Multispecific antibody with specificity for ROR1 and CD3

Abstract

The present invention relates to a multispecific antibody comprising one or two binding domains that specifically bind to the extracellular domain of ROR1 (ROR1-BD) and one binding domain that specifically binds to CD3 (CD3-BD), wherein the multispecific antibody does not comprise an immunoglobulin Fc region. The present invention further relates to nucleic acids encoding said multispecific antibody, vector(s) comprising said nucleic acid, host cell(s) comprising said nucleic acid or said vector(s), and methods for producing said multispecific antibody. Additionally, the present invention relates to pharmaceutical compositions comprising said multispecific antibody, and methods of their use.

Description

[Technical field]

[0001] The present invention relates to a multispecific antibody comprising one or two binding domains that specifically bind to the extracellular domain of ROR1 (ROR1-BD) and one binding domain that specifically binds to CD3 (CD3-BD), wherein the multispecific antibody does not comprise an immunoglobulin Fc region. The present invention further relates to nucleic acids encoding said multispecific antibody, vector(s) comprising said nucleic acid, host cell(s) comprising said nucleic acid or said vector(s), and methods for producing said multispecific antibody. Additionally, the present invention relates to pharmaceutical compositions comprising said multispecific antibody, and methods of their use. [Background technology]

[0002] The tyrosine protein kinase transmembrane receptor ROR1 is a member of the receptor tyrosine kinase-like orphan receptor (ROR) family. The extracellular domain of the ROR1 receptor is composed of three distinct domains: a membrane-distal immunoglobulin-like domain, an intervening Frizzled domain, and a membrane-proximal Kringle domain. ROR1 is an oncofetal antigen required during embryonic development. Mice with homozygous knockout of ROR1 die perinatally due to respiratory system dysfunction (Non-Patent Document 1). Expression of ROR1 in normal adult tissues is restricted to certain cells of adipose tissue, pancreas, and lung, and a subset of intermediate B cells (Non-Patent Document 2; Non-Patent Document 3; Non-Patent Document 4). However, it is expressed on several hematological and solid cancers, as well as on a subset of cancer stem cells. Furthermore, several studies have demonstrated that ROR1 plays an important role in cancer progression and metastasis (see, for example, Non-Patent Document 5). Therefore, ROR1 is an attractive target for tumor-specific therapy.

[0003] ROR1 is a cell surface protein that mediates signals through the binding of its ligands, presumably Wnt5a and NKX1-2. Wnt5a has been shown to bind to the Frizzled domain on the extracellular portion of ROR1 and, in transfected cells, has been shown to regulate NF-κB activation and proliferation of normal and lung tumor cell lines. NKX1-2 binding to ROR1 has been shown to play a role in the survival of lung cancer cell lines by both kinase-dependent and kinase-independent mechanisms.

[0004] Cluster of differentiation 3 (CD3) is a type I transmembrane protein expressed on the surface of certain T cells, which is required for T cell activation. It participates in the T cell receptor (TCR) complex and interacts with other domains of this complex. In mammals, CD3 possesses three distinct chains (γ, δ, and ε) and either the ζ2 (CD247) chain or the ζ / η chain. CD3 and the ζ chain, together with the TCR, form what is known as the T cell receptor complex. Antibodies against CD3, such as antibodies against CD3ε, have been shown to induce clustering of CD3 on T cells, leading to T cell activation.

[0005] Several bispecific antibodies targeting ROR1 and CD3 have been described in the prior art.

[0006] For example, US Patent No. 5,999,633 discloses bispecific anti-ROR1 x CD3 antibodies, or DuoBody® molecules, that specifically bind to either the Ig-like domain, the Frizzled domain, or the Kringle domain of the extracellular domain of ROR1. The DuoBody® molecules exhibited EC values ​​ranging from 0.2 to 17 nM for T cell-directed killing of ROR1-expressing modified SK-MES-1 cells. 50The improved bispecific anti-ROR1xCD3 Duobody® molecules RORxCD3D1 and ROR1xCD3D2 showed EC values ​​ranging from 0.2 to 1 nM for T cell-directed killing of ROR1-expressing cancer cells MAVER-1, JeKo-1, and Z-138. 50 values, and the maximum kill is typically less than 50%.

[0007] Patent document 2 discloses an anti-ROR1 x CD3 bispecific diabody consisting of two bridged and disulfide-stabilized Fv binding domains with specificity for two target antigens and connected to an immunoglobulin Fc region (DART® format). Furthermore, T cell-directed cytotoxicity tests of some of these DART antibodies against different ROR1-expressing cancer cell lines are disclosed. The EC of killing of these cancer cells was 50 Values ​​typically range from 10 to 100 pM, with maximal killing of 10 to 40%, with some exceptions; however, only one example has been shown where maximal killing was slightly higher than 60%. For the best DART antibody, DART-D, the EC 50 Values ​​are reported.

[0008] Patent document 3 discloses an anti-ROR1×CD3 bispecific BiTE, i.e., a bispecific antibody consisting of two scFv fragments connected via a short linker. The best BiTE exemplified is clone F, which shows significant cytotoxicity against ROR1-positive B-1643 and B-7 neuroblastoma cells at a concentration of 0.01 microgram / ml.

[0009] US Patent No. 5,999,633 discloses anti-ROR1 x CD3 bispecific antibodies with five different bispecific IgG-based constructs. Twenty bispecific antibodies based on these IgGs were prepared and analyzed for their internalization and their ability to induce T cell-mediated killing of various ROR1-expressing cancer cells. All 20 constructs were internalized, albeit at different rates. The 20 constructs typically showed EC values ​​ranging from a few pM to 50 pM for T cell-mediated killing of the ROR1-expressing cancer cells. 50 values, with maximum kill ranging from 10 to 50%. Only one case has been reported in which the maximum kill exceeded 60%.

[0010] US Patent No. 5,999,663 and subsequent patent application US Patent No. 5,999,663 disclose bispecific antibodies comprising one Fab fragment specifically binding to human CD3ε and one or two additional Fab fragments specifically binding to the extracellular domain of human ROR1, wherein said bispecific antibodies exhibit a low rate of internalization in cell-based assays. Bispecific antibody constructs that are monovalent or bivalent to ROR1 and comprise an immunoglobulin Fc region have been exemplified and subjected to internalization experiments. The results of the internalization experiments indicate that the antibody constructs that are bivalent to ROR1 exhibit a faster internalization rate in primary B-CLL cells than the corresponding antibody constructs that are monovalent to ROR1. T cell-directed cytotoxicity studies with these bispecific antibodies against ROR1-expressing RPMI8226MM target cells demonstrate a maximum target cell lysis of 30 to 40% at an antibody concentration of 100 pM. US Patent No. 6,399,436 further discloses T cell-directed cytotoxicity studies with the improved antibody Mab2-TCBcv, which is bivalent against ROR1, tested against several target cells expressing different levels of ROR1. The data show that Mab2-TCBcv exhibits 20 to 50% maximum target cell lysis for Colo-704 and OVCAR-5 ovarian cancer cells, which express intermediate levels of ROR1, and 20% maximum target cell lysis for SKOV-3 ovarian cancer cells, which express low levels of ROR1.

[0011] As is evident from the above references, prior art anti-ROR1xCD3 antibodies often exhibit significantly reduced potency against cancer cells expressing low levels of ROR1. Moreover, these prior art antibodies typically exhibit low percentages of maximal killing, often well below 50%, in cytotoxicity assays against many ROR1-expressing cancer cells, especially against low ROR1-expressing cancer cells, indicating that their mode of action is not ideal.

[0012] Thus, there is a clear need for new and improved therapeutic anti-ROR1xCD3 antibodies.

[0013] The therapeutic antibodies should have high potency against high ROR1 expressing cancer cells and against cancer cells expressing ROR1 at lower levels. Furthermore, they should exhibit a reasonable percentage of maximal killing, e.g., 40% or more, for these cancer cells. At the same time, they should exhibit a tolerable safety profile to keep dose-limiting side effects at a low level. Furthermore, to facilitate developability and manufacturability in high yields, it is desirable for the therapeutic antibodies to have excellent biophysical properties, such as high stability. [Prior art documents] [Patent documents]

[0014] [Patent Document 1] International Publication No. 2017 / 127499 Brochure [Patent Document 2] International Publication No. 2017 / 142928 Brochure [Patent Document 3] International Publication No. 2019 / 008379 Brochure [Patent Document 4] International Publication No. 2020 / 237173 Brochure [Patent Document 5] International Publication No. 2014 / 167022 Brochure [Patent Document 6] International Publication No. 2016 / 055592 Brochure [Non-patent literature]

[0015] [Non-Patent Document 1] Borcherding et al., 2014 [Non-Patent Document 2] Baskar et al., 2008 [Non-Patent Document 3] Hudecek et al., 2010 [Non-Patent Document 4] Bicocca et al., 2012 [Non-Patent Document 5] Cui et al., 2013 Summary of the Invention

[0016] It is an object of the present invention to provide a medicament for improving the treatment of cancers expressing ROR1. In particular, it is an object of the present invention to provide a novel and improved therapeutic antibody capable of efficiently eradicating even cancer cells expressing low levels of ROR1. It is a further object of the present invention to identify anti-ROR1xCD3 antibodies with an improved mode of action, i.e., exhibiting a high percentage of maximum killing in cytotoxicity assays against ROR1-expressing cancer cells. Said novel and improved therapeutic antibody should furthermore exhibit a tolerable safety profile, i.e., not affect healthy cells or cause dose-limiting toxicity by systemic activation of T cells through non-specific binding to CD3.

[0017] Here, the inventors have found that multispecific antibodies comprising one or two binding domains specifically targeting ROR1 (ROR1-BD) and one binding domain specifically targeting CD3 (CD3-BD), in particular binding domains characterized by specific sequence features, as defined herein, exhibit surprisingly high potency in killing cancer cells expressing ROR1 at high to low levels. Moreover, it has been surprisingly found that said multispecific antibodies comprising one ROR1-BD, as defined herein, typically exhibit maximal killing of 60% or more against cancer cells expressing ROR1 at high to low levels in T cell-directed cytotoxicity assays.

[0018] Thus, in a first aspect, the present invention provides a method for producing a composition comprising: a) one or two binding domains (ROR1-BD) that specifically bind to the extracellular domain of ROR1; and b) comprises one binding domain that specifically binds to CD3 (CD3-BD); Where: - the multispecific antibody does not contain an immunoglobulin Fc region; - said ROR1-BDs, independently of each other, HCDR1 sequence of SEQ ID NO: 1, HCDR2 sequence of SEQ ID NO:2, HCDR3 sequence of SEQ ID NO:3, LCDR1 sequence of SEQ ID NO: 4, The LCDR2 sequence of SEQ ID NO:5, and LCDR3 sequence of SEQ ID NO:6 A set consisting of and / or HCDR1 sequence of SEQ ID NO: 13 or 14, HCDR2 sequence of SEQ ID NO: 15, HCDR3 sequence of SEQ ID NO: 16, LCDR1 sequence of SEQ ID NO: 17, The LCDR2 sequence of SEQ ID NO: 18, and LCDR3 sequence of SEQ ID NO: 19 comprising a set of CDR sequences selected from the set consisting of - said CD3-BD comprising: HCDR1 sequence of SEQ ID NO: 30, HCDR2 sequence of SEQ ID NO: 31, HCDR3 sequence of SEQ ID NO: 32, LCDR1 sequence of SEQ ID NO: 33, The LCDR2 sequence of SEQ ID NO: 34, and LCDR3 sequence of SEQ ID NO: 35 The present invention relates to a multispecific antibody comprising a set of CDR sequences consisting of:

[0019] In a second aspect, the present invention relates to specific ROR1 binding domains.

[0020] In a third aspect, the present invention relates to a nucleic acid or two nucleic acids encoding a multispecific antibody or a specific ROR1-binding domain of the invention.

[0021] In a fourth aspect, the present invention relates to a vector or two vectors comprising a nucleic acid or two nucleic acids according to the invention.

[0022] In a fifth aspect, the present invention relates to a host cell(s) comprising the vector or two vectors of the invention.

[0023] In a sixth aspect, the present invention relates to a method for producing a multispecific antibody of the invention, comprising (i) providing a nucleic acid or two nucleic acids of the invention or a vector or two vectors of the invention, expressing said nucleic acid(s) or said vector(s) and collecting said multispecific antibody or said specific binding domains from the expression system, or (ii) providing a host cell(s) of the invention, culturing said host cell(s) and collecting said multispecific antibody or said specific binding domains from the cell culture.

[0024] In a seventh aspect, the present invention relates to a pharmaceutical composition comprising a multispecific antibody of the invention and a pharma- ceutically acceptable carrier.

[0025] In an eighth aspect, the present invention relates to a multispecific antibody according to the invention for use as a medicament.

[0026] In a ninth aspect, the present invention relates to a multispecific antibody according to the invention for use in the treatment of a disease, more particularly a human disease selected from cancers, in particular cancers expressing ROR1.

[0027] In a tenth aspect, the present invention relates to a method for the treatment of a disease, in particular a human disease, more particularly a human disease selected from cancer, comprising the step of administering a multispecific antibody of the invention to a patient in need thereof.

[0028] The aspects, advantageous features and preferred embodiments of the invention, each taken alone or in combination, summarized in the following paragraphs further contribute to solving the object of the invention:

[0029] 1. Multispecific antibodies are: a) one or two binding domains (ROR1-BD) that specifically bind to the extracellular domain of ROR1; and b) comprises one binding domain that specifically binds to CD3 (CD3-BD); Where: - said ROR1-BDs, independently of each other, HCDR1 sequence of SEQ ID NO: 1, HCDR2 sequence of SEQ ID NO:2, HCDR3 sequence of SEQ ID NO:3, LCDR1 sequence of SEQ ID NO: 4, The LCDR2 sequence of SEQ ID NO:5, and LCDR3 sequence of SEQ ID NO:6 A set consisting of and / or HCDR1 sequence of SEQ ID NO: 13 or 14, HCDR2 sequence of SEQ ID NO: 15, HCDR3 sequence of SEQ ID NO: 16, LCDR1 sequence of SEQ ID NO: 17, The LCDR2 sequence of SEQ ID NO: 18, and LCDR3 sequence of SEQ ID NO: 19, comprising a set of CDR sequences selected from the set consisting of - said CD3-BD comprising: HCDR1 sequence of SEQ ID NO: 30, HCDR2 sequence of SEQ ID NO: 31, HCDR3 sequence of SEQ ID NO: 32, LCDR1 sequence of SEQ ID NO: 33, The LCDR2 sequence of SEQ ID NO: 34, and LCDR3 sequence of SEQ ID NO: 35 The set of CDR sequences consists of:

[0030] 2. Multispecific antibodies are a) one or two binding domains (ROR1-BD) that specifically bind to the extracellular domain of ROR1; and b) comprises one binding domain that specifically binds to CD3 (CD3-BD); Where: - said ROR1-BDs, independently of each other, HCDR1 sequence of SEQ ID NO: 1, HCDR2 sequence of SEQ ID NO:2, HCDR3 sequence of SEQ ID NO:3, LCDR1 sequence of SEQ ID NO: 4, The LCDR2 sequence of SEQ ID NO:5, and LCDR3 sequence of SEQ ID NO:6 A set consisting of or HCDR1 sequence of SEQ ID NO: 13 or 14, HCDR2 sequence of SEQ ID NO: 15, HCDR3 sequence of SEQ ID NO: 16, LCDR1 sequence of SEQ ID NO: 17, The LCDR2 sequence of SEQ ID NO: 18, and LCDR3 sequence of SEQ ID NO: 19 comprising a set of CDR sequences selected from the set consisting of - said CD3-BD comprising: HCDR1 sequence of SEQ ID NO: 30, HCDR2 sequence of SEQ ID NO: 31, HCDR3 sequence of SEQ ID NO: 32, LCDR1 sequence of SEQ ID NO: 33, The LCDR2 sequence of SEQ ID NO: 34, and LCDR3 sequence of SEQ ID NO: 35 2. The multispecific antibody of item 1, comprising a set of CDR sequences consisting of:

[0031] 3. The multispecific antibody has the following characteristics 1) to 3): 1) the multispecific antibody does not contain an immunoglobulin Fc region; 2) the multispecific antibody does not contain a CH1 and / or CL region; 3) said multispecific antibody is humanized, in particular said multispecific antibody is humanized and comprises rabbit-derived CDRs; 3. The multispecific antibody of item 1 or 2, having one or more of the following:

[0032] 4. The multispecific antibody of any one of the preceding items, wherein said one or two ROR1-BDs bind to an Ig-like domain of ROR1 and / or do not block binding of Wnt5α to ROR1.

[0033] 5. The multispecific antibody of any one of the preceding items, in which said multispecific antibody comprises two ROR1-BDs, said two ROR1-BDs comprise two different sets of CDRs, i.e. one ROR1-BD comprises the CDRs of SEQ ID NO: 1 to 6 and the other ROR1-BD comprises the CDRs of SEQ ID NO: 13 / 14 to 19, or said two ROR1-BDs comprise the same set of CDRs, i.e. both ROR1-BDs comprise either the CDRs of SEQ ID NO: 1 to 6 or the CDRs of SEQ ID NO: 13 / 14 to 19, in particular said two ROR1-BDs comprise the same set of CDRs.

[0034] 6. The multispecific antibody of any one of the preceding items, wherein the multispecific antibody is trispecific.

[0035] 7. The multispecific antibody of item 6, wherein the multispecific antibody comprises one binding domain that specifically binds to human serum albumin (hSA-BD).

[0036] 8. The multispecific antibody of any one of the preceding items, wherein the multispecific antibody is trispecific and trivalent or trispecific and tetravalent.

[0037] 9. The multispecific antibody of any one of the preceding items, wherein the format of said multispecific antibody is selected from the group consisting of tandem tri-scFv (triple body), triabody, scDb-scFv, tandem tri-scFv or scDb-scFv fused to the N- and / or C-terminus of a heterodimerization domain other than the heterodimeric Fc domain, and MATCH.

[0038] 10. The multispecific antibody of any one of the preceding items, wherein the format of said multispecific antibody is selected from scDb-scFv, scMATCH3, MATCH3, and MATCH4.

[0039] 11. When each of the one or two ROR1-BDs is in scFv format, a monovalent dissociation constant (K) of 1 pM to 2 nM as measured by surface plasmon resonance (SPR) D ) with a K of 1 pM to 1 nM, especially 1 to 500 pM D , which binds to the extracellular domain of human ROR1, b.EC50 between 5 pM and 10 nM 50 , especially with an EC of 5 pM to 5 nM 50 , especially with an EC of 5 pM to 4 nM 50 , especially with an EC of 5 pM to 3 nM 50 , which binds to human ROR1-expressing MDA-MB-231 cells, c. a melting temperature (Tm) as determined by differential scanning fluorimetry (DSF) of at least 58°C, particularly at least 59°C, particularly at least 60°C, particularly at least 61°C, particularly wherein said scFv is formulated in 50 mM phosphate-citrate buffer with 150 mM NaCl at pH 6.4; d. the scFv has a loss of monomer content after 4 weeks of storage at 4° C. of less than 3%, particularly less than 2%, particularly less than 1%, when the scFv is at a starting concentration of 10 mg / ml, particularly where the scFv is formulated in 50 mM phosphate-citrate buffer with 150 mM NaCl at pH 6.4, and / or e. the scFv has a loss of monomer content after 4 weeks of storage at 40° C. of less than 10% when the scFv is at a starting concentration of 10 mg / ml, particularly where the scFv is formulated in 50 mM phosphate-citrate buffer with 150 mM NaCl at pH 6.4, and / or f. The multispecific antibody of any of the preceding items, having less than 1% loss in protein content after 4 weeks of storage at 4° C. or 40° C. when the scFv is at a starting concentration of 10 mg / ml, particularly wherein the scFv is formulated in 50 mM phosphate-citrate buffer with 150 mM NaCl at pH 6.4.

[0040] 12. The multispecific antibody of any one of the preceding items, wherein said one or two ROR1-BDs comprise VH1a, VH1b, VH3, or VH4 domain framework sequences FR1 to FR4, in particular VH3 or VH4 domain framework sequences FR1 to FR4, in particular VH3 domain framework sequences FR1 to FR4.

[0041] 13. The one or two ROR1-BDs further have the following characteristics 1) and 2): 1) a VL domain comprising framework regions FR1, FR2 and FR3 selected from the VK subtype, in particular from the VK1 and VK3 subtypes, in particular of the VK1 subtype, and a VλFR4 selected from VλFR4, in particular of the VλFR4 having an amino acid sequence having at least 70, 80, 90 percent identity to any of SEQ ID NOs: 76 to 83, more particularly a VλFR4 selected from any of SEQ ID NOs: 76 to 83, in particular a VλFR4 according to SEQ ID NO: 76 or 83; 2) a VH domain comprising VH framework regions FR1, FR2, FR3 and FR4 selected from VH framework subtypes, in particular from VH framework subtypes VH1a, VH1b, VH3 and VH4, in particular from VH framework subtypes VH3 and VH4, in particular of the VH3 subtype, said VH framework regions FR1, FR2, FR3 and FR4 having substitutions (AHo numbering) of arginine (R) at amino acid position 12, threonine (T) at amino acid position 103 and glutamine (Q) at amino acid position 144. The multispecific antibody of any one of the preceding items, having one or both of the following:

[0042] 14. The one or two ROR1-BDs are a) a VH sequence that is at least 90, 91, 92, 93, 94, 95, 96, 97, 98, or 99 percent identical to any one of the amino acid sequences selected from SEQ ID NOs: 7 and 10, and b) a VL sequence that is at least 90, 91, 92, 93, 94, 95, 96, 97, 98, or 99 percent identical to any one of the amino acid sequences selected from SEQ ID NOs: 9 and 12; or a) a VH sequence that is at least 90, 91, 92, 93, 94, 95, 96, 97, 98, or 99 percent identical to any one of the amino acid sequences selected from SEQ ID NOs: 20, 23, 26, and 28; and b) a VL sequence that is at least 90, 91, 92, 93, 94, 95, 96, 97, 98, or 99 percent identical to any one of the amino acid sequences selected from SEQ ID NOs: 22, 25, 27, and 29; or a) a VH sequence that is at least 90, 91, 92, 93, 94, 95, 96, 97, 98, or 99 percent identical to any one of the amino acid sequences selected from SEQ ID NOs: 8 and 11, and b) a VL sequence that is at least 90, 91, 92, 93, 94, 95, 96, 97, 98, or 99 percent identical to any one of the amino acid sequences selected from SEQ ID NOs: 9 and 12; or a) a VH sequence that is at least 90, 91, 92, 93, 94, 95, 96, 97, 98, or 99 percent identical to any one of the amino acid sequences selected from SEQ ID NOs: 21 and 24, and b) a VL sequence that is at least 90, 91, 92, 93, 94, 95, 96, 97, 98, or 99 percent identical to any one of the amino acid sequences selected from SEQ ID NOs: 22 and 25. The multispecific antibody of any one of the preceding items, comprising:

[0043] 15. The one or two ROR1-BDs are a) a VH sequence that is at least 90, 91, 92, 93, 94, 95, 96, 97, 98, or 99 percent identical to any one of the amino acid sequences selected from SEQ ID NOs: 7, 20, and 26; and b) a VL sequence that is at least 90, 91, 92, 93, 94, 95, 96, 97, 98, or 99 percent identical to any one of the amino acid sequences selected from SEQ ID NOs: 9, 22, and 27; or 1) a VH sequence that is at least 90, 91, 92, 93, 94, 95, 96, 97, 98, or 99 percent identical to any one of the amino acid sequences selected from SEQ ID NOs: 8, 11, and 24; and 2) a VL sequence that is at least 90, 91, 92, 93, 94, 95, 96, 97, 98, or 99 percent identical to any one of the amino acid sequences selected from SEQ ID NOs: 9, 12, and 25; The multispecific antibody of any one of the preceding items, comprising:

[0044] 16. The one or two ROR1-BDs are a) a VH sequence selected from SEQ ID NOs: 7 and 10; b) a VL sequence selected from SEQ ID NOs: 9 and 12; or a) a VH sequence selected from SEQ ID NOs: 20, 23, 26, and 28; b) a VL sequence selected from SEQ ID NOs: 22, 25, 27, and 29; or a) a VH sequence selected from SEQ ID NOs: 8, 11, and 14; b) a VL sequence selected from SEQ ID NOs: 9, 12, and 25 15. The multispecific antibody of any one of items 1 to 14, comprising:

[0045] 17. The one or two ROR1-BDs are a) the VH sequence of SEQ ID NO: 7 and the VL sequence of SEQ ID NO: 9, or b) the VH sequence of SEQ ID NO: 10 and the VL sequence of SEQ ID NO: 12, or c) the VH sequence of SEQ ID NO: 20 and the VL sequence of SEQ ID NO: 22, or d) the VH sequence of SEQ ID NO: 23 and the VL sequence of SEQ ID NO: 25, or e) the VH sequence of SEQ ID NO: 8 and the VL sequence of SEQ ID NO: 9, or f) the VH sequence of SEQ ID NO: 11 and the VL sequence of SEQ ID NO: 12, or g) the VH sequence of SEQ ID NO: 21 and the VL sequence of SEQ ID NO: 22, or h) the VH sequence of SEQ ID NO: 24 and the VL sequence of SEQ ID NO: 25, or i) the VH sequence of SEQ ID NO: 26 and the VL sequence of SEQ ID NO: 27, or j) comprising the VH sequence of SEQ ID NO: 28 and the VL sequence of SEQ ID NO: 29, In particular, wherein said one or two ROR1-BDs are a) the VH sequence of SEQ ID NO: 7 and the VL sequence of SEQ ID NO: 9, or b) the VH sequence of SEQ ID NO: 10 and the VL sequence of SEQ ID NO: 12, or c) the VH sequence of SEQ ID NO: 23 and the VL sequence of SEQ ID NO: 25, or d) the VH sequence of SEQ ID NO: 8 and the VL sequence of SEQ ID NO: 9, or e) the VH sequence of SEQ ID NO: 11 and the VL sequence of SEQ ID NO: 12, or f) comprising the VH sequence of SEQ ID NO: 24 and the VL sequence of SEQ ID NO: 25; The multispecific antibody of any one of the preceding items.

[0046] 18. The multispecific antibody of any one of the preceding items, wherein said CD3-BD binds to CD3ε.

[0047] 19. The CD3-BD, when in scFv format, comprises: a) A dissociation constant (K) of less than 50 nM as measured by SPR D ) with a K of 0.1 to 50 nM, in particular 0.1 to 30 nM, in particular 0.1 to 15 nM, in particular 0.1 to 10 nM D , which binds to human CD3ε, b) a K of less than 50 nM as measured by SPR D and especially a K of 0.1 to 50 nM, especially 0.1 to 30 nM, especially 0.1 to 15 nM, especially 0.1 to 10 nM. D , which binds to Macaca fascicularis (cynomolgus) CD3, c) has a melting temperature (Tm) as determined by differential scanning fluorimetry (DSF) of at least 60°C, in particular at least 65°C, and more particularly at least 67°C. The multispecific antibody of any one of the preceding items.

[0048] 20. The multispecific antibody of any one of the preceding items, wherein said CD3-BD comprises VH1a, VH1b, VH3, or VH4 domain framework sequences FR1 to FR4, in particular VH3 or VH4 domain framework sequences FR1 to FR4, in particular VH3 domain framework sequences FR1 to FR4.

[0049] 21. The CD3-BD has the following characteristics 1) and 2) 1) a VL domain comprising framework regions FR1, FR2 and FR3 selected from the VK subtype, in particular from the VK1 and VK3 subtypes, in particular of the VK1 subtype, and a VλFR4 selected from VλFR4, in particular of the VλFR4 having an amino acid sequence having at least 70, 80, 90 percent identity to any of SEQ ID NOs: 76 to 83, more particularly a VλFR4 selected from any of SEQ ID NOs: 76 to 83, in particular a VλFR4 according to SEQ ID NO: 76 or 83; 2) a VH domain comprising VH framework regions FR1, FR2, FR3 and FR4 selected from VH framework subtypes, in particular from VH framework subtypes VH1a, VH1b, VH3 and VH4, in particular from VH framework subtypes VH3 and VH4, in particular of the VH3 subtype, said VH framework regions FR1, FR2, FR3 and FR4 having substitutions (AHo numbering) of arginine (R) at amino acid position 12, threonine (T) at amino acid position 103 and glutamine (Q) at amino acid position 144. 21. The multispecific antibody of items 18 to 20, having one or both of the following:

[0050] 22. The CD3-BD comprises: a) a VH sequence that is at least 90, 91, 92, 93, 94, 95, 96, 97, 98, or 99 percent identical to any one of the amino acid sequences selected from SEQ ID NO: 36 or 37, and b) a VL sequence that is at least 90, 91, 92, 93, 94, 95, 96, 97, 98, or 99 percent identical to any one of the amino acid sequences selected from SEQ ID NO: 38. The multispecific antibody of any one of the preceding items, comprising:

[0051] 23. The CD3-BD comprises: a) a VH domain comprising the amino acid sequence of SEQ ID NO: 36 or 37, and b) a VL domain comprising the amino acid sequence of SEQ ID NO: 38 23. The multispecific antibody of item 22, comprising

[0052] 24. The multispecific antibody of item 22 or 23, wherein the VH sequence comprises the C51 amino acid residue (AHo numbering) and the VL sequence comprises the C141 amino acid residue (AHo numbering).

[0053] 25. The hSA-BD comprises: (i) the HCDR1, HCDR2, and HCDR3 sequences of SEQ ID NOs: 39, 40, and 41, respectively, and human VH framework sequences FR1, FR2, FR3, and FR4 selected from VH framework subtypes, in particular from VH framework subtypes VH1a, VH1b, VH3, and VH4, in particular from VH framework subtypes VH3 and VH4, in particular of the VH3 subtype, wherein said VH framework regions FR1, FR2, FR3, and FR4 optionally have substitutions (AHo numbering) of arginine (R) at amino acid position 12, threonine (T) at amino acid position 103, and glutamine (Q) at amino acid position 144; and (ii) a human antibody VL framework sequence FR1, FR2, and FR3 selected from the Vκ subtype, particularly from the Vκ1 and Vκ3 subtypes, and in particular the Vκ1 subtype, of SEQ ID NO: 42, 43, and 44, respectively, and a VλFR4 sequence, particularly a VλFR4 sequence selected from any of SEQ ID NO: 76 to SEQ ID NO: 83, in particular a VλFR4 sequence according to SEQ ID NO: 76 or 83; or (i) the HCDR1, HCDR2, and HCDR3 sequences of SEQ ID NOs: 49, 50, and 51, respectively, and human VH framework sequences FR1, FR2, FR3, and FR4 selected from VH framework subtypes, in particular from VH framework subtypes VH1a, VH1b, VH3, and VH4, in particular from VH framework subtypes VH3 and VH4, in particular being of the VH3 subtype, wherein said VH framework regions FR1, FR2, FR3, and FR4 optionally have substitutions (AHo numbering) of an arginine (R) at amino acid position 12, a threonine (T) at amino acid position 103, and a glutamine (Q) at amino acid position 144; (ii) a human antibody VL framework sequence FR1, FR2, and FR3 selected from the Vκ subtype, particularly from the Vκ1 and Vκ3 subtypes, and in particular the Vκ1 subtype, and a VλFR4 sequence, particularly a VλFR4 sequence selected from any of SEQ ID NO: 76 to SEQ ID NO: 83, in particular the VλFR4 sequence according to SEQ ID NO: 76 or 83; or (i) the HCDR1, HCDR2, and HCDR3 sequences of SEQ ID NOs: 59, 60, and 61, respectively, and human VH framework sequences FR1, FR2, FR3, and FR4 selected from VH framework subtypes, in particular from VH framework subtypes VH1a, VH1b, VH3, and VH4, in particular from VH framework subtypes VH3 and VH4, in particular being of the VH3 subtype, wherein said VH framework regions FR1, FR2, FR3, and FR4 optionally have substitutions (AHo numbering) of an arginine (R) at amino acid position 12, a threonine (T) at amino acid position 103, and a glutamine (Q) at amino acid position 144; (ii) human antibody VL framework sequences FR1, FR2, and FR3 selected from the Vκ subtype, particularly from the Vκ1 and Vκ3 subtypes, and particularly the Vκ1 subtype, of SEQ ID NO: 62, 63, and 74, respectively, and a VλFR4 sequence, particularly a VλFR4 sequence selected from any of SEQ ID NO: 76 to SEQ ID NO: 83, particularly a VλFR4 sequence according to SEQ ID NO: 76 or 83. 8. The multispecific antibody of item 7, comprising

[0054] 26. The hSA-BD comprises: (i) a VH domain comprising the amino acid sequence of SEQ ID NO: 45 and a VL domain comprising the amino acid sequence of SEQ ID NO: 46, or (ii) a VH domain comprising the amino acid sequence of SEQ ID NO: 47 and a VL domain comprising the amino acid sequence of SEQ ID NO: 48; or (iii) a VH domain comprising the amino acid sequence of SEQ ID NO: 55 and a VL domain comprising the amino acid sequence of SEQ ID NO: 56; or (iv) a VH domain comprising the amino acid sequence of SEQ ID NO: 57 and a VL domain comprising the amino acid sequence of SEQ ID NO: 58; or (v) a VH domain comprising the amino acid sequence of SEQ ID NO: 65 and a VL domain comprising the amino acid sequence of SEQ ID NO: 67; or (vi) a VH domain comprising the amino acid sequence of SEQ ID NO: 66 and a VL domain comprising the amino acid sequence of SEQ ID NO: 67; or (vii) a VH domain comprising the amino acid sequence of SEQ ID NO: 68 and a VL domain comprising the amino acid sequence of SEQ ID NO: 70; or (viii) 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 26. The multispecific antibody of item 7 or 25, comprising

[0055] 27. The above hSA-BD is (i) a VH domain comprising the amino acid sequence of SEQ ID NO: 47 and a VL domain comprising the amino acid sequence of SEQ ID NO: 48, or (ii) a VH domain comprising the amino acid sequence of SEQ ID NO: 57 and a VL domain comprising the amino acid sequence of SEQ ID NO: 58; or (iii) a VH domain comprising the amino acid sequence of SEQ ID NO: 68 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: 69 and a VL domain comprising the amino acid sequence of SEQ ID NO: 70 26. The multispecific antibody of item 7 or 25, comprising

[0056] 28. The multispecific antibody of any one of the preceding items, wherein said binding domains are independently selected from the group consisting of Fab, F(ab)2, Fv, scFv, dsFv, and scAb, in particular from the group consisting of Fv, scFv, and dsFv.

[0057] 29. Each of the binding domains independently a) a cognate pair of VL and VH domains (Fv fragment), or b) Cognate pairs of VL and VH domains linked by an oligo- or polypeptide linker (scFv fragments) 29. The multispecific antibody of item 28, selected from the group consisting of

[0058] 30. Multispecific antibodies include: a) one ROR1-BD, b) one CD3-BD, and c) One hSA-BD 30. The multispecific antibody of any one of items 1 to 29.

[0059] 31. The multispecific antibody is a single chain protein, the single chain protein being 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 comprising an amino acid sequence consisting of 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; wherein the single chain protein further comprises (viii) a third binding domain formed by a third VL domain and a third VH domain connected via a fourth polypeptide linker, and C-terminally or N-terminally fused to said amino acid sequence via a fifth polypeptide linker; wherein the three binding domains have the following specificities: a) One binding domain specifically binds to human ROR1 (ROR1-BD); b) another binding domain specifically binds to CD3 (CD3-BD), and c) The remaining binding domain specifically binds to hSA (hSA-BD) 31. The multispecific antibody of item 30, comprising:

[0060] 32. The three binding domains have the following specificities: a) one of the first and second binding domains specifically binds to human serum albumin (hSA-BD); b) the other of the first and second binding domains specifically binds to human CD3 (CD3-BD), and c) the third binding domain specifically binds to ROR1 (ROR1-BD); having In particular, the three binding domains have the following specificities: a) the first binding domain specifically binds to human serum albumin (hSA-BD); b) the second binding domain specifically binds to human CD3 (CD3-BD), and c) The third binding domain specifically binds to ROR1 (ROR1-BD) 32. The multispecific antibody of item 31, having the formula:

[0061] 33. The multispecific antibody is a heterodimeric protein comprising a first and a second single chain protein; wherein the first single-chain protein comprises a first amino acid sequence consisting of (from N- to C-terminus): (ia) a first VL domain, (iia) a first polypeptide linker, and (iiia) a second VL domain, wherein the second single chain protein comprises a second amino acid sequence consisting of (from N- to C-terminus): (ib) a first VH domain, (iib) a second polypeptide linker, and (iiib) a second VH domain, wherein 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; wherein at least one of said first and said second single chain proteins further comprises: (iv) a third binding domain formed by a third VL domain and a third VH domain connected via a third polypeptide linker, wherein said third binding domain is fused to said first or said second amino acid sequence via a fourth polypeptide linker; wherein said three binding domains have the following specificities: a) one binding domain specifically binds to human CD3 (CD3-BD); b) another binding domain specifically binds to ROR1 (ROR1-BD), and c) The remaining binding domain specifically binds human serum albumin (hSA-BD) The multispecific antibody of item 30.

[0062] 34. The multispecific antibody comprises: a monovalent dissociation constant (K) of less than 5 nM as measured by surface plasmon resonance (SPR) D ) with a monovalent K of 1 pM to 5 nM D and binds to the extracellular domain of human ROR1 with a KD of 1 pM to 1 nM, particularly 1 to 500 pM; b) A (monovalent) K of less than 20 nM as measured by SPR D and a monovalent K of 0.1 to 20 nM, in particular 0.1 to 15 nM, in particular 0.1 to 10 nM. D , which binds to human CD3ε, C. fascicularis (cynomolgus monkey) CD3ε, specifically, a monovalent K of 30 nM or less as measured by SPR. D , especially 0.1 to 30 nM, especially 0.1 to 20 nM monovalent K D binds to cynomolgus CD3ε and d.EC50 from 0.01 to 50 nM 50 , especially with an EC of 0.01 to 30 nM 50 , especially with an EC of 0.01 to 20 nM 50 , especially with an EC of 0.01 to 10 nM 50 , especially with an EC of 0.01 to 5 nM 50, which binds to human ROR1-expressing MDA-MB-231 cells, e. An EC for killing human ROR1-expressing MDA-MB-231 target cells of less than 5 nM, particularly 1 pM to 5 nM, particularly 1 pM to 2 nM, particularly 1 pM to 1 nM, as determined in a T cell-driven cytotoxicity assay against said target cells. 50 having f. a melting temperature (Tm) of at least 55°C, preferably at least 58°C, more preferably at least 60°C, as determined by differential scanning fluorimetry, particularly wherein said antigen-binding fragment is in 50 mM phosphate-citrate buffer with 300 mM D(+)-sucrose at pH 6.5; g. the scFv has less than 1% loss of monomer content after 4 weeks of storage at 4° C. when the scFv is at a starting concentration of 1 mg / ml, e.g., less than 0.5%, less than 0.3%, less than 0.2%, particularly where the scFv is formulated in 50 mM phosphate-citrate buffer with 300 mM D(+)-sucrose at pH 6.5; and / or h. The multispecific antibody of any one of items 30 to 33, having a loss of monomer content after 4 weeks of storage at 40° C. of less than 5%, such as less than 4%, when said scFv is at a starting concentration of 1 mg / ml, in particular wherein said scFv is formulated in 50 mM phosphate-citrate buffer with 300 mM D(+)-sucrose at pH 6.5.

[0063] 35. The multispecific antibody of any one of items 30, 31, 33, and 34, wherein the multispecific antibody is selected from the antibodies defined by SEQ ID NOs: 87, 88, 89, 90, 91, 92, 93, 94, 107, and 108.

[0064] 36. A multispecific antibody is a) Two ROR1-BDs; b) one CD3-BD, and c) One hSA-BD 30. The multispecific antibody of any one of items 1 to 29, comprising:

[0065] 37. The multispecific antibody of item 36, wherein the two ROR1-BDs bind to the same epitope on the extracellular domain of ROR1 or to different epitopes on the extracellular domain of ROR1, in particular to the same epitope on the extracellular domain of ROR1.

[0066] 38. The multispecific antibody is a heterodimeric protein comprising a first and a second single chain protein; wherein the first single-chain protein comprises (from the N-terminus to the C-terminus): (ia) a first VL domain, (iia) a first polypeptide linker, and (iiia) a second VL domain, a first amino acid sequence consisting of wherein the second single-chain protein comprises (from the N- to C-terminus) (ib) a first VH domain, (iib) a second polypeptide linker, and (iiib) a second VH domain, and a second amino acid sequence consisting of wherein 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; wherein at least one of the first and second single-chain proteins further comprises (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 said first or said second amino acid sequence via a fourth polypeptide linker; wherein at least one of the first and second single-chain proteins further comprises (v) a fourth binding domain formed by a fourth VL domain and a fourth VH domain connected via a fifth polypeptide linker and fused to said first or said second amino acid sequence via a sixth polypeptide linker; Here, the four binding domains have the following specificities: a) one binding domain specifically binds to human CD3 (CD3-BD); b) two additional binding domains specifically bind to ROR1 (ROR1-BD), and c) The remaining binding domain specifically binds human serum albumin (hSA-BD) 38. The multispecific antibody of item 36 or 37, comprising:

[0067] 39. The multispecific antibody of item 38, wherein the third binding domain is fused to either the first or the second amino acid sequence and the fourth binding domain is fused to the other of the two said amino acid sequences.

[0068] 40. The four binding domains have the following specificities: a) the first binding domain specifically binds to human CD3 (CD3-BD); b) the second binding domain specifically binds human serum albumin (hSA-BD), and c) The third and fourth binding domains specifically bind to ROR1 (ROR1-BD) 40. The multispecific antibody of item 38 or 39, having the formula:

[0069] 41. The multispecific antibody comprises: a monovalent dissociation constant (K) of less than 10 nM as measured by surface plasmon resonance (SPR) D ) with a monovalent K of 10 pM to 5 nM D , especially 10 pM to 7 nM, especially 10 pM to 5 nM K D , which binds to the extracellular domain of human ROR1, b. A monovalent K of less than 20 nM as measured by SPR D Monovalent K Dand a monovalent K of, in particular, 0.1 to 20 nM, in particular 0.1 to 15 nM, in particular 0.1 to 10 nM. D , which binds to human CD3ε, C. fascicularis (cynomolgus monkey) CD3ε, specifically, a monovalent K of 30 nM or less as measured by SPR. D , especially 0.1 to 30 nM, especially 0.1 to 20 nM monovalent K D , binds to cynomolgus CD3ε, d.EC50 from 0.01 to 10 nM 50 , especially with an EC of 0.01 to 10 nM 50 , especially with an EC of 0.01 to 5 nM 50 , especially with an EC of 0.01 to 3 nM 50 , especially with an EC of 0.01 to 2 nM 50 , which binds to human ROR1-expressing MDA-MB-231. e. an EC of less than 500 pM, particularly between 0.1 pM and 500 pM, particularly between 0.1 pM and 200 pM, particularly between 0.1 and 100 pM, particularly between 0.1 and 50 pM, for killing human ROR1-expressing MDA-MB-231 target cells as determined by a T cell-driven cytotoxicity assay against said target cells; 50 having f. An EC for killing human ROR1-expressing JEKO-1 target cells of less than 20 pM, particularly 0.01 pM to 20 pM, particularly 0.01 pM to 10 pM, particularly 0.01 to 5 pM, particularly 0.01 to 3 pM, as determined by a T cell-driven cytotoxicity assay against said target cells. 50 having g. An EC for killing human ROR1-expressing SKOV-3 target cells of less than 1000 pM, particularly 0.1 pM to 1000 pM, particularly 0.1 pM to 500 pM, particularly 0.1 to 200 pM, particularly 0.1 to 100 pM, as determined by a T cell-driven cytotoxicity assay against said target cells. 50 having h. has a melting temperature (Tm) as determined by differential scanning fluorimetry of at least 53°C, preferably at least 55°C, more preferably at least 58°C, particularly wherein said antigen-binding fragment is formulated in 50 mM phosphate-citrate buffer with 300 mM sucrose at pH 6.5; i. said multispecific antibody has a loss of monomer content after 2 weeks of storage at 4° C. of less than 1%, in particular less than 0.5%, when said multispecific antibody is at a starting concentration of 1 mg / ml, in particular wherein said multispecific antibody is formulated in 50 mM phosphate-citrate buffer with 300 mM D(+)-sucrose at pH 6.5, and / or j. said multispecific antibody has less than 5% loss of monomer content after at least 4 weeks of storage at 40° C. when said multispecific antibody is at a starting concentration of 1 mg / ml, in particular wherein said multispecific antibody is formulated in 50 mM phosphate-citrate buffer with 300 mM D(+)-sucrose at pH 6.5; 41. The multispecific antibody of any one of items 36 to 40.

[0070] 42. The first single-chain protein comprises the amino acid sequence of SEQ ID NO: 95, more specifically consisting of the amino acid sequence of SEQ ID NO: 95; and The second single chain protein comprises the amino acid sequence of SEQ ID NO: 96, more specifically consisting of the amino acid sequence of SEQ ID NO: 96; or The first single-chain protein comprises the amino acid sequence of SEQ ID NO:97, more specifically consisting of the amino acid sequence of SEQ ID NO:97; and The second single chain protein comprises the amino acid sequence of SEQ ID NO: 98, more specifically consisting of the amino acid sequence of SEQ ID NO: 98; or The first single-chain protein comprises the amino acid sequence of SEQ ID NO: 99, more specifically consisting of the amino acid sequence of SEQ ID NO: 99; and The second single chain protein comprises the amino acid sequence of SEQ ID NO: 100, more specifically consisting of the amino acid sequence of SEQ ID NO: 100; or The first single-chain protein comprises the amino acid sequence of SEQ ID NO: 101, more specifically consisting of the amino acid sequence of SEQ ID NO: 101; and The second single chain protein comprises the amino acid sequence of SEQ ID NO: 102, more specifically consisting of the amino acid sequence of SEQ ID NO: 102; or The first single-chain protein comprises the amino acid sequence of SEQ ID NO: 103, more specifically consisting of the amino acid sequence of SEQ ID NO: 103; and The second single chain protein comprises the amino acid sequence of SEQ ID NO: 104, more specifically consisting of the amino acid sequence of SEQ ID NO: 104; or The first single-chain protein comprises the amino acid sequence of SEQ ID NO: 105, more specifically consisting of the amino acid sequence of SEQ ID NO: 105; and The second single chain protein comprises the amino acid sequence of SEQ ID NO: 106, more specifically consisting of the amino acid sequence of SEQ ID NO: 106; or The first single-chain protein comprises the amino acid sequence of SEQ ID NO: 109, more specifically consisting of the amino acid sequence of SEQ ID NO: 109; and The second single chain protein comprises the amino acid sequence of SEQ ID NO: 110, more specifically consisting of the amino acid sequence of SEQ ID NO: 110; or The first single-chain protein comprises the amino acid sequence of SEQ ID NO: 111, more specifically consisting of the amino acid sequence of SEQ ID NO: 111; and The second single chain protein comprises the amino acid sequence of SEQ ID NO: 112, more specifically consisting of the amino acid sequence of SEQ ID NO: 112; or The first single-chain protein comprises the amino acid sequence of SEQ ID NO: 113, more specifically consisting of the amino acid sequence of SEQ ID NO: 113; and said second single chain protein comprises the amino acid sequence of SEQ ID NO: 114, more specifically consisting of the amino acid sequence of SEQ ID NO: 114; or The first single-chain protein comprises the amino acid sequence of SEQ ID NO: 115, more specifically consisting of the amino acid sequence of SEQ ID NO: 115; and The second single chain protein comprises the amino acid sequence of SEQ ID NO: 116, more specifically consisting of the amino acid sequence of SEQ ID NO: 116. 42. The multispecific antibody of any one of items 36 to 41.

[0071] 43. The multispecific antibody of any one of items 32 and 38 to 41, wherein said heterodimeric protein does not comprise a cognate pair of first and second protein interaction domains other than said first and second VL and VH domains, wherein said first protein interaction domain is comprised on said first single chain protein and wherein said second protein interaction domain is comprised on said second single chain protein.

[0072] 44. The multispecific antibody of any one of items 43, 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.

[0073] 45. The multispecific antibody of any one of items 43, 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.

[0074] 46. ​​The multispecific antibody of any one of items 30 to 45, wherein any binding domain contained on said single-chain protein or heterodimeric protein is an immunoglobulin variable domain and is located on said first and second single-chain proteins, typically connected via a linker.

[0075] 47. The multispecific antibody of any one of the preceding items, wherein at least one of said antibody variable domains comprises a CDR region derived from a parent rabbit antibody.

[0076] 48. The multispecific antibody of any one of the preceding items, wherein the ROR1-BD, or, in the case where two ROR1-BDs are present in the multispecific antibody, at least one of the ROR1-BDs and the CD3-BD, are capable of simultaneously binding to their respective antigens, or, in the case where two ROR1-BDs are present in the multispecific antibody, both the ROR1-BDs and the CD3-BD are capable of simultaneously binding to their respective antigens.

[0077] 49. ROR1-BD as defined by any one of items 1, 4, and 11 to 17.

[0078] 50. a) One or two ROR1-BDs as defined in item 49; b) at least one binding domain that specifically binds to a target distinct from ROR1 A multispecific antibody comprising:

[0079] 51. A nucleic acid or two nucleic acids encoding the multispecific antibody of any one of items 1 to 48 and 50 or the ROR1-BD of item 49.

[0080] 52. A vector or two vectors comprising the nucleic acid or two nucleic acids of item 51.

[0081] 53. A host cell(s) containing the vector or two vectors of item 52.

[0082] 54. A method for producing a multispecific antibody according to any one of items 1 to 48 and 50 or a ROR1-BD according to item 49, comprising (i) providing a nucleic acid or two nucleic acids according to item 51, or a vector or two vectors according to item 52, expressing said nucleic acid(s) or said vector(s) and collecting said multispecific antibody or said ROR1-BD from the expression system, or (ii) providing a host cell(s) according to item 53, culturing said host cell(s) and collecting said multispecific antibody or said ROR1-BD from the cell culture.

[0083] 55. A pharmaceutical composition comprising the multispecific antibody of any one of items 1 to 48 and 50 and a pharma- ceutically acceptable carrier.

[0084] 56. The multispecific antibody of any one of items 1 to 48 and 50, for use as a medicament.

[0085] 57. The multispecific antibody of any one of items 1 to 48 and 50, for use in the treatment of a disease, in particular a human disease, more in particular a human disease selected from cancer.

[0086] 58. The multispecific antibody of any one of items 1 to 48 and 50 for use in treating the disease of item 57, wherein the disease is a ROR1-expressing cancer.

[0087] 59. The multispecific antibody of item 58, wherein the ROR1-expressing cancer is selected from chronic lymphocytic leukemia / small lymphocytic lymphoma (CLL / SLL), acute myeloid leukemia (AML), acute lymphoblastic leukemia (ALL), mantle cell lymphoma (MCL), hairy cell leukemia, follicular lymphoma (FL), marginal zone lymphoma (MZL), diffuse large B-cell lymphoma (DLBCL), non-Hodgkin's lymphoma (NHL), Richter's syndrome (RS), lung cancer, pancreatic cancer, prostate cancer, colon cancer, bladder cancer, breast cancer, ovarian cancer, glioblastoma, testicular cancer, uterine cancer, adrenal cancer, melanoma, neuroblastoma, sarcoma, and renal cancer.

[0088] 60. A method for the treatment of a disease, in particular a human disease, more particularly a human disease selected from cancer, comprising administering to a patient in need thereof a multispecific antibody according to any one of items 1 to 48.

[0089] 61. The method of item 60, wherein the disease is a ROR1-expressing cancer.

[0090] 62. The method of item 61, wherein the ROR1-expressing cancer is selected from chronic lymphocytic leukemia / small lymphocytic lymphoma (CLL / SLL), acute myeloid leukemia (AML), acute lymphoblastic leukemia (ALL), mantle cell lymphoma (MCL), hairy cell leukemia, follicular lymphoma (FL), marginal zone lymphoma (MZL), diffuse large B-cell lymphoma (DLBCL), non-Hodgkin's lymphoma (NHL), Richter's syndrome (RS), lung cancer, pancreatic cancer, prostate cancer, colon cancer, bladder cancer, breast cancer, ovarian cancer, glioblastoma, testicular cancer, uterine cancer, adrenal cancer, melanoma, neuroblastoma, sarcoma, and kidney cancer. [Brief description of the drawings]

[0091] [Figure 1]Representative concentration-response curves of anti-ROR1 scFv PRO2060 (A, B) and PRO2062 against human ROR1 expressed on MDA-MB-231 cells and against ROR1-negative MCF-7 cells by flow cytometry are shown. PRO2060 and PRO2062 demonstrated binding to ROR1-positive MDA-MB-231 cells (A, C), whereas no binding was observed to ROR1-negative MCF-7 cells (B, D). Compared to the reference Fab fragment PRO2213, PRO2060 showed approximately 3-fold lower EC50 binding and PRO2062 showed approximately 3-fold higher EC50 binding. Data points in the figures represent the average of two technical replicates. [Diagram 2] Concentration-response curves of the optimized anti-ROR1 scFvs PRO2271 (A, B) and PRO2291 and PRO2292 (C, D) against human ROR1 expressed on MDA-MB-231 cells and against ROR1-negative MCF-7 cells by flow cytometry are shown. All scFvs demonstrated binding to ROR1-positive MDA-MB-231 cells (A, C), but no binding was found to ROR1-negative MCF-7 cells (B, D). Compared to PRO2062, optimized PRO2271 showed approximately 2-fold lower EC50 binding. PRO2060 and PRO2062 variants PRO2291 and PRO2292, which include a VL-VH interdomain disulfide bond for stabilization, demonstrated similar EC50 binding when compared to PRO2060 and PRO2062. Data points in the figure represent the average of two technical replicates. [Diagram 3] Schematic diagram of the trispecific scMATCH3 format. [Figure 4]Concentration-response curves of anti-ROR1xhSAxCD3 scMATCH3 molecules PRO2286, PRO2287 (A, B), and PRO2257 (C, D) against human ROR1 expressed on MDA-MB-231 cells and against ROR1-negative MCF-7 cells by flow cytometry. All three scMATCH3 molecules demonstrated binding to ROR1-positive MDA-MB-231 cells (A, C), but no binding was found to ROR1-negative MCF-7 cells (B, D). PRO2213 was included as a reference anti-ROR1 antibody. Data points in the figure represent the average of two technical replicates. [Diagram 5] Concentration-response curves of anti-ROR1xhSAxCD3 scMATCH3 molecules PRO2507, PRO2508, PRO2509, PRO2510, and PRO2668 against human ROR1 expressed on MDA-MB-231 cells and against ROR1-negative MCF-7 cells by flow cytometry. All scMATCH3 molecules demonstrated binding to ROR1-positive MDA-MB-231 cells (A, B, C), but no binding was found to ROR1-negative MCF-7 cells (D). PRO2668, an optimized variant of PRO2510, demonstrated binding to MDA-MB-231 (E) cells, but no binding was found to ROR1-negative MCF-7 cells (F). PRO2213 was included as a reference anti-ROR1 antibody. Data points in the figure represent the average of two technical replicates. [Figure 6] ROR1 expression and binding site quantification of MDA-MB-231 and MCF7 cells are shown. (A) ROR1 expression levels were determined by ROR1 antibody staining of the indicated tumor cells followed by flow cytometry. Blue histograms indicate ROR1 expression, control is indicated in red. (B) ROR1 surface levels were quantified using the Quantum Simply Cellular anti-mouse IgG kit. Numbers indicate the calculated number of binding sites on the cell surface. [Figure 7]Cytotoxicity and CD8 T cell activation mediated by ROR1-targeted scMATCH3 molecules. (A) PBMCs were co-cultured with either MDA-MB-231 cells (top) or MCF7 cells (bottom) at an effector-to-target ratio of 10:1 for 40 hours. Controls include target cells alone (spontaneous lysis) and target cells treated with 1% Triton at the beginning or end of the experiment. (B) CD8 T cell activation and (C) CD4 T cell activation as indicated by the frequency of CD69+ cells from the dataset depicted in (A). One representative experiment of three is shown. [Figure 8] Cytotoxicity and CD8 T cell activation mediated by ROR1-targeted scMATCH3 molecules. (A) PBMCs were co-cultured with either MDA-MB-231 cells (top) or MCF7 cells (bottom) at an effector-to-target ratio of 10:1 for 40 hours. Controls include target cells alone (spontaneous lysis) and target cells treated with 1% Triton at the beginning or end of the experiment. (B) CD8 T cell activation and (C) CD4 T cell activation as indicated by the frequency of CD69+ cells from the data set depicted in (A). One experiment of two is shown. [Figure 9] A schematic diagram of the MATCH4 format is shown. [Figure 10(1)]Concentration-response curves of bivalent anti-ROR1 MATCH4 molecules PRO2589, PRO2590, PRO2591, and PRO2592 against human ROR1 expressed on MDA-MB-231 cells (A, B), against human ROR1 expressed on JIMT-1 cells (C, D), and against ROR1-negative MCF-7 cells (E) by flow cytometry are shown. All MATCH4 molecules demonstrated binding to ROR1-positive MDA-MB-231 cells (A, B) and ROR1-positive JIMT-1 cells (C, D), but no binding was found to ROR1-negative MCF-7 cells (E). PRO2590's optimized variant PRO2670 demonstrated binding to MDA-MB-231 (F) cells, but no binding was found to ROR1-negative MCF-7 cells (G). Fab fragment PRO2213 was included as a reference anti-ROR1 antibody. Data points in the figure represent the average of two technical replicates. [Figure 10(2)] Concentration-response curves of bivalent anti-ROR1 MATCH4 molecules PRO2589, PRO2590, PRO2591, and PRO2592 against human ROR1 expressed on MDA-MB-231 cells (A, B), against human ROR1 expressed on JIMT-1 cells (C, D), and against ROR1-negative MCF-7 cells (E) by flow cytometry are shown. All MATCH4 molecules demonstrated binding to ROR1-positive MDA-MB-231 cells (A, B) and ROR1-positive JIMT-1 cells (C, D), but no binding was found to ROR1-negative MCF-7 cells (E). PRO2590's optimized variant PRO2670 demonstrated binding to MDA-MB-231 (F) cells, but no binding was found to ROR1-negative MCF-7 cells (G). Fab fragment PRO2213 was included as a reference anti-ROR1 antibody. Data points in the figure represent the average of two technical replicates. [Figure 11] ROR1 antibody staining of the indicated tumor cells followed by flow cytometry determined ROR1 surface expression levels. Blue histograms indicate ROR1 expression, negative control is indicated in red. [Figure 12(1)]Cytotoxicity mediated by scMATCH3 and MATCH4 molecules of interest among several ROR1-expressing cell lines. Pan-T cells were co-cultured with MDA-MB-231 cells (A, upper left), JIMT-1 (A, lower left), SKOV-3 (A, lower right), or Jeko-1 cells (A, upper right) at an effector-to-target ratio of 10:1 for 40 hours. Controls include target cells alone (spontaneous lysis) and target cells treated with 1% Triton at the beginning or end of the experiment. MATCH4 molecules are PRO2589, PRO2590, PRO2591, and PRO2592, curves are color-coded by domain, scMATCH3 molecules are PRO2507, PRO2510, and PRO2557. One experiment out of four is shown. MCF7 were used as reference cells (B), i.e., ROR1-negative cells, and were treated as described for A. Error bars represent the variation between technical replicates. [Figure 12 (2)] Cytotoxicity mediated by scMATCH3 and MATCH4 molecules of interest among several ROR1-expressing cell lines. Pan-T cells were co-cultured with MDA-MB-231 cells (A, upper left), JIMT-1 (A, lower left), SKOV-3 (A, lower right), or Jeko-1 cells (A, upper right) at an effector-to-target ratio of 10:1 for 40 hours. Controls include target cells alone (spontaneous lysis) and target cells treated with 1% Triton at the beginning or end of the experiment. MATCH4 molecules are PRO2589, PRO2590, PRO2591, and PRO2592, curves are color-coded by domain, scMATCH3 molecules are PRO2507, PRO2510, and PRO2557. One experiment out of four is shown. MCF7 were used as reference cells (B), i.e., ROR1-negative cells, and were treated as described for A. Error bars represent the variation between technical replicates. [Figure 13(1)]Figure 1 shows T cell activation mediated by MATCH molecules of interest among several ROR1 expressing cell lines. Pan T cells were co-cultured with MDA-MB-231 (A), JIMT-1 (B), or SKOV-3 (C) cells at an effector-to-target ratio of 10:1 for 40 h. Supernatants were analyzed for LDH release and cells were analyzed by flow cytometry for the presence of activated T cells as indicated by the frequency of CD69 expressing cells. The upper graph corresponds to CD8 T cells and the lower to CD4 T cells. MATCH4 molecules are PRO2589, PRO2590, PRO2591, and PRO2592, curves are color coded by domain, scMATCH3 molecules are PRO2507, PRO2510, and PRO2557. One experiment out of three (SKOV-3) or four (MDA-MB-231, JIMT-1) is shown. Error bars represent the variation between technical replicates. [Figure 13(2)] Figure 1 shows T cell activation mediated by MATCH molecules of interest among several ROR1 expressing cell lines. Pan T cells were co-cultured with MDA-MB-231 (A), JIMT-1 (B), or SKOV-3 (C) cells at an effector-to-target ratio of 10:1 for 40 h. Supernatants were analyzed for LDH release and cells were analyzed by flow cytometry for the presence of activated T cells as indicated by the frequency of CD69 expressing cells. The upper graph corresponds to CD8 T cells and the lower to CD4 T cells. MATCH4 molecules are PRO2589, PRO2590, PRO2591, and PRO2592, curves are color coded by domain, scMATCH3 molecules are PRO2507, PRO2510, and PRO2557. One experiment out of three (SKOV-3) or four (MDA-MB-231, JIMT-1) is shown. Error bars represent the variation between technical replicates. [Figure 14(1)]The absorbance levels of pre-existing ADA in 21 human serum samples are shown for PRO2668 (A), PRO2669 (B), PRO2510 (C), and PRO2589 (D), as determined by the ELISA-based pre-existing ADA binding assay described in Example 10. Measurements were performed with spiked and unspiked serum samples (confirmatory assay setup). Additionally, the corresponding reduction (inhibition (%)) of absorbance levels of spiked human serum samples for said molecules is shown (E). [Figure 14(2)] The absorbance levels of pre-existing ADA in 21 human serum samples are shown for PRO2668 (A), PRO2669 (B), PRO2510 (C), and PRO2589 (D), as determined by the ELISA-based pre-existing ADA binding assay described in Example 10. Measurements were performed with spiked and unspiked serum samples (confirmatory assay setup). Additionally, the corresponding reduction (inhibition (%)) of absorbance levels of spiked human serum samples for said molecules is shown (E). [Figure 14(3)] The absorbance levels of pre-existing ADA in 21 human serum samples are shown for PRO2668 (A), PRO2669 (B), PRO2510 (C), and PRO2589 (D), as determined by the ELISA-based pre-existing ADA binding assay described in Example 10. Measurements were performed with spiked and unspiked serum samples (confirmatory assay setup). Additionally, the corresponding reduction (inhibition (%)) of absorbance levels of spiked human serum samples for said molecules is shown (E). [Figure 15(1)]Cytotoxicity and CD8 T cell activation mediated by ROR1-targeted scMATCH3 and scMATCH4 molecules. Pan-T cells were co-cultured with either MDA-MB-231 cells (A top) or JIMT-1 cells (A bottom) at an effector-to-target ratio of 10:1 for 40 hours. Controls include target cells alone (spontaneous lysis) and target cells treated with 1% Triton at the beginning or end of the experiment. CD8 T cell activation (B) as indicated by the frequency of CD8+CD69+ cells from the dataset depicted in A. One representative experiment of at least two independent experiments is shown. [Figure 15(2)] Cytotoxicity and CD8 T cell activation mediated by ROR1-targeted scMATCH3 and scMATCH4 molecules. Pan-T cells were co-cultured with either MDA-MB-231 cells (A top) or JIMT-1 cells (A bottom) at an effector-to-target ratio of 10:1 for 40 hours. Controls include target cells alone (spontaneous lysis) and target cells treated with 1% Triton at the beginning or end of the experiment. CD8 T cell activation (B) as indicated by the frequency of CD8+CD69+ cells from the dataset depicted in A. One representative experiment of at least two independent experiments is shown. [Figure 16(1)] Figure 2 shows that PRO2668 mediated T cell-mediated depletion cytotoxicity of exemplary solid tumor cell lines (A). Graph illustrating EC50 values ​​from different solid tumor cell lines in dependence of ROR1 receptor density (B). T cell-mediated depletion cytotoxicity of hematological tumor cell lines (C). Tumor cell lines were co-cultured with T cells for 40 h and cytotoxicity was assessed using lactate dehydrogenase release relative to control. The average number of cell surface ROR1 receptors and the average EC50 values ​​for cytotoxicity are shown in the graph insert. [Figure 16(2)]Figure 2 shows that PRO2668 mediated T cell-mediated depletion cytotoxicity of exemplary solid tumor cell lines (A). Graph illustrating EC50 values ​​from different solid tumor cell lines in dependence of ROR1 receptor density (B). T cell-mediated depletion cytotoxicity of hematological tumor cell lines (C). Tumor cell lines were co-cultured with T cells for 40 h and cytotoxicity was assessed using lactate dehydrogenase release relative to control. The average number of cell surface ROR1 receptors and the average EC50 values ​​for cytotoxicity are shown in the graph insert. [Figure 17] CD8 activation and proliferation mediated by PRO2668 on day 6 using MDA-MB-231 or JIMT-1 as solid tumor target cells (A). CD8 activation and proliferation mediated by Z-138 as hematological tumor target cells on day 3 (B). Targets and CellTrace Violet-labeled healthy T cells were co-cultured at an E:T ratio of 5:1. The frequency of proliferating cells was determined by CellTrace Violet dilution relative to the control. Activated cells were identified by CD25 labeling. One representative experiment of two is shown. [Figure 18(1)]Histograms of CLL cells and actively dividing Jeko-1 cells are shown (A). CLL cell samples do not proliferate. 5-ethynyl-2'-deoxyuridine (EdU) incorporation was assessed for the actively dividing Jeko-1 tumor cell line (dark grey histogram) compared to CLL patient PBMCs (light grey histogram). EdU incorporation indicates cell division. ROR1 surface expression and quantification is shown for CLL patient PBMCs (white histogram, relative to black negative control) (B). PRO2668 binding gated on CLL patient tumor cells (C). PBMCs from CLL patients were co-cultured with allogeneic healthy T cells at a 5:1 E:T ratio for 40 h. Specific killing (D) was determined by the percentage of Annexin V and Live / Dead dye positive CLL cells by treatment. T cells from D were stained for the activation marker CD69 (E). Supernatants from D were assessed for cytokine release using a cytometric bead array-based multiplexing system (F). Data shown are from one of four independent donors / experiments. [Figure 18(2)]Histograms of CLL cells and actively dividing Jeko-1 cells are shown (A). CLL cell samples do not proliferate. 5-ethynyl-2'-deoxyuridine (EdU) incorporation was assessed for the actively dividing Jeko-1 tumor cell line (dark grey histogram) compared to CLL patient PBMCs (light grey histogram). EdU incorporation indicates cell division. ROR1 surface expression and quantification is shown for CLL patient PBMCs (white histogram, relative to black negative control) (B). PRO2668 binding gated on CLL patient tumor cells (C). PBMCs from CLL patients were co-cultured with allogeneic healthy T cells at a 5:1 E:T ratio for 40 h. Specific killing (D) was determined by the percentage of Annexin V and Live / Dead dye positive CLL cells by treatment. T cells from D were stained for the activation marker CD69 (E). Supernatants from D were assessed for cytokine release using a cytometric bead array-based multiplexing system (F). Data shown are from one of four independent donors / experiments. [Figure 19] Longitudinal tumor growth inhibition (median) in the presence of different doses of scMATCH3 PRO2668 and scMATCH4 PRO2670 molecules. Immunodeficient mice were implanted subcutaneously with Jeko-1 cells, followed by PBMC engraftment 3 days later. After tumor volumes reached 80-100 mm3, animals were randomized and dosing was initiated. Dosing was performed every 5 days. Longitudinal data were analyzed using a mixed effects model followed by Tukey's multiple comparison test. ns: not significant, **p ≤ 0.01, *p ≤ 0.05, relative to control IgG (palivizumab) at the end of the experiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0092] Many hematological and solid cancers express ROR1, making ROR1-expressing cancers attractive targets for cancer therapy. However, existing bispecific anti-ROR1×CD3 therapeutic antibodies developed so far generally do not have the expected breadth of efficacy against these cancer types. In particular, these therapeutic bispecific anti-ROR1×CD3 antibodies often have moderate potency against cancer cells expressing low levels of ROR1. Moreover, these prior art antibodies typically exhibit a low percentage of maximal killing, often well below 50%, against many ROR1-expressing cancer cells in cytotoxicity assays, indicating that their mode of action is not ideal. Therefore, there is a need in the medical field for improved anti-ROR1×CD3-based therapeutics with higher efficacy to address a broad range of ROR1-expressing cancers, including those expressing ROR1 at moderate to low levels.

[0093] The present invention provides multispecific antibodies comprising one or two binding domains that specifically bind to the extracellular domain of ROR1 (ROR1-BD) and one binding domain that specifically binds to CD3 (CD3-BD), where the CDR regions of ROR1-BD and CD3-BD have the definitions given above. The multispecific antibodies of the present invention can potently kill cancer cells that express high levels of ROR1 and cancer cells that express ROR1 only at low levels. Some of the multispecific antibodies of the present invention, such as PRO2589 and PRO2591, have sub-pM EC for killing high ROR1-expressing JEKO-1 cells. 50 Several of these multispecific antibodies, such as PRO2589 and PRO2590, demonstrated EC values ​​of less than 50 pM for killing low ROR1-expressing SKOV-3 ovarian cancer cells as determined by a T cell-driven cytotoxicity assay against the target cells. 50 (e.g., Figures 12 and 15).

[0094] Furthermore, the multispecific antibodies of the present invention comprising one ROR1-BD as defined herein show maximal killing of 70% or more of different ROR1-expressing cancer cells as determined by T cell-driven cytotoxicity assay against said cancer cells. For example, PRO2510 shows maximal killing of about 80% for MDA-MB-231, JEKO-1, JIMT-1 and SKOV-3 cancer cells expressing high to low levels of ROR1 (e.g., Figures 12 and 15).

[0095] In addition, the anti-ROR1xCD3 multispecific antibodies of the invention exhibit very advantageous biophysical properties, in particular good storage stability.

[0096] Therefore, the multispecific antibodies of the present invention offer distinct therapeutic advantages over conventional therapies.

[0097] 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.

[0098] The terms "comprise" and "comprising" are used herein in their open and non-limiting sense unless otherwise noted. Thus, for such latter embodiments, the term "comprise" encompasses the narrower term "consisting of."

[0099] Unless otherwise indicated herein or clearly contradicted by context, the terms "a" and "an" and "the" and similar references should be construed in the context of describing the present invention (especially in the context of the following claims) to cover both the singular and the plural. For example, the term "a cell" encompasses a plurality of cells, including mixtures thereof. Where the plural is used with respect to compounds, salts, and the like, this is taken to mean also a single compound, salt, or the like.

[0100] In one embodiment, the present invention provides a method for producing a method for treating a pulmonary arthritis, comprising: a) one or two binding domains (ROR1-BD) that specifically bind to the extracellular domain of ROR1; and b) one binding domain that specifically binds to CD3 (CD3-BD) Including, Here, said ROR1-BD and said CD3-BD relate to a multispecific antibody comprising the CDR sequences defined above.

[0101] In a further aspect, the present invention provides a method for producing a method for the treatment of a cancer a) one or two binding domains (ROR1-BD) that specifically bind to the extracellular domain of ROR1; and b) one binding domain that specifically binds to CD3 (CD3-BD) Including, wherein said ROR1-BD and said CD3-BD comprise the CDR sequences defined above, and wherein the multispecific antibody does not comprise an immunoglobulin Fc region.

[0102] The term "antibody" and the like as used herein includes 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 (including, without limitation, 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 is composed of a heavy chain variable region (abbreviated herein as VH) and a heavy chain constant region. The heavy chain constant region is composed of three domains, CH1, CH2, and CH3. Each light chain is composed of a light chain variable region (abbreviated herein as VL) and a light chain constant region. The light chain constant region is composed of one domain, CL. The VH and VL regions can be further subdivided into regions of hypervariability called complementarity determining regions (CDRs), flanked by more conserved regions called framework regions (FRs). Each VH and VL is composed of three CDRs and four FRs arranged in the following order from amino terminus to carboxy terminus: 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 region of the antibody can mediate the binding of immunoglobulins 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.

[0103] As used herein, the term "immunoglobulin Fc region" or "Fc region" is used to define the C-terminal region of an immunoglobulin heavy chain, i.e., the CH2 and CH3 domains of the heavy chain constant region. The term "Fc region" encompasses native sequence Fc regions and variant Fc regions, i.e., Fc regions that have been engineered to exhibit certain desired properties, such as modulated Fc receptor binding function and / or reduced or inhibited Fab arm exchange. An example of such an engineered Fc region is the Knobs-into-Hole (KiH) technology (see, e.g., Ridgway et al., Protein Engineering, Design and Selection 9:617-21 (1996) and Spiess et al., Journal of Biological Chemistry 288(37):26583-93 (2013)). Native sequence Fc regions include human lgG1, lgG2 (lgG2A, IgG2B), lgG3, and lgG4. "Fc receptor" or "FcR" refers to a receptor that binds to the Fc region of an antibody. In particular, FcR is a native sequence human FcR (gamma receptor) that binds IgG antibodies and includes receptors of the FcγRI, FcyRII, and FcγRIII subclasses, including allelic variants and alternatively spliced ​​forms of these receptors. FcγRII receptors include FcγRIIA ("activating receptors") and FcγRIIB ("inhibiting receptors"), which have similar amino acid sequences that differ primarily in their cytoplasmic domains. Activating receptor FcγRIIA contains an immunoreceptor tyrosine-based activation motif (ITAM) in its cytoplasmic domain. The inhibitory receptor FcγRIIB contains an immunoreceptor tyrosine-based inhibition motif (ITIM) in its cytoplasmic domain (see M. Daeron, Annual Review of Immunology, 5:203-234 (1997)).FcRs have been reviewed in Ravetch and Kinet, Annual Review of Immunology, 9:457-92 (1991); Capet et al., ImmunoMethods, 4:25-34 (1994); and deHaas et al., Journal of Laboratory and Clinical Medicine, 126:330-41 (1995). Other FcRs, including those to be identified in the future, are encompassed by the term "FcR" herein. The term "Fc receptor" or "FcR" also encompasses the neonatal receptor, FcRn, which is responsible for the transfer of maternal IgG to the fetus. Guyer et al., Journal of Immunology, vol. 117:587 (1976), and Kim et al., Journal of Immunology, vol. 24:249 (1994). Methods for measuring binding to FcRn are known (e.g., Ghetie and Ward, Immunology Today, vol. 18:(12):592-8 (1997); Ghetie et al., Nature Biotechnology, vol. 15:637-40 (1997); Hinton et al., Journal of Biological Chemistry, vol. 10:111-115 (1997)). See, Chemistry, TJI (Vol. 8): p. 6213-6 (2004); WO2004 / 92219 (Hinton et al.). The in vivo binding to FcRn and serum half-life of human FcRn high affinity binding polypeptides can be assayed, for example, in transgenic mice or transfected human cell lines expressing human FcRn, or in primates to which the polypeptides having variant Fc regions are administered.WO2004 / 42072 (Presta) describes antibody variants with improved or diminished binding to FcR. See also, e.g., Shields et al., Journal of Biological Chemistry, 9(2):6591-6604 (2001).

[0104] The terms "binding domain", "antigen-binding fragment thereof", "antigen-binding portion" and the like of an antibody as used herein refer to one or more fragments of an intact antibody that retain the ability to specifically bind to a given antigen (e.g., ROR1, CD3, hSA). The antigen-binding function of an antibody can be performed by a fragment of an intact antibody. Specifically, in the case of the multispecific antibodies of the present invention, the terms "binding domain", "antigen-binding fragment thereof", "antigen-binding portion" and the like as used herein refer to Fab fragments, i.e., monovalent fragments consisting of the VL, VH, CL, and CH1 domains, F(ab)2 fragments, i.e., bivalent fragments comprising two Fab fragments linked by a disulfide bridge at the hinge region, Fv fragments consisting of the VL and VH domains of a single arm of an antibody, disulfide-stabilized Fv fragments (dsFv), single-chain Fv fragments (scFv) and scAbs, i.e., single chains consisting of the VL, VH, and CL domains of a single arm of an antibody. Preferably, the binding domains of the antibodies of the invention are independently selected from Fv fragments, scFv fragments, and single chain Fv fragments (scFv). In a particular embodiment, the binding domains of the antibodies of the invention are independently selected from single chain Fv fragments (scFv). In another particular embodiment, the VL and VH domains of the scFv fragment are stabilized by an inter-domain disulfide bond, in particular, said VH sequence comprises a single cysteine ​​residue at position 51 (AHo numbering) and said VL sequence comprises a single cysteine ​​residue at position 141 (AHo numbering).

[0105] In one embodiment, the multispecific antibodies of the invention do not comprise a CH1 and / or CL region. In a particular embodiment, the multispecific antibodies of the invention do not comprise an immunoglobulin Fc region or a CH1 and CL region.

[0106] The term "complementarity determining region" ("CDR") refers to an amino acid sequence having boundaries determined using any of a number of well-known schemes, including those described by Kabat et al. (1991), Sequences of Proteins of Immunological Interest, 5th ed., Public Health Service, National Institutes of Health, Bethesda, MD ("Kabat" numbering scheme); Al-Lazikani et al. (1997), Journal of Molecular Biology, Vol. 273, pp. 927-948 ("Chothia" numbering scheme); ImMunoGenTics (IMGT) numbering (Lefranc, M.-P., The Immunologist, vol. 11, no. 112, pp. 1111-1125, 2002). Immunologist, Vol. 7, pp. 132-136 (1999); Lefranc, M.-P. et al., Developmental & Comparative Immunology, Vol. 27, pp. 55-77 (2003) ("IMGT" numbering scheme); and the numbering scheme described in Honegger & Pluckthun, Journal of Molecular Biology, Vol. 309, pp. 657-670 (2001) ("AHo" numbering). For example, in the classical format, in Kabat, the CDR amino acid residues on the heavy chain variable domain (VH) are numbered 31-35 (HCDR1), 50-65 (HCDR2), and 95-102 (HCDR3); the CDR amino acid residues on the light chain variable domain (VL) are numbered 24-34 (LCDR1), 50-56 (LCDR2), and 89-97 (LCDR3). In Chotia, the CDR amino acids on the VH are numbered 26-32 (HCDR1), 52-56 (HCDR2), and 95-102 (HCDR3); the amino acid residues on the VL are numbered 24-34 (LCDR1), 50-56 (LCDR2), and 89-97 (LCDR3).Combining the CDR definitions of both Kabat and Chotias, the CDRs consist of amino acid residues 26-35 (HCDR1), 50-65 (HCDR2), and 95-102 (HCDR3) on the human VH, and amino acid residues 24-34 (LCDR1), 50-56 (LCDR2), and 89-97 (LCDR3) on the human VL. In IMGT, the CDR amino acid residues on the VH are numbered approximately 26-35 (HCDR1), 51-57 (HCDR2), and 93-102 (HCDR3), and the CDR amino acid residues on the VL are numbered approximately 27-32 (LCDR1), 50-52 (LCDR2), and 89-97 (LCDR3) (numbering according to "Kabat"). In IMGT, the CDRs of an antibody can be determined using the program IMGT / DomainGapAlign.

[0107] In the context of the present invention, unless specifically stated otherwise, the numbering system suggested by Honegger & Pluckthun ("AHo") is used (Honegger & Pluckthun, Journal of Molecular Biology, 309 (2001), 657-670). 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 according to Honegger & Pluckthun takes into account the length diversity found in naturally occurring antibodies, both in different VH and VL subfamilies, especially in the CDRs, and allows for sequence gaps. Thus, in a given antibody variable domain, normally not all positions 1 to 149 will be occupied by an amino acid residue.

[0108] 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. The term "specifically bind" or "specific" as used herein refers to a measurable and reproducible interaction, such as the binding of an antibody with a target, that determines the presence of the target in the presence of a heterogeneous population of molecules, including biological molecules. For example, an antibody that specifically binds to a target (which may be an epitope) is an antibody that binds to this target with greater affinity, avidity, more readily, and / or with greater duration than it binds to other targets. In its most general form (and when no set reference is made), "specific binding" refers to the ability of an antibody to distinguish between the target of interest and unrelated molecules, as determined, for example, according to a specificity assay method known in the art. Such methods include, but are not limited to, Western blot, ELISA, RIA, ECL, IRMA, SPR (surface plasmon resonance) test, and peptide scan. For example, a standard ELISA assay can be performed. Scoring can be performed by standard color development (e.g. tetramethylbenzidine with secondary antibody and hydrogen peroxide with horseradish peroxidase). The reaction of a particular well is scored by optical density at, for example, 450 nm. A typical background (=negative reaction) can be about 0.1 OD. A typical positive reaction can be about 1 OD. This means that the ratio between positive and negative scores can be 10-fold or more. In a further example, an SPR assay can be performed, where a difference between background and signal of at least 10-fold, in particular at least 100-fold, indicates specific binding. Typically, the determination of binding specificity is performed by using not a single reference molecule, but a set of about 3 to 5 unrelated molecules, such as milk powder, transferrin, or the like.

[0109] Preferably, the antibody of the present invention is an isolated antibody. The term "isolated antibody" as used herein refers to an antibody that is substantially free of other antibodies with different antigen specificity (e.g., an isolated antibody that specifically binds to ROR1 and CD3 is substantially free of antibodies that specifically bind to antigens other than ROR1 and CD3, or an isolated antibody that specifically binds to ROR1, CD3, and hSA is substantially free of antibodies that specifically bind to antigens other than ROR1, CD3, and hSA). Moreover, an isolated antibody can be substantially free of other cellular materials and / or chemicals.

[0110] Preferably, the antibody of the present invention is a monoclonal antibody. As used herein, the term "monoclonal antibody" refers to an antibody that has substantially the same amino acid sequence or is derived from the same genetic source. A monoclonal antibody exhibits binding specificity and affinity to a specific epitope, or binding specificity and affinity to a specific epitope.

[0111] The antibodies of the present invention include, but are not limited to, chimeric antibodies, human antibodies, and humanized antibodies.

[0112] The term "chimeric antibody" (or antigen-binding fragment thereof) as used herein refers to (a) an antibody molecule (or antigen-binding fragment thereof) in which the constant region or a portion thereof has been modified, replaced, or exchanged such that the antigen-binding site (variable region) is linked to a constant region of a different or modified class, effector function, and / or species, or (b) an antibody molecule (or antigen-binding fragment thereof) in which the variable region or a portion thereof has been modified, replaced, or exchanged by a variable region having a different or modified antigen specificity. For example, a mouse antibody can be modified by replacing its constant region with a constant region from a human immunoglobulin. Due to the replacement with a human constant region, the chimeric antibody can retain its specificity in recognizing the antigen while having a reduced antigenicity in humans compared to the original mouse antibody.

[0113] 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 human sequences, e.g., human germline sequences or mutated versions of human germline sequences. Human antibodies and antigen-binding fragments thereof of the present invention may include amino acid residues not encoded by human sequences (e.g., mutations introduced by in vitro random or site-specific mutagenesis or by in vivo somatic mutation). This definition of a human antibody specifically excludes humanized antibodies that contain 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, Journal of Molecular Biology, 227:381 (1992); Marks et al., Journal of Molecular Biology, 222:581 (1991)). Methods described in Cole et al., Monoclonal Antibodies and Cancer Therapy, Alan R. Liss, p. 77 (1985); Boemer et al., Journal of Immunology, vol. 147(l): pp. 86-95 (1991) are also available for the preparation of human monoclonal antibodies. See also van Dijk and van de Winkel, Current Opinion in Pharmacology, vol. 5: pp. 368-74 (2001).Human antibodies can be prepared by administering antigen to transgenic animals, such as immunized XenoMouse, that have been engineered to produce such antibodies in response to antigen challenge, but whose endogenous loci have been disabled (see, e.g., US 6,075,181 and US 6,150,584 for XenoMouse™ technology). See also, e.g., Li et al., Proceedings of the National Academy of Sciences of the United States of America, vol. 103:3557-3562 (2006), for human antibodies produced by human B-cell hybridoma technology.

[0114] The term "humanized" antibody (or antigen-binding fragment thereof) as used herein refers to an antibody (or antigen-binding fragment thereof) that is less immunogenic in humans while retaining the reactivity of non-human antibodies. This can be achieved, for example, by retaining non-human CDR regions and replacing the remaining parts of the antibody with their human counterparts (i.e., constant and variable region framework portions). Additional framework region modifications can be made on the human framework sequences and on the CDR sequences derived from the germline of another mammalian species. The humanized antibody of the present invention can include amino acid residues that are not encoded by human sequences (e.g., mutations introduced by in vitro random or site-specific mutagenesis or by in vivo somatic mutation, or conservative substitutions to promote stability or manufacturing). See, e.g., Morrison et al., Proceedings of the National Academy of Sciences, Vol. 81:6851-6855, 1984; Morrison and Oi, Advances in Immunology, Vol. 44:65-92, 1988; Verhoeyen et al., Science, Vol. 239:1534-1536, 1988; Padlan, Molecular Immunology, Vol. 28:489-498, 1991; and Padlan, Molecular Immunology, Vol. 31:169-217, 1994. Other examples of human engineering technologies include, but are not limited to, the Xoma technology disclosed in US Pat. No. 5,766,886.

[0115] The term "recombinant humanized antibody" as used herein includes all human antibodies that are prepared, expressed, created or isolated by recombinant means, e.g., antibodies isolated from a host cell that has been transformed to express the humanized antibody, e.g., from a transfectoma, and antibodies prepared, expressed, created or isolated by any other means involving splicing of all or part of the human immunoglobulin gene sequence to other DNA sequences.

[0116] Preferably, the multispecific antibodies of the invention are humanized, more preferably, the multispecific antibodies of the invention are humanized and comprise rabbit-derived CDRs.

[0117] 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., ROR1 and CD3). The term "multispecific antibody" includes bispecific, trispecific, tetraspecific, pentaspecific, and hexaspecific. Preferably, the multispecific antibody of the present invention is a bispecific antibody or a trispecific antibody, particularly a trispecific antibody. The term "bispecific antibody" as used herein refers to an antibody that binds to at least two different epitopes on two different targets (e.g., ROR1 and CD3). The term "trispecific antibody" as used herein refers to an antibody that binds to at least three different epitopes on three different targets (e.g., ROR1, CD3, and hSA).

[0118] 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" and "linear" epitopes are distinguished in that the binding to the former but not the latter is lost in the presence of denaturing solvents.

[0119] The term "conformational epitope" as used herein refers to amino acid residues of an antigen that come together on the surface when the polypeptide chain folds to form the native protein.

[0120] The term "linear epitope" refers to an epitope in which all of the points of interaction between a protein and an interacting molecule (eg, an antibody) occur linearly along the primary amino acid sequence of the (continuous) protein.

[0121] The term "recognize" as used herein refers to that antigen-binding fragment of an antibody that finds and interacts with (eg, binds to) that conformational epitope.

[0122] The term "avidity" as used herein refers to a measure that provides information on the overall stability or strength of antibody-antigen complex. It is controlled by three main factors: antibody epitope affinity, the valency of both antigen and antibody, and the structural arrangement of the interacting parts. Ultimately, these factors determine the specificity of the antibody, that is, the probability that a particular antibody will bind to a precise antigen epitope.

[0123] The multispecific antibodies of the present invention comprise one or two ROR1-BDs.

[0124] The term "ROR1" refers in particular to human ROR1 having UniProt ID number Q01973. Advantageously, the ROR1-BD of the present invention targets human ROR1, in particular the extracellular domain (ECD) of ROR1, the sequence of which is shown in Table 6 (SEQ ID NO: 117). In a particular embodiment, the ROR1-BD binds to the Ig-like domain of ROR1. The Ig-like domain of ROR1 is defined by residues 42 to 147 of SEQ ID NO: 117 in accordance with UniProt ID number Q01973. In another particular embodiment, the ROR1-BD does not block the binding of Wnt5α to ROR1.

[0125] Suitably, ROR1-BD, when in scFv format, is characterised by one or more of the following parameters: a. A monovalent dissociation constant (K) between 1 pM and 2 nM as measured by surface plasmon resonance (SPR) D ) with a K of 1 pM to 1 nM, especially 1 to 500 pM D and binds to the extracellular domain of human ROR1. b.EC50 between 5 pM and 10 nM 50 , especially with an EC of 5 pM to 5 nM 50 , especially with an EC of 5 pM to 4 nM 50 , especially with an EC of 5 pM to 3 nM 50 and binds to human ROR1-expressing MDA-MB-231 cells.

[0126] As used herein, the term "MDA-MB-231 cells" refers to a breast cancer cell line that expresses ROR1. MDA-MB-231 cells are commercially available and serve as a model cell line for breast cancer (adenocarcinoma).

[0127] The term "affinity" as used herein refers to the strength of the interaction between an antibody and an antigen at a single antigenic site. At each antigenic site, the variable region of an antibody "arm" interacts with the antigen at multiple sites through weak non-covalent forces. The more interactions, the stronger the affinity.

[0128] "Binding affinity" generally refers to the strength of the total non-covalent interactions between a single binding site of a molecule (e.g., an antibody) and its binding partner (e.g., an antigen, or more specifically, an epitope on an antigen). Unless otherwise indicated, "binding affinity," "binds to," "binds to," or "binding to," as used herein, refer to the inherent binding 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 conventional 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. Various methods for measuring binding affinity are known in the art, any of which may be used for the purposes of the present invention. Specific illustrative and exemplary embodiments for measuring binding affinity, i.e., binding strength, are described below.

[0129] As used herein, the term "K assoc ", "K a " or "K on " is intended to refer to the association rate of a particular antibody-antigen interaction, and the term "K dis ", "K d " or "K off " is intended to refer to the off-rate of a particular antibody-antigen interaction. In one embodiment, the term "K D " is intended to refer to the dissociation constant, which is K a K for d The ratio of (i.e., K d / K a ) and expressed as molar concentration (M). D " or "K D "KD value" or "KD value" is measured in one embodiment by using a surface plasmon resonance assay.

[0130] Affinity for recombinant human ROR1 was determined by surface plasmon resonance (SPR) measurements as described in paragraphs

[0176] and

[0192] (scFv), and

[0219] and

[0242] to

[0244] (multispecific molecules). Affinity for recombinant human CD3, recombinant cynomolgus CD3, and recombinant marmoset CD3 was determined by surface plasmon resonance (SPR) measurements as described in paragraphs

[0208] (scFv), and

[0221] and

[0247] (multispecific molecules). Affinity for recombinant human serum albumin (hSA), cynomolgus monkey serum albumin (cSA), and mouse serum albumin (mSA) was determined by surface plasmon resonance (SPR) measurements as described in paragraphs

[0213] (scFv), and

[0223] and

[0249] (multispecific molecules).

[0131] In some embodiments, the ROR1-BD of the invention or the ROR1-BD present in a multispecific antibody of the invention does not block binding of Wnt5α to ROR1.

[0132] Advantageously, the ROR1-BD of the multispecific antibody of the invention, when in scFv format, is further characterized by one or more of the following parameters: c. a melting temperature (Tm) as determined by differential scanning fluorimetry (DSF) of at least 58°C, particularly at least 59°C, particularly at least 60°C, particularly at least 61°C, particularly wherein said scFv is formulated in 50 mM phosphate-citrate buffer with 150 mM NaCl at pH 6.4; d. the scFv has a loss of monomer content after 4 weeks of storage at 4° C. of less than 3%, particularly less than 2%, particularly less than 1%, when the scFv is at a starting concentration of 10 mg / ml, particularly where the scFv is formulated in 50 mM phosphate-citrate buffer with 150 mM NaCl at pH 6.4, and / or e. the scFv has a loss of monomer content after 4 weeks of storage at 40° C. of less than 10% when the scFv is at a starting concentration of 10 mg / ml, particularly where the scFv is formulated in 50 mM phosphate-citrate buffer with 150 mM NaCl at pH 6.4, and / or f. The scFv has less than 1% loss in protein content after 4 weeks of storage at 4° C. or 40° C. when the scFv is at a starting concentration of 10 mg / ml, particularly where the scFv is formulated in 50 mM phosphate-citrate buffer with 150 mM NaCl at pH 6.4.

[0133] For most of the measurements, the nanoDSF (Prometheus NT.48, NaonTemper) method was used to determine the thermodynamic properties of scFvs and multispecific antibodies, e.g., as described in paragraphs

[0197] and

[0206] (scFvs) and paragraphs

[0234] and

[0261] (multispecific formats). In the case of CD3-BD (scFv), the DSF method was used as described in Egan et al., MAbs, vol. 9(1) (2017), p. 68-84; Niesen et al., Nature Protocols, vol. 2(9) (2007), p. 2212-2221. In this method, the transition midpoint of thermal unfolding of scFv constructs is determined by differential scanning fluorimetry using the fluorescent dye SYPRO® Orange (see Wong & Raleigh, Protein Science 25 (2016) 1834-1840). Samples are prepared at a final protein concentration of 50 μg / ml in pH 6.4 phosphate-citrate buffer and contain a final concentration of 5× SYPRO® Orange in a total volume of 100 μl. 25 microliters of prepared sample are added in triplicate to a white-walled AB Gene PCR plate. The assay is performed by a qPCR machine used as a thermal cycler and fluorescence emission is detected using a custom dye calibration routine in the software. The PCR plate containing the test samples is subjected to a temperature ramp from 25° C. to 96° C. in 1° C. increments with a 30 s pause after each temperature increment. The total assay time is approximately 2 h. Tm was calculated using the mathematical second derivative method with the software GraphPad Prism to calculate the inflection point of the curve. The reported Tm is the average of three measurements.

[0134] The loss of monomer content is determined by area under the curve calculation of the SE-HPLC chromatogram. SE-HPLC is a separation technique based on a solid stationary phase and a liquid mobile phase as outlined by the United States Pharmacopeia (USP) Chapter 621. This method separates molecules based on their size and shape and utilizes a hydrophobic stationary phase and an aqueous mobile phase. The separation of molecules is determined by the void volume (V0) and total permeate volume (V T ) occurs between the chromatograms. Measurements by SE-HPLC are performed by a Chrommaster HPLC system (Hitachi High-Technologies Corporation) equipped with automatic sample injection and a UV detector set at a detection wavelength of 280 nm. The instrument is controlled by the software EZChrom Elite (Agilent Technologies, version 3.3.2, SP2), which also supports the analysis of the resulting chromatograms. Protein samples are clarified by centrifugation and kept at a temperature of 4-6 °C on the autosampler prior to injection. For the analysis of scFv samples, a Shodex KW403-4F column (Showa Denko Inc., #F6989202) is used with a standardized buffered saline mobile phase (50 mM sodium phosphate pH 6.5, 300 mM sodium chloride) at a recommended flow rate of 0.35 ml / min. The target sample load per injection was 5 μg. Samples are detected by a UV detector at a wavelength of 280 nm and data are recorded by a suitable software suite. The resulting chromatogram is V0 to V T range, thereby eliminating matrix-related peaks with elution times >10 min.

[0135] Suitably, the ROR1-BD of the multispecific antibody of the invention is a binding domain provided in the present disclosure. The ROR1-BD of the multispecific antibody of the invention includes, but is not limited to, the humanized ROR1-BD whose sequence is listed in Table 1.

[0136] In addition, the multispecific antibodies of the invention comprise one CD3-BD. In a particular embodiment, said CD3-BD binds to CD3ε. In particular, the multispecific antibodies of the invention comprise one CD3BD that targets human and cynomolgus (Macaca fascicularis) CD3ε.

[0137] Suitably, the CD3-BD used in the present invention, when in scFv format, is characterised by one or more of the following parameters: a) Dissociation constant (K) of less than 50 nM as measured by SPR D ) with a K of 0.1 to 50 nM, in particular 0.1 to 30 nM, in particular 0.1 to 15 nM, in particular 0.1 to 10 nM D , which binds to human CD3ε, b) K<50 nM as measured by SPR D and especially a K of 0.1 to 50 nM, especially 0.1 to 30 nM, especially 0.1 to 15 nM, especially 0.1 to 10 nM. D , binds to Macaca fascicularis (cynomolgus monkey) CD3, and c) has a melting temperature (Tm) as determined by differential scanning fluorimetry (DSF) of at least 60°C, particularly at least 65°C, and more particularly at least 67°C.

[0138] Suitably, the CD3-BD of the multispecific antibody of the invention is a binding domain as provided in this disclosure, including but not limited to the humanized CD3-BD whose sequence is listed in Table 2.

[0139] To increase the number of specificities / functionalities at the same or lower molecular weight, it is advantageous to use antibodies including antibody fragments such as Fv and Fab fragments and 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. Although such fragments appear to exhibit several advantages compared to whole immunoglobulins, they also suffer from an increased clearance rate from serum because they lack the Fc domain that confers a long half-life in vivo (Medasan et al., 1997, Journal of Immunology, vol. 158: pp. 2211-2217). Molecules with lower molecular weights penetrate into target tissues (e.g. solid tumors) more efficiently and therefore hold the promise of improved efficacy at the same or lower doses.

[0140] Here, we have 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. To the extent that it cannot be predicted a priori that all three or four binding domains will remain functional without sterically or otherwise inhibiting each other in a complex multi-target, multi-cellular in vivo context, this finding is surprising.

[0141] The term "hSA" specifically refers to human serum albumin with UniProt ID number P02768. Human serum albumin (hSA) is an abundant protein of 66.4 kDa (50% of total protein) in human serum composed of 585 amino acids (Sugio, Protein Engineering, Design and Selection, Vol. 12, 1999, p. 439-446). The multifunctional hSA protein has been associated with a structure that allows it to bind and transport several metabolites, such as fatty acids, metal ions, bilirubin, and some drugs (Fanali, Molecular Aspects of Medicine, Vol. 33, 2012, p. 209-290). The concentration of hSA in serum is around 3.5-5 g / dl. Albumin-binding antibodies and fragments thereof can be used, for example, to extend the in vivo serum half-life of drugs or proteins conjugated to them.

[0142] In some embodiments, the hSA-BD is derived from a monoclonal antibody or antibody fragment.

[0143] Suitable hSA-BDs comprised in the multispecific antibodies of the invention are the binding domains provided in this disclosure. hSA-BDs comprised in the multispecific antibodies of the invention include, but are not limited to, the humanized hSA binding domains whose sequences are listed in Table 3.

[0144] In a particular embodiment, the hSA-BD comprised in the multispecific antibody of the present invention specifically binds to hSA and cynomolgus serum albumin (cSA).

[0145] In other specific embodiments, the hSA-BD comprised in the multispecific antibody of the invention specifically binds to hSA, cSA, and mouse serum albumin (mSA). In these specific embodiments, the hSA-BD comprised in the multispecific antibody of the invention, when in scFv format, is characterized by one or more of the following parameters: a) Dissociation constants (K) of 0.1 to 10 nM, in particular 0.1 to 5 nM, measured by SPR at pH 5.5 and pH 7.4 D ) and binds to human serum albumin (hSA), b) Dissociation constants (K) of 0.1 to 10 nM, particularly 0.1 to 5 nM, measured by SPR at pH 5.5 and pH 7.4 D ) which binds to Macaca fascicularis (crab-eater) serum albumin (cSA) and c) dissociation constants (K) of 0.1 to 20 nM, particularly 0.1 to 10 nM, measured by SPR at pH 5.5 and pH 7.4 D ) and binds to mouse serum albumin (mSA).

[0146] Other suitable hSA-BDs for use in the multispecific antibodies of the invention include (i) polypeptides that bind serum albumin (see, e.g., Smith et al., 2001, Bioconjugate Chemistry, vol. 12: 750-756; EP 0486525; US 6,267,964; WO 2004 / 001064; WO 2002 / 076489; and WO 2001 / 45746); (ii) polypeptides that bind serum albumin (see, e.g., Holt et al., Protein Engineering, Design & Selection, vol. 12: 750-756; EP 0486525; US 6,267,964; WO 2004 / 001064; WO 2002 / 076489; and WO 2001 / 45746); Selection), vol 21, issue 5, pp 283-288, WO2004 / 003019, WO2008 / 096158, WO2005 / 118642, WO2006 / 0591056, and WO2011 / 006915; (iii) the anti-serum albumin antibodies described in WO2009 / 040562, WO2010 / 035012, and WO2011 / 086091.

[0147] In certain embodiments, the multispecific antibodies of the invention comprise (i) one ROR1-BD, (ii) one CD3-BD, and (iii) one hSA-BD, i.e., the multispecific antibodies of these particular embodiments are monovalent for all three ROR1, CD3, and hSA specificities.

[0148] In further particular embodiments, the multispecific antibodies of the invention comprise (i) two ROR1-BDs, (ii) one CD3-BD, and (iii) one hSA-BD, i.e., the multispecific antibodies of these particular embodiments are bivalent for the ROR1 specificity and monovalent for both the CD3 and hSA specificities.

[0149] The term "multivalent antibody" refers to a single binding molecule with more than one valency, where "valency" is described as the number of antigen-binding moieties that bind to an epitope on a target molecule. Thus, a single binding molecule can bind to more than one binding site on a target molecule and / or to more than one target molecule due to the presence of more than one copy of the corresponding antigen-binding moiety. Examples of multivalent antibodies include, but are not limited to, bivalent antibodies, trivalent antibodies, tetravalent antibodies, pentavalent antibodies, hexavalent antibodies, and the like.

[0150] As used herein, the term "monovalent antibody" refers to an antibody that binds to a single target molecule, more specifically, to a single epitope on a target molecule. Also, as used herein, the term "binding domain" or "monovalent binding domain" refers to a binding domain that binds to a single epitope on a target molecule.

[0151] In the case where the multispecific antibody of the present invention comprises two ROR1-BDs, the two ROR1-BDs bind to either the same epitope or different epitopes on the extracellular domain of ROR1. Preferably, the two ROR1-BDs bind to the same epitope on the extracellular domain of ROR1.

[0152] The term "same epitope" as used herein refers to individual protein determinants on a protein capable of specific binding to more than one antibody, where for each of said antibodies, the individual protein determinants are identical, i.e., consist of the same chemically active surface groups of molecules such as amino acids or sugar side chains that have the same three-dimensional structural characteristics and the same charge characteristics. The term "different epitopes" as used herein in the context of specific protein targets refers to individual protein determinants on a protein, where each of the individual protein determinants is capable of specific binding to a different antibody, where these individual protein determinants are not identical for the different antibodies, i.e., consist of non-identical chemically active surface groups of molecules such as amino acids or sugar side chains that have different three-dimensional structural characteristics and different charge characteristics. These different epitopes may or may not overlap.

[0153] In certain embodiments, the multispecific antibodies of the invention are bispecific and bivalent.

[0154] In a further particular embodiment, the multispecific antibodies of the invention are bispecific and trivalent.

[0155] In a further particular embodiment, the multispecific antibodies of the invention are trispecific and trivalent.

[0156] In a further particular embodiment, the multispecific antibodies of the invention are trispecific and tetravalent.

[0157] Other variable domains for use in the present invention include amino acid sequences which have been mutated but still have at least 80, 85, 90, 91, 92, 93, 94, 95, 96, 97, 98, or 99 percent identity of the CDR regions to the CDR regions shown in the sequences set out in Tables 1, 2, and 3. Other variable domains for use in the present invention include variant amino acid sequences in which not more than 1, 2, 3, 4, or 5 amino acids have been mutated in the CDR regions when compared to the CDR regions shown in the sequences set out in Tables 1, 2, and 3.

[0158] Advantageously, the VH domains of the binding domains of the multispecific antibodies of the invention belong to or are selected from the human antibody VH family. In a preferred embodiment, the VH domains of one or two of the ROR1-BD, the CD3-BD and, if present, the hSA-BD of the multispecific antibodies of the invention belong to the VH framework subtypes VH1a, VH1b, VH3 or VH4. In one particular embodiment, said VH domains belong to the VH framework subtypes VH3 or VH4, in particular to the VH framework subtype VH3.

[0159] In the context of the present invention, the term "belonging to a VHx framework subtype (or VκX / Vλ framework subtype)" or "selected from a VHx framework subtype (or VκX / Vλ framework subtype)" means that the VH / VL framework sequences FR1 to FR4 (also referred to herein as VH / VL framework regions FR1 to FR4) show the highest degree of homology to said human antibody VHx or VκX / Vλ framework subtypes.

[0160] Examples of VH1a, VH1b, VH3, and VH4 sequences, as well as examples of other VHx sequences, are given by Knappik et al., Journal of Molecular Biology, vol. 296 (2000) p. 57-86 or WO2019 / 057787. A specific example of a VH domain belonging to the VH3 family is represented by SEQ ID NO: 71, and specific examples of VH domains belonging to the VH1a, VH1b, or VH4 framework subtypes are represented by SEQ ID NOs: 72, 73, and 74 (Table 4, framework regions are marked in non-bold). In particular, the framework regions FR1 to FR4 are taken from SEQ ID NO: 71 belonging to the VH3 family (Table 4, regions are marked in non-bold). Suitably, a VH belonging to the VH3 family as used herein is a VH comprising FR1 to FR4 having at least 85%, particularly at least 90%, more particularly at least 95% sequence identity to FR1 to FR4 of SEQ ID NO: 71. Alternative examples of VH3 and VH4 sequences, as well as examples of other VHx sequences, can be found in Knappik et al., Journal of Molecular Biology, Vol. 296 (2000) p. 57-86 or WO2019 / 057787.

[0161] Suitably, the VL domain of the binding domain used in the present invention comprises Vκ framework FR1, FR2 and FR3, particularly Vκ1 or Vκ3 framework, particularly Vκ1 framework FR1 to FR3, and framework FR4 (which is selected from VκFR4). In the case where said binding domain is in scFv format, said binding domain comprises Vκ framework FR1, FR2 and FR3, particularly Vκ1 or Vκ3 framework, particularly Vκ1 framework FR1 to FR3, and framework FR4 (which is selected from VλFR4).

[0162] A suitable Vκ1 framework FR1 to FR3 and an exemplary VλFR4 are set forth in SEQ ID NO: 75 (Table 4, framework regions are not bolded). Alternative examples of Vκ1 sequences and examples of Vκ2, Vκ3, or Vκ4 sequences can be found in Knappik et al., Journal of Molecular Biology, vol. 296 (2000) pp. 57-86. A suitable Vκ1 framework FR1 to FR3 comprises an amino acid sequence having at least 85, 90, 95 percent identity to the amino acid sequence corresponding to FR1 to FR3 taken from SEQ ID NO: 75 (Table 4, framework regions are not bolded). Suitable VλFR4s are as set forth in SEQ ID NO:76 to SEQ ID NO:82 and SEQ ID NO:83 and comprise a single cysteine ​​residue, in particular in some cases a second single cysteine ​​is present on the corresponding VH chain, in particular at position 51 (AHo numbering) of the VH, for the formation of an inter-domain disulfide bond. In one embodiment, the VL domain of the binding domain of the multispecific antibody of the invention comprises a VλFR4 comprising an amino acid sequence having at least 85, 90, 95 percent identity to an amino acid sequence selected from any of SEQ ID NO:76 to SEQ ID NO:83, in particular to SEQ ID NO:76 or 83, when in scFv format.

[0163] In particular, the binding domain of the multispecific antibody of the invention comprises a VH domain listed in Tables 1, 2, and 3. Preferably, the binding domain used in the invention comprises a VH amino acid sequence listed in one of Tables 1, 2, and 3, in which no more than 10 amino acids on the framework sequences (i.e. sequences that are not CDR sequences) have been mutated (wherein the mutations are, as various non-limiting examples, additions, substitutions, or deletions). Preferably, the binding domain of the multispecific antibody of the invention comprises a VH amino acid sequence listed in one of Tables 1, 2, and 3, in which no more than 7 amino acids, particularly no more than 5, 4, 3, or 2 amino acids, particularly no more than 1 amino acid on the framework sequences (i.e. sequences that are not CDR sequences) have been mutated (wherein the mutations are, as various non-limiting examples, additions, substitutions, or deletions). Other binding domains for use in the present invention include amino acid sequences which have been mutated but still have at least 90, 91, 92, 93, 94, 95, 96, 97, 98, or 99 percent identity of the VH region to the VH region shown in the corresponding sequence set forth in one of Tables 1, 2, and 3. Also included are VH domains comprising at least positions 5 to 140 (AHo numbering), particularly at least positions 3 to 145, and more particularly at least positions 2 to 147, of one of the sequences set forth in Tables 1, 2, and 3, provided that such VH domains exhibit the functional characteristics defined in items 11 and 19 above.

[0164] In a particular embodiment, one or two of the ROR1-BD, CD3-BD and, if present, the hSA-BD of a multispecific antibody of the invention comprise a VH domain as defined above, wherein the framework regions FR1, FR2, FR3 and FR4 of said VH domain have substitutions (AHo numbering) of arginine (R) at amino acid position 12, threonine (T) at amino acid position 103 and glutamine (Q) at amino acid position 144. Said VH domains may be referred to herein as "variant VH domains" or "modified VH domains" or "improved VH domains". Similarly, binding domains comprising said VH domains may be referred to herein as "variant binding domains" or "modified binding domains" or "improved binding domains". Specific examples of such variant VH domains are the VH domains of SEQ ID NOs: 8, 11, 21, 24, 37, 66 and 69.

[0165] It has been found by the inventors that the modified ROR1, CD3, and hSA-BD, when in scFv format, exhibit significantly reduced immunogenicity when compared to versions of the binding domains that do not contain the substitutions indicated above in the VH framework regions. More specifically, the modified ROR1, CD3, and hSA-BD, when in scFv format, exhibit reduced binding to pre-existing anti-drug antibodies (ADA) present in human serum when compared to versions of the binding domains that do not contain the substitutions indicated above in the VH framework regions, in particular reduced binding to pre-existing ADA as determined by a pre-existing ADA binding assay, as defined in more detail in Example 10.

[0166] In particular, it has been found by the inventors that the multispecific antibodies of the invention comprising said modified ROR1, CD3 and hSA-BD exhibit significantly reduced immunogenicity when compared to versions of the multispecific antibodies of the invention that do not comprise said modified binding domains. More specifically, the multispecific antibodies of the invention comprising said modified ROR1, CD3 and hSA-BD show reduced binding to pre-existing anti-drug antibodies (ADA) present in human serum, in particular to pre-existing ADA as determined by a pre-existing ADA binding assay, more particularly by the pre-existing ADA binding assay defined in Example 10, when compared to versions of the multispecific antibodies of the invention that do not comprise said modified binding domains.

[0167] Immunogenicity, i.e., the tendency of a therapeutic protein to induce an antibody response in a patient, can be predicted, for example, by its ability to be recognized by anti-drug antibodies (ADAs) already present in the human serum of healthy and untreated individuals, referred to herein as "pre-existing ADAs."

[0168] Therefore, for purposes of the present invention, the term "immunogenicity" as used herein refers to the ability of a Therapeutic Protein, e.g., an antibody, antibody fragment, or antibody binding domain, to be recognized by pre-existing ADA in a human serum sample. Without intending to be bound by theory, it is believed that pre-existing ADA binding, and the induction of ADA formation during therapeutic treatment, is linked to the occurrence of B cell and / or T cell epitopes on the Therapeutic Protein. The extent of such immunogenicity can be determined by ELISA assays and expressed relative to the total number of human sera tested (percentage of positive serum samples) by the percentage of human serum samples that contain a measurable amount of pre-existing ADA and / or ADA formed during therapeutic treatment that recognizes or binds the Therapeutic Protein in question. The reduction in immunogenicity between a Therapeutic Protein and a corresponding Therapeutic Protein that has been modified for the purpose of reducing its immunogenicity can be measured by comparing the percentage of positive serum samples for the modified Therapeutic Protein with the percentage of positive serum samples for the original Therapeutic Protein. A lower number or percentage of positive serum samples for the modified Therapeutic Protein indicates reduced immunogenicity relative to the original Therapeutic Protein.

[0169] A serum sample is considered to contain a measurable amount of pre-existing ADA when the ELISA signal is above a certain threshold. This threshold is also referred to herein as screening cut point (SCP). SCP can be calculated as defined below, or can be set to any value relative to the maximum ELISA signal obtained for the serum tested (e.g., 30%, 25%, 20%, 15%, 10%, or 5% of the maximum ELISA signal obtained for the serum tested). Preferably, SCP is calculated as defined below.

[0170] In particular, the binding domain of the multispecific antibody of the invention comprises a VL domain listed in one of Tables 1, 2, and 3. Preferably, the binding domain of the multispecific antibody of the invention comprises a VL amino acid sequence listed in one of Tables 1, 2, and 3, in which no more than about 10 amino acids on the framework sequences (i.e. sequences that are not CDR sequences) have been mutated (wherein the mutations are, as various non-limiting examples, additions, substitutions, or deletions). Preferably, the binding domain of the multispecific antibody of the invention comprises a VL amino acid sequence listed in one of Tables 1, 2, and 3, in which no more than about 10 amino acids on the framework sequences (i.e. sequences that are not CDR sequences) have been mutated (wherein the mutations are, as various non-limiting examples, additions, substitutions, or deletions). Other binding domains for use in the present invention include amino acid sequences which are mutated but still have at least 90, 91, 92, 93, 94, 95, 96, 97, 98, or 99 percent identity of the VL region to the VL regions shown in the sequences set out in Tables 1, 2, and 3. Also included are VL domains comprising at least positions 5 to 140 (AHo numbering), particularly at least positions 3 to 145, and more particularly at least positions 2 to 147, of one of the sequences set out in Tables 1, 2, and 3, provided that such VL domains exhibit the functional characteristics defined in items 11 and 19 above.

[0171] In the context of the present invention, the term "binding domain as used in the present invention" relates to a binding domain, e.g. a binding domain independent of the context of multispecificity, and in particular to a binding domain comprised on a multispecific construct, e.g. one of the binding domains comprised on a bispecific, trispecific or tetraspecific construct.

[0172] Advantageously, the binding domains of the multispecific antibody of the invention are independently selected from the group consisting of Fab, F(ab)2, Fv, scFv, dsFv and scAb, in particular from the group consisting of Fv, scFv and dsFv. In particular, the binding domains of the multispecific antibody of the invention are scFv.

[0173] Suitably, the binding domains of the multispecific antibody of the invention are operably linked. The binding domains of the multispecific antibody of the invention are capable of binding simultaneously to their respective antigens or receptors. The term "simultaneously" as used in this context refers to the simultaneous binding of at least one of the ROR1-BD and the CD3-BD to their respective antigens. Alternatively, in the case where two ROR1-BDs are present in the multispecific antibody, the term "simultaneously" refers to the simultaneous binding of both ROR1-BD and the CD3-BD to their respective antigens.

[0174] The multispecific antibody of the present invention comprises one or two ROR1-BDs, one CD3-BD, and optionally one hSA-BD, wherein said one or two ROR1-BDs, said CD3-BD, and said optional hSA-BD are operably linked to each other.

[0175] The term "operably linked" as used herein refers to two molecules (e.g., polypeptides, domains, binding domains) attached in a manner that allows each molecule to retain functional activity. Two molecules can be "operably linked" whether they are attached directly or indirectly (e.g., via a linker, via a moiety, via a linker to a moiety). The term "linker" refers to a peptide or other moiety that is optionally located between the binding domains or antibody fragments used in the present invention. A number of strategies can be used to covalently link molecules together. These include, but are not limited to, a polypeptide linkage between the N- and C-termini of a protein or protein domain, linkage via a disulfide bond, and linkage via a chemical cross-linking reagent. In one aspect of this embodiment, the linker is a peptide bond created by recombinant technology or peptide synthesis. Choosing a suitable 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 proteolysis and oxidation. Furthermore, the linker may contain amino acid residues that provide flexibility.

[0176] 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, which connects two domains, each attached to one end of the linker. The polypeptide linker should have a length that is sufficient to link two molecules in such a way that they assume the correct conformation relative to each other such that they retain the desired activity. In certain 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). In addition, the amino acid residues selected for inclusion in the polypeptide linker should exhibit properties that do not significantly interfere with the activity of the polypeptide. Thus, the linker peptide should not exhibit charges that would conflict with the activity of the polypeptide as a whole or that would interfere with internal folding or form bonds or other interactions with the amino acid residues of one or more of the monomers that would seriously impede receptor monomer domain binding. In certain embodiments, the polypeptide linker is an unstructured polypeptide. Useful linkers include glycine-serine or GS linkers. A polymer of glycine and serine linked together by a "Gly-Ser" or "GS" linker (e.g., (Gly-Ser) n , (GSGGS) n [SEQ ID NO: 118], (GGGGS) n [SEQ ID NO: 119], and (GGGS) n[SEQ ID NO: 120], where n is an integer of at least 1), glycine-alanine polymers, alanine-serine polymers, and other flexible linkers, such as the tether of the Shaker potassium channel, and various other flexible linkers that will be appreciated by those skilled in the art. Glycine-serine polymers are preferred because oligopeptides containing these amino acids are relatively unstructured and thus may be capable of serving as neutral tethers between components. Secondly, serine is hydrophilic and thus capable of solubilizing what may be globular glycine chains. Thirdly, similar chains have been shown to be effective in connecting subunits of recombinant proteins, such as single chain antibodies.

[0177] In one group of embodiments, the multispecific antibody is a format selected from any suitable format known in the art that is at least trispecific and does not include immunoglobulin Fc region(s) and CH1 and / or CL region. This includes, as non-limiting examples, tandem tri-scFv (triple body), triabody, scDb-scFv, tandem tri-scFv or scDb-scFv fused to the N- and / or C-terminus of a heterodimerization domain other than the heterodimeric Fc domain, and formats based on MATCH (described in WO2016 / 0202457, Egan T. et al., MABS Vol. 9 (2017) p. 68-84). In particular, the multispecific antibody format is selected from scDb-scFv, scMATCH3, MATCH3, and MATCH4.

[0178] The term "scDB" or "single-chain diabody" refers to an antibody fragment with two antigen-binding sites, which fragments comprise a VH connected to a VL on 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 to create two antigen-binding sites. Single-chain diabodies can be bivalent or bispecific, preferably bispecific. Single-chain diabodies are described in detail, for example, in EP404097, WO93 / 01161, Hudson et al., Nature Medicine, 9:129-134 (2003), and Hollinger et al., Proceedings of the National Academy of Sciences, 90:6444-6448 (1993). Triabodies are described in Hudson et al., Nature Medicine, 9:129-134 (2003).

[0179] Bispecific scDbs, in particular bispecific monomeric scDbs, can be in particular linked by linkers L1, L2 and L3 as follows: 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-VL B, two variable heavy domains (VH) or fragments thereof and two variable light domains (VL) or fragments thereof connected in the order VLB-L1-VLA-L2-VHA-L3-VHB, or VLB-L1-VHA-L2-VLA-L3-VHB, where the VLA and VHA domains combine to form an antigen-binding site for a first antigen and the VLB and VHBs combine to form an antigen-binding site for a second antigen.

[0180] The linker L1 is in particular a peptide of 2 to 10 amino acids, more particularly 3 to 7 amino acids, most particularly 5 amino acids, and the linker L3 is in particular a peptide of 1 to 10 amino acids, more particularly 2 to 7 amino acids, most particularly 5 amino acids. In a particular embodiment, the linker L1 and / or L3 is a peptide of 1 or 2 units (GGGGS) consisting of 4 glycine amino acid residues and 1 serine amino acid residue. n [SEQ ID NO: 121], where n=1 [SEQ ID NO: 85] or 2 [SEQ ID NO: 86], in particular n=1.

[0181] The middle linker L2 is in particular a peptide of 10 to 40 amino acids, more particularly 15 to 30 amino acids, most particularly 20 to 25 amino acids. In a particular embodiment, said linker L2 is one or more units (GGGGS) consisting of four glycine amino acid residues and one serine amino acid residue. n [SEQ ID NO: 122], where n=1, 2, 3, 4, 5, 6, 7, or 8, particularly n=4 [SEQ ID NO: 84].

[0182] In one embodiment, the multispecific antibody of the present invention comprises one ROR1-BD, one CD3-BD, and one hSA-BD, and is a single chain protein with scDb-scFv or scMATCH3 format. 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) by a flexible Gly-Ser linker. In one embodiment, said flexible Gly-Ser linker is a peptide of 2 to 40 amino acids, for example 2 to 35, 2 to 30, 2 to 25, 2 to 20, 2 to 15, 2 to 10 amino acids, in particular 10 amino acids. In a particular embodiment, said linker is one or more units (GGGGS) consisting of four glycine amino acid residues and one serine amino acid residue. n [SEQ ID NO: 122], where n=1, 2, 3, 4, 5, 6, 7 or 8, in particular n=2 [SEQ ID NO: 86]. In particular, the single chain multispecific antibody of this embodiment is in the scMATCH3 format.

[0183] Specific, but non-limiting, examples of scMATCH3 multispecific antibodies of the invention are PRO2286, PRO2287, PRO22507, PRO2508, PRO2509, PRO2510, PRO2557, PRO2596, PRO2667, and PRO2668, whose sequences are listed in Table 5.

[0184] In another embodiment, the multispecific antibody of the present invention comprises two ROR1-BDs, one CD3-BD, and one hSA-BD and is a heterodimeric protein of the MATCH4 format described in WO2016 / 0202457; Egan T. et al., MABS Vol. 9 (2017) p. 68-84.

[0185] Specific, but non-limiting, examples of heterodimeric MATCH4 multispecific antibodies of the present invention are PRO2589, PRO2590, PRO2591, PRO2592, PRO2658, PRO2659, PRO2669, and PRO2670, whose sequences are listed in Table 5.

[0186] The multispecific antibodies of the present invention or their fragments or binding domains thereof, such as ROR1-BD, may be produced using any convenient antibody production method known in the art (for example, for the production of bispecific constructs, see Fischer, N. & Leger, O., Pathobiology, 74 (2007) pp. 3-14; for bispecific diabodies and tandem scFvs, see Hornig, N. & Farber-Schwarz, A., Methods in Molecular Biology, 907 (2012) pp. 713-727 and WO99 / 57150). Further examples of suitable methods for the preparation of bispecific constructs include, among others, Genmab (see Labrijn et al., Proceedings of the National Academy of Sciences, Vol. 110 (2013) pp. 5145-5150) and Merus (see de Kruif et al., Biotechnology and Bioengineering, Vol. 106 (2010) pp. 741-750) technologies.

[0187] These methods typically involve the generation of monoclonal antibodies and the combination of the antigen-binding domains of two or more different monoclonal antibodies, or fragments or portions thereof, to provide bispecific or multispecific constructs using known molecular cloning techniques.

[0188] Multispecific antibodies of the invention can be prepared by conjugating the component binding specificities using methods known in the art. For example, each binding specificity of the 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, carbodiimides, 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, Journal of Experimental Medicine 160:1686; Liu, MA et al., 1985, Proceedings of the National Academy of Sciences 82:8648). Other methods include those described in Paulus, 1985, Behring Institute Mitteilungen, No. 78, pp. 118-132; Brennan et al., 1985, Science, vol. 229: pp. 81-83, and Glennie et al., 1987, Journal of Immunology, vol. 139: pp. 2367-2375. Conjugating agents are SATA and sulfo-SMCC, both available from Pierce Chemical Co. (Rockford, Ill., USA).

[0189] When the binding specificities are antibodies, they can be conjugated by sulfhydryl bonding of the C-terminal hinge regions of the two heavy chains. In certain embodiments, prior to conjugation, the hinge region is modified to contain an odd number of sulfhydryl residues, e.g., one.

[0190] 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 where the bispecific molecule is a mAb x Fab, mAb x scFv, mAb x dsFv, or mAb x Fv fusion protein. Methods for preparing multispecific antibodies and molecules are described, for example, in US 5,260,203, US 5,455,030, US 4,881,175, US 5,132,405, US 5,091,513, US 5,476,786, US 5,013,653, US 5,258,498, and US 5,482,858.

[0191] The binding of multispecific antibodies to their specific targets can be confirmed, for example, by enzyme-linked immunosorbent assay (ELISA), radioimmunoassay (REA), FACS analysis, bioassays (e.g., growth inhibition), or Western blot assays. Each of these assays generally detects the presence of a particular protein-antibody complex of interest by using a labeled reagent (e.g., an antibody) specific for the complex of interest.

[0192] In a further embodiment, the present invention provides a nucleic acid or two nucleic acids encoding a multispecific antibody of the invention or a fragment thereof or a binding domain thereof, such as ROR1-BD. Such a nucleic acid may be optimized for expression by a mammalian cell.

[0193] The term "nucleic acid" is used interchangeably herein with the term "polynucleotide(s)" and refers to one or more deoxyribonucleotides or ribonucleotides and polymers thereof, in either single- or double-stranded form. The term encompasses nucleic acids containing known nucleotide analogs or modified backbone residues or linkages, which are synthetic, naturally occurring, and non-naturally occurring. They have similar binding properties as the reference nucleic acid, and they are metabolized in a similar manner as the reference nucleotide. Examples of such analogs include, without limitation, phosphorothioates, phosphoramidates, methyl phosphonates, chiral methyl phosphonates, 2-O-methyl ribonucleotides, peptide nucleic acids (PNAs). Unless otherwise indicated, a particular nucleic acid sequence also implicitly encompasses its conservatively modified variants (e.g., degenerate codon substitutions) and complementary sequences, as well as the sequence explicitly indicated. Specifically, as detailed below, degenerate codon substitutions can be achieved by creating sequences in which the third position of one or more selected (or all) codons is substituted with mixed-base and / or deoxyinosine residues (Batzer et al., Nucleic Acids Research 19:5081, 1991; Ohtsuka et al., Journal of Biological Chemistry 260:2605-2608, 1985; and Rossolini et al., Molecular and Cellular Probes 8:91-98, 1994).

[0194] The present invention provides substantially purified nucleic acid molecules encoding polypeptides comprising segments or domains of the multispecific antibodies described above, such as ROR1-BD. When expressed from an appropriate expression vector, the polypeptides encoded by these nucleic acid molecules are capable of displaying the antigen-binding ability of the multispecific antibodies or ROR1-BD of the present invention.

[0195] Polynucleotide sequences can be produced by de novo solid-phase DNA synthesis or by PCR mutagenesis of existing sequences (e.g., sequences described in the Examples below) encoding the multispecific antibodies of the invention or their fragments or binding domains thereof, such as the ROR1-BD of the invention. Direct chemical synthesis of nucleic acids can be achieved by methods known in the art. For example, the phosphotriester method of Narang et al., 1979, Methods in Enzymology 68:90; the phosphodiester method of Brown et al., 1979, Methods in Enzymology 68:109; the diethyl phosphoramidite method of Beaucage et al., Tetrahedron Letters 22:1859, 1981; and the solid support method of US 4,458,066. Introducing mutations into a polynucleotide sequence by PCR can be performed, 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. (eds.), Academic Press, San Diego, Calif., 1990; Mattila et al., Nucleic Acids Research, vol. 19:967, 1991; and Eckert et al., PCR Methods and Applications, vol. 1:17, 1991.

[0196] 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, such as ROR1-BD.

[0197] The term "vector" is intended to refer to a polynucleotide molecule capable of transporting another polynucleotide to which it is 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 into which additional DNA segments can be ligated into the viral genome. Certain 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 are thereby replicated together with the host genome.

[0198] Moreover, certain vectors are capable of directing the expression of genes to which they are operably linked. Such vectors are referred to herein as "recombinant expression vectors" (or simply "expression vectors"). In general, expression vectors useful in recombinant DNA technology are often in the form of plasmids. As plasmids are the most commonly used form of vector, "plasmid" and "vector" may be used interchangeably herein. However, the invention is intended to encompass 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 term "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 transcription control sequence to a transcribed sequence. For example, a promoter or enhancer sequence is operably linked to a coding sequence if it stimulates or regulates the transcription of the coding sequence in an appropriate host cell or other expression system. Generally, a promoter transcription control sequence that is operably linked to a transcribed sequence is physically contiguous with the transcribed sequence. That is, they are cis-acting. However, some transcriptional control sequences, such as enhancers, need not be physically contiguous or in close proximity to the coding sequences whose transcription they enhance.

[0199] A variety of expression vectors can be used to express polynucleotides encoding multispecific antibody chain(s) or their fragments or their binding domains, such as ROR1-BD. Both viral-based and non-viral expression vectors can be used to produce antibodies by mammalian host cells. Non-viral vectors and systems include plasmids, episomal vectors that typically carry expression cassettes for expressing proteins or RNA, and human artificial chromosomes (see, for example, Harrington et al., Nature Genetics 15:345, 1997). For example, non-viral vectors useful for expressing IL-4R-binding polynucleotides and polypeptides in mammalian (e.g., human) cells include pThioHisA, B, and C, pcDNA3.1 / His, pEBVHisA, B, and C (Invitrogen, San Diego, USA), MPSV vectors, and numerous 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) based vectors. See Brent et al., supra; Smith, Annual Review of Microbiology 49:807, 1995; and Rosenfeld et al., Cell 68:143, 1992.

[0200] The choice of expression vector depends on the intended host cell in which the vector is to be expressed. Typically, expression vectors contain a promoter and other control sequences (e.g., enhancers) that are operably linked to the polynucleotide encoding the multispecific antibody chain or fragment. In one embodiment, an inducible promoter is used to prevent expression of the inserted sequence except under induction conditions. Inducible promoters include, for example, arabinose, lacZ, metallothionein promoters, or heat shock promoters. Cultures of transformed organisms can be expanded without biasing the population for coding sequences whose expression products are better tolerated by the host cell under non-inducing conditions. In addition to promoters, other control elements may also be required or desired for efficient expression of the multispecific antibody chain or fragment. These elements typically include the ATG initiation codon and adjacent ribosome binding sites or other sequences. In addition, the efficiency of expression can be enhanced by the inclusion of enhancers appropriate for the cell system being used (see, e.g., Scharf et al., Results and Problems in Cell Differentiation 20:125, 1994; and Bittner et al., Methods in Enzymology 153:516, 1987). For example, the SV40 enhancer or CMV enhancer can be used to increase expression in mammalian host cells.

[0201] The vectors to be used typically encode the light and heavy chain variable domains of the multispecific antibody. In certain cases, they also encode the constant regions or parts thereof. Such vectors allow the expression of the variable regions with the constant regions as a fusion protein, thereby leading to the production of intact antibodies and their antigen-binding fragments. Typically, such constant regions are human.

[0202] 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 to the progeny of such a cell. Because certain modifications due to either mutation or environmental influences may occur in successive generations, such progeny may not actually be identical to the parent cell, but are still included within the scope of the term "host cell" as used herein.

[0203] Host cells for harboring and expressing the multispecific antibodies of the invention or fragments thereof or binding domains thereof, such as ROR1-BD, can be either prokaryotic or eukaryotic. E. coli is one useful prokaryotic host for cloning and expressing the polynucleotides of the invention. Other microbial hosts suitable for use include bacilli, such as Bacillus subtilis, and other Enterobacteriaceae, such as Salmonella, Serratia, and various Pseudomonas species. In these prokaryotic hosts, expression vectors can also be made, which typically contain expression control sequences (e.g., an origin of replication) compatible with the host cell. In addition, there will be any number of a variety of well-known promoters, such as the lactose promoter system, the tryptophan (trp) promoter system, the beta-lactamase promoter system, or promoter systems from lambda phage. The promoters typically control expression, optionally have operator sequences, and have ribosome binding site sequences and the like for initiating and completing transcription and translation. Other microbes, such as yeast, can also be used to express the multispecific antibodies of the invention. Insect cells in combination with baculovirus vectors may also be used.

[0204] 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, such as ROR1-BD. For example, they can be either hybridoma cell lines expressing endogenous immunoglobulin genes, or mammalian cell lines harboring exogenous expression vectors. These include any normal mortal or normal or abnormal immortal animal or human cells. For example, a number of 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 cultures to express polypeptides is generally discussed, for example, in Winnacker, FROM GENES TO CLONES, VCH Publishers, NY, NY, 1987. Expression vectors for mammalian host cells can include expression control sequences, such as origins of replication, promoters, and enhancers (see, e.g., Queen et al., Immunological Reviews 89:49-68, 1986), as well as necessary processing information sites, such as ribosome binding sites, RNA splice sites, polyadenylation sites, and transcription termination sequences. These expression vectors usually contain a promoter derived from a mammalian gene or from a mammalian virus. Suitable promoters can be constitutive, cell type-specific, stage-specific, and / or regulatable or controllable. Useful promoters include, but are not limited to, metallothionein promoters, constitutive adenovirus major late promoters, dexamethasone-inducible MMTV promoters, SV40 promoters, MRPpolIII promoters, constitutive MPSV promoters, tetracycline-inducible CMV promoters (e.g., human immediate early CMV promoters), constitutive CMV promoters, and promoter-enhancer combinations known in the art.

[0205] Methods for introducing an expression vector containing a polynucleotide of interest vary depending on the type of cellular host. For example, calcium chloride transfection is commonly utilized for prokaryotic cells, while calcium phosphate treatment or electroporation may be used for other cellular hosts (see generally, Green, MR and Sambrook, J., Molecular Cloning: A Laboratory Manual (4th ed.), Cold Spring Harbor Laboratory Press (2012)). Other methods include, for example, electroporation, calcium phosphate treatment, liposome-mediated transformation, injection and microinjection, ballistic techniques, virosomes, immunoliposomes, polycation-nucleic acid conjugates, naked DNA, artificial virions, fusion to herpes virus structural protein VP22 (Elliot and O'Hare, Cell 88:223, 1997), drug-enhanced uptake of DNA, and ex vivo transduction. For long-term, high-yield production of recombinant proteins, stable expression will often be desired. For example, cell lines stably expressing the multispecific antibodies or fragments thereof or binding domains thereof of the present invention, such as ROR1-BD, can be prepared using the expression vectors of the present invention containing viral origins of replication or endogenous expression elements and a selectable marker gene. After introduction of the vector, the cells can be grown in rich medium for 1 to 2 days before they are switched to selective medium. The purpose of the selectable marker is to confer resistance to selection, and its presence allows growth of cells that successfully express the introduced sequences on a selective medium. Resistant, stably transfected cells can be propagated using tissue culture techniques appropriate to the cell type.Therefore, the present invention provides a method for producing a multispecific antibody or antigen-binding fragment thereof, such as ROR1-BD, of the present invention, wherein said method comprises the step of culturing a host cell comprising a nucleic acid or vector encoding a multispecific antibody or antigen-binding fragment thereof, such as ROR1-BD, of the present invention, whereby said multispecific antibody or antigen-binding fragment thereof, such as ROR1-BD, of the present disclosure is expressed.

[0206] In one embodiment, the present invention relates to a method for producing a multispecific antibody or ROR1-BD of the present invention, comprising the step of culturing a host cell expressing a nucleic acid encoding a multispecific antibody or ROR1-BD of the present invention. In particular, the present invention relates to a method for producing a multispecific antibody or ROR1-BD of the present invention, comprising (i) providing a nucleic acid or two nucleic acids encoding a multispecific antibody or ROR1-BD of the present invention, or one vector or two vectors encoding a multispecific antibody or ROR1-BD of the present invention, expressing said nucleic acid(s) or said vector(s) and collecting said multispecific antibody or said ROR1-BD from the expression system, or (ii) providing a host cell(s) expressing a nucleic acid(s) encoding a multispecific antibody or ROR1-BD of the present invention, culturing said host cell(s) and collecting said multispecific antibody or said ROR1-BD from the cell culture.

[0207] In a further aspect, the present invention relates to a pharmaceutical composition comprising a multispecific antibody of the present invention and a pharma- ceutically acceptable carrier. By "pharmaceutically acceptable carrier" is meant a medium or diluent that does not interfere with the structure of the antibody. A pharma- ceutically acceptable carrier enhances or stabilizes the composition or facilitates the preparation of the composition. Pharmaceutically acceptable carriers include physiologically compatible solvents, dispersion media, coatings, antibacterial and antifungal agents, isotonic and absorption delaying agents, and the like.

[0208] Certain of such carriers allow the pharmaceutical composition to be formulated as, for example, tablets, pills, dragees, capsules, liquids, gels, syrups, slurries, suspensions, and lozenges for oral ingestion by a subject. Certain of such carriers allow the pharmaceutical composition to be formulated for injection, infusion, or topical administration. For example, a pharma- ceutically acceptable carrier can be a sterile aqueous solution.

[0209] The pharmaceutical composition of the present invention can be administered by various methods known in the art. The route and / or mode of administration varies depending on the desired results. Administration can be intravenous, intramuscular, intraperitoneal, or subcutaneous, or can be administered 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), particularly for intramuscular or subcutaneous administration. Depending on the route of administration, the active compound, i.e., the multispecific antibody of the present invention, can be coated with a material to protect it from the action of acids and other natural conditions that may inactivate the compound.

[0210] The pharmaceutical compositions of the present invention may 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 Edition, 2000; and Sustained and Controlled Release Drug Delivery Systems, edited by J.R. Robinson, Marcel Dekker Inc., New York, 1978. The pharmaceutical compositions are preferably manufactured under GMP conditions. Typically, a therapeutically effective dose or effective dose of the multispecific antibodies of the present invention is used in the pharmaceutical compositions of the present invention. The multispecific antibodies of the present invention are formulated into pharma- ceutically acceptable dosage forms by conventional methods known to those skilled in the art. The dosage regimen is adjusted to provide the optimum desired response (e.g., therapeutic response). For example, a single bolus may be administered, several divided doses may be administered over time, or the dose may be reduced or increased accordingly as indicated by the exigencies of the therapeutic situation. For the ease of administration and uniformity of dosage, it is particularly advantageous to prepare parenteral compositions in unit dosage form.As used herein, unit dosage form refers to a physically separate unit suitable as a single dosage for the subject to be treated.Each unit contains a predetermined amount of active compound calculated to produce desired therapeutic effect together with required pharmaceutical carrier.

[0211] The actual dosage level of the active ingredient in the pharmaceutical composition of the present invention can be varied to obtain an amount of the active ingredient that is effective to achieve the desired therapeutic response for a particular patient, composition, and mode of administration without being toxic to the patient. The dosage level selected will depend on various pharmacokinetic factors, including the activity of the particular composition of the present invention or their esters, salts, or amides used, the route of administration, the time of administration, the excretion rate of the particular compound being used, the duration of treatment, other drugs, compounds, and / or materials used in combination with the particular composition being used, the age, sex, weight, condition, general health, and medical history of the patient to be treated, and the like.

[0212] The multispecific antibodies of the invention are usually administered on multiple occasions. The intervals between doses can be weekly, monthly, or yearly. The intervals can also be irregular as indicated by measuring the blood levels of the multispecific antibodies of the invention in the patient. Alternatively, the multispecific antibodies of the invention can be administered as a sustained release formulation, in which case less frequent administration is required. The dosage and frequency will vary depending on the half-life of the antibody in the patient. In general, humanized antibodies 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. In prophylactic applications, relatively low dosages are administered at relatively infrequent intervals over an extended period of time. Some patients continue to receive treatment for the rest of their lives. In therapeutic applications, relatively high dosages are sometimes required at relatively short intervals until the progression of the disease is reduced or halted, preferably until the patient shows partial or complete remission of the symptoms of the disease. The patient can then be administered a prophylactic regimen.

[0213] In one aspect, the present invention relates to a multispecific antibody of the invention or a pharmaceutical composition of the invention for use as a medicament. In a preferred embodiment, the present invention provides a multispecific antibody or a pharmaceutical composition for use in the treatment of a proliferative disease, in particular a cancer, more particularly a ROR1-expressing cancer, in a subject in need thereof.

[0214] In another aspect, the present invention provides a pharmaceutical composition for use in the manufacture of a medicament for the treatment of a proliferative disease, particularly cancer, more particularly a ROR1-expressing cancer.

[0215] In another embodiment, the present invention relates to the use of the multispecific antibody or the pharmaceutical composition for treating a proliferative disease, in particular cancer, more particularly ROR1-expressing cancer, in a subject in need thereof.

[0216] In another aspect, the present invention relates to a method for treating a subject, comprising administering to the subject a therapeutically effective amount of a multispecific antibody of the invention. In a preferred embodiment, the present invention relates to a method for the treatment of a proliferative disease, in particular cancer, more particularly ROR1-expressing cancer, in a subject, comprising administering to the subject a therapeutically effective amount of a multispecific antibody of the invention.

[0217] The term "subject" includes human and non-human animals.

[0218] The term "animal" includes all vertebrates, e.g., non-human mammals and non-mammals, e.g., non-human primates, sheep, dogs, cows, chickens, amphibians, and reptiles. Except where noted, the terms "patient" or "subject" are used interchangeably herein.

[0219] As used herein, the terms "treatment," "treating," "treat," "treated," and the like refer to obtaining a desired pharmacological and / or physiological effect. The effect may be therapeutic in terms of partial or complete cure of a disease and / or harmful effects attributable to the disease, or slowing disease progression. As used herein, "treatment" covers any treatment of a mammalian, e.g., human, disease, including (a) inhibiting the disease, i.e., halting its development, and (b) relieving the disease, i.e., causing regression of the disease.

[0220] The term "therapeutically effective amount" or "effective amount" refers to the amount of an agent that, when administered to a mammal or other subject for treating a disease, is sufficient to affect such treatment for the disease. A "therapeutically effective amount" will vary depending on the agent, the disease and its severity, as well as the age, weight, etc., of the subject to be treated.

[0221] In one embodiment, the proliferative disease is cancer. The term "cancer" refers to a disease characterized by rapid and uncontrolled growth of abnormal cells. Cancer cells can spread locally or from the bloodstream and lymphatic system to other parts of the body. The terms "tumor" and "cancer" are used interchangeably herein, e.g., both terms encompass solid and liquid tumors, e.g., diffuse or circulating tumors. The term "cancer" or "tumor" as used herein includes precancerous and malignant cancers and tumors. The term "cancer" is used herein to mean a broad range of tumors, including all solid and hematological malignancies. In particular, the cancer is a ROR1-expressing cancer.

[0222]

[0176] Non-limiting examples of ROR1-expressing cancers are lymphocytic leukemia / small lymphocytic lymphoma (CLL / SLL), acute myeloid leukemia (AML), acute lymphoblastic leukemia (ALL), mantle cell lymphoma (MCL), hairy cell leukemia, follicular lymphoma (FL), marginal zone lymphoma (MZL), diffuse large B-cell lymphoma (DLBCL), Richter's syndrome (RS), lung cancer, pancreatic cancer, prostate cancer, colon cancer, bladder cancer, breast cancer, ovarian cancer, glioblastoma, testicular cancer, uterine cancer, adrenal cancer, melanoma, neuroblastoma, sarcoma, and renal cancer.

[0223] The multispecific antibodies or pharmaceutical compositions of the invention also inhibit the growth of solid and liquid tumors. The multispecific antibodies or pharmaceutical compositions of the invention are also suitable for preventing the metastatic spread of tumors and the growth or development of micrometastases in subjects with cancer. SEQUENCE LISTING (unless otherwise specified, mutations are designated according to the AHo numbering scheme, CDRs are defined according to the CDR definitions of Numab)

[0224]

Table 1-a

Table 1-b

Table 1-c

[0225]

Table 2

[0226]

Table 3-a

Table 3-b

Table 3-c

[0227]

Table 4

[0228]

Table 5-a

Table 5-b

Table 5-c

Table 5-d

Table 5-e

Table 5-f

[0229] [Table 6]

[0230] In the event of a conflict within the text of this application between the text of the specification (eg, Tables 1 through 6) and the sequence listing, the text of the specification will control.

[0231] It is understood that certain features of the invention that are described for clarity in the context of separate embodiments may also be provided in combination in a single embodiment. Conversely, various features of the invention that are described for brevity in the context of a single embodiment may also be provided separately or in any suitable subcombination. All combinations of the embodiments relating to the present invention are specifically embraced by the present invention and are disclosed herein as if each of the combinations were individually and explicitly disclosed herein. In addition, all subcombinations of the various embodiments and elements thereof are also specifically embraced by the present invention and are disclosed herein as if each of such subcombinations were individually and explicitly disclosed herein.

[0232] The present invention should not 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 be encompassed within the scope of the appended claims.

[0233] To the extent possible under their respective patent laws, all patents, applications, publications, test methods, literature, and other materials cited herein are incorporated herein by reference.

[0234] The following examples illustrate the invention described above, but are not intended to limit the scope of the invention in any way. Thus, other test models known to those skilled in the art can also determine the beneficial effects of the claimed invention. EXAMPLES

[0235] Example 1: Generation and testing of anti-ROR1 molecules: Project Aims The goal of the project was to generate humanized monoclonal antibody fragments that specifically bind to the extracellular domain of human ROR1. An additional goal was to identify anti-ROR1 antibody fragments that are sufficiently potent and stable for incorporation into a multispecific antibody format.

[0236] 1.1.Immunization Nine rabbits were immunized with human ROR1 extracellular domain (ECD). Six rabbits were immunized according to Numab's standard protocol, consisting of four antigen injections over a 70 day period, and three additional rabbits were immunized according to the extension protocol, which consisted of the same initial schedule as the standard protocol, with the final antigen injection given four days before the end of the protocol on day 112.

[0237] 1.2. Sorting and hit identification 3,740 single-cell B cells from six rabbits immunized according to Numab's standard immunization protocol were sorted by FACS with the ROR1 extracellular domain. Single-cell B cells were cultured and B cell supernatants containing the secreted antibodies of interest were collected over a period of four weeks. From those 3,740 sorted B cells, 993 clones showed binding to recombinant human ROR1 by ELISA. The positive clones were then tested for their ability to bind to ROR1-expressing cells, and 627 clones showed binding to cancer cells (MDA-MB-231) expressing high ROR1 levels, and 99 of them showed affinity to human ROR1ECD lower than 800 pM. The heavy and light chain sequences from those clones were retrieved, and 15 ROR1-binding clones with the best binding affinity measured by SPR were selected and produced as recombinant rabbit IgG.

[0238] 1.3. Production of Recombinant Rabbit IgG After selection of clones for hit confirmation, rabbit antibodies were cloned, expressed, and purified for further characterization. Cloning of the corresponding light and heavy chain variable domains involved in vitro ligation of DNA fragments onto suitable mammalian expression vectors (pFUSE, Invivogen). Expression vectors for rabbit antibody heavy and light chains were transfected into mammalian suspension cell lines (CHO-S) for transient heterologous expression (50 ml scale, CHOgro Expression Kit from Mirus). Secreted rabbit IgG was then affinity purified (Protein A affinity purification), buffer exchanged into PBS, pH 7.4, and the final product was analyzed by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE, data not shown), UV absorbance at 280 nm, and size-exclusion high performance liquid chromatography (SE-HPLC) to verify identity, content, and purity. Of the 15 selected rabbit IgGs, 13 were successfully produced.

[0239] 1.4. Hit confirmation of recombinant rabbit IgG (rIgG) Thirteen successfully produced rIgGs were assessed for binding affinity to ROR1-expressing cells by CELISA and to recombinant ROR1 by SPR. In addition, the binding region of rIgG on ROR1 was defined by assessing the binding of rabbit IgG to three chimeric ROR1 / ROR2 variants.

[0240] Binding to ROR1-expressing cells by CELISA Binding to ROR1 expressing cells was evaluated using MDA-MB-231 cells (ATCC, cat.HTB-26, human breast adenocarcinoma cells expressing high levels of human ROR1) and MCF-7 (ATCC, cat.HTB-22, human breast cancer cells not expressing human ROR1). 40,000 cells were distributed into flat-bottom tissue culture treated 96-well plates. The following day, the plates were washed three times with 450 μl of washing buffer (PBS, 0.2% BSA) per well in overflow mode, and 50 μl of each point of serial dilutions of each tested rabbit IgG and reference anti-ROR1 antibody PRO1842 (human IgG antibody comprising VH / VL sequences of SEQ ID NOs: 129 and 130 as disclosed in patent application US2017 / 0306018A1) were added, and the plates were incubated for 1.5 h at room temperature (RT) under gentle agitation. After washing three times with 450 μl wash buffer, 50 μl of HRP-coupled goat anti-rabbit IgG or rabbit anti-human IgG antibody was added to each well. After 1 h incubation at RT on a nutation mixer, the plates were washed three times with 450 μl wash buffer per well prior to the addition of 50 μl TMB (3,3',5,5'-tetramethylbenzidine, KPL, Cat. No. 53-00-00). After 10 min of color development, the enzymatic reaction was stopped by the addition of 50 μl 1 M HCl per well and the plates were read at 450 nm with 690 nm as the reference wavelength. For recombinant production of PRO1842, mammalian expression plasmid (pcDNA3.1, host: CHO-S) was ordered from Gene Universal (Newark DE, USA). Expression and purification of PRO1842 was performed in a similar manner as described in section 1.3 and yielded sufficient material and purity (>98% monomer content by SE-HPLC) for the intended use.

[0241] Eleven clones showed binding to ROR1-expressing cells, and eight of these clones had similar or better EC 50 and bound to ROR1 at 100 ng / L (see Table 6).

[0242] Binding affinity by SPR The binding kinetics (including affinity) of rabbit IgG to recombinant human ROR1 protein (His tag, AcroBiosystems) was determined by SPR analysis with a Mass2 device (Sierra Sensors, Bruker). Rabbit IgG molecules were captured on specific spots via anti-rabbit IgG antibodies (goat anti-rabbit IgG-Fc, Bethyl) covalently immobilized on a carboxymethylated dextran surface (HCA sensor chip, Sierra Sensors, Bruker), and a titration series of recombinant human ROR1 protein was injected as analyte. After each analyte injection cycle, every spot on the sensor chip was regenerated (glycine-HCl, pH 1.5) and new rabbit IgG was subsequently recaptured. Binding kinetics to human ROR1 was measured using a multi-cycle kinetic assay with 10 analyte concentrations ranging from 0.176 to 90 nM diluted in running buffer (HEPES-buffered saline, 0.05% Tween-20, pH 7.5, Bioconcept). Apparent dissociation (k d ) and meeting (k a ) rate constant and apparent dissociation equilibrium constant (K D ) was calculated using the one-to-one Langmuir binding model by the Sierra Sensors analysis software (Sierra Analyzer, Sierra Sensors, Bruker), and the quality of the fit was evaluated using the relative chi 2 The binding level was monitored based on the theoretical R max The stability was calculated as the maximum stability bond achieved normalized to

[0243] Affinity for recombinant ROR1 was determined for 11 rabbit IgGs. No binding was observed for one clone, and the best binding affinity measured was 1.06 nM. The results are summarized in Table 7. The reference anti-ROR1 antibody PRO1776 (human IgG4 antibody comprising VH / VL sequences of SEQ ID NO: 5 and 7 as disclosed in patent application WO2014 / 031174) was also measured in a similar setup. For recombinant production of PRO1776, a mammalian expression plasmid (pcDNA3.1, host: CHO-S) was ordered from Gene Universal (Newark DE, USA). Expression and purification of PRO1776 was performed in a similar manner as described in section 1.3 above, yielding sufficient material and purity (>98% monomer content by SE-HPLC) for the intended use.

[0244] Identification of rabbit IgG (rIgG)-binding regions on ROR1 To define the binding region of rIgG on ROR1, the degree of binding to HEK293T cells transiently transfected with three chimeric ROR1 / ROR2 variants was evaluated by CELISA. The ROR1 / ROR2 variants were generated by exchanging each of the three domains of ROR1, namely, the Ig-like domain of ROR1 (region 1), the Frizzled domain of ROR1 (region 2), and the Kringle domain of ROR1 (region 3), with the corresponding domains in ROR2. To exclude potential ROR2 binders, binding to ROR2 was first evaluated by ELISA, but no ROR2 binders were identified using this assay.

[0245] Plates were then coated with 25,000 cells per well on flat-bottom poly-D-lysine treated 96-well plates. The following day, cells were transfected with the corresponding constructs and incubated at 37°C, 5% CO2. After 24 h, cells were washed with 450 μl wash buffer (PBS, 0.2% BSA) and samples were added (10 μg / ml rIgG) and incubated for 1.5 h at RT under gentle agitation. After three washes with 450 μl wash buffer, 50 μl of HRP-coupled goat anti-rabbit IgG antibody was added to each well. After 1 h of incubation at RT on a nutation mixer, plates were washed three times with 450 μl wash buffer per well prior to the addition of 50 μl TMB (3,3',5,5'-tetramethylbenzidine, KPL, Cat. No. 53-00-00). After 10 min, the color development of the enzyme reaction was stopped by adding 50 μl of 1 M HCl per well, and the plates were read at 450 nm with 690 nm as the reference wavelength. The binding levels of rIgG were calculated relative to the binding of rabbit IgG to wtROR1. A clear reduction in the binding level of rIgG to a particular variant compared to the wtROR1 protein would indicate the localization of the rIgG epitope on the segment of human ROR1 replaced by the respective ROR2 sequence.

[0246] Well-defined binding regions were identified for 12 rabbit IgGs. Nine rIgGs bound to region 1 (the Ig-like domain of ROR1) and three antibodies bound to region 2 (the Frizzled domain of ROR1). None of the rIgGs bound to region 3 (the Kringle domain of ROR1). The results are summarized in Table 7.

[0247] [Table 7]

[0248] Eight clones showed binding affinity to ROR1-expressing MDA-MB-231 cells comparable to or better than the reference anti-ROR1 antibody PRO1842. From these eight clones, the seven best clones were selected and expressed as two versions of a humanized single-chain Fv on the Numab framework.

[0249] 1.5.ScFv production ScFv creation The goal of sequence humanization is to reduce the risk of immunogenicity for our molecule of interest and to stabilize the antibody variable domain (Fv fragment) to enable its use as a building block for the assembly of multispecific formats. Thus, humanization involves the transfer of key residues from the donor (rabbit) complementarity determining region (CDR) sequence onto the human acceptor framework to increase humanness and improve stability without significantly impacting functionality.

[0250]

[0197] Seven rabbit monoclonal antibody clones were selected for lead ROR1 scFv domain generation as shown in Table 7. Humanization of these clones involved the transfer of rabbit CDRs into one of Numab's proprietary human variable domain acceptor scaffolds. In this process, the amino acid sequences of the six CDR regions were characterized using Numab's CDR definitions (Table 8) and grafted onto Numab's proprietary and highly stable fully human VH3 / Vk1 lambda cap acceptor framework. These constructs are known as "CDR grafts".

[0251] [Table 8]

[0252] The exclusive grafting of rabbit CDRs onto the human acceptor framework is the most basic grafting strategy, referred to herein as "CDR grafting". In addition to CDR grafting, additional graft variants were designed that contain a defined pattern of rabbit framework residues, referred to herein as "full grafting". The details of these graft variants are as follows: CDR grafting: Grafting of rabbit CDRs onto the acceptor framework of Numab, using the CDR definitions of Numab. No back mutations of donor framework residues. Full graft: The largest graft, containing donor Fv core residues, donor VL / VH interface residues, and donor framework residues that potentially interact with antigen.

[0253] Table 9 summarizes the CDR and full grafts of the ROR1 humanized domains. The rabbit framework residues grafted onto the human framework are listed separately for the VL and VH domains (using AHo numbering).

[0254] Mammalian expression plasmids (pcDNA3.1, host: CHO-S) of the humanized scFv constructs were ordered from Gene Universal in 1 mg scale and used for transient transfection of CHO-S cells as described below.

[0255] scFv manufacturing Expression of mammalian scFv constructs was performed in CHO-S cells using the CHOgro transient transfection kit (Mirus). Cultures were harvested by centrifugation and then filtered after up to 7 days of expression at 37°C (or when cell viability <70%). Proteins were purified from the clarified culture supernatants by protein L affinity chromatography. With the exception of the scFv from clone 55-06-E06, all molecules showed fractions with >95% monomer content after capture, assessed by SE-HPLC analysis. All molecules were directly rebuffered by dialysis into 50 mM phosphate-citrate buffer with 150 mM NaCl at pH 6.4. The CDR graft of clone 55-42-E05 (PRO2065) showed low expression titers and therefore could not be taken forward. For quality control of the produced material, standard analytical methods, e.g. SE-HPLC, UV 280 , and SDS-PAGE were applied. The preparation of humanized anti-ROR1scFv molecules is summarized in Table 10.

[0256] Since one scFv could not be expressed (55-42-E05-sc01 / PRO2065), only 13 of the 14 scFvs were subjected to subsequent pharmacodynamic characterization.

[0257] [Table 9]

[0258] [Table 10]

[0259] 1.6.ScFv characterization Cell binding to ROR1-expressing MDA-MB-231 cells by flow cytometry MDA-MB-231 cells (ATCC, cat.HTB-26, human breast adenocarcinoma cells expressing high levels of human ROR1) and MCF-7 (ATCC, cat.HTB-22, human breast adenocarcinoma cells not expressing human ROR1) were harvested and cell numbers were determined. Suspension cells were centrifuged at 400×g for 5 min, and 100 μl of suspension cells (50,000 cells) diluted in PBS-EB (1×DPBS, 2% FCS, HI, 2 mM EDTA) were added to designated wells of a non-binding 96-well plate. After three washing steps with PBS-EB, cells were centrifuged and the wash buffer was aspirated. Three-fold serial dilutions of the scFv and the reference anti-ROR1 Fab fragment PRO2213 (human Fab fragment comprising the VH / VL sequences of SEQ ID NO: 129 and 130 as disclosed in patent application US2017 / 0306018A1) were prepared starting from a concentration of 39 nM (scFv) or 63 nM (PRO2213) and then added to the plates with the cells. After 1 h of incubation at 4° C., the plates were washed three times with 100 μl of PBS-EB. The cells incubated with the scFv were resuspended in 100 μl of Numab's framework-specific detection antibody (PRO2268, rabbit IgG) and incubated for 30 min at 4° C. Following a washing step, PRO2268 was detected by the addition of an anti-rabbit IgG antibody labeled with APC at a concentration of 2 μg / ml and incubated for 1 h at 4° C. Cells incubated with the reference Fab fragment PRO2213 were resuspended in 100 μl of goat anti-human F(ab')2-AlexaFluor647-conjugated antibody (Jackson Immuno Research, cat.109-606-097) at a concentration of 2.8 μg / ml and incubated at 4°C for 1 h. Cells were then washed again three times with 100 μl PBS-EB per well. Cell pellets were resuspended in 50 μl PBS-EB and analyzed by a NovoCyte2060 flow cytometer device. The fluorescence intensity of the APC channel was recorded for each sample and the geometric mean of fluorescence intensity (MFI) was calculated.Data were single referenced (subtracted for the fluorescence intensity found in cells incubated with detection reagent alone) and subsequent concentration-response curves were fitted using 4-PLfit (GraphPad Prism software). PRO2213 was produced from a papain digest of PRO1842, a human IgG. PRO2213 could be produced in sufficient quantity and purity (>98% monomer content by SE-HPLC) for the intended use.

[0260] The apparent binding affinity to cell surface human ROR1 was assessed, and the four scFvs had EC for binding to human ROR1-expressing cells that was at most 3-fold lower than the reference Fab fragment PRO2213. 50 The two scFvs showed no binding to cells expressing human ROR1 (Table 11). To test for non-specific binding, all scFvs were tested for binding to human ROR1-negative MCF-7 cells by flow cytometry, and none of the scFvs tested demonstrated binding to MCF-7 cells. The anti-ROR1 scFvs PRO2060 and PRO2062 were assessed for apparent binding affinity to plasma membrane-based human ROR1 in three and two independent flow cytometry experiments, respectively. The average EC obtained in these experiments for PRO2060 and PRO2062 was 0.01 and 0.02, respectively. 50 and the mean relative EC 50 Values ​​are also shown in Table 11. Both scFvs demonstrated specific binding to human ROR1 positive MDA-MB-231 cells, whereas no binding was found to ROR1 negative MCF-7 cells (see FIG. 1). Compared to the reference Fab fragment PRO2213, PRO2060 showed approximately 3-fold lower apparent binding affinity and PRO2062 showed approximately 3-fold higher apparent binding affinity.

[0261] [Table 11]

[0262] PRO2060 and PRO2062 were measured in three and two independent experiments, respectively, and the average EC 50 , average rel.EC 50 , the mean rel.max binding values ​​were calculated. PRO2213 was used as the reference anti-ROR1 antibody.

[0263] Binding affinity by SPR The binding kinetics (including affinity) of scFv to recombinant human ROR1 protein (Fc tag, AcroBiosystems) was determined by SPR analysis with a T200 device (Biacore, Cytiva). Recombinant human ROR1 molecules were covalently immobilized on a carboxymethylated dextran surface (CM5 sensor chip, Biacore, Cytiva), and a titration series of each scFv was injected as an analyte. After each analyte injection cycle, every flow path on the sensor chip was regenerated (glycine, pH 2.0) and a new concentration of scFv was injected. The binding kinetics to human ROR1 was measured at nine analyte concentrations ranging from 0.005 to 30 nM diluted in running buffer (HEPES-buffered saline, 0.05% Tween-20, pH 7.5, Bioconcept) using a multi-cycle kinetic assay. The apparent dissociation (k d ) and meeting (k a ) rate constant and apparent dissociation equilibrium constant (K D ) was calculated using a one-to-one Langmuir binding model by Biacore analysis software (Biacore Evaluation Software version 3.2, Cytiva), and the quality of the fit was monitored based on the Chi and U values, which are measures of the quality of the curve fitting. The binding level was calculated based on the theoretical R max The stability of the sigma-based ...

[0264] All 13 scFvs were evaluated for their binding affinity to ROR1-expressing cells and recombinant hROR1. The data were correlated and are presented in Table 12.

[0265] Affinities for hROR1 were determined for eight scFvs. Affinities ranged from 160 nM to 0.098 nM. Three scFvs showed better affinity than the Fab fragment reference PRO2213. For three scFvs, affinity could not be determined due to technical reasons, and for two scFvs, no binding was observed (Table 12).

[0266] [Table 12]

[0267] Domains that demonstrated non-binding (SPR and cell binding) or equivocal data were excluded from further analysis. Four scFvs derived from three different clones that showed the best binding to human ROR1 expressing cells and the highest affinity to recombinant human ROR1ECD by SPR were selected for further biophysical characterization. These domains include PRO2060, PRO2062, PRO2059, and PRO2066.

[0268] Because cynomolgus ROR1 is 100% identical to human ROR1, no further binding assays were required to demonstrate cross-species binding of these domains.

[0269] 1.7. Biophysical characterization of scFv domains (storage stability and DSF) Storage stability The humanized scFv was subjected to a 4-week stability study in which the scFv was formulated at 10 mg / ml in an aqueous buffer (50 mM NaCiP, 150 mM NaCl, pH 6.4) and stored at temperatures <80° C., 4° C., and 40° C. over 28 days. The monomeric and oligomeric fractions in the formulations were assessed by integration of SE-HPLC peak areas at different times during the course of the study. Table 13 summarizes the percentage of monomer content and the percentage of monomer loss relative to day 0. The change in protein concentration was monitored over the course of the study by UV-Vis measurements at 280 nm. However, no notable changes in protein concentration were observed for any of the samples relative to day 0, so the data are not shown. Thermal stability was analyzed by nDSF (NanoTemper) to determine the onset of unfolding (T onset ) and the unfolding midpoint (T m ) was determined. The results are shown in Table 13.

[0270] PRO2059 and PRO2060 (derivatives of clone 55-38-D07) show a good stability profile, with only mild monomer loss at all tested temperatures and the best overall thermal stability of the four molecules. PRO2062 (derivative of clone 55-39-G02) is inferior to PRO2059 and PRO2060 in the extent of monomer loss and thermal stability at 40° C. PRO2066 (derivative of clone 55-42-E05) shows significant monomer loss at 4° C. and 40° C. and can therefore be considered the least stable molecule in this study (despite its acceptable thermal stability with a Tm of 70.6° C.).

[0271] In summary, PRO2060 and PRO2062 were carried forward as lead domains. These domains were selected over other domains based on a combination of pharmacodynamic and biophysical characteristics. Both molecules bind to different epitopes in ROR1ECD region 1, allowing for additional candidate diversity. These lead domains were carried forward for optimization by protein engineering to improve T cell score and stability as outlined in the next section.

[0272] 1.8. Creation of optimized single-chain domains PRO2062 has a relatively high in silico calculated T cell score (1,662), which indicates a higher probability of stimulating an immune response when introduced into humans. To reduce the T cell score, one of the CDRs (CDR1H) was modified. In rabbits, the first half of CDR1H is more conserved and the second half is more variable. Unlike all other common CDR definitions (IMGT, Chotia, and Honegger), the Kabat CDR definition only encompasses the variable second half of CDR1H. To minimize the risk of blocking binding, two mutations were introduced into the conserved CDR1H portion, because it is likely that the variable CDR1H portion is more important for the interaction with the antigen. By analyzing sequences from the same project, we were able to identify a CDR1H sequence from another clone that lowered the T cell score and differed by only two amino acids in the conserved CDR1H portion. In addition, this CDR1H sequence is similar to the PRO2062 variable CDR1H sequence, indicating that two mutations (L27I and S28D) in the conserved CDR1H portion may fit well into the protein structure. Furthermore, in silico mutagenesis indicates that the two mutations are not destabilizing. In addition to the T cell reducing mutations, the so-called outer loop of the donor heavy chain was grafted, because it is important for the correct positioning of CDR3H and may serve as a compensation for the modifications on CDR1H. Briefly, a segment on FW3H (AHo residues 82-87) was replaced by KTST (R82K, D83T, N84S, N87T (S85 and K86 on the framework were deleted)), AHo89 was changed from Val to Leu (VH3 consensus), and AHo90 was changed from Tyr to Asp (donor). The resulting T cell score of PRO2271, which has T cell score reducing mutations on the CDR1H and outer loops, is 1 160. One additional variant (PRO2272, 55-39-G02-sc04) was created with the same T cell score reducing mutations and outer loops as PRO2271, but with fewer donor residues than PRO2271.PRO2292, ie, PRO2062, which has a disulfide between AHo141VL-51VH, has a relatively high T cell score value (1,662).

[0273] [Table 13]

[0274] PRO2060 had a T cell score of 949 and therefore no T cell score optimization was required. PRO2291, which has a disulfide between AHo141VL-51VH, has the same T cell score value as PRO2060 of 949. The optimized variants are summarized in Table 14.

[0275] [Table 14]

[0276] 1.9. Characterization of the optimized single-chain domain Cellular binding of optimized scFv to ROR1-expressing MDA-MB-231 cells by flow cytometry Plasma membrane binding of the optimized scFvs to human ROR1-expressing MDA-MB-231 cells by flow cytometry was performed as described above. Briefly, 5-fold serial dilutions of the optimized scFvs of anti-ROR1 clone 55-39-G02 (PRO2271, PRO2292) and clone 55-38-D07 (PRO2291), as well as the reference Fab fragment PRO2213, were prepared starting at a concentration of 40 nM and then added to the plate with cells. After 1 h of incubation at 4°C, the plate was washed and incubated with specific detection antibodies. The fluorescence intensity of the APC channel was recorded for each sample using a NovoCyte2060 flow cytometer, and the geometric mean MFI of the fluorescence intensity was calculated.

[0277] The apparent binding affinities of the optimized scFvs PRO2271, PRO2291, and PRO2292 to plasma membrane human ROR1 were assessed by flow cytometry experiments. The calculated EC 50 ,rel.EC 50 , and rel.max binding values ​​are shown in Table 15. As shown in Figure 2, all scFvs demonstrated binding to ROR1 positive MDA-MB-231 cells, but no binding was found to ROR1 negative MCF-7 cells. PRO2271, a variant of PRO2062 with T cell score optimization and outer loop grafting, showed approximately 2-fold lower apparent binding affinity when compared to the original scFv PRO2062. PRO2291 and PRO2292, variants of PRO2060 and PRO2062 that incorporate a VL-VH interdomain disulfide bond for stabilization, demonstrated apparent binding affinity comparable to that of the parent scFvs PRO2060 and PRO2062.

[0278] [Table 15]

[0279] Binding affinity by SPR The binding kinetics (including affinity) of the optimized scFvs PRO2071, PRO2291, and PRO2292 to recombinant human ROR1 protein (Fc tag, AcroBiosystems) was determined by SPR analysis with a T200 device (Biacore, Cytiva) as described in section 1.6.

[0280] Affinity for human ROR1 was determined along with a non-optimized version as a comparison. The results of the SPR analysis are summarized in Table 16.

[0281] [Table 16]

[0282] 1.10. Biophysical characterization of scFv domains (storage stability and DSF) Optimized variants of humanized scFv were subjected to a 4-week stability study, in which the scFv was formulated at 10 mg / ml in aqueous buffer (50 mM NaCiP, 150 mM NaCl, pH 6.4) and stored at temperatures <-80°C, 4°C, and 40°C over a 4-week period. Monomeric and oligomeric fractions in the formulations were assessed by integration of SE-HPLC peak areas at different time points during the course of the study. Table 17 summarizes the % monomer content and % monomer loss relative to d0. Changes in protein concentration were monitored over the course of the study by UV-Vis measurements at 280 nm. No notable differences were observed for any of the samples relative to d0, so data are not shown. Thermal stability was analyzed by nDSF (NanoTemper) and the onset of unfolding (T onset ) and the unfolding midpoint (T m ) was determined. The results are shown in Table 17.

[0283] The optimized derivative of clone 55-39-G02 (PRO2271) shows similar stability to PRO2062. Both the derivatives of clones 55-38-D07 and 55-39-G02 carrying the VL-VH disulfide (PRO2291 and PRO2292) show improved stability compared to their disulfide-less counterparts (PRO2060 and PRO2062) and can be considered the overall most stable scFv domains.

[0284] Example 2: Generation and testing of anti-CD3 molecules: The identification, selection and humanization of anti-CD3 rabbit IgG clone 28-21-D09, as well as the production and characterization of the corresponding humanized anti-CD3 scFv 28-21-D09-sc04 (PRO726) were performed as described in patent application WO2018 / 224441, which is incorporated herein by reference. The characterization of anti-CD3 scFv PRO726 is briefly outlined below.

[0285] 2.1. Characterization of anti-CD3 scFv 28-21-D09-sc04 (PRO726) Binding affinity and species cross-reactivity (FS) The binding kinetics (including affinity) of the selected binding domain 28-21-D09-sc04 (PRO726) to recombinant human CD3 epsilon extracellular domain protein (hCD3ε, His tag, SinoBiological) was determined by SPR analysis with a T200 device (Biacore, Cytiva). Recombinant hCD3ε molecules were covalently immobilized on a carboxymethylated dextran surface (CM5 sensor chip, Biacore, Cytiva) and a titration series of PRO726 was injected as the analyte. After each analyte injection cycle, all flow paths on the sensor chip were regenerated (glycine, pH 2.0) and a new concentration of PRO726 was injected.

[0286] [Table 17]

[0287] [Table 18]

[0288] Binding kinetics to hCD3ε was measured using a multi-cycle kinetic assay with eight analyte concentrations ranging from 0.7 to 90 nM diluted 1:2 in running buffer (HEPES-buffered saline, 0.05% Tween-20, pH 7.5, Bioconcept). Apparent dissociation (k d1 ) and meeting (k a1 ) rate constant, second-order reaction constant (k a2 and k d2 ) and the apparent dissociation equilibrium constant (K D ) was calculated using a two-state binding model by Biacore analysis software (Biacore Evaluation Software version 3.2, Cytiva), and the quality of the fit was monitored based on the relative chi2. The binding level was calculated as the maximum stable binding achieved normalized to the theoretical Rmax.

[0289] Binding kinetics to recombinant cynomolgus CD3ε (cCD3ε, His tag, SinoBiological) was measured as mentioned above for hCD3ε, except that cCD3ε was used instead of hCD3ε. The results are summarized in Table 18.

[0290] Biophysical characterization The humanized CD3 binding domain was subjected to a 4-week stability study in which the scFv was formulated at 10 mg / ml in an aqueous buffer (50 mM NaCiP, 150 mM NaCl, pH 6.4) and stored at temperatures <-80°C, 4°C, and 40°C over a 4-week period. The monomeric and oligomeric fractions in the formulations were assessed by integration of SE-HPLC peak areas at different time points during the course of the study. Table 19 summarizes the % monomer content and the % monomer loss relative to d0. The change in protein concentration was monitored over the course of the study by UV-Vis measurements at 280 nm. There was no notable loss of protein content observed for any of the samples relative to d0. Data not shown.

[0291] The thermal unfolding data obtained from differential scanning fluorimetry (DSF) measurements are shown in Table 19. The resulting thermal unfolding midpoint (Tm) and onset temperature of unfolding (Tonset) were determined by fitting the data to the Boltzmann equation.

[0292] [Table 19]

[0293] Example 3: Generation and testing of anti-hSA molecules: The identification, selection, humanization, and production and characterization of the humanized anti-hSA binding domains 19-01-H04-sc03 (PRO0325) and 23-13-A01-sc03 (PRO0459) were performed as described in patent application EP192069599, which is incorporated herein by reference. The identification, selection, humanization, and production and characterization of the humanized anti-hSA binding domain 19-04-A10-sc02 (PRO2155) were performed analogously to the procedure described in patent application EP192069599. The characterization of the anti-CD3 scFv PRO2155 is briefly outlined below.

[0294] 3.1. Characterization of anti-hSAscFv 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 at both pH 7.4 and pH 5.5 by SPR analysis with a T200 device (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 an analyte. After each analyte injection cycle, all flow paths on the sensor chip were regenerated (glycine, pH 2.0) and a new concentration of scFv molecules was injected. Binding kinetics to hSA was measured using a multi-cycle kinetic assay with 11 concentrations ranging from 0.044 to 45 nM diluted (1:2) in the appropriate running buffer (PBS, 0.05% Tween-20, or PBS, 0.05% Tween-20, pH 5.5). d ) and meeting (k a ) rate constant and apparent dissociation equilibrium constant (K D ) was calculated using a one-to-one Langmuir binding model by Biacore analysis software (Biacore Evaluation Software version 3.2, Cytiva), and the quality of the fit was monitored based on the relative chi2. The binding level was calculated as the maximum stable binding achieved normalized to the theoretical Rmax.

[0295] 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, with the difference 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 20.

[0296] [Table 20]

[0297] Biophysical characterization 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) were subjected to a 4-week stability study, in which the scFvs were formulated at 10 mg / ml in aqueous buffer (50 mM NaCiP, 150 mM NaCl, pH 6.4) and stored at temperatures <-80°C, 4°C, and 40°C over a 4-week period. The monomeric and oligomeric fractions in the formulations were assessed by integration of SE-HPLC peak areas at different times during the course of the study. Table 21 summarizes the % monomer content and % monomer loss relative to d0. The change in protein concentration was monitored over the course of the study by UV-Vis measurements at 280 nm. No significant protein content loss was observed for any of the samples relative to d0, so data are not shown. Thermal stability was analyzed by nDSF (NanoTemper) to determine the onset of unfolding (Tonset) and the midpoint of unfolding (Tm). DSF results are shown in Table 21.

[0298] [Table 21]

[0299] Example 4: Creation and production of scMATCH3 anti-ROR1 x CD3 x hSA multispecific antibodies The HSA, CD3, and ROR1 binding domains were combined in the trispecific scMATCH3 format, which consists exclusively of variable domains connected by GS linkers of different lengths as illustrated in Figure 3. In this format, the split variable domains are located on a single peptide chain (sc), which assemble into a fully functional trispecific molecule. Similar to scFv domains, VL / VH disulfide bonds can be incorporated to stabilize the individual domains. ScMATCH3 molecules can be recombinantly expressed by mammalian cells, and conventional affinity chromatography steps can be used for their purification.

[0300] Expression of mammalian scMATCH3 constructs was performed in CHO-S cells using the CHOgro Transient Transfection Kit (Mirus). Cultures were harvested after 5-7 days of expression at 37°C (when cell viability <70% was reached) by centrifugation and then filtered. Proteins were purified from clarified culture supernatants by protein L affinity chromatography followed by size-exclusion chromatography (SEC) in 50 mM phosphate-citrate buffer with 300 mM sucrose at pH 6.5. The monomer content of the SEC fractions was assessed by SE-HPLC analysis and fractions with a monomer content >95% were pooled. For quality control of the produced material, standard analytical methods, e.g. SE-HPLC, UV, HPLC ... 280 , and SDS-PAGE were applied. An overview of the molecular domain composition and production of the scMATCH3 molecule is shown in Table 22.

[0301]

[0234] The scMATCH3 molecules PRO2667 and RPO2668, containing modified or improved ROR1, CD3, and hSA binding domains, were produced at 1 L scale at Evitria AG (Schlieren, Switzerland) using their proprietary mammalian expression system. Proteins were purified from clarified culture supernatants by Protein L (CaptoL, Cytiva) affinity chromatography followed by SEC in 50 mM phosphate-citrate buffer with 300 mM sucrose at pH 6.5. The monomer content of the SEC fractions was assessed by SE-HPLC analysis and fractions with a monomer content >95% were pooled. For quality control of the produced material, standard analytical methods, e.g. SE-HPLC, UV, HPLC ... 280 and SDS-PAGE were applied. An overview of the preparation of these scMATCH3 molecules is shown in Table 23. The thermal stability of the molecules was evaluated by nDSF using a Prometheus NT.48 device (NanoTemper).

[0302] [Table 22]

[0303] [Table 23]

[0304] Example 5: Pharmacodynamic characterization of scMATCH3 molecules 5.1.Binding affinity by SPR The binding kinetics of nine multispecific scMATCH3 molecules to the three targets human ROR1 (hROR1), human CD3ε (hCD3ε), and hSA were examined by SPR.

[0305] Binding affinity to hROR1 by SPR The binding kinetics (including affinity) of nine scMATCH3 molecules to recombinant human ROR1 protein (Fc tag, AcroBiosystems) was determined by SPR analysis with a T200 device (Biacore, Cytiva) as described in section 1.6. Recombinant human ROR1 molecules were covalently immobilized on a carboxymethylated dextran surface (CM5 sensor chip, Biacore, Cytiva), and a titration series of each MATCH3 molecule was injected as an analyte. After each analyte injection cycle, every flow path on the sensor chip was regenerated (glycine pH 2.0 and 3 M MgCl2) and a new concentration of MATCH3 molecules was injected. The binding kinetics to human ROR1 was measured using a multi-cycle kinetic assay with nine analyte concentrations ranging from 0.088 to 45 nM diluted 1:2 in running buffer (HEPES-buffered saline, 0.05% Tween-20, pH 7.5, Bioconcept). Apparent dissociation (k d ) and meeting (k a ) rate constant and apparent dissociation equilibrium constant (K D ) was calculated using a one-to-one Langmuir binding model by Biacore analysis software (Biacore Evaluation Software version 3.2, Cytiva), and the quality of the fit was monitored based on Chi2 and U values, which are measures of the quality of the curve fitting. The binding level was calculated as the maximum stable binding achieved normalized to the theoretical Rmax.

[0306] Binding kinetics for hROR1 were determined for nine scMATCH3, with affinities ranging from 1.6 nM to 0.12 nM (Table 24). The hROR1 affinity of all tested scMATCH3 molecules was similar to their respective parental scFvs.

[0307] [Table 24]

[0308] Binding affinity to hCD3ε by SPR The binding kinetics (including affinity) of nine selected scMATCH3 molecules to recombinant human CD3 epsilon extracellular domain protein (hCD3ε, His-tag, SinoBiological) was determined by SPR analysis with a T200 device (Biacore, Cytiva). Recombinant hCD3ε molecules were covalently immobilized on a carboxymethylated dextran surface (CM5 sensor chip, Biacore, Cytiva), and a titration series of each MATCH3 molecule was injected as an analyte. After each analyte injection cycle, every flow path on the sensor chip was regenerated (glycine pH 2.0 and 3 M MgCl2) and a new concentration of MATCH3 molecules was injected. The binding kinetics to hCD3ε was measured using a multi-cycle kinetic assay with nine analyte concentrations ranging from 0.088 to 45 nM diluted 1:2 in running buffer (HEPES-buffered saline, 0.05% Tween-20, pH 7.5, Bioconcept). Apparent dissociation (k d1 ) and meeting (k a1 ) rate constant, second-order reaction constant (k a2 and k d2 ), and the apparent dissociation equilibrium constant (K D ) was calculated using a two-state binding model by Biacore analysis software (Biacore Evaluation Software version 3.2, Cytiva), and the quality of the fit was monitored based on the relative chi2. The binding level was calculated as the maximum stable binding achieved normalized to the theoretical Rmax.

[0309] Binding kinetics to hCD3ε was determined for nine MATCH3, with affinities ranging from 9.4 nM to 1.5 nM (Table 25). The hCD3ε affinity of all tested MATCH3 molecules was similar to that of the parental scFv 28-21-D09-sc04.

[0310] [Table 25]

[0311] Binding affinity to hSA by SPR The binding kinetics (including affinity) of seven selected scMATCH3 molecules to human serum albumin (hSA, Sigma-Aldrich) was determined by SPR analysis with a T200 device (Biacore, Cytiva). hSA molecules were covalently immobilized on a carboxymethylated dextran surface (CM5 sensor chip, Biacore, Cytiva) and a titration series of each MATCH3 molecule was injected as analyte (Exp. ND032-0033). After each analyte injection cycle, every flow path on the sensor chip was regenerated (glycine, pH 2.0) and a new concentration of MATCH3 molecules was injected. The binding kinetics to hSA was measured using a multi-cycle kinetic assay with 10 analyte concentrations ranging from 0.352 to 180 nM diluted 1:2 with the corresponding running buffer (PBS, 0.05% Tween-20, pH 5.5). The apparent dissociation (k d ) and meeting (k a ) rate constant and apparent dissociation equilibrium constant (K D ) was calculated using a one-to-one Langmuir binding model by Biacore analysis software (Biacore Evaluation Software version 3.2, Cytiva), and the quality of the fit was monitored based on the relative Chi2 and U values, which are measures of the quality of the curve fitting. The binding level was calculated as the maximum stable binding achieved normalized to the theoretical Rmax.

[0312] Binding kinetics to hSA was determined for seven MATCH3 molecules (Table 26). The hSA affinity at pH 5.5 of all tested MATCH3 molecules is deemed sufficient to extend the systemic half-life of the MATCH3 molecules.

[0313] [Table 26]

[0314] 5.2. Cellular binding of scMATCH3 molecules to human ROR1 by flow cytometry

[0242] Cell binding of anti-ROR1 scMATCH3 molecules to ROR1-expressing MDA-MB-231 cells by flow cytometry was performed as described in section 1.6. Briefly, 5-fold serial dilutions of scMATCH3 molecules starting from 40 nM (PRO2286, PRO2287) or 120 nM (PRO2507, PRO2508, PRO2509, PRO2510, PRO2557) and of the reference Fab fragment PRO2213 starting from a concentration of 40 nM or 120 nM were prepared and then added to the plate with cells. After 1 h of incubation at 4°C in the presence of 1 mg / ml human serum albumin (hSA), the plate was washed and incubated with specific detection antibodies. The fluorescence intensity of the APC channel was recorded for each sample using a NovoCyte2060 flow cytometer, and the geometric mean MFI of the fluorescence intensity was calculated.

[0315] Binding of anti-ROR1xhSAxCD3scMATCH3 molecules to plasma membrane human ROR1 was assessed by flow cytometry. The calculated EC 50 ,rel.EC 50 , and rel.max binding values ​​are shown in Table 27. As shown in Figure 4, scMATCH3 molecules PRO2286 and PRO2287, both molecules harboring the same anti-ROR1-BD 55-39-G02-sc02 but different anti-hSA-BD, demonstrated binding to ROR1-positive MDA-MB-231 cells, but no binding was found to ROR1-negative MCF-7 cells. The rel.EC of the parental scFv PRO2062 (anti-ROR1 domain 55-39-G02-sc02) 50 Compared to the values, both scMATCH3 molecules had approximately 5-fold lower EC 50 Binding was demonstrated (rel.EC of PRO2286 50 =0.63, rel.EC of PRO2287 50=0.47, average rel.EC of PRO2062 50 = 2.77). This reduced EC 50 It is hypothesized that the binding is due to steric hindrance caused by hSA bound to the scMATCH3 molecule. Along those lines, other scMATCH3 molecules harboring either the anti-ROR1 domain 55-39-G02-sc02 (PRO2509) or its optimized variants (PRO2510: 55-39-G02-sc05, PRO2557: 55-39-G02-sc06) showed EC50 values ​​that were 5- to 10-fold reduced when compared to their respective scFvs. 50 In detail, the rel.EC of PRO2509, PRO2510, and PRO2557 were 50 The values ​​were 0.53 (2.77 for the corresponding scFv PRO2062), 0.3 (2.61 for PRO2292), and 0.21, respectively.

[0316]

[0244] The optimized variant of PRO2510, PRO2668, was measured using the same experimental setup by flow cytometry. In this experiment, PRO2668 showed binding to MDA-MB-231 cells comparable to PRO2510 as indicated by similar binding titers and rel.max binding values. Dose-response curves of these anti-ROR1xhSAxCD3 scMATCH3 molecules can be found in Figures 4 and 5. For scMATCH3 molecules assembled with anti-ROR1 domains 55-38-D07-sc02 (PRO2507) and 55-38-D07-sc06 (PRO2508), their EC 50 Binding was also found to be reduced when compared to their respective scFvs. 50 The values ​​were 0.03 (0.33 for the corresponding scFv PRO2060) and 0.04 (0.27 for PRO2291), respectively. The concentration-response curves of these anti-ROR1xhSAxCD3scMATCH3 molecules can be seen in FIG.

[0317] In summary, all scMATCH3 molecules demonstrated specific binding to human ROR1-expressing cells and had higher EC 50 Binding was found for molecules harboring anti-ROR1 clone 55-39-G02 when compared to a series of molecules assembled with anti-ROR1 clone 55-38-D07. Five to ten-fold lower binding was found for scMATCH3 molecules when compared to their corresponding scFvs, likely due to steric hindrance caused by hSA bound to the scMATCH3 molecules. The results of the cell binding data are summarized in Table 27. Calculated EC 50 ,rel.EC 50 , and rel. maximal binding values ​​are shown. Fab fragment PRO2213 was used as reference anti-ROR1 antibody.

[0318] [Table 27]

[0319] Example 6: Potency assessment: target cell cytotoxicity and T cell activation mediated by scMATCH3 molecules To understand the capacity of these scMATCH3 molecules to promote tumor cell killing and CD8+ T cell activation, we performed a series of cytotoxicity assays to compare the potency and T cell activation profiles among our lead candidates.

[0320] First, we tested the expression levels of ROR1 on several cell lines along with quantification of relative receptor levels on the surface (Figure 6). Cells were incubated with ZombieNIR Live / Dead-APC-Cy7 (Biolegend, 1:500) for 20 min at 4°C, followed by a washing step, then incubation with anti-ROR1-PE (Biolegend, 1:25) for 20 min at 4°C, followed by washing and acquisition on a Novocyte flow cytometer (ACEA). Figure 6A demonstrates surface ROR1 expression on MDA-MB-231 and MCF7 cells. Receptor levels were quantified using the Quantum Simply Cellular Anti-Mouse IgG kit (Bangs Laboratories) according to the manufacturer's instructions. Beads and cells were acquired on an AttuneNxT flow cytometer (Thermo Fisher) and data were analyzed using FlowJo software (BD). The relative surface level of ROR1 was higher on MDA-MB-231 cells compared to MCF7 cells. The surface level of ROR1 on MCF7 cells was very low and considered to be comparable to healthy adult human tissues. Therefore, MCF7 cells were used as a negative control (Figure 6B).

[0321] These assays were performed using peripheral blood mononuclear cells (PBMCs) isolated from buffy coats using SepMate tubes on Lymphoprep (StemCell Technologies) to create density gradients according to the manufacturer's instructions.

[0322]

[0249] Isolated PBMCs were co-cultured with ROR1 positive and negative tumor cells of interest at an effector to target ratio of 10:1 for 40 hours. After co-culture, 100 μl of supernatant was aspirated and tested for lactate dehydrogenase (LDH) release from dying cells using a Roche cytotoxicity detection kit. The remaining supernatant was frozen at -80°C for downstream multiplex cytokine analysis. Cells were labeled with the following reagents or antibodies (all from Biolegend) for flow cytometric analysis of specific T cell activation: Live / Dead fixable Aqua, anti-human CD4-APC-Cy7 (clone OKT4), anti-human CD11c-PE-Cy7 (clone Bu15), anti-human CD8-PerCP-Cy5.5 (clone SK1), and anti-human CD69-PE (clone FN50). Cells were acquired by flow cytometer (Attune, Thermo Fisher Scientific) and analyzed for upregulation of CD69 as a measure of both CD4 and CD8 T cell activation using FlowJo software (BD).

[0323] All scMATCH3 molecules tested resulted in specific killing of ROR1-positive cells, such as MDA-MB-231, with no to minimal activity observed against the ROR1-negative cell line MCF7 (Figure 7A). PRO2060-derived scMATCH3 molecules PRO2507 and PRO2508 displayed slightly less potency compared to PRO2062-derived molecules PRO2509 and PRO2510. This was consistent with the higher affinity observed for PRO2062. CD8 and CD4 T cell activation very closely matched the cytotoxicity data (Figures 7B and 7C).

[0324] The scMATCH3 molecule PRO2557, whose domains are derived from PRO2062 and T cell score-optimized to reduce immunogenicity, also mediated specific killing of ROR1-positive cells (Figure 8A). As with other scMATCH3 molecules, CD8 and CD4 T cell activation correlated with the observed cytotoxicity (Figures 8B and 8C, respectively). Table 28 shows the resulting mean EC20 specific killing of MDA-MB-231 cells for the scMATCH3 molecules across experiments. 50 Summarizes values.

[0325] [Table 28]

[0326] Example 7: scMATCH3 molecular biophysical characterization ScMATCH3 molecules were subjected to a 4-week stability study, in which the molecules were formulated at 1 mg / ml in aqueous buffer (50 mM NaCiP, 300 mM sucrose, pH 6.5) and stored at <-80°C, 4°C, and 40°C over 4 weeks. Monomeric and oligomeric fractions in the formulations were assessed by integration of SE-HPLC peak areas at different time points during the course of the study. Table 29 summarizes the % monomer content and % monomer loss relative to d0. Changes in protein concentration were monitored over the course of the study by UV-Vis measurements at 280 nm and are shown in Table 30. Thermal stability was analyzed by nDSF (NanoTemper) and the onset of unfolding (T onset ) and the unfolding midpoint (T m The nanoDSF unfolding midpoint (T m ) The results are shown in the last column of Table 29.

[0327] [Table 29]

[0328] [Table 30]

[0329] Molecules containing the anti-ROR1 domain from 55-38-D07 (PRO2507 and PRO2508) and the anti-ROR1 domain from 55-39-G02 (PRO2509 and PRO2510) show very similar monomer stability over the course of 28 d at a concentration of 1 mg / ml. There was no significant change in monomer content at temperatures of -80°C and 4°C, as well as by repeated freeze-thawing (5x) performed before SE-HPLC / UV measurement at d28 / -80°C.

[0330] SDS-PAGE analysis of day 0 (d0) and day 28 (d28) samples revealed the presence of low molecular weight (LMW) species in all (reduced) samples at d0 of the study, besides the main band corresponding to the calculated molecular weight of the molecule. The reason for their occurrence is not entirely clear, but may be related to the generally non-optimized expression and purification conditions used to produce the molecule or to unfavourable SDS-PAGE sample preparation conditions. In general, no significant increase in LWM species in the 4°C and 40°C samples could be detected over time when comparing d28 samples to d0 samples. In all samples incubated under accelerated conditions (40°C) over 28 days, the fraction of LMW species clearly increased, indicating a time- and temperature-dependent fragmentation of the scMATCH3 molecule. Since all molecules show a similar favourable stability profile, exclusion of molecules based on accelerated stability data is not expected.

[0331] PRO2509 and PRO2510 appear to be slightly better than PRO2507 and PRO2508 in terms of thermal stability as assessed by nDSF because their onset of unfolding (T onset ) is considerably higher for molecules containing the 55-39-G02-derived ROR1 domain.

[0332] Example 8: Creation and production of MATCH4 anti-ROR1 x CD3 x hSA multispecific antibodies The ROR1-binding domain from the scMATCH3 molecule was incorporated into the MATCH4 molecule format to further enhance potency through the use of two ROR1-binding domains, which would result in bivalent targeting similar to conventional full-length antibodies, but with the added benefits of T cell targeting (CD3 domain) and extended half-life (hSA domain).

[0333] MATCH4 is a format developed by Numab, which consists exclusively of variable domains connected by GS linkers of different lengths that allow only specific pairing of matching domain pairs. The MATCH4 format can be recombinantly expressed from mammalian cells. For purification, a conventional affinity chromatography step can be used. The architecture of the ND032-MATCH4 molecule is illustrated in Figure 9. This format requires that the dimeric subunit consists of a core of two split variable domain pairs, each respective subunit possessing either two VL domains or two VH domains located in tandem, thereby driving the heterodimerization of the two protein chains. The dimer-forming tandem variable domains on each MATCH4 chain are organized in an antiparallel N-terminus-C-terminus orientation with their counterpart chains. Both chains are co-expressed by mammalian cells into a fully functional tetraspecific molecule. A conventional Gly-Ser linker between the variable domains was used to connect them. Furthermore, the antiparallel MATCH4 format is amenable to the introduction of a disulfide bridge in one of the core domains as indicated in Figure 9. Additional VL / VH disulfide bonds can be included for stabilization of individual domains.

[0334] Expression of mammalian MATCH4 constructs was performed at 1 l scale at Evitria AG (Schlieren, Switzerland) using their proprietary mammalian expression system. Proteins were purified from clarified culture supernatants by protein L affinity chromatography followed by SEC in 50 mM phosphate-citrate buffer with 300 mM sucrose at pH 6.5. The monomer content of the SEC fractions was assessed by SE-HPLC analysis and fractions with a monomer content >95% were pooled. For quality control of the produced material, standard analytical methods, e.g. SE-HPLC, UV, HPLC ... 280 , and SDS-PAGE were applied. The molecular composition and preparation summary of the MATCH4 molecules are shown in Table 31. The preparation summary of the MATCH4 molecules PRO2669 and RPO2670 containing modified or improved ROR1, CD3, and hSA binding domains is shown in Table 32.

[0335] [Table 31]

[0336] [Table 32]

[0337] Example 9: Pharmacodynamic characterization of MATCH4 molecules 9.1. Binding affinity by SPR The binding kinetics of the four multispecific MATCH4 molecules to the three targets hROR1, hCD3ε, and hSA were examined by SPR.

[0338] Binding affinity to hROR1 by SPR The binding kinetics (including affinity) of four sMATCH4 molecules to recombinant human ROR1 protein with or without avidity was determined in two different assay designs by SPR analysis with a T200 device (Biacore, Cytiva).

[0339]

[0261] The avidity-free binding affinity was determined by capturing MATCH4 molecules on the surface using a proprietary anti-framework rabbit IgG (rIgG21-57-20-B11) immobilized on a carboxymethylated dextran surface (CM5 sensor chip, Biacore, Cytiva). A titration series of human ROR1 (Peprotech160-054) molecules was injected as analyte. After each capture-analyte injection cycle, every flow path on the sensor chip was regenerated (3M MgCl2), MATCH4 molecules were captured again, and then different concentrations of human ROR1 were captured. The binding kinetics to human ROR1 was measured using a multi-cycle kinetic assay with 10 analyte concentrations ranging from 0.088 to 90 nM diluted 1:2 with running buffer (HEPES-buffered saline, 0.05% Tween-20, pH 7.5, Bioconcept).

[0340] Avidity binding was determined similarly to the procedure described in section 1.6. Recombinant human ROR1 molecules (Fc tag, AcroBiosystems RO1-H5250) were covalently immobilized on a carboxymethylated dextran surface (CM5 sensor chip, Biacore, Cytiva), and a titration series of each MATCH4 molecule was injected as analyte. After each analyte injection cycle, every flow path on the sensor chip was regenerated (glycine pH 2.0 and 3 M MgCl2) and a new concentration of MATCH4 molecules was injected. Binding kinetics to human ROR1 was measured using a multi-cycle kinetic assay with nine analyte concentrations ranging from 0.088 to 45 nM diluted 1:2 in running buffer (HEPES-buffered saline, 0.05% Tween-20, pH 7.5, Bioconcept).

[0341] In both assays, the apparent dissociation (k d ) and meeting (k a ) rate constant and apparent dissociation equilibrium constant (K D) was calculated using a one-to-one Langmuir binding model by Biacore analysis software (Biacore Evaluation Software version 3.2, Cytiva), and the quality of the fit was monitored based on Chi2 and U values, which are measures of the quality of the curve fitting. The binding level was calculated as the maximum stable binding achieved normalized to the theoretical Rmax.

[0342] Binding kinetics to hROR1 was determined for four MATCH4 molecules with affinities ranging from 5 nM to 1.2 nM. hROR1 affinities with avidity were determined for four MATCH4 molecules and ranged from 136 pM to 22 pM. SPR results are summarized in Table 33.

[0343] Binding affinity to hCD3ε by SPR The binding kinetics (including affinity) of four MATCH4 molecules to recombinant human CD3 epsilon extracellular domain protein (hCD3ε, His tag, SinoBiological) was determined by SPR analysis with a T200 device (Biacore, Cytiva). Recombinant hCD3ε molecules were covalently immobilized on a carboxymethylated dextran surface (CM5 sensor chip, Biacore, Cytiva), and a titration series of each MATCH4 molecule was injected as an analyte. After each analyte injection cycle, every flow path on the sensor chip was regenerated (glycine pH 2.0 and 3 M MgCl2) and a new concentration of MATCH4 molecules was injected. The binding kinetics to hCD3ε was measured using a multi-cycle kinetic assay with nine analyte concentrations ranging from 0.088 to 45 nM diluted 1:2 in running buffer (HEPES-buffered saline, 0.05% Tween-20, pH 7.5, Bioconcept). Apparent dissociation (k d1 ) and meeting (k a1 ) rate constant, second-order reaction constant (k a2 and k d2 ) and the apparent dissociation equilibrium constant (K D) was calculated using a two-state binding model by Biacore analysis software (Biacore Evaluation Software version 3.2, Cytiva), and the quality of the fit was monitored based on the relative chi2. The binding level was calculated as the maximum stable binding achieved normalized to the theoretical Rmax.

[0344] Binding kinetics to hCD3ε was determined for four MATCH4 with affinities ranging from 6.3 nM to 4.7 nM (Table 34). The hCD3ε affinity of all tested MATCH4 molecules was similar to that of the parental scFv 28-21-D09-sc04.

[0345] Binding affinity to hSA by SPR The binding kinetics (including affinity) of four MATCH4 molecules to human serum albumin (hSA, Sigma-Aldrich A3782) was determined by SPR analysis with a T200 device (Biacore, Cytiva). Human serum albumin molecules were covalently immobilized on a carboxymethylated dextran surface (CM5 sensor chip, Biacore, Cytiva), and a titration series of each MATCH4 molecule was injected as an analyte. After each analyte injection cycle, every flow path on the sensor chip was regenerated (pulse of glycine pH 1.5 and 3 M MgCl2) and a new concentration of MATCH4 molecules was injected. The binding kinetics to hSA was measured using a multi-cycle kinetic assay with 10 analyte concentrations ranging from 0.18 to 90 nM diluted 1:2 in running buffer (PBS-P+pH 5.5, PBS pH 5.5 with HCl, 0.05% Tween 20). The apparent dissociation (k d ) and meeting (k a ) rate constant and apparent dissociation equilibrium constant (K D) was calculated using a one-to-one Langmuir binding model by Biacore analysis software (Biacore Evaluation Software version 3.2, Cytiva), and the quality of the fit was monitored based on Chi2 and U values, which are measures of the quality of the curve fitting. The binding level was calculated as the maximum stable binding achieved normalized to the theoretical Rmax.

[0346] Binding rates to hSA were determined for the four MATCH4 with affinities ranging from 6.7 nM to 5.5 nM (Table 35). The hSA affinity of all tested MATCH4 molecules was similar to that of the parent scFv 19-04-A10-sc02.

[0347] 9.2. Cellular binding of MATCH4 molecules to human ROR1 by flow cytometry Cell binding of anti-ROR1MATCH4 molecules (PRO2589, PRO2590, PRO2591, PRO2592, and PRO2658) to ROR1-expressing MDA-MB-231 cells by flow cytometry was performed as described above. Briefly, 5-fold serial dilutions of MATCH4 molecules starting at 120 nM and 5-fold serial dilutions of reference Fab fragment PRO2213 starting at a concentration of 120 nM were prepared and then added to the plate with cells. After 1 h of incubation at 4°C in the presence of 1 mg / ml human serum albumin (hSA), the plate was washed and incubated with specific detection antibodies. The fluorescence intensity of the APC channel was recorded for each sample using a NovoCyte2060 flow cytometer, and the geometric mean MFI of the fluorescence intensity was calculated.

[0348] Binding of MATCH4 molecules to plasma membrane-based human ROR1 was assessed by flow cytometry. The calculated EC 50 ,rel.EC 50, and rel. max binding values ​​are shown in Table 36. Fab fragment PRO2213 was used as the reference anti-ROR1 antibody. All MATCH4 molecules demonstrated binding to ROR1 positive MDA-MB-231 cells, but no binding was found to ROR1 negative MCF-7 cells (see Figure 10). EC 50 Binding was determined based on the EC 50 For example, the rel.EC value of PRO2589 was increased by 10-fold. 50 is 0.45, which is the rel.EC of PRO2507 in scMATCH3. 50 (0.03), which is roughly 10 times better than the EC 50 Binding was increased 5-fold (rel.EC 50 PRO2557 rel.EC 50 The increased binding is likely due to the bivalent nature of the MATCH4 molecule, which results in an avidity effect. Among the MATCH4 molecules, PRO2591 (ROR1 domain: 55-39-G02-sc03) had the best EC 50 The binding of PRO2590 to the Fab fragment PRO2213 was approximately 2.5 times better than that of the reference Fab fragment PRO2213. The optimized variant of PRO2590, PRO2670, was measured using the same experimental setup by flow cytometry. In this experiment, we found that PRO2670 binds to the Fab fragment PRO2213 with a similar rel.EC 50 We found that PRO2590 bound to MDA-MB-231 cells with comparable potency to PRO2590 as indicated by the values.

[0349] In summary, all MATCH4 molecules demonstrated specific binding to human ROR1 expressing cells, and higher binding was found for molecules incorporating anti-ROR1 clone 55-39-G02 when compared to a series of molecules assembled with anti-ROR1 clone 55-38-D07. Possibly due to an avidity effect, all bivalent MATCH4 molecules showed increased EC 50 The binding was demonstrated.

[0350] [Table 33]

[0351] [Table 34]

[0352] [Table 35]

[0353] [Table 36]

[0354] 9.3. Target cell cytotoxicity and T cell activation mediated by MATCH4 molecules To understand the ability of MATCH4 molecules to promote tumor cell killing and CD8+ T cell activation, a series of cytotoxicity assays were performed to compare the potency and T cell activation profiles of our lead candidates.

[0355] First, the expression levels of ROR1 on several cell lines were tested in parallel. Briefly, cells were incubated with ZombieNIR Live / Dead-APC-Cy7 (Biolegend, 1:500) for 20 min at 4°C, followed by a washing step, then incubation with anti-ROR1-PE (Biolegend, 1:25) for 40 min at 4°C, followed by washing and acquisition on an Attune flow cytometer (Thermo Fisher Scientific). The measured surface ROR1 expression on MDA-MB-231, Jeko-1, JIMT-1, SKOV-3, and MCF7 cells is shown in Figure 11. The highest ROR1 levels were found on MDA-MB-231 and Jeko-1 cells. JIMT-1 and SKOV-3 cells showed low expression levels of ROR1, and MCF7 cells showed very low ROR1 expression levels. Therefore, MCF7 cells were used as a negative control for ROR1.

[0356] Cytotoxicity assays were performed using pan-T cells isolated from peripheral blood mononuclear cells (PBMCs) isolated from buffy coats. PBMCs were isolated using SepMate tubes on Lymphoprep (StemCell Technologies) to create density gradients according to the manufacturer's instructions. After PBMC isolation, pan-T cells were isolated using a pan-T isolation kit from Miltenyi according to the manufacturer's instructions.

[0357] Isolated pan-T cells were co-cultured with ROR1 positive and negative tumor cells of interest at an effector-to-target ratio of 10:1 for 40 hours. After co-culture, 100 μl of supernatant was aspirated and tested for lactate dehydrogenase (LDH) released from dying cells using a Roche cytotoxicity detection kit. The remaining supernatant was frozen at -80°C for downstream multiplex cytokine analysis. The remaining cells were labeled with the following reagents or antibodies (all from Biolegend): Live / Dead fixable Aqua, anti-human CD4-APC-Cy7 (clone OKT4), anti-human CD11c-PE-Cy7 (clone Bu15), anti-human CD8-PerCP-Cy5.5 (clone SK1), and anti-human CD69-PE (clone FN50). Cells were acquired by flow cytometer (Attune, Thermo Fisher Scientific) and analyzed for upregulation of CD69 as a measure of both CD4 and CD8 T cell activation using FlowJo software (BD).

[0358] All MATCH4 molecules tested induced specific killing of highly ROR1 positive cells such as MDA-MB-231 and Jeko-1 (see FIG. 12A), with no to minimal activity observed against the ROR1 negative cell MCF7 (see FIG. 12B). 50 The EC values ​​were better than the corresponding scMATCH3 (PRO2589 and PRO2590 vs. PRO2507, PRO2591 and PRO2592 vs. PRO2557 / PRO2510). However, in low ROR1 expressing cells, i.e., JIMT-1 and SKOV-3 (Fig. 12A, bottom), PRO2589 and PRO2590 (ROR1 domain derived from PRO2060) showed slightly better EC values ​​compared to PRO2591 and PRO2592 (ROR1 domain PRO2271). 50 The PRO2060-derived MATCH4 molecule has a slightly better EC compared with the scMATCH3 molecule PRO2507 on low ROR1 expressing cells. 50The values ​​also appear to show a difference between the cytotoxicity and cytotoxicity of the scMATCH3 molecules derived from PRO2271, however this difference does not appear to be present when comparing the MATCH4 and scMATCH3 molecules derived from PRO2271. Interestingly, on cells expressing low levels of ROR1, the MATCH4 molecule appears to show an overall lower maximal killing compared to the scMATCH3 molecule (FIG. 12A, bottom). The cytotoxicity results are summarized in Table 37.

[0359] Furthermore, T cell activation, as measured by an increase in the frequency of cells expressing CD69, was examined. CD8-based (Figure 13, top) and CD4-based (Figure 13, bottom) T cell activation correlated very closely with the cytotoxicity data. The highest frequencies of activated CD8 and CD4 T cells were obtained with MDA-MB-231 cells, which also express the highest levels of ROR1 (Figure 13A). PRO2060-based MATCH4 molecules showed a lower EC 50 However, these T cell populations had an overall higher frequency of activated cells when examining scMATCH3 molecules compared to MATCH4 molecules. The results are summarized in Tables 38 and 39.

[0360] The greatest difference in activity of MATCH4 molecules appears to be with low ROR1 expressing cells, namely SKOV-3 and JIMT-1 cells. Due to the expected heterogeneity of ROR1 expression levels between patients and between tumor indications, it would be advantageous to effectively target tumors with lower ROR1 expression levels. Based on these data, PRO2592 appears to have the lowest potency (cytotoxicity and T cell activation) and the lowest maximum.

[0361] 9.4. Biophysical Characterization of Lead MATCH4 Molecules Storage stability and melting temperature MATCH4 molecules were subjected to a 14-day stability study, in which the molecules were formulated at 1 mg / ml in aqueous buffer (50 mM phosphate-citrate buffer with 300 mM sucrose at pH 6.5) and stored at <-80°C, 4°C, and 40°C over 14 days. Monomeric and oligomeric fractions in the formulations were assessed by integration of SE-HPLC peak areas at different time points over the course of the study. Table 40 summarizes the % monomer content and % monomer loss relative to day 0. Changes in protein concentration were monitored over the course of the study by UV-Vis measurements at 280 nm and are shown in Table 41. Thermal stability was analyzed by nDSF (NanoTemper) and the onset of unfolding (T onset ) and the unfolding midpoint (T m ) was determined. m The results are shown in Table 40.

[0362] All four MATCH4 molecules displayed an excellent stability profile, showing little loss of monomer or protein content after 14 days of incubation. There was no notable change in monomer content at temperatures of -80°C and 4°C, as well as by repeated freeze-thawing (5x) performed prior to SE-HPLC / UV measurement on the 14 day / -80°C samples.

[0363] Example 10: Existing ADA binding assay 10.1. General Assay Procedure: A method was developed at Numab to detect pre-existing anti-drug antibodies in human serum using a direct assay format.

[0364] 96-well half-area plates were coated with 100ng / ml of test molecules (MATCH3 or scFv format) for 2 hours at room temperature. The plates were blocked with PBS containing 0.2% Tween and 1% BSA for 1 hour. Then, individual human sera were added to the corresponding wells either unspiked (screening assay) or spiked (confirmation assay) with the same molecules coated at a dilution of 1:20 (5% serum) or 1:100 (1% serum). Spiking concentrations ranged from 60 to 115nM, and the spiked samples were preincubated for 1 hour. Then, antibodies bound to the coated molecules on the plate were detected by 100ng / ml rabbit anti-human IgG-HRP for 1 hour. TMB substrate was added as a substrate, and after a short incubation, the enzymatic reaction was stopped by 1M HCl. The optical density of each well was read at 450nm. All steps were performed at room temperature. Between each step, the plate was washed three times with 450 μl wash buffer. With the exception of the blocking and washing steps, all assay components were added in a volume of 25 μl / well and used in duplicate. For the incubation steps, the ELISA plate was placed on a rotating mixer (40 rpm).

[0365] Typically, the first round of measurements was performed with unspiked human serum (screening assay). Then, a screening cut point (SCP) was calculated for each plate. Unspiked samples with a signal below the SCP were called "screening negative" and were not considered in the confirmatory assay. Unspiked samples with a signal above the SCP were called "screening positive".

[0366] [Table 37]

[0367] [Table 38]

[0368] [Table 39]

[0369] [Table 40]

[0370] [Table 41]

[0371] For most of these "screening positive" samples, a second round of measurements was performed with spiked human serum (confirmatory assay) to determine whether the initially detected binding of antibodies in the respective serum samples was specific for the test molecule. A reduction in the adsorption signal in the spiked wells indicates that the signal observed in the unspiked wells in the initial screening assay is specific for the coated molecule on the plate. The resulting percent inhibition (% inhibition) required to confirm specificity was set at 20 to 30% or the confirmation cut point was calculated in percentage (% CCP). In the latter case, only samples with percent inhibition exceeding the CCP were confirmed as positive. The % inhibition was calculated as the reduction of the initial signal obtained with unspiked serum (screening assay) as follows: % CCP = initial signal (1 - (spiked serum / unspiked serum signal)) * 100

[0372] In an alternative procedure, no initial screening assay was performed. Instead, test samples were analyzed directly using the confirmatory assay procedure. However, data analysis included the same calculations, including at least the calculation of SCP and %CCP.

[0373] 10.2. Determination / Calculation of Screening Cut Points (SCP), Normalization Factors (NF), and Floating Cut Points (FCP) For each test compound, 40 individual human serum samples from healthy untreated subjects were analyzed. In other cases, 20 individual human serum samples from healthy untreated subjects were analyzed.

[0374] Screening Cut Point (SCP): The screening cut point (SCP) is the threshold at which a signal is considered positive (screening positive). It is calculated to include 5% false positive sera.

[0375] The screening cut point (SCP) is calculated as follows: SCP=Average N+1.645×SDN During the ceremony, - "Average N" corresponds to the average signal from all unspiked individual sera measured for a particular test compound; - "SDN" corresponds to the standard deviation calculated from all unspiked individual sera measured for a particular test compound.

[0376] Normalization factor (NF): The normalization factor (NF) is calculated as follows: NF = SCP-negative control mean During the ceremony, - "SCP" is as defined above, - "Negative control mean" corresponds to the average signal from negative controls (pooled individual human sera, identical on each plate) analyzed in duplicate (i.e. two wells) per plate.

[0377] At least two NC samples (i.e., four wells) must be considered for the calculation of the normalization factor.

[0378] Floating Cut Point (FCP) Calculation After the determination of SCP and NF, the floating cut point (FCP) of each plate was used as the reference cut point. The FCP takes into account the analytical variability of each analytical unit by normalizing the SCP by the negative control of the plate. The FCP is calculated for each analytical unit as follows: FCP=NF+Average NC During the ceremony, - "Average NC" refers to the average signal of the negative control samples, "NF" refers to the normalization factor defined above.

[0379] 10.3. Improved Multispecific Antibody Variants in Existing ADA Binding Assay Results The variants PRO2668 and PRO2669, as well as the corresponding unmodified references PRO2510 and PRO2589 for comparison, were measured for their immunogenic properties using the existing ADA binding assay described above. The measurements were performed directly with 20 human serum samples (19 in the case of PRO2589) in a confirmatory assay set-up.

[0380] The screening positive sera were then further analyzed by considering a %CCP of 30%. The number of positive serum samples for each molecule tested is summarized in Table 42. A graph of the absorption levels of pre-existing ADA in human serum and reduction in absorbance levels (ADA binding inhibition) of spiked human serum is shown in Figure 14 for PRO2741 and reference PRO2660.

[0381] [Table 42]

[0382] Example 11: Target cell cytotoxicity and T cell activation mediated by scMATCH3 and MATCH4 variants PRO2667, PRO2668, PRO2669, and PRO2670 method These assays were performed with human peripheral blood mononuclear cells (PBMCs) isolated from buffy coats using SepMate tubes on Lymphoprep (StemCell Technologies) to create density gradients according to the manufacturer's instructions. Pan T cells were isolated from PBMCs using a pan T cell isolation kit from Miltenyi Biotec according to the manufacturer's instructions.

[0383] Isolated pan-T were co-cultured with ROR1 positive and negative tumor cells of interest at either an effector-to-target ratio of 10:1 for adherent cells or 5:1 for suspension cells for 40 hours. After co-culture, 100 μl of supernatant was aspirated and tested for lactate dehydrogenase (LDH) release from dying cells using a Roche cytotoxicity detection kit. The remaining supernatant was frozen at -80°C for downstream multiplex cytokine analysis. Cells were labeled with the following reagents or antibodies (all from Biolegend) for flow cytometric analysis of specific T cell activation: Live / Dead fixable Aqua, anti-human CD4-APC-Cy7 (clone OKT4), anti-human CD11c-PE-Cy7 (clone Bu15), anti-human CD8-PerCP-Cy5.5 (clone SK1), and anti-human CD69-PE (clone FN50). Cells were acquired by flow cytometer (Attune, Thermo Fisher Scientific) and analyzed for upregulation of CD69 as a measure of both CD4 and CD8 T cell activation using FlowJo software (BD).

[0384] result The use of all tested scMATCH3 (PRO2667 and PRO2668) and MATCH4 (PRO2669 and PRO2670) molecules resulted in specific killing of ROR1-positive MDA-MB-231 and JIMT-1 cells, as shown in Figure 15A. The potency of MATCH4 molecules was superior to that of scMATCH3 molecules. However, the level of maximum lysis achieved against the lower ROR1-expressing JIMT-1 cells was higher when using scMATCH3 molecules. Similar results were obtained when examining CD8 T cell activation, as measured by the frequency of CD69-expressing CD8 T cells (Figure 15B).

[0385] Example 12: Potency of PRO2668 to induce specific T cell-mediated lysis of ROR1-positive solid and hematological cancer cell lines Cytotoxicity mediated by PRO2668 was evaluated with cell lines representing solid (Figures 16A and 16B) and hematological malignancies (Figure 16C). Cytotoxicity was observed using ROR1-positive cell lines as targets, whereas the ROR1-negative line HCC1954 was not lysed, indicating the ROR1-dependency of PRO2668 (Figure 16A). The potency of specific lysis generally improved with increasing ROR1 density (Figure 16B). Specific killing of mantle cell lymphoma lines Mino and Z-138 was also observed (Figure 16C). These data indicate that PRO2668 is capable of mediating specific killing against a broad range of ROR1-expressing tumor cell lines.

[0386] Example 13: Potency of PRO2668 in mediating ROR1-dependent CD8 T cell activation and proliferation method Isolated healthy pan-T cells were labeled with CellTrace Violet according to the manufacturer's instructions and co-cultured with ROR1-positive tumor cells at an effector-to-target ratio of 5:1 for 3–6 days. After 3 and / or 6 days of co-culture, cells were labeled with the following reagents or antibodies (all from Biolegend) for flow cytometric analysis of specific T cell activation: Live / Dead fixable Aqua, anti-human CD4-APC-Cy7 (clone OKT4), anti-human CD19-PE (clone HIB19), anti-human CD8-PerCP-Cy5.5 (clone SK1), anti-human CD3-AlexaFluor488 (clone OKT3), and anti-human CD25-BV650 (clone BC96). Cells were acquired by a flow cytometer (Attune, Thermo Fisher Scientific) and analyzed using FlowJo software (BD). Dividing cells had reduced levels of fluorescence compared to non-dividing cells (dye dilution). Upregulation of CD25 counts was used as a measure of CD8 T cell activation.

[0387] result CD8 T cell activation was observed by co-culture with ROR1-expressing breast cancer cell lines MDA-MB-231 and JIMT-1 (Figure 17A). The degree of activation appeared to correlate with ROR1 expression, as a higher frequency of activated T cells was observed with higher ROR1 expression. CD8 T cell proliferation was observed in response to both ROR1-expressing cell lines when examining the frequency of dividing cells (cells with reduced CellTrace violet fluorescence) or dividing cell counts per microliter. The degree of proliferation was higher when targets with higher ROR1 expression were used. CD8 T cell activation and proliferation was also observed in response to mantle cell lymphoma line Z-138 (Figure 17B). These data indicate that engagement of ROR1-expressing targets and T cells with PRO2668 leads to T cell activation and proliferation.

[0388] Example 14: Potency of PRO2668 to mediate specific lysis of non-mitotic low ROR1 expressing primary human CLL patient samples method These assays were performed using healthy human pan-T cells isolated from CLL patient samples (IQ Biosciences) and peripheral blood mononuclear cells (PBMCs). PBMCs were isolated from buffy coats using SepMate tubes on Lymphoprep (StemCell Technologies) to create density gradients according to the manufacturer's instructions. Pan-T cells were isolated from PBMCs using Miltenyi Biotec's pan-T cell isolation kit according to the manufacturer's instructions.

[0389] CLL patient samples were assessed for their ability to divide by examining 5-ethynyl-2'-deoxyuridine (EdU) incorporation using the Click-It Plus flow cytometry assay kit and following the manufacturer's instructions. Dividing cells were positively labeled compared to non-dividing cells. Receptor density was assessed using the protocol outlined above, for example as outlined in section 9.3. Cellular binding of PRO2668 was assessed by titrating increasing concentrations of PRO2668 in the presence of human serum albumin, followed by extensive washing and detection with a fluorescently labeled secondary reagent designed to detect the framework. Samples were acquired by flow cytometer and analyzed using FlowJo.

[0390] Isolated pan-T were co-cultured with CLL patient samples for 40 hours at a 5:1 effector target ratio. After co-culture, the remaining supernatant was frozen at -80°C for downstream multiplex cytokine analysis. Cells were labeled with the following reagents or antibodies (all from Biolegend) for flow cytometric analysis of cell death and specific T cell activation: Live / Dead fixable Aqua, anti-human CD4-APC-Cy7 (clone OKT4), anti-human CD3-AlexaFluor488 (clone OKT3), anti-human CD8-PerCP-Cy5.5 (clone SK1), and anti-human CD69-BV650 (clone FN50), anti-human CD19-BV421 (clone SJ25C1), Annexin V-APC, anti-human ROR1-PE (clone 2A2). Cells were acquired by flow cytometer (Attune, Thermo Fisher Scientific) and analyzed for upregulation of CD69 as a measure of both CD4 and CD8 T cell activation using FlowJo software (BD). Dying CLL cells were gated as CD19+Annexin V+L / D+. Cytokine release from frozen supernatants was assessed using the LEGENDplex™ Multi-Analyte Flow Assay Kit Human Basic Immune Response Panel 13plex according to the manufacturer's instructions.

[0391] result The lack of proliferation of CLL PBMCs in the periphery creates a challenge for drugs that require active proliferation for their activity (Hu et al., Blood Advances vol. 5 (2021) p. 3152-3162). The lack of proliferation of CLL patient samples from peripheral blood was confirmed using EdU incorporation (Figure 18A). In contrast to the actively dividing Jeko-1 cell line, CLL patient samples did not divide. These samples expressed low levels of ROR1 (8,000 receptors per cell, Figure 18B). Saturation of ROR1 receptors was observed in these cells (Figure 18C). Specific lysis (Figure 18D) and concomitant CD8 and CD4 activation was observed by co-culture with healthy allogeneic pan-T cells and CLL patient samples (Figure 18E). Cytokine release was modest for some cytokines (IFNg and TNFa) and not others, such as IL-6 (Figure 18F). These data indicate that PRO2668 is capable of binding and specifically lysing non-dividing CLL patient samples, resulting in T cell activation and modest levels of cytokine release.

[0392] Example 15: In vivo efficacy of PRO2668 and PRO2670 method Frozen PBMCs were thawed in a 37°C water bath, centrifuged, resuspended in PBS, and stored on ice for inoculation. JeKo-1 tumor cells were maintained in vitro in RPMI1640 medium supplemented with 20% fetal bovine serum at 37°C in an atmosphere of 5% CO2. Cells were harvested in the exponential growth phase and quantified using a cell counter prior to tumor inoculation.

[0393] Each treatment group consisted of 5 animals. Each mouse received 5 × 10 6 mice in 0.1 ml of PBS mixed with Matrigel (1:1) for tumor development. 6 1×10 JeKo-1 tumor cells were inoculated subcutaneously in the upper right flank area. 7 PBMCs were transplanted intraperitoneally 3 days after tumor inoculation. 3 Randomization was performed when

[0394] Dosing began after randomization (day 0) and was performed every 5 days thereafter. After tumor inoculation, animals were checked daily for morbidity and mortality. During routine monitoring, animals were checked for tumor growth and any effect of treatment on behavior, e.g., mobility, food and water consumption, weight gain / loss (body weight was measured twice a week after randomization), eye / hair tangles, and any other abnormalities. Mortality and observed clinical signs were recorded in detail for each individual animal.

[0395] Tumor volumes were measured twice weekly after randomization in two dimensions using calipers and volumes were expressed in mm using the formula: V = (L × W × W) / 2. 3 where V is the tumor volume, L is the tumor length (the longest tumor dimension), and W is the tumor width (the longest tumor dimension perpendicular to L). A tumor volume of ≥ 3,000 mm was defined as a loss of 20% of body weight compared to the body weight on the first day of treatment. 3 Individual mice with a mean tumor volume of ≥ 2,000 mm (individual mice were sacrificed) or a group 3 In that case (all mice of the same group were sacrificed), the animals were sacrificed.

[0396] result PRO2668 was tested using the Jeko-1 human xenograft model in an in vivo efficacy study. Jeko-1 cells were implanted subcutaneously and human PBMCs were administered 3 days after tumor engraftment. As can be illustrated from FIG. 19, dosing of PRO2668 at 1 mg / kg and 0.2 mg / kg resulted in significant tumor growth inhibition compared to the irrelevant protein control Palivizumab. No difference in efficacy was observed with PRO2668 at 0.04 mg / kg versus the control. Antitumor efficacy was observed with PRO2670 at the 1 mg / kg dose, but the lower doses did not demonstrate significantly different tumor volumes versus the control. While both molecules appear to be efficacious, PRO2668 also demonstrates efficacy at the lower dose of 0.2 mg / kg.

[0397] Example 16: Biophysical characterization of scMATCH3 and MATCH4 variants PRO2667, PRO2668, PRO2669, and PRO2670 Storage stability and melting temperature MATCH4 molecules were subjected to a 28-day stability study, in which the molecules were formulated at 1 mg / ml in aqueous buffer (50 mM phosphate-citrate buffer with 300 mM sucrose at pH 6.5) and stored at <-80°C, 4°C, and 40°C for 14 days. Monomeric and oligomeric fractions in the formulations were assessed by integration of SE-HPLC peak areas at different time points during the course of the study. Table 43 summarizes the % monomer content and % monomer loss relative to day 0. Changes in protein concentration were monitored over the course of the study by UV-Vis measurements at 280 nm and are shown in Table 44. Thermal stability was analyzed by nDSF (NanoTemper) and the onset of unfolding (T onset ) and the unfolding midpoint (T m ) was determined. m The results are shown in Table 43.

[0398] All four MATCH4 molecules displayed excellent stability profiles and did not show significant monomer or protein content loss after 28 days of incubation. There was no significant change in monomer content at temperatures of -80°C and 4°C, as well as repeated freeze-thawing (5x) performed before SE-HPLC / UV measurement on the 28th day / -80°C samples, and there was only a minor loss of monomer content after 28 days of storage at 40°C.

[0399] [Table 43]

[0400] [Table 44]

Claims

1. a) one or two binding domains that specifically bind to the extracellular domain of ROR1 (ROR1-BD); and b) one binding domain that specifically binds to CD3 (CD3-BD); Including, Where: - multispecific antibodies do not contain an immunoglobulin Fc region, said ROR1-BDs, independently of each other, (I) (i) the HCDR1 sequence of SEQ ID NO:1; HCDR2 sequence of SEQ ID NO:2, HCDR3 sequence of SEQ ID NO:3, LCDR1 sequence of SEQ ID NO:4, The LCDR2 sequence of SEQ ID NO:5, and a set of CDR sequences consisting of the LCDR3 sequence of SEQ ID NO: 6, and (ii) a VH sequence that is at least 90, 91, 92, 93, 94, 95, 96, 97, 98, or 99 percent identical to any one of the amino acid sequences selected from SEQ ID NOs: 7 and 10, and A VL sequence that is at least 90, 91, 92, 93, 94, 95, 96, 97, 98, or 99 percent identical to any one of the amino acid sequences selected from SEQ ID NOs: 9 and 12; and / or (II) (i) the HCDR1 sequence of SEQ ID NO: 13 or 14; HCDR2 sequence of SEQ ID NO: 15, HCDR3 sequence of SEQ ID NO: 16, LCDR1 sequence of SEQ ID NO: 17, The LCDR2 sequence of SEQ ID NO: 18, and a set of CDR sequences consisting of the LCDR3 sequence of SEQ ID NO: 19, and (ii) a VH sequence that is at least 90, 91, 92, 93, 94, 95, 96, 97, 98, or 99 percent identical to any one of the amino acid sequences selected from SEQ ID NOs: 20, 23, 26, and 28; and a VL sequence that is at least 90, 91, 92, 93, 94, 95, 96, 97, 98, or 99 percent identical to any one of the amino acid sequences selected from SEQ ID NOs: 22, 25, 27, and 29; Including, said CD3-BD comprising (i) the HCDR1 sequence of SEQ ID NO: 30; HCDR2 sequence of SEQ ID NO: 31, HCDR3 sequence of SEQ ID NO: 32, LCDR1 sequence of SEQ ID NO: 33, The LCDR2 sequence of SEQ ID NO: 34, and a set of CDR sequences consisting of the LCDR3 sequence of SEQ ID NO: 35, and (ii) a VH sequence that is at least 90, 91, 92, 93, 94, 95, 96, 97, 98, or 99 percent identical to the amino acid sequence of SEQ ID NO: 36, and a VL sequence that is at least 90, 91, 92, 93, 94, 95, 96, 97, 98, or 99 percent identical to the amino acid sequence of SEQ ID NO: 38; Including, Multispecific antibodies.

2. The multispecific antibody of claim 1, wherein in the case where said multispecific antibody comprises two ROR1-BDs, said two ROR1-BDs comprise the same set of CDRs, i.e., both ROR1-BDs comprise either the CDRs of SEQ ID NOs: 1 to 6 or the CDRs of SEQ ID NOs: 13 / 14 to 19.

3. The multispecific antibody of claim 1, wherein the antibody further comprises one binding domain that specifically binds to human serum albumin (hSA-BD).

4. The multispecific antibody of claim 1 , wherein the multispecific antibody does not comprise a CH1 and / or CL region.

5. The hSA-BD comprises a VH1a, VH1b, VH3, or VH4 domain framework sequence FR1 to FR4, or 4. The multispecific antibody of claim 3, wherein said hSA-BD comprises a VL domain comprising VL framework regions FR1, FR2, and FR3 selected from the Vκ subtype, and a VL framework FR4 that is a VλFR4.

6. The multispecific antibody of claim 1, wherein the multispecific antibody is trispecific.

7. The multispecific antibody described in Claim 6, wherein the multispecific antibody is trispecific and trivalent, or trispecific and tetravalent.

8. The multispecific antibody of claim 1, wherein the format of the multispecific antibody is selected from scDb-scFv, scMATCH3, MATCH3, and MATCH4.

9. The one or two ROR1-BDs a) the VH sequence of SEQ ID NO: 7 and the VL sequence of SEQ ID NO: 9, or b) the VH sequence of SEQ ID NO: 10 and the VL sequence of SEQ ID NO: 12, or c) the VH sequence of SEQ ID NO: 20 and the VL sequence of SEQ ID NO: 22, or d) the VH sequence of SEQ ID NO: 23 and the VL sequence of SEQ ID NO: 25, or e) the VH sequence of SEQ ID NO: 8 and the VL sequence of SEQ ID NO: 9, or f) the VH sequence of SEQ ID NO: 21 and the VL sequence of SEQ ID NO: 22, or g) the VH sequence of SEQ ID NO: 26 and the VL sequence of SEQ ID NO: 27, or h) the VH sequence of SEQ ID NO: 28 and the VL sequence of SEQ ID NO: 29 The multispecific antibody of claim 1 , comprising:

10. The one or two ROR1-BDs a) the VH sequence of SEQ ID NO: 7 and the VL sequence of SEQ ID NO: 9, or b) the VH sequence of SEQ ID NO: 10 and the VL sequence of SEQ ID NO: 12, or c) the VH sequence of SEQ ID NO: 23 and the VL sequence of SEQ ID NO: 25, or d) the VH sequence of SEQ ID NO: 8 and the VL sequence of SEQ ID NO: 9 The multispecific antibody of claim 1 , comprising:

11. The above CD3-BD a) a VH domain comprising the amino acid sequence of SEQ ID NO: 36, and b) a VL domain comprising the amino acid sequence of SEQ ID NO: 38 The multispecific antibody of claim 1 , comprising:

12. The above-mentioned hSA-BD (i) a VH domain comprising the amino acid sequence of SEQ ID NO: 47 and a VL domain comprising the amino acid sequence of SEQ ID NO: 48; or (ii) a VH domain comprising the amino acid sequence of SEQ ID NO: 45 and a VL domain comprising the amino acid sequence of SEQ ID NO: 46; or (iii) a VH domain comprising the amino acid sequence of SEQ ID NO: 55 and a VL domain comprising the amino acid sequence of SEQ ID NO: 56; or (iv) a VH domain comprising the amino acid sequence of SEQ ID NO: 57 and a VL domain comprising the amino acid sequence of SEQ ID NO: 58; or (v) a VH domain comprising the amino acid sequence of SEQ ID NO: 65 and a VL domain comprising the amino acid sequence of SEQ ID NO: 67; or (vi) a VH domain comprising the amino acid sequence of SEQ ID NO: 68 and a VL domain comprising the amino acid sequence of SEQ ID NO: 70; The multispecific antibody of claim 5 , comprising:

13. The hSA-BD (i) a VH domain comprising the amino acid sequence of SEQ ID NO: 65 and a VL domain comprising the amino acid sequence of SEQ ID NO: 67; or (ii) a VH domain comprising the amino acid sequence of SEQ ID NO: 68 and a VL domain comprising the amino acid sequence of SEQ ID NO: 70 The multispecific antibody of claim 5 , comprising:

14. The ROR1-BD described in claim 1.

15. ROR1-BD described in claim 9.

16. A nucleic acid or two nucleic acids encoding the multispecific antibody of claim 1.

17. A nucleic acid or two nucleic acids encoding the ROR1-BD of claim 14.

18. A vector or two vectors comprising the nucleic acid or two nucleic acids according to claim 16.

19. A vector or two vectors comprising the nucleic acid or two nucleic acids described in claim 17.

20. 20. One or more host cells comprising the vector or two vectors of claim 18.

21. One or more host cells comprising the vector or two vectors described in claim 19.

22. A pharmaceutical composition comprising the multispecific antibody of claim 1 and a pharma- ceutically acceptable carrier for use in the treatment of a ROR1-expressing cancer.

23. The pharmaceutical composition of claim 22 for use in treating a ROR1-expressing cancer selected from chronic lymphocytic leukemia / small lymphocytic lymphoma (CLL / SLL), acute myeloid leukemia (AML), acute lymphoblastic leukemia (ALL), mantle cell lymphoma (MCL), hairy cell leukemia, follicular lymphoma (FL), marginal zone lymphoma (MZL), non-Hodgkin's lymphoma (NHL), diffuse large B-cell lymphoma (DLBCL), Richter's syndrome (RS), lung cancer, pancreatic cancer, prostate cancer, colon cancer, bladder cancer, breast cancer, ovarian cancer, glioblastoma, testicular cancer, uterine cancer, adrenal cancer, melanoma, neuroblastoma, sarcoma, and renal cancer.

Images