RS52492B - MOLECULES WHICH INTERSPECIALLY SPECIFICALLY BISPECIFICALLY Bind - Google Patents

Abstract

A polypeptide comprising a first binding domain, which is an antibody capable of binding to the epitope of the human CD3ε epitope and Callithrix jacchus, Saguinus oedipus or Saimiri sciureus, comprising that epitope part of the amino acid sequence contained in the group consisting of SEQ ID no: 2, 4 , 6, or 8 and containing at least the amino acid sequence Gln-Asp-Gly-Asn-Glu, and another binding domain capable of binding to human and / or non-chimpanzee EGFR, Her2 / neu or IgE primates. another 17 patent claims.

Description

MOLECULES THAT

INTERSPECIES SPECIFIC

BIND BISPECIFICALLY

[0001] The present invention relates to polypeptides containing the first human binding domain that represents an antibody capable of binding to the epitope CD3 (epsilon) of humans and primates other than chimpanzees and the second binding domain capable of binding to EGFR, Her2/neu or IgE of humans and/or primates other than chimpanzees, as well as to the processes for obtaining said polypeptides. The invention further relates to nucleic acids encoding polypeptides, to vectors containing them, and to host cells containing the vector. In another aspect, the invention provides pharmaceutical compositions containing said polypeptides and medicinal uses of the polypeptides.

[0002] T cell recognition is mediated by clonotically distributed alpha beta and gamma delta T cell receptors (TcR) that react with the peptide portion of the peptide MHC molecule (pMHC).

(Davis & Bjorkman, Nature 334 (1988), 395 - 402). Antigen-specific TcR chains do not have signaling domains, but are coupled to the conserved CD3 multisubunit signaling apparatus (Call, Cell 111 (2002), 967 - 979, Alarcon, Immunol. Rev. 191 (2003), 38 - 46, Malissen Immunol. Rev. 191 (2003), 7 - 27). The mechanism by which TcR binding directly communicates with the signaling apparatus remains a fundamental question in T cell biology (Alarcon, loc. cit.; Davis, Cell 110 (2002), 285 - 287). Persistent T cell responses are known to involve coreceptor engagement, TcR oligomerization, and higher order arrangement of TcR-pMHC complexes at the immune synapse (Daviš & van der Mervwe, Curr. Biol. 11 (2001), R289 - R291, Davis, Nat. Immunol. 4

(2003), 217 - 224). However, very early TcR signaling occurs in the absence of these events and can induce a ligand-induced conformational fold in CD3 epsilon (Alarcon, loc. cit., Davis (2002), loc. cit. Gil, J. Biol. Chem. 276 (2001), 11174 - 11179, Gil, Cell 109 (2002), 901-912). The epsilon, gamma, delta, and zeta subunits of the signaling complex associate with each other to form the CD3 epsilon-gamma heterodimer, the CD3 epsilon-delta heterodimer, and the CD3 zeta-zeta homodimer (Call, loc. cit.). Various studies have shown that CD3 molecules are important for proper cell surface expression of the alpha-beta TcR and normal T cell development.

(Berkhout, J. Biol. Chem. 263 (1988), 8528 - 8536, Wang, J. Exp. Med. 188 (1998), 1375 - 1380, Kappes, Curr. Opin. Immunol. 7 (1995), 441 -447). Solved structure of ectodomains

of murine CD3 epsilon gamma heterodimer fragments showed that epsilon gamma subunits are C2-set Ig domains that interact with each other to form an unusual side-to-side dimer configuration (Sun, Cell 105 (2001), 913-923). Although the cysteine-rich branch appears to play an important role in driving CD3 dimerization (Su, loc. cit., Borroto, J. Biol. Chem. 273 (1998), 12807 -12816), interaction by the extracellular domains of CD3 epsilon and CD3 gamma is sufficient to bind these proteins to TcR beta (Manolios, Eur. J. Immunol. 24 (1994), 84-92, Manolios & Li, Immunol. Cell Biol. 73 (1995), 532-536). still controversial, the dominant TcR stoichiometry most likely contains one alpha beta TcR, one CD3 epsilon gamma heterodimer, one CD3 epsilon delta heterodimer and one CD3 zeta homodimer (Call, loc. cit.). Given the central role of the human CD3 epsilon gamma heterodimer in the immune response, the crystal structure of this complex bound to the therapeutic antibody OKT3 was recently elucidated (Kjer-Nielsen, PNAS 101, (2004), 7675 - 7680).

[0003] A number of therapeutic strategies modulate T cell immunity by targeting TcR signaling, particularly anti-human CD3 monoclonal antibodies (mAbs) that are widely used in immunosuppressive regimens. The CD3-specific murine mAb OKT3 was the first mAb licensed for human use (Sgro, Toxicology 105 (1995), 23-29) and is widely used clinically as an immunosuppressive agent in transplantation (Chatenoud, Clin. Transplant 7 (1993), 422-430, Chatenoud, Nat. Rev. Immunol. 3 (2003), 123-132; Kumar, Transplant. 30 (1998), 1351 - 1352), type 1 diabetes (Chatenoud (2003), et al.), and psoriasis (Utset, J. Rheumatol. 29 (2002), 1907 -1913). Furthermore, anti-CD3 mAbs can induce partial T cell signaling and clonal anergy (Smith, J. Exp. Med. 185 (1997), 1413-1422). OKT3 has been described in the literature as a potent T cell mitogen (Van Wauve, J. Immunol. 124 (1980), 2708 -18 ) as well as a potent T cell killer (Wong, Transplantation 50 (1990), 683 - 9 ). OKT3 exerts both of these activities in a time-dependent manner, following early T cell activation leading to cytokine release; after further administration OKT3 later blocks all known functions of T cells. Due to this subsequent blocking of T cell function, OKT3 has found wide application as an immunosuppressant in therapy regimens to reduce or even abolish allograft tissue rejection.

[0004] OKT3 affects allograft tissue rejection most likely by blocking the function of all T cells, which play an important role in acute rejection. OKT3 interacts with and blocks the function of the CD3 complex in the membrane of human T cells, which is associated with the T cell antigen recognition (TCR) structure and is essential for signal transduction. Which TCR/CD3 subunit binds to OKT3 has been the subject of much investigation. However, some evidence of 0KT3 specificity for the epsilon subunit of the TCR/CD3 complex has been reported (Tunnacliffe, Int. Immunol. 1 (1989), 546-50; Kjer-Nielsen, PNAS 101, (2004), 7675-7680). Further evidence has shown that OKT3 binding to the TCR/CD3 complex requires the presence of other subunits of this complex (Salmeron, J. Immunol. 147 (1991), 3047-52).

[0005] Other known antibodies specific for the CD3 molecule are listed in Tunnacliffe, Int. Immunol. 1 (1989), 546-50. As previously indicated, such CD3-specific antibodies are capable of inducing various T cell responses, such as lymphokine production (Von Wussow, J. Immunol. 127(1981), 1197; Palacious, J. Immunol. 128(1982), 337), proliferation (Van Wauve, J. Immunol. 124). (1980), 2708-18) and induction of suppressor T cells (Kunicka, in "Lymphocyte Tvping II" 1 (1986), 223). That is, depending on the experimental conditions, a CD3-specific monoclonal antibody can inhibit or induce cytotoxicity (Leevvenberg, J. Immunol. 134

(1985), 3770; Phillips, J. Immunol. 136 (1986) 1579; Platsoucas, Proc. Natl. Acad. Sci. USA 78

(1981), 4500; Itoh, Cell. Immunol. 108 (1987), 283-96; Mentzer, J. Immunol. 135 (1985), 34; Landegren, J. Exp. Med. 155 (1982), 1579; Choi (2001), Eur. J. Immunol. 31, 94 -106; Xu

(2000), Cell Immunol. 200, 16-26; Kimball (1995), Transpl. Immunol. 3, 212 - 221).

[0006] Although many CD3 antibodies described in the art are reported to recognize the CD3 epsilon subunit of the CD3 complex, most of them actually bind to conformational epitopes and, thus, only recognize CD3 epsilon in the native TCR context. Conformational epitopes are characterized by the presence of two or more separate amino acid residues that are separated in the primary sequence, but come together on the surface of the molecule when the polypeptide is folded into the native protein/antigen (Sela, (1969) Science 166,1365 and Laver, (1990) Cell 61, 553-6). The conformational epitopes bound by CD3 epsilon antibodies described in the literature can be separated into two groups. In a larger group, said epitopes are created by two CD3 subunits, e.g. of CD3 epsilon chain and CD3 gamma or

[0007] WO 2007/033230 describes antibodies to the EMGGITQTPYKVSISGT CD3 sequence corresponding to residue 6-22 of human CD3 epsilon

[0008] CD3 delta chain. For example, the commonly used CD3 epsilon monoclonal antibodies OKT3, WT31, UCHT1, 7D6 and Leu-4 were found in multiple studies to not bind to cells that were individually transfected with the CD3-epsilon chain. However, these antibody stained cells were double transfected with a combination of CD3 epsilon plus or CD3 gamma or CD3 delta (Tunnacliffe, loc. cit.; Law, Int. Immunol. 14(2002), 389-400; Salmeron, J. Immunol. 147 (1991), 3047-52; Coulie, Eur. J. Immunol. 21 (1991), 1703-9). In another, smaller group, a conformational epitope is created within the CD3 epsilon subunit itself. A member of this group is, for example, Ab APA 1/1 which is generated against denatured CD3 epsilon (Risueno, Blood 106 (2005), 601-8). In total, most CD3 epsilon antibodies described in the art recognize conformational epitopes that are located on two or more CD3 subunits. The separate amino acid residues that make up the three-dimensional structure of these epitopes can thus be located either on the CD3 epsilon subunit itself or on the CD3 epsilon subunit and on other CD3 subunits such as CD3 gamma or CD3 delta.

[0009] Another problem with CD3 antibodies is that many CD3 antibodies have been found to be species specific. Anti-CD3 monoclonal antibodies—as is generally true of any other monoclonal antibody—function by highly specific recognition of their target molecules. They recognize only one site, or epitope, on their target CD3 molecule. For example, one of the most widely used and best characterized monoclonal antibodies specific for the CD3 complex is OKT-3. This antibody reacts with chimpanzee CD3 but not with CD3 homologues from other primates, such as macaques, or dog CDR3 (Sandusk et al., J. Med. Primatol. 15 (1986), 441-451). The anti-CD3 monoclonal antibody UCHT-1 also reacts with chimpanzee CD3 but not macaque CD3 (own data). On the other hand, there are also examples of monoclonal antibodies that recognize macaque antigens but not their human counterparts. One example of this group is the monoclonal antibody FN-18, directed against macaque CD3 (Uda et al., J. Med. Primatol. 30 (2001), 141-147). Interestingly, peripheral lymphocytes from about 12% of cynomolgus monkeys were found to fail to react with the anti-Rhesus monkey CD3 monoclonal antibody (FN-18) due to a polymorphism of the CD3 antigen in macaques. Uda et al. described a two amino acid substitution in the cynomolgus monkey CD3 sequence, which does not react with FN-18 antibodies, compared to animal-derived CD3, which reacts with FN-18 antibodies (Uda et al., J Med Primatol. 32 (2003), 105-10; Uda et al., J Med Primatol. 33 (2004), 34-7).

[0010] This discrimination ability, ie. the interspecies specificity inherent in CD3 monoclonal antibodies and their fragments is a significant obstacle to their development as therapeutic agents for the treatment of human diseases. In order to receive a license for use, each new drug candidate must go through rigorous testing. This testing can be divided into preclinical and clinical phases. While the latter - further divided into generally known clinical stages I, II and III - is performed on human patients, the first is performed on animals. The purpose of preclinical testing is to confirm that the drug candidate has the desired activity and most importantly, that it is safe. Only when pre-clinical testing has established safety in animals and possible efficacy of the drug candidate, will the appropriate regulatory authority approve it for clinical testing in humans. A drug candidate can be tested for safety in animals in the following three ways, (i) in relevant species, i.e. species where the drug candidate can recognize orthologous antigens, (ii) in transgenic animals that contain human antigens, and (iii) using surrogates for the drug candidate that can bind orthologous antigens present in animals. The limitations of transgenic animals are that this technology is usually limited to rodents. There are considerable differences in physiology between rodents and humans and safety results cannot be easily extrapolated to humans. Limitations of surrogates for a drug candidate are different material compositions compared to the current drug candidate and often rodent animals are used with the aforementioned limitations. Therefore, preclinical data obtained in rodents are of limited predictive value regarding the potential drug under investigation. The approach of choice for safety testing is the use of relevant species, preferably a lower primate. A limitation described in the art now being placed on CD3 binding molecules suitable for therapeutic intervention in humans is that the relevant species are higher primates, especially chimpanzees. Chimpanzees are considered an endangered species and because of their resemblance to humans, the use of such animals for drug safety testing is banned in Europe and very restricted elsewhere.

[0011] The present invention relates to polypeptides containing a first binding domain that is an antibody capable of binding to the epitope of the CD3e chain of human and Callithrix jacchus, Saguinus

oedipus or Saimiri sciureus, wherein the epitope is part of an amino acid sequence found in a group consisting of SEQ ID NO:2, 4, 6, or 8 and contains at least the amino acid sequence Gln-Asp-Gly-Asn-Glu, and another binding domain capable of binding to human or non-chimpanzee EGFR, Her2/neu, or IgE. The sequences as set forth in SEQ ID NO. 2, 4, 6 and 8, their fragments are context-independent CD3 epitopes.

[0012] The advantage of the present invention is the provision of a polypeptide containing the first human binding domain, which is an antibody that exhibits interspecies specificity for the CD3e (epsilon) chain of humans and primates other than chimpanzees, which can be used both for preclinical determination of the safety, activity and/or pharmacokinetic profile of these primarily human binding domains in primates, and -in identical form- as a medicine for humans. The same molecule can be used in preclinical tests on animals as well as in clinical trials on humans. This provides highly comparable results and greatly increased predictive power in animal studies compared to species-specific surrogate molecules. In the present invention, the N-terminal 1-27 amino acid residue of the polypeptide fragment of the extracellular domain of CD3 epsilon was surprisingly identified and - contrary to all other known epitopes of CD3 epsilon described in the art - it retains its three-dimensional structural integrity when removed from its natural environment in the CD3 complex (and combined with a heterologous amino acid sequence such as EpCAM or an immunoglobulin Fc fragment).

[0013] The context-independent CD3 epitope provided in the present invention corresponds to the first 27 N-terminal amino acids of CD3 epsilon or a series of functional fragments of these 27 amino acids. The phrase "context-independent," as used herein in connection with a CD3 epitope means that binding of the binding antibody molecule(s) described herein does not result in a change or modification of the conformation, sequence, or structure surrounding the antigenic determinant or epitope. Contrary to this. A CD3 epitope recognized by a conventional CD3 binding molecule (eg, as described in WO 99/54440 or WO 04/106380) is localized on the CD3 epsilon C-terminal chain to the N-terminal 1-27 amino acids of the context-independent epitope, wherein it has the correct conformation if it is incorporated within the remainder of the epsilon chain and held in the correct position by heterodimerization of the epsilon chain either with CD3 gamma or delta chain.

[0014] Anti-CD3 binding molecules as part of a bispecific binding molecule as provided herein and generated (and directed) to a context-independent CD3 epitope provided a surprising clinical improvement over T cell redistribution and, thus, a more favorable safety profile. Without being bound by theory, since the CD3 epitope is context-independent, creating an autonomous self-sufficient subdomain without much effect on the rest of the CD3 complex, the CD3 binding molecules provided herein induce fewer allosteric changes in CD3 conformation than conventional CD3 binding molecules (such as molecules provided in WO 99/54440), which recognize context-dependent CD3 epitopes.

[0015] The context-independence of the CD3 epitopes of the CD3 binding molecules of the invention as part of a bispecific binding molecule is associated with less redistribution of T cells during the initial phase of treatment with the CD3 binding molecules of the invention leading to a better safety profile of the CD3 binding molecules of the invention compared to conventional CD3 binding molecules known in the art, which recognize context-dependent CD3 epitopes. In particular, since the redistribution of T cells during the initial phase of treatment with CD3 binding molecules is a major risk factor for adverse CNS events, the CD3 binding molecules of the invention by recognizing a context-specific rather than a context-dependent CD3 epitope have a substantial safety advantage over CD3 binding molecules known in the art. Patients with such CNS adverse events associated with T cell redistribution during the initial phase of treatment with conventional CD3 binding molecules suffer from confusion and disorientation, in some cases from urinary incontinence. Confusion is a change in mental status in which the patient is unable to think at their usual level of clarity. The patient usually has difficulty concentrating and his thinking is not only clouded and unclear, but often significantly slowed down. Patients with CNS adverse events associated with T cell redistribution during the initial phase of treatment with conventional CD3 binding molecules may also suffer from memory loss. Often, the confusion leads to a loss of the ability to recognize people and/or places, or to state times or dates. Feelings of disorientation are common with confusions, and the ability to make decisions is impaired. CNS adverse events associated with T cell redistribution during the initial phase of treatment with conventional CD3 binding molecules may additionally include slurred speech and/or difficulty finding speech. This disorder can impair both the expression and understanding of language, as well as reading and writing. In addition to urinary incontinence, dizziness and fainting may also accompany CNS adverse events associated with T cell redistribution during the initial phase of treatment with conventional CD3 binding molecules in some patients.

[0016] The maintenance of the three-dimensional structure within the mentioned fragment of the N-terminal polypeptide of 27 amino acids of CD3 epsilon can be used to create binding domains, primarily human, which bind to the fragment of the N-terminal polypeptide of CD3 epsilon in vitro and to the native (CD3 epsilon subunit) CD3 complex on T cells in vivo with the same binding affinity. These data clearly indicate that the N-terminal fragment, as described here, builds a tertiary conformation, which is similar to the structure that normally exists in vivo. A very sensitive test for the importance of the structural integrity of amino acids 1-27 of the N-terminal polypeptide CD3 epsilon fragment was performed. Individual amino acids of amino acids 1-27 of the N-terminal fragment of the CD3 epsilon polypeptide were changed to alanine (alanine scan) to test the sensitivity of amino acids 1-27 of the N-terminal polypeptide fragment of CD3 epsilon to smaller breaks. CD3-specific antibody molecules as part of the bispecific binding molecule of the invention were used to test binding to alanine mutants of amino acids 1-27 of the N-terminal polypeptide fragment of CD3 epsilon (see attached Example 5). Individual changes of the first five amino acid residues at the very N-terminal end of the fragment and two amino acids at positions 23 and 25 of amino acids 1-27 of the N-terminal polypeptide fragment of CD3 epsilon were critical for the binding of antibody molecules. Substitution of amino acid residues in the region of positions 1-5 containing residues Q (glutamine at position 1), D (aspartic acid at position 2), G (glycine at position 3), N (asparagine at position 4), and E (glutamic acid at position 5) to alanine abolished the binding of the primarily human binding molecules of the invention to said fragment. Thus, at least for some time, the primarily human binding molecules of the invention, two amino acid residues at the C-terminus of the mentioned fragment (threonine at position 23 and isoleucine at position 25) reduced the binding energy for the primarily human binding molecules of the invention.

[0017] Unexpectedly, thus isolated, primarily human, binding molecules were found to not only recognize the human N-terminal fragment of CD3 epsilon, but also the corresponding homologous fragments of CD3 epsilon from various primates, including New World monkeys.

(Marmoset, Callithrix jacchus; Saguinus oedipus; Saimiri sciureus) and Old World monkeys (Macaca fascicularis, also known as Cinomolgus monkey, or Macaca mulatta, also known as Rhesus monkey). In this way, the multi-primate specificity of CD3-binding molecules was detected. The following sequence analyzes confirmed that humans and primates share a series of highly homologous sequences at the N-terminus of the CD3 epsilon extracellular domain.

[0018] In the present invention, as defined in the claims, it was found possible to generate, primarily human, binding molecules specific for CD3 epsilon, whereby the identical molecule can be used in preclinical testing on animals, as well as in clinical trials, or even in human therapy. chimpanzee counterparts), also bind to non-chimpanzee primate homologues of said antigens, including New World monkeys and Old World monkeys. As shown in the Examples that follow, said CD3 epsilon specific, primarily human, binding molecules can be integrated into single chain bispecific antibodies to generate therapeutic agents against various diseases, including, without limitation, cancer and immune disorders. Thus, the need to construct a surrogate CD3 epsilon binding domain or bispecific single-chain antibody, including those for testing in phylogenetically distant (from human) species, is eliminated. As a result, the same molecule can be used in preclinical testing on animals as it is intended for use in humans in clinical testing, as well as for subsequent licensing and therapeutic administration of the drug. The ability to use the same molecule for preclinical animal testing as in later human application virtually eliminates, or at least greatly reduces, the danger that data obtained in preclinical animal testing will have limited applicability to human cases.. In short, obtaining preclinical safety data in animals using the same molecule that will actually be applied in humans does much to ensure the applicability of the data to the relevant human scenario. In contrast to this, in the conventional approach when surrogate molecules are used, said surrogate molecules have to be adjusted for the animal testing system used with preclinical safety evaluation. Thus, the molecule to be used in human therapy differs in sequence and possibly in structure from the surrogate molecule used in preclinical testing, in pharmacokinetic parameters and/or biological activity; the consequence is that the data obtained in preclinical testing on animals have limited applicability / transferability to the case code. The use of surrogate molecules requires the construction, production, purification and characterization of an entirely new construct. This leads to additional development costs and additional time needed to obtain that molecule. Globally, surrogates must be developed separately in addition to the active drug to be used in human therapy, so two lines of development must be performed for the two molecules. Therefore, a major advantage of a human binding molecule or antibody-based construct described herein that exhibits interspecies specificity is that the identical molecule can be used for human therapy and preclinical animal testing.

[0019] It is preferred for the polypeptide of the invention that the first binding domain representing an antibody is capable of binding to an epitope of the CD3 epsilon chain of a human and a non-chimpanzee primate of human origin.

[0020] In addition, due to the human origin of the binding human molecules of the invention, the generation of an immune reaction against said binding molecules is excluded to the greatest extent possible after the administration of the binding molecules to humans.

[0021] Another great advantage of the primarily human, CD3 epsilon-specific human binding molecules, as part of the bispecific binding molecules of the invention, is that they can be used for preclinical testing on various primates. The behavior of a drug candidate in animals should ideally be indicative of the expected behavior of the drug when administered to humans. As a result, data obtained from such preclinical testing should therefore be highly predictive of human cases. However, as learned from the tragic outcome of the recent Phase I clinical trial of TGN1412 (a CD28 monoclonal antibody), a drug candidate may work differently in a primate species than in humans. While no adverse effects, or only limited adverse effects, were observed in preclinical testing of the said antibody, in animal studies performed on cynomolgus monkeys, six people experienced multiple organ damage when administered the said antibody (Lancet 368 (2006), 2206-7). The results of these adverse events suggest that it may not be sufficient to limit preclinical trials to only one species (primates). The fact that the CD3 epsilon specific human binding molecules of the invention bind to New and Old World monkey series may help to overcome the problems encountered in the aforementioned case. Accordingly, the subject invention, as defined in the claims, provides ways and methods for minimizing species differences that occur when human therapeutic drugs are developed and tested.

[0022] With primarily human, interspecies-specific CD3 epsilon binding domains as part of the bispecific binding molecules of the invention, it is also no longer necessary to adapt animal tests to a drug candidate intended for human administration, such as, e.g. creation of transgenic animals. CD3 epsilon-specific, primarily human, binding molecules (or bispecific single-chain antibodies containing them) that exhibit interspecies specificity according to the uses and methods of the invention can be directly used for preclinical testing in non-chimpanzee primates, without any genetic manipulation of the animals. As is well known to one of ordinary skill in the art, approaches in which a test animal is adapted to a drug candidate always carry the risk that results obtained in preclinical safety testing are less representative and predictive of humans due to animal modification. For example, in transgenic animals, the proteins encoded by the transgene are often overexpressed. Therefore, the data obtained for the biological activity of the antigen against this antigen protein may be limited in their predictive value for humans in whom the protein is expressed at a much lower, more physiological level.

[0023] A further advantage of using CD3 epsilon-specific, primarily human, binding molecules (or bispecific single-chain antibodies containing them) that exhibit interspecies specificity is the fact that chimpanzees, as an endangered species, are avoided for animal testing. Chimpanzees are the closest relatives of humans and have recently been grouped into the hominid family based on genome sequence data (Wildman et al., PNAS 100 (2003), 7181). Therefore, data obtained from chimpanzees are generally considered highly predictive of humans. However, due to their status as an endangered species, the number of chimpanzees that can be used for medical experiments is very limited. As indicated earlier, maintaining chimpanzees for animal testing is also expensive and ethically problematic. The use of CD3 epsilon-specific, primarily human, binding molecules (or bispecific single-chain antibodies that contain them) avoids both ethical objections and the burden of financial preclinical testing, without prejudicing quality, i.e. applicability of data obtained by animal testing. In light of this, the use of CD3 epsilon-specific, primarily human, binding molecules (or bispecific single-chain antibodies containing them) provides a reasonable alternative to chimpanzee studies.

[0024] Another advantage of CD3 epsilon-specific, primarily human, binding molecules (or bispecific single-chain antibodies containing them) is the ability to collect multiple blood samples when used as part of preclinical animal testing, for example, during pharmacokinetic studies. Collecting multiple blood samples is easier in non-chimpanzee primates than in lower animals, e.g. of mice.. Taking multiple blood samples allows continuous testing of blood parameters to determine the biological effects caused by the introduction of primarily human binding molecules (or bispecific single-chain antibodies) of the invention. In addition, the collection of multiple blood samples allows the investigator to determine the pharmacokinetic profile of, for example, human binding molecules (or bispecific single chain antibodies) as defined herein. Furthermore, potential side effects, which may be caused primarily by human, binding molecules (or bispecific single-chain antibodies), and which are reflected in blood parameters, can be measured in different blood samples taken during the administration of said antibody. This allows determination of the potential toxicity profile of primarily human, binding molecules (or bispecific single-chain antibodies) as defined herein.

[0025] The advantages of primarily human, binding molecules (or bispecific single-chain antibodies) as defined here, which exhibit interspecies specificity can be briefly summarized as follows: First, primarily human, binding molecules (or bispecific single-chain antibodies), as defined here, used in preclinical testing are the same as those used in human therapy. Therefore, it is no longer necessary to develop two independent molecules, which may differ in pharmacokinetic properties and biological activity. This is a very big advantage because, for example, pharmacokinetic results can be more directly transferred and applied to humans than is the case with e.g. conventional approach with a surrogate.

Second, the use of primarily human binding molecules (or bispecific single-chain antibodies), as defined herein, to obtain human therapeutics costs less and requires less effort than a surrogate.

Third, primarily human binding molecules (or bispecific single-chain antibodies), as defined herein, can be used for preclinical testing not only in a single primate species, but in a series of different primate species, thus limiting the risk of potential differences between primates and humans.

Fourth, the use of chimpanzees as an endangered species for animal testing was avoided.

Fifth, it is possible to draw multiple blood samples for extensive pharmacokinetic studies.

Sixth, due to the human origin, primarily human, of the binding molecules according to the preferred embodiment of the invention, the creation of an immune reaction to said binding molecules is minimized when they are administered to humans. Elicitation of an immune response with drug candidate-specific antibodies performed in a non-human species, eg mouse, leading to the development of human-anti-human antibodies (HAMAs) to therapeutic molecules of murine origin, is excluded.

[0026] The term "protein" is well known in the art and describes biological compounds. Proteins contain one or more chains of amino acids (polypeptides), whereby the amino acids are linked to each other by peptide bonds. The term "polypeptide" as used herein describes a group of molecules consisting of more than 30 amino acids. According to the invention, a group of polypeptides comprises "proteins" as long as the proteins consist of a single polypeptide. Also by definition, the term "polypeptide" describes protein fragments as long as these fragments consist of more than 30 amino acids. Polypeptides can further build multimers such as dimers, trimers and higher oligomers, i.e. those consisting of more than one polypeptide molecule. Polypeptide molecules that build such dimers, trimers, etc. they can be identical or different. The corresponding higher order structures of these multimers are therefore called homo- or heterodimers, homo- or heterotrimers, etc. An example of a heteromultimer is an antibody molecule, which, in its native form, consists of two identical light polypeptide chains and two identical heavy polypeptide chains. The terms "polypeptide" and "protein" also refer to naturally modified polypeptides/proteins where the modification is caused by e.g. post-translational modifications such as glycosylation, acetylation, phosphorylation and the like. Such modifications are well known in the art.

[0027] As used herein, "human" and "human" refer to the species Homo sapiens. For medical use of the constructs described herein, a human patient will be treated with the same molecule.

[0028] The term "human origin" as used in the context of molecules of the invention describes molecules derived from human libraries or having a structure/sequence corresponding to a human equivalent. Therefore, proteins that have an amino acid sequence that matches the analogous sequence in humans, eg an antibody fragment that has an amino acid sequence in an assembly that matches the sequence of a human embryo, is considered a molecule of human origin.

[0029] As used herein, "non-chimpanzee primate" or "non-chimp primate" or their grammatical variants) refers to any primate that is not a chimpanzee, ie. animals that do not belong to the genus Pan, and include the species Pan paniscus and Pan troglodytes, also known as Anthropopithecus troglodytesili Simia satyrus."Primate", "species of primate", "primates" or their grammatical variants means the order of placental mammals which is divided into two suborders prosimians and anthropoids and includes man, apes, apes and lemurs. In particular, the term "primates" as used herein includes the suborder Strepsirrhini (non-tarsarian prosimians), including the suborder Lemuriformes (which includes the superfamilies Cheirogoleidae and Lemuroidge), the suborder Chiromyiformes (which includes the family Daubentoniidae), and the suborder Lorisiformes (which includes the families Lorisidae and Galagidae). "Primates," as used herein, also includes the suborder Haplorrhini, including the suborder Tarsiiformes (which includes family Tarsiidae), suborder Simiiformes (which includes Platyrrhini, or New World monkeys, and Catarrhini, including Cercopithecidae, or Old World monkeys).

[0030] Species other than chimpanzees can be understood within the meaning of the invention as lemur, tarsier, gibbon, marmoset (belonging to New World monkeys of the family Cebidae) or Old World monkey (belonging to the superfamily Cercopithecoidea).

[0031] As used herein, "old world monkey, (recumbent monkey)" includes any monkey from the superfamily Cercopithecoidea, which itself is divided into families: Cercopithecinae, which are mostly African but include various species of macaques that are Asian and North African, and Colobinae, which include Asian genera but also African colobus monkeys.

[0032] Especially from the superfamily Cercopithecinae, a suitable non-chimpanzee primate can be from the tribe Cercopithecini, from the genus Allenopithecus (Allen's swamp monkey, Allenopithecus nigroviridis); from the genus Miopithecus (Angolan talapoin, Miopithecus talapoin; Gabonese talapoin, Miopithecus ogouensis); from the genus Erythrocebus (Patas monkey, Erythrocebus patas) from the genus Chlorocebus (green monkey, Chlorocebus sabaceus; grivet, Chlorocebus aethiops; squirrel monkey from the Bale Mountains, Chlorocebus djamdjamensis; tantalus monkey, Chlorocebus tantalus; squirrel monkey, Chlorocebus pygerythrus; malbruk, Chlorocebus cinosuros); or from the genus Cercopithecus (dryas monkey or salongo monkey, Cercopithecus dryas; diana monkey, Cercopithecus diana; roloway monkey, Cercopithecus roloway, greater white-nosed monkey, Cercopithecus nictitans; blue monkey, Cercopithecus mitis; silver monkey, Cercopithecus doggetti; golden monkey, Cercopithecus kandti; Sykes monkey, Cercopithecus albogularis; Mona monkey, Cercopithecus mona; Campbell's mona monkey, Cercopithecus campbelli; Lowe's mona monkey, Cercopithecus lowei; crested mona monkey, Cercopithecus pogonias; Wolf's mona monkey, Cercopithecus vvolfi; Dent's mona monkey, Cercopithecus denti; lesser white-nosed monkey, Cercopithecus petaurista; white-necked guenon, Cercopithecus erythrogaster, Sclater's guenon, Cercopithecus sclateri; red-eared guenon, Cercopithecus erythrotis; whiskered guenon, Cercopithecus cephus; red-tailed guenon, Cercopithecus ascanius; L'Hoest's monkey, Cercopithecus Ihoesti; Preuss's guenon, Cercopithecus erythrotis preussi; sun-tailed monkey, Cercopithecus solatus; Hamlyn's monkey or owl-faced monkey, Cercopithecus hamlyni; De Brazza monkey, Cercopithecus neglectus).

[0033] Alternatively, suitable non-chimpanzee primates, also of the superfamily Cercopithecinaeali of the tribe Papionini, may be of the genus Macaca (Barbary macaques, Macaca sylvanus; lion-tailed macaques, Macaca silenus; southern pig-tailed macaques or Beruk, Macaca nemestrina; northern pig-tailed macaques, Macaca leonina; Pagai macaques or Bokkoi, Macaca pagensis; Siberian macaques, Macaca siberu; black macaques, Macaca maura; booted macaques Macaque, Macaca ochreata; Tonkean macaques, Macaca tonkeana; Heck's macaques, Macaca hecki; Gorontalo macaques, Macaca nigriscens; celebes crested macaques or black "man-shaped monkeys". macaques, Macaca arctoides; rhesus macaques, Macaca mulatta; Taiwanese macaques, Macaca cyclopis; Japanese macaques, Macaca fuscata; Sri Lankan macaques, Macaca sinica; Indian macaques, Macaca radiata; Barbary macaques, Macaca sylvanus; Asan macaques, Macaca assamensis; Tibetan macaques or Milne-Edwards's macaques, Macaca thibetana; Arunachal macaques or munzala, Macaca munzala) ; from genus Lophocebus

Lophocebus albigena; Lophocebus albigena albigena; Lophocebus albigena osmani;

Lophocebus albigena johnstoni; black crested mangabey, Lophocebus aterrimus; Opdenbosch's mangabey, Lophocebus opdenboschi; Lophocebus kipunji) ; from the genus Papio (maned baboon, Papio hamadryas; Guinea baboon, Papio papio; olive baboon, Papio anubis; yellow baboon, Papio cinocephalus; bear baboon, Papio ursinus); from the genus Theropithecus (Gelada baboon, Theropithecus gelada); from the genus Cercocebus (sooty mangabey, Cercocebus atys; Cercocebus atys atys; Cercocebus. atys lunulatus; white-necked mangabey, Cercocebus torquatus; agile mangabey, Cercocebus agilis; golden-breasted mangabey, Cercocebus chrysogaster; tan mangabey, Cercocebus sanjei); or from the genus Mandrillus (Mandrillus, Mandrillus sphinx; Mandrillus leucophaeus).

[0034] Most preferred are Macaca fascicularis (also known as cynomolgus monkey and, therefore, called "cynomolgus" in the Examples) iMacaca mulatta (rhesus monkey, called "rhesus").

[0035] From the superfamily Colobinae, a suitable primate besides the chimpanzee can be from the African group, from the genus Colobus (black colobus, Colobus satanas; Angolan colobus, Colobus angolensis; royal colobus, Colobus polykomos; bear colobus, Colobus vellerosus; northern black and white colobus, Colobus guereza); from the genus Piliocolobus (western red colobus, Piliocolobus badius;

Piliocolobus badius badius; Piliocolobus badius temminckii; Piliocolobus badius vvaldronae;

Pennant's Colobus,Piliocolobus pennantii; Piliocolobus pennantii pennantii; Piliocolobus pennantii epieni; Piliocolobus pennantii bouvieri;Preuss's red colobus,Piliocolobus preussi;Thollon's red colobus,Piliocolobus tholioni,Central African red colobus,Piliocolobus

foai; Piliocolobus foai foai; Piliocolobus foai elliott, Piliocolobus foai oustaleti; Piliocolobus foai

semlikiensis; Piliocolobus foai parmentierorum; Uganda red colobus, Piliocolobus tephrosceles; Uzungwan red colobus, Piliocolobus gordonorum; Zanzibar red colobus, Piliocolobus kirkii; Tanian red colobus, Piliocolobus rufomitratus); or from the genus Procolobus (olive colobus, Procolobus verus). From the superfamily Colobinae, suitable primates that are not

chimpanzees can alternatively be from the Langur group (English leaf monkey), from the genus Semnopithecus (Nepal gray langur, Semnopithecus schistaceus; Kashmiri gray langur, Semnopithecus ajax; Taraje gray langur, Semnopithecus hector, Hanuman's langur, Semnopithecus entellus; black-footed gray langur, Semnopithecus hypoleucos; gray langur of the Southern Plains, Semnopithecus dussumieri; tufted gray langur, Semnopithecus priam) ; from T. vetulus group or genusTrachypithecus (purple-legged langur, Trachypithecus vetulus; Nilgiri langur, Trachypithecus johnh); from T. cristatus group of genus Trachypithecus (Javanese lutung, Trachypithecus auratus; silvery leaf monkey or silvery lutung, Trachypithecus cristatus; Indochinese lutung, Trachypithecus germaini; Tenasserim lutung, Trachypithecus barbel) ; from 7. obscurusgenus groups

Trachypithecus (engl. Dusky Leafmajmuniliengl. SpectacIed Leaf'majmun, Trachypithecus

obscurus; Phayre's Leaf Monkey, Trachypithecus phayrei) ; from T. pileatus of the genus Trachypithecus (Capped langur, Trachypithecus pileatus; Shortridge's langur, Trachypithecus shortridgei; Gee's golden langur, Trachypithecus gee), from the group T. francoisi of the genus Trachypithecus (Francois's langur, Trachypithecus francoisi; Hatinh langur, Trachypithecus hatinhensis; white-headed langur, Trachypithecus potiocephalus; Lao langur, Trachypithecus laotum; Delacour's langur, Trachypithecus delacouri; Indochinese black langur, Trachypithecus ebenus); or from the genus Presbytis (Sumatran surili, Presbytis melalophos; Banded surili, Presbytis femoralis; Sarawak surili, Presbytis chrysomelas; VVhite-thighed surili, Presbytis siamensis; VVhite-fronted surili, Presbytis frontata; Javanese surili, Presbytis comata; Thomas's langur, Presbytis thomasi; Hose's langur, Presbytis hosei; Maroon Leaf monkey, Presbytis rubicunda; Mentavvai langur or joja, Presbytis potenziani; Natuna Island surili, Presbytis natunae).

[0036] From the superfamily Colobinae, a suitable primate besides the chimpanzee can alternatively be from the group of monkeys with specific noses, from the genus Pygathrix (Red-shanked Douc, Pygathrix nemaeus; Black-shanked Douc, Pygathrix nigripes; Gray-shanked Douc, Pygathrix cinerea); from the genus Rhinopithecus (golden short-nosed monkey, Rhinopithecus roxellana; bieti; gray short-nosed monkey, Rhinopithecus brelicht, Tonkinese short-nosed langur, Rhinopithecus avunculus) ; from the genus Našališ (Proboscis Monkey, Našališ larvatus); or from the genus Simias (pig-tailed langur, Simias concolor).

[0037] As used herein, the term "marmoset" refers to any New World monkey (long-tailed monkey) rhodasCallithrix, for example those belonging to the Atlantic marmosets of the subgenus Callithrix(sic!) (common marmoset, Callithrix (Callithrix) jacehus; black woolly marmoset, Callithrix (Callithrix) penicillata; Wied's marmoset, Callithrix (Callithrix) kuhlii; Marmozet, Callithrix ( Callithrix) geoffroyi; Buffy-headed Marmozet, Callithrix ( Callithrix) aurita) marmoset, Callithrix ( Mico) argentata; white marmoset, Callithrix ( Mico) leucippe; Emilia's marmoset, Callithrix ( Mico) emiliae; black-headed marmoset, Callithrix ( Mico) nigriceps; Marca's marmoset, Callithrix ( Mico) marcai; black-tailed marmoset, Callithrix ( Mico) melanura; Santarem marmoset, Callithrix ( Mico) humeralifera; Maues Marmoset, Callithrix ( Mico) mauesi; golden-white marmoset, Callithrix ( Mico) chrysoleuca; Hershkovitz's marmoset, Callithrix ( Mico) intermedia; Satere marmoset, Callithrix ( Mico) sateret); Roosmalen's dwarf marmoset belonging to the suborder Callibella ( Callithrix ( Callibelia) humilis); or pygmy marmoset belonging to the suborder. Cebuella (Callithrix (Cebuella) pygmaea).

[0038] Other genera of New World monkeys include tamarins of the genus Saguinus (which include the S.oedipus-group, the S. midas group, the S.nighcollis group, the S.mystax group, the S.bicolor group, and the S.inustus group) and squirrel monkeys of the genus Saimiri (eg Saimiri sciureus, Saimiri oerstedii, Saimiri

ustus, Saimiri boliviensis, Saimiri vanzolini)

[0039] The term "binding domain" characterizes in connection with the present invention a polypeptide domain that specifically binds/interacts with the specified target structure/antigen/epitope. Thus, the binding domain is the "antigen interaction site". The term "antigen interaction site" defines, in accordance with the present invention, a polypeptide motif, which is able to specifically react with a specific antigen or a specific group of antigens, e.g. identical antigen in different species. Said binding/interaction also defines "specific recognition". The term "specific recognition" means according to the invention, that the antibody molecule is able to specifically interact with and/or bind to at least two, preferably at least three, more preferably at least four, amino acids of the antigen, e.g. Human CD3 antigen, as defined herein. An example of such binding is the specificity of the "principle of lock and key". Thus, specific motifs in the amino acid sequence of the binding domain and the antigen bind to each other as a result of their primary, secondary or tertiary structure, and as a result of secondary modifications of said structure. The specific interaction of the antigen interaction site with its specific antigen can also lead to the simple binding of this site to the antigen. Furthermore, specific interaction of an antigen interaction site with its specific antigen can alternatively lead to signal initiation, e.g. due to causing a change in antigen conformation, antigen oligomerization, etc. A preferred example of a binding domain according to the present invention is an antibody. The binding domain can be a monoclonal or polyclonal antibody or derived from a monoclonal or polyclonal antibody.

[0040] The term "antibody" includes derivatives or functional fragments thereof that still retain binding specificity. Techniques for obtaining antibodies are well known in the art and are described, e.g. in Harlow and Lane "Antibodies, A Laboratory Manual", Cold Spring Harbor Laboratory Press, 1988 and Harlow and Lane "Using Antibodies: A Laboratory Manual" Cold Spring Harbor Laboratory Press, 1999. The term "antibody" also includes immunoglobulins (Ig's) of various classes (ie IgA, IgG, IgM, IgD and IgE) and subclasses (such as lgG1, lgG2 etc). These antibodies can be used, e.g. for immunoprecipitation, affinity purification, and immunolocalization of polypeptides or fusion proteins of the invention, as well as for monitoring the presence and amount of such polypeptides, for example, in cultures of recombinant prokaryotic or eukaryotic cells of organisms.

[0041] The definition of the term "antibody" also includes embodiments such as chimeric, single-chain antibodies, and humanized antibodies, as well as antibody fragments, such as Fab fragments, among others. Antibody fragments or derivatives further include F(ab')2, Fv, scFv fragments or single domain antibodies, single variable domain antibodies or single variable domain immunoglobulin containing only one variable domain, which may be VH or VL, which specifically binds an antigen or epitope independent of other V regions or domains, see e.g. Harlow and Lane (1988) and

(1999), loc. cit. Such single variable domain immunoglobulins contain not only an isolated single variable domain antibody polypeptide, but also larger polypeptides containing one or more single variable domain antibody monomers of the polypeptide sequence.

[0042] Different procedures are known in the art and can be used to obtain such antibodies and/or fragments. Thus, (antibody) derivatives can be obtained in a peptidomimetic manner. Furthermore, the techniques described for obtaining single chain antibodies (see, inter alia, US Patent 4,946,778) can be adapted to obtain single chain antibodies for the selected polypeptide(s). Also, transgenic animals can be used to express humanized antibodies specific for the polypeptides and fusion proteins of the present invention. To obtain monoclonal antibodies, any technique that provides antibodies obtained by continuous cell line culture can be used. Examples of such techniques include the hybridoma technique (Kohler and Milstein Nature 256 (1975), 495-497), the trioma technique, the human B-cell hybridoma technique (Kozbor, Immunology Today 4 (1983), 72), and the EBV-hybridoma technique for obtaining human monoclonal antibodies (Cole et al., Monoclonal Antibodies and Cancer Therapy, Alan R. Liss, Inc. (1985), 77 - 96). Surface plasmon resonance measurement as used in the BlAcore system can be used to increase the efficiency of phage antibodies that bind to epitopes of a target polypeptide, such as CD3 epsilon (Schier, Human Antitela Hybridomas 7 (1996), 97-105; Malmborg, J. Immunol. Methods 183 (1995), 7-13). It is also intended in the context of this invention, that the term "antibody" includes antibody constructs, which can be expressed in a host as defined in the following text, e.g. antibody constructs that can be transfected and/or transduced via, inter alia, viruses or plasmid vectors.

[0043] The term "specific interaction", as used herein in accordance with the subject invention, means that the binding (domain) molecule does not regulate or significantly cross-react with polypeptides that have a similar structure as those bound to the binding molecule, and that may be expressed by the same cells as the polypeptide of interest. Cross-reactivity of a panel of binding molecules to be tested can be tested, for example, by assaying the binding of said panel of binding molecules under conventional conditions (see, e.g., Harlow and Lane, Antibodies: A Laboratory Manual, Cold Spring Harbor Laboratory Press, 1988 and Using Antibodies: A Laboratory Manual, Cold Spring Harbor Laboratory Press, 1999). Examples of specific interaction of a binding domain with a specific antigen include the specificity of the ligand for its receptor. The given definition particularly includes the interaction of ligands, which induce a signal after binding to its specific receptor. An example of this interaction, which is particularly covered by the aforementioned definition, is the interaction of an antigenic determinant (epitope) with the binding domain (antigen binding site) of an antibody.

[0044] The term "cross-species specificity" or "interspecies specificity" as used herein refers to binding of the binding domain described herein to the same target molecule in humans and non-chimpanzee primates. Thus, "interspecies specificity" or "cross-species specificity" should be understood as interspecies reactivity for the same molecule X expressed in different species, but not for a molecule other than X. Interspecies specificity of a monoclonal antibody that recognizes e.g. CD3 epsilon human, and CD3 epsilon non-chimpanzee primates, e.g. CD3 epsilon macaques can be determined, for example by FACS analysis. FACS analysis was performed in such a way that the corresponding monoclonal antibody was tested for binding to human cells and non-chimpanzee primate cells, e.g. macaque cells expressing said CD3 epsilon antigens of humans and non-chimpanzee primates. A suitable test is shown in the following examples.

[0045] As used herein, CD3 epsilon means a molecule that is expressed as part of a T cell receptor and has the meaning commonly assigned to it in the art. In humans, it includes in individual or independently combined form all known CD3 subunits, for example CD3 epsilon, CD3 delta, CD3 gamma, CD3 zeta, CD3 alpha and CD3 beta. CD3 antigens of non-chimpanzee primates as set forth herein are, for example, Macaca fascicularisCD3 iMacaca mulattaCD3. In Macaca fascicularis, it includes CD3 epsilon FN-18 negative and CD3 epsilon FN-18 positive, CD3 gamma and CD3 delta. In Macaca mulatta, it includes CD3 epsilon, CD3 gamma and CD3 delta. Preferably, said CD3 as used herein is CD3 epsilon.

[0046] Human CD3 epsilon I listed in GenBank Accession No. NM_000733 and comprising SEQ ID NO. 1. The human CD3 gamma is designated in GenBank Accession NO. NM_000073. Human CD3 delta is listed in GenBank Accession No. NM_000732.

[0047] CD3 epsilon "FN-18 negative" of Macaca fascicularis (ie CD3 epsilon not recognized by monoclonal antibody FN-18 due to polymorphism as mentioned earlier) is listed in GenBank Accession no. AB073994.

[0048] CD3 epsilon "FN-18 positive" of Macaca fascicularis (ie, CD3 epsilon recognized by monoclonal antibody FN-18) is listed in GenBank Accession No. AB073993. CD3 gammaMacaca fascicularis listed in GenBank Accession No. AB073992. CD3 deltaMacaca fascicularis is listed under GenBank Accession No. AB073991.

[0049] Nucleic acid sequences and amino acid sequences of the corresponding CD3 epsilon, gamma and delta homologs of Macaca mulatta can be identified and isolated by recombinant techniques described in the literature (Sambrook et al. Molecular Cloning: A Laboratory Manual; Cold Spring Harbor Laboratory Press, 3rd edition 2001). This applies mutatis mutandis to CD3 epsilon, gamma and delta homologues of primates other than chimpanzee as defined herein. Amino acid sequence identification in Callithrix jacchus, Saimiri sciureusiSaguinus oedipus\ described in the attached examples. The amino acid sequence of the CD3 epsilon extracellular domain of Callithrix jacchus is described in SEQ ID NO: 3, that of Saguinus oedipus is described in SEQ ID NO: 5 and that of Saimiri sciureus is described in SEQ ID NO: 7.

[0050] In accordance with the above, the term "epitope" defines an antigenic determinant, which is specifically bound/identified by a binding molecule as previously defined. Binding domains or molecules can specifically bind or react with conformational or continuum epitopes, which are unique to the target structure, e.g. CD3 epsilon chain of human and non-chimpanzee primates. A conformational or discontinuous epitope is characterized for polypeptide antigens by the presence of two or more separate amino acid residues that are separated in the primary sequence but come together on the surface of the molecule when the polypeptide folds into the native protein/antigen (Sela,

(1969) Science 166, 1365 and Laver, (1990) Cell 61, 553-6). Two or more separate amino acid residues that contribute to an epitope are present on separate sections of one or more polypeptide chains. These residues come together on the surface of the molecule when the polypeptide chain(s) folds into a three-dimensional structure to create an epitope. In contrast, a continuous or linear epitope consists of two or more separate amino acid residues, which are present in one linear segment of the polypeptide chain. Within the scope of the present invention, the CD3 epitope "which does not depend on the context" refers to the conformation of said epitope. Such a context-independent epitope localized on the CD3 epsilon chain can develop its correct conformation only if it is embedded within the remainder of the epsilon chain and held in the correct position by heterodimerization of the epsilon chain with either the CD3 gamma or the delta chain. In contrast, the context-independent CD3 epitope provided herein refers to the N-terminal residue 1-27 amino acids of the CD3 epsilon polypeptide or functional fragment thereof. This N-terminal 1-27 amino acid polypeptide residue or a functional fragment thereof maintains the integrity of its three-dimensional and correct conformation when dissociated from its native environment in the CD3 complex. The context-independence of the N-terminal residue 1-27 amino acids of the polypeptide or its functional fragment, which is part of the extracellular domain of CD3 epsilon, thus represents an epitope that is completely different from the epitope of CD3 epsilon described in connection with the method for obtaining human binding molecules in WO 2004/106380. When using the mentioned method, only recombinant CD3 epsilon is used. The conformation of this expressed recombinant CD3 epsilon differed from that of its native form, that is, the form in which the CD3-epsilon subunit of the TCR/CD3 complex exists as part of a non-covalent complex with either the CD3-delta or the CD3-gamma subunit of the TCR/CD3 complex. When such expressed recombinant CD3-epsilon protein is used as an antigen to select antibodies from an antibody library, antibodies specific for this antigen are identified in the library, although such a library does not contain antibodies specific for autoantigens. This is due to the fact that exclusively expressed recombinant CD3-epsilon protein does not exist in vivo; it is not an autoantigen. Therefore, the subpopulation of B cells expressing antibodies specific for this protein is not depleted in vivo; an antibody library made from such B cells will not contain the genetic material for antibodies specific for the only expressed recombinant CD3-epsilon protein.

[0051] However, since the N-terminal residue of 1-27 amino acids of the polypeptide or its functional fragment is an epitope, which folds into its native form, binding domains according to the present invention cannot be identified by methods based on the approach described in WO 04/106380. Therefore, tests must verify that binding molecules such as those described in WO 04/106380 are not capable of binding to the N-terminal residues 1-27 of amino acids of the CD3 epsilon chain. Thus, conventional anti-CD3 binding molecules or anti-CD3 antibody molecules (eg, as described in WO 99/54440) bind the CD3 epsilon chain at a position more C-terminally located than the N-terminal residue 1-27 amino acids of the context-independent polypeptide or functional fragment thereof provided herein. The previously known antibody molecules OKT3 and UCHT-1 are also specific for the epsilon-subunit of the TCR/CD3 complex between amino acid residues 35 to 85 and, therefore, the epitope of these antibodies is also more C-terminally located. In addition to the horn, UCHT-1 binds to the CD3 epsilon chain in the region between amino acid residues 43 to 77 (Tunnacliffe, Int. Immunol. 1 (1989), 546 - 50; Kjer-Nielsen, PNAS 101, (2004), 7675 - 7680; Salmeron, J. Immunol. 147 (1991), 3047 - 52). Thus, previously known anti-CD3 molecules do not bind to, and are not directed to, the context-independent N-terminal residue 1-27 amino acid epitope defined herein (or a functional fragment thereof).

[0052] To create a primarily human binding domain containing polypeptide of the invention, e.g. in a bispecific single chain antibody as defined herein, can be used e.g. monoclonal antibodies that bind to both human and non-chimpanzee CD3 epsilon (eg, macaque CD3 epsilon).

[0053] In a preferred embodiment of the polypeptide of the invention, the non-chimpanzee primate is an Old World monkey. In an even more preferred embodiment of the polypeptide, the Old World monkey is a monkey of the genus Papio genus Macaca. Most preferably, the monkey of the macaque genus is the Assamese macaque (Macaca assamensis), Barbary macaque {Macaca sylvanus); long-tailed macaques or Java macaques (Macaca fascicularis), lion-tailed macaques (Macaca silenus); pig-tailed macaques (Macaca nemestrina); rhesus macaques (Macaca mulatta), Tibetan macaques (Macaca thibetana), Tonkean macaques (Macaca tonkeana), Sri Lankan macaques (Macaca sinica), short-tailed or red-faced macaques or bear macaques (Macaca arctoides or black macaques (Macaca maura).Most preferably, a monkey from the Papio genus is

Hamadryas Baboon, Papio hamadryas\Guinea Baboon, Papio papio; Olive Baboon, Papio

anubis; Yellow Baboon, Papio cynocephalus; Chacma Baboon, Papio ursinus.

[0054] In an alternative preferred embodiment of the polypeptide of the invention, the non-chimpanzee primate is a New World monkey. In a more preferred embodiment of the polypeptide, the New World monkey is a monkey of the genus Callithrix (marmoset), genus Saguinus, or genus Saimiri. Most importantly, the New World monkey from the genus Callithrix is Callithrix jacchus, the monkey from the genus Saguinus is Saguinus oedipus, the monkey from the genus Saimiri is Saimiri sciureus.

[0055] As previously defined herein, the polypeptide of the invention binds by the first binding domain to the epitope of the human and primate non-chimpanzee CD3£(epsilon) chain, wherein the epitope is part of an amino acid sequence contained in a group consisting of 27 amino acid residues as set forth in SEQ ID No. 2, 4, 6, or 8.

[0056] In accordance with the present invention, it is desirable for the polypeptide of the invention that the said epitope is part of an amino acid sequence containing 26, 25, 24, 23, 22, 21, 20, 19, 18, 17, 16, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6 or 5 amino acid.

[0057] Even more preferably, said epitope contains the amino acid sequence Gln-Asp-Gly-Asn-Glu (Q-D-G-N-E-D).

[0058] Within the scope of the present invention, "functional fragment of N-terminal residues 1-27 amino acids" means that said functional fragment is still a context-independent epitope that maintains the integrity of its three-dimensional structure when outside its natural environment in the CD3 complex (and fused to a heterologous amino acid sequence such as EpCAM or part of immunoglobulin Fc, e.g. as shown in Example 3.1). The maintenance of the three-dimensional structure within the N-terminal 27 amino acid polypeptide or its functional CD3 epsilon fragment can be used to create binding domains that bind to the N-terminal CD3 epsilon polypeptide fragment in vitro and to the natural (CD3 epsilon subunit) CD3 complex for T cells in vivo with the same binding affinity. Within the scope of the present invention, a functional fragment of N-terminal residues 1-27 amino acids indicates that the CD3 binding molecules provided herein can still bind to such functional fragments in a context-independent manner. One of ordinary skill in the art is aware of methods for epitope mapping to determine which amino acid residues of the epitope are recognized by such anti-CD3 binding molecules (eg, alanine scanning or pepp spot analysis).

[0059] In a preferred embodiment of the invention, the polypeptide of the invention comprises a (first) binding domain as defined herein and a second binding domain capable of binding to a cell surface antigen.

.[0060]The term "cell surface antigen" as used herein refers to a molecule displayed on the surface of a cell. In most cases, this molecule will be located in or on the plasma membrane of the cell so that at least part of this molecule remains accessible from the outside of the cell in a tertiary form. An example of a cell surface molecule, without limitation, that is located in the plasma membrane is a transmembrane protein, which contains, in its tertiary conformation, regions of hydrophilicity and hydrophobicity. Here, at least one hydrophobic region allows the cell surface molecule to embed, or insert, into the hydrophobic plasma membrane of the cell, while hydrophilic regions extend on both sides of the plasma membrane into the cytoplasm and extracellular space. Examples, without limitation, of cell surface molecules that are located at the plasma membrane are proteins that are modified at the cysteine residue to carry a palmitoyl group, proteins that are modified at the C-terminal cysteine residue to carry a famesyl group, or proteins that are modified at the C-terminus to carry a glycosyl phosphatidyl inositol ("GPI") anchor. These groups enable covalent attachment of proteins to the outer surface of the plasma membrane, where they remain available for recognition by extracellular molecules such as antibodies. Examples of cell surface antigens include EGFR, EGFRvlll, , MCSP, carbonic anhydrase IX (CAIX), CD30, CD33, Her2/neu, IgE, CD44v6, and Muc-1. In addition, examples of suitable cell surface antibodies include antigens which

are characteristic of a specific disease or weakness, i.e. cancer, autoimmune diseases or infectious diseases including viral infections. Accordingly, the term "cell surface antigen" explicitly includes viral proteins such as native, unprocessed viral proteins found on the surface of an infected cell (described inter alia for the envelope proteins of hepatitis B, C and HIV-1 viruses).

[0061] One defensive function of cytotoxic T cells is the destruction of virus-infected cells, therefore, the unique property of the bispecific binding molecules of the invention to activate and redirect cytotoxic T cells regardless of their native specificity, has a major impact on a broad field of chronic viral infections. For most of these cases, elimination of the persistently infected cells is the only chance of cure. Currently, adaptive T cell therapies have been developed against chronic CMV and EBV infections (Roonev, C.M., et al., Use of gene-modified virus-specific T lymphocytes to control Epstein-Barr-virus-related Ivmphoproliferation. Lancet, 1995. 345 (8941): p. 9-13; VValter, E.A., et al., Reconstitution of cellular immunity against cvmegalovirus in recipients of allogeneic bone marrow by transfer of T-cell clones from the donor. N Engl J Med, 1995. 333(16): p. 1038-44.

[0062] Chronic hepatitis B infection is clearly one indication that warrants great attention. Worldwide, between 350 and 400 million people are infected with HBV. The current treatment of chronic HBV hepatitis relies on interferon y and nucleoside or nucleotide analogues, it is a long-term therapy with significant side effects such as causing acute hepatitis (hepatitis flares), fever, myalgia, thrombocytopenia and depression.. Although there are now more than 4 approved therapeutic regimens, elimination of the virus is rarely achieved. Persistent inflammation in chronic hepatitis B leads to liver cirrhosis and hepatocellular carcinoma in more than 25% of patients. Furthermore, up to 40% of patients with chronic hepatitis will die from serious complications, accounting for 0.6 to 1.0 million deaths worldwide.

[0063] HBV, the prototype Hepadna virus, is an enveloped virus whose relaxed circular (rc) genome is reverse transcribed into an RNApregenome. After infection, the rc DNA is inserted into the hepatocyte nucleus where it is complemented with covalently closed circular DNA (cccDNA) containing 4 overlapping reading frames. It serves as the transcriptional template for the pregenomic RNA and three subgenomic RNAs. The pregenome RNA functions as an mRNA to move the viral core and polymerase protein. Infected cells continuously produce HBV surface protein (HBsAg) from cccDNA even when HBV replication is stopped. HBsAg consists of small surface proteins (S) with a very small proportion of medium and large (L) surface proteins. Both HBV S and L targeted are targeted to endoplasmic reticulum (ER) membranes from which they are transported to vesicle membranes via trans golgi organelles to the plasma membrane (Gorelick, F.S. and C. Shugrue, Exiting the endoplasmic reticulum. Mol Cell Endocrinol, 2001. 177 (1-2): p. 13 - 8). S and L proteins are permanently expressed on the surface of HBV replicating hepatocytes as recently shown (Chu, CM. and Y.F. Liaw, Membrane staining for hepatitis B surface antigen on hepatocvtes: a sensitive and specific marker of active viral replication in hepatitis B. J Clin Pathol, 1995. 48(5): p. 470 - 3).

[0064] Prototype viruses that exhibit coat proteins on the cell surface are hepatitis B virus (HBV), hepatitis C virus (HCV) and HIV-1. and they represent a large burden of disease globally. For the HIV-1 virus indication, it was recently shown that T cells modified by a chimeric TCR with an Fv antibody construct directed at the gp120 envelope protein can kill target cells infected with HIV-1 (Masiero, S., et al., T-cell engineering by a chimeric T-cell receptor with antibody-type specificity for the HIV-1 gp120. Gene Ther, 2005.12 (4): p. 299 - 310). Of the hepadnaviruses, hepatitis B virus (HBV) expresses the HBsAg envelope protein complex that is continuously generated from episomal cccDNA even when HBV replication has subsided.

[0065] Expression of intact S and L HBV proteins on the cell surface makes them accessible to antibodies that are the hallmark of seroconversion when patients recover from the acute phase of infection and when it changes from circulating HBsAg to antiHBs. If seroconversion does not occur, up to 30% of hepatocytes continue to express HBV S protein even after highly active and long-term antiviral therapy. Thus, apart from T lymphocytes that recognize specific intracellularly processed HBV peptides and presented by MHC molecules on the cell surface, recruitment of other forms of T cells is feasible on intact surface proteins such as S and L antigens available on the outer cell membrane. By using a fragment of a single-chain antibody that recognizes the small (S) and large envelope (L) proteins of the hepatitis B virus, artificial T-cell receptors were created that enable the targeting of grafted T-cells to infected hepatocytes and, after activation of these T-cells by contact with the antigen, the release of cytokines and the killing of infected hepatocytes.

[0066] Limitations of this approach are, (i) that T-cells need to be manipulated in vitro, (ii) retroviruses used for T-cell receptor transfer can cause insertional mutagenesis in T-cells, and (iii) once T-cell transfer is performed, the cytotoxic response cannot be limited..

[0067] To overcome these limitations, bispecific single-chain antibody molecules containing a first binding domain specific for human and non-chimpanzee CD3 epsilon antigen (as provided herein in the context of the present invention), as well as a second binding domain specific for HBV or HCV, envelope proteins of infected hepatocytes can be generated and fall within the scope of this invention. Within the scope of this invention it is also preferred that the second binding domain binds to the surface of human cell antigens and/or their non-chimpanzee primate counterparts, selected from EGFR, Her2/neu or IgE.

[0068] To create a second binding domain of the polypeptide of the invention, e.g. of a bispecific single-chain antibody, as defined herein, monoclonal antibodies that bind to surface antigens of both human and non-chimpanzee primate cells may be used. Suitable binding domains for a bispecific polypeptide as defined herein may e.g. be derived from interspecies-specific monoclonal antibodies by recombinant methods described in the literature. A monoclonal antibody that binds to a human cell surface antigen and to a non-chimpanzee non-chimpanzee homologue of said cell surface antigen can be tested by FACS assays as previously noted. It is clear to one of ordinary skill in the art that interspecies specific antibodies can also be generated by hybridoma techniques described in the literature (Milstein and Kohler, Nature 256 (1975), 495-7). For example, mice can be alternately immunized with human and primate non-chimpanzee CD33. From these mice, hybridoma cells producing an interspecies-specific antibody were isolated by the hybridoma technique and analyzed by FACS as previously described. The generation and analysis of bispecific polypeptides, such as bispecific single chain antibodies that exhibit interspecies specificity, as defined herein, are demonstrated in the following examples. The advantages of bispecific single-chain antibodies that exhibit interspecies specificity include the points listed below.

[0069] It is particularly preferred for the polypeptide of the invention that the first binding domain capable of binding to a human and non-chimpanzee epitope of the CD3e chain comprises a VL region comprising CDR-L1, CDR-L2 and CDR-L3 selected from: 1. (a) CDR-L1 as set forth in SEQ ID NO. 27, CDR-L2 as set forth in SEQ ID NO. 28 and CDR-L3 as set forth in SEQ ID NO. 29; 2. (b) CDR-L1 as set forth in SEQ ID NO. 117, CDR-L2 as set forth in SEQ ID NO. 118 and CDR-L3 as set forth in SEQ ID NO. 119; and 3. (c) CDR-L1 as set forth in SEQ ID NO. 153, CDR-L2 as set forth in SEQ ID NO. 154 and CDR-L3 as set forth in SEQ ID NO. 155.

[0070] The variable regions, ie the variable light chain ("L" of "VL") and the variable heavy chain ("H" of "VH") are considered in the art to provide the binding domains of the antibody. The first variable regions harbor complementarity-determining regions.

[0071] The term "complementarity determining region (CDR)" is well known in the art and dictates the antigenic specificity of an antibody. The term "CDR-L" or "L CDR" refers to CDRs in VL, while the term "CDR-H" or "H CDR" refers to CDRs in VH.

[0072] In an alternative preferred embodiment of the polypeptide of the invention, the first binding domain capable of binding to the human and non-chimpanzee CD3e chain comprises a VH region comprising CDR-H 1, CDR-H2 and CDR-H3 selected from: 1. (a) CDR-H 1 as set forth in SEQ ID NO. 12, CDR-H2 as set forth in SEQ ID NO. 13 and CDR-H3 as set forth in SEQ ID NO. 14; 2. (b) CDR-H 1 as set forth in SEQ ID NO. 30, CDR-H2 as set forth in SEQ ID NO. 31 and CDR-H3 as set forth in SEQ ID NO. 32; 3. (c) CDR-H 1 as set forth in SEQ ID NO. 48, CDR-H2 as set forth in SEQ ID NO. 49 and CDR-H3 as set forth in SEQ ID NO. 50; 4. (d) CDR-H 1 as set forth in SEQ ID NO. 66, CDR-H2 as set forth in SEQ ID NO. 67 and CDR-H3 as set forth in SEQ ID NO. 68; 5. (e) CDR-H 1 as set forth in SEQ ID NO. 84, CDR-H2 as set forth in SEQ ID NO. 85 and CDR-H3 as set forth in SEQ ID NO. 86; 6. (f) CDR-H 1 as set forth in SEQ ID NO. 102, CDR-H2 as set forth in SEQ ID NO. 103 and CDR-H3 as set forth in SEQ ID NO. 104; 7. (g) CDR-H 1 as set forth in SEQ ID NO. 120, CDR-H2 as set forth in SEQ ID NO. 121 and CDR-H3 as set forth in SEQ ID NO. 122; 8. (h) CDR-H 1 as set forth in SEQ ID NO. 138, CDR-H2 as set forth in SEQ ID NO. 139 and CDR-H3 as set forth in SEQ ID NO. 140; 9. (i) CDR-H 1 as set forth in SEQ ID NO. 156, CDR-H2 as set forth in SEQ ID NO. 157 and CDR-H3 as set forth in SEQ ID NO. 158; and 10. (j) CDR-H 1 as set forth in SEQ ID NO. 174, CDR-H2 as set forth in SEQ ID NO. 175 and CDR-H3 as set forth in SEQ ID NO.176.

[0073] It is further preferred that the first binding domain capable of binding to an epitope of the human and non-chimpanzee primate CD3e chain comprises a VL region selected from the group consisting of a VL region as set forth in SEQ ID NO. 35, 39, 125, 129, 161 or 165.

[0074] Alternatively, it is preferred that the first binding domain capable of binding to a human and non-chimpanzee epitope of the CD3e chain comprises a VH region selected from the group consisting of VH regions as set forth in SEQ ID NO. 15,19, 33, 37, 51, 55, 69, 73, 87, 91, 105, 109, 123, 127, 141, 145, 159, 163, 177 or 181.

[0075] Even more preferably, the polypeptide of the invention is characterized by a first binding domain capable of binding to an epitope of the CD3e chain of a human and non-chimpanzee primate comprising a VL region and a VH region selected from the group consisting of: 1. (a) a VL region as set forth in SEQ ID NO. 17 or 21 and the VH region as specified in

SEO.IDBR. 15 or 19;

2. (b) VL region as set forth in SEQ ID NO. 35 or 39 and the VH region as specified in

SEQ ID NO: 33 or 37;

3. (c) the VL region as set forth in SEQ ID NO. 53 or 57 and the VH region as specified in

SEQ ID NO: 51 or 55;

4. (d) VL region as set forth in SEQ ID NO. 71 or 75 and the VH region as specified in

SEQ ID NO: 69 or 73;

5. (e) VL region as set forth in SEQ ID NO. 89 or 93 and the VH region as specified in

SEOID No. 87 or 91;

6. (f) the VL region as set forth in SEQ ID NO. 107 or 111 and the VH region as specified in

SEOID No. 105 or 109;

7. (g) VL region as set forth in SEQ ID NO. 125 or 129 and the VH region as specified in

SEQ ID NO: 123 or 127;

8. (h) VL region as set forth in SEQ ID NO. 143 or 147 and the VH region as specified in

SEOID No. 141 or 145;

9. (i) VL region as set forth in SEQ ID NO. 161 or 165 and the VH region as specified in

SEOIDBR. 159 or 163; and

10. (j) VL region as set forth in SEQ ID NO. 179 or 183 and the VH region as specified in SEOIDBR. 177 or 181.

[0076] According to a preferred embodiment of the polypeptide of the invention, the VH-region and VL-region pairs are in single chain antibody (scFv) format. The VH and VL regions are classified in the order VH-VL or VL-VH. Preferably, the VH-region is located N-terminal to the junction sequence. The VL-region is located C-terminal to the junction sequence.

[0077] A preferred embodiment of the previously described polypeptide of the invention is characterized by a first binding domain capable of binding to the epitope of the human and non-chimpanzee primate CD3e chain comprising an amino acid sequence selected from the group consisting of SEQ ID NOs: 23, 25, 41,43, 59,61, 77, 79, 95, 97, 113, 115, 131, 133, 149, 151, 167, 169, 185 or 187.

[0078] The invention further relates to the previously described polypeptide, where the second binding domain binds to a cell surface antigen, which is primarily a tumor antigen.

[0079] The term "tumor antigen" as used herein can be understood to mean those antigens that appear on tumor cells. These antigens can appear on the cell surface with an extracellular part that is often combined with a transmembrane and cytoplasmic part of the molecule. These antigens can sometimes appear only on tumor cells and never on normal cells. Tumor antigens may be exclusively expressed on tumor cells or may represent a tumor-specific mutation compared to normal cells. In this case, they are called tumor-specific antigens. Antigens that are presented by tumor cells and normal cells are more common, and are called tumor-associated antigens. These tumor-associated antigens may be overexpressed compared to normal cells or are available for antibody binding in tumor cells due to the less compact structure of tumor tissue compared to normal tissue. Non-limiting examples of tumor antigens, as used herein, are EGFR (Liu, Br. J. Cancer 82/12 (2000), 1991 -1999; Bonner, Semin. Radiat. Oncol. 12 (2002), 11-20; Kivota, Oncology 63/1 (2002), 92 - 98; Kuan, Brain Tumor Pathol. 17/2 (2000), 71 - 78).

[0080] EGFR (also known as c-erb1 or HER1) belongs to the erbB family of receptor tyrosine kinases. When activated by binding of ligands from the EGF family of growth factors, EGFR homodimerizes or heterodimerizes with another EGFR or with another member of the erbB family, initiating a signaling cascade through mitogen-activated protein kinases and other transcription factors leading to proliferation, differentiation, and renewal (Olayioye, EMBO J. 19 (2000), 3159 - 67). EGFR is overexpressed in many epithelial cancers, including colorectal, breast, lung, head and neck cancers (Mendelsohn, J. Clin. Oncol. 21 (2003), 2787 - 99; Mendelsohn, J. Clin. Oncol. 20 (18, Suppl.) (2002), 1S - 13S; Prewett, Clin. Cancer Res. 88

(2002), 994-1003). Overexpression and/or mutation of EGFR in malignant cells leads to constitutive activation of the kinase, resulting in proliferation, angiogenesis, invasion, metastasis and inhibition of apoptosis (Mendelsohn (2003, loc. cit.; Ciardiello, Clin. Cancer Res. 7

(2001), 2958 - 70; Perez-Soler, Oncologist 9 (2004), 58 - 67). Monoclonal antibodies targeting the extracellular ligand binding domain or the intracellular tyrosine kinase signaling cascade of EGFR have shown efficacy as an antitumor target (Laskin, Cancer Treat. Reviewv 30

(2004), 1 - 17). For example, cetuximab (Erbitux), a humanized monoclonal antibody to EGFR, which competitively inhibits the extracellular domain of EGFR to inhibit ligand activation of the receptor, was approved by the Food and Drug Administration (FDA) in 2004 for the treatment of metastatic colon cancer in combination with the topoisomerase inhibitor irinotecan.

[0081] In a preferred embodiment of the invention, the polypeptide is a bispecific single-chain antibody molecule. The problems described here regarding the development of surrogate molecules for preclinical testing are further compounded if the drug candidate is a bispecific antibody, eg, a bispecific single-chain antibody. Such a bispecific antibody requires that both recognized antigens are interspecies specific for the specified animal species, in order to allow drug safety testing in such animals.

[0082] As stated hereinbefore, the subject invention provides polypeptides comprising a first binding domain that is an antibody capable of binding to an epitope of the CD3e chain of a human and non-chimpanzee primate and a second binding domain capable of binding to a cell surface antigen selected from EGFR, Her2/neu or IgE, wherein the second binding domain preferably also binds to a human and non-chimpanzee cell surface antigen. The advantage of a bispecific single-chain antibody as a drug candidate, which meets the requirements of the preferred polypeptide of the invention, is the use of such a molecule in preclinical testing on animals, as well as in clinical trials and even for therapy in humans. In the interspecies-specific bispecific molecule according to the invention, the binding domain capable of binding to the CD3 epsilon chain epitope of human and non-chimpanzee primates is located in the order VH-VL or VL-VH at the N-terminus or C-terminus of the bispecific molecule. Examples of interspecies-specific bispecific molecules according to the invention in different arrangements of the VH- and VL-chains of the first and second binding domains are described in the attached examples.

[0083] As used herein, "bispecific single chain antibody" means a single polypeptide chain containing two binding domains. Each binding domain contains one antibody heavy chain variable domain ("VH region"), wherein the VH region of the first binding domain specifically binds to a CD3e molecule, and the VH region of the second binding domain specifically binds to a cell surface antigen, as further defined in the text that follows. The two binding domains are optionally linked to each other by a short polypeptide spacer. A non-limiting example of a polypeptide spacer is Gly-Gly-Gly-Gly-Ser (G-G-G-S) and repeats thereof. Each binding domain may further comprise one antibody light chain variable domain ("VL region"), wherein the VH region and the VL region within both the first and second binding domains are connected to each other by a polypeptide linker for example, of the type described and protected in EP 623679 B1, but in any case of sufficient length to allow the VH region and the VL region of the first binding domain and the VH region and the VL region of the second binding domain to pair with each other such that, together, able to bind specifically to the corresponding first and second molecules.

[0084] According to a preferred embodiment of the invention, the previously characterized bispecific single-chain antibody molecule contains a group of the following sequences such as CDR H1, CDR H2, CDR H3, CDR L1, CDR L2 and CDR L3 in the second binding domain selected from SEQ ID NO: 441 - 446, SEQ ID NO: 453 - 458, SEQ ID NO: 463 - 468, SEQ ID NO: 481 - 486.

[0085] A particularly preferred embodiment of the invention relates to the previously defined polypeptide, where the bispecific single-chain antibody molecule contains sequences selected from: 1. (a) amino acid sequences as set forth in any of SEQ ID NOs:389, 391, 393, 395, 397, 399, 409, 411, 413, 415, 417, 419, 429, 431, 433, 435, 437, 439, 447, 449, 451, 469, 471, 473, 475, 477, 479, 495, 497, 499, 501, 503 and 505; and 2. (b) the amino acid sequences encoded by the nucleic acid sequence as set forth in any of SEQ ID NOs: 390, 392, 394, 396, 398, 400, 410, 412, 414, 416, 418, 420, 430, 432, 434, 436, 438, 440, 448, 450, 452, 470, 472, 474, 476, 478, 480, 496, 498, 500, 502, 504 and 506.

[0086] In a preferred embodiment of the invention, the bispecific single chain antibody is interspecies specific for CD3 epsilon and for a tumor antigen recognized by its second binding domain.

[0087] In an alternative embodiment, the present invention provides a nucleic acid sequence encoding the previously described polypeptide of the invention.

[0088] The present invention also relates to a vector containing the nucleic acid molecule of the present invention.

[0089] Many suitable vectors are known to one of ordinary skill in molecular biology, the choice of which will depend on the desired function and includes plasmids, cosmids, viruses, bacteriophages and other vectors commonly used in genetic engineering. Methods well known to one of ordinary skill can be used to construct various plasmids and vectors; see, for example, the techniques described in Sambrook et al. (loc cit.) and Ausubel, Current Protocols in Molecular Biology, Green Publishing Associates and Wiley Interscience, N.Y. (1989), (1994). Alternatively, polynucleotides and vectors of the invention can be reconstituted into liposomes for delivery to target cells. As will be discussed in detail later, a cloning vector was used to isolate individual DNA sequences. Relevant sequences can be transferred into expression vectors when expression of a particular polypeptide is required. Typical cloning vectors include pBluescript SK, pGEM, pUC9, pBR322 and pGBT9. Typical expression vectors include pTRE, pCAL-n-EK, pESP-1, pOP13CAT. Preferably, said vectors contain a nucleic acid sequence that is a regulatory sequence operably linked to a nucleic acid sequence defined herein.

[0090] The term "regulatory sequence" refers to DNA sequences that are required to influence the expression of the coding sequences to which they are ligated. The nature of such control sequences varies with the host organism. In prokaryotes, control sequences generally include the promoter, ribosome binding site, and terminators. In eukaryotes, control sequences generally include promoters, terminators and, in some cases, enhancers, transactivators or transcription factors. The term "control sequence" is intended to include, at a minimum, all components whose presence is required for expression, and may also include additional suitable components.

[0091] The term "operably linked" refers to positions where the components described above are in a relationship that allows them to function in the manner intended for them. The control sequence is "operably linked" to the coding sequence in such a way that expression is achieved under conditions compatible with the control sequence. In the case where the control sequence is a promoter, it is obvious to one of ordinary skill in the art that the nucleic acid is primarily used.

[0092] Thus, said vector is primarily an expression vector. An "expression vector" is a construct that can be used to transform a selected host and provide expression of a coding sequence therein. Expression vectors can be, for example, cloning vectors, binary vectors or integration vectors. Expression involves the transcription of nucleic acid molecules primarily into translatable mRNA. Regulatory elements that ensure expression in prokaryotic and/or eukaryotic cells are well known to one of ordinary skill in the art. When it comes to eukaryotic cells, they contain normal promoters that ensure transcription initiation and optionally poly-A signals that ensure transcription termination and transcript stabilization. Possible regulatory elements that enable expression in prokaryotic host cells include, nor., the PL,lac, trbili tac promoter in E. coli, and examples of regulatory elements that enable expression in eukaryotic host cells are the AOX1 or GAL1 promoter in yeast or the CMV-, SV40-, RSV-promoter (Rous sarcoma virus), CMV-enhancer, SV40-enhancer or globin intron in mammalian and other animal cells.

[0093] In addition to the element responsible for transcription, such regulatory elements may also contain transcription termination signals, such as the SV40-poly-A site or the tk-poly-A site, downstream of the polynucleotide. In addition, depending on the expression system, used leader sequences capable of directing the polypeptide to the cellular compartment or secreting it into the medium can be added to the coding sequence of said nucleic acid sequence and are well known in the art, see also the attached Examples. One or more leader sequences are assembled in an appropriate phase from the sequence for translation, initiation and termination, and preferably, a leader sequence that is able to direct the secretion of the synthesized protein, or part thereof, into the periplasmic space or the extracellular medium. Optionally, the heterologous sequence may encode a fusion protein including an identification peptide at the end that confers desired characteristics, eg, stabilization or reduced purification of the expressed recombinant product; see supra. In this connection, suitable expression vectors are known in the art, such as Okayama-Berg cDNA expression vector pcDV1 (Pharmacia), pCDM8, pRc/CMV, pcDNAl, pcDNA3 (In-vitrogene), pEF-DHFR, pEF-ADA or pEF-neo (Mack et al. PNAS (1995) 92, 7021 - 7025 and Raum et al. Cancer Immunol Immunother (2001) 50(3), 141 -150) or pSPORTI (GIBCO BRL).

[0094] Preferably, the expression control sequences will be eukaryotic promoter systems in vectors capable of transforming or transfecting eukaryotic host cells, but control sequences for prokaryotic hosts can also be used. Once the vector has been incorporated into a suitable host, it is maintained under conditions suitable for high level expression of the nucleotide sequences and, if desired, collection and purification of the polypeptides of the invention can follow, see eg the attached examples.

[0095] An alternative expression system that can be used to express a cell cycle interacting protein is the insect system. In one such system, Autographa californica nuclear polyhedrosis virus (AcNPV) was used as a vector for the expression of foreign genes in Spodoptera frugiperda cells or Trichoplusial larvae. The coding sequence of said nucleic acid molecule can be cloned into a nonessential region of a virus, such as a polyhedrin gene, and placed under the control of a polyhedrin promoter. Successful insertion of said coding sequence will keep the polyhedrin gene inactive and produce a recombinant virus lacking a protein coat. Recombinant viruses were then used to infect S. frugiperdacelli or Trichoplusial larvae expressing the protein of the invention (Smith, J. Virol. 46 (1983), 584; Engelhard, Proc. Nat. Acad. Sci. USA 91 (1994), 3224-3227).

[0096] Additional regulatory elements may include transcriptional as well as translational enhancers. Conveniently, the translocation vectors described earlier contain selectable and detectable markers.

[0097] Gene markers useful for selecting transformed cells and, e.g., plant tissues and plants are well known to the person of ordinary skill in the art and include, for example, antimetabolite resistance as a basis of selection for dhfr, which confers resistance to methotrexate (Reiss, Plant Phvsiol. (Life Sci. Adv.) 13

(1994), 143-149); npt, which confers resistance to the aminoglycoside neomycin, kanamycin and paromycin (Herrera-Estrella, EMBO J. 2 (1983), 987-995) and hygro, which confers resistance to hygromycin (Marsh, Gene 32 (1984), 481-485). Additional selectable genes have been described, namely trpB, which allows cells to use indole instead of tryptophan; hisD, which allows cells to use histinol instead of histidine (Hartman, Proc. Natl. Acad. Sci. USA 85 (1988), 8047); mannose-6-phosphate isomerase which enables cells to utilize mannose (WO 94/20627) and ODC (ornithine decarboxylase) which confers resistance to the ornithine decarboxylase inhibitor, 2-(difluoromethyl)-DL-omithin, DFMO (McConlogue, 1987, In: Current Communications in Molecular Biology, Cold Spring Harbor Laboratory ed.) or deaminase from Aspergillus terreus which confers resistance.

Blasticidin-u S (Tamura, Biosci. Biotechnol. Biochem. 59 (1995), 2336 - 2338).

[0098] Useful detectable markers are also known to those of ordinary skill in the art and are commercially available. Advantageously, said marker gene is one that encodes luciferase (Giacomin, Pl. Sci. 116 (1996), 59-72; Scikantha, J. Bact. 178 (1996), 121), green fluorescent protein (Gerdes, FEBS Lett. 389 (1996), 44-47) or fi-glucuronidase (Jefferson, EMBO J. 66). (1987), 3901 - 3907). This embodiment is particularly useful for simple and rapid screening of cells, tissues and organisms containing said vector.

[0099] As described earlier, said nucleic acid molecule can be used alone or as part of a vector to express the polypeptide of the invention in cells, for example for purification, but also for the purpose of gene therapy. Nucleic acid molecules or vectors containing one or more DNA sequences encoding any previously described polypeptide of the invention are introduced into cells that in turn produce the polypeptide of interest. Gene therapy, which is based on the introduction of therapeutic genes into cells by ex-vivo or in-vivo techniques, is one of the most important applications of gene transfer. Suitable vectors, methods or systems for gene delivery for in-vitro or in-vivo gene therapy are described in the literature and known to one of ordinary skill in the art, see, e.g. Giordano, Nature Medicine 2 (1996), 534-539; Schaper, Circ. Res. 79 (1996), 911 - 919; Anderson, Science 256 (1992), 808-813; Verma, Nature 389 (1994), 239; Isner, Lancet 348 (1996), 370-374; Muhlhauser, Circ. Res. 77 (1995), 1077-1086; Onodera, Blood 91 (1998), 30-36; Verma, Gene Ther. 5

(1998), 692 - 699; Nabel, Ann. N.Y. Acad. Sci. 811 (1997), 289 - 292; Verzeletti, Hum. Gene Ther. 9 (1998), 2243 - 51; Wang, Nature Medicine 2 (1996), 714-716; WO 94/29469; WO 97/00957, US 5,580,859; US 5,589,466; or Schaper, Current Opinion in Biotechnology 7 (1996), 63 56 - 0. Said nucleic acid molecules and vectors can be designed for direct introduction or for introduction into the cell via liposomes, or viral vectors (eg, adenoviral, retroviral). Preferably said cell is a germ cell line, an embryonic cell, or an egg cell or derived from these, most preferably said cell is a stem cell. An example of an embryonic stem cell may be, inter alia, the stem cell described in Nagy, Proc. Mati. Acad. Sci. USA 90 (1993), 8424 - 8428.

[0100] The invention also provides as a host a cell transformed or transfected with a vector of the invention. Such a cell can be obtained by introducing the previously described vector of the invention or the previously described nucleic acid molecule of the invention into a host cell. The presence of at least one vector or at least one nucleic acid molecule in a host cell can mediate expression of the gene encoding the previously described single-cell antibody construct

[0101] The described nucleic acid molecule or vector of the invention, which has been introduced into a host cell can either be integrated into the genome of the host cell or can be maintained extrachromosomally.

[0102] The host cell can be any prokaryotic or eukaryotic cell.

[0103] The term "prokaryote" is intended to include all bacteria, which can be transformed or transfected with DNA or RNA molecules to express the proteins of the invention. Prokaryotic host cells may include gram-negative as well as gram-positive bacteria such as, for example, E. coli, S. tvphimurium, Serratia marcescens, and Bacillus subtilis. The term "eukaryotic" is intended to include yeast, higher plants, insects, and primarily mammalian cells. Depending on the host cell used in the recombinant production procedure, the protein encoded by the polynucleotide of the present invention may be glycosylated or may be non-glycosylated. The use of a plasmid or virus containing the coding sequences of the polypeptide of the invention and genetically fused to an N-terminal FLAG-tag and/or a C-terminal His-tag is particularly preferred. Preferably, the length of said FLAG-tag is about 4 to 8 amino acids, most preferably 8 amino acids. The previously described polynucleotide can be used to transform or transfect host cells using any of the techniques commonly known to one of ordinary skill in the art. Furthermore, methods for obtaining fused, operably linked genes and expressing them in, e.g. mammalian cells and bacteria are well known in the art (Sambrook, loc cit.).

[0104] Preferably, the host cell is a bacterium or an insect, fungus, plant or animal cell.

[0105] It is especially envisaged that said host cell may be a mammalian cell. Particularly preferred host cells include CHO cells, COS cells, myeloma cell lines such as SP2/0 or NS/0. As illustrated in the appended claims, particularly preferred host cells are CHO cells.

[0106] A more preferred host cell is a human cell or human cell line, eg c6 (Kroos, Biotechnol. Prog., 2003, 19:163-168).

[0107] In a further embodiment, the subject invention thus relates to a process for obtaining the polypeptide of the invention, wherein said process comprises growing the cell of the host of the invention under conditions that enable the expression of the polypeptide of the invention and the recovery of the obtained polypeptide from the culture.

[0108] Transformed host cells can be grown in fermentors and cultured in accordance with techniques known in the art to achieve optimal cell size. The polypeptide of the invention can then be isolated from the growth medium of cell lysates or cell membrane fractions. Isolation and purification of e.g. polypeptides of the invention expressed in microorganisms can be performed by any conventional means, such as, for example, preparative chromatographic separation and immunological separation, such as that involving the use of monoclonal or polyclonal antibodies directed, e.g. according to the polypeptide tag of the invention or as described in the attached examples.

[0109] It is known in the art that the host cell culture conditions that enable expression depend on the host cell system and the expression system/vector used in such a process. The parameters that need to be modified to achieve conditions that allow expression of the recombinant polypeptide are known in the art. Thus, suitable conditions can be determined by one of ordinary skill in the absence of additional inventive challenges.

[0110] Once expressed, the polypeptide of the invention can be purified according to standard procedures of the art, including ammonium sulfate precipitation, affinity columns, column chromatography, gel electrophoresis in si. see Scopes, "Protein Purification", Springer-Verlag, N.Y.

(1982). For pharmaceutical use, substantially pure polypeptides of at least about 90 to 95% homogeneity are preferred, and 98 to 99% homogeneity or more are most preferred. Once purified, partially or to the desired homogeneity, the polypeptide of the invention can then be used therapeutically (including extracorporeally) or in experimental development and performance procedures. In addition, exemplary methods for isolating polypeptides of the invention from culture are described in detail in the accompanying examples.

[0111] In addition, the<p>invention provides a composition comprising a polypeptide of the invention or a polypeptide produced by the previously described process. Preferably, said composition is a pharmaceutical composition.

[0112] In accordance with the invention, the term "pharmaceutical composition" refers to a composition for administration to a patient, preferably a human. A particularly preferred pharmaceutical composition of the present invention comprises binding molecules directed to and generated by context-independent CD3 epitopes. Preferably, the pharmaceutical composition contains suitable formulations of carriers, stabilizers and/or excipients. In a preferred embodiment, the pharmaceutical composition comprises a composition for parenteral, transdermal, intraluminal, intraarterial, intrathecal and/or intranasal administration or by direct injection into tissue. In particular, administration of said composition to the patient by infusion or injection is envisaged. Administration of suitable com<p>ositions can be performed in different ways, e.g. intravenously, intraperitoneally, subcutaneously, intramuscularly, topically or intradermally. In particular, the present invention provides compositions suitable for continuous administration. As a non-limiting example, continuous, ie. continuous administration can be realized with a small pump system, worn by the patient, to measure the influx of the therapeutic agent into the patient's body. A pharmaceutical composition containing binding molecules directed against and generated against a context-independent CD3 epitope of the invention can be administered using said pump system. Such pump systems are generally known in the art, and typically relate to the periodic change of cartridges containing a therapeutic agent to be infused. When replacing a cartridge in such a pump system, a temporary interruption of the otherwise continuous flow of therapeutic agent into the patient's body may occur. In such a case, the administration phase before cartridge replacement and the administration phase after cartridge replacement together within the pharmaceutical meaning and method of the invention together constitute one "continuous administration" of such therapeutic agent.

[0113] Continuous or continuous administration of these binding molecules directed to and generated by the context-independent CD3 epitope of the present invention can be intravenous or subcutaneous via a liquid delivery device or a small pump system that includes a mechanism for driving fluid from a reservoir and a mechanism for driving this mechanism. Pump systems for subcutaneous administration may include a needle or cannula to pierce the patient's skin and deliver a suitable composition into the patient's body. These pump systems can be attached or attached to the patient's skin independent of a vein, artery, or blood vessel, allowing direct contact between the pump system and the patient's skin. The pump system can be attached to the patient's skin for 24 hours to several days. The pump system can be small in size with a small volume reservoir. As a non-limiting example, the reservoir volume for a suitable pharmaceutical composition to be administered may be between 0.1 and 50 ml.

[0114] Continuous administration can be transdermal via a patch on the skin, which is replaced at intervals. One of ordinary skill in the art is familiar with drug delivery patch systems suitable for this purpose. It should be noted that transdermal administration is particularly suitable for continuous administration, since the replacement of a worn patch can be performed simultaneously with the placement of a new, second patch, for example, on the surface of the skin right next to the first worn patch and immediately before removing the first worn patch. There will be no interruption of flow or loss of cell power. [01^5? A composition of the present invention comprising binding molecules directed to and directed to a context-independent CD3 epitope may additionally comprise a pharmaceutically acceptable carrier. Examples of suitable pharmaceutical carriers are well known to those of ordinary skill in the art and include solutions, e.g. phosphate buffered saline solutions, water, emulsions such as water/oil emulsions, various types of wetting agents, sterile solutions, liposomes, etc. Compositions containing such carriers can be formulated by well known conventional methods. Formulations may contain hydrocarbons, buffer solutions, amino acids and/or surfactants. The hydrocarbons can be non-reducing sugars, preferably trehalose, sucrose, octasulfate, sorbitol or xylitol. Such formulations can be used for continuous administration, which can be intravenous or subcutaneous with and/or without a pump system. Amino acids can be charged amino acids, preferably lysine, lysine acetate, arginine, glutamate and/or histidine. Surfactants can be detergents, preferably with a molecular weight of >1.2 KD and/or polyether, preferably with a molecular weight of >3 KD. Non-limiting examples of preferred detergents are Tween 20, Tween 40, Tween 60, Tween 80, or Tween 85. Non-limiting examples of preferred polyethers are PEG 3000, PEG 3350, PEG 4000, or PEG 5000. Buffer systems used in the commercial assay may have a preferred pH of 5-9 and may contain citrate, succinate, phosphate, histidine and acetate. The compositions of the subject pronal can be administered to a subject at a suitable dose which can be determined e.g. by examining the effects of increasing doses, by administering increasing doses of a polypeptide of the invention that exhibits the interspecies specificity described herein to non-chimpanzee primates, for example macaques. As previously noted, a polypeptide of the invention exhibiting the interspecies specificity described herein may conveniently be used in identical form in non-chimpanzee non-human primate preclinical testing and as a human drug. These compositions can also be administered with other protein and non-protein drugs. These drugs can be administered simultaneously with compositions containing a polypeptide of the invention as defined herein, or separately before or after administration of said polypeptide at timed intervals and doses. The dosage regimen will be determined by the attending physician and clinical factors. As is well known in the medical field, dosages for any given patient depend on many factors, including the patient's size, body surface area, age, the particular compound to be administered, gender, time and method of administration, general health, and other concomitant medications. Preparations for parenteral administration include sterile aqueous and non-aqueous solutions, suspensions, and emulsions. Examples of non-aqueous solutions are propylene glycol, polyethylene glycol, vegetable oils such as olive oil, and injectable organic esters such as ethyl oleate. Aqueous vehicles include water, alcoholic/aqueous solutions, emulsions or suspensions, including saline and buffered media Parenteral vehicles include sodium chloride solutions, Ringer's dextrose, dextrose and sodium chloride, and Ringer's lactate or certain oils. Intravenous vehicles include fluids and nutrient supplements, electrolyte supplements (such as those based on Ringer's dextrose), and the like. Preservatives and other additives may also be present, such as, for example, antimicrobials, antioxidants, chelating agents, inert gases, and the like. In addition, the composition of the present invention may contain protein carriers, such as, for example, albumin serum or immunoglobulin, preferably of human origin. It is envisaged that the composition of the invention may contain, in addition to the polypeptide of the invention defined here, additional biologically active agents, depending on the intended use of the composition. Such agents can be drugs that act on the gastrointestinal system, drugs that act as cytostatics, drugs that prevent hyperuricemia, drugs that inhibit immunoreactions (eg corticosteroids), drugs that modulate the inflammatory response, drugs that act on the circulatory system and/or agents such as cytokines known in the art.

[0116] The biological activity of the pharmaceutical compositions defined herein can be determined, for example, by cytotoxicity assays, as described in the following examples, in WO 99/54440 or Schlereth et al. (Cancer Immunol. Immunother. 20 (2005), 1 - 12). "Efficacy" or "in vivo efficacy" as used herein refers to response to therapy with a pharmaceutical composition of the invention, using, e.g., standardized NCI response criteria. The success or effectiveness of vivotherapy using the pharmaceutical composition of the invention refers to the effectiveness of the composition for the intended purpose, ie the ability of the composition to cause the desired effect, ie. destruction of pathogenic cells, eg tumor cells. Effectiveness in vivo can be controlled by established standard methods for appropriate disease entities including, without limitation, white blood cell count, differential cell count, fluorescence activated cell sorting, bone marrow aspiration. In addition, specific clinical hernia parameters of various diseases and other established standard methods can be used. Furthermore, computed tomography, X-ray, nuclear magnetic resonance imaging (eg, determination of response based on National Cancer Institute criteria [Cheson BD, Horning SJ, Coiffier B, Shipp MA, Fisher Rl, Connors JM, Lister TA, Vose J, Grillo-l.opez A, Hagenbeek A, Cabanillas F, Klippensten D. Hiddemann W, Castellino R, Harris NL, Armitage JO, Charter W, Hoppe R, Canellos GP. Report of an international workshop for non-Hodgkin's Ivmphomas. J Clin Oncol. 1999: 1244], white blood cell count, fluorescence activated cell sorting, bone marrow aspiration. lymph nodes, and various parameters lymphoma-specific clinical assays (eg lactate dehydrogenase) and other established standard methods.

[0117] Another major challenge in drug development, such as the pharmaceutical compositions of the invention, is the predictable modulation of pharmacokinetic properties. So far, the pharmacokinetic profile of the candidate drug has been established, i.e. the profile of pharmacokinetic parameters that influence the ability of a particular drug to treat the specified condition. Pharmacokinetic parameters of a drug that affect the drug's ability to treat certain diseases include, without limitation: half-life, volume of distribution, hepatic presystemic metabolism, and degree of blood serum binding. The effectiveness of the given medicinal product can be influenced by each of the mentioned parameters.

[0118] "Half-life" means the time for which 50% of the administered drug is eliminated by biological processes, e.g. metabolism, excretion, etc.

[0119] "Hepatic first-pass metabolism" means the tendency of the drug to be metabolized upon first contact with the liver, i.e. during its first passage through the liver. "Extent of distribution" means the degree of retention of the drug in various parts of the body, such as, for example, intracellular and extracellular spaces, tissues and organs, etc. and the distribution of the drug in these parts.

[0120] "Degree of blood serum binding" refers to the tendency of a drug to react with and bind to blood serum proteins, such as albumin, resulting in a reduction or loss of the drug's biological activity.

[0121] Pharmacokinetic parameters also include bioavailability, latency (Tlag), Tmax, absorption rates, and even onset and/or Cmax for a given amount of drug administered. "Bioavailability" refers to the amount of a drug in parts of the blood.

[0122] "Delay" refers to the delay in time between the administration of a drug and its detection and measurement in blood or plasma.

[0123] "Tmax" is the time after which the maximum blood concentration of the drug is reached, and "Cmax" is the maximum blood concentration obtained by the given drug. The time required to reach the concentration of the drug in the blood or tissue for its biological effect is affected by all parameters. Pharmacokinetic parameters of bispecific single-chain antibodies, a preferred embodiment of the polypeptide of the invention exhibiting interspecies specificity, which can be determined in preclinical testing in animals other than chimpanzees, as shown in the preceding text, are also given e.g. in the publication of Schlereth et al. (Cancer Immunol. Immunother. 20 (2005), 1 -12). [01241 The term "toxicity" as used herein refers to the toxic effects of a drug manifested in adverse events or serious adverse events. These side effects may refer to loss of tolerance to the drug in general and/or loss of local tolerance after administration. Toxicity may also include drug-induced teratogenic or carcinogenic effects.

[0125] The terms "safety", "in vivo safety" or "tolerability" as used herein define the administration of a drug without causing serious adverse events after administration (local tolerability) and over a long period of drug administration. "Safety", "in vivo safety" or "tolerability" can be assessed eg at regular intervals during treatment and during the follow-up period. Measurements include clinical assessment, eg. organ manifestations, and scanning laboratory abnormalities. Clinical assessment can be performed and deviation from normal findings determined/coded according to NCI-CTC and/or MedDRA standards can be determined. Organ manifestations may include criteria such as allergy/immunology, blood/bone marrow, cardiac arrhythmia, coagulation, and the like, as noted e.g. in Common Terminology Criteria for adverse events v3.0 (CTCAE). Laboratory parameters that may be tested include, for example, hematology, clinical hernia, coagulation profile, and urinalysis and examination of other body fluids such as serum, plasma, lymphoid or spinal fluid, cerebrospinal fluid, and the like. Safety can be assessed in this way, e.g. physical examination, imaging techniques (i.e., ultrasound, X-ray, CT scan, magnetic resonance imaging (MRI), other measurements with technical devices (i.e., electrocardiogram), vital signs, measurement of laboratory parameters and determination of adverse events. For example, adverse events in primates other than chimpanzees when using and methods according to the invention can be examined by histopathological and/or histochemical methods.

[0126] The term "effective and non-toxic dose" as used herein refers to a tolerable dose of a bispecific single chain antibody as defined herein, which is high enough to cause depletion of pathological cells, tumor elimination, tumor shrinkage, or disease stabilization without, or substantially without, major adverse effects. Such effective and non-toxic doses can be determined e.g. dose increase trials that are described in the literature and should be below the dose that causes serious adverse side effects (dose limiting toxicity, DLT, from English dose limiting toxicity). [01271 The aforementioned terms are discussed in eg Preclinical safety evaluation of biotechnology-derived pharmaceuticals S6; ICH Harmonized Tripartite Guideline; ICH Steering Committee meeting on July 16, 1997.

[0128] Furthermore, the invention relates to pharmaceutical compositions containing a polypeptide of the present invention (i.e., a polypeptide containing at least one binding domain capable of binding to an epitope of the CD3 epsilon chain of a human and a non-chimpanzee primate, wherein the epitope is part of an amino acid sequence found in a group consisting of SEQ ID No. 2, 4, 6, or 8 in accordance with this product or produced by a process according to this invention for the prevention, treatment, and amelioration of diseases selected from proliferative disease, tumor disease or immune disorder Preferably, said pharmaceutical composition additionally contains suitable formulations of carriers, stabilizers and/or excipients.

[0129] Another aspect of the invention relates to the use of a polypeptide as defined here or produced according to the process defined here, for the preparation of a pharmaceutical composition for the prevention, treatment or amelioration of diseases. Primarily, said disease is a proliferative disease, a tumor disease or an immune disorder. It is further preferred that the said tumor disease is a malignant disease, primarily cancer.

[0130] In another preferred embodiment of the use of the polypeptide of the invention, the said pharmaceutical composition is suitable for administration in combination with an additional drug, i.e. as part of co-therapy. In such co-therapy, the active agent may optionally be included in the same pharmaceutical composition as the polypeptide of the invention, or may be included as a separate pharmaceutical composition. In the latter case, said particular pharmaceutical composition is suitable for administration before, simultaneously with, or after the administration of the pharmaceutical composition containing the polypeptide of the invention. The additional drug or pharmaceutical composition can be a non-protein compound or a protein compound. In the case where the additional drug is a protein compound, it is convenient for the protein compound to be able to provide an activation signal for effector cells of the immune system.

[0131] Preferably, said proteinaceous compound or non-proteinaceous compound may be administered simultaneously or non-concurrently with a polypeptide of the invention, a nucleic acid as hereinbefore defined, a vector as hereinbefore defined, or a host as hereinbefore defined.

[0132] Another aspect of the invention relates to a pharmaceutical composition of the invention for use in a method for treating or ameliorating a disease of a subject in need thereof, wherein said method comprises the step of administering an effective amount of a pharmaceutical composition of the invention. Primarily, said disease is a disease is a proliferative disease, a tumor disease or an immunological disorder. It is even more preferable that the said tumor disease is a malignant disease, primarily cancer.

[0133] In another preferred embodiment of the above, the pharmaceutical composition is suitable for administration in combination with an additional drug, i.e. as part of co-therapy. In such co-therapy, the active agent may optionally be included in the same pharmaceutical composition as the polypeptide of the invention, or may be included as a separate pharmaceutical composition. In the latter case, said particular pharmaceutical composition is suitable for administration before, simultaneously with, or after the administration of the pharmaceutical composition containing the polypeptide of the invention. The additional drug or pharmaceutical composition can be a non-protein compound or a protein compound. In the case where the additional drug is a protein compound, it is convenient that the protein compound is capable of providing an activation signal for effector cells of the immune system.

[0134] Preferably, said proteinaceous compound or non-proteinaceous compound may be administered simultaneously or non-concurrently with a polypeptide of the invention, a nucleic acid as hereinbefore defined, a vector as hereinbefore defined, or a host as hereinbefore defined.

[0135] It is preferable for the previously described method of the invention that the said subject is a human.

[0136] In another aspect, the invention relates to a kit comprising a polypeptide of the invention, a nucleic acid molecule of the invention, a vector of the invention, or a host of the invention.

[0137J The subject invention further provides the following list of points:

Item 1. A method for identifying one or more polypeptides containing an interspecies-specific binding domain capable of binding to the human and non-chimpanzee CD3 epsilon (CD3e) epitope, the method comprising the steps of: 1. (a) contacting one or more polypeptides with an N-terminal fragment of the CD3e extracellular domain of a maximum of 27 amino acids comprising the amino acid sequence Gln-Asp-Gly-Asn-Glu-Glu-Met-Gly (SEQ ID NO. 381) or Gln-Asp-Gly-Asn-Glu-Glu-lle-Gly (SEQ ID NO. 382), attached via its C-terminus to the solid phase;

2. (b) elution of the bound polypeptide(s) from this fragment, i

3. (c) isolating one or more polypeptides from the eluate under (b).

Preferably, the polypeptide(s) identified by the method of the invention are of human origin.

This "method for identification of polypeptides" is understood as a method for isolating one or more different polypeptides with the same specificity for a fragment of the extracellular domain of CD3s containing at its N-terminus the amino acid sequence Gln-Asp-Gly-Asn-Glu-Met-Gly (SEQ ID NO. 381) or Gln-Asp-Gly-Asn-Glu-Glu-lle-Gly (SEQ ID NO. 382) from a plurality of candidate polypeptides, as and a method for purifying polypeptides from solution. A non-limiting embodiment of a method for isolating one or more different polypeptides with the same specificity for a CD3e extracellular domain fragment includes methods for selecting antigen-specific binding entities, e.g. panning methods as commonly used for hybridoma screening, screening of transiently/stable transfected clones of eukaryotic host cells or in phage methods. A non-limiting example of this second method for purifying polypeptides from solution is e.g. purifying a recombinantly expressed polypeptide from a culture supernatant or preparing such a culture.

As previously stated, the fragment used in this method is the N-terminal fragment of the extracellular domain of the primate CD3e molecule. The amino acid sequence of the extracellular domain of CD3£ molecules from various primates is set forth in SEQ ID NOs: 1, 3, 5 and 7. Two forms of the N-terminal octamer are listed in SEQ ID NOs: 381 and 382. Preferably, this N-terminus is freely available for binding the polypeptide to be identified by the method of the invention. The term "freely available" is understood in the context of the invention to mean free from additional motifs such as His-tag. The interference of such a His-tag with the binding molecules identified by this method is described in the attached Example 6.

According to the previously mentioned method, the mentioned fragment is attached via its C-terminus to the solid phase. One of ordinary skill in the art will readily and without hesitation select a suitable solid phase support depending on the embodiment of the method of the invention used. Examples of solid supports include, without limitation, matrixes in the form of beads (eg agarose beads, sepharose beads, polystyrene beads, dextran beads), plates (culture plates or MultiVVell plates) as well as chips known e.g. from Biacore<®>. The selection of ways and methods for attaching/immobilizing the fragment to the specified solid substrate depends on the selection of the solid substrate. A commonly used method for attachment/immobilization is coupling via an N-hydroxysuccinimide (NHS) ester. The hernia underlying this coupling as well as alternative methods for attachment/immobilization are known to the person skilled in the art, e.g. from Hermanson "Bioconjugate Techniques", Academic Press, Inc.

(1996). To attach to. and immobilization on chromatographic supports, the following agents are usually used: NHS-activated Sepharose (eg HiTrap-NHS from GE Life Science - Amersham), CnBr-activated Sepharose (eg GE Life Science Amersham), NHS-activated dextran beads (Sigma) or activated polymethacrylate. These reagents can also be used in a one-step procedure. Moreover, dextran beads containing iron oxide (eg, available from Miltenvi) can be used in a one-step process. These beads can be used in combination with a magnet to separate the beads from the solution. Polypeptides can be immobilized on a Biacore chip (eg, CMS chips) using NHS-activated carboxymethyldextran. Further examples of suitable solid supports are amine reactive MultiVVell plates (eg Nunc Immobilizer™ plates). According to the methods mentioned above, the mentioned fragment of the extracellular domain of CD3 epsilon can be directly coupled to a solid substrate or through a series of amino acids, which can be linkers or other protein/polypeptide groups. Alternatively, the extracellular domain of CD3 epsilon may be indirectly coupled via one or more adapter molecules. Ways and methods for elution of a peptide bound to an immobilized epitope are well known in the art, as are methods for extracting the identified polypeptide(s) from the eluate. In accordance with the invention, the method for extracting one or more different polypeptides with the same specificity for the fragment of the extracellular domain of CD3£, which contains at its N-end the amino acid sequence Gln-Asp-Gly-Asn-Glu-Glu-X-Gly from the multitude of polypeptide candidates, can contain one or more steps of the following methods for the selection of antigen-specific entities: CD3e specific binding molecules can be selected from the repertoire of antibodies. a phage display library can be created based on standard procedures, as, for example, described in "Phage Display: A Laboratory Manual"; Ed. Barbas, Burton, Scott &Silverman; Cold Spring Harbor Laboratory Press, 2001. The format of an antibody fragment in an antibody library can be an scFv, but in general it can also be a Fab fragment or even a single domain antibody fragment. Naive antibody fragment libraries can be used to isolate antibody fragments. To select entities with potentially low immunogenic binding for subsequent use in therapeutics, a library of human antibody fragments may be suitable for direct selection of human antibody fragments. In some cases, they can form the basis of synthetic antibody libraries (Knappik et al. J Mol. Biol. 2000, 296:57 ff). A suitable format can be a Fab, scFv (as described in the following text) or an antibody domain (dAbs, reviewed in Ho!t et al., Trends Biotechnol. 2003, 21:484 ff).

It is also known in the art that in many cases there is no available source of immune human antibody to the target antigen. Therefore, animals are immunized with the target antigen and libraries of corresponding antibodies isolated from animal tissue such as e.g. spleen or PBMCs. The N-terminal fragment can be biotinylated or covalently linked to proteins such as KLH or bovine serum albumin (BSA). According to the usual approach, rodents are used for immunization. Some non-human immune antibody repertoires may be particularly suitable for other reasons, e.g. due to the presence of single domain antibodies (VHH) derived from cameloid species (as described in Muvldermans, J Biotechnol. 74:277; De Genst et al. Dev Como Immunol. 2006, 30:187 ff). Therefore, the appropriate antibody library format can be Fab, scFv (as described in the text that follows), or single-domain antibodies (VHHs).

In one possible approach, 10 week old F1 mice from a balb/c x C57black cross can be immunized with ceiim cells e.g. orim that express the N-terminal transmembrane EpCAM N. which occurs as a translational fusion of N-terminal amino acids 1 to 27 of the mature CD3 epsilon chain. Alternatively, mice can be immunized with a 1-27 CD3 epsilon-Fc fusion protein (an appropriate approach is described in the attached Example 2). After booster immunization (immunization), blood samples can be taken and the serum titer of antibodies to CD3-positive T cells can be tested, eg by FACS analysis. Usually, serum titers are significantly higher in immunized than in non-immunized animals. Immunized animals can form the basis for the construction of immune antibody libraries. Examples of such data include libraries with exposed phases. Such libraries can be created based on standard procedures, as described, for example, in "Phage Display: A Laboratory Manual"; Ed. Barbas, Burton, Scott &Silverman; Cold Spring Harbor Laboratory Press, 2001.

Non-human antibodies can also be humanized using exposed phages thanks to the creation of libraries with more diverse antibodies that can then be enriched with binding molecules during selection. In the exposed phage approach, any of the sets of phages that expose antibody libraries form the basis for selecting binding elements of the corresponding antigen as a target molecule. Because of the exposed antibody fragments on the phage surface, this general method is called phage display. One preferred method of selection is the use of small proteins such as the filamentous domain of phage N2 that is translationally fused to the N-terminus of a phage-displayed scFv. Another exposure method known in the art that can be used to isolate binding entities is the ribosome exposure method (reviewed in Groves & Osboum, Expert Opin Biol Ther. 2005, 5:125 ff; Lipovsek & Pluckthun, J Immunol Methods 2004, 290:52 ff).

To demonstrate the binding of scFv phage particles to the 1-27 CD3e-Fc fusion protein, a phage library containing the cloned scFv repertoire can be harvested from an appropriate culture supernatant with PEG (polyethylene glycol). ScFv phage particles can be incubated with an immunolabeled CD3e Fc fusion protein. Immunostained CD3e Fc fusion protein can be applied to the solid phase. Binding entities can be eluted and the eluate used to infect fresh uninfected bacterial hosts. Bacterial hosts successfully transduced with a phagemid copy encoding a human scFv-fragment can be re-selected for carbenicillin resistance and then infected with e.g. VCMS 13 helper phage, to initiate a second round of antibody exposure and in vitro selection. Normally a total of 4 to 5 rounds of selection are performed Binding of isolated binding entities can be tested to CD3epsilon positive Jurkat cells, HPBall cells, PBMCs or transfected eukaryotic cells carrying an N-terminal CD3e sequence fused to surface-exposed EpCAM using flow cytometry (see attached Example 4).

Point 2. The method of point 1, wherein the polypeptide(s) comprise an identified binding domain as a first binding domain and a second binding domain capable of binding to a surface

antigen cells.

To create a second binding domain of the polypeptide identified by the aforementioned method, e.g. of a bispecific single-chain antibody as defined herein, monoclonal antibodies that bind to the cell surface of both human and non-chimpanzee antigens may be used. Suitable binding domains for a bispecific polypeptide as defined herein may, e.g. be derived from interspecies-specific monoclonal antibodies by recombinant methods described in the art. A monoclonal antibody that binds to the human cell surface antigen and to the non-chimpanzee cell surface homologue of the antigen can be tested by FACS assays as previously noted. Hybridoma techniques as described in the literature (Milstein and Kohler, Nature 256 (1975), 495-7) can also be used to generate interspecies specific antibodies. For example, mice can alternatively be immunized with human and non-chimpanzee CD33. From these mice, hybridoma cells producing the interspecies-specific antibody can be isolated by the hybridoma technique and analyzed by FACS as previously stated. The generation and analysis of bispecific polypeptides, such as bispecific single chain antibodies that exhibit interspecies specificity as described herein, are shown in the Examples that follow. Advantages of bispecific single-chain antibodies that exhibit interspecies specificity include the points listed in the text that follows.

Item 3. The method of item 2, wherein the second binding domain binds to the cell surface of a human and/or primate other than chimpanzee antigen.

Item 4. The method of any item 1 to 3, wherein the first binding domain is an antibody.

Item 5. The method of item 4, wherein the antibody is a single-chain antibody.

Item 6. The method of any item 2 to 5, wherein the second binding domain is an antibody.

Item 7. The method of any item 1 to 6, wherein the CD3e extracellular domain fragment comprises one or more polypeptide fragments having an amino acid sequence of any one of those listed in SEQ ID No. 2, 4, 6 or 8.

Dot;. 3. The method from point 7, where said fragment has 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27 amino acid residues.

Item 9. The method of any item 1 to 8, wherein the identification method is a method of screening multiple polypeptides containing an interspecies-specific binding domain capable of binding to a human and non-chimpanzee CD3e epitope.

Item 10. The method of any item 1 to 8, wherein the method of identification is a method of purifying/isolating a polypeptide comprising an interspecies-specific binding domain capable of binding to an epitope of human and non-chimpanzee non-human primate CD3e.

Item 11. Use of an N-terminal fragment of the CD3e extracellular domain of a maximum of 27 amino acids containing the amino acid sequence Gin-Asp-Gly-Asn-Giu-Glu-Met-Gly (SEQ ID NO. 381) or Gln-Asp-Gly-Asn-Glu-Glu-lle-Gly (SEQ ID NO. 382) to create an interspecies specific binding domain.

In accordance with the stated use of the invention, it is preferred that the generated interspecies specific binding domain is a domain of human origin.

Item 12. The use according to item 11, wherein the interspecies specific binding domain is an antibody.

Item 13. Use according to item 12, wherein the antibody is a single chain antibody.

Item 14. Use according to items 12 to 13, wherein the antibody is a bispecific antibody.

[0138] These and other embodiments are described and included in the description and Examples of the subject invention. Recombinant techniques and methods in immunology are described e.g. in Sambrook et al. Molecular Cloning: A Laboratory Manual; Cold Spring Harbor Laboratory Press, 3rd edition 2001; Lefkovits; Immunology Methods Manual; The Comprehensive Sourcebook of Techniques; Academic Press, 1997; Golemis; Protein-Protein Interactions: A Molecular Cloning Manual; Cold Spring Laboratory Press, 2002. Additional literature concerning any of the antibodies, methods, uses and compounds to be used in accordance with the present invention can be found in public libraries and databases, using, for example, electronic devices. Other databases and addresses such as hf'o://www.ncbi.nlm.nih.gov/ or listed on the EMBL-services homepage under http://www embl.de/services/index.html are known to the person skilled in the art and can also be obtained using, eg, http://www. google.com.

[0139] The figures show:

Picture 1

Fusion of N-terminal amino acids 1-27 of primate CD3 epsilon with a heterologous soluble protein

Picture 2

The figure shows the mean absorbance values obtained from four parallel samples each measured by an ELISA assay that detects the presence of a construct with the N-terminal amino acids 1-27 of the mature human CD3 epsilon chain fused to the hinge region and the Fc gamma part of human lgG1 and a C-terminal marker with 6 Histidines in the supernatant of transiently transfected 293 cells. The first column, labeled "27 aa huCD3E" shows the mean absorbance values for construct, the second column, labeled "irrel. SN" shows mean values for the supernatant of 293 cells, transfected with an irrelevant construct, as a negative control. Comparison of the values obtained for the construct with the values obtained for the negative control clearly shows the presence of the reference construct.

Picture 3

The figure shows the average absorbance values obtained on four parallel samples measured by an ELISA probe that detects the binding of molecules that bind to anti-CD3 interspecies-specifically, in the form of crude preparations of periplasmically expressed single-chain antibodies, to a construct with N-terminal amino acids of 1-27 of the mature human CD3 epsilon chain fused with the hinge region and Fc gamma part of human lgG1 and the C-terminal marker His6. From left to right, columns show mean absorbance values for specificities labeled A2J HLP, I2C HLP, E2M HLP, F70 HLP, G4H HLP, H2C HLP, E1L HLP, F12Q HLP, F6A HLP, and H1E HLP. The far?right column, labeled "neg. contr." shows mean absorbance values for preparations of mouse single-chain antibodies to human CD3 as negative controls. Comparison of the values obtained for the anti-CD3 specificity with the values obtained for the negative control clearly shows the strong binding of the anti-CD3 specificity to the N-terminal amino acids 1-27 of the mature human CD3 epsilon chain. ;Figure 4;Fusion of the N-terminal amino acids of 1-27 primate CD3 epsilon with a heterologous membrane-bound protein. ;Figure 5;Histograms of various transfectants tested with a FACS probe that detects the presence of a recombinant transmembrane fused protein consisting of cynomolgus EpCAM and the N-terminal 1-27 amino acids of the CD3 epsilon chain of humans, marmoset monkeys, tamarins, squirrel monkeys, and domestic pigs. Left to right and top to bottom histograms show results for transfectants expressing human 27-mer, marmoset 27-mer, tamarin 27-mer, squirrel monkey 27-mer, and pig 27-mer constructs, respectively. In the individual plots, the thin line represents the sample incubated in PBS with 2% FCS, instead of the anti-Flag M2 antibody, as a negative control, and the thick line represents the sample incubated with the anti-Flag M2 antibody. For each construct, the histograms show the binding of the anti-Flag M2 antibody to the transfectants, clearly showing the expression of the recombinant constructs on the transfectants. ;Figure 6 ;Histogram of different transfectants tested with a FACS probe that detects the binding of interspecies-specific anti-CD3 binding molecules in the form of crude preparations of periplasmically expressed single-chain antibodies to N-terminal amino acids 1-27 of CD3 epsilon chain of human, marmoset, tamarin, or squirrel monkey, fused with cynomolgus EpCAM. ;Figure 6A: ;Histograms from left to right and from top to bottom show the results for transfectants expressing 1-27 CD3-EpCAM with 27-mer human probed with CD3-specific binding molecules designated as H2C HLP, F12Q HLP, E2M HLP, and G4H ;HLP, respectively. ;Figure 6B: ;Histograms from left to right and from top to bottom show the results for transfectants expressing 1-27 CD3-EpCAM with 27-mer marmoset probed with CD3-specific binding molecules labeled H2C HLP, F12Q HLP, E2M HLP, and G4H HLP, respectively. ;Figure 6C:;Histograms from left to right and from top to bottom show the results for transfectants expressing 1-27 CD3-EpCAM with 27-mer tamarin probed with CD3-specific binding molecules, labeled as H2C HLP, F12Q HLP, E2M HLP, and G4H HLP, respectively. ;Figure 6D:;Histograms from left to right and from top to bottom show the results for transfectants expressing 1-27 CD3-EpCAM with a 27-mer squirrel monkey probed with CD3-specific binding molecules labeled H2C HLP, F12Q HLP, E2M HLP, and G4H HLP, respectively. ;Figure 6E:;Histograms from left to right and from top to bottom show the results for transfectants expressing 1-27 CD3-EpCAM with a 27-mer pig tested with CD3-specific binding molecules, designated as H2C HLP, F12Q HLP, E2M HLP, and G4H ;HLP, respectively. ;In the individual views, a thin line represents a sample incubated with a preparation of murine single-chain antibody to human CD3, as a negative control, and a thick line represents a sample incubated with the corresponding indicated antibody to CD3. Considering the absence of binding to transfectants with the pig 27-mer, and the expression levels of the constructs shown in Figure 5, the histograms show the specific and strong binding of the tested anti-CD3 specificities of the fully interspecies-specific human bispecific single-chain antibodies to cells expressing recombinant transmembrane fusion proteins containing the N-terminal amino acids 1-27 of the human, marmoset, tamarin, squirrel monkey CD3 epsilon chain fused with the cynomolgus EpCAM, and thus demonstrate interspecies specificity of anti-CD3 binding molecules for multiple primates. ;Figure 7;FACS assay for detection of human CD3 epsilon on transfected murine EL4 T cells. Graphical analysis shows histograms. Thick line shows transfected cells incubated with anti-human CD3 antibody UCHT-1. The thin line represents cells incubated with a mouse lgG1 isotype control. UCHT1 anti-CD3 antibody staining clearly demonstrates expression of the human CD3 epsilon chain on the surface of transfected murine EL4 T cells. ;Figure 8;Binding of interspecies-specific anti-CD3 antibodies to alanine mutants in an alanine-probed experiment. In the individual figures, the columns show, from right to left, the calculated binding values in arbitrary units on a logarithmic scale for wild-type (WT) transfectants and alanine mutants from positions 1 to 27. Binding values were calculated from the following formula: ;In this equation, value_ Sample means a value in arbitrary binding units that describes the degree of binding of anti-CD3 antibodies to a specific alanine mutant, as shown in this picture. Sample represents the geometric mean fluorescence value obtained for the specific anti-CD3 antibody tested for the specific transfectant by alanine search.neg_ Contr.represents the geometric mean fluorescence value obtained for the negative control tested for the specific alanine mutant,UCHT-*\represents the geometric mean fluorescence value obtained for the UCHT-1 antibody tested for the specific alanine mutant,WTrepresents the geometric mean fluorescence value obtained for the specific anti-CD3 antibody tested for the wild-type transfectant, x specifies the given transfectant, y specifies the corresponding anti-CD3 antibody awt specifies that the given transfectant is wild type. Individual positions of alanine in the mutant are indicated by a letter representing the amino acid in the wild type and the number of the given position.

Figure 8A:

The figure shows the results for the interspecies-specific anti-CD3 antibody A2J HLP expressed as a chimeric IgG molecule. Reduced binding activity is found for mutations with alanine at position 4 (asparagine), at position 23 (threonine) and at position 25 (isoleucine). Complete loss of binding is found for mutations to alanine at position 1 (glutamine), at position 2 (aspartate), at position 3 (glycine), and at position 5 (glutamate).

Figure 8B:

The figure shows the results for the interspecies-specific anti-CD3 antibody E2M HLP, which is expressed as a chimeric IgG molecule. Reduced binding activity is found for mutations with alanine at position 4 (asparagine), at position 23 (threonine) and at position 25 (isoleucine). Complete loss of binding is found for mutations to alanine at position 1 (glutamine), at position 2 (aspartate), at position 3 (glycine), and at position 5 (glutamate).

Figure 8C:

The figure shows the results for the interspecies-specific anti-CD3 antibody H2C HLP, which is expressed as a chimeric IgG molecule. Reduced binding activity is found for mutations with alanine at position 4 (asparagine). Complete loss of binding is found for mutations to alanine-glutamine at position 1 (glutamine), at position 2 (aspartate), at position 3 (glycine) and at position 5 (glutamate).

Figure 8D:

shows results for the interspecies-specific anti-CD3 antibody F12Q HLP, tested as a periplasmically expressed single-chain antibody. Complete loss of binding is found for mutations to alanine at position 1 (glutamine), at position 2 (aspartate), at position 3 (glycine), and at position 5 (glutamate).

Figure 9

FACS assay detecting binding of the interspecies-specific anti-CD3 binding molecule H2C HLP to human CD3, with and without an N-terminal His6 tag.

Histograms were obtained on EL4 cells, transfected with wild-type human CD3 epsilon chain (left histogram) or CD3 epsilon chain with N-terminal His6 marker (right histogram), tested with a FACS probe that detects the binding of the interspecies-specific binding molecule H2C HLP. Samples were incubated with the corresponding isotype as a negative control (thin line), human anti-CD3 antibody UCHT-1 as a positive control (dotted line) and interspecies-specific anti-CD3 antibody H2C HLP in the form of a chimeric IgG molecule (thick line).

Histogram shows comparable UCHT-1 antibody binding to both transfectants compared to the isotype control, indicating expression of both recombinant constructs. The histogram also shows the binding of the anti-CD3 binding molecule H2C HLP only to the natural CD3 epsilon chain, but not to the His6-human CD3 epsilon chain. These results indicate that the free N-terminus is essential for the binding of the interspecies-specific anti-CD3 binding molecule H2C HLP.

Figure 10

Binding to saturation of EGFR-21-63 LH x H2C to human CD3-positive PBMC to determine KD value of CD3 binding to cells by FACS analysis. The assay was performed as described in Example 7.

Took pictures

FACS binding analysis of certain interspecies-specific bispecific single-chain constructs to human EGFR-transfected CHO cells, the human CD3+ T cell line HPB-ALL, cynomolgus EGFR-transfected CHO cells, and the macaque T cell line 4119 LnPx. FACS staining was performed as described in Example 12. The thick line represents cells incubated with 2 µg/ml of purified protein, which were then incubated with anti-his antibody and labeled PE detection antibody. The thin histogram line represents the negative control: cells incubated only with anti-his antibody and detection antibody.

Figure 12

Cytotoxic activity induced by a certain interspecies-specific single-chain construct specific for EGFR, redirected to the indicated cell lines. A) and B) Stimulated CD4-/CD56- human PBMCs were used as effector cells, CHO cells, transfected with human EGFR as target cells. The assay was performed as described in Example 13.

Figure 13

Cytotoxic activity induced by specific interspecies-specific EGFR-specific single-chain construct redirected to indicated cell lines.A) and B) T-cell line 4119 LnPx macaque was used as effector cells, CHO cells, EGFR-transfected cynomolgus as target cells. The assay was performed as described in Example 13.

Figure 14

FACS binding analysis of interspecies-specific bispecific single-stranded constructs to CHO cells transfected with human MCSP D3, human

CD3+ T cell line HPB-ALL, MCSP D3 cynomolgus transfected CHO cells and macaque T cell line 4119 LnPx. FACS staining was performed as described in Example 17. The thick line represents cells incubated with 2 µg/ml of purified protein, which were then incubated with anti-his antibody and labeled PE detection antibody. The thin histogram line represents the negative control: cells incubated only with anti-his antibody and detection antibody

Siika15 FACS binding analysis of interspecies-specific bispecific single-chain constructs for human MCSP D3-transfected CHO cells, human CD3+ T cell line HPB-ALL, cynomolgus MCSP D3-transfected CHO cells, and macaque T cell line 4119 LnPx. FACS staining was performed as described in Example 17. The thick line represents cells incubated with 2 µg/ml of purified protein, which were then incubated with anti-his antibody and labeled PE detection antibody. The thin histogram line represents the negative control: cells incubated only with anti-his antibody and detection antibody

Figure 16

FACS binding analysis of interspecies-specific bispecific single-chain constructs to human MCSP D3-transfected CHO cells, human CD3+ T cell line HPB-ALL, cynomolgus MCSP D3-transfected CHO cells, and macaque T cell line 4119 LnPx. FACS staining was performed as described in Example 17. The thick line represents cells incubated with 2 µg/ml of purified monomeric protein, which were then incubated with anti-his antibody and labeled PE detection antibody. The thin histogram line represents the negative control: cells incubated only with anti-his antibody and detection antibody

Figure 17

Cytotoxic activity induced by certain interspecies-specific MCSP-specific single-chain constructs redirected to the indicated target cell lines. A) Stimulated CD4-/CD56- human PBMCs were used as effector cells, CHO cells, transfected with human MCSP D3 as target cells. B) T-cell line 4119 LnPx macaque was used as effector cells, CHO cells, transfected MCSP D3 cynomolgus as target cells. The assay was performed as described in Example 18.

Figure 18

Cytotoxic activity induced by certain interspecies-specific MCSP-specific single-chain constructs redirected to the indicated target cell lines. A) and B) T-cell line 4119 LnPx macaque was used as effector cells, CHO cells, transfected MCSP D3 cynomolgus as target cells. The assay was performed as described in Example 18.

Figure 19

Cytotoxic activity induced by certain interspecies-specific MCSP-specific single-chain constructs, redirected to the indicated target cell lines.. A) and B) Stimulated CD4-/CD56- human PBMCs were used as effector cells, CHO cells, transfected with human MCSP D3 as target cells. The assay was performed as described in Example 18.

Figure 20

Cytotoxic activity induced by certain interspecies-specific MCSP-specific single-chain constructs redirected to the indicated target cell lines. A) Stimulated CD4-/CD56- human PBMCs were used as effector cells, CHO cells, transfected with human MCSP D3 as target cells. B) T-cell line 4119 LnPx macaque was used as effector cells, CHO cells, transfected MCSP D3 cynomolgus as target cells. The assay was performed as described in Example 18.

Figure 21

Cytotoxic activity induced by certain interspecies-specific MCSP-specific single-chain constructs redirected to the indicated target cell lines. A) Stimulated CD4-/CD56- human PBMCs were used as effector cells, CHO cells, transfected with human MCSP D3 as target cells. B) T-cell line 4119 LnPx macaque was used as effector cells, CHO cells, transfected MCSP D3 cynomolgus as target cells. The assay was performed as described in Example 18.

Figure 22

Plasma stability of MCSP and CD3 interspecies-specific bispecific single-chain antibodies tested by measuring the cytotoxic activity induced by samples of specific single-chain constructs incubated with 50% human plasma at 37°C and 4°C for 24 hours either with the addition of 50% human plasma immediately prior to cytotoxicity testing or without the addition of plasma. CHO cells transfected with human MCSP were used as the target cell line, and stimulated CD4-/CD56- human PBMCs were used as effector cells. The assay was performed as described in Example 19.

Figure 23

FACS binding analysis of interspecies-specific bispecific single-chain constructs to human HER2-transfected CHO cells, the human CD3+ T cell line HPB-ALL, human HER2-transfected CHO cells, and the macaque T cell line 4119 LnPx. FACS staining was performed as described in Example 23.4. The thick line represents cells incubated with 2 µg/ml of the purified bispecific single-stranded construct. Thin line represents negative controls. PBS with 2% FCS was used as a negative control. For each interspecies-specific, bispecific single-stranded construct, the histograms show the specific binding of the construct to human and macaque HER2 and human and macaque CD3

Figure 24

The diagrams show the results of chromium release assays measuring the cytotoxic activity induced by certain interspecies-specific HER2-specific single-chain constructs directed to the indicated target cell lines. The specified effector cells were also used. Tests were performed as described in Example 23.5. The diagrams clearly show for each shown construct the strong recruitment of cytotoxic activity of effector cells towards CHO cells transfected with human and macaque HER2.

Figure 25

Analysis of periplasmic preparations containing scFv protein fragments with a Flag marker from selected clones using CD3-specific ELISA. Periplasmic preparations of protein fragments of soluble scFv were added to ELISA plate wells, which were coated with soluble human CD3 epsilon (ak 1-27)-Fc fusion protein and which were additionally blocked with PBS with 3% BSA. Detection was performed with a monoclonal labeled antibody, after which streptavidin conjugated with peroxidase was added. ELISA was developed with ABTS substrate solution. OD values (y axis) were measured at 405 nm using an ELISA reader. Clone names are represented on the x-axis.

Figure 26

Analysis of periplasmic preparations containing scFv protein fragments with Flag marker from selected clones using ELISA test. The same periplasmic preparations of soluble scFv protein fragments as in Fig. 25 were added to ELISA plate wells, which were not coated with human CD3 epsilon (ak 1-27)-Fc fusion protein, but with HUlgGI (Sigma) and blocked with 3% BSA in PBS. Detection was performed with a monoclonal labeled antibody, after which streptavidin conjugated with peroxidase was added. ELISA was developed with ABTS substrate solution. OD values (y axis) were measured at 405 nm using an ELISA reader. Clone names are represented on the x-axis.

[0140] The present invention is further described by way of the following illustrative, non-limiting examples which provide a better understanding of the present invention and its many advantages.

EXAMPLES

1. Identification of CD3epsilon sequences from non-human primate blood samples

[0141] To identify CD3epsilon, blood samples were taken from the following primates: Callithrixjacchus, Saguinus oedipusiSaimiris scivreus. Fresh whole blood samples treated with heparin were prepared for isolation of total cellular RNA according to the manufacturer's procedure (OlAamp RNA Blood Mini Kit, Oiagen). Extracted mRNA was transcribed into cDNA according to published procedures. Briefly, 10 µl of precipitated RNA was incubated for 10 min with 1.2 µl of 10x hexanucleotide mix (Roche) at 70 °C, and stored on ice. The reaction mixture consisted of 4 µl buffer 5x superscript II, 0.2 µl 0.1 M dithiothritol, 0.8 µl superscript II (Invitrogen), 1.2 µl deoxyribonucleoside triphosphate (25 µM), 0.8 µl RNase Inhibitor (Roche), and 1.8 µl DNase- and RNase-free water (Roth). The reaction mixture was incubated at room temperature for 10 min, then at 42 °C for 50 min and at 90 °C for 5 min. The reaction was cooled on ice before the addition of 0.8 µl RNaseH (1 U/µl, Roche) and incubated for 20 min at 37 °C.

[0142] First strand cDNAs from each species were subjected to 35 separate cycles of polymerase chain reaction using Taq DNA polymerase (Sigma) and the following primer combinations designed from database searches, left primer

(forward primer) 5'-AGAGTTCTGGGCCTCTGC-3' (SEQ ID NO: 377); right stop (reverse primer) 5-CGGATGGGCTCATAGTCTG-3' (SEQ ID NO: 378);. Amplified bands of 550 bp were gel purified (Gel Extraction Kit, Oiagen) and sequenced (Seq iserve, Vaterstetten/Germany, see sequence list).

CD3epsilonCallithrix jacchus - a

Nucleotides

[0143]

Amino acids

[0144]

CD3epsilonSa<g>uinus oedious - a

Nucleotides

[0145]

Amino acids

[0146]

CD3epsilonSaimiris ciureus - a

Nucleotides

[0147]

Amino acids

[0148]

2. Generation of fragments (scFv) of an interspecies-specific single-chain antibody that binds to N-terminal amino acids 1-27 of human CD3epsilon and various non-chimpanzee primates

2.1. Immunization of mice using the N-terminus of CD3epsilon separated from its native CD3-context by fusion with a soluble heterologous protein

[0149] Ten-week-old F1 mice, derived from a cross between balb/c x C57cmi, were immunized with a CD3epsilon-Fc fusion protein having the N-terminal 1-27 amino acids of the mature CD3epsilon chain (1-27 CD3-Fc) of human and Ilisaimiris sciureus (South American squirrel monkey). To this end, each mouse was intraperitoneally injected with 40 ug of 1-27 CD3-Fc fusion protein with 10 nmol of thioate-modified CpG-Oligonucleotides (5'-tccatgacgttcctgatgct-3') in 300 ul of PBS. Mice received a booster immunization in the same manner after 21, 42 and, optionally, 63 days. Ten days after the first booster immunization, blood samples were taken and the titer of antibodies to 1-27 CD3-Fc fused protein iwa was determined in the serum by ELISA. In addition, the titer of the CD3-positive human T cell line HPBall was tested by flow cytometry according to standard procedures. Serum titers were significantly higher in immunized animals than in non-immunized ones.

2.2. Generating a scFv immune mouse antibody library: Constructing a combinatorial

antibody libraries and phage display

[0150] Three days after the last injection, mouse spleen cells were harvested for total RNA according to standard procedures.

[0151] A DNA fragment library of the immunoglobulin (Ig) variable region (VK) of the light chain (cap) and the variable region (VH) of the mouse Ig heavy chain was constructed by RT-PCR on mouse spleen RNA using VK and VH specific primers. cDNA was synthesized by standard procedures.

[0152] The spacers were designed to create a recognition site for 5'-XhoI and 3'-BstEII in amplified V-fragments of the heavy chain and a recognition site for 5'-Sacl and 3'-Spel in amplified VK DNA fragments.

[0153] For PCR-amplification of VH DNA-fragments each of eight different 5'-VH-family specific limiters (MVH1(GC)AG GTG CAG CTC GAG GAG TCA GGA CCT; MVH2

GAG GTC CAG CTC GAG CAG TCT GGA CCT; MVH3 CAG GTC CAA CTC GAG CAG CCT GGG GCT; MVH4 GAG GTT CAG CTC GAG CAG TCT GGG GCA; MVH5 GA(AG) GTG AAG CTC GAG GAG TCT GGA GGA; MVH6 GAG GTG AAG CTT CTC GAG TCT GGA GGT; MVH7 GAA GTG AAG CTC GAG GAG TCT GGG GGA; MVH8 GAG GTT CAG CTC GAG CAG TCT GGA GCT) was combined with a single 3'-VH stop (3'MuVHBstEII tga gga gac ggt gac cgt ggt ccc ttg gcc cea g); for PCR amplification of VK-chain fragments of seven different 5'-VK-family limiters (MUVK1 CCA GTT CCG AGC TCG TTG TGA CTC AGG AAT CT; MUVK2 CCA GTT CCG AGC TCG TGT TGA CGC AGC CGC CC; MUVK3 CCA GTT CCG AGC TCG TGC TCA CCC AGT CTC CA; MUVK4 CCA GTT CCG AGC TCC AGA TGA CCC AGT CTC CA; MUVK5 CCA GTG AGC TCG TGA; MUVK6 CCA GAT GTG TCG TCA; MUVK7 CCA GTT CCG AGC TCG CA) is combined with a single 3'-VK stop (3'MuVkHindlll/BsiW1 tgg tgc act agt cgt acg ttt gat ete aag ctt ggt ccc).

[0154] The following program was used for PCR amplification: denaturation at 94°C for 20 sec; renaturation of the limiter at 52°C for 50 sec, extension of the limiter at 72°C for 60 sec and 40 cycles, after which comes the final extension at 72<O0>C, 10 min.

[0155] 450 ng of light kappa chain fragments (Sac1-Spel treated) were ligated with 1400 ng phagemid pComb3H5Bhis (large fragment treated with Sac1-Spel enzymes). The resulting combinatorial library of antibodies was transformed by electroporation of electrocompetent Escherichia coli XL1Blue cells in 300 ul (2.5 kV, cuvette with a distance between electrodes of 0.2 cm, 25 uFD, 200 Ohm, Biorad gene-pulser) yielding a library with more than 10<7> independent clones. After one hour of phenotypic expression, positive transformants for carbenicillin resistance encoded by pComb3H5BHis vector were selected in 100 ml of liquid culture in super broth (SB, from super broth), overnight. Cells were harvested by centrifugation and plasmid was isolated using a commercial plasmid isolation kit (OJagen).

[0156] 2800 ng of this plasmid DNA containing the VK-library (treated with Xhol-BstEII enzymes, large fragments) was ligated with 900 ng of V-fragments of the heavy chain (treated with Xhol-BstEII enzymes) and two aliquots of 300 ul of electrocompetentE were again transformed with them. coliXL1 Blue cells by electroporation (2.5 kV, 0.2 cm gap cuvette, 25 uFD, 200 Ohm) to obtain a total VH-VK scFv (single-stranded variable fragment) library of more than 10<7> independent clones.

[0157] After phenotypic expression and slow adaptation to carbenicillin, cells of E. coli containing the antibody library were transferred to the selection medium SB-carbenicillin (50 ug/mL). CellsE. colicells containing the antibody library were then infected with an infectious dose of 10<12> particles of helper phage VCSM13, which led to the production and secretion of filamentous phage M13, where the phage particles contain single-stranded pComb3H5BHis-DNA encoding a mouse scFv-fragment and express the corresponding scFv-protein as a translational fusion with protein III of the phage coat. This pool of phages displaying an antibody library was later used to select antigen-binding bodies.

2.3. Selection of molecules that specifically bind to CD3 based on the exposed ones

phase

[0158] A phage library containing the cloned scFv repertoire was collected from the supernatant by PEG8000/NaCl precipitation and centrifugation. About 10<11> to 10<12>scFv phage particles were resuspended in 0.4 ml PBS/0.1% BSA and incubated with 10<5> to 10<7> Jurkat cells (a CD3-positive T-cell line) for 1 hour on ice with slow mixing. These Jurkat cells were previously cultured in RPMI medium supplemented with fetal bovine serum (10%), glutamine, and penicillin/streptomycin, harvested by centrifugation, washed in PBS, and resuspended in PBS/1% FCS (containing Na azide). Phase scFvs that did not specifically bind to Jurkat cells were eliminated by up to five washes with PBS/1% FCS (containing Na azide). After washing, the bound material was eluted from the cells by resuspending the cells in HCl-glycine pH 2.2 (10 min incubation followed by vortexing) and after neutralization with 2 M Tris pH 12, the eluate was used to infect a fresh, uninfected culture of E. coli XL1 Blue (OD600> 0.5). CultureE. coli containing E. coli cells successfully transduced with a phagemid copy encoding the human scfv-fragment were again selected for carbenicillin resistance and then infected with VCMS 13 helper phage to initiate a second round of antibody exposure and in vitro selection. Generally, a total of 4 to 5 rounds of selection were performed.

2.4. Screening for molecules that specifically bind to CD3-

[0159] Plasmid DNA, corresponding to the 4th and 5th rounds of panning, was isolated from cultures of E. coli post-selection. To obtain the soluble scFv-protein, fragments of VH-VL-DNA were cut from the plasmid (Xhol-Spel). These fragments were cloned through the same restriction sites into plasmid pComb3H5BFIag/His, which differ from the original plasmid pComb3H5BHis in that the expression construct (eg scFv) includes a Flag-tag (TGD YKDDDDK) between scFv and His6-tag, and additional phage proteins are deleted. After ligation, 100 µl of heat shock competent E. coli TG1 or XLI blue were transformed with each set (different rounds of panning) of plasmid DNA and plated on LB-agar with carbenicillin. Single colonies were transferred to 100 µl of LB Carb (50 µg/ml).

[0160] E coli transformed with Comb3H5BHis containing VL- and VH-segments produce soluble scFv in sufficient quantities after excision of the lili gene fragment inducible with 1 mM IPTG. Thanks to a suitable signal sequence, the scFv chain is exported to the periplasm where it is packaged into a functional conformation.

[0161] Single bacterial colonies of E.coli TG1 were taken from transformation plates for periplasmic preparations on a small scale and grown in SB-medium (e.g. 10 ml) to which 20 mM MgCl 2 and carbenicillin 50 µg/ml were added and, after collection, redissolved in PBS (e.g. 1 ml). After four rounds of freezing at -70°C and thawing at 37°C, the bacterial outer membrane disrupted by temperature shock and soluble periplasmic proteins, including cFv, were released into the supernatant. After removal of whole cells and cell debris by centrifugation, the supernatant containing human CD3-scFv antibodies was collected for further studies.

2.5. Identifying molecules that specifically bind to CD3

[0162] The binding of isolated scFv was checked by flow cytometry on eukaryotic cells that express on their surface a heterologous protein that exposes the first 27 N-terminal amino acids of CD3epsilon at its N-end.

[0163] As described in Example 4, the first 1-27 amino acids of the N-terminus of the mature CD3 epsilon chain of the human T cell receptor complex (amino acid sequence: QDGNEEMGGITQTPYKVSISGTTVILT) were fused to the N-terminus of the transmembrane protein EpCAM so that the N-terminus was located on the outside of the cell surface. In addition, the FLAG epitope was inserted between the N-terminus of the 1-27 CD3 epsilon sequence and the EpCAM sequence. This fusion product is expressed in human embryonic kidney (HEK) and Chinese hamster ovary (CHO) cells.

[0164] Eukaryotic cells exposing the first 27 amino acids of the N-terminus of the mature CD3epsilon of other primate species were obtained in the same way for Saimiri sciureus (squirrel monkey) (N-terminal amino acid sequence of CD3epsilon: QDGNEEIGDTTQNPYKVSISGTTVTLT), for Callithrix jacchus (N-terminal amino acid sequence of CD3epsilon: QDGNEEMGDTTQNPYKVSISGTTVTLT) and for Saguinus oedipus (N-terminal amino acid sequence of CD3epsilon: QDGNEEMGDTTQNPYKVSISGTTVTLT).

[0165] For flow cytometry, 2.5x10<5> cells were incubated with 50 µl of supernatant or with 5 µg/ml of purified constructs in 50 µl of PBS with 2% FCS. Binding of the constructs is detected with an anti-His antibody (Penta-His Antibody, without BSA, Oiagen GmbH, Hilden, FRG) at 2 µg/ml in 50 µl PBS with 2% FCS. As a reagent in the second step, the F(ab')2 fragment purified by affinity chromatography, conjugated with R-phycoerythrin, goat anti-mouse IgG (specific for the Fc gamma fragment), diluted 1:100 in 50 µl PBS with 2% FCS (Dianova, Hamburg, FRG) was used. Samples were measured on a FACSscan (BD biosciences, Heidelberg, FRG).

[0166] Binding was always confirmed by flow cytometry, as described in the previous paragraph, on primary T-cells of humans and various primates (eg saimiris sciureus, callithrix

Jacchus, bloody Oedipus).

2.6. Creation of human/humanized scFv-specific equivalents for CD3epsilon,

except for human

[0167] The VH region of the murine anti-CD3 scFv was compared to the sequences of human embryonic antibodies. The human embryonic antibody VH sequence that has the highest homology with other VHs was selected and a direct comparison of the two amino acid sequences was performed. Within that framework there are a number of amino acids in other VHs that differ from the amino acids of the human VH region within that framework ("different framework positions"). Some of these amino acids may contribute to the binding and activity of a given antibody to its target.

[0168] To construct a library containing mouse CDRs differing in every frame position from the selected human VH sequence, both possibilities (both human and maternal mouse amino acids), native oligonucleotides were synthesized. These oligonucleotides are incorporated into different amino acid positions in humans with a probability of 75% and in mice with a probability of 25%. For one human VH, for example, six of these oligonucleotides must be synthesized to cover the end sequence of about 20 nucleotides. To that end, every other limiter was an anti-limiter (antisense example). The restriction sites required for subsequent cloning were removed from the oligonucleotides.

[0169] These spacers can be 60 to 90 nucleotides in length, depending on the number of spacers needed to cover the entire V sequence.

[0170] Such, e.g., six limiters were mixed in equal amounts (eg, 1 ul of each limiter (stopper stock solutions 20 to 100 uM) for 20 ul PCR reactions) and added to a PCR mixture consisting of PCR buffer, nucleotides and Taq polymerase. This mixture is incubated in the PCR apparatus for 3 minutes at 94 °C, at 65 °C for 1 minute, at 62 °C for 1 minute, at 59 °C for 1 minute, at 56 °C for 1 minute, at 52 °C for 1 minute, at 50 °C for 1 minute and at 72 °C for 10 minutes. The product is then electrophoresed in an agarose gel and products of length 200 to 400 are isolated from the gel by standard methods.

[0171] Such a PCR product is then used as a template for a standard PCR reaction using primers that include N-terminal and C-terminal restriction sites suitable for cloning. A DNA fragment of the correct size (about 350 nucleotides for VH) was isolated by electrophoresis on agarose gel using standard methods. In this way, sufficient VH DNA fragment was amplified. This VH fragment represents now a set of VH fragments, each of which has a different amount of murine and human amino acids at different frame positions (a set of humanized VHs). The same procedure was performed for the VL region of mouse anti-CD3 scFv (set of humanized VLs).

[0172] The set of humanized VHs was then combined with the set of humanized VLs in the phage display vector pComb3H5Bhis to obtain a library of functional scFvs from which, after display on filamentous phage, CD3-binding molecules were selected, screened, identified and validated as already described for other (mouse) derived anti-CD3 scFvs. Individual clones were then analyzed for suitable traits and amino acid sequences. The scFvs that had the highest amino acid sequence homology with human embryonic V-segments were chosen, especially those where at least one CDR of CDR I and II VH and CDR I and II VLkapa or CDR I and II VLIambda has more than 80% identical amino acid sequence with the closest CDR of all human embryonic V-segments. Anti-CD3 scFvs were translated into recombinant bispecific single chain antibodies, as described in Examples 10 and 16 below, and further characterized.

3. Creation of a recombinant fusion protein with 1-27 amino acids from the N-terminus of CD3

human epsilon chain fused to the Fc-portion of an lgG1 (1-27 CD3-Fc).

3.1. Cloning and expression of 1-27 CD3-Fc

[0173] Coding sequence for amino acids 1-27. The N-terminus of the human CD3 epsilon chain, fused to the hinge region and the Fc gamma region of human immunoglobulin lgG1, as well as to the 6 Histidine Tag marker was obtained by gene synthesis using standard procedures (cDNA sequences and amino acid sequences of the recombinant fused protein are listed as SEQ ID Nos. 350 and 349). The synthesis of the gene fragment was designed to first have the Kozak site required for eukaryotic expression of the construct, followed by the 19 amino acids of the immunoglobulin head peptide, followed, in proper reading frame, by the coding sequence of the first 27 amino acids of the extracellular portion of the mature human CD3 epsilon chain, then the coding sequence of the hinge region and the Fc gamma portion of human lgG1, then the coding sequence of marker 6 Histidine and stop codon (Figure 1). Gene fragment synthesis was also designed to introduce restriction sites at the beginning and end of the cDNA encoding the fusion protein. The introduction of restriction sites, EcoRI at the 5' end and SalI at the 3' end, is used for the following cloning procedures: The synthesized gene fragment was cloned via EcoRI and SalI sites into a plasmid designated as pEF-DHFR (pEF-DHFR is described in Mack et al. Proc. Natl. Acad. Sci. USA 92 (1995) 7021-7025) by standard procedures. The sequence-verified plasmid was used to transfect the FreeStvle 293 Expression System (Invitrogen GmbH, Karlsruhe, Germany) according to the manufacturer's procedure. After 3 days of growing the CD38 culture, the supernatants of the transfectants were collected and checked for the presence of the recombined construct by ELISA. Goat anti-human IgG antibody specific for the Fc-gamma fragment (obtained from Jackson ImmunoResearch Europe Ltd., Newmarket, Suffolk, UK) was diluted in PBS to 5 pg/ml and coated onto wells in a 96-well MaxiSorp ELISA plate (Nunc GmbH & Co. KG, Wiesbaden, Germany). For each well, 100 µl was used overnight at 4 °C. The slides were washed with PBS with 0.05% Tween 20 (PBS/Tween) and blocked with 3% BSA in PBS (bovine albumin, fraction V, Sigma-Aldrich Chemie GmbH, Taufkirchen, Germany) for 60 minutes at room temperature (RT). The wells were then washed again with PBS/Tween and incubated with the cell supernatants for 60 min at RT. After washing, the slides were incubated with peroxidase-conjugated anti-His6 antibody (Roche Diagnostics GmbH, Roche Applied Science, Mannheim, Germany), diluted 1:500 in uPBS with 1% BSA, for 60 minutes at RT. The wells were then washed with 200 µl of PBS/Tween and 100 µl of SIGMAFAST OPD [o-phenylenediamine dihydrochloride] substrate solution (Sigma-Aldrich Chemie GmbH, Taufkirchen, Germany) was added according to the manufacturer's instructions. The reaction was stopped by adding 100 µl of 1 M H2SO4. The color reaction was measured on a PowerWaveX microplate spectrophotometer (BioTek Instruments, Inc., Winooski, Vermont, USA) at 490 nm and subtracted the background absorbance value at 620 nm. As shown in Figure 2, the presence of the construct, compared to the irrelevant supernatants of mock-transfected HEK 293 cells, used as negative controls, was clearly detectable.

3.2. Assay for the binding of intraspecies-specific single-chain antibodies to 1-27 CD3-Fc.

[0174] Binding of crude preparations of interspecies-specific CD3 epsilon-specific single chain antibodies expressed in the periplasm to 1-27 CD3-Fc was checked by ELISA.

Goat antibody to human IgG, an antibody specific for the Fc-gamma fragment (Jackson ImmunoResearch Europe Ltd., Newmarket, Suffolk, UK) was diluted in PBS to 5 pg/ml and coated on the wells of a 96-well MaxiSorp ELISA plate (Nunc GmbH & Co. KG, Wiesbaden, Germany). For each well, 100 µl was used overnight at 4 °C. The slides were washed with PBS with 0.05% Tvveen 20 (PBS/Tvveen) and blocked with 3% BSA in PBS (bovine albumin, fraction V, Sigma-Aldrich Chemie GmbH, Taufkirchen, Germany) for 60 min at RT. The wells were then washed with PBS/Tween and incubated with crude preparations of interspecies-specific single-chain antibodies expressed in the periplasm, as described above, for 60 min at RT. After washing with PBS/Tween, the slides were incubated with peroxidase-conjugated anti-Flag M2 antibody (Sigma-Aldrich Chemie GmbH, Taufkirchen, Germany), diluted 1:10,000 in PBS with 1% BSA, for 60 minutes at RT. The wells were then washed with PBS/Tween and incubated with 100 µl of SIGMAFAST OPD [o-phenylenediamine dihydrochloride] substrate solution (Sigma-Aldrich Chemie GmbH, Taufkirchen, Germany) according to the manufacturer's instructions. The color reaction was stopped by adding 100 µl of 1 M H2SO4 and measured on a PowerWaveX microplate spectrophotometer (BioTek Instruments, Inc., vVinooski, Vermont, USA) at 490 nm and subtracted the background absorbance value at 620 nm. In comparison with mouse anti CD3 single chain antibody, strong binding of interspecies specific human CD3 epsilon single chain antibody was found for the 1-27 CD3-Fc construct (Figure 3).

4. Creation of recombinant transmembrane fusion proteins from 1-27. amino

acids from the N-terminus of CD3 epsilon from various non-chimpanzee primates fused with

Cynomolgus monkey EpCAM (1-27 CD3-EpCAM).

4.1. Cloning and expression of 1-27 CD3-EpCAM

[0175] CD3 epsilon has been isolated from various primates, except chimpanzees (marmoset, tamarin, squirrel monkey) and from pigs. The coding sequences of 1-27 N-terminal amino acids of the CD3 epsilon chain of an adult human, a common marmoset (Callithrix jacchus), a wigged tamarin (Saguinusoedipus), a common squirrel monkey (Saimiri sciureus) and a domestic pig (Sus scrofa; used as a negative control) fused to the N-terminus of the cynomolgus EpCAM, labeled with a Flag marker, were obtained by gene synthesis using standard procedures. The cDNA sequences and amino acid sequences of the recombined fusion proteins are listed as SEQ ID NO. 351 to 360). The gene syntheses of the fragments were designed to contain the first site for BsrGI, which allows for the fusion in the correct reading frame of the coding sequence for the 19 amino acids of the immunoglobulin head peptide, which is already present in the target expression vector, followed, in the same reading frame, by the coding sequence of the first N-terminal 1-27 amino acids of the extracellular part of the mature CD3 epsilon chains, and then, in the same reading frame, by the coding sequence of the Flag marker and, in the same reading frame, the coding sequence of the cynomolgus transmembrane protein EpCAM (Figure 4). Gene synthesis fragments are also designed to introduce a restriction site at the end of the cDNA encoding the fusion protein. The introduced restriction sites BsrGI at the 5' end and SalI at the 3' end were used in the following cloning procedures. The fragments obtained by gene synthesis were cloned via BsrGI and SalI into a derivative of the plasmid designated as pEF DHFR (pEF-DHFR is described in Mack et al. Proc. Natl. Acad. Sci. USA 92 (1995) 7021-7025), which already contains the coding sequence for 19 amino acids of the immunoglobulin head peptide, by standard procedures. Plasmids, whose sequences were verified, were used for transient transfection of 293-HEK cells using MATra-A Reagent (IBA GmbH, Gottingen, Germany) and 12 µg of plasmid DNA for adherent 293-HEK cells in 175 ml culture flasks, according to the manufacturer's instructions. After 3 days of growing in culture, the transfectants were tested for expression of recombinant transmembrane protein on the cell surface using FACS, according to standard procedures. To this end, 2.5x10<5> cells were incubated with anti-Flag M2 antibody (Sigma-Aldrich Chemie GmbH, Taufkirchen, Germany), 5 µg/ml, in PBS with 2% FCS. Bound antibody was detected by affinity chromatography-purified F(ab')2 fragment goat anti-mouse IgG, specific for the Fc-gamma fragment, conjugated to R-phycoerythrin, 1:100 in PBS with 2% FCS (Jackson ImmunoResearch Europe Ltd., Newmarket, Suffolk, UK). Samples were measured on a FACScalibur (BD biosciences, Heidelberg, Germany). The expression of a fusion recombinant transmembrane protein, marked with Flag, consisting of EpCAM cynomolgus and 1.-27. N-terminal amino acids of CD3 epsilon chain of human, marmoset, tamarin, squirrel monkey, or pig on transfected cells (Figure 5).

4.2. Binding of interspecies-specific single-chain anti-CD3 antibodies to 1-27 CD3-EpCAM

[0176] Binding of crude preparations of periplasmically expressed interspecies-specific single-chain anti-CD3 antibodies to 1-27 N-terminal amino acids of human, marmoset, tamarin, or squirrel monkey CD3 epsilon chains, fused with cynomolgus Ep-CAM was tested by FACS assay according to standard procedures. To this end, 2.5x10<5> cells were incubated with crude preparations of periplasmically expressed interspecies-specific single-chain antibodies to CD3 (the preparations were obtained as described above and by standard procedures) and single-chain mouse antibody to human CD3, as a negative control. Penta-His antibody (Oiagen GmbH, Hildesheim, Germany), 5 µg/ml in 50 µl PBS with 2% FCS, was used as secondary antibody. Antibody binding was detected by affinity chromatography-purified F(ab')2 fragment of goat anti-mouse IgG, specific for the Fc-gamma fragment, conjugated to R-phycoerythrin 1:100 in PBS with 2% FCS (Jackson ImmunoResearch Europe Ltd., Newmarket, Suffolk, UK). Samples were measured on a FACScalibur (BD biosciences, Heidelberg, Germany). As shown in Figure 6 (A to E), single chain antibodies were found to bind to transfectants expressing recombinant transmembrane fusion proteins consisting of the 1-27 N-terminal amino acids of human, marmoset, tamarin, or squirrel monkey CD3 epsilon fused to cynomolgus EpCAM. No binding of interspecies-specific single-chain antibodies was found to a fusion protein consisting of the 1-27 N-terminus of porcine CD3 epsilon fused to cynomolgus EpCAM, used as a negative control. Multi-primate interspecies specificity of single-chain anti-CD3 antibodies was demonstrated. The signals obtained with the anti-Flag M2 antibody and the interspecies-specific single-chain antibodies were comparable, indicating a strong binding activity of the interspecies-specific single-chain antibodies to the N-terminal amino acids 1-27 of CD3 epsilon.

5. Analysis of the binding of interspecies-specific anti-CD3 single-chain antibodies by screening

alanine of mouse cells transfected with human CD3 epsilon chain and its alanine

mutants

5.1. Cloning and expression of human native CD3 epsilon

[0177] The coding sequence of the human CD3 epsilon chain was obtained by gene synthesis using standard procedures (the cDNA sequences and amino acid sequences of the human CD3 epsilon chain are listed as SEQ ID Nos. 362 and 361). The gene fragment synthesis was designed to have the Kozak site required for eukaryotic expression of the construct and restriction sites at the beginning and end of the cDNA encoding human CD3 epsilon. The introduced restriction sites EcoRI at the 5' end and SalI at the 3' end were used in subsequent cloning procedures. The synthesized gene fragment was then cloned by standard procedures, via EcoRI and SalI into a plasmid designated as pEF NEO. pEF NEO was derived from pEF DHFR (Mack et al. Proc. Natl. Acad. Sci. USA 92 (1995) 7021-7025) by replacing the cDNA for DHFR with the cDNA for neomycin resistance, by conventional molecular cloning. A plasmid of verified sequence was used to transfect the murine T cell line EL4 (ATCC No. TIB-39) grown in RPMI with stabilized L-glutamine, supplemented with 10% FCS, 1% penicillin/streptomycin, 1% HEPES, 1% pyruvate, 1 non-essential amino acids (all from Biochrom AG Berlin, Germany) at 37 °C, 95% humidity and 7% CO2. Transfection was performed using SuperFect Transfection Reagent (Qiagen GmbH, Hilden, Germany) and 2 µg of plasmid DNA according to the manufacturer's instructions. After 24 hours, the cells were washed with PBS and regrown in the indicated culture medium, with 600 µg/ml G418 for selection (PAA Laboratories GmbH, Pasching, Austria). 16 to 20 days after transfection, resistant cells are seen to predominate. After an additional 7 to 14 days, cells were tested for human CD3 epsilon expression by FACS analysis, according to standard procedures. 2.5x10<5> cells were incubated with UCHT-1 (BD biosciences, Heidelberg, Germany), a human CD3 antibody, 5 µg/ml in PBS with 2% FCS. Antibody binding was detected by affinity chromatography-purified F(ab')2 fragment of goat anti-mouse IgG, specific for the Fc-gamma fragment, conjugated to R-phycoerythrin diluted 1:100 in PBS with 2% FCS (Jackson ImmunoResearch Europe Ltd., Newmarket, Suffolk, UK). Samples were measured on a FACScalibur (BD biosciences, Heidelberg, Germany). Figure 7 shows the expression of natural human CD3 on transfected EL4 cells.

5.2. Cloning and expression of interspecies-specific single-chain antibodies to CD3 as lgG1 antibodies.

[0178] To improve binding detection methods of interspecies specific single chain anti-CD3 antibodies, H2C HLP, A2J HLP and E2M HLP were translated into IgG1 antibodies with mouse IgG1 and human lambda constant regions. Sequences. cDNAs encoding the heavy and light chains of the corresponding IgG antibodies were obtained by gene synthesis according to standard procedures. The synthesis of the gene fragment was designed to first have the Kozak site required for eukaryotic expression of the construct, followed by the 19 amino acids of the immunoglobulin head peptide (SEQ ID NOS: 364 and 363) and then, in the proper reading frame, the coding sequence of the corresponding heavy chain variable region or the corresponding light chain variable region, followed, in the same reading frame, by the coding sequence of the mouse constant region. lgG1 (SEQ ID Nos. 366 and 365) or the human lambda light chain constant region coding sequence (SEQ ID Nos. 368 and 367). Restriction sites were introduced at the beginning and end of the cDNA for the fusion protein. EcoRI restriction sites at the 5' end and SalI at the 3' end were used for the following cloning procedures: Synthesized gene fragments were cloned via EcoRI and SalI sites into a plasmid designated pEF-DHFR (Mack et al. Proc. Natl. Acad. Sci. USA 92 (1995) 7021-7025) for heavy chain constructs and pEF ADA (pEF ADA is described in Raum et al., Cancer immunoi immunother., 50(3), (2001), 141-50) for light chain constructs, by standard procedures. The sequence-verified plasmids were used for transfection in the FreeStvle 293 Expression System (Invitrogen GmbH, Karlsruhe, Germany) according to the manufacturer's instructions. After 3 days of culture, transfectant supernatants were collected and used for the alanine screening experiment.

5.3. Cloning and expression of alanine mutants of human CD3 epsilon for alanine search

[0179] Gene synthesis resulted in 27 cDNA fragments encoding the human CD3 epsilon chain, with one codon changed from the natural human CD3 epsilon sequence to a codon for alanine (GCC), for each of the 1-27 amino acids of the extracellular domain of the mature human GD3 epsilon chain. Except for that altered codon, the cDNA fragments were identical to the aforementioned cDNA fragment for native human CD3. In each construct, only one codon was substituted in relation to the described cDNA fragment for natural human CD3. EcoRI and SalI restriction sites were introduced into the cDNA fragments at identical sites as in the construct with the native fragment. All alanine search constructs were cloned into pEF NEO and sequence-verified plasmids from the transfected EL4 cell. Transfection and selection of transfectants was performed as described. As a result, a panel of expression constructs was obtained in which the first amino acid of the human CD3 epsilon chain, glutamine (Q, Gln) at position 1 was replaced by alanine. The last amino acid replaced by alanine was threonine (T, Thr) at position 27 of mature, wild-type human CD3 epsilon. For each amino acid between glutamine 1 and threonine 27, the corresponding transfectants were generated with a wild-type amino acid substitution to alanine.

5.4. Alanine search experiment

[0180] Chimeric IgG antibodies, described in 2) and interspecies specific single chain antibodies specific for CD3 epsilon were tested in an alanine search experiment. Antibody binding to EL4 cell lines transfected with constructs with alanine mutants of human CD3 epsilon, described in 3) was checked by FACS assay according to standard procedures. 2.5x10<5> cells of the respective transfectants were incubated with 50 ul of cell culture supernatant containing chimeric IgG antibodies or with 50 ul of crude preparations of periplasmically expressed single-chain antibodies. For samples incubated with crude preparations of periplasmically expressed single-chain antibodies, anti-Flag M2 antibody (Sigma-Aldrich Chemie GmbH, Taufkirchen, Germany), 5 µg/ml in 50 µl PBS with 2% FCS, was used as the secondary antibody. For samples incubated with chimeric IgG antibodies, no secondary antibody was required. For all samples, binding of the antibody molecule was detected by affinity chromatography-purified F(ab')2 fragment of goat anti-mouse IgG, specific for the Fc-gamma fragment, conjugated to R-phycoerythrin 1:100 in PBS with 2% FCS (Jackson ImmunoResearch Europe Ltd., Newmarket, Suffolk, UK). Samples were measured on a FACScalibur (BD biosciences, Heidelberg, Germany). Differential binding of chimeric IgG molecules or interspecies-specific single-chain antibodies to EL4 cell lines transfected with alanine mutants of human CD3 epsilon was detected. Either an isotype control or a crude preparation of a periplasmically expressed single-chain antibody of irrelevant specificity was used as a negative control. Antibody UCHT-1 was used as a positive control for expression levels of alanine mutants of human CD3 epsilon. EL4 cell lines transfected with alanine mutants for the amino acids tyrosine at position 15, valine at position 17, isoleucine at position 19, valine at position 24, or leucine at position 26 of the mature CD3 epsilon chain were not evaluated due to low expression levels (data not shown). The binding of interspecies-specific single-chain antibodies and single-chain antibodies in the form of chimeric IgG to EL4 cell lines transfected with alanine mutants of human CD3 epsilon is shown in Figure 8 (A-D) as relative binding in arbitrary units where the geometric mean fluorescence of the corresponding negative controls is subtracted from each geometric mean fluorescence of the sample. To compensate for differences in expression levels, all sample values of some transfectants were then divided by the geometric mean UCHT-1 antibody values for the respective transfectants. To compare the specificity with the value of the wild-type sample, all sample values of the respective specificities were finally divided by the value for the wild-type sample. Thus, the value of the natural sample is arbitrarily taken as 1 arbitrary binding unit.

[0181] Calculations according to the following formula were used:

[0182] In this equation, value_ Sample denotes a value in arbitrary binding units indicating the degree of binding of a specific anti-CD3 antibody to a specific alanine mutant as shown in Figure 8 (A-D), Sample denotes the geometric mean of the fluorescence value obtained for the specific anti-CD3 antibody tested (analyzed) on the specific alanine-scanning transfectant, neg_ contr. denotes the geometric mean of the fluorescence value obtained for negative control testing on the specific alanine mutant, UCHT-1 denotes the geometric mean of the fluorescence value obtained for the UCHT-1 antibody tested for the specific alanine mutant, WT denotes the geometric mean of the fluorescence value obtained for the specific anti-CD3 antibody tested on the wild-type transfectant, x denotes the corresponding transfectant, y denotes the corresponding anti-CD3 antibody and wt denotes the corresponding wild-type transfectant.

[0183] As seen in Figure 8 (A-D), IgG antibody A2J HLP shows a complete loss of binding to the amino acids asparagine at position 4, threonine at position 23 and isoleucine at position 25 of the mature CD3 epsilon chain. A complete loss of IgG antibody binding to A2J HLP was observed for the amino acids glutamine at position 1, aspartate at position 2, glycine at position 3, and glutamate at position 5 of the mature CD3 epsilon chain. The IgG antibody E2M HLP shows a complete loss of binding to the amino acids asparagine at position 4, threonine at position 23 and isoleucine at position 25 of the mature CD3 epsilon chain. The IgG antibody E2M HLP shows a complete loss of binding to the amino acids glutamine at position 1, aspartate at position 2, glycine at position 3, and glutamate at position 5 of the mature CD3 epsilon chain. The H2C HLP IgG antibody shows moderate loss of binding to the amino acids asparagine at position 4 of the mature CD3 epsilon chain and shows complete loss of binding to the amino acids glutamine at position 1, aspartate at position 2, glycine at position 3, and glutamate at position 5 of the mature CD3 epsilon chain. The F12Q HLP single chain antibody shows a complete loss of binding to the amino acids glutamine at position 1, aspartate at position 2, glycine at position 3 of the mature CD3 epsilon chain to glutamate at position 5 of the mature CD3 epsilon chain.

6. Analysis of binding interpscies specific anti-CD3 binding molecules H2C HLP to

human CD3 epsilon chain with and without the N-terminal His6 tag transfected into the mouse T-cell line EL4

6.1. Cloning and expression of the human CDS epsilon chain with N-terminal six

histidine tag (His6 tag)

[0184] A cDNA fragment encoding the human CD3 epsilon chain with an N-terminal His6 tag was obtained by gene synthesis. The synthesis of the gene fragment is labeled so that it first has the Kozak site required for eukaryotic expression of the construct, followed in the correct reading frame by the coding sequence of the 19 amino acid immunoglobulin sequence peptide, followed by the coding sequence of the His6 tag followed in the correct reading frame by the coding sequence of the mature human CD3 epsilon chain (the cDNA and amino acid sequences of the construct are given as SEQ ID NOs 380 and 379). The gene fragment synthesis was also designed to contain restriction sites at the beginning and end of the cDNA. The introduced restriction sites EcoRI at the 5' end and SalI at the 3" end were used in the following cloning procedures. The synthesis of the gene fragment was then cloned via EcoRI and SalI into a plasmid designated as pEF-NEO (as previously described) according to standard protocols. The plasmid of confirmed sequence was used to transfect the EL4 mouse T cell line. Transfection and selection of transfectants were performed as previously described. After 34 days, culture cells, the transfectants were used for the test described below.

6.2. Binding of the interspecies-specific anti-CD3 binding molecule H2C HLP to the human CD3 epsilon chain with and without the N-terminal His6 tag

[0185] A chimeric IgG antibody with binding specificity for H2C HLP specific for CD3 epsilon was tested for binding to human CD3 epsilon with or without an N-terminal His6 tag. Antibody binding to EL4 cell line transfected with His6-human CD3 epsilon and wild-type human CD3 epsilon respectively was tested by FACS assay according to standard protocols. 2.5x10<5> cells of the transfectants were incubated with 50 ul of cell culture supernatant containing the chimeric IgG antibody or 50 ul of the corresponding control antibody at a concentration of 5 ug/ml in PBS with 2% FCS. The corresponding isotype control was used as a negative control, and the CD3 specific antibody UCHT-1 was used as a positive control for expressing the construct. Bound antibody was detected by affinity chromatography-purified F(ab')2 fragment goat anti-mouse IgG, specific for the Fc-gamma fragment, conjugated to R-phycoerythrin 1:100 in PBS with 2% FCS (Jackson ImmunoResearch Europe Ltd., Newmarket, Suffolk, UK). Samples were measured on a FACSCalibur (BD biosciences, Heidelberg, Germany). A comparison was made with an EL4 cell line transfected with wild-type human CD3 epsilon, where a clear loss of binding of the chimeric IgG with H2C HLP binding specificity for human-CD3 epsilon with an N-terminal His6 tag was detected. These results indicate that the free N-terminus of CD3 epsilon is crucial for the binding of interspecies-specific anti-CD3 with the binding specificity of H2C HLP to the human CD3 epsilon chain (Figure 9).

7. Determination of binding constant KD of bispecific single-chain antibody

interspecies-specific EGFR primate CD3 (EGFR LH x H2C HLP) primates for fusion

protein 1-27 CD3-Fc using Surface Plasmon Resonance measurements in relation to

binding to CD3 expressing PBMC as measured by fluorescent sorting

cell (eng. Fluorescence Activated Cell Sorter (FACS)

7.1. Measurement of surface plasmon resonance (Plasmon Surface Resonance

Measurement

[0186] To determine the binding affinity of fully interspecies-specific bispecific single-chain antibodies EGFR-21-63 LH x H2C HLP to amino acids 1-27 of the N-terminus of the human CD3 epsilon chain, surface plasmon resonance (SPR-cyanostics) measurements were performed with a recombinant fusion protein consisting of N-terminal amino acids 1-27 of the mature human CD3 epsilon chain fused with by the Fc part of human lgG1 (1-27 CD3-Fc). At this end, a Biacore Carboxymethyl-DextranCM5 chip (Biacore, Uppsala, Sweden) was installed on a Biacore 2000® system (Biacore, Uppsala, Sweden). One flow cell was activated by the addition of a solution of N-(3-Dimesylaminopropyl)-N-ethylcarbodiimide hydrochloride/N-Hydroxysuccinimide according to standard procedures. Solution. of the 1-27 CD3-Fc fusion protein was then added creating a stable covalent bond of the protein to the dextran layer of the Biacore chip. Unbound protein was removed by intensive washing followed by blocking of unreacted remaining NHS-activated carboxy groups by addition of ethanolamine solution. The success of protein coupling was confirmed by a stronger signal measured as Response Units compared to the signal before coupling. The reference cell was prepared as described but without the addition of the protein solution.

[0187] Purified bispecific antibody EGFR-21-63 LH x H2C HLP was extensively dialyzed against HBS-EP buffer (Biacore, Uppsala, Svveden) in a Slide-A-Lyzer® Mini Dialysis Unit (Pierce, Rockford-II, USA). Protein concentration after dialysis was determined by absorbance at UV280 nm giving a concentration of 43 µg/ml.

[0188] The protein solution was transferred to a 96-well plate and serially diluted by adding 1:1 HBS-EP buffer to the next 10 wells.

[0189]Surface plasmon resonance measurements were performed by taking individual samples from all 11 deposits. Flow cells were regenerated by addition of acetate buffer between measurements to release bound protein.

[0190] The binding signals of the bispecific antibody molecules were obtained by subtracting the signal originating from the reference cell from the signal obtained by measuring the cell conjugated with 1-27 CD3-Fc protein. Association and dissociation curves were measured as Response Units and recorded. Binding constants were calculated using Biacore® curve fitting software based on the Langmuir model.

[0191] The calculated KD binding constants from the first five concentrations were determined to be 1.52 x10"<7>M.

7.2. Determination of CD3 binding constant by FACS measurement

[0192] In order to test the interspecies-specific binding affinity of the bispecific antibody molecules in relation to the strength of binding to native human CD3, an additional FACS binding saturation analysis was performed. The selected EGFR-21-63 LH x H2C HLP bispecific antibody molecule was used to prepare a series of serial dilutions with a factor of 1:1.5 and a starting concentration of 63.3 ug/ml. Bispecific antibody molecules were incubated at these different concentrations with 1.25 x 10<5>human PBMCs each sample for 1 h at 4°C followed by two washing steps in PBS at 4°C. Detection of bound bispecific antibody molecules was performed using Penta-His antibody (Oiagen GmbH, Hildesheim, Germany) at a concentration of 5 µg/ml in 50 µl of PBS with 2% FCS. After incubation for 45 minutes at 4°C and two washing steps, Penta-His antibody binding was detected by affinity chromatography-purified F(ab')2 fragment of goat anti-mouse IgG, specific for the Fc-gamma fragment, conjugated to R-phycoerythrin diluted 1:100 in PBS with 2% FCS (Jackson ImmunoResearch Europe Ltd., Newmarket, Suffolk, UK). Flow cytometry was performed on a FACS-Canto II machine, FACS Diva software was used for data collection and analysis (Becton Dickinson biosciences, Heidelberg). FACS staining and fluorescence intensity measurements were performed as described in Current Protocols in Immunology (Coligan, Kruisbeek, Margulies, Shevach and Strober, Wiley-lnterscience, 2002). The mean values of the obtained fluorescence intensity were plotted as a function of the concentration of the bispecific antibody molecule used and analyzed by biomathematical software in the form of one-sided binding analysis (hyperbola). The software calculated the corresponding KD values that describe the binding of the ligand (bispecific antibody molecule) to the receptor (subtraction of CD3 positive PBMC) that follows the law of mass action. The basic formula is as follows: Y = Bmax x X / (Kd+X) where Bmax is the maximum binding. KD is the ligand concentration required to reach half maximal binding. FACS staining was performed in duplicate, R<2> values were better than 0.95.

[0193] The determined half maximal binding to the molecule of the bispecific antibody EGFR-21-63 LH x H2C HLP was achieved at a concentration of 8472 ng/ml which corresponds to 154 nM (1.54 x 10"<7>M) given a molecular weight of 55000 Daltons (Figure 10). Thus, the affinity of EGFR-21-63 LH x H2C HLP for the N-terminal amino acids 1-27 of the human CD3 epsilon chain separated from its native CD3-context was shown to equal the affinity of EGFR-21-63 LH x H2C HLP for native CD3 of intact T cells.

8. Creation of CHO cells transfected with human EGFR

[0194] The human EGFR-positive cell line, A431 (epidermoid carcinoma cell line, CRL-1555, American Type Culture Collection, Rockville, MD) was used to obtain total RNA which was isolated according to the instructions provided in the kit (Oiagen, RNeasy Mini Kit, Hilden, Germany). The obtained RNA was used for random reverse transcription cDNA synthesis. The following oligonucleotides were used to clone the entire sequence of the human EGFR antigen:

[0195] The coding sequence was amplified by PCR (denaturation at 94°C for 5 min, renaturation at 58°C for 1 min, extension at 72°C for 2 min for the first cycle; denaturation at 94°C for 1 min, renaturation at 58°C for 1 min, extension at 72°C for 2 min for 30 cycles; terminal extension at 72°C for 5 min). The PCR product was then digested with XbaI and SalI, ligated to appropriately processed pEF-DHFR expression vectors (Raum et al., Cancer Immunol. Immunother. 2001; 50:141-150), and transformed into E. coli. The previously described procedure was performed according to standard protocols (Sambrook, Molecular Cloning; A Laboratory Manual, 3rd edition, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, New York (2001)). The clone whose sequence was verified (SEQ ID 370, Amino acid sequence SEQ ID 369) was transfected with DHFR-deficient CHO cells, constructed for eukaryotic expression. Eukaryotic protein expression in CHO cells lacking DHFR was performed as described by Kaufmann R.J. (1990) Methods Enzymol. 185, 537-566. Gene amplification of the construct was induced by increasing concentrations of methotrexate (MTX) to a final concentration of 20 nM

MTX.

9. Generation of CHO cells expressing the extracellular domain of cynomolgus EGFR

[0196] The cynomolgus EGFR extracellular domain cDNA sequence was obtained using a set of two PCRs on cynomolgus monkey colon cDNA (Cat#: C1534090-Cy-BC; obtained from BioCat GmbH, Heidelberg, Germany) under the following reaction conditions: 1 cycle at 94°C for 3 minutes followed by 35 cycles at 94°C for 1 minute. 53°C for 1 minute and 72°C for 2 minutes followed by a final cycle of 72°C for 3 minutes. The following primers were used:

[0197] Those PCRs yielded two overlapping fragments (A: 1-869, B: 848-1923), which were isolated and sequenced according to standard protocols using PCR primers, thus forming a 1923 bp portion of the cynomolgus EGFR cDNA sequence from the third nucleotide codon +1 of the mature protein to the 21<st>codon of the transmembrane domain. To generate the cynomolgus EGFRa expression construct, a cDNA fragment was obtained by gene synthesis according to standard protocols (the cDNA and amino acid sequence of the construct is given as SEQ ID Nos 372 and 371). In this construct, the cynomolgus EGFR coding sequence from amino acids +2 to +641 of the mature EGFR protein is fused to the human EGFR coding sequence by replacing the coding sequence of amino acids +2 to +641. The gene fragment synthesis was also designed to first have the Kozak site required for eukaryotic expression of the construct and restriction sites at the start and end of the cDNA encoding essentially the extracellular domain of cynomolgus EGFR fused to the transmembrane and intracellular domains of human EGFR. Furthermore, a conservative mutation was introduced at amino acid 627 (the 4<lh>amino acid of the transmembrane domain) mutating valine to leucine, yielding a restriction site (Sphl) for cloning purposes. The introduced restriction sites XbaI at the 5' end and Saii at the 3' end were used in the following cloning procedures. The gene fragment synthesis was then cloned via XbaI and SalI into a plasmid designated pEF-DHFR (pEF-DHFR is described in Mack et al. Proc. Natl. Acad. Sci. USA 92 (1995) 7021-7025). A sequence-confirmed clone of this plasmid was used to transfect CHO/dhfr- cells as previously described.

10. Creation of EGFRiCD3 species-specific bispecific single-chain molecules

10.1. Cloning of interspecies specific binding molecules

[0198] In general, each of the bispecific monolayer antibody molecules, comprising a domain having interspecies-specific binding specificity for human and non-chimpanzee CD3 epsilon as well as a domain having interspecies-specific binding specificity for human and non-chimpanzee EGFR, are designed as shown in Table 1 below:

[0199] The above constructs containing variable light chains (L) and variable heavy chains (H) of the species-specific human and cynomolgus EGFR domains were obtained by gene synthesis. Syntheses of gene fragments are designed so that it first has the Kozak site required for eukaryotic expression of the construct, then the 19 amino acid immunoglobulin head peptide, followed in the same reading frame by the coding sequence of the respective bispecific single-chain antibody molecule, and then in the same reading frame, by the coding sequence of 6 histidine tags and a stop codon. Gene fragment synthesis was also designed to introduce appropriate N- and C-terminal restriction sites. Synthesis of the gene fragment was cloned across these restriction sites into a plasmid designated pEF-DHFR (pEF-DHFR is described in Raum et al. Cancer Immunol Immunother 50 (2001) 141-150) according to standard protocols (Sambrook, Molecular Cloning; A Laboratory Manual, 3rd edition, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, New York (2001)). The clone with the determined nucleotide sequence was transfected into Chinese hamster ovary (CHO) cells without dihydrofolate reductase (DHFR) for eukaryotic expression of the construct.

[0200] CHO-cells without DHFR (ATCC No. CRL 9096) were transfected with constructs stably or transiently, by electroporation or alternatively, HEK 293 (engl. human embryonal kidney cells, ATCC Number: CRL-1573) transiently, according to standard protocols.

10.2. Expression and purification of bispecific single-chain antibody molecules

[0201] Bispecific single-chain antibody molecules were expressed in Chinese hamster ovary (CHO) cells. Expression of eukaryotic proteins in DHFR-null CHO cells was performed as described by Kaufmann RJ. (1990) Methods Enzvmol. 185, 537-566. Gene amplification of the construct was induced by increasing the final concentrations of MTX to 20 nM. After two passages

stationary cultures, cells were grown in rotary flasks in nucleoside-free HyQ PF CHO liquid soy medium (with 4.0 mM L-Glutamine with 0.1% Pluronic F - 68; HvCIone) for 7 days before harvesting. Cells were removed by centrifugation and the supernatant containing the expressed protein was stored at -20°C. Alternatively, the constructs were transiently expressed in HEK 293 cells. Transfection was performed with 293fectin reagent (Invitrogen, #12347-019) according to the manufacturer's protocol.

[0202] Akta® Explorer Systems (GE Health Systems) and Unicom® Software were used for chromatography. Metal chelate chromatography ("IMAC") was performed using Fractogel EMD chelate® (Merck) coated with ZnCl 2 according to the manufacturer's protocol. The column was equilibrated with buffer A (20 mM sodium phosphate buffer pH 7.2, 0.1 M NaCl) and the cell culture supernatant (500 ml) was applied to the column (10 ml) at a flow rate of 3 ml/min. The column was washed with buffer A to remove unbound sample. Bound protein was eluted using a two-step gradient of buffer B (20 mM sodium phosphate buffer pH 7.2, 0.1 M NaCl, 0.5 M Imidazole) as:

Step 1: 20% buffer B in 6 column volumes

Step 2: 100% buffer B in 6 column volumes

[0203] Fractions of the eluted protein from step 2 were collected for further purification. All chemicals were of research grade and purchased from Sigma (Deisenhofen) or Merck (Darmstadt).

[0204] Gel filtration chromatography was performed on a HiLoad 16/60 Superdex 200 prep grade column (GE/Amersham) equilibrated with Equibuffer (25 mM Citrate, 200 mM Lysine, 5% Glycerol, pH 7.2). Eluted protein samples (flow rate 1 ml/min) were subjected to standard SDS-PAGE and Western Blot for detection. Before purification, the column was calibrated for molecular weight determination (molecular weight marker kit, Sigma MW GF-200). Protein concentrations were determined by OD280 nm.

[0205] Purified bispecific single-chain antibody protein was analyzed by SDS PAGE under reducing conditions performed on precast 4-12% Bis Tris gels (Invitrogen). Sample preparation and application were performed according to the manufacturer's protocol. Molecular weight was determined with MultiMark protein standard (Invitrogen). The gel was stained with colloidal Coomassie (Invitrogen protocol). The purity of the isolated protein was >95% as determined by SDS-PAGE.

[0206] The bispecific single chain antibody has a molecular weight of about 52 kDa under native conditions as determined by gel filtration in PBS. All constructs were purified according to this method.

[0207] VVestem Blot was performed using Optitran® BA-S83 membrane and Invitrogen Blot Module according to the manufacturer's protocol. Antibodies were directed against His Tag (Penta His, Qiagen) and goat-anti-mouse Ig labeled with alkaline phosphatase (AP) (Sigma), and BCIP/NBT (Sigma) as substrate. A single band was detected at 52 kD corresponding to the purified bispecific single chain antibody.

11. Determination of the binding constant of fully interspecies-specific bispecifics

of single-chain antibodies for the fusion protein 1-27 CD3-Fc by surface resonance measurements

plasmon

[0208] To determine the binding affinity of molecules of bispecific single-chain antibodies interspecies-specific for primate EGFR and primate CD3 for amino acids 1-27 of the N-terminus of the mature human CD3 epsilon chain, surface plasmon resonance measurements were performed with a recombinant fusion protein consisting of N-terminal amino acids 1-27 of the human CD3 epsilon chain fused to the Fc-part of human IgG1 (1-27 CD3-Fc). For this purpose, a Biacore Carboxymethyl-Dekstran CM5 chip (Biacore, Uppsala, Svveden) was installed on a Biacore 2000® system (Biacore, Uppsala, Svveden). The flow cell was activated with a solution of N-(3-Dimethylaminopropyl)-N'-ethylcarbodiimide hydrochloride/N-Hydroxysuccinimide according to standard procedures. A solution of the 1-27 CD3-Fcje fusion protein is then added to form a stable covalent bond of the protein to the dextran layer of the Biacore chip. Unbound protein was removed by extensive washing followed by blocking of the remaining unreacted NHS-activated carboxy groups by addition of ethanolamine solution. The success of protein splicing is confirmed by the detection of a higher signal measured as a response unit compared to the signal before splicing. The reference cell was prepared as described but without the addition of protein solution.

[0209] Purified single-chain bispecific antibodies given below were adjusted to a concentration of 5 µg/ml by adding ith HBS-EP buffer (Biacore, Uppsala, Svveden) and transferred to a 96-well plate, each of the antibodies in a volume of 150 µl.

[0210] Surface plasmon resonance measurements were performed for all samples and flow cells were regenerated by addition of acetate buffer between measurements to release bound protein (all according to standard protocols).

[0211] The binding signals of the bispecific single chain antibodies were obtained by subtracting the reference cell signal from the signal obtained by measuring cells conjugated with 1-27 CD3-Fc protein.

[0212] The association and disassociation curves were measured as response units and recorded. Binding constants were calculated using Biacore® curve-fitting software, which is based on the Langmuir model. The calculated affinities of the tested fully interspecies-specific bispecific single-chain molecules for N-terminal amino acids 1-27 of human CD3 epsilon are given as KD values and range from 2.54x10"<6>M to 2.49x10"<7>M. "LH" refers to the arrangement of the variable domains as VL-VH. "HL" refers to the arrangement of the variable domains in as VH-VL. G4H, F70, A2J, E1L, E2M, H 1 E and F6A refer to different interspecies specific CD3 binding molecules. 12. Analysis of the binding of EGFR and CD3 interspecies-specific bispecific antibodies,

flow cytometry

[0213] In order to test the functionality of the interspecies-specific Bispecific antibody construct in terms of binding capacity to human and cynomolgus EGFR and CD3, respectively, FACS analysis was performed. For this purpose CHO cells transfected with human EGFR as described in Example 8 and human CD3 positive T cells of the leukemic cell line HPB-ALL (DSMZ, Braunschweig, ACC483) were used to test binding to human antigens. Binding reactivity to cynomolgus antigens was tested using the cynomolgus EGFR transfectant obtained as described in Example 9 and the 4119LnPx macaque T cell line (obtained by courtesy of Prof Fickenscher, Hvgiene Institute, Virology, Erlangen-Nuemberg; published in Knappe A, et al., and Fickenscher H., Blood 2000, 95, 3256-61) 200,000 cells of the respective cell population were incubated for 30 min on ice with 50 µl of purified interspecies-specific bispecific antibody construct protein (2 µg/ml). Alternatively, the supernatant of a cell culture transiently producing proteins was used. Cells were washed twice in PBS and binding of the construct was detected with mouse Penta His antibody (Oiagen; diluted 1:20 in 50 µl PBS with 2% FCS). After washing, bound anti-His antibodies were detected with Fc gamma-specific antibodies (Dianova) conjugated to phycoerythrin diluted 1:100 in PBS with 2% FCS. Fresh growth medium was used as a negative control.

[0214] Flow cytometry was performed on a FACS-Calibur machine, CellOuest software

used for data performance and analysis (Becton Dickinson biosciences, Heidelberg). FACS staining and fluorescence intensity measurements were performed as described in Current Protocols in Immunology (Coligan, Kruisbeek, Margulies, Shevach and Strober, Wiley-lnterscience, 2002).

[0215] The binding ability of several bispecific single-chain molecules that are specific for EGFR and interspecies specific for human and non-chimpanzee CD3 were clearly detected as shown in Figure 11. In FACS analysis, all constructs showed binding to CD3 and EGFR relative to culture medium and first and second antibody detection as negative controls. Interspecies specificities of bispecific antibodies for human and cynomolgus CD3 and EGFR antigens are presented.

13. Bioactivity of EGFR and CD3 interspecies specific bispecific single chain

antibodies

[0216] The bioactivity of the obtained bispecific single-chain antibodies was analyzed by chromium 51 (<51>Cr) released in an in vitro cytotoxicity test using the EGFR positive cell line described in Examples 8 and 9. As effector cells, stimulated CD8 positive human T cells or T cell line 4119LnPx macaques were used.

Stimulated CD8+ T cells were obtained as follows:

[0217] Petri dishes (145 mm diameter, Greiner) were precoated with a commercially available anti-CD3 specific antibody at a final concentration of 1 µg/ml for 1 h at 37°C. Unbound protein was removed by a single washing step with PBS. Fresh PBMCs were isolated from peripheral blood (30 - 50 ml human blood) by Ficoll gradient centrifugation according to standard protocols. 3 - 5 x 10<7>PBMCs were added to a pre-coated petri dish in 120 ml RPMI 1640 /10% FCS / IL-2 20 U/ml (Proleukin, Chiron) and stimulated for 2 days. On day 3, cells were harvested, washed once with RPM11640. IL-2 was added at a final concentration of 20 U/ml and cultured again for one day. CD8+ cytotoxic T lymphocytes (CTLs) were isolated by depletion of CD4+ T cells and CD56+ NK cells.

[0218] Target cells were washed twice with PBS and labeled with 11.1 MBq<51>Cr in a final volume of 100ul RPMI with 50% FCS for 45 minutes at 37°C. Subsequently labeled target cells were washed 3 times with 5 ml RPMI and then used in the cytotoxicity assay. The test/assay was performed on a 96-well plate in a total volume of 250 µl to which RPMI (as above) was added in an E:T ratio of 10:1. 1 ug/ml, interspecies-specific bispecific single-chain antibody molecules and their 20-fold dilutions were used. Alternatively, the cell culture supernatant of peolase-derived proteins was serially diluted in 1:2 steps. The test time was 18 h and cytotoxicity was measured as the relative values of released chromium in the supernatant in relation to the difference between maximal lysis (with the addition of Triton-X) and spontaneous lysis (without effector cells). All measurements were performed in quadruplicate. Measurements of nrom activity in the supernatants were performed with a Wizard 3" gammacounter (PerkinElmer Life Sciences GmbH, Cologne, Germany). Analysis of experimental data was performed with Prism 4 for Windows (version 4.02, GraphPad Software Inc., San Diego, California, USA). Sigmoidal dose-response curves typically have R<2> values of >0.90 as determined by the software. EC50 values calculated by the analytical program were used to compare bioactivity.

[0219] As shown in Figures 12 and 13, all of the resulting interspecies-specific bispecific single-chain antibody constructs exhibit cytotoxic activity against human CD8+ cell-challenged human EGFR-positive target cells and cynomolgus EGFR-positive target cells challenged with the 4119LnPx macaque T cell line. A bispecific single-chain antibody with different target specificities was used as a negative control.

14. Cloning and expression of C-terminal, transmembrane and shortened extracellular

human MCSP domain

[0220] The coding sequence of the C-terminal, transmembrane and truncated extracellular domain of human MCSP (amino acid 1538 - 2322) was obtained by gene synthesis according to standard protocols (cDNA sequence and amino acid sequence of the recombinant construct for the expression of the C-terminal, transmembrane and truncated extracellular domain of human MCSP (designated as human D3 and human D3) under SEQ ID NOs 374 and 373). The gene synthesis fragment was designed so that it first has the Kozak site required for eukaryotic expression of the construct, then the coding sequence of 19 amino acids of the immunoglobulin head peptide, then in the reading frame the FLAG marker, and then in the same reading frame a sequence containing several restriction sites for cloning and coding for 9 amino acid artificial linkers (SRTRSGSOL), then in the same reading frame the coding sequence of the C-terminal, transmembrane and of truncated human MCSP extracellular domains and a stop codon. Restriction sites are introduced at the beginning and at the end of the DNA fragment. Restriction sites EcoRI at the 5' end and SalI at the 3' end were used in the following cloning procedures. The fragment was digested with EcoRI and SalI and cloned into pEF-DHFR (pEF-DHFR is described in Mack et al. Proc. Natl. Acad. Sci. USA 92 (1995) 7021-7025) according to standard protocol. A plasmid of verified sequence was used for transfection of CHO/dhfr- cells (ATCC No. CRL 9096). Cells were grown in glutamine-stabilized RPMI 1640 supplemented with 10% FCS, 1% penicillin/streptomycin (all from Biochrom AG Berlin, Germany) and nucleoside from cell culture reagent stock solution (Sigma-Aldrich Chemie GmbH, Taufkirchen, Germany) to a final concentration of 10 µg/ml Adenosine, 10 µg/ml Deoxyadenosine and 10 µg/ml Thymidine, in incubator at 37°C, 95% humidity and 7% C02. Transfection was performed using PolyFect Transfection Reagent (OjagenGmbH, Hilden, Germany) and 5 µg of plasmid DNA according to the manufacturer's instructions. After culturing for 24 h, cells were washed once with PBS and recultured in RPMI 1640 with stabilized glutamine and 1% penicillin/streptomycin. Thus, the medium does not contain nucleosides and this part was applied to the transfected cells. Approximately 14 days after transfection, an increase in resistant cells was observed. After an additional 7 to 14 days, transfectants were tested for construct expression by FACS analysis. 2.5x10<5> cells were incubated with 50 µl of anti-Flag-M2 antibody (Sigma-Aldrich Chemie GmbH, Taufkirchen, Germany) diluted to 5 µg/ml in PBS with 2% FCS. Bound antibody was detected by affinity chromatography-purified F(ab')2 fragment of goat anti-mouse IgG, specific for the Fc-gamma fragment, conjugated to R-phycoerythrin specially diluted 1:100 in PBS with 2% FCS (ImmunoResearch Europe Ltd., Newmarket, Suffolk, UK). Samples were measured on a FACScalibur (BD biosciences, Heidelberg, Germany).

15. Cloning and expression of C-terminal, transmembrane-truncated extracellular

MCSP macaque domain

[0221] The cDNA sequence of the C-terminal, transmembrane and truncated extracellular domains of macaque MCSP (designated macaque D3) was obtained using a set of three PCRs on macaque monkey skin cDNA (Cat No. C1534218-Cy-BC; BioCat GmbH, Heidelberg, Germany) under the following reaction conditions: 1 cycle at 94°C, 3 min., 40 cycles at 94°C for 0.5 min., 52°C for 0.5 min. i72"C for 1.75 min., terminal cycle at 72°C for 3 min. The following primers were used:

[0222] These PCRs yielded three overlapping fragments (A: 1-1329, B: 1229-2428, C: 1782-2547)

which were isolated and sequenced according to standard protocols using PCR primers and thus obtained portions of 2547 bp of macaque MCSP cDNA sequence (the cDNA sequence and amino acid sequence of this portion of macaque MCSP are listed as SEQ ID NOs 376 and 375) from 74 bp upstream of the C-terminal domain coding sequence to 121 bp downstream of the stop codon. The following PCR uses the following reaction conditions: 1 cycle at 94°C for 3 min, 10 cycles at 94°C for 1 min, 52°C for 1 min and 72°C for 2.5 min, a terminal cycle at 72°C for 3 min was applied to connect the PCR products of the above reactions A and B. The following primers were used:

[0223] Primers for this PCR were designed to introduce restriction sites at the beginning and end of the cDNA fragment encoding the C-terminal, transmembrane, and truncated extracellular domains of macaque MCSP. The introduced restriction sites MpheI at the 5' end and SalI at the 3' end were used in the following cloning procedures. The PCR fragment was then cloned via Mfel and Sali into a Bluescript plasmid containing the EcoRI/Mfel fragment of the aforementioned pEF-DHFR plasmid (pEF-DHFR is described in Raum et al. Cancer Immunol Immunother 50 (2001) 141-150) by replacing the C-terminal, transmembrane and truncated extracellular domains of human MCSP. Synthesis of a gene fragment containing the coding sequences of the frontal peptide immunoglobulin and a Flag marker as well as an artificial linker (SRTRSGSOL) in frame with the 5' end of a cDNA fragment encoding the C-terminal, transmembrane and truncated extracellular domains of macaque MCSP. This vector was used to transfect CHO/dhfr- cells (ATCC No. CRL 9096). Cells were grown in RPMI 1640 with stabilized glutamine supplemented with 10% FCS, 1% penicillin/streptomycin (all from Biochrom AG Berlin, Germany) and nucleosides from cell culture reagent stock solution (Sigma-Aldrich Chemie GmbH, Taufkirchen, Germany) to a final concentration of 10 µg/ml Adenosine, 10 µg/ml Deoxyadenosine and 10 µg/ml Thymidine in an incubator. at 37°C, 95% humidity and 7% C02. Transfection was performed with PolyFect Transfection Reagent (Ojagen GmbH, Hilden, Germany) and 5 µg of plasmid DNA according to the manufacturer's instructions. After growing for 24 h, cells were washed once with PBS and grown again in RPM11640 with stabilized glutamine and 1% penicillin/streptomycin. Thus, the culture medium does not contain nucleosides and, accordingly, a fraction is applied only to the transfected cells. About 14 days after transfection, the growth of resistant cells was observed. After another 7 to 14 days, the transfectants were tested for expression of the recombinant construct via FACS. 2.5x10<5> cells were incubated with 50 µl of anti-Flag-M2 antibody (Sigma-Aldrich Chemie GmbH, Taufkirchen, Germany) diluted to 5 µg/ml in PBS with 2% FCS. R-phycoerythrin, diluted to 1:100 in PBS with 2% FCS (Jackson ImmunoResearch Europe Ltd., Newmarket, Suffolk, UK). Samples were measured on a FACScalibur (BD biosciences, Heidelberg, Germany).

16. Obtaining and characterization of MCSP and CD3 interspecies-specific bispecific single-chain molecules

[0224] Each of the bispecific single-chain antibody molecules containing the CD3 epsilon domain interspecies specific for human and non-chimpanzee as well as the binding domain MCSP interspecies specific for human and non-chimpanzee were designed as given in the following Table 2:

[0225] The above constructs containing the variable heavy chain (VH) and variable light chain (VL) domains interspecies specific for human and macaque MCSP D3 and the VH and VL domains interspecies specific for human and macaque CD3 were obtained by gene synthesis. The synthesis of the gene fragment is designed so that it first has the Kozak site required for eukaryotic expression of the construct, then the 19 amino acid immunoglobulin head peptide, then in the same reading frame the coding sequence of the corresponding bispecific single-chain antibody molecule, and then in the same reading frame the coding sequence of the histidine6-tag and the stop codon. Synthesis of the gene fragment was also designed to introduce suitable N- and C-terminal restriction sites. Synthesis of the gene fragment was cloned through these restriction sites into a plasmid designated as pEF-DHFR (pEF-DHFR is described in Raum et al. Cancer Immunol Immunother 50 (2001) 141-150) according to a standard protocol (Sambrook, Molecular Cloning; A Laboratory Manual, 3rd edition, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, New York (2001)). Different CHO cells without DHFR were transfected with the constructs, stably or transiently (ATCC No. CRL 9096) by electroporation or alternatively, HEK 293 cells (engl. human embryonic kidney cells, translation: human embryonic kidney cells, ATCC Number: CRL-1573) transiently, according to standard protocols.

[0226] Eukaryotic protein expression in DHFR-null CHO cells was performed as described by Kaufmann RJ. (1990) Methods Enzvmol. 185, 537-566. Gene amplification of the constructs was induced by the addition of increasing concentrations of methotrexate (MTX) up to a final concentration of 20 nM MTX. After two passages of stationary culture, cells were grown in spinner flasks in nucleoside-free liquid soy medium HyQ PF CHO.(with 4.0 mM L-Glutamine with 0.1% Pluronic F-68; HvCIone) for 7 days before collection. Cells were removed by centrifugation and the supernatant containing the expressed protein was left at -20°C.

[0227] Akta® Explorer Systems (GE Health Systems) and Unicom® Software were used for chromatography. Metal chelate chromatography (immobilized metal affinity chromatography; "IMAC") was performed using Fractogel EMD chelate® (Merck) coated with ZnCl2 according to the manufacturer's instructions. The column was equilibrated with buffer A (20 mM sodium phosphate buffer pH 7.2, 0.1 M NaCl) and cell culture supernatant (500 ml) was applied to the column (10 ml) at a flow rate of 3 ml/min, the column was washed with buffer A to remove unbound sample. Bound protein was eluted using a two-step gradient of buffer B (20 mM sodium phosphate buffer pH 7.2, 0.1 M NaCl, 0.5 M Imidazole) according to the following:

Step 1: 20% buffer B in 6 column volumes

Step 2: 100% buffer B in 6 column volumes

[0228] The eluted protein fractions from step 2 were collected for further purification.. All chemicals were of research grade and purchased from Sigma (Deisenhofen) or Merck (Darmstadt).

[0229] Gel filtration chromatography was performed on a HiLoad 16/60 Superdex 200 prep grade column (GE/Amersham) equilibrated with Equibuffer (25 mM Citrate, 200 mM Lysine, 5% Glycerol, pH 7.2). Eluted protein samples (flow rate 1 ml/min) were subjected to standard SDS-PAGE and Western Blot for detection. Before purification, the column was equilibrated for molecular weight determination (molecular weight marker kit, Sigma MVV GF-200). Protein concentrations were determined by OD280 nm.

[0230] Purified bispecific single-chain antibody protein was analyzed by SDS PAGE under reducing conditions performed on precast 4-12% Bis Tris gels (Invitrogen). Sample preparation and application were performed according to the manufacturer's protocol. Molecular weight was determined with MultiMark protein standard (Invitrogen). The gel was stained with colloidal Coomassie (Invitrogen protocol). the purity of the isolated protein is >95% as determined by SDS-PAGE.

[0231] The bispecific single chain antibody has a molecular weight of about 52 kDa under native conditions as determined by gel filtration in PBS. All constructs were purified according to this method.

[0232] VVestem Blot was performed using Optitran® BA-S83 membrane and Invitrogen Blot Module according to the manufacturer's protocol. Hisd Tag antibody (Penta His, Qiagen) was used to detect the protein of the bispecific single-chain antibody. Goat-anti-mouse Ig labeled with alkaline phosphatase (AP) (Sigma) was used as the secondary antibody and BCIP/NBT (Sigma) as the substrate. A single band was detected at 52 kD corresponding to the purified bispecific single-chain antibody.

[0233] Alternatively, various CHO cells lacking DHFR were transiently transfected. Briefly, 4 x 105 cells per construct were grown in 3 ml of RPM11640 glutamine-stabilized medium supplemented with 10% fetal bovine serum, 1% penicillin/streptomycin, and nucleosides from cell culture reagent stock solution (Sigma-Aldrich Chemie GmbH, Taufkirchen, Germany) to a final concentration of 10 ug/ml Adenosine, 10 ug/ml Deoxyadenosine, and 10 ug/ml Thymidine, in an incubator at 37°C, 95% humidity and 7% CO 2 one day before transfection. Transfection was performed with Fugene 6 Transfection Reagent (Roche, # 11815091001) according to the manufacturer's instructions. 94 µl of OptiMEM medium (Invitrogen) and 6 µl of Fugene 6 were mixed and incubated for 5 min at room temperature. Then, 1.5 µg of DNA per construct was added, the mixture was incubated for 15 minutes at room temperature. Meanwhile, CHO cells without DHFR were washed with 1x PBS and resuspended in 1.5 ml of RPMI 1640 medium. The transfection mixture was diluted with 600 µl of RPM11640 medium, added to the cells and incubated overnight at 37°C, 95% humidity and 7% CO 2 . The day after transfection, the incubation volume of each accession was increased to 5 ml of RPM11640 medium. The supernatant was collected after 3 days of incubation.

17. Analysis of the binding of MCSPiCD3 interspecies-specific bispecific antibodies

flow cytometry

[0234] To test the functionality of the interspecies-specific bispecific antibody constructs for their capacity to bind to MCSP D3 and human and macaque monkey CD3, respectively, FACS analysis was performed. For this purpose, CHO cells transfected with human MCSP D3 (as described in Example 14) and a CD3 positive human HPB-ALL leukemia T cell cell line (DSMZ, Braunschweig, ACC483) were used to test human antigen binding. Binding reactivity to macaque antigens was tested using the resulting macaque MCSP D3 transfectants (described in Example 15) and the macaque T cell line 4119LnPx (obtained by courtesy of Prof. Fickenscher, Hvgiene Institute, Virology, Erlangen-Nuernberg; published in Knappe A, et al., and Fickenscher H., Blood 2000, 95, 3256-61). 200,000 cells of the respective cell lines were incubated for 30 min on ice with 50 ul of purified protein of interspecies-specific bispecific antibody constructs (2 ug/ml) or cell culture supernatant of transfected cells expressing interspecies-specific bispecific antibody constructs. Cells were washed twice in PBS with 2% FCS and binding of the construct was detected using mouse anti-His antibody (Penta His antibody; Oiagen; diluted 1:20 in 50 µl PBS with 2% FCS). After washing, bound anti-His antibodies were detected using picoerythrin-conjugated Fc gamma-specific antibody (Dianova) diluted 1:100 in PBS with 2% FCS. The supernatant of untransfected CHO cells was used as a negative control for binding to T cell lines. A single-chain construct with irrelevant target specificity was used as a negative control for binding to MCSP-D3 transfected CHO cells.

[0235] Flow cytometry was performed on a FACS-Calibur apparatus; CellOuest software was used for data acquisition and analysis (Becton Dickinson biosciences, Heidelberg). FACS staining and fluorescence intensity measurements were performed as described in Current Protocols in Immunology (Coligan, Kruisbeek, Margulies, Shevach and Strober, Wiley-Interscience, 2002).

[0236] Bispecific binding of single-chain molecules previously given; that are interspecies specific for MCSP D3 and interspecies specific for human and macaque CD3 were clearly detected as shown in Figures 14, 15, 16 and 58. In FACS analysis, all constructs showed binding to CD3 and MCSP D3 relative to the respective negative controls. Interspecies specificity of bispecific antibodies for human and macaque CD3 and MCSP D3 antigens has been demonstrated.

18. Bioactivity of MCSPiCD3 interspecies-specific bispecific single-chain antibodies

[0237] As shown in Figures 17 to 21, all obtained constructs of interspecies specific bispecific single chain antibodies reveal cytotoxic activities on MSCP positive human target cells challenged with human CD8+ cells and MSCP positive cynomolgus target cells challenged with T cell line 4119LnPx macaque monkey. A bispecific single-chain antibody with different target specificities was used as a negative control.

19. Stability of MCSP and CD3 interspecies-specific bispecific single-chain antibodies

in plasma

[0238] The stability of the obtained single-chain antibodies in human plasma was analyzed by incubating bispecific single-chain antibodies in 50% human plasma at 37°C and at 4°C for 24 h and subsequent bioactivity testing. Bioactivity was investigated in an in vitro assay/cytotoxicity assay with chromium 51 (<51>Cr) release using an MCSP positive CHO cell line (expressing MCSP as cloned according to Example 14 or 15) as target and stimulated CD8 positive human T cells as effector cells.

[0239] EC50 values calculated by the analytical program as previously described were used to compare the bioactivity of bispecific single-chain antibodies incubated with 50% human plasma for 24 h at 37°C and 4°C, respectively, with bispecific single-chain antibodies without added plasma or mixed with the same amount of plasma immediately before testing.

[0240] As shown in Figure 22 and Table 3, the bioactivity of G4 H-L x I2C H-L, G4 H-L x H2C H-L, and G4 H-L x F12Q H-L bispecific antibodies was not significantly reduced relative to controls with the addition of plasma immediately prior to bioactivity testing.

20. Creation of interspecies-specific bispecific single-chain EGFR and CD3 molecules

man

[0241] Bispecific single-chain antibody molecules with an interspecies-specific binding domain for human and cynomolgus CD3 as well as an interspecies-specific binding domain for human EGFR are given in Table 4 below:

[0242] The above constructs containing the variable heavy chain (VH) and variable light chain (VL) domains interspecies specific for human and cynomolgus EGFR were obtained by gene synthesis. The synthesis of the gene fragment was designed to first have the Kozak site required for eukaryotic expression of the construct, followed by the 19 amino acid immunoglobulin head peptide, followed in the same reading frame by the coding sequence of the respective bispecific monolayer antibody molecule, and then in the same reading frame by the coding sequence of 6 histidine tags and a stop codon.

[0243] Synthesis of the gene fragment was also designed to introduce suitable N- and C-terminal restriction sites. Synthesis of the gene fragment was cloned through these restriction sites into a plasmid designated as pEF-DHFR (pEF-DHFR is described in Raum et al. Cancer Immunol Immunother 50 (2001) 141-150) according to a standard protocol (Sambrook, Molecular Cloning; A Laboratory Manual, 3rd edition, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, New York (2001)). The constructs were stably transfected into CHO cells lacking DHFR (ATCC No. CRL 9096) and certainly obtained and purified as described in Example 10.21. Obtaining interspecies-specific bispecific single-chain molecules EGFR and

CD3 and man

[0244] Bispecific single-chain antibody molecules with an interspecies-specific binding domain for human and cynomolgus monkey CD3 as well as an interspecies-specific binding domain for human EGFR were designed as given in the following Table 5:

[0245] The above constructs containing variable heavy chain (VH) and variable light chain (VL) domains interspecies specific for human and cynomolgus EGFR were obtained by gene synthesis. The synthesis of a gene fragment designed so that it first has the Kozak site required for eukaryotic expression of the construct, then the 19 amino acid immunoglobulin head peptide, followed in the same reading frame by the coding sequence of the respective bispecific single-chain antibody molecule, and then in the same reading frame, the coding sequence of 6 histidine tags and a stop codon.

[0246] Synthesis of the gene fragment was also designed to introduce suitable N- and C-terminal restriction sites. Synthesis of the gene fragment was cloned through these restriction sites

sites into a plasmid designated pEF-DHFR (pEF-DHFR is described in Raum et al. Cancer Immunol Immunother 50 (2001) 141-150) according to standard protocols (Sambrook, Molecular Cloning; A Laboratory Manual, 3rd edition, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, New York (2001)). The constructs were stably transfected into CHO cells lacking DHFR (ATCC No. CRL 9096) and certainly obtained and purified as described in Example 10.

22. Obtaining interspecies-specific bispecific Her2/neui single-chain molecules

Human CD3

[0247] Bispecific single-chain antibody molecules with an interspecies-specific binding domain for human and cynomolgus CD3 as well as an interspecies-specific binding domain for human Her2/neu were designed as given in Table 6 below:

Table 6: Formats of Her2/neui CD3 interspecies-specific bispecific single-chain antibodies

[0248] The above constructs containing the variable heavy chain (VH) and variable light chain (VL) domains interspecies specific for human and cynomolgus HER2/neu were obtained by gene synthesis. The synthesis of the gene fragment was designed so that it first has the Kozak site required for eukaryotic expression of the construct, then the 19 amino acid immunoglobulin head peptide, followed in the same reading frame by the coding sequence of the respective bispecific single-chain antibody molecule, and then in the same reading frame, by the coding sequence of 6 histidine tags and a stop codon.

[0249] Synthesis of the gene fragment was also designed to introduce suitable N- and C-terminal restriction sites. Synthesis of the gene fragment was cloned through these restriction sites

sites into a plasmid designated pEF-DHFR (pEF-DHFR is described in Raum et al. Cancer Immunol Immunother 50 (2001) 141-150) according to standard protocols (Sambrook, Molecular Cloning; A Laboratory Manual, 3rd edition, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, New York (2001)). Various CHO cells lacking DHFR (ATCC No. CRL 9096) were stably transfected with the constructs and also obtained and purified as described in Example 10.

23.1. Generation of CHO cells expressing human HER2

[0250] The human HER2 coding sequence as published in GenBank (accession number X03363) was obtained by gene synthesis according to standard protocols. The gene synthesis fragments were designed to contain the coding sequence of the human HER2 protein, including its leader peptide (the cDNA and amino acid sequence of the construct are provided as SEQ ID Nos. 459 and 460). The gene synthesis fragment is also designed to introduce restriction sites at the beginning and end of the fragment. The introduced restriction sites, XbaI at the 5' end and SalI at the 3' end, were used in the cloning procedures that follow. The gene synthesis fragment was cloned via XbaI and SalI into a plasmid designated pEFDHFR (pEFDHFR is described in Raum et al. Cancer Immunol Immunother 50 (2001) 141-150) according to standard protocols. The foregoing procedures were performed according to standard protocols [Sambrook, Molecular Cloning; A

Laboratory Manual, 3rd edition, Cold Spring Harbor Laboratory Press, Cold Spring Harbor,

New York (2001)).A clone with confirmed nucleotide sequence was transfected into DHFR-deficient CHO cells for eukaryotic expression of the construct. Eukaryotic protein expression in CHO cells lacking DHFR was performed as described by Kaufmann RJ.

(1990) Methods Enzymol. 185, 537-566. Amplification of the construct gene was induced by increasing concentrations of methotrexate (MTX) up to a final concentration of 20 nM MTX.

23.2. Generation of CHO cells expressing the extracellular domain of macaque monkey Her2

[0251] The coding sequence for human Her2 as previously described was modified to encode amino acids 123 to 1038 of the macaque monkey Her2 protein as published in GenBank (accession number XP_001090430). The coding sequence for this chimeric protein was obtained by gene synthesis according to standard chemical protocols (cDNA and amino acid sequence of the construct are given as SEQ ID Nos 461 and 462). The gene fragment synthesis was also designed to first have the Kozak site required for eukaryotic expression of the construct and restriction sites at the beginning and end of the fragment. The introduced restriction sites XbaI at the 5' end and Sal at the 3' end were used in the cloning procedures as follows. Synthesis of the gene fragment was then cloned via XbaI and SalI into a plasmid designated pEFDHFR (pEFDHFR is described in Raum et al. Cancer Immunol Immunother 50 (2001) 141-150). A sequence-confirmed clone of this plasmid was used to transfect transfect CHO/dhfr- cells as previously described.

23.3. Creation of interspecies-specific bispecific single-chain HER2 molecules and

CD3

3.1. Cloning of interspecies specific binding molecules

[0252] In general, bispecific single-chain antibody molecules, where each molecule contains a domain with binding specificity interspecies specific for human and macaque CD3epsilon, as well as domains with binding activity interspecies specific for human and macaque HER2, are designed as described in Table 7:

[0253] The above constructs contain the variable domains of the light chains (L) of the heavy chains.

(H) Interspecies-specific human and macaque HER2 and CD3-specific VH and VL combinations interspecies-specific for human and macaque CD3 were obtained by gene synthesis. Gene synthesis

fragment is designed so that it first has the Kozak site required for eukaryotic expression of the construct, then the 19 amino acid immunoglobulin head peptide, followed in the same reading frame by the coding sequence of the respective bispecific single-chain antibody molecule, and then in the same reading frame, the coding sequence of 6 histidine tags and a stop codon.

Gene fragment synthesis is also designed to introduce appropriate restriction sites at the beginning and end of the fragment. The introduced restriction sites were used in the cloning procedures that follow. Synthesis of the gene fragment was cloned across these restriction sites into a plasmid designated pEFDHFR (pEFDHFR is described in Raum et al. Cancer Immunol Immunother 50 (2001) 141-150) according to standard protocols. The previously described procedures were performed according to standard protocols (Sambrook, Molecular Cloning; A

Laboratory Manual, 3rd edition, Cold Spring Harbor Laboratory Press, Cold Spring Harbor,

New York (2001)).A clone with confirmed nucleotide sequence was transfected into CHO cells lacking DHFR for eukaryotic expression of the construct. Eukaryotic protein expression in DHFR-deficient CHO cells was performed as described by Kaufmann RJ. (1990) Methods Enzvmol. 185, 537-566. Gene amplification of the construct was induced by increasing concentrations of methotrexate (MTX) to a final concentration of 20 nM MTX.

3.2. Expression and purification of a bispecific single-chain antibody molecule

[0254] Bispecific single-chain antibody molecules are expressed in Chinese hamster ovary cells (CHO). Eukaryotic expression of the protein in CHO cells lacking DHFR was performed as described by Kaufmann R. J. (1990) Methods Enzymol. 185, 537-566. Gene amplification of the construct was induced by the addition of increasing concentrations of MTX to a final concentration of 20 nM MTX. After two passages of stationary culture, cells were grown in spinner flasks in nucleoside-free liquid soy medium HyQ PF CHO (with 4.0 mM L-Glutamine with 0.1% Pluronic F - 68; HvCIone) for 7 days before harvesting. Cells were removed by centrifugation and the supernatant containing the expressed protein was stored at -80°C. Transfection was performed with 293fectin reagent (Invitrogen, #12347-019) according to the manufacturer's instructions.

[0255] Akta® Explorer Systems (GE Health Systems) and Unicorn® Software were used for chromatography. Metal chelate chromatography ("IMAC") was performed using Fractogel EMD chelate® (Merck) coated with ZnCl 2 according to the manufacturer's protocol. The column was equilibrated with buffer A (20 mM sodium phosphate buffer pH 7.2, 0.1 M NaCl) and the cell culture supernatant (500 ml) was applied to the column (10 ml) at a flow rate of 3 ml/min. The column was washed with buffer A to remove unbound sample. Bound protein was eluted using a two-step gradient of buffer B (20 mM sodium phosphate buffer pH 7.2, 0.1 M NaCl, 0.5 M Imidazole) as:

Step 1: 20% buffer B in 6 column volumes

Step 2: 100% buffer B in 6 column volumes

[0256] Fractions of the eluted protein from step 2 were collected for further purification. All chemicals were of research grade and purchased from Sigma (Deisenhofen) or Merck (Darmstadt).

[0257] Gel filtration chromatography was performed on a HiLoad 16/60 Superdex 200 prep grade column (GE/Amersham) equilibrated with Equibuffer (25 mM Citrate, 200 mM Lysine, 5% Glycerol, pH 7.2). Eluted protein samples (flow rate 1 ml/min) were subjected to standard SDS-PAGE and Western Blot for detection. Before purification, the column was calibrated for molecular weight determination (molecular weight marker kit, Sigma MW GF-200). Protein concentrations were determined by OD280 nm.

[0258] Purified bispecific single chain antibody protein was analyzed by SDS PAGE under reducing conditions performed on sapre-casf 4-12%Bis Tris gels (Invitrogen). Sample preparation and application were performed according to the manufacturer's protocol. Molecular weight was determined with MultiMark protein standard (Invitrogen). The gel was stained with colloidal Coomassie (Invitrogen protocol). The purity of the isolated protein was >95% as determined by SDS-PAGE.

[0259] ]The bispecific single chain antibody has a molecular weight of about 52 kDa under native conditions as determined by gel filtration in PBS. All constructs were purified according to this method.

[0260] Western Blot was performed using Optitran® BA-S83 membrane and Invitrogen Blot Module according to the manufacturer's protocol. His Tag antibody (Penta His, Oiagen) was used to detect the protein of the bispecific single-chain antibody. Goat-anti-mouse Ig antibody labeled with alkaline phosphatase (AP) (Sigma) was used as the secondary antibody and BCIP/NBT (Sigma) as the substrate. A single band was detected at 52 kD corresponding to the purified bispecific single chain antibody.

23.4. Analysis of the binding of HER2 and CD3 interspecies-specific bispecific antibodies by flow cytometry

[0261] In order to test the functionality of the interspecies-specific Bispecific antibody constructs in terms of binding capacity for human and macaque HER2 and CD3, respectively, FACS analysis was performed. For this purpose CHO cells transfected with human HER2 as described in Example 23.1 and human CD3 positive T cells of the leukemic cell line HPB-ALL (DSMZ, Braunschweig, ACC483) were used to test binding to human antigens. Binding reactivity to macaque antigens was tested using the resulting macaque HER2 transfectant as described in Example 23.2 and the macaque T cell line 4119LnPx (obtained by courtesy of Prof Fickenscher, Hvgiene Institute, Virology, Erlangen-Nuernberg; published in Knappe A, et al., and Fickenscher H., Blood 2000, 95, 3256-61) 200,000 cells of the respective cell population were incubated for 30 min on ice with 50 µl of purified interspecies-specific bispecific antibody construct protein (2 µg/ml). Cells were washed twice in PBS with 2% FCS and binding of the construct was detected with mouse anti-His antibody (Qiagen; diluted 1:20 in 50 µl PBS with 2% FCS). After washing, bound anti-His antibodies were detected with Fc gamma-specific antibodies (Dianova) conjugated to phycoerythrin, diluted 1:100 in PBS with 2% FCS. PBS with 2% FCS was used as a negative control for the binding of T cell lines as well as CHO cells transfected with HER2.

[0262] Flow cytometry was performed on a FACS-Calibur machine, CellOuest software

used for data performance and analysis (Becton Dickinson biosciences, Heidelberg). FACS staining and fluorescence intensity measurements were performed as described in Current Protocols in Immunology (Coligan, Kruisbeek, Margulies, Shevach and Strober, Wiley-lnterscience, 2002).

[0263] Bispecific binding of single chain molecules that are interspecies specific for HER2 and interspecies specific for human and non-chimpanzee CD3 was clearly detected as shown in Figure 23. In FACS analysis, all constructs showed binding to CD3 and HER2 relative to the respective negative controls. Interspecies specificity of bispecific antibodies for human and macaque CD3 and HER2 antigens is shown.

23.5. Bioactivity of HER2iCD3 interspecies-specific bispecific single-stranded

antibodies

[0264] The bioactivity of the obtained bispecific single-chain antibodies was analyzed by chromium 51 (<51>Cr) released in an in vitro cytotoxicity assay using the HER2 positive cell line described in Examples 23.1 and 23.2. Stimulated human CD4/CD56 without PBMC or the 4119LnPx macaque T cell line were used as effector cells as indicated in the respective figures.

[0265] Obtaining stimulated CD4/CD56 deficient PBMC was performed as follows: Petri dish (85 mm diameter, Nunc) was coated with commercially available anti-CD3 specific antibody (eg OKT3, Othoclone) at a final concentration of 1 µg/ml for 1 h at 37°C. Unbound protein was removed by a single wash step with PBS. Fresh PBMCs were isolated from peripheral blood (30 - 50 ml human blood) by Ficoll gradient centrifugation according to standard protocols. 3 - 5 x 107 PBMCs were added to a pre-coated petri dish in 50 ml RPMI 1640 with stabilized glutamine/10% FCS / IL-2 20 U/ml (Proleukin, Chiron) and stimulated for 2 days. On day 3, cells were harvested and washed with RPM11640. IL-2 was added to a final concentration of 20 U/ml and the cells were cultured again for one day in the same cell medium as previously given. By depleting CD4+ T cells and CD56+ NK cells according to standard protocols, CD8+ cytotoxic T lymphocytes (CTLs) were enriched, target cells were washed twice with PBS and labeled with 11.1 MBq<51>Cr in a final volume of 100 ul RPMI with 50% FCS for 45 minutes at 37°C. Subsequently, the labeled target cells were washed 3 times with 5 ml RPMI i.p. as used in the cytotoxicity assay. The assay was performed in a 96-well plate in a total volume of 250 µl supplemented with RPMI (as above) with an E:T ratio of 1:1 or 10:1, as indicated in the respective figures. 1 µg/ml molecule of interspecies-specific bispecific single-chain antibodies and their 15-21 fivefold dilutions were applied. The analysis time is 18 h and cytotoxicity was measured as the relative value of released chromium in the supernatant in relation to the difference between maximal lysis (addition of Triton-X) and spontaneous lysis (without effector cells). All measurements were performed in quadruplicates. Chromium activity measurements in supernatants were performed with a Wizard 3" gamma counter (Perkin Elmer Life Sciences GmbH, Cologne, Germany). Analysis of experimental data was performed with Prism 4 for Windows (version 4.02, GraphPad Software Inc., San Diego, California, USA). Sigmoidal dose-response curves typically have values of R<2>>0.90 as determined by the software. EC50 values were calculated by regression analysis. used to determine bioactivity.

[0266] As shown in Figure 24, interspecies-specific bispecific single-chain antibody constructs show cytotoxic activity on HER2-positive human target cells stimulated with human PBMCs lacking CD4/CD56 and on HER2-positive human target cells induced with the 4119LnPx macaque monkey T cell line.

Example 24: Cloning and expression of human and macaque IgE in the form in which it is bound to

membrane

[0267] Mouse cell line J558L (obtained through Interlab Project, Istituto Nazionale per la Ricerca sul Cancro, Genova, Italy, ECACC 88032902), a myeloma cell line that spontaneously loses heavy chain variants but synthesizes and secretes lambda light chain, was used to complement the membrane-bound human and macaque heavy chain variants, respectively. In order to obtain such constructs, synthetic molecules were obtained by gene synthesis according to standard protocols (the nucleotide sequences of the constructs are now listed as SEQ ID Nos. 507 and 508). In these constructs, the coding sequence for the human and macaque c epsilon chain were fused to the transmembrane region of human IgE, respectively. The built-in specificity of the VH chain is directed towards the hapten (4-hydroxy-3-nitro-phenyl)acetyl) (NP). The synthesis of the gene fragment was designed to first have the Kozak site required for eukaryotic expression of the construct and the immunoglobulin head and restriction sites at the beginning and end of the DNA. The introduced restriction sites EcoRI at the 5" end and Sal at the 3' end were used during the cloning step into an expression plasmid designated as pEFDHFR. After confirmation of the sequence (macaque: XM_001116734 macaca mulatta Ig epsilon C region, mRNA; human: NC_000014 Homo sapiens chromosome 14, complete sequence, National Center for Biotechnology Information, http:// www. ncbi. nlm. nih. gov/entrez) plasmids were used for transfection of CHO/dhfr- cells as described previously. Eukaryotic expression of the protein in DHFR-deficient CHO cells was performed as described by Kaufmann RJ (1990) Methods Enzymol. Gene amplification was induced by increasing concentrations of methotrexate (MTX). MTX.

Example 25: Creation of interspecies-specific bispecific single-chain IgEi molecules

CD3

[0268] In general, bispecific monoclonal antibody molecules, each containing a domain having binding specificity for human and cynomolgus CD3, as well as a domain having binding specificity for human and macaque IgE antigen, are designed as given in Table 8 below:

[0269] The aforementioned constructs containing light chain (L) and heavy chain (H) variable domains interspecies specific for IgE and human and macaque CD3 specific VH and VL combinations interspecies specific for human and macaque CD3 were obtained by gene synthesis. The synthesis of the gene fragment was designed so that it first has the Kozak site required for eukaryotic expression of the construct, then the 19 amino acid immunoglobulin head peptide, followed in the same reading frame by the coding sequence of the respective bispecific monolayer antibody molecule, and then in the same reading frame by the coding sequence of 6 histidine tags and a stop codon. Gene fragment synthesis is also designed to introduce appropriate restriction sites at the beginning and end of the fragment. The introduced restriction sites were used in the following cloning procedures. Synthesis of the gene fragment was cloned through these restriction sites into a plasmid designated pEFDHFR according to standard protocols. A clone with confirmed nucleotide sequence was transferred into DHFR-deficient CHO cells for eukaryotic expression of the construct. Eukaryotic expression of the protein in various CHO cells lacking DHFR was performed as described by Kaufmann RJ. (1990) Methods Enzymol. 185, 537-566. Gene amplification of the construct was induced by increasing concentrations of methotrexate (MTX) to a final concentration of 20 nM MTX. Alternatively, DHFK-deficient CHO cells were transiently transfected with the constructs, according to a standard protocol.

[0270] FACS binding experiments were performed with human IgE transfected J558L cell line to determine human IgE binding capacity. Interspecies specificity for IgE-positive macaque cells was tested by arraying J558L cells transfected with macaque IgE. The same changes in cell lines are applied in the cytotoxicity test performed with IgE and CD3 interspecies specific bispecific single chain antibodies. In addition, tests were performed as described in Examples 4 and 5.

[0271] As indicated in Figure 23, the obtained interspecies specific bispecific single chain IgE and CD3 antibodies show binding to human and cynomolgus antigen and thus prove to be fully interspecies specific.

[0272] As shown in Figure 24, all of the obtained interspecies-specific bispecific single-chain antibody constructs showed cytotoxic activity on IgE positive human target cells induced by human CD8+ cells and IgE positive macaque monkey target cells challenged by the 4119LnPx macaque monkey T cell line. As a negative control, an irrelevant bispecific single-chain antibody was used.

Example 26: Specific binding of scFv clones to the N-terminus of human CD3 epsilon

26.1. Bacterial expression of scFv constructs in E. coliXL1 Blue

[0273] As previously stated, E. coli XL1Blue transformed with pComb3H5Bhis/Flag containing the VL- and VH-segment produce soluble scFv in sufficient amounts after excision of the fragment III gene and induction with 1 mM IPTG. The scFv-chain is exported to the periplasm where it folds into a functional conformation.

[0274] The following scFv clones were selected for this experiment: i) ScFvs 4-10, 3-106, 3-114, 3-148, 4-48, 3-190 and 3-271 as described in WO 2004/106380. ii) ScFvs from human H2C anti-CD3epsilon binding clones. F12Q and 12C as described herein.

[0275] For periplasmic preparations, bacterial cells transformed with appropriate scFv-containing plasmids allowing periplasmic expression were grown on SB-medium supplemented with 20 mM MgCl 2 and carbenicillin 50 µg/ml redissolved in PBS after harvesting. After four rounds of freezing at -70°C and thawing at 37°C, the bacterial outer membrane was disrupted by osmotic shock and soluble periplasmic proteins including scFvs were released into the supernatant. After elimination of intact cells and cell debris, centrifugation, the supernatant containing human anti-human CD3-scFvs was collected and used in further testing. Crude supernatants containing scFv will further determine periplasmic preparations (PPP).

26.2. Binding of scFvs to (aa 1-27)- Fc fusion protein human CD3 epsilon

[0276] ELISA assays were performed by coating 96-well plastic plates (Nunc, maxisorb) with (aa 1-27)-Fc fusion protein human CD3 epsilon, usually at 4°C overnight. The coating antigen solution was then removed, the slides were washed once with PBS/0.05% Tween 20 and then blocked with PBS/3% BSA for at least one hour. After removal of the blocking solution, solutions of PPPs and controls were added to the wells and incubated typically for one hour at room temperature. The wells were then washed three times with PBS/0.05% Tvveen 20. Detection of scFvs bound to the immobilized antigen was performed using Biotin-labeled anti FLAG-tag antibody (M2 anti Flag-Bio, Sigma, usually at a final concentration of 1 µg/ml PBS) and detected with Streptavidin labeled peroxidase (Dianova, 1 µg/ml PBS). The signal was developed by the addition of ABTS substrate solution and measured at a wavelength of 405 nm. Non-specific binding of test samples to blocking agents and/or human IgG1 portion CD3 epsilon (aa 1-27)-human Fc fusion protein was examined by performing an identical assay with identical reagents and an identical time interval on ELISA plates coated with human IgG1 (Sigma). PBS was used as a negative control.

[0277] As shown in Figure 25, scFvs H2C, F12Q and I2C show strong binding signals to human CD3 epsilon (aa 1-27)-Fc fusion protein. Human scFvs 3-106, 3-114, 3-148, 3-190, 3-271, 4-10 and 4-48 (as described in WO 2004/106380) do not show any significant binding above negative control levels.

[0278] TO avoid the possibility of positive binding of scFvs H2C, F12Q and 12C to the wells coated with human CD3 epsilon (aa 1-27)- Fc fusion protein that may occur due to binding to BSA (used as a blocking agent) and/or human lgG1 Fc-gamma-portion of human CD3 epsilon (aa 1-27)- Fc fusion protein, another was performed in parallel ELISA test. In the second ELISA, all parameters are identical to those in the first ELISA, except that in the second ELISA human IgG1 (Sigma) is coated instead of human CD3 epsilon (aa 1-27)-Fc fusion protein. As shown in Figure 26, none of the scFvs tested showed significant binding to BSA and/or human lgG1 above the noise level. Taken together, these results allow us to conclude that scFvs4-10, 3-271, 3-148, 3-190, 4-48, 3-106 and 3-114 do not bind specifically to the (aa 1-27)-region of human CD3 epsilon, while scFvs H2C, F12Q and I2C clearly show specific binding to the N-terminal 27 amino acids. CD3 epsilon human.

REFERENCES CITED IN THE DESCRIPTION

This list of references cited by the Applicant is for the reader only. He is not part of this

of the European patent, easy references were collected with great care, it is impossible to exclude

errors or omissions, and EPO disclaims all liability in this regard.

Patent literature cited in the description

•WO2007033230Ar00071• VVO9420627A f00971 •VVO9954440Ar0013lf00141r00511f01161• VV09429469Af00991•WO04106380A r0013l 100511 f00511•VVO9700957A f00991•US4946778Af00421•US5580859A r00991• WO2004106380Af00501T02741TO2771• US5589466AT00991• EP623679B1f00831

Claims (18)

1. A polypeptide containing a first binding domain, which is an antibody capable of binding to an epitope of the CD3e chain of human and Callithrix jacchus, Saguinus oedipus or Saimiri sciureus, characterized in that this epitope is part of the amino acid sequence contained in the group formed by SEQ ID No: 2, 4, 6, or 8 and contains at least the amino acid sequence Gln-Asp-Gly-Asn-Glu, and a second binding domain that is able to bind to EGFR, Her2/neu or human and/or non-chimpanzee primate IgE 2. Polypeptide as defined in claim 1, characterized in that the epitope is part of the amino acid sequence contained in the group consisting of SEQ ID NOs:2, 4, 6, and 8 and contains at least one amino acid sequence Gln-Asp-Gly-Asn-Glu. 3. The polypeptide of claim 1 or 2, characterized in that the first binding domain comprises a VL region comprising CDR-L1, CDR-L2 and CDR-L3 selected from: (a) CDR-L1 as in SEQ ID No: 27, CDR-L2 as in SEQ ID No: 28 and CDR-L3 as in SEQ ID No: 29; (b) CDR-L1 as in SEQ ID No:117, CDR-L2 as in SEQ ID No:118 and CDR-L3 as in SEQ ID No:119; and (c) CDR-L1 as in SEQ ID NO:153, CDR-L2 as in SEQ ID NO:154, and CDR-L3 as in SEQ ID NO:155. 4. Polypeptide according to one of claims 1 or 2, characterized in that the first binding domain contains a VL region comprising CDR-HL, CDR-H2 and CDR-H3 selected from: (a) CDR-H 1 as in SEQ ID NO:12, CDR-H2 as in SEQ ID NO:13 and CDR-H3 as in SEQ ID NO:14; (b) CDR-H1 as in SEQ ID NO:30, CDR-H2 as in SEQ ID NO:31 and CDR-H3 as in SEQ ID NO:32; (c) CDR-H1 as in SEQ ID NO:48, CDR-H2 as in SEQ ID NO:49 and CDR-H3 as in SEQ. ID NO:50; (d) CDR-H 1 as in SEQ ID NO:66, CDR-H2 as in SEQ ID NO:67 and CDR-H3 as in SEQ ID NO:68; (e) CDR-H 1 as in SEQ ID NO:84, CDR-H2 as in SEQ ID NO:85 and CDR-H3 as in SEQ ID NO:86; (f) CDR-H 1 as in SEQ ID NO: 102, CDR-H2 as in SEQ ID NO: 103 and CDR-H3 as in SEQ ID NO: 104; (g) CDR-H 1 as in SEQ ID NO:120, CDR-H2 as in SEQ ID NO:121 and CDR-H3 as in SEQ ID NO:122; (h) CDR-H 1 as in SEQ ID NO:138, CDR-H2 as in SEQ ID NO:139 and CDR-H3 as in SEQ ID NO:140; (i) CDR-H 1 as in SEQ ID NO: 156, CDR-H2 as in SEQ ID NO: 157 and CDR-H3 as in SEQ ID NO: 158; and (j) CDR-H 1 as in SEQ ID NO:174, CDR-H2 as in SEQ ID NO:175 and CDR-H3 as in SEQ ID NO:176. 5. The polypeptide according to one of claims 1 to 3, wherein the first binding domain comprises a VL region selected from the group consisting of a VL region as in SEQ ID NO:35, 39, 125, 129,161 or 165. 6. The polypeptide according to one of claims 1 to 4, wherein the first binding domain comprises a VH region selected from the group consisting of VH regions as in SEQ ID NO: 15, 19, 33, 37, 51, 55, 69, 73, 87, 91, 105, 109, 123, 127, 141, 145, 159, 163, 177 or 181. 7. The polypeptide according to one of claims 1 to 6, wherein the first binding domain comprises a VL region and a VH region selected from the group consisting of: (a) a VL region as in SEQ ID NO:17 or 21 and a VH region as in SEQ ID NO:15 or 19; (b) a VL region as in SEQ ID NO:35 or 39 and a VH region as in SEQ ID NO:33 or 37; (c) a VL region as in SEQ ID NO:53 or 57 and a VH region as in SEQ ID NO:51 or 55; (d) a VL region as in SEQ ID NO:71 or 75 and a VH region as in SEQ ID NO:69 or 73; (e) the VL region as in SEQ ID NO:89 or 93 and the VH region as in SEPA ID NO:87 or 91; (f) a VL region as in SEQ ID NO:107 or 111 and a VH region as in SEQ ID NO:105 or 109; (g) a VL region as in SEQ ID NO:125 or 129 and a VH region as in SEQ ID NO:123 or 127; (h) a VL region as in SEQ ID NO:143 or 147 and a VH region as in SEQ ID NO:141 or 145; (i) a VL region as in SEQ ID NO:161 or 165 and a VH region as in SEQ ID NO:159 or 163; and (j) a VL region as depicted in SEQ ID NO:179 or 183 and a VH region as depicted in SEQ ID NO:177 or 181. 8. The polypeptide according to claim 7, in which the first binding domain contains an amino acid sequence selected from the group consisting of SEQ ID NO:23, 25, 41, 43, 59, 61, 77, 79, 95, 97, 113, 115, 131, 133, 149, 151, 167, 169, 185 or 187. 9. The polypeptide according to one of claims 1 to 8, wherein said polypeptide is a bispecific single-chain antibody molecule. 10. The polypeptide according to claim 9, in which the bispecific single-chain antibody molecule contains a group of the following sequences as CDR H1, CDR H2, CDR H3, CDR L1, CDR L2 and CDR L3 in the second binding domain selected from SEQ ID NOs:441 - 446, SEQ ID NOs:453 - 458, SEQ ID NOs:463 - 468, SEQ ID NOs:481 - 486. 11. The polypeptide of claim 9, wherein the bispecific single chain antibody molecule comprises a sequence selected from: (a) an amino acid sequence as set forth in SEQ ID NOs:389, 391, 393, 395, 397, 399, 409, 411, 413, 415, 417, 419, 429, 431, 433, 435, 437, 439, 447, 449, 451, 469, 471, 473, 475, 477, 479, 495, 497, 499, 501, 503 and 505; and (b) an amino acid sequence encoded by a nucleic acid sequence as set forth in one of SEQ ID NOs:390, 392, 394, 396, 398, 400, 410, 412, 414, 416, 418, 420, 430, 432, 434, 436, 438, 440, 448, 450, 452, 470, 472, 474, 476, 478, 480, 496, 498, 500, 502, 504 and 506. 12. A nucleic acid sequence encoding a polypeptide as defined in one of claims 1 to 11. 13. A vector comprising a nucleic acid sequence as defined in claim 12. 14. A host cell transformed or transfected with the vector defined in claim 13 15. A process for the production of a polypeptide according to any one of claims 1 to 11, wherein said process comprises culturing the host cells defined in claim 14, under conditions that allow expression of the polypeptides defined in any one of claims 1 to 11 and recovering the produced polypeptide from the culture 16. A pharmaceutical composition containing a polypeptide according to one of claims 1 to 11, or which is produced according to the process of claim 15. 17. A polypeptide according to one of claims 1 to 11, or produced according to the process of claim 15 for use in the prevention, treatment or mitigation of a disease selected from proliferative disease, tumor disease or immunological disorders. 18. A kit comprising a polypeptide as defined in one of claims 1 to 11, a nucleic acid molecule as defined in claim 12, a vector as defined in claim 13, or a host cell as defined in claim 14.