P99H) and SIN-351 (K104H) , reduced CD47 binding at both physiological pH and acidic pH (FIG. 14C-D and FIG. 22B-E) , a property applicable to help further lower CD47-binding at physiological pH when combined with the backbone mutation (K53H, R69H, Q52H, K68H or Q52H+K68H) ; 3) Subgroup 3 mutations, such as in SIN-339 (V33H) , SIN-340 (P35H) and SIN-305 (K96H) , reduce substantially or abolish CD47 binding at both physiological pH and acidic pH (FIG. 22B and FIG. 14A-C) ; 4) Subgroup 4 mutations, such as in SIN-341 (Q37H) , SIN-349 (R46H) , SIN-344 (E70H) and SIN-306 (M72H) , didn’ t notably affect CD47 binding at both physiological pH and acidic pH (FIG. 22B-D and FIG. 14A-D) . However, when Q37H was combined with the backbone mutation (K53H or R69H) , surprisingly, it further lowered CD47 binding at physiological and acidic pH (FIG. 23A-B) , indicating that single mutation of SIRP IgV that doesn’ t impact CD47 binding on its own could still affect CD47 binding when combined with the backbone mutation (K53H, R69H, Q52H, K68H or Q52H+K68H) ; 5) Subgroup 5 mutations, such as in SIN-353 (I31W) , SIN-352 (I31Y) and SIN-350 (R59H) , increased CD47 binding at both physiological pH and acidic pH (FIG. 22E) . In addition, the deglycosylation mutation at residue N80 (N80A, N80G, N80S, N80Q) is reported to remove glycosylation of SIRP IgV and increase CD47 binding (see US1080082B2) . Thus, this subgroup of mutations is applicable to help further increase CD47-binding at an acidic pH when combined with the backbone mutation (K53H, R69H, Q52H, K68H, Q52H+K68H, K53H+K68H, R69H+K68H, K53H+Q52H, R69H+Q52H, K53H+K68H+Q52H, or R69H+K68H+Q52H) . 6) Subgroup 6 mutations, such as in SIN-338 (I31H) , increased CD47 binding at physiological pH (FIG. 22B) ; Additionally, the single mutations in Subgroup 7 and 9 of Table 19 were reported to reduce or increase CD47 binding respectively (see US11021694B2 and US10800821B2) , a property applicable to help further decrease CD47-binding at physiological pH or increase CD47-binding respectively at an acidic pH when combined with the backbone mutation (K53H, R69H, Q52H, K68H, Q52H+K68H, K53H+K68H, R69H+K68H, K53H+Q52H, R69H+Q52H, K53H+K68H+Q52H, or R69H+K68H+Q52H) . Together, these results showcase mutations at a large range of additional residues of SIRP IgV domain for developing combo mutations with the backbone mutation (K53H, R69H, Q52H, K68H, Q52H+K68H, K53H+K68H, R69H+K68H, K53H+Q52H, R69H+Q52H, K53H+K68H+Q52H, or R69H+K68H+Q52H) to further modulate CD47 binding at physiological and / or acidic pH, as shown in Table 19. Certain exemplary combo mutations are also provided in SEQ ID NO: 22 to 113. Additionally, the backbone mutation (K53H, R69H, Q52H, K68H, Q52H+K68H, K53H+K68H, R69H+K68H, K53H+Q52H, R69H+Q52H, K53H+K68H+Q52H, or R69H+K68H+Q52H) or the combo mutations as shown in Table 19 herein are amenable for combination with additional mutations of SIRP IgV domains known in the arts. For example, a backbone mutation (K53H, R69H, Q52H, K68H, Q52H+K68H, K53H+K68H, R69H+K68H, K53H+Q52H, R69H+Q52H, K53H+K68H+Q52H, or R69H+K68H+Q52H) or a combo mutations as shown in Table 19 herein is applicable to be added to any SIRP IgV-comprising protein such as ALX148, 1D4, 1A5, 2D3, 2A10, 2B5, 2A2, 2F5, FB3, FD6, FA4, CV1, AS1, AS2, SG3847, CD001, CD002, 106, IMM02, IMM0306, IMM2902, IMM2502, IMM4701C, IMM5601, IMM6101, IMM2505, HX009, DSP-107, JMT601, SL-172154, SG-404, SG-2501 or SG12473 (as set forth in Table 21) to generate de novo or modulate existing pH-sensitive CD47-binding of these protein. Table 19. Combination of additional mutation (s) with a backbone mutation Note: The “ / ” indicates different mutation changes at the corresponding amino acid position across SIRPα, SIRPβ and SIRPγIgV domains, due to different parent amino acid residues and / or position numbering for the mutation at the corresponding position, according to the SIRPα, SIRPβand SIRPγIgV amino acid sequences provided in SEQ ID NO: 57 to 75. Table 20. Production of additional single-mutation variant SIRPαV2 IgV-IgG1 Fc fusion protein Table 21. List of exemplary protein known in the arts that comprise a SIRPαIgV domain Example 20. pH-differential CD47-binding of multivalent fusion protein comprising variant SIRP IgV domains
[0195] In addition to the Fc-fusion protein comprising bivalent SIRP IgV monomers linked to the N-terminal of Fc, antibody fusion protein comprising tetravalent SIRP IgV monomers linked to the N-terminal of light chain and / or heavy chain of an antibody were also constructed as illustrated in Table 22, and were produced and analyzed by SDS-PAGE and SEC-HPLC, similarly as done in prior Examples. Table 22. Production of bispecific antibody fusion protein with tetravalent SIRP IgV variant domain
[0196] An ELISA assay was performed as similarly done in Example 13 and an FACS binding assay on Raji cells was performed similarly as done in Example 12, to analyze the CD47 binding of the bispecific antibody fusion protein PC2 to PC5 with tetravalent SIRP IgV variant domains, along with counterpart SIRP IgV-Fc fusion protein SIN300 to SIN302 with bivalent SIRP IgV variant domains, at an acidic pH versus at a physiological pH.
[0197] The ELISA result in FIG. 24A and FACS binding results in FIG. 24B-C showed all the protein with the SIRP IgV variant domains exhibited higher binding to CD47 at acidic pH 6.0 than at physiological pH. Moreover, for the protein with SIRP IgV variant domain comprising the same mutation (e.g. single mutation of K53H or R69H) but at different valency, the results in FIG. 24 showed that the protein comprising tetravalent SIRP IgV variant monomers with K53H mutation (PC2, PC4) or R69H mutation (PC3, PC5) exhibited increased CD47-binding at both acidic and physiological pH than the fusion protein comprising bivalent SIRP IgV variant monomers with the same K53H mutation (SIN-301) or R69H mutation (SIN-302) , and the increase at an acidic pH was generally larger or similar than the increase at physiological pH. Meanwhile, the tetravalent antibody fusion protein comprising four SIRP IgV monomers separately linked to the N-terminal of the VH and VL of the antibody as illustrated in format FV-44, exhibited generally higher CD47-binding at acidic and / or physiological pH than the antibody fusion protein comprising four SIRP IgV monomers with the same mutation wherein two monomers were serially linked together as illustrated in format FV-42, for example the protein PC4 versus PC2 comprising the same mutation of K53H, and the protein PC5 versus PC3 comprising the same mutation of R69H (FIG. 24) . Example 21. Construction of trispecific antibody fusion protein against PD-L1 / PD-1, VEGF and CD47 with SIRP variants with higher CD47-binding at an acidic pH than at physiological pH
[0198] To facilitate preferential binding of a trispecific antibody fusion protein to CD47 in tumor with acidic pH over ubiquitous CD47 in normal tissues with physiological pH, trispecific antibody fusion protein against PD-L1 / PD-1, VEGF and CD47 with higher CD47-binding at an acidic pH than at physiological pH were constructed by replacing the wild type SIRP IgV domain of the trispecific antibody fusion protein against PD-L1 / PD-1, VEGF and CD47 constructed in earlier Examples and Embodiment Sets, with an acid pH-sensitive SIRP IgV variant identified in Example 11-20 that exhibited higher binding to CD47 at an acidic pH than at physiological pH. As an example, a SIRPαV2 IgV variant domain comprising a mutation of K53H was first used to construct exemplary trispecific antibody fusion protein with bivalent SIRPαV2 IgV variant monomers (PVC145 in format FV-11, PVC147 and PVC149 in format FV-17 and PVC151 in format FV-19) and with tetravalent SIRPαV2 IgV variant monomers (PVC152 in FV-31, PVC153 in FV-34 and PVC162 in FV-37) as illustrated in Table 23, which were produced and analyzed by SDS-PAGE and SEC-HPLC analysis, similarly as done in Example 1-5. Except PVC151 and PVC149 showed a low SEC-HPLC purity of 44%and 76%after a single step protein A purification, all the other protein showed high SEC-HPLC purity. These results suggested generally consistent expression profile between a trispecific antibody protein with a SIRPαV2 IgV variant domain and its counterpart protein with a wild type SIRPαV2 IgV domain, as among the counterpart protein with a wild type SIRPαV2 IgV domain for these 6 proteins, only the counterpart protein for PVC151 (namely PVC23) also showed a low SEC-HPLC purity.
[0199] An ELISA assay was performed similarly as done in Example 13 to analyze the CD47 binding of the protein at an acidic pH 6.0 versus at a physiological pH 7.3. The ELISA results in FIG. 25 showed that PVC152 and PVC153 exhibited higher binding to CD47 at acidic pH 6.0 than at the physiological pH 7.3. Furthermore, both PVC152 and PVC153 exhibited CD47 binding similar to the benchmark bispecific antibody fusion protein PC2 with the same SIRP IgV variant domain at both pH 6.0 and pH 7.3, indicating consistent profile of pH-sensitive CD47 binding between the trispecific antibody protein against PD-L1, VEGF and CD47 with SIRP IgV variant domain and the counterpart bispecific antibody protein against PD-L1 and CD47 with the same SIRP IgV variant domain. Table 23. Production of trispecific antibody fusion protein against PD-L1, VEGF and CD47 with SIRP variants
[0200] While the present embodiments and examples have been particularly shown and described with reference to example embodiments herein, it will be understood by those of ordinary skill in the art that various changes in form and details may be made therein without departing from the spirit and scope of the present embodiments and examples as defined by the following claims. Those skilled in the art will recognize, or be able to ascertain using no more than routine experimentation, numerous equivalents to the specific procedures described herein. Such equivalents are considered to be within the scope of the present invention and are covered by the following claims. The contents of all non-patent literature publications, patents, and patent applications cited throughout this application are hereby incorporated by reference. The appropriate components, processes, and methods of those patents, applications and other documents may be selected for the present invention and embodiments thereof. REFERENCE Ansell, S.M., M.B. Maris, A.M. Lesokhin, R.W. Chen, I.W. Flinn, A. Sawas, M.D. Minden, D. Villa,M. M. Percival, A.S. Advani, J.M. Foran, S.M. Horwitz, M.G. Mei, J. Zain, K.J. Savage, C. Querfeld, O.E. Akilov, L. D.S. Johnson, T. Catalano, P.S. Petrova, R.A. Uger, E.L. Sievers, A. Milea, K. Roberge, Y. Shou and O.A. O'Connor (2021) . "Phase I Study of the CD47 Blocker TTI-621 in Patients with Relapsed or Refractory Hematologic Malignancies. " Clin Cancer Res 27 (8) : 2190-2199. Apte, R.S., D.S. Chen and N. Ferrara (2019) . "VEGF in Signaling and Disease: Beyond Discovery and Development. " Cell 176 (6) : 1248-1264. Barclay, A.N. and M.H. Brown (2006) . "The SIRP family of receptors and immune regulation. " Nat Rev Immunol 6 (6) : 457-464. Gillies, R.J., N. Raghunand, G.S. Karczmar and Z.M. Bhujwalla (2002) . "MRI of the tumor microenvironment. " J Magn Reson Imaging 16 (4) : 430-450. Golubovskaya, V., R. Berahovich, H. Zhou, S. Xu, H. Harto, L. Li, C.C. Chao, M. M. Mao and L. Wu (2017) . "CD47-CAR-T Cells Effectively Kill Target Cancer Cells and Block Pancreatic Tumor Growth. " Cancers (Basel) 9 (10) . Hatherley, D., S.C. Graham, J. Turner, K. Harlos, D.I. Stuart and A.N. Barclay (2008) . "Paired receptor specificity explained by structures of signal regulatory proteins alone and complexed with CD47. " Mol Cell 31 (2) : 266-277. Hatherley, D., S.M. Lea, S. Johnson and A.N. Barclay (2014) . "Polymorphisms in the human inhibitory signal-regulatory protein alpha do not affect binding to its ligand CD47. " J Biol Chem 289 (14) : 10024-10028. Lee, W.Y., D.A. Weber, O. Laur, E.A. Severson, I. McCall, R.P. Jen, A.C. Chin, T. Wu, K.M. Gernert and C.A. Parkos (2007) . "Novel structural determinants on SIRP alpha that mediate binding to CD47. " J Immunol 179 (11) : 7741-7750. Liu, X., L. Liu, Z. Ren, K. Yang, H. Xu, Y. Luan, K. Fu, J. Guo, H. Peng, M. Zhu and Y.X. Fu (2018) . "Dual Targeting of Innate and Adaptive Checkpoints on Tumor Cells Limits Immune Evasion. " Cell Rep 24 (8) : 2101-2111. Liu, Y., Q. Tong, Y. Zhou, H.W. Lee, J.J. Yang, H.J. Buhring, Y.T. Chen, B. Ha, C.X. Chen, Y. Yang and K. Zen (2007) . "Functional elements on SIRPalpha IgV domain mediate cell surface binding to CD47. " J Mol Biol 365 (3) : 680-693. Parvez, A., F. Choudhary, P. Mudgal, R. Khan, K.A. Qureshi, H. Farooqi and A. Aspatwar (2023) . "PD-1 and PD-L1: architects of immune symphony and immunotherapy breakthroughs in cancer treatment. " Front Immunol 14: 1296341. Puro, R.J., M.N. Bouchlaka, R.R. Hiebsch, B.J. Capoccia, M.J. Donio, P.T. Manning, W.A. Frazier, R.W. Karr and D.S. Pereira (2020) . "Development of AO-176, a Next-Generation Humanized Anti-CD47 Antibody with Novel Anticancer Properties and Negligible Red Blood Cell Binding. " Mol Cancer Ther 19 (3) : 835-846. Voets, E., M. Parade, D. Lutje Hulsik, S. Spijkers, W. Janssen, J. Rens, I. Reinieren-Beeren, G. van den Tillaart, S. van Duijnhoven, L. Driessen, M. Habraken, P. van Zandvoort, J. Kreijtz, P. Vink, A. van Elsas and H. van Eenennaam (2019) . "Functional characterization of the selective pan-allele anti-SIRPalpha antibody ADU-1805 that blocks the SIRPalpha-CD47 innate immune checkpoint. " J Immunother Cancer 7 (1) : 340. Wang, H., Y. Sun, X. Zhou, C. Chen, L. Jiao, W. Li, S. Gou, Y. Li, J. Du, G. Chen, W. Zhai, Y. Wu, Y. Qi and Y. Gao (2020) . "CD47 / SIRPalpha blocking peptide identification and synergistic effect with irradiation for cancer immunotherapy. " J Immunother Cancer 8 (2) . Zhang, W., Q. Huang, W. Xiao, Y. Zhao, J. Pi, H. Xu, H. Zhao, J. Xu, C. E. Evans and H. Jin (2020) . "Advances in Anti-Tumor Treatments Targeting the CD47 / SIRPalpha Axis. " Front Immunol 11: 18. Zhong, T., L. Zhang, Z. Huang, X. Pang, C. Jin, W. Liu, J. Du, W. Yin, N. Chen, J. Min, M. Xia and B. Li (2025) . "Design of a fragment crystallizable-engineered tetravalent bispecific antibody targeting programmed cell death-1 and vascular endothelial growth factor with cooperative biological effects. " iScience 28 (3) : 111722.
Claims
1.A multispecific molecule comprising:1)an antibody that binds to PD-L1 or PD-1; and2)an anti-VEGF single-chain domain, wherein the anti-VEGF single-chain domain comprises an amino acid sequence selected from the group consisting of SEQ ID NO: 17 to 30 and 451 to 453 and 451 to 455;wherein the anti-VEGF single-chain domain is linked through a linker to the N-terminal or C-terminal of the heavy chain or light chain ofthe antibody that binds to PD-L1 or PD-1.2.The multispecific molecule of claim 1, wherein the antibody that binds to PD-L1 or PD-1 comprises a VH of SEQ ID NO: 32 and a VL amino acid sequence of SEQ ID NO: 33; a VH amino acid sequence of SEQ ID NO:34 and a VL amino acid sequence of SEQ ID NO: 35; a VH amino acid sequence of SEQ ID NO: 36 and a VL amino acid sequence of SEQ ID NO: 37; a VH amino acid sequence of SEQ ID NO: 38 and a VL amino acid sequence of SEQ ID NO: 39; a VH amino acid sequence of SEQ ID NO: 40 and a VL amino acid sequence of SEQ ID NO: 41; or a VH amino acid sequence of SEQ ID NO: 42 and a VL amino acid sequence of SEQ ID NO: 43.3.The multispecific molecule any one of claims 1-2, wherein the linker comprises an amino acid sequence selected from group consisting of SEQ ID NO: 9-14 and 472-480.4.The multispecific molecule of claim 1, wherein the antibody that binds to PD-L1 or PD-1 comprises a VH of SEQ ID NO: 32 and a VL amino acid sequence of SEQ ID NO: 33, and the anti-VEGF single-chain domain comprises an amino acid sequence of SEQ ID NO: 17, 18, or 451, wherein the anti-VEGF single-chain domain is linked through a linker preferably to the C-terminal ofthe heavy chain ofthe antibody.5.The multispecific molecule of claim 1, wherein the antibody that binds to PD-L1 or PD-1 comprises a VH of SEQ ID NO: 34 and a VL amino acid sequence of SEQ ID NO: 35, and the anti-VEGF single-chain domain comprises an amino acid sequence of SEQ ID NO: 17, 18, or 451, wherein the anti-VEGF single-chain domain is linked through a linker preferably to the C-terminal ofthe heavy chain ofthe antibody.6.The multispecific molecule any one of claims 4-5, wherein the anti-VEGF single-chain domain comprises an amino acid sequence of SEQ ID NO: 17.7.The multispecific molecule any one of claims 4-6, wherein the linker comprises preferably 15 amino acids or less.8.The multispecific molecule any one of claims 4-6, wherein the linker comprises 10 amino acids or less.9.The multispecific molecule any one of claims 4-6, wherein the linker comprises 8 amino acids or less.10.The multispecific molecule any one of claims 4-6, wherein the linker comprises 5 amino acids or less.11.The multispecific molecule any one of claims 4-6, wherein the linker comprises 5 amino acids.12.The multispecific molecule any one of claims 4-6, wherein the linker comprises an amino acid sequence selected from the group consisting of SEQ ID NO: 9-10 and 472-480.13.The multispecific molecule any one of claims 4-6, wherein the linker comprises an amino acid sequence of SEQ ID NO: 473 or 476.14.The multispecific molecule any one of claims 1-3, wherein the linker comprises preferably an amino acid sequence of SEQ ID NO: 13 when the anti-VEGF single-chain domain is linked to the C-terminal of the light chain ofthe backbone anti-PD-L1 / PD-1 antibody.15.The multispecific molecule any one of claims 1-2, wherein the anti-VEGF single-chain domain is linked through a linker to the C-terminal of the heavy chain or light chain of the antibody that binds to PD-L1 or PD-1, when the anti-VEGF single-chain domain comprises an amino acid sequence of SEQ ID NO: 25 or 26.16.The multispecific molecule any one of claims 1-2, wherein the antibody that binds to PD-L1 or PD-1 comprises a VH amino acid sequence of SEQ ID NO: 32 and a VL amino acid sequence of SEQ ID NO: 33, when the anti-VEGF single-chain domain comprises an amino acid sequence of SEQ ID NO: 27, 28, 29 or 30.17.The multispecific molecule any one of claims 1-2, wherein the antibody that binds to PD-L1 or PD-1 comprises a VH amino acid sequence of SEQ ID NO: 34 and a VL amino acid sequence of SEQ ID NO: 35, when the anti-VEGF single-chain domain comprises an amino acid sequence of SEQ ID NO: 27, 28, 29 or 30.18.The multispecific molecule any one of claims 1-2, wherein the antibody that binds to PD-L1 or PD-1 comprises a VH amino acid sequence of SEQ ID NO: 38 and a VL amino acid sequence of SEQ ID NO: 39, when the anti-VEGF single-chain domain comprises an amino acid sequence of SEQ ID NO: 27, 28, 29 or 30.19.The multispecific molecule any one of claims 1-2, wherein the antibody that binds to PD-L1 or PD-1 comprises a VH amino acid sequence of SEQ ID NO: 40 and a VL amino acid sequence of SEQ ID NO: 41, when the anti-VEGF single-chain domain comprises an amino acid sequence of SEQ ID NO: 27, 28, 29 or 30.20.The multispecific molecule any one of claims 1-2, wherein the antibody that binds to PD-L1 or PD-1 comprises a VH amino acid sequence of SEQ ID NO: 42 and a VL amino acid sequence of SEQ ID NO: 43, when the anti-VEGF single-chain domain comprises an amino acid sequence of SEQ ID NO: 27, 28, 29 or 30.21.A multispecific molecule comprising:1)an antibody that binds to VEGF; and2)an anti-PD-L1 or anti-PD-1 single-chain domain, wherein the anti-PD-L1 or anti-PD-1 single-chain domain is linked through a linker to the N-terminal or C-terminal ofthe heavy chain or light chain ofthe antibody that binds to VEGF.22.The multispecific molecule of claim 21, wherein the antibody that binds to VEGF comprises a VH amino acid sequence of SEQ ID NO: 15 and a VL amino acid sequence of SEQ ID NO: 16; and the anti-PD-L1 or anti-PD-1 single-chain domain comprises an amino acid sequence selected from the group consisting of SEQ ID NO: 46 to 54 and 455.23.The multispecific molecule any one of claims 22, wherein the anti-PD-L1 or anti-PD-1 single-chain domain comprises an amino acid sequence of SEQ ID NO: 455 and is preferably linked to the C-terminal of the heavy chain ofthe antibody that binds to VEGF.24.The multispecific molecule of claim 22, wherein the anti-PD-L1 or anti-PD-1 single-chain domain comprises an amino acid sequence of SEQ ID NO: 46.25.The multispecific molecule of claim 22, wherein the anti-PD-L1 or anti-PD-1 single-chain domain comprises an amino acid sequence of SEQ ID NO: 47.26.The multispecific molecule of claim 22, wherein the anti-PD-L1 or anti-PD-1 single-chain domain comprises an amino acid sequence of SEQ ID NO: 50.27.The multispecific molecule of claim 22, wherein the anti-PD-L1 or anti-PD-1 single-chain domain comprises an amino acid sequence of SEQ ID NO: 51.28.The multispecific molecule of claim 22, wherein the anti-PD-L1 or anti-PD-1 single-chain domain comprises an amino acid sequence of SEQ ID NO: 52.29.The multispecific molecule of claim 22, wherein the anti-PD-L1 or anti-PD-1 single-chain domain comprises an amino acid sequence of SEQ ID NO: 53.30.The multispecific molecule any one of claims 21-29, wherein the linker linking the anti-PD-L or anti-PD-1 single-chain domain to the antibody that binds to VEGF comprises an amino acid sequence selected from group consisting of SEQ ID NO: 9-14 and 472-480.31.The multispecific molecule any one of claims 1-20, wherein the multispecific molecule comprises a CD47-binding domain, wherein the CD47-binding domain is linked through a linker to the the antibody that binds to PD-L1 or PD-L1 or the anti-VEGF single-chain domain.32.The multispecific molecule of claim 31, wherein the CD47-binding domain comprises a SIRP IgV monomer comprising an amino acid sequence selected from the group consisting of SEQ ID NO: 57 to 75.33.The multispecific molecule of claim 31, wherein the CD47-binding domain comprises a SIRP IgV multimer polypeptide comprising two SIRP IgV monomers linked in tandem through a linker of SEQ ID NO: 11 wherein each of the SIRP IgV monomers comprises an amino acid sequence selected from the group consisting of SEQ ID NO: 57 to 75.34.The multispecific molecule any one of claims 32-33, wherein the CD47-binding SIRP IgV monomer comprises an amino acid sequence of SEQ ID NO: 57.35.The multispecific molecule any one of claims 32-33, wherein the CD47-binding SIRP IgV monomer comprises an amino acid sequence of SEQ ID NO: 58.36.The multispecific molecule any one of claims 32-35, wherein the CD47-binding domain comprises a SIRPαor SIRPβIgV monomer comprising a mutation of I31Y or I31W, wherein the SIRP IgV variant monomer comprising the mutation exhibits enhanced binding to CD47 than the parent SIRP IgV monomer without the mutation.37.The multispecific molecule any one of claims 32-33, wherein the CD47-binding domain comprises a SIRPγmonomer comprising a mutation of L31Y or L31W, wherein the SIRP IgV variant monomer comprising the mutation exhibits enhanced binding to CD47 than the parent SIRP IgV monomer without the mutation.38.The multispecific molecule any one of claims 32-33, wherein the CD47-binding SIRP IgV monomer comprises a mutation of R59H, R59H+I31Y, R59H+I31W, R59H+L31Y or R59H+L31W, wherein the SIRP IgV variant monomer comprising the mutation exhibits enhanced binding to CD47 than the parent SIRP IgV monomer without the mutation.39.The multispecific molecule any one of claims 31-38, wherein the the antibody that binds to PD-L1 or PD-L1 comprises a Fc ofhuman IgG4 (SEQ ID NO: 1) , IgG2 (SEQ ID NO: 2) or silent IgG1 (SEQ ID NO: 4, 5,6 or 7) .40.The multispecific molecule any one of claims 31-38wherein the the antibody that binds to PD-L1 or PD-L1 comprises a Fc ofhuman IgG4 (SEQ ID NO: 1) .41.The multispecific molecule any one of claims 32-35, wherein the CD47-binding domain comprises a SIRPαor SIRPβIgV monomer comprising a mutation selected from the group consisting of K53H, R69H, K68H, Q52H, I31E, I31D, Q52H+K68H, K53H+K68H, R69H+K68H, K53H+Q52H, R69H+Q52H, K53H+I31E, R69H+I31E, K68H+I31E, Q52H+I31E, K53H+K68H+Q52H, R69H+K68H+Q52H, Q52H+K68H+I31E, K53H+I31D, R69H+I31D, K68H+I31D, Q52H+I31D, and Q52H+K68H+I31D, wherein the SIRP IgV variant monomer comprising the mutation exhibits higher binding to CD47 at an acidic pH than at physiological pH, wherein the acidic pH is<7.0 and the physiological pH is between 7.2 and 7.5.42.The multispecific molecule any one of claims 32-33, wherein the CD47-binding domain comprises a SIRPγIgV monomer comprising a mutation selected from the group consisting ofK53H, R69H, K68H, Q52H, L31E, L31D, Q52H+K68H, K53H+K68H, R69H+K68H, K53H+Q52H, R69H+Q52H, K53H+L31E, R69H+L31E, K68H+L31E, Q52H+L31E, K53H+K68H+Q52H, R69H+K68H+Q52H, Q52H+K68H+L31E, K53H+L31D, R69H+L31D, K68H+L31D, Q52H+L31D, Q52H+K68H+L31D, K53H+N101D, K53H+N101D+L37Q, K53H+N101D+L31E and K53H+N101D+L31D, wherein the SIRP IgV variant monomer comprising the mutation exhibits higher binding to CD47 at an acidic pH than at physiological pH, wherein the acidic pH is<7.0 and the physiological pH is between 7.2 and 7.5.43.The multispecific molecule any one of claims 41-42, wherein the CD47-binding domain comprises a SIRP IgV monomer comprising a mutation ofK53H.44.The multispecific molecule any one of claims 41-42, wherein the CD47-binding domain comprises a SIRP IgV monomer comprising a mutation ofR69H.45.The multispecific molecule any one of claims 41-42, wherein the CD47-binding domain comprises a SIRP IgV monomer comprising a mutation of Q52H.46.The multispecific molecule any one of claims 41-42, wherein the CD47-binding domain comprises a SIRP IgV monomer comprising a mutation ofK68H.47.The multispecific molecule any one of claims 41-42, wherein the CD47-binding domain comprises a SIRP IgV monomer comprising a mutation of Q52H+K68H.48.The multispecific molecule any one of claims 41-42, wherein the CD47-binding domain comprises a SIRP IgV monomer comprising a mutation ofK53H+I31E or K53H+L31E.49.The multispecific molecule of claim 42, wherein the CD47-binding SIRPγIgV monomer comprises a mutation of K53H+N101D.50.The multispecific molecule of claim 42, wherein the CD47-binding SIRPγIgV monomer comprises a mutation ofK53H+L37Q.51.The multispecific molecule of claim 42, wherein the CD47-binding SIRPγIgV monomer comprises a mutation ofK53H+N101D+L37Q.52.The multispecific molecule of claim 42, wherein the CD47-binding SIRPγIgV monomer comprises a mutation of K53H+N101D+L31E.53.The multispecific molecule of claim 42, wherein the CD47-binding SIRPγIgV monomer comprises a mutation of K53H+N101D+L31D.54.The multispecific molecule any one of claims 41-53, wherein the anti-PD-L1 or anti-PD-1 single-chain domain comprises a Fc of human IgG4 (SEQ ID NO: 1) , active IgG1 (SEQ ID NO: 8) , silent IgG1 (SEQ ID NO:4, 5, 6 or 7) or IgG2 (SEQ ID NO: 2) .55.The multispecific molecule any one of claims 41-53, wherein the antibody that binds to PD-L1 comprises a Fc of active human IgG1 (SEQ ID NO: 8) .56.The multispecific molecule any one of claims 41-53, wherein the antibody that binds to PD-L1 or PD-1 comprises a Fc ofhuman IgG4 (SEQ ID NO: 1) .57.The multispecific molecule any one of claims 31-56, wherein the CD47-binding domain is linked through a linker to the N-terminal of the light chain of the antibody that binds to PD-L1 or PD-1 and the anti-VEGF single-chain domain is linked through a linker to the C-terminal of the light chain or heavy chain of the antibody that binds to PD-L1 or PD-1.58.The multispecific molecule any one of claims 31-56, wherein the CD47-binding domain is linked through a linker to the N-terminal of the heavy chain of the antibody that binds to PD-L1 or PD-1 and the anti-VEGF single-chain domain is linked through a linker to the C-terminal of the light chain or heavy chain of the antibody that binds to PD-L1 or PD-1.59.The multispecific molecule any one of claims 31-56, wherein the CD47-binding domain is linked through a linker to the N-terminal of the light chain of the antibody that binds to PD-L1 or PD-1 and the anti-VEGF single-chain domain is linked through a linker to the N-terminal of the heavy chain ofthe antibody that binds to PD-L1 or PD-1.60.The multispecific molecule any one of claims 31-56, wherein the CD47-binding domain is linked through a linker to the N-terminal of the heavy chain of the antibody that binds to PD-L1 or PD-1 and the anti-VEGF single-chain domain is linked through a linker to the N-terminal ofthe light chain ofthe antibody that binds to PD-L1 or PD-1.61.The multispecific molecule any one of claims 31-60, wherein the linker comprises an amino acid sequence selected from group consisting of SEQ ID NO: 9-14 and 472-480.62.The multispecific molecule of 61, wherein the linker comprises an amino acid sequence of SEQ ID NO: 11.63.The multispecific molecule any one of claims 31-62, wherein the anti-VEGF single-chain domain comprises an amino acid sequence of SEQ ID NO: 17, 18, 24, 25, 26, 27, 28, 29, 30 or 451, when the antibody that binds to PD-L1 comprises a VH amino acid sequence of SEQ ID NO: 32 and a VL amino acid sequence of SEQ ID NO: 33.64.The multispecific molecule any one of claims 31-62, wherein the anti-VEGF single-chain domain comprises an amino acid sequence of SEQ ID NO: 17, 18, 24, 25, 26, 27, 28, 29, 30 or 451, when the antibody that binds to PD-L1 comprises a VH amino acid sequence of SEQ ID NO: 34 and a VL amino acid sequence of SEQ ID NO: 35.65.The multispecific molecule any one of claims 31-62, wherein the anti-VEGF single-chain domain comprises an amino acid sequence of SEQ ID NO: 17, 18, 24, 25, 26, 27, 28, 29, 30 or 451, when the antibody that binds to PD-1 comprises a VH amino acid sequence of SEQ ID NO: 38 and a VL amino acid sequence of SEQ ID NO: 39.66.The multispecific molecule any one of claims 31-62, wherein the anti-VEGF single-chain domain comprises an amino acid sequence of SEQ ID NO: 17, 18, 24, 25, 26, 27, 28, 29, 30 or 451, when the antibody that binds to PD-1 comprises a VH amino acid sequence of SEQ ID NO: 40 and a VL amino acid sequence of SEQ ID NO: 41.67.The multispecific molecule any one of claims 63-66, wherein the anti-VEGF single-chain domain comprises an amino acid sequence of SEQ ID NO: 17.68.The multispecific molecule any one of claims 21-30, wherein the multispecific molecule comprises a CD47-binding domain, wherein the CD47-binding domain is linked through a linker to the antibody that binds to VEGF, or to the anti-PD-L1 or anti-PD-1 single-chain domain.69.The multispecific molecule of claim 68, wherein the CD47-binding domain comprises a SIRP IgV monomer comprising an amino acid sequence selected from the group consisting of SEQ ID NO: 57 to 75.70.The multispecific molecule of claim 68, wherein the CD47-binding domain comprises a SIRP IgV multimer polypeptide comprising two SIRP IgV monomers linked in tandem through a linker of SEQ ID NO: 11 wherein each of the SIRP IgV monomers comprises an amino acid sequence selected from the group consisting of SEQ ID NO: 57 to 75.71.The multispecific molecule any one of claims 68-70, wherein the CD47-binding SIRP IgV monomer comprises an amino acid sequence of SEQ ID NO: 57.72.The multispecific molecule any one of claims 68-70, wherein the CD47-binding SIRP IgV monomer comprises an amino acid sequence of SEQ ID NO: 58.73.The multispecific molecule any one of claims 68-72, wherein the CD47-binding domain comprises a SIRPαor SIRPβIgV monomer comprising a mutation of I31Y or I31W, wherein the SIRP IgV variant monomer comprising the mutation exhibits enhanced binding to CD47 than the parent SIRP IgV monomer without the mutation.74.The multispecific molecule any one of claims 68-70, wherein the CD47-binding domain comprises a SIRPγmonomer comprising a mutation of L31Y or L31W, wherein the SIRP IgV variant monomer comprising the mutation exhibits enhanced binding to CD47 than the parent SIRP IgV monomer without the mutation.75.The multispecific molecule any one of claims 68-72, wherein the CD47-binding SIRP IgV monomer further comprises a mutation of R59H, R59H+I31Y, R59H+I31W, R59H+L31Y or R59H+L31W, wherein the SIRP IgV variant monomer comprising the mutation exhibits enhanced binding to CD47 than the parent SIRP IgV monomer without the mutation.76.The multispecific molecule any one of claims 68-75, wherein the antibody that binds to VEGF comprises a Fc ofhuman IgG4 (SEQ ID NO: 1) , IgG2 (SEQ ID NO: 2) or silent IgG1 (SEQ ID NO: 4, 5, 6 or 7) .77.The multispecific molecule any one of claims 62-69, wherein the antibody that binds to VEGF comprises a Fc ofhuman IgG4 (SEQ ID NO: 1) .78.The multispecific molecule any one of claims 68-72, wherein the CD47-binding domain comprises a SIRPαor SIRPβIgV monomer comprising a mutation selected from the group consisting of K53H, R69H, K68H, Q52H, I31E, I31D, Q52H+K68H, K53H+K68H, R69H+K68H, K53H+Q52H, R69H+Q52H, K53H+I31E, R69H+I31E, K68H+I31E, Q52H+I31E, K53H+K68H+Q52H, R69H+K68H+Q52H, Q52H+K68H+I31E, K53H+I31D, R69H+I31D, K68H+I31D, Q52H+I31D, and Q52H+K68H+I31D, wherein the SIRP IgV variant monomer comprising the mutation exhibits higher binding to CD47 at an acidic pH than at physiological pH, wherein the acidic pH is<7.0 and the physiological pH is between 7.2 and 7.5.79.The multispecific molecule any one of claims 68-70, wherein the CD47-binding domain comprises a SIRPγIgV monomer comprising a mutation selected from the group consisting ofK53H, R69H, K68H, Q52H, L31E, L31D, Q52H+K68H, K53H+K68H, R69H+K68H, K53H+Q52H, R69H+Q52H, K53H+L31E, R69H+L31E, K68H+L31E, Q52H+L31E, K53H+K68H+Q52H, R69H+K68H+Q52H, Q52H+K68H+L31E, K53H+L31D, R69H+L31D, K68H+L31D, Q52H+L31D, Q52H+K68H+L31D, K53H+N101D, K53H+N101D+L37Q, K53H+N101D+L31E and K53H+N101D+L31D, wherein the SIRP IgV variant monomer comprising the mutation exhibits higher binding to CD47 at an acidic pH than at physiological pH, wherein the acidic pH is<7.0 and the physiological pH is between 7.2 and 7.5.80.The multispecific molecule any one of claims 78-79, wherein the CD47-binding domain comprises a SIRP IgV monomer comprising a mutation ofK53H.81.The multispecific molecule any one of claims 78-79, wherein the CD47-binding domain comprises a SIRP IgV monomer comprising a mutation ofR69H.82.The multispecific molecule any one of claims 78-79, wherein the CD47-binding domain comprises a SIRP IgV monomer comprising a mutation of Q52H.83.The multispecific molecule any one of claims 78-79, wherein the CD47-binding domain comprises a SIRP IgV monomer comprising a mutation ofK68H.84.The multispecific molecule any one of claims 78-79, wherein the CD47-binding domain comprises a SIRP IgV monomer comprising a mutation of Q52H+K68H.85.The multispecific molecule any one of claims 78-79, wherein the CD47-binding domain comprises a SIRP IgV monomer comprising a mutation ofK53H+I31E or K53H+L31E.86.The multispecific molecule of claim 79, wherein the CD47-binding SIRPγIgV monomer comprises a mutation of K53H+N101D.87.The multispecific molecule of claim 79, wherein the CD47-binding SIRPγIgV monomer comprises a mutation ofK53H+L37Q.88.The multispecific molecule of claim 79, wherein the CD47-binding SIRPγIgV monomer comprises a mutation ofK53H+N101D+L37Q.89.The multispecific molecule of claim 79, wherein the CD47-binding SIRPγIgV monomer comprises a mutation of K53H+N101D+L31E.90.The multispecific molecule of claim 79, wherein the CD47-binding SIRPγIgV monomer comprises a mutation of K53H+N101D+L31D.91.The multispecific molecule any one of claims 78-90, wherein the antibody that binds to VEGF comprises a Fc of human IgG4 (SEQ ID NO: 1) , active IgG1 (SEQ ID NO: 8) , silent IgG1 (SEQ ID NO: 4, 5, 6 or 7) or IgG2 (SEQ ID NO: 2) .92.The multispecific molecule any one of claims 78-90, wherein the antibody that binds to VEGF comprises a Fc ofhuman IgG4 (SEQ ID NO: 1) .93.The multispecific molecule any one of claims 78-90, wherein the antibody that binds to VEGF comprises a Fc of active human IgG1 (SEQ ID NO: 8) when the anti-PD-L1 or anti-PD-1 single-chain domain is an anti-PD-L1 single-chain domain.94.The multispecific molecule any one of claims 68-93, wherein the CD47-binding domain is linked through a linker to the N-terminal of the light chain of the antibody that binds to VEGF and the anti-PD-L1 or anti-PD-1 single-chain domain is linked through a linker to the C-terminal of the heavy chain of the antibody that binds to VEGF.95.The multispecific molecule any one of claims 68-93, wherein the CD47-binding domain is linked through a linker to the N-terminal of the heavy chain of the antibody that binds to VEGF and the anti-PD-L1 or anti-PD-1 single-chain domain is linked through a linker to the C-terminal of the heavy chain of the antibody that binds to VEGF.96.The multispecific molecule any one of claims 68-93, wherein the anti-PD-L1 or anti-PD-1 single-chain domain is linked through a linker to the C-terminal of the heavy chain ofthe antibody that binds to VEGF and the CD47-binding domain is linked through a linker to the C-terminal of the anti-PD-L1 or anti-PD-1 single-chain domain.97.The multispecific molecule any one of claims 68-93, wherein the CD47-binding domain is linked through a linker to the C-terminal of the heavy chain of the antibody that binds to VEGF and the anti-PD-L1 or anti-PD-1 single-chain domain is linked through a linker to the C-terminal ofthe CD47-binding domain.98.The multispecific molecule any one of claims 68-93, wherein the anti-PD-L1 or anti-PD-1 single-chain domain is linked through a linker to the N-terminal ofthe heavy chain ofthe antibody that binds to VEGF and the CD47-binding domain is linked through a linker to the C-terminal ofthe heavy chain of the antibody that binds to VEGF.99.The multispecific molecule any one of claims 68-93, wherein the anti-PD-L1 or anti-PD-1 single-chain domain is linked through a linker to the N-terminal of the light chain of the antibody that binds to VEGF and the CD47-binding domain is linked through a linker to the C-terminal ofthe heavy chain of the antibody that binds to VEGF.100.The multispecific molecule any one of claims 68-93, wherein the CD47-binding domain is linked through a linker to the N-terminal of the light chain ofthe antibody that binds to VEGF and the anti-PD-L1 or anti-PD-1 single-chain domain is linked through a linker to the N-terminal of the heavy chain of the antibody that binds to VEGF.101.The multispecific molecule any one of claims 68-93, wherein the CD47-binding domain is linked through a linker to the N-terminal of the heavy chain of the antibody that binds to VEGF and the anti-PD-L1 or anti-PD-1 single-chain domain is linked through a linker to the N-terminal ofthe light chain ofthe antibody that binds to VEGF.102.The multispecific molecule any one of claims 68-101, wherein the linker comprises an amino acid sequence selected from group consisting of SEQ ID NO: 9-14 and 472-480.103.The multispecific molecule any one of claims 68-101, wherein the linker comprises an amino acid sequence of SEQ ID NO: 11.104.The multispecific molecule any one of claims 68-103, wherein the anti-PD-L1 or anti-PD-1 single-chain domain comprises an amino acid sequence of46.105.The multispecific molecule any one of claims 68-103, wherein the anti-PD-L1 or anti-PD-1 single-chain domain comprises an amino acid sequence of455.106.The multispecific molecule any one of claims 68-103, wherein the anti-PD-L1 or anti-PD-1 single-chain domain comprises an amino acid sequence of SEQ ID NO: 50.107.A multispecific molecule comprising:1) a CD47-binding domain, wherein the CD47-binding domain comprises a SIRP IgV monomer or a SIRP IgV multimer polypeptide comprising two SIRP IgV monomers linked in tandem through a linker of SEQ ID NO: 11, wherein the SIRP IgV monomer comprising an amino acid sequence selected from the group consisting of SEQ ID NO: 57 to 75.2) a second domain that binds to VEGF; and3) a third domain that binds to PD-L1 or PD-1.108.The multispecific molecule of claim 107, wherein the CD47-binding domain comprises a SIRPαor SIRPβIgV monomer comprising a mutation of I31Y or I31W.109.The multispecific molecule of claim 107, wherein the CD47-binding domain comprises a SIRPγIgV monomer comprising a mutation ofL31Y or L31W.110.The multispecific molecule of claim 107, wherein the CD47-binding domain comprises a SIRP IgV monomer comprising a mutation ofR59H, R59H+I31Y, R59H+I31W, R59H+L31Y or R59H+L31W.111.The multispecific molecule of claim 107, wherein the CD47-binding domain comprises a SIRPαor SIRPβIgV monomer comprising a mutation selected from the group consisting of K53H, R69H, K68H, Q52H, I31E, I31D, Q52H+K68H, K53H+K68H, R69H+K68H, K53H+Q52H, R69H+Q52H, K53H+I31E, R69H+I31E, K68H+I31E, Q52H+I31E, K53H+K68H+Q52H, R69H+K68H+Q52H, Q52H+K68H+I31E, K53H+I31D, R69H+I31D, K68H+I31D, Q52H+I31D, and Q52H+K68H+I31D.112.The multispecific molecule of claim 107, wherein the CD47-binding domain comprises a SIRPγIgV monomer comprising a mutation selected from the group consisting of K53H, R69H, K68H, Q52H, L31E, L31D, Q52H+K68H, K53H+K68H, R69H+K68H, K53H+Q52H, R69H+Q52H, K53H+L31E, R69H+L31E, K68H+L31E, Q52H+L31E, K53H+K68H+Q52H, R69H+K68H+Q52H, Q52H+K68H+L31E, K53H+L31D, R69H+L31D, K68H+L31D, Q52H+L31D, Q52H+K68H+L31D, K53H+N101D, K53H+L307Q, K53H+N101D+L37Q, K53H+N101D+L31E, K53H+N101D+L31D, K53H+L37Q+L31E, K53H+L37Q+L31D, K53H+N101D+L37Q+L31E and K53H+N101D+L37Q+L31D.113.The multispecific molecule any one of claims 111-112, wherein the CD47-binding domain comprises a SIRP IgV monomer comprising a mutation ofK53H.114.The multispecific molecule any one of claims 111-112, wherein the CD47-binding domain comprises a SIRP IgV monomer comprising a mutation ofR69H.115.The multispecific molecule any one of claims 111-112, wherein the CD47-binding domain comprises a SIRP IgV monomer comprising a mutation of Q52H.116.The multispecific molecule any one of claims 111-112, wherein the CD47-binding domain comprises a SIRP IgV monomer comprising a mutation ofK68H.117.The multispecific molecule any one of claims 111-112, wherein the CD47-binding domain comprises a SIRP IgV monomer comprising a mutation of Q52H+K68H.118.The multispecific molecule any one of claims 111-112, wherein the CD47-binding domain comprises a SIRP IgV monomer comprising a mutation ofK53H+I31E or K53H+L31E.119.The multispecific molecule of claim 112, wherein the CD47-binding domain comprises a SIRPγIgV monomer comprising a mutation ofK53H+N101D.120.The multispecific molecule of claim 112, wherein the CD47-binding domain comprises a SIRPγIgV monomer comprising a mutation ofK53H+L37Q.121.The multispecific molecule of claim 112, wherein the CD47-binding domain comprises a SIRPγIgV monomer comprising a mutation ofK53H+N101D+L37Q.122.The multispecific molecule of claim 112, wherein the CD47-binding domain comprises a SIRPγIgV monomer comprising a mutation ofK53H+N101D+l31E.123.The multispecific molecule of claim 112, wherein the CD47-binding domain comprises a SIRPγIgV monomer comprising a mutation ofK53H+N101D+l31D.124.The multispecific molecule any one of claims 107-123, wherein the second domain that binds to VEGF is an anti-VEGF single-chain domain comprising an amino acid sequence selected from the group consisting of SEQ ID NO: 17 to 30 and 451 to 453, and the third domain is an antibody that binds to PD-L1 or PD-1 comprising a VH amino acid sequence of SEQ ID NO: 32 and a VL amino acid sequence of SEQ ID NO: 33, A VH sequence of SEQ ID NO: 34 and a VL sequence of SEQ ID NO: 35, a VH sequence of SEQ ID NO: 36 and a VL sequence of SEQ ID NO: 37, a VH sequence of SEQ ID NO: 38 and a VL sequence of SEQ ID NO: 39, A VH sequence of SEQ ID NO: 40 and a VL sequence of SEQ ID NO: 41, or a VH sequence of SEQ ID NO: 42 and a VL sequence of SEQ ID NO: 43, wherein both the CD47-binding domain and the anti-VEGF single-chain domain are linked through a linker to the antibody that binds to PD-L1 or PD-1.125.The multispecific molecule of claim 124, wherein selected from the group consisting of SEQ ID NO: 9-14 and 472-480the linker comprises an amino acid sequence selected from group consisting of SEQ ID NO: 9-14 and 472-480.126.The multispecific molecule of claim 124, wherein the linker comprises an amino acid sequence of SEQ ID NO: 11.127.The multispecific molecule of claim 124, wherein the second domain that binds to VEGF is an anti-VEGF single-chain domain comprising an amino acid sequence of SEQ ID NO: 17.128.The multispecific molecule of claim 124, wherein the second domain that binds to VEGF is an anti-VEGF single-chain domain comprising an amino acid sequence of SEQ ID NO: 26.129.The multispecific molecule of claim 124, wherein the second domain that binds to VEGF is an anti-VEGF single-chain domain comprising an amino acid sequence of SEQ ID NO: 27.130.The multispecific molecule of claim 124, wherein the second domain that binds to VEGF is an anti-VEGF single-chain domain comprising an amino acid sequence of SEQ ID NO: 29.131.The multispecific molecule of claim 124, wherein the second domain that binds to VEGF is an anti-VEGF single-chain domain comprising an amino acid sequence of SEQ ID NO: 30.132.The multispecific molecule of claim 124, wherein the third domain that binds to PD-L1 or PD-1 is an antibody comprising a VH amino acid sequence of SEQ ID NO: 32 and a VL amino acid sequence of SEQ ID NO: 33.133.The multispecific molecule of claim 124, wherein the third domain that binds to PD-L1 or PD-1 is an antibody comprising a VH amino acid sequence of SEQ ID NO: 34 and a VL amino acid sequence of SEQ ID NO: 35.134.The multispecific molecule of claim 124, wherein the third domain that binds to PD-L1 or PD-1is an antibody comprising a VH amino acid sequence of SEQ ID NO: 38 and a VL amino acid sequence of SEQ ID NO: 39.135.The multispecific molecule of claim 124, wherein the third domain that binds to PD-L1 or PD-1 is an antibody comprising a VH amino acid sequence of SEQ ID NO: 40 and a VL amino acid sequence of SEQ ID NO: 41.136.The multispecific molecule of claim 124, wherein the third domain that binds to PD-L1 or PD-1 is an antibody comprising a VH amino acid sequence of SEQ ID NO: 42 and a VL amino acid sequence of SEQ ID NO: 43.137.The multispecific molecule any one of claims 124-136, wherein the second domain that binds to VEGF is linked through a linker to the N-terminal or C-terminal of the heavy chain of the third domain of antibody that binds to PD-L1 or PD-1, and the CD47-binding domain is linked through a linker to the N-terminal or C-terminal ofthe light chain ofthe third domain ofantibody that binds to PD-L1 or PD-1.138.The multispecific molecule any one of claims 124-136, wherein the second domain that binds to VEGF is linked through a linker to the N-terminal or C-terminal of the light chain of the third domain of antibody that binds to PD-L1 or PD-1, and the CD47-binding domain is linked through a linker to the N-terminal or C-terminal ofthe heavy chain ofthe third domain of antibody that binds to PD-L1 or PD-1.139.The multispecific molecule any one of claims 137-138, wherein the CD47-binding domain is linked through a linker to the N-terminal of the light chain of the third domain of antibody that binds to PD-L1 or PD-1 and the second domain that binds to VEGF is linked through a linker to the C-terminal ofthe light chain or heavy chain ofthe third domain of antibody that binds to PD-L1 or PD-L1.140.The multispecific molecule any one of claims 137-138, wherein the CD47-binding domain is linked through a linker to the N-terminal of the heavy chain of the third domain of antibody that binds to PD-L1 or PD-1 and the second domain that binds to VEGF is linked through a linker to the C-terminal ofthe light chain or heavy chain ofthe third domain of antibody that binds to PD-L1 or PD-L1.141.The multispecific molecule any one of claims 137-138, wherein the CD47-binding domain is linked through a linker to the N-terminal of the light chain of the third domain of antibody that binds to PD-L1 or PD-1 and the second domain that binds to VEGF is linked through a linker to the N-terminal of the heavy chain ofthe third domain of antibody that binds to PD-L1 or PD-L1.142.The multispecific molecule any one of claims 137-138, wherein the CD47-binding domain is linked through a linker to the N-terminal of the heavy chain of the third domain of antibody that binds to PD-L1 or PD-1 and the second domain that binds to VEGF is linked through a linker to the N-terminal ofthe light chain ofthe third domain ofantibody that binds to PD-L1 or PD-L1.143.The multispecific molecule any one of claims 107-123, wherein the second domain that binds to VEGF is an anti-VEGF antibody comprising a VH sequence of SEQ ID NO: 15 and a VL sequence of SEQ ID NO: 16,and the third domain that binds to PD-L1 or PD-1 comprises a single-chain domain comprising an amino acid sequence selected from the group consisting of SEQ ID NO: 46 to 54 and 455, wherein both the CD47-binding domain and the third domain comprising a single-chain domain that binds to PD-L1 or PD-1 are linked through a linker to N-terminal or C-terminal of the heavy chain or light chain ofthe second domain of antibody that binds to VEGF.144.The multispecific molecule of claim 143, wherein the linker comprises an amino acid sequence selected from group consisting of SEQ ID NO: 9-14 and 472-480.145.The multispecific molecule of claim 143, wherein the linker comprises an amino acid sequence of SEQ ID NO: 11.146.The multispecific molecule of claim 143, wherein the third domain that binds to PD-L1 or PD-1 comprises an anti-PD-L1 or anti-PD-1 single-chain domain comprising an amino acid sequence of455.147.The multispecific molecule of claim 143, wherein the third domain that binds to PD-L1 or PD-1 comprises an anti-PD-L1 or anti-PD-1 single-chain domain comprising an amino acid sequence of46.148.The multispecific molecule of claim 143 wherein the third domain that binds to PD-L1 or PD-1 comprises an anti-PD-L1 or anti-PD-1 single-chain domain comprising an amino acid sequence of SEQ ID NO: 47.149.The multispecific molecule of claim 143, wherein the third domain that binds to PD-L1 or PD-1 is an anti-PD-L1 or anti-PD-1 single-chain domain comprising an amino acid sequence of SEQ ID NO: 48.150.The multispecific molecule of claim 143, wherein the third domain that binds to PD-L1 or PD-1 is an anti-PD-L1 or anti-PD-1 single-chain domain comprising an amino acid sequence of SEQ ID NO: 49.151.The multispecific molecule of claim 143, wherein the third domain that binds to PD-L1 or PD-1 is an anti-PD-L1 or anti-PD-1 single-chain domain comprising an amino acid sequence of SEQ ID NO: 50.152.The multispecific molecule of claim 143, wherein the third domain that binds to PD-L1 or PD-1 is an anti-PD-L1 or anti-PD-1 single-chain domain comprising an amino acid sequence of SEQ ID NO: 51.153.The multispecific molecule of claim 143, wherein the third domain that binds to PD-L1 or PD-1 is an anti-PD-L1 or anti-PD-1 single-chain domain comprising an amino acid sequence of SEQ ID NO: 52.154.The multispecific molecule of claim 143, wherein the third domain that binds to PD-L1 or PD-1 is an anti-PD-L1 or anti-PD-1 single-chain domain comprising an amino acid sequence of SEQ ID NO: 53.155.The multispecific molecule of claim 143, wherein the third domain that binds to PD-L1 or PD-1is an anti-PD-L1 or anti-PD-1 single-chain domain comprising an amino acid sequence of SEQ ID NO: 54.156.The multispecific molecule any one of claims 143-155, wherein the CD47-binding domain is linked through a linker to the N-terminal of the light chain ofthe second domain of antibody that binds to VEGF and the third domain that binds to PD-L1 or PD-1 is linked through a linker to the C-terminal of the heavy chain ofthe second domain of antibody that binds to VEGF.157.The multispecific molecule any one of claims 143-155, wherein the CD47-binding domain is linked through a linker to the N-terminal of the heavy chain of the second domain of antibody that binds to VEGF and the third domain that binds to PD-L1 or PD-1 is linked through a linker to the C-terminal of the heavy chain ofthe second domain of antibody that binds to VEGF.158.The multispecific molecule any one of claims 143-155, wherein the third domain that binds to PD-L1 or PD-1 is linked through a linker to the C-terminal of the heavy chain of the second domain of antibody that binds to VEGF and the CD47-binding domain is linked through a linker to the C-terminal of the third domain that binds to PD-L1 or PD-1.159.The multispecific molecule any one of claims 143-155, wherein the CD47-binding domain is linked through a linker to the C-terminal of the heavy chain of the second domain of antibody that binds to VEGF and the third domain that binds to PD-L1 or PD-1 is linked through a linker to the C-terminal of the CD47-binding SIRP IgV domain.160.The multispecific molecule any one of claims 143-155, wherein the third domain that binds to PD-L1 or PD-1 is linked through a linker to the N-terminal of the heavy chain of the second domain of antibody that binds to VEGF and the CD47-binding domain is linked through a linker to the C-terminal of the heavy chain ofthe second domain of antibody that binds to VEGF.161.The multispecific molecule any one of claims 143-155, wherein the third domain that binds to PD-L1 or PD-1 is linked through a linker to the N-terminal of the light chain of the second domain of antibody that binds to VEGF and the CD47-binding domain is linked through a linker to the C-terminal of the heavy chain ofthe second domain of antibody that binds to VEGF.162.The multispecific molecule any one of claims 143-155, wherein the CD47-binding domain is linked through a linker to the N-terminal of the light chain ofthe second domain of antibody that binds to VEGF and the third domain that binds to PD-L1 or PD-1 is linked through a linker to the N-terminal of the heavy chain ofthe second domain of antibody that binds to VEGF.163.The multispecific molecule of claim 124, wherein the third domain of antibody that binds to PD-L1 or PD-1 comprises a Fc of human IgG4 (SEQ ID NO: 1) , silent IgG1 (SEQ ID NO: 4, 5, 6 or 7) , IgG2 (SEQ ID NO: 2) or active IgG1 (SEQ ID NO: 8) .164.The multispecific molecule of claim 124, wherein the third domain of antibody that binds to PD-L1 or PD-1comprises a Fc ofhuman IgG4 (SEQ ID NO: 1) .165.The multispecific molecule claim 124, wherein the third domain of antibody that binds to PD-L1 or PD-1 comprises a Fc of silent human IgG1 (SEQ ID NO: 5) .166.The multispecific molecule claim 124, wherein the third domain of antibody that binds to PD-L1 or PD-1 comprises a Fc of active human IgG1 (SEQ ID NO: 8) when the third domain of antibody binds to PD-L1 and the CD47-binding domain comprises a SIRP IgV domain comprising a mutation any one of claims 111-123.167.The multispecific molecule of claim 143, wherein the second domain of antibody that binds to VEGF comprises a Fc ofhuman IgG4 (SEQ ID NO: 1) , silent IgG1 (SEQ ID NO: 4, 5, 6 or 7) , IgG2 (SEQ ID NO: 2) or active IgG1 (SEQ ID NO: 8) .168.The multispecific molecule of claim 143, wherein the second domain of antibody that binds to VEGF comprises a Fc ofhuman IgG4 (SEQ ID NO: 1) .169.The multispecific molecule claim 143, wherein the second domain of antibody that binds to VEGF comprises a Fc of silent human IgG1 (SEQ ID NO: 5) .170.The multispecific molecule claim 143, wherein the second domain of antibody that binds to VEGF comprises a Fc of active human IgG1 (SEQ ID NO: 8) when the third domain binds to PD-L1 and the CD47-binding domain comprises a SIRP IgV domain comprising a mutation any one ofclaims 111-123.171.A conjugate comprising a multispecific molecule of any one any one of claims 1-170.172.A nucleic acid comprising a sequence encoding a polypeptide of a multispecific molecule of any one any one of claims 1-170.173.An expression vector, comprising a nucleic acid of claim 172.174.A modified cell comprising a nucleic acid of claim 172 and / or an expression vector of claim 173.175.A pharmaceutical composition comprising a multispecific molecule of any one any one of claims 1-170, a conjugate of claim 171, a nucleic acid of claim 172, a vector of claim 173, a modified cell of claim 174, and a pharmaceutically acceptable carrier.176.A method of treating a cancer in a mammal comprising administering an effective amount of a multispecific molecule of any one any one of claims 1-170, a conjugate of claim 171, a nucleic acid of claim 172, a vector of claim 173, a modified cell of claim 174, and / or a pharmaceutical composition of claim 175, to a mammal in need thereof.177.The method of claim 176, wherein the cancer comprises ovarian cancer, endometrial cancer, uterine cancer, cervical cancer, breast cancer, gastric cancer, esophageal cancer, lung cancer, head and neck cancer, colorectal cancer, liver cancer, bone cancer, sarcoma, osteosarcoma, brain cancer, multiple myeloma, acute myeloid leukemia (AML) , myelodysplastic syndrome, adrenal cancer, anal cancer, bile duct cancer, bladder cancer, basal cell skin cancer, chondrosarcoma, Ewing’s sarcoma, eye cancer, gallbladder cancer, gastrointestinal carcinoid tumors, gastrointestinal stromal tumor (GIST) , glioma, glioblastoma, hepatocellular cancer, kaposi sarcoma, kidney cancer, laryngeal and hypopharyngeal cancer, leiomyosarcoma, non-small cell lung cancer (NSCLC) , small cell lung cancer (SCLC) , lung carcinoid tumor, mesothelioma, Merkel cell carcinoma, melanoma, nasal cavity and paranasal sinus cancer, nasopharyngeal cancer, neuroendocrine cancer, neuroblastoma, oral cavity and oropharyngeal cancer, pancreatic cancer, penile cancer, pituitary tumors, prostate cancer, renal cancer, retinoblastoma, rhabdomyosarcoma, salivary gland cancer, squamous cell skin cancer, small intestine cancer, stomach cancer, testicular cancer, thymus cancer, thyroid cancer, uterine sarcoma, urinary tract cancer, vaginal cancer, vulvar cancer, Wilms’ tumor, diffuse large B-cell lymphoma (DLBCL) , follicular lymphoma, hairy-cell leukemia, Hodgkin’s lymphoma, leukemia, acute lymphocytic leukemia (ALL) , chronic lymphocytic leukemia (CLL) , chronic myeloid leukemia (CML) , chronic myelomonocytic leukemia (CMML) , lymphoma, Mantle cell lymphoma (MCL) , marginal zone lymphoma (MZL) , myeloma, Non-Hodgkin lymphoma (NHL) , plasmacytoma, or Waldenstrom macroglobulinemia.178.The method of claim 176, wherein the mammal is a human.179.A method of combination therapy in human comprising administering a therapeutically effective amount of a multispecific molecule, a conjugate, a nucleic acid, an expression vector, a modified cell, and / or a pharmaceutical composition ofpreceding claims, and a therapeutically effective amount of another therapy.180.The method of combination therapy of claim 179, wherein another therapy comprises administration of one or more of chemotherapy, molecular targeted therapy, biologics therapy, immunotherapy and radiotherapy.
Citation Information
Patent Citations
Anti-PD-L1 / VEGF bifunctional antibody and application thereof
CN109942712A
Anti-VEGF-anti-PD1 bispecific antibody with brand-new sequence
CN110563849A
Humanized CD47 antibody or antigen binding fragment thereof and application
CN112679611A
Anti-PD-1-anti-VEGFA bispecific antibodies, pharmaceutical compositions and uses thereof
CN112830972A
CD47 / 4-1BB binding protein and medical application thereof
CN118108853A