Multispecific antibodies

Abstract

The present invention relates to a multispecific antibody comprising two antibody-based binding domains (MSLN-BDs) that specifically bind to mesothelin; and at least one antibody-based binding domain (CD3-BD) that specifically binds to CD3; the multispecific antibody does not comprise an immunoglobulin Fc region polypeptide, and each of the MSLN-BDs has a monovalent dissociation constant (K) in the range of 0.5 to 20 nM as measured by SPR. D ) which binds to mesothelin (MSLN). The present invention further relates to nucleic acid sequences encoding said multispecific antibodies, vectors comprising said nucleic acid sequences, host cells comprising said nucleic acid sequences or said vectors, and methods for producing said multispecific antibodies. In addition, the present invention relates to pharmaceutical compositions comprising said multispecific antibodies and methods of using them.

Description

[Technical Field]

[0001] The present invention relates to a multispecific antibody comprising two antibody-based binding domains (MSLN-BDs) that specifically bind to mesothelin; and at least one antibody-based binding domain (CD3-BD) that specifically binds to CD3; the multispecific antibody does not comprise an immunoglobulin Fc region polypeptide, and each of the MSLN-BDs has a monovalent dissociation constant (K) in the range of 0.5 to 20 nM as measured by SPR. D ) which binds to mesothelin (MSLN). The present invention further relates to nucleic acid sequences encoding said multispecific antibodies, vectors comprising said nucleic acid sequences, host cells comprising said nucleic acid sequences or said vectors, and methods for producing said multispecific antibodies. In addition, the present invention relates to pharmaceutical compositions comprising said multispecific antibodies and methods of using them. [Background technology]

[0002] Background of the Invention Despite significant advances in cancer treatment, cancer remains a major unmet medical need. Some of the most substantial advances in cancer treatment in recent years have been driven by the emergence of various classes of molecular immunotherapies, non-limiting examples of which include monoclonal antibodies (mAbs), bispecific antibodies (bsAbs), recombinant proteins, and chimeric antigen receptor-T cell (CAR-T cell) therapy. These therapies induce antitumor immunity by a) actively targeting immune effector cells to tumor-resident cells, b) stimulating immune effector cells, and / or c) alleviating tumor-mediated immunosuppression. These immunotherapies generally exploit the overexpression of specific antigens by tumor-resident cells (e.g., malignant cells, cells of the tumor vasculature, stromal cells, immune cells, etc.) to direct their pharmacological activity toward the tumor rather than extratumoral loci. Among these antigens, tumor-associated antigens (TAAs) comprise cell surface proteins selectively overexpressed by malignant cells. By binding with high affinity to TAAs, immunotherapies can restrict their immunomodulatory activity to some extent to the immune synapse between tumor cells and immune effector cells.

[0003] One common class of TAA-conjugated immunotherapies are mAbs, which induce antitumor immunity by opsonizing tumor cells and triggering antibody-dependent cellular cytotoxicity (ADCC) by cells expressing Fcγ receptors (FcγR), primarily natural killer (NK) cells. Other TAA-conjugated immunotherapies utilize cytotoxic T lymphocytes (CTLs) to induce targeted depletion of malignant cells (e.g., CAR-T cells), as well as bsAbs that simultaneously engage the T cell antigen CD3 (TAA / CD3 bsAb).

[0004] Although the therapeutic utility of TAA-(re)directed CTLs and conventional TAA / CD3 bsAbs has been validated in the clinic, dose-limiting toxicities (DLTs) often prevent administration at the maximally effective dose (MED) or lead to treatment interruption, thereby limiting efficacy.

[0005] One reason for DLT is that conventional TAA / CD3 bsAbs are commonly associated with cytokine release syndrome (CRS), presumably due to excessive activity of the anti-CD3 domain. The extratumoral activity of immunotherapy can result in the secretion of pro-inflammatory cytokines in healthy tissues, potentially resulting in an undesirable safety profile. Furthermore, while TAA / CD3 bsAbs potently deplete TAA-overexpressing cells, they do so by recruiting and stimulating CTLs regardless of whether such cells express tumor-antigen-recognizing T cell receptors (TCRs) (i.e., tumor-reactive T cells). Therefore, TAA / CD3 bsAbs pose a potential safety risk due to their somewhat indiscriminate stimulation of CTLs rather than stimulating or reactivating the host's natural antitumor immunity.

[0006] Although the exact pathways by which such DLTs arise may vary, the risk of immunotherapy-related toxicity can be minimized or eliminated by enhancing tumor localization of pharmacological activity.

[0007] TAAs (e.g., oncofetal antigens) that are expressed almost exclusively on the surface of cancer cells are called clean TAAs. TAAs that are also expressed on the surface of normal, non-cancerous cells—typically at lower levels compared to cancer cells—are considered unclean TAAs. Due to the enormous potency of the TAA / CD3 bsAb approach, unclean TAAs pose a challenge because they can damage non-tumor cells that also express them. Mesothelin (MSLN) is an example of an unclean TAA; it is expressed not only on the surface of tumor cells but also, at lower levels, in various other tissues. Therefore, when targeting unclean TAAs, new therapies are needed to improve the selectivity of TAA / CD3 bsAb approaches toward tumor tissue and minimize off-tumor / on-target effects. This is particularly true for the MSLN / CD3 bsAb approach.

[0008] MSLN +Mesothelin (MSLN) has been proposed as a targetable tumor-associated antigen (TAA) for the treatment of solid tumors (e.g., mesothelioma). Many other types of cancer also show a similar pattern of MSLN. + These include certain forms of ovarian and pancreatic cancer, as well as triple-negative breast cancer. Current standard treatments for mesothelioma include tumor resection, chemotherapy, and radiation therapy, as well as palliative measures (e.g., fluid removal and pain management). Immunotherapies for growing tumors include the use of PD-1 / PD-L1 blockers (e.g., pembrolizumab and nivolumab) or anti-CTLA4 antibodies (e.g., ipilimumab) to stimulate the immune system, and VEGF inhibitors (e.g., bevacizumab) to block angiogenesis. While these therapies have achieved clinical success, they are associated with a greater risk of systemic side effects. Therefore, there is a need for specific approaches targeting MSLNs.

[0009] Numerous preclinical and early-phase clinical studies have been or are being conducted to evaluate the feasibility of targeting MSLN with several different approaches, including antibody-based drugs and CAR-T cells. These include the anti-MSLN fragment SS1P immunotoxin, used with chemotherapeutic agents (e.g., pemetrexed or cisplatin) or coupled to PE38; the chimeric monoclonal antibody amatuximab, which induces ADCC; and the antibody-drug conjugates anetumab-ravtansine (BAY 94-9343: anti-MSLN plus tubulin inhibitor DM4) or DMOT4039A (anti-MSLN plus antimitotic monomethylauristatin E), which target MSLN-expressing cells. HPN536, a multispecific engager (anti-MSLN + anti-CD3 + anti-albumin) with improved half-life, redirects T cells to kill MSLN-expressing targets in vitro and in vivo and appears to be well tolerated by cynomolgus monkeys. Several anti-MSLN chimeric antigen receptor (CAR)-T therapies have also been well tolerated, including transiently transfected CAR-T with mRNA (RNA CARTmeso) and CAR-T with an engineered suicide gene (iCasp9m28z). Responses to most anti-MSLN therapies to date have been modest, highlighting the challenges inherent in treating solid tumors.

[0010] For anti-MSLN therapy, this problem is further complicated by the fact that MSLN is released into the serum of cancer patients (where it is called soluble mesothelin-related protein (SMRP)). High-affinity antibodies against MSLN also bind strongly to SMRP, significantly reducing its activity and the effective dose against cancer cells.

[0011] Therefore, there is a need for novel molecules that can effectively localize to tumors and stimulate T cell responses in the presence of SMRPs.

[0012] Multispecific antibodies with at least three binding domains—two of which (MSLN-BD) specifically bind mesothelin and one of which (CD3-BD) specifically bind CD3—with well-balanced affinities between the two MSLN-BDs, could theoretically address many of the above-mentioned limitations in safety and efficacy. Such multispecific antibodies could theoretically induce high tumor localization and improved selectivity, potentially providing safer and more effective therapies for a variety of cancers. Furthermore, such molecules would further limit the need for co-administration of additional immunotherapies to boost patient responses, thereby facilitating development and minimizing treatment costs. However, the realization of therapeutic multispecific antibodies has been complicated by challenges related to their molecular architecture, the characteristics of their component antigen-binding domains, their manufacturability, and / or poor biophysical properties. Taken together, there remains a clear need for novel multispecific antibodies that have increased localization to tumor cells, induce effective T cell activation with a tolerable toxicity profile, and possess biophysical properties that make them suitable for pharmaceutical development.

[0013] In addition, despite the fact that numerous antibodies specific for MSLN and / or CD3 already exist, the complex and specialized requirements for such multispecific antibodies necessitate the development of novel antibody domains with tailored properties.

[0014] Thus, despite numerous treatment options available for patients suffering from cancer, there remains a need for effective and safe therapeutic agents and their preferred targeted use. While immunomodulatory biologics offer a promising approach in the treatment of cancer due to their mode of action, the lack of global immune stimulation and any restriction of this immune modulation to pathologically relevant cells and sites can result in numerous side effects and significant toxicity, potentially leading to increased patient morbidity and mortality. Therefore, it is an object of the present invention to provide drugs that improve the treatment of proliferative diseases, particularly cancer. Summary of the Invention

[0015] SUMMARY OF THE INVENTION It is an object of the present invention to provide drugs that improve the treatment of proliferative diseases, particularly cancer. In particular, it was an object of the present invention to provide drugs with increased on-target efficacy and thereby an improved toxicological profile.

[0016] In a first aspect, the present invention provides a method for producing a medicament for the treatment of a pulmonary arthritis, comprising: a) Two antibody-based binding domains (MSLN-BD) that specifically bind to mesothelin; b) comprises at least one antibody-based binding domain (CD3-BD) that specifically binds to CD3; Each of the MSLN-BDs does not contain an immunoglobulin Fc region polypeptide and has a monovalent dissociation constant (K) in the range of 0.5 to 20 nM as measured by SPR. D ) relates to a multispecific antibody that binds to mesothelin (MSLN).

[0017] In a second aspect, the present invention relates to specific MSLN-binding domains.

[0018] In a third aspect, the present invention relates to one or two nucleic acid sequences encoding a multispecific antibody or a specific MSLN-binding domain of the invention.

[0019] In a fourth aspect, the present invention relates to a vector or two vectors comprising said nucleic acid sequence or said two nucleic acid sequences of the invention.

[0020] In a fifth aspect, the present invention relates to a host cell or host cells comprising said vector or said two vectors of the invention.

[0021] In a sixth aspect, the present invention relates to a method for producing a multispecific antibody or specific binding domain of the invention, the method comprising (i) providing a nucleic acid sequence or two nucleic acid sequences of the invention, or a vector or two vectors of the invention, expressing said nucleic acid sequence or two nucleic acid sequences, or said vector or two vectors and recovering said multispecific antibody from the expression system, or (ii) providing a host cell or multiple host cells of the invention and culturing said host cell or multiple host cells; and recovering said multispecific antibody from the cell culture.

[0022] In a seventh aspect, the present invention relates to a pharmaceutical composition comprising a multispecific antibody of the invention and a pharmaceutically acceptable carrier.

[0023] In an eighth aspect, the present invention relates to a multispecific antibody comprising three or four binding domains for use in the treatment of a disease, in particular a human disease, more particularly a human disease selected from cancer, an inflammatory disease, and an autoimmune disease.

[0024] In a ninth aspect, the present invention relates to a multispecific antibody that is a heterodimeric protein comprising three or four binding domains for use in the treatment of a disease, in particular a human disease, more particularly a human disease selected from cancer, an inflammatory disease, and an autoimmune disease.

[0025] In a tenth aspect, the present invention relates to a method for treating a disease, in particular a human disease, more particularly a human disease selected from cancer, inflammatory diseases, and autoimmune diseases, comprising the step of administering an above-mentioned single chain multispecific antibody of the invention comprising three or four binding domains.

[0026] In an eleventh aspect, the present invention relates to a method for treating a disease, in particular a human disease, more particularly a human disease selected from cancer, inflammatory diseases, and autoimmune diseases, the method comprising administering an above-mentioned heterodimeric multispecific antibody of the invention comprising three or four binding domains.

[0027] The aspects, advantageous features and preferred embodiments of the present invention summarized in the following items contribute, either alone or in combination, to solving the object of the present invention: 1.a) Two antibody-based binding domains (MSLN-BDs) that specifically bind to mesothelin; b) comprises at least two antibody-based binding domains (CD3-BDs) that specifically bind to CD3; Each of the MSLN-BDs does not contain an immunoglobulin Fc region polypeptide and has a monovalent dissociation constant (K) in the range of 0.5 to 20 nM as measured by SPR. D ) a multispecific antibody that binds to mesothelin (MSLN). 2. EC for killing of target cells with MSLN expression levels 6-8 times higher than MeT-5A cells (ATCC CRL-9444) as measured by flow cytometry 50 is not increased by more than 25-fold in the presence of at least 200 ng / ml, particularly at least 300 ng / ml, particularly at least 400 ng / ml, particularly at least 500 ng / ml of soluble mesothelin, as determined in a T cell-driven cytotoxicity assay against said target cells. 3. Capable of killing target cells with a 6-8 fold higher MSLN expression level than MeT-5A cells (ATCC CRL-9444) as measured by flow cytometry, and an EC 300 / 400 / 500 for killing MeT-5A cells as measured by a T cell-driven cytotoxicity assay against the target cells and the MeT-5A cells. 50 EC at least 10 times, particularly at least 20 times, particularly at least 25 times smaller than 50 Item 1. A multispecific antibody having the formula: 4. Each of the MSLN-BDs has a monovalent dissociation constant (K) in the range of 0.5 to 15 nM, particularly in the range of 0.6 to 10 nM, and particularly in the range of 0.7 to 5 nM, as measured by SPR. D Item 4. The multispecific antibody according to any one of Items 1 to 3, which binds to mesothelin (MSLN) at the site of . 5.a) Two antibody-based binding domains (MSLN-BD) that specifically bind to mesothelin; b) comprises at least one antibody-based binding domain (CD3-BD) that specifically binds to CD3; Each of the MSLN-BDs does not contain an immunoglobulin Fc region polypeptide and has a monovalent dissociation constant (K) in the range of 0.1 to 5 nM as measured by SPR. D ) a multispecific antibody that binds to mesothelin (MSLN). 6. EC for killing of target cells with 6-8 fold higher MSLN expression levels than MeT-5A cells (ATCC CRL-9444) as measured by flow cytometry 50 is not increased more than 50-fold in the presence of at least 200 ng / ml, particularly at least 300 ng / ml, particularly at least 400 ng / ml, particularly at least 500 ng / ml of soluble mesothelin, as determined in a T cell-driven cytotoxicity assay against said target cells. 7. Each of the MSLN-BDs has a monovalent dissociation constant (K) in the range of 0.1 to 3 nM, particularly in the range of 0.15 to 2 nM, and particularly in the range of 0.2 to 1 nM, as measured by SPR. D 7. The multispecific antibody according to item 5 or 6, which binds to mesothelin (MSLN) at the site of 8. The multispecific antibody according to any one of Items 1 to 7, wherein each of the MSLN-BDs specifically binds to human mesothelin. 9. The multispecific antibody according to any one of items 1 to 8, which binds to region I, region II, and / or region III of MSLN, preferably region I and / or region II of MSLN, particularly region I of MSLN. 10. The multispecific antibody according to any one of items 1 to 9, wherein the two MSLN-BDs bind to the same epitope on MSLN. 11. The MSLN-BD is (i) the HCDR1, HCDR2, and HCDR3 sequences of SEQ ID NOs: 1, 2 (or 10), and 3, respectively, and the LCDR1, LCDR2, and LCDR3 sequences of SEQ ID NOs: 4, 5, and 6, respectively; or the HCDR1, HCDR2, and HCDR3 sequences of SEQ ID NOs: 11, 12, and 13, respectively, and the LCDR1, LCDR2, and LCDR3 sequences of SEQ ID NOs: 14, 15, and 16, respectively; and (ii) the framework sequences FR1 to FR4 of the VH3 or VH4 domain; in particular, the framework sequences FR1 to FR4 of the VH3 domain; and (iii) The multispecific antibody according to any one of items 1 to 4 and 8 to 10, comprising a VL domain comprising a VL framework comprising Vκ frameworks FR1, FR2, and FR3 (particularly FR1 to FR3 of Vκ1 or Vκ3, particularly Vκ1 FR1 to FR3) and a framework FR4 (a Vλ FR4 selected from a Vκ FR4 and a Vλ FR4, particularly a Vλ FR4 comprising an amino acid sequence that is at least 70, 80, or 90 percent identical to any of SEQ ID NOs: 132 to 139, more particularly a Vλ FR4 selected from any of SEQ ID NOs: 132 to 139, particularly a Vλ FR4 according to SEQ ID NO: 132 or 139). 12. The MSLN-BD is a.1) the HCDR1, HCDR2, and HCDR3 sequences of SEQ ID NOs: 1, 2 (or 10), and 3, respectively; b.1) the LCDR1, LCDR2, and LCDR3 sequences of SEQ ID NOs: 4, 5, and 6, respectively; c.1) a VH sequence that is at least 80, 85, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99, or 100 percent identical to the amino acid sequence of SEQ ID NO: 7, and d.1) comprises a VL sequence that is at least 80, 85, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99, or 100 percent identical to the amino acid sequence of SEQ ID NO: 9; or a.2) the HCDR1, HCDR2, and HCDR3 sequences of SEQ ID NOs: 1, 2 (or 10), and 3, respectively; b.2) the LCDR1, LCDR2, and LCDR3 sequences of SEQ ID NOs: 4, 5, and 6, respectively; c.2) a VH sequence that is at least 80, 85, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99, or 100 percent identical to the amino acid sequence of SEQ ID NO:8, and d.2) comprises a VL sequence that is at least 80, 85, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99, or 100 percent identical to the amino acid sequence of SEQ ID NO:9; or a.3) the HCDR1, HCDR2, and HCDR3 sequences of SEQ ID NOs: 11, 12, and 13, respectively; b.3) LCDR1, LCDR2, and LCDR3 sequences of SEQ ID NOs: 14, 15, and 16, respectively; c.3) a VH sequence that is at least 80, 85, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99, or 100 percent identical to the amino acid sequence of SEQ ID NO: 17; and d.3) comprises a VL sequence that is at least 80, 85, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99, or 100 percent identical to the amino acid sequence of SEQ ID NO: 18; or a.4) the HCDR1, HCDR2, and HCDR3 sequences of SEQ ID NOs: 11, 12, and 13, respectively; b.4) LCDR1, LCDR2, and LCDR3 sequences of SEQ ID NOs: 14, 15, and 16, respectively; c.4) a VH sequence that is at least 80, 85, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99, or 100 percent identical to the amino acid sequence of SEQ ID NO: 19; and d.4) comprises a VL sequence that is at least 80, 85, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99, or 100 percent identical to the amino acid sequence of SEQ ID NO: 21; or a.5) the HCDR1, HCDR2, and HCDR3 sequences of SEQ ID NOs: 11, 12, and 13, respectively; b.5) ​​LCDR1, LCDR2, and LCDR3 sequences of SEQ ID NOs: 14, 15, and 16, respectively; c.5) a VH sequence that is at least 80, 85, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99, or 100 percent identical to the amino acid sequence of SEQ ID NO: 20; and d.5) comprises a VL sequence that is at least 80, 85, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99, or 100 percent identical to the amino acid sequence of SEQ ID NO: 21; or a.6) the HCDR1, HCDR2, and HCDR3 sequences of SEQ ID NOs: 11, 12, and 13, respectively; b.6) LCDR1, LCDR2, and LCDR3 sequences of SEQ ID NOs: 14, 15, and 16, respectively; c.6) a VH sequence that is at least 80, 85, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99, or 100 percent identical to the amino acid sequence of SEQ ID NO: 22; and d.6) comprises a VL sequence that is at least 80, 85, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99, or 100 percent identical to the amino acid sequence of SEQ ID NO: 24; or a.7) the HCDR1, HCDR2, and HCDR3 sequences of SEQ ID NOs: 11, 12, and 13, respectively; b.7) LCDR1, LCDR2, and LCDR3 sequences of SEQ ID NOs: 14, 15, and 16, respectively; c.7) a VH sequence that is at least 80, 85, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99, or 100 percent identical to the amino acid sequence of SEQ ID NO: 23, and d.7) The multispecific antibody of any one of clauses 1 to 4 and 8 to 11, comprising a VL sequence that is at least 80, 85, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99, or 100 percent identical to the amino acid sequence of SEQ ID NO: 24. 13. The MSLN-BD is (i) the HCDR1, HCDR2, and HCDR3 sequences of SEQ ID NOs: 25, 26, and 27, respectively, and the LCDR1, LCDR2, and LCDR3 sequences of SEQ ID NOs: 28, 29, and 30, respectively; and (ii) the framework sequences FR1 to FR4 of the VH3 or VH4 domain; in particular, the framework sequences FR1 to FR4 of the VH3 domain; and (iii) The multispecific antibody according to any one of Items 5 to 10, comprising a VL domain comprising a VL framework comprising Vκ frameworks FR1, FR2, and FR3 (particularly FR1 to FR3 of Vκ1 or Vκ3, particularly Vκ1 FR1 to FR3) and a framework FR4 (a Vλ FR4 selected from a Vκ FR4 and a Vλ FR4, particularly a Vλ FR4 comprising an amino acid sequence that is at least 70, 80, or 90 percent identical to any one of SEQ ID NOs: 132 to 139, more particularly a Vλ FR4 selected from any one of SEQ ID NOs: 132 to 139, particularly a Vλ FR4 according to SEQ ID NO: 132 or 139). 14. The MSLN-BD is a.1) the HCDR1, HCDR2, and HCDR3 sequences of SEQ ID NOs: 25, 26, and 27, respectively; b.1) the LCDR1, LCDR2, and LCDR3 sequences of SEQ ID NOs: 28, 29, and 30, respectively; c.1) a VH sequence that is at least 80, 85, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99, or 100 percent identical to the amino acid sequence of SEQ ID NO: 31; and d.1) comprises a VL sequence that is at least 80, 85, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99, or 100 percent identical to the amino acid sequence of SEQ ID NO: 33; or a.2) the HCDR1, HCDR2, and HCDR3 sequences of SEQ ID NOs: 25, 26, and 27, respectively; b.2) LCDR1, LCDR2, and LCDR3 sequences of SEQ ID NOs: 28, 29, and 30, respectively; c.2) a VH sequence that is at least 80, 85, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99, or 100 percent identical to the amino acid sequence of SEQ ID NO: 32; and d.2) comprises a VL sequence that is at least 80, 85, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99, or 100 percent identical to the amino acid sequence of SEQ ID NO: 33; or a.3) the HCDR1, HCDR2, and HCDR3 sequences of SEQ ID NOs: 25, 26, and 27, respectively; b.3) LCDR1, LCDR2, and LCDR3 sequences of SEQ ID NOs: 28, 29, and 30, respectively; c.3) a VH sequence that is at least 80, 85, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99, or 100 percent identical to the amino acid sequence of SEQ ID NO: 34, and d.3) comprises a VL sequence that is at least 80, 85, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99, or 100 percent identical to the amino acid sequence of SEQ ID NO: 36; or a.4) the HCDR1, HCDR2, and HCDR3 sequences of SEQ ID NOs: 25, 26, and 27, respectively; b.4) LCDR1, LCDR2, and LCDR3 sequences of SEQ ID NOs: 28, 29, and 30, respectively; c.4) a VH sequence that is at least 80, 85, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99, or 100 percent identical to the amino acid sequence of SEQ ID NO: 35, and d.4) The multispecific antibody of any one of clauses 5 to 10 and 13, comprising a VL sequence that is at least 80, 85, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99, or 100 percent identical to the amino acid sequence of SEQ ID NO: 36. 15. The multispecific antibody of any one of items 11 to 14, wherein the VH domain comprises the C51 amino acid residue (AHo numbering) and the VL domain comprises the C141 amino acid residue (AHo numbering). 16. Each of the MSLN-BDs cross-reacts with Macaca fascicularis (cynomolgus monkey) MSLN, particularly with cynomolgus monkey MSLN, and has a monovalent K in the range of 0.2 to 75 nM, particularly in the range of 0.3 to 60 nM, particularly in the range of 0.4 to 50 nM, particularly in the range of 0.5 to 40 nM, as measured by SPR. D Item 16. The multispecific antibody according to any one of Items 8 to 15, which binds to the target antigen via the following structure: 17. The multispecific antibody according to any one of items 1 to 16, wherein the CD3-BD is bound to CD3ε. 18. The multispecific antibody according to any one of clauses 1 to 17, which comprises one or two CD3-BDs, in particular one CD3-BD. 19. The CD3-BD binds to CD3ε with a monovalent K of less than 50 nM as measured by SPR. D , especially monovalent K of 0.5 to 50 nM D , especially 1 to 40 nM, especially 2 to 30 nM monovalent K D Item 19. The multispecific antibody according to any one of Items 1 to 18, which binds to the target antigen via the following structure: 20. The CD3-BD is (i) the HCDR1, HCDR2, and HCDR3 sequences of SEQ ID NOs: 45, 46, and 47, respectively, in a human antibody VH framework (particularly a VH3 framework); (ii) The multispecific antibody according to any one of Items 1 to 19, comprising the LCDR1 sequence, LCDR2 sequence, and LCDR3 sequence of SEQ ID NOs: 48, 49, and 50, respectively, in a human antibody VL framework, wherein the VL framework comprises Vκ framework FR1, FR2, and FR3 (particularly a Vκ1 framework) and framework FR4 (selected from a Vκ FR4 and a Vλ framework 4). 21. The CD3-BD is (i) a VH domain comprising the amino acid sequence of SEQ ID NO: 51, 140, 141, or 142; and (ii) The multispecific antibody of clause 20, comprising a VL domain comprising the amino acid sequence of SEQ ID NO: 52. 22. The multispecific antibody of any one of clauses 1 to 21, further comprising at least one human serum albumin binding domain (hSA-BD), in particular one hSA-BD. 23. The hSA-BD, (i) the HCDR1, HCDR2, and HCDR3 sequences of SEQ ID NOs: 53, 54, and 55, respectively, in a human antibody VH framework (particularly a VH3 framework); (ii) a human antibody VL framework comprising the LCDR1 sequence, LCDR2 sequence, and LCDR3 sequence of SEQ ID NOs: 56, 57, and 58, respectively, wherein the VL framework comprises Vκ framework FR1, FR2, and FR3 (particularly a Vκ1 framework) and framework FR4 (selected from Vκ FR4 (particularly Vκ1 FR4) and Vλ framework 4); or (i) the HCDR1, HCDR2, and HCDR3 sequences of SEQ ID NOs: 63, 64, and 65, respectively, in a human antibody VH framework (particularly a VH3 framework); (ii) a human antibody VL framework comprising the LCDR1 sequence, LCDR2 sequence, and LCDR3 sequence of SEQ ID NOs: 66, 67, and 68, respectively, wherein the VL framework comprises Vκ framework FR1, FR2, and FR3 (particularly a Vκ1 framework) and framework FR4 (selected from Vκ FR4 (particularly Vκ1 FR4) and Vλ framework 4); or (i) the HCDR1, HCDR2, and HCDR3 sequences of SEQ ID NOs: 63, 64, and 65, respectively, in a human antibody VH framework (particularly a VH3 framework); (ii) The multispecific antibody of clause 22, comprising the LCDR1 sequence, LCDR2 sequence, and LCDR3 sequence of SEQ ID NOs: 66, 67, and 68, respectively, in a human antibody VL framework, wherein the VL framework comprises Vκ framework FR1, FR2, and FR3 (particularly a Vκ1 framework) and framework FR4 (selected from Vκ FR4 (particularly Vκ1 FR4) and Vλ framework 4). 24. The hSA-BD is (i) a VH domain comprising the amino acid sequence of SEQ ID NO: 59 and a VL domain comprising the amino acid sequence of SEQ ID NO: 60; or (ii) a VH domain comprising the amino acid sequence of SEQ ID NO: 61 and a VL domain comprising the amino acid sequence of SEQ ID NO: 62; or (iii) a VH domain comprising the amino acid sequence of SEQ ID NO: 69 and a VL domain comprising the amino acid sequence of SEQ ID NO: 70; or (iv) a VH domain comprising the amino acid sequence of SEQ ID NO: 71 and a VL domain comprising the amino acid sequence of SEQ ID NO: 72; or (v) a VH domain comprising the amino acid sequence of SEQ ID NO: 79 or 143 and a VL domain comprising the amino acid sequence of SEQ ID NO: 80 or 144; or (vi) The multispecific antibody of clause 22 or 23, comprising a VH domain comprising the amino acid sequence of SEQ ID NO: 81 or 145 and a VL domain comprising the amino acid sequence of SEQ ID NO: 82 or 146. 25. The multispecific antibody of paragraph 24, wherein the VH domain comprises the C51 amino acid residue (AHo numbering) and the VL domain comprises the C141 amino acid residue (AHo numbering). 26. The multispecific antibody according to any one of items 1 to 25, wherein the binding domains are independently selected from the group consisting of Fab, Fv, scFv, dsFv, scAb, and STAB. 27. Each of the binding domains comprises: (a) a cognate pair of VL and VH domains (Fv fragment); or (b) a cognate pair of VL and VH domains (scFv fragments) linked by an oligopeptide or polypeptide linker 27. The multispecific antibody of clause 26, wherein the antibody is independently selected from: 28. The multispecific antibody according to any one of Items 1 to 27, wherein the multispecific antibody is in a format selected from the group consisting of a tandem scDb (Tandab), a linear dimeric scDb (LD-scDb), a cyclic dimeric scDb (CD-scDb), a tandem tri-scFv, a tribody (Fab-(scFv)2), Fab-Fv2, a triabody, a scDb-scFv, a tetrabody, a di-diabody, CODV, a tandem di-scFv, a tandem tri-scFv, Fab-(scFv)2, Fab-Fv2, or a CODV fused to the N-terminus and / or C-terminus of a heterodimerization domain other than a heterodimeric Fc domain, and MATCH. 29. The multispecific antibody according to any one of items 1 to 28, which does not contain a CH1 region and / or a CL region. 30. The multispecific antibody according to any one of items 1 to 29, which is an scDb-scFv, a tribody, or a MATCH, in particular in the MATCH format or tribody format, more particularly in the MATCH format, and more particularly in MATCH3 or MATCH4. 31. The multispecific antibody according to any one of items 1 to 30, which is a single-chain protein. 32. The single-chain proteins are arranged in the following order: (i) a first VL domain; (ii) a first polypeptide linker; (iii) a first VH domain; (iv) a second polypeptide linker; (v) a second VL domain; (vi) a third polypeptide linker, and (vii) a second VH domain wherein said first VL domain associates with said second VH domain to form a first binding domain, and said second VL domain associates with said first VH domain to form a second binding domain; The single-chain protein (viii) a third binding domain formed by a third VL domain and a third VH domain connected through a fourth polypeptide linker, and fused at its C-terminus or N-terminus to the amino acid sequence through a fifth polypeptide linker; further comprising The three binding domains have the following specificities: a) the first binding domain specifically binds to human CD3 (CD3-BD); b) The multispecific antibody of paragraph 31, wherein the second and third binding domains specifically bind to mesothelin (MSLN-BD). 33. The multispecific antibody of paragraph 32, wherein the single-chain protein is formed by a fourth VL domain and a fourth VH domain connected by a sixth polypeptide linker, and further comprises an hSA-BD fused at its C-terminus or N-terminus to the amino acid sequence through a seventh polypeptide linker. 34. A heterodimeric protein comprising a first and a second single-chain protein; The first single-chain protein comprises (from the N-terminus to the C-terminus): (ia) the first VL domain; (iia) a first polypeptide linker, and (iiia) second VL domain a first amino acid sequence consisting of The second single-chain protein comprises (from the N-terminus to the C-terminus): (ib) the first VH domain; (iib) a second polypeptide linker, and (iiib) a second VH domain a second amino acid sequence consisting of said first VL domain associates with either said first or said second VH domain to form a first binding domain, and said second VL domain associates with the other of said VH domains to form a second binding domain; At least one of the first and second single-chain proteins is (iv) a third binding domain formed by a third VL domain and a third VH domain connected via a third polypeptide linker, and fused to the first or second amino acid sequence via a fourth polypeptide linker; Optionally, at least one of the first and second single-chain proteins comprises: (v) a fourth binding domain formed by a fourth VL domain and a fourth VH domain connected via a fifth polypeptide linker, the fourth binding domain being fused to the first or second amino acid sequence via a sixth polypeptide linker; The three, and optionally four, binding domains have the following specificities: a) One antibody-based binding domain (CD3-BD) specifically binds to human CD3; b) two additional antibody-based binding domains (MSLN-BD) specifically bind mesothelin, and when the optional fourth binding domain is present: c) The multispecific antibody according to any one of items 1 to 30, wherein the remaining binding domain (hSA-BD) specifically binds to human serum albumin. 35. The multispecific antibody of paragraph 34, wherein the optional fourth domain is present, one of the first and second binding domains is CD3-BD, and the other of the first and second binding domains is hSA-BD. 36. The multispecific antibody of paragraph 34 or 35, wherein the optional fourth domain is present, and wherein the third domain is fused to either the first or second amino acid sequence, and the fourth domain is fused to the other of the two amino acid sequences. 37. The multispecific antibody of any one of items 34 to 36, wherein any binding domain contained in the heterodimeric protein consists solely of immunoglobulin variable domains located in the first and second single-chain proteins. 38. The multispecific antibody of any one of items 34 to 37, wherein the heterodimeric protein does not comprise a cognate pair of first and second protein interaction domains other than the first and second VL and VH domains, the first protein interaction domain is comprised in the first single-chain protein, and the second protein interaction domain is comprised in the second single-chain protein. 39. The multispecific antibody of any one of clauses 34 to 38, wherein the first single-chain protein and the second single-chain protein heterodimerize in a parallel orientation, i.e., the first VL domain associates with the first VH domain and the second VL domain associates with the second VH domain. 40. The multispecific antibody of any one of clauses 34 to 38, wherein the first single-chain protein and the second single-chain protein heterodimerize in an antiparallel orientation, i.e., the first VL domain associates with the second VH domain, and the second VL domain associates with the first VH domain. 41. The first single-chain protein comprises an amino acid sequence that is at least 90, 95, 96, 97, 98, or 99 percent identical to the amino acid sequence of SEQ ID NO: 83, more particularly comprises the amino acid sequence of SEQ ID NO: 83, and in particular consists of the amino acid sequence of SEQ ID NO: 83; and the second single-chain protein comprises an amino acid sequence that is at least 90, 95, 96, 97, 98, or 99 percent identical to the amino acid sequence of SEQ ID NO: 84, more particularly comprises, and in particular consists of, the amino acid sequence of SEQ ID NO: 84; or the first single-chain protein comprises an amino acid sequence that is at least 90, 95, 96, 97, 98, or 99 percent identical to the amino acid sequence of SEQ ID NO: 85, more particularly comprises the amino acid sequence of SEQ ID NO: 85, and in particular consists of the amino acid sequence of SEQ ID NO: 85; and the second single-chain protein comprises an amino acid sequence that is at least 90, 95, 96, 97, 98, or 99 percent identical to the amino acid sequence of SEQ ID NO: 86, more particularly comprises, and in particular consists of, the amino acid sequence of SEQ ID NO: 86; or the first single-chain protein comprises an amino acid sequence that is at least 90, 95, 96, 97, 98, or 99 percent identical to the amino acid sequence of SEQ ID NO: 87, more particularly comprises the amino acid sequence of SEQ ID NO: 87, and in particular consists of the amino acid sequence of SEQ ID NO: 87; and the second single-chain protein comprises an amino acid sequence that is at least 90, 95, 96, 97, 98, or 99 percent identical to the amino acid sequence of SEQ ID NO: 88, more particularly comprises, and in particular consists of, the amino acid sequence of SEQ ID NO: 88; or the first single-chain protein comprises an amino acid sequence that is at least 90, 95, 96, 97, 98, or 99 percent identical to the amino acid sequence of SEQ ID NO: 89, more particularly comprises the amino acid sequence of SEQ ID NO: 89, and in particular consists of the amino acid sequence of SEQ ID NO: 89; and the second single-chain protein comprises an amino acid sequence that is at least 90, 95, 96, 97, 98, or 99 percent identical to the amino acid sequence of SEQ ID NO: 90, more particularly comprises the amino acid sequence of SEQ ID NO: 90, and in particular consists of the amino acid sequence of SEQ ID NO: 90; or the first single-chain protein comprises an amino acid sequence that is at least 90, 95, 96, 97, 98, or 99 percent identical to the amino acid sequence of SEQ ID NO: 91, more particularly comprises the amino acid sequence of SEQ ID NO: 91, and in particular consists of the amino acid sequence of SEQ ID NO: 91; and the second single-chain protein comprises an amino acid sequence that is at least 90, 95, 96, 97, 98, or 99 percent identical to the amino acid sequence of SEQ ID NO: 92, more particularly comprises, and in particular consists of, the amino acid sequence of SEQ ID NO: 92; or the first single-chain protein comprises an amino acid sequence that is at least 90, 95, 96, 97, 98, or 99 percent identical to the amino acid sequence of SEQ ID NO: 93, more particularly comprises the amino acid sequence of SEQ ID NO: 93, and in particular consists of the amino acid sequence of SEQ ID NO: 93; and the second single-chain protein comprises an amino acid sequence that is at least 90, 95, 96, 97, 98, or 99 percent identical to the amino acid sequence of SEQ ID NO: 94, more particularly comprises, and in particular consists of, the amino acid sequence of SEQ ID NO: 94; or the first single-chain protein comprises an amino acid sequence that is at least 90, 95, 96, 97, 98, or 99 percent identical to the amino acid sequence of SEQ ID NO: 95, more particularly comprises the amino acid sequence of SEQ ID NO: 95, and in particular consists of the amino acid sequence of SEQ ID NO: 95; and the second single-chain protein comprises an amino acid sequence that is at least 90, 95, 96, 97, 98, or 99 percent identical to the amino acid sequence of SEQ ID NO: 96, more particularly comprises, and in particular consists of, the amino acid sequence of SEQ ID NO: 96; or the first single-chain protein comprises an amino acid sequence that is at least 90, 95, 96, 97, 98, or 99 percent identical to the amino acid sequence of SEQ ID NO: 97, more particularly comprises the amino acid sequence of SEQ ID NO: 97, and in particular consists of the amino acid sequence of SEQ ID NO: 97; and the second single-chain protein comprises an amino acid sequence that is at least 90, 95, 96, 97, 98, or 99 percent identical to the amino acid sequence of SEQ ID NO: 98, more particularly comprises, and in particular consists of, the amino acid sequence of SEQ ID NO: 98; or the first single-chain protein comprises an amino acid sequence that is at least 90, 95, 96, 97, 98, or 99 percent identical to the amino acid sequence of SEQ ID NO: 99, more particularly comprises the amino acid sequence of SEQ ID NO: 99, and in particular consists of the amino acid sequence of SEQ ID NO: 99; and the second single-chain protein comprises an amino acid sequence that is at least 90, 95, 96, 97, 98, or 99 percent identical to the amino acid sequence of SEQ ID NO: 100, more particularly comprises, and in particular consists of, the amino acid sequence of SEQ ID NO: 100; or the first single-chain protein comprises an amino acid sequence that is at least 90, 95, 96, 97, 98, or 99 percent identical to the amino acid sequence of SEQ ID NO: 101, more particularly comprises the amino acid sequence of SEQ ID NO: 101, and in particular consists of the amino acid sequence of SEQ ID NO: 101; and the second single-chain protein comprises an amino acid sequence that is at least 90, 95, 96, 97, 98, or 99 percent identical to the amino acid sequence of SEQ ID NO: 102, more particularly comprises, and in particular consists of, the amino acid sequence of SEQ ID NO: 102; or the first single-chain protein comprises an amino acid sequence that is at least 90, 95, 96, 97, 98, or 99 percent identical to the amino acid sequence of SEQ ID NO: 103, more particularly comprises the amino acid sequence of SEQ ID NO: 103, and in particular consists of the amino acid sequence of SEQ ID NO: 103; and the second single-chain protein comprises an amino acid sequence that is at least 90, 95, 96, 97, 98, or 99 percent identical to the amino acid sequence of SEQ ID NO: 104, more particularly comprises, and in particular consists of, the amino acid sequence of SEQ ID NO: 104; or the first single-chain protein comprises an amino acid sequence that is at least 90, 95, 96, 97, 98, or 99 percent identical to the amino acid sequence of SEQ ID NO: 105, more particularly comprises the amino acid sequence of SEQ ID NO: 105, and in particular consists of the amino acid sequence of SEQ ID NO: 105; and the second single-chain protein comprises an amino acid sequence that is at least 90, 95, 96, 97, 98, or 99 percent identical to the amino acid sequence of SEQ ID NO: 106, more particularly comprises, and in particular consists of, the amino acid sequence of SEQ ID NO: 106; or the first single-chain protein comprises an amino acid sequence that is at least 90, 95, 96, 97, 98, or 99 percent identical to the amino acid sequence of SEQ ID NO: 107, more particularly comprises the amino acid sequence of SEQ ID NO: 107, and in particular consists of the amino acid sequence of SEQ ID NO: 107; and the second single-chain protein comprises an amino acid sequence that is at least 90, 95, 96, 97, 98, or 99 percent identical to the amino acid sequence of SEQ ID NO: 108, more particularly comprises, and in particular consists of, the amino acid sequence of SEQ ID NO: 108; or the first single-chain protein comprises an amino acid sequence that is at least 90, 95, 96, 97, 98, or 99 percent identical to the amino acid sequence of SEQ ID NO: 109, more particularly comprises the amino acid sequence of SEQ ID NO: 109, and in particular consists of the amino acid sequence of SEQ ID NO: 109; and the second single-chain protein comprises an amino acid sequence that is at least 90, 95, 96, 97, 98, or 99 percent identical to the amino acid sequence of SEQ ID NO: 110, more particularly comprises the amino acid sequence of SEQ ID NO: 110, and in particular consists of the amino acid sequence of SEQ ID NO: 110; or the first single-chain protein comprises an amino acid sequence that is at least 90, 95, 96, 97, 98, or 99 percent identical to the amino acid sequence of SEQ ID NO: 111, more particularly comprises the amino acid sequence of SEQ ID NO: 111, and in particular consists of the amino acid sequence of SEQ ID NO: 111; and the second single-chain protein comprises an amino acid sequence that is at least 90, 95, 96, 97, 98, or 99 percent identical to the amino acid sequence of SEQ ID NO: 112, more particularly comprises, and in particular consists of, the amino acid sequence of SEQ ID NO: 112; or the first single-chain protein comprises an amino acid sequence that is at least 90, 95, 96, 97, 98, or 99 percent identical to the amino acid sequence of SEQ ID NO: 113, more particularly comprises the amino acid sequence of SEQ ID NO: 113, and in particular consists of the amino acid sequence of SEQ ID NO: 113; and the second single-chain protein comprises an amino acid sequence that is at least 90, 95, 96, 97, 98, or 99 percent identical to the amino acid sequence of SEQ ID NO: 114, more particularly comprises, and in particular consists of, the amino acid sequence of SEQ ID NO: 114; or the first single-chain protein comprises an amino acid sequence that is at least 90, 95, 96, 97, 98, or 99 percent identical to the amino acid sequence of SEQ ID NO: 115, more particularly comprises the amino acid sequence of SEQ ID NO: 115, and in particular consists of the amino acid sequence of SEQ ID NO: 115; and the second single-chain protein comprises an amino acid sequence that is at least 90, 95, 96, 97, 98, or 99 percent identical to the amino acid sequence of SEQ ID NO: 116, more particularly comprises, and in particular consists of, the amino acid sequence of SEQ ID NO: 116; or the first single-chain protein comprises an amino acid sequence that is at least 90, 95, 96, 97, 98, or 99 percent identical to the amino acid sequence of SEQ ID NO: 117, more particularly comprises the amino acid sequence of SEQ ID NO: 117, and in particular consists of the amino acid sequence of SEQ ID NO: 117; and the second single-chain protein comprises an amino acid sequence that is at least 90, 95, 96, 97, 98, or 99 percent identical to the amino acid sequence of SEQ ID NO: 118, more particularly comprises, and in particular consists of, the amino acid sequence of SEQ ID NO: 118; or the first single-chain protein comprises an amino acid sequence that is at least 90, 95, 96, 97, 98, or 99 percent identical to the amino acid sequence of SEQ ID NO: 119, more particularly comprises the amino acid sequence of SEQ ID NO: 119, and in particular consists of the amino acid sequence of SEQ ID NO: 119; and the second single-chain protein comprises an amino acid sequence that is at least 90, 95, 96, 97, 98, or 99 percent identical to the amino acid sequence of SEQ ID NO: 120, more particularly comprises, and in particular consists of, the amino acid sequence of SEQ ID NO: 120; or the first single-chain protein comprises an amino acid sequence that is at least 90, 95, 96, 97, 98, or 99 percent identical to the amino acid sequence of SEQ ID NO: 121, more particularly comprises the amino acid sequence of SEQ ID NO: 121, and in particular consists of the amino acid sequence of SEQ ID NO: 121; and the second single-chain protein comprises an amino acid sequence that is at least 90, 95, 96, 97, 98, or 99 percent identical to the amino acid sequence of SEQ ID NO: 122, more particularly comprises, and in particular consists of, the amino acid sequence of SEQ ID NO: 122; or the first single-chain protein comprises an amino acid sequence that is at least 90, 95, 96, 97, 98, or 99 percent identical to the amino acid sequence of SEQ ID NO: 123, more particularly comprises the amino acid sequence of SEQ ID NO: 123, and in particular consists of the amino acid sequence of SEQ ID NO: 123; and 41. The multispecific antibody according to any one of Items 34 to 40, wherein the second single-chain protein comprises an amino acid sequence that is at least 90, 95, 96, 97, 98, or 99 percent identical to the amino acid sequence of SEQ ID NO: 124, more particularly comprises the amino acid sequence of SEQ ID NO: 124, and in particular consists of the amino acid sequence of SEQ ID NO: 124. 42. The multispecific antibody of any one of items 1 to 41, wherein at least one of the antibody variable domains comprises a CDR region derived from a parent rabbit antibody. 43. The multispecific antibody of any one of clauses 1 to 42, wherein at least one of the MSLN-BDs and the CD3-BD are capable of simultaneously binding to their respective antigens, in particular, wherein both the MSLN-BD and the CD3-BD are capable of binding to their respective antigens. 44. An MSLN-binding domain as defined in any one of paragraphs 11 to 15. 45. One or two nucleic acid sequences encoding the multispecific antibody of any one of items 1 to 43 or the MSLN-binding domain of item 44. 46. ​​A vector or two vectors comprising the nucleic acid sequence or two nucleic acid sequences of paragraph 45. 47. A host cell or host cells comprising one vector or two vectors of paragraph 46. 48. A method for producing a multispecific antibody according to any one of paragraphs 1 to 43 or an MSLN-binding domain according to paragraph 44, comprising: (i) providing a nucleic acid sequence or two nucleic acid sequences according to paragraph 45, or a vector or two vectors according to paragraph 46, expressing the nucleic acid sequence or sequences, or the vector or sequences, and recovering the multispecific antibody or the MSLN-binding domain from the expression system; or (ii) providing a host cell or host cells according to paragraph 47 and culturing the host cell or host cells; and recovering the multispecific antibody or the MSLN-binding domain from the cell culture. 49. A pharmaceutical composition comprising the multispecific antibody according to any one of items 1 to 43 and a pharmaceutically acceptable carrier. 50. The multispecific antibody of any one of clauses 1 to 43 for use in the treatment of a disease, particularly a human disease, more particularly a human disease selected from cancer, inflammatory diseases, and autoimmune diseases. 51. The multispecific antibody of any one of clauses 1 to 43 for use in treating a disease according to clause 50, wherein the disease is cancer, in particular a cancer selected from mesothelioma, pancreatic cancer, and ovarian cancer. 52. A method for treating a disease, particularly a human disease, more particularly a human disease selected from cancer, inflammatory diseases, and autoimmune diseases, comprising the step of administering the multispecific antibody of any one of items 1 to 43. 53. The method according to paragraph 52, wherein the disease is cancer, particularly a cancer selected from mesothelioma, pancreatic cancer, and ovarian cancer. [Brief explanation of the drawings]

[0028] [Figure 1] Figure 1 shows the binding of anti-MSLN scFvs PRO1783, PRO1922, PRO1925, PRO2306, and PRO2309, as well as the reference antibody amatuximab, to the plasma membrane of cells derived from the H226 cell line, which expresses high levels of human MSLN. The binding of PRO1783, PRO1922, PRO1925, PRO2306, PRO2309, and amatuximab to the H226 cell line was examined by competitive ELISA (cELISA). In the cELISA, HRP-coupled protein L and an anti-human IgG antibody coupled to HRP were used to detect PRO1783, PRO1922, PRO1925, PRO2306, PRO2309, and amatuximab bound to H226 cells, respectively. Optical density (OD450nm-690nm) is shown as a function of antibody concentration (in nM). Note that only increasing concentrations were fitted. The EC50 value for PRO1783 is roughly 6-fold greater than that seen for the reference antibody amatuximab. EC50 values ​​for PRO1925, PRO2306, and PRO2309 are in the same range as PRO1783, while the EC50 value for PRO1922 is closer to that of amatuximab.

[0029] [Figure 2]Figure 2 shows the binding of anti-MSLN scFv PRO1783 and the reference antibody amatuximab to the plasma membrane of cells from a CHO cell line expressing cynomolgus monkey MSLN. Binding of PRO1783 and amatuximab to the plasma membrane of a CHO cell line expressing cynomolgus monkey MSLN was examined by cELISA. Binding of PRO1783 and amatuximab was detected with protein L coupled to HRP and an anti-human IgG antibody coupled to HRP, respectively. Optical density (OD450nm-690nm) is shown as a function of antibody concentration (in nM). It was demonstrated that PRO1783, as well as amatuximab, binds to cynomolgus monkey MSLN.

[0030] [Figure 3] Figure 3 shows the inhibition of human MSLN / MUC16 interaction by anti-MSLN scFvs PRO1783, PRO1922, and PRO1925. The potency of PRO1783, PRO1922, PRO1925, and amatuximab in inhibiting human MSLN / MUC16 interaction was examined by cELISA. Optical density (OD450nm-690nm) is shown as a function of antibody concentration (in nM). The IC50 of PRO1783 was 0.5 nM, whereas the reference antibody amatuximab more potently neutralized the MSLN / MUC16 interaction (IC50 = 0.014 nM). PRO1925 has an IC50 similar to that of PRO1783, while the IC50 of PRO1922 is close to that of amatuximab.

[0031] [Figure 4] Figure 4 shows the chimeric variant of hMSLN / mMSLN. The domains highlighted in dark gray are the segments of the human MSLN sequence that were replaced with the corresponding mouse sequence. Segment VI corresponds to the C-terminal part of the MSLN extracellular domain closest to the plasma membrane. Segments I and II correspond to region I of MSLN; segments III and IV correspond to region II of MSLN; and segments V and VI correspond to region III of MSLN.

[0032] [Figure 5]Figure 5 illustrates different MATCH formats. (Left) Antiparallel MATCH4 format architecture, in which the split heterodimer-forming variable domains on each chain are organized in reverse N- to C-terminal order as their cognate variable domains on the complementary MATCH chain. (Right) Representative architecture of the scMATCH3 format, in which the split variable domains are located on a single peptide chain and assembled into a trispecific molecule. Alternative configurations (e.g., VL2-VL1-VH1-VH2-scFv) are also within the scope of the scMATCH3 format. The Gly-Ser linker used to connect the domains is indicated by a line. Table 17 describes the domains contained in the different molecules generated and their location within the molecule (Domains 1-4). As defined in the claims, the numbering of the domains does not correlate with the numbering of the binding domains.

[0033] [Figure 6] Figure 6 shows the cytotoxic activity and effect of PRO2000 and PRO1872 on CD8+ T cell activation in the presence of human serum albumin. (A) Specific killing of cancer cells with high MSLN expression (H226 cells). For cancer cells expressing high levels of mesothelin, PRO2000 exhibited 75-fold greater target cell killing than PRO1872. (B) Specific killing of mesothelial cells with low MSLN expression (MeT-5A cells). For cells derived from healthy mesothelial tissue (MeT-5A; ATCC CRL-9444) expressing low levels of mesothelin, the monovalent mesothelin-binding protein PRO1872 exhibited the highest killing activity. (C) CD8+ T cell activation in the presence of cancer cells with high MSLN expression (H226 cells) and (D) CD8+ T cell activation in the presence of mesothelial cells with low MSLN expression (MeT-5A cells). Similar data were observed for CD8+ T cell activation using PBMCs from donor #1. Target cells and CD8+ T cells were analyzed by flow cytometry 40 hours after incubation with each molecule, and data were fitted using a sigmoidal 4PL fit (GraphPad Prism).

[0034] [Figure 7] Figure 7 shows the cytotoxic activity and effect of PRO2000 and PRO1872 on CD8+ T cell activation in the presence of human serum albumin. (A) Specific killing of cancer cells with low MSLN expression (H292 cells). In cancer cells expressing low levels of mesothelin, the target cell killing observed with PRO1872 is 7-fold greater than that of PRO2000. (B) Specific killing of cancer cells with intermediate MSLN expression (HPAC cells). In cancer cells expressing intermediate levels of mesothelin, both the monovalent and bivalent mesothelin binders PRO1872 and PRO2000 demonstrate similar killing activity. (C) CD8+ T cell activation in the presence of cancer cells with low MSLN expression (H292 cells) and (D) CD8+ T cell activation in the presence of cancer cells with intermediate MSLN expression (HPAC cells). Similar data were observed for CD8+ T cell activation, with the exception that PRO2000 was significantly more potent (4-fold) than PRO1872 in the presence of HPAC cells. PBMCs from donor #1 were used. Target cells and CD8+ T cells were analyzed by flow cytometry 40 hours after incubation with each molecule, and data were fitted using a sigmoidal 4PL fit (GraphPad Prism).

[0035] [Figure 8]Figure 8 shows the cytotoxic activity and effect of PRO2000 and PRO1872 on CD8+ T cell activation in the absence or presence of sMSLN. (A-C) Cytotoxic activity of PRO2000 and PRO1872 against H226 target cells. Specific killing of H226 cells in the absence of sMSLN (A), in the presence of 50 ng / ml sMSLN (B), or in the presence of 500 ng / ml sMSLN (C). The killing activity of PRO2000 is less affected by increasing concentrations of sMSLN compared to PRO1872 (D-F). Similar data are observed for CD8+ T cell activation under corresponding conditions. CD8+ T cell activation in the presence of H226 cells without sMSLN (D), in the presence of 50 ng / ml sMSLN (E), or in the presence of 500 ng / ml sMSLN (F). PBMCs from donor #2 were used. Target cells and CD8+ T cells were analyzed by flow cytometry 40 hours after the start of incubation with each molecule, and data were fitted using a sigmoidal 4PL fit (GraphPad Prism).

[0036] [Figure 9] Figure 9 shows the cytotoxic activity of PRO2000, PRO2100, and PRO1872 in the absence or presence of 100 ng / ml sMSLN: specific killing of H226 cells in the absence (A) or presence (B) of 100 ng / ml sMSLN; specific killing of Met-5A cells (ATCC CRL-9444) in the absence (C) or presence (D) of 100 ng / ml sMSLN. PBMCs from donor #3 were used. Target cells were analyzed by flow cytometry 40 hours after the start of incubation with each molecule, and data were fitted using a sigmoidal 4PL fit (GraphPad Prism).

[0037] [Figure 10]Figure 10 shows the cytotoxic activity of PRO2562, PRO2566, PRO2567, and PRO2660 against high MSLN-expressing cancer cells (H226 cells), intermediate MSLN-expressing cancer cells (OVCAR-3 cells), and low MSLN-expressing cancer cells (Met-5A cells). Palivizumab was used as a negative control ("Control").

[0038] [Figure 11] Figure 11 shows a summary of the EC50 values ​​of PRO2562, PRO2566, PRO2567, PRO2660, and the monovalent reference antibody PRO1872 for specific killing of H226 cells (A), OVCAR3 cells (B, left), and Met-5A cells (B, right).

[0039] [Figure 12] FIG. 12 shows the cytotoxic activity of PRO2562, PRO2566, PRO2567, PRO2660, and the monovalent reference antibody PRO1872 against H226 cells in the absence or presence of 50 ng / ml sMSLN or 500 ng / ml sMSLN.

[0040] [Figure 13] FIG. 13 shows a summary of the EC50 values ​​of PRO2562, PRO2566, PRO2567, PRO2660, and the monovalent reference antibody PRO1872 for specific killing of H226 cells in the absence or presence of 50 ng / ml sMSLN or 500 ng / ml sMSLN.

[0041] [Figure 14-1]Figure 14 shows the binding of MATCH molecules to target cell lines expressing different levels of mesothelin on the cell surface. Binding of PRO2000, PRO2100, and PRO1872 to (A) high-mesothelin-expressing H226 cells, (C) intermediate-mesothelin-expressing HPAC cells, (B) low-mesothelin-expressing H292 cancer cells, and (D) low-mesothelin-expressing mesothelin-expressing mesothelin-containing cells MeT-5A (ATCC CRL-9444) was assessed by flow cytometry. Binding of PRO2562, PRO2566, PRO2567, and PRO2660 to (E) high-mesothelin-expressing H226 cells, (F) intermediate-mesothelin-expressing OVCAR-3 cells, and (G) low-mesothelin-expressing mesothelin-containing cells MeT-5A (ATCC CRL-9444) was assessed by flow cytometry. Data were fitted using a sigmoidal 4PL fit (GraphPad Prism). [Figure 14-2] Figure 14 shows the binding of MATCH molecules to target cell lines expressing different levels of mesothelin on the cell surface. Binding of PRO2000, PRO2100, and PRO1872 to (A) high-mesothelin-expressing H226 cells, (C) intermediate-mesothelin-expressing HPAC cells, (B) low-mesothelin-expressing H292 cancer cells, and (D) low-mesothelin-expressing mesothelin-expressing mesothelin-containing cells MeT-5A (ATCC CRL-9444) was assessed by flow cytometry. Binding of PRO2562, PRO2566, PRO2567, and PRO2660 to (E) high-mesothelin-expressing H226 cells, (F) intermediate-mesothelin-expressing OVCAR-3 cells, and (G) low-mesothelin-expressing mesothelin-containing cells MeT-5A (ATCC CRL-9444) was assessed by flow cytometry. Data were fitted using a sigmoidal 4PL fit (GraphPad Prism).

[0042] [Figure 15]Figure 15 shows that treatment with the molecule PRO2000 (biMSLN.CD3) results in tumor growth inhibition in an H292 xenograft model compared to control conditions. (A) Longitudinal analysis of tumor growth in the presence or absence of biMSLN.CD3 treatment. The line indicates the median. Animals were simultaneously implanted subcutaneously with 1 x 10 H292 tumor cells and 1 x 10 PBMCs. Treatment was administered intravenously starting on day 5 and repeated every 5 days until the end of the experiment. (B) Data at day 40 shown as a scatter plot. Each point corresponds to one animal, and data are presented as the mean and standard deviation. Two-way repeated measures ANOVA followed by Tukey's multiple comparison test was used to indicate the significance of each data set compared to the palivizumab control (Ctrl, lower line) or no treatment (upper line). ns = not significant; *, p < 0.05; **, p < 0.01; ***, p < 0.001. The gray dotted line indicates 0 on the y-axis.

[0043] [Figure 16-1] Figure 16 shows that treatment with the molecule PRO2000(biMSLN.CD3) results in tumor growth inhibition in a human pancreatic cancer (HPAC) xenograft model compared to control conditions. (A) Longitudinal analysis of tumor growth in the presence of increasing concentrations of PRO2000(biMSLN.CD3). The line indicates the median. Animals were simultaneously implanted subcutaneously with 1 x 10 HPAC tumor cells and 2.5 x 10 PBMCs, and treatment was administered intravenously starting on day 5 and repeated every 5 days until the end of the experiment. [Figure 16-2] (B) Corresponding longitudinal analysis of tumor growth in the presence of increasing concentrations of PRO2000(biMSLN.CD3) compared to increasing concentrations of PRO1872(MSLN.CD3). Lines indicate median values. Palivizumab was used as a negative control. *p<0.05; ****p<0.0001; *****p<0.00001. (C) Subsection of the graph shown in (B) showing the time course of only the lowest dose. DETAILED DESCRIPTION OF THE INVENTION

[0044] Known MSLN / CD3 bsAb-based immunotherapies typically suffer from dose-limiting toxicity and limited efficacy in vivo. Thus, there is a need in the medical field for novel MSLN / CD3 bsAb-based immunotherapies that exhibit greater efficacy with reduced or no dose-limiting toxicity than currently available approaches.

[0045] The present invention provides multispecific antibodies that contain a combination of two mesothelin-binding domains (MSLN-BDs) and at least one binding domain for CD3 (CD3-BD). The affinity of the MSLN-BDs for MSLN is tuned to enable efficient localization to the surface of target cells with high MSLN expression, while significantly minimizing the impact on healthy mesothelial cells with low MSLN expression. The presence of two MSLN-BDs embedded within a well-defined, compact multidomain antibody architecture lacking immunoglobulin Fc region polypeptides, combined with well-balanced MSLN and CD3 binding affinities, allows these multispecific antibodies to exhibit high on-target efficacy while minimizing off-tumor side effects. The compact, bivalent design of the multispecific antibodies of the present invention (unachievable with bivalent multispecific antibodies based on classical IgG architecture) as well as their well-tuned MSLN and CD3 binding affinities are crucial features for achieving the desired selectivity and efficacy profile.

[0046] The multispecific antibodies of the present invention can bind to target cells through two MSLN-BDs in a manner that is highly dependent on antigen density by exploiting the avidity effect. At the same time, the multispecific antibodies of the present invention can induce T-cell activation and kill tumor cells by binding to CD3 through the CD3-BD. The multispecific antibodies of the present invention have enhanced selectivity for highly MSLN-expressing cells, leading to efficient tumor localization, enabling treatment without the dose-limiting toxicity caused by nonspecific activation of T cells.

[0047] Additionally, it was surprisingly found that the efficacy of killing highly MSLN-expressing target cells was not significantly reduced in the presence of high levels of soluble mesothelin, which are often observed in patient serum. Furthermore, multispecific antibodies of the present invention, comprising (a) two MSLN-binding domains and (b) at least one CD3-BD and having the above-described design and antigen-binding affinity, exhibited additional beneficial properties, as shown in the Examples and accompanying figures. Furthermore, the optional addition of a half-life-extending anti-hSA domain not only allows for convenient administration but also facilitates delivery of the molecule to the tumor microenvironment.

[0048] Thus, the multispecific antibodies of the present invention offer distinct therapeutic advantages over conventional compositions and therapies.

[0049] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0050] The terms "comprising" and "including" are used herein in their open-ended sense unless expressly stated otherwise. Thus, with respect to such latter embodiments, the term "comprising" encompasses the narrower term "consisting of."

[0051] In the context of describing the present invention (particularly in the context of the claims below), the terms "a," "an," "the," and similar references are to be construed as including both the singular and the plural, unless otherwise indicated herein or clearly contradicted by the context. For example, the term "a cell" includes a plurality of cells (including mixtures thereof). When the plural is used for compounds, salts, etc., this is understood to mean a single compound, salt, etc.

[0052] In one aspect, the present invention provides a method for producing a medicament for the treatment of a pulmonary arthritis. a) Two antibody-based binding domains (MSLN-BD) that specifically bind to mesothelin; b) comprises at least one antibody-based binding domain (CD3-BD) that specifically binds to CD3; Each of the MSLN-BDs does not contain an immunoglobulin Fc region polypeptide and has a monovalent dissociation constant (K) in the range of 0.1 to 20 nM as measured by SPR. D ) relates to a multispecific antibody that binds to mesothelin (MSLN).

[0053] As used herein, terms such as "antibody" encompass whole antibodies or single chains thereof; and any antigen-binding fragment (i.e., "antigen-binding portion") or single chains thereof; and molecules comprising antibody CDRs, VH regions, or VL regions (non-limiting examples of which include multispecific antibodies). A naturally occurring "whole antibody" is a glycoprotein comprising at least two heavy (H) chains and two light (L) chains interconnected by disulfide bonds. Each heavy chain consists of a heavy chain variable region (abbreviated herein as VH) and a heavy chain constant region. The heavy chain constant region consists of three domains, CH1, CH2, and CH3. Each light chain consists of a light chain variable region (abbreviated herein as VL) and a light chain constant region. The light chain constant region consists of one domain, CL. The VH and VL regions can be further divided into regions of hypervariability called complementarity-determining regions (CDRs), interspersed with more conserved regions called framework regions (FRs). Each VH and VL consists of three CDRs and four FRs, arranged from amino-terminus to carboxy-terminus in the following order: FR1, CDR1, FR2, CDR2, FR3, CDR3, FR4. The variable regions of the heavy and light chains contain binding domains that interact with antigens. The constant regions of the antibody can mediate the binding of the immunoglobulin to host tissues or factors, including various cells of the immune system (e.g., effector cells) and the first component (Clq) of the classical complement system.

[0054] The term "immunoglobulin Fc region" as used herein refers to the CH2 and CH3 domains of the heavy chain constant region.

[0055] The terms "binding domain," "antigen-binding fragment thereof," "antigen-binding portion" of an antibody, and the like, are used herein to refer to one or more fragments of an intact antibody that retain the ability to specifically bind to a given antigen (e.g., MSLN, CD3, hSA). The antigen-binding function of an antibody can be performed by fragments of an intact antibody. In some embodiments, the selection of binding domains for the multispecific antibodies of the invention can be made from a Fab fragment (a monovalent fragment consisting of the VL, VH, CL, and CH1 domains); a F(ab)2 fragment (a bivalent fragment comprising two Fab fragments linked by a disulfide bridge at the hinge region); a Fd fragment consisting of the VH and CH1 domains; a Fv fragment consisting of the VL and VH domains of a single arm of an antibody; or a single-domain antibody (dAb) fragment consisting of the VH domain (Ward et al., 1989 Nature ). 341:544-546; isolated complementarity determining regions (CDRs), single chain Fvs, dsFvs, scAbs, STABs, single domain antibodies (sdAbs or dAbs), single domain heavy chain antibodies and single domain light chain antibodies, VHHs, VNARs, single domain antibodies based on VNAR structures from sharks, and binding domains based on alternative scaffolds (non-limiting examples include ankyrin-based domains, finomers, avimers, anticalins, fibronectin, and binding sites incorporated into the constant region of an antibody (e.g., f-star technology (F-star modular antibody technology™)). The binding domain of the present invention is suitably a single chain Fv fragment (scFv) or a single antibody variable domain. In a preferred embodiment, the binding domain of the present invention is a single chain Fv fragment (scFv). In a particular embodiment, the two variable domains of the antigen-binding fragment are stabilized by an interdomain disulfide bond, as in an Fv or scFv fragment, and in particular the VH domain contains a single cysteine ​​residue at position 51 (AHo numbering) and the VL domain contains a single cysteine ​​residue at position 141 (AHo numbering).

[0056] The term "complementarity determining region" ("CDR") means an amino acid sequence whose boundaries are determined using any of a number of well-known schemes. Such schemes include those described by Kabat et al. (1991), "Sequences of Proteins of Immunological Interest," 5th Ed. Public Health Service, National Institutes of Health, Bethesda, Maryland ("Kabat" numbering scheme); Al-Lazikani et al., (1997) JMB 273, 927-948 ("Chothia" numbering scheme); ImMunoGenTics (IMGT) numbering (Lefranc, M.-P., The Immunologist, 7, 132-136 (1999); Lefranc, M.-P. et al., Dev. Comp. Immunol., 27, 55-77 (2003)) ("IMGT" numbering scheme); and Honegger & Pluckthun, J. Mol. Biol. 309 (2001) The numbering scheme is set forth in U.S. Pat. No. 6,657-670 ("AHo" numbering). For example, with respect to the classical format, under Kabat, the CDR amino acid residues in the heavy chain variable domain (VH) are numbered 31-35 (HCDR1), 50-65 (HCDR2), and 95-102 (HCDR3); the CDR amino acid residues in the light chain variable domain (VL) are numbered 24-34 (LCDR1), 50-56 (LCDR2), and 89-97 (LCDR3). Under Chothia, the CDR amino acid residues in the VH are numbered 26-32 (HCDR1), 52-56 (HCDR2), and 95-102 (HCDR3); the amino acid residues in the VL are numbered 24-34 (LCDR1), 50-56 (LCDR2), and 89-97 (LCDR3).Combining the CDR definitions of both Kabat and Chothia, the CDRs consist of amino acid residues 26-35 (HCDR1), 50-65 (HCDR2), and 95-102 (HCDR3) in human VH, and amino acid residues 24-34 (LCDR1), 50-56 (LCDR2), and 89-97 (LCDR3) in human VL. Under IMGT, the CDR amino acid residues in VH are numbered approximately 26-35 (HCDR1), 51-57 (HCDR2), and 93-102 (HCDR3), and those in VL are numbered approximately 27-32 (LCDR1), 50-52 (LCDR2), and 89-97 (LCDR3) (numbering according to "Kabat"). Under IMGT, the CDRs of an antibody can be determined using the program IMGT / DomainGapAlign.

[0057] In the context of the present invention, the numbering system proposed by Honegger & Pluckthun ("AHo") is used (Honegger & Pluckthun, J. Mol. Biol. 309 (2001) 657-670), unless otherwise specified. In particular, the following residues are defined as CDRs according to the AHo numbering scheme: LCDR1 (also referred to as CDR-L1): L24-L42; LCDR2 (also referred to as CDR-L2): L58-L72; LCDR3 (also referred to as CDR-L3): L107-L138; HCDR1 (also referred to as CDR-H1): H27-H42; HCDR2 (also referred to as CDR-H2): H57-H76; HCDR3 (also referred to as CDR-H3): H108-H138. For clarity, the numbering system by Honegger & Plückthun takes into account the length diversity found in natural antibodies (both the different VH and VL subfamilies and the CDRs, especially the CDRs), and provides for gaps in the sequence, so that in a given antibody variable domain, positions 1 to 149 are not usually all occupied by amino acid residues.

[0058] The term "binding specificity" as used herein refers to the ability of an individual antibody to react with one antigenic determinant and not with a different antigenic determinant. As used herein, the terms "specifically bind" or "specific for" refer to a measurable and reproducible interaction (e.g., binding between a target and an antibody) that determines the presence of a target in the presence of a heterogeneous population of molecules, including biological molecules. For example, an antibody that specifically binds to a target (possibly an epitope) is one that binds to this target with greater affinity, avidity, more readily, and / or longer than it binds to other targets. In its most general form (and when no defined criteria are mentioned), "specific binding" refers to the ability of an antibody to distinguish between a target of interest and unrelated molecules when tested, for example, according to specificity assays known in the art. Non-limiting examples of such assays include Western blot, ELISA, RIA, ECL, IRMA, SPR (surface plasmon resonance) testing, and peptide scanning. For example, a standard ELISA assay can be performed. Numerical values ​​can be determined by standard colorimetric methods (e.g., secondary antibodies and horseradish peroxide, and tetramethylbenzidine and hydrogen peroxide). The reaction in a well is quantified, for example, by the optical density at 450 nm. A typical background (=negative response) will be about 0.1 OD; a typical positive response will be about 1 OD. This means that the ratio of positive to negative values ​​can be 10-fold or greater. In a further example, an SPR assay can be performed, where a difference of at least 10-fold, and particularly at least 100-fold, between background and signal indicates specific binding. Typically, binding specificity is determined not using a single reference molecule, but by using a set of about 3-5 unrelated molecules (e.g., milk powder, transferrin, etc.).

[0059] The antibodies of the present invention are suitably isolated antibodies. The term "isolated antibody," as used herein, refers to an antibody that is substantially free of other antibodies with different antigen specificities (e.g., an isolated antibody that specifically binds to MSLN and CD3 is substantially free of antibodies that specifically bind to antigens other than MSLN and CD3, and an isolated antibody that specifically binds to MSLN, CD3, and human serum albumin is substantially free of antibodies that specifically bind to antigens other than MSLN, CD3, and human serum albumin). Furthermore, an isolated antibody may be substantially free of other cellular material and / or chemicals.

[0060] The antibodies of the present invention are suitably monoclonal antibodies. The terms "monoclonal antibody" or "monoclonal antibody composition" as used herein refer to antibodies having substantially the same amino acid sequence or derived from the same genetic source. Monoclonal antibody compositions exhibit binding specificity and affinity for a particular epitope or epitopes.

[0061] Non-limiting examples of antibodies of the present invention include chimeric antibodies, human antibodies, and humanized antibodies.

[0062] The term "chimeric antibody" (or antigen-binding fragment thereof) refers to (a) an antibody molecule (or antigen-binding fragment thereof) in which the constant region or a portion thereof has been altered, substituted, or exchanged so that the antigen-binding site (variable region) is linked to a constant region of a different or altered class, effector function, and / or species, or to an entirely different molecule (e.g., an enzyme, toxin, hormone, growth factor, drug, etc.) that confers new properties to the chimeric antibody; or (b) an antibody molecule (or antigen-binding fragment thereof) in which the variable region or a portion thereof has been altered, substituted, or exchanged with a variable region having a different or altered antigen specificity. For example, a murine antibody can be modified by replacing its constant region with a constant region from a human immunoglobulin. The replacement with a human constant region allows the chimeric antibody to retain its specificity in antigen recognition, but reduces its antigenicity in humans compared to the original murine antibody.

[0063] The term "human antibody" (or antigen-binding fragment thereof), as used herein, is intended to include antibodies (and antigen-binding fragments thereof) having variable regions in which both the framework and CDR regions are derived from sequences of human origin. Furthermore, if the antibody contains a constant region, the constant region also is derived from such a human sequence (e.g., a human germline sequence or a mutated version of a human germline sequence). The human antibodies and antigen-binding fragments thereof of the present invention may contain amino acid residues not encoded by human sequences (e.g., mutations introduced by random or site-specific mutagenesis in vitro or by somatic mutation in vivo). This definition of a human antibody specifically excludes humanized antibodies comprising non-human antigen-binding residues. Human antibodies can be produced using a variety of techniques known in the art, including phage display libraries (Hoogenboom and Winter, J. Mol. Biol, 227:381 (1991); Marks et al., J. Mol. Biol, 222:581 (1991)). Also available for preparing human monoclonal antibodies are methods described in Cole et al., Monoclonal Antibody and Cancer Therapy, Alan R. Liss, p. 77 (1985); Boemer et al., J. Immunol, 147(1):86-95 (1991). See also van Dijk and van de Winkel, Curr. Opin. Pharmacol, 5: 368-74 (2001). Human antibodies can be prepared by administering antigen to transgenic animals (e.g., immunized xenomouse (see, e.g., U.S. Patent Nos. 6,075,181 and 6,150,584 for XENOMOUSE™ technology)) that have been engineered to produce such antibodies in response to antigen challenge, but in which the endogenous gene locus has been rendered nonfunctional. See also, e.g., Li et al., Proc. Natl. Acad. Sci. USA, 103:3557-3562 (2006) regarding human antibodies produced by human B-cell hybridoma technology.

[0064] A "humanized" antibody (or antigen-binding fragment thereof), as used herein, is an antibody (or antigen-binding fragment thereof) that retains the reactivity of a non-human antibody while being less immunogenic in humans. This can be achieved, for example, by retaining the non-human CDR regions while replacing the remainder of the antibody with its human counterparts (i.e., the constant regions as well as the framework portions of the variable regions). Additional framework region modifications can be made within the human framework sequences as well as within CDR sequences derived from the germline of another mammalian species. The humanized antibodies of the invention may contain amino acid residues not encoded by human sequences (e.g., mutations introduced by random or site-directed mutagenesis in vitro or by somatic mutation in vivo, or conservative substitutions that promote stability or manufacturing). See, e.g., Morrison et al., Proc. Natl. Acad. Sci. USA, 81:6851-6855, 1984; Morrison and Oi, Adv. Immunol., 44:65-92, 1988; Verhoeyen et al., Science, 239:1534-1536, 1988; Padlan, Molec. Immun., 28:489-498, 1991; and Padlan, Molec. Immun., 31:169-217, 1994. Other non-limiting examples of human engineering techniques include the Xoma technology disclosed in U.S. Patent No. 5,766,886.

[0065] The term "recombinant humanized antibody," as used herein, encompasses any human antibody prepared, expressed, created, or isolated by recombinant means (such as antibodies isolated from host cells (e.g., transfectomas) that have been transformed to express a humanized antibody, as well as antibodies prepared, expressed, created, or isolated by any other means, including splicing all or part of one human immunoglobulin gene sequence to another DNA sequence).

[0066] Suitably, the antibody or antigen-binding fragment thereof of the present invention is humanized. Suitably, the antibody or antigen-binding fragment thereof of the present invention is humanized and comprises CDRs from rabbit.

[0067] The term "multispecific antibody" as used herein refers to an antibody that binds to two or more different epitopes on at least two or more different targets (e.g., MSLN and CD3). The term "multispecific antibody" includes bispecific, trispecific, tetraspecific, pentaspecific, and hexaspecific antibodies. The term "bispecific antibody" as used herein refers to an antibody that binds to two different epitopes on two different targets (e.g., MSLN and CD3). The term "trispecific antibody" as used herein refers to an antibody that binds to three different epitopes on three different targets (e.g., MSLN, CD3, and hSA).

[0068] The term "epitope" refers to a protein determinant capable of specific binding to an antibody. Epitopes usually consist of chemically active surface groupings of molecules (such as amino acids or sugar side chains) and usually have specific three-dimensional structural characteristics, as well as specific charge characteristics. "Conformational" epitopes are distinguished from "linear" epitopes by the loss of binding to the former but not the latter in the presence of denaturing solvents.

[0069] The term "conformational epitope" as used herein refers to amino acid residues of an antigen that are grouped together on the surface when the polypeptide chain folds to form the native protein.

[0070] The term "linear epitope" refers to an epitope that comprises all of the points of interaction that occur between a protein and an interacting molecule (such as an antibody) linearly along the primary amino acid sequence (continuous) of a protein.

[0071] The term "recognize" as used herein refers to an antibody antigen-binding fragment thereof that finds and interacts with (eg, binds to) the corresponding conformational epitope.

[0072] As used herein, the term "affinity" refers to the strength of the interaction between an antibody and an antigen at a single antigenic site. Within each antigenic site, the variable regions of an antibody "arm" interact with the antigen at multiple sites through weak non-covalent forces; the greater the interaction, the stronger the affinity.

[0073] "Affinity" generally refers to the strength of the sum of all non-covalent interactions between a single binding site of a molecule (e.g., an antibody) and its binding partner (e.g., an antigen). Unless otherwise specified, as used herein, "affinity," "binds," "binds to," or "binding to" refers to the inherent affinity that reflects a 1:1 interaction between members of a binding pair (e.g., an antibody fragment and an antigen). The affinity of a molecule X for its partner Y is generally measured by the dissociation constant (K D ) Affinity can be measured by common methods known in the art, including those described herein. Low affinity antibodies generally bind antigens slowly and tend to dissociate easily, whereas high affinity antibodies generally bind antigens faster and tend to remain bound longer. A variety of methods for measuring affinity are known in the art, any of which can be utilized for purposes of the present invention. Specific illustrative and representative embodiments for measuring affinity (i.e., binding) strength are described below.

[0074] The terms "K," "Ka," or "K" are intended herein to refer to the association rate of a particular antibody-antigen interaction, while the terms "K," "K," or "K" are intended herein to refer to the dissociation rate of a particular antibody-antigen interaction. In one embodiment, "K" is used to refer to the association rate of a particular antibody-antigen interaction. D The term "K" is intended herein to mean the dissociation constant obtained from the ratio of K to K (i.e., K / K), expressed as a molar concentration (M). D " or "K DIn one embodiment, the "KD value" or "KD value" is measured using a surface plasmon resonance assay. Affinity for recombinant human mesothelin (human MSLN) and recombinant cynomolgus monkey MSLN (cyno MSLN) was determined by surface plasmon resonance (SPR) measurements as described in paragraph

[0168] . Affinity for recombinant human CD3 was measured by SPR as described in paragraph

[0196] .

[0075] Suitably, the multispecific antibodies of the invention are bivalent with respect to MSLN specificity.

[0076] Multispecific antibodies of the invention are suitably monovalent, bivalent, or multivalent with respect to CD3 specificity. In one embodiment, multispecific antibodies of the invention are bivalent with respect to CD3 specificity. In a preferred embodiment, multispecific antibodies of the invention are monovalent with respect to CD3 specificity.

[0077] The term "multivalent antibody" refers to a single binding molecule with two or more valencies, where "valency" is described as the number of antigen-binding moieties that bind to the same epitope on the target molecule. Thus, a single binding molecule can bind to two or more binding sites on a target molecule. Non-limiting examples of multivalent antibodies include bivalent antibodies, trivalent antibodies, tetravalent antibodies, pentavalent antibodies, etc.

[0078] The term "monovalent antibody" as used herein refers to an antibody that binds to a single epitope on a target molecule (such as CD3). Also, the terms "binding domain" or "monovalent binding domain" as used herein refer to a binding domain that binds to a single epitope on a target molecule (such as CD3).

[0079] The term "bivalent antibody" as used herein means an antibody that binds to two epitopes on the same two target molecules (such as the MSLN target molecule).

[0080] The two MSLN-BDs of the multispecific antibody of the present invention bind to any region of the extracellular portion of MSLN (e.g., region I, region II, and / or region III of MSLN). Preferably, the two MSLN-BDs of the multispecific antibody of the present invention bind to region I and / or region II of MSLN, particularly region I of MSLN. Region I is the portion of MSLN that is most distal from the cell surface to which MSLN adheres.

[0081] The two MSLN-BDs of the multispecific antibody of the present invention bind to the same or different epitopes on the MSLN target molecule. Preferably, the two MSLN-BDs of the multispecific antibody of the present invention bind to the same epitope on the MSLN target molecule. The term "same epitope" as used herein means individual protein determinants on the surface of the same protein that can specifically bind to an antibody, each of which is the same, i.e., consisting of the same chemically active surface groups of the molecule (such as amino acids or sugar side chains with the same three-dimensional structural characteristics and charge characteristics). The term "different epitopes," as used herein in relation to a specific protein target, means individual protein determinants on the same protein that can specifically bind to an antibody, each of which is not the same, i.e., consisting of different chemically active surface groups of the molecule (such as amino acids or sugar side chains with different three-dimensional structural characteristics and charge characteristics). These different epitopes may or may not overlap.

[0082] The inventors of the present invention have now surprisingly found that e.g. trispecific molecules (biMSLNs) 高KD xCD3xhSA) PRO2000, PRO2562, PRO2565, PRO2566, and PRO2567 highly efficiently killed target cells with MSLN expression levels approximately 7-fold greater than healthy MeT-5A cells (ATCC CRL-9444) as determined by flow cytometry, and EC2000 was significantly higher than that of healthy MeT-5A cells as determined in a T cell-driven cytotoxicity assay against the target cells and the MeT-5A cells. 50EC at least 25 times smaller than 50 (See, e.g., Table 31.) Thus, PRO2000, PRO2562, PRO2566, and PRO2567 exhibit very high killing potency toward high MSLN-expressing target molecules, but much less killing potency toward healthy cells, indicating a potentially broad therapeutic window for treatment with PRO2000, PRO2562, PRO2566, and PRO2567. In contrast, PRO2000, PRO2562, PRO2566, and PRO2567 have over 5-fold greater affinity (K) than MSLN-BD for killing the high MSLN-expressing target cells and the healthy Met-5A cells. D ) and the trispecific reference molecule PRO1872 (MSLN 低KD The efficacy of the trispecific molecules PRO2000, PRO2562, PRO2566, and PRO2567 (biMSLN) in killing target cells with MSLN expression levels approximately 7-fold greater than the healthy MeT-5A cells, as determined by flow cytometry. 高KD xCD3xhSA) EC 50 We found that the activity of soluble mesothelin (sMSLN) increased by no more than 6-fold in the presence of 50 ng / ml and no more than 40-fold in the presence of 500 ng / ml of soluble mesothelin (sMSLN) when measured in a T cell-driven cytotoxicity assay against the target cells. On the other hand, the trispecific reference molecule PRO1872 (MSLN) for killing the target cells was significantly increased. 低KD xCD3xhSA) EC 50The values ​​increased nearly 8-fold in the presence of 50 ng / ml sMSLN and more than 75-fold in the presence of 500 ng / ml sMSLN. Thus, the significant killing potency of the trispecific molecules PRO2000, PRO2562, PRO2566, and PRO2567 against highly MSLN-expressing target cells is only slightly affected by high concentrations of sMSLN. On the other hand, the killing potency of the trispecific molecules PRO2000, PRO2562, PRO2566, and PRO2567 against healthy cells is further reduced in the presence of high concentrations of sMSLN (data not shown). This indicates that the therapeutic window for their clinical use is further broadened by the presence of sMSLN. This is an important finding because high plasma levels of soluble mesothelin-related protein (SMRP) are often observed in patients. We obtained similar favorable results regarding the CD8+ T cell activation potency of PRO2000. These findings are all the more surprising because it would not have been expected a priori that all four binding domains would remain functional in a complex multi-target, multi-cellular context without interfering with each other, sterically or otherwise.

[0083] Suitable MSLN-BDs for use in the multispecific antibodies of the invention are the binding domains provided in this disclosure. Non-limiting examples of mesothelin-BDs of the invention include the humanized MSLN-binding domains whose sequences are listed in Table 1.

[0084] CD3-BDs suitable for use in the multispecific antibodies of the invention are the binding domains provided in this disclosure. Non-limiting examples of CD3-BDs of the invention include the humanized CD3-binding domains whose sequences are listed in Table 3.

[0085] Suitably, the multispecific antibodies of the invention have two different specificities (MSLN and CD3). Suitably, the multispecific antibodies of the invention are bispecific antibodies, bivalent with respect to MSLN. The multispecific antibodies of the invention can comprise one additional specificity (triabody) or multiple specificities (tetraspecific, pentaspecific, or hexaspecific antibodies). In one embodiment, the multispecific antibody is bispecific (MSLN and CD3). In another embodiment, the multispecific antibody is trispecific (MSLN, CD3, and hSA).

[0086] Suitably, the antibodies of the invention do not comprise immunoglobulin Fc region polypeptides.

[0087] To increase the number of specificities / functions at the same or smaller molecular weight, it is advantageous to use antibodies, including antibody fragments (e.g., Fv, Fab, Fab', and F(ab')2 fragments, as well as other antibody fragments). These smaller molecules retain the antigen-binding activity of whole antibodies and may also exhibit improved tissue penetration and pharmacokinetic properties compared to whole immunoglobulin molecules. While such fragments appear to offer many advantages over whole immunoglobulin molecules, they also suffer from increased serum clearance due to the lack of the Fc domain, which confers a long half-life in vivo (Medasan et al., 1997, J. Immunol. 158:2211-2217). Smaller molecular weight molecules penetrate target tissues (e.g., solid tumors) more efficiently, potentially resulting in improved efficacy at the same or lower doses.

[0088] The inventors have surprisingly found that the addition of a human serum albumin binding domain (hSA-BD) to a multispecific antibody of the invention does not interfere with the ability of the other binding domains to bind to their respective targets. This finding is all the more surprising, as it would not be expected a priori that all four binding domains would remain functional without sterically or otherwise interfering with each other in a complex multi-target, multi-cellular in vivo context.

[0089] Suitably, the multispecific antibody of the invention may comprise additional binding domains with specificity for human serum albumin. In one embodiment, the multispecific antibody comprises (i) two MSLN-BDs; (ii) at least one CD3-BD; and (iii) at least one hSA-BD.

[0090] The term "hSA" specifically refers to human serum albumin, which has UniProt ID number P02768. Human serum albumin (hSA) is a 66.4 kDa protein (50% of all proteins) abundant in human serum and consists of 585 amino acids (Sugio, Protein Eng, Vol. 12, 1999, 439-446). The multifunctional hSA protein contains structures that allow the binding and transport of numerous metabolites (e.g., fatty acids), metal ions, bilirubin, and several drugs (Fanali, Molecular Aspects of Medicine, Vol. 33, 2012, 209-290). The concentration of HSA in serum is approximately 3.5-5 g / dL. Albumin-binding antibodies and their fragments can be used, for example, to extend the in vivo serum half-life of drugs or proteins complexed with them.

[0091] In some embodiments, the hSA-BD is derived from a monoclonal antibody or antibody fragment.

[0092] Suitable hSA-BDs for use in the multispecific antibodies of the invention are the binding domains provided in this disclosure. Non-limiting examples of hSA-BDs of the invention include the humanized hSA binding domains whose sequences are listed in Table 4.

[0093] In particular, the hSA-BD of the present invention specifically binds to human serum albumin.

[0094] Other hSA-BDs suitable for use in the multispecific antibodies of the invention include (i) serum albumin-binding polypeptides (see, e.g., Smith et al., 2001, Bioconjugate Chem. 12:750-756; European Patent No. EP0486525; U.S. Patent No. 6267964; WO2004 / 001064; WO2002 / 076489; and WO2001 / 45746); (ii) anti-serum albumin-binding single variable domains (Holt et al., Protein Engineering, Design & Selection, vol 21, 5, pp283-288, WO2004 / 003019, WO2008 / 096158, WO2005 / 118642, WO2006 / 0591056, and WO2011 / 006915); (iii) anti-serum albumin antibodies (described in WO2009 / 040562, WO2010 / 035012, and WO2011 / 086091).

[0095] Other variable domains of the invention contain amino acid sequences that are mutated but still have at least 80, 85, 90, 91, 92, 93, 94, 95, 96, 97, 98, or 99 percent identity in the CDR regions to the CDR regions shown in the sequences set out in Tables 1, 3, and 4. Other variable domains of the invention contain mutated amino acid sequences in which no more than 1, 2, 3, 4, or 5 amino acids are mutated in the CDR regions when compared to the CDR regions shown in the sequences set out in Tables 1, 3, and 4.

[0096] Suitably, the VH domain of the binding domain of the present invention belongs to the VH3 or VH4 family. In one embodiment, the binding domain of the present invention comprises a VH domain belonging to the VH3 family. In the context of the present invention, the expression "belonging to the VHx family (or VLx family)" means that the framework sequences FR1 to FR3 show the greatest homology to said VHx family (or VLx, respectively). Examples of VH and VL families are given in Knappik et al., J. Mol. Biol. 296 (2000) 57-86 or in WO2019 / 057787. A particular example of a VH domain belonging to the VH3 family is represented by SEQ ID NO: 129, and a particular example of a VH domain belonging to the VH4 family is represented by SEQ ID NO: 130. In particular, the framework regions FR1 to FR3 from SEQ ID NO: 129 belong to the VH3 family (Table 7, regions not bolded). A VH belonging to the VH3 family, as used herein, is suitably a VH comprising FR1-FR3 that is at least 85%, particularly at least 90%, and more particularly at least 95% identical to FR1-FR3 of SEQ ID NO: 129. Alternative examples of VH3 sequences and VH4 sequences can be found in Knappik et al., J. Mol. Biol. 296 (2000) 57-86 or WO2019 / 057787. The hSA-BD of the present invention suitably comprises Vκ framework FR1, FR2, and FR3 (particularly Vκ1 or Vκ3 frameworks, particularly Vκ1 framework FR1-3) and framework FR4 (selected from Vκ FR4 and Vλ FR4, with Vλ FR4 being particularly selected). Suitable Vκ1 framework FR1-3, as well as a representative Vλ FR4, are shown in SEQ ID NO: 131 (Table 7, FR regions are shown in non-bold). Alternative examples of Vκ1 sequences, as well as examples of Vκ2, Vκ3, or Vκ4 sequences, can be found in Knappik et al., J. Mol. Biol. 296 (2000) 57-86.Suitable Vκ1 framework FR1-3 comprise amino acid sequences corresponding to FR1-3 and at least 70, 80, 90, or 95 percent identical to the amino acid sequence from SEQ ID NO: 131 (Table 7, FR regions are not bolded). Suitable Vλ FR4s are set forth in SEQ ID NOs: 132-138 and 139, which comprise a single cysteine ​​residue, particularly if a second single cysteine ​​is present, in the corresponding VH chain, particularly at position 51 of VH (AHo numbering), for interdomain disulfide bond formation. In one embodiment, a VL domain of the invention comprises a Vλ FR4 that is at least 70, 80, or 90 percent identical to an amino acid sequence selected from any of SEQ ID NOs: 132-139 (particularly SEQ ID NO: 132 or 139).

[0097] Binding domains of the invention include VH domains listed in Tables 1, 3, and 4. Suitably, binding domains of the invention comprise a VH amino acid sequence listed in one of Tables 1, 3, and 4, in which 20 or fewer amino acids in the framework sequences (e.g., sequences that are not CDRs) have been mutated (mutations are, by way of various non-limiting examples, additions, substitutions, or deletions). Suitably, binding domains of the invention comprise a VH amino acid sequence listed in one of Tables 1, 3, and 4, in which 15 or fewer amino acids, particularly 10 or fewer amino acids, particularly 5 or fewer amino acids, have been mutated (mutations are, by way of various non-limiting examples, additions, substitutions, or deletions) in the framework sequences (e.g., sequences that are not CDRs). Other binding domains of the invention are mutated but still contain at least 80, 85, 90, 91, 92, 93, 94, 95, 96, 97, 98, or 99 percent identical amino acids in the VH region to a VH region shown in a corresponding sequence set forth in one of Tables 1, 3, and 4 (including VH domains comprising at least positions 5-140 (AHo numbering), particularly at least positions 3-145, of one of the sequences set forth in Tables 1, 3, and 4).

[0098] In particular, binding domains of the invention comprise a VL domain listed in one of Tables 1, 3, and 4. Suitably, binding domains of the invention comprise a VL amino acid sequence listed in one of Tables 1, 3, and 4, in which 20 or fewer amino acids in the framework sequences (e.g., sequences that are not CDRs) have been mutated (mutations are, by way of various non-limiting examples, additions, substitutions, or deletions). Suitably, binding domains of the invention comprise a VL amino acid sequence listed in one of Tables 1, 3, and 4, in which 15 or fewer amino acids, particularly 10 or fewer amino acids, particularly 5 or fewer amino acids, have been mutated (mutations are, by way of various non-limiting examples, additions, substitutions, or deletions) in the framework sequences (e.g., sequences that are not CDRs). Other binding domains of the invention are mutated but still contain at least 80, 85, 90, 91, 92, 93, 94, 95, 96, 97, 98, or 99 percent amino acid identity within the VL region to a VL region shown in a corresponding sequence set forth in one of Tables 1, 3, and 4 (including VL domains comprising at least positions 5-140 (AHo numbering), particularly at least positions 3-145, of one of the sequences set forth in Tables 1, 3, and 4).

[0099] In the context of the present invention, the expression "binding domain of the invention" relates both to a binding domain as such, i.e. independent of the multispecific context, and to a binding domain comprised in a multispecific construct (e.g. one of the binding domains comprised in a bispecific, trispecific or tetraspecific construct), in particular to the latter.

[0100] Suitably, the binding domain of the present invention is selected from the group consisting of a Fab, Fv, scFv, dsFv, scAb and STAB.

[0101] Suitably, the binding domain of the present invention is an scFv antibody fragment.

[0102] The multispecific antibodies of the invention can be in any suitable format.

[0103] Suitably, the binding domains of a multispecific antibody are operably linked. The binding domains of a multispecific antibody of the present invention can simultaneously bind to their respective antigens or receptors. The term "simultaneously" when used in this context refers to simultaneous binding of at least one MSLN-BD and one CD3-BD. In special cases, such as target cells with a high density of MSLN on the cell surface, simultaneous binding of all three binding domains, i.e., both MSLN-BD and CD3-BD, may also be possible.

[0104] The multispecific antibody of the present invention comprises two MSLN-BDs and at least one CD3-BD, and the MSLN-BD and the CD3-BD are operably linked to each other.

[0105] The phrase "operably linked" as used herein refers to two molecules (e.g., polypeptides, domains, binding domains) attached such that each retains functional activity. Two molecules can be "operably linked" whether they are directly attached or indirectly attached (e.g., via a linker, a moiety, or a linker to a moiety). The term "linker" refers to a peptide or other moiety, optionally located between a binding domain or an antibody fragment of the present invention. Numerous strategies can be used to covalently link molecules to each other. Non-limiting examples include N- to C-terminal polypeptide linkages between proteins or protein domains, linkages via disulfide bonds, and linkages via chemical cross-linking reagents. In one aspect of this embodiment, the linker is a peptide bond generated by recombinant technology or peptide synthesis. The selection of an appropriate linker for a particular case in which two polypeptide chains are to be joined depends on various parameters, including, but not limited to, the properties of the two polypeptide chains (e.g., whether they naturally oligomerize), the distance between the N- and C-termini to be joined, if known, and / or the stability of the linker against proteolytic degradation and oxidation. Additionally, the linker can contain amino acid residues that provide flexibility.

[0106] In the context of the present invention, the term "polypeptide linker" refers to a linker consisting of a chain of amino acid residues linked by peptide bonds that connects two domains, with each domain attached to one end of the linker. The polypeptide linker must be long enough to link the two molecules so that they assume the correct conformation relative to each other and retain the desired activity. In particular embodiments, the polypeptide linker has a continuous chain of 2 to 30 amino acid residues (e.g., 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, or 30 amino acid residues). Additionally, the amino acid residues selected for inclusion in the polypeptide linker must exhibit properties that do not significantly interfere with the activity of the polypeptide. Therefore, the linker peptide as a whole must not exhibit a charge. Charges are believed to significantly impede the binding of receptor monomer domains by being incompatible with the activity of the polypeptide, by interfering with internal folding, or by forming bonds or other interactions with amino acid residues in one or more of the monomers. In particular embodiments, the polypeptide linker is a simple polypeptide. Useful linkers include glycine-serine or GS linkers. A "Gly-Ser" or "GS" linker, as will be appreciated by those skilled in the art, is a polymer of tandem glycines and serine residues (e.g., (Gly-Ser) n , (GSGGS) n , (GGGGS) n , and (GGGS) n(where n is an integer of at least 1), glycine-alanine polymers, alanine-serine polymers, and other flexible linkers (such as the connecting chains for Shaker potassium channels) and a variety of other flexible linkers. Glycine-serine polymers are preferred because both of these amino acids are relatively simple and can function as neutral connecting chains between components. Second, serine is hydrophilic, allowing the solubilization of what could be globular glycine chains. Third, similar chains have been shown to be effective for joining subunits of recombinant proteins (such as single-chain antibodies).

[0107] The multispecific antibody is suitably in a format selected from any suitable multispecific (e.g., at least bispecific) format known in the art that does not comprise an immunoglobulin Fc region, non-limiting examples of which include tandem scDb (Tandab), linear dimeric scDb (LD-scDb), cyclic dimeric scDb (CD-scDb), tandem tri-scFv, tribody (Fab-(scFv)2), Fab-Fv2, triabody, scDb-scFv, tetrabody, di-diabody, CODV, tandem-di-scFv, tandem tri-scFv, Fab-(scFv)2, Fab-Fv2, or CODV fused to the N-terminus and / or C-terminus of a heterodimerization domain other than the heterodimeric Fc domain, and MATCH (WO2016 / 0202457; Egan T. et al., MABS 9 (2017) 68-84), and formats based on DuoBodies (bispecific IgG prepared by Duobody technology) (MAbs. 2017 Feb / Mar;9(2):182-212. doi:10.1080 / 19420862.2016.1268307). Particularly suitable multispecific antibodies are single-chain diabodies (scDb)-scFv or MATCH.

[0108] In one embodiment, the multispecific antibody of the invention does not comprise a CH1 region and / or a CL region.

[0109] In another embodiment, the multispecific antibody of the present invention is in a format selected from the list consisting of scDb-scFv, triabody, and tribody. Particularly suitable for use herein is scDb-scFv, in particular wherein one of the MSLN-BD and the CD3-BD is in the form of an scDb and the second MSLN-BD is an scFv operably linked to the scDb.

[0110] The term "diabody" refers to an antibody fragment having two antigen-binding sites, which fragment comprises a VH connected to a VL in the same polypeptide chain (VH-VL). By using a linker that is too short to allow pairing between the two domains on the same chain, the domains are forced to pair with complementary domains on another chain, creating two antigen-binding sites. Diabodies can be bivalent or bispecific. Diabodies are more fully described, for example, in European Patent No. EP 404097, WO 93 / 01161, Hudson et al., Nat. Med. 9:129-134 (2003), and Hollinger et al., Proc. Natl. Acad. Sci. USA 90: 6444-6448 (1993). Triabodies and tetrabodies are also described in Hudson et al., Nat. Med. 9:129-134 (2003).

[0111] A bispecific scDb (especially a bispecific monomeric scDb) particularly comprises two variable heavy chain domains (VH) or fragments thereof and two variable light chain domains (VL) or fragments thereof connected by linkers L1, L2, and L3, in the following order: VHA-L1-VLB-L2-VHB-L3-VLA, VHA-L1-VHB-L2-VLB-L3-VLA, VLA-L1-VLB-L2-VHB-L3-VHA, VLA-L1-VHB-L2 -VLB-L3-VHA, VHB-L1-VLA-L2-VHA-L3-VLB, VHB-L1-VHA-L2-VLA-L3-VLB, VLB-L1-VLA-L2-VHA-L3-VHB, or VLB-L1-VHA-L2-VLA-L3-VHB, in which the VLA and VHA domains combine to form an antigen-binding site for a first antigen and the VLB and VHB combine to form an antigen-binding site for a second antigen.

[0112] Linker L1 is particularly a peptide of 2 to 10 amino acids, more particularly 3 to 7 amino acids, and most particularly 5 amino acids, and linker L3 is particularly a peptide of 1 to 10 amino acids, more particularly 2 to 7 amino acids, and most particularly 5 amino acids. In particular embodiments, linker L1 and / or L3 comprises one or two units of four (4) glycine amino acid residues and one (1) serine amino acid residue (GGGGS). n (But n=1 or 2, especially n=1).

[0113] The central linker L2 is particularly a peptide of 10 to 40 amino acids, more particularly 15 to 30 amino acids, and most particularly 20 to 25 amino acids. In a particular embodiment, said linker L2 comprises one or more units of four (4) glycine amino acid residues and one (1) serine amino acid residue (GGGGS). n (where n=1, 2, 3, 4, 5, 6, 7, or 8, and especially n=4).

[0114] In one embodiment, the multispecific antibody of the invention is an scDb-scFv. The term "scDb-scFv" refers to an antibody format in which a single-chain Fv (scFv) fragment is fused to a single-chain diabody (scDb) via a flexible Gly-Ser linker. In one embodiment, the flexible Gly-Ser linker is a peptide consisting of 2 to 40 amino acids (e.g., 2 to 35, 2 to 30, 2 to 25, 2 to 20, 2 to 15, or 2 to 10 amino acids, particularly 10 amino acids). In a particular embodiment, the linker comprises one or more units consisting of four (4) glycine amino acid residues and one (1) serine amino acid residue (GGGGS). n (where n=1, 2, 3, 4, 5, 6, 7, or 8, and especially n=2).

[0115] In one embodiment of the invention, the multispecific antibodies of the invention are in the MATCH format as described in WO2016 / 0202457; Egan T., et al., MABS 9 (2017) 68-84. In particular, in this embodiment, the multispecific antibodies of the invention are in the MATCH3 or MATCH4 format.

[0116] The multispecific antibodies of the present invention can be produced using any convenient antibody production method known in the art (see, for example, Fischer, N. & Leger, O., Pathobiology 74 (2007) 3-14 for the generation of bispecific constructs; Hornig, N. & Farber-Schwarz, A., Methods Mol. Biol. 907 (2012) 713-727 for bispecific diabodies and tandem scFvs, and WO99 / 57150). Specific examples of methods suitable for the preparation of bispecific constructs of the present invention are further encompassed in particular by the Genmab (see Labrijn et al., Proc. Natl. Acad. Sci. USA 110 (2013) 5145-5150) and Merus (see de Kruif et al., Biotechnol. Bioeng. 106 (2010) 741-750) technologies. Methods for generating bispecific antibodies containing functional antibody Fc portions are also well known in the art (see, e.g., Zhu et al., Cancer Lett. 86 (1994) 127-134; Suresh et al., Methods Enzymol. 121 (1986) 210-228).

[0117] These methods typically involve generating monoclonal antibodies, for example, by fusing myeloma cells with spleen cells from mice immunized with the desired antigen using hybridoma technology (see, e.g., Yokoyama et al., Curr. Protoc. Immunol. Chapter 2, Unit 2.5, 2006), or by recombinant antibody engineering (repertoire cloning, or phage display / yeast display) (see, e.g., Chames & Baty, FEMS Microbiol. Letters 189 (2000) 1-8), and combining the antigen-binding domains, or fragments or portions thereof, of two or more different monoclonal antibodies using known molecular cloning techniques to obtain bispecific or multispecific constructs.

[0118] Multispecific molecules of the invention can be prepared by conjugating the constituent binding specificities using methods known in the art. For example, each binding specificity of a bispecific molecule can be generated separately and then conjugated to one another. When the binding specificities are proteins or peptides, a variety of coupling or cross-linking agents can be used for covalent conjugation. Examples of cross-linking agents include protein A, carbodiimide, N-succinimidyl 5-acetyl-thioacetate (SATA), 5,5'-dithiobis(2-nitrobenzoic acid) (DTNB), o-phenylenedimaleimide (oPDM), N-succinimidyl 3-(2-pyridyldithio)propionate (SPDP), and sulfosuccinimidyl 4-(N-maleimidomethyl)cyclohexane-1-carboxylate (sulfo-SMCC) (see, e.g., Karpovsky et al., 1984 J. Exp. Med. 160:1686; Liu, MA et al., 1985 Proc. Natl. Acad. Sci. USA 82:8648). Other methods include those described by Paulus, 1985 Behring Ins. Mitt. No. 78, 118-132; Brennan et al., 1985 Science 229:81-83; and Glennie et al., 1987 J. Immunol. 139:2367-2375. Complexing agents are SATA and sulfo-SMCC, both available from Pierce Chemical Co. (Rockford, 111).

[0119] When the binding specificities are multiple antibodies, they can be conjugated via sulfhydryl bonding in the C-terminal hinge regions of the two heavy chains. In one particular embodiment, the hinge region is modified to contain an odd number of sulfhydryl residues (e.g., one) prior to conjugation.

[0120] Alternatively, two or more binding specificities can be encoded in the same vector and expressed and assembled in the same host cell. This method is particularly useful when the bispecific molecule is a mAb X mAb, mAb X Fab, Fab XF(ab')2, or ligand X Fab fusion protein. The multispecific antibody of the present invention can be a single-chain molecule containing one single-chain antibody and one binding determinant, or a single-chain multispecific antibody containing two binding determinants. The multispecific antibody can comprise at least two single-chain molecules. Methods for preparing multispecific antibodies and molecules are described, for example, in U.S. Patent Nos. 5,260,203; 5,455,030; 4,881,175; 5,132,405; 5,091,513; 5,476,786; 5,013,653; 5,258,498; and 5,482,858.

[0121] Binding of a multispecific antibody to its specific target can be confirmed, for example, by enzyme-linked immunosorbent assay (ELISA), radioimmunoassay (REA), FACS analysis, bioassay (e.g., growth inhibition), or Western blot assay. Each of these assays generally detects the presence of a protein-antibody complex of particular interest by employing a labeled reagent (e.g., an antibody) specific for the complex of interest.

[0122] In a further aspect, the present invention provides a nucleic acid encoding a multispecific antibody or fragment thereof or a binding domain thereof of the present invention. Such nucleic acid sequences can be optimized for expression in mammalian cells.

[0123] The term "nucleic acid" is used interchangeably herein with the term "polynucleotide" and refers to one or more deoxyribonucleotides or ribonucleotides and polymers thereof in single- or double-stranded form. The term encompasses nucleic acids containing known nucleotide analogs or modified backbone residues or linkages, which are synthetic, natural, and non-natural, and which have similar binding properties and are metabolized in a manner similar to the reference nucleic acid. Non-limiting examples of such analogs include phosphorothioates, phosphoramidates, methyl phosphonates, chiral-methyl phosphates, 2-O-methyl ribonucleotides, and peptide-nucleic acids (PNAs). Unless otherwise specified, a particular nucleic acid sequence implicitly encompasses the explicitly indicated sequence, as well as conservatively modified variants thereof (e.g., degenerate codon substitutions) and complementary sequences. Specifically, as detailed below, degenerate codon substitutions can be achieved by generating sequences in which the third position of one or more selected codons (or all codons) is substituted with mixed-base and / or deoxyinosine residues (Batzer et al., Nucleic Acid Res. 19:5081, 1991; Ohtsuka et al., J. Biol. Chem. 260:2605-2608, 1985; and Rossolini et al., Mol. Cell. Probes 8:91-98, 1994).

[0124] The present invention provides substantially purified nucleic acid molecules encoding polypeptides comprising segments or domains of the above-described multispecific antibodies, which, when expressed by an appropriate expression vector, are capable of exhibiting one or more antigen-binding capabilities of the multispecific antibodies of the invention.

[0125] The present invention also provides polynucleotides encoding at least one CDR region, and typically all three CDR regions, of the binding domains of the multispecific antibodies of the invention shown in Tables 1, 3, and 4. Due to the degeneracy of the code, multiple nucleic acid sequences will encode each of the amino acid sequences of an immunoglobulin.

[0126] Polynucleotide sequences can be generated by de novo solid-phase DNA synthesis or PCR mutagenesis of existing sequences encoding the multispecific antibodies of the invention, or fragments thereof, or binding domains thereof (e.g., the sequences described in the Examples below). Direct chemical synthesis of nucleic acids can be achieved by methods known in the art (such as the phosphotriester method of Narang et al., 1979, Meth. Enzymol. 68:90; the phosphodiester method of Brown et al., Meth. Enzymol. 68:109, 1979; the diethylphosphoramidite method of Beaucage et al., Tetra. Lett., 22:1859, 1981; and the solid-support method of U.S. Pat. No. 4,458,066). Mutations in polynucleotide sequences by PCR can be carried out, for example, as described in PCR Technology: Principles and Applications for DNA Amplification, H.A. Erlich (Ed.), Freeman Press, NY, NY, 1992; PCR Protocols: A Guide to Methods and Applications, Innis et al. (Ed.), Academic Press, San Diego, Calif., 1990; Mattila et al., Nucleic Acids Res. 19:967, 1991; and Eckert et al., PCR Methods and Applications 1:17, 1991.

[0127] The present invention also provides expression vectors and host cells for producing the multispecific antibodies of the present invention, or fragments thereof, or binding domains thereof.

[0128] The term "vector" is intended to mean a polynucleotide molecule capable of transporting another polynucleotide to which it has been linked. One type of vector is a "plasmid," which refers to a circular double-stranded DNA loop into which additional DNA segments can be ligated. Another type of vector is a viral vector, wherein additional DNA segments can be ligated into the viral genome. Some vectors are capable of autonomous replication in a host cell into which they are introduced (e.g., bacterial vectors having a bacterial origin of replication and episomal mammalian vectors). Other vectors (e.g., non-episomal mammalian vectors) can be integrated into the genome of a host cell upon introduction into the host cell, and thereby are replicated along with the host genome.

[0129] Furthermore, a vector is capable of directing the expression of genes that are operably linked to it. Such vectors are referred to herein as "recombinant expression vectors" (or simply, "expression vectors"). In general, expression vectors useful in recombinant DNA techniques are often in the form of plasmids. Because plasmids are the most commonly used form of vector, the terms "plasmid" and "vector" can be used interchangeably herein. However, the invention is intended to encompass such other forms of expression vectors, such as viral vectors (e.g., replication-defective retroviruses, adenoviruses, and adeno-associated viruses), which serve equivalent functions. In this particular context, the phrase "operably linked" refers to a functional relationship between two or more polynucleotide (e.g., DNA) segments. Typically, it refers to the functional relationship of a transcriptional regulatory sequence to a transcribed sequence. For example, a promoter or enhancer sequence is operably linked to a coding sequence if it stimulates or modulates the transcription of the coding sequence in an appropriate host cell or other expression system. Generally, a promoter transcriptional regulatory sequence operably linked to a transcribed sequence is physically contiguous with the transcribed sequence, i.e., cis-acting. However, some transcriptional regulatory sequences (such as enhancers) need not be physically contiguous or located near the coding sequences whose transcription they enhance.

[0130] A variety of expression vectors can be used to express polynucleotides encoding multispecific antibody chains or binding fragments. Both viral and non-viral expression vectors can be used to produce antibodies in mammalian host cells. Non-viral vectors and systems include plasmids, episomal vectors (typically containing expression cassettes for protein or RNA expression), and human artificial chromosomes (see, e.g., Harrington et al., Nat Genet. 15:345, 1997). For example, non-viral vectors useful for expressing MSLN-binding polynucleotides and polypeptides in mammalian (e.g., human) cells include pThioHis A, B, and C, pcDNA3.1 / His, pEBVHis A, B, and C (Invitrogen, San Diego, CA), MPS V vectors, and many other vectors known in the art for expressing other proteins. Useful viral vectors include vectors based on retroviruses, adenoviruses, adeno-associated viruses, herpes viruses, SV40, papilloma viruses, HBP Epstein-Barr virus, vaccinia virus vectors, and Semliki Forest virus (SFV) vectors. See Brent et al., supra; Smith, Annu. Rev. Microbiol. 49:807, 1995; and Rosenfeld et al., Cell 68:143, 1992.

[0131] The choice of expression vector depends on the host cell in which the vector is to be expressed. Typically, expression vectors contain a promoter and other regulatory sequences (e.g., enhancers) operably linked to the polynucleotide encoding the chains or fragments of the multispecific antibody. In one embodiment, an inducible promoter is used to prevent expression of the inserted sequence except under inducing conditions. Inducible promoters include, for example, arabinose, lacZ, metallothionein promoters, or heat shock promoters. Cultures of transformed organisms can grow a population of coding sequences in an unbiased manner under non-inducing conditions, and the expression products are better tolerated by the host cells. In addition to a promoter, other regulatory elements may be required or desired for efficient expression of the chains or fragments of the multispecific antibody. These elements typically contain an ATG initiation codon and adjacent ribosome binding site or other sequences. In addition, the efficiency of expression can be increased by including enhancers appropriate for the cell system used (see, e.g., Scharf et al., Results Probl. Cell Differ. 20:125, 1994; Bittner et al., Meth. Enzymol., 153:516, 1987). For example, the SV40 enhancer or CMV enhancer can be used to increase expression in mammalian host cells.

[0132] The expression vector may also provide the location of a secretory signal sequence for forming a fusion protein with the polypeptide encoded by the inserted sequence of the multispecific antibody of the invention, or a fragment thereof, or its binding domain. In most cases, the inserted sequence of the multispecific antibody of the invention, or a fragment thereof, or its binding domain is linked to a signal sequence and then incorporated into a vector. Sometimes, the vector used to receive the sequences encoding the binding domains of the light and heavy chain variable domains of the multispecific antibody also encodes the constant regions or parts thereof.

[0133] The term "recombinant host cell" (or simply "host cell") refers to a cell into which a recombinant expression vector has been introduced. It should be understood that such terms are intended to refer not only to the particular subject cell but also to the progeny of such a cell. Because modifications due to mutations or environmental influences may occur in successive generations, such progeny may not, in fact, be identical to the parent cell, but are still included within the scope of the term "host cell" as used herein.

[0134] Host cells for incorporating and expressing the multispecific antibodies, or fragments thereof, or binding domains thereof, of the present invention can be prokaryotic or eukaryotic. Escherichia coli (E. coli) is one prokaryotic host useful for cloning and expressing the polynucleotides of the present invention. Other suitable microbial hosts include bacilli (e.g., Bacillus subtilis) and other Enterobacteriaceae (e.g., Salmonella, Serratia, and various Pseudomonas species). Expression vectors can also be made in these prokaryotic hosts, typically containing expression control sequences compatible with the host cell (e.g., an origin of replication). Additionally, any of a number of well-known promoters will be present, such as the lactose promoter system, the tryptophan (trp) promoter system, the beta-lactamase promoter system, or promoter systems from phage lambda. The promoter typically controls expression, optionally with an operator sequence, and contains, for example, ribosome binding site sequences for initiating and completing transcription and translation. Other microorganisms, such as yeast, can also be used to express the MSLN-binding polypeptides of the present invention. Insect cells can also be used in combination with baculovirus vectors.

[0135] In one embodiment, mammalian host cells are used to express and produce the multispecific antibodies, or fragments thereof, or binding domains thereof, of the invention. For example, the host cells can be hybridoma cell lines expressing endogenous immunoglobulin genes or mammalian cell lines harboring exogenous expression vectors. Host cells include any normal, mortal, or immortal animal or human cell, normal or abnormal. For example, numerous suitable host cell lines capable of secreting intact immunoglobulins have been developed, including CHO cell lines, various Cos cell lines, HeLa cells, myeloma cell lines, transformed B cells, and hybridomas. The use of mammalian tissue cell culture to express polypeptides is generally discussed in, for example, Winnacker, FROM GENES TO CLONES, VCH Publishers, NY, NY, 1987. Expression vectors for mammalian host cells can include expression control sequences, such as an origin of replication, a promoter, and an enhancer (see, e.g., Queen, et al., Immunol. Rev. 89:49-68, 1986), and necessary information processing sites, such as ribosome binding sites, RNA splice sites, polyadenylation sites, and transcription termination sequences. These expression vectors usually contain promoters derived from mammalian genes or mammalian viruses. Suitable promoters can be constitutive, cell type-specific, stage-specific, and / or tunable or regulatable. Non-limiting examples of useful promoters include the metallothionein promoter, the constitutive adenovirus major late promoter, the dexamethasone-inducible MMTV promoter, the SV40 promoter, the MRP polIII promoter, the constitutive MPS V promoter, the tetracycline-inducible CMV promoter (such as the human immediate-early CMV promoter), the constitutive CMV promoter, and promoter-enhancer combinations known in the art.

[0136] Methods for introducing expression vectors containing polynucleotide sequences of interest vary depending on the type of cellular host. For example, calcium chloride transfection is commonly utilized in prokaryotic cells, whereas calcium phosphate treatment or electroporation can be used in other cellular hosts. (See generally Sambrook, et al., supra.) Other methods include, for example, electroporation, calcium phosphate treatment, liposome-mediated transformation, injection and microinjection, ballistic techniques, virosomes, immunoliposomes, polycationic nucleic acid complexes, naked DNA, artificial virions, fusion to the herpesvirus structural protein VP22 (Elliot and O'Hare, Cell 88:223, 1997), drug-enhanced DNA uptake, and in vitro transduction. For long-term, high-yield production of recombinant proteins, stable expression will often be desirable. For example, cell lines stably expressing the multispecific antibodies, or fragments thereof, or binding domains thereof of the present invention can be prepared using expression vectors of the present invention containing viral origins of replication or endogenous expression elements and a selectable marker gene. After the introduction of the vector, the cells are allowed to grow in an enriched medium for 1-2 days and then switched to a selective medium. The purpose of the selectable marker is to confer resistance to selection; its presence allows the growth of cells capable of expressing the introduced sequences in a selective medium. Resistant, stably transfected cells can be grown using tissue culture techniques appropriate to the cell type. Thus, the present invention provides a method for producing an antibody or antigen-binding fragment thereof of the present invention, comprising the step of culturing a host cell containing a nucleic acid or vector encoding an antibody or antigen-binding fragment thereof of the present invention, thereby expressing the antibody or antigen-binding fragment thereof of the present disclosure.

[0137] In one aspect, the present invention relates to a method for producing a multispecific antibody of the present invention, or a binding domain thereof, or a fragment thereof, the method comprising the step of culturing a host cell expressing a nucleic acid encoding the multispecific antibody of the present invention, or a binding domain thereof, or a fragment thereof. In particular, the present invention relates to a method for producing a multispecific antibody of the present invention, or a binding domain thereof, or a fragment thereof, the method comprising: (i) providing one or two nucleic acid sequences encoding the multispecific antibody of the present invention, or a binding domain thereof, or one or two vectors encoding the multispecific antibody of the present invention, expressing the one or two nucleic acid sequences or the one or two vectors and recovering the multispecific antibody or the binding domain from the expression system, or (ii) providing a host cell or multiple host cells expressing a nucleic acid encoding the multispecific antibody of the present invention, or a binding domain thereof, and culturing the one or multiple host cells; and recovering the multispecific antibody or the binding domain from the cell culture.

[0138] In a further aspect, the present invention relates to a pharmaceutical composition comprising a multispecific antibody of the present invention and a pharmaceutically acceptable carrier. The pharmaceutically acceptable carrier enhances or stabilizes the composition or facilitates its preparation. Pharmaceutically acceptable carriers include physiologically compatible solvents, dispersion media, coatings, antibacterial and antifungal agents, isotonic and absorption delaying agents, and the like.

[0139] The pharmaceutical compositions of the present invention can be administered by a variety of methods known in the art. The route and / or mode of administration vary depending on the desired results. Administration can be intravenous, intramuscular, intraperitoneal, or subcutaneous, or proximal to the target site. The pharmaceutically acceptable carrier should be suitable for intravenous, intramuscular, subcutaneous, parenteral, spinal, or epidermal administration (e.g., by injection or infusion). Depending on the route of administration, the active compound (i.e., the multispecific antibody of the present invention) can be coated in a material to protect it from the action of acids and other natural conditions that may inactivate the compound.

[0140] Pharmaceutical compositions of the present invention can be prepared according to methods well known and routinely practiced in the art. See, for example, Remington: The Science and Practice of Pharmacy, Mack Publishing Co., 20th ed., 2000; and Sustained and Controlled Release Drug Delivery Systems, J.R. Robinson, ed., Marcel Dekker, Inc., New York, 1978. Pharmaceutical compositions are preferably manufactured under GMP conditions. Typically, a therapeutically effective dose or effective dosage of a multispecific antibody of the present invention is used in the pharmaceutical compositions of the present invention. The multispecific antibody of the present invention can be formulated into a pharmaceutically acceptable dosage form by conventional methods known to those of skill in the art. Dosage regimens can be adjusted to provide the optimal desired response (e.g., a therapeutic response). For example, a single bolus can be administered, several divided doses can be administered over time, or the dose can be proportionally reduced or increased as indicated by the exigencies of the therapeutic situation. Parenteral compositions are particularly advantageously formulated in unit dosage form for ease of administration and uniformity of dosage. Dosage unit form, as used herein, means physically discrete units suited as unitary dosages for subjects to be treated; each unit containing a predetermined quantity of active compound calculated to produce the desired therapeutic effect in combination with the required pharmaceutical carrier.

[0141] The actual dosage level of the active ingredient in the pharmaceutical compositions of the present invention can be varied so that, for a particular patient, composition, and mode of administration, an amount of the active ingredient is effective to achieve the desired therapeutic response without causing toxicity to the patient. The selected dosage level will depend on a variety of pharmacokinetic factors, including the activity of the particular composition of the present invention, or its ester, salt, or amide, employed, the route of administration, the time of administration, the rate of excretion of the particular compound employed, the duration of treatment, other drugs, compounds and / or materials used in combination with the particular composition employed, and the age, sex, weight, general health, and previous medical history of the patient being treated.

[0142] Multispecific antibodies of the invention are typically administered on multiple occasions. The intervals between single doses can be weekly, monthly, or yearly. The intervals can also be irregular, depending on the values ​​indicated by measuring the patient's blood levels of the multispecific antibodies of the invention. Alternatively, the multispecific antibodies of the invention can be administered as a sustained-release formulation, which requires less frequent administration. The dosage and frequency will vary depending on the half-life of the antibody in the patient's body. Humanized antibodies generally exhibit a longer half-life than chimeric and non-human antibodies. The dosage and frequency of administration can vary depending on whether the treatment is prophylactic or therapeutic. For prophylactic applications, relatively low doses are administered over a long period of time at less frequent intervals. Some patients continue to receive treatment for the rest of their lives. For therapeutic applications, relatively high doses at relatively short intervals may sometimes be required until the progression of the disease slows or stops, preferably until the patient shows partial or complete improvement in the symptoms of the disease. Thereafter, the patient can be administered a prophylactic regimen.

[0143] In one aspect, the invention relates to a multispecific antibody of the invention or a pharmaceutical composition of the invention for use as a medicament. In a suitable embodiment, the invention provides a multispecific antibody or a pharmaceutical composition for use in treating a proliferative disease, particularly cancer, in a subject in need thereof.

[0144] In another aspect, the present invention provides a multispecific antibody or pharmaceutical composition for use in the manufacture of a medicament for the treatment of a proliferative disease, particularly cancer.

[0145] In another aspect, the present invention relates to the use of a multispecific antibody or a pharmaceutical composition for treating a proliferative disease, particularly cancer, in a subject in need thereof.

[0146] In a further aspect, the present invention relates to the use of a multispecific antibody or a pharmaceutical composition in the manufacture of a medicament for treating a proliferative disease, particularly cancer, in a subject in need thereof.

[0147] In another aspect, the invention relates to a method of treating a subject, the method comprising administering to the subject a therapeutically effective amount of a multispecific antibody of the invention. In a suitable embodiment, the invention relates to a method of treating a proliferative disease (particularly cancer) in a subject, the method comprising administering to the subject a therapeutically effective amount of a multispecific antibody of the invention.

[0148] The term "subject" includes humans and non-human animals. Non-human animals include all vertebrates, e.g., mammals and non-mammals, such as non-human primates, sheep, dogs, cows, chickens, amphibians, and reptiles. Except where noted, the terms "patient" and "subject" are used interchangeably herein.

[0149] The terms "treatment," "treating," "treat," "treated," and the like, as used herein, refer to obtaining a desired pharmacological and / or physiological effect. The effect may be therapeutic in the sense of a partial or complete cure of a disease and / or deleterious effects attributable to the disease, or a delay in disease progression. "Treatment," as used herein, covers any treatment of a disease in a mammal (e.g., a human), including (a) inhibiting the disease, i.e., arresting its development; and (b) palliating the disease, i.e., causing regression of the disease.

[0150] The phrases "therapeutically effective amount" or "effective amount" mean the amount of an agent that, when administered to a mammal or other subject for treating a disease, is sufficient to effect such treatment for the disease. The "therapeutically effective amount" will vary depending on the agent, the disease and its severity, and the age, weight, etc., of the subject being treated.

[0151] In one embodiment, the proliferative disease is cancer. The term "cancer" refers to a disease characterized by the rapid and uncontrolled growth of abnormal cells. Cancer cells can spread locally or to other parts of the body through the bloodstream and lymphatic system. The terms "tumor" and "cancer" are used interchangeably herein, e.g., both terms encompass solid and liquid, e.g., diffuse or circulating, tumors. As used herein, the term "cancer" or "tumor" includes pre-malignant cancers and tumors as well as malignant cancers and tumors. The term "cancer" is used herein to refer to a broad spectrum of tumors, including all solid and hematologic malignancies. Non-limiting examples of such tumors include benign or malignant (especially malignant) malignant tumors, solid tumors, brain tumors, kidney cancer, liver cancer, adrenal gland cancer, bladder cancer, breast cancer, gastric cancer (e.g., stomach tumors), esophageal cancer, ovarian cancer, cervical cancer, colon cancer, rectal cancer, prostate cancer, pancreatic cancer, lung cancer (e.g., non-small cell lung cancer and small cell lung cancer), vaginal cancer, thyroid cancer, melanoma (e.g., unresectable or metastatic melanoma), renal cell carcinoma, sarcoma, glioblastoma, multiple myeloma, or gastrointestinal cancer (especially colon cancer or colorectal cancer). adenoma), tumors of the head and neck, endometrial cancer, Cowden syndrome, Lhermitte-Duclos disease, Banayan-Zonana syndrome, benign prostatic hyperplasia, neoplasia (especially of epithelial character, preferably breast cancer or squamous cell carcinoma), chronic lymphocytic leukemia, chronic myeloid leukemia (e.g., Philadelphia chromosome-positive chronic myeloid leukemia), acute lymphocytic leukemia (e.g., Philadelphia chromosome-positive acute lymphocytic leukemia), non-Hodgkin's lymphoma, plasma cell myeloma, Hodgkin's lymphoma, leukemia, and any combination thereof. In a preferred embodiment, the cancer is selected from mesothelioma, pancreatic cancer, and ovarian cancer.

[0152] The multispecific antibodies of the invention or the compositions of the invention inhibit the growth of not only solid tumors but also liquid tumors. In a further embodiment, the proliferative disease is a solid tumor. The term "solid tumor" particularly refers to breast cancer, ovarian cancer, colon cancer, rectal cancer, prostate cancer, stomach cancer (especially gastric cancer), cervical cancer, lung cancer (e.g., non-small cell lung cancer and small cell lung cancer), and head and neck tumors. Furthermore, depending on the type of tumor and the specific combination used, a reduction in tumor volume can be achieved. The multispecific antibodies of the invention or the compositions of the invention are also suitable for inhibiting the spread of tumors by metastasis and the growth or development of micrometastases in cancer-bearing subjects. Sequence Listing (mutations designated according to the AHo numbering scheme, CDRs defined according to Numab CDR definitions) [Table 1-1] [Table 1-2] [Table 2] [Table 3] [Table 4-1] [Table 4-2] [Table 4-3] [Table 5-1] [Table 5-2] [Table 5-3] [Table 5-4] [Table 5-5] [Table 5-6] [Table 5-7] [Table 6] [Table 7]

[0153] Throughout the text of this application, in the event of a discrepancy between the text of the specification (e.g., Tables 1-7) and the sequence listing, the text of the specification will control.

[0154] It will be appreciated that certain features of the invention, which are, for clarity, described in the context of separate embodiments, may also be provided in combination in a single embodiment. Conversely, various features of the invention, which are, for brevity, described in the context of a single embodiment, may also be provided separately or in any suitable subcombination. All combinations of embodiments of the invention are expressly embraced by the present invention, and each and every combination is disclosed herein as if it were individually and explicitly disclosed. In addition, all subcombinations of the various embodiments and elements thereof are expressly embraced by the present invention, and each and every such subcombination is disclosed herein as if it were individually and explicitly disclosed.

[0155] The present invention is not to be limited in scope by the specific embodiments described herein. Indeed, various modifications of the invention in addition to those described herein will become apparent to those skilled in the art from the foregoing description. Such modifications are intended to fall within the scope of the appended claims.

[0156] All patents, applications, publications, test methods, literature, and other materials cited herein are incorporated by reference to the extent possible under their respective patent laws.

[0157] The following examples illustrate the above invention, but are not intended to limit the scope of the invention in any way. Other current test models known to those skilled in the art may also reveal the beneficial effects of the claimed invention. [Example]

[0158] Example 1: Generation and Pharmacodynamic Characterization of Anti-MSLN Molecules In the first step, anti-MSLN antibody fragments with intermediate to low affinity for MSLN should be identified. These fragments should be suitable for use in multispecific antibody formats, particularly the MATCH3 and MATCH4 antibody formats.

[0159] Identification, selection, and generation of anti-MSLN binding domains of the present invention The humanized scFv anti-MSLN binding domains of the present invention were identified, selected, humanized, and produced in a manner similar to the scFv anti-CD3 binding domains described in patent application PCT / EP2018 / 064630 (incorporated herein by reference).

[0160] From some of the identified monoclonal antibodies with the desired properties (particularly the desired affinity), scFv molecules were generated according to the following procedure.

[0161] Humanization and expression: Rabbit antibodies were humanized by CDR grafting and, optionally, specific rabbit framework residues onto a lambda-capped Vk1 / VH3 Fv scaffold. Each scFv was designed in the following orientation: N-terminus-VL-peptide linker-VH-C-terminus (peptide linker: (G4S)4).

[0162] Recombinant amino acid sequences were synthesized de novo, and scFv constructs were expressed in CHO-S cells using the CHOgro transient transfection kit (Mirus). Cultures were harvested by centrifugation at 37°C after 5–7 days of expression (cell viability <70%), and proteins were purified from clarified culture supernatants by protein L or A affinity chromatography, followed, if necessary, by a final purification step by size-exclusion chromatography (SEC) using a Superdex S200 column.

[0163] Quality control: Standard analytical methods (such as SE-HPLC, UV280, and SDS-PAGE) were utilized to quality control the produced materials.

[0164] SE-HPLC Samples for SE-HPLC analysis were passed through either a Shodex™ (Showa Denko, Catalog No. 554-1740) KW402.5-4F column (for scFv analysis) or a Shodex™ (Showa Denko, Catalog No. 554-1741) KW403-4F column (for MATCH protein analysis) with running buffer (Shodex™ KW402.5-4F: 250 mM NaCl, 50 mM NaOAc (Cat. No. A 1045), pH 6.0; Shodex™ KW403-4F: 35 mM NaH2PO4 (Cat. No. A 3905), 15 mM Na2HPO4 (Cat. No. A1372), 300 mM NaCl, pH 6.0) at a flow rate of 0.35 mL / min. Eluted proteins were detected by absorbance at λ=280 nm.

[0165] SDS-PAGE Protein identity and degradation were assessed by SDS-PAGE analysis; denatured proteins were loaded onto MiniPROTEAN TGX™ precast gels (Bio-Rad Laboratories, catalog no. 4569036) and electrophoresed proteins were stained with Coomassie Brilliant Blue solution. Molecular weight standards: BioRad Precision™ Plus (catalog no. 161-03 / 04).

[0166] The production data for the resulting scFv molecules are summarized in Table 8.

[0167] Pharmacodynamic characterization of anti-mesothelin scFv antibodies PRO1783, PRO1925, PRO2306, PRO2309 (low affinity), and PRO1922 The key pharmacodynamic properties of the humanized anti-mesothelin scFv antibody PRO1783 were evaluated (including binding kinetics and affinity to recombinant human MSLN and recombinant cynomolgus monkey MSLN by SPR, binding to the plasma membrane of human MSLN-expressing cell lines and cynomolgus monkey MSLN-expressing cell lines by cELISA, and inhibition of MSLN / MUC16 by cELISA). In addition, the key pharmacodynamic properties of the humanized anti-MSLN scFvs PRO1922, PRO1925, PRO2306, and PRO2309 were evaluated (including binding kinetics and affinity to recombinant human MSLN by SPR, binding to the plasma membrane of human MSLN-expressing cell lines by cELISA, and inhibition of MSLN / MUC16 by cELISA (PRO1922 and PRO1925)). The results are summarized in Tables 9–13.

[0168] Affinity of SPR to human MSLN and cynomolgus monkey MSLN The affinity of scFv PRO1783 (derived from monoclonal antibody 54-01-G02) for recombinant human MSLN and recombinant cynomolgus monkey MSLN was determined by SPR analysis on a T200 instrument (Biacore, GE Healthcare). In this experiment, recombinant human MSLN and recombinant cynomolgus monkey MSLN (purchased from Peprotech and Sino Biological, respectively) were immobilized on the surface of different flow cells of a CM5 sensor chip using a standard amine coupling procedure. The scFv antibody PRO1783 was then injected into the flow cell at concentrations ranging from 90 to 0.12 nM for 5 minutes, and the proteins were allowed to dissociate for 12 minutes. Dissociation (k d ) and meeting (k a ) and the equilibrium dissociation constant (K D The affinity of scFv PRO1922 and PRO1925 was evaluated by SPR as described above, but over a concentration range of 15–0.12 nM. The affinity of scFv PRO2306 and PRO2309 was evaluated by SPR as described above, but over a concentration range of 90–0.35 nM.

[0169] As shown in Table 9, PRO1783 binds recombinant human MSLN with low nanomolar affinity (K D =2.91 nM). In SPR, PRO1922, PRO1925, PRO2306, and PRO2309 bound to recombinant human MSLN with significant affinities in the subnanomolar range. SPR measurements also demonstrated binding of PRO1783 to recombinant cynomolgus monkey MSLN, but with reduced affinity (K D =30.06nM, Table 10).

[0170] Binding to MSLN-expressing cell lines by cELISA Binding to cells (H226 cell line) expressing high levels of human MSLN The binding of the anti-MSLN scFv antibody PRO1783 to MSLN on the plasma membrane was assessed by cELISA in H226 cancer cells. Briefly, 20,000 NCI-H226 cells expressing MSLN or HEK293T (MSLN-negative) were dispensed into a 96-well flat-bottom tissue culture plate. The next day, the plate was washed three times with 450 μl of wash buffer (PBS, 0.2% BSA) per well in overflow mode. 50 μl of serial dilutions of PRO1783 and the anti-MSLN reference antibody amatuximab were added and incubated for 1.5 hours at room temperature (RT) with gentle agitation. After washing three times with 450 μl of wash buffer, 50 μl of HRP-coupled Protein L or HRP-coupled anti-human IgG antibody was added to each well. After 1 hour of incubation at room temperature on a nutation mixer, the plate was washed three times with 450 μl of wash buffer per well before adding 50 μl of TMB (3,3',5,5'-tetramethylbenzidine, KPL). After 10 minutes of color development, the enzyme reaction was stopped by adding 50 μl of 1 M HCl per well, and the plate was read at 450 nm with 690 nm as the reference wavelength.

[0171] The results of experiments assessing the binding of PRO1783 to the plasma membrane of the H226 cell line, which expresses high levels of MSLN, are shown in Table 11. EC 50 was found to be at a concentration of 1.44 nM, which is roughly 6-fold worse (relative EC 50 Compare values, Table 11). EC for binding of PRO1925, PRO2306, and PRO2309 to H226 cells 50 The EC values ​​for PRO1922 binding to H226 cells were found to be similar to those for PRO1783. 50The values ​​were found to be similar to those of amatuximab. When mesothelin-negative HEK293T cells were tested by cELISA, no binding of amatuximab, PRO1783, PRO1922, PRO1925, PRO2306, or PRO2309 was detected (data not shown). The concentration-response curves of PRO1783, PRO1922, PRO1925, PRO2306, PRO2309, and amatuximab in cELISA using the H226 cell line are shown in Figure 1.

[0172] Binding to cells expressing cynomolgus monkey MSLN (CHO recombinant cell line) The cross-reactivity of the anti-MSLN scFv antibody PRO1783 with cynomolgus monkey MSLN was examined in a cELISA using a recombinant CHO cell line expressing cynomolgus monkey MSLN. Twenty thousand cynomolgus monkey MSLN-expressing CHO cells or CHO-K1 cells (cynomolgus monkey MSLN-negative) were dispensed into a 96-well flat-bottom tissue culture plate. The next day, the plate was washed, and serial dilutions of PRO1783 and the anti-MSLN reference antibody amatuximab were added as described in the cELISA protocol using the H226 cell line. After 1.5 hours of incubation at room temperature with gentle agitation, the plate was washed again, and binding of PRO1783 and amatuximab, respectively, was detected by adding HRP-coupled protein L or HRP-coupled anti-human IgG antibody. After 1 hour of incubation at room temperature on a nutation mixer, the plate was washed, and TMB was added to each well. After 10 minutes of color development, the enzymatic reaction was stopped by adding 50 μl per well of 1 M HCl and the plates were read at 450 nm with 690 nm as the reference wavelength.

[0173] The results of the cELISA using a CHO cell line expressing cynomolgus monkey MSLN are shown in Table 12. EC of PRO1783 on binding of cynomolgus monkey MSLN to the plasma membrane 50 was found to be at a concentration of 12 nM, which is comparable to the reference antibody amatuximab (relative EC 50= 0.03). On the other hand, when compared with the binding of human MSLN to the plasma membrane, an increase in the half-maximal binding concentration of PRO1783 is evident. This is consistent with the results of SPR analysis, demonstrating the decreased affinity of PRO1783 for recombinant cynomolgus monkey MSLN protein. The concentration-response curves of PRO1783 and amatuximab in cELISA using a CHO cell line expressing cynomolgus monkey MSLN are shown in Figure 2. When CHO-K1 wild-type cells were examined by cELISA, neither amatuximab nor PRO1783 binding was detected (data not shown).

[0174] Neutralization of MSLN / MUC16 interaction by competitive ELISA The efficacy of the anti-MSLN scFv antibodies PRO1783, PRO1922, and PRO1925 in blocking the MSLN / MUC16 interaction was evaluated by competitive ELISA. ELISA plates were coated by adding 50 μl of PBS containing 1 μg / ml MUC16 overnight at 4°C. The next day, the plates were washed three times with 450 μl of wash buffer per well in overflow mode, and 300 μl of blocking buffer was added to each well for 1 hour at room temperature on a nutation mixer. Biotinylated MSLN was then diluted in blocking buffer to reach a final concentration of 1 ng / ml. Next, PRO1783, PRO1922, PRO1925, and amatuximab were titrated into the blocking buffer containing biotinylated MSLN and incubated for 1 hour at room temperature on a nutation mixer. The ELISA plate was washed three times with 450 μl of wash buffer per well in overflow mode, and 50 μl of each concentration of PRO1783, PRO1922, PRO1925, and amatuximab from the titration curve was added to the ELISA plate in duplicate. The plate was incubated at room temperature for 1.5 hours with gentle agitation. After washing three times with 450 μl of wash buffer per well, 50 μl of 10 ng / ml streptavidin-polyHRP40 was added to each well of the ELISA plate. After incubation at room temperature for 1 hour, the plate was washed three times with 450 μl of wash buffer and developed for 5–10 minutes after the addition of 50 μl of TMB. Finally, the enzyme reaction was stopped by adding 50 μl of 1 M HCl, and the plate was read at 450 nm with a reference wavelength of 690 nm.

[0175] The results of the competitive ELISA are shown in Table 13. IC of blocking human MSLN / MUC16 interaction by PRO1783 50 was found to be at a concentration of 0.5 nM, which is the relative IC 50As shown by the IC values, PRO1783 is less potent than the reference antibody amatuximab in neutralizing the human MSLN / MUC16 interaction. PRO1922 and PRO1925 have significantly lower IC values ​​than PRO1783. 50 These concentrations are believed to be sufficient to block the human MSLN / MUC16 interaction. The concentration-response curves of PRO1783, PRO1922, PRO1925, and amatuximab in competitive ELISA are shown in Figure 3.

[0176] Generation and pharmacodynamic characterization of the reference anti-MSLN molecule PRO1795 The anti-MSLN binding domain PRO1795, which has high affinity for MSLN, is used as the reference binding domain.

[0177] The humanized reference anti-MSLN binding domain PRO1795 was identified, selected, humanized, and generated in a manner similar to the anti-MSLN binding domains of the invention and the anti-CD3 molecules described herein.

[0178] We also evaluated the key pharmacodynamic properties of PRO1795, including measuring binding kinetics and affinity to recombinant human and cynomolgus MSLN in SPR, assessing binding to the plasma membrane of human and cynomolgus MSLN-expressing cell lines in cELISA, and assessing inhibition of MSLN / MUC16 in cELISA. The results are summarized in Tables 9-13. [Table 8] [Table 9] [Table 10] [Table 11] [Table 12] [Table 13]

[0179] Epitope mapping of anti-MSLN rabbit IgG clone 54-01-G02 (predecessor clone of low affinity anti-MSLN scFv domain PRO1783): Binding to human / mouse MLSN variants by cELISA To precisely identify the binding region of the selected anti-MSLN rabbit IgG, we assessed the binding level of seven human / mouse variants (V5-tagged) of the extracellular domain (ECD) of MSLN to transiently transfected HEK293T cells by cELISA (Figure 4). Flat-bottom poly-D-lysine-treated 96-well plates were coated with 25,000 cells per well. The following day, the cells were transfected with the corresponding constructs and incubated at 37°C and 5% CO2. After 24 h, the cells were washed with 450 μl of wash buffer (PBS, 0.2% BSA), and samples (250 ng / ml rIgG or serial dilutions of anti-V5 tag antibodies) were added for 1.5 h at room temperature (RT) with gentle agitation. After washing three times with 450 μl of wash buffer, 50 μl of HRP-coupled goat IgG or rabbit IgG antibodies was added to each well. After 1 hour of incubation at room temperature on a nutating mixer, the plate was washed three times with 450 μl of wash buffer per well, followed by the addition of 50 μl of TMB (3,3',5,5'-tetramethylbenzidine, KPL, Catalog No. 53-00-00). After 10 minutes of color development, the enzyme reaction was stopped by adding 50 μl of 1 M HCl per well, and the plate was read at 450 nm using 690 nm as the reference wavelength. Binding levels were calculated relative to the anti-V5 antibody. A clear decrease in rIgG binding levels to specific variants compared to the reference antibody (anti-V5 tag) is considered to indicate the localization of the rIgG epitope within the segment of human MSLN substituted with the respective mouse sequence.

[0180] When anti-MSLN rabbit IgG clone 54-01-G02 (the precursor clone of the low-affinity anti-MSLN scFv domain PRO1783) was tested by cELISA, reduced binding to the chimeric human / mouse variant V1 (the most distal region of the ECD of human MSLN) was observed (a 65% reduction in binding compared to the V5 reference antibody; Table 14), whereas binding of 54-01-G02 to all other variants was greater than 90%. These data suggest that the V1 region of human MSLN is the critical region for 54-01-G02 binding. However, other regions of the ECD of human MSLN are also involved in 54-01-G02 binding, as 54-01-G02 can still bind substantially to human MSLN even in the absence of the V1 region (Table 14). For rabbit IgG 5422-H03 (the precursor clone of the high-affinity anti-MSLN scFv domain PRO1795), it appears that it is not possible to identify the binding region using a chimeric human / mouse variant of the ECD of human MSLN in cELISA (data not shown). [Table 14]

[0181] Example 2: Generation and testing of anti-CD3 molecules The humanized anti-CD3 binding domain 28-21-D09 sc04 was identified, selected, humanized, and also produced and characterized as described in patent application PCT / EP2018 / 064630, which is incorporated herein by reference.

[0182] Example 3: Generation and testing of anti-hSA molecules The humanized anti-hS binding domains 19-01-H04-sc03 and 23-13-A01-sc03 were identified, selected, humanized, and produced and characterized as described in patent application EP19206959.9 (incorporated herein by reference). The humanized anti-hS binding domain 19-04-A10-sc02 (PRO2155) was identified, selected, humanized, and produced and characterized in a similar manner to the procedures described in patent application EP19206959.9. The characterization of the anti-hSA scFv PRO2155 is briefly outlined below.

[0183] Characterization of anti-hSA scFv 19-04-A10-sc02 (PRO2155) Binding affinity and species cross-reactivity The binding kinetics (including affinity) of the selected domain 19-04-A10-sc02 to human serum albumin (hSA, Sigma-Aldrich A3782) was determined by SPR analysis at both pH 7.4 and pH 5.5 on a T200 instrument (Biacore, Cytiva). hSA molecules were covalently immobilized on a carboxymethylated dextran surface (CM5 sensor chip, Biacore, Cytiva), and a titration series of each scFv molecule was injected as analyte. After each analyte injection cycle, all flow channels on the sensor chip were regenerated (glycine pH 2.0), and a new concentration of scFv molecule was injected. The binding kinetics to hSA at 11 concentrations ranging from 0.044 to 45 nM (1:2) in the relevant running buffer (PBS 0.05% Tween®-20, or PBS 0.05% Tween®-20, pH 5.5) was measured using a multi-cycle kinetics assay. Apparent dissociation (k d ) and meeting (k a ) and the apparent dissociation equilibrium constant (K D ) was calculated using Biacore analysis software (Biacore evaluation software version 3.2, Cytiva) using a 1:1 Langmuir binding model, and the quality of the fit was monitored based on relative chi. Binding levels were calculated as the most stable binding achieved normalized to the theoretical Rmax value.

[0184] The binding kinetics of selected scFvs were also determined for cynomolgus monkey serum albumin (cSA, Molecular Innovations CYSA) and mouse serum albumin (mSA, Sigma-Aldrich A3559) as described above, except that cSA or mSA was used instead of hSA. The binding kinetics for hSA, mSA, and cSA at pH 5.5 and pH 7.4 are summarized in Table 15. [Table 15]

[0185] Biophysical characterization A 4-week stability study was performed on HSA-domains 19-04-A10-sc02 (PRO2155) and 19-04-A10-sc06 (sc02 domain with VL-VH disulfide, VL-T141C / VH-G51C, AHo numbering; PRO2317). The scFv was formulated at 10 mg / ml in aqueous buffer (50 mM NaCiP, 150 mM NaCl, pH 6.4) and stored at temperatures below -80°C, 4°C, and 40°C for 4 weeks. The proportion of monomer and oligomer in the formulation was assessed by integration of SE-HPLC peak areas at different time points throughout the study. Table 16 summarizes the monomer content (in %) and % monomer loss relative to d0. The change in protein concentration was monitored throughout the study by UV-visible measurement at 280 nm. No notable loss of protein content was observed in any sample compared to d0, so the data are not shown. Thermal stability was analyzed by nDSF (NanoTemper) to determine the onset of unfolding (T onset) and midpoint of unfolding (T m ). DSF results are shown in Table 16. [Table 16]

[0186] Example 4: Multispecific constructs of the present invention (biMSLN 低親和性Generation and Pharmacodynamic Characterization of CD3 x CD3 x hSA Constructs: Molecular Architecture MATCH is a format invented by Numab that consists of only variable domains connected by different linkers that allow the specific pairing of only matching domain pairs (Egan TJ et al., A novel multispecific heterodimeric antibody format enabling modular assembly of variable domain fragments. MABS 9 (2017) 68-84). This format is particularly well suited for easily screening different combinations of antigen-binding domains for optimal cooperativity. MATCH can be recombinantly expressed from mammalian cells. Purification can be performed using conventional affinity chromatography steps.

[0187] The architecture of the MATCH molecule is shown in Figure 5. The MATCH4 format requires that the dimeric subunits consist of a core of two split variable domain pairs, each with either two tandemly arranged VL or two VH domains, which drive heterodimerization of the two protein chains. The tandem variable domains that dimerize on each MATCH4 chain are arranged in an antiparallel N-terminal-C-terminal orientation with their counterparts. Both chains are co-expressed in mammalian cells to form fully functional tetraspecific molecules. A traditional Gly-Ser linker between the variable domains was used to connect them, as shown in Figure 5. Different linker lengths are typically used in MATCH molecules (see the sequence listing for MATCH molecules in Table 5). Furthermore, the antiparallel MATCH4 format facilitates the introduction of disulfide bridges into one of the core domains, as shown in Figure 5. The corresponding MATCH3 format (not shown) is similarly constructed and organized, except that only one scFv binding domain is attached to the core of two split variable domain pairs instead of two scFv binding domains as in MATCH4.

[0188] Like the MATCH4 and MATCH3 formats, the scMATCH3 format consists of only the variable domains connected by a different linker, as shown in Figure 5 (right). However, in this format, the split variable domains are located on a single peptide chain (sc) and assembled into a fully functional trispecific molecule, as shown in Figure 5 (right). Like the MATCH4 and MATCH3 formats, the scMATCH3 molecule can also be recombinantly expressed in mammalian cells and purified using conventional affinity chromatography steps.

[0189] By combining two to three humanized rabbit antibodies on a λ-capped Fv scaffold, antiparallel quadrispecific MATCH4 molecules according to the invention with only one high affinity MSLN-BD and a reference trispecific scMATCH3 molecule were designed as summarized in Table 17.

[0190] manufacturing MATCH constructs were expressed in CHO-S cells using the CHOgro transient transfection kit (Mirus). Cultures were harvested by centrifugation at 37°C after 5–7 days of expression (cell viability less than 70%), and proteins were purified from the clarified culture supernatants by protein L or A affinity chromatography, followed, if necessary, by a final purification step by size-exclusion chromatography (SEC) using a Superdex S200 column in 50 mM phosphate-citrate buffer, pH 6.5, and 300 mM sucrose. The monomer content of SEC fractions was assessed by SE-HPLC analysis, and fractions with a monomer content greater than 95% were pooled. For quality control of the produced material, standard analytical methods (SE-HPLC, UV) were used. 280 , and SDS-PAGE, etc.) were used.

[0191] Details regarding the preparation of the resulting molecules are summarized in Table 18.

[0192] Pharmacodynamic characterization of multispecific antibodies The following section describes the characterization of representative multispecific molecules that are monovalent or bivalent for human MSLN, monovalent for human CD3ε, and monovalent for human serum albumin (hSA). The bivalent anti-MSLN antibodies PRO2000, PRO2100, PRO2562, PRO2566, PRO2567, and PRO2660 (i.e., biMSLNxCD3xhSA) and the monovalent anti-MSLN antibody PRO1872 (i.e., MSLNxCD3xhSA) are shown. 低KD xCD3xhSA) was investigated by SPR, and its binding kinetics and affinity to recombinant human MSLN and human CD3ε were assessed by SPR.

[0193] Affinity for human MSLN in SPR The affinity of the multispecific anti-MSLN antibodies to recombinant human MSLN was determined by SPR analysis on a T200 instrument (Biacore, GE Healthcare). In this experiment, recombinant human MSLN (purchased from Peprotech) was immobilized on the surface of a CM5 sensor chip as described above. The multispecific antibodies were injected into the flow cell, and the binding kinetics as well as the equilibrium dissociation constant (K D ) was calculated as above.

[0194] The affinity of multispecific anti-MSLN antibodies to recombinant human MSLN in the absence of avidity was determined by SPR analysis on a T200 instrument (Biacore, GE Healthcare). In this experiment, a dedicated anti-framework rabbit IgG (PRO2679) was immobilized on the surface of a CM5 sensor chip as described above. The multispecific antibodies were captured on each flow cell by a 20-second injection. Recombinant human MSLN (purchased from Peprotech) was then injected into the flow cell, and the binding kinetics as well as the equilibrium dissociation constant (K D ) was calculated as above.

[0195] As shown in Table 19, the monovalent anti-MSLN antibody PRO1872 exhibits an affinity (K) in the subnanomolar range for recombinant human MSLN in SPR. D= 0.187 nM). Similar affinity to human MSLN was found for the corresponding anti-MSLN scFv antibody PRO1795 (K D = 0.321 nM, domain 54-22-H03-sc01, data not shown). The bivalent anti-MSLN antibody PRO2000 exhibited activity in the low nanomolar range (K D = 1.06 nM), which was similar to the corresponding anti-MSLN scFv antibody PRO1783 (K D = 2.91 nM, which is superior to the affinity found for domain 54-01-G02-sc01 (Table 9). PRO2562, PRO2566, and PRO2567 exhibited very similar affinities to human MSLN in this assay, with K D The values ​​are in the small nM range between 1.39 and 1.52 nM.

[0196] Affinity to human CD3ε in SPR The affinity of the multispecific anti-MSLN antibody for recombinant human CD3ε was determined by SPR analysis on a T200 instrument (Biacore, GE Healthcare). In this experiment, human recombinant CD3ε protein (Sino Biological) was immobilized on the surface of a CM5 sensor chip (GE healthcare) by amine coupling. Serial dilutions of the anti-MSLN multispecific antibody in HBS-T+ buffer (10 mM HEPES, 150 mM NaCl, and 0.05% Tween® 20, pH 7.4) were injected into the flow cell at a flow rate of 30 μl / min for 5 min. The antibody was allowed to dissociate from CD3ε on the CM5 chip for 12 min. After each injection cycle, the surface was regenerated with one injection of 10 mM glycine HCl, pH 2. The apparent dissociation (k d ) and meeting (k a ) and the apparent dissociation equilibrium constant (K D) was calculated using Biacore analysis software (BIAevaluation, GE Healthcare) using a 1:1 Langmuir binding model, and the quality of the fit was monitored based on the chi-square and U values ​​(one indicator of the quality of the curve fit). Because the fit using the 1:1 Langmuir binding model showed suboptimal quality of curve fit, a two-state reaction model was used to calculate the K D This model describes the 1:1 binding of the analyte to the immobilized ligand followed by a conformational change that stabilizes the complex.

[0197] As shown in Table 20, the anti-MSLN antibodies PRO1872 and PRO2000 are both monovalent for CD3ε, have the same anti-CD3 domain (28-21-D09-sc04), and show similar affinities in the nanomolar range for recombinant human CD3ε in SPR (PRO1872, K D = 12.1 nM; PRO2000, K D = 20.0 nM). PRO2562, PRO2566, PRO2567, and PRO2660 showed somewhat better affinity for recombinant human CD3ε, with K D The values ​​are in the small nM range between 2.97 and 6.45 nM.

[0198] Affinity for human serum albumin at pH 5.5 in SPR The binding kinetics of human serum albumin (hSA, Sigma Aldrich, catalog A3782) was evaluated by SPR on a T200 instrument (Biacore, GE Healthcare). HSA was immobilized on the surface of a sensor chip (CM5 sensor chip, GE healthcare) by amine coupling. Serial dilutions of anti-MSLN polyspecific antibodies ranging from 0.7 to 180 nM in a pH 5.5 running buffer (PBS-Tween® 20) were injected into the flow cell for 5 min. The dissociation time was set to 12 min. The apparent dissociation (k d ) and meeting (k a ) and the apparent dissociation equilibrium constant (K D) was calculated using the 1:1 Langmuir binding model as described above.

[0199] As shown in Table 21, the anti-MSLN antibody PRO1872 (anti-hSA domain: 23-13-A01-sc02) exhibited an affinity (K) in the subnanomolar range for recombinant hSA in SPR. D = 0.175 nM). PRO2562, PRO2566, PRO2567, and PRO2660 showed somewhat lower affinity for recombinant hSA, with a K D The values ​​are in the small nM range between 5.71 and 8.80 nM. [Table 17] [Table 18] [Table 19] [Table 20] [Table 21]

[0200] Biophysical characterization of representative MATCH4 molecules Storage stability and melting point by nDSF A 28-day stability study was performed on the MATCH4 molecule, where the molecule was formulated at 1 mg / mL in aqueous buffer (50 mM phosphate-citrate buffer at pH 6.5 and 300 mM sucrose) and stored at <-80°C, 4°C, and 40°C for 28 days. The proportion of monomer and oligomer in the formulation was assessed by integration of SE-HPLC peak areas at different time points throughout the study. Table 22 summarizes the monomer content (in %) and % monomer loss compared to day 0. The change in protein concentration throughout the study was monitored by UV-visible measurements at 280 nm and is shown in Table 23. Thermal stability was analyzed by nDSF (NanoTemper) to determine the onset of unfolding (T オンセット ), the midpoint of unfolding (T m ), and scattering onset temperature were determined. T m The results are shown in Table 24, including standard deviations (SD). All four MATCH4 molecules exhibit good stability profiles, showing only minor loss of monomer or protein content after 28 days of incubation. There are no notable changes in monomer content at temperatures of -80°C and 4°C, or when the 28-day / -80°C samples were subjected to repeated freeze-thawing (5x) prior to SE-HPLC / UV measurements. [Table 22] [Table 23] [Table 24]

[0201] Example 5: Measuring mesothelin density on the cell surface of cell lines: Introduction One goal was to compare the ability of multispecific molecules that were monovalent or bivalent with respect to mesothelin binding to target cell lines that displayed different levels of mesothelin on the cell surface. Therefore, mesothelin expression on the plasma membrane was quantified in different cell lines.

[0202] method The antibody binding capacity (ABC) of cancer cell lines expressing various levels of mesothelin and healthy mesothelial tissue was assessed by flow cytometry (FC) using the Quantum Simply Cellular anti-human IgG kit (Bangs Laboratories). Briefly, 1 mg of anti-mesothelin antibody (7D9.3, Genentech) was conjugated to Alexa Fluor 488 using the Lightning-Link Rapid Conjugation Kit (Expedeon) according to the manufacturer's instructions. Receptor density values ​​are reported as antibody binding capacity (ABC). ABC values ​​were derived from a standard curve generated using Quantum Simply Cellular bead anti-human IgG (Bangs Laboratories, Inc.). These beads consist of four populations of microspheres, each conjugated to a different number of anti-human IgG molecules per bead. As a first step, increasing concentrations of Alexa Fluor 488-labeled anti-mesothelin antibody were tested on the bead population with the highest amount of binding sites to determine the saturating antibody concentration, which was used during quantification as described in the manufacturer's protocol. The beads and test samples were then stained with the corresponding saturating concentration of Alexa Fluor 488-labeled anti-mesothelin antibody according to the manufacturer's instructions and run on the same day using the same photomultiplier settings as the test samples. To calculate ABC values, the geometric means for the four Quantum Simply Cellular bead populations were analyzed using NovoExpress software (ACEA Biosciences). A standard curve was generated by linear regression using a QuickCal v. 2.3 Excel spreadsheet-based analysis template (Bangs Laboratories, Inc.). R-squared values ​​were typically 0.99 or greater. ABC values ​​for samples labeled with Alexa Fluor 488-anti-mesothelin antibody were interpolated from the standard curve.

[0203] result Mesothelin density on the plasma membrane of three cancer cell lines (H226, H292, and HPAC) and one cell line derived from healthy mesothelial tissue (MeT-5A; (ATCC® CRL-9444™); supplier: ATCC) was determined using Quantum Simply Cellular beads. The data obtained are shown in Table 25. H226 cells showed the highest level of expression, followed by the HPAC cell line, which had fourfold less expression. Comparable mesothelin expression levels were found in the H292 and MeT-5A cell lines, but were 8- to 10-fold less than the expression observed on the cell surface of H226 cells. [Table 25]

[0204] Example 6: Cytotoxicity Assay (T Cell Driven Target Cell Depletion): Introduction biMSLN 高KD MSLN revealed that xCD3xhSA selectively targets and kills T cells expressing mesothelin. 低KD Cytotoxicity assays were performed in the presence of human PBMCs using cell lines expressing different densities of mesothelin on the cell surface to evaluate their efficacy compared to xCD3xhSA. Additionally, the effect of the presence of soluble mesothelin (sMSLN) on the efficacy of the molecules was also evaluated in this assay. 低 KD Simultaneous binding of the xCD3xhSA trispecific molecule leads to cross-linking of CD3ε on T cells, activating a signaling cascade that initiates T cell activation (upregulation of CD69, cytokine secretion) and release of cytotoxic granules, ultimately resulting in target cell killing.

[0205] method blood cell sorting Human peripheral blood mononuclear cells (PBMCs) were isolated from fresh blood of healthy volunteers using lymphocyte separation medium Lymphoprep (Stemcell Technologies) according to the manufacturer's instructions. In this set of experiments, blood from three different donors (Donor #1, Donor #2, and Donor #3) was used. The characteristics of individual donor blood vary significantly, particularly with respect to the quantity and reactivity of CD8+ T cells contained therein. As a result, the killing potential varies widely among these blood-derived CD8+ T cell samples. Consequently, as observed in the examples disclosed herein, different killing and CD8+ T cell activation potentials can be obtained using the same test molecule in the presence of the same target cells.

[0206] Briefly, blood was diluted 1:2 with human PBMC isolation buffer (PBS, 2% FCS, 2 mM EDTA) and applied to Leucosep tubes containing the recommended volume of Lymphoprep medium. The Leucosep tubes were centrifuged at 800 × g for 30 minutes at room temperature without interruption. The PBMC-containing cell layer was then collected and washed twice with human PBMC isolation buffer, and red blood cells were lysed using red blood cell lysis buffer at room temperature for 5 minutes. The isolated human cells were then washed once with the respective isolation buffer and once with assay medium (RPMI-1640, 10% FCS). After platelet removal, the isolated PBMCs were diluted to 3 × 10 per ml in assay medium. 6 The cells were resuspended at a density of 10 viable cells.

[0207] Flow cytometry-based in vitro cytotoxicity assay (FC assay) and CD8+ T cell activation: Three cancer cell lines, H226 cells (high mesothelin density), HPAC cells (intermediate mesothelin density), and H292 cells (low mesothelin density), as well as the MeT-5A cell line (low mesothelin density) derived from healthy mesothelial tissue, were used as target cells. Five thousand viable target cells previously labeled with PKH67 were diluted in 75 μl of assay medium (RPMI-1640, 10% FCS) and added to a 96-well plate. When applicable, assay buffer containing 50, 100, or 500 ng / ml of soluble mesothelin was used. 25 μl of 6-fold concentrated test protein was diluted in assay medium and added to the appropriate wells. 150,000 viable effector cells (PBMCs) were diluted in 50 μl of assay medium and added to each well (E:T ratio 30:1), and the plate was mixed on a nutation mixer at room temperature and then incubated at 37°C and 5% CO. After 40 hours, cells were trypsinized, resuspended in staining buffer (PBS, 2% BCS, 2 mM EDTA), and transferred to a non-binding plate.

[0208] Cells were stained for different markers, including CD69, CD8, CD4, CD11c, and Annexin-V. For analysis, we focused on apoptotic and dead target cells and activated CD8+ T cells. Target cells were identified by green fluorescence (PKH67), and their viability was analyzed using Annexin-V APC. Effector cells (CD8+ cells) were identified by detecting CD8 on their surface (anti-CD8 PerCP-Cy5.5). CD8+ T cell activation was finally detected by quantifying CD69 expression (anti-CD69 PE). CD4 was used to distinguish CD8+ from CD4+ T cells. Monocytes and dendritic cells were stained with CD11c to improve target cell gating. For all markers except Annexin-V, cells were incubated for 30 minutes at room temperature with gentle agitation. Cells were washed once with staining buffer and once with Annexin-binding buffer, and Annexin-V staining was performed for 30 minutes at room temperature with agitation. Cells were washed once with Annexin-V binding buffer, and flow cytometry analysis was performed on a Novocyte flow cytometer.

[0209] The lysis rate of specific target cells was calculated according to the following formula: Specific lysis of target cells (unit: %) = [1 - target cell viability in sample / average viability in control samples] x 100 The percentage of activated CD8+ T cells corresponds to the proportion of CD69+ CD8+ T cells.

[0210] LDH release-based cytotoxicity: The release of LDH (lactate dehydrogenase) from the cytosol is one indicator of cell death. A colorimetric LDH release assay (Roche) was established to examine the cytotoxicity mediated by lead molecules of interest. Two cancer cell lines, H226 cells (high mesothelin density) and OVCAR-3 cells (intermediate mesothelin density), as well as the MeT-5A cell line (low mesothelin density) derived from healthy mesothelial tissue, were used as target cells. 10,000 live target cells were added to a 96-well plate and allowed to adhere overnight. The next day, 300,000 live effector cells (PBMCs) were added to each well in hSA-containing buffer (E:T ratio 30:1). Each molecule shown in the figure was added in 5-fold dilutions starting at 50 nM. Where applicable, final concentrations of 0 ng / mL, 50 ng / mL, and 500 ng / mL sMSLN were added to wells. After 40 hours, supernatants were removed for LDH release assays and cells were stained for T cell markers (including activation).

[0211] The percentage of specific lysis was calculated as follows: 1. Subtract the average OD value of the background medium from the total OD value 2. Calculate % cytotoxicity = ((sample - spontaneous kill) / (maximum kill - spontaneous kill)) x 100 NOTE: Spontaneous killing: Mean OD value of effector + target cells without treatment (untreated). NOTE: Maximum killing: Mean OD value of target cells treated with 1% Triton for 40 h.

[0212] result MATCH Molecule PRO2000 (MATCH-4:biMSLN 高KD xCD3xhSA) and PRO1872 (scMATCH-3:MSLN 低KDThe cytotoxic potency and efficacy of IgG1 (xCD3xhSA) on CD8+ T cell activation were assessed using a flow cytometry-based cytotoxicity assay. Data obtained using the high-mesothelin-expressing cell line H226 and the low-mesothelin-expressing MeT-5A cells derived from healthy tissue are presented in Tables 26 and 27, and concentration-response curves for the MATCH molecules are presented in Figure 6. Both molecules exhibit high potency against the high-mesothelin-expressing H226 cells. The bivalent mesothelin targeting molecule PRO2000 is 75-fold more potent than the monovalent mesothelin targeting molecule PRO1872. Conversely, against MeT-5A cells expressing low levels of mesothelin, the monovalent mesothelin-binding molecule PRO1872 exhibits 16-fold greater potency in target cell killing than PRO2000. Against the high-expressing cells, PRO2000 exhibits a killing potency (EC of 0.07 pM). 50 ) against MeT-5A cells, PRO2000 has an EC of 144.70 pM, while PRO1872 has an EC of 5.31 pM. 50 and PRO1872 shows 8.88 pM. Similar data are observed for CD8+ T cell activation in each condition.

[0213] Additionally, the cytotoxic activity and efficacy of PRO2000 and PRO1872 on CD8+ T cell activation were examined in two other target cancer cell lines, HPAC cells and H292 cells, which express intermediate and low levels of mesothelin, respectively (Table 27 and Figure 7). As previously observed, PRO1872 is more potent against cells with low mesothelin expression than the mesothelin-targeting bivalent molecule PRO2000. Conversely, PRO2000 and PRO1872 have similar potencies against HPAC cells, which exhibit intermediate mesothelin density. Against HPAC, PRO2000 had a killing potency of 40.75 pM and PRO1872 had a killing potency of 30.26 pM, whereas against low-mesothelin-expressing H292 cells, PRO2000 had an EC50 of 652.2 pM. 50 and PRO1872 at 91.03 pM. Similar data were observed for CD8+ T cell activation in each condition, except that PRO2000 was four times more potent than PRO1872 in the presence of HPAC cells.

[0214] Several studies have reported serum concentrations of soluble mesothelin of several hundred ng / ml in cancer patients, so we evaluated the effect of the presence of soluble mesothelin on the efficacy of molecules to kill target cells.

[0215] The cytotoxic potency and efficacy of the MATCH molecules PRO2000 and PRO1872 on CD8+ T cell activation were compared using high-mesothelin-expressing H226 cells in the absence or presence of 50 ng / ml or 500 ng / ml soluble mesothelin (sMSLN). The resulting data are shown in Tables 28 and 29, and the concentration-response curves for the molecules are shown in Figure 8. The potency of both molecules was negatively affected by soluble mesothelin in a dose-dependent manner. Bivalent mesothelin targeting PRO2000 (which has a lower monovalent affinity for mesothelin than PRO1872) exhibits a 17-fold decrease in killing potency in the presence of 500 ng / ml sMSLN compared to the potency observed in the absence of sMSLN. Conversely, PRO1872, a molecule monovalent for mesothelin with superior monovalent affinity for mesothelin, exhibits 106-fold less potency in the presence of 500 ng / ml sMSLN compared to its potency in the absence of sMSLN. Similar data are observed for CD8+ T cell activation in each condition.

[0216] Additionally, we characterized PRO2100, a variant of PRO2000, and showed that both molecules were equally effective at killing target cells. In PRO2100, potential glycosylation sites were mutated to prevent glycosylation.

[0217] The cytotoxic potency of the MATCH4 molecules PRO2000 and PRO2100 was compared in the presence of high mesothelin-expressing H226 cells and low mesothelin-expressing MeT-5A mesothelial cells using a flow cytometry-based cytotoxicity assay. PRO1872 was also included. The resulting data are shown in Table 30, and concentration-response curves are shown in Figure 9. None of the molecules tested exhibited killing of low mesothelin-expressing MeT-5A cells. In the presence of H226 target cells, PRO2000 and PRO2100 exhibit very similar potencies in the absence or presence of sMSLN. In the absence of sMSLN, potency is 0.7 pM for PRO2000 and 1.54 pM for PRO2100. In the presence of 100 ng / ml sMSLN, the potency shifts by 2-fold (3.56 pM for PRO2100) and 5-fold (3.45 pM for PRO2000). Conversely, in the absence of sMSLN, PRO1872, a molecule with superior monovalent affinity, is 10-fold and 20-fold less potent (EC ) than PRO2100 and PRO2000, respectively. 50 PRO1872 = 13.38 pM). Furthermore, PRO1872 is 10-fold less potent in the presence of 100 ng / ml sMSLN.

[0218] Based on the data obtained above, further representative MATCH molecules PRO2567, PRO2566, PRO2562, and PRO2660 (MATCH-4:biMSLN 高KD The cytotoxic potential of PRO2567, PRO2566, and PRO2562 was assessed by LDH release as outlined in the methods section. When examining PRO2567, PRO2566, and PRO2562, these molecules demonstrated greater potency than PRO2660 against high MSLN-expressing H226 cells, with similar dose-response curves (Figure 10). A slight decrease in potency of these molecules was observed in intermediate OVCAR-3 expressing cells, and a further decrease in activity was seen when Met-5A cells were used as the low MSLN-expressing target (Figure 10). These data are summarized in Figure 11, where the EC values ​​across experiments are shown. 50 Values ​​are given including comparison with PRO1872 as a reference. EC of PRO2567, PRO2566, PRO2562, and PRO266050 It is observed that the values ​​appear to be smaller than those observed with PRO1872 when using H226 cells as targets (Figure 11A). A similar trend is observed for PRO2567, PRO2566, and PRO2562 against OVCAR-3 cells (Figure 11B). Mean Met-5A EC 50 and average H226 EC 50 When examining the range of activity by calculating the x-fold difference between values, it is observed that the range of activity for PRO2567, PRO2566, and PRO2562 is significantly wider than that for PRO1872 (see Table 31).

[0219] The potency and dose-response curves of the lead molecules were investigated in the presence of 50 ng / mL and 500 ng / mL sMSLN (Figure 12). 高KD The xCD3xhSA molecules were observed to maintain their potency in the presence of sMSLN. For example, PRO2566 and PRO2567 maintain their potency in the presence of sMSLN compared to PRO1872, even at high sMSLN concentrations (x-fold change in potency from 0 to 500 ng / mL: 14.2-fold and 23.8-fold vs. 78.2-fold; see Table 32). In summary Figure 13, absolute EC 50 The difference in EC values ​​between PRO2566 and PRO2567 is demonstrated. 50 The EC value of monovalent PRO1872 was relatively unaffected by the addition of sMSLN. 50 The values ​​are relatively more affected by the addition of sMSLN. PRO2660 appears to be more sensitive to sMSLN, but is still less sensitive than PRO1872. These data support the MATCH4 biMSLN 高KD We demonstrate that the xCD3xhS format is superior in terms of retaining full antitumor efficacy in the presence of sMSLN.

[0220] Example 7: Binding of MATCH molecules to target cells Introduction To support the cytotoxicity data obtained using target cells expressing different cell surface densities of mesothelin (H226, HPAC, H292, OVCAR-3, and MeT-5A), the cellular binding of MATCH molecules to at least two of these cell lines was assessed by flow cytometry. Representative MATCH4 molecules, PRO2000, PRO2100, PRO2562, PRO2566, PRO2567, and PRO2660, were tested to confirm that both molecules have similar binding properties. scMATCH3 PRO1872 was also included for comparison.

[0221] method Cells were washed twice with 100 μl of PBS and incubated with 5-fold serial dilutions of PRO1872, PRO2000, PRO2100, PRO2562, PRO2566, PRO2567, or PRO2660 ranging from 50,000 to 0.005 pM in staining buffer (PBS, 2% heat-inactivated BCS, 2 mM EDTA). Cells were washed twice with staining buffer, and binding of MATCH4 molecules was visualized with protein L-PE (2 μg / ml). Plates were incubated for 30 min at room temperature on a nutation mixer, washed twice with staining buffer, centrifuged at 200 g for 5 min, and resuspended in a final volume of 50 μl of staining buffer. PE signals from 20,000 events per well were analyzed by flow cytometry using a Novocyte flow cytometer instrument. Data were analyzed using NovoExpress software (ACEA Biosciences). The mean fluorescence intensity (MFI) values ​​of MATCH molecules were corrected for nonspecific binding by subtracting the blank (antibody at zero concentration). ΔMFI data were analyzed with a four-parameter logistic curve fit using GraphPad Prism data analysis software (GraphPad Software) to determine the 50% target cell binding (EC 50 The concentration of the molecule of interest required to reach .DELTA. ...

[0222] result The binding of PRO2000, PRO2100, and PRO1872 to different cell lines was assessed by flow cytometry. The concentration at which half-maximal binding was observed (EC 50 The binding activity and maximum binding intensity (MFI) achieved are shown in Table 33, and the corresponding titration curves are shown in Figure 14 (A-D). PRO2000 and PRO2100 exhibit comparable binding data to all cell lines tested. Compared to PRO1872, PRO2000 and PRO2100 exhibit 3-fold greater binding to the high mesothelin-expressing cancer cell line H226, 1.5-2-fold greater binding to the intermediate mesothelin-expressing cancer cell line HPAC, and 2-fold less binding to the low mesothelin-expressing cancer cell line H292. These data correlate with those obtained in the cytotoxicity assay. The maximum binding observed for each molecule tested confirms the ranking of cell lines with respect to cell surface mesothelin expression. Additionally, cellular binding of the MATCH4 molecules PRO2567, PRO2566, PRO2562, and PRO2660 (biMSLNxCD3xhSA) to different cell lines was also assessed as described above. The concentration at which half-maximal binding was observed (EC 50 The EC values ​​and maximum binding values ​​(MFI) reached are shown in Table 34, and the corresponding titration curves are shown in Figure 14 (E-G). The lead MATCH4 molecules PRO2567, PRO2566, and PRO2562 exhibited high EC values ​​on high MSLN-expressing H226 cells. 50 Binding was demonstrated to be comparable to PRO2000 and PRO2100 (compare Table 33 with Table 34). When assessed for binding to OVCAR-3 cells, which express intermediate levels of MSLN, the EC 50 A further decrease in binding was observed. EC 50 The reduced binding is due to the lower expression level of MSLN on OVCAR-3 cells (as indicated by the maximum MFI values ​​obtained), reducing the contribution of avidity effects to binding strength. For low MSLN-expressing Met-5A cells, a substantial decrease in maximum MFI values ​​was found when compared to H226 and OVCAR-3 cells, indicating a significant decrease in EC 50Values ​​are EC 50 Similar results were obtained with PRO2000 and PRO2100. MSLN on Met-5A cells localized in membrane microdomains, leading to tight, avidity-driven binding of MATCH4 molecules and resulting in small ECs. 50 It can be assumed that the maximum MFI value remains small because of the overall low expression of MSLN on the surface of these cells, whereas the maximum MFI value remains small. In summary, the EC of the lead MATCH4 molecules PRO2567, PRO2566, and PRO2562 50 Binding is comparable to the cell binding data obtained with PRO2000 and PRO2100. The reduced binding strength to intermediate MSLN-expressing OVCAR-3 cells, likely due to a loss of avidity, was also seen in cytotoxicity experiments, where the MATCH4 molecule was found to be less potent at killing OVCAR-3 cells compared to killing high MSLN-expressing H226 cells. [Table 26] [Table 27] [Table 28] [Table 29] [Table 30] [Table 31] [Table 32] [Table 33] [Table 34]

[0223] Example 8: PRO2000 (biMSLN 高KD In vivo tumor growth inhibition using (xCD3xhSA): Introduction PRO2000(biMSLN 高KD To determine the ability of mesothelin-expressing cell line (xCD3xhSA) to effectively control tumor growth compared to control animals, two in vivo mesothelin-expressing cell line xenograft experiments were performed at Charles River Laboratories. One experiment examined tumor growth inhibition in an H292 xenograft model, and the other examined tumor growth inhibition in an HPAC xenograft model.

[0224] method animal Female NCG mice from Charles River Laboratories were raised and housed under conditions suitable for functioning humanized mice. Animals were used at 8–12 weeks of age in both studies.

[0225] research design 1 × 10 subcutaneous injections into the flank of the animals in the treatment groups (n = 5–6 per H292 group, n = 10 per HPAC group). 7 1 x 10 H292 NSCLC tumor cells and 1 x 10 7 PBMCs, or 1 x 10 7 2.5 x 10 HPAC tumor cells and 2.5 x 10 6 PBMCs were simultaneously transplanted. Five days later, the animals were intravenously administered the molecule of interest, with booster doses administered every five days until the end of the experiment. At regular intervals throughout the experiment, the animals were monitored for tumor growth using caliper measurements, as well as for weight loss. The mean tumor volume in the control group was 800 mm 3 Animals were euthanized at 0 days or 40 days, whichever came first. Animals were monitored and euthanized in accordance with Charles River Laboratories animal health and welfare regulations.

[0226] result As described in the Methods section, a PBMC / H292 co-implantation model was used to evaluate the efficacy of PRO2000 (biMSLN) in promoting tumor growth inhibition. 高KD The efficacy of PRO2000 (biMSLN.CD3xhSA) was evaluated in H292 cells, which express intermediate levels of MSLN and are established from non-small cell lung cancer. Multiple dose levels of PRO2000 were administered intravenously, as shown in Figure 15. For comparison, tumor cells implanted in the absence of PBMCs (no treatment) were used, as well as palivizumab (anti-RSV antibody) as a control IgG treatment. Tumor growth was observed in the control condition (light gray line, Figure 15A). Treatment with the PRO2000 (biMSLN.CD3) molecule inhibited tumor growth at 1 mg / kg and 5 mg / kg compared to the control (black and dark gray lines, respectively, in Figure 15A). A two-way ANOVA followed by Tukey's multiple comparison test was used to determine the significance of treatment; data from day 40 are shown in Figure 15B, with each dot representing an individual animal. The two higher doses (1 mg / kg and 5 mg / kg) resulted in significantly smaller tumor volumes compared to palivizumab-treated animals (ctrl) and untreated animals. The lowest dose (0.2 mg / kg) appeared to be suboptimal, as these comparisons were not significant compared to palivizumab-treated animals. Overall, there appeared to be no adverse effects on animal health, as animal weights remained relatively stable throughout the experiment (data not shown). Taken together, these data support the efficacy of the multispecific antibody PRO2000 (biMSLN 高KD We demonstrate that xCD3xhSA) has tumor growth inhibitory activity in vivo and is a promising conceptual candidate for cancer immunotherapy.

[0227] As described in the methods section, a PBMC / HPAC co-transplant model was used to evaluate the efficacy of PRO2000 (biMSLN) in promoting tumor growth inhibition. 高KDThe efficacy of the lowest dose of PRO2000 (biMSLN) was further evaluated. HPAC tumor cells express higher levels of MSLN compared to H292. Similar to the H292 model, growth was observed in the control condition (Figure 16, Palivizumab) and dose-dependent tumor growth inhibition was observed in the test condition. Two-way ANOVA followed by Tukey's multiple comparison test was used to determine the significance of the treatments. 高KD We observed that the monovalent MSLN-targeting molecule PRO1872 (biMSLNxCD3xhSA) resulted in significant tumor growth inhibition at multiple time points compared to the lowest dose of the monovalent MSLN-targeting molecule PRO1872 (MSLNxCD3xhSA). This data supports the efficacy of PRO2000 (biMSLNxCD3xhSA) against cells expressing higher levels of MSLN. 高KD 1 shows that the avidity-based activity of IgG1 (xCD3xhSA) results in improved efficacy.

Claims

1. A multispecific antibody, a) two antibody-based binding domains (MSLN-BD) that specifically bind to mesothelin; b) one antibody-based binding domain (CD3-BD) that specifically binds to CD3; c) one human serum albumin binding domain (hSA-BD); Including, wherein said multispecific antibody does not comprise an immunoglobulin Fc region polypeptide, and each of said MSLN-BDs has a monovalent dissociation constant (K) in the range of 0.5 to 20 nM, particularly in the range of 0.6 to 10 nM, as measured by SPR. Ｄ ) binds to mesothelin (MSLN), and (i) a first single-chain protein comprising the amino acid sequence of SEQ ID NO: 83 and a second single-chain protein comprising the amino acid sequence of SEQ ID NO: 84; (ii) a first single-chain protein comprising the amino acid sequence of SEQ ID NO: 101 and a second single-chain protein comprising the amino acid sequence of SEQ ID NO: 102; (iii) a first single-chain protein comprising the amino acid sequence of SEQ ID NO: 109 and a second single-chain protein comprising the amino acid sequence of SEQ ID NO: 110; or (iv) a first single-chain protein comprising the amino acid sequence of SEQ ID NO: 111 and a second single-chain protein comprising the amino acid sequence of SEQ ID NO: 112; A multispecific antibody comprising:

2. 1. One or two nucleic acids encoding the multispecific antibody of claim 1.

3. A vector or two vectors comprising the nucleic acid or two nucleic acids according to claim 2.

4. A host cell or a plurality of host cells comprising a vector or two vectors according to claim 3.

5. 10. An in vitro method for producing a multispecific antibody according to claim 1, comprising: (i) providing one or two nucleic acids according to claim 2 or one or two vectors according to claim 3, expressing said one or two nucleic acids or said one or two vectors and recovering said multispecific antibody from the expression system, or (ii) providing one or more host cells according to claim 4 and culturing said one or more host cells; and recovering said multispecific antibody from the cell culture. A method comprising:

6. A pharmaceutical composition comprising the multispecific antibody of claim 1 and a pharmaceutically acceptable carrier.

7. 2. The multispecific antibody according to claim 1 for use in the treatment of a disease, in particular a human disease, more particularly a human disease selected from cancer, in particular a cancer selected from mesothelioma, pancreatic cancer and ovarian cancer, an inflammatory disease, and an autoimmune disease.

Images