The present invention relates to anti-RORl antibodies, including bispecific and multispecific antibodies, Immunoligand-Toxin Conjugates targeting RORl, and anti RORl CARs and CAR cells.

Introduction

For several neoplastic diseases, no efficacious therapeutic approaches exist. This applies, inter alia, to chronic lymphocytic leukemia (CLL), acute lymphocytic leukemia (ALL), non- Hodgkin lymphomas (NHL), and myeloid malignancies, as well as to solid cancers including colon, lung, breast, ovarian and pancreatic cancers.

It is an object of the present invention to provide new therapeutic approaches to address these diseases.

Preferred embodiments

It is to be understood that embodiments disclosed herein are not meant to be understood as individual embodiments which would not relate to one another. Features discussed with one embodiment are meant to be disclosed also in connection with other embodiments shown herein. If, in one case, a specific feature is not disclosed with one embodiment, but with another, the skilled person would understand that does not necessary mean that said feature is not meant to be disclosed with said other embodiment. The skilled person would understand that it is the gist of this application to disclose said feature also for the other embodiment, but that just for purposes of clarity and to keep the specification in a manageable volume this has not been done.

According to a first embodiment of the invention, a human or humanized antibody, or an antibody-based binding protein, or a modified antibody format retaining target binding capacity, or an antibody derivative or fragment retaining target binding capacity is provided, which targets Receptor tyrosine kinase-like orphan receptor-1 (ROR1).

In a second embodiment also an Immunoligand-Drug conjugate of said first embodiment with a functional moiety covalently coupled to a human or humanized antibody, or an antibody- based binding protein, or a modified antibody format retaining target binding capacity, or an antibody derivative or fragment retaining target binding capacity is provided, which targets Receptor tyrosine kinase-like orphan receptor-1 (ROR1). This conjugate is preferably an antibody drug conjugate (ADC), to which preferably a small molecular weight cellular toxin is conjugated, preferably site-specifically, and preferably by, but not limited to sortase- enzyme mediated conjugation technology (SMAC -technology) disclosed in WO2014140317.

In a third embodiment, also mammalian cells carrying receptors comprising a human or humanized antibody, or an antibody-based binding protein, or a modified antibody format retaining target binding capacity, or an antibody derivative or fragment retaining target binding capacity for Receptor tyrosine kinase-like orphan receptor-1 (ROR1) is provided. Such mammalian cells are preferably T cells of the immune system, carrying preferably chimeric antigen receptors (CARs) comprising said human or humanized antibody, or an antibody-based binding protein, or a modified antibody format retaining target binding capacity, or an antibody derivative or fragment retaining target binding capacity for Receptor tyrosine kinase-like orphan receptor-1 (ROR1). In a further preferred embodiment, these mammalian cells are therefore CAR T cells comprising said human or humanized antibody, or an antibody-based binding protein, or a modified antibody format retaining target binding capacity, or an antibody derivative or fragment retaining target binding capacity for Receptor tyrosine kinase-like orphan receptor-1 (ROR1)

Receptor tyrosine kinase-like orphan receptor-1 (ROR1), is a member of the receptor tyrosine kinase (RTK) family of cell surface molecules, that is highly expressed during embryonic and fetal development and which is important for embryonic and fetal development, because ROR-1 knock-out mice die neonatally (Nomi M et al. (2001) Mol Cell Biol. 21, 8329-35). In postpartum human tissue, ROR1 expression was found to be mostly absent. Low expression has been observed in adipose tissue, and to a minor extent in pancreas, lung, and a subset of intermediate B cells. However, importantly, elevated ROR1 expression has been found in several cancers, including hematological malignancies as chronic lymphocytic leukemia (CLL), acute lymphocytic leukemia (ALL), mantle cell lymphoma, non-Hodgkin lymphomas (NHL), and myeloid malignancies, as well as in solid cancers, including neuroblastoma, sarcoma, renal cell carcinoma, melanoma, colon cancer, lung cancer, breast cancer, ovarian cancer, head and neck cancer, and pancreatic cancer. Therefore, the expression patter of ROR1 is referred to as onco-fetal expression pattern and therefore is of high medical interest as a target for particular cancers and tumors (Borcherding N et al (2014) Protein Cell 5, 496- 502; Hojjat-Farsangi et al (2014) Seminars in Cancer Biology, 29, 21-31).

In line with the virtual absence of ROR1 expression in postpartum tissue, no toxicity was observed when non-human primates that were treated with ROR1 -specific cytotoxic T cells suggesting that specific targeting of ROR1 in humans for cancer therapy is safe (Berger C et al. (2015) Cancer Immunol Res. 3, 206-16). Thus far, no RORl-specific antibodies have clinically been approved for cancer therapy.

To be applicable for the therapy of human disease, antibodies targeting ROR-1, and Immunoligand-Toxin-Conjugates targeting ROR-1, should show minimal immunogenicity, thus minimizing the possibility to be cleared from the system by the immune-system of patients and thus resulting in prolonged serum half-life and, consequently, higher efficiency. Further, reduced immunogenicity avoids unwanted and sometimes even life threatening side effects, like immunogenic reactions.

Therefore the present invention provides human and humanized anti-ROR-1 antibodies, which are expected to result in minimal immunogenicity and with efficacy in the treatment of neoplastic conditions.

In one specific embodiment, humanized anti-ROR-1 antibodies, or antibody-based binding proteins, modified antibody formats retaining target binding capacity, antibody derivatives or fragmentss retaining target binding capacity are provided. The term "humanized antibody" refers to a chimeric antibody that contains sequences derived from human and non-human (e.g., rabbit) immunoglobulins such that substantially all of the CDR regions are of non-human origin, while substantially all of the FR regions correspond to those of a human immunoglobulin sequence. In one embodiment, the antibody has been humanized from a rodent or rabbit parent antibody, i.e., comprises CDR regions tghat are of rodent or rabbit origin.

The humanized antibodies have been developed on the basis of functionally characterized anti-RORl rodent antibodies from mouse and rabbit origin, named R11, R12 (both rabbit) (WO 2012/075158 Al) and 2A2 (mouse) (WO 2010/124188 Al), which all bind different extracellular domains of the ROR1 molecule, namely the Ig, the frizzled and the Kringle domains of ROR1 (Yang J et al. (2011) PLoS ONE 6, e21018; Baskar S et al. (2012) mAbs 4, 1-13).

These antibodies have been humanized by proprietary algorithms, on the basis of sequence alignments to large next-generation sequencing data from human antibody repertoires and selecting human frameworks with maximal similarity to human IgG heavy and light chain sequences. Collections of identified human framework regions with desired sequence homology to the parental antibody clones have been selected and combined with the CDRs of high affinity rodent mAbs R11, R12 and 2A2.

Collections of up to 183 predicted humanized antibodies have been gene synthesized and libraries of individual IgG heavy and light chains have stably been expressed and displayed on the surface of mammalian cells using transposition-mediated antibody Display (Transpo- mAb) as described (Waldmeier et al. (2016) MAbs 8(4): 726-740, WO2014013026). Individual combinations of humanized antibody heavy and light chains have been selected based on ROR1 target binding efficiency and best clones were analyzed for sequence recovery. Based on the screening of larger repertoires of humanized IgG heavy and light chains, high-affinity humanized mAbs could be identifies without further affinity maturation.

The provided humanized anti-ROR-1 antibodies show, in some cases, unexpectedly significant improvements in the binding affinity towards ROR1 relative to the parental antibodies from which they were derived. This is unexpected because humanization of antibodies usually leads to a reduction or loss in affinity affinity, that tediously needs to be compensated by additional affinity maturation (Baca et al. (1997) J. Biol. Chem. 271 : 10678- 84). a preferred embodiment, the antibody comprises at least the 3 CDR sequences

In another preferred embodiment, the antibody comprises at least the 3 CDR sequences:

In both cases, it is to be understood that the definition of the CDR ("Complementarity

Determining Region") is based on the "IMGT unique numbering for all IG and TR V- REGIONs of all species: interest for structure and evolution". Further, "CDR LC" means Light Chain CDR, while "CDR HC" means Heavy Chain CDR.

In a preferred embodiment, the antibody comprises at least one heavy chain or light chain variable region sequence that is 95 % identical, preferably 96 or even 97 % identical, more preferably 98 % or even 99 % identical, and most preferably 100 % to a sequence selected from the group consisting of:

„VR HC" means Heavy Chain Variable Sequence, while "VR LC" means Light Chain Variable Sequence. a preferred embodiment, the antibody is humanized from

• mouse anti ROR antibody mAb 2A2,

• rabbit anti ROR antibody mAb R11 , and/or

• rabbit anti ROR antibody mAb R12, and/or is selected from the group consisting of

• hu2A2b-Q 11 (also called msQ 11-3)

• hu2 A2-D23 (also called msD23 - 19)

• hu2A2-D 16 (also called msD 16-7)

preferably, but not limited to having better than 11 nM affinity

and/or antibodies sharing at least 95%, preferably 96 or even 97 % identical, more preferably 98 %> or even 99 %> identical, and most preferably 100 %> amino acid sequence identify with any of the antibodies mentioned above.

In one embodiment, the antibody has an affinity of better than 1.0 nM to human ROR1.

In one other embodiment, the antibody humanized from rabbit anti ROR antibody mAb R12 has a better affinity than its parent, i.e., rabbit monoclonal antibody R12.

In a preferred embodiment, the antibody has at least one of the characteristics set forth in table 1 or table 2.

According to yet another embodiment of the invention, a human or humanized antibody is provided, or an antibody-based binding protein, modified antibody format retaining target binding capacity, antibody derivative or fragment retaining target binding capacity, which

(i) has a binding affinity for ROR1 that is at least as high or substantially as high as the binding affinity of an antibody, antibody-based binding protein, modified antibody format, antibody derivative or fragment according to any of the antibodies described above, and/or

(ii) competes with an antibody, antibody-based binding protein, modified antibody format, antibody derivative or fragment according to any of the antibodies described above for binding to ROR1. In one embodiment of the antibody, antibody-based binding protein, modified antibody format, antibody derivative or fragment according to the above description, the Receptor tyrosine kinase-like orphan receptor- 1 (ROR1 ) is human ROR1.

According to another embodiment of the invention, the antibody-based binding protein, modified antibody format, antibody derivative or fragment of any of the aforementioned claims is a bispecific antibody or a multispecific antibody.

The terms "bispecific antibody" and "multispecific antibody" refers to an antibody having the capacity to bind to two, or more, distinct epitopes either on a single antigen or two different antigens, out of which one is ROR1. Bispecific antibodies of the present invention can be produced via biological methods, such as somatic hybridization; or genetic methods, such as the expression of a non-native DNA sequence encoding the desired antibody structure in an organism; chemical methods, such as chemical conjugation of two antibodies; or a combination thereof (Kontermann, R.E. In: Bispecific Antibodies. Kontermann RE (ed.), Springer Heidelberg Dordrecht London New York, pp. 1 -28 (201 1 )).

Chemically conjugated bispecific antibodies arise from the chemical coupling of two existing antibodies or antibody fragments. Typical couplings include cross-linking two different full- length antibodies, cross-linking two different Fab' fragments to produce a bispecific F(ab')2, and cross-linking a F(ab')2 fragment with a different Fab' fragment to produce a bispecific F(ab')3. For chemical conjugation, oxidative reassociation strategies can be used. Current methodologies include the use of the homo- or heterobifunctional cross-linking reagents (Id.). Heterobifunctional cross-linking reagents have reactivity toward two distinct reactive groups on, for example, antibody molecules. Examples of heterobifunctional cross-linking reagents include SPDP (N-succinimidyl-3-(2- pyridyldithio)propionate), SATA (succinimidyl acetylthioacetate), SMCC (succinimidyl trans-4-(maleimidylmethyl) cyclohexane-1 carboxylate), EDAC (1 -ethyl-3- (3- dimethylaminopropyl) carbodiimide), PEAS (N-((2- pyridyldithio)ethyl)-4- azidosalicylamide), ATFB, SE (4-azido-2,3,5,6-tetrafluorobenzoic acid, succinimidyl ester), benzophenone-4-maleimide, benzophenone-4-isothiocyanate, 4- benzoylbenzoic acid, succinimidyl ester, iodoacetamide azide, iodoacetamide alkyne, Click- iT maleimide DIBO alkyne, azido (PEO)4 propionic acid, succinimidyl ester, alkyne, succinimidyl ester, Click-iT succinimidyl ester DIBO alkyne, Sulfo- SBED (Sulfo-N- hydroxysuccinimidyl-2-(6-[biotinamido]-2-(p-azido benzamido)- hexanoamido) ethyl-1 ,3'- dithioproprionate), photoreactive amino acids {e.g., L-Photo- Leucine and L-Photo- Methionine), NHS-haloacetyl crosslinkers such as, for example, Sulfo-SIAB, SIAB, SBAP, SIA, NHS- maleimide crosslinkers such as, for example, Sulfo-SMCC, SM(PEG)n series crosslinkers, SMCC, LC-SMCC, Sulfo- EMCS, EMCS, Sulfo-GMBS, GMBS, Sulfo-KMUS, Sulfo-MBS, MBS, Sulfo-SMPB, SMPB, AMAS, BMPS, SMPH, PEG12-SPDP, PEG4- SPDP, Sulfo-LC-SPDP, LC- SPDP, SMPT, DCC (N, N'-Dicyclohexylcarbodiimide), EDC (1 -Ethyl-3-(3- dimethylaminopropyl) carbodiimide), NHS (N-hydroxysuccinimide), Sulfo-NHS (N- hydroxysulfosuccinimide), BMPH, EMCH, KMUH, MPBH, PDPH, and PMPI.

Homobifunctional cross-linking reagents have reactivity toward the same reactive group on a molecule, for example, an antibody. Examples of homobifunctional cross-linking reagents include DTNB (5,5' -dithiobis(2-nitrobenzoic acid), o-PDM (o-phenylenedimaleimide), DMA (dimethyl adipimidate), DMP (dimethyl pimelimidate), DMS (dimethyl suberimidate), DTBP (dithiobispropionimidate), BS(PEG)5, BS(PEG)9, BS3, BSOCOES, DSG, DSP, DSS, DST, DTSSP, EGS, Sulfo-EGS, TSAT, DFDNB, BM(PEG)n crosslinkers, BMB, BMDB, BMH, BMOE, DTME, and TMEA. [0030] Somatic hybridization is the fusion of two distinct hybridoma (a fusion of B cells that produce a specific antibody and myeloma cells) cell lines, producing a quadroma capable of generating two different antibody heavy (VHA and VHB) and light chains (VLA and VLB). (Kontermann, R.E. In: Bispecific Antibodies. Kontermann RE (ed.), Springer Heidelberg Dordrecht London New York, pp. 1 -28 (201 1 )). These heavy and light chains combine randomly within the cell, resulting in bispecific antibodies (a VHA combined with a VLA and a VHB combined with a VLB), as well as some nonfunctional (e.g. two VHAs combined with two VLBs) and monospecific (two VHAs combined with two VLAs) antibodies. The bispecific antibodies can then be purified using, for example, two different affinity chromatography columns. Similar to monospecific antibodies, bispecific antibodies may also contain an Fc region that elicits Fc-mediated effects downstream of antigen binding. These effects may be reduced by, for example, proteolytically cleaving the Fc region from the bispecific antibody by pepsin digestion, resulting in bispecific F(ab')2 molecules (Id.).

Bispecific antibodies may also be generated via genetic means, e.g., in vitro expression of a plasmid containing a DNA sequence corresponding to the desired antibody structure. See, e.g., Kontermann, R.E. In: Bispecific Antibodies. Kontermann RE (ed.), Springer Heidelberg Dordrecht London New York, pp. 1 -28 (201 1 ). Such bispecific antibodies are discussed in greater detail below.

A bispecific antibody of the present invention may be bivalent, trivalent, or tetravalent. As used herein, "valent", "valence", "valencies", or other grammatical variations thereof, mean the number of antigen binding sites in an antibody molecule.

Further provided are a) an isolated nucleic acid sequence, or a set thereof, that encodes an antibody, antibody- based binding protein, modified antibody format, antibody derivative or fragment according to the above description, or a bispecific antibody or a multispecific antibody according to the above description,

b) a vector comprising at least one such nucleic acid sequence,

c) an isolated cell expressing an antibody, antibody-based binding protein, modified antibody format, antibody derivative or fragment according to the above description, or a bispecific antibody or a multispecific antibody according to the above description, d) and/or comprising a nucleic acid sequence, or a set thereof, according to the above description, or a vector according to the above description, and

e) a method of producing an antibody, antibody-based binding protein, modified

antibody format, antibody derivative or fragment according to the above description, comprising culturing of a cell according to the above description, and purification of the antibody, antibody-based binding protein, modified antibody format, antibody derivative or fragment.

According to another aspect of the invention, an Immunoligand-Drug Conjugate having the general formula A - (L)n - (T) m is provided, in which

• A is an Immunoligand targeting Receptor tyrosine kinase-like orphan receptor- 1

(ROR1),

• L is a linker,

• T is a toxin and in which n and m are integers between >1 and < 10 In this construct, (L)n can mean several linkers which form a unitary chain that conjugates one toxin to the one Immunoligand, and/or several linkers which connect several toxins to the one Immunoligand. Likewise, (L)n can mean several linkers which conjugate two

Subdomains of the same Immunologand to two toxin molecules.

The resulting Immunoligand-Toxin-Conjugate would thus have a Toxin/Immunoligand ratio of >1 and < 10. Preferably, n and m are integers between >1 and < 4. The resulting

Immunoligand-Toxin-Conjugate would thus have an Toxin/Immunoligand ratio of >1 and < 4.

As used herein, the term "immunoligand" is meant to define an entity, an agent or a molecule that has affinity to a given target, e.g., a receptor, a cell surface protein, a cytokine or the like. Such Immunoligand may optionally block or dampen agonist-mediated responses, or inhibit receptor-agonist interaction. Most importantly, however, the immunoligand may serve as a shuttle to deliver a payload to a specific site, which is defined by the target recognized by said immunoligand. Thus, an Immunoligand targeting a receptor, delivers its payload to a site which is characterized by abundance of said receptor.

In a further preferred embodiment, the Immunoligand is at least one selected from the group consisting of an

• antibody,

• antibody-based binding protein

• modified antibody format retaining target binding capacity,

• antibody derivative or fragment retaining target binding capacity, and/or

• bispecific antibody or a multispecific antibody. "Antibodies", also synonymously called "immunoglobulins" (Ig), are generally comprising four polypeptide chains, two heavy (H) chains and two light (L) chains, and are therefore multimeric proteins, or an equivalent Ig homologue thereof (e.g., a camelid antibody, which comprises only a heavy chain, single domain antibodies (dAbs) which can be either be derived from a heavy or light chain); including full length functional mutants, variants, or derivatives thereof (including, but not limited to, murine, chimeric, humanized and fully human antibodies, which retain the essential epitope binding features of an Ig molecule, and including dual specific, bispecific, multispecific, and dual variable domain immunoglobulins; Immunoglobulin molecules can be of any class (e.g., IgG, IgE, IgM, IgD, IgA, and IgY), or subclass (e.g., IgGl, IgG2, IgG3, IgG4, IgAl, and IgA2) and allotype.

An "antibody-based binding protein", as used herein, may represent any protein that contains at least one antibody-derived VH, VL, or CH immunoglobulin domain in the context of other non-immunoglobulin, or non-antibody derived components. Such antibody-based proteins include, but are not limited to (i) F _c -fusion proteins of binding proteins, including receptors or receptor components with all or parts of the immunoglobulin CH domains, (ii) binding proteins, in which VH and or VL domains are coupled to alternative molecular scaffolds, or (iii) molecules, in which immunoglobulin VH, and/or VL, and/or CH domains are combined and/or assembled in a fashion not normally found in naturally occuring antibodies or antibody fragments.

An "antibody drug conjugate" (ADC), as used herein, relates to either an antibody, or an antibody fragment, or an antibody-based binding protein, coupled to a small molecular weight active pharmaceutical ingredient (API), including, but not limited to a toxin (including e.g., but not limited to, tubulin inhibitors, actin binders, R A polymerase inhibitors, DNA- intercalating and modifying/damaging drugs), a kinase inhibitor, or any API that interferes with a particular cellular pathway that is essential for the survival of a cell and/or essential for a particular physiologic cellular pathway.

An "antibody derivative or fragment", as used herein, relates to a molecule comprising at least one polypeptide chain derived from an antibody that is not full length, including, but not limited to (i) a Fab fragment, which is a monovalent fragment consisting of the variable light (VL), variable heavy (VH), constant light (CL) and constant heavy 1 (CHI) domains; (ii) a F(ab')2 fragment, which is a bivalent fragment comprising two Fab fragments linked by a disulfide bridge at the hinge region; (iii) a heavy chain portion of a F _a b (Fa) fragment, which consists of the VH and CHI domains; (iv) a variable fragment (F _v ) fragment, which consists of the VL and VH domains of a single arm of an antibody, (v) a domain antibody (dAb) fragment, which comprises a single variable domain; (vi) an isolated complementarity determining region (CDR); (vii) a single chain F _v Fragment (scF _v ); (viii) a diabody, which is a bivalent, bispecific antibody in which VH and VL domains are expressed on a single polypeptide chain, but using a linker that is too short to allow for pairing between the two domains on the same chain, thereby forcing the domains to pair with the complementarity domains of another chain and creating two antigen binding sites; (ix) a linear antibody, which comprises a pair of tandem F _v segments (VH-CH1-VH-CH1) which, together with complementarity light chain polypeptides, form a pair of antigen binding regions; (X) Dual- Variable Domain Immunoglobulin (xl) other non-full length portions of immunoglobulin heavy and/or light chains, or mutants, variants, or derivatives thereof, alone or in any combination.

The term "modified antibody format", as used herein, encompasses antibody-drug- conjugates, Polyalkylene oxide-modified scFv, Monobodies, Diabodies, Camelid Antibodies, Domain Antibodies, bi- or trispecific antibodies, IgA, or two IgG structures joined by a J chain and a secretory component, shark antibodies, new world primate framework + non-new world primate CDR, IgG4 antibodies with hinge region removed, IgG with two additional binding sites engineered into the CH3 domains, antibodies with altered Fc region to enhance affinity for Fc gamma receptors, dimerised constructs comprising CH3+VL+VH, and the like. The term "antibody mimetic", as used herein, refers to proteins not belonging to the immunoglobulin family, and even non-proteins such as aptamers, or synthetic polymers. Some types have an antibody-like beta-sheet structure. Potential advantages of "antibody mimetics" or "alternative scaffolds" over antibodies are better solubility, higher tissue penetration, higher stability towards heat and enzymes, and comparatively low production costs.

Another preferred embodiment is an Immunoligand comprising at least one antibody or antibody fragment with binding capacity to ROR1 as set forth in the above disclosure.

An "Immunoligand-Drug Conjugate" (IDC), as used herein, relates to a molecule that comprises a binding moiety of a humanized anti ROR1 antibody or antibody-based binding protein as disclosed herein, coupled to a small molecular weight active pharmaceutical ingredient (API), including, but not limited to a toxin (including e.g., but not limited to, tubulin inhibitors, actin binders, RNA polymerase inhibitors, DNA-intercalating and modifying/damaging drugs), a kinase inhibitor, or any API that interferes with a particular cellular pathway that is essential for the survival of a cell and/or essential for a particular physiologic cellular pathway.

Another preferred embodiment is an Immunoligand-Drug Conjugate as disclosed above comprising covalent a linker between an Immunoligand and preferably a small molecular weight active pharmaceutical ingredient (API).

In another preferred embodiment, said linker is at least one selected from the group consisting of

• an oligopeptide linker, optionally comprising cleavable spacers, that may be cleaved by changes in pH, redox potential and or specific intracellular enzymes and/or

• a maleimide linker, optionally comprising cleavable spacers, that may be cleaved by changes in pH, redox potential and or specific intracellular enzymes

In a preferred embodiment, the linker comprises, or consists of, at least one selected from the group consisting of: an oligopeptide linker (including cleavable and non-cleavable oligopeptide linkers), a hydrazine linker, a thiourea linker, a self-immolative linker, a succinimidyl trans-4-(maleimidylmethyl)cyclohexane-l-carboxylate (SMCC) linker, a maleimide linker, a disulfide linker, a thioether linker, and/or a maleimide linker.

The skilled person understands that further linkers may be suitable. Such linkers may be non- cleavable or may be cleaved by changes in pH, redox potential or specific intracellular or tumor tissue associated enzymes. Cleavable oligopeptide linkers include protease- or matrix metalloprotease-cleavable linkers. It is understood that the linker may comprise combinations of the above. For example, the linker may be a valine-citruline PAB linker.

In a preferred embodiment, the linker comprises an oligopeptide that is recognized by sortase enzymes, including but not limited to amino acid sequences selected from LPXSG _n , LPXAG _n , LPXTGn, LAXTGn, LAETGn, LPXTAn or NPQTG _n with n being an integer between > 1 and < 21, and X being any amino acid.

In still another preferred embodiment, the linker comprises an oligopeptide of the sequence LPXTGn with n being an integer between > 1 and < 20, and X being any naturally occurring amino acid.

In another preferred embodiment, the linker is conjugated to the C-terminus of at least one subdomain of the Immunoligand.

In another preferred embodiment, prior to conjugation,

• the immunoligand bears a sortase recognition tag used or conjugated to the C-terminus of at least one subdomain thereof, and

• the toxin comprises a short glycine stretch with a length of 1-20 glycine residues, preferably with a length of 3 to 5 amino acids.

Preferably, the sortase recognition tag is:

• LPXTG or LPXAG, which is recognized by Staphylococcus aureus sortase A;

• LPXSG, which is recognized by Staphylococcus aureus sortase A or an engineered sortase A 4S-9 from Staphylococcus aureus;

• LAXTG, and particularly LAETG, which is recognized by engineered sortase A 2A-9 from Staphylococcus aureus;

• LPXTA, which is recognized by Streptococcus pyogenes sortase A; or

• NPQTN, which is recognized by Staphylococcus aureus sortase B.

The following table shows the recognition tags and the peptides derived therefrom to be part of the linker (with n being an integer between > 1 and < 21, and X being any amino acid):

Engineered sortases, including A 2A-9 and A 4S-9 from Staphylococcus aureus, are described in Dorr BM et al, PNAS 2014; 111, 13343-8., and Chen et al, PNAS 2011; 108(28); 11399- 11404.

As background and to exemplify the general concept of sortase transpeptidation, Sortase A, for example, uses an oligo-glycine-stretch as a nucleophile to catalyze a transpeptidation by which the terminal amino group of the oligo-glycine effects a nucleophilic attack on the peptide bond joining the last two C-terminal residues of the sortase tag. This results in breakage of that peptide bond and the formation of a new peptide bond between the C- terminally second-to-last residue of the sortase tag and the N-terminal glycine of the oligo- glycine peptide, i.e. resulting in a transpeptidation.

Prior to sortase conjugation, the sortase tag may, at its C-terminus, furthermore carry other tags, like His-tags, Myc-tags or Strep-tags (see Fig. 4a of WO2014/140317, the content of which is incorporated by reference herein). However, because the peptide bond between the 4th and 5th amino acid of the sortase tag is cleaved upon sortase A mediated conjugation, these additional tags do not appear in the conjugated product.

Sortase tag may, for example, be fused to a C-terminus of a binding protein, or to a domain or subunit thereof, by genetic fusion, and are co-expressed therewith. In another preferred embodiment, the sortase tag may directly be appended to the last naturally occurring C- terminal amino acid of the immunoglobulin light chains or heavy chains, which in case of the human immunoglobulin kappa light chain is the C-terminal cysteine residue, and which in the case of the human immunoglobulin IgGl heavy chain may be the C-terminal lysine residue encoded by human Fcyl cDNA. However, another preferred embodiment is also to directly append the sortase tag to the second last C-terminal glycine residue encoded by human Fcyl cDNA, because usually terminal lysine residues of antibody heavy chains are clipped off by posttranslational modification in mammalian cells. Therefore, in more than 90% of the cases naturally occurring human IgGl lacks the C-terminal lysine residues of the IgGl heavy chains.

Therefore, one preferred embodiment of the invention is to omit the C-terminal lysine amino acid residues of human IgGl heavy chain constant regions in expression constructs for sortase recognition-motif tagged Igyl heavy chains. Another preferred embodiment is to include the C-terminal lysine amino acid residues of human IgGl heavy chain constant regions in expression constructs for sortase recognition-motif tagged Igyl heavy chains.

In another preferred embodiment the sortase tag may be appended to the C-terminus of a human immunoglobulin IgGl heavy chain where the C-terminal lysine residue encoded by human Fcyl cDNA is replaced by an amino acid residue other than lysine to prevent unproductive reactions of sortase with the ε -amino group of said C-terminal lysine residue leading to inter-heavy chain crosslinking.

We have described previously that in some cases (e.g. at the C-terminus of the Ig kappa light chains, see: Beerli et al. (2015) PloS One 10, el31 177) it is beneficial to add additional amino acids between the C-terminus of the binding protein and the sortase tag. This has been shown to improve sortase enzyme conjugation efficiencies of payloads to the binding protein. In the case of Ig kappa light chains, it was observed that by adding 5 amino acids between the last C-terminal cysteine amino acid of the Ig kappa light chain and the sortase pentapeptide motif improved the kinetic of conjugation, so that the C-termini of Ig kappa light chains and Ig heavy chains could be conjugated with similar kinetics (see: Beerli et al. (2015) PloS One 10, el31177). Therefore, it is another preferred embodiment that optionally > 1 and < 11 amino acids are added in between the last C-terminal amino acid of a binding protein or antibody subunit and the sortase tag.

Further, the immunoligand can comprise, C-terminally of the sortase tag, other tags, like for instance, but not limited to a His tag, a Myc tag and/or a Strep II tag. See WO2014140317 A2 for more details, the subject matter of which is incorporated by reference herein

In another preferred embodiment, the toxin is at least one selected from the group consisting of • maytansinoids,

• auristatins,

• anthracyclins, preferably PNU-derived anthracyclins

• calicheamicins,

• tubulysins

• duocarmycins

• radioisotopes

• liposomes comprising a toxid paylooad,

• protein toxins

• taxanes, and/or

• pyrrolbenzodiazepines

Examples for preferred maytansinoid toxins are shown in Fig. 1 and 2. The anthracycline derivatives disclosed herein also referred to as "PNU" are derivatives of PNU- 159682, which is a metabolite of the anthracycline nemorubicin and has for the first time been disclosed by Quintierei et al. 2005. PNU- 159682 is shown Fig. 5. A preferred anthracycline derivative is depicted in Fig. 3 (a).

Immunoligand Drug Conjugates comprising anthracycline derivatives are disclosed in WO2016102697 and applications claiming the priority thereof, the content of which is incorporated by reference herein.

Preferably, the Immunoligand-Drug Conjugates comprises two or more different toxins. In such way, the cell killing activity can be enhanced, by avoiding resistances against single toxins comprised in the Immunoligand-Drug Conjugates

In another preferred embodiment, the Immunoligand-Drug Conjugate has a cell killing activity as set forth in Fig. 8 or Fig. 12.

In another embodiment, the Immunoligand-Drug Conjugate is created by sortase-mediated conjugation of (i) an Immunoligand carrying one or more sortase recognition tags and (ii) one or more toxins carrying an oligoglycine tag. According to another aspect of the invention, a method of producing an Immunoligand-Drug Conjugate according to any of the aforementioned disclosure is provided, which method comprises the following steps: a) providing an Immunoligand according to the list set forth above, which Immunoligand carries a sortase recognition tag,

b) providing one or more toxins carrying an oligoglycine tag, and

c) conjugating the Immunoligand and the toxin by means of sortase-mediated

conjugation.

The method of conjugating an Immunoligand to a payload by means of a sortase or a split intein is disclosed in full detail in WO2014140317, the subject matter of which is incorporated by reference herein.

According to yet another rembodiment, a ROR1 specific chimeric antigen receptor (CAR) is provided, comprising a) at least one antibody, antibody-based binding protein, modified antibody format or antibody derivative or fragment according to the above description, or

b) a bi-or multispecific antibody according to the above description, which is fused or conjugated to at least one transmembrane region and at least one

intracellular domain.

Chimeric antigen receptors (CAR), sometimes also called artificial T cell receptors, are engineered receptors, which graft an arbitrary specificity onto an immune effector cell. Typically, these receptors are used to graft the specificity of a monoclonal antibody onto a T cell; with transfer of their coding sequence facilitated by retroviral vectors. The receptors are called chimeric because they are composed of parts from different sources.

CARs are potential candidates as a therapy for cancer, using a technique called adoptive cell transfer. T cells are removed from a patient and modified so that they express CARs specific to the patient's particular cancer, by specifically binding to a cancer-specific antigen, as is the case in ROR1. The T cells, which can then recognize and kill the cancer cells, are reintroduced into the patient.

The structure of the prototypic CAR is modular, designed to accommodate various functional domains and thereby to enable choice of specificity and controlled activation of T cells. In the context of the present invention, a CAR comprises an antibody-like binding domain derived from an antibody, antibody-based binding protein, modified antibody format, antibody derivative or fragment, which targets ROR1. Such entity can be, e.g., but is not limited to, a single chain variable fragment (scFv) that combines the specificity and binding residues of both the heavy and light chain variable regions of a monoclonal antibody in a single polypeptide chain, fused or conjugated to at least one transmembrane region and at least one intracellular domain.

Preferably, said transmembrane region comprises a CD8a transmembrane domain. Preferably, said CAR further comprises a hinge region disposed between the transmembrane domain and the antibody, antibody-based binding protein, modified antibody format retaining target binding capacity, or antibody derivative. Preferably, said intracellular domain comprises a T- cell receptor signaling domain. More preferably, said signaling domain comprises or is derived from a zeta chain of a CD3-zeta chain. Preferably, said intracellular domain further comprises one or more intracellular signaling domain of a T cell costimulatory molecule.

A preferred intracellular signaling domain of a T cell costimulatory molecule is selected from the group consisting of 4-lBB, CD-28, OX40 and/or CD278/ICOS. Combination of two or more of these domains are specifically preferred.

According to another embodiment of the invention, a cell comprising such chimeric antigen receptor is provided.

Said cell is preferably an engineered T cell, also called "CAR T cell". CAR T cells are genetically engineered T cells armed with CARs whose extracellular recognition unit is comprised of an antibody-derived recognition domain and whose intracellular region is derived from lymphocyte stimulating moiety(ies). By arming T cells with such chimeric receptors, the engineered cell is redirected with a predefined specificity to any desired target antigen, in a non-HLA restricted manner. CAR constructs are introduced ex vivo into T cells from peripheral lymphocytes of a given patient using retroviral or lentiviral vectors or transposable elements. Following infusion of the resulting CAR-engineered T cells back into the patient, they traffick, reach their target site, and upon interaction with their target cell or tissue, they undergo activation and perform their predefined effector function. Therapeutic targets for the CAR approach include cancer and HIV-infected cells, or autoimmune effector cells. Alternatively, said cell is preferably an engineered natural killer cell (NK cell).

Another aspect of the invention is the use of the antibody-based binding protein, modified antibody format retaining target binding capacity, antibody derivative or fragment of any of claims according to the above description, the bi-or multispecific antibody according to the above description, the Immunoligand-Drug Conjugate according to the above description, or the CAR or cell according to the above description, for the treatment of a patient that is

• suffering from,

• at risk of developing, and/or

• being diagnosed for a neoplastic disease.

In a preferred embodiment, the neoplastic disease is at least one selected from the group consisting of

• blood cancer, such as lymphomas and leukemias, like chronic lymphocytic leukemia (CLL), acute lymphocytic leukemia (ALL), non-Hodgkin lymphomas (NHL), Mantle Cell lymphpma (MCL), acute myeloid leukemia (AML), and

• solid cancers such as cancer of colon, lung, breast, ovarian and pancreatic cancers

According to a further aspect of the invention, a pharmaceutical composition is provided, which comprises the antibody or antibody-based binding protein, modified antibody format retaining target binding capacity, antibody derivative or fragment according to the above description, the bi-or multispecific antibody according to the above description, the Immunoligand-Drug Conjugate according to the above description, or the CAR or cell according to the above description, together with one or more pharmaceutically acceptable ingredients. According to a further aspect of the invention, a method of killing or inhibiting the growth of a cell expressing ROR1 in vitro or in a patient is provided, which method comprises administering to the cell a pharmaceutically effective amount or dosis of (i) the antibody or antibody-based binding protein, modified antibody format retaining target binding capacity, antibody derivative or fragment according to the above description, the bi-or multispecific antibody according to the above description, the Immunoligand-Drug Conjugate according to the above description, or the CAR or cell according to the above description, or (ii) of a pharmaceutical composition according to the above description

Preferably, wherein the cell expressing ROR1 is a cancer cell, preferably, blood cancer, such as lymphomas and leukemias, like chronic lymphocytic leukemia (CLL), acute lymphocytic leukemia (ALL), non-Hodgkin lymphomas (NHL), Mantle Cell lymphpma (MCL), acute myeloid leukemia (AML), or solid cancer such as cancer of colon, lung, breast, ovarian and pancreatic cancer.

Further preferably, the ROR1 is human ROR1.

Experiments and Figures

While the invention has been illustrated and described in detail in the drawings and foregoing description, such illustration and description are to be considered illustrative or exemplary and not restrictive; the invention is not limited to the disclosed embodiments. Other variations to the disclosed embodiments can be understood and effected by those skilled in the art in practicing the claimed invention, from a study of the drawings, the disclosure, and the appended claims. In the claims, the word "comprising" does not exclude other elements or steps, and the indefinite article "a" or "an" does not exclude a plurality. The mere fact that certain measures are recited in mutually different dependent claims does not indicate that a combination of these measures cannot be used to advantage. Any reference signs in the claims should not be construed as limiting the scope.

All amino acid sequences disclosed herein are shown from N-terminus to C-terminus; all nucleic acid sequences disclosed herein are shown 5'->3'.

EXPERIMENTAL RESULTS Generation and selection of humanized anti-RORl antibodies via Transpo-mAb Display screening

In order to humanized mouse 2A2 (Baskar S et al. (2012) mAbs 4, 1-13) and two rabbit monoclonal antibodies (R12 and R11 (Yang J et al. (2011) PLoS ONE 6, e21018) specific for the cancer cell surface marker ROR1, we have decided to humanized variable regions of said antibodies by first aligning the parental antibody sequences by proprietary algorithms to a proprietary database containing next-generation sequencing data of human variable region sequences for human IgG antibodies. Based on the proprietary algorithm, best matches were identified and CDRs of the parental mAbs were in silico grafted into selected human VH and VL frameworks. This has resulted in to the in silico prediction of 64 humanized VH sequences for the VH of mAb 2A2 and 49 humanized VL sequences for the VL of mAb 2A2. In the case of Rl l, 101 humanized variants of VH and 82 humanized variants of VL have been generated in silico. Lastly, 25 humanized VH and 25 humanized VL sequences were predicted in the case of R12. As described herein libraries of transposable IgG heavy and light chain expression vectors have been cloned. The vector libraries have been stably transposed into Ll l host cells as described in herein in order to generate cellular libraries by transposition in which individual cells express a single combination of a humanized ROR1 IgG heavy and light chains. Based on binding of cells to ROR1 recombinant antigen by FACS (Fig. 9 (a)), individual binders were selected by single-cell FACS sorting and supernatants from individual clones were screened by SPR for high-affinity binders (Fig. 9 (b)). ROR1 affinities of sixty (2A2) and thirty (Rl l) cell clone supernatants were determined using single- concentration measurements. We subsequently chose to recover the sequences from those five clones for each humanization that showed the highest affinities, as described in herein. After sequence recovery and validation of individual HC/LC pairs by transient transfection into HEK293T cells and analysis of supernatants for antigen binding by ELISA, we chose to express and purify at a larger scale those three mAbs that showed the best binding (data not shown).

For the humanization of the R12 clone, we chose to screen Ll l cell clone supernatants for most promising clones by ELISA. To rate clones according to their strength of antigen binding, we determined their individual antigen-binding : IgG-titer ratios (Fig. 10). We subsequently chose to recover the sequences from those sixteen cell clones that showed the highest antigen-binding : IgG-titer ratios, i.e. showed highest binding normalized to IgG-titer. After sequence recovery and validation of individual HC/LC pairs by transient transfection of the respective antibody expression constructs into HEK293T cells and analysis of supematants for antigen-binding : IgG-titer ratios by ELISA, we proceeded to measure affinities of those eight mAbs that showed the highest antigen-binding : IgG-titer ratio in ELISA by surface plasmon resonance (SPR), using HEK293T supematants obtained from transient transfections and including supematants from HEK293T cells transiently transfected with expression constructs coding for parental mAb R12 in parallel as a reference. This analysis demonstrated that the affinities of the isolated humanized anti-RORl mAbs were remarkably superior to those of the parental mAb (Table 2). To confirm this observation, three humanized R12 mAbs were expressed at large scale in HEK293T cells, purified and determination of affinities by SPR was repeated with these purified mAbs (Fig. 13).

This analysis confirmed a substantial gain in monovalent affinity of humanized versus parental mAb R12 of more than 5 -fold.

Therefore, multi-cycle Surface Plasmon Resonance (SPR) for the top humanized anti-RORl humanized mAbs demonstrated that the affinities of the selected humanized anti-RORl mAbs were in the same range of the respective parental mAbs in the case of 2A2 and R11 (Figure 9 (c) and Table 1) or even significantly higher affine in the case of R12 (Table 2 and Figure 13 (a) and (b)).

This finding is extremely surprising, because, antibody humanization does regularly lead to a significant reduction or complete loss of binding affinity, which then has to be re-gained by subsequent affinity maturation.

Analysis of the degree of humanization

Analysis of the degree of humanization among these clones was also performed. To do so, we determined the similarity in percent of each chain's framework regions to those of the human germline sequence that was most closely related to the entire variable region sequence of the humanized mAbs (Fig. 11). Overall, the average degree of humanization of the library and isolated mAbs compared well to humanized antibodies that have been clinically approved (Fig. 11). Evaluation of in vitro cell killing activity of ADCs with humanized ROR1 mAbs

To evaluate the potency of the isolated humanized antibody clones as antibody-drug- conjugates (ADCs), antibodies were expressed and purified as LPETG-tagged versions to allow conjugation of anthracyclin-derived toxins (Figure 6) employing SMAC -technology™. 2A2-based anti-RORl antibodies conjugated to the cytostatic drug depicted in Figure 6 (PNU-EDA-Glys) potently affected cell viability of the ROR1 -positive 697 ALL cell line, while cold competition of the ADCs with unconjugated antibody significantly reduced cytotoxicity suggesting that cytotoxicity was indeed target-mediated (Fig. 8 (b)). ADCs based on humanized versions showed highly similar cell killing efficiencies as well as target- specificity. Likewise, ADCs based on parental and humanized R12 antibodies were able to specifically and potently kill RORl-positive human 697 ALL cancer cells (Fig. 12 (b), as well as murine cancer cells overexpressing human ROR1 (EMT6-ROR1) but not the parental cell line (EMT6) not expressing human ROR1 (Fig. 12 (c)).

ADCs and their in vivo efficacy

Mice treated with SMAC-conjugated hu2A2b-Ql 1 -based ADC showed prolonged survival relative to those treated with isotype control ADC (Fig. 8 (c)). In particular, the median survival time of mice treated with the hu2A2b-Ql 1-based ADC was of 28.3 days, whereas it was of 23.5 days in the isotype control group. No significant body weight loss difference was observed between the groups (data not shown).

EXAMPLES

The following section describes experimental details how the anti-RORl antibodies disclosed herein were created.

Generation of EBNA-based chimeric antibody expression constructs

Parental anti-RORl mouse mAb 2A2 (Baskar S et al. (2012) mAbs 4, 1-13), rabbit mAb R12 and rabbit mAb R11 (Yang J et al. (2011) PLoS ONE 6, e21018) were produced as chimeric full-length IgGl antibodies with human constant regions as follows. Variable region coding sequences including flanking restriction sites 5'NotI/3'-NheI (VH) and Notl/BsiWI (Vkappa) or Notl/Kasl (Vlambda) were produced by total gene synthesis (GenScript, Piscataway, USA) using MNFGLRLiFLVLTLKGVQC as leader sequence. Human IgH-gamma 1 and IgL-kappa (2A2 and R11) or -lambda (R12) constant region sequences for soluble antibody expression bearing appended c-terminal tags (IgH-chain-tag: LPETG-G-WSHPQFEK; IgL- chain-tag: GGGGS- LPETG-G-WSHPQFEK) were produced by total gene synthesis (GenScript, Piscataway, USA) including flanking restriction sites 5'NheI/3'BstBI (IgH-gamma 1) and 5'BsiwI/3'BstBI (IgL-kappa) or 5'KasI/3'BstBI (IgL-lambda). Antibody variable and constant region fragments were double-digested individually according to their flanking restriction sites and were assembled by ligation in the expression vector pCB14b (Figure 14) that was previously linearized by double-digestion with Notl/BstBI restriction enzymes. pCB14b is a derivative of the episomal mammalian expression vector pCEP4 (Invitrogen), carrying the EBV replication origin and encoding the EBV nuclear antigen (EBNA-1) to permit extrachromosomal replication, and contains a puromycin selection marker in place of the original hygromycin B resistance gene. Figure 15 shows plasmid maps of the resulting pCB14b-based expression constructs for heavy chain (2A2 and R12, Figure 15(a) and (c), respectively), kappa light chain (2A2, Figure 15(b)) and lambda light chain (R12, Figure 15(d)) expression. Rl l constructs (kappa light chain) are analogous to 2A2 constructs, however bearing R11 variable regions.

Expression and purification of antigens

Strep II-tagged ROR1 -extracellular domain was produced as follows: the nucleotide sequence encoding the expracellular domain of human ROR1 (NP 005003) was N-terminally fused to a signal sequence (MNFGLRLIFLVLTLKGVQC) and C-terminally fused with a sequence encoding a strepll-tag (GWSHPQFEK) (Fig. 16). The entire nucleotide sequences with flanking 5'NotI and 3'HindIII sites were produced by total gene synthesis (GenScript, Piscataway, USA), assembled in the proprietary mammalian expression vector pEvi5 linearized by double-digest with Notl/Hindlll restriction enzymes by Evitria (Schlieren, Switzerland) and verified by DNA sequencing. Expression of the proteins was performed in suspension-adapted CHO Kl cells by Evitria (Schlieren, Switzerland). Supernatants from pools of transfected CHO Kl cells were harvested by centrifugation and sterile filtered (0.2 μιη) before FPLC-based affinity purification using StrepTactin columns (IB A GmbH, Goettingen, Germany).

Construction of vectors for PiggyBac Transposon expression of IgGs All DNA syntheses to generate plasmid backbones and antibody sequences were performed by GenScript (Piscataway, USA). The amino-acid sequence of hyperactive PiggyBac transposase (hyPB) according to (Yusa K et al. (2011) Proc Natl Acad Sci USA 108, 1531- 1536) was codon-optimized for murine expression, synthesized with flanking restriction sites and cloned into the transient expression vector pcDNA3.1 (Fig. 17). Backbones for generating transposable antibody expression constructs (pPB) were assembled from modular parts with flanking restriction sites that were synthesized or derived from sequence-verified commercially available vectors, and are described in detail in Patent WO2014013026A1.

Antibody ORFs were assembled in transposable vector backbones as follows by two- or three-way cloning: antibody variable regions along with the leader sequence MNFGLRLiFLVLTLKGVQC were introduced using 5 'NotI/3 'Nhel (IgHV) and 5'NotI/3'BsiWI (IgkappaV) or 5'NotI/3'KasI (IglambdaV) restriction sites, in-frame with constant region fragments using 5'NheI/3 'BstBI (IgHC-gamma 1) or 5'BsiWI/3'BstBI (IgKC) or 5'KasI/3'BstBI (IgLC) restriction sites.

Construction of expression libraries for humanized ROR-1 IgG heavy and light chains

Humanized framework regions for rabbit anti ROR1 antibodies Rl l and R12 (Yang J et al. (2011) PLoS ONE 6, e21018) and murine anti-RORl antibody 2A2 (Baskar S et al. (2012) mAbs 4, 1-13) have been aligned with millions of human VH and VL sequences derived by next-generation sequencing (NGS) from healthy donors and by way of proprietary algorithms, most homologous human frameworks relative to the rodent parental mAbs have been selected. Selected frameworks have been assembled by proprietary algorithms in silico with the CDRs of the parental mAbs of rodent anti-RORl mAbs R11, R12, and 2A2.

The selected human variable regions along with the leader sequence MNFGLRLIFLVLTLKGVQC were synthesized by Gen9, Inc. (Cambridge, USA).

For humanization of 2A2, 64 VH x 49 VL variants were synthesized, for humanization of Rl l, 101 VH x 82 VL variants were synthesized and for humanization of R12, 25 VH x 25 VL were synthesized in total. Aliquots of synthesized VHS and VLS were pooled to equimolar amounts and amplified by PCR using forward primer univ-Notl-SP-F, comprising a Notl restriction enzyme recognition site for later directional cloning of the VH and VL sequence libraries.

(GAGGAGGCGGCCGCCATGAACTTTGGG ) and reverse primers huCGl-B, comprising a Nhel restriction enzyme recognition site for later directional cloning of the VH sequence library.

(AAGACCGATGGGCCCTTGGTG )

Reverse primer huCK-B

(GAAGACAGATGGTGCAGCCAC ) was used as for amplification of VK sequence libraries, comprising a BsiWI restriction enzyme recognition site for later directional cloning of the VK sequence library.

Reverse primer huCL-B

(GGAAACAGAGTCACGCTTGG ) was used as for amplification of νλ sequence libraries, comprising a Kasl restriction enzyme recognition site for later directional cloning of the νλ sequence library.

PCR amplified library fragments were column-purified, digested using Notl/Nhel (IgHV) or

Notl/BsiWI (IgkappaV) or Notl/KasI (IglambdaV), and cloned into transposable vectors described in above by 2-way (HC constructs) or 3-way cloning (LC constructs). Vector fragments for HC constructs were prepared by digestion of pPB-Hygro-HCgl-gen (Fig. 18

(a)) with Notl/Nhel. Vector fragments for LC constructs were prepared by digestion of pPB-

Puro-LC (Fig. 18 (b) with Notl/BstBI restriction enzymes. Constant region sequences including flanking restriction sites 5'BsiWI/3'-BstBI (Ckappa) and 5'KasI/3'-BstBI

(Clambda) were produced by total gene synthesis and cloned into pUC57 (GenScript,

Piscataway, USA). Constant region fragments for assembly of libraries were generated by digestion of DNA maxi-preps of the respective plasmids containing kappa or lambda constant regions with BsiWI/BstBI (to generate kappa constant region fragments) or KasI/BstBI (to generate lambda constant region fragments). Library ligations were transformed into Neb5- alpha electrocompetent cells (Neb, Ipswich, USA), pre-cultured for 1 hour, amplified in selective LB-media containing O.lmg/ml ampicillin overnight, and plasmid DNA was isolated using NucleoBond Xtra Maxi Plus kit (Macherey&Nagel, Dueren, Germany). Library sizes were determined by plating out serial dilutions of the pre-culture onto selective agar plates (titration plates) and obtained clone numbers were calculated to estimate library complexities. At least 12 clones from titration plates were analyzed by restriction digest and sequencing of variable regions using primer pPBseql3 (GGCCAGCTTGGCACTTGATG).

Expression, Display and Identification of best humanized ROR1 mAbs

For the Display of humanized antibody libraries Ll l murine preB cells were used, which represent an in-house generated subclone of the Abelson murine leukemia virus (A-MuLV) transformed preB cell line 63-12 isolated from RAG-2 deficient mice (Shinkai Y et al. (1992) Cell, 68(5), 855-867). The cells were cultured in SF-IMDM media supplemented with 2% fetal calf serum, 2mM L-Glutamine, 100 IU Penicillin, O.lmg/ml Streptomycin (Amimed, BioConcept Ltd., Allschwil, Switzerland) and 50μΜ b-mercaptoethanol (Amresco, Solon, USA) in screwcap bottles (Sarstedt, Numbrecht, Germany) at 37° under 7.5% C0 ₂ .

EMT6 cells, a kind gift of Prof. A. Zippelius (University of Basel) and 293T cells were both grown in DMEM supplemented with 10% FCS, 2mM L-Glutamine, 100 IU Penicillin, O. lmg/ml Streptomycin and 0.25μg/ml Fungizone (Amimed) at 37° under 5% CO2. 697 cells were grown in RPMI supplemented with 10% FCS, 2mM L-Glutamine, 100 IU Penicillin, O. lmg/ml Streptomycin and 0.25μg/ml Fungizone (Amimed) at 37° under 7.5% CO2.

Transposition and selection of Lll cells expressing humanized anti-RORl IgG libraries:

One day before electroporation, Ll l cells were seeded at a density of 0.2E+6 cells/ml to obtain log-phase growing cells the next day. The entire procedure of electroporation was performed at room temperature. Cells were harvested by centrifugation at 1200rpm for 6min and resuspended in plain RPMI medium to a concentration of 8E+7 cells/ml. Per cuvette, 25 μg of total DNA was diluted in 400μl RPMI (using HC/LC/tranposase weight ratios as shown in Figure S2B) and 400μl cell suspension was combined with diluted DNA and transferred to a 0.4cm gap gene pulser cuvette (BioRad, Hercules, USA). Electroporation was done with a BioRad GenePulser II equipped with capacitance extender set to 300V and 950μΡ. After incubation for 5-10 min in cuvettes in order to allow pores to close, cells were washed once in complete SF-IMDM growth medium, resuspended and seeded into T175 tissue culture flasks at a total volume of 64ml of complete growth medium. For selection, ^g/ml Puromycin and 800μg/ml Hygromycin (0240.4 and CP12.2, respectively; Carl Roth, Karlsruhe, Germany) were added simultaneously, and selection was allowed to proceed for 4- 5 days without exchange of medium or subculturing, until selection was complete.

FACS Staining and FACS sorting of cellular libraries

Cells were stained on ice in FACS-buffer (PBS supplemented with 2% FCS) at a concentration of 1E+7 cells/ml. Wash steps were performed by pelleting cells by centrifugation at 1300rpm for 3min, resuspending in FACS-buffer using a 5x volume of staining reactions, pelleting again and resuspending in FACS buffer using lx volume of staining reaction.

For analysis of surface-antibody expression, cells were stained using 1 :200 diluted Ig- kappaLC-APC labeled antibody (MH10515, Life Technologies, Carlsbad, USA) for 30 minutes. Cells were washed once and analysed by flow cytometry on a FACSCalibur instrument (Becton-Dickinson, Franklin Lakes, USA).

For staining of cellular libraries expressing IgG libraries, previously determined limiting concentrations of antigen (0.25 μg/ml ROR1 -strep) and 1 :250 diluted anti-Human-IgG (Fc gamma-specific) PE (ebioscience 12-4998-82) were added and cells were incubated on ice in the dark for 30 minutes. After washing cells once, 1 :500 diluted Strep-mAb classic Oyster 645 (2-1555-050 iba, Goettingen, Germany) was used to detect strep-tagged antigen bound to cells for 30 minutes. After a final wash, cells were filtered using cell strainer cap FACS tubes (BD Falcon). Sorting was performed on a FACSAriall instrument (Becton-Dickinson, Franklin Lakes, USA).

Screening of anti-RORl IgG expressing Lll clones by ELISA

Antigen-binding analysis by ELISA was performed by coating of Nunc-Immuno MaxiSorp 96-well plates (Thermo Scientific, Waltham, USA) with antigen diluted in coating buffer (100 mM bicarbonate/carbonate buffer) over night at 4°C. For sandwich ELISA, plates were coated with 2μg/ml AffmiPure F(ab')2 fragment donkey anti-human IgG (Jackson Immunoresearch, West Grove, USA) diluted in coating buffer over night at 4°C. After coating, plates were washed twice with PBS/0.05%Tween-20 (PBS-T), blocked with PBS-T containing 3% bovine serum albumin (BSA) (Carl Roth, Karlsruhe, Germany) for 1 hour at 37°C and washed again 5 times with PBS/T. Ll l clone supematants were pre-diluted 3-fold, while supematants from 293T cells were pre-diluted 50-fold. Parental mAbs diluted to 0^g/ml were used to generate standard curves. 3.5-fold serial dilutions of samples were added and plates were incubated for 1 hour at 37°C. After 5 washes with PBS/T, HRPO- conjugated F(ab)2 anti-human FC-gamma (Jackson Immunoresearch, West Grove, USA) was added at 10'000-fold dilution in PBS/T containing 1% BSA, and plates were incubated for 1 hour at 37°C. Finally, plates were washed 5 times with PBS/T and 50μl of Sigmafast OPD Peroxidase substrate (Sigma-Aldrich, St. Louis, USA) were added. Reactions were stopped by adding 50μl of 2M H2SO4. Absorption was measured at 490nm. OD50 values of standards with known concentrations and samples determined by 4-point curve fitting models were used to calculate EC50.

Analysis of affinity by surface plasmon resonance (SPR)

Affinities were determined using a Biacore T200 instrument (GE Healthcare, Buckinghamshire, UK) and data was evaluated using Biacore Evaluation T200 V2.0 software. To capture mAbs, goat a-human Fc-gamma-specific IgG (Jackson ImmunoResearch, # 109- 005-098) was covalently immobilized on a CM5 chip (GE Healthcare, # BR-1005-30).

For measurement of Ll l clone supematants, antibodies were diluted to O^g/ml with complete SF-IMDM cell culture medium and captured for 120s with a flow of ΙΟμΙ/min. ROR1 -strep was diluted to 40nM in running buffer. Association was measured at a flow rate of 30μl/min for 120s, and dissociation was followed for 200s. Curves were fitted using 30s dissociation due to upper plateau formation at later timepoints. Capture levels ranged from 29.1 RU to 57.7 RU.

For determination of anti-RORl affinities of selected 2A2- and Rl l -based humanized clones and comparison to parental mAbs, purified mAbs were diluted to 0^g/ml with running buffer and captured for 120s with a flow of ΙΟμΙ/min. RORl-strep was diluted in running buffer using 2-fold serial dilutions ranging from 20nM to 2.5nM. Association was measured at a flow of 30μl/min for 120s, and dissociation was followed for 200s. Curves were fitted using 30s dissociation due to upper plateau formation at later timepoints. Capture levels ranged from 29.1 RU to 57.7 RU.

For initial determination of huR12 clone affinities and parental R12 using transient HEK293T supernantants, IgG expression vectors were transiently transfected into 293T cells and IgG titers in supernatants were quantified by ELISA. For SPR, supernatants were diluted to 1 μg/ml IgG with running buffer (HBS-EP+ pH 7.4 (10 mM HEPES, 150 mM NaCl, 3 mM EDTA, 0.05 % Tween 20)) and captured for 200s with a flow of 30μl/min. ROR1 -strep was diluted in running buffer using 2-fold serial dilutions ranging from 20nM to 2.5nM. Association was measured at a flow of 30μl/min for 120s, and dissociation was followed for 1000s. Capture levels ranged from 26.9 RU to 54.1 RU.

For determination of huR12 clone affinities and parental R12 using purified IgG, mAbs were diluted to 2μg/ml IgG with running buffer (HBS-EP+ pH 7.4 (10 mM HEPES, 150 mM NaCl, 3 mM EDTA, 0.05 % Tween 20)) and captured for 200s with a flow of 30μl/min. ROR1 -strep was diluted in running buffer using 2-fold serial dilutions ranging from 1.25nM to 20nM. Association was measured at a flow of 30μl/min for 120s, and dissociation was followed for 1000s. Capture levels ranged from 26.9 RU to 54.1 RU.

All measurements were performed at 25°C. All curves were fitted using a 1 : 1 binding model with RI=0. Regeneration was done for 90s using lOOmM H3PO4 at a flow of 30 μΐ/min.

Recovery and identification of VH and VL sequences from Lll clones displaying strong ROR1 binding specificity

RNA was extracted from ~2E+6 cells grown in 24-well plates using Tri-Reagent (Sigma- Aldrich, St. Louis, USA) and reverse transcribed with ProtoScriptll Reverse transcriptase (Neb, Ipswich, USA) using random nonamers, according to manufacturer's instructions. Variable regions were amplified by PCR using Q5 DNA polymerase (Neb, Ipswich, USA) by means of forward primer EFlaNotI_F (CCATTTCAGGTGTCGTGAGC) and reverse primers CG- revseq-1 (GTTCGGGGAAGTAGTCCTTG) for VH and Intron-rev-1

(GTGGATGTCAGTAAGACCAATAGGTGCC) for VL. PCR products were purified by column purification (Macherey&Nagel, Dueren, Germany) according to the manufacturers instructions, digested with restriction enzymes and purified by agarose-gel purification. Recovered variable regions were cloned into pCB14b, along with human Ig-gammal or Ig- kappa constant regions by 2- or 3-way cloning, respectively. Variable regions from several bacterial clones were sequenced by Sanger sequencing (performed at Microsynth AG, Balgach, Switzerland) using sequencing primer CMVseq2 (GCAGTGTAGTCTGAGCAGTAC) and were aligned to library sequences using Geneious Software (Biomatters, New Zealand).

Expression of recombinant IgG antibodies in 293T cells

Expression of recombinant antibodies was performed by transfecting pCB14b-based expression constructs into HEK-293T cells and harvesting of cell supernatants.

For transient antibody expression, cells were transfected in 6-well-plates using Lipofectamine LTX plus (LifeTechnologies, Carlsbad, USA). Per well, 2.5μg of total DNA was transfected, and fresh growth medium was added the next day and conditioned for 4 days. Supernatants were sterile-filtered and stored at -20°C until analysis.

For large-scale, semi-stable antibody expression, cells were transfected in 10cm dishes using Lipofectamine LTX plus, enriched by selection with 2μg/ml Puromycin (0240.4, Carl Roth, Karlsruhe, Germany), expanded to 14cm dishes coated with poly-L lysine and maintained in DMEM/F12 serum-free medium (Gibco) containing 16^g/ml N-Acetyl-L-Cysteine, lOmg/ml L-Glutathione and ^g/ml Puromycin. Supernatants containing antibodies were harvested twice a week, sterile-filtered and stored at 4°C until purification. Purification by FPLC was performed using an Akta purifier instrument (GE Lifesciences). After passing supernatants over Amsphere protein A columns (JWT203CE, JSR Micro, Sunnyvale, USA), followed by washing with PBS, antibodies were eluted with 0.1M Glycine pH 2.5 and immediately neutralized with 1M Tris pH 8.0. Buffer exchange with PBS was performed using Amicon Ultra-4 Centrifugal Filters (Merck Millipore).

Generation of EMT-6 cell breast cancer cells overexpressing human ROR1 target antigen

Murine EMT6 cells overexpressing human ROR1 antigen were generated by transposition. An expression construct for expression of full-length human ROR1 (NP 005003.2) was created using a PiggyBac-transposable expression vector containing a puromycin slection marker (pPB-PGK-Puro-RORl) (Figure 19). This vector was derived by replacing IRES- driven expression of antibiotic selection markers with selectable marker expression driven by a separate phosphoglycerate kinase promoter (PGK). Electroporation of EMT6 cells with the pPB-PGK-Puro-RORl expression vector and a PiggyBac expression vector pcDNA3.1-mPB (Figure 17) was performed the same way as for LI 1 cells, except that EMT6 cell pellets were resuspended to 5E+6 cells/ml in RPMI, and 20μg of total DNA was used (pPB-PGK-Puro- RORl :transposase =3:2). After electroporation, cells were seeded on a 6-well plate and selection was performed using 3μg/ml Puromycin for 6 days. Single cell clones were isolated by staining transposed and selected cells as follows: Cells were stained with 2μg/ml anti- ROR1 antibody 2A2 (mouse-human IgG chimeric) diluted in FACS-buffer (PBS/2%FCS) for 30 min on ice. After washing with FACS-buffer, the primary antibody was detected using PE- coupled Fc-specific anti-human-IgG diluted 1 :200 in FACS buffer for 30 min on ice. After a final wash, single cells were sorted into 96-well plates using a FACSAriall instrument with FACSDiva software (Becton-Dickinson, Franklin Lakes, USA). After expansion, clones were verified to homogenously express antigen by staining as above and flow cytometry analysis.

Sortase-mediated antibody conjugation

Sortase-mediated antibody conjugation was performed essentially as described (Beerli et al, PloS One, 10(7), e0131177, 2015). Briefly, sortase-tagged mAbs [10μΜ] were incubated with oligo-glycine modified toxin [200μΜ] in the presence of 3μΜ Sortase A in 50mM HEPES, 150mM NaCl, ImM CaCl ₂ , 10% Glycerol, pH 7.5 for 3.5h at 25°C. The reaction was stopped by passing it through an rProtein A GraviTrap column (GE Healthcare) equilibrated with 25mM HEPES, 150mM NaCl, pH 7.5, followed by washing with 20 column volumes (CVs) of wash buffer (25 mM HEPES pH 7.5, 150 mM NaCl, 10% v/v Glycerol). Bound conjugate was eluted with 5 CVs of elution buffer (0.1M Glycine, 50mM NaCl, pH 2.5), with 0.5 CV fractions collected into tubes containing 25% v/v 1M HEPES pH 8 to neutralize the acid. Protein-containing fractions were pooled and formulated using ZebaSpin desalting columns (GE Healthcare, Thermo) according to the manufacturer's instructions. ADCs based on chimeric and humanized versions of mAbs 2A2 and Rl l were formulated in PBS without Ca ²⁺ and Mg ²⁺ (Amimed-Bioconcept). ADCs based on chimeric mAb R12 were formulated in 20mM HEPES, pH6.8, 175 mM Sucrose, 0.02% Tween20, while ADCs based on humanized R12 mAbs were formulated in 15 mM Histidine, pH 6.0, 175 mM Sucrose, 0.02% Tween20.

In vitro cell killing assays

Cytotoxicity of ADCs was investigated in cell killing assays essentially as described (Beerli et al, 2015). Briefly, cells were plated on 96 well plates in 75μΙ, growth medium and grown at 37°C in a humidified incubator in a 7.5% C0 ₂ atmosphere. Cells were seeded at Γ000 cells/well (EMT6 and derived clones thereof) or 10Ό00 cells/well (for testing 2A2 and hu2A2 ADCs on 697 cells) or 75*000 cells (for testing huR12-4 ADC on 697 cells) into 96-well plates. The following day, 25μL, of 3.5-fold serial dilutions of each ADC in growth medium was added, resulting in final ADC concentrations from 20μg/mL to 0.02ng/ml. For competition, ADC serial dilutions were done in growth medium containing unconjugated mAb ("comp") at a concentration of 200μg/mL. After 4 additional days, plates were removed from the incubator and equilibrated at room temperature. After approximately 30 minutes, 50μL medium was carefully removed from each well and replaced with 50μL CellTiter-Glo® Luminescent Solution (Promega, Cat.No G7570). After shaking the plates at 450rpm for 5 min followed by lOmin incubation in the dark without shaking, luminescence was measured on a Tecan Spark 10M with an integration time of 250 milliseconds per well.

Disseminated xenograft model of human B cell leukemia cells (697) in NOD-SCID mice

The efficacy of an antibody drug conjugate (ADC) based on hu2A2b-Ql 1 , SMAC-conjugated with the glycine modified toxin of Fig. 6, was investigated and compared to a group treated with isotype control-ADC (trastuzumab, targeting HER2) in a disseminated xenograft model of female NOD-SCID mice carrying tumors derived from human 697 leukemia cancer cells. Twelve 9-week old mice, each weighing at least 20 g, were inoculated with 697 tumor cells (5xl0 ⁶ cell/animal, in 200 μL, PBS (Dulbecco ^' s Phosphate Buffered Saline)) on study day 0. Test ADC, formulated in PBS, was administered (i.v.) in a group of 6 mice at 1.0 mg/kg on days 7, 14 and 21 , with administration of mouse IgG (Privigen IVIG (CSL Behring, Lot. 4335500037), 30 mg/kg) 20 hours before each ADC administration. One isotype control ADC (HER2 -targeting trastuzumab, SMAC-conjugated with the glycine modified toxin of Fig. 6), was administered under the same conditions and protocol in a group of 6 mice. Any animal showing clinical signs of moderate pain, moderate distress or any degree of suffering was humanely euthanized. Furthermore; any animal was humanely euthanized if it exhibited clinical signs that exceed the limits of the study specific humane endpoints according to the European and Danish legislation on animals in experimental studies. Body weight of each animal was measured three times weekly throughout the study. Data analysis was performed using the software PRISM.

FIGURES

Fig. 1: Chemical structures of the Gly ₅ modified maytansine toxins used for SMAC- technology™ immunoligand conjugation.

In the upper part of Fig. 1 displays the maytansinoid is DM1 ([N ² ' -deacetyl-N ² '-(3-mercapto- l-oxopropyl)-maytansine], containing the so-called SMCC linker to which the oligo-glycine peptide (Gly _n ), in order to allow conjugation by SMAC -technology™, This SMCC linker is only an optional component for the SMAC -technology™ conjugated immunoligand toxin conjugates, and of no importance for the conjugation of the payload.

Instead of DM1, other optional linker structures, like the SPDB linker of the maytansinoid payload DM4 ([N20-deacetyl-N20-(4-mercapto-4-methyl-l-oxo- pentyl)-maytansine] may optionally be included, see Fig. 2.

In the lower part of Fig. 1 , the maytansinoid is maytansine itself, which in the unconjugated form has the structure of Fig. 2 (a), may be used to generate the oligo-glycine peptide (Gly _n ) derivative depicted here, which has formed the basis for the immunoligand maytansine conjugates analyzed herein.

Fig. 2: Three Maytansinoids that can be used in the context of the present invention. Fig. 5 (a): Maytansine, Fig. 5 (b): DM1 ([N ² ' -deacetyl-N ² '-(3-mercapto-l-oxopropyl)-maytansine], Fig. 5 (c): DM4 ([N20-deacetyl-N20-(4-mercapto-4-methyl-l-oxo- pentyl)-maytansine].

Fig. 3 (a): An anthracycline (PNU) derivative that can be used with the Immunoligand-Toxin- conjugate according to the invention. The derivative may comprise at its wavy line a chemical structure comprising an oligo-glycine peptide (Gly _n ) coupled to said anthracyline derivative in such a way that the oligo-glycine (Gly _n ) peptide has a free amino terminus, and wherein n is an integer between > 1 and < 21. The derivative is derived from anthracycline PNU- 159682 having the formula (v) as depicted in Fig. 5.

Fig. 3 (b): An oligo-glycine peptide (Gly _n ) is coupled to the anthracyline derivative as seen in Fig. 3 (a) by means of an ethylenediamino linker (EDA), which ethylenediamino linker is coupled to the anthracycline derivative by means of a first amide bond, while it is conjugated to the carboxyterminus of the oligo-glycine peptide by means of a second amide bond, said ethylenediamino linker and oligo-glycine peptide.

Fig. 4 (a): Another anthracycline (PNU) derivative that can be used with the Immunoligand- Toxin-conjugate according to the invention. The derivative may comprise at its wavy line a chemical structure comprising an oligo-glycine peptide (Gly _n ) coupled to said anthracyline derivative in such a way that the oligo-glycine (Gly _n ) peptide has a free amino terminus, and wherein n is an integer between > 1 and < 21. The derivative is derived from anthracycline PNU-159682 having the formula (v) as depicted in Fig. 5.

Fig. 4 (b): An oligo-glycine peptide (Gly _n ) is coupled to the anthracyline derivative as seen in Fig. 4 (a) by means of an ethylenediamino linker (EDA), which ethylenediamino linker is coupled to the anthracycline derivative by means of a first amide bond, while it is conjugated to the carboxyterminus of the oligo-glycine peptide by means of a second amide bond, said ethylenediamino linker and oligo-glycine peptide.

Fig. 5: Structure of PNU-159682 as described in the prior art (e.g. WO2009099741, or Quintieri L et al (2005) Clin Cancer Res. 11, 1608-17.

Fig. 6: Structure of PNU-EDA-Glys as utilized for the SMAC -technology conjugation to C- terminally LPETG sortase tagged monoclonal antibodies using sortase enzyme as disclosed in the examples herein.

Fig. 7: Schematic drawing of site-specific sortase mediated antibody conjugation (SMAC- technology). The monoclonal antibodies need to be produced with C-terminal LPXTG sortase tags. The toxic payload needs to be produced to contain an oligoglycine peptide stretch (Gly _n - stretch) with a certain number of glycine (n>l and < 21, preferably n>3 and < 10, most preferably n=5) residues in a row. Sortase A enzyme from Staph, aureus specifically recognizes the LPXTG pentapeptide motif and catalyzes the transpeptidation of the oligo- glycine peptide stretch to the threonine-glycine peptide bond of LPXTG, thereby generating a new stabile peptide bond between the threonine and the N-terminal glycine of the oligo- glycine stretch.

Fig. 8: To evaluate performance of humanized 2A2 antibodies and to compare with potency of the parental, chimeric mAb in the form of antibody-drug conjugates (ADCs). To generate ADCs antibodies were expressed heavy and light chains C-terminally tagged with sequences comprising a sortase A recognition motif, and were coupled to the Gly5 -modified anthracyclin PNU159682-derivative using Sortase-mediated conjugation (WO2014140317) and were then tested in in vitro killing assays using ROR1 -expressing ALL cancer cell lines. 697 ALL expressed ROR1 as revealed by staining with 2A2 anti-RORl antibody and analysis by flow cytometry (Figure 8 (a)). When cell killing was performed, both chimeric 2A2-PNU and humanized versions efficiently killed 697 cells, and this effect was target-specific because it could be competed with unconjugated antibody ("cold competition"). After four days of treatment, viable cells were quantified using a luminescent cell viability assay. Each data point represents the mean of duplicates and error bars represent SD (Figure 8 (b)). Efficacy of humanized 2A2 ADC was evaluated in vivo in a disseminated mouse model established with human 697 ALL cell line. Kaplan-Meyer survival curves are shown and compared with a group treated with an isotype control (Figure 8 (c)). Mice treated with the humanized 2A2 ADC display a prolonged survival relative to the control.

Fig 9: Humanization of mouse 2A2 and rabbit R11 anti-RORl antibodies by Transpo-mAb. Transposable humanized libraries consisted of 64 VH x 49 VL (mouse) and 101 VH x 82 VL (rabbit) sequences and were transposed using a DNA weight ratio of 4:2:1 (HC:LC:transposase). The cellular library was enriched for antibody expression by antibiotic selection and labelled for antigen binding using soluble antigen (ROR1 -strep) at previously determined limiting concentrations. Antigen binding was detected using a fluorophore labeled anti-strep-tag antibody (StrepMAB-645). To normalize for antibody surface expression levels, libraries were co-stained with anti-IgG-PE. Antigen-binding single-cells were then isolated by FACS. Figure 9 (a) shows flow cytometry plots of humanized cellular libraries analyzed for antigen-binding (x-axis) and antibody surface expression (y-axis) as analyzed immediately prior to sorting. After single-cell sorting, clone supernatants were expanded, thereby generating antibody-containing conditioned supernatants which were screened by SPR. Figure 9 (b) shows an isoaffinity plot displaying k _a and kd determined by SPR with single-cell sorted clone supernatants (unicates) or purified parental mAbs (triplicates). Antibody sequences of 5 clones showing best KDs were recovered by RT-PCR and deconvoluted/validated by ELISA. Sequences not matching library design were not pursued further. The top 3 clones from this analysis were expressed in 293T cells, purified and affinities were measured by SPR. Figure 9 (c) shows an isoaffinity plot of these purified top 3 mAbs of each screen.

Fig. 10: Humanization of rabbit R12 antibody by Transpo-mAb. Transposable humanized libraries consisted of 25 VH x 25 VL sequences and were transposed into Ll l cells using a DNA weight ratio of 0.25:0.125: 1 (HC:LC:transposase). Cells transposed with the humanized R12 cDNA library were enriched for antibody surface expression by antibiotic selection and then single-cell sorted for antigen binding. The resulting clones were expanded and culture supernatants were analysed by ELISA. Shown is a scatter-plot of single-cell clone supernatants analyzed in parallel for antigen binding and IgG titer by ELISA. Serial dilutions of clonal supernatants grown in 96-well plates were directly used for assessing binding to ELISA plates that were coated with limiting concentrations of RORl to minimize avidity effects. IgG levels were determined by sandwich ELISA. EC50 values were calculated using standard curves obtained with known concentrations of parental mAb R12. Those sixteen clones with highest RORl -binding :IgG-titer ratios were chosen for antibody variable region sequence recovery.

Fig. 11: Comparison of generated humanized antibodies alongside clinically approved humanized mAbs with human germline genes. Variable regions of indicated antibodies were subjected to Ig-Blast database search (http://www.ncbi.nlm.nih.gov/igblast/) for the closest human germline sequence each, and sequence identity within framework regions 1 , 2 and 3 were determined. Average identity with human germline over all three frameworks was considered as a measure of humanization grade and is shown in percent. For libraries, mean values over all sequences within the library are shown. FDA approval status and sequences of reference humanized antibodies and were retrieved from http://imgt.org/.

Fig. 12: Performance of humanized R12 antibodies and comparison with potency of the parental, chimeric mAb in the form of antibody-drug conjugates (ADCs). To generate ADCs, antibodies were expressed with heavy- and light-chains comprising a c-terminal sortase A recognition motif, and were coupled to a Gly5-modified anthracycline (as per Figure 6) using Sortase-mediated conjugation (WO2014140317). The resulting ADCs were tested in in vitro killing assays using human 697 ALL cells expressing endogenous levels of RORl (Figure 8 (a)), and using clones of murine EMT6 cells overexpressing human RORl (EMT6-ROR1 clone 6, EMT6-ROR1 clone 14), or parental EMT6 cells {EMT WT). To verify the absence or presence of antigen surface expression on EMT6 cell variants, cells were stained with 2μg/ml 2A2 anti-RORl antibody and were subsequently analysed by flow cytometry (Figure 12 (a)). To assess the potency of the ADCs generated, human ALL cancer cells (697), as well as murine EMT6 cells overexpressing human RORl (EMT6-ROR1 clone 6, EMT6-ROR1 clone 14), or parental EMT6 cells {EMT WT) were exposed to ADCs for 4 days and cell viability was quantified using a luminescence-based cell viability assay (Figure 12 (b) and (c), respectively). In all cases, cell viability of RORl -expressing cells was reduced (in positive correlation with RORl expression) when treated with humanized R12-based ADCs, whereas treatment with control ADCs solely resulted in non-specific killing at highest doses.

Fig. 13: Comparison of affinities of humanized and parental R12 mAbs determined by SPR using purified antibody. The three humanized R12 mAbs indicated were expressed at large scale in HEK293T cells, purified by Protein A affinity chromatography and subjected to SPR analysis for determination of affinity towards soluble RORl -strep. Antibodies were captured using a goat a-human Fc-gamma-specific IgG. RORl -strep was diluted in running buffer using 2-fold serial dilutions ranging from 1.25nM to 20nM. Association was measured 120s, and dissociation was followed for 1000s. Capture levels ranged from 26.9 RU to 54.1 RU. This analysis confirms a substantial gain in affinity of humanized versus parental mAb R12 of more than 5 -fold.

Fig. 14: EBNA-based expression vector pCB14b for transient or semi-stable antibody expression in HEK293T cells. Shown are the principal components of the vector and the restriction sites used for the described clonings.

Fig. 15: Transposable antibody expression constructs. Shown are the principal components of the vector and the restriction sites used for the described clonings, (a) Transposable 2A2 heavy chain expression vector, (b) Transposable 2A2 light chain expression vector, (c) Transposable R12 heavy chain expression vector, (d) Transposable R12 light chain expression vector. Fig. 16: Sequence of Strep-tagged human ROR1. Shown is the amino acid sequence of the extracellular domain of human ROR1, N-terminally tagged with a strep-tag (WSHPQFEK).

Fig. 17: Construct for transient expression of hyperactive PiggyBac transposase (hyPB).

Fig. 18: Vectors used as backbones to generate transposable heavy chain libraries. Shown are the principal components of the vector and the restriction sites used for the described clonings, (a) Vector used for insertion of humanized heavy chain variable region fragments. Heavy chain variable region fragmens were introduced using Notl/Nhel sites, (b) Vector used for insertion of humanized light chain variable region fragments along with constant region fragments of kappa or lambda type. Kappa light chain variable region fragments digested with Notl/BsiWI were introduced into this vector linearized with Notl/BstBI, along with kappa light chain constant region fragments digested with BsiWI/BstBI. Lambda light chain variable region fragments digested with Notl/KasI were introduced into this vector linearized with Notl/BstBI, along with lambda light chain constant region fragments digested with KasI/BstBI.

Fig. 19: Transposable vector for expression of full-length ROR1. Shown are the principal components of the vector and the restriction sites used for the described clonings.

Incorporation by reference

This specification comprises several references to the sortase mediated antibody conjugation (SMAC). This technology is fully disclosed in WO2014140317, content of which is incorporated by reference herein for enabling purposes.