Bispecific antibody constructs that bind to DLL3 and CD3

Abstract

The present invention relates to bispecific antibody constructs that bind to DLL3 and CD3 comprising a first binding domain that binds to human DLL3 on the surface of a target cell and a second binding domain that binds to human CD3 on the surface of a T cell. In addition, the invention provides a polynucleotide encoding the antibody construct, a vector comprising the polynucleotide and a host cell transformed or transfected with the polynucleotide or vector. Furthermore, the invention provides a method for producing the antibody construct of the invention, a medical use of the antibody construct and a kit comprising the antibody construct.

Description

[0001] This application is a divisional application of an invention patent application, the application date of the parent application is August 1, 2016, the application number is 202210020142.9, and the invention name is "Bispecific antibody construct binding to DLL3 and CD3".

[0002] The application with the application date of August 1, 2016, application number 202210020142.9, and invention name “Bispecific antibody construct binding to DLL3 and CD3” is a divisional application of the application with the application date of August 1, 2016, application number 201680040701.5, and invention name “Bispecific antibody construct binding to DLL3 and CD3”. Technical Field

[0003] The present invention relates to a bispecific antibody construct comprising a first binding domain of human DLL3 that binds to the surface of a target cell and a second binding domain of human CD3 that binds to the surface of a T cell. In addition, the present invention provides a polynucleotide encoding the antibody construct, a vector comprising the polynucleotide, and a host cell transformed or transfected with the polynucleotide or vector. In addition, the present invention provides a method for producing the antibody construct of the present invention, the medical use of the antibody construct, and a kit comprising the antibody construct. Background Art

[0004] Small cell lung cancer (SCLC) is an aggressive form of lung cancer with a poor prognosis and limited treatment options, accounting for approximately 15% of all newly diagnosed lung cancers and approximately 25,000 new cases in the United States each year, and 180,000 new cases worldwide. Over the past few decades, survival rates have remained low, with only 5% of SCLC patients surviving five years, largely due to the lack of new therapies that can combat this form of lung cancer. Most patients have extensive stage disease, while about one-third of patients have limited stage disease, which is defined by the presence of a tumor in only one side of the chest and is suitable for a single radiation field. These stages affect the available treatment options, with limited stage disease being treated with chemotherapy and radiation, and extensive stage disease being treated with chemotherapy alone. Disseminated, metastatic tumors with lymphoma-like features are the hallmarks of SCLC. The first known diagnosis of SCLC patients described it as a disease of the lymphatic system, which was not identified as lung cancer before 1926, and it emphasized some unique properties of SCLC tumors compared with other solid tumors.

[0005] Patients typically respond well to current first-line therapy, which includes etoposide and cisplatin, but invariably relapse rapidly, developing chemotherapy-resistant disease for which there are currently no available treatment options. The prognosis in the relapsed refractory setting is extremely poor, with rapid disease progression and a short median survival of less than six months. In addition, SCLC patients have a high rate of comorbidities, including hypertension, heart disease, diabetes, and paraneoplastic syndromes. Together with the typically higher age of SCLC patients, these comorbidities affect the ability of patients to tolerate harsh chemotherapy regimens, further limiting treatment options.

[0006] Bispecific antibody formats comprising a scFv that recognizes CD3 expressed on T cells and another scFv that recognizes a tumor-associated antigen have demonstrated promising clinical efficacy, with high response rates in hematological malignancies such as refractory B-ALL (Topp, MS et al. Blood, 2012. 120(26): p. 5185-5187), leading to the approval of Blincyto. Although the efficacy of T cell-engaging therapy has not been demonstrated in solid tumor indications, SCLC may represent a promising solid tumor indication for CD3 x tumor target bispecific antibody formats given the disseminated nature of the disease. Therefore, bispecific T cell engagers that direct T cells against specific tumor antigens present new opportunities as new therapeutic options in the treatment of SCLC.

[0007] DLL3 was currently identified as a SCLC-specific tumor antigen by comparing the prevalence of DLL3 mRNA in a panel of primary patient tumors and a large panel of normal tissues by next generation sequencing. DLL3 expression levels in SCLC tumors were modest but highly prevalent, with approximately 90% of the tumors analyzed showing evidence of DLL3 expression by RNA-seq. In contrast to SCLC tumors, normal tissues showed very low expression of DLL3 transcripts, with minor expression detected in the testis, optic nerve, and cerebellum. Comparison of SCLC cell lines to tumors by RNA-seq showed similar expression levels, while cell surface quantification of DLL3 expression on SCLC cell lines indicated expression levels below 5000 DLL3 / cell, with typical expression levels below 2000 DLL3 / cell. Expression of DLL3 protein was confirmed by IHC, with 86% of SCLC tumors showing positive staining for DLL3, with a homogeneous and membranous staining pattern. All other normal tissues stained negative for DLL3, except for very faint staining in the cerebellum.

[0008] DLL3 is an atypical Notch ligand that functions to inhibit Notch signaling in a cell-autonomous manner, cis-binding to Notch, thereby blocking cell-to-cell interactions and internalization of Notch in target cells (a hallmark of typical Notch signaling). The main role of DLL3 is somitogenesis during embryonic development. DLL3 knockout mice display segmental defects in axial skeleton, cranial brain, and neuronal development. Somitic patterning defects have also been found in people with certain germline DLL3 mutations, causing a condition known as spondylolisthesis.

[0009] DLL3 has been previously proposed in methods for the diagnosis and treatment of gliomas other than SCLC using antibody-drug conjugates (ADCs) (WO 2013 / 126746). The use of ADC-based approaches for DLL3 may have limitations given the low expression levels of the protein on the cell surface and the reduced performance of the ADC against the target in the presence of low expression. In addition, ADC molecules often display toxicity associated with the free warhead (possibly as a result of linker degradation), leading to maximum tolerated dose limitations and potential effects on efficacy unrelated to the target chosen for the antibody. Given the desired sensitivity of T cells to their targets, this is unlikely to be a problem for T cell-engaging bispecific molecules (engineered to simultaneously engage DLL3 and CD3) and have shown potent in vitro cytotoxicity on cell lines expressing hundreds of target proteins per cell. In addition, the generally smaller size of bispecific T cell-engaging antibody constructs relative to normal antibodies (full-length IgG) improves tissue penetration and enhances efficacy due to more efficient engagement of the DLL3 and CD3 targets, leading to improved synapse formation between T cells and target tumor cells.

[0010] SCLC remains a high unmet medical need, and new treatment options are needed to improve the outlook for this sizeable patient population. The bispecific antibody formats discussed above have been clinically validated, and thus antibody constructs targeting DLL3 and CD3 represent a promising new possibility for the treatment of SCLC and an opportunity to improve the survival of patients suffering from this indication. As there remains a need for additional available options for the treatment of tumors or cancer diseases associated with overexpression of DLL3, means and methods are provided herein to address this problem in the form of bispecific antibody constructs having one binding domain for DLL3 and a second binding domain for CD3 on T cells. Summary of the invention

[0011] Thus, in a first aspect, the present invention provides a bispecific antibody construct comprising a first binding domain that binds to human DLL3 on the surface of a target cell and a second binding domain that binds to human CD3 on the surface of a T cell, wherein the first binding domain binds to an epitope of DLL3 contained within the region depicted by SEQ ID NO: 260.

[0012] It must be noted that, as used herein, the singular forms "a", "an", and "the" include plural references unless the context clearly indicates otherwise. Thus, for example, reference to "an agent" includes one or more of such different agents, and reference to "the method" includes reference to equivalent steps and methods known to those of ordinary skill in the art that may be modified or substituted for the methods described herein.

[0013] Unless otherwise indicated, the term "at least" preceding a series of elements should be understood to refer to every element in the series. Those skilled in the art will recognize or be able to ascertain using only routine experimentation many equivalents to the specific embodiments of the invention described herein. The present invention is intended to encompass such equivalents.

[0014] Whenever used herein, the term "and / or" includes the meanings of "and", "or" and "all or any other combinations of the elements connected by the term".

[0015] As used herein, the term "about" or "approximately" means within ±20%, preferably within ±15%, more preferably within ±10% and most preferably within ±5% of a given value or range.

[0016] Throughout this specification and in the claims that follow, unless the context requires otherwise, the word "comprise" and variations such as "comprises" and "comprising" will be understood to imply the inclusion of a stated integer or step or group of integers or steps but not the exclusion of any other integer or step or group of integers or steps. When used herein, the term "comprises" may be replaced by the terms "containing" or "including", or, at times as used herein, by the term "having".

[0017] When used herein, "consisting of excludes any element, step, or ingredient not specified in the claimed elements. When used herein, "consisting essentially of does not exclude materials or steps that do not materially affect the basic and novel characteristics of the claim.

[0018] In various instances herein, any of the terms "comprising," "consisting essentially of," and "consisting of may be replaced with any of the other two terms.

[0019] The term "antibody construct" refers to a molecule in which the structure and / or function are based on the structure and / or function of an antibody (e.g., full-length or complete immunoglobulin molecule). Therefore, the antibody construct is capable of binding to its specific target or antigen. In addition, the antibody construct according to the present invention comprises the minimum structural requirement for allowing target binding of an antibody. This minimum requirement can be defined, for example, by the presence of at least three light chain CDRs (i.e., CDR1, CDR2, and CDR3 in the VL region) and / or three heavy chain CDRs (i.e., CDR1, CDR2, and CDR3 in the VH region), preferably all six CDRs. The antibody based on the construct according to the present invention includes, for example, a monoclonal antibody, a recombinant antibody, a chimeric antibody, a deimmunized antibody, a humanized antibody, and a human antibody.

[0020] Full-length or complete antibodies are within the definition of "antibody constructs" according to the present invention, and also include camelid antibodies and other immunoglobulin antibodies produced by biotechnology or protein engineering methods or processes. These full-length antibodies can be, for example, monoclonal antibodies, recombinant antibodies, chimeric antibodies, deimmunized antibodies, humanized antibodies and human antibodies. Fragments of full-length antibodies (such as VH, VHH, VL, (s) dAb, Fv, Fd, Fab, Fab', F(ab')2 or "rIgG" ("half antibodies")) are also within the definition of "antibody constructs". The antibody construct according to the present invention may also be a modified fragment of an antibody, also referred to as an antibody variant, such as scFv; di-scFv or bi(s)-scFv; scFv-Fc; scFv-zipper; scFab; Fab2; Fab3; diabody; single-chain diabody; tandem diabody; tandem di-scFv; tandem tri-scFv; "minibody", which is exemplified by the following structures: (VH-VL-CH3)2, (scFv-CH3)2, ((scFv)2-CH3+CH3), ((scFv)2-CH3) or (scFv-CH3-scFv)2; multifunctional antibodies, such as trifunctional antibodies or tetrafunctional antibodies; and single domain antibodies, such as nanobodies, or single variable domain antibodies containing only one variable domain (which may be VHH, VH or VL) that specifically binds to an antigen or epitope independently of other V regions or domains. More preferred forms of the antibody construct according to the present invention are cross bodies, maxi bodies, hetero-Fc constructs and single-Fc constructs. Examples of these forms will be described below.

[0021] A binding domain may generally include an antibody light chain variable region (VL) and an antibody heavy chain variable region (VH); however, it need not include both. For example, a Fd fragment has two VH regions and generally retains some antigen binding function of a complete antigen binding domain. Additional examples of antibody fragments, antibody variants, or binding domain formats include (1) a Fab fragment, which is a monovalent fragment having VL, VH, CL, and CH1 domains; (2) a F(ab')2 fragment, which is a bivalent fragment having two Fab fragments connected by a disulfide bridge at the hinge region; (3) a Fd fragment, which has two VH and CH1 domains; (4) a Fv fragment, which has the VL and VH domains of a single arm of an antibody; (5) a dAb fragment (Ward et al. (1989) Nature 341:544-546), which has a VH domain; (6) isolated complementarity determining regions (CDRs) and (7) single-chain Fv (scFv), the latter being preferred (e.g., derived from a scFv library). Examples of embodiments of antibody constructs according to the invention are e.g. described in WO 00 / 006605, WO 2005 / 040220, WO 2008 / 119567, WO 2010 / 037838, WO 2013 / 026837, WO 2013 / 026833, US 2014 / 0308285, US 2014 / 0302037, WO 2014 / 144722, WO 2014 / 151910 and WO 2015 / 048272.

[0022] In addition, the definition of the term "antibody construct" includes monovalent, bivalent and multivalent constructs, and therefore includes monospecific constructs that specifically bind to only one antigen structure and bispecific and multispecific constructs that specifically bind to more than one antigen structure (e.g., two, three or more) via different binding domains. In addition, the definition of the term "antibody construct" includes molecules consisting of only one polypeptide chain and molecules consisting of more than one polypeptide chain, which chains can be the same (homodimers, homotrimers or homooligomers) or different (heterodimers, heterotrimers or heterooligomers). Examples of antibodies and variants or derivatives thereof as determined above are described in Harlow and Lane, Antibodies a laboratory manual, CSHL Press (1988); and Using Antibodies: a laboratory manual, CSHL Press (1999); Kontermann and Dübel, Antibody Engineering, Springer, 2nd edition, 2010; and Little, Recombinant Antibodies for Immunotherapy, Cambridge University Press 2009.

[0023] The antibody constructs of the present invention are preferably "in vitro generated antibody constructs". This term refers to an antibody construct according to the above definition, wherein all or part of the variable region (e.g., at least one CDR) is produced in non-immune cell selection (e.g., in vitro phage display) protein chip or any other method that can test the ability of candidate sequences to bind to antigens. Therefore, the term preferably excludes sequences that are produced only by genome rearrangement in immune cells in animals. "Recombinant antibodies" are antibodies prepared by using recombinant DNA technology or genetic engineering.

[0024] As used herein, the term "monoclonal antibody" (mAb) or monoclonal antibody construct refers to an antibody obtained from a population of substantially homogeneous antibodies, i.e., the individual antibodies constituting the population are identical except for possible naturally occurring mutations and / or post-translational modifications (e.g., isomerization, amidation) that may occur in small amounts. Compared to conventional (polyclonal) antibody preparations that typically include different antibodies directed against different determinants (or epitopes), monoclonal antibodies are highly specific to a single antigenic site or determinant on an antigen. In addition to their specificity, monoclonal antibodies are advantageous in that they are synthesized by hybridoma culture and are therefore not contaminated by other immunoglobulins. The modifier "monoclonal" indicates that the antibody is characterized by being obtained from a substantially homogeneous population of antibodies, and should not be construed as requiring the antibody to be produced by any particular method.

[0025] For the preparation of monoclonal antibodies, any technology that provides antibodies produced by continuous cell line culture can be used. For example, the monoclonal antibody to be used can be prepared by the hybridoma method originally described by Koehler et al., Nature, 256:495 (1975), or can be prepared by recombinant DNA methods (see, for example, U.S. Patent No. 4,816,567). Examples of other technologies for preparing human monoclonal antibodies include trioma technology, human B cell hybridoma technology (Kozbor, Immunology Today 4 (1983), 72) and EBV hybridoma technology (Cole et al., Monoclonal Antibodies and Cancer Therapy, Alan R. Liss, Inc. (1985), 77-96).

[0026] Hybridomas can then be screened using standard methods such as enzyme-linked immunosorbent assay (ELISA) and surface plasmon resonance (BIACORE). TM ) analysis to identify one or more hybridomas that produce antibodies that specifically bind to a particular antigen. Any form of the relevant antigen can be used as an immunogen, such as a recombinant antigen, a naturally occurring form, any variant or fragment thereof, and antigenic peptides thereof. Surface plasmon resonance, as used in the BIAcore system, can be used to increase the efficacy of phage antibodies that bind to epitopes of target antigens, such as DLL3 or CD3ε (Schier, Human Antibodies Hybridomas 7 (1996), 97-105; Malmborg, J. Immunol. Methods 183 (1995), 7-13).

[0027] Another exemplary method for preparing monoclonal antibodies includes screening protein expression libraries, such as phage display or ribosome display libraries. Phage display is described, for example, in Ladner et al., U.S. Pat. No. 5,223,409; Smith (1985) Science 228:1315-1317, Clackson et al., Nature, 352:624-628 (1991) and Marks et al., J. Mol. Biol., 222:581-597 (1991).

[0028] In addition to using display libraries, non-human animals, such as rodents (such as mice, hamsters, rabbits or rats), can be immunized using relevant antigens. In one embodiment, the non-human animal includes at least a portion of a human immunoglobulin gene. For example, a mouse strain lacking mouse antibody production can be engineered with a large fragment of a human Ig (immunoglobulin) locus. Using hybridoma technology, antigen-specific monoclonal antibodies derived from genes with the desired specificity can be prepared and selected. See, for example, XENOMOUSE TM , Green et al. (1994) Nature Genetics 7: 13-21, US 2003-0070185, WO 96 / 34096 and WO 96 / 33735.

[0029] Monoclonal antibodies can also be obtained from non-human animals and then modified using recombinant DNA techniques known in the art, e.g., humanized, deimmunized, chimerized, etc. Examples of modified antibody constructs include humanized variants of non-human antibodies, "affinity matured" antibodies (see, e.g., Hawkins et al. J. Mol. Biol. 254, 889-896 (1992) and Lowman et al., Biochemistry 30, 10832-10837 (1991)), and antibody mutants with altered effector function (see, e.g., U.S. Pat. No. 5,648,260; Kontermann and Dübel (2010), loc. cit., and Little (2009), loc. cit.).

[0030] In immunology, affinity maturation is the process of making B cells produce antibodies with increased affinity for antigens during the immune response process. When repeatedly exposed to the same antigen, the host will produce antibodies with successively increasing affinity. Same as the natural prototype, in vitro affinity maturation is based on the principle of mutation and selection. In vitro affinity maturation has been successfully used to optimize antibodies, antibody constructs and antibody fragments. Random mutations in CDR are introduced using radiation, chemical mutagens or error-prone PCR. In addition, genetic diversity can be increased by chain shuffling. Two or three rounds of mutations and selection using display methods (such as phage display) usually produce antibody fragments with affinity in the low nanomolar range.

[0031] A preferred type of amino acid substitution variant of the antibody construct includes replacing one or more hypervariable region residues of a parent antibody (e.g., a humanized or human antibody). Typically, the resulting variant selected for further development should have improved biological properties relative to the parent antibody from which it was generated. A suitable way to generate such substitution variants involves affinity maturation using phage display. In short, several hypervariable region sites (e.g., 6-7 sites) are mutated to generate all possible amino acid substitutions at each site. The antibody variants thus generated are displayed in a monovalent manner from filamentous phage particles as a fusion with the gene III product of M13 packaged within each particle. The phage-displayed variants are then screened for their biological activity (e.g., binding affinity) as disclosed herein. To identify candidate hypervariable region sites for modification, alanine scanning mutagenesis can be performed to identify hypervariable region residues that significantly contribute to antigen binding. Alternatively or additionally, it may be beneficial to analyze the crystal structure of the antigen-antibody complex to identify contact points between the binding domain and, for example, human DLL3. Such contact residues and adjacent residues are candidates for substitution according to the techniques detailed herein. Following generation of such variants, the panel of variants is screened as described herein, and antibodies with superior properties in one or more relevant assays may be selected for further development.

[0032] The monoclonal antibodies and antibody constructs of the present invention specifically include "chimeric" antibodies (immunoglobulins) in which a portion of the heavy and / or light chain is identical or homologous to the corresponding sequence in an antibody derived from a particular species or belonging to a particular antibody class or subclass, while the remainder of the chain is identical or homologous to the corresponding sequence in an antibody derived from another species or belonging to another antibody class or subclass, and fragments of such antibodies (as long as they exhibit the desired biological activity) (U.S. Pat. No. 4,816,567; Natl. Acad. Sci. USA. 81: 6851-6855 (1984)). Chimeric antibodies of interest herein include "primitized" antibodies, which comprise variable domain antigen-binding sequences derived from non-human primates (e.g., Old World Monkey, ape, etc.) and human constant region sequences. A variety of methods for preparing chimeric antibodies have been described. See, e.g., Morrison et al., Proc. Natl. Acad. ScL USA 81:6851, 1985; Takeda et al., Nature 314:452, 1985; Cabilly et al., U.S. Pat. No. 4,816,567; Boss et al., U.S. Pat. No. 4,816,397; Tanaguchi et al., EP 0171496; EP 0173494; and GB 2177096.

[0033] Antibodies, antibody constructs, antibody fragments or antibody variants may also be modified by specific deletion of human T-cell epitopes (a process known as "deimmunization"), for example, by methods disclosed in WO 98 / 52976 or WO 00 / 34317. Briefly, the heavy and light chain variable domains of the antibodies may be analyzed for peptides that bind to class II MHC; these peptides represent potential T-cell epitopes (as defined in WO 98 / 52976 and WO 00 / 34317). To detect potential T-cell epitopes, a computer modeling method known as "peptide threading" may be applied, and in addition, motifs present in VH and VL sequences in databases of human class II MHC binding peptides may be searched, as described in WO 98 / 52976 and WO 00 / 34317. These motifs bind to any of the 18 major class II MHC DR allotypes and therefore constitute potential T-cell epitopes. The detected potential T cell epitopes can be eliminated by replacing a few amino acid residues in the variable domain or preferably by single amino substitution. Typically, conservative substitutions are performed. Typically, but not exclusively, amino acids common to positions in human germline antibody sequences can be used. Human germline sequences are disclosed in, for example, Tomlinson et al. (1992) J. Mol. Biol. 227: 776-798; Cook, GP et al. (1995) Immunol. Today Vol. 16 No. 5: 237-242; and Tomlinson et al. (1995) EMBO J. 14: 14: 4628-4638. The VBASE directory provides a comprehensive directory of human immunoglobulin variable region sequences (compiled by Tomlinson, LA. et al. MRC Centre for Protein Engineering, Cambridge, UK). These sequences can be used as a source of human sequences, for example for framework regions and CDRs. A common human framework region can also be used, such as described in U.S. Patent No. 6,300,064.

[0034] "Humanized" antibodies, antibody constructs, variants or fragments thereof (such as Fv, Fab, Fab', F(ab')2 or other antigen binding sequences of antibodies) are antibodies or immunoglobulins of mostly human sequences that contain minimal sequences derived from non-human immunoglobulins. To a great extent, humanized antibodies are human immunoglobulins (recipient antibodies) in which residues from the hypervariable regions (also CDRs) of the receptor are replaced with residues from the hypervariable regions of non-human (e.g., rodent) species (donor antibodies) (such as mice, rats, hamsters or rabbits with the desired specificity, affinity and capacity). In some cases, the Fv framework region (FR) residues of the human immunoglobulin are replaced with corresponding non-human residues. In addition, "humanized antibodies" as used herein may also include residues not found in the receptor antibody or the donor antibody. These modifications are performed to further improve and optimize antibody performance. Humanized antibodies may also include at least a portion of an immunoglobulin constant region (Fc), typically at least a portion of a constant region of a human immunoglobulin. For further details, see Jones et al., Nature, 321:522-525 (1986); Reichmann et al., Nature, 332:323-329 (1988); and Presta, Curr. Op. Struct. Biol., 2:593-596 (1992).

[0035] Humanized antibodies or fragments thereof can be produced by replacing the sequences of Fv variable domains that are not directly involved in antigen binding with equivalent sequences from human Fv variable domains. Exemplary methods for producing humanized antibodies or fragments thereof are provided by the following documents: Morrison (1985) Science 229: 1202-1207; Oi et al., (1986) BioTechniques 4: 214; and US 5,585,089, US 5,693,761, US 5,693,762, US 5,859,205 and US 6,407,213. These methods include separating, manipulating and expressing nucleic acid sequences encoding all or part of the immunoglobulin Fv variable domains from at least one of the heavy chain or light chain. Such nucleic acids can be obtained from hybridomas that produce antibodies against a predetermined target as described above, as well as from other sources. The recombinant DNA encoding the humanized antibody molecule can then be cloned into a suitable expression vector.

[0036] Humanized antibodies can also be produced using transgenic animals (such as mice) that express human heavy and light chain genes but cannot express endogenous mouse immunoglobulin heavy and light chain genes. Winter describes an exemplary CDR transplantation method (U.S. Patent No. 5,225,539) that can be used to prepare humanized antibodies as described herein. All CDRs of specific human antibodies can be replaced with at least a portion of non-human CDRs, or only some CDRs can be replaced with non-human CDRs. Only the CDRs required to replace the humanized antibody are bound to the desired number of predetermined antigens.

[0037] Humanized antibodies can be optimized by introducing conservative substitutions, consensus sequence substitutions, germline substitutions and / or back mutations. Such altered immunoglobulin molecules can be prepared by any of several techniques known in the art (e.g., Teng et al., Proc. Natl. Acad. Sci. USA, 80: 7308-7312, 1983; Kozbor et al., Immunology Today, 4: 7279, 1983; Olsson et al., Meth. Enzymol., 92: 3-16, 1982 and EP 239 400).

[0038] The terms "human antibody", "human antibody construct" and "human binding domain" include antibodies, antibody constructs and binding domains having antibody regions (such as variable and constant regions or domains) that substantially correspond to human germline immunoglobulin sequences known in the art, including, for example, those described by Kabat et al., (1991) (loc. cit.). The human antibodies, antibody constructs or binding domains of the present invention may include, for example, in the CDRs, especially in CDR3, amino acid residues that are not encoded by human germline immunoglobulin sequences (e.g., mutations introduced by random or site-specific mutagenesis in vitro or by somatic mutations in vivo). The human antibody, antibody construct or binding domain may have at least one, two, three, four, five or more positions substituted with amino acid residues that are not encoded by human germline immunoglobulin sequences. The definitions of human antibodies, antibody constructs and binding domains as used herein also encompass fully human antibodies, which include only non-artificial and / or genetically altered human sequences of antibodies, such as those obtainable by using techniques or systems such as Xenomouse.

[0039] In some embodiments, the antibody construct of the present invention is an "isolated" or "substantially pure" antibody construct. "Isolated" or "substantially pure" when used to describe the antibody construct disclosed herein means that the antibody construct has been identified, separated and / or recovered from the components of its production environment. Preferably, the antibody construct is not or substantially not associated with all other components from its production environment. The pollutant components of its production environment (such as the pollutant components produced by recombinant transfected cells) are substances that usually interfere with the diagnostic or therapeutic use of the polypeptide, and may include enzymes, hormones and other proteins or non-protein solutes. The antibody construct may, for example, account for at least about 5% by weight or at least about 50% by weight of the total protein in a given sample. It should be understood that, as the case may be, the isolated protein may account for 5% to 99.9% by weight of the total protein content. The polypeptide can be prepared at a significantly higher concentration by using an inducible promoter or a high expression promoter so that it is prepared at an increased concentration level. This definition includes preparing antibody constructs in a variety of organisms and / or host cells known in the art. In preferred embodiments, the antibody construct will be purified (1) to an extent sufficient to obtain at least 15 residues of N-terminal or internal amino acid sequence by use of a spinning cup sequenator, or (2) to homogeneity by SDS-PAGE under non-reducing or reducing conditions using Coomassie blue or, preferably, silver stain. Typically, however, the isolated antibody construct will be prepared by at least one purification step.

[0040] In connection with the present invention, the term "binding domain" characterizes: a domain that (specifically) binds to / interacts with / recognizes a given target epitope or a given target site on a target molecule (antigen) (here: DLL3 and CD3, respectively). The structure and function of the first binding domain (recognition of DLL3) and preferably also the structure and / or function of the second binding domain (recognition of CD3) are based on the structure and / or function of an antibody (e.g. a full-length or complete immunoglobulin molecule). According to the present invention, the first binding domain is characterized by the presence of three light chain CDRs (i.e. CDR1, CDR2 and CDR3 of the VL region) and / or three heavy chain CDRs (i.e. CDR1, CDR2 and CDR3 of the VH region). The second binding domain preferably also comprises the minimal structural requirements of an antibody that allow target binding. More preferably, the second binding domain comprises at least three light chain CDRs (i.e. CDR1, CDR2 and CDR3 of the VL region) and / or three heavy chain CDRs (i.e. CDR1, CDR2 and CDR3 of the VH region). It is envisaged that the first and / or second binding domain is produced or obtainable by phage display or library screening methods rather than by transplanting CDR sequences from pre-existing (monoclonal) antibodies into a scaffold.

[0041] According to the present invention, the binding domain is in the form of one or more polypeptides. Such polypeptides may include protein parts and non-protein parts (e.g., chemical linkers or chemical cross-linking agents, such as glutaraldehyde). Proteins (including fragments thereof, preferably biologically active fragments and peptides, generally having less than 30 amino acids) contain two or more amino acids (producing amino acid chains) coupled to each other via covalent peptide bonds. The term "polypeptide" as used herein describes a group of molecules, which is generally composed of more than 30 amino acids. Polypeptides can further form polymers, such as dimers, trimers, and higher oligomers, i.e., are composed of more than one polypeptide molecule. The polypeptide molecules forming such dimers, trimers, etc. can be the same or different. The corresponding higher-order structures of such polymers are therefore referred to as homodimers or heterodimers, homotrimers or heterotrimers, etc. An example of a heteropolymer is an antibody molecule, which is composed of two identical polypeptide light chains and two identical polypeptide heavy chains in its naturally occurring form. The terms "peptide", "polypeptide" and "protein" also refer to naturally modified peptides / polypeptides / proteins, wherein the modification is achieved, for example, by post-translational modifications such as glycosylation, acetylation, phosphorylation, etc. "Peptide", "polypeptide" or "protein" as referred to herein may also be chemically modified, such as pegylated. Such modifications are well known in the art and are described below.

[0042] Preferably, the binding domain that binds to DLL3 and / or the binding domain that binds to CD3 is a human binding domain. Antibodies and antibody constructs comprising at least one human binding domain avoid some of the problems associated with antibodies or antibody constructs having variable and / or constant regions from non-human, such as rodents (e.g., murine, rat, hamster, or rabbit). The presence of such rodent-derived proteins may cause rapid clearance of the antibody or antibody construct or may cause the patient to produce an immune response against the antibody or antibody construct. To avoid the use of antibodies or antibody constructs derived from rodents, human or fully human antibodies / antibody constructs may be produced by introducing human antibody functions into rodents so that the rodents produce fully human antibodies.

[0043] The ability to clone and reconstruct megabase-sized human loci in YACs and introduce them into the mouse germline provides a powerful method to elucidate the functional components of very large or roughly mapped loci and to generate useful models of human disease. In addition, the use of such techniques to replace mouse loci with human equivalents can provide unique insights into the expression and regulation of human gene products during development, their connections with other systems, and their involvement in disease induction and progression.

[0044] An important practical application of such strategy is the "humanization" of the mouse humoral immune system. The introduction of human immunoglobulin (Ig) loci into mice in which endogenous Ig genes have been inactivated provides the opportunity to study the potential mechanism of programmed expression and antibody assembly and its role in B cell development. In addition, such strategy can provide an ideal source for producing fully human monoclonal antibodies (mAbs), which is an important milestone towards the prospect of antibody therapy for realizing human diseases. It is expected that fully human antibodies or antibody constructs minimize the immunogenicity and allergic reactions inherent to mouse mAbs or mouse-derived mAbs, and therefore improve the efficacy and safety of the antibodies / antibody constructs administered. It is expected that the use of fully human antibodies or antibody constructs can provide substantial advantages in the treatment of chronic and recurrent human diseases (such as inflammation, autoimmunity and cancer) that require repeated administration of compounds.

[0045] One approach toward this goal is to engineer mouse antibody production-deficient mouse strains with large fragments of human Ig loci, and it is expected that such mice will produce large pedigrees of human antibodies in the absence of mouse antibodies. Large human Ig fragments will maintain large variable gene diversity and appropriate regulation of antibody production and expression. By utilizing mouse mechanisms to achieve antibody diversification and selection and the lack of immune tolerance to human proteins, the human antibody spectrum regenerated in these mouse strains will produce high-affinity antibodies for any antigen of interest (including human antigens). Using hybridoma technology, antigen-specific human mAbs with desired specificity can be easily produced and selected. This general strategy was demonstrated in conjunction with the production of the first XenoMouse mouse strain (see Green et al., Nature Genetics 7: 13-21 (1994)). XenoMouse strains were engineered with yeast artificial chromosomes (YACs) containing 245kb and 190kb size germline structural fragments of human heavy chain loci and κ light chain loci, respectively, and the chromosomes contained core variable and constant region sequences. YAC containing human Ig has been confirmed to be compatible with the mouse system for antibody rearrangement and expression, and can replace inactivated mouse Ig genes. This is demonstrated by its ability to induce B cell development, produce adult-like human lineages of fully human antibodies, and produce antigen-specific human mAbs. These results also show that the introduction of a larger portion of the human Ig locus containing a larger amount of V genes, additional regulatory elements, and human Ig constant regions can basically outline the complete lineage unique to the human humoral response to infection and immunization. The research of Green et al. has recently extended to the introduction of more than about 80% human antibody lineages by introducing megabase-sized germline YAC fragments of human heavy chain loci and kappa light chain loci, respectively. See Mendez et al. Nature Genetics 15:146-156 (1997) and U.S. Patent Application No. 08 / 759,620.

[0046] The production of XenoMouse mice is further discussed and described in U.S. Patent Application Nos. 07 / 466,008, 07 / 610,515, 07 / 919,297, 07 / 922,649, 08 / 031,801, 08 / 112,848, 08 / 234,145, 08 / 376,279, 08 / 430,938, 08 / 464,584, 08 / 464,582, , Serial No. 08 / 463,191, Serial No. 08 / 462,837, Serial No. 08 / 486,853, Serial No. 08 / 486,857, Serial No. 08 / 486,859, Serial No. 08 / 462,513, Serial No. 08 / 724,752 and Serial No. 08 / 759,620; and U.S. Patent Nos. 6,162,963, 6,150,584, 6,114,598, 6,075,181 and 5,939,598 and Japanese Patent Nos. 3 068180B2, 3 068506B2 and 3 068 507B2. See also Mendez et al. Nature Genetics 15: 146-156 (1997) and Green and Jakobovits J. Exp. Med. 188: 483-495 (1998), EP 0 463 151 Bl, WO 94 / 02602, WO 96 / 34096, WO 98 / 24893, WO 00 / 76310 and WO 03 / 47336.

[0047] In an alternative approach, others, including GenPharm International, Inc., have used the "minilocus" approach. In the minilocus approach, an exogenous Ig locus is mimicked by including fragments (individual genes) from the Ig locus. Thus, one or more VH genes, one or more DH genes, one or more JH genes, a μ constant region, and a second constant region (preferably a gamma constant region) form a construct for insertion into an animal. This method is described in U.S. Pat. No. 5,545,807 to Surani et al. and U.S. Pat. Nos. 5,545,806, 5,625,825, 5,625,126, 5,633,425, 5,661,016, 5,770,429, 5,789,650, 5,814,318, 5,877,397, 5,874,299 and 6,255,458 each to Lonberg and Kay, 5,591,669 and 6,023,010 to Krimpenfort and Berns, and 6,023,010 to Berns et al. U.S. Patent Nos. 5,612,205, 5,721,367, and 5,789,215, and U.S. Patent No. 5,643,763 to Choi and Dunn, and GenPharm International U.S. Patent Application Nos. 07 / 574,748, 07 / 575,962, 07 / 810,279, 07 / 853,408, 07 / 904,068, 07 / 990,860, 08 / 053,131, 08 / 096,762, 08 / 155,301, 08 / 161,739, 08 / 165,699, 08 / 209,741. See also EP 0546073B1, WO 92 / 03918, WO 92 / 22645, WO 92 / 22647, WO 92 / 22670, WO 93 / 12227, WO 94 / 00569, WO 94 / 25585, WO 96 / 14436, WO 97 / 13852 and WO 98 / 24884 and U.S. Pat. No. 5,981,175. See also Taylor et al. (1992), Chen et al. (1993), Tuaillon et al. (1993), Choi et al. (1993), Lonberg et al. (1994), Taylor et al. (1994) and Tuaillon et al. (1995), Fishwild et al. (1996).

[0048] Kirin has also demonstrated the production of human antibodies from mice into which large fragments of chromosomes or complete chromosomes have been introduced by microcell fusion. See European Patent Application Nos. 773 288 and 843 961. Xenerex Biosciences is developing a technology that may produce human antibodies. In this technology, SCID mice are reconstituted with human lymphocytes (e.g., B cells and / or T cells). The mice are then immunized with an antigen and the mice can produce an immune response to the antigen. See U.S. Patents Nos. 5,476,996, 5,698,767, and 5,958,765.

[0049] Human anti-mouse antibody (HAMA) responses have led the industry to prepare chimeric or humanized antibodies. However, it is expected that certain human anti-chimeric antibody (HACA) responses will be observed, especially in chronic or multiple dose use of antibodies. Therefore, it would be desirable to provide antibody constructs comprising a human binding domain to DLL3 and / or a human binding domain to CD3 in order to eliminate the problem and / or effects of HAMA or HACA responses.

[0050] According to the present invention, the terms "(specifically) bind to", "(specifically) recognize", "(specifically) target" and "(specifically) react with" mean that the binding domain interacts or specifically interacts with a given epitope or a given target site on a target molecule (antigen) (here: DLL3 and CD3, respectively).

[0051] The term "epitope" refers to a site on an antigen to which a binding domain, an antibody or immunoglobulin, or a derivative, fragment or variant of an antibody or immunoglobulin specifically binds. An "epitope" is antigenic, and for this reason the term epitope is sometimes also referred to herein as an "antigenic structure" or "antigenic determinant". Thus, a binding domain is an "antigenic interaction site". Such binding / interaction is also understood to define "specific recognition".

[0052] An "epitope" can be formed by contiguous or non-contiguous amino acids juxtaposed by the tertiary folding of a protein. A "linear epitope" is an epitope whose primary sequence of amino acids comprises the identified epitope. A linear epitope typically comprises at least 3 or at least 4, and more typically at least 5 or at least 6 or at least 7, e.g., about 8 to about 10 amino acids in a unique sequence.

[0053] In contrast to linear epitopes, a "conformational epitope" is an epitope in which the primary sequence of amino acids constituting the epitope is not the sole defining component of the epitope recognized (e.g., an epitope in which the primary sequence of amino acids is not necessarily recognized by a binding domain). Typically, a conformational epitope comprises an increased number of amino acids relative to a linear epitope. With respect to recognition of a conformational epitope, the binding domain recognizes the three-dimensional structure of an antigen, preferably a peptide or protein, or a fragment thereof (in the context of the present invention, the antigen structure for the first binding domain is contained within the DLL3 protein). For example, when a protein molecule folds to form a three-dimensional structure, certain amino acids and / or the polypeptide backbone that form the conformational epitope become juxtaposed, enabling the antibody to recognize the epitope. Methods for determining the conformation of an epitope include, but are not limited to, x-ray crystallography, two-dimensional nuclear magnetic resonance (2D-NMR) spectroscopy, and site-directed spin labeling and electron paramagnetic resonance (EPR) spectroscopy.

[0054] An epitope mapping method is described below: When a region (contiguous stretch of amino acids) in a human DLL3 protein is exchanged / replaced with its corresponding region of a non-human and non-primate DLL3 antigen (e.g. mouse DLL3, but other species such as chicken, rat, hamster, rabbit, etc. are also contemplated), a reduction in binding of the binding domain is expected to occur unless the binding domain is cross-reactive to the non-human, non-primate DLL3 used. The reduction is preferably at least 10%, 20%, 30%, 40% or 50% compared to binding to the corresponding region in the human DLL3 protein; more preferably at least 60%, 70% or 80%, and most preferably 90%, 95% or even 100%, wherein binding to the corresponding region in the human DLL3 protein is set to 100%. It is contemplated that the aforementioned human DLL3 / non-human DLL3 chimera is expressed in CHO cells. It is also contemplated that the human DLL3 / non-human DLL3 chimera is fused to the transmembrane domain and / or cytoplasmic domain of a different membrane-bound protein (such as EpCAM).

[0055] In an alternative or additional approach to epitope mapping, several truncated forms of the human DLL3 extracellular domain can be generated in order to determine the specific region recognized by the binding domain. In these truncated forms, starting from the N-terminus, different extracellular DLL3 domains / subdomains or regions are progressively deleted. The truncated DLL3 forms generated and used in the context of the present invention are depicted in Figure 1It is envisaged that truncated DLL3 forms are expressed in CHO cells. It is also envisaged that truncated DLL3 forms are fused to transmembrane domains and / or cytoplasmic domains of different membrane-bound proteins (such as EpCAM). It is also envisaged that truncated DLL3 forms encompass a signal peptide domain at its N-terminus, such as a signal peptide derived from a mouse IgG heavy chain signal peptide. It is further envisaged that truncated DLL3 forms encompass a v5 domain at its N-terminus (after the signal peptide), which allows verification of its correct expression on the cell surface. It is expected that these truncated DLL3 forms will experience a reduction or loss of binding, which forms no longer encompass the DLL3 region recognized by the binding domain. The reduction in binding is preferably at least 10%, 20%, 30%, 40%, 50%; more preferably at least 60%, 70%, 80% and most preferably 90%, 95% or even 100%, wherein binding to the complete human DLL3 protein (or its extracellular region or domain) is set to 100%. Methods for testing this loss of binding are described in Example 2.

[0056] Another method to determine the contribution of specific residues of the target antigen to the recognition of an antibody construct or binding domain is alanine scanning (see, e.g., Morrison KL and Weiss GA. Cur Opin Chem Biol. 2001 Jun; 5(3):302-7), in which each residue to be analyzed is replaced with alanine, e.g., via site-directed mutagenesis. Alanine is used because of its non-bulky, chemically inert methyl functional group, yet still mimics the secondary structural references that many other amino acids have. Sometimes, bulky amino acids such as valine or leucine can be used in situations where the size of the mutated residue needs to be maintained. Alanine scanning is a mature technique that has been used for a long time.

[0057] The interaction between a binding domain and an epitope or a region containing an epitope implies that the binding domain exhibits a measurable affinity for the epitope / region containing the epitope on a particular protein or antigen (here: DLL3 and CD3, respectively), and generally does not exhibit significant reactivity with proteins or antigens other than DLL3 or CD3. "Measurable affinity" includes binding with a binding affinity of about 10 -6 Preferably, when the binding affinity is about 10 -12 Up to 10 -8 M, 10 -12 Up to 10 -9 M, 10 -12 Up to 10 -10 M, 10 -11 Up to 10 -8 M, preferably about 10 -11 Up to 10 -9M, the binding is considered specific. Whether a binding domain specifically reacts with a target or specifically binds to a target can be particularly easily tested by comparing the reaction of the binding domain with the target protein or antigen with the reaction of the binding domain with a protein or antigen other than DLL3 or CD3. Preferably, the binding domain of the invention does not essentially or substantially bind to a protein or antigen other than DLL3 or CD3 (i.e., the first binding domain is not capable of binding to a protein other than DLL3 and the second binding domain is not capable of binding to a protein other than CD3).

[0058] The term "essentially / substantially does not bind" or "is not capable of binding" means that the binding domain of the present invention does not bind to proteins or antigens other than DLL3 or CD3, i.e., exhibits no more than 30%, preferably no more than 20%, more preferably no more than 10%, and particularly preferably no more than 9%, 8%, 7%, 6% or 5% reactivity with proteins or antigens other than DLL3 or CD3, wherein binding to DLL3 or CD3 is set to 100%, respectively.

[0059] It is also envisaged that the antibody constructs of the invention bind to human DLL3 isoforms having one or both of the following DLL3 point mutations: F172C and L218P. See Example 5.

[0060] It is believed that specific binding is achieved by the specific motifs in the amino acid sequence of the binding domain and the antigen. Therefore, the binding is achieved because of its primary, secondary and / or tertiary structure and the secondary modification of the structure. The specific interaction of the antigen interaction site with its specific antigen can result in a simple combination of the site with the antigen. In addition, the specific interaction of the antigen interaction site with its specific antigen can alternatively or additionally result in a triggering signal, for example due to inducing antigen conformational changes, antigen oligomerization, etc.

[0061] The antibody constructs according to the invention bind to an epitope of DLL3 which is contained in the region depicted as SEQ ID NO: 260, corresponding to an amino acid stretch encompassing the region EGF-3 and EGF-4. Additional groups of anti-DLL3 binders were also generated and their DLL3 binding specificity was identified during epitope mapping (see Example 2).

[0062] The largest group of generated binders recognized epitopes within the DSL domain. However, none of these antibody constructs met the initial 18-hour 51 Criteria for adequate cytotoxic activity in a cytotoxic assay of Cr using stimulated human CD8+ T cells as effector cells and huDLL3-transfected CHO cells as target cells.

[0063] In another initial 48 hour FACS-based cytotoxicity assay (using unstimulated human PBMCs as effector cells and huDLL3-transfected CHO cells as target cells), these binders recognizing a DLL3 epitope within the N-terminus of the protein had EC50 values ​​between 1455 and 1580 pM. In the context of the present invention, these values ​​are not considered sufficient for bispecific antibodies provided for therapeutic use in directing the patient's immune system, more specifically T cell cytotoxic activity, against cancer cells.

[0064] Finally, another group of binders was generated and characterized for their cytotoxic activity in multiple assays. Epitope mapping of these binders revealed specificity for DLL3 epitopes contained within the EGF-5 region and to some extent also within the EGF-6 region (see Example 2 for details). The overall picture of the data generated in the different cytotoxicity assays (see Examples 8.3, 8.5, 8.6 and 8.7) for these binders named DLL3-18, DLL3-19, DLL3-20 and DLL3-21 reveals the following: Although not all binders failed in all assays with favorable EC50 values, the entire group clearly underperformed if compared to the antibody constructs according to the invention. This observation is highlighted by the darker shading of the results shown in Tables 6-9.

[0065] In summary, it can be clearly stated that the antibody construct according to the present invention, which binds to an epitope of DLL3 contained in the region as depicted by SEQ ID NO: 260, shows the best activity performance so far compared to various other groups of DLL3 binders with different epitope specificities. In other words, the antibody construct according to the present invention presents a favorable epitope-activity relationship, thus supporting a potent bispecific antibody construct-mediated cytotoxic activity.

[0066] In another aspect, the present invention provides a bispecific antibody construct comprising a first binding domain that binds to human DLL3 on the surface of a target cell and a second binding domain that binds to human CD3 on the surface of a T cell, wherein the first binding domain binds to an epitope of DLL3 contained within the region depicted by SEQ ID NO:258.

[0067] Preferably, the first binding domain of the bispecific antibody construct of the present invention comprises a VH region comprising CDR-H1, CDR-H2 and CDR-H3 and a VL region comprising CDR-L1, CDR-L2 and CDR-L3, which are selected from:

[0068] a) CDR-H1 as depicted in SEQ ID NO:31, CDR-H2 as depicted in SEQ ID NO:32, CDR-H3 as depicted in SEQ ID NO:33, CDR-L1 as depicted in SEQ ID NO:34, CDR-L2 as depicted in SEQ ID NO:35, and CDR-L3 as depicted in SEQ ID NO:36;

[0069] b) CDR-H1 as depicted in SEQ ID NO:41, CDR-H2 as depicted in SEQ ID NO:42, CDR-H3 as depicted in SEQ ID NO:43, CDR-L1 as depicted in SEQ ID NO:44, CDR-L2 as depicted in SEQ ID NO:45, and CDR-L3 as depicted in SEQ ID NO:46;

[0070] c) CDR-H1 as depicted in SEQ ID NO:51, CDR-H2 as depicted in SEQ ID NO:52, CDR-H3 as depicted in SEQ ID NO:53, CDR-L1 as depicted in SEQ ID NO:54, CDR-L2 as depicted in SEQ ID NO:55, and CDR-L3 as depicted in SEQ ID NO:56;

[0071] d) CDR-H1 as depicted in SEQ ID NO:61, CDR-H2 as depicted in SEQ ID NO:62, CDR-H3 as depicted in SEQ ID NO:63, CDR-L1 as depicted in SEQ ID NO:64, CDR-L2 as depicted in SEQ ID NO:65, and CDR-L3 as depicted in SEQ ID NO:66;

[0072] e) CDR-H1 as depicted in SEQ ID NO:71, CDR-H2 as depicted in SEQ ID NO:72, CDR-H3 as depicted in SEQ ID NO:73, CDR-L1 as depicted in SEQ ID NO:74, CDR-L2 as depicted in SEQ ID NO:75, and CDR-L3 as depicted in SEQ ID NO:76;

[0073] f) CDR-H1 as depicted in SEQ ID NO:81, CDR-H2 as depicted in SEQ ID NO:82, CDR-H3 as depicted in SEQ ID NO:83, CDR-L1 as depicted in SEQ ID NO:84, CDR-L2 as depicted in SEQ ID NO:85, and CDR-L3 as depicted in SEQ ID NO:86;

[0074] g) CDR-H1 as depicted in SEQ ID NO:91, CDR-H2 as depicted in SEQ ID NO:92, CDR-H3 as depicted in SEQ ID NO:93, CDR-L1 as depicted in SEQ ID NO:94, CDR-L2 as depicted in SEQ ID NO:95, and CDR-L3 as depicted in SEQ ID NO:96;

[0075] h) CDR-H1 as depicted in SEQ ID NO: 101, CDR-H2 as depicted in SEQ ID NO: 102, CDR-H3 as depicted in SEQ ID NO: 103, CDR-L1 as depicted in SEQ ID NO: 104, CDR-L2 as depicted in SEQ ID NO: 105, and CDR-L3 as depicted in SEQ ID NO: 106; and

[0076] i) CDR-H1 as depicted in SEQ ID NO: 111, CDR-H2 as depicted in SEQ ID NO: 112, CDR-H3 as depicted in SEQ ID NO: 113, CDR-L1 as depicted in SEQ ID NO: 114, CDR-L2 as depicted in SEQ ID NO: 115 and CDR-L3 as depicted in SEQ ID NO: 116.

[0077] The term "variable" refers to the portion of an antibody or immunoglobulin domain that exhibits variability in its sequence and is involved in determining the specificity and binding affinity of a particular antibody (i.e., the "variable domain"). The variable heavy chain (VH) pairs with the variable light chain (VL) to form a single antigen binding site.

[0078] Variability is not evenly distributed throughout the variable domain of an antibody; it is concentrated in subdomains of each heavy and light chain variable region. These subdomains are called "hypervariable regions" or "complementarity determining regions" (CDRs). The more conservative (i.e., non-hypervariable) portions of the variable domains are called "framework" regions (FRMs or FRs), and provide a scaffold for the six CDRs in three-dimensional space to form an antigen binding surface. The variable domains of naturally occurring heavy and light chains each contain four FRM regions (FR1, FR2, FR3, and FR4), which primarily adopt a β-folded configuration, connected by three hypervariable regions, thereby forming a loop that connects the β-folded structure and in some cases forms a part of the β-folded structure. The hypervariable regions in each chain are tightly bound together with the hypervariable regions of the other chain through the FRMs, contributing to the formation of the antigen binding site (see Kabat et al., cited above).

[0079] The term "CDR" and its plural form "CDRs" refer to the complementarity determining regions, three of which constitute the binding characteristics of the light chain variable region (CDR-L1, CDR-L2 and CDR-L3) and three of which constitute the binding characteristics of the heavy chain variable region (CDR-H1, CDR-H2 and CDR-H3). The CDR contains most of the residues responsible for the specific interaction of the antibody with the antigen and therefore contributes to the functional activity of the antibody molecule: it is the main determinant of antigen specificity.

[0080] The CDR boundaries and lengths defined precisely are subject to different classification and numbering systems. CDRs can therefore be referred to by Kabat, Chothia, contacts or any other boundary definitions (including numbering systems described herein). Although the boundaries are different, each of these systems has some degree of overlap in constituting the so-called "hypervariable region" within the variable sequence. Therefore, the CDR definitions according to these systems can differ in length and in the boundary regions relative to adjacent framework regions. See, for example, Kabat (a method based on interspecies cross-sequence variability), Chothia (a method based on crystallographic studies of antigen-antibody complexes) and / or MacCallum (Kabat et al., cited above; Chothia et al., J.Mol.Biol, 1987, 196: 901-917; and MacCallum et al., J.Mol.Biol, 1996, 262: 732). Another standard for characterizing an antigen binding site is the AbM definition used by Oxford Molecular's AbM antibody modeling software. See, e.g., Protein Sequence and Structure Analysis of Antibody Variable Domains. In: Antibody Engineering Lab Manual (Editors: Duebel, S. and Kontermann, R., Springer-Verlag, Heidelberg). To the extent that two residue identification techniques define overlapping but non-identical regions, they may be combined to define hybrid CDRs. However, numbering according to the so-called Kabat system is preferred.

[0081] Typically, CDRs form loop structures that can be classified as typical structures. The term "typical structure" refers to the main chain conformation adopted by the antigen binding (CDR) loop. From comparative structural studies, it was found that five of the six antigen binding loops have only limited pedigrees of available conformations. Each typical structure can be characterized by the torsion angle of the polypeptide main chain. Therefore, the corresponding loops between antibodies can have very similar three-dimensional structures, although there is high amino acid sequence variability in most of the loops (Chothia and Lesk, J.Mol.Biol., 1987, 196: 901; Chothia et al., Nature, 1989, 342: 877; Martin and Thornton, J.Mol.Biol, 1996, 263: 800). In addition, there is a relationship between the adopted ring structure and the amino acid sequence around it. The conformation of a specific typical class is determined by the length of the loop and the amino acid residues located at key positions within the loop and within the conservative framework (i.e., outside the loop). It can therefore be assigned to a specific typical class based on the presence of these key amino acid residues.

[0082] The term "typical structure" may also include considerations about the linear sequence of antibodies, such as those compiled by Kabat (Kabat et al., citation above). The Kabat numbering scheme (system) is a widely used standard for numbering the amino acid residues of the variable domains of antibodies in a consistent manner, and is also a preferred scheme for use in the present invention as mentioned elsewhere herein. Additional structural considerations may also be used to determine the typical structure of antibodies. For example, those differences that the Kabat numbering cannot fully reflect can be described by the numbering system of Chothia et al. and / or revealed by other techniques (e.g., crystallography and two-dimensional or three-dimensional computer modeling). Therefore, a given antibody sequence can be attributed to a typical category that allows, in particular, identification of appropriate framework sequences (chassis sequences) (e.g., based on the need to include a variety of typical structures in a library). The Kabat numbering and structural considerations of the antibody amino acid sequences described by Chothia et al. (citation above) and their implications for interpreting typical aspects of antibody structure are described in the document. The subunit structures and three-dimensional configurations of different classes of immunoglobulins are well known in the art. For a review of antibody structure, see Antibodies: A Laboratory Manual, Cold Spring Harbor Laboratory, Harlow et al., eds., 1988.

[0083] The CDR3 of light chain and especially the CDR3 of heavy chain can constitute the most important determinant of antigen binding in light chain and heavy chain variable region. In some antibody constructs, heavy chain CDR3 seems to constitute the main contact area between antigen and antibody. In vitro selection schemes with only CDR3 changes can be used to change the binding properties of antibodies or determine which residues contribute to antigen binding. Therefore, CDR3 is usually the largest source of molecular diversity in antibody binding sites. For example, H3 can be as short as two amino acid residues or greater than 26 amino acids.

[0084] In a classical full-length antibody or immunoglobulin, each light (L) chain is linked to a heavy (H) chain by a covalent disulfide bond, while the two H chains are linked to each other by one or more disulfide bonds depending on the H chain isotype. The CH domain closest to VH is usually referred to as CH1. The constant ("C") domain does not directly participate in antigen binding, but exhibits various effector functions, such as antibody-dependent cell-mediated cytotoxicity and complement activation. The Fc region of an antibody is contained within the heavy chain constant domain and is, for example, capable of interacting with Fc receptors localized on the cell surface.

[0085] The sequences of antibody genes after assembly and somatic mutation are highly variable, and these altered genes are estimated to encode 10 10 Different antibody molecules (Immunoglobulin Genes, 2nd edition, Jonio et al., ed., Academic Press, San Diego, CA, 1995). Therefore, the immune system provides a pedigree of immunoglobulins. The term "pedigree" refers to at least one nucleotide sequence derived entirely or in part from at least one sequence encoding at least one immunoglobulin. The sequence can be produced by the in vivo rearrangement of the V, D and J segments of the heavy chain and the V and J segments of the light chain. Alternatively, the sequence can be produced from cells in response to which rearrangement (e.g., in vitro stimulation) occurs. Alternatively, some or all of the sequences can be obtained by DNA splicing, nucleotide synthesis, mutagenesis and other methods, see, for example, U.S. Patent No. 5,565,332. The pedigree may include only one sequence or may include multiple sequences, including sequences in a collection of diverse heredity.

[0086] A preferred antibody construct according to the invention may also be defined as a bispecific antibody construct comprising a first (preferably human) binding domain that binds to human DLL3 on the surface of a target cell and a second binding domain that binds to human CD3 on the surface of a T cell, wherein the first binding domain binds to the same epitope of DLL3 as an antibody selected from the group consisting of DLL3-4, DLL3-5, DLL3-6, DLL3-7, DLL3-8, DLL3-9 and DLL3-10, i.e. an antibody comprising a VH region comprising CDR-H1, CDR-H2 and CDR-H3 and a VL region comprising CDR-L1, CDR-L2 and CDR-L3, selected from the group consisting of:

[0087] a) CDR-H1 as depicted in SEQ ID NO:31, CDR-H2 as depicted in SEQ ID NO:32, CDR-H3 as depicted in SEQ ID NO:33, CDR-L1 as depicted in SEQ ID NO:34, CDR-L2 as depicted in SEQ ID NO:35, and CDR-L3 as depicted in SEQ ID NO:36;

[0088] b) CDR-H1 as depicted in SEQ ID NO:41, CDR-H2 as depicted in SEQ ID NO:42, CDR-H3 as depicted in SEQ ID NO:43, CDR-L1 as depicted in SEQ ID NO:44, CDR-L2 as depicted in SEQ ID NO:45, and CDR-L3 as depicted in SEQ ID NO:46;

[0089] c) CDR-H1 as depicted in SEQ ID NO:51, CDR-H2 as depicted in SEQ ID NO:52, CDR-H3 as depicted in SEQ ID NO:53, CDR-L1 as depicted in SEQ ID NO:54, CDR-L2 as depicted in SEQ ID NO:55, and CDR-L3 as depicted in SEQ ID NO:56;

[0090] d) CDR-H1 as depicted in SEQ ID NO:61, CDR-H2 as depicted in SEQ ID NO:62, CDR-H3 as depicted in SEQ ID NO:63, CDR-L1 as depicted in SEQ ID NO:64, CDR-L2 as depicted in SEQ ID NO:65, and CDR-L3 as depicted in SEQ ID NO:66;

[0091] e) CDR-H1 as depicted in SEQ ID NO:71, CDR-H2 as depicted in SEQ ID NO:72, CDR-H3 as depicted in SEQ ID NO:73, CDR-L1 as depicted in SEQ ID NO:74, CDR-L2 as depicted in SEQ ID NO:75, and CDR-L3 as depicted in SEQ ID NO:76;

[0092] f) CDR-H1 as depicted in SEQ ID NO:81, CDR-H2 as depicted in SEQ ID NO:82, CDR-H3 as depicted in SEQ ID NO:83, CDR-L1 as depicted in SEQ ID NO:84, CDR-L2 as depicted in SEQ ID NO:85, and CDR-L3 as depicted in SEQ ID NO:86; and

[0093] g) CDR-H1 as depicted in SEQ ID NO:91, CDR-H2 as depicted in SEQ ID NO:92, CDR-H3 as depicted in SEQ ID NO:93, CDR-L1 as depicted in SEQ ID NO:94, CDR-L2 as depicted in SEQ ID NO:95 and CDR-L3 as depicted in SEQ ID NO:96.

[0094] Whether an antibody construct binds to the same DLL3 epitope as another given antibody construct can be measured, for example, by epitope mapping with a chimeric or truncated target molecule, for example as described above and in Example 2.

[0095] Preferred antibody constructs according to the present invention may also be defined as bispecific antibody constructs comprising a first (preferably human) binding domain that binds to human DLL3 on the surface of target cells and a second binding domain that binds to human CD3 on the surface of T cells, wherein the first binding domain competes for binding with an antibody selected from the group consisting of DLL3-4, DLL3-5, DLL3-6, DLL3-7, DLL3-8, DLL3-9 and DLL3-10, i.e. an antibody comprising a VH region comprising CDR-H1, CDR-H2 and CDR-H3 selected from the above and a VL region comprising CDR-L1, CDR-L2 and CDR-L3 selected from the above.

[0096] Whether an antibody construct competes for binding with another given antibody construct can be measured in a competitive assay such as a competitive ELISA or a cell-based competitive assay. Avidin-coupled microparticles (beads) can also be used. Similar to avidin-coated ELISA plates, these beads can each be used as a substrate that can be assayed when reacting with biotinylated proteins. Antigen is coated onto beads, followed by pre-coating with the first antibody. A second antibody is added and any additional binding is determined. Readout is performed via flow cytometry.

[0097] In one embodiment of the invention, the first binding domain of the antibody construct of the invention comprises a VH region selected from the group consisting of the VH regions depicted by SEQ ID NO:37, SEQ ID NO:47, SEQ ID NO:57, SEQ ID NO:67, SEQ ID NO:77, SEQ ID NO:87, SEQ ID NO:97, SEQ ID NO:107, SEQ ID NO:117, SEQ ID NO:435 and SEQ ID NO:529.

[0098] In another embodiment, the first binding domain of the antibody construct of the present invention comprises a VL region selected from the group consisting of the VL regions depicted by SEQ ID NO:38, SEQ ID NO:48, SEQ ID NO:58, SEQ ID NO:68, SEQ ID NO:78, SEQ ID NO:88, SEQ ID NO:98, SEQ ID NO:108, SEQ ID NO:118, SEQ ID NO:436 and SEQ ID NO:530.

[0099] In another embodiment, the first binding domain of the antibody construct of the present invention comprises a VH region and a VL region selected from the pairings of the VH region and the VL region depicted by the following items: SEQ ID NO:37+38, SEQ ID NO:47+48; SEQ ID NO:57+58; SEQ ID NO:67+68; SEQ ID NO:77+78; SEQ ID NO:87+88; SEQ ID NO:97+98; SEQ ID NO:107+108; SEQ ID NO:117+118; SEQ ID NO:435+436; and SEQ ID NO:529+530.

[0100] In another embodiment, the first binding domain of the antibody construct of the present invention comprises a polypeptide selected from the group consisting of the polypeptides depicted in SEQ ID NO:39, SEQ ID NO:49, SEQ ID NO:59, SEQ ID NO:69, SEQ ID NO:79, SEQ ID NO:89, SEQ ID NO:99, SEQ ID NO:109, SEQ ID NO:119, SEQ ID NO:437 and SEQ ID NO:531.

[0101] The above first binding domain (specified by its CDRs, VH and VL regions and combinations thereof) is characterized as a binding domain that binds to a DLL3 epitope contained within the region depicted as SEQ ID NO:258.

[0102] The term "bispecific" as used herein refers to an antibody construct that is "at least bispecific", i.e. it comprises at least a first binding domain and a second binding domain, wherein the first binding domain binds to one antigen or target (here: DLL3) and the second binding domain binds to another antigen or target (here: CD3). Thus, the antibody construct according to the invention comprises specificities for at least two different antigens or targets. The term "bispecific antibody construct" of the invention also encompasses multispecific antibody constructs, such as trispecific antibody constructs, the latter comprising three binding domains, or constructs with more than three (e.g. four, five ...) specificities.

[0103] Given that the antibody constructs according to the invention are (at least) bispecific, they do not occur naturally and they are clearly different from naturally occurring products. Therefore, a "bispecific" antibody construct or immunoglobulin is an artificial hybrid antibody or immunoglobulin having at least two different binding sites with different specificities. Bispecific antibody constructs can be produced by a variety of methods including hybridoma fusion or Fab' fragment connection. See, for example, Songsivilai and Lachmann, Clin. Exp. Immunol. 79: 315-321 (1990).

[0104] At least two binding domains and variable domains of the antibody construct of the present invention may or may not contain a peptide linker (spacer peptide). According to the present invention, the term "peptide linker" comprises an amino acid sequence that connects the amino acid sequence of one (variable and / or binding) domain and another (variable and / or binding) domain of the antibody construct of the present invention to each other. The basic technical feature of such peptide linkers is that they do not include any polymerization activity. Suitable peptide linkers include those described in U.S. Patent Nos. 4,751,180 and 4,935,233 or WO 88 / 09344. Peptide linkers may also be used to connect other domains or modules or regions (such as half-life extension domains) to the antibody construct of the present invention.

[0105] In the case of using a joint, this joint preferably has a certain length and sequence to ensure that each of the first and second domains can retain its different binding specificities independently of each other. For connecting the peptide joint of at least two binding domains (or two variable domains) in the antibody construct of the present invention, it is preferably a peptide joint comprising only a few amino acid residues (e.g., 12 amino acid residues or less). Therefore, a peptide joint of 12, 11, 10, 9, 8, 7, 6 or 5 amino acid residues is preferred. The envisioned peptide joint with less than 5 amino acids comprises 4, 3, 2 or 1 amino acid, wherein the joint rich in Gly is preferred. In the context of the "peptide joint", particularly preferred "single" amino acid is Gly. Therefore, the peptide joint can be composed of a single amino acid Gly. Another preferred embodiment of a peptide joint is characterized in that the amino acid sequence Gly-Gly-Gly-Gly-Ser, i.e., Gly4Ser (SEQ ID NO: 286), or its polymer, i.e., (Gly4Ser) x, wherein x is 1 or a larger integer (e.g., 2 or 3). Preferred linkers are depicted in SEQ ID NO: 285-293. The characteristics of such peptide linkers (without promoting secondary structure) are known in the art and are described, for example, in Dall'Acqua et al. (Biochem. (1998) 37, 9266-9273); Cheadle et al. (Mol Immunol (1992) 29, 21-30) and Raag and Whitlow (FASEB (1995) 9 (1), 73-80). In addition, peptide linkers that do not promote any secondary structure are preferred. The connection of the domains to each other can be provided, for example, by genetic engineering, as described in the Examples. Methods for preparing fused and operably linked bispecific single chain constructs and expressing the constructs in mammalian cells or bacteria are well known in the art (e.g. WO 99 / 54440 or Sambrook et al., Molecular Cloning: A Laboratory Manual, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, New York, 2001).

[0106] As described above, the present invention provides a preferred embodiment, wherein the antibody construct is in a form selected from the group consisting of: (scFv)2, scFv-single domain mAb, bifunctional antibody and oligomers of any of the foregoing forms. The term "in the form of..." does not exclude that the construct can be further modified, for example by connection or fusion to other parts, as described herein.

[0107] According to a particularly preferred embodiment and as recorded in the accompanying examples, the antibody constructs of the invention are "bispecific single-chain antibody constructs", more preferably bispecific "single-chain Fv" (scFv). Although the two domains (VL and VH) of the Fv fragment are encoded by separate genes, they can be joined using recombinant methods by synthetic linkers, as described above, which enable them to be made into a single protein chain, in which the VL and VH regions are paired to form a monovalent molecule; see, for example, Huston et al. (1988) Proc. Natl. Acad. Sci USA 85: 5879-5883). These antibody fragments are obtained using conventional techniques known to those skilled in the art, and the function of the fragments is evaluated in the same manner as for intact or full-length antibodies. Single-chain variable fragments (scFv) are therefore fusion proteins of the heavy chain (VH) and light chain (VL) variable regions of immunoglobulins, usually connected with a short linker peptide of about 10 to about 25 amino acids, preferably about 15 to 20 amino acids. The linker is usually rich in glycine for flexibility, and serine or threonine for solubility, and can connect the N-terminus of VH to the C-terminus of VL, or vice versa. This protein retains the specificity of the original immunoglobulin despite the removal of the constant region and the introduction of the linker.

[0108] Bispecific single chain molecules are known in the art and are described in WO 99 / 54440, Mack, J. Immunol. (1997), 158, 3965-3970, Mack, PNAS, (1995), 92, 7021-7025, Kufer, Cancer Immunol. Immunother., (1997), 45, 193-197, Blood, (2000), 95, 6, 2098-2103, Brühl, Immunol., (2001), 166, 2420-2426, Kipriyanov, J. Mol. Biol., (1999), 293, 41-56. The techniques described for the production of single-chain antibodies (see, inter alia, U.S. Pat. No. 4,946,778; Kontermann and Dübel (2010), loc. cit., and Little (2009), loc. cit.) can be adapted to produce single-chain antibody constructs that specifically recognize a selected target.

[0109] A bivalent (also referred to as bivalent) or bispecific single chain variable fragment (bi-scFv or di-scFv) of the form (scFv)2 can be engineered by connecting two scFv molecules (e.g., with a linker as described above). If the two scFv molecules have the same binding specificity, the resulting (scFv)2 molecule will preferably be referred to as bivalent (i.e., it has two valencies for the same target epitope). If the two scFv molecules have different binding specificities, the resulting (scFv)2 molecule will preferably be referred to as bispecific. The connection can be made by producing a single peptide chain with two VH regions and two VL regions, thereby obtaining a tandem scFv (see, e.g., Kufer P. et al., (2004) Trends in Biotechnology 22 (5): 238-244). Another possibility is to produce an scFv molecule with a linker peptide that is too short for the two variable regions to fold together (e.g., about five amino acids), thereby forcing the scFv to dimerize. This type is called a bifunctional antibody (see, e.g., Hollinger, Philipp et al., (July 1993) Proceedings of the National Academy of Sciences of the United States of America 90(14):6444-8.).

[0110] According to another preferred embodiment of the antibody construct of the present invention, the heavy chain (VH) and light chain (VL) of the binding domain (binding to the target antigen DLL3 or CD3) are not directly connected via a peptide linker as described above, but form a binding domain as described for the bifunctional antibody. Thus, the VH of the CD3 binding domain may be fused to the VL of the DLL3 binding domain via a peptide linker, and the VH of the DLL3 binding domain may be fused to the VL of the CD3 binding domain via such a peptide linker.

[0111] Single domain antibodies contain only one (monomeric) antibody variable domain that is capable of selectively binding to a specific antigen independently of other V regions or domains. The first single domain antibodies were engineered from heavy chain antibodies found in camels, and they are called V H H fragment. Cartilaginous fish also have heavy chain antibodies (IgNARs) from which V NARSingle domain antibodies of fragments. An alternative method is to split the dimeric variable domains from common immunoglobulins (e.g., from humans or rodents) into monomers, thereby obtaining VH or VL in the form of single domain Abs. Although most of the current research on single domain antibodies is based on heavy chain variable domains, nano antibodies derived from light chains also show specific binding to target epitopes. Examples of single domain antibodies are referred to as sdAbs, nano antibodies, or single variable domain antibodies.

[0112] Thus, (single domain mAb)2 is a monoclonal antibody construct composed of (at least) two single domain monoclonal antibodies, each of which is individually selected from the group consisting of VH, VL, V H H and V NAR The linker is preferably in the form of a peptide linker. Similarly, a "scFv single domain mAb" is a monoclonal antibody construct consisting of at least one single domain antibody as described above and one scFv molecule as described above. Likewise, the linker is preferably in the form of a peptide linker.

[0113] It is further contemplated that the present invention provides a bispecific antibody construct comprising a first binding domain that binds to human DLL3 on the surface of a target cell and a second binding domain that binds to human CD3 on the surface of a T cell, wherein the first binding domain binds to an epitope of DLL3 contained within the region depicted by SEQ ID NO:259.

[0114] Therefore, in another aspect of the invention, the first binding domain of the bispecific antibody construct comprises a VH region comprising CDR-H1, CDR-H2 and CDR-H3 and a VL region comprising CDR-L1, CDR-L2 and CDR-L3, which are selected from:

[0115] a) CDR-H1 as depicted in SEQ ID NO: 121, CDR-H2 as depicted in SEQ ID NO: 122, CDR-H3 as depicted in SEQ ID NO: 123, CDR-L1 as depicted in SEQ ID NO: 124, CDR-L2 as depicted in SEQ ID NO: 125, and CDR-L3 as depicted in SEQ ID NO: 126;

[0116] b) CDR-H1 as depicted in SEQ ID NO: 131, CDR-H2 as depicted in SEQ ID NO: 132, CDR-H3 as depicted in SEQ ID NO: 133, CDR-L1 as depicted in SEQ ID NO: 134, CDR-L2 as depicted in SEQ ID NO: 135, and CDR-L3 as depicted in SEQ ID NO: 136;

[0117] c) CDR-H1 as depicted in SEQ ID NO: 141, CDR-H2 as depicted in SEQ ID NO: 142, CDR-H3 as depicted in SEQ ID NO: 143, CDR-L1 as depicted in SEQ ID NO: 144, CDR-L2 as depicted in SEQ ID NO: 145, and CDR-L3 as depicted in SEQ ID NO: 146;

[0118] d) CDR-H1 as depicted in SEQ ID NO: 151, CDR-H2 as depicted in SEQ ID NO: 152, CDR-H3 as depicted in SEQ ID NO: 153, CDR-L1 as depicted in SEQ ID NO: 154, CDR-L2 as depicted in SEQ ID NO: 155, and CDR-L3 as depicted in SEQ ID NO: 156; and

[0119] e) CDR-H1 as depicted in SEQ ID NO: 161, CDR-H2 as depicted in SEQ ID NO: 162, CDR-H3 as depicted in SEQ ID NO: 163, CDR-L1 as depicted in SEQ ID NO: 164, CDR-L2 as depicted in SEQ ID NO: 165, and CDR-L3 as depicted in SEQ ID NO: 166;

[0120] f) CDR-H1 as depicted in SEQ ID NO: 131, CDR-H2 as depicted in SEQ ID NO: 439, CDR-H3 as depicted in SEQ ID NO: 133, CDR-L1 as depicted in SEQ ID NO: 134, CDR-L2 as depicted in SEQ ID NO: 135, and CDR-L3 as depicted in SEQ ID NO: 136;

[0121] g) CDR-H1 as depicted in SEQ ID NO: 131, CDR-H2 as depicted in SEQ ID NO: 440, CDR-H3 as depicted in SEQ ID NO: 133, CDR-L1 as depicted in SEQ ID NO: 134, CDR-L2 as depicted in SEQ ID NO: 135, and CDR-L3 as depicted in SEQ ID NO: 136;

[0122] h) CDR-H1 as depicted in SEQ ID NO: 131, CDR-H2 as depicted in SEQ ID NO: 132, CDR-H3 as depicted in SEQ ID NO: 133, CDR-L1 as depicted in SEQ ID NO: 441, CDR-L2 as depicted in SEQ ID NO: 135, and CDR-L3 as depicted in SEQ ID NO: 136;

[0123] i) CDR-H1 as depicted in SEQ ID NO: 131, CDR-H2 as depicted in SEQ ID NO: 132, CDR-H3 as depicted in SEQ ID NO: 133, CDR-L1 as depicted in SEQ ID NO: 442, CDR-L2 as depicted in SEQ ID NO: 135, and CDR-L3 as depicted in SEQ ID NO: 136;

[0124] j) CDR-H1 as depicted in SEQ ID NO: 131, CDR-H2 as depicted in SEQ ID NO: 132, CDR-H3 as depicted in SEQ ID NO: 133, CDR-L1 as depicted in SEQ ID NO: 443, CDR-L2 as depicted in SEQ ID NO: 135, and CDR-L3 as depicted in SEQ ID NO: 136;

[0125] k) CDR-H1 as depicted in SEQ ID NO: 131, CDR-H2 as depicted in SEQ ID NO: 132, CDR-H3 as depicted in SEQ ID NO: 133, CDR-L1 as depicted in SEQ ID NO: 444, CDR-L2 as depicted in SEQ ID NO: 135, and CDR-L3 as depicted in SEQ ID NO: 136;

[0126] 1) CDR-H1 as depicted in SEQ ID NO: 131, CDR-H2 as depicted in SEQ ID NO: 439, CDR-H3 as depicted in SEQ ID NO: 133, CDR-L1 as depicted in SEQ ID NO: 441, CDR-L2 as depicted in SEQ ID NO: 135, and CDR-L3 as depicted in SEQ ID NO: 136; and

[0127] m) CDR-H1 as depicted in SEQ ID NO: 131, CDR-H2 as depicted in SEQ ID NO: 440, CDR-H3 as depicted in SEQ ID NO: 133, CDR-L1 as depicted in SEQ ID NO: 442, CDR-L2 as depicted in SEQ ID NO: 135, and CDR-L3 as depicted in SEQ ID NO: 136.

[0128] In one embodiment, the first binding domain of the antibody construct of the invention comprises a VH region selected from the group consisting of the VH regions depicted in SEQ ID NO: 127, SEQ ID NO: 137, SEQ ID NO: 147, SEQ ID NO: 157, SEQ ID NO: 167, SEQ ID NO: 445, SEQ ID NO: 446, SEQ ID NO: 447, SEQ ID NO: 448, SEQ ID NO: 449, SEQ ID NO: 450, SEQ ID NO: 451, SEQ ID NO: 452, SEQ ID NO: 453, SEQ ID NO: 454, and SEQ ID NO: 455.

[0129] In another embodiment, the first binding domain of the antibody construct of the invention comprises a VL region selected from the group consisting of the VL regions depicted by SEQ ID NO: 128, SEQ ID NO: 138, SEQ ID NO: 148, SEQ ID NO: 158, SEQ ID NO: 168, SEQ ID NO: 456, SEQ ID NO: 457, SEQ ID NO: 458, SEQ ID NO: 459, SEQ ID NO: 460, SEQ ID NO: 461, SEQ ID NO: 462, SEQ ID NO: 463, SEQ ID NO: 464, SEQ ID NO: 465, SEQ ID NO: 466, SEQ ID NO: 467, SEQ ID NO: 468, SEQ ID NO: 469, and SEQ ID NO: 470.

[0130] In another embodiment, the first binding domain of the antibody construct of the present invention comprises a VH region and a VL region selected from the pairing of the VH region and the VL region depicted by the following items: SEQ ID NO: 127+128; SEQ ID NO: 137+138; SEQ ID NO: 147+148; SEQ ID NO: 157+158; SEQ ID NO: 167+168; SEQ ID NO; 137+456; SEQ ID NO; 137+457; SEQ ID NO: 137+458; SEQ ID NO: 137+459; SEQ ID NO: 137+460; SEQ ID NO: 445+138; SEQ ID NO: 446+138; SEQ ID NO: 447+138; SEQ ID NO: 445+460; SEQ ID NO: 448+461; SEQ ID NO: 449+462; SEQ ID NO: 450+463; SEQ ID NO: 451+452; SEQ ID NO: 453+454; SEQ ID NO: 454+455; SEQ ID NO: 455+456; SEQ ID NO: 453+456 SEQ ID NO: 450+464; SEQ ID NO: 450+465; SEQ ID NO: 450+466; SEQ ID NO: 450+467; SEQ ID NO: 450+468; SEQ ID NO: 451+463; NO:454+469; and SEQ ID NO:455+470.

[0131] In another embodiment, the first binding domain of the antibody construct of the invention comprises a polypeptide selected from the group consisting of the polypeptides depicted in SEQ ID NO: 129, SEQ ID NO: 139, SEQ ID NO: 149, SEQ ID NO: 159, SEQ ID NO: 169, SEQ ID NO: 471, SEQ ID NO: 472, SEQ ID NO: 473, SEQ ID NO: 474, SEQ ID NO: 475, SEQ ID NO: 476, SEQ ID NO: 477, SEQ ID NO: 478, SEQ ID NO: 479, SEQ ID NO: 480, SEQ ID NO: 481, SEQ ID NO: 482, SEQ ID NO: 483, SEQ ID NO: 484, SEQ ID NO: 485, SEQ ID NO: 486, SEQ ID NO: 487, SEQ ID NO: 488, SEQ ID NO: 489, SEQ ID NO: 490, SEQ ID NO: 491, SEQ ID NO: 492, and SEQ ID NO: 493. NO:493.

[0132] The above first binding domain (specified by its CDRs, VH and VL regions and combinations thereof) is characterized as a binding domain that binds to a DLL3 epitope contained within the region depicted as SEQ ID NO:259.

[0133] Another preferred antibody construct according to the invention may also be defined as a bispecific antibody construct comprising a first (preferably human) binding domain that binds to human DLL3 on the surface of a target cell and a second binding domain that binds to human CD3 on the surface of a T cell, wherein the first binding domain binds to the same epitope of DLL3 as an antibody selected from the group consisting of DLL3-13, DLL3-14 and DLL3-15, i.e. an antibody comprising a VH region comprising CDR-H1, CDR-H2 and CDR-H3 and a VL region comprising CDR-L1, CDR-L2 and CDR-L3, selected from the group consisting of:

[0134] a) CDR-H1 as depicted in SEQ ID NO: 121, CDR-H2 as depicted in SEQ ID NO: 122, CDR-H3 as depicted in SEQ ID NO: 123, CDR-L1 as depicted in SEQ ID NO: 124, CDR-L2 as depicted in SEQ ID NO: 125, and CDR-L3 as depicted in SEQ ID NO: 126;

[0135] b) CDR-H1 as depicted in SEQ ID NO: 131, CDR-H2 as depicted in SEQ ID NO: 132, CDR-H3 as depicted in SEQ ID NO: 133, CDR-L1 as depicted in SEQ ID NO: 134, CDR-L2 as depicted in SEQ ID NO: 135, and CDR-L3 as depicted in SEQ ID NO: 136; and

[0136] c) CDR-H1 as depicted in SEQ ID NO: 141, CDR-H2 as depicted in SEQ ID NO: 142, CDR-H3 as depicted in SEQ ID NO: 143, CDR-L1 as depicted in SEQ ID NO: 144, CDR-L2 as depicted in SEQ ID NO: 145, and CDR-L3 as depicted in SEQ ID NO: 146.

[0137] Another preferred antibody construct according to the present invention can also be defined as a bispecific antibody construct, which comprises a first (preferably human) binding domain that binds to human DLL3 on the surface of target cells and a second binding domain that binds to human CD3 on the surface of T cells, wherein the first binding domain competes for binding with an antibody selected from DLL3-13, DLL3-14 and DLL3-15, i.e. an antibody comprising a VH region comprising CDR-H1, CDR-H2 and CDR-H3 selected from the above and a VL region comprising CDR-L1, CDR-L2 and CDR-L3 selected from the above.

[0138] It is also envisioned that the antibody construct of the present invention has another function in addition to its function of binding to the target molecules DLL3 and CD3. In this form, the antibody construct is a trifunctional or multifunctional antibody construct that acts by targeting target cells via binding to DLL3, modulating cytotoxic T cell activity via CD3 binding, and providing another function (such as a fully functional Fc constant domain that modulates antibody-dependent cellular toxicity) via recruitment of effector cells such as NK cells, markers (fluorescence, etc.), therapeutic agents such as toxins or radionuclides, and / or substances that extend serum half-life, etc.

[0139] The example of the material of the serum half-life of the antibody construct of the present invention is extended includes the peptide, protein or protein domain that is fused or otherwise connected to the antibody construct. The group of peptide, protein or protein domain includes the peptide (referring to WO 2009 / 127691) that is attached to other proteins (such as serum albumin) with preferred pharmacokinetic properties in human body. The alternative concept of the peptide of this type of half-life extension includes the peptide that is attached to neonatal Fc receptor (FcRn, referring to WO2007 / 098420), which can also be used in the construct of the present invention. The concept of connecting a larger domain or complete protein of a protein includes, for example, fusion of human serum albumin, variants or mutants of human serum albumin (see WO 2011 / 051489, WO 2012 / 059486, WO 2012 / 150319, WO 2013 / 135896, WO 2014 / 072481, WO 2013 / 075066) or its domains and fusion of the constant region (Fc domain) of an immunoglobulin and its variants. Such variants of the Fc domain can be optimized / modified to allow the desired pairing of dimers or multimers, thereby eliminating Fc receptor binding (e.g., Fcγ receptors) or for other reasons. Another concept known in the art to extend the half-life of small protein compounds in the human body is the pegylation of compounds such as the antibody constructs of the present invention.

[0140] In a preferred embodiment, the bispecific antibody construct according to the invention may be linked (e.g., via a peptide bond) to a fusion partner (such as a protein or polypeptide or peptide), e.g., for the purpose of extending the serum half-life of the construct. These fusion partners may be selected from human serum albumin ("HSA" or "HALB") and sequence variants thereof, peptides that bind to HSA, peptides that bind to FcRn ("FcRn BP"), or constructs comprising an (antibody-derived) Fc region. Exemplary sequences of these fusion partners are depicted in SEQ ID NOs: 295-341. Typically, the fusion partner may be linked directly (e.g., via a peptide bond) or via a peptide bond such as (GGGGS) n (wherein "n" is an integer of 2 or greater, such as 2 or 3 or 4) is connected to the N-terminus or C-terminus of the bispecific antibody construct according to the present invention. Suitable peptide linkers are depicted in SEQ ID NOs: 285-293.

[0141] Therefore, preferred antibody constructs according to the invention comprise:

[0142] (a) a polypeptide comprising, starting from the N-terminus, in the following order:

[0143] - A polypeptide having an amino acid sequence selected from the group consisting of: SEQ ID NO:39, SEQ ID NO:49, SEQ ID NO:59, SEQ ID NO:69, SEQ ID NO:79, SEQ ID NO:89, SEQ ID NO:99, SEQ ID NO:109, SEQ ID NO:119, SEQ ID NO:129, SEQ ID NO:139, SEQ ID NO:149, SEQ ID NO:159, SEQ ID NO:169, SEQ ID NO:437, SEQ ID NO:471, SEQ ID NO:472, SEQ ID NO:473, SEQ ID NO:474, SEQ ID NO:475, SEQ ID NO:476, SEQ ID NO:477, SEQ ID NO:478, SEQ ID NO:479, SEQ ID NO:480, SEQ ID NO:481, SEQ ID NO:482, SEQ ID NO:483, SEQ ID NO:484, SEQ ID NO:485, SEQ ID NO:486, SEQ ID NO:487, SEQ ID NO:488, SEQ ID NO:489, SEQ ID NO:490, SEQ ID NO:491, SEQ ID NO:492, SEQ ID NO:493 and SEQ ID NO:531;

[0144] a peptide linker having an amino acid sequence selected from the group consisting of SEQ ID NOs: 285-293; and

[0145] a polypeptide having an amino acid sequence selected from the group consisting of SEQ ID NO:350, SEQ ID NO:359, SEQ ID NO:368, SEQ ID NO:377, SEQ ID NO:386, SEQ ID NO:395, SEQ ID NO:404, SEQ ID NO:413, SEQ ID NO:422, SEQ ID NO:431 and SEQ ID NO:434; and

[0146] Optional His tag, such as depicted in SEQ ID NO: 294;

[0147] (b) a polypeptide comprising, starting from the N-terminus, in the following order:

[0148] - A polypeptide having an amino acid sequence selected from the group consisting of: SEQ ID NO:39, SEQ ID NO:49, SEQ ID NO:59, SEQ ID NO:69, SEQ ID NO:79, SEQ ID NO:89, SEQ ID NO:99, SEQ ID NO:109, SEQ ID NO:119, SEQ ID NO:129, SEQ ID NO:139, SEQ ID NO:149, SEQ ID NO: 159, SEQ ID NO: 169, SEQ ID NO: 437, SEQ ID NO: 471, SEQ ID NO: 472, SEQ ID NO: 473, SEQ ID NO: 474, SEQ ID NO: 475, SEQ ID NO: 476, SEQ ID NO: 477, SEQ ID NO: 478, SEQ ID NO: 479, SEQ ID NO: 480, SEQ ID NO:481, SEQ ID NO:482, SEQ ID NO:483, SEQ ID NO:484, SEQ ID NO:485, SEQ ID NO:486, SEQ ID NO:487, SEQ ID NO:488, SEQ ID NO:489, SEQ ID NO:490, SEQ ID NO:491, SEQ ID NO:492, SEQ ID NO:493 and SEQ ID NO:531;

[0149] A peptide linker having an amino acid sequence selected from the group consisting of SEQ ID NOs: 285-293;

[0150] a polypeptide having an amino acid sequence selected from the group consisting of SEQ ID NO:350, SEQ ID NO:359, SEQ ID NO:368, SEQ ID NO:377, SEQ ID NO:386, SEQ ID NO:395, SEQ ID NO:404, SEQ ID NO:413, SEQ ID NO:422, SEQ ID NO:431 and SEQ ID NO:434;

[0151] · an optional peptide linker having an amino acid sequence selected from SEQ ID NOs: 285-293;

[0152] A polypeptide having an amino acid sequence selected from the group consisting of SEQ ID NOs: 295 and 301-330; and

[0153] Optional His tag, such as depicted in SEQ ID NO: 294;

[0154] (c) a polypeptide comprising, starting from the N-terminus, in the following order:

[0155] A polypeptide having the amino acid sequence QRFVTGHFGGLX1PANG (SEQ ID NO: 296), wherein X1 is Y or H; and

[0156] - A polypeptide having an amino acid sequence selected from the group consisting of: SEQ ID NO:39, SEQ ID NO:49, SEQ ID NO:59, SEQ ID NO:69, SEQ ID NO:79, SEQ ID NO:89, SEQ ID NO:99, SEQ ID NO:109, SEQ ID NO:119, SEQ ID NO:129, SEQ ID NO:139, SEQ ID NO:149, SEQ ID NO:159, SEQ ID NO:169, SEQ ID NO:437, SEQ ID NO:471, SEQ ID NO:472, SEQ ID NO:473, SEQ ID NO:474, SEQ ID NO:475, SEQ ID NO:476, SEQ ID NO:477, SEQ ID NO:478, SEQ ID NO:479, SEQ ID NO:480, SEQ ID NO:481, SEQ ID NO:482, SEQ ID NO:483, SEQ ID NO:484, SEQ ID NO:485, SEQ ID NO:486, SEQ ID NO:487, SEQ ID NO:488, SEQ ID NO:489, SEQ ID NO:490, SEQ ID NO:491, SEQ ID NO:492, SEQ ID NO:493 and SEQ ID NO:531;

[0157] A peptide linker having an amino acid sequence selected from the group consisting of SEQ ID NOs: 285-293;

[0158] a polypeptide having an amino acid sequence selected from the group consisting of SEQ ID NO:350, SEQ ID NO:359, SEQ ID NO:368, SEQ ID NO:377, SEQ ID NO:386, SEQ ID NO:395, SEQ ID NO:404, SEQ ID NO:413, SEQ ID NO:422, SEQ ID NO:431 and SEQ ID NO:434;

[0159] A polypeptide having the amino acid sequence QRFVTGHFGGLHPANG (SEQ ID NO: 298) or QRFCTGHFGGLHPCNG (SEQ ID NO: 300); and

[0160] Optional His tag, such as depicted in SEQ ID NO: 294;

[0161] (d) a polypeptide comprising, starting from the N-terminus, in the following order:

[0162] a polypeptide having an amino acid sequence selected from the group consisting of SEQ ID NO:348, SEQ ID NO:357, SEQ ID NO:366, SEQ ID NO:375, SEQ ID NO:384, SEQ ID NO:393, SEQ ID NO:402, SEQ ID NO:411, SEQ ID NO:420, SEQ ID NO:429, and SEQ ID NO:432;

[0163] A peptide linker having an amino acid sequence as depicted in SEQ ID NO: 292;

[0164] a polypeptide having an amino acid sequence selected from the group consisting of SEQ ID NO:38, SEQ ID NO:48, SEQ ID NO:58, SEQ ID NO:68, SEQ ID NO:78, SEQ ID NO:88, SEQ ID NO:98, SEQ ID NO:108, SEQ ID NO:118, SEQ ID NO:128, SEQ ID NO:138, SEQ ID NO:148, SEQ ID NO:158, SEQ ID NO:168, SEQ ID NO:436, SEQ ID NO:456, SEQ ID NO:457, SEQ ID NO:458, SEQ ID NO:459, SEQ ID NO:460, SEQ ID NO:461, SEQ ID NO:462, SEQ ID NO:463, SEQ ID NO:464, SEQ ID NO:465, SEQ ID NO:466, SEQ ID NO:467, SEQ ID NO:468, SEQ ID NO:469, SEQ ID NO:470, SEQ ID NO:471, SEQ ID NO:472, SEQ ID NO:473, SEQ ID NO:474, SEQ ID NO:475 NO:470 and SEQ ID NO:530;

[0165] A polypeptide having the amino acid sequence depicted in SEQ ID NO: 331; and

[0166] A polypeptide comprising, starting from the N-terminus, in the following order:

[0167] a polypeptide having an amino acid sequence selected from the group consisting of SEQ ID NO:37, SEQ ID NO:47, SEQ ID NO:57, SEQ ID NO:67, SEQ ID NO:77, SEQ ID NO:87, SEQ ID NO:97, SEQ ID NO:107, SEQ ID NO:117, SEQ ID NO:127, SEQ ID NO:137, SEQ ID NO:147, SEQ ID NO:157, SEQ ID NO:167, SEQ ID NO:435, SEQ ID NO:445, SEQ ID NO:446, SEQ ID NO:447, SEQ ID NO:448, SEQ ID NO:449, SEQ ID NO:450, SEQ ID NO:451, SEQ ID NO:452, SEQ ID NO:453, SEQ ID NO:454, and SEQ ID NO:455, and SEQ ID NO:529;

[0168] A peptide linker having an amino acid sequence as depicted in SEQ ID NO: 292;

[0169] a polypeptide having an amino acid sequence selected from the group consisting of SEQ ID NO:349, SEQ ID NO:358, SEQ ID NO:367, SEQ ID NO:376, SEQ ID NO:385, SEQ ID NO:394, SEQ ID NO:403, SEQ ID NO:412, SEQ ID NO:421, SEQ ID NO:430, and SEQ ID NO:433 followed by a serine residue at the C-terminus; and

[0170] A polypeptide having the amino acid sequence depicted by SEQ ID NO: 332;

[0171] (e) a polypeptide comprising, starting from the N-terminus, in the following order:

[0172] a polypeptide having an amino acid sequence selected from the group consisting of SEQ ID NO:348, SEQ ID NO:357, SEQ ID NO:366, SEQ ID NO:375, SEQ ID NO:384, SEQ ID NO:393, SEQ ID NO:402, SEQ ID NO:411, SEQ ID NO:420, SEQ ID NO:429, and SEQ ID NO:432;

[0173] A peptide linker having an amino acid sequence as depicted in SEQ ID NO: 292;

[0174] - A polypeptide having an amino acid sequence selected from: SEQ ID NO:38, SEQ ID NO:48, SEQ ID NO:58, SEQ ID NO:68, SEQ ID NO:78, SEQ ID NO:88, SEQ ID NO:98, SEQ ID NO:108, SEQ ID NO:118, SEQ ID NO:128, SEQ ID NO:138, SEQ ID NO:148, SEQ ID NO:158, SEQ ID NO:168, SEQ ID NO:436, SEQ ID NO:456, SEQ ID NO:457, SEQ ID NO:458, SEQ ID NO:459, SEQ ID NO:460, SEQ ID NO:461, SEQ ID NO:462, SEQ ID NO:463, SEQ ID NO:464, SEQ ID NO:465, SEQ ID NO:466, SEQ ID NO:467, SEQ ID NO:468, SEQ ID NO:469, SEQ ID NO:470 and SEQ IDNO: 530; and

[0175] A polypeptide having the amino acid sequence depicted by SEQ ID NO: 333; and

[0176] A polypeptide comprising, starting from the N-terminus, in the following order:

[0177] a polypeptide having an amino acid sequence selected from the group consisting of SEQ ID NO:37, SEQ ID NO:47, SEQ ID NO:57, SEQ ID NO:67, SEQ ID NO:77, SEQ ID NO:87, SEQ ID NO:97, SEQ ID NO:107, SEQ ID NO:117, SEQ ID NO:127, SEQ ID NO:137, SEQ ID NO:147, SEQ ID NO:157, SEQ ID NO:167, SEQ ID NO:435, SEQ ID NO:445, SEQ ID NO:446, SEQ ID NO:447, SEQ ID NO:448, SEQ ID NO:449, SEQ ID NO:450, SEQ ID NO:451, SEQ ID NO:452, SEQ ID NO:453, SEQ ID NO:454, and SEQ ID NO:455, and SEQ ID NO:529;

[0178] A peptide linker having an amino acid sequence as depicted in SEQ ID NO: 292;

[0179] a polypeptide having an amino acid sequence selected from the group consisting of SEQ ID NO:349, SEQ ID NO:358, SEQ ID NO:367, SEQ ID NO:376, SEQ ID NO:385, SEQ ID NO:394, SEQ ID NO:403, SEQ ID NO:412, SEQ ID NO:421, SEQ ID NO:430, and SEQ ID NO:433 followed by a serine residue at the C-terminus; and

[0180] A polypeptide having the amino acid sequence depicted by SEQ ID NO: 334;

[0181] (f) a polypeptide comprising, starting from the N-terminus, in the following order:

[0182] - A polypeptide having an amino acid sequence selected from the group consisting of SEQ ID NO:39, SEQ ID NO:49, SEQ ID NO:59, SEQ ID NO:69, SEQ ID NO:79, SEQ ID NO:89, SEQ ID NO:99, SEQ ID NO:109, SEQ ID NO:119, SEQ ID NO:129, SEQ ID NO:139, SEQ ID NO:149, SEQ ID NO:159, SEQ ID NO:169, SEQ ID NO:437, SEQ ID NO:471, SEQ ID NO:472, SEQ ID NO:473, SEQ ID NO:474, SEQ ID NO:475, SEQ ID NO:476, SEQ ID NO:477, SEQ ID NO:478, SEQ ID NO:479, SEQ ID NO:480, SEQ ID NO:481, SEQ ID NO:482, SEQ ID NO:483, SEQ ID NO:484, SEQ ID NO:485, SEQ ID NO:486, SEQ ID NO:487, SEQ ID NO:488, SEQ ID NO:489, SEQ ID NO:490, SEQ ID NO:491, SEQ ID NO:492, SEQ ID NO:493 and SEQ ID NO:531;

[0183] A peptide linker having an amino acid sequence selected from the group consisting of SEQ ID NOs: 285-293;

[0184] a polypeptide having an amino acid sequence selected from the group consisting of SEQ ID NO:350, SEQ ID NO:359, SEQ ID NO:368, SEQ ID NO:377, SEQ ID NO:386, SEQ ID NO:395, SEQ ID NO:404, SEQ ID NO:413, SEQ ID NO:422, SEQ ID NO:431 and SEQ ID NO:434;

[0185] A polypeptide having the amino acid sequence depicted by SEQ ID NO: 335; and a polypeptide having the amino acid sequence depicted by SEQ ID NO: 336;

[0186] (g) a polypeptide comprising, starting from the N-terminus, in the following order:

[0187] - A polypeptide having an amino acid sequence selected from the group consisting of: SEQ ID NO:39, SEQ ID NO:49, SEQ ID NO:59, SEQ ID NO:69, SEQ ID NO:79, SEQ ID NO:89, SEQ ID NO:99, SEQ ID NO:109, SEQ ID NO:119, SEQ ID NO:129, SEQ ID NO:139, SEQ ID NO:149, SEQ ID NO:159, SEQ ID NO:169, SEQ ID NO:437, SEQ ID NO:471, SEQ ID NO:472, SEQ ID NO:473, SEQ ID NO:474, SEQ ID NO:475, SEQ ID NO:476, SEQ ID NO:477, SEQ ID NO:478, SEQ ID NO:479, SEQ ID NO:480, SEQ ID NO:481, SEQ ID NO:482, SEQ ID NO:483, SEQ ID NO:484, SEQ ID NO:485, SEQ and

[0188] A polypeptide having the amino acid sequence depicted by SEQ ID NO: 337; and

[0189] A polypeptide comprising, starting from the N-terminus, in the following order:

[0190] a polypeptide having an amino acid sequence selected from the group consisting of SEQ ID NO:350, SEQ ID NO:359, SEQ ID NO:368, SEQ ID NO:377, SEQ ID NO:386, SEQ ID NO:395, SEQ ID NO:404, SEQ ID NO:413, SEQ ID NO:422, SEQ ID NO:431 and SEQ ID NO:434; and

[0191] A polypeptide having the amino acid sequence depicted by SEQ ID NO: 338;

[0192] (h) a polypeptide comprising, starting from the N-terminus, in the following order:

[0193] - A polypeptide having an amino acid sequence selected from the group consisting of: SEQ ID NO:39, SEQ ID NO:49, SEQ ID NO:59, SEQ ID NO:69, SEQ ID NO:79, SEQ ID NO:89, SEQ ID NO:99, SEQ ID NO:109, SEQ ID NO:119, SEQ ID NO:129, SEQ ID NO:139, SEQ ID NO:149, SEQ ID NO:159, SEQ ID NO:169, SEQ ID NO:437, SEQ ID NO:471, SEQ ID NO:472, SEQ ID NO:473, SEQ ID NO:474, SEQ ID NO:475, SEQ ID NO:476, SEQ ID NO:477, SEQ ID NO:478, SEQ ID NO:479, SEQ ID NO:480, SEQ ID NO:481, SEQ ID NO:482, SEQ ID NO:483, SEQ ID NO:484, SEQ ID NO:485, SEQ and

[0194] A polypeptide having the amino acid sequence depicted by SEQ ID NO: 339; and

[0195] A polypeptide comprising, starting from the N-terminus, in the following order:

[0196] a polypeptide having an amino acid sequence selected from the group consisting of SEQ ID NO:350, SEQ ID NO:359, SEQ ID NO:368, SEQ ID NO:377, SEQ ID NO:386, SEQ ID NO:395, SEQ ID NO:404, SEQ ID NO:413, SEQ ID NO:422, SEQ ID NO:431 and SEQ ID NO:434; and

[0197] A polypeptide having the amino acid sequence depicted in SEQ ID NO: 340; or

[0198] (i) a polypeptide comprising, starting from the N-terminus, in the following order:

[0199] - A polypeptide having an amino acid sequence selected from the group consisting of: SEQ ID NO:39, SEQ ID NO:49, SEQ ID NO:59, SEQ ID NO:69, SEQ ID NO:79, SEQ ID NO:89, SEQ ID NO:99, SEQ ID NO:109, SEQ ID NO:119, SEQ ID NO:129, SEQ ID NO:139, SEQ ID NO:149, SEQ ID NO:159, SEQ ID NO:169, SEQ ID NO:437, SEQ ID NO:471, SEQ ID NO:472, SEQ ID NO:473, SEQ ID NO:474, SEQ ID NO:475, SEQ ID NO:476, SEQ ID NO:477, SEQ ID NO:478, SEQ ID NO:479, SEQ ID NO:480, SEQ ID NO:481, SEQ ID NO:482, SEQ ID NO:483, SEQ ID NO:484, SEQ ID NO:485, SEQ ID NO:486, SEQ ID NO:487, SEQ ID NO:488, SEQ ID NO:489, SEQ ID NO:490, SEQ ID NO:491, SEQ ID NO:492, SEQ ID NO:493 and SEQ ID NO:531;

[0200] A peptide linker having an amino acid sequence selected from the group consisting of SEQ ID NOs: 285-293;

[0201] a polypeptide having an amino acid sequence selected from the group consisting of SEQ ID NO:350, SEQ ID NO:359, SEQ ID NO:368, SEQ ID NO:377, SEQ ID NO:386, SEQ ID NO:395, SEQ ID NO:404, SEQ ID NO:413, SEQ ID NO:422, SEQ ID NO:431 and SEQ ID NO:434; and

[0202] A polypeptide having the amino acid sequence depicted by SEQ ID NO: 341.

[0203] For example, preferred bispecific antibody constructs of the present invention comprise or consist of a polypeptide selected from the group consisting of the polypeptides depicted in:

[0204] SEQ ID NO:224, SEQ ID NO:225, SEQ ID NO:226, SEQ IDNO:227, SEQ ID NO:228, SEQ ID NO:229, SEQ ID NO:230, SEQ ID NO:231, SEQ ID NO:232, SEQ ID NO:233, SEQ IDNO:234, SEQ ID NO:235, SEQ ID NO:236, SEQ ID NO:237, SEQ ID NO:242, SEQ ID NO:243, SEQ ID NO:244, SEQ ID NO:245, SEQ ID NO:246, SEQ ID NO:247, SEQ ID NO:248, SEQ ID NO:249, SEQ ID NO:250 and SEQ ID NO:251.

[0205] As described above, several preferred antibody constructs of the invention are modified by fusion to another moiety, such as albumin or an albumin variant. If these fusion constructs are characterized for their properties, such as, in particular, their target affinity or cytotoxic activity, the skilled artisan will appreciate that similar fusion constructs or unmodified bispecific antibody constructs can be expected to have similar (or possibly even better) properties. For example, if a bispecific antibody construct fused to albumin has a measurable or desired cytotoxic activity or target affinity, it can be expected that the same / similar or even higher cytotoxic activity / target affinity will be observed for the same construct without albumin.

[0206] According to another preferred embodiment, the bispecific antibody construct of the present invention comprises (in addition to the two binding domains) a third domain containing two polypeptide monomers, each of which comprises a hinge, a CH2 and a CH3 domain, wherein the two polypeptides (or polypeptide monomers) are fused to each other via a peptide linker. Preferably, the third domain comprises in order from N-terminal to C-terminal: hinge-CH2-CH3-linker-hinge-CH2-CH3. The preferred amino acid sequence of the third domain is depicted in SEQ ID NO:541-548. Each of the polypeptide monomers preferably has an amino acid sequence selected from SEQ ID NO:533-540 or having at least 90% identity with these sequences. In another preferred embodiment, the first and second binding domains of the bispecific antibody construct of the present invention are fused to the third domain via a peptide linker, and the peptide linker is, for example, selected from SEQ ID NO:285, 286, 288, 289, 290, 292 and 293.

[0207] According to the present invention, "hinge" is the hinge region of IgG. This region can be identified by analogy using Kabat numbering, see Kabat position 223-243. According to the above, the minimum requirement for "hinge" is the amino acid residues corresponding to the IgG1 sequence segment from D231 to P243 according to Kabat numbering. The terms CH2 and CH3 refer to immunoglobulin heavy chain constant regions 2 and 3. These regions can also be identified by analogy using Kabat numbering, see Kabat position 244-360 of CH2 and Kabat position 361-478 of CH3. It should be understood that there are some variations between immunoglobulins with respect to their IgG1 Fc region, IgG2 Fc region, IgG3 Fc region, IgG4 Fc region, IgM Fc region, IgA Fc region, IgD Fc region and IgE Fc region (see, e.g., Padlan, Molecular Immunology, 31 (3), 169-217 (1993)). The term Fc monomer refers to the last two heavy chain constant regions of IgA, IgD and IgG, and the last three heavy chain constant regions of IgE and IgM. Fc monomers may also include the flexible hinge at the N-terminal of these domains. For IgA and IgM, Fc monomers may include J chains. For IgG, the Fc moiety comprises the hinge between immunoglobulin domains CH2 and CH3 and the first two domains and CH2. Although the boundaries of the Fc moiety of immunoglobulin can vary, the example of the human IgG heavy chain Fc moiety comprising functional hinge, CH2 and CH3 domains may be defined as, for example, comprising residues (hinge domains) D231 to (CH3 domain C-terminal) P476 or D231 to L476 for IgG4, respectively, wherein numbering is according to Kabat.

[0208] The antibody construct of the present invention may therefore comprise, in order from N-terminus to C-terminus:

[0209] (a) a first binding domain;

[0210] (b) a peptide linker having an amino acid sequence selected from the group consisting of SEQ ID NOs: 286, 292 and 293;

[0211] (c) a second binding domain;

[0212] (d) a peptide linker having an amino acid sequence selected from the group consisting of SEQ ID NO: 285, 286, 288, 289, 290, 292 and 293;

[0213] (e) a first polypeptide monomer of the third domain (comprising hinge, CH2 and CH3 domains);

[0214] (f) a peptide linker having an amino acid sequence selected from the group consisting of SEQ ID NOs: 550, 551, 552 and 553; and

[0215] (g) A second polypeptide monomer of the third domain (comprising hinge, CH2 and CH3 domains).

[0216] It is also preferred that the antibody construct of the present invention comprises, in order from N-terminus to C-terminus:

[0217] a first binding domain having an amino acid sequence selected from the group consisting of SEQ ID NO:39, SEQ ID NO:49, SEQ ID NO:59, SEQ ID NO:69, SEQ ID NO:79, SEQ ID NO:89, SEQ ID NO:99, SEQ ID NO:109, SEQ ID NO:119, SEQ ID NO:129, SEQ ID NO:139, SEQ ID NO:149, SEQ ID NO:159, SEQ ID NO:169, SEQ ID NO:437, SEQ ID NO:471, SEQ ID NO:472, SEQ ID NO:473, SEQ ID NO:474, SEQ ID NO:475, SEQ ID NO:476, SEQ ID NO:477, SEQ ID NO:478, SEQ ID NO:479, SEQ ID NO:480, SEQ ID NO:481, SEQ ID NO:482, SEQ ID NO:483, SEQ ID NO:484, SEQ ID NO:485 NO:485, SEQ ID NO:486, SEQ ID NO:487, SEQ ID NO:488, SEQ ID NO:489, SEQ ID NO:490, SEQ ID NO:491, SEQ ID NO:492, SEQ ID NO:493 and SEQ ID NO:531;

[0218] A peptide linker having an amino acid sequence selected from the group consisting of SEQ ID NOs: 286, 292 and 293;

[0219] a second binding domain having an amino acid sequence selected from the group consisting of SEQ ID NO:350, SEQ ID NO:359, SEQ ID NO:368, SEQ ID NO:377, SEQ ID NO:386, SEQ ID NO:395, SEQ ID NO:404, SEQ ID NO:413, SEQ ID NO:422, SEQ ID NO:431, and SEQ ID NO:434;

[0220] a peptide linker having an amino acid sequence selected from the group consisting of SEQ ID NO: 285, 286, 288, 289, 290, 292 and 293; and

[0221] A third domain having an amino acid sequence selected from the group consisting of SEQ ID NOs: 541-548.

[0222] Therefore, in a preferred embodiment, the antibody construct of the present invention comprises or consists of a polypeptide selected from the group consisting of the polypeptides depicted in SEQ ID NO:517, SEQ ID NO:518, SEQ ID NO:519, SEQ ID NO:520, SEQ ID NO:521, SEQ ID NO:522, SEQ ID NO:523, SEQ ID NO:524, SEQ ID NO:525, SEQ ID NO:526, SEQ ID NO:527 and SEQ ID NO:528.

[0223] The sequence listing (Table 18) also provides sequence variations for binders named DLL3-4 and DLL3-14. Point mutations inserted into these sequence variants are identified based on the position of this mutation within the corresponding scFv molecule. It should be understood that alternative ways of identifying these positions are also possible depending on the reference polypeptide (which can also be a CDR region or a VH / VL region). For example, the variant named DLL3-4-001 has a G44C-G243C double mutation in its scFv molecule (SEQ ID NO: 437). This translates into a G44C mutation in the corresponding VH chain (SEQ ID NO: 435) and a G101C mutation in the corresponding VL chain (SEQ ID NO: 436).

[0224] Covalent modifications of antibody constructs are also included within the scope of the present invention and are usually, but not necessarily always, performed post-translationally. For example, several types of covalent modifications of antibody constructs are introduced into the molecule by reacting specific amino acid residues of the antibody construct with an organic derivatizing agent that reacts with selected side chains or N-terminal or C-terminal residues.

[0225] Cysteinyl residues are most commonly reacted with α-haloacetates (and corresponding amines) such as chloroacetic acid or chloroacetamide to produce carboxymethyl or carboxamidomethyl derivatives. Cysteinyl residues are also derivatized by reaction with bromotrifluoroacetone, α-bromo-β-(5-imidazolyl)propionic acid, chloroacetyl phosphate, N-alkylmaleimide, 3-nitro-2-pyridyl disulfide, methyl 2-pyridyl disulfide, p-chloromercuric benzoate, 2-chloromercuric-4-nitrophenol, or chloro-7-nitrobenzo-2-oxa-1,3-diazole.

[0226] Histidyl residues are derivatized by reaction with diethylpyrocarbonate at pH 5.5-7.0 because this reagent is relatively specific for histidyl side chains. Para-bromophenacyl bromide is also useful; the reaction is preferably carried out in 0.1 M sodium cacodylate at pH 6.0. Lysyl and amino terminal residues are reacted with succinic acid or other carboxylic acid anhydrides. Derivatization with these reagents has the effect of reversing the charge of the lysinyl residues. Other reagents that can be used to derivatize residues containing α-amino groups include imidoesters such as methyl pyridoxal imide; pyridoxal phosphate; pyridoxal; chloroborohydride; trinitrobenzenesulfonic acid; O-methylisourea; 2,4-pentanedione and transaminase-catalyzed reactions with glyoxylate.

[0227] Arginyl residues are modified by reaction with one or several conventional reagents, among which are phenylglyoxal, 2,3-butanedione, 1,2-cyclohexanedione and ninhydrin. Due to the high pKa of the guanidine functional group, the derivatization of arginine residues requires the reaction to be carried out under alkaline conditions. In addition, these reagents react with the groups of lysine as well as with the arginine epsilon-amino group.

[0228] Specific modifications can be made to tyrosyl residues, with particular interest in introducing spectral labels into tyrosyl residues by reaction with aromatic diazo compounds or tetranitromethane. Most commonly, N-acetylimidazole and tetranitromethane are used to form O-acetyl tyrosyl species and 3-nitro derivatives, respectively. Tyrosyl residues are modified using 125 I or 131 I iodination to prepare labeled proteins for radioimmunoassay, the chloramine T method described above is suitable.

[0229] The carboxyl side groups (aspartyl or glutamyl) are selectively modified by reaction with carbodiimide (R'-N=C=N--R'), wherein R and R' are optionally different alkyl groups, such as 1-cyclohexyl-3-(2-morpholinyl-4-ethyl)carbodiimide or 1-ethyl-3-(4-azonia-4,4-dimethylpentyl)carbodiimide. In addition, aspartyl and glutamyl residues are converted to asparaginyl and glutaminyl residues by reaction with ammonium ions.

[0230] Derivatization with bifunctional agents can be used to crosslink the antibody constructs of the invention to a water-insoluble support matrix or surface used in a variety of methods. Commonly used crosslinking agents include, for example, 1,1-bis(diazoacetyl)-2-phenylethane, glutaraldehyde, N-hydroxysuccinimide esters (e.g., esters with 4-azidosalicylic acid), homobifunctional imide esters (including disuccinimidyl esters such as 3,3'-dithiobis(succinimidyl propionate) and bifunctional maleimides such as bis-N-maleimido-1,8-octane. Such as 3-[(p-azidophenyl)dithio]propionyl Derivatization agents of amine methyl esters generate photoactivatable intermediates capable of forming crosslinks in the presence of light. Alternatively, reactive water-insoluble matrices (such as cyanogen bromide activated carbohydrates) and reactive substrates as described in U.S. Pat. Nos. 3,969,287, 3,691,016, 4,195,128, 4,247,642, 4,229,537, and 4,330,440 are used for protein immobilization.

[0231] Glutaminyl and asparaginyl residues are often deamidated to form corresponding glutamyl and aspartyl residues, respectively. Alternatively, these residues are deamidated under moderately acidic conditions. Any form of these residues falls within the scope of the present invention.

[0232] Other modifications include hydroxylation of proline and lysine; phosphorylation of the hydroxyl groups of seryl or threonyl residues; methylation of the α-amino groups of lysine, arginine, and histidine side chains (TECreighton, Proteins: Structure and Molecular Properties, WH Freeman & Co., San Francisco, 1983, pp. 79-86); acetylation of the N-terminal amine; and amidation of any C-terminal carboxyl group.

[0233] Another type of covalent modification of antibody constructs included within the scope of the present invention includes altering the glycosylation pattern of the protein. As is known in the art, the glycosylation pattern may depend on the sequence of the protein (e.g., the presence or absence of specific glycosylated amino acid residues, discussed below) or the host cell or organism in which the protein is produced. Specific expression systems are discussed below.

[0234] Glycosylation of polypeptides is usually N-linked or O-linked. N-linked refers to the attachment of the carbohydrate moiety to the side chain of an asparagine residue. The tripeptide sequences asparagine-X-serine and asparagine-X-threonine (where X is any amino acid except proline) are recognition sequences for enzymatic attachment of the carbohydrate moiety to the asparagine side chain. Therefore, the presence of any of these tripeptide sequences in a polypeptide can generate potential glycosylation sites. O-linked glycosylation refers to the attachment of one of the sugars N-acetylgalactosamine, galactose or xylose to a hydroxyamino acid (serine or threonine is most commonly used, although 5-hydroxyproline or 5-hydroxylysine may also be used).

[0235] Addition of glycosylation sites to the antibody construct is conveniently accomplished by altering the amino acid sequence so that it contains one or more of the above tripeptide sequences (for N-linked glycosylation sites). The alteration may also be made by adding or replacing one or more serine or threonine residues to the starting sequence (for O-linked glycosylation sites). For convenience, the amino acid sequence of the antibody construct is preferably altered by changes at the DNA level, particularly by mutating the DNA encoding the polypeptide at a preselected base to produce codons that will translate into the desired amino acids.

[0236] Another means of increasing the number of carbohydrate moieties on the antibody construct is by chemical or enzymatic coupling of glycosides to proteins. These procedures are advantageous because they do not require production of proteins in host cells with glycosylation capabilities for N-linked and O-linked glycosylation. Depending on the coupling mode used, sugars may be linked to (a) arginine and histidine, (b) free carboxyl groups, (c) free sulfhydryl groups, such as sulfhydryl groups of cysteine, (d) free hydroxyl groups, such as hydroxyl groups of serine, threonine or hydroxyproline, (e) aromatic residues, such as aromatic residues of phenylalanine, tyrosine or tryptophan, or (f) amide groups of glutamine. These methods are described in WO 87 / 05330 and Aplin and Wriston, 1981, CRC Crit. Rev. Biochem., pp. 259-306.

[0237] Removal of carbohydrate moieties present on the starting antibody construct can be achieved chemically or enzymatically. Chemical deglycosylation requires exposure of the protein to a triflic acid compound or an equivalent compound. This treatment causes most or all sugars other than the linked sugar (N-acetylglucosamine or N-acetylgalactosamine) to be cleaved, leaving intact polypeptides. Chemical deglycosylation is described by Hakimuddin et al., 1987, Arch.Biochem.Biophys.259:52 and Edge et al., 1981, Anal.Biochem.118:131. Enzymatic cleavage of carbohydrate moieties on polypeptides can be achieved by using a variety of endoglycosidases and exoglycosidases as described by Thotakura et al., 1987, Meth.Enzymol.138:350. Glycosylation at potential glycosylation sites can be prevented by using tunicamycin compounds as described by Duskin et al., 1982, J.Biol.Chem.257:3105. Tunicamycin blocks the formation of protein-N-glycosidic bonds.

[0238] Other modifications of antibody constructs are also contemplated herein. For example, another type of covalent modification of antibody constructs includes connecting antibody constructs to various non-protein polymers in a manner described in U.S. Patent Nos. 4,640,835, 4,496,689, 4,301,144, 4,670,417, 4,791,192 or 4,179,337, including but not limited to various polyols, such as polyethylene glycol, polypropylene glycol, polyoxyalkylene or copolymers of polyethylene glycol and polypropylene glycol. In addition, as known in the art, amino acid substitutions can be performed at multiple positions within the antibody construct, for example to promote the addition of polymers (such as PEG).

[0239] In some embodiments, the covalent modification of the antibody constructs of the present invention includes the addition of one or more labels. The labeling group can be coupled to the antibody construct via spacer arms of various lengths to reduce potential steric hindrance. Various methods of labeling proteins are known in the art and can be used to perform the present invention. The term "label" or "labeling group" refers to any detectable label. In general, labels belong to a variety of categories, depending on the assay for detecting the label, the following examples include but are not limited to:

[0240] a) isotopic labeling, which may be a radioactive isotope or a heavy isotope, such as a radioisotope or a radionuclide (e.g. 3 H. 14 C. 15 N. 35 S. 89 Zr, 90 Y. 99 Tc, 111 In,125 I. 131 I)

[0241] b) Magnetic labels (e.g. magnetic particles)

[0242] c) Redox-active part

[0243] d) Optical dyes (including but not limited to chromophores, phosphors and fluorophores), such as fluorescent groups (e.g. FITC, rhodamine, lanthanide phosphors), chemiluminescent groups, and fluorophores that can be "small molecule" fluorescent or protein fluorescent

[0244] e) Enzyme groups (e.g. horseradish peroxidase, β-galactosidase, luciferase, alkaline phosphatase)

[0245] f) Biotinylated groups

[0246] g) a predetermined polypeptide epitope recognized by the secondary reporter protein (e.g., a leucine zipper pair sequence, a binding site for a secondary antibody, a metal binding domain, an epitope tag, etc.)

[0247] "Fluorescent label" means any molecule that can be detected via its intrinsic fluorescent properties. Suitable fluorescent labels include, but are not limited to, fluorescein, rhodamine, tetramethylrhodamine, eosin, erythrosine, coumarin, methylcoumarin, pyrene, Malacite Green, stilbene, Lucifer Yellow, Cascade Blue J, Texas Red, IAEDANS, EDANS, BODIPY FL, LC Red 640, Cy 5, Cy 5.5, LC Red 705, Oregon Green, Alexa-Fluor dyes (Alexa Fluor 350, Alexa Fluor 430, Alexa Fluor 488, Alexa Fluor 546, Alexa Fluor 568, Alexa Fluor 594, Alexa Fluor 633, Alexa Fluor 660, Alexa Fluor 680), Cascade Blue, Cascade Yellow, and R-phycoerythrin (PE) (Molecular ELISA). Probes, Eugene, OR), FITC, rhodamine and Texas Red (Pierce, Rockford, IL), Cy5, Cy5.5, Cy7 (Amersham Life Science, Pittsburgh, PA). Suitable optical dyes (including fluorophores) are described in the Molecular Probes Handbook by Richard P. Haugland.

[0248] Suitable protein fluorescent markers also include, but are not limited to, green fluorescent protein, including GFP from Renilla, Ptilosarcus or Aequorea species (Chalfie et al., 1994, Science 263:802-805), EGFP (Clontech Laboratories, Inc., Genbank Accession No. U55762), blue fluorescent protein (BFP, Quantum Biotechnologies, Inc. 1801 de Maisonneuve Blvd. West, 8th Floor, Montreal, Quebec, Canada H3H 1J9; Stauber, 1998, Biotechniques 24:462-471; Heim et al., 1996, Curr. Biol. 6:178-182), enhanced yellow fluorescent protein (EYFP, Clontech Laboratories, Inc.), luciferase (Ichiki et al., 1993, J. Immunol. 150:5408-5417), β-galactosidase (Nolan et al., 1988, Proc. Natl. Acad. Sci. USA 85:2603-2607) and Renilla (WO92 / 15673, WO95 / 07463, W 098 / 14605, WO98 / 26277, WO99 / 49019, U.S. Patent Nos. 5,292,658, 5,418,155, 5,683,888, 5,741,668, 5,777,079, 5,804,387, 5,874,304, 5,876,995, 5,925,558).

[0249] Leucine zipper domains are peptides that promote the oligomerization of proteins in which they are present. Leucine zippers were originally identified in several DNA binding proteins (Landschulz et al., 1988, Science 240: 1759), and have since been found in a variety of different proteins. Known leucine zippers include naturally occurring peptides and derivatives thereof that are dimerized or trimerized. Examples of leucine zipper domains suitable for producing soluble oligomeric proteins are described in PCT application WO 94 / 10308, and leucine zippers derived from lung surfactant protein D (SPD) are described in Hoppe et al., 1994, FEBS Letters 344: 191. The use of modified leucine zippers to allow stable trimerization of heterologous proteins fused thereto is described in Fanslow et al., 1994, Semin. Immunol. 6: 267-78. In one method, a recombinant fusion protein comprising a DLL3 antibody fragment or derivative fused to a leucine zipper peptide is expressed in a suitable host cell, and the formed soluble oligomeric DLL3 antibody fragment or derivative is recovered from the culture supernatant.

[0250] The antibody construct of the present invention may also include other domains, which, for example, contribute to the separation of molecules or are related to the adapted pharmacokinetic properties of the molecules. The domain that contributes to the separation of antibody constructs may be selected from a peptide motif or a secondary introduced part, which may be captured in a separation method (e.g., a separation column). The non-limiting embodiments of such other domains include peptide motifs, referred to as Myc tags, HAT tags, HA tags, TAP tags, GST tags, chitin binding domains (CBD tags), maltose binding proteins (MBP tags), Flag tags, Strep tags and variants thereof (e.g., StrepII tags) and His tags. All antibody constructs disclosed herein characterized by the identified CDRs preferably include a His tag domain, which is generally referred to as a repeating sequence of continuous His residues in the amino acid sequence of a molecule, preferably five and more preferably six His residues (six histidines). The His tag may be located at, for example, the N-terminal or C-terminal of the antibody construct, and it is preferably located at the C-terminal. Most preferably, the six histidine tags (HHHHHH) are connected to the C-terminal of the antibody construct according to the present invention via a peptide bond.

[0251] The first binding domain of the antibody construct of the invention binds to human DLL3 on the surface of a target cell. A preferred amino acid sequence of human DLL3 is represented by SEQ ID NO: 252. It should be understood that in the context of the present invention, the term "on the surface" means that the binding domain specifically binds to an epitope contained within the extracellular domain of DLL3 (DLL3 ECD). Thus, when the first binding domain according to the invention is expressed by a naturally expressing cell or cell line and / or a cell or cell line transformed or (stably / transiently) transfected with DLL3, it preferably binds to DLL3. In a preferred embodiment, the first binding domain also binds to DLL3 when DLL3 is used as a "target" or "ligand" molecule in an in vitro binding assay such as BIAcore or Scatchard. A "target cell" may be any prokaryotic or eukaryotic cell expressing DLL3 on its surface; preferably, the target cell is a cell that is part of the human or animal body, such as a specific cancer or tumor cell expressing DLL3.

[0252] The term "DLL3 ECD" refers to a form of DLL3 that is substantially free of the transmembrane and cytoplasmic domains of DLL3. The skilled artisan will appreciate that the transmembrane domain identified for the DLL3 polypeptides of the invention is identified according to criteria routinely used in the art to identify hydrophobic domain types. The precise boundaries of the transmembrane domain may vary, but most likely will vary by no more than about 5 amino acids at either end of the domains specifically mentioned herein. A preferred human DLL3 ECD is shown in SEQ ID NO:253.

[0253] The affinity of the first binding domain of human DLL3 is preferably ≤20 nM, more preferably ≤10 nM, even more preferably ≤5 nM, even more preferably ≤2 nM, even more preferably ≤1 nM, even more preferably ≤0.6 nM, even more preferably ≤0.5 nM, and most preferably ≤0.4 nM. Affinity can be measured, for example, in a BIAcore assay or in a Scatchard assay, for example, as described in the Examples. Other methods of determining affinity are also well known to the skilled person.

[0254] T cells or T lymphocytes are a type of lymphocyte (itself a type of white blood cell) that plays a central role in cell-mediated immunity. There are several T cell subsets, each with different functions. T cells can be distinguished from other lymphocytes, such as B cells and NK cells, by the presence of a T cell receptor (TCR) on the cell surface. The TCR is responsible for recognizing antigens bound to major histocompatibility complex (MHC) molecules and consists of two different protein chains. In 95% of T cells, the TCR consists of an alpha (α) and beta (β) chain. When the TCR engages with an antigenic peptide and MHC (peptide / MHC complex), the T lymphocyte is activated through a series of biochemical events mediated by associated enzymes, co-receptors, special adapter molecules, and activated or released transcription factors.

[0255] The CD3 receptor complex is a protein complex and is composed of four chains. In mammals, the complex contains CD3γ (gamma) chain, CD3δ (delta) chain and two CD3ε (epsilon) chains. These chains are associated with T cell receptor (TCR) and so-called ζ (zeta) chain to form T cell receptor CD3 complex and generate activation signals in T lymphocytes. CD3γ (gamma), CD3δ (delta) and CD3ε (epsilon) chains are highly related cell surface proteins of the immunoglobulin superfamily containing a single extracellular immunoglobulin domain. The intracellular tail of the CD3 molecule contains a single conservative motif called an activation motif based on immunoreceptor tyrosine or ITAM, which is crucial to the signal transduction ability of TCR. The CD3ε molecule is a polypeptide encoded by the CD3E gene on chromosome 11 in humans. The most preferred epitope of CD3ε is contained in amino acid residues 1-27 of the human CD3ε extracellular domain.

[0256] Recruiting T cells via multispecific, at least bispecific antibody constructs to redirect target cell lysis involves the formation of cytolytic synapses and the delivery of perforins and granzymes. Engaged T cells are capable of continuous target cell lysis and are not affected by immune escape mechanisms that interfere with peptide antigen processing and presentation or clonal T cell differentiation; see, e.g., WO2007 / 042261.

[0257] Cytotoxicity mediated by the DLL3xCD3 bispecific antibody construct can be measured in various ways. See Examples 8.1 to 8.7. Effector cells may be, for example, stimulated enriched (human) CD8-positive T cells or unstimulated (human) peripheral blood mononuclear cells (PBMCs). If the target cells are of macaque origin or express macaque DLL3 or are transfected with macaque DLL3, the effector cells should also be of macaque origin, such as a macaque T cell line, such as 4119LnPx. The target cells should express DLL3 (at least its extracellular domain), such as human or macaque DLL3. The target cells may be cell lines (such as CHO) stably or transiently transfected with DLL3 (e.g., human or macaque DLL3). Alternatively, the target cells may be DLL3-positive naturally expressing cell lines, such as the human lung carcinoma cell line SHP-77. Generally, target cell lines expressing higher levels of DLL3 on the cell surface are expected to have lower EC50 values. The effector cell: target cell (E: T) ratio is typically about 10: 1, but may vary. The cytotoxic activity of the DLL3xCD3 bispecific antibody construct can be measured in a 51-chromium release assay (incubation time of about 18 hours) or a FACS-based cytotoxicity assay (incubation time of about 48 hours). It is also possible to modify the incubation time (cytotoxic response) of the assay. Other methods of measuring cytotoxicity are well known to the skilled person and include MTT or MTS assays, ATP-based assays (including bioluminescent assays), sulforhodamine B (SRB) assays, WST assays, clonogenic assays, and ECIS techniques.

[0258] The cytotoxic activity mediated by the DLL3xCD3 bispecific antibody constructs of the invention is preferably measured in a cell-based cytotoxicity assay. It can also be measured in a 51-chromium release assay. 50 The value corresponds to the half-maximal effective concentration (the concentration of the antibody construct that induces a cytotoxic response that is halfway between the baseline and the maximum value). Preferably, the EC of the DLL3xCD3 bispecific antibody construct is 50 Values ​​are ≤5000pM or ≤4000pM, more preferably ≤3000pM or ≤2000pM, even more preferably ≤1000pM or ≤500pM, even more preferably ≤400pM or ≤300pM, even more preferably ≤200pM, even more preferably ≤100pM, even more preferably ≤50pM, even more preferably ≤20pM or ≤10pM, and most preferably ≤5pM.

[0259] The EC given above 50The EC50 values ​​can be measured in different assays. One skilled in the art will appreciate that it is expected that EC50 values ​​may be lower when stimulated / enriched CD8+ T cells are used as effector cells compared to unstimulated PBMCs. In addition, it may be expected that EC50 values ​​may be lower when target cells express a large amount of target antigen compared to low target expression rats. For example, when stimulated / enriched human CD8+ T cells are used as effector cells (and DLL3-transfected cells such as CHO cells or DLL3-positive human lung cancer cell line SHP-77 are used as target cells), the EC50 values ​​of the DLL3xCD3 bispecific antibody construct may be lower than those of the DLL3xCD3 bispecific antibody construct. 50 The value is preferably ≤1000 pM, more preferably ≤500 pM, even more preferably ≤250 pM, even more preferably ≤100 pM, even more preferably ≤50 pM, even more preferably ≤10 pM, and most preferably ≤5 pM. When human PBMCs are used as effector cells, the EC of the DLL3xCD3 bispecific antibody construct 50 The value is preferably ≤5000 pM or ≤4000 pM (particularly when the target cell is a DLL3-positive human lung cancer cell line SHP-77), more preferably ≤2000 pM (particularly when the target cell is a DLL3-transfected cell, such as a CHO cell), more preferably ≤1000 pM or ≤500 pM, even more preferably ≤200 pM, even more preferably ≤150 pM, even more preferably ≤100 pM, and most preferably ≤50 pM or lower. When a macaque T cell line such as LnPx4119 is used as an effector cell and a macaque DLL3-transfected cell line such as a CHO cell is used as a target cell line, the EC value of the DLL3xCD3 bispecific antibody construct is preferably ≤5000 pM or ≤4000 pM (particularly when the target cell is a DLL3-positive human lung cancer cell line SHP-77), more preferably ≤2000 pM (particularly when the target cell is a DLL3-transfected cell, such as a CHO cell), more preferably ≤1000 pM or ≤500 pM, even more preferably ≤200 pM, even more preferably ≤150 pM, even more preferably ≤100 pM, and most preferably ≤50 pM or lower. 50 The value is preferably ≤2000 pM or ≤1500 pM, more preferably ≤1000 pM or ≤500 pM, even more preferably ≤300 pM or ≤250 pM, even more preferably ≤100 pM, and most preferably ≤50 pM.

[0260] Preferably, the DLL3xCD3 bispecific antibody construct of the invention does not induce / mediate or does not substantially induce / mediate lysis of DLL3 negative cells such as CHO cells. The terms "does not induce lysis", "does not substantially induce lysis", "does not mediate lysis" or "does not substantially mediate lysis" mean that the antibody construct of the invention does not induce or mediate more than 30%, preferably no more than 20%, more preferably no more than 10%, particularly preferably no more than 9%, 8%, 7%, 6% or 5% lysis of DLL3 negative cells, wherein lysis of the DLL3 positive human lung cancer cell line SHP-77 (see above) is set as 100%. This generally applies to antibody construct concentrations of up to 500 nM. The skilled person knows how to measure cell lysis without further ado. Furthermore, the specification teaches specific instructions on how to measure cell lysis.

[0261] The difference in cytotoxic activity between the monomeric and dimeric isoforms of each DLL3xCD3 bispecific antibody construct is referred to as the "potency gap". This potency gap can be measured, for example, in terms of the EC of the monomeric and dimeric forms of the molecule. 50 For calculation of the ratio between the values, see Example 15. The potency difference of the DLL3xCD3 bispecific antibody constructs of the invention is preferably ≤5, more preferably ≤4, even more preferably ≤3, even more preferably ≤2 and most preferably ≤1.

[0262] The first and / or second (or any further) binding domain of the antibody construct of the invention preferably has cross-species specificity for members of the class Mammalian of the primates. Cross-species specific CD3 binding domains are described, for example, in WO 2008 / 119567. According to one embodiment, the first and / or second binding domain will bind to DLL3 / CD3 of primates, including (but not limited to) New World primates (such as common marmosets (Callithrix jacchus), cotton-top tamarins (Saguinus Oedipus) or squirrel monkeys (Saimiri sciureus)), Old World primates (such as baboons and macaques), gibbons, orangutans and non-human homininae, in addition to binding to human DLL3 and human CD3, respectively. It is envisaged that the first binding domain of the antibody construct of the invention (which binds to human DLL3 on the surface of target cells) also binds to at least macaque DLL3, and / or the second binding domain (which binds to human CD3 on the surface of T cells) also binds to at least macaque CD3. The preferred macaque is cynomolgus monkey (Macaca fascicularis). Rhesus monkeys (Macaca mulatta, Rhesus) are also envisaged.

[0263] Preferred bispecific antibody constructs of the invention comprise a first binding domain that binds to human DLL3 on the surface of a target cell and a second binding domain that binds to human CD3 and at least macaque CD3 on the surface of a T cell. In one aspect of this embodiment, the first binding domain binds to an epitope of DLL3 contained within the region depicted as SEQ ID NO: 260.

[0264] In one aspect of the invention, the first binding domain binds to human DLL3 and further binds to macaque DLL3 (such as cynomolgus monkey DLL3), and more preferably binds to macaque DLL3 ECD. A preferred cynomolgus monkey DLL3 is depicted in SEQ ID NO: 271. A preferred macaque DLL3 ECD is depicted in SEQ ID NO: 272. The affinity of the first binding domain of macaque DLL3 is preferably ≤15nM, more preferably ≤10nM, even more preferably ≤5nM, even more preferably ≤1nM, even more preferably ≤0.5nM, even more preferably ≤0.1nM, and most preferably ≤0.05nM or even ≤0.01nM.

[0265] Preferably, the antibody constructs according to the invention bind to macaque DLL3 with a difference in affinity compared to human DLL3 [maDLL3:hu DLL3] (as determined, for example, by BiaCore or by Scatchard analysis) preferably between 0.1 and 10, more preferably between 0.2 and 5, even more preferably between 0.3 and 4, even more preferably between 0.5 and 3 or between 0.5 and 2.5, and most preferably between 0.5 and 2 or between 0.6 and 2. See Examples 3 and 4.

[0266] In one embodiment of the antibody construct of the present invention, the second binding domain binds to human CD3ε and common marmoset, cotton-top tamarin or squirrel monkey CD3ε. Preferably, the second binding domain binds to extracellular epitopes of these CD3ε chains. It is also envisioned that the second binding domain binds to extracellular epitopes of human and rhesus CD3ε chains. The most preferred epitope of CD3ε is contained within amino acid residues 1-27 of the extracellular domain of human CD3ε. Even more specifically, the epitope comprises at least the amino acid sequence Gln-Asp-Gly-Asn-Glu. Common marmosets and cotton-top tamarins are both New World primates belonging to the family Callitrichidae, while squirrel monkeys are New World primates belonging to the family Cebidae.

[0267] It is particularly preferred for the antibody construct of the present invention that the second binding domain that binds to human CD3 on the surface of T cells comprises a VL region comprising CDR-L1, CDR-L2 and CDR-L3 selected from the group consisting of:

[0268] (a) CDR-L1 as depicted in SEQ ID NO:27 of WO 2008 / 119567, CDR-L2 as depicted in SEQ ID NO:28 of WO 2008 / 119567 and CDR-L3 as depicted in SEQ ID NO:29 of WO 2008 / 119567;

[0269] (b) CDR-L1 as depicted in SEQ ID NO: 117 of WO 2008 / 119567, CDR-L2 as depicted in SEQ ID NO: 118 of WO 2008 / 119567, and CDR-L3 as depicted in SEQ ID NO: 119 of WO 2008 / 119567; and

[0270] (c) CDR-L1 as depicted in SEQ ID NO: 153 of WO 2008 / 119567, CDR-L2 as depicted in SEQ ID NO: 154 of WO 2008 / 119567 and CDR-L3 as depicted in SEQ ID NO: 155 of WO 2008 / 119567.

[0271] In an alternative preferred embodiment of the antibody construct of the present invention, the second binding domain that binds to human CD3 on the surface of T cells comprises a VH region comprising CDR-H1, CDR-H2 and CDR-H3 selected from the group consisting of:

[0272] (a) CDR-H1 as depicted in SEQ ID NO: 12 of WO 2008 / 119567, CDR-H2 as depicted in SEQ ID NO: 13 of WO 2008 / 119567, and CDR-H3 as depicted in SEQ ID NO: 14 of WO 2008 / 119567;

[0273] (b) CDR-H1 as depicted in SEQ ID NO:30 of WO 2008 / 119567, CDR-H2 as depicted in SEQ ID NO:31 of WO 2008 / 119567, and CDR-H3 as depicted in SEQ ID NO:32 of WO 2008 / 119567;

[0274] (c) CDR-H1 as depicted in SEQ ID NO:48 of WO 2008 / 119567, CDR-H2 as depicted in SEQ ID NO:49 of WO 2008 / 119567, and CDR-H3 as depicted in SEQ ID NO:50 of WO 2008 / 119567;

[0275] (d) CDR-H1 as depicted in SEQ ID NO:66 of WO 2008 / 119567, CDR-H2 as depicted in SEQ ID NO:67 of WO 2008 / 119567, and CDR-H3 as depicted in SEQ ID NO:68 of WO 2008 / 119567;

[0276] (e) CDR-H1 as depicted in SEQ ID NO:84 of WO 2008 / 119567, CDR-H2 as depicted in SEQ ID NO:85 of WO 2008 / 119567, and CDR-H3 as depicted in SEQ ID NO:86 of WO 2008 / 119567;

[0277] (f) CDR-H1 as depicted in SEQ ID NO: 102 of WO 2008 / 119567, CDR-H2 as depicted in SEQ ID NO: 103 of WO 2008 / 119567, and CDR-H3 as depicted in SEQ ID NO: 104 of WO 2008 / 119567;

[0278] (g) CDR-H1 as depicted in SEQ ID NO: 120 of WO 2008 / 119567, CDR-H2 as depicted in SEQ ID NO: 121 of WO 2008 / 119567, and CDR-H3 as depicted in SEQ ID NO: 122 of WO 2008 / 119567;

[0279] (h) CDR-H1 as depicted in SEQ ID NO: 138 of WO 2008 / 119567, CDR-H2 as depicted in SEQ ID NO: 139 of WO 2008 / 119567, and CDR-H3 as depicted in SEQ ID NO: 140 of WO 2008 / 119567;

[0280] (i) CDR-H1 as depicted in SEQ ID NO: 156 of WO 2008 / 119567, CDR-H2 as depicted in SEQ ID NO: 157 of WO 2008 / 119567, and CDR-H3 as depicted in SEQ ID NO: 158 of WO 2008 / 119567; and

[0281] (j) CDR-H1 as depicted in SEQ ID NO: 174 of WO 2008 / 119567, CDR-H2 as depicted in SEQ ID NO: 175 of WO 2008 / 119567 and CDR-H3 as depicted in SEQ ID NO: 176 of WO 2008 / 119567.

[0282] More preferably for the antibody construct of the present invention, the second binding domain that binds to human CD3 on the surface of T cells comprises a VL region selected from the group consisting of the VL regions depicted in SEQ ID NO: 35, 39, 125, 129, 161 or 165 of WO 2008 / 119567.

[0283] Alternatively, preferably, the second binding domain that binds to human CD3 on the surface of T cells comprises a VH region selected from the following: a VH region as depicted in SEQ ID NO: 15, 19, 33, 37, 51, 55, 69, 73, 87, 91, 105, 109, 123, 127, 141, 145, 159, 163, 177 or 181 of WO 2008 / 119567.

[0284] More preferably, the antibody construct of the present invention is characterized by a second binding domain that binds to human CD3 on the surface of T cells, the second binding domain comprising a VL region and a VH region selected from the group consisting of:

[0285] (a) a VL region as depicted by SEQ ID NO: 17 or 21 of WO 2008 / 119567 and a VH region as depicted by SEQ ID NO: 15 or 19 of WO 2008 / 119567;

[0286] (b) a VL region as depicted by SEQ ID NO: 35 or 39 of WO 2008 / 119567 and a VH region as depicted by SEQ ID NO: 33 or 37 of WO 2008 / 119567;

[0287] (c) a VL region as depicted by SEQ ID NO: 53 or 57 of WO 2008 / 119567 and a VH region as depicted by SEQ ID NO: 51 or 55 of WO 2008 / 119567;

[0288] (d) a VL region as depicted by SEQ ID NO: 71 or 75 of WO 2008 / 119567 and a VH region as depicted by SEQ ID NO: 69 or 73 of WO 2008 / 119567;

[0289] (e) a VL region as depicted by SEQ ID NO: 89 or 93 of WO 2008 / 119567 and a VH region as depicted by SEQ ID NO: 87 or 91 of WO 2008 / 119567;

[0290] (f) a VL region as depicted by SEQ ID NO: 107 or 111 of WO 2008 / 119567 and a VH region as depicted by SEQ ID NO: 105 or 109 of WO 2008 / 119567;

[0291] (g) a VL region as depicted by SEQ ID NO: 125 or 129 of WO 2008 / 119567 and a VH region as depicted by SEQ ID NO: 123 or 127 of WO 2008 / 119567;

[0292] (h) a VL region as depicted by SEQ ID NO: 143 or 147 of WO 2008 / 119567 and a VH region as depicted by SEQ ID NO: 141 or 145 of WO 2008 / 119567;

[0293] (i) a VL region as depicted by SEQ ID NO: 161 or 165 of WO 2008 / 119567 and a VH region as depicted by SEQ ID NO: 159 or 163 of WO 2008 / 119567; and

[0294] (j) a VL region as depicted in SEQ ID NO: 179 or 183 of WO 2008 / 119567 and a VH region as depicted in SEQ ID NO: 177 or 181 of WO 2008 / 119567.

[0295] According to a preferred embodiment of the antibody construct of the present invention, the binding domain and in particular the second binding domain (which binds to human CD3 on the surface of T cells) has the following form: the pairing of the VH region and the VL region is in the form of a single-chain antibody (scFv). The VH and VL regions are arranged in the order of VH-VL or VL-VH. Preferably, the VH region is located at the N-terminus of the linker sequence and the VL region is located at the C-terminus of the linker sequence.

[0296] A preferred embodiment of the above-mentioned antibody construct of the present invention is characterized in that the second binding domain that binds to human CD3 on the surface of T cells comprises an amino acid sequence selected from the group consisting of SEQ ID NO: 23, 25, 41, 43, 59, 61, 77, 79, 95, 97, 113, 115, 131, 133, 149, 151, 167, 169, 185 or 187 of WO 2008 / 119567.

[0297] Thus, in one embodiment, the antibody constructs of the invention comprise a polypeptide selected from the group consisting of the polypeptides depicted in SEQ ID NO: 40, SEQ ID NO: 50, SEQ ID NO: 60, SEQ ID NO: 70, SEQ ID NO: 80, SEQ ID NO: 90, SEQ ID NO: 100, SEQ ID NO: 110, SEQ ID NO: 120, SEQ ID NO: 211, SEQ ID NO: 212, SEQ ID NO: 213, SEQ ID NO: 214, SEQ ID NO: 215, SEQ ID NO: 216, SEQ ID NO: 217, SEQ ID NO: 438, and SEQ ID NO: 532. These antibody constructs have a first binding domain that binds to an epitope of DLL3 contained within the region depicted as SEQ ID NO: 258.

[0298] In an alternative embodiment, the antibody construct of the invention comprises a polypeptide selected from the group consisting of the polypeptides depicted in SEQ ID NO: 130, SEQ ID NO: 140, SEQ ID NO: 150, SEQ ID NO: 160, SEQ ID NO: 170, SEQ ID NO: 218, SEQ ID NO: 219, SEQ ID NO: 220, SEQ ID NO: 494, SEQ ID NO: 495, SEQ ID NO: 496, SEQ ID NO: 497, SEQ ID NO: 498, SEQ ID NO: 499, SEQ ID NO: 500, SEQ ID NO: 501, SEQ ID NO: 502, SEQ ID NO: 503, SEQ ID NO: 504, SEQ ID NO: 505, SEQ ID NO: 506, SEQ ID NO: 507, SEQ ID NO: 508, SEQ ID NO: 509, SEQ ID NO: 510, SEQ ID NO: 511, SEQ ID NO: 512, SEQ ID NO: 513, SEQ ID NO: 514, SEQ ID NO: 515 NO:514, SEQ ID NO:515 and SEQ ID NO:516. These antibody constructs have a first binding domain that binds to an epitope of DLL3 contained within the region depicted as SEQ ID NO:259.

[0299] Amino acid sequence modifications of the antibody constructs described herein are also contemplated. For example, it may be desirable to improve the binding affinity and / or other biological properties of the antibody construct. Amino acid sequence variants of the antibody construct are prepared by introducing appropriate nucleotide changes into the antibody construct nucleic acid or by peptide synthesis. All of the following amino acid sequence modifications should result in an antibody construct that still retains the desired biological activity (binding to DLL3 and binding to CD3) of the unmodified parent molecule.

[0300] The term "amino acid" or "amino acid residue" generally refers to an amino acid having its art-recognized definition, such as an amino acid selected from the group consisting of alanine (Ala or A); arginine (Arg or R); asparagine (Asn or N); aspartic acid (Asp or D); cysteine ​​(Cys or C); glutamine (Gln or Q); glutamic acid (Glu or E); glycine (Gly or G); histidine (His or H); isoleucine (He or I); leucine (Leu or L); lysine (Lys or K); methionine (Met or M); phenylalanine (Phe or F); proline (Pro or P); serine (Ser or S); threonine (Thr or T); tryptophan (Trp or W); tyrosine (Tyr or Y); and valine (Val or V), although modified, synthetic, or rare amino acids may be used as desired. In general, amino acids can be grouped as having nonpolar side chains (e.g., Ala, Cys, He, Leu, Met, Phe, Pro, Val), negatively charged side chains (e.g., Asp, Glu), positively charged side chains (e.g., Arg, His, Lys), or uncharged polar side chains (e.g., Asn, Cys, Gln, Gly, His, Met, Phe, Ser, Thr, Trp, and Tyr).

[0301] Amino acid modifications include, for example, deletions and / or insertions and / or substitutions of residues within the amino acid sequence of the antibody construct. Any combination of deletions, insertions and substitutions is performed to obtain the final construct, as long as the final construct has the desired characteristics. Amino acid changes may also alter post-translational processes of the antibody construct, such as changing the number or position of glycosylation sites.

[0302] For example, 1, 2, 3, 4, 5 or 6 amino acids may be inserted or deleted in each of the CDRs (depending, of course, on their length), while 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20 or 25 amino acids may be inserted or deleted in each of the FRs. Preferably, amino acid sequence insertions include amino and / or carboxyl terminal fusions of 1, 2, 3, 4, 5, 6, 7, 8, 9 or 10 residues in length inserted into polypeptides containing one hundred or more residues, as well as intrasequence insertions with single or multiple amino acid residues. Insertion variants of the antibody constructs of the present invention include fusions to the N-terminus or C-terminus of the antibody construct of an enzyme or fusions to a polypeptide that prolongs the serum half-life of the antibody construct.

[0303] The most concerned sites of substitution mutagenesis include CDRs of heavy and / or light chains, especially hypervariable regions, but also encompass FR changes in heavy and / or light chains. The substitution is preferably a conservative substitution as described herein. Preferably, depending on the length of the CDR or FR, 1, 2, 3, 4, 5, 6, 7, 8, 9 or 10 amino acids in the CDR can be substituted, while 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20 or 25 amino acids in the framework region (FR) can be substituted. For example, if the CDR sequence encompasses 6 amino acids, it is conceivable to replace one, two or three of these amino acids. Similarly, if the CDR sequence encompasses 15 amino acids, it is conceivable to replace one, two, three, four, five or six of these amino acids.

[0304] A useful method for identifying certain residues or regions in an antibody construct as preferred locations for mutagenesis is called "alanine scanning mutagenesis," as described by Cunningham and Wells, Science, 244: 1081-1085 (1989). Here, a residue or group of target residues (e.g., charged residues such as arg, asp, his, lys, and glu) within an antibody construct is identified and replaced with a neutral or negatively charged amino acid (most preferably alanine or polyalanine) to affect the interaction of the amino acid with the epitope.

[0305] Subsequently, by introducing additional or other variants at or for the substitution site to optimize those amino acid positions that show functional sensitivity to the substitution. Therefore, when the site or region for introducing amino acid sequence changes is predetermined, the nature of the mutation itself does not need to be predetermined. For example, to analyze or optimize the performance of the mutation at a given site, alanine scanning or random mutagenesis can be performed at the target codon or region, and the expressed antibody construct variants are screened for the optimal combination of desired activity. The technology of performing substitution mutations at predetermined sites of DNA with known sequence is well known, such as M13 primer mutagenesis and PCR mutagenesis. Screening of mutants is performed using an antigen binding activity assay (such as DLL3 or CD3 binding).

[0306] In general, if an amino acid is substituted in one or more or all CDRs of the heavy and / or light chain, it is preferred that the subsequently obtained "substituted" sequence has at least 60% or 65%, more preferably 70% or 75%, even more preferably 80% or 85% and especially preferably 90% or 95% identity with the "original" CDR sequence. This means that the degree of identity with the "substituted" sequence depends on the length of the CDR. For example, a CDR with 5 amino acids preferably has 80% identity with its substituted sequence such that at least one amino acid is substituted. Thus, the CDRs of an antibody construct may have different degrees of identity with their substituted sequences, for example CDRL1 may have 80% and CDRL3 may have 90%.

[0307] Preferred substitutions (or replacements) are conservative substitutions. However, any substitutions (including non-conservative substitutions or one or more of the "exemplary substitutions" listed in Table 1 below) are contemplated, as long as the antibody construct retains its ability to bind to DLL3 via the first binding domain and to CD3 or CD3ε via the second binding domain and / or its CDRs are identical to the subsequently substituted sequences (at least 60% or 65%, more preferably 70% or 75%, even more preferably 80% or 85% and particularly preferably 90% or 95% identical to the "original" CDR sequences).

[0308] Conservative substitutions are shown under the heading "Preferred Substitutions" in Table 1. If such substitutions result in a change in biological activity, then more substantial changes, denominated "exemplary substitutions" in Table 1, or as further described below when referring to amino acid classes, may be introduced and the products screened for the desired characteristics.

[0309] Table 1: Amino Acid Substitutions

[0310] original Exemplary substitutions Preferred substitutions Ala(A) val, leu, ile val Arg(R) lys, gln, asn lys Asn(N) gln, his, asp, lys, arg gln Asp(D) glu、asn glu Cys(C) ser、ala ser Gln(Q) asn、glu asn Glu(E) asp、gln asp Gly(G) Ala ala His(H) asn, gln, lys, arg arg Ile(I) leu, val, met, ala, phe leu Leu(L) norleucine, ile, val, met, ala ile Lys(K) arg, gln, asn arg Met(M) leu, phe, ile leu Phe(F) leu, val, ile, ala, tyr tyr Pro(P) Ala ala Ser(S) Thr thr Thr(T) Ser ser Trp(W) tyr、phe tyr Tyr(Y) trp, phe, thr, ser phe Val(V) ile, leu, met, phe, ala leu

[0311] Substantial changes in the biological properties of the antibody constructs of the invention can be achieved by selecting substitutions that differ significantly in their effect on maintaining (a) the structure of the polypeptide backbone in the area of ​​the substitution, e.g., a folded or helical conformation, (b) the charge or hydrophobicity of the molecule at the target site, or (c) the bulk of the side chain. Naturally occurring residues can be grouped according to common side chain properties: (1) hydrophobic: norleucine, met, ala, val, leu, ile; (2) neutral hydrophilic: cys, ser, thr; (3) acidic: asp, glu; (4) basic: asn, gin, his, lys, arg; (5) residues affecting chain orientation: gly, pro; and (6) aromatic: trp, tyr, phe.

[0312] Non-conservative substitutions entail exchanging a member of one of these classes for another class. Any cysteine ​​residue that is not involved in maintaining the correct conformation of the antibody construct may generally be substituted with serine to increase the oxidative stability of the molecule and prevent abnormal cross-linking. Conversely, cysteine ​​bonds may be added to the antibody to improve its stability (especially when the antibody is an antibody fragment such as an Fv fragment).

[0313] For amino acid sequences, sequence identity and / or similarity are determined using standard techniques known in the art, including, but not limited to, the local sequence identity algorithm of Smith and Waterman, 1981, Adv. Appl. Math. 2:482; the sequence identity alignment algorithm of Needleman and Wunsch, 1970, J. Mol. Biol. 48:443; the similarity search method of Pearson and Lipman, 1988, Proc. Nat. Acad. Sci. USA 85:2444; computerized implementations of these algorithms (GAP, BESTFIT, FASTA, and TFASTA in the Wisconsin Genetics Software Package, Genetics Computer Group, 575 Science Drive, Madison, Wis.); the best fit sequence program described by Devereux et al., 1984, Nucl. Acid Res. 12:387-395, preferably using the default settings or by inspection. Preferably, the percent identity is calculated by FastDB based on the following parameters: mismatch penalty 1; gap penalty 1; gap size penalty 0.33; and ligation penalty 30, "Current Methods in Sequence Comparison and Analysis," Macromolecule Sequencing and Synthesis, Selected Methods and Applications, pp. 127-149 (1988), Alan R. Liss, Inc.

[0314] An example of an available algorithm is PILEUP. PILEUP uses progressive pairwise alignments to generate a multiple sequence alignment from a group of related sequences. It can also draw a tree diagram showing the clustering relationships used to generate the alignment. PILEUP uses a simplified method of the progressive alignment method of Feng and Doolittle, 1987, J. Mol. Evol. 35: 351-360; this method is similar to the method described by Higgins and Sharp, 1989, CABIOS 5: 151-153. Available PILEUP parameters include a default gap weight of 3.00, a default gap length weight of 0.10, and a weighted end gap.

[0315] Another example of an algorithm that can be used is the BLAST algorithm, described in: Altschul et al., 1990, J. Mol. Biol. 215:403-410; Altschul et al., 1997, Nucleic Aci ds Res. 25:3389-3402; and Karin et al., 1993, Proc. Natl. Acad. Sci. USA 90:5873-5787. A particularly useful BLAST program is the WU-BLAST-2 program, which is obtained from Altschul et al., 1996, Methods in Enzymology 266:460-480. WU-BLAST-2 uses several search parameters, most of which are set to default values. The adjustable parameters are set using the following values: overlap interval = 1, overlap fraction = 0.125, word threshold (T) = II. The HSP S and HSP S2 parameters are dynamic values ​​and are established by the program itself depending on the composition of the specific sequence and the composition of the specific database being searched for the sequence of interest; however, the values ​​can be adjusted to increase sensitivity.

[0316] Another algorithm that can be used is gapped BLAST reported by Altschul et al., 1993, Nucl. Acids Res. 25:3389-3402. Gapped BLAST uses BLOSUM-62 replacement scoring; the threshold T parameter is set to 9; the two-hit method is used to trigger ungapped extensions, charge gap length k, at a cost of 10+k; Xu is set to 16, and Xg is set to 40 for the database search phase and 67 for the algorithm output phase. Gap alignments are triggered by a score corresponding to about 22 bits.

[0317] Typically, the amino acid homology, similarity or identity between each variant CDR is at least 60% of the sequence described herein, and more typically at least 65% or 70%, more preferably at least 75% or 80%, even more preferably at least 85%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, 99% and almost 100% preferably increasing homology or identity. In a similar manner, the "percentage (%) of nucleic acid sequence identity" relative to the protein-binding nucleotide sequence identified herein is defined as the percentage of nucleotide residues in the candidate sequence that are identical to the nucleotide residues in the coding sequence of the antibody construct. The specific method uses the BLASTN module of WU-BLAST-2 set to default parameters, wherein the overlap interval and overlap fraction are set to 1 and 0.125, respectively.

[0318] Typically, the nucleic acid sequence homology, similarity or identity between the nucleotide sequence encoding each variant CDR and the nucleotide sequence depicted herein is at least 60%, and more typically has at least 65%, 70%, 75%, 80%, 81%, 82%, 83%, 84%, 85%, 86%, 87%, 88%, 89%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98% or 99% and almost 100% homology or identity. Thus, a "variant CDR" is a CDR having a specified homology, similarity or identity to a parent CDR of the present invention, and sharing the biological function of the parent CDR, including but not limited to at least 60%, 65%, 70%, 75%, 80%, 81%, 82%, 83%, 84%, 85%, 86%, 87%, 88%, 89%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98% or 99% of the specificity and / or activity.

[0319] In one embodiment, the percent identity with the human germline of the antibody construct according to the invention is ≥70% or ≥75%, more preferably ≥80% or ≥85%, even more preferably ≥90%, and most preferably ≥91%, ≥92%, ≥93%, ≥94%, ≥95% or even ≥96%. See Example 7. The identity with the human antibody germline gene product is considered to be an important feature for reducing the risk of therapeutic proteins causing an immune response to the drug in patients during treatment. Hwang & Foote ("Immunogenicity of engineered antibodies"; Methods 36 (2005) 3-10) demonstrated that the reduction of the non-human part of the drug antibody construct leads to a reduced risk of inducing anti-drug antibodies in patients during treatment. By comparing a large number of clinically evaluated antibody drugs and corresponding immunogenicity data, it is shown that humanizing the V region of the antibody makes the immunogenicity of the protein lower (average 5.1% of patients) than antibodies with unchanged non-human V regions (average 23.59% of patients). Therefore, a high degree of identity with human sequences is desirable for V-region based protein therapeutics in the form of antibody constructs. For this purpose of determining germline identity, the V region of VL can be aligned with the amino acid sequences of human germline V segments and J segments using Vector NTI software (http: / / vbase.mrc-cpe.cam.ac.uk / ), and the amino acid sequence can be calculated as a percentage by dividing the number of identical amino acid residues by the total number of amino acid residues of VL. This calculation can also be performed on the VH segment (http: / / vbase.mrc-cpe.cam.ac.uk / ), except that VH CDR3 may be excluded due to the high diversity of VH CDR3 and the lack of existing human germline VH CDR3 alignment partners. Recombinant techniques can then be used to improve sequence identity with human antibody germline genes.

[0320] In another embodiment, the bispecific antibody constructs of the invention exhibit high monomer yields under standard research scale conditions (e.g., in a standard two-step purification process). Preferably, the monomer yield of the antibody constructs according to the invention is ≥0.25 mg / L supernatant, more preferably ≥0.5 mg / L, even more preferably ≥1 mg / L, and most preferably ≥3 mg / L supernatant.

[0321] Likewise, the yield of dimeric antibody construct isotypes and thus the monomer percentage of the antibody construct can be determined (i.e., monomer: (monomer + dimer)). The productivity of monomers and dimeric antibody constructs and the calculated monomer percentage can be obtained, for example, in the SEC purification step of the culture supernatant prepared at a standardized research scale in a roller bottle. In one embodiment, the monomer percentage of the antibody construct is ≥80%, more preferably ≥85%, even more preferably ≥90%, and most preferably ≥95%.

[0322] In one embodiment, the antibody construct has a preferred plasma stability (ratio of EC50 with plasma to EC50 without plasma) of ≤5 or ≤4, more preferably ≤3.5 or ≤3, even more preferably ≤2.5 or ≤2, and most preferably ≤1.5 or ≤1. The plasma stability of the antibody construct can be tested by incubating the construct in human plasma at 37°C for 24 hours and then measuring the EC50 in a 51-chromium release cytotoxicity assay. The effector cells in the cytotoxicity assay can be stimulated enriched human CD8 positive T cells. The target cells can be, for example, CHO cells transfected with human DLL3. The effector cell: target cell (E:T) ratio can be selected to be 10:1. The human plasma pool for this purpose is derived from the blood of healthy donors collected by syringes coated with EDTA. Cellular components are removed by centrifugation, and the upper plasma phase is collected and then pooled. As a control, the antibody construct is diluted and then immediately subjected to a cytotoxicity assay in RPMI-1640 medium. Plasma stability was calculated as the ratio of EC50 (after plasma incubation) to EC50 (control). See Example 11.

[0323] It is also preferred that the conversion of monomers of the antibody constructs of the present invention to dimers is low. The conversion can be measured under different conditions and analyzed by high performance size exclusion chromatography. See Example 9. For example, incubation of monomer isoforms of antibody constructs can be performed at 37°C for 7 days and in an incubator to a concentration of, for example, 100 μg / ml or 250 μg / ml. Under these conditions, it is preferred that the antibody constructs of the present invention show a dimer percentage of ≤5%, more preferably ≤4%, even more preferably ≤3%, even more preferably ≤2.5%, even more preferably ≤2%, even more preferably ≤1.5%, and most preferably ≤1% or ≤0.5%, or even 0%.

[0324] It is also preferred that the bispecific antibody constructs of the present invention present extremely low dimer conversion after multiple freeze / thaw cycles. For example, the antibody construct monomer is adjusted to a concentration of 250 μg / ml in a general formulation buffer and subjected to three freeze / thaw cycles (freezing at -80°C for 30 minutes, then thawing at room temperature for 30 minutes), and then subjected to efficient SEC to determine the percentage of the initial monomer antibody construct that has been converted into a dimeric antibody construct. It is preferred that the dimer percentage of the bispecific antibody construct, for example, after three freeze / thaw cycles is ≤5%, more preferably ≤4%, even more preferably ≤3%, even more preferably ≤2.5%, even more preferably ≤2%, even more preferably ≤1.5%, and most preferably ≤1% or even ≤0.5%.

[0325] The bispecific antibody constructs of the present invention preferably show favorable thermal stability, wherein the aggregation temperature is ≥45°C or ≥50°C, more preferably ≥52°C or ≥54°C, even more preferably ≥56°C or ≥57°C, and most preferably ≥58°C or ≥59°C. The thermal stability parameter can be determined as follows based on the antibody aggregation temperature: the antibody solution at a concentration of 250 μg / ml is transferred to a disposable cuvette and placed in a dynamic light scattering (DLS) device. The sample is heated from 40°C to 70°C at a heating rate of 0.5°C / min, wherein the measured radius is continuously acquired. The aggregation temperature of the antibody is calculated using the increase in radius indicating protein melting and aggregation. See Example 10.

[0326] Alternatively, the temperature melting curve can be determined by differential scanning calorimetry (DSC) to determine the intrinsic biophysical protein stability of the antibody construct. These experiments were performed using a MicroCal LLC (Northampton, MA, USA) VP-DSC device. The energy absorption of the sample containing the antibody construct was recorded from 20°C to 90°C compared to the sample containing only the formulation buffer. The antibody construct was adjusted to a final concentration of 250 μg / ml in, for example, SEC running buffer. To record the corresponding melting curve, the entire sample temperature was gradually increased. At each temperature T, the energy absorption of the sample and the formulation buffer reference was recorded. A graph of the energy absorption difference Cp (kcal / mole / °C) of the sample minus the reference was plotted relative to the corresponding temperature. The melting temperature is defined as the temperature at which the first maximum value of energy absorption occurs.

[0327] It is furthermore envisaged that the DLL3xCD3 bispecific antibody constructs of the invention do not cross-react (i.e. do not substantially bind) with the human DLL3 paralogs DLL1 and / or DLL4. It is furthermore envisaged that the DLL3xCD3 bispecific antibody constructs of the invention do not cross-react (i.e. do not substantially bind) with the macaque / cynomolgus monkey DLL3 paralogs DLL1 and / or DLL4. See Example 6.

[0328] It is also envisaged that the turbidity of the DLL3xCD3 bispecific antibody constructs of the invention (as measured by OD340 after purification of the monomeric antibody constructs concentrated to 2.5 mg / ml and overnight incubation) is ≤ 0.2, preferably ≤ 0.15, more preferably ≤ 0.12, even more preferably ≤ 0.1, and most preferably ≤ 0.08. See Example 12.

[0329] It is further envisaged that the DLL3xCD3 bispecific antibody constructs of the invention are not internalized or not significantly internalized by target cells. The internalization rate can be determined, for example, as described in Example 16. Preferably, the internalization rate (e.g., measured as a reduction in cytotoxicity) is ≤20% after 2 hours of (pre)incubation of the antibody construct with target cells, more preferably ≤15%, even more preferably ≤10% and most preferably ≤5%.

[0330] It is further envisaged that shed or soluble DLL3 does not significantly impair the efficacy or biological activity of the DLL3xCD3 bispecific antibody constructs of the invention. This can be measured, for example, in a cytotoxicity assay, where soluble DLL3 is added to the assay at increasing concentrations, for example, at 0 nM - 0.3 nM - 0.7 nM - 1 nM - 3 nM - 7 nM - 12 nM. The EC50 value of the antibody construct tested should not be significantly increased in the presence of soluble DLL3. See Example 17.

[0331] In another embodiment, antibody construct according to the present invention is stable under acidic pH value.Under non-physiological pH value such as pH 5.5 (pH value required for operation cation exchange chromatography), antibody construct becomes more tolerant, and the recovery rate of antibody construct eluted from ion exchange column relative to the total amount of protein loaded is higher.The antibody construct recovery rate of ion (e.g. cation) exchange column at pH 5.5 is preferably ≥30%, more preferably ≥40%, more preferably ≥50%, even more preferably ≥60%, even more preferably ≥70%, even more preferably ≥80%, even more preferably ≥90%, even more preferably ≥95% and most preferably ≥99%.Referring to Example 13.

[0332] It is further envisaged that the bispecific antibody constructs of the invention exhibit therapeutic efficacy or anti-tumor activity. This can be assessed, for example, in studies as disclosed in the following examples of advanced human tumor xenograft models:

[0333] On day 1 of the study, 5×10 cells of a human DLL3-positive cancer cell line (e.g., SHP-77) were plated. 6 The cells were injected subcutaneously into the right side of the back of female NOD / SCID mice. When the average tumor volume reached about 100 mm 3 When human CD3-positive T cells expanded in vitro were transplanted into mice, approximately 2×10 7The cells were injected into the peritoneal cavity of the animals. The mice in the vehicle control group 1 did not receive effector cells and were used as non-transplanted controls to compare with the vehicle control group 2 (receiving effector cells) to monitor the effect of T cells alone on tumor growth. Antibody treatment was performed when the average tumor volume reached approximately 200 mm 3 The mean tumor size of each treatment group on the day of treatment start should not be statistically different from any other group (analysis of variance). Mice were treated for about 15 to 20 days by intravenous injection of 0.5 mg / kg / day of the DLL3xCD3 bispecific antibody construct. Tumors were measured by caliper during the study and progression was assessed by comparing tumor volumes (TV) between groups. Tumor growth inhibition T / C [%] was determined by calculating TV as T / C% = 100 × (median TV of analysis group) / (median TV of control group 2).

[0334] The skilled person knows how to modify or adjust certain parameters of this study, such as the number of tumor cells injected, the injection site, the number of transplanted human T cells, the amount of bispecific antibody construct to be administered and the schedule, while still achieving meaningful and reproducible results. Preferably, tumor growth inhibition T / C [%] ≤ 70 or ≤ 60, more preferably ≤ 50 or ≤ 40, even more preferably ≤ 30 or ≤ 20 and most preferably ≤ 10 or ≤ 5 or even ≤ 2.5.

[0335] The present invention further provides a polynucleotide / nucleic acid molecule encoding the antibody construct of the present invention.

[0336] Polynucleotide is a biopolymer consisting of 13 or more nucleotide monomers covalently bonded in a chain. DNA (such as cDNA) and RNA (such as mRNA) are examples of polynucleotides with different biological functions. Nucleotide is an organic molecule that serves as a monomer or subunit of a nucleic acid molecule (such as DNA or RNA). Nucleic acid molecules or polynucleotides can be double-stranded and single-stranded, linear and circular. It is preferably contained in a vector, which is preferably contained in a host cell. The host cell can express an antibody construct, for example, after transformation or transfection with a vector or polynucleotide of the present invention. For this purpose, a polynucleotide or nucleic acid molecule is operably connected to a control sequence.

[0337] The genetic code is a set of rules used to translate information encoded in genetic material (nucleic acids) into proteins. Biological decoding in living cells is achieved by ribosomes, which use tRNA molecules to carry amino acids and read three nucleotides of mRNA at a time to connect amino acids in the order specified by mRNA. The code defines how the sequence of these nucleotide triplets (called codons) specifies which amino acid will be added later during protein synthesis. The codons of three nucleotides in the nucleic acid sequence specify a single amino acid, but there are some exceptions. Because the vast majority of genes are encoded with exactly the same code, this specific code is usually referred to as a typical or standard genetic code. Although the genetic code determines the protein sequence of a given coding region, other genomic regions can affect the time and location of these protein production.

[0338] Furthermore, the present invention provides a vector comprising the polynucleotide / nucleic acid molecule of the present invention.

[0339] Vector is a nucleic acid molecule used as a vehicle for transferring (external) genetic material into a cell. The term "vector" encompasses, but is not limited to, plasmids, viruses, cosmids, and artificial chromosomes. In general, engineered vectors include a replication origin, a multiple cloning site, and a selectable marker. The vector itself is typically a nucleotide sequence, typically a DNA sequence, which includes an insertion sequence (transgene) and a larger sequence that serves as the "backbone" of the vector. Modern vectors can encompass additional features in addition to the transgenic insertion sequence and the backbone: promoters, genetic markers, antibiotic resistance, reporter genes, targeting sequences, protein purification tags. A vector referred to as an expression vector (expression construct) is specifically used to express a transgene in a target cell, and typically has a control sequence.

[0340] The term "control sequence" refers to DNA sequences required for expression of an operably linked coding sequence in a particular host organism. Suitable control sequences for prokaryotes include, for example, promoters, optional operator sequences, and ribosome binding sites. Eukaryotic cells are known to utilize promoters, polyadenylation signals, and enhancers.

[0341] A nucleic acid is "operably linked" when it is placed in a functional relationship with another nucleic acid sequence. For example, if the DNA of a presequence or secretory leader is expressed as a preprotein that participates in the secretion of a polypeptide, it is operably linked to the DNA of the polypeptide; if a promoter or enhancer affects the transcription of a coding sequence, it is operably linked to the sequence; or if a ribosome binding site is positioned so as to facilitate translation, it is operably linked to a coding sequence. In general, "operably linked" means that the DNA sequence being linked is continuous and, in the case of a secretory leader, continuous and in reading frame. However, enhancers do not have to be continuous. Connection is achieved by connection at suitable restriction sites. If such sites do not exist, synthetic oligonucleotide linkers or joints are used according to conventional practice.

[0342] "Transfection" is the process of intentionally introducing nucleic acid molecules or polynucleotides (including vectors) into target cells. The term is mainly used for non-viral methods in eukaryotic cells. Transduction is generally used to describe the transfer of viral-mediated nucleic acid molecules or polynucleotides. The transfection of animal cells generally involves opening a transient hole or "hole" in the cell membrane to allow the absorption of substances. Transfection can be carried out using calcium phosphate, by electroporation, by cell extrusion, or by mixing cationic lipids with substances that produce liposomes (which fuse with the cell membrane and deposit their cargo inside).

[0343] The term "transformation" is used to describe the non-viral transfer of nucleic acid molecules or polynucleotides (including vectors) into bacteria and non-animal eukaryotic cells (including plant cells). Therefore, transformation is a genetic change of a bacterial or non-animal eukaryotic cell caused by the direct uptake and subsequent incorporation of exogenous genetic material (nucleic acid molecules) from its environment through the cell membrane. Transformation can be achieved by artificial means. For transformation to occur, the cell or bacterium must be in a competent state, which can occur as a time-limited response to environmental conditions (such as starvation and cell density).

[0344] Furthermore, the present invention provides a host cell transformed or transfected with the polynucleotide / nucleic acid molecule or with the vector of the present invention.

[0345] As used herein, the term "host cell" or "receptor cell" is intended to include any individual cell or cell culture that may be or has been a receptor for a vector, an exogenous nucleic acid molecule, and a polynucleotide encoding an antibody construct of the present invention and / or a receptor for the antibody construct itself. The introduction of the corresponding substance into the cell is carried out by means of methods such as transformation and transfection. The term "host cell" is also intended to include the progeny or possible progeny of a single cell. Because certain modifications may occur in subsequent generations due to natural, accidental or deliberate mutations or due to environmental influences, in fact such progeny may not be exactly the same as the parent cell (in morphology or genome or total DNA complementary sequence), but is still included in the scope of the term as used herein. Suitable host cells include prokaryotic or eukaryotic cells, and also include, but are not limited to, bacteria, yeast cells, fungal cells, plant cells, and animal cells, such as insect cells and mammalian cells, for example, rodents, rats, macaques, or humans.

[0346] Antibody constructs of the present invention can be produced in bacteria. After expression, antibody constructs of the present invention are separated from E. coli cell paste with soluble parts and can be purified by, for example, affinity chromatography and / or size exclusion. Final purification can be carried out similarly to the process of purifying antibodies expressed in, for example, CHO cells.

[0347] In addition to prokaryotes, eukaryotic microorganisms such as filamentous fungi or yeast are also suitable cloning or expression hosts for the antibody constructs of the invention. Among the lower eukaryotic host microorganisms, Saccharomyces cerevisiae or common baker's yeast is most commonly used. However, a variety of other genera, species and strains are generally available and suitable for use herein, such as Schizosaccharomyces pombe; Kluyveromyce hosts, such as K. lactis, K. fragilis (ATCC 12424), K. bulgaricus (ATCC 16045), K. wickeramii (ATCC 24178), K. waltii (ATCC 56500), K. drosophilarum (ATCC 36906), K. thermotolerans and K. marxianus; Yarrowia (EP 402 226); Pichia pastoris (EP 183 226); 070); Candida; Trichoderma reesia (EP 244 234); Neurosporacrassa; Schwanniomyces, such as Schwanniomyces occidentalis; and filamentous fungi, such as Neurospora, Penicillium, Tolypocladium and Aspergillus hosts, such as A. nidulans and A. niger.

[0348] Suitable host cells for expressing the glycosylated antibody constructs of the present invention are derived from multicellular organisms. Examples of invertebrate cells include plants and insect cells. A variety of baculovirus strains and variants have been identified, as well as corresponding permissive insect host cells from hosts such as Spodoptera frugiperda (caterpillars), Aedes aegypti (mosquitoes), Aedes albopictus (mosquitoes), Drosophila melanogaster (fruit flies) and Bombyx mori (Bombyx mori). A variety of virus strains for transfection are publicly available, such as the L-1 variant of Autographa californica NPV and the Bm-5 strain of Bombyx mori NPV, and such viruses can be used herein as viruses according to the present invention, particularly for transfecting Spodoptera frugiperda cells.

[0349] Plant cell cultures of cotton, corn, potato, soybean, petunia, tomato, Arabidopsis, and tobacco can also be used as hosts. Cloning and expression vectors that can be used to produce proteins in plant cell culture are known to those skilled in the art. See, for example, Hiatt et al., Nature (1989) 342: 76-78, Owen et al. (1992) Bio / Technology 10: 790-794, Artsaenko et al. (1995) The Plant J 8: 745-750 and Fecker et al. (1996) Plant Mol Biol 32: 979-986.

[0350] However, vertebrate cells have received the most attention, and propagation of vertebrate cells in culture (tissue culture) has become a routine procedure. Examples of useful mammalian host cell lines are monkey kidney CV1 cell line transformed by SV40 (COS-7, ATCC CRL 1651); human embryonic kidney cell line (293 or 293 cells subcloned for growth in suspension culture, Graham et al., J. Gen Virol. 36:59 (1977)); baby hamster kidney cells (BHK, ATCC CCL 10); Chinese hamster ovary cells / -DHFR (CHO, Urlaub et al., Proc. Natl. Acad. Sci. USA 77:4216 (1980)); mouse Settley cells (TM4, Mather, Biol. Reprod. 23:243-251 (1980)); monkey kidney cells (CVI ATCC CCL 70); African green monkey kidney cells (VERO-76, ATCC CRL 1587); human cervical carcinoma cells (HELA, ATCC CCL 2); canine kidney cells (MDCK, ATCC CCL 34); Buffalo rat liver cells (BRL 3A, ATCC CRL 1442); human lung cells (W138, ATCC CCL 75); human hepatocytes (Hep G2, 1413 8065); mouse mammary tumor (MMT 060562, ATCC CCL5 1); TRI cells (Mather et al., Annals NY Acad. Sci. (1982) 383:44-68); MRC 5 cells; FS4 cells; and human hepatoma cell line (Hep G2).

[0351] In another embodiment, the invention provides a method of producing an antibody construct of the invention, the method comprising culturing a host cell of the invention under conditions such that the antibody construct of the invention is expressed, and recovering the produced antibody construct from the culture.

[0352] As used herein, the term "culture" refers to maintaining cells in vitro in a culture medium under suitable conditions, allowing them to differentiate, grow, proliferate and / or reproduce. The term "expression" includes any step involved in producing the antibody constructs of the present invention, including but not limited to transcription, post-transcriptional modification, translation, post-translational modification and secretion.

[0353] When using recombinant technology, antibody constructs can be produced in cells, in the periplasmic space or directly secreted into the culture medium. If the antibody construct is produced in the cell, then as a first step, the particle fragments of the host cell or the cleaved fragments are removed, for example, by centrifugation or ultrafiltration. Carter et al., Bio / Technology 10: 163-167 (1992) describes the procedure for separating the antibodies secreted into the periplasmic space of Escherichia coli. In short, the cell slurry is thawed in the presence of sodium acetate (pH 3.5), EDTA and phenylmethylsulfonyl fluoride (PMSF) for about 30 minutes. Cell debris can be removed by centrifugation. In the case where the antibody is secreted into the culture medium, the supernatant of such expression system is usually first concentrated using a commercially available protein concentration filter (e.g., Amicon or MilliporePellicon ultrafiltration unit). Protease inhibitors such as PMSF can be included in any of the aforementioned steps to inhibit protein decomposition, and antibiotics can be included to prevent the growth of foreign contaminants.

[0354] Antibody constructs of the invention prepared from host cells can be recovered or purified using, for example, hydroxyapatite chromatography, gel electrophoresis, dialysis, and affinity chromatography. Other protein purification techniques such as ion exchange column fractionation, ethanol precipitation, reverse phase HPLC, silica chromatography, heparin SEPHAROSE chromatography, etc., may also be utilized depending on the antibody to be recovered. TM Chromatography, chromatography on anion or cation exchange resins (such as polyaspartic acid columns), chromatofocusing, SDS-PAGE, and ammonium sulfate precipitation. If the antibody construct of the invention comprises a CH3 domain, Bakerbond ABX resin (JT Baker, Phillipsburg, NJ) can be used for purification.

[0355] Affinity chromatography is a preferred purification technique. The matrix to which the affinity ligand is attached is most commonly agarose, but other matrices are also available. Mechanically stable matrices (such as controlled pore glass or poly(styrenedivinyl)benzene) have faster flow rates and shorter processing times than those achievable with agarose.

[0356] Furthermore, the present invention provides a pharmaceutical composition comprising the antibody construct of the present invention or the antibody construct produced according to the method of the present invention.

[0357] As used herein, the term "pharmaceutical composition" relates to a composition suitable for administration to a patient (preferably a human patient). Particularly preferred pharmaceutical compositions of the present invention comprise preferably a therapeutically effective amount of one or more antibody constructs of the present invention. Preferably, the pharmaceutical composition also comprises a suitable formulation of one or more (pharmaceutically effective) carriers, stabilizers, excipients, diluents, solubilizers, surfactants, emulsifiers, preservatives and / or adjuvants. The acceptable ingredients of the composition are preferably non-toxic to the receptor at the dosage and concentration employed. The pharmaceutical compositions of the present invention include, but are not limited to, liquid, frozen and lyophilized compositions.

[0358] The compositions of the present invention may include a pharmaceutically acceptable carrier. In general, as used herein, "pharmaceutically acceptable carrier" means any and all aqueous and non-aqueous solutions, sterile solutions, solvents, buffers (e.g., phosphate buffered saline (PBS) solutions), water, suspensions, emulsions (such as oil / water emulsions), various types of wetting agents, liposomes, dispersion media, and coatings that are compatible with drug administration (particularly parenteral administration). The use of such media and agents in pharmaceutical compositions is well known in the art, and compositions comprising such carriers can be formulated by well-known conventional methods.

[0359] Certain embodiments provide pharmaceutical compositions comprising an antibody construct of the invention and one or more additional excipients, such as those described illustratively in this section and elsewhere herein. In this regard, excipients may be used in the invention for a variety of purposes, such as adjusting the physical, chemical or biological properties of the formulation, such as adjusting viscosity, and / or methods of the invention to improve effectiveness and / or stabilize such formulations and methods to combat degradation and corruption due to stresses that occur, for example, during manufacture, transportation, storage, preparation prior to use, administration, and thereafter.

[0360] In certain embodiments, the pharmaceutical composition may contain formulation materials for adjusting, maintaining or preserving, for example, the pH value, osmotic pressure, viscosity, clarity, color, isotonicity, odor, sterility, stability, dissolution or release rate, absorption or penetration of the composition (see REMINGTON'S PHARMACEUTICAL SCIENCES, 18th edition, (AR Genrmo ed.), 1990, Mack Publishing Company). In such embodiments, suitable formulation materials may include, but are not limited to:

[0361] Amino acids such as glycine, alanine, glutamine, asparagine, threonine, proline, 2-phenylalanine, including charged amino acids, preferably lysine, lysine acetate, arginine, glutamic acid and / or histidine

[0362] Antimicrobial agents, such as antibacterial and antifungal agents

[0363] Antioxidants, such as ascorbic acid, methionine, sodium sulfite or sodium bisulfite;

[0364] Buffers, buffer systems and buffering agents for maintaining the composition at physiological pH or slightly lower pH, typically in the pH range of about 5 to about 8 or 9; examples of buffers are borate, bicarbonate, Tris-HCl, citrate, phosphate or other organic acid, succinate, phosphate, histidine and acetate; for example, Tris buffer at about pH 7.0-8.5 or acetate buffer at about pH 4.0-5.5;

[0365] Non-aqueous solvents such as propylene glycol, polyethylene glycol, vegetable oils (such as olive oil), and injectable organic esters (such as ethyl oleate);

[0366] Aqueous carriers, including water, alcoholic / aqueous solutions, emulsions or suspensions, including saline and buffered media;

[0367] Biodegradable polymers, such as polyesters;

[0368] Bulking agents, such as mannitol or glycine;

[0369] Chelating agents, such as ethylenediaminetetraacetic acid (EDTA);

[0370] Isotonic and absorption delaying agents;

[0371] Complexing agents, such as caffeine, polyvinylpyrrolidone, β-cyclodextrin or hydroxypropyl-β-cyclodextrin)

[0372] Fillers

[0373] Monosaccharides, disaccharides and other carbohydrates (such as glucose, mannose or dextrin); the carbohydrate may be a non-reducing sugar, preferably trehalose, sucrose, octasulfate, sorbitol or xylitol;

[0374] (Low molecular weight) proteins, polypeptides or protein carriers, such as human or bovine serum albumin, gelatin or immunoglobulins, preferably of human origin;

[0375] Coloring and flavoring agents;

[0376] Sulfur-containing reducing agents such as glutathione, lipoic acid, sodium thioglycolate, thioglycerol, [α]-monothioglycerol, and sodium thiosulfate

[0377] Diluent;

[0378] Emulsifiers;

[0379] Hydrophilic polymers, such as polyvinylpyrrolidone)

[0380] Salt-forming counterions, such as sodium;

[0381] preservatives, such as antimicrobials, antioxidants, chelating agents, inert gases, etc.; examples are: benzalkonium chloride, benzoic acid, salicylic acid, thimerosal, phenethyl alcohol, methylparaben, propylparaben, chlorhexidine, sorbic acid or hydrogen peroxide);

[0382] Metal complexes, such as Zn-protein complexes;

[0383] Solvents and co-solvents (such as glycerol, propylene glycol or polyethylene glycol);

[0384] Sugars and sugar alcohols, such as trehalose, sucrose, octasulfate, mannitol, sorbitol or xylitol, stachyose, mannose, sorbose, xylose, ribose, inositol, galactose, lactitol, ribitol, myoinisitol, dulcitol, glycerol, cyclic alcohols (e.g., inositol), polyethylene glycol; and polyhydroxy sugar alcohols;

[0385] Suspension concentrate;

[0386] Surfactants or wetting agents such as pluronic, PEG, sorbitan esters, polysorbates (such as polysorbate 20, polysorbate), triton, tromethamine, lecithin, cholesterol, tyloxapal; the surfactant may be a detergent with a molecular weight preferably >1.2 KD and / or a polyether with a molecular weight preferably >3 KD; non-limiting examples of preferred detergents are Tween 20, Tween 40, Tween 60, Tween 80 and Tween 85; non-limiting examples of preferred polyethers are PEG 3000, PEG 3350, PEG 4000 and PEG 5000;

[0387] Stability enhancers, such as sucrose or sorbitol;

[0388] Tonicity enhancing agents, such as alkali metal halides (preferably sodium chloride or potassium chloride), mannitol sorbitol;

[0389] Parenteral delivery vehicles include sodium chloride solution, Ringer's dextrose, dextrose and sodium chloride, lactated Ringer's solution, or fixed oils;

[0390] • Intravenous delivery vehicles, including fluid and nutrient replenishers, electrolyte replenishers (such as Ringer's dextrose based replenishers).

[0391] It will be apparent to one skilled in the art that different components of a pharmaceutical composition (e.g., those listed above) may have different roles, e.g., amino acids may act as buffers, stabilizers, and / or antioxidants; mannitol may act as a bulking agent and / or a tonicity enhancer; sodium chloride may act as a delivery vehicle and / or a tonicity enhancer; and so forth.

[0392] It is contemplated that, in addition to the polypeptide of the present invention as defined herein, the compositions of the present invention may also include additional biologically active agents, depending specifically on the intended use of the composition. Such agents may be drugs acting on the gastrointestinal system, drugs acting as cytostatics, drugs preventing hyperuricemia, drugs (e.g., corticosteroids) that suppress immune responses, drugs regulating inflammatory responses, drugs acting on the circulatory system, and / or drugs such as cytokines known in the art. It is also contemplated that the antibody constructs of the present invention are applied to co-therapy, i.e., in combination with another anticancer agent.

[0393] In certain embodiments, the best pharmaceutical composition will be determined by those skilled in the art according to, for example, the expected route of administration, delivery form and the required dosage.See, for example, REMINGTON'S PHARMACEUTICAL SCIENCES, the same. In certain embodiments, such compositions can affect the physical state, stability, in vivo release rate and in vivo clearance rate of the antibody construct of the present invention. In certain embodiments, the main vehicle or carrier in the pharmaceutical composition can be aqueous or non-aqueous in nature. For example, suitable vehicles or carriers can be water for injection, physiological saline solution or artificial cerebrospinal fluid, which may be supplemented with other substances common in the composition for parenteral administration. Neutral buffered saline or saline mixed with serum albumin are other exemplary vehicles. In certain embodiments, the antibody construct composition of the present invention can be prepared for storage by mixing the selected composition with the desired purity with optional formulation (REMINGTON'S PHARMACEUTICAL SCIENCES, the same) in the form of lyophilized cake or aqueous solution. In addition, in certain embodiments, the antibody construct of the present invention can be formulated as lyophilized powder using suitable excipients (such as sucrose).

[0394] When considering parenteral administration, the therapeutic composition used in the present invention can be provided in the form of a pyrogen-free, parenterally acceptable aqueous solution comprising the desired antibody construct of the present invention in a pharmaceutically acceptable vehicle. The vehicle particularly suitable for parenteral injection is sterile distilled water, and the antibody construct of the present invention is formulated as a sterile, isotonic solution appropriately preserved therein. In certain embodiments, the preparation may involve the desired molecule and be prepared together with a reagent such as an injectable microsphere, a biologically erodible particle, a polymeric compound (such as polylactic acid or polyglycolic acid), a bead, or a liposome, which may provide control or sustained release of a product that may be delivered via a depot injection. In certain embodiments, hyaluronic acid may also be used, which has a persistent effect that promotes circulation. In certain embodiments, an implantable drug delivery device may be used to introduce the desired antibody construct.

[0395] Other pharmaceutical compositions will be apparent to those skilled in the art, including formulations comprising the antibody constructs of the present invention in sustained or controlled delivery / release formulations. It is also known to those skilled in the art to prepare a variety of other sustained or controlled delivery methods (such as liposome carriers, bioerodible microparticles or porous beads and reservoir injections). See, for example, International Patent Application No. PCT / US93 / 00829, which describes the controlled release of porous polymer microparticles for delivering pharmaceutical compositions. Sustained release formulations may include a semipermeable polymer matrix in the form of a shaped article (e.g., a film or microcapsule). Sustained release matrices may include polyesters, hydrogels, polylactides (such as disclosed in U.S. Pat. No. 3,773,919 and European Patent Application Publication No. EP 058481), copolymers of L-glutamic acid and γ-ethyl L-glutamate (Sidman et al., 1983, Biopolymers 2: 547-556), poly (2-hydroxyethyl-methacrylate) (Langer et al., 1981, J. Biomed. Mater. Res. 15: 167-277 and Langer, 1982, Chem. Tech. 12: 98-105), ethylene vinyl acetate (Langer et al., 1981, supra) or poly-D (-) -3-hydroxybutyric acid (European Patent Application Publication No. EP 133,988). Sustained release compositions may also include liposomes that can be prepared by any of several methods known in the art. See, eg, Eppstein et al., 1985, Proc. Natl. Acad. Sci. USA 82:3688-3692; European Patent Application Publication Nos. EP 036,676, EP 088,046, and EP 143,949.

[0396] The antibody constructs can also be encapsulated in microcapsules prepared, for example, by coacervation techniques or by interfacial polymerization (e.g., hydroxymethylcellulose or gelatin microcapsules and poly(methyl methacrylate) microcapsules, respectively), in colloidal drug delivery systems (e.g., liposomes, albumin microspheres, microemulsions, nanoparticles and nanocapsules), or in macroemulsions. Such techniques are disclosed in Remington's Pharmaceutical Sciences, 16th edition, Oslo, A., ed., (1980).

[0397] Pharmaceutical compositions for in vivo administration are usually provided in the form of sterile preparations. Sterilization can be achieved by filtering through a sterile filtration membrane. When the composition is lyophilized, sterilization using this method can be performed before or after lyophilization and rehydration. Compositions for parenteral administration can be stored in lyophilized form or in solution form. Parenteral compositions are generally placed in containers with sterile access ports, such as intravenous solution bags or vials with stoppers that can be pierced by hypodermic needles.

[0398] Another aspect of the invention comprises a self-buffering formulation of the antibody construct of the invention, which can be used as a pharmaceutical composition as described in International Patent Application WO 06138181A2 (PCT / US2006 / 022599). Numerous discussions are available on protein stabilization and formulation materials and methods suitable for use therein, such as Arakawa et al., "Solvent interactions in pharmaceutical formulations," Pharm Res. 8(3):285-91 (1991); Kendrick et al., "Physical stabilization of proteins in aqueous solution," In: RATIONAL DESIGN OF STABLE PROTEIN FORMULATIONS: THEORY AND PRACTICE, Carpenter and Manning, eds. Pharmaceutical Biotechnology. 13:61-84 (2002) and Randolph et al., "Surfactant-protein interactions," Pharm Biotechnol. 13:159-75 (2002), particularly with respect to excipients and methods for use therein in the self-buffering protein formulations of the present invention, particularly with respect to protein drug products and methods for veterinary and / or human medical use.

[0399] Salt can be used for example to adjust the ionic strength and / or isotonicity of preparation and / or improve the solubility and / or physical stability of protein or other ingredients of composition according to the present invention according to some embodiments of the present invention. As is well known, ions can stabilize proteins in native state by being bonded to charged residues on protein surface and by shielding charged and polar groups in protein and reducing the intensity of their electrostatic interaction, attraction and repulsion interaction. Ions can also stabilize proteins in denatured state by being bonded to denatured peptide bonds (--CONH) of protein in particular. In addition, the interaction of ions with charged groups and polar groups in protein can also reduce intermolecular electrostatic interaction, thus preventing or reducing protein aggregation and insolubility.

[0400] Ionic species vary significantly in their effects on proteins. Many classes of ions and their effects on proteins have been developed for use in formulating pharmaceutical compositions according to the present invention. One example is the Hofmeister series, which arranges ions and polar non-ionic solutes by their effects on the conformational stability of proteins in solution. Stabilizing a solute is called "kosmotropic". Destabilizing a solute is called "chaotropic". Kosmotropic agents are typically used in high concentrations (e.g., >1 molar ammonium sulfate) to precipitate proteins from solution ("salting out"). Chaotropic agents are typically used to denature and / or dissolve proteins ("salting out"). The relative effectiveness of an ion on "salting out" and "salting out" defines its position in the Hofmeister series.

[0401] Free amino acids can be used as bulking agents, stabilizers and antioxidants and other standard uses in the antibody construct formulations of the present invention according to various embodiments of the present invention. Lysine, proline, serine and alanine can be used to stabilize the protein in the formulation. Glycine can be used for freeze drying to ensure the correct cake structure and characteristics. Arginine can be used to inhibit protein aggregation in liquid and freeze-dried formulations. Methionine can be used as an antioxidant.

[0402] Polyols include sugars, such as mannitol, sucrose and sorbitol and polyols such as glycerol and propylene glycol, and (for the purpose of this article) polyethylene glycol (PEG) and related substances. Polyols are lyophilic. They are stabilizers suitable for use in liquid and lyophilized preparations to protect proteins from physical and chemical degradation processes. Polyols can also be used to adjust the tension of the preparation. The polyol that can be used in the selected embodiment of the present invention is mannitol, which is commonly used to ensure the structural stability of the cake in the lyophilized preparation. It ensures the structural stability of the cake. It is generally used together with a lyoprotectant (such as sucrose). Sorbitol and sucrose are preferred agents for adjusting tension and serve as stabilizers to avoid freeze-thaw stress during transportation or to avoid agglomeration during the manufacturing process. Reducing sugars (which contain free aldehyde groups or ketone groups) such as glucose and lactose can saccharify surface lysine and arginine residues. Therefore, it is generally not among the preferred polyols used according to the present invention. Furthermore, sugars that form such reactive species, such as sucrose, which hydrolyzes to fructose and glucose under acidic conditions and thus causes glycosylation, are also not among the preferred polyols of the present invention in this regard. PEG can be used to stabilize proteins and act as a cryoprotectant and in this regard can be used in the present invention.

[0403] Embodiments of the antibody construct formulations of the present invention further comprise surfactants. Protein molecules can be easily adsorbed onto the surface and denatured and subsequently aggregated at gas-liquid, solid-liquid and liquid-liquid interfaces. These effects are generally inversely proportional to protein concentration. These harmful interactions are generally inversely proportional to protein concentration and are generally aggravated by physical agitation such as that produced during product transportation and handling. Surfactants are generally used to prevent, minimize or reduce surface adsorption. In this regard, surfactants suitable for use in the present invention include other fatty acid esters of polysorbate 20, polysorbate 80, sorbitan polyethoxylates and poloxamer 188. Surfactants are also commonly used to control protein conformational stability. In this regard, the use of surfactants has protein specificity, because any given surfactant will generally stabilize some proteins and destabilize other proteins.

[0404] Polysorbates are susceptible to oxidative degradation and often contain sufficient amounts of peroxide when supplied to cause oxidation of protein residue side chains (especially methionine). Therefore, polysorbates should be used with caution and when used, should be used at their lowest effective concentration. In this regard, polysorbates exemplify the general rule that excipients should be used at their lowest effective concentration.

[0405] Embodiments of the antibody construct formulations of the present invention further comprise one or more antioxidants. To a certain extent, harmful oxidation of proteins can be prevented in pharmaceutical formulations by maintaining appropriate levels of ambient oxygen and temperature and by avoiding exposure. Antioxidant excipients can also be used to prevent oxidative degradation of proteins. In this regard, available antioxidants are reducing agents, oxygen / free radical scavengers and chelating agents. Antioxidants used for therapeutic protein formulations according to the present invention are preferably water-soluble and maintain their activity throughout the shelf life of the product. In this regard, EDTA is a preferred antioxidant according to the present invention. Antioxidants can damage proteins. For example, reducing agents (specifically, such as glutathione) can destroy intramolecular disulfide bonds. Therefore, antioxidants selected for the present invention are particularly eliminated or fully reduced in the possibility of damaging proteins in the formulation itself.

[0406] The formulations according to the invention may contain metal ions which are protein cofactors and are required for the formation of protein coordination complexes, such as zinc required for the formation of certain insulin suspensions. Metal ions may also inhibit some processes that degrade proteins. However, metal ions also catalyze physical and chemical processes that degrade proteins. Magnesium ions (10-120 mM) can be used to inhibit the isomerization of aspartic acid to isoaspartic acid. Ca +2 ions (up to 100 mM) can improve the stability of human DNA enzyme. +2 , Mn +2 and Zn +2 Can destabilize rhDNase. Similarly, Ca +2 and Sr +2 It can stabilize factor VIII, which can be affected by Mg +2 , Mn +2 and Zn +2 , Cu +2 and Fe +2 Unstable, and its aggregation can be caused by Al +3 Ions increase.

[0407] Embodiments of the antibody construct formulations of the present invention further include one or more preservatives. When developing a multi-dose parenteral formulation that involves more than one extraction from the same container, a preservative is required. Its primary function is to inhibit microbial growth and ensure product sterility throughout the shelf life or drug use period. Commonly used preservatives include benzyl alcohol, phenol, and meta-cresol. Although preservatives have a long history of use with small molecule parenteral formulations, developing protein formulations including preservatives can still be challenging. Preservatives almost always have an effect (aggregation) that destabilizes proteins, and this has become a major factor limiting their use in multi-dose protein formulations. So far, most protein drugs that have been formulated are only used for single use. However, when multi-dose formulations become possible, they have the additional advantage of making patients convenient and increasing sales. A good example is human growth hormone (hGH), where the development of preservatives has made more convenient, multiple-use injection pens commercially available. At least four such pen devices containing hGH preservatives are currently available on the market. Norditropin (liquid, Novo Nordisk), Nutropin AQ (liquid, Genentech), and Genotropin (lyophilized - dual chamber cartridge, Pharmacia & Upjohn) contain phenol, while Somatrope (Eli Lilly) is formulated with m-cresol. Several aspects need to be considered during the formulation and development of a preservative dosage form. The effective preservative concentration in the drug product must be optimized. This requires testing a given preservative in a dosage form such that a concentration range confers antimicrobial effectiveness without compromising protein stability.

[0408] As expected, developing liquid preparations containing preservatives is more challenging than freeze-dried preparations. Freeze-dried products can be freeze-dried without preservatives and rehydrated with diluents containing preservatives when used. This shortens the time when preservatives are in contact with proteins, greatly reducing the risk of related stability to a minimum. In the case of liquid preparations, preservative effectiveness and stability should be maintained throughout the product shelf life (approximately 18 to 24 months). The key point to note is that preservative effectiveness should be shown in the final preparation containing active drugs and all excipient components.

[0409] The antibody constructs disclosed herein can also be formulated as immunoliposomes. "Liposomes" are small vesicles composed of various types of lipids, phospholipids and / or surfactants that can be used to deliver drugs to mammals. The components of the liposomes are usually arranged in a bilayer form, similar to the lipid arrangement of biological membranes. Liposomes containing antibody constructs are prepared by methods known in the art, such as Epstein et al., Proc. Natl. Acad. Sci. USA 82: 3688 (1985); Hwang et al., Proc. Natl Acad. Sci. USA. 77: 4030 (1980); U.S. Patents No. 4,485,045 and No. 4,544,545; and WO 97 / 38731. Liposomes with increased circulation time are disclosed in U.S. Patent No. 5,013,556. Particularly suitable liposomes can be produced by reverse phase evaporation with a lipid composition comprising phosphatidylcholine, cholesterol and PEG-derivatized phosphatidylethanolamine (PEG-PE). The liposomes are extruded through a filter with a defined pore size to produce liposomes with the desired diameter. The Fab' fragments of the antibody constructs of the present invention can be conjugated to the liposomes via a disulfide interconversion reaction as described in Martin et al. J. Biol. Chem. 257: 286-288 (1982). The liposomes may optionally contain a chemotherapeutic agent. See Gabizon et al. J. National Cancer Inst. 81 (19) 1484 (1989).

[0410] Once the pharmaceutical composition has been formulated, it can be stored in a sterile vial in the form of a solution, suspension, gel, emulsion, solid, crystal, or dehydrated or lyophilized powder. Such preparations can be stored in a ready-to-use form or in a reconstituted form (e.g., lyophilized form) prior to administration.

[0411] The biological activity of the pharmaceutical composition defined herein can be determined, for example, by cytotoxicity assays, as described in the following examples, in WO 99 / 54440, or by Schlereth et al. (Cancer Immunol. Immunother. 20 (2005), 1-12). As used herein, "efficacy" or "in vivo efficacy" refers to the response to a therapy performed by a pharmaceutical composition of the present invention using, for example, standardized NCI response criteria. The success or in vivo efficacy of a therapy using a pharmaceutical composition of the present invention refers to the effectiveness of the composition for its intended purpose, i.e., the ability of the composition to cause its desired effect (i.e., the elimination of pathological cells such as tumor cells). In vivo efficacy can be monitored by standard methods established for the corresponding disease entity, including but not limited to white blood cell counts, differentials, fluorescence activated cell sorting, bone marrow aspiration. In addition, various disease-specific clinical chemistry parameters and other established standard methods can be used. In addition, computer-assisted tomography, X-ray, and magnetic resonance tomography (e.g., for response assessment based on National Cancer Institute criteria [Cheson BD, Horning SJ, Coiffier B, Shipp MA, Fisher RI, Connors JM, Lister TA, Vose J, Grillo-Lopez A, Hagenbeek A, Cabanillas F, Klippensten D, Hiddemann W, Castellino R, Harris NL, Armitage JO, Carter W, Hoppe R, Canellos GP. Report of an international workshop to standardize response criteria for non-Hodgkin's lymphomas. NCI Sponsored International Working Group. J Clin Oncol. 1999 Apr;17(4):1244]), positron emission tomography, white blood cell count, differential, fluorescence activated cell sorting, bone marrow aspirate, lymph node biopsy / histology, and various lymphoma-specific clinical chemistry parameters (e.g., lactate dehydrogenase), and other established standard methods.

[0412] Another major challenge in the development of drugs, such as the pharmaceutical compositions of the present invention, is the predictable modulation of pharmacokinetic properties. To this end, a pharmacokinetic profile of a candidate drug, i.e., a profile of pharmacokinetic parameters that affect the ability of a particular drug to treat a given condition, can be established. Pharmacokinetic parameters of a drug that affect the ability of a drug to treat a certain disease entity include, but are not limited to, half-life, volume of distribution, hepatic first-pass metabolism, and serum binding. The efficacy of a given agent can be affected by each of the parameters mentioned above.

[0413] "Half-life" means the time at which 50% of the administered drug is eliminated by biological processes such as metabolism, excretion, etc. So-called "hepatic first-pass metabolism" refers to the tendency of a drug to be metabolized after first contact with the liver, i.e., during the first pass through the liver. "Volume of distribution" means the degree of retention of a drug throughout the various compartments of the body (e.g., intracellular and extracellular spaces, tissues and organs, etc.), and the distribution of the drug within these compartments. "Extent of serum binding" means the tendency of a drug to interact and bind to serum proteins (such as albumin), resulting in a reduction or loss of the drug's biological activity.

[0414] Pharmacokinetic parameters also include bioavailability, lag time (Tlag), Tmax, absorption rate, more effective and / or Cmax of a given amount of administered drug. "Bioavailability" means the amount of the drug in the blood compartment. "Lag time" means the time delay between drug administration and its detection and measurability in blood or plasma. "Tmax" is the time before reaching the maximum blood drug concentration, and "Cmax" is the blood concentration obtained to the maximum extent using a given drug. The time to achieve the required drug blood or tissue concentration for biological effect is affected by all parameters. The pharmacokinetic parameters of the bispecific antibody constructs showing cross-species specificity (which can be measured in preclinical animal tests of non-chimpanzee primates as outlined above) are also described in publications such as Schlereth et al. (Cancer Immunol. Immunother. 20 (2005), 1-12).

[0415] One embodiment provides an antibody construct of the present invention or an antibody construct produced according to the method of the present invention for use in preventing, treating or ameliorating a tumor or cancer disease or a metastatic cancer disease.

[0416] The formulations described herein can be used as pharmaceutical compositions for treating, ameliorating and / or preventing pathological medical conditions as described herein in patients in need thereof. The term "treatment" refers to both therapeutic treatment and preventive or prophylactic measures. Treatment includes applying or administering a formulation to the body, isolated tissue or cells from a patient suffering from a disease / disorder, having symptoms of a disease / disorder or being susceptible to a disease / disorder with the purpose of curing, rehabilitating, alleviating, mitigating, altering, remedying, improving, modifying or influencing a disease, a symptom of a disease or a predisposition to a disease, a symptom of a disease / disorder or a predisposition to a disease / disorder.

[0417] The term "improvement" as used herein refers to any improvement in the disease condition of a patient suffering from a tumor or cancer or metastatic cancer as described below, which is performed by administering an antibody construct according to the present invention to a subject in need thereof. Such improvement can also be considered as a slowing down or stopping of the progression of a patient's tumor or cancer or metastatic cancer. The term "prevention" as used herein means avoiding the occurrence or reoccurrence of a patient suffering from a tumor or cancer or metastatic cancer as described below, which is performed by administering an antibody construct according to the present invention to a subject in need thereof.

[0418] The term "disease" refers to any condition that would benefit from treatment with an antibody construct or pharmaceutical composition as described herein. It includes chronic and acute disorders or diseases, including those pathological conditions that predispose a mammal to the disease in question.

[0419] "Neoplasia" is an abnormal growth of tissue, usually but not always forming a mass. When a mass is also formed, it is generally referred to as a "tumor". Neoplasia or tumors can be benign, potentially malignant (precancerous) or malignant. Malignant neoplasms are generally referred to as cancer. They usually invade and destroy surrounding tissues and can form metastases, i.e., they spread to other parts, tissues or organs of the body. Therefore, the term "metastatic cancer" encompasses transfer to other tissues or organs outside the original tumor. Lymphomas and leukemias are lymphatic neoplasms. For purposes of the present invention, they are also encompassed by the terms "tumor" or "cancer".

[0420] In a preferred embodiment of the present invention, the tumor or cancer disease is selected from the group including, but not limited to (or consisting of): lung cancer, preferably SCLC; breast, cervical, colon, colorectal, endometrial, head and neck, liver, ovarian, pancreatic, prostate, skin, stomach, testicular, thyroid, adrenal, kidney, bladder, uterine, esophageal, urothelial and brain tumors or cancers; lymphoma; carcinoma; and sarcoma; and metastatic cancer diseases derived from any of the foregoing.

[0421] More specifically, the tumor or cancer disease can be selected from the following: small cell lung cancer (SCLC), non-small cell lung cancer (NSCLC), glioma, glioblastoma, melanoma, neuroendocrine prostate cancer, neuroendocrine pancreatic cancer, hepatoblastoma and hepatocellular carcinoma. The metastatic cancer disease can be derived from any of the aforementioned diseases.

[0422] The present invention also provides a method for treating or ameliorating tumors or cancer diseases or metastatic cancer diseases, comprising the step of administering the antibody construct of the present invention or the antibody construct produced according to the method of the present invention to a subject in need thereof.

[0423] The term "subject in need thereof" or "subject in need of treatment" includes subjects already suffering from the disorder as well as subjects in which the disorder is to be prevented. Subjects in need thereof or "patients" include humans and other mammalian subjects receiving prophylactic or therapeutic treatment.

[0424] The antibody constructs of the invention will generally be designed for a specific route and method of administration, a specific dosage and frequency of administration, a specific treatment of a specific disease, particularly within a certain range of bioavailability and persistence. The materials of the composition are preferably formulated at concentrations acceptable to the site of administration.

[0425] Thus, formulations and compositions can be designed according to the present invention to be delivered by any suitable route of administration. In the context of the present invention, routes of administration include but are not limited to

[0426] Topical routes (such as transepithelial, inhalation, nasal, ocular, otic, vaginal, transmucosal);

[0427] Enteral routes (such as oral, enteral, sublingual, sublabial, buccal, rectal); and

[0428] - Parenteral routes (such as intravenous, intraarterial, intraosseous, intramuscular, intracerebral, intraventricular, epidural, intrathecal, subcutaneous, intraperitoneal, extra-amniotic, intraarticular, intracardiac, intradermal, intralesional, intrauterine, intravesical, intravitreal, transdermal, intranasal, transmucosal, intrasynovial, intraluminal).

[0429] The pharmaceutical composition and antibody construct of the present invention are particularly suitable for parenteral administration, such as subcutaneous or intravenous delivery, for example, by injection (such as push injection) or by infusion (such as continuous infusion). The pharmaceutical composition can be administered using a medical device. Examples of medical devices for administering pharmaceutical compositions are described in U.S. Patents Nos. 4,475,196, 4,439,196, 4,447,224, 4,447,233, 4,486,194, 4,487,603, 4,596,556, 4,790,824, 4,941,880, 5,064,413, 5,312,335, 5,312,335, 5,383,851 and 5,399,163.

[0430] Specifically, the present invention provides uninterrupted administration of suitable compositions. As a non-limiting example, uninterrupted or substantially uninterrupted, i.e., continuous administration can be achieved by metering the inflow of therapeutic agents into the patient's body through a small pump system worn by the patient. The pharmaceutical composition comprising the antibody construct of the present invention can be administered using the pump system. Such pump systems are generally well known in the art and generally rely on regular replacement of cartridges containing therapeutic agents to be infused. When the cartridges in such pump systems are replaced, the originally uninterrupted flow of therapeutic agents to the patient's body can then be temporarily interrupted. In this case, within the meaning of the pharmaceutical means and methods of the present invention, the administration phase before the cartridge replacement and the administration phase after the cartridge replacement are still considered to constitute together a "uninterrupted administration" of such therapeutic agents.

[0431] Continuous or uninterrupted administration of the antibody construct of the present invention can be intravenous or subcutaneous administration, which is by means of a fluid delivery device or a small pump system including a fluid drive mechanism for driving the fluid to leave the reservoir and an actuating mechanism for actuating the drive mechanism. The pump system for subcutaneous administration may include a needle or a catheter for passing through the patient's skin and delivering a suitable composition to the patient's body. The pump system may be directly fixed or attached to the patient's skin independently of a vein, an artery or a blood vessel, so that direct contact can be made between the pump system and the patient's skin. The pump system may be attached to the patient's skin for 24 hours to several days. In the case of a small volume reservoir, the pump system may have a small size. As a non-limiting example, the volume of the reservoir for a suitable pharmaceutical composition to be administered may be between 0.1 and 50 ml.

[0432] Continuous application can also be percutaneous application by means of the patch worn on the skin and replaced at certain time intervals. Those skilled in the art know the patch system for drug delivery suitable for this purpose. It should be noted that percutaneous application is particularly suitable for uninterrupted application, because changing the first used-up patch can be advantageously achieved simultaneously by, for example, placing a new second patch on the surface of the skin of the first used-up patch in close proximity and then immediately removing the first used-up patch. The problem of flow interruption or power supply battery failure will not occur.

[0433] If the pharmaceutical composition has been lyophilized, the lyophilized material is first reconstituted in an appropriate liquid prior to administration. The lyophilized material can be reconstituted in, for example, bacteriostatic water for injection (BWFI), saline, phosphate buffered saline (PBS), or the same formulation in which the protein was prior to lyophilization.

[0434] The compositions of the present invention can be administered to a subject at a suitable dose, which can be determined, for example, by administering increasing doses of an antibody construct of the present invention that exhibits cross-species specificity to a non-chimpanzee primate (e.g., macaque) through a dose escalation study. As described above, the antibody constructs of the present invention that exhibit cross-species specificity as described herein can be advantageously used in the same form for preclinical trials of non-chimpanzee primates and for use as human drugs. The dosing regimen will be determined by the attending physician and clinical factors. As is well known in the medical field, for any one patient, the dosage depends on many factors, including the patient's size, body surface area, age, specific compound to be administered, sex, time and route of administration, general health, and other drugs administered simultaneously.

[0435] The term "effective dose" is defined as an amount sufficient to achieve or at least partially achieve the desired effect. The term "therapeutically effective dose" is defined as an amount sufficient to cure or at least partially arrest the disease and its complications in patients who have already suffered from the disease. The amount or dosage effective for this purpose will depend on: the condition to be treated (indications), the antibody construct delivered, the treatment situation and goal, the severity of the disease, previous therapy, the patient's clinical history and response to the therapeutic agent, the route of administration, the patient's body shape (weight, body surface or organ size) and / or condition (age and general health) and the general state of the patient's own immune system. The appropriate dosage can be adjusted according to the judgment of the attending physician so that it can be administered to the patient once or through a series of administrations, and the best therapeutic effect is obtained.

[0436] Typical dosages may range from about 0.1 μg / kg up to about 30 mg / kg or more, depending on the factors mentioned above. In specific embodiments, dosages may range from 1.0 μg / kg to about 20 mg / kg, optionally 10 μg / kg to about 10 mg / kg, or 100 μg / kg to about 5 mg / kg.

[0437] A therapeutically effective amount of an antibody construct of the invention preferably reduces the severity of disease symptoms, increases the frequency or duration of disease symptom-free periods, or prevents impairment or disability resulting from disease distress. For treating tumors expressing DLL3, a therapeutically effective amount of an antibody construct of the invention (e.g., an anti-DLL3 / anti-CD3 antibody construct) preferably inhibits cell growth or tumor growth by at least about 20%, at least about 40%, at least about 50%, at least about 60%, at least about 70%, at least about 80%, or at least about 90% relative to untreated patients. The ability of a compound to inhibit tumor growth can be assessed in an animal model predictive of efficacy in human tumors.

[0438] The pharmaceutical composition can be administered as the sole therapeutic agent or in combination with other therapies (such as anti-cancer therapies, e.g., other protein and non-protein drugs) as needed. These drugs can be administered simultaneously with the composition comprising the antibody construct of the invention as defined herein at appropriately defined time intervals and dosages, or separately before or after administration of the antibody construct.

[0439] The term "effective and nontoxic dose" as used herein refers to a tolerable dose of the antibody construct of the present invention, which is high enough to cause pathological cell elimination, tumor removal, tumor reduction or disease stabilization without or substantially without major toxic effects. Such effective and nontoxic doses can be determined, for example, by dose escalation studies described in the art and should be lower than doses that cause serious adverse events (dose-limiting toxicity; DLT).

[0440] As used herein, the term "toxicity" refers to the toxic effects of a drug that manifest as adverse events or serious adverse events. These adverse events may refer to a general lack of drug tolerance and / or a lack of local tolerance after administration. Toxicity may also include teratogenic or carcinogenic effects caused by the drug.

[0441] As used herein, the term "safety", "in vivo safety" or "tolerance" defines the administration of a drug without causing serious adverse events immediately after administration (local tolerance) and without causing serious adverse events during the longer-term administration of the drug. "Safety", "in vivo safety" or "tolerance" can be evaluated, for example, at regular intervals during treatment and follow-up. Measurements include clinical assessments, such as organ performance, and screening for laboratory abnormalities. Clinical assessments can be performed, and deviations compared to normal findings are recorded / coded according to NCI-CTC and / or MedDRA standards. Organ performance may include standards such as allergies / immunology, blood / bone marrow, arrhythmias, coagulation, such as described in, for example, Common Terminology Criteria for Adverse Events v3.0 (CTCAE). Testable laboratory parameters include, for example, hematology, clinical chemistry, coagulation profiles and urinalysis, as well as examinations of other body fluids such as serum, plasma, lymph or spinal fluid, fluid, etc. Thus, safety can be assessed, for example, by physical examination, imaging techniques (i.e. ultrasound, x-ray, CT scan, magnetic resonance imaging (MRI), other measurements using technical devices (i.e. electrocardiogram), vital signs, by measuring laboratory parameters and recording adverse events. For example, in the uses and methods according to the invention, adverse events in non-chimpanzee primates can be examined by histopathological and / or histochemical methods.

[0442] The above terms are also mentioned, for example, in Preclinical safety evaluation of biotechnology-derived pharmaceuticals S6; ICH Harmonised Tripartite Guideline; ICH Steering Committee meeting, July 16, 1997.

[0443] In another embodiment, the present invention provides a kit comprising an antibody construct of the present invention, an antibody construct produced according to a method of the present invention, a polypeptide of the present invention, a vector of the present invention and / or a host cell of the present invention.

[0444] In the context of the present invention, the term "kit" means two or more components packaged together in a container, receptacle or other object, one of which corresponds to an antibody construct, pharmaceutical composition, vector or host cell of the present invention. Thus, a kit can be described as a group of products and / or appliances sufficient to achieve a certain purpose, which can be sold as a single unit.

[0445] The kit may comprise one or more receptacles (such as vials, ampoules, containers, syringes, bottles, bags) of any suitable shape, size and material (preferably waterproof, e.g., plastic or glass) containing an appropriate dose of the antibody construct or pharmaceutical composition of the invention for administration (see above). The kit may additionally contain instructions for use (e.g., in the form of a brochure or instruction manual), means for administering the antibody construct of the invention (such as a syringe, pump, infusion set, etc.), means for rehydrating the antibody construct of the invention and / or means for diluting the antibody construct of the invention.

[0446] The present invention also provides kits for single-dose administration units. The kits of the present invention may also contain a first receiver comprising a dried / lyophilized antibody construct and a second receiver comprising an aqueous formulation. In certain embodiments of the present invention, kits containing single-chamber and multi-chamber prefilled syringes (e.g., liquid syringes and lyosyringes) are provided. BRIEF DESCRIPTION OF THE DRAWINGS

[0447] The attached figure shows:

[0448] Figure 1

[0449] Schematic representation of DLL3 ECD / truncated DLL3 construct expressed on CHO cells for epitope mapping. The transmembrane and intracellular domains are derived from EpCAM. See Example 1.

[0450] Figure 2

[0451] Epitope mapping of anti-DLL3 antibody constructs. Examples of bispecific antibody constructs recognizing the EGF-3 domain and the EGF-4 domain, respectively, as detected by epitope mapping. The x-axis depicts PE-A (PE=phycoerythrin, A=signal area), while the y-axis depicts counts. See Example 2.

[0452] Figure 3

[0453] Cross-reactivity of anti-DLL3 antibody constructs as detected by flow cytometry: binding to human and macaque DLL3 and CD3. The x-axis depicts FL2-H, while the y-axis depicts counts. See Example 5.

[0454] Figure 4

[0455] Analysis of anti-DLL3 antibody constructs by flow cytometry: No binding to human paralogs DLL1 and DLL4. The x-axis depicts FL1-H and the y-axis depicts counts. See Example 6.

[0456] Figure 5

[0457] Stability of anti-DLL3 antibody constructs after 96 hours incubation in human plasma. See Example 11.

[0458] Figure 6

[0459] Potency differences between monomeric and dimeric isoforms of anti-DLL3 antibody constructs. See Example 15.

[0460] Figure 7

[0461] In vitro internalization assays were performed as described in Example 16. Antibody constructs were pre-incubated on target cells (SHP-77) in the absence of T cells to measure the loss of antibody constructs due to internalization. The results showed that no significant internalization occurred. Results achieved with a control target (whose internalization is well known) are shown on the right for comparison.

[0462] Figure 8

[0463] SHP-77 Model( Fig. 8A ) and WM266-4 models ( Figure 8B ) in mice xenografts. Tumor volume is plotted against time. See Example 18.

[0464] Fig. 9

[0465] As determined in Example 20, different HLE antibody constructs ( Fig.9A Albumin fusions and Fig. 9B Pharmacokinetics of Fc fusions in .

[0466] Specific embodiment Example

[0467] The following examples illustrate the present invention. These examples should not be construed as limiting the scope of the present invention. The present invention is limited only by the claims.

[0468] Example 1

[0469] Generation of CHO cells expressing wild-type and truncated DLL3

[0470] For the purposes of Examples 1 and 2, the extracellular domain of the DLL3 antigen can be subdivided into different subdomains or regions defined by the following amino acid positions:

[0471] Signal peptide: 1-26

[0472] N-terminal: 27-175

[0473] DSL: 176-215

[0474] EGF-1: 216-249

[0475] EGF-2: 274-310

[0476] EGF-3: 312-351

[0477] EGF-4: 353-389

[0478] EGF-5: 391-427

[0479] EGF-6: 429-465

[0480] For the construction of truncated DLL3 molecules for epitope mapping, the sequences of the corresponding eight extracellular domains of human DLL3 (signal peptide plus N-terminus, DSL, EGF1, EGF2, EGF3, EGF4, EGF5 and EGF6) were deleted from the antigen step by step, starting from the N-terminus. The following molecules were generated; see also Figure 1 :

[0481] Hu DLL3 ECD / complete ECD SEQ ID NO: 263

[0482] Hu DLL3 ECD / Max DSL SEQ ID NO: 264

[0483] ·Hu DLL3 ECD / Up to EGF-1 SEQ ID NO:265

[0484] ·Hu DLL3 ECD / Up to EGF-2 SEQ ID NO:266

[0485] ·Hu DLL3 ECD / Up to EGF-3 SEQ ID NO:267

[0486] ·Hu DLL3 ECD / up to EGF-4 SEQ ID NO:268

[0487] ·Hu DLL3 ECD / up to EGF-5 SEQ ID NO:269

[0488] Hu DLL3 ECD / EGF-6 only SEQ ID NO: 270

[0489] For the generation of CHO cells expressing human, cynomolgus monkey ("cyno") and truncated human N-terminal V5-tagged DLL3, the corresponding coding sequences of human DLL3-ECD (SEQ ID NO: 253; see also GeneBank Accession No. NM_016941), cyno DLL3-ECD (SEQ ID NO: 272) and seven truncated human N-terminal V5-tagged DLL3-ECD forms (see above) were cloned into a plasmid called pEF-DHFR (pEF-DHFR is described in Raum et al. Cancer Immunol Immunother 50 (2001) 141-150). For cell surface expression of human and cyno DLL3, the original signal peptide was used, and for cell surface expression of truncated human N-terminal DLL3, the mouse IgG heavy chain signal peptide was used followed by a V5 tag. All mentioned DLL3-ECD sequences are followed in frame by the coding sequence of an artificial Ser / Gly-linker, followed by the domain derived from the transmembrane / intracellular domain of human EpCAM (amino acids 266-314 of the sequence published in GenBank Accession No. NM_002354). All cloning procedures were performed according to standard protocols (Sambrook, Molecular Cloning; A Laboratory Manual, 3rd Edition, Cold Spring Harbour Laboratory Press, Cold Spring Harbour, New York (2001)). For each construct, the corresponding plasmid was transfected into DHFR-deficient CHO cells for eukaryotic expression as described by Kaufman RJ (1990) Methods Enzymol. 185, 537-566.

[0490] Expression of human and cynoDLL3 on CHO cells was verified in a FACS assay using a monoclonal mouse IgG2b anti-human DLL3 antibody. Expression of seven truncated forms of human DLL3-ECD was verified using a monoclonal mouse IgG2a anti-v5 tag antibody. Bound monoclonal antibodies were detected with anti-mouse IgG Fc-γ-PE. As a negative control, cells were incubated with an isotype control antibody instead of the primary antibody. Samples were measured by flow cytometry.

[0491] Example 2

[0492] Epitope mapping of anti-DLL3 antibody constructs

[0493] Cells transfected with human DLL3 and truncated human DLL3 molecules (see Example 1) were stained with crude, undiluted periplasmic extracts containing a bispecific DLL3xCD3 antibody construct fused to human albumin (variant 1), wherein the CD3 binding domain was designated I2C, in PBS / 1.5% FCS. Bound molecules were detected with an in-house mouse monoclonal anti-CD3 binding domain antibody (50 μl) followed by anti-mouse IgG Fc-γ-PE (1:100, 50 μl; Jackson Immunoresearch #115-116-071). All antibodies were diluted in PBS / 1.5% FCS. As a negative control, cells were incubated with PBS / 2% FCS instead of periplasmic extracts. Samples were measured by flow cytometry.

[0494] The regions recognized by the corresponding DLL3 binding domains are shown in the sequence listing (Table 18). Binders specific for epitopes located within the N-terminus, DSL domain and different EGF domains of DLL3 were mapped.

[0495] Figure 2 Two exemplary binders are shown that bind to a DLL3 epitope contained within the EGF-3 region (FACS signal is lost in the corresponding truncated DLL3 constructs without EGF-3). Figure 2 Also shown are exemplary binders that bind to a DLL3 epitope contained within the EGF-3 region (FACS signal lost in the corresponding truncated DLL3 construct without EGF-4).

[0496] Some binders are described as being specific for an epitope located within the region designated EGF-5 / [6]. Square brackets indicate specificity for a truncated DLL3 construct that leaves only the EGF-6 domain ( Figure 1 ), the FACS signal of the binder was reduced (i.e., neither completely present nor completely lost).

[0497] The bispecific DLL3xCD3 constructs used in the following examples were selected from those with "I2C" as CD3 binding domain and with a C-terminal fusion to wild-type human serum albumin, see e.g. SEQ ID NOs: 224-230, 233-235, 238-241.

[0498] Example 3

[0499] Biacore-based determination of antibody affinity to human and cynomolgus monkey DLL3

[0500] Biacore analysis experiments were performed using recombinant human / cyno DLL3-ECD fusion protein with chicken albumin to determine target binding of the antibody constructs of the invention.

[0501] In detail, a CM5 sensor chip (GE Healthcare) was fixed with approximately 600-800 RU of the corresponding recombinant antigen using acetate buffer (pH 4.5) according to the manufacturer's manual. The DLL3xCD3 bispecific antibody construct samples were loaded with a dilution series of the following concentrations: 50 nM, 25 nM, 12.5 nM, 6.25 nM and 3.13 nM diluted in HBS-EP running buffer (GE Healthcare). The flow rate was 30 μl / min for 3 minutes, followed by the application of HBS-EP running buffer at a flow rate of 30 μl / ml for 8 to 20 minutes. The regeneration of the chip was performed using 10 mM glycine, 10 mM NaCl (pH 1.5) solution. The data set was analyzed using BiaEval software. Two independent experiments were generally performed.

[0502] The DLL3xCD3 bispecific antibody constructs according to the invention showed very high affinities in the sub-nanomolar range for human DLL3 (except for DLL3-13, which had KD values ​​in the very low single-digit nanomolar range). Equilibrium binding to macaque DLL3 also showed affinities in a similar range. The affinity values ​​and the calculated affinity differences are shown in Table 2.

[0503] Table 2: Affinity of DLL3xCD3 bispecific antibody constructs to human and macaque DLL3 as determined by Biacore analysis, and calculated interspecies affinity differences.

[0504]

[0505] Furthermore, binding of the bispecific antibody constructs to human CD3 and cynomolgus CD3 was demonstrated in the Biacore assay in the low 2-digit nanomolar range (data not shown).

[0506] Example 4

[0507] Scatchard-based analysis of the affinity of the DLL3xCD3 bispecific antibody construct to human and macaque DLL3 on target antigen-positive cells and determination of interspecies affinity differences

[0508] The affinity of the DLL3xCD3 bispecific antibody constructs to CHO cells transfected with human or macaque DLL3 was also determined by Scatchard analysis, the most reliable method to measure potential affinity differences between human and macaque DLL3. For Scatchard analysis, saturation binding experiments were performed using a monovalent detection system to accurately determine the monovalent binding of the DLL3xCD3 bispecific antibody constructs to the respective cell lines.

[0509] 2×10 of the corresponding cell line 4 Each cell (CHO cell line expressing human DLL3 recombinantly, CHO cell line expressing macaque DLL3 recombinantly) was incubated with 50 μl of a three-fold dilution series starting at 10-20 nM (twelve dilutions at 1:2) of the corresponding DLL3xCD3 bispecific antibody construct (until saturation was reached), followed by incubation for 16 h at 4°C with stirring and one remaining washing step. Next, the cells were incubated for another hour with 30 μl of the CD3xALEXA488 conjugate solution. After one washing step, the cells were resuspended in 150 μl of FACS buffer containing 3.5% formaldehyde, incubated for another 15 minutes, centrifuged, resuspended in FACS buffer and analyzed using a FACS CantoII machine and FACS Diva software. Data were obtained from two independent experiments (each in triplicate). The corresponding Scatchard analysis was calculated to infer the maximum binding (Bmax). The concentration of the DLL3xCD3 bispecific antibody construct at half-maximal binding was determined, which reflects the corresponding KD. The values ​​from triplicate measurements were plotted as hyperbolic and sigmoidal curves to indicate the appropriate concentrations ranging from minimal to optimal binding.

[0510] The values ​​depicted in Table 3 are derived from two independent experiments for each DLL3xCD3 bispecific antibody construct. Cell-based Scatchard analysis demonstrated that the DLL3xCD3 bispecific antibody constructs of the invention had subnanomolar affinities for human DLL3 and mac DLL3, and presented a small cynomolgus / human species affinity difference of approximately 1.

[0511] Table 3: Affinity (KD) of DLL3xCD3 bispecific antibody constructs as determined in a cell-based Scatchard assay, and calculated affinity difference KD Cynomolgus DLL3 / KD Human DLL3. Antibody constructs were measured in two independent experiments (each in triplicate).

[0512]

[0513] Example 5

[0514] Bispecific binding and cross-species reactivity

[0515] To confirm binding to human DLL3 and CD3 as well as cyno DLL3 and CD3, the bispecific antibody constructs of the invention were tested by flow cytometry using the following cells:

[0516] CHO cells transfected with human DLL3, an artificial DLL3 isoform (characterized by the point mutations F172C and L218P) and macaque DLL3, respectively,

[0517] Human DLL3 positive human lung cancer cell line SHP-77,

[0518] The human T-cell leukemia cell line HPB-all (DSMZ, Braunschweig, ACC483) expressing CD3, and

[0519] ·LnPx 4119 T cell line expressing cynomolgus monkey CD3

[0520] For flow cytometry, 200,000 cells of the corresponding cell line were incubated with 50 μl of purified bispecific antibody constructs at a concentration of 5 μg / ml for 60 minutes at 4°C. The cells were washed twice with PBS / 2% FCS at 4°C and then incubated with internal mouse antibodies (2 μg / ml) specific for the CD3 binding portion of the bispecific antibody construct for 30 minutes. After washing, the bound mouse antibodies were detected with goat anti-mouse Fcγ-PE (1:100) at 4°C for 30 minutes. The samples were measured by flow cytometry. Untransfected CHO cells were used as negative controls.

[0521] The results are shown in Figure 3. The DLL3xCD3 bispecific antibody construct of the invention stained CHO cells transfected with human DLL3, artificial DLL3 isoforms and cyno DLL3, and it also stained the human DLL3 positive human lung cancer cell line SHP-77 (native expressor). Human and cyno T cell lines expressing CD3 were also recognized by the bispecific antibody construct. In addition, there was no staining of negative control cells (untransfected CHO, data shown in Example 6).

[0522] Example 6

[0523] Confirmation of the absence of binding to human paralogs

[0524] Human DLL3 paralogs DLL1 and DLL4 were stably transfected into CHO cells. The sequences of the paralogs used in the examples of the present invention are identified in the sequence listing (SEQ ID NO: 283 and 284). Protein expression was confirmed in FACS analysis with antibodies specific for the respective paralogs: for DLL4, the antibodies were anti-human DLL1 MAB1818 (R&D; 5 μg / ml) and anti-human DLL4 MAB1506 (R&D; 5 μg / ml).

[0525] Flow cytometry was performed as described in Example 5, except that goat anti-mouse FITC (1:100) was used to detect bound mouse antibodies. The results are shown in Figure 4 The analysis confirmed that none of the DLL3xCD3 bispecific antibody constructs of the invention cross-reacted with the human DLL3 paralogs DLL1 and DLL4.

[0526] Example 7

[0527] Identity with human lineage

[0528] To analyze the identity / similarity of the antibody constructs to the sequences of human antibody germline genes, the DLL3 binding domains of the present invention were aligned as follows: the complete VL including all CDRs was aligned; the complete VH including CDR 1 and 2 but not including CDR3 was aligned against the human antibody germline gene (Vbase). More details can be found in the specification of the present application. The results are shown in Table 4 below:

[0529] Table 4: VH and VL identities to human germline

[0530]

[0531]

[0532] Example 8

[0533] Cytotoxic activity

[0534] The potency of the DLL3xCD3 bispecific antibody constructs of the invention to redirect effector T cells against target cells expressing DLL3 was analyzed in five in vitro cytotoxicity assays:

[0535] • The potency of the DLL3xCD3 bispecific antibody construct to redirect stimulated human CD8+ effector T cells towards human DLL3-transfected CHO cells was measured in an 18-hour 51-chromium release assay.

[0536] The potency of the DLL3xCD3 bispecific antibody construct to redirect stimulated human CD8+ effector T cells against the DLL3-positive human lung cancer cell line SHP-77 was measured in an 18-hour 51-chromium release assay.

[0537] • The potency of the DLL3xCD3 bispecific antibody constructs to redirect T cells in unstimulated human PBMCs towards human DLL3-transfected CHO cells was measured in a 48-hour FACS-based cytotoxicity assay.

[0538] • The potency of the DLL3xCD3 bispecific antibody constructs to redirect T cells in unstimulated human PBMCs against the DLL3-positive human cell line SHP-77 was measured in a 48-hour FACS-based cytotoxicity assay.

[0539] • To confirm that the cross-reactive DLL3xCD3 bispecific antibody constructs were able to redirect macaque T cells against macaque DLL3-transfected CHO cells, a 48-hour FACS-based cytotoxicity assay was performed using a macaque T cell line as effector T cells.

[0540] Example 8.1

[0541] Chromium release assay using stimulated human T cells

[0542] CD8 enriched + T cells were stimulated. Culture dishes (145 mm diameter, Greiner bio-one GmbH, Kremsmünster) were coated with a commercial anti-CD3 specific antibody (OKT3, Orthoclone) at a final concentration of 1 μg / ml for 1 h at 37°C. Unbound proteins were removed by a washing step with PBS. 3–5 × 10 7 Individual PBMCs were added to 120 ml pre-coated culture dishes with stabilized glutamine / 10% FCS / IL-2 20 U / ml ( Chiron) in RPMI 1640 and stimulated for 2 days. On the third day, the cells were collected and washed once with RPMI 1640. IL-2 was added to a final concentration of 20 U / ml and the cells were cultured for another day in the same cell culture medium as above. Dynal-Beads was used according to the manufacturer's protocol to deplete CD4 + T cells and CD56 + NK cells make CD8 + Cytotoxic T lymphocyte (CTL) enrichment.

[0543] CHO target cells transfected with Cyno DLL3 or human DLL3 were washed twice with PBS and incubated with 11.1 MBq 51 Cr labeling for 60 minutes. Subsequently, the labeled target cells were washed 3 times with 5 ml RPMI and then used for cytotoxicity assay. The assay was performed in a 96-well plate in a total volume of 200 μl of supplemented RPMI with an E:T ratio of 10:1. Purified bispecific antibody constructs and their three-fold dilutions were used at a starting concentration of 0.01-1 μg / ml. The incubation time for the assay was 18 hours. Cytotoxicity was determined as the relative value of the chromium released in the supernatant relative to the difference between maximum lysis (addition of Triton-X) and spontaneous lysis (no effector cells). All measurements were performed in quadruplicate. The measurement of chromium activity in the supernatant was performed in a Wizard 3" gamma counter (Perkin Elmer Life Sciences GmbH, The results were analyzed using Prism 5 for Windows (version 5.0, GraphPad Software Inc., San Diego, California, USA). The cytotoxic activity was compared using the EC50 values ​​calculated from the sigmoidal dose-response curves by the analysis program.

[0544] Example 8.2

[0545] Efficacy of redirecting stimulated human effector T cells against CHO cells transfected with human DLL3

[0546] In 51-Cr( 51 The cytotoxic activity of the DLL3xCD3 bispecific antibody constructs according to the invention was analyzed in a CHO cell (C) release cytotoxicity assay using CHO cells transfected with human DLL3 as target cells and stimulated human CD8+ T cells as effector cells. The experiment was performed as described in Example 8.1.

[0547] The results are shown in Table 5. The DLL3xCD3 bispecific antibody constructs showed very potent cytotoxic activity against human DLL3-transfected CHO cells in the single picomolar range.

[0548] Table 5: 51-Chromium ( 51 Cr) EC50 values ​​[pM] of the DLL3xCD3 bispecific antibody constructs analyzed in the release cytotoxicity assay.

[0549]

[0550] Example 8.3

[0551] Efficacy of redirecting stimulated human effector T cells against the DLL3-positive human lung cancer cell line SHP-77

[0552] In 51-Cr( 51 The cytotoxic activity of the DLL3xCD3 bispecific antibody constructs was analyzed in a Cr release cytotoxicity assay using the DLL3-positive human lung cancer cell line SHP-77 as a source of target cells and stimulated human CD8+ T cells as effector cells. The assay was performed as described in Example 8.1.

[0553] According to the results of 51-chromium release assay using stimulated enriched human CD8+ T lymphocytes as effector cells and CHO cells transfected with human DLL3 as target cells, the DLL3xCD3 bispecific antibody constructs of the present invention also have potent cytotoxic activity against naturally expressing target cells (see Table 6).

[0554] Table 6: 18-hour 51-chromium ( 51 Cr) EC50 values ​​[pM] of the DLL3xCD3 bispecific antibody constructs analyzed in the release cytotoxicity assay.

[0555] Rows 1-10: Antibody constructs according to the invention that bind to a DLL3 epitope contained within the region depicted as SEQ ID NO: 260 (Rows 1-7: Antibody constructs that bind to a DLL3 epitope contained within the EGF-3 region. Rows 8-10: Antibody constructs that bind to a DLL3 epitope contained within the EGF-4 region). Rows 11-14: Antibody constructs that bind to a DLL3 epitope contained within EGF-5 / [EGF-6].

[0556]

[0557] Example 8.4

[0558] FACS-based cytotoxicity assay using unstimulated human PBMCs

[0559] Isolation of effector cells

[0560] By Ficoll density gradient centrifugation, human peripheral blood mononuclear cells (PBMC) are prepared from the lymphocyte preparation (buffy coat) (byproduct of the blood bank collecting blood for infusion). Buffy coat is provided by local blood bank and prepares PBMC on the same day with blood collection. After Ficoll density centrifugation and fully washing with Dulbecco's PBS (Dulbecco's PBS), remaining red blood cells are removed from PBMC by incubation with red blood cell lysis buffer (155mM NH4Cl, 10mM KHCO3, 100 μM EDTA). After centrifugation of PBMC at 100x g, platelets are removed via supernatant. Remaining lymphocytes mainly include B and T lymphocytes, NK cells and monocytes. PBMC are maintained at 37°C / 5% CO2 in RPMI culture medium (Gibco) with 10% FCS (Gibco) to cultivate.

[0561] Consume CD14 + and CD56 + cell

[0562] For CD14 + For depletion of NK cells, human CD14 MicroBeads (Milteny Biotec, MACS, #130-050-201) were used, and for depletion of NK cells, human CD56 MicroBeads (MACS, #130-050-401) were used. PBMCs were counted and centrifuged at 300 x g for 10 minutes at room temperature. The supernatant was discarded and the cell pellet was resuspended in MACS separation buffer [80 μL / 10 7 cells; PBS (Invitrogen, #20012-043), 0.5% (v / v) FBS (Gibco, #10270-106), 2 mM EDTA (Sigma-Aldrich, #E-6511). CD14 MicroBeads and CD56 MicroBeads (20 μL / 10 7 cells) and incubate at 4-8°C for 15 minutes. 7 After centrifugation (see above), the supernatant was discarded and the cells were resuspended in MACS separation buffer (500 μL / 10 8Cells). CD14 / CD56 negative cells were then isolated using LS columns (Miltenyi Biotec, #130-042-401). PBMCs without CD14+ / CD56+ cells were cultured in RPMI complete medium, RPMI1640 (BiochromAG, #FG1215) supplemented with 10% FBS (Biochrom AG, #S0115), 1x non-essential amino acids (Biochrom AG, #K0293), 10 mM Hepes buffer (Biochrom AG, #L1613), 1 mM sodium pyruvate (Biochrom AG, #L0473) and 100 U / mL penicillin / streptomycin (Biochrom AG, #A2213) at 37°C in an incubator until needed.

[0563] Target cell labeling

[0564] For analysis of cell lysis by flow cytometry, the fluorescent membrane dye DiOC 18 CHO cells transfected with human DLL3 or macaque DLL3 were labeled with DiO (Molecular Probes, #V22886) as target cells and distinguished from effector cells. Briefly, cells were harvested, washed once with PBS and incubated in PBS containing 2% (v / v) FBS and the membrane dye DiO (5 μL / 10 μL). 6 cells) in PBS adjusted to 10 6 After incubation at 37°C for 3 minutes, the cells were washed twice in complete RPMI medium and the cell number was adjusted to 1.25 × 10 5 Cell viability was determined using 0.5% (v / v) isotonic Eosin G solution (Roth, #45380).

[0565] Flow cytometry-based analysis

[0566] This assay is designed to quantify the lysis of CHO cells transfected with cyno or human DLL3 in the presence of serial dilutions of DLL3 bispecific antibody constructs. +160 μl of this suspension was transferred to each well of a 96-well plate. 40 μL of serial dilutions of the DLL3xCD3 bispecific antibody construct and a negative control bispecific antibody (a CD3-based bispecific antibody construct that recognizes an unrelated target antigen) or RPMI complete medium were added as additional negative controls. The bispecific antibody-mediated cytotoxicity reaction was carried out in a 7% CO2 humidified incubator for 48 hours. The cells were then transferred to a new 96-well plate and the loss of target cell membrane integrity was monitored by adding propidium iodide (PI) at a final concentration of 1 μg / mL. PI is a membrane-impermeable dye that is usually excluded by living cells, while dead cells absorb it and become identifiable by fluorescence emission.

[0567] Samples were measured on a FACSCanto II instrument by flow cytometry and analyzed by FACSDiva software (both from Becton Dickinson). Target cells were identified as DiO positive cells. PI negative target cells were classified as live target cells. The percentage of cytotoxicity was calculated according to the following formula:

[0568]

[0569] n = number of events

[0570] The percentage of cytotoxicity was plotted against the corresponding bispecific antibody concentration using GraphPad Prism 5 software (Graph Pad Software, San Diego). The dose-response curves were analyzed using a four-parameter logistic regression model to estimate a sigmoidal dose-response curve with a fixed hill slope and to calculate EC50 values.

[0571] Example 8.5

[0572] Efficacy of redirecting unstimulated human PBMCs toward CHO cells transfected with human DLL3

[0573] The cytotoxic activity of the DLL3xCD3 bispecific antibody constructs was analyzed in a FACS-based cytotoxicity assay using CHO cells transfected with human DLL3 as target cells and unstimulated human PBMCs as effector cells. The assay was performed as described in Example 8.4 above.

[0574] The results of FACS-based cytotoxicity assays using unstimulated human PBMCs as effector cells and CHO cells transfected with human DLL3 as targets are shown in Table 7.

[0575] Table 7: EC50 values ​​[pM] of DLL3xCD3 bispecific antibody constructs measured in a 48 h FACS-based cytotoxicity assay with unstimulated human PMBCs as effector cells and CHO cells transfected with human DLL3 as target cells.

[0576] Rows 1-10: Antibody constructs according to the invention that bind to a DLL3 epitope contained within the region depicted as SEQ ID NO: 260 (Rows 1-7: Antibody constructs that bind to a DLL3 epitope contained within the EGF-3 region. Rows 8-10: Antibody constructs that bind to a DLL3 epitope contained within the EGF-4 region). Rows 11-14: Antibody constructs that bind to a DLL3 epitope contained within the EGF-5 / [EGF-6] region.

[0577]

[0578]

[0579] It is expected that EC50 values ​​will generally be higher in cytotoxicity assays using unstimulated PBMCs as effector cells compared to cytotoxicity assays using stimulated human CD8+ T cells (see Example 8.2).

[0580] Example 8.6

[0581] Efficacy of redirecting unstimulated human PBMCs toward the DLL3-positive lung cancer cell line SHP-77

[0582] In addition, the cytotoxic activity of the DLL3xCD3 bispecific antibody constructs was analyzed in a FACS-based cytotoxicity assay using the DLL3-positive human lung cancer cell line SHP-77 as a source of target cells and unstimulated human PBMCs as effector cells. The assay was performed as described in Example 8.4 above. The results are shown in Table 8.

[0583] Table 8: EC50 values ​​[pM] of DLL3xCD3 bispecific antibody constructs measured in a 48 h FACS-based cytotoxicity assay using unstimulated human PBMCs as effector cells and the human cell line SHP-77 as a source of target cells.

[0584] Rows 1-10: Antibody constructs according to the invention that bind to a DLL3 epitope contained within the region depicted as SEQ ID NO: 260 (Rows 1-7: Antibody constructs that bind to a DLL3 epitope contained within the EGF-3 region. Rows 8-10: Antibody constructs that bind to a DLL3 epitope contained within the EGF-4 region). Rows 11-14: Antibody constructs that bind to a DLL3 epitope contained within the EGF-5 / [EGF-6] region.

[0585]

[0586] Example 8.7

[0587] Efficacy of redirecting macaque T cells against CHO cells expressing macaque DLL3

[0588] Finally, the cytotoxic activity of the DLL3xCD3 bispecific antibody constructs was analyzed in a FACS-based cytotoxicity assay using CHO cells transfected with cynomolgus (cyno) DLL3 as target cells and the cynomolgus T cell line 4119LnPx (Knappe et al. Blood 95:3256-61 (2000)) as a source of effector cells. Target cell labeling and flow cytometry-based analysis of cytotoxic activity of CHO cells transfected with cynomolgus DLL3 were performed as described above.

[0589] The results are shown in Table 9. Cynomolgus T cells from the cell line 4119LnPx were induced to effectively kill CHO cells transfected with cynomolgus DLL3 by the DLL3xCD3 bispecific antibody constructs according to the invention. In this assay, the antibody constructs were potent with 2-digit picomolar EC50 values, confirming that they were extremely active in the cynomolgus system.

[0590] Table 9: EC50 values ​​[pM] of DLL3xCD3 bispecific antibody constructs measured in a 48 h FACS-based cytotoxicity assay with the macaque T cell line 4119LnPx as effector cells and CHO cells transfected with macaque DLL3 as target cells.

[0591] Rows 1-10: Antibody constructs according to the invention that bind to a DLL3 epitope contained within the region depicted as SEQ ID NO: 260 (Rows 1-7: Antibody constructs that bind to a DLL3 epitope contained within the EGF-3 region. Rows 8-10: Antibody constructs that bind to a DLL3 epitope contained within the EGF-4 region). Rows 11-14: Antibody constructs that bind to a DLL3 epitope contained within the EGF-5 / [EGF-6] region.

[0592]

[0593] Example 9

[0594] Conversion of monomers to dimers after (i) three freeze / thaw cycles and (ii) 7 days of incubation at 250 μg / ml

[0595] Bispecific DLL3xCD3 antibody monomer constructs were subjected to different stress conditions and then high performance SEC to determine the percentage of the initial monomeric antibody construct that had been converted to a dimeric antibody construct.

[0596] (i) 25 μg of monomeric antibody construct was adjusted to a concentration of 250 μg / ml with universal formulation buffer, then frozen at -80°C for 30 minutes and then thawed at room temperature for 30 minutes. After three freeze / thaw cycles, the dimer content was determined by HP-SEC.

[0597] (ii) 25 μg of monomeric antibody construct was adjusted to a concentration of 250 μg / ml with universal formulation buffer before incubation at 37° C. for 7 days. Dimer content was determined by HP-SEC.

[0598] A high-resolution SEC column, TSK Gel G3000 SWXL (Tosoh, Tokyo-Japan), was connected to an A905 autosampler. Purifier 10FPLC (GE Lifesciences). Column equilibration and running buffer consisted of 100 mM KH2PO4-200 mM Na2SO4 (adjusted to pH 6.6). Antibody solution (25 μg protein) was applied to the equilibrated column and eluted at a flow rate of 0.75 ml / min at a maximum pressure of 7 MPa. The entire run was monitored by absorbance at 280, 254 and 210 nm. The analysis was performed by peak integration of the 210 nm signal recorded in the evaluation table run by Unicorn software. The dimer content was calculated by dividing the area of ​​the dimer peak by the total area of ​​the monomer plus dimer peak.

[0599] The results are shown in the following Table 10. The DLL3xCD3 bispecific antibody constructs of the invention exhibited a dimer percentage of 0.0% after three freeze / thaw cycles and a dimer percentage of ≤ 2% after 7 days of incubation at 37°C.

[0600] Table 10: Percentage of monomeric versus dimeric DLL3xCD3 bispecific antibodies as determined by high performance size exclusion chromatography (HP-SEC).

[0601]

[0602]

[0603] Example 10

[0604] Thermal stability

[0605] The antibody aggregation temperature is determined as follows: 40 μl of a 250 μg / ml antibody construct solution is transferred to a disposable cuvette and placed in a Wyatt dynamic light scattering device DynaPro Nanostar (Wyatt). The sample is heated from 40°C to 70°C at a heating rate of 0.5°C / min, with the measured radius continuously acquired. The software package delivered with the DLS device uses the increase in radius indicating protein melting and aggregation to calculate the aggregation temperature of the antibody construct.

[0606] All tested DLL3xCD3 bispecific antibody constructs of the invention showed a thermostability with an aggregation temperature ≥ 45°C, as shown in the following Table 11. The group of antibody constructs binding to the epitope of DLL3 contained within the EGF-4 region (as depicted by SEQ ID NO: 259) even had a thermostability of ≥ 50°C, more precisely ≥ 54°C.

[0607] Table 11: Thermal stability of bispecific antibody constructs as determined by DLS (dynamic light scattering).

[0608]

[0609]

[0610] Embodiment 11

[0611] Stability after 24 h incubation in human plasma

[0612] Purified bispecific antibody constructs were incubated in human plasma pools at a ratio of 1:5 at a final concentration of 2-20 μg / ml for 96 hours at 37°C. After plasma incubation, the antibody constructs were compared in a 51-chromium release assay using stimulated enriched human CD8+ T cells and human DLL3 transfected CHO cells at a starting concentration of 0.01-0.1 μg / ml and an effector cell to target cell (E:T) ratio of 10:1 (assay as described in Example 8.1). Unincubated freshly thawed bispecific antibody constructs were included as controls.

[0613] The results are shown in Table 12 below; exemplary results for two antibody constructs DLL-4 and DLL-14 are also shown in Figure 5 All tested antibody constructs had very favorable plasma stability (EC 50 Plasma / EC 50 The group of antibody constructs binding to the epitope of DLL3 contained within the region of EGF-4 (as depicted by SEQ ID NO: 259) even had a plasma stability of ≤ 1.5, more precisely ≤ 1.1.

[0614] Table 12: EC50 values ​​and calculated plasma / control values ​​for antibody constructs with and without plasma incubation

[0615]

[0616]

[0617] Example 12

[0618] Turbidity at 2500μg / ml antibody concentration

[0619] 1 ml of the purified antibody construct solution at a concentration of 250 μg / ml was concentrated to 2500 μg / ml by rotating the concentrator unit.After storage at 5°C for 16 hours, the turbidity of the antibody solution was determined by OD340 nm optical absorption measurement against universal formulation buffer.

[0620] The results are shown in Table 13 below. All tested antibody constructs had a very favorable turbidity of ≤ 0.1, except for one construct that had a turbidity slightly above 0.1. The group of antibody constructs that bind to the epitope of DLL3 contained within the EGF-4 region (as depicted by SEQ ID NO: 259) even had a turbidity of ≤ 0.08.

[0621] Table 13: Turbidity of antibody constructs after concentration to 2.5 mg / ml overnight

[0622]

[0623] Embodiment 13

[0624] Protein homogeneity by high-resolution cation exchange chromatography

[0625] The protein homogeneity of the antibody constructs of the invention was analyzed by high resolution cation exchange chromatography CIEX.

[0626] 50 μg of antibody construct monomer was diluted with 50 ml of binding buffer A (20 mM sodium dihydrogen phosphate, 30 mM NaCl, 0.01% sodium octanoate, pH 5.5) and 40 ml of this solution was applied to the Micro FPLC device (GE Healthcare, Germany) 1ml BioPro SP-F column (YMC, Germany). After sample binding, a washing step was performed with binding buffer. For protein elution, a linear increasing salt gradient of 50% buffer B was applied using buffer B (20mM sodium dihydrogen phosphate, 1000mM NaCl, 0.01% sodium octanoate, pH 5.5) over 10 column volumes. The entire run was monitored by absorbance at 280, 254 and 210nm. The Unicorn software runs the analysis by peak integration of the 280 nm signal recorded in the evaluation table.

[0627] The results are shown in the following Table 14. All antibody constructs tested had a very favorable homogeneity of ≥80% (area under the curve (= AUC) of the main peak). The group of antibody constructs binding to the epitope of DLL3 contained within the EGF-3 region (as depicted by SEQ ID NO: 258) even had a homogeneity of ≥90%.

[0628] Table 14: Protein homogeneity of antibody constructs (AUC % of main peak)

[0629]

[0630]

[0631] Embodiment 14

[0632] Surface hydrophobicity as measured by HIC Butyl

[0633] The surface hydrophobicity of the bispecific antibody constructs of the invention was tested in hydrophobic interaction chromatography HIC in flow-through mode.

[0634] 50 μg of antibody construct monomer was diluted to a final volume of 500 μl with universal formulation buffer (10 mM citric acid, 75 mM lysine HCl, 4% trehalose, pH 7.0) and applied to the Purifier FPLC system (GE Healthcare, Germany) with a 1 ml Butyl Sepharose FF column (GE Healthcare, Germany). The absorbance at 280, 254 and 210 nm was monitored throughout the run. The Unicorn software runs the analysis by peak integration of the 280 nm signal recorded in the evaluation table. The elution behavior is evaluated by comparing the area and the rate of rise and fall of the protein signal, thereby indicating the strength of the interaction of the BiTE albumin fusion with the matrix.

[0635] The antibody constructs had good elution properties, most of which were rapid and complete.

[0636] Embodiment 15

[0637] Potency differences between monomeric and dimeric isoforms of bispecific antibody constructs

[0638] To determine the difference in cytotoxic activity between monomer and dimer isoforms of each DLL3xCD3 bispecific antibody construct (referred to as potency difference), an 18-hour 51-chromium release cytotoxicity assay (Example 8.1) was performed with purified bispecific antibody construct monomers and dimers as described above. Effector cells were stimulated enriched human CD8+ T cells. Target cells were CHO cells transfected with hu DLL3. The ratio of effector cells to target cells (E:T) was 10:1. Potency difference was calculated as the ratio between EC50 values.

[0639] The results are shown in Table 15 below; exemplary results for two antibody constructs DLL-4 and DLL-14 are also shown in Figure 6 The potency differences of the tested DLL3xCD3 bispecific antibody constructs ranged between 0.2 and 1.0. Thus, there were no dimers that were substantially more active than their corresponding monomers.

[0640] Table 15: Potency differences between monomeric and dimeric isoforms

[0641]

[0642] Example 16

[0643] In vitro internalization assay

[0644] The change in potency of the DLL3xCD3 bispecific antibody construct was measured as a function of pre-incubation of the construct on target cells in the absence of T cells. If the antibody construct is internalized, it should undergo lysosomal degradation. The effective concentration should decrease over time, and therefore the apparent potency should also decrease. This effect was observed with other targets, for which this is a known phenomenon, but no effect was observed with the DLL3xCD3 bispecific antibody construct. The assay was performed as follows:

[0645] T cells were counted and diluted to 1 × 10 in assay medium. 5SHP-77 cells were counted and plated at 2500 cells / well (cpw). Antibody constructs were serially diluted 1:2 (with Bravo) at a starting concentration of 100 nM. Antibody constructs were added to the culture assay plates and allowed to incubate for 0, 1, or 2 hours before T cells were added. T cells were then plated at 25,000 cpw and the analytes were incubated at 37°C for 48 hours. The SHP-77 cell survival was analyzed using the ELISA system (25 μl / well). The results are shown in Figure 7 There was no suggestion of significant internalization of the antibody construct DLL3-4 x CD3(I2C).

[0646] Embodiment 17

[0647] Shedding studies

[0648] To analyze whether the cytotoxic activity of the DLL3xCD3 bispecific antibody constructs of the invention is significantly impaired by the presence of shed DLL, the following assay was performed: T cells were counted and diluted to 1×10 5 SHP-77 cells were counted and diluted to 1.25×10 cells / ml in assay medium with increasing concentrations of soluble DLL3 between 0.3 nM and 12 nM. 5 SHP-77 cells were plated at 2500 cells / well (cpw) and T cells were added at 25000 cpw. Antibody constructs were serially diluted 1:2 (with Bravo) and added to the culture assay (with Bravo). Incubated at 37°C for 48 hours. The SHP-77 cell survival was analyzed using the ELISA system (25 μl / well).

[0649] Embodiment 18

[0650] Mouse xenograft efficacy studies

[0651] The anti-tumor activity of the HLE DLL3xCD3 bispecific antibody construct (SEQ ID NO: 517) was tested in a model of female NOD / SCID mice injected subcutaneously with 5×10 6 Individual DLL3-positive SCLC (SHP-77luc) or 5×10 6 Individual DLL3-positive melanoma (WM 266-4) cells were injected intraperitoneally on day 12 with effector cells (2 × 10 7In vitro expanded and activated live human CD3+ T cells). Treatment started on day 16 (WM 266-4) or day 18 (SHP-77luc). Antibody constructs were administered four times every five days (q5d×4) by intravenous push. Treatment groups were as follows:

[0652] –SCLC model (SHP-77luc) / 7 mice / group

[0653] Vehicle-treated group with T cells

[0654] DLL3xCD3 bispecific antibody construct: 10 mg / kg per administration

[0655] –Melanoma model (WM266-4) / 9 mice / group

[0656] Vehicle-treated group with T cells

[0657] DLL3xCD3 bispecific antibody construct: 10 mg / kg per administration

[0658] DLL3xCD3 bispecific antibody construct: 2 mg / kg per administration

[0659] Tumors were measured by caliper during the study and progression was assessed by comparing tumor volumes (TV) between groups. Tumor growth inhibition T / C [%] on day x was determined by calculating tumor volume as T / C (%) = 100 × (median TV of analysis group) / (median TV of control group), and the calculated values ​​are depicted in the following Table 16:

[0660] Table 16: T / C values ​​for mouse xenograft studies with SHP-77luc cells and WM266-4 cells

[0661]

[0662] In addition, the results are shown in Fig. 8A and Figure 8B Significant tumor growth inhibition was demonstrated in both tumor models and at both tested dose levels of 2 mg / kg and 10 mg / kg.

[0663] Embodiment 19

[0664] Exploratory toxicology study in cynomolgus monkeys

[0665] An exploratory toxicology study was conducted with a non-half-life extended DLL3xCD3 bispecific antibody construct (SEQ ID NO: 554). Three female cynomolgus monkeys were dosed via continuous intravenous infusion for 16 days (5, 15, and 45 μg / kg / day for 3 days each, followed by 100 μg / kg / day for 7 days). No clinical observations, body temperature, food consumption, or body weight changes related to the test article were observed.

[0666] Consistent with expectations for the pharmacology of the DLL3xCD3 bispecific antibody construct, circulating T lymphocyte populations (total T-lymphocytes, T-helper and T-cytotoxic lymphocytes, NK cells, B-lymphocytes, and CD25+ activated T-lymphocytes) decreased on the first day of dosing and remained low throughout the duration of the study in all animals. Activation markers (CD69 and CD25 on activated T cells) increased on day 1, but not at subsequent time points in the study.

[0667] In summary, the DLL3xCD3 bispecific antibody construct showed promising results even at the highest dose tested (100 μg / kg / day, >300×EC 50 ) also tolerates very well.

[0668] Embodiment 20

[0669] PK study in cynomolgus monkeys

[0670] The PK study in cynomolgus monkeys was performed with naive male cynomolgus monkeys. Three different albumin-fused DLL3xCD3 (I2C) bispecific antibody constructs (DLL3-4, DLL3-6 and DLL3-14) were administered as a single intravenous bolus at a concentration of 12 μg / kg. Two animals were used per group for each molecule.

[0671] A separate cynomolgus monkey PK study was performed under the same conditions but using DLL3-4 in a different Fc fusion format.

[0672] Blood samples were collected before dosing and at 0.05, 0.5, 1, 4, 8, 24, 48, 72, 120, 168, 240 and 336 hours after dosing. Serum was prepared for determination of serum concentrations of the molecules in the immunoassay. The assay was performed by capturing the antibody construct via the target portion of the antibody construct, while antibodies to the CD3 binding portion of the construct were used for detection. PK parameters were determined using serum concentration-time curves. Pharmacokinetic parameters were determined using standard non-compartmental analysis (NCA) methods. The following PK parameters were evaluated: AUC inf (area under the serum concentration-time curve), V ss (steady-state volume of distribution), CL (systemic clearance), and terminal t 1 / 2(Terminal half-life). For all antibody constructs, serum levels were quantifiable in all animals at all time points after their administration. No clinical observations were performed in any of the treated animals.

[0673] The pharmacokinetics of the tested antibody constructs are shown in Figure 9 and the PK parameters are summarized in Table 17 below as mean values ​​for n=2.

[0674] Albumin-fused constructs showed favorable PK profiles consistent with a once- or twice-weekly dosing schedule in human patients. Even more favorable PK profiles supporting once-weekly dosing or even every other week dosing were observed with Fc-fused constructs (scFc).

[0675] Table 17: Pharmacokinetic parameters of HLE variants of DLL3xCD3 bispecific antibody constructs in cynomolgus monkeys

[0676]

[0677]

[0678]

[0679]

[0680]

[0681]

[0682]

[0683]

[0684]

[0685]

[0686]

[0687]

[0688]

[0689]

[0690]

[0691]

[0692]

[0693]

[0694]

[0695]

[0696]

[0697]

[0698]

[0699]

[0700]

[0701]

[0702]

[0703]

[0704]

[0705]

[0706]

[0707]

[0708]

[0709]

[0710]

[0711]

[0712]

[0713]

[0714]

[0715]

[0716]

[0717]

[0718]

[0719]

[0720]

[0721]

[0722]

[0723]

[0724]

[0725]

[0726]

[0727]

[0728]

[0729]

[0730]

[0731]

[0732]

[0733]

[0734]

[0735]

[0736]

[0737]

[0738]

[0739]

[0740]

[0741]

[0742]

[0743]

[0744]

[0745]

[0746]

[0747]

[0748]

[0749]

[0750]

[0751]

[0752]

[0753]

[0754]

[0755]

[0756]

[0757]

[0758]

[0759]

[0760]

[0761]

[0762]

[0763]

[0764]

[0765]

[0766]

[0767]

[0768]

[0769]

[0770]

[0771]

[0772]

[0773]

[0774]

[0775]

[0776]

[0777]

[0778]

[0779]

[0780]

[0781]

[0782]

[0783]

[0784]

[0785]

[0786]

[0787]

[0788]

[0789]

[0790]

[0791]

[0792]

[0793]

[0794]

[0795]

[0796]

[0797]

[0798]

[0799]

[0800]

[0801]

[0802]

[0803]

[0804]

[0805]

[0806]

[0807]

[0808]

[0809]

[0810]

[0811]

[0812]

[0813]

[0814]

[0815]

[0816]

[0817]

[0818]

[0819]

[0820]

Claims

1. A bispecific antibody construct comprising a first binding domain that binds to human DLL3 on the surface of a target cell and a second binding domain that binds to human CD3 on the surface of a T cell, wherein the first binding domain binds to the region of DLL3 contained within SEQ ID NO: 260, and wherein the first binding domain comprises a VH region and a VL region, the VH region comprising CDR-H1, CDR-H2 and CDR-H3 selected from the following items, and the VL region comprising CDR-L1, CDR-L2 and CDR-L3 selected from the following items: a) a CDR-H1 comprising the amino acid sequence of SEQ ID NO: 171, a CDR-H2 comprising the amino acid sequence of SEQ ID NO: 172, a CDR-H3 comprising the amino acid sequence of SEQ ID NO: 173, a CDR-L1 comprising the amino acid sequence of SEQ ID NO: 174, a CDR-L2 comprising the amino acid sequence of SEQ ID NO: 175, and a CDR-L3 comprising the amino acid sequence of SEQ ID NO: 176; b) a CDR-H1 comprising the amino acid sequence of SEQ ID NO: 181, a CDR-H2 comprising the amino acid sequence of SEQ ID NO: 182, a CDR-H3 comprising the amino acid sequence of SEQ ID NO: 183, a CDR-L1 comprising the amino acid sequence of SEQ ID NO: 184, a CDR-L2 comprising the amino acid sequence of SEQ ID NO: 185, and a CDR-L3 comprising the amino acid sequence of SEQ ID NO: 186; c) a CDR-H1 comprising the amino acid sequence of SEQ ID NO: 191, a CDR-H2 comprising the amino acid sequence of SEQ ID NO: 192, a CDR-H3 comprising the amino acid sequence of SEQ ID NO: 193, a CDR-L1 comprising the amino acid sequence of SEQ ID NO: 194, a CDR-L2 comprising the amino acid sequence of SEQ ID NO: 195, and a CDR-L3 comprising the amino acid sequence of SEQ ID NO: 196; and d) CDR-H1 comprising the amino acid sequence of SEQ ID NO:201, CDR-H2 comprising the amino acid sequence of SEQ ID NO:202, CDR-H3 comprising the amino acid sequence of SEQ ID NO:203, CDR-L1 comprising the amino acid sequence of SEQ ID NO:204, CDR-L2 comprising the amino acid sequence of SEQ ID NO:205 and CDR-L3 comprising the amino acid sequence of SEQ ID NO:

206.

2. The antibody construct of claim 1, wherein the first binding domain further binds to macaque DLL3.

3. The antibody construct of claim 1, wherein the second binding domain binds to human CD3ε and binds to rhesus CD3ε.

4. The antibody construct of claim 1, wherein the first binding domain comprises a VH region selected from the group consisting of SEQ ID NO: 177, SEQ ID NO: 187, SEQ ID NO: 197, and SEQ ID NO:

207.

5. The antibody construct of claim 4, wherein the first binding domain comprises a VL region selected from SEQ ID NO: 178, SEQ ID NO: 188, SEQ ID NO: 198 and SEQ ID NO:

208.

6. The antibody construct of claim 5, wherein the first binding domain comprises a VH region and a VL region, wherein the VH region and the VL region comprise an amino acid sequence pair selected from the following: SEQ ID NO: 177 and 178; SEQ ID NO: 187 and 188; SEQ ID NO: 197 and 198; and SEQ ID NO: 207 and 208, respectively.

7. The antibody construct of claim 6, wherein the first binding domain comprises a polypeptide selected from the group consisting of: SEQ ID NO: 179, SEQ ID NO: 189, SEQ ID NO: 199 and SEQ ID NO:

209.

8. The antibody construct according to any one of claims 1 to 7, comprising a polypeptide selected from the group consisting of: SEQ ID NO: 180, SEQ ID NO: 190, SEQ ID NO: 200 and SEQ ID NO:

210.

9. The antibody construct according to claim 1, wherein the antibody construct is in the form of (scFv)2.

10. The antibody construct according to claim 1, wherein the antibody construct is in the form of a scFv-single domain mAb.

11. The antibody construct of claim 1, wherein the first domain, the second domain, or both are in Fab format.

12. A bispecific antibody construct comprising a first binding domain that binds to human DLL3 on the surface of a target cell and a second binding domain that binds to human CD3 on the surface of a T cell, wherein the first binding domain comprises a VH region and a VL region, the VH region comprising CDR-H1, CDR-H2 and CDR-H3, the VL region comprising CDR-L1, CDR-L2 and CDR-L3, wherein the CDR-H1, CDR-H2 and CDR-H3 comprise the amino acid sequences of SEQ ID NO:31, SEQ ID NO:32 and SEQ ID NO:33, respectively, and wherein the CDR-L1, CDR-L2 and CDR-L3 comprise the amino acid sequences of SEQ ID NO:34, SEQ ID NO:35 and SEQ ID NO:36, respectively.

13. The antibody construct of claim 12, wherein the VH of the first domain comprises a sequence at least 90% identical to SEQ ID NO:435, and wherein the VL of the first domain comprises a sequence at least 90% identical to SEQ ID NO:

436.

14. The antibody construct of claim 11 or 12, wherein the VH of the first domain comprises a sequence at least 90% identical to SEQ ID NO: 435, and wherein the VL of the first domain comprises a sequence at least 94% identical to SEQ ID NO:

436.

15. A polynucleotide encoding the antibody construct of any one of claims 1 to 14. A vector comprising the polynucleotide according to claim 15 .

17. A host cell transformed or transfected with the polynucleotide according to claim 15 or with the vector according to claim 16.

18. A method for producing an antibody construct according to any one of claims 1 to 14, the method comprising culturing the host cell according to claim 17 under conditions allowing expression of the antibody construct, and recovering the produced antibody construct from the culture.

19. A pharmaceutical composition comprising the antibody construct according to any one of claims 1 to 14 or the antibody construct produced according to the method of claim 18 and one or more carriers.

20. A kit comprising the pharmaceutical composition according to claim 19.

21. The kit of claim 20, further comprising instructions for preventing, treating or ameliorating a tumor or cancer expressing DLL3.

22. Use of an antibody construct according to any one of claims 1 to 14 in the preparation of a medicament for treating a tumor, a cancer or a metastatic cancer in a subject in need thereof, wherein the tumor, the cancer or the metastatic cancer expresses DLL3.

23. The use according to claim 22, wherein the tumor or cancer disease is (a) a tumor or cancer of the lung, breast, cervix, colon, colorectal, endometrium, head and neck, liver, ovary, pancreas, prostate, skin, stomach, testis, thyroid, adrenal gland, kidney, bladder, uterus, esophagus, urothelium or brain; (b) a lymphoma, carcinoma or sarcoma; or (c) a metastatic cancer disease derived from any of the foregoing.

24. The use according to claim 23, wherein the tumor or cancer is: SCLC, NSCLC, glioma, glioblastoma, melanoma, neuroendocrine prostate cancer, neuroendocrine pancreatic cancer, hepatoblastoma and hepatocellular carcinoma or a metastatic cancer disease derived from any of the foregoing.

Images