NZ769422A

NZ769422A - Dll3-cd3 bispecific antibodies

Info

Publication number: NZ769422A
Application number: NZ769422A
Authority: NZ
Inventors: Paul Adam; Michael Dziegelewski; Rajkumar Ganesan; Philip Nicholas Gorman; Pankaj Gupta; Priyanka Gupta; Susanne Hipp; Justin Scheer; Vladimir H Voynov
Original assignee: Boehringer Ingelheim International Gmbh
Priority date: 2018-06-09
Filing date: 2019-06-07

Abstract

The present disclosure relates to DLL3-CD3 bispecific antibodies, characterized by their sequences. The application further relates to uses of the antibodies and their compositions, in particular for therapeutic purposes in the field of cancer diseases.

Description

DLL3-CD3 BISPECIFIC ANTIBODIES OUND OF THE INVENTION TECHNICAL FIELD The present invention relates to multi-speciﬁc binding proteins comprising a ﬁrst antigen binding unit speciﬁc for DLL3 and a second antigen binding unit speciﬁc for CD3. The ion also relates to nucleic acids encoding such g proteins, to methods for ing such g proteins; host cells expressing or capable of expressing such binding proteins, compositions comprising such binding ns and to uses of such binding proteins or such compositions, in particular for therapeutic purposes in the ﬁeld of cancer diseases.

BACKGROUND INFORMATION DLL3 is a type I transmembrane protein that belongs to the Notch ligand family. DLL3 is primarily involved in somitogenesis, where it predominantly is localized in the Golgi apparatus and acts as an inhibitor ofNotch signalling, in contrast to the other DLL family members, DLLl and DLL4, which are localized on the cell surface. Only in cancer cells, where DLL3 is overexpressed, some DLL3 molecules escape to the cell surface (Dylla, Molecular & Cellular Oncology 2016, Vol. 3, No. 2). DLL3 has been identiﬁed as a tumor —speciﬁc target by using l methods, e. g. LC/MS analysis of tumor tissues (W0 2014/125273), RNA sequencing (), and immunohistochemistry (Saunders et al., Sci Transl Med. 2015 Aug 26;?(302):302ral36; Saunders et al., AACR 2017; ) analysis.

DLL3 expression has been described in neuroendocrine tumors such as large cell ndocrine carcinoma (LCNEC), small cell lung cancer , small cell bladder cancer, neuroendocrine prostate cancer, neuroendocrine pancreatic cancer and glioblastoma (Saunders et al., Sci Transl Med. 2015 Aug 26,7(3 02):302ra136; Saunders et al., AACR 2017).

SCLC represents an indication of extremely high medical need. SCLC accounts for 13% of Lung Cancer diagnoses and has a worse sis than NSCLC. tional treatments WO 34220 2019/064942 for these cancers mainly include chemotherapy, radiotherapy, surgery or combinations thereof, there are no targeted therapies available yet. While the initial response rate to chemotherapy is high, many patients quickly e with chemo-resistant disease, for which there are no treatment options available. Although there have been improvements in treating these cancers over the last years, overall survival rates for these tumor types remain unchanged, thus, there is an unmet medical need for more targeted and potent therapies.

One ch for a targeted therapy is provided by antibody drug conjugates (ADCs) However, for DLL3 this gy is not preferred due to low cell surface expression of DLL3 and high rate of resistance to chemotherapy. As the ty of patients relapse after chemotherapy treatment and due to the low expression of DLL3 on the cell surface, targeting DLL3 with ADCs might have limitations. In addition ADC approaches often have rget toxicities caused by free drug as a result of linker instability or ation.

There have been attempts to combine DLL3 targeting with targeting of other ns. For example, WO2017/021349 describes a bispecific antibody construct combining binding to human DLL3 on the surface of a target cell and binding to human CD3 on the surface of a T cell. It is however unproven whether such approaches will be successful and to date, no targeted therapies for SCLC and glioblastoma as well as other DLL3 expressing tumors are available.

In view of the poor outlook for such cancer patients, there is a need to identify more efﬁcacious therapies, particularly efﬁcacious therapies with improved bility. Thus, it is an object of the invention to provide pharmacologically active agents, compositions and/or methods of treatment that provide certain advantages compared to the agents, compositions and/or methods currently used and/or known in the art. These advantages include in vivo y, ed therapeutic and pharmacological properties, increased specificity, improved safety profile, less side effects, reduced immunogenicity, and other advantageous properties such as improved ease of preparation or reduced costs of goods, higher stability/longer ife, need for less frequent administration ns especially as compared to candidate drugs already known in the art.

BRIEF SUMMARY OF THE INVENTION The present invention is based on a bispeciﬁc T cell engaging approach employing multi- c binding proteins with a binding arm to CD3 on T cells and a binding arm to DLL3 on the cell surface of tumor cells. h simultaneously binding to T cells and tumor cells, the T cell engagers force formation of a cytolytic synapse n the two cells and so, redirect the T cell activity selectively to the targeted tumor cells.

In one aSpect, the invention provides a multi-speciﬁc binding protein comprising a ﬁrst antigen binding unit speciﬁcally binding to DLL3 and a second antigen g unit specifically binding to CD3, said ﬁrst antigen binding unit speciﬁcally binding to DLL3 ed from the group ting of i) to xviii): i) an antigen binding unit comprising light chain CDRs comprising the amino acid sequences of SEQ ID NO:1 (CDRl), SEQ ID NO:2 (CDR2) and SEQ ID NO:3 (CDR3) and heavy chain CDRs comprising the amino acid sequences of SEQ ID N024 (CDRl), SEQ ID NO:5 (CDR2) and SEQ ID NO:6 (CDR3); ii) an antigen binding unit comprising light chain CDRs sing the amino acid sequences of SEQ ID NO:7 , SEQ ID NO:8 (CDR2) and SEQ ID NO:9 (CDR3) and heavy chain CDRs comprising the amino acid sequences of SEQ ID NO:10 (CDRl), SEQ ID NO:11 (CDR2) and SEQ ID NO: 12 (CDR3); iii) an antigen binding unit comprising light chain CDRs comprising the amino acid sequences of SEQ ID NO:13 (CDRl), SEQ ID NO: 14 (CDR2) and SEQ ID NO:15 (CDR3) and heavy chain CDRs comprising the amino acid ces of SEQ ID NO:16 (CDRI), SEQ ID NO:17 (CDR2) and SEQ ID NO:18 (CDR3); iv) an antigen binding unit comprising light chain CDRs comprising the amino acid sequences of SEQ ID NO:l9 (CDRl), SEQ ID NO:20 (CDR2) and SEQ ID NO:21 (CDR3) and heavy chain CDRs comprising the amino acid sequences of SEQ ID NO:22 (CDRI), SEQ ID NO:23 (CDR2) and SEQ ID NO:24 (CDR3); an antigen binding unit comprising light chain CDRs comprising the amino acid sequences of SEQ ID NO:25 , SEQ ID NO:26 (CDR2) and SEQ ID NO:27 (CDR3) and heavy chain CDRs comprising the amino acid sequences of SEQ ID NO:28 (CDRl), SEQ ID NO:29 (CDR2) and SEQ ID NO:30 (CDR3); Vi) an n binding unit comprising light chain CDRs comprising the amino acid sequences of SEQ ID NO:3l (CDRl), SEQ ID NO:32 (CDR2) and SEQ ID NO:33 (CDR3) and heavy chain CDRs sing the amino acid sequences of SEQ ID NO:34 (CDRl), SEQ ID NO:35 (CDR2) and SEQ ID NO:36 ; Vii) an antigen binding unit comprising light chain CDRs comprising the amino acid sequences of SEQ ID NO:133 (CDRl), SEQ ID NO:134 (CDR2) and SEQ ID NO:135 (CDR3) and heavy chain CDRs comprising the amino acid sequences of SEQ ID NO:136 (CDRl), SEQ ID NO:137 (CDR2) and SEQ ID NO:138 (CDR3); viii) an antigen binding unit comprising light chain CDRs comprising the amino acid sequences of SEQ ID NO:139 (CDRl), SEQ ID NO:140 (CDR2) and SEQ ID NO:141 (CDR3) and heavy chain CDRs comprising the amino acid sequences of SEQ ID NO:142 (CDRl), SEQ ID NO:143 (CDR2) and SEQ ID NO:144 (CDR3); ix) an n binding unit comprising light chain CDRs comprising the amino acid sequences of SEQ ID NO:145 (CDRl), SEQ ID NO:146 (CDR2) and SEQ ID NO:147 (CDR3) and heavy chain CDRs comprising the amino acid ces of SEQ ID NO:148 (CDRl), SEQ ID NO:149 (CDR2) and SEQ ID NO:150 (CDR3); an n binding unit comprising light chain CDRs comprising the amino acid sequences of SEQ ID NO:151 (CDRl), SEQ ID NO:152 (CDR2) and SEQ ID NO:153 (CDR3) and heavy chain CDRs comprising the amino acid sequences of SEQ ID NO:154 (CDRl), SEQ ID NO:155 (CDR2) and SEQ ID NO:156 (CDR3); Xi) an antigen binding unit comprising light chain CDRs comprising the amino acid sequences of SEQ ID NO:157 (CDRl), SEQ ID NO:158 (CDR2) and SEQ ID NO:159 (CDR3) and heavy chain CDRs comprising the amino acid sequences of SEQ ID NO:160 (CDRl), SEQ ID NO:161 (CDR2) and SEQ ID NO:162 (CDR3); xii) an antigen g unit comprising light chain CDRs comprising the amino acid sequences of SEQ ID NO:163 (CDRl), SEQ ID NO:164 (CDR2) and SEQ ID NO:165 (CDR3) and heavy chain CDRs comprising the amino acid sequences of SEQ ID NO:166 (CDRl), SEQ ID NO:167 (CDR2) and SEQ ID NO:168 (CDR3); xiii) an antigen binding unit comprising light chain CDRS comprising the amino acid sequences of SEQ ID NO:169 (CDRl), SEQ ID NO:170 (CDR2) and SEQ ID NO:171 (CDR3) and heavy chain CDRS comprising the amino acid sequences of SEQ ID NO:172 (CDRl), SEQ ID NO:173 (CDR2) and SEQ ID NO:174 (CDR3); xiv) an antigen g unit comprising light chain CDRS comprising the amino acid sequences of SEQ ID NO:175 (CDRl), SEQ ID NO:176 (CDR2) and SEQ ID NO:177 (CDR3) and heavy chain CDRS comprising the amino acid sequences of SEQ ID NO:178 (CDRl), SEQ ID NO:179 (CDR2) and SEQ ID NO:180 (CDR3); xv) an antigen binding unit comprising light chain CDRS comprising the amino acid sequences of SEQ ID NO:181 (CDRl), SEQ ID NO:182 (CDR2) and SEQ ID NO:183 (CDR3) and heavy chain CDRS comprising the amino acid sequences of SEQ ID NO:184 (CDRl), SEQ ID NO:185 (CDR2) and SEQ ID NO:186 (CDR3); xvi) an n binding unit comprising light chain CDRS comprising the amino acid ces of SEQ ID NO:187 (CDRl), SEQ ID NO:188 (CDR2) and SEQ ID NO:189 (CDR3) and heavy chain CDRS comprising the amino acid sequences of SEQ ID NO:190 (CDRl), SEQ ID NO:191 (CDR2) and SEQ ID NO:192 (CDR3); xvii) an n binding unit comprising light chain CDRS sing the amino acid sequences of SEQ ID NO:193 (CDRl), SEQ ID NO:194 (CDR2) and SEQ ID NO:195 (CDR3) and heavy chain CDRS comprising the amino acid sequences of SEQ ID NO:196 (CDRl), SEQ ID NO:197 (CDR2) and SEQ ID NO:198 (CDR3); and xviii) an antigen binding unit comprising light chain CDRS comprising the amino acid sequences of SEQ ID NO:199 (CDRl), SEQ ID NO:200 (CDR2) and SEQ ID NO:201 (CDR3) and heavy chain CDRS comprising the amino acid ces of SEQ ID NO:202 (CDRl), SEQ ID NO:203 (CDR2) and SEQ ID NO:204 (CDR3).

WO 34220 In some ments of the binding n of the invention, the ﬁrst antigen g unit speciﬁcally binding to DLL3 comprises a ﬁrst light chain variable domain and a ﬁrst heavy chain variable domain and is selected from the group consisting of i) to xviii): i) an antigen binding unit comprising a light chain variable domain comprising the amino acid sequence of SEQ ID NO:37 and heavy chain variable domain comprising the amino acid sequence of SEQ ID NO:3 8; ii) an antigen binding unit comprising a light chain variable domain comprising the amino acid sequences of SEQ ID NO:39 and heavy chain variable domain comprising the amino acid sequences of SEQ ID NO:40; iii) an antigen binding unit comprising a light chain le domain comprising the amino acid sequence of SEQ ID NO:4l and heavy chain variable domain comprising the amino acid sequence of SEQ ID NO:42; iv) an antigen binding unit comprising a light chain variable domain comprising the amino acid sequence of SEQ ID NO:43 and heavy chain variable domain comprising the amino acid ce of SEQ ID NO:44; v) an n binding unit comprising a light chain variable domain comprising the amino acid sequence of SEQ ID NO:45 and heavy chain variable domain comprising the amino acid sequence of SEQ ID NO:46; vi) an antigen binding unit comprising a light chain variable domain comprising the amino acid sequence of SEQ ID NO:47 and heavy chain variable domain sing the amino acid sequence of SEQ ID NO:48; vii) an antigen binding unit comprising a light chain variable domain comprising the amino acid sequence of SEQ ID NO:205 and heavy chain variable domain comprising the amino acid sequence of SEQ ID NO:206; viii) an antigen binding unit comprising a light chain le domain comprising the amino acid sequence of SEQ ID NO:207 and heavy chain variable domain comprising the amino acid sequence of SEQ ID NO:208; ix) an antigen binding unit sing a light chain variable domain comprising the amino acid sequence of SEQ ID NO:209 and heavy chain variable domain comprising the amino acid sequence of SEQ ID NO:210; an antigen binding unit comprising a light chain variable domain comprising the amino acid sequence of SEQ ID NO:211 and heavy chain variable domain comprising the amino acid sequence of SEQ ID NO:212; xi) an antigen binding unit comprising a light chain variable domain comprising the amino acid sequence of SEQ ID NO:213 and heavy chain variable domain comprising the amino acid sequence of SEQ ID ; IO xii) an antigen binding unit comprising a light chain variable domain comprising the amino acid sequence of SEQ ID NO:215 and heavy chain variable domain comprising the amino acid sequence of SEQ ID ; xiii) an antigen g unit comprising a light chain variable domain comprising the amino acid sequence of SEQ ID NO:217 and heavy chain variable domain comprising the amino acid sequence of SEQ ID NO:218; xiv) an antigen binding unit comprising a light chain le domain comprising the amino acid sequence of SEQ ID NO:219 and heavy chain variable domain comprising the amino acid sequence of SEQ ID NO:220; xv) an antigen g unit comprising a light chain variable domain comprising the amino acid sequence of SEQ ID NO:221 and heavy chain variable domain comprising the amino acid ce of SEQ ID NO:222; xvi) an antigen binding unit comprising a light chain variable domain comprising the amino acid sequence of SEQ ID NO:223 and heavy chain variable domain comprising the amino acid ce of SEQ ID NO:224; xvii) an antigen g unit comprising a light chain variable domain comprising the amino acid sequence of SEQ ID NO:225 and heavy chain variable domain comprising the amino acid sequence of SEQ ID NO:226; and xviii) an antigen binding unit comprising a light chain le domain comprising the amino acid sequence of SEQ ID NO:227 and heavy chain variable domain comprising the amino acid sequence of SEQ ID NO:228. 2019/064942 In some embodiments of the binding protein of the invention, the second antigen binding unit speciﬁcally binding to CD3 is selected from the group consisting of i) to iii): i) an antigen binding unit sing light chain CDRs comprising the amino acid sequences of SEQ ID NO:55 (CDRl), SEQ ID NO:56 (CDR2) and SEQ ID NO:57 (CDR3) and heavy chain CDRs comprising the amino acid sequences of SEQ ID NO:58 (CDRl), SEQ ID NO:59 (CDR2) and SEQ ID NO:6O ; ii) an antigen g unit comprising light chain CDRs comprising the amino acid sequences of SEQ ID NO:61 (CDRl), SEQ ID NO:62 (CDR2) and SEQ ID NO:63 (CDR3) and heavy chain CDRs comprising the amino acid sequences of SEQ ID NO:64 (CDRl), SEQ ID NO:65 (CDR2) and SEQ ID NO:66 (CDR3); and iii) an n binding unit comprising light chain CDRs comprising the amino acid sequences of SEQ ID NO:96 (CDRl), SEQ ID NO:97 (CDR2) and SEQ ID NO:98 (CDR3) and heavy chain CDRs comprising the amino acid sequences of SEQ ID NO:99 (CDRl), SEQ ID NO:100 (CDR2) and SEQ ID NO:101 (CDR3).

In some embodiments of the binding protein of the invention, the second antigen binding unit speciﬁcally binding to CD3 ses a second light chain variable domain and a second heavy chain variable domain and is selected from the group consisting of: i) an antigen binding unit comprising a light chain variable domain comprising the amino acid sequence of SEQ ID NO:67 and a heavy chain variable domain comprising the amino acid sequence of SEQ ID NO:68; ii) an antigen binding unit comprising a light chain variable domain comprising the amino acid sequence of SEQ ID NO:69 and a heavy chain variable domain comprising the amino acid sequence of SEQ ID NO:70; and iii) an antigen binding unit comprising a light chain variable domain comprising the amino acid sequence of SEQ ID NO:102 and a heavy chain variable domain comprising the amino acid sequence of SEQ ID NO: 103.

In some embodiments of the invention, the ﬁrst antigen binding unit speciﬁcally binding to DLL3 comprises from its N— to C-terminus: a ﬁrst light chain variable , a ﬁrst light chain constant domain, a ﬁrst e linker, a ﬁrst heavy chain variable domain and a ﬁrst heavy chain constant CH1 domain; and the second antigen binding unit speciﬁcally binding to CD3 comprises from its N— to C—terminus: a second light chain le domain, a second light chain constant domain, a second peptide linker, a second heavy chain variable domain and a second heavy chain constant CH1 domain. In some embodiments of the invention, the ﬁrst and/or second peptide linker comprises 26 to 42 amino acids, preferably any one of 30 to 40 amino acids, 34 to 40 amino acids, or 36 to 39 amino acids, more preferably 38 amino acids. In some embodiments of the invention, the ﬁrst linker and/or second linker is a Gly—Ser , preferably comprising the amino acid sequence of SEQ ID NO:89, more preferably said ﬁrst and second peptide linker comprise the same sequence.

In some embodiments, the binding protein of the invention further comprises a ﬁrst and a second Fc domain, said ﬁrst PC domain ntly linked to said ﬁrst antigen binding unit, and said second Fc domain covalently linked to said second antigen binding unit.

In some embodiments of the invention, i) the ﬁrst PC domain comprises a ne (Y) at position 366 [T3 66Y], and the second Fc domain ses a threonine (T) at position 407 [Y407T], or ii) the ﬁrst Fc domain comprises a phan (W) at position 366 ], and the second Fc domain comprises a serine (S) at position 366 [T366S], an alanine (A) at position 368 [L368A] and a valine (V) at position 407 [Y407V], or iii) the second Fc domain comprises a tyrosine (Y) at position 366 [T366Y], and the ﬁrst PC domain comprises a threonine (T) at position 407 [Y407T], or iv) the second Fc domain comprises a tryptophan (W) at position 366 [T366W], and the ﬁrst Fc domain comprises a serine (S) at position 366 [T366S], an alanine (A) at position 368 ] and a valine (V) at position 407 [Y407V], preferably the ﬁrst or the second Fc domain further comprises an arginine at on 435 [H435R] and a phenylalanine at position 436 [Y436F]. In some embodiments, the ﬁrst and/or second Fc domain further ses an alanine at position 234[L234A] and at position 235 ]. In some embodiments, the ﬁrst light chain constant domain and the second light chain constant domain comprise a human kappa or lambda domain.

In preferred embodiments, the g protein of the invention comprises a ﬁrst polypeptide chain speciﬁcally binding to DLL3 comprising an amino acid sequence selected from the group consisting of SEQ ID NO:73, SEQ ID NO:74, SEQ ID NO:75, SEQ ID NO:76, SEQ ID NO:77, SEQ ID NO:78, SEQ ID NO:241, 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, SEQ ID NO: 251, and SEQ ID NO: 252 WO 34220 and a second polypeptide chain speciﬁcally binding to CD3 comprising the amino acid sequence of SEQ ID NO:79, SEQ ID NO:80 or SEQ ID .

In a further aspect, the invention provides an ed nucleic acid molecule i) encoding a ﬁrst antigen binding unit and/or a second antigen binding unit of a protein of the invention, optionally further encoding a ﬁrst and/or a second Fc domain, or ii) encoding a ﬁrst and/or a second polypeptide chain of a protein of the invention. In further aspects provided herein are expression vectors sing the c acid molecule of the ion, host cells transfected with such expression vectors, and methods of manufacturing a protein of the invention.

In a further aspect of the ion, provided herein is a multi-speciﬁc binding protein comprising a ﬁrst ptide chain speciﬁcally binding to DLL3 and a second ptide chain speciﬁcally binding to CD3, where the ﬁrst polypeptide chain comprises a ﬁrst light chain, a ﬁrst linker, and a ﬁrst heavy chain and the second polypeptide chain comprises a second light chain, a second linker, and a second heavy chain, preferably the C-terminus of said ﬁrst light chain is covalently bound to the N—terminus of said ﬁrst heavy chain via said ﬁrst peptide linker and the C-terminus of said second light chain is covalently bound to the inus of said second heavy chain Via said second peptide linker.

In some embodiments of the g protein of the invention, the ﬁrst polypeptide chain speciﬁcally binding to DLL3 comprises a light chain variable domain and heavy chain variable domain comprising CDR sequences and/or VH/VL sequences as deﬁned for the antigen binding units of DLL3#1, DLL3#2, DLL3#3, DLL3#4, DLL3#5, DLL3#6, DLL3#7, DLL3#8, DLL3#9, DLL3#10, DLL3#ll, DLL3#12, DLL3#13, DLL3#14, DLL3#15, DLL3#16, DLL3#17, or DLL3#18 described herein. In some embodiments, the second polypeptide chain speciﬁcally binding to CD3 comprises a light chain variable and heavy chain variable domain comprising CDR sequences and/or VH/VL sequences as deﬁned for the n binding units of CD3#1, CD3#2, or CD3#3 described herein.

Further s, embodiments, uses and methods involving the binding proteins of the invention will become clear from the following detailed description of the invention and from the appended claims. _10_ The ion provides for novel binding proteins that allow a more efﬁcient treatment of DLL3 expressing cancers, such as SCLC, glioblastoma or a DLL3 expressing neuroendocrine tumor.

BRIEF DESCRIPTION OF THE DRAWINGS Figure 1: Schematic entation of a bispeciﬁc binding protein of the invention Figure 2: Schematic representation of the DLL3 full length protein and DLL3 domain constructs expressed on HEK293 cells for epitope domain mapping. For the DLL3 domain ucts, the transmembrane and the intracellular domains are derived from EpCAM.

Figure 3A: DLL3 epitope domain mapping of seven exemplary DLL3/CD3 binding proteins. ary DLL3/CD3 binding proteins recognizing the membrane proximal peptide, EGF4, EGFl and DSL domains of DLL3. The y-axis depicts , the x-axis depicts PE—A (phycoerythrin signal area).

Figure 3B: DLL3 epitope domain mapping of six exemplary DLL3/CD3 binding ns.

Exemplary DLL3/CD3 binding proteins recognizing the EGFl, EGF3, EGF4 or EGF6 of DLL3. The y—axis depicts counts, the x-axis depicts PE-A erythrin signal).

Figure 3C: DLL3 epitope domain mapping of six exemplary DLL3/CD3 binding proteins. ary DLL3/CD3 binding proteins recognizing the DSL domains or neither DSL, EGF, nor membrane proximal peptide of DLL3. The y-axis depicts counts, the x-axis depicts PE—A (phycoerythrin signal).

Figure 4: Binding of seven exemplary DLL3/CD3 binding proteins to cell lines expressing human and cyno DLL3. The y—axis s counts, the x-axis depicts PE-A (phycoerythrin signal area).

Figure 5: Binding of seven exemplary DLL3/CD3 g ns to cell lines expressing human DLLl and DLL4, no binding to human DLLl and DLL4 detected. The y—axis depicts , the x-axis depicts PE-A (phycoerythrin signal area).

Figure 6A: Binding of seven exemplary DLL3/CD3 binding proteins to SCLC cell lines and human T cells by ﬂow cytometry analysis.

Figure 6 B: Binding of exemplary DLL3/CD3 binding proteins (directed to a peptide that is neither the DSL nor the EGFl-6 nor the membrane proximal peptide) to SHP77 cells. _11_ Figure 6 C: Binding of exemplary DLL3/CD3 binding proteins (directed against the DSL domain) to SHP77 cells.

Figure 6 D: Binding of exemplary DLL3/CD3 binding proteins (directed against the EGFl domain) to SHP77 cells.

Figure 6 E: Binding of ary DLL3/CD3 binding ns (directed against the EGF3 domain) to SHP77 cells.

Figure 6 F: Binding of exemplary D3 binding proteins (directed against the EGF4 domain) to SHP77 cells.

Figure 6 G: Binding of exemplary D3 binding proteins (directed against the DSL domain) to SHP77 cells.

Figure 6 H: Binding of exemplary DLL3/CD3 binding proteins ted against the membrane al peptide ) to SHP77 cells.

Figure 6 1: Binding of exemplary DLL3/CD3 binding proteins (directed to a peptide that is neither the DSL nor the EGFl-6 nor the membrane proximal peptide) to NCI—H82 cells.

Figure 6 J: Binding of exemplary DLL3/CD3 binding proteins (directed against the DSL domain) to NCI-H82 cells.

Figure 6 K: Binding of exemplary DLL3/CD3 binding proteins (directed t the EGFl domain) to NCI—H82 cells.

Figure 6 L: Binding of exemplary DLL3/CD3 binding proteins (directed against the EGF3 domain) to 2 cells.

Figure 6 M: Binding of exemplary DLL3/CD3 binding ns ted against the EGF4 domain) to NCI—H82 cells.

Figure 6 N: Binding of exemplary D3 binding proteins (directed against the DSL domain) to NCI—H82 cells.

Figure 6 0: Binding of exemplary DLL3/CD3 binding proteins (directed against the membrane proximal peptide domain) to NCI-H82 cells.

Figure 6 P: Binding of exemplary DLL3/CD3 binding ns (directed to a peptide that is neither the DSL nor the EGFl-6 nor the membrane proximal peptide) to T cells. _12_ Figure 6 Q: Binding of exemplary D3 binding proteins (directed against the DSL domain) to T cells.

Figure 6 R: Binding of exemplary DLL3/CD3 binding proteins (directed against the EGFl domain) to T cells.

Figure 6 S: g of exemplary DLL3/CD3 binding proteins (directed t the EGF3 domain) to T cells.

Figure 6 T: Binding of exemplary DLL3/CD3 binding proteins (directed against the EGF4 domain) to T cells.

Figure 6 U: g of exemplary DLL3/CD3 binding proteins ted against the EGF6 domain) to T cells.

Figure 6 V: g of exemplary D3 binding ns (directed against the membrane proximal peptide domain) to T cells.

Figure 7 A: Potency in lysing cells of seven D3 binding proteins redirecting non- stimulated PBMCs towards human SCLC cell lines.

Figure 7 B: Potency in lysing cells of exemplary DLL3/CD3 binding proteins (directed to a peptide that is neither the DSL nor the EGF1-6 nor the membrane proximal peptide) redirecting non-stimulated T cells towards human SHP77 cells.

Figure 7 C: Potency in lysing cells of exemplary DLL3/CD3 binding proteins ted against the DSL domain) redirecting non-stimulated T cells towards human SHP77 cells.

Figure 7 D: Potency in lysing cells of exemplary DLL3/CD3 binding proteins (directed against the EGFl domain redirecting non-stimulated T cells towards human SHP77 cells.

Figure 7 E: Potency in lysing cells of exemplary DLL3/CD3 binding proteins (directed against the EGF3 domain) redirecting non—stimulated T cells towards human SHP77 cells.

Figure 7 F: Potency in lysing cells of exemplary DLL3/CD3 binding proteins (directed against the EGF4 domain) redirecting non-stimulated T cells s human SHP77 cells.

Figure 7 G: Potency in lysing cells of exemplary DLL3/CD3 binding proteins (directed against the EGF6 domain) redirecting non-stimulated T cells towards human SHP77 cells.

Figure 7 H: Potency in lysing cells of exemplary DLL3/CD3 binding ns (directed against the membrane proximal peptide redirecting imulated T cells towards human SHP77 cells. _13_ Figure 7 I: Potency in lysing cells of exemplary DLL3/CD3 binding proteins (directed to a peptide that is neither the DSL nor the EGF1-6 nor the membrane proximal peptide) redirecting non-stimulated T cells towards human NCI-H82 cells.

Figure 7 J: Potency in lysing cells of exemplary DLL3/CD3 binding proteins (directed against the DSL domain) redirecting imulated T cells towards human NCI—H82 cells.

Figure 7 K: Potency in lysing cells of exemplary DLL3/CD3 g proteins (directed against the EGFl domain redirecting non-stimulated T cells towards human NCI-H82 cells.

Figure 7 L: Potency in lysing cells of exemplary DLL3/CD3 binding proteins (directed against the EGF3 domain) redirecting non—stimulated T cells towards human NCI-H82 cells.

Figure 7 M: Potency in lysing cells of ary DLL3/CD3 binding proteins (directed against the EGF4 domain) redirecting non-stimulated T cells towards human NCI-H82 cells.

Figure 7 N: Potency in lysing cells of exemplary D3 binding ns (directed against the EGF6 domain) redirecting non-stimulated T cells towards human 2 cells.

Figure 7 O: Potency in lysing cells of exemplary D3 binding ns (directed against the membrane proximal peptide cting non-stimulated T cells towards human NCI—H82 cells.

Figure 8: In vitro internalization assay with two ary DLL3/CD3 binding proteins.

DLL3/CD3 binding proteins were pre-incubated with DLL3-positive SHP77 cells for 2 and 4 hours before co—incubation for 48 hours with human PBMCs.

Figure 9: Pharmacokinetics of two exemplary DLL3/CD3 binding proteins. Mean cokinetic s of DLL3#3/CD3#1 (open squares) and DLL3#3/CD3#2 (closed squares) in male C57BL/6 mice following a single 1 mg/kg intravenous dose.

Figure 10: Anti-tumor activity of two exemplary DLL3/CD3 binding proteins. The y—axis depicts median tumor volumes and the x—axis depicts time.

Figure 11: Potency in lysing cells of an exemplary DLL3/CD3 binding protein redirecting imulated PBMCstowards human SHP77 and RKO—E6 cells. _14_ Figure 12: Potency in activation of T cells in presence of SHP77 cells of an ary DLL3/CD3 binding protein Figure 13: Potency in degranulation of T cells in presence of SHP77 cells of an exemplary D3 binding protein Figure 14A: Potency secretion of Interferon gamma by PBMCsin presence of SHP77l cells of an exemplary DLL3/CD3 binding protein Figure 143: Potency secretion of MCP—l by PBMCsin presence of SHP77 cells of an exemplary DLL3/CD3 g protein Figure 15: Potency in proliferation of T cells in presence of SHP77 cells of an exemplary DLL3/CD3 binding protein Figure 16: Potency in lysing cells of an exemplary DLL3/CD3 binding protein redirecting imulated pan T cells and na't've T cells towards human SHP77 cells Figure 17: Potency in lysing cells of an exemplary DLL3/CD3 g protein redirecting non—stimulated CD4+ and CD4+ central memory T cells towards human SHP77 cells Figure 18: Potency in lysing cells of an exemplary DLL3/CD3 binding protein redirecting non—stimulated CD4+ and CD4+ effector memory T cells towards human SHP77 cells Figure 19: Potency in lysing cells of an exemplary DLL3/CD3 binding protein redirecting non-stimulated CD8+ and CD8+CD45RA+ effector memory T cells s human SHP77 cells Figure 20: y in lysing cells of an exemplary DLL3/CD3 binding protein redirecting non-stimulated CD8+ and CD8+ memory T cells towards human SHP77 Figure 21: T cell inﬁltration in SHP77 xenograft tumor tissue with an exemplary DLL3/CD3 binding protein.

Figure 22: Binding of three exemplary anti-DLL3 antibodies to recombinant DLL3 protein in ELISA assay. The y-axis depicts , the x-axis depicts concentration of DLL3 protein.

Figure 23: Binding of six exemplary anti-DLL3 antibodies to DLL3-positive SCLC cell lines determined by ﬂow cytometry. The y—axis depicts , the x—axis depicts concentration of anti-DLL3 antibody.

Figure 24: Representative staining of a SCLC tissue sample with 5 exemplary anti-DLL3 dies. Anti-DLL3 antibodies DLL3#1 and DLL3#5 showed that the tumor cells exhibit a te and/or diffuse cytoplasmic and/or nous DLL3 staining. Images _15_ have been electronically generated by LEICA SCN400 automated scanner with a magniﬁcation of 10x.

Figure 25: Representative staining of cell pellets from SCLC cell lines with different DLL3 expression levels. Anti-DLL3 antibody DLL3#5 DETAILED DESCRIPTION OF THE INVENTION USED TERMS AND DEFINITIONS The above and other aspects and embodiments of the invention will become clear from the further description herein, in which: Unless indicated or deﬁned ise, all terms used have their usual meaning in the art, which will be clear to the skilled person. nce is for example made to the rd handbooks, such as Sambrook et al, "Molecular Cloning: A Laboratory Manual" (2nd Ed.), Vols. 1—3, Cold Spring Harbor Laboratory Press (1989); Lewin, "Genes IV", Oxford sity Press, New York, (1990), and Roitt et al., "Immunology" (2nd Ed.), Gower Medical Publishing, London, New York (1989), as well as to the general ound art cited herein. Furthermore, unless indicated otherwise, all methods, steps, techniques and manipulations that are not speciﬁcally described in detail can be performed and have been performed in a manner known per se, as will be clear to the skilled person. nce is for example again made to the standard handbooks, to the general background art referred to above and to the further references cited n.

When used herein the term “comprising” and variations thereof such as “comprises” and “comprise” can be substituted with the term “containing” or ding” or “having.” The term "sequence" as used herein (for example in terms like "heavy/light chain II N antibody sequenceH “ sequence variable domain sequence37 ant domain sequence” , , , or "protein sequence"), should generally be understood to include both the relevant amino acid sequence as well as nucleic acid sequences or nucleotide sequences encoding the same, unless the context requires a more limited interpretation.

An “antigen binding unit” as used herein refers to a polypeptide capable of binding to its speciﬁc target or antigen and comprising the l structural requirements derived from an antibody (typically present in an antibody) which allow for target binding. Thus, an antigen g unit comprises at least the presence of three light chain and three heavy —l6— chain CDR sequences, preferably at least a light chain variable domain and a heavy chain variable domain.

The generalized structure of an antibody or immunoglobulin is well known to those of skill in the art. These molecules are heterotetrameric glycoproteins, typically of about 150,000 daltons, composed of two cal light (L) chains and two identical heavy (H) chains and are typically referred to as full length antibodies. Each light chain is covalently linked to a heavy chain by one disulﬁde bond to form a heterodimer, and the heterotetrameric le is formed through a covalent disulflde linkage between the two identical heavy chains of the dimers. Although the light and heavy chains are linked together by one disulﬁde bond, the number of disulflde linkages between the two heavy chains varies by immunoglobulin isotype. Each heavy and light chain also has regularly spaced intrachain disulﬁde bridges. Each heavy chain has at the inus a variable domain (VH), followed by three or four (in case of IgE) constant s (CH1, CH2, CH3, and CH4), as well as a hinge region between CH1 and CH2. Each light chain has two domains, an N- terminal variable domain (VL) and a C-terminal constant domain (CL). The VL domain associates non-covalently with the VH domain, s the CL domain is ly covalently linked to the CH1 domain via a disulfide bond. Particular amino acid residues are believed to form an interface between the light and heavy chain variable domains (Chothia et al., 1985, J. Mol. Biol. 186:651—663). Variable s are also referred to herein as variable regions or Fv and denote the part that confers specificity to an antibody for the antigen by carrying the antigen-binding site.

The “light chain variable domain” (or “light chain le region”) and “heavy chain variable domain” (or “heavy chain variable region”) as used herein have the same general structure and each domain essentially consists of four framework (FR) regions whose sequences are widely conserved, which are referred to in the art and hereinbelow as "framework region 1" or "FRI"; as "framework region 2" or"FR2"; as "framework region 3" or "FR3"; and as "framework region 4" or "FR4", respectively; which framework regions are upted by three hypervariable regions, HVRs (or CDRs), which are referred to in the art and herein below as ementarity determining region l"or "CDRl"; as "complementarity determining region 2" or "CDR2"; and as "complementarity ining region 3" or "CDR3", respectively. Thus, the general structure or sequence of an immunoglobulin variable domain can be indicated as follows: FRl - CDRl - FR2 — CDR2 - FR3 - CDR3 - FR4. The framework regions adopt a beta-sheet conformation and _17_ WO 34220 the CDRs may form loops connecting the beta-sheet structure. The CDRs in each chain are held in their three-dimensional structure by the framework s and form together with the CDRs from the other chain the antigen binding site.

Within the context of this invention, reference to CDR’s is based on the ion of CCG, also referred to as IMGT (Lefranc MP, Pommie’ C, Ruiz M, elli V, Foulquier E, Truong L, Thouvenin—Contet V, Lefranc G. “IMGT unique numbering for immunoglobulin and T cell receptor variable domains and Ig superfamily V-like domains.“ Dev Comp Immunol. 2003 Jan;27(l):55-77; Giudicelli V, Brochet X, Lefranc MP. “IMGT/V-QUEST: IMGT rdized analysis of the globulin (IG) and T cell receptor (TR) nucleotide sequences“. Cold Spring Harb Protoc. 2011;2011(6):695—715. An alternative deﬁnition of CDRs known in the art is based on Chothia (Chothia and Lesk, J. Mol. Biol. 1987, 196: 901—917), together with Kabat (E.A. Kabat, T.T. Wu, H. Bilofsky, M. Reid- Miller and H. Perry, ce of Proteins of logical Interest, National Institutes of Health, Bethesda (1983)).

The term “constant domains” or “constant ” as used within the current application denotes the sum of the domains of an antibody other than the variable region. Such constant domains and regions are well known in the state of the art and e.g. described by Kabat et a1. (“Sequence of proteins of immunological interest”, US Public Health Services, NIH Bethesda, MD, Publication No. 91).

The "Fc part" or “Fc domain” of an antibody is not involved directly in binding of an antibody to an antigen, but ts various effector functions. An "Fc part/domain of an antibody" is a term well known to the d artisan and deﬁned on the basis of papain cleavage of antibodies. Depending on the amino acid sequence of the constant region of their heavy chains, antibodies or immunoglobulins are divided in the classes: IgA, IgD, IgE, IgG and IgM. According to the heavy chain constant s the different classes of immunoglobulins are called 0t, 5, 8, y, and u respectively. Several of these may be further divided into subclasses pes), e.g. IgGl, IgG2, IgG3, and IgG4, IgAl, and IgA2. The Fc part of an dy is directly involved in ADCC (antibody dependent cell—mediated cytotoxicity) and CDC (complement—dependent cytotoxicity) based on complement activation, Clq binding and Fc receptor binding. Complement activation (CDC) is initiated by binding of complement factor Clq to the F0 part of most IgG antibody subclasses. While the inﬂuence of an antibody on the complement system is dependent on certain conditions, —18- binding to Clq is caused by deﬁned binding sites in the Fc part. Such binding sites are e. g.

L234, L235, D270, N297, E318, K320, K322, P331 and P329 (numbering according to EU numbering (Edelman et al, Proc Natl Acad Sci U S A. 1969 May;63(l):78—85)). Most crucial among these residues in mediating Clq and chamma receptor binding in IgGl are L234 and L235 (Hezareh et al., J. Virology 75 (2001) 12161- 12168, Shields et a1 (2001) JBC, 276 (9): 6591—6604). Antibodies of subclass IgG1 and IgG3 usually show complement activation and Clq and C3 binding, s IgG2 and IgG4 do not te the ment system and do not bind Clq and C3.

The term “antibody” or “antibody molecule” (used synonymously herein) encompasses a monoclonal antibody, a polyclonal dy, a human antibody, a humanized antibody, a chimeric antibody, peciflc antibodies (e.g., bispeciflc antibodies), a fragment of an antibody, in particular a Fv, Fab, Fab’, or F(ab’)2 fragment, a single chain antibody, in particular a single chain variable nt (scFv), a single chain Fab fragment (scFab), a Small Modular Immunopharmaceutical , a domain antibody, a nanobody, a diabody. The dy may have an effector function, such as ADCC or CDC, that is usually mediated by the Fc part (antibody constant ) of the antibody, or it may have no effector function, e. g. by lacking a PC part or having a blocked, masked Fc part, in essence a PC part that is not or insufﬁciently recognized by immune cells or immune system components, like the complement system.

Monoclonal antibodies (mAb) are monospeciﬁc antibodies that are identical in amino acid sequence. They may be produced by hybridoma technology from a hybrid cell line (called oma) representing a clone of a fusion of a specific antibody-producing B cell with a myeloma (B cell cancer) cell (Kohler G, Milstein C. Continuous es of fused cells secreting antibody of predeﬁned specificity. Nature 1975;256:4957). Alternatively, monoclonal antibodies may be produced by recombinant sion in host cells (Norderhaug L, Olafsen T, Michaelsen TE, Sandlie I. (May 1997). "Versatile vectors for transient and stable expression of recombinant antibody molecules in mammalian cells." J l Methods 204 (1): 77—87; see also below). A binant antibody” or “recombinant binding protein” is an antibody or binding n which has been produced by a recombinantly engineered host cell. It is optionally isolated or purified.

Full length antibodies can be treated with enzymes such as papain or pepsin to generate useful antibody fragments. Papain digestion is used to produces two identical antigen— _19_ binding antibody fragments called "Fab" fragments, each with a single antigen—binding site, and a al "Fc" fragment. The Fab fragment also contains the constant domain of the light chain and the CH1 domain of the heavy chain. Pepsin treatment yields a F(ab')2 fragment that has two antigen-binding sites and is still capable of cross-linking antigen.

Fab' fragments differ from Fab fragments by the presence of additional residues including one or more cysteines from the antibody hinge region at the C-terminus of the CH1 domain. F(ab')2 antibody fragments are pairs of Fab‘ fragments linked by cysteine residues in the hinge region. Other chemical couplings of antibody fragments are also known.

"Fv" fragment contains a te antigen-recognition and g site ting of a dimer of one heavy and one light chain le domain in tight, non-covalent ation.

In this conﬁguration, the three CDRs of each variable domain interact to deﬁne an antigen- biding site on the surface of the VH—VL dimer. Collectively, the six CDRs confer antigen- g specificity to the dy.

A "single—chain FV" or "scFv" antibody fragment is a single chain FV variant comprising the VH and VL s of an antibody where the domains are present in a single polypeptide chain. The single chain Fv is capable of recognizing and binding an antigen.

The scFV polypeptide may optionally also contain a ptide linker positioned between the VH and VL domains in order to facilitate formation of a desired three-dimensional ure for antigen binding by the scFV (see, e.g., Pluckthun, 1994, In The cology of monoclonal Antibodies, Vol. 113, Rosenburg and Moore eds., Springer—Verlag, New York, pp. 269-315).

A "single-chain Fab" or "scFab" antibody fragment is a single chain Fab variant comprising the VL, CL, VH and CH1 domains of an antibody where the domains are present in a single polypeptide chain. The single chain Fab is capable of recognizing and binding an antigen. The scFab polypeptide may optionally also contain a polypeptide linker positioned between the CL and VH domains (Hust et al (2007) BMC Biotechnology).

For application in man, it is often desirable to reduce immunogenicity of therapeutic molecules, such as antibodies or binding ns comprising an antigen binding unit as described herein, originally derived from other species, like mouse. This can be done by construction of chimeric dies/binding proteins, or by a process called “humanization”. In this context, a “chimeric antibody”; or “chimeric antigen binding unit” _20_ WO 34220 is understood to be an antibody or an n binding unit comprising a sequence part (e. g. a variable domain) derived from one species (e. g. mouse) fused to a sequence part (e.g. the constant domains) d from a different species (e. g. human). In this context, a “humanized antibody”, “a humanized binding protein” or a “humanized antigen binding unit” is an antibody, a protein or antigen binding unit comprising a le domain originally derived from a non-human s, wherein n amino acids have been mutated to make the overall sequence of that variable domain more closely resemble a sequence of a human variable domain. Methods of humanization of antibodies are well- known in the art (Billetta R, Lobuglio AF. “Chimeric antibodies”. Int Rev Immunol. l993;lO(2—3): 165-76; Riechmann L, Clark M, Waldmann H, Winter G . "Reshaping human antibodies for therapy". Nature: 332:323).

An “optimized antibody” or an “optimized antigen binding unit or protein” is a specific type of humanized dy or humanized antigen binding unit/protein which includes an immunoglobulin amino acid sequence variant, or fragment thereof, which is capable of binding to a predetermined n and which comprises one or more FRs having substantially the amino acid sequence of a human immunoglobulin and one or more CDRs having substantially the amino acid sequence of a non-human immunoglobulin. This non- human amino acid sequence often referred to as an "import" sequence is typically taken from an "import" antibody domain, particularly a variable domain. In general, an optimized antibody includes at least the CDRs (or HVLs) of a non-human antibody or derived from a non-human antibody, inserted between the FRs of a human heavy or light chain variable domain. It will be understood that certain mouse FR residues may be important to the function of the optimized antibodies and therefore certain of the human ne sequence heavy and light chain variable domains es are modified to be the same as those of the corresponding mouse ce. During this s undesired amino acids may also be removed or changed, for e to avoid deamidation, undesirable charges or lipophilicity or non—specific binding. An “optimized dy”, an “optimized antibody fragment” or “optimized” may sometimes be referred to as “humanized antibody”, “humanized antibody nt” or “humanized”, or as “sequence-optimized”.

Furthermore, technologies have been developed for creating antibodies or VH/VL domains based on sequences derived from the human genome, for example by phage display or use of transgenic animals (WWW. Ablexis.com/technology—alivamab.php; WO 90/05144; D.

Marks, H.R. Hoogenboom, T.P. Bonnert, J. McCafferty, A.D. Grifﬁths and G. Winter _21_ 2019/064942 (1991) "By—passing immunisation. Human antibodies from V—gene libraries displayed on phage." J.Mol.Biol., 222, 581—597; Knappik et al., J. M01. Biol. 296: 57—86, 2000; S.

Carmen and L. Jermutus, "Concepts in antibody phage display". Briefings in Functional Genomics and Proteomics 2002 1(2):189-203; Lonberg N, Huszar D. "Human antibodies from transgenic mice". Int Rev Immunol. 1995;13(1):65-93.; Brﬁggemann M, Taussig MJ.

"Production of human antibody repertoires in transgenic mice”. Curr Opin hnol. 1997 Aug;8(4):455-8.). Such antibodies or antigen binding units or VH/VL domains are “human antibodies,” “human n binding units,” or “human VH/VL domains” in the context of the present invention.

The term "human antibody", “human antigen binding unit”, or “ human VH/VL ” as used herein, is intended to include antibodies, antigen g units or VH/VL domains having variable (and constant, if applicable) regions d from human germline immunoglobulin sequences. The human antibodies, antigen binding units, proteins or VH/VL domains of the present technology may include amino acid residues not encoded by human germline immunoglobulin sequences (e.g., mutations introduced by random or site- speciﬁc mutagenesis in vitro or by somatic mutation in vivo). However, the term "human antibody", “human n g unit”, or “ human VH/VL domain”as used herein, is not intended to e antibodies in which CDR sequences derived from the germline of r lian species), such as a mouse, rat or rabbit, have been grafted onto human framework sequences. Thus, as used herein, the term "human antibody", “human antigen binding unit”, or “human VH/VL ” refer to an antibody, antigen binding unit or VH/VL domain in which substantially every part of the protein (e.g., CDR, framework, CL, CH domains (e.g., CHI is substantially non- , CH2, CH3), hinge, VL, VH) immunogenic in humans, with only minor sequence changes or variations. Such changes or variations optionally and preferably retain or reduce the immunogenicity in humans or other species relative to non—modiﬁed antibodies or antigen binding units.

Thus, a human antibody, human n binding unit or human VH/VL domain is distinct from e.g., a chimeric or humanized antibody. It is d out that a human antibody, human antigen binding unit or human VH/VL domain can be produced by a non—human animal or prokaryotic or eukaryotic cell that is capable of expressing functionally rearranged human immunoglobulin (e.g., heavy chain and/or light chain) genes. _22_ The term er” refers to a homogenous form of an antibody or a multispeciflc protein as described herein. For example, for a full-length antibody, monomer means a monomeric antibody having two identical heavy chains and two identical light . In the context of the t invention, a monomer means a protein of the present invention having a single antigen binding unit speciﬁc for DLL3, and a single antigen binding unit specific for CD3 as described herein. For example, a r of a g protein bed herein may have two chains, a ﬁrst chain comprising a single chain Fab with a first antigen binding unit and ally a first PC domain and a second chain comprising a single chain Fab with a second antigen binding unit and optionally a second Fc domain.

An epitope is a region of an antigen that is bound by an antibody or n binding moiety (e.g. the antigen binding unit of the proteins described herein). The term “epitope” includes any polypeptide determinant capable of specific binding to an antibody or antigen binding moiety. In certain embodiments, epitope determinants include chemically active surface groupings of molecules such as amino acids, glycan side , phosphoryl, or sulfonyl, and, in certain embodiments, may have speciﬁc three dimensional structural characteristics, and/or speciﬁc charge characteristics. Conformational and non- conformational epitopes are distinguished in that the g to the former but not the latter is lost in the presence of denaturing solvents.

An n binding molecule/protein (such as an globulin, an antibody, an antigen binding unit, or a fragment of such antigen binding molecule/protein) that can “bind”, "bind to", “speciﬁcally bind”, or "Speciﬁcaz’b/ bind to", that "has aﬂiniU/for", “is specific for” and/or that "has specificilyfor" a certain epitope, antigen or protein (or for at least one part, fragment or epitope thereof) is said to be "against" or "directed againsi" said epitope, antigen or protein or is a "binding" molecule/protein with respect to such epitope, antigen or protein.

As used herein, the terms “binding” and “specific binding” refer to the binding of the antibody or antigen binding moiety (such as an immunoglobulin, an dy, an antigen binding unit, or a fragment of such antigen binding le/protein) to an epitope of the n in an in vitro assay, preferably in a plasmon resonance assay ((Malmqvist M., "Surface plasmon resonance for detection and measurement of antibody-antigen affinity and kinetics", Curr Opin Immunol. 1993 Apr;5(2):282-6.)) with puriﬁed wild—type antigen. Antibody y can also be ed using kinetic exclusion assay (KinExA) _23_ technology (Darling, R.J., and Brault P—A., “Kinetic exclusion assay technology: Characterization of Molecular Interactions.” ASSAY and Drug Development Technologies. 2004, Dec 2(6): 647-657).

Generally, the term "speciﬁcity" refers to the number of ent types of antigens or epitopes to which a particular antigen binding molecule/protein (such as an globulin, an antibody, an antigen binding unit, or a fragment of such n binding molecule/protein) can bind. The city of an antigen-binding molecule/protein can be determined based on its affinity and/or avidity. The afﬁnity, ented by the equilibrium constant for the dissociation of an n with an antigen-binding protein (KD), is a measure for the binding strength between an epitope and an antigen—binding site on the antigen-binding molecule/protein: the lesser the value of the K13, the stronger the binding strength between an epitope and the antigen—binding molecule/protein (alternatively, the afﬁnity can also be expressed as the affinity constant (KA), which is l/KD). As will be clear to the skilled person (for example on the basis of the further disclosure herein), affinity can be determined in a manner known per se, ing on the c antigen of interest. y is the measure of the th of binding between an antigen-binding molecule/protein (such as an immunoglobulin, an antibody, an antigen binding unit, or fragment of such n binding molecule/protein) and the pertinent antigen. Avidity is related to both the affinity between an epitope and its n binding site on the antigen—binding molecule/protein and the number of pertinent g sites present on the antigen-binding molecule/protein.

The term "isolated," as used herein, refers to material that is removed from its original or native environment (e.g. the l environment if it is naturally occurring). For example, a naturally-occurring polynucleotide or polypeptide present in a living animal is not isolated, but the same polynucleotide or polypeptide, separated by human intervention from some or all of the co-existing materials in the natural system, is isolated. Such polynucleotides could be part of a vector and/or such polynucleotides or polypeptides could be part of a composition, and still be isolated in that such vector or composition is not part of the environment in which it is found in nature. For example, a nucleic acid, protein/polypeptide molecule is considered to be "(in) essentially isolated (form)" - when compared to its native biological source and/or the reaction medium or cultivation medium from which it has been obtained - when it has been separated from at least one other ent with which it is usually associated in said source or medium, such as another _24_ nucleic acid, another protein/polypeptide, another biological ent or macromolecule or at least one contaminant, impurity or minor component. In particular, a nucleic acid or protein/polypeptide molecule is considered "essentially isolated" when it has been d at least , in ular at least 10- fold, more in particular at least ld, and up to old or more. A nucleic acid or protein/polypeptide molecule that is "in essentially isolated form" is preferably essentially homogeneous, as determined using a suitable technique, such as a suitable chromatographical technique, e. g., polyacrylamide- gelelectrophoresis.

As used , the terms "identical" or "percent identity, " in the context of two or more nucleic acids or polypeptide sequences, refer to two or more sequences or subsequences that are the same or have a speciﬁed percentage of nucleotides or amino acid residues that are the same, when compared and d for maximum correspondence. To determine the percent identity, the sequences are aligned for optimal comparison purposes (e.g., gaps can be introduced in the sequence of a ﬁrst amino acid or nucleic acid sequence for optimal alignment with a second amino or nucleic acid sequence). The amino acid residues or nucleotides at corresponding amino acid positions or nucleotide positions are then compared. When a position in the ﬁrst ce is occupied by the same amino acid residue or nucleotide as the corresponding on in the second sequence, then the molecules are identical at that position. The t identity between the two sequences is a on of the number of identical positions shared by the sequences (i.e., % identity=# of identical positions/total # of positions (e.g., overlapping positions)x100). In some embodiments, the two sequences that are compared are the same length after gaps are introduced within the sequences, as appropriate (e.g., excluding additional sequence extending beyond the sequences being compared). For example, when variable region sequences are ed, the leader and/or constant domain sequences are not considered.

For sequence comparisons between two sequences, a "corresponding" CDR refers to a CDR in the same on in both sequences (e.g., CDR—H1 of each sequence).

The determination of percent identity or percent similarity between two sequences can be accomplished using a mathematical thm. A preferred, non-limiting example of a mathematical algorithm utilized for the comparison of two sequences is the thm of Karlin and Altschul, 1990, Proc. Natl. Acad. Sci. USA 87:2264—2268, modiﬁed as in Karlin and Altschul, 1993, Proc. Natl. Acad. Sci. USA 90:5873—5877. Such an algorithm is incorporated into the NBLAST and XBLAST programs of Altschul et al., 1990, J. M01. _25_ Biol. 215:403 —410. BLAST nucleotide searches can be performed with the NBLAST program, score=100, wordlength=12, to obtain nucleotide sequences homologous to a nucleic acid encoding a n of interest. BLAST protein searches can be performed with the XBLAST program, score=5 0, ngth=3, to obtain amino acid sequences gous to a protein of st. To obtain gapped alignments for comparison purposes, Gapped BLAST can be ed as described in Altschul et al., 1997, c Acids Res. :3389—3402. Alternatively, PSI—Blast can be used to m an iterated search which detects distant onships between molecules (Id). When utilizing BLAST, Gapped BLAST, and PSI—Blast programs, the default parameters of the respective programs (e. g., XBLAST and NBLAST) can be used. Another preferred, non-limiting example of a mathematical algorithm utilized for the comparison of sequences is the algorithm of Myers and Miller, CABIOS (1989). Such an algorithm is incorporated into the ALIGN program (version 2.0) which is part of the GCG sequence alignment software package. When utilizing the ALIGN program for comparing amino acid sequences, a PAM120 weight residue table, a gap length y of 12, and a gap penalty of 4 can be used. Additional algorithms for sequence analysis are known in the art and include ADVANCE and ADAM as described in Torellis and Robotti, 1994, Comput. Appl. Biosci. 10:3—5; and FASTA described in Pearson and Lipman, 1988, Proc. Natl. Acad. Sci. USA 85:2444-8. Within FASTA, ktup is a control option that sets the sensitivity and speed of the search. If ktup=2, similar regions in the two sequences being compared are found by looking at pairs of aligned residues; if ktup=l, single aligned amino acids are examined. ktup can be set to 2 or 1 for protein sequences, or from 1 to 6 for DNA sequences. The default if ktup is not specified is 2 for proteins and 6 for DNA. Alternatively, protein sequence alignment may be carried out using the CLUSTAL W thm, as described by Higgins et al., 1996, Methods l. 266:3 83—402.

The term “covalently linked” as used herein means either a direct covalent bond between es, or an indirect association where two es are not directly bonded but are both covalently bonded to an intermediate molecule or domain, e. g. an intermediate domain of an immunoglobulin.

The terms "compete" or "cross—compete" are used interchangeably herein to refer to the ability of an antibody molecule to ere with binding of an antibody molecule, e.g., an anti-DLL3 dy molecule of the invention, to a target, e.g., human DLL3. The interference with binding can be direct or indirect (e.g., through an eric modulation of —26— the antibody molecule or the target). The extent to which an antibody molecule is able to interfere with the binding of another antibody molecule to the , and ore r it can be said to compete, can be determined using a competition binding assay, for example, a FACS assay, an ELISA or BIACORE assay. In some embodiments, a competition binding assay is a quantitative competition assay. In some embodiments, a ﬁrst anti-DLL3 antibody molecule is said to compete for binding to the target with a second anti- DLL3 antibody molecule when the binding of the ﬁrst antibody molecule to the target is reduced by 10% or more, e.g., 20% or more, 30% or more, 40% or more, 50% or more, 55% or more, 60% or more, 65% or more, 70% or more, 75% or more, 80% or more, 85% or more, 90% or more, 95% or more, 98% or more, 99% or more in a competition binding assay (e. g., a competition assay described herein).

It is to be inferred without explicit recitation and unless otherwise intended, that when the present technology relates to a polypeptide, protein, polynucleotide or antibody, an equivalent or a biologically equivalent of such is intended within the scope of the present technology.

As used herein, the term "biological equivalent f” is intended to be synonymous with "equivalent thereof” when referring to a reference protein, antibody, polypeptide, polynucleotide or nucleic acid, and intends those having l homology while still maintaining desired ure or functionality. Unless speciﬁcally recited herein, it is plated that any nucleic acid, polynucleotide, polypeptide, protein or antibody mentioned herein also includes equivalents thereof. For example, an equivalent intends at least about 80 % homology or identity and alternatively, at least about 85 %, or atively at least about 90 %, or alternatively at least about 95 %, or alternatively 98 % percent homology or identity and exhibits ntially equivalent biological activity to the reference protein, polypeptide, antibody or nucleic acid.

As used herein, the term "detectable label" refers to a molecule, compound or composition that can produce a detectable , i.e. physical or chemical signal including a metric, cent, electrical, radioactive and chemiluminescent signal, which can be measured by visual or instrumental methods, and which indicates the presence and/or quantity/concentration of the label in a sample.

A “detectable signal” can be generated by various mechanism ing tion, emission and/or scattering of a photon (including radio frequency, microwave frequency, _27_ infrared frequency, visible frequency and ultra—violet frequency photons) and includes but is not limited to colorimetric, ﬂuorescent, electrical, radioactive and chemiluminescent signals.

As used herein, "expression" refers to the s by which polynucleotides are transcribed into mRNA and/or the process by which the ribed mRNA is subsequently being translated into es, polypeptides, or proteins. If the polynucleotide is derived from genomic DNA, expression may e splicing of the mRNA in an eukaryotic cell. The expression level of a gene may be determined by measuring the amount of mRNA or protein in a cell or tissue sample.

As used herein the term “biological sample” means a sample material derived from or contacted by living cells. The term is intended to include s, cells and biological ﬂuids isolated from a subject. As used herein, the term e sample" shall refer to a cellular sample that preserves the cross-sectional spatial relationship between the cells as they existed within the t from which the sample was obtained. Biological samples can also be obtained from es of internal organs or from cancers.

“Histochemistry” and cytochemistry” are techniques used to identify a le within the t of intact cells by labeling the samples with an agent that binds speciﬁcally to the molecule in a manner than can be visualized on a microscope. ”Immunohistochemistry” and “immunocytochemistry” are types of histochemistry and cytochemistry that use antibodies to label the molecules.

MULTI—SPECIFIC BINDING PROTEINS OF THE INVENTION The present ion provides multi-speciﬁc binding proteins comprising at least one antigen binding unit speciﬁcally binding to DLL3 (a ﬁrst antigen binding unit), and at least one antigen binding unit speciﬁcally binding to CD3 (a second antigen binding unit). Such (multi-speciﬁc) binding proteins are also referred to herein as (multi—speciﬁc) binding molecules or DLL3/CD3 g proteins or DLL3/CD3 binding molecules.

The ors have surprisingly found that multi-speciﬁc binding proteins of the invention induce selective lysis of DLL3—positive SCLC cell lines in the presence of T cells and are already active at low effector to target cell ratios. Importantly, the binding proteins of the —28- invention do not cause T cell activation, T cell proliferation, and Cytokine secretion in the absence of DLL3—positive cells or lysis of DLL3 —negative cells.

For avoidance of doubt, DLL3 as used herein refers to human DLL3 of UniProt Q9NYJ7 and the nucleic acid sequence encoding that n. CD3 as used herein refers to human CD3epsilon (UniProt P07766) and CD3 gamma (Uniprot: P09693) complexes (human CD38y xes).

Targeting DLL3 with a bispeciﬁc T cell engaging approach is expected to provide advantages over an ADC approach, as redirecting T cells is not inﬂuenced by resistance to chemotherapy and low expression levels on the cell e are less critical for this mode of action. T cell engagers are multi-speciﬁc binding proteins with g arms to CD3 on T cells and binding arms to an antigen on the cell surface oftumor cells. Through simultaneously binding to T cells and tumor cells, the T cell engagers force formation of a cytolytic synapse between the two cells and so, redirect the T cell activity selectively to the targeted tumor cells.

In one aspect, the multi-specific binding protein of the invention comprises a first antigen binding unit speciﬁcally g to DLL3 and a second antigen binding unit speciﬁcally binding to CD3, wherein said first g unit is selected from the group consisting of i) to xviii): i) an antigen binding unit comprising light chain CDRs comprising the amino acid sequences of SEQ ID NO:1 (CDRl), SEQ ID NO:2 (CDR2) and SEQ ID NO:3 (CDR3) and heavy chain CDRs comprising the amino acid sequences of SEQ ID NO:4 , SEQ ID NO:5 (CDR2) and SEQ ID NO:6 (CDR3) (antigen binding unit ); ii) an antigen binding unit comprising light chain CDRs comprising the amino acid ces of SEQ ID NO:7 (CDRl), SEQ ID N028 (CDR2) and SEQ ID NO:9 (CDR3) and heavy chain CDRs comprising the amino acid sequences of SEQ ID NO:10 (CDRl), SEQ ID NO:11 (CDR2) and SEQ ID NO:12 (CDR3) (antigen binding unit DLL3#2); iii) an antigen binding unit comprising light chain CDRs sing the amino acid ces of SEQ ID NO:l3 (CDRl), SEQ ID NO:l4 (CDR2) and SEQ ID N0215 (CDR3) and heavy chain CDRs comprising the amino acid _29_ WO 20191234220 sequences of SEQ ID NO:16 (CDRl), SEQ ID NO]? (CDR2) and SEQ ID NO:18 (CDR3) (antigen binding unit DLL3#3); iv) an antigen binding unit comprising light chain CDRs comprising the amino acid sequences of SEQ ID NO:19 (CDRl), SEQ ID NO:20 (CDR2) and SEQ ID NO:2l (CDR3) and heavy chain CDRs comprising the amino acid sequences of SEQ ID NO:22 (CDRl), SEQ ID NO:23 (CDR2) and SEQ ID NO:24 (CDR3) (antigen binding unit DLL3#4); an antigen binding unit sing light chain CDRs comprising the amino acid ces of SEQ ID NO:25 (CDRl), SEQ ID NO:26 (CDR2) and SEQ ID NO:27 (CDR3) and heavy chain CDRs comprising the amino acid sequences of SEQ ID NO:28 (CDRl), SEQ ID NO:29 (CDR2) and SEQ ID NO:30 (CDR3) (antigen g unit DLL3#5); vi) an antigen binding unit comprising light chain CDRs comprising the amino acid sequences of SEQ ID NO:3l (CDRl), SEQ ID NO:32 (CDR2) and SEQ ID NO:33 (CDR3) and heavy chain CDRs sing the amino acid sequences of SEQ ID NO:34 (CDRl), SEQ ID NO:35 (CDR2) and SEQ ID NO:36 (CDR3) (antigen binding unit DLL3#6); vii) an antigen binding unit comprising light chain CDRs comprising the amino acid sequences of SEQ ID NO:133 (CDRl), SEQ ID NO:134 (CDR2) and SEQ ID NO:135 (CDR3) and heavy chain CDRs comprising the amino acid ces of SEQ ID NO:136 (CDRl), SEQ ID NO:137 (CDR2) and SEQ ID NO:138 (CDR3) (antigen binding unit DLL3#7); viii) an antigen binding unit comprising light chain CDRs comprising the amino acid sequences of SEQ ID NO:139 (CDRI), SEQ ID NO:140 (CDR2) and SEQ ID NO:14l (CDR3) and heavy chain CDRs comprising the amino acid sequences of SEQ ID NO:142 (CDRl), SEQ ID NO:143 (CDR2) and SEQ ID NO: 144 (CDR3) (antigen binding unit ); ix) an antigen binding unit comprising light chain CDRs comprising the amino acid sequences of SEQ ID NO:145 (CDRl), SEQ ID NO:146 (CDR2) and SEQ ID NO:147 (CDR3) and heavy chain CDRs comprising the amino acid sequences of SEQ ID NO:148 (CDRl), SEQ ID NO:149 (CDR2) and SEQ ID NO:150 (CDR3) (antigen binding unit DLL3#9); _30_ WO 20191234220 X) an antigen g unit comprising light chain CDRs comprising the amino acid sequences of SEQ ID NO:151 (CDRl), SEQ ID NO:152 (CDR2) and SEQ ID NO:153 (CDR3) and heavy chain CDRs comprising the amino acid sequences of SEQ ID NO:154 (CDRl), SEQ ID NO:155 (CDR2) and SEQ ID NO:156 (CDR3) (antigen binding unit DLL3#10); xi) an antigen binding unit sing light chain CDRs comprising the amino acid sequences of SEQ ID NO:157 (CDRl), SEQ ID NO:158 (CDR2) and SEQ ID NO:159 (CDR3) and heavy chain CDRs comprising the amino acid sequences of SEQ ID NO:160 (CDRl), SEQ ID NO:161 (CDR2) and SEQ ID NO:162 (CDR3) (antigen binding unit DLL3#1 1); xii) an antigen binding unit comprising light chain CDRs comprising the amino acid sequences of SEQ ID NO:163 (CDRl), SEQ ID NO:164 (CDR2) and SEQ ID NO:165 (CDR3) and heavy chain CDRs comprising the amino acid sequences of SEQ ID NO:166 (CDRl), SEQ ID NO:167 (CDR2) and SEQ ID NO:168 (CDR3) (antigen binding unit DLL3#12); xiii) an antigen binding unit comprising light chain CDRs comprising the amino acid sequences of SEQ ID NO:169 (CDRl), SEQ ID NO:170 (CDR2) and SEQ ID NO:171 (CDR3) and heavy chain CDRs comprising the amino acid sequences of SEQ ID NO:172 (CDRl), SEQ ID NO:173 (CDR2) and SEQ ID NO:174 (CDR3) (antigen binding unit DLL3#13); xiv) an antigen binding unit comprising light chain CDRs comprising the amino acid ces of SEQ ID NO:175 (CDRl), SEQ ID NO:176 (CDR2) and SEQ ID NO:177 (CDR3) and heavy chain CDRs comprising the amino acid sequences of SEQ ID NO:178 (CDRl), SEQ ID NO:179 (CDR2) and SEQ ID NO:180 (CDR3) (antigen binding unit 4); xv) an antigen g unit comprising light chain CDRs comprising the amino acid ces of SEQ ID NO:181 (CDRl), SEQ ID NO:182 (CDR2) and SEQ ID NO:183 (CDR3) and heavy chain CDRs comprising the amino acid sequences of SEQ ID NO:184 (CDRl), SEQ ID NO:185 (CDR2) and SEQ ID NO:186 (CDR3) (antigen binding unit DLL3#15); xvi) an antigen binding unit comprising light chain CDRs comprising the amino acid sequences of SEQ ID NO:187 (CDRl), SEQ ID NO:188 (CDR2) and _31_ 2019/064942 SEQ ID NO:189 (CDR3) and heavy chain CDRs comprising the amino acid sequences of SEQ ID NO:190 (CDRl), SEQ ID NO:191 (CDR2) and SEQ ID NO:192 (CDR3) (antigen binding unit DLL3#16); xvii) an n binding unit comprising light chain CDRs comprising the amino acid sequences of SEQ ID NO:193 (CDRl), SEQ ID NO:194(CDR2) and SEQ ID NO:l95 (CDR3) and heavy chain CDRs comprising the amino acid sequences of SEQ ID NO:196 (CDRl), SEQ ID NO:197 (CDR2) and SEQ ID NO:198 (CDR3) (antigen binding unit DLL3#17); and xviii) an n binding unit comprising light chain CDRs comprising the amino acid sequences of SEQ ID NO:199 (CDRI), SEQ ID NO:200 (CDR2) and SEQ ID NO:201 (CDR3) and heavy chain CDRs comprising the amino acid ces of SEQ ID NO:202 (CDRl), SEQ ID NO:203 (CDR2) and SEQ ID NO:204 (CDR3) (antigen binding unit DLL3#18).

Preferably, the ﬁrst antigen binding unit of the binding n of the invention is any one of i) to iii) as deﬁned by the CDR sequences above.

For the avoidance of doubt, each of the speciﬁc embodiments listed herein can also be considered independent aspects of the invention.

In some embodiments of the binding protein of the invention, said second antigen binding unit speciﬁcally binding to CD3 is selected from the group consisting of i) to iii): i) an antigen binding unit comprising light chain CDRs comprising the amino acid sequences of SEQ ID NO:55 (CDRl), SEQ ID NO:56 (CDR2) and SEQ ID NO:57 (CDR3) and heavy chain CDRs comprising the amino acid sequences of SEQ ID NO:58 (CDRl), SEQ ID NO:59 (CDR2) and SEQ ID NO:60 (CDR3) (antigen binding unit CD3#1); ii) an antigen binding unit comprising light chain CDRs comprising the amino acid ces of SEQ ID NO:61 (CDRl), SEQ ID NO:62 (CDR2) and SEQ ID NO:63 (CDR3) and heavy chain CDRs comprising the amino acid sequences of SEQ ID NO:64 (CDRl), SEQ ID NO:65 (CDR2) and SEQ ID NO:66 (CDR3) (antigen binding unit CD3#2); and iii) an antigen binding unit comprising light chain CDRs sing the amino acid sequences of SEQ ID NO:96 (CDRl), SEQ ID NO:97 (CDR2) and SEQ ID _32_ N0298 (CDR3) and heavy chain CDRs comprising the amino acid sequences of SEQ ID NO:99 (CDRl), SEQ ID NO:100 (CDR2) and SEQ ID NO:101 (CDR3) (antigen binding unit CD3#3).

The ﬁrst antigen binding units i) to xviii) as outlined above are termed DLL3#1, DLL3#2, DLL3#3, DLL3#4, DLL3#5, DLL3#6, DLL3#7, DLL3#8, DLL3#9, DLL3#10, DLL3#11, DLL3#12, DLL3#13, DLL3#14, DLL3#15, 6, DLL3#17 and DLL3#18, respectively and the second antigen binding units i) to iii) as outlined above are termed CD3#1, CD3#2 and CD3#3, respectively. Provided herein is a sequence table which y allows identiﬁcation of individual amino acid sequences to speciﬁc antigen binding units and full length binding proteins of the present invention. A summary is ed in Table l in Example 2.

The terms “ﬁrst” and “second” with respect to antigen binding units in general, as used herein, is solely intended to indicate that these units are two different units (as they bind to different target antigens). Thus, these terms shall not be understood to refer to the exact order or sequence of the units within the binding n of the invention.

In some embodiments, the binding protein of the invention comprises a ﬁrst antigen binding unit selected from the group consisting of DLL3#1, , DLL3#3, DLL3#4, DLL3#5, DLL3#6, DLL3#7, DLL3#8, DLL3#9, DLL3#10, DLL3#11, DLL3#12, DLL3#13, 4, DLL3#15, DLL3#16, DLL3#17 and DLL3#18 as deﬁned by the respective CDR sequences described above and a second antigen binding unit of CD3#1 as deﬁned by the respective CDR sequences described above. In preferred embodiments, the binding protein of the ion comprises a ﬁrst antigen binding unit selected from the group consisting of DLL3#1, DLL3#2, and DLL3#3 as deﬁned by the respective CDR sequences described above and a second antigen binding unit of CD3#1 as deﬁned by the respective CDR sequences described above.

In some ments, the binding protein of the invention comprises a ﬁrst antigen binding unit selected from the group consisting of DLL3#1, DLL3#2, DLL3#3, DLL3#4, DLL3#5, DLL3#6, DLL3#7, DLL3#8, DLL3#9, DLL3#10, 1, 2, DLL3#l3, DLL3#14, DLL3#15, DLL3#l6, DLL3#l7 and DLL3#l8 as deﬁned by the respective CDR sequences described above and a second antigen binding unit of CD3#2 as deﬁned by the tive CDR sequences described above. In preferred embodiments, the g n of the invention ses a ﬁrst antigen binding unit selected from the _33_ WO 34220 group consisting of DLL3#1, , and DLL3#3 as deﬁned by the respective CDR sequences described above and a second n g unit of CD3#2 as deﬁned by the respective CDR sequences described above.

In some embodiments, the binding protein of the invention comprises a ﬁrst antigen binding unit selected from the group ting of , DLL3#2, DLL3#3, DLL3#4, DLL3#5, DLL3#6, DLL3#7, DLL3#8, DLL3#9, DLL3#10, DLL3#11, DLL3#12, DLL3#l3, DLL3#14, DLL3#15, DLL3#l6, 7 and DLL3#l8 as deﬁned by the respective CDR sequences described above and a second antigen binding unit of CD3#3 as deﬁned by the respective CDR sequences described above. In preferred embodiments, the binding n of the invention comprises a ﬁrst n g unit selected from the group consisting of DLL3#1, DLL3#2, and DLL3#3 as deﬁned by the respective CDR sequences described above and a second n binding unit of CD3#3 as deﬁned by the respective CDR ces described above.

In one preferred embodiment, the binding protein of the invention comprises a ﬁrst antigen binding unit speciﬁcally binding to DLL3, comprising light chain CDRs comprising the amino acid sequences of SEQ ID N021 (CDRl), SEQ ID N022 (CDR2) and SEQ ID N023 (CDR3) and heavy chain CDRs comprising the amino acid sequences of SEQ ID N024 (CDRl), SEQ ID N025 (CDR2) and SEQ ID N026 (CDR3) and a second antigen binding unit speciﬁcally binding to CD3, comprising light chain CDRs comprising the amino acid sequences of SEQ ID NO:55 (CDRl), SEQ ID NO:56 (CDR2) and SEQ ID NO:57 (CDR3) and heavy chain CDRs comprising the amino acid sequences of SEQ ID N0258 (CDRl), SEQ ID NO:59 (CDR2) and SEQ ID N0260 (CDR3).

In one preferred embodiment, the binding protein of the invention comprises a ﬁrst antigen binding unit speciﬁcally binding to DLL3, comprising light chain CDRs comprising the amino acid sequences of SEQ ID N027 (CDRl), SEQ ID N028 (CDR2) and SEQ ID NO:9 (CDR3) and heavy chain CDRs comprising the amino acid sequences of SEQ ID N0210 (CDRl), SEQ ID N0: 11 (CDR2) and SEQ ID N0:12 (CDR3) and a second antigen binding unit speciﬁcally binding to CD3, comprising light chain CDRs comprising the amino acid sequences of SEQ ID NO:55 (CDRl), SEQ ID NO:56 (CDR2) and SEQ ID N0257 (CDR3) and heavy chain CDRs comprising the amino acid sequences of SEQ ID NO:58 (CDRl), SEQ ID NO:59 (CDR2) and SEQ ID N0260 (CDR3). _34_ In one red embodiment, the binding protein of the invention comprises a ﬁrst antigen binding unit speciﬁcally binding to DLL3, comprising light chain CDRs comprising the amino acid ces of SEQ ID NO: 13 (CDRl), SEQ ID NO:14 (CDR2) and SEQ ID NO:15 (CDR3) and heavy chain CDRs comprising the amino acid sequences of SEQ ID NO:16 (CDRl), SEQ ID NO: 17 (CDR2) and SEQ ID NO:18 (CDR3) and a second antigen binding unit speciﬁcally binding to CD3, comprising light chain CDRs comprising the amino acid sequences of SEQ ID N0255 (CDRl), SEQ ID NO:56 (CDR2) and SEQ ID N0:57 (CDR3) and heavy chain CDRs comprising the amino acid sequences of SEQ ID N0258 (CDRl), SEQ ID N0259 (CDR2) and SEQ ID NO:60 (CDR3).

In one red embodiment, the binding protein of the invention comprises a ﬁrst antigen binding unit speciﬁcally binding to DLL3, comprising light chain CDRs comprising the amino acid sequences of SEQ ID N0:l (CDRl), SEQ ID N022 (CDR2) and SEQ ID N023 (CDR3) and heavy chain CDRs comprising the amino acid sequences of SEQ ID NO:4 (CDRl), SEQ ID N025 (CDR2) and SEQ ID NO:6 (CDR3) and a second antigen binding unit speciﬁcally binding to CD3 comprising light chain CDRs comprising the amino acid sequences of SEQ ID NO:61 (CDRl), SEQ ID NO:62 (CDR2) and SEQ ID NO:63 (CDR3) and heavy chain CDRs comprising the amino acid sequences of SEQ ID NO:64 , SEQ ID NO:65 (CDR2) and SEQ ID NO:66 .

In one preferred embodiment, the binding protein of the invention ses a ﬁrst antigen g unit speciﬁcally binding to DLL3, comprising light chain CDRs comprising the amino acid sequences of SEQ ID NO:7 (CDRl), SEQ ID NO:8 (CDR2) and SEQ ID NO:9 (CDR3) and heavy chain CDRs comprising the amino acid sequences of SEQ ID NO:10 (CDRl), SEQ ID NO:11 (CDR2) and SEQ ID NO:12 (CDR3) and a second antigen binding unit speciﬁcally g to CD3, comprising light chain CDRs comprising the amino acid sequences of SEQ ID N0261 (CDRl), SEQ ID NO:62 (CDR2) and SEQ ID NO:63 (CDR3) and heavy chain CDRs comprising the amino acid sequences of SEQ ID NO:64 (CDRl), SEQ ID NO:65 (CDR2) and SEQ ID NO:66 (CDR3).

In one preferred embodiment, the binding protein of the invention comprises a ﬁrst antigen g unit speciﬁcally binding to DLL3, comprising light chain CDRs comprising the amino acid sequences of SEQ ID NO: 13 (CDRl), SEQ ID NO:14 (CDR2) and SEQ ID NO:15 (CDR3) and heavy chain CDRs comprising the amino acid sequences of SEQ ID NO:16 (CDRl), SEQ ID NO: 17 (CDR2) and SEQ ID NO:18 (CDR3) and a second antigen _35_ binding unit cally binding to CD3, comprising light chain CDRs comprising the amino acid sequences of SEQ ID NO:61 (CDRl), SEQ ID NO:62 (CDR2) and SEQ ID NO:63 (CDR3) and heavy chain CDRs sing the amino acid sequences of SEQ ID NO:64 (CDRl), SEQ ID NO:65 (CDR2) and SEQ ID NO:66 (CDR3).

In addition to the CDR sequences as set out herein, the antigen binding units of the binding proteins of the invention include immunoglobulin framework region (FR) sequences.

These ces are preferably not immunogenic in humans, and are therefore preferably human or humanized FR sequences. Suitable human or humanized FR sequences are known in the art. Specifically preferred FR sequences can be taken from the embodiments shown herein, disclosing the te n binding units and thereby CDR sequences as well as FR sequences.

In red embodiments of the binding proteins of the invention, the ﬁrst and the second binding unit each comprise a light chain variable domain and a heavy chain variable domain derived from an dy le, said light/heavy chain variable domains deﬁned by the CDR sequences of any one of DLL3#1, DLL3#2, DLL3#3, DLL3#4, DLL3#5, , DLL3#7, DLL3#8, DLL3#9, DLL3#10, DLL3#11, DLL3#12, DLL3#13, DLL3#14, DLL3#15, DLL3#16, DLL3#17 or 8 for the first antigen binding unit and said light/heavy chain variable s deﬁned by the CDR sequences of any one of CD3#1, CD3#2 or CD3#3 for the second antigen binding unit. In some embodiments of the binding protein of the invention, the VH and/or VL domain of the antigen binding units of any one or more of DLL3#1, DLL3#2, DLL3#3, DLL3#4, DLL3#5, DLL3#6, DLL3#7, DLL3#8, DLL3#9, DLL3#10, DLL3#11, DLL3#12, DLL3#13, DLL3#14, DLL3#15, DLL3#16, DLL3#17, DLL3#18, CD3#1, CD3#2 or CD3#3 is a human or humanized VH and/or VL domain.

In preferred embodiments of the invention, the light/heavy chain le domains of the first antigen binding unit are further defined as follows i) a light chain variable domain comprising the amino acid sequences of SEQ ID N037 and heavy chain variable domain comprising the amino acid sequences of SEQ ID NO:38 (antigen g unit DLL3#1); or ii) a light chain variable domain comprising the amino acid sequences of SEQ ID N039 and heavy chain variable domain comprising the amino acid sequences of SEQ ID NO:40 (antigen g unit DLL3#2); or _36_ iii) a light chain variable domain comprising the amino acid sequences of SEQ ID NO:41 and heavy chain variable domain comprising the amino acid ces of SEQ ID NO:42 (antigen binding unit DLL3#3); or iv) a light chain le domain comprising the amino acid sequences of SEQ ID NO:43 and heavy chain variable domain comprising the amino acid sequences of SEQ ID NO:44 (antigen binding unit DLL3#4); or a light chain variable domain comprising the amino acid sequences of SEQ ID NO:45 and heavy chain variable domain comprising the amino acid sequences of SEQ ID NO:46 (antigen binding unit DLL3#5); or vi) a light chain variable domain comprising the amino acid sequences of SEQ ID NO:47 and heavy chain variable domain comprising the amino acid sequences of SEQ ID NO:48 (antigen binding unit DLL3#6); or vii) a light chain variable domain comprising the amino acid sequence of SEQ ID NO:205 and heavy chain variable domain comprising the amino acid ce of SEQ ID NO:206 (antigen g unit DLL3#7); or viii) a light chain variable domain comprising the amino acid sequence of SEQ ID NO:207 and heavy chain variable domain comprising the amino acid sequence of SEQ ID NO:208 (antigen binding unit DLL3#8); or ix) a light chain le domain comprising the amino acid sequence of SEQ ID NO:209 and heavy chain variable domain comprising the amino acid sequence of SEQ ID NO:210 (antigen binding unit DLL3#9); or a light chain variable domain comprising the amino acid sequence of SEQ ID N02 11 and heavy chain variable domain comprising the amino acid sequence of SEQ ID N022 12 (antigen binding unit DLL3#10); or Xi) a light chain le domain comprising the amino acid sequence of SEQ ID NO:213 and heavy chain variable domain comprising the amino acid ce of SEQ ID N02 14 (antigen binding unit DLL3#l l); or xii) a light chain variable domain comprising the amino acid sequence of SEQ ID N022 15 and heavy chain variable domain comprising the amino acid sequence of SEQ ID N02 16 (antigen binding unit DLL3#l2); or xiii) a light chain variable domain comprising the amino acid sequence of SEQ ID N02 17 and heavy chain variable domain sing the amino acid sequence of SEQ ID NO:218 (antigen g unit DLL3#l3); or _37_ xiv) a light chain variable domain comprising the amino acid sequence of SEQ ID NO:219 and heavy chain variable domain comprising the amino acid sequence of SEQ ID NO:220 (antigen binding unit DLL3#14); or xv) a light chain variable domain sing the amino acid sequence of SEQ ID NO:221 and heavy chain variable domain comprising the amino acid sequence of SEQ ID NO:222 (antigen binding unit DLL3#15); or xvi) a light chain variable domain comprising the amino acid sequence of SEQ ID NO:223 and heavy chain variable domain comprising the amino acid sequence of SEQ ID NO:224 (antigen binding unit DLL3#16); or xvii) a light chain variable domain sing the amino acid sequence of SEQ ID NO:225 and heavy chain variable domain comprising the amino acid sequence of SEQ ID NO:226 en binding unit DLL3#l7); or xviii) sing a light chain variable domain comprising the amino acid ce of SEQ ID NO:227 and heavy chain variable domain comprising the amino acid sequence of SEQ ID NO:228 (antigen binding unit DLL3#18).

Preferably, the ﬁrst antigen binding unit of the binding protein of the invention is any one of i) to iii) as defined by the VL and VH sequences above.

In preferred embodiments of the invention, the light/heavy chain variable domains of the second antigen binding unit are further defined as follows i) a light chain variable domain comprising the amino acid sequence of SEQ ID NO:67 and a heavy chain le domain sing the amino acid sequence of SEQ ID NO:68 (antigen binding unit CD3#l); or ii) a light chain variable domain comprising the amino acid sequence of SEQ ID NO:69 and a heavy chain variable domain comprising the amino acid sequence of SEQ ID NO:70 (antigen binding unit CD3#2); or iii) a light chain variable domain sing the amino acid sequence of SEQ ID NO:102 and a heavy chain variable domain comprising the amino acid sequence of SEQ ID NO:103 (antigen binding unit CD3#3).

In some embodiments, the binding n of the ion comprises a combination of a ﬁrst and a second antigen binding unit selected from the group consisting of DLL3#l/CD3#1, DLL3#2/CD3#1, DLL3#3/CD3#1, DLL3#4/CD3#1, DLL3#5/CD3#1, DLL3#6/CD3#1, /CD3#1, DLL3#8/CD3#1, DLL3#9/CD3#1, DLL3#10/CD3#1, _38_ DLL3#11/CD3#1, DLL3#12/CD3#1, DLL3#13/CD3#1, DLL3#14/CD3#1, DLL3#15/CD3#1, DLL3#16/CD3#1, DLL3#l7/CD3#1, DLL3#18/CD3#1, DLL3#1/CD3#2, /CD3#2, DLL3#3/CD3#2, DLL3#4/CD3#2, DLL3#5/CD3#2, DLL3#6/CD3#2, /CD3#2, DLL3#8/CD3#2, DLL3#9/CD3#2, DLL3#lO/CD3#2, 1/CD3#2, DLL3#l2/CD3#2, DLL3#13/CD3#2, DLL3#14/CD3#2, DLL3#15/CD3#2, DLL3#16/CD3#2, DLL3#l7/CD3#2, DLL3#18/CD3#2,DLL3#l/CD3#3, DLL3#2/CD3#3, DLL3#3/CD3#3, DLL3#4/CD3#3, DLL3#5/CD3#3, /CD3#3, DLL3#7/CD3#3, DLL3#8/CD3#3, DLL3#9/CD3#3, DLL3#10/CD3#3, DLL3#1 l/CD3#3, DLL3#12/CD3#3, DLL3#l3/CD3#3, DLL3#l4/CD3#3, 5/CD3#3, DLL3#l6/CD3#3, DLL3#l7/CD3#3, DLL3#l 8/CD3#3, the ﬁrst and second antigen binding unit being deﬁned by the CDR and/or VH and VL sequences of the antigen binding units as described above. In preferred embodiments, the binding protein of the invention comprises a combination of a ﬁrst and a second antigen binding unit selected from the group consisting of DLL3#l/CD3#1, DLL3#2/CD3#1, DLL3#3/CD3#1, DLL3#1/CD3#2, /CD3#2, DLL3#3/CD3#2, DLL3#l/CD3#3, DLL3#2/CD3#3, and DLL3#3/CD3#3, the ﬁrst and second antigen binding unit being deﬁned by the CDR and/or VH and VL sequences of the antigen binding units as described above.

In one red ment, the binding protein of the invention comprises (i) a ﬁrst n binding unit speciﬁcally binding to DLL3 comprising a light chain variable domain comprising the amino acid sequences of SEQ ID NO:37 and a heavy chain variable domain comprising the amino acid sequences of SEQ ID NO:38 and (ii) a second antigen g unit speciﬁcally binding to CD3 comprising a light chain variable domain comprising the amino acid sequence of SEQ ID NO:67 and a heavy chain variable domain comprising the amino acid sequence of SEQ ID NO:68.

In one preferred embodiment, the binding protein of the invention comprises (i) a ﬁrst antigen binding unit speciﬁcally binding to DLL3 comprising a light chain le domain sing the amino acid sequences of SEQ ID N039 and a heavy chain variable domain comprising the amino acid sequences of SEQ ID NO:40 and (ii) a second antigen binding unit speciﬁcally binding to CD3 comprising a light chain variable domain comprising the amino acid sequence of SEQ ID NO:67 and a heavy chain variable domain comprising the amino acid sequence of SEQ ID NO:68. _39_ In one preferred embodiment, the binding protein of the invention comprises (i) a ﬁrst antigen binding unit cally binding to DLL3 comprising a light chain variable domain comprising the amino acid sequences of SEQ ID NO:41 and heavy chain le domain comprising the amino acid sequences of SEQ ID NO:42 and (ii) a second antigen binding unit speciﬁcally binding to CD3 comprising a light chain le domain sing the amino acid sequence of SEQ ID NO:67 and a heavy chain le domain comprising the amino acid sequence of SEQ ID NO:68.

In one red embodiment, the binding protein of the invention comprises (i) a ﬁrst n binding unit speciﬁcally g to DLL3 comprising a light chain variable domain sing the amino acid sequences of SEQ ID NO:37 and heavy chain variable domain comprising the amino acid sequences of SEQ ID NO:38 and (ii) a second n binding unit speciﬁcally binding to CD3 comprising a light chain variable domain comprising the amino acid sequence of SEQ ID NO:69 and a heavy chain variable domain comprising the amino acid sequence of SEQ ID NO:70.

In one preferred embodiment, the binding protein of the invention ses (i) a ﬁrst antigen g unit speciﬁcally binding to DLL3 comprising a light chain variable domain comprising the amino acid sequences of SEQ ID NO:39 and heavy chain variable domain comprising the amino acid sequences of SEQ ID NO:40 and (ii) a second antigen binding unit speciﬁcally g to CD3 comprising a light chain variable domain comprising the amino acid sequence of SEQ ID NO:69 and a heavy chain variable domain comprising the amino acid sequence of SEQ ID NO:70.

In one preferred embodiment, the binding protein of the invention comprises (i) a ﬁrst antigen binding unit speciﬁcally binding to DLL3 comprising a light chain variable domain comprising the amino acid sequences of SEQ ID NO:4l and heavy chain variable domain comprising the amino acid sequences of SEQ ID NO:42 and a second antigen binding unit speciﬁcally binding to CD3 comprising a light chain variable domain comprising the amino acid sequence of SEQ ID NO:69 and a heavy chain variable domain comprising the amino acid sequence of SEQ ID NO:70.

In some embodiments, the binding protein of the invention comprises i) a ﬁrst antigen binding unit speciﬁcally binding to DLL3 (e.g. any one of DLL3#1, DLL3#2, DLL3#3, DLL3#4, DLL3#5, DLL3#6, DLL3#7, DLL3#8, DLL3#9, DLL3#10, DLL3#ll, 2, DLL3#13, DLL3#14, DLL3#15, DLL3#16, DLL3#17 or DLL3#18 as deﬁned _40_ by the respective CDR or VH/VL sequences above) which comprises a ﬁrst light chain le domain ntly linked, either directly or indirectly, to a ﬁrst heavy chain variable domain with a ﬁrst peptide linker and/or ii) a second antigen binding unit speciﬁcally g to CD3 (e.g. any one of CD3#1, CD3#2, or CD3#3 as deﬁned by the respective CDR or VH/VL sequences above) which comprises a second light chain variable domain covalently linked, either directly or indirectly, to a second heavy chain variable domain with a second peptide linker.

In some embodiments of the binding proteins of the invention, the ﬁrst and/or the second antigen binding unit further comprises a CL and a CH1 domain like in a light/heavy Fab domain of a conventional antibody molecule, thus said ﬁrst g unit comprises a) a VL domain (e.g., deﬁned by the light chain CDR (LCCDR) or VL sequences of any one of DLL3#l, DLL3#2, , DLL3#4, DLL3#5, DLL3#6, DLL3#7, DLL3#8, DLL3#9, DLL3#10, DLL3#11, DLL3#12, DLL3#13, DLL3#14, DLL3#15, DLL3#16, DLL3#17 or 8) covalently linked (directly or indirectly bound) to a ﬁrst CL domain and b) a VH domain (e. g., deﬁned by the heavy chain CDR (HCCDR) or VH sequences of any one of DLL3#1, DLL3#2, DLL3#3, DLL3#4, DLL3#5, DLL3#6 DLL3#7, DLL3#8, DLL3#9, DLL3#10, DLL3#11, DLL3#12, DLL3#13, DLL3#14, DLL3#15, DLL3#16, DLL3#17 or DLL3#18) covalently linked (directly or ctly bound) to a ﬁrst CH1 domain and/or said second antigen binding unit comprises a) a VL domain (e. g., deﬁned by the LCCDR or VL sequences of any one of CD3#1, CD3#2 or CD3#3) covalently linked (directly or indirectly bound) to a second CL domain and b) a VH domain (e.g., deﬁned by HCCDR or VH sequences of any one of CD3#1, CD3#2 or CD3#3) covalently linked (directly or indirectly bound) to a second CHl domain.

In the context of the present invention, a CL domain is the constant domain of an antibody light chain, for example a kappa (K) or a lambda (9») light chain. An example of a constant region of a kappa light chain is shown in SEQ ID NO:87. An e of a constant region of a lambda light chain is shown in SEQ ID NO:88. In some embodiments, the ﬁrst and the second CL domain are the same, e.g. the ﬁrst and the second CL domain are both a kappa light chain nt domain or the ﬁrst and the second CL domain are both a lambda light chain constant domain. In some embodiments, the ﬁrst and the second CL domain are different, e.g., the ﬁrst CL domain is a constant kappa domain and the second CL domain is a constant lambda domain or vice versa. _41_ WO 34220 In the t of the present invention, a CH1 domain is the ﬁrst constant domain of an antibody heavy chain. An e of a constant CH1 domain is shown in SEQ ID NO:253.

In preferred embodiments of the binding proteins of the ion, the ﬁrst antigen binding unit (e. g., any one of DLL3#1, DLL3#2, DLL3#3, DLL3#4, DLL3#5, DLL3#6, DLL3#7, DLL3#8, DLL3#9, DLL3#10, DLL3#11, DLL3#12, DLL3#13, DLL3#14, DLL3#15, DLL3#16, 7 or DLL3#18 deﬁned by the CDR and/or VH/VL sequences as outlined above) of the binding proteins of the ion comprises from N— to C-terminus: a ﬁrst light chain variable domain, a ﬁrst CL domain, a ﬁrst linker peptide, a ﬁrst VH domain and a ﬁrst CH1 domain, and/or the second binding unit (e.g., any one of CD3#1, CD3#2 or CD3#3 deﬁned by the CDR and/or VH/VL sequences as outlined above) of the binding proteins of the invention comprises from N— to inus: a second light chain variable domain, a second CL domain, a second linker peptide, a second VH domain and a second CH1 domain. In this embodiment, the ﬁrst and/or the second binding unit have the structure of a single chain Fab. For both, the ﬁrst and/or the second antigen binding unit, when forming a single chain Fab, the order can be ed such that from N- to C- terminus the antigen binding unit comprises: VH-CHl -[linker peptide]—VL-CL. In some embodiments of the protein of the ion when the ﬁrst and/or second antigen binding unit comprise a Fab or a single chain Fab, the constant domains can be of the same type (e.g. both CL domains are kappa or lambda light chain nt domains) or of different types (the ﬁrst CL domain is a kappa and the second CL domain is a lambda light chain constant domain or vice versa).

In one aspect, a linker used in a binding protein of the present invention comprises 26 to 42 amino acids, for example 30 to 40 amino acids. In a further aspect, a linker used in a protein of the present ion comprises 34 to 40 amino acids, for example 36 to 39 amino acids, for example 38 amino acids. In some embodiments, the linker comprises a sequence of any one of SEQ ID N05: 89, 90, 91, 92, 93, 94, or 95, preferably SEQ ID NO: The linker sequence may be a naturally occurring sequence or a non—naturally occurring sequence. If used for therapeutic purposes, the linker is preferably non—immunogenic in the subject to which the binding protein of the ion is administered. _42_ One useful group of linker sequences are linkers derived from the hinge region of heavy chain dies as described in WOl996/34103 and WOl994/04678. Other examples are poly-alanine linker sequences such as Ala-Ala-Ala.

Further preferred examples of linker sequences are Gly/Ser linkers of different length such as (glyxsery)z linkers, including e.g. er)3, (gly4ser)5, (gly4ser)7, (gly3ser)3, (gly3ser)5, (gly3ser)7, (gly3ser2)3, (gly3ser2)5, and (gly3ser2)7 or a linker of any one of SEQ ID Nos: 89 to 95 .

In some embodiments of the g proteins of the invention, the VL domain of the first n binding unit (e.g., defined by the light chain CDR (LCCDR) or VL sequences of DLL3#1, DLL3#2, DLL3#3, DLL3#4, DLL3#5, DLL3#6, DLL3#7, DLL3#8, DLL3#9, DLL3#10, DLL3#11, DLL3#12, DLL3#13, DLL3#14, DLL3#15, DLL3#16, DLL3#17 or DLL3#18) is ntly linked (e.g., directly bound) via a ﬁrst Gly/Ser linker (e.g., Gly/Ser linker of any one of 26 to 42 amino acids, 30 to 40 amino acids, 34 to 40 amino acids, or 36 to 39 amino acids, preferably 38 amino acids) to the VH domain of the ﬁrst n g unit (e.g., defined by the heavy chain CDR ) or VH sequences of DLL3#1, DLL3#2, DLL3#3, DLL3#4, DLL3#5, DLL3#6, DLL3#7, DLL3#8, DLL3#9, DLL3#10, DLL3#11, DLL3#12, DLL3#13, 4, DLL3#15, DLL3#16, DLL3#17 or DLL3#18); and the VL domain of the second antigen binding unit (e.g., deﬁned by the light chain CDR (LCCDR) or VL sequences of CD3#1, CD3#2 or CD3#3) is ntly linked (e. g. directly bound) via a second Gly/Ser linker (e.g., Gly/Ser linker of any one of 26 to 42 amino acids, 30 to 40 amino acids, 34 to 40 amino acids, or 36 to 39 amino acids, preferably 38 amino acids) to the VH domain of the second antigen binding unit (e.g., deﬁned by the heavy chain CDR (HCCDR) or VH sequences of CD3#1, CD3#2 or CD3#3). More preferably, the ﬁrst and the second linker are the same. Even more preferably, the ﬁrst and the second linker each comprise the amino acid sequence of SEQ ID NO:89.

In some embodiments, the binding protein of the ion comprises a ﬁrst single chain Fab forming a first antigen binding unit speciﬁc for DLL3 and comprising the sequence selected from the group consisting of SEQ ID NO:49, SEQ ID NO:50, SEQ ID NO:51, SEQ ID NO:52, SEQ ID NO:53, SEQ ID NO:54, SEQ ID NO:229, SEQ ID NO:230, SEQ ID NO:231, SEQ ID NO:232, SEQ ID NO:233, SEQ ID , SEQ ID NO:235, SEQ ID NO:236, SEQ ID NO:237, SEQ ID NO:238, SEQ ID NO:239, and SEQ ID NO:240 and _43_ 2019/064942 a second single chain Fab of SEQ ID NO:71 forming a second antigen binding unit speciﬁc for CD3. In a preferred embodiment, the binding protein of the invention comprises a ﬁrst single chain Fab forming a ﬁrst antigen binding unit comprising the sequence of SEQ ID NO:49 and a second single chain Fab comprising the sequence of SEQ ID NO:71 forming a second antigen binding unit. In a preferred embodiment, the binding protein of the invention comprises a ﬁrst single chain Fab forming a ﬁrst antigen binding unit comprising the sequence of SEQ ID NO:50 and a second single chain Fab comprising the sequence of SEQ ID NO:71 forming a second antigen g unit. In a preferred embodiment, the binding protein of the invention comprises a ﬁrst single chain Fab forming a ﬁrst antigen binding unit comprising the sequence of SEQ ID NO:51 and a second single chain Fab comprising the sequence of SEQ ID NO:71 forming a second antigen g unit.

In some ments, the binding n of the invention comprises a ﬁrst single chain Fab forming a ﬁrst antigen binding unit speciﬁc for DLL3 and comprising the sequence selected from the group consisting of SEQ ID NO:49, SEQ ID NO:50, SEQ ID NO:51, SEQ ID NO:52, SEQ ID NO:53, SEQ ID NO:54 SEQ ID , SEQ ID NO:230, SEQ ID NO:231, SEQ ID NO:232, SEQ ID NO:233, SEQ ID NO:234, SEQ ID NO:235, SEQ ID NO:236, SEQ ID NO:237, SEQ ID NO:238, SEQ ID NO:239, and SEQ ID NO:240 and a second single chain Fab of SEQ ID NO:72 forming a second antigen binding unit speciﬁc for CD3. In a preferred ment, the binding protein of the invention comprises a ﬁrst single chain Fab forming a ﬁrst antigen binding unit comprising the sequence of SEQ ID NO:49 and a second single chain Fab comprising the sequence of SEQ ID NO:72 forming a second antigen binding unit. In a preferred embodiment, the binding n of the invention comprises a ﬁrst single chain Fab forming a ﬁrst antigen binding unit comprising the sequence of SEQ ID NO:50 and a second single chain Fab comprising the sequence of SEQ ID NO:72 forming a second antigen binding unit. In a preferred embodiment, the binding protein of the invention comprises a ﬁrst single chain Fab forming a ﬁrst antigen binding unit comprising the sequence of SEQ ID NO:51 and a second single chain Fab comprising the ce of SEQ ID NO:72 forming a second antigen binding unit.

In some ments, the binding n of the invention ses a ﬁrst single chain Fab forming a ﬁrst antigen binding unit speciﬁc for DLL3 and sing the ce selected from the group consisting of SEQ ID NO:49, SEQ ID NO:50, SEQ ID NO:51, _44_ SEQ ID N0252, SEQ ID N0253, SEQ ID NO:54 SEQ ID N02229, SEQ ID NO:230, SEQ ID NO:231, SEQ ID NO:232, SEQ ID NO:233, SEQ ID NO:234, SEQ ID NO:235, SEQ ID NO:236, SEQ ID NO:237, SEQ ID NO:23 8, SEQ ID NO:239, and SEQ ID NO:240 and a second single chain Fab of SEQ ID NO: 104 forming a second n binding unit speciﬁc for CD3. In a red embodiment, the binding protein of the invention comprises a ﬁrst single chain Fab forming a ﬁrst antigen binding unit comprising the sequence of SEQ ID NO:49 and a second single chain Fab comprising the sequence of SEQ ID NO:104 forming a second antigen binding unit. In a preferred embodiment, the binding n of the invention comprises a ﬁrst single chain Fab forming a ﬁrst antigen g unit comprising the sequence of SEQ ID NO:50 and a second single chain Fab comprising the sequence of SEQ ID NO:104 forming a second antigen binding unit. In a preferred embodiment, the binding protein of the invention comprises a ﬁrst single chain Fab g a ﬁrst n binding unit comprising the sequence of SEQ ID NO:51 and a second single chain Fab comprising the sequence of SEQ ID NO:104 forming a second antigen binding unit.

In some embodiments, the ﬁrst antigen binding unit (e.g., DLL3#1, DLL3#2, DLL3#3, DLL3#4, DLL3#5, DLL3#6, DLL3#7, DLL3#8, DLL3#9, DLL3#10, DLL3#11, DLL3#12, DLL3#13, DLL3#14, DLL3#15, DLL3#16, DLL3#17 or DLL3#18 as deﬁned by the CDR and/or VH/VL sequences outlined above) and/or the second antigen binding unit (e.g. CD3#1, CD3#2 or CD3#3) comprises a VL domain covalently linked (e. g. directly bound) to a CL domain and a VH domain linked to a CH1 domain (i.e. er forming a Fab fragment derived from an antibody), and said CH1 domain is further covalently linked (e. g. directly bound) to an EC domain thereby forming an arm of a conventional Y shaped antibody molecule with one light and one heavy chain. In some ments, the ﬁrst and the second antigen binding unit each form a Fab nt, i.e. a ﬁrst and a second Fab fragment, which is covalently linked (e.g. directly bound) to a ﬁrst and a second Fc domain, respectively, thereby forming a conventional heterotetrameric iﬁc and bivalent antibody molecule.

In preferred embodiments, the ﬁrst antigen binding unit (e.g., any one of DLL3#1, DLL3#2, DLL3#3, DLL3#4, DLL3#5, DLL3#6, DLL3#7, DLL3#8, DLL3#9, 0, DLL3#l l, DLL3#12, 3, DLL3#l4, DLL3#15, DLL3#l6, 7 or DLL3#18 as deﬁned by the CDR and/or VH/VL sequences ed above) and/or the second antigen binding unit (e. g. any one of CD3#1, CD3#2 or CD3#3) comprise a single chain Fab, i.e. _45_ an antibody light chain (VL—CL) covalently linked to the VH—CHl domain of a heavy chain via a peptide linker (e. g., Gly/Ser linker of any one of 26 to 42 amino acids, 30 to 40 amino acids, 34 to 40 amino acids, or 36 to 39 amino acids, preferably 38 amino acids, even more preferably a linker of SEQ ID NO:89). In preferred embodiments, the binding protein of the invention comprises a ﬁrst polypeptide chain comprising a ﬁrst single chain Fab speciﬁcally binding to DLL3 (e. g., any one of DLL3#l, DLL3#2, DLL3#3, DLL3#4, DLL3#5, DLL3#6, , DLL3#8, , DLL3#10, DLL3#11, DLL3#12, DLL3#13, 4, DLL3#15, DLL3#16, DLL3#17 or DLL3#18 as deﬁned by the respective CDR or VH/VL sequences above, preferably the ﬁrst antigen binding unit is DLL3#1, DLL3#2, DLL3#3) and a ﬁrst PC domain (this polypeptide chain herein referred to also as “DLL3 chain”) and a second polypeptide chain comprising a second single chain Fab speciﬁcally binding to CD3 (e.g. any one of CD3#1 CD3#2 or CD3#3 as deﬁned by the respective CDR or VH/VL sequences above) and a second Fc domain (this polypeptide chain herein referred to herein also as “CD3 chain”). In some embodiments, the ﬁrst and the second Fc domain are the same. In red ments, the ﬁrst and the second Fc domains are different. The resulting g proteins of the invention bear a full PC and have two independent binding sites, a ﬁrst binding unit for DLL3 and a second binding unit for CD3. In some embodiments, the ﬁrst n binding unit consists of a single polypeptide chain (a DLL3 chain). In some embodiments, the second antigen binding unit ts of a single ptide chain (a CD3 chain). In some embodiments, both the ﬁrst and the second antigen binding unit consist of a single polypeptide chain, respectively.

In some embodiments, the ﬁrst antigen binding unit of the n of the invention is formed by a ﬁrst polypeptide chain (a DLL3 chain) and the second antigen binding unit is formed by a second polypeptide chain (a CD3 chain). Thus, in some embodiments, the binding n of the ion comprises two different polypeptide chains, each comprising an antigen binding unit with different speciﬁcity, the ptide chains covalently linked to each other, either via disulﬁde bonds or potentially via a peptide linker. In some embodiments, the binding protein of the invention is a bispeciﬁc, bivalent heterodimeric n comprising two ptide chains, one polypeptide chain (a ﬁrst polypeptide chain or DLL3 chain) comprising an antigen binding unit speciﬁcally binding to DLL3 (e. g. any one of DLL3#1, DLL3#2, DLL3#3, DLL3#4, DLL3#5, DLL3#6, DLL3#7, DLL3#8, DLL3#9, DLL3#10, DLL3#1 l, DLL3#12, DLL3#13, DLL3#14, DLL3#15, DLL3#16, DLL3#17 or DLL3#18 deﬁned by the respective CDR and/or —46- VH/VL ces) and another polypeptide chain (a second polypeptide chain or CD3 chain) comprising an antigen binding unit cally binding to CD3 (e. g. any one of CD3#1, CD3#2 or CD3#3).

In the context of the present invention, an Fc domain is for example derived from the heavy chain of an IgG, for e an IgG1,IgG2 or IgG4_ For example, an Fc domain of the present invention is a Fc domain of a heavy chain of an IgG1 or IgG4 and comprises a hinge region and two constant domains (CH2 and CH3). Examples of Fc domains (including a hinge region) are shown in SEQ ID NOsz81 and 84.

The numbering of the amino acids in the amino acid chains of a protein of the present invention is herein according to the EU numbering system (Edelman, Cunningham et a1. 1969), unless ise speciﬁed. This means that the amino acid numbers indicated herein correspond to the ons in a heavy chain of the ponding pe (e.g.

IgG1 or IgG4), according to the EU numbering system, unless otherwise speciﬁed.

In some embodiments, the ﬁrst Fc domain and the second Fc domain in a protein of the present invention each comprises one or more amino acid changes which reduce the formation of homodimers of the ﬁrst or second polypeptide chains instead of heterodimers of a ﬁrst and a second polypeptide chain. Through these changes, a "protrusion" is generated in one of the Fc domains by replacing one or more, small amino acid side chains from the ace of one of the heavy chains with larger side chains (e. g. tyrosine or tryptophan). Compensatory "cavities" of identical or similar size are created on the interface of the other PC domain by replacing large amino acid side chains with smaller ones (e. g. alanine or ine). This provides a mechanism for increasing the yield of the heterodimer over other unwanted end-products such as mers, in particular homodimers of the Fc domain with the “protrusion” (see for example Ridgway et a1.

Protein Eng, 1996. 9(7): p. 617—21; Atwell et a1, JMB, 1997, 270, 26-35). In some ments, such amino acid changes are a tyrosine (Y) at position 366 [T366Y] of the ﬁrst PC domain and a threonine (T) at position 407 [Y407T] of the second Fc . In some embodiments, the ﬁrst PC domain comprises a serine (S) at position 366 [T3668] and the second Fc domain comprises a tryptophan (W) at position 366 [T366W], an alanine (A) at position 368 [L368A] and a valine (V) at position 407 [Y407V]. In preferred embodiments, the ﬁrst PC domain comprises a tryptophan (W) at position 366 [T3 66W] and the second Fc domain comprises a serine (S) at position 366 [T3668], an alanine (A) at _47_ position 368 ] and a valine (V) at position 407 [Y407V]. For example, position 366 of the Fc domain ing to EU numbering, corresponding to the amino acid position 146 in the human IgG1 Fc sequence of SEQ ID NO:81, is changed from T at position 146 in SEQ ID NO:81 to W at position 146 in SEQ ID NO:82; and ons 366, 368 and 407 according to EU numbering, corresponding to the amino acid positions 146, 148 and 187, respectively, in SEQ ID NO:81, are changed from T, L and Y at these positions in SEQ ID NO:81 to S, A and V at these positions in SEQ ID NO:83. In any of these embodiments, the amino acid changes described for the ﬁrst PC domain may be located in the second Fc domain and the respective amino acid changes for the second Fc domain may be located in the ﬁrst PC domain. In other words, the term “ﬁrst” and “second” can be exchanged in these embodiments. In some embodiments, such a PC domain is an Fc domain derived from the heavy chain of an IgG1 or IgG4.

In some embodiments, the ﬁrst PC domain comprises a cysteine (C) at position 354 [S354C] in addition to the tryptOphan (W) at position 366 [T3 66W] and the second Fc domain comprises a ne (C) at position 349 [Y349C] in addition to the serine (S) at position 366 [T366S], the alanine (A) at position 368 [L368A] and the valine (V) at position 407 [Y407V]. In one aspect, such Fc domain is an EC domain derived from the heavy chain of an IgG4.

In some ments, the ﬁrst PC domain or the second Fc domain in a binding protein of the present invention r comprises one or more amino acid changes which reduce the binding of the Fc domain to protein A. In some embodiments, such amino acid changes are an arginine at position 435 [H43 SR] and a phenylalanine at position 436 [Y43 6F] of one of the Fc domains. Both changes are derived from the sequence of human IgG3 (IgG3 does not bind to protein A). These two mutations are located in the CH3 domain and are orated in one of the Fc domains to reduce binding to Protein A (see for example erg et al. J l Methods, 1997. : p. 25—34). These two changes facilitate the removal of homodimers of heavy chains comprising these changes during n puriﬁcation.

In some embodiments, in a binding protein of the present invention, the Fc domain, which comprises a threonine (T) at position 407 [Y407T], further comprises an arginine at position 435 [H435R] and a phenylalanine at position 436 [Y436F]. In this case, the other heavy chain comprises a tyrosine (Y) at on 366 [T366Y], but does not include the —48- two changes at positions 435 and 436. Alternatively, in some ments, in a protein of the present invention, the Fc domain, which comprises a serine (S) at position 366 [T3 66S], an alanine (A) at position 368 [L368A] and a valine (V) at position 407 [Y407V], further comprises an arginine at position 435 [H435R] and a phenylalanine at position 436 [Y43 6F]. In this case, the other PC domain comprises a phan (W) at position 366 [T3 66W], but does not include the two changes at ons 435 and 436. Thus, the Fc domain comprising the amino acid change resulting in a y” as described above also comprises the amino acid changes, which reduce binding to Protein A. Homodimers comprising this PC domain are removed through reduced binding to Protein A. The production of mers of the other PC domain, which comprises the “protrusion”, is reduced by the presence of the “protrusion”.

In some embodiments, the Fc domain of a protein of the present ion may or may not further comprises YTE mutations (M252Y/S254T/T256E, EU numbering (Dall'Acqua, Kiener et al. 2006)). These mutations have been shown to improve the pharmacokinetic properties of Fc s through preferential enhancement of binding afﬁnity for neonatal FcRn receptor at pH 6.0.

In some embodiments, the ﬁrst and/or the second Fc domain of the present invention derived from an IgGl also includes the “KO” mutations (L234A, L235A). In a further aspect, the ﬁrst and/or the second Fc domain of the present invention derived from an IgG4 also includes the Pro hinge mutation (S228P).

In preferred embodiments of the invention, the binding n comprises i) a ﬁrst polypeptide chain comprising the amino acid sequence of SEQ ID NO:73 and a second polypeptide chain comprising the amino acid sequence of SEQ ID NO:79 (DLL3#l/CD3#1), or ii) a ﬁrst polypeptide chain comprising the amino acid sequence of SEQ ID NO:74 and a second polypeptide chain comprising the amino acid sequence of SEQ ID NO:79 (DLL3#2/CD3#1), or iii) a ﬁrst polypeptide chain comprising the amino acid ce of SEQ ID NO:75 and a second polypeptide chain comprising the amino acid sequence of SEQ ID NO:79 (DLL3#3/CD3#1), or iv) a ﬁrst polypeptide chain comprising the amino acid ce of SEQ ID NO:76 and a second polypeptide chain comprising the amino acid sequence of SEQ ID NO:79 (DLL3#4/CD3#1), or v) a ﬁrst polypeptide chain comprising the amino acid ce of SEQ ID NO:77 and a second polypeptide chain comprising the amino acid sequence of SEQ ID NO:79 _49_ (DLL3#5/CD3#1), or vi) a ﬁrst polypeptide chain comprising the amino acid sequence of SEQ ID NO:78 and a second polypeptide chain comprising the amino acid sequence of SEQ ID NO:79 (DLL3#6/CD3#1); or Vii) a ﬁrst ptide chain comprising the amino acid sequence of SEQ ID NO:241 and a second polypeptide chain sing the amino acid sequence of SEQ ID NO:79 (DLL3#7/CD3#1); or viii) a ﬁrst polypeptide chain comprising the amino acid sequence of SEQ ID NO:242 and a second polypeptide chain comprising the amino acid sequence of SEQ ID NO:79 (DLL3#8/CD3#1); or ix) a ﬁrst polypeptide chain comprising the amino acid sequence of SEQ ID NO:243 and a second polypeptide chain comprising the amino acid sequence of SEQ ID NO:79 w (DLL3#9/CD3#1); or x) a ﬁrst polypeptide chain sing the amino acid sequence of SEQ ID NO:244 and a second polypeptide chain comprising the amino acid sequence of SEQ ID NO:79 (DLL3#10/CD3#1); or xi) a ﬁrst polypeptide chain comprising the amino acid sequence of SEQ ID NO:245 and a second polypeptide chain comprising the amino acid sequence of SEQ ID NO:79 11/CD3#1); or xii) a ﬁrst polypeptide chain H sing the amino acid sequence of SEQ ID NO:246 and a second polypeptide chain sing the amino acid sequence of SEQ ID NO:79 (DLL3#12/CD3#1); or xiii) a ﬁrst polypeptide chain comprising the amino acid sequence of SEQ ID NO:247 and a second polypeptide chain comprising the amino acid sequence of SEQ ID NO:79 (DLL3#l3/CD3#1); or xiv) a ﬁrst polypeptide chain comprising the amino acid sequence of SEQ ID NO:248 and a second polypeptide chain comprising the amino acid sequence of SEQ ID NO:79 (DLL3#l4/CD3#1); or xv) a ﬁrst polypeptide chain comprising the amino acid sequence of SEQ ID NO:249 and a second polypeptide chain comprising the amino acid sequence of SEQ ID NO:79 (DLL3#15/CD3#1), or xvi) a ﬁrst polypeptide chain comprising the amino acid sequence of SEQ ID NO:250 and a second polypeptide chain comprising the amino acid sequence of SEQ ID NO:79 (DLL3#l6/CD3#1); or xvii) a ﬁrst polypeptide chain comprising the amino acid sequence of SEQ ID N025] and a second polypeptide chain comprising the amino acid ce of SEQ ID NO:79 l7/CD3#1); or xviii) a ﬁrst polypeptide chain comprising the amino acid sequence of SEQ ID NO:252 and a second polypeptide chain comprising the amino acid sequence of SEQ ID NO:79 18/CD3#1). Preferably, the first and second polypeptide chain are linked via one or more disulﬁde bonds and form an antibody like structure (Figure 1) r to a conventional Y—shaped antibody molecule. _50_ In another preferred embodiment, the binding protein comprises a ﬁrst polypeptide chain speciﬁc for DLL3 comprising an amino acid sequence ed from the group ting of SEQ ID NO:73, SEQ ID NO:74, and SEQ ID NO:75, and a second polypeptide chain speciﬁc for CD3 comprising the amino acid sequence of SEQ ID NO:79.

In one preferred embodiment, the binding protein comprises a ﬁrst polypeptide chain speciﬁc for DLL3 comprising an amino acid sequence of SEQ ID NO:73 and a second polypeptide chain speciﬁc for CD3 comprising the amino acid sequence of SEQ ID NO:79.

In one preferred embodiment, the binding protein comprises a ﬁrst polypeptide chain speciﬁc for DLL3 sing an amino acid sequence of SEQ ID NO:74 and a second polypeptide chain speciﬁc for CD3 comprising the amino acid sequence of SEQ ID NO:79.

In one preferred ment, the binding protein comprises a ﬁrst polypeptide chain speciﬁc for DLL3 comprising an amino acid sequence of SEQ ID NO:75 and a second polypeptide chain speciﬁc for CD3 comprising the amino acid sequence of SEQ ID NO:79.

In preferred embodiments of the invention, the binding protein comprises i) a ﬁrst ptide chain comprising the amino acid sequence of SEQ ID NO:73 and a second polypeptide chain comprising the amino acid sequence of SEQ ID NO:80 (DLL3#1/CD3#2), or ii) a ﬁrst polypeptide chain comprising the amino acid sequence of SEQ ID NO:74 and a second polypeptide chain comprising the amino acid sequence of SEQ ID NO:80 (DLL3#2/CD3#2), or iii) a ﬁrst polypeptide chain comprising the amino acid ce of SEQ ID NO:75 and a second polypeptide chain comprising the amino acid ce of SEQ ID NO:80 (DLL3#3/CD3#2), or iv) a ﬁrst polypeptide chain comprising the amino acid sequence of SEQ ID NO:76 and a second polypeptide chain comprising the amino acid sequence of SEQ ID NO:80 (DLL3#4/CD3#2), or V) a ﬁrst polypeptide chain comprising the amino acid sequence of SEQ ID NO:77 and a second ptide chain comprising the amino acid sequence of SEQ ID NO:80 /CD3#2), or vi) a ﬁrst polypeptide chain comprising the amino acid sequence of SEQ ID NO:78 and a second polypeptide chain comprising the amino acid sequence of SEQ ID NO:80 6/CD3#2); or vii) a ﬁrst ptide chain comprising the amino acid sequence of SEQ ID NO:241 and a second polypeptide chain comprising the amino acid sequence of SEQ ID NO:80 (DLL3#7/CD3#2); or viii) a ﬁrst polypeptide chain comprising the amino acid sequence of SEQ ID NO:242 and a second polypeptide chain comprising the amino acid sequence of SEQ ID NO:80 (DLL3#8/CD3#2); or ix) a ﬁrst _51_ polypeptide chain comprising the amino acid ce of SEQ ID NO:243 and a second polypeptide chain comprising the amino acid sequence of SEQ ID NO:80 (DLL3#9/CD3#2); or x) a ﬁrst polypeptide chain comprising the amino acid sequence of SEQ ID NO:244 and a second polypeptide chain sing the amino acid sequence of SEQ ID NO:80 (DLL3#10/CD3#2); or xi) a ﬁrst polypeptide chain comprising the amino acid sequence of SEQ ID NO:245 and a second ptide chain comprising the amino acid sequence of SEQ ID NO:80 (DLL3#11/CD3#2); or xii) a ﬁrst ptide chain sing the amino acid sequence of SEQ ID NO:246 and a second polypeptide chain comprising the amino acid ce of SEQ ID NO:80 (DLL3#l2/CD3#2); or xiii) a ﬁrst polypeptide chain comprising the amino acid sequence of SEQ ID NO:247 and a second polypeptide chain comprising the amino acid sequence of SEQ ID NO:80 (DLL3#l3/CD3#2); or xiv) a ﬁrst polypeptide chain sing the amino acid ce of SEQ ID NO:248 and a second ptide chain comprising the amino acid sequence of SEQ ID NO:80 (DLL3#14/CD3#2); or xv) a ﬁrst polypeptide chain comprising the amino acid sequence of SEQ ID NO:249 and a second polypeptide chain comprising the amino acid sequence of SEQ ID NO:80 (DLL3#15/CD3#2), or xvi) a ﬁrst polypeptide chain comprising the amino acid sequence of SEQ ID NO:250 and a second polypeptide chain comprising the amino acid sequence of SEQ ID NO:80 (DLL3#16/CD3#2); or xvii) a ﬁrst polypeptide chain comprising the amino acid sequence of SEQ ID NO:251 and a second polypeptide chain comprising the amino acid sequence of SEQ ID NO:80 (DLL3#l7/CD3#2); or xviii) a ﬁrst polypeptide chain comprising the amino acid sequence of SEQ ID NO:252 and a second polypeptide chain comprising the amino acid sequence of SEQ ID NO:80 (DLL3#l8/CD3#2). Preferably, the first and second polypeptide chain are linked via one or more disulﬁde bonds and form an antibody like structure (Figure 1) similar to a conventional Y—shaped antibody le.

In another preferred embodiment, the binding protein comprises a ﬁrst polypeptide chain speciﬁc for DLL3 comprising an amino acid sequence selected from the group consisting of SEQ ID NO:73, SEQ ID NO:74, and SEQ ID NO:75, and a second polypeptide chain Speciﬁc for CD3 comprising the amino acid sequence of SEQ ID NO:80.

In one preferred embodiment, the binding protein comprises a ﬁrst ptide chain speciﬁc for DLL3 comprising an amino acid sequence of SEQ ID NO:73 and a second polypeptide chain speciﬁc for CD3 comprising the amino acid sequence of SEQ ID NO:80. _52_ In one red embodiment, the binding protein comprises a ﬁrst polypeptide chain speciﬁc for DLL3 comprising an amino acid sequence of SEQ ID NO:74 and a second polypeptide chain Speciﬁc for CD3 comprising the amino acid sequence of SEQ ID NO:80.

In one preferred embodiment, the binding protein comprises a ﬁrst ptide chain speciﬁc for DLL3 comprising an amino acid sequence of SEQ ID NO:75 and a second polypeptide chain speciﬁc for CD3 comprising the amino acid sequence of SEQ ID NO:80.

In preferred embodiments of the invention, the binding protein comprises i) a ﬁrst polypeptide chain comprising the amino acid sequence of SEQ ID NO:73 and a second polypeptide chain comprising the amino acid sequence of SEQ ID NO:105 (DLL3#1/CD3#3), or ii) a ﬁrst polypeptide chain comprising the amino acid sequence of SEQ ID NO:74 and a second polypeptide chain comprising the amino acid sequence of SEQ ID NO:105 (DLL3#2/CD3#3), or iii) a ﬁrst polypeptide chain comprising the amino acid sequence of SEQ ID NO:75 and a second polypeptide chain comprising the amino acid sequence of SEQ ID NO:105 (DLL3#3/CD3#3), or iv) a ﬁrst polypeptide chain comprising the amino acid sequence of SEQ ID NO:76 and a second polypeptide chain comprising the amino acid sequence of SEQ ID NO:105 4/CD3#3), or V) a ﬁrst polypeptide chain comprising the amino acid ce of SEQ ID NO:77 and a second polypeptide chain comprising the amino acid sequence of SEQ ID NO:105 (DLL3#5/CD3#3), or Vi) a ﬁrst polypeptide chain comprising the amino acid sequence of SEQ ID NO:78 and a second polypeptide chain comprising the amino acid sequence of SEQ ID NO:105 (DLL3#6/CD3#3); or Vii) a ﬁrst polypeptide chain sing the amino acid sequence of SEQ ID NO:241 and a second polypeptide chain comprising the amino acid sequence of SEQ ID NO:105 (DLL3#7/CD3#3); or Viii) a ﬁrst polypeptide chain comprising the amino acid ce of SEQ ID NO:242 and a second ptide chain comprising the amino acid sequence of SEQ ID NO:105 (DLL3#8/CD3#3); or ix) a ﬁrst polypeptide chain comprising the amino acid sequence of SEQ ID NO:243 and a second polypeptide chain comprising the amino acid sequence of SEQ ID NO:105 (DLL3#9/CD3#3); or x) a ﬁrst polypeptide chain comprising the amino acid sequence of SEQ ID NO:244 and a second polypeptide chain sing the amino acid sequence of SEQ ID NO: 105 (DLL3#10/CD3#3); or xi) a ﬁrst polypeptide chain comprising the amino acid sequence of SEQ ID NO:245 and a second polypeptide chain sing the amino acid sequence of SEQ ID NO:105 11/CD3#3); or xii) a ﬁrst ptide chain comprising the amino acid sequence of SEQ ID NO:246 and a second polypeptide chain _53_ comprising the amino acid sequence of SEQ ID NO:105 (DLL3#12/CD3#3); or xiii) a ﬁrst polypeptide chain comprising the amino acid sequence of SEQ ID NO:247 and a second polypeptide chain comprising the amino acid sequence of SEQ ID NO:105 (DLL3#l3/CD3#3); or xiv) a ﬁrst polypeptide chain sing the amino acid sequence of SEQ ID NO:248 and a second polypeptide chain comprising the amino acid sequence of SEQ ID NO:105 (DLL3#l4/CD3#3); or xv) a ﬁrst polypeptide chain comprising the amino acid sequence of SEQ ID NO:249 and a second polypeptide chain comprising the amino acid sequence of SEQ ID NO:105 15/CD3#3), or xvi) a ﬁrst polypeptide chain comprising the amino acid ce of SEQ ID NO:250 and a second polypeptide chain comprising the amino acid ce of SEQ ID NO:105 l6/CD3#3); or xvii) a ﬁrst polypeptide chain comprising the amino acid sequence of SEQ ID NO:251 and a second polypeptide chain comprising the amino acid sequence of SEQ ID NO:105 (DLL3#l7/CD3#3); or xviii) a ﬁrst ptide chain comprising the amino acid sequence of SEQ ID NO:252 and a second polypeptide chain comprising the amino acid sequence of SEQ ID NO:105 (DLL3#l8/CD3#3). Preferably, the ﬁrst and second polypeptide chain are linked via one or more disulﬁde bonds and form an antibody like structure (Figure 1) similar to a conventional Y-shaped antibody molecule.

In another preferred embodiment, the binding protein comprises a ﬁrst polypeptide chain c for DLL3 comprising an amino acid sequence selected from the group consisting of SEQ ID NO:73, SEQ ID NO:74, and SEQ ID NO:75, and a second polypeptide chain speciﬁc for CD3 comprising the amino acid sequence of SEQ ID NO:105.

In one preferred embodiment, the binding protein comprises a ﬁrst polypeptide chain speciﬁc for DLL3 comprising an amino acid sequence of SEQ ID NO:73 and a second polypeptide chain speciﬁc for CD3 comprising the amino acid sequence of SEQ ID NO:105.

In one red embodiment, the binding protein comprises a ﬁrst polypeptide chain speciﬁc for DLL3 sing an amino acid sequence of SEQ ID NO:74 and a second polypeptide chain speciﬁc for CD3 comprising the amino acid sequence of SEQ ID NO:105.

In one preferred ment, the binding protein ses a ﬁrst polypeptide chain speciﬁc for DLL3 comprising an amino acid sequence of SEQ ID NO:75 and a second _54_ polypeptide chain speciﬁc for CD3 comprising the amino acid sequence of SEQ ID NO:105.

For all of the above ments it shall be understood that, by using the term "comprising", it is intended to also include an ment in which the respective domain or molecule "consists of’ the amino acid sequence as indicated.

In a further aspect, the present invention provides a protein comprising a ﬁrst polypeptide chain speciﬁcally binding to DLL3 and a second ptide chain speciﬁcally binding to CD3, wherein the ﬁrst chain comprises a ﬁrst light chain covalently linked (e. g. ly bound) to a ﬁrst linker, which is itself covalently linked (e. g. directly bound) to a ﬁrst heavy chain, and wherein the second chain Speciﬁcally binding to CD3 ses a second light chain covalently linked (e.g. directly bound) to a second linker, which is itself covalently linked (e.g. directly bound) to a second heavy chain.

In some embodiments, starting from its N—terminus, the ﬁrst polypeptide chain ses a ﬁrst light chain variable region cally g to DLL3, a ﬁrst light chain constant region, a ﬁrst linker, a ﬁrst heavy chain variable region speciﬁc for DLL3 and a ﬁrst heavy chain constant region. In some embodiments, ng from its N—terminus, the second polypeptide chain comprises a second light chain variable region speciﬁcally binding to CD3, a second light chain constant region, a second linker, a second heavy chain variable region Speciﬁc for CD3 and a second heavy chain constant region.

The resulting proteins bear a full PC, which is marginally larger than an IgG and has two independent binding sites (e.g., each binding site being monovalent for the respective antigen), a ﬁrst binding site for DLL3 and a second binding site for CD3. Preferably, the ﬁrst and second polypeptide chain are linked via one or more disulﬁde bonds. As such, the ns of the invention are antibody-like structures, having the Y shaped structure of a conventional full length antibody (see Figure 1). This bispeciﬁc format greatly reduces heterogeneity after expression and puriﬁcation (e.g. by avoiding mispairing of light and heavy variable domains with different g speciﬁcities), while maintaining the functional properties of the binding moieties within a structure less likely to generate unwanted immunogenic reactions. This also enables good expression of heterodimeric proteins, e.g. in mammalian cells.

In some ments of the protein of the invention, the ﬁrst polypeptide chain speciﬁcally binding to DLL3 comprises a ﬁrst light chain variable domain and a ﬁrst heavy _55_ chain variable domain, which comprise CDR sequences selected from the group consisting ofi) to xviii): i) light chain CDRs comprising the amino acid sequences of SEQ ID NO:1 (CDRI), SEQ ID NO:2 (CDR2) and SEQ ID NO:3 (CDR3) and a heavy chain CDRs com- prising the amino acid sequences of SEQ ID NO:4 (CDRl), SEQ ID NO:5 (CDR2) and SEQ ID NO:6 ; ii) light chain CDRs comprising the amino acid sequences of SEQ ID NO:7 (CDRl), SEQ ID N028 (CDR2) and SEQ ID NO:9 (CDR3) and heavy chain CDRs comprising the amino acid sequences of SEQ ID NO: 10 (CDRl), SEQ ID NO: 11 (CDR2) and SEQ ID NO:12 (CDR3); iii) light chain CDRs comprising the amino acid sequences of SEQ ID NO:l3 (CDRI), SEQ ID NO: 14 (CDR2) and SEQ ID NO:15 (CDR3) and heavy chain CDRs comprising the amino acid sequences of SEQ ID NO: 16 (CDRl), SEQ ID NO:l7 (CDR2) and SEQ ID NO:18 (CDR3); iv) light chain CDRs comprising the amino acid sequences of SEQ ID NO:l9 (CDRl), SEQ ID NO:20 (CDR2) and SEQ ID NO:21 (CDR3) and heavy chain CDRs comprising the amino acid ces of SEQ ID NO:22 (CDRl), SEQ ID NO:23 (CDR2) and SEQ ID NO:24 (CDR3); v) light chain CDRs comprising the amino acid sequences of SEQ ID NO:25 (CDRl), SEQ ID NO:26 (CDR2) and SEQ ID NO:27 (CDR3) and heavy chain CDRs comprising the amino acid sequences of SEQ ID NO:28 (CDRl), SEQ ID NO:29 (CDR2) and SEQ ID NO:30 (CDR3); and vi) light chain CDRs comprising the amino acid sequences of SEQ ID NO:31 (CDRI), SEQ ID NO:32 (CDR2) and SEQ ID NO:33 (CDR3) and heavy chain CDRs ising the amino acid ces of SEQ ID NO:34 (CDRl), SEQ ID NO:35 (CDR2) and SEQ ID NO:36 (CDR3); vii) light chain CDRs comprising the amino acid sequences of SEQ ID NO: 133 (CDRl), SEQ ID NO:134 (CDR2) and SEQ ID NO:135 (CDR3) and heavy chain CDRs comprising the amino acid sequences of SEQ ID NO:l36 (CDRl), SEQ ID NO:137 (CDR2) and SEQ ID NO: 138 (CDR3); —56— viii) light chain CDRs comprising the amino acid sequences of SEQ ID NO: 139 (CDRl), SEQ ID NO: 140 (CDR2) and SEQ ID NO: 141 (CDR3) and heavy chain CDRs comprising the amino acid sequences of SEQ ID NO:142 (CDRl), SEQ ID NO:143 (CDR2) and SEQ ID NO: 144 (CDR3); ix) light chain CDRs sing the amino acid sequences of SEQ ID NO: 145 (CDRl), SEQ ID NO: 146 (CDR2) and SEQ ID NO: 147 (CDR3) and heavy chain CDRs comprising the amino acid sequences of SEQ ID NO: 148 (CDRl), SEQ ID NO:149 (CDR2) and SEQ ID NO:150 (CDR3); light chain CDRs comprising the amino acid sequences of SEQ ID NO: 151 , SEQ ID NO: 152 (CDR2) and SEQ ID NO: 153 (CDR3) and heavy chain CDRs comprising the amino acid sequences of SEQ ID NO: 154 (CDRl), SEQ ID NO:155 (CDR2) and SEQ ID NO: 156 (CDR3); Xi) light chain CDRs comprising the amino acid sequences of SEQ ID NO: 157 (CDRl), SEQ ID NO: 158 (CDR2) and SEQ ID NO: 159 (CDR3) and heavy chain CDRs comprising the amino acid sequences of SEQ ID NO: 160 (CDRl), SEQ ID NO:161 (CDR2) and SEQ ID NO: 162 (CDR3); xii) light chain CDRs comprising the amino acid sequences of SEQ ID NO: 163 (CDRl), SEQ ID NO: 164 (CDR2) and SEQ ID NO: 165 (CDR3) and heavy chain CDRs comprising the amino acid ces of SEQ ID NO:166 (CDRl), SEQ ID NO:167 (CDR2) and SEQ ID NO:168 (CDR3); xiii) light chain CDRs comprising the amino acid sequences of SEQ ID NO: 169 (CDRl), SEQ ID NO:170 (CDR2) and SEQ ID NO: 171 (CDR3) and heavy chain CDRs comprising the amino acid sequences of SEQ ID NO: 172 , SEQ ID NO:173 (CDR2) and SEQ ID NO: 174 (CDR3); xiv) comprising light chain CDRs comprising the amino acid sequences of SEQ ID NO:175 (CDRl), SEQ ID NO: 176 (CDR2) and SEQ ID NO: 177 (CDR3) and heavy chain CDRs comprising the amino acid sequences of SEQ ID NO: 178 (CDRl), SEQ ID NO: 179 (CDR2) and SEQ ID NO: 180 (CDR3); xv) light chain CDRs comprising the amino acid sequences of SEQ ID NO: 181 (CDRl), SEQ ID NO:182 (CDR2) and SEQ ID NO: 183 (CDR3) and heavy chain _57_ WO 34220 CDRs comprising the amino acid sequences of SEQ ID NO:184 (CDRl), SEQ ID NO:185 (CDR2) and SEQ ID NO:186 (CDR3); xvi) light chain CDRs comprising the amino acid ces of SEQ ID NO: 187 (CDRl), SEQ ID NO:188 (CDR2) and SEQ ID NO: 189 (CDR3) and heavy chain CDRs comprising the amino acid sequences of SEQ ID NO: 190 (CDRl), SEQ ID NO:191 (CDR2) and SEQ ID NO:192 (CDR3); xvii) light chain CDRs comprising the amino acid sequences of SEQ ID NO: 193 (CDRl), SEQ ID NO:194(CDR2) and SEQ ID NO:195 (CDR3) and heavy chain CDRs comprising the amino acid sequences of SEQ ID NO: 196 (CDRl), SEQ ID NO:197 (CDR2) and SEQ ID NO: 198 (CDR3); and xviii) light chain CDRs comprising the amino acid sequences of SEQ ID NO: 199 (CDRl), SEQ ID NO:200 (CDR2) and SEQ ID NO:201 (CDR3) and heavy chain CDRs comprising the amino acid sequences of SEQ ID NO:202 (CDRl), SEQ ID NO:203 (CDR2) and SEQ ID NO:204 (CDR3).

The respective light/heavy chain variable domains deﬁned by these CDR sequences are termed DLL3#1, DLL3#2, DLL3#3, DLL3#4, DLL3#5, DLL3#6, , DLL3#8, DLL3#9, 0, DLL3#11, DLL3#12, 3, DLL3#14, DLL3#15, DLL3#16, DLL3#17, and DLL3#18, respectively. Preferably, the CDR sequences are selected from the group consisting of i) to iii) (DLL3#1, DLL3#2, DLL3#3) as deﬁned above.

In preferred embodiments of the binding protein of the invention, said second polypeptide chain speciﬁcally binding to CD3 comprises a second light chain variable domain and second heavy chain variable domain, which ses CDR sequences selected from the group ting of: i) light chain CDRs comprising the amino acid sequences of SEQ ID NO:55 (CDRl), SEQ ID NO:56 (CDR2) and SEQ ID NO:57 (CDR3) and heavy chain CDRs com-prising the amino acid sequences of SEQ ID NO:58 (CDRl), SEQ ID NO:59 (CDR2) and SEQ ID NO:60 (CDR3); ii) light chain CDRs comprising the amino acid sequences of SEQ ID NO:61 (CDRl), SEQ ID NO:62 (CDR2) and SEQ ID NO:63 (CDR3) and heavy chain CDRs ising the amino acid sequences of SEQ ID NO:64 (CDRl), SEQ ID NO:65 (CDR2) and SEQ ID NO:66 (CDR3); and —58— WO 34220 iii) light chain CDRs comprising the amino acid sequences of SEQ ID NO:96 (CDRl), SEQ ID NO:97 (CDR2) and SEQ ID NO:98 (CDR3) and heavy chain CDRS ising the amino acid sequences of SEQ ID NO:99 (CDRl), SEQ ID N0:100 (CDR2) and SEQ ID N0:101 (CDR3).

The respective light/heavy chain variable domains deﬁned by these CDR sequences are termed CD3#l, CD3#2 and CD3#3, respectively.

In one preferred embodiment, the binding protein of the invention comprises (i) a ﬁrst polypeptide chain speciﬁcally binding to DLL3, comprising a ﬁrst light chain variable domain with light chain CDRs comprising the amino acid sequences of SEQ ID N021 (CDRl), SEQ ID N012 (CDR2) and SEQ ID N0:3 (CDR3) and a ﬁrst heavy chain variable domain with heavy chain CDRs comprising the amino acid sequences of SEQ ID N024 (CDRl), SEQ ID N025 (CDR2) and SEQ ID NO:6 (CDR3); and (ii) a second ptide chain speciﬁcally binding to CD3, comprising a second light chain variable domain with light chain CDRs comprising the amino acid ces of SEQ ID N0:55 (CDRl), SEQ ID N0:56 (CDR2) and SEQ ID N0:57 (CDR3) and a second heavy chain variable domain with heavy chain CDRS comprising the amino acid sequences of SEQ ID NO:58 (CDRl), SEQ ID NO:59 (CDR2) and SEQ ID NO:60 (CDR3).

In one red embodiment, the binding protein of the ion comprises (i) a ﬁrst polypeptide chain speciﬁcally binding to DLL3, comprising a ﬁrst light chain variable domain with light chain CDRs sing the amino acid sequences of SEQ ID N027 (CDRl), SEQ ID N018 (CDR2) and SEQ ID N029 (CDR3) and a ﬁrst heavy chain variable domain with heavy chain CDRs comprising the amino acid ces of SEQ ID NO: 10 (CDRl), SEQ ID N0: 11 (CDR2) and SEQ ID NO: 12 (CDR3); and (ii) a second polypeptide chain speciﬁcally binding to CD3, comprising a second light chain variable domain with light chain CDRs comprising the amino acid sequences of SEQ ID N0:55 (CDRl), SEQ ID N0:56 (CDR2) and SEQ ID N0:57 (CDR3) and a second heavy chain variable domain with heavy chain CDRs com-prising the amino acid sequences of SEQ ID N0258 (CDRl), SEQ ID N0:59 (CDR2) and SEQ ID NO:60 (CDR3).

In one preferred embodiment, the binding protein of the invention comprises (i) a ﬁrst polypeptide chain Speciﬁcally binding to DLL3, comprising a ﬁrst light chain variable domain with light chain CDRs comprising the amino acid sequences of SEQ ID N0:l3 (CDRl), SEQ ID NO: 14 (CDR2) and SEQ ID NO: 15 (CDR3) and a ﬁrst heavy chain _59_ variable domain with heavy chain CDRs comprising the amino acid sequences of SEQ ID NO: 16 , SEQ ID NO: 17 (CDR2) and SEQ ID NO: 18 (CDR3); and (ii) a second polypeptide chain speciﬁcally binding to CD3, comprising a second light chain variable domain with light chain CDRs comprising the amino acid ces of SEQ ID NO:55 (CDRl), SEQ ID NO:56 (CDR2) and SEQ ID NO:57 (CDR3) and a second heavy chain variable domain with heavy chain CDRs comprising the amino acid sequences of SEQ ID NO:58 (CDRl), SEQ ID NO:59 (CDR2) and SEQ ID NO:60 (CDR3).

In one preferred embodiment, the g protein of the invention comprises (i) a ﬁrst ptide chain speciﬁcally g to DLL3, comprising a ﬁrst light chain variable domain with light chain CDRs comprising the amino acid sequences of SEQ ID NO:1 (CDRl), SEQ ID NO:2 (CDR2) and SEQ ID NO:3 (CDR3) and a ﬁrst heavy chain variable domain with heavy chain CDRs comprising the amino acid sequences of SEQ ID NO:4 (CDRl), SEQ ID NO:5 (CDR2) and SEQ ID NO:6 (CDR3); and (ii) a second polypeptide chain speciﬁcally binding to CD3, comprising a second light chain variable domain with light chain CDRs comprising the amino acid ces of SEQ ID NO:61 (CDRl), SEQ ID NO:62 (CDR2) and SEQ ID NO:63 (CDR3) and a second heavy chain variable domain with heavy chain CDRs comprising the amino acid sequences of SEQ ID NO:64 (CDRl), SEQ ID NO:65 (CDR2) and SEQ ID NO:66 (CDR3).

In one preferred ment, the binding protein of the invention comprises (i) a ﬁrst polypeptide chain speciﬁcally binding to DLL3, comprising a ﬁrst light chain variable domain with light chain CDRs comprising the amino acid sequences of SEQ ID N027 (CDRl), SEQ ID NO:8 (CDR2) and SEQ ID NO:9 (CDR3) and a ﬁrst heavy chain variable domain with heavy chain CDRs comprising the amino acid sequences of SEQ ID NO: 10 (CDRl), SEQ ID NO: 11 (CDR2) and SEQ ID NO:12 (CDR3); and (ii) a second polypeptide chain speciﬁcally g to CD3, comprising a second light chain variable domain with light chain CDRs comprising the amino acid sequences of SEQ ID NO:61 (CDRl), SEQ ID NO:62 (CDR2) and SEQ ID NO:63 (CDR3) and a second heavy chain variable domain with heavy chain CDRs com-prising the amino acid sequences of SEQ ID NO:64 (CDRl), SEQ ID NO:65 (CDR2) and SEQ ID NO:66 .

In one preferred embodiment, the binding protein of the invention comprises (i) a ﬁrst polypeptide chain speciﬁcally binding to DLL3, comprising a ﬁrst light chain variable domain with light chain CDRs comprising the amino acid sequences of SEQ ID NO:13 —60- (CDRl), SEQ ID NO: 14 (CDR2) and SEQ ID NO: 15 (CDR3) and a ﬁrst heavy chain variable domain with heavy chain CDRs comprising the amino acid ces of SEQ ID NO: 16 , SEQ ID NO: 17 (CDRZ) and SEQ ID NO:18 (CDR3); and (ii) a second polypeptide chain speciﬁcally binding to CD3, comprising a second light chain variable domain with light chain CDRs comprising the amino acid sequences of SEQ ID NO:61 (CDRl), SEQ ID NO:62 (CDR2) and SEQ ID NO:63 (CDR3) and a second heavy chain variable domain with heavy chain CDRs com-prising the amino acid ces of SEQ ID NO:64 (CDRl), SEQ ID NO:65 (CDR2) and SEQ ID NO:66 (CDR3).

In preferred embodiments of the protein of the invention, said ﬁrst polypeptide chain speciﬁcally g to DLL3 ses a light chain variable domain (a ﬁrst light chain variable domain) and a heavy chain variable domain (a ﬁrst heavy chain variable domain) selected from the group consisting of i) to xviii): i) a light chain variable domain comprising the amino acid sequences of SEQ ID NO:37 and heavy chain variable domain comprising the amino acid sequences of SEQ ID NO:38 (DLL3#l); ii) a light chain le domain comprising the amino acid sequences of SEQ ID NO:39 and heavy chain le domain comprising the amino acid ces of SEQ ID NO:40 (DLL3#2); iii) a light chain variable domain comprising the amino acid sequences of SEQ ID NO:41 and heavy chain variable domain comprising the amino acid sequences of SEQ ID NO:42 (DLL3#3); iv) a light chain variable domain comprising the amino acid sequences of SEQ ID NO:43 and heavy chain variable domain comprising the amino acid sequences of SEQ ID NO:44 (DLL3#4); v) a light chain variable domain comprising the amino acid sequences of SEQ ID NO:45 and heavy chain variable domain comprising the amino acid sequences of SEQ ID NO:46 (DLL3#5); vi) a light chain variable domain comprising the amino acid sequences of SEQ ID NO:47 and heavy chain variable domain comprising the amino acid sequences of SEQ ID NO:48 (DLL3#6); —61- 2019/064942 Vii) a light chain variable domain sing the amino acid sequence of SEQ ID NO:205 and heavy chain variable domain comprising the amino acid ce of SEQ ID NO:206 (DLL3#7); viii) a light chain variable domain comprising the amino acid sequence of SEQ ID NO:207 and heavy chain variable domain comprising the amino acid sequence of SEQ ID NO:208 (DLL3#8); ix) a light chain variable domain comprising the amino acid ce of SEQ ID NO:209 and heavy chain variable domain comprising the amino acid sequence of SEQ ID NO:210 (DLL3#9); a light chain variable domain comprising the amino acid sequence of SEQ ID NO:211 and heavy chain variable domain comprising the amino acid sequence of SEQ ID NO:212 (DLL3#10); xi) a light chain variable domain comprising the amino acid sequence of SEQ ID NO:213 and heavy chain variable domain comprising the amino acid sequence of SEQ ID NO:214(DLL3#11); xii) a light chain variable domain comprising the amino acid ce of SEQ ID NO:215 and heavy chain variable domain comprising the amino acid sequence of SEQ ID NO:216 (DLL3#12); xiii) a light chain variable domain comprising the amino acid sequence of SEQ ID NO:217 and heavy chain variable domain comprising the amino acid sequence of SEQ ID NO:218 (DLL3#13); xiv) a light chain le domain comprising the amino acid sequence of SEQ ID NO:219 and heavy chain variable domain comprising the amino acid sequence of SEQ ID NO:220 14); xv) a light chain variable domain comprising the amino acid sequence of SEQ ID NO:221 and heavy chain variable domain comprising the amino acid sequence of SEQ ID NO:222 (DLL3#15); xvi) a light chain variable domain comprising the amino acid sequence of SEQ ID NO:223 and heavy chain variable domain comprising the amino acid sequence of SEQ ID NO:224 (DLL3#16); —62- 2019/064942 xvii) a light chain variable domain sing the amino acid sequence of SEQ ID NO:225 and heavy chain variable domain comprising the amino acid sequence of SEQ ID NO:226 (DLL3#17); and xviii) a light chain variable domain sing the amino acid sequence of SEQ ID NO:227 and heavy chain variable domain comprising the amino acid sequence of SEQ ID NO:228 (DLL3#18).

Preferably, the light chain variable and heavy chain variable domain sequences are ed from the group consisting of i) to iii) (DLL3#1, DLL3#2, DLL3#3) as deﬁned above.

In preferred embodiments of the protein of the ion, said second polypeptide chain speciﬁcally binding to CD3 comprises a light chain variable domain (a second light chain variable domain) and a heavy chain variable domain (a second heavy chain variable domain) selected from the group consisting of: i) a light chain variable domain comprising the amino acid sequence of SEQ ID NO:67 and a heavy chain variable domain comprising the amino acid sequence of SEQ ID NOZ68 (CD3#1); ii) a light chain variable domain comprising the amino acid sequence of SEQ ID NO:69 and a heavy chain variable domain comprising the amino acid ce of SEQ ID NO:70 (CD3#2); and iii) a light chain variable domain comprising the amino acid sequence of SEQ ID NO:102 and a heavy chain variable domain sing the amino acid sequence of SEQ ID NO: 103 (CD3#3).

In some embodiments, the g protein of the invention comprises a ﬁrst and a second polypeptide chain comprising CDR and/or VH and VL sequences of the light/heavy chain variable domains selected from the list consisting of DLL3#l/CD3#1, DLL3#2/CD3#1, DLL3#3/CD3#1, DLL3#4/CD3#1, DLL3#5/CD3#1, DLL3#6/CD3#1, DLL3#7/CD3#1, DLL3#8/CD3#1, DLL3#9/CD3#1, DLL3#10/CD3#1, DLL3#1l/CD3#1, DLL3#12/CD3#1, DLL3#13/CD3#1, 4/CD3#1, DLL3#15/CD3#1, DLL3#16/CD3#1, DLL3#l7/CD3#1, DLL3#18/CD3#1, DLL3#1/CD3#2, DLL3#2/CD3#2, DLL3#3/CD3#2, DLL3#4/CD3#2, DLL3#5/CD3#2, DLL3#6/CD3#2, DLL3#7/CD3#2, DLL3#8/CD3#2, DLL3#9/CD3#2, DLL3#lO/CD3#2, DLL3#1l/CD3#2, DLL3#12/CD3#2, DLL3#13/CD3#2, DLL3#14/CD3#2, DLL3#15/CD3#2, DLL3#16/CD3#2, DLL3#l7/CD3#2, 8/CD3#2, DLL3#1/CD3#3, _63_ DLL3#2/CD3#3, DLL3#3/CD3#3, DLL3#4/CD3#3, DLL3#5/CD3#3 and DLL3#6/CD3#3, DLL3#7/CD3#3, DLL3#8/CD3#3, DLL3#9/CD3#3, DLL3#10/CD3#3, DLL3#1l/CD3#3, DLL3#12/CD3#3, DLL3#13/CD3#3, DLL3#14/CD3#3, /CD3#3, DLL3#16/CD3#3, DLL3#l7/CD3#3, DLL3#18/CD3#3. In preferred embodiments, the g protein of the ion comprises a ﬁrst and a second polypeptide chain comprising CDR and/or VH and VL sequences of the light/heavy chain variable domains selected from the list consisting of DLL3#l/CD3#1, DLL3#2/CD3#1, DLL3#3/CD3#1, DLL3#l/CD3#2, DLL3#2/CD3#2, DLL3#3/CD3#2, DLL3#l/CD3#3, DLL3#2/CD3#3, and DLL3#3/CD3#3.

In one preferred embodiment, the binding protein of the invention comprises (i) a ﬁrst polypeptide chain speciﬁcally binding to DLL3, comprising a light chain variable domain of SEQ ID NO:37 and heavy chain variable domain of SEQ ID NO:38; and (ii) a second polypeptide chain speciﬁcally binding to CD3, comprising a light chain variable domain of SEQ ID NO:67 and a heavy chain variable domain of SEQ ID NO:68.

In one preferred embodiment, the binding protein of the invention comprises (i) a ﬁrst polypeptide chain speciﬁcally binding to DLL3, comprising a light chain variable domain of SEQ ID NO:39 and heavy chain variable domain of SEQ ID NO:40; and (ii) a second polypeptide chain speciﬁcally binding to CD3, comprising a light chain variable domain of SEQ ID NO:67 and a heavy chain variable domain of SEQ ID NO:68.

In one preferred embodiment, the binding protein of the invention comprises (i) a ﬁrst polypeptide chain speciﬁcally binding to DLL3, comprising a light chain variable domain of SEQ ID NO:41 and heavy chain variable domain of SEQ ID NO:42; and (ii) a second polypeptide chain speciﬁcally binding to CD3, comprising a light chain variable domain of SEQ ID NO:67 and a heavy chain variable domain of SEQ ID NO:68.

In one preferred embodiment, the g protein of the ion ses (i) a ﬁrst polypeptide chain speciﬁcally binding to DLL3, sing a light chain variable domain of SEQ ID N037 and heavy chain variable domain of SEQ ID NO:3 8; and (ii) a second ptide chain speciﬁcally binding to CD3, comprising a light chain le domain of SEQ ID NO:69 and a heavy chain le domain of SEQ ID NO:70.

In one preferred embodiment, the binding protein of the invention comprises (i) a ﬁrst polypeptide chain speciﬁcally binding to DLL3, comprising a light chain variable domain of SEQ ID N039 and heavy chain variable domain of SEQ ID NO:40; and (ii) a second —64- polypeptide chain speciﬁcally binding to CD3, comprising a light chain variable domain of SEQ ID NO:69 and a heavy chain variable domain of SEQ ID NO:70.

In one red embodiment, the g n of the invention comprises (i) a ﬁrst polypeptide chain speciﬁcally binding to DLL3, comprising a light chain variable domain of SEQ ID NO:41 and heavy chain le domain of SEQ ID NO:42; and (ii) a second polypeptide chain speciﬁcally binding to CD3, comprising a light chain variable domain of SEQ ID NO:69 and a heavy chain variable domain of SEQ ID NO:70.

In some embodiments of the invention, the ﬁrst and/or the second linker peptide comprise a linker as described above, e. g. a linker derived from a hinge region, a poly-alanine linker or a Gly/Ser linker, wherein the linker comprises 26 to 42 amino acids, for example any one of 30 to 40 amino acids, 34 to 40 amino acids, or 36 to 39 amino acids, preferably 38 amino acids.

In preferred embodiments, the ﬁrst polypeptide chain speciﬁc for DLL3 comprises the amino acid sequence of any one of SEQ ID NO:49, SEQ ID NO:50, SEQ ID NO:51, SEQ ID NO:52, SEQ ID NO:53, SEQ ID NO:54, SEQ ID NO:229, SEQ ID , SEQ ID NO:231, SEQ ID NO:232, SEQ ID NO:233, SEQ ID NO:234, SEQ ID NO:235, SEQ ID NO:236, SEQ ID NO:237, SEQ ID NO:238, SEQ ID NO:239, or SEQ ID NO:240 and the second polypeptide chain speciﬁc for CD3 comprises the amino acid sequence of SEQ ID NO:7l.

In some embodiments, the ﬁrst ptide chain speciﬁc for DLL3 comprises the amino acid sequence of any one of SEQ ID NO:49, SEQ ID NO:50, SEQ ID NO:51, SEQ ID NO:52, SEQ ID NO:53, SEQ ID NO:54, SEQ ID NO:229, SEQ ID NO:230, SEQ ID NO:231, SEQ ID NO:232, SEQ ID NO:233, SEQ ID NO:234, SEQ ID , SEQ ID NO:236, SEQ ID NO:237, SEQ ID , SEQ ID NO:239, or SEQ ID NO:240 and the second polypeptide chain speciﬁc for CD3 comprises the amino acid sequence of SEQ ID NO:72.

In some embodiments, the ﬁrst polypeptide chain speciﬁc for DLL3 comprises the amino acid ce of any one of SEQ ID NO:49, SEQ ID NO:50, SEQ ID NO:51, SEQ ID NO:52, SEQ ID NO:53, SEQ ID NO:54, SEQ ID NO:229, SEQ ID NO:230, SEQ ID NO:231, SEQ ID NO:232, SEQ ID NO:233, SEQ ID NO:234, SEQ ID NO:235, SEQ ID NO:236, SEQ ID NO:237, SEQ ID NO:23 8, SEQ ID NO:239, or SEQ ID NO:240 and the —65— second ptide chain speciﬁc for CD3 comprises the amino acid sequence of SEQ ID NO:104.

In one preferred embodiment, the ﬁrst polypeptide chain speciﬁc for DLL3 comprises the amino acid sequence of SEQ ID NO:49 and the second polypeptide chain speciﬁc for CD3 comprises the amino acid sequence of SEQ ID NO:71.

In one preferred embodiment, the ﬁrst polypeptide chain speciﬁc for DLL3 comprises the amino acid sequence of SEQ ID NO:50 and the second polypeptide chain speciﬁc for CD3 comprises the amino acid sequence of SEQ ID NO:71.

In one preferred ment, the ﬁrst ptide chain speciﬁc for DLL3 comprises the amino acid ce of SEQ ID NO:51 and the second polypeptide chain speciﬁc for CD3 comprises the amino acid sequence of SEQ ID NO:71.

In one preferred embodiment, the ﬁrst polypeptide chain speciﬁc for DLL3 comprises the amino acid sequence of SEQ ID NO:49 and the second polypeptide chain speciﬁc for CD3 comprises the amino acid sequence of SEQ ID NO:72.

In one preferred embodiment, the ﬁrst polypeptide chain speciﬁc for DLL3 comprises the amino acid sequence of SEQ ID NO:50 and the second polypeptide chain speciﬁc for CD3 comprises the amino acid sequence of SEQ ID NO:72.

In one preferred embodiment, the ﬁrst polypeptide chain speciﬁc for DLL3 comprises the amino acid sequence of SEQ ID NO:51 and the second polypeptide chain speciﬁc for CD3 comprises the amino acid sequence of SEQ ID NO:72.

In some embodiments of binding protein of the invention, the ﬁrst and second polypeptide chain comprises an EC domain derived from the heavy chain of an IgG, for example an IgGl, IgG2 or IgG4. For example, an EC domain of the t invention is a PC domain of a heavy chain of an IgGl or IgG4 and comprises a hinge region and two constant domains (CH2 and CH3). Examples of Fc domains ofhuman IgGs are shown in SEQ ID NO:81 and In some embodiments of the binding protein of the invention, the heavy chain comprises one or more amino acid changes. For example, such amino acid changes are a ne (Y) at position 366 ] of the ﬁrst heavy chain and a threonine (T) at position 407 [Y407T] of the second heavy chain. In some ments, the ﬁrst heavy chain comprises a serine (S) at on 366 [T3 66S] and the second heavy chain comprises a tryptophan —66- (W) at position 366 [T366W], an alanine (A) at position 368 [L368A] and a valine (V) at position 407 [Y407V]. In preferred embodiments, the ﬁrst heavy chain comprises a tryptophan (W) at position 366 [T366W] and the second heavy chain comprises a serine (S) at position 366 [T366S], an alanine (A) at on 368 [L368A] and a valine (V) at position 407 [Y407V]. For example, position 366 of the Fc domain according to EU numbering, corresponding to the amino acid position 146 in the human IgG1 Fc sequence of SEQ ID NO:81, is changed from T at position 146 in SEQ ID NO:81 to W at position 146 in SEQ ID NO:82; and positions 366, 368 and 407 according to EU numbering, ponding to the amino acid positions 146, 148 and 187, respectively, in SEQ ID NO:81, are changed from T, L and Y at these positions in SEQ ID NO:81 to S, A and V at these positions in SEQ ID NO:83. In any of these embodiments, the amino acid changes described for the ﬁrst heavy chain may be located in the second heavy chain and the respective amino acid changes for the second heavy chain may be located in the ﬁrst heavy chain. In other words, the term “ﬁrst” and “second” can be ged in these embodiments. In some embodiments, the heavy chain is derived from the heavy chain of an IgG1 or IgG4.

In some embodiments, the ﬁrst heavy chain or the second heavy chain in a protein of the present invention further comprises one or more amino acid changes which reduce the binding of the heavy chain to protein A. In some embodiments, such amino acid changes are an arginine at position 435 [H435R] and a phenylalanine at position 436 [Y436F] of one of the heavy chains.

In some embodiments, in a protein of the present invention, the heavy chain, which comprises a threonine (T) at position 407 [Y407T], further ses an ne at on 435 [H43 SR] and a phenylalanine at position 436 [Y43 6F]. In this case, the other heavy chain comprises a ne (Y) at position 366 [T366Y], but does not include the two changes at positions 435 and 43 6. atively, in some embodiments, in a protein of the present invention, the heavy chain, which comprises a serine (S) at position 366 [T3 66S], an alanine (A) at position 368 [L368A] and a valine (V) at position 407 [Y407V], further comprises an arginine at position 435 [H435R] and a phenylalanine at position 436 [Y436F]. In this case, the other heavy chain comprises a tryptophan (W) at position 366 [T3 66W], but does not include the two changes at ons 435 and 436. Thus, the heavy chain sing the amino acid change resulting in a “cavity” as bed above also comprises the amino acid changes, which reduce binding to n A. Homodimers —67— comprising these heavy chains are removed through reduced binding to Protein A. The production of homodimers of the other heavy chain, which ses the “protrusion”, is reduced by the presence of the “protrusion”.

In some embodiments, the heavy chain of a protein of the present invention may or may not further comprise YTE mutations (M252Y/SZS4T/T25 6E, EU numbering (Dall'Acqua, Kiener et a1. 2006)). These mutations have been shown to improve the pharmacokinetic properties of heavy chain through ential enhancement of g afﬁnity for neonatal FcRn receptor at pH 6.0.

In some embodiments, the ﬁrst and/or the second heavy chain of the present invention derived from an IgGl also includes the “KO” mutations (L234A, L235A). In a further aspect, the ﬁrst and/or the second heavy chain of the present invention derived from an IgG4 also includes the Pro hinge mutation (SZ28P).

In a further aspect, the proteins of the invention comprise a ﬁrst antigen binding unit or polypeptide chain c for DLL3 with an afﬁnity of preferably 3 lOnM, more preferably S_lnM, even more preferable S 0.1nM, even more preferably £0.01 nM to human and cynomolgus monkey DLL3. The afﬁnity can be measured in a SPR (BIAcore) assay using recombinant DLL3 -protein, as described, e. g. in the examples or other methods that are well known for the skilled person. The proteins comprise a second antigen binding unit or polypeptide chain with an y of ably 5500 nM, more preferably 5 100 nM, even more preferably 3 10 nM to human and cynomolgus monkey CD3EV complex.

In a r aspect, the DLL3/CD3 binding proteins of the invention do not bind to DLL3- negative cells and do not cross-react with the human and DLL3 paralogues DLLl and DLL4, as shown in Example 5.

In a further aspect, the DLL3/CD3 binding proteins of the present invention comprise a ﬁrst antigen binding unit or a ﬁrst ptide chain cally binding to the membrane al peptide of the DLL3 protein. In a further aspect, the proteins of the invention display weak binding to DLL3 protein expressing cells, e. g. no saturation of protein binding to the cell surface is achieved up to a concentration of 100 nM (see e. g. Figure 6).

In a r aspect, the DLL3/CD3 binding proteins of the present invention are capable of ing cytotoxicity against tumor cells by providing optimal steric conditions for the formation of a tic synapse n a tumor cell expressing DLL3 and a T cell, in —68- order to redirect the T cell activity selectively to the targeted tumor cells, leading to tumor cell lysis.

Various methods can be used to measure the cytotoxicity mediated by the DLL3/CD3 binding proteins of the present invention. For example, xicity can be measured using the method described in example 10. Effector cells can be e.g. stimulated or unstimulated (human or cynomolgus monkey) T cells or their subsets (e. g. CD4, CD8) or unstimulated (human or cynomolgus monkey) peripheral blood clear cells ). The target cells should express at least the extracellular domain of (human or cynomolgus monkey) DLL3 and can be cells with endogenous (natural) DLL3 sion, such as human small cell lung carcinoma cell lines SHP77, NCI-H82, alternatively also recombinant cells that s either the full-length DLL3 or the ellular domain of DLL3. The effector to target cell ratio (ET) is usually about 10:1 but can vary. Cytotoxic activity of DLL3/CD3 binding molecules can be determined e. g. in a LDH—release assay after 48 or 72 hours of incubation. Modiﬁcations in incubation time and read-out used for determination of cytotoxicity are possible and known to the skilled person. Read—out systems for cytotoxicity can comprise MTT/MTS assays, ATP-based , FACS-based assays, 51- um release assays, sulforhodamine B (SRB) assays, colorimetric (WST) assays, clonogenic assays, ECIS technology and bioluminescent assays.

The cytotoxic activity mediated by D3 binding proteins of the t invention is preferably measured in a cell—based cytotoxicity assay. The cytotoxicity is represented by the EC90 values measured in the cytotoxicity assay. The skilled person is aware that an EC90 can be expected to be lower when puriﬁed T cells are used as effector cells, compared with PBMCs, the skilled person is also aware that the EC90 can be even lower when stimulated T cells are used. It can rmore expected that the EC90 values are lower when the target cells express a high number of DLL3 on the cell surface compared to cell expressing a low number of DLL3 molecules on the cell surface. The EC90 of the DLL3/CD3 binding protein is preferably 5 10 nM, more preferably 5 5 nM and even more preferablyg 1 nM.

Preferably, the multi-speciﬁc binding proteins of the invention do not induce/mediate lysis ofDLL3 negative cells. The term “do not induce/mediate lysis” of egative cells means that an DLL3/CD3 binding molecule does not induce or mediate lysis of more than %, ably not more than 20%, more preferably not more than 10% and particular not —69- more than 5% or DLL3-negative cells, s lysis of the DLL3 -positive lung carcinoma cell line is set to be 100%. This usually applies for concentrations of the binding protein of up to 1000 nM.

Preferably, the DLL3/CD3 binding proteins of the invention are not alized by the targeted cells. The rate of internalization can be assayed e. g. as described in Example 11.

Preferably, the internalization rate (e.g. measured as a decrease in cytotoxicity) is S 50% after a 4 hour pre-incubation of the DLL3/CD3 binding proteins with the target cells, more ably 5 40% and even more preferably 33 0%.

Furthermore, the DLL3/CD3 binding proteins of the invention are shown to be stable with a monomer content above 95% (e.g., at least 98%, see example 14), have favorable pharmacokinetic properties and good downstream manufacturability and are further ed to have good bio—distribution (see e.g., example 12). The proteins of the present invention furthermore have a favorable immunogenicity proﬁle (see example 16) and have good stability in—vitro and o (see e.g., examples 12 and 15). Furthermore, the DLL3/CD3 binding proteins of the invention (e.g., /CD3#1, DLL3#3/CD3#2) show ble efﬁcacy in a humanized in viva xenograft mouse model. DLL3/CD3 binding proteins induced strong tumor regression starting already after the ﬁrst dose of DLL3/CD3 binding proteins. Furthermore the DLL3/CD3 binding ns of the invention induce tumor regression at very low doses of 0.25 mg/kg administered once weekly (q7d), r supporting their eutic applicability. In particular, the DLL3/CD3 binding proteins of the ion induce selective T cell proliferation, T cell tion, T cell degranulation and cytokine secretion (see example 18) only in the presence of DLL3- positive target cells and not in the presence of DLL3-negative target cells, and r signiﬁcantly increase T cell inﬁltration into tumor tissue (see example 20). Furthermore example 19 demonstrates that DLL3/CD3 binding ns mediate CD4+ as well as CD8+ T cell redirected lysis. In particular, na'ive T cells as well as CD4+ effector memory, CD4+ central memory, CD8+CD45RA+ effector and CD8+ memory cells contribute to the T cell redirected lysis of DLL3 —expression tumor cells.

A further aspect of the present invention provides isolated nucleic acid molecules encoding the ﬁrst and/or the second n binding unit of a multi-speciﬁc binding protein of the invention. In some embodiments, the nucleic acid molecules further encode a ﬁrst and/or a second Fc domain as described herein, the ﬁrst and/or second Fc domain linked to the 3’ _70_ end of the nucleic acid le encoding the ﬁrst and/or second antigen binding unit, tively. In some embodiments, the nucleic acid molecule encodes i) a ﬁrst polypeptide chain comprising a ﬁrst single chain Fab speciﬁc for DLL3 (e.g. any one of DLL3#1, DLL3#2, DLL3#3, DLL3#4, DLL3#5, DLL3 #6, DLL3#7, DLL3#8, DLL3#9, 0, DLL3#11, DLL3#12, DLL3#13, DLL3#14, DLL3#15, DLL3#16, DLL3#17 and DLL3#18), and optionally a ﬁrst PC domain and/or ii) a second polypeptide chain comprising a second single chain Fab speciﬁc for CD3 (e. g., any one of CD3#1, CD3#2 and CD3#3) and optionally a second Fc domain.

Preferably the nucleic acid molecule comprises a nucleotide ce encoding a ﬁrst single chain Fab of any one of SEQ ID NO:49, SEQ ID NO:50, SEQ ID NO:51, SEQ ID NO:52, SEQ ID NO:53, SEQ ID NO:54, SEQ ID NO:229, SEQ ID NO:230, SEQ ID NO:231, SEQ ID NO:232, SEQ ID NO:233, SEQ ID NO:234, SEQ ID NO:235, SEQ ID NO:236, SEQ ID NO:237, SEQ ID NO:238, SEQ ID NO:239, or SEQ ID NO:240 and/or a second single chain Fab of SEQ ID NO:71, SEQ ID NO:72 or SEQ ID NO: 104. In some embodiments, the nucleic acid molecule ses a nucleotide sequence encoding a ﬁrst polypeptide chain of any one of SEQ ID NO:73, SEQ ID NO:74, SEQ ID NO:75, SEQ ID NO:76, SEQ ID NO:77, SEQ ID NO:78, SEQ ID NO:241, 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, SEQ ID NO: 251, or SEQ ID NO: 252 and/or a second polypeptide chain speciﬁc for CD3 comprising the amino acid sequence of SEQ ID NO:79, SEQ ID NO:80 or SEQ ID NO: 105.

A further aspect of the invention es an expression vector containing a DNA molecule comprising the nucleotide sequence encoding the ﬁrst and/or second antigen binding domain (e.g. a ﬁrst and/or second single chain Fab of the ion). Preferably the expression vector ses, in addition, a nucleic acid molecule, preferably a DNA molecule, encoding a ﬁrst and/or second Fc domain, linked to the nucleic acid molecule, preferably the DNA le, encoding the ﬁrst and/or second antigen binding domain (e.g. ﬁrst and/or second single chain Fab, respectively). As such, the expression vector comprises a nucleotide ce encoding a polypeptide chain comprising a ﬁrst single chain Fab linked to a ﬁrst Fc domain and/or a nucleotide sequence encoding a polypeptide chain comprising a second single chain Fab linked to a second Fc domain. _71_ In a red embodiment, the expression vector contains a DNA molecule comprising the nucleotide ce ng the ﬁrst and/or the second polypeptide chain of the invention. In a preferred embodiment, the expression vector comprises the nucleotide sequence encoding a ﬁrst polypeptide chain of in any one of SEQ ID NO:73, SEQ ID NO:74, SEQ ID NO:75, SEQ ID NO:76, SEQ ID NO:77, SEQ ID NO:78, SEQ ID NO:241, 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, SEQ ID NO: 251, or SEQ ID NO: 252 and/or a second polypeptide chain comprising SEQ ID NO:79.

In further preferred embodiments, the expression vector comprises the nucleotide sequence encoding a ﬁrst polypeptide chain of any one of SEQ ID NO:73, SEQ ID NO:74, SEQ ID NO:75, SEQ ID NO:76, SEQ ID NO:77, SEQ ID NO:78, SEQ ID NO:241, 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, SEQ ID NO: 251, or SEQ ID NO: 252 and/or a second polypeptide chain sing SEQ ID NO:80.

In further red embodiments, the expression vector comprises the nucleotide sequence encoding a ﬁrst polypeptide chain of any one of SEQ ID NO:73, SEQ ID NO:74, SEQ ID NO:75, SEQ ID NO:76, SEQ ID NO:77, SEQ ID NO:78, SEQ ID NO:241, 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, SEQ ID NO: 251, or SEQ ID NO: 252 and/or a second polypeptide chain comprising SEQ ID NO:105.

In a cally preferred embodiment, two expression vectors may be used, one of them for expression of the ﬁrst polypeptide chain speciﬁc for DLL3, the other one for expression of the second polypeptide chain speciﬁc for CD3, which two expression vectors may then both be transfected into a host cell for recombinant protein expression.

Preferably, the expression vector will be a vector comprising said nucleic acid molecule or molecules, operably linked to at least one regulatory sequence, wherein such regulatory sequence may be a promoter, enhancer, or terminator sequence, and most preferably a heterologous promotor, enhancer, or terminator ce.

In another aspect, the invention relates to a host cell having an expression vector ng a ﬁrst polypeptide chain c for DLL3 of the invention and an expression vector encoding a second polypeptide chain speciﬁc for CD3 of the invention. _72_ ing to a particularly preferred embodiment, said host cells are otic cells such as mammalian cells. In another embodiment, such host cells are bacterial cells. Other useful cells are yeast cells or other fungal cells.

Suitable mammalian cells include for e CHO cells, BHK cells, HeLa cells, COS cells, and the like. However, amphibian cells, insect cells, plant cells, and any other cells used in the art for the expression of heterologous proteins can be used as well.

ANTI—DLL3 ANTIBODIES DLL3 is normally expressed on intracellular membranes, including those of the Golgi w apparatus in foetal brain, and plays a key role in somitogenesis in the paraxial mesoderm (Geffers et al., J Cell Biol. 2007;178:465; n et al., Hum. Mol. Genet. 2011;20:905— 916). Marked induction of DLL3 expression in some tumor types including SCLC, LCNEC and glioblastoma results in localization to the cell surface: this together with the absence of detectable cell surface DLL3 in non-malignant cells opens a new window of H opportunity for tumour-cell speciﬁc therapy and for the use of anti-DLL3 antibodies as diagnostic and stic tools.

Several anti-DLL3 antibodies (such as LS—cl67440, an; AP21739PU—N, Acris; LS- C148700, LSBio) are commercially available. r, these antibodies do not perform well in immunohistochemistry (IHC), FACS and ELISA assays and fail to speciﬁcally bind to DLL3 protein expressed on cells, particularly tumor cells. Anti-DLL3 antibodies have also been reported, e.g., in 111733 and suggested for diagnostic use in glioma patients, but no data of IHC assays are shown. Thus, the use of anti-DLL3 antibodies as reliable diagnostic tools to accurately measure DLL3 expression in patients, e.g. in tumor cells/tissues and/or to assess the efficacy of DLL3 targeted therapies remains challenging.

Thus, there is a need to identify alternative anti-DLL3 antibodies, which can be used for accurate and speciﬁc detection of DLL3 protein expression in various assays such as FACS, ELISA, precipitation, Western blotting, ELISA, radioimmunoassay, ﬂow cytometry, IHC and immunometric assays in any kind of biological sample and can be used as reliable diagnostic reagents. _73_ Therefore, a further aspect of the ion provides anti-DLL3 antibody molecules comprising i) light chain CDRs comprising the amino acid sequences of SEQ ID NO:1 (CDRl), SEQ ID NO:2 (CDR2) and SEQ ID NO:3 (CDR3) and heavy chain CDRs comprising the amino acid sequences of SEQ ID NO:4 (CDRl), SEQ ID NO:5 (CDR2) and SEQ ID NO:6 (CDR3); or ii) light chain CDRs comprising the amino acid sequences of SEQ ID NO:7 (CDRl), SEQ ID NO:8 (CDR2) and SEQ ID N029 (CDR3) and heavy chain CDRs comprising the amino acid ces of SEQ ID NO: 10 (CDRl), SEQ ID NO:ll (CDR2) and SEQ ID NO:12 (CDR3); or iii) light chain CDRs comprising the amino acid sequences of SEQ ID NO:13 (CDRl), SEQ ID NO:14 (CDR2) and SEQ ID NO:15 (CDR3) and heavy chain CDRs comprising the amino acid sequences of SEQ ID NO:16 , SEQ ID NO:17 (CDR2) and SEQ ID NO: 18 (CDR3); or iv) light chain CDRs sing the amino acid sequences of SEQ ID NO:19 (CDRl), SEQ ID NO:20 (CDR2) and SEQ ID NO:21 (CDR3) and heavy chain CDRs comprising the amino acid sequences of SEQ ID N0222 (CDRl), SEQ ID NO:23 (CDR2) and SEQ ID NO:24 (CDR3); or light chain CDRs comprising the amino acid sequences of SEQ ID NO:25 (CDRl), SEQ ID NO:26 (CDR2) and SEQ ID NO:27 (CDR3) and heavy chain CDRs comprising the amino acid sequences of SEQ ID N028 (CDRl), SEQ ID NO:29 (CDR2) and SEQ ID NO:30 (CDR3); or vi) light chain CDRs comprising the amino acid sequences of SEQ ID NO:31 (CDRl), SEQ ID NO:32 (CDR2) and SEQ ID NO:33 (CDR3) and heavy chain CDRs comprising the amino acid sequences of SEQ ID NO:34 (CDRl), SEQ ID NO:35 (CDR2) and SEQ ID NO:36 (CDR3); or vii) light chain CDRs comprising the amino acid sequences of SEQ ID NO: 133 (CDRl), SEQ ID NO: 134 (CDR2) and SEQ ID NO: 135 (CDR3) and heavy chain CDRs comprising the amino acid sequences of SEQ ID NO: 136 , SEQ ID NO:137 (CDR2) and SEQ ID NO:138 (CDR3); or _74_ viii) light chain CDRs sing the amino acid sequences of SEQ ID NO: 139 (CDRl), SEQ ID NO: 140 (CDR2) and SEQ ID NO: 141 (CDR3) and heavy chain CDRs comprising the amino acid sequences of SEQ ID NO: 142 (CDRl), SEQ ID NO:143 (CDR2) and SEQ ID NO: 144 (CDR3); or ix) light chain CDRs comprising the amino acid sequences of SEQ ID NO: 145 , SEQ ID NO: 146 (CDR2) and SEQ ID NO: 147 (CDR3) and heavy chain CDRs sing the amino acid sequences of SEQ ID NO: 148 (CDRl), SEQ ID NO:149 (CDR2) and SEQ ID NO:150 (CDR3); or light chain CDRs comprising the amino acid sequences of SEQ ID NO: 151 (CDRl), SEQ ID NO: 152 (CDR2) and SEQ ID NO: 153 (CDR3) and heavy chain CDRs comprising the amino acid sequences of SEQ ID NO: 154 (CDRl), SEQ ID NO:155 (CDR2) and SEQ ID NO:156 (CDR3); or xi) light chain CDRs comprising the amino acid sequences of SEQ ID NO: 157 (CDRl), SEQ ID NO: 158 (CDR2) and SEQ ID NO: 159 (CDR3) and heavy chain CDRs comprising the amino acid sequences of SEQ ID NO: 160 (CDRl), SEQ ID NO:161 (CDR2) and SEQ ID NO:162 (CDR3); or xii) light chain CDRs comprising the amino acid sequences of SEQ ID NO: 163 (CDRl), SEQ ID NO: 164 (CDR2) and SEQ ID NO: 165 (CDR3) and heavy chain CDRs comprising the amino acid sequences of SEQ ID NO: 166 (CDRl), SEQ ID NO:167 (CDR2) and SEQ ID NO:168 (CDR3); or xiii) light chain CDRs comprising the amino acid sequences of SEQ ID NO: 169 (CDRl), SEQ ID NO: 170 (CDR2) and SEQ ID NO: 171 (CDR3) and heavy chain CDRs comprising the amino acid sequences of SEQ ID NO: 172 (CDRl), SEQ ID NO:173 (CDR2) and SEQ ID NO:174 (CDR3); or xiv) comprising light chain CDRs comprising the amino acid sequences of SEQ ID NO:175 (CDRl), SEQ ID NO: 176 (CDR2) and SEQ ID NO: 177 (CDR3) and heavy chain CDRs comprising the amino acid sequences of SEQ ID NO: 178 (CDRl), SEQ ID NO: 179 (CDR2) and SEQ ID NO:180 (CDR3); or xv) light chain CDRs comprising the amino acid sequences of SEQ ID NO:181 (CDRl), SEQ ID NO: 182 (CDR2) and SEQ ID NO: 183 (CDR3) and heavy chain CDRs comprising the amino acid ces of SEQ ID NO:184 (CDRI), SEQ ID NO:185 (CDR2) and SEQ ID NO:186 (CDR3); or _75_ xvi) light chain CDRs comprising the amino acid ces of SEQ ID NO:187 (CDRl), SEQ ID NO: 188 (CDR2) and SEQ ID NO: 189 (CDR3) and heavy chain CDRs comprising the amino acid sequences of SEQ ID NO: 190 , SEQ ID NO:191 (CDR2) and SEQ ID NO:192 (CDR3); or xvii) light chain CDRs comprising the amino acid ces of SEQ ID NO: 193 (CDRl), SEQ ID NO:194(CDR2) and SEQ ID NO:195 (CDR3) and heavy chain CDRs comprising the amino acid ces of SEQ ID NO: 196 (CDRl), SEQ ID NO:197 (CDR2) and SEQ ID NO:198 (CDR3); or xviii) light chain CDRs sing the amino acid sequences of SEQ ID NO: 199 w (CDRl), SEQ ID NO:200 (CDR2) and SEQ ID NO:201 (CDR3) and heavy chain CDRs comprising the amino acid sequences of SEQ ID NO:202 (CDRl), SEQ ID NO:203 (CDR2) and SEQ ID NO:204 (CDR3).

The antibodies i) to xviii) as outlined above are termed DLL3#1, DLL3#2, DLL3#3, DLL3#4, DLL3#5, DLL3#6, DLL3#7, DLL3#8, DLL3#9, DLL3#10, DLL3#11, E DLL3#12, DLL3#13, DLL3#14, DLL3#15, DLL3#16, DLL3#17 or DLL3#18 respectively. Provided herein is a sequence table which readily allows identiﬁcation of individual amino acid sequences to speciﬁc antibodies of the present invention.

In preferred embodiments of the invention, the antibody molecule ses CDR sequences as deﬁned in i) or v) above corresponding to DLL3#1 or DLL3#5.

In some embodiments, the anti-DLL3 dy of the invention is a chimeric, a humanized or a human antibody molecule. In some embodiments, the antibody molecule is a monoclonal antibody Fab, F(ab)2, Fv or scFv. . In some embodiments, the anti-DLL3 antibody molecule of the invention comprises a heavy chain constant region selected from the group consisting of IgG1, IgG2, IgG3, IgG4, IgM, IgA and IgE constant regions. In some embodiments, the light chain nt region of the anti-DLL3 antibody molecule of the invention is kappa or lambda.

In some ments, the anti-DLL3 antibody of the invention has a heavy chain variable domain sing an amino acid sequence at least 85% identical to any one of SEQ ID NOs:38, 40, 42, 44, 46, 48, 206, 208, 210, 212, 214, 216, 218, 220, 222, 224, 226, and 228.

Preferably, the antibody molecule has a heavy chain variable domain comprising an amino acid sequence of SEQ ID NOs:38, 40, 42, 44, 46, 48, 206, 208, 210, 212, 214, 216, 218, 220, 222, 224, 226, or 228. _76_ In some embodiments, the anti-DLL3 antibody molecule has a light chain variable domain comprising an amino acid sequence at least 85% identical to any one of SEQ ID NOs:37, 39, 41, 43, 45, 47, 205, 207,209, 211, 213, 215, 217, 219, 221, 223, 225, and 227.

Preferably, the antibody molecule has a light chain variable domain sing an amino acid sequence of SEQ ID NOsz37, 39, 41, 43, 45, 47, 205, 207, 209, 211, 213, 215, 217, 219, 221, 223, 225, or 227.

Methods of calculating amino acid ce identities are well known in the art and further discussed herein in the Deﬁnitions section of the speciﬁcation.

In some embodiments, the LL3 antibody molecule has i) a heavy chain variable domain comprising the amino acid sequence of SEQ ID NO:38 and alight chain variable domain comprising the amino acid sequence of SEQ ID NO:37 (DLL3#1), or ii) a heavy chain variable domain sing the amino acid sequence of SEQ ID NO: 40 and a light chain variable domain comprising the amino acid sequence of SEQ ID NO: 39 (DLL3#2); or iii) a heavy chain variable domain comprising the amino acid sequence of SEQ ID NO:42 and a light chain variable domain comprising the amino acid sequence of SEQ ID NO:41 (DLL3#3), or iv) a heavy chain variable domain comprising the amino acid sequence of SEQ ID NO:44 and a light chain variable domain comprising the amino acid sequence of SEQ ID NO:43 (DLL3#4); or v) a heavy chain le domain comprising the amino acid sequence of SEQ ID NO:46 and a light chain variable domain comprising the amino acid ce of SEQ ID NO:45 (DLL3#5); or vi) a heavy chain variable domain comprising the amino acid sequence of SEQ ID NO:48 and a light chain variable domain comprising the amino acid ce of SEQ ID NO:47 (DLL3#6); or vii) a light chain variable domain comprising the amino acid sequence of SEQ ID NO:205 and heavy chain variable domain sing the amino acid sequence of SEQ ID NO:206 (DLL3#7); or viii) a light chain variable domain comprising the amino acid sequence of SEQ ID NO:207 and heavy chain variable domain comprising the amino acid sequence of SEQ ID NO:208 (DLL3#8); or ix) a light chain variable domain comprising the amino acid sequence of SEQ ID NO:209 and heavy chain variable domain comprising the amino acid sequence of SEQ ID NO:210 (DLL3#9); or x) a light chain variable domain sing the amino acid sequence of SEQ ID NO:211 and heavy chain variable domain comprising the amino acid sequence of SEQ ID NO:212 (DLL3#10); or xi)a light chain le domain comprising the amino acid sequence of SEQ ID NO:213 and heavy chain variable domain comprising the amino acid sequence of SEQ ID NO:214 (DLL3#11); or xii) a light chain _77_ variable domain comprising the amino acid sequence of SEQ ID NO:215 and heavy chain variable domain comprising the amino acid sequence of SEQ ID NO:216 12); or xiii) a light chain variable domain comprising the amino acid sequence of SEQ ID NO:217 and heavy chain variable domain sing the amino acid sequence of SEQ ID NO:218 (DLL3#13); or xiv) a light chain le domain comprising the amino acid sequence of SEQ ID NO:219 and heavy chain variable domain comprising the amino acid sequence of SEQ ID NO:220 (DLL3#14); or xv) a light chain variable domain comprising the amino acid sequence of SEQ ID NO:221 and heavy chain variable domain comprising the amino acid sequence of SEQ ID NO:222 (DLL3#15); or xvi) a light chain variable domain comprising the amino acid sequence of SEQ ID NO:223 and heavy chain variable domain sing the amino acid sequence of SEQ ID NO:224 (DLL3#16); or xvii) a light chain variable domain comprising the amino acid sequence of SEQ ID NO:225 and heavy chain variable domain comprising the amino acid sequence of SEQ ID NO:226 (DLL3#17); or xviii) a light chain variable domain comprising the amino acid sequence of SEQ ID NO:227 and heavy chain variable domain comprising the amino acid ce of SEQ ID NO:228 (DLL3#18).

In some embodiments, the anti-DLL3 antibody of the invention is a mouse monoclonal antibody. In the context of this ion a mouse monoclonal antibody includes an antibody where the VH and VL are obtained from immunization of mice with human DLL3 protein, subsequent selection of le VH and VL sequences binding with certain afﬁnity to human DLL3, and then further joining such VH and VL sequences to constant domains which are derived from mouse (e.g. from mouse IgG2a) by recombinant techniques; and which are produced by recombinant expression in host cells. Further encompassed by the invention are chimeric antibodies, e.g., comprising variable and constant regions from different species. In some embodiments, the antibody molecule of the invention is a ic antibody comprising VH and VL domains derived from mouse as bed above and further comprising constant domains derived from another species such as human, , rat, goat, donkey. In some ments, the chimeric antibody ses VH and VL domains derived from mouse and r humanized or sequence optimized as deﬁned above and further comprises constant domains derived from another species. In some embodiments, the chimeric antibody comprises VH and VL domains derived from a transgenic animal (e. g. a mouse) comprising human IgG sequences, thus comprises human VH and VL sequences, and further comprises constant domains derived —78— from another species. In any of the embodiments of chimeric antibodies as outlined above, the heavy chain constant region is a mouse, human, rabbit, rat, goat or donkey heavy chain region.

In some embodiments, the anti-DLL3 antibody molecule of the invention has a constant domain selected from the group consisting of IgG1, IgG2, IgG3, IgG4, IgM, IgA and IgE constant s. In a preferred embodiment, the anti-DLL3 antibody has a constant domain of IgG2a, preferably comprising the sequence of SEQ ID NO:254. In some embodiments, the anti-DLL3 antibody molecule has a light chain constant domain which is kappa or lambda, preferably the light chain nt domain is a kappa light chain constant w domain, preferably comprising the sequence of SEQ ID NO:255.

In some embodiments, the anti-DLL3 antibody molecule is capable of binding to human and cynomolgus monkey DLL3 with a iation constant (KD) of preferably 5 10nM, more preferably S_lnM, even more preferable S 0.1nM, even more preferably £0.01 nM.

The afﬁnity (KD value) can be measured in a SPR (BIAcore) assay using recombinant E DLL3 —protein, as described, e. g. in the examples or other methods that are well known for the skilled person.

In some embodiments, the anti—DLL3 antibody molecule does not bind to mouse DLL3.

In some embodiments, the anti-DLL3 antibody le is capable of detecting DLL3 expression (e. g. asmic and surface protein expression) on tissue samples (e.g., parafﬁn embedded/formalin-ﬁxed tissue samples) such as tumor tissue s, or ed cell lines. Optionally, the parafﬁn ed/formalin ﬁxed cell culture or tissue samples are further treated with an epitope retriever such as Proteinase K.

In some embodiments, the anti-DLL3 antibody molecule of the invention binds to the same e as any one of the DLL3#1, DLL2#2, DLL3#3, DLL3#4, DLL3#5, DLL3#6, DLL3#7, DLL3#8, DLL3#9, DLL3#10, DLL3#11, DLL3#12, DLL3#13, DLL3#14, DLL3#15, DLL3#16, DLL3#17, or DLL3#18 antibody. In some embodiments, the anti- DLL3 dy molecule of the invention competes for binding to human DLL3 with any one of the DLL3#1, DLL2#2, DLL3#3, DLL3#4, DLL3#5, DLL3#6, DLL3#7, DLL3#8, DLL3#9, 0, DLL3#11, DLL3#12, DLL3#13, DLL3#14, DLL3#15, DLL3#16, DLL3#17, or DLL3#18 antibody molecules. _79_ Another aspect of the present invention provides ed nucleic acid molecules ng the heavy chain variable domain and/or the light chain variable domain of an anti-DLL3 antibody molecule of the invention.

Preferably the c acid molecule comprises a nucleotide sequence ng the heavy chain variable domain of any one of SEQ ID NOsz38, 40, 42, 44, 46, 48, 206, 208, 210, 212, 214, 216, 218, 220, 222, 224, 226, or 228. Preferably the nucleic acid molecule comprises a tide sequence ng the light chain variable domain of any one of SEQ ID NOs: 37, 39, 41, 43, 45, 47, 205, 207, 209, 211, 213, 215, 217, 219, 221, 223, 225, or 227.

A further aspect of the invention provides an expression vector containing a DNA molecule comprising the nucleotide sequence encoding the heavy chain variable domain and/or the light chain variable domain of an anti-DLL3 antibody molecule of the invention.

Preferably the expression vector comprises, in addition, a nucleic acid molecule, preferably a DNA molecule, ng the constant domains of a heavy chain and/or the constant domain of a light chain, respectively, linked to the nucleic acid molecule, preferably the DNA molecule, encoding the heavy chain variable domain and/or the light chain variable domain, respectively.

In a specifically preferred embodiment, two expression vectors may be used, one of them for expression of the heavy chain, the other one for sion of the light chain, which two expression vectors may then both be transfected into a host cell for recombinant protein expression.

Preferably, the expression vector will be a vector comprising said nucleic acid molecule or molecules, operably linked to at least one regulatory sequence, wherein such tory sequence may be a promoter, enhancer, or ator sequence, and most preferably a heterologous promotor, enhancer, or terminator sequence.

In another aspect, the invention relates to a host cell having an sion vector encoding a heavy chain of an anti-DLL3 antibody molecule of the invention and an expression vector encoding a light chain of an LL3 antibody le of the invention.

According to a particularly preferred embodiment, said host cells are eukaryotic cells such as mammalian cells. In another embodiment, such host cells are bacterial cells. Other useful cells are yeast cells or other fungal cells. —80- Suitable mammalian cells include for example CHO cells, BHK cells, HeLa cells, COS cells, and the like. However, amphibian cells, insect cells, plant cells, and any other cells used in the art for the expression of heterologous proteins can be used as well.

METHODS OF MANUFACTURE AND PURIFICATION The invention further provides methods of manufacturing a multi-speciflc binding protein of the invention, such methods generally comprising the steps of: — culturing host cells comprising an expression vector comprising a nucleic acid encoding a binding protein of the invention under conditions that allow formation of the binding protein of the invention; and, — recovering the binding n expressed by the host cells from the culture; and - optionally further purifying and/or modifying and/or formulating the g protein of the invention.

The invention further provides methods of cturing an anti-DLL3 antibody of the ion, such methods generally comprising the steps of: - culturing host cells comprising an expression vector sing a nucleic acid encoding an antibody molecule of the invention under ions that allow formation of the antibody molecule; and, — recovering the antibody molecule expressed by the host cells from the culture; and - optionally further ing and/or modifying and/or formulating the antibody molecule of the invention.

A nucleic acid of the ion can e. g. be a DNA molecule sing coding sequences as well as regulatory sequences and optionally natural or artiﬁcial introns, or can be a cDNA molecule. It may have its original codons or may have an optimized codon usage that has been speciﬁcally adapted for expression in the intended host cell or host organism.

According to one embodiment of the invention, the c acid of the invention is in ially isolated form, as defined above.

The nucleic acids of the ion may be prepared or obtained in a manner known per se (e.g. by automated DNA synthesis and/or recombinant DNA technology), based on the ation on the amino acid sequences for the proteins of the invention given herein. —81- The nucleic acid of the invention will typically be incorporated into an sion , i.e. a vector that can provide for expression of the protein when transfected into a suitable host cell or other expression system.

For manufacturing the binding proteins or antibodies of the invention, the d artisan may choose from a great variety of expression systems well known in the art, e. g. those reviewed by Kipriyanow and Le Gall, 2004.

Expression s include plasmids, retroviruses, cosmids, EBV derived episomes, and the like. The expression vector and expression control sequences are selected to be compatible with the host cell. The nucleotide sequence encoding the ﬁrst antigen binding unit (e. g. the DLL3 specific single chain Fab or the full length DLL3 chain of the binding protein of the ion) and the nucleotide sequence encoding the second antigen binding unit (e. g. CD3 specific single chain Fab or the full length CD3 chain of the binding protein of the invention) of the DLL3/CD3 binding protein can be ed into separate vectors. In certain embodiments, both DNA sequences are inserted into the same expression vector.

The nucleotide sequence encoding the light chain of a DLL3 antibody and the nucleotide sequence encoding the heavy chain of a DLL3 antibody can be inserted into separate vectors. In certain embodiments, both DNA sequences are inserted into the same expression vector.

Convenient vectors are those that encode a functionally complete human CH (constant heavy) immunoglobulin sequence, with riate restriction sites engineered so that any antigen binding unit such as a single chain Fab sequence or any heavy/light chain variable domain can be easily ed and expressed, as described above. For the antibody heavy chain, it can be, without limitation, any IgG e (IgGl, IgG2, IgG3, IgG4) or other globulins, ing allelic variants.

The recombinant expression vector may also encode a signal peptide that facilitates ion of the full length CD3 or DLL3 chain from a host cell or of the light/heavy chain of an anti-DLL3 antibody. The DNA encoding the protein chain may be cloned into the vector such that the signal peptide is linked in—frame to the amino terminus of the mature full length chain DNA. The signal peptide may be an immunoglobulin signal peptide or a heterologous peptide from a non-immunoglobulin protein. Alternatively, the DNA sequence ng the full length chains of the protein of the ion may already contain a signal peptide sequence. —82- In addition to the D3 chain encoding DNA sequences or the heavy/light chain of a DLL3 antibody encoding DNA ces, the recombinant expression vectors typically carries regulatory sequences, optionally heterologous tory sequences, including promoters, enhancers, ation and polyadenylation signals and other expression control ts that control the expression of the protein chains in a host cell. Examples for er sequences lified for expression in mammalian cells) are promoters and/or enhancers derived from CMV (such as the CMV Simian Virus 40 (SV40) promoter/enhancer), adenovirus, (e. g., the adenovirus major late promoter (AdMLP)), polyoma and strong mammalian promoters such as native immunoglobulin and actin promoters. Examples for polyadenylation signals are BGH polyA, SV40 late or early polyA; alternatively, 3 ’UTRs of immunoglobulin genes etc. can be used.

The inant expression vectors may also carry ces that te replication of the vector in host cells (e. g. origins of replication) and selectable marker genes. Nucleic acid molecules encoding the full length chain with the first antigen g unit (single chain Fab and Fe domain) or an antigen-binding n thereof and/or the full length chain with the second antigen binding unit (single chain Fab and Fe domain) or an antigen- binding portion thereof, and vectors comprising these DNA molecules can be introduced into host cells, e.g. bacterial cells or higher eukaryotic cells, e.g. mammalian cells, according to transfection methods well known in the art, including liposome—mediated transfection, polycation-mediated transfection, protoplast fusion, microinj ections, calcium phosphate precipitation, electroporation or transfer by viral vectors.

Preferably, the DNA molecules encoding the DLL3 and CD3 chain of the protein of the invention are present on two expression vectors which are co—transfected into the host cell, preferably a mammalian cell.

Mammalian cell lines available as hosts for expression are well known in the art and include, inter alia, Chinese hamster ovary (CHO) cells, NSO, SP2/0 cells, HeLa cells, baby hamster kidney (BHK) cells, monkey kidney cells (COS), human carcinoma cells (e. g., Hep G2 and A—549 cells), 3T3 cells or the derivatives/progenies of any such cell line.

Other mammalian cells, including but not limited to human, mice, rat, monkey and rodent cells lines, or other eukaryotic cells, including but not limited to yeast, insect and plant cells, or prokaryotic cells such as bacteria may be used. —83- The proteins of the invention are produced by culturing the host cells for a period of time sufﬁcient to allow for sion of the protein in the host cells. Protein les are ably recovered from the e medium as a secreted polypeptide or it can be recovered from host cell lysates if for example expressed without a secretory signal. It is necessary to purify the protein molecules using standard protein puriﬁcation methods used for recombinant proteins and host cell proteins in a way that ntially homogenous preparations of the protein are obtained. By way of example, state—of—the art puriﬁcation methods useful for obtaining n molecules of the invention include, as a ﬁrst step, removal of cells and/or particulate cell debris from the culture medium or lysate. The protein is then puriﬁed from contaminant soluble proteins, polypeptides and nucleic acids, for example, by fractionation on immunoafﬁnity or ion-exchange s, ethanol precipitation, reverse phase HPLC, ex chromatography, chromatography on silica or on a cation exchange resin. As a ﬁnal step in the process for obtaining a protein molecule preparation, the puriﬁed protein molecule may be dried, e.g. lyophilized, as described below for therapeutic applications.

The present invention relates to binding proteins that have binding speciﬁcities for at least two different targets. In relation to the present invention, the binding molecules are d from antibodies. Techniques for making binding molecules include, but are not d to, inant co-expression of two immunoglobulin chains having different speciﬁcities (see Milstein and Cuello, Nature 305: 537 (1983)), WO 93/08829, and Traunecker et al., EMBO J. 10: 3655 (1991)), and "knob—in—hole" engineering (see, e.g., US. Patent No. 5,731, 168; Atwell et a1, JMB, 1997, 270, 26—35). g proteins of the invention may also be made by engineering electrostatic steering effects for making antibody Fc—heterodimeric molecules (A1); cross— g two or more antibodies or nts (see, e.g., US Patent No. 4,676,980, and Brennan et al., e, 229: 81 (1985)); using leucine zippers to produce ciﬁc proteins (see, e.g., Kostelny et al., Immunol., 148(5): 1547—1553 (1992)); using "diabody" technology for making bispeciﬁc antibody fragments (see, e. g., Hollinger et al., Proc. Natl. Acad. Sci. USA, 90:6444- 6448 (1993)); and using single-chain Fv (sFV) dimers (see, e.g. Gruber et al., /.

Immunol., 152:5368 (1994)); and preparing trispeciﬁc antibodies as described, e. g., in Tutt et a1. /. Immunol. 147: 60 (1991).

The compositions (e.g., multi-speciﬁc binding proteins and anti-DLL3 antibodies) and methods disclosed herein encompass polypeptides and c acids having the sequences —84- speciﬁed, or sequences substantially identical or similar thereto, e. g., ces at least 85%, 90%, 95% identical or higher to the sequence speciﬁed. In the t of an amino acid ce, the term "substantially identical" is used herein to refer to a ﬁrst amino acid sequence that contains a sufﬁcient or minimum number of amino acid residues that are i) identical to, or ii) conservative substitutions of aligned amino acid residues in a second amino acid sequence such that the ﬁrst and second amino acid sequences can have a common structural domain and/or common functional activity. For example, amino acid sequences that contain a common structural domain have at least about 85%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98% or 99% identity to a reference sequence, e.g., a sequence provided . In the context of tide sequence, the term "substantially identical" is used herein to refer to a ﬁrst nucleic acid ce that contains a sufﬁcient or minimum number of nucleotides that are identical to aligned nucleotides in a second nucleic acid sequence such that the ﬁrst and second nucleotide sequences encode a polypeptide having common functional activity, or encode a common structural polypeptide domain or a common functional ptide activity, for example, nucleotide sequences having at least about 85%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98% or 99% ty to a reference sequence.

The nucleic acid molecules of the invention include, but are not limited to, the DNA molecules ng the polypeptide sequences shown in the ce listing. Also, the present invention also relates to nucleic acid molecules that hybridize to the DNA molecules encoding the polypeptide sequences shown in the sequence listing under high ency binding and washing conditions, as deﬁned in . Preferred molecules (from an mRNA perspective) are those that have at least 75% or 80% (preferably at least 85%, more preferably at least 90% and most preferably at least 95%) homology or sequence identity with one of the DNA molecules described herein. By way of example, in view of expressing the dies in eukaryotic cells, the DNA sequences shown in the sequence listing have been designed to match codon usage in eukaryotic cells. If it is desired to express the antibodies in E. coli, these sequences can be changed to match E. coli codon usage. Variants of DNA molecules of the invention can be constructed in l different ways, as described e. g. in .

The proteins of the invention may have a modiﬁed N—terminal sequence, e.g. a deletion of one or more of the N-terminal amino acids, or an exchange of e. g. the ﬁrst, N-terminal amino acid (e. g. glutamate to alanine), to optimize the molecule for being expressed by —85- using certain expression s (such as speciﬁc vectors or host cells), or for being expressed as inclusion bodies or in soluble form, or for being secreted into the medium or the periplasmic space or for being contained Within the cell, or for ng a more homogenous product. The polypeptides of the ion may have a modiﬁed C-terminal sequence, such as an additional alanine, and/or further amino acid ges in the C- terminal part or at other deﬁned positions within any of the framework regions, as explained e.g. in WO2012/l7574l, WO201 1/075861, or WO2013/024059, in order to e.g. further enhance stability or reduce immunogenicity of such ptides.

For the avoidance of doubt, all of the embodiments relating to pharmaceutical compositions, kits, treatment methods, medical uses, combinations, methods of administration and dosages as described herein are contemplated for any of the multi- speciﬁc binding ns bed herein, either alone or in combination with further therapeutic agents (as speciﬁed in more detail .

PHARMACEUTICAL COMPOSITIONS; METHODS OF ADMINISTRATION; DOSAGES The invention further relates to pharmaceutical compositions for the treatment of a disease (as speciﬁed in more detail below), wherein such compositions comprise at least one multi— speciﬁc binding protein of the invention. The invention further encompasses methods of treating a disease (as ed in more detail below) using at least one multi—speciﬁc protein of the invention or pharmaceutical composition as set out below, and further encompasses the preparation of a medicament for the ent of such disease by using such binding protein of the invention or pharmaceutical composition.

The binding proteins of the invention (e.g., any one of DLL3#l/CD3#1, DLL3#2/CD3#1, DLL3#3/CD3#1, DLL3#4/CD3#1, DLL3#5/CD3#1, DLL3#6/CD3#1, DLL3#7/CD3#1, DLL3#8/CD3#1, DLL3#9/CD3#1, DLL3#10/CD3#1, DLL3#1 l/CD3#1, DLL3#l2/CD3#1, DLL3#l3/CD3#1, DLL3#l4/CD3#1, DLL3#15/CD3#1, DLL3#l6/CD3#1, DLL3#17/CD3#1, DLL3#18/CD3#1, DLL3#l/CD3#2, /CD3#2, DLL3#3/CD3#2, DLL3#4/CD3#2, DLL3#5/CD3#2, DLL3#6/CD3#2, DLL3#7/CD3#2, /CD3#2, DLL3#9/CD3#2, DLL3#10/CD3#2, DLL3#1 l/CD3#2, DLL3#l2/CD3#2, DLL3#l3/CD3#2, DLL3#l4/CD3#2, DLL3#15/CD3#2, DLL3#l6/CD3#2, DLL3#17/CD3#2, DLL3#18/CD3#2, DLL3#l/CD3#3, —86- DLL3#2/CD3#3, DLL3#3/CD3#3, DLL3#4/CD3#3, DLL3#5/CD3#3 and DLL3#6/CD3#3, DLL3#7/CD3#3, DLL3#8/CD3#3, DLL3#9/CD3#3, DLL3#10/CD3#3, DLL3#1 l/CD3#3, DLL3#12/CD3#3, DLL3#13/CD3#3, DLL3#14/CD3#3, DLL3#15/CD3#3, DLL3#16/CD3#3, DLL3#l7/CD3#3, DLL3#1 3, preferably any one of DLL3#1/CD3#1, DLL3#1/CD3#2, DLL3#1/CD3#3, DLL3#2/CD3#1, DLL3#2/CD3#2, DLL3#2/CD3#3, DLL3#3/CD3#1, DLL3#3/CD3#2, DLL3#3/CD3#3 ) and/or the compositions comprising the same can be administered to a patient in need thereof in any suitable manner, depending on the c pharmaceutical formulation or composition to be used. Thus, the binding proteins of the ion and/or the compositions comprising the same can for e be administered intravenously (iv), subcutaneously (s.c.), intramuscularly , intraperitoneally (i.p.), transdermally, orally, sublingually (e. g. in the form of a sublingual , spray or drop placed under the tongue and adsorbed through the mucus membranes into the capillary network under the tongue), (intra-)nasally (e. g. in the form of a nasal spray and/or as an aerosol), topically, by means of a suppository, by inhalation, or any other suitable manner in an effective amount or dose. The binding protein can be administered by infusion, bolus or injection. In preferred embodiments, the administration is by intravenous on or subcutaneous injection.

The binding proteins of the ion and/or the compositions comprising the same are administered according to a regimen of treatment that is suitable for treating and/or alleviating the disease, disorder or condition to be treated or alleviated. The clinician will generally be able to determine a suitable treatment regimen, depending on factors such as the e, disorder or ion to be treated or alleviated, the severity of the disease, the severity of the symptoms thereof, the speciﬁc g protein of the invention to be used, the speciﬁc route of administration and pharmaceutical formulation or composition to be used, the age, , weight, diet, general condition of the patient, and similar factors well known to the clinician. Generally, the treatment regimen will comprise the administration of one or more binding proteins of the ion, or of one or more compositions comprising the same, in therapeutically effective amounts or doses. lly, for the ent and/or alleviation of the diseases, disorders and conditions mentioned herein and depending on the speciﬁc e, disorder or condition to be treated, the potency of the speciﬁc binding protein of the invention to be used, the speciﬁc route of administration and the speciﬁc pharmaceutical formulation or ition used, the binding ns of the invention will generally be administered in an amount between —87— 0.005 and 20.0 mg per am ofbody weight and dose, preferably between 0.05 and .0 mg/kg/dose, either continuously (e. g. by infusion) or more preferably as single doses (such as e.g. twice a week, , or monthly doses; cf. below), but can signiﬁcantly vary, especially, depending on the before-mentioned parameters. Thus, in some cases it may be sufficient to use less than the m dose given above, whereas in other cases the upper limit may have to be exceeded. When administering large amounts it may be advisable to divide them up into a number of smaller doses spread over the day.

Depending on the specific binding protein of the invention and its specific pharmacokinetic and other properties, it may be administered daily, every second, third, fourth, ﬁfth or sixth day, weekly, monthly, and the like. An administration regimen could include long-term, weekly treatment. By "long-term" is meant at least two weeks and ably months, or years of duration.

The efficacy of the multi-specific n of the invention, and of compositions comprising the same, can be tested using any suitable in vitro assay, cell-based assay, in vivo assay and/or animal model known per se, or any ation thereof, ing on the specific disease involved. Suitable assays and animal models will be clear to the skilled person, and for example include the assays and animal models used in the Examples below.

FORMULATIONS For pharmaceutical use, the binding proteins of the invention may be formulated as a pharmaceutical preparation sing (i) at least one binding protein of the invention (e.g., any one of DLL3#l/CD3#1, DLL3#2/CD3#1, DLL3#3/CD3#1, DLL3#4/CD3#1, DLL3#5/CD3#1, /CD3#1, DLL3#7/CD3#1, DLL3#8/CD3#1, DLL3#9/CD3#1, DLL3#10/CD3#1, DLL3#1 l/CD3#1, DLL3#12/CD3#1, DLL3#l3/CD3#1, DLL3#l4/CD3#1, DLL3#15/CD3#1, DLL3#l6/CD3#1, DLL3#l7/CD3#1, DLL3#l8/CD3#1, DLL3#1/CD3#2, DLL3#2/CD3#2, DLL3#3/CD3#2, DLL3#4/CD3#2, DLL3#5/CD3#2, DLL3#6/CD3#2, DLL3#7/CD3#2, DLL3#8/CD3#2, DLL3#9/CD3#2, DLL3#10/CD3#2, DLL3#1 l/CD3#2, DLL3#12/CD3#2, DLL3#l3/CD3#2, DLL3#l4/CD3#2, 5/CD3#2, DLL3#l6/CD3#2, DLL3#l7/CD3#2, DLL3#l8/CD3#2, DLL3#1/CD3#3, /CD3#3, DLL3#3/CD3#3, /CD3#3, DLL3#5/CD3#3 and DLL3#6/CD3#3, DLL3#7/CD3#3, DLL3#8/CD3#3, DLL3#9/CD3#3, DLL3#10/CD3#3, DLL3#1 l/CD3#3, DLL3#12/CD3#3, —88- DLL3#l3/CD3#3, DLL3#14/CD3#3, DLL3#15/CD3#3, DLL3#16/CD3#3, DLL3#l7/CD3#3, DLL3#l8/CD3#3, preferably any one of DLL3#1/CD3#1, DLL3#1/CD3#2, DLL3#l/CD3#3, DLL3#2/CD3#1, /CD3#2, DLL3#2/CD3#3, DLL3#3/CD3#1, DLL3#3/CD3#2, DLL3#3/CD3#3) and (ii) at least one pharmaceutically acceptable carrier, diluent, excipient, adjuvant, and/or stabilizer, and (iii) ally one or more further pharmacologically active polypeptides and/or compounds. By "pharmaceutically acceptable" is meant that the respective material does not show any biological or otherwise undesirable effects when administered to an dual and does not interact in a deleterious manner with any of the other components of the pharmaceutical ition (such as e.g. the pharmaceutically active ingredient) in which it is contained. c examples can be found in rd handbooks, such as e. g.

Remington's Pharmaceutical Sciences, 18th Ed., Mack Publishing Company, USA (1990).

For example, the binding proteins of the invention may be formulated and administered in any manner known per se for conventional antibodies and antibody fragments and other pharmaceutically active proteins. Thus, according to a further embodiment, the invention relates to a pharmaceutical composition or preparation that contains at least one binding protein of the invention and at least one pharmaceutically acceptable carrier, diluent, excipient, adjuvant and/or stabilizer, and optionally one or more further cologically active substances, in the form of lyophilized or otherwise dried formulations or aqueous or non-aqueous solutions or suspensions.

Pharmaceutical preparations for parenteral administration, such as intravenous, uscular, subcutaneous injection or intravenous on may for example be sterile solutions, suspensions, dispersions, emulsions, or s which comprise the active ingredient and which are suitable, optionally after a further dissolution or dilution step, for infusion or injection. Suitable carriers or diluents for such preparations for example e, without limitation, sterile water and pharmaceutically acceptable aqueous buffers and solutions such as physiological phosphate—buffered , 's solutions, dextrose solution, and Hank's solution; water oils; ol; ethanol; glycols such as propylene glycol, as well as mineral oils, animal oils and vegetable oils, for example peanut oil, n oil, as well as suitable mixtures thereof.

Solutions of the binding proteins of the invention may also contain a preservative to prevent the growth of microorganisms, such as antibacterial and ngal agents, for example, p-hydroxybenzoates, parabens, chlorobutanol, phenol, sorbic acid, thiomersal, —89— (alkali metal salts of) ethylenediamine tetraacetic acid, and the like. In many cases, it will be preferable to include isotonic agents, for example, , buffers or sodium chloride.

Optionally, emulsiﬁers and/or dispersants may be used. The proper ﬂuidity can be maintained, for example, by the ion of liposomes, by the maintenance of the required particle size in the case of dispersions or by the use of surfactants. Other agents delaying absorption, for example, aluminum earate and gelatin, may also be added.

The ons may be ﬁlled into ion vials, ampoules, on bottles, and the like.

In all cases, the ultimate dosage form must be sterile, ﬂuid and stable under the conditions of manufacture and storage. Sterile inj ectable solutions are prepared by incorporating the active compound in the required amount in the appropriate solvent with various of the other ingredients enumerated above, as required, followed by ﬁlter sterilization. In the case of sterile powders for the preparation of sterile injectable solutions, the preferred methods ofpreparation are vacuum drying and the freeze drying techniques, which yield a powder of the active ingredient plus any additional desired ingredient present in the previously sterile—ﬁltered solutions.

Usually, aqueous solutions or suspensions will be preferred. Generally, suitable formulations for therapeutic proteins such as the binding proteins of the invention are ed protein solutions, such as solutions including the protein in a le concentration (such as from 0.001 to 400 mg/ml, preferably from 0.005 to 200 mg/ml, more preferably 0.01 to 200 mg/ml, more preferably 1.0 - 100 mg/ml, such as 1.0 mg/ml (i.v. administration) or 100 mg/ml (s.c. administration) and an aqueous buffer such as: - phosphate buffered saline, pH 7.4, - other ate buffers, pH 6.2 to 8.2, - acetate s, pH 3.2 to 7.5, preferably pH 4.8 to 5.5 - ine buffers, pH 5.5 to 7.0, — succinate buffers, pH 3.2 to 6.6, and - citrate s, pH 2.1 to 6.2, and, optionally, salts (e.g. NaCl) and/or sugars (such as e. g. sucrose and trehalose) and/or other polyalcohols (such as e. g. ol and glycerol) for providing isotonicity of the solution.

In addition, other agents such as a detergent, e.g. 0.02 % Tween—20 or Tween-80, may be included in such solutions. Formulations for subcutaneous application may include signiﬁcantly higher concentrations of the antibody of the invention, such as up to _90_ 100 mg/ml or even above 100 mg/ml. r, it will be clear to the person skilled in the art that the ingredients and the amounts f as given above do only represent one, red option. Alternatives and variations thereof will be immediately apparent to the skilled person, or can easily be conceived starting from the above disclosure. The above described formulations can optionally be provided as lyophilized formulation that is to be reconstituted in a solution, e.g. in water for injection (WFI). ing to a further aspect of the invention, a binding protein of the invention may be used in combination with a device useful for the administration of protein, such as a syringe, injector pen, micropump, or other device.

METHOD OF TREATMENT A further aspect of the invention provides a method of treating cancer comprising administering to a patient in need thereof a therapeutically effective amount of the g protein of the invention.

A further aspect of the invention provides a binding n of the invention for use in a method of treating cancer.

A further aspect of the ion is the use of the binding protein of the invention for preparing a pharmaceutical composition for treating cancer.

For the avoidance of doubt, the medical use aspects of the invention may comprise any of the speciﬁc binding proteins of the invention as described above (e.g., any one of DLL3#1/CD3#1, DLL3#2/CD3#1, DLL3#3/CD3#1, DLL3#4/CD3#1, DLL3#5/CD3#1, /CD3#1, DLL3#7/CD3#1, DLL3#8/CD3#1, DLL3#9/CD3#1, DLL3#10/CD3#1, DLL3#11/CD3#1, DLL3#12/CD3#1, DLL3#13/CD3#1, DLL3#l4/CD3#1, DLL3#15/CD3#1, DLL3#l6/CD3#1, DLL3#17/CD3#1, DLL3#18/CD3#1, DLL3#1/CD3#2, DLL3#2/CD3#2, /CD3#2, DLL3#4/CD3#2, DLL3#5/CD3#2, DLL3#6/CD3#2, DLL3#7/CD3#2, DLL3#8/CD3#2, DLL3#9/CD3#2, DLL3#10/CD3#2, DLL3#11/CD3#2, DLL3#12/CD3#2, DLL3#13/CD3#2, 4/CD3#2, DLL3#15/CD3#2, 6/CD3#2, DLL3#17/CD3#2, DLL3#18/CD3#2, DLL3#l/CD3#3, DLL3#2/CD3#3, /CD3#3, DLL3#4/CD3#3, DLL3#5/CD3#3 and DLL3#6/CD3#3, DLL3#7/CD3#3, DLL3#8/CD3#3, DLL3#9/CD3#3, DLL3#10/CD3#3, DLL3#11/CD3#3, DLL3#12/CD3#3, DLL3#l3/CD3#3, DLL3#14/CD3#3, DLL3#15/CD3#3, DLL3#l6/CD3#3, DLL3#l7/CD3#3, _91_ DLL3#1 8/CD3#3, preferably any one of DLL3#l/CD3#1, DLL3#l/CD3#2, DLL3#l/CD3#3, DLL3#2/CD3#1, DLL3#2/CD3#2, DLL3#2/CD3#3, DLL3#3/CD3#1, DLL3#3/CD3#2, DLL3#3/CD3#3 ).

As used herein, the term "cancer" is meant to include all types of cancerous growths or oncogenic processes, metastatic tissues or malignantly transformed cells, tissues, or organs, irrespective of histopathologic type or stage of invasiveness. ary cancers whose growth can be inhibited using the multi-specific binding proteins described herein are any DLL3 sing tumors, preferably SCLC, LCNEC, glioma, astoma, melanoma, or other DLL3 expressing neuroendocrine tumors (e.g., neuroendocrine prostate cancer or neuroendocrine pancreatic , small cell bladder cancer).

Neuroendocrine tumors (NETs) arise from the dispersed endocrine system and can occur in many different areas of the body. Traditionally, NETs have been classified by their anatomic site of origin and are typically highly aggressive. They are most often located in the gastrointestinal tract, pancreas or the lungs (small cell lung carcinoma and large cell neuroendocrine carcinoma) as well as in s, or in the genitourinary tract (bladder, prostate, ovary, cervix, and endometrium). NETs e certain tumors of the gastrointestinal tract and of the pancreatic islet cells, certain thymus and lung tumors, and medullary carcinoma of the llicular cells of the thyroid.

Additional cancers whose growth can be inhibited using the multi-specific binding proteins described herein are pseudo neuroendocrine tumors (pNETs) that share certain pic, phenotypic or mical characteristics with ionally defined neuroendocrine tumors .

In some embodiments, the following cancers, tumors, and other proliferative diseases may be treated with speciﬁc binding proteins of the invention: lung cancer; preferably SCLC, NSCLC or LCNEC; breast; cervical; colon; colorectal; endometrial; head and neck; liver (hepatoblastoma or hepatocellular carcinoma); ovarian; pancreatic; prostate; skin; gastric; testis; thyroid; adrenal; renal; bladder; uterine; esophageal; urothelial cancer; brain tumor; lymphoma; Ewing sarcoma; and other neuroendocrine and small blue round cell tumors.

In a preferred embodiment of the invention the cancer is small cell lung cancer (SCLC) or glioblastoma. _92_ All cancers, , neoplasms, etc., mentioned above which are characterized by their speciﬁc location/origin in the body are meant to include both the primary tumors and the metastatic tumors derived therefrom.

It is possible that a patient is more likely to d to treatment with a binding protein of the invention (as described herein) if that patient has a cancer which is characterized by having a high expression of DLL3. Thus, in some embodiments, the cancer to be treated with the binding ns of the invention is a cancer with high expression of DLL3, e. g., DLL3 expression is higher than the average expression in cancer cells of a population of patients suffering from the same type of a DLL3 expressing cancer.

The binding proteins of the invention may be used in therapeutic regimens in the context of ﬁrst line, second line, or any further line treatments and maintenance treatment.

The binding proteins of the invention may be used for the prevention, short—term or long- term treatment of the above—mentioned diseases, optionally also in combination with radiotherapy, one or more additional therapeutic agents and/or surgery.

In preferred embodiments, the protein of the invention is used for the treatment of cancer in combination with a PD-l antagonist, such as an anti-PD-l antibody or an anti-PDL-l antibody. Preferably said anti-PD-l antibody is selected from the group consisting of pembrolizumab, nivolumab, zumab, PDl-l, PD1-2, PD1-3, PDl-4, and PD1-5 as described herein (as deﬁned by the sequences in Table A below) and in W02017/198741 (incorporated herein by reference). ably said anti-PDL-l dy is selected from the group consisting of atezolizumab, ab and durvalumab. In particular red embodiments, the binding protein of the ion (preferably any one of DLL3#1/CD3#1, DLL3#1/CD3#2, DLL3#1/CD3#3, DLL3#2/CD3#1, DLL3#2/CD3#2, DLL3#2/CD3#3, DLL3#3/CD3#1, DLL3#3/CD3#2, or DLL3#3/CD3#3) is used for the treatment of cancer in combination with PDl-l. In ular preferred ments, the binding protein of the invention (preferably any one of DLL3#l/CD3#1, DLL3#l/CD3#2, DLL3#l/CD3#3, DLL3#2/CD3#1, DLL3#2/CD3#2, DLL3#2/CD3#3, /CD3#1, DLL3#3/CD3#2, or DLL3#3/CD3#3) is used for the treatment of cancer in combination with PDl—2. In particular preferred embodiments, the binding protein of the invention (preferably any one of DLL3#1/CD3#1, /CD3#2, DLL3#1/CD3#3, DLL3#2/CD3#1, DLL3#2/CD3#2, DLL3#2/CD3#3, DLL3#3/CD3#1, DLL3#3/CD3#2, DLL3#3/CD3#3) is used for the treatment of cancer in combination with PDl-3. In ular preferred embodiments, the _93_ binding protein of the invention (preferably any one of DLL3#1/CD3#1, /CD3#2, DLL3#l/CD3#3, DLL3#2/CD3#1, DLL3#2/CD3#2, DLL3#2/CD3#3, DLL3#3/CD3#1, DLL3#3/CD3#2, DLL3#3/CD3#3) is used for the treatment of cancer in combination with PDl-4. In particular preferred embodiments, the binding protein of the ion (preferably any one of DLL3#1/CD3#1, DLL3#1/CD3#2, /CD3#3, DLL3#2/CD3#1, DLL3#2/CD3#2, DLL3#2/CD3#3, DLL3#3/CD3#1, /CD3#2, DLL3#3/CD3#3) is used for the treatment of cancer in combination with PD1-5.

Table A: Amino acid sequences and SEQ ID NOs of heavy chain and light chain sequences of anti—PDl antibodies PD 1 —1, PD 1 —2, PD 1 —3, PD1—4, PD1—5.

SEQ ID Brief Sequence NUInbeTi description of sequence SEQ ID PDl-l HC EVMLVESGGGLVQPGGSLRLSCTASGFTFSASAMSWV N02256 RQAPGKGLEWVAYISGGGGDTYYSSSVKGRFTISRDN AKNSLYLQMNSLRAEDTAVYYCARHSNVNYYAMDY WGQGTLVTVSSASTKGPSVFPLAPCSRSTSESTAALGC LVKDYFPEPVTVSWNSGALTSGVHTFPAVLQSSGLYSL PSSSLGTKTYTCNVDHKPSNTKVDKRVESKYG PPCPPCPAPEFLGGPSVFLFPPKPKDTLMISRTPEVTCVV VDVSQEDPEVQFNWYVDGVEVHNAKTKPREEQFNST YRVVSVLTVLHQDWLNGKEYKCKVSNKGLPSSIEKTIS KAKGQPREPQVYTLPPSQEEMTKNQVSLTCLVKGFYPS DIAVEWESNGQPENNYKTTPPVLDSDGSFFLYSRLTVD KSRWQEGNVFSCSVMHEALHNHYTQKSLSLSLG SEQ ID PDl-l LC EIVLTQSPATLSLSPGERATMSCRASENIDTSGISFMNW N01257 YQQKPGQAPKLLIYVASNQGSGIPARFSGSGSGTDFTLT ISRLEPEDFAVYYCQQSKEVPWTFGQGTKLEIKRTVAA PPSDEQLKSGTASVVCLLNNFYPREAKVQWKV DNALQSGNSQESVTEQDSKDSTYSLSSTLTLSKADYEK HKVYACEVTHQGLSSPVTKSFNRGEC SEQ ID PD 1 —2 HC EVMLVESGGGLVQPGGSLRLSCTASGFTFSASAMSWV _94_ SEQ ID Brief Sequence Number: description of sequence NO:258 RQAPGKGLEWVAYISGGGGDTYYSSSVKGRFTISRDN AKNSLYLQMNSLRAEDTAVYYCARHSNPNYYAMDY WGQGTLVTVSSASTKGPSVFPLAPCSRSTSESTAALGC LVKDYFPEPVTVSWNSGALTSGVHTFPAVLQSSGLYSL SSVVTVPSSSLGTKTYTCNVDHKPSNTKVDKRVESKYG PAPEFLGGPSVFLFPPKPKDTLMISRTPEVTCVV VDVSQEDPEVQFNWYVDGVEVHNAKTKPREEQFNST YRVVSVLTVLHQDWLNGKEYKCKVSNKGLPSSIEKTIS KAKGQPREPQVYTLPPSQEEMTKNQVSLTCLVKGFYPS DIAVEWESNGQPENNYKTTPPVLDSDGSFFLYSRLTVD KSRWQEGNVFSCSVMHEALHNHYTQKSLSLSLG SEQ ID PD1—2 LC SPATLSLSPGERATMSCRASENIDTSGISFMNW NO:259 YQQKPGQAPKLLIYVASNQGSGIPARFSGSGSGTDFTLT ISRLEPEDFAVYYCQQSKEVPWTFGQGTKLEIKRTVAA PSVFIFPPSDEQLKSGTASVVCLLNNFYPREAKVQWKV DNALQSGNSQESVTEQDSKDSTYSLSSTLTLSKADYEK HKVYACEVTHQGLSSPVTKSFNRGEC SEQ ID PD1—3 HC EVMLVESGGGLVQPGGSLRLSCTASGFTFSKSAMSWV NO:260 RQAPGKGLEWVAYISGGGGDTYYSSSVKGRFTISRDN LQMNSLRAEDTAVYYCARHSNVNYYAMDY WGQGTLVTVSSASTKGPSVFPLAPCSRSTSESTAALGC LVKDYFPEPVTVSWNSGALTSGVHTFPAVLQSSGLYSL SSVVTVPSSSLGTKTYTCNVDHKPSNTKVDKRVESKYG PPCPPCPAPEFLGGPSVFLFPPKPKDTLMISRTPEVTCVV VDVSQEDPEVQFNWYVDGVEVHNAKTKPREEQFNST YRVVSVLTVLHQDWLNGKEYKCKVSNKGLPSSIEKTIS KAKGQPREPQVYTLPPSQEEMTKNQVSLTCLVKGFYPS DIAVEWESNGQPENNYKTTPPVLDSDGSFFLYSRLTVD KSRWQEGNVFSCSVMHEALHNHYTQKSLSLSLG _95_ SEQ ID Brief Sequence Number: description of sequence SEQ ID PD1—3 LC EIVLTQSPATLSLSPGERATMSCRASENIDVSGISFMNW NO:261 YQQKPGQAPKLLIYVASNQGSGIPARFSGSGSGTDFTLT ISRLEPEDFAVYYCQQSKEVPWTFGQGTKLEIKRTVAA PSVFIFPPSDEQLKSGTASVVCLLNNFYPREAKVQWKV DNALQSGNSQESVTEQDSKDSTYSLSSTLTLSKADYEK HKVYACEVTHQGLSSPVTKSFNRGEC SEQ ID PD1-4 HC EVMLVESGGGLVQPGGSLRLSCTASGFTFSKSAMSWV NO:262 RQAPGKGLEWVAYISGGGGDTYYSSSVKGRFTISRDN AKNSLYLQMNSLRAEDTAVYYCARHSNVNYYAMDY WGQGTLVTVSSASTKGPSVFPLAPCSRSTSESTAALGC LVKDYFPEPVTVSWNSGALTSGVHTFPAVLQSSGLYSL SSVVTVPSSSLGTKTYTCNVDHKPSNTKVDKRVESKYG PAPEFLGGPSVFLFPPKPKDTLMISRTPEVTCVV VDVSQEDPEVQFNWYVDGVEVHNAKTKPREEQFNST YRVVSVLTVLHQDWLNGKEYKCKVSNKGLPSSIEKTIS KAKGQPREPQVYTLPPSQEEMTKNQVSLTCLVKGFYPS DIAVEWESNGQPENNYKTTPPVLDSDGSFFLYSRLTVD GNVFSCSVMHEALHNHYTQKSLSLSLG SEQ ID PD1—4 LC EIVLTQSPATLSLSPGERATMSCRASENIDVSGISFMNW NO:263 QAPKLLIYVASNQGSGIPARFSGSGSGTDFTLT ISRLEPEDFAVYYCQQSKEVPWTFGQGTKLEIKRTVAA PSVFIFPPSDEQLKSGTASVVCLLNNFYPREAKVQWKV DNALQSGNSQESVTEQDSKDSTYSLSSTLTLSKADYEK HKVYACEVTHQGLSSPVTKSFNRGEC SEQ ID PDl-S HC EVMLVESGGGLVQPGGSLRLSCTASGFTFSKSAMSWV NO:264 RQAPGKGLEWVAYISGGGGDTYYSSSVKGRFTISRDN AKNSLYLQMNSLRAEDTAVYYCARHSNVNYYAMDY WGQGTLVTVSSASTKGPSVFPLAPCSRSTSESTAALGC LVKDYFPEPVTVSWNSGALTSGVHTFPAVLQSSGLYSL _96_ SEQ ID Brief Sequence Number: ption of sequence SSVVTVPSSSLGTKTYTCNVDHKPSNTKVDKRVESKYG PPCPPCPAPEFLGGPSVFLFPPKPKDTLMISRTPEVTCVV VDVSQEDPEVQFNWYVDGVEVHNAKTKPREEQFNST YRVVSVLTVLHQDWLNGKEYKCKVSNKGLPSSIEKTIS KAKGQPREPQVYTLPPSQEEMTKNQVSLTCLVKGFYPS DIAVEWESNGQPENNYKTTPPVLDSDGSFFLYSRLTVD KSRWQEGNVFSCSVMHEALHNHYTQKSLSLSLG SEQ ID PD1-5 LC EIVLTQSPATLSLSPGERATMSCRASENIDVSGISFMNW N02265 QAPKLLIYVASNQGSGIPARFSGSGSGTDFTLT ISRLEPEDFAVYYCQQSKEVPWTFGQGTKLEIKRTVAA PPSDEQLKSGTASVVCLLNNFYPREAKVQWKV DNALQSGNSQESVTEQDSKDSTYSLSSTLTLSKADYEK HKVYACEVTHQGLSSPVTKSFNRGEC According to these preferred embodiments and any other of the aspects of the present invention, antibodies PDl—l, PDl—2, PDl—3, PDl—4 and PDl-5 are antibody molecules as disclosed in /198741, and are deﬁned by the sequences as shown in Table A above.

Accordingly, PDl—l has a heavy chain comprising the amino acid sequence of SEQ ID NO:256 and a light chain comprising the amino acid sequence of SEQ ID NO:257; PDl-2 has a heavy chain comprising the amino acid sequence of SEQ ID NO:258 and a light chain comprising the amino acid sequence of SEQ ID NO:259; PDl-3 has a heavy chain sing the amino acid sequence of SEQ ID NO:260 and a light chain comprising the amino acid sequence of SEQ ID NO:261; PDl-4 has a heavy chain comprising the amino acid sequence of SEQ ID NO:262 and a light chain comprising the amino acid sequence of SEQ ID NO:263; and PDl-5 has a heavy chain comprising the amino acid sequence of SEQ ID NO:264 and a light chain comprising the amino acid sequence of SEQ ID NO:265. _97_ The above also includes the use of the binding proteins of the invention in various methods of treating the above es by administering a therapeutically effective dose to a patient in need thereof, as well as the use of these binding proteins for the manufacture of medicaments for the treatment of such es, as well as pharmaceutical compositions including such binding proteins of the invention, as well as the preparation and/or manufacture of medicaments including such binding proteins of the invention, and the like.

COMBINATIONS WITH OTHER ACTIVE SUBSTANCES OR TREATMENTS A binding n of the invention may be used on its own or in combination with one or more onal therapeutic agents, in particular in combination with a chemotherapeutic agent like DNA damaging agents, a eutically active compound that inhibits angiogenesis, a signal uction pathway inhibitor, an EGFR inhibitor, an immune modulator, an immune checkpoint inhibitor, a c checkpoint inhibitor or a al therapy agent.

The additional eutic agent may be administered simultaneously with, optionally as a component of the same pharmaceutical preparation, or before or after administration of the DLL3/CD3 binding protein.

Cytostatic and/or cytotoxic active substances which may be administered in combination with binding molecules of the invention include, without being restricted thereto, hormones, hormone analogues and antihormones, aromatase inhibitors, LHRH agonists and antagonists, tors of growth factors (growth factors such as for example et derived growth factor (PDGF), fibroblast growth factor (FGF), vascular endothelial growth factor (VEGF), epidermal growth factor (EGF), ne-like growth factors (IGF), human epidermal growth factor (HER, e.g. HER2, HER3, HER4) and hepatocyte growth factor (HGF)), inhibitors are for e )growth factor antibodies, (anti-)growth factor receptor antibodies and tyrosine kinase inhibitors, such as for example cetuximab, geﬁtinib, ib, nintedanib, imatinib, lapatinib, bosutinib and trastuzumab; antimetabolites (e.g. antifolates such as methotrexate, raltitrexed, pyrimidine analogues such as 5—ﬂuorouracil (5 —FU), gemcitabine, irinotecan, doxorubicin, TAS-102, capecitabine and gemcitabine, purine and adenosine analogues such as mercaptopurine, thioguanine, cladribine and pentostatin, bine (ara C), ﬂudarabine); antitumor antibiotics (e. g. anthracyclins); platinum derivatives (e. g. cisplatin, oxaliplatin, —98- carboplatin); tion agents (e. g. estramustin, meclorethamine, melphalan, chlorambucil, busulphan, dacarbazin, cyclophosphamide, ifosfamide, temozolomide, nitrosoureas such as for example carrnustin and lomustin, thiotepa); antimitotic agents (e. g.

Vinca alkaloids such as for example Vinblastine, Vindesin, Vinorelbin and Vincristine; and taxanes such as paclitaxel, docetaxel); angiogenesis inhibitors, including bevacizumab, ramucirumab and aﬂibercept, tubuline inhibitors; DNA synthesis inhibitors, PARP inhibitors, omerase tors (e.g. epipodophyllotoxins such as for example ide and hos, teniposide, amsacrin, topotecan, ecan, mitoxantrone), serine/threonine kinase inhibitors (e. g. PDKl inhibitors, Raf inhibitors, A-Raf inhibitors, B-Raf inhibitors, C—Raf inhibitors, mTOR inhibitors, mTORCl/2 inhibitors, PI3K tors, PI3K0L inhibitors, dual mTOlUPI3K inhibitors, STK33 inhibitors, AKT inhibitors, PLKl inhibitors (such as volasertib), tors of CDKs, including CDK9 inhibitors, Aurora kinase inhibitors), tyrosine kinase inhibitors (e.g. PTKZ/FAK inhibitors), protein protein interaction inhibitors, MEK inhibitors, ERK inhibitors, FLT3 inhibitors, BRD4 inhibitors, IGF-lR inhibitors, Bcl-xL inhibitors, Bcl—2 inhibitors, Bcl—2/Bcl-xL inhibitors, ErbB or inhibitors, BCR-ABL inhibitors, ABL inhibitors, Src inhibitors, rapamycin analogs (e. g. everolimus, temsirolimus, ridaforolimus, sirolimus), en synthesis inhibitors, androgen receptor inhibitors, DNMT tors, HDAC inhibitors, ANGl/2 tors, CYPl7 inhibitors, radiopharmaceuticals, therapeutic agents such as immune checkpoint inhibitors (e. g. CTLA4, PDl, PD-Ll, LAG3, and TIM3 binding molecules / immunoglobulins, such as ipilimumab, nivolumab, pembrolizumab) and various chemotherapeutic agents such as amifostin, anagrelid, clodronat, ﬁlgrastin, interferon, interferon alpha, leucovorin, rituximab, procarbazine, levamisole, mesna, ne, pamidronate and porflmer; proteasome inhibitors (such as Bortezomib); Smac and BH3 mimetics; agents restoring p53 functionality including mdm2-p53 antagonist; inhibitors of the Wnt/beta—catenin signaling y; and/or cyclin—dependent kinase 9 inhibitors.

Particularly preferred are treatments with the binding molecules of the invention in combination with one or more immunotherapeutic , including anti-PD-l and anti- PD—Ll agents and anti LAG3 agents: Exemplary anti-PDl agents include but are not limited to anti-PD-l antibody PDR-OOl, pembrolizumab, nivolumab, pidilizumab and PDl-l, PDl-2, PDl-3, PD1-4 and PDl-5 as sed herein (Table A) and in WO2017/198741. Exemplary anti— PDL-l agents include but are not limited to _99_ atezolizumab, avelumab and durvalumab. In preferred embodiments, the binding molecule of the invention (preferably any one of DLL3#1/CD3#1, DLL3#l/CD3#2, DLL3#1/CD3#3, /CD3#1, DLL3#2/CD3#2, DLL3#2/CD3#3, /CD3#1, DLL3#3/CD3#2, DLL3#3/CD3#3) is combined with PDl—l. In preferred embodiments, the binding molecule of the invention (preferably any one of DLL3#1/CD3#1, /CD3#2, DLL3#1/CD3#3, DLL3#2/CD3#1, DLL3#2/CD3#2, DLL3#2/CD3#3, DLL3#3/CD3#1, DLL3#3/CD3#2, DLL3#3/CD3#3) is combined with PD1—2. In preferred embodiments, the binding molecule of the invention (preferably any one of DLL3#1/CD3#1, DLL3#1/CD3#2, DLL3#l/CD3#3, DLL3#2/CD3#1, DLL3#2/CD3#2, DLL3#2/CD3#3, DLL3#3/CD3#1, DLL3#3/CD3#2, DLL3#3/CD3#3) is combined with PDl -3. In preferred embodiments, the binding molecule of the invention (preferably any one ofDLL3#1/CD3#1, DLL3#1/CD3#2, DLL3#1/CD3#3, DLL3#2/CD3#1, DLL3#2/CD3#2, DLL3#2/CD3#3, DLL3#3/CD3#1, DLL3#3/CD3#2, DLL3#3/CD3#3) is combined with PD1—4. In preferred embodiments, the binding molecule of the invention (preferably any one of DLL3#1/CD3#1, DLL3#l/CD3#2, DLL3#1/CD3#3, DLL3#2/CD3#1, DLL3#2/CD3#2, DLL3#2/CD3#3, DLL3#3/CD3#1, DLL3#3/CD3#2, DLL3#3/CD3#3) is ed with PD1—5.

In n embodiments, the additional eutic agent may be a further therapeutic agent, such as modulators of: TIM-1, TIM-3, TIM-4, PD-L2, LAG3, CTLA—4, Galectin 9, Galectin—1, CD69, CD113, GPR56, CD48, GARP, CAECAM—l, BTLA, TIGIT, CD160, LAIRl, 2B4, CEACAM, CD39, TGFB, IL—10, Fas ligand, ICOS, B7 family (B7—1, B7—2, B7-H1 ), B7—DC (PD—L2), B7—H2 (ICOS—L), B7—H3, B7— H4, B7—H5 ), or .

In some embodiments, the onal immunotherapeutic agent is a member of the TNF family of molecules that bind to cognate TNF receptor family members, which include CD40 and CD40L, OX—40, OX—40L, CD70, CD27L, CD30, CD30L, 4—1BBL, CD137, GITR, TRAIL/ApoZ-L, TRAILRl/DR4, TRAILR2/DR5, TRAILR3, TRAILR4, OPG, RANK, RANKL, TWEAKIVFn14, TWEAK, BAFFR, EDAR, XEDAR, TACI, APRIL, BCMA, LIGHT, DcR3, HVEM, VEGI/TLIA, DR3, EDAR, EDAl, XEDAR, EDA2, TNFRl, Lymphotoxin (x/TNFB, TNFR2, TNFOt, LTBR, Lymphotoxin (x152, FAS, FASL, RELT, DR6, TROY, NGFR.

In some embodiments, the additional therapeutic agent is a SMAC mimetic. SMAC —100— mimetics are compounds which bind to Inhibitor of sis Proteins (IAPs) and have immune tory function and mediate the induction of systemic cytokines (e. g. IL—6, TNFO. etc.) and ines (e.g. MCP-l) when administered to animals or humans. In some embodiments, the SMAC mimetic is (i) LCL161, i.e. compound A in example 1 of (page 28/29; [122]), or a pharmaceutically acceptable salt thereof; (ii) the SMAC mimetic known as Debio-l 143, or a ceutically acceptable salt f; (iii) the SMAC mimetic known as pant, or a pharmaceutically acceptable salt thereof; (iv) the SMAC mimetic known as ASTX-660, or a ceutically acceptable salt thereof; or (V) the SMAC mimetic known as CUDC-427, or a pharmaceutically acceptable salt thereof.

In some embodiments, the additional immunotherapeutic agent is selected from (i) antagonists of cytokines that inhibit T cell activation (e.g., IL-6, IL-10, TGF-B, VEGF; "immunosuppressive cytokines") and/or (ii) agonists of cytokines that stimulate T cell activation and/or cytokines such as IL2, for stimulating an immune response, e.g., for treating proliferative diseases, such as cancer.

In some embodiments, the additional immunotherapeutic agent is an agonist of a protein that ates T cell activation, such as CD28, GITRL, OX40L, CD27, and CD28H.

In some ments, the additional eutic agent is an oncolytic virus including but not limited to an oncolytic virus derived from vaccinia Virus, adenovirus, (AdV), herpes simplex virus (HSVl or HSV2), reovirus, myxoma virus (MYXV), poliovirus, vesicular stomatitis Virus (VSV), Maraba Virus, varicella Virus, measles Virus (MV), or Newcastle disease virus (NDV).

DLAGNOSTIC USES The anti-DLL3 antibody molecules of the invention are useful in diagnostic and prognostic methods and can be employed for labelling, localizing, or fying cells or tissues expressing DLL3 (e.g. in ELISA assays, FACS analysis, immunohistology or the like) by attaching a dye, a drug or another molecule with binding specificity for a ent antigen.

For example, a detectable label can be conjugated to the anti-DLL3 antibody molecule of the invention or a secondary reagent binding ically to the antibody molecule of the invention and being ated with a detectable label (e.g. a secondary antibody) can be used in such diagnostic methods. In some embodiments, DLL3 speciﬁc antibodies —lOl— speciﬁcally bind to DLL3 expressed in cells, either in the cytoplasm and/or at the cell surface, and are used for localizing and/or identifying such cells. In some embodiments, the DLL3 c antibodies provided herein are used for identifying cells or tissues expressing DLL3 (e. g. tumor cells).

Antibodies of the invention are selected such that they have a high level of epitope binding city and high binding afﬁnity to the DLL3 polypeptide. In l, the higher the binding y of an antibody, the more stringent wash ions can be performed in an immunoassay to remove nonspeciﬁcally bound material without removing the target polypeptide.

Thus, further ed herein are methods of detecting DLL3 in a biological sample.

Speciﬁcally, in one aspect the ion provides a method of detecting DLL3 in a sample (e.g. a tissue sample, a tumor tissue sample) comprising (a) contacting the sample with an antibody of the invention; and (b) detecting anti—DLL3 antibody bound to DLL3.

A biological sample can be obtained/isolated from any tissue (including biopsies), cell or body ﬂuid of a t. In a preferred embodiment, the sample is a tissue sample, preferably a tumor tissue sample. In a speciﬁc embodiment, the sample is a ﬁxed (tumor) tissue sample, preferably a formalin-ﬁxed, parafﬁn—embedded (FFPE) tissue sample, more ably a formalin—ﬁxed, parafﬁn—embedded (FFPE) tumor tissue sample.

The samples are typically divided into several portions and afﬁxed to a medium for microscopic analysis, such as a microscope slide. Where the sample is a tissue sample, the l portions may be tissue sections. In some embodiments, serial sections are taken from FFPE tissue samples. In some ments, serial sections are taken from a plurality of different sites of a FFPE block, which can be done to capture both intra—section heterogeneity and intra-block heterogeneity. In some embodiments, serial sections are taken from a plurality of different biopsy samples taken from different locations in the same tumor, which can be done to capture both section heterogeneity and intra—tumor heterogeneity. Typically, tissue samples ﬁrst undergo deparafﬁnization, antigen retrieval (e.g. with Proteinase K) before antigen ion. The tissue sample can be obtained/isolated from a variety of tissues (e. g. from a bi0psy), speciﬁcally tissues obtained from tumors including but not limit to SCLC, LCNEC, glioma, glioblastoma, —102— melanoma, or other neuroendocrine tumors (e.g., neuroendocrine prostate cancer or neuroendocrine pancreatic cancer, small cell bladder cancer).

Further exemplary tumors from which a e) sample can be obtained/isolated include lung cancer; preferably SCLC, NSCLC or LCNEC; ; cervical; colon; colorectal; endometrial; head and neck; liver (hepatoblastoma or hepatocellular oma); ovarian; pancreatic; prostate; skin; gastric; testis; thyroid; adrenal; renal; bladder; uterine; esophageal; urothelial cancer; brain tumor; lymphoma; Ewing sarcoma; and other neuroendocrine (including pseudo—endocrine tumors) and small blue round cell tumors.

Exemplary methods of DLL3 detection in a sample are immunocytochemistry (ICC), immunohistochemistry (IHC), Western Blotting, Flow cytometry and/or ELISA.

In a preferred embodiment, DLL3 is detected using IHC.

In a c aspect of the ion, the method of detecting DLL3 in a tissue sample comprises a) contacting a tissue sample (e. g. a tumor tissue such as SCLC, astoma or ndocrine tumors), preferably said tissue sample is a ﬁxed tissue sample (e. g., formalin-fixed and paraffin embedded), with an antibody of the invention (e. g.

DLL3#1 or DLL3#5) b) permitting formation of antibody-antigen complexes in the sample, and c) detecting the anti-DLL3 antibody bound to DLL3.

In some embodiments of the methods of the invention, the anti-DLL3 antibody is detected by a able signal, preferably, the detectable signal is generated by a detectable label.

In a red embodiment of the method of the invention, the detectable signal is detected in an IHC assay. In some embodiments, the detectable signal is detected in an ELISA assay, by ﬂow cytometry or by Western blot. In some embodiments, the methods described herein can be used to detect DLL3 present on the cell surface of a cell (e. g. a tumor cell), for example using ﬂow cytometry. In some embodiments, the methods described herein can be used to detect the presence of DLL3 in a tissue sample (e. g. tumor tissue sample) using ELISA, Western blot or IHC.

In some embodiments, the able label is directly conjugated to the anti—DLL3 antibody and thus is deposited on the sample upon binding of the anti-DLL3 antibody to DLL3, this is generally referred to as a direct labeling method). Direct ng methods —103— are often more directly quantiﬁable, but often suffer from a lack of sensitivity. In other embodiments, tion of the detectable label is effected by the use of a secondary detection reagent binding speciﬁcally to the antibody molecule of the invention and being conjugated with a detectable label (e. g. a secondary antibody), this is generally referred to as an indirect labeling method. In some embodiments, the speciﬁc secondary detection reagent may be a species—speciﬁc secondary antibody, an apten antibody g to a hapten—conjugated anti-DLL3 antibody, or a biotin—binding protein bound to a biotinylated anti-DLL3 antibody).

Particular examples of detectable labels that can be conjugated either to the anti-DLL3 antibody of the invention or to a secondary detection reagent, include chromogenic, ﬂuorescent, orescent, luminescent and radioactive molecules and materials, catalysts (such as enzymes) that convert one substance into another substance to provide a detectable difference (such as by converting a colorless substance into a colored substance or vice versa, or by producing a precipitate or increasing sample turbidity), haptens that can be detected through antibody-hapten binding interactions using additional detectably labeled dy conjugates, and paramagnetic and magnetic molecules or materials.

For example, the detectable label can be an enzyme such as adish peroxidase (HRP), alkaline atase (AP), acid phosphatase, glucose oxidase, B—galactosidase, B— glucuronidase, and B—lactamase; a ﬂuorphore such as ceins, phores, ins, BODIPY dyes, resoruﬁns, and rhodamines; nanoparticles such as quantum dots (US. Pat. Nos. 6,815,064, 6,682,596 and 6,649,138); metal chelates such as DOTA and DPTA chelates of radioactive or paramagnetic metal ions like Gd <3 +>; and liposomes, for example, liposomes containing trapped ﬂuorescent molecules.

Where the detectable label includes an enzyme, a detectable substrate such as a chromogen, a enic compound, or a luminogenic compound can be used in combination with the enzyme to generate a detectable signal. Particular examples of genic compounds include diaminobenzidine (DAB), 4-nitrophenylphospate , fast red, bromochloroindolyl ate (BCIP), nitro blue tetrazolium (NBT), BCIP/NBT, fast red, AP Orange, AP blue, tetramethylbenzidine (TMB), 2,2'— azino—di-[3- ethylbenzothiazoline sulphonate] (ABTS), o-dianisidine, 4- chloronaphthol (4-CN), nitrophenyl- -D-galactopyranoside (ONPG), o- phenylenediamine (OPD), 5—bromo—4- indoly - -galactopyranoside ), umbelliferyl- —D-galactopyranoside (MU-Gal), p-nitrophenyl- a-D-galactopyranoside (PNP), 5—bromo-4—chloro—3-indolyl- —D- — l 04— WO 34220 glucuronide (X—Gluc), 3-aminoethyl ol (AEC), fuchsin, trotetrazolium (INT), tetrazolium blue and tetrazolium violet.

In some examples, the detectable moiety is a hore, which belongs to several common chemical classes including coumarins, ﬂuoresceins (or ﬂuorescein derivatives and analogs), rhodamines, resoruﬁns, luminophores and cyanines.

In other ments, the detectable moiety is a molecule detectable via brightfleld microscopy, such as dyes including diaminobenzidine (DAB), 4-(dimethylamino) azobenzene—4'- sulfonamide (DABSYL), tetramethylrhodamine (DISCOVERY Purple), N,N'—biscarboxypentyl—5,5'—disulfonato—indo— dicarbocyanine (Cy5), and Rhodamine 110 (Rhodamine).

Haptens are small molecules that are speciﬁcally bound by antibodies, although by themselves they will not elicit an immune response in an animal and must ﬁrst be attached to a larger carrier molecule such as a protein to generate an immune response. Examples of s include di—nitrophenyl, biotin, digoxigenin, and ﬂuorescein.

Further ed herein in one aspect are methods of diagnosing or fying a tumor as DLL3 expressing tumor (expressing DLL3 in the cytoplasm and/or the cell surface), comprising detecting DLL3 in a sample (e.g. a tumor tissue sample) of the subject using the anti-DLL3 antibodies (e. g. DLL3#1 or DLL3#5) of the invention (e.g. using IHC on a tumor tissue sample). In a further aspect, provided herein are methods of selecting a DLL3 targeted therapy for a subject, comprising detecting DLL3 in a sample (e.g. a tumor tissue sample) of the t using the anti-DLL3 dies (e.g. DLL3#1 or DLL3#5) of the invention (e.g. using IHC on a tumor tissue sample). In a r aspect, provided herein are methods of monitoring the therapeutic effect (e.g. of a DLL3 ted therapy) in a subject, comprising ing DLL3 in a sample (e.g. a tumor tissue sample) of the subject using the anti-DLL3 antibodies (e.g. DLL3#1 or DLL3#5) of the invention (e.g. using IHC on a tumor tissue sample).

The anti-DLL3 antibody molecules of the invention can also be used for targeting cells. In some embodiments, the DLL3 specific antibodies provided herein are used for delivering a drug or cytotoxic agent to a target cell (e.g. a tumor cell expressing DLL3) by attaching such drug or cytotoxic agent to said DLL3 antibody, thereby, for example, killing said target cell. —105— KITS The ion also encompasses kits comprising at least a multi-specific binding protein of the invention (e.g., any one of DLL3#l/CD3#1, DLL3#2/CD3#1, DLL3#3/CD3#1, /CD3#1, DLL3#5/CD3#1, DLL3#6/CD3#1, DLL3#7/CD3#1, DLL3#8/CD3#1, DLL3#9/CD3#1, DLL3#10/CD3#1, DLL3#11/CD3#1, DLL3#12/CD3#1, DLL3#13/CD3#1, DLL3#14/CD3#1, DLL3#15/CD3#1, 6/CD3#1, DLL3#17/CD3#1, DLL3#18/CD3#1, DLL3#1/CD3#2, DLL3#2/CD3#2, DLL3#3/CD3#2, DLL3#4/CD3#2, DLL3#5/CD3#2, DLL3#6/CD3#2, DLL3#7/CD3#2, DLL3#8/CD3#2, DLL3#9/CD3#2, DLL3#10/CD3#2, DLL3#11/CD3#2, DLL3#12/CD3#2, DLL3#13/CD3#2, DLL3#14/CD3#2, DLL3#15/CD3#2, DLL3#16/CD3#2, DLL3#17/CD3#2, DLL3#18/CD3#2, /CD3#3, /CD3#3, DLL3#3/CD3#3, DLL3#4/CD3#3, DLL3#5/CD3#3 and DLL3#6/CD3#3, DLL3#7/CD3#3, DLL3#8/CD3#3, DLL3#9/CD3#3, DLL3#10/CD3#3, DLL3#11/CD3#3, 2/CD3#3, DLL3#l3/CD3#3, 4/CD3#3, DLL3#15/CD3#3, DLL3#16/CD3#3, DLL3#17/CD3#3, 8/CD3#3, preferably any one of DLL3#1/CD3#1, DLL3#1/CD3#2, DLL3#1/CD3#3, /CD3#1, DLL3#2/CD3#2, DLL3#2/CD3#3, DLL3#3/CD3#1, DLL3#3/CD3#2, DLL3#3/CD3#3) and optionally one or more other components selected from the group consisting of other drugs used for the treatment of the diseases and disorders as described above.

In one embodiment, the kit es a composition containing an effective amount of a binding protein of the invention in unit dosage form.

The invention also encompasses kits comprising at least a multi-specific binding protein of the invention, and one or more other components ed from the group consisting of other drugs used for the treatment of the diseases and disorders as described above.

In one embodiment, the kit includes a composition containing an effective amount of a multi-speciﬁc binding protein of the invention in unit dosage form (preferably any one of DLL3#l/CD3#1, DLL3#1/CD3#2, DLL3#l/CD3#3, DLL3#2/CD3#1, DLL3#2/CD3#2, DLL3#2/CD3#3, DLL3#3/CD3#1, DLL3#3/CD3#2, DLL3#3/CD3#3). In a further embodiment the kit includes both a composition containing an effective amount of a multi- c binding protein of the invention in unit dosage form (preferably any one of DLL3#l/CD3#1, DLL3#l/CD3#2, DLL3#1/CD3#3, /CD3#1, DLL3#2/CD3#2, DLL3#2/CD3#3, DLL3#3/CD3#1, DLL3#3/CD3#2, DLL3#3/CD3#3) and a composition —106— containing an effective amount of a PD-l antagonist in unit dosage form, such as an anti PD-l antibody, most preferably PDl-l and PDl-5 as described , PD1-2, PD1-3, PDl-4, herein (e.g. Table A) and in W02017/l9874l.

In some embodiments, the kit comprises a sterile container which contains such a composition; such containers can be boxes, es, bottles, vials, tubes, bags, pouches, blister-packs, or other le container forms known in the art. Such containers can be made of plastic, glass, laminated paper, metal foil, or other materials suitable for holding medicaments. Further, the kit may comprise the pharmaceutical composition in a first container with the binding protein of the invention in lyophilized form and a second ner with a pharmaceutically able diluent (e.g., sterile water) for injection. The pharmaceutically able diluent can be used for reconstitution or dilution of the binding protein.

If desired, a specific binding n of the invention, is provided together with instructions for administering the multi-specific binding proteins to a subject having cancer. The instructions will generally e information about the use of the composition for the treatment or prevention of a . In other embodiments, the instructions include at least one of the following: description of the therapeutic agent; dosage schedule and stration for treatment or prevention of cancer or symptoms thereof; precautions; warnings; indications; counter—indications; overdosage information; adverse reactions; animal cology; clinical studies; and/or references. The instructions may be printed directly on the container (when present), or as a label d to the container, or as a separate sheet, pamphlet, card, or folder supplied in or with the container.

As set forth , the present disclosure further provides diagnostic methods for determining the expression of DLL3 in a biological sample. In one aspect, the present disclosure provides kits for performing these methods as well as instructions for carrying out the s of the present disclosure such as collecting tissue and/or performing the detection assay, and/or analyzing the results.

The kit comprises, or atively consists essentially of, or yet further consists of, a DLL3 antibody composition of the invention (e. g., monoclonal antibodies) and instructions for use. The kits are useful for detecting the presence of DLL3 polypeptides in a biological sample e. g., any body ﬂuid or biopsy samples of body tissue. In some embodiments, the kit —107— can comprise: one or more DLL3 antibodies e of binding a DLL3 polypeptide in a sample (e.g., anti-DLL3 dy DLL3#l or DLL3#5); means for ing the DLL3 polypeptide in the sample; and optionally means for ing the signal/amount of the DLL3 ptide in the sample with a control. One or more of the anti-DLL3 antibodies of the invention may be labeled. The kit components, (e. g., reagents) can be packaged in a suitable container. The kit can further se instructions for using the kit to detect the DLL3 polypeptides. In certain embodiments, the kit comprises an anti-DLL3 antibody of the invention (a first antibody) and a second, different antibody which binds to either the DLL3 polypeptide or the ﬁrst antibody and is conjugated to a detectable label.

The kit can also comprise, e.g., a buffering agent, a preservative or a protein—stabilizing agent. The kit can further comprise components necessary for detecting the detectable— label, e.g., an enzyme or a substrate. The kit can also contain a control sample or a series of control samples, which can be assayed and compared to the test sample. Each component of the kit can be enclosed within an dual container and all of the various containers can be within a single package, along with instructions for reting the results of the assays performed using the kit. The kits of the present disclosure may contain a written product on or in the kit container. The written product describes how to use the ts contained in the kit.

As amenable, these suggested kit ents may be packaged in a manner customary for use by those of skill in the art. For example, these suggested kit components may be provided in solution or as a liquid dispersion or the like.

EXAMPLES The following examples rate the invention. These examples should not be construed as to limit the scope of this invention.

Example 1: Design and construction of CD3/DLL3 binding proteins The present inventors have developed multi—specif1c binding proteins that bind DLL3 and CD3 and that induce T-cell activation leading to lysis of DLL3 -expressing tumor cells. The molecular design used has an IgG dy scaffold and an IgG-like structure. It features knob-in-hole technology in the Fe for hetero-dimerization of the Knob (anti-DLL3) and —108— Hole (anti—CD3) arms. In addition, the binding n has ﬂexible peptide sequences between the light and the corresponding heavy chain in each arm. Thus, the binding protein comprises two arms, one binding to CD3, the other one binding to DLL3, each arm comprising a single chain Fab and an Fc region.

Preferably the binding molecule is bispeciﬁc and bivalent (monovalent for each of the two targets).

Preparation of binding domains that recognize DLL3 and CD3 using high throughput V gene ry from hybridomas and cultured single B cells.

To obtain anti-DLL3 binders, hybridomas or single B cells derived from DLL3 immunized wild-type and AlivaMab humanized mice (Ablexis, San Francisco, CA, USA: ab transgenic mouse rm with human immunoglobulin loci) were cultured in Vitro.

Supernatants were screened for reactivity t recombinant human DLL3, by AlphaLISA (PerkinElmer, Waltham, MA, USA), and against SHP77 cells (ATCC®, CRL— 2195TM) expressing human DLL3, by Flow Cytometry.

Immunoglobulin (Ig) VH and VL genes were then ampliﬁed from fied positive . To isolate RNA from hybridomas, about 2x106 cells from single clones were pelleted and used as source al. For single B cells, 100 to 500 cells expanded from singularly ed B cells were used as source material. RNA was isolated using RNeasy Plus (Qiagen, Hilden, Germany). cDNA was then sized using Smarter cDNA synthesis kit (Clontech, Mountain View, CA) according to manufacturer’s instructions. To facilitate cDNA synthesis, oligodT was used to prime reverse transcription of all messenger RNAs ed by “5’ capping” with a Smarter IIA oligonucleotide.

Subsequent ampliﬁcation of the VH and VL fragments was performed using a 2-step PCR amplification using 5’ primers targeting the Smarter IIA cap and 3 ’ primers targeting consensus regions in CHl. Brieﬂy, each Soul PCR reaction consists of 20uM of forward and reverse primer mixes, 25 ul of PrimeStar Max DNA polymerase premix (Clontech), 2ul riﬁed cDNA, and 21 ul of double-distilled H20. The cycling program starts at 94 0C for 3 min, followed by 35 cycles (94°C for 30 Sec, 50°C for l min, 68 °C for l min), and ends at 72 °C for 7 min. The second round PCR was performed with VL and VH 2nd round primers containing 15bp complementary extensions that “overlap” respective regions in their respective pTT5 mother vector (VH and VL). Second round PCR was —109— performed with the following program: 94 0C for 3 min; 35 cycles (94 °C for 30 Sec, 50°C for l min, 68 0C for l min), and ends at 72 °C for 7 min.

In—Fusion® HD Cloning Kit (Clontech, USA.) was used for directional cloning of VL gene into a pTTS hngK vector and VH gene into a pTTS hngGlKO vector. To facilitate In—Fusion® HD Cloning, PCR products were puriﬁed and treated with g Enhancer before In—Fusion HD Cloning. Cloning and transformation were performed according to manufacturer’s protocol (Clontech, U.S.A.). Mini—prep DNAs were subjected to Sanger sequencing to confirm that te V-gene fragments were obtained.

Using this methodology, pairs of Ig VH and VL genes encoding binding s with specificity for DLL3 were prepared. inant antibodies were produced by transient transfection of CHO-E37 cells with the corresponding heavy and light chain-encoding Conﬁrmatory screening of recombinant antibodies Supernatants containing expressed recombinant antibodies were assayed by ﬂow cytometry for binding to cell lines sing human or cyno DLL3. Briefly, cells were incubated with recombinant supernatants, washed, and bound mAbs from the supernatants were detected with anti—human—IgG—APC (Jackson ImmunoResearch 109—136—098).

Signal-to-background ratios (S/B) were calculated by dividing the median ﬂuorescence intensity (MFI) of the sample by that of e control (variable regions against an unrelated protein and different constant region backbones). e Plasmon Resonance (SPR) on Biacore 400 was performed on recombinant supernatants. Brieﬂy, the non-optimized IgGs in the HTP supernatants were captured via Protein A/G onto the sensor surface for 60 sec at 10 ul/min. Binding of 100 nM human DLL3 to the captured IgGs was monitored for 180 sec of association at 30 ul/min, ed by 120 sec of dissociation in the HBS-EP buffer. Regeneration of the Protein A/G surface was performed with Glycine pH 2.1 in between each binding cycle. The following materials were used in this assay: Protein t: recombinantly expressed human DLL3.

System running buffer: HBS-EP (10 mM HEPES pH 7.4, 150 mM NaCl, 3 mM EDTA, and 0.005% V/V rbate P20). ing reagent: Protein A/G G (ThermoFisher Scientific, Waltham, MA, USA), with speciﬁcity towards all human IgG isotypes. —llO— Clones of interest (with Kd<300pM) were selected for peciﬁc ting.

Multispeciﬁc binding proteins were generated and further evaluated in mechanistic and functional screening (such as cell binding, cytotoxicity and T cell activation as described below).

Humanization of DLL3 and CD3 s Sequences of DLL3 binders as described above as well as CD3 binders described in the ture (Pessano et al., EMBO J. 1985 Feb; 4(2): 337-44; Salmerén A et al., J l. 1991 Nov 1;147(9):3047—52) were humanized and/or optimized. Sequence optimization/humanization of antibodies is a methodology to engineer antibodies raised in non-human species (against a speciﬁc antigen/ epitope) for use as therapeutics that resemble antibodies produced in humans and thereby eliminating potential adverse effects such as genicity while retaining the specificity. The sequence zation/humanization approach utilized here was as described by Singh et a1, 2015 (Singh S et al., mAbs 2015: 7(4):778-91). In brief, closely matching human germlines were ﬁed in silica, and optimization/humanization variants were evaluated using a phage screening method. Final lead candidate sequences were selected based on g, percent human score and Epivax (in silico predictive tool for potential genicity) score.

Construction of bispeciﬁc proteins binding DLL3 and CD3 The variable regions of the DLL3 and CD3 s were cloned into the expression vector pTT5 (National Research Council, Canada), using common molecular biology techniques to form bispecific binding proteins with one DLL3 specific binding unit comprising a single chain Fab binding to DLL3 and an Fc region (such binding unit also referred to herein as “DLL3 arm” or “DLL3 chain”) and a CD3 speciﬁc binding unit comprising a single chain Fab binding to CD3 and an Fc region (such binding unit also referred to herein as “CD3 arm” or “CD3 chain. The Fc regions of the DLL3 and CD3 arms include either “Knob” or “Hole” mutations (Atwell et a1, JMB, 1997, 270, 26—35) and the respective chains are referred to as Knob or Hole chains. For fragment DNA ly, a Gibson—assembly and NEBuilder HiFi DNA assembly approaches were used, following manufacturer’s protocols (New England Biolabs, Ipswich, MA, USA). DNA mini-preps were sequenced. —lll— Each expression vector contains eukaryotic promoter elements for the chain—encoding gene (DLL3 or CD3 arm/chain), i.e., the gene encoding the signal sequence and the light and heavy chain, an expression cassette for a prokaryotic selection marker gene such as ampicillin, and an origin of replication. These DNA plasmids were propagated in ampicillin resistant E. coli colonies and cultures and were puriﬁed.

Example 2: Expression and puriﬁcation of bispeciﬁc, bivalent binding proteins binding DLL3 and CD3 Bispeciﬁc molecules binding DLL3 and CD3 were produced by transient transfection of CHO—E cells with the pTTS vectors carrying the DLL3/CD3—chain—encoding genes.

Brieﬂy, transfected CHO—E cells g in suspension in serum—free media were cultivated in shake ﬂasks under agitation at 140 rpm, 37°C and 5% C02 and kept at ions of exponential growth. On the day of transfection, cells were chemically transfected with Knob-chain plasmid and Hole-chain plasmid in 1:3 mass ratio. They were then seeded at l to 2x10A6 cells/ml in l L of Gibco® FreeStyleTM CHO expression medium echnologies, NY, US). Cells were then incubated under orbital shaking for days with one-time feeding of 200 ml commercial feed solution to allow expression of the proteins. Antibody titers in the cell culture supernatants were determined using an Octet® instrument (Pall ForteBio, CA, US) and protA biosensor tips according to manufacturer’s ctions.

Recombinant DLL3/CD3 g proteins were puriﬁed from culture supernatant in a two- step process: ﬁrst by n A afﬁnity chromatography using MabSelectTM column (GE Healthcare); second, by Cation exchange chromatography using a Poros 50 HS column (Applied tems, Carlsbad, CA, USA). The two—step puriﬁed material was stored in ﬁnal buffer of 50mM Sodium Acetate and IOOmM NaCl, pH 5.0 Purity and degree of heterogeneity of the s were assessed by analytical size-exclusion chromatography, mass spectrometry and analytical ultracentrifugation. Samples that were advanced for functional testing comprised two-step puriﬁed material, with about 99% monomer t.

Table 1: Amino acid ces and SEQ ID NOs of CDRs, VH, VL, scFabs, DLL3—arm and m sequences of the ns/antibody constructs described herein: —112— 2019/064942 SEQ ID Brief ce Number description ofsequence SEQID DLLﬁﬂ RASQSVSSNFLV h021 LCCDR1 SEQID DLLﬁﬂ GASTRAS BOQ LCCDRZ SEQ ID DLL3#1 QQYGDSPYT ROS LCCDR3 SEQ ID DLL3#1 GNTFTNYYMH BOA HCCDR1 SEQID DLLﬁﬂ HDPSVGSKSYAQKFLG h05 HCCDRZ SEQID DLLﬁﬂ AGKRFGESYFDY hOﬁ HCCDR3 SEQ ID DLL3#2 RASQGISNYLA hOﬂ LCCDR1 SEQ ID DLL3#2 AASSLQS BO£ LCCDRZ SEQ ID DLL3#2 LQHNSSPYT h09 LCCDR3 SEQ ID DLL3#2 GYTFTSYYMH ROAO HCCDR1 SEQID DLLﬁﬂ HNPSGGSTSYAQKFQG ROﬂl HCCDRZ SEQ ID DLL3#2 GEAVGGNYYYYGMDV ROAZ HCCDR3 SEQ ID DLL3#3 RASQGISNYLV R0ﬂ3 LCCDR1 SEQID DLLﬁB AVSSLYS R0ﬂ4 LCCDR2 SEQID DLLﬁB LQHDSYPYT ROAS LCCDR3 SEQID DLLﬁB GYTFTSYYVH R0ﬂ6 HCCDR1 SEQID DLLEB HNPGGGTTSYAQKFLG —113— SEQ ID Brief Sequence Number description of sequence N0:17 HCCDR2 SEQ ID DLL3#3 NYFYYGMDV N0:18 HCCDR3 SEQ ID DLL3#4 RASKSVSSFGYSFMH N0:19 LCCDRl SEQ ID DLL3#4 LASNLES N020 LCCDR2 SEQ ID DLL3#4 QHSRELPWT N021 LCCDR3 SEQ ID DLL3#4 VYTFTSYFMY N022 HCCDRl SEQ ID DLL3#4 EISPTNGNSNLNERFKN N023 HCCDR2 SEQ ID DLL3#4 GGDGYLDY N024 HCCDR3 SEQ ID DLL3#5 QASQDISNYLN N025 LCCDRl SEQ ID DLL3#5 DASNLET N026 LCCDR2 SEQ ID DLL3#5 QQYDNLPTWT N027 LCCDR3 SEQ ID DLL3#5 YYIH N028 HCCDRl SEQ ID DLL3#5 WINPNSGGTNYAQKFQG N029 HCCDR2 SEQ ID DLL3#5 GWDY N0230 HCCDR3 SEQ ID DLL3#6 RASQDISNYFA N0231 LCCDRI SEQ ID DLL3#6 AASTLQS N0232 LCCDR2 SEQ ID DLL3#6 QQLNSYPYT N0233 LCCDR3 —114— SEQ ID Brief Sequence Number ption ofsequence SEQ ID DLL3#6 GGSISSYFWS NO:34 HCCDR1 SEQ ID DLL3#6 STNYNPSLNS NO:35 HCCDRZ SEQ ID DLL3#6 RGDWGGFDI NO:36 HCCDR3 SEQ ID DLL3#1 EIVLTQSPGTLSLSPGERATLSCRASQSVSSNFLVWYQQKPG NO:37 VL QAPRPLIYGASTRASGIPDRFSGSGSGADFTLTISRLEPEDFAL YYCQQYGDSPYTFGQGTTLEIK SEQ ID DLL3#1 QVQLVQSGAEVKKPGASVKVSCKASGNTFTNYYMHWVRQ N0238 VH APGPGLEWMGIIDPSVGSKSYAQKFLGRVTIARDTSTSTVFL DLYSLRSEDTAVYFCARAGKRFGESYFDYWGQGTLVTVSS SEQ ID DLL3#2 DIQMTQSPSAMSASVGDRVTITCRASQGISNYLAWFQQKPG NO:39 VL KVPEPLIYAASSLQSGVPSRFSGSGSVTEFTLTISSLQPEDFAT NSSPYTFGQGTKLEIK SEQ ID DLL3#2 QVQLVQSGAEVKKPGASVKVSCKASGYTFTSYYMHWVRQ NO:40 VH APGQGLEWMGIINPSGGSTSYAQKFQGRVTMTRDTSTSTW MELSSLRSEDTAVYYCARGEAVGGNYYYYGMDVWGQGTT VTVSS SEQ ID DLL3#3 DIQMTQSPSAMSASVGDRVTITCRASQGISNYLVWFQQKPG NO:41 VL KAPKRLIYAVSSLYSGVPSRFSGSGSGTEFTLTISSLQPEDFAT YYCLQHDSYPYTFGQGTKLEIK SEQ ID DLL3#3 QVQLVQSGAEVKKPGASVKVSCKASGYTFTSYYVHWVRQ NO:42 VH APGQGLEWMVIINPGGGTTSYAQKFLGRVTMTRDTSTNTVY MELKSLRSEDTAVYYCARGEAVTGNYFYYGMDVWGQGTT VTVSS SEQ ID DLL3#4 DIVLTQSPASLAVSLGQRATISCRASKSVSSFGYSFMHWYQQ NO:43 VL KPGQPPKLLIYLASNLESGVPARFSGSGSGTDFTLNIHPVEEE DAATYYCQHSRELPWTFGGGTKLEIK SEQ ID DLL3#4 QVQLQQSGTELVKPGASVKLSCKASVYTFTSYFMYWVKQR NO:44 VH PGHGLEWIGEISPTNGNSNLNERFKNKATLTVDKSSSTAYM QLSSLTSEDSAVYYCTRGGDGYLDYWGQGTTLTVSS SEQ ID DLL3#5 DIQMTQSPSSLSASVGDRVTVTCQASQDISNYLNWYQQKPG —115— SEQ ID Brief Sequence Number description of sequence NO:45 VL KAPKLLIYDASNLETGVPSRFSGSGSGTDFTFTISSLQPEDIAT YYCQQYDNLPTWTFGQGTKVEIK SEQ ID DLL3#5 QVQLVQSGAEVKKPGASVKVSCKASGFTFTGYYIHWVRQA NO:46 VH PGQGLEWMGWINPNSGGTNYAQKFQGRVTMTRDSSINTAF MELSRLTSDDTAVYYCAAGWDYWGQGTLVTVSS SEQ ID DLL3#6 DIQLTQSPSFLSTSVGDRVTITCRASQDISNYFAWYQQKPGK NO:47 VL YAASTLQSGVPSRFSGGGSGTEFTLTISSLQPEDFAT YYCQQLNSYPYTFGQGTKLEIK SEQ ID DLL3#6 QVQLQESGPGLVKPSETLSLTCTVSGGSISSYFWSWIRQPAG NO:48 VH KGLEWIGRIYTSGSTNYNPSLNSRLTMSVDTSKNQFSLKLSS VTAADTAV Y Y CARRGDWGGFDIWGQGTVVTVSS SEQ ID DLL3#1 EIVLTQSPGTLSLSPGERATLSCRASQSVSSNFLVWYQQKPG NO:49 scFab QAPRPLIYGASTRASGIPDRFSGSGSGADFTLTISRLEPEDFAL GDSPYTFGQGTTLEIKRTVAAPSVFIFPPSDEQLKS GTASVVCLLNNFYPREAKVQWKVDNALQSGNSQESVTEQD SKDSTYSLSSTLTLSKADYEKHKVYACEVTHQGLSSPVTKSF NRGECGGGGSEGKSSGSGSESKSTEGKSSGSGSESKSTGGGG QSGAEVKKPGASVKVSCKASGNTFTNYYMHWVR QAPGPGLEWMGIIDPSVGSKSYAQKFLGRVTIARDTSTSTVF LDLYSLRSEDTAVYFCARAGKRFGESYFDYWGQGTLVTVSS ASTKGPSVFPLAPSSKSTSGGTAALGCLVKDYFPEPVTVSWN SGALTSGVHTFPAVLQSSGLYSLSSVVTVPSSSLGTQTYICNV NHKPSNTKVDKRVEPKSC SEQ ID DLL3#2 DIQMTQSPSAMSASVGDRVTITCRASQGISNYLAWFQQKPG NO:50 scFab KVPEPLIYAASSLQSGVPSRFSGSGSVTEFTLTISSLQPEDFAT YYCLQHNSSPYTFGQGTKLEIKRTVAAPSVFIFPPSDEQLKSG TASVVCLLNNFYPREAKVQWKVDNALQSGNSQESVTEQDS KDSTYSLSSTLTLSKADYEKHKVYACEVTHQGLSSPVTKSF NRGECGGGGSEGKSSGSGSESKSTEGKSSGSGSESKSTGGGG SQVQLVQSGAEVKKPGASVKVSCKASGYTFTSYYMHWVR QAPGQGLEWMGIINPSGGSTSYAQKFQGRVTMTRDTSTSTV YMELSSLRSEDTAVYYCARGEAVGGNYYYYGMDVWGQGT TVTVSSASTKGPSVFPLAPSSKSTSGGTAALGCLVKDYFPEP —116— 2019/064942 SEQ ID Brief Sequence Number description of sequence VTVSWNSGALTSGVHTFPAVLQSSGLYSLSSVVTVPSSSLGT QTYICNVNHKPSNTKVDKRVEPKSC SEQ ID DLL3#3 SPSAMSASVGDRVTITCRASQGISNYLVWFQQKPG N025 1 scFab KAPKRLIYAVSSLYSGVPSRFSGSGSGTEFTLTISSLQPEDFAT YYCLQHDSYPYTFGQGTKLEIKRTVAAPSVFIFPPSDEQLKS GTASVVCLLNNFYPREAKVQWKVDNALQSGNSQESVTEQD SKDSTYSLSSTLTLSKADYEKHKVYACEVTHQGLSSPVTKSF GGGSEGKSSGSGSESKSTEGKSSGSGSESKSTGGGG SQVQLVQSGAEVKKPGASVKVSCKASGYTFTSYYVHWVRQ APGQGLEWMVIINPGGGTTSYAQKFLGRVTMTRDTSTNTVY MELKSLRSEDTAVYYCARGEAVTGNYFYYGMDVWGQGTT VTVSSASTKGPSVFPLAPSSKSTSGGTAALGCLVKDYFPEPV TVSWNSGALTSGVHTFPAVLQSSGLYSLSSWTVPSSSLGTQ TYICNVNHKPSNTKVDKRVEPKSC SEQ ID DLL3#4 DIVLTQSPASLAVSLGQRATISCRASKSVSSFGYSFMHWYQQ NO:52 scFab KPGQPPKLLIYLASNLESGVPARFSGSGSGTDFTLNIHPVEEE DAATYYCQHSRELPWTFGGGTKLEIKRTVAAPSVFIFPPSDE QLKSGTASVVCLLNNFYPREAKVQWKVDNALQSGNSQESV TEQDSKDSTYSLSSTLTLSKADYEKHKVYACEVTHQGLSSP VTKSFNRGECGGGGSEGKSSGSGSESKSTEGKSSGSGSESKS TGGGGSQVQLQQSGTELVKPGASVKLSCKASVYTFTSYFMY WVKQRPGHGLEWIGEISPTNGNSNLNERFKNKATLTVDKSS STAYMQLSSLTSEDSAVYYCTRGGDGYLDYWGQGTTLTVS SASTKGPSVFPLAPSSKSTSGGTAALGCLVKDYFPEPVTVSW NSGALTSGVHTFPAVLQSSGLYSLSSVVTVPSSSLGTQTYICN VNHKPSNTKVDKRVEPKSC SEQ ID DLL3#5 DIQMTQSPSSLSASVGDRVTVTCQASQDISNYLNWYQQKPG NO:53 scFab KAPKLLIYDASNLETGVPSRFSGSGSGTDFTFTISSLQPEDIAT YYCQQYDNLPTWTFGQGTKVEIKRTVAAPSVFIFPPSDEQLK SGTASVVCLLNNFYPREAKVQWKVDNALQSGNSQESVTEQ DSKDSTYSLSSTLTLSKADYEKHKVYACEVTHQGLSSPVTK SFNRGECGGGGSEGKSSGSGSESKSTEGKSSGSGSESKSTGG GGSQVQLVQSGAEVKKPGASVKVSCKASGFTFTGYYIHWV —117— W0 20191234220 SEQ ID Brief Sequence Number ption ofsequence RQAPGQGLEWWKHVDWNSGGTNYAQKFQGRVTMTRDS$N TAFMELSRLTSDDTAVYYCAAGWDYWGQGTLVTVSSASTK GPSVFPLAPSSKSTSGGTAALGCLVKIBTTEPVTVSVWNSGAL TSGVHTFPAVLQSSGLYSLSSVVTVPSSSLGTQTYTCNVTHHK PSNTKVDKRVEPKSC SEQ ID DLL3#6 DIQLTQSPSFLSTSVGDRVTITCRASQDISNYFAWYQQKPGK N054 scFab APKLLIYAASTLQSGVPSRFSGGGSGTEFTLTISSLQPEDFAT YYCQQLNSYPYTFGQGTKLEHKRTVAAPSVFHTPSDEQLKS GTASVVCLLNNFYPREAKNTBNKVDNALQSGNSQESVTEQD SKDSTYSLSSTLTLSKADYEKHKVYACEVTHQGLSSPVTKSF NRGECGGGGSEGKSSGSGSESKSTEGKSSGSGSESKSTGGGG SQVQLQESGPGLVKP$HISLTCTVSGGSSSYFWSWHRQPA GKGLEWIGRIYTSGSTNYNPSLNSRLTMSVDTSKNQFSLKLS SVTAADTAVYYCARRLHNNGGFDDNGQGTVVTVSSASTKGP PSSKSTSGGTAALGCLVKDYFPEPVTVSWNSGALTS GVHTFPAVLQSSGLYSLSSVVTVPSSSLGTQTYHHVVNHKPS RVEPKSC SEQID CD3#1 RSSTGAVTTSNYAN N055 LCCDRI SEQID CD3#1 GTNKRAP N056 LCCDR2 SEQ ID CD3#1 ALWYSNLWV N057 LCCDR3 SEQID CD3#1 GFTFNTYALHQ N058 HCCDRI SEQID CD3#1 IURSKXTDWLATYYADSVKD N059 HCCDR2 SEQID CD3#1 HGNFGNSYVSWWAY N050 HCCDR3 SEQID CD3#2 RSSTGAVTTSNYVUQ N051 LCCDRI SEQID CD3#2 GTNKRAP N052 LCCDR2 —118— SEQ ID Brief Sequence Number description of sequence SEQ ID CD3#2 ALWYSNLWV I\O:63 LCCDR3 SEQ ID CD3#2 GFTFNTYAMN I\O:64 HCCDR1 SEQ ID CD3#2 RIRSKYINYATYYADSVKD I\O:65 HCCDRZ SEQ ID CD3#2 SYVSWFAY I\O:66 HCCDR3 SEQ ID CD3#1 EAVVTQEPSLTVSPGGTVTLTCRSSTGAVTTSNYANWVQEK I\O:67 VL PGQLPRGLIGGTNKRAPWVPARFSGSLLGGKAALTLSGAQP EDEAEYFCALWYSNLWVFGGGTKLTVL SEQ ID CD3#1 SGGGLVQPGGSLKLSCAASGFTFNTYAMNWVRQA N0268 VH PGKGLEWVARIRSKYNNYATYYADSVKDRFTISRDDSKNTA LKTEDTAVYYCVRHGNFGNSYVSWFAYWGQGTL VTVSA SEQ ID CD3#2 EAVVTQEPSLTVSPGGTVTLTCRSSTGAVTTSNYANWVQEK NO:69 VL PGQLPRGLIGGTNKRAPWVPARFSGSLLGGKAALTLSGAQP EDEAEYFCALWYSNLWVFGGGTKLTVL SEQ ID CD3#2 EVQLVESGGGLVQPGGSLKLSCAASGFTFNTYAMNWVRQA NO:7O VH PGKGLEWVARIRSKYINYATYYADSVKDRFTISRDDSKNTA YLQMNNLKTEDTAVYYCVRHGNFGNSYVSWFAYWGQGTL VTVSA SEQ ID CD3#1 EAVVTQEPSLTVSPGGTVTLTCRSSTGAVTTSNYANWVQEK NO:71 scFab PGQLPRGLIGGTNKRAPWVPARFSGSLLGGKAALTLSGAQP EDEAEYFCALWYSNLWVFGGGTKLTVLGQPKAAPSVTLFPP SSEELQANKATLVCLISDFYPGAVKVAWKADGSPVNTGVET TTPSKQSNNKYAASSYLSLTPEQWKSHRSYSCQVTHEGSTV EKTVAPAECSGGGGSEGKSSGSGSESKSTEGKSSGSGSESKS TGGGGSEVQLVESGGGLVQPGGSLKLSCAASGFTFNTYAM NWVRQAPGKGLEWVARIRSKYNNYATYYADSVKDRFTISR DDSKNTAYLQMNNLKTEDTAVYYCVRHGNFGNSYVSWFA YWGQGTLVTVSAASTKGPSVFPLAPSSKSTSGGTAALGCLV KDYFPEPVTVSWNSGALTSGVHTFPAVLQSSGLYSLSSVVT —119— SEQ ID Brief Sequence Number ption of sequence VPSSSLGTQTYICNVNHKPSNTKVDKRVEPKSC SEQ ID CD3#2 EAVVTQEPSLTVSPGGTVTLTCRSSTGAVTTSNYANWVQEK NO:72 scFab PGQLPRGLIGGTNKRAPWVPARFSGSLLGGKAALTLSGAQP EDEAEYFCALWYSNLWVFGGGTKLTVLGQPKAAPSVTLFPP SSEELQANKATLVCLISDFYPGAVKVAWKADGSPVNTGVET TTPSKQSNNKYAASSYLSLTPEQWKSHRSYSCQVTHEGSTV EKTVAPAECSGGGGSEGKSSGSGSESKSTEGKSSGSGSESKS TGGGGSEVQLVESGGGLVQPGGSLKLSCAASGFTFNTYAM NWVRQAPGKGLEWVARIRSKYINYATYYADSVKDRFTISRD DSKNTAYLQMNNLKTEDTAVYYCVRHGNFGNSYVSWFAY WGQGTLVTVSAASTKGPSVFPLAPSSKSTSGGTAALGCLVK DYFPEPVTVSWNSGALTSGVHTFPAVLQSSGLYSLSSVVTVP SSSLGTQTYICNVNHKPSNTKVDKRVEPKSC SEQ ID DLL3#1 EIVLTQSPGTLSLSPGERATLSCRASQSVSSNFLVWYQQKPG NO:73 chain QAPRPLIYGASTRASGIPDRFSGSGSGADFTLTISRLEPEDFAL YYCQQYGDSPYTFGQGTTLEIKRTVAAPSVFIFPPSDEQLKS GTASVVCLLNNFYPREAKVQWKVDNALQSGNSQESVTEQD SKDSTYSLSSTLTLSKADYEKHKVYACEVTHQGLSSPVTKSF NRGECGGGGSEGKSSGSGSESKSTEGKSSGSGSESKSTGGGG SQVQLVQSGAEVKKPGASVKVSCKASGNTFTNYYMHWVR QAPGPGLEWMGIIDPSVGSKSYAQKFLGRVTIARDTSTSTVF RSEDTAVYFCARAGKRFGESYFDYWGQGTLVTVSS ASTKGPSVFPLAPSSKSTSGGTAALGCLVKDYFPEPVTVSWN SGALTSGVHTFPAVLQSSGLYSLSSVVTVPSSSLGTQTYICNV TKVDKRVEPKSCDKTHTCPPCPAPEAAGGPSVFLFP PKPKDTLMISRTPEVTCVVVDVSHEDPEVKFNWYVDGVEV HNAKTKPREEQYNSTYRVVSVLTVLHQDWLNGKEYKCKVS NKALPAPIEKTISKAKGQPREPQVYTLPPSREEMTKNQVSLW CLVKGFYPSDIAVEWESNGQPENNYKTTPPVLDSDGSFFLYS KLTVDKSRWQQGNVFSCSVMHEALHNHYTQKSLSLSPG SEQ ID DLL3#2 DIQMTQSPSAMSASVGDRVTITCRASQGISNYLAWFQQKPG NO:74 chain KVPEPLIYAASSLQSGVPSRFSGSGSVTEFTLTISSLQPEDFAT YYCLQHNSSPYTFGQGTKLEIKRTVAAPSVFIFPPSDEQLKSG —120— SEQ ID Brief Sequence Number description of sequence TASVVCLLNNFYPREAKVQWKVDNALQSGNSQESVTEQDS KDSTYSLSSTLTLSKADYEKHKVYACEVTHQGLSSPVTKSF NRGECGGGGSEGKSSGSGSESKSTEGKSSGSGSESKSTGGGG SQVQLVQSGAEVKKPGASVKVSCKASGYTFTSYYMHWVR QAPGQGLEWMGIINPSGGSTSYAQKFQGRVTMTRDTSTSTV YMELSSLRSEDTAVYYCARGEAVGGNYYYYGMDVWGQGT ASTKGPSVFPLAPSSKSTSGGTAALGCLVKDYFPEP VTVSWNSGALTSGVHTFPAVLQSSGLYSLSSVVTVPSSSLGT QTYICNVNHKPSNTKVDKRVEPKSCDKTHTCPPCPAPEAAG GPSVFLFPPKPKDTLMISRTPEVTCVVVDVSHEDPEVKFNWY VDGVEVHNAKTKPREEQYNSTYRVVSVLTVLHQDWLNGK EYKCKVSNKALPAPIEKTISKAKGQPREPQVYTLPPSREEMT KNQVSLWCLVKGFYPSDIAVEWESNGQPENNYKTTPPVLDS DGSFFLYSKLTVDKSRWQQGNVFSCSVMHEALHNHYTQKS LSLSPG SEQ ID DLL3#3 SPSAMSASVGDRVTITCRASQGISNYLVWFQQKPG NO:75 chain IYAVSSLYSGVPSRFSGSGSGTEFTLTISSLQPEDFAT YYCLQHDSYPYTFGQGTKLEIKRTVAAPSVFIFPPSDEQLKS GTASVVCLLNNFYPREAKVQWKVDNALQSGNSQESVTEQD SKDSTYSLSSTLTLSKADYEKHKVYACEVTHQGLSSPVTKSF NRGECGGGGSEGKSSGSGSESKSTEGKSSGSGSESKSTGGGG SQVQLVQSGAEVKKPGASVKVSCKASGYTFTSYYVHWVRQ APGQGLEWMVIINPGGGTTSYAQKFLGRVTMTRDTSTNTVY MELKSLRSEDTAVYYCARGEAVTGNYFYYGMDVWGQGTT VTVSSASTKGPSVFPLAPSSKSTSGGTAALGCLVKDYFPEPV TVSWNSGALTSGVHTFPAVLQSSGLYSLSSVVTVPSSSLGTQ TYICNVNHKPSNTKVDKRVEPKSCDKTHTCPPCPAPEAAGG PSVFLFPPKPKDTLMISRTPEVTCVVVDVSHEDPEVKFNWYV DGVEVHNAKTKPREEQYNSTYRVVSVLTVLHQDWLNGKE YKCKVSNKALPAPIEKTISKAKGQPREPQVYTLPPSREEMTK NQVSLWCLVKGFYPSDIAVEWESNGQPENNYKTTPPVLDSD GSFFLYSKLTVDKSRWQQGNVFSCSVMHEALHNHYTQKSL SLSPG —121— SEQ ID Brief Sequence Number description of sequence SEQ ID DLL3#4 DIVLTQSPASLAVSLGQRATISCRASKSVSSFGYSFMHWYQQ NO:76 chain KPGQPPKLLIYLASNLESGVPARFSGSGSGTDFTLNIHPVEEE DAATYYCQHSRELPWTFGGGTKLEIKRTVAAPSVFIFPPSDE QLKSGTASVVCLLNNFYPREAKVQWKVDNALQSGNSQESV TEQDSKDSTYSLSSTLTLSKADYEKHKVYACEVTHQGLSSP VTKSFNRGECGGGGSEGKSSGSGSESKSTEGKSSGSGSESKS TGGGGSQVQLQQSGTELVKPGASVKLSCKASVYTFTSYFMY WVKQRPGHGLEWIGEISPTNGNSNLNERFKNKATLTVDKSS STAYMQLSSLTSEDSAVYYCTRGGDGYLDYWGQGTTLTVS SASTKGPSVFPLAPSSKSTSGGTAALGCLVKDYFPEPVTVSW NSGALTSGVHTFPAVLQSSGLYSLSSVVTVPSSSLGTQTYICN VNHKPSNTKVDKRVEPKSCDKTHTCPPCPAPEAAGGPSVFL DTLMISRTPEVTCVVVDVSHEDPEVKFNWYVDGVE VHNAKTKPREEQYNSTYRVVSVLTVLHQDWLNGKEYKCK VSNKALPAPIEKTISKAKGQPREPQVYTLPPSREEMTKNQVS LWCLVKGFYPSDIAVEWESNGQPENNYKTTPPVLDSDGSFF LYSKLTVDKSRWQQGNVFSCSVMHEALHNHYTQKSLSLSP SEQ ID DLL3#5 DIQMTQSPSSLSASVGDRVTVTCQASQDISNYLNWYQQKPG NO:77 chain KAPKLLIYDASNLETGVPSRFSGSGSGTDFTFTISSLQPEDIAT YYCQQYDNLPTWTFGQGTKVEIKRTVAAPSVFIFPPSDEQLK SGTASVVCLLNNFYPREAKVQWKVDNALQSGNSQESVTEQ DSKDSTYSLSSTLTLSKADYEKHKVYACEVTHQGLSSPVTK SFNRGECGGGGSEGKSSGSGSESKSTEGKSSGSGSESKSTGG LVQSGAEVKKPGASVKVSCKASGFTFTGYYIHWV RQAPGQGLEWMGWINPNSGGTNYAQKFQGRVTMTRDSSIN TAFMELSRLTSDDTAVYYCAAGWDYWGQGTLVTVSSASTK GPSVFPLAPSSKSTSGGTAALGCLVKDYFPEPVTVSWNSGAL TSGVHTFPAVLQSSGLYSLSSVVTVPSSSLGTQTYICNVNHK PSNTKVDKRVEPKSCDKTHTCPPCPAPEAAGGPSVFLFPPKP KDTLMISRTPEVTCVVVDVSHEDPEVKFNWYVDGVEVHNA KTKPREEQYNSTYRVVSVLTVLHQDWLNGKEYKCKVSNKA LPAPIEKTISKAKGQPREPQVYTLPPSREEMTKNQVSLWCLV —122— SEQ ID Brief Sequence Number description of sequence KGFYPSDIAVEWESNGQPENNYKTTPPVLDSDGSFFLYSKLT QQGNVFSCSVMHEALHNHYTQKSLSLSPG SEQ ID DLL3#6 DIQLTQSPSFLSTSVGDRVTITCRASQDISNYFAWYQQKPGK NO:78 chain APKLLIYAASTLQSGVPSRFSGGGSGTEFTLTISSLQPEDFAT YYCQQLNSYPYTFGQGTKLEIKRTVAAPSVFIFPPSDEQLKS GTASVVCLLNNFYPREAKVQWKVDNALQSGNSQESVTEQD SKDSTYSLSSTLTLSKADYEKHKVYACEVTHQGLSSPVTKSF NRGECGGGGSEGKSSGSGSESKSTEGKSSGSGSESKSTGGGG SQVQLQESGPGLVKPSETLSLTCTVSGGSISSYFWSWIRQPA GKGLEWIGRIYTSGSTNYNPSLNSRLTMSVDTSKNQFSLKLS SVTAADTAVYYCARRGDWGGFDIWGQGTVVTVSSASTKGP SVFPLAPSSKSTSGGTAALGCLVKDYFPEPVTVSWNSGALTS AVLQSSGLYSLSSWTVPSSSLGTQTYICNVNHKPS NTKVDKRVEPKSCDKTHTCPPCPAPEAAGGPSVFLFPPKPKD TLMISRTPEVTCVVVDVSHEDPEVKFNWYVDGVEVHNAKT KPREEQYNSTYRVVSVLTVLHQDWLNGKEYKCKVSNKALP APIEKTISKAKGQPREPQVYTLPPSREEMTKNQVSLWCLVKG AVEWESNGQPENNYKTTPPVLDSDGSFFLYSKLTVD KSRWQQGNVFSCSVMHEALHNHYTQKSLSLSPG SEQ ID CD3#1 EAVVTQEPSLTVSPGGTVTLTCRSSTGAVTTSNYANWVQEK NO:79 chain PGQLPRGLIGGTNKRAPWVPARFSGSLLGGKAALTLSGAQP EDEAEYFCALWYSNLWVFGGGTKLTVLGQPKAAPSVTLFPP SSEELQANKATLVCLISDFYPGAVKVAWKADGSPVNTGVET TTPSKQSNNKYAASSYLSLTPEQWKSHRSYSCQVTHEGSTV EKTVAPAECSGGGGSEGKSSGSGSESKSTEGKSSGSGSESKS TGGGGSEVQLVESGGGLVQPGGSLKLSCAASGFTFNTYAM NWVRQAPGKGLEWVARIRSKYNNYATYYADSVKDRFTISR DDSKNTAYLQMNNLKTEDTAVYYCVRHGNFGNSYVSWFA YWGQGTLVTVSAASTKGPSVFPLAPSSKSTSGGTAALGCLV KDYFPEPVTVSWNSGALTSGVHTFPAVLQSSGLYSLSSVVT VPSSSLGTQTYICNVNHKPSNTKVDKRVEPKSCDKTHTCPPC PAPEAAGGPSVFLFPPKPKDTLMISRTPEVTCVVVDVSHEDP EVKFNWYVDGVEVHNAKTKPREEQYNSTYRVVSVLTVLHQ —123— SEQ ID Brief Sequence Number description of sequence DWLNGKEYKCKVSNKALPAPIEKTISKAKGQPREPQVYTLP PSREEMTKNQVSLSCAVKGFYPSDIAVEWESNGQPENNYKT TPPVLDSDGSFFLVSKLTVDKSRWQQGNVFSCSVMHEALHN LSLSPG SEQ ID CD3#2 EAVVTQEPSLTVSPGGTVTLTCRSSTGAVTTSNYANWVQEK NO:80 chain PGQLPRGLIGGTNKRAPWVPARFSGSLLGGKAALTLSGAQP EDEAEYFCALWYSNLWVFGGGTKLTVLGQPKAAPSVTLFPP SSEELQANKATLVCLISDFYPGAVKVAWKADGSPVNTGVET TTPSKQSNNKYAASSYLSLTPEQWKSHRSYSCQVTHEGSTV EKTVAPAECSGGGGSEGKSSGSGSESKSTEGKSSGSGSESKS TGGGGSEVQLVESGGGLVQPGGSLKLSCAASGFTFNTYAM NWVRQAPGKGLEWVARIRSKYINYATYYADSVKDRFTISRD DSKNTAYLQMNNLKTEDTAVYYCVRHGNFGNSYVSWFAY WGQGTLVTVSAASTKGPSVFPLAPSSKSTSGGTAALGCLVK DYFPEPVTVSWNSGALTSGVHTFPAVLQSSGLYSLSSVVTVP QTYICNVNHKPSNTKVDKRVEPKSCDKTHTCPPCPA PEAAGGPSVFLFPPKPKDTLMISRTPEVTCVVVDVSHEDPEV KFNWYVDGVEVHNAKTKPREEQYNSTYRVVSVLTVLHQD WLNGKEYKCKVSNKALPAPIEKTISKAKGQPREPQVYTLPPS REEMTKNQVSLSCAVKGFYPSDIAVEWESNGQPENNYKTTP PVLDSDGSFFLVSKLTVDKSRWQQGNVFSCSVMHEALHNRF TQKSLSLSPG SEQ ID Fc domain* DKTHTCPPCPAPELLGGPSVFLFPPKPKDTLMISRTPEVTCVV NO:81 (IgGl) DPEVKFNWYVDGVEVHNAKTKPREEQYNSTYRV VSVLTVLHQDWLNGKEYKCKVSNKALPAPIEKTISKAKGQP REPQVYTLPPSREEMTKNQVSLTCLVKGFYPSDIAVEWESN GQPENNYKTTPPVLDSDGSFFLYSKLTVDKSRWQQGNVFSC SVMHEALHNHYTQKSLSLSPG SEQ ID Fc W domain DKTHTCPPCPAPEAAGGPSVFLFPPKPKDTLMISRTPEVTCV NO:82 (IgG 1 , VVDVSHEDPEVKFNWYVDGVEVHNAKTKPREEQYNSTYR LALA) VVSVLTVLHQDWLNGKEYKCKVSNKALPAPIEKTISKAKGQ PREPQVYTLPPSREEMTKNQVSLWCLVKGFYPSDIAVEWES NGQPENNYKTTPPVLDSDGSFFLYSKLTVDKSRWQQGNVFS —124— SEQ ID Brief Sequence Number description ence CSVMHEALHNHYTQKSLSLSPG SEQ ID Fc SAV DKTHTCPPCPAPEAAGGPSVFLFPPKPKDTLMISRTPEVTCV NO:83 domain VVDVSHEDPEVKFNWYVDGVEVHNAKTKPREEQYNSTYR (IgG 1 , VVSVLTVLHQDWLNGKEYKCKVSNKALPAPIEKTISKAKGQ RF/LALA) PREPQVYTLPPSREEMTKNQVSLSCAVKGFYPSDIAVEWESN GQPENNYKTTPPVLDSDGSFFLVSKLTVDKSRWQQGNVFSC SVMHEALHNRFTQKSLSLSPG SEQ ID Fc domain ESKYGPPCPPCPAPEFLGGPSVFLFPPKPKDTLMISRTPEVTC NO:84 (IgG4Pro) VVVDVSQEDPEVQFNWYVDGVEVHNAKTKPREEQFNSTYR VVSVLTVLHQDWLNGKEYKCKVSNKGLPSSIEKTISKAKGQ PREPQVYTLPPSQEEMTKNQVSLTCLVKGFYPSDIAVEWESN GQPENNYKTTPPVLDSDGSFFLYSRLTVDKSRWQEGNVFSC SVMHEALHNHYTQKSLSLSLG SEQ ID Fc W domain ESKYGPPCPPCPAPEFLGGPSVFLFPPKPKDTLMISRTPEVTC NO:85 (IgG4Pro) EDPEVQFNWYVDGVEVHNAKTKPREEQFNSTYR VVSVLTVLHQDWLNGKEYKCKVSNKGLPSSIEKTISKAKGQ PREPQVYTLPPSQEEMTKNQVSLWCLVKGFYPSDIAVEWES NGQPENNYKTTPPVLDSDGSFFLYSRLTVDKSRWQEGNVFS ALHNHYTQKSLSLSLG SEQ ID Fc SAV ESKYGPPCPPCPAPEFLGGPSVFLFPPKPKDTLMISRTPEVTC NO:86 domain WVDVSQEDPEVQFNWYVDGVEVHNAKTKPREEQFNSTYR (IgG4Pro, VVSVLTVLHQDWLNGKEYKCKVSNKGLPSSIEKTISKAKGQ RF) YTLPPSQEEMTKNQVSLSCAVKGFYPSDIAVEWES NGQPENNYKTTPPVLDSDGSFFLVSRLTVDKSRWQEGNVFS CSVMHEALHNRFTQKSLSLSLG SEQ ID constant RTVAAPSVFIFPPSDEQLKSGTASVVCLLNNFYPREAKVQW NO:87 region of a KVDNALQSGNSQESVTEQDSKDSTYSLSSTLTLSKADYEKH kappa light KVYACEVTHQGLSSPVTKSFNRGEC chain SEQ ID constant GQPKAAPSVTLFPPSSEELQANKATLVCLISDFYPGAVKVA NO:88 region of a WKADGSPVNTGVETTTPSKQSNNKYAASSYLSLTPEQWKS lambda light HRSYSCQVTHEGSTVEKTVAPAECS chain —l25— SEQ ID Brief Sequence Number ption ofsequence SEQ ID Linker GGGGSEGKSSGSGSESKSTEGKSSGSGSESKSTGGGGS 1\O:89 SEQ ID Linker GGGGSGGGGSGGSGGSGGGGGS 1\O:90 SEQ ID Linker GGGGSGGGGSGGGGSGGGGSGGGGGS R0291 SEQ ID Linker GGGGSGGGGGGSGGGGGGSGGGGSGGGGGS R0192 SEQ ID Linker GGGGSGGGGSGGGSGGGSGGGGSGGGGSGGGGGS 1\O:93 SEQ ID Linker GGGGSGGGGSGGGSGGGSGGGSGGGGSGGGGSGGGGGS R0194 SEQ ID Linker GGGGSGGGGSGGGSGGGSGGGSGGGGSGGGGGSGGGSGG R0195 GGS SEQ ID CD3#3 RSSTGAVTTSNYAN R0196 LCCDR1 SEQ ID CD3#3 GTNKRAP R0197 LCCDRZ SEQ ID CD3#3 ALWYSNLWV R0298 LCCDR3 SEQ ID CD3#3 GFTFNTYAMN R0299 HCCDR1 SEQ ID CD3#3 RIRSKYANYATYYADSVKD RO:100 HCCDRZ SEQ ID CD3#3 HGNFGNSYVSWFAY RO:101 HCCDR3 SEQ ID CD3#3 EAVVTQEPSLTVSPGGTVTLTCRSSTGAVTTSNYANWVQEK R0: 1 02 VL PGQLPRGLIGGTNKRAPWVPARFSGSLLGGKAALTLSGAQP EDEAEYFCALWYSNLWVFGGGTKLTVL SEQ ID CD3#3 SGGGLVQPGGSLKLSCAASGFTFNTYAMNWVRQA NO: 1 03 HL PGKGLEWVARIRSKYANYATYYADSVKDRFTISRDDSKNTA YLQMNNLKTEDTAVYYCVRHGNFGNSYVSWFAYWGQGTL VTVSA —126— SEQ ID Brief Sequence Number description of sequence SEQ ID CD3#3 EAVVTQEPSLTVSPGGTVTLTCRSSTGAVTTSNYANWVQEK NO: 1 04 scFab PGQLPRGLIGGTNKRAPWVPARFSGSLLGGKAALTLSGAQP EDEAEYFCALWYSNLWVFGGGTKLTVLGQPKAAPSVTLFPP SSEELQANKATLVCLISDFYPGAVKVAWKADGSPVNTGVET TTPSKQSNNKYAASSYLSLTPEQWKSHRSYSCQVTHEGSTV EKTVAPAECSGGGGSEGKSSGSGSESKSTEGKSSGSGSESKS TGGGGSEVQLVESGGGLVQPGGSLKLSCAASGFTFNTYAM NWVRQAPGKGLEWVARIRSKYANYATYYADSVKDRFTISR DDSKNTAYLQMNNLKTEDTAVYYCVRHGNFGNSYVSWFA YWGQGTLVTVSAASTKGPSVFPLAPSSKSTSGGTAALGCLV KDYFPEPVTVSWNSGALTSGVHTFPAVLQSSGLYSLSSVVT VPSSSLGTQTYICNVNHKPSNTKVDKRVEPKSC SEQ ID CD3#3 EAVVTQEPSLTVSPGGTVTLTCRSSTGAVTTSNYANWVQEK N0:105 chain PGQLPRGLIGGTNKRAPWVPARFSGSLLGGKAALTLSGAQP EDEAEYFCALWYSNLWVFGGGTKLTVLGQPKAAPSVTLFPP SSEELQANKATLVCLISDFYPGAVKVAWKADGSPVNTGVET TTPSKQSNNKYAASSYLSLTPEQWKSHRSYSCQVTHEGSTV EKTVAPAECSGGGGSEGKSSGSGSESKSTEGKSSGSGSESKS EVQLVESGGGLVQPGGSLKLSCAASGFTFNTYAM NWVRQAPGKGLEWVARIRSKYANYATYYADSVKDRFTISR DDSKNTAYLQMNNLKTEDTAVYYCVRHGNFGNSYVSWFA YWGQGTLVTVSAASTKGPSVFPLAPSSKSTSGGTAALGCLV KDYFPEPVTVSWNSGALTSGVHTFPAVLQSSGLYSLSSVVT VPSSSLGTQTYICNVNHKPSNTKVDKRVEPKSCDKTHTCPPC PAPEAAGGPSVFLFPPKPKDTLMISRTPEVTCVVVDVSHEDP EVKFNWYVDGVEVHNAKTKPREEQYNSTYRVVSVLTVLHQ EYKCKVSNKALPAPIEKTISKAKGQPREPQVYTLP TKNQVSLSCAVKGFYPSDIAVEWESNGQPENNYKT TPPVLDSDGSFFLVSKLTVDKSRWQQGNVFSCSVMHEALHN RFTQKSLSLSPG SEQ ID DLL3#7 RASQSVNSNFLA N0:133 LCCDRI SEQ ID DLL3#7 GTSSRAT —127— SEQ ID Brief Sequence Number description ofsequence NO:134 LCCDR2 SEQ ID DLL3#7 QQYGSSPWT NO:135 LCCDR3 SEQ ID DLL3#7 YGMF NO:136 HCCDR1 SEQ ID DLL3#7 VIWLDGDDEDYVDSVKG NO:137 HCCDR2 SEQ ID DLL3#7 VLDY NO:138 HCCDR3 SEQ ID DLL3#8 KSSQSVLDTSNNKNYLV NO:139 LCCDR1 SEQ ID DLL3#8 WASTRES NO: 140 LCCDR2 SEQ ID DLL3#8 QHYYNSPYT NO: 141 LCCDR3 SEQ ID DLL3#8 GYTFTDYYMH NO: 142 HCCDR1 SEQ ID DLL3#8 WINPNSGGTNYEQKFQG NO: 143 HCCDR2 SEQ ID DLL3#8 DAWIPMDY NO: 144 HCCDR3 SEQ ID DLL3#9 RASQSISRSYLA NO: 145 LCCDR1 SEQ ID DLL3#9 GASSRAT NO: 146 LCCDR2 SEQ ID DLL3#9 QQYGTSPIT NO: 147 LCCDR3 SEQ ID DLL3#9 GGSISSYYWS NO: 148 HCCDRI SEQ ID DLL3#9 NTNYNPSLKS NO: 149 HCCDR2 SEQ ID DLL3#9 IGVAGFYFDY N02150 HCCDR3 —128— SEQ ID Brief Sequence Number description ence SEQ ID DLL3#10 RASQSLNSIFLA BOJSI LCCDR1 SEQID DLLﬁﬂO GASSRAT BOASZ LCCDRZ SEQ ID DLL3#1O QQYGGSMNT hOﬂ53 LCCDR3 SEQ ID DLL3#10 GYTFTGYYMH hOﬂ54 HCCDR1 SEQ ID 0 WINPNSGGTIFAQRFQG BOASS HCCDRZ SEQID DLLﬁﬂO DFGDTVGNAFDI hOﬂ56 HCCDR3 SEQID , SASSSVTYHI h0ﬂ57 LCCDR1 SEQID DLLyﬂl RTSYLAS h0ﬂ58 LCCDRZ SEQID DLLﬁﬂl QQRSSYPRT h0ﬂ59 LCCDR3 SEQID DLLﬁﬂl GYAFSDYWHT .\O:160 HCCDR1 SEQID DLL3#H. IHYPGSGSTKSSEKFKN R0ﬂ61 HCCDRZ SEQID DLLﬁﬂl LYYYGSHYLDT X0ﬂ62 HCCDR3 SEQID DLLﬁﬂZ SASSSVTYHI R0ﬂ63 LCCDR1 SEQID DLLﬁﬂZ RTSYLAS R0ﬂ64 LCCDR2 SEQID DLLﬁﬂ2 QQRSSYPRT R0ﬂ65 LCCDR3 SEQID DLLﬁﬂ2 GYAFSDYWHT .\O:166 HCCDR1 SEQID DLLﬁﬂ2 IMYPGSGSTKSSEKFKN —129— SEQ ID Brief Sequence Number description ofsequence NOﬂ67 HCCDR2 SEQID DLLﬁﬂZ LYYYGSYYLDT NOﬂ68 HCCDR3 SEQ ID DLL3#13 RSSQSIVHSNGNTYLE NOﬂ69 LCCDRl SEQID DLLﬁﬂ3 KVSNRFS NOﬂ70 LCCDR2 SEQID DLLyﬂ3 FQGSHVPYT NOﬂ71 LCCDR3 SEQID DLLyﬂ3 GYTFTNYGVT N0ﬂ72 HCCDRl SEQID DLLyﬂ3 “HNTYSGAPTYADDFNG N0ﬂ73 HCCDR2 SEQID DLLyﬂ3 LDDYDLYYFDY NOﬂ74 HCCDR3 SEQID DLLyﬂ4 KASQSVDYDGDSYNDJ NOﬂ75 LCCDRl SEQID DLLyﬂ4 AASTLES NOﬂ76 LCCDR2 SEQ ID DLL3#14 PWT NOﬂ77 LCCDR3 SEQID DLL3ﬂ4 GYTFTDYWTH N0ﬂ78 HCCDRl SEQID DLL3ﬂ4 XHYPGNSYTAYNQKFKD NOﬂ79 HCCDR2 SEQ ID DLL3#14 DY NOﬂ80 HCCDR3 SEQID DLLyﬂ5 KASQSVDYDGDSYLN NOﬂ81 LCCDRI SEQID DLLyﬂ5 AASNLES NOﬂ82 LCCDR2 SEQ ID DLL3#15 QQSSEDPRT NOﬂ83 LCCDR3 —130— 2019/064942 SEQ ID Brief Sequence Number description ofsequence SEQID DLLﬁﬂS GYTFTNYGLHQ I\O:184 HCCDR1 SEQID DLLﬁﬂS “HNTYTGEPTYADDFKG BOASS HCCDRZ SEQ ID DLL3#15 FHFSSNGDAMDN I\O:186 HCCDR3 SEQ ID DLL3#16 RASQSVSDWLA h0ﬂ87 LCCDR1 SEQID DLLﬁﬂ6 RASSLES h0ﬂ88 LCCDRZ SEQ ID 6 QLYNSYSPT h0ﬂ89 LCCDR3 SEQ ID DLL3#16 GFTFSSYWMT I\O:190 HCCDR1 SEQID DLLyﬂ6 PHKEDGSEKYYVDSVKG h0ﬂ91 HCCDRZ SEQ ID DLL3#16 DWGYFDY h0ﬂ92 HCCDR3 SEQ ID DLL3#17 RASENIYYSLA R0ﬂ93 LCCDR1 SEQ ID DLL3#17 NTNSLED R0ﬂ94 LCCDR2 SEQID DLLﬁﬂ7 KQAYDFPLT R0ﬂ95 LCCDR3 SEQ ID DLL3#17 YYIH .\O:196 HCCDR1 SEQID DLLﬁﬂ7 “HYPGDGSTNNNEKFKG .\O:197 HCCDRZ SEQ ID DLL3#17 GEGNAMDD R0ﬂ98 HCCDR3 SEQ ID DLL3#18 RASENIYYSLA R0ﬂ99 LCCDR1 SEQID DLLﬁﬂ8 'NANSLED —131— SEQ ID Brief Sequence Number description ofsequence NO:200 LCCDR2 SEQ ID 8 KQAYDVPLT NO:201 LCCDR3 SEQ ID DLL3#18 GYTFTAYFIH NO:202 HCCDRI SEQ ID DLL3#18 YIDPFNDDTNYNVKFKG NO:203 HCCDR2 SEQ ID DLL3#18 GTSATLDY NO:204 HCCDR3 SEQ ID DLL3#7 EIVLTQSPDTLSLSPGERATLSCRASQSVNSNFLAWYQQKPG NO:205 VL QTPRLLIFGTSSRATGIPDRFSGSGSGTDFTLTISRLEPEDFAV YYCQQYGSSPWTFGQGTKVEIR SEQ ID DLL3#7 QVQLVESGGGVVQPGRSLRLSCAASGFTFSSYGMFWVRQA NO:206 VH WVAVIWLDGDDEDYVDSVKGRFTISRDDSKNTLY LQMNSLRVDDTAIYYCARVLDYWGQGTLVTVSS SEQ ID DLL3#8 DIVMAQSPDSLAVSLGERATINCKSSQSVLDTSNNKNYLVW NO2207 VL YQQKPGQPPKLLIYWASTRESGVPDRFSGSGSGTDFTLTISSL QAEDVAVYYCQHYYNSPYTFGQGTKLEIK SEQ ID DLL3#8 QVQLVQSGAEVKKPGASVKVSCKASGYTFTDYYMHWVRQ NO:208 VH APGQGLEWMGWINPNSGGTNYEQKFQGRVTMTRDTSISTA YMELNRLRSDDTAVYYCTRDAVVIPMDYWGQGTLVTVSS SEQ ID DLL3#9 EIVLTQSPGTLSLSPGERATLSCRASQSISRSYLAWYQQKPGQ NO:209 VL APRLLIYGASSRATGIPDRFSGSGSGTDFTLTISRLEPEDFAVY YCQQYGTSPITFGQGTRLEIK SEQ ID DLL3#9 QVHLQESGPGLVKPSETLSLTCTVSGGSISSYYWSWIRQTPG NO:21 0 VH KGLDWIGYRYYSGNTNYNPSLKSRVTISLDMSNNQFSLKLS SVTAADTAIYYCASIGVAGFYFDYWGQGTLVTVSS SEQ ID DLL3#10 EIVLTQSPGTLSLSPGERATLSCRASQSLNSIFLAWYQQKPGQ NO:21 1 VL APWLLIYGASSRATGIPDRFSGSGSGTDFTLTISRLEPEDFAV YFCQQYGGSMNTFGQGTKLEIK SEQ ID 0 QVQLVQSGAEVKKPGASVKVSCKASGYTFTGYYMHWVRL NO:212 VH APGQGLEWMGWINPNSGGTIFAQRFQGRVTMTRDTSISTVY MDLNRLRSDDTAVYYCARDFGDTVGNAFDIWGQGTMVTV —132— SEQ ID Brief Sequence Number description ofsequence SEQ ID DLL3#11 QIVLTQSPAIMSASPGEKVTITCSASSSVTYIHWFQQNPGTSP NO:21 3 VL KLWIYRTSYLASGVPARFSGSGSGTSYSLTISRMEAEDAATY YCQQRSSYPRTFGGGTKLEIK SEQ ID DLL3#11 QVQLQQPGAELVQPGSSVKMSCKASGYAFSDYWITWVKQR NO:214 VH PGQGLEWIGDIYPGSGSTKSSEKFKNKATLTADTSSSKAYIQ FSSLTPEDSAVYYCVSLYYYGSHYLDTWGQGTTLTVSS SEQ ID DLL3#12 QVVLTQSPAIMSASPGEKVTITCSASSSVTYIHWFQQNPGTSP NO:21 5 VL KLWIYRTSYLASGVPARFSGSGSGTSYSLTISRMEAEDAATY YCQQRSSYPRTFGGGTKLEIK SEQ ID DLL3#12 QVQLQQPGAEFVQPGSSVKMSCKASGYAFSDYWITWVKQR NO:21 6 VH PGQGLEWIGDIYPGSGSTKSSEKFKNRATLTADTSSSTAYIQF SSLTPEDSAVYYCVSLYYYGSYYLDTWGQGTTLTVSS SEQ ID DLL3#13 DVLMTQTPLSLPVSLGDQAAISCRSSQSIVHSNGNTYLEWYL NO:21 7 VL QKPGQSPKVLIYKVSNRFSGVPDRFSGSGSGTDFTLKISRVE AEDLGVYYCFQGSHVPYTFGGGTKLEIK SEQ ID DLL3#13 QIQLVQSGPELKKPGETVKISCKASGYTFTNYGVTWVKQAP NO:21 8 VH GKGLKWMGWINTYSGAPTYADDFNGRFALSLETSASTAYL QINNLKNEDTATYFCARLDDYDLYYFDYWGQGTALTVSS SEQ ID DLL3#14 DIVLTQSPASLSVSLGQRATISCKASQSVDYDGDSYMNWYQ NO:21 9 VL QKPGQPPKLLIYAASTLESGIPARFSGSGSGTDFTLNIHPVEEE DAATYYCQQSDEDPWTFGGGTKLEIK SEQ ID DLL3#14 QIQLQQSGPELVKPGVKISCKASGYTFTDYYIHWMKQRPGQ NO:220 VH GLEWIGYIYPGNSYTAYNQKFKDKATLTADNPSSTAYMQLS SLTSEDSAVYFCARSGGSAMDYWGQGTSVTVSS SEQ ID DLL3#15 DIVLTQSPASLAVSLGQRATISCKASQSVDYDGDSYLNWYQ NO:221 VL PKLLIYAASNLESGIPARFSGSGSGTDFTLNIHPVEE YCQQSSEDPRTFGGGTKLEIK SEQ ID DLL3#15 QIQLVQSGPELKKPGETVKISCKASGYTFTNYGMNWVKQAP NO:222 VH GKGLKWMGWINTYTGEPTYADDFKGRFAFSLETSASTAYL NEDMATYFCTKFHFSSNGDAMDNWGQGTSVTVSS SEQ ID DLL3#16 DIQMTQSPSTLSASVGDRVTITCRASQSVSDWLAWYQQKPG NO2223 VL KAPKFLIYRASSLESGVPSRFSGSGSGTEFTLTISSLQPADFAT —133— SEQ ID Brief ce Number description of sequence YYCQLYNSYSPTFGQGTKVEIK SEQ ID DLL3#16 EVHLVESGGGLVQPGGSLRLSCAASGFTFSSYWMTWVRQA NO:224 VH PGKGLEWVANIKEDGSEKYYVDSVKGRFTISRDNAKNSLYL QMNSLRAEDTALYYCARDWGYFDYWGQGTLVTVSS SEQ ID DLL3#17 DIQMTQSPASLAASVGETVTITCRASENIYYSLAWYQQKQG NO:225 VL KSPQLLIYNTNSLEDGVPSRFSGSGSGTQYSMKINSMQPEDT ATYFCKQAYDFPLTFGAGTKLELK SEQ ID DLL3#17 QIQLQQSGPEVVKPGASVKISCKASGYTFISYYIHWVKQRPG NO:226 VH QGLEWIGWIYPGDGSTNNNEKFKGKTTLTADKSSSTAYMLL SSLTSEDSAVYFCARGEGNAMDDWGQGTSVTVSS SEQ ID 8 DIQMTQSPASLAASVGETVTITCRASENIYYSLAWYQQKQG NO:227 VL KSPQLLIYNANSLEDGVPSRFSGSGSGTQYSMKINNMQPEDT ATYFCKQAYDVPLTFGAGTKLELK SEQ ID DLL3#18 QVQLQQSGPDLVKPGASVKMSCEASGYTFTAYFIHWVKQK N02228 VH PGQGLEWIGYIDPFNDDTNYNVKFKGKATLTSDTSSSIAYME LSSLTSEDSSFYYCARGTSATLDYWGHGTTLTVSS SEQ ID DLL3#7 EIVLTQSPDTLSLSPGERATLSCRASQSVNSNFLAWYQQKPG NO:229 scFab QTPRLLIFGTSSRATGIPDRFSGSGSGTDFTLTISRLEPEDFAV YYCQQYGSSPWTFGQGTKVEIRRTVAAPSVFIFPPSDEQLKS GTASVVCLLNNFYPREAKVQWKVDNALQSGNSQESVTEQD SKDSTYSLSSTLTLSKADYEKHKVYACEVTHQGLSSPVTKSF NRGECGGGGSEGKSSGSGSESKSTEGKSSGSGSESKSTGGGG SQVQLVESGGGVVQPGRSLRLSCAASGFTFSSYGMFWVRQ APGKGLEWVAVIWLDGDDEDYVDSVKGRFTISRDDSKNTL YLQMNSLRVDDTAIYYCARVLDYWGQGTLVTVSSASTKGP SVFPLAPSSKSTSGGTAALGCLVKDYFPEPVTVSWNSGALTS GVHTFPAVLQSSGLYSLSSVVTVPSSSLGTQTYICNVNHKPS NTKVDKRVEPKSC SEQ ID DLL3#8 DIVMAQSPDSLAVSLGERATINCKSSQSVLDTSNNKNYLVW NO:230 scFab YQQKPGQPPKLLIYWASTRESGVPDRFSGSGSGTDFTLTISSL QAEDVAVYYCQHYYNSPYTFGQGTKLEIKRTVAAPSVFIFPP SGTASVVCLLNNFYPREAKVQWKVDNALQSGNSQ ESVTEQDSKDSTYSLSSTLTLSKADYEKHKVYACEVTHQGL —134— SEQ ID Brief Sequence Number description of sequence SSPVTKSFNRGECGGGGSEGKSSGSGSESKSTEGKSSGSGSES KSTGGGGSQVQLVQSGAEVKKPGASVKVSCKASGYTFTDY YMHWVRQAPGQGLEWMGWINPNSGGTNYEQKFQGRVTM STAYMELNRLRSDDTAVYYCTRDAVVIPMDYWGQ GTLVTVSSASTKGPSVFPLAPSSKSTSGGTAALGCLVKDYFP EPVTVSWNSGALTSGVHTFPAVLQSSGLYSLSSVVTVPSSSL GTQTYICNVNHKPSNTKVDKRVEPKSC SEQ ID DLL3#9 EIVLTQSPGTLSLSPGERATLSCRASQSISRSYLAWYQQKPGQ NO:231 scFab APRLLIYGASSRATGIPDRFSGSGSGTDFTLTISRLEPEDFAVY YCQQYGTSPITFGQGTRLEIKRTVAAPSVFIFPPSDEQLKSGT ASWCLLNNFYPREAKVQWKVDNALQSGNSQESVTEQDSK DSTYSLSSTLTLSKADYEKHKVYACEVTHQGLSSPVTKSFNR GECGGGGSEGKSSGSGSESKSTEGKSSGSGSESKSTGGGGSQ VHLQESGPGLVKPSETLSLTCTVSGGSISSYYWSWIRQTPGK GLDWIGYRYYSGNTNYNPSLKSRVTISLDMSNNQFSLKLSS VTAADTAIYYCASIGVAGFYFDYWGQGTLVTVSSASTKGPS VFPLAPSSKSTSGGTAALGCLVKDYFPEPVTVSWNSGALTSG VHTFPAVLQSSGLYSLSSVVTVPSSSLGTQTYICNVNHKPSN VEPKSC SEQ ID DLL3#10 EIVLTQSPGTLSLSPGERATLSCRASQSLNSIFLAWYQQKPGQ NO:232 scFab APWLLIYGASSRATGIPDRFSGSGSGTDFTLTISRLEPEDFAV YFCQQYGGSMNTFGQGTKLEIKRTVAAPSVFIFPPSDEQLKS GTASVVCLLNNFYPREAKVQWKVDNALQSGNSQESVTEQD SKDSTYSLSSTLTLSKADYEKHKVYACEVTHQGLSSPVTKSF NRGECGGGGSEGKSSGSGSESKSTEGKSSGSGSESKSTGGGG SQVQLVQSGAEVKKPGASVKVSCKASGYTFTGYYMHWVR LAPGQGLEWMGWINPNSGGTIFAQRFQGRVTMTRDTSISTV YMDLNRLRSDDTAVYYCARDFGDTVGNAFDIWGQGTMVT VSSASTKGPSVFPLAPSSKSTSGGTAALGCLVKDYFPEPVTV LTSGVHTFPAVLQSSGLYSLSSVVTVPSSSLGTQTY ICNVNHKPSNTKVDKRVEPKSC SEQ ID DLL3#1 1 QIVLTQSPAIMSASPGEKVTITCSASSSVTYIHWFQQNPGTSP N02233 scFab KLWIYRTSYLASGVPARFSGSGSGTSYSLTISRMEAEDAATY —135— SEQ ID Brief Sequence Number description of sequence YCQQRSSYPRTFGGGTKLEIKRTVAAPSVFIFPPSDEQLKSGT ASVVCLLNNFYPREAKVQWKVDNALQSGNSQESVTEQDSK DSTYSLSSTLTLSKADYEKHKVYACEVTHQGLSSPVTKSFNR GECGGGGSEGKSSGSGSESKSTEGKSSGSGSESKSTGGGGSQ VQLQQPGAELVQPGSSVKMSCKASGYAFSDYWITWVKQRP IGDIYPGSGSTKSSEKFKNKATLTADTSSSKAYIQF SSLTPEDSAVYYCVSLYYYGSHYLDTWGQGTTLTVSSASTK LAPSSKSTSGGTAALGCLVKDYFPEPVTVSWNSGAL TSGVHTFPAVLQSSGLYSLSSVVTVPSSSLGTQTYICNVNHK PSNTKVDKRVEPKSC SEQ ID 2 QVVLTQSPAIMSASPGEKVTITCSASSSVTYIHWFQQNPGTSP N0:234 scFab KLWIYRTSYLASGVPARFSGSGSGTSYSLTISRMEAEDAATY YCQQRSSYPRTFGGGTKLEIKRTVAAPSVFIFPPSDEQLKSGT ASVVCLLNNFYPREAKVQWKVDNALQSGNSQESVTEQDSK DSTYSLSSTLTLSKADYEKHKVYACEVTHQGLSSPVTKSFNR GECGGGGSEGKSSGSGSESKSTEGKSSGSGSESKSTGGGGSQ VQLQQPGAEFVQPGSSVKMSCKASGYAFSDYWITWVKQRP GQGLEWIGDIYPGSGSTKSSEKFKNRATLTADTSSSTAYIQFS SLTPEDSAVYYCVSLYYYGSYYLDTWGQGTTLTVSSASTKG PSVFPLAPSSKSTSGGTAALGCLVKDYFPEPVTVSWNSGALT SGVHTFPAVLQSSGLYSLSSVVTVPSSSLGTQTYICNVNHKP SNTKVDKRVEPKSC SEQ ID DLL3#13 DVLMTQTPLSLPVSLGDQAAISCRSSQSIVHSNGNTYLEWYL NO:235 scFab QKPGQSPKVLIYKVSNRFSGVPDRFSGSGSGTDFTLKISRVE AEDLGVYYCFQGSHVPYTFGGGTKLEIKRTVAAPSVFIFPPS GTASVVCLLNNFYPREAKVQWKVDNALQSGNSQE SVTEQDSKDSTYSLSSTLTLSKADYEKHKVYACEVTHQGLS SPVTKSFNRGECGGGGSEGKSSGSGSESKSTEGKSSGSGSES KSTGGGGSQIQLVQSGPELKKPGETVKISCKASGYTFTNYGV TWVKQAPGKGLKWMGWINTYSGAPTYADDFNGRFALSLE TSASTAYLQINNLKNEDTATYFCARLDDYDLYYFDYWGQG TALTVSSASTKGPSVFPLAPSSKSTSGGTAALGCLVKDYFPE PVTVSWNSGALTSGVHTFPAVLQSSGLYSLSSVVTVPSSSLG —136— SEQ ID Brief ce Number description of sequence TQTYICNVNHKPSNTKVDKRVEPKSC SEQ ID DLL3#14 DIVLTQSPASLSVSLGQRATISCKASQSVDYDGDSYMNWYQ NO:236 scFab QKPGQPPKLLIYAASTLESGIPARFSGSGSGTDFTLNIHPVEEE DAATYYCQQSDEDPWTFGGGTKLEIKRTVAAPSVFIFPPSDE QLKSGTASWCLLNNFYPREAKVQWKVDNALQSGNSQESV TEQDSKDSTYSLSSTLTLSKADYEKHKVYACEVTHQGLSSP VTKSFNRGECGGGGSEGKSSGSGSESKSTEGKSSGSGSESKS TGGGGSQIQLQQSGPELVKPGVKISCKASGYTFTDYYIHWM KQRPGQGLEWIGYIYPGNSYTAYNQKFKDKATLTADNPSST AYMQLSSLTSEDSAVYFCARSGGSAMDYWGQGTSVTVSSA STKGPSVFPLAPSSKSTSGGTAALGCLVKDYFPEPVTVSWNS GALTSGVHTFPAVLQSSGLYSLSSVVTVPSSSLGTQTYICNV NHKPSNTKVDKRVEPKSC SEQ ID DLL3#15 DIVLTQSPASLAVSLGQRATISCKASQSVDYDGDSYLNWYQ NO:237 scFab QKPGQPPKLLIYAASNLESGIPARFSGSGSGTDFTLNIHPVEE EDAATYYCQQSSEDPRTFGGGTKLEIKRTVAAPSVFIFPPSDE QLKSGTASWCLLNNFYPREAKVQWKVDNALQSGNSQESV DSTYSLSSTLTLSKADYEKHKVYACEVTHQGLSSP VTKSFNRGECGGGGSEGKSSGSGSESKSTEGKSSGSGSESKS TGGGGSQIQLVQSGPELKKPGETVKISCKASGYTFTNYGMN WVKQAPGKGLKWMGWINTYTGEPTYADDFKGRFAFSLETS QINNLKNEDMATYFCTKFHFSSNGDAMDNWGQGT SVTVSSASTKGPSVFPLAPSSKSTSGGTAALGCLVKDYFPEP VTVSWNSGALTSGVHTFPAVLQSSGLYSLSSVVTVPSSSLGT QTYICNVNHKPSNTKVDKRVEPKSC SEQ ID DLL3#16 SPSTLSASVGDRVTITCRASQSVSDWLAWYQQKPG NO:23 8 scFab KAPKFLIYRASSLESGVPSRFSGSGSGTEFTLTISSLQPADFAT YYCQLYNSYSPTFGQGTKVEIKRTVAAPSVFIFPPSDEQLKS GTASVVCLLNNFYPREAKVQWKVDNALQSGNSQESVTEQD SKDSTYSLSSTLTLSKADYEKHKVYACEVTHQGLSSPVTKSF NRGECGGGGSEGKSSGSGSESKSTEGKSSGSGSESKSTGGGG SEVHLVESGGGLVQPGGSLRLSCAASGFTFSSYWMTWVRQ APGKGLEWVANIKEDGSEKYYVDSVKGRFTISRDNAKNSLY —137— SEQ ID Brief Sequence Number description of sequence LQMNSLRAEDTALYYCARDWGYFDYWGQGTLVTVSSAST KGPSVFPLAPSSKSTSGGTAALGCLVKDYFPEPVTVSWNSG ALTSGVHTFPAVLQSSGLYSLSSVVTVPSSSLGTQTYICNVN HKPSNTKVDKRVEPKSC SEQ ID DLL3#17 DIQMTQSPASLAASVGETVTITCRASENIYYSLAWYQQKQG NO:239 scFab KSPQLLIYNTNSLEDGVPSRFSGSGSGTQYSMKINSMQPEDT ATYFCKQAYDFPLTFGAGTKLELKRTVAAPSVFIFPPSDEQL KSGTASVVCLLNNFYPREAKVQWKVDNALQSGNSQESVTE TYSLSSTLTLSKADYEKHKVYACEVTHQGLSSPVT KSFNRGECGGGGSEGKSSGSGSESKSTEGKSSGSGSESKSTG GGGSQIQLQQSGPEVVKPGASVKISCKASGYTFISYYIHWVK QRPGQGLEWIGWIYPGDGSTNNNEKFKGKTTLTADKSSSTA YMLLSSLTSEDSAVYFCARGEGNAMDDWGQGTSVTVSSAS FPLAPSSKSTSGGTAALGCLVKDYFPEPVTVSWNSG HTFPAVLQSSGLYSLSSVVTVPSSSLGTQTYICNVN HKPSNTKVDKRVEPKSC SEQ ID 8 DIQMTQSPASLAASVGETVTITCRASENIYYSLAWYQQKQG NO:240 scFab KSPQLLIYNANSLEDGVPSRFSGSGSGTQYSMKINNMQPEDT ATYFCKQAYDVPLTFGAGTKLELKRTVAAPSVFIFPPSDEQL KSGTASVVCLLNNFYPREAKVQWKVDNALQSGNSQESVTE QDSKDSTYSLSSTLTLSKADYEKHKVYACEVTHQGLSSPVT KSFNRGECGGGGSEGKSSGSGSESKSTEGKSSGSGSESKSTG GGGSQVQLQQSGPDLVKPGASVKMSCEASGYTFTAYFIHW VKQKPGQGLEWIGYIDPFNDDTNYNVKFKGKATLTSDTSSSI AYMELSSLTSEDSSFYYCARGTSATLDYWGHGTTLTVSSAS TKGPSVFPLAPSSKSTSGGTAALGCLVKDYFPEPVTVSWNSG ALTSGVHTFPAVLQSSGLYSLSSVVTVPSSSLGTQTYICNVN HKPSNTKVDKRVEPKSC SEQ ID DLL3#7 EIVLTQSPDTLSLSPGERATLSCRASQSVNSNFLAWYQQKPG NO:24 1 chain QTPRLLIFGTSSRATGIPDRFSGSGSGTDFTLTISRLEPEDFAV YYCQQYGSSPWTFGQGTKVEIRRTVAAPSVFIFPPSDEQLKS GTASVVCLLNNFYPREAKVQWKVDNALQSGNSQESVTEQD SKDSTYSLSSTLTLSKADYEKHKVYACEVTHQGLSSPVTKSF —138— SEQ ID Brief Sequence Number description of sequence NRGECGGGGSEGKSSGSGSESKSTEGKSSGSGSESKSTGGGG SQVQLVESGGGVVQPGRSLRLSCAASGFTFSSYGMFWVRQ APGKGLEWVAVIWLDGDDEDYVDSVKGRFTISRDDSKNTL YLQMNSLRVDDTAIYYCARVLDYWGQGTLVTVSSASTKGP SVFPLAPSSKSTSGGTAALGCLVKDYFPEPVTVSWNSGALTS GVHTFPAVLQSSGLYSLSSVVTVPSSSLGTQTYICNVNHKPS NTKVDKRVEPKSCDKTHTCPPCPAPEAAGGPSVFLFPPKPKD TLMISRTPEVTCVVVDVSHEDPEVKFNWYVDGVEVHNAKT KPREEQYNSTYRVVSVLTVLHQDWLNGKEYKCKVSNKALP ISKAKGQPREPQVYTLPPSREEMTKNQVSLWCLVKG FYPSDIAVEWESNGQPENNYKTTPPVLDSDGSFFLYSKLTVD KSRWQQGNVFSCSVMHEALHNHYTQKSLSLSPG SEQ ID DLL3#8 DIVMAQSPDSLAVSLGERATINCKSSQSVLDTSNNKNYLVW NO:242 chain YQQKPGQPPKLLIYWASTRESGVPDRFSGSGSGTDFTLTISSL QAEDVAVYYCQHYYNSPYTFGQGTKLEIKRTVAAPSVFIFPP SGTASVVCLLNNFYPREAKVQWKVDNALQSGNSQ ESVTEQDSKDSTYSLSSTLTLSKADYEKHKVYACEVTHQGL SSPVTKSFNRGECGGGGSEGKSSGSGSESKSTEGKSSGSGSES KSTGGGGSQVQLVQSGAEVKKPGASVKVSCKASGYTFTDY YMHWVRQAPGQGLEWMGWINPNSGGTNYEQKFQGRVTM TRDTSISTAYMELNRLRSDDTAVYYCTRDAVVIPMDYWGQ GTLVTVSSASTKGPSVFPLAPSSKSTSGGTAALGCLVKDYFP EPVTVSWNSGALTSGVHTFPAVLQSSGLYSLSSVVTVPSSSL GTQTYICNVNHKPSNTKVDKRVEPKSCDKTHTCPPCPAPEA AGGPSVFLFPPKPKDTLMISRTPEVTCVWDVSHEDPEVKFN WYVDGVEVHNAKTKPREEQYNSTYRVVSVLTVLHQDWLN KVSNKALPAPIEKTISKAKGQPREPQVYTLPPSREE MTKNQVSLWCLVKGFYPSDIAVEWESNGQPENNYKTTPPV LDSDGSFFLYSKLTVDKSRWQQGNVFSCSVMHEALHNHYT QKSLSLSPG SEQ ID DLL3#9 EIVLTQSPGTLSLSPGERATLSCRASQSISRSYLAWYQQKPGQ NO:243 chain APRLLIYGASSRATGIPDRFSGSGSGTDFTLTISRLEPEDFAVY YCQQYGTSPITFGQGTRLEIKRTVAAPSVFIFPPSDEQLKSGT —139— SEQ ID Brief Sequence Number description of sequence LNNFYPREAKVQWKVDNALQSGNSQESVTEQDSK DSTYSLSSTLTLSKADYEKHKVYACEVTHQGLSSPVTKSFNR GECGGGGSEGKSSGSGSESKSTEGKSSGSGSESKSTGGGGSQ VHLQESGPGLVKPSETLSLTCTVSGGSISSYYWSWIRQTPGK GLDWIGYRYYSGNTNYNPSLKSRVTISLDMSNNQFSLKLSS VTAADTAIYYCASIGVAGFYFDYWGQGTLVTVSSASTKGPS VFPLAPSSKSTSGGTAALGCLVKDYFPEPVTVSWNSGALTSG VHTFPAVLQSSGLYSLSSVVTVPSSSLGTQTYICNVNHKPSN TKVDKRVEPKSCDKTHTCPPCPAPEAAGGPSVFLFPPKPKDT LMISRTPEVTCVVVDVSHEDPEVKFNWYVDGVEVHNAKTK PREEQYNSTYRVVSVLTVLHQDWLNGKEYKCKVSNKALPA PIEKTISKAKGQPREPQVYTLPPSREEMTKNQVSLWCLVKGF VEWESNGQPENNYKTTPPVLDSDGSFFLYSKLTVDK SRWQQGNVFSCSVMHEALHNHYTQKSLSLSPG SEQ ID 0 EIVLTQSPGTLSLSPGERATLSCRASQSLNSIFLAWYQQKPGQ NO:244 chain APWLLIYGASSRATGIPDRFSGSGSGTDFTLTISRLEPEDFAV YFCQQYGGSMNTFGQGTKLEIKRTVAAPSVFIFPPSDEQLKS GTASVVCLLNNFYPREAKVQWKVDNALQSGNSQESVTEQD SKDSTYSLSSTLTLSKADYEKHKVYACEVTHQGLSSPVTKSF NRGECGGGGSEGKSSGSGSESKSTEGKSSGSGSESKSTGGGG SQVQLVQSGAEVKKPGASVKVSCKASGYTFTGYYMHWVR LAPGQGLEWMGWINPNSGGTIFAQRFQGRVTMTRDTSISTV YMDLNRLRSDDTAVYYCARDFGDTVGNAFDIWGQGTMVT VSSASTKGPSVFPLAPSSKSTSGGTAALGCLVKDYFPEPVTV SWNSGALTSGVHTFPAVLQSSGLYSLSSVVTVPSSSLGTQTY ICNVNHKPSNTKVDKRVEPKSCDKTHTCPPCPAPEAAGGPS VFLFPPKPKDTLMISRTPEVTCVVVDVSHEDPEVKFNWYVD GVEVHNAKTKPREEQYNSTYRVVSVLTVLHQDWLNGKEY KCKVSNKALPAPIEKTISKAKGQPREPQVYTLPPSREEMTKN QVSLWCLVKGFYPSDIAVEWESNGQPENNYKTTPPVLDSDG SFFLYSKLTVDKSRWQQGNVFSCSVMHEALHNHYTQKSLSL SEQ ID DLL3#1 1 QIVLTQSPAIMSASPGEKVTITCSASSSVTYIHWFQQNPGTSP —140— SEQ ID Brief Sequence Number description of sequence NO:245 chain TSYLASGVPARFSGSGSGTSYSLTISRMEAEDAATY YCQQRSSYPRTFGGGTKLEIKRTVAAPSVFIFPPSDEQLKSGT ASWCLLNNFYPREAKVQWKVDNALQSGNSQESVTEQDSK DSTYSLSSTLTLSKADYEKHKVYACEVTHQGLSSPVTKSFNR GECGGGGSEGKSSGSGSESKSTEGKSSGSGSESKSTGGGGSQ VQLQQPGAELVQPGSSVKMSCKASGYAFSDYWITWVKQRP GQGLEWIGDIYPGSGSTKSSEKFKNKATLTADTSSSKAYIQF SSLTPEDSAVYYCVSLYYYGSHYLDTWGQGTTLTVSSASTK GPSVFPLAPSSKSTSGGTAALGCLVKDYFPEPVTVSWNSGAL TSGVHTFPAVLQSSGLYSLSSWTVPSSSLGTQTYICNVNHK PSNTKVDKRVEPKSCDKTHTCPPCPAPEAAGGPSVFLFPPKP KDTLMISRTPEVTCVVVDVSHEDPEVKFNWYVDGVEVHNA KTKPREEQYNSTYRVVSVLTVLHQDWLNGKEYKCKVSNKA LPAPIEKTISKAKGQPREPQVYTLPPSREEMTKNQVSLWCLV KGFYPSDIAVEWESNGQPENNYKTTPPVLDSDGSFFLYSKLT VDKSRWQQGNVFSCSVMHEALHNHYTQKSLSLSPG SEQ ID DLL3#12 QVVLTQSPAIMSASPGEKVTITCSASSSVTYIHWFQQNPGTSP NO:246 chain KLWIYRTSYLASGVPARFSGSGSGTSYSLTISRMEAEDAATY YCQQRSSYPRTFGGGTKLEIKRTVAAPSVFIFPPSDEQLKSGT ASVVCLLNNFYPREAKVQWKVDNALQSGNSQESVTEQDSK DSTYSLSSTLTLSKADYEKHKVYACEVTHQGLSSPVTKSFNR GECGGGGSEGKSSGSGSESKSTEGKSSGSGSESKSTGGGGSQ VQLQQPGAEFVQPGSSVKMSCKASGYAFSDYWITWVKQRP GQGLEWIGDIYPGSGSTKSSEKFKNRATLTADTSSSTAYIQFS SLTPEDSAVYYCVSLYYYGSYYLDTWGQGTTLTVSSASTKG PSVFPLAPSSKSTSGGTAALGCLVKDYFPEPVTVSWNSGALT SGVHTFPAVLQSSGLYSLSSVVTVPSSSLGTQTYICNVNHKP KRVEPKSCDKTHTCPPCPAPEAAGGPSVFLFPPKPK DTLMISRTPEVTCVVVDVSHEDPEVKFNWYVDGVEVHNAK TKPREEQYNSTYRVVSVLTVLHQDWLNGKEYKCKVSNKAL PAPIEKTISKAKGQPREPQVYTLPPSREEMTKNQVSLWCLVK GFYPSDIAVEWESNGQPENNYKTTPPVLDSDGSFFLYSKLTV DKSRWQQGNVFSCSVMHEALHNHYTQKSLSLSPG —141— SEQ ID Brief Sequence Number description of sequence SEQ ID DLL3#13 DVLMTQTPLSLPVSLGDQAAISCRSSQSIVHSNGNTYLEWYL NO:247 chain QKPGQSPKVLIYKVSNRFSGVPDRFSGSGSGTDFTLKISRVE AEDLGVYYCFQGSHVPYTFGGGTKLEIKRTVAAPSVFIFPPS DEQLKSGTASVVCLLNNFYPREAKVQWKVDNALQSGNSQE SVTEQDSKDSTYSLSSTLTLSKADYEKHKVYACEVTHQGLS SPVTKSFNRGECGGGGSEGKSSGSGSESKSTEGKSSGSGSES KSTGGGGSQIQLVQSGPELKKPGETVKISCKASGYTFTNYGV TWVKQAPGKGLKWMGWINTYSGAPTYADDFNGRFALSLE TSASTAYLQINNLKNEDTATYFCARLDDYDLYYFDYWGQG TALTVSSASTKGPSVFPLAPSSKSTSGGTAALGCLVKDYFPE PVTVSWNSGALTSGVHTFPAVLQSSGLYSLSSVVTVPSSSLG TQTYICNVNHKPSNTKVDKRVEPKSCDKTHTCPPCPAPEAA GGPSVFLFPPKPKDTLMISRTPEVTCVVVDVSHEDPEVKFNW YVDGVEVHNAKTKPREEQYNSTYRVVSVLTVLHQDWLNG KEYKCKVSNKALPAPIEKTISKAKGQPREPQVYTLPPSREEM TKNQVSLWCLVKGFYPSDIAVEWESNGQPENNYKTTPPVLD SDGSFFLYSKLTVDKSRWQQGNVFSCSVMHEALHNHYTQK SLSLSPG SEQ ID DLL3#14 DIVLTQSPASLSVSLGQRATISCKASQSVDYDGDSYMNWYQ NO:248 chain PKLLIYAASTLESGIPARFSGSGSGTDFTLNIHPVEEE DAATYYCQQSDEDPWTFGGGTKLEIKRTVAAPSVFIFPPSDE QLKSGTASWCLLNNFYPREAKVQWKVDNALQSGNSQESV TEQDSKDSTYSLSSTLTLSKADYEKHKVYACEVTHQGLSSP VTKSFNRGECGGGGSEGKSSGSGSESKSTEGKSSGSGSESKS TGGGGSQIQLQQSGPELVKPGVKISCKASGYTFTDYYIHWM GLEWIGYIYPGNSYTAYNQKFKDKATLTADNPSST AYMQLSSLTSEDSAVYFCARSGGSAMDYWGQGTSVTVSSA STKGPSVFPLAPSSKSTSGGTAALGCLVKDYFPEPVTVSWNS GALTSGVHTFPAVLQSSGLYSLSSVVTVPSSSLGTQTYICNV NHKPSNTKVDKRVEPKSCDKTHTCPPCPAPEAAGGPSVFLFP LMISRTPEVTCVVVDVSHEDPEVKFNWYVDGVEV PREEQYNSTYRVVSVLTVLHQDWLNGKEYKCKVS NKALPAPIEKTISKAKGQPREPQVYTLPPSREEMTKNQVSLW —142— SEQ ID Brief Sequence Number description of sequence CLVKGFYPSDIAVEWESNGQPENNYKTTPPVLDSDGSFFLYS KLTVDKSRWQQGNVFSCSVMHEALHNHYTQKSLSLSPG SEQ ID DLL3#15 DIVLTQSPASLAVSLGQRATISCKASQSVDYDGDSYLNWYQ NO:249 chain QKPGQPPKLLIYAASNLESGIPARFSGSGSGTDFTLNIHPVEE EDAATYYCQQSSEDPRTFGGGTKLEIKRTVAAPSVFIFPPSDE QLKSGTASVVCLLNNFYPREAKVQWKVDNALQSGNSQESV TEQDSKDSTYSLSSTLTLSKADYEKHKVYACEVTHQGLSSP VTKSFNRGECGGGGSEGKSSGSGSESKSTEGKSSGSGSESKS TGGGGSQIQLVQSGPELKKPGETVKISCKASGYTFTNYGMN WVKQAPGKGLKWMGWTNTYTGEPTYADDFKGRFAFSLETS ASTAYLQINNLKNEDMATYFCTKFHFSSNGDAMDNWGQGT SVTVSSASTKGPSVFPLAPSSKSTSGGTAALGCLVKDYFPEP VTVSWNSGALTSGVHTFPAVLQSSGLYSLSSVVTVPSSSLGT QTYICNVNHKPSNTKVDKRVEPKSCDKTHTCPPCPAPEAAG GPSVFLFPPKPKDTLMISRTPEVTCVVVDVSHEDPEVKFNWY HNAKTKPREEQYNSTYRVVSVLTVLHQDWLNGK EYKCKVSNKALPAPIEKTISKAKGQPREPQVYTLPPSREEMT KNQVSLWCLVKGFYPSDIAVEWESNGQPENNYKTTPPVLDS DGSFFLYSKLTVDKSRWQQGNVFSCSVMHEALHNHYTQKS LSLSPG SEQ ID DLL3#16 DIQMTQSPSTLSASVGDRVTITCRASQSVSDWLAWYQQKPG NO:250 chain KAPKFLIYRASSLESGVPSRFSGSGSGTEFTLTISSLQPADFAT YYCQLYNSYSPTFGQGTKVEIKRTVAAPSVFIFPPSDEQLKS GTASVVCLLNNFYPREAKVQWKVDNALQSGNSQESVTEQD SLSSTLTLSKADYEKHKVYACEVTHQGLSSPVTKSF NRGECGGGGSEGKSSGSGSESKSTEGKSSGSGSESKSTGGGG SEVHLVESGGGLVQPGGSLRLSCAASGFTFSSYWMTWVRQ APGKGLEWVANIKEDGSEKYYVDSVKGRFTISRDNAKNSLY LQMNSLRAEDTALYYCARDWGYFDYWGQGTLVTVSSAST KGPSVFPLAPSSKSTSGGTAALGCLVKDYFPEPVTVSWNSG ALTSGVHTFPAVLQSSGLYSLSSVVTVPSSSLGTQTYICNVN KVDKRVEPKSCDKTHTCPPCPAPEAAGGPSVFLFPP KPKDTLMISRTPEVTCVVVDVSHEDPEVKFNWYVDGVEVH —143— SEQ ID Brief Sequence Number description of sequence NAKTKPREEQYNSTYRVVSVLTVLHQDWLNGKEYKCKVSN KALPAPIEKTISKAKGQPREPQVYTLPPSREEMTKNQVSLWC PSDIAVEWESNGQPENNYKTTPPVLDSDGSFFLYSK LTVDKSRWQQGNVFSCSVMHEALHNHYTQKSLSLSPG SEQ ID DLL3#17 DIQMTQSPASLAASVGETVTITCRASENIYYSLAWYQQKQG N0:251 chain KSPQLLIYNTNSLEDGVPSRFSGSGSGTQYSMKJNSMQPEDT ATYFCKQAYDFPLTFGAGTKLELKRTVAAPSVFIFPPSDEQL KSGTASVVCLLNNFYPREAKVQWKVDNALQSGNSQESVTE QDSKDSTYSLSSTLTLSKADYEKHKVYACEVTHQGLSSPVT ECGGGGSEGKSSGSGSESKSTEGKSSGSGSESKSTG GGGSQIQLQQSGPEVVKPGASVKISCKASGYTFISYYIHWVK QRPGQGLEWIGWIYPGDGSTNNNEKFKGKTTLTADKSSSTA YMLLSSLTSEDSAVYFCARGEGNAMDDWGQGTSVTVSSAS TKGPSVFPLAPSSKSTSGGTAALGCLVKDYFPEPVTVSWNSG ALTSGVHTFPAVLQSSGLYSLSSVVTVPSSSLGTQTYICNVN HKPSNTKVDKRVEPKSCDKTHTCPPCPAPEAAGGPSVFLFPP KPKDTLMISRTPEVTCVVVDVSHEDPEVKFNWYVDGVEVH NAKTKPREEQYNSTYRVVSVLTVLHQDWLNGKEYKCKVSN KALPAPIEKTISKAKGQPREPQVYTLPPSREEMTKNQVSLWC LVKGFYPSDIAVEWESNGQPENNYKTTPPVLDSDGSFFLYSK LTVDKSRWQQGNVFSCSVMHEALHNHYTQKSLSLSPG SEQ ID DLL3#18 DIQMTQSPASLAASVGETVTITCRASENIYYSLAWYQQKQG N0:252 chain KSPQLLIYNANSLEDGVPSRFSGSGSGTQYSMKINNMQPEDT ATYFCKQAYDVPLTFGAGTKLELKRTVAAPSVFIFPPSDEQL KSGTASVVCLLNNFYPREAKVQWKVDNALQSGNSQESVTE QDSKDSTYSLSSTLTLSKADYEKHKVYACEVTHQGLSSPVT ECGGGGSEGKSSGSGSESKSTEGKSSGSGSESKSTG GGGSQVQLQQSGPDLVKPGASVKMSCEASGYTFTAYFIHW VKQKPGQGLEWIGYIDPFNDDTNYNVKFKGKATLTSDTSSSI AYMELSSLTSEDSSFYYCARGTSATLDYWGHGTTLTVSSAS TKGPSVFPLAPSSKSTSGGTAALGCLVKDYFPEPVTVSWNSG ALTSGVHTFPAVLQSSGLYSLSSVVTVPSSSLGTQTYICNVN HKPSNTKVDKRVEPKSCDKTHTCPPCPAPEAAGGPSVFLFPP —144— SEQ ID Brief Sequence Number description of sequence KPKDTLMISRTPEVTCVVVDVSHEDPEVKFNWYVDGVEVH NAKTKPREEQYNSTYRVVSVLTVLHQDWLNGKEYKCKVSN KALPAPIEKTISKAKGQPREPQVYTLPPSREEMTKNQVSLWC LVKGFYPSDIAVEWESNGQPENNYKTTPPVLDSDGSFFLYSK LTVDKSRWQQGNVFSCSVMHEALHNHYTQKSLSLSPG SEQ ID Constant ASTKGPSVFPLAPSSKSTSGGTAALGCLVKDYFPEPVTVSWN NO:253 region of SGALTSGVHTFPAVLQSSGLYSLSSVVTVPSSSLGTQTYICNV heavy chain NHKPSNTKVDKRVEPKSC CH 1 SEQ ID Mouse AKTTAPSVYPLAPVCGDTTGSSVTLGCLVKGYFPEPVTLTW NO:254 IgGZa HC NSGSLSSGVHTFPAVLQSDLYTLSSSVTVTSSTWPSQSITCNV AHPASSTKVDKKIEPRGPTIKPCPPCKCPAPNLLGGPSVFIFPP KIKDVLMISLSPIVTCVVVDVSEDDPDVQISWFVNNVEVHTA QTQTHREDYNSTLRVVSALPIQHQDWMSGKEFKCKVNNKD LPAPIERTISKPKGSVRAPQVYVLPPPEEEMTKKQVTLTCMV DIYVEWTNNGKTELNYKNTEPVLDSDGSYFMYSK NWVERNSYSCSVVHEGLHNHHTTKSFSRTPGK SEQ ID Mouse RADAAPTVSIFPPSSEQLTSGGASVVCFLNNFYPKDINVKWK NO:255 Kappa LC IDGSERQNGVLNSWTDQDSKDSTYSMSSTLTLTKDEYERHN SYTCEATHKTSTSPIVKSFNRNEC *Underlined sequence indicates hinge region Example 3: Production of recombinant proteins ' Human DLL3-His A cell line to e Human DLL3 -His was generated using HEK—293 cells (Thermo Fisher), the Lenti-X Lentiviral System (Clontech), and plasmid encoding Human DLL3- His (human DLL3 Accession No. Q9NYJ7 huDLL3—His: SEQ ID NO:106). For expression, cells were cultured and ed at 37C, 5% C02, and shaking at 140 rpm. On Day 0 of expression, cell were pelleted and re—suspended in Expi 293 media. On day 3 of sion the conditioned culture supernatant was ted by pelleting the cells for 40 s at 4700 rpm. Protease inhibitors were added to the biomass before puriﬁcation.

Expression was conﬁrmed by western blot. The conditioned culture supernatant was —145— ed with 0.5 mM TCEP, 0.02% CHAPS, 10 mM ole. Puriﬁcation was carried out on a HisTrap Ni excel column and Buffer A: 50 mM MES, 50 mM NaCl, 0.5 mM TCEP, 0.02% CHAPS, pH 6.5. The protein of interest was eluted in Buffer A supplemented with 0.5M Imidazole, pH 8.5, using an elution gradient from 20 mM imidazole to 500 mM imidazole. The pooled fractions were dialyzed in buffer250 mM MES, 50 mM NaCl, 1 mM TCEP, 0.02% CHAPS, 0.2M Arginine, 3% glycerol, pH 6.5.

The puriﬁed material was qualiﬁed by mass spectrometry and analytical ultra- fugation.

° Cyno DLL3-His A cell line to produce Cyno DLL3 -His was generated using HEK-293 cells (Thermo Fisher), the Lenti-X Lentiviral System (Clontech), and plasmid encoding Cyno DLL3 -His (Cyno DLL3 Accession No: XM_0055 89196.1 (RefSeq); Cyno DLL3 —His: SEQ ID NO:107). For expression, cells were cultured and expanded at 37°C, 5% CO2, and shaking at 140 rpm. On Day 0 of expression, cells were pelleted and re-suspended in Expi 293 media. On day 3 of expression the conditioned culture supernatant was harvested by ing the cells for 40 minutes at 4700 rpm. Protease inhibitors were added to the biomass before ation. Expression was conﬁrmed by western blot. The conditioned culture supernatant was ed with 0.5 mM TCEP, 0.02% CHAPS, 10 mM imidazole. ation was carried out on a HisTrap Ni excel column and Buffer A: 50 mM MES, 50 mM NaCl, 0.5 mM TCEP, 0.02% CHAPS, pH 6.5. The protein of interest was eluted in Buffer A supplemented with 0.5M Imidazole, pH 8.5, using an elution gradient from 20 mM imidazole to 500 mM imidazole. The pooled fractions were dialyzed in buffer: 50 mM MES, 50 mM NaCl, 1 mM TCEP, 0.02% CHAPS, 0.2M Arginine, 3% glycerol, pH 6.5.

The puriﬁed material was qualiﬁed by mass spectrometry and analytical ultra- fugation. 0 Human CD3 E+G Hch-6xHis (E+G indicates sy subunits) A cell line to e Human CD3 E+G Hch—6xHis was generated using HEK—293 cells (Thermo Fisher), the Lenti—X Lentiviral System (Clontech), and plasmid encoding Human CD3 E+G Hch—6xHis (human CD3E Accession No: P07766; human CD3E+G-Hch—His: SEQ ID NO: 108). For sion, cells were cultured and expanded in Freestyle 293 media, at 37C, humidiﬁed 8% CO2 environment, and shaking at 135 rpm. The conditioned culture supernatant was harvested at Day 6 by centrifugation for 30 minutes at 9300xg. sion was monitored by SDS—PAGE and Western Blotting. The conditioned culture — 1 46— WO 34220 supernatant was adjusted with 0.2M Sucrose, 5% ol, 0.01% CHAPS, and 10mM Imidazole. The pH was then adjusted to 7.2. Puriﬁcation was carried out in a two—step process: afﬁnity puriﬁcation using Ni/NTA resin (overnight incubation at 4C, and elution with 250mM Imidazole); followed by size—exclusion chromatography on a Superdex 200 column in destination buffer PBS with 0.2M Sucrose, 5% glycerol, 0.01% CHAPS, lmM TCEP, pH7.2. The pooled al was concentrated using a 10K MWCO PES ne viva cell 100 centrifugation device prior to ﬁnal analysis and storage. The puriﬁed material was qualiﬁed by mass spectrometry and analytical ultra-centrifugation.

° Cyno CD3 E+G His (E+G indicates S’Y subunits) A cell line to produce Cyno CD3 E+G Hch—6xHis was generated using HEK—293 cells (Thermo Fisher), the Lenti—X Lentiviral System (Clontech), and plasmid encoding Cyno CD3 E+G Hch—6xHis (Cynomolgus CD3E Accession No: Q95L15 <, cyno CD3 E+G hch—His: SEQ ID NO:109). For expression, cells were cultured and expanded in Freestyle 293 media, at 37C, humidified 8% C02 environment, and shaking at 135 rpm. The conditioned culture supernatant was harvested at Day 6 by centrifugation for 30 minutes at 9300xg. Expression was monitored by SDS-PAGE and Western Blotting. The conditioned culture supernatant was ed with 0.2M e, 5% glycerol, 0.01% CHAPS, and 10mM Imidazole. The pH was then ed to 7.2. Puriﬁcation was carried out in a two— step s: y puriﬁcation using Ni/NTA resin (overnight incubation at 4C, and elution with 250mM Imidazole); ed by size-exclusion chromatography on a Superdex 200 column in destination buffer PBS with 0.2M Sucrose, 5% ol, 0.01% CHAPS, lmM TCEP, pH7.2. The pooled material was concentrated using a 10K MWCO PES membrane viva cell 100 centrifugation device prior to ﬁnal analysis and storage. The puriﬁed material was qualiﬁed by mass spectrometry and analytical ultra-centrifugation.

Example 4A: SPR based determination of afﬁnities to recombinant DLL3 and CD3 S’Y subunits and interspecies cross-reactivity The g afﬁnity of puriﬁed DLL3/CD3 bispeciﬁc constructs for recombinant human and cyno DLL3—ECD and Fc—fusion proteins of human and cyno CD3 8y subunits was determined by Surface Plasmon nce (SPR), using a ProteOn XPR36 instrument (Bio-Rad).The running buffer and all dilutions (except where stated) were done in PBS—T— EDTA. The GLM sensor chip (Bio—Rad) was normalized and pre-conditioned as per the manufacturer’s recommendations. The sensor chip was activated with equal mixture of EDC/s-NHS in the horizontal direction for 300 sec at a ﬂow rate of 30 ul/min and —147— immobilized with protein A/G (20 ug/ml in 10 mM acetate pH 4.5) in the horizontal direction for 300 sec at a ﬂowrate of 30 ul/min resulting in ~ 5000 RU of protein A/G on the surface. The sensor chip was vated with 1M ethanolamine HCl in the horizontal direction for 300 sec at a flowrate of 30 . The sensor chip was stabilized with 18 sec of 0.85% phosphoric acid at a ﬂowrate of 100 til/min 3 times horizontally and 3 times vertically.

For binding kinetic determination to DLL3, bispeciﬁc molecules were captured individually on the protein A/G surface vertically for 250 sec at a ﬂowrate of 25 ul/min.

The baseline was stabilized by injecting EDTA for 60 sec at a ﬂowrate of 40 ul/min horizontally. The analyte (hu DLL3 or cy DLL3) was injected horizontally over the captured antibody for 600 sec at a ﬂowrate of 40 ul/min and a dissociation for 2700 sec.

The concentrations of the analyte were 20 nM, 10 nM, 5 nM, 2.5 nM, and 1.25 nM. The surface was regenerated by injecting 0.85% phosphoric acid solution for 18 sec at a ﬂowrate of 100 ul/min one time horizontally and one time vertically.

For binding kinetic determination to CD3, hu CD387 —Fc and cy CD387 —Fc (dissolved in pH 4.5 acetate) was immobilized directly onto the sensor surface h amine coupling for 140 sec at a e of 30 ul/min vertically. The baseline was stabilized by injecting PBS-T-EDTA for 60 sec at a ﬂowrate of 40 ul/min ntally. The analyte (the bispeciﬁc molecule) was injected horizontally over the captured CD3 87—Fc construct for 600 sec at a ﬂowrate of 40 pl/l’l’lll’l and a dissociation for 600 sec. The concentrations of the analyte were 250 nM, 125 nM, 62.5 nM, 31.3 nM, and 15.6 nM or 2 uM, 1 “M, 0.5 “M, 0.25 uM, and 0.125 nM. The surface was regenerated by injecting 0.85% phosphoric acid solution for 18 sec at a ﬂowrate of 100 ul/min one time horizontally and one time vertically.

The inter spot (interactions with sensor e) and blank (PBS-T—EDTA) were subtracted from the raw data. Sensorgrams were then ﬁt to 1:1 kinetic model to provide rate constants (ka, kd) as well as ty (KD) value for hu DLL3 and cy DLL3, and fit globally to equilibrium to provide affinity (KD) value for hu CD387 and cy CD387.

DLL3/CD3 binding proteins bed herein showed affinities to DLL3 in the pM range. ies to the CD3 87 subunit were in the low nM range. The interspecies gap between human and cyno is balanced. Affinities of three exemplary D3 g proteins (DLL3/CD3 binding proteins comprising a DLL3 chain of SEQ ID NO:75 and a CD3 —148— chain of SEQ ID NO:79, SEQ ID NO:80 or SEQ ID NO:105, respectively) are shown in Table 2A.

Table 2A: ies (KD) of DLL3/CD3 binding proteins to human and cynomolgus monkey DLL3 and CD3 8}! subunit as determined by SPR analysis, and interspecies gaps binding protein Example 48: SPR based determination of afﬁnities to recombinant DLL3 and CD3 87 The g afﬁnity of puriﬁed DLL3/CD3 bispeciﬁc constructs for recombinant human DLL3 —ECD and Fc—fusion n of human CD3 8y subunits was determined by Surface Plasmon Resonance (SPR), using a ProteOn XPR36 instrument (Bio—Rad). The method was similar as in Example 4A, with a few differences as outlined below.

For binding kinetic determination to DLL3, anti—DLL3/CD3 les were captured individually on the protein A/G e vertically for 200 sec at a ﬂowrate of 30 ill/min.

The baseline was stabilized by injecting PBS-T-EDTA for 60 sec at a ﬂowrate of 30 ill/min horizontally. The analyte (hu DLL3) was injected horizontally over the captured antibody for 300 sec at a ﬂowrate of 30 ill/min and a dissociation for 1800 sec. The concentrations of the analyte were 20 nM, 10 nM, 5 nM, 2.5 nM, and 1.25 nM. The surface was regenerated by injecting 0.85% phosphoric acid solution for 18 sec at a ﬂowrate of 100 ul/min one time horizontally and one time vertically.

For binding kinetic determination to CD3, HuCD3E_HuCD3 G—Fc (dissolved in pH 4.5 e) was immobilized directly onto the sensor surface through amine coupling for 360 sec at a ﬂowrate of 30 ul/min vertically. The baseline was stabilized by injecting PBS-T- EDTA for 60 sec at a e of 30 ill/min horizontally. The analytes (all anti-DLL3/CD3 molecules) were injected horizontally over the captured dy for 600 sec at a ﬂowrate of 30 ul/min and a dissociation for 1800 see. The concentrations of the e were 250 nM, 83.3 11M, 27.8 nM, 9.3 nM, and 3.1 nM. The surface was rated by injecting —149— 0.85% phosphoric acid solution for 18 sec at a ﬂowrate of 100 ul/min one time horizontally and one time vertically.

Afﬁnities of exemplary DLL3/CD3 binding proteins (DLL3/CD3 binding ns comprising a DLL3 chain of SEQ ID NO:73, SEQ ID NO:74, SEQ ID NO:75, SEQ ID NO:76, SEQ ID NO:77, SEQ ID NO:78, SEQ ID NO:241, SEQ ID NO:242, SEQ ID NO:243, SEQ ID , SEQ ID NO:245, SEQ ID NO:246, SEQ ID NO:247, SEQ ID NO:248, SEQ ID NO:249, SEQ ID NO:250, SEQ ID , or SEQ ID NO:252 and a CD3 chain of SEQ ID NO:79, a DLL3 binding protein comprising a DLL3 chain of SEQ ID NO:75 and a CD3 chain of SEQ ID NO:80, and a DLL3 binding protein comprising a DLL3 chain of SEQ ID NO:75 and a CD3 chain of SEQ ID NO: 105) are shown in 2B.

Table 2B: Afﬁnities (KD) of DLL3/CD3 binding proteins to human DLL3 and CD3 8y subunit as determined by SPR analysis DLL3/CD3 Recomb. binding protein DLL3#1/ 52 pM I CD3#1 DLL3#2/ 91 pM CD3#1 DLL3#3/ 75 pM CD3#1 DLL3#3/ 120 pM CD3#2 DLL3#3/ 116 pM CD3#3 / < 20 pM CD3#1 DLL3#5/ < 20 pM CD3#1 DLL3#6/ 152 pM CD3#1 DLL3#7/ 637 pM CD3#1 DLL3#8/ 40 pM CD3#1 DLL3#9/ 46 pM CD3#1 DLL3#10/ 141 pM I CD3#1 DLL3#1 1/ 37 pM CD3#1 DLL3#12/ < 20 pM CD3#1 DLL3#13/ 23 pM CD3#1 —150— DLL3/CD3 Recomb. binding protein CD3#l CD3#1 CD3#l CD3#l CD3#1 Example 5: Generation of HEK293 cells expressing the ellular domain and its subdomains on the cell surface For generation of stable HEK293 cells expressing full-length human DLL3 (Uniprot: Q9NYJ7), cynomolgus DLL3 (UPIOOO3AB95BD), human DLLl (Q00548) and human DLL4 (Q9NR61), respectively, the respective coding sequence was cloned into pcDNA3.l (Thermo Fisher iﬁc) and a FLAG-tag was inserted n the signal sequence and the extracellular . The expression on the cell surface was veriﬁed by using an anti- FLAG antibody (clone M2, Sigma) ed by a monoclonal anti mouse lgGl (Acris).

The extracellular domain of DLL3 consists of different subdomains: - Hu DLL3 Signal peptide: aa 1-26 ° Hu DLL3 N—terminal ECD domain: aa 27-175 - Hu DLL3 DSL: aa 176—215 0 Hu DLL3 EGFl: aa 216—249 0 Hu DLL3 EGF2: aa 274—3 10 - Hu DLL3 EGF3: aa 1 0 Hu DLL3 EGF4: aa 353—3 89 - Hu DLL3 EGFS: aa 391—427 ° Hu DLL3 EGF6: aa 429—465 - Membrane proximal peptide: aa 466—492 The following subdomains of DLL3 were expressed on the cell surface of HEK293 cells: - Hu DLL3 : Uniprot: Q9NYJ7 aa 216-310 ° Hu DLL3 EGF 2+3: Uniprot: Q9NYJ7 aa 274—351 - Hu DLL3 EGF3+4: Uniprot: Q9NYJ7 aa 312-389 ° Hu DLL3 EGF4+5: Uniprot: Q9NYJ7 aa 353—427 - Hu DLL3 EGF5+6: Uniprot: Q9NYJ7 aa 391-465 —151— - Hu DLL3 EGF6+membrane proximal peptide: Uniprot: Q9NYJ7 aa 429-492 For the generation of HEK293 cells expressing 6-His-myc tagged subdomains of human DLL3, the respective coding sequences were cloned into pcDNA 3.1 (Thermo Fisher Scientiﬁc). Constructs contain an N—terminal mouse IgG Vk g sequence, followed by a idin-myc tag and the respective human DLL3 . To ensure cell surface localization of the human DLL3 domains, all constructs were followed by a Ser/Gly— Linker, the transmembrane domain (aa266-288) and the intracellular domain (aa 289—3 14) ofhuman EpCAM (P16422). The expression of the subdomains was veriﬁed using mouse monoclonal IgG2a antibodies against s domains as described in WO 2013126746.

Bound monoclonal antibody was detected with an goat anti—mouse IgG (F(ab’)2)—RPE (Jackson Immuno Research). The samples were ed by flow cytometry. The sion of human DLLl and DLL4 was conﬁrmed by FACS analysis using anti-DLLl (R&D s, MAB1818) and anti—DLL4 R&D Systems, MAB1506) antibodies followed by PE-labeled anti-mouse or anti-rat secondary antibodies (see Figure 5, right panel).

Table 3: Sequences of various DLL3 ains expressed on HEK293 cells ID Construct Sequence SEQ ID FLAG- MVSPRMSGLLSQTVILALIFLPQTRPDYKDDDDKAGVFELQIHSF N021 10 tagged GPGPGPGAPRSPCSARLPCRLFFRVCLKPGLSEEAAESPCALGAAL human SARGPVYTEQPGAPAPDLPLPDGLLQVPFRDAWPGTFSFIIETWRE DLL3 full ELGDQIGGPAWSLLARVAGRRRLAAGGPWARDIQRAGAWELRF length SYRARCEPPAVGTACTRLCRPRSAPSRCGPGLRPCAPLEDECEAP LVCRAGCSPEHGFCEQPGECRCLEGWTGPLCTVPVSTSSCLSPRG PSSATTGCLVPGPGPCDGNPCANGGSCSETPRSFECTCPRGFYGL VTCADGPCFNGGLCVGGADPDSAYICHCPPGFQGSNCE KRVDRCSLQPCRNGGLCLDLGHALRCRCRAGFAGPRCEHDLDD CAGRACANGGTCVEGGGAHRCSCALGFGGRDCRERADPCAARP CAHGGRCYAHFSGLVCACAPGYMGARCEFPVHPDGASALPAAP PGLRPGDPQRYLLPPALGLLVAAGVAGAALLLVHVRRRGHSQD AGSRLLAGTPEPSV DVDPQGIYVISAPSIYAREVATPLFPPLHTGRAGQRQHLLFPYPSSI LSVK —152— WO 34220 2019/064942 ID uct Sequence SEQ ID FLAG- MVSPRMSRLLSQTVILALIFIPQARPDYKDDDDKAGVFELQIHSFG NO:111 tagged cyno PGPGPGAPRSPCSARGPCRLFFRVCLKPGLSEEAAESPCALGAALS DLL3 full ARGPVYTEQPEAPAPDLPLPNGLLQVPFRDAWPGTFSLIIETWREE length LGDQIGGPAWSLLARVTRRRRLAAGGPWARDIQRAGAWELRFS YRARCELPAVGTACTRLCRPRSAPSRCGPGLRPCAPLEDECEAPP VCRAGCSLEHGFCEQPGECRCLEGWTGPLCMVPVSTSSCLGLRG PSSTTTGCLVPGPGPCDGNPCANGGSCSETPGSFECTCPRGFYGLR CEVSGVTCADGPCFNGGLCVGGADPDSAYICHCPPGFQGSNCEK RVDRCSLQPCRNGGLCLDLGHALRCRCRAGFAGPRCEHDLDDC AGRACANGGTCVEGGGAHRCSCALGFGGRNCRERADPCAARPC AHGGRCYAHFSGLVCACAPGYMGARCEFPVHPDGVSALPAAPP GLRPGDPQRYLLPPALGLLVAAGVAGAALLLVHVRRRGHAQDA GSRLLAGTPEPSV VDSRGIYVISAPSIYAREVAMPLFPPLHTGRAGQRQNLLFPFPSSlL SEQ ID FLAG- MGSRCALALAVLSALLCDYKDDDDKQVWSSGVFELKLQEFVNK NO:112 tagged KGLLGNRNCCRGGAGPPPCACRTFFRVCLKHYQASVSPEPPCTY human GSAVTPVLGVDSFSLPDGGGADSAFSNPIRFPFGFTWPGTFSLIIEA DLLl full LHTDSPDDLATENPERLISRLATQRHLTVGEEWSQDLHSSGRTDL length KYSYRFVCDEHYYGEGCSVFCRPRDDAFGHFTCGERGEKVCNPG WKGPYCTEPICLPGCDEQHGFCDKPGECKCRVGWQGRYCDECIR YPGCLHGTCQQPWQCNCQEGWGGLFCNQDLNYCTHHKPCKNG ATCTNTGQGSYTCSCRPGYTGATCELGIDECDPSPCKNGGSCTDL ENSYSCTCPPGFYGKICELSAMTCADGPCFNGGRCSDSPDGGYSC RCPVGYSGFNCEKKIDYCSSSPCSNGAKCVDLGDAYLCRCQAGF SGRHCDDNVDDCASSPCANGGTCRDGVNDFSCTCPPGYTGRNCS EHAPCHNGATCHERGHRYVCECARGYGGPNCQFLLPEL PPGPAVVDLTEKLEGQGGPFPWVAVCAGVILVLMLLLGCAAVV VCVRLRLQKHRPPADPCRGETETMNNLANCQREKDISVSIIGATQ IKNTNKKADFHGDHSADKNGFKARYPAVDYNLVQDLKGDDTAV RDAHSKRDTKCQPQGSSGEEKGTPTTLRGGEASERKRPDSGCSTS KDTKYQSVYVISEEKDECVIATEV SEQ ID FLAG- MAAASRSASGWALLLLVALWQQRAAGDYKDDDDKSGVFQLQL NO:113 tagged QEFINERGVLASGRPCEPGCRTFFRVCLKHFQAVVSPGPCTFGTV human STPVLGTNSFAVRDDSSGGGRNPLQLPFNFTWPGTFSLIIEAWHAP DLL4 full GDDLRPEALPPDALISKIAIQGSLAVGQNWLLDEQTSTLTRLRYSY length RVICSDNYYGDNCSRLCKKRNDHFGHYVCQPDGNLSCLPGWTG EYCQQPICLSGCHEQNGYCSKPAECLCRPGWQGRLCNECIPHNG CRHGTCSTPWQCTCDEGWGGLFCDQDLNYCTHHSPCKNGATCS NSGQRSYTCTCRPGYTGVDCELELSECDSNPCRNGGSCKDQEDG PGYYGLHCEHSTLSCADSPCFNGGSCRERNQGANYACE CPPNFTGSNCEKKVDRCTSNPCANGGQCLNRGPSRMCRCRPGFT GTYCELHVSDCARNPCAHGGTCHDLENGLMCTCPAGFSGRRCE VRTSIDACASSPCFNRATCYTDLSTDTFVCNCPYGFVGSRCEFPV GLPPSFPWVAVSLGVGLAVLLVLLGMVAVAVRQLRLRRPDDGS REAMNNLSDFQKDNLIPAAQLKNTNQKKELEVDCGLDKSNCGK QQNHTLDYNLAPGPLGRGTMPGKFPHSDKSLGEKAPLRLHSEKP ECRISAICSPRDSMYQSVCLISEERNECVIATEV —153— ID Construct Sequence SEQ ID FLAG tag DYKDDDDK I\O:114 SEQ ID Vk leader METDTLLLWVLLLWVPGSTGD I\O:115 SEQ ID yc EQKLISEEDL 1\O:1 16 tag SEQ ID Ser/Gly SGGGGS I\O:1 17 linker SEQ ID EpCAM IVVVVIAVVAGIVVLVI I\O:1 18 transmemb rane domain SEQ ID EpCAM SRKKRMAKYEKAEIKEMGEMHRELNA N021 19 intracellula r domain SEQ ID EGFl APLVCRAGCSPEHGFCEQPGECRCLEGWTGPLCT NO: 120 SEQ ID Hu DLL3 APLVCRAGCSPEHGFCEQPGECRCLEGWTGPLCTVPVSTSSCLSP NO:121 EGF 1+2 RGPSSATTGCLVPGPGPCDGNPCANGGSCSETPRSFECTCPRGFY GLRCE SEQ ID EGF2 GPGPCDGNPCANGGSCSETPRSFECTCPRGFYGLRCE NO: 122 SEQ ID Hu DLL3 GPGPCDGNPCANGGSCSETPRSFECTCPRGFYGLRCEVSGVTCAD NO:123 EGF 2+3 GPCFNGGLCVGGADPDSAYICHCPPGFQGSNCE SEQ ID EGF3 SGVTCADGPCFNGGLCVGGADPDSAYICHCPPGFQGSNCE V0: 124 SEQ ID Hu DLL3 SGVTCADGPCFNGGLCVGGADPDSAYICHCPPGFQGSNCEKRVD V0: 125 EGF 3+4 RCSLQPCRNGGLCLDLGHALRCRCRAGFAGPRCE SEQ ID EGF4 RVDRCSLQPCRNGGLCLDLGHALRCRCRAGFAGPRCE V0: 126 SEQ ID Hu DLL3 RVDRCSLQPCRNGGLCLDLGHALRCRCRAGFAGPRCEHDLDDC V0: 127 EGF 4+5 AGRACANGGTCVEGGGAHRCSCALGFGGRDCR SEQ ID EGFS DLDDCAGRACANGGTCVEGGGAHRCSCALGFGGRDCR V0: 128 SEQ ID Hu DLL3 DDCAGRACANGGTCVEGGGAHRCSCALGFGGRDCRERADPCAA NO:129 EGF 5+6 RPCAHGGRCYAHFSGLVCACAPGYMGARCE —154— ID Construct Sequence SEQ ID EGF6 RADPCAARPCAHGGRCYAHFSGLVCACAPGYMGARCE NO: 1 30 SEQ ID Hu DLL3 RADPCAARPCAHGGRCYAHFSGLVCACAPGYMGARCEFPVHPD NO:131 EGF6- GASALPAAPPGLRPGDPQRYL membrane proximal peptide SEQ ID Membrane FPVHPDGASALPAAPPGLRPGDPQRYL NOzl32 proximal peptide Example 6: e domain mapping of DLL3/CD3 binding proteins Epitope domains were ined by binding of D3 binding ns to recombinant cell lines expressing subdomains of DLL3 (see Figure 2). Generation of these cell lines is described under Example 5.

DLL3/CD3 binding proteins were produced by transient transfection of CHO-E cells with the pTT5 vectors carrying the chain—encoding genes, as described in Example 2.

Cells expressing ains of DLL3 were stained with ep puriﬁed DLL3/CD3 binding proteins at 1.6 nM in FACS buffer (PBS/0.5%BSA/0.05% sodium azide). Bound molecules were detected with PE—conjugated anti—human secondary antibody (Sigma— Aldrich, #P8047). As negative control, cells were incubated with ary antibody only.

The samples were measured by ﬂow—cytometry. Figure 3A—C show exemplary binders (DLL3/CD3 binding ns comprising a DLL3 chain of SEQ ID NO:73, SEQ ID NO:74, SEQ ID NO:75, SEQ ID NO:76, SEQ ID NO:77, SEQ ID NO:78, SEQ ID NO:241, SEQ ID NO:242, SEQ ID NO:243, SEQ ID NO:244, SEQ ID , SEQ ID NO:246, SEQ ID NO:247, SEQ ID NO:248, SEQ ID NO:249, SEQ ID NO:250, SEQ ID NO:251 or SEQ ID NO:252 and a CD3 chain of SEQ ID NO:79, and a DLL3 binding protein comprising a DLL3 chain of SEQ ID NO:75 and a CD3 chain of SEQ ID NO:80) binding to different regions of DLL3. The sequences of the epitope domains are listed in Example 5, Table 3. —155— WO 34220 Example 7: Interspecies cross-reactivity Interspecies cross—reactivity was determined by binding of DLL3/CD3 binding proteins to NCI—H82 (HTB—l75TM) and SHP77 (ATCC®, CRL—2195TM) cells, two SCLC cell lines, as well as to a recombinant cell line expressing cynomolgus DLL3 (generation of cell line described in Example 5).

DLL3/CD3 g proteins were produced by transient ection of CHO—E cells with the pTTS vectors carrying the chain-encoding genes, as described in Example 2. Cells sing human or cynomolgus DLL3 were stained with 1.6 nM of two-step puriﬁed DLL3/CD3 binding proteins with increasing concentrations in FACS buffer (PBS/0.5%BSA/0.05% sodium azide). Bound molecules were detected with PE—conjugated anti-human secondary antibody (Sigma—Aldrich, #P8047). As negative control, cells were incubated with secondary antibody only. The samples were measured by ﬂow-cytometry.

Figure 4 shows g to human and cyno DLL3-expressing cells of seven exemplary DLL3/CD3 binding proteins (DLL3/CD3 binding proteins comprising a DLL3 chain of SEQ ID NO:73 and a CD3 chain of SEQ ID NO:79, a DLL3 chain of SEQ ID NO:74 and a CD3 chain of SEQ ID NO:79, a DLL3 chain of SEQ ID NO:75 and a CD3 chain of SEQ ID NO:79, a DLL3 chain of SEQ ID NO:75 and a CD3 chain of SEQ ID NO:80, a DLL3 chain of SEQ ID NO:76 and a CD3 chain of SEQ ID NO:79, a DLL3 chain of SEQ ID NO:77 and a CD3 chain of SEQ ID NO:79 or a DLL3 chain of SEQ ID NO:78 and a CD3 chain of SEQ ID NO:79).

Example 8: Conﬁrmation of the absence of binding to human DLLl and DLL4 Absence of cross-reactivity to human DLLl and DLL4 was determined by binding of DLL3/CD3 binding ns to recombinant cells expressing human DLLl or DLL4 (generation of cell lines described in Example 5).

DLL3/CD3 binding proteins were produced by transient ection of CHO-E cells with the pTTS vectors carrying the encoding genes, as described in Example 2.

Cells expressing DLLl and DLL4 were stained with 1.6 nM of two-step puriﬁed DLL3/CD3 binding ns with increasing concentrations in FACS buffer .5%BSA/0.05% sodium . Bound molecules were detected with PE—conjugated anti-human secondary antibody (Sigma—Aldrich, #P8047). As negative control, cells were —156— ted with secondary antibody only. The samples were measured by ﬂow—cytometry.

Expression of human DLLl and DLL4 was conﬁrmed by FACS analysis using anti-DLLl (R&D s, MAB1818) and LL4 R&D Systems, MAB1506) antibodies followed by PE-labeled anti—mouse or anti—rat secondary antibodies. Figure 5 shows binding of seven exemplary DLL3/CD3 binding proteins (DLL3/CD3 binding proteins comprising a DLL3 chain of SEQ ID NO:73 and a CD3 chain of SEQ ID NO:79, a DLL3 chain of SEQ ID NO:74 and a CD3 chain of SEQ ID NO:79, a DLL3 chain of SEQ ID NO:75 and a CD3 chain of SEQ ID NO:79, a DLL3 chain of SEQ ID NO:75 and a CD3 chain of SEQ ID NO:80, a DLL3 chain of SEQ ID NO:76 and a CD3 chain of SEQ ID NO:79, a DLL3 chain of SEQ ID NO:77 and a CD3 chain of SEQ ID NO:79 or a DLL3 chain of SEQ ID NO:78 and a CD3 chain of SEQ ID NO:79) to human DLLl and DLL4- expressing cells.

Example 9: Binding of DLL3/CD3 binding proteins to SCLC cell lines and T cells Binding of DLL3/CD3 binding proteins to human SCLC cell lines was tested by ﬂow cytometry ofNCI—H82 (HTB—175TM) and SHP77 (ATCC®, CRL—2195TM). DLL3/CD3 binding proteins were produced, as described in Example 2.

Human peripheral blood mononuclear cells (PBMCs) were prepared by Ficoll density nt centrifugation from enriched lymphocyte preparations (buffy coats), a side product ofblood banks collecting blood for transfusions. All buffy coats were obtained after informed t in accordance with the Declaration of Helsinki and with approval of the cantonal ethical tee in Austria and PBMCs were prepared the same day of tion. Therefore, mononuclear cells were isolated by Ficoll y gradient centrifugation (35 min without brake at 1400 rpm) and extensive washes with PBS.

Remaining erythrocytes were removed by incubating for 3 minutes in ACK lysis buffer (Thermo Fisher Scientific, A1049201), followed by washing in PBS, before suspension in assay medium ning RPMI 1640 GlutaMAX (Gibco #61870—010), 5% human AB serum AB (Gemini, GemCell cat # 100—5 12 LOT#H565001) + 1% MEM—NEAA (Gibco #11140-035), 10mM HEPES (Affymetrix #7365 9), 10 uM beta—Mercaptoethanol (Gibco #21985—023) and sodium pyruvat (Gibco #11360—039).

T—cells were ed by ve selection using the Pan T Cell Isolation Kit II (Miltenyi Biotec #130—091—156). In brief, cells were resuspend in 40 ul buffer PBS/0,5% BSA —157— (Gibco ref#04l-94553 M)/2mM EDTA (Invitrogen ref# 15575-03 8)per 10 Mio cells and incubated with 10 ul of -Antibody il per 10 Mio cells for 5 min at 4 °C.

Subsequently, 30 pl buffer and 20 pl anti-biotin MicroBeds/ 10 million cells were added and incubated for 10 min at 4°C. Subsequently the mixture was placed in a pre-rinsed 25LS column (Miltenyi Biotec #130401) in the ic ﬁeld of suitable MACS separator (Miltenyi Biotec). Flow—through was collected and washed in assay medium.

Cells (T cells or human SCLC cells) were stained with increasing concentrations of two- step purified DLL3/CD3 binding proteins with increasing concentrations in FACS buffer (PBS/0.5%BSA/0.05% sodium azide). Bound molecules were detected with PE—conjugated anti—human secondary antibody (Sigma-Aldrich, ). DLL3/CD3 binding proteins targeting DSL (DLL3#6), EGFl (DLL3#5) or EGF4 (DLL#4) domains show stronger binding to SCLC cell lines sing DLL3, compared to DLL3/CD3 binding ns which target the membrane proximal peptide (DLL3#1, DLL3#2, DLL3#3). Binding to SHP77, NCI-H82 cells and T cells of ary DLL3/CD3 binding proteins (DLL3/CD3 binding proteins comprising a DLL3 chain of SEQ ID NO:73, SEQ ID NO:74, SEQ ID NO:75, SEQ ID NO:76, SEQ ID NO:77, SEQ ID NO:78, SEQ ID NO:241, 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, SEQ ID NO:251 or SEQ ID NO:252 and a CD3 chain of SEQ ID NO:79, and a DLL3 g protein comprising a DLL3 chain of SEQ ID NO:75 and a CD3 chain of SEQ ID NO:80) against ent epitope domains is shown in Figure 6A—V.

Example 10A: Potency of redirecting imulated PBMCs against human SCLC cell lines Potency of non-stimulated PBMCs against SCLC cell lines was determined using lactate- dehydrogenase (LDH) release as readout for cell lysis. In this assay, DLL3 -positive SCLC cell lines, SHP77 and NCI—H82, were co—cultured with human PBMCs as or cells and increasing concentrations of DLL3/CD3 binding proteins for 72 hours at an effector to target cell ratio of 10:1. DLL3/CD3 binding proteins (DLL3/CD3 binding proteins comprising a DLL3 chain of SEQ ID NO:73, SEQ ID NO:74, SEQ ID NO:75, SEQ ID NO:76, SEQ ID NO:77, SEQ ID NO:78, and a CD3 chain of SEQ ID NO:79, and a DLL3 binding protein comprising a DLL3 chain of SEQ ID NO:75 and a CD3 chain of SEQ ID —158— NO:80) were produced by transient transfection of CHO—E cells with the pTT5 vectors carrying the encoding genes, as described in Example 2.

Human eral blood mononuclear cells (PBMCs) were prepared by Ficoll density gradient centrifugation from ed lymphocyte preparations (buffy coats), a side product of blood banks collecting blood for transfusions. All buffy coats were obtained after informed consent in accordance with the ation of Helsinki and with approval of the cantonal ethical committee in a and PBMCs were prepared the same day of collection. Therefore, mononuclear cells were isolated by Ficoll density gradient centrifugation (35 min without brake at 1400 rpm) and extensive washes with PBS.

Remaining erythrocytes were removed by incubating for 3 minutes in ACK lysis buffer (Thermo Fisher Scientiﬁc, A1049201), followed by washing in PBS, before sion in assay medium containing RPMI 1640 GlutaMAX (Gibco #61870-010), 5% human AB serum AB i, GemCell cat # 100-5 12 LOT#H565001) + 1% AA (Gibco #11140-035), lOmM HEPES (Affymetrix #7365 9), 10 uM beta-Mercaptoethanol (Gibco #21985—023) and sodium pyruvat (Gibco #11360—039).

Subsequently, target cells, SHP77 and NCI—H82, and PBMCs at a ratio of 1:10 were incubated with DLL3/CD3 binding proteins at concentrations from 0.0001 nM to 100 nM for 72 hours.

Cytotoxic activity was determined using the Cytotoxicity Detection KitPLUS (Roche), following the manufacturer’s instructions. In brief, this method is based on the usage of the e of LDH from dead or plasma —membrane damaged cells. Cell culture supernatant is incubated with the reaction mixture from the kit for 30 minutes and the formation of Formazan, as a result of LDH activity is measured in a spectrophotometer at 500nm as surrogate for cell lysis. Percentage of cytotoxicity relative to the maximal lysis control was calculated according to the ing formula: measured value — ound Cytotoxicity (relative to control) = max1mal lySIS — mlnlmal lySIS_ _ _ . .

Background: Target cells + Effector cells Maximal lysis: Target cells + 5% Triton X—100 Minimal lysis: Target cells —159— W0 2019l234220 Using GraphPad Prism 5 software, the percentage of cytotoxicity relative to the maximal lysis control was plotted against the ponding DLL3/CD3 binding protein concentrations. Dose response curves were analysed with the four-parameter logistic equation model for evaluation of sigmoidal dose—response curve and EC90 values were calculated (EC90 values are shown in Table 4A).

Figure 7A shows an example of potency in cell lysis of seven exemplary DLL3/CD3 binding proteins (DLL3/CD3 binding proteins comprising a DLL3 chain of SEQ ID NO:73, SEQ ID NO:74, SEQ ID NO:75, SEQ ID NO:76, SEQ ID NO:77, or SEQ ID NO:78, and a CD3 chain of SEQ ID NO:79, and a DLL3 binding protein comprising a DLL3 chain of SEQ ID NO:75 and a CD3 chain of SEQ ID NO:80) binding to different epitope domains.

Table 4A: EC90 values [nM] of DLL3/CD3 binding proteins as measured in a 72 hour cytotoxicity assay with non—stimulated PBMCs as effector cells and SCLC cell lines, NCI— H82 and SHP77 as target cells.

DLL3/CD3 binding . NCI H82 SHP77 proteins /CD3#1 1 .7 i—‘ J> DLL3#2/CD3#1 2 .9 waits/com -_ HUJ - U'I-J> /CD3#2 3 .9 16.5 no saturation at 100 DLL3#4/CD3#1 _ no saturation at 100 nM no saturation at 100 DLL3#5/CD3#1 _ no saturation at 100 nM no saturation at 100 DLL3#6/CD3#1 _ no saturation at 100 nM e 10B: Potency of redirecting non-stimulated T cells against human SCLC cell lines T cells were isolated as described in Example 9. Subsequently puriﬁed T cells were used as effector cells in a cytotoxicity assay with SHP77 and NCI-H82 cells as target cells as described in e 10A. Dose response curves are shown in Figures 7B—H (SHP77 cells) and Figure 71-0 (NCI-H82 cells). ated EC90 values are shown in Table 4B.

Figure 7B—O show examples of potency in cell lysis of 19 ary DLL3/CD3 binding proteins (DLL3/CD3 g proteins comprising a DLL3 chain of SEQ ID NO:73, SEQ —160— ID NO:74, SEQ ID NO:75, SEQ ID NO:76, SEQ ID NO:7?, SEQ ID No:78, SEQ ID NO:241, 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, SEQ ID NO:251, SEQ ID NO:252 Table 4B: ECgovalues [nM] of D3 binding proteins as measured in a 72 hour cytotoxicity assay with puriﬁed T cells as effector cells and SCLC and NCI-H82 cell lines as target cells.

D3 binding NCI-H82 SHP77 proteins /CD3#1 no activity no activity DLL3#6/CD3#1 no activity no activity DLL3#8/CD3#1 no activity no activity DLL3#9/CD3#1 no activity no activity DLL3#l6/CD3#1 no activity no activity DLL3#l7/CD3#1 no activity no activity Example 11: In vitro internalization assay Changes in the potency of the DLL3/CD3 g proteins after pre—incubated with DLL3- positive SHP77 cells as target cells were measured. If the D3 binding protein is —l6l— WO 34220 internalized, the effective concentration should decrease with this and thus the apparent potency should decrease.

D3 g proteins were produced as described in Example 2. Human eral blood mononuclear cells (PBMCs) were prepared as described in Example 10.

SHP77 target cells were plated and incubated with DLL3/CD3 binding proteins at concentrations from 0.0001 to 100 nM for 2 and 4 hours before adding non-stimulated PBMCs for 48 hours. The effector to target cell ratio was 10: l. Cytotoxic activity was determined using the Cytotoxicity Detection KitPLUS (Roche) as described in Example 10.

Figure 8 shows no ence in potency and efficacy of lysing cells of two exemplary DLL3/CD3 binding proteins CD3 binding protein comprising a DLL3 chain of SEQ ID NO:75 and a CD3 chain of SEQ ID NO:79, and DLL3/CD3 binding protein comprising a DLL3 chain of SEQ ID NO:75 and a CD3 chain of SEQ ID NO:80). Potency and efficacy have been observed ed to non—pre—incubated samples (0h). Results suggest that no signiﬁcant internalization occurs with DLL3/CD3 binding proteins.

Example 12: Mouse PK study The PK of DLL3/CD3 g proteins was evaluated in C57BL/6 mice ing a single 1 mg/kg iv. dose. Serum concentrations of DLL3/CD3 binding proteins were determined using a DLL3 capture/CD3 detection assay.

In brief, male C57BL/6 mice received a single 1 mg/kg intravenous (IV) dose (n = 3 per molecule). Blood samples were collected se and 0.15, 2, 8, 24, 72, 96, 168, 240 and 336 hours post—dose. Serum drug levels were measured with an MSD—based ligand binding assay, using biotinylated DLL3 as the capture reagent and sulfo-tagged CD3 as the detection reagent. Pharmacokinetic (PK) parameters were ated from serum concentration time-profiles using non-compartmental analysis. The following PK parameters were assessed: AUCtlast (area under the serum concentration-time curve from time zero to the last quantiﬁable time-point), AUCinf (area under the serum concentration- time curve extrapolated to y), CL (systemic clearance), Vss (steady—state volume of bution) and t1/2 (terminal half-life).

Mean (SD) serum concentration time-profiles for exemplary DLL3/CD3 binding proteins (DLL3/CD3 binding protein comprising a DLL3 chain of SEQ ID NO:75 and a CD3 chain of SEQ ID NO:79, and DLL3/CD3 binding protein comprising a DLL3 chain of SEQ ID —162— NO:75 and a CD3 chain of SEQ ID NO:80) are ized in Figure 9. Mean (SD) PK parameters for these exemplary D3 binding proteins are summarized in Table 5.

Table 5: Mean (SD) PK parameters of exemplary DLL3/CD3 g proteins in male C57BL/6 mice following a single 1 mg/kg enous dose DLL3 CD3 bmdm{wroteins' ' AUCO-Iast AUCu.inf CL Vss t 1/2 ("W/ml) (“W/ml) (mL/kg) (days) (mL/dav/kg) 178 10.6 DLL3#3/CD3#1 1180(127) 1760(179) 13.7(1.48) (57-7) (467) 109 4.97 DLL3#3/CD3#2 70) 1440(154) 16.8(1.74) (19.5) (0.78) Example 13: In vivo xenograft y study Efﬁcacy s were performed using a human xenograft mouse model reconstituted with human T cells. In detail, human SHP77 small cell lung cancer cells (2 x 107) were injected subcutaneously (S. c.) into the right dorsal ﬂank of sub-lethally irradiated (2 Gy, day -1) female NOD.Cg—Prkdcscjd IZng‘ij“g/IicTac mice (Day 1). In el, human CD3 positive T cells (isolated from healthy human blood donor) were expanded in vitro.

Human peripheral blood mononuclear cells (PBMCS) were prepared as described in Example 10. s were isolated by negative selection using the Pan T Cell Isolation Kit 11 (Miltenyi Biotec #130-091—156). In brief, cells were resuspend in 40 ul buffer 5% BSA (gibco ref#041—94553 M)/2mM EDTA (Invitrogen ref# 15575—03 8)per 10 Mio cells and incubated with 10 ul of Biotin-Antibody cocktail per 10 Mio cells for 5 min at 4 °C.

Subsequently, 30 ul buffer and 20 ul anti—biotin MicroBeds/10 million cells were added and incubated for 10 min at 4°C. Subsequently the mixture was placed in a pre-rinsed 25LS column (Miltenyi Biotec #130401) in the magnetic ﬁeld of suitable MACS separator (Miltenyi Biotec). Flow—through was collected and washed in assay medium.

Subsequently T cells were expanded using the T Cell Activation/Expansion Kit human (Miltenyi Biotec Cat#l30-09l-441, Lot#5170720843) for 20 days. In brief, anti-Biotin MACSiBeadTM Particles are loaded with CD2-, CD3-, CD28 Biotin and are transferred to —l63— the puriﬁed T cells in a ratio of 2 cells per particle and incubated in presence of 20 Units recombinant IL—2 (R&D#202—IL—050/CF) at a y of 0.5—1 106 ml for 20 days.

Cells were supplemented with 20 Units fresh IL-2 every three days. Three days before injection into the animals, T cells were were restimulated with anti-Biotin MACSiBeadTM Particles are loaded with CD2-, CD3-, CD28 Biotin at a ratio of 1 bead per 4 cells for additional three days. Finally, beads were removed with a MACSiMAG Separator (Miltenyi Biotec) and T cells were washed in PBMCs.

On day 14, animals were randomized into treatment groups based on tumor volume and 2 x 107 human T cells were injected intra—peritoneally (i.p.). Treatment was started on day 17 and DLL3/CD3 binding protein(DLL3/CD3 binding protein comprising a DLL3 chain of SEQ ID NO:75 and a CD3 chain of SEQ ID NO:79, and DLL3/CD3 binding protein comprising a DLL3 chain of SEQ ID NO:75 and a CD3 chain of SEQ ID NO:80) or Vehicle buffer (SOmM NaOAc, lOOmM NaCl, pH 5.0) was administered in a q7d dosing regimen by intravenous (2'. v.) bolus injections into the lateral tail vein. Tumor growth was monitored by external caliper measurements and tumor volumes were calculated using a standard hemi-ellipsoid formula. Human T cell engraftment was assessed in the spleen by immunohistochemistry (IHC) staining for human CD3 at the end of the study. Only those animals g human T—cell engraftment at the end of the study were included in the tical analysis. Animals reaching sacrifice criteria were euthanized early during the s for ethical s. ent of tumor-bearing mice with DLL3/CD3 binding proteins once weekly i.v. at 0.25 mg/kg induced icant tumor regression (Figure 10).

Example 14: Percent monomer content of DLL3/CD3 binding proteins Percent monomer was determined for exemplary DLL3/CD3 binding ns (DLL3/CD3 binding ns comprising a DLL3 chain of SEQ ID NO:73, SEQ ID NO:74, SEQ ID NO:75, SEQ ID NO:76, SEQ ID NO:77, SEQ ID NO:78, SEQ ID NO:241, SEQ ID NO:242, SEQ ID NO:243, SEQ ID , SEQ ID NO:245, SEQ ID , SEQ ID NO:247, SEQ ID NO:248, SEQ ID NO:249, SEQ ID NO:250, SEQ ID NO:251, or SEQ ID NO:252 and a CD3 chain of SEQ ID NO:79, a DLL3 binding protein comprising a DLL3 chain of SEQ ID NO:75 and a CD3 chain of SEQ ID NO:80, and a DLL3 binding protein comprising a DLL3 chain of SEQ ID NO:75 and a CD3 chain of SEQ ID NO: 105) by Analytical Size Exclusion Chromatography (aSEC) (shown in Table 6). aSEC was run on a Waters (Milfrod, MA, USA) Acquity UPLC system using a Protein BEH SEC column —164— 2019/064942 200A, 1.7pm, 4.6x150mm (Cat# 186005225). Running conditions were as follows: Mobile phase: 50mM Sodium Phosphate, 200mM Arginine and 0.05% Sodium Azide; Flow rate: 0.5ml/min; Runtime: 5 minutes; Sample loading amount: 100g; Peak detection: ; Automated processing method of chromatograms.

Table 6: Percent monomer after ﬁrst and second puriﬁcation step DLL3/CD3 binding Percent monomer after Percent monomer after proteins 1St step of puriﬁcation 2nd step of ation DLL3#1/CD3#1 76.4 99.4 DLL3#2/CD3#1 54.4 98.0 /CD3#1 69.5 99.8 DLL3#3/CD3#2 77.6 99.5 DLL3#3/CD3#3 70.8 98.3 DLL3#4/CD3#1 72.2 99.1 DLL3#5/CD3#1 70.5 99.1 DLL3#6/CD3#1 55.2 99.9 DLL3#7/CD3#1 78.1 99.0 DLL3#8/CD3#1 54.8 99.5 DLL3#9/CD3#1 72.0 99.0 DLL3#10/CD3#1 70.2 98.3 1/CD3#1 68.2 94.7 DLL3#12/CD3#1 66.2 98.7 DLL3#13/CD3#1 52.8 99.9 DLL3#14/CD3#1 52.6 87.0 DLL3#15/CD3#1 61.5 na DLL3#16/CD3#1 71.8 99.5 DLL3#17/CD3#1 60.2 94.7 DLL3#18/CD3#1 61.5 98.2 Example 15A: Thermostability Thermostability was determined by Differential Scanning Calorimetry (DSC) and results of the first melting transitions (Tml) of D3 binding proteins are shown in Table 7.

DSC l melts e information regarding the thermal stability of a protein in solution relative to a buffer control. Thermal unfolding and aggregation of a 1 mg/ml solution of proteins in 20 mM Citrate, 115 mM NaCl, pH 6 was monitored from 20 °C to 110 °C at a scan rate of 60 °C/hr Via an automated capillary DSC. —165— Table 7A DLL3/CD3 binding proteins Tml DLL3#3/CD3#1 64 DLL3#3/CD3#2 60 DLL3#3/CD3#3 63 DLL3#4/CD3#1 63 Example 15B: Thermostability Thermostability was determined by Thermal Shift Analysis (TSA) and results of the ﬁrst melting transitions (Tml) of D3 binding proteins (DLL3/CD3 binding proteins comprising a DLL3 chain of SEQ ID NO:73, SEQ ID NO:74, SEQ ID NO:75, SEQ ID NO:76, SEQ ID NO:77, SEQ ID NO:78, SEQ ID NO:241, SEQ ID , 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, SEQ ID NO:251 or SEQ ID NO:252 and a CD3 chain of SEQ ID NO:79, a DLL3 binding n sing a DLL3 chain of SEQ ID NO:75 and a CD3 chain of SEQ ID NO:80, and a DLL3 binding protein comprising a DLL3 chain of SEQ ID NO:75 and a CD3 chain of SEQ ID NO: 105) are shown in Table 7B. The ﬂuorescence intensity proﬁle as a on of temperature was acquired using a QuantStudio 6 Flex real-time PCR system (Applied Biosystems, Waltham, MA) with SYPRO Orange (Invitrogen, Carlsbad, CA) as the extrinsic ﬂuorophore. Sample was diluted to 0.4 mg/ml in 10 mM histidine, pH 6.0 with 40 mM sodium chloride and 0.02% sodium azide. The melt curve was generated with a thermal ramp from 25°C to 95°C at a rate of 2°C/min, with data collected imately every 0.4°C h the ‘ROX’ ﬁlter set (Ex: 580 110 nm, Em: 623 il4 nm). Data were analyzed using Protein Thermal Shift Software n 1.3 (ThermoFisher Scientiﬁc, Waltham, MA).

Table 7B DLL3/CD3 binding proteins Tml (°C) DLL3#1/CD3#1 65.6 DLL3#2/CD3#1 65.5 DLL3#3/CD3#1 65.6 DLL3#3/CD3#2 63.0 DLL3#3/CD3#3 65.5 —l66— DLL3/CD3 binding proteins Tml (°C) DLL3#4/CD3#1 65.8 /CD3#1 65.6 DLL3#6/CD3#1 65.8 DLL3#7/CD3#1 65.6 DLL3#8/CD3#1 65.8 DLL3#9/CD3#1 65.9 DLL3#10/CD3#1 64.6 DLL3#1 l/CD3#1 66.0 DLL3#12/CD3#1 65.9 DLL3#l3/CD3#1 65.8 DLL3#14/CD3#1 65.9 DLL3#l5/CD3#1 na DLL3#l6/CD3#1 65.8 DLL3#l7/CD3#1 65.8 DLL3#l8/CD3#1 65.8 Example 16: Predicted Immunogenicity scores in silico by Epivax: genicity of sequences was evaluated in silico with a mathematical algorithm.

Speciﬁcally, EpiMatrix Treg—adjusted Scores (Einax Inc., Providence RI)) as a measure of immunogenicity scores, were ined for DLL3/CD3 binding proteins (DLL3/CD3 binding proteins comprising a DLL3 chain of SEQ ID NO:?3, SEQ ID NO:74, SEQ ID NO:75, SEQ ID NO:76, SEQ ID NO:77, SEQ ID NO:78, SEQ ID , SEQ ID N02242, SEQ ID N02243, SEQ ID NO:244, SEQ ID , SEQ ID NO:246, SEQ ID NO:247, SEQ ID NO:248, SEQ ID NO:249, SEQ ID NO:250, SEQ ID NO:251 or SEQ ID N02252 and a CD3 chain of SEQ ID NO:79, a DLL3 binding protein comprising a DLL3 chain of SEQ ID NO:75 and a CD3 chain of SEQ ID NO:80, and a DLL3 binding protein comprising a DLL3 chain of SEQ ID NO:75 and a CD3 chain of SEQ ID NO: 105) and compared to the scores of various Fc sequences. These scores are taking T—cell epitopes and Treg es into consideration. The lower the immunogenicity score, the less likely a sequence to be immunogenic. In general, a negative score is considered low risk of immunogenicity, while a highly positive score is viewed as indication for potential immunogenicity. As shown in the tables below, D3 binding proteins described herein have very low immunogenicity scores, indicating that the risk of being genic is low for these binding proteins. —167— Table 8: Adjusted Epivax scores of DLL3/CD3 binding proteins DLL3/CD3 binding VH VL Full polypeptide ns chain (VL-CL-linker- VH-CH1-hinge-CH2- CH3) DLL3#1 DLL3 chain -36.88 -59.85 -43.31 DLL3#2 DLL3 chain -46.99 31.40 -33.53 DLL3#3 DLL3 chain -34.44 2.84 -35.50 DLL3#4 DLL3 chain 25.97 28.92 -21.93 DLL3#5 DLL3 chain -59.87 -46.45 -44.95 DLL3#6 DLL3 chain 29.70 -l4.55 -27.26 CD3#1 CD3 chain -9.68 -50.52 -35.25 CD3#2 CD3 chain -15,54 -50.52 -36.23 CD3#3 CD3 chain -13.49 -50.52 -35.88 DLL3#7 DLL3 chain -48.55 -34.92 -41.60 DLL3#8 DLL3 chain -54.39 -37.79 -42.89 DLL3#9 DLL3 chain 35.95 -60.37 -31.83 DLL3#10 DLL3 chain -33.82 -34.60 -40.73 1 DLL3 chain -24.00 26.66 -30.50 DLL3#12 DLL3 chain -30.43 25.68 -31.67 DLL3#13 DLL3 chain 8.64 18.63 -25.96 DLL3#14 DLL3 chain 1.58 -10.63 -31.72 DLL3#15 DLL3 chain 8.93 -8.31 -29.91 DLL3#16 DLL3 chain -55.14 13.98 -35.59 DLL3#17 DLL3 chain 11.89 12.55 -26.52 DLL3#18 DLL3 chain 5.82 15.85 -27.01 Table 9: Adjusted Epivax scores of Fc domains Fc Protein Chain Adjusted Epivax score Fc-IgGl -WT -25.64 Fc-IgGl -LALA -29.83 Fc-IgGl -LALA-KNOB -31.76 Fc-IgGl-LALA-HOLE -18.01 Example 17A: eciﬁc binding to surfaces The speciﬁcity of the DLL3/CD3 binding proteins of the invention was r tested in a SPR—based assay using highly charged proteins. A non-speciﬁc binding assay was —168— 2019/064942 developed using biosensor technology to determine if binding proteins have signiﬁcant binding to unrelated d ns. In this assay, DLL3/CD3 binding proteins were passed over two SPR surfaces, one coated with a negatively d protein (Trypsin Inhibitor) and one coated with a positively charged protein (Lysozyme). When a protein displays signiﬁcant non-speciﬁc binding to these surfaces, it is likely that the cause of binding is the presence of positive or negative charged surface patches on the candidate.

Non-speciﬁc binding of proteins may translate to poor pharmacokinetics (PK) and biodistribution and may also have downstream manufacturability s.

The experiment was performed on a Biacore T200. The dilution, surface preparation, and binding experiments were performed at 25 0C in 1X HBS-EP buffer prepared from 10X HBS-EP. The ﬂow rate for both the immobilization protocol and binding ment was at 5 uL/min.

To prepare the surface for the non-specific binding experiment, chicken egg white lysozyme and trypsin inhibitor from glycine max n were coupled ly to a series S CM5 chip with the surface density of 000 RU using the amine coupling kit according to the cture instructions.

Samples were prepared at 1 uM in 1X HBS—EP buffer. The samples were injected over activated surfaces with a 10 min association and 10 min dissociation. The data was collected using Biacore T200 Control Software version 2.0.1 and analyzed using Biacore T200 Evaluation Software version 3.0.

The DLL3/CD3 binding proteins did not bind to these highly charged surfaces. Table 10A shows absence of binding to the two highly charged ns, Trypsin tor and Lysozyme, with two exemplary DLL3/CD3 binding proteins.

Table 10A: Non-specific g DLL3/CD3 Binding Protein. _ _ Lysozyme (posntlve)_ _ Tryp. Inhlbltor (negative). _ _ DLL3#3/CD3#1 DLL3#3/CD3#2 Example 17B: Non-speciﬁc binding to surfaces The specificity of the DLL3/CD3 binding ns (DLL3/CD3 binding proteins comprising a DLL3 chain of SEQ ID NO:73, SEQ ID NO:74, SEQ ID NO:75, SEQ ID —l69— NO:76, SEQ ID NO:77, SEQ ID N0278, SEQ ID , SEQ ID NO:242, SEQ ID , SEQ ID NO:244, SEQ ID NO:245, SEQ ID , SEQ ID NO:247, SEQ ID N02248, SEQ ID N02249, SEQ ID NO:250, SEQ ID NO:251 or SEQ ID NO:252, and a CD3 chain of SEQ ID NO:79, a DLL3 binding n comprising a DLL3 chain of SEQ ID NO:75 and a CD3 chain of SEQ ID NO:80, and a DLL3 binding protein comprising a DLL3 chain of SEQ ID NO:75 and a CD3 chain of SEQ ID NO: 105) was further tested in an SPR—based assay using highly charged proteins, as described in Examplel7a, with a 10 min association and 15 min dissociation. The results are shown in Table 10B.

Table 10B: Low RU numbers indicate no signiﬁcant binding to unrelated charged proteins Non-specific binding (RU) DLL3/CD3 Binding Tryp. Inhibitor Protein Lysozyme (positive) (negative) DLL3#1/CD3#1 4 1 DLL3#2/CD3#1 7 2 DLL3#3/CD3#1 54 85 DLL3#3/CD3#2 44 36 DLL3#3/CD3#3 17 28 /CD3#1 16 37 DLL3#5/CD3#1 4 0 DLL3#6/CD3#1 6 0 DLL3#7/CD3#1 9 3 DLL3#8/CD3#1 39 18 DLL3#9/CD3#1 15 27 DLL3#10/CD3#1 25 12 DLL3#ll/CD3#1 ll 13 DLL3#l2/CD3#1 8 9 DLL3#l3/CD3#1 ll 3 4/CD3#1 l8 6 DLL3#l5/CD3#1 na na DLL3#16/CD3#1 6 0 DLL3#l7/CD3#1 l5 8 DLL3#l8/CD3#1 23 81 —l70— e 18: Induction of lysis, T cell activation, T cell degranulation, T cell proliferation and cytokine secretion in presence of DLL3-positive and DLL3-negative tumor cells PBMCs were puriﬁed as described in Example 9. To determine T cell tion, T cell degranulation and cytokine secretion, a cytotoxicity assay with PBMCs and DLL3 -positive SHP77 or egative RKO—E6 cells as target cells was setup as described in Example 10B. Potency of cell lysis by redirecting T cells towards human SHP77 or RKO—E6 cells was determined as described in Example 10A. To determine T cell activation, and T cell ulation cells were centrifuged and stained with antibodies against CD4 (BD #550630), CD8 (BD #557834), CD25 (BD#340907), subsequently the cells were permabilized using the Fixation/Permeabilization Solution (BD 4) and stained with antibodies against Perforin (BioLegend #308120) and Granzyme B (BD #560221) and measured by ﬂow—cytometry. Cytokine levels in supematants were determined by U- PLEX Biomarker Group 1 (hu) Assays (MSD, #K15067L-2, Kit).

To determine the proliferation of T cells, PBMCs were labeled with 5 uM Cell TraceTM CFSE (Invitrogen, ) and T cell stained with an anti-CD3 antibody (BioLegend cat#: 317336) . Subsequently the labeled PBMCs were incubated with SHP77 or RKO-E6 cells at a ratio of 10:1 and increasing concentrations of an D3 binding protein for 6 days.

Figures 11-15 show induction of T cell redirected lysis of SHP77 or RKO-E6 cells (Figure 11) and induction of dose-dependent T cell activation (Figure 12), T cell degranulation (Figurel3), secretion of cytokines (Figure 14A: Interferon gamma; 14B: MCP-l), and proliferation of T cells (Figure 15) in presence of SHP77 or RO-E6 cells and an DLL3/CD3 binding protein.

Example 19: Redirection of T cell subsets to SHP77 cells PBMCs were ed as bed in Example 10a. Subsequently T cell subsets were isolated using the ing reagents and protocols: Table 11: ts and protocols for ion of T cell subsets T cell subset Used reagents and protocols Native T cells Miltenyi Biotech; #130—097—095 CD4+ T cells Miltenyi Biotech; #130533 —171— T cell subset Used reagents and protocols CD4+ or Memory T cells Miltenyi Biotech; #130—094-125 CD4+ Central Memory T cells Miltenyi Biotech; #130302 CD8+ T cells Miltenyi Biotech; #130—096—495 CD8+CD45RA+ Effector T cells Miltenyi Biotech; #130-094—485 CD8+ Memory T cells Miltenyi Biotech; #130412 Figures 16—20 shows potency of redirecting of Pan T cells (Figure 16), naive T cells (Figure 16), CD4+T cells (Figures 17, 18), CD4+ Effector Memory T cells (Figure 18), CD4+ Central Memory T cells (Figure 17), CD8+T cells es 19, 20), CD8+CD45RA+ Effector T cells (Figure 19), CD8+ Memory T cells e 20) against human SHP77 cells of an DLL3/CD3 g protein.

Example 20: T cell inﬁltration in SHP77 xenograft tumor tissue with an exemplary DLL3/CD3 binding protein.

Remaining tumor s from mice in the study described in Example 13 were prepared, fixed in formalin and embedded in n. Subsequently tissue sections were prepared and stained for CD3 sion on T cells (2GV6 Ventana). T cell inﬁltration in SHP77 xenograft tumor tissue with an ary DLL3/CD3 binding protein is shown in Figure 21. The scoring in Table 12 was used to quantify CD3 expression in xenograft tumor tissues.

Table 12: Scoring for quantification of infiltrating CD3 -positive T cells moderate cell numbers or clusters in stroma and margin; low numbers of intra- epithelial cells igh cell numbers or clusters in stroma and margin; low numbers of intraepithelial cells —l72— Example 21: Production of LL3 antibodies To obtain anti-DLL3 binders, hybridomas or single B cells derived from DLL3 immunized wild-type and AlivaMab humanized mice (Ablexis, San Francisco, CA, USA: AlivaMab transgenic mouse platform with human immunoglobulin loci) were ed in vitro.

Supernatants were screened for reactivity against recombinant human DLL3, by AlphaLISA (PerkinElmer, Waltham, MA, USA), and against SHP-77 cells (ATCC®, CRL-2195TM) expressing human DLL3, by Flow try.

Immunoglobulin (Ig) VH and VL genes were then ampliﬁed from identified positive clones. To isolate RNA from hybridomas, about 2x106 cells from single clones were pelleted and used as source material. For single B cells, 100 to 500 cells expanded from singularly ed B cells were used as source material. RNA was isolated using RNeasy Plus (Qiagen, Hilden, Germany). cDNA was then synthesized using Smarter cDNA synthesis kit (Clontech, Mountain View, CA) according to manufacturer’s instructions. To facilitate cDNA synthesis, oligodT was used to prime e transcription of all messenger RNAs ed by “5’ capping” with a Smarter IIA oligonucleotide.

Subsequent ampliﬁcation of the VH and VL fragments was performed using a 2-step PCR amplification using 5’ primers ing the Smarter IIA cap and 3 ’ primers targeting consensus s in CH1. Brieﬂy, each 50ul PCR reaction consists of 20uM of forward and reverse primer mixes, 25 ul of PrimeStar Max DNA polymerase premix (Clontech), 2 ul of unpuriﬁed cDNA, and 21 ul of double-distilled H20. The cycling program starts at 94 °C for 3 min, followed by 35 cycles (94°C for 30 Sec, 50°C for l min, 68 °C for l min), and ends at 72 °C for 7 min. The second round PCR was performed with VL and VH 2nd round primers containing l5bp complementary extensions that “overlap” respective regions in their tive pTT5 mother vector (VH and VL). Second round PCR was performed with the following program: 94 °C for 3 min; 35 cycles (94 °C for 30 Sec, 50°C for l min, 68 °C for l min), and ends at 72 °C for 7 min.

In—Fusion® HD Cloning Kit (Clontech, USA.) was used for directional cloning ofVL gene into pTT5 hngK vector and VH gene into pTT5 hngGlKO vector. To tate In- Fusion® HD Cloning, PCR ts were purified and treated with Cloning Enhancer before In—Fusion HD g. Cloning and transformation were performed according to manufacturer’s protocol (Clontech, U.S.A.). Mini—prep DNAs were subjected to Sanger sequencing to conﬁrm that complete V-gene fragments were obtained. —l73— Using this methodology, pairs of Ig VH and VL genes ng binding domains with speciﬁcity for DLL3 were prepared. inant antibodies were produced by transient transfection of CHO-E37 cells with the corresponding heavy and light chain-encoding plasmids.

Conﬁrmatory screening of recombinant antibodies Supernatants containing expressed recombinant antibodies were assayed by ﬂow cytometry for g to cell lines expressing human DLL3. Brieﬂy, cells were incubated with recombinant atants, washed, and bound mAbs from the supernatants were detected with anti-human-IgG-APC (Jackson Research 109—136—098). Signal—to— background ratios (S/B) were calculated by dividing the median ﬂuorescence intensity (MFI) of the sample by that of isotype control.

Surface Plasmon nce (SPR) on Biacore 400 was performed on recombinant supernatants. Brieﬂy, the non-optimized IgGs in the HTP supernatants were captured Via Protein A/G onto the sensor e for 60 sec at 10 . Binding of 100 11M human DLL3 to the captured IgGs was monitored for 180 sec of association at 30 ul/min, followed by 120 sec of dissociation in the HBS-EP buffer. Regeneration of the n A/G e was performed with Glycine pH 2.1 in between each binding cycle. The following materials were used in this assay: n reagent: recombinantly expressed human DLL3.

System running buffer: HBS-EP (10 mM HEPES pH 7.4, 150 mM NaCl, 3 mM EDTA, and 0.005% V/V polysorbate P20). Capturing reagent: Protein A/G, with specificity towards all human IgG isotypes.

Clones of interest (with Kd<3 OOpM) were selected and further evaluated in various detection assays as described below.

Example 22: Binding of anti-DLL3 antibodies to recombinant human DLL3 protein g of antibodies was performed in an ELISA. In brief, anti-DLL3 antibodies were coated at a concentration of 2 itle at 4°C over night. After washing the plates were blocked with ng buffer (Gibco, Gibco#043 —90309A) for one hour at room temperature, followed by washing and incubation with recombinant DLL3 protein at increasing concentrations. For detection the bound DLL3 protein was incubated with a polyclonal anti-DLL3 antibody (R&D Systems, AB4315) followed by a SULFO—TAG labelled anti-goat antibody (R32AG-1) for one hour. The bound antibody was quantiﬁed —174— by using a Read Buffer (MSD) in a Sector imager 6000 (MSD). Figure 22 shows binding to three exemplary anti-DLL3 antibodies to recombinant protein. LL3 antibody targeting the membrane proximal peptide domain (DLL3#3) shows stronger binding to recombinant DLL3 protein, ed to anti-DLL3 antibodies which target the EGF4(DLL3#4) or EGFl domain (DLL3#S).

Example 23: Binding of anti-DLL3 antibodies to SCLC cell lines Cells (T cells or human SCLC cells) were stained with increasing concentrations of anti- DLL3 antibody with increasing concentrations in FACS buffer (PBS/0.5%BSA/0.05% sodium . Bound molecules were detected with PE—conjugated ouse secondary antibody (Jackson Immuno Research, 115—116-072) by FACS analysis. Anti-DLL3 dies targeting DSL (DLL3#6), EGFl (DLL3#5) or EGF4 4) domains show stronger binding to SCLC cell lines expressing DLL3, compared to anti-DLL3 antibodies which target the C—terminal peptide (DLL3#1, , ) as shown in Figure 23.

Example 24: (IHC) Anti-DLL3 antibodies were tested on a DLL3 mRNA positive SCLC sample at 20ug/ml concentration using Ventana Discovery ultra (Ventana Medical Systems, Inc, Arizona) following manufacturer’s instructions. In brief, IHC was performed on formalin fixed parafﬁn embedded tissue and sections (4 um) where stained with anti-DLL3 dies.

IHC assay was performed on Ventana Discovery Ultra rm using RUO Discovery Universal protocol. Anti—DLL3 dy staining was optimised using Proteinase K as Epitope Retrieval for 12 minutes. Anti-DLL3 antibody was incubated for 60 minutes, followed by anti—mouse HQ detection system and anti HQ HRP both for 12 minutes.

Negative controls (IgG antibody) were performed for each tissue section, treated identically to the test slides.

The human SCLC cell line SHP77 were used as positive control. Digital images of whole- tissue sections were ed using a Leica SCN400 histology r (Leica Microsystems, Milton Keynes, UK).

Immunohistochemistry analysis of a human SCLC sample with exemplary anti-DLL3 antibodies is shown in Figure 24. —l75— Example 25: Determination of sion levels on SCLC cell lines Cell surface expression of anti-DLL3 antibodies was quantiﬁed using the QIFIKIT (K0078; Dako). In brief, SCLC cell lines , NCI-H82 and NCI-H2286) were labeled with the anti-DLL3 antibody (DL309, ) or irrelevant antibody (isotype control) at 5 ug/ml. In a te vial. Subsequently cells and beads provided in the kit were labeled in el with ﬂuorescein-conjugated anti—mouse ary antibodies.

Samples were measured in a BD CantoII Fluorocytometer and consequently, the ﬂuorescence is correlated with the number of bound anti—DLL3 antibodies on the cells and on the beads. Table 12 shows DLL3 molecules expressed on the cell surface of three SCLC cell lines.

Table 13: DLL3 molecules expressed on the cell surface of SCLC cell lines Cell surface DLL3 SCLC CELL LINE molecules/cell SHP77 (ATCC®, CRL- CI—H2286 129 Example 26: Determination of sensitivity of IHC protocol on SCLC cell line blocks To evaluate the sensitivity of the IHC protocol, SCLC cell lines (SI-11377, NCI—H82 and NCI—H2286) were processed like tissues in hospitals and ﬁxed in Formalin and uently embedded in parafﬁn.

SCLC cell lines were cultured according to ATCC instructions. Cells were scraped from the plate and ﬁxed in Formalin and subsequently added to solubilized Histogel (Thermo Fisher scientiﬁc, HG-4000—012) and incubated over night at 4°C, followed by a standard parafﬁn ing ure performed in routine pathology laboratory. In brief, cell block is incubated in l and Isopropanol before ing in Parafﬁn (roti-Plast, #66425). Sections of the cell pellet blocks were performed and sections were stained with the protocol as described in Example 24. The results shown in Figure 25 show that an IHC protocol with anti-DLL3 antibody DLL3#5 is able to detect cells with very low DLL3 expression. —l76— WO 20191234220

Claims

1. l. A protein comprising a ﬁrst antigen binding unit Speciﬁcally binding to DLL3 and a second antigen binding unit speciﬁcally binding to CD3, wherein said ﬁrst antigen binding unit speciﬁcally binding to DLL3 is selected from the group consisting of i) to xviii): i) an antigen binding unit comprising light chain CDRs comprising the amino acid ces of SEQ ID N0:l (CDRl), SEQ ID N0:2 (CDR2) and SEQ ID NO:3 (CDR3) and heavy chain CDRs comprising the amino acid sequences 10 of SEQ ID NO:4 (CDRl), SEQ ID N025 (CDR2) and SEQ ID N0:6 (CDR3); ii) an antigen binding unit comprising light chain CDRs sing the amino acid sequences of SEQ ID N017 (CDRl), SEQ ID N028 (CDR2) and SEQ ID N0:9 (CDR3) and heavy chain CDRs comprising the amino acid sequences of SEQ ID N0210 (CDRI), SEQ ID N0:ll (CDR2) and SEQ ID N0212 15 ; iii) an antigen binding unit comprising light chain CDRs comprising the amino acid ces of SEQ ID N0:l3 (CDRl), SEQ ID N0:14 (CDR2) and SEQ ID N0:15 (CDR3) and heavy chain CDRs comprising the amino acid sequences of SEQ ID N0:l6 (CDRl), SEQ ID N0:l7I (CDR2) and SEQ ID 20 N0:18 (CDR3); iv) an antigen binding unit comprising light chain CDRs comprising the amino acid sequences of SEQ ID N0:l9 , SEQ ID N0:20 (CDR2) and SEQ ID N0:21 (CDR3) and heavy chain CDRs comprising the amino acid sequences of SEQ ID N0:22 (CDRl), SEQ ID N0:23 (CDR2) and SEQ ID 25 N0:24 (CDR3); an antigen binding unit comprising light chain CDRs comprising the amino acid sequences of SEQ ID N0:25 (CDRl), SEQ ID N0:26 (CDR2) and SEQ ID N0:27 (CDR3) and heavy chain CDRs comprising the amino acid sequences of SEQ ID N0:28 , SEQ ID N0:29 (CDR2) and SEQ ID 30 NO:30 (CDR3); vi) an antigen binding unit comprising light chain CDRs comprising the amino acid ces of SEQ ID N0231 (CDRI), SEQ ID N0232 (CDR2) and SEQ —l77— WO 20191234220 ID NO:33 (CDR3) and heavy chain CDRs comprising the amino acid sequences of SEQ ID NO:34 , SEQ ID NO:35 (CDR2) and SEQ ID NO:36 (CDR3); vii) an antigen binding unit comprising light chain CDRs comprising the amino acid sequences of SEQ ID NO:133 (CDRl), SEQ ID NO:134 (CDR2) and SEQ ID NO:135 (CDR3) and heavy chain CDRs comprising the amino acid sequences of SEQ ID NO:136 (CDRl), SEQ ID NO:137 (CDR2) and SEQ ID NO:138 (CDR3); viii) an antigen binding unit comprising light chain CDRs comprising the amino 10 acid sequences of SEQ ID NO:139 (CDRl), SEQ ID NO:140 (CDR2) and SEQ ID NO:14l (CDR3) and heavy chain CDRs comprising the amino acid sequences of SEQ ID NO:142 (CDRI), SEQ ID NO:143 (CDR2) and SEQ ID NO:144 (CDR3); ix) an antigen binding unit comprising light chain CDRs comprising the amino 15 acid sequences of SEQ ID NO:145 (CDRl), SEQ ID NO:146 (CDR2) and SEQ ID NO:147 (CDR3) and heavy chain CDRs comprising the amino acid sequences of SEQ ID NO:148 (CDRI), SEQ ID NO:149 (CDR2) and SEQ ID NO:150 (CDR3); an antigen binding unit comprising light chain CDRs comprising the amino 20 acid ces of SEQ ID NO:151 (CDRl), SEQ ID NO:152 (CDR2) and SEQ ID NO:153 (CDR3) and heavy chain CDRs comprising the amino acid ces of SEQ ID NO:154 (CDRl), SEQ ID NO:155 (CDR2) and SEQ ID NO:156 (CDR3); xi) an antigen binding unit comprising light chain CDRs comprising the amino 25 acid ces of SEQ ID NO:157 (CDRl), SEQ ID NO:158 (CDR2) and SEQ ID NO:159 (CDR3) and heavy chain CDRs comprising the amino acid sequences of SEQ ID NO:16O (CDRl), SEQ ID NO:161 (CDR2) and SEQ ID NO:162 (CDR3); xii) an n binding unit sing light chain CDRs comprising the amino 30 acid sequences of SEQ ID NO:163 (CDRl), SEQ ID NO:164 (CDR2) and SEQ ID NO:165 (CDR3) and heavy chain CDRs comprising the amino acid —l78— WO 20191234220